<%BANNER%>

Best Management Practice Development with the Ceres-Maize Model for Sweet Corn Production in North Florida

Permanent Link: http://ufdc.ufl.edu/UFE0022077/00001

Material Information

Title: Best Management Practice Development with the Ceres-Maize Model for Sweet Corn Production in North Florida
Physical Description: 1 online resource (329 p.)
Language: english
Creator: He, Jianqiang
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: bmp, ceres, corn, dssat, glue, irrigation, nitrogen, parameter, sensitivity, uncertainty
Agricultural and Biological Engineering -- Dissertations, Academic -- UF
Genre: Agricultural and Biological Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Increasing nitrogen loads within the Suwannee River Basin of North Florida has become a major concern. Nitrogen fertilizer application in field crop production is proved to be the most import nitrogen contribution in this region. Florida ranks highest in the nation in the production and value of fresh market sweet corn. Thus it is necessary to develop research based nitrogen best management practices (N-BMPs) to reduce nitrogen leaching while keeping an acceptable yield in sweet corn production. This study is an attempt to utilize the CERES-Maize mode of the Decision Support System for Agrotechnology Transfer (DSSAT) model as a platform to develop potential BMPs for sweet corn production in North Florida. The results show that the non-restricted and restricted one-at-a-time (OAT) method can be used to conduct global sensitivity analysis for the CERES-Maize so as to select the most influential parameters for model calibration. The generalized likelihood uncertainty estimation (GLUE) method was proved to be a powerful tool for model parameter estimation, since the uncertainties in model input parameters were significantly reduced after GLUE was used to estimate the model input parameters. The uncertainties in model outputs were reduced correspondingly. The comparison between the model simulated and field observed results of the seven treatments in a field plot experiment of sweet corn in 2006, shows that the model did a good job in predicting dry yield and phenology dates. The results of BMP development with the calibrated CERES-Maize model show that if the growers could apply both irrigation water and nitrogen fertilizer more frequently but with smaller amounts in each application, this would result in an acceptable yield and a lower level of nitrogen leaching. The results showed a total nitrogen amount between 196 and 224 kg N ha-1 would be enough for sweet corn production in North Florida, which confirmed that the recommendation nitrogen amount (224 kg N ha-1) by Institute of Food and Agricultural Sciences ?IFAS?, Univerisity of Florida, was reasonable. The results of uncertainty analysis of the CERES-Maize model for sweet corn simulation show that the weather was the dominant uncertainty contributor. This was because after two rounds of GLUE parameter estimation procedure, the uncertainties existing in input parameters were minimized.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jianqiang He.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Dukes, Michael D.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022077:00001

Permanent Link: http://ufdc.ufl.edu/UFE0022077/00001

Material Information

Title: Best Management Practice Development with the Ceres-Maize Model for Sweet Corn Production in North Florida
Physical Description: 1 online resource (329 p.)
Language: english
Creator: He, Jianqiang
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: bmp, ceres, corn, dssat, glue, irrigation, nitrogen, parameter, sensitivity, uncertainty
Agricultural and Biological Engineering -- Dissertations, Academic -- UF
Genre: Agricultural and Biological Engineering thesis, Ph.D.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Increasing nitrogen loads within the Suwannee River Basin of North Florida has become a major concern. Nitrogen fertilizer application in field crop production is proved to be the most import nitrogen contribution in this region. Florida ranks highest in the nation in the production and value of fresh market sweet corn. Thus it is necessary to develop research based nitrogen best management practices (N-BMPs) to reduce nitrogen leaching while keeping an acceptable yield in sweet corn production. This study is an attempt to utilize the CERES-Maize mode of the Decision Support System for Agrotechnology Transfer (DSSAT) model as a platform to develop potential BMPs for sweet corn production in North Florida. The results show that the non-restricted and restricted one-at-a-time (OAT) method can be used to conduct global sensitivity analysis for the CERES-Maize so as to select the most influential parameters for model calibration. The generalized likelihood uncertainty estimation (GLUE) method was proved to be a powerful tool for model parameter estimation, since the uncertainties in model input parameters were significantly reduced after GLUE was used to estimate the model input parameters. The uncertainties in model outputs were reduced correspondingly. The comparison between the model simulated and field observed results of the seven treatments in a field plot experiment of sweet corn in 2006, shows that the model did a good job in predicting dry yield and phenology dates. The results of BMP development with the calibrated CERES-Maize model show that if the growers could apply both irrigation water and nitrogen fertilizer more frequently but with smaller amounts in each application, this would result in an acceptable yield and a lower level of nitrogen leaching. The results showed a total nitrogen amount between 196 and 224 kg N ha-1 would be enough for sweet corn production in North Florida, which confirmed that the recommendation nitrogen amount (224 kg N ha-1) by Institute of Food and Agricultural Sciences ?IFAS?, Univerisity of Florida, was reasonable. The results of uncertainty analysis of the CERES-Maize model for sweet corn simulation show that the weather was the dominant uncertainty contributor. This was because after two rounds of GLUE parameter estimation procedure, the uncertainties existing in input parameters were minimized.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Jianqiang He.
Thesis: Thesis (Ph.D.)--University of Florida, 2008.
Local: Adviser: Dukes, Michael D.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0022077:00001


This item has the following downloads:


