<%BANNER%>

Detection of Turfgrass Stress Using Ground Based Remote Sensing

Permanent Link: http://ufdc.ufl.edu/UFE0021869/00001

Material Information

Title: Detection of Turfgrass Stress Using Ground Based Remote Sensing
Physical Description: 1 online resource (96 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: Environmental Horticulture -- Dissertations, Academic -- UF
Genre: Horticultural Science thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Many turf managers and researchers still use time and labor intensive techniques to manage and assess turf that originate from decades ago. Many of these management practices have been proven repeatedly to work in a variety of situations to assess turfgrass stress, but may be time consuming and inconsistent. However, there are new developments in ways of assessing and mapping stress that increase the efficiency by which one manages and assesses turf. With increasing water and other environmental restrictions turf managers and researchers need be aware of technological advances in their industry. Two field experiments were conducted at the University of Florida, West Florida Research and Education Center near Jay, FL. The first was the detection of leaf nitrogen concentration and soil volumetric water content using remote sensing technology. Plots were arranged in a randomized complete block design and treatments were arranged in a split plot factorial design with four N levels split across four irrigation regimes with three replications per treatment. The second experiment used remote sensing technology to assess biomass production created by applications of plant growth regulators. Plots were arranged in a factorial design observing three different PGRs, each at four incremental levels with three replications per treatment. Spectral measurements were taken from two remote sensing devices. A ground based vehicle mounted optical sensor Crop Circle (Model ACS-210) (Holland Scientific) which produces its own light to compute a normalized difference vegetation index (NDVI) which is calculated as (880nm-650nm) / (880nm+650nm). A hand-held hyperspectral spectroradiometer (FieldSpec Pro; Analytical Spectral Devices, Inc., Boulder, CO) with a spectral range of 300 to 2500 nm was also used to collect reflectance data. All reflectance readings were taken in full sunlight between the hours of 1100 and 1400 Central Standard Time (CST) to minimize variance caused by solar radiation. Plots were ground-truthed using time domain reflectometry for soil moisture content (VWC) and clipping samples obtained from plots were dried, weighed for biomass, and tested for Total Kjeldahl N content. Visual ratings were also taken. The results from the nitrogen and irrigation experiment indicate reflectance data best modeled soil VWC, visual ratings, and some treatment effects from N. Highest correlations for VWC were achieved from PLS regression on ASD data (r2=0.71) and the Crop Circle device (r2=0.70). The results from the PGR experiment modeled the biomass reduction expected from increasing rates of flurprimidol and trinexapac-ethyl compared to the untreated control, but not ethephon. Similar trends were observed in NDVI and leaf area index (LAI) indices computed from Crop Circle and ASD instruments, as well as several stress indices computed from the ASD device. If remote sensing technology could be used to adequately assess turf parameters, and was coupled with the global positioning system (GPS) and geographic information systems (GIS) technologies, irrigations and chemical applications could be applied on a site specific basis. In this setting, turf managers could potentially reduce inputs thereby reducing cost and environmental impact.
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.
Thesis: Thesis (M.S.)--University of Florida, 2008.
Local: Adviser: Unruh, Joseph B.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0021869:00001

Permanent Link: http://ufdc.ufl.edu/UFE0021869/00001

Material Information

Title: Detection of Turfgrass Stress Using Ground Based Remote Sensing
Physical Description: 1 online resource (96 p.)
Language: english
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2008

Subjects

Subjects / Keywords: Environmental Horticulture -- Dissertations, Academic -- UF
Genre: Horticultural Science thesis, M.S.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: Many turf managers and researchers still use time and labor intensive techniques to manage and assess turf that originate from decades ago. Many of these management practices have been proven repeatedly to work in a variety of situations to assess turfgrass stress, but may be time consuming and inconsistent. However, there are new developments in ways of assessing and mapping stress that increase the efficiency by which one manages and assesses turf. With increasing water and other environmental restrictions turf managers and researchers need be aware of technological advances in their industry. Two field experiments were conducted at the University of Florida, West Florida Research and Education Center near Jay, FL. The first was the detection of leaf nitrogen concentration and soil volumetric water content using remote sensing technology. Plots were arranged in a randomized complete block design and treatments were arranged in a split plot factorial design with four N levels split across four irrigation regimes with three replications per treatment. The second experiment used remote sensing technology to assess biomass production created by applications of plant growth regulators. Plots were arranged in a factorial design observing three different PGRs, each at four incremental levels with three replications per treatment. Spectral measurements were taken from two remote sensing devices. A ground based vehicle mounted optical sensor Crop Circle (Model ACS-210) (Holland Scientific) which produces its own light to compute a normalized difference vegetation index (NDVI) which is calculated as (880nm-650nm) / (880nm+650nm). A hand-held hyperspectral spectroradiometer (FieldSpec Pro; Analytical Spectral Devices, Inc., Boulder, CO) with a spectral range of 300 to 2500 nm was also used to collect reflectance data. All reflectance readings were taken in full sunlight between the hours of 1100 and 1400 Central Standard Time (CST) to minimize variance caused by solar radiation. Plots were ground-truthed using time domain reflectometry for soil moisture content (VWC) and clipping samples obtained from plots were dried, weighed for biomass, and tested for Total Kjeldahl N content. Visual ratings were also taken. The results from the nitrogen and irrigation experiment indicate reflectance data best modeled soil VWC, visual ratings, and some treatment effects from N. Highest correlations for VWC were achieved from PLS regression on ASD data (r2=0.71) and the Crop Circle device (r2=0.70). The results from the PGR experiment modeled the biomass reduction expected from increasing rates of flurprimidol and trinexapac-ethyl compared to the untreated control, but not ethephon. Similar trends were observed in NDVI and leaf area index (LAI) indices computed from Crop Circle and ASD instruments, as well as several stress indices computed from the ASD device. If remote sensing technology could be used to adequately assess turf parameters, and was coupled with the global positioning system (GPS) and geographic information systems (GIS) technologies, irrigations and chemical applications could be applied on a site specific basis. In this setting, turf managers could potentially reduce inputs thereby reducing cost and environmental impact.
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.
Thesis: Thesis (M.S.)--University of Florida, 2008.
Local: Adviser: Unruh, Joseph B.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2008
System ID: UFE0021869:00001


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 E20101108_AAAADA INGEST_TIME 2010-11-09T02:43:28Z PACKAGE UFE0021869_00001
AGREEMENT_INFO ACCOUNT UF PROJECT UFDC
FILES
FILE SIZE 1053954 DFID F20101108_AACGUH ORIGIN DEPOSITOR PATH frank_j_Page_21.tif GLOBAL false PRESERVATION BIT MESSAGE_DIGEST ALGORITHM MD5
ef45f58628de91bf0d521c4a2013c9ac
SHA-1
5d0f77a6b70d5f033593f6a112c5a07e59c05d1d
121667 F20101108_AACGTT frank_j_Page_62.jp2
2a1f0b511244428144826ccab775c71f
b4aea716c05f9fb1e71e1e24d71baea0e1aa7e6a
117503 F20101108_AACGTU frank_j_Page_65.jp2
019f53ff8c549f06b69967f9fa8749cd
6e5644aa75c538b774dacff05cf36e85400efa3a
F20101108_AACGUI frank_j_Page_22.tif
fe63da7c5fa67b4a0573fd727515ce99
438156ec9f77d9a31967c8ec3e173ca2ff31c0a9
102575 F20101108_AACGTV frank_j_Page_69.jp2
6762dac7e504ae7f0a7bd4dc4c62f55e
673ac75e1e0dc64ca9c75fb29bd3ce0e28b12760
F20101108_AACGUJ frank_j_Page_28.tif
faa6f0ae74e478b21d0d7a5822dcd0d9
7b80d28801777bd4b97e44590aa4c6fca8f17d21
104408 F20101108_AACGTW frank_j_Page_71.jp2
519e38e2d880a25b783523cfbb6dffcc
ff35ca16e0625575adc80108291eccec1b1661d4
F20101108_AACGUK frank_j_Page_31.tif
31667345389b9086e25f1d84198f2f33
ef0f6d15e12f2f1facfb2ded6fc06ed9348db515
12056 F20101108_AACGTX frank_j_Page_73.jp2
ea5ab6dbd6cdd48056b7450d9eac07c2
70fc2f4fae8051dc5973203cce944d23d240e3cb
25271604 F20101108_AACGUL frank_j_Page_34.tif
86f1d806380b363d1182830533d228bd
b57d7969f4618034cd87c70f78f2b3214bbc7802
74853 F20101108_AACGTY frank_j_Page_74.jp2
49c8f15557425051f007be607384962f
712b4c67a17044663d659f02dd9c4b9270da5252
F20101108_AACGVA frank_j_Page_95.tif
0cea14f3de9ea6caf732ae8221140f0e
f0937f73e233c0393808336865c93e2a7cd518eb
F20101108_AACGUM frank_j_Page_39.tif
1523c156eabe8705ccda8a5e5351f817
259c171bb6f45d99c0a387aae2419e1d6e23748d
72999 F20101108_AACGTZ frank_j_Page_76.jp2
7f85880c4184b351e7553d1b81f3cfd7
044ff07618bffb5fadc4bfe5b5a7b3878f698f1f
984 F20101108_AACGVB frank_j_Page_02.pro
e18c20e4d615083eeac8f8c6919d7e13
1eb43476bdc7d7be5a2b1fcc5fa26b89b5eed5e7
F20101108_AACGUN frank_j_Page_40.tif
8bae413b07253c9298ed80c84dc191c2
b9f59297ef417d05df48833e2541bac9c1975a7c
94491 F20101108_AACGVC frank_j_Page_08.pro
efb7e8eb9a8e04636ca4cb8330d7a12c
bb42e75fcb01f7e8eb98cbbc76542981f798274e
F20101108_AACGUO frank_j_Page_41.tif
b2b610dec74961f34dd528700e70de42
3746d8a75ba0a8ed8d41a07c73c3d7218126e876
63999 F20101108_AACGVD frank_j_Page_09.pro
fa080df68696241a2a9a57eda921a089
575ca6bba7afbafaa4ae86cdb707bf2b37e4fedf
1054428 F20101108_AACGUP frank_j_Page_47.tif
23d262db9bb14c1d21460782f4e7de7f
8b8bc750ef2cf6be52333c863fbb64df8e8c02ff
49492 F20101108_AACGVE frank_j_Page_12.pro
244193277108e9125578581ad134b5a4
c2c05c2ee7b1fa195bf7c284128d2ca542f087d5
F20101108_AACGUQ frank_j_Page_53.tif
966cd7414221741d88d196536d7861d9
0825022f114d81a3cc7732ef92da01967290f2f7
55284 F20101108_AACGVF frank_j_Page_23.pro
cd0dd12484bb37b01e7dffcb5b29154a
ba6fb22f028e47e8c102942a7145b495bbfb8689
F20101108_AACGUR frank_j_Page_55.tif
7e521fdca5cce17dcc7a272bdeffa66c
4220082dd6c1cec86910e24d74ffcfdd3f5c0c33
53664 F20101108_AACGVG frank_j_Page_28.pro
ecadfe2a7b9ac53e20f5c14783ea0db0
a765b31d98cd29428d6d7f4433454086723595f5
F20101108_AACGUS frank_j_Page_61.tif
55b0dec7117cb5d6b350c37b32299a85
654dbd2b50550eff72235ac3c24de7e181700983
52820 F20101108_AACGVH frank_j_Page_30.pro
11a324fa4c3c75cd62be92272777b475
67a280778309f81ecaa9e286820cbdfc19ec3316
F20101108_AACGUT frank_j_Page_64.tif
dc4ae5d9c5ef9a857236ca238418a0af
92aa0e582d781c06911cdeccf8d045b1f71f78f4
56890 F20101108_AACGVI frank_j_Page_34.pro
1a99a31a042bb065bca2d760d55e2506
f8178eddedd9100abcc13f58cbfb191d2c8d024c
F20101108_AACGUU frank_j_Page_65.tif
8c62acfeee97ba3f184255e0406a56aa
ea65b2dadd32655c45ef1d229ba15b1dbfb107b7
55037 F20101108_AACGVJ frank_j_Page_36.pro
4044e3e1b8d8bffaa7623534b8610658
2d5a9e48fb2dc1e7c533c802ac9b93bf2023b933
F20101108_AACGUV frank_j_Page_68.tif
b6c235ea4425140f6a84b29966008ee8