Full Text
xml version 1.0 encoding UTF-8
REPORT xmlns http:www.fcla.edudlsmddaitss xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.fcla.edudlsmddaitssdaitssReport.xsd
INGEST IEID E20110331_AAAAEY INGEST_TIME 2011-04-01T03:18:50Z PACKAGE UFE0022077_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 6495 DFID F20110331_AADLFN ORIGIN DEPOSITOR PATH he_j_Page_160thm.jpg GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
0d4d59360ced77c6d6e6fd32ba3b8ce0
SHA-1
f28484b55efde62ea363bb975bc684433dd8af6f
106631 F20110331_AADKIU he_j_Page_184.jpg
0b4ff2da02155df97b83608d7503c492
a29c902da26a29389011d6250c0d4e7f9109329a
115380 F20110331_AADHXH he_j_Page_149.jp2
e18eef0f76f98b85730739bcb867d562
5dd295e2f1c83c73736a440912b1beeea8dbe940
2009 F20110331_AADIBX he_j_Page_240.txt
394c34ea49c7af986a27cc1b8f0cdffc
91008e2593e3ecaadaff22d47ed3a2ac83e0bcb0
35237 F20110331_AADIUA he_j_Page_188.QC.jpg
f80fd70f3910f636beb7b0b383e8d534
106df4d2ac6a3474c369ad98ae268241981e15d4
3624 F20110331_AADLFO he_j_Page_161thm.jpg
4969369be6029f5858018ab2e8d57383
7160748fd81a77d45b41d4aacf9ff429af855fbb
109137 F20110331_AADKIV he_j_Page_185.jpg
c5fe9ab68f938a439a30dfe0db56a934
9d77c096b74d573d13fcdb072796f612f56b50e3
20699 F20110331_AADHXI he_j_Page_306.QC.jpg
e8f38b483a1e3d73f9e4876434835975
083381e6e45959d672473cd83bec906bfaa473fe
117657 F20110331_AADIBY he_j_Page_220.jp2
34be45c0c5d69df32d27059e3e1af431
2e2d89d9854625d6e1bd33a80fefa2e2ab6262d1
58330 F20110331_AADIUB he_j_Page_258.jpg
a0fede8463404655f9af34953638557f
335e9a0e77920171ea03d1e8be462a3b7bea9d99
4490 F20110331_AADLFP he_j_Page_164thm.jpg
edce48b63e77cfca40d4ee1e034c6d7a
3641081de54814b451b50208cd9f520059dfb817
97131 F20110331_AADKIW he_j_Page_186.jpg
231aa9e79729dafd93692de87fc09737
01afdd6c3f2380b3686dd075bcbf58e6be9fbf33
1053954 F20110331_AADHXJ he_j_Page_034.tif
d68d2583d6a17b01872f37fe0c23f345
3e1f0ec71448827e2ad5f8a1fe9834203c6f4456
41297 F20110331_AADIBZ he_j_Page_019.pro
c156f2ecd497346b196479d6d8c65c1d
05c2643663320fc42b28156973ac06ed77aa4288
34123 F20110331_AADIUC he_j_Page_186.QC.jpg
ebc9e495e8cbb7c8cedf67f784ed66cb
c740388bc06ee6f24c1abebe959e1d95acaa5505
6603 F20110331_AADLFQ he_j_Page_165thm.jpg
31a7ace60ca5b834f3b766e03d705b43
eb91211aa8be89d1485bfd4e12dac91e4e4e754a
100527 F20110331_AADKIX he_j_Page_187.jpg
269903cbd6bd990c47cbfc29ef759e7c
0991dd725efef61f75cbd4711b62a5563610e7c0
2823 F20110331_AADHXK he_j_Page_326.txt
e5b7fa58b93f56a76b6cfab106942ab4
7600f2fb7f5e20d167d26358029342688c2234b5
110239 F20110331_AADIUD he_j_Page_045.jp2
a0bcff3230aca050eef32bcb16b8830a
52b448500366470ed99b2d6de709bc80bcb57891
8349 F20110331_AADLFR he_j_Page_172thm.jpg
02f29c59d3c36bc89a0df5f683c45454
31fbe4e3d126dfc5871aa70a58f2385a8c670947
107500 F20110331_AADKIY he_j_Page_188.jpg
fdd86ac5aaf7f573a553328996772155
167de350182799de0faed4f97805a2b35ccf57d3
1746 F20110331_AADHXL he_j_Page_089.txt
9c916d004c2fd2b3d8795ed6eec65f21
7cceb66b5a0fcd621a388c531fc7bd6aa80ac1b0
29164 F20110331_AADIUE he_j_Page_062.pro
d563a88aa4c733102c80cad7a76dd796
82b40e305cd4fe7d06fd3816bf09393b12696381
7824 F20110331_AADLFS he_j_Page_173thm.jpg
5c76cdf9353aef12a735b174c4fdc2fd
31882f64849a647f6bce4d63d45afbf3b430353f
107676 F20110331_AADKIZ he_j_Page_189.jpg
11b70fd9fb619631bd19a6fd7418d15c
a6ce7513963ac518d5d36dcc87188978dd6315de
112200 F20110331_AADHXM he_j_Page_193.jp2
1b95ff7c84ebdd699939345c270d4fe2
ce3cc6fac857fc9e1c9dffebcd7bb53b4f198416
114875 F20110331_AADIUF he_j_Page_253.jp2
5091c3475dcc9e250db8a38234ef8ef4
a3e040cc71d3a0f614e604728a4512570f1a8839
7880 F20110331_AADLFT he_j_Page_174thm.jpg
3917f2ada92d9ed8c5d1a46bc1d2ec76
80fc17f4d475d5a00f3e21378d9ef8ec0937d68f
109045 F20110331_AADHXN he_j_Page_139.jp2
c07105b6d249f87bb61f94038e2efed0
c2731567505d138f5eee0b7506e2de9b4ba54325
153 F20110331_AADIUG he_j_Page_302.txt
c0616390cc53b8b0af3f2c03aa9f608c
0f83f59728d92ff07675a8656a0cb17de4f02b1a
9077 F20110331_AADLFU he_j_Page_176thm.jpg
ac0e2e2ca9e826b4e94f11ebd67b8708
b9dcdac6bbb5e3e63d1936460be9f1b7e6fed6a8
1923 F20110331_AADJRA he_j_Page_090.txt
307bc4bdafff35e0cb6c9f5389fa6a91
a248cca13c5bb219cd79d09a1e9b9e817dd26523
35323 F20110331_AADHXO he_j_Page_007.QC.jpg
4752c2cf0e8a728e627600671a170d6c
9af8f288f1eb371bc912e3c64723b6e7c7bf0419
19954 F20110331_AADIUH he_j_Page_288.QC.jpg
a5c4f84cf45e52999088117a1457e1e2
9aee444ee36299470b263c150779ecb7731847fa
9172 F20110331_AADLFV he_j_Page_178thm.jpg
c4da9f4c86b38bafec43c210b1b571aa
7e64800b65e0b7c21c1826cc9ea6025867244be9
1984 F20110331_AADJRB he_j_Page_091.txt
2b9eea05bf467fc1f773c08f799bd1b3
0814893e32f5b3f5931c08510ec960d85dca2a42
18131 F20110331_AADHXP he_j_Page_310.QC.jpg
9fc521fc2ac8bc09bb7361f80d232f05
6a2c46192177f5ce25f4f7495d0e95108790f7a0
287476 F20110331_AADIUI he_j_Page_037.jp2
9d015bc765b78a51f3f0a25fa150537c
507ce3fddf06f32a3b81bd89157c8445c0d8d4f5
8591 F20110331_AADLFW he_j_Page_182thm.jpg
e20f1cf6df246fca052392e844aa364e
8a4cfab273e1a944e96274fe0449a1e1df317f63
1578 F20110331_AADJRC he_j_Page_092.txt
99b455fc26622e4afd1091dc8f339fa4
d016db24c4da14c02d339bbd9f267a9a3977b5f4
8535 F20110331_AADHXQ he_j_Page_001.QC.jpg
9578ff2cbc059a4bae4774454d889582
eec107cd0dcf4465444e4cb67a9b03a4221c2b69
120339 F20110331_AADIUJ he_j_Page_007.pro
31954d5dde60ee07fdf8625f20a7eaef
34b9a0da87d69915f42174e36c87821072fdfb60
8917 F20110331_AADLFX he_j_Page_185thm.jpg
2c72a6676eb23ca4c8e5652981ecb856
d6bcd17a1040037b396c1f3d3ac3852dea9f7346
1896 F20110331_AADJRD he_j_Page_093.txt
f61c70f00f196181da71c1a85de4b9e4
4d6c09db3ffd2e6dbae0e5bd594770963c3ae499
1910 F20110331_AADHXR he_j_Page_099.txt
4a35cd25a723aef4175a31125c62f79e
36656e35b418c337ee692be015de060f457e4d61
111247 F20110331_AADIUK he_j_Page_169.jp2
b60f98c4ab05412edbc4fc1ae28126d7
ddb84224e971f3e5bbef19c2e61a5a6c09349bb4
8253 F20110331_AADLFY he_j_Page_186thm.jpg
9785208581b8ab584b4bece46724c751
c3d211544237771ed8ae65ba1b286546d0127c35
2241 F20110331_AADJRE he_j_Page_097.txt
b40131624ce94c4531b4367ec70ea6ba
dc6ad2c52ddc774df9e8473736b99b4c8ecae5f3
2136 F20110331_AADIHA he_j_Page_026.txt
b9c7d121c506039644a491f2eb6b238e
a9b9b8a016a62cbd0876ef416ebd5db88c22bc7f
112968 F20110331_AADHXS he_j_Page_188.jp2
579d30bd82f30a1c67e6c27a157646fa
2afa027b5cea2ab3c4eb57605dda232f107a48df
8551 F20110331_AADIUL he_j_Page_216thm.jpg
9dab15849b14342ba2fff2677c875824
6d8a57891413dca7b04c3335ca67e662dd04da18
8879 F20110331_AADLFZ he_j_Page_188thm.jpg
74050c52fcf2a47d5d1bda3c9a819144
3c1e5a5d11aedd72aeac3315bfb725a092ca0216
2042 F20110331_AADJRF he_j_Page_100.txt
92ba10a9b41a87912301a5c311a488a4
4deb6d00df48b64ce308d49bd162ceea5ad895b6
31391 F20110331_AADIHB he_j_Page_095.QC.jpg
5fb94fe988dd3c3fde10a013a6e34f5a
934c7855cd769794a15a3a4f4196f5332b6dffb0
3443 F20110331_AADHXT he_j_Page_271thm.jpg
6c600253bd9b55ff8e678b51f32192ba
3dfdf837c18fc90824321a1a089d55fbbdad35dc
97974 F20110331_AADIUM he_j_Page_039.jpg
05409c2b2db6cde7c0065217295b7a5b
4f398fe0783d01dec95e4d15cbaea4ca996a44f5
2223 F20110331_AADJRG he_j_Page_101.txt
1685bed9ae6d1049a7f0c46f0ee84dc0
a01ae64a0e7f2223513af0975a1abd78027ee571
9188 F20110331_AADIHC he_j_Page_299thm.jpg
50fb089e8cb9c850eb58607b24975e9c
06e9ee034b1cf7e817e32970c86a002ddc4de743
31281 F20110331_AADHXU he_j_Page_047.QC.jpg
28628a9acea1a36a99483d28d7085999
8f8f6e5439831cb1e2829b48ae61b268ffce74d0
36724 F20110331_AADIUN he_j_Page_267.pro
e813fb787a156dbc7903fb6d80896c16
813c56b26da73a06c3e5716fb015bc04318185a7
113777 F20110331_AADKOA he_j_Page_102.jp2
401c98b51973575538430731f0620d63
ca68f0e997b81e517b6377ba91f11b505c02a12b
2063 F20110331_AADJRH he_j_Page_104.txt
a15887501231286c092c7ddc560b8bc3
753c034f915ad6b3ed00fb01eddf26fe179c753e
6108 F20110331_AADIHD he_j_Page_156.pro
82a182cd336a2d5e73d791f69c53a3a3
1da997f4faa49b8b7c00766fe7ac1af7aeebd68e
F20110331_AADHXV he_j_Page_277.tif
caefceecd1eae6350525a836873930b2
5a6ac4dde6d10d855845566dcca504121135c1a6
72892 F20110331_AADIUO he_j_Page_257.jp2
03dc645d3037b1ca468fa56cb6611c1a
324ddf6843b99acdbb26768b49c1e556647526de
111211 F20110331_AADKOB he_j_Page_103.jp2
6b26039cdbe621b9efa6f01af6a9bd8c
e9a588bc5fa2311692087e64e8412386c82eaf24
1999 F20110331_AADJRI he_j_Page_105.txt
68227630a3773c71f58b680305d29004
394b14f81ed5ce5695ee8d54d9316e5fe8501381
89820 F20110331_AADIHE he_j_Page_090.jp2
90cab8b288cc0f8f5f6e1ac544593ec2
549ed4d060b6ad227db653f274ceee5054bf65bc
8701 F20110331_AADHXW he_j_Page_081thm.jpg
f3a279db3086f881a7aa8a0df630655d
c52f62b702045cb5ae90ad2890e7355d79b956db
25271604 F20110331_AADIUP he_j_Page_320.tif
f372117898ed85bd2946bd0de3a445d2
755f573caa5d62042239da993afd475a7468262b
113526 F20110331_AADKOC he_j_Page_104.jp2
e4c2ccce2ea8c6552b4624412ac16943
8d3782e0d1b55bb8d86f776a76ae47d9a5ece9a2
F20110331_AADJRJ he_j_Page_106.txt
223d5e07e01d2b49d5180c25193a9c08
c24b80b8760cfeddfa642c9a3f33faf829db284a
1051864 F20110331_AADIHF he_j_Page_298.jp2
15030b2b5670989976e6f42640d607f3
5287f9a5fe2e1d90c8819a01f03078a0b0046ceb
F20110331_AADHXX he_j_Page_282.tif
b005b5e79c225e19af74eadc80399ae4
921c2e771ca702529dfe5a35fd37e3369b517966
54100 F20110331_AADIUQ he_j_Page_046.pro
c2d157654749de5bd1e8903c61a5f0f0
a3fb00a2ce79bccd35380b3b30d3361db62d8919
107625 F20110331_AADKOD he_j_Page_105.jp2
605976281cea720af663295ebabbb1c2
80ad86eedd00cd4ba9093be7d378c270e3599646
1359 F20110331_AADJRK he_j_Page_108.txt
3e767d9e19a46a0d4315cd9b4372f2fa
a61073ddcdfb9cfab855589c6a81834a71561a58
2031 F20110331_AADIHG he_j_Page_003.QC.jpg
b2841195ea347041e7560f0db1b496fa
b60143db768a888dddc27bd1f8768be96d9e7dc6
56799 F20110331_AADHXY he_j_Page_283.jpg
f12399a51523eaceeca1a186e930da8a
31a17886585e175e41e04e9ffe6e255cc765752b
79645 F20110331_AADIUR he_j_Page_167.jp2
31c8c4b25bc8470274f4623ec12c8de0
e6cd270c61cb122a4497a007f1fce5297a6d6e41
113822 F20110331_AADKOE he_j_Page_106.jp2
9161cbbff36e6f14c1790121c9655e0b
385559b8b335d1a8dd06d463f71784fbfdff7fc9
674 F20110331_AADJRL he_j_Page_109.txt
492aa77050fbefd48770e3b1b9bbad6e
9a782a6fa1501596ff414e62d85f65ff5348813b
2214 F20110331_AADIHH he_j_Page_098.txt
d68d8ad3419c626c3ed6ed6ce611acb7
c108372c056bfcbd1fc88b874e99b82d3bef1fe2
4878 F20110331_AADJEA he_j_Page_285thm.jpg
eb40f403783b10b80e25bb326c2be072
55dba3e349d72a3a60e14b70b5129b9023a37298
109751 F20110331_AADIUS he_j_Page_190.jp2
8a20b391a010a3dbed9e035ccac3b24e
b5bbdf4f883e861bd390d98187fa4910ec6ca7f2
539068 F20110331_AADKOF he_j_Page_109.jp2
42450c536ffa2bbdca80d22f27985906
f47fe4b86a1d4c61eba3f2f06d3af25e85353072
1073 F20110331_AADJRM he_j_Page_113.txt
e167c5a81131dc7aca73bb09035e2ce8
397ca3217584b57f4a36d685a66b4668362d678a
162333 F20110331_AADHXZ he_j_Page_008.jpg
61f47df773b2229ee289c9fd7f745725
37face7b112468b4ed3f500cc03f0eddee662430
F20110331_AADJEB he_j_Page_309.tif
2e94d948211be3cfdabc9f231adce0c3
324fb14db2769bd8359af74f80d4e9a781e86718
26767 F20110331_AADIUT he_j_Page_010.pro
db17aec9c9143172d7b2eafc5cdb5d8d
aa6b398a7d2f6b1954615b7ca17e3a7a0d3b1e6a
460126 F20110331_AADKOG he_j_Page_113.jp2
cbe31009b10741c6fb098b826f44bc19
8be41911168a8dfc40bbb1625345f00e503361df
1378 F20110331_AADJRN he_j_Page_114.txt
df7acbeb086b2d18224f4e9a0c19aab4
36ccb225a9cc77bcc0a2611f3c6b8d1eb2034057
5707 F20110331_AADIHI he_j_Page_205thm.jpg
6bf4806349becfdea8b72a3101430d56
c3278de6e655a1af31a8787d58558e6e5152a040
36189 F20110331_AADJEC he_j_Page_073.QC.jpg
bb91dd30fc59791e6499b530af50c9f7
3da90dcb9d0613e334b0a84b36e9db98ed3d198b
4216 F20110331_AADIUU he_j_Page_328thm.jpg
e8a5f73db37a71c39d63a5d37b5f410d
f46dbeafdc1b0719d21b271131618ff3171aedfd
467601 F20110331_AADKOH he_j_Page_114.jp2
2d34e9c4e8d5b7d130f82d80276afa73
0982674be4894462116a7a9a0a1b2bb4384597be
1485 F20110331_AADJRO he_j_Page_116.txt
7ca21b163982bfc13a37db958b6b5e99
3d03caa87c063d685a0c2cd640135f6c3b3ca5ee
130913 F20110331_AADIHJ he_j_Page_325.jp2
a04768b8c1774101e45af92b5f835443
760f34a56266f8154b0c7246c5cef55af5efda64
106759 F20110331_AADJED he_j_Page_297.jpg
96978ff98598e4fb6d680a16d189897e
b7b5a6f582d8af90939946ed911f262c9422b8dc
617089 F20110331_AADKOI he_j_Page_118.jp2
9fe206fac6fe3518943f94131d6d73fa
aca04e22556961308564359dadc08af3a268243f
121473 F20110331_AADIHK he_j_Page_017.jpg
d8a8839794056d21ef4019924a00ea01
cfa3cb2b09ec38219799fa9dbed05ca9263bdd3a
4616 F20110331_AADIUV he_j_Page_275thm.jpg
48423b37810a52c516fac88aca277004
ca64d1e864bc0b9df30b279ead4e6f443ec31531
1540 F20110331_AADJRP he_j_Page_117.txt
bec808f261b28cc3a80a5d7d63bd4629
ae725d4423cd47040fd37d681614460fccc4981f
1051732 F20110331_AADKOJ he_j_Page_120.jp2
96e9cc3cac1a89c2a45a23f5882c0034
5973303ae9edc927515ab03044aa658ef80c2dbf
891 F20110331_AADIHL he_j_Page_112.txt
0c22dd9c74396f3fd96fb12900133239
67c0ad486881508a2298658331c8033341d05af1
2657 F20110331_AADJEE he_j_Page_166.txt
59f441d5f20c8ddd4a0821d152c5bfa4
afdf05efabbd68067a2b9844684d63efce80847b
44771 F20110331_AADIUW he_j_Page_153.pro
ffcd29530cb704a0b0d351d1e9a4f8a4
5a824cfdfff65dcc0bf79ce95ff2e52d39883a4a
1082 F20110331_AADJRQ he_j_Page_118.txt
c6a39d40fca41b257c4601f89b7a1bd2
ff5726803db0027fa290b50bf8374fa58ba2f8fd
103200 F20110331_AADKOK he_j_Page_122.jp2
86cf9454e87a4b5b62106b539ca3843a
52d27db3770267b3c6bb326b1d7595e8b681b035
F20110331_AADIHM he_j_Page_114.tif
b8e257ff09136bc322c5200fac21a494
8017abb5be226cae0fe32cc1225c131d91684a58
F20110331_AADJEF he_j_Page_310.tif
ef2d2ce1622eae74c1c6f0a980542915
be610c340717b83a6959ec8828db2969da784afa
114642 F20110331_AADIUX he_j_Page_073.jp2
e8a0c7b03afd133849e9a14c19f092fd
3650c6af3d08989299dca3522c3d4c6526ee1560
508101 F20110331_AADKOL he_j_Page_124.jp2
a2629b1e3a1c70b84d5248c77ae09dbb
5f8cfad212f923c96e3c51115be9de7edeccc583
61555 F20110331_AADIHN he_j_Page_273.jpg
8a84a9bc3af05e1c53beb1d8779fcee6
86a46fd6fb2cf2da499fa815682129dc49072436
51469 F20110331_AADJEG he_j_Page_261.jpg
f2c86e39b79028ebc2399a4e5a289902
1638497ad1a608a6fcea7c1f24cd9c792ec38360
14744 F20110331_AADIUY he_j_Page_197.pro
b6180acf1773b3c27208a8e827e71ffc
ca490d4e0c1a2224d50ece72e9ad3729d4df7147
605 F20110331_AADJRR he_j_Page_121.txt
73173d7081c6c1aa473514b00bc51366
3a027b7011b2495dcb42f11892e40f2565155e96
40724 F20110331_AADKOM he_j_Page_126.jp2
d49bde31a3085a2947d46bac2edb6e48
6cf95243c8c3dd61ea2e3a32392632a08c23c873
65528 F20110331_AADIHO he_j_Page_281.jpg
d2e4649ca2f5b0588ceec77775c9ce72
564e1b5420c400a845f5d38a620833cdf0dd2d1d
3951 F20110331_AADJEH he_j_Page_268thm.jpg
97c1f2dd94ffcd05f033c2d0c560ddfa
f6bd5d4747539c8164e713b0de12bbe8fc258320
22173 F20110331_AADIUZ he_j_Page_163.QC.jpg
e7c23fcb8aea14b537a185b3ec7a073f
e44e18ea61c0f2c1547c68ce872c86f2477dde45
56078 F20110331_AADKBA he_j_Page_176.pro
1c4a4724b2ce91f15bd95ae0777931ca
449dfd3f6fd5216f01dd880b949f535e459433e6
3164 F20110331_AADJRS he_j_Page_122.txt
0ad4c079f54eb4bb83469c57554665c0
f01d54c99229c42d15d6f6bfdb5c4b6a7c6ab1d9
62862 F20110331_AADIHP he_j_Page_291.jpg
ecc798ffa7be5b5538885416fdb3bdfd
ae2450418133e4ade77fe1de14c1225a52d6cdd2
35463 F20110331_AADJEI he_j_Page_208.QC.jpg
2f1552589415a796183932440177b440
8404aeb179b1150d9ac47f883c064514b753240d
57247 F20110331_AADKBB he_j_Page_177.pro
53987be8e5249063442f33ceeb9d2126
9b228f41d02a04a35271f8a3ee29b0b7608346ac
3648 F20110331_AADJRT he_j_Page_123.txt
89d064229b58366d50bd8082accf71da
72dd893ad4ebbdc4f909bab194f0ecc9d0d94317
646014 F20110331_AADKON he_j_Page_127.jp2
afb5ac6141abf6479f605eecc8755cf9
2a459d202986232b9bf937d3b2e2bcf875c48c8b
35852 F20110331_AADIHQ he_j_Page_178.QC.jpg
e17ef2f3e937874b6619ef9d31e5be30
6360432aa491f6cf6a57a75840326a02731db0ae
4437 F20110331_AADJEJ he_j_Page_290thm.jpg
f1885b31b71215b4017bf996cd884e3a
e9d0c4d9bb1502ef419d0a953a95170d2a3a246b
51755 F20110331_AADKBC he_j_Page_180.pro
4662390474e020ea2a5d82f90a3a5f1b
2696a2c50df9d8d7e6491a4ccd016b7002af2067
1733 F20110331_AADJRU he_j_Page_124.txt
99ad5c4be1e26463b0f0379d2d53af97
d65134062794aecbe4ccbc93c351b5f7dce8825c
117701 F20110331_AADKOO he_j_Page_129.jp2
b12dbedce1733a9aee1f6050ec592b81
82147192762057a7445bc16d92dc046788253fde
114146 F20110331_AADIHR he_j_Page_175.jpg
7c1db35de9783fedbfc8ce30ec6f317d
6849b0eaef63ab35c104d492b8b3c0db608c73ab
2625 F20110331_AADJEK he_j_Page_316.txt
c4282552d5c10f06ac8ccad7a3f185fc
793de744f5080856a71d0f4240a819a950d3a04d
50512 F20110331_AADKBD he_j_Page_181.pro
b970274413cbedb69aa3a8d034afb7a2
f074de8d2b470f15b759c92f773b9ef2fd9c23a5
1237 F20110331_AADJRV he_j_Page_126.txt
2cea2e33f83baed6422edb4ee0d65710
48a657cb24676305ca968e04b66ff1c479805e91
114152 F20110331_AADKOP he_j_Page_131.jp2
91ff14d3e5e1e34289d3b89df90562ff
48d8d9d51bf6e4aa13aa5e9154d03ee73697d235
1051970 F20110331_AADIHS he_j_Page_299.jp2
257f1cdf90989bb3b9650ac4133d4bba
bfe91ec82659cf222e45e05e3ae8008e24b58133
8721 F20110331_AADJEL he_j_Page_009thm.jpg
2ca450a318cdb41d9b5bfc41e67ae096
1f8113476ae9c894b8929f96bbefde1806c3608b
50012 F20110331_AADKBE he_j_Page_182.pro
4f8c041b83240069cbe7a0c308c5cf5c
54ae0c337ed4787a8557951f44301a49144de176
1688 F20110331_AADJRW he_j_Page_127.txt
ca75238d435d5041b21373ab6154e8f3
9c9f313a3e51d509673dda9da5b052eea1836f33
115172 F20110331_AADKOQ he_j_Page_132.jp2
b64ca54416ad629c56e81e477d3d927b
374f55eef398a0b0af8e716938228d123d9fb9e5
53301 F20110331_AADIHT he_j_Page_052.pro
256a0b6615530f53ca7444649cfc9489
330ad08103b4ac29143481f26a5d6e9d0bd4965e
604360 F20110331_AADJEM he_j_Page_121.jp2
f6cbd7b69efe673fcd859de16c6955f3
d5a90f98d42e64cae1fef41623beb34e9e7ad190
54299 F20110331_AADKBF he_j_Page_183.pro
63aae84a55d9f95afc258aa71d3734b5
ead02026ac2253da9c1758e5af50eedcf4609847
2229 F20110331_AADJRX he_j_Page_129.txt
d61a6e7ad1885b1f9faeb86000f077ce
9e8e0914d6d81236d2673dcb0c5599c844753518
117937 F20110331_AADKOR he_j_Page_133.jp2
317124eab8bc05dd5a16933f989bee14
955928d9db46afad91ace2d386c2f1498ebb2b31
103708 F20110331_AADIHU he_j_Page_210.jpg
f77dfcaef0d7809e4f3d3d72e2371833
f513a8dca12a36594ff0e067e8d3f68ffebb68c9
106796 F20110331_AADJEN he_j_Page_193.jpg
bbafc9f63bd1dd7b1787333a7ebed2a2
0a76808cbf9ef1828b2741dfa2e75cf8d4236488
48398 F20110331_AADKBG he_j_Page_186.pro
c0b3aaebec2b32c8ae8cbd1e2077330b
ea296addf38c391484624fafadf7f17192f013bd
2171 F20110331_AADJRY he_j_Page_132.txt
314bddfdce57a0ea2353412e3e645cc1
61123a240b9cdeb4384c1556389f276cc4e83764
108371 F20110331_AADKOS he_j_Page_135.jp2
23b1044e2d14d3a27fb858f6bf24bf73
e06b916b22cc8e9c83680730648ade70076442c8
2129 F20110331_AADIHV he_j_Page_219.txt
c1db0d8451ac27bcf3ff8bc32c7a1daa
af0b050e3d736c8040f951979326ec12e8caae4c
36469 F20110331_AADJEO he_j_Page_279.pro
fc8836b8075348049032426074b3881c
802bd1ca574cd28397abecddf1db798d04c37767
52874 F20110331_AADKBH he_j_Page_188.pro
0d225cb5a3e3e4a6e75d79939a783a39
8de081d5a0cea0490943e3d0b23ed3121db90cd8
2183 F20110331_AADJRZ he_j_Page_133.txt
b792bd3415f69a80f1fdb7b75390f65a
8da0d85a60433415fedc3a31f7011a1cd7f74ba0
88963 F20110331_AADKOT he_j_Page_137.jp2
2a32e5895da8ce55c2900583255ce6ca
7b3e4f833cae129ea8c4e7f4b0f4da66b32f5c5e
32297 F20110331_AADIHW he_j_Page_091.QC.jpg
c9e33099a0fceb63a29eff78df7f144d
f29bc668f0089c12862986f623e315adafcd482b
100 F20110331_AADJEP he_j_Page_037.txt
f3b67b57f249a27bfb0c1618eb11c085
c376700cc0817a47e000769b2912bd56661a391d
53709 F20110331_AADKBI he_j_Page_189.pro
0eeb4a81e8da5280346c110572783d1a
cec45f66eb920378a9bd0cfe273caa9db6a4eb72
104701 F20110331_AADKOU he_j_Page_142.jp2
c21f56e978e8998314fc3147fb533587
ab98a0aaf4b0d36631c6fdff273fc7976941ee0a
34837 F20110331_AADIHX he_j_Page_055.QC.jpg
8b51ead22c88a44140af471de625a6ca
568117edbd758fb64042fd1121a0ea30d53a7669
107885 F20110331_AADJEQ he_j_Page_171.jp2
1e0f153182ea03b9782bf901d28f09c9
84170628a856334b36b868bc7a83225b30196087
49293 F20110331_AADKBJ he_j_Page_191.pro
b975dc51c123824ddca894cbc3bc6c99
18c4c650c7e7b36935994ec0fdd77aaa6cb64ed1
100031 F20110331_AADKOV he_j_Page_145.jp2
64595f37941b20244ab8fe4a2f6b929e
39916c99aad7bb2020fe286e70f941bf7f44b524
14515 F20110331_AADIHY he_j_Page_310.pro
f888df0734217d96fb8d80140fd861a8
e5899c82126fee202468b0e040c185d9b97fce03
1457 F20110331_AADJER he_j_Page_273.txt
5050c571edc9cbd4af2a13b23bcfcfd7
0927bccff363ecc2feb0adead806d05c3e58ba48
24260 F20110331_AADKBK he_j_Page_194.pro
f127426c33227ec21fde778dc07c28fb
819abeaad29dfe512ce7793e68b02970c3a4dc66
115486 F20110331_AADKOW he_j_Page_146.jp2
0aacebad4dd776252a940accdd7e9bd7
c18debf0abf208be662e07bed9e601790fab12e8
9301 F20110331_AADIHZ he_j_Page_177thm.jpg
67b8b8f4e3aeacb73d37edbdad7370b0
77ff0c6b031f703c1624b348f3b29638f75cdb7a
2146 F20110331_AADJES he_j_Page_073.txt
0546a9966b759513552ca9a6268449ff
3036c3c6497011747ccb499c1aef353c15e76784
12450 F20110331_AADKBL he_j_Page_195.pro
958b99dd9725f440247a42dacc157aa6
51936ebc0d9bb88041923a69e30fd034c5df9852
114942 F20110331_AADKOX he_j_Page_147.jp2
57e796ab1c83ef7bd21a28f7c3f70849
0419e56607927935bd3f0a6bf8a0026947837029
F20110331_AADJET he_j_Page_036.tif
e9c17f2f9117a76038d392c935393aca
5e49bb5d0f164b5a9f9e51e6be7217fbf12cdb9e
16230 F20110331_AADKBM he_j_Page_198.pro
9cd4e65f37ac2a5ddcedf77c56bc5036
c00854482295f769a9045a3c777d7752cb319a97
115479 F20110331_AADKOY he_j_Page_150.jp2
39f3f6a87591d3e9a27fdf050899f1fe
89adb18303190650cc0a7e178211c410da367cb6
34547 F20110331_AADJEU he_j_Page_190.QC.jpg
55bf53a9e4f892a1ff736b2a54f1fb74
50e71eee57d90c48c2d2e2d5c42579503de5c3cd
6903 F20110331_AADKBN he_j_Page_199.pro
653af5564ac79ebce9e9b8dcf9404396
04d1623cbaf422c31ccfb65f4be2ce0cdc77b57d
96876 F20110331_AADKOZ he_j_Page_153.jp2
7687932c02afca6ccbcf5b62048b301d
eb2a4f17e39b4c8f6f5b29e384efb0daade13321
98555 F20110331_AADJEV he_j_Page_143.jpg
46790da820dc64bfb64e34bc4b93b377
5131c31a605d00f66a888458b4c970023c0fb75c
40980 F20110331_AADKBO he_j_Page_200.pro
dc34003d2845658057f5ee080b1f27d5
93a58bb1ae046d1227dc82d7d9348fddf4de700a
412912 F20110331_AADJEW he_j_Page_155.jp2
2388c5f9a2fd6d39497c96331b6d1f4f
07493414a0dfa7f47cc9675e9c371daf524701a5
46894 F20110331_AADKBP he_j_Page_201.pro
67a3b33205aba345b0e72acb031a0174
7a1a25f50706998bd99f4bce1a3a411d67951d59
27628 F20110331_AADJEX he_j_Page_019.QC.jpg
669b713e09ed4885c1789e1930d22fd1
e70a6852918666a823ef27136cf1e2df4464a28c
105335 F20110331_AADJXA he_j_Page_006.pro
3246c6d9422afcc7970c5ff976cfabca
f2b10e9eb8d26e510555250f798069228f21e81c
40652 F20110331_AADKBQ he_j_Page_203.pro
d0de5cc7ca40f49ed07cd2c27629b659
a037e9b7ad305bf12fbd9ae703b8c2342c43d762
70913 F20110331_AADJEY he_j_Page_222.jp2
c3e18cfab4e1cd732f161a8cabeb9cf1
4a7d4ef489644fca708ca700f933b310a94f89f5
111511 F20110331_AADJXB he_j_Page_009.pro
cd23d483ac49bbb21fdb40ccfb743e4e
f98d2687f2025ec6a9a7b13281db8602ce003d04
40879 F20110331_AADKBR he_j_Page_204.pro
6904386abdfd4380c26ccc6a407ed987
d94995ea664fb63ef008b3908bd2fd47fee33a1c
F20110331_AADJEZ he_j_Page_258.tif
d362a1e6c16d0d3d09a9eaa9e6e8d87a
a3ad470c8a1550eaea8c11cd8d39ac6253017186
71121 F20110331_AADJXC he_j_Page_011.pro
b433e1edfbacb9b2128aecbeb32517ee
0cae31625e48ab1dc3ea439e6cb45d18e00ce9dc
53940 F20110331_AADKBS he_j_Page_206.pro
ae963181be76bdd114332083bd106ead
534cba054d7bd0003cddad8c15e23846670b0355
72668 F20110331_AADJXD he_j_Page_012.pro
86bd35c8a56f6d90fd1b5274558e9692
fcc52fbae077d9711d6122f411c3855fe1140f21
52312 F20110331_AADKBT he_j_Page_208.pro
89b8c6069fbaf4a7177b62a7a8291421
5f673f6c10773da1dd035d415d0cbfcc44e575fd
F20110331_AADINA he_j_Page_130.tif
92c26c322c1eadd2bec6ca7551a3c874
ca9bacf47e385c82758e225b15656759a734f8b7
75269 F20110331_AADJXE he_j_Page_013.pro
63f37383af51f1bc0ffbc60b84207ef9
d23fef47dc507a76e2e8d665334fe0447dc93d2a
52287 F20110331_AADKBU he_j_Page_209.pro
d27703da4f0d818a1fc5faf7f9ddbb9b
12662015e4f6d1e46829e75a506ebd882bffc761
34504 F20110331_AADINB he_j_Page_216.QC.jpg
95d0fe73cebbc115e3ed7443d54691f6
a73009df143900cd03f681c3a9fe8a98a8538930
79619 F20110331_AADJXF he_j_Page_016.pro
7c061006e55091706f6170c172005344
eb9f6b28c3f702c091125a339db4ecd5bbb670e6
51058 F20110331_AADKBV he_j_Page_210.pro
86c2ae963f70cac3a4571cb1945f4dfc
05993d3e76343385e0c4cbb51294fcdca5563155
1414 F20110331_AADINC he_j_Page_249.txt
c097ff5a3b55881e3e7a380273075f23
335620f8178a5435aaaf6f097e3a240af03fb512
80133 F20110331_AADJXG he_j_Page_017.pro
1c8dbab9bfeff837625340c06c1a2d29
654f7fe2aec8a16c1d81e0e1c59d6acde5d2d0f7
48336 F20110331_AADKBW he_j_Page_212.pro
afee81df26886b5cd23a2037d93542e7
1bcb33d92a79296f35eb898e4531764334efc6ac
7763 F20110331_AADKUA he_j_Page_292.QC.jpg
be096519c1b3a2b30dcd85d928271085
2139c80965b65c917712bd1f06cee61fcafffbaa
8301 F20110331_AADIND he_j_Page_056thm.jpg
d06edf9712084a3b3ffee7542bcbc661
3bf9008c91b524f3096828deddd36e36a0935ef0
60266 F20110331_AADJXH he_j_Page_018.pro
ff9cb23530d739f33a519884833ac83c
2a4cf5ec6dac786279aeff9333ceb4de7d52ec3f
50293 F20110331_AADKBX he_j_Page_213.pro
3ca749e9074256a8fe0ad18f8f98732c
69da0a27efb5d310a7c8cc74039fc43c0126be86
36711 F20110331_AADKUB he_j_Page_179.QC.jpg
fa5f352ebeecb68573873e2fbd7006bf
a13ed10350ef01eb53b4bd976327d05bd1baed9e
1051982 F20110331_AADINE he_j_Page_009.jp2
3d3a9196a5c9d980a63e4d644b56642f
1446d636f6561f7f807ad350c137cdc19acc5bd0
31968 F20110331_AADJXI he_j_Page_020.pro
e18ea107028210012097f23ffcd5c7dd
e4a5789b4cb868c3d7f48cb7646d41667dd83370
49436 F20110331_AADKBY he_j_Page_214.pro
e79b1eee9803e1bed3bd885b57c40df9
292c04e6013de7e117f2103424bf20bd9e867920
36973 F20110331_AADKUC he_j_Page_031.QC.jpg
a23ba32f8b52bfa5f58908c8fd5d814f
433d99e1c8ebefd26c7df2a544c196071523d25e
67342 F20110331_AADINF he_j_Page_322.pro
7ef575284bc41258d0084bf73ef59266
2cfb2c3d005fb2e83ee5bc4488f10a3c9aa88520
54775 F20110331_AADJXJ he_j_Page_022.pro
558bcd2d4f44474d088f2d463e056866
a33146005f5bfa24527defc18eafe837ab458c94
55743 F20110331_AADKBZ he_j_Page_215.pro
cf753215d07e82f9ac400f903233e2bd
16db03819574024cd0408a9a7bf057525f144270
34633 F20110331_AADKUD he_j_Page_193.QC.jpg
a62f2ca7df0a881316de5400f167aee8
182114003045a396407dfff72263056c399dccd3
18196 F20110331_AADING he_j_Page_223.QC.jpg
571e4587282f2ab1ae86f33b4d6292f1
a639b3aa10848addd6b6e91fa7a077fa69330f0b
54363 F20110331_AADJXK he_j_Page_027.pro
d9b7756e79ea45e68db73bbe14aee8bf
f791c60496818145001d06ffbcc9ed5b3935bce9
6619 F20110331_AADKUE he_j_Page_119thm.jpg
80ccb4e00ebcd20671f3f724977cb14d
306b1d734b24d4dadd076429a62bdd0ba583610a
F20110331_AADINH he_j_Page_295.tif
8bad2b0b7d26b0f2b5c8e6e9c71eff9f
b04a1e8854e57ce0e76c0e19bc31704b2347141d
F20110331_AADJKA he_j_Page_102.tif
5697e46a0008b1825fe7d8911838387f
81252af6ae5a80873ea0c1c9bf19ffbc548f6fc3
53318 F20110331_AADJXL he_j_Page_028.pro
316a601c955558b9357d6589327d7589
60b3b057d3dd0a1779cafcc7d3236335a0b2e821
8467 F20110331_AADKUF he_j_Page_103thm.jpg
c114df615ea0d66a0b3e6b04aec31610
b65cbbccc7f46bf1a445502406b76c3a53d269b6
8188 F20110331_AADINI he_j_Page_091thm.jpg
d321861f2904c68e507f16fd05d74b4b
e32a65159d039a9c8e317863c01abcd4461f0d5f
F20110331_AADJKB he_j_Page_103.tif
92426b5fe7ba08dcc061822f23677fd0
4f1a08f8aa96cda6c173df33a60116a3d74f0cd3
49215 F20110331_AADJXM he_j_Page_029.pro
47987ffadd42dc027187f42da2bc78af
8012d442973ec04eb5d5c518f0181ec895c1e4d7
4425 F20110331_AADKUG he_j_Page_257thm.jpg
2c8a0e2127274b8eb516a88d84a55515
e168e11e3d520ae68147de3bbc61b68363600d72
95235 F20110331_AADINJ he_j_Page_221.jpg
24dc03977b2ca7dcaff0ce065e191e17
381273929325ba5e899a1a7541f6726911493e9b
F20110331_AADJKC he_j_Page_105.tif
d129486cefe1ebd6fe34d55f1fb0f0ba
4c2dc5e3d9303085e60baf4b3aa52986ccf6b554
65809 F20110331_AADJXN he_j_Page_030.pro
e7cf946956179ce5719e456fe445db2c
65debea15791e4393acf2f964e1d13104fdb47b4
9510 F20110331_AADKUH he_j_Page_323thm.jpg
d2c48793ccbb9853e0aa9cc4c4d52d48
95c307d9c6c31d9393a022c5e10f3c759cc32abb
92003 F20110331_AADINK he_j_Page_067.jpg
fac451a5f18499726984fa786dda4783
e7d63df1f95b34c82ef7fe8ebc78ab3314356115
F20110331_AADJKD he_j_Page_108.tif
026122a6a16123e0fe6fc7c670eb3181
520aa390622d9ce244d8169f4e85adaf76d316ec
55258 F20110331_AADJXO he_j_Page_031.pro
346f18d3aed5f76f42508e0b358f3ec0
3ea9db66d8b876075358175c0e8e334820d8f10f
32058 F20110331_AADKUI he_j_Page_029.QC.jpg
a75c7167f3e911cdcb178fa14124962a
ff1533827994881849b465dd5e603cd39dfadac2
7686 F20110331_AADINL he_j_Page_207.QC.jpg
e32e5d25b4a36d19e46dcae90ae9f62f
fbf72050eb2a51e9c4d0f99aab5ddd8da21f87f1
F20110331_AADJKE he_j_Page_109.tif
4f26702aaced50ce16b59589f31c41e5
ee4d5e010c56ad4506a51f755ff1b4d8674b50ea
8423998 F20110331_AADIAA he_j_Page_201.tif
e888b1fe9f2908a3a94eb36959e4ac28
3d6118e452708fda2ae40dc27a62d5dad1730385
54746 F20110331_AADJXP he_j_Page_033.pro
74ae87cba5de81403b05cfba33187a9a
f620d013da03be3f2f8f9c3c5bfeb93763b6668c
35806 F20110331_AADKUJ he_j_Page_025.QC.jpg
edcb6c878ca567cf24a98807aa63816d
27b1c82236e4627c83efc95747abb43d7980f658
35052 F20110331_AADINM he_j_Page_286.pro
f265b96a401abc637cb8a52db1bd6d46
3fe6b31f834fbdcf2b1e477ab96f2bd9f23f8b52
F20110331_AADJKF he_j_Page_110.tif
1e23085931dfd143b7cb0f1dc00bf976
0e750c6523b7604084043d6eeb28092a5eafc417
32328 F20110331_AADJXQ he_j_Page_036.pro
13fc6e25c36ef0f6293b43ba37137564
e44b32d1c9f1183db83a4453e38971d83532c04c
22038 F20110331_AADKUK he_j_Page_205.QC.jpg
dcb8b457a6c239a7e40d1a0099edbf0f
29191921b8b252443010e96f1fbe5ac0c9081bf1
35820 F20110331_AADINN he_j_Page_183.QC.jpg
010bd31a9b39e802e003e981cf3bba02
83327a6be600b90728fbcef3aac9e26c3e165360
F20110331_AADJKG he_j_Page_111.tif
8c985723380d577723e4e5ef07e5849a
1cc548966b159f9fc1e4175e8e2509f131b11c82
9306 F20110331_AADIAB he_j_Page_296thm.jpg
e826a52ad67aa463d8044ae520e84a10
fa8311a3a69e7f0ff567780fecd6254e8a1b5f2c
1348 F20110331_AADJXR he_j_Page_037.pro
f626bdf96544971a3d544e68173b5a20
e62b855ccf29ac6336e65af0e184a8d9afc8f31a
8804 F20110331_AADKUL he_j_Page_027thm.jpg
6cd5021061ca9b9795d45a97ebcbdcd4
696f8200d996c4d7c71ea8082ad469e373580b93
F20110331_AADJKH he_j_Page_112.tif
19abcb4a346496b0a235422419d0353b
422788d94f42bb6689cd7ca30e7cc5c608d86bef
39673 F20110331_AADKHA he_j_Page_111.jpg
acf932ba0c46b65013644fb387c5d44e
79a01fdb1e88978ff1758aa44d5ca1a7e615bf86
2319 F20110331_AADIAC he_j_Page_233.txt
db5b73ee7da0cd0664c7bb8a22bc870d
b244e0fbc88d97cc1eef4b9e76fc52c6ef0bca4b
29244 F20110331_AADJXS he_j_Page_038.pro
3d1deeb8ee19b5f023b92c10e3437f50
d3ccc1013775ad23aea5d3779dfa7408144a39da
27374 F20110331_AADKUM he_j_Page_136.QC.jpg
2ae7fb81f780028118fd33f4449312ed
12e3e5c42501c7c311f49d26b9df5e67502b1020
4044 F20110331_AADINO he_j_Page_270thm.jpg
d7317d2b5b08b15c9d2cbe8ff660d7fa
24c3e3f86e84a30a6331fc3b5741ff2f9394675d
F20110331_AADJKI he_j_Page_113.tif
37d5d4e0847279d77535058b43116a00
9a69c1d1fe9318924c5626c365925019f0250eef
50188 F20110331_AADKHB he_j_Page_113.jpg
4070832bfa0d1ac48fb3067bc328cde8
269fad61de5a247e60587854566314bc49938be3
37787 F20110331_AADIAD he_j_Page_083.pro
4b9353d143e19f0d8e88c9429f98fded
d3d345a65328516f1ad85af23c4fa5702ff4ade5
47745 F20110331_AADJXT he_j_Page_039.pro
9e27b8b06f3e44caf0748a537f23a44f
5fd424ab59d6b7f51b9a8d1e54438eced348c534
858 F20110331_AADKUN he_j_Page_003thm.jpg
43952aa5bb872d1a26e64f42edac6c57
0ba8c9b08539715de25ecacfc5448942de30156d
8268 F20110331_AADINP he_j_Page_008thm.jpg
c06ea3f483d7f0877683e6c2f2f5d556
80884eccad06a0aa46a8b8b39280f4cec30ef483
F20110331_AADJKJ he_j_Page_116.tif
101ff62636505d982e9dc7b1ed04dee6
99c7836fbf6ee8dd62d4a5e25f406c115499c8e3
52572 F20110331_AADKHC he_j_Page_114.jpg
178b31dc13a5fab60e0f74d61077fab9
498f8995c85970828502d701ef22f33a5907e387
57764 F20110331_AADIAE he_j_Page_117.jpg
91e291630207fb0e798e3715a3e117d2
f2e97270fd3bc6ad8db6712f3c928c1a9f219f7a
51289 F20110331_AADJXU he_j_Page_040.pro
88c49818ea9b3170fcf1f11f74fdc657
bab5d6fc7d0a24238655374b0ce57efdbadd6b78
8246 F20110331_AADKUO he_j_Page_099thm.jpg
c803277892b5f3181ff457c4a94e8f67
072ec27a4791bb525297bd0b884b5c73ffdeb87a
5425 F20110331_AADINQ he_j_Page_315thm.jpg
439666afdf2879ae91053a6677cc64f6
191a1a3a50142426c679183361d5c9e45e1d4e3f
56185 F20110331_AADKHD he_j_Page_115.jpg
df1e4e6a767227725484d5d0acce53bc
011b9dc766df5c46bd3246caf6de3188951200af
110720 F20110331_AADIAF he_j_Page_216.jp2
bbd8b2a5b4dcd18b5500fa1befac3f1e
a069895573c532fdbc92eed15ee2126e5c59cd4d
47983 F20110331_AADJXV he_j_Page_041.pro
187035274645cd7e8c88690faf081950
7a9fea4fad0bca210359a908188ef910f8ccca5f
8793 F20110331_AADKUP he_j_Page_133thm.jpg
0e9ec0c4ad1b66178fe7a1c1daa432d8
e45651217a06228ab90f87ddf08226614bd1c4f2
2084 F20110331_AADINR he_j_Page_107.txt
92e1284523c210a135c44f37fec22777
7a6b248051a23219a2c28d2eac4a6cf8967e7227
F20110331_AADJKK he_j_Page_117.tif
dc9788d976eb53cdfcb80765479e7d85
475432cbcbd1924768fdd9cc140d861f382de766
56671 F20110331_AADKHE he_j_Page_116.jpg
cd5acc7364a256576e87c294e2eb7d76
44d42e6e632489dcb1e1d4935c1ea351f3680d3e
582296 F20110331_AADIAG he_j_Page_312.jp2
4731fd92d7a77ff5725233077051170f
bab644d8c34177744be7e246efa0166d3100cb0b
56389 F20110331_AADJXW he_j_Page_044.pro
b900c471a7c24959af7a54946eabdacb
9a6b1f6ba599971d4ce3b2e29ac98ed99831057c
17683 F20110331_AADKUQ he_j_Page_289.QC.jpg
78970e07120d4c89d5c545a2655bafc4
37dd40d75106e462eeff095804eb562e92f653be
F20110331_AADINS he_j_Page_150.tif
258ecb5f4353032178af410c2586116a
1c57bfc65cbe44b038afc9a23e5fec129e471ec5
F20110331_AADJKL he_j_Page_119.tif
e91652a0a1cb30e6a98d9cf801036cee
7955717026d481001a688e7b498420f8b2558f4e
62692 F20110331_AADKHF he_j_Page_119.jpg
77f36098a2c05aa8e4161399bea28258
45a96f046434adbbca1464e165ad666b473ae575
F20110331_AADIAH he_j_Page_257.tif
96167bdd221f247a9e879e9cb307306f
7454a69c8c7f7cc611f5ec16eb901566db8e889c
14905 F20110331_AADKUR he_j_Page_038.QC.jpg
7be3e806073ba3c20e2c26600f0bcf26
cd9fa55156e1dbc639e71914c78074ad393c443f
29139 F20110331_AADINT he_j_Page_054.QC.jpg
914f43944ba88d9077658814c0c327d1
95d50ff62306b461e421fa5357218bf62bbfe6f1
F20110331_AADJKM he_j_Page_120.tif
5b974f468470c09d55c915df3559d793
0cfde4d06baa29529d3fb563ebd67a5805d41749
F20110331_AADIAI he_j_Page_132.tif
cccc7b079d3615c1def8927fff385346
550fc84ab0c4a86825e2dc3ca0dc9d5aab32d2ad
49110 F20110331_AADJXX he_j_Page_048.pro
e596fc35ecb426a8c1006114efa3e6c1
c820cdd6509c81ec2869299343ae55484c514e92
8784 F20110331_AADLEA he_j_Page_055thm.jpg
0d126f7e77d24893ab500c826e438edc
b164c56b476bf33c6f64cd03d06ca7841604e661
6762 F20110331_AADKUS he_j_Page_049thm.jpg
377b7d0f636a86ce29192b7787b32fd1
69254288cbd34e539ed441d44c83a8c459a3816a
1051896 F20110331_AADINU he_j_Page_300.jp2
fd3eb5f2055f8c158dfbf684b7d32fab
86b09dc8d3ef6533b96ed5998137fb9d32b67e7f
F20110331_AADJKN he_j_Page_122.tif
d82008ae4c8f448aabca77ede7da18f4
2cb1627303e4a6d89cd7a77e32711f430bc10cd3
65900 F20110331_AADKHG he_j_Page_120.jpg
6904a83d5c72d29017a807eb7571d2da
39f3caa4e6f0ec686d3224cdb5f37b7c7a6c089f
121078 F20110331_AADIAJ he_j_Page_177.jp2
d372fe63e606c691d9418141c124db58
f7fbe66076d0667ef1824661026957c9280129df
42860 F20110331_AADJXY he_j_Page_049.pro
eb06d05c2d977a983e13d25563db9c46
0e6526ac3746058d978d62a2d860f7dd0ed8024b
9202 F20110331_AADLEB he_j_Page_058thm.jpg
d0a7e9dba28799237417a206e802d097
3e3e55d963466e2fecd2dd07a2df0dc025621063
24168 F20110331_AADINV he_j_Page_083.QC.jpg
e578052a942ae78855e1d5cb134da3ec
6fb3638734191e93f554200bff23075d6ecff402
F20110331_AADJKO he_j_Page_124.tif
47dde56ce9f10dec003cde83d5704d07
0ebb4a536a319d0517a6872c3aa2fdfef2174a9b
99378 F20110331_AADKHH he_j_Page_122.jpg
9eef75bc2d9663570d7355ef69b93c30
7959763f393f6dfc8eedc7464ecfbc053a365fb2
10011 F20110331_AADIAK he_j_Page_320thm.jpg
0d5814b2e23dcb54ebb3db0f8734330b
87251e42033ce70a33e3f923c634b45a06072b0e
48098 F20110331_AADJXZ he_j_Page_050.pro
3a89ecf05849717c4b067b970df6eaa1
5f526bb0dd126a0dee103e6ab2f0d18e0ca928f9
8838 F20110331_AADKUT he_j_Page_129thm.jpg
08552a7831d1b6b1e6ffafa7accedef6
3e1b2aed4f4ae6881a05bac5a6dbd7ae90ede78f
1051689 F20110331_AADINW he_j_Page_297.jp2
ae1ae0e1a0f0700ec1a07f8b72d1a158
0a239196b5806a347816c3d129181241a6c08256
F20110331_AADJKP he_j_Page_126.tif
7a0e5fe0c4946e5e526c63419c917930
ce36a19b90d6a922386e0b1064b1d80c45c64fe1
128348 F20110331_AADKHI he_j_Page_123.jpg
cdf642512800cdaf6b153e3c35c62891
3b34a75f113ef1d2a5f0b748bc08331763f31f34
53778 F20110331_AADIAL he_j_Page_217.pro
d33d9d8d408d42ef8b64fe55aa60d8e3
3dd12d7209a2bfa94fd36d9775da99519f8051f4
6492 F20110331_AADLEC he_j_Page_062thm.jpg
f0fa83083fdd29099f4128b3ff8ca76a
fb761d63511acf087e6f0ed265f73afbed845024
36227 F20110331_AADKUU he_j_Page_123.QC.jpg
c49e2284ad8e3c5f1d3fc873f14b0d57
cbb6ad69d4700c35f84b8bd49ae62b30b7b864a6
5408 F20110331_AADINX he_j_Page_309thm.jpg
2801d0b6a0b5a9673b6b716e57b75d21
a690ef5c4b9a26424929fd859e9ad25275783d1f
F20110331_AADJKQ he_j_Page_128.tif
fd9513598cdbbd6a5e48d9fd95de0475
7c9efe49dd30ba6759f35d4ba2044b792bd0c356
50410 F20110331_AADKHJ he_j_Page_124.jpg
2158bab9f9817380d055f1f624f220dc
09af973da638b6a8d4c93f7efcb016b22b47580e
52354 F20110331_AADIAM he_j_Page_096.pro
38f9c7e11e126bdf7a3b0d4a9cc94862
b368142f97da34fb17a4be17f5a5c3deb209ff67
3449 F20110331_AADLED he_j_Page_066thm.jpg
18dd7bca4c4e2469f0636c079be028d0
92efcf2da5bc2d99fec2c4ee6d50a978a61b6e2b
36835 F20110331_AADKUV he_j_Page_030.QC.jpg
1ef1782c59a12944ccb73cd89cd2b23e
8684cc0756edb94c3389872391d4724a31d0e0df
F20110331_AADJKR he_j_Page_129.tif
65eacd3f7f597725d2b3e8d0ac4a6fe5
13e8007b1221f02afe6c19a044fa703e87e2ef08
40535 F20110331_AADKHK he_j_Page_126.jpg
d40d7991073dc8032567123641721542
d013375deb2c75f6c92f7e4188a4ce6ef5efc1a8
6312 F20110331_AADIAN he_j_Page_118thm.jpg
6d82abd87eb16bba5e289d382b7db79a
813c31f1d7236292347406c315e16307383156d6
4511 F20110331_AADINY he_j_Page_251thm.jpg
773812efc4ab7d066a107f498fba28ef
2e1da49cdce5404910688be15382e6edb1590284
8927 F20110331_AADLEE he_j_Page_073thm.jpg
1aedfd43ad38c7aa2fc3315f72c9cd0e
f8444354c2ca99667e5e359f3a8ad79ded6fc7cf
9770 F20110331_AADKUW he_j_Page_327thm.jpg
2948947e43aa027606231db898066796
957684923a51c4c06e650adffd797a32355ef0b5
F20110331_AADJKS he_j_Page_131.tif
f7a354f245516b0107d53af7ec2c3a36
a542be8a864044a56dc38c7e56ef27a32682011a
44229 F20110331_AADKHL he_j_Page_128.jpg
c6b21613413c49deb03a2931032ba1f2
ab4a919c5838906c677cf2232c34b4af67cf19ff
65171 F20110331_AADIAO he_j_Page_258.jp2
87bb451f3e35d0eba3f0b6098401f15e
2194a6abb5904a6dbc61a88b128f871a7a8c19c3
8425398 F20110331_AADINZ he_j_Page_127.tif
5cb9c90d90857fd668a4d3af007b19d1
ed7b794b9ea35337fa8a586f25f2cbe582d48723
7518 F20110331_AADLEF he_j_Page_077thm.jpg
22728cb37734474be4d313fc3f24249b
68b9e826f560e5167f472fe64c4b570cd4c6dadf
3703 F20110331_AADKUX he_j_Page_126thm.jpg
4b57fd827c437cc3e870003ce4bb773e
728a8c7c3847e88895198d78fe059476b5572a0b
F20110331_AADJKT he_j_Page_133.tif
3b25121c1c62234bda0432c7a6ee69b4
65cdb17111a347a382e1bf04eec66069c5c64a78
110000 F20110331_AADKHM he_j_Page_130.jpg
e23ecc575d9d8957e4f7791e8a436288
96da3a731bfaff18dc1e1c9469ed9c0d3aae01c6
1970 F20110331_AADIAP he_j_Page_079.txt
d213c0cab60cbcb0d4b25e9a15e68cf3
dcda04d26196ec8c8c620e72265c8f7f591a4cb3
7904 F20110331_AADLEG he_j_Page_078thm.jpg
7d375663c49adc11d902f3cdd1df01a7
92f2123dcb5f77393112992f40fd1721eecbdc39
36170 F20110331_AADKUY he_j_Page_022.QC.jpg
b5eb9bd0fd040259999c5c9a5c6de0a1
e7bdb176f540c63303892f1cafaa22839aa875f7
F20110331_AADJKU he_j_Page_134.tif
ecedf2431a863c35a2710e1a14dde10f
fc1cfb30e4c6849618f2736f3ccf2ed970fd746c
110983 F20110331_AADKHN he_j_Page_132.jpg
b69bafc49065f11363b05a7489167434
117b8292fb9c9b55b00be00d07fca77be7fa6756
114958 F20110331_AADIAQ he_j_Page_094.jp2
d1e1c7f2479a57dd9493da5006cd35a8
06c24960dc847c6792af4599a54705861fb02684
8489 F20110331_AADLEH he_j_Page_079thm.jpg
9d5f514ff661956c79fbcef0f6c93585
d2fecdf25a328ecacba223a346ff0cbea87778bb
4890 F20110331_AADKUZ he_j_Page_197thm.jpg
afd69b4cd2a3902955d6bc50f2f2c996
e05eb9e9ddc51988061e54ea06517c02cf0d30e8
F20110331_AADJKV he_j_Page_135.tif
26682111df4a010dd8c0ef5b31e7ca31
c773e789592907dd852299d10716a3a11072fc26
103735 F20110331_AADKHO he_j_Page_135.jpg
a9015bee998b73161e0380ef5266e937
ef0585b805d1138d62f1f18e1d1a0f389a8d3a37
F20110331_AADIAR he_j_Page_002.tif
9aaf0ae51248025427c36fdd9431cd2f
4f210d8245dd93a3656fa44635346034922168cc
7121 F20110331_AADLEI he_j_Page_080thm.jpg
61450ade20e7bc673a54a8f8d48aeb46
f15db395997a3a074bdc2a521a5adaed65b19221
F20110331_AADJKW he_j_Page_136.tif
f04f565ad9ed8212c8a2855dcd2204d6
e461896aee2cf55f4475d2778a9b67d13f2bef8d
88730 F20110331_AADKHP he_j_Page_136.jpg
ab3efe44773c4d9e336906bae3228201
762aa999813cbeb37fe0e09576351f626eadc0aa
1345 F20110331_AADIAS he_j_Page_275.txt
8e8803895484e0fda425b70086711405
936a67fed260b52f715bcb6c2a19d2fed3c9dcc9
7377 F20110331_AADLEJ he_j_Page_086thm.jpg
a6de5b617d2b107af0c8a8011af52a65
1248c3795882e58df208ab67ed1272222f6afb96
F20110331_AADJKX he_j_Page_137.tif
bc881b6ec3bb2198cf7ea4c21a96eee1
9a25917e021b35c0b5204d42e21b338e66ded735
102911 F20110331_AADKHQ he_j_Page_138.jpg
dc2f2d347129402175c51e515a5f8dbb
7b06e6c253bee636287e91e69fe570af84488c38
2034 F20110331_AADIAT he_j_Page_181.txt
ff799d71732287172c20589fedec70f4
e4f7757c597adc3234c9ddef811204759464b518
7075 F20110331_AADLEK he_j_Page_089thm.jpg
7e3df2eaafea030fff8fdebe79abeec1
c8dcbf16f19b95f704dca295f0194fba9b05fba0
F20110331_AADJKY he_j_Page_139.tif
3ee9ff9cdfef75fe099236f5dcd84eb4
b5df1d919d86776ac012d2a02ca48ff69623b10d
102927 F20110331_AADKHR he_j_Page_139.jpg
edeb90e3e472b29acd742f3f280d9147
fd8e6a390d3cfd8305d935af968ced1fada1c1af
8064 F20110331_AADIAU he_j_Page_156thm.jpg
79c778fdcd912b1314d9d945d93cfc3d
16d9780d2684f2fc99d4e9946e07718af89c8082
5938 F20110331_AADLEL he_j_Page_092thm.jpg
48f06d041c9b873e6e6398a7f511e2c1
00c257b639e297e07bcdc57e4d6ddc6b63afab78
F20110331_AADJKZ he_j_Page_142.tif
ce2f4ae621bb45cfbb0bf3298e0d7451
64d268fa4752a81788b51653c782f255ffc1aecc
103346 F20110331_AADKHS he_j_Page_140.jpg
8b2e74ff47b0fde40f1f21fadc4bda39
9578ecce36db79c43d9a4338c5c722beb6cbecee
5741 F20110331_AADIAV he_j_Page_108thm.jpg
5d2e5a33ecc1e2e8984889e9a44056c2
79cd11395d6935fe9a15a203cf4f848c30cdf7af
9079 F20110331_AADLEM he_j_Page_098thm.jpg
b555e001b631e77b971b83609137dc76
d5849f0d059e8911f38e6dcfcec98bc1dae06f3f
87475 F20110331_AADKHT he_j_Page_141.jpg
ab2109732228edd9fbe4d048f17ec916
2c32bf418a7e861b14e74938075ed7bd9d7009ca
2050 F20110331_AADIAW he_j_Page_023.txt
1acc772c296a98a9d10c589f9df47612
8f1977b8fe881837b103f0e19d0a5a4a66b5ddbe
8914 F20110331_AADLEN he_j_Page_101thm.jpg
2a5f3c6786e576deea964f0da8744f91
ade6a6dc2800b534595361407f8ad05ace9b4c2a
99086 F20110331_AADKHU he_j_Page_142.jpg
17e5664fb3e5a202f05c83a6fa673d73
23776e43ed4555121723b21b50f2c801b0164076
F20110331_AADIAX he_j_Page_158.tif
b3d3fdd76f51577ffba575dfd84724f5
3468fe7f374b34a120cd69252a36a2f542542a17
1261 F20110331_AADITA he_j_Page_155.txt
c7854f127724fd71207f1f7560524ba7
0b8b6786009e5bc5a8030f26e06f0b675469b82d
8192 F20110331_AADLEO he_j_Page_105thm.jpg
9865ede92efc7b6c5d9a8ae72109ee14
b8dd796f2fdf08693763faeca514803f1be6a129
92798 F20110331_AADKHV he_j_Page_145.jpg
20a3c4a35ee8ec261c8914c3911f446d
32b61311886f35abd68e467c502d5604aaba8c8a
1168 F20110331_AADIAY he_j_Page_120.txt
1f057a8775f1b66827b6100c3eac5c54
e28efce2967f3bd11fc53201d03ed92d9c2b539e
34688 F20110331_AADITB he_j_Page_276.pro
2c7e545be34f0ccf9d603aea59a90122
43db2b8c7c0985059e6b4c88730470e6245740cc
8629 F20110331_AADLEP he_j_Page_106thm.jpg
dda016cbc1465c2194463006b2ec1104
24f281323fdc1603727b18421f56273b36c0e937
109213 F20110331_AADKHW he_j_Page_146.jpg
73f2696e2f94b9e94a0de3a83f011b0b
e4d45f304b1c9cdd65ad391c51c300e1c649b64d
61180 F20110331_AADIAZ he_j_Page_168.jpg
0c20536a5fddc9663e74ead112423282
2aab3c957988bccd294158b68a0ab8e532104651
8250 F20110331_AADITC he_j_Page_192thm.jpg
556d297ecb0171697e10c7a08e5367b4
2537c80d0ca17ded65c18007c913a03a01a0e399
4262 F20110331_AADLEQ he_j_Page_111thm.jpg
f988db0853fa406a007a65938660e415
1799197e969a46f3756011394b2ad212a1b682d3
108116 F20110331_AADKHX he_j_Page_147.jpg
fd45e5ee0cfdb4807d7ed6618de4251a
240221a6ec8ed073daeb3a6071bfdd5c23f8babd
8687 F20110331_AADITD he_j_Page_023thm.jpg
aa929a9f68e23bb136fe1d76baaf7fb8
6373b420e221f155c590d2c69d9ff91ee1f6ffe5
4539 F20110331_AADLER he_j_Page_112thm.jpg
179a3800d32a85e5f55255245ccce8bd
17110b738420b43987760885a798d3354432edc3
106196 F20110331_AADKHY he_j_Page_148.jpg
92be1e9738d13a96c2320ac5c6fe232a
9c148752ca5eaba82b21b3a89be820fbd5de1a1c
1391 F20110331_AADITE he_j_Page_062.txt
07b29c2d84f93cb0621fdb3c0454ecb3
a2226e327d4363095957b57f3351ef6dfe20fbb8
5595 F20110331_AADLES he_j_Page_113thm.jpg
9544de3ce4a76bb10275db63a78c3c1c
018f674cf64a29fe436820815a9eda3dcfb24520
110234 F20110331_AADKHZ he_j_Page_150.jpg
0ebe32e69539a2cce9e6ec8b0d6cabea
a06bad4cd54faf34cad8fc153f784d6678811d51
119746 F20110331_AADITF he_j_Page_031.jp2
210327b1965d32b39eab72402b0130e4
3e3b0d18298ffd7e36a31db79f8397acc397ef6a
6642 F20110331_AADLET he_j_Page_117thm.jpg
ce5fb46f009934f9b8010ba4c55cbbf8
eeaa4c56c889387bcc09a5a8f1447f38c79c492d
120444 F20110331_AADITG he_j_Page_295.jpg
0ec15a9d7a27b803ff50608a20360307
0f5d8b54e2e5dc136d60d23289b9435e83680a11
9197 F20110331_AADLEU he_j_Page_123thm.jpg
017050ed534ce7bacd191c14a224429d
d5739f03ef3a415679c99c785baabdcba8b6d995
1993 F20110331_AADJQA he_j_Page_043.txt
628427dd8af8be0156ccbf73bc554238
8e1b026d63185080c92c5e316ac96f094899568c
F20110331_AADITH he_j_Page_224.tif
92dd78bb732a1b934af5e1497fda9a30
e4f82153f1dd033ada34a49712b602c9ed4a73b0
7618 F20110331_AADLEV he_j_Page_125thm.jpg
152f6ec9cf324abba28ea8d7a28c1a89
0b577058686c4a37787f39e6c766eebe5e60afe5
2213 F20110331_AADJQB he_j_Page_044.txt
85f41d7429cd27f0a095e4a81693db0b
99d3cb871f37c382093920aa188724675f2d2854
104775 F20110331_AADITI he_j_Page_134.jpg
57591f4da2f76b7f7ece1afead9e0d31
c17963be3e8d2e2e783881c62e4b110087fe234b
2770 F20110331_AADLEW he_j_Page_127thm.jpg
6348b2b928f40fb8396a31cb79841462
18b39c46810874dbf86f222aa4c26ea5b22b44d8
2125 F20110331_AADJQC he_j_Page_046.txt
a693ba8edcb46d1ef5deceb0c437d6a9
c59b5e9cfad838508cf5fc370408beaed51df363
108990 F20110331_AADITJ he_j_Page_152.jp2
eb478eab5a6f2182a29a12e67e00d5d7
68127f92a5100b817bbd8029f4466fcd8129cce1
3950 F20110331_AADLEX he_j_Page_128thm.jpg
aa993a23fbb43d54e15b13b8a4854150
5a7579fefb1d65930f5501d343ccfbc69c3cdae3
1903 F20110331_AADJQD he_j_Page_047.txt
a596d3317ec465807e25954be61bada1
c9ca3a3c0ede2aa9251744d3b7fd989cd8ac38d6
107455 F20110331_AADITK he_j_Page_029.jp2
e28abea1c2c437211671319396585221
b92d175b826d8dfdbd748b89b2739f33b1457050
8755 F20110331_AADLEY he_j_Page_130thm.jpg
d62f06b97f2da7948c9ef0c5dfb3beb6
12f18c074cc5a8317e385fe77e355c80ae22fd72
1823 F20110331_AADJQE he_j_Page_049.txt
d5d16a1691d2b35e5ee8fefd37cab573
16fb15682f351c7b84e6d3693c40ab1f3dc3475f
37603 F20110331_AADIGA he_j_Page_259.pro
6f7b66fc122a1a641fa2626d3e4acb1a
d2a3ff71729824a4900e449ecdd2ae384fdc7f5e
7727 F20110331_AADITL he_j_Page_082thm.jpg
0c525af6f3ed459d1183fab4c72504df
030d5e389d3f126978b149a22aa75d34bc905a45
8881 F20110331_AADLEZ he_j_Page_131thm.jpg
927ab0f5a327713b987b587f08c4f91c
56f7d8ed1dfa1cef00e7a71f9c6e47fec52edfe8
2108 F20110331_AADJQF he_j_Page_051.txt
e383a08ed25401965524de0d93b4bb9c
720af78a86f7f24105768bb7499c0b250afc15fc
43188 F20110331_AADIGB he_j_Page_054.pro
58451e9266ad9f588661665ec9f18cf8
beb09f4a227ab31abdd56396abdc5db20da89cd9
4237 F20110331_AADITM he_j_Page_230.pro
0ca3157da2fb369b11be86add6a80a13
84593ffbf6ff7869570e6eff7101b43b4d3e3d99
2098 F20110331_AADJQG he_j_Page_052.txt
c3d52c6c5c1d66ed21792b7932591d35
691b22e76d3c8065057e799f99318671865fb9ff
4052 F20110331_AADIGC he_j_Page_267thm.jpg
f5258683541303e74f820feee65b2a64
237c2f8a73ec27a5b1396fd8acdd3ecd035f5611
2054 F20110331_AADITN he_j_Page_060.txt
628d2ca5067e9bf5ddb34f37d61931f5
f545fea6710a5d65d30a2153663ce209318e8e95
103266 F20110331_AADKNA he_j_Page_056.jp2
0638b194dfd1db199f6760adac5a9a5c
e1d59b2fafe90646d53bc2bea177f4f4e1344f2b
2196 F20110331_AADJQH he_j_Page_053.txt
80c1c9aaa9a0fe954b4aaff437fd9e51
49e2c6c72a3c57ee164f26b54228532808cdf880
F20110331_AADIGD he_j_Page_085.tif
aec8a069c8f884949b03f76d0d4d235b
5f0ba328ab798a432b81c3721d22de47722d1d5a
99285 F20110331_AADITO he_j_Page_212.jpg
49a7673cd2a09770877c0225ff8a89df
0243d5b785603137659a60b7153dd6fcb48558d8
108181 F20110331_AADKNB he_j_Page_059.jp2
83abc9b1e4bd5199632f38fa4cc15531
7bbcfbb5de34699de51023bf62be9e115274e7f1
1731 F20110331_AADJQI he_j_Page_054.txt
453658f34577b331c48c6513f7faedf1
e3debe4997c89d2b0b25f7a4e424057aa91e0804
7900 F20110331_AADIGE he_j_Page_153thm.jpg
166f55f5190e987bea6f55fbe9fa4561
75ffb864da02556144944bbd47913db8387e11d6
F20110331_AADITP he_j_Page_084.tif
16f027763762317354c6a13e3853b90d
5d91c156c766f49eea83b92b113bf103f2fef45b
113357 F20110331_AADKNC he_j_Page_060.jp2
452852cc354f94fab4f84dc631070a00
c4b8de5f5b6235d328273884fe260814598496d4
2018 F20110331_AADJQJ he_j_Page_055.txt
a366b711bab279f58f104c8c65cf40de
535cf90e65d46483fe883be69abe803fc8a281b7
105037 F20110331_AADIGF he_j_Page_076.jpg
4bbe44872a3b3f76a408438a550f072c
3ffc4baf0dca3ffd5f4a5626475c3a2410b1fd74
F20110331_AADITQ he_j_Page_323.tif
637276b99696adcb1fd87d6faf8ed5c3
935e43e3a08d4e615163fd89a1daa4cf375aa8d2
848200 F20110331_AADKND he_j_Page_062.jp2
c4e9110b3328743a3a81fcd76b04449d
7662338b44522fb52cb12c401716b733166b2299
1877 F20110331_AADJQK he_j_Page_056.txt
02f3a761122edc942ec0773dd32a2986
220a02cd6bd59eba880c2094f6fa79477f07c9b7
6019 F20110331_AADIGG he_j_Page_120thm.jpg
ab0bf97d53029b2e2b78307b36c830a6
b6d3af23b884e9a58736e9c745027a06ee3e8952
F20110331_AADITR he_j_Page_121.tif
21a53dd0c0c88850bc3d63a64e313ca8
80385d10cc1e97578869aa67fe01b9b16af05c0a
844148 F20110331_AADKNE he_j_Page_063.jp2
f64459c3d693221959a08e26ec8d1e6c
e155cc300d7b2ba8aadfbd2e5d9305e8be700592
2189 F20110331_AADJQL he_j_Page_057.txt
4dee5e9ce5bf35367e23760673c9eecc
93a168863f6d401700136cfb2fd48267221815c3
75561 F20110331_AADJDA he_j_Page_281.jp2
07173a01816bcdff80e701d614a5b364
fbb3da5909302ea28c9526045e59f6405fdca14e
6556 F20110331_AADITS he_j_Page_158thm.jpg
8d0e1e68ebe53ce0a7dd490def09118b
1f4d0342c7814fd2258e65b50e90df0098c33395
842514 F20110331_AADKNF he_j_Page_064.jp2
1f58fb47e8756f8cd1b428722ba163d5
110f42e0c844283e6a9cb8d07c59d6688d989774
2261 F20110331_AADJQM he_j_Page_058.txt
fc001d4f425700cb408d4813f32b2971
e0e039de3469d3cece3ec2d38ad659de97c30a8e
52876 F20110331_AADIGH he_j_Page_025.pro
7e81cca0db5bd4d2d6bc621d4ac017bc
05348039f5d2187975d966afe5ae1cc53b4aea9b
29055 F20110331_AADJDB he_j_Page_293.QC.jpg
785edfef2b86ef9ec81de806bd034850
e62ff464075945099152c4fa072f6da1dbc715d5
2079 F20110331_AADITT he_j_Page_071.txt
9b913ef3e7b432e7e7e7c6d0ff2a03dc
131645ae5d6663ff61b4253494d44da5d3d81361
26400 F20110331_AADKNG he_j_Page_065.jp2
ce4214d8508e50b9e9a86786c32a86c1
e1f2ce5b169b9e6d7beb11e4229efe1b4a934d5c
542 F20110331_AADJQN he_j_Page_061.txt
b8eaa2d509fc7476c49fa865694fd632
58e3451e4485bb65c8a41c1a0b5634c3d62b0aca
4362 F20110331_AADIGI he_j_Page_249thm.jpg
ae53266399f9b6a84f88d559c3498320
d7e62178d5a65e9a13d5502f3c73e22c2dccc8da
5501 F20110331_AADJDC he_j_Page_064thm.jpg
8051e028ac26b3bacc442deb91ecb93f
2cf5e531fce748438ea6986bb0163d7ca95ac2f3
91990 F20110331_AADKNH he_j_Page_067.jp2
0b20a9d62d5bcb4c0a3734523a8e9dac
6316604511426c41b6a6c1129a803225a9a432cb
1227 F20110331_AADJQO he_j_Page_064.txt
482ddaf44f0764e36d32e184ef260393
4f7c295bc9b7d8a9da783313b017df8f1a492031
2078 F20110331_AADIGJ he_j_Page_096.txt
dc4d5f22c35ba14f8f76f44aa342e622
a679207349a7dc1ae97b706db1cbb378031204fa
18972 F20110331_AADITU he_j_Page_277.QC.jpg
6b4cc59812a1e57513cd2fd255697a3c
6aa1865980d3dade2a5e0235c2cd70759ffefa8c
107280 F20110331_AADKNI he_j_Page_070.jp2
487129771b43b97a95d2bb8675838c88
02bd1f4fa6b7a6e8381c6bcf092aded7ae62c7f4
586 F20110331_AADJQP he_j_Page_065.txt
98c39feb3f7f4ac54b09c74d475e5531
97df340a61240a7c765934e19d9791ff493ec76d
115843 F20110331_AADIGK he_j_Page_178.jp2
65dcd0768854e628a8253ab0df5f3875
6bd8b4922eae95ee2af237f6d81ab124e5166758
545747 F20110331_AADJDD he_j_Page_160.jp2
5971afc2eeeb6f57d85c9338303698a7
e401856923c7f0aab16f185c46b71fc937766e4b
34017 F20110331_AADITV he_j_Page_139.QC.jpg
db22314f7a64d807151e08e9417a134f
51531e760b574cf9ddc0d45da4ec9c5bbcb334b2
97117 F20110331_AADKNJ he_j_Page_072.jp2
9d00f8155f8f005af21477201b82d180
589c4cbedb8431450a134432d2cc17264061db1c
1555 F20110331_AADIGL he_j_Page_232.txt
ae584a3a911115669087bcae25f6414c
d606455b91cb4a81d3f5829a7f26ce2332b25d52
32086 F20110331_AADJDE he_j_Page_232.pro
7c1e52bb86d22a01141f5b940ffee938
b2281b58db286e0f201b30e24eb953269136c1b7
3244 F20110331_AADITW he_j_Page_017.txt
8d657c24ec645b45c6ea64074b1c76ee
6a019365a7fde84f1db0630e55e8ce923c7cb360
114304 F20110331_AADKNK he_j_Page_074.jp2
e8db03cb2a4b0acd13fb45633970fa81
caf7fba6d939184a4f7d1795b45438f08b09378f
2294 F20110331_AADJQQ he_j_Page_066.txt
14a9a59a0505292c64ff9b0dbe45f50a
a19c10c53c281d3221b139d641b031f3ef24b850
49418 F20110331_AADIGM he_j_Page_021.pro
d4bf658abf7121b519874f51351995be
4d0a426562a1a24ac96949a17afcec0697679de8
F20110331_AADJDF he_j_Page_209.tif
44fed301e7342aaa5bc5b6817cd279b7
793cc7b5cd2abe889a9274249b5f7498eeab57cf
22078 F20110331_AADITX he_j_Page_116.pro
898638d6e0c916aeac5cfc5bc6a723bc
bda6c61be39a821afd64dc9f46f05fb9e4a6890a
101151 F20110331_AADKNL he_j_Page_075.jp2
3acd22c63546c672614d73d44333acd8
82e63f194182380ff705cc3cba7749a003daf298
2066 F20110331_AADJQR he_j_Page_068.txt
f555dd225eb8b3f6dd3fce9dc30eaca5
77f08568bc435289f73a125046832ae08d2bfc94
31810 F20110331_AADIGN he_j_Page_099.QC.jpg
90a406a1b1f4f36f8b06ec1eb49b6969
61b724c2b0d8edd61cb6049fb7913c0c94abe373
67419 F20110331_AADJDG he_j_Page_036.jpg
5da39092791631f51999736a68962d2d
6a6efc3abfb65fd54eeb5e361e0ebc5b13c2988d
503 F20110331_AADITY he_j_Page_001.txt
ea528f088cfd61286c27360cbd57fd93
02d929ebe22283e43db0e8061eb52736438b447f
2120 F20110331_AADJQS he_j_Page_074.txt
70ebef14e8715aaf917b222daa087dcb
cfdf587b481a7e174500b71fe8ad955cf4b2780d
15979 F20110331_AADIGO he_j_Page_195.QC.jpg
b3640bb2116bbeabdd79fca4fb990233
941757ab6903bd60465842d25e28a686e409b352
43817 F20110331_AADJDH he_j_Page_136.pro
118a1f44d692a82123f7b9d7d52ca41a
c9a394f47ab0687574c3d943d2795c023267f8c9
17327 F20110331_AADITZ he_j_Page_113.QC.jpg
41ebd636f3318ddc8ee241af78ee939d
788d40871fc1f2ba1ca7e93914a528c4be1dd6d3
51962 F20110331_AADKAA he_j_Page_134.pro
cfa3cdfe3bbc22eb4b4de32968bf62c3
7ecacf64f6c07e87e8e12a54069c5e688d44cc8a
112965 F20110331_AADKNM he_j_Page_076.jp2
fa9e10be5d5dd20c6223cd56a811fefb
3435ea061dca4a1ceaa2e56dc433ac476b9b0819
1732 F20110331_AADJQT he_j_Page_080.txt
a6c90751126d405e9fdd88e9df0c8fd6
5fbd5a36d3dd70d4b602afc75888a74a3febcfc5
30603 F20110331_AADIGP he_j_Page_271.pro
bc5eff7ffe6db8e95a1c6dfff6563171
f3bfc2a4fdb2b7b91a20828ea0e8910099a9d1b3
3331 F20110331_AADJDI he_j_Page_121thm.jpg
c8cc48858d539479e4b5a9a0235c8ca8
5352315707df1837a674641ed901ece03f954ecb
41696 F20110331_AADKAB he_j_Page_141.pro
ac966b7b9a5350ec41ef9d1031bb9827
9393751b90128a9a06136c4cd04529f03e8968d6
85405 F20110331_AADKNN he_j_Page_077.jp2
08061350ccf7d3eb81c6e8656fe6d280
999e252f3fa9ff057720e4ebaaae54632b067f15
2115 F20110331_AADJQU he_j_Page_081.txt
53a8ff64354890202896ebeb80ac4268
a3e5a7a594ad1b838078568cc60d4025641ac205
114301 F20110331_AADIGQ he_j_Page_097.jpg
4d653267f868d0e71e28537508f054f5
afa1df56b1bef6342286cbcf1a445cae6dc52a74
37011 F20110331_AADJDJ he_j_Page_044.QC.jpg
6f298b6a9e0797fcf5cfd9fb31456d94
fe396ea4c177aa1d31854a85b3cd013c0ab7809a
48277 F20110331_AADKAC he_j_Page_142.pro
601b92f5b9eaef481bca5f6878c6a16b
2e76b4684536d7042250a80fb2a21f7648ef32ec
93587 F20110331_AADKNO he_j_Page_082.jp2
6c648af84155f249563fc6612bdfeb8c
30180fa7286cb345fac299c970b4682e693d69ca
1935 F20110331_AADJQV he_j_Page_082.txt
ac2ac6352b02accccd86fe2430b747f3
f1c080158a15094fc753bb19042ff83b97effcb8
34551 F20110331_AADIGR he_j_Page_258.pro
b9afac670a8529e7b3bf678f988d532d
33e9edab1ed002a296ad2e20fdaba7f35176f28a
48405 F20110331_AADJDK he_j_Page_192.pro
22334bde44dc68f829a943e00dbd7d54
1c54b8d963c81cc49b8e1e0081200fd6a94b1e52
49613 F20110331_AADKAD he_j_Page_143.pro
566745788d976a72de96229445351f1e
c213f718c7a131544ea90329c2d90dfaedc19bf2
82065 F20110331_AADKNP he_j_Page_083.jp2
8ba828f041d951afc6db0939918a3aa5
2b8f818d44b61121fa826eca89ed6ba2ca7732df
1703 F20110331_AADJQW he_j_Page_083.txt
3e02547662d55049d8f7313dd6d4b8e3
b3d489469316dc340feff4260a83803e715a9509
111283 F20110331_AADIGS he_j_Page_134.jp2
eec08a6d1fb775d15ccfcf907393691e
88457cba1dc35e2af1aeb9e72c91656f2608599f
8750 F20110331_AADJDL he_j_Page_170thm.jpg
44952b05adde6ca3eb42148bbd85d974
5d9e5dd5b5335dba089502086c839a087db31f5e
52886 F20110331_AADKAE he_j_Page_144.pro
0776b2301cd7b59589e1aa92f71e556b
976f6aa8e4e2a5f71f4a2572e1fa33c3605ff7b7
78561 F20110331_AADKNQ he_j_Page_084.jp2
b04a144f14120728234007d6e87977bb
aab6732679b8449a375b689f87337e86ffcba1bb
1868 F20110331_AADJQX he_j_Page_086.txt
d0e293da309deb9c5e450d7be9cf7b21
c1a6144660a7fba61e31e67574ce1a43545e9847
106490 F20110331_AADIGT he_j_Page_055.jpg
4352b031b904d043760ee4cafe589b4f
63777721c349c850a8e699f9e50dec1eb22b33c0
2156 F20110331_AADJDM he_j_Page_208.txt
3bd68fedcd67325b60354aa6fa43dea6
3a3aa87377a224c620d847344e91991bb16bcec5
47827 F20110331_AADKAF he_j_Page_145.pro
44a706ad0083c496251720ff825467bc
5410e85a1352e72b26b81b1dfbd0d42bc356b9f1
91676 F20110331_AADKNR he_j_Page_085.jp2
aad9b196be1eb8904418f4d0862a5d05
631a997d2f1716da5ab0bca272d61228bb1f4d12
1835 F20110331_AADJQY he_j_Page_087.txt
e99ee023b0730ac296ff1c353d593088
c08b7ecfca0a9171aa4216d1618f791bee423b13
8164 F20110331_AADIGU he_j_Page_047thm.jpg
d7db3add517685f7fb792706216ce9cd
c07e7d9bf754f06f6c3861d873b7122b699870c7
8873 F20110331_AADJDN he_j_Page_316thm.jpg
f56b20ab283ae9c89e14afd724ef644c
b1bf8a2eceb90ee613fc9478bd1a80d5a32b0380
52948 F20110331_AADKAG he_j_Page_148.pro
6dddecaee98381b808b5cc6f323a8003
b0a4463a1564d8cf38ca34b3b060fd4f4bde38cd
92249 F20110331_AADKNS he_j_Page_087.jp2
543c8ca413f51001808792d9a3d53430
a3f95406a3974f104da3876ad4795bf67840d389
F20110331_AADJQZ he_j_Page_088.txt
3384c132ecc48f5c9bbda50a66de7f01
fb8f30f1240770c1d6acd04a6e0485b24fc51ee8
F20110331_AADIGV he_j_Page_141.tif
317bf413ae1c4ec1868ec817395e8322
355ed7788cd66de311b2d22acc883e44e1eafa74
8618 F20110331_AADJDO he_j_Page_076thm.jpg
11b15dce317315cd4b8bdc41e239858a
99e8e5b8d352c9803e1e399498e9ddceb5a2ee4e
54173 F20110331_AADKAH he_j_Page_150.pro
3019e0b60a5af3e1f66c763071aa955f
e2fb4d2d98909da1b9d0c77ed9978c4ea732a8d4
107463 F20110331_AADKNT he_j_Page_091.jp2
b278b2db7b6f512d5e7d6cb50661499e
6486521cd8da8d16f47dad7298fb340f4c2b5544
29995 F20110331_AADIGW he_j_Page_087.QC.jpg
44ae02a0591f6c7d1195d5ea2dfcaa90
f081b5f5319e9dcc27a1a76bc0c7492673b81b3f
35079 F20110331_AADJDP he_j_Page_068.QC.jpg
8890b92937f2c9619f8ebd91298c5a94
f37f2785ab803f7b1a64f0ef5436b2986fd34aa1
55177 F20110331_AADKAI he_j_Page_151.pro
b59718aa4d317527ec10c7aa2b7c30c0
4c2f03b0ced3891941f2c9116393a7866f2dd8e0
69752 F20110331_AADKNU he_j_Page_092.jp2
15a2c91e7c73c41e19794f0c1e693e79
4f0f99903b9decccc4f12ed40ae9854a6ecc2d5f
2804 F20110331_AADIGX he_j_Page_015.txt
330d642a3dd90b3f1bf159771c99c4b0
394e6acea58f10da9a02449a80af5f3c10612980
64768 F20110331_AADIZA he_j_Page_308.jpg
1fbd42b3ba9ebe8ce207de6bec545348
912bcccb147667372170077b90bde3f92a0c5a34
41188 F20110331_AADJDQ he_j_Page_112.jpg
6b19bd9866b3b961611a47354fbfd4e4
72d662ef01f5dda68fc23fd713a4e318458c8d82
34579 F20110331_AADKAJ he_j_Page_154.pro
0a10a55aa73ffd8ebe22340c6f1f9d82
b1ed501336b78f1f475fe22e164b4fbf65d16cbb
104123 F20110331_AADKNV he_j_Page_093.jp2
77fa515709ad1452d307b7ff5a401620
6da5bc45134ca2eb4f580aa99087096879874f4a
F20110331_AADIGY he_j_Page_270.tif
e4668937c3016ccc53fc84e5685b9310
e475c79c246c982afc53d970a967f9a9504bee6b
2188 F20110331_AADIZB he_j_Page_215.txt
285ddc5e13d13d8e77fca9d367edea95
c3a5d3356e1768339bc0598a231f8e6582586e60
17258 F20110331_AADJDR he_j_Page_286.QC.jpg
3c37d2ec2a791c4b9298cf3991884acd
8ea82820f6b364dc5e1be87a73614447db9bd529
21284 F20110331_AADKAK he_j_Page_155.pro
3a95e875a0b3acf8313cac9351af5649
135375ca0cc5d1d9a7563ea3087756d7b4e714f3
110913 F20110331_AADKNW he_j_Page_096.jp2
89440a3265952e2dda49a7be30bcb800
42ca8b751e2551a70ab7c231c50c1b76f53f85c2
F20110331_AADIGZ he_j_Page_151.tif
c0db4509288f142808cf7813c107970c
bd8faf369af0d870eb8b9d063cd25827b3fd3b5d
F20110331_AADIZC he_j_Page_123.tif
bb1fc713cb851400acac5f8c01295e5d
eb1cd253e56f2b9b1323d3b7772f6681100d7c6b
4297 F20110331_AADJDS he_j_Page_273thm.jpg
6a544a279da294a395398c3b7c4302f2
78ed02e8ba59c1acb2840a270cc3549892ed7bb1
14614 F20110331_AADKAL he_j_Page_157.pro
48e6b502fbbc9b2caf219b95e4e72e9e
4b6200ef34d51ece02137f10bbe369d6a4f6c2a0
119346 F20110331_AADKNX he_j_Page_097.jp2
5f840ec0ff6953f373891b3600cf132f
c05459c86fee1c92cdf25ecbd1b07ea380e52fe8
76130 F20110331_AADIZD he_j_Page_231.jpg
d8d3ee5395207d4d4c057bf6bb5cdbb8
d3ffe69d53e40680bc9ee37edb7b4ea41eabbce3
2053 F20110331_AADJDT he_j_Page_076.txt
a491f87f64d44b7f34ff8960d3583794
6d2db5b43741b97ee6a14a3d4f60030a725f9e00
16913 F20110331_AADKAM he_j_Page_158.pro
3dfddcc48590d10cd99f859f2519039c
824979ada753305298c14f392c852ebb52bfc2a1
102428 F20110331_AADKNY he_j_Page_099.jp2
eabc30f33e7b558017a0c778de61d54d
7706d1cfa5a188f5d831116782828556d81f8aeb
686 F20110331_AADIZE he_j_Page_310.txt
cb24d6cd6139c878ba02347be116460f
78ecbf0d88ed759d0924586d21ead427feb9869e
F20110331_AADJDU he_j_Page_193.tif
efa50d922e2b6a1cd7697bbcecf09858
5647115a4b6079dd711f5886cb57aac74c8c0830
19075 F20110331_AADKAN he_j_Page_159.pro
f35a507ad4ae085e5748317ed55fc7f8
368129ded81403e6c6fd7a830603b01fd8fbb287
118520 F20110331_AADKNZ he_j_Page_101.jp2
a6417acf84a67c201d362b21fe6974e7
df12327a058dc66b1bd591a2aac3749bec69c82d
1326 F20110331_AADIZF he_j_Page_290.txt
078434b35752bd6b732e190a6bca87e4
bb383f533cbc5c54d687938e5684f0dfae7ddb8e
20414 F20110331_AADJDV he_j_Page_160.QC.jpg
c1eebcc0e4c7fcf7b7fbd056dcc9b7c3
71455e21baa01f5ba9f722a3a8f38e35a896cbc3
43414 F20110331_AADKAO he_j_Page_162.pro
9dbd0fe8cc738c4a5d993db554ee710e
e47d072bbbe6ba4db6203cae72a7283e8eb53530
34871 F20110331_AADIZG he_j_Page_148.QC.jpg
46110708cd185ee418e1a6a8555cac1c
e289cd20ec4ef99cd4846006d2103fd350c3e8e5
9105 F20110331_AADJDW he_j_Page_011thm.jpg
59a13560e01b47365db69c92804f5d79
4d2e52f8a95c9dc9a5fab06f6334122aa965f7ee
41365 F20110331_AADKAP he_j_Page_163.pro
06b78d01f59e52ed4abb656788187f96
5b11c33d1a71a0ca31914aa02a56b71cede51aa1
F20110331_AADIZH he_j_Page_007.tif
da13e94f426dd34ab8b45cc74d4b7d34
866c13592bbc3b18f322e6c6d7a93ddf5bc3d608
70933 F20110331_AADJDX he_j_Page_273.jp2
7a21a24d6e5af5725ab8769ab375d03b
1517bdefd8e33809c7a934850313b68ef0c1cc1f
572 F20110331_AADJWA he_j_Page_292.txt
1ad721c4294f3ceeef2bac9aca3beefd
aac1f2703dc4cef6ac0647186ed291c532320cab
34246 F20110331_AADKAQ he_j_Page_164.pro
8fe2ee63abe512f6d459bfdc5f5e12e5
d313628023d788d81970e54a8d84fa76030441e1
112928 F20110331_AADIZI he_j_Page_133.jpg
0f435ab1ee93d2c1523ea7d23c4a9530
e455c3a524c57e6d56efd4b4172fc956f7d5f391
556471 F20110331_AADJDY he_j_Page_157.jp2
2833e8e5d850e7e1f8b6212583d6117e
7312556d0ddeae5cbe59d82dcdcb775b9489c5a5
170 F20110331_AADJWB he_j_Page_293.txt
cf3ddd248d98aa7c450f26eea5a0a6a2
f2ca6c397c837cb93199993364bee55d57f31bde
56741 F20110331_AADKAR he_j_Page_165.pro
28d82356a8733527c6b66417cf46a977
a7d1540dc3188c7afd6866609a00f7d57a92f625
1846 F20110331_AADIZJ he_j_Page_075.txt
4eea1b2043f99050902fde2655bb9658
1983a9475ade25f171a5dd654ae7312a9db507e1
F20110331_AADJDZ he_j_Page_326.tif
9d55b0924f650f93110d9b8160cae594
6f9e9db0add9ff253f1ad986d8e0a55ddfc926f2
204 F20110331_AADJWC he_j_Page_294.txt
ea968ce8deddcf9bdb287bd2cff097a5
d4bc47c58658aff71358e2ffddc96e30a8c2272b
45673 F20110331_AADKAS he_j_Page_166.pro
b537c5fd9f6ce3e69faa83f076151a27
6ed75a31182fd0a6753cd53e5920204f5c014879
5448 F20110331_AADIZK he_j_Page_200thm.jpg
061191a8695562af3270c6faaf31d30c
a651d843a6fa43d715cfd7feb1e15281aee3c1a8
161 F20110331_AADJWD he_j_Page_295.txt
6bf265a02c941ce05e6c3edf25c7a01f
cb02e66366b9df0187a4984cfa70250fb0326240
52268 F20110331_AADKAT he_j_Page_167.pro
63bc6bf9d51a6e4eb6b4ee29b100a9c0
353a4f685bdf1d418f4d2d429107890b48655f27
66579 F20110331_AADIMA he_j_Page_257.jpg
915337fa3bc4e6426942762e1fd2e2f3
b32a592386bd4634cc642e267e31a05252c358e9
51813 F20110331_AADIZL he_j_Page_023.pro
f0bb1923472f212b4df27eaefcd83342
f970714c7d218540ae87f3fcfef0b5bf463ae636
666 F20110331_AADJWE he_j_Page_296.txt
c41a51530dbd734aa20606929328a929
be58915cd5a7a4144ffa0cb16c9aeb0d5e9a2d3f
51304 F20110331_AADKAU he_j_Page_169.pro
bdf8b4c14863dc0d680d8c9be3e5f0fb
ace03571e1c9e3b67c6def8a90b9bd075db40915
1947 F20110331_AADIMB he_j_Page_048.txt
9f65f16ec346434b0eae114e06926f0f
24a0f90014532f9f61ceba8856d6da2d19f8c670
F20110331_AADIZM he_j_Page_223.tif
f521d5e6a9c47598ffc41731cdaeb444
f4ce54c6ef69178185c537f6d2932cb24b703a09
136 F20110331_AADJWF he_j_Page_299.txt
f295544acd5bc574c98327aeb4bd92d2
0d6441725116696b5278a29236f9fcde829590ee
54077 F20110331_AADKAV he_j_Page_170.pro
13757a7c104ab72e82d5a2072b781274
781da2d87ca86433965f36a5164ce390a4bb9bfb
F20110331_AADIMC he_j_Page_046.tif
ca6c007d59bb32254e82116a5e8867a3
23ed72398a93d54fa4d0cbaf31fa6ac52112ac13
26245 F20110331_AADIZN he_j_Page_137.QC.jpg
0346f9a8a769e5f7d142201bb4a9c246
b68fc58ab0f976adf376880bb895364783457c33
124 F20110331_AADJWG he_j_Page_300.txt
4e86147352fd333b60c27626cc5c70c6
34476eed9b7899f3c26288a68003054809c1dfd5
50612 F20110331_AADKAW he_j_Page_171.pro
fda4db51f783dfb14bd48f0d766a38c4
f9dca4a8e7b05d5c103f1ff810a1cfb3fe432fb9
535177 F20110331_AADKTA he_j_Page_311.jp2
85b10b730af8da7ef90c9135d72f486b
6606a24faeb2ee627faa563842d18ce765572e65
54367 F20110331_AADIMD he_j_Page_026.pro
f0f9abd1d6acaaacf0119787035892b3
dee0d1ec42c36d867ff089c10fbe83751477b108
7964 F20110331_AADIZO he_j_Page_302thm.jpg
071433fdf7a443b935d1bda3c5c02d46
0664dc3a232b720d0f3625baf2883d079c4e551b
1068 F20110331_AADJWH he_j_Page_303.txt
f2c1eb1b57505382e9dd51d792d4d4ba
719d11548694a651d79b0e207caad814a875b069
48328 F20110331_AADKAX he_j_Page_172.pro
1087464811ac56e8bc6bf1f4730f22bf
32c524f1d1d9fbc2f2317d728e530c84dba93e22
577610 F20110331_AADKTB he_j_Page_314.jp2
4423c0b1bd78cde9e5e58bed94a2d816
3cf3ad9f4c2179b6f91e9ac0c98e0a8467db7478
8801 F20110331_AADIME he_j_Page_031thm.jpg
d2636e678795921769ce8299dc7b60a9
5bf9ac1a0317b8a1fab39840fe15ffd2f3cd6c87
33400 F20110331_AADIZP he_j_Page_289.pro
76e2638ada19664263485deafcdd7899
e91f3cce1c4184e962ca895ab8cc3688535031b2
101 F20110331_AADJWI he_j_Page_304.txt
1a5dd7459a26927123398587a6ed6b56
64b6407ff0593287e5a3a3475ec67c9e4361a0b9
44292 F20110331_AADKAY he_j_Page_173.pro
e06dfa18b617c814a9496fe94a11301c
3f0413b9a4c32d63c00d4c044695ac39ba7c2335
568219 F20110331_AADKTC he_j_Page_315.jp2
b7cfdb30f5872e9dbbca928e2a8b91d1
1a7142d12639e7bd670e53dced08d056d36bb403
109093 F20110331_AADIMF he_j_Page_208.jpg
9da0c293d8e07bd3502e7362f4ec605c
8bb81a983e84e270d456463d3d5aeba02d627855
8079 F20110331_AADIZQ he_j_Page_087thm.jpg
1e328f78622794a1a1853308d802d949
b5e57fa44dc4e24c72835a02d65a538c77180bfb
881 F20110331_AADJWJ he_j_Page_307.txt
af548957dfbb709b33dc9f513830a150
f76885bc97fed3efdfb6bef68f32e6fc21cff599
56406 F20110331_AADKAZ he_j_Page_175.pro
445bf6ef8a1fb8990735d7ee38eb0154
f1e440a59120ea161302f34d933a746417cfa3b3
132578 F20110331_AADKTD he_j_Page_316.jp2
c44a0b51dbae735996d946910771bd24
475959f259d2f7dba9203c3fd0ae57842f0441c5
8825 F20110331_AADIMG he_j_Page_240thm.jpg
2f78bd39c1809699bc2947f6532b683e
1c2ed965bd32e8326d05929ec70454b9a2b16d33
54496 F20110331_AADIZR he_j_Page_254.pro
dc923da025960a1ec829c7b73381a3a7
cf56bafe76ba0b7eda5fc9ddb94de577b6b1ce57
640 F20110331_AADJWK he_j_Page_309.txt
936003b575e713eddba9b1ebf2120683
506b9882ecad4e41283c94fa55e98ac5f98ff94e
1051798 F20110331_AADKTE he_j_Page_317.jp2
2eb8aa57c8c8ed7f0f02d67003697bb9
bade72bc0b0552cd02ed675822d3cddcc2231e32
F20110331_AADIMH he_j_Page_031.tif
20b031b13816b5131ec08d09aee512f2
d7cce93e2457284c81dad1ee839f55d476a71015
F20110331_AADJJA he_j_Page_047.tif
6878bae25986af67b33e6a12b1c47c74
d49405b655951f72e98f2089d15af71d0b0af4b0
5959 F20110331_AADIZS he_j_Page_063thm.jpg
dad16474d7905f1a639c20bb5162ef70
dc4d23df066600da43ef226dc4137769a4eeb9dd
808 F20110331_AADJWL he_j_Page_311.txt
51964a14f7b2954e7677caf5a28e9125
d436025658c1745c32f8d5039ba11102abdfda39
1051967 F20110331_AADKTF he_j_Page_319.jp2
86239536ab79041486d77fd6bb96c873
9e43aae7e76b0e0e31f7cc3f34020f5b6abae62d
52185 F20110331_AADIMI he_j_Page_135.pro
4810f7f7cfde4e32a90beb28e4153a1d
4218a2b911b107c5fec3d34fc5ea1cae1b6d6764
F20110331_AADJJB he_j_Page_048.tif
8f51f73787511f864649f040f6a431b8
303fadeb144863ff3f25f6020f3cf3e4cabf0407
25438 F20110331_AADIZT he_j_Page_303.pro
4317de64c1ef0dd155edd851a6d3177c
4ac860b06a9bd73af7a80ecb5bfa6f970a79ddd3
1165 F20110331_AADJWM he_j_Page_313.txt
d7d53cfd008b53e313eebc4507ad4254
7d469ab08103e1a6f4ec3e582e583355be590d6d
140143 F20110331_AADKTG he_j_Page_322.jp2
fe0b20f8e782e02b9fabb6e4d07c1ea0
7ee146f0fc2c650a84dd99a087997b08ad57dfd8
1195 F20110331_AADIMJ he_j_Page_128.txt
11ab6f811e2d917f522a56a1a6d10838
b41d9781f803e47399e30c2bc3eee48c40c3b236
F20110331_AADJJC he_j_Page_049.tif
cfa92de205ab0f41867724eb222cf800
d1f96ea4ac52a7f4fcdb0e486a65b3654d84350b
F20110331_AADIZU he_j_Page_101.tif
18d5332acfdafa0ac09c6882ea25ea8d
cf8041557af2e7e28b07085674527d70e7cc90a5
736 F20110331_AADJWN he_j_Page_315.txt
a5a245431a88ff05e7cc6d4a00b5e1f7
4a242bb6d0878260e9c2cbdba441b5755188c0c3
132181 F20110331_AADKTH he_j_Page_324.jp2
d5275110b5e4a3ed8fb2a9a391e3f842
d7d42d9d2f7f1eac12ec0736603baab9d8c254dd
57343 F20110331_AADIMK he_j_Page_314.jpg
20289928a8b332b323e2c0dbef7f2529
6b443191a46c02da6ff371ddae34374beccca79d
F20110331_AADJJD he_j_Page_050.tif
56c2fdca6bc4c8360ff083a1817fb8db
7b173b1c3d5075efd19f62d272f61ad6ebc6e8f7
2785 F20110331_AADIZV he_j_Page_318.txt
b87d449dd84a70f639805ebfdd80c560
05ea6789d89f244086fe68abe54b9dc7dfb94621
2987 F20110331_AADJWO he_j_Page_317.txt
1ffea9256e7ab415729ce52c115fbf3b
e70c3b8eab871e68339f56cc13fde5604e087488
1051966 F20110331_AADKTI he_j_Page_326.jp2
f864361f9dd1dfa81479f9950987999c
a11c5514879f1b348a2bd139084372eabe10b4f8
1051969 F20110331_AADIML he_j_Page_320.jp2
58a1e80f08d8a8c07cfff546bd0c3e55
c243be45c40ba240c27d14de8ecc35738a332aea
F20110331_AADJJE he_j_Page_052.tif
ee053ae6d35f799dc3108e4c373bdadb
71964ccd6ed569cd2b5e431981bbec96f76516f6
50246 F20110331_AADIZW he_j_Page_139.pro
83e7710ed3c3e270617a2f40c75c2987
1c88f6f418f50c87cd9411c3efd73feb01b666c1
2630 F20110331_AADJWP he_j_Page_319.txt
9111900458f7db220a82568877ff6928
f3b734c0d923e9733a07ebe2c83b8924d4a837d7
1051981 F20110331_AADKTJ he_j_Page_327.jp2
0b96df6228273eb41e3d412eb18290e1
09987b3fbe334c4803f87346d0e6226a4ed91284
4780 F20110331_AADIMM he_j_Page_291thm.jpg
d7d7ea59b8a8bc221592145f3cbc0b52
ac54c19371b0da8e0b65d0ed0aa0dd56f2f05ff3
F20110331_AADJJF he_j_Page_055.tif
f8b9c843f085b7bf2393058bea6e56e1
1c836a1071f4b4296d0f0ad5b7adff32f56ceb0c
1389 F20110331_AADIZX he_j_Page_287.txt
d25636674be3a5c1c2b6614b253b3837
e3f5eb2e732df13ba4ecd9a98b5b6e5599c43475
3053 F20110331_AADJWQ he_j_Page_320.txt
0369e4bdb6004f52a66d6138bef9348a
653adab1ecc12e62e8034c72119f4bf9e7dee498
604367 F20110331_AADKTK he_j_Page_328.jp2
3e3f79047590619753632a44f492771b
986e40ec3536618838517fe7e181b47eb37f0475
F20110331_AADJJG he_j_Page_056.tif
057d7f6c250e05d031307fe61645b3e1
5e9f6cc43f17913d18e7dde2d1abc6b298d80a1e
1838 F20110331_AADIZY he_j_Page_042.txt
39abb7fe80e3e9eaa7ec6ce6ffccfbc0
794240336b6f159efc3bead545b2a0154f2cd8c5
2906 F20110331_AADJWR he_j_Page_321.txt
6fde82b7c423bbec3ff72f849ad4d319
334ff686018eedad589a271fa0a7a0f8595eed7d
8208 F20110331_AADKTL he_j_Page_140thm.jpg
bd7f88fcc576f671770d3a55a40aa29b
0d865bf6101eef1d8b60f67c9e68054c5add4601
8284 F20110331_AADIMN he_j_Page_029thm.jpg
c0598b9b6f80dfc342feea5b2f300153
638020443ac9b8e71cea668beb215e19ad541d1e
F20110331_AADJJH he_j_Page_057.tif
5d238e85cbb12f9730bd32653135e8e2
84179a59e7ad8720cfdfd59c4576e2a5afda2a08
36367 F20110331_AADIZZ he_j_Page_121.jpg
dcb2adce71d9ce1c57606bcb628a1dbb
b73c938b24bc9394bb30db8f61102dc0c8981e51
52166 F20110331_AADKGA he_j_Page_066.jpg
d835313fec0eac9d32ebce1c30d8f18c
8444fa0c5ad056fbae91b522e3493bcfebbcbf05
2745 F20110331_AADJWS he_j_Page_322.txt
87e82e4f8f31e4f6a1c95bc586073390
ff4cc4626cdf801b28e7a07f085c6592f431a89b
33088 F20110331_AADKTM he_j_Page_004.QC.jpg
98258eec250e512bc6ec9ce06a11fd62
bafd39cbb71ed4fdeda09e30f9c1eebf905640fe
17057 F20110331_AADIMO he_j_Page_194.QC.jpg
66f16ac51c695c347a5af8b4fd32da7b
7a7c0d67951438ff713debe5bd63c1fd4fbc308a
F20110331_AADJJI he_j_Page_058.tif
d64bb22ec0c327ee11852449efc2c23e
f55fb47bcea98abb424c50237ae8829e1d837bcc
105282 F20110331_AADKGB he_j_Page_068.jpg
246976db814923d71fdba401e36db3a1
4a31fe2bc410631497cbdb5ee26e0a4d8d6d443a
F20110331_AADJWT he_j_Page_323.txt
e81a04ef61513b3b6e94153bc5422cae
20f290d5af5da723bce2f2e5696b1ebaa9dc1ecc
8753 F20110331_AADKTN he_j_Page_044thm.jpg
a3e257cbd35ca369f2502d918782bba8
0bb3de9527cda3371dabbdede9aa3ab2c3b215e0
1442 F20110331_AADIMP he_j_Page_271.txt
1f8a475e56ec74b13eb87843e41e817f
8d05dd081809dd6b21164f87237994b0eef1ffa8
105775 F20110331_AADKGC he_j_Page_069.jpg
579eddff72b7bf873aee0d463438358b
e6c9d2fd207b65879dadeac3ade11d981ff92625
2610 F20110331_AADJWU he_j_Page_324.txt
8b93745d870bbbe9975103e255e0ddca
c81f5e8bca9d2f412ed77edeea63802688342349
40655 F20110331_AADKTO he_j_Page_327.QC.jpg
c9785eac855b9ff88752e3203b43020e
1ad643f630a79bf187544968bb61d756cb2f3ad5
36182 F20110331_AADIMQ he_j_Page_146.QC.jpg
1f5ec4a0e1b1e8f846bd4dd52900985d
68d8ed51e0472e3b6399e9e42821aa85d9585f23
F20110331_AADJJJ he_j_Page_059.tif
18939496986d3bb790faf2c98c3307dc
0641531bdb8f3a25a73185164bf8eaba6f858e2c
108767 F20110331_AADKGD he_j_Page_073.jpg
7fb52607a6d77d9df56d8ab26c8e390a
28813db6b81bbc09bc3a89d663e2ef6d11909c44
2980 F20110331_AADJWV he_j_Page_327.txt
ce9af88af585ecb5c4d8c2cb44bca94e
cc5803681064d13c49b6d70672016f180a5dcca6
4835 F20110331_AADKTP he_j_Page_233thm.jpg
cbee99851e6296eb3b8f1db5fd4fefc1
505f5a40d5f6510392141ce35f125fd66cf73645
F20110331_AADIMR he_j_Page_138.tif
9d6411883db52c8f1f253636235af90c
9a550adf572bfeec00c33851e6887b97db7677e0
F20110331_AADJJK he_j_Page_060.tif
ef98cfb4db2fabf1f7bad15967c18cd2
a6a92c3e3ea6576d46c406c03a847af166fd5745
95104 F20110331_AADKGE he_j_Page_075.jpg
95a553c99c780ba2906e6cfa882cc66e
848b64453855b3f9c3d663b40bd0bc65b4730f75
8964 F20110331_AADKTQ he_j_Page_215thm.jpg
b70027bb3b69c6bc2313ff926cd48429
e702780702af85a5151715168fac829f118541cf
8940 F20110331_AADIMS he_j_Page_022thm.jpg
5ef5fc4e6dfba5c271688a2d59c1b5af
8a85a9faf1e5c7b522994ba9e2ae6bc0cdcded55
F20110331_AADJJL he_j_Page_064.tif
300c32716588e8d8281a609f163b7acc
1f0d7f1df7c228fd191c9de6905bdbcf69ca8e5a
8814 F20110331_AADJWW he_j_Page_001.pro
a26897c2c89ce2ada5317854f754cd2e
4e85d55ee7917ed55969cb64c0d71bb7e62456dd
4066 F20110331_AADKTR he_j_Page_266thm.jpg
3d00584aae85ee81f59e09f1eb1c31cf
b05f6477a2062e0fef4ec0b8c3b6b17bc226e0c4
114555 F20110331_AADIMT he_j_Page_023.jp2
94853de9eb62267ba3755b88fc3d7b9c
f97cb1c424b622f86caa14a846122ab89749f51d
F20110331_AADJJM he_j_Page_069.tif
e49465f2f5be956374581b6d0e175b4f
f47aeba5a727e24d4827e9b621567b3e8c4531f3
82038 F20110331_AADKGF he_j_Page_077.jpg
902f32ab20dc8803a6173edae1a5f0d0
78e56d3c7fcfc6590dfdd506948be43f6cbcaf03
2492 F20110331_AADJWX he_j_Page_003.pro
1d8d197430dd55aa98658a6b5ec64d35
9c7343dcadb4eb8885105481d4a7158f6ffd0f38
38413 F20110331_AADLDA he_j_Page_322.QC.jpg
821155e76c3191697c7582e292d0f61a
ef22b0d7b8d14a02d70128ca3af93f81374b46c1
37411 F20110331_AADIMU he_j_Page_177.QC.jpg
65a32dc187e46cc18768b8f290f408fd
369b6faac6e05cbbad1ea725333566458f1bda1c
F20110331_AADJJN he_j_Page_072.tif
1136f69031a3e362a6e86c45e86a655d
859b8729b3885c83d0f97423963535254160b9b9
96644 F20110331_AADKGG he_j_Page_078.jpg
322c8f1ac784e5a4616453d767eb65dc
42e78ac23767d1bce92a9e0196cf5395d31248af
49119 F20110331_AADJWY he_j_Page_004.pro
3ac12e2e878c16503398ec154cb8893c
bdedc413f2f041483d630150adb703c86bcdd0d3
15289 F20110331_AADKTS he_j_Page_206.QC.jpg
e31187d8a9de706137e79620ddda8068
1b0474905658e3e148af8bb3e240c6fe4c9b9392
F20110331_AADIMV he_j_Page_302.tif
37bb1b81d0ce05c51b0c41024255f923
4af667e95f31de257ae140f1197ece716b6ffe75
F20110331_AADJJO he_j_Page_073.tif
0d6b074463a4f897378523c6de5a8de4
cc2da6bc94d28d9f886ff5d5de1af312af8ec20d
80550 F20110331_AADKGH he_j_Page_080.jpg
00a84bae6e59bbae7a271671ab3861da
be0083b3d1b56059c8048463cd9378aa9b3a228f
24249 F20110331_AADJWZ he_j_Page_005.pro
38c6cf64d49e4d805addb72862d8b282
46527bd7c70ac4608997bde959d1db73d1ad5ac5
38402 F20110331_AADLDB he_j_Page_323.QC.jpg
ac6cd5c5cf8dc4e36b33e39bc45b62d1
ba3bd5cf649ad5a523190320f95fa8a1764479b2
36932 F20110331_AADKTT he_j_Page_294.QC.jpg
d7e2d868a3d804125d34b6c1d86d4bd6
2de75d177599790865f70aab841ae774c4e1933e
F20110331_AADIMW he_j_Page_173.tif
6513c4f9c6f6f3e4dc903a586ba5cb8c
fe5f4a7c8de7328676a63f2374fa1cbbd3714472
F20110331_AADJJP he_j_Page_076.tif
d4032e0947bca2a80dd32fad53ae987f
1f9903969249d9e275bf7973dc9bf94a9203479b
88862 F20110331_AADKGI he_j_Page_082.jpg
96b5956d05ba9d576a0d5357dee5b6fb
f0be222593d0299d4e368887f2b04d0f427fcf6e
36366 F20110331_AADLDC he_j_Page_324.QC.jpg
294d37d19f07e18263761373ff78fd91
068157174905df385bb1f6169dc0fa62b95bf67b
5898 F20110331_AADKTU he_j_Page_259thm.jpg
b36af4c75a74707062a6efcd16305ef4
884f8ccdff2d8f9c3725dbd66adadcab01da96fc
36841 F20110331_AADIMX he_j_Page_277.pro
738794482ba76e6ddc2bbbd699886bab
84c3deb8d14ae6bcf48153a12a0d2d469dda0864
F20110331_AADJJQ he_j_Page_078.tif
ff81627d63be5a4e5cd1dae00605f254
20c545e66b067606e6827d400459173c9187a806
74139 F20110331_AADKGJ he_j_Page_084.jpg
78dab8b1d112a64923268afd3d583b2a
dc08afdb83f322b2654dc05ae63be0be85284fb8
16529 F20110331_AADLDD he_j_Page_328.QC.jpg
f72c89d7b9c6345205a7925262ef64cc
2e27ceee7ca7cc1ba1ee41efa3ca2455be6cef05
36626 F20110331_AADKTV he_j_Page_150.QC.jpg
75bf62d180c3e99ab0cfb7bbbf730e52
0620d0ae90cf45dd521ebdc672eed9e3a9da7c84
88904 F20110331_AADIMY he_j_Page_019.jpg
d877463b6deeed9d1be9b8a7a5dbc835
6ed8f0b76379bf65c35f55997054123151e5eda1
F20110331_AADJJR he_j_Page_081.tif
1ddb2436cede1a35d941feb999406c09
fb8bd96bda6ea081b4aceeb82d1bdfd40d4311bc
87470 F20110331_AADKGK he_j_Page_085.jpg
0a94bd4dffe92915da0b3fd36fe2da93
f2f0777ced75a1e3c5fc13d07d217bc9dae081b7
2026 F20110331_AADLDE he_j_Page_001thm.jpg
1ba1497455948cf2c51ae59fe7795b38
fd1ba346998d6d43f5dc402e2b0b8edbb96cc59e
41749 F20110331_AADKTW he_j_Page_298.QC.jpg
fe75b464be0debbd443c41b1d8cc63c8
e6d7e95de70341f7d213528292fa3c2ba41e3984
5686 F20110331_AADIMZ he_j_Page_201thm.jpg
655ed5c50e91b9b9a12ad45735267e94
c1aabcaf3597e828b67e442d51bfd3cf50569d3f
F20110331_AADJJS he_j_Page_082.tif
eb6775c31b9d51cd535ec12914482ffd
4929107cbeccec48e044c20b97831fd2b12735d7
89608 F20110331_AADKGL he_j_Page_087.jpg
6c0ebe5ff82515d819c2cec7c7333b0b
dd24dae4dbe942b179b3d7262a16a1f872e61444
547 F20110331_AADLDF he_j_Page_002thm.jpg
3a6a239cde7ef786f1c572d1f4c66930
4c9d895720b65cc979fe6128e6e10b524f823a77
7571 F20110331_AADKTX he_j_Page_085thm.jpg
d77c14e40e3edb1203f6173347c4fe89
dd5fa7b99ec0a8a351ff42ffe85da5243a63335b
F20110331_AADJJT he_j_Page_083.tif
c94f5d51c6526d4097344a85db4cae21
dade6f445c2ec731a9a213e4d55946341c0d5371
77077 F20110331_AADKGM he_j_Page_089.jpg
fbbf709d88cecb1c3e8e1d93cedd2c8b
10296a94e9f9f9b23936a43ef36f573bd1cc4fc4
7216 F20110331_AADLDG he_j_Page_006thm.jpg
37a2d2885cba346b245658e0c54f04fc
64746b47082dc7522a71a9999ebccf8f3010b638
4843 F20110331_AADKTY he_j_Page_262thm.jpg
a2b5c4154e352b8bbc068444c669f2de
beebd46977e480263c05290622a8b06e203fabdf
F20110331_AADJJU he_j_Page_086.tif
5a1fd82eb403ec639f70308ff69620c1
bed29442a0741df23e6360e20603e53bde586dc1
85061 F20110331_AADKGN he_j_Page_090.jpg
957710fb9e6de7aee47b485ad0982967
ee9ceaaf8b560ef5a13aa100a2d695b803af6cbc
3972 F20110331_AADLDH he_j_Page_010thm.jpg
18b593519723527494052ed6773735da
29bb8a5648e5a7b7b887d6be4a52b762ee40c0f8
36599 F20110331_AADKTZ he_j_Page_176.QC.jpg
cfd2b044eeb1f8b0cc1e51c66f277382
3911a3a2befce2ac977db6abfeb566de41c5e223
F20110331_AADJJV he_j_Page_089.tif
c1a4f1d92dc5507a310ca16a9835ccdf
da1611164d9f6e6edb1997a4f8cbd1e77d935afe
100476 F20110331_AADKGO he_j_Page_091.jpg
368fecfda7de020c03f7aaf4a4b7d6b7
4f844d2f118ebeabec6bdb5c6898a6d1042bdef4
8775 F20110331_AADLDI he_j_Page_013thm.jpg
2930b3837a6a9b7c519e8958d6f8e088
f57e8565fc1be1e489b2a47390ce57f4ca44590a
F20110331_AADJJW he_j_Page_091.tif
a187ee36c7cb35b15aec8d2bd2c098fe
9ffba32a81e986a5211f945523e3da4f18747151
98079 F20110331_AADKGP he_j_Page_093.jpg
4a9a1d600f5a4a46f334bbda222e2954
ec1e114d7214008b35bef77790602828768a1b19
8055 F20110331_AADLDJ he_j_Page_014thm.jpg
99f06c78364eeec1813fe607acf0d78e
b554098e7e73b50d578f70421ed5980759ca2989
F20110331_AADJJX he_j_Page_094.tif
22b003c12e2f61eaa89beacd8f0790f2
0399ab43ef22aafe0616d6226388b3c4ee14031c
92890 F20110331_AADKGQ he_j_Page_095.jpg
b8a373791bb146b2706f15d2d85d5172
00e32a526c4b5d0b5e469beaf72f91a33e45b029
8021 F20110331_AADLDK he_j_Page_017thm.jpg
c96e2f8416f7e0bd790157fc81e630dd
e9ea7e79789e1c322b16e019c462e86b66036f3b
F20110331_AADJJY he_j_Page_097.tif
7506dc464d8f615e28ee1ee575413e0c
2fa7a51522c561383695a01fb39580e4850cc758
114060 F20110331_AADKGR he_j_Page_098.jpg
e8a2121364520996365d17e15436bb58
0a7d059aa8fd5a3737cea9aecade1ccb809bf18f
6951 F20110331_AADLDL he_j_Page_018thm.jpg
4ec3c488047d276b42320682d1a093fb
23b7f4a556d698c5f1d16d7ff38a89389551a55f
F20110331_AADJJZ he_j_Page_100.tif
c3025e1aaf79db48da4c183461907c99
89d3854b161076eed3e1eb3b3cc44019bd0fd6f9
102189 F20110331_AADKGS he_j_Page_100.jpg
9283544d9a2b45afd4ef1cafa6a336e1
bb13059121aaa245e922f2d8405662c8563e7b8d
8022 F20110331_AADLDM he_j_Page_021thm.jpg
485ed91d64bcf1af380cf1537761e6dd
05824cde6f89edf73c9e2412b3a75a9420fa9e12
113618 F20110331_AADKGT he_j_Page_101.jpg
c15b8618e638e3a2e71eadd4bff140be
00dae40a22defc698083afdfb0d344dac80132c9
8930 F20110331_AADLDN he_j_Page_025thm.jpg
a761685151e6d47c4dd347e710a7e42c
b2e18b01985a36d767be67b70fa92a7d7585f6a6
103243 F20110331_AADKGU he_j_Page_103.jpg
586b65d0551b2adc0dd47d59a0284f08
8be6811adeed474c15e87e773036bb322e4cfd44
4893 F20110331_AADISA he_j_Page_008.txt
fab069dac42ca352ce4d483023efd53d
b8118d6d0b174779fc8ccc9880bdf9af08472dc5
8644 F20110331_AADLDO he_j_Page_028thm.jpg
8f6e3edbc63898e71f5e21ab75b90463
5d1f91bd6e5e08fff84335eae5d5fdbcbdfc5ae1
106818 F20110331_AADKGV he_j_Page_104.jpg
675d9e87d03d89066d0b9829b238f31b
922f4249e8594375b87bef7e7f2dee4ab485ff33
41315 F20110331_AADISB he_j_Page_127.jpg
e74b65600096a62292cb3c6311c753c7
edca4d314961b1139eb43261643785bc2f252414
9243 F20110331_AADLDP he_j_Page_030thm.jpg
e62f1d749400a5f69b471ae29995f95f
0f652609ad83ec3f2fa860254b8660b5daa5e01d
107706 F20110331_AADKGW he_j_Page_106.jpg
e8b12313c24a6200c12226250a765c96
1fd412429e90cac959a47f46bc4d9339c06c74ec
15392 F20110331_AADISC he_j_Page_313.pro
12b8111a96b6b2f16951ef274ebc574d
6f53a99d0b43f388117b80e646a3899d5ce44cb7
8921 F20110331_AADLDQ he_j_Page_032thm.jpg
f2c754a2111d2bf4bf7c62b6f859d809
09f43d9931be9aa04e1e190fe6b97c6776de3b59
34284 F20110331_AADKZA he_j_Page_076.QC.jpg
b2ea5014cea2cc7d68e37c02c2425514
052d7aeeb15600aa19a71f99206d8997327c2fbc
106159 F20110331_AADKGX he_j_Page_107.jpg
072744ad5a6e1527198c7d5a3c5bdd81
19685337d07279b9d4b7db57a073a0829f37f6e5
17511 F20110331_AADISD he_j_Page_005.QC.jpg
5f055d65281aeb06601facf86726938b
5707f1bbe8aaf890a7edd3578c413d115fc9f37a
8507 F20110331_AADLDR he_j_Page_034thm.jpg
da785b2c4bf8beb77702430bdc63e32c
5bbe52a767be5c6f22c6fd2c4486a9efc82dfabe
22534 F20110331_AADKZB he_j_Page_084.QC.jpg
1fdabf953052e9eb0cd32dd33ac08d4c
945a09091329f9b39b85a0c14db5029976585a2b
52353 F20110331_AADKGY he_j_Page_109.jpg
9b4e7cd5d434db4f973ef7a63daab53d
ca0582e039e56e3fb38ac357e554ee70c3c6f995
1922 F20110331_AADISE he_j_Page_221.txt
1d641cb25d54909f7ea2e72c62de3b14
b60f261dd985794fc79d328c7996aeaad3fbc7a5
5538 F20110331_AADLDS he_j_Page_036thm.jpg
0e189946f915002a969059ac9b8e675e
e58ca3db5e1a312e53f7f1da7b15e95f2c6c2a11
26226 F20110331_AADKZC he_j_Page_085.QC.jpg
c3089f37a79ec595181ff76fa7c0e899
594f97e71e88c088b4bcfcf8ced4cd8ad263a51a
39503 F20110331_AADKGZ he_j_Page_110.jpg
33debf4e8c906ca56c7d37e6fbbe5f13
3fcaeff7fbb32992d842012e765d830f0d248224
114193 F20110331_AADISF he_j_Page_052.jp2
b9b91d9f6c409aee82fba34740b44c97
b4c4efda36c02a62a24aa349a93101c9d88ea66b
3075 F20110331_AADLDT he_j_Page_037thm.jpg
40fcd738ca83d71c6bca30ce60bc862a
e410912d4e9d5b0d684737691ecccdfb5f5993bd
28883 F20110331_AADKZD he_j_Page_088.QC.jpg
20fde2f85bee508d8a78067fd7b326cc
4bf0fab1ceabb7b0f0e03f1e05408d242bae80fa
28039 F20110331_AADISG he_j_Page_086.QC.jpg
f494e462a9606747743696aed92b4326
611640211d32c51e5351f2be65328306b73c2e3c
7273 F20110331_AADLDU he_j_Page_042thm.jpg
98a1a936b3eb9a4eaf70aebd1f37dbc4
92a9297b851a76c5fc4fba93e0eb848f806fc5b2
26516 F20110331_AADKZE he_j_Page_090.QC.jpg
e21228efa8dbe8f52463f22f5953c064
43f95096f469ab4b7cdb0896bc528674f48b0bb7
F20110331_AADJPA he_j_Page_324.tif
2641e4f90878a1d5075075aa69ec2e67
57ed2973376c4bf7ca80fd91144933f666908f4a
974 F20110331_AADISH he_j_Page_005.txt
54c3a428f7dea99a6b0877b5b0a3bab9
4a1923ede1629c3dfe9fc61b3f6bdc7f6a8c7419
8401 F20110331_AADLDV he_j_Page_043thm.jpg
1910800c526996c56034a91b3bfeb7c2
b6b549cbf1ec14ae2e11d0eecaad414c4563c16e
37452 F20110331_AADKZF he_j_Page_097.QC.jpg
4ca04427f99e94c6ec6bbc636b2fa89a
89795514aebb89e239f5f63584bba21f53447908
F20110331_AADJPB he_j_Page_325.tif
bd696497ff511f1510d98a67daeeae50
733c2a6e9dd87ea641d9fd703d1041f86d126441
F20110331_AADISI he_j_Page_194thm.jpg
8ed50b64b7a8dfebf6752ef6b101e4e8
a9960584ac8182268d013ba9b43739162c2de76c
8423 F20110331_AADLDW he_j_Page_045thm.jpg
4764a6fc018fea69af042e18013a1f33
28ffecdd9aca50a3e45bf55a2176aea74d89bc56
36999 F20110331_AADKZG he_j_Page_098.QC.jpg
2687d05cc4b82b35d8b9054682b5dea2
3dd5013d4832b9301f1ab4fca3e44ec0c191a7bc
F20110331_AADJPC he_j_Page_327.tif
09d29bb9d6574c2e1deff8b0ab536ef8
50fd6f9e4c45745d12d28e7c41dc4b6de8c9eeaf
17149 F20110331_AADISJ he_j_Page_164.QC.jpg
53c4bcf5ea2ea7e7848e9f0c4c3344be
527f9c015fdfeb1e54abf0ba949362f571c66814
8130 F20110331_AADLDX he_j_Page_048thm.jpg
d37d27b16964946b89697f6abe171c22
bdcb7e2389636f61b54e319b98cd15ebd8251ce2
37799 F20110331_AADKZH he_j_Page_101.QC.jpg
a269a12becd7e50c0ad80c3d0d5d4173
7c3696940b1dfe00bd0e9ad5549f895f241d4d1b
F20110331_AADJPD he_j_Page_328.tif
ee505391db21180935d707bfd77e92cb
b798a798f2ba0ab1ee09f59d9f2623967a540f02
112125 F20110331_AADISK he_j_Page_129.jpg
8efd45a24740670cbfd13b35778d3890
c89e9649b2e613cff7b682942c6669e2f1986ead
8749 F20110331_AADLDY he_j_Page_051thm.jpg
53efe28830b38afcd3640d1e5f763b06
0839e484104fb18fb242ad84ca3c7f24e8fed6bf
36501 F20110331_AADKZI he_j_Page_102.QC.jpg
470fe1915dee9fc5d957645ad106d94a
92b35895f25246d5a8e511965eacd413477c562d
1981 F20110331_AADJPE he_j_Page_004.txt
5ce42c49745741997edf1bb8233bbbc9
10f18ad33f1c5e0881962a580f6ace24239f11de
53830 F20110331_AADIFA he_j_Page_219.pro
c9f7cba63db2a207d754471c716d496f
bb8a0d2bafca5cbaf65435f3ca1e8ca0b0d035b2
8977 F20110331_AADISL he_j_Page_148thm.jpg
ac4d6758ea261318ce4c51efc2136fca
c6eca8b9093f2d9aa54e13729e01a01947ecbb2e
7297 F20110331_AADLDZ he_j_Page_054thm.jpg
7c21d4a053a1e5bd58539f9e2062b3a9
b1ea0a938cf0783b9a22ccddc274c1b9cc6e4473
33211 F20110331_AADKZJ he_j_Page_103.QC.jpg
6ae8212908ef4c1b3a4cc384c0b954c7
883f351e73f1e40ae6ef0ac72db52e8ea1abd357
4470 F20110331_AADJPF he_j_Page_009.txt
8ca11d24331c92d6e46d9ecaaa4d102b
7250af7098242c8964379afbe3f7a70f59491fd6
21644 F20110331_AADIFB he_j_Page_250.QC.jpg
90112da496af9c09a3bc7d47fc41f372
a96f1835cbb86fbad43172098dc16b4ec173e7c9
F20110331_AADISM he_j_Page_008.tif
45e7d155b20eaaafe11a9450943436b8
c1ec699357c090e28bd5a7aa4395fdfdeafa38ff
34627 F20110331_AADKZK he_j_Page_104.QC.jpg
2cadbe2ca52f35070ddcdfeebee01b38
2d515449f62e7896b4ff87e65a41f8a75ed331c5
2857 F20110331_AADJPG he_j_Page_012.txt
d1d362cccdf38dcfdb2ce6585a25e4a6
221522b58b53217d08531ec1a17075847a2a21b1
108506 F20110331_AADIFC he_j_Page_043.jp2
eba38d3fd2c8e654e81069583219310b
48901fafdf0ecf65cebcd0685b6223f64e01b89d
1398 F20110331_AADISN he_j_Page_274.txt
d7c11a719e16ff121fff19442684e06d
0bf93526440ee8f0d28b6f3a5ceaf9240d6f9238
32618 F20110331_AADKZL he_j_Page_105.QC.jpg
127049ab970d98619f1384e0fc8913c0
b471b658704182b5f40d4644d6329a3f9a74ffd7
54987 F20110331_AADKMA he_j_Page_005.jp2
c9c55e4c9e66346ff9e5c44c4003755a
c5db05baa7ff43944da91cc0a9f5efd924704cc3
3008 F20110331_AADJPH he_j_Page_013.txt
27d11f9700665323125a43fa58ce413d
f0486d8113caf097dce03da10e188d7254a16a34
F20110331_AADIFD he_j_Page_204.tif
37fbd97a5719a0f56062bf55b115d333
0b654dc21e1b9a3cbcb6cef33e6fe288fd523018
4649 F20110331_AADISO he_j_Page_287thm.jpg
33b712845043ee17c5226eafb376063d
769bb3282434d5533d04be51c7f6ce489211d226
34718 F20110331_AADKZM he_j_Page_107.QC.jpg
3b9663344ea8e1c86134a20d39395558
33be92350bfbb16169f9808f6af9de2c3c942fb2
1051960 F20110331_AADKMB he_j_Page_010.jp2
7620c84daf4e52f23c9deea2e66ae67e
3e86fc5fc80edd6504b1bbb82b324f60bc8b2901
2609 F20110331_AADJPI he_j_Page_014.txt
742bd2ffa6e6ed3995dfce2d54359568
6eda1145ea501aab1e37c6adb0cfd1dc623d2a9e
62433 F20110331_AADIFE he_j_Page_162.jp2
2f82fe8aa3bbd55b2ab1385473ddcba7
771077be4fdd1ee95493e8692a8499fcbb5bc231
7630 F20110331_AADISP he_j_Page_072thm.jpg
09b2b1f4a4344e9b14a5835d1a5d71c1
eb1c937f9ecf2ce5e6c70bf417e3aaaed69429fb
22971 F20110331_AADKZN he_j_Page_108.QC.jpg
51028d8adcb15aa518e095d5df770be8
5fe743ffcf3c5fd19bbd295301d9678e821cab30
1051986 F20110331_AADKMC he_j_Page_011.jp2
c9fa708eb0f6521f6334ba6f5be4c7b6
0e718422fce50c2c3547caf040a6785140da2e1e
3140 F20110331_AADJPJ he_j_Page_016.txt
b4ac32c207168f3a2d041059b251bf48
9cda46c173d81c55cab12cddbb55691a97392ddb
28080 F20110331_AADIFF he_j_Page_270.pro
cbff1c1fc284a3ab56aecd97aa3cdb67
caa411d63c5f6cd22a2da20670da76c82f133aa7
33052 F20110331_AADISQ he_j_Page_300.QC.jpg
174e4e7391cac7d584eab9eed30f4385
fc3daa1446b5117b3c484fa1b9369848f559a2df
17559 F20110331_AADKZO he_j_Page_109.QC.jpg
488118433bc8fe825ef98e60bebfc639
92216d3c6f7a587c9359049814e3cf1ad3fe8df2
1051963 F20110331_AADKMD he_j_Page_012.jp2
cd9f1821af83d7b20458302942d277fa
7c16e9cdb90f964f69ab13996496b7fa7346730b
1840 F20110331_AADJPK he_j_Page_019.txt
bd2dbf9c38da37a2df50b7035b89ab6d
00848bf7ba3b4ecf1180807cec284f2100549725
51641 F20110331_AADISR he_j_Page_070.pro
aeb8500c32943ba511ffb8261a7bfee4
d0a32177b2cfd79c59df3d5bfe1e707584ec69c4
13954 F20110331_AADKZP he_j_Page_112.QC.jpg
617883f32d5d9adf8f35640e6053cc24
1dad52e4f074d6b5833b8697865ffbcb653d3304
F20110331_AADKME he_j_Page_013.jp2
d068a58b8d3751578e62b05216ed4a14
ea16d1a62492f9e3605a10ebd4bace61ad371deb
1290 F20110331_AADJPL he_j_Page_020.txt
277f29047ebbed66a4bda3f47581f4d2
6202442ea4a3eb82e125f9ac7bd531490bf29a2a
35857 F20110331_AADIFG he_j_Page_051.QC.jpg
f5a08cc11e9513a0848eebe1d5df2479
b2da942319038bc70f3497096ff3ba3ff5ff095a
18300 F20110331_AADJCA he_j_Page_168.QC.jpg
e31c566c4b8d693eba61ca6da33e10f5
101a9bd027541fb92b758ec77c50c91d118e64a1
F20110331_AADISS he_j_Page_234.tif
6a738af75a93190bda7d0a48d3614e5f
3045f46a4d73fd02d2cc7e864e9c8f7f2d54876d
20698 F20110331_AADKZQ he_j_Page_115.QC.jpg
d90b04834d5b5b967e97955e4a8e08ba
4b3a38da56df9193c5412ef112233bee83b84a46
F20110331_AADKMF he_j_Page_014.jp2
7035ab4e2e5a52b18bb4a7866863dd46
6714779d76c95db6ca2855a22aa587e2f911a523
2058 F20110331_AADJPM he_j_Page_021.txt
4200f0987e1776888306c26f552b2785
5753928af285330efe6d3858471751887cdc4bb7
F20110331_AADIFH he_j_Page_140.tif
d4b38b5489359603ce371ffed953ae25
ce3914580ac45b9a4bfe3b44249c18814ca0e8df
8297 F20110331_AADJCB he_j_Page_142thm.jpg
bc34e1797df9b48a654a1cbe3a54cf8f
f0cb6763275f5fdf476b3d858fd758208f6aa514
20776 F20110331_AADKZR he_j_Page_116.QC.jpg
4285bf2368c33e08f6591a52c396b46a
628cedb0753ecdfc0a1754d4b77e2a3c8d7a40f0
1051975 F20110331_AADKMG he_j_Page_015.jp2
55959e4f09a5ee4fb6896807ea9a1176
d7d5335f23929394c0fc7c08c01d789d50eb47a2
2149 F20110331_AADJPN he_j_Page_024.txt
a576f0749493ede8a94f39c22c1a2520
d93103c17d0a5ebcbcbce00c3940478230c12d74
112455 F20110331_AADIFI he_j_Page_215.jpg
cc1cece4ae99f43c0902264a02ec10a0
c3053e18c32d75034eca950f7fdaabb71435d835
F20110331_AADIST he_j_Page_230.tif
ad019ba04564cff25ab84849bfcfcbdb
ed1d23fe3c45c18613416d88ba23cd6b7f8947ce
20975 F20110331_AADKZS he_j_Page_117.QC.jpg
a35938229446e1102893c907bb8ba3e9
b47c4d14f11568a8c0f90798901c5eeea4861db2
1051979 F20110331_AADKMH he_j_Page_016.jp2
cc07c5109f86ed5ca14bc5ae184741d1
fbaaaf8896c0d0bcd7c7a21eb24ea9abf7b02175
2088 F20110331_AADJPO he_j_Page_025.txt
2c692727abe861992f69774437e6a973
547888f1d5b47d15b657cae29c3828fe21a13e87
56500 F20110331_AADIFJ he_j_Page_244.jpg
d847aca1bd6c9d644586b07240b8a6de
dace3c7c54328a7ff30620b7adf4e4e45563a189
18297 F20110331_AADJCC he_j_Page_158.QC.jpg
f6b5de46deeca05b9ee36dc499ca353c
d779f4b510c4672c74956f68ed3f99d9efb17514
6296 F20110331_AADISU he_j_Page_306thm.jpg
7ebe2fbdd013bd7da8a4f24703389db3
cf3ea8f6c07a10b636ff437b1a8ca14f612899a0
19980 F20110331_AADKZT he_j_Page_118.QC.jpg
48ff264bae751a77aeec9fa6b6b98e2a
dc6c18512bab306bd0a7f4a07b640b1972bb255b
F20110331_AADKMI he_j_Page_017.jp2
83628139a475b6deab48cf5bc5889284
8f0553be53df812efdc4f117dfdf9479e7755f36
651 F20110331_AADIFK he_j_Page_195.txt
33b3f269a75ad8f9f77997ddea011ffe
1bdc907fc75a34050313e4ed40a1931995d4032d
106167 F20110331_AADJCD he_j_Page_144.jpg
753c7c5cabd0f88b826def29907273e1
69369e6054ad6a49a2b9758bff3ae4a7ef94ec26
105947 F20110331_AADISV he_j_Page_071.jpg
a8acb37ac6ea01520b5a7fcdd1b64990
c0540804cfb787df8124e9f0ee6b3da8fb149886
21715 F20110331_AADKZU he_j_Page_119.QC.jpg
ef4e7077f0aff0725f7ecb4a2447cc77
c16070a68af69c392861f86fb3d918a1dbc4076c
92149 F20110331_AADKMJ he_j_Page_019.jp2
eeab3927570e887fdf9f9061a65d65e9
3ba91ae0f87c49b426e21124328dc3f0f51728e1
2141 F20110331_AADJPP he_j_Page_027.txt
75f8e87c7b663c23d9f458c77e61ea41
6ce340d1b6fc2b2cc78c690c5d712c5e3519a17d
23008 F20110331_AADIFL he_j_Page_308.QC.jpg
e789bdd71a66619d016feeb0e77bf520
d531a633b3ce19c33b004eded7181fc7bdb6b520
35938 F20110331_AADJCE he_j_Page_329.jp2
f10841a2aa46bd62b1ba61de507eb1d2
800bd2edcfa3596db0ed71335c7a37a468fd4378
65273 F20110331_AADISW he_j_Page_014.pro
bf26afe5b1893af406d8a63819514204
7710dbc5b124de3a2f697b9d149c9ffbd048fc96
18675 F20110331_AADKZV he_j_Page_120.QC.jpg
39a5e2780281c0dbe67764ddff446669
97020d793308ccc1c5e84a90564ae0fbabbba35d
106913 F20110331_AADKMK he_j_Page_021.jp2
c16ebe4603b81b930f71ef0d31a64331
058cb5e41510756900fe3fa0be136301c2c8514c
2105 F20110331_AADJPQ he_j_Page_028.txt
520a8627835ae5ccc1dcae1b53cfd4b2
65ec84c580f3ef2ea7f5f112dfe22254cfac7303
F20110331_AADIFM he_j_Page_285.tif
0296bb6c57fa64f876b4c7db90d1e6f0
0d343f227a24e473cd6ee83fa192b6879519d567
14718 F20110331_AADJCF he_j_Page_314.pro
a4ce54ffe9bf63103a926c55487c88d0
1bd0c1f11be788efcaf642feea1db03de18be504
F20110331_AADISX he_j_Page_297.tif
1ca86224fea321d06b972e116f0e44c2
b1701247228a569520a98cea837ad0608e0fa606
10162 F20110331_AADKZW he_j_Page_121.QC.jpg
999aadd2cb62658f12ce23b51aad67d4
cb0e879b79554d48ce7fefa6e8e23c87dd0e23a8
F20110331_AADJPR he_j_Page_029.txt
e4216a11cf57356a312ef997fda388c7
6bbbaf2d028ed107b8b75b19a0dd609c7d4a08c0
F20110331_AADIFN he_j_Page_203.tif
bac07619343f8cf06f7176116437700d
8b52cf8f999b81b05c5b58240537c120d343f864
99757 F20110331_AADJCG he_j_Page_253.jpg
28ea1cd75fd645c517f5e17a881fad7f
f8bd1119143d141d3794477efd4319c136b1ca8d
109203 F20110331_AADISY he_j_Page_183.jpg
6bf4bb1219f21f314088a7b61ca1120a
0342ff17dc240633c356d23b33cf891b219dd16f
28663 F20110331_AADKZX he_j_Page_122.QC.jpg
0ca544b29fa91b276363810e0e0e82e1
553245c66df657df4917d0cd2154086fccf0a5d4
115930 F20110331_AADKML he_j_Page_025.jp2
3b9148c6bddea2e5a1632e5a5c9befb7
a36cdca3a0fd73039a2ca1f8346e06b246f3e3d9
2763 F20110331_AADJPS he_j_Page_030.txt
f9a2628168e3bcbfc70b17aa85137fba
09e0b8471840bb4129971b32a095bc926f64dabc
24588 F20110331_AADIFO he_j_Page_167.QC.jpg
c6d7bf53de854462a81360a5b646071e
874b940bc47e671cfd507927f463c0c785c2f618
54313 F20110331_AADJCH he_j_Page_146.pro
5e90c1f96bb26ae01a12d1acc89d2b69
11657eccacbc4e4a764bd3dc2b129da83ce1713a
65863 F20110331_AADISZ he_j_Page_092.jpg
b4ddc0da6ca0579baf8e0aacc3bbb2a4
03782ec7d7b473d0d3f27e81d2fa3168830d751c
118492 F20110331_AADKMM he_j_Page_026.jp2
7801d26adb9ad232c5314a3a20e0e253
6d4dd560887a1cc1e71dc85369c7891f1654a4cc
F20110331_AADJPT he_j_Page_032.txt
a305f66a0ebb22cbf84a242bdfb47814
b13c3853b7df792883a41b9c8c71067ce6b64520
28379 F20110331_AADIFP he_j_Page_067.QC.jpg
da384b61c825e66d6925b248f2a8a398
87444f8b1f0a8f871f5b5b67f1cf624f789293b5
120949 F20110331_AADJCI he_j_Page_058.jp2
fea83b1c74d9c88571b6350e852dd872
8232e2fc0b67bc8ecbe00e4190ad36abba28081a
26779 F20110331_AADKZY he_j_Page_125.QC.jpg
0eb4a274639c85756ed33233445c5d42
fded3a067eae9c75b63dda54cd55c2a2a19ec2a6
115938 F20110331_AADKMN he_j_Page_027.jp2
bfd451a2e86b8172d816c75cc25b7eec
010787709aef2acf61f044a0fa704eed135b6b78
2087 F20110331_AADJPU he_j_Page_034.txt
f9873ac712ce97af4a73e95b5f851a39
bfaa2249729a588bda8c956397f709aa5393e7ed
8530 F20110331_AADIFQ he_j_Page_004thm.jpg
e41455ff3df4e9735935d6b46a405d33
b996d968997fbba850cbee2a2fd36f373d83e74a
41515 F20110331_AADJCJ he_j_Page_137.pro
2edec5fcd696b7b55eb9a109c94d1904
f1de498d5700c247f2cd3dd13fd203901f2175c4
12984 F20110331_AADKZZ he_j_Page_126.QC.jpg
97c7a465bf72d70fa652cf36508efebe
19f044b7ce0128c64c10e9417396794568195235
114115 F20110331_AADKMO he_j_Page_028.jp2
97a1e50941177ba9735db32c929e50f1
761c51b4fbb702ca9e8813931005448d3d22cb3f
1936 F20110331_AADJPV he_j_Page_035.txt
d4121ac41e183b5234fce8cf85b3e04d
15f348a26d22bc46cecb259ebe95ceb8fc12bd5e
49899 F20110331_AADIFR he_j_Page_091.pro
f61e9d5b8e32147fb3b485dc213b05f2
3cc46c91b876e313bab6664570b282cb7c0e2088
2165 F20110331_AADJCK he_j_Page_031.txt
bfa94f68f6be8c9cc06def27ec742ac0
9adde23192b129d147c1d3e9a7e47e6aace124aa
136398 F20110331_AADKMP he_j_Page_030.jp2
a965d251478e327c354c7116ae08e288
3306b5c0fa4f94b21d307c56775be9d9e9cbcf3f
1638 F20110331_AADJPW he_j_Page_038.txt
8860cfdb039ee1b6e818704184a4cda2
e35cfb3775fe13bd9025e06f777203d36439cf23
109096 F20110331_AADIFS he_j_Page_217.jpg
e67ab4e96a966da9b4f7f0e8f3fb23b3
57a1df16df87e584633e8272d2fc8212c804a42e
42733 F20110331_AADJCL he_j_Page_042.pro
6f314292ddd03866a65e82ae8df5f9b9
f56d439aef9bdaa1872a702634ad8f64bdca82f2
118697 F20110331_AADKMQ he_j_Page_032.jp2
38739b38fa564b6722864d11216e8830
f75422db83ea44e3563107c0a212ccd8b6efc24c
2000 F20110331_AADJPX he_j_Page_039.txt
8f78d9b160a986353d7000233c1489b4
e6dcbe2a89739899afb1c08e89c97d4dd27bb02e
7333 F20110331_AADIFT he_j_Page_136thm.jpg
1e36810b5ba42a05bd37596d60d9cc61
8aa71860ac5274b3657c429063c5774c522ee397
51210 F20110331_AADJCM he_j_Page_194.jpg
4633fa81f5904da1b92664389292758e
b14ba63ae7699372a0b91550a1fbbc754d5c4148
100934 F20110331_AADKMR he_j_Page_035.jp2
7f633be7fb58ea982535b52bcd664dd0
b17493ead27623fc056e0d749d7ae5eaea04f1ae
2033 F20110331_AADJPY he_j_Page_040.txt
dead9bb1bb11e27ae2a556ebc7eaf7f6
94a9d7e714591aee49079c0a37c3c8151a3c3fbf
F20110331_AADIFU he_j_Page_241.tif
643eae78f1c0672f7b474520e6e44150
84b402b24407ab1f85d377ddcd82a9868d7ef1c1
728 F20110331_AADJCN he_j_Page_304thm.jpg
20fc669fa6943fff2760e4743039262b
d84a71b6fd384333f521226cecebca97c776e4f7
71084 F20110331_AADKMS he_j_Page_036.jp2
1805392047dabd87165587e134b7c001
6c126bcfb5a723235216f1ee0e268c789882d8bb
1897 F20110331_AADJPZ he_j_Page_041.txt
b46bcd66f047a4f74dc44b7c05417dd1
05c2a04a511f28350bef7dcbfc3fb617bd562571
1354 F20110331_AADIFV he_j_Page_264.txt
6fbe7cae8951ec5db541e2efc52ce2e2
e3775c93c3bc1cfdf6935d492dfee5c34d1b235f
6578 F20110331_AADJCO he_j_Page_137thm.jpg
e8296677714a0081f9877d99f1cda0b5
da7c0f4cb9481cc20076aa1a44f3b5a4817790a3
46419 F20110331_AADKMT he_j_Page_038.jp2
64e4c0ffeb28e3858aa0790b09d6ba19
4b72ed28d84c5908399c42646daa902830f3698e
2038 F20110331_AADIFW he_j_Page_045.txt
720102a00ddf345139c2eee91f541549
0d51df64d3cff1252afb440957a16fb36eba652e
5383 F20110331_AADJCP he_j_Page_162thm.jpg
615f822adf4753b41d4396ca68fc8122
c092cba04b83abe06c4e232adea91907cc135d8e
103650 F20110331_AADKMU he_j_Page_039.jp2
406d5a7eef72e5627882838a549ffb54
4675f67672f525e674dcf1b530c510f2a1771f28
2620 F20110331_AADIFX he_j_Page_199thm.jpg
b25095c22562edace700c4caa46794a3
2be55c60aa91f90ca8fda09414e57930f399df7a
F20110331_AADIYA he_j_Page_286.tif
66f791f6e581c988203e584b393654f3
7fbaf35976b464384679e2a8d4724b63d971a189
9193 F20110331_AADJCQ he_j_Page_175thm.jpg
2ea4bc2a0bd84ef798957c50d59c44f0
85d42e7a23aa61d1089ffed00d6e778bf41a6f57
110510 F20110331_AADKMV he_j_Page_040.jp2
db60be7ed7e67f641c9e472214e3875b
35f6c33bb2679892eb87969574b1bf846fc46812
2465 F20110331_AADIFY he_j_Page_018.txt
525c5824611e820bb957364a1fa544fc
07d7b498e8b261b587db8a941665f332035d3135
13105 F20110331_AADIYB he_j_Page_065.pro
43b0ba5c42ff90392718a6e8ed8408af
e0761dc6e91f88f3f3185c311f757f9705c46631
F20110331_AADJCR he_j_Page_093.tif
300a481685b49318f20be4ece483f16f
7cc4e407f547cbc151a77320d6aeaee75ad0a308
105292 F20110331_AADKMW he_j_Page_048.jp2
6112e4e8fbf73ef2587e5a72353768a6
b22097b4908e952bd19bc1ceeb594a4f025944b3
19267 F20110331_AADIFZ he_j_Page_244.QC.jpg
a7142405c8468462f4259f2ccc6a6797
0812b217141e73307350da6ef189215fc93d2572
7705 F20110331_AADIYC he_j_Page_050thm.jpg
e4c66b29e37d5bda211d66c5a8d57d1c
ab5b4b872e263e12796cb1c2d6d5ecdf102d2e18
383062 F20110331_AADJCS he_j_Page_111.jp2
06b7f9ef7d56f0dc5ee6c6975e135435
07aa2581a27f4949d584728a3910c273aa9ba2d6
96291 F20110331_AADKMX he_j_Page_050.jp2
1caebc88db4b9344560c72e35db594b1
31248e1045322e2803747c1df7d5fa1adecbea21
F20110331_AADIYD he_j_Page_187.tif
e01c629b142cd7ee8a5a84f57a334702
84bfa87c02d8dd4174e19f7bc89d26a60462dbeb
840 F20110331_AADJCT he_j_Page_198.txt
5ccbd4fa5a9bc997fb34c5b51293268b
6ee99ef72ab686f21df9ada111bdefd7a66c8a64
114274 F20110331_AADKMY he_j_Page_051.jp2
8222b9a56e4a36c176f836204a2b0edf
49ce60eb32b6da75393d5c080c71a15e5ca9811f
98078 F20110331_AADIYE he_j_Page_056.jpg
cd07f79b7f2b181cf73a1580434b5268
741776bd5b5513f904a917f82ee2a429838bde3b
26585 F20110331_AADJCU he_j_Page_252.QC.jpg
330090b1cb233ea703d948b805e2601e
951213b633cce97f5b1ceeb685affadc1d634e4a
94741 F20110331_AADKMZ he_j_Page_054.jp2
b781aa90c57adbf3d58949eaed68a81a
f63a55a865334ce1f6850c670d26f3ee592a5b85
98005 F20110331_AADIYF he_j_Page_099.jpg
1016a51f6a8d9b1c08b3b7817ff63b3e
c97bfc38cdd40e03e68f2bc539a3a84c1ac25f16
F20110331_AADJCV he_j_Page_061.tif
1767b499ddc1dccae2d91c6d3460f060
c094041d5985af0cb6c13a0eaca605bc5e5cb929
F20110331_AADIYG he_j_Page_239.tif
9aa53fbef8ec916060f8b6950f793bb8
3ed6fd7e57f534a25b9a48cdb3c14edf7b9e7e93
112218 F20110331_AADJCW he_j_Page_184.jp2
b3d014cc8c0d5f94eaf978076d3c7ddb
728187fd3c06593b681366b04b74098abdac1755
49505 F20110331_AADIYH he_j_Page_043.pro
e4cd9f329cc2463b8ffa3be0deb1d4c1
4f037a1bb85b541310dcd545dcb116d5f87df2c8
105043 F20110331_AADJCX he_j_Page_243.jp2
a111ac65a467cd6fa1763b01ed5a1b82
bd2e88b2a208db158a84557f52008cc42f792c47
1679 F20110331_AADJVA he_j_Page_246.txt
98fa6bd620ef17195aadc26b8f176cbd
71ebec932344ab45178262db3eda22dfce6bfc05
8303 F20110331_AADIYI he_j_Page_171thm.jpg
24d1ba825b25844a0e48a2d0ae803b47
07c75726e973618d9afb0f8e731d906a1ed872f4
26409 F20110331_AADJCY he_j_Page_260.QC.jpg
a17a11eef56a66a74d7680259c210029
22c8ff87560d8e4f47de8dc8f5e0acff76c50ccb
F20110331_AADJVB he_j_Page_247.txt
c3f3fbeedc4b2e4d70ec76821405a2b9
52b22b44e1686122423a672b4c962883d339404b
4937 F20110331_AADIYJ he_j_Page_195thm.jpg
7e4b7852095432c92db2e3a16bba241b
74619b8451add31d4900025cabfcadc8e43f3432
70869 F20110331_AADJCZ he_j_Page_262.jp2
3ccd0fe99d7f4145cef1682d90136d66
d051475a882f1f528db4518aa26f00db7415e964
1566 F20110331_AADJVC he_j_Page_250.txt
1397fc5815f3ef06ac7ba531ba2c0a46
57ccda5c136c60cd023d1cc90856dbb576b8e565
119832 F20110331_AADIYK he_j_Page_098.jp2
0b63c6413e9476443242b1314616ca77
0d9153c046b4eb8d3f10f4fda151b5054e363466
F20110331_AADJVD he_j_Page_251.txt
51607ee43b192a5bff2ed1053aa7027c
31732f93688911259f9546369de930bb7274e131
33737 F20110331_AADILA he_j_Page_021.QC.jpg
398bebb14460f6d6057197ad131ea73f
7ddc9064b42efaf7ebec1066e6e196ed588f53c7
104988 F20110331_AADIYL he_j_Page_078.jp2
f7e603b24101286661a84c1479b94afa
52d4932e0f85a9bb2110fcc204521628cbf87670
1701 F20110331_AADJVE he_j_Page_255.txt
4d03b1d66c69078ae631edb29849b7c4
c0c3a7d690846f3c571e4c596347b292f069f5bc
637619 F20110331_AADILB he_j_Page_306.jp2
be59d49515050ba63ae0e19c9ac4586a
d4d1b14cba82485cf25d18e8e61f680ba47aa90d
F20110331_AADIYM he_j_Page_256.tif
141d5c6ec4349ffea44b67bc3b5c8207
655fddd22323913f37e546d4167c6071f9296410
1554 F20110331_AADJVF he_j_Page_257.txt
fce18f242f10b67287a43328f06d6b68
9a84c29138f1875796d47223551d8405aeb79c44
F20110331_AADILC he_j_Page_106.tif
33a2f03dda44e7feff89088304af0988
92088cff081994cb0a62b30b1586019559582b2b
1051976 F20110331_AADIYN he_j_Page_007.jp2
6d58a392a8e10b800b2d5f2ef0b2d7d8
db2f82bede7041d3308f1d461385746ff0b758a4
1451 F20110331_AADJVG he_j_Page_258.txt
0fe59b269b508b21aec4a4404f9e1872
c63d6208ad3ef07eac63026a1f9096a861ce538a
64285 F20110331_AADKSA he_j_Page_276.jp2
03260d79d530ffded9ddebaedcd898f6
a3baf7b6eed4b68e990458477e280bb796e8ec6d
303938 F20110331_AADILD he_j_Page_226.jp2
01c679b58d2a995631516bdeb4e96d7c
530760dc06cfc8c12c871a5f970880167a70b6ba
112038 F20110331_AADIYO he_j_Page_218.jp2
f28b5c95a2c1828294418597599d790d
835f0adfd0dc6f0bd6bd1271f073b7e3cfbda8be
1515 F20110331_AADJVH he_j_Page_259.txt
e1f285cc91736ce942de428546bb282e
9c28acd412068ca08fa1495f8a736c2b336ac626
75771 F20110331_AADKSB he_j_Page_277.jp2
2bf7f66d98aaf212c9f8c93456fb7a48
5700db9771e5eaa56e024f889e27051459ef450b
67643 F20110331_AADILE he_j_Page_249.jp2
1bf69e65bd08ca467b63bcd411d1518e
3149f8c7d1e9c48c3e65ca76239db57a7c687acc
23213 F20110331_AADIYP he_j_Page_256.QC.jpg
85294be5bde9dbb9d831268ed4278ac5
f9c95c2479fc93d1a3e21e6e3bfa56e3ddd2dea5
1834 F20110331_AADJVI he_j_Page_260.txt
b3d923bda9328e9c69221261a2ab4adb
283249cf55898593074caaecf64e822d15b0f636
70631 F20110331_AADKSC he_j_Page_278.jp2
1b686d16a2d1756996fd9503f9c3dea1
5c38225598b6e736edc83dbc4c99c012748e2019
56384 F20110331_AADILF he_j_Page_289.jpg
d61aecb90b488ec1f9a03b272af69e93
c9a5965a360d73d5d0b004865d592193629aef60
26206 F20110331_AADIYQ he_j_Page_248.QC.jpg
54145e8d436b62111667ee43ee7ffd8d
26d2afcc6912520c62d0bc5a02b8d9b4c8c7d8de
1527 F20110331_AADJVJ he_j_Page_262.txt
1ed9285f374a8391bafeb3658b39b4df
c3cf978461ef70693e2aefa1214332994d6f0fc6
65004 F20110331_AADKSD he_j_Page_280.jp2
366eafe2b58eac13cae557254b44b09f
f51e95f7a60f00e099aea496061cca11d2485541
69489 F20110331_AADILG he_j_Page_166.jpg
b8b72992b38620525087e93c5edca78f
88f9c15714b0c4e9ac760a76835557633a858792
F20110331_AADIYR he_j_Page_022.txt
d00f2d6ebfbc09d0fc85f2e60fd7a7e7
bb22fa689bb6aaf791088b02bbe9470ecae5bb8b
1351 F20110331_AADJVK he_j_Page_265.txt
89f1f6489b05f211b03d5e6d8dd668b7
45ecc0d5d5fb7bb95bdb1d1aad9962c126a5db98
75117 F20110331_AADKSE he_j_Page_282.jp2
59a8521aa80a230e32f2ca9d1cf68ca3
c6ba3f3297a0db9763d9f5fc5d00174bbfd00e00
7660 F20110331_AADILH he_j_Page_075thm.jpg
bccf6c4c19c1243ec0351622e1cd4f56
5c770ef8c8af7903386ccfa19d5f25e543e6a23d
375959 F20110331_AADJIA UFE0022077_00001.mets FULL
aeee1cd189f2ed07e2fa15fbf50c9416
4b7be3d95324ae6448f8648b304ba45337a292b9
8340 F20110331_AADIYS he_j_Page_242thm.jpg
e00d00fa1e263c4638a39140e26c7ead
15b59bf644ef1034446a553deacc2b7cc8769e49
1244 F20110331_AADJVL he_j_Page_266.txt
cf872bbf1b828e316938712dedeee5b2
bda4fa0a4c44d2533a6e04bb26f3c59aa3f3d056
65842 F20110331_AADKSF he_j_Page_283.jp2
57481035affe26dd5fb32ae7e7482e7f
9c739e237133c442cf706211054977c66797b5e3
1934 F20110331_AADILI he_j_Page_078.txt
3c0ad83c93e95b98ad21175a2ceed277
6aea97d6e8a3744ec073a4253c0524dca30c0b35
8727 F20110331_AADIYT he_j_Page_183thm.jpg
f1725d6e7e5ddc71be043bb5e704c9ca
cf3c80e0574d150da000921b063122815435c4a8
F20110331_AADJVM he_j_Page_267.txt
425e78b4110771e230de569d0a3dc435
1737dfd55e05cb4b76ad8c81092c1f01ef3db4f5
72956 F20110331_AADKSG he_j_Page_284.jp2
fb2d71633f781bdd26e637926a4f181c
79c6f7a5d480a335447b836ad53f61c2167e9c3b
26837 F20110331_AADILJ he_j_Page_292.jpg
7a10300cebabc8efdfa726614b5043fe
0d8f260554ad96dff7abe4cdb5556c89cd5c7dfb
75356 F20110331_AADIYU he_j_Page_320.pro
6c76d833234809c42b753fc16f3a81e3
85e8931b9b1d2f07840d799523df75f0ec235ed4
1397 F20110331_AADJVN he_j_Page_268.txt
c11911d61f6318ff7c9dbbdb4492b69b
5adf3f98509098046b887198564f86f2ea6d3ed7
76200 F20110331_AADKSH he_j_Page_285.jp2
52b7a18b0e397c2d19ab78ef848770da
bc3f0a1fba453f0e6e82756c610e7ddb04c84ab8
31383 F20110331_AADILK he_j_Page_093.QC.jpg
96531a5ee2821eace9d20897d5862a86
581b80f0a4c0d9ffe42e3f1f7d9c1b677bc384af
F20110331_AADJID he_j_Page_001.tif
7460bec371f9003db4b817c044ef1403
bd1ee6caec2f9f7d4c5f28a94fa91838166b52e0
886684 F20110331_AADIYV he_j_Page_066.jp2
68bdd12cc1c4c0840cba7b7eb3f965f3
f18f11f4c5c3aebb89a34d789c78891cbd77528b
1298 F20110331_AADJVO he_j_Page_269.txt
8bd3483a918f6782f58e459be7b40efd
246c1b7a00bbb476222207d847ff58823e5a6955
66184 F20110331_AADKSI he_j_Page_286.jp2
fa220a55f421a9d634495b151efe2b3a
8650025f0a372bfb5726715c1937bf0f918f5298
F20110331_AADILL he_j_Page_099.tif
082302d464bc118041ac84e8c579c5d5
563ac4d8748d3ebe3bf271dbec161f8338e48910
F20110331_AADJIE he_j_Page_003.tif
a48311f7ff72efa746e61f113d9b2304
624e77e15260544a926d2290e260095b6d6331ad
134373 F20110331_AADIYW he_j_Page_319.jpg
8cb2a1f515ecd8e3a839a6d9984ee610
48c7781e1208a335d86f25d15736777b8ab63f8c
1182 F20110331_AADJVP he_j_Page_270.txt
c65873e52bd07a14948fb387c10df063
f0b75db73d061e112e09b0c882fc4941769d5023
72957 F20110331_AADKSJ he_j_Page_287.jp2
6f9c9507c9ae178c8462f958c8fc520f
f9b27cef9378337cf62945886cecb647de012607
F20110331_AADJIF he_j_Page_004.tif
3579f266bab7036b5b12db30c67f9fb9
3c01ab3cd2d85495450e6fbc70c030e4d7d5379a
622 F20110331_AADIYX he_j_Page_329.txt
c5d3df5162d4c48cf6c94818d9d7999a
f0a329bdc52997811925c333185312e969d4d17b
1078 F20110331_AADJVQ he_j_Page_272.txt
3a317517f639f51062952fddd66fc92e
9f687d7f008a8ed4fa8f5a713a05940c61cb9765
74387 F20110331_AADKSK he_j_Page_288.jp2
56fe4e434e185dfa460c166c12b9be9b
adf71d8773e461c46a8031adfd6ee8e5a519bb46
66065 F20110331_AADILM he_j_Page_200.jpg
cae9c557bbb4e0a833bc2d120ef635ca
9842bee22fcf320dfc1f00d27803bd60fcf29159
F20110331_AADJIG he_j_Page_005.tif
d905519f67cbc970439d2b1bb21ee171
f4867017e2caf373e5578ebc63ccd09823c38e58
9050 F20110331_AADIYY he_j_Page_057thm.jpg
8356f4ae4808fdab814fcca75bc22325
248daba9548c6cbeb28ef4db8c88787a29eccd1e
1374 F20110331_AADJVR he_j_Page_276.txt
bccb6771e6f8c58fb47c892b612e14dc
6944ac5d1d2ea0042f2b49c4bfffe80580243a61
64004 F20110331_AADKSL he_j_Page_289.jp2
b92a2d819772db04963ac013696749c0
4dfd8d1d95b489adbdc660d973776c832ca5d4c3
2055 F20110331_AADILN he_j_Page_180.txt
5009155b2e1947d91921075951759f23
b65a4f54e17633d5481e71bdd8b92e0aedb939ed
F20110331_AADJIH he_j_Page_006.tif
7cb61e136cd20e47e1df4a08a1cb7529
efda959d4b5566010ccc35ee721d09e7715ed7cf
104778 F20110331_AADKFA he_j_Page_018.jpg
a2de8806469afaf1f10ed53879c72e93
c4a512c30b08db3ce298969c7596d97136f2b8ad
1440 F20110331_AADJVS he_j_Page_277.txt
e097082043975ce56e043c52c8840be6
77095d9ed41ef38ac9d680dcd2afb916c76602a7
62803 F20110331_AADKSM he_j_Page_290.jp2
a9f6f11dfd504a201657348a6244ef8b
a972e750247c98eb64e5bc4a3859309747c6184a
87567 F20110331_AADILO he_j_Page_125.jpg
f022f47c61ee2f73c309fd983a546bb3
ed4d55d4c59d1dfcbbbb5c5fcc378f5e5ff2e1f5
7648 F20110331_AADIYZ he_j_Page_003.jpg
7e87ce3be4c6de633ba69fce1372c054
5e295ed6c295fa6600c5f08d89fa6d11ac44b7f7
67881 F20110331_AADKFB he_j_Page_020.jpg
f6fae227c596cc7a084389669660faa1
82625ecb51d3a6486b9dee9df2214053f6d3cfd8
1438 F20110331_AADJVT he_j_Page_279.txt
8bc11f6d76177a055f8db00ad359da74
4d3ba7d9b65878384b2e6e77fcc4c50d51f06017
70212 F20110331_AADKSN he_j_Page_291.jp2
9249f260671e1568ae3f574c94d02d4a
d372478f33f65db304b18e4868b1d299e75bca23
110977 F20110331_AADILP he_j_Page_068.jp2
60c991fca02c75bc0ded7adcf05f4b21
eaec1a3e7af0cd8944850e1e9ef146d795ddd5b4
F20110331_AADJII he_j_Page_009.tif
d085f2e6b768697e708dd47254246822
ae95692f955193825bac70245f332045d27e9c8d
102341 F20110331_AADKFC he_j_Page_021.jpg
957c52d565bca9febb47b75b7d3234d0
9eaa8255ed8a737e262bac9562118ec36cab5b82
F20110331_AADJVU he_j_Page_280.txt
c186d98c00d4d4244fd1c8e2ccf7d386
fa30bf59469aa1ac77561ecf29136373dc8c489b
29801 F20110331_AADKSO he_j_Page_292.jp2
baee1eca225c22f79495fe413c662c9d
34fb518f0506f3c2ac5163e32befbefeb55b21dd
13617 F20110331_AADILQ he_j_Page_061.pro
effe23c385926d4f514409269edec004
40401b5f3a0beb15983256ff24841541c4e311d1
F20110331_AADJIJ he_j_Page_010.tif
07f7a31d7b950e26843fec0899a3db7c
afe066a6145c86b6c2a41adbd5d19ac272b1fd14
111205 F20110331_AADKFD he_j_Page_022.jpg
4ec63ce68419604133099e4e6633284d
347ff2d75d7aec48e6073c669509848967eb4777
1051808 F20110331_AADKSP he_j_Page_293.jp2
9d480d7a60008b8836e3f5ba4bc1afe5
e9baa0a9854bc62b4ffd5e3ce0c3df3bc52adfbe
54624 F20110331_AADILR he_j_Page_094.pro
92a1a18c70ebe256ea30abf052e863b1
92ce7881b76a21f63262fee0e810fe90c56216aa
F20110331_AADJIK he_j_Page_013.tif
5dafa92dfc5c7b85525331199aee7175
e88c777f398b9bb47a041cba43fbc1e28fe91363
1428 F20110331_AADJVV he_j_Page_281.txt
ec7516f3ee5b3d0ef23f4e6b9f341f70
fe18de528e07fb79e9579634752af7f4e784ae7e
1051964 F20110331_AADKSQ he_j_Page_295.jp2
8e4613254bc2290cc907da5032325744
899fb90adfaa6102208062ec6423c51f95349749
47509 F20110331_AADILS he_j_Page_221.pro
a5cf9f3576a19bee52da597cdee318e5
bb9bbb49b214a90720516df227129e024f28fdf2
F20110331_AADJIL he_j_Page_015.tif
ca424a721dedf33166afcb876736f491
e233ed572c20552c67ca1ca9776cca31e52b7e08
112709 F20110331_AADKFE he_j_Page_024.jpg
929f8dfd9f2c5930253677c1a645b4b2
3f6f0ca4ede1a9007d2b2debe4afca4ba5eb13ae
1379 F20110331_AADJVW he_j_Page_283.txt
96b6b51aacc74d10b123c4ae2f46e91a
321dfd6c636fcf347bb003383f189adaaa332ab3
F20110331_AADILT he_j_Page_038.tif
8d2fc7ccc89349fe5b28293152132a99
fcad005ea7875e247d3f1294e80807a566a1f2c0
F20110331_AADJIM he_j_Page_017.tif
4ff81a40ce7b3b7acf5b0ee7b50bece4
8386930bafdaebd5a34ba635104df8a481742705
109081 F20110331_AADKFF he_j_Page_025.jpg
9db409aef0559e8e52983f77767c7aa6
84af6ad1551774a4a9d3b279655b7aa806bf5b02
1412 F20110331_AADJVX he_j_Page_284.txt
ffac8ffd2a18b39d6b3588b4d189af16
e60e2c81bae06864140d93deaf4f19db84cad966
1051928 F20110331_AADKSR he_j_Page_296.jp2
a443d859e74699fb81c7fbab6d919354
7ebd25fc2ef094a909a0c98c22657465fa7650a0
8460 F20110331_AADILU he_j_Page_059thm.jpg
33b5ae58fd5be6840ee93a923bf4951a
b9b546be986819c221c1ee6988305cf6cdabea18
F20110331_AADJIN he_j_Page_018.tif
4094dbb7c7e038a76508f6d35dea2bac
b17cd1a619dee0099e9d005d11042a6a7af684a0
111326 F20110331_AADKFG he_j_Page_027.jpg
4c689bc2d59a987072fda31035a54c0f
19a5265655e89e1cb32873ab1149fce8aed41ca3
F20110331_AADJVY he_j_Page_286.txt
e6246befa43156ac0cc3b481ce9d5d42
d7eb8f2ec9e64b0c3705b09747ec57afd7713e28
16644 F20110331_AADLCA he_j_Page_269.QC.jpg
72829e67778a7bc6c64ad201639dbedc
2dd05a1264db76138f1fe69d66f1c5a402f9e0b3
1051968 F20110331_AADKSS he_j_Page_301.jp2
8cb2e8dbf1c79155c4ac954b47b58cf0
920793457df15620d0a58138fa5ea8a37c4b46ca
7690 F20110331_AADILV he_j_Page_297thm.jpg
5e03d0cbcc69443b81cd4e6854db38f2
4428d856dba25966844b29ced6e952b5e654f731
F20110331_AADJIO he_j_Page_019.tif
9784e104ce21971ccd24d7c3e7a44b39
f9d2615fdcc757bb578652c03e3aa3d87739daec
105950 F20110331_AADKFH he_j_Page_028.jpg
b6896aa40788f8c235b83e686e165d88
afac27dd4e5792c604c411eb10ab5d73d633d373
1476 F20110331_AADJVZ he_j_Page_291.txt
6e9db814e0f7e160539d34c2581f7c60
e60b98e9ee251ff158ed5d141da3530897daaa9b
14458 F20110331_AADLCB he_j_Page_270.QC.jpg
8df22d344b6b02aa9683f0cc1be07974
b30076ef62ad629f4abdf6002371e6495aa9690d
1051955 F20110331_AADKST he_j_Page_302.jp2
6f00fe71c4fbed701a322f58aa44ebd3
b7d15320f28b2d981e75fb294ffdaf050d7f4026
84894 F20110331_AADILW he_j_Page_080.jp2
430cdc48646fd72903befa4f78c6e73b
3bc355db4149d5aac6ce750116ba55ea9f9683da
F20110331_AADJIP he_j_Page_022.tif
dcea146ec63655269eeefacb5a4956ae
0d13c39193d99d10e64ff2b3b44b94aa38e3ef9e
100912 F20110331_AADKFI he_j_Page_029.jpg
d5697330122512d5044bd03c11a841f6
4e0c7c74e35bc320db7d47a4f1b72f60e1db9f3a
14295 F20110331_AADLCC he_j_Page_271.QC.jpg
adb3175c364b244a5a4a0d322654f08a
b9e6b90a6ca6cc49c143d09d29c1c91f6b43e692
54076 F20110331_AADKSU he_j_Page_303.jp2
03da3555fc1cba3f3bb4b3eec91757c1
f9bab4a8dee9991b73128c66c76c6506a62ef1c3
28929 F20110331_AADILX he_j_Page_141.QC.jpg
3186c88fa488302d411837fef2360aec
15dfa4dd98b6b0e6cc581baabf23001602d68703
F20110331_AADJIQ he_j_Page_024.tif
f61880c640127206843ecc70c15ee27f
13d989bbfcadb2375390d5913ae2bcc1f76b385f
112229 F20110331_AADKFJ he_j_Page_031.jpg
a99209d1aec427fc610b969b57dd9629
a5c8c5cf441ed322db7307f64e40e29bdffa3ae9
15078 F20110331_AADLCD he_j_Page_272.QC.jpg
25f18926d01437b9335a7f5f11de6d20
5bcd1caff8b2393bd2aaa4b5613f709553b95bc5
7718 F20110331_AADKSV he_j_Page_304.jp2
a2e4542ade4243f9c55d27afa7e6563e
2f01d2b3046fa0770a39bc1ca16028d7c17ef01d
F20110331_AADILY he_j_Page_271.tif
cfe7ddde7f635e1305e4218dfce4930e
9c0375e4293b185ec13a00cba52fab493b2cdfaa
F20110331_AADJIR he_j_Page_025.tif
34aecf4eb1eceafc68e09814b99e48f4
b4b8e433121b3b7218b58d06bc33015c8bb5e2aa
93688 F20110331_AADKFK he_j_Page_035.jpg
795228fbabd5ffaf6a72f2b51bf31348
9bee7058c18b9afe0d9b0c9555bab9eedf94d383
15783 F20110331_AADLCE he_j_Page_274.QC.jpg
d994bd2aefd6afb752d592fd4f95d3b5
7414fd810c51319b4a0e2382cd45e5684d686f30
575994 F20110331_AADKSW he_j_Page_307.jp2
01bdca41970c390395d10b7282feb24b
df13d3444994f0166b18c4c36a989d957d8f1046
14088 F20110331_AADILZ he_j_Page_224.QC.jpg
79e2a921e17f16d219e034b911b7c67b
a6f7dbcae53afd0fe6e0aec0723c437c323aac50
F20110331_AADJIS he_j_Page_026.tif
283b7320dd165f574c64cfa7c5ed45f5
aa422fbc89ba74cd2ba327193d365ce7a2fab38f
45302 F20110331_AADKFL he_j_Page_038.jpg
fe01a09ec78441bffa7dc5a995935f7c
235490ddf57676d95f2de31fea7e8b29b8e8b07b
16829 F20110331_AADLCF he_j_Page_276.QC.jpg
27d3dbb0c1524f4f3e14ac1f7201a7ee
3f9eeb10435b5855c1f1e0803f0d6acc4a3480d2
679601 F20110331_AADKSX he_j_Page_308.jp2
d05d8fadf9c6a855cfce1783504788fb
098dd0f40a050208f83a1fecefa5577e5e200e09
F20110331_AADJIT he_j_Page_027.tif
db4af9e016229f494d6343dcc4828695
278f5733ab1b22622a75730115ae9830fd0d5072
102924 F20110331_AADKFM he_j_Page_043.jpg
d7d9ca2d63bf96eaea2471ca030116bc
521bc528e09978e41ab527c561acd408461b0674
18124 F20110331_AADLCG he_j_Page_278.QC.jpg
ff05a41e7fbd8f01033ed0ea52ec4019
1151597c3df4fc856d04422c6ee5f6fb49bfb9b7
581432 F20110331_AADKSY he_j_Page_309.jp2
b21eb761b9fb3ef86f5814505e0173a3
020ac355c2b9619cdc1469782edb29f274821020
F20110331_AADJIU he_j_Page_032.tif
bfe0dc83241df90b6411ad1512c7cd63
3be317ed0505d1c8c9a3631c56c3261b27b6466a
113299 F20110331_AADKFN he_j_Page_044.jpg
33d488e4106fd41e4782df76c250b9dd
e25a4cc36cd0e763491543f253bfaf8b0c07146c
18279 F20110331_AADLCH he_j_Page_279.QC.jpg
49dce21c9e25959f13c7655fd52f9dac
c2b87c37461fb40a95005c11c2d347f2e5b153e1
574691 F20110331_AADKSZ he_j_Page_310.jp2
3c617fb883756b7cff1908a343fe0e09
bcd1b99d86ede627ff3ac268618e8d95f381c4eb
F20110331_AADJIV he_j_Page_033.tif
2ae5488bd7b2fe70536b03b4cbb0ddfe
182619e10a347b3756fa542447a8fd3fd1129f6a
105507 F20110331_AADKFO he_j_Page_045.jpg
197eec8f4ed3bc66b9e2439d0fc497c2
f7155aab1de81dd4e03d540e571ce9aec91ab67e
16533 F20110331_AADLCI he_j_Page_280.QC.jpg
48c157cb6f97c1dcf1774cf002679697
372b56c6b5b5b054d0ad95b8fa69db79256ec1eb
F20110331_AADJIW he_j_Page_037.tif
a0497c6d3d98a4cbdca24f9f05959f13
8e890926e5a59af05c8d8189501e96c21b97d937
90505 F20110331_AADKFP he_j_Page_047.jpg
514b4d1257f930fb0c63995cb0776f68
9986d5716dc5c16be5736556a97b7fe0f2fcd005
18409 F20110331_AADLCJ he_j_Page_282.QC.jpg
f799feff0fc20c3e89edb68be5ec0642
78292928c7da12bf61c86b687b292820614a515b
F20110331_AADJIX he_j_Page_040.tif
2c617f5810cec367616ef5481aebb18b
cb5aa50e7d25e2534905cb2fe09082fccee3c774
81833 F20110331_AADKFQ he_j_Page_049.jpg
1e9812d2e89181b40bb27ea4d6dbcf70
1cfcbc58d3cd0807b9d4092b88a9c35a9fe6c677
16385 F20110331_AADLCK he_j_Page_283.QC.jpg
cbc8d61d4a70badd09fa94f2807a7e78
c15d050a1087e0ef718f9b005b0680b377ad7ab3
F20110331_AADJIY he_j_Page_043.tif
b576185ecede311a352b1270932bf494
d26e66e24c347fe7af19be1a5eb2f66757c573b3
106845 F20110331_AADKFR he_j_Page_051.jpg
e6c476eb7773053443e3ad1427010e5b
9f741b4dd7efd431af3f33e34e87f5749054cd13
18634 F20110331_AADLCL he_j_Page_284.QC.jpg
8f940357ec9fa3a3c404bf412b51a960
781cc19e72bb0e76bf590cb5b759c54f33edb0ac
F20110331_AADJIZ he_j_Page_045.tif
bb8373f2ed4ce11bb5c40faadd6aa528
c608de0d212e2023d8e9801d4a323418365c5bb8
108031 F20110331_AADKFS he_j_Page_052.jpg
e544a804d2afa9b4760c8ce05d5399f5
ffdb2e1dfd6b5c9f12aa5263628645584f9ea799
17895 F20110331_AADLCM he_j_Page_290.QC.jpg
55371f7b4011c0178fd4e883458bdc84
cf4d8187b5928568feccf06e28f6e3b8d66a94e5
109239 F20110331_AADKFT he_j_Page_053.jpg
5f44152094644e732337f6d6b1fd6532
4ec0949c05c8e2700b924f11ea2fe0b5a9eeed78
19118 F20110331_AADLCN he_j_Page_291.QC.jpg
8d561fe4550290e772e92dae9694ab71
80251238a1edf743b6e68d07d728ef500c67f3b2
88096 F20110331_AADKFU he_j_Page_054.jpg
e97917f1cc1cd41e81a2465fa049e9a3
a8739fa17daad4a2f0b7767274ec479c547ffb7c
14158 F20110331_AADIRA he_j_Page_196.pro
521a3e406c092941bd70d8ba0fd30032
28a0e4e2e7578344bb8462a1618e3c0fd2a97e4e
33263 F20110331_AADLCO he_j_Page_295.QC.jpg
0a1935bccac488e426ee95ce94b084c0
0ec695f6bf24acc5f0205e379e0a6aa977b59bf0
F20110331_AADKFV he_j_Page_059.jpg
3b134b1e63ab6baf10e477388cd7b791
c7760d5ae5d580b956de643f4045da4666614067
48199 F20110331_AADIRB he_j_Page_196.jpg
b235b0848db402e1d5ca9e0ef2aaa271
8ba8848a0b1c92833692a7173304b0f260eaa100
28967 F20110331_AADLCP he_j_Page_297.QC.jpg
aa1b70e4b242042d8a8ef1c61c954af1
1edccc1cb537d304bfcf2b104608fb6283de1cfa
109426 F20110331_AADKFW he_j_Page_060.jpg
b0c6324e07654446c7bd9dbebe6dacf1
84f9879fa4b0d4f77c593e737123ff9b6d0a0bbe
85158 F20110331_AADIRC he_j_Page_137.jpg
61cc8b95c775ffb465aa695391778b3f
09ba3ea58e7592a543456b4d61ba6699f92d24e8
31163 F20110331_AADLCQ he_j_Page_301.QC.jpg
87b99d81956f3dd672a302049015cb73
2f497cfe6cd7ff0d12f30343cea7c29193761c43
32240 F20110331_AADKYA he_j_Page_143.QC.jpg
52784104a2cfd2965531d4856b628507
93a55a93553c5475bd48aacb194416dabb492e22
31337 F20110331_AADKFX he_j_Page_061.jpg
e5f37dd10ed9eea409d0e41e84cd8dbb
68bf564bc8ede3cb806dda5e1ea952133c4b2c44
21506 F20110331_AADIRD he_j_Page_092.QC.jpg
33f8b6604ee95f553cd5a7777353cce4
cf0e539642620cf7f3d0f4c446c8f5c4bed587f9
25568 F20110331_AADLCR he_j_Page_302.QC.jpg
0fcae43ca641b86d300b5905532b6953
7c1508b76dce775d78761b3beb2f30b62481debe
35553 F20110331_AADKYB he_j_Page_325.QC.jpg
23548fddec231a07c84d8ea1f50d8d23
9591f8e59a758c2d45f1f130fea68214f0578158
73519 F20110331_AADKFY he_j_Page_062.jpg
aa68c77100e8a75271780bbe859b1df1
e2b90d87bd1f0b5c163a968d855c69dbe1427819
7691 F20110331_AADIRE he_j_Page_039thm.jpg
7b4317b6634b15a08a0b9a1bea83bc1c
b412f04deeae7e07ceb348b4023bca50c864635d
15702 F20110331_AADLCS he_j_Page_303.QC.jpg
96e7c808301b6b2e4982c88f3cf4798c
90b3cf2040429270c3041e294bd8f4822d098edc
F20110331_AADKYC he_j_Page_045.QC.jpg
209a01d137adfb6e5e27f39e6f9a9150
9fff3709a4953e5140c7abdee4f49e19e83c7d89
62084 F20110331_AADKFZ he_j_Page_063.jpg
256e3dbda3b8b6994df66f913abfac79
728f851040ce56dd8f5c75ad717059624255198d
73005 F20110331_AADIRF he_j_Page_235.jpg
b01bfaf5d859110dc3780e1f723cd61b
e083278a3fb30ce58740933badc8f695c261e565
2285 F20110331_AADLCT he_j_Page_304.QC.jpg
e1824b5f73417493eedb616550e42794
5c800ae77f59c951c7a62885fb18e9821c29b21d
36257 F20110331_AADKYD he_j_Page_239.QC.jpg
0178d7df362cc84c49e0ce013ef91518
45403052e2146856eaa79b972bcfb98bf9b9e647
44406 F20110331_AADIRG he_j_Page_245.pro
c3840dbcc2284b5c82ff18f0c6c6fe92
11a2e8ba2373f0be0d8b131343693030170c832f
25810 F20110331_AADLCU he_j_Page_305.QC.jpg
7396adf1b0c4ec48c0d1c0ba4b2b4ecc
853df41862690b6a75cbb06884a3baffe9386689
9465 F20110331_AADKYE he_j_Page_012thm.jpg
e698eb8c4d94730e2708798a09fa6ffe
2e4d74602e842081a1798d12f8ffdfb408eb3d3f
F20110331_AADJOA he_j_Page_278.tif
d9e815da54e5659619bf09e25608435d
a60fc46aebe2d1a2466746c74c1e220934ade90a
110807 F20110331_AADIRH he_j_Page_026.jpg
af83debb3d11c77273b99f5cf362260c
8e88f1ccf74ac69f57674f935ec72908bf8218f4
18288 F20110331_AADLCV he_j_Page_309.QC.jpg
1438fbb281c1748cd6e6c5aabe94da84
152f1f2563f55c2d0b8bb8c8f4f232922cc8f5d6
9615 F20110331_AADKYF he_j_Page_326thm.jpg
332c48241fdacd5c698d903fbb13f094
e99ba8e76a3188758ea75341957076c2f639912c
F20110331_AADJOB he_j_Page_279.tif
4c3a306ddcc4104aa2d5b444c01b79e4
4e6b959f27b0564689b2ad53b918b0cd71f01f25
102193 F20110331_AADIRI he_j_Page_214.jpg
352b75ed9b72b0c01330ee73a868b426
3e4f8fef65d23e5fa3fb293795c927183c0fd6b3
18072 F20110331_AADLCW he_j_Page_311.QC.jpg
2e3e3fd667892d67c35fe427cfba5961
456cf9c6b3d4744e85115aad0de9b2ff4a6367b5
518736 F20110331_AADKYG UFE0022077_00001.xml
651b735ac7b71ca982e05afcf7e58360
774fa421b69fa534e80a7408510bb9e75ef8e8db
F20110331_AADJOC he_j_Page_280.tif
5921fc2e122140de8710a74b6be62b8a
e4085a9f6138453160554e994f373437ccebf0ef
F20110331_AADIRJ he_j_Page_096.tif
e582a568a2745f685c4c2412aac73c11
994f7d4783ba65fa5292a989685c243e371a2fff
39548 F20110331_AADLCX he_j_Page_317.QC.jpg
5d8b6b9fc60e85f48b0c0c76b21fa7e4
e88ab249a5710f093554c281c599f50360733681
29886 F20110331_AADKYH he_j_Page_006.QC.jpg
a08273c8c45a583d5bf1f8d22cbfdf2c
fd3b0f37fac59cfc4d374113165b07abf3b0a4f0
F20110331_AADJOD he_j_Page_281.tif
1681a586587f447836da033ae51860ca
3ab87404e7afef164fa7fcc85e58c06f827373b4
F20110331_AADIRK he_j_Page_202.pro
d41d33921c26b2f2a5b8782012baab18
2d46e5b2e686ee535af100a7cff9bf92c40cdc37
38460 F20110331_AADLCY he_j_Page_319.QC.jpg
29470c3594d0070359179bee2ebc9ecb
83251f90ae200283de9aa224a6b5f70a89228394
41147 F20110331_AADKYI he_j_Page_012.QC.jpg
d5533882e4863f176d8e1acc5407e52e
2160d751ffd3ac2804bb682b564cc46ff9d79f68
F20110331_AADJOE he_j_Page_283.tif
9500798a3ec0c296a5ebc949d3a3e773
8898b607901530c5b18e2ccd78b94e41d3dd7522
18855 F20110331_AADIEA he_j_Page_313.QC.jpg
1015ad1a29235f51786ffc6c4be44c51
3125ebce87eb9dcbf474a7b9bc6ab3912c9d5c91
74600 F20110331_AADIRL he_j_Page_108.jp2
767a95eae3cdf02cb12841dc4c9d1671
832035528fa55ad3802f673b9f7da8c2dc8592ac
40187 F20110331_AADLCZ he_j_Page_321.QC.jpg
2d6be0433d3e5a2424d4952a0a215053
e6e511499a448fca948503a7834116d1d65facf9
36966 F20110331_AADKYJ he_j_Page_013.QC.jpg
dd21d3beac9746d743c0cd3f83f24c85
c338ea009a7ec1943692f0b1a7a7b205654b939a
F20110331_AADJOF he_j_Page_284.tif
5480c86e9bb03ea64abe148bb3196fde
c8589789b290eb74d1f904ddb746e90b77e0bdad
51958 F20110331_AADIEB he_j_Page_184.pro
72f21565de8f0962a47ad996851f2438
5c4f2589c01716867a01975a124e56c0d09fd4f2
8034 F20110331_AADIRM he_j_Page_007thm.jpg
a3b35cd842d8eb47096fd32e58e91ce6
06ba2e861b3a907d5b05ad28a40350bf61b1a461
34151 F20110331_AADKYK he_j_Page_014.QC.jpg
01a14e44a94deb4830aaccf1c157ab65
bd4854122ea8be2066cce7b0ce16ef7909dca3c3
F20110331_AADJOG he_j_Page_287.tif
735104d66436ba32c17e60869e986ff3
f5085a64792df369efc028d2a94129609e815bf6
98913 F20110331_AADIEC he_j_Page_070.jpg
6160918285f1fce3a1779b5ab5d32059
dfd46b4d9af5803f007f4284c7516af9ad2a108a
11217 F20110331_AADIRN he_j_Page_329.QC.jpg
daf1762b3d20fe6d051f4bdaf298eb1b
ff1dc3704ebc913db2f4677e83313285e396e2d1
28462 F20110331_AADKYL he_j_Page_018.QC.jpg
07a7eed8edb22b96a2ed8a9eaa097456
ba6e0020f9cc2e39bc982f8cf288ad421a0288af
124615 F20110331_AADKLA he_j_Page_299.jpg
583395d1b33af69275c6175d43b33cbc
bc56025efdc6c89bb2f79a94daeb4dd5e9fdf50c
F20110331_AADJOH he_j_Page_289.tif
fd14f28f06682e2f4bc5ef735a327345
571c67fb3296bf1a70429bb76e46830e1f25145c
35338 F20110331_AADIED he_j_Page_296.QC.jpg
410afabfd703f660df7741188d496885
837b95708e43be826a34350a32d1de2dbb19ab72
39729 F20110331_AADIRO he_j_Page_168.pro
c68899772331f2584d7d083e462470b2
6ef99f356ac06684b0314a19e3947c285599d8c9
34326 F20110331_AADKYM he_j_Page_023.QC.jpg
05ba13895a4c0389bc1ed7e2843aee4d
dc6d8419e198c7105fa88f65fb7e56a663caa403
101621 F20110331_AADKLB he_j_Page_301.jpg
0792acbeac23a4d5b3fac14e081e204e
3be4744af3568881fd386ea7ec52d43920fd927c
F20110331_AADJOI he_j_Page_290.tif
5751b470362f437638129e6f94f29aac
ba8059ad02413518b2abdd609b4c7722c75c8019
33664 F20110331_AADIEE he_j_Page_181.QC.jpg
ae13f325f119182ece8b8f4d86a1dd63
959f97fef1b246d0b975cdebd5f6d7e5af455d47
71868 F20110331_AADIRP he_j_Page_020.jp2
a3c0e87bd0f6b7b0f4d7b44a9e8924b9
1e18be3dc0bb28fdb84a482b4ab3cd091d514efb
35769 F20110331_AADKYN he_j_Page_026.QC.jpg
a6eda8644fa09497fb7417c9a45ab34b
7aad5b489d9d8f20fb3e72e9c5f729e550b3e1e1
82144 F20110331_AADKLC he_j_Page_302.jpg
39e711e7e31d0261e76d20de46cbbd46
3792efee98afd4997b4273b49a138b5b6accd87a
F20110331_AADJOJ he_j_Page_291.tif
5eceb2c6e697c048b05547cecadf854c
70043b403e4655b412f52cfaddd96501462d5fb9
56647 F20110331_AADIRQ he_j_Page_290.jpg
178765bc86bee5803f749d8a31403737
357f48b38ca858c2a63335914f35fa5fcea5b9ea
36369 F20110331_AADKYO he_j_Page_027.QC.jpg
bfe209a2908f92c0fbf9830970b405c4
c41e94f323f9f004da70456dd5fdcb86c68e133f
52020 F20110331_AADKLD he_j_Page_303.jpg
d263a1ebc6b1d8db467dc33449a437d4
b8cb29650800cf38f0f1456b97f196ec984c4980
F20110331_AADJOK he_j_Page_292.tif
fccd8ae2e1cf6ce70fcdb12e18d73cb2
6fef24b71691a0ec12ecf61402907d0a80deec13
505798 F20110331_AADIEF he_j_Page_116.jp2
c6123f074edfe632cb37f72d3e76ad59
2c54855ef9660591ca116ae9c6e6631fe08f7079
18982 F20110331_AADIRR he_j_Page_262.QC.jpg
2daeb6600172e3cd9c2727a76799d3d3
807cc1bdc91a4576ba58151964d72c5f237e3d67
35359 F20110331_AADKYP he_j_Page_028.QC.jpg
a7b7d49cff6bbc2cd189c5fd0b95e7b7
c2a32b76595acbc210ac42b8a5b2fe6711aa7d56
7002 F20110331_AADKLE he_j_Page_304.jpg
40f918bc9d682f050be36b269a4a1bf2
3cbf6e973e5abeac547aab9b9db08c6aea01a2e5
F20110331_AADJOL he_j_Page_293.tif
d6eecc93c54e6044d5b6d46516747fac
3c315c7c5c8457bff7f4d61fa3d343bccefc4c28
89612 F20110331_AADIEG he_j_Page_042.jp2
8015fc1d4ac72c6f1d265479b5c26426
6276c33461d241dd931c21d7970d44bbbf21b6b2
234 F20110331_AADJBA he_j_Page_230.txt
5dc237e8fc83f241364f7ef464763ab3
cd60ecb4f11191d129f2aa39a9ef220f4b703ad2
37244 F20110331_AADKYQ he_j_Page_032.QC.jpg
145cdc969982f247a354bcede4470828
6a68276ab13576d09ef1ad7b8c57eb097da9ee53
75465 F20110331_AADKLF he_j_Page_305.jpg
46a88a0b54a6efa5360c9d472504edec
6079325f62dc8d8995affc05e9d317d8407522ce
F20110331_AADJOM he_j_Page_296.tif
a2804b345738498be4b6a778a3d7bcfd
1039cc709acc79bfe961b0cdc1d4eb1524e3625f
5438 F20110331_AADIEH he_j_Page_313thm.jpg
00e9a5e90186bf867a91dbb6f7bac13a
ed5804c09e788dcf988ed1042676c496fad99825
119340 F20110331_AADIRS he_j_Page_053.jp2
a31d3c15f6ca2c44421053619855d0f0
145ce1712e4e8d3323a64b2ec39120b534b1008a
34182 F20110331_AADKYR he_j_Page_034.QC.jpg
d52ff9b851ca6ad7a588c2fc3d66f87b
1f246c2b11a9f60577a271dfb52c568f5fe2540d
59436 F20110331_AADKLG he_j_Page_306.jpg
81b540ed00fb01d3329d05ec54cc5439
2961aa237aa76c0b73e9f78c50800ef55ef97766
F20110331_AADJON he_j_Page_298.tif
d63f3e05200b30504a80ef08757a2a9f
93946932361d127d075441a06aaff311ebbdba04
7679 F20110331_AADIEI he_j_Page_088thm.jpg
7fe5d3e55df5cd06ce6e3d5622ef7749
800a91fd2da3247bd86ec4fbc32e4e3c2491875f
32757 F20110331_AADJBB he_j_Page_264.pro
44d1693c107c712bca654928c425e69c
8bc2b5a4f99565e707729924c491e873e22be70f
83393 F20110331_AADIRT he_j_Page_202.jpg
67c2323e52e5c1faadafe222834d08a2
6c8c0a409e904d60b84990872eb46290676d0a09
21736 F20110331_AADKYS he_j_Page_036.QC.jpg
f8eb30184c86d85358b26832dfca4c96
37982ffd3c8bfac19f8fcd4182dec7da5f69f223
54028 F20110331_AADKLH he_j_Page_307.jpg
4d79952d212a75ee4d2f78f7dcdb540e
174cb6ce2ce9ffae8cc7ba03c1ed54e1ae06f0b1
F20110331_AADIEJ he_j_Page_029.tif
7721b5d04689ee74e54c5be9d006a1f4
736961508f92bbee1529d8df0c873151580125b6
2398 F20110331_AADJBC he_j_Page_125.txt
7219ecaaffaad17c4944ff7518ec4c9a
d75cf9e470aa7ba1c820b4389f52963472656c5e
42819 F20110331_AADIRU he_j_Page_320.QC.jpg
aab765298159077f2eca2e1e6471d1bc
8d43f4893a798a8f9a62f2a18cdc0bcd7b115a77
10349 F20110331_AADKYT he_j_Page_037.QC.jpg
0457331cf376c5a247cb451bd21fb671
b470c4d1ae12ea02c0ada0f6135b0ee046cf5bd3
54052 F20110331_AADKLI he_j_Page_309.jpg
a805ac46ac0240fa01b81c216590af8d
46060f2b345df8e2d6139cb487bae0668eba7086
F20110331_AADJOO he_j_Page_299.tif
8aa251b86fb55eea42fa7f921e467c95
02998bf4e999317c2adbd1da2c925a3b4a9204ff
3255 F20110331_AADIEK he_j_Page_067.txt
c1db3e6e504ccb0b630174e96f3a6f86
3897d215c1dc7771bdc5dcf86495c8ebf735fccb
F20110331_AADJBD he_j_Page_042.tif
87345d17e6030974391d2074f74fdfee
3300449dc605c2c56bde695098c653776919d09f
15566 F20110331_AADIRV he_j_Page_258.QC.jpg
c3ceda070b27edb5e55f84f0912a2864
d51d55f7ec56427c7001a20a3e0c55cec01246b3
33793 F20110331_AADKYU he_j_Page_040.QC.jpg
192256668ab4874127f35546e24894cb
f77a4dbbdf16f622b875268f80e02a09713120ca
54211 F20110331_AADKLJ he_j_Page_310.jpg
a4850d1fa9495a1078371e95b1339ee7
c9895567fce871479825b36c375cc8654a9fffb1
F20110331_AADJOP he_j_Page_301.tif
c99f8943a45f644a7dfcbb2096a15ef2
d72a10341e38232ce5143be4c3d3416450310317
38165 F20110331_AADIEL he_j_Page_009.QC.jpg
429366df75cf2188089d41c12a98d2b4
eceff9b6501b0565df95982373b73bb9c10c9ee9
8712438 F20110331_AADJBE he_j_Page_066.tif
39967482b490942f03193cce1e89413a
e1b8634451d8a0955e41b4067c99a4a6c1956f6f
33981 F20110331_AADIRW he_j_Page_108.pro
db9eed99c38f05954a2d4bc12b527998
0a523f5dc0057b8fa1373cb19397cb6cecbee3e0
33780 F20110331_AADKYV he_j_Page_043.QC.jpg
7680a94ca4e8e4beedc1b691854a69bc
39736ea3052796de4b400763143f41dadfe5d209
F20110331_AADJOQ he_j_Page_303.tif
532c80236b0243a95f19526bd3279307
997db898106884df97f7af4e2c6e0cec0625d3d3
114180 F20110331_AADIEM he_j_Page_034.jp2
7ea05e2bd91ad6cf004d8064396be61b
bece1a0b474a8a08e029091de5710c91b75acf4a
116935 F20110331_AADJBF he_j_Page_300.jpg
528d8ef391688abd79dd94296bc70e8d
33dfd223bfdae3002c165204a686d66f91eeffaa
F20110331_AADIRX he_j_Page_313.tif
3bef716d2a4f94669e1d3b51c256241c
bac366d8d44a36fef53971f4a97e0174e7aa380f
34906 F20110331_AADKYW he_j_Page_052.QC.jpg
3f4f48b45c2d41aaba4fd9c357c2973d
c7a3f74c3d853b92d661592c1803609521e08df0
52995 F20110331_AADKLK he_j_Page_311.jpg
8284bf948ef573d3ecebe343a1104485
0364fe32f2a8c7ce95ef4bdf7b725da55b4fa6a6
F20110331_AADJOR he_j_Page_305.tif
4d23e84117129660078f4d6ba1bccc1e
a6c4ba5993bc075f57b6bd74b553e3d301ad35da
1147 F20110331_AADIEN he_j_Page_261.txt
49b5daa1b4940d7a90aca7ca07d76e7a
c5ced4fd4e472644c184e366db545262623d6dd4
101603 F20110331_AADJBG he_j_Page_105.jpg
f95dc1f07f2ad649c191aab24bfe1ae0
0e2011957203234ec39c71b093d05cedacbeaadb
F20110331_AADIRY he_j_Page_251.tif