c5fe2e453224d613339e269f75eff78dc23c9a15
48640 F20101108_AACGVK frank_j_Page_38.pro
fdd788752022983b3f58eba010cd862e
e1b8ff56c821561065e12ae6680091514a645a4d
F20101108_AACGUW frank_j_Page_72.tif
0259f876c2ef2d6ffccd31ce49662cb0
96c7948e2c5cc49a534cdd2cce8a3fa9b6f05de9
52149 F20101108_AACGVL frank_j_Page_42.pro
cc504ddd999d488f462476aad2948cc4
902c434d062907a2b53c06bf415de37a40c9e18f
F20101108_AACGUX frank_j_Page_77.tif
7b33dcf8e60ef3da2b700cbc4440b748
ef0f50c0c89c82b7d15e37a4bc7c9d4752103a98
3922 F20101108_AACGWA frank_j_Page_08.txt
06dad9601d59491671288b14485a078a
b3e2348375601097f9216a97d56093e3b9de7fde
50796 F20101108_AACGVM frank_j_Page_44.pro
9572bb601f438296b3379e9bfc00f9e9
2d1bbbfbcfe5bb76666c533fd1f58410a744d619
F20101108_AACGUY frank_j_Page_79.tif
adcb48d8e40a54feb57516776a8f59ba
2b9bcceb45eae624087a8901294c827b1a177126
2666 F20101108_AACGWB frank_j_Page_09.txt
d555880d459b8b2ccf4671468fc63c8e
4476d13a3e4a6f3c4662725d1743786211a3a25b
38091 F20101108_AACGVN frank_j_Page_50.pro
55e80c25862aed5aca40f5cfff0cc02e
dbda513ff6e762a145385f5b9dc984832c1ba6d2
F20101108_AACGUZ frank_j_Page_86.tif
e0702e08d87769a13550bb2f8a21f31d
04df444b88c92d95fc3e0e828a667851d10887b6
1960 F20101108_AACGWC frank_j_Page_12.txt
05e83c6a142a7744a91bb76356ab9ecc
dda7e34a58dc0b3450b2139324462091fda155a8
34694 F20101108_AACGVO frank_j_Page_54.pro
dd12375e4d17b6049d0cb069d6663ac3
74dba37f66ab78d99622c80619ab3f2f79cac633
2311 F20101108_AACGWD frank_j_Page_17.txt
c3da29c37c88fb7783318b977c1ece9b
68ead6ea54403616368de0d2c511a8fa7aca3bca
17754 F20101108_AACGVP frank_j_Page_59.pro
9f736e874400e326a7d4b1f96e24acf5
be24e3f34ecf63c88f9e255e81df09c74e33830e
2100 F20101108_AACGWE frank_j_Page_25.txt
4a49bd6762a9c23dbbc9b6f096ab722b
02ed254ce73674e11983abc81c44e3beac0a4d16
49694 F20101108_AACGVQ frank_j_Page_67.pro
2ffd1469ae4f7ab7534813a17c41450d
c68a91ab6877322b2554ca604a873c23a8475f1f
2195 F20101108_AACGWF frank_j_Page_28.txt
340125f9a15d08ac0bf4b6ace1b5c93f
8d7cd6c471d321688f471416b16ef88ad59e61b0
54520 F20101108_AACGVR frank_j_Page_68.pro
052d704fe588f5d3913ac20d46ee3ce3
12b2ea085dbd2e8de627da7aaff10123453f6e25
2162 F20101108_AACGWG frank_j_Page_31.txt
66747297857a6fce6ac740ea813e667d
e08df112f3db761b1a24738de6ab91d4bf5552d5
47259 F20101108_AACGVS frank_j_Page_71.pro
025e60c3fb744ff94cab9b3f5b86d01b
6da4ad804efae02a6ec739eeb35199db6d8f51d7
2118 F20101108_AACGWH frank_j_Page_33.txt
a758275f939f59d24e6bb0cdc45697cb
955f3fff2dad721fffad60784904c1b5920af318
44078 F20101108_AACGVT frank_j_Page_75.pro
5b81b9c6bbdee0546b766a6e9aa162c8
27a311bbe67722a2d54e2d0e45672db541c881ef
2091 F20101108_AACGWI frank_j_Page_35.txt
cde0eb2132cea94cc3e22d786442d7d5
cb1b6b459bcbe2c2f515bb81a44d2e5da1ad4f55
27956 F20101108_AACGVU frank_j_Page_77.pro
578988b080ca5caae9ee50365c3d93d8
a31795e2c77fe868d495741df96663d4283a1da0
2159 F20101108_AACGWJ frank_j_Page_36.txt
508a281bf1ed1b7c89979c4d525635da
8b7736d6f3ab52c15fe203bd0b7333eb1b1127e4
6514 F20101108_AACGVV frank_j_Page_85.pro
cd6dc3b9e1a7142837debeef29dd1b42
58d58772408eccc4d8e6ca73a75044bd26461815
1931 F20101108_AACGWK frank_j_Page_40.txt
93aea7a6a0d990915d62d1b267c614a4
3bd6aaeac34f05570b8c84341c5079d58c570f39
32401 F20101108_AACGVW frank_j_Page_87.pro
19b9ffd772a05a1a98107d7db1afc0b0
0187d20f6ea3e351052c4ab3ad1e43555670d1e4
2014 F20101108_AACGWL frank_j_Page_41.txt
ef73eb2e813906eca36723c9d550b6eb
0e2358b0501f143b42dbd07853a9e6f5de6cce7d
21053 F20101108_AACGVX frank_j_Page_91.pro
a56c6aee5f5f493700de361966f94a54
d96b753af2e606c1dcd0ddb57d133a41992c9aa3
2123 F20101108_AACGWM frank_j_Page_44.txt
bf48a416e1a7f5239bbc81459bec55fa
abc814cc50548dc1536c802a6bd76c97029fdd0f
61142 F20101108_AACGVY frank_j_Page_93.pro
72c9e35352b74f0391ce33cc3711602e
4947b88a2e2f4926eb675c60939ed45c08846906
2419 F20101108_AACGXA frank_j_Page_94.txt
61ed1f50c557d755ef6a8d07196b3b68
5b37287b77bf72cd089ace560b79477e7b1ff4f3
1690 F20101108_AACGWN frank_j_Page_50.txt
607514ada5dfe82d05b1bc5a56953770
67a4a87a5f8ffa8a55d47b8dfe903409955c8c79
426 F20101108_AACGVZ frank_j_Page_01.txt
feb41c78f56baf98cfd1a21487af538a
6bc49659e0954e7693f5b355605079c258d705a0
2321 F20101108_AACGXB frank_j_Page_01thm.jpg
ec26b1effb1b1f78b8ad213f0939598b
7d694900f8b8f1e9569c203e68e57fcecccdd5e7
1790 F20101108_AACGWO frank_j_Page_57.txt
df4331e5abe5b2b472028c7e1f9f297b
a9596175c1dbe76f8c54b6f144fac734944cd762
8878 F20101108_AACGXC frank_j_Page_76.QC.jpg
ae621faacd1696efe5438f5d706cf63a
940b273941d26b519dc04fd0b3e69939667a37a7
1027 F20101108_AACGWP frank_j_Page_59.txt
099eb52d63d901549c5b4499ba369c05
4fe9c7c1988a21e11a6d29f1b7676f7d28a60fe5
23383 F20101108_AACGXD frank_j_Page_40.QC.jpg
308889112b639f1ddd8b1bde3f5bae6f
a5e155931af21d1816e3b11a5151f329f8224c11
2155 F20101108_AACGWQ frank_j_Page_64.txt
dd7dafba63e6d1e070de31164d0e88cc
dbbf7a65bcfe5915c49787ad7434ad77b13bed17
6904 F20101108_AACGXE frank_j_Page_64thm.jpg
61cfbcb2ad58fd9824342f2405060b91
c1102121e25e8333e8cd617d95dc99d3d5aa3805
1853 F20101108_AACGWR frank_j_Page_75.txt
2dcbdfdff73de9118eb16819d8b8cc5c
59498e763ed7d8b58757bed71d9b4f6dc5f746d9
24863 F20101108_AACGXF frank_j_Page_82.QC.jpg
b51f1493a5aba43b150de81b21d4132d
935263a425b625f43f7584e1554bf15e64b9ed64
2166 F20101108_AACGWS frank_j_Page_82.txt
96b31157e2d18ec9dce9ba142a128348
e3d2cd89dd142d6eedb6f0b3b389b1c8a1065455
25288 F20101108_AACGXG frank_j_Page_21.QC.jpg
140e0de0411d4ce1efc661e3ecc0451c
3cd0c68e81aec6505ba348bc4ce893b7d6457818
223 F20101108_AACGWT frank_j_Page_84.txt
7b0c52ea57f60ee01f81a3b52d165dcb
664c88a4ba70d2c97fb8302804b60f11d287bf46
F20101108_AACGAA frank_j_Page_45.tif
a73e86023caaf1a7236a2914bdce692d
1d418622230fa7baef0b55610d44c9fb44c2c7fb
6417 F20101108_AACGXH frank_j_Page_71thm.jpg
3d7e84a579017f00483c807da2d86bae
d9031fddfd16b9fbec5b0c90b3ee1643cb96f4af
872 F20101108_AACGWU frank_j_Page_86.txt
42c655727e570ab955c637ae60c7d7f0
632312d0dc6eb61980616b61d178c771a7b626ae
114756 F20101108_AACGAB frank_j_Page_16.jp2
8f774af45f9c921aa80dcc66903afaef
54d8f48f0dcf62d342208f01c330923834a2185b
25982 F20101108_AACGXI frank_j_Page_64.QC.jpg
29cfa7da25eac07a95892aaf3a2ccd43
e5ed1b3b5f65c572ca8921d8ca6b022d306aca57
1856 F20101108_AACGWV frank_j_Page_88.txt
ecb6226e0dc4aa9a0b1a18562d40901b
777619a33f8f7b4100693fade5d105b0693ec4bf
1975 F20101108_AACGAC frank_j_Page_67.txt
0fbde4cc6b4a83256c3edefc8d78b41e
c6e41ef950fd23d3db955f7a61078dccad2cd1a4
27705 F20101108_AACGXJ frank_j_Page_92.QC.jpg
91e26cd5d3ea97edc20f90ee1076cd03
3b6cde59c9be6025740271fefe7084750b7d0c4e
407 F20101108_AACGWW frank_j_Page_89.txt
bacc2aa8503943f6b7a2306b1bc967f1
d6d21845eeb8af18520380e52b403ffa0e739099
52909 F20101108_AACGAD frank_j_Page_82.pro
bd03848728c53ae50f36cbf4c4eb5fc7
71e670ff30123c11d115cd244db956c73de1fdae
6536 F20101108_AACGXK frank_j_Page_70thm.jpg
75acb4e4ded08142050dc4e6ce25a572
b63d3005d4dc13a7b11f0d551a704aabfbce906c
1004 F20101108_AACGWX frank_j_Page_91.txt
4aa8bbaecd92cf349f9baee6f5248d4d
5b6ec6d1a3dcf7176aa667b4598e542cc4a08fd9
6050 F20101108_AACGAE frank_j_Page_11thm.jpg
6da0028dd090ce9f900208602fbf3538
db9a38f0af71ea779c269840a0557ca224716725
6922 F20101108_AACGXL frank_j_Page_63thm.jpg
66d728bac7c9653542081f0f2ddfa44a
6b7a908ac8171a7502aa10c57fd080044f6e11ca
2438 F20101108_AACGWY frank_j_Page_92.txt
79c00004ce901bdc63f1bf32a9162fb0
a6828a9d51e26e9a8a059b785012520b6546ff82
68727 F20101108_AACGAF frank_j_Page_71.jpg
10eea73cd31781d1822141343faf9873
8e416d7af8cd3a8f9e602f951c6b4bbb7562b9b2
10620 F20101108_AACGYA frank_j_Page_51.QC.jpg
70f1266fce00f66935d69765cad3286e
d7349dc0583694aeb0dc05c282fb64962b2935b2
7782 F20101108_AACGXM frank_j_Page_58.QC.jpg
9d527b974e14f0bdc0db5da6b4d76e89
1a55abf547ab93a63b8a94b0042d019426a87300
2472 F20101108_AACGWZ frank_j_Page_93.txt
005063ecab1641a6aac8da2975731603
34af9b74d51bfa547565badfa1360d6ece91c1f6
F20101108_AACGAG frank_j_Page_50.tif
ea19ce4f7d098bfa4e4e330805678f88
e7bacfd42d753366839a6f016b1c958687c1d9fd
7926 F20101108_AACGYB frank_j_Page_80.QC.jpg
900e01faf397e4c4ac70c59db31d4ac4
acc493cac6d9699d8eab738b82dded16eab2c1ba
3156 F20101108_AACGXN frank_j_Page_86thm.jpg
b8970dd69f437fda2d8a64d40dfb3b05
b52cc39c60b290663307fb957c203eb7e14b5def