c121b4ea1ed76f44d5bc80fd45b85b8e
35099e099059d63f7a0f38b7ca2a77fdba947447
57537 F20110331_AADKLL he_j_Page_312.jpg
f9ce37cee277d608f3776dff8891d440
0d1663c7b812e71a90a5d70539ae98739a086144
F20110331_AADJOS he_j_Page_306.tif
e7ec3d92b957b350b54593fecd3bfa8e
60817578b6fbce08229ed8af39329dd1482768eb
29802 F20110331_AADIEO he_j_Page_173.QC.jpg
a239358bb5ca7b1ec6c90f9325dc7281
43cf57d5346efbd7560fe25f138a7931effaafc2
107537 F20110331_AADJBH he_j_Page_079.jp2
0b602206ee89c43cc5500b7871b832b6
4ed3cb22b9445aa4b9402817b43499c793f7029e
14698 F20110331_AADIRZ he_j_Page_329.pro
c8d4e29d53f6e76002b145dfc8422580
14a7eac21666c9d740a7f849f541f32acff62298
34799 F20110331_AADKYX he_j_Page_071.QC.jpg
6a4d994c9a03d7c3264909cee304db92
21ab2b53d0d9d1ad71e23edf24e21761dd446688
58320 F20110331_AADKLM he_j_Page_313.jpg
154aa6a2f5a9d4d863e1b3cd0a506c0f
8536a2dc6284fa79478f10b647bd803cf486f83d
F20110331_AADJOT he_j_Page_308.tif
c5f23ec1e4c586aa95d59fd94f8a7568
ad0fe6324b5fe3f11f6864a9321b4bed020a918b
F20110331_AADIEP he_j_Page_191.tif
ba119a3825a2dae3b695477d05d5a947
b46a187c34e4c231363def1c2d01509121f327ad
17455 F20110331_AADJBI he_j_Page_120.pro
ef2d4644ea92f8b7eabcf000d19be271
98c77ca99d26311d2d969c8d756eea9175a6fab1
35401 F20110331_AADKYY he_j_Page_074.QC.jpg
ab8dcfbe7896a52bddcc02db487bcb27
78efb974cd9484180359be7a04b2937067140d2d
129179 F20110331_AADKLN he_j_Page_316.jpg
7e881a9a72f0f2d06f11be48a4166fe4
f3fea0b105791ea4fe85fa57a1c7837eb5f4036c
F20110331_AADJOU he_j_Page_312.tif
b8c21d42f5bf1f0ea5f6dde42db9c07f
55a218069ea759d38f41e0c242f9707166db021f
106955 F20110331_AADIEQ he_j_Page_100.jp2
9efd1ba210b7fc69f14fbe7343680af7
137832de05b4caaac08e6bd7243ac7100d0b8a1b
F20110331_AADJBJ he_j_Page_051.tif
331e004233dbf87ec59b999920c19b34
e1ec83731e6b8bce5d561dd41ff926ac376d1759
31340 F20110331_AADKYZ he_j_Page_075.QC.jpg
1a5325cdf96601c45a63373170fe5016
264fd30a08baa4ea274177800a559a0b6040879c
143423 F20110331_AADKLO he_j_Page_317.jpg
74c6ee4b7ad203635cead5eb1792040b
fcced93ff5554df08f7a19e96727fb0c1e72547d
F20110331_AADJOV he_j_Page_314.tif
32757e4805247ae21989e0135eca4407
372c851d36454704a5467fd61911ea342369e653
31942 F20110331_AADIER he_j_Page_212.QC.jpg
748edc276e45c52933d955ef62f9b091
918402bbfd6467292a7096e848edc4c909624701
92184 F20110331_AADJBK he_j_Page_050.jpg
b64d2357dce7d6a3810c85a4b5cc76ed
aa7a5eadb75e3470020263b1799522aaa0cd2fd8
142288 F20110331_AADKLP he_j_Page_318.jpg
c1b06337d82c9139922e58b6b0758102
a8dd5006ea019c83dc928bdb49c78776b27f0e31
F20110331_AADJOW he_j_Page_317.tif
863c4aef0e1e5b1a042daed30fdb67c2
3b39ab4bb0c4d42e69e36272655d5d5401502d6a
116429 F20110331_AADIES he_j_Page_151.jp2
90ab0974af835f651e93c53e328834c2
d486bf6d62c114847b954ed896b298cddb9fbff4
36576 F20110331_AADJBL he_j_Page_282.pro
2ac5f1098ecfdead78f7f386a7fa15bf
2dc0faab13c9a4b4b318d852c88d7397df604cbd
143095 F20110331_AADKLQ he_j_Page_321.jpg
ef07abb99b2887e64d33b897af35cc0d
4dd12a6eeb6f07145517278bbbf28b05bd6c1abd
F20110331_AADJOX he_j_Page_318.tif
67bb7e641aac6025e419ad9a38021e5b
5c6db0fc2b04d36aa266f9137ea22c56ff57c28a
118012 F20110331_AADIET he_j_Page_033.jp2
70755a66017e7972142e97ef2e4d44db
5d3eca30dcb5e4e41afefb66db346fc6ebd01b59
2219 F20110331_AADJBM he_j_Page_253.txt
eddbbdd17b15db768addc515cd05afe0
9856d1ffe515dd9133a569e0a82206ae4af7eab3
133706 F20110331_AADKLR he_j_Page_323.jpg
7bea908d4dcd1ae2044a7ad508177ca0
39a0a7ed10aff6e1b64341bd61111ff7ce959202
F20110331_AADJOY he_j_Page_319.tif
fd4acee70ba196015add9185a957f7a7
9c00190ccac1eb5e9570fe00ed89b263232c3b1a
514152 F20110331_AADIEU he_j_Page_117.jp2
a61f195b7ed6d2478a66e2b2bb59a351
8c108be543d8b21307cedf13db203977fb0273eb
F20110331_AADJBN he_j_Page_170.tif
393a9b8c46ccef2e3cd337a8079f598c
3cf3c7808dadc174a2fe820df1e4f70c3277a193
128402 F20110331_AADKLS he_j_Page_324.jpg
f22c58d8ebd6cc692531880391560ce6
833385d79e7c4adb6fe2cd5c6c5c657f267be6c2
F20110331_AADJOZ he_j_Page_321.tif
f640f6b31a43ebe847ac0b72ef1602f3
cb87522a1e9fb0a5089e906f183321d54ca89202
8327 F20110331_AADIEV he_j_Page_214thm.jpg
45ed5026734790da0801c7d352fe04d0
133d0fd1ba74d3c0dd67d4b340298769c9b4e869
F20110331_AADJBO he_j_Page_261.tif
1b5e37372acbf2b479c5ed41bd344494
14ab13a5577df4718a4588695759dcc2d5cd59d8
124151 F20110331_AADKLT he_j_Page_325.jpg
640d8f6bdf9a28b19fbf43b4d693977f
af2cc98aa2fe20a63f3a68226be6e7e8e9337b52
34033 F20110331_AADIEW he_j_Page_249.pro
ba152bc90b85bf6689b81ba82a6897f7
5e1ca024d81c47086ec0508f36ec77dac75b14a2
4698 F20110331_AADJBP he_j_Page_281thm.jpg
691e191322c28a8c935072d2bf49a14b
c37c33871a4c74b26ebc6540790e7bf3afc7a8dd
143780 F20110331_AADKLU he_j_Page_326.jpg
b6ed58140f43ecf5d444906607807887
e1daa3247f0d27d8e39159f41c5de3e688882391
88407 F20110331_AADIEX he_j_Page_260.jpg
eee27425dc9572be517829a6f65f0ba9
28aad38815a39d4d9f34b90b429f350949ee7b59
107580 F20110331_AADIXA he_j_Page_138.jp2
f5c5d3bbd92653122bbf8e74916e406a
36a1fc7c1a5efce887d78c37535b38ff2ca7a4ff
8689 F20110331_AADJBQ he_j_Page_102thm.jpg
dd35091c3476d359328a5f9c6f078c67
a1a1b608cb65f401bdb1512e5e778aaf7b7f9839
144123 F20110331_AADKLV he_j_Page_327.jpg
401bd2c57b5a7574d9776813ddb3f7ff
16c451fb364e3627aa17f441925b0c5fcdc0d1d6
60632 F20110331_AADIEY he_j_Page_278.jpg
fa97b4d25dcba54fb59910a33332d9dc
081f487c7937355c11a96488b648c5273aaaa3bf
1316 F20110331_AADIXB he_j_Page_312.txt
cc0f3e877f1dca51571dc7c0456e57b8
ca6b527f95252664fbb3054e4f8b7ee16268f27d
9573 F20110331_AADJBR he_j_Page_321thm.jpg
f261c4e48c8627b66c2c1c7700518b73
fee0e6621f6d493a6d1cc46d8500fe0a98846584
56678 F20110331_AADKLW he_j_Page_328.jpg
af4adf60f3b0c6f98721ffacec270615
a9bc97f2d9a6f780bd609c093069123ea657e4f1
F20110331_AADIEZ he_j_Page_247.tif
c9ba0bfec950ed65486a4dbfc17e9b73
d06455c14a61c59c20d73249a10f688644f775f1
F20110331_AADIXC he_j_Page_222.tif
4b54e760a3bd58f6c82ef22888e30ee8
8c469abc7dcaa09efd1996c494dd2adae9029131
2111 F20110331_AADJBS he_j_Page_254.txt
3a4be72b1f907fcb2f93f33a8b2a39d8
57fc9fa01af757051e333fd532d7ea06b29990d8
34696 F20110331_AADKLX he_j_Page_329.jpg
f2ad74576acbdfdc8731e4327b2486a4
84ecc661d18d5f98a074ecfcebcb546b35fcd3ca
47661 F20110331_AADIXD he_j_Page_205.pro
ec84d2cb232caf44c7dd92bb6b29a72a
fb46b24f1564839aaa1bb72feba50d2669c76e35
55500 F20110331_AADJBT he_j_Page_053.pro
2af83e1c404faabb8a2880067271ba60
52671e5aec87666a3e60fde9a51c8c9e31a1bf30
4950 F20110331_AADKLY he_j_Page_002.jp2
9e3ce707b565347596e41bb3d05141dc
c24edd474021334e4b22f64141fa6bd42c809acf
F20110331_AADIXE he_j_Page_157.tif
8a059292a77fb53f3888344632b646f5
61c33e8600a054291face386d40d142528ea5ead
1773 F20110331_AADJBU he_j_Page_072.txt
50362e812e8c4036a670c82787ceffdb
a56996bab9688f34324ca12a361cdab73a65d76c
106020 F20110331_AADKLZ he_j_Page_004.jp2
70b0f50b6889e1d48bb25672db90f519
a8090c1bbe638ce9c2d6742eaf7bef3b29d3f3fe
106307 F20110331_AADIXF he_j_Page_081.jpg
85f73864465b4f65624a403395c4a56b
e88aff5b2bc058e8ee3c168f139f15477747d29a
967 F20110331_AADJBV he_j_Page_194.txt
8e89aaa5d8828ae5fff3e6458f7e5a1b
4776262cee974a3169b973d5e4f795d99583b11b
2022 F20110331_AADIXG he_j_Page_103.txt
ce66b8a48e3bd7ecd6db60e445b01bab
3e32e7f42b8b475e7d9e95eab8fcfce3d5b7ffe1
F20110331_AADJBW he_j_Page_221.tif
e0c505a44e6f9725cb4d5605527dcb44
2d94541e32ed30e787cdc8f494cdf731f00d3bd9
36193 F20110331_AADIXH he_j_Page_281.pro
897b34f83bfaf4b37f604425c25500ba
382c314101f6adb5eaaa34673d71307f745b8a72
F20110331_AADJBX he_j_Page_181.tif
caec10a6e49e82e6c738da4ae2543540
641c88cd10f23c80012c7296a024ccd9ea314f3c
2358 F20110331_AADJUA he_j_Page_206.txt
038fd39b4dd35abeab908bf8237f315d
24a043158aa776a140a4757040f4a3d33a063b2b
14793 F20110331_AADIXI he_j_Page_197.QC.jpg
8f188b2c301d34f6e31d8322f60f4a93
b619e55cdfb3aabc3fe30caa5a3bc158a3726322
44174 F20110331_AADJBY he_j_Page_197.jpg
d137537a5da33a455f005c39dcd4130f
8d9f5b5df923380a1300cf3df39fc95bc6912f7a
669 F20110331_AADJUB he_j_Page_207.txt
6624cf99fecbd2984e0507e551ccb5ec
d0cecd06e59845588483eae49ecee22e768b0ef1
F20110331_AADIXJ he_j_Page_184.tif
545ec9094b40816b5abe12af9314b527
b2c9914d30e58ca8b490942cd69628bef358a094
13538 F20110331_AADJBZ he_j_Page_111.QC.jpg
21e5bb97b346791caecea1c43fc7a8a2
04ef5242496ecb33daa70e819cf548ff7656c8ce
2068 F20110331_AADJUC he_j_Page_209.txt
241b94048c97add1ed3df2c1d3ac2d27
551e12c89c92f7e237f4a33758d8da5ae98f7ead
F20110331_AADIXK he_j_Page_195.tif
ca60c960f2b70f8de849a071fbcea6fd
f7371ece2e1322b12f183cdcc5b53ee3e30600f9
2024 F20110331_AADJUD he_j_Page_210.txt
48d70a57fb913b7502bd5ac6723060b8
2001f55ffeaa291d7e4d5501fca363260e5ef745
81046 F20110331_AADIKA he_j_Page_089.jp2
961aaff97aa57c43f025ed130c591845
06784eff33261c17aede9937056ab1f32b9657bc
62273 F20110331_AADIXL he_j_Page_122.pro
3921e46757605aec03e87b4304d972bd
8fd3eb07ab701ea9c5719434465a19b07a506ecc
1927 F20110331_AADJUE he_j_Page_211.txt
f1765a16780c2fcad6257c6ae68e5174
ef1ab984c9fb5caf56bce7b33ee2a76f16f3f0c1
65389 F20110331_AADIKB he_j_Page_262.jpg
547cd43b8941fdad41b8bfdc9e953a3e
64c60d71fb256cc3f68bf538163a3868298baa30
109316 F20110331_AADIXM he_j_Page_094.jpg
d67663212357894f6447d1a39160725e
e20ce19999e1cb7bac616f90b88edcafd8c2aef3
1921 F20110331_AADJUF he_j_Page_212.txt
a2b20467f5d6ea64f68ab0ad4334f223
753cdd49b4668e86b19aa6e0b1dbdd10f7d7682f
35130 F20110331_AADIKC he_j_Page_262.pro
bf016253f567c04a0938fd8e33381f11
315b0aa9b611bf37223f3a207b6773a0322a3ba1
48064 F20110331_AADIXN he_j_Page_099.pro
4192130d18e7b71d941463c32ecff538
f2159aa61cfbe2506400d9b7aabb4a30e64394dc
1996 F20110331_AADJUG he_j_Page_213.txt
bd440483689b15b3f44be8e8576572f4
71273a198a6bee16ed2ab887df0474c9286987e2
81520 F20110331_AADKRA he_j_Page_233.jp2
852a4648c501cd75aaf357513bd3f630
2c1c4bf87f0e2be18767e9f690541fe38cf8a078
5247 F20110331_AADIKD he_j_Page_228thm.jpg
dde8a997997dffe94b44fe094867e0b1
95d0f033d540850551aa6dbf239b3d92b4baf3dc
91534 F20110331_AADIXO he_j_Page_141.jp2
cf183aaea2c7c224e45a7ad21877fd54
9d31ef893848478003c02313aea0589b0987f222
F20110331_AADJUH he_j_Page_214.txt
c583336373f1eec68ee877c9de6b7e83
37ca45ac3e3f35c5262fd1dab75545fae9f57bff
79091 F20110331_AADKRB he_j_Page_235.jp2
40da76ee20877b06dcd01285fe178917
2cb7fde6ed67c3d9bd41b036c43928d672f4e8c0
70528 F20110331_AADIKE he_j_Page_108.jpg
050fc573e497cd993d254e02a59aed93
e5517433ae2d6e5138a363364cbbc37121922061
50257 F20110331_AADIXP he_j_Page_105.pro
a6f16b94a8766a5167b7aacbc12b99b0
cba4027d72e74a5887f0e02b0e2c1caaa36e186b
2124 F20110331_AADJUI he_j_Page_217.txt
f7f6982670eb019855d51358fcda27c1
efb4d394d1e97ae5f46cb3c69324001179fd3d08
118953 F20110331_AADKRC he_j_Page_236.jp2
c114c32ceeee706d62d74facf2474ddd
2327c1e89ba8c79ea1e4bc2472e8598c21dff4f4
113065 F20110331_AADIKF he_j_Page_081.jp2
e2441823fe3c2e83a2652363cd40a453
1f188c828408e427998b3ba2ac06c58b0ced6d99
1971 F20110331_AADIXQ he_j_Page_143.txt
12c1a9cf06fc50b87a56b092b7608083
2fbc1866ba7d3048bc2fcb4a6f15f3a6e29675f3
2045 F20110331_AADJUJ he_j_Page_218.txt
d13ea3ce78757033668744e3fb8f6f56
344075ee6e97e0a8fb4cfaefc6c826b82154ea14
118320 F20110331_AADKRD he_j_Page_237.jp2
5c7df9729fe99fc8b81404b3d09d7e2c
4b634dfd2e76fb3bebed265891a88e3cca760fe8
870825 F20110331_AADIKG he_j_Page_305.jp2
6e2237e41eb0f2642699823177edab38
cd25ca747d4285b4bc24f978927344b33ea5c794
21992 F20110331_AADIXR he_j_Page_255.QC.jpg
13f8d504c6f04d0fc3a57a95c7a83598
58fdeaf8edfbc6a1822beb4dead587520ba8a7fc
2144 F20110331_AADJUK he_j_Page_220.txt
2fca32d79c5204f17abc0508728d7238
30230f67e878515ec84378f22671b558db254a16
119160 F20110331_AADKRE he_j_Page_238.jp2
8d78e3675c3f5a3a949b4296cc3715c6
763d62d5ae968a106cd128844e3b31362bd0c9ed
33744 F20110331_AADIKH he_j_Page_069.QC.jpg
1634e42e170c5c7eff634f0e8d384d72
3d43d949dee9cf32f65a042ce26e7748cf307744
1938 F20110331_AADJHA he_j_Page_095.txt
12cd071b7c73b7641767277bf448bab0
08da0522505c530dd1bb5d12c1eab8e071c5bdc8
F20110331_AADIXS he_j_Page_232.tif
ad0512d84a20fdce8a8c73cf9419b0ae
83012a32a128dcdba9b4e5780b051965e558527e
F20110331_AADJUL he_j_Page_222.txt
fac65fac8a077b890142e2317ce42ab2
b8c86180f5994ad33bb3d64fbb095057bdc4fe3f
108616 F20110331_AADKRF he_j_Page_240.jp2
c2f90faa617a0daeb4aed415526aecdb
5136ff7ba0ee2c53c590d338d926a6579c885871
5385 F20110331_AADIKI he_j_Page_314thm.jpg
8b79cff3def09765fbfa4874e3c1fb19
a363f503b9943e33dd974f55fc2a377a547e4516
9213 F20110331_AADJHB he_j_Page_026thm.jpg
7dfb514b731b7446ac41cb221cf586d6
abce3e9f09a57f022a6f46f89ac992841b689e91
4189 F20110331_AADIXT he_j_Page_110thm.jpg
51527e653bb6716bbf5bb7f43f1898b4
52c1ea9a21ff23c9b46c7e8243d1c6d1f846a2ee
768 F20110331_AADJUM he_j_Page_223.txt
2d6cefa5469b5f9be1c34f388cbab271
58d844bc5b90b523c1ebffce16a51fe6354faebd
118050 F20110331_AADKRG he_j_Page_241.jp2
c72ef3092a8f479bd1da4fbc74433f62
34d4f73d75bdd4a063bd82c9aee824c0369984ce
47721 F20110331_AADIKJ he_j_Page_211.pro
1a68e9857eb5e970450af2fa420b8cc6
a2ac0e3681fca4f5e535b1ed78183f9154ac089b
7449 F20110331_AADJHC he_j_Page_203thm.jpg
fd418b2971ba8e5791e85e418ecec584
52e7d3ceebbabf5727626609488df1d81d2aa7e7
F20110331_AADIXU he_j_Page_202.tif
c2f1139fb28b625568a835975d643320
412c863c62a5da013800a2a61a6eb72c97cb59c9
314 F20110331_AADJUN he_j_Page_226.txt
07ed64eac5c62f59aa7b9ca91656ba2f
b3a76cf161d629925913e3a750bc20ed817ccf92
58783 F20110331_AADKRH he_j_Page_244.jp2
b04560f79280beccb85d4aa144b8cc8a
4c1b426e7161039ff1bd42c64d435b0c590060ae
111916 F20110331_AADIKK he_j_Page_107.jp2
54d5a4bd98b1327d34d87cd1e599eee3
45706c0c2e91cf8fbb54893cf69425a92963fc80
2151 F20110331_AADJHD he_j_Page_094.txt
388bb978e3e1f9bae91e9a1b0ba04b53
6537554a8ea464dcd4204cff4fde6a4f777c3b61
5914 F20110331_AADIXV he_j_Page_247thm.jpg
94dde19c9e8c0636bb49bc9c0581b3bb
ab04bd36c26d8f84303f638f4c21139d2856f96b
F20110331_AADJUO he_j_Page_227.txt
43c6cbba59a03321aa4b8af7245520ec
a186c62b7cba1edaec5b9fc137f3b64d2e507a92
85024 F20110331_AADKRI he_j_Page_245.jp2
c66622841cba4866698837196ef6cf45
457235514b411d6ec5335ca4ff202ea875688ad5
F20110331_AADJHE he_j_Page_316.tif
54ef10be0b7f1be64bd41208805a52b0
da8e188a61d31739a523770a7dd6566f18399846
34998 F20110331_AADIXW he_j_Page_184.QC.jpg
74c9e092f5980e8c5d18a38f580c9f9d
8d819707edeb087f40cc51dece7255938e4ce29c
336 F20110331_AADJUP he_j_Page_228.txt
530dddb5b607fd99ebd756c39a749360
24d497fc9fa2e97f72ec77b6268f637ffdc07a93
82370 F20110331_AADKRJ he_j_Page_246.jp2
12168ad27bbb6d252e4342d8d4210c7c
8b7b7827ac803b00ab14d27684b8df913f2ca026
106895 F20110331_AADIKL he_j_Page_191.jp2
ccdb3a49f42d8e73b33f0ee9943e03a4
40d51d470c8d7dd1fad6d04f6dc585d5dbecc144
108045 F20110331_AADJHF he_j_Page_074.jpg
855b1e3cd2212ecb83b4c4f6991c4069
b6e0857069b19a95b3a798433bfc4afd152d64a3
57326 F20110331_AADIXX he_j_Page_268.jpg
283bb1aaed7ce37faf40c6a37437aee2
62455941940521eac46e83c69291429e3ff9592e
333 F20110331_AADJUQ he_j_Page_229.txt
1e862fe1a30489acf471e0a531ac8e1d
92cdb723b573b0dcd8ef59f63eb407337a2e4483
104478 F20110331_AADKRK he_j_Page_248.jp2
a385bd0387089c6badee5e8445af714b
9d09f1a5f69cfd0b14296f5c515f88bc068618bd
27392 F20110331_AADIKM he_j_Page_001.jpg
e12b2bb5cbe1e84d5eca2767c43d565a
7ef946417d3ca565840a6fd9008737744c2546cb
90988 F20110331_AADJHG he_j_Page_248.jpg
41ebd4c9b961b7bd00abce309cde08d1
78171bcebcb1c7ade207d2ba69bf606721459832
1246 F20110331_AADJUR he_j_Page_231.txt
8b6e0edcfec41f72c5abb1baebf9b324
c4d62d838e1f1a1c4881c064cd50c58b07c7af21
62176 F20110331_AADKRL he_j_Page_251.jp2
d96b158d814d8579a7f2317f4fa52ee9
0c67ab27b3c1ae7e30321f87ac5121aba0e354a6
51719 F20110331_AADIKN he_j_Page_045.pro
e3b4357c470ace78c5ac051e3f8fa2df
9bcba6ebc18123ad33c6d6b85829d955f84d9043
46184 F20110331_AADIXY he_j_Page_035.pro
c1b36a3c1dd65139f8a3c1eba1d8ed08
a592bfd1ebbb7600bf4c147cedd7a5558b7348fe
5170 F20110331_AADKEA he_j_Page_301.pro
d1f96e0b242a4de0038954fd444064c9
083e2a1f77a0da39862564cba99041840b3fd414
3527 F20110331_AADJUS he_j_Page_234.txt
7f04e9bdf27bbefc1588cb2f5d5b22ea
6a22c03a953b111c8bb709a1f7e3fa199a529fc8
114042 F20110331_AADKRM he_j_Page_252.jp2
3e08657cd3c93e13acc8a20185aaa898
bc5652dd73faed936ba388d72880bf84a50e682c
36737 F20110331_AADIKO he_j_Page_316.QC.jpg
d87f01f910555c57af04586993919d30
0525a2667dac6e048e4e113898df4279d21b8407
9027 F20110331_AADJHH he_j_Page_003.jp2
a02589c778f479c49a104214a477a196
3dfb62c4ecdce3f94f970e555b1ff6695f1c519c
38210 F20110331_AADIXZ he_j_Page_058.QC.jpg
fb207afa112517d8bcfed2618e2c53be
20ae9911efd07a95ea749970442ed864cd75dbcf
2113 F20110331_AADKEB he_j_Page_302.pro
5efcaaf2731b51111db20871c895a6bb
6fa6c33d6f51ed46e930e19e33d931eec6881756
2396 F20110331_AADJUT he_j_Page_236.txt
87a267ac12dd5a3bc55917fefea4c13c
a456997a6abf2252b69396878bc4da7430048534
85090 F20110331_AADKRN he_j_Page_255.jp2
40216b3ff2b43a82e20e18b773429752
c1dc26519d600c67174758dbec3d9dd6e10b008c
F20110331_AADIKP he_j_Page_249.tif
efdbc4c8f6e3efb5fbbf4896a963a9e2
810ed4b292c692547df223d0a86da8a96f826425
45432 F20110331_AADJHI he_j_Page_047.pro
f63e020160660fb238fb31f03aa7853e
476c3665b520fe179ef51b96d271b092daadd70c
2329 F20110331_AADKEC he_j_Page_304.pro
02c2c17c30b0c4b35a08cbb58ea488cf
dfa11147649856e16da9c659db6a4e83fe445fb1
94113 F20110331_AADKRO he_j_Page_256.jp2
56c74a821c7b96d9c41f9815f877739e
ddadf0db0f61230853fd8c2ab4dc7b1f24eae208
83479 F20110331_AADIKQ he_j_Page_250.jp2
8d9500d15b7173424aac0b35d8b76096
5261bf752af1a5ab729dbf711ff636837e079be8
8520 F20110331_AADJHJ he_j_Page_237thm.jpg
b5e40ffe8d07e81ae89cd2f307fe7400
a3cebbaeffdb9d40de21f009a6c3835ccea78a6d
2401 F20110331_AADJUU he_j_Page_237.txt
366b36bb04a47eb1e34963c3d7a48f0e
e824636f697df1080054480d916ea5ec5281231d
81893 F20110331_AADKRP he_j_Page_259.jp2
13859ce94e0c26b26b0c1d00897334ab
145a4335cdcc64de09cf2932ca5ea7a2e75c3031
979 F20110331_AADIKR he_j_Page_308.txt
f68f99656f9bf0cfe7fcf5ccc5b4978e
84f60583019feee20cf812f97c72200ab3059a66
5613 F20110331_AADJHK he_j_Page_163thm.jpg
a8eed04b86ab1ff00af114e0a4707d39
f95f08e4b705b2a223ac74b0e35b865399cfa86c
18646 F20110331_AADKED he_j_Page_306.pro
fe5f4a1ae34de91631c5275777221c43
2691148960ed13c9dcbd88bdb3bfe1c570780bb1
2178 F20110331_AADJUV he_j_Page_239.txt
81d58e742dd2df41f11936ce90b09835
a5f58dd9e2cb3ce4cb83abf4cd8a376c46387de8
106342 F20110331_AADIKS he_j_Page_234.jpg
225e08d84d2f9ccb915edc95d02e2b92
fb1893674e00b7be4a8f607f88988d76b7002fac
35097 F20110331_AADJHL he_j_Page_287.pro
0755dd9a83f77e964c17e31a6f35d96d
9acc72c53694647794c0b804b86611fe95ff85d4
13832 F20110331_AADKEE he_j_Page_307.pro
95f9ddd0fd63254618f26816fd058ffd
cdbaf1a0bae4289d4f2ed8329d9bf3f8dba9c938
1957 F20110331_AADJUW he_j_Page_242.txt
1778c7efa578671713bf1554b6ad7d09
d3f509ce4b704dda73066520c7f1d51ae2b4f261
64901 F20110331_AADKRQ he_j_Page_264.jp2
1a9ba6a67261dca2682c87043a27681b
5ff65092049214aea3ae7852880eefe9c3376e46
99750 F20110331_AADIKT he_j_Page_095.jp2
79e1671d293664bca9b1dbb093b76822
accc7f10f0a45429f956eceeafec5c9fd57b9d1c
7042 F20110331_AADJHM he_j_Page_305thm.jpg
27c0481368f67da8a0c8b7f4635777fc
5430167fc643ee221749cee3859b60edc3038522
14256 F20110331_AADKEF he_j_Page_309.pro
58e842c9abb6e27df8a77165347c4815
e1953b30ecf68d90c985c7d29ecce3173f04c96f
1974 F20110331_AADJUX he_j_Page_243.txt
a39ff4da9f2943ae4ba0c016594cd0a3
9c6c6ac75580812c5b6365e024468475c1bd5431
67826 F20110331_AADKRR he_j_Page_265.jp2
c18357c01153ba1c39e5795ae95f465a
94a8a2033aac0e28fea09beccdb8c649f7c3082f
116592 F20110331_AADIKU he_j_Page_130.jp2
5dd8ddfa0628af7795065eee8eb26c79
8909b950a95fd79306aeb94103a36ab7163c92a5
622528 F20110331_AADJHN he_j_Page_204.jp2
e6576cf313ef0b69948e696a8c6b8acd
665e80896390650cba8e2bf9f88005a3c46b4b9b
14200 F20110331_AADKEG he_j_Page_311.pro
bc9b6a8da9906eddd0c87848116f9065
a9fe0b67b8554d3a3fbd749df34f317336c7864a
1049 F20110331_AADJUY he_j_Page_244.txt
1857d4de32cba06e068f08a1bb17a291
1d33e4e7334dabbcc9b3082d21c7a3f4bdcacba1
21808 F20110331_AADLBA he_j_Page_201.QC.jpg
6f2544c282133645ffc056f60170e00a
2dd46c1a936a6732e35e55d0b05a1078e662b5e3
58345 F20110331_AADKRS he_j_Page_266.jp2
c52154ecaa684513948a8162f7cd9ad6
42efb999c57e7b8e74e5f396aa12b3c833c3dad0
1067 F20110331_AADIKV he_j_Page_010.txt
1365bc2ed895703ebbe1b9c0e570eb0a
58dcebcc14c6ccc22ff7d32fb12bfd0978729715
8533 F20110331_AADJHO he_j_Page_104thm.jpg
cac31df17ab0f86498038840bd3778e9
c3c22e98aae16960033411c877b936b6875cff53
17425 F20110331_AADKEH he_j_Page_312.pro
c56e16b5a0e763f4cbdfbf733268b6e4
64099d2deefc5db51a2baf3af5615e127160c534
2040 F20110331_AADJUZ he_j_Page_245.txt
0117e5d99d9c22f58d07ef73731ed5d4
68c493a6d217280c5aea74359ca2be9c8ca65c29
24781 F20110331_AADLBB he_j_Page_203.QC.jpg
1a4b25432325c56b5f37e2ed76288fc5
8adfeb607557912700d4730f6889f22a1e0a8aba
68748 F20110331_AADKRT he_j_Page_267.jp2
40da6fd5a48cff182f0bbc635430a40d
bda55403a5bb7130f293c94f34c884ce3ec7f6fe
117484 F20110331_AADIKW he_j_Page_179.jp2
d1c352e3f2123a771fa41c54dcaa6467
250862493b2daa26911e163303d016dee7af225b
60019 F20110331_AADJHP he_j_Page_162.jpg
0d3ad8fb24653016f4271e95a5444db2
35b5ffce495dd4dbfd07c33e9311fc22556c7ed7
16136 F20110331_AADKEI he_j_Page_315.pro
dfab78e42aa77ca56d37ca87ed968c1d
a33a41d142924c197bc7ce3d190b6bf4939699bc
34012 F20110331_AADLBC he_j_Page_209.QC.jpg
fd8ecdcc223f29e69462e4e62b4e312c
5d7f73ea74c99084c46066b12b8d464ef532ce82
64414 F20110331_AADKRU he_j_Page_268.jp2
f0389d10fc43a1f78636f5b7c4ff3d9b
39df115175f5c3b510ee187ef17cae93d4552ce0
F20110331_AADIKX he_j_Page_216.tif
694fe899452d40dbab8f3ef1d34738d1
c42d557e89579065a65eeb13f2b49129e5e097e8
98473 F20110331_AADJHQ he_j_Page_243.jpg
362a7e2af2d6796b79b62ddf61b23a5f
5e7f40e645c8ef7066483d1f2149779731ec7c77
73021 F20110331_AADKEJ he_j_Page_317.pro
9707bfcedd7af72b6745e94e05c08e80
7623e1dfaee9bd1217accc4416c2f0c8161debc8
33120 F20110331_AADLBD he_j_Page_211.QC.jpg
69e0481d04ca5ff9c7901fe024a04ea8
74d062343b7a6967736a1a99b535e95ac281f6ce
63011 F20110331_AADKRV he_j_Page_269.jp2
1a519ad4412bcadbc3c1fd113a3e8829
1d9bf43163337e474551abbbfb239b69714eeedf
27957 F20110331_AADIKY he_j_Page_080.QC.jpg
065851fac93554a3a04d24bf6c363138
0e44c169c09258bb759f6320134e7e66d10d6ab5
F20110331_AADJHR he_j_Page_259.tif
aa1aba9df8fb1fa452ec43484ce9dcc4
d393ce22f746d7a36dfc789677c265c252f5e0c6
67892 F20110331_AADKEK he_j_Page_318.pro
29f5689143c059349531027b3d6c1488
ed2a0f297a579ca3c8d9a6dbb0e0624102e3939a
33910 F20110331_AADLBE he_j_Page_213.QC.jpg
55e25e057246576bb6ce67f6917d0f44
4439677889eb027a92a9cef68b0e3a6ffd8382a3
57626 F20110331_AADKRW he_j_Page_270.jp2
d8342a6d654bfa58f079dffd2b6ec77a
6ca23590094f43e08070326ee1f7d33226835fdc
1287 F20110331_AADIKZ he_j_Page_036.txt
c76c52d4e909fc0c60aa18afa8fede23
d4ff35100a7dd689f03403bc3ad9601a7d003b59
16703 F20110331_AADJHS he_j_Page_305.pro
6c2a6690a80ad58713206b01480afa71
77c0bc3d55cc7a6cf4305d6e36aa5128e55faa1e
64408 F20110331_AADKEL he_j_Page_319.pro
9afd06c0c2015e99a6f18f687f5071cd
8849892b75c645af6b1497a0820289bca92bf04b
32895 F20110331_AADLBF he_j_Page_214.QC.jpg
5f934fd0574268d55450de21e2b1ff44
f320f9b3ee9be5f1d5b36a30af63ac8f24ad7b4f
53823 F20110331_AADKRX he_j_Page_272.jp2
ce01dde9692b3358611a43e232995866
95b7055e0eb2d137b1a9d308ea9111e74968f3cb
2176 F20110331_AADJHT he_j_Page_241.txt
1e45143a89f48e21b02b6d142c1f8df8
eb218bc771a2c61a0539734fcbec8c2606c4b5ce
71126 F20110331_AADKEM he_j_Page_321.pro
c85cca9f4223f64ca3617ac67bb1a738
7992c28c889dce55a6f99765295f3cf82909bcb1
35204 F20110331_AADLBG he_j_Page_217.QC.jpg
0e97a64f11f5cf9fb0530eb249e81e49
48d4bfb904552cfc37da37985a7b4a697aceb4ff
59909 F20110331_AADKRY he_j_Page_274.jp2
88089d6d57ce6de9dc68dc28a4b2419c
b53bc6adf7d281178943f055acccb59483683f8d
8947 F20110331_AADJHU he_j_Page_239thm.jpg
5c91d02d42ea428720a1bcb1eec5dc4a
d68b1083d538738fdff7566dd07a29743f980c6d
68272 F20110331_AADKEN he_j_Page_323.pro
96cbb51fe99b12b3d6a417aa987b4e1f
4285a7dcbb2065b0751c14c09df7576409432985
31056 F20110331_AADLBH he_j_Page_221.QC.jpg
e9164ad5bcaf8c7bebb3d485448f7b45
3b516763f96b3c77cdca639a71fdc2112b760d0a
71653 F20110331_AADKRZ he_j_Page_275.jp2
5fd313cdd529b31fe6ab47f8bd71989d
9f5c02199c62294ce509d84ef50bc162592764e2
1360 F20110331_AADJHV he_j_Page_115.txt
df46fa96883e070b34bec3a12fd212ba
19ff85e8f7e3f91379b0f1cf0f834e53c6cfd69d
63645 F20110331_AADKEO he_j_Page_324.pro
c581f28f4cffc28e93802fe451da4425
3009f7625ea74eb0e7f42b213e9eb5c65093c5e2
8269 F20110331_AADLBI he_j_Page_226.QC.jpg
d9046a343ff68f4428edbca66aa9c2a5
1ec4e543cbadabac6014e8ba87bd032e5669c525
F20110331_AADJHW he_j_Page_074.tif
17e91c392aeff309ebebee383e9b7b97
b809ce11635d5b2392a5b01bee6bff45956c7a47
63856 F20110331_AADKEP he_j_Page_325.pro
1c4351b5632ebfb48fbef4c9e97eceb9
83ca0dcba74610261df91532a1390429b887ee09
F20110331_AADLBJ he_j_Page_227.QC.jpg
8e29667efd8e30b5c626ea15ed8cd444
7485ce327864a21a1a595516d8fd3fbcbb9dc618
499031 F20110331_AADJHX he_j_Page_115.jp2
efcf1ed48eaa6de0bc183f65f4060c31
637d70cc2e39fc826d93c09d5c097260e11fd57a
68827 F20110331_AADKEQ he_j_Page_326.pro
63eec6d3670b75dba4347f935404f875
0ede68167ebec244e9ec563a6e6c70d6e2bf05ff
15334 F20110331_AADLBK he_j_Page_228.QC.jpg
736511dbd998670a025db6e0006610e2
3ea4758c5167ce763f0ad2561570ea0a3790925e
10573 F20110331_AADJHY he_j_Page_061.QC.jpg
bf5c01e59acacd079732bd4228348fb8
5e6dec609e9ea57a1b14024547e1179bcdf3b342
72738 F20110331_AADKER he_j_Page_327.pro
9ee8d2a04b92e1580ac762db726f2d58
19b6fe900431c169dff79f5f06bd3cd27c00952b
24306 F20110331_AADLBL he_j_Page_231.QC.jpg
d23a63bf517681b382b56fc6e4db7788
820920fa4c022f6fd510c2e1f792c730772bec41
1980 F20110331_AADJHZ he_j_Page_182.txt
d75eb8d6bdea4ebf797b8c8878b6c1e0
91f38a5e8f96673764d169d53f581930ddd652a0
25111 F20110331_AADKES he_j_Page_328.pro
8d68939b80ed48bd99766f65f1414a03
f3f234ab94d532695eb78f6d94ca85a9e5b31eed
19432 F20110331_AADLBM he_j_Page_233.QC.jpg
b3793db4a0763bb4ecd2c0d5c495f9f7
8198c9c27c11a074e3c2b43a01b21ef5ff2490bc
52819 F20110331_AADKET he_j_Page_005.jpg
f227e8e6f89a08b64e0276cd878f9260
756052f1dd810cb110738d76abbe2698104c9c29
35119 F20110331_AADLBN he_j_Page_236.QC.jpg
a6f0fd45e62a6bab3b41a3823741655c
3e48164f7ba2f1eb453a26996b88aab372883d2a
134579 F20110331_AADKEU he_j_Page_006.jpg
dd9f40b1fd0cf9469f19bc5e667f91bf
3493105da99c7874ac616573b053f757b0992228
F20110331_AADIQA he_j_Page_246.tif
9a34ab7fb5a52d6c86697ef656e64ee1
5c38c6ba8d63d01f7bf0eadba43397aae4025413
34249 F20110331_AADLBO he_j_Page_237.QC.jpg
6c452cf773d4e35773dbed20ed93a084
f8612c6403f1149cbd7bf7cf3aadeb2fc13f96aa
160917 F20110331_AADKEV he_j_Page_009.jpg
895cb8f0285abccdc4ffcb452c8c2c33
9b0978ca14986b6b12e3d86a646db8332f8c7d3a
118258 F20110331_AADIQB he_j_Page_008.pro
8660beaa4a598804525d6595c324e2a4
01e445a61feff3ea0034213d24684dbd4e4a6b38
33057 F20110331_AADLBP he_j_Page_243.QC.jpg
ed5fb78c779db43526487748c493fdb0
bf552ca5ce2967b4c0f04267ae7c6fbfd9d1f461
58593 F20110331_AADKEW he_j_Page_010.jpg
a7666699f148821587b5e64db50074dc
5b405fe58a490ceb5084bdd39a8f3a6bfbeeb518
64230 F20110331_AADIQC he_j_Page_316.pro
1cdb43083ae027a7fc95fa5e8befeb76
8c2c9588d502983eeb526663574beef61f3e29bf
22010 F20110331_AADLBQ he_j_Page_246.QC.jpg
864943207264a3bcbf4a018133b26b6d
8b99f41d3d87288725b2ca20124291496ed5aaac
17747 F20110331_AADKXA he_j_Page_273.QC.jpg
9836670cca37b38402b5eafaff8c3f53
0000ae2dcd65081204024682d540b29103453bdc
132405 F20110331_AADKEX he_j_Page_011.jpg
24134e4a2ce2c424098cbed71cc3810e
948aa9c18ecb35e849a29916408b1ff4f7bb2c1d
29572 F20110331_AADIQD he_j_Page_153.QC.jpg
77acdd76a8311bb051bd0a128df630aa
a3cbcba87228f17d11c0122ca21555a9a579d995
23484 F20110331_AADLBR he_j_Page_247.QC.jpg
c30c6f4d7311cb8e39aba1f08d7f53e4
b1ffcf1e6172d265554222e81da6ebe61a00033b
8639 F20110331_AADKXB he_j_Page_046thm.jpg
15a363aa15bd3e1f1db6478f1b11d105
3fa335256e594461b0900c7788c8e2885ac00b36
138553 F20110331_AADKEY he_j_Page_012.jpg
32ea579edb92b828ec357cfd002be954
275a0262493f7627ef45226b4bc8c3b2f081f79a
104023 F20110331_AADIQE he_j_Page_040.jpg
70e99f66da83dcded48dda045712df09
9ee67cccc136bba97422992349e9ea6b9f773b17
17979 F20110331_AADLBS he_j_Page_249.QC.jpg
ef2be627663c3738f3208952800626d2
98735de3bbdf26d64befbea287f843b6e618cb3f
28172 F20110331_AADKXC he_j_Page_049.QC.jpg
99bf529c2269533890b31c153748dfb7
c5555db204a21e331e0d81f5995846d62fcd3444
127096 F20110331_AADKEZ he_j_Page_015.jpg
c1c3ef33c1f04f2babbc8618be54da50
71123c756ae3aa73967b14ee9f486196227519fd
F20110331_AADIQF he_j_Page_263.txt
3f4124b591cccbea53979aa2e0d891e4
ab84b3aac6c637f738b3d2bb1461822b282756b7
25743 F20110331_AADLBT he_j_Page_253.QC.jpg
4285bfdc444c5018dc85c32a56b1dd76
ed3a3dbf5a7b99b126aa88614e722236eb4ed055
2933 F20110331_AADKXD he_j_Page_230thm.jpg
2637fe308d41a7621de3c361a050f63a
168e12c5ec9c42401692b4285e4a10c554441a3d
8446 F20110331_AADIQG he_j_Page_293thm.jpg
afae9568d792733086794fcc279a29ee
5815873ea7b03f01c6170a56c6671fe60fc7ce5c
27849 F20110331_AADLBU he_j_Page_254.QC.jpg
51f3d60db1044d82a66524711f841715
1e8362a1f300cc96ab201e458a5893e0f4deef9b
9149 F20110331_AADKXE he_j_Page_146thm.jpg
781248d1f1f764928c53b850ca719a2f
9d03c168cdf0278c99d77df608c77c09ada1aba2
F20110331_AADJNA he_j_Page_233.tif
8fdb415b983977fe6a859c67da891fbb
841d96888d9b5aefeda78618e3f8b9bea2ded680
29985 F20110331_AADIQH he_j_Page_117.pro
c7b1776280dc3c271295dc113d537ef3
640db73eae8aafb1e775f3e8b193215d247f2f9d
18508 F20110331_AADLBV he_j_Page_257.QC.jpg
f3a89864422ce612a92cc841db9fd001
0378f300a30570863018d88207882527d5286209
7881 F20110331_AADKXF he_j_Page_035thm.jpg
8691a0bf69ad67c32203eeb4c4ed3cf0
4dbd02dd48231e4af74e042a816c8dfd05e30df8
F20110331_AADJNB he_j_Page_235.tif
28cf40cfcde320be6d4b5a7ce5395853
14b3efda6268e573f51b95a47832234c173f6aed
117645 F20110331_AADIQI he_j_Page_170.jp2
61e041961f2f3b34a7ea90210b344a0a
994a156cdc9a38c977b9d0c22fcbe886b92d5725
14829 F20110331_AADLBW he_j_Page_261.QC.jpg
2018b4bb1372e1d94955120d3476e4a9
7988503262c98985b20901d630575919dd1c22c2
20014 F20110331_AADKXG he_j_Page_200.QC.jpg
5a2cb5d75d430f733af47ae50c14748c
dcaa6f522be0dbcfe5f7335ccd2df10545529447
F20110331_AADJNC he_j_Page_236.tif
f88af7d26cbd8d597489c3e33e86235c
2da9627b2edbc1a38dd8aa9d23e5ef70a0e53844
35918 F20110331_AADIQJ he_j_Page_015.QC.jpg
50fbebec32ff4f401dc742384cf1bad6
8d0cc7296ff1e502f62311da4d7b0b0c0ebcd065
18012 F20110331_AADLBX he_j_Page_265.QC.jpg
8d1d3a178d4847db6ad88130b47a0246
51f8279d8b9a466982ee76eb578835327549debe
8781 F20110331_AADKXH he_j_Page_069thm.jpg
b7cbe48a47cba184da0ec7d593ea0a87
e5924777afefafb6d9bd0f592fbc923aba24f35f
F20110331_AADJND he_j_Page_237.tif
17fc8fba12d44c09f8f51e7ea2e94bea
11bc25d5615a980332b34177453318c6baa85279
91155 F20110331_AADIQK he_j_Page_086.jp2
307d4d4de2deb7b2e8ec0822b8d00917
005ec45ddb4cc8d7587474918dd936d760d5e98e
14808 F20110331_AADLBY he_j_Page_266.QC.jpg
3ae9ec6195d5f5bbfa1ef36ce1e59478
1b7d8449c4909d9041da587637c4708acc926e53
5766 F20110331_AADKXI he_j_Page_020thm.jpg
a4144920e6e9e76649a0044b60b92198
25e5b40ac331029a224df839a8b46ab7141bf19c
F20110331_AADJNE he_j_Page_240.tif
893100aa8913142307ea902f61818958
6ce3558db34a18011b091af4950000505a4b0150
981 F20110331_AADIDA he_j_Page_306.txt
0253903c901b3df56140990fe7d281ba
4b73ddab3ee84ba7f860142545fe22f364dd854e
F20110331_AADIQL he_j_Page_115.tif
4f1a9b13c8356e594a085d52a02a8708
ad84d30ff78c1b65768afe74b14406ece9c1d577
15660 F20110331_AADLBZ he_j_Page_268.QC.jpg
280492cb39533ed9047eaf6869d42eee
8465c3117b92349186d5fcb6cdc5b30db64fc078
34870 F20110331_AADKXJ he_j_Page_106.QC.jpg
b252546e05b67de9a2420c151bf3d208
07c2cb140acec603e74220f9e6860ae2e7d69180
F20110331_AADJNF he_j_Page_242.tif
a6dab613ac7fa7adb802494c205a9c5f
c8e2e83fd85bf6a1ed8a87c66b25e58cdade18b9
109377 F20110331_AADIDB he_j_Page_144.jp2
805dea4cbd972b9833fd92caeef5dc19
6b31996b91df527ba66016fa79b078faec87cc1c
44821 F20110331_AADIQM he_j_Page_174.pro
360a4ea32e826a3c378f7cf428901df3
1450ad03aed7673187c4de98bff7b391652ea7cf
31846 F20110331_AADKXK he_j_Page_142.QC.jpg
eb3f112218becdd59f10b278f9c8924d
08568bc343ec45585c757aab1bef4751fbeb4eb5
F20110331_AADJNG he_j_Page_244.tif
1a635e0a55020dda7e89a52ecb998b78
1aa2ace129ae32dbca66aec44d79b5b15815cfff
104946 F20110331_AADIDC he_j_Page_041.jp2
027f5675d6df38efdcc88f67d9e4d879
31c8ee0431184b06c6938e02d566e9b08a2000eb
35486 F20110331_AADIQN he_j_Page_131.QC.jpg
50635a1ad48a53e940428031beaa16ca
e5a514fe307f8a2343f965993fc15e3c55b39c2b
36407 F20110331_AADKXL he_j_Page_053.QC.jpg
9d4bf6b7deba85ed244cecbe27840761
4de606d120acc8ba79b0414c62b8b2d41db5f7f9
F20110331_AADJNH he_j_Page_245.tif
0980cfcc0774266d9a02e0b204b0e98b
fd79d3fcfeadc0b03189eeecd445eabab515c769
102012 F20110331_AADKKA he_j_Page_242.jpg
3b9787edd81d93a0a99e99a0c74defe7
d6f8444253df94d8d080cccb70c173c71b309804
2237 F20110331_AADIDD he_j_Page_252.txt
f8b7b991d9cfaea4dede9174cfc0d51d
63f2a511cb331c7a540758de223fccab57d1fd27
18714 F20110331_AADIQO he_j_Page_285.QC.jpg
ea7723df3891ad92862463f1d3e2bf68
2200b550b31e3a3687de4c7431655c0bf48445df
17036 F20110331_AADKXM he_j_Page_232.QC.jpg
7c4c56c55acc05c2027ba5e09119601b
b7af74d5946d541c0ed36cff3061aee24feea7df
F20110331_AADJNI he_j_Page_248.tif
e1517aa1df0bcaa17d171cd3b6f6da5d
071a49cfa062c544d9f6f76964d4167ceab7453b
78149 F20110331_AADKKB he_j_Page_245.jpg
778bb74c59a575cbd33d61b7b0592335
52cfbeaf1f41418dda7f9c64c0120eaef01b81ec
673 F20110331_AADIQP he_j_Page_111.txt
72954ef646fbde55ab042cdd29c2c985
73c913e92c6bd3ed5118dfe887ec4a975ef64878
34325 F20110331_AADKXN he_j_Page_096.QC.jpg
f60501b2d08870d4f5298ef38f19f848
6b8a4ebb1ffe0f937f7cfb0eba8727b535bdf150
F20110331_AADJNJ he_j_Page_250.tif
8578085d62165463e3c612256b0ef824
79eddd7ac2a163e1ed4b6aad263d9a23f7b08186
62894 F20110331_AADKKC he_j_Page_249.jpg
93973adbb990b376bc091d752476e0c7
1b28abe9f6731999a6809d75e333d6a018780780
116274 F20110331_AADIDE he_j_Page_058.jpg
0d679955533df432174b32958551f8a2
50ba53d224a068bb0499facbb54dae6d9da61d24
106785 F20110331_AADIQQ he_j_Page_131.jpg
fd8bf2e38499a223c2170954960d17d4
bb904f037a8102c5f0c6f3de323dfa19f82c6793
35206 F20110331_AADKXO he_j_Page_059.QC.jpg
a9b0279366bfcd7513f9bd88b5f0b05b
fbf5a98eb14272eb670a7f38a3e131226cd7fccf
F20110331_AADJNK he_j_Page_252.tif
0b6abb8a84c77373a51490d2ff170677
539a981c6970522ea1d8aa876c5ad95586d21e24
73753 F20110331_AADKKD he_j_Page_250.jpg
de6a4f55baf91997482f9da98907403e
7a2fca1b767f443a72e4f91d01cee82dfb7ea0f4
108044 F20110331_AADIDF he_j_Page_219.jpg
22bcea3bd2bd907fc8bcb4eea643f319
93d95c52fd00b2edf241fc3dd8233b5a361ac35a
31463 F20110331_AADKXP he_j_Page_050.QC.jpg
adfea80e99d76892f407807401dae0e8
db57c1f0aa7451a48d2dedca7decbf7c2a84780b
F20110331_AADJNL he_j_Page_253.tif
169c54f9357bcdafbdc8de6120b74f66
d7b3d3033b7b85fb05e51a3df1d5ab123a070167
56323 F20110331_AADKKE he_j_Page_251.jpg
ce54eae29321969b93fdd558d56869f8
2830a5460f7446fbc5fd7fc0af08366094651f6c
119916 F20110331_AADIDG he_j_Page_014.jpg
20db393265d85bc070767419a94ae9bc
e83b979236a8f5cb498fac4deae617dec68666f4
84 F20110331_AADIQR he_j_Page_002.txt
cd53594093a2eb9ecc3a72b3876e4c69
ceac012cfa4129082d49dc800413f7f0efc0c90f
7100 F20110331_AADKXQ he_j_Page_254thm.jpg
e1a48a1f06607b2463dd810d4d8f6769
d2c1d31b39608050f3ab95f72c2104dfa581f552
F20110331_AADJNM he_j_Page_254.tif
7fc6b6eb5f7093404253e3322512a8cf
ee92df3aefc604344bb2269bf1709112dd462b95
74491 F20110331_AADKKF he_j_Page_255.jpg
5f0555955d9dd914010675826200f68f
4beee608c478719943ba50e62641aaa6550b19cb
5553 F20110331_AADIDH he_j_Page_166thm.jpg
8a76ed79d42194ec9f094bf2f5acdbfe
c72515e90764becd2ae5c5b096279e77e501d724
59516 F20110331_AADJAA he_j_Page_271.jp2
5cff81dbea97e1d6a4a7841f466be9ef
44e39fd74eeaaf0a2423d33131e38e60f91c238a
36443 F20110331_AADIQS he_j_Page_046.QC.jpg
981194ddc0c15b2111bac770da9ff5d7
8bdd35c1caf00e17d96918b2ab538085cf75b93c
31626 F20110331_AADKXR he_j_Page_048.QC.jpg
237cb171951f9970187013c41a1fde36
71164c80f481771e1a74966690768101c9d7db6e
82043 F20110331_AADKKG he_j_Page_256.jpg
8a570b7e058b934cc18d44146fb35be0
0de04fd52c588d01a54f21176a54a04408d6c249
F20110331_AADIDI he_j_Page_123.jp2
59fe87147d41520b60f4d484fb157b4b
f09f6bb46714889d3918a23fc2bec364c37c2f05
599066 F20110331_AADJAB he_j_Page_313.jp2
6d2f81800c05f7b148dbd42287c343ee
33e2abc033f4cceb427f7164c88661ea8ba9437d
20393 F20110331_AADIQT he_j_Page_160.pro
958717596981819a00e3082ca0853993
746068fcb27eaf68644bf42a4c34d11d2db8a227
5790 F20110331_AADLHA he_j_Page_250thm.jpg
a11348ca20d8b52a2abb478666d1f639
69055fd9e85b95f9a64f2c5b7a21a33f4481a65b
18871 F20110331_AADKXS he_j_Page_312.QC.jpg
06e8350c8fbb8763bba2996aa4f33e6b
dbd514f870c6edc989f2c6a4f73f61002f8add12
F20110331_AADJNN he_j_Page_255.tif
ae47e5abd0006aa3d73e92a27ce077e1
3558e0f0452776f427d32209baa728b51fd04953
73486 F20110331_AADKKH he_j_Page_259.jpg
000dea205d997936c03131e759fc0211
ba92259b7e06d4b0f51318465f9c34024d55e0f0
103697 F20110331_AADIDJ he_j_Page_152.jpg
71c4f24c70041823950643edb2f37d91
5d2aa68ec6607299f8dbbde835a1ef6f3d67fcd9
11185 F20110331_AADJAC he_j_Page_127.QC.jpg
1a3093b35ea224b31b38de608cf99eed
223dd8dd5c8615eb3745f83f7093746e583e36ff
F20110331_AADIQU he_j_Page_329.tif
b0bd22a84229de7847a1c093486d9e1f
e73fb23505f8e90e904eaee2fe7aba7f354a715f
6220 F20110331_AADLHB he_j_Page_253thm.jpg
bb8cc8753ed66e3e2411417fb03a0c27
a950487bfa9569d4c61ba975c41709226da7b6d6
7739 F20110331_AADKXT he_j_Page_199.QC.jpg
5ef667cf13ae7b650d557b6df055eb93
85be53f5de65144876da4b8c408841138953aae5
F20110331_AADJNO he_j_Page_260.tif
b05843fbcb81a6223ca327e0b442ea7d
339e9583d009e6d195834ec17bc91b08cdeda1b7
74579 F20110331_AADKKI he_j_Page_263.jpg
8a125a13c5f6a35874ad858188eec0ca
cdd19b14a75046034dc0b0ab7ee3afe976c203cc
112660 F20110331_AADIDK he_j_Page_237.jpg
5aa2f825d61b7588c4f046c56cb400fe
eb5df63c811aeb722a828ce985774c403f7f8373
4089 F20110331_AADJAD he_j_Page_269thm.jpg
232b218e4a74b02c44b903beab556b83
9e549dff4e6c95299cdc5cf69f7da4cd694e394d
2613 F20110331_AADIQV he_j_Page_325.txt
5ed740329abc063e567aa9922f814c3e
c7be5f1b225d846bdacc66c28576a6e98e07f589
6339 F20110331_AADLHC he_j_Page_260thm.jpg
56750f0e0b78243c1d95025ca6aa2109
9bfd70b4a186bdfe747ed6575f622be228b590ba
8918 F20110331_AADKXU he_j_Page_325thm.jpg
e41c943cc86b9a812dde1cb92ebd1970
bd2d5e61c36e40aa4743eba23781fe36804f8174
F20110331_AADJNP he_j_Page_262.tif
9d74e5b19da49fde890a52f73f825bdd
e11a341362754d4a954b35d151960629f07399bf
1054428 F20110331_AADIDL he_j_Page_206.tif
5bdb8f59cb27df886e57faddadd245e4
3807a83ba3cb1bf49140bc30cb10841ed4400c43
152 F20110331_AADJAE he_j_Page_003.txt
f43e9e8f48e73752389b0ec579755cce
a2eba47c583c74a616707c97fb28c1b85bc45654
111244 F20110331_AADIQW he_j_Page_149.jpg
f16ffe3c45d1a5f8165e4982caf17c42
bbc2da5ebc119110bfd369776692c1e50e4b1016
6193 F20110331_AADLHD he_j_Page_263thm.jpg
4c220c7d4f4cfb6f655534069c2677d8
0eebf09bfbeb60944307dbc19768666718224148
8524 F20110331_AADKXV he_j_Page_070thm.jpg
e8ab372aeab0b8ae7b27fb05fae8662b
3605b1f1da670154dd0561769ee2b9069c7fcb14
F20110331_AADJNQ he_j_Page_263.tif
6586fb3df2fede9582a92706ade56ae4
cbad9902b995d48d3a290c0508a9d7726129ca5a
50795 F20110331_AADKKJ he_j_Page_266.jpg
0c9fa8250e4ae42f5242026d7a95668a
b75557702624e8cbb2fb604744692e2c769ff6dd
4776 F20110331_AADIDM he_j_Page_264thm.jpg
486d277df4123eea31847849bc28507c
cb808373d9a0dfde717da6643c9d5d039e304370
57438 F20110331_AADJAF he_j_Page_261.jp2
726bb136a1137ceb601caa910e96579f
aaadec9f57ff0a3606f0066d7271e25a4ee33cbd
30714 F20110331_AADIQX he_j_Page_035.QC.jpg
7f77a06f28327b92b9339cf615daaa65
d7d9088158c98f099e398061d36c95152e5f25f1
3872 F20110331_AADLHE he_j_Page_272thm.jpg
68781651bb6fd4511b5b1920677244a8
877ccb7baaca70b5b054ebbf0fda461094bbe2ab
F20110331_AADJNR he_j_Page_264.tif
54adcfaaf11762e10ab26330b93aa50c
3872384cd873dd904e9a6619b00e110185dc949f
61273 F20110331_AADKKK he_j_Page_267.jpg
e31501d3c52d471048f1a41232d2a016
c7795e761e38e7575464831f26bac610f3a5ef33
25723 F20110331_AADIDN he_j_Page_065.jpg
05c381ea4cdd68bc9be086991879a10b
0da56a0250e113d5eba974a4c5b136ce205206c8
36741 F20110331_AADJAG he_j_Page_220.QC.jpg
329f2377ca820a02a9e79a179d4a8b1f
9590cef9bdc8460a244e3cc3b64dd48c3603a4dc
33269 F20110331_AADIQY he_j_Page_240.QC.jpg
4af838fc423ce4a11761f724892b4b34
5148cab4ff872bf5fc0c274a89d5fc1aad329643
5194 F20110331_AADKXW he_j_Page_109thm.jpg
43bbb57e28b6c7f889e2be4fb92896be
2ae861588f87c7bc4c8d64faf65e1f6ef94762e1
F20110331_AADJNS he_j_Page_265.tif
77133fa0a413efe2c891e22f8f1ce5a9
69ed6f213e501079105af50246181643a71c7146
55422 F20110331_AADKKL he_j_Page_269.jpg
10f2fce90169f1fcffb7cf659b21a487
10fcfddf33006f2e2f242bd7581ecdca3354de48
5564 F20110331_AADIDO he_j_Page_256thm.jpg
872d79df651b44d284fb60672ad5db1b
5a8c1370f8a89203531beebcf3d70dba48e3fdd5
54343 F20110331_AADJAH he_j_Page_073.pro
bc3d3e741fe1190328db4806f10ef3ea
8705d239c2d91dd7be9619f89f818b74e2419083
107273 F20110331_AADIQZ he_j_Page_023.jpg
b5c18ed0256d229f0e0cdb5da51ed39f
d936cc53e6530a9bc707d2159571e10a1b4c50a7
3766 F20110331_AADLHF he_j_Page_274thm.jpg
86f2a018096241950978c441e16d49a1
d70c10353d7fee43a0c66d6b5262c01f2453d7aa
5010 F20110331_AADKXX he_j_Page_288thm.jpg
969ab12c70df173125ce041819bd7c7b
5dfd518b5e108379107cf102ce3507248dfff102
F20110331_AADJNT he_j_Page_266.tif
c1d0fd0503bb64fa55704daa5356aede
d28120e116f0a63f8f5b4bc4c079771780e1a317
53795 F20110331_AADKKM he_j_Page_271.jpg
ef7dff52965227d1987500c47dee0208
83ee15ad378bae7409a979b008f17591932b7cae
98652 F20110331_AADIDP he_j_Page_048.jpg
a9056c2f4fbabae2cd684e2b64cb5938
8d4b90c8d938b7bdfe2bc47586657eacda643318
F20110331_AADJAI he_j_Page_223.pro
7493c90e84b87320f3490d6395fdcd37
13011731300968b09682b5fbde42eeaeadb56d50
4463 F20110331_AADLHG he_j_Page_276thm.jpg
a1203b9743b041e1ea4d98ff66dc50b4
1d771d68bd627cd6d8ba332099eb709afa85a33a
37255 F20110331_AADKXY he_j_Page_129.QC.jpg
fa16cfd9fa69901d8d03e09199113b6a
15826928eebf90639584e474151a9fa8ea505814
F20110331_AADJNU he_j_Page_267.tif
63000b23efcdce7b64b624777324beee
57255e0d76b7602a7f6f101be61400971dae08f5
50017 F20110331_AADKKN he_j_Page_272.jpg
bf4c04cbdc1e107f7a637a1b052bd133
3b9c9bb2c6d30e4cc513c1281d5d5ab37e69a2a0
140561 F20110331_AADHZA he_j_Page_323.jp2
f117b8f920ebef222814c5c9fbeda512
b962699f556eb32139a017d31036d22be5e8105f
8986 F20110331_AADIDQ he_j_Page_053thm.jpg
a163f62c72aa920bca8b17979be17d7d
db9eeeaa0c14803d22124a6ab7d33e4260ca5e54
352 F20110331_AADJAJ he_j_Page_298.txt
24f4608144ce3f06a50fbb771f1af1c0
6092aa4a5de1e780b333281874e7ab752a048629
4675 F20110331_AADLHH he_j_Page_277thm.jpg
d51a997364fd01ca23b4ecbcd0e7ffa3
f67a9f708137303ea589466321a7dc2ccc1c646d
9075 F20110331_AADKXZ he_j_Page_016thm.jpg
395517def5ff447fde4fa2f3d0ab5d89
bf3eb16b2f4fbaf06ecd74901b72faaa77433727
54436 F20110331_AADKKO he_j_Page_274.jpg
3f64836792ecf09badcca3dca3cdd400
97b27e6f8b864f0b7e53a2563b6513359d7d728b
F20110331_AADJNV he_j_Page_268.tif
d9719d430b33fbca9a23cbd86916da75
da6f14dff355a29e3891b3c37faf053153be3bd1
110542 F20110331_AADHZB he_j_Page_033.jpg
59a3fd19fe61a0b9754c0656723ef0af
7b24ae57caea55684e0ac41c6711dbb463becd55
F20110331_AADIDR he_j_Page_030.tif
b60cf4ced0b7391cea43412965619994
57e9d2a0291df0d1aaeab566c92f8638058617f4
60449 F20110331_AADJAK he_j_Page_160.jpg
a08ab3e00b6005069c189158875fba28
6bee6c54661ad76bbae4b42bbe82d6997128aa91
4357 F20110331_AADLHI he_j_Page_278thm.jpg
47ffadf9bf6561ecdb514fad6ebb09e5
6606421a15bc9d29d02788d81980990c13cccf88
61326 F20110331_AADKKP he_j_Page_275.jpg
a57378ee39963f492e60b9fc9aad2cbe
5c1cacd8d584306bdc8a344207dba3314bf42302
F20110331_AADJNW he_j_Page_269.tif
f2bfa784f882537f3511d76b0f85b3c5
725fff6c4a7b7d8898fba5e8327eae6ffd2f9790
13930 F20110331_AADHZC he_j_Page_066.QC.jpg
1bff53a22a8c881464a349589c289a5f
df08676067bd19352822cd0e1f53798b6ee9ab45
1334 F20110331_AADIDS he_j_Page_002.QC.jpg
a93fdc371953eba867adc13fc4199cac
8c4ef9dbcff2c28dbba8b3bfcb95a5fdd1221c45
51409 F20110331_AADJAL he_j_Page_152.pro
a947018bb51e270b0edc70fba233a708
6719bf16e738589e22b82f6732e1efa4ab90890d
4558 F20110331_AADLHJ he_j_Page_279thm.jpg
2a96d923a108ef99031a236a261c107d
667bf053ab78ca37ee2e0eef52748693edc3d8d0
58607 F20110331_AADKKQ he_j_Page_276.jpg
faaea2de0130ca64558e24b684e44166
1cc643de8244aba1527115341f26c563e75bc3a2
F20110331_AADJNX he_j_Page_273.tif
b6a4f9c76c42de656045121ac11d3dca
bf070a5da682a60785dd45790bbd942115fcc967
46063 F20110331_AADHZD he_j_Page_198.jpg
55a30715396ca3a27e434c2403fe1dcd
e626454a73344b223c15a94ff11e4e9981e23040
1051932 F20110331_AADIDT he_j_Page_294.jp2
01afd2566db46ac1640eadfb712a4ffe
798cd565bbc550c6045f7549b26fb88f22336b6d
F20110331_AADJAM he_j_Page_243.tif
d21ca66d082374ee252ea157387da57c
0f84c7f0f2fe602e5c00d1b79031e3822170c83f
4282 F20110331_AADLHK he_j_Page_280thm.jpg
9244e5fa837c3447d288ac96aa76fe81
71152fb29c0c95d789a75e01da432bd8f8ff82f0
65191 F20110331_AADKKR he_j_Page_277.jpg
53d3f6c73a3067892b1db4d9a55f1932
a4ca2601f730ee3b72bd829be924c050880619b5
F20110331_AADJNY he_j_Page_274.tif
b525b6f11b593c4badb4db232603b89f
5f111265e2712ca3981b1e028431aac1caf6b30e
73644 F20110331_AADHZE he_j_Page_246.jpg
17e09279b197009ffd2fd83064f3426a
7cd17297da0dc692c6d3f9a2867519c2a2ae2f46
34256 F20110331_AADIDU he_j_Page_182.QC.jpg
8bb1e73a49a850302d32b8e01b9ed678
152ae0abf6ae6215f675123844e47a11f5a4f8b4
57571 F20110331_AADJAN he_j_Page_315.jpg
661fcc55d6d026064d3225907fd6ebf5
ee74aaa12535eee0c3749056f6d1194d6b1a1759
4770 F20110331_AADLHL he_j_Page_282thm.jpg
471536d40017c04139e6c43c2165e7df
c47d35acb6db1e447de4cc3189a93293c77890ae
66334 F20110331_AADKKS he_j_Page_279.jpg
82f3ead180be11609d19a5644ca9b84e
3bc60bad4e6188a79f11d475dca5b314258a5da6
F20110331_AADJNZ he_j_Page_275.tif
0967a99c12ca95ff866dfeb624bd948f
e68f695632abd8320ab956021d8efe04b7c97db9
7516 F20110331_AADHZF he_j_Page_122thm.jpg
f123a02950b2c4875ede5e6750d62342
58fa90fd8dc056cd94bf7e30476a1cd71c47e211
104373 F20110331_AADIDV he_j_Page_254.jpg
8e348a91fd1e0e8e80e93a4355f262d1
e97cb58bf06895f691cb6a351a8c8729d8991d30
116833 F20110331_AADJAO he_j_Page_254.jp2
1649cf320e262264df1bf5084d69d16d
f5dd02ae355011e40d73915799597ada5b113719
4434 F20110331_AADLHM he_j_Page_283thm.jpg
90000f2da0da1fb679d00495d108b6fb
8250bb82d9d25581da3e41e6fdd49ac5b1a459e7
57966 F20110331_AADKKT he_j_Page_280.jpg
ab784534b0a4aa8f2ca21ec7d9cf9064
7c7770b9ed48d29904737a59cf98831879c0d058
31747 F20110331_AADHZG he_j_Page_056.QC.jpg
8a18be5339e8a8f06e38a73ac2356cf3
4dc62be9555ebe5162225e84b64056490c4b87e6
18522 F20110331_AADIDW he_j_Page_264.QC.jpg
2a57df7a39a5979cfbf65c1b701a03c0
0ef4bef246ac3798421821d978d6750c0caa5870
592756 F20110331_AADJAP he_j_Page_119.jp2
f57f0d19bd98f74cffbd9b95710c687e
1ac416264a75c41cb52fac60e22e5033fc84353e
4500 F20110331_AADLHN he_j_Page_284thm.jpg
995e48bc5e3a65eb454b99b12123964b
b443ff9fa3f928cc9b6fa428af1946a967dd7013
63894 F20110331_AADKKU he_j_Page_282.jpg
2a486a71234d5140d0a22af35f68d20e
6ba319c1d138f3f27efb8d018c3474c4fd110ad0
F20110331_AADHZH he_j_Page_062.tif
1b3392a133a82e911f3306fdd3232381
e624b74d5caaa6f340cdbbcd8e811f372ff83af9
74007 F20110331_AADIDX he_j_Page_279.jp2
2fd4e559242e040737be8c0e0042c8c4
842e940f171cab66ba37ffd7f1597f6a2bfbd4f6
F20110331_AADIWA he_j_Page_322.tif
d77f5e4f71dee53a11a15a1288af12f1
3736fdbd915ded5d66b6ae0684d779ff5c6591c9
105232 F20110331_AADJAQ he_j_Page_143.jp2
68a4b8cad8e427470bb3d44587d45292
7d331f24a23e719d562de246a25cedb05592a010
F20110331_AADLHO he_j_Page_286thm.jpg
3e5621758d14a172aeea4e7f95eb7a62
b0c1c546a0bf94460f8d33aae60f96347c61c956
63596 F20110331_AADKKV he_j_Page_284.jpg
4d15e2a33e2ec1b7008dcef1f9374a0a
6b12ee191e72d416242b65747d01be05d7ea8a7b
9084 F20110331_AADHZI he_j_Page_033thm.jpg
6115978eae6f3e33cbc0145be1b7f500
a41c80441b155d85914505060f81496c32fd0a45
53659 F20110331_AADIDY he_j_Page_074.pro
8662fee75fd939e5d7e5023346b26d2e
5b1a2488d1e9c20a66493f86289d8c100b3bf5cf
119813 F20110331_AADIWB he_j_Page_024.jp2
8475690213ba892d3b40023bfebf35de
2162364f4808f9bc0840816f8e1fca786baf0e82
F20110331_AADJAR he_j_Page_021.tif
67f2a6e3f9508ac9456e0ab0e4f1c720
8d77d2af01aa468c0896d306766fb6f9999a931c
4626 F20110331_AADLHP he_j_Page_289thm.jpg
f13b20cb9e23582d34e476387b27a18f
313dbb027f2e494be76250564693f84c280c8791
63408 F20110331_AADKKW he_j_Page_285.jpg
e29c47f56e1a5f5e26819a0d37e50de5
ab20ca5d91bd72bad46deb5d4c54e3e8c1236cc2
37823 F20110331_AADHZJ he_j_Page_238.QC.jpg
45280ccc5f10bc2049a69f57948b26a3
b7b366e06205b6b702daa5c2d6ee5ce65f205dd8
1444 F20110331_AADIDZ he_j_Page_282.txt
d84d2d88f1994e575184289942146244
728d58a2ef46b4db1f87b7e3709fdf2108cbd27a
8987 F20110331_AADIWC he_j_Page_238thm.jpg
c1de5e7adb58ebc3fa172671bc16cb4d
429c83c5c76e9633518504f361f82252f708ce0b
72289 F20110331_AADJAS he_j_Page_154.jpg
dad0aed4c1c8e51b4a19cdab9a3cc6f9
7561cc79fff66311663e64808aed2958cddcdc56
2175 F20110331_AADLHQ he_j_Page_292thm.jpg
9aafba317c6231cb63ca407830f18272
daca7e6e412baa5e3a13d0a92ca3c360e07d9803
90795 F20110331_AADKKX he_j_Page_293.jpg
b6594bb06d26f593909af31520f1180d
313259092908c6d184e20ccda8d40a43e8d8a9a3
51187 F20110331_AADHZK he_j_Page_190.pro
c36819a4b8431f1f54b7cac6b409ae8b
fe6d0d46b89ddc101cba3ec2ba977f2b4c7002c0
97570 F20110331_AADIWD he_j_Page_041.jpg
e7c982b7ccd79dac09ac3a981f28586f
d251dfb1fd8b83eadeaadc2c78b6862a9e9ca48a
8709 F20110331_AADJAT he_j_Page_094thm.jpg
8de2b518a29fd151bcfbf4c0c2e7d823
3e1b264a96cbfba1a9b2091623eab8e6a26af44d
10486 F20110331_AADLHR he_j_Page_294thm.jpg
d7fae3841c3e253737cdeba2a52bac0f
bb81c9129e2e44a88586ee3c4215511bbb815a62
118988 F20110331_AADKKY he_j_Page_296.jpg
6d3b819f202a2067b6a4d31c97c5d9b0
e02a6477da02c9aa6ffa5503d7cd01e3e3389ded
154539 F20110331_AADHZL he_j_Page_320.jpg
3838d6f84047e146e32d2128968a586c
ac3c3ceee5a1041061a2f6295b3e457419e15e7f
54310 F20110331_AADIWE he_j_Page_179.pro
fe1e54068632b2880fe5d857aaedd6f6
dda8fb94a98018857bbe83bbfb86b9b1c8960e13
F20110331_AADJAU he_j_Page_035.tif
bc6ebd3e90014856cc1384f651cba7a2
d834dfd839999e7c4794dc0a9e611d5966db369a
8856 F20110331_AADLHS he_j_Page_295thm.jpg
7d5dae4496e1742f93de16380628d38d
631a729f6aa9aee07fe4a42b317a7a0bcde84bb3
139837 F20110331_AADKKZ he_j_Page_298.jpg
16c50e42c7264657534acbf31ce0bcba
f0a15511cdb4f6074e203107b37272772bc113db
810 F20110331_AADHZM he_j_Page_197.txt
57f64091a826f7f2bdaf3e4a5fcc4be2
557eb730d75d3bbc83d912a0497f4365b15358b9
648 F20110331_AADIWF he_j_Page_161.txt
45e4cd16f0157644c6e486514172af9c
b6423c56941cd67db2578bc12cb0f1c58fb4b1cf
108722 F20110331_AADJAV he_j_Page_046.jpg
7bf4a35e24cea7ed6bf70b26a95e82f8
de7ad45a7e0afaa8a04732ea09bccac81be517ed
8795 F20110331_AADLHT he_j_Page_301thm.jpg
029e10f4cb129bd6a86d59b1ed6dde79
af66d2aafce70a6c401167c24b74ea589d86a664
22877 F20110331_AADHZN he_j_Page_263.QC.jpg
c1671a2cdf9eac211faaa926e8b8a6f7
2609c313e7a8a9e4dc7f605839e25bc0438e81d2
1991 F20110331_AADIWG he_j_Page_065thm.jpg
f54d461327a923ffe03e5e1cd4ed798a
1b43b32db1de7562ec48592b835cb159f7666b81
101080 F20110331_AADJAW he_j_Page_004.jpg
91ceaefa5c0084fa3edbd9a8de07f816
815b45733d2aea02d3dd12de1d863ec5d49958ae
4168 F20110331_AADLHU he_j_Page_303thm.jpg
aa7f85ae8a811e31180ed875e3da161f
137ff0437194ead12e87f8a87f7876b069a2a65a
1454 F20110331_AADHZO he_j_Page_285.txt
8db5263188ed40ef6fbdcf4520683521
8238488ef9cb42ed110b700556bec4023ef6b6d6
F20110331_AADIWH he_j_Page_052thm.jpg
81a9afddfe8815b94409130d6a75c710
177c3e46582dee53d24e2d926d145c4517e01c3a
15459 F20110331_AADJAX he_j_Page_229.QC.jpg
3ffeec8f59cac07f8cdf753972b69f4e
b997a0ce8d6338b4e1fa08b5c3132535156cf127
3780 F20110331_AADJTA he_j_Page_165.txt
ac017def2da97d68d2f77f7af49b83a2
f4082b2804c7527f8189b481e23b206b267cdb39
6160 F20110331_AADLHV he_j_Page_308thm.jpg
24f06cbdf85fc02f6d9725f79a723867
d9c9d8db946729d50d0f570781238c8ffbede2a0
18108 F20110331_AADHZP he_j_Page_287.QC.jpg
39add8a80429e02ccaf6ff8c336a8ed5
67533b92941bdd80b8a50ff1579c5fa940cad6f6
7905 F20110331_AADIWI he_j_Page_221thm.jpg
a11b9e4e24b8bd332ba202990831bfe4
8ad5adbcc531048b1ba241c242e212445eb2fb81
2868 F20110331_AADJAY he_j_Page_204.txt
47fcbc36454e7d67d277b7ce01ae21bd
2abeb40598e86c835b46d8e4b4a4ac23e84dd23b
3107 F20110331_AADJTB he_j_Page_167.txt
e3f9fd739e318ca7f044270a1808827c
d3bcfeb8d006b99b5c93f8e3a241ccbe427889ba
5460 F20110331_AADLHW he_j_Page_310thm.jpg
36866795f79f62fa0628645b57f4685c
e00a8797239a0255d0c191e8e2ffa65d2749a214
59125 F20110331_AADHZQ he_j_Page_286.jpg
bbf6a846bd8aefdf3663808cf5cc9efe
41755ca7e7337a9f98fad1d2c8b73e652ea42b82
778 F20110331_AADIWJ he_j_Page_002.pro
b71fb043d414aa86480bf98ab4758971
297644ad82c697c35135016674881a1b4ebe1687
56748 F20110331_AADJAZ he_j_Page_097.pro
4e9fe687538b191b9f1118c7f66100dd
0f40a09e1f370987369a1a4e6b01cfeb453eaaf1
2138 F20110331_AADJTC he_j_Page_168.txt
dde1c2874a186dc65cf457a0d34ef4d2
a335c3452f24d539845b0f356accfebb691e5e6b
5390 F20110331_AADLHX he_j_Page_311thm.jpg
84c29fea394086f23ada2d311044a320
4ab8503abd0ac743e95ad3f35d8ff83e464cf1f9
55168 F20110331_AADHZR he_j_Page_132.pro
c68e740e6b22a8c99915825cbf9dd410
1c0aeaecdb7f9bd0de2929e581da8eb17e8f5871
6382 F20110331_AADIWK he_j_Page_116thm.jpg
6a5dd817d0a137ea1ce4848790c3ee64
e262b1448903db211333787a9d4214aa4e2b8234
F20110331_AADJTD he_j_Page_169.txt
4b1cf244afdf34bebb9b8b90c638bc65
70436e3d0166e61aa45b2dd477f42891a6053766
5280 F20110331_AADLHY he_j_Page_312thm.jpg
40d44bb277ee1c07791dd2b3c58ff583
52a060d393cd31ef368743c45538c660632578da
F20110331_AADIJA he_j_Page_288.tif
388f5f0c4c6f2aee44eb27a5c4c9620b
2908db2248ac83a5aff1fe81827ab8c55854441a
32145 F20110331_AADHZS he_j_Page_078.QC.jpg
b69796431dcc775cfb3c351b1a0c462c
c16671ed48f33c224b50cba946e60b9b3df6bd8c
100944 F20110331_AADIWL he_j_Page_181.jpg
a731ac0e216fbcc69aff40fc0756b408
10e38087896b72902fc135af1fa6bf25dc710e94
2134 F20110331_AADJTE he_j_Page_170.txt
9fd608749a402bea72f866c2123fada5
6949ccc624671393977ec4dd350111047d8c9537
2760 F20110331_AADLHZ he_j_Page_329thm.jpg
6ae5a2389d8ec7e6e02a516fbbf65fd3
906720864b845a1da51d0483f3f8987d56949ee8
F20110331_AADIJB he_j_Page_104.tif
71a54b05dfabdeb004639446a37a096d
91543d3a67f0a720c37b308ad812319a20eb9f6f
61410 F20110331_AADHZT he_j_Page_227.jpg
056afedec2d107d65d8da1c9ca6daa69
ff1fefae9cb2dacaea18842f595cb1ea449828b8
F20110331_AADIWM he_j_Page_053.tif
4ac10070fc2d374b5f4da2591fd1f13c
2cf01876de243688188fa961a2962061c1d0773f
1919 F20110331_AADJTF he_j_Page_172.txt
66c0a727c7ebf8763fd382ce8564b3de
e7fe4995b6dd5dd586a05e93351c4a808da052ea
55505 F20110331_AADIJC he_j_Page_133.pro
cddf506167fdbf05695960a740920d5c
3941f917f7f27c6b3d7e7343587e962142a36fd9
F20110331_AADHZU he_j_Page_065.tif
bb245c64843bdd3c87cc8036254c6c57
9de5471a7cdec5e2c776f174b0412a3e8e917fc4
52730 F20110331_AADIWN he_j_Page_193.pro
37051c6ac395b3c66b9fe9938ee03363
d6c06a5f10d1e415c4e3a833b36414eeb0727787
1768 F20110331_AADJTG he_j_Page_173.txt
5826fdb8e847d9baa139ad0ae85db5a0
c1a62b949d3415772d4ebac6f0e8a0e8236b0255
469984 F20110331_AADKQA he_j_Page_198.jp2
160880d29e4768fc2429cf6d113cd613
2e8cf20c76adbc9cdc4b861960a9560030e68bc3
1841 F20110331_AADIJD he_j_Page_164.txt
5ae25f1497ad9de61edec03c434866d5
516c2a0dd185d37e5c5fdeadfbc56c9968960cc4
70943 F20110331_AADHZV he_j_Page_015.pro
90c996e7992a20b7d9fc78409847df3d
4402cce469c458866bef7bd01f8afc1fdc7d4cc0
2834 F20110331_AADIWO he_j_Page_295.pro
fd302de9eb523d8011e3f7a51306b975
baed8b361e3d35abaab0d8dbd8deb29b2cba595f
1940 F20110331_AADJTH he_j_Page_174.txt
de2265b3ac30c7eed0578f7c7457af44
c3928f6387c504ec96279c533a0853c5045c8a1a
218469 F20110331_AADKQB he_j_Page_199.jp2
6874c4dbb710030df617bf1cedffaf77
4cc3a674024119184ae480f0e517b86e206c86e5
1051959 F20110331_AADIJE he_j_Page_008.jp2
c43aa672ef7a3bcd605600729df000e8
7829454a3ca398e33470f5c02e67454186ad5468
104586 F20110331_AADHZW he_j_Page_247.jp2
f040aa32e46efb769790ce0b8ba5631e
41e3e1e2345cf3dc0cc1e5c03f002f31ef294c0b
36146 F20110331_AADIWP he_j_Page_094.QC.jpg
f6c2bf4c057df5330e9bcf1be4457e81
fa738698b1d6aff60cff04a267cdbb634dc333f2
2230 F20110331_AADJTI he_j_Page_175.txt
36ca8c046f4ab1240ae2e1b01e6574ee
04439fbba97784a9d430a27c335ec4f7241f555a
65296 F20110331_AADKQC he_j_Page_200.jp2
5a7e2ccd2746faf3317d7b24cc08e4d7
afddd1b49b9917ea966671b89110ecddd13f7b7b
2014 F20110331_AADIJF he_j_Page_059.txt
49c915e96b5f9abfbd9417263599d048
52aa37552cbc55a7e581051fee34cc3770fad054
90384 F20110331_AADHZX he_j_Page_247.jpg
913afacddd73b538c1ccfe0d840a422b
fc18cd13682e268ca6d57173e509db138e78efb1
54037 F20110331_AADIWQ he_j_Page_147.pro
819affef01f67709b471ec03c2266a1a
2d68b66d7773d1baf4b1d73045aa2949d878d6b8
2206 F20110331_AADJTJ he_j_Page_176.txt
6715a9546ae52123bd3726508af83012
0f30cfb4d55118ea4760c707049148f98094307b
741700 F20110331_AADKQD he_j_Page_201.jp2
dad287439b4e22b884f1976cd5af22f1
45e39e818cb4d5c785a92ad31b62d57963e10daa
8643 F20110331_AADIJG he_j_Page_300thm.jpg
32e76b55904702222a32d3ba69d468b5
66fd987144a1f96da62d82a9a2128086604286fa
89872 F20110331_AADHZY he_j_Page_088.jpg
cee9b5982d2e4d69ff18bb1faf12b5aa
5c0ba505f2dbae58569a4a8d0744ce9e61d52e72
48854 F20110331_AADIWR he_j_Page_248.pro
73809d08dadc3f24cc5a2b85e17f5b84
c6d1797200358eff569b67768495b987d0e202af
F20110331_AADJTK he_j_Page_178.txt
2f65c37f0c1190a05641da3d4314bfe5
7f65495ef3c804e0c67878e2f3b11cfe195d559f
87767 F20110331_AADKQE he_j_Page_202.jp2
54df57a87da5582faa1c2d716d18c42a
c4846e5c49ff2c79f8319ab24f414649d60dc6d2
F20110331_AADIJH he_j_Page_186.tif
edc6b26574905ba5f998daf359e41c2c
e39988e36012da2600cd84349a426ab6af0e3600
135122 F20110331_AADHZZ he_j_Page_016.jpg
c70d07062ce263952fa0f75340ce4e32
fcb9f56ec8adfe6a1c2564781e10a94495b1f8b7
8893 F20110331_AADJGA he_j_Page_107thm.jpg
7536c088eec45516bea36df88f6f4ec4
b42b944613cd8ddf1b654a65018147d35e25edbc
6494 F20110331_AADIWS he_j_Page_115thm.jpg
15095d5ec0dc187511ba4626b303b872
59ef5fa9be873db5bc0a64eab83fafbe690d1170
2147 F20110331_AADJTL he_j_Page_179.txt
915177f2c7845fcc9874c5419d922ebf
fc9f0c07992826069569eff14b4b34be477aa87f
76225 F20110331_AADKQF he_j_Page_205.jp2
66b986037b6f9599910d1416d7ab0c9a
5d97e7ff60909de635715602463679746293d376
8817 F20110331_AADIJI he_j_Page_180thm.jpg
0693470c8a75d29a774d725e2cdfd6be
754408aa7d01c347cc837c54b4e296cce2b1b465
50740 F20110331_AADJGB he_j_Page_240.pro
fd2fc488c4fe4bdd25db01fe30dff405
d4940b709f559982f78c18861895131cbd257943
F20110331_AADIWT he_j_Page_102.txt
c0c2e0e6ad622da6c39a0b26ca73c96d
10167ff8187015d769e529b1bf2ded363600315f
2140 F20110331_AADJTM he_j_Page_185.txt
2526964d3d7e7bb244108b13f6c614e3
b1bc6e792920187f5d9bc4c244cd70ae8368ef28
83323 F20110331_AADKQG he_j_Page_206.jp2
e288a345e55717c5c5ea93a0fa085f1e
3e35068655ecb4f8b8da7c95fe4119ce82dc5d54
2130 F20110331_AADIJJ he_j_Page_070.txt
ca08149739ac1f2dcfaf220097937b1a
d350c91139b2f5b22bacc2201ddcacf968678795
8410 F20110331_AADJGC he_j_Page_191thm.jpg
f674b229f89d1a25b3a7977d43b54d09
0a7f8ee1c4a0b197c2063a75af8fd4ae73df0e5f
112361 F20110331_AADIWU he_j_Page_032.jpg
7a7b78cc5b6c46a7496be6470a25f527
7ee271730de94820f75fe8778e9754df78c23c7a
1983 F20110331_AADJTN he_j_Page_187.txt
55d2cd0712223c178bc497ade0859e06
27cd4fcc97a05b311ce8ffe5bfdf6c148ee305a0
24083 F20110331_AADKQH he_j_Page_207.jp2
2c7ea815ab27662fd51e2ee0ec548cd9
78926449d519aa42c60f8d59a8c41a21f60449a4
3734 F20110331_AADJGD he_j_Page_258thm.jpg
e076d78539458d0e04a018fa452148fc
b0f206471b408076c01f39a3b1f214a38af4d76b
F20110331_AADIWV he_j_Page_092.tif
9ecea4986fa911077024d80ca678a067
063760057e1b446a026e3315c6f2f0839e6684fb
2090 F20110331_AADJTO he_j_Page_188.txt
0057f4cd2164825d4f10ede695616f02
7cba4bd3cc840c35e3f5cf7235570175bbde5f91
115067 F20110331_AADKQI he_j_Page_208.jp2
97ba260837884ef30b222cd8b08cebf1
e567e6ac2a31daed88460834e39fa21cb24ae32a
55631 F20110331_AADIJK he_j_Page_032.pro
e94b88c611f42cea9ea34a253bec2097
7b922e9f6aee99ab2ca98680e3d694935424e924
8622 F20110331_AADJGE he_j_Page_218thm.jpg
9b1fcb7548051869f9f9878e4ee6da94
07dfcf9ad218a8617164fb71ae0c67eebd75bd4c
1818 F20110331_AADIWW he_j_Page_137.txt
d175f6411c0c1be13511a74d02ebde5a
9cdc2ad8f66622b39ec4ee583149d8cc4ec21744
2118 F20110331_AADJTP he_j_Page_189.txt
c48ec1c67b163ccb466729f0396b4684
21a2a6d1c538e674d092b10314c8338594d4b5c5
113941 F20110331_AADKQJ he_j_Page_209.jp2
cdd61b5f962f0fca0e7dc1772c4bf6af
3fbcbb62f1401dbc46c9805f39aa4d1b2406402b
66760 F20110331_AADIJL he_j_Page_288.jpg
a0abfd2cd2cec38244447dfbb7101f21
a272cbe008af37f4699a936743ac6750e6ad2758
4992 F20110331_AADJGF he_j_Page_007.txt
274652ceadd3a63ca460d2eff9e703fd
be40f2379dfe2e51b9e68bf584e0fd972fb0448a
2021 F20110331_AADJTQ he_j_Page_190.txt
aec9b8b60f9a272d800cdceb4285454e
55757dca0c04a2e45214168db57a63b7501911fc
110925 F20110331_AADKQK he_j_Page_210.jp2
8f9293ac91ab6644469e5f3433bb2eba
07889a1a721caadcb10a57cd3927eed3f1ffbf86
1931 F20110331_AADIJM he_j_Page_186.txt
9b068251f2cadc82b711a0cc2e45f32d
ee7736c6b882af267f453a8b2999faa31b763219
49966 F20110331_AADIWX he_j_Page_187.pro
858622c53b20b0d99bfc8c01eb8a824f
3203743882dc1d75ba4cf86ad4f50f8b168a801e
F20110331_AADJTR he_j_Page_191.txt
541c9a261e8bac1d474f29562eea0914
48548a47fc9880cb831bcce1b34d4f135ddfa37c
104989 F20110331_AADKQL he_j_Page_212.jp2
f98772a71a2eeaf2e23cb1e841f665e4
f0b064536d2f7640c3617d59c4aa38747ee9576a
F20110331_AADIJN he_j_Page_304.tif
4b5634777f50510654c949612e238b3e
1c519af49890d62949de472815e2439675ecbff4
2010 F20110331_AADJGG he_j_Page_050.txt
e2674a1590501986b987e1932f38f0dd
c377ec6f40edc24713be9c39e7b61767789fa6a0
10819 F20110331_AADIWY he_j_Page_298thm.jpg
d4a0f586c0ae91e2265a7e86db6e220b
a842d3342772f9f2fd45e1ecd9fd4c46e2f8e5a0
46822 F20110331_AADKDA he_j_Page_260.pro
1323e79441569889cfed7391b823b52e
780c4624cf930a77baf37cc7eb9f9a54f72090ea
F20110331_AADJTS he_j_Page_192.txt
0181f1e874d7c13d89951ec7dfe75229
97cc65eb1590d03b32643febb02d57f60d528bd3
110263 F20110331_AADKQM he_j_Page_213.jp2
c2f88e26b153291b2ef03992ed88c847
361838c5b6ac96e1724d81823d955d9ac63ec40e
F20110331_AADIJO he_j_Page_095thm.jpg
0aa8663c4b79a25964118b26a1a5c246
1aedfb83ca1e17bdbae23a564aa900520d760b2e
4519 F20110331_AADJGH he_j_Page_265thm.jpg
8287297bc7a2bd512b8673f4263671ac
7aa5a7e961d1fb22f57dc1e8ec33237cb19fbc3e
31800 F20110331_AADIWZ he_j_Page_041.QC.jpg
d3696a338a9a67077076d33b33f08b79
ace3b55913df5fc40d8bade5f99f1112450a38f3
27339 F20110331_AADKDB he_j_Page_261.pro
5f04bd98ecbc25b0d2f5da3e00220d38
865cd81b41d3015fafbb98dcab6039b4d0fc6f23
107085 F20110331_AADKQN he_j_Page_214.jp2
acb46b9bb581cb52c0d78e383594c3ff
67f5dd4086f1a9689a5dd9ced8d79d8da49f43b6
857 F20110331_AADIJP he_j_Page_305.txt
4a8ec6f51e4d600b2aa9cfe59db08168
7c1c84a444a474f1d2095e8d64c8f48be8526045
93773 F20110331_AADJGI he_j_Page_088.jp2
11f20ffaf5819a5752436b14f381fa1f
0712222b4c76425632e231de4bfbf5886c0b60e6
F20110331_AADJTT he_j_Page_193.txt
b12706ee83611b8d35179b7e4a2bfb46
62bb06e2f9217ca98770b46a159dd598fd8c9a24
119599 F20110331_AADKQO he_j_Page_215.jp2
17250f4cd690b57066b08e43b0e71ff1
b696524f5c9ebf1285ffdf939eae04a49c3cc6cd
103123 F20110331_AADIJQ he_j_Page_221.jp2
6d1481e39006b44fea49b646df00b0c9
5e68671621f6c4d63119309905be4aab71eec07e
115032 F20110331_AADJGJ he_j_Page_046.jp2
69d25d58f36ff8658e732729dd5b82b8
9b0280c0a18e4577993bfc152c38bed9bfed594e
33074 F20110331_AADKDC he_j_Page_265.pro
5eeb4df50ddb00f9b2e9af4218821cd2
47572e93025000d46b4ef0d9fc033c43e5cbdc99
717 F20110331_AADJTU he_j_Page_196.txt
e5a8929d1aa27c5515c73ab0b5485fce
fc8e1c5805d7cddea81721dc55bdfe635b7c3a64
3692 F20110331_AADIJR he_j_Page_002.jpg
7b889bfa3740024efd72f7cab22abfa8
ca8ae9206cc6588865549805a24fd31902da3249
104917 F20110331_AADJGK he_j_Page_211.jp2
5fdedd5247d4b5db6a09a863aeec8dfe
8c4562644f6acb74ece4bef5ad6014630e92ff1c
28665 F20110331_AADKDD he_j_Page_266.pro
6805d9aceb3f21ec0dc93f36faceb411
ca1c93ecaa6d328621c27b5ac521c8adc3982f55
424 F20110331_AADJTV he_j_Page_199.txt
9395aa4ea19867963fcf0fad0c2dd221
6542e137cd1a6a6d95b0ec1e7346f20da3d034b3
115294 F20110331_AADKQP he_j_Page_217.jp2
bfc43c3fe45c6c46e3f477d3066eefa9
2ee8d4890872ad8e8d3e1e82955cc6f397aa9530
2059 F20110331_AADIJS he_j_Page_184.txt
cd39f36ef258f5bba8383329181f6bd5
5078c1de6a7a67866d435aa24c422eea052a7cc6
159940 F20110331_AADJGL he_j_Page_007.jpg
6a249b121ddb0b2c03082ecf6e73f5d9
64c937d355598bf5922f24b5fe25925181209c12
33019 F20110331_AADKDE he_j_Page_268.pro
1222adceaa6feca94a54f6f872c3181f
125399a64c6278cf80cea51e44197182d81b8905
2561 F20110331_AADJTW he_j_Page_201.txt
9d71400e74b676ed8ab9c2f280a14afe
34bac0849ac202cf204aa04d7e13b0e5f96c613e
113984 F20110331_AADKQQ he_j_Page_219.jp2
b72291564c4b1ce6c25578e3845cdcfd
4d01164597ed621ed56826e7f4b1b4095710afb0
36038 F20110331_AADIJT he_j_Page_299.QC.jpg
e087615cae01236156c3d79e156b12ed
f524f28e1981c6b95c523f10c4601ef781135e3a
7663 F20110331_AADJGM he_j_Page_067thm.jpg
d30fae84843d57a0ffa8cde16f5a5007
dfc0b97b165db7ea8aec5a895333d41b247a950e
31263 F20110331_AADKDF he_j_Page_269.pro
ae982d96d9aa1447aec15ca6693072cd
09dac756adf58f0950ae0f6831ff2cf81e6cb74c
2850 F20110331_AADJTX he_j_Page_202.txt
269bb18ca3e2808656ab0f34088fbd2e
a0eb41af170dc0984b8616284584416c2d0c6a4c
756429 F20110331_AADKQR he_j_Page_223.jp2
0f1db141526abed8b485d6ae412ff71b
5f9a12895f04463196d099171df630a44b1d1913
21657 F20110331_AADIJU he_j_Page_020.QC.jpg
4073b66e3e6d8d09b83fb0a7e71f08d5
daabfc10850fe97abff5e51badb6966c15b82cda
1542 F20110331_AADJGN he_j_Page_288.txt
5df3755c27ee42711d6f17514d65d0b6
a6a122a92fb0869ac389b52ada9b5ad3c741a7ab
27004 F20110331_AADKDG he_j_Page_272.pro
d2c4b6451b25b153220d3ba881e2a679
0a3236c3c4359e1e9e411d47d8fb7fec07c7f165
1966 F20110331_AADJTY he_j_Page_203.txt
f5804017859931ca6a5618c3f1e81e85
f08b9bf0be70a76963b7fc3b373c0b8c67ca4db4
13886 F20110331_AADLAA he_j_Page_128.QC.jpg
d0883307b2ffba344d33cd5b481b9e5f
9116f2c16db63646bbdbe6fa100c7c86bc2dea05
572791 F20110331_AADKQS he_j_Page_224.jp2
f0702f8a2da91201f78dd996831b0461
b7ded51e1b6c60c1afa64911815a1ecfdcc60172
36952 F20110331_AADIJV he_j_Page_215.QC.jpg
b24bf66a9b5d756d95bae5aa6cbe1c18
e3cd2f4f50c0e710ef216b98880ebde81866e00b
117959 F20110331_AADJGO he_j_Page_294.jpg
beffa7db40f9f3f784dfdba119c0fdec
81861ecd0e7860c4d87d2b0698e7f2ef2e50a14a
35946 F20110331_AADKDH he_j_Page_273.pro
7797d994c81f4cfede1f8107609e928c
30409ac9422dc1a4e480ba121ac0c5939a88f5ee
2482 F20110331_AADJTZ he_j_Page_205.txt
c7728a60ea1643f3a1dc9b3cda3b4099
b0456e37d9d1b44429215326ad336196a16ed99b
35903 F20110331_AADLAB he_j_Page_130.QC.jpg
dd1fe5093af214b32ee92e340dd459a8
0dd852734269d4bf8c98fa3bd967da71c6971dcd
578701 F20110331_AADKQT he_j_Page_225.jp2
d108c1a2aeb6a298980492d5d0d75b79
3519aa765c7061d86d997a6676c9a8b4004bf3d0
F20110331_AADIJW he_j_Page_068.tif
11a5a1aeba2e5f9003fd2b2be35f3b7f
4ffd956ac7225eeac7ec7b362986a41c40f13a59
F20110331_AADJGP he_j_Page_315.tif
5b498477d9759f31f6e376e9088e7001
1fbeb9d0aaf676dc7b45cc944ce890452febc30a
30118 F20110331_AADKDI he_j_Page_274.pro
d05af0d217e8082fdf129479931b8e4d
8388f2e94de0ea06dffbf31486e858690f827882
35564 F20110331_AADLAC he_j_Page_132.QC.jpg
428d69e18dcad58e6b53563ff6fcaa24
3b54ab507c8bc9aa9067bed3b88118488d080407
788498 F20110331_AADKQU he_j_Page_227.jp2
808c34b2f6147413d9fabe576b82a779
b71c4e5bf606197d2e2dcbd2cb3431b5c2527a34
F20110331_AADIJX he_j_Page_174.QC.jpg
f9a938864fd1098491b4e59be9d065d4
0099378fcd589c996465d875d5a54fde275bd9ac
73945 F20110331_AADJGQ he_j_Page_201.jpg
206e644315daa236df7538d7d14e8dae
97fe63408c995ddf56d8b7913d5213246678fd27
33488 F20110331_AADKDJ he_j_Page_275.pro
d86a5d52beea0594c98ae8e17da8572e
abeccbe67bb9b92b5fd47c89537971bd34942b8b
34989 F20110331_AADLAD he_j_Page_134.QC.jpg
c7e51589b672944cde2d2e24ca58a43e
2283da928cd38167e034aa4738b4d9792e1530cf
607577 F20110331_AADKQV he_j_Page_228.jp2
0b833c6959762c43a7fce2473844193d
f88e7ee1ea4b143ff2e148cddf91dcee32dce913
18461 F20110331_AADIJY he_j_Page_307.QC.jpg
108b278d69c88dcb8e848a78f064eafd
6a6fb5da8498f1d0c7d595a76b38d8f3db202f35
90969 F20110331_AADJGR he_j_Page_072.jpg
a2950620434bce291fb2ce4b1e628dd9
9bdc5d1cc1963d27ce9d7af207aba6c6698db35b
34709 F20110331_AADKDK he_j_Page_278.pro
d04c3527260423bf74ac4bb4b194464a
7661aade7391c2c0eb5c6a50046796f44e01c356
34463 F20110331_AADLAE he_j_Page_135.QC.jpg
b56dd544565f858ef73bf7a8545c31c8
3fb3113500f6630f5a829835670786a04842756b
607539 F20110331_AADKQW he_j_Page_229.jp2
cfb725a8e45a8d29c2c9183cf29048b1
98b95cd40805d4ce9f9aee00b40c1c2f1e03aa78
366336 F20110331_AADIJZ he_j_Page_110.jp2
abacf39116c3249622b7c57fa834ff78
c81048f009e53e4160ab5dc9c09fef52f3380db2
702034 F20110331_AADJGS he_j_Page_203.jp2
e6aab2acade1416f6b900462f99c246b
993c1ef7e89b12aeb79d8e20aab9d0816764e500
34432 F20110331_AADKDL he_j_Page_280.pro
e29e3e8b629d418363bb5c20a4bd4054
84c07b3ca002e473b5e1f00585d0747eef7f3119
35124 F20110331_AADLAF he_j_Page_138.QC.jpg
f88aec0f601ce072fc60faef13c298da
83c29c5198c767c1347f81b734c1c9db97a5bf5a
310120 F20110331_AADKQX he_j_Page_230.jp2
e540655b505e17f4d8e7e400a3a8a464
9df2830ffa4ee411857e37f11a12038002a08045
55941 F20110331_AADJGT he_j_Page_238.pro
fd8b2b34c01138f150d1c5a6b3efb512
df41898ca5fbe6a678fcd40f2d177c38a5df1ba9
34805 F20110331_AADKDM he_j_Page_283.pro
762686042d04475f152fd46a58bfe5bb
db29b6cc28cfd4cac7a64ed339b799dc1c8b7095
34880 F20110331_AADLAG he_j_Page_144.QC.jpg
83e2bb38469a08ef078c1353e20c0557
8b784cce8e0b13461553a8968295333ad31d9645
1041348 F20110331_AADKQY he_j_Page_231.jp2
df24e06f3b487757a15977e47f4ca791
a4870c94d617bbdc94dfe9ca1a57b18184a9cba4
F20110331_AADJGU he_j_Page_171.tif
64eb535ac34df2481b44e0a9f7ec4e1f
d3067636de6a9e16a381442a1d5a7cec6f3fb844
35766 F20110331_AADKDN he_j_Page_284.pro
8f9f1907d7cdc92ebae6e4ade8710c22
91fbe4e05ae2848fe6fa1bcce59fbdb19fa0c07c
35495 F20110331_AADLAH he_j_Page_147.QC.jpg
ffd8f7d18ab8aa7a9146da4104368985
0ddf069fc9827df8db2d84bdf21e4e9b1d52a48b
55709 F20110331_AADKQZ he_j_Page_232.jp2
2a22c3a845f17df974052a27343c4624
c3cd70e332edc80f735825e28b962ac7d7e11921
51842 F20110331_AADJGV he_j_Page_218.pro
5c7f2dc13a31094015fa20016a704099
7a49eba179c56d12562dadaf5e7f4fc62c7e9602
36808 F20110331_AADKDO he_j_Page_285.pro
8f3635bc860f1f467c752ea869d2d77e
df60e2cf1b782f986be025b902138c16107d3172
36651 F20110331_AADLAI he_j_Page_149.QC.jpg
9731f2db1e4ec7b85a76c8900ef9a9c1
193a908da539961a3b3b11740352770a3738c06d
F20110331_AADJGW he_j_Page_012.tif
4e5a12ec30a4b2ba45a489eab92df4e2
b16258b72d0031bc4dac038a07f41e08e8ffaddf
38908 F20110331_AADKDP he_j_Page_288.pro
50d4a9d26cb263f2806d8f1bd417e1e3
23bccd7ab0b050b6dd9893a3534b2c31b36674c6
36260 F20110331_AADLAJ he_j_Page_151.QC.jpg
6d8ce6cbacfbcddd728f484caf075544
899b21a7f163c7fcc991f604cd88cbc94a8b6a64
34843 F20110331_AADJGX he_j_Page_081.QC.jpg
b59ce03b9dcd7c2a9c0106cc77f2aa57
b709823ba762d5a2009bdb82a6ea3345ef4dc6cd
32708 F20110331_AADJZA he_j_Page_092.pro
147c6b50887821813d30aa8945b942da
408d9379cdb1cea6f003741b84c49d1aaf4f5416
31787 F20110331_AADKDQ he_j_Page_290.pro
c6ebdcd2de589526bf6c7722df32d725
0746e740baa6099fecc323b1d2946f4be2b1327d
24105 F20110331_AADLAK he_j_Page_154.QC.jpg
c76b3818464b3eda9e8aa40208d676ba
48c4b632934b2cbe5ca15036e202c1333816cc35
F20110331_AADJGY he_j_Page_149.tif
ce9833d5b8e7b9c6197a098b1910c69a
9de3f7f7f3f169b2809ab5db599df095cc5a2d59
47846 F20110331_AADJZB he_j_Page_093.pro
7fd7f73d783eab2a609bbba8bd26cfab
9720f602b589c8943d60f128e01e0ea20f588181
35936 F20110331_AADKDR he_j_Page_291.pro
e1b8a48aa7fdd7efcc64c78b268960b0
c366aebf3e10431e2ded449254fa7d97b4549379
12796 F20110331_AADLAL he_j_Page_155.QC.jpg
3f33f576cc2b2f9433a16d4b81dc2470
70836ca2e5b7912b379264b1c513191491245318
131679 F20110331_AADJGZ he_j_Page_013.jpg
87de07bd66189245398dad9a3d73e76d
f82fb9d33a2db45eaa687f130a1a553537bb1a88
46591 F20110331_AADJZC he_j_Page_095.pro
b7ecb5161e9081f2ce12fbb067b1acac
602e632564ddd11109675f3b51fe9acefe211b61
13872 F20110331_AADKDS he_j_Page_292.pro
0aa98e08dcee8adf7dda23e4634011db
d46e352ece5f20715ec7942dd6c169e244eb4534
21045 F20110331_AADLAM he_j_Page_156.QC.jpg
4fca98332a9f7479bac43e890e2a1631
934474912d973477436a6b4d5b031ab900bc1ac7
50134 F20110331_AADJZD he_j_Page_100.pro
cecd5cc774ed0ed7b2dc5c503a81ea54
36de80b7172d5b3015d0fe6fb08944626e8e3b98
3963 F20110331_AADKDT he_j_Page_293.pro
f31eea85181eab1135b9e7714b2d93dc
cbe574408dc9a09f3b117aad3637f79f04f38e5d
20363 F20110331_AADLAN he_j_Page_159.QC.jpg
b2012072e1c3c8d671ae0c555852c8ac
ce1c85db80c1c6f693e8cd98be294c56a4221400
32819 F20110331_AADIPA he_j_Page_017.QC.jpg
57707ceddc4738e262e3a24f0af13c93
e40d930d52fce0e4ae33ba79319b7c6471a66eaa
56492 F20110331_AADJZE he_j_Page_101.pro
589e9d47aae9d5223b208f06cb0e0e4c
038c02696562f8c065a42fa4c9130b3bf48ba525
4028 F20110331_AADKDU he_j_Page_294.pro
bc3cc4972758acd4b317e891382b14a8
86d8e33e195be51d54befd17db666647ddc4ebba
23581 F20110331_AADLAO he_j_Page_165.QC.jpg
eeb590ace65d552fc5797dc971e92227
12f83f07e6307ffb02f95d58c02b228c269d6331
53373 F20110331_AADJZF he_j_Page_102.pro
98cb0a19fbe98e378e29ebe3ab241c13
fa28a7b51dc9f70cfd2e1d52034c1a47061b89d5
12897 F20110331_AADKDV he_j_Page_296.pro
ca9e0358b44ee2d150471bfc2f4594f9
cfbddc427eebdb5329438c55798ce710e5a6351a
12754 F20110331_AADIPB he_j_Page_207.pro
0cda6fd7b04012c4d3292e6e7635c6a4
0923657df9cef9b8a55c6692894c4b04aa0d9d5a
20790 F20110331_AADLAP he_j_Page_166.QC.jpg
bdd7c9262be7c7dc31f9391ffc4d820e
e56878ba97262f76c42b35f566dda650dd9d414c
51438 F20110331_AADJZG he_j_Page_103.pro
86e1ad30dd94efae0084c9ffdbad9853
94c6e71297739aec9e0f3d72b59d8bd5a4d5d3d5
F20110331_AADKDW he_j_Page_297.pro
04ec6fdb4b90faae28fc2abc47caf194
44d58082d444d1f0b916020d033bb976caf9eb9d
F20110331_AADIPC he_j_Page_028.tif
18019342e8b339d5d51ec95a4aaaddf4
9c8fbbcc9273ab83a97255044aabc6ebae955306
34970 F20110331_AADLAQ he_j_Page_169.QC.jpg
428bd310e792112b0a561b132c3a44e6
3b45b22c14101e52d28adf5e6ace9dd8cfe4f344
8492 F20110331_AADKWA he_j_Page_040thm.jpg
1fd26d4c41a7484d855ff26ef0edc0eb
68bf17775b5c7bc525e974ba7ce888e39f6bd7f9
52180 F20110331_AADJZH he_j_Page_104.pro
31ca8684d5f982ea6f55da22200443a7
aed189af75b2957ebdf249d198872ce8a0f44911
7575 F20110331_AADKDX he_j_Page_298.pro
adb55b26e5d4894bbeecee15ab70790c
7c9009900578fd722c32700363ccc124506028af
9010 F20110331_AADIPD he_j_Page_097thm.jpg
bd8a5c5193fbfedc70f419ceeced2c77
c17d604aaa120c374b02f91bed370a8d1450b068
35570 F20110331_AADLAR he_j_Page_170.QC.jpg
374b9ddbc0e8b2b0505d6708634b6067
9238fc136d94f607a770c03e824c15372debb31f
7434 F20110331_AADKWB he_j_Page_090thm.jpg
c8a2127ccb54fd97211a61814a5cf655
829388e5264d131ebecb03bd4a279ad92c1d613a
53176 F20110331_AADJZI he_j_Page_106.pro
b5595ec59eec2670bec1ad8d6e18cd08
f9a9d5098325fec5c88c7153bf2264eaa09f0ad4
F20110331_AADKDY he_j_Page_299.pro
b032be93081e5430e4c977f42deb0dd6
6f6ef362682ac8741397310085067f1cf4251694
9449 F20110331_AADIPE he_j_Page_322thm.jpg
07650e9fd79329bc4c516b28b9647ef7
d691abcc0391a7ac628046a01377576dfaddbbd3
32422 F20110331_AADLAS he_j_Page_172.QC.jpg
8a787333087fe2a2aa738eab07f91eb0
276ff5d056aeaa63fea430096219f49a2e41a783
36719 F20110331_AADKWC he_j_Page_057.QC.jpg
cd0ebb46d0c57f09d62663737d9caeb5
7eeb0bb2c96aad26fbe7c97d1dd5cfbb478c605d
F20110331_AADJZJ he_j_Page_107.pro
1c2d114f16da4b68a631b48de2c67b0a
3c8440e1062c53e6fdfdc679833ce80245a46cc8
1820 F20110331_AADKDZ he_j_Page_300.pro
ad9d7b3c8cf7e70d24fc25e6bb32f84f
4e8d4acf4232efdd2fd2d52f397135e15a7e8804
1631 F20110331_AADIPF he_j_Page_119.txt
27c6e9532508240e58ed2bcd24038a7f
35eed28eb89f03e912ad6d0ee5afe08eafcb40f5
36977 F20110331_AADLAT he_j_Page_175.QC.jpg
5f0c8653ecd4237d13dcfdca25c82f00
fdd963b4566bc25dd49df6fd2029b7b18bf82352
8557 F20110331_AADKWD he_j_Page_190thm.jpg
f8cc09153e43bf7427e394eef47c1060
0acd67bd599879b8121782976b8379a4c9025a2e
13281 F20110331_AADJZK he_j_Page_109.pro
301021bdd568f6678df917737624f848
25bbb87129bf27fe45d96ec8b84e7740f125ac2f
109386 F20110331_AADIPG he_j_Page_178.jpg
998c67b22a279ee6ea5f3b814cec4ce1
fe2b121b225fd3adfba59cea193e5ba2fdec0e87
33171 F20110331_AADLAU he_j_Page_180.QC.jpg
8a450e040c4001ceb9b561be6a9ffd56
b70ad69f0aae0a37aa18774fdebbfc77d007e7b1
33176 F20110331_AADKWE he_j_Page_210.QC.jpg
e370dcbfc6882d5dff04e683214ecef6
6cf4023d99236bd0c3a6cd32d96a4cab1784dd31
F20110331_AADJMA he_j_Page_183.tif
574e798d5f793226d20bf36e2fad9381
5e1d934aabc98ef7603e3b7b3cc9d716488cd579
12911 F20110331_AADJZL he_j_Page_111.pro
55d92adf989f43dc9cf016136fc9dcaf
8ba5c9cbeb5c7c21f0bca6a41d703c3518969d36
33808 F20110331_AADIPH he_j_Page_079.QC.jpg
1fb4f92177c6fda728eb8bb442d5af80
3301909dcaee8c3a22cc6be656e580e0abca117b
35581 F20110331_AADLAV he_j_Page_185.QC.jpg
a91ce31ce1c45f6ffb3aa69c05b77868
54b59ecf314acff7180d91b4a331de72d3602e28
16479 F20110331_AADKWF he_j_Page_010.QC.jpg
264618e134e8cdd24161e08947abd9c1
25547288c466c28bdef8759c5e3347bd7b06da98
F20110331_AADJMB he_j_Page_185.tif
443edb46e3530fdaff8ee1511e0c2575
9755d4f3358678aab9a029a9dcf1b108465854ac
15186 F20110331_AADJZM he_j_Page_112.pro
9e961bc8568e2fe7294fc18c83372ac5
d27fd2ec1fe1a0d5da23682a09a77c029a41c440
F20110331_AADIPI he_j_Page_171.txt
197f7a01274cc4f46ac5d399ae97354c
509e50099d547b10dfc570429278f7f68d3d595b
32649 F20110331_AADLAW he_j_Page_187.QC.jpg
173ad99bd5a790994c9e2aba59430cb1
59468ae09f737381df5d8a738160c7caf9e97f84
18365 F20110331_AADKWG he_j_Page_275.QC.jpg
7665cbb169cc7a8ebb9dd3fe5505c0af
f34da44e5fa7707392c5680f8156dcab8f454d18
F20110331_AADJMC he_j_Page_188.tif
7e3c418ad18ac611071c63270cab5387
f22ce9b251b6099b3038e78f5518d9deb25dc63f
19390 F20110331_AADJZN he_j_Page_113.pro
dc8d3e6584adfdb3849be6649bc3f496
4bc857246816b861af0d13978266f817ab1c5a12
56582 F20110331_AADIPJ he_j_Page_098.pro
f0d34e204252666f863296ba942beb31
908f4fd8f49ba1a71a1c544ed1120a6488ec34c6
32480 F20110331_AADLAX he_j_Page_192.QC.jpg
0f13853c40e9401431d781e803740452
58a431695a05972a4e1230f4cd814959a1df0409
22355 F20110331_AADKWH he_j_Page_222.QC.jpg
f5698a6e9a276cb4945417dc4c48fcd0
4c4f3673668a8c8a9b3009dd6488f4f12428f65c
F20110331_AADJMD he_j_Page_189.tif
c205e57d86c60b1484014a0519a6f535
b931cc6554e10c729b9ef0ce62c551c2dbf74d81
23075 F20110331_AADJZO he_j_Page_114.pro
4292a72f9c6c51851ec9ccb1fdc80eb6
2aead6ca3d8e841d4e175b5d8423deed1e8c0b0d
F20110331_AADIPK he_j_Page_162.tif
8abf3034a84f7723e9c0d7273a9e9ab3
9dc96498693b105e3a3ab9e608e5b6b81219d047
15987 F20110331_AADLAY he_j_Page_196.QC.jpg
1fa73040f603c4c9b2c370c14757c84a
7c05e88447ecf49c15fbdf00a8281143dc735f66
3935 F20110331_AADKWI he_j_Page_261thm.jpg
e4580e21bb2f2eb85a6b4ec6f6a348be
9c719bbb9694180209f93d740c1af8604dfa1b79
F20110331_AADJME he_j_Page_190.tif
66aab6b30ecb314b948d51a041db015c
1efa73582719a6f96ffa25e1519d4b38e6ac82d2
60373 F20110331_AADICA he_j_Page_159.jpg
4bb4b3fbb0f957f911f38d0e83da9b2f
777db68dad2f142fca7b9530c8b1bba2b04084d9
23853 F20110331_AADJZP he_j_Page_115.pro
b91a709ecf1f712d7381ba24ccd54b0f
f0c38e05c264ccddd708401b144f4ad5c0dab834
F20110331_AADIPL he_j_Page_223thm.jpg
9395c3313a196150c4d2eeb25490fce9
6980da8488483c4ad540b6f268b91a10442299ee
15472 F20110331_AADLAZ he_j_Page_198.QC.jpg
383aa6ded0ab2ec6b64ba9beb03f3f32
b29eb59daa26695dc116524d67b921797c94277b
F20110331_AADKWJ he_j_Page_070.QC.jpg
f6586ce8310d8c1d85a04af8201877e9
9a3544f8af2c8ca3298033314879e4b1d340440b
F20110331_AADJMF he_j_Page_192.tif
07506d754078739069b1bcbac0b0347b
046ffb79d7fb49ce9f90439cdb58f8d0990e6ae6
35701 F20110331_AADICB he_j_Page_033.QC.jpg
4afcca5354584b1a5bf5f59c209ec540
893931d1076c14885f32ed0aab8094918b7b549a
28586 F20110331_AADJZQ he_j_Page_119.pro
e342826befe7054a83659baf2fedb114
dd9a19e455c2f65e4476dfa497b753df971b8be5
6505 F20110331_AADIPM he_j_Page_167thm.jpg
6747b03ca35c516dcdd5f89dd15d1098
bb9c4b2ffc5f0abeba757b95ba305c2abe8a0031
9125 F20110331_AADKWK he_j_Page_151thm.jpg
fe673f4ede208d7e4fb8ee3cd1e7a976
9bcc7fdd6f3a164adccb6941a19cb70f88ff5498
F20110331_AADJMG he_j_Page_194.tif
f7361930f5e9266d66d65704cd067878
d8003b36ff8d00ed1655c4b0d29e4535f9b43548
10201 F20110331_AADICC he_j_Page_161.pro
afc41a29098ad49eab6081940067f12f
4f5877eda7af839d25b28003ffb8f0b850b2a72d
8585 F20110331_AADJZR he_j_Page_121.pro
e5afbccd463588c3c44814620550f199
ee0d0cb52a50a3e2272a01db2d6ba9965f388ddc
9934 F20110331_AADIPN he_j_Page_317thm.jpg
bc297694da696ad0b0513d940be5273e
bb2fe6b5126d611f9f34f4d6b1b44cac6ce35014
26495 F20110331_AADKWL he_j_Page_202.QC.jpg
ddb8fe565f5a6c3a64c8c6d481a8aa8d
0635de8f7509c47dcd403b9834f0bf5175c9d5af
F20110331_AADJMH he_j_Page_196.tif
1bfbac4ccc811384ab33939083bebf76
ed612f887fb0b7ed53b4d7af3251faaa2e1351a0
100810 F20110331_AADKJA he_j_Page_191.jpg
494a062283bfbf12cf44d2d25c4b6d69
13e3bc1ca739e39996c20793921a9d1300f49951
79036 F20110331_AADJZS he_j_Page_123.pro
6bac7fd719cd3ba39da5c1698431919b
58c5ca17678166349e4832d04cc361f7bdfc42e7
26458 F20110331_AADIPO he_j_Page_042.QC.jpg
bb18bf9059c3e0ae61090ff4e68ee498
dd2da6be7b20fa60acb433964c4d67fa6a20f8e0
6625 F20110331_AADKWM he_j_Page_252thm.jpg
ff1ce9563f63e62bb6142f98241942b0
c2d0bc81c8f07e834ad54cd87b88e8da4a023b61
F20110331_AADJMI he_j_Page_197.tif
bd74f37385a74ea6f7df7050c4e65045
1cb61fdd1f414e8c83c4eae6d22999f6df313368
98219 F20110331_AADKJB he_j_Page_192.jpg
23c65b5ef82ff7b1827725cfff16bcbc
66693370645d6a0d5e377c6ae32e346adff35f57
106832 F20110331_AADICD he_j_Page_242.jp2
f83456fffb7b5bccb5141e64b37e07ea
0bdf6dcb78bd98fa4275910cc94a60c037425280
31364 F20110331_AADJZT he_j_Page_124.pro
bf32fd9d9b41a7e51b4ae7849ab23faa
9347758f9f1b13f8fb7181e280cf9867abd96e66
F20110331_AADIPP he_j_Page_276.tif
a1bd4421e227be351886d47c59d2865c
3dc3368558ada29a74cc46b2884e6bdbcde3a9db
30216 F20110331_AADKWN he_j_Page_072.QC.jpg
5bf092bde6a0d91f4dbf9330468e9f45
957b94eba86e2b050bcf3892be4edf8eeae53e32
F20110331_AADJMJ he_j_Page_198.tif
4eca57e88e5c49e8b995c6005265eb29
0697efbcda7848339a916366b2ad70527798f125
41584 F20110331_AADKJC he_j_Page_195.jpg
6974d3a1df77b6f221fc742f013d087c
c29cedbb67743de1b6509524d9f9eebc69664ad6
2110 F20110331_AADICE he_j_Page_069.txt
01e1ed319ca72745372c68cee7bdff39
e726192a4ef9ad3f6415c7279f5478d3cafa5edb
48450 F20110331_AADJZU he_j_Page_125.pro
b24768b146385ce8d0723e7593c3b839
08a5e23ab6a40b8cefd5765eabe6e43970960e95
4304 F20110331_AADKWO he_j_Page_005thm.jpg
d334a4b6eeba3286defd590d3eeef02c
e6551bc81c90af708a4d83673b5108bd59e3c1d1
F20110331_AADJMK he_j_Page_199.tif
38fadfece22008e89c9b0cc438a17329
5b2cdf5a48d004d440b373ada6f4697e0240ee4a
23601 F20110331_AADKJD he_j_Page_199.jpg
5ac323d3fbd6d607a6384da78c2d4e13
d962dfa86a533514caac605eefb96b0f49f7083b
F20110331_AADICF he_j_Page_067.tif
fb62f8c7552c6f87a5fd0dd55e2172ba
421b3fe38444fec5260090e2e3ad2b4bf2d2e4c1
23955 F20110331_AADJZV he_j_Page_126.pro
47585dcb27fcf989b9f342e68dca8824
5db0a0d07f4b891b1a1e101246a16a91336ba0d9
35076 F20110331_AADIPQ he_j_Page_218.QC.jpg
4604d7d82c7fa8ae0e3abd6db913d642
ad506281a3dba4c23c4ed25436615e3272765ddd
8389 F20110331_AADKWP he_j_Page_093thm.jpg
b55e0c3e25d37e2d1262faa471dcb514
616e1da8903eb949e5797348c3c580d1cafd4c92
F20110331_AADJML he_j_Page_205.tif
51d50f6e6a33e823f388b6888ab3d1e4
463a6dbec54cffa983927b6c35ce3e9ed4eba5db
69892 F20110331_AADKJE he_j_Page_203.jpg
0e240fb823aead738358652a086bddfe
cff39334425692067c143a73df08928eec985738
2207 F20110331_AADICG he_j_Page_238.txt
323ddfff39116b123587588bf9eecb94
deef7bda2261f2e66a048d380b1e9e8ba437ddaf
35069 F20110331_AADJZW he_j_Page_127.pro
73b4c0088b0daceb338d75be91df5b79
31a70acb3ce1e539f941b71570c229a0d7c1a491
690 F20110331_AADIPR he_j_Page_314.txt
c033e9196e855ca290688132d114b07f
eada1a83b1c97c1d5631061489e9fcd77188dbef
8949 F20110331_AADKWQ he_j_Page_074thm.jpg
b565f122694fd65367f6bec383a5157d
b4279c00dc337b046e35a91990ccdd9b563de231
63479 F20110331_AADKJF he_j_Page_204.jpg
7d0c035fbabc3c067c5c1c466304d514
a318f3fe4b8f07746bda7629440c63bfb522eeea
F20110331_AADICH he_j_Page_063.tif
5794a4edd01a75485fba7e988a2ffc35
ee2019bcda1228e9a68b0d4b299d0131a9bb162d
22700 F20110331_AADJZX he_j_Page_128.pro
4aabed402da09677d69fa2aa564e2e0f
43b034a2da98ee3eeef2f6b201138e8d16b857b0
117252 F20110331_AADIPS he_j_Page_057.jp2
374666078d01ae8d04772d015a0c717f
efe05bca23ff0b5b9617428137770c852ad8e58c
35811 F20110331_AADKWR he_j_Page_189.QC.jpg
66351a4cbb121c843e7b507535ea5dbc
b314d9a6afbe4e67ae03edab41551433981e148f
F20110331_AADJMM he_j_Page_207.tif
07b61e336571240643d67529acfc4937
a2737c55172192a0698482f46292955efabfc624
75627 F20110331_AADKJG he_j_Page_205.jpg
c2be81792a0ab2a6a823a63a89f8b72a
0dd0315f5385b370353eab01ce3f691f5ea0281c
2251 F20110331_AADICI he_j_Page_177.txt
80b2ae6bf286d89b694905f6cf2315b3
9ed4da1ded9d8f2d6a01d319736986d3928003de
54104 F20110331_AADJZY he_j_Page_129.pro
da8f6b352434307aeb0d9ea2c7fd9ed5
b6b8528ed03f4155772fbcb35a840f11443237ca
18307 F20110331_AADIPT he_j_Page_162.QC.jpg
da5b88037edb9c188237848be9bd9bee
ca1640b3994d5494ce9cc1f4ede736f2781d6d03
5134 F20110331_AADLGA he_j_Page_196thm.jpg
c2280776669706350875fbb99cbdde1b
8f8cff2c030078f90c098791ffabaa0f7e9464cc
16906 F20110331_AADKWS he_j_Page_251.QC.jpg
cf8e63a156e99c838286af34e8889a80
3770d08116035f546f4a280c8de0063739cf99ab
F20110331_AADJMN he_j_Page_208.tif
41567d2dac831beed58e63fd99ed5e60
b03e1963fc0db6dc32324c939603210d7082b151
51092 F20110331_AADKJH he_j_Page_206.jpg
04fceb36aca737f9b78375414de82cac
2c9d6c3b5d29f42a13c4f318c15313b76a5b8cd9
82679 F20110331_AADICJ he_j_Page_042.jpg
591b7d90e75f9048f83816ed51463680
ff54df4236cc28ff0f95b780b52cde9efd7205f2
8462 F20110331_AADIPU he_j_Page_068thm.jpg
53e62cd997ccc2e27290af1d32a7977f
dbb3b5b74a631370e738acae2d2ebb0dcd4b06fe
4944 F20110331_AADLGB he_j_Page_198thm.jpg
de5668bc5b1ff4580505718eb6738331
e488cd62ae1af1fb08195a63062fbbc8d03a27fe
4865 F20110331_AADKWT he_j_Page_225thm.jpg
0ff86fce3934152e4b283928ee7a4ce9
4b68619f074f69efbd4f7aa9049a1b69ef264bf5
F20110331_AADJMO he_j_Page_210.tif
4b4e27d438d2f648cdb77da6a042af53
accc5fdf32b0874a9f5118ac58f87c9f32befd48
F20110331_AADICK he_j_Page_071.tif
29dd89470792e446c1e32685ddf688d5
505fc58c5f2090d75f22f74853fc9f02dbd5168a
52826 F20110331_AADJZZ he_j_Page_131.pro
6ece0da564d1a8cd50a0af0a9dc1b85b
8a39f0cf88f3fc84277988e3ad3ff9c8fd9a4235
53989 F20110331_AADIPV he_j_Page_130.pro
8b9767091f0bc9eeea16e796f382eb66
9d887c740850c0bc38122dea7779b9c47bfb3807
7076 F20110331_AADLGC he_j_Page_202thm.jpg
99d1166e6aa3843bc2d05879dba58635
8a9228aebc4ba2fc5f9b0bedb9dcddd935478d01
20053 F20110331_AADKWU he_j_Page_204.QC.jpg
47f78116a35004d5751a25caa0c076d8
c5a265bf2311b34854a649cb0e559451fce1be32
F20110331_AADJMP he_j_Page_211.tif
fcf0b830b9c431758ca5d6cddb9c386a
e8fc26b6e9db717bb554034a8fb2a271d6f6b510
25129 F20110331_AADKJI he_j_Page_207.jpg
67e9942cf6aacb4858d27807aadc504c
ba9aa919e9a893e1fda6f504676ca38ae9ad676b
51221 F20110331_AADICL he_j_Page_055.pro
f595ee0c5aa739cdf5d8a95b1a53463a
f4472cacce36698ffc538b807ad09098a4ba4c94
32816 F20110331_AADIPW he_j_Page_061.jp2
0b1b785a10f0b907aaf2a2a0cacd59e3
5532fe1ff71cc246180485431da83ce525d1bfb8
5432 F20110331_AADLGD he_j_Page_204thm.jpg
58b6b4ca25d0143c765fe6a917ceff68
772240a090109ca82cf70e7675f374714d1bb61e
F20110331_AADJMQ he_j_Page_212.tif
af9e2c52908b1dfd58a18198198187e9
660564d71d292450045404465920df7032ba2a40
106687 F20110331_AADKJJ he_j_Page_209.jpg
69fd2030d540a08dc19ebbd0c5547e3e
552b83c2f95c15b48c28fb88b8c4326e9d5406a9
F20110331_AADICM he_j_Page_294.tif
027a4efb0347bab23b4f1a637317fcf8
421fc807a785549231fa7ca3ae4e949b2da0394e
F20110331_AADIPX he_j_Page_174.tif
97d6013702eec379acc9c49c6760c493
d806ec133df06fe46611cf781592dd5dd0438bd0
5977 F20110331_AADKWV he_j_Page_114thm.jpg
4a48cdc4920c3e7aa60e7cd06f2e70a0
1b6a33ac0a915a6e980a45f6faf5ba84fab278f5
F20110331_AADJMR he_j_Page_213.tif
db1cb27d15690820e4d9c9d4502b05ff
92d17cd32268ae0400426aa006dfc4efb26c2a40
99816 F20110331_AADKJK he_j_Page_211.jpg
45698da9a5b3e2e7e93d1501eaf70275
b62c4b3d51715feeed554f8a71bf38943ae76393
F20110331_AADICN he_j_Page_039.tif
f72daff484e10b5377aaeb1b6fa1c462
7365e14a1a16bad0bc68e5a88f4b132f2bed3e36
F20110331_AADIPY he_j_Page_307.tif
ad52c36ae9a7caa666deaeb8cc13a918
ff99b0a8d3554a887271d595baf1cf85a1b4d865
4203 F20110331_AADLGE he_j_Page_206thm.jpg
9df743613bc25fda13996f09ce9c2334
c635737decdc41b11a16352f2e4c057664bb1dab
36050 F20110331_AADKWW he_j_Page_008.QC.jpg
f1934171575ce04a2b20142fffe450f0
1024b882ad569f9905aae874a51d0bc5a0d5dd7f
F20110331_AADJMS he_j_Page_214.tif
299c8c03cd6e8f8a58c266cd04ac5808
d9df528b8d0e1463fabe1c81a59bca5e7b3d340d
103674 F20110331_AADKJL he_j_Page_213.jpg
15b3b7382686f786440fbef18fea8448
b53c9ef6fb391f91634c4984cc003bd31c36189b
4430 F20110331_AADICO he_j_Page_006.txt
061c6d34a8b9eff952723cbc5713d96f
3fd49bec90331972f0feabf8252b16785a67efc4
1051977 F20110331_AADIPZ he_j_Page_006.jp2
08b92bc5eab50bde3eb6bcf5ab0bc8ad
7e9ba5319f15f67aa8241f57fa009f26bc8ac156
8847 F20110331_AADLGF he_j_Page_208thm.jpg
2233c9801653f8f281ae649dd9c347cc
c9bd207ef3f7325818a88e48a9c0cb42fca7f3ed
F20110331_AADKWX he_j_Page_234.QC.jpg
7a2362c5546da94df3e1b903dab606d9
05851a8472151023f47f46195fb4a0a0eab19924
F20110331_AADJMT he_j_Page_217.tif
e4625627a1088f747e2067c1c5c4d227
367de06a3886a9a12e4ce84f399cfd0014267c0e
106003 F20110331_AADKJM he_j_Page_218.jpg
34b32dc5fce9224308ba2688cb9ae8e4
14f38c863be9151352f4e2d39877cf33641fe730
312 F20110331_AADICP he_j_Page_301.txt
0e5ae39e22153f643a0738117d267cc0
6b2232e7af177af21e08262c20624aac5459eb99
8428 F20110331_AADLGG he_j_Page_209thm.jpg
b6bc15071a1204246bc8f173fb07e4b4
e88066f0b7f9d240b1e9ea74a02ddb3dbfbe1ab8
22543 F20110331_AADKWY he_j_Page_235.QC.jpg
61016d9ae6be6fb1483d79bc31c9ff12
90b48e945bc18e81b4f146a2e2cccb549dff375e
F20110331_AADJMU he_j_Page_218.tif
3882498b4fbfa32427bedde4290a0bb0
b00004b1e865301b59d6f44829d14c2a91d4a714
110532 F20110331_AADKJN he_j_Page_220.jpg
5a8d2bac2726d70c58158280e8c30761
16e74ea66eff25aca34a6f6bdd697aca06c93b4b
116027 F20110331_AADHYA he_j_Page_239.jp2
505aee3ca6c8986499e00b7d0f45d2b3
1f27d9c77eb9dae75935eff705532f3128ace612
F20110331_AADICQ he_j_Page_216.txt
709b34b37babf77ff3dd0e64418f49a9
1c9541e818d7bd0b703a367db4138112d9d29e51
8355 F20110331_AADLGH he_j_Page_210thm.jpg
bf96bc5209aba5c76edaa08854114e12
11fc3f27e757c949dcee2199d180926e15fd389f
30876 F20110331_AADKWZ he_j_Page_082.QC.jpg
caf7bb757f750a61b12fa4670a3477ff
2866564b252a8f1bd6acbed35f041063a270d585
F20110331_AADJMV he_j_Page_219.tif
ae47c7a20fb1eabe6c4502a31ace14ed
5e4e49d64cb593c4aed64b14fe4b4122cf5b6f10
67229 F20110331_AADKJO he_j_Page_222.jpg
0b95a265c9a7287cc1f35255eb6aceaf
9b584f5d2c1a16a3fef4c3d101d5f0720562e42f
1214 F20110331_AADHYB he_j_Page_063.txt
b37f989db1223e916f2497afe2613bcd
3647b39bf951d8ba1ef157d1b567dc91eb23c2be
F20110331_AADICR he_j_Page_011.tif
58f5752bd2cf1121c9d82846f3caa139
3cea166513c400c6fccabd770798c182bfe635c4
8348 F20110331_AADLGI he_j_Page_211thm.jpg
7548b42df4591a81f6bd396d386c237b
5460d4310413f227362fe39c305ce92c0069fba1
F20110331_AADJMW he_j_Page_225.tif
45cd72a0a68ad016ba28032c69159c5d
1e0d4e6fde2512cd73e85b45a7bfcd7f45f4daed
59468 F20110331_AADKJP he_j_Page_223.jpg
d54e4a95b6c4f6c092f9cdb8e5c9db14
fd82677406de12eb474271c475d2adfc61740b32
F20110331_AADHYC he_j_Page_125.tif
9928f1d7b7b2f8c70bf1a9d5681a7443
038a8c7195a20f9f7aad6994a32f7f68b81d6ba6
26584 F20110331_AADICS he_j_Page_001.jp2
d0a8d09f3cd9dd33acff6ef70e1389f6
f9950a4973db826a56b1a469bba89c6ca98dcd66
8184 F20110331_AADLGJ he_j_Page_212thm.jpg
3b7d8562f9153c931b7569c6791bb847
2f2d56cc71db37d6c2d864024e4c34e2d9892093
F20110331_AADJMX he_j_Page_226.tif
5c00aee53f2aa9e9b401baaf4f3980ae
e74e0c98731a2fb0f59978c079f943a569954eb1
43465 F20110331_AADKJQ he_j_Page_224.jpg
7274ee938a929254a9634001c3c3ad71
b03c78c699204e807d7f380e4ac4ad96ee3a4e69
11073 F20110331_AADHYD he_j_Page_161.QC.jpg
bb5d8460229f4532b9599930a9f414d2
fabe0b79b0faa63067f16d6e1ec06419d665015c
33027 F20110331_AADICT he_j_Page_171.QC.jpg
3ac4979f12fa293fe58a0d8eaaaa89bb
21bd5a2b5f3facee32d194a4fa19618fd3853744
8392 F20110331_AADLGK he_j_Page_213thm.jpg
65a1211a116a398c551e8310d2fb487e
ddcb7eff7e22391d93a87ae6d6d0b9185a764400
F20110331_AADJMY he_j_Page_227.tif
7d18b3e0cc2d5d25283945412e1e1291
00804be21b9ec3302a6218e4fcfc50d59a8d23ff
25530 F20110331_AADKJR he_j_Page_226.jpg
0d3a1f06bb2fb90728de80c8eb732b16
e48c4af543cbf9b3f803828416d25018594f1545
18276 F20110331_AADHYE he_j_Page_314.QC.jpg
31a1cc05eea8872cf15ed9321d104f58
882ad54b5fe367558c8664d69c6564320a712dfe
F20110331_AADICU he_j_Page_228.tif
def9768ac6cdf9f24661b3d5b6f9b4f1
9194cd88542b6f46cd08d7a08eefb60b2b231a19
F20110331_AADLGL he_j_Page_217thm.jpg
d36e25b7d13283718bbec5dbc9423399
08251dab11367cb949ba44f97249fc5b83a87d63
F20110331_AADJMZ he_j_Page_229.tif
92517765ba8cac8196060e452f47ae2b
5005e0119e47cc9445407a48ac22e5adc123efbd
46250 F20110331_AADKJS he_j_Page_228.jpg
a874b2c035d2412bd8d81bc0f112999a
a523d6b478b4a0ef78380eeaa493e49efd5facdb
5549 F20110331_AADHYF he_j_Page_255thm.jpg
3a2c8690e050fddcc4ea0aed24bd3457
908b72d0bf5af4924382d1abb554cf08a46c7f97
62250 F20110331_AADICV he_j_Page_287.jpg
b93c848e637287751ed34e47d88ae659
541c757b539b69dfca1b487a7aae03eb19964275
F20110331_AADLGM he_j_Page_219thm.jpg
84580612fb59afe2dda1d9f32d537340
570afeb5a453a23272682ae00a57b0e32aca28c3
46329 F20110331_AADKJT he_j_Page_229.jpg
f6a960d9e35d41c84ad15e927fe3ea8d
fde5d7ac645777f92e5b08e62d49435f663dcdef
54522 F20110331_AADHYG he_j_Page_149.pro
9f6ae629f255c69dee376801337a673e
7cf57c3b15e796a42038e723a888deb92d53dfd0
18978 F20110331_AADICW he_j_Page_064.QC.jpg
cf8b48f03ea76f873e96b6ca36bbaa47
053a06c0d3e4413d2f6e53ab3cb0dfeaf73064c1
9073 F20110331_AADLGN he_j_Page_220thm.jpg
9e0b9b4d29f2e9027cf037086aa8fd00
64c0c15721132c9b203b64755debcff438a53703
25840 F20110331_AADKJU he_j_Page_230.jpg
45d98395c193f2c64fcc734e4615c2b4
6d82527aaccb0f4a3ab6a358cd77f15df923f02c
49721 F20110331_AADHYH he_j_Page_138.pro
8c1e32f7eb50001ec0ddfdfc780a5a93
794dbce40ec9b94d6a0e06956c7dbff24240fc81
1699 F20110331_AADICX he_j_Page_077.txt
c55756416ba9428dbf22a35aceb27815
3a5cd507bce851c715deeb4926c2e229924e3eb8
116999 F20110331_AADIVA he_j_Page_022.jp2
02327360d6733b236b8b181d5cf22af3
3c1edc2ce4e5c9eb8272dd18d80c1ded410e0fa6
F20110331_AADLGO he_j_Page_222thm.jpg
df0ed06e6d17c4a9dd1feeb7701df332
73bb560c7ca4db5c0e6a5a31b9134994761c6a9a
55749 F20110331_AADKJV he_j_Page_232.jpg
b4bfde328dbe75cad6243c4c66f1d25e
b1236df16848480c67ab15e608b1a40cd0422031
23692 F20110331_AADHYI he_j_Page_062.QC.jpg
137049199541bba0afe295c5b57bb46b
e66bbcac618b6c835ae0d79f4edb9fc4f76f4780
90985 F20110331_AADICY he_j_Page_174.jpg
7ad6c0aecae00f43b74be5d9ebcc438b
514e3cd3166d64c89b713995fcef80974cf43f60
113220 F20110331_AADIVB he_j_Page_236.jpg
938b33b27f711751fc6f50af3cd17e9a
8ec04d442da83699d5f786063c2c163c7c78ddb6
4806 F20110331_AADLGP he_j_Page_224thm.jpg
e48a7b7e9719e33de641e379efe15e9d
c3c7ddc1e9f5b24659077ebf7a15691bd24c8d87
75406 F20110331_AADKJW he_j_Page_233.jpg
658354c1871f47777d6bf8c733eb1685
5825a028e4550f449027ca0dc871411ebc318241
3968 F20110331_AADHYJ he_j_Page_038thm.jpg
291e4bc229d142b728fae8c9e7179f70
36ee00579e5e2cbd7ad089c14bcc45df6557797b
137 F20110331_AADICZ he_j_Page_297.txt
4d320b228494d2a597b858ec1f49b9eb
fe80b3b917a01406a33d58ef1022667f49bb75a4
62160 F20110331_AADIVC he_j_Page_265.jpg
f2419a3830e0f440cad9eaa51056e4b1
25e3748dd3330bca0b089ef21e0c36fdfa63214b
6163 F20110331_AADLGQ he_j_Page_227thm.jpg
b3b8ed4d81d971bfc7fd9514b31ca318
7809a81d0259dba6d01a290125c000ede19d36fe
114222 F20110331_AADKJX he_j_Page_238.jpg
608d807912b25ef22f2c42efc4ad772f
cb06877dd0b9237934b336d6c04a7e604a95475b
110442 F20110331_AADHYK he_j_Page_140.jp2
639b5dcd0052e6245ca1431c673a2809
8b59d8d9c625a681d0849c5565af867b685d3fcd
110005 F20110331_AADIVD he_j_Page_055.jp2
48c9a499b3dae38cf34e74305dac95be
6722d1ba15f4d72976da57fec3d9eec5b1f45afb
4433 F20110331_AADLGR he_j_Page_232thm.jpg
02790fbd3109307d422ca416bbcaac0f
17e216b085bd58a34925e579013183ea174010fb
103132 F20110331_AADKJY he_j_Page_240.jpg
c5c2438ede31b89fc2639d27588ce821
bed1a7550a05965444e97189c3ee97811add905f
F20110331_AADHYL he_j_Page_014.tif
44593414bd9bc3a572f5ebbaf5c5c43f
3a701bf6331901e2ae69352d502d1d5d5095f2b6
1955 F20110331_AADIVE he_j_Page_248.txt
fddc1b75325f7374eeb8a36810727bc0
de4a83e480a44844abed984256e35a564ca676e8
7098 F20110331_AADLGS he_j_Page_234thm.jpg
80acb301ef36d335f5a1457ffd918ae1
fa8bc8637eaadbc93fdc19db523117ebc260fb90
110890 F20110331_AADKJZ he_j_Page_241.jpg
6791cec50f10115e868232bb0c7a57f6
d82af18f964a9dd03c9efd9601e5cdb02201d91a
F20110331_AADHYM he_j_Page_220.tif
64f1047cd68cecbb87cbb68ba0a85501
3bd8b16a0ff5c69efc79754f336dfa382689a4e4
43774 F20110331_AADIVF he_j_Page_225.jpg
97324f3ac7989814467ed8680e1fcfe2
ba2d6c84afb0fbf82f5eb84580b606e2a9a9cb5a
5446 F20110331_AADLGT he_j_Page_235thm.jpg
67ef43a7082c1a66c85a37d8cf56ebc4
c7a399369eec22807d70177427d6e236340e8fa0
106538 F20110331_AADHYN he_j_Page_034.jpg
3f5c32a7f39508fe2950a2b4e63e9f7b
6d0f24e99bc1efda2a78daa7f51349fa66ba679b
12837 F20110331_AADIVG he_j_Page_110.QC.jpg
b6190cec8a260e3377478971007c358f
2c8d2d3b6ea9391c9bcd420365d2903b81d1502b
F20110331_AADLGU he_j_Page_236thm.jpg
e9561298024a7043e3ec8edf3684f7b0
e9178a8cfb42b955e4a1391e2fe5691e570b903a
2049 F20110331_AADJSA he_j_Page_134.txt
93ec4070faa3cd03c744059d496a260b
03842bb6d654a917bc2cb4319012cbda55a935b8
20509 F20110331_AADHYO he_j_Page_308.pro
58ef24f1b4f2bca5419afd3d90b7620c
e6fa903939ac75f8b8ec71d3b7b292553d10166d
8960 F20110331_AADIVH he_j_Page_241thm.jpg
78277539dba2041d46524ac130d59ef8
30f8ad09dd9ecb29626dfbc958500f922ed924b7
8561 F20110331_AADLGV he_j_Page_243thm.jpg
d3576d3cc47cae39bb109842379b210c
28977eb820d6b98a3ada5f2eb27c28a79e3cfaa8
F20110331_AADJSB he_j_Page_135.txt
e1a081289207b3b2b0c34e5c1c62c688
096f85828d0ce4d7e1815e064a3ed82236a2b559
7136 F20110331_AADHYP he_j_Page_231thm.jpg
69ddee19ceb988cfa8eb8f37c6852562
18b711259c7c44f52bbc04d5067e2f9c0c691368
51309 F20110331_AADIVI he_j_Page_270.jpg
b763232d707e9926f51bb82c2243b9ec
1d3fc1ced00a6161db14e15970d088f0065291b8
4580 F20110331_AADLGW he_j_Page_244thm.jpg
0335f2738c88103e1968591fcd79c04a
40870a080cf7472452f7d80d97292b506ebe331d
1850 F20110331_AADJSC he_j_Page_136.txt
72266d751f21409796c38c6d76bf87f1
91d7d7f3cc55b4b205de388226c77ea5539a689a
420480 F20110331_AADHYQ he_j_Page_128.jp2
6bc7313829f5313c59bb74342053cbcc
06fc22c5e0ca84f7e7983882d939bf74364fc2a6
31151 F20110331_AADIVJ he_j_Page_145.QC.jpg
f2c62be394e8cbd4b510eb6480f61d1b
b7ae63334707c80f7595734ccee534afefeb7804
5630 F20110331_AADLGX he_j_Page_245thm.jpg
1e25d400c33b1aa60072434d4df22e72
89b1be3e9d7031c49d4dd15955a8c2403f15a3e7
1880 F20110331_AADHYR he_j_Page_256.txt
2781ef061168e8603be1116006b54698
a3ca13567343da9edda3eade4c11bd81b8d9f006
40554 F20110331_AADIVK he_j_Page_080.pro
12b1b4b0d5bb8efb2f1aaffd8e2ba8c6
7f2248390716d74a393fcefda405278535a22d4e
F20110331_AADJSD he_j_Page_138.txt
4e796d1e5b64d6dc69c45da58114859f
6ade1debe7bb7085b5377464a5aab279e5828ac7
5471 F20110331_AADLGY he_j_Page_246thm.jpg
f8a25d0d70773c0f5bbaa53aaf97f912
00d9c65619e8e6a6ba513368b58d9eaf3a1efb7c
8965 F20110331_AADIIA he_j_Page_193thm.jpg
89f320c2c3a599b774bab1443ff82ed3
92bdde135e7d05365b4776db2aeb8c91d113b5e6
8504 F20110331_AADHYS he_j_Page_189thm.jpg
8adac5246e895f058553821f1527dfb2
70eb03d585a22cf248958b51b0f84e91d748749d
104809 F20110331_AADIVL he_j_Page_192.jp2
29e388934c1b5c0f8e747421a5992db4
7bb34acb413799d2aabde650e77c0f8d7fe4319f
1998 F20110331_AADJSE he_j_Page_139.txt
e12af68612b833b579b544e863b93b64
467f20e4a657fbcaa28e4626b4fabc42025ab818
6671 F20110331_AADLGZ he_j_Page_248thm.jpg
35ec351f706c3f827701c5c6335840d2
e589aa11192ea25fe9c6eb7ee44a77e240c32e6f
F20110331_AADIIB he_j_Page_070.tif
5e41c71ee5b6aa4c052a9ab192e9a6f8
032d1f3ceda55a2bff764db34338f7994ada6363
F20110331_AADHYT he_j_Page_044.tif
59c1f512c18ac5baa71011a74cd4420d
7c7ae46b317520bebda0cecb579193b2ba019461
F20110331_AADIVM he_j_Page_016.tif
3245c8636fce132ea627c887e37953e7
43863d761a93c467993166ef52fcb0f71de5f335
1990 F20110331_AADJSF he_j_Page_140.txt
47988d5c7dbaa7809a2f18a3410d2c9a
38c997a80591e68a001b9c956114dbc6f1e61fc8
113494 F20110331_AADIIC he_j_Page_071.jp2
8282d906f6570fa11341ee0e162d45ec
823c91a9ea734d34b98c88ae2fde892542e19660
F20110331_AADHYU he_j_Page_200.tif
e07441f16cc48459fdeb69e53c819caa
701b38a3a8526c5558153780ff4c87ef59b4235d
112088 F20110331_AADIVN he_j_Page_148.jp2
0e331876ec8daa944e6022a4804418ba
c04250bf83e7297b858ae23f2e7f11549ac06b15
1671 F20110331_AADJSG he_j_Page_141.txt
bf6ad876beb48eb6bb5d017d27c84e64
2bd0133f982362c3d8aef531425f5cee59e49666
76937 F20110331_AADKPA he_j_Page_154.jp2
ca8ed237ed8801e83378088b5208a015
a0f25dea45b11933ec2951a081e612e5b81cbd11
115083 F20110331_AADIID he_j_Page_234.jp2
c09c304823c4d9d63045595981b5506b
02d5dc87ccc862c313df8a5dcff943c0995f87aa
54801 F20110331_AADHYV he_j_Page_178.pro
894b8e9922f4b07ca10837f2f09e2de1
0f08837452ac7d120e4aea332e776043d193f1e4
F20110331_AADIVO he_j_Page_300.tif
b5852af3eba0d061bc4d982750a61b85
e0081edfcf6dce5d4f73cfbdb5389d15a36c0111
1959 F20110331_AADJSH he_j_Page_142.txt
02491581021f3801e8aafb315ca38dc2
fddb8cceeff3d1ea312f0af719a5af2f238c9069
740041 F20110331_AADKPB he_j_Page_156.jp2
762823b13275d9c05306080fbb796399
e79c708b49d7aee1e0328c01c7838956cf9b83c3
55141 F20110331_AADIIE he_j_Page_241.pro
9ba21b95bbf443fdf8eb8d93f9d44be9
69e7f72d9621a00eb43f67fdcf156f079475fd21
F20110331_AADHYW he_j_Page_272.tif
7d6fbba03dea4d6df082b07a17541417
8ffc6001b1c8f3dd65ddb5062caa21083fbb9aea
F20110331_AADIVP he_j_Page_225.QC.jpg
d36bd0624d6e221d543823c4967d789b
0a8cea03059cfde9e96289e9c3bf185151d2ce64
2094 F20110331_AADJSI he_j_Page_144.txt
6ea7809cb78183f15317154cc25f1f81
6350b446f91bead6c81ed2975630a32789738c9e
468782 F20110331_AADKPC he_j_Page_158.jp2
e7605add034d0ecc0383f4337656b99b
044fa579592f6f2add8da6c4b8102c71857b7f43
110322 F20110331_AADIIF he_j_Page_239.jpg
79d9c49a7006ce58de81e0c971e169b5
c975f73cf884856c66612201304b315f9f251c08
36953 F20110331_AADHYX he_j_Page_016.QC.jpg
84fc148dfb42508230bcff603b5f4492
c961fc287ed8aba9afb311fd65bc8936fcdef27c
20641 F20110331_AADIVQ he_j_Page_063.QC.jpg
815bf1047d2ef4d7e13f27eff882e63c
5e159ed16e32c112816eeb0af07ce4bde082b451
F20110331_AADJSJ he_j_Page_145.txt
eeba62e47f309610b954e5050cce3101
b244e7d07433a52c94d98162d25d606286a53384
544310 F20110331_AADKPD he_j_Page_159.jp2
c7c0395a71f9369f66cdeb89da0ee503
fd729c757d34c0f2fe6b3b424b240523ca81fecd
F20110331_AADIIG he_j_Page_118.tif
c3c98e9aaf7ba1c45966f54936c64e27
068fc605116dcce9846da7781728ad12c74fd8e0
1364 F20110331_AADHYY he_j_Page_278.txt
89cb4ffa9c00b7fb52d5306381c0e631
6bfefd200f81f41dff7c5bbd791c27ff25deb19b
F20110331_AADIVR he_j_Page_023.tif
ea9750fab1d84be6f8d52daf1273c8da
c7eb0cc9a49a9503c937ad466b1bd3d2ca1ccc0d
2132 F20110331_AADJSK he_j_Page_146.txt
2f7c1eee5d4c071d13be064b507ad9c9
e0c5255c9821ea48c0eb55dea4d9e588819bd827
274859 F20110331_AADKPE he_j_Page_161.jp2
4ef2a2f2357abeaa4c13eeabc2dc7d7a
be04f6440303d84bfedfe7820b9fd1249863ce43
88625 F20110331_AADIIH he_j_Page_049.jp2
03ccc6874bf79e15da2cea885c1d3a85
1487d179b68cc654fcebf2d0b6384b520387410f
60039 F20110331_AADHYZ he_j_Page_064.jpg
2db954044f250cb8a903e291e3602c27
a4fecdd770f197ed7a48af108921824af623c17c
95439 F20110331_AADJFA he_j_Page_047.jp2
4529d5aaa902271b870969a9ec0f0fef
0d2dcdd558e68d5e02caad00cb88eb964d23fca0
72163 F20110331_AADIVS he_j_Page_083.jpg
c88859f397c39848668b6a0722080bfc
b590e793947eb2cf197c97620472dd17fd552f16
2123 F20110331_AADJSL he_j_Page_147.txt
976a880dc70b20530c7d26423a0d5ad4
48f7a04aa82ab9e7624f734b265892975ece1460
73787 F20110331_AADKPF he_j_Page_163.jp2
cae777a93d195d5d47b58ec7e40fbcbf
fd7bf08ea12c99bc09740052ffe22577fd4120fd
26483 F20110331_AADIII he_j_Page_089.QC.jpg
032a8a609cc7bd80921195475fb579d7
593f7e954484b093112167e5ac5107995ac8a0c5
37487 F20110331_AADJFB he_j_Page_133.QC.jpg
b17e94a1ddd994caa2348ff3d1077707
89e54c545fb257122952f9b7b6bcac8bfc679b3f
52966 F20110331_AADIVT he_j_Page_034.pro
766f5907c505823c93b84b559fbbee0e
30c11c2baf2aa5398c99bc19cef319a2f0b55636
2086 F20110331_AADJSM he_j_Page_148.txt
03a7a33af42ef20fdc3db3f623caa6a8
f7fd712967c8319076738f6885d734f124dbd3a8
56411 F20110331_AADKPG he_j_Page_164.jp2
f7bcc65feb6e74a07383107392ea56a7
06126d5dd03b1f5571d4948e695f182a665233c3
9527 F20110331_AADJFC he_j_Page_318thm.jpg
7adf3a95af47869a56a1455e604a0101
b76a3f3e4383f8136e6c360986d09074504f1285
F20110331_AADIVU he_j_Page_161.tif
05b1e92ca77ef1cab3c9cd026caa953f
df78ea9615a1e570e62761dce1057f1d597d269f
F20110331_AADJSN he_j_Page_149.txt
10d055f8ddf4a0619d2a9d550e7e71cd
d123517c9387ec6f620e6048601a7bb94b26024e
80833 F20110331_AADKPH he_j_Page_165.jp2
82e9275b97df622f805737efedaa9f53
79d31ef0e24c5ee2957e5db8e526273ac60a6f43
33613 F20110331_AADIIJ he_j_Page_191.QC.jpg
60bbd338f900e9f031156652d4747b5e
8251f77e4e902150ebec9f0866367eae006982fa
8605 F20110331_AADJFD he_j_Page_181thm.jpg
810c0e6b91ed263c2bc457e9e6a64aed
1b14d083abdfe58f92c8324144da71ae9c719153
120243 F20110331_AADIVV he_j_Page_044.jp2
e9a21eb18d3b375d79165a7e30c8a13f
253fa5cc00911cb1d7acca8b23a0931cd2998c24
2135 F20110331_AADJSO he_j_Page_150.txt
fef0dab2ebd89b4c98f57801e6546b83
6cc0be5870116aaf1fb528ac079600062dc23506
70127 F20110331_AADKPI he_j_Page_166.jp2
b47d239115a15a644250aaf1ac26f0ee
439d6f587b208bc2ef507fb1c37975ab02fd8006
9128 F20110331_AADIIK he_j_Page_324thm.jpg
87f0551a982b29801964549d9db2705a
d472a0606725944121fe60913f2bd938c3579205
8806 F20110331_AADJFE he_j_Page_169thm.jpg
f8b33839b3ef9c378ad8f9759bed93f6
0bf30153130fd159963628d8df80b4dafa6c92d0
2172 F20110331_AADJSP he_j_Page_151.txt
44f6b1e15a3bd646c204c19fa0b233f6
4867643d22a469300a63927ed6189cd96fdf0105
61889 F20110331_AADKPJ he_j_Page_168.jp2
b5c70d5edefbaa39647afa9180affccf
143b21fa616443efc6d59ef22e4455f705d39744
50358 F20110331_AADIIL he_j_Page_140.pro
cd934a8c8fd17d89cbc2a3207962f0af
727e4d7cb86c0123cc9edb94c682b4c9833966cc
F20110331_AADIVW he_j_Page_321.jp2
ae8a2e6fed2ada5bc7975dc68b6567f1
d7d328fc8353cf615e7600628eed9b9418e1c119
F20110331_AADJSQ he_j_Page_152.txt
2e2006429653ea4dc904a8c09e607374
45210998c3298366c5e30913ed5d01b95c95ac3d
104143 F20110331_AADKPK he_j_Page_172.jp2
97c809690aa2133e5344397998c642b4
544c0ef603c1ee7f52b35d43842e8baff75baaee
93498 F20110331_AADIIM he_j_Page_136.jp2
582bfaf79a772ba58f17d91fc2137a13
23b8584faf25b31dd30fb8811bdf47bf303d6b7f
F20110331_AADJFF he_j_Page_080.tif
f82764cda83c21cb6fbf65c991cbbb11
b497096d9ecc932cb03b7586adf164a08dc6c943
135996 F20110331_AADIVX he_j_Page_322.jpg
809e5926e9643d6d5076ed7043085713
f142cdc9fcc7b619e03834eaf5b715e965ba5009
1830 F20110331_AADJSR he_j_Page_153.txt
b9213cf1d677f36eba020c9f7a40e110
4116591c13b0f46890f8248ebdd6532c6dd84734
98653 F20110331_AADKPL he_j_Page_173.jp2
9a3591761f5d817817041895005d34db
9fe7f1e4b29c3e829592f425ce97f8e2e356c9b4
2516 F20110331_AADIIN he_j_Page_235.txt
37aeabad227bb39b4738e2ad8b3674e8
d63895dc3ccbb9dc27747a36a59ef843035552e2
F20110331_AADJFG he_j_Page_098.tif
1adabd5ac57e2d63fa5f54ce2fa42786
18426101c529ff8adb0634df35a8f426818a080f
4604 F20110331_AADIVY he_j_Page_124thm.jpg
7948cb719060a0268ec8ed1fbb5eb071
0ee8517dde4279ad0fbfdecb133041beb805c40c
50939 F20110331_AADKCA he_j_Page_216.pro
4fbcb0bdfed7e708d893a65ecd74b98e
9e4c5bd7aa985b422f28ade4babd9dfd34b6657d
96510 F20110331_AADKPM he_j_Page_174.jp2
d36327adcedf35aaf5f4879ca3edbb1b
42a0d44e140be25b35442a8196848903ce1fb617
F20110331_AADIIO he_j_Page_090.tif
e65f8c782696700c152524f2bdd73932
64392d834092afd309ed9edb63ac6435e2e778b0
F20110331_AADJFH he_j_Page_207thm.jpg
e59f3012232b4f89eac37cd8cf83225d
4149a28079e45ee6ca22effc5fe7c8c4ec98b848
F20110331_AADIVZ he_j_Page_075.tif
76b17a28315edd0bd810048eff8166cd
1beef5b457e1cabb2b2492d543872b9601c66c73
1388 F20110331_AADJSS he_j_Page_154.txt
6f84ca4e804ab2a8825b0067285ac930
b7ac7f2a99fa21795edf703454fbfbc027432cf6
121379 F20110331_AADKPN he_j_Page_175.jp2
187eff45aeac8fa9601e24bed0e541f0
aaf0812d012b548a025526dccf380fca6d907151
33093 F20110331_AADIIP he_j_Page_242.QC.jpg
02719f6d476342f7ae59109889ebdd51
cbd1260b7fea2e71cca7b2cdbcbeaab347030a89
19017 F20110331_AADJFI he_j_Page_114.QC.jpg
add6e34b1d0da3ffc7d4bc35f1461aca
5e6802f934e37146de959937c8dea868ecdc36d0
54525 F20110331_AADKCB he_j_Page_220.pro
6f7025ea871f57b410719222b567151d
2d40d551973a7d2db81136ae73f1edf92bb6e439
313 F20110331_AADJST he_j_Page_156.txt
f903cf97fc413adadeb5338afa87ba75
028f537934a2dba0299f07328f48987d9f69cdad
F20110331_AADIIQ he_j_Page_087.tif
6c895d156972cd16301fe6c14a46a274
031634da85e12b5d23266233aec605c4abea182b
F20110331_AADJFJ he_j_Page_153.tif
372ec5971c7e617044528c8e5fd3d621
b9d39a996373f2d030676c76b0cd7326661c65ba
31875 F20110331_AADKCC he_j_Page_222.pro
b1ea0d02477bad29715648b2f49e972f
8587a5449390e7d418052370ae89f2adb319ca99
639 F20110331_AADJSU he_j_Page_157.txt
b2840e2801c91f01c3205dc5a3725042
9da06fe046cb30b56a5d56b230829eed69f14e8d
117336 F20110331_AADKPO he_j_Page_176.jp2
989a5f0403dacbd1b3c3a0ba43586068
5263f759d36d24b8b9de748085c9b7f693ec4f53
1909 F20110331_AADIIR he_j_Page_084.txt
d4083afc35ef10334849bae07aa3875f
6ec215b7966b78ee5c5c227461815c637c5319c9
18780 F20110331_AADJFK he_j_Page_315.QC.jpg
d2f7da31480fb2bf8f7d2dd7c1d821c3
40bd88286b1aa1ee7484a1cf59ac9146e6af3d6b
5799 F20110331_AADKCD he_j_Page_224.pro
882e6caa107513d1538468c52a4e925a
18ab7d517bbe763d84127e4f4866dd3e7b44dfd9
1009 F20110331_AADJSV he_j_Page_158.txt
543acd0abef94cfc77b29a280eb3b14a
66f79821a3e6efebeee831bfc5e99b347279d34f
109447 F20110331_AADKPP he_j_Page_180.jp2
72af082155e7ae2ec4b8f3b302771348
b9bee01b187e4b92bac878455bf606891565d2ff
101567 F20110331_AADIIS he_j_Page_079.jpg
7db3f2702fc2f084002cc2db3b0ef9ef
a5e887f887c948d6e2fe4e9f0165716be0d3cf97
F20110331_AADJFL he_j_Page_238.tif
7bdd8133e05a81ef29fd23fe7462b37d
2d97a77593391d9fe5aa854df739628541395916
7006 F20110331_AADKCE he_j_Page_225.pro
766c7d79e5c5116afd42d45b347a9fae
c89a548a97a16b95cb8427b3c0b48fd427014959
961 F20110331_AADJSW he_j_Page_159.txt
5eda92cdb045a6e85058c28af00b9854
d7b35c48ce8e3b514faa5268b864bbe45f8e5034
107914 F20110331_AADKPQ he_j_Page_181.jp2
4760fdcf1137d7072e520e3e8311fc6e
64382047c00a607640667625a36cc902ac7ac846
9462 F20110331_AADIIT he_j_Page_110.pro
27d0427abd994867cc7a8a85e92ba5bd
1af5095e899d448e3cb1fdbdc2a1dc212e294cd9
32672 F20110331_AADJFM he_j_Page_039.QC.jpg
9addccb081746dc45cc1e80b01e60620
3a281546c817ea54ebe51f3416d1934fc080b526
5466 F20110331_AADKCF he_j_Page_226.pro
bae118b542714529bbf33c795835792c
bec4d0b69f5c4d62ecf3e20ff66b581116ed876d
910 F20110331_AADJSX he_j_Page_160.txt
78ad8596a31efa438521b758db2a544a
5b1d65615688d5042a8ff0dfac111ca91b6893b2
111346 F20110331_AADKPR he_j_Page_182.jp2
fa047c3ac042c71c241fd25506f0a43d
27f950f64f5f6510d3b78ecb72cfe3add02f42ff
104358 F20110331_AADIIU he_j_Page_186.jp2
d478a504430ee4749e5e4c4227a84f06
f9f6ad559353371ff63316a2bd9c4d02bc365c67
1051985 F20110331_AADJFN he_j_Page_018.jp2
ec2d4c4fe432e32798067f4b51fa2e9b
e080bcdb1869f6399c5f1b5ff737560178f3edbb
13611 F20110331_AADKCG he_j_Page_227.pro
48ac373b09f7b669bd2cd9d88849aafa
93b48109c2d191323d01dd37c7d51cdccdfaeec2
2461 F20110331_AADJSY he_j_Page_162.txt
1ec517b0706e2fb5b2b18f577aeef2f5
558b8ed734994f41383e6cf6316d9143e980d061
115665 F20110331_AADKPS he_j_Page_183.jp2
d423bad739b689f0fea2c31158671e00
881e047c21fcc638b8575d407d8c7ec06ac9d544
8957 F20110331_AADIIV he_j_Page_179thm.jpg
7e340b3009126c7af7fe444e881ce22b
ace69536988cf75ecabb86c81b659c44965b932f
112198 F20110331_AADJFO he_j_Page_057.jpg
b8aa244f1343a1b83a412842716e14b1
dd1842c3ccae20b155b8c6d86ffc82fd7c00de56
5159 F20110331_AADKCH he_j_Page_228.pro
f546cbfbadd72fdc740bce905948d5e5
0b3b1f632b4759f111a48155ec844512fccb5122
F20110331_AADJSZ he_j_Page_163.txt
f336c4286bc792c929111a4cf424144d
86fa65e8d01b0905f2a7389343f149184b8e5c23
114404 F20110331_AADKPT he_j_Page_185.jp2
0950a69ee7687ef22b096d36237e2aeb
bcca0701d0a44ad3a8c9ff241740ee396ebca124
6846 F20110331_AADIIW he_j_Page_141thm.jpg
e03e0ddb4ab9a6d5aca41563be09a7f5
b0432605519dc1727d166aaa87c7b653c1722f46
F20110331_AADJFP he_j_Page_041.tif
fa6ae9e9a5470a564bb868283679f10b
17ad0022e274dd2d91c89dc5509497c831f251df
5076 F20110331_AADKCI he_j_Page_229.pro
12d4e843b5e60bc99b2dc31b178fc7f6
d51ed8070abb474207f235ffa985aae1d5bd0933
106409 F20110331_AADKPU he_j_Page_187.jp2
1c3b8b073aba11d8af057d344939b8eb
5f670bd991fc439d75df537055b5a93c549fce17
36657 F20110331_AADIIX he_j_Page_241.QC.jpg
1a95f8639561139a7862e3a35198d311
e694c43947c2cb2228d9dcc3a4debb0c66d08987
54275 F20110331_AADJFQ he_j_Page_185.pro
5c140a5573165c893a04b7d16fe4d97c
aef3efa54bf3ecc8132ddc45390a1492af56e507
20797 F20110331_AADKCJ he_j_Page_231.pro
6fecfa2d4c039c6884167585c1d27c30
d46158c95eefd34b640fce3d23573abbc415db8a
113796 F20110331_AADKPV he_j_Page_189.jp2
eff0da3ed48ce1355389131d194a6ec8
197d924dac9512a8861fc1f1ee621ba0603381dc
7997 F20110331_AADIIY he_j_Page_100thm.jpg
49b7312823753feaf1f5c7edd383ca3a
19f2dee9c25fcb42a813c75bb1bf1b7567f2ee4d
8536 F20110331_AADJFR he_j_Page_096thm.jpg
4a49f86fbbec889319ac5b3816c2cbce
49e5693f6dbc1c6f9e36a9b8106d6296a0fd39f6
46024 F20110331_AADKCK he_j_Page_233.pro
72e850d8662aaaf8ec590fa00959d769
f4ee50b8db94612af98bd4830a7b56b52ed51381
53383 F20110331_AADKPW he_j_Page_194.jp2
45249946621a0d69d0d461e0a9ecc0a2
0d8841fca02ce374a9af956ae5d0c741fc9df0e4
57296 F20110331_AADIIZ he_j_Page_118.jpg
a10b7ea8fc6c77d0bd4f094a92d06124
25eb1488db3ed253a3dc352afece52095e678334
1041 F20110331_AADJFS he_j_Page_328.txt
6c8187a6fd255bfad5e2c8455c14e2da
b55344e65b05eb1539774a54a4dbce33d20e440a
67537 F20110331_AADKCL he_j_Page_234.pro
d0467559a433ab64a348b163ae9aac17
52ce4b7bb57ad9a0a39b3647abd4ac340e254ef2
401352 F20110331_AADKPX he_j_Page_195.jp2
4ec6b51ff11f4879e2cc600091c7075a
b9ccd9df5fd878398992c58cdeeacf2a9eeb7569
378 F20110331_AADJFT he_j_Page_225.txt
36f43527b873c03eefe750b1e783f3d5
7e3671a0447efef14e79a550ddf5195d019158d3
47037 F20110331_AADKCM he_j_Page_235.pro
940a2cde94e9dce86f7aeb27744c1552
9dbfb9e80fe685cab1babf7aa7427fdad9ec78fd
524509 F20110331_AADKPY he_j_Page_196.jp2
4eff1c34858f4423ad4cae16d9368d1b
a8cc3a828dc1e5de7680889017d688a1f987c32a
31155 F20110331_AADJFU he_j_Page_037.jpg
17f70da512fe873cd75f42872a012f1d
63c429cbd9a7223b1aacf776621bf34637e52ac5
55963 F20110331_AADKCN he_j_Page_236.pro
267bdde4a93085384bf27ed380f40001
a47c9846457263cede37f713c598a33c2f8fe7fe
452294 F20110331_AADKPZ he_j_Page_197.jp2
7ee546e474455416b63d3ce7581d119f
ed17437be6e6571bb060947adfd6bafe38cf7c17
F20110331_AADJFV he_j_Page_077.tif
44d0977209b692cc447611104fcdf708
3fdabb47c556aa07f3242e46d428a5c7855e760a
54967 F20110331_AADKCO he_j_Page_239.pro
6183a1ed390f276662ddee38038e57a3
2eaed46b42ed5ede7daed96c08cf91e1bea2eaed
404771 F20110331_AADJFW he_j_Page_112.jp2
f43421173d09c33584ce239f2d39984d
0e40701beeb07e2c387b535e779c2fbfeaab894d
49081 F20110331_AADKCP he_j_Page_242.pro
1f2ce7a620df56be2ae7f2fd9a96a6c8
4989dcc11ab1ee6109ecac3cb3aaf414df0121f0
F20110331_AADJFX he_j_Page_020.tif
eb496637ee4f5ee93348573e4f35db0b
cf2d61c29d1dbadfb94fb951eb2c984f541519ee
53362 F20110331_AADJYA he_j_Page_051.pro
d06487d9a91207282152631d7dd2d4d0
21b15d976008fe519312dfb8af4c3554e67b16a5
48658 F20110331_AADKCQ he_j_Page_243.pro
20d48880f4bc3d5b7b6dedad9c55f331
753afc41baa6211597799d85ab3b1728c3626440
2133 F20110331_AADJFY he_j_Page_130.txt
ec773667f64b6f29a659a3697ac134fa
46459c9dd266213101eb50d9cbc248e31f412480
47069 F20110331_AADJYB he_j_Page_056.pro
0004ee940dfe38caa253eab83dc41cfb
1e67320a10e31747c2ea21901ef1b7b60aad0076
26173 F20110331_AADKCR he_j_Page_244.pro
2e5964b641882451c17c386e01a7c1ca
4b087d8d0e2e80598eac0099f894d5e7d3924e86
48434 F20110331_AADJFZ he_j_Page_263.pro
e82a8e9386a6b236cfde5777f4f24a4f
605215bd212dede4eaa855d53a2143bf72195966
55297 F20110331_AADJYC he_j_Page_057.pro
0bb21e8d15b3484fe09cdbf83a0d07d3
15cc3033df8323b30e639860b003ad0bb6d9db8e
38907 F20110331_AADKCS he_j_Page_246.pro
a8fdf0eb62e1c9d8ec898026c4f861f4
35d4cbd9c19d27371a5ae92ad741887e7ebe7c24
57431 F20110331_AADJYD he_j_Page_058.pro
8e03941cb466b18e136ca6c9d13f91b2
3f69f4ce6202aaac4272930b3d0424f85423df9d
50585 F20110331_AADKCT he_j_Page_247.pro
cc32ec9d84a41d64aa3d8b7b6cef89b0
f71d439cc9c879ad8a0c463413e42f047ec94479
107229 F20110331_AADIOA he_j_Page_102.jpg
b59717052174a2bbbb3b92768c5c9479
b1b46208d05ed6f50a6f69fa3b1d1b6fbb87d99f
49914 F20110331_AADJYE he_j_Page_059.pro
6cb4c083b61cb74938ac368c19039bf7
02263d9c0aae9e59f92b2f73c8fb4d2b95ef1cdd
39681 F20110331_AADKCU he_j_Page_250.pro
2cbf578b60c917b6b3b7c8651d17a3c0
d2b2079f7f2eba4958d05210c8a648672f180d1d
37451 F20110331_AADIOB he_j_Page_024.QC.jpg
0a5bd1cc16e1c69f444780276d098bd2
6a5e36f4115303593db8db598a866cbac0136ddd
52022 F20110331_AADJYF he_j_Page_060.pro
346d4eb6d42f0d7bcf7585bbe1e6bc58
c884b10fdcc94aa1c8beb8ab3569c97ebec978bd
29772 F20110331_AADKCV he_j_Page_251.pro
ce25093b4c61d9cdff0dcf55e900e714
045703f3521aab0617d3c5183898100bbf66b338
21883 F20110331_AADIOC he_j_Page_118.pro
940b66d1bd71a1a44bf51ad187b42fe2
26d9cb2bc2148c8b5bd8d3da3ce38f3eef1c1c8d
F20110331_AADJYG he_j_Page_063.pro
496b61ca39c317a8fb8fdc613a8d6445
4df252549c5640ade4900709c099a8cb4ffe345d
56312 F20110331_AADKCW he_j_Page_252.pro
9fbcfcb693b97fdb237e388a35360206
b271c8ae116a3c416057192a3b69ed94c21a5010
19259 F20110331_AADKVA he_j_Page_281.QC.jpg
8713f8de5143b4d341c3983bec4a97f4
a6db73d605a9597949f5b54b1b57215e59738b30
98224 F20110331_AADIOD he_j_Page_260.jp2
7548be0d8cad18dc3784fb25996f0d4d
c4f31c027ee5bde37c6aa7e2a91df8532b7b40a0
18478 F20110331_AADJYH he_j_Page_064.pro
18e2e768192a1758398a0a7af882adac
8dac05cb365a18ead99891dd1e2e5b497cb214f4
55936 F20110331_AADKCX he_j_Page_253.pro
322d1a8a764a1d0a7c5f74b31378ac56
f0ea66c585844ab94ad6e26cd8125a7b262cc4b5
F20110331_AADKVB he_j_Page_024thm.jpg
fab71ca22949f388c1d0ecd0545f7565
05cc8a540cbd44dffdcc3bbf2a4a0ab2b314d53f
93074 F20110331_AADIOE he_j_Page_125.jp2
0ba65e6262ccdd6a560462e55dea76ce
dd0fdd45d0e2cc7724a46b1a148ae2a2dce41caf
49464 F20110331_AADJYI he_j_Page_066.pro
abca6ffff9fec4da9306ea23e91483e8
58a4a36c523e7c0636e9dd6b27cc27d1e05f3593
46528 F20110331_AADKCY he_j_Page_256.pro
a3bcaf7275b44f19d922f95af9099145
e2664b24237a33facd3de83c84f23e1b198499ef
33226 F20110331_AADKVC he_j_Page_100.QC.jpg
cc0ef8bf6572492261e36360804b56ff
bccadde5dfc2d4a43a931fa4d451c2aeb3fbe92e
2915 F20110331_AADIOF he_j_Page_226thm.jpg
e9353da0f4293b3db7ac342ee47039f6
5f009f564e126df676ada0b87e1cb8ca76fd2eb6
61643 F20110331_AADJYJ he_j_Page_067.pro
310b2360c6c090a6b7485daeff96e1d8
ff4c64d3359ca80f05410c936caccbc8e5190f94
37565 F20110331_AADKCZ he_j_Page_257.pro
664f397a617c18798fc9f4ea836d19ea
16865fdd989824cfb1b2efc8b538f9bf1ed048d4
7884 F20110331_AADKVD he_j_Page_041thm.jpg
2d6d7b33de78c49285b40105851cce00
5d4483579619c8801eac3ab397a4a573a9e354d6
F20110331_AADIOG he_j_Page_289.txt
88daa7f80100ea0620137da126a5e53c
2a4f8d523e628a9f20e6a9c3d9fa790075518584
50689 F20110331_AADJYK he_j_Page_068.pro
d65b73c23ba5b823cdf558fb8f0e49d3
56cb9cb19de06f1fbe93fca2e0cdc37e061e91ae
22792 F20110331_AADKVE he_j_Page_259.QC.jpg
0150616559a87627339c4b06546de8a4
ba676b1f6d18f911a8a28ba5615dd015ca73e558
1913 F20110331_AADIOH he_j_Page_085.txt
d293616cf56b88e1e220e0c685152647
92e7c9a2eb20994025c4e8bb107f21b3eb80a4ef
F20110331_AADJLA he_j_Page_143.tif
781deb552aed433bc585e3c250caa0bd
4abe69eef05901f9b0fb7023cc8286615d7d76b5
53761 F20110331_AADJYL he_j_Page_069.pro
350e22465c3e3cf9ea63c548b1b16b87
3d5c6dae23f26db42cfd568bd7c893a524b258ce
2621 F20110331_AADKVF he_j_Page_061thm.jpg
d4a0e3d7130d2205978a1e941f6ac7a3
54bb547561e4e783d35cf446b595c9996ac5d55a
85676 F20110331_AADIOI he_j_Page_086.jpg
16282d8f76b148cc7f13fd4268069e3b
58f380c2d27c2077fc327f239a019b823966c4a3
F20110331_AADJLB he_j_Page_144.tif
406a7fe94f1c172b880645ed6231d3c2
eeff17d86564dbc90dcb191d7c13eb5228438ee4
52727 F20110331_AADJYM he_j_Page_071.pro
ae359676372a421736d7b59dddd0175f
0eda765557239678cedbfae661906ea27a87ec43
5178 F20110331_AADKVG he_j_Page_229thm.jpg
d9860b28101d787fe1c52bb515b11ec7
210f7fc7f387bb0f951bb594ec8b0986ab38e866
51765 F20110331_AADIOJ he_j_Page_156.jpg
2a220025a77ef778a39a413ca0ee77c7
d23205dd53e82632d2b783b797b957fe3d53f72d
F20110331_AADJLC he_j_Page_145.tif
14323846e88c320156eeecdfbea3f644
a3c8c82b2b397fcd0d2576d3e985358e12592dc6
44338 F20110331_AADJYN he_j_Page_072.pro
b68d4e71124e9630ed4347c1d44fd752
311d4198f996cfc81991fa0b8c71bd3a9a53c13e
16481 F20110331_AADKVH he_j_Page_124.QC.jpg
9f3a8cebd949179e0b406a13ed05e43a
541e714e266ef99ea27031062070b6e878fe0a3f
34263 F20110331_AADIOK he_j_Page_152.QC.jpg
4a6dbe16d7d973dbbdc678049777a568
69793bcc04a671396b4d7df7041f834ae7710c53
F20110331_AADJLD he_j_Page_146.tif
8cbc868f6191f347b7776b9c2e15e593
77bd21646e80d6e8201ecd2fa9375c34aad92aae
46283 F20110331_AADJYO he_j_Page_075.pro
4287aa55233689bfdfe9885009edccf7
f0ba94ad2a4341bf1e14f3cb10fa2408d6e5615f
22126 F20110331_AADKVI he_j_Page_245.QC.jpg
6e2fb0e3f74d565682494c852a91f89a
efcc106350433a12089d369056b573f02f2c7146
496 F20110331_AADIOL he_j_Page_110.txt
d953d11f1c647ec59d927eaa59647f4d
90b2ad14e388e4e622641c3e1fda6cb85c92b8f2
F20110331_AADJLE he_j_Page_147.tif
ea17f8d82eb7ea5af5f4678e0add3de8
db1e88a995d34402d0956e265568f24681a0dc00
101635 F20110331_AADIBA he_j_Page_216.jpg
e590f80589f8ce94284a734df23316d3
c6d88ceca7862b155c9ed04a058e6a827b125ada
51734 F20110331_AADJYP he_j_Page_076.pro
0400433cc7b8cc2ac71144e05789158e
280cfd64d5b16711028b58ec48e24b7271305d02
8866 F20110331_AADKVJ he_j_Page_147thm.jpg
29fc24413c111ac87d0dd05871fe21f6
75c4d1799828d6af4d4b480a655f30501b4ce81c
F20110331_AADIOM he_j_Page_079.tif
244ad57fa9b11f5a60d5c7ef42876233
08f4bd3d34c370d6ac71b38313503dcffb7515b1
F20110331_AADJLF he_j_Page_148.tif
8a650410039cdb412e4d09254b54b2d2
1d3028670b7d908a2c34267e43770c1a3f6039c0
2863 F20110331_AADIBB he_j_Page_011.txt
42e8e4abf69c2d5aa7bf3008734de142
e85f514e583937578aa13821b350e940a9600a8f
39268 F20110331_AADJYQ he_j_Page_077.pro
39f78298f96f00c7dea3ab803bff223f
ec07d4ed449a50cdeacdef94f003c9135d6a4bc8
37359 F20110331_AADKVK he_j_Page_011.QC.jpg
4e0bad53572661d0bcf26db0b01bf05b
05f2da39929b69aa9b3dc16379527bd2ee58f025
F20110331_AADJLG he_j_Page_152.tif
36ba47a367e8697ece34b970ee2f1d00
0e6e7c8cd2835a67cf6601073d29b89331255692
48742 F20110331_AADJYR he_j_Page_078.pro
10cbd8f213c6d3d428be7cded6ea92ef
dd8e7c17b9cb39d37e880b673e61d88089a2fbf8
9282 F20110331_AADION he_j_Page_319thm.jpg
82e1f151d75295c2eb4eef5a1d531c6d
bc6e07fd03bd6f75e7e724d42db29e04f297f507
7337 F20110331_AADKVL he_j_Page_083thm.jpg
0919221253f05d6daab283b16d8ed864
0728cafb6e895f212c9796be557cd2d66936260a
F20110331_AADJLH he_j_Page_155.tif
f6888d5cd145e27c6fdcda7a39cf36be
4ca7ab7ca5c6facbd5aa63464ceafd2861cd5a9e
112014 F20110331_AADKIA he_j_Page_151.jpg
603e6f3cd3fd6eda0bfde0e41d6ab13d
9adb06172c629e94bea4b49158b81712040afc8d
39305 F20110331_AADIBC he_j_Page_318.QC.jpg
5693ae50a38795ac3abe08d488a8c819
afea48a77924d7ed88da2dd243bd64e1fb09a167
49685 F20110331_AADJYS he_j_Page_079.pro
e3f548a09963a53f87b1054a96dd5f85
cdb95cf039b6110e361f83b4998a58197863efd7
25431 F20110331_AADIOO he_j_Page_077.QC.jpg
373333cb8b64d9d1583dd1a27be95b24
93c7318f1a333209af150def26796b6f2d9856a6
34068 F20110331_AADKVM he_j_Page_140.QC.jpg
da0dd0575f8739f365eb0bf808ada582
2fef84ac0bd0f1568d23a81b999dc911f458754c
F20110331_AADJLI he_j_Page_156.tif
d7f2813968afc9f27575cfceec9e2e5c
0fb8fe89cddf335d001b50eeba157de5345cdbf5
90867 F20110331_AADKIB he_j_Page_153.jpg
c92f6a8d1dd9a27b2e100ea22e969e21
b93eac3ce8f2fd54a5ad72a0090b147547458d9b
106039 F20110331_AADIBD he_j_Page_190.jpg
3e0a698f580954c14491ee6247758996
d127816b6d44ad25046e345d56ca2650d2adaa5b
44139 F20110331_AADJYT he_j_Page_082.pro
d2752d2ad3c69f5e73c1c62e679096b1
cf5c93834470d71ef505562d4ae21b38e103a4d8
8635 F20110331_AADKVN he_j_Page_230.QC.jpg
382571c79d76fb8d8d0118c873a35b7d
3010089557e219214780bf30eebff6f96d0baeba
F20110331_AADJLJ he_j_Page_159.tif
6b226c9b9e90ca3e069b60f9fbd21482
a868836647d06f4fcd2108a46d82b6c0beeeea7e
40137 F20110331_AADKIC he_j_Page_155.jpg
1c4840a9f65bec1421399f73ef28aeb0
76f9730fb20421ebc101bb743367825f36bd7833
9074 F20110331_AADIBE he_j_Page_060thm.jpg
89e9d7bd955e462e831f32b0b368c11c
e486e0f107f076e114b8f303023a1fd9e2272bb1
38004 F20110331_AADJYU he_j_Page_084.pro
3e1aac691bdcea0124bcb02ef9ffb5f3
598f5da77be4ad0e8f953e11bab8b5dca0cc1453
54491 F20110331_AADIOP he_j_Page_024.pro
0d6e9e9113cd8bf9c3030f09cb8ed986
bb33e0892e829d9881c1b8a2471b5c405cd5db43
40006 F20110331_AADKVO he_j_Page_326.QC.jpg
1d1606e74336d07767c704640272973e
b0019de5a44392f510de44f4f0005ada747244c2
F20110331_AADJLK he_j_Page_160.tif
ba3935e77a0bd4bae7a8a942f2598ba3
1f1000c64f303376d45fc556b10b946c58ee0c44
47113 F20110331_AADKID he_j_Page_157.jpg
53474bb4e8c15ef8b64af4a0ecd668d5
978adc271acbced2a650641dc840dc768a7ca64a
2114 F20110331_AADIBF he_j_Page_131.txt
e57cbf3d47cf06f069995db50b5cda2c
66e50348719b6c5fbb40452d996c8cbc3a65a8aa
43810 F20110331_AADJYV he_j_Page_085.pro
7e424edf916c194a9aa6d3d5f9210d4a
4d7a0f6bf649cbe1d1c71e11f3fb55c16304604a
85987 F20110331_AADIOQ he_j_Page_263.jp2
eafc270d27c6845cf35bc5cca48d33a4
1fd6e58965fc9e02d018ca7b103ecb00916de885
5716 F20110331_AADKVP he_j_Page_307thm.jpg
9ad7061ac3b03ba1d55341d46f81d41e
4afc4877a828a6b04647911c5af3a7a21dbb7b55
54849 F20110331_AADKIE he_j_Page_158.jpg
77c9a2911d9d7373400eb505a705bbbb
49e490c3508d1d4d846145404fe07955fce99288
41587 F20110331_AADIBG he_j_Page_255.pro
98a80bc812a19f6081d791f2167de1c9
ffb2c81bfd12c0776c58374fca2bd5067cfeda7b
42994 F20110331_AADJYW he_j_Page_086.pro
09a503147117273b45d27c2af38ae619
2a727c063c610c68bbc89393ef718a246cf261d6
113308 F20110331_AADIOR he_j_Page_069.jp2
8fe868a06fa8c2f592c7d40cbdd27d51
fdb210990f41c67003ddddf881d306a0d6ecddc9
7017 F20110331_AADKVQ he_j_Page_019thm.jpg
6b3d0c15e6c6fe76ce5a5b3d6847655f
e71907dfa8a6e918125671e11289fec849a13994
F20110331_AADJLL he_j_Page_163.tif
702693ba7087f1ecfd3bc8e8177efdd4
4a806a734418ff0dbf358bd48f5ded3bc84928a3
33355 F20110331_AADKIF he_j_Page_161.jpg
415451177359ef92358996a4b159c8c0
a2b85b247eaf8eec9d13deb064a401528748c689
F20110331_AADIBH he_j_Page_095.tif
964ea271a46b9772cf84d2193897ae6d
e82e35d5376d1742efe6600b8ca7ebe92232fee3
41357 F20110331_AADJYX he_j_Page_087.pro
90758a7335b6822cb08937f737ce032a
6e4076c409c988e637932d6d623a373c4be398f7
2313 F20110331_AADIOS he_j_Page_200.txt
9146664fb79a2a9455a7ced7ab9a2d44
4fd2b3fdf241ead442bff179a39111eec9283820
8656 F20110331_AADKVR he_j_Page_015thm.jpg
9a3c6e89f3fe8479efb2088e3c3d7a58
2c447bd29a2ac7fd9b5e4d41d78db30755ef90f7
F20110331_AADJLM he_j_Page_164.tif
33a2c42aa8d6a08ca7aa3309b389ad3b
105e75a7ff7819504d96a936b00acfae0f2a5786
73130 F20110331_AADKIG he_j_Page_163.jpg
f75cb027e483f882603d4a7886750008
3f28244fe6d95517fa9e00a78dc35a979864bf19
8288 F20110331_AADIBI he_j_Page_152thm.jpg
19877c26e20630a9a87a5e4ff8365087
c259f66ab945807181f51c5bf14b89f1e37457e1
55824 F20110331_AADIOT he_j_Page_237.pro
7de66a54e99a8e2a8d67c2f2073a5e41
aa7797c85348544e165ac83ac4af6488f791da74
F20110331_AADLFA he_j_Page_134thm.jpg
e7fae5444450c1a8a69f22a2b6d136da
5807fd2c54332cc22ad209d895637b0e1a6870ec
4952 F20110331_AADKVS he_j_Page_168thm.jpg
51be192267ffa0eadf3d1c2189e7ce9f
57acc54e891116b43ba554bf3442220cc10c1d30
F20110331_AADJLN he_j_Page_165.tif
784bfdbd8040ad0bb5c94c842ab33f65
6b8c0006acc80b207e24b5dfc07cb43c50a661df
99097 F20110331_AADIBJ he_j_Page_252.jpg
b081e9ceffeb4f7ef427a8e5286d3c79
e9acd7dc403a7d8a75a191ca5c1226778f866eba
43209 F20110331_AADJYY he_j_Page_088.pro
4fb16d5e8d48032c4f27cf2bfd448edf
187d9f183af43b86f91571d11ce48febf7e46e50
128161 F20110331_AADIOU he_j_Page_030.jpg
5e10fb48738cc2d10bb88d3bfe01a35d
9214b8b7a303288e98a47ab2adb84f111404ae77
8407 F20110331_AADLFB he_j_Page_135thm.jpg
26936f6ee177265e026f5887e41678d0
1decb0509b3665de6bd769e89fe3d19908839408
8360 F20110331_AADKVT he_j_Page_184thm.jpg
e686e299c89d455b7d5dfd5f5078118b
6355cb2fff9887d3f4ae07380af26a5453eeb61c
F20110331_AADJLO he_j_Page_166.tif
9e555d1fa6d73b025b961fc41fdf2bc4
0ec25377279112c8f72496418dd1303bdac5e124
54709 F20110331_AADKIH he_j_Page_164.jpg
a09e126838610b5cf7fb53c00a82d918
cdf61f75d9091c8d733a8b5e4c1ccd65e2f79d7f
6434103 F20110331_AADIBK he_j.pdf
fe79d4f457e04407cd8835decc47b3bf
157f3eea5063d120e4d5ad4264c90132a6933a87
38262 F20110331_AADJYZ he_j_Page_089.pro
119ee3b277c2821a3db5da15d2e71bc2
5f6352cc5321ea311cdd0286393f9027694aa92d
F20110331_AADIOV he_j_Page_054.tif
358ad4566cf599f20aaeed11bb114ff0
5deb2ea2bbfd3a66164571d983020dd7813e1b40
8514 F20110331_AADLFC he_j_Page_138thm.jpg
3cce3189fc4f6e7686af3b2924278d62
c491806678e6b345350963f4ae5045f474088bb8
F20110331_AADJLP he_j_Page_167.tif
92776117849f57f88e67335f89d7417f
ccc9a316c3fd4bd2c57a25c734a8418d5629426b
75832 F20110331_AADKII he_j_Page_165.jpg
2696b92249092cb5e1bae025d77d6a5f
031cf364aafb000deefafceb630767d3f96f61f1
F20110331_AADIBL he_j_Page_215.tif
45f9365a74636d317d8143fc19529ddd
1f308352c405d3629f307f05680f2eb5ecd69a24
F20110331_AADIOW he_j_Page_231.tif
28e1020c2e9c525c9475e878059fb995
3aa3502e7a37b92cb7cba173caddef4176650506
15056 F20110331_AADKVU he_j_Page_157.QC.jpg
b5b49ebfe4c534a3fe46ef1cad41a94c
0df2957bd0b70576709c8de4a0905e5881020b63
F20110331_AADJLQ he_j_Page_168.tif
13f1d38f3aecbe416171f048d5761db6
459bdd415220ebbe1df489c554c4133734663821
80287 F20110331_AADKIJ he_j_Page_167.jpg
56f5b59df2828543b0a013b3f52e1656
c9d14acf61b748a102f9f0121a7043d3608e62ba
F20110331_AADIBM he_j_Page_154.tif
b7b8193942dd95ff0835e690100981c9
d51161d3b665d94b591712173d5fbf3dd8419956
7890 F20110331_AADIOX he_j_Page_065.QC.jpg
1e66c0fb88ed2d9b20fb376e00f00b28
58a4863447719741caf7f196a3b3e634e30cc172
8543 F20110331_AADLFD he_j_Page_139thm.jpg
15407e73d97c263908f66b160b1a7655
769d4ecdc0c04b56684e075cee6f5d55b0c53f06
9045 F20110331_AADKVV he_j_Page_132thm.jpg
880c01e4e97676ea4b029cc687e00bd6
eb6974223324d614ca585f96dbcdaccda84aa478
F20110331_AADJLR he_j_Page_169.tif
daae6d2ccd231cbf108d83b06995de6a
47c3cd8913b9cd300c409bbeeff194adf19439c2
106158 F20110331_AADKIK he_j_Page_169.jpg
c08f0cef7076d0b72905b638dc67b91c
b01c3d5d47c1a9e9056cbbaf748df952df48f8ce
F20110331_AADIBN he_j_Page_088.tif
0785a22212bbfaf2979fbff0c42cd9ce
6497823b4da047489ebee9901144c7d17420b970
53547 F20110331_AADIOY he_j_Page_081.pro
7454c17216726ca94aed146155e95014
bd00cfe4e6dbbdc8f5d58f8058e03623812558ca
8289 F20110331_AADLFE he_j_Page_143thm.jpg
0832f1d8440beb03bf231b39a42c6ee3
212d98359fdb768f47ab10a917fe9a800284df02
6766 F20110331_AADKVW he_j_Page_084thm.jpg
c2c4b454d82de7654a698f9377e2bf52
cd0e5a2c110b3e428dc93b716c52cd595d55dfb0
F20110331_AADJLS he_j_Page_172.tif
246250784ad87b9a6e113009a4358e0d
f14bbb4348f50159a1a02539c4e45089ad394498
109140 F20110331_AADKIL he_j_Page_170.jpg
bb969e8ce3ee5a3aeb26f5b12cb45a1a
7d821ae1f50200032a4c9ecba232c949a3e0631d
58454 F20110331_AADIBO he_j_Page_264.jpg
b6ffb67e0646058565fcb2a5da378a68
3364c1f032beb5ff4dc77749a8bca53349d940ad
43349 F20110331_AADIOZ he_j_Page_090.pro
e25776e3fd500a19b8ab5f1f8736d3ca
af57a4168c3d1f41d6a2e5744d95da31a833ae0c
8684 F20110331_AADLFF he_j_Page_144thm.jpg
860825c77c31935f756de2528ccf2ce5
e5eb32941b1be11676efc3cc6da9672647213828
F20110331_AADKVX he_j_Page_071thm.jpg
5f1dc704e49999901f4f7a25b4b2389a
05bfdead9c43eaa2f7a1f6402fae8a1486719c85
F20110331_AADJLT he_j_Page_175.tif
e1b7f3614b4b471f25fe48d57a236080
ec70d8a5d4309b8b2e44791bc9e8d3392d0b0deb
101862 F20110331_AADKIM he_j_Page_171.jpg
0b1c34faf8877e2ef3b36766465e5105
90575f9f0f309415302272c807b05e7de546f613
278 F20110331_AADIBP he_j_Page_224.txt
4db2978771fb7623d0e4297ea0d625d4
e6651fb698d8c8bcc4e4f22d4d01ed544db3d8a8
8167 F20110331_AADLFG he_j_Page_145thm.jpg
0b4464e61148c5aa99fc2943b5cf7a9f
eacf1bb80d3a631a8e0a3830b4da64db25e4686f
15823 F20110331_AADKVY he_j_Page_267.QC.jpg
fd0ec08d8c50ac27ff61bcb051049c61
df531fa96822f05a5e2834fefca247d40fa0495c
F20110331_AADJLU he_j_Page_176.tif
7aa94458f105afbf55ef962dc08f28bb
8d6778f6cf05b2ddebcfdd17dd4f6deeb7ba1005
100616 F20110331_AADKIN he_j_Page_172.jpg
4d1a034d28e78ab786cc878b27c4a6f2
1cd4c133531d36f536a76add0c4a92ce1321dd2d
2158 F20110331_AADIBQ he_j_Page_033.txt
17a1f83d2ca0c8728b72d64575688e8a
72978dc875ead9860e6a2a184820c4199da6bca1
9106 F20110331_AADLFH he_j_Page_149thm.jpg
433c04a2c572b807d40ea7d921ad9e1a
384121d5804e87c1b21e98f16904b92d96397b48
8197 F20110331_AADKVZ he_j_Page_187thm.jpg
ae32866153a6410b42ae0cb86d986420
5286e3c7e6cfa0fa2d8b8fefbe1526eac50ef073
F20110331_AADJLV he_j_Page_177.tif
108f7d7dec594ed29c5c65bc6aa271a9
7fa710d1571eb4a276cecbab7d96b41d0bc4d211
92868 F20110331_AADKIO he_j_Page_173.jpg
e1ff19c03052cf7939f1a607ce3536f0
6fed1b64b8a919c7f24efe387490ab18d1851e47
2145 F20110331_AADIBR he_j_Page_183.txt
9deaf1d89bc2229654d240ee6e3c1517
d308b210a6b26abc77099dffc45eb6df4e64f4d3
8791 F20110331_AADLFI he_j_Page_150thm.jpg
716ada85300c514de9275c5cc14246f3
f298ad4609967c6b0b6cb2edce96a426e04e4a03
F20110331_AADJLW he_j_Page_178.tif
6271b2ad89b348e9d07a7da5834667d9
41c5d17c2b9473b9833089cf75a165a90f25e39c
111430 F20110331_AADKIP he_j_Page_176.jpg
a3beb3158df52b5ae809d8d65ceaa97b
2412ad9aea1f5b4c82f4883601ab36da94ca78c0
105186 F20110331_AADIBS he_j_Page_096.jpg
3abe06c47b9c6e162cb3b766a83eee53
54e8d34c06c651425d0b1b8652042435f658a4e9
5980 F20110331_AADLFJ he_j_Page_154thm.jpg
f526eb7fb9e5a10e0ecc57b4d331bd77
e4c0feb2f231431bdac9125d7bf9a0abf0a0c313
F20110331_AADJLX he_j_Page_179.tif
78cd04a7f5742aedc985e59836a2b8da
36e0e0947ef2bb0bce30b46f29826ed789a50927
114296 F20110331_AADKIQ he_j_Page_177.jpg
c330e16fd34ff3b8ca2b501d21af5d5f
198f2212d883778d26748a0b2aa0de6094a14582
F20110331_AADIBT he_j_Page_107.tif
b04408aae976993a4c318d6e8672bb7b
bc160ed8f422f2ddc4e12d133c876eb55a826a93
3479 F20110331_AADLFK he_j_Page_155thm.jpg
ad95cc6d9b9c9773dd137106add9e499
26c00fc562a7edaa7191e254bb673962399dbc53
F20110331_AADJLY he_j_Page_180.tif
a54aee304f17b344164249eb27a1b032
86a5fb0502968bf0f27d16742139112ed50676d5
110583 F20110331_AADKIR he_j_Page_179.jpg
3e8881b191dd5b80ec5617c9c081d0ec
b79cb250dbf42369f462e36b392828a98ed86906
35868 F20110331_AADIBU he_j_Page_060.QC.jpg
c0eca4fb7f6572e40f8627a4dff1dcc9
359ecff1976b07aefb798446985ea13d990cb09d
4376 F20110331_AADLFL he_j_Page_157thm.jpg
9399acdba6ab9ab29f77f2aa210a3daa
bac6c729555a6b3f3d5e5126052e60faa934d4ae
F20110331_AADJLZ he_j_Page_182.tif
6a2f10c5e4e0bd2a44f33d159c852174
dc4bdfac7fd5247fa2f1edeca2c261261e9ed477
103898 F20110331_AADKIS he_j_Page_180.jpg
8f1a1828cb80ab5c323bf45f7fdda635
d98c83f4c8f0cb4ad58e264d8344792539abda22
35282 F20110331_AADIBV he_j_Page_219.QC.jpg
166f5a00f1b11d35af3300fa4d0ddbe3
f0ed88b61bfabb0dd91386bb1a632b3808fd8cc8
6586 F20110331_AADLFM he_j_Page_159thm.jpg
a57c41256dec45724f8a918fa63b8113
31e634ef5b2e0996c7d99d4027704674cd1e9a7d
104080 F20110331_AADKIT he_j_Page_182.jpg
4f057a5ebd707fb7d68fd1cca91a943c
500db5659a36e76e0bab6e1aabc606e7630a257e
F20110331_AADHXG he_j_Page_311.tif
192a2a05b6abc93f2949e27f714e29ef
72c6575b44afc640c5ecb15ec52632765e9be540
1051984 F20110331_AADIBW he_j_Page_318.jp2
bf4669a2b645247cc4a985c9d6647b6c
2b4104824b2effd06cdc6c58576082440e8044dd