F20101108_AACGAH frank_j_Page_27.tif
dba10638e168cd44af9cafe00e817139
9e205adeb9b9ab3325743c4c6a7e3bf1b8d22355
26667 F20101108_AACGYC frank_j_Page_17.QC.jpg
7c38288031c8aceb36b8fbc8f2b467db
9381a1d414a82dd579542b733fe2557435dc6f0a
5858 F20101108_AACGXO frank_j_Page_95thm.jpg
e92082ea102a7aa94278a119702e9a35
60b1e2e0b9af505c5a0b9b4b27e87f67ef24f7e1
143438 F20101108_AACGYD UFE0021869_00001.xml FULL
a343ac2669f92bab1e7114b8e2b8d213
c103d445ea69a1d5ba952b71d87b8094e127daec
10149 F20101108_AACGXP frank_j_Page_45.QC.jpg
0bbb86164c6ecf0cca28de84159d7fc7
cae9b5c2fd9e74b46baad67fea5632bb317e8764
34271 F20101108_AACGAI frank_j_Page_55.pro
2fa4b2a482a2a04d23e0e9d743b51d3a
128bf5cdbed1409e067b79ec6f0c46962cd87dd6
6871 F20101108_AACGYE frank_j_Page_01.QC.jpg
66335ae0d753bb99fca06828ceb3139c
270a10883181d86d9e795f343150e3bac91a40a5
6755 F20101108_AACGXQ frank_j_Page_19thm.jpg
09e0409704c9d3fb7946153d3e1dde7b
78b929f1fa5ed564a46f12726125a7fb864837b7
F20101108_AACGAJ frank_j_Page_96.tif
4274d80a2be5d5542d3f9600b7952bf9
21783337244c66b76a9f30b8e0e237a6a50c9bfe
7268 F20101108_AACGYF frank_j_Page_17thm.jpg
dbad4f3aff1f3dc78756d7c1bb342c5d
3a7d7290c4cbe2b98aa71a1cb66a5be1508da77f
6998 F20101108_AACGXR frank_j_Page_66thm.jpg
8eba8cb219c1d974da9f660d5db08352
a74ba4f1516182788fb4a2cb11ab648b53ef5783
6581 F20101108_AACGAK frank_j_Page_20thm.jpg
9159dda620977a85a300363ac00e9009
1dde100f532df0646643dbf0f051287424888d30
25593 F20101108_AACGYG frank_j_Page_26.QC.jpg
17fdf536afddf7edce51d28885680072
10260be19e8b794c0f8666a339d0e744d3de1c5e
18455 F20101108_AACGXS frank_j_Page_03.QC.jpg
fb522f45bd9451faefa8127cdd2e9f1f
40ddc017f718dcb64de6119def1ce43250e97585
6624 F20101108_AACGAL frank_j_Page_12thm.jpg
76b3ad72ea170ab8b296f74803073463
c2cbc03d1ab07fa76b1c240ec6caafb72c3ee2f6
28342 F20101108_AACGYH frank_j_Page_34.QC.jpg
45a579295060578071d6077e729bd160
69043067592c8b376b046c384fab7b929d78ab37
24939 F20101108_AACGXT frank_j_Page_63.QC.jpg
d73c73549d7c0111057495ab02b437f8
59f3ab59c18fbcc8c70f0bbfa028c3244b750b33
33327 F20101108_AACGBA frank_j_Page_54.jpg
9e29639006b9a0794f9b697b38dc8f04
a915a45a4887b4c776c9f89ee63138f42987f6fb
1248 F20101108_AACGAM frank_j_Page_43.txt
fdbecb57e998a76ea818bb12bb357889
b7b8ff8e856307317d1f785419525bc2ed4051d7
2912 F20101108_AACGYI frank_j_Page_51thm.jpg
0e0adb517e125b0223772666dd28903e
a6fe757d0f159212a6a1c3cfaf5028c3ea11481a
7240 F20101108_AACGXU frank_j_Page_62thm.jpg
7597a683b16c3beba9d1a8ad8e269dbf
e6104ee0271902bc10bc7219b0f4e732bd610d63
39603 F20101108_AACGBB frank_j_Page_46.pro
8444b73ea71f3a9d252cf3dc050208c1
b8ab5745f24dca27a9487b9a614cd231a4013a57
24308 F20101108_AACGAN frank_j_Page_67.QC.jpg
3bcd1fa15b6b9cd994bc4a52dd3da7e6
ed42177a32c5241b32811c217046cb2483f50cd0
8885 F20101108_AACGYJ frank_j_Page_54.QC.jpg
5f2c80fa8ccbb96e2fa756d8f1b7dab9
563f2f5e4bcff01293ca87c9ff19259c16d02ca4
25799 F20101108_AACGXV frank_j_Page_23.QC.jpg
31924ad512389e135cc0c0895cd01b72
1e4b8682d5d620e9605d09b5379d59753b2a1542
72537 F20101108_AACGBC frank_j_Page_67.jpg
a11e472cb8af6cbdc0ea55066ebd7eee
40b4fe85635c1988eed8430d0c493f6950f2eab3
75662 F20101108_AACGAO frank_j_Page_82.jpg
8b93c6337fe11eb36afb6d7ec8c8a180
1a28b706f0e23291271bbe9a9d8645fbcbfd8017
10058 F20101108_AACGYK frank_j_Page_55.QC.jpg
76f11deb46d809aac5b1fca73c9dd176
edf528489a132c9efacb083b811d0c56a4058732
6792 F20101108_AACGXW frank_j_Page_65thm.jpg
1851adfe97d872dc3d5cf3f5f8baaa98
030b0b4a9765dda41676840bcc1fb46aa8a91aff
24505 F20101108_AACGBD frank_j_Page_58.pro
11da164b01db4ffa21435c33cbe1f1cf
aaf41abed6fd65992ffe8718e6254e73bff33c63
3492 F20101108_AACGAP frank_j_Page_06.txt
74921d5931c7d649da0f495a12840950
0829ca00041ffbe8f5986d31e8e7e03c26909158
2866 F20101108_AACGYL frank_j_Page_61thm.jpg
f987a754ada0d7937e0e68c02255a789
c8a5510137916aa2d45fdd16b448a085e51d6bb5
1879 F20101108_AACGXX frank_j_Page_84thm.jpg
743105c799f540de5427ce3229c8c13d
40efbf7e56398d86e10e838c6c6d96e64870df03
9549 F20101108_AACGBE frank_j_Page_74.QC.jpg
b609574952c97006bbab4a40c7a30a3e
4bb06a3ca100e5bdc8eb364bf2b66f0e568d7913