BEST MANAGEMENT PRACTICE DEVELOPMENT WITH THE CERES-MAIZE MODEL
FOR SWEET CORN PRODUCTION IN NORTH FLORIDA




















By

JIANQIANG HE


A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
DOCTOR OF PHILOSOPHY

UNIVERSITY OF FLORIDA

2008
































2008 Jianqiang He


































To my parents, brother, and the teachers who taught me at different stages in my life









ACKNOWLEDGMENTS

I believe my most important achievement in the past four years was learning how to do

academic research and how to cooperate with other people. I recognized that this crucial

improvement in my life is owed to the contributions of many people.

First and foremost, I am greatly indebted to my supervisor, Dr. Michael D. Dukes, for his

insightful guidance, continuous encouragement, and unselfish support in my research over the

past more than four years. His thoughtful coaching with all aspects of my research was a

guarantee of the success of this endeavor. His enthusiasm and preciseness have left an

everlasting impression on me. I will never forget how he never even neglected a tiny typo when

he was revising my writing. Without his help, it would not have been possible for me to

complete this research.

I would like to express sincere appreciation to Dr. Wendy Graham and Dr. James Jones for

their insightful and invaluable advice on resolving all kinds of technical problems in my research.

I will never forget their sacrifices, spending their valuable time to meet with me many times.

Their enthusiasm for research and pursuit for excellence will be an example to me forever.

Without their combined supervision of each step throughout the model simulation, this study

would not have been possible.

I am grateful to Dr. Jasmeet Judge for her help in the field experiment and advice in model

simulation, to Dr. George Hochmuth for his advice in sweet corn BMP development, and to Dr.

Michael Annable for his advice in nitrogen movement. I deeply benefited from their suggestions

and advice on a multitude of perspectives regarding my research.

I would like to thank Dawn Lucas and Dr. Donald Graetz for their help in laboratory

experiments. Dawn helped me analyze thousands of soil and water samples. I am thankful they









let me occupy their laboratory for my experiment at least three weeks in every summer from

2004 to 2006.

Special thanks go to Danny Burch, staff of the Agricultural and Biological Engineering

(ABE) Department, and people in the Plant Science Research and Education Unit, University of

Florida. Without their help, the burdensome field work would not have been completed. I would

also like to thank the undergraduate students (Steve, Clay, Frank, Habtu, Sanjiv, Patrick, Thi,

David, etc.). They helped me conduct the field and laboratory experiments. Now they are my

good friends. I am also grateful to several officemates for their friendship and encouragement,

and faculty, staff and students in the ABE Department for the harmonious academic environment.

Finally I particularly appreciate my parents and younger brother for their unconditional

love, understanding, patience and encouragement. I love and miss them.









TABLE OF CONTENTS

page

ACKNOWLEDGMENT S 4....___ ... ... ...............
A C K N O W L E D G M E N T S ............................................................................................................... 4

L IST O F T A B L E S .......... ......... ............................................. ................................. 11

LIST OF FIGURES .................................. .. .... ..... ................. 14

ABSTRAC T ................................................... ............... 19

1 INTRODUCTION ............... .......................................................... 21

1.1 Study Background .............................. ....... ....... .. ......... .. ......21
1.1.1 N itrate Pollution in N north Florida ........................................ ....... ............... 21
1.1.2 Sw eet C orn Production in Florida..................................................... ................... 23
1.1.3 Total Maximum Daily Loads and Best Management Practice..............................24
1.1.4 Best Management Practices for Sweet Corn Production.......................................27
1.1.5 Best Management Practice Development...........................................................31
1 .2 O b j e c tiv e s ................................................................................................................... 3 5
1.3 D issertation Outline ............... ................. ........... ................. .. ...... 35

2 GLOBAL SENSITIVITY ANALYSIS OF CERES-MAIZE MODEL WITH ONE-AT-
A -TIM E M ETTH OD .................. ............ ........ ..................................... .. 39

2 .1 Introdu action ...............................39...........................
2 .1.1 Sen sitivity A naly sis ............................................................. ........ ........ 39
2.1.2. L ocal Sensitivity A analysis .................................. ............... ............... 41
2.1.3 Global Sensitivity Analysis ............................................................................42
2 .2 M materials and M methods .......................................................................... .....................43
2.2.1. M odel Description .................. .......... .............. .. .... ................. 43
2.2.1.1 CERES-M aize m odel ............................................................................43
2.2.1.2 Soil w ater sub-m odule............................................ ........... ............... 44
2.2.1.3 Soil nitrogen sub-m odule ........................................ ......... ............... 46
2.2.2 N on-restricted OAT M ethod ............................................................................ 47
2.2.3 Normalization of Input Parameters ............................................................ 48
2.2.4 R restricted O A T M ethod ........................................ ..................................... 49
2.2.5 One-at-a-time (OAT) Method for CERES-Maize Model ....................................51
2.2.6 Field Experim ent .................................... .. .. ..... .. ............54
2.3 R results and D discussion .............. .................................... .... .................55
2.3.1 Non-restricted OAT Results .............. .......... ............... 55
2.3.1.1 Response profiles ...................... ...... .... ....... ...............55
2.3.1.2 Correlation coefficient m atrix ................................................... ................56
2.3.1.3 Influential parameter selection based on non-restricted OAT method ........57
2.3.2 Influential Parameter Selection Based on Restricted OAT Method .....................58
2 .4 Sum m ary and conclu sions ....................................................................... ................... 59



6









3 PARAMETER ESTIMATION FOR CERES-MAIZE MODEL WITH THE GLUE
M E T H O D ......................................................................................................................... 6 8

3 .1 In tro d u c tio n ................................................................................................................. 6 8
3.1.1 P aram eter E stim action ................................................................... ..... ................68
3.1.2 GLU E M ethod ............... ................. ........... ............... .. ...... 71
3.2 M ethod and M materials ........................................................................................ 73
3.2.1 Field Experim ent .................................... .. .. ..... .. ............73
3.2.2 M ain P procedure of G L U E ........................................................... .....................77
3.2.3 Selection of Input Param eters..................................................................... ..... 77
3.2.4 Prior D distribution ............................. ............................ ...... ........ ......78
3.2.5 Model Run with Generated Parameter Vectors...............................79
3.2.6 Determination of Number of Model Runs.................................................81
3.2.7 Likelihood Function and Likelihood Value ................................. ............... 81
3.2.7.1 A available likelihood functions ................................ .. ... ............... .... 81
3.2.7.2 Selection of likelihood function and method of likelihood value
com bination ..................... .................... ............... ............ 88
3.2.7.3 Comparison of distributions of input parameters...................................91
3.2.7.4 Comparison of distributions of outputs................................ ... ..................91
3.2.8 Estim action of Posterior Distribution............................... ....................... ....... 92
3.2.9 GLU E Sim ulation .................. ............... ................. ................... 93
3.2.10 GLUE Verification ............................. ............ ............. ............... 93
3.2.11 Expected Values of Posterior Distribution.....................................................94
3.3 R results and D discussion ........................ .................... ... ........... ......... 95
3.3.1 Results of Prior D istribution.................................................... ...................95
3.3.2 Results of Number of M odel Runs................................................ .... .................96
3.3.3 Results of Likelihood Function and Method of Likelihood Value
C o m b in atio n ................. ........... .. .......... ... .............................................. 9 7
3.3.3.1 Comparison of distributions of input parameters............... ...................97
3.3.3.2 Comparison of distributions of model outputs................ ..................99
3.3.4 Distributions of Selected Param eters......................................... ............... 101
3.3.5 PD F Plot of Selected Param eters................................................ ............... 102
3.3.6 Distributions of Outputs ........................... .. ... ........................... 103
3.3.7 Joint Distribution between Yield and Nitrogen Leaching................................104
3.3.8 G L U E V verification ............................................................................ ............. 104
3.3.9 Result of Expected Values of Posterior Distribution ........ ...........................106
3 .4 C o n clu sio n s....................................................... ................ 10 6

4 FIELD EXPERIMENT OF SWEET CORN AND SIMULATION WITH
CALIBRATED CERES-MAIZE MODEL .............................................................. 129

4.1 Introduction ......................................................................................................... ..... 129
4 .2 M material and M methods .......................................................................... ......... ........... 13 1
4.2.1 E xperim ent Site and D esign ..................................................................... ...... 13 1
4.2.2 N itrogen Fertilizer A application ........................................ ........................ 133
4.2.3 Irrigation Scheduling ................................... ................................................... 136



7









4.2.4 Soil, Biomass, and Yield Sampling ............................................. ...............138
4.2.5 CERES-M aize M odel Simulation ............................................. ............... 140
4.3 R results and D discussion ........................... ......... .............................. ................... 142
4.3.1 Quantity of Sweet Corn Yield ................. .................................... 142
4.3.2 Quality of Sweet Corn Yield ............. ............ ............... 143
4.3.3 Nitrogen Balance Estimation ............ ............. ............... 145
4 .3.3.1 N nitrogen input ........................... ...... .............................. .. .... .......... 145
4.3.3.2 N itrogen output ...................... ................ .................. ........ 147
4 .3 .3 .3 N itrog en b alan ce ........................................................................................ 14 8
4.3.4 Comparison between Model Simulations and Field Observations.....................149
4.3.4.1 Comparison between dry matter yields................................................149
4.3.4.2 Comparison between phenology dates................................................... 150
4.3.4.3 Comparison between potential nitrogen leaching .................................... 150
4 .4 C o n clu sio n s....................................................... ................ 15 3

5 BEST MANAGEMENT PRACTICE DEVELOPMENT WITH CERES-MAIZE
MODEL FOR SWEET CORN PRODUCTION IN NORTH FLORIDA........................... 169

5 .1 In tro d u ctio n ............................................................................................... ............. 16 9
5.2 M materials and M methodology ............................................................................ ............171
5.2.1 Experim ent Site ......................... ................................. 171
5.2.2 C rop M odel C alibration............................................... ............................. 172
5.2.3 B M P Sim ulations......................................................................................... 173
5.2.4 Determination of Acceptable Yield.......... ............... ...............179
5.3 Results and Discussion ......... .................. ................ .. .............. 181
5.3.1 Effects of Irrigation .............. ...... ................ .. ........ .. ... ....181
5.3.2 Effects of Nitrogen Fertilizer............................................ 184
5.3.2.1 Total nitrogen fertilizer am ount ...................................... ............... 184
5.3.2.2 Nitrogen fertilizer split......... ......................................... ....... ..... .......... 186
5.3.2.3 Amount of nitrogen fertilizer in each application.............. ...............186
5.3.3 Selection of Potential B M Ps............................................... ............... ....188
5.3.4 Evaluation and Implementation of Potential BMPs ..........................................189
5.4 Sum m ary and C onclu sions ...................... .. .. ............. .............................................. 19 1

6 UNCERTAINTY ANALYSIS OF POTENTIAL SWEET CORN BMPS UNDER
WEATHER AND INPUT PARAMETER VARIABILITY .........................................208

6.1 Introduction ............. ..............................................208
6.2 M materials and M methods ...................... ......... ........................................ ....................2 11
6.2.1 Field Experiment and Weather Data .......................................... 211
6.2.2 Uncertainty of Input Parameters.....................................................................212
6.2.3 Selected Potential BM Ps .............................................................. ............... 213
6.2.4 A Grower Practice of N Fertilizer and Irrigation Management .........................213
6.2 .5 M onte C arlo Sim ulation ............................................................ .....................2 15
6.3 R results and D iscu ssion ......................................................................... ....................2 16
6.3.1 B M P C om prison .............. ........................................................................216
6.3.2 O utput U uncertainty Plot................................................ ............................ 218


8









6.3.3 Output Uncertainty over Time Range of 1958-1990.............................220
6.4 Sum m ary and C conclusions ..................................................................... ..................22 1

7 CONCLUSIONS AND FUTURE WORK.................... .. .................. .............. 236

7.1 Sum m ary and Research Contributions ........................................ ....................... 236
7 .2 C o n c lu sio n s........................... .. ..................................... ..... .... .... ...... ...............2 3 7
7.2.1 Global Sensitivity Analysis of CERES-Maize Model with One-at-a-time
(O A T ) M ethod ................. ................ ............ ......................... ................ 237
7.2.2 Parameter Estimation for CERES-Maize Model with GLUE Method ..............238
7.2.3 Field Plot Experiment of Sweet Corn and Simulation with Calibrated CERES-
M aize M odel .............................................................. ..........................239
7.2.4 Best Management Practices Development with CERES-Maize Model for
Sweet Corn Production in N orth Florida.......................................... ......... ......240
7.2.5 Uncertainty Analysis of Potential Sweet Corn BMPs under Weather and Input
P aram eter V ariability ......... .. ............ ........................................................242
7 .3 F utu re W ork ...........................................................................24 3

A P P E N D IX ........................ ........ ......... .............. ................................... 245

A INPUT AND OUTPUT PARAMETERS OF CERES-MAIZE MODEL IN DSSAT.........245

B MATLAB CODE FOR GLOBAL SENSITIVITY ANALYSIS WITH THE
RESTRICTED OAT M ETHOD .......................................................... ............... 246

B. 1 M ain Function ........................................................... .... ................ 246
B.2 Sensitivity Analysis of Genotype Parameter ...................................... ............... 246
B .3 G enotype F ile C hange.......................................................................... ....................248
B.4 Genotype Param eter Space ..................................... .. ......................................249
B.5 Processing Sensitivity Analysis Results of Genotype Parameter ............................... 251
B.6 Sensitivity Analysis of Soil Parameter................................. ...............252
B .7 Soil File Change ...... ............. ... ....... ................. .............. 255
B .8 Soil Param eter Space ............................................ ........ ............ ........ ......259
B.9 Processing Sensitivity Analysis Results of Soil Parameter .......................................261

C MATLAB CODE FOR GLUE PROCESS................................................................ 262

C 1 M ain F u n action ..................................................................................... ....................2 62
C.2 Generation of Random Numbers ............................. ......... ..............................263
C.3 Function "m vnrnd" ............. ....................................................... .. ....264
C.4 Parameter Setup for Genotype and Soil...................................................................... 265
C .5 Change of Soil File ............................................ ............. ........... 266
C .6 C change of G enotype F ile ....................................................................... ..................268
C .7 Sum m ary O utput Processing ............................................................... .....................269
C .8 Plant N itrogen O utput Processing...................................................................... ...... 270
C.9 Soil Nitrogen Output Processing.................... ......... ........................... 273
C 10 Param eter PD F Plot.......... .... ........... ................. .............. ...... ...............275
C. 11 3-D Plot of Joint Distribution of Yield and Nitrogen Leaching .............................288









D PICTURES OF FIELD EXPERIM EN T.................................................... .....................293

E SAS CODE FOR ANOVA OF YIELD QUANTITY AND QUALITY..............................303

F NITRATE AND AMMONIUM CONCENTRATIONS IN MONITORING WELLS IN
BLOCK 1 IN THE PLANT SCIENCE RESEARCH AND EDUCATION UNIT
U N IV E R SITY O F FL O R ID A ................................................................... .....................305

G TOTAL KJELDAHL NITROGEN CONCENTRATION OF LEAVES AND STEMS
OF SWEET CORN IN FIELD EXPERIMENT IN PLOTS IN 2006 ...............................306

H NITRATE AND AMMONIUM NITROGEN CONCENTRATION OF SOIL IN FIELD
EXPERIMENT OF SWEET CORN IN PLOTS IN 2006............... .................3.08

L IST O F R E F E R E N C E S ..................................................................................... .................. 16

B IO G R A PH IC A L SK E T C H ............................................................................. ....................329









LIST OF TABLES


Table page

1-1 Sweet corn harvested for sale in Florida in 2002 and 1997 (USDA-NASS, 1998,
2 0 0 2 ) ........................................................................................ . 3 8

1-2 Nitrogen fertilizer application for sweet corn in Florida (USDA-NASS, 1993, 1995,
1999b, 2003, 2006) .........................................................................38

2-1 Genotype coefficient for the DSSAT CERES-Maize model .................. ................65

2-2 Covariance coefficient matrix of genotype and soil parameters of the DSSAT model .....66

2-3 Criteria for input parameter determinationa ............................................ ................... 67

2-4 Selected parameters for GLUE simulation based on the non-restricted OAT method
and covariance coefficient m atrix ............................................ ............................. 67

2-5 Mean and variance of absolute elementary effects of genotype parameters ...................67

2-6 Mean and variance of absolute elementary effects of soil parameters ...........................67

2-7 Selected parameters for model calibration based on the restricted OAT method..............67

3-1 Average soil physical properties of the experiment site (from 24 sampling locations)... 122

3-2 Selected parameters for GLUE method due to sensitivity analysis of predicted dry
matter yield and accumulative nitrogen leaching (See Chapter 2 for details)................122

3-3 Covariance matrix of the prior distribution ........................................... ...............122

3-4 Results of Jarque-Bera test of the input parameters b .................... ..........................122

3-5 Mean values and standard deviations (STDEV) of first-round posterior distributions
derived from different likelihood functions and likelihood combinations .....................123

3-6 Mean values and standard deviations (STDEV) of model outputs derived from first-
round posterior distributions .................................................................. .................124

3-7 Fundamental statistical properties of prior, first posterior and second posterior
distributions derived from L 1C2............................................. ............................. 125

3-8 Measured and estimated mean values of soil properties of the field experiment site......125

3-9 Selected parameter set for GLUE verification............................................................126

3-10 Generated duplicates of observations for GLUE verification............... .... ..............126









3-11 Means and standard deviations of the selected parameters in GLUE verificationa .........127

3-12 Means and standard deviations of model outputs in GLUE verification ......................128

3-13 Expectation values of second posterior distribution of selected parameters .................. 128

4-1 Soil properties of the experim ent site .......... ................................................ .......... 162

4-2 DUiq values of 4 different numbers of drip tapes at 3 depths at t=30min..................... 162

4-3 Fertigation schedules of field plot experiment in 2006 .........................................162

4-4 Crop coefficients of sweet corn at different stages of development.............................162

4-5 Second posterior distribution of the selected parameters ...............................................163

4-6 Measured and estimated mean values of soil properties of the field experiment site......163

4-7 ANOVA results of total yield of sweet corn ..................................... ...............163

4-8 Irrigation and nitrogen treatment effects on yield quantity ..........................................164

4-9 ANOVA results of total ears of sweet corn ............................................................ 164

4-10 Irrigation and nitrogen treatment effects on yield quality ............................................165

4-11 Nitrogen budget of a replicate of treatment F111 in Block 1 of the plot experiment ......165

4-12 Estimated nitrogen leaching of seven treatment in field plot experiment .....................166

4-13 ANOVA results of nitrogen leaching estimated from N balance .................................. 166

4-14 Irrigation and nitrogen treatment effects on cumulative nitrogen leaching estimated
from N balance..................................... ................................. .......... 166

4-15 Simulated and measured dry yields in field plot experiment in 2006.............................167

4-16 Simulated and measured anthesis and maturity dates in field plot experiment .............167

4-17 Nitrogen balance of model simulation of treatment F1 .............................................167

4-18 Simulated potential nitrogen leaching of the seven treatment in field plot experiment ..168

4-19 Simulated and estimated accumulative nitrogen leaching in field plot experiment ........168

5-1 Expectation values of second posterior distribution of selected parametersa ..................200

5-2 Soil properties of the experim ent site ......... ...................................... .......... ........200

5-3 Calculation of total available soil water (ASW) in the soil profile..............................200









5-4 Irrigation treatments based on different MAD values ............................................. 200

5-5 N itrogen splits used in BM P simulation ........................................ ....... ............... 201

5-6 Nitrogen splits used in single factor simulation.................................... ..................201

5-7 Acreage, yield, production, and value of Florida sweet corn 1998-2006
(U SD A N A SS, 2007)........... .................................................................. ................... 202

5-8 Fresh yields of selected white sweet corn varieties in Clanton Ala. 1995-1996
(Sim onne et al. 1999) ............... ................. ...................... ......... 202

5-9 Fresh yields of sweet corn experiment in Springfield Tenn. 1993-1995 (Mullins et al.,
1 9 9 9 ) ................... ........................................................... ................ 2 0 2

5-10 Fresh yields of sweet corn experiment in Eden Valley and Freeville, NY, 1998-2001
(Rangaraj an et al., 2002) ............... ................. ........... ............ ......... 203

5-11 Fresh yields of sweet corn experiment in Belle Glade, Florida, in spring of 2001
(S h u ler, 2 0 0 2 ) ......................................................................... 2 0 4

5-12 Summary of sweet corn yield in field experiments conducted in Florida (Hochmuth
and Cordasco, 2000) ................................. ............... .. ............204

5-13 Selected irrigation strategies ........................................................................ 205

5-14 Ranking of dry yield (HWAH) and nitrogen leaching (NLCM) under different N
fertilizer application splits........................................................................ ..................205

5-15 Selected factors of N fertilizer application strategies .................................................205

5-16 Ranking of average nitrogen leaching (NLCM) of combination management over 33
years (1958-1990) ............................... ..................................................206

5-17 Selected potential BM Ps for sweet corn production............................................ 207

6-1 Second posterior distribution of the selected parameters (from Chapter 3) ....................232

6-2 Six selected potential BMPs for sweet corn production (from Chapter 5)................232

6-3 N fertilizer management in the "EPA319 Project". ......................................................233

6-4 Irrigation management in the "EPA319 Project".......................................................234

6-5 Mean and standard deviation (STDEV) of simulated corn dry yield and nitrogen
leaching both under different uncertainty scenarios a..................................................... 235









LIST OF FIGURES


Figure pe

1-1 D iagram of research structure............................................................................. ...... 37

2-1 Scheme of non-restricted OAT method. ........... ...................................... .................62

2-2 Schem e of restricted OAT method .............................................................................. 62

2-3 Response profiles of sweet corn yield to six normalized genotype parameters ................63

2-4 Response profiles of sweet corn yield to nine normalized soil parameters .....................63

2-5 Response profiles for the nitrogen leaching to six normalized genotype parameters........64

2-6 Response profiles for the nitrogen leaching to nine normalized soil parameters .............64

3-1 Diagram of Block 1 of field experiment ..... ...................... ...............109

3-2 Influence of number of model runs on mean values of P1 ............................................. 109

3-3 Influence of number of model runs on standard deviations of P1 .......................... .. 110

3-4 Influence of number of model runs on mean values of SLRO ................................. 110

3-5 Influence of number of model runs on standard deviations of SLRO ..........................111

3-6 Influence of number of model runs on mean values of simulated dry yields..................11

3-7 Influence of number of model runs on standard deviations of simulated dry yields....... 112

3-8 Influence of number of model runs on mean values of simulated nitrogen leaching...... 112

3-9 Influence of number of model runs on standard deviations of simulated nitrogen
le a c h in g ................... ......................................................... ................ 1 1 3

3-10 Parametre PI: probability distribution ..... ...................... ...............113

3-11 Parametre P5: probability distribution................ ........ .................... ... ............ 114

3-12 Parametre PHINT: probability distribution ............................ .....................114

3-13 Parametre SLDR: probability distribution.............. .... .................. ............... 115

3-14 Parametre SLRO: probability distribution.............. .... .................. ............... 115

3-15 Parametre SLLL: probability distribution............................ ............... 116









3-16 Parametre SDUL: probability distribution....... ...............................116

3-17 Parametre SSAT: probability distribution ..... ....................................117

3-18 Parametre SLPF: probability distribution............ ....................... ...................117

3-19 Histogram of predicted dry matter yields ................................ ...............118

3-20 Histogram of predicted anthesis dates ......................... ......................118

3-21 Histogram of predicted maturity dates.......................... .......................119

3-22 Histogram of predicted cumulative nitrogen leaching............................................... 119

3-23 Joint distribution between yield and nitrogen leaching under prior distribution of
input param eters .............. .... ......... .............. .............................120

3-24 Joint distribution between yield and nitrogen leaching under the first posterior
distribution of input param eters ....................... ......... ........................ ............... 120

3-25 Joint distribution between yield and nitrogen leaching under the second posterior
distribution of input param eters ....................... ......... ........................ ............... 121

4-1 Experiment plot arrangement layout..... .. ............................................................ 155

4-2 Soil moisture at t=30 minutes with 1, 2, 3 and 4 drip tapes......................................156

4-3 Drip tape arrangement in each row interval................................... ...............157

4-4 Drip tape arrangement and sampling zone in each plot............................................157

4-5 Fresh yield under different N fertilizer levels under II. .......................................... 158

4-6 Yield under different N fertilizer levels under 12 .................................... ..................... 158

4-7 Number of ears per unit area under different N fertilizer levels under II ....................159

4-8 Number of ears per unit area under different N fertilizer levels under 12 ..................... 159

4-9 Number of ears per unit area under different irrigation levels under F ......................160

4-10 Number of ears per unit area under different irrigation levels under F2 .......................160

4-11 Number of ears per unit area under different irrigation levels under F3 ....................... 161

5-1 Response curves of yield to different remaining ASW ...........................................195

5-2 Response curves of nitrogen leaching to different remaining ASW.............................195









5-3 Response curves of yield to different irrigation depths .............. .... ...............196

5-4 Response curves of nitrogen leaching to different irrigation depths ............................ 196

5-5 Rainfall and accumulated irrigations in East Half of Blockl in 2006 ...........................197

5-6 Response curves of yield to different N fertilizer levels.............................................. 197

5-7 Response curves of nitrogen leaching to different N fertilizer levels............................198

5-8 Dry yield vs. different N fertilizer application amount.................................................198

5-9 Nitrogen leaching vs. different N fertilizer application amount.................................... 199

6-1 Histogram and cumulative distribution of predicted average annual dry yield of the
six selected potential BMPs and the actual grower practice both under weather and
input param eter uncertainty.. .............................. ... ........................................ 223

6-2 Histogram and cumulative distribution of predicted average annual nitrogen leaching
(NLCM) of the six selected potential BMPs and the actual grower practice both
under weather and input parameter uncertainty.......................................................227

6-3 Simulated 10% and 90% confidence limits of average annual yields of BMP1 both
under weather and input parameter uncertainty.......................................................231

6-4 Simulated 10% and 90% confidence limits of average annual nitrogen leaching of
BMP1 both under weather and input parameter uncertainty ............ .................231

D-1 Components of nitrogen fertilizer solution........................................ ...............293

D -2 F ertigation control table .......................................................................... ................... 293

D -3 Fertigation system installation ........................................................................ 294

D-4 Main fertigation lines, injection holes, peristaltic pump, and solution bucket................294

D -5 Sub-m ain fertigation lines........................................................................ .. ................295

D-6 Drip tapes and sub-m ain fertigation line.................................... ........................ 295

D -7 D rip tape distribution in one row ................................................................................. 296

D-8 Irrigation with the linear move irrigation system ................................. ............... 296

D -9 Sw eet corn planting .......... ........................................................................ ....... ............... 297

D -10 Sw eet corn em ergence .......................................................................... .....................297

D-11 Comparison between no-nirogen-applied plot (near) and nitrogen-applied plot (far)....298









D -12 Sw eet corn tasseling .......... ...................................................................... ........ .. ....... .. 298

D -13 Sw eet corn m maturity .............................................................................. ..................... 299

D -14 Sw eet corn harvest ................................. ............. .............. .... ...... .. 299

D -15 P lant sam pling .......... ............................................................................ .......... ......300

D -16 Soil sam pling ...............................................................300

D -17 Y field sam pling ........................ .................................... .. ........ ............. 301

D -18 Y field w weighing ........... .................................................................................. ... ....... .. 30 1

D -19 Y field grading ...............................................................302

D -20 R research partner ............... .............................. ............. ............ ............ 302

F-l Average nitrate concentration in the monitoring wells on the west part and east part
o f B lo ck 1 .......... ............................................................................ 3 0 5

F-2 Average ammonium concentration in the monitoring wells on the west part and east
part of B lock 1 ...........................................................................305

G-1 Average total Kjeldahl nitrogen (TKN) concentration of leaves of sweet corn under
irrigation lev el II ......................................................................... 306

G-2 Average total Kjeldahl nitrogen (TKN) concentration of leaves of sweet corn under
irrigation level I2 ..........................................................................306

G-3 Average total Kjeldahl nitrogen (TKN) concentration of stems of sweet corn under
irrigation lev el I ......................................................................... 307

G-4 Average total Kjeldahl nitrogen (TKN) concentration of stems of sweet corn under
irrigation level I2 ..........................................................................307

H-1 Average nitrate nitrogen concentration of soil at layer 1 (0-15 cm) under irrigation
lev e l I 1 ......................................................................................... 3 0 8

H-2 Average nitrate nitrogen concentration of soil at layer 1 (0-15 cm) under irrigation
lev el 12 ........................................................ .................................. 3 0 8

H-3 Average nitrate nitrogen concentration of soil at layer 2 (15-30 cm) under irrigation
lev e l I 1 ......................................................................................... 3 0 9

H-4 Average nitrate nitrogen concentration of soil at layer 2 (15-30 cm) under irrigation
lev el 12 ......................................................... ................................. 3 0 9




17









H-5 Average nitrate nitrogen concentration of soil at layer 3 (30-60 cm) under irrigation
lev e l I 1 .............. ........................................................................... 3 1 0

H-6 Average nitrate nitrogen concentration of soil at layer 3 (30-60 cm) under irrigation
lev el 12 ........................................................................................ 3 10

H-7 Average nitrate nitrogen concentration of soil at layer 4 (60-90 cm) under irrigation
le v e l I 1 ................... ........................................................................ 3 1 1

H-8 Average nitrate nitrogen concentration of soil at layer 4 (60-90 cm) under irrigation
lev e l 12 ......... ......................................................................................... . 1 1

H-9 Average ammonium nitrogen concentration of soil at layer 1 (0-15 cm) under
irrigation lev el II ......................................................................... 3 12

H-10 Average ammonium nitrogen concentration of soil at layer 1 (0-15 cm) under
irrigation lev el I2 .........................................................................3 12

H-11 Average ammonium nitrogen concentration of soil at layer 2 (15-30 cm) under
irrigation lev el II ......................................................................... 3 13

H-12 Average ammonium nitrogen concentration of soil at layer 2 (15-30 cm) under
irrigation lev el I2 .........................................................................3 13

H-13 Average ammonium nitrogen concentration of soil at layer 3 (30-60 cm) under
irrigation lev el II ......................................................................... 3 14

H-14 Average ammonium nitrogen concentration of soil at layer 3 (30-60 cm) under
irrigation lev el I2 .........................................................................3 14

H-15 Average ammonium nitrogen concentration of soil at layer 4 (60-90 cm) under
irrigation lev el II ......................................................................... 3 15

H-16 Average ammonium nitrogen concentration of soil at layer 4 (60-90 cm) under
irrigation lev el I2 .........................................................................3 15









Abstract of Dissertation Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Degree of Doctor of Philosophy


BEST MANAGEMENT PRACTICE DEVELOPMENT WITH THE CERES-MAIZE MODEL
FOR SWEET CORN PRODUCTION IN NORTH FLORIDA

By

Jianqiang He

August 2008

Chair: Michael D. Dukes
Major: Agricultural and Biological Engineering

Increasing nitrogen loads within the Suwannee River Basin of North Florida has become a

major concern. Nitrogen fertilizer application in field crop production is proved to be the most

import nitrogen contribution in this region. Florida ranks highest in the nation in the production

and value of fresh market sweet corn. Thus it is necessary to develop research based nitrogen

best management practices (N-BMPs) to reduce nitrogen leaching while keeping an acceptable

yield in sweet corn production.

This study is an attempt to utilize the CERES-Maize mode of the Decision Support System

for Agrotechnology Transfer (DSSAT) model as a platform to develop potential BMPs for sweet

corn production in North Florida.

The results show that the non-restricted and restricted one-at-a-time (OAT) method can be

used to conduct global sensitivity analysis for the CERES-Maize so as to select the most

influential parameters for model calibration. The generalized likelihood uncertainty estimation

(GLUE) method was proved to be a powerful tool for model parameter estimation, since the

uncertainties in model input parameters were significantly reduced after GLUE was used to









estimate the model input parameters. The uncertainties in model outputs were reduced

correspondingly.

The comparison between the model simulated and field observed results of the seven

treatments in a field plot experiment of sweet corn in 2006, shows that the model did a good job

in predicting dry yield and phenology dates.

The results of BMP development with the calibrated CERES-Maize model show that if the

growers could apply both irrigation water and nitrogen fertilizer more frequently but with

smaller amounts in each application, this would result in an acceptable yield and a lower level of

nitrogen leaching. The results showed a total nitrogen amount between 196 and 224 kg N ha-1

would be enough for sweet corn production in North Florida, which confirmed that the

recommendation nitrogen amount (224 kg N ha-1) by Institute of Food and Agricultural Sciences

(IFAS) Univerisity of Florida, was reasonable.

The results of uncertainty analysis of the CERES-Maize model for sweet corn simulation

show that the weather was the dominant uncertainty contributor. This was because after two

rounds of GLUE parameter estimation procedure, the uncertainties existing in input parameters

were minimized.









CHAPTER 1
INTRODUCTION

1.1 Study Background

1.1.1 Nitrate Pollution in North Florida

Increasing nitrogen loads within the Suwannee River Basin of North Florida has recently

become a major concern. According to the "Surface Water Quality and Biological Annual Report

2003" (Suwannee River Water Management District, 2004), in 2003, 4,165 metric tons of

nitrate-nitrogen and 1,733 metric tons of phosphorus were transported to the Gulf of Mexico by

the Aucilla, Econfina, Fenholloway, Suwannee, and Waccasassa Rivers. The Suwannee River

Basin alone accounted for 4,069 metric tons of nitrate nitrogen and 1,476 metric tons of total

phosphorus.

In 1995, a study was conducted to determine how springs and other ground-water inflow

affect the quantity and quality of water in the Suwannee River (Pittman et al., 1997). They

studied a 53-km stretch of the Suwannee River from Dowling Park, Fla., to Branford, Fla. Water

samples for nitrate concentrations (dissolved nitrite plus nitrate as nitrogen) and discharge data

were collected at 11 springs and 3 river sites during the 3-day period in July 1995 during base

flow in the river. They found that nitrate (NO3-1) loads increased downstream from 2,300 to

6,000 kg day-1, an increase of 160% in the study reach, and that 54% of nitrate load increase was

supplied by the groundwater inflow. Eighty-nine percent of the nitrate load increase occurred in

the lower two-thirds of the stretch. Ham and Hatzell (1996) found that nitrate concentration in

Suwannee River increased at a rate of 0.02 mg L-1 per year over a twenty-year period from 1971

to 1991, with an average concentration for the 20-year period of 0.5 mg L1.

Leaching of nitrate nitrogen is economically and environmentally undesirable (Katyal et al.,

1985; Poss and Saragoni, 1992; Theocharopoulos et al., 1993). Nitrate that leaches below the









crop root zone represents the loss of a valuable plant nutrient, and hence, increases agricultural

costs. If nitrate enters groundwater supplies, it can also impose risks to both human health and

the environment. Consumption by humans and animals through drinking water with high nitrate

levels has been associated with several health problems. The most serious is methemoglobinemia

or blue baby syndrome (02 deficiency in blood) in infants. High nitrate concentrations in

drinking water are detrimental to the health of infants especially during the first 6 months of life.

Additionally, groundwater with high nitrate levels that discharge into sensitive surface waters

can contribute to long-term eutrophication of these water bodies (Asadi, et al., 2002). For this

reason, the US Environmental Protection Agency (EPA) has set a maximum contaminant level

requiring the nitrate-nitrogen concentration not exceed 10 mg N L-1 and the nitrite-nitrogen

concentration not exceed 1 mg N L-1 in public water supplies (U.S. Dept. Health, Education, and

Welfare, 1962).

Nitrates leached into the groundwater of Suwannee River Basin are believed to come from

several sources including animal wastes, chemical fertilizer, industrial, and domestic sewage.

The Middle Suwannee River Basin, which includes Lafayette and Suwannee counties, has

hundreds of residential and commercial septic systems in rural areas, about 300 row crop and

vegetable farms, 44 dairies with more than 25,000 animals and 150 poultry operations with more

than 38 million birds. Suwannee County is the leading poultry production area in Florid (Woods,

2005). According to the report by Katz et al. (1999), in Suwannee County, the relative

contribution of N from fertilizers increased from about 23% in 1955 to more than 60% in 1980.

During 1955-1995, the contribution of estimated N inputs from animal wastes (poultry, dairy and

beef cows, and swine) ranged from about 21 to 42% of the total estimated N inputs. It is obvious

that N fertilizer is the most important N contributor in this region.









1.1.2 Sweet Corn Production in Florida

Florida ranks highest in the nation in the production and value of fresh market sweet corn

(Zea mays L.), typically accounting for approximately 25% of both national sweet corn

production and of U.S. cash receipts for fresh sales (FASS, 2002; USDA-NASS, 1997, 1999a).

Sweet corn has typically ranked as one of Florida's five most valuable vegetable crops.

During the 2000-01 production seasons, sweet corn was the second ranked vegetable crop in

terms of acreage and fifth ranked in total value. Harvested acreage for sweet corn represented

14.9% of the state's total vegetable acreage during that season, while production value

represented 8% of the total production value of all Florida vegetables (FASS, 2002). Average

yield ranged from approximately 8,200 kg ha-1 fresh sweet corn yield in 1969-1970, to 16,400 kg

ha1 in 2000-2001 (FASS 2001, 2002).

The principal fresh sweet corn production region in Florida is the Everglades area (Palm

Beach County), which during the 1999-2000 season produced 63% of the state crop. The

southeastern/southwestern area (Miami-Dade, Collier, and Hendry Counties) were responsible

for 25% of the state's production. The west/north area (Suwannee and Jackson Counties)

accounted for about 7% of the sweet corn production. Sweet corn was also grown in the central

area around Lake Apopka, but this region only produced about 5% of the crop since the muck

soils in this area have been taken out of production (FASS, 2001).

Table 1-1 shows the harvested acreage of sweet corn in Florida in 1997 and 2002 (USDA-

NASS, 1998, 2002). It can be seen that in 1997 there were 413 sweet corn producing farms in

Florida, with a total planted area of 17,791 ha. In 2002, number of farms decreased to 340, while

the total planted area decreased tol5,768 ha. In both years, the large farms, especially the ones

that were greater than 200 ha consisted of the most part of total area.









Table 1-2 shows the applications of chemical nitrogen fertilizer in Florida from 1992 to

2006 (USDA-NASS, 1993, 1995, 1999b, 2003, 2006). In 1998 and 2002, all sweet corn acreage

in Florida received nitrogen applications totaling 1.83 and 2.61million kg, respectively. Between

1992 and 2006, from 81 to 100 % of sweet corn acreage in Florida received an average of 2.0 to

10.0 applications of nitrogen seasonally. An average range of 46 to 62 kg N ha-1 had been used at

each application, with a statewide annual total N application ranging from 1.64 to 5.48 million

kg. It should be noticed that in 2006 though only 86% of total planted area received nitrogen

fertilizer, the total number ofN applications increased dramatically to 10 times a season.

Consequently, the N rate per crop year in 2006 increased to 475 kg ha-1, which was almost 3

times as that of 2002. The total applied chemical nitrogen fertilizer to sweet corn also doubled

from 2002 to 2006.

Adequate water is especially important in sweet corn production during periods of silking

and tasseling and of ear development (Hochmuth et al., 1996). Most of Florida's sweet corn is

grown under irrigation. In 1997, 53% of farms and 71% of sweet corn acreage was irrigated

(USDA-NASS, 1998). About 92% of sweet corn growers in Florida surveyed in 1993 reported

that they checked soil moisture and plant need to determine irrigation needs, while 8% used an

established schedule modified to meet plant needs. Furthermore, only 8% were using a

mechanical system to monitor soil moisture, and of those not using a mechanical system, 30%

considered it too expensive, 30% reported not knowing of a good and inexpensive system, 30%

cited limited water supply, and 10% said that lack of time prevented them from adopting a

mechanical system (Larson et al., 1999).

1.1.3 Total Maximum Daily Loads and Best Management Practice

In 1972, Congress passed the Clean Water Act (CWA) which set forth federal

requirements for identification of polluted or impaired water bodies. These rules were passed









down to the states by the U.S. Environmental Protection Agency (EPA), which requires states to

establish a prioritized list of impaired water bodies and to develop estimated loads that the water

bodies could receive of each pollutant while meeting water quality standards (DeBusk, 2001).

These estimated loads determined for each water body are called Total Maximum Daily

Loads (TMDLs). TMDLs are defined as the maximum amount of a pollutant that a water body

can receive and still meet the water quality standards as established by the 1972 Clean Water Act.

Section 303(d) of the act requires states to submit lists of surface waters that do not meet

applicable water quality standards and to establish TMDLs for these waters on a prioritized

schedule.

In response to state TMDL requirements, the Florida Watershed Restoration Act (FWRA)

was passed in 1999. This act established the Florida Department of Environmental Protection

(FDEP) as the lead agency in coordinating the implementation of the TMDL allocation through

water quality protection programs. These programs include non-regulatory and incentive-based

programs, including best management practices (BMPs), cost sharing, waste minimization,

pollution prevention, and public education. This act also required the Florida Department of

Agriculture and Consumer Services (FDACS) to develop and adopt rules pursuant to suitable

interim measures, best management practices, or other measures necessary to achieve the level of

pollution reduction established by the FDEP for agricultural pollutant sources. These practices

and measures may be implemented by those parties responsible for agricultural pollutant sources

and the department, the water management districts, and the FDACS shall assist with

implementation (Florida Statutes, s.403.067, 1999).

The FDEP also should develop TMDL calculations for each water body or water body

segment according to the priority ranking and schedule unless the impairment of such waters is









due solely to activities other than point and non-point sources of pollution. When a water body is

identified as impaired and a TMDL is established, pollutant loads are divided among the

different stakeholders (agriculture and urban). Hence, the TMDL shall include establishment of

reasonable and equitable allocations of the total maximum daily load between or among point

and non-point sources that will alone, or in conjunction with other management and restoration

activities, and achieve water quality standards for the pollutant causing impairment. The

allocations may establish the maximum amount of the water pollutant that may be discharged or

released into the water body or water body segment in combination with other discharges or

releases. Allocations may also be made to individual basins and sources or as a whole to all

basins and sources or categories of sources of inflow to the water body or water body segments

(Florida Statutes, s.403.067, 1999).

Normally, each stakeholder would implement a set of management practices that are

expected to reduce its contribution to meet its designated load. These practices are commonly

referred to as BMPs and can be defined as a practice or combination of practices determined by

the coordinating agencies, based on research, field-testing, and expert review, to be the most

effective and practicable on-location means, including economical and technological

considerations, for improving water quality in agricultural and urban discharges. Although some

water bodies do not have designated TMDLs as of yet, and therefore do not legally require

BMPs, many agricultural BMP manuals are being developed. FDEP, FDACS, and the Institute

of Food and Agricultural Science (IFAS) at the University of Florida have partnered with local

agencies and stakeholders to develop BMP manuals (Migliaccio and Boman, 2006).

The primary benefit for growers implementing agricultural BMPs (even without a

designated TMDL) is that if a BMP program is in place, an agricultural producer is considered to









be operating under a presumption of compliance with water quality standards. This protects the

farmer from liabilities to the state when water quality standards are not met (IFAS-UF, 2006).

According to the "Water Quality/Quantity Best Management Practices for Florida

Vegetable and Agronomic Crops" (FDACS, 2005), all farming operations using this BMP

manual shall reasonably attempt to implement the recommended BMPs in order to establish a

baseline set of BMPs to ensure a reduction in pollutant loading to impaired receiving waters.

Depending on the farm's site specific conditions, all of these baseline BMPs need not be

implemented. Only BMPs applicable for a particular location and production system should be

implemented. This Tier-1 or first level of BMP protection also includes many of the practices

that are identified as "essential" under USDA-NRCS conservation planning procedures.

Irrigation scheduling and optimum fertilizer management are two of the proposed set of

minimum BMPs that are suggested to be implemented.

1.1.4 Best Management Practices for Sweet Corn Production

In this research, focus was on the most common cultural practices that directly affect the N

cycle, N fertilization and irrigation. Fertilization is the cultural practice that can directly

influence the N cycle in the root zone of sweet corn. Fertilization affects not only plant uptake,

but also mineralization, nitrification, denitrification, and ammonia volatilization (Cockx and

Simonne, 2003). Mineralization will not be significant in sandy soil due to the low organic

matter content, but will be significant in organic soils. However, approximately 50% of total N-

fertilizer applied can be taken up by the crop (Bundy and Andradki, 2005), i.e. about 50% of the

total applied N-fertilizer would be lost by leaching, volatilization, denitrification, etc.

Irrigation is another important factor. Florida is among the wettest states in the U.S. with

most areas receiving an average of 1,270 mm of rain annually (Black, 2003). However, rainfall









distribution is not adequate for vegetable production and irrigation must be used since rainfall is

always unevenly distributed in time and space (Cockx and Simonne, 2003).

Irrigation scheduling is used to apply the proper amount of water to a crop at the proper

time. The characteristics of the irrigation system, crop needs, soil properties, and atmospheric

conditions must all be considered to properly schedule irrigations. Poor timing or insufficient

water application can result in crop stress and reduced yields from inappropriate amounts of

available water and/or nutrients. Excessive water applications may reduce yield and quality, are a

waste of water, and increase the risk of nutrient leaching (Maynard and Olson, 2001).

Irrigation must be scheduled according to water availability and crop need. Irrigation

scheduling requires knowing when to irrigate and how much water to apply. When to irrigate can

be determined from plant or soil indicators or water balance techniques. How much water to

apply can be based on soil water measurements or water balance techniques (Fangmeier etc.,

2006).

Monitoring soil status always means checking soil water tension (SWT). SWT represents

the magnitude of the suction (negative pressure) the plant roots have to create to free soil water

from the attraction of the soil, and move it into root cells. The dryer the soil, the higher the

suction needed, hence, the higher SWT. SWT can be measured in the field with moisture sensors

or tensiometers (Olson and Simonne, 2005).

Crop water requirement information is needed when establishing a soil water budget to

forecast irrigation events. The sum of the water lost from the soil surface (evaporation) and water

used by plants (transpiration) is called evaportranspiration (ET). There are many factors that

affect the rate of ET, including plant species, weather factors, and the amount and quality of

water available to the plant. Generally, reference ET (ETo) is determined for use as a base level.









Crop water use (ETc) is related to ETo by a crop coefficient (Kc) that is the ratio of ETc to ETo

(Irrigation Association, 2001).

Water usage also varies with soil dryness. Plants can remove water more easily from a wet

soil. To account for this, a concept called readily available water (RAW) has been developed

(Keller and Bliesner, 1990). It defines the amount of water that is more easily remove by the

plant. Another associated term, maximum allowable depletion (MAD) relates RAW with

available water (AW), which is the water that can be stored in soil and be available for growing

crops. Usually, the value of MAD is given for a particular plant and the RAW is then computed

with equation EAW=AW X MAD. The MAD values can be expressed as percentage and usually

range from 0.4 to 0.6 (Rochester, 1995).

BMPs are specific cultural practices that aim at reducing the loads of specific compounds

while increasing or maintaining economical yields (Simonne and Hochmuth, 2003). The

implementation of BMPs may be a key factor in reducing the consequences of alterations of the

N cycle in sweet corn fields. Implementation of BMPs at the farm level is a key to maintaining

the quality and the quantity of ground and surface water.

Li and Yost (2000) stated that the application rates, timing, and method of both N

fertilization and irrigation are important tools that determine and control the fate and behavior of

N in soil-plant systems. For example, multiple applications with small amounts of fertilizer (e.g.

split application) usually enhance plant uptake and reduce potential nitrate leaching, although

increasing costs.

Waskom (1994) summarized BMPs for nitrogen fertilization for crops such as corn, sugar

beet, and beans as follows:

(1) Time application of N fertilizer to coincide as closely as possible to the period of
maximum crop uptake;









(2) Use sidedress or in-season fertilizer application for at least 40% of the total N applied to
irrigated spring planted crops or fields with severe leaching hazard;

(3) Apply N fertilizer where it can be most efficiently taken up by the crop:
a) Ridge banded fertilizer used in conjunction with alternate row furrow irrigation can
reduce downward movement of N;
b) Multiple, small applications of N through sprinkler irrigation systems can increase
fertilizer efficiency and reduce total N fertilizer application;
c) Fertilizers applied on irrigated fields with high surface loss potential should be
subsurface banded or incorporated immediately after application;
d) Nitrogen applied in irrigation water should be metered with an appropriate device
that is properly calibrated. Due to the increased possibility of leaching or runoff, N
fertilizer through conventional flood or furrow irrigation system is strongly
discouraged.

(4) The following recommendations apply to cropland fields where the leaching potential is
moderate to severe:
a) Follow alfalfa or other legumes with high N use crops (such as small grains, sugar
beets, or corn) that efficiently use N fixed by the legume;
b) Follow shallow-rooted crops with low N use efficiency in the rotation by a deep-
rooted, high N use crop that scavenges excess N (such as corn, sugar beets, or
alfalfa). Analyze subsoil samples for residual nitrate to determine carryover credit
to the subsequent crop.

Bauder and Waskom (2003) summarized the BMPs for corn in Colorado. The BMPs

include: (1) use sidedress or in-season fertilizer application for at least 40% of the total N applied

to irrigated crops with sandy soils; (2) use fall planted cover crops such as rye or triticale to

scavenge excess N left in the soil after poor crop; (3) mix and store N fertilizer at least 30 m (100

feet) away from wells or any water supply; (4) if applying manure, incorporate manure as soon

as possible after application to minimize volatilization losses, reduce odor, and prevent runoff,

and (5) apply only enough irrigation water to fill the effective crop root zone.

Hochmuth (2000) recommended the nitrogen management practices for vegetable

production in Florida as follows: (1) knowing the crop nutrient requirement (CNR) for N and

targeting this amount for total crop N fertilization; (2) setting realistic yield goals; (3) using

polyethylene mulch, where practical, to protect N from leaching; (4) selecting controlled-release









N fertilizers when practical and economical; (5) calibrating fertilizer applicators accurately and

making adjustments to equipment so that the correct amount of N is applied in the correct

position of the root zone or production bed, near the root system; (6) applying N at periods

during the growing season when crop N uptake is most active; (7) using fertigation where

possible to "spoon-feed" N to crops during the season; (8) managing irrigation water properly to

avoid leaching and to keep water and N in the root zone; and (9) using tissue-testing or petiole

sap testing to monitor crop-N status and to determine adjustments needed in the N-fertilization

program. In addition, he also suggested 224 kg N ha-1 as nitrogen recommendations for sweet

corn production on sandy mineral soils in Florida.

In the "Vegetable Production Guide for Florida 2003-2004", Olson and Simonne (2005)

suggested that 20% to 25% of N should be applied at planting, then sidedress band the remaining

N in one or two applications during the early part of growth cycle. After midseason, N can be

applied through center pivot irrigation systems at rates of 11 to 22 kg N ha-1 in several

applications.

1.1.5 Best Management Practice Development

Best management practices related with irrigation and N fertilizer application have been

developed with field plot experiments. For example, a study was conducted in an acid-sulfate

soil in the central region of Thailand, in 1999 and 2000 to assess the influence of different rates

ofN fertigation on corn yield and nitrate leaching. The corn varieties planted in the two years

were super sweet corn Agro variety (Zea mays L.) and the Suwan 3851 single-cross hybrid (Zea

mays L.), respectively. The nitrogen source was urea and there were four N fertigation treatments

that included 0 (control), 100, 150 and 200 kg N ha-1, each having three replications arranged in

a randomized complete block design (RCBD). Soil was irrigated to field capacity at 50%

available soil moisture depletion regime throughout the season. The average maximum corn









grain yield of 3,520 kg ha 1 was obtained at 200 kg N ha 1 in 1999 and 5,420 kg ha was

obtained at 150 kg N ha in 2000. But the statistical analysis did not show any significant

differences in grain yield between N200 and N150 treatments in either year. The nitrate leaching

was calculated from the equation LN = DPR x C, where DPR was the water drainage, and C was

nitrate nitrogen concentration in soil water measured by a soil water sampler. The results of

leaching calculation showed that the highest leaching values were obtained in N200 treatments in

both years with 23 and 5.3 kg N ha in 1999 and 2000, respectively. The lowest yield of 0.55

and 0.98 t ha 1 were obtained at 0 kg N ha 1 in 1999 and 2000, respectively (Asadi et al., 2002).

Sweet corn fertilization research has been conducted in Florida for more than thirty years.

During the 35-year period from 1962 to 1996 yields have increased. Sustained high yields can be

expected with fertilization practices designed to supply crop nutrient requirements (Volk, 1962;

Robertson, 1962; Rudert and Locascio, 1979; Hochmuth et al., 1992; Hochmuth, 1994; White et

al., 1996).

Hochmuth and Cordasco (2000) summarized the field research of nitrogen fertilizer

application in sweet corn production that occur on the mineral soils of the north, west, southwest,

and central regions of Florida. Of the fifteen summarized experiments, fourteen resulted in

optimum yields with N rates at or below the nitrogen fertilizer application rate of 168 kg N ha-1.

However, additional studies are needed to evaluate yield responses to nitrogen rates above 168

kg N ha-1. Plants fertilized with 190, 381 or 526 kg N ha-1 on marl and rockland soils resulted in

yields equivalent to those fertilized with 168 kg N ha-l. Split N application increased yield 14%

in a 1962 experiment compared to yields from plants fertilized in a single application (Volk,

1962). The remaining experiments were fertilized with the split method, recommended for un-

mulched crops where leaching and fertilizer bur might occur with the single application method.









Nitrogen recovery was improved when fertilizer was banded in the root zone to one side or to

both sides of the plant row. In some experiments, the length of ear blank-tip area decreased with

N rates from 0 to 168 kg N ha-1, yield of cull ears decreased, and yield of fancy and No. 1 grade

ears increased with168 kg N ha-1 compared to yields with lower N treatments.

However, development and certification of site-specific guidelines for optimal timing,

water, and nitrogen requirements requires extensive and expensive field experiments. Since it is

impossible to test all the interactions between the amount of water and nitrogen during the

seasons, use of simulation models can greatly facilitate the evaluation of different production

practices and/or environments and thereby streamline the decision-making process (Rinaldi et al.,

2007). Several examples of using crop models to test different practices for different crops are

summarized as follows.

Paz et al. (1999) stated that past efforts to correlate yield from small field plots to soil type,

elevation, fertility, and other factors had been only partially successful for characterizing spatial

variability in corn yield. Furthermore, methods to determine optimum nitrogen rate in grids

across fields depended upon the ability to accurately predict yield variability and corn response

to nitrogen. They developed a technique to use the CERES-Maize crop growth model to

characterize corn (Zea mays L.) yield variability. The model was calibrated using 3 years of data

from 224 grids in a 16 ha field near Boone, IA. The model gave excellent predictions of yield

trend along transects in the field, explaining approximately 57% of the yield variability. Once the

model was calibrated for each grid cell, optimum nitrogen rate to maximize net return was

computed for each location using 22 years of historical weather data.

The model for potato growth (LINTUL-NPOTATO) was used to explore N uptake, tuber

yield and residual soil mineral N (RSMN) of a potato crop (Solanum tuberosum L.) for 30 years









of historical weather data, as influenced by: (1) the time of slurry application; (2) cultivar

maturity; (3) the N/P ratio of the manure; and (4) historical N use. Results indicated that a

spring-applied slurry is to be preferred over an autumn-applied slurry in order to avoid over-

winter N losses. Patterns of N uptake suggest that organic N with a large proportion of mineral N

and applied shortly after emergence, could improve potato yields in organic farming (Van

Delden et al, 2003).

Rinaldi et al. (2007) used the CROPGRO model to predict the growth of processing potato

(Lycopersicon esculentum Mill.) in Southern Italy. One data set of 2002 was used to calibrate the

model, while three independent data sets were used to validate the model. Subsequently this

model was combined with 53 years of local historical weather data and it was used as a research

tool to evaluate the benefits, risks and costs of 23 different interactive irrigation and/or N-

management scenarios. Irrigation water was applied (1) on reported dates with 3 and 5 days

intervals and application rates of 15 and 25 mm or (2) with automatic irrigation initiated at

residual soil moisture levels in the upper 30 cm of the soil profile of 25, 50, or 75%. Three

amount levels of N application (100, 200 and 300 kg ha-1 as ammonium nitrate) were considered.

Based on simulation results it is concluded that irrigation scenario with low amount but with

frequent applications ("3-day 15 mm" scenario) resulted in high value of irrigation water use

efficiency; frequent irrigation applications combined with low N rates reduced crop stress and

represented the best scenario from both a production and environmental point of view (low N

leaching).

Thorp et al. (2006) used the CERES-Maize crop growth model to study the corn (Zea mays

L.) yield response and the nitrogen (N) dynamics of a cornfield in central Iowa, USA. The model

was calibrated to minimize error between simulated and measured yield over five growing









seasons. Model simulations were then completed for 13 spring-applied N rates in each of 100

grid cells with varying soil properties. For each N rate and grid cell, simulations were repeated

for 37 years of historical weather information collected near the study site. Model runs provided

the crop yield and unused N in the soil at harvest for all combinations of N rate, grid cell, and

weather year. The overall goal of this work was to develop a methodology for directly

contrasting the production and environmental concerns of N management in agricultural systems.

In this way, N management plans can be designed to achieve a proper balance between

production and environmental goals.

1.2 Objectives

The current project is an attempt to utilize the CERES-Maize model of the Decision

Support System for Agrotechnology Transfer (DSSAT) model to develop potential N-BMPs for

sweet corn production in North Florida. The objectives include:

* Study on the response of sweet corn yield quantity and quality to different irrigation and
nitrogen application levels;

* Study of the nitrogen fate and balance in sweet corn production;

* Global sensitivity analysis of the CERES-Maize model;

* Application of the generalized likelihood uncertainty estimation (GLUE) method in
parameter estimation of the CERES-Maize model;

* Utilization of the CERES-Maize model to develop potential N-BMPs for sweet corn
production;

* Uncertainty analysis for the developed potential BMPs both under weather and input
parameter uncertainties.

1.3 Dissertation Outline

This current research both involves field experimentation and crop model simulation. The

main purpose of field experiments is to provide necessary data for model simulations such as









sensitivity analysis, model calibration, and model verification. At the same time, field

experiments also provide enough materials to study the response of sweet corn yield to different

nitrogen fertilizer and irrigation levels and the fate of fertilizer nitrogen in sweet corn production.

The model simulation is the core part of the research. The main research structure could be

shown in Figure 1-1.

In general, Chapter 2 of this dissertation will present a sensitivity analysis of the crop

model, in which the behavior of the crop model is investigated and the most sensitive input

parameters are selected for calibration. In Chapter 3, the generalized likelihood uncertainty

estimation (GLUE) parameter estimation process will be described as a procedure for model

calibration. Chapter 4 will compare the model outputs and field observations in a procedure of

model verification. In Chapter 5, the procedures of BMP development will be discussed, where

some potential BMPs will be selected with the calibrated model. In Chapter 6 uncertainty

analysis will be conducted for the selected potential BMPs. Finally in the last chapter (Chapter 7),

some research conclusions and suggestions about future work will be provided.






























Figure 1-1. Diagram of research structure










Table 1-1. Sweet corn harvested for sale in Florida in 2002 and 1997 (USDA-NASS, 1998, 2002)
2002 1997
Level
Farms Area (ha.) Farms Area (ha.)
0-2 ha. 225 114 268 132
2-20 ha. 73 403 94 436
20-100 ha. 15 872 12 616
100-200 ha. 13 2,010 14 1,824
More than 200
14 25
ha 12,368 14,708
Total 340 15,768 413 17,791


Table 1-2. Nitrogen fertilizer application for sweet corn in Florida (USDA-NASS, 1993, 1995,
1999b, 2003, 2006)
Planted Area Applications Rate per Rate per Total
Year Area Applied Application Crop Year Applied
ha. % Number kg ha-1 kg ha-' kg
1992 20,800 81 3.0 46 137 2,300,000
1994 17,200 90 2.3 47 106 1,648,000
1998 16,800 100 2.0 54 109 1,831,000
2002 16,400 100 2.5 62 159 2,606,000
2006 13,400 86 10.0 47 475 5,480,000









CHAPTER 2
GLOBAL SENSITIVITY ANALYSIS OF CERES-MAIZE MODEL WITH ONE-AT-A-TIME
METHOD

2.1 Introduction

2.1.1 Sensitivity Analysis

A crop model has been described as a "quantitative scheme for predicting the growth,

development and yield of a crop, given a set of genotype coefficients and relevant environmental

variables" (Monteith, 1996). A crop model is the result of a long and complex construction

process, involving data at multiple states for understanding basic process, elaborating model

structure, estimating parameters and evaluating prediction quality. However, there is a need to

study the model on its own, with an emphasis on its behavior rather than its coherence with a

given data set. This is where sensitivity analysis becomes useful for the modeler and model user

(Monod et al., 2006).

The sensitivity analysis determines how sensitive the output of a crop model is, with

respect to the elements of the model which are subject to uncertainty or variability. This is useful

as a guiding tool when the model is under development as well as to understand model behavior

when it is used for prediction or for decision support. For dynamic models, sensitivity analysis is

closely related to the study of error propagation, i.e. the influence that the lack of precision on

model input will have on the output.

Because sensitivity analysis usually relies on simulations, it is also closely related to the

methods associated with computer experiments. A computer experiment is a set of simulation

runs designed in order to efficiently explore the model responses when the input varies within

given ranges (Sacks et al., 1989; Welch et al., 1992). The goals in computer experiments

identified by Koehler and Owen (1996) include optimization of the model response, visualization









of the model behavior, approximation by a simpler model or estimation of the average, variance,

or probability of the response to exceed some threshold.

Within a given model, model equations, parameters and input variables are all subject to

variability or uncertainty. The following 3 reasons make it an inevitably necessary step to do

sensitivity analysis before any model simulation.

First, choices have to be made regarding the model structure and on the functional

relationships between input variables and output variables. These choices may sometimes be

quite subjective and it is not always clear what their consequences will be. For example,

Martinez et al. (2001) performed a sensitivity analysis to determine the effects of the number of

soil layers on the output of a land surface-atmosphere model. For spatial models, there was

frequently a need to evaluate how the scale chosen for input variables affects the precision of the

model output (see e.g. Salvador et al., 2001).

Second, parameter values result from estimation procedures or sometimes from

bibliographic reviews or expert opinion. Their precision is however limited by the variability and

possible lack of adequacy of the available data. Some parameters may also naturally vary. The

uncertainty and natural variability of parameters are the central point of many sensitivity

analyses. Barlund and Tattari (2001), for example, studied the influence of model parameters on

the predictions of field-scale phosphorus losses, in order to get better insight into the

management model ICECREAM. Ruget et al. (2002) performed sensitivity analysis on

parameters of the crop simulation model STICS, in order to determine the main parameters that

need to be estimated precisely.

Third, additional and major sources of variability in a model output are the input variables.

Lack of precision when measuring or estimating input variables need to be quantified when









making predictions from a model or when using it for decision support. Rahn et al. (2001)

compared contrasted input scenarios for HRI WELL-N model on crop fertilizer requirements

through a sensitivity analysis. They identified the main factors that need to be measured

precisely to provide robust recommendations on fertilization. Contrasted settings of the input

variables were used for performing sensitivity analyses assuming different scenarios by Dubus

and Brown (2002).

As shown by the examples above, sensitivity analysis may have various objectives, such as:

(1) to check that the model output behaves as expected when the input varies; (2) to identify

which parameters have a small or a large influence on the output; (3) to identify which

parameters need to be estimated more accurately; (4) to detect and quantify interaction effects

between parameters, between input variables, or between parameters and input variables; (5) to

determine possible simplification of the model; and (6) to identify input variables which need to

be measured with maximum accuracy (Monod et al., 2006). Some of these objectives have close

links with other methods associated with modeling, like model construction, parameter

estimation or model use for decision support.

2.1.2. Local Sensitivity Analysis

Local sensitivity analysis is based on the local derivatives of a model output Y = f(Z)with

respect to a single input factor Z, which indicates how fast the output increases or decreases

locally around given values ofZ The derivatives can sometimes be calculated analytically, but

they are usually calculated numerically for complex models. Problems may arise if the derivative

of the model does not exist at some points. In addition, the derivatives may depend strongly on

the Z value (Monod et al., 2006).









The local (first-order) sensitivity coefficient S,/""C (Zk) is defined as the partial derivative of

the output variable Y with respect factor Z,, calculated at the scenario Zk:


S ca (Zk)= f (Z (2-1)
8Z

This criterion is equivalent to the slope of the calculated model output in the parameter

space, and S, "" (Zk) criterion is an absolute measure of sensitivity, which depends on the scales

or measurement units of Y and Z, A standardized version, called the relative sensitivity, is

defined by following equation (Monod et al., 2006):

so f(Z) Zk,
SO(Zk ) x(Z) (2-2)
OZ, f(Zk)

Local sensitivity analysis can be used to study the role of some parameters or input

variables in the model. But this method is less useful than global sensitivity when the purpose of

the analysis it to study the effect of uncertainty of several factors on model outputs.

2.1.3 Global Sensitivity Analysis

In global sensitivity analysis, the output variability is evaluated for input factors varying

within their entire domains. This provides a more realistic and comprehensive view of the model

behavior.

There are several methods for global sensitivity analysis, such as the one-at-a-time (OAT)

method (Morris, 1991), factorial design and analysis of variance (Monod et al., 2006), intensive

sampling and variance-based method (Monod et al., 2006), and the Fourier amplitude sensitivity

test (FAST) method (Chan et al., 2000). The OAT method is more straightforward and less

complicated in application compared to the other methods mentioned above.









The objective of this research is to conduct global sensitivity analysis for the CERES-

Maize model with the one-at-a-time (OAT) method so as to: (1) determine the sensitivity of the

model outputs (dry matter yield, kg ha-1 and cumulative nitrogen leaching, kg N ha-1) with

respect to changes in soil and genotype input parameters, and (2) identify the most influential

input parameters that need to be calibrated in future research.

2.2 Materials and Methods

2.2.1. Model Description

2.2.1.1 CERES-Maize model

The crop model CERES-Maize, used for this research is embedded in the Decision Support

System for Agrotechnology Transfer (DSSAT) software (Jones et al., 2003), version 4.0. To run

the model, several input files must be compiled that contain information about the experiment

site, soil, climate and genotype (Tsuji et al., 1994).

At the heart of the DSSAT revisions is a cropping system model (DSSAT-CSM), which

incorporates all crops as modules using a single soil model (Jones et al., 2003). The CERES-

Maize, Wheat and Barley models were modified for integration into the modular DSSAT-CSM.

For these CERES models, the plant life cycle is divided into several phases, which are similar

among the crops. Rate of development is governed by thermal time, or growing degree days

(GDD), which is computed based on the daily maximum and minimum temperatures. The GDD

required to progress from one growth stage to another are either defined as a user input, or are

computed internally based on user inputs and assumptions about duration of intermediate stages.

The genotype coefficients for the DSSAT CERES-Maize, Wheat and Barley models are listed in

Table 2-1.

Daily plant growth is computed by converting daily intercepted photosynthetically active

radiation (PAR) into plant dry matter using a crop-specific radiation use efficiency parameter.









Light interception is computed as a function LAI, plant population, and row spacing. The amount

of new dry matter available for growth each day may also be modified by the most limiting of

water or nitrogen stress, and temperature, and is sensitive to atmospheric CO2 concentration.

Above ground biomass has priority for carbohydrate, and at the end of each day, carbohydrate

not used for above ground biomass is allocated to roots. Roots must receive, however, a specified

state-dependent minimum of the daily carbohydrate available for growth. Leaf area is converted

into new leaf weight using empirical functions (Jones et al., 2003).

Kernel numbers per plant are computed during flowering based on the cultivar's genotype

potential, canopy weight, average rate of carbohydrate accumulation during flowering, and

temperature, water and nitrogen stresses. Potential kernel number is a user-defined input for

specific cultivars. Once the beginning of grain fill is reached, the model computes daily grain

growth rate based on a user-specified cultivar input defined as the potential kernel growth rate

(mg kernel-ld-1). Daily growth rate is modified by temperature and assimilate availability. If the

daily pool of carbon is insufficient to allow growth at the potential rate, a fraction of carbon can

be remobilized from the vegetative to reproductive sinks each day. Kernels are allowed to grow

until physiological maturity is reached. If the plant runs out of resources, however, growth is

terminated prior to physiological maturity. Likewise, if the grain growth rate is reduced below a

threshold value for several days, growth is also terminated (Jones and Kiniry, 1986; Ritchie and

Otter, 1985; Ritchie et al., 1998).

2.2.1.2 Soil water sub-module

The soil water balance model developed for CERES-Wheat by Ritchie and Otter, (1985)

was adapted for use by all of the DSSAT v3.5 crop models (Jones, et al, 2003). This one-

dimensional model computes the daily changes in soil water content by soil layer due to

infiltration of rainfall and irrigation, vertical drainage, unsaturated flow, soil evaporation, and









root water uptake processes. The model uses a "tipping bucket" approach for computing soil

water drainage when a layer's water content is above a drained upper limit parameter, or field

capacity. Upward unsaturated flow is also computed using a conservative estimate of the soil

water diffusivity and differences in volumetric soil water content of adjacent layers (Ritchie,

1998).

Soil water infiltration during a day is computed by subtracting surface runoff from rainfall

that occurs on that day. The SCS (Soil Conservation Services) method is used to partition rainfall

into runoff and infiltration, based on a "curve number" that attempts to account for texture, slope,

and tillage. When irrigation is applied, the amount applied is added to the amount of rainfall for

the day to compute infiltration and runoff Drainage of liquid water through the profile is first

calculated based on an overall soil drainage parameter assumed to be constant with depth. The

amount of water passing through any layer is then compared with the saturated hydraulic

conductivity of that layer, if this parameter is provided. If the saturated hydraulic conductivity of

any layer is less than computed vertical drainage through that layer, actual drainage is limited to

the conductivity value, and water accumulates above the layer. This feature allows the model to

simulate poorly drained soils and perched water tables. For example, a soil may have a layer with

very low or no drainage at the bottom of the profile. Vertical drainage from the profile would not

occur or it would be very low, limited by the saturated hydraulic conductivity value of the

bottom layer (Jones et al., 2003).

Evaporation of water from the soil surface and root water uptake (transpiration) from each

layer are computed in the soil-plant-atmosphere model (SPAM) with the Priestley-Taylor

equation (Priestly and Taylor, 1972) and communicated to this soil water balance module. Each









day, the soil water content of each layer is updated by adding or subtracting daily flows of water

to or from the layer due to each process (Jones et al., 2003).

2.2.1.3 Soil nitrogen sub-module

The nitrogen balance model simulates the processes of organic matter turnover with the

associated mineralization and/or immobilization of nitrogen, nitrification, denitrification,

hydrolysis of urea, ammonia volatilization, N plant uptake and translocation to the different

organs during crop cycle. Transport of nitrate occurs at the same rate as the flow of water

(Booltink et al., 1996).

The CERES N model of the DSSAT model has two forms, one for upland cereal crops and

one for flooded soil rice cropping systems. Both versions simulate the turnover of soil organic

matter and the decay of crop residues with the associated mineralization and/or immobilization

of N. Nitrification of ammonium and N losses associated with denitrification are estimated by

both models. The lowland version adds to this a floodwater chemistry routine which simulates

the fluxes of ammonia N and urea between floodwater and soil, and calculates ammonia and

volatilization losses. Both models incorporate a plant N component that simulates N uptake and

distribution within the plant and remobilization during grain filling and plant growth responses to

plant N status. The models are closely coupled with the CERES water balance and crop growth

routines (Tsuji et al., 1998).

Since the soil nitrogen model used within the DSSAT is intrinsically linked to the water

balance model, those parameters used by the water balance model that define ranges of soil water

availability, soil drainage and deep percolation characteristics and layer depth increments are

also required by the N model. The inputs for the soil water balance model are described by

Ritchie and Otter (1985). The N model itself requires input data that describe the initial amount

of mineral N present in the soil profile and information that will enable the estimation of how









much N will be mineralized from soil organic matter, the potassium chloride extractable nitrate

and ammonium present in each of the layers. The soil bulk density is used in the calculations of

concentrations of N from mass (Tsuji et al., 1998).

2.2.2 Non-restricted OAT Method

The most intuitive method to conduct a sensitivity analysis is to vary one factor at a time,

while the other factors are fixed at their nominal values. The relationship between the values z, of

factor Z, and the responses f(z0,1,...z0-,, 1' zo,1+ ...Zo,s) determines an OAT response profile,

where S is the total number of parameters. In practice, each input factor Z, takes k equispaced

z z
values from Zmi, to z max, with an increment of = max,, m',, The model responses
(k 1)

f(Zo,1,...Zo,, 1z, z,,Zo,z+...zo,s) are then calculated for each of the k discretized values ofZ,

(Monod et al., 2006). The main idea of non-restricted OAT method can also be briefly shown in

Figure 2-1.

In Figure 2-1, the space of possible values of parameter Z, from the minimum to the

maximum value could be divided and represented as al x k matrix, where k is the dimension of

the space. Other parameters were assigned with their nominal value, which were the mean values

derived from DSSAT model database in current research. In the example shown in Figure 2-1,

the first element or the minimum value of the matrix was selected as the model input parameter.

The model is run with the discretized values ofZ, and the relevant outputs are recorded.

After trying all of the available values of Z, in its discretized space, the process is repeated for

other input parameters. When the value ofZ, varies, all other parameters keep their norminal









values. If the number of sensitivity parameters is not too large, graphical representations are the

best way to summarize the response profiles.

The number of k must be chosen carefully when the model is non-linear and particularly

when it is non-monotonic. Provided the value of k is odd, the number of model simulations to

calculate all profiles is equal tos(k -1) + 1. Whenk is small and the model is non-linear, the non-

linear effects, as well as maxima or minima, may be undetected, which may lead to under-

estimating sensitivity indices such as the index of Bauer and Hamby (1991). However, when k

is too large, the computing time may become very long if there are many input factors and the

model is complex.

However, the non-restricted OAT method does not provide any information about

covariance and interaction between the input parameters, which might also contribute to

uncertainty in predictions. When only selecting parameters just according to the response

profiles, some important parameters might be missed. To compensate for the drawback

mentioned above, a correlation coefficient matrix was established for the genotype and soil

parameters concerned to assist in parameter selection.

2.2.3 Normalization of Input Parameters

The results of global sensitivity analysis with non-restricted OAT method can be shown as

response curves of outputs concerned to input parameters. However, input parameters always

have different units and ranges. It is difficult to present all of the response curves in one figure

when under such different ranges and units. If a common range is used for all parameters, the

presentation could be simplified.

The simplification could be done by normalizing the levels of the factors so that they vary

between -1 and +1 or between 0 and 1. Normalized values z, of an input factor Z, can easily be









calculated from the un-normalized values through the following relationships (Monod et al.,

2006):


C Z (Zmax() + Zmln() )/2(23)
zl ~= ( (2-3)
(Zmax(l) Zmn())/2

or


Zc =_ Z Zmn() (2-4)
(Zmax(l) Zmn(1))

In this research, with Equation (2-3), the response curves of dry matter yield and

cumulative nitrogen leaching to the genotype and soil parameters in their whole domains were

represented.

2.2.4 Restricted OAT Method

Unfortunately, the non-restricted OAT method does not tell model users anything about the

sensitivity contribution by the interactions between input parameters, since this method holds

other parameters fixed at their nominal values while only changing one factor. This problem can

be solved by the restricted OAT method, which was exploited by Morris (1991).

The main idea of the Morris restricted OAT method can be briefly explained by Figure 2-2.

The space of each parameter is divided intok equal sections, shown as a k x 1 matrix. For all of

the s factors, a k x s matrix is constructed. It was assumed the input paremeters were independent.

One value is then randomly picked up for each parameter from its own space. The s randomly

picked values of relevant parameters construct a possible parameter scenario. Then the model

can be run with this established scenario to calculate the elementary effect or local sensitivities

for each of the parameters by just changing one parameter while keeping other parameters fixed

at their nominal values of this specific scenario. After this scenario, another randomly









established scenario can be used to repeat the same process until sufficient local sensitivity

values are available for each parameter.

The random selection of parameter values to construct a scenario can be realized with the

method below. First, a 1 x s matrix of random integers is generated. Each element of the integer

matrix follows a uniform distribution of [1, k]. Then use these random integers as addresses to

select values for the parameters from their individual spaces. For example (as shown in Figure 2-

2), the first element of the integer matrix, R1, is 2. Then the second element of the space ofZ, is

selected as the nominal value for Z in the scenario. The same process is repeated for other

parameters to construct the scenario. It is obvious that more scenarios were constructed, more

reliable the results would be. However, the running time should also be considered, because tens

of thousands of model runs were required. In this research, 2,000 scenarios were constructed

finally.

Morris defined the elementary effect of the ith input factor for a given scenario

Zo = ( o,'1,...z0 Zo,1 zo,2, o,+...z ,S ) as:


S(Z = Zf(O,1,...ZO,,-1'ZO,, + A Z 0,1+1 Z s)-f,S() Z- f(zo,li, Zo, lz, Z, +l, z .. ') Z, 2-
d Zo) ~ (2-5)
A


where zo,, + A is a perturbed value ofz0,,, and A is a predetermined multiple ofl /(k -1).

The number of parameter scenarios depends on the reliability of the result of sensitivity

analysis and the time of model running. In theory, more scenarios are tested, more reliable the

sensitivity analysis result. However, if too many scenarios and parameters are involved, it will

require huge number of model running, which will be very time consuming and delay future

work. In this research, 2,000 scenarios were constructed.









After calculating d, (Z0) for sufficient scenarios, the resulting distribution of the

elementary effects of the ith factor is then characterized by its mean and variance. A high mean

indicates a factor with an important influence on the output. A high variance indicates either a

factor interacting with another factor or a factor whose effect is non-linear.

2.2.5 One-at-a-time (OAT) Method for CERES-Maize Model

In the current research, all of the soil and genotype input parameters of the CERES-Maize

model were investigated both with the non-restricted and restricted OAT method to identify the

most influential parameters for future model calibration. The six genotype input parameters of

the model are listed in Table 2-1. The nine soil parameters include the following: soil water

saturation (SSAT), drained lower limit (SLLL), drained upper limit (SDUL), bulk density

(SBDM), soil albedo (SALB), evaporation limit (SLU1), runoff curve number (SLRO), drainage

rate (SLDP), and fertility factor (SLPF). See Appendix A for details about the definitions and

units of these parameters.

This procedure provides useful information for the model developers, e.g. which input

parameters need more accurate measurement or calculation. The outputs of concern are dry

matter ear yield (kg ha-1) and cumulative nitrogen leaching (kg ha-1), since they are the two main

factors in potential best management practices (BMPs) development in this study (Chapter 5).

The main procedures of global sensitivity analysis of the CERES-Maize model with the

restricted OAT method in this study are outlined here. First, the sampling spaces for each of the

six genotype and nine soil parameters were established according to the values available in the

DSSAT model database. For each parameter, the range between the minimum and maximum

value was divided into 100 equal sections, with an increment of = (Zmax Zm ) /(100 -1).

These 100 values were saved as a vector. The same process was repeated for other parameters.









However, for SLLL, SDUL and SSAT, the scenarios of SLLL>SDUL or SDUL>SSAT were

avoided since they could cause the model to stop running. For example, if SLLL>SDUL, it

means the soil available water (SDUL-SLLL) would be negative, which conflicts with the basic

physical principles. Therefore the minimum value of SLLL is set as its own minimum value, but

the maximum value is set as the minimum value of SDUL. In the same way, the minimum value

of SSAT is set as the maximum value of SDUL.

Next, 15 random integer numbers following a uniform distribution of [1,100] were

generated, since there were a total of 15 genotype and soil parameters under investigation. These

15 random integers were assigned to each of the 15 parameters as the addresses for selecting

values from their individual matrix space. The 15 selected numbers from the spaces of input

parameters constructed a scenario of parameter set to run the model.

Third, the values of the scenario were used to change corresponding values in the soil and

genotype files and replace their original files in the correct installation directory ofDSSAT

model. As the model was run, outputs were saved.

Fourth, a perturbation was given to the ithparameter, for example a 5% increment from the

initial value. The perturbation value should not be too large. Otherwise the result will fail to

approximate the definition of local sensitivity (Equation 2-1 and 2-2). It also should be not too

small. Otherwise some of the model outputs, especially nitrogen leaching, will not change when

only under a very small perturbation of the input parameters. In this research, it was found that

5% was a good choice for perturbation after comparing the results under a perturbation of 3%,

5%, 10%, and 20%. Thus, a new parameter vector was generated. The values in this new

parameter vector were used to change the genotype or soil file again. The model was then rerun

and the process was repeated. Equation (2-5) was used to calculate the elementary effect of









ith parameter under this specific scenario. The same procedure was repeated to calculate the

elementary effects for all of the input parameters under this specific input parameter scenario.

Fifth, the elementary effects for the parameters under other randomly established scenarios

were calculated, as well as the mean and variance values of the elementary effects for each

parameter. Since the values of elementary effects might be either positive or negative, indicating

a parameter may increase the output in some places or decrease the output in other places when

the value of the parameter increases, the mean values alone might be misleading. For example, a

parameter that simultaneously has large positive and negative values of elementary effect might

have a very low mean value of elementary effect. Therefore in the current research, the mean and

variance of the absolute values of the elementary effects were calculated.

Finally, the mean and variance of the absolute values of the elementary effects of each

parameter were compared to determine which parameters have greater influences on the output.

The global sensitivity analysis with the restricted OAT method was conducted with Matlab

program (Appendix B for more detailed codes). A total of 2,000 scenarios were randomly

generated, and hence 2000 x (15 + 2) = 34,000 model runs were conducted.

The main procedures of non-restricted OAT method were similar to those of restricted

OAT method. However, the number of model runs was smaller. As described above, since the

domain of each input parameter was evenly separated into 100 sections and 15 input soil and

genotype parameters were investigated, thus only 100 x 15 = 1,500 model runs were required.

In the current research, both the restricted and non-restricted OAT method were used to

conduct global sensitivity analyses for the dry matter yield and nitrogen leaching responding to

input parameters. However, for the restricted OAT method, the sensitivity analyses were

conducted separately for the genotype and soil parameters. It means when doing sensitivity









analysis for the genotype parameters, the soil parameters were fixed at their nominal values, and

vice versa. This is because the genotype and soil parameters are two completely different kinds

of input parameters. Thus it was assumed that the influence of interaction or correlation between

them on the results of sensitivity analysis would be neglectable.

2.2.6 Field Experiment

When doing sensitivity analysis, except for the soil and genotype input parameters,

additional information, such as planting date, harvest, irrigation, and, potassium application,

nitrogen fertilizer application etc., was required as fundamental inputs to run the model. This

information was obtained from the field experiment in this study.

The field experiments were conducted at the Plant Science Research and Education Unit,

the University of Florida in the spring of 2005. The unit is located in Pine Acres (29.4094N,

82.1777W, 20.746 meters above sea level), Marion County, Florida, U.S. (Judge et al., 2005).

There were two experiment field identified as Blockl and Plots. The variety of sweet corn

planted was Saturn SH2.

Finally, field management information obtained in Blockl in 2005 was used as

fundamental inputs for model testing. Weather data, including daily solar radiation, maximum

temperature, minimum temperature, and rainfall, were also required as the necessary driving

force for model testing. In this study, these data were obtained from the weather database of the

Florida Automated Weather Network (FAWN) in 2005 at the Citra, where the unit is located.

See Chapter 3 for more information about the field experiment in Blockl.









2.3 Results and Discussion


2.3.1 Non-restricted OAT Results

2.3.1.1 Response profiles

Drawing response profiles is often useful, at least in preliminary stages. The response

profiles of dry matter yield to genotype and soil parameters are shown in Figure 2-3 and 2-4

below, while the response profiles for nitrogen leaching are shown in Figure 2-5 and 2-6 for the

results of global sensitivity analysis with the non-restricted OAT method. Please be noticed these

figures do not represent any actual growth scenario of sweet corn, because the nominal values of

the input parameters were set as the mean values of them, which were derived from the DSSAT

database.

From these four figures it can be seen that genotypes P1, P5, PHINT and soil parameters

SLLL, SDUL and SLPF had strong influences on yield, while soil parameters SLLL, SDUL,

SLDR and SLRO had strong influences on nitrogen leaching. For example, when P1 is the mean

value, the predicted dry yield was about 8,400 kg ha-1. Then the predicted dry yield decreased to

about 6,200 kg ha-1 when P1 decreased to its minimum value, with a decrement of 26%. And it

increased to almost 11,800 kg ha-1 when P1 increased to its maximum value, with an increment

of 40%. However, for the genotype parameters (P2, G2, and G3) and soil parameters (SLDR,

SLRO and SBDM etc.), the predicted dry yields were almost kept at the same values as about

8,400 kg ha-l. These simulated dry yields were much higher than the field plot experiment results

(Chapter 4), since the model was not calibrated yet. The genotype parameters did not describe

the genetic characteristics of the real corn in the experiment.

Similar results were observed for nitrogen leaching. For example, the amount of nitrogen

leaching decreased from 220 kg N ha-1 to about 40 kg N ha- when the value of SDUL increased

from its minimum value (0.145 to 0.374 cm3/cm3) to the mean value,. And the amount of









nitrogen leaching increased from 40 kg N ha-1 to about 190 kg N ha-1, when the value of SLLL

decreased from its mean value to the maximum value. However, the amounts of predicted

nitrogen leaching under other soil parameters were all approximately 40 kg N ha-1 with minimal

changes due to input parameter changes.

Those parameters that showed high influence on model outputs when their values changed

should be considered with priority when select parameters for future model calibration with the

generalized likelihood uncertainty estimation (GLUE) method (Chapter 3). However, these

figures do not provide information about covariance and interaction between parameters, which

might also contribute to uncertainty in predictions. When only selecting parameters just

according to the response profiles, some important parameters might be missed.

2.3.1.2 Correlation coefficient matrix

As described in Section 2.2.2 and 2.3.1.1, the non-restricted OAT method did not provide

any information about the covariance of the input parameters. Some important parameters

probably will be missed if not considering the influence of covariance between the input

parameters. Thus, a correlation coefficient matrix was established to assist in influential

parameter selection.

In the current research, the parameter values in the database of the DSSAT model were

used to calculate the covariance values between each pair of the parameters. It was assumed that

there was no covariance between cultivar and soil parameters since they are completely different

kinds of parameters.

The calculated results were used to establish a correlation coefficient matrix as shown in

Table 2-2, where it can be observed that soil parameters SLLL, SDUL and SSAT are highly

correlated to each other.









2.3.1.3 Influential parameter selection based on non-restricted OAT method

Based on the response profiles of input genotype and soil parameters to dry matter yield

and cumulative and the correlation coefficient matrix, the influential input parameters either to

dry matter yield or nitrogen leaching could be selected for further calibration.

The criteria of selection were listed in Table 2-3. According to the criteria, if a parameter

was highly sensitive, it was selected. If a parameter was not very sensitive, but it had a high

correlation with a highly sensitive one, it was also selected. For example, from Figure 2-3, it is

easy to see that P1, P5, and PHIN were the most sensitive genotype parameter to dry matter yield,

since the response curves of them have the steepest slope. Thus P1, P5, and PHIN were selected.

In Figure 2-4 soil parameter SLLL, SDUL, and SLPF showed the highest sensitivity to dry

matter yield, so they were also selected.

Similarly, in Figure 2-5, P1, P5, and PHIN were the most sensitive genotype parameters to

cumulative nitrogen leaching, though the sensitivities were much lower than the influential soil

parameters. In Figure 2-6, more soil parameters showed significant influence on nitrogen

leaching, including SLLL, SDUL, SLDR, SLRO. These parameters were also selected.

For these four figures, it can be seen that soil parameter SSAT only showed a little bit

sensitivity to dry matter yield and nitrogen leaching. However, SSAT had a very high correlation

with the sensitive soil parameter SLLL and SDUL. As shown in Table 2-2, the covariance

coefficients between SSAT and SLLL and between SSAT and SDUL, were 0.576 and 0.647

respectively. To consider the probable sensitivity contribution by the covariance, soil parameter

SSAT was also selected. Thus, in the end, totally nine parameters selected based on the non-

restricted OAT method and the correlation coefficient matrix were listed in Table 2-4.

In general, one more parameter SSAT was selected than only considering the response

curves but not considering the covariance.









2.3.2 Influential Parameter Selection Based on Restricted OAT Method

As described in Section 2.2.4 that the results of the restricted OAT method were presented

as absolute elementary effect values, which have no units. The mean values and standard

deviations of the elementary effect values of the input parameters were calculated. If an input

parameter has a high mean value, e.g. greater than 1.0, it means the parameter is sensitive to the

output concerned. If the parameter has a high value of standard deviation, it means the parameter

is non-linear or highly correlated with other parameters.

The mean values and standard deviations of the elementary elements of the genotype

parameters corresponding to predicted dry matter yield and cumulative nitrogen leaching, which

are the two main factors for BMP development, are summarized in Table 2-5. It can be seen that

P1 was the most sensitive factor for dry matter yield, with P5 and PHINT following.

PHINT had a large value of variance for dry matter yield, which implies that the response

curve of PHINT to dry yield may be nonlinear, for example, existing plateau or threshold regions.

Actually from Figure 2-3, it can be seen that the response curve of dry matter yield to PHINT

was dentate. P5 and PHINT are also sensitive to nitrogen leaching, comparing with other

parameters. Therefore P1, P5 and PHINT should be selected when doing model calibration.

Table 2-6 shows the results of global sensitivity analysis with the restricted OAT method

of soil parameters corresponding to dry matter yield and nitrogen leaching. Most soil parameters

have little influence on yield, except for SLPF, SLLL and SDUL. This is true because SLPF is

the fertility factor that reflects the influence of micronutrients such as copper and zinc. SDUL is

the soil water holding capacity, while SLLL is the soil permanent wilting point. These two

parameters determine the soil available water (SAW) in the soil profile, which is always defined

as the difference between SDUL and SLLL. If the value of SAW is small, the corn may suffer

from water stress when the amount of irrigation is fixed, thus reducing yield.









SLLL, SDUL, and SLRO have strong influence on nitrogen leaching. Nitrogen leaching is

accompanied by water movement in soil profile, with more water infiltrating into soil profile,

more nitrogen will be leached. When the value of SDUL is lower, less water can be held by soil,

and more nitrogen will be leached. SLRO is the soil runoff curve number, which influences

water runoff on soil surface. If less water was lost by runoff, more would be lost by infiltration

and more nitrogen would be leached consequently.

Though the mean and variance of elementary elements to nitrogen leaching of SLDR were

lower than SLLL, SDUL, and SLRO, the values were still considerable (0.88 and 8.28) as shown

in Table 2-6. From Figure 2-6, it can also be seen that SLDR was sensitive to nitrogen leaching.

And the SLDR represents the soil drainage coefficient, while soil drainage will definitely

influence soil water movement and final nitrogen leaching. Thus, SLDR was also selected.

Considering both the influence on yield and nitrogen leaching, SLDR, SLRO, SLPF, SLLL, and

SDUL should be selected.

In general, genotype parameter ofPl, P5 and PHINT and the soil parameter of SLDR,

SLRO, SLPF, SLLL, and SDUL should be selected based on the results of restricted OAT

method (Table 2-7).

If comparing the selected parameters listed in Table 2-4 and 2-7, it is interesting to that the

two methods almost share the same results, except for parameter SSAT. Since SSAT was so

highly correlated with other sensitive parameters such as SLLL and SDUL, this parameter should

be selected. Finally, the input parameters listed in Table 2-4 were used for future research in

GLUE simulation.

2.4 Summary and conclusions

In this research, the non-restricted and restricted OAT methods were used to conduct

global sensitivity analysis for the CERES-Maize model. The outputs of concern were dry matter









yield and accumulative nitrogen leaching, because they are the two main factors for BMP

development in this study (Chapter 5). Some conclusions were drawn as follows.

Genotypes parameters P1, P5, PHINT and soil parameters SDUL, SLLL and SLPF have

strong influence on dry matter yield. The mean values of absolute elementary effect of Pl, P5

and PHINT to dry yield were 14.50, 2.14, and 1.95, while other genetic parameters all have a

mean value less than 1.0. This means that genotype parameters P1, P5, and PHINT had 2 to 14

times more influence on dry matter yield than the other genotype parameters. The mean values of

absolute elementary effect of SDUL, SLLL, and SLPF were 4.46, 2.07, and 1.56, while the

values of other soil parameter are all less than 1.0.

Genetic parameters P5 and PHINT and soil parameters SDUL, SLLL, and SLRO have

strong influence on nitrogen leaching. The mean values of absolute elementary effect of P5 and

PHINT to nitrogen leaching were 1.61 and 2.19, while other genotype parameters all have a

mean value less than 1.0. This means that genotype parameters P5 and PHINT had 1.6 to about 2

times more influence on dry matter yield than the other genotype parameters. The mean values of

absolute elementary effect of SDUL, SLLL, and SLRO to nitrogen leaching were 7.63, 4.51, and

2.65, while other soil parameters all have a mean value less than 1.0.

Soil parameters SLLL, SDUL and SSAT were highly correlated to each other. The

covariance coefficient between SLLL and SDUL was 0.935, while the coefficient between SLLL

and SSAT, SDUL and SSAT were 0.576 and 0.647.

Nine parameters were selected for future model calibration with GLUE method (Chapter 3).

They were P1, P5, PHINT, SLDR, SLRO, SLPF, SLLL, SDUL and SSAT. Genotype parameters

P1, P5, and PHINT and soil parameters SLLL, SDUL, SLDR and SLRO were selected because

they have highest mean values of absolute elementary effect corresponding to either dry yield or









nitrogen leaching. The soil parameter SSAT was selected because it was highly correlated with

sensitive parameters SDUL and SLLL. Though the mean and variance of elementary elements to

nitrogen leaching of SLDR were lower than SLLL, SDUL, and SLRO, the values were still

considerable, much greater than other rest soil parameters. And the SLDR represents the soil

drainage coefficient, while soil drainage will definitely influence soil water movement and final

nitrogen leaching. Thus, SLDR was also selected.











Zi(1,1) Z(1,3) Zi(1,4) ... ... ... ... Zi(1,k)
(Min.) (Max.)







ZI(O) Z2(0) Z3(0) Z4(0) ... ... Zi= ... ... Zs(0)
Zi(1,1)


Figure 2-1. Scheme of non-restricted OAT method. Zi is the ith input parameter. Zs(O)
represents the nominal value of input parameter Zs.


Z1(1,1) Z2(1,1) Z3(1,1) Z4(1,1) Zs(1,1)
(Min) (Min.) (Min.) (M in.) '" 0IM,11n .
Z1(2,1) Z2(2,1) Z3(2,1) -- Z4.2 1) --- Zs(2,1)

Z1(3,1) Z2(3,1) Z3(3,1) Z4(3,1) "." Zs(3,1)
Z1(4,1) Z2(4,1) Z3(4,1) Z1(4,1) --- Zs(4,1)










Max. Max. (Max.1 Ma (ax.





Z1(2,1) Z2(4,1) Z3(1,1) Z4(2,1) ... Zs(k,1)

Figure 2-2. Scheme of restricted OAT method. Z is an input parameter. Zs(k, 1) represents the value of
the element (k,1) of the space matrix of input parameter Zs. R is the random integer
following uniform distribution and used as address for nominal value selection in the
domain of the input parameter.





















ci,


ci,

C
*o


-1.00 -0.80 -0.60 -0.40 -0.20 0.00 0.20 0.40 0.60 0.80 1.00
Normalized Genotype Parameters

P1 P5 G2 G3 PHIN -- P2

Figure 2-3. Response profiles of sweet corn yield to six normalized genotype parameters









10000


S.----4000.................................................----






2000


2000




-1.00 -0.80 -0.60 -0.40 -0.20 0.00 0.20 0.40 0.60 0.80 1.00
Normalized Soil Parameters

SALB --SLU1 SLDR SLRO -w-SLPF SLLL -i-SDUL SSAT SBDM

Figure 2-4. Response profiles of sweet corn yield to nine normalized soil parameters











63


I -UU


10000









4000


2000
I I I /' I I


























100


50
--.--...........--------_ ---=




-1.00 -0.80 -0.60 -0.40 -0.20 0.00 0.20 0.40 0.60 0.80 1.00
Normalized Genotype Parameter

-P1 -P2 P5 G2 -G3 -PHIN

Figure 2-5. Response profiles for the nitrogen leaching to six normalized genotype parameters






250



200



150
C-)








0 -- -0 -,,,,III 0-----
!= ----- 'fc---------------------------------------I






-1.00 -0.80 -0.60 -0.40 -0.20 0.00 0.20 0.40 0.60 0.80 1.00
Normalized Soil Parameter

-SALB --SLU1 SLDR SLRO -w-SLPF -SLLL -i-SDUL -SSAT SBDM

Figure 2-6. Response profiles for the nitrogen leaching to nine normalized soil parameters











64


~rn


I. .










Table 2-1. Genotype coefficient for the DSSAT CERES-Maize model
No. Parameter Definition
1 P1 Degree days (base 8 C) from emergence to end of juvenile phase
2 P2 Photoperiod sensitivity coefficient (0-1.0)
3 P5 Degree days (base 8 C) from silking to physiological maturity
4 G2 Potential kernel number
5 G3 Potential kernel growth rate mg/(kemel d)
6 PHINT Degree days required for a leaf tip to emerge (phyllochron interval) (C d)












Table 2-2. Covariance coefficient matrix of genotype and soil parameters of the DSSAT model
P1 P2 P5 G2 G3 PHINT SALB SLU1 SLDR SLRO SLPF SLLL SDUL SSAT SBDM
P1 1.000 0.359 0.354 -0.276 -0.135 0.227 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
P2 0.359 1.000 -0.078 -0.270 0.059 0.337 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
P5 0.354 -0.078 1.000 -0.189 -0.453 0.141 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
G2 -0.276 -0.270 -0.189 1.000 0.183 -0.029 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
G3 -0.135 0.059 -0.453 0.183 1.000 -0.031 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
PHINT 0.227 0.337 0.141 -0.029 -0.031 1.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000 0.000
SALB 0.000 0.000 0.000 0.000 0.000 0.000 1.000 -0.314 0.268 -0.063 -0.203 -0.482 -0.557 -0.439 0.063
SLU1 0.000 0.000 0.000 0.000 0.000 0.000 -0.314 1.000 0.044 0.485 0.119 0.275 0.323 0.312 -0.171
SLDR 0.000 0.000 0.000 0.000 0.000 0.000 0.268 0.044 1.000 -0.153 -0.082 -0.187 -0.156 -0.271 0.026
SLRO 0.000 0.000 0.000 0.000 0.000 0.000 -0.063 0.485 -0.153 1.000 -0.015 0.242 0.273 0.238 -0.019
SLPF 0.000 0.000 0.000 0.000 0.000 0.000 -0.203 0.119 -0.082 -0.015 1.000 0.156 0.194 0.224 -0.119
SLLL 0.000 0.000 0.000 0.000 0.000 0.000 -0.482 0.275 -0.187 0.242 0.156 1.000 0.935 0.576 -0.328
SDUL 0.000 0.000 0.000 0.000 0.000 0.000 -0.557 0.323 -0.156 0.273 0.194 0.935 1.000 0.647 -0.352
SSAT 0.000 0.000 0.000 0.000 0.000 0.000 -0.439 0.312 -0.271 0.238 0.224 0.576 0.647 1.000 -0.471
SBDM 0.000 0.000 0.000 0.000 0.000 0.000 0.063 -0.171 0.026 -0.019 -0.119 -0.328 -0.352 -0.471 1.000










Table 2-3. Criteria for input parameter determination


High Correlation Low Correlation No Correlation


High Sensitivity + + +
Low Sensitivity + -
No Sensitivity
a +=Accepted, -=Rejected; High Sensitivity=High relative sensitivity;
High Correlation =High correlation coefficient to high sensitive parameters, defined as greater than 0.5 in this
research.



Table 2-4. Selected parameters for GLUE simulation based on the non-restricted OAT method
and covariance coefficient matrix
P1 P5 PHINT SLPF SLDR SLRO SDUL SLLL SSAT
Parameter 3 3 3 3 3 3
Paramet Cd Cd Cd cm3/cm3 cm3/cm3 cm3/cm3
Value 225.10 763.60 41.20 0.96 0.46 73.00 0.25 0.13 0.38
a Cd means degree day.




Table 2-5. Mean and variance of absolute elementary effects of genotype parameters
Parameter Unit Dry Yield Nitrogen Leaching
Mean Variance Mean Variance
P1 Cd 1.71 4.14 0.67 4.15
P2 -0.22 3.68 0.19 2.18
P5 Cd 2.15 19.14 1.61 13.54
G2 0.73 20.29 0.84 10.88
G3 mg day' 0.96 20.92 0.85 10.67
PHINT Cd 1.96 25.56 2.20 11.80



Table 2-6. Mean and variance of absolute elementary effects of soil parameters
Parameter Unit Dry Yield Nitrogen Leaching
Mean Variance Mean Variance
SALB 0.03 0.01 0.22 3.05
SLU1 0.06 0.09 0.24 3.11
SLDR 0.36 1.61 0.88 8.28
SLRO 0.26 2.19 2.65 30.11
SLPF 1.56 2.74 0.25 2.16
SLLL m3/m3 2.07 18.92 4.51 35.82
SDUL m3/m3 4.46 333.58 7.63 36.26
SSAT m3/m3 0.12 0.55 0.34 4.04
SBDM g/cm3 0.00 0.00 0.00 0.00



Table 2-7. Selected parameters for model calibration based on the restricted OAT method
P1 P5 PHINT SLPF SLDR SLRO SDUL SLLL


Parameter


u u u /C cm /cm
Value 225.10 763.60 41.20 0.96 0.46 73.00 0.25 0.13


S3 3 3 3


t 1 t 1









CHAPTER 3
PARAMETER ESTIMATION FOR CERES-MAIZE MODEL WITH THE GLUE METHOD

3.1 Introduction

3.1.1 Parameter Estimation

Proper estimation of model parameters is required for ensuring accurate model predictions

and good model based decision rules (Makowski et al., 2002). If a crop model was considered as

an equation system, the parameters could be considered as the unknowns and the observed data

could be considered as the constants. The process of model parameter estimation could be

considered as the process of solving the equation system. Since crop models usually have many

parameters, there are often more parameters than the number of observations and the number of

equations is smaller than the number of the unknowns. Thus, it is generally numerically

impossible to estimate all the parameters of the crop models. On the other hand, crop models are

based on equations that describe the processes involved in crop growth and development, and

generally there is limited information about these processes in model application. For example,

information might be needed about the thermal time to flowering, which can only come from

controlled environmental experiments, or information about maximum rate of root elongation

from specific experiments on this aspect of crop growth. Thus the problem of parameter

estimation for crop models is not a straightforward regression problem. Rather the problem is

using both field data and information about growth and development to estimate model

parameters (Makowski et al., 2006).

The objective of parameter estimation is to determine the values or range of values of

model parameters. A parameter is a numerical value that is not a measured or observed input

variable (Makowski et al., 2006). The same quantity may or may not be a parameter depending

on circumstances. For example, initial soil mineral nitrogen may be measured, in which case it is









an input variable. In other cases it may not be measured, in which case it is a parameter that has

to be estimated.

It is useful to distinguish two approaches of parameter estimation, the frequentist and the

Bayesian. The frequentist uses estimation methods to approximate the true parameter values 0 by

using only a sample of data. For the frequentist, parameters are not random variables but are

fixed. Prior information on parameter values are not taken into account. Different types of

frequentist methods (maximum likelihood, least squares, etc) were developed in the 1920s and

1930s by R.A. Fisher, J. Neyman, and E. Pearson notably (Makowski et al., 2006). The

application of a frequentist method to a particular dataset gives a point estimate of the model

parameters and the function that relates point estimates to datasets is called an estimator

(Makowski et al., 2006).

The Bayesian method estimates parameters from two different types of information, a

sample of data (like the frequentist) and prior information about parameter values. The result of

the application of a Bayesian method is a probability distribution of parameter values. All

Bayesian methods proceed in two steps. The first step is to define a parameter probability

distribution based on literature or expert knowledge. This distribution is called prior parameter

distribution and reflects the initial state of knowledge about parameter values. The prior

distribution can be, for example, a uniform distribution with lower and upper bounds derived

form expert knowledge or a normal distribution. The second step consists in calculating a new

parameter probability distribution from both the prior distribution and the available data, such as

observed crop yields, soil moisture, and biomass nutrient concentration etc. This new distribution,

called posterior distribution, is computed by using the Bayes theorem. The posterior distribution

can be used in different ways. Point estimates of parameters can be taken as the expected value








or, alternatively, the mode of the posterior distribution. The posterior parameter distribution can

also be used for generating the probability distribution of the model outputs, for instance, the

distribution of yield (Makowski et al., 2006).

Bayesian methods are becoming increasingly popular for estimating parameters for

complex mathematical models (e.g. Campbell et al., 1999), because this approach provides a

coherent framework for dealing with uncertainty. This is also due to the increase in the speed of

computer calculation and the recent development of new algorithms (Malakoff, 1999).

The principle is to start with a prior probability distribution of the model parameters whose

density is noted P(O). This prior distribution describes our belief about the parameter values

before we observe the set of measurements Y. In practice, P(O) is based on past studies, expert

knowledge, and literature. The Bayesian methods then tell us how to update this belief about

using the measurements Y to give the posterior parameter density P(O I Y) (density of0 based on

the dataY ). What we now believe about is captured inP( I Y).

The posterior parameter distribution is given by Bayesian theorem, as shown in following

Equation (3-1):

( P(Y I ) P(0) (3-1)
P(P | Y)= (3-1)
P(Y)

where Y is the vector of measurements, 0 is the parameter, P(0)is the prior distribution of

the parameter, P(O I Y) is the posterior distribution of parameter based on measurement Y, P(Y)

is a constant of proportionality determined by the requirement that the integral ofP(0 Y) over the

parameter space equals 1, and P(Y I 0)is a likelihood function of Y under prior distribution of 0.

The likelihood is the probability of the dataY given the parameterO. Its value is determined from

the probability distribution of the errors between modeled and observed data. It is readily seen









that both the prior distribution and the new data affect the posterior parameter distribution

(Makowski et al., 2002).

The Bayesian method has several advantages: (1) a parameter can be estimated from

different types of information (data, literature, expert knowledge); (2) the posterior probability

distribution can be used to implement uncertainty analysis methods; and (3) the posterior

probability distribution can be used for optimizing decisions in face of uncertainty.

3.1.2 GLUE Method

Due to the complexity of agronomic models (non-linearity and a large number of

parameters), it is almost impossible to directly calculate analytically the posterior parameter

distribution. However, the growing power of computers and the development of new methods of

numerical calculation make the Bayesian approach accessible even with complex models

(Campbell et al, 1999; Franks, et al., 1998; Harmon and Challenor, 1997; Malakoff, 1999).

One increasingly popular method is the Generalized Likelihood Uncertainty Estimation

method (GLUE) (Beven and Binley, 1992; Franks et al., 1998; Shulz et al., 1999). The principle

of this method is to discretize the parameter space by generating a large number of parameter

values from the prior distribution. Likelihood values are then calculated at each parameter value

with likelihood functions and measurements. Weights or probabilities are calculated with the

Bayesian equation. Finally, posterior distribution of the parameter was estimated with the

weights.

The GLUE method is based on the equifinality concept. In the study of Beven and Freer

(2001), it is argued that, given current levels of understanding and measurement technologies, it

may be endemic to mechanistic modeling of complex environmental systems that there are many

different model structures and many different parameter sets within a chosen model structure that

may be behavioral or acceptable in reproducing the observed behavior of the system. Indeed, to









focus attention on a rejection of the concept of the optimal model in favor of multiple

possibilities for producing simulations that are acceptable simulators in some sense, this idea has

been called equifinality (Beven, 1993).

One implication of rejecting the concept of a single optimal parameter set and accepting

the concept of equifinality is that the uncertainty associated with the use of a model in prediction

might be wide, since if there are several different acceptable model structures or many

acceptable parameter sets scattered throughout the parameter space, all of which are consistent in

some sense with the calibration data (Beven and Freer, 2001).

This appears to lead quite naturally to a form of Bayesian averaging of parameter sets and

predictions, in which prior distributions of parameters are assessed in terms of some likelihood

measure relative to the observations and a posterior distribution is calculated that can then be

used in prediction. This is the basis of the GLUE methodology proposed by Beven and Binley

(1992).

The principle of the GLUE methodology (Beven and Binley, 1992) is to approximate the

posterior parameter distribution P(O9 Y) by a discrete probability distribution (0,, p,), i = 1,..., N,

N
p, = 1, where p, is the probability associated with the parameter vector ,, and N is the total


number of available parameter vectors. The procedure of this method was described in the paper

of Makowski (2002).

The objective of this current research is to use the GLUE method to estimate the genotype

and soil parameters of the CERES-Maize component of the DSSAT model under sweet corn

(Zea mays L.) production in North Florida.









3.2 Method and Materials


3.2.1 Field Experiment

In this study, field experiments were necessary to conduct the GLUE simulation. First, the

CERES-Maize model needs fundamental input data for model run, such as weather data, planting

date, irrigation depths and timing, nitrogen fertilizer application rates and timing, and other

management information. Second six kinds of field observation data, such as yield, anthesis date,

maturity date, corn leafN concentration, soil moisture, and soil nitrate concentration, were

required to compare the simulated and measured outputs so as to calculate the likelihood values.

The sweet corn field experiments were conducted at the Plant Science Research and

Education Unit, the University of Florida in the spring of 2005 and 2006. The unit is located near

Citra (29.4094N, 82.1777W, 20.746 meters above sea level), Marion County, Florida. The

experiment field was identified as Blockl and the variety of sweet corn planted was Saturn SH2.

In this study, the data collected in the experiment field Block 1 (Figure 3-1) were used for

the GLUE process for model parameter estimation. In Block 1, there were only two treatments

each year, the high-nitrogen-level treatment and the low-nitrogen-level treatment, while the

irrigation level was the same. The size of Block 1 was about 9.0 acres and divided in half for

these two treatments.

The soil of the experiment field is coarse and is mapped as Lake Sand, Candler Variant,

Tavares Variant, and Millhopper Variant 1 etc., which mainly belong to Quartzipsamments

(Entisol). Soil characterization was done for 24 sites at 3 depths of 0-15 cm, 15-30 cm, and 30-60

cm. The samples were analyzed at the Soil and Water Science Department of the University of

Florida.The permanent wilting point (PWP) was measured as the soil moisture at a soil pressure

of 15.3 bar, field capacity (FC) as the soil moisture at 0.1 bar, and soil saturation as the soil

moisture at Obar. In this study, the small soil core method was used to measure the values of









PWP and FC (Klute, 1986). This method requires soil sampler and core rings for obtaining

undisturbed soil cores, pressure plate apparatus or similar device, moisture cans, balance, drying

oven, and spatula etc. The procedures of soil core analysis can be found in "Methods of Soil

Analysis Part 1: Physical and Mineralogical Methods" (Klute, 1986). The main measured

properties of the soil at the experiment site are summarized in Table 3-1.

The nitrogen fertilizer used in the experiment was a composite of several nitrogen

compounds. The total nitrogen mass concentration was about 32%, including 7.9% nitrate

nitrogen, 7.9% ammoniacal nitrogen, and 16.2% urea nitrogen. The density of the fertilizer

solution was 1.294 kg L-1, while the concentration of nitrogen in this solution was 0.414 kg N L1.

In the field experiments of 2005 and 2006, soil and biomass samplings were conducted to

evaluate the nitrogen status in soil profile and corn tissue. Yield sampling was conducted at

harvest to determine the final yield of each treatment.

Soil sampling was roughly at biweekly intervals during the growth season according to

sampling positions in the field map (Figure 3-1). Soil samples were collected in each of the eight

locations (W1 through W4, El through E4 as described in Figure 3-1) at 4 depths of 0-15 cm,

15-30 cm, 30-60 cm, and 60-90 cm. The samples were analyzed at the Department of Soil and

Water Science University of Florida for KCL extractable nitrate and ammonium concentrations

and moisture content as well.

Gravimetric soil moisture content was determined by calculating the ratio of mass of water

to that of the dry soil. The mass of water is the difference between wet soil sample and the dry

soil sample. Traditionally, the most frequently used definition for a dry soil is the mass of a soil

sample after it has come to constant weight in an oven at a temperature between 100 and 110 C.

Then the gravimetric soil moisture content (6d ) can be converted to the volumetric soil moisture









content (0vb ) by use of the formula of0vb = (Pb / p)dw, where Pb is the bulk density of the soil,

and pw is the density of water (Klute, 1986).

The analysis of soil nitrate and ammonium concentration included two main procedures: (1)

extraction of exchangeable ammonium, nitrate; and (2) determination of nitrate and ammonium

concentration with colorimetric method (Page et al, 1982).

The procedure of extraction is described as follows. Place 3 g of soil in a wide-mouth

bottle, and add 30 ml of 1M KC1. Stopper the bottle, and shake it on a mechanical shaker for 1

hour. Allow the soil-KCl suspension to settle until the supernatant liquid is clear (usually about

30 min). Then use a vacuum filter with a pore size of 0.45 pm to filter the solution.

Nitrate and ammonium concentrations were measured by colorimetric methods. The

special apparatus required for nitrate concentration determination was Rapid Flow Analyzer

(RFA), ALPKEM 300 Series (OI Corporation, College Station, TX). The apparatus for

ammonium concentration determination was Technicon Industrial Method AA II (Technicon

Instrument Corporation, Tarrytown, NY).

Biomass sampling was conducted at the eight locations close to the soil sampling. The

sampling frequency was also once every two weeks. In each sampling, a whole plant that had an

average height in the sampling area was collected. The sample was then processed and analyzed

in the lab. The analysis of the plant samples included measurement of the moisture and total

Kjeldahl nitrogen (TKN) of different plant parts. Each plant was divided into leaves, stems,

husks, cobs, and kernels, then weighed wet. Plant roots were not considered here, because of the

negligible amount of nitrogen in the roots (Albert, 2002).









Fresh mass of each biomass sample was measured first. Then the samples were dried in the

oven for 48 years at a constant temperature of 60 C for 48 hours. The dry mass of each sample

was measure. Then the biomass moisture was calculated.

The Kjeldahl procedures generally employed for determination of total N involve two

steps: (1) digestion of the sample to convert organic N to NH4+-N, and (2) determination of

NH4+-N in the digest. The digestion is usually performed by heating the sample with H2S04

containing substances that promote oxidation of organic matter and conversion of organic N to

NH4+-N. The substances generally favored are salts such as K2SO4 or Na2SO4, which increase

the temperature of digestion, and catalysts such as Hg, Cu, or Se, which increase the rate of

oxidation of organic matter by H2S04 (Page et al, 1982). In this study, the substances were

K2S04 and CuSO4. The determination of NH4+-N in the digest was conducted with colorimetric

method in the Analytical Research Laboratory (ARL), Institute of Food and Agricultural

Sciences, the University of Florida.

Yield sampling was conducted at the end of the experiment season at a sweet corn

physiological maturity. This date was about 70 to 80 days after planting. Ears in a sampling zone,

which consisted of a 6.1 m (20 feet) section of two rows near each of the eight sampling location,

were completely collected whether the kernels were fully filled or not. The total plant numbers in

this zone were also counted. Then the collected ears were weighed and classified into three

classes, US #1, US #2, and Cull according to the USDA sweet corn classification standards

(USDA, 1962).

The dates and methods of planting, tillage, irrigation, fertigation, pesticide and herbicide

application, and harvest were collected. Some critical dates for sweet corn growth, such as

tasseling, silking, physiological maturity, were also recorded. They were important management









data. These data were primarily obtained from in situ observations made by the managers of the

farm.

3.2.2 Main Procedure of GLUE

In general the main procedures of the parameter estimation with the GLUE method are as

follows:

* Select the input parameters for estimation with the GLUE method;

* Determine the prior distributions of the selected input parameters;

* Determine the number of model runs, N ;

* Randomly generate N vectors 0, i = ,..., N, from the prior parameter distribution P(O);

* Run the model N times with the N generated parameter vectors 0, ;

* Calculate the likelihood values P(Y o), associated with the different generated parameter
vectors 0 ;

P(Y 0,)-P(0,)
* Calculate the probability p, with p, = ( P(
SP(Y 0,).-P(0,)
j=1

* Use the pairs (0, p, ), i = ,..., N, to determine various characteristics of the posterior
distribution, such as mean, variance, and covariance of the input parameters.

3.2.3 Selection of Input Parameters

Complex dynamic crop models include many parameters. For example, the STICS model

(Brisson et al., 1998) includes more than 200 parameters. This problem is often called over-

parameterization. And in many crop models, it is impossible to estimate simultaneously all the

parameters because several parameters are unidentifiable due to the structure of the model

equations. Lack of identifiability occurs when several sets of parameters lead to the same model

prediction (Makowski et al., 2006).









A common practice is to select a subset of parameters, to estimate those parameters from

measurement data, and to set the others equal to predefined values. The implementation of this

approach requires one to decide which among all the parameters will be adjusted to the data and

to choose a method for estimating the values of the selected parameters. Four methods are

proposed for selecting parameters by Makowski et al. (2006) as follows: (1) selection based on

literature, (2) selection to avoid identifiability problems, (3) sensitivity analysis, and (4)statistical

choice of parameters to estimate.

In the current research, the sensitivity analysis method was used to select input parameters.

The principle method is to calculate a sensitivity index for each parameter and to select

parameters with high sensitivity index values. This method allows modelers to identify the

parameters that have a strong influence on the model output variables of interest. Only these

parameters will be estimated with the measured data and others are fixed to values provided by

the literature.

In Chapter 2, the restricted and non-restricted one-at-a-time (OAT) methods were used to

conduct global sensitivity of the input parameters (including soil and genotype parameters) to

model outputs (dry matter yield, kg ha-1 and accumulative nitrogen leaching, kg ha-1). These two

methods gave similar results of parameter selection. The selected parameters are specified in

Table 3-2.

3.2.4 Prior Distribution

In this current research, the parameter values in the database of the DSSAT model were

used to derive the prior distribution of input parameters. The form of distribution and the

statistical properties, such as the mean value, variance, maximum, minimum values of each

parameter were calculated with the available parameter values.









The normal distribution was considered as the first choice, because it is the most common

distribution. In addition, the statistical parameters, mean value (/p) and variance (02 ), are easy

to obtain. To determine whether the selected parameters follow normal distributions, a normality

test was conducted. The Jarque-Bera test (Judge et al, 1982) was used in this research. This test

evaluates the hypothesis the random variable x has a normal distribution with unspecified mean

and variance, against the alternative that x does not have a normal distribution.

3.2.5 Model Run with Generated Parameter Vectors

According to the results of normality test mentioned in Section 3.2.4, a multivariate normal

distribution was assigned to all selected parameters except for SLPF. SLPF was assigned a

uniform distribution of [0.7, 1.0].

A Matlab program titled "mvnrnd.m" (See Appendix C) was used to generate random

parameter vectors. The function R = MVNRND (MU, SIGMA, CASES) returns a matrix of

random numbers chosen from the multivariate normal distribution with mean vector, MU, and

covariance matrix, SIGMA. Here CASES is the number of rows in R, or the number of the

generated parameter sets. SIGMA is a square positive definite matrix with size equal to the

length of MU. Table 3-2 and 3-3 show the mean vector and covariance matrix of the prior

distribution of the selected parameters to run the function above. The mean vector and

covariance matrix were all obtained by calculating the mean values, variances, and covarriance

of the available parameter values in the database of DSSAT model.

For layered parameters, such as SLLL, SDUL, and SSAT, random values for each layer

had to be assigned. There are two ways to resolve this problem. First, each layer of a layered

parameter could be considered as an individual input parameter with individual random numbers









generated. However, this method would make the covariance matrix very large and difficult to

handle.

The second method reasonably assumes that there exist perfect correlations among the soil

layers. Then for each generated random number for layer 1 of an input parameter, there exists a

perturbation defined as follows:

xl, ul
E, = (3-2)
0-1

Where xl, is the ith generated random number for a soil property for layer 1, pl ando-1

are the mean and standard deviation of the soil property of layer 1. Then for the soil property of

layer 2, the ith random number, x2, can be calculated with equation (3-3).

x2, = /2+E, -c2 (3-3)

where p2 is the mean value, whileo2 is the standard deviation of the soil property of

layer 2. The same approach was used to calculate the input values for layer 3, layer 4, layer 5 etc.

In this current research, the soil profile was divided into five layers, as follows: 0-5 cm, 5-15 cm,

15-30 cm, 30-60 cm, and 60-90 cm.

With the method above, the values of the 5 layers of SLLL, SDUL and SSAT can be

generated with their own different perturbation values.

Then the model was run with these parameter vectors and the following outputs were

recorded: dry yield (HWAH, kg ha-1), anthesis date (ADAT, days after planting), maturity date

(MDAT, days after planting), cumulative nitrogen leaching (NLCM, kg ha-1), soil nitrate

nitrogen of four layers (mg g-1), soil moisture of four layers (%) and leaf total nitrogen content

(%).









3.2.6 Determination of Number of Model Runs

Determining an acceptable number of model runs is very important. Enough simulations

must be conducted to guarantee reliable statistical characteristics of the model input parameters

and the model outputs, but the amount of time needs to be considered at the same time. In theory,

more simulations were conducted, more reliable the statistical properties of the model outputs.

When enough simulations have been carried out, the means or standard deviations of the

generated parameter values and model outputs should all converge to constants. This occurs

when the means and standard deviations cease to change as the number of model runs continues

to increase. These criteria were used to determine the minimum number of model runs for this

research.

However, it should be noticed that the minimum number of model runs mentioned here is

not something to guarantee the reliability of the posterior distribution in the GLUE process for

parameter estimation, but only something to guarantee the reliability of model input parameters

and model outputs when starting the process. The number of model runs for a reliable posterior

distribution will depend on the range and distribution of the prior distribution, and the number of

observations involved. Generally, if the range of the prior distribution is wide, more model runs

will be required to increase the occurrence probability of the optimal parameter sets in the

smaller range that is more close to the actual values of the parameters. And if more observations

are involved in the GLUE process, more model runs will also be required because the occurrence

probability of the parameter sets that can optimize all observations will decrease.

3.2.7 Likelihood Function and Likelihood Value

3.2.7.1 Available likelihood functions

Based on the simulation and measurement results, the likelihood values of each parameter

vector 6, were calculated with a selected likelihood function. As with any calibration procedure,









the GLUE methodology requires the definition of some measure of goodness-of-fit, in this case

the likelihood measure, in comparing observations and predictions of the model. The likelihood

measure or the likelihood function must have some specific characteristics. It should be zero for

all simulations that are considered to exhibit behavior dissimilar to the system under study, and it

should increase monotonically as the similarity in behavior increases (Beven and Binley, 1992).

Several likelihood functions have previously been used in GLUE simulations by different

people. Some examples are introduced below:

(1) Likelihood function based on auto-correlated Gaussian error model (Romanowicz et al.,

1996):

L[YT, |,0,x] (2;U2) .(I a2)2 *exp (1/2u2){( +( -1 )
LL[Y, 2 2, JJ([ ?/

(3-4)

where = (,u, o, a) is the parameter vector; p,c,,a represent the coefficients of the

likelihood function; XT is model input; is the input parameter set; Y, is model output; 2 is

the variance of model prediction error; p is the mean of model prediction error; e, is the model

prediction error at different time steps; and r is the number of time steps in the simulation.

(2) Likelihood function based on inverse error variance with shaping factor N (Beven and

Binley., 1992):

L ([A( XT,0,=)] ( (3-5)

where M(e x, O,) indicates the ith model structure, conditioned on input XT and

observationO ; a-2 is the model prediction error variance; and N is shaping factor.

(3) Likelihood function based on Nash and Sutcliffe efficiency criterion with shaping

factor N (Freer et al., 1996):








( 2N
L[M(X,O,)]= 1-" for C

where M(e Ix,,o) indicates the ith model structure, conditioned on input X, and

observation O; o is the model prediction error variance; 02 is the observation variance; and

N is shaping factor.

(4) Likelihood function based on exponential transformation of error variance with shaping

factor N (Freer et al., 1996):

L[(O | XTO)] = exp(- No) (3-7)

where M(e Ixo,) indicates the ith model structure, conditioned on input XT and

observation O; C-2 is the model prediction error variance; and N is shaping factor.

(5) Likelihood function based on minimum mean square error (MSE) (Wang et al., 2005):

L[, ]= ex MSIE, (i = 1,2,3...N) (3-8)
Smmin(MSE))

where 0, is the ithparameter vector; 0 is the measured or observed value; MSE, is the mean

square model prediction error for theith parameter set; min(MSE) is the minimum value of

MSE,; and N is the number of parameter vectors for equation.

(6) Maximum likelihood function (Makowski et al., 2006):


L[O ]f= 1Y ----- exp ))2 (3-9)
z=1 2 r 2 o

where 0 is the parameter vector; 0 is the measured or observed value; o-2 is the variance

of the observations; 0, is the measured or observed value for model simulation scenario; ]" is

the corresponding value calculated by the model; and M is the number of observations.









The model prediction error was calculated with following equation:

S=0-P (3-10)

In this research, considering the normal distribution of the selected input parameters except

for SLPF and the availability of observations, another available likelihood function was defined

(Personal communication with Dr. Shrikant Jagtap, Department of Agricultural and Biological

Engineering, the University of Florida):

MSE
L(O, 10)== exp 2- (i = 1,2,3...N) (3-11)



MSE, -- P(O,)- 0 (3-12)
MJ=1

where 0, is the ith parameter set; MSE, is the mean square model prediction error for

theith parameter set; c- is the variance of observations; N is the number of parameter vectors;

M is the number of observation replicates; P(O,)is the single predicted value with input

parameter set 0,; and 0, is jth the replicate value of the observation.

The likelihood function (Equation 3-11) was derived from the commonly used maximum

likelihood function (Equation 3-9). Since the observation variance o2 was the same for the

different replicates of an observation, equation (3-9) can be rearranged as such:
M
S 1 (O -(, P_(0))2 1 (O P, (0))2 (3-13)
L[ | 0] ex 2 o2 -2 exp (3-13)
(T 20 2o


M1 (0, p, (0))2 1 3-14)
L[O 1 0] 2 2 = i exp 2- 2 2j ex2(
2 o









2( (M _- P ())2M
L[Oo 1 e (01 0 ())2 1 eXP MSE.M (3-15)
L[e\0\]= 1 exp -1 2 M = -1 exp 2-M (3-15)
2 2002 "M 2M 20 )


If want to find the parameter vector 0 that can maximize this rearranged maximum

likelihood function (Equation 3-15), it is equivalent to find the same parameter vector to

maximize the following equation (3-16), since the variance of observation o-2 and number of

replicates of the observation A, were constants for each model run. Thus the likelihood function

shown in Equation 3-19 can be simplified to the likelihood function shown in Equation 3-16.


L(O O)= exp 2_SE (3-16)


Equation (3-16) was the likelihood function for one parameter vectorO. When this

equation was used for every parameter vector, 0,, then:


L(O, 0)=exp 2-ME (i = 1,2,3...N) (3-17)


Equation (3-17) became the likelihood function previously defined in equation (3-11). The

procedure above shows how the likelihood function (Equation 3-11) was derived. This likelihood

function can also be considered as a variant of Equation 3-8, by replacing the min(MSE)

with 2"o2. When the real value of o2 is unknown, then min(MSE) is used as an estimation for

the observation variance. But when the value of o2 is available, it is better to use it directly

(Personal communication with Dr. James Jones, department of Agricultural and Biological

Engineering, the University of Florida).

The last likelihood function is (Personal communication with Dr. Wendy Graham,

Department of Agricultural and Biological Engineering, the University of Florida):










L[O 0]= exp ))2 (3-18)
2 20


O= Oj (3-19)
S=1

where 0 is the mean value of all replications of the observation. This likelihood function is

a variant of (3-9), which used the mean value of the observations instead of calculating the

product of several observations.

In this research, there were six observations from three types of field experiment results.

The first type was an integrated observation, which means there was only one observation value

in the entire crop growth season, such as yield at maturity (kg ha-1), anthesis date (days after

planting), and maturity date (days after planting). The second type was temporal variant

observations, which had several observation values at different days during the growth season,

such as leaf nitrogen concentration (%). The third type was both temporal and spatial variant

observations, which have different observation values in different soil layers and on different

days, such as soil nitrate content (mg g-l) and soil volumetric moisture content (%) in four soil

layers.

A method of combining the individual likelihood values was required, for the general case

of multiple sites or types of observations contributing to an overall likelihood weight for each

simulation. There are also a number of different methods for doing this. Examples of likelihood

measure combination equations (before renormalization) are listed below:

(1) Bayes' multiplication (e.g. Beven and Binley, 1992; Romanowicz et al., 1994, 1996):

L[M(O,)] oc L, [M(O,)]- L, [(0, Y1O, )] (3-20)








where M(O1, Y1,O1) indicates theith model simulation results, conditioned on a new value

O1, of parameter 0,, input data Y,, and observationO1; L, M([i, Y
of M( 1, | Y1 O1); L, [M(,)] is the prior likelihood value of model prediction conditioned on

parameter 0,. The posterior likelihood value of parameter O, L[M(O,)], was defined

proportional to the production of the LI M(I, I Y
(2) Weighted addition (e.g. Zak et al., 1997):

L[M(O,)]x cc ,L,[M(O,)]+ CdL, M(, I Y,0)] (3-21)

where vT and 7l are weighting coefficients for different periods or different variables;

M(, | Y1,O1) indicates theith model simulation results, conditioned on a new value O1, of

parameter O,, input data Y1, and observation 1; L, [M(O, I Ol)] is the likelihood value of

M(1, | Y1,O1); L [M(O,)] is the prior likelihood value of model prediction conditioned on

parameter 0,. The posterior likelihood value of parameter O, L[M(O,)], was defined

proportional to the weighted sum of the L [M( O, I Y,01)], and L0[M(O,)].

(3) Fuzzy union, fuzzy intersection, weighted fuzzy combination (e.g. Aronica et al., 1997):

L[M(O,)] il,[I,,[M(O,)LI [, (O YO, 01 (3-22)

L[M(O,)]ccMax[Lo[M(O,)L, [(O, Y,OY, 0 (3-23)

L[M(O,)]x ,ioc [,,[M(O, )1L, [(O1, Y,1)O + +,l)Max[L [M(O, )1 L, [M( 1 Y,,01)

(3-24)

where vT and i7 are weighting coefficients for different periods or different variables;

M(6, I Y10,1) indicates theith model simulation results, conditioned on a new value 1, of









parameter O,, input data Y[, and observation O; L1, [M(, Y, O)] is the likelihood value of

M(O,, | Y ,O1); L0 [M(,)] is the prior likelihood value of model prediction conditioned on

parameter O,. In Equation (3-22), the posterior likelihood value of parameter O, L[M(,)], was

defined proportional to the minimum value among L, [M(1, YO)] and Lo[M(,)]. In

Equation (3-23), L[M(,)], was defined proportional to the maximum value among

LZ M(1, I Y1,0)] and Lo[M(O,)]. And in Equation (3-24), L[M(O,)], was defined proportional

to the weighted sum of minimum and maximum value among L [M(( Y1,01)] and Lo [M(,)].

(4) Aggregated function suggested by Wang et al. (2005):

K 1/2
Lcombined= [ W L(O O )2
k=1

K
W = 1 (3-25)
k=1

where Lcombined is the combined likelihood value of parameter vector0; L(O, Ok) is the

likelihood value derived from observation Ok; K is the number of observation types; and W, is

the weight of the likelihood value. The total sum of W should equal 1.

3.2.7.2 Selection of likelihood function and method of likelihood value combination

From the descriptions above, it can be seen that many likelihood functions and methods of

likelihood value combination exist. However, to determine the best one that can reduce the

parameter uncertainties most significantly and give the best outputs, likelihood functions and

likelihood combination methods were investigated.

As discussed in Section 3.2.2, the input parameters follow a multivariate normal

distribution. So it is reasonable to choose the likelihood functions that are derived from normal








distribution. In addition, the availability of observation data should also be considered. Hence,

four types of likelihood functions were chosen and investigated with the same model outputs.

The four likelihood functions, identified as L1 (Equation 3-9), L2 (Equation 3-18), L3 (Equation

3-8) and L4 (Equation 3-11), are as follows:


L[O, 0]= 1 exp '(i= 1,2,3...N) (L1)
J=1 2102o
S2-o 2P(0, )- .


L[ 0]= exp P(O)2 (i= 1,2,3...N) (L2)
J22 7o 2 _o


L[O, 0]= exp- MSE1 ,(i =1,2,3...N) (L3)
Smin(MSE)

MSE
L(O, O)= exp 2-SE (i= 1,2,3...N) (L4)
S2c J

where 0, is the ith parameter set; P(O,)is the model output under parameter set 0,; O is the

observation; O0 is the jth replicate of 0; (7 is the variance of observations; O is the mean

value of the observation replicates; MSE, is the mean square model prediction error for the ith

parameter set; min(MSE) is the minimum value of MSE, ; N is the number of parameter sets;

and M is the number of observation replicates.

Another factor is the method of likelihood value combination that integrates the likelihood

values derived from different observations (dry matter yield, anthesis date, maturity date, leaf

TKN concentration, soil nitrate concentration, and soil volumetric moisture) together. Three

types of methods, identified as Cl, C2 and C3 respectively, were investigated:









K
Z L,[M ( Y, )]
Lcombined 1 (C 1 )
K

Lcombned = L,[M( Y, O)] (C2)
1=1


Leombned K L, [M(O I Y,Of1 (C 3)
L F; -1 72

where Lcomned is the combined likelihood value; M(o | Y, O) indicates the model

simulation results, conditioned on parameter 0, input data Y, and observation O;

L, [M(O Y, O)] is the likelihood value calculated from different observations; and K is the

number of observation types.

Equation (Cl) is a special case of the combination function (3-21), where the weighting

coefficients of all terms were equally set as l/K. Equation (C2) is the same as the combination

function (3-20). Equation (C3) is a special case of combination function (3-25), where the

weighting coefficients were all set as / K. In each equation, K is the total number of likelihood

values derived from different observations.

As described previously, there were three types of observations in this study. For the

integrated observation, K is just the number of observation types. For example, if only consider

the dry matter yield, anthesis date, and maturity date, the value of Kwas three since only three

kinds of integrate observations were considered.

However, for temporally variant observation, such as leaf nitrogen concentration, there

were five observations at five different dates during the growth season. It is necessary to

calculate some kind of combined likelihood value of for leaf nitrogen concentration first with the

methods described above before combining it with other likelihood values. In this case, the value









of K equals five, the number of observations in the growth season. The both temporally and

spatially variant observations, such as soil nitrate concentration and water concentration, could

also be handled to calculate some combined likelihood values first.

After each of the six types of observations has an individual likelihood value, the final

combined likelihood value can be calculated with the methods described above.

3.2.7.3 Comparison of distributions of input parameters

A 4 x 3 complete factorial (four likelihood functions and three methods of likelihood

combination) experiment design was used to find the best likelihood function and method of

likelihood value combination that can most significantly reduce the uncertainties in parameter

distributions and model outputs, where the likelihood function was considered as one factor,

while the method for likelihood value combination was another factor.

The best likelihood function and method of likelihood value combination should be the

ones that would have the lowest uncertainties or variances in the posterior distributions of model

input parameters.

3.2.7.4 Comparison of distributions of outputs

After comparing the uncertainties in the posterior distributions of model input parameters,

it was also necessary to compare the uncertainties in model outputs, because it was believed that

the best likelihood function and method of likelihood value combination could be the ones that

can produce outputs that are closest to observation data.

In this study, the mean values and standard deviations of different model outputs such as

yield, anthesis date, maturity date, and accumulative nitrogen leaching, which were obtained

from different posterior distributions derived from different likelihood functions and methods of

likelihood combination, were compared to the measured values in field experiment.









3.2.8 Estimation of Posterior Distribution

With the calculated likelihood values, the value of probability p, was calculated with

following equation:

L(O, O)
p(O,)= (3-26)
ZL(O, O)

In this study, all of the generated random parameter sets were involved when calculating

the probability. No parameter sets were truncated according to their combined likelihood values.

In some literature, the probability p, is also called likelihood weight (Wang et al., 2005).

There are many pairs (0, p, I i = 1,..., N) available, which describe the posterior distribution of 0

and then can be used to estimate expected value for each of the selected parameters. In addition,

the variance and covariance among parameters can be determined, with the following equation:

N
fipost = P( ), (3-27)
i=i

N
2 post = p(o,)-(o, i ), (3-28)
i=1


Cov(X,Y) = p(O,).(0 X-post -).(Y,- (3-29)
i=i

where pos ost, and Cov(X, Y) are the estimated mean value, variance, and covariance

between two parameters of posterior distribution. These three estimated statistical values will

help to reconstruct a new prior distribution for future research.









3.2.9 GLUE Simulation

Two rounds of GLUE process for model parameter estimation were conducted respectively

with the input variables (weather and field management) and observation data of field

experiments in Block 1 in the spring of 2005 and 2006.

In the first round of GLUE, the first posterior distribution of parameters was derived from

the calculated N pairs (08, p I i = 1,..., N), where 8, is the ith parameter set, p, is the calculated

probability of the ithparameter set, conditioned on the observation O. Then this first posterior

distribution was used as the new prior distribution for the second round of GLUE. Then a second

posterior distribution was obtained. This second posterior distribution was used for model

verification, development of best management practices, and uncertainty analysis in the rest part

of the dissertation.

The procedures of random number generation, model running, and result saving were all

automatically realized with Matlab programs (see Appendix C for details). The calculation of

likelihood values and derivation of posterior distribution were conducted with spreadsheets of

Microsoft Excel.

3.2.10 GLUE Verification

Though the soil and genotype parameters could be estimated with the GLUE method, the

reliability and accuracy of this method might still be suspect if without direct comparison

between the estimated parameter values and the really measured parameter values. Estimated

values of some soil parameters, such as SLLL, SDUL and SSAT, were compared with the field

measured ones to see whether they were correctly calibrated. However, for some parameters,

especially the genotype parameters, there was no experiment designed in this study to obtain

their measured values. To obtain more confidence in the GLUE method, in the selected









likelihood function, and in the method of likelihood value combination, a verification procedure

was conducted.

The steps of this verification procedure were as follows: (1) select a parameter set in the

second round of GLUE that gave the highest likelihood value; (2) run the model with this

selected parameter set under the field experiment condition in Block 1 in 2005 and record the

outputs; (3) use the variances of the observations (dry matter yield, anthesis date, maturity date,

corn leaf N concentration, soil moisture, and soil nitrate concentration) as variances and their

corresponding model outputs as mean values, to generate four replicates for each observation; (4)

conduct first-round GLUE with the generated replications of different observations, using the

prior distribution derived from DSSAT database; (5) run the model with this selected parameter

set under the conditions of 2006, and also make a record of the outputs; (6) repeat GLUE with

the first-round posterior distribution as the prior distribution; (7) compare the second-round

posterior distribution with the initially-selected parameter set (Personal communication with Dr.

Wendy Graham and James Jones, Department of Agricultural and Biological Engineering, the

University of Florida).

If the second-round posterior distribution could approach the initial parameter set very well,

which means the mean value of posterior distribution should be very close to the initial

parameter set, and the variance or uncertainty of the posterior distribution should be reduced to a

small level.Then it can be concluded that the GLUE method and the selected likelihood function

are reliable for input parameter estimation. It was assumed in this procedure that the observations

followed a normal distribution with given means and variances.

3.2.11 Expected Values of Posterior Distribution

In this study, the GLUE estimation was used for model calibration. The calibrated model

was used as a computer platform to explore different combinations of nitrogen fertilizer levels









and irrigation levels for sweet corn production in North Florida. Some of these combinations

were selected as potential BMPs (See Chapter 5). The selected potential BMPs were then tested

for their uncertainties caused both by weather uncertainties and parameter uncertainties (See

Chapter 6).

When exploring many possible treatments to find some potential BMPs, a nominal

parameter set is needed to carry out pre-selections of the BMPs. In other words, the prediction

uncertainties caused by input parameter uncertainties were temporarily neglected. Since a second

posterior distribution was already available after two rounds of GLUE estimations in this study,

it was reasonable to use the expectations of the distribution as the nominal values to run the

model. The expectations of the selected parameters were obtained with following equation:

N
E(0)= p(0 )x0, (3-30)
i=1

3.3 Results and Discussion

3.3.1 Results of Prior Distribution

Table 3-4 shows the results of normality test with the Jarque-Bera method for the selected

parameters. It can be seen that parameter SLDR, SLLL, SDUL, and SSAT all followed a normal

distribution under a significance level of 0.05. For parameter P1 and P5, though they failed to

strictly follow a normal distribution under a significance level of 0.05, they had larger p-values,

which means if they could follow a normal distribution if under a lower significance level, for

example 0.01. At the same time the covariance between the parameters were also considered, it

is reasonable to select some kind of distribution that can represent covariance. The multivariate

normal distribution would be a good choice. So parameter PHINT and SLRO were also assigned

a normal distribution.









SLPF completely failed to follow a normal distribution. It is a parameter that represents the

influence of micronutrients such as zinc (Zn) and copper (Cu). Its values were always set as 1

when its actual value is unknown. In this research, a uniform distribution of [0.7, 1.0] was

assigned for SLPF, where 0.7 and 1.0 were the minimum and maximum of SLPF, respectively.

Thus, finally except for SLPF, a multivariate normal distribution was used as the prior

distribution for the selected input parameters.

3.3.2 Results of Number of Model Runs

Enough number of model runs should be carried out to guarantee the statistical properties

of the input parameters, in other words, to guarantee the prior distribution. Thus, the stability of

the generated random values of the selected input parameters was tested to determine the number

of model runs. Each of the selected input parameters was tested. For convenience, only P1 was

used as an example for the genotype parameters. The results were shown in Figure 3-2 and 3-3.

Figure 3-2 shows the response curve of mean value of different numbers of generated

values of Pl vs. number of model runs or parameter sets, while Figure 3-3 shows the response

curve of standard deviations of Pl vs. number of model runs. From these figures, it can found

that after about 1,000 randomly generated values of Pl, the mean values of generated P1 would

reach a constant of 225, which was the mean value of Pl in the prior distribution (225.10). The

standard deviation of generated Pldistribution did not reach a constant of 68 until about 2,000

model runs, which is comparable to real value of standard deviation of 67.5.

The soil parameter SLRO was also used as a representative for the test of generated soil

parameter stability. Figures 3-4 and 3-5 show that the minimum reliable number of model runs

should be about 2,000, since additional parameters sets yielded little change in the values of

mean and standard deviation of SLRO.









Thus, it can be concluded that at least 2,000 randomly generated parameter sets or model

runs could guarantee the parameters follow the prior distribution.

Next, the stability of statistical properties of model outputs was tested. Figures 3-6 through

3-9 show the mean values and standard deviations of predicted yields (kg ha-1) and nitrogen

leaching (kg ha-1) under different numbers of model runs with generated random parameter sets.

From the four figures above, it can be seen that after about 3,000 model runs, the four

statistics all reached constant values. However, before 3,000, most notably before 1,000 runs,

values varied dramatically. It could be concluded that at least 3,000 simulations should be

conducted to generate reliable results. Thus, the number of 3,000 model runs was chosen since

this value satisfied both input and output stability. However, it should be noticed that the

minimum number of model runs mentioned here is not something to guarantee the reliability of

the posterior distribution in the GLUE process for parameter estimation, but only something to

guarantee the reliability of model input parameters and model outputs when starting the process.

3.3.3 Results of Likelihood Function and Method of Likelihood Value Combination

3.3.3.1 Comparison of distributions of input parameters

From Table 3-5, it can be seen that under any of the likelihood functions, L1, L2, L3 or L4,

the way to combine the likelihood values derived from different observations had a very strong

influence on the corresponding posterior distributions, since they influenced the standard

deviations of the posterior distributions to varying degrees.

For the combination methods of Cl and C3, the standard deviations or the uncertainties of

most input parameters did not decrease, and even increased in some cases. For example, under

L1 the standard deviation of Pl became 112.31 under Cl and 112.01 under C2, but the standard

deviation of Pl in the prior distribution was only 67.83. Similar trend occurred for P5, PHINT,

etc. Thus, likelihood combination methods Cl and C3 failed as a tool to reduce the parameter









uncertainties. This is because these two methods were not strict enough to eliminate some

parameter sets that simultaneously have extremely good predictions for some outputs and poor

predictions for the others. For example, one parameter set had gained a likelihood value of 0.9 in

predicting the dry matter yield, but only 0.0001 in predicting the maturity date, which means this

parameter set did a very good job in predicting the dry matter yield, but a poor job in predicting

the maturity date.Then under Cl and C3, the combined likelihood values would be 0.450 and

0.636, respectively. This parameter set might be selected when deriving the posterior

distributions of the input parameters, since the combined likelihood values were considerable.

However, under C2 the combined likelihood value would only be 0.00009, which was very small

and would probably be neglected when deriving the posterior distribution. If this parameter set

was selected, the ranges of the parameters that control corn yield might be refined, but the ranges

of the parameters that control maturity date might be coarser at the same time because more poor

parameter values were selected when contracting the posterior distribution. Finally, the

uncertainties of the input parameters could not be reduced.

Under C2, the results were much better than under Cl and C3. The standard deviations of

almost every selected parameter decreased. This method, which is defined as a factorial product,

had the most powerful ability to eliminate unsatisfactory parameter sets as mentioned above.

Thus in future research, C2 was used as the standard method to combine different likelihood

values. Under C2, there was no great difference among the posterior distributions derived from

L1, L2, L3 and L4, especially between L1 and L2, L3 and L4. For example, the mean value of

the prior distribution of P1 was 225.10. Then it changed to 144.49, 142.12, 166.20, and 166.45

under L1, L2, L3 and L4, respectively. The results were not surprising since the forms of the

likelihood function of L1 and L2 were similar, as were L3 and L4. The standard deviation of the









prior distribution of Pl was 67.83. Then it decreased to 23.39, 12.98, 38.04, and 37.49 under L1,

L2, L3 and L4 respectively. It can be seen that likelihood function L2 had the lowest value of

standard deviation or lowest uncertainty in the posterior distribution. However, this result may be

a little bit misleading, because L2 used the average observation O instead of the replicates of the

observation, in other words, this likelihood function under-represent the uncertainties in

observations.

In general, it might be concluded that the likelihood functions did not have dramatic

influence on the posterior distributions for the same method of likelihood value combination, if

the functions were reasonably defined and close to each other.

3.3.3.2 Comparison of distributions of model outputs

As shown in Table 3-5, there was some difference between the results from L2 and L3,

especially for genotype parameters P1 and P5. For example, the mean value of Pl under L1 and

L2 was 144.49 and 142.12, which were close to each other. However, it was 166.20 and 166.45

under L3 and L4, which was higher than L1 and L2. The same situation can be seen for P5.

These differences could heavily influence the length of growth stages and the dry matter

yield at last. Therefore when the best method of likelihood value combination, C2, was

determined, it was still necessary to find the likelihood function that was most efficient in the

GLUE procedure.

The model was run 3,000 times in a second round with the input parameter distributions

determined by L1C2, L2C2, L3C2 and L4C2 (Table 3-5). Then the outputs were compared with

the observed values to determine which likelihood function should be selected. The outputs are

listed in Table 3-6. To quantify the agreement between the outputs derived from different









likelihood functions and observations of field experiment, a measure called absolute relative

error (ARE) was defined as follows:


ARE=I (3-35)
Y

where Y is the measured value, and Y' is the model predicted value.

From Table 3-6, it is easy to see that L1C2 most precisely matched the observations,

especially in yield. The input parameter distribution of L2C2 over-predicted the yield, but had

the lowest mean value of nitrogen leaching of 73.32 kg hal-. For example, the ARE value of

yield of L2C2 compared with measured value is 0.15, which was higher than L1C2, L3C2, and

L4C2. No ARE value is available for nitrogen leaching, since there was no direct measurement

for it in this study.

However, the input parameter distribution of L3C2 and L4C2 underestimated the yield

with an ARE value of 0.03 and 0.05, but over-predicted the anthesis date and maturity date with

a substantial ARE value of 0.24. Therefore, it can be concluded that the L1C2 (likelihood

function 1 and likelihood value combination 2) should give the best results in the process of

parameter estimation with the GLUE method. Actually this is really the most theoretical

approach. However it was also obvious that L1C2 overestimated the anthesis date (about 5 days

longer) and maturity date (about 8 days longer).

The prior distribution was derived from the database of the DSSAT model. The range and

uncertainty in the prior distribution were very large. The occurrence probability of behavioral

parameter sets that were very close to the actual values was low. And six types of observations

were involved in the GLUE process, among which some were temporally and spatially variant.

Thus, it required the GLUE method to optimize multiple objectives simultaneously, which also

decreased the occurrence probability of behavioral parameter sets. In the first round of GLUE









estimation, the generated parameter sets were widely distributed. Most generated parameter sets

failed to have a considerable likelihood value. Only about 10 parameter sets, which had

considerable likelihood values, were finally selected to construct the posterior distributions.

Other parameter sets were automatically eliminated by the GLUE process.

However, it seems ten parameter sets were not very enough to construct the posterior

distribution. There were two possible ways to compensate this drawback. The first way was to

increase the number of model runs. For example, if the number of model runs could be increased

to ten times of the initial one, the number of behavioral parameter sets might also increase to ten

times. However, this method would also greatly increase the model running time. Another

method was to conduct a second round of GLUE process. After the first round of GLUE, the

range and variance of the posterior distribution of the input parameters would be decreased, the

occurrence probability of behavioral parameter sets with the new distributions would be

increased significantly. Consequently, the second posterior distribution would be more smooth

and precise. In this research, the second method was used.

In the second round of GLUE process, the first posterior distribution was used as the new

prior distribution. The observations of field experiment in 2006 in Block 1 were used to construct

the second posterior distributions. From this point on, the default strategy of likelihood function

and likelihood value combination will be L1C2. L1 is the likelihood function 1 as shown in

Equation (3-9), which is directly derived from the density function of normal distribution. C2 is

the method 2 of likelihood combination, which is based on Bayes' multiplication.

3.3.4 Distributions of Selected Parameters

Table 3-7 shows the statistical properties of the prior, the first, and the second posterior

distributions of the input parameters after two rounds of GLUE simulation using likelihood

function of L1 and method of likelihood combination of C2. By comparison, it can be seen that









the initial mean values of the selected parameters changed. The ranges defined by the minimum

and maximum values also became narrower. The values of standard deviations decreased, which

means the uncertainties of the input parameters were decreased dramatically. For example, the

initial mean value of genotype parameter of Pl was 225.10. It changed to 144.49 and then to

99.17 in the two rounds of GLUE simulation. The initial range was [110.00, 450.00], then the

range of first posterior distribution narrowed to [136.93, 216.80], and the range of second

posterior distribution continuously narrowed to [77.68, 118.22]. The standard deviation of Pl for

the prior, first posterior and second posterior distributions was 67.83, 23.39, and 8.22,

respectively. Similar changes could also be found in other parameters, either in genotype or soil.

The measured mean values in field experiment and estimated mean values with the GLUE

method of the soil parameters are given in Table 3-8. Interestingly, the mean values of estimated

and measured soil parameters were pretty close to each other. For example, the mean value of

calibrated SDUL in the second posterior distribution was 0.104 cm3/ cm3, while the mean value

of measure SDUL was 0.110 cm3/ cm3. The error was only about 0.006 cm3/ cm3. Similar results

were observed in SLLL and SSAT.

In general, the uncertainties of the selected 9 input parameters were dramatically decreased

after two rounds of GLUE estimations.

3.3.5 PDF Plot of Selected Parameters

The prior distribution, first and second posterior distribution of the selected input

parameters were plotted with their histograms to show the changes of their ranges and

distribution forms (Figure 3-10 to 3-18).

From these figures, the ranges and distributions of the selected input parameters during the

two rounds of GLUE can be estimated. In the first round of GLUE, most of the distributions did

not follow a normal distribution. This was because very few qualified parameter sets were









selected under a very restrictive likelihood function. But in the second round of GLUE, most of

the distributions tended to follow a normal distribution. This result occurred because the ranges

of the new prior distributions were heavily narrowed. More acceptable parameter sets that could

give good outputs were generated and selected finally to smooth the final distributions became

smoother.

3.3.6 Distributions of Outputs

The four figures below (Figures 3-19 to 3-22) show the distributions of the predicted yields

(dry matter, kg ha-1), anthesis dates (days after planting), maturity dates (days after planting), and

accumulative nitrogen leaching (NLCM, kg ha-1), respectively. Red curves in the figures are the

fitted normal distribution curves. These figures show the trend that after two rounds of GLUE

estimation, the uncertainties of predicted outputs, such as yields, anthesis date, maturity date, and

nitrogen leaching, were also reduced. The uncertainties of the input parameters were

significantly reduced since the mean values shifted toward the field measured values (especially

for soil parameters), and standard deviations were noticeably reduced. Consequently, the

uncertainties in outputs also decreased. For example, under the second posterior distribution, the

mean value of the predicted dry matter yield was near 3,000 kg ha-1, which approximated the

measured yield of sweet corn production in North Florida (See Chapter 4 for information on

sweet corn yield). For anthesis dates and maturity dates, most parameter sets gave a prediction of

55 and 80 days after planting, respectively. These values were also equal to observations in the

field experiment of 2006. In general, the output uncertainties were dramatically reduced after

two round of GLUE simulation, which strengthened the confidence in the model behavior and to

use the posterior distributions of the selected input parameters for future research.









3.3.7 Joint Distribution between Yield and Nitrogen Leaching

Yield and nitrogen leaching were two main concerns in this study to develop potential best

management practices (BMPs) (see Chapter 5). Therefore it was necessary to know how these

two correlated to each other. A 3-D plot of the joint distribution between yield and nitrogen

leaching was developed. Figures 3-23, 3-24 and 3-25 show the 3-D plots under prior, first

posterior and second posterior distribution of the selected input parameters, respectively.

In Figure 3-23, the predicted yields spread out in a very wide range with a mean value of

close to 6,000 kg ha-1, which was much greater than the observed values. For nitrogen leaching,

the values mainly focused around 0 to 20 kg N ha-1. They were also far from the estimated values

of nitrogen leaching in field plot experiment in 2006, which were 90 to 170 kg N ha-1 (Table 4-

16 in Chapter 4).

In Figures 3-24 and 3-25, the ranges of yield and nitrogen leaching narrowed, meaning

the uncertainties of these two outputs decreased. Finally, yields were around 3,000 kg ha-1, which

were more close to the measured values in field experiment (Table 4-16 in Chapter 4). And the

values of nitrogen leaching were mainly between 50 to 100 kg ha-l. Though they were still less

than the estimated nitrogen leaching amounts in field plot experiment, they became more close to

the estimated values. See Chapter 4 for more information about sweet corn yield and nitrogen

leaching.

3.3.8 GLUE Verification

Though the GLUE procedure did a good job in soil parameter estimation because the

estimated SLLL, SDUL and SSAT values were close to measured values (Table 3-8), the

accuracy of the estimated genotype parameters may not have been accurate. For example,

genotype parameter P1 still had a coefficient of variation of 0.083. However, there was no

experiment designed in this study to directly measure those genotype parameters. To obtain more









confidence in the GLUE method and more confidence in the likelihood function (L1) and the

method of likelihood value combination (C2), a verification procedure was conducted.

The selected parameter set was shown in Table 3-9. Four replicates for each observation

were generated with a normal distribution. The mean value of the normal distribution was the

model output derived from this selected parameter set. The variance of the normal distribution

was the variance of observation.

Since the leaf nitrogen content, soil nitrate and moisture content were both temporally

and/or spatially variant, it is inconvenient to list them all here. Therefore only the generated

yields, anthesis dates (ADAT), and maturity dates (MDAT), were specified in Table 3-10.

The results of input parameters of the GLUE verification were summarized in Table 3-11.

It can be seen that after two rounds of GLUE, the uncertainties of the selected parameters

continuously decreased. All mean values gradually approached the measured values listed in

Table 3-9, and all standard deviations decreased gradually. The mean values, especially for the

selected parameters in the second-round GLUE, were similar to the generated replicates.

For example, the initial mean value of P1 in the prior distribution was 225.1. Then it

became 140.4 after the first-round GLUE simulation and 97.3 after the second-round GLUE

simulation. The value of 97.3 was close to the initially selected value of P1 (95.1). The value of

absolute relative error was only about 2.3%. The highest value of RAE occurred in soil

parameter SLDR, which was about 8.1%.

In general, the results were acceptable because it was impossible to make the estimated

values of parameters completely converge to the selected ones, since error always existed in the

observations. This confirmed that the GLUE method was efficient in parameter estimation. After









a strict GLUE process, the mean values of the posterior distributions of the sensitive parameters

approached the actual values, if those values can be measured.

Table 3-12 summarizes the single values for the integrated observations in the growth

season, the yield, anthesis date, and maturity date in the GLUE verification. After two rounds of

GLUE, the uncertainties of the model outputs all decreased. For example, the value of standard

deviation of predicted yields decreased from 2173 to 268. The same change can be observed in

predicted anthesis date (ADAT) and maturity date (MDAT). All mean values gradually

approached the measured values (last column of Table 3-12), and all standard deviations

gradually decreased. The values of CV all decreased to a very low level. For example, the value

of CV of predicted yields decreased from 0.316 under the prior distribution to about 0.077 under

the second posterior distribution.

The mean values of the predicted anthesis dates and maturity dates after two rounds of

GLUE simulation were especially close to the means of the observation values. The values of

RAE were almost 0. The results again confirmed the validity of the GLUE method.

3.3.9 Result of Expected Values of Posterior Distribution

After two rounds of GLUE estimations in this study, a second posterior distribution was

available. Then the expectations of the distribution were calculated with Equation (3-29) to act as

nominal parameter set for future study. The results are listed in Table 3-13.

3.4 Conclusions

In this study, the generalized likelihood uncertainty estimation (GLUE) method was used

to estimate the genotype and soil parameters of the CERES-Maize model of DSSAT. Two years

of field experiment data (2005 and 2006) in Block 1 were used in a two-round GLUE process of

parameter estimation. In the first round of GLUE, the prior distribution was obtained from the

database of DSSAT. The model was run with data of 2005 and the prior distribution. The first









posterior distribution was derived. Then in the second round of GLUE, the first posterior

distribution was used as the new prior distribution. The model was run with data of 2006 and the

new prior distribution. Then the second posterior distribution was obtained.

It was found that all of the selected input parameters (P1, P5, PHINT, SLDR, SLRO,

SDUL, SLLL, and SSAT) can approximate to or follow a normal distribution, except for SLPF,

which was because SLPF was set as 1 when the actual value of it was unknown.

It was necessary to know the number of model runs that must be conducted so as to

guarantee the reliability of model inputs and outputs. Finally it was found that at least 3,000

random parameter sets should be generated and 3,000 model runs should be conducted. It should

be noticed that 3,000 is not the number of model runs to get enough behavioral parameter sets to

construct the posterior distribution.

Though many likelihood functions and methods of likelihood value combination had been

suggested in the literatures, it was found that the likelihood functions and methods of likelihood

value combination could have a very strong influence on the posterior distributions. The

likelihood function L1 (Equation LI), which is based on the probability density function of

normal distribution, and method of combination C2 (Equation C2), which is based on

multiplication, was the best choice for this study.

After two rounds of GLUE simulations, the uncertainty in input parameters and model

outputs were substantially reduced. For example, the standard deviation of P1 for the prior, first

posterior and second posterior distributions were 67.83, 23.39, and 8.22, respectively. Similar

trends occurred for other input parameters. In comparison, the mean values of estimated and

measured soil parameters were very close to each other. The mean value of calibrated SDUL in

the second posterior distribution was 0.104 cm3/ cm3, while the mean value of measure SDUL









was 0.110 cm3/ cm3. The error was only about 0.006 cm3/ cm3. Similar results were observed in

SLLL and SSAT with an error of -0.009 and 0.014 cm3/ cm3, respectively.

To guarantee the reliability of the GLUE method, a process of GLUE verification was

conducted. The verification involved selecting a parameter set, running model with the parameter

set, generating new replicates for the corresponding outputs, conducting two rounds of GLUE

simulation, and comparing the selected parameter set with the second posterior distribution

According to the results, it can be seen that after two rounds of GLUE, the uncertainties of the

model outputs all decreased, and all mean values gradually approached the selected true values.

For example, the value of initially selected P1 was 95.1, while the mean value of the second

posterior distribution of P1 was 97.3, with an error only of 2.2. Similar trends occurred for other

input parameters. The expectations of the posterior distributions should be used as the nominal

values to continue future research in the development of best management practices.

In general the results of this study confirmed that the GLUE method was a powerful tool to

estimate the model input parameters, and strengthened the model users' confidence in their

research.



































I 622.4'"
Figure 3-1. Diagram of Block 1 of field experiment. Symbols of El to E4 represent the four soil
and plant sampling sites on the east part, while W1 to W4 represent the west part.
Symbols of Well #lto Well #4 represent the four monitoring wells for groundwater.


240


235


S230
(0
S225
CU
a 220


215


210


0 1000 2000 3000 4000 5000 6000


7000 8000 9000 10000


Number of model runs

Figure 3-2. Influence of number of model runs on mean values of Pl


























50 I I I I
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Number of model runs

Figure 3-3. Influence of number of model runs on standard deviations of P1


0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Number of model runs

Figure 3-4. Influence of number of model runs on mean values of SLRO










20
19
0 18
-J
4 17
0
S16
.o
15
14
S13
0 12
11
10
0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Number of model runs

Figure 3-5. Influence of number of model runs on standard deviations of SLRO


8400

- 8200
-c
) 8000

7800

5 7600

c 7400

7200

7000


0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Number of model runs

Figure 3-6. Influence of number of model runs on mean values of simulated dry yields


~--------












3000

;-
& 2500


2000
-I
0
2 1500
o lOO
0
1000

-a
C 500


0
I I----I---I----I---I---I----I---I-

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Number of model runs


Figure 3-7. Influence of number of model runs on standard deviations of simulated dry yields





50

45

40
,-c
o 35

30

z 25
0 20
C
m 15
10

5

0
0 -------------------------------

0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000

Number of model runs


Figure 3-8. Influence of number of model runs on mean values of simulated nitrogen leaching














S40



30
^ 35

| 30
z
5 25

. 20

S15

S10

S5

0


0 1000 2000 3000 4000 5000 6000 7000 8000 9000 10000
Number of model runs


Figure 3-9. Influence of number of model runs on standard deviations of simulated nitrogen
leaching





(A)
1I


| 0.5-
a-


50 100 150 200 250 300 350 400 450 500
Number of P1
(B)


1

c 0.5
0
.P
a-


50 100 150 200 250 300 350 400 450 500
Number of P1
(C)


1


S0.5
a


50 100 150 200 250 300 350 400 450 500
Number of P1


Figure 3-10. Parametre P1: probability distribution under (A) prior distribution; (B) first
posterior distribution, and (C) second posterior distributions


p p


A~---~--





























00 200 300 400 500 600 700 800 900 1000
Number of P5
(C)


100 200 300 400 500 600 700 800 900 1000
Number of P5


100 200 300 400 500 600 700
Number of P5
(B)


800 900 1000


Figure 3-11. Parametre P5: probability distribution under (A) prior distribution; (B) first
posterior distribution, and (C) second posterior distributions


10 20 30 40 50 60
Number of PHIN
(B)


10 20 30 40 50 60
Number of PHIN
(C)


1


0.5
aC


10 20 30 40 50 60
Number of PHIN


Figure 3-12. Parametre PHINT: probability distribution under (A) prior distribution; (B) first
posterior distribution, and (C) second posterior distributions


0 1


1

0.5

0
0



1


0.5
0
0
0(


0


1


S0.5


0
(


1

0.5
a


I 0.5-
.Q-

0-
0















I 0.5-
01

0
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Number of SLDR
(B)


0.5-


0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Number of SLDR
(C)





0 J I
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1
Number of SLDR


Figure 3-13. Parametre SLDR: probability distribution under (A) prior distribution; (B) first
posterior distribution, and (C) second posterior distributions


0.5
a-


-I



,...m m B m-----_
3 10 20 30 40 50 60 70 80 90 100
Number of SLRO


0.5 -


0 10 20 30 40 50 60 70 80 90 100
Number of SLRO
(C)


0.5-


0 10 20 30 40 50 60 70 80 90 100
Number of SLRO

Figure 3-14. Parametre SLRO: probability distribution under (A) prior distribution; (B) first
posterior distribution, and (C) second posterior distributions















0.5-

0 0.05 0.1
0 0.05 0.1


0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Number of SLLL
(B)


0.5


0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Number of SLLL
(c)


0.5 -


0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Number of SLLL



posterior distribution, and (C) second posterior distributions


(A)
1

0.5-


0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Number of SDUL
(B)
1

0.5-


0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Number of SDUL

1 I(C)


0.5-


0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5
Number of SDUL

Figure 3-16. Parametre SDUL: probability distribution under (A) prior distribution; (B) first
posterior distribution, and (C) second posterior distributions



















0.25 0.3 0.35 0.4 0.45 0.5
Number of SSAT
(B)


II


0.55 0.6 0.65 0.7


I I I I


I I I I I I
.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.
Number of SSAT
(C)
I I I I I I I I


7


0.5


0 -
0.2 0.25 0.3 0.35


0.4 0.45 0.5
Number of SSAT


0.55 0.6 0.65 0.7


Figure 3-17. Parametre SSAT: probability distribution under (A) prior distribution; (B) first
posterior distribution, and (C) second posterior distributions


1


1 0.5
Q"_


0.65 0.7 0.75 0.8
Number of SLPF
(B)


0.65 0.7 0.75 0.8
Number of SLPF
(C)


0.85 0.9 0.95


0.85 0.9 0.95


0.65 0.7 0.75 0.8 0.85 0.9 0.95
Number of SLPF


Figure 3-18. Parametre SLPF: probability distribution under (A) prior distribution; (B) first
posterior distribution, and (C) second posterior distributions


S0.5-


0-
0.2


1

0.5

0
0


1


0.5
a


1


0.5
.o
0
CL


-

-




















5000 10000


10000


5000 10000
Predicted Yield (kg/ha)


Figure 3-19. Histogram of predicted dry matter yields under (A) prior distribution; (B) first
posterior distribution, and (C) second posterior distributions


3000

E 2000
z
1 1000
0
0~


20 40 60 80 100 120 140 160 180 200
(B)


20 40 60 80


120 140 160 180 200


20 40 60 80 100 120 140
Predicted Anthesis Date (days after planting)


160 180 200


Figure 3-20. Histogram of predicted anthesis dates under (A) prior distribution; (B) first
posterior distribution, and (C) second posterior distributions


n 400L
E
z
200
0

0


n 400
E
z
S200
0
0
0



0 400
E
:3
z
S200

S0
0


15000


15000


15000


2000-

1000-

0-
0


E 2000-
z
| 1000-
0
0-
0


Ad













3000

E 2000
Z
z
S1000
0

0 20 40 60 80 100 120 140 160 180 200
(B)
3000

S2000

S1000
:3
0

0 20 40 60 80 100 120 140 160 180 200
(C)
..3000 ,

I 2000
z
0 1000
S3I

0 20 40 60 80 100 120 140 160 180 200
Predicted Maturity Date (days after planting)

Figure 3-21. Histogram of predicted maturity dates under (A) prior distribution; (B) first
posterior distribution, and (C) second posterior distributions



3000

E 2000

1000


0 20 40 60 80 100 120 140 160 180 200
(B)


, 300(

E 200(
z
S100
0



, 300(

E 200(
z
| 100(
3
0
0


SI I I-

0 -

0 20 40 60 80 100 120 140 160 180 200
(C)


0-

0 -

0 -

0 -


.EnlnlinEnnn.-.


0 20 40 60 80 100 120 140 160 180 200
Predicted Accumulative Nitrogen Leaching (kg/ha)

Figure 3-22. Histogram of predicted cumulative nitrogen leaching under (A) prior distribution;
(B) first posterior distribution, and (C) second posterior distributions
















0.1

0.08 -

S0.06

S0.04 -

0.02


300
250
15000
150 10000
100 --. -
0 5000
50
0 0
Accumulative Nitrogen leaching (kg/ha) Dry Matter Yield (kg/ha)


Figure 3-23. Joint distribution between yield and nitrogen leaching under prior distribution of
input parameters


0.1 -

0.08

S0.06

2 0.04

0.02-

00
300


15000


100 W00

Accumulative Nitrogen leaching (kg/ha) 0 Dry Matter Yield (kg/ha)

Figure 3-24. Joint distribution between yield and nitrogen leaching under the first posterior
distribution of input parameters




120


L
















0.1

0.08

0.06

2 0.04

0.02

300

20( 15000
10000
100 5000

Accumulative Nitrogen leaching (kg/ha) Dry Matter Yield (kg/ha)

Figure 3-25. Joint distribution between yield and nitrogen leaching under the second posterior
distribution of input parameters










Table 3-1. Average soil physical properties of the experiment site (from 24 sampling locations)
Bulk
Depth e Clay Silt Sand BulkPWP FC Saturation
Texture Density
(cm) () (%) (%) (g/cm3) (cm3/cm3) (cm3/cm3) (cm3/cm3)
15 Sandy soil 2.75 1.92 95.33 1.67 0.051 0.110 0.313
30 Sandy soil 2.56 2.35 95.08 1.69 0.061 0.117 0.317
60 Sandy soil 2.36 1.76 95.88 1.67 0.077 0.118 0.357


Table 3-2. Selected parameters for GLUE method due to sensitivity analysis of predicted dry
matter yield and accumulative nitrogen leaching (See Chapter 2 for details)a
P1 P5 PHINT SLPF SLDR SLRO SDUL SLLL SSAT
Parameter
CPd Cd Cd cm3/cm3 cm3/cm3 cm3/cm3
Value 225.10 763.60 41.20 0.96 0.46 73.00 0.25 0.13 0.38
a oCd means degree day.


Table 3-3. Covariance matrix of the prior distribution
P1 P5 PHINT SLDR SLRO SDUL SLLL SSAT
P1 4561.712 2373.905 61.862 0.00 0.000 0.000 0.000 0.000
P5 2373.905 9679.386 55.854 0.000 0.000 0.000 0.000 0.000
PHINT 61.862 55.854 15.975 0.000 0.000 0.000 0.000 0.000
SLDR 0.000 0.000 0.000 0.036 -0.339 -0.003 -0.003 -0.005
SLRO 0.000 0.000 0.000 -0.339 132.383 0.314 0.236 0.259
SDUL 0.000 0.000 0.000 -0.003 0.314 0.010 0.008 0.006
SDLL 0.000 0.000 0.000 -0.003 0.236 0.008 0.007 0.005
SSAT 0.000 0.000 0.000 -0.005 0.259 0.006 0.005 0.009



Table 3-4. Results of Jarque-Bera test of the input parameters ab
Parameter H p-value JBSTAT CV
P1 1 0.012 8.862 5.992
P5 1 0.014 8.566 5.992
PHINT 1 1.398e-004 17.751 5.992
SLDR 0 0.079 5.083 5.992
SLRO 1 0.001 13.153 5.992
SLPF 1 0.000 9.733e+003 5.992
SLLL 0 0.080 5.063 5.992
SDUL 0 0.248 2.786 5.992
SSAT 0 0.083 5.576 5.992
a
SH=: reject the hypothesis that x has a normal distribution; while H=0: accept the hypothesis.
bCV is the critical value, and JBSTAT is the constructed statistic in the Jarque-Bera test. If JBSTAT>CV, then
H=I, while JBSTAT









Table 3-5. Mean values and standard deviations (STDEV) of first-round posterior distributions
a
derived from different likelihood functions and likelihood combinations
Under L1
Prior Distribution Under C1 Under C2 Under C3
Mean STDEV Mean STDEV Mean STDEV Mean STDEV


225.10
763.60
41.17
0.46
73.00
0.26
0.14
0.39
0.96


67.83
98.80
4.01
0.19
11.56
0.10
0.08
0.09
0.11


176.90
698.38
41.38
0.52
72.33
0.19
0.11
0.33
0.83


112.31
120.07
4.78
0.22
11.00
0.10
0.07
0.11
0.09


144.49
630.78
40.77
0.73
77.40
0.10
0.06
0.30
0.92


23.39
27.68
0.57
0.01
10.63
0.00
0.00
0.02
0.04


Under L2


Prior Distribution


Mean
225.10
763.60
41.17
0.46
73.00
0.26
0.14
0.39
0.96


STDEV
67.83
98.80
4.01
0.19
11.56
0.10
0.08
0.09
0.11


Prior Distribution


Mean
225.10
763.60
41.17
0.46
73.00
0.26
0.14
0.39
0.96


STDEV
67.83
98.80
4.01
0.19
11.56
0.10
0.08
0.09
0.11


Under C1


Mean
178.45
700.20
40.75
0.52
71.27
0.20
0.11
0.32
0.83


Under C1


Mean
176.72
698.61
41.23
0.52
72.23
0.19
0.11
0.33
0.83


STDEV Mean
105.42 142.12
124.40 611.96
4.47 40.06
0.23 0.69
11.56 78.24
0.11 0.10
0.08 0.06
0.11 0.28
0.09 0.91
Under L3


STDEV
110.12
121.44
4.78
0.22
11.09
0.11
0.08
0.11
0.09


Under C2


Under C2


Mean
166.20
653.98
41.20
0.73
67.50
0.11
0.06
0.28
0.88


STDEV
12.98
24.67
1.32
0.10
6.48
0.00
0.01
0.06
0.06


STDEV
38.04
47.55
1.13
0.04
17.29
0.00
0.01
0.04
0.07


Under L4


Prior Distribution


Mean
225.10
763.60
41.17
0.46
73.00
0.26
0.14
0.39
0.96


STDEV
67.83
98.80
4.01
0.19
11.56
0.10
0.08
0.09
0.11


Under C1
Mean STDEV
75.31 40.52
663.15 98.37
37.71 4.22
0.45 0.19
72.77 10.80
0.25 0.10
0.13 0.08
0.37 0.10
0.85 0.08


Under C2
Mean STDEV
166.45 37.49
650.43 50.16
41.06 1.36
0.72 0.07
67.39 17.01


0.11
0.06
0.27
0.87


0.00
0.01
0.05
0.07


P1
P5
PHINT
SLDR
SLRO
SDUL
SLLL
SSAT
SLPF


0.25
0.13
0.37
0.85


0.10
0.08
0.10
0.08


178.34 112.01
700.44 119.00
41.46 4.75
0.52 0.22
72.31 11.02
0.19 0.10
0.11 0.07
0.33 0.11
0.83 0.09

Under C3
Mean STDEV
181.11 104.62
704.36 122.21
40.88 4.44
0.53 0.23
71.21 11.60
0.20 0.11
0.11 0.08
0.32 0.11
0.83 0.09

Under C3
Mean STDEV
178.95 109.72
701.90 119.85
41.34 4.75
0.52 0.22
72.21 11.12
0.19 0.11
0.11 0.07
0.33 0.11
0.83 0.09

Under C3
Mean STDEV
75.94 41.25
669.97 97.07
37.80 4.23
0.45 0.19
72.69 10.84


P1
P5
PHINT
SLDR
SLRO
SDUL
SLLL
SSAT
SLPF


P1
P5
PHINT
SLDR
SLRO
SDUL
SLLL
SSAT
SLPF


P1
P5
PHINT
SLDR
SLRO
SDUL
SLLL
SSAT
SLPF


"Under L means deriving posterior distribution with the likelihood function L1, "Under Cl "means
deriving posterior distribution under the method of likelihood value combination C1, and so on and so forth.