3619 F20101108_AACGAQ frank_j_Page_44thm.jpg
ffdc2221c6455f4bcf0f2bc6cfbcdcce
2ace979c4f773a90d7a20a038fd5e497218f6ce3
23086 F20101108_AACGYM frank_j_Page_70.QC.jpg
405d864ee7ceda81ade3377a2592749d
e418cba64921f82b8004edb1cf9f25800fa9f189
5976 F20101108_AACGXY frank_j_Page_13thm.jpg
5b6db2f39a57815fd49d8b92bb79da76
996ad21708b6f6fbfc20f53b6aee598f6681fcef
76763 F20101108_AACGBF frank_j_Page_33.jpg
a54c108b0559859119bf081d3ec0753d
6c1f462db6fbe2e44c53068a77c50341ddb1562b
F20101108_AACGAR frank_j_Page_29.tif
4e4c3ad303616f21006960d6ffacaa75
d23cba8eabe047a9c7eed57b069dec36c639d050
F20101108_AACGYN frank_j_Page_79thm.jpg
1b4ad17b8215615ddca67f9b2216e819
fc9f9ba18b1fc69b31a1369a646600b74bd50252
25384 F20101108_AACGXZ frank_j_Page_39.QC.jpg
22914b90f6154aac083900bf36099005
66d48888408abd6d233ace83d6687f79f3bc02ad
25025 F20101108_AACGBG frank_j_Page_33.QC.jpg
cb49b0bbc8fbf50839eccdf7dc084eff
92bde38d0a87a2e7b37f07e30795b2440117aac8
6754 F20101108_AACGAS frank_j_Page_79.QC.jpg
24bcdebd54e689afe934934051d579cf
62f94757f22cee9bd4ed743478d9e143aa2c1bdd
27674 F20101108_AACGYO frank_j_Page_94.QC.jpg
d6b710e469a98d124943dfbf2e25b734
6fe41f2d2bbdd28eed7c5992cfd16c50b1ba9851
44402 F20101108_AACGBH frank_j_Page_86.jp2
04cecb1418556d402c9ea783441c37ce
79259a14298afe5456ca90de86c7c588c76caee4
5479 F20101108_AACGBI frank_j_Page_84.QC.jpg
ca718263ad4f19c09b29d421ee87f409
535b14bff3f7049c5dfc6fcd93bf8df1d2e20f31
F20101108_AACGAT frank_j_Page_76.tif
7205c5a698a4e66f3c23c6957e49381b
36618ee7ec0f0fd7a156bf9a07af89d5d48b90a6
24615 F20101108_AACGAU frank_j_Page_24.QC.jpg
367e059320f468ed233c1ab85dbf9725
8a0fbc9512497c1a73b6cc500a462119fa32d9e4
10192 F20101108_AACGBJ frank_j_Page_48.QC.jpg
5f16336b0ab0a3a6d8350ec34cb601fc
4e5b1e1071c50cfd51f7feaac7434866c94a3485
6800 F20101108_AACGAV frank_j_Page_25thm.jpg
ffef39dcf6293a80246993cd83e87572
b550a8a8158336c09002b7902ce24a16d25c2085
75029 F20101108_AACGBK frank_j_Page_16.jpg
bd978dfbd65f664f3999d846a7b7fbee
9c3967fe3560b69ca4432b8b455effb6b554351b
2093 F20101108_AACGAW frank_j_Page_16.txt
bf8eede5848e6bdf5d999204e2acf5a9
5aa25ea32a3bb26d6b3302bc2a5c5fcf7fcc5ee1
11368 F20101108_AACGCA frank_j_Page_50.QC.jpg
2b2b7df76131ff5327dc43065d27d5a0
6291ad8cf0c9f55009e0e7373c31e03a6911689e
1196 F20101108_AACGBL frank_j_Page_49.txt
d03a52355e78fc8d3c6f05e205139a91
322d4520a94e764d0e99d6fc04ebc97a8e0071f9
1842 F20101108_AACGAX frank_j_Page_95.txt
623ddb49d236a95baaf05dab1d5935aa
e994113a73e638e21b06e99786996d32df19416f
F20101108_AACGCB frank_j_Page_03.tif
213130fe018cd4d4ea0ea81841d0e7b6
84803d9cda4513066af15967454d0c59a599beaa
1171 F20101108_AACGBM frank_j_Page_77.txt
b818eee9ebc6ad063d2c6dbf4db675c9
8c540404b3e80c6cc77d1a4290ecdb9976909c3c
7537 F20101108_AACGAY frank_j_Page_78.QC.jpg
bf17fcb609ed2b85945e43678e30422a
063481128db7859031825c24e3c0765cb73dde9e
F20101108_AACGCC frank_j_Page_15.tif
aeccf9fff5af18b20a2eec70aa69a1ed
a2fdaeb384bdc895ce19ae4682d1e925d85e5b20
54756 F20101108_AACGBN frank_j_Page_64.pro
9174054152bceabcdeab15f63aac8beb
b4c396a8e4fbb36730c55d7473e364dcd8e15470
5719 F20101108_AACGAZ frank_j_Page_04thm.jpg
4ce9b28feecab189a12513b5ca2e7564
01797bc8feecd51b3eaf4f4295e2cd923cba9d6d
58206 F20101108_AACGCD frank_j_Page_17.pro
c2f5af42ca14a464f3629306fd27244d
1c11f9a69fca641bd054f4f07f14875e1bea77c9
23680 F20101108_AACGBO frank_j_Page_09.QC.jpg
68b2523ebce269608d2f5c5a714ae23d
dedde2a870fe7a518c00a12e226da4ad5b10ab58
F20101108_AACGCE frank_j_Page_84.tif
4d4b2b508dc1e8a02a3a6204dfea07c6
e5cd7d4bf0a675626effdbe2db2b2e2d8550e32d
F20101108_AACGBP frank_j_Page_49.tif
46df0c3005e05835ff40a996afbfdb96
1631ea5de395bbd4888636627a90bd5c0eee3973
F20101108_AACGCF frank_j_Page_08.tif
dfb3a4f8bf6cea6faf325ac2317a2de8
1383fac43607eee9b5ec0286b96b102907ca3e95
31042 F20101108_AACGBQ frank_j_Page_49.jpg
d0e2ba19b612ed7d8ae3009d9c2c61d8
b171093c4fe3d1fe2ad036c234f916f165866234
19195 F20101108_AACGCG frank_j_Page_79.pro
d8dccb674befc35f127892032dbfe759
9e91b7630aefe895037c6febf977e5f2f8680a0f
78639 F20101108_AACGBR frank_j_Page_14.jpg
dde38b7a4b2826f597d2be6e2c0fff01
891ffc60365f0ea62f1c2942fa40100fa59a1cae
F20101108_AACGCH frank_j_Page_26.tif
dcf9fd86e4fae92d40b3c1f1448141e0
dccc70f91db330cb4111e56a978b1f3014580383
6400 F20101108_AACGBS frank_j_Page_38thm.jpg
9945c2bb9fa79e8146b5af33e80be270
4b2c1af273cd00b860ba0ddc0cb5b78e7e371571
8136 F20101108_AACGCI frank_j_Page_77.QC.jpg
82a30b8a016b14ba34f2c2e261f0e7b2
c9b8741c755e18931e45fc2a540fc8ba883512e5
1051983 F20101108_AACGBT frank_j_Page_07.jp2
8d7d67a41337e2833da6159c5bf49088
6accfc3b93c6780a386b03396464e6c964ebd6a6
69290 F20101108_AACGCJ frank_j_Page_69.jpg
4fb0340ff00350c5d4485ea91e3b3cb7
7bbaf8428263fc0675a2c77df3e69850b89eabe5
2993 F20101108_AACGBU frank_j_Page_56thm.jpg
ddcbf7c65d9bc184a5d69a4bd93e828f
3111d0ea208d25ec923f21a96f0bee0503534f8d
1010 F20101108_AACGBV frank_j_Page_52.txt
b7e92c0afc1b83f2c2b85ce14016ead0
756ad72b830969469c22a49e7ec0e5e0797cdaae
F20101108_AACGCK frank_j_Page_25.tif
a5e443735e1fa153d0e0abf447a3c84b
5ea3244f4ea9349e1411f4316a28162b1a9c6dcb
59651 F20101108_AACGBW frank_j_Page_94.pro
dab0c521c00c8983af3ce6a0388756d5
38ce2c0938a6b2e71796a4d15239d2e6bdd2781e
2106 F20101108_AACGCL frank_j_Page_21.txt
3ef6d48e8cd7caff63793ffb56e5255f
340d874d24ce9f09673ead7ef603299726f70d2a
360431 F20101108_AACGBX frank_j_Page_81.jp2
9642ec8d8af0b3ddb78e5766d0ce2127
d12c2ab7ad4b01a678897900e83b6c619eff7d4c
81124 F20101108_AACGDA frank_j_Page_45.jp2
55f6d7f7c14ac2f70d18d4860c51702e
d20d5f7cfdb1c94946c165af7547a876042d71f7
F20101108_AACGCM frank_j_Page_63.tif
1e11569e7ae7067e5efbb1dc2a4e77fd
2c2355ed9b616af4fa8355d43d59be77db26d9b4
24142 F20101108_AACGBY frank_j_Page_01.jp2
c0d7dc4b3fccaf7e2295c36a98556865
f049bd95cfe252b40d3f3861ac7608c874962b04
603 F20101108_AACGDB frank_j_Page_61.txt
3a851180cd12d5ca28857dafc8ed1f07
c45b07c536c50e5ae5b2560dc48f6f35555ebfb4
F20101108_AACGCN frank_j_Page_35.tif
1685bb35f32439bce08ed121bc5959c2
d76c5c92d2148b4dc898a68bac4ac0c455c6dd68
26244 F20101108_AACGBZ frank_j_Page_78.jpg
6527381b1afc15855e9a3e036fcabd6f
c44782d7c98d6815bbd2efe489919a4e3522db62
1006 F20101108_AACGDC frank_j_Page_90.txt
0acb8b0a99ea7c977fff63eb7e802fc6
03702904a9283c96d0f50b03809c0b85b96c4d20
26339 F20101108_AACGCO frank_j_Page_72.QC.jpg
000158a399f4c8aa614821f02d5d5deb
233491e37155049ab37e518156705ef24cbe0656
F20101108_AACGDD frank_j_Page_89.tif
889149eb142a6c42bcd8c19ef3a37cc1
766693a04b319bfccdc068e206560487f4445c2a
25573 F20101108_AACGCP frank_j_Page_35.QC.jpg
83c7711fa15a70f05ef5d09031aacc29
ff1204d9e4d901252058266f3ef74956532e0270
2101 F20101108_AACGDE frank_j_Page_42.txt
9fb87f8e724d209073294df3fb59c827
3702e291546e092f5bfd06898716be3c9501fa3d
F20101108_AACGCQ frank_j_Page_75.tif
e4a35648e891ea07499b7b87464da7f0
db93c8918eac635e745599c5b66de2de7edaa518
55384 F20101108_AACGDF frank_j_Page_62.pro
6bfc89aa672466d97ddd9436083730e5
bf5c2f4aaf58534226aba001db3766af2d43b5d9
2261 F20101108_AACGCR frank_j_Page_34.txt
57931ba13abad65564fa0ba91f78c388
7929d62409d344679ed83d266ff3d326f8734127
95547 F20101108_AACGDG frank_j_Page_75.jp2
f79f9b0a760bb323a01f345535e81bce
9ad02cdc8cc1ddecae8809ac56f8294429c0427d
10072 F20101108_AACGCS frank_j_Page_02.jpg
cb6b4b5e08a285012090f8de86a5d949
4ad4cfac0b02276bf50dc4086fe9ef727a78276d
78665 F20101108_AACGDH frank_j_Page_64.jpg
b3d60e8b5d142a2ab76f45d7a96fbaae
a3bca9b3295641b5afb3975e4e24a2757e9b5db5
96709 F20101108_AACGCT frank_j_Page_57.jp2
2514ac0d9bffb210d4f36195caa05a89
821393c34c3cfffb56e2eb7e75797b781bcfa9ef
10465 F20101108_AACGDI frank_j_Page_90.QC.jpg
5a3888905f0b232e79fbabe1abff10d2
4ff6faf58690ea7b5f486602a5bc2ceb65cfd187
22649 F20101108_AACGCU frank_j_Page_38.QC.jpg
9e7eeb2dc89bf40b0e04f2d670b32e11
b669ee750fbe54bcc8487b008833201003ebb225
48663 F20101108_AACGDJ frank_j_Page_87.jpg
e72db31b256b3426853369e91d0f74c6
fa39dab3ad7097605c912b4110cad99d1522cd1c
2256 F20101108_AACGCV frank_j_Page_66.txt
2f5f58e06bda5eda6611cead2185dd31
f6cc38612d3b825f939255ae7f13bd7e699bc861
3911 F20101108_AACGDK frank_j_Page_73.pro
479969f731f63e91492f109011fc611a
1e539e7888eb08fb8c8fdb5b43f6933e01fedeaf