Table 3-6. Mean values and standard deviations (STDEV)
round posterior distributionsab


of model outputs derived from first-


Outputs Yield Anthesis Date Maturity Date Nitrogen Leaching
Unit kg ha Days Days kg ha'-
Mean 3217.20 55.35 88.09 110.88
L1C2 ARE 0.01 0.09 0.11
STDEV 582.94 3.26 4.39 42.07
Mean 3724.53 57.94 93.52 73.32
L2C2 ARE 0.15 0.14 0.18
STDEV 1473.95 1.33 0.83 56.09
Mean 3148.53 60.92 98.59 138.14
L3C2 ARE 0.03 0.20 0.24
STDEV 1162.27 5.28 6.98 63.32
Mean 3080.4 61.05 98.4 137.67
L4C2
ARE 0.05 0.20 0.24
STDEV 1386.31 5.18 6.93 73.14
Mean 3234.79 50.75 79.5
Measured
Measured STDEV 120.9 2.22 3.7
a ARE is the absolute relative error, which is defined as ARE=IY-Y'I/Y, where Y is the measured value and Y'
is the predicted value of model output.
a Signal '-' means no absolute relative error was available since there was no direct measurement of nitrogen
leaching in this study.










Table 3-7. Fundamental statistical properties of prior, first posterior and second posterior
distributions derived from L1C2
Prior


P1
P5
PHINT
SLDR
SLRO
SDUL
SLLL
SSAT
SLPF



P1
P5
PHINT
SLDR
SLRO
SDUL
SLLL
SSAT
SLPF



P1
P5
PHINT
SLDR
SLRO
SDUL
SLLL
SSAT
SLPF


Min
110.000
580.000
30.000
0.000
30.000
0.086
0.020
0.230
0.700

Min
136.926
550.355
36.777
0.443
44.543
0.102
0.023
0.141
0.759

Min
77.676
553.141
39.162
0.708
41.492
0.097
0.053
0.235
0.760


Table 3-8. Measured and estimated mean values of soil properties of the field experiment site
SLLL (cm3/ cm3) SDUL (cm3/ cm3) SSAT (cm3/ cm')
Measured Estimated Measured Estimated Measured Estimated
Mean 0.051 0.060 0.110 0.104 0.314 0.300
STDEV 0.031 0.002 0.044 0.002 0.070 0.021
CV 60.8% 3.3% 40.0% 1.9% 22.3% 7.0%


Max Mean
450.000 225.096
1000.000 763.595
50.000 41.175
1.000 0.463
95.000 72.995
0.470 0.263
0.351 0.138
0.700 0.388
1.000 0.962
First Posterior
Max Mean
216.797 144.492
716.360 630.781
42.524 40.772
0.750 0.731
80.836 77.398
0.109 0.104
0.070 0.062
0.305 0.298
0.933 0.919
Second posterior
Max Mean
182.175 99.169
676.212 577.201
41.712 39.676
0.752 0.732
99.850 78.143
0.109 0.104
0.068 0.060
0.362 0.300
0.932 0.872


Standard Deviation
67.826
98.800
4.014
0.192
11.561
0.100
0.084
0.094
0.114

Standard Deviation
23.387
27.684
0.567
0.006
10.627
0.002
0.003
0.021
0.042


Standard Deviation
8.217
9.746
0.202
0.006
9.660
0.002
0.002
0.021
0.041


CV
30.1%
12.9%
9.8%
41.4%
15.8%
37.9%
61.1%
24.2%
11.9%

CV
16.2%
4.4%
1.4%
0.9%
13.7%
1.7%
4.4%
7.2%
4.6%


CV
8.3%
1.7%
0.5%
0.9%
12.4%
1.6%
4.0%
7.0%
4.7%










Table 3-9. Selected parameter set for GLUE verification
P1 P5 PHINT SLDR SLRO SDUL SLLL SSAT SLPF
Parameter
CPd Cd Cd cm'/cm3 cm3/cm3 cm3/cm3 -
Value 95.1191 572.0396 39.5679 0.7392 89.4470 0.1037 0.0604 0.3190 0.9312
a Cd means degree day.


Table 3-10. Generated duplicates of observations for


GLUE verification


2005 Measured Generated Replication
Observation Mean STDEV 1 2 3 4
Yield 3451 59 3046 3100 3385 3686
ADAT 55 3 60 50 56 50
MDAT 85 4 87 82 84 84
2006 Measured Generated Replication
Observation Mean STDEV 1 2 3 4
Yield 3206 121 3170 3052 3236 3360
ADAT 50 2 49 51 50 51
MDAT 85 4 87 80 88 82










Table 3-11. Means and standard deviations of the selected parameters in GLUE verification
Prior Distribution First-round GLUE Second-round GLUE Se d
Selected ARE
Mean STDEV CV Mean STDEV CV Mean STDEV CV
P1 225.096 67.826 0.301 140.43 10.689 0.076 97.334 1.929 0.02 95.119 2.3%
P5 763.595 98.8 0.129 617.347 24.137 0.039 566.077 2.279 0.004 572.04 1.0%
PHINT 41.175 4.014 0.097 40.375 1.416 0.035 38.723 0.498 0.013 39.568 2.1%
SLDR 0.463 0.192 0.414 0.707 0.082 0.116 0.799 0.033 0.041 0.739 8.1%
SLRO 72.995 11.561 0.158 78.879 5.905 0.075 85.258 1.061 0.012 89.447 4.7%
SDUL 0.263 0.1 0.379 0.103 0.001 0.006 0.104 0.001 0.005 0.104 0.0%
SLLL 0.138 0.084 0.611 0.057 0.011 0.188 0.057 0.002 0.03 0.06 5.0%
SSAT 0.388 0.094 0.242 0.292 0.047 0.16 0.322 0.014 0.042 0.319 0.9%
SLPF 0.962 0.114 0.119 0.915 0.052 0.057 0.91 0.004 0.004 0.931 2.3%
aARE was the absolute relative error between the selected values and the estimated values of the parameters after two rounds of GLUE process;
"Selected" means the parameter set was selected from the behavioral parameter sets in the second round of GLUE process; the selected parameter set was
used as the "true" value in GLUE verification.










Table 3-12. Means and standard deviations of model outputs in GLUE verification
Prior First Posterior Second Posterior Measured in 2006
Yield MEAN 6867 3442 3471 3206
STDEV 2173 1396 268 121
CV 0.316 0.406 0.077 0.038
RAE 114.2% 7.4% 8.3%
ADAT MEAN 76 58 50 50


STDEV
CV
RAE
MDAT MEAN
STDEV
CV
RAE


9
0.118
52.0%
117
10
0.084
37.6%


0.059
16.0%
94
4
0.048
10.6%


0.009
0.0%
85
0
0.005
0.0%


2
0.044

85
4
0.044


Table 3-13. Expectation values of second posterior distribution of selected parametersa
P1 P5 PHINT SLDR SLRO SDUL SLLL SSAT SLPF
Parameter
Paramet Cd d Cd d Cd cm3/cm3 cm3/cm3 cm3/cm3 -
Expectation 99.17 577.20 39.68 0.73 78.14 0.10 0.06 0.30 0.87
a Cd means degree day.









CHAPTER 4
FIELD PLOT EXPERIMENT OF SWEET CORN AND SIMULATION WITH CALIBRATED
CERES-MAIZE MODEL

4.1 Introduction

Leaching of nitrate nitrogen is economically and environmentally undesirable (Asadi, et al.,

2002). Due to the large acreage of sweet corn planted and the relative large amount ofN

fertilizer application to this crop, something must be done to control this situation. A proactive,

incentive-based program of developing crop specific Best management practices (BMPs) in

Florida began in 1994 as a result of an amendment to the Florida Fertilizer Law approved by the

state legislature (Alva et al., 2005). This amendment authorized the Florida Department of

Agriculture and Consumer Services (DACS) to develop research based crop specific N-BMPs.

The U.S. Environmental Protection Agency (EPA) defines a BMP as "methods, measures

or practices selected by an agency to meet its non-point source control needs" (Code of Federal

Regulations, 1994). BMPs generally refer to practices determined to be the most effective

practical means for preventing or reducing the amount of pollution generated by non-point

sources to a level compatible with quality goals (Center et al., 1996).

Bottcher et al. (1995) defined the specific BMP for Florida, where the environmental

impact and economics of the farming operations were maintained as fairly important. In their

definition, the BMPs were those on-farm activities designed to reduce nutrient losses in drainage

waters to an environmentally acceptable level, while simultaneously maintaining an

economically viable farming operation for the grower. Practices that have a high potential for

negatively impacting the financial profitability of a farm should not, therefore, be considered

BMPs. In the case where the economic cost of implementing certain BMPs puts an excessive

financial burden on the farmer, such practices should be considered as BMPs only if external

funds are available to return an acceptable level of profitability to the farm (Bottcher et al., 1995).









Research, concentrating on finding the optimum N rate and N placement method for sweet

corn production, has been conducted in several experiments in Florida, some of those were in

Gainesville by Rudert and Locascio in 1976 and 1977 (Rudert and Locascio, 1979), in Quincy,

North Florida Research and Education Center, in 1990 (Rhoads, 1990), and at the Suwannee

Valley Agricultural Research and Education Center near Live Oak (Hochmuth and Donley, 1992;

Kidder et al., 1989). All of the experiments provided information about how much N fertilizer

should be used to optimize sweet corn yield. But there were some limitations to these

experiments. First, some of the experiments only concentrated on N fertilizer itself and excluded

irrigation, which can simultaneously influence nitrogen leaching and corn yield. Second, most of

the experiments failed to explain the effects of N fertilizer and irrigation on ear quality

characteristics, such as ear grade according to their lengths and diameters.

The fate and budget of nitrogen in the agricultural systems of sweet corn in Florida is also

an important issue and should receive much attention for both the agricultural and the

environmental aspect if research based BMPs are to be developed. The N budget or balance is

often evaluated by comparing various N inputs and outputs in soil-crop systems by considering

changes of soil mineral N (Sogbedji et al., 2000). Research on the N balance that takes into

account mineralization and inorganic N in soils can provide more detailed information on the N

cycles and losses by integrating soil N process into the total N budgets (Liu et al., 2003). There

are limitations to the calculation of the N balance, however, because it is difficult to measure

each component of the N budget accurately in relation to soil processes (Jarvis, 1996; Sogbedji et

al., 2000).

With the development of computer technology, crop models have become a strong tool for

exploration of possible management strategies in crop production. A crop model has been









described as a "quantitative scheme for predicting the growth, development and yield of a crop,

given a set of genetic coefficients and relevant environmental variables" (Monteith, 1996).

Models are not perfect, and can at best only represent a current understanding of biological

systems; yet they do highlight the areas where information and understanding are lacking (Boote

et al., 1996). With these caveats, crop models can be used to predict crop growth, development

and yield as a function of soil, climate, weather, and crop management conditions (Ghaffari et al.,

2001). The CERES-Maize corn growth and yield model (Jones and Kiniry, 1986; Tsuji et al.,

1994) in the Decision Support System for Agrotechnology Transfer (DSSAT) model, V4.0, is a

popular crop model. This crop model can be used to simulate field experiments. Then the

reliability of the model can be evaluated according to the results of comparison between the

simulated results and observed results.

The objectives of this research were the following: (1) explore the response of yield

quantity and quality of sweet corn to different irrigation and fertilization levels; (2) study the fate

of nitrogen fertilizer in sweet corn production; and (3) simulate the field plot treatments with

CERES-Maize mode of DSSAT model and compare the outputs with observations so as to

evaluate the model.

4.2 Material and Methods

4.2.1 Experiment Site and Design

A field plot experiment was conducted in the spring of 2006 at the Plant Science Research

and Education Unit, the University of Florida. The unit is located in Pine Acres (29.4094N,

82.1777W, 20.746 meters above sea level), Marion County, Florida, U.S. (Judge et al., 2005).

The soil of the experiment field is very sandy. It consists of Lake Sand, Candler Variant,

Tavares Variant, and Millhopper Variant 1 etc, which mainly belong to Quartzipsamments

(Entisol). Soil samples were collected at 24 sites at 3 depths of 0-15 cm, 15-30 cm, and 30-60 cm.









The samples were sent to the lab of the Soil and Water Science Department of the University of

Florida and analyzed for physical properties. The permanent wilting point (PWP) was measured

as the soil moisture at a soil pressure of 15.3 bar, field capacity (FC) as the soil moisture at 0.1

bar, and soil saturation as the soil moisture at 0 bar. The main measured properties of the soil at

the experiment site are summarized in Table 4-1.

The plot experiment was designed as a two-factor split-plot experiment, since fertilizer and

irrigation rates, which simultaneously affect nitrogen leaching and corn yield, were tested. This

experiment consisted of two irrigation levels and three rates of N fertilizer application. The two

irrigation levels were I0 and 1.5 x I0, where I0 is the irrigation schedule based on daily water

balance in soil profile. They were identified as II and 12 respectively. Three fertilizer application

levels were 185, 247 and 309 kg N hal-, which were identified as Fl, F2 and F3. Subsequently,

there were six combination treatments as F1i1, F2I1, F3I1, F1I2, F2I2, and F3I2 (Figure 4-1).

As shown in Figure 4-1, there were 4 blocks. In each block, a single replicate of a complete

factorial experiment by irrigation and nitrogen levels was included. However, the treatment

combinations within a block were not completely randomized. Each block in the design was

divided into two whole plots (I1 and 12). Then each whole plot was divided into three subplots or

split-plots (F F2 and F3). So irrigation levels were considered as main treatments, with

fertilizer levels as subplot treatment. Four blocks resulted in four replications for each of the six

treatment combinations.

To estimate the growth and yield of sweet corn under extreme conditions, some extra

treatments were also arranged as controls in this study. A non-irrigated, IO, and zero nitrogen

fertilizer level, FO, were added to the experimental design. The additional treatments were

derived using FO and IO along with other nitrogen and irrigation levels. These treatments were









F110, F210, F3IO, and FOI1. Treatment FOI1 had 3 replicates, which were arranged in a single

column. However the combinations, F1I0, F210, and F3IO, were not replicated (Figure 4-1).

Since these treatments were not randomly arranged in the field or did not have enough replicates

to meet the statistical requirement of a successful field experiment, only the results from the first

eight columns (Figure 4-1, from left to right) of field were used for statistical analysis, while the

results from the rest two columns were only treated as a reference. Other aspects of fertility (such

as the application of phosphorus (P), potassium (K), and micronutrient) and management (such

as planting, harvesting, and pest control) were the same across all the treatments.

Each plot consisted of eight rows of sweet corn (Zea mays L., Saturn SH2) corn with a row

spacing of 76 cm. The plot length was 15.2 m. The planting date was March 14, 2006, or Julian

day 73. The corn was planted at a depth of 3.8 cm with a planting population density of 59,000

plants ha1 (24,000 plan ac-1).

Weather data was extremely important for production management, especially for

irrigation scheduling. The values of daily reference evapotranspiration (ETo) and precipitation

were used to schedule the timing and depth of irrigation events. Daily weather data, including

daily ETo, rainfall, minimum temperature, and maximum temperature for Pine Acres was

directly obtained from the weather database of the Florida Automated Weather Network (FAWN)

at the Citra site where Pine Acres is located. The methodology of reference evapotranspiration

adopted by FAWN was described in Section 4.2.3.

4.2.2 Nitrogen Fertilizer Application

Growers typically apply fertilizer through the sprinkler irrigation system, such as the center

pivot system or linear move irrigation system. However, in this experiment the experiment

design made it difficult to apply fertilizer via the linear move irrigation system. First, the length

of a span of the system was more than 40 m, while the width of each experiment plot was only









15.2 m. One span could almost cover the three plots in each column (Figure 4-1). However,

different rates of nitrogen fertilizer were required for the three plots as the experiment design,

while one span could only apply one rate. Second, the linear system was cumbersome to control,

making accurate application of fertilizer difficult on small plots. Thus, the nitrogen fertilizer was

applied into the field through a drip tape system instead in this research to simulate the sprinkler.

Nitrogen fertilizer was also applied as solution with high application uniformity, which could

guarantee the experiment results would not be impacted.

As shown in Figure 4-1, there were twenty seven sub-main lines conveying fertilizer

solution and water to each fertigated plot. These twenty seven sub-main lines were connected to

three main lines. When applying fertigation, nitrogen fertilizer solution was injected with an

injection pump into the drip tape system through the injection hole. The injection pump used in

this experiment was an Easy-Load II MASTERFLEX peristaltic pump (Cole-Parmer Instrument

Company, Vernon Hills, Illinois).

The nitrogen fertilizer used in the experiment was a composite of several nitrogen

compounds. The total nitrogen mass was about 32% of the solution, including 7.9% nitrate

nitrogen, 7.9% ammoniacal nitrogen, and 16.2% urea nitrogen. The concentration of total

fertilizer solution was 1.29 kg L-1, while the concentration of nitrogen in this solution was 0.41

kg N L1.

When arranging the drip tapes between each row, fertigation uniformity was considered.

The uniformity mainly depends on the number of drip tapes between rows. A model simulation

was conducted with the HYDRUS-2D computer program to decide how many drip tapes were

needed to obtain adequate uniformity. The dimension of simulation profile was 76x50 cm, where

76 cm is the row width and 50 cm is the soil profile depth. Each emitter of a drip tape was









represented as a point with a constant pressure of 0 bar and placed at the top of the profile. Initial

soil pressure was set as 0.2 bar. The bottom of the profile was set as free drainage. Figure 4-2

shows the simulated soil moisture in the soil profile after 30 minutes with one, two, three, and

four drip tapes, respectively between rows. The different blue colors from light to dark represent

different soil moisture from higher to lower.

In this research, the low quarter distribution uniformity (DUlq) value as defined in Equation

(4-1), were used to quantify the uniformity between the various numbers of drip tapes. The

calculated DUlq values for four different numbers of drip tapes at three depths of Dl (10cm), D2

(20cm), and D3 (30cm) at a time of 30 minutes are listed in Table 4-2.

DU Average Minimum 25%
DU = x 100% (4-1)
S Average Total

where Average _Minimum 25% is the average of lower 25% of soil water contents,

and Average Total is the average of total soil water contents. The values of soil water contents

at different locations were provided as outputs of HYDRUS-2D.

The uniformities at the depth of D3 were all 1.00, which means the water applied did not

reach this layer, yet. The water moisture at deep layers was assumed homogeneous. For layer Dl

and D2, it can be found that with four drip tapes, highest fertigation uniformity could be obtain

ed. These four drip tapes were evenly fixed in each of the row interval.

The final arrangement of drip tapes in each row interval was shown as Figure 4-3. Four

drip tapes were evenly fixed in the interval to guarantee the uniformity of fertigation. For

example, as shown in Table 4-1, the DUlq at Dl (10cm) at 30 minutes was only 0.42 with only

two drip tapes, while it increased to 0.97 with four drip tapes. There were seven inter-row areas

in each plot. However, only the central five row intervals were arranged with drip tapes since the









outer two rows were border rows. See Appendix D for the photos of fertigation system, corn

growth, and sampling in field experiment.

The final N fertilizer application schedules in 2006 are given in Table 4-3. The middle

fertilizer level (F2) was about 10% higher than the nitrogen fertilizer level recommended by

Institute of Food and Agricultural Sciences (IFAS), which is about 224 kg N ha-1 (Hochmuth,

2000). Fl was 75% of F2, while F3 was 125% of F2. The first N application was carried out

during planting, while all other applications were applied in weekly applications beginning three

weeks after planting with an injection pump.

4.2.3 Irrigation Scheduling

In this experiment, the irrigation schedule was prepared with the water balance for the soil

profile. This schedule was defined as the standard irrigation for this study. The water content in

the effective root zone was estimated by using following dynamic water balance equation:

WCt = WCt1 + IRR + RAIN ETc (DP + RO) (4-2)
where
WCt= Soil water content today, mm
WCt-i= Soil water content yesterday, mm
IRR= Irrigation depth since yesterday, mm
RAIN= Rain since yesterday, mm
ETc = Crop ET, mm
DP= Deep percolation, mm
RO=Runoff, mm

Supposing there is no water wasted in each irrigation event, i.e. there is no DP or RO, and

IRR and RAIN are all known. Hence, the irrigation level, WCt WCt_, is a function of ETc:

ETc =ETO xKc xKs (4-3)

The reference crop ET or reference ET, denoted as ETo, is the evapotranspiration from the

reference surface. The reference surface is a hypothetical grass reference crop with an assumed









crop height of 0.12 m, a fixed surface resistance of 70 s m-1 and an albedo of 0.23 (Allen et al.,

1998).

Equation 4-3 adjusts ETo by the crop coefficient (Kc) and the stress coefficient (Ks). In

practice, the Ks was set as 1 for this project due to the well-watered nature of the crop. The daily

ETo and rainfall values were obtained from Florida Automated Weather Network (FAWN).

In the FAWN system, the daily ETo values were calculate using the IFAS Penman method

(Jones et al., 1984). The working Penman equation is give by following equation:


S(l- a)R, -oT '(0.56- 0.08e 1.42 0.42
A+y R_
ET =

+ [0.263x(0.5+0.0062u)x(ea -ed)] (4-4)
A+y
where
ET =daily potential evapotranspiration, mm day-'
R =total incoming solar radiation, cal cm-2 day1
Rs =total daily cloudless sky radiation, cal cm-2 day1
T =average air temperature in K
e =vapor pressure of air= (emax + emn )/2, mb
emax =maximum vapor pressure of air during a day, mb
e. = minimum vapor pressure of air during a day, mb
ed =vapor pressure at dew point temperature (Td ), mb
u2 =wind speed at a height of 2m, km day-1
A =slope of saturated vapor pressure curve of air, mb C-1
S=psychrometric constant, 0.66 mb C-1
A =latent heat of vaporization of water-(59.59-0.055Tavg), cal cm-2 mm'
Lvg (Tmax + Tmi n)/2, OC
Tma =maximum daily temperature, C
Tn =minimum daily temperature, C

The crop coefficient of sweet corn depends on the growth stages. The following crop

coefficients (Table 4-4) recommended by Bauder and Waskom (2003), were used to determined









ETc used by sweet corn for various stages of development. Similar information can also be

found in the "Vegetable Production Guide for Florida" (Olson and Simonne, 2005).

4.2.4 Soil, Biomass, and Yield Sampling

In the field experiments soil and biomass samplings were done to evaluate the nitrogen

status in soil profile and corn tissue. Finally yield sampling was conducted to evaluate the yield.

The sampling position in each plot was shown in Figure 4-4.

Soil Sampling was conducted approximately biweekly during the growth season. Soil

samples were collected in each of the plots at 4 depths of 0-15 cm, 15-30 cm, 30-60 cm, and 60-

90 cm. The samples were analyzed at the Department of Soil and Water Science University of

Florida for KCL extractable nitrate, ammonium concentrations, and moisture content.

Gravimetric soil moisture content is determined by calculating the ratio of mass of water to

that of the wet soil. The mass of water is calculated by subtracting the mass of dry soil sample

from the mass of the wet one. Traditionally, the most frequently used definition for a dry soil is

the mass of a soil sample after it has come to constant weight in an oven at a temperature

between 100 and 110 C. Then the gravimetric soil moisture content (0dw) can be converted to

the volumetric soil moisture content (vb ) by use of the formula ofv = (Pb Pw )Odw, where Pb

is the bulk density of the soil, and p, is the density of water (Klute, 1986).

The analysis of extractable nitrate and ammonium included two main procedures: (1)

extraction of exchangeable ammonium, nitrate; and (2) determination of nitrate and ammonium

concentration with colorimetric method (Page et al, 1982).

The procedure of extraction is described as follows. Place 3 g of soil in a wide-mouth

bottle, and add 30 ml of 1M KC1. Stopper the bottle, and shake it on a mechanical shaker for 1









hour. Allow the soil-KCl suspension to settle until the supernatant liquid is clear (usually about

30 min). Then use a vacuum filter with a pore size of 0.45 pm to filter the solution.

Nitrate and ammonium concentrations were measured by colorimetric methods. The

special apparatus required for nitrate concentration determination was Rapid Flow Analyzer

(RFA), ALPKEM 300 Series (OI Corporation, College Station, TX). The apparatus for

ammonium concentration determination was Technicon Industrial Method AA II (Technicon

Instrument Corporation, Tarrytown, NY).

Biomass sampling was performed near the soil sampling locations. The sampling

frequency was once every two weeks. When sampling, the collected crop plants were stored in a

cooler on ice for transport. Then the samples were stored in the freezer with a temperature

around 0 C before processing.

A whole plant that had an average height in the plot was collected and divided into leaves,

stems, husks, cobs, and kernels for fresh weight and dry weight determination. The analysis of

the plant samples included measurement of the moisture and total Kjeldahl nitrogen (TKN) of

different plant parts. Plant roots were not considered in this project, because of the negligible

amount of nitrogen in the roots (Albert, 2002).

Fresh mass of each biomass sample was measured first. Then the samples were dried in the

oven for 48 years at a constant temperature of 60 C for 48 hours. The dry mass of each sample

was measure. Then the biomass moisture was calculated.

The Kjeldahl procedures generally employed for determination of total N involve two

steps: (1) digestion of the sample to convert organic N to NH4 -N, and (2) determination of

NH4 -N in the digest. The digestion is usually performed by heating the sample with H2SO4

containing substances that promote oxidation of organic matter and conversion of organic N to









NH4 -N. The substances generally favored are salts such as K2SO4 or Na2SO4, which increase

the temperature of digestion, and catalysts such as Hg, Cu, or Se, which increase the rate of

oxidation of organic matter by H2S04 (Page et al, 1982). In this study, the substances were

K2S04 and CuSO4. The determination of NH4 -N in the digest was conducted with colorimetric

method in the Analytical Research Laboratory (ARL), Institute of Food and Agricultural

Sciences, the University of Florida.

Yield sampling was conducted at the end of the experiment season, at the time of

physiological maturation of sweet corn, 70 to 80 days after planting. Ears in a sampling zone,

which consisted of a 6.1-meter (20 feet) section of the 2 center rows of each sampling site, were

completely collected whether the kernels were fully filled or not. The total plant numbers in this

zone were also counted. Then the collected ears were weighed and classified into three classes,

US #1, US #2, and Cull according to the classification standard of USDA on the quality of sweet

corn (USDA, 1962).

4.2.5 CERES-Maize Model Simulation

The CERES-Maize model of DSSAT (Jones et al., 2003) was calibrated with the

generalized likelihood uncertainty estimation (GLUE) method (see Chapter 3 for details). In this

method, the second posterior distributions of the selected sensitive input parameters were used

(Table 4-5). See Appendix A for the definition and units of the parameters in the table.

Genotype parameters P1, P5, and PHINT described the genetic properties of the sweet corn

planted. Their values could be obtained through controlled experiment. However, there was no

experiment designed in this study to directly measure those genotype parameters. So a GLUE

verification procedure was conducted in Chapter 3 to guarantee the accuracy of these genotype

parameters (See Section 3.3.8 Chapter 3 for details).









The values of some soil parameters were measured in the experiment site (Table 4-1). If

compared the measured and estimated mean values of soil parameter SLLL, SDUL, and SSAT

(Table 4-6), it can be found that the mean values of estimated and measured soil parameters were

pretty close to each other. For example, the mean value of calibrated SDUL in the second

posterior distribution was 0.104 cm3/ cm3, while the mean value of measure SDUL was 0.110

cm3/ cm3. The error was only about 0.006 cm3/ cm3. Similar results were observed in SLLL and

SSAT.

In general, the uncertainties of the selected 9 input parameters were dramatically decreased

after two rounds of GLUE estimations. The second posterior distributions (Table 4-5) can be

used to simulate the real field experiment of sweet corn, since the output uncertainties were

reduced after GLUE simulations (see Chapter 3 for details). In this research, the new mean value

vector and covariance matrix derived from the second round of GLUE process were used to

generate random parameters.

In this study, the seven treatments (FOI1, Fll, F211, F311, F112, F2I2, and F3I2)

mentioned in Section 4.2.1 were run with the CERES-Maize model under the weather and

management conditions of field plot experiment in 2006. For each treatment, 3,000 simulations

were conducted with 3,000 different parameter sets that were randomly generated by the

posterior distributions. Then the results were recorded and the mean and variance of yield,

anthesis date, maturity date, and accumulative nitrogen leaching of each treatment were

calculated.









4.3 Results and Discussion

4.3.1 Quantity of Sweet Corn Yield

Quantity of sweet corn yield was defined as the fresh mass of the ears collected in a unit

area in this study. A complete ear included husks, kernels, and cob. The total weight of US #1

and US #2 yield was defined as marketable yield.

As described in Section 4.2.1, the field experiment was designed as a split-plot experiment.

An ANOVA analysis was performed to determine treatment effects on yield quantity with SAS

program (SAS Inst. Inc., 1996). The results of ANOVA were specified in Table 4-7. See

Appendix E for detailed SAS program.

It can be found in Table 4-7 that irrigation levels (P=0.1068) and interactions between

irrigation and nitrogen fertilizer (P=0.7434) were not significant. However nitrogen fertilizer

levels showed significance influence on yield quantity of sweet corn. A similar ANOVA analysis

was also performed for marketable yields, which showed the interaction was also not significant.

The irrigation and nitrogen treatment effects on total yield and marketable yield are shown

in Table 4-8. It can be found that whether for total yield or marketable yield, irrigation level did

not show significant influence, though the average total yield increased from 18,618 kg ha-1 to

20,091 kg ha-1, and marketable yield increased from 16,681 kg ha-1 to 18,431 kg ha-1.

However, nitrogen level showed significant influence on both total yield and marketable

yield. There was a significant difference between the total or marketable yields under Fl and F2,

which means when increasing the nitrogen application from 185 to 247 kg ha-l, the yields were

increased significantly. But, there was not a significant difference between F2 and F3, thus

increasing nitrogen fertilizer beyond 247 kg N ha-1 did not significantly increase yield.









Figure 4-5 and 4-6 show the difference of total and market fresh yields between the

different nitrogen fertilizer levels under individual irrigation level II and 12. The same trend as

described above can be found from these figures.

For all of the histograms in this current publication, the upper error bar shows the

maximum value of the four duplicates of the treatment, while the lower error bar shows the

minimum one.

4.3.2 Quality of Sweet Corn Yield

According to the USDA (1962), "U.S. No.1" consists of ears of sweet corn of similar color

characteristics that are fresh and free from damage by freezing, cross pollination, denting, worms,

birds, fermentation, smut or other disease or other means. Each ear must have at least an average

of 10.2 cm of the cob covered with undamaged kernels, in addition to any good kernels that

would necessarily be lost in the usual method of trimming to remove damaged kernels. "U.S.

No.2" shall also meet the same color characteristics as "U.S. No. I", but each ear must have at

least an average of 7.6 cm of the cob covered with kernels. "Culls" consists of ears of sweet corn

that fail to meet the requirement of"U.S. No.2" grade.

As for yield quantity, ANOVA analyses were also conducted for total corn ears, US #1

ears, US #2 ears and culls harvested per hectare. For convenience, only the ANOVA results of

total ears per hectare were listed in Table 4-9.

It is easy to find that at a significance level of 0.05, no factor whether irrigation level, or

nitrogen level, or interaction between irrigation and nitrogen treatments, had a significant

influence on the total ears numbers of sweet corn harvest on a unit acreage of field. Actually, this

is reasonable, because the number of ears that can show up in a field should be mainly controlled

by the genetic properties of corn, rather than by field management.









In the results of ANOVA of US #1, US #2, and cull, interactions between irrigation and

nitrogen treatment also did not show significant influence. Thus, the effects of irrigation and

nitrogen levels on yield quality were listed in Table 4-10.

From Table 4-10, it can be seen that irrigation level did not significantly influence yield

quality though the number of US #1 per unit acreage increased a bit from I1 to 12. Fertilizer level

did not show significant influence on the number of US #2 and cull per unit acreage, either.

However, as it was expected, nitrogen level showed significant influence on the number of US

#1 per unit acreage. From Fl to F2, the total number of US #1 ears increased from 42,764 ears

ha-1 to 52,612 ears ha-1, which was an increment of almost 10,000 ears ha-l, i.e. the nitrogen level

improved the yield quality significantly. There was no significant difference between the US #1

ears under F2 and F3, which means the stimulus of nitrogen fertilizer on yield quality will be

limited after a special point of nitrogen application.

Figures about the influence of irrigation levels and nitrogen fertilize levels on yield quality

were drawn as well to visualize the trend of quality improvement. Figure 4-7 and 4-8 show the

number of ears under different nitrogen fertilizer levels, respectively under II and 12. Figure 4-9,

4-10 and 4-11 show the number of ears under different irrigation levels, respectively under F1,

F2 and F3. In these figures, the data of the control plots were also presented for reference,

however they were not used in ANOVA or Duncan's multiple range test.

Interestingly, it can be found in Figure 4-7 that when there was no nitrogen applied, even

though there was irrigation, the harvest ears were all "culls", which could not be sold in the

market. In Figures 4-9 to 4-11, it can also be found that when no irrigation was applied, the yield

quality was greatly deteriorated. Though there were ears available, most or all of them were









"Culls". Thus, it can be concluded that both adequate irrigation and nitrogen fertilizer are

necessary to guarantee both yield quantity and quality.

These figures also show that the number limit of "total ears" of a unit area was about

110,000 ears ha-1. The ear number limit of "U.S. No. 1" of a unit area was about 60,500 ears ha1.

The ear number limit of "U.S. No. 2" of a unit area was about 28,000 ears ha-l. There was no

obvious ear number limit for "Culls".

4.3.3 Nitrogen Balance Estimation

In this current research, the equation used by Meisinger and Randall (1991) to calculate

long-term potentially leachable total nitrogen, Npi was used as the fundamental equation to

estimate nitrogen leaching in the field experiment. The equation is:

Npl = N,,, t Noutput Nst (4-5)

where Ninput and Noutput are N entering and leaving the field between the top of the crop

canopy and the bottom of the soil sampling zone (90cm below the soil surface) respectively, and

ANA, is the change in N storage. Np1 was used as the budget-derived estimation of nitrogen

loading to groundwater during a crop growth season.

The components of the right side of Equation (4-5) were investigated before using the

equation to estimate the potential nitrogen leaching in sweet corn production.

4.3.3.1 Nitrogen input

There were four possible nitrogen sources in the corn production system. The first was the

N fertilizer applied (Table 4-3), which was the largest nitrogen source. The second source was

the initial organic or inorganic nitrogen already present in the soil profile. This could be

determined by the results of initial soil nitrogen sampling. The third source was the corn seeds,

because there was organic N present in them as protein, amino acids, and nucleic acids.









According to the research of Meisinger and Randall (1991), the seed of sweet corn could supply

an N input of 0.3 kg N ha-1. The forth source was the N from atmospheric deposition. Small

portions of atmospheric N2 could return to the soil in rainfall or through the effects of lightning.

An estimated 1013 g per year of N2 could be fixed and transformed in ammonia by lighting in the

world. According to the research of Li et al. (2002), the annual atmospheric N deposition rate in

Florida was about 11 kg N ha-1 year-. Since the whole growth season of sweet corn in north

Florida was only about 70 to 80 days, about one fifth of a year, the atmospheric N deposition in

the experiments could be estimated as about one fifth of the annual deposition, which was about

2.4 kg N ha-.

The fifth source was the N dissolved in irrigation water. Near the plot experiment site, four

wells were developed to monitor the nitrate and ammonium concentration in the groundwater.

Thus the nitrogen concentration data collected from these wells during the experiment could be a

good reference to estimate the nitrogen concentration for irrigation water. The average N-NO3

concentration was about 3.65 mg L-1, while the average N-NH4 concentration was about 0.22 mg

L1. The depth of irrigation level I1 was about 21.0 cm, while the depth of 12 was about 27.8 cm.

Thus, the total nitrogen contained in irrigation water of I1 and 12 were about 8.1 and 10.8 kg N

ha-1, respectively. See Appendix F for the details about changes of nitrate and ammonium

concentrations in these monitoring wells.

The last possible N source could be the N that was biochemically fixed in the soil by

specialized micro-organisms including bacteria, actinomycetes, and cyanobacteria. This process

is called nitrogen fixation. It occurs in plants that harbor nitrogen-fixing bacteria within their root

nodules (Cockx and Simonne, 2003). The best-studied example ofN fixation is the association

between legumes and bacteria in the genus rhizobium. The main legume crops commercially









grown in Florida are peanuts, snap bean and pink-eyed and black-eyed pea (Cockx and Simonne,

2003). But in this current research, the field was kept fallow before experiment and no legume

was planted during experiment. Thus this N source was considered negligible when establishing

the N balance.

4.3.3.2 Nitrogen output

The nitrogen output of the corn production system includes several components. The most

important N output was the N in corn tissues. This part of N could be estimated with the TKN

concentration of corn tissues, the weight of corn biomass and plant density at harvest. See

Appendix G for more information about the TKN concentration of leaves and stems of sweet

corn during the growth season.

The second N output was the final inorganic nitrogen in the soil profile, or the final

residual N, which was determined by analyzing the nitrate and ammonium concentrations in the

initial and final soil samples. The N stored in the 90-cm top soil profile at the first soil sampling

conducted at planting was defined as the initial N storage. The N stored in the 90-cm top soil

profile at the last soil sampling conducted at harvest was defined as the final N storage. Then the

net N residual in the 90-cm top soil profile was defined as the difference between the final N

storage and initial N storage. See Appendix H for more information about the nitrate and

ammonium nitrogen concentrations in soil profile during the growth season.

Another possible N output was the gaseous loss. It includes several physical and chemical

processes, such as volatilization of ammonia and denitrification of nitrate from farmland. Since

instruments were not installed during field experiments to measure the gaseous loss caused by

volatilization and denitrification, an estimation based on literatures was performed. According to

Liu et al. (2003), the miscellaneous gaseous N loss was about 4-7% of the total N fertilizer

application as urea in maize production. In this research, the total nitrogen mass was about 32%









of the solution, including 7.9% nitrate nitrogen, 7.9% ammoniacal nitrogen, and 16.2% urea

nitrogen. The urea nitrogen covers more than half of the total nitrogen. The nitrate nitrogen and

ammoniacal nitrogen can also be lost through denitrification and NH3 volatilization. Thus, the

gaseous loss in this experiment was estimated as 6% of total N fertilizer applied.

4.3.3.3 Nitrogen balance

Based on the data and assumptions mentioned above, N balance was established with

Equation (4-5) for the sweet corn experiments in 2006. Table 4-11 shows the N balance of a

replicate in Block 1 of the treatment F1I1 in 2006. Similar procedures were also conducted for

all of the replicates of other 5 treatments, F211, F3I, F 112, F2I2 and F3I2. The final estimated

amounts of potential N leaching of the seven treatments were summarized in Table 4-12.

Then an ANONA analysis was conducted to analyze the influence of irrigation level,

fertilizer level, and their interaction on nitrogen leaching. The result is summarized in Table 4-13.

From Table 4-13, it can be found that under the confidence level of 0.05, both irrigation and

nitrogen fertilizer levels showed significant influence on nitrogen leaching, especially the

nitrogen fertilizer levels. This confirmed that common assumption that more water applied, more

nitrogen will be leached, and more nitrogen fertilizer applied, more nitrogen will be leached as

well.

Since the interaction does not show significant influence on nitrogen leaching, the average

nitrogen leaching amounts estimated from N balance under different irrigation and N fertilizer

levels were summarized in Table 4-14. From the results in Table 4-14, the increase of nitrogen

leaching caused by the increasing irrigation water and nitrogen fertilizer application was more

obvious. Under II, the average potential nitrogen leaching was 150.26 ka ha-l, then it increased

to 167.00 ka ha-1 when under 12. The average amount of potential nitrogen leaching was 124.17,









146.89, and 204.83 ka ha-1 respectively when under Fl, F2, and F3. The increment was higher

than under different irrigation levels.

4.3.4 Comparison between Model Simulations and Field Observations

4.3.4.1 Comparison between dry matter yields

The seven treatments (FOI1, F ll, F211, F311, F112, F212, and F3I2) that were investigated

in field plot experiment were also simulated with the CERES-Maize model under the weather

and management conditions of the field plot experiment in 2006. The model simulation results

were recorded and the mean and variance of yield, anthesis date, maturity date, and accumulative

nitrogen leaching of each treatment were calculated.

Table 4-15 shows the simulations and measurements of dry yields of the seven treatments.

The simulated results were relatively close to the measured results for all treatments except for

FOI1 and F 112. The absolute relative errors (ARE) between the measured and simulated yields

were all near or less than 10% except for treatment FOI1 and F 12. The simulated mean dry

matter yield of treatment F211 was 3023 kg ha-1., while the measured one was 2902 kg ha-1. The

difference was only about 100 kg ha-1, or 4% of the measured yield.

However, for treatment FOI1 and F 112, the simulated yields were higher than the measured

results. For FOI1, there was some nitrogen contributed by senesced organic matter and soil

nitrogen (nitrate and ammonium) in the soil profile in the initial conditions of the CERES-Maize

model. For example, the model assumed the nitrogen from dead organic matter was 7 kg N ha-1.

The initial nitrate-N and ammonium-N concentration was 0.1 g N Mg-1 and 0.5 g N Mg-1, which

was equal to about 1.3 and 6.7 kg N ha-l. And there was also a starter N application of 15 kg N

ha-1 at planting. Thus, the available N for sweet corn growth was about 30 kg N ha-1 even when

there was no additional nitrogen fertilizer application. The model probably underestimated the

influence ofN stress on corn yield and finally had a higher dry matter yield than the field









experiment. However, for F 112, the source of difference is not clearly known. Probably it was

due to the uncertainties in initial soil N and organic matter condition.

It can also be seen that the measured yields had relatively higher values of standard

deviations compared to the simulated values. This is because the observed yields suffered many

kinds of uncertainties and non-uniformity in weather and management, while the CERES-Maize

model just assumed these factors were homogeneous throughout the area.

4.3.4.2 Comparison between phenology dates

The simulated and observed anthesis and maturity dates of sweet corn were summarized in

Table 4-16. The observed the anthesis and maturity dates for different treatments were the same.

So the standard deviations of them were just set as zero and not listed. It were found that the

calibrated model performed very well in predicting these important phenology dates. This is true

because the model was just calibrated with the data collected from a similar field experiment

near the plots with the same corn genotype, nitrogen fertilizer, soil, etc.

4.3.4.3 Comparison between potential nitrogen leaching

Before comparing the simulated and estimated, it was necessary to check the N balance in

the model simulation. The main N input of model simulation included: inorganic N applied

(identified as NICM in DSSAT) or nitrogen fertilizer, initial nitrate-N in soil profile, initial

ammonium-N in soil profile, and nitrogen from senesced plant matter. In this research, the

nitrogen fertilizer application amounts (NICM) for each application were listed in Table 4-3. The

initial nitrate-N and ammonium-N concentration was 0.1 g N Mg-1 and 0.5 g N Mg-1, which

equal to about 1.3 and 6.7 kg N ha-'.The nitrogen obtained from dead organic matter was

calculated as 7 kg N ha-1 by the model. The main N output of model simulation included three

main components: nitrogen uptake during season (NUCM), nitrogen leached during the season

(NLCM), and inorganic N at maturity in soil (NIAM).









In this research, treatment FiI1 was used as an example to show the nitrogen balance in the

simulation of sweet corn growth with the CERES-Maize model. The result was shown in Table

4-17. It can be seen that when under FIl1, about 94 kg N ha-1 from the total application of 184

kg N ha-1 was utilized by sweet corn biomass. The N utilization efficiency was about 51%. The

amount of nitrogen leaching during the season was about 32 kg N ha-l, while the amount of

inorganic nitrogen at maturity in soil was 73 kg N ha1.

Similar balance calculations were also conducted for other six treatments (F2I1, F3I1, F112,

F2I2, F3I2, and FOI1). Since the amount of nitrogen leaching was concerned in this research, the

amounts of NLCM and NIAM were summarized in Table 4-18. It can be seen that NLCM only

covered a part of the N output except for NUCM. A significant part of N output was calculated

as NIAM in soil profile. NLCM represented the nitrogen that had already been leached into

groundwater, while NIAM represented the inorganic nitrogen (nitrate and ammonium) that was

still in the soil profile at corn maturity. Thus if we only consider NCLM as the value of potential

nitrogen leaching in this research, it might be unrepresentative and misleading. The NIAM also

would be subject to leaching after harvest due to rainfall. From a long-term point of view, the

total potential N leaching should be the sum of NLCM and NIAM.

A comparison was conducted between the simulated and estimated potential nitrogen

leaching in Table 4-19. It can be seen that the for treatment FIl1, F2I2, and F3I2, the results

were close since they all had a value of absolute relative error (ARE) less than 10%. However,

for treatment FOI1, F2I1, F3I1, and 1112, the difference was greater, since all values of ARE were

all greater than 20%. This difference was probably caused by the uncertainties in the process of

estimation of nitrogen leaching in field experiment since many components in the nitrogen

balance were not directly measured and just obtained from literature or other source. For









example, in the estimation of nitrogen leaching for treatment F1ll, the sum of nitrogen input by

corn seed, atmospheric deposition, and irrigation water was about 10.8 kg N ha-', which was

only about 5.8% of the input as nitrogen fertilizer (185 kg N ha-1). It is obvious that nitrogen

fertilizer contributed the main nitrogen input. The amount of nitrogen fertilizer applied was

measured and reliable. The uncertainty contribution by corn seed, atmospheric deposition, and

irrigation water should be insignificant. However, for nitrogen output, the amount of plant

utilization, net soil nitrogen residue, and gaseous loss, was 51.4, -15.9, and 11.1 kg N ha-1,

respectively. The estimation of gaseous loss might contribute some uncertainty when

determining nitrogen output. The value of net soil nitrogen residual was negative, which means

some of the initial soil N residue was used by the plant. The inaccuracy in soil nitrogen

concentration measurement could also contribute some uncertainty. The biomass TKN of each

plot was estimated with only one average plant, so uncertainty was inevitable. In general, the

uncertainties in nitrogen output could weaken the reliability of the estimation of nitrogen

leaching in this study. If all of the uncertainties in the process of estimation were considered, the

difference between the simulated and estimated potential nitrogen leaching would not be

surprising and unacceptable. Thus to obtain more confidence on the value of nitrogen leaching, a

direct measurement should be conducted, for example, with lysimeter.

It should also be noticed that if comparing the NLCM in model simulation (Table 4-18)

with the estimated potential nitrogen leaching (Table 4-12), the difference would be large and

hard to interpret. And under FOI1 when there was no nitrogen applied, there was still some

nitrogen leaching, which was due to the initial inorganic N in soil profile and N from dead plant

matter.









In general, there was a little higher difference between the simulated and estimated

potential nitrogen leaching for some treatments. The calibrated CERES-Maize model did a pretty

good job at predicting the growth and influence of management on yields of sweet corn. It could

be a good tool to help explore some potential optimal management practices for sweet corn

production in North Florida.

4.4 Conclusions

Sweet corn is a very important economic crop in Florida. Nitrogen fertilizer applications

and irrigation levels had dramatic effects on sweet corn quantity and quality. Field plot

experiment and model simulation with the CERES-Maize model were conducted in North

Florida in 2006 to explore the relationships among them. Several conclusions could be drawn as

follows.

N fertilizer level was significant in improving both fresh total yield and fresh marketable

yield, while irrigation level and interaction between N fertilizer level and irrigation level was

not significant. N fertilizer level was not significant in increasing total ears, or US #2, or cull

ears per unit area, but significant in improving the number of US #1 ears. Irrigation level and

the interaction between irrigation and N fertilizer did not show any significance for yield

quality.

The results show that 247 kg N ha-1 was adequate to guarantee the yield quantity and

quality in sweet corn production. More nitrogen application, e.g. 309 kg N ha-1, did not

significantly improve yield quantity and quality. Irrigation level II, which was based on daily

ETo, could guarantee the yield quantity and quality, since irrigation level 12 (1.5 X I1) did not

significantly improve yield quantity and quality.









According to the results of ANOVA of nitrogen leaching estimated from nitrogen balance,

both irrigation and nitrogen fertilizer levels showed significant influence on nitrogen leaching.

This confirmed that common assumption that more water applied, more nitrogen will be leached,

and more nitrogen fertilizer applied, more nitrogen will be leached as well. However, the

interaction does not show significant influence on nitrogen leaching.

After comparing the simulated and observed dry matter yields, anthesis dates and maturity

dates, and estimated nitrogen leaching of the seven treatments in field plot experiment of sweet

corn in 2006, it shows that the model did a good job in predicting dry yield and phenology dates.

There was a little larger difference between the simulated and estimated amount of potential

nitrogen leaching for some treatments. This is probably mainly because of the uncertainties in

the process of estimation of potential N leaching in field plot experiment. Thus the results were

not surprising or unacceptable.

In general, the calibrated CERES-Maize model did a very good job at predicting the

growth and influence of management on yields of sweet corn. It can be a good tool to help

explore some potential optimal management practices for sweet corn production in North Florida.















Block 1 Block 2 Block 3 Block 4 Control N
ImgatonMclhe

Movement Direction

oflrrigationMachine F F F2 F






51' F3 F2 F1 F3 F1 F2 F3 F3 FO F2
.--35 -- -T- -
S F3 F2 F2 F2











Notes:
1. Totally 30 plots including factorial experiment 21 x 3F x 4Block=24 plots, and6 control plots.
2. I =rigationbased on dailyET value, 12=1 .5 l, ad I=no iigatio.
3. F1=fertilizer level i, F2=fertilizer level 2, F3=fertilizerlevel 3, while fertilizer levels randcmizedin each irrigation plots.
4. FO= no fertilizer aipplicati.
5. The layoutpattem f sumain lines of F1 and F3 is as saf s as F2

Figure 4-1. Experiment plot arrangement layout







































4W


'4W


Figure 4-2. Soil moisture at t=30 minutes with 1, 2, 3 and 4 drip tapes. The different blue colors
from light to dark represent different soil moistures from higher to lower.


_ I r_


_ ~ I I_\ ~










Tape 3


Tape 4


Figure 4-3. Drip tape arrangement in each row interval


SN





Notes:
1. The scheme in only half of the
real plot
2. R=Row of sweet cor. There are
6 rows of sweet corn in eachplot
3. There are 4 drip lines between
two rows of sweet corn
4. Drip line layout pattern in other 4
intervals is as same as the interval
betweenR1 andR2.


Figure 4-4. Drip tape arrangement and sampling zone in each plot


A740


Tape 1


Tape 2


U


U M


1


- - . . - - ----


~i~iC~. V


U












25000


20000


S15000
*A

10000

LL
5000


0


11
Fertilizer Levels

U Total Yield U Market Yield

Figure 4-5. Fresh yield under different N fertilizer levels under II. The upper error bar shows
the maximum value of the four duplicates of the treatment, while the lower error bar
shows the minimum one.


25000


20000

r(

) 15000
o

S10000
U)
LL
5000


0


12
Fertilizer Levels

U Total Yield U Market Yield

Figure 4-6. Yield under different N fertilizer levels under 12. The upper error bar shows the
maximum value of the four duplicates of the treatment, while the lower error bar
shows the minimum one.










140000


120000

S100000

80000

00
40000
620000
0



0


~Iri


Fertilizer Levels
STotal US#1 0 US#2 O Cull
Figure 4-7. Number of ears per unit area under different N fertilizer levels under II. The upper
error bar shows the maximum value of the four duplicates of the treatment, while the
lower error bar shows the minimum one.


140000

120000

= 100000

S80000
w
Lu
60000

E 40000
Z
20000

0


F2
12
Fertilizer Levels


F3


U Total U US #1 O US #2 O Cull
Figure 4-8. Number of ears per unit area under different N fertilizer levels under 12. The upper
error bar shows the maximum value of the four duplicates of the treatment, while the
lower error bar shows the minimum one.


F1
F1











140000

S120000
-c
E 100000
0)
80000

W 60000
0
S40000
E
z 20000


11 12 10


F1
Irrigation Levels

E Total E US #1 O US #2 O Cull


Figure 4-9. Number of ears per unit area under different irrigation levels under F 1. The upper
error bar shows the maximum value of the four duplicates of the treatment, while the
lower error bar shows the minimum one.


140000

120000

100000

80000

60000

40000

20000


12

F2
Irrigation Levels


E Total E US #1 O US #2 O Cull

Figure 4-10. Number of ears per unit area under different irrigation levels under F2. The upper
error bar shows the maximum value of the four duplicates of the treatment, while the
lower error bar shows the minimum one.











140000

S120000
x--
r 100000

80000
LU
60000

n 40000
E
z 20000

0
11 12 10

F3
Irrigation Levels

E Total E US #1 O US #2 O Cull

Figure 4-11. Number of ears per unit area under different irrigation levels under F3. The upper
error bar shows the maximum value of the four duplicates of the treatment, while the
lower error bar shows the minimum one.










Table 4-1. Soil properties of the experiment site
Depth Clay Silt Sand Bulk
Depth ture Clay Silt Sand PWP FC Saturation
(cm) (%) (%) (%) (g/cm3) (cm3/cm3) (cm3/cm3) (cm3/cm3)
0-15 Sandy soil 2.75 1.92 95.33 1.67 0.051 0.110 0.313
15-30 Sandy soil 2.56 2.35 95.08 1.69 0.061 0.117 0.317
30-60 Sandy soil 2.36 1.76 95.88 1.67 0.077 0.118 0.357


Table 4-2. DUiq values of 4 different numbers of drip tapes at 3 depths at t=30min
Depth 1 Drip Tape 2 Drip Tapes 3 Drip Tapes 4 Drip Tapes
Dl (10cm) 0.58 0.42 0.82 0.97
D2 (20cm) 0.65 0.51 0.55 0.99
D3 (30cm) 1.00 1.00 1.00 1.00


Table 4-3. Fertigation schedules of field plot experiment in 2006
Date F1 (kg N ha-1) F2 (kg N ha-) F3 (kg N ha')
3/14/06 15 15 15
4/7/06 27 41 55
4/12/06 21 28 35
4/19/06 21 28 35
4/26/06 21 28 35
5/3/06 21 28 35
5/10/06 21 28 35
5/17/06 21 28 35
5/24/06 17 23 29
Total 185 247 309


Table 4-4. Crop coefficients of sweet corn at different stages of development
Growth Stage Time (weeks after planting) KC
1 Planting-2 0.15
2 3-4 0.30
3 5-6 0.50
4 7-8 0.65
5 9-10 1.00
6 11-Harvest 0.90










Table 4-5. Second posterior distribution of the selected parameters
Parameter Unit Min Max Mean Standard Deviation CV
P1 Cd 77.6758 182.1748 99.1689 8.2169 0.0829
P5 Cd 553.1408 676.2120 577.2011 9.7462 0.0169
PHINT Cd 39.1615 41.7123 39.6760 0.2021 0.0051
SLDR 0.7079 0.7521 0.7316 0.0063 0.0086
SLRO 41.4916 99.8501 78.1428 9.6603 0.1236
SDUL cm3/cm3 0.0970 0.1093 0.1044 0.0016 0.0155
SLLL cm3/cm3 0.0526 0.0684 0.0601 0.0024 0.0401
SSAT cm3/cm3 0.2352 0.3624 0.3002 0.0209 0.0695
SLPF 0.7595 0.9322 0.8720 0.0414 0.0474



Table 4-6. Measured and estimated mean values of soil properties of the field experiment site
SLLL (cm3/ cm3) SDUL (cm3/ cm3) SSAT (cm3/ cm3)
Measured Estimated Measured Estimated Measured Estimated
Mean 0.051 0.060 0.110 0.104 0.314 0.300
STDEV 0.031 0.002 0.044 0.002 0.07 0.021



Table 4-7. ANOVA results of total yield of sweet corn
Source of Variance Degrees of Freedom Sum of Squares F Value P-Value CV
Block 3 41846554.24 3.26 0.0596
Irrigation (I) 1 13012402.31 3.04 0.1068
Irrigation Error 3 5059041.81 0.39 0.7597
Nitrogen (N) 2 57707644.3 6.74 0.0109
IxNa 2 2602660.19 0.3 0.7434
Error 12 51375030.4
Total 23 171603333.2 10.69
a Interaction between irrigation and nitrogen fertilizer treatments.










Table 4-8. Irrigation and nitrogen treatment effects on yield quantity
Total Yield (kg ha ') Marketable Yield (kg ha ')
Irrigation Level
11 18,618 aa 16,681 a
12 20,091 a 18,431 a
Nitrogen Level
Fl 17,182 Bb 15,255 B
F2 20,181 A 18,584 A
F3 20,701 A 18,828 A
I x N Interaction NS NS
CV 10.69 11.56
NS: non-significant.
a Means with columns followed by the same lowercase letters are not significantly different (p < 0.05 )
according to t-test;
b Means with columns followed by the same uppercase letters are not significantly different (p < 0.05)
according to Duncan's multiple range test.


Table 4-9. ANOVA results of total ears of sweet corn
Source of Variance Degrees of Freedom Sum of Squares F Value P-Value CV
Block 3 665639481.3 1.78 0.2041
Irrigation (I) 1 9511908.6 0.08 0.7869
Irrigation Error 3 9511908.6 0.03 0.9942
Nitrogen (N) 2 686895685.2 2.76 0.1034
IxNa 2 60662682.4 0.24 0.7876
Error 12 1494146132
Total 23 2926367798 11.27
a Interaction between irrigation and nitrogen fertilizer treatments.










Table 4-10. Irrigation and nitrogen treatment effects on yield quality
Total Ears US #1 US #2 Cull
(ears ha-1) (ears ha-1) (ears ha-1) (ears ha 1)
Irrigation Level
I1 99,648 aa 46,317 a 27,160 a 26,171 a
12 98,389 a 53,152 a 23,563 a 21,674 a
Nitrogen Level
Fl 92,274 B b 42,764 B 24,957 A 21,674 A
F2 99,423 AB 52,612 A 25,901 A 20,910 A
F3 105,359 A 53,826 A 25,227 A 26,306 A
I x N Interaction NS NS NS NS
CV 11.27 16.10 23.73 34.25
NS: non-significant.
a Means with columns followed by the same lowercase letters are not significantly different (p < 0.05 )
according to t-test;
b Means with columns followed by the same uppercase letters are not significantly different (p < 0.05 )
according to Duncan's multiple range test.


Table 4-11. Nitrogen budget of a replicate of treatment FiI1 in Block 1 of the plot experiment
Component Item Part Value Unit
N Fertilizer 184.94 kg ha1
Input Seed 0.3 kg ha1
Atmospheric Deposition 2.41 kg ha-1
Irrigation Water Nitrate 7.66 kg ha-1
Irrigation Water Ammonium 0.45 kg ha'1
0-15cm -0.88 kg ha1
15-30cm -1.76 kg ha1
Soil Nitrate 30-60cm 0.50 kg ha'
60-90cm -0.34 kg ha'
Net Res l Subtotal -2.47 kg ha1
Net Residual i
0-15cm -5.27 kgha1
15-30cm -4.41 kg ha1
Soil Ammonium 30-60cm -9.23 kg ha1
60-90cm 5.46 kg ha1
Subtotal -13.46 kg ha1
Volatilization and 11 1 -
Gaseous Loss Volatilization and 11.10 kg ha1
denitrification
Cobs 5.14 kg ha1
Husks 2.47 kg ha1
Output Kernels 28.37 kg ha-
Plant Uptake .
Plant Uptake Leaves 12.94 kg ha-
Stems 2.51 kg ha1
Subtotal 51.42 kg ha1
Potential
Potential 149.17 kg ha-
Leaching










Table 4-12. Estimated nitrogen leaching of seven treatment in field plot experiment
Treatment Mean (kg ha ') STDEV (kg ha')
FOIl 17.28 7.11
Fll 115.15 17.89
F2I1 139.97 19.11
F3I1 195.65 15.26
F1I2 133.18 23.15
F2I2 153.81 15.38
F3I2 214.01 13.20


Table 4-13. ANOVA results of nitrogen leaching estimated from N balance
Source of Variance Degrees of Freedom Sum of Squares F Value P-Value CV
Block 3 1013.03103 1.03 0.4145
Irrigation (I) 1 1681.80989 5.12 0.0429
Irrigation Error 3 643.05542 0.65 0.5962
Nitrogen (N) 2 27681.78488 42.17 <.0001
IxNa 2 25.42077 0.04 0.9621
Error 12 3938.75866
Total 23 34983.86065 11.42
a Interaction between irrigation and nitrogen fertilizer treatments.


Table 4-14. Irrigation and nitrogen treatment effects on cumulative nitrogen leaching estimated
from N balance
Accumulative nitrogen leaching (kg ha 1)
Irrigation Level
11 150.26 b a
12 167.00 a
Nitrogen Level
Fl 124.17 C b
F2 146.89 B
F3 204.83 A
I x N Interaction NS
CV 22.57
NS: non-significant.
a Means with columns followed by the same lowercase letters are not significantly different (p < 0.05 )
according to t-test;
b Means with columns followed by the same uppercase letters are not significantly different (p < 0.05)
according to Duncan's multiple range test.










Table 4-15. Simulated and measured dry yields in field plot experiment in 2006
T t Simulated (kg ha-1) Measured (kg ha 1)
Treatment
Mean STDEV CV Mean STDEV CV AREa
FOI1 1438 679 47% 152 33 21% 846%
Fll 2843 664 23% 2533 276 11% 12%
F2I1 3023 655 22% 2902 389 13% 4%
F3I1 3024 655 22% 2943 220 7% 3%
F1I2 3377 933 28% 2621 231 9% 29%
F2I2 3419 949 28% 3152 463 15% 8%
F3I2 3447 960 28% 3268 324 10% 5%
a ARE is the absolute relative error, defined as ARE=IY-Y'I/Y, where Y is the measured value and Y' is the
predicted value of model output.


Table 4-16. Simulated and measured anthesis and maturity dates in field plot experiment
Anthesis Date Maturity Date
Simulated Observed Simulated Observed
Treatment
Mean STDEV Mean STDEV
FOI1 51 1 51 81 2 80
Fll 51 1 51 81 2 80
F2I1 51 1 51 81 2 80
F3I1 51 1 51 81 2 80
F1I2 51 1 51 81 2 80
F2I2 51 1 51 81 2 80
F3I2 51 1 51 81 2 80



Table 4-17. Nitrogen balance of model simulation of treatment F111
Component Item Value Unit
Inorganic N Applied or N Fertilizer (UNIM) 184 kg ha-1
Initial Soil Nitrate-N 1.3 kg ha-1
Input Initial Soil Ammonium-N 6.7 kg ha-1
N from Senesced Plant Matter 7 kg ha-1
Subtotal 199 kg ha-1
N Uptake during Season (NUCM) 94 kg ha-1
N Leached during Season (NLCM) 32 kg ha 1
Output
Inorganic N at Maturity (NIAM) 73 kg ha 1
Subtotal 199 kg ha-1
Potential N Leaching NLCM+NIAM 115 kg ha'










Table 4-18. Simulated potential
N Leached during
Season (NLCM)
Treatment kg ha
Mean


nitrogen leaching of the seven treatment in field plot experiment
Inorganic N at Potential N Leaching
Maturity (NIAM) (NLCM+NIAM)
kg ha kg ha -


Mean


Mean


STDEV


FOI1 9.04 11.92 20.96 2.11
Fll 31.95 73.21 105.16 9.93
F2I1 32.58 144.81 177.39 9.74
F3I1 33.12 206.14 239.27 9.69
F1I2 55.48 47.13 102.61 15.72
F2I2 66.23 97.42 163.64 15.86
F3I2 76.48 148.36 224.85 16.04



Table 4-19. Simulated and estimated accumulative nitrogen leaching in field plot experiment
Treatment Simulated (kg ha-1) Estimated (kg ha ) AREa
Treatment AREa
Mean STDEV CV Mean STDEV CV
FOIl 20.96 2.11 10% 17.28 7.11 41% 21%
Fll 105.16 9.93 9% 115.15 17.89 20% 9%
F2I1 177.39 9.74 5% 139.97 19.11 11% 27%
F3I1 239.27 9.69 4% 195.65 15.26 7% 22%
F1I2 102.61 15.72 15% 133.18 23.15 13% 23%
F2I2 163.64 15.86 10% 153.81 15.38 12% 6%
F3I2 224.85 16.04 7% 214.01 13.20 7% 5%


a ARE is the absolute relative error, defined as ARE:
predicted value of model output.


=|Y-Y'I/Y, where Y is the measured value and Y' is the









CHAPTER 5
BEST MANAGEMENT PRACTICE DEVELOPMENT WITH CERES-MAIZE MODEL FOR
SWEET CORN PRODUCTION IN NORTH FLORIDA

5.1 Introduction

Best management practices (BMPs) are specific cultural practices that are aimed at

reducing the loads of specific compounds while increasing or maintaining economical yields

(Simonne and Hochmuth, 2003). The implementation of BMPs may be key in reducing the

consequences of alterations of the N cycle in sweet corn fields. Implementation of BMPs at the

farm level is a key to maintaining the quality and the quantity of ground and surface water

(Simonne and Hochmuth, 2003).

The planned BMPs related with irrigation and N fertilizer application could be obtained

with traditional field plot experiments. But development and certification of site-specific

guidelines for optimal timing and water and nitrogen requirements requires extensive and

expensive field experiments. Since it is impossible to test all the interactions between the amount

of water and nitrogen during the seasons, use of simulation models can greatly facilitate the

evaluation of different production practices and/or environments and thereby streamline the

decision-making process (Rinaldi et al., 2007).

Due to the important roles of nitrogen application and irrigation in sweet corn production

in Florida, these two factors were studied to obtain relevant potential BMPs for sweet corn

production in North Florida. For irrigation, the focus was on irrigation rate and timing. The rate

was mainly dependent on the water balance in the relevant soil profile. The timing of irrigation

could be determined both by water balance or soil moisture status. For nitrogen fertilizer, the two

factors were total application amount and application split. The nitrogen application split was the

number of times that N fertilizer was applied, while the application amount was how much

fertilizer was applied in each application.









Previous BMP development depended on sampling experiments. Some initial traditional

efforts to study yield variability of crops have focused on taking static measurements of soil,

management, or plant properties and regressing these values against grid level yields (Jones et al.,

1989; Cambardella et al., 1996; Khakural et al., 1996; Sudduth et al., 1996). However, these

efforts have proven to be illusive in determining causes of yield variability. The reason for this is

because crop yield is influenced by temporal interactions of management, soil properties, and

environment. Traditional analytical techniques, which regress static measurements against yield

do not account for temporal interactions of stress on crop growth and yield (Paz, et al., 1999). In

addition, experiment time and cost should also be considered.

With the development of computer technology, crop models fall into our eyesight as a

strong tool for exploration of possible BMPs for crop production. A crop model has been

described as a "quantitative scheme for predicting the growth, development and yield of a crop,

given a set of genetic coefficients and relevant environmental variables" (Monteith, 1996).

Models are not perfect, and can at best only represent a current understanding of biological

systems; yet they do highlight where information and understanding are lacking (Boote et al.,

1996). With these caveats, crop models can be used to predict crop growth, development and

yield as a function of soil, climate, weather, and crop management conditions (Ghaffari et al.,

2001). The CERES-Maize corn growth and yield model (Jones and Kiniry, 1986; Tsuji et al.,

1994) in the Decision Support System for Agrotechnology Transfer (DSSAT) model, V4.0, is a

popular crop model.

This current research demonstrates the use of the CERES-Maize model to develop some

potential BMPs for sweet corn (Zea mays L.) production on sandy soil in North Florida. The

objective was to select management combinations of different irrigation and nitrogen fertilizer









levels that can simultaneously obtain acceptable yield and lower nitrogen leaching as potential

BMPs for future study.

5.2 Materials and Methodology

5.2.1 Experiment Site

In this study, a field experiment was necessary. First field data such as yield, anthesis date,

maturity date, corn leafN concentration, soil moisture, and soil nitrate concentration, were

required for model calibration with the generalized likelihood uncertainty estimation (GLUE)

method. Second some information such as planting date, corn population density, planting depth,

micro nutrient application, pest control etc., were required data to run the model.

The field experiments of sweet corn were conducted at the Plant Science Research and

Education Unit, the University of Florida in the spring of 2005 and 2006. The unit is located near

Citra (29.4094N, 82.1777W, 20.746 meters above sea level), Marion County, Florida. The

experiment field was identified as Blockl. The variety of sweet corn planted was Saturn SH2

(Judge et al., 2005).

In this study, the data collected in the experiment field identified as Block 1 (Figure 3-1 in

Chapter 3) were used for the GLUE simulation. In Block 1, there were only two treatments each

year, the high-nitrogen-level treatment and the low-nitrogen-level treatment, while the irrigation

level was the same. The size of Block 1 was about 9.0 acres and divided into two even parts for

these two treatments.

The nitrogen fertilizer used in the experiment was a composite solution of several nitrogen

compounds. The total nitrogen mass concentration was about 32%, including 7.9% nitrate

nitrogen, 7.9% ammoniacal nitrogen, and 16.2% urea nitrogen. The concentration of total

fertilizer solution was 1.294 kg L-1, while the concentration of nitrogen in this solution was 0.414

kg N L-1. When applying fertigation, nitrogen fertilizer solution was injected with an injection









pump into the drip tape system through the injection hole. The injection pump used in this

experiment was an Easy-Load II MASTERFLEX pump (Cole-Parmer Instrument Company,

Vernon Hills, Illinois).

A total of 230 kg N ha-1 was applied in the East Half with a starter application of 15 kg N

ha-1, while the rest was applied in eight even application. And 335 kg N ha-1 was applied in the

West Half in the same way as in the East Half. Other practices including irrigation were the same

for the two parts.

Since model input parameter calibration had already been done with the GLUE method in

Chapter 3, only the necessary information collected in the experiment of 2006 was used as input

for model running.

5.2.2 Crop Model Calibration

The crop model CERES-Maize used in this study is embedded in the Decision Support

System for Agrotechnology Transfer (DSSAT) software (Jones et al., 2003), version 4.0. It was

used to simulate the growth of sweet corn to find potential BMPs.

First, the model was calibrated. In this study, both the genetic parameters of the sweet corn

variety and the soil parameters were estimated through the generalized likelihood uncertainty

estimation (GLUE) method (See Chapter 3 for details about GLUE method and procedures).

In contrast to the commonly used methods of model calibration, the GLUE method

provided a posterior distribution of the input parameters rather than a unique parameter set that

can optimize all of the observations and relevant predictions. This study is a pre-selection of

potential BMPs for sweet corn production, which means the uncertainties in predictions caused

by uncertainties of input parameter should be temporarily neglected so as to reduce the number

of model runs and facilitate the study. Thus, the expectation values of the most sensitive input









parameters (Table 5-1) derived in Chapter 3 were used as the nominal parameter set to conduct

BMP simulations.

5.2.3 BMP Simulations

In this study, common cultural practices such as planting, weed management, disease

management, micro nutrient application, and harvest etc., were used for sweet corn production in

North Florida. More information is available in the "Vegetable Production Guide for Florida

2003-2004" (Olson and Simonne, 2005).

However, the focus of this study was on the cultural practices that most directly affect the

nitrogen cycle and yield of sweet corn production according to the definition of nitrogen BMP in

this study. So the computer experiment was designed mainly considering irrigation and N

fertilizer application.

Generally, the procedures of the BMP development consisted of the following steps: (1)

model simulation with different irrigation strategies; (2) model simulation with different

irrigation strategies; (3) model simulation with different irrigation and N fertilizer combinations;

and (4) identification of potential BMPs.

The irrigation events were scheduled with two methods. First, irrigation water was applied

automatically with specific irrigation depths that were derived from soil properties when 10, 20,

30, 40, 50, 60, 70, 80 and 90% of the maximum available water (MAW) in the upper 50 cm of

the soil profile was remaining, i.e. from 90 to 10% of the MAW was depleted. The DSSAT

model requires model users to provide the irrigation depth and remaining MAW. The soil profile

depth was set as 50 cm because more than 70% of sweet corn roots are concentrated in the top

two feet of soil (Bauder and Waskom, 2003).









To determine the irrigation depth in each irrigation event, the irrigation scheduling method

based on maximum allowable depletion (MAD) of the total available soil water (ASW) was

applied (Panda, 2004).

The ASW was taken as the difference between root zone water storage at field capacity

(FC) and permanent wilting point (PWP). Since soil is always inhomogeneous and can be treated

as layered medium, the value of ASW of a layered soil profile can be calculated with Equation

(5-1):

N
ASW = (FC, -PW)x RZ, (5-1)


where RZ, is the root zone depth of ith soil layer. N is the number of soil layers.

The soil of the experiment field is very sandy. It consists of Lake Sand, Candler Variant,

Tavares Variant, and Millhopper Variant 1 etc. Soil experiment was conducted at 24 sites at 3

depths of 0-15 cm, 15-30 cm, and 30-60 cm. Then the samples were sent to the lab of the Soil

and Water Science Department of the University of Florida. According to the definitions, the

permanent wilting point (PWP) was measured as the soil moisture at a soil pressure of 15.3 bar,

field capacity (FC) as the soil moisture at 0.1 bar, and soil saturation as the soil moisture at 0 bar.

The main measured properties of the soil at the experiment site are summarized in Table 5-2.

According to the FC and PWP values in Table 5-2, the 50 cm soil profile could be

separated into 3 layers: 0-15 cm, 15-30 cm, 30-50 cm. The calculated ASW value was 25 mm as

shown in Table 5-3.

The MAD is the maximum amount of depletion that can occur without stress to the plant.

Hence, the readily available water (RAW) is:

N
RAW = MADx ASW = MADx (FC,-PWE)x RZ, (5-3)
1









If the MAD value is 50%, it simply means, before half of the available water in the root

zone is depleted (either was evaporated, transpired, or has transpired or has traveled outside of

the root zone), supplemental irrigation is added to "refill" the reservoir. Allowing the depletion

amount to drop down below 50% can lead to plant stress for some plant material and once a plant

has reached PWP, no amount of water can be applied for recovery.

If an irrigation event was triggered at a threshold of 40% of the MAW remaining in the soil

profile in the DSSAT model, it equals to a MAD value of 60%, which means 60% of the ASW

could be used by sweet corn without stress. A supplementary irrigation should be triggered to

compensate this MAD. The necessary irrigation depth should equal to the value of RAW.

All of the irrigation depths could be calculated with the same method. With this scheduling

method, there were a total of 9 possible irrigation scenarios, identified as I to 9 (Table 5-4).

In addition, a set of irrigation schedules based on fixed days and depths were also tested. In

each schedule, irrigation water was applied with a fixed irrigation depth for 1 time (Wednesday),

2 times (Tuesday and Friday), or 3 times (Monday, Wednesday and Friday) every week during

the growth season. This type schedule was meant to represent typical sweet corn grower

practices. As mentioned in Table 5-3, the value of ASW was about 25 mm. If the precipitation

depth in each irrigation event was higher than 25 mm, deep percolation and runoff would happen.

Thus, the fixed irrigation depths here were set as 20, 40, 60, 80 and 100% of the ASW value,

which were 5, 10, 15, 20 and 25 mm. Hence, there were 5 x 3 = 15 possible irrigation scenarios

both considering irrigation times and irrigation depth, identified as 110 to 124 (Table 5-5).

The irrigation schedule based on water balance in soil profile was not considered, because

this schedule is dependent on daily rainfall and ET values, which would change year to year. In

this study, any irrigation schedule was run with the measured weather data of 33 years (1958-









1990). It is hard to automatically set the irrigation events for each year. And at the same time,

values of reference ET were also not available.

When doing irrigation simulations, the N application schedule was fixed as the one of East

Half of Block 1 in 2006 (See Section 5.2.1 for details), where a total of about 230 kg N ha-1 was

applied in eight applications. The initial values of soil volumetric moisture for each layer were

all set as 8.6%. The initial nitrate-N and ammonium-N concentration was 0.1 g N Mg-1 and 0.5 g

N Mg-1, which equal to about 1.3 and 6.7 kg N ha-1.The previous crop was set as cotton since it

was planted in Blcok 1 in the end of 2005 before sweet corn. The values of root weight, nodule

weight, and residue nitrogen and phosphorus, were all set as zero.

For nitrogen fertilizer, three factors should be considered. First is the total amount fertilizer

nitrogen. Totally 21 nitrogen fertilizer levels were simulated. They were the amounts from 0 kg

N ha-1 to 561 kg N ha-1 with increments of 28 kg N ha-1 (0 lb N acre-' to 500 lb N acre-' with

increments of 25 lb N acre-'). Among these levels, 224 kg N ha-1 (200 lb N acre-'), is

recommended by Institute of Food and Agricultural Sciences (IFAS), University of Florida

(Hochmuth, 2000). This level was defined as the standard nitrogen level for this research. Other

levels were derived from this recommended level. These N levels were identified as N1 toN21.

The second factor is when to apply the given amount of fertilizer nitrogen. Since inorganic

nitrogen is not stable in soil and becomes less available for crop uptake over time, application

time is important. Corn absorbs the majority of its nitrogen during rapid growth between 8-leaf

and dough growth stages (Bauder and Waskom, 2003). If nitrogen is insufficient during this

period, yield loss will occur. Application of nitrogen immediately before or during this period

will result in higher uptake by the crop and less nitrate lost to leaching or transformations to

unavailable forms. An application schedule that applies a small amount of nitrogen early in the









season (pre-plant or starter) followed by later, in-season applications of higher amounts of

nitrogen is ideal. This schedule takes care of the small, but important, early season nitrogen

needs and maximizes uptake by applying nitrogen during the rapid growth and nitrogen

requirement period, i.e. the nitrogen application should coincide with corn nutrient demand.

In this study, the growth season of sweet corn in North Florida was defined as three main

growth stages: small plant stage, large plant stage, and ear development stage as recommended

by the "Vegetable Production Guide for Florida 2003-2004" (Olson and Simonne, 2005). The

small plant stage roughly includes the first 4 weeks after planting, or from planting to 12 leaves

fully emerged. The large plant stage roughly includes week 5 to 7, or from 12 leaves to 20 leaves

fully emerged (tasseling/silking). The remaining 4 weeks (from tasseling to maturity) can be

considered as the ear development stage. These 3 stages will be used to determine the amounts of

N fertilizer applied at different time periods during the whole growth season.

A nitrogen split was defined as the quotient between the amount of nitrogen applied in

each stage and the total N amount except for the starter N fertilizer. This factor was considered

and tested to find the optimal time of nitrogen fertilizer application so that the time of N

application could coincide with the nitrogen need of sweet corn plant. Finally, 30 kinds ofN split

could be set up as shown in Table 5-6. For example, "1:0:0" means apply all of given amount N

except for the 15 kg N ha-1 as starter (see Section 5.2.1) during the small leaf stage and nothing

for the later periods. These splits were identified as S1 to S30. Based on the real nitrogen

application schedule in field experiment, the split of "0:1/2:1/2" was defined as the standard split.

The third factor is how much N fertilizer should be applied in each fertigation event. For

example, 224 kg N ha-1 was applied with a split of 0:1/4:3/4, which means 56 kg N ha-1 was

applied in the large leaf stage, while 168 kg N ha-1 was applied in the ear development stage. The









168 kg N ha-1 could be applied just in one event in week 8, or two events in week 7 and 9. It is

obvious that frequent application with small amount could be a good choice, because less

nitrogen would be leached. But the cost will also increase when the number of application

increased. To determine the optimal N fertilizer application in each event, "application amount"

should be considered.

In this study, 20 kinds of "application amount" ranging from 5 to 100 kg N ha-1 with a step

5 kg N ha-1 were investigated. All these kinds of"application amount" were identified

as Al to A20. And the "application amount" of 40 kg N ha-1 was set as the standard, because it

was close tot the "application amount" in field experiment.

When doing N simulations, the irrigation schedule was set to the actual field experiment in

2006, which was set up according to daily ET value and water balance in the top soil profile.

When doing single factor (such as total amount, split or application amount) analysis for N

application, the other factors were set at their defined standard levels. The initial values of soil

volumetric moisture for each layer were all set as 8.6%. The initial nitrate-N and ammonium-N

concentration was 0.1 g N Mg-1 and 0.5 g N Mg-1. The previous crop was set as cotton. The

values of root weight, nodule weight, and residue nitrogen and phosphorus, were all set as zero.

For each of the irrigation and nitrogen fertilizer treatment, the model was run under 33

years (1958-1990) of historical weather data at Gainesville, FL, the USA, which is about 32 km

from the experiment site of this study. These weather data were provided by the McNair Bostick

Simulation Lab of the Department of Agricultural and Biological Engineering, the University of

Florida. Then the average values of the dry yield and nitrogen leaching over these 33 years were

used to show their responses to different management strategies. In this way, the uncertainties of

simulation results caused by climate variability were considered. In another words, the potential









BMPs selected could be considered applicable in weather conditions represented by the

simulated time period.

5.2.4 Determination of Acceptable Yield

It is necessary to know the acceptable yield when determining the potential BMPs from the

combination scenarios of irrigation and N fertilizer practices. The acceptable yield in this study

was defined as the lowest yield that would be accepted by farmers. If the simulated yield was

below the acceptable yield, the strategy of irrigation and N fertilizer application was considered

as failure. Since there was no established "acceptable yield" for sweet corn production in Florida

found in the literature, an estimate had to be made according to currently existing sources.

One possible way is the National Agricultural Statistics Service (NASS) of the U.S.

Department of Agriculture (USDA). USDA proposed reports on "Florida: Acreage, yield,

production, and value of Florida vegetables, melons, potatoes, blueberries, and strawberries"

every year. In these reports, important information about sweet corn production in Florida could

be found, such as total planted and harvested acreage, yield, price and total value etc. Table 5-6

summarizes the information about sweet corn production from 1998-2007 (USDA NASS, 2007).

From Table 5-7, it can be seen that the average fresh yield of sweet corn in the past 9 years

in Florida was 16,939 kg ha-1 (15,111 lb acre-'). The average moisture of fresh ears of sweet corn

obtained from field experiment was as high as 84.0%, so the average dry yield was about 2,418

kg ha-1. The highest marketable fresh yield in the field experiment of this study with sufficient N

and water application could be 20,346 kg ha-1, and a corresponding dry matter yield of 3,255 kg

ha-l, which was almost 800 kg ha-l greater than the average dry matter yield obtained from the

statistics.

Another way to estimate the acceptable yield of sweet corn production in Florida is to

survey the results of related experiments conducted in Florida or southern United States.









In the study of yield, ear characteristics, and consumer acceptance of selected white sweet

corn varieties by Simonne et al. (1999), they summarized the yields of 10 varieties of white

sweet corn planted in Clanton, Ala., in 1995 and 1996 (Table 5-8). No information about N

fertilizer application was provided. The data in Table 5-7 shows the yield is strongly dependent

on sweet corn variety. For "Even Sweeter", the yield could be as high as 14,726 kg ha-1, but for

"Rising Star", it is only 8,291 kg ha-1, less than 60% of "Even Sweeter".

Mullins et al. (1999) studied the response of selected sweet corn cultivars to nitrogen

fertilization at Springfield, Tenn., in 1993, 1994 and 1995. Their results are shown in Table 5-9.

It can be seen that N fertilizer rate, cultivar, and year (climate) all had some kind of influence on

sweet corn yields. In general the yields in their experiment were lower than the ones listed in

Table 5-6.

Rangarajan et al. (2002) conducted research at Eden Valley and Freeville, NY, from 1998

to 2001. They concentrated on the in-row spacing and cultivar on sweet corn yield. The yields of

sweet corn in their experiment are summarized in Table 5-10. The yields shown in this table are

close to those listed in Table 5-7. It also shows that in-row spacing did not have a strong

influence on yield.

Shuler (2002) studied the effect of within-row plant spacing on sweet corn grown on muck

soil at Belle Glade, Florida in spring and fall 2001. In his research, 4 within-row spacing

treatments of 0.15, 0.18, 0.20 or 0.23 meters (6, 7, 8 and 9 inches), and 5 varieties were used.

Unfortunately, there was no information available about N fertilizer applications, either. The

yields obtained in his research are shown in Table 5-11.

Hochmuth and Cordasco (2000) summarized the field experiments conducted in Florida

from 1961 to 1997 to evaluate the sweet corn yield responses to varying rates of fertilizers. The









yields are shown in Table 5-12. The average fresh yield in the experiments summarized in Table

5-12 was only 15,596 kg ha-1. It was only a little bit higher than the average fresh yield (15,111

kg ha-1) specified in Table 5-6 from the Florida Agricultural Statistics.

From the research summarized above, it can be found that sweet corn yield was influenced

by many factors, such as location, variety, within-row spacing, N fertilizer level, etc. In many

cases, the average fresh yields were lower than 20,000 kg ha-1. The results specified in Table 5-

11 could be a good reference, since the research site was in south Florida and research time

period was in 2001, which were all comparable to the site and time of this current research. The

average fresh yield in Table 5-10 was 20,251 kg ha-1. This average yield is close to the results in

field experiments in this study, where the average yield was about 20,000 ka ha-1 (see Chapter 4

for details).

Finally, the fresh yield of 21,000 kg ha-1 and corresponding dry yield of 3,400 kg ha-1 at

average ear moisture of 84.0% were selected as a good estimation for the acceptable yield for

this study.

5.3 Results and Discussion

5.3.1 Effects of Irrigation

First, the response of sweet corn yield (dry matter, kg ha-1) and accumulative nitrogen

leaching (kg ha-1) to different irrigation strategies were investigated. As mentioned in Section

5.2.3, the first 9 irrigation strategies identified as Il to19 (Table 5-4), were designed as

automatic irrigation with specific depths triggered by remaining available soil water refill the soil

profile with a depth of 50 cm.

Figure 5-1 and 5-2 show the response curves of yield and nitrogen leaching to different

remaining ASW moistures. In Figure 5-1, it can be seen that if the irrigation event was triggered

at a threshold of 70% ASW remained (or a MAD of 30%) with a precipitation depth of 7.5 mm









(Table 5-4), the predicted dry yield arrived its maximum value as 3,867 kg ha-1. If an irrigation

event was triggered at a very low level of remaining ASW, e.g. 10%, then there would be long

interval between two irrigation events. Corn growth would suffer from water stress, and yield

would be reduced significantly to less than 1,000 kg ha-1, which is only about one third of the

acceptable dry matter yield mentioned in Section 5.2.4. Under a threshold of 20%, 30%, 40%, or

50% remaining ASW (or a MAD of 80%, 70%, 60%, or 50%), there appeared to be water stress

since the yields were all below 3,000 kg ha-1.

When the irrigation event was triggered at higher levels of remaining ASW, e.g. 80% and

90% (or a MAD of 20% and 10%), it meant more frequent irrigations with very small depth of

irrigation (5.0 and 2.5 mm respectively). The predicted dry material yield would decrease a little

bit, which might be caused by more nitrogen leaching and less nitrogen available for plant uptake.

Figure 5-2 shows an obvious trend of increasing of nitrogen leaching if irrigation events

were more frequent. For example, if the irrigation event of 22.5 mm was triggered when

remaining ASW wasl0%, the accumulative nitrogen leaching (NLCM) value was approximately

32 kg ha-l. But it could be as high as almost 120kg ha-l, if the event was triggered with a

threshold of 80% or 90% remaining ASW (or a MAD of 20% or 10%) and a precipitation depth

of 5 or 2.5 mm.

Considering the yield and nitrogen leaching, it was observed that a threshold of a MAD

value of 40 or 30% with a precipitation depth of 10 or 7.5 mm would be a good choice for

irrigation scheduling, since they result in a higher yield though they failed to obtain the lowest

nitrogen leaching level. This conclusion could be confirmed by the values of MAD provided by

James et al. (1982), where the MAD value of sweet corn was 50%.









Haman and Smajstrla (1997) mentioned that Florida's sandy soils are well known for their

inability to hold water. Very little water is stored in the root zone. A general rule for vegetable

irrigation is to provide irrigation before 50% of this water is used in order to avoid plant stress,

which means a MAD value of 50%. They also pointed out that if possible, 33% depletion should

be used for scheduling drip irrigation, which requires frequent (once or more per day) and short

water applications. Thus, a MAD value of 40% or 30% in this study should be reasonable.

In addition, if irrigating at a low MAD value such as 10%, it requires a low irrigation depth

(2.5 mm). These depths probably can not match the lower irrigation depth limit of the linear or

center pivot irrigation systems, which are usually used by farmers in North Florida for crop

irrigation. At the same time, the frequency of irrigation events will also be very high, probably

less than 24 hours. However, it might not be practical for a center pivot system, because it could

take more than 24 hours for a 400-meter system to finish an irrigation circle (Keller and Bliesner,

1990).

Next, the irrigation strategies designed as automatic irrigation on fixed days with specific

depth were investigated (Figure 5-3 and 5-4). When enough water was applied, e.g. 2 or 3

irrigations per week with a depth of 15, 20, or 25 mm, the yield didn't increase, and in some

cases decreased. There was no significant difference between the yields of 2 and 3 irrigations a

week when irrigation depths were 15, 20 and 25 mm, but yield could be significantly reduced

when there was just 1-irrigation a week due to water stress.

The assumption that when more water is applied more nitrogen is leached was confirmed

by Figure 5-4. For example, at 2 or 3 irrigation events a week with a depth of 25 mm, the

simulated nitrogen leaching could be about 150 kg ha-1, while the real nitrogen application was

only about 230 kg ha-1, indicating more than 65% of the applied nitrogen was leached.









Evaluating the amounts of yield and nitrogen leaching, it is apparent that two-irrigations a

week with a precipitation depth of 15 mm, or three-irrigations a week with a precipitation depth

of 10 mm would achieve a relatively higher level of yield and lower nitrogen leaching.

Compared with the irrigation schedules with a MAD value of 30% or 40%, these two fixed-date-

and-depth methods share the similar predicted dry matter yield (3,500 to 40,000 kg ha-1) and

amount of nitrogen leaching (80 to 100 kg ha-1).

Finally based on the criterion of higher yield and lower nitrogen leaching, six irrigation

strategies (Table 5-13) were selected as the optimal ones for future combination simulations.

Among these irrigation schedules, Irrigation 2, 3, 5 and 6 had close predicted dry matter yield

and amount of nitrogen leaching as mentioned above. While Irrigation 1 had a lower yield (about

2,500 kg ha-1) and lower nitrogen leaching (about 65 kg ha-1), and Irrigation 4 had a similar dry

matter yield (about 3,700 kg ha-1) and a higher amount of nitrogen leaching (about 120 kg ha-1).

5.3.2 Effects of Nitrogen Fertilizer

5.3.2.1 Total nitrogen fertilizer amount

Total amount of fertilizer nitrogen required for sweet corn production was a core issue in

developing research-based N BMPs. The recommended nitrogen amount by the Institute of Food

and Agricultural Sciences (IFAS), University of Florida (Hochmuth, 2000) was 224 kg N ha-1

(200 lb N ac-1). What are the implications if this recommendation is reduced? How will the yield

and accumulative nitrogen leaching change with different nitrogen fertilizer levels? It will be

necessary to explore the influence of total N amount first.

As mentioned in Section 5.2.3, when doing nitrogen simulations, the irrigation strategy

was the actual one of Block 1 in 2006, which was scheduled according to the daily ET value. The

accumulated irrigation amount and rainfall data in 2006 are shown in Figure 5-5.









Figure 5-6 and 5-7 show the response curves of yield and nitrogen leaching to different

nitrogen fertilizer levels. The irrigation strategy was the actual field experiment in Block 1 in

2006. Here the nitrogen fertilizer was applied with a split of 0:1/2:1/2 (nothing in the small leaf

stage, 1/2 of total nitrogen except for the starter N fertilizer in the larger leaf stage, and 1/2 in the

ear development stage) and 40 kg N ha-1 for "application amount", A starter nitrogen application

of 17 kg N ha-1 was set for all of the simulations.

The two curves in these two figures (Figure 5-6 and 5-7) confirmed the fact as more

nitrogen is applied, more will be leached. After 196 kg N ha-1 or 175 lb N acre-' (red star in

Figure 5-6 and 5-7), the yield increment caused by added N fertilizer decreased and finally

approached zero, where the predicted dry matter yield was near 3,400 kg ha-l. At the same time,

nitrogen leaching kept steadily increased when more nitrogen was applied. For example, the

amount of nitrogen leaching increased from 82 to 266 kg N ha-l when nitrogen application level

increased from 196 to 561 kg N ha-1. When the fertilizer level was zero, there was still some

yield and nitrogen leaching (Figure 5-6 and 5-7). This is because the model assumed the nitrogen

from dead organic matter was 7 kg N ha-l. The initial nitrate-N and ammonium-N concentration

was 0.1 g N Mg-1 and 0.5 g N Mg-1, which equal to about 1.3 and 6.7 kg N ha-l. And there was

alos a starter N application of 15 kg N ha-1 at planting. In other words, there was some nitrogen

available except for N fertilizer application.

It seemed 168 kg N ha-1 (green star in Figure 5-6 and 5-7) would be enough for sweet corn

growth and produce comparably less nitrogen leaching. However, the model simulation assumed

the fertilizer application efficiency as 1.0, which means no nitrogen was wasted. It is not true in

actual production. The efficiency must be less than 1.0. Thus, the actual amount of nitrogen

fertilizer should be greater than 168 kg N ha-1.









To explore other possible N fertilizer strategies, 6 nitrogen amounts, as 140, 168, 196, 224,

252 and 280 kg N ha-1 (or 125, 150, 175, 200, 225 and 250 lb N acre-'), were selected as optimal

total amounts for future combination simulations.

5.3.2.2 Nitrogen fertilizer split

The whole growth season of sweet corn could be divided into three stages, the small leaf

stage, large leaf stage, and ear development stage. Nitrogen can arbitrarily be applied during the

growth season, but the best way is to make nitrogen application coincide with the N requirement

of sweet corn.

As shown in Table 5-6, 30 splits identified as S1lto S30 were simulated. The top-10 splits

that have the highest yields or lowest nitrogen leaching amounts are summarized in Table 5-14.

The results in Table 5-14 show that nitrogen fertilizer split did not show a significance influence

on yield if there was application of N during the small leaf stage or large leaf stage, since the

predicted dry matter yields were all about 3,400 ka ha-l. However, splitting N applications

showed a significant influence on N leaching. The best splits were "0:1/4:3/4", "0:1/3:2/3",

because these fertigation schedules could best coincide the nitrogen need of sweet corn growth,

especially from tasseling to maturity.

The results showed that 3 splits S3, S4 and S5 (0-1/4-3/4, 0-1/3-2/3, 0-1/2-1/2) all rank in

top-10 either in yield or nitrogen leaching. High yield and low nitrogen leaching was the

objective of this research for BMP. Therefore, these 3 splits were selected as optimal ones for

combination simulations.

5.3.2.3 Amount of nitrogen fertilizer in each application

The amount of each N fertilizer application determines how many times the total N

fertilizer would be applied into the field. As mentioned in Section 5.2.3, 20 different kinds of









"application amount" ranging from 5 to 100 kg N ha- with a step 5 kg N ha-1 were simulated in

this study. These "application amount" were identified as Al to A20.

In Figure 5-8, the response curve of yield to "application amount" shows that if the

application of the total nitrogen fertilizer was less then 70 kg N ha-1 in each event (red stars in

Figure 5-8 and 5-9), the yield would stay almost the same. In Figure 5-9, the response curve of

nitrogen leaching to splits was not a straight line. Lowest nitrogen leaching would be obtained

when applying just 5 or 10 kg N ha-1 in each fertigation event. The main trend showed that the

accumulative N leaching would increase with the increasing of "application amount" (Figure 5-

9).

However, too little N fertilizer applied in application amount could result in too many

fertigation events, which would increase production cost. It seems 30, 40 or 50 kg N ha-1 could

be the best "application amount", if the production cost was considered in addition to yield and

nitrogen leaching. Finally, these three kinds of "application amount" were selected as optimal

ones for future combination simulations.

After the single factor analysis above, the main factors for best N fertilizer management

strategies were selected (Table 5-15). The first factor was the total N amount. Six different N

fertilizer levels ranged from 140 to 280 kg N ha-1 were selected. The second factor was the

nitrogen application split strategy. Three kinds of split were selected since they were believed to

coincide with the nitrogen need of sweet corn growth. And three kinds of N application amount

were selected both considering nitrogen leaching and labor cost.

Thus for a complete factorial experiment design to test all of the possible combinations,

these three selected factors could result in 6 x 3 x 3 = 54 kinds of fertilizer application strategies









among which the optimal one of N fertilizer was assumed to exist. These N fertilizer strategies

would also be combined with the selected irrigation strategies to look for the potential BMPs.

5.3.3 Selection of Potential BMPs

As mentioned in previous sections, there is a total of 6 irrigation strategies (Table 5-13)

and 54 N fertilizer strategies (Table 5-15). So there will be 6 x 54 = 324 possible management

scenarios. AS mentioned in Section 5.2.3, all of these scenarios were simulated under the 33-year

continuous historical weather conditions (1958-1990) of Gainesville.

All of the combination treatments that had a simulated yield (HWAH, kg ha-) above the

acceptable yield, 3,400 kg ha-1 (Section 5.2.4) were selected. Then these selected treatments were

ranked according to their nitrogen leaching values (NLCM, kg ha-1). Table 5-16 specifies the top

20 of these treatments that had the lowest NLCM values.

For irrigation strategies, it seemed that 5.0 mm irrigation triggered by a MAD of 20% and

7.5 mm irrigation at a MAD of 30% would be the best irrigation strategies. Actually this also

confirmed the assumption that frequent irrigation with small amount of water could reduce

nitrogen leaching due to the less water loss by deep percolation. In practice, this requires the

linear or center pivot irrigation systems continuously running round by round with a high speed

and low irrigation depth. Actually, this is what the farmers usually do in sweet corn production

especially from tasseling to maturity.

For N fertilizer, the ranking shows that the optimal total amount of nitrogen application

was 196 kg N ha-1 or 224 kg N ha-1. The splits were dominantly the ones of"0:1/4:3/4" and

"0:1/3:2/3". Only one combination with a split of"0:1/2:1/2" fell in the top-20 combinations

selected. It seemed 30, 40 or 50 kg N ha-1 all could be the best "application amount". There were

10 combinations that had an "application amount" of 30 kg N ha-1 in the top-20 combinations.









This confirmed another assumption that frequent application of N fertilizer with small amount

could reduce nitrogen leaching.

In general from the combination simulation results, it could be concluded that if growers

can apply both irrigation water and N fertilizer in more frequent applications but with smaller

amounts in each event, it will result in an acceptable yield and a lower level of nitrogen leaching.

But this requires farmers to run their irrigation and fertigation system more often, which could

increase the production cost due to the increase of labor, electricity, etc.

Finally, considering the yield, nitrogen leaching and operation cost, the following six

combination treatments (Table 5-17) were selected as potential BMPs for future study. For

example, for potential BMP1 it means an irrigation of 5.0 mm should be started when the MAD

value was 20%. Totally 196 kg N ha-1 should be applied. Except for a starter application of 15 kg

N ha-1, nothing should be applied during the small leaf stage, 1/4 of the total N during the large

leaf stage, and 3/4 during the fruit development stage. In each N application, only 30 kg N ha-1

should be applied.

These selected potential BMPs will be used to conduct uncertainty analysis both under

input variable and weather variations in following research in Chapter 6.

5.3.4 Evaluation and Implementation of Potential BMPs

By definition, BMPs are practices or combination of practices determined by the

coordinating agencies, based on research, field-testing, and expert review, to be the most

effective and practicable on-location means, including economic and technological

considerations, for improving water quality in agricultural and urban discharges. BMPs are

typically implemented as a "BMP treatment train" that includes a combination of nonstructural

and structural practices that have been determined to be effective for reducing or preventing

pollution. BMPs must be: technically feasible, economically viable, and socially acceptable.









For the six developed potential BMPs listed in Table 5-17, they had a total nitrogen

application amount of 196 or 224 kg N ha-1 (175 or 200 lb. N ac.-1), which was lower or equal to

the recommended nitrogen level of 200 lb. N ac.-1 by IFAS, the Univerisity of Florida

(Hochmuth, 2000). The crop model simulation results confirmed that the recommendation of

IFAS is correct. In actual production, farmers always apply more nitrogen than the recommended

value to guarantee their yields. For example, in the EPA319 demonstration project (Hochmuth,

2003) in Suwannee Farms, O'Brien, FL, total nitrogen application was 302 kg N ha-1, which was

almost 50% higher than the recommended N application amount. Thus, it can be concluded that

if the developed potential BMPs could be implemented in reality, more nitrogen fertilizer would

be saved and consequently less nitrogen would be leached.

However, it should be noticed that even if these potential BMPs were adopted, it does not

mean the groundwater quality will be definitely protected. For example, the nitrate-N standard

for drinking water is 10 mg N L-1 (U.S. Dept. Health, Education, and Welfare, 1962). The total

rainfall and irrigation water depth was 64.0 mm and 270.5 mm respectively, in the field

experiment of Block 1 in 2006. If it is designed to make the solution concentration of nitrogen

leaching equal to or less than 10 mg N L1, only about 33.5 kg N ha-1 was allowed for leaching.

From Table 5-16, it can be seen that no potential treatment could approach this standard because

the simulated average nitrogen leaching in season (NLCM) were all greater than 34 kg N ha-1.

And considering the inorganic N left in soil profile (NIAM), which was subject to leaching in a

long term, the potential nitrogen leaching would be even higher. In this case, these potential

BMPs probably would not be real BMPs any more, since they failed to definitely prevent

nitrogen pollution.









Thus, it can be concluded that the nitrogen pollution is almost inevitable when planting

sweet corn on the sandy soils in North Florida. However, the developed potential BMPs could

significantly reduce the amount of actual nitrogen fertilizer application and consequently

improve the pollution.

Ultimately, the developed potential BMPs need to be incorporated into actual farm

production. To fully integrate these developed potential BMPs into a meaningful farm

production plan requires an on-farm assessment and a quality assurance program. For on-farm

assessment, all growers should perform an environmental assessment of their crop production

operations, which will aid in identifying which BMPs should be considered to achieve the

greatest economic and environmental benefit. Having a viable quality assurance program is very

important to ensure that BMP implementation is occurring on track. The quality assurance

program also serves to build overall program credibility and further provides assurance that

BMPs are constructed or installed as designed (Florida Department of Agriculture and Consumer

Services, 2005).

Thus, the simulation of potential BMPs with the CERES-Maize model is just a first step in

BMP development for sweet corn production. If want to make these "potential BMPs" to be "real

BMPs", more works such as education, assessment, and quality assurance have to be done in the

future.

5.4 Summary and Conclusions

In this study, the CERES-Maize module of the DSSAT model was utilized as a platform to

develop BMPs for sweet corn production in North Florida. The model was calibrated with the

GLUE method. The expectation values of the posterior distributions of the input parameters were

used as the nominal parameter set to conduct the simulations. Each treatment was simulated

under 33 years' historical weather data.









A total of 24 irrigation treatments, 21 nitrogen fertilizer levels, 30 nitrogen splits, and 20

kinds of application amount were simulated. Finally six potential BMPs were selected according

to their dry matter yields and amounts of nitrogen leaching. Some conclusions were drawn as

follows.

Irrigation frequency and amount had strong influence on corn yield. For example, if the

irrigation event was triggered by lower remaining available soil moisture (such as 10% and 20%),

which means a longer interval between two irrigation events, the yield would be significantly

reduced to less than 1,000 kg ha-1 due to water stress, which is only about one third of the

acceptable dry matter yield.

The trend of increasing nitrogen leaching was obvious if irrigation events were more

frequent and more water was applied in each event. For example, at 2 or 3 irrigation events a

week with a depth of 25 mm, the simulated nitrogen leaching could be about 150 kg ha-1, while

the real nitrogen application was only about 230 kg ha-l, indicating more than 65% of the applied

nitrogen was leached.

More nitrogen applied resulted in more being leached. After 168 kg N ha-1, the yield

increment caused by increasing of N fertilizer approached zero, where the predicted dry matter

yield was near 3,400 kg ha-l. At the same time, nitrogen leaching kept steadily increased when

more nitrogen was applied. For example, the amount of nitrogen leaching increased from 82 to

266 kg N ha-1 when nitrogen application level increased from 196 to 561 kg N ha-1.

Nitrogen fertilizer split did not show a significant influence on yield if there was

application of N during the small leaf stage or large leaf stage. However, splitting N applications

showed a significant influence on N leaching. Except for a starter N application of 15 kg N ha- ,









the best splits were 0:1/4:3/4 and 0:1/3:2/3, because these fertigation schedules could best

coincide the nitrogen need of sweet corn growth, especially from tasseling to maturity.

A small "application amount" could not increase yield very much if it was less than 70 kg

N ha-1, but it could decrease N leaching. However, too little N fertilizer applied in application

amount could result in too many fertigation events, which would increase production cost. It

seems 30, 40 or 50 kg N ha-1 could be the best "application amount", if the production cost was

considered in addition to yield and nitrogen leaching.

If grower could apply both irrigation water and nitrogen fertilizer more frequently but with

smaller amounts in each application, this would result in an acceptable yield and a lower level of

nitrogen leaching.

The nitrogen pollution is almost inevitable when planting sweet corn on the sandy soils in

North Florida. However, the potential BMPs could significantly reduce the amount of actual

nitrogen fertilizer application and consequently improve the pollution. The simulation of

potential BMPs with the CERES-Maize model is just a step in BMP development for sweet corn

production. If want to make these "potential BMPs" to be "real BMPs", more works such as

education, assessment, and quality assurance have to be done.

The CERES-Maize model verified itself as a powerful tool to develop practical strategies

for agricultural production. It provided a convenient and economical way to obtain useful

information on the interactions between crop, soil, weather and field management strategies.

However, the current model still has some disadvantages. First, the current CERES-maize

model can only predict the yield quantity, but not the quality. According to the classification

standard ofUSDA, sweet corn can be classified into (USDA, 1962) three levels, US #1, US #2

and Cull. The sum of US #1 and US #2 is the marketable yield. The current model could not do









such classification. For example, in the prediction results, even when there was completely no

nitrogen application, there was still some level of yield, near 500 kg ha-1. However, from field

production experience in this region, when there is no nitrogen, the quality will be all culls with

no marketable yield (see yield quality results in Chapter 4 for reference).

Second, no model is perfect. Uncertainty always exists in the prediction results. If the

uncertainty is too great, the simulation results will be misleading. For example, the model was

only run with a set of nominal values of input parameters. What will be the response of the

selected BMPs under the uncertainties of input parameters and weather uncertainty? To answer

this question, the selected BMPs must be investigated for their output uncertainties caused by

weather and input parameter uncertainties. This will be the main topic of Chapter 6.











4500
4000
3500
3000
2500
2000
1500
1000
500
n


0 10 20 30 40 50 60 70
Remaining available soil water (%)

Figure 5-1. Response curves of yield to different remaining ASW


80 90 100


Too Dry
0 1 I 1 1 1 1 1 1 1I1
0 10 20 30 40 50 60 70 80 90
Remaining available soil water (%)

Figure 5-2. Response curves of nitrogen leaching to different remaining ASW


Too Wet


Too Dry


Too Wet











4500
4000

3500

- 3000
S2500
S2000

S1500
1000

500
0


5 10 15 20 25
Irrigation Depth (mm)


-*- 1 irrigation/week 2 irrigations/week -a- 3 irrigations/week

Figure 5-3. Response curves of yield to different irrigation depths


200
180
' 160
1 140
0 120
0 100
S80
2 60
z 40
20
0


5 10 15 20 25
Irrigation Depth (mm)


-- 1 irrigation/week 2 irrigations/week 3 irrigations/week

Figure 5-4. Response curves of nitrogen leaching to different irrigation depths















250
E
E
c 200
o
0

S150


S100
n'0
0;


0 I 1 i 1 1 i i _** i
014



Date

Rainfall -- Accumulated Irrigation

Figure 5-5. Rainfall and accumulated irrigations in East Half of Blockl in 2006


4000

3500


3000

2500

2000

1500

1000


100 200 300 400 500


N fertilizer levels (kg/ha)

Figure 5-6. Response curves of yield to different N fertilizer levels. Green dot indicates the
nitrogen fertilizer level of 168 kg N ha-1, while red dot indicates the nitrogen fertilizer
level of 196 kg N ha.














250

S200


S150

g 100

50

0
0 100 200 300 400 500 600
N fertilizer level (kg/ha)

Figure 5-7. Response curves of nitrogen leaching to different N fertilizer levels. Green dot
indicates the nitrogen fertilizer level of 168 kg N ha-1, while red dot indicates the
nitrogen fertilizer level of 196 kg N ha-1.


4000

3500

; 3000

S2500
CS
S2000

S1500

S1000

500

0


20 40 60 80 100


Application Amount (kg/ha)


Figure 5-8. Dry yield vs. different N fertilizer application amount. Red dot indicates the
"application amount" of 70 kg N ha-1 in each event.














120

100

S80

60

40

20

0
0 20 40 60 80 100 120
Application Amount (kg/ha)

Figure 5-9. Nitrogen leaching vs. different N fertilizer application amount. Red dot indicates the
"application amount" of 70 kg N ha-1 in each event.










Table 5-1. Expectation values of second posterior distribution of selected parametersa
P1 P5 PHINT SLDR SLRO SDUL SLLL SSAT SLPF
Parameter
Paramete d Cd Cd cm/cm cm/cm3 cm3/cm3
Expectation 99.17 577.20 39.68 0.73 78.14 0.10 0.06 0.30 0.87
a Cd means degree day.




Table 5-2. Soil properties of the experiment site
Bulk
Depth Teture Clay Silt Sand Dens PWP FC Saturation
Texture Density
(cm) (%) (%) (%) (g/cm3) (cm3/cm3) (cm3/cm3) (cm3/cm3)
0-15 Sandy soil 2.75 1.92 95.33 1.67 0.051 0.110 0.313
15-30 Sandy soil 2.56 2.35 95.08 1.69 0.061 0.117 0.317
30-60 Sandy soil 2.36 1.76 95.88 1.67 0.077 0.118 0.357


Table 5-3. Calculation of total available soil water (ASW) in the soil profile
Layer (cm) FC PWP Soil Depth (mm) ASW (mm)
0-15 0.110 0.051 150.0 8.7
15-30 0.117 0.061 150.0 8.4
30-50 0.118 0.077 200.0 8.3
Sum -- 25.4



Table 5-4. Irrigation treatments based on different MAD values
Treatment MAD ASW (mm) Irrigation Depth (mm)
I1 90% 25.0 22.5
12 80% 25.0 20.0
13 70% 25.0 17.5
14 60% 25.0 15.0
15 50% 25.0 12.5
16 40% 25.0 10.0
17 30% 25.0 7.5
18 20% 25.0 5.0
19 10% 25.0 2.5










Table 5-5.
Treatment
110
Il1
112
113
114
115
116
117
118
119
120
121
122
123
124


Nitrogen splits used in BMP simulation
Number of Irrgiation I
per week
1
2 Mond
3 Monday, V
1
2 Mond
3 Monday, V
1
2 Mond
3 Monday, V
1
2 Mond
3 Monday, V
1
2 Mond
3 Monday, V


rigation Day
Wednesday
lay and Thursday
Wednesday, and Friday
Wednesday
lay and Thursday
Wednesday, and Friday
Wednesday
lay and Thursday
Wednesday, and Friday
Wednesday
lay and Thursday
Wednesday, and Friday
Wednesday
lay and Thursday
Wednesday, and Friday


Table 5-6. Nitrogen splits used in single factor simulation
No. Split Description No. Split Description No. Split Description
1 S1 0-0-1 a 11 Sll 1/5-1/5-3/5 21 S21 1/3-2/3-0
2 S2 0-1/5-4/5 12 S12 1/5-2/5-2/5 22 S22 1/2-0-1/2
3 S3 0-1/4-3/4 13 S13 1/5-3/5-1/5 23 S23 1/2-1/2-0
4 S4 0-1/3-2/3 14 S14 1/5-4/5-0 24 S24 2/3-0-1/3
5 S5 0-1/2-1/2 15 S15 1/4-0-3/4 25 S25 2/3-1/3-0
6 S6 0-2/3-1/3 16 S16 1/4-1/4-2/4 26 S26 3/4-0-1/4
7 S7 0-3/4-1/4 17 S17 1/4-2/4-1/4 27 S27 3/4-1/4-0
8 S8 0-4/5-1/5 18 S18 1/4-3/4-0 28 S28 4/5-0-1/5
9 S9 0-1-0 19 S19 1/3-0-2/3 29 S29 4/5-1/5-0
10 S10 1/5-0-4/5 20 S20 1/3-1/3-1/3 30 S30 1-0-0
a "0-0-1" means nothing was applied in the small leaf stage, nothing in the large leaf stage, and all nitrogen in
the fruit development stage except for a starter N application of 15 kg N ha1.


Irrigation Depth
(mm)
5
5
5
10
10
10
15
15
15
20
20
20
25
25
25










Table 5-7. Acreage, yield, production, and value of Florida sweet corn 1998-2006 (USDANASS,
2007)


Year
1998
1999
2000
2001
2002
2003
2004
2005
2006
Average


Acreage (acres)
Planted Harvested
41600 40300
39200 37800
40900 37400
40200 37900
41600 40800
39400 38800
38900 38700
35100 33600
33000 26300
38878 36844


Yield
(lb acre 1)
14500
14000
15000
14000
14000
14500
15500
16000
18500
15111


Production
(million lb.)
584.4
529.2
561
530.6
571.2
562.6
599.9
537.6
486.6
551.456


Table 5-8. Fresh yields of selected white sweet corn varieties in Clanton Ala. 1995-1996
(Simonne et al. 1999)
No. Year Place Cultivar Yield (kg ha1)
1 1995-19967 Clanton, Ala. Even Sweeter 14,726
2 1995-19967 Clanton, Ala. Treasure 14,264
3 1995-19967 Clanton, Ala. Snow White 12,400
4 1995-19967 Clanton, Ala. Snow Belle 11,432
5 1995-19967 Clanton, Ala. Fantasia 11,342
6 1995-19967 Clanton, Ala. Starshine 10,495
7 1995-19967 Clanton, Ala. Silver Queen 9,180
8 1995-19967 Clanton, Ala. FMX 413 8,925
9 1995-19967 Clanton, Ala. Silverado 8,675
10 1995-19967 Clanton, Ala. Rising Star 8,291


Table 5-9. Fresh yields of sweet corn experiment in
1999)


Springfield Tenn. 1993-1995 (Mullins et al.,


N N Yield Yield
Parameter
(lb acre -) (kg ha -) (tons acre- ) (kg ha1)
N rate (lb/acre) 0 0 2.50 5,600
50 56 3.10 6,944
100 112 3.6 8,064
150 168 3.7 8,288
Cultivar
Silver Queen 100 112 3.2 7,168
Incredible 100 112 4 8,960
Chanllenger 100 112 2.5 5,600
Year
1993 100 112 2.9 6,496
1994 100 112 4.1 9,184
1995 100 112 2.7 6,048










Table 5-10. Fresh yields of sweet corn experiment in Eden Valley and Freeville, NY, 1998-2001
(Rangarajan et al., 2002)
1998

PlIn-row spacing N N Yield Yield
Place (inches) (lb acre-1) (kg ha-1) (tons acre-1) (kg ha 1)

Eden Vallev.NY 7 120 134.4 8.30 18.592


120
120
1999


In-row spacing
(inches)


134.4
134.4


8.00
7.00


N N Yield
(lb acre-1) (kg ha-1) (tons acre 1)


Freeville, NY


8.30
7.90
7.10


Cultivar

Sweet Symphony
Temptation


In-row spacing
(inches)


N N Yield Yield
(lb acre-1) (kg ha-1) (tons acre-1) (kg ha 1)


100
100
2000


7.50
7.00


16,800
15.680


N N Yield Yield
(lb acre-1) (kg ha-1) (tons acre-1) (kg ha 1)


Freeville, NY


134.4
134.4


Cultivar


Temptation
Sweet Symphony
Seneca Spring


In-row spacing
(inches)


6.50
6.50


14,560
14,560


N N Yield Yield
(lb acre-1) (kg ha ) (tons acre ') (kg ha 1)


120
120
121
2001


134.4
134.4
135.52


7.10
6.10
6.30


15,904
13,664
14,112


N N Yield Yield
(lb acre ') (kg ha1) (tons acre ') (kg ha 1)


Freeville, NY


Cultivar

Temptation
Sweet Symphony
Seneca Spring


N
(lb/acre)

120
120
121


N Yield Yield
(kg/ha) (tons/acre) (kg/ha)


134.4
134.4
135.52


6.70
6.90
5.70


15,008
15,456
12,768


Place


17,920
15,680


Yield
(kg ha 1)

18,592
17,696
15,904


Place


Place


134.4
134.4


6.50
6.30


14,560
14,112


I










Table 5-11. Fresh yields of sweet corn experiment in Belle Glade, Florida, in spring of 2001
(Shuler, 2002)
Spacing Yield Yield
Variety
(inches) (42 lb crt-1) (kg ha 1)
A&C '945'
9 364 17,123
8 391 18,393
7 304 14,300
6 424 19,945
Rogers '9686'
9 468 22,015
8 441 20,745
7 497 23,379
6 478 22,485
A&C
'Summer Sweet 8102 BC'
9 384 18,063
8 422 19,851
7 466 21,921
6 527 24,790
Average
9 405 19,051
8 418 19,663
7 449 21,121
6 476 22,391
'945' -391 18,393
'9686' 471 22,156
'8102' 449 21,121



Table 5-12. Summary of sweet corn yield in field experiments conducted in Florida (Hochmuth
and Cordasco, 2000)
N Rate Yield
No. Year Location N Rate Yield ) Source
(kg ha ) (kg ha )
1 1961, 1962 Gainesville, FL 125 11,252 Volk, 1962
2 1961, 1962 Gainesville, FL 168 14,360 Robertson, 1962
3 1976, 1977 Gainesville, FL 224 14,595 Rudert and Locascio, 1979
4 1991 Live Oak, FL 168 15,537 Hochmuth et al., 1992
5 1993 Gainesville, FL 168 17,279 Hochmuth, 1994
6 1996 Sanford, FL 252 15,349 White et al., 1996
7 1997 Gainesville, FL 168 21,375 Hochmuth, 1997a
8 1997 Gainesville, FL 168 15,019 Hochmuth, 1997b










Table 5-13. Selected irrigation strategic
No. Title
1 Irrigation 1
2 Irrigation 2
3 Irrigation 3
4 Irrigation 4
5 Irrigation 5
6 Irrigation 6


Description
12.5 mm with a MAD of 50%
10.0 mm with a MAD of 40%
7.5 mm with a MAD of 30%
5.0 mm with a MAD of 20%
2 irrigation per week with a depth of 15 mm
3 irrigation per week with a depth of 10 mm


Table 5-14. Ranking of dry yield (HWAH) and nitrogen leaching (NLCM) under different N
fertilizer application splits
HWAH Ranking NLCM Ranking
HWAH NLCM
Split Description (g h 1 Split Description 1(kg h
(kg ha ) (kg ha )
S12 1/5-2/5-2/5 3499 S4 0-1/3-2/3 80
S19 1/3-0-2/3 3494 S1 0-0-1 82
S4 0-1/3-2/3 3453 S2 0-1/5-4/5 84
S5 0-1/2-1/2 3447 S10 1/5-0-4/5 84
S20 1/3-1/3-1/3 3432 S3 0-1/4-3/4 87
S6 0-2/3-1/3 3429 S15 1/4-0-3/4 87
S3 0-1/4-3/4 3414 S5 0-1/2-1/2 88
S16 1/4-1/4-2/4 3406 S19 1/3-0-2/3 91
S11 1/5-1/5-3/5 3403 S11 1/5-1/5-3/5 94
S17 1/4-2/4-1/4 3376 S12 1/5-2/5-2/5 101



Table 5-15. Selected factors of N fertilizer application strategies
Total Amount Split Application amount
Title Description Title Description Title Description
N Fertilizer 1 140 kg N ha1 Split 1 0:1/4:3/4 Application amount 1 30 kg N ha1
N Fertilizer 2 168 kg N ha1 Split 2 0:1/3:2/3 Application amount 2 40 kg N ha1
N Fertilizer 3 196 kg N ha- Split 3 0:1/2:1/2 Application amount 3 50 kg N ha1
N Fertilizer 4 212 kg N ha
N Fertilizer 5 252 kg N ha-
N Fertilizer 6 280 kg N ha-1










Table 5-16. Ranking of average nitrogen leaching (NLCM) of combination management over 33 years (1958-1990)
Total Application HWAH HWAH NLCM NLCM
No. Irrigation Nitrogen Nitrogen Amount Mean STDEV Mean STDEV Percent of
1 mT1 1 Split 1 1T1Leachmng
kg N ha Split kg N ha1 kg N ha1 kg N ha1 kg N ha kg N ha Leaching
1 5.0 mm-MAD 20% 196 0-1/4-3/4 30 3495 887 35 24 18%
2 7.5 mm-MAD 30% 196 0-1/4-3/4 30 3515 667 37 26 19%
3 5.0 mm-MAD 20% 196 0-1/4-3/4 40 3549 875 37 25 19%
4 5.0 mm-MAD 20% 196 0-1/3-2/3 30 3511 893 37 25 19%
5 5.0 mm-MAD 20% 196 0-1/4-3/4 50 3548 869 38 26 19%
6 5.0 mm-MAD 20% 196 0-1/3-2/3 40 3544 900 38 26 19%
7 5.0 mm-MAD 20% 196 0-1/4-3/4 40 3532 725 38 27 19%
8 5.0 mm-MAD 20% 224 0-1/4-3/4 30 3511 890 39 28 17%
9 7.5 mm-MAD 30% 196 0-1/3-2/3 30 3513 663 39 26 20%
10 7.5 mm-MAD 30% 196 0-1/4-3/4 50 3550 752 40 27 20%
11 7.5 mm-MAD 30% 196 0-1/3-2/3 40 3516 699 40 27 20%
12 7.5 mm-MAD 30% 224 0-1/4-3/4 30 3516 666 40 29 18%
13 10.0 mm-MAD 60% 196 0-1/4-3/4 30 3506 769 40 26 20%
14 5.0 mm-MAD20% 196 0-1/3-2/3 50 3556 888 41 27 21%
15 5.0 mm-MAD20% 224 0-1/4-3/4 40 3555 902 41 30 18%
16 5.0 mm-MAD20% 224 0-1/3-2/3 30 3524 899 41 29 18%
17 10.0 mm-MAD 60% 196 0-1/4-3/4 40 3589 829 42 27 21%
18 5.0 mm-MAD20% 196 0-1/2-1/2 30 3540 898 42 27 21%
19 10.0 mm-MAD 60% 196 0-1/3-2/3 30 3499 765 42 27 21%
20 7.5 mm-MAD 30% 196 0-1/3-2/3 50 3498 708 42 28 21%


206










Table 5-17. Selected potential BMPs for sweet corn production
Nitrogen Level Nitrogen Split Application amount
No. Irrigation
kg N ha-1 kg N ha-1
1 5.0 mm-MAD 20% 196 0-1/4-3/4 30
2 5.0 mm-MAD 20% 196 0-1/3-2/3 30
3 7.5 mm-MAD 30% 196 0-1/4-3/4 40
4 7.5 mm-MAD 30% 196 0-1/3-2/3 30
5 5.0 mm-MAD 20% 224 0-1/4-3/4 30
6 7.5 mm-MAD 30% 224 0-1/4-3/4 30









CHAPTER 6
UNCERTAINTY ANALYSIS OF POTENTIAL SWEET CORN BMPS UNDER WEATHER
AND INPUT PARAMETER VARIABILITY

6.1 Introduction

Techniques of system analysis and crop growth modeling are increasingly being used in

agriculture for estimating production potential, agro-technology transfer, designing plant types,

strategic and tactical decisions, and setting research priorities (Teng & Penning de Vries, 1992;

Penning de Vries & Teng, 1993). Dynamic process based models simulate daily increase in crop

growth through a number of processes such as photosynthesis, dry matter partitioning, crop

development and transpiration as affected by soil and weather factors and crop management

(Aggarwal, 1995).

However, uncertainties in model outputs always exist. Models can at best only represent a

current understanding of biological systems; yet they do highlight where information and

understanding are lacking (Boote et al., 1996). This is because models are all simplified

representations of reality. Even if the structures of the model equations are perfect, there will be

errors in model output due to inaccuracies both in the initial conditions and in the values of

model parameters and forcing functions (Pei and Wang, 2003). In other words, uncertainty in

model prediction is unavoidable. Clarification of the inherent uncertainties and quantification of

uncertainties in modeling results is thus critical for improving model prediction methods and

identifying effective management strategies (Van Straten and Keesman, 1991). Linkov and

Burmistrov (2005) defined uncertainty of models as following three broad categories: (1)

Parameter uncertainty-uncertainty in the value of input parameters in a model; (2) Model

uncertainty-uncertainty about a model structure (i.e., the relevance of simplifying assumptions

and mathematical equations); (3) Scenario uncertainty-uncertainty regarding missing or

incomplete information to fully define the system under study.









Much of the uncertainty in model outputs can be ascribed to incomplete information on

input values relating to crop, soil and weather factors, and agronomic management date required

to run the model (Burrough, 1989; Richter & Sondgerdth, 1990). Crop parameter values could be

significantly uncertain due to imperfect knowledge of those caused by random errors related to

size and number of observations and systematic errors related to bias in the experimental,

measurement, observation and calibration procedures. In addition, crop input parameters may

exhibit spatial and temporal variability. Recognizing model parameter variability, biologists

generally report model simulation results with standard deviations or standard errors that

describe variation associated with the measured variable (Aggarwal, 1995).

Soil parameters required by the crop models have also shown spatial and temporal

variation and might have considerable measurement errors. These inputs are often estimated, e.g.

using Geographical Information Systems (Richardson, 1984; Nix, 1987). The stochastic nature of

many soil parameters are expected to result in uncertainty of the outputs of crop models

(Aggarwal, 1995).

The weather during the growing season affects growth and development through

accumulative dynamic growth, and the final value of crop characteristics of interest, e.g. grain

yield (Lawless and Semenov, 2005). Weather has been shown to have a strong influence on the

most suitable crop type and to a certain extent the most suitable cultivar at a given site (Jagtap et

al., 2002).

Heinmann et al. (2002) showed that the accuracy of rainfall observations was critical for

the simulation of crop yield and that the variability of simulated estimates was directly correlated

to the accuracy of model inputs. Xie et al. (2003) evaluated the importance of input variables on

the yield estimates made for maize and sorghum by the ALMANAC model. They concluded that,


209









in a dry land environment, rainfall and then solar radiation were the most important of the

meteorological variables for non-irrigated crops.

Solar radiation is a key variable since it is used, amongst other things, as part of the

estimation of evapotranspiration (ET) and biomass accumulation. Bellocchi et al. (2003) tested

the impacts of three air-temperature-based methods for estimating solar radiation data on the

estimates of reference crop ET and subsequent determination of above ground biomass at 20

locations worldwide. The results showed that each source had different levels of performance, in

terms of yield estimates, with each geographical location and season patterns.

In previous research, the CERES-Maize model of the DSSAT model (Jones et al., 2003)

was used to develop nitrogen (N) best management practices (BMPs) for sweet corn (Zea mays

L.) production on the sandy soil in North Florida (see Chapter 5 details). Six management

combinations of different irrigation and N fertilizer application strategies were selected as

potential BMPs. The CERES-Maize model requires soil parameters, genotype parameters, and

four kinds of climatic parameters as follows: daily rainfall, minimum daily temperature,

maximum daily temperature, and daily solar radiation. As discussed above, these potential BMPs

inevitably suffered from the uncertainties caused by these input parameters, since all of them

were selected with a set of nominal input parameters. Will these selected potential BMPs work

under other possible weather conditions, especially some extreme climate situations? What will

be the variance of the relevant model outputs? To answer these questions, an uncertainty analysis

of model outputs should be done.

The main objective of this research was to quantify the total uncertainties in simulated dry

matter yields (HWAH, kg ha-1) and accumulative nitrogen leaching (NLCM, kg ha-1) of the six


210









selected potential BMPs when simulating them with the CERES-Maize model. Two sources of

uncertainties were concerned: the input parameters (genotype and soil) and the weather data.

In addition, a similar simulation was also conducted for one real field management strategy

with the CERES-Maize model so as to compare the difference between the potential BMPs and

the real management practice.

6.2 Materials and Methods

6.2.1 Field Experiment and Weather Data

Except for the information about irrigation and nitrogen fertilizer application provided by

the selected potential BMPs (see Chapter 5 for details), some additional management

information was required by the CERES-Maize model as fundamental inputs to conduct

simulations. These inputs include planting date, planting population density, planting depth,

micro nutrient application, and harvest date etc.

In this study, this kind of fundamental information was obtained from the field experiment

in Block 1 in the spring of 2006 at the Plant Science Research and Education Unit, the University

of Florida. The unit is located in Pine Acres (29.4094N, 82.1777W, 20.746 meters above sea

level), Marion County, Florida, U.S. (Judge et al., 2005). See Chapter 3 and 5 for details about

the field experiment in Block 1.

The climate in Florida is subtropical and is characterized by long, warm summers and mild

winters. However, there are large variations between locations and from year to year. For

example, average annual rainfall (1971-2000) is 1,448 mm at DeLand, 1295 mm at Sanford, and

1270 mm at Ocala. Most of the rainfall occurs in June-September, with some months having as

much as 510 mm of rainfall. About 70-75 percent of the rainfall commonly returns to the

atmosphere as evapotranspiration (Sumner, 1996; Knowles, 1996).









A complete weather data set, which included precipitation, maximum temperature,

minimum temperature, and solar radiation for model simulations, was not available for Citra, FL,

where the experiment site is located. In the absence of site-specific data for his specific

experiment field, a model user has several choices of weather data source: (1) historical weather

data from nearby alternative weather stations; (2) artificial data from stochastic weather

generators e.g. LARS-WG (Barrow and Semenov, 1995), ClimGen (Stockle et al., 2001).

In this study, 33 years (1958-1990) of historical weather data at Gainesville, FL, the USA,

which is about 32 km from the experiment site of this research, were chosen as the nearest

complete weather data. These measured weather data were provided by the McNair Bostick

Simulation Lab of the Department of Agricultural and Biological Engineering, the University of

Florida.

6.2.2 Uncertainty of Input Parameters

In previous research, the generalized likelihood uncertainty estimation (GLUE) method

was used to calibrate the CERES-Maize model of the DSSAT model (see Chapter 3 for details).

In contrast to the commonly used methods of model calibration, the GLUE method gives a

distribution of the input parameters rather than a unique parameter set that can optimize all of the

observations and relevant predictions. Thus, the calibrated model still contains uncertainty

caused by input parameters, but the degree of uncertainty is significantly reduced by the GLUE

simulations.

The reduced uncertainty of the input parameters of soil and genotype can be presented by

their variance and mean value as shown in Table 6-1. The distribution of the selected parameter

is a multivariate normal distribution, except for parameter SLPF, which was assigned a uniform

distribution.









6.2.3 Selected Potential BMPs

Six potential BMPs were selected in previous research (Table 6-2, See Chapter 5 for

details). The details of each BMP were explained as follows. For example, in BMP1 irrigation

"5.0 mm-MAD 20%" meant that a 5.0 mm-irrigation was triggered by a value of maximum

amount of depletion (MAD) of 20% in the top 50 cm soil profile, i.e. 80% of the available soil

water (ASW) was remaining in that spoil profile. The ASW was defined as the water between

water holding capacity and permanent wilting point of the soil.

Nitrogen amount is the total fertilizer N applied in each growth season of sweet corn.

Nitrogen split determines how much of the total N fertilizer should be applied during the small

leaf stage, large leaf stage, and ear development stage, respectively. Application amount is how

much N fertilizer should be applied into the field in each fertilization event.

6.2.4 A Grower Practice of N Fertilizer and Irrigation Management

It is necessary to compare these potential BMPs with the actual N fertilizer and irrigation

management strategies utilized by sweet corn growers for their simulated dry matter yield (kg ha-

1) and accumulative amount of nitrogen leaching (kg ha-1), so as to determine whether these

potential BMPs have advantages in decreasing N leaching and maintaining an acceptable yield.

An example of N fertilizer and irrigation program was borrowed from the 2003 sweet corn

crop in an EPA319 demonstration project (Hochmuth, 2003). The project was conducted in

Suwannee Farms, O'Brien, FL, which is about 150 km from Pine Acres, the field experiment site

of this study. Most of the soil at Suwannee Farms is considered a Blanton Fine Sand or a Penney

Fine Sand according to the Suwannee County Soil Survey. Both of these soils are very similar,

and for practical purposes can be considered the same for nutrient and irrigation management.









The IFAS nitrogen fertilizer recommendation for sweet corn is 224 kg N ha-1, allowing for

an additional 34 kg N ha-1 for a leaching rain. However, in the "EPA319 project", total nitrogen

application of 302 kg N ha-1 was targeted.

Liquid nitrogen sources that can be injected through the irrigation system were

recommended and could greatly improve nutrient efficiency of placement and utilization.

Applications through the irrigation system should traditionally target 11 to 22 kg N ha-l, however

in the "EPA319 project", the goal was to maintain applications for approximate 45 kg N ha-1 per

fertigation event. The detailed information about the N management is specified in Table 6-3.

When reducing total nitrogen rates from a production program, irrigation management

becomes critical, and will become the major factor that keeps the nutrient management program

on track. Nitrogen is easily leached, especially on the sandy soils common to much of Florida

(Hochmuth and Hanlon, 2000).

The following irrigation recommendations (Table 6-4) are designed to provide adequate

moisture for the crop while minimizing leaching potential of nitrates. These recommendations

are also dependent on a highly efficient irrigation system that can apply specific irrigation

amounts. The guide is designed without consideration of rainfall. Any sufficient moisture

contribution from rainfall should replace the irrigation recommendation suggested for that day or

time period. Actual irrigation amounts should be increased or decreased depending on actual

weather conditions and crop growth rate.

This actual grower practice of nitrogen fertilizer and irrigation management was simulated

with the CERES-Maize model under the same fundamental inputs derived from the field

experiment in Block 1, Citra, FL. The model was run with the same input parameter distribution


214









obtained in Chapter 3 and the weather data described in Section 6.2.1. Then the uncertainties in

the predicted dry yield and nitrogen leaching were analyzed.

Finally the results of uncertainty analysis were compared with the six selected potential

BMPs mentioned above in order to make sure that the selected potential BMPs could actually

reduce nitrogen leaching amount while maintaining an acceptable yield.

6.2.5 Monte Carlo Simulation

Several methods have been used to account for uncertainty, such as Kalman filtering (Peter,

1979; Ahsam and O'Connor, 1994), first-order analysis (FOA) (Chaubey et al., 1999; Haan and

Skaggs, 2003a, 2003b), Monte Carlo simulation (MCS) (Hession et al., 1996; Haan and Skaggs,

2003a, 2003b; Ogle et al., 2003), Latin hypercube sampling (LHS) (Pebesman and Heuvelink,

1999), and generalized likelihood uncertainty estimation (GLUE) (Beven and Binley, 1992;

Beven, 1993). Among these methods, the Monte Carlo method is a "brute-force" approach to

estimate the probability density function of output variables from the probability density

functions of input variables (Hanna et al., 1997). It is the most commonly used non-structured

method. The Monte Carlo technique generates an estimate of the overall uncertainty in the

predictions due to all the uncertainties in the input parameters, regardless of interactions and

quantity of parameters (Macdonald and Strachan, 2001).

In Monte Carlo simulations, the input parameters are described by probability distributions,

and a single set of input data is randomly generated based on the distributions. This single data

set is run through the model and an output data set is obtained. The results of the run are stored

and a new set of input data is generated. Multiple simulations, typically thousands, are carried

out until the results of a new run do not affect the probability distribution of the output variable.

The number of simulations depends on the number and variability of input parameters, and the

required confidence in the output probability distribution (Graettinger and Dowding, 2001).









A Monte Carlo simulation was carried out for the six selected potential BMPs and one

actual management practice (described in Section 6.2.3 and 6.2.4), under the uncertainties of

input parameters and weather. The main procedure were as follows: (1) generate 1,000 random

parameter sets according to the statistical properties specified in Table 1; (2) run the model for

the six potential BMPs and the actual grower practice with these 1,000 parameter sets under 33

years' (1958-1990) measured weather data and record the relevant outputs, so 231,000

simulations were performed; and (3) process the output files and plot the results with the

software Matlab.

6.3 Results and Discussion

6.3.1 BMP Comparison

The mean values and standard deviations of simulated sweet corn dry yield and nitrogen

leaching amounts were summarized for three uncertainty scenarios: under only parameter

uncertainties, under only weather uncertainties, and under both parameter and weather

uncertainties (Table 6-5). When only considering parameter uncertainties, the weather condition

was fixed as 1958, since it is the first yield of simulation. When only considering weather

uncertainties, the parameter set was set as the nominal set, which was derived from the second-

round posterior distribution of GLUE simulation (Table 6-1).

To tell which pairs of outputs of the treatments are different, so as to determine which one

was more suitable, a one-way analysis of variance (ANOVA) was conducted for the scenarios of

under only parameter uncertainty, under only weather uncertainty, and under both parameter and

weather uncertainties (Table 6-5).

When under only parameter uncertainty, the six potential BMPs and the actual grower

practice show some difference in predicting dry matter yield and amount of nitrogen leaching.

The actual grower practice has the highest value of nitrogen leaching and a moderate dry yield.


216









However, for the scenarios of under only weather uncertainty and under both parameter and

weather uncertainties, the actual grower practice shows a significant difference in nitrogen

leaching, much higher than the six potential BMPs, while there is no difference among the BMPs.

For dry yield, all the BMPs and actual grower practice show no great difference, i.e. they give

the similar yields.

It can be seen in Table 6-5 that the weather was the dominant uncertainty contributor. This

was because after two rounds of GLUE simulation, the uncertainties existing in input parameters

were minimized (see Chapter 3 for details). However, the uncertainties of climate could not be

reduced artificially. Hence, when simulating the selected BMPs both with respect to parameter

and weather uncertainties, weather contributed the most part of uncertainties in model outputs.

When both considering weather and parameter influences, the predicted means of dry

matter yield per treatment ranged from about 3,310 kg ha-1 to 3,505 kg ha-1. The values of

coefficient of variation (CV) of the predicted yields only using the generated weather data were

all in the range of 25%. For predicted cumulative nitrogen leaching, the predicted mean values

per treatment ranged from about 30 to 76 kg ha-l. The values of CV were all in the range of 80%,

more than three times the variability of predicted dry matter yields.

It could be concluded that weather variability could cause higher uncertainty in model

outputs of nitrogen leaching than in yields. This was because nitrogen leaching was more

sensitive to weather conditions (especially rainfall), than yield.

According to the ANOVA result, it can also be concluded that the real case that applied

270 lb Nacre-' could not increase yield, but could increase nitrogen leaching significantly, i.e.

the selected potential BMPs did a better job in reducing nitrogen leaching while obtaining an

acceptable yield. For realistic production, it seems that BMP3 and BMP4 could be good choices









since they had the highest yields and relatively higher reliability. In BMP3 and BMP4, 196 kg N

ha-1 (175 lb N acre-1) was required, but this amount of nitrogen fertilizer was obtained based on

the assumption that the fertilizer application efficiency was 1.0. Actually this efficiency could be

not obtained due to irrigation uniformity, crop canopy interception, and effects of wind. So in

realistic production, an extra increment should be considered, for example a 10% increase. This

result supports the IFAS (Institute of Food and Agricultural Sciences of the University Of

Florida) recommendation of N fertilizer for sweet corn production, which is 224 kg N ha-1 (200

lb N acre-1) (Hochmuth, 2000).

6.3.2 Output Uncertainty Plot

The distributions of the predicted annual dry corn yields for each treatment over the 33-

year simulation period are shown in Figure 6-1, while the distributions of the predicted average

annual accumulative nitrogen leaching for each BMP treatment are shown in Figure 6-2.

These figures visualize the uncertainties of model output when both considering

uncertainties in input parameters (genotype and soil) and weather data. In each figure, each

predicted corn yield or nitrogen leaching of one year was an average over 1,000 different

simulations with 1,000 different sets of randomly generated input parameters were used. The

90% confidence interval (CI) estimated from the 5% and 95% quantiles of the cumulative

distribution function (CDF) were used as the uncertainty limits of the predictions (Haan and

Skaggs, 2003; Sabbagh and Fox, 1999).

From Figure 6-1, it can be seen that the distributions of the predicted dry yields under the

six potential BMPs and the actual grower practice all approximately follow a normal distribution,

especially for BMP3, BMP4, and BMP6. The 90% CI range of dry yield for BMP1, BMP2,

BMP5, and actual grower practice is 1,500 kg ha-1 (from about 2,800 to 4,300 kg ha-1). For









BMP3, BMP4, and BMP6, the value of 90% CI range of dry yield is also 1,500 kg ha-1, but

changes from about 2,800 to 4,300 kg ha-1.

The values of 50% quantile of the seven practices are all between 3,300 and 3,400 ka ha-1,

which is a little bit higher than the measured dry yield in field experiment in this study (Table 4-

15 in Chapter 4), science there was no water stress under these BMPs.

From Figure 6-2, it can be seen that the distributions of the predicted amounts of nitrogen

leaching under the six potential BMPs and the actual grower practice all fail to follow a normal

distribution. All of the distributions skew to the left side. Thus, it is better to find another kind of

distribution to describe the predicted nitrogen leaching in the DSSAT model.

The 90% CI range of nitrogen leaching during the season for the six selected potential

BMPs is wide, more than 70 kg N ha-1 (from about 10 to 80 kg ha-1). The 50% quantile is all

around 30 kg N ha-1. The predicted mean values (Table 6-5) of these BMPs were much lower

than the estimated ones in field experiment (Table 4-19 in Chapter 4). This is because the

predicted nitrogen leaching here was only the nitrogen completely leached during season, which

did not include the organic nitrogen left in the soil after maturity. If considering the sum of

nitrogen leaching during the season and inorganic nitrogen in soil profile in model prediction as

the potential nitrogen leaching, the predicted and estimated amounts of potential nitrogen

leaching would be much closer (See Chapter 4 details).

The 90% CI range of nitrogen leaching for the actual grower practice is much wider, more

than 120 kg N ha-1 (from about 10 to 130 kg ha-1). The 50% quantile is around 55 kg N ha-1. The

mean value nitrogen leaching of grower practice (Table 6-5) is also much lower than the

measured values (Table 4-19 in Chapter 4) due to the same reason mentioned above. However,

the result can also support the assumption that when more nitrogen is applied, more will be


219









leached, i.e. the selected BMPs can reduce nitrogen leaching compared with the actual grower

practice.

6.3.3 Output Uncertainty over Time Range of 1958-1990

It is necessary to show how the corn yield and nitrogen leaching changed in the simulation

years, since it could give model users an idea of what level the yield and nitrogen would change

if a special BMP strategy was used across in a range of actual weather conditions. For

convenience, only the selected BMP1 was chosen as an example. It is sure that the similar

analysis can be conducted for other selected BMPs and the actual grower practices. Figure 6-3

shows the simulated 10% and 90% confidence limits of the average yearly corn yields for the

study period for treatment BMP1, while Figure 6-4 shows similar information of the average

yearly nitrogen leaching.

From Figure 6-3, it can be seen that the average yields ranged between 2,500 kg ha-1 and

4,500 kg ha-l, considering both the weather and input parameter uncertainties. Cumulative

nitrogen leaching ranged between 10 and 90 kg N ha- (Figure 6-4). There exist great differences

among the amounts of cumulative nitrogen leaching in different years. For example, the nitrogen

leaching amount could be as high as about 90 kg N ha- in 1959, but it decreased to 40 kg N ha-

in 1960 and 1961. A 50 kg N ha- difference exists between these continuous years.

The variations in the curves in Figure 6-4 can be explained by the uncertainties in weather

conditions. For example, in Figure 6-4, there are two obvious peaks in nitrogen leaching in 1959

and 1984. The respective cumulative rainfall in the growth season of sweet corn (90 days after

planting date of March 8th) in 1959 and 1984 was 591 and 341 mm. In 1964 and 1981, when

there was the lowest amount of nitrogen leaching, the respective cumulative rainfall was 193 and

130 mm. These dry year cumulative rainfall amounts were only about one third of those in 1959

and 1984. For other years (e.g. 1971 and 1980), the cumulative rainfall was between 210 and 310


220









mm. These curves confirmed the sensitivity of nitrogen leaching to weather conditions,

especially rainfall.

However, the yield variations shown in Figure 6-3 were difficult to explain with a single

factor. The collective influence includes temporal variations of temperature, solar radiation, and

rainfall. For example, in 1987 the predicted yield was low (less than 2,500 kg ha-1). One reason

was probably because of the low temperature in the early growth season. The average maximum

and minimum temperature in the first 2 weeks of the growth season of 1987 was about 23.2 C

and 9.1 C, respectively, but the corresponding temperatures in 1986 was about 26.1 C and 12.5

C, i.e. 3 C higher than 1987. The cold weather could inevitably retard or harm the corn

seedlings, and finally reduce the yield. And the accumulated rainfall in growth season was about

324.2 mm that year, which resulted in a large amount of nitrogen leached and unavailable for

corn consumption.

6.4 Summary and Conclusions

Outputs of crop models may be uncertain depending on the range of uncertainty of the

input parameters (genotype and soil) and weather data. However, crop models still remained

important in applications related to estimation of production potentials, strategic and tactical

decisions and agricultural technology transfer, since they are efficient, quantitative tools for the

integration of complex and dynamic interactions of crops with climatic, soil, and agronomic

environments.

In this study, six selected potential best management practices and a real N fertilizer

application and irrigation management case were investigated for uncertainties of yield and

accumulative nitrogen leaching caused by weather and input parameter uncertainty. Some

conclusions were drawn as follows.









The weather was the dominant uncertainty contributor to model outputs such as dry matter

yield and nitrogen leaching during season, which was because after two rounds of GLUE

simulation, the uncertainties existing in input parameters were minimized (see Chapter 3 for

details). However, the uncertainties of climate could not be reduced artificially.

Weather variability could cause higher uncertainty in model outputs of nitrogen leaching

than in yields. This is because nitrogen leaching was more sensitive to weather conditions

(especially rainfall), than yield.

After comparison, the selected BMP3 (an irrigation of 7.5 mm with a MAD value of 30%,

a total of 196 kg N ha-1 with a split of 0-1/4-3/4 and an application amount of 40 kg N ha-1) and

BMP4 (an irrigation of 7.5 mm with a MAD value of 30%, a total of 196 kg N ha-1 with a split of

0-1/3-2/3 and an application amount of 30 kg N ha-1) could be good choices for real sweet corn

production, compared with other BMPs and the actual grower practice.

The simulation results support the recommendation of IFAS about N fertilizer for sweet

corn production, which is 224 kg N ha-1 (200 lb N acre-1) if fertilizer application efficiency was

considered.



























01 1500 2000
1000 1500 2000


Predicted Yield (kg/ha)


BMP2
(A)
i i I


2532 2884 3236 3588 3940 4292


3000 3500
Predicted Yield (kg/ha)


Figure 6-1. Histogram (A) and cumulative distribution (B) of predicted average annual dry yield

of the six selected potential BMPs and the actual grower practice both under weather

and input parameter uncertainty. The red curve in each histogram is the fitted normal

distribution curve. The 5%, 50% and 95% quantiles are shown as dots in (B).


BMP1
(A)
I I I I I


000 5500
5000 5500













BMP3
(A)


0 1
1000 1500 2000


Predicted Yield (kg/ha)


BMP4
(A)


4500 5000 5500


10 1500 2000


Predicted Yield (kg/ha)


Figure 6-1. Continued


224


5000 5500














BMP5
(A)
I I I I I


01 150
1000 1500 2000


Predicted Yield (kg/ha)


BMP6
(A)


4500 5000 5500


)0 1500 2000


Predicted Yield (kg/ha)


Figure 6-1. Continued


000 5500
5000 5500














Grower Practice
(A)
I I I I I


01 15 110
1000 100 2000
1000 1500 2000


Predicted Yield (kg/ha)


Figure 6-1. Continued


226


5 I
6000 5500











BMP1
(A)


RB


Predicted nitrogen leaching (kg/ha)


BMP2
(A)
30.


90 100


Predicted nitrogen leaching (kg/ha)


Figure 6-2. Histogram (A) and cumulative distribution (B) of predicted average annual nitrogen
leaching (NLCM) of the six selected potential BMPs and the actual grower practice
both under weather and input parameter uncertainty. The red curve in each histogram
is the fitted normal distribution curve. The 5%, 50% and 95% quantiles are shown as
dots in (B).














BMP3
(A)


70 80 90 11


Predicted nitrogen leaching (kg/ha)


BMP4
(A)
30.,


80 90 1i


Predicted nitrogen leaching (kg/ha)


Figure 6-2. Continued













BMP5
(A)


80 90 11


Predicted nitrogen leaching (kg/ha)


BMP6
(A)
30.,


80 90 1i


Predicted nitrogen leaching (kg/ha)


Figure 6-2. Continued


229













Grower Practice
(A)
SI I I


120 140 160 180 21


Predicted nitrogen leaching (kg/ha)


Figure 6-2. Continued


230











5000


4500
4000 .. "

o 3500. .. *



A ': \ 2500 A "" .... / 'V,
:A




2000

1500

1000
1958 1963 1968 1973 1978 1983 1988

Year

-- Mean ---- -- 90% Quantile ---- 10% Quantile

Figure 6-3. Simulated 10% and 90% confidence limits of average annual yields of BMP1 both
under weather and input parameter uncertainty


1958


1963


1968


1978


1983


1988


Year

--- Mean ---.--- 90% Quantile --- --- 10% Quantile


Figure 6-4. Simulated 10% and 90% confidence limits of average annual nitrogen leaching of
BMP1 both under weather and input parameter uncertainty


/IL







,,,E
", ; i ; l ~ ,,


, L ,
**










Table 6-1. Second posterior distribution of the selected parameters (from Chapter 3)
Parameter Unit Min Max Mean Standard Deviation CV
P1 Cd 77.676 182.175 99.169 8.217 8.3%
P5 Cd 553.141 676.212 577.201 9.746 1.7%
PHINT Cd 39.162 41.712 39.676 0.202 0.5%
SLDR 0.708 0.752 0.732 0.006 0.9%
SLRO 41.492 99.850 78.143 9.660 12.4%
SDUL cm3/cm3 0.097 0.109 0.104 0.002 1.6%
SLLL cm3/cm3 0.053 0.068 0.060 0.002 4.0%
SSAT cm3/cm3 0.235 0.362 0.300 0.021 7.0%
SLPF 0.760 0.932 0.872 0.041 4.7%



Table 6-2. Six selected potential BMPs for sweet corn production (from Chapter 5)
BMP Irr n Nitrogen Amount Nitrogen Application Amount
BMP Irrigati(kg N ha -) Split (kg N ha 1)
BMP1 5.0 mm-MAD 20% 196 0-1/4-3/4 30
BMP2 5.0 mm-MAD 20% 196 0-1/3-2/3 30
BMP3 7.5 mm-MAD 30% 196 0-1/4-3/4 40
BMP4 7.5 mm-MAD 30% 196 0-1/3-2/3 30
BMP5 5.0 mm-MAD 20% 224 0-1/4-3/4 30
BMP6 7.5 mm-MAD 30% 224 0-1/4-3/4 30










Table 6-3. N fertilizer management in the "EPA319 Project"
No. Time Management Description
Nitrogen based fertilizers should not be broadcast to the soil surface
before planting. This will be highly susceptible to leaching.
Traditional applications of nitrogen at planting are approximately 17
2 Planting kg N ha1 (15 lb N acre 1). However applications of 34 kg N ha1 (30
lb N acre-1) are acceptable.
By now the root system is developing rapidly and an application of
approximately 18 kg (40 lbs) nitrogen can be banded beside the corn
plants. The bands should be applied approximately 5 to 10 cm (2 to 4
Approximately inches) to the side of the corn row. It is recommended that this
21 days application be applied as a dry granular or liquid nitrogen solution. If
after planting applying liquid, applying approximately 5 to 10 cm (2 to 4 inches)
(DAP) below the soil sources will help reduce volatilization. Nitrogen
applications made through irrigation is not recommended, as the
small plants do not yet have sufficient roots to utilize broadcast
nitrogen efficiently.
Nitrogen applications made through the irrigation systems can now
improve nitrogen use efficiency if accompanied by appropriate
irrigation. Applications of approximate 45 kg N ha-1 (40 lb N acre 1)
4 Approximately per application can be made on weekly basis.
36 DAP Approximate nitrogen management should target between 3 and 5 of
these applications during the season. In this project, it is targeting a
maximum of 5 applications made through the irrigation system
applying approximately 45 kg N ha' each.
S Within 1 week Nitrogen applications made during the final week of ear development
from harvest and maturity are not beneficial.










Table 6-4. Irrigation management in the "EPA319 Project"
No. Time Management Description
Herbicides applied pre-plant that have a 'water-in' requirement
should receive no more than 0.64 cm (0.25 inches), even on a very
1 Pre-pla dry soil. An irrigation of this amount may wet the soil to a depth of at
Herbicides
least 4 inches and may penetrate much deeper, depending on current
soil moisture levels.
To facilitate planting, the soil should be moist enough to allow proper
2 Pre-plant seed placement. Apply no more than 0.51 cm (0.20 inches) if rainfall
has not provided sufficient moisture.
A light application of approximate 0.39 cm (0.15 inches) immediately
3 Post Plant after planting will allow the soil surface to 'crust'. This will help
retain soil moisture as the seeds begin to germinate and grow.
As the seeds germinate and growth begins, water requirements are
very small, and soil moisture levels within the root zone may remain
adequate for some time. Soil moisture should be checked 2 or 3 times
4 Pre-emergence per week.
Irrigation during this period usually applies 'surface moisture' to
prevent the soil surface from drying out excessively and reducing
wind erosion. These applications should apply a minimum amount.
As the plants emerge, and photosynthesis begins, water use increases
gradually with plant size.
Post-emergence Irrigation during this period should be scheduled every 3 or 4th day,
5 Approx. with rates of 0.38 to 0.64 cm (0.15 to 0.25 inches) per application.
10 to 18 DAP (Example: 0.38 cm at 10 DAP, 0.51 cm at 14 DAP, and 0.64 cm at 18
DAP.)
Soil moisture should be checked every other day, if not daily.
The sweet corn plants are growing rapidly and water use is increasing
accordingly.
6 19-38 DAP Irrigation events should be scheduled every 2nd or 3rd day, applying
0.64 to o.76 cm (0.25 to 0.30 inches) per application.
Soil moisture should be checked daily.
Irrigation should now be applied daily to meet the daily water uptake
by the sweet corn.
During this period, water use is still increasing, but is averaging just
7 39-48 DAP below 0.51 cm per day.
With a highly efficient irrigation system, applications of 0.51 to 0.64
cm (0.20 to 0.25 inches) per day should be adequate to keep up with
the crop requirements.
S 49-58 DAP Irrigation should be increased to around 0.64 to 0.69 cm (0.25 to 0.27
inches) per day.
59 DAP
g59 DAP Irrigation should be approximately 0.76 to 0.84 cm (0.30 to 0.33
9 through .
H are inches) per day.
Harvest


234










Table 6-5. Mean and standard deviation (STDEV) of simulated corn dry yield and nitrogen
leaching both under different uncertainty scenarios a
Treatment Dry Yield (kg ha-') Nitrogen leaching (kg ha 1)
Mean STDEV CV Mean STDEV CV
Under parameter uncertainties
BMP1 3230 a 491 15.2% 48 c 9 18.6%
BMP2 3268 a 497 15.2% 52 b 10 18.6%
BMP3 3064 b 454 14.8% 51 bc 9 17.2%
BMP4 2993 b 451 15.1% 51 bc 9 18.0%
BMP5 3252 a 495 15.2% 51 bc 9 17.9%
BMP6 2991 b 449 15.0% 50 c 8 16.9%
Grower Practice 3059 b 513 14.0% 101 a 22 22.4%
Under weather uncertainties
BMP1 3495 a 903 25.8% 35 b 24 68.3%
BMP2 3511 a 908 25.9% 37b 25 67.8%
BMP3 3532 a 737 20.9% 38 b 27 68.9%
BMP4 3513 a 674 19.2% 39 b 26 67.7%
BMP5 3511 a 905 25.8% 39 b 28 72.0%
BMP6 3516 a 677 19.3% 40 b 29 72.8%
Grower Practice 3186 a 801 25.1% 77 a 50 64.3%
Under both parameter set and weather uncertainties
BMP1 3396 a 900 25.9% 33 b 25 75.7%
BMP2 3505 a 912 26.0% 30 b 25 81.9%
BMP3 3334 a 822 24.7% 32 b 26 82.0%
BMP4 3309 a 794 24.0% 32 b 26 82.7%
BMP5 3491 a 907 26.0% 31 b 27 86.2%
BMP6 3308 a 891 26.9% 33 b 28 86.8%
Grower Practice 3482 a 813 23.4% 76 a 48 87.2%
a Values within columns followed by the same lower case letters were not significantly different at a 5% level,
according to the Tukey's Studentized Range (HSD) test.









CHAPTER 7
CONCLUSIONS AND FUTURE WORK

7.1 Summary and Research Contributions

Increasing nitrogen loads within the Suwannee River Basin of North Florida and other

areas has become a major concern. Leaching of nitrogen is economically and environmentally

undesirable. Nitrogen fertilizer application in field crop production is believed to be the most

import nitrogen contribution in this region. Something must be done to improve this situation.

Florida ranks highest in the nation in the production and value of fresh market sweet corn,

typically accounting for approximately 25% of both national sweet corn production and of U.S.

cash receipts for fresh sales. Thus it is necessary to develop research based nitrogen best

management practices (N-BMPs) to reduce nitrogen leaching while keeping an acceptable yield

in sweet corn production.

Crop models are becoming attractive for BMP development because field plot experiments

have spatial and temporal limitations, and are expensive and time-consuming. This study utilized

the CERES-Maize mode of the Decision Support System for Agrotechnology Transfer (DSSAT)

model as a platform to develop potential BMPs for sweet corn production in North Florida. The

research involved both field experiments and crop model simulations.

The main contributions of this research to the fields of crop modeling and land and water

resource engineering include follows:

(1) It is the first time to use the non-restricted and restricted one-at-a-time (OAT)
method for global sensitivity analysis for the CERES-Maize model of DSSAT.
These methods were proved to be effective tool to investigate the behavior of the
model and to select influential parameters for calibration.

(2) The generalized likelihood uncertainty estimation (GLUE) as a method for
model parameter estimation was used for hydrological models before. This is the
first time the method was successfully used in a crop model (CERES-Maize
model). The results showed that this method could significantly reduce the
uncertainties in model input parameters and consequently reduce the









uncertainties in model outputs. This research also tested the influence of
different likelihood functions and method of likelihood combination on GLUE
results. It also tried a procedure of GLUE verification to prove that GLUE is a
valid method. In general, this research has provided a paradigm for model
parameter estimation with the Bayesian method.

(3) It is the first time to use the calibrated CERES-Maize model as a computer
platform to conduct crop model experiments to explore potential best
management practices (BMPs) for nitrogen management.

(4) Uncertainty analysis was conducted for the selected potential BMPs, which
showed the different contributions of input parameters and weather conditions to
model output uncertainties. It will provide information for model users how
large the output uncertainty will be when using the model.

7.2 Conclusions

7.2.1 Global Sensitivity Analysis of CERES-Maize Model with One-at-a-time (OAT)
Method

In this research, global sensitivity analysis was used as a tool to select the most influential

input parameters for model calibration. Both non-restricted and restricted OAT methods were

used to conduct global sensitivity analysis for the CERES-Maize model. Some conclusions were

drawn as follows.

First, genotype parameters P1 (degree days from emergence to end of juvenile phase), P5

(degree days from silking to physiological maturity), PHINT (degree days required for a leaf tip to emerge)

and soil parameters SDUL (soil drained upper limit), SLLL (soil drained lower limit) and SLPF

(soil fertility factor) have a strong influence on dry matter yield. Second, genetic parameters P5

and PHINT and soil parameters SDUL, SLLL, and SLRO (soil runoff curve number) have strong

influence on nitrogen leaching. Third, soil parameters SLLL, SDUL and SSAT (soil saturation)

were highly correlated with each other.

Finally, nine parameters were selected for future model calibration with generalized

likelihood uncertainty estimation (GLUE) method (Chapter 3). They were P1, P5, PHINT, SLDR

(soil drainage rate), SLRO, SLPF, SLLL, SDUL and SSAT.









7.2.2 Parameter Estimation for CERES-Maize Model with GLUE Method

In this part of research, the GLUE method was used to estimate the influential genotype

and soil parameters, which were selected in global sensitivity analysis of the CERES-Maize

model. Some conclusions were drawn as follows.

According to the normality test with the Jarque-Bera method, it was found that all of the

selected parameters are close to or follow a normal distribution, except for SLPF. To determine

the number of model runs, it was found that at least 3,000 random parameter sets should be

generated and 3,000 model runs should be completed to guarantee reliable model simulation

results. To determine the best likelihood function and method of likelihood value combination

for this study, it was found that the likelihood functions and methods of likelihood value

combination could have very strong influence on the posterior distributions. The likelihood

function L1 (Equation L1 in Chapter 3) and method of combination C2 (Equation C2 in chapter

3) was the best choice, since the combination L1C2 had the lowest relative error, 0.01 for dry

matter yield, 0.09 for anthesis date, and 0.11 for maturity date (see Table 3-6 for details).

After two rounds of GLUE simulations, the uncertainty in input parameters and model

outputs were substantially reduced. For example, the value of standard deviation of input

parameter P1 for the prior, first posterior and second posterior distributions changed from 67.83

to 23.39, and to 8.22, respectively. For anthesis dates and maturity dates, the predictions were 55

and 80 days after planting after two rounds of GLUE simulation, which were very close to the

real field observations 51 and 80, respectively (see Table 3-6 for details).

The mean values of estimated and measured soil parameters were very close to each other.

For example, the mean value of calibrated SDUL in the second posterior distribution was 0.104

cm3/ cm3, while the mean value of measure SDUL was 0.110 cm3/ cm3. The error was only about

0.006 cm3/ cm3. Similar results were observed in SLLL and SSAT.









According to the results of GLUE verification, it can be seen that after two rounds of

GLUE, the uncertainties of the model outputs all decreased, and all mean values gradually

approached the selected true values. The expectations of the posterior distributions were used as

the nominal values to continue future research in the development of best management practices.

In general, the results of this study confirmed that the GLUE method was a powerful tool

to estimate the model input parameters.

7.2.3 Field Plot Experiment of Sweet Corn and Simulation with Calibrated CERES-Maize
Model

A field plot experiment was conducted at the Plant Science Research and Education Unit,

the University of Florida in the spring of 2006 to explore the influence of fertilizer applications

and irrigation levels on sweet corn quantity and quality. The nitrogen fertilizer and irrigation

treatments in field plot experiment were also simulated with the CERES-Maize model with the

expectation values of the posterior distributions of the selected influential parameters as the

nominal values. Thus, this part of research could be considered as a procedure of model

verification. Several conclusions were drawn.

Increasing nitrogen fertilizer from 185 to 247 kg N ha-1 significantly increased both fresh

total yield and marketable yield. Increasing irrigation rate from II (irrigation level based on daily

ET value and soil profile water balance) to 12 (1.5 times of I1), was not significant for fresh total

yield or for fresh marketable yield. The increase in N fertilizer rate was not significant in

increasing total ears, US #2, or cull ears per unit area, but increased the number of US #1 ears.

Both irrigation and nitrogen fertilizer levels showed significant influence on nitrogen

leaching. When irrigation level increased from I1 to 12, the average amount of estimated nitrogen

leaching increased from 150 to 167 kg N ha-1. When nitrogen fertilizer level increased from 247

to 309 kg N ha-l, the average amount of nitrogen leaching increased from 124 to 205 kg N ha-.









See Table 4-13 for details. This confirmed the common assumption that more water applied,

more nitrogen will be leached, and more nitrogen fertilizer applied, more nitrogen will be

leached as well.

After comparing the simulated and observed dry matter yields, anthesis dates and maturity

dates, and estimated nitrogen leaching of the seven treatments in field plot experiment of sweet

corn in 2006, it appears that the model did a good job in predicting dry yield and phenology

dates. From Table 4-15, it can be seen that the relative errors between the measured and

simulated yields were all near or less than 10% except for treatment FOI1 and F 12. From Table

4-16, it can be seen that the measured and simulated anthesis dates were that same, while there

was only one day difference in the maturity dates.

However there was a great difference if comparing the simulated nitrogen leaching during

season and the estimated potential nitrogen leaching. This is because the model calculated a

significant part of nitrogen output in the system as inorganic nitrogen left in the soil profile after

maturity. Thus, it is reasonable to use the sum of nitrogen leaching during the season and the

inorganic nitrogen left in soil profile after maturity as the potential nitrogen leaching in

prediction. If compare the predicted and measured potential nitrogen leaching (as shown in Table

4-19 in Chapter 4), the difference would be small, though there was still some uncertainties due

to the procedure of nitrogen leaching estimation.

7.2.4 Best Management Practices Development with CERES-Maize Model for Sweet Corn
Production in North Florida

In this study, the CERES-Maize module was utilized as a platform to develop BMPs for

sweet corn production in North Florida. The expectation values of the posterior distributions of

the input parameters derived in GLUE simulation were used as the nominal parameter set to









conduct the simulations. Each irrigation, nitrogen, or irrigation and nitrogen combination

treatment was simulated using 33 years of historical weather data for North Florida.

Irrigation frequency and amount significantly influenced corn yield. For example, when

apply water with a maximum allowable depletion (MAD) value of 10%, corn growth would

suffer from water stress, and yield would be reduced significantly to less than 1,000 kg ha-1.

Second, the trend of increasing nitrogen leaching was obvious if irrigation events were

more frequent and more water was applied in each event. For example, if the irrigation event of

22.5 mm was triggered with a MAD value of 90%, the nitrogen leaching amount was

approximately 32 kg ha-1. But it could be as high as almost 120kg ha-1, if the event was triggered

with a threshold of a MAD of 20% or 10% and a precipitation depth of 5 or 2.5 mm.

Third, more nitrogen applied resulted in more being leached. For example, the amount of

nitrogen leaching increased from 82 to 266 kg N ha-1 when nitrogen application level increased

from 196 to 561 kg N ha-.

Fourth, nitrogen fertilizer split did not show a significance influence on yield if there was

application of N during the small leaf stage or large leaf stage. However, splitting N applications

showed a significant influence on N leaching. The best splits were 0:1/4:3/4 (nothing in the small

leaf stage, 1/4 of the total nitrogen except for the starter nitrogen in large leaf stage, and 3/4 in

the ear development stage) and 0:1/3:2/3 (nothing in the small leaf stage, 1/3 of the total nitrogen

except for the starter nitrogen in large leaf stage, and 2/3 in the ear development stage), because

these fertigation schedules could best meet the nitrogen need of sweet corn growth, especially

from tasseling to maturity. A small "application amount" could not increase yield very much if it

was less than 70 kg N ha-l, but it could decrease N leaching from 110 kg N ha-1 to about 60 kg N

ha-1 when the value of "application amount" decreased from 70 kg N ha1 to 10 kg N ha-1









However, it seems 30, 40 or 50 kg N ha-1 could be good choices of "application amount", if the

production cost was considered in addition to yield and nitrogen leaching.

Finally, if growers could apply both irrigation water and nitrogen fertilizer more frequently

but with smaller amounts in each application, this would result in an acceptable yield and a lower

level of nitrogen leaching.

7.2.5 Uncertainty Analysis of Potential Sweet Corn BMPs under Weather and Input
Parameter Variability

In this study, six selected potential best management practices obtained in BMP

development (Chapter 5) and an actual N fertilizer application and irrigation management case

were investigated for uncertainties of yield and accumulative nitrogen leaching caused by

weather and input parameter uncertainty.

The weather was the dominant uncertainty contributor. This was because after two rounds

of GLUE simulation, the uncertainties existing in input parameters were minimized (see Chapter

3 for details). However, the uncertainties of climate could not be reduced artificially.

Second, weather variability could cause higher uncertainty in model outputs of nitrogen

leaching than in yields. This is because nitrogen leaching was more sensitive to weather

conditions (especially rainfall), than yield.

Third, after comparison, the selected BMP3 (an irrigation of 7.5 mm with a value of

maximum allowable depletion (MAD) of 30%, a total of 196 kg N ha-1 with a split of 0-1/4-3/4

and an application amount of 40 kg N ha-1) and BMP4 (an irrigation of 7.5 mm with a MAD

value of 30%, a total of 196 kg N ha-1 with a split of 0-1/3-2/3 and an application amount of 30

kg N ha-1) could be good choices for real sweet corn production, compared with other BMPs

(BMP1, BMP2, BMP5, and BMP6) and the actual grower practice in the EPA319 demonstration

project.









Finally, the simulation results supported the recommendation of IFAS about N fertilizer for

sweet corn production, which is 224 kg N ha-1 (200 lb N acre-1) if fertilizer application efficiency

is considered.

7.3 Future Work

The CERES-Maize model verified itself as a powerful tool to develop strategies for

agricultural production. It provided a convenient and economical way to obtain useful

information on the interactions between crop, soil, weather and field management strategies.

This study provided a useful paradigm of model sensitivity analysis, model calibration with

GLUE method, and BMP development. However, the current research still has some

disadvantages, requiring future work.

First, the current CERES-maize model can only predict the yield quantity, but not the

quality. According to the USDA sweet corn quality classification standard, the cobs of sweet

corn can be separated into three levels, US #1, US #2 and Cull. The US #1 and US #2 can be

sold in the market. Their sum is called the marketable yield. The current model can not make

such classification, and as such it would be a good research topic to add the quality component

into the current crop model.

Second, this research only focused on timing and amount of N fertilizer application, but it

did not consider other factors such as N fertilizer types (such as ammonia, nitrate, urea etc.), N

fertilizer application method (banding on surface, banding beneath surface, broadcast and

incorporated, or broadcast and not incorporated), and controlled release N fertilizer. These

factors all can influence N fertilizer use efficiency and leaching.

Third, this study only investigated potential N-based BMPs. The BMPs addressing

phosphorus, crop rotation, cover crop, and intercrop could all be simulated with the properly

calibrated crop model.









Fourth, this research did not integrate economic analysis of the developed BMPs. This

could strongly influence the adoption of these BMPs by growers. If a BMP is too expensive,

though it could decrease the N leaching significantly, it may not be adopted by growers. How to

balance the economic and social benefits would be an important topic.

Fifth, as described in Section 7.2.3 there was a great difference if comparing the simulated

nitrogen leaching during season (NLCM) and the estimated potential nitrogen leaching. This is

because the model calculated a significant part of nitrogen output in the system as inorganic

nitrogen left in the soil profile after maturity (NIAM). According to the results of nitrogen

leaching estimation in field experiment, it seems the model allocated too much nitrogen to

NIAM. Thus, in the future, it is also necessary to find the correct ratio between NLCM and

NIAM in model nitrogen prediction. This will be a good topic to improve model performance.










APPENDIX A
INPUT AND OUTPUT PARAMETERS OF CERES-MAIZE MODEL IN DSSAT


Symbol Definition
Thermal time from seedling emergence to the end
P1 of the juvenile phase (expressed in degree days
above a base temperature of 8 C) during which the
plant is not responsive to changes in photoperiod.
Extent to which development (expressed as days)
is delayed for each hour increase in photoperiod
P2 above the longest photoperiod at which
development proceeds at a maximum rate (which
is considered to be 12.5 hours).
Thermal time from silking to physiological
P5 maturity (expressed in degree days above a base
temperature of 8 C).
G2 Maximum possible number of kernels per plant.
G3 Kernel filling rate during the linear grain filling
stage and under optimum conditions.
Phylochron interval; the interval in thermal time
PHINT (degree days) between successive leaf tip
appearances.
SLLL Drained lower limit
SDUL Drained upper limit
SSAT Soil saturation water content
SBDM Soil bulk density
SALB Soil albedo
SLU1 Soil evaporation limit
SLRO Soil runoff curve number
SLDP Soil drainage rate
SLPF Growth reduction/ Fertility factor
HWAH Dry matter at maturity
NLCM Cumulative nitrogen leaching


Unit
degree days
above a base
temperature of
8C


Values based upon


DSSAT Database


DSSAT Database


degree days
above a base
temperature of
8C


mg day '

degree days

m3/m3
m3/m3
m 3/m3
g/cm3





kg ha-'
kg ha1


DSSAT Database

DSSAT Database
DSSAT Database

DSSAT Database

DSSAT Database
DSSAT Database
DSSAT Database
DSSAT Database
DSSAT Database
DSSAT Database
DSSAT Database
DSSAT Database
DSSAT Database
Simulation
Simulation










APPENDIX B
MATLAB CODE FOR GLOBAL SENSITIVITY ANALYSIS WITH THE RESTRICTED OAT
METHOD

B.1 Main Function

%%Main program to do soil and genotype parameter sensitivity analysis at the same time

n='Please input N to determine the numbers of simulations,N* 100';
disp(")
disp(n)
n=input('N=');
Num=n;

system('copy C:\MATLAB7\work\Sensitivity\MZCER040_Plate.cul
C:\DSSAT4\Genotype\MZCER040.cul');%%Fix the genotyoe file first
SoilSensitivityAnalysis(Num);

system('copy C:\MATLAB7\work\Sensitivity\soilPlate.sol C:\DSSAT4\Soil\soil.sol');%%Fix the soil file first
GenoSensitivityAnalysis(Num);

B.2 Sensitivity Analysis of Genotype Parameter

function GenoSensitivityAnalysis(N)

Num=N;

ResultRZero=zeros(0,2);
ResultRP 1Zero=zeros(0,2);
ResultRP2Zero=zeros(0,2);
ResultRP5Zero=zeros(0,2);
ResultRG2Zero=zeros(0,2);
ResultRG3Zero=zeros(0,2);
ResultRPHINTZero=zeros(0,2);

for i=l:l:Num

addpath C:\MATLAB7\work\Sensitivity;
[R,RP1,RP2,RP5,RG2,RG3,RPHINT]=GenoParameterSpace;

% SoilChange;%%Change soil file

GenoChange(R);%%Change genotype file with a set of norminal value
system('..\DSCSM040.EXE B D4Batch.DV4');
system('copy summary.out Output\SummaryGeno.txt');
[Result]=GenoSummaryProcess2005;
ResultRZero=[ResultRZero;Result];

GenoChange(RP1);%%Change genotype file with a set of norminal value but increase P1 by 5%
system('..\DSCSM040.EXE B D4Batch.DV4');
system('copy summary.out Output\SummaryGeno.txt');
[Result]=GenoSummaryProcess2005;










ResultRP Zero= [ResultRP Zero;Result];


GenoChange(RP2);%%Change genotype file with a set of norminal value but increase P2 by 5%
system('..\DSCSM040.EXE B D4Batch.DV4');
system('copy summary.out Output\SummaryGeno.txt');
[Result]=GenoSummaryProcess2005;
ResultRP2Zero=[ResultRP2Zero;Result];

GenoChange(RP5);%%Change genotype file with a set of norminal value but increase P5 by 5%
system('..\DSCSM040.EXE B D4Batch.DV4');
system('copy summary.out Output\SummaryGeno.txt');
[Result]=GenoSummaryProcess2005;
ResultRP5Zero=[ResultRP5Zero;Result];

GenoChange(RG2);%%Change genotype file with a set of norminal value but increase G2 by 5%
system('..\DSCSM040.EXE B D4Batch.DV4');
system('copy summary.out Output\SummaryGeno.txt');
[Result]=GenoSummaryProcess2005;
ResultRG2Zero= [ResultRG2Zero;Result];

GenoChange(RG3);%%Change genotype file with a set of norminal value but increase G3 by 5%
system('..\DSCSM040.EXE B D4Batch.DV4');
system('copy summary.out Output\SummaryGeno.txt');
[Result]=GenoSummaryProcess2005;
ResultRG3Zero= [ResultRG3Zero;Result];

GenoChange(RPHINT);%%Change genotype file with a set of norminal value but increase PHINT by 5%
system('..\DSCSM040.EXE B D4Batch.DV4');
system('copy summary.out Output\SummaryGeno.txt');
[Result]=GenoSummaryProcess2005;
ResultRPHINTZero=[ResultRPHINTZero;Result];

end

ResultRZero;
ResultRP Zero;
ResultRP2Zero;
ResultRP5Zero;
ResultRG2Zero;
ResultRG3Zero;
ResultRPHINTZero;

P1Change=ResultRPlZero-ResultRZero;
P2Change=ResultRP2Zero-ResultRZero;
P5Change=ResultRP5Zero-ResultRZero;
G2Change=ResultRG2Zero-ResultRZero;
G3Change=ResultRG3Zero-ResultRZero;
PHINTChange=ResultRPHINTZero-ResultRZero;

dP =P1Change./ResultRZero/0.05;%%Elementary effect of P1
dP2=P2Change./ResultRZero/0.05;%%Elementary effect of P2
dP5=P5Change./ResultRZero/0.05;%%Elementary effect of P5
dG2=G2Change./ResultRZero/0.05;%%Elementary effect of G2










dG3=G3Change./ResultRZero/0.05;%%Elementary effect of G3
dPHINT=PHINTChange./ResultRZero/0.05;%%Elementary effect of PHINT

dPlmeanYield=mean(dPl(:,1));
dPlvarianceYield=var(dP1(:,1));
dPlmeanNLCM=mean(dPl(:,2));
dPlvarianceNLCM=var(dPl(:,2));%%Mean and variance of elementary effect of P1 for HWAH and NLCM

dP2meanYield=mean(dP2(:,1));
dP2varianceYield=var(dP2(:,1));
dP2meanNLCM=mean(dP2(:,2));
dP2varianceNLCM=var(dP2(:,2));%%Mean and variance of elementary effect of P2 for HWAH and NLCM

dP5meanYield=mean(dP5(:,1));
dP5varianceYield=var(dP5(:,1));
dP5meanNLCM=mean(dP5(:,2));
dP5varianceNLCM=var(dP5(:,2));%%Mean and variance of elementary effect of P5 for HWAH and NLCM

dG2meanYield=mean(dG2(:,1));
dG2varianceYield=var(dG2(:,1));
dG2meanNLCM=mean(dG2(:,2));
dG2varianceNLCM=var(dG2(:,2));%%Mean and variance of elementary effect of G2 for HWAH and NLCM

dG3meanYield=mean(dG3(:,1));
dG3varianceYield=var(dG3(:,1));
dG3meanNLCM=mean(dG3(:,2));
dG3varianceNLCM=var(dG3(:,2));%%Mean and variance of elementary effect of G3 for HWAH and NLCM

dPHINTmeanYield=mean(dPHINT(:,1));
dPHINTvarianceYield=var(dPHINT(:, 1));
dPHINTmeanNLCM=mean(dPHINT(:,2));
dPHINTvariance_NLCM=var(dPHINT(:,2));%%Mean and variance of elementary effect of PHINT for
HWAH and NLCM


dSetGeno=[dP ,dP2,dP5,dG2,dG3,dPHINT];

fidnew-fopen('C:\MATLAB7\work\Sensitivity\dSetGeno.txt','w+');
for i=1:Num
fprintf(fidnew,'%10.9f %10.9f %10.9f %10.9f %10.9f %10.9f %10.9f %10.9f %10.9f %10.9f %10.9f
%10.9f\n',...
dSetGeno(i,1),dSetGeno(i,2),dSetGeno(i,3),dSetGeno(i,4),dSetGeno(i,5),dSetGeno(i,6),...
dSetGeno(i,7),dSetGeno(i,8),dSetGeno(i,9),dSetGeno(i,10),dSetGeno(i, 11),dSetGeno(i,12));
end
fclose(fid new);%%Make a txt file for the elementary effect values for each parameter

B.3 Genotype File Change

%%Modify Genotype File

function GenoChange(R)

%R=[1,2,3,4,5,6]











LineParal=47;
fidMZCER040=fopen('MZCER040_Plate.cul','r');
fidnew=fopen('C:\DSSAT4\Genotype\MZCERO40.cul','w+');

fori=1:50
line-fgetl(fidMZCER040);
if ~ischar(line), break, end

%%For P1, P2, P5, G2 and G3 and PHIN
if i==LineParal
linel=line(1:31);
line2=line(32:36);
line3=line(37:38);
line4=line(39:43);
line5=line(44);
line6=line(45:49);
line7=line(50);
line8=line(51:55);
line9=line(56:57);
line 10=line(58:60);
line 11=line(61:62);
line 12=line(63:66);
linel3=line(67:end);%%Devide the each line of genotype file into 13 sections

line2str=-num2str(R(1,1),'%5. 1f);%%P
line4str=-num2str(R(1,2),'%5.1f);%%P2
line6str=-num2str(R(1,3),'%4.1f);%%P5
line8str=num2str(R(1,4),'%4. 1f);%%G2
line 1 Ostr=num2str(R(1,5),'%4. 1f);%%G3
line 12str=num2str(R(1,6),'%4. If);%%PHINT

NewLine=[linel line2str line3 line4str line line6str line7 line8str line9 linelOstr line 1 linel2str linel3];
fprintf(fidnew,'%s\n',NewLine);

else
fprintf(fidnew,'%s\n',line);
end
end

fclose(fidMZCER040);
fclose(fidnew);

B.4 Genotype Parameter Space

%%Generate the paramter space for each parameter, then do sampling from
%%these spaces to form a series of nominal scenarios

function [R,RP 1,RP2,RP5,RG2,RG3,RPHINT]=GenoParameterSpace

N=100;% Section Number for every parameter

P1Min=5;










P1Max=450;
DeltaPl=P1Max-P1Min;
P1 Space=[P1Min:(1/(N-1))*DeltaP1 :P1Max];%Generation of sampling space of P1

P2Min=0;
P2Max=2;
DeltaP2=P2Max-P2Min;
P2Space= [P2Min:(1/(N-1))*DeltaP2:P2Max];%Generation of sampling space of P2

P5Min=580;
P5Max=990;
DeltaP5=P5Max-P5Min;
P5Space= [P5Min:(1/(N-1))*DeltaP5 :P5Max];%Generation of sampling space of P5

G2Min=248;
G2Max=990;
DeltaG2=G2Max-G2Min;
G2Space= [G2Min:(1/(N-1))*DeltaG2:G2Max];%Generation of sampling space of G2

G3Min=5;
G3Max=16.5;
DeltaG3=G3Max-G3Min;
G3Space=[G3Min:(1/(N-1))*DeltaG3:G3Max];%Generation of sampling space of G3

PHINTMin=30;
PHINTMax=50;
DeltaPHINT=PHINTMax-PHINTMin;
PHINTSpace= [PHINTMin:(1/(N- 1))*DeltaPHINT :PHINTMax];%Generation of sampling space of PHIN

Random = unifmd(1,100,[1 6]);%Generate the random number as the subscript of each parameter...
%selected from its own space

IntRandom=intl 6(Random);

R=[P1 )),P2(IntRandom()),P2Space(IntRandom(1,2)),P5Space(IntRandom(1,3)),...
G2Space(IntRandom( 1,4)),G3Space(IntRandom( 1,5)),PHINTSpace(IntRandom( ,6))];%0ne sapling of
parameter...
%%set


Incre=0.05; %Degree of increment

IncrePl=[P1Space(IntRandom(1,1))*Incre, 0, 0, 0, 0, 0];
RP 1 =R+IncreP 1 ;%%Increment of P1

IncreP2=[0, P2Space(IntRandom(1,2))*Incre, 0, 0, 0, 0];
RP2=R+IncreP2;%%Increment of P2

IncreP5=[0, 0, P5Space(IntRandom(1,3))*Incre, 0, 0, 0];
RP5=R+IncreP5 %%Increment of P5

IncreG2=[0, 0, 0, G2Space(IntRandom(1,4))*Incre, 0, 0];
RG2=R+IncreG2;%%Increment of G2











IncreG3=[0, 0, 0, 0, G3Space(IntRandom(1,5))*Incre, 0];
RG3=R+IncreG3;%%Increment of G3

IncrePHINT=[0, 0, 0, 0, 0, PHINTSpace(IntRandom(1,6))*Incre];
RPHINT=R+IncrePHINT;%%Increment of PHINT

B.5 Processing Sensitivity Analysis Results of Genotype Parameter

%%Processing the Sensitivity simulation results, selecting the data needed

function [Result]=GenoSummaryProcess2005

N=5;

%%%%To get the values of anthesis days, maturity days, yield and nitrogen
%%%%leaching.

fid_Summary=fopen('C:\DSSAT4\Maize\Output\SummaryGeno.txt','r');

LineParal=5;

fori=1:(N+2)
line-fgetl(fidSummary);
if ~ischar(line), break, end

if i==LineParal
linel=line(1:66);
line2=line(67:73);
line3=line(74);
line4=line(75:81);
line5=line(82);
line6=line(83:89);
line7=line(90:110);
line8=line(111:115);
line9=line(116:219);
line10=line(220:222);
line 11 =ine(223: end);

PDAT=str2num(line2);%%Planting date
ADAT=str2num(line4);%%oAnthesis date
MDAT=str2num(line6);%%Maturity date
HWAH=str2num(line8);%%Yield
NLCM=str2num(line10);%%Nitrogen leaching

% Result= [PDAT,ADAT,MDAT,HWAH,NLCM];
Result=[HWAH,NLCM];
end
end

fclose(fidSummary);










B.6 Sensitivity Analysis of Soil Parameter


function SoilSensitivityAnalysis(N)

Num=N;

ResultRZero=zeros(0,2);
ResultRSALBZero=zeros(0,2);%%Storage for the results of SALB 1
ResultRSLUlZero=zeros(0,2);%%Storage for the results of SLU1 2
ResultRSLDRZero=zeros(0,2);%%Storage for the results of SLDR 3
ResultRSLROZero=zeros(0,2);%%Storage for the results of SLRO 4
ResultRSLPFZero=zeros(0,2);%%Storage for the results of SLPF 5
ResultRSLLLZero=zeros(0,2);%%Storage for the results of SLLL 6
ResultRSDULZero=zeros(0,2);%%Storage for the results of SDUL 7
ResultRSSATZero=zeros(0,2);%%Storage for the results of SDUL 8
ResultRSBDMZero=zeros(0,2);%%Storage for the results of SDUL 9

for i=l:l:Num

addpath C:\MATLAB7\work\Sensitivity;
[R, RSALB, RSLU1, RSLDR, RSLRO, RSLPF, RSLLL, RSDUL, RSSAT, RSBDM]=SoilParameterSpace;

% SoilChange;%%Change soil file

SoilChange(R);%%Change soil file with a set of norminal value
system('..\DSCSM040.EXE B D4Batch.DV4');
system('copy summary.out Output\SummarySoil.txt');
[Result]=SoilSummaryProcess2005;
ResultRZero=[ResultRZero;Result];

SoilChange(RSALB);%%Change soil file with a set of norminal value but increase SALB by 5%
system('..\DSCSM040.EXE B D4Batch.DV4');
system('copy summary.out Output\SummarySoil.txt');
[Result]=SoilSummaryProcess2005;
ResultRSALBZero=[ResultRSALBZero;Result];

SoilChange(RSLU1);%%Change soil file with a set of norminal value but increase SLU1 by 5%
system('..\DSCSM040.EXE B D4Batch.DV4');
system('copy summary.out Output\SummarySoil.txt');
[Result]=SoilSummaryProcess2005;
ResultRSLUlZero= [ResultRSLU Zero;Result];

SoilChange(RSLDR);%%Change soil file with a set of norminal value but increase SLDR by 5%
system('..\DSCSM040.EXE B D4Batch.DV4');
system('copy summary.out Output\SummarySoil.txt');
[Result]=SoilSummaryProcess2005;
ResultRSLDRZero=[ResultRSLDRZero;Result];

SoilChange(RSLRO);%%Change soil file with a set of norminal value but increase SLRO by 5%
system('..\DSCSM040.EXE B D4Batch.DV4');
system('copy summary.out Output\SummarySoil.txt');
[Result]=SoilSummaryProcess2005;
ResultRSLROZero=[ResultRSLROZero;Result];











SoilChange(RSLPF);%%Change soil file with a set of norminal value but increase SLPF by 5%
system('..\DSCSM040.EXE B D4Batch.DV4');
system('copy summary.out Output\SummarySoil.txt');
[Result]=SoilSummaryProcess2005;
ResultRSLPFZero=[ResultRSLPFZero;Result];

SoilChange(RSLLL);%%Change soil file with a set of norminal value but increase SLLL by 5%
system('..\DSCSM040.EXE B D4Batch.DV4');
system('copy summary.out Output\SummarySoil.txt');
[Result]=SoilSummaryProcess2005;
ResultRSLLLZero=[ResultRSLLLZero;Result];

SoilChange(RSDUL);%%Change soil file with a set of norminal value but increase SDUL by 5%
system('..\DSCSM040.EXE B D4Batch.DV4');
system('copy summary.out Output\SummarySoil.txt');
[Result]=SoilSummaryProcess2005;
ResultRSDULZero=[ResultRSDULZero;Result];

SoilChange(RSSAT);%%Change soil file with a set of norminal value but increase SSAT by 5%
system('..\DSCSM040.EXE B D4Batch.DV4');
system('copy summary.out Output\SummarySoil.txt');
[Result]=SoilSummaryProcess2005;
ResultRSSATZero=[ResultRSSATZero;Result];

SoilChange(RSBDM);%%Change soil file with a set of norminal value but increase SBDM by 5%
system('..\DSCSM040.EXE B D4Batch.DV4');
system('copy summary.out Output\SummarySoil.txt');
[Result]=SoilSummaryProcess2005;
ResultRSBDMZero=[ResultRSBDMZero;Result];


end

ResultRZero;
ResultRSALBZero;%%Storage for the results of SALB 1
ResultRSLUlZero;%%Storage for the results of SLU1 2
ResultRSLDRZero;%%Storage for the results of SLDR 3
ResultRSLROZero;%%Storage for the results of SLRO 4
ResultRSLPFZero;%%Storage for the results of SLPF 5
ResultRSLLLZero;%%Storage for the results of SLLL 6
ResultRSDULZero;%%Storage for the results of SDUL 7
ResultRSSATZero;%%Storage for the results of SDUL 8
ResultRSBDMZero;%%Storage for the results of SDUL 9

SALBChange=ResultRSALBZero-ResultRZero;
SLU1Change=ResultRSLUlZero-ResultRZero;
SLDRChange=ResultRSLDRZero-ResultRZero;
SLROChange=ResultRSLROZero-ResultRZero;
SLPFChange=ResultRSLPFZero-ResultRZero;
SLLLChange=ResultRSLLLZero-ResultRZero;
SDULChange=ResultRSDULZero-ResultRZero;
SSATChange=ResultRSSATZero-ResultRZero;










SBDMChange=ResultRSBDMZero-ResultRZero;


dSALB=SALBChange./ResultRZero/0.05;%%Elementary effect of SALB
dSLUl=SLU1Change./ResultRZero/0.05;%%Elementary effect of SLUl
dSLDR=SLDRChange./ResultRZero/0.05;%%Elementary effect of SLDR
dSLRO=SLROChange./ResultRZero/0.05;%%Elementary effect of SLRO
dSLPF=SLPFChange./ResultRZero/0.05;%%Elementary effect of SLPF
dSLLL=SLLLChange./ResultRZero/0.05;%%Elementary effect of SLLL
dSDUL=SDULChange./ResultRZero/0.05;%%Elementary effect of SDUL
dSSAT=SSATChange./ResultRZero/0.05;%%Elementary effect of SSAT
dSBDM=SBDMChange./ResultRZero/0.05;%%Elementary effect of SBDM

dSALBmeanYield=mean(dSALB(:,1));
dSALBvarianceYield=var(dSALB(:,1));
dSALBmean_NLCM=mean(dSALB(:,2));
dSALBvariance_NLCM=var(dSALB(:,2));%%Mean and variance of elementary effect of SALB for HWAH
and NLCM

dSLUlmeanYield=mean(dSLU1(:,1));
dSLUlvarianceYield=var(dSLUl(:,l));
dSLUlmeanNLCM=mean(dSLU1 (:,2));
dSLUlvarianceNLCM=var(dSLUl(:,2));%%Mean and variance of elementary effect of SLU1 for HWAH
and NLCM

dSLDRmeanYield=mean(dSLDR(:,1));
dSLDRvarianceYield=var(dSLDR(:,1));
dSLDRmean_NLCM=mean(dSLDR(:,2));
dSLDRvariance_NLCM=var(dSLDR(:,2));%%Mean and variance of elementary effect of SLDR for HWAH
and NLCM

dSLROmeanYield=mean(dSLRO(:,1));
dSLROvarianceYield=var(dSLRO(:,1));
dSLROmean_NLCM=mean(dSLRO(:,2));
dSLROvariance_NLCM=var(dSLRO(:,2));%%Mean and variance of elementary effect of SLRO for HWAH
and NLCM

dSLPFmeanYield=mean(dSLPF(:,1));
dSLPFvarianceYield=var(dSLPF(:,1));
dSLPFmean_NLCM=mean(dSLPF(:,2));
dSLPFvariance_NLCM=var(dSLPF(:,2));%%Mean and variance of elementary effect of SLPF for HWAH and
NLCM

dSLLLmeanYield=mean(dSLLL(:,1));
dSLLLvarianceYield=var(dSLLL(:, 1));
dSLLLmeanNLCM=mean(dSLLL(:,2));
dSLLLvarianceNLCM=var(dSLLL(:,2));%%Mean and variance of elementary effect of SLLL for HWAH
and NLCM

dSDULmeanYield=mean(dSDUL(:,1));
dSDULvarianceYield=var(dSDUL(:,1));
dSDULmeanNLCM=mean(dSDUL(:,2));
dSDULvarianceNLCM=var(dSDUL(:,2));%%Mean and variance of elementary effect of SDUL for HWAH
and NLCM











dSSATmeanYield=mean(dSSAT(:,1));
dSSATvarianceYield=var(dSSAT(:,l));
dSSATmeanNLCM=mean(dSSAT(:,2));
dSSATvariance_NLCM=var(dSSAT(:,2));%%Mean and variance of elementary effect of SSAT for HWAH
and NLCM

dSBDMmeanYield=mean(dSBDM(:,1));
dSBDMvarianceYield=var(dSBDM(:,1));
dSBDMmeanNLCM=mean(dSBDM(:,2));
dSBDMvariance_NLCM=var(dSBDM(:,2));%%Mean and variance of elementary effect of SBDM for HWAH
and NLCM


dSetSoil=[dSALB, dSLU1, dSLDR, dSLRO, dSLPF, dSLLL, dSDUL, dSSAT, dSBDM];

fidnew-fopen('C:\MATLAB7\work\Sensitivity\dSetSoil.txt','w+');
for i=1:Num
fprintf(fidnew,'%10.9f %10.9f %10.9f %10.9f %10.9f %10.9f %10.9f %10.9f %10.9f %10.9f %10.9f %10.9f
%10.9f %10.9f %10.9f %10.9f %10.9f %10.9f\n',...
dSetSoil(i,1),dSetSoil(i,2),dSetSoil(i,3),dSetSoil(i,4),dSetSoil(i,5),dSetSoil(i,6),...
dSetSoil(i,7),dSetSoil(i,8),dSetSoil(i,9),dSetSoil(i, 10),dSetSoil(i, 11),dSetSoil(i, 12),...
dSetSoil(i, 13),dSetSoil(i, 14),dSetSoil(i, 15),dSetSoil(i, 16),dSetSoil(i, 17),dSetSoil(i, 18));
end
fclose(fid_new);%%Make a txt file for the elementary effect values for each parameter

B.7 Soil File Change

%%Modify Soil File

function SoilChange(R)

LineParal=60;
LinePara2=62;
fid_soil=fopen('soilPlate.sol','r');
fidnew=fopen('C:\DSSAT4\Soil\soil.sol','w+');

for i=1:70
line=fgetl(fid_soil);
if ~ischar(line), break, end

%%For SALB, SLU1, SLDR SLRO and SLPF
if i==LineParal
linel=line(1:8);
line2=line(9:12);
line3=line(13:15);
line4=line(16:18);
line5=line(19:20);
line6=line(21:24);
line7=line(25:26);
line8=line(27:30);
line9=line(31:32);
line 10=line(33:36);










line 11
line 12=
line 13=


=ine(37:38);
1line(39:42);
l=ine(243:end);


line2str-num2str(R(1,1),'%4.2f);
line4str=-num2str(R(1,2),'%3. 1f);
line6str=num2str(R(1,3),'%4.2f);
line8str-num2str(R(1,4),'%4. 1f);
line 12str=num2str(R(1,5),'%4.2f);

NewLine=[linel line2str line3 line4str line5 line6str line7 line8str line9 linel0 line 1 linel2str linel3];
fprintf(fidnew,'%s\n',NewLine);

%%Generate the values of SLLL, SDUL, SSAT and SBDM for other 4 layers according to the random value
of the first layer
ParaSLLL=R(1,6);
MeanSLLL=[0.12979 0.13355 0.13740 0.14300 0.14471];
StdevSLLL=[0.08797 0.08901 0.09332 0.10033 0.10152];
PertSLLL=(ParaSLLL-MeanSLLL(1))/StdevSLLL(1);

ParaSDUL=R(1,7);
MeanSDUL=[0.25205 0.25550 0.25824 0.25836 0.25716];
StdevSDUL=[0.10966 0.11195 0.11560 0.12133 0.12151];
PertSDUL=(ParaSDUL-MeanSDUL(1))/StdevSDUL(1);

ParaSSAT=R(1,8);
MeanSSAT=[0.37986 0.37619 0.37831 0.37816 0.37383];
StdevSSAT=[0.09691 0.09334 0.09803 0.09329 0.09050];
PertSSAT=(ParaSSAT-MeanSSAT(1))/StdevSSAT(1);


ParaSBDM=R(1,9);
MeanSBDM=[1.327068966 1.352068966 1.370344828
StdevSBDM=[0.213196299 0.215103065 0.212660387
PertSBDM=(ParaSSAT-MeanSBDM(1))/StdevSBDM(1);
%%Mean, STDEV and perturbation of each layer


1.382241379
0.217957368


1.393448276];
0.227569552];


forj=l:5
Para2SLLL(j ,)=MeanSLLL()+PertSLLL*StdevSLLL();
Para2SDUL(j,1)=MeanSDUL(j)+PertSDUL*StdevSDUL(j);
Para2SSAT(j, 1)=MeanSSAT(j)+PertSSAT* StdevSSAT(j);
Para2SBDM(j, )=MeanSBDM(j)+PertSBDM*StdevSBDM(j);

if Para2SLLL(j,1)==0
Para2SLLL(j,1)=Para2SLLL(j,1)+0.001;
Para2SDUL(j,1)=Para2SDUL(j,1)+0.002;
Para2SSAT(j,1)=Para2SSAT(j,1)+0.003;
end

if Para2SLLL(j,1)==Para2SDUL(j,1)
Para2SDUL(j,1)=Para2SDUL(j,1)+0.001;
Para2SSAT(j,1)=Para2SSAT(j,1)+0.002;
end










if Para2SDUL(j,1)==Para2SSAT(j,1)
Para2SSAT(j,1)=Para2SSAT(j,1)+0.001;
end
end

Para2=ones(5,4);%% Define the parameter matrix 5X3, for SLLL, SDUL, SSAT and SBDM

Para2Temp(:,:)=[Para2SLLL(:,1) Para2SDUL(:,I) Para2SSAT(:,I) Para2SBDM(:,I)];
Para2(:,:)=abs(Para2Temp(:,:));
B=sort(Para2(:,:),2);
Para2(:,:)=B;%%% Adjust the values of SLLL, SDUL and SSAT

%%For Layer 1 of SLLL SDUL and SSAT
elseif i==LinePara2
linel=line(1:13);
line2=line(14:18);
line3=line(19);
line4=line(20:24);
line5=line(25);
line6=line(26:30);
line7=line(31:44);
line8=line(45:48);
line9=line(49:end);

line2str=-num2str(Para2(1,1)+0.001,'%5.3f);
line4str=num2str(Para2(1,2)+0.002,'%5.3f);
line6str=-num2str(Para2(1,3)+0.003,'%5.3f);
line8str=num2str(Para2(1,4),'%4.2f);

NewLine=[line 1 line2str line3 line4str line5 line6str line7 line8str line9];
fprintf(fidnew,'%s\n',NewLine);

%%For Layer 2 of SLLL SDUL and SSAT
elseif i==LinePara2+1
linel=line(1:13);
line2=line(14:18);
line3=line(19);
line4=line(20:24);
line5=line(25);
line6=line(26:30);
line7=line(31:44);
line8=line(45:48);
line9=line(49:end);

line2str=-num2str(Para2(2,1)+0.001,'%5.3f);
line4str=num2str(Para2(2,2)+0.002,'%5.3f);
line6str=num2str(Para2(2,3)+0.003,'%5.3f);
line8str=num2str(Para2(2,4),'%4.2f);

NewLine=[line 1 line2str line3 line4str line5 line6str line7 line8str line9];
fprintf(fidnew,'%s\n',NewLine);

%%For Layer 3 of SLLL SDUL and SSAT










elseif i==LinePara2+2
linel=line(1:13);
line2=line(14:18);
line3=line(19);
line4=line(20:24);
line5=line(25);
line6=line(26:30);
line7=line(31:44);
line8=line(45:48);
line9=line(49:end);

line2str=-num2str(Para2(3,1)+0.001,'%5.3f);
line4str=-num2str(Para2(3,2)+0.002,'%5.3f);
line6str=num2str(Para2(3,3)+0.003,'%5.3f);
line8str=num2str(Para2(3,4),'%4.2f);

NewLine=[line 1 line2str line3 line4str line5 line6str line7 line8str line9];
fprintf(fidnew,'%s\n',NewLine);

%%For Layer 4 of SLLL SDUL and SSAT
elseif i==LinePara2+3
linel=line(1:13);
line2=line(14:18);
line3=line(19);
line4=line(20:24);
line5=line(25);
line6=line(26:30);
line7=line(31:44);
line8=line(45:48);
line9=line(49:end);

line2str=-num2str(Para2(4,1)+0.001,'%5.3f);
line4str=-num2str(Para2(4,2)+0.002,'%5.3f);
line6str=num2str(Para2(4,3)+0.003,'%5.3f);
line8str=num2str(Para2(4,4),'%4.2f);

NewLine=[line 1 line2str line3 line4str line5 line6str line7 line8str line9];
fprintf(fidnew,'%s\n',NewLine);

%%For Layer 5 of SLLL SDUL and SSAT
elseif i==LinePara2+4
linel=line(1:13);
line2=line(14:18);
line3=line(19);
line4=line(20:24);
line5=line(25);
line6=line(26:30);
line7=line(31:44);
line8=line(45:48);
line9=line(49:end);

line2str=-num2str(Para2(4,1)+0.001,'%5.3f);
line4str=-num2str(Para2(4,2)+0.002,'%5.3f);










line6str=num2str(Para2(4,3)+0.003,'%5.3f);
line8str=num2str(Para2(4,4),'%4.2f);

NewLine=[line 1 line2str line3 line4str line5 line6str line7 line8str line9];
fprintf(fidnew,'%s\n',NewLine);

else
fprintf(fidnew,'%s\n',line);
end
end

fclose(fid_soil);
fclose(fidnew);

B.8 Soil Parameter Space

%%Generate the paramter space for each parameter, then do sampling from
%%these spaces to form a series of nominal scenarios

function [R, RSALB, RSLU1, RSLDR, RSLRO, RSLPF, RSLLL, RSDUL, RSSAT,
RSBDM]=SoilParameterSpace

N=100;% Section Number for every parameter

SALBMin=0.07;
SALBMax=0.18;
DeltaSALB=SALBMax-SALBMin;
SALBSpace=[SALBMin:(1/(N-1))*DeltaSALB:SALBMax];%Generation of sampling space of SALB

SLU1Min=2;
SLU1Max=12.7;
DeltaSLUl=SLU1Max-SLU1Min;
SLU1 Space= [SLU1Min:(1/(N-1))*DeltaSLU1: SLU1Max];%Generation of sampling space of SLU1

SLDRMin=0;
SLDRMax=1;
DeltaSLDR=SLDRMax-SLDRMin;
SLDRSpace=[SLDRMin:(1/(N-1))*DeltaSLDR:SLDRMax];%Generation of sampling space of SLDR

SLROMin=30;
SLROMax=95;
DeltaSLRO=SLROMax-SLROMin;
SLROSpace=[SLROMin:(1/(N-1))*DeltaSLRO:SLROMax];%Generation of sampling space of SLRO

SLPFMin=0.7;
SLPFMax=1;
DeltaSLPF=SLPFMax-SLPFMin;
SLPFSpace= [SLPFMin:(1/(N-1))*DeltaSLPF:SLPFMax];%Generation of sampling space of SLPF

SLLLMin=0.02;
SLLLMax=0.252;
DeltaSLLL=SLLLMax-SLLLMin;
SLLLSpace= [SLLLMin: (1/(N- 1))*DeltaSLLL: SLLLMax];%Generation of sampling space of SLLL











SDULMin=0.253;
SDULMax=0.374;
DeltaSDUL=SDULMax-SDULMin;
SDULSpace=[SDULMin:(1/(N-1))*DeltaSDUL:SDULMax];%Generation of sampling space of SDUL

SSATMin=0.375;
SSATMax=0.7;
DeltaSSAT=SSATMax-SSATMin;
SSATSpace=[SSATMin:(1/(N- 1))*DeltaSSAT: SSATMax];%Generation of sampling space of SSAT

SBDMMin=0.7;
SBDMMax=1.66;
DeltaSBDM=SBDMMax-SBDMMin;
SBDMSpace=[SBDMMin:(1/(N-1))*DeltaSBDM:SBDMMax];%Generation of sampling space of BDM

Random = unifmd(1,100,[1 9]);%Generate the random number as the subscript of each parameter...
%selected from its own space

IntRandom=intl 6(Random);

R=[SALBSpace(IntRandom(1,1)),SLU1Space(IntRandom(1,2)),SLDRSpace(IntRandom(1,3)),..
SLROSpace(IntRandom(1,4)),SLPFSpace(IntRandom(1,5)),SLLLSpace(IntRandom(1,6)),...
SDULSpace(IntRandom(1,7)),SSATSpace(IntRandom(l,8)),SBDMSpace(IntRandom(1,9))];%One sapling
of parameter...
%%set


Incre=0.05; %Degree of increment

IncreSALB=[SALBSpace(IntRandom(1,1))*Incre, 0, 0, 0, 0, 0, 0, 0, 0];
RSALB=R+IncreSALB;%%Increment of SALB

IncreSLUl=[0, SLU1Space(IntRandom(1,2))*Incre, 0, 0, 0, 0, 0, 0, 0];
RSLUl=R+IncreSLU1;%%Increment of SLU1

IncreSLDR=[0, 0, SLDRSpace(IntRandom(1,3))*Incre, 0, 0, 0, 0, 0, 0];
RSLDR=R+IncreSLDR;%%Increment of SLDR

IncreSLRO=[0, 0, 0, SLROSpace(IntRandom(1,4))*Incre, 0, 0, 0, 0, 0];
RSLRO=R+IncreSLRO;%%Increment of SLRO

IncreSLPF=[0, 0, 0, 0, SLPFSpace(IntRandom(1,5))*Incre, 0, 0, 0, 0];
RSLPF=R+IncreSLPF;%%Increment of SLPF

IncreSLLL=[0, 0, 0, 0, 0, SLLLSpace(IntRandom(1,6))*Incre, 0, 0, 0];
RSLLL=R+IncreSLLL;%%Increment of SLLL

IncreSDUL=[0, 0, 0, 0, 0, 0, SDULSpace(IntRandom(1,7))*Incre, 0, 0];
RSDUL=R+IncreSDUL;%%Increment of SDUL

IncreSSAT=[0, 0, 0, 0, 0, 0, 0, SSATSpace(IntRandom(1,8))*Incre, 0];
RSSAT=R+IncreSSAT;%%Increment of SSAT











IncreSBDM=[0, 0, 0, 0, 0, 0, 0, 0, SBDMSpace(IntRandom(1,9))*Incre];
RSBDM=R+IncreSBDM;%%Increment of SBDM

B.9 Processing Sensitivity Analysis Results of Soil Parameter

%%Processing the Sensitivity simulation results, selecting the data needed
function [Result]=SoilSummaryProcess2005

N=5;
%%%%To get the values of anthesis days, maturity days, yield and nitrogen
%%%%leaching.

fid_Summary=fopen('C:\DSSAT4\Maize\Output\SummarySoil.txt','r');

LineParal=5;

fori=1:(N+2)
line-fgetl(fidSummary);
if ~ischar(line), break, end

if i==LineParal
linel=line(1:66);
line2=line(67:73);
line3=line(74);
line4=line(75:81);
line5=line(82);
line6=line(83:89);
line7=line(90:110);
line8=line(111:115);
line9=line(116:219);
line10=line(220:222);
line 11 =ine(223: end);

PDAT=str2num(line2);%%Planting date
ADAT=str2num(line4);%%oAnthesis date
MDAT=str2num(line6);%%Maturity date
HWAH=str2num(line8);%%Yield
NLCM=str2num(line 10);%%Nitrogen leaching

% Result= [PDAT,ADAT,MDAT,HWAH,NLCM];
Result=[HWAH,NLCM];
end
end

fclose(fid_Summary);










APPENDIX C
MATLAB CODE FOR GLUE PROCESS

C.1 Main Function

%GLUE simulations, totally Num* 100 times

n='Please input N to determine the numbers of simulations,N* 100';
disp(")
disp(n)
n=input('N=');
Num=n

RG(Num);
ParalZero=zeros(0,3);
Para2Zero=zeros(0,3);
Para3Zero=zeros(0,3);

for i=l:l:Num

fprintf('This is the %gth batch of simulation.\n',i);

addpath C:\MATLAB7\work\Soil;
[Paral,Para2,Para3]=ParaSetup(i);

Para2(1,:,:);
for m=1:100
for n=1:3
Para22(m,n)=Para2(1,n,m);
end
end

ParalZero=[ParalZero;Paral];
Para2Zero=[Para2Zero;Para22];
Para3Zero=[Para3Zero;Para3] ;%%Collect generated parameter sets

SoilChange;%%Change soil file
GenoChange;%%Change genotype file

%addpath C:\DSSAT4\Maize
system('..\DSCSM040.EXE B D4Batch.DV4');
system('copy Output\summary_output.txt+summary.out Output\summaryoutput.txt');
system('copy Output\PlantN output.txt+PlantN.out Output\PlantN output.txt');
system('copy Output\SoilN_output.txt+SoilN.out Output\SoilN_output.txt');
end
ParaSet=[Para3Zero,ParalZero,Para2Zero];

fidnew=fopen('C:\MATLAB7\work\Soil\Paraset.txt','w+');
for i=:Num*100
fprintf(fidnew,'% 10.9f10.9f %10.9f %10.9f %10.9f%10.9f10.9f %10.9f%10.9f\n',...
ParaSet(i,1),ParaSet(i,2),ParaSet(i,3),ParaSet(i,4),ParaSet(i,5),ParaSet(i,6),...
ParaSet(i,7),ParaSet(i,8),ParaSet(i,9));
end










fclose(fid_new);%%Make a txt file for the generated parameter sets


C.2 Generation of Random Numbers

% Random number generator. This is the function to generate the random numbers that follow an assumed
% multivariate normal distribution.

% n='Please input the lines of random numbers';
% disp(")
% disp(n)
% n=input(N=');
% N=n

function RG(Num)
N=100*Num;


% %Prior
% A=[4561.71
% 2373.91
% 61.86
%0 0
%0 0
%0 0
%0 0
%0 0
%];
0/o


2373.91 61.86 0
9679.39 55.85 0
55.85 15.97 0 0
0 0.0364 -0.3392 -0.0030
0 -0.3392 132.3833
0 -0.0030 0.3141 0.0098
0 -0.0030 0.2355 0.0078
0 -0.0049 0.2587 0.0061


0 0 0 0;


0
0
-0.0030
0.3141
0.0078
0.0070
0.0046


0 0
0 0;
-0.0049;
0.2355 0.2587;
0.0061;
0.0046;
0.0088


% B=[225.10 763.60 41.20 0.460 73.000 0.252 0.130 0.380];

%%First Posterior


% A=[372.9828
% -229.3543
% -21.7743
%0 0
%0 0
%0 0
%0 0
%0 0
%];
%


-229.3543 -21.7743
2597.0686 -29.9287
-29.9287 11.4105
0 0.0256 -1.0992 0.0032
0 -1.0992 79.6101
0 0.0032 -0.0985 0.0016
0 0.0016 -0.0499 0.0015
0 -0.0035 0.1952 0.0003


%B=[59.7269 520.0481


36.3871


0 0 0 0 0;
0 0 0 0 0;
0 0 0 0 0;


0.0016
-0.0985
0.0015
0.0026
-0.0012


-0.0035;
-0.0499 0.1952;
0.0003;
-0.0012;
0.0045 ;


0.4441 71.6747


0.2321 0.1369 0.3806];


%%Second Posterior
A=[25.0172 -1.5125 -7.5566 0 0
-1.5125 1121.8859 -21.2470 0
-7.5566 -21.2470 8.9119 0 0
0 0 0 0.0205 -0.8059 0.0027
0 0 0 -0.8059 69.3364
0 0 0 0.0027 -0.0923 0.0011
0 0 0 0.0013 -0.0374 0.0008
0 0 0 -0.0021 0.0855 0.0003


0 0 0;
0 0 0 0;
0 0 0;


0.0013
-0.0923
0.0008
0.0012
-0.0005


-0.0021;
-0.0374 0.0855;
0.0003;
-0.0005;
0.0025;


0;










36.5520 0.4665 70.8025


R=mvnrnd(B,A,N);

%First Posterior
% min=0.8238;
% max=0.9797;
% SLPF=unifmd(min,max,N,1)

%Second Posterior
min=0.8239;
max=0.9791;
SLPF=unifmd(min,max,N,1);

format long
R=abs(R);
SLPF;

fidnew-fopen('C:\MATLAB7\work\Soil\Randoms.txt','w+');
fori=1:N
fprintf(fid new,'% 10.9f10.9f %10.9f%10.9f %10.9f%10.9f10.9f %10.9f %10.9f\n',...
R(i,1),R(i,2),R(i,3),R(i,4),R(i,5),R(i,6),R(i,7),R(i,8),SLPF(i));
end
fclose(fidnew);

C.3 Function "mvnrnd"

function r = mvnmd(mu,sigma,cases);
%MVNRND Random matrices from the multivariate normal distribution.
% R = MVNRND(MU,SIGMA,CASES) returns a matrix of random numbers
% chosen from the multivariate normal distribution with mean vector,
% MU, and covariance matrix, SIGMA. CASES is the number of rows in R.
%
% SIGMA is a square positive definite matrix with size equal to
% the length of MU

% B.A. Jones 7-6-94 % Copyright(c) 1993-95 by The MathWorks, Inc.


[ml nl] = size(mu);
c = max([ml nl]);
ifml .* nl = c
error('Mu must be a vector.');
end

[m n] = size(sigma);
if m n
error('Sigma must be square');
end

ifm c
error('The length of mu must equal the number of rows in sigma.');
end


B=[68.3377 524.2715


0.2299 0.1252 0.3868];











[T p] = chol(sigma);
ifp 0
error('Sigma must be a positive definite matrix.');
end


ifml == c
mu = mu';
end

mu = mu(ones(cases,1),:);

r = randn(cases,c) T + mu;

C.4 Parameter Setup for Genotype and Soil

%Processing the generated random numbers to genotype and soil

function [Paral,Para2,Para3]=ParaSetup(i)

B=load('C:\MATLAB7\work\Soil\Randoms.txt');

A=B((i-1)*100+1:i*100,:);

J=find(A(:,5)>=100);
A(J,5)=99;%%Change the runoff curve number that is greater than 100 to 99.

Paral=[A(:,4) A(:,5) A(:,9)];

ParaSLLL=[A(:,7)];
ParaSDUL=[A(:,6)];
ParaSSAT=[A(:,8)];

MeanSLLL=[0.12979 0.13355 0.13740 0.14300 0.14471];
StdevSLLL=[0.08797 0.08901 0.09332 0.10033 0.10152];
PertSLLL=(ParaSLLL-MeanSLLL(1))/StdevSLLL(1);

MeanSDUL=[0.25205 0.25550 0.25824 0.25836 0.25716];
StdevSDUL=[0.10966 0.11195 0.11560 0.12133 0.12151];
PertSDUL=(ParaSDUL-MeanSDUL(1))/StdevSDUL(1);

MeanSSAT=[0.37986 0.37619 0.37831 0.37816 0.37383];
StdevSSAT=[0.09691 0.09334 0.09803 0.09329 0.09050];
PertSSAT=(ParaSSAT-MeanSSAT(1))/StdevSSAT(1);

for i=1:100
forj=l:5
Para2SLLL(j,i)=MeanSLLL(j)+PertSLLL(i)* StdevSLLL(j);
Para2SDUL( ,i)=MeanSDUL(j)+PertSDUL(i)* StdevSDUL(j);
Para2S SAT (,i)=MeanSSAT(j)+PertS SAT(i)* StdevSSAT(j);

if Para2SLLL(,i)==0










Para2SLLL(j,i)=Para2SLLL(j,i)+0.001;
Para2SDUL(j,i)=Para2SDUL(j,i)+0.002;
Para2SSAT(j,i)=Para2SSAT(j,i)+0.003;
end

if Para2SLLL(,i)==Para2SDUL(j,i)
Para2SDUL(j,i)=Para2SDUL(j,i)+0.001;
Para2SSAT(j,i)=Para2SSAT(j,i)+0.002;
end

if Para2SDUL(j,i)==Para2SSAT(j,i)
Para2SSAT(j,i)=Para2SSAT(j,i)+0.001;
end
end
end

Para2=ones(5,3,100);

fori=1:100
Para2Temp(:,:,i)=[Para2SLLL(:,i) Para2SDUL(:,i) Para2SSAT(:,i)];
Para2(:,:,i)=abs(Para2Temp(:,:,i));
B=sort(Para2(:,:,i),2);
Para2(:,:,i)=B;
end

%%Genotype
I=find(A(:,2)>=1000);%%Change the P5 value that is greater than 1000 to 990.
A(I,2)=990;
Para3=[A(:,l) A(:,2) A(:,3)];

save parameter Paral Para2 Para3;

C.5 Change of Soil File

%%Modify Soil File

function SoilChange
load parameter;
LineParal=60:13:1347;
LinePara2=LineParal+2;
fid_soil=fopen('soilPlate.sol','r');
fidnew=fopen('C:\DSSAT4\Soil\soil.sol','w+');
k=l;
for i=1:2000
line=fgetl(fid_soil);
if ~ischar(line), break, end

%%For SLDR SLRO and SLPF
if k<=100 & i==LineParal(k)
line =line(1:20);
line2=line(21:24);
line3=line(25:26);
line4=line(27:30);










line5=line(31:38);
line6=line(39:42);
line7=line(43 :end);
line2str=num2str(Paral(k,1),'%4.2f);
line4str=num2str(Paral (k,2),'%4. If);
line6str=num2str(Paral (k,3),'%4.2f);
NewLine=[linel line2str line3 line4str line5 line6str line7;
fprintf(fidnew,'%s\n',NewLine);


%%For Layer 1 of SLLL SDUL and SSAT
elseif k<=100 & i==LinePara2(k)
linel=line(1:13);
line2=line(14:18);
line3=line(19);
line4=line(20:24);
line5=line(25);
line6=line(26:30);
line7=line(31 :end);

line2str=-num2str(Para2(1,1,k)+0.001,'%5.3f);
line4str=num2str(Para2(1,2,k)+0.002,'%5.3f);
line6str=-num2str(Para2(1,3,k)+0.003,'%5.3f);
NewLine=[linel line2str line3 line4str line5 line6str line7;
fprintf(fidnew,'%s\n',NewLine);

%%For Layer 2 of SLLL SDUL and SSAT
elseif k<=100 & i==LinePara2(k)+l
linel=line(1:13);
line2=line(14:18);
line3=line(19);
line4=line(20:24);
line5=line(25);
line6=line(26:30);
line7=line(31 :end);

line2str=num2str(Para2(2,1,k)+0.001,'%5.3f);
line4str=num2str(Para2(2,2,k)+0.002,'%5.3f);
line6str=num2str(Para2(2,3,k)+0.003,'%5.3f);
NewLine=[linel line2str line3 line4str line5 line6str line7;
fprintf(fidnew,'%s\n',NewLine);

%%For Layer 3 of SLLL SDUL and SSAT
elseif k<=100 & i==LinePara2(k)+2
linel=line(1:13);
line2=line(14:18);
line3=line(19);
line4=line(20:24);
line5=line(25);
line6=line(26:30);
line7=line(31 :end);

line2str=-num2str(Para2(3,1,k)+0.001,'%5.3f);










line4str=num2str(Para2(3,2,k)+0.002,'%5.3f);
line6str=num2str(Para2(3,3,k)+0.003,'%5.3f);
NewLine=[linel line2str line3 line4str line5 line6str line7;
fprintf(fidnew,'%s\n',NewLine);

%%For Layer 4 of SLLL SDUL and SSAT
elseif k<=100 & i==LinePara2(k)+3
linel=line(1:13);
line2=line(14:18);
line3=line(19);
line4=line(20:24);
line5=line(25);
line6=line(26:30);
line7=line(31 :end);

line2str=-num2str(Para2(4,1,k)+0.001,'%5.3f);
line4str=num2str(Para2(4,2,k)+0.002,'%5.3f);
line6str=num2str(Para2(4,3,k)+0.003,'%5.3f);
NewLine=[linel line2str line3 line4str line5 line6str line7;
fprintf(fidnew,'%s\n',NewLine);

%%For Layer 5 of SLLL SDUL and SSAT
elseif k<=100 & i==LinePara2(k)+4
linel=line(1:13);
line2=line(14:18);
line3=line(19);
line4=line(20:24);
line5=line(25);
line6=line(26:30);
line7=line(31 :end);

line2str=num2str(Para2(5,1,k)+0.001,'%5.3f);
line4str=num2str(Para2(5,2,k)+0.002,'%5.3f);
line6str=num2str(Para2(5,3,k)+0.003,'%5.3f);
NewLine=[linel line2str line3 line4str line line6str line7;
fprintf(fidnew,'%s\n',NewLine);

k=k+1;
else
fprintf(fidnew,'%s\n',line);
end
end

fclose(fid_soil);
fclose(fidnew);

C.6 Change of Genotype File

%%Modify Genotype File

function GenoChange
load parameter;
LineParal=47:1:146;










fidMZCER040=fopen('MZCER040_Plate.cul','r');
fidnew=fopen('C:\DSSAT4\Genotype\MZCERO40.cul','w+');
k=l;
fori=1:200
line=fgetl(fidMZCER040);
if ~ischar(line), break, end

%%For P1, P5 and PHIN
if k<=100 & i==LineParal(k)
linel=line(1:31);
line2=line(32:36);
line3=line(37:44);
line4=line(45:49);
line5=line(50:62);
line6=line(63:66);
line7=line(67:end);
line2str=num2str(Para3(k,1),'%5.1f);
line4str=num2str(Para3(k,2),'%5. 1f);
line6str=num2str(Para3(k,3),'%4.1f);
NewLine=[linel line2str line3 line4str line5 line6str line7;
fprintf(fidnew,'%s\n',NewLine);

k=k+l;%%/Next Line
else
fprintf(fidnew,'%s\n',line);
end
end

fclose(fidMZCER040);
fclose(fidnew);

C.7 Summary Output Processing

%%Processing the GLUE simulation results, selecting the data needed

N=10400;

%%%%To get the values of anthesis days, maturity days, yield and nitrogen
%%%%leaching.

fid_Summary=fopen('C:\DSSAT4\Maize\Output\Summary_Output.txt','r');
fidnew=fopen('C:\DSSAT4\Maize\Output\Summary.txt','w+');

k=1;%%Subscriot of Line Matrix

fori=l:(N+5)
line=fgetl(fidSummary);
if ~ischar(line), break, end

if length(line)==283
if line(67:73)=='2005068'
linel=line(1:66);
line2=line(67:73);










line3=line(74);
line4=line(75:81);
line5=line(82);
line6=line(83:89);
line7=line(90:110);
line8=line(111:115);
line9=line(116:219);
line10=line(220:222);
line 11 =ine(223: end);

PDAT=line2;%%Planting date
ADAT=num2str(line4);%%Anthesis date
MDAT=line6;%%Maturity date
HWAH=line8;%%Yield
NLCM=line10;%%/Nitrogen leaching

NewLine=[PDAT,' ',ADAT,' ',MDAT,' ',HWAH,' ',NLCM];
fprintf(fidnew,'%s\n',NewLine);
end
end
end

fclose(fid_Summary);
fclose(fidnew);

C.8 Plant Nitrogen Output Processing

%%Processing the GLUE simulation results, selecting the data needed

N=1300000;

%%%%To get the values of anthesis days, maturity days, yield and nitrogen
%%%%leaching.

fidPlantN=fopen('C:\DSSAT4\Maize\Output\PlantN Output.txt','r');

LN90=zeros(0,1);
LN97=zeros(0,1);
LN112=zeros(0,1);
LN126=zeros(0,1);
LN136=zeros(0,1);
LN140=zeros(0,1);
LN147=zeros(0,1);
LN153=zeros(0,1);%%Empty matrix for leaf nitogen concentration at different days


fori=l:(N+5)
line-fgetl(fid_PlantN);
if ~ischar(line), break, end

if length(line)== 114
if line(2:5)=='2005'










linel=line(1);
line2=line(2:5);
line3=line(6);
line4=line(7:9);
line5=line(10:87);
line6=line(88:91);
line7=line(91:end);

YearStr=line2;%%Planting year as string
DayStr=line4;%%Sampling day as string
LNStr=line6;%%Maturity date as string

Yea=str2num(YearStr);%%Planting year as number
Day=str2num(DayStr);%%Sampling day as number
LN=str2num(LNStr);%%Leaf nitrogen concentration

if Day==90
LN90=[LN90;LN];
else if Day==97
LN97=[LN97;LN];
else if Day==112
LN112=[LN112;LN];
else if Day==126
LN126=[LN126;LN];
else if Day==136
LN136=[LN136;LN];
else if Day==140
LN140=[LN140;LN];
else if Day==147
LN147=[LN147;LN];
else if Day==153
LN153=[LN153;LN];
end
end
end
end
end
end
end
end

end

end
end

% fidnew=fopen('C:\DSSAT4\Maize\Output\PlantN.txt','w+');
fidnew l-fopen('C:\DSSAT4\Maize\Output\PlantN90.txt','w+');
fidnew2=fopen('C:\DSSAT4\Maize\Output\PlantN97.txt','w+');
fidnew3=fopen('C:\DSSAT4\Maize\Output\PlantN112.txt','w+');
fidnew4=fopen('C:\DSSAT4\Maize\Output\PlantN126.txt','w+');
fidnew5=fopen('C:\DSSAT4\Maize\Output\PlantN136.txt','w+');
fidnew6=fopen('C:\DSSAT4\Maize\Output\PlantN140.txt','w+');










fidnew7=fopen('C:\DSSAT4\Maize\Output\PlantN147.txt','w+');
fidnew8=fopen('C:\DSSAT4\Maize\Output\PlantN153.txt','w+');

for i=1:length(LN90)
% fprintf(fidnew,'%4. If %4. If %4. If %4. If %4. If %4. If %4. If %4. lfn',...
% LN90,LN97,LN112,LN126,LN136,LN140,LN147,LN153);
fprintf(fidnew 1,'%4.1f\n',LN90(i,1));
end

for i=1:length(LN97)
fprintf(fidnew2,'%4. f\n',LN97(i,1));
end

for i=l:length(LN112)
fprintf(fidnew3,'%4. f\n',LN112(i,l));
end

for i=l:length(LN126)
fprintf(fidnew4,'%4.1f\n',LN126(i,l));
end

for i=l:length(LN136)
fprintf(fidnew5,'%4.1f\n',LN136(i,l));
end

for i=l:length(LN140)
fprintf(fidnew6,'%4.1f\n',LN140(i,l));
end

for i=l:length(LN147)
fprintf(fidnew7,'%4.1f\n',LN147(i,l));
end

for i=l:length(LN153)
fprintf(fidnew8,'%4.1f\n',LN153(i,l));
end

fclose(fidnewl);
fclose(fidnew2);
fclose(fidnew3);
fclose(fidnew4);
fclose(fidnew5);
fclose(fidnew6);
fclose(fidnew7);
fclose(fidnew8);
%%Make a txt file for the generated parameter sets


fclose(fidPlantN);
% fclose(fid PlantN);










C.9 Soil Nitrogen Output Processing


%%Processing the GLUE simulation results, selecting the data needed

N=1300000;

%%%%To get the values of anthesis days, maturity days, yield and nitrogen
%%%%leaching.

fid_SoilN=fopen('C:\DSSAT4\Maize\Output\SoilN_Output.txt','r');

SN90=zeros(0,4);
SN97=zeros(0,4);
SN 112=zeros(0,4);
SN126=zeros(0,4);
SN136=zeros(0,4);
SN140=zeros(0,4);
SN147=zeros(0,4);
SN153=zeros(0,4);%%Empty matrix for soil nitogen concentration at different days in four layers


fori=1:N
line-fgetl(fid_SoilN);
if ~ischar(line), break, end

if length(line)==224
if line(2:5)=='2005'
line =line(1:6);
line2=line(7:9);
line3=line(10:70);
line4=line(71:76);
line5=line(77:82);
line6=line(83:88);
line7=line(89:94);
line8=line(95:100);
line9=line(101:end);

DayStr=line2;%%Sampling day as string
L1Str=line4;%%Soil NItrogen in Layer 1 as string
L2Str=line5;%%Soil NItrogen in Layer 2 as string
L3Str=line6;%%Soil NItrogen in Layer 3 as string
L4Str=line8;%%Soil NItrogen in Layer 4 as string

Day=str2num(DayStr);%%Sampling day as number
L1=str2num(L 1Str);%%Soil Nitrogen in Layer 1 as number
L2=str2num(L2Str);%%Soil NItrogen in Layer 2 as number
L3=str2num(L3 Str);%%Soil NItrogen in Layer 3 as number
L4=str2num(L4Str);%%Soil NItrogen in Layer 4 as number
SN=[L1,L2,L3,L4];

if Day==90
SN90=[SN90;SN];
else if Day==97










SN97=[SN97;SN];
else if Day==112
SN112=[SN112;SN];
else if Day==126
SN126=[SN126;SN];
else if Day==136
SN136=[SN136;SN];
else if Day==140
SN140=[SN140;SN];
else if Day==147
SN147=[SN147;SN];
else if Day==153
SN153=[SN153;SN];
end
end
end
end
end
end
end
end

end

end
end

fidnew l=fopen('C:\DSSAT4\Maize\Output\SoilN90.txt','w+');
fidnew2=fopen('C:\DSSAT4\Maize\Output\SoilN97.txt','w+');
fidnew3-fopen('C:\DSSAT4\Maize\Output\SoilN 112.txt','w+');
fidnew4=fopen('C:\DSSAT4\Maize\Output\SoilN126.txt','w+');
fidnew5-fopen('C:\DSSAT4\Maize\Output\SoilN136.txt','w+');
fidnew6=fopen('C:\DSSAT4\Maize\Output\SoilN140.txt','w+');
fidnew7=fopen('C:\DSSAT4\Maize\Output\SoilN147.txt','w+');
fidnew8=fopen('C:\DSSAT4\Maize\Output\SoilN153.txt','w+');

for i=l:length(SN90)
fprintf(fidnewl,'%5. If %5. If %5. If %5. lf\n',SN90(i,:));
end

for i=l:length(SN97)
fprintf(fidnew2,'%5. If %5. If %5. If %5. lf\n',SN97(i,:));
end

for i=l:length(SN112)
fprintf(fidnew3,'%5. If %5. If %5. If %5. lf\n',SN112(i,:));
end

for i=l:length(SN126)
fprintf(fidnew4,'%5. If %5. If %5. If %5. lf\n',SN126(i,:));
end

for i=l:length(SN136)










fprintf(fidnew5,'%5. If %5. If %5. If %5. If\n',SN136(i,:));
end

for i=l:length(SN140)
fprintf(fidnew6,'%5. If %5. If %5. If %5. lf\n',SN140(i,:));
end

for i=l:length(SN147)
fprintf(fidnew7,'%5. If %5. If %5. If %5. lf\n',SN147(i,:));
end

for i=l:length(SN153)
fprintf(fidnew8,'%5. If %5. If %5. If %5. lf\n',SN153(i,:));
end

fclose(fidnewl);
fclose(fidnew2);
fclose(fidnew3);
fclose(fidnew4);
fclose(fidnew5);
fclose(fidnew6);
fclose(fidnew7);
fclose(fidnew8);
%%Make a txt file for the soil nitrogen in 4 different layer on different
%%days

fclose(fid_SoilN);

C.10 Parameter PDF Plot

%%%%%%%%Plot histograms for the selected parameters

%%%%%%%%%%%Histograms of P1%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
clear
A=load('HistogramPrior.txt');
B=load('HistogramFirstPosterior.txt');
C=load('Histogram_SecondPosterior.txt');

figure (1)
break_point=[0:25:500];
l=length(break_point);

subplot(3,1,1);%%Histogram under prior distribution
[n,x]=hist(A(:,1),20);
bar(x,n/10000);

h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of P ');
ylabel('Probability');
title('\fontsize{ 11 }PDF of P1 under prior distribution')%%PDF of P1.
axis([break_point(l) break_point(l) 0 1.0]);











subplot(3,1,2);%%Histogram under first posterior distribution
x=B(:,l);
n=B(:,10);

m=length(n);

a=(break_point(l)+break_point(2))/2;
%%l=length(break_point);
b=break_point(2)-break_point( 1);
c=(break_point(1- 1)+break_point(1))/2;

middle_point=[a:b:c];

for i= 1 :length(break_point)-1
I-find(x>=break_point(i) & x<=break_point(i+l));
Pp_group=n(I);
P_group(i)=sum(Pp_group);
end

bar(middle_point,P_group, 1 ,'b');
h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of P1');
ylabel('Probability');
title('\fontsize{ 11 }PDF of P1 under first posterior distribution')%%PDF of P1.
axis([break_point(1) break_point(1) 0 1.0]);

subplot(3,1,3);%%Histogram under second posterior distribution
x=C(:,l);
n=C(:,10);

m=length(n);

a=(break_point(l)+break_point(2))/2;
%%l=length(break_point);
b=break_point(2)-break_point( 1);
c=(break_point(1- 1)+break_point(1))/2;

middle_point=[a:b:c];

for i= 1 :length(break_point)-1
I-find(x>=break_point(i) & x<=break_point(i+l));
Ppgroup=n(I);
P_group(i)=sum(Pp_group);
end

bar(middle_point,P_group, 1 ,'b');
h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of P1');
ylabel('Probability');
title('\fontsize{ 11 }PDF of P1 under second posterior distribution')%%PDF of P1.










axis([break_point(1) break_point(l) 0 1.0]);


%%%%%%%%%%%Histograms of P5%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
clear
A=load('HistogramPrior.txt');
B=load('HistogramFirstPosterior.txt');
C=load('Histogram_SecondPosterior.txt');

figure (2)
breakpoint=[0:50:1000];
l=length(break_point);

subplot(3,1,1);%%Histogram under prior distribution
[n,x]=hist(A(:,2),20);
bar(x,n/10000);

h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of P5');
ylabel('Probability');
title('\fontsize{ 11 }PDF of P5 under prior distribution')%%PDF of P5.
axis([break_point(1) break_point(l) 0 1.0]);

subplot(3,1,2);%%Histogram under first posterior distribution
x=B(:,2);
n=B(:,10);

m=length(n);

a=(break_point(1)+breakpoint(2))/2;
%%l=length(break_point);
b=breakpoint(2)-breakpoint(1);
c=(breakpoint(1- 1)+break_point(1))/2;

middlepoint=[a:b:c];

for i= 1 :length(break_point)-1
I-find(x>=breakpoint(i) & x<=break_point(i+l));
Pp_group=n(I);
P_group(i)=sum(Pp_group);
end

bar(middlepoint,P_group, 1 ,'b');
h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of P5');
ylabel('Probability');
title('\fontsize{ 11 }PDF of P5 under first posterior distribution')%%PDF of P5.
axis([break_point(1) break_point(1) 0 1.0]);

subplot(3,1,3);%%Histogram under second posterior distribution










x=C(:,2);
n=C(:,10);

m=length(n);

a=(break_point(1)+break_point(2))/2;
%%l=length(break_point);
b=break_point(2)-break_point( 1);
c=(breakpoint(1-1 )+break_point(1))/2;

middle_point=[a:b:c];

for i= 1 :length(break_point)-1
I=find(x>=breakpoint(i) & x<=break_point(i+l));
Pp_group=n(I);
P_group(i)=sum(Pp_group);
end

bar(middle_point,P_group, 1 ,'b');
h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of P5');
ylabel('Probability');
title('\fontsize{ 11 }PDF of P5 under second posterior distribution')%%PDF of P5.
axis([break_point(1) break_point(1) 0 1.0]);


%%%%%%%%%%%Histograms of PHIN%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
clear
A=load('HistogramPrior.txt');
B=load('HistogramFirstPosterior.txt');
C=load('Histogram_SecondPosterior.txt');

figure (3)
break_point=[0:3:60];
l=length(break_point);

subplot(3,1,1);%%Histogram under prior distribution
[n,x]=hist(A(:,3),20);
bar(x,n/10000);

h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of PHIN');
ylabel('Probability');
title('\fontsize{ 11 }PDF of PHIN under prior distribution')%%PDF of PHIN.
axis([break_point(1) break_point(1) 0 1.0]);

subplot(3,1,2);%%Histogram under first posterior distribution
x=B(:,3);
n=B(:,10);










m=length(n);


a=(break_point(l)+break_point(2))/2;
%%l=length(break_point);
b=break_point(2)-break_point( 1);
c=(breakpoint(1-1 )+break_point(1))/2;

middle_point=[a:b:c];

for i= 1 :length(break_point)-1
I-find(x>=breakpoint(i) & x<=break_point(i+l));
Pp_group=n(I);
P_group(i)=sum(Pp_group);
end

bar(middle_point,P_group, 1 ,'b');
h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of PHIN');
ylabel('Probability');
title('\fontsize{ 11 }PDF of PHIN under first posterior distribution')%%PDF of PHIN.
axis([break_point(1) break_point(1) 0 1.0]);

subplot(3,1,3);%%Histogram under second posterior distribution
x=C(:,3);
n=C(:,10);

m=length(n);

a=(break_point(l)+break_point(2))/2;
%%l=length(break_point);
b=break_point(2)-break_point( 1);
c=(break_point(1- 1)+break_point(1))/2;

middle_point=[a:b:c];

for i= 1 :length(break_point)-1
I-find(x>=breakpoint(i) & x<=break_point(i+l));
Ppgroup=n(I);
P_group(i)=sum(Pp_group);
end

bar(middle_point,P_group, 1 ,'b');
h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of PHIN');
ylabel('Probability');
title('\fontsize{ 11 }PDF of PHIN under second posterior distribution')%%PDF of PHIN.
axis([break_point(1) break_point(1) 0 1.0]);


%%%%%%%%%%%Histograms of SLDRo%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%










clear
A=load('HistogramPrior.txt');
B=load('HistogramFirstPosterior.txt');
C=load('Histogram_SecondPosterior.txt');

figure (4)
break_point=[0:0.05:1];
l=length(break_point);

subplot(3,1,1);%%Histogram under prior distribution
[n,x]=hist(A(:,4),20);
bar(x,n/10000);

h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of SLDR');
ylabel('Probability');
title('\fontsize{ 11 }PDF of SLDR under prior distribution')%%PDF of SLDR.
axis([break_point(1) break_point(l) 0 1.0]);

subplot(3,1,2);%%Histogram under first posterior distribution
x=B(:,4);
n=B(:,10);

m=length(n);

a=(break_point(1)+break_point(2))/2;
%%l=length(break_point);
b=break_point(2)-break_point( 1);
c=(break_point(1- 1)+break_point(1))/2;

middle_point=[a:b:c];

for i= 1 :length(break_point)-1
I-find(x>=break_point(i) & x<=break_point(i+l));
Pp_group=n(I);
P_group(i)=sum(Pp_group);
end

bar(middle_point,P_group, 1 ,'b');
h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of SLDR');
ylabel('Probability');
title('\fontsize{ 11 }PDF of SLDR under first posterior distribution')%%PDF of SLDR.
axis([break_point(1) break_point(1) 0 1.0]);

subplot(3,1,3);%%Histogram under second posterior distribution
x=C(:,4);
n=C(:,10);

m=length(n);










a=(breakpoint(1)+breakpoint(2))/2;
%%l=length(break_point);
b=breakpoint(2)-breakpoint(1);
c=(breakpoint(1-1 )+breakpoint(l))/2;

middlepoint=[a:b:c];

for i= 1 :length(breakpoint)-1
I-find(x>=breakpoint(i) & x<=break_point(i+l));
Pp_group=n(I);
P_group(i)=sum(Pp_group);
end

bar(middle_point,P_group, 1 ,'b');
h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of SLDR');
ylabel('Probability');
title('\fontsize{ 11 }PDF of SLDR under second posterior distribution')%%PDF of SLDR.
axis([breakpoint(1) breakpoint(l) 0 1.0]);


%%%%%%%%%%%Histograms of SLRO%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
clear
A=load('HistogramPrior.txt');
B=load('HistogramFirstPosterior.txt');
C=load('Histogram_SecondPosterior.txt');

figure (5)
breakpoint=[0:5:100];
l=length(breakpoint);

subplot(3,1,1);%%Histogram under prior distribution
[n,x]=hist(A(:,5),20);
bar(x,n/10000);

h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of SLRO');
ylabel('Probability');
title('\fontsize{ 11 }PDF of SLRO under prior distribution')%%PDF of SLRO.
axis([breakpoint(1) breakpoint(l) 0 1.0]);

subplot(3,1,2);%%Histogram under first posterior distribution
x=B(:,5);
n=B(:,10);

m=length(n);

a=(breakpoint(1)+breakpoint(2))/2;
%%l=length(break_point);
b=breakpoint(2)-breakpoint(1);










c=(break_point(1- 1)+break_point(1))/2;


middle_point=[a:b:c];

for i= 1 :length(break_point)-1
I-find(x>=break_point(i) & x<=break_point(i+l));
Pp_group=n(I);
P_group(i)=sum(Pp_group);
end

bar(middle_point,P_group, 1 ,'b');
h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of SLRO');
ylabel('Probability');
title('\fontsize{ 11 }PDF of SLRO under first posterior distribution')%%PDF of SLRO.
axis([break_point(1) break_point(1) 0 1.0]);

subplot(3,1,3);%%Histogram under second posterior distribution
x=C(:,5);
n=C(:,10);

m=length(n);

a=(break_point(1)+break_point(2))/2;
%%l=length(break_point);
b=break_point(2)-break_point( 1);
c=(break_point(1- 1)+break_point(1))/2;

middle_point=[a:b:c];

for i= 1 :length(break_point)-1
I-find(x>=breakpoint(i) & x<=break_point(i+l));
Ppgroup=n(I);
P_group(i)=sum(Pp_group);
end

bar(middle_point,P_group, 1 ,'b');
h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of SLRO');
ylabel('Probability');
title('\fontsize{ 11 }PDF of SLRO under second posterior distribution')%%PDF of SLRO.
axis([break_point(1) break_point(1) 0 1.0]);


%%%%%%%%%%%Histograms of SDUL%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
clear
A=load('HistogramPrior.txt');
B=load('HistogramFirstPosterior.txt');
C=load('Histogram_SecondPosterior.txt');










figure (6)
break_point=[0:0.025:0.5];
l=length(break_point);

subplot(3,1,1);%%Histogram under prior distribution
[n,x]=hist(A(:,6),20);
bar(x,n/10000);

h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of SDUL');
ylabel('Probability');
title('\fontsize{ 11 }PDF of SDUL under prior distribution')%%PDF of SDUL.
axis([break_point(1) break_point(1) 0 1.0]);

subplot(3,1,2);%%Histogram under first posterior distribution
x=B(:,6);
n=B(:,10);

m=length(n);

a=(break_point(1)+break_point(2))/2;
%%l=length(break_point);
b=break_point(2)-break_point(1);
c=(break_point(1- 1)+break_point(1))/2;

middle_point=[a:b:c];

for i= 1 :length(break_point)-1
I-find(x>=break_point(i) & x<=break_point(i+l));
Ppgroup=n(I);
P_group(i)=sum(Pp_group);
end

bar(middle_point,P_group, 1 ,'b');
h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of SDUL');
ylabel('Probability');
title('\fontsize{ 11 }PDF of SDUL under first posterior distribution')%%PDF of SDUL.
axis([break_point(1) break_point(1) 0 1.0]);

subplot(3,1,3);%%Histogram under second posterior distribution
x=C(:,6);
n=C(:,10);

m=length(n);

a=(break_point(1)+break_point(2))/2;
%%l=length(break_point);
b=break_point(2)-break_point(1);
c=(break_point(1- 1)+break_point(1))/2;










middlepoint=[a:b:c];


for i= 1 :length(break_point)-1
I-find(x>=breakpoint(i) & x<=break_point(i+l));
Pp_group=n(I);
P_group(i)=sum(Pp_group);
end

bar(middle_point,P_group, 1 ,'b');
h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of SDUL');
ylabel('Probability');
title('\fontsize{ 11 }PDF of SDUL under second posterior distribution')%%PDF of SDUL.
axis([break_point(1) break_point(1) 0 1.0]);


%%%%%%%%%%%Histograms of SLLL%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
clear
A=load('HistogramPrior.txt');
B=load('HistogramFirstPosterior.txt');
C=load('Histogram_SecondPosterior.txt');

figure (7)
break_point=[0:0.025:0.5];
l=length(break_point);

subplot(3,1,1);%%Histogram under prior distribution
[n,x]=hist(A(:,7),20);
bar(x,n/10000);

h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of SLLL');
ylabel('Probability');
title('\fontsize{ 11 }PDF of SLLL under prior distribution')%%PDF of SLLL.
axis([break_point(1) break_point(l) 0 1.0]);

subplot(3,1,2);%%Histogram under first posterior distribution
x=B(:,7);
n=B(:,10);

m=length(n);

a=(break_point(1)+break_point(2))/2;
%%l=length(break_point);
b=break_point(2)-break_point( 1);
c=(break_point(1- 1)+break_point(1))/2;

middle_point=[a:b:c];

for i= 1 :length(break_point)-1










I-find(x>=break_point(i) & x<=break_point(i+l));
Pp_group=n(I);
P_group(i)=sum(Pp_group);
end

bar(middle_point,P_group, 1 ,'b');
h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of SLLL');
ylabel('Probability');
title('\fontsize{ 11 }PDF of SLLL under first posterior distribution')%%PDF of SLLL.
axis([break_point(1) break_point(1) 0 1.0]);

subplot(3,1,3);%%Histogram under second posterior distribution
x=C(:,7);
n=C(:,10);

m=length(n);

a=(breakpoint(1)+breakpoint(2))/2;
%%l=length(break_point);
b=breakpoint(2)-breakpoint(1);
c=(breakpoint(1- 1)+breakpoint(l))/2;

middlepoint=[a:b:c];

for i= 1 :length(breakpoint)-1
I-find(x>=breakpoint(i) & x<=break_point(i+l));
Ppgroup=n(I);
P_group(i)=sum(Pp_group);
end

bar(middle_point,P_group, 1 ,'b');
h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of SLLL');
ylabel('Probability');
title('\fontsize{ 11 }PDF of SLLL under second posterior distribution')%%PDF of SLLL.
axis([breakpoint(1) breakpoint(l) 0 1.0]);


%%%%%%%%%%%Histograms of SSAT%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
clear
A=load('HistogramPrior.txt');
B=load('HistogramFirstPosterior.txt');
C=load('Histogram_SecondPosterior.txt');

figure (8)
breakpoint=[0.2:0.025:0.7];
l=length(break_point);

subplot(3,1,1);%%Histogram under prior distribution










[n,x]=hist(A(:,8),20);
bar(x,n/10000);

h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel(Number of SSAT');
ylabel('Probability');
title('\fontsize{ 11 }PDF of SSAT under prior distribution')%%PDF of SSAT.
axis([break_point(1) break_point(1) 0 1.0]);

subplot(3,1,2);%%Histogram under first posterior distribution
x=B(:,8);
n=B(:,10);

m=length(n);

a=(break_point(l)+break_point(2))/2;
%%l=length(break_point);
b=break_point(2)-break_point( 1);
c=(break_point(1- 1)+break_point(1))/2;

middle_point=[a:b:c];

for i= 1 :length(break_point)-1
I-find(x>=break_point(i) & x<=break_point(i+l));
Ppgroup=n(I);
P_group(i)=sum(Pp_group);
end

bar(middle_point,P_group, 1 ,'b');
h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel(Number of SSAT');
ylabel('Probability');
title('\fontsize{ 11 }PDF of SSAT under first posterior distribution')%%PDF of SSAT.
axis([break_point(1) break_point(1) 0 1.0]);

subplot(3,1,3);%%Histogram under second posterior distribution
x=C(:,8);
n=C(:,10);

m=length(n);

a=(break_point(l)+break_point(2))/2;
%%l=length(break_point);
b=break_point(2)-break_point( 1);
c=(break_point(1- 1)+break_point(1))/2;

middle_point=[a:b:c];

for i= 1 :length(break_point)-1
I-find(x>=break_point(i) & x<=break_point(i+l));
Ppgroup=n(I);










P_group(i)=sum(Pp_group);
end

bar(middle_point,P_group, 1 ,'b');
h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of SSAT');
ylabel('Probability');
title('\fontsize{ 11 }PDF of SSAT under second posterior distribution')%%PDF of SSAT.
axis([break_point(1) break_point(1) 0 1.0]);


%%%%%%%%%%%Histograms of SLPF%%%%%%%%%%%%%%%%
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
clear
A=load('HistogramPrior.txt');
B=load('HistogramFirstPosterior.txt');
C=load('Histogram_SecondPosterior.txt');

figure (9)
break_point=[0.6:0.02:1.0];
l=length(break_point);

subplot(3,1,1);%%Histogram under prior distribution
[n,x]=hist(A(:,9),20);
bar(x,n/10000);

h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of SLPF');
ylabel('Probability');
title('\fontsize{ 11 }PDF of SLPF under prior distribution')%%PDF of SLPF.
axis([break_point(1) break_point(l) 0 1.0]);

subplot(3,1,2);%%Histogram under first posterior distribution
x=B(:,9);
n=B(:,10);

m=length(n);

a=(break_point(1)+break_point(2))/2;
%%l=length(break_point);
b=break_point(2)-break_point( 1);
c=(break_point(1- 1)+break_point(1))/2;

middle_point=[a:b:c];

for i= 1 :length(break_point)-1
I-find(x>=breakpoint(i) & x<=break_point(i+l));
Ppgroup=n(I);
P_group(i)=sum(Pp_group);
end










bar(middlepoint,P_group, 1 ,'b');
h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of SLPF');
ylabel('Probability');
title('\fontsize{ 11 }PDF of SLPF under first posterior distribution')%%PDF of SLPF.
axis([breakpoint(1) breakpoint(l) 0 1.0]);

subplot(3,1,3);%%Histogram under second posterior distribution
x=C(:,9);
n=C(:,10);

m=length(n);

a=(breakpoint(1)+breakpoint(2))/2;
%%l=length(break_point);
b=breakpoint(2)-breakpoint(1);
c=(breakpoint(1- 1)+breakpoint(l))/2;

middlepoint=[a:b:c];

for i= 1 :length(breakpoint)-1
I-find(x>=breakpoint(i) & x<=break_point(i+l));
Ppgroup=n(I);
P_group(i)=sum(Pp_group);
end

bar(middlepoint,P_group, 1 ,'b');
h=findobj (gca,'Type','patch');
set(h,'FaceColor','blue','EdgeColor','w')
xlabel('Number of SLPF');
ylabel('Probability');
title('\fontsize{ 11 }PDF of SLPF under second posterior distribution')%%PDF of SLPF.
axis([breakpoint(1) breakpoint(l) 0 1.0]);

C.11 3-D Plot of Joint Distribution of Yield and Nitrogen Leaching

clear;
%%Yield and nitrogen leaching joint distribution under prior distribution
%%of input parameters
A=load('Joint Prior.txt');

figure (1)

Yield=A(:,l);
NLCM=A(:,2);%%Get the data of yield and nitrogen leaching (NLCM)
NumData=length(Yield);%%Get the number of total data

breakpointl=[0:500:15000];
L =length(breakpointl);%%Set the grid vector for yield

breakpoint2=[0:10:300];
L2=length(breakpoint2);%%Set the grid vector for NLCM











al =(break_point (1)+breakpoint (2))/2;
b 1 =break_point (2)-break_point (1);
cl=(break_point (L-l1)+break_pointl(L1))/2;
middlepointl=[al:bl :cl];%%Set the middle point vector for yield

a2=(break_point2(1)+break_point2(2))/2;
b2=break_point2(2)-break_point2(1);
c2=(break_point2(L2-1)+break_point2(L2))/2;
middle_point2=[a2:b2:c2];%%Set the middle point vector for NLCM


I=zeros(0,1);

for i=l:Ll-l
Il=find(Yield>=break_pointl(i) & Yield
forj=I:L2-1
I2=find(NLCM>=break_point2(j) & NLCM
for m=l:length(I1)
for n= :length(I2)
if (Il(m)==I2(n))
I=[I;I2(n)];
end
end
end

count(j ,i)=length(I);
I=zeros(0,1);

end

end

middle_pointl=[al:bl:cl];
middle_point2=[a2:b2:c2];
count; %% Use these three things to plot the joint distribution between yield and nitrogen leaching

[X,Y]=meshgrid(middle_pointl,middle_point2);
Z=count/NumData;
surfc(X,Y,Z)

xlabel('Yield (kg/ha)');
ylabel(Nitrogen leaching (NLCM, kg/ha)');
Zlabel('Probability');
title('\fontsize{ 11 }Joint Distribution between yield and NLCM under prior distribution of parameter')%%PDF
of P1.
axis([0 15000 0 300 0 0.2]);

clear;
%%Yield and nitrogen leaching joint distribution under first posterior distribution
%%of input parameters










A=load('JointFirstPosterior.txt');


figure (2)

Yield=A(:,l);
NLCM=A(:,2);%%Get the data of yield and nitrogen leaching (NLCM)
NumData=length(Yield);%%Get the number of total data

breakpointl=[0:500:15000];
L =length(breakpointl);%%Set the grid vector for yield

breakpoint2=[0:10:300];
L2=length(breakpoint2);%%Set the grid vector for NLCM

al=(breakpointl (1)+breakpoint (2))/2;
b 1 =break_point (2)-break_point (1);
cl=(break_point 1 (L -1)+break_point (L 1))/2;
middlepointl=[al:bl :cl];%%Set the middle point vector for yield

a2=(breakpoint2(1)+breakpoint2(2))/2;
b2=breakpoint2(2)-breakpoint2(1);
c2=(breakpoint2(L2-1)+breakpoint2(L2))/2;
middlepoint2=[a2:b2:c2];%%Set the middle point vector for NLCM


I=zeros(0,1);

for i=l:Ll-l
Il=find(Yield>=breakpointl(i) & Yield
forj=I:L2-1
I2=find(NLCM>=breakpoint2(j) & NLCM
for m=l:length(I1)
for n= :length(I2)
if (Il(m)==I2(n))
I=[I;I2(n)];
end
end
end

count(j ,i)=length(I);
I=zeros(0,1);

end

end

middlepointl=[al:bl:cl];
middlepoint2=[a2:b2:c2];
count; %% Use these three things to plot the joint distribution between yield and nitrogen leaching

[X,Y]=meshgrid(middlepointl,middlepoint2);










Z=count/NumData;
surfc(X,Y,Z)

xlabel('Yield (kg/ha)');
ylabel(Nitrogen leaching (NLCM, kg/ha)');
Zlabel('Probability');
title('\fontsize{ 11 }Joint Distribution between yield and NLCM under first posterior distribution of
parameter')%%PDF of P1.
axis([0 15000 0 300 0 0.2]);

clear;
%%Yield and nitrogen leaching joint distribution under second posterior distribution
%%of input parameters

A=load('JointSecondPosterior.txt');

figure (3)

Yield=A(:,l);
NLCM=A(:,2);%%Get the data of yield and nitrogen leaching (NLCM)
NumData=length(Yield);%%Get the number of total data

breakpointl=[0:500:15000];
L =length(break_pointl);%%Set the grid vector for yield

break_point2=[0:10:300];
L2=length(break_point2);%%Set the grid vector for NLCM

al=(break_pointl (1)+break_point (2))/2;
b 1 =break_point (2)-break_point (1);
cl=(break_point 1 (L -1)+break_point (L 1))/2;
middlepointl=[al:bl :cl];%%Set the middle point vector for yield

a2=(break_point2(1)+break_point2(2))/2;
b2=break_point2(2)-break_point2(1);
c2=(break_point2(L2-1)+break_point2(L2))/2;
middle_point2=[a2:b2:c2];%%Set the middle point vector for NLCM


I=zeros(0,1);

for i=l:Ll-l
Il-find(Yield>=break_pointl(i) & Yield
forj=I:L2-1
I2=find(NLCM>=break_point2(j) & NLCM
for m=l:length(I1)
for n= 1:length(I2)
if (Il(m)==I2(n))
I=[I;I2(n)];
end
end












count(j ,i)=length(I);
I=zeros(0,1);

end

end

middlepointl=[al:bl:cl];
middlepoint2=[a2:b2:c2];
count; %% Use these three things to plot the joint distribution between yield and nitrogen leaching

[X,Y]=meshgrid(middlepointl,middlepoint2);
Z=count/Num Data;
surfc(X,Y,Z)

xlabel('Yield (kg/ha)');
ylabel(Nitrogen leaching (NLCM, kg/ha)');
Zlabel('Probability');
title('\fontsize{ 11 }Joint Distribution between yield and NLCM under second posterior distribution of
parameter')%%PDF of P1.
axis([0 15000 0 300 0 0.21);









APPENDIX D
PICTURES OF FIELD EXPERIMENT


Figure D-1. Components of nitrogen fertilizer solution


Figure D-2. Fertigation control table

































Figure D-3. Fertigation system installation


Figure D-4. Main fertigation lines, injection holes, peristaltic pump, and solution bucket

































Figure D-5. Sub-main fertigation lines


Figure D-6. Drip tapes and sub-main fertigation line



































Figure D-7. Drip tape distribution in one row


;.- r --- jl~'"~-';; -. -... ,.,,:: 4...i" .-T..,, ... :..
.~ ~~~~~";-- : '-- "'2. ,-':" ,., .':" .


..-.... ......- __ .. ; ...:..: .. ...... .- -,, ,..
.' .. ...: : .. ...,
~~LLL~:"" ~ i
,.L o ".
lh LIrd .. .. i -
-"J~ *e '" "Y ,,".i


Figure D-8. Irrigation with the linear move irrigation system


































Figure D-9. Sweet corn planting


Figure D-10. Sweet corn emergence





















&%L 10 -A A!b1^ A 0 &. ai AV 1W -c I-) -EL
Figure D-1 1. Comparison between no-nirogen-applied plot (near) and nitrogen-applied plot (far)


,- -- "--
W -FIRSO^, ism dv '-'r -


Figure D-12. Sweet corn tasseling


































Figure D-13. Sweet corn maturity


Figure D-14. Sweet corn harvest

































Figure D-15. Plant sampling


Figure D-16. Soil sampling










J



C

Atl






. st SW
*- bc 'rsrc.
t. '---.hi

C7?yAStz
ozIi






OWo~-"~~~;a zi..-


Figure D-17. Yield sampling


Figure D-18. Yield weighing









































Figure D-19. Yield grading


I-.-.


Figure D-20. Research partner










APPENDIX E
SAS CODE FOR ANOVA OF YIELD QUANTITY AND QUALITY

proc import
datafile='C:\Jianqiang He\PhD Study\PhD Research\SAS Data Analysis\Yield Analysis of Plots\Yields of
Plots in 2006.xls'
out=Yield DBMS=excel2000 REPLACE;
SHEET="Yield";
Getnames=yes;
run;
proc print data=Yield;
run;

proc anova data=Yield;
class Block I N;
model TotalYield= Block I Block*I N I*N;
test h=I e=Block*I;
MEANS I|N/DUNCAN;
run;

proc anova data=Yield;
class Block I N;
model MarketYield= Block I Block*I N I*N;
test h=I e=Block*I;
MEANS I|N/DUNCAN;
run;

proc anova data=Yield;
class Block I N;
model TotalEar= Block I Block*I N I*N;
test h=I e=Block*I;
MEANS I|N/DUNCAN;
run;

proc anova data=Yield;
class Block I N;
model US1= Block I Block*I N I*N;
test h=I e=Block*I;
MEANS I|N/DUNCAN;
run;

proc anova data=Yield;
class Block I N;
model US2= Block I Block*I N I*N;
test h=I e=Block*I;
MEANS I|N/DUNCAN;
run;

proc anova data=Yield;
class Block I N;










model Cull= Block I Block*I N I*N;
test h=I e=Block*I;
MEANS IIN/DUNCAN;
run;










APPENDIX F
NITRATE AND AMMONIUM CONCENTRATIONS IN MONITORING WELLS IN
BLOCK IN THE PLANT SCIENCE RESEARCH AND EDUCATION UNIT UNIVERSITY
OF FLORIDA


4/1/04 10/18/04 5/6/05 11/22/05 6/10/06 12/27/06 7/15/07
Date
--Average of East --Average of West


Figure F-1. Average nitrate concentration in the monitoring wells on the west part and east part
of Block 1



1
0.9


4/1/04 10/18/04 5/6/05 11/22/05 6/10/06 12/27/06 7/15/07
Date

--Average of East -u-Average of West

Figure F-2. Average ammonium concentration in the monitoring wells on the west part and east
part of Block 1


305











APPENDIX G
TOTAL KJELDAHL NITROGEN CONCENTRATION OF LEAVES AND STEMS OF
SWEET CORN IN FIELD EXPERIMENT IN PLOTS IN 2006


50


40


30


20






0
10


0 IIII


4/3/06 4/13/06 4/23/06


5/3/06 5/13/06 5/23/06
Date


6/2/06 6/12/06


F111 -*-F211 ---F311

Figure G-1. Average total Kjeldahl nitrogen (TKN) concentration of leaves of sweet corn under
irrigation level I1


50 1 1


40


30

-J
b 20
z
I-
10


0
I--



o


4/3/06 4/13/06 4/23/06


5/3/06 5/13/06 5/23/06
Date


6/2/06 6/12/06


F112 -- F212 --- F312

Figure G-2. Average total Kjeldahl nitrogen (TKN) concentration of leaves of sweet corn under
irrigation level 12


306















50


- 40

E
30

z
1 20
1-

10


0 i
4/3/06 4/13/06 4/23/06 5/3/06 5/13/06 5/23/06 6/2/06 6/12/06
Date

F111 ---F211 -u--F311


Figure G-3. Average total Kjeldahl nitrogen (TKN) concentration of stems of sweet corn under
irrigation level I1





60


50


S40


3 30

0
S20


10


n


4/3/06 4/13/06 4/23/06


5/3/06 5/13/06 5/23/06
Date


6/2/06 6/12/06


F112 -*- F212 --- F312


Figure G-4. Average total Kjeldahl nitrogen (TKN) concentration of stems of sweet corn under
irrigation level I2














NITRATE




20

18

16

J 14

12

10

8

06
z


APPENDIX H
AND AMMONIUM NITROGEN CONCENTRATION OF SOIL IN FIELD
EXPERIMENT OF SWEET CORN IN PLOTS IN 2006


3/4/06 3/14/06 3/24/06 4/3/06 4/13/06 4/23/06 5/3/06 5/13/06 5/23/06 6/2/06 6/12/06
Date


-*-F1I1 F211


F311 FOll


Figure H-1. Average nitrate nitrogen concentration of soil at layer 1 (0-15 cm) under irrigation
level II


6

4

2

0
0 -----------------------------------------------------




3/4/06 3/14/06 3/24/06 4/3/06 4/13/06 4/23/06 5/3/06 5/13/06 5/23/06 6/2/06 6/12/06
Date


-*-F112 -- F212


F312


Figure H-2. Average nitrate nitrogen concentration of soil at layer 1 (0-15 cm) under irrigation
level 12












20

18
16

14
12

10

8
6

4
2

0
3/4/


5/3/06


Date


-- F1I1 --F211


F311 FOIl


Figure H-3. Average nitrate nitrogen concentration of soil at layer 2 (15-30 cm) under irrigation
level II


20

18

16

14

12

10

8
a


4

2

0
3/4/


06


3/24/06


4/13/06


5/3/06


5/23/06


6/12/06


Date


-*- F112 -- F212


Figure H-4. Average nitrate nitrogen concentration of soil at layer 2 (15-30 cm) under irrigation
level 12


309


4/13/06


06


3/24/06


5/23/06


6/12/06


. I...





























3 3 0 I


Date


-- F1I1 -- F211


F311 FOIl


Figure H-5. Average nitrate nitrogen concentration of soil at layer 3 (30-60 cm) under irrigation
level II


3/24/06


4/13/06


5/3/06


5/23/06


6/12/06


Date


-*- F112 --- F212


Figure H-6. Average nitrate nitrogen concentration of soil at layer 3 (30-60 cm) under irrigation
level 12


20

18

S16
0)
E 14

o 12

10
0
z
c 6
0
Z 4

2

0


3/4/06


3/4/06


3/24/06


4/13/06


5/3/06


5/23/06


6/12/06












20
18
16

14
12
10
8
6
-


4
2
0
3/4/


06


3/24/06


4/13/06


5/3/06


5/23/06


6/12/06


Date

-- F1I1 F211 F311 F011


Figure H-7. Average nitrate nitrogen concentration of soil at layer 4 (60-90 cm) under irrigation
level II


20
18-
18
14-
16
14
12
10
8
6
-


4
2
0
3/4/1


06


3/24/06


4/13/06


5/3/06


5/23/06


6/12/06


Date

-*-F112 -F212 F312


Figure H-8. Average nitrate nitrogen concentration of soil at layer 4 (60-90 cm) under irrigation
level 12












20

18

16

E 14
C-
.o 12

S10

0 8
0
z
4 6
z 4 ----

2

0
3/4/06 3/14/06 3/24/06 4/3/06 4/13/06 4/23/06 5/3/06 5/13/06 5/23/06 6/2/06 6/12/06
Date

-F111 --- F211 F311 F011


Figure H-9. Average ammonium nitrogen concentration of soil at layer 1 (0-15 cm) under
irrigation level I1





20

18

S16

E14

S12

a 10

0 8

z 6
I

z4

2

0
0 --------------------------------------



















3/4/06 3/14/06 3/24/06 4/3/06 4/13/06 4/23/06 5/3/06 5/13/06 5/23/06 6/2/06 6/12/06

Date

-F- F112 -- F212 F312


Figure H-10. Average ammonium nitrogen concentration of soil at layer 1 (0-15 cm) under
irrigation level 12


312












20
18
)16
E
14
12

10
o
8


14
z 2
2

0
3/4/06 3/14/06 3/24/06 4/3/06 4/13/06 4/23/06 5/3/06 5/13/06 5/23/06 6/2/06 6/12/06
Date

-- F1I1 -- F211 F311 FOIl


Figure H-11. Average ammonium nitrogen concentration of soil at layer 2 (15-30 cm) under
irrigation level I1





20

18

16

E 14
.2 12

10
0 8
z
4 6
I
z 4

2

0
3/4/06 3/14/06 3/24/06 4/3/06 4/13/06 4/23/06 5/3/06 5/13/06 5/23/06 6/2/06 6/12/06
Date

-F112 -- F212 F312


Figure H-12. Average ammonium nitrogen concentration of soil at layer 2 (15-30 cm) under
irrigation level 12





















8
6
4


0
3/4/06 3/14/06 3/24/06 4/3/06 4/13/06 4/23/06 5/3/06 5/13/06 5/23/06 6/2/06 6/12/06
Date


-- F1I1 --F211


F311 F011


Figure H-13. Average ammonium nitrogen concentration of soil at layer 3 (30-60 cm) under
irrigation level I1


'& 16
0)
S14
o2 12

S10
0 8
S6

z4
z 4


3/4/06 3/14/06 3/24/06 4/3/06 4/13/06 4/23/06 5/3/06 5/13/06 5/23/06 6/2/06 6/12/06
Date


-*- F112 F212


Figure H-14. Average ammonium nitrogen concentration of soil at layer 3 (30-60 cm) under
irrigation level 12


314











20
18
16
14
12
10
8
6
4
2
0
3/4/06 3/14/06 3/24/06 4/3/06 4/13/06 4/23/06 5/3/06 5/13/06 5/23/06 6/2/06 6/12/06
Date


-- F1I1 --F211


F311 FOIl


Figure H-15. Average ammonium nitrogen concentration of soil at layer 4 (60-90 cm) under
irrigation level I1


20
18
c 16
E
S14
o
S12
i 10
0
o 8
C-
o
z 6
4
z
2
0


3/4/06 3/14/06 3/24/06 4/3/06 4/13/06 4/23/06 5/3/06 5/13/06 5/23/06 6/2/06 6/12/06
Date


-- F112 F212


F312


Figure H-16. Average ammonium nitrogen concentration of soil at layer 4 (60-90 cm) under
irrigation level 12


315









LIST OF REFERENCES


Aggarwal, p.K., 1995. Uncertainties in crop, soil and weather inputs used in growth models:
implications for simulated outputs and their applications. Agricultural Systems 48, 361-
284.

Ahsam, M. and O'Connor, K.M., 1994. A reappraisal of the Kalman filtering technique, as
applied in river flow forecasting. J. Hydrology 161: 197-226.

Albert, M.A., 2002. Monitoring and Modeling the Fate and Transport of Nitrogen in the Vadose
Zone Beneath a Suwannee River Basin Vegetable Farm. Thesis, Agricultural and
Biological Engineering Department, University of Florida, Gainesville, FL.

Allen, R.G., Pereira, L.S., Raes, D., Smith, M., 1998. Crop evapotranspiration Guidelines for
computing crop water requirements FAO Irrigation and drainage paper 56. FAO Food
and Agriculture Organization of the United Nations, Rome.

Alva, A.K., Paramasivam, S., Obreza, T.A., Schumann, A.W., 2005. Nitrogen best management
practice for citrus trees I. Fruit yield, quality, and leaf nutritional status. Scientia
Horticulturae. Scientia Horticultureae. 107:233-244

Aronica, G., Hankin, B.G., Beven, K.J., 1998. Uncertainty and equifinality in calibrating
distributed roughness coefficients in a flood propagation model with limited data. Adv.
Water Resour. 22 (4), 349-365.

Asadi, M. E., Clemente, R.S., Gupta, A.D., Loof, R., Hansen, G.K., 2002. Impacts of fertigation
via sprinkler irrigation on nitrate leaching and corn yield in an acid-sulfate soil in
Thailand. Agricultural Water Management 52 (2002) 197-213.

Barlund, I., Tattarri, S., 2001. Ranking of parameters on the basis of their contribution to model
uncertainty. Ecological Modeling 142, 11-23.

Barrow, E.M., Semenov, M.A., 1995. Climate-change scenarios with high spatial and temporal
resolution for agricultural applications. Forestry 84, 349-360.

Bauder, T., Waskom, R., 2003. Best Management Practices for Colorado Corn. Colorado State
University Cooperative Extension Bulletin XCM574A, February 2003.

Bellocchi, G., Donatelli, M., Fila, G., 2003. Calculating reference evapotranspiration and crop
biomass using estimated radiation inputs. Italian Journal of Agronomy 7, 95-102.

Beven, K.J., 1993. Prophecy, reality and uncertainty in distributed hydrological modeling. Adv.
Water Resour. 16, 41-51.

Beven, K., and Binley, A., 1992. The future of distributed models: model calibration and
uncertainty prediction, Hydrol. Process. 6: 279-298.









Beven, K., and Freer, J., 2001. Equifinality, data assimilation, and uncertainty estimation in
mechanistic modeling of complex environmental systems using the GLUE methodology.
Journal of Hydrology, 249, 11-29.

Black, R.J., 2003. Florida Climate Data. EES5, Environmental Horticulture Department, Florida
Cooperative Extension Service, Institute of Food and Agricultural Sciences, University of
Florida. Accessed at: http://edis.ifas.ufl.edu/EH105

Booltink, H.W.G., Thornton, P.K., Verhagen, J., Bouma, J., 1996. Application of simulation
models and weather generators to optimize farm management strategies. In: Robert, P.C.,
Rust, R.H., Larson, W. (Eds.), Proceedings of the third conference on precision
agriculture. Minneapolis, USA, June 23-26 (SSSA Special Publication).

Boote, K.J., Jones, J.W, Pickering, N.B.,1996. Potential uses and limitations of crop models.
Agronomy J. 88, 704-716.

Bottcher A.B., Tremwel, T.K., Campbell, K.L., 1995. Best management practices for water
quality improvement in the Lake Okeechobee Watershed. Ecological Engineering. 5,
341-356.

Brisson, N., Mary, B., Ripoche, D., Jeuffroy, M-H., Ruget, F., Nicoullaud, B., Gate, P.,
Devienne-Barret, F., Antonioletto, R., Durr, C., Richard, G., Beaudoin, N., Recous, S.,
Tayot, X., Plenet, D., Cellier, P., Machet, J-M., Meynard, J-M., Delecolle, R., 1998.
STICS: a genetic model for the simulation of crops and their water and nitrogen balances.
I. Theory and parameterization applied to wheat and corn. Agronomie 18, 311-346.

Bundy, L.G., Andraski, T.W., 2005. Recovery of fertilizer nitrogen in crop residues and cover
crops on an irrigated sandy soil. Soil Science Society of American Journal 69: 640-648.

Burrough, P.A., 1989. Matching spatial databases and quantitative models in land resource
management. Soil Use and Manage.5. 3-8.

Cambardella, C. A., Colvin, T. S., Karlen, D. L., Logsdon, S. D., Berry, E. C., Radke, J.K.,
Kapspar,T. C., Parkin, T. B., and Jaynes, D. B., 1996. Soil property contributions to yield
variation pattern. In: Robert, P.C., Rust, R.H., Larson, W.E. (Eds.), Proceedings of the
Third International Conference on Precision Agriculture. ASA, CSSA, SSSA, Inc.,
Madison, WI, PP. 224-417.

Campbell, E.P., Fox, D.R., Bates, B.C., 1999. A Bayesian approach to parameter estimation and
pooling in nonlinear flood event models, Water Resour. Res. 35, 211-220.

Center, T. J., Houston, J. E., Fuchs, C., Zeddies, J., 1996. Employing BMPs to reduce
agricultural water pollution: economics, regulatory institutions and policy concerns.
Drake Law Rev. 45: 125-141.

Chan, K., Tarantola, S., Saltelli, A., Sobol, I.M., 2000. Variance-based methods. In sensitivity
analysis, A Saltatelli, K. Chan, and E.M. Scott (eds.), 167-197. Wiley, New York.









Chaubey, I., Haan, C.T., Grunwald, S., and Salisbury, J.M., 1999. Uncertainty in model
parameters due to spatial variation of rainfall. J. Hydrology 220 (1-2): 48-61.

Cockx, E., Simonne, E.H., 2003. Reduction of the impact of fertilization and irrigation on
processes in the nitrogen cycle in vegetable fields with BMPs. HS948, Horticultural
Sciences Department, Florida Cooperative Extension Service, Institute of Food and
Agricultural Sciences. University of Florida.

Code of Federal Regulations, 1994. Title 40, Part 131.

DeBusk, W.F., 2001. Overview of the Total Maximum Daily Load (TMDL) Program. SL188,
Soil and Water Sciences Department, Florida Cooperative Extension Service, Institute of
Food and Agricultural Sciences. University of Florida.

Dubus, I.G., Brown, C.D., 2002. Sensitivity and first-step uncertainty analyses for the
preferential flow model MACRO. Journal of Environmental Quality 3, 227-240.

Fangmeier, D.D., Elliot, W.J., Workman, S.R., Huffman, R.L., Schwab, G.O., 2006. Soil and
Water Conservation Engineering. Fifth Edition, Thomon Delmar Learning, Thomson
Corporation.

Florida Agricultural Statistics Service, 2001 May. 1999-2000 Vegetable Summary. Florida
Agricultural Statistics Service, Orlando, FL. Accessed at:
http://www.nass.usda.gov/fl/rtocO.htm

Florida Agricultural Statistics Service, 2002 February. 2000-2001 Vegetable Acreage,
Production, and Value. Florida Agricultural Statistics Service, Orlando, FL. Accessed at:
http://www.nass.usda.gov/fl/rtocO.htm

Florida Department of Agriculture and Consumer Services, 2005. Water Quality/Quantity Best
Management Practices for Florida Vegetable and Agronomic Crops.

Florida Statutes, s.403.067, 1999. Establishment and implementation of total maximum daily
loads. Florida Senate. Accessed at:
http://www.flsenate.gov/statutes/index.cfm?Appmode=Display Statute&Search_String
=&URL=Ch0403/SEC067.HTM&Title=->1999->Ch0403->Section%20067#0403.067

Franks, S.W., Gineste, P., Beven, K.J., and Merot. P., 1998. On constraining the predictions of a
distributed model: The incorporation of fuzzy estimates of saturated areas into the
calibration process, Water Resour. Res. 34 787-797.

Freer, J., Beven, K.J., Ambroise, B., 1996. Bayesian estimation of uncertainty in runoff
prediction and the value of data: an application of the GLUE approach. Water Resour.
Res. 32(7), 2161-2173

Ghaffari A., Cook, H.F., Lee, H.C., 2001. Simulating winter wheat yields under temperate
conditions: exploring different management scenarios. European Journal of Agronomy 15
(2001) 231-240.









Graettinger, A.J., and Dowding, C.H., 2001. Quantifying exploration sufficiency while
accommodating judgement. Mathematical Geology 33 (2), 133-153.

Haan, P.K., and Skaggs, R.W, 2003a. Effects of parameter uncertainty on DRAINMOD
predictions: I. Hydrology and yield. Trans. ASAE 46(4): 1061-1067.

Haan, P.K., and Skaggs, R.W, 2003b. Effects of parameter uncertainty on DRAINMOD
predictions: II. Nitrogen loss. Trans. ASAE 46(4): 1069-1075.

Ham, L.K., Hatzell, H.H., 1996. Analysis of Nutrients in the Surface Waters of the Georgia-
Florida Coastal Plain Study Unit, 1970-91. U.S. Geological Survey Water Resources
Investigations Report 96-4037.

Haman, D.Z., and Smajstrla A. G., 1997. Scheduling Tips For Drip Irrigation of Vegetables.
AE259, the Agricultural and Biological Engineering Department, Florida Cooperative
Extension Service, Institute of Food and Agricultural Sciences, University of Florida.
Accessed at: http://edis.ifas.ufl.edu/AE092#TABLE_1

Hanna, S.R., Chang, J.C., and Fernau,M.E., 1997. Monte Carlo estimates of uncertainties in
predictions by a photochemical grid model (UAM-IV) due to uncertainties in input
variables. Atmospheric Environment 1997;32(21):3619-28.

Harmon, R., Challenor P., 1997. A Markov chain Monte Carlo method for estimation and

Heinmann, A.B., Hoogenboom, G., Chojnicki, B., 2002. The impact of potential errors in rainfall
observations on the simulation of crop growth, development, and yield. Ecological
Modeling 157, 1-21.

Hession, W.C., Storm, D.E., and Haan, C.T., 1996. Two-phase uncertainty analysis: An example
using the Universal Soil Loss Equation. Trans. ASAE 39 (4): 1309-1319.

Hochmuth, G., 1994. Sweet corn response to N and K fertilization and to vitamins as growth
enhancers. Fla. Agr. Expt. Sta. Res. Rep. Suwannee Valley AREC 94-01.

Hochmuth, G., 1997a. Response of sweet corn and snapbean to Growplex Humate. Fla. Agr.
Expt. Sta. Res. Rep. Suwannee Valley AREC 97-7.

Hochmuth, G., 1997b. Snapbean and sweet corn response to N rate and furrow-placed Growplex
humate. Fla. Agr. Expt. Sta. Res. Rep. Suwannee Valley AREC 97-21.

Hochmuth, G., 2000. Nitrogen Management Practices for Vegetable Production in Florida.
Circular 1222, Florida Cooperative Extension Service, Institute of Food and Agricultural
Sciences, University of Florida. First Published: April 2000.

Hochmuth, G., 2003. Nutrient and irrigation management guide spring sweet corn-2003.
Unpublished.









Hochmuth, G., Cordasco, K., 2000. A summary of N, P, and K research with sweet corn in
Florida. Horticultural Sciences Department, HS-758, Florida Cooperative Extension
Services, Institute of Food and Agricultural Sciences, University of Florida.

Hochmuth, G., Donley, M., 1992. Nitrogen fertilization of sweet corn on a sandy soil in northern
Florida. Fla. Agr. Expt. Sta. Res. Rep. Suwannee Valley AREC 92.

Hochmuth, G.J., and Hanlon, E.A., 2000. Commercial Vegetable Fertilization Principles.
Circular 225-E, Florida Cooperative Extension Service, Institute of Food and Agricultural
Sciences, University of Florida. Accessed at: http://edis.ifas.ufl.edu/CV009

Hochmuth, G., Hochmuth B., and Donley, M., 1992. Nitrogen fertilization of sweet corn on a
sandy soil in northern Florida. Fla. Agr. Expt. Sta. Res. Rep. Suwannee Valley AREC 92.

Hochmuth, G.J., Maynard, D.N., Vavrina, C.S., Stall, W.M., Kucharek, T.A., Johnson, F.A. and
Taylor, T.G.,1996 July. Sweet Corn Production in Florida. Horticultural
Sciences.Department HS 737, Florida Cooperative Extension Service, Institute of Food
and Agricultural Sciences, University of Florida. Accessed at:
http://edis.ifas.ufl.edu/scripts/htmlgen.exe?DOCUMENT_CV135

IFAS-UF (Institute of Food and Agricultural Sciences University of Florida), 2006. Citrus
BMPs. Accessed on: January 20, 2006; Accessed at: http://citrusbmp.ifas.ufl.edu/

Irrigation Association, 2001. Certified Landscape Irrigation Auditor. The Irrigation Association.

Jagtap, S.S., Jones, J.W., Hildebrand, P., Letson, D., O'Brien, J.J., Podesta, G., Zieten, D.,
Zazueta, F., 2002. Responding to stakeholder's demands for climate information: from
research to applications in Florida. Agric. Syst. 74, 415-430.

James, L.G., Erpeneck, J.M., Bassett, D.L., and Middleton, J.E., 1982. Irrigation requirement for
Washintong-estimates and methodology. Research bulletin XB0925. Agricultural
Research Center, Washington State University, Pullman. 37pp.

Jarvis, S.C., 1996. Future trends in nitrogen research. Plant Soil 181, 47-56.

Jones, J.W., Allen, L.H., Shih, S.F., Roger, J.S., Hammond, L.C., Smajstrla, A.G., Martsolf, J.D.,
1984. Estimated and measured evapotranspiration for Florida climate, crops and soils.
Bulletin 840 (Tech). IFAS, University Of Florida, Gainesville, FL.

Jones, J., Hoogenboom, G., Porter C., Boote, K., Batchelor, W., Hunt, L., Wilkens, P., Singh, U.,
Gijsman A., and Ritchie, J., 2003. The DSSAT model cropping system model. European
J. Agronomy 18 (3-4): 235-265.

Jones, C.A., Kiniry, J.R., 1986. CERES-Maize: A simulation model of maize growth and
development. Texas A&M University Press, College Station, Texas.

Jones, A.J., Mielke, L.M., Bartles, C.A., and Miller, C.A., 1989. Relationship of landscape
position and properties to crop production. J. Soil Water Conserv. 44, 328-332.









Judge, G.G., Hill, R.C., Griffiths, W.E., Lutkepohl, H., and Lee, T.C., 1982. Introduction to the
Theory and Practice of Economics. New York, Wiley.

Judge, J., Casanova, J.J., Lin, T., Tien, K.C., Jang, M., Lanni, O., and Miller, L.W., 2005. Field
observations during the second microwave, water, and energy experiment (MicroWEX-2):
from March through June 3, 2004. Circular No. 1480. Gainesville, FL. University of
Florida, Center for Remote Sensing. Accessed at: http://edis.ifas.ufl.edu/AE360.

Katyal, J.C., Singh, B., Vlek, P.L.G., Grasweel, E.T., 1985. Fate and efficiency of nitrogen
fertilizers applied to wetland rice. II. Punjab, India. Fertil. Res. 6, 279-290.

Katz, B.G., Hornsby, H. D., Bohlke, J.F., Mokray, M.F., 1999. Sources and Chronology of
Nitrate Contamination in Spring Waters, Suwannee River Basin, Florida. U.S. Geological
Survey, Water-Resources Investigations Report 99-4252, Prepared in cooperation with
the Suwannee River Water Management District.

Keller, J, and Bliesner, R.D., 1990. Sprinkle and Trickle Irrigation. The Blackburn Press,
Caldwell, New Jersey 07006 U.S.A.

Khakural, B. R., Robert, P. C., and Mulla, D. J., 1996. Relating corn/soybean yield to variability
in soil and landscape characteristics. In: Robert, P.C., Rust, R.H., Larson, W.E. (Eds.),
Proceedings of the Third International Conference on Precision Agriculture. ASA,
CSSA,SSSA, Inc., Madison, WI, PP. 117-128.

Kidder, G., Hanlon, E. A., Hochmuth, G. J., 1989. IFAS standardized fertilization
recommendations for vegetable crops. Fla. Coop. Ext. Serv. Spec. Ser. SS-SOS-907.

Klute, A., 1986. Methods of Soil Analysis Part 1: Physical and Mineralogical Methods. Second
Edition. Ameican Society of Agronomy, Inc., Soil Science Society of America, Inc.,
Publisher, Madison, Wisconsin USA. Number 9 (Part 1) in the series.

Knowles, L. Jr., 1996, Estimation of evapotranspiration in the Rainbow Springs and Silver
Springs Basins in north-central Florida: U.S. Geological Survey Water-Resources
Investigations Report 96-4024.

Koehler, J.R., Owen, A.B., 1996. Computer experiments. In Handbook of Statistics, Vol.13,
Design and Analysis of Experiments, S. Ghosh and C.R. Rao (eds.), 261-308. Elsevier,
Amsterdam.

Larson, B.C., Mossler, M.A., Nesheim, O.N., 1999. Florida Crop/Pest Management Profiles:
Sweet Corn. Food Science and Human Nutrition Department, Florida Cooperative
Extension Service, Institute of Food and Agricultural Sciences, University of Florida.
Accessed at: http://edis.ifas.ufl.edu/PI034

Lawless, C., and Semenov, M.A., 2005. Assessing lead-time for predicting wheat growth using a
crop simulation model. Agricultural and Forest Meteorology 135, 302-313.









Li, M., Yost, R.S., 2000. Management-oriented modeling: optimizing nitrogen management with
artificial intelligence. Agricultural Systems 65 (2000) 1-27.

Li, Y.C., Alva, A.K., Calvert, D.V., Banks, D.J., 2002. Atmospheric deposition of nitrogen in a
high lightning intensity area. Commun. Soil Sci. Plant Anal. 33, 1671-1677.

Linkov, I. and Burmistrov, D., 2005. Sources of uncertainty in model predictions: lessons
learned from the IAEA Forest and Fruit Working Group model intercomparisons. Journal
of Environmental Radioactivity. Article in press.

Liu, X., Ju, X., Zhang, F., Pan, J., Christie, P., 2003. Nitrogen dynamics and budgets in a winter
wheat-maize cropping system in the North China Plain. Field Crops Research 83, 111-
124.

Macdonald, L., and Strachan, P., 2001. Practical application of uncertainty analysis. Energy and
buildings 33 (2001) 219-227.

Makowski, D., Hillier, J., Wallach, D., and Andrieu, B., 2006. Working with dynamic crop
models: evaluation, analysis, parameterization, and application, Wallach, D, Makowski,
D., and Jones, J.W., (eds.)Elsevier B.V., pp. 55, pp.101-103.

Makowski, D., Wallach, D., and Tremblay, M. 2002. Using a Bayesian approach to parameter
estimation: comparison of the GLUE and MCMC methods. Agronomie 22:191-203.

Malakoff D., 1999. Bayes offers a 'New' way to make sense of numbers, Science 286, 1460-
1464.

Martinez, J.E., Duchon, C.E., Crosson, W.L., 2001. Effect of the number of soil layers on a
modeled surface water budget. Water Resources Research 37, 367-377.

Maynard, D.N., Olson, S.M., 2001. Vegetable production guide for Florida 2001-2002.
Extension, Institute of Food and Agricultural Science, the University Of Florida.

Meisinger, J.J., Randall, G.W., 1991. Estimating nitrogen budgets for soil-crop systems. In:
Follett, R.F. (Ed.), Managing Nitrogen for Groundwater Quality and Farm Profitability.
ASA, CSSA, SSSA, Madison, WI, pp.85-124.

Migliaccio, K.W., and Boman, B.J., 2006. Total Maximum Daily Loads and Agricultural BMPs
in Florida. Department of Agricultural and Biological Engineering, Florida Cooperative
Extension Service, Institute of Food and Agricultural Sciences, University of Florida.
Accessed at: http://edis.ifas.ufl.edu

Monod, H., Naud, C., and Makowski, D., 2006. Working with dynamic crop models: evaluation,
analysis, parameterization, and application, Wallach, D, Makowski, D., and Jones, J.W.,
(eds.) Elsevier B.V., pp. 55-57, pp.60, and pp.67-86.

Monteith, J. L.,1996. The quest for balance in crop modeling. Agronomy J. 88, 695-697.









Morris, M.D., 1991. Factorial sampling plans for preliminary computational experiments.
Technometrics, Vol.33, No. 2., pp.161-174.

Mullins, C.A., Straw, R.A., Pitt. B., Onks, D.O., Mullen, M.D., Reynolds, J., Kirchner, M., 1999.
Response of selected sweet corn cultivars to nitrogen fertilization. Horttechnology,
January-March 1999 9(1).

Nix, H., 1987. The role of crop modeling, minimum data sets and geographical information
systems in the transfer of agricultural technology. In Agricultural Environments.
Characterization, Classification and Mapping, ed. A.H. Bunting. CAB International, UK,
pp. 113-117.

Ogle, S.M., Breidt, F.J. Eve, M.D., and Paustian, K, 2003. Uncertainty in estimating land use and
management impacts on soil organic carbon storage for U.S. agricultural lands between
1982 and 1997. Global Change Biology 9 (11): 1521-1542.

Olson, S.M., and Simonne, E., 2005. Vegetable production guide for Florida 2003-2004.
Extension, Institute of Food and Agricultural Science, the University Of Florida.

Page, A.L, Miller, E.H., and Keeney, D.R., 1982. Methods of soil analysis: Part 2 Chemical and
microbiological properties. Second edition. American Society of Agronomy, Inc. Soil
Science Society of America, Inc. Publisher, Madison, Wisconsin, USA. p595 and p673.

Panda, R.K., Behera, S.K., and Kashyap, P.S., 2004. Effective management of irrigation water
for maize under stressed conditions. Agricultural Water Management 66 (2004) 181-203.

Paz, J.O., Batchelor, W.D., Babcock, B.A., Colvin,T.S., Logsdon, S.D., Kaspar, T.C., and Karlen,
D.L., 1999. Model-based technique to determine variable rate nitrogen for corn.
Agricultural System 61, 69-75.

Pebesma, E.J., and Heuvelink., G.B.M., 1999. Latin hypercube sampling of Gaussian random
fields. Technometrics 41 (4): 303-312.

Pei, H. and Wang, Y., 2003. Eutrophication research of West Lake, Hangzhou, China: modeling
under uncertainty. Water Research 37 (2003) 416-428.

Penning de Vries, F.W.T., & Teng, P. S., 1993. Systems approach to agricultural development,
eds. Kluwer Academic Publishers, the Netherlands.

Peter, S.M., 1979. Stochastic Models, Estimation, and Control. Vol. 1. New York, N.Y.:
Academic Press.

Pittman, J.R., Hatzell, H.H., Oaksford, E.T., 1997. Spring Contributions to Water Quality and
Nitrate Loads in the Suwannee River during Base Flow in July 1995. U.S Geological
Survey Water-Resources Investigations Report 97-4152.

Poss, R., Saragoni, H., 1992. Leaching of nitrate, calcium and magnesium under maize
cultivation on an oxisol in Togo. Fertil. Res. 33, 123-133.









Priestly, C.H.B., Taylor, H.B., 1972. On the assessment of surface heat flux and evaporation
using large-scale parameters. Monthly Weather Review 100: 81-92.

Rahn, C., Mead, A., Draycott, A., Lillywhite, R., Salo, T., 2001. A sensitivity analysis of the
prediction of the nitrogen fertilizer requirement of cauliflower crops using the HRI
WELL-N computer model. Journal of Agricultural Science 137, 55-69.

Rangarajan, A., Ingall, B., Orfanedes, M., Wolfe, D., 2002. In-row spacing and cultivar affects
ear yield and quality of early-planted sweet corn. HortTechnology, July-September 2002,
12(3).

Rhoads, F. M., 1990. Sweet corn research in north Florida. Fla. Agri. Expt. Sta. Research Report,
Quincy.

Richardson, C.W., 1984. Weather generators for crop management model. Trans Amer. Soc.
Agric. Eng., 28, 1602-6.

Richter, O. & Sondgerdth, D., 1990. Parameter estimation in ecology. The link between data and
models. VCH Publishers, Weinheim, FRG, pp.218.

Rinaldi, M., Ventrella, D., Gagliano, C., 2007. Comparison of nitrogen and irrigation strategies
in tomato using CROPGRO model. A case study from Southern Italy. Agricultural Water
Management 87, 91-105.

Ritchie, J.T., 1998. Soil water balance and plant stress. In: Tsuji, G.Y., Hoogenboom, G.,
Thornton, P.K. (Eds.), Understanding options for agricultural production. Kluwer
Academic Publishes, Dordrecht, the Netherlands, pp.41-54.

Ritchie, J.T., and Otter, S., 1985. Description and performance of CERES-Wheat: a user-oriented
wheat yield model. In: ARS Wheat Yield Project. ARS-38. Natl Tech Info Serv,
Springfield, Missouri, pp. 159-175.

Ritchie, J.T., Singh, U., Godwin, D.C., Bowen, W.T., 1998. Cereal growth, development and
yield. In: Tsuji, G.Y., Hoogenboom, G., Thornton, P.K. (Eds.), Understanding options for
agricultural production. Lluwer academic publishers, Dordrecht, the Netherlands, pp.79-
98.

Robertson, W.K., 1962. A study of sources, rates and methods of application of nitrogen on
sweet corn. Proc. Fla, State Hort. Soc. 75:249-253.

Rochester, E.W., 1995. Landscape Irrigation Design. The Rochester Group, Auburn, AL.ASAE
Publication No.8.

Romanowicz, R., Beven, K.J., Tawn, J., 1994. Evaluation of predictive uncertainty in non-linear
hydrological models using a Bayesian approach. In: Barnett, V., Turkman, K.F. (Eds.).
Statistics for the Environment II. Water Related Issues. Wiley, New York, pp. 297-317.









Romanowicz, R., Beven, K.J., Tawn, J., 1996. Bayesian calibration of flood inundation models.
In: Anderson, M.G., Walling, D.E., Bates, P.D. (Eds.). Flood Plain Processes. Wiley,
Chichester.

Rudert, B.D., and Locascio, S.J., 1979. Growth and tissue composition of sweet corn as affected
by N source, nitrapyrin, and season. J. Amer. Soc. Hort. Sci. 104:520-523.

Sabbagh, G.J., and Fox, G.A., 1999. Statistical method for evaluation of a water table
management model. Trans. ASAE 42(3): 713-719.

Sacks, J., Welch, W., Mitchell, T., Wynn, H.P., 1989. Design and analysis of computer
experiments. Statistical Science 4, 409-435.

Salvador, R., Pinol, J., Tarantola, S., Pla. E., 2001. Global sensitivity analysis and scale effects of
a fire propagation model used over Mediterranean shrubland. Ecological Modeling 136,
175-189.

SAS Inst. Inc., 1996. SAS Statistics User's Guide Release Version 6. SAS Inst. Inc., Cary, NC.

Shuler, K. D., 2002. Effect of within-row plant spacing on sweet corn grown on muck soil, Belle
Glade, Florida, Spring and Fall 2001. Proc. Fla. State Hort. Soc. 115: 226-232. 2002.

Shulz K., Beven K., Huwe B., Equifinality and the problem of robust calibration in nitrogen
budget simulations, Soil Sci. Soc. Am. J. 63 (1999) 1934-1941.

Simonne, E., and Hochmuth, G.,2003. Supplemental fertilizer application in the BMP area for
vegetable. UF/IFAS, Fla. Coop. Ext. Serv., HS906. Accessed at:
http://edis.ifas.ufl.edu/HS163

Simonne, E., Simonne, A., and Boozer, R., 1999. Yield, ear characteristics, and consumer
acceptance of selected white sweet corn varieties in the southeastern United States.
HortTechnology, April-June 1999 9(2).

Sogbedji, J.M., van Es, J.M., Yang, C.L., Geohring, L.D., Magdoff, F.R., 2000. Nitrate leaching
and nitrogen budget as affected by nitrogen rate and soil type. J. Environ. Qual. 29, 1813-
1820.

Stockle, C.O., Nelson, R., Donatelli, M., Castellvi, F., 2001. ClimGen: a flexible weather
generation program. In: Proceedings 2nd International Symposium Modeling Cropping
Systems, 16-18 July, Florence, Italy. European Society of Agronomy, pp. 229-230.

Sudduth, K.A., Drummond, S.T., Birrell, S. J. and Kitchen, N.R., 1996. Analysis of spatial
factions influencing crop yield. In: Robert, P.C., Rust, R.H., Larson, W.E. (Eds.),
Proceedings of the Third International Conference on Precision Agriculture. ASA,
CSSA,SSSA, Inc., Madison, WI, PP. 129-140.









Sumner, D.M., 1996, Evapotranspiration from successional vegetation in adeforested area of the
Lake Wales Ridge, Florida: U.S. Geological Survey Water-Resources Investigations
Report 96-4244.

Suwannee River Water Management District, 2004. Surface Water Quality and Biological
Annual Report 2003. Suwannee River Water Management District, FL.

Teng, P. S. & Penning de Vries, F.W.T., 1992. Systems approach to agricultural development.
Elsevier Applied Science Ltd, London, pp.309.

Theocharopoulos, S.P., Karayianni, M., Gatzogiani, P., Afentaki, A., Aggelides, S., 1993.
Nitrogen leaching from soils in the kopais area of Greece. Soil Use Manage. 9, 76-84.

Thorp, K.R., Batchelor, W.D., Paz, J.O., Steward, B.L., Caragea, P.C., 2006. Methodology to
link production and environmental risks of precision nitrogen management strategies in
corn. Agricultural Systems 89 (2006) 272-298.

Tsuji, G.Y., Jones, J.W., Balas, S., (Eds.), 1994. DSSAT v3, vols. 1-3. University of Hawaii,
Honolulu, Hawaii.Tsuji, G.Y., Jones, J.W., Balas, S., (Eds.), 1994. DSSAT v3, vols. 1-3.
University of Hawaii, Honolulu, Hawaii.

Tsuji, G.Y., Hoogenboom, G., Thornton, P.K., (Eds.), 1998. Systems approach for sustainable
agricultural development: understanding options for agricultural production. Kluwer
Academic Publishers in cooperation with ICASA.

U .S. Department of Agriculture, 1962. United States standards for grades of sweet corn for
processing. Agricultural Marketing Service, United States Department of Agriculture.
Effective May 15, 1962 (Reprinted January 1997). Accessed at:
http://www.ams.usda.gov/standards/vpcorsw.pdf

U.S. Dept. of Agriculture/National Agricultural Statistics Service, 1993. Agricultural Chemical
Usage 1992 Vegetables Summary. Economic Research Service, Washington, D.C.

U .S. Dept. of Agriculture/National Agricultural Statistics Service, 1995. Agricultural Chemical
Usage, Vegetable Crop Summary 1994. Agricultural Statistics Board, National
Agricultural Statistics Service, U.S. Department of Agriculture. Accessed at:
http://usda.mannlib.cornell.edu/reports/nassr/other/pcu-bb/

U.S. Dept. of Agriculture/National Agricultural Statistics Service, 1997. 1996 U.S. Vegetable
Commodity Rankings Top Ten States and United States Based on Production. Accessed
at: http://www.usda.gov/nass/pubs/ranking/vrank97.htm

U.S. Dept. of Agriculture/National Agricultural Statistics Service, 1998. 1997 Census of
Agriculture Volume 1: National, State and County Tables. National Agricultural
Statistics Service. Accessed at:
http://www.nass.usda.gov/census/census97/volumel/vollpubs.htm









U.S. Dept. of Agriculture/ National Agricultural Statistics Service, 1999a. Vegetables 1998
Summary. Vg. 1-2 (99) c. Accessed at:
http://usda.mannlib.cornell.edu/reports/nassr/fruit/pvg-bban/vgan0199.txt

U .S. Dept. of Agriculture/National Agricultural Statistics Service, 1999b. Agricultural Chemical
Usage, Vegetables, 1998 Summary. Agricultural Statistics Board, National Agricultural
Statistics Service, U.S. Department of Agriculture. Accessed at:
http://usda.mannlib.cornell.edu/reports/nassr/other/pcu-bb/

U.S. Dept. of Agriculture/National Agricultural Statistics Service, 2002. 2002 Census of
Agriculture Volume 1, Chapter 1: Florida State Level Data. National Agricultural
Statistics Service. Accessed at:
http://www.nass.usda.gov/census/census02/volumel/fl/indexl.htm

U.S. Dept. of Agriculture/National Agricultural Statistics Service, 2003. Agricultural Chemical
Usage, Vegetables, 2002 Summary. Agricultural Statistics Board, National Agricultural
Statistics Service, U.S. Department of Agriculture. Accessed at:
http://usda.mannlib.cornell.edu/reports/nassr/other/pcu-bb/

U.S. Dept. of Agriculture/National Agricultural Statistics Service, 2007. Agricultural Chemical
Usage, Vegetables, 2006 Summary. Agricultural Statistics Board, National Agricultural
Statistics Service, U.S. Department of Agriculture. Accessed at:
http://usda.mannlib.cornell.edu/reports/nassr/other/pcu-bb/

U.S. Dept. of Health, Education, and Welfare, 1962. Public health service drinking water
standards. Publ. 956 U.S. Dep. Helth, Educ., and Welfare, Washing, DC.

Van Delden, A., Schroder, J.J., Kropff, M.J., Grashoff, C., Booij, R., 2003. Simulated potato
yield, and crop and soil nitrogen dynamics under different organic nitrogen management
strategies in the Netherlands. Agriculture, Ecosystems and Environment 96 (2003) 77-95.

Van Straten, G and Keesman, K.J., 1991. Uncertainty propagation and speculation in projective
forecasts of environmental change: A lake-eutrophication example. Journal of
Forecasting. Vol. 10, no. 1-2, pp. 163-190. 1991.

Volk, G.M., 1962. Comparison of covered to non-covered side dressing of urea, ammonium
nitrate, uran, feran and calcium nitrate on sweet corn. Proc. Fla. State Hort. Soc. 75:170-
175.

Wallach, D., Goffinet, B., 1987. Mean squared error of prediction in models for studying
ecological and agronomic systems. Biometrics 43, 561-573.

Wang, X., He, X, Williams, J.R., Isaurralde, R. C., and Atwood, J. D. 2005. Sensitivity and
uncertainty analyses of crop yields and soil organic mcarbon simulated with EPIC. Trans
ASAE 48(3):1041-1054.

Waskom, R.M., 1994. Best Management Practices for Nitrogen Fertilization. Colorado State
University Cooperative Extension, Bulletin #XCM-172, August 1994.









Welch, W.J., Buck, R.J., Sacks, j., Wynn, H.P., Mitchell, T.J., Morris, M.D., 1992. Screening,
predicting, and computer experiments. Technometrics 4, 15-25.

White, J.M., Tyson, R.V., Hanlon, E.A., Hochmuth, G.J., and Neal, C.A., 1996. Plant petiole sap
testing for nitrogen and potassium in sweet corn grown on mineral soil. Proc. Fla. State
Hort. Soc. 109:149-151.

Woods, C., 2005. A New Tune: Years of Research on Nitrate Pollution and Agricultural
Solutions to Combat It All Spell Relief for the Historic Suwannee River. Explore
Research at the University of Florida. Summer 2005, Volume 10, Issue 2. Accessed at:
http://www.rgp.ufl.edu/publications/explore/v10n2/story3.html

Xie, Y., Kiniry, J.R., Williams, J.R., 2003. The ALMANAC model's sensitivity to input
variables. Agricultural Systems 78, 1-16.

Zak, S.K., Beven, K.J., Reynolds, B., 1997. Uncertainty in the estimation of critical loads: a
practical methodology. Soil, Water Air Pollut.









BIOGRAPHICAL SKETCH

Jianqiang He was born in Tianshui, Gansu Province, China, in 1977. He received his

Bachelor of Engineering in mechanical engineering from the Jilin University, China in 2000 and

Master of Engineering in irrigation machinery design from Chinese Academy of Agricultural

Mechanization and Science in 2003. In January 2004, he came to Univerisity of Florida to pursue

his Ph.D. degree in the Agricultural and Biological Engineering Department under the

supervision of Dr. Michael D. Dukes. In August 2008, he received his Ph.D. degree.





PAGE 1

BEST MANAGEMENT PRACTICE DEVELOPMENT WITH THE CERES-MAIZE MODEL FOR SWEET CORN PRODUCTION IN NORTH FLORIDA By JIANQIANG HE A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2008 1

PAGE 2

2008 Jianqiang He 2

PAGE 3

To my parents, brother, and the teachers who taught me at different stages in my life 3

PAGE 4

ACKNOWLEDGMENTS I believe my most important achievement in the past four years was learning how to do academic research and how to cooperate with other people. I recognized that this crucial improvement in my life is owed to the contributions of many people. First and foremost, I am greatly indebted to my supervisor, Dr. Michael D. Dukes, for his insightful guidance, continuous encouragement, and unselfish support in my research over the past more than four years. His thoughtful coaching with all aspects of my research was a guarantee of the success of this endeavor. His enthusiasm and preciseness have left an everlasting impression on me. I will never forget how he never even neglected a tiny typo when he was revising my writing. Without his help, it would not have been possible for me to complete this research. I would like to express sincere appreciation to Dr. Wendy Graham and Dr. James Jones for their insightful and invaluable advice on resolving all kinds of technical problems in my research. I will never forget their sacrifices, spending their valuable time to meet with me many times. Their enthusiasm for research and pursuit for excellence will be an example to me forever. Without their combined supervision of each step throughout the model simulation, this study would not have been possible. I am grateful to Dr. Jasmeet Judge for her help in the field experiment and advice in model simulation, to Dr. George Hochmuth for his advice in sweet corn BMP development, and to Dr. Michael Annable for his advice in nitrogen movement. I deeply benefited from their suggestions and advice on a multitude of perspectives regarding my research. I would like to thank Dawn Lucas and Dr. Donald Graetz for their help in laboratory experiments. Dawn helped me analyze thousands of soil and water samples. I am thankful they 4

PAGE 5

let me occupy their laboratory for my experiment at least three weeks in every summer from 2004 to 2006. Special thanks go to Danny Burch, staff of the Agricultural and Biological Engineering (ABE) Department, and people in the Plant Science Research and Education Unit, University of Florida. Without their help, the burdensome field work would not have been completed. I would also like to thank the undergraduate students (Steve, Clay, Frank, Habtu, Sanjiv, Patrick, Thi, David, etc.). They helped me conduct the field and laboratory experiments. Now they are my good friends. I am also grateful to several officemates for their friendship and encouragement, and faculty, staff and students in the ABE Department for the harmonious academic environment. Finally I particularly appreciate my parents and younger brother for their unconditional love, understanding, patience and encouragement. I love and miss them. 5

PAGE 6

TABLE OF CONTENTS page ACKNOWLEDGMENTS ...............................................................................................................4 LIST OF TABLES .........................................................................................................................11 LIST OF FIGURES .......................................................................................................................14 ABSTRACT ...................................................................................................................................19 1 INTRODUCTION..................................................................................................................21 1.1 Study Background ............................................................................................................21 1.1.1 Nitrate Pollution in North Florida ..........................................................................21 1.1.2 Sweet Corn Production in Florida ..........................................................................23 1.1.3 Total Maximum Daily Loads and Best Management Practice ...............................24 1.1.4 Best Management Practices for Sweet Corn Production ........................................27 1.1.5 Best Management Practice Development ...............................................................31 1.2 Objectives .........................................................................................................................35 1.3 Dissertation Outline ..........................................................................................................35 2 GLOBAL SENSITIVITY ANALYSIS OF CERES-MAIZE MODEL WITH ONE-AT-A-TIME METTHOD..............................................................................................................39 2.1 Introduction .......................................................................................................................39 2.1.1 Sensitivity Analysis ................................................................................................39 2.1.2. Local Sensitivity Analysis .....................................................................................41 2.1.3 Global Sensitivity Analysis ....................................................................................42 2.2 Materials and Methods .....................................................................................................43 2.2.1. Model Description .................................................................................................43 2.2.1.1 CERES-Maize model ...................................................................................43 2.2.1.2 Soil water sub-module ..................................................................................44 2.2.1.3 Soil nitrogen sub-module .............................................................................46 2.2.2 Non-restricted OAT Method ..................................................................................47 2.2.3 Normalization of Input Parameters ........................................................................48 2.2.4 Restricted OAT Method .........................................................................................49 2.2.5 One-at-a-time (OAT) Method for CERES-Maize Model ......................................51 2.2.6 Field Experiment ....................................................................................................54 2.3 Results and Discussion .....................................................................................................55 2.3.1 Non-restricted OAT Results ...................................................................................55 2.3.1.1 Response profiles .........................................................................................55 2.3.1.2 Correlation coefficient matrix ......................................................................56 2.3.1.3 Influential parameter selection based on non-restricted OAT method ........57 2.3.2 Influential Parameter Selection Based on Restricted OAT Method .......................58 2.4 Summary and conclusions ................................................................................................59 6

PAGE 7

3 PARAMETER ESTIMATION FOR CERES-MAIZE MODEL WITH THE GLUE METHOD...............................................................................................................................68 3.1 Introduction .......................................................................................................................68 3.1.1 Parameter Estimation ..............................................................................................68 3.1.2 GLUE Method ........................................................................................................71 3.2 Method and Materials .......................................................................................................73 3.2.1 Field Experiment ....................................................................................................73 3.2.2 Main Procedure of GLUE ......................................................................................77 3.2.3 Selection of Input Parameters .................................................................................77 3.2.4 Prior Distribution ....................................................................................................78 3.2.5 Model Run with Generated Parameter Vectors ......................................................79 3.2.6 Determination of Number of Model Runs ..............................................................81 3.2.7 Likelihood Function and Likelihood Value ...........................................................81 3.2.7.1 Available likelihood functions .....................................................................81 3.2.7.2 Selection of likelihood function and method of likelihood value combination ..................................................................................................88 3.2.7.3 Comparison of distributions of input parameters .........................................91 3.2.7.4 Comparison of distributions of outputs ........................................................91 3.2.8 Estimation of Posterior Distribution .......................................................................92 3.2.9 GLUE Simulation ...................................................................................................93 3.2.10 GLUE Verification ...............................................................................................93 3.2.11 Expected Values of Posterior Distribution ...........................................................94 3.3 Results and Discussion .....................................................................................................95 3.3.1 Results of Prior Distribution ...................................................................................95 3.3.2 Results of Number of Model Runs .........................................................................96 3.3.3 Results of Likelihood Function and Method of Likelihood Value Combination .............................................................................................................97 3.3.3.1 Comparison of distributions of input parameters .........................................97 3.3.3.2 Comparison of distributions of model outputs .............................................99 3.3.4 Distributions of Selected Parameters ....................................................................101 3.3.5 PDF Plot of Selected Parameters ..........................................................................102 3.3.6 Distributions of Outputs .......................................................................................103 3.3.7 Joint Distribution between Yield and Nitrogen Leaching ....................................104 3.3.8 GLUE Verification ...............................................................................................104 3.3.9 Result of Expected Values of Posterior Distribution ...........................................106 3.4 Conclusions .....................................................................................................................106 4 FIELD EXPERIMENT OF SWEET CORN AND SIMULATION WITH CALIBRATED CERES-MAIZE MODEL..........................................................................129 4.1 Introduction .....................................................................................................................129 4.2 Material and Methods .....................................................................................................131 4.2.1 Experiment Site and Design .................................................................................131 4.2.2 Nitrogen Fertilizer Application ............................................................................133 4.2.3 Irrigation Scheduling ............................................................................................136 7

PAGE 8

4.2.4 Soil, Biomass, and Yield Sampling......................................................................138 4.2.5 CERES-Maize Model Simulation........................................................................140 4.3 Results and Discussion...................................................................................................142 4.3.1 Quantity of Sweet Corn Yield..............................................................................142 4.3.2 Quality of Sweet Corn Yield................................................................................143 4.3.3 Nitrogen Balance Estimation................................................................................145 4.3.3.1 Nitrogen input............................................................................................145 4.3.3.2 Nitrogen output..........................................................................................147 4.3.3.3 Nitrogen balance........................................................................................148 4.3.4 Comparison between Model Simulations and Field Observations.......................149 4.3.4.1 Comparison between dry matter yields......................................................149 4.3.4.2 Comparison between phenology dates.......................................................150 4.3.4.3 Comparison between potential nitrogen leaching......................................150 4.4 Conclusions.....................................................................................................................153 5 BEST MANAGEMENT PRACTICE DEVELOPMENT WITH CERES-MAIZE MODEL FOR SWEET CORN PRODUCTION IN NORTH FLORIDA........................... 169 5.1 Introduction.....................................................................................................................169 5.2 Materials and Methodology............................................................................................171 5.2.1 Experiment Site....................................................................................................171 5.2.2 Crop Model Calibration........................................................................................172 5.2.3 BMP Simulations..................................................................................................173 5.2.4 Determination of Acceptable Yield......................................................................179 5.3 Results and Discussion...................................................................................................181 5.3.1 Effects of Irrigation..............................................................................................181 5.3.2 Effects of Nitrogen Fertilizer................................................................................184 5.3.2.1 Total nitrogen fertilizer amount.................................................................184 5.3.2.2 Nitrogen fertilizer split...............................................................................186 5.3.2.3 Amount of nitrogen fertilizer in each application......................................186 5.3.3 Selection of Potential BMPs.................................................................................188 5.3.4 Evaluation and Implementation of Potential BMPs.............................................189 5.4 Summary and Conclusions.............................................................................................191 6 UNCERTAINTY ANALYSIS OF POTENTIAL SWEET CORN BMPS UNDER WEATHER AND INPUT PARAMETER VARIABILITY................................................208 6.1 Introduction.....................................................................................................................208 6.2 Materials and Methods...................................................................................................211 6.2.1 Field Experiment and Weather Data....................................................................211 6.2.2 Uncertainty of Input Parameters...........................................................................212 6.2.3 Selected Potential BMPs......................................................................................213 6.2.4 A Grower Practice of N Fertilizer and Irrigation Management...........................213 6.2.5 Monte Carlo Simulation.......................................................................................215 6.3 Results and Discussion...................................................................................................216 6.3.1 BMP Comparison.................................................................................................216 6.3.2 Output Uncertainty Plot........................................................................................218 8

PAGE 9

6.3.3 Output Uncertainty over Time Range of 1958-1990............................................220 6.4 Summary and Conclusions.............................................................................................221 7 CONCLUSIONS AND FUTURE WORK.......................................................................... 236 7.1 Summary and Research Contributions...........................................................................236 7.2 Conclusions.....................................................................................................................237 7.2.1 Global Sensitivity Analysis of CERES-Maize Model with One-at-a-time (OAT) Method.......................................................................................................237 7.2.2 Parameter Estimation for CERES-Maize Model with GLUE Method................238 7.2.3 Field Plot Experiment of Sweet Corn and Simulation with Calibrated CERES-Maize Model..........................................................................................................239 7.2.4 Best Management Practices Development with CERES-Maize Model for Sweet Corn Production in North Florida...............................................................240 7.2.5 Uncertainty Analysis of Potential Sweet Corn BMPs under Weather and Input Parameter Variability.............................................................................................242 7.3 Future Work....................................................................................................................243 APPENDIX..................................................................................................................................245 A INPUT AND OUTPUT PARAMETERS OF CERES-MAIZE MODEL IN DSSAT.........245 B MATLAB CODE FOR GLOBAL SENSITIVITY ANALYSIS WITH THE RESTRICTED OAT METHOD...........................................................................................246 B.1 Main Function................................................................................................................246 B.2 Sensitivity Analysis of Genotype Parameter.................................................................246 B.3 Genotype File Change....................................................................................................248 B.4 Genotype Parameter Space............................................................................................249 B.5 Processing Sensitivity Analysis Results of Genotype Parameter..................................251 B.6 Sensitivity Analysis of Soil Parameter...........................................................................252 B.7 Soil File Change.............................................................................................................255 B.8 Soil Parameter Space.....................................................................................................259 B.9 Processing Sensitivity Analysis Results of Soil Parameter...........................................261 C MATLAB CODE FOR GLUE PROCESS...........................................................................262 C.1 Main Function................................................................................................................262 C.2 Generation of Random Numbers...................................................................................263 C.3 Function mvnrnd........................................................................................................264 C.4 Parameter Setup for Genotype and Soil.........................................................................265 C.5 Change of Soil File........................................................................................................266 C.6 Change of Genotype File...............................................................................................268 C.7 Summary Output Processing..........................................................................................269 C.8 Plant Nitrogen Output Processing..................................................................................270 C.9 Soil Nitrogen Output Processing....................................................................................273 C.10 Parameter PDF Plot......................................................................................................275 C.11 3-D Plot of Joint Distribution of Yield and Nitrogen Leaching..................................288 9

PAGE 10

D PICTURES OF FIELD EXPERIMENT...............................................................................293 E SAS CODE FOR ANOVA OF YIELD QUANTITY AND QUALITY..............................303 F NITRATE AND AMMONIUM CONCENTRATIONS IN MONITORING WELLS IN BLOCK 1 IN THE PLANT SCIENCE RESEARCH AND EDUCATION UNIT UNIVERSITY OF FLORIDA..............................................................................................305 G TOTAL KJELDAHL NITROGEN CONCENTRATION OF LEAVES AND STEMS OF SWEET CORN IN FIELD EXPERIMENT IN PLOTS IN 2006..................................306 H NITRATE AND AMMONIUM NITROGEN CONCENTRATION OF SOIL IN FIELD EXPERIMENT OF SWEET CORN IN PLOTS IN 2006....................................................308 LIST OF REFERENCES.............................................................................................................316 BIOGRAPHICAL SKETCH.......................................................................................................329 10

PAGE 11

LIST OF TABLES Table page 1-1 Sweet corn harvested for sale in Florida in 2002 and 1997 (USDA-NASS, 1998, 2002)..................................................................................................................................38 1-2 Nitrogen fertilizer application for sweet corn in Florida (USDA-NASS, 1993, 1995, 1999b, 2003, 2006)............................................................................................................38 2-1 Genotype coefficient for the DSSAT CERES-Maize model.............................................65 2-2 Covariance coefficient matrix of genotype and soil parameters of the DSSAT model.....66 2-3 Criteria for input parameter determination a .......................................................................67 2-4 Selected parameters for GLUE simulation based on the non-restricted OAT method and covariance coefficient matrix a .....................................................................................67 2-5 Mean and variance of absolute elementary effects of genotype parameters.....................67 2-6 Mean and variance of absolute elementary effects of soil parameters..............................67 2-7 Selected parameters for model calibration based on the restricted OAT method..............67 3-1 Average soil physical properties of the experiment site (from 24 sampling locations)...122 3-2 Selected parameters for GLUE method due to sensitivity analysis of predicted dry matter yield and accumulative nitrogen leaching (See Chapter 2 for details) a ................122 3-3 Covariance matrix of the prior distribution.....................................................................122 3-4 Results of Jarque-Bera test of the input parameters a b ....................................................122 3-5 Mean values and standard deviations (STDEV) of first-round posterior distributions derived from different likelihood functions and likelihood combinations a .....................123 3-6 Mean values and standard deviations (STDEV) of model outputs derived from first-round posterior distributions ab .........................................................................................124 3-7 Fundamental statistical properties of prior, first posterior and second posterior distributions derived from L1C2......................................................................................125 3-8 Measured and estimated mean values of soil properties of the field experiment site......125 3-9 Selected parameter set for GLUE verification a ................................................................126 3-10 Generated duplicates of observations for GLUE verification..........................................126 11

PAGE 12

3-11 Means and standard deviations of the selected parameters in GLUE verification a .........127 3-12 Means and standard deviations of model outputs in GLUE verification.........................128 3-13 Expectation values of second posterior distribution of selected parameters a ..................128 4-1 Soil properties of the experiment site..............................................................................162 4-2 DU lq values of 4 different numbers of drip tapes at 3 depths at t=30min........................162 4-3 Fertigation schedules of field plot experiment in 2006...................................................162 4-4 Crop coefficients of sweet corn at different stages of development................................162 4-5 Second posterior distribution of the selected parameters................................................163 4-6 Measured and estimated mean values of soil properties of the field experiment site......163 4-7 ANOVA results of total yield of sweet corn....................................................................163 4-8 Irrigation and nitrogen treatment effects on yield quantity.............................................164 4-9 ANOVA results of total ears of sweet corn.....................................................................164 4-10 Irrigation and nitrogen treatment effects on yield quality...............................................165 4-11 Nitrogen budget of a replicate of treatment F1I1 in Block 1 of the plot experiment......165 4-12 Estimated nitrogen leaching of seven treatment in field plot experiment.......................166 4-13 ANOVA results of nitrogen leaching estimated from N balance....................................166 4-14 Irrigation and nitrogen treatment effects on cumulative nitrogen leaching estimated from N balance.................................................................................................................166 4-15 Simulated and measured dry yields in field plot experiment in 2006..............................167 4-16 Simulated and measured anthesis and maturity dates in field plot experiment...............167 4-17 Nitrogen balance of model simulation of treatment F1I1................................................167 4-18 Simulated potential nitrogen leaching of the seven treatment in field plot experiment..168 4-19 Simulated and estimated accumulative nitrogen leaching in field plot experiment........168 5-1 Expectation values of second posterior distribution of selected parameters a ..................200 5-2 Soil properties of the experiment site..............................................................................200 5-3 Calculation of total available soil water (ASW) in the soil profile..................................200 12

PAGE 13

5-4 Irrigation treatments based on different MAD values.....................................................200 5-5 Nitrogen splits used in BMP simulation..........................................................................201 5-6 Nitrogen splits used in single factor simulation...............................................................201 5-7 Acreage, yield, production, and value of Florida sweet corn 1998-2006 (USDANASS, 2007)........................................................................................................202 5-8 Fresh yields of selected white sweet corn varieties in Clanton Ala. 1995-1996 (Simonne et al. 1999).......................................................................................................202 5-9 Fresh yields of sweet corn experiment in Springfield Tenn. 1993-1995 (Mullins et al., 1999)................................................................................................................................202 5-10 Fresh yields of sweet corn experiment in Eden Valley and Freeville, NY, 1998-2001 (Rangarajan et al., 2002)..................................................................................................203 5-11 Fresh yields of sweet corn experiment in Belle Glade, Florida, in spring of 2001 (Shuler, 2002)..................................................................................................................204 5-12 Summary of sweet corn yield in field experiments conducted in Florida (Hochmuth and Cordasco, 2000)........................................................................................................204 5-13 Selected irrigation strategies............................................................................................205 5-14 Ranking of dry yield (HWAH) and nitrogen leaching (NLCM) under different N fertilizer application splits................................................................................................205 5-15 Selected factors of N fertilizer application strategies......................................................205 5-16 Ranking of average nitrogen leaching (NLCM) of combination management over 33 years (1958-1990)............................................................................................................206 5-17 Selected potential BMPs for sweet corn production........................................................207 6-1 Second posterior distribution of the selected parameters (from Chapter 3)....................232 6-2 Six selected potential BMPs for sweet corn production (from Chapter 5)......................232 6-3 N fertilizer management in the EPA319 Project..........................................................233 6-4 Irrigation management in the EPA319 Project.............................................................234 6-5 Mean and standard deviation (STDEV) of simulated corn dry yield and nitrogen leaching both under different uncertainty scenarios a ......................................................235 13

PAGE 14

LIST OF FIGURES Figure page 1-1 Diagram of research structure............................................................................................37 2-1 Scheme of non-restricted OAT method.............................................................................62 2-2 Scheme of restricted OAT method....................................................................................62 2-3 Response profiles of sweet corn yield to six normalized genotype parameters................63 2-4 Response profiles of sweet corn yield to nine normalized soil parameters.......................63 2-5 Response profiles for the nitrogen leaching to six normalized genotype parameters........64 2-6 Response profiles for the nitrogen leaching to nine normalized soil parameters..............64 3-1 Diagram of Block 1 of field experiment..........................................................................109 3-2 Influence of number of model runs on mean values of P1..............................................109 3-3 Influence of number of model runs on standard deviations of P1...................................110 3-4 Influence of number of model runs on mean values of SLRO........................................110 3-5 Influence of number of model runs on standard deviations of SLRO.............................111 3-6 Influence of number of model runs on mean values of simulated dry yields..................111 3-7 Influence of number of model runs on standard deviations of simulated dry yields.......112 3-8 Influence of number of model runs on mean values of simulated nitrogen leaching......112 3-9 Influence of number of model runs on standard deviations of simulated nitrogen leaching............................................................................................................................113 3-10 Parametre P1: probability distribution ............................................................................113 3-11 Parametre P5: probability distribution.............................................................................114 3-12 Parametre PHINT: probability distribution.....................................................................114 3-13 Parametre SLDR: probability distribution.......................................................................115 3-14 Parametre SLRO: probability distribution.......................................................................115 3-15 Parametre SLLL: probability distribution........................................................................116 14

PAGE 15

3-16 Parametre SDUL: probability distribution.......................................................................116 3-17 Parametre SSAT: probability distribution.......................................................................117 3-18 Parametre SLPF: probability distribution........................................................................117 3-19 Histogram of predicted dry matter yields........................................................................118 3-20 Histogram of predicted anthesis dates.............................................................................118 3-21 Histogram of predicted maturity dates.............................................................................119 3-22 Histogram of predicted cumulative nitrogen leaching.....................................................119 3-23 Joint distribution between yield and nitrogen leaching under prior distribution of input parameters...............................................................................................................120 3-24 Joint distribution between yield and nitrogen leaching under the first posterior distribution of input parameters.......................................................................................120 3-25 Joint distribution between yield and nitrogen leaching under the second posterior distribution of input parameters.......................................................................................121 4-1 Experiment plot arrangement layout................................................................................155 4-2 Soil moisture at t=30 minutes with 1, 2, 3 and 4 drip tapes.............................................156 4-3 Drip tape arrangement in each row interval.....................................................................157 4-4 Drip tape arrangement and sampling zone in each plot...................................................157 4-5 Fresh yield under different N fertilizer levels under I1...................................................158 4-6 Yield under different N fertilizer levels under I2............................................................158 4-7 Number of ears per unit area under different N fertilizer levels under I1.......................159 4-8 Number of ears per unit area under different N fertilizer levels under I2.......................159 4-9 Number of ears per unit area under different irrigation levels under F1.........................160 4-10 Number of ears per unit area under different irrigation levels under F2.........................160 4-11 Number of ears per unit area under different irrigation levels under F3.........................161 5-1 Response curves of yield to different remaining ASW...................................................195 5-2 Response curves of nitrogen leaching to different remaining ASW................................195 15

PAGE 16

5-3 Response curves of yield to different irrigation depths...................................................196 5-4 Response curves of nitrogen leaching to different irrigation depths...............................196 5-5 Rainfall and accumulated irrigations in East Half of Block1 in 2006.............................197 5-6 Response curves of yield to different N fertilizer levels..................................................197 5-7 Response curves of nitrogen leaching to different N fertilizer levels..............................198 5-8 Dry yield vs. different N fertilizer application amount....................................................198 5-9 Nitrogen leaching vs. different N fertilizer application amount......................................199 6-1 Histogram and cumulative distribution of predicted average annual dry yield of the six selected potential BMPs and the actual grower practice both under weather and input parameter uncertainty.............................................................................................223 6-2 Histogram and cumulative distribution of predicted average annual nitrogen leaching (NLCM) of the six selected potential BMPs and the actual grower practice both under weather and input parameter uncertainty...............................................................227 6-3 Simulated 10% and 90% confidence limits of average annual yields of BMP1 both under weather and input parameter uncertainty...............................................................231 6-4 Simulated 10% and 90% confidence limits of average annual nitrogen leaching of BMP1 both under weather and input parameter uncertainty...........................................231 D-1 Components of nitrogen fertilizer solution......................................................................293 D-2 Fertigation control table...................................................................................................293 D-3 Fertigation system installation.........................................................................................294 D-4 Main fertigation lines, injection holes, peristaltic pump, and solution bucket................294 D-5 Sub-main fertigation lines................................................................................................295 D-6 Drip tapes and sub-main fertigation line..........................................................................295 D-7 Drip tape distribution in one row.....................................................................................296 D-8 Irrigation with the linear move irrigation system............................................................296 D-9 Sweet corn planting..........................................................................................................297 D-10 Sweet corn emergence.....................................................................................................297 D-11 Comparison between no-nirogen-applied plot (near) and nitrogen-applied plot (far)....298 16

PAGE 17

D-12 Sweet corn tasseling.........................................................................................................298 D-13 Sweet corn maturity.........................................................................................................299 D-14 Sweet corn harvest...........................................................................................................299 D-15 Plant sampling..................................................................................................................300 D-16 Soil sampling...................................................................................................................300 D-17 Yield sampling.................................................................................................................301 D-18 Yield weighing.................................................................................................................301 D-19 Yield grading...................................................................................................................302 D-20 Research partner...............................................................................................................302 F-1 Average nitrate concentration in the monitoring wells on the west part and east part of Block 1.........................................................................................................................305 F-2 Average ammonium concentration in the monitoring wells on the west part and east part of Block 1.................................................................................................................305 G-1 Average total Kjeldahl nitrogen (TKN) concentration of leaves of sweet corn under irrigation level I1..............................................................................................................306 G-2 Average total Kjeldahl nitrogen (TKN) concentration of leaves of sweet corn under irrigation level I2..............................................................................................................306 G-3 Average total Kjeldahl nitrogen (TKN) concentration of stems of sweet corn under irrigation level I1..............................................................................................................307 G-4 Average total Kjeldahl nitrogen (TKN) concentration of stems of sweet corn under irrigation level I2..............................................................................................................307 H-1 Average nitrate nitrogen concentration of soil at layer 1 (0-15 cm) under irrigation level I1.............................................................................................................................308 H-2 Average nitrate nitrogen concentration of soil at layer 1 (0-15 cm) under irrigation level I2.............................................................................................................................308 H-3 Average nitrate nitrogen concentration of soil at layer 2 (15-30 cm) under irrigation level I1.............................................................................................................................309 H-4 Average nitrate nitrogen concentration of soil at layer 2 (15-30 cm) under irrigation level I2.............................................................................................................................309 17

PAGE 18

H-5 Average nitrate nitrogen concentration of soil at layer 3 (30-60 cm) under irrigation level I1.............................................................................................................................310 H-6 Average nitrate nitrogen concentration of soil at layer 3 (30-60 cm) under irrigation level I2.............................................................................................................................310 H-7 Average nitrate nitrogen concentration of soil at layer 4 (60-90 cm) under irrigation level I1.............................................................................................................................311 H-8 Average nitrate nitrogen concentration of soil at layer 4 (60-90 cm) under irrigation level I2.............................................................................................................................311 H-9 Average ammonium nitrogen concentration of soil at layer 1 (0-15 cm) under irrigation level I1..............................................................................................................312 H-10 Average ammonium nitrogen concentration of soil at layer 1 (0-15 cm) under irrigation level I2..............................................................................................................312 H-11 Average ammonium nitrogen concentration of soil at layer 2 (15-30 cm) under irrigation level I1..............................................................................................................313 H-12 Average ammonium nitrogen concentration of soil at layer 2 (15-30 cm) under irrigation level I2..............................................................................................................313 H-13 Average ammonium nitrogen concentration of soil at layer 3 (30-60 cm) under irrigation level I1..............................................................................................................314 H-14 Average ammonium nitrogen concentration of soil at layer 3 (30-60 cm) under irrigation level I2..............................................................................................................314 H-15 Average ammonium nitrogen concentration of soil at layer 4 (60-90 cm) under irrigation level I1..............................................................................................................315 H-16 Average ammonium nitrogen concentration of soil at layer 4 (60-90 cm) under irrigation level I2..............................................................................................................315 18

PAGE 19

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy BEST MANAGEMENT PRACTICE DEVELOPMENT WITH THE CERES-MAIZE MODEL FOR SWEET CORN PRODUCTION IN NORTH FLORIDA By Jianqiang He August 2008 Chair: Michael D. Dukes Major: Agricultural and Biological Engineering Increasing nitrogen loads within the Suwannee River Basin of North Florida has become a major concern. Nitrogen fertilizer application in field crop production is proved to be the most import nitrogen contribution in this region. Florida ranks highest in the nation in the production and value of fresh market sweet corn. Thus it is necessary to develop research based nitrogen best management practices (N-BMPs) to reduce nitrogen leaching while keeping an acceptable yield in sweet corn production. This study is an attempt to utilize the CERES-Maize mode of the Decision Support System for Agrotechnology Transfer (DSSAT) model as a platform to develop potential BMPs for sweet corn production in North Florida. The results show that the non-restricted and restricted one-at-a-time (OAT) method can be used to conduct global sensitivity analysis for the CERES-Maize so as to select the most influential parameters for model calibration. The generalized likelihood uncertainty estimation (GLUE) method was proved to be a powerful tool for model parameter estimation, since the uncertainties in model input parameters were significantly reduced after GLUE was used to 19

PAGE 20

estimate the model input parameters. The uncertainties in model outputs were reduced correspondingly. The comparison between the model simulated and field observed results of the seven treatments in a field plot experiment of sweet corn in 2006, shows that the model did a good job in predicting dry yield and phenology dates. The results of BMP development with the calibrated CERES-Maize model show that if the growers could apply both irrigation water and nitrogen fertilizer more frequently but with smaller amounts in each application, this would result in an acceptable yield and a lower level of nitrogen leaching. The results showed a total nitrogen amount between 196 and 224 kg N ha -1 would be enough for sweet corn production in North Florida, which confirmed that the recommendation nitrogen amount (224 kg N ha -1 ) by Institute of Food and Agricultural Sciences IFAS, Univerisity of Florida, was reasonable. The results of uncertainty analysis of the CERES-Maize model for sweet corn simulation show that the weather was the dominant uncertainty contributor. This was because after two rounds of GLUE parameter estimation procedure, the uncertainties existing in input parameters were minimized. 20

PAGE 21

CHAPTER 1 INTRODUCTION 1.1 Study Background 1.1.1 Nitrate Pollution in North Florida Increasing nitrogen loads within the Suwannee River Basin of North Florida has recently become a major concern. According to the Surface Water Quality and Biological Annual Report 2003 (Suwannee River Water Management District, 2004), in 2003, 4,165 metric tons of nitrate-nitrogen and 1,733 metric tons of phosphorus were transported to the Gulf of Mexico by the Aucilla, Econfina, Fenholloway, Suwannee, and Waccasassa Rivers. The Suwannee River Basin alone accounted for 4,069 metric tons of nitrate nitrogen and 1,476 metric tons of total phosphorus. In 1995, a study was conducted to determine how springs and other ground-water inflow affect the quantity and quality of water in the Suwannee River (Pittman et al., 1997). They studied a 53-km stretch of the Suwannee River from Dowling Park, Fla., to Branford, Fla. Water samples for nitrate concentrations (dissolved nitrite plus nitrate as nitrogen) and discharge data were collected at 11 springs and 3 river sites during the 3-day period in July 1995 during base flow in the river. They found that nitrate (NO 3 -1 ) loads increased downstream from 2,300 to 6,000 kg day -1 an increase of 160% in the study reach, and that 54% of nitrate load increase was supplied by the groundwater inflow. Eighty-nine percent of the nitrate load increase occurred in the lower two-thirds of the stretch. Ham and Hatzell (1996) found that nitrate concentration in Suwannee River increased at a rate of 0.02 mg L -1 per year over a twenty-year period from 1971 to 1991, with an average concentration for the 20-year period of 0.5 mg L -1 Leaching of nitrate nitrogen is economically and environmentally undesirable (Katyal et al., 1985; Poss and Saragoni, 1992; Theocharopoulos et al., 1993). Nitrate that leaches below the 21

PAGE 22

crop root zone represents the loss of a valuable plant nutrient, and hence, increases agricultural costs. If nitrate enters groundwater supplies, it can also impose risks to both human health and the environment. Consumption by humans and animals through drinking water with high nitrate levels has been associated with several health problems. The most serious is methemoglobinemia or blue baby syndrome (O 2 deficiency in blood) in infants. High nitrate concentrations in drinking water are detrimental to the health of infants especially during the first 6 months of life. Additionally, groundwater with high nitrate levels that discharge into sensitive surface waters can contribute to long-term eutrophication of these water bodies (Asadi, et al., 2002). For this reason, the US Environmental Protection Agency (EPA) has set a maximum contaminant level requiring the nitrate-nitrogen concentration not exceed 10 mg N L -1 and the nitrite-nitrogen concentration not exceed 1 mg N L -1 in public water supplies (U.S. Dept. Health, Education, and Welfare, 1962). Nitrates leached into the groundwater of Suwannee River Basin are believed to come from several sources including animal wastes, chemical fertilizer, industrial, and domestic sewage. The Middle Suwannee River Basin, which includes Lafayette and Suwannee counties, has hundreds of residential and commercial septic systems in rural areas, about 300 row crop and vegetable farms, 44 dairies with more than 25,000 animals and 150 poultry operations with more than 38 million birds. Suwannee County is the leading poultry production area in Florid (Woods, 2005). According to the report by Katz et al. (1999), in Suwannee County, the relative contribution of N from fertilizers increased from about 23% in 1955 to more than 60% in 1980. During 1955-1995, the contribution of estimated N inputs from animal wastes (poultry, dairy and beef cows, and swine) ranged from about 21 to 42% of the total estimated N inputs. It is obvious that N fertilizer is the most important N contributor in this region. 22

PAGE 23

1.1.2 Sweet Corn Production in Florida Florida ranks highest in the nation in the production and value of fresh market sweet corn (Zea mays L.), typically accounting for approximately 25% of both national sweet corn production and of U.S. cash receipts for fresh sales (FASS, 2002; USDA-NASS, 1997, 1999a). Sweet corn has typically ranked as one of Florida's five most valuable vegetable crops. During the 2000-01 production seasons, sweet corn was the second ranked vegetable crop in terms of acreage and fifth ranked in total value. Harvested acreage for sweet corn represented 14.9% of the state's total vegetable acreage during that season, while production value represented 8% of the total production value of all Florida vegetables (FASS, 2002). Average yield ranged from approximately 8,200 kg ha -1 fresh sweet corn yield in 1969-1970, to 16,400 kg ha -1 in 2000-2001 (FASS 2001, 2002). The principal fresh sweet corn production region in Florida is the Everglades area (Palm Beach County), which during the 1999-2000 season produced 63% of the state crop. The southeastern/southwestern area (Miami-Dade, Collier, and Hendry Counties) were responsible for 25% of the state's production. The west/north area (Suwannee and Jackson Counties) accounted for about 7% of the sweet corn production. Sweet corn was also grown in the central area around Lake Apopka, but this region only produced about 5% of the crop since the muck soils in this area have been taken out of production (FASS, 2001). Table 1-1 shows the harvested acreage of sweet corn in Florida in 1997 and 2002 (USDA-NASS, 1998, 2002). It can be seen that in 1997 there were 413 sweet corn producing farms in Florida, with a total planted area of 17,791 ha. In 2002, number of farms decreased to 340, while the total planted area decreased to15,768 ha. In both years, the large farms, especially the ones that were greater than 200 ha consisted of the most part of total area. 23

PAGE 24

Table 1-2 shows the applications of chemical nitrogen fertilizer in Florida from 1992 to 2006 (USDA-NASS, 1993, 1995, 1999b, 2003, 2006). In 1998 and 2002, all sweet corn acreage in Florida received nitrogen applications totaling 1.83 and 2.61million kg, respectively. Between 1992 and 2006, from 81 to 100 % of sweet corn acreage in Florida received an average of 2.0 to 10.0 applications of nitrogen seasonally. An average range of 46 to 62 kg N ha -1 had been used at each application, with a statewide annual total N application ranging from 1.64 to 5.48 million kg. It should be noticed that in 2006 though only 86% of total planted area received nitrogen fertilizer, the total number of N applications increased dramatically to 10 times a season. Consequently, the N rate per crop year in 2006 increased to 475 kg ha -1 which was almost 3 times as that of 2002. The total applied chemical nitrogen fertilizer to sweet corn also doubled from 2002 to 2006. Adequate water is especially important in sweet corn production during periods of silking and tasseling and of ear development (Hochmuth et al., 1996). Most of Florida's sweet corn is grown under irrigation. In 1997, 53% of farms and 71% of sweet corn acreage was irrigated (USDA-NASS, 1998). About 92% of sweet corn growers in Florida surveyed in 1993 reported that they checked soil moisture and plant need to determine irrigation needs, while 8% used an established schedule modified to meet plant needs. Furthermore, only 8% were using a mechanical system to monitor soil moisture, and of those not using a mechanical system, 30% considered it too expensive, 30% reported not knowing of a good and inexpensive system, 30% cited limited water supply, and 10% said that lack of time prevented them from adopting a mechanical system (Larson et al., 1999). 1.1.3 Total Maximum Daily Loads and Best Management Practice In 1972, Congress passed the Clean Water Act (CWA) which set forth federal requirements for identification of polluted or impaired water bodies. These rules were passed 24

PAGE 25

down to the states by the U.S. Environmental Protection Agency (EPA), which requires states to establish a prioritized list of impaired water bodies and to develop estimated loads that the water bodies could receive of each pollutant while meeting water quality standards (DeBusk, 2001). These estimated loads determined for each water body are called Total Maximum Daily Loads (TMDLs). TMDLs are defined as the maximum amount of a pollutant that a water body can receive and still meet the water quality standards as established by the 1972 Clean Water Act. Section 303(d) of the act requires states to submit lists of surface waters that do not meet applicable water quality standards and to establish TMDLs for these waters on a prioritized schedule. In response to state TMDL requirements, the Florida Watershed Restoration Act (FWRA) was passed in 1999. This act established the Florida Department of Environmental Protection (FDEP) as the lead agency in coordinating the implementation of the TMDL allocation through water quality protection programs. These programs include non-regulatory and incentive-based programs, including best management practices (BMPs), cost sharing, waste minimization, pollution prevention, and public education. This act also required the Florida Department of Agriculture and Consumer Services (FDACS) to develop and adopt rules pursuant to suitable interim measures, best management practices, or other measures necessary to achieve the level of pollution reduction established by the FDEP for agricultural pollutant sources. These practices and measures may be implemented by those parties responsible for agricultural pollutant sources and the department, the water management districts, and the FDACS shall assist with implementation (Florida Statutes, s.403.067, 1999). The FDEP also should develop TMDL calculations for each water body or water body segment according to the priority ranking and schedule unless the impairment of such waters is 25

PAGE 26

due solely to activities other than point and non-point sources of pollution. When a water body is identified as impaired and a TMDL is established, pollutant loads are divided among the different stakeholders (agriculture and urban). Hence, the TMDL shall include establishment of reasonable and equitable allocations of the total maximum daily load between or among point and non-point sources that will alone, or in conjunction with other management and restoration activities, and achieve water quality standards for the pollutant causing impairment. The allocations may establish the maximum amount of the water pollutant that may be discharged or released into the water body or water body segment in combination with other discharges or releases. Allocations may also be made to individual basins and sources or as a whole to all basins and sources or categories of sources of inflow to the water body or water body segments (Florida Statutes, s.403.067, 1999). Normally, each stakeholder would implement a set of management practices that are expected to reduce its contribution to meet its designated load. These practices are commonly referred to as BMPs and can be defined as a practice or combination of practices determined by the coordinating agencies, based on research, field-testing, and expert review, to be the most effective and practicable on-location means, including economical and technological considerations, for improving water quality in agricultural and urban discharges. Although some water bodies do not have designated TMDLs as of yet, and therefore do not legally require BMPs, many agricultural BMP manuals are being developed. FDEP, FDACS, and the Institute of Food and Agricultural Science (IFAS) at the University of Florida have partnered with local agencies and stakeholders to develop BMP manuals (Migliaccio and Boman, 2006). The primary benefit for growers implementing agricultural BMPs (even without a designated TMDL) is that if a BMP program is in place, an agricultural producer is considered to 26

PAGE 27

be operating under a presumption of compliance with water quality standards. This protects the farmer from liabilities to the state when water quality standards are not met (IFAS-UF, 2006). According to the Water Quality/Quantity Best Management Practices for Florida Vegetable and Agronomic Crops (FDACS, 2005), all farming operations using this BMP manual shall reasonably attempt to implement the recommended BMPs in order to establish a baseline set of BMPs to ensure a reduction in pollutant loading to impaired receiving waters. Depending on the farms site specific conditions, all of these baseline BMPs need not be implemented. Only BMPs applicable for a particular location and production system should be implemented. This Tier-1 or first level of BMP protection also includes many of the practices that are identified as essential under USDA-NRCS conservation planning procedures. Irrigation scheduling and optimum fertilizer management are two of the proposed set of minimum BMPs that are suggested to be implemented. 1.1.4 Best Management Practices for Sweet Corn Production In this research, focus was on the most common cultural practices that directly affect the N cycle, N fertilization and irrigation. Fertilization is the cultural practice that can directly influence the N cycle in the root zone of sweet corn. Fertilization affects not only plant uptake, but also mineralization, nitrification, denitrification, and ammonia volatilization (Cockx and Simonne, 2003). Mineralization will not be significant in sandy soil due to the low organic matter content, but will be significant in organic soils. However, approximately 50% of total N-fertilizer applied can be taken up by the crop (Bundy and Andradki, 2005), i.e. about 50% of the total applied N-fertilizer would be lost by leaching, volatilization, denitrification, etc. Irrigation is another important factor. Florida is among the wettest states in the U.S. with most areas receiving an average of 1,270 mm of rain annually (Black, 2003). However, rainfall 27

PAGE 28

distribution is not adequate for vegetable production and irrigation must be used since rainfall is always unevenly distributed in time and space (Cockx and Simonne, 2003). Irrigation scheduling is used to apply the proper amount of water to a crop at the proper time. The characteristics of the irrigation system, crop needs, soil properties, and atmospheric conditions must all be considered to properly schedule irrigations. Poor timing or insufficient water application can result in crop stress and reduced yields from inappropriate amounts of available water and/or nutrients. Excessive water applications may reduce yield and quality, are a waste of water, and increase the risk of nutrient leaching (Maynard and Olson, 2001). Irrigation must be scheduled according to water availability and crop need. Irrigation scheduling requires knowing when to irrigate and how much water to apply. When to irrigate can be determined from plant or soil indicators or water balance techniques. How much water to apply can be based on soil water measurements or water balance techniques (Fangmeier etc., 2006). Monitoring soil status always means checking soil water tension (SWT). SWT represents the magnitude of the suction (negative pressure) the plant roots have to create to free soil water from the attraction of the soil, and move it into root cells. The dryer the soil, the higher the suction needed, hence, the higher SWT. SWT can be measured in the field with moisture sensors or tensiometers (Olson and Simonne, 2005). Crop water requirement information is needed when establishing a soil water budget to forecast irrigation events. The sum of the water lost from the soil surface (evaporation) and water used by plants (transpiration) is called evaportranspiration (ET). There are many factors that affect the rate of ET, including plant species, weather factors, and the amount and quality of water available to the plant. Generally, reference ET (ET 0 ) is determined for use as a base level. 28

PAGE 29

Crop water use (ET C ) is related to ET 0 by a crop coefficient (K C ) that is the ratio of ET C to ET 0 (Irrigation Association, 2001). Water usage also varies with soil dryness. Plants can remove water more easily from a wet soil. To account for this, a concept called readily available water (RAW) has been developed (Keller and Bliesner, 1990). It defines the amount of water that is more easily remove by the plant. Another associated term, maximum allowable depletion (MAD) relates RAW with available water (AW), which is the water that can be stored in soil and be available for growing crops. Usually, the value of MAD is given for a particular plant and the RAW is then computed with equation EAW=AWMAD. The MAD values can be expressed as percentage and usually range from 0.4 to 0.6 (Rochester, 1995). BMPs are specific cultural practices that aim at reducing the loads of specific compounds while increasing or maintaining economical yields (Simonne and Hochmuth, 2003). The implementation of BMPs may be a key factor in reducing the consequences of alterations of the N cycle in sweet corn fields. Implementation of BMPs at the farm level is a key to maintaining the quality and the quantity of ground and surface water. Li and Yost (2000) stated that the application rates, timing, and method of both N fertilization and irrigation are important tools that determine and control the fate and behavior of N in soil-plant systems. For example, multiple applications with small amounts of fertilizer (e.g. split application) usually enhance plant uptake and reduce potential nitrate leaching, although increasing costs. Waskom (1994) summarized BMPs for nitrogen fertilization for crops such as corn, sugar beet, and beans as follows: (1) Time application of N fertilizer to coincide as closely as possible to the period of maximum crop uptake; 29

PAGE 30

(2) Use sidedress or in-season fertilizer application for at least 40% of the total N applied to irrigated spring planted crops or fields with severe leaching hazard; (3) Apply N fertilizer where it can be most efficiently taken up by the crop: a) Ridge banded fertilizer used in conjunction with alternate row furrow irrigation can reduce downward movement of N; b) Multiple, small applications of N through sprinkler irrigation systems can increase fertilizer efficiency and reduce total N fertilizer application; c) Fertilizers applied on irrigated fields with high surface loss potential should be subsurface banded or incorporated immediately after application; d) Nitrogen applied in irrigation water should be metered with an appropriate device that is properly calibrated. Due to the increased possibility of leaching or runoff, N fertilizer through conventional flood or furrow irrigation system is strongly discouraged. (4) The following recommendations apply to cropland fields where the leaching potential is moderate to severe: a) Follow alfalfa or other legumes with high N use crops (such as small grains, sugar beets, or corn) that efficiently use N fixed by the legume; b) Follow shallow-rooted crops with low N use efficiency in the rotation by a deep-rooted, high N use crop that scavenges excess N (such as corn, sugar beets, or alfalfa). Analyze subsoil samples for residual nitrate to determine carryover credit to the subsequent crop. Bauder and Waskom (2003) summarized the BMPs for corn in Colorado. The BMPs include: (1) use sidedress or in-season fertilizer application for at least 40% of the total N applied to irrigated crops with sandy soils; (2) use fall planted cover crops such as rye or triticale to scavenge excess N left in the soil after poor crop; (3) mix and store N fertilizer at least 30 m (100 feet) away from wells or any water supply; (4) if applying manure, incorporate manure as soon as possible after application to minimize volatilization losses, reduce odor, and prevent runoff, and (5) apply only enough irrigation water to fill the effective crop root zone. Hochmuth (2000) recommended the nitrogen management practices for vegetable production in Florida as follows: (1) knowing the crop nutrient requirement (CNR) for N and targeting this amount for total crop N fertilization; (2) setting realistic yield goals; (3) using polyethylene mulch, where practical, to protect N from leaching; (4) selecting controlled-release 30

PAGE 31

N fertilizers when practical and economical; (5) calibrating fertilizer applicators accurately and making adjustments to equipment so that the correct amount of N is applied in the correct position of the root zone or production bed, near the root system; (6) applying N at periods during the growing season when crop N uptake is most active; (7) using fertigation where possible to "spoon-feed" N to crops during the season; (8) managing irrigation water properly to avoid leaching and to keep water and N in the root zone; and (9) using tissue-testing or petiole sap testing to monitor crop-N status and to determine adjustments needed in the N-fertilization program. In addition, he also suggested 224 kg N ha -1 as nitrogen recommendations for sweet corn production on sandy mineral soils in Florida. In the Vegetable Production Guide for Florida 2003-2004, Olson and Simonne (2005) suggested that 20% to 25% of N should be applied at planting, then sidedress band the remaining N in one or two applications during the early part of growth cycle. After midseason, N can be applied through center pivot irrigation systems at rates of 11 to 22 kg N ha -1 in several applications. 1.1.5 Best Management Practice Development Best management practices related with irrigation and N fertilizer application have been developed with field plot experiments. For example, a study was conducted in an acid-sulfate soil in the central region of Thailand, in 1999 and 2000 to assess the influence of different rates of N fertigation on corn yield and nitrate leaching. The corn varieties planted in the two years were super sweet corn Agro variety (Zea mays L.) and the Suwan 3851 single-cross hybrid (Zea mays L.), respectively. The nitrogen source was urea and there were four N fertigation treatments that included 0 (control), 100, 150 and 200 kg N ha -1 each having three replications arranged in a randomized complete block design (RCBD). Soil was irrigated to field capacity at 50% available soil moisture depletion regime throughout the season. The average maximum corn 31

PAGE 32

grain yield of 3,520 kg ha 1 was obtained at 200 kg N ha 1 in 1999 and 5,420 kg ha 1 was obtained at 150 kg N ha 1 in 2000. But the statistical analysis did not show any significant differences in grain yield between N200 and N150 treatments in either year. The nitrate leaching was calculated from the equationDPN C RL whereDPRwas the water drainage, andnitrate nitrogen concentration in soil water measured by a soil water sampler. The C was in 2). results of leaching calculation showed that the highest leaching values were obtained in N200 treatmentsboth years with 23 and 5.3 kg N ha 1 in 1999 and 2000, respectively. The lowest yield of 0.55 and 0.98 t ha 1 were obtained at 0 kg N ha 1 in 1999 and 2000, respectively (Asadi et al., 200 Sweet corn fertilization research has been conducted in Florida for more than thirty years. During the 35-year period from 1962 to 1996 yields have increased. Sustained high yields can be expected with fertilization practices designed to supply crop nutrient requirements (Volk, 1962; Robertson, 1962; Rudert and Locascio, 1979; Hochmuth et al., 1992; Hochmuth, 1994; White et al., 1996). Hochmuth and Cordasco (2000) summarized the field research of nitrogen fertilizer application in sweet corn production that occur on the mineral soils of the north, west, southwest, and central regions of Florida. Of the fifteen summarized experiments, fourteen resulted in optimum yields with N rates at or below the nitrogen fertilizer application rate of 168 kg N ha -1 However, additional studies are needed to evaluate yield responses to nitrogen rates above 168 kg N ha -1 Plants fertilized with 190, 381 or 526 kg N ha -1 on marl and rockland soils resulted in yields equivalent to those fertilized with 168 kg N ha -1 Split N application increased yield 14% in a 1962 experiment compared to yields from plants fertilized in a single application (Volk, 1962). The remaining experiments were fertilized with the split method, recommended for un-mulched crops where leaching and fertilizer burn might occur with the single application method. 32

PAGE 33

Nitrogen recovery was improved when fertilizer was banded in the root zone to one sid