1994 F20101108_AACGCW frank_j_Page_83.txt
451caf8338d2b0f1bdfb2e580ec18f33
8d2f74904b658eec0b71d91fe7b1eff8c7fca249
11586 F20101108_AACGCX frank_j_Page_91.QC.jpg
318599c3d61595ee2a3bcc3702185394
496fa532c4f5abf3a3b7a477b0dbe9553d43ff4c
6713 F20101108_AACGEA frank_j_Page_82thm.jpg
879fc4f66a49360a9e069095efca3e4b
cbaf08114565d7b2ede613a02fa3625b26daa293
77693 F20101108_AACGDL frank_j_Page_26.jpg
8f411526497baff8c83eafda5062dde7
a3f4f636a71cc8c556f74fe78c8e820818a90eee
53765 F20101108_AACGCY frank_j_Page_33.pro
59a6fb024c15f4c4adb371acb2d3ff9a
e64078a0df18e5d059d421c5b26737778952acca
111241 F20101108_AACGEB frank_j_Page_10.jpg
8630a86efc6866e60d053527b8a79c41
10a0f88ea9c045bd9543949b057314f1aa47d6d6
56136 F20101108_AACGDM frank_j_Page_58.jp2
871ae7e77a17bcecf25cf74af04562f2
a83000f56f00bb9d2b73e2901398a57bb3c6cb84
6670 F20101108_AACGCZ frank_j_Page_83thm.jpg
17847f68d092cefd5d5f7f53c2071442
28d548bcbad7d822ff02ccb3f15498dd490658ec
34349 F20101108_AACGEC frank_j_Page_45.pro
fcb7fd16c1a0cb2e5b942511fafb1518
de90b765772caef14e5085db17845867e9e9ecc4
23157 F20101108_AACGDN frank_j_Page_79.jpg
760673c53ae49f7731568ea39a14047a
a703c50d6d2e2c7091a16f10c387f47b2699b407
1492 F20101108_AACGED frank_j_Page_48.txt
5b61dd6b2d393042e89d87b50ef3daa2
b2140134f4ce1d75772197ec3f710527da89d67f
78935 F20101108_AACGDO frank_j_Page_66.jpg
839044fcdd2c0eab8b0e456296b21299
4551c539de14b13d0bc6bec32985003cddab58c4
31235 F20101108_AACGEE frank_j_Page_43.pro
ba3aff60535f8005d4343a6420dbd1b1
fc65363e975f13785a4bf0fa32c91336258bd5b9
35384 F20101108_AACGDP frank_j_Page_90.jpg
a05872cca514f93df761364244ef0b00
02c45dd16ac09cf9d5007ebf042b24b156c65192
1051973 F20101108_AACGEF frank_j_Page_09.jp2
5473555e24f62f57f281a5a9b2ac0329
c2b1fc64aba4469ea775eb26c768bb99626b7aee
159 F20101108_AACGDQ frank_j_Page_73.txt
7694c020c84fd1beeb5f43ce246a9413
49877560b1b551ccdf5efe8fd4cbf096a816be8c
F20101108_AACGEG frank_j_Page_18.tif
e16873b46857365549945f4dfd3e5e96
97e2f6ccc760fd02b3c59b614a6ead91d1e04c05
142781 F20101108_AACFZA frank_j_Page_08.jpg
daae26fcde9547281d5ce39e32f43b4f
263beec18025c72d89f723818d534ef411a295aa
26069 F20101108_AACGDR frank_j_Page_22.QC.jpg
24ce154fd404a4592d28a6ed75cdc6c8
e7f51a253681aa01cd491cb528b82122dce3af82
53330 F20101108_AACGEH frank_j_Page_25.pro
e91d9ec999ab7cb7ea7aefc87db03b6e
0749b8bbc5cb3b82b33e1258548cd4e708a19792
24587 F20101108_AACFZB frank_j_Page_30.QC.jpg
8762d2e8b83ab80c24f761b2be711a53
1372479c1a652ca54355de6a63370cc9688b9311
26725 F20101108_AACGDS frank_j_Page_47.pro
d9fc94a7fdbf4543b02ad8bcfdf82afa
ef08bcdd72c037fb3e60a10e9d0135bd95ce5d47
2250 F20101108_AACGEI frank_j_Page_62.txt
753d06e9671266afe6ce5ebd781a6a28
4de7548a1ddc91730cc4d6b754e09f6e7922b5a8
508174 F20101108_AACGDT frank_j_Page_91.jp2
16220355c9f0f77c1e067a7454ee3d9c
b3d9982407590f206fdc572b5f71a79b185a38be
119529 F20101108_AACFYO frank_j_Page_31.jp2
81c1c6a3393a3d32a965bac1e1877bdf
77bb11dbf9c48c9f05e39c4098f19e7eba1f5edd
90015 F20101108_AACGEJ frank_j_Page_46.jp2
51565fd09525587e5e432ee23d2b24e9
7e7016849c848095d88761b0577ab3e36be91ca5
81476 F20101108_AACFZC frank_j_Page_17.jpg
303bb75bcbaece2891f4ec96580b2ddc
906814dc03ee3c117546d58d3d0924a54311d238
24530 F20101108_AACGDU frank_j_Page_19.QC.jpg
a5ab9aafe36c266c402f2d198a7c20d7
5ceac3de5cb0d99c5bfdd22600efb26f16042f17
F20101108_AACFYP frank_j_Page_32.tif
e5833cbe4add274e2188a3d92b4e4193
c24bcf391f354f5fb9d77c29f8c49cefaf22787e
120844 F20101108_AACGEK frank_j_Page_64.jp2
e0c127cd63224f2113299b6a48bfa3ab
30a71d9cf4b982670c0d93d22f6ed611d4e2c517
14254 F20101108_AACFZD frank_j_Page_96.QC.jpg
2bd33cd09ceae419400ed7d15f1621e9
7a452780df9fab81c201b6fb76d6e168fd1f00d7
34243 F20101108_AACGDV frank_j_Page_86.jpg
8e5093df9d0f9ed398344cabd8dce226
82728846f37d5993a03b48959ddfc14c4d441a31
15987 F20101108_AACFYQ frank_j_Page_43.QC.jpg
4665dc12a274e840cd34a9c17a6689bd
e56517aa8dc0bf0645f0676ddfd954129e203edf
22778 F20101108_AACGEL frank_j_Page_85.jpg
170fcc772f3e9091339c836402d40eac
2cccba6815ed1347b28d5afebfd3aacd74596b6b
F20101108_AACFZE frank_j_Page_38.tif
9d9754d4e98ef6e8544c98ef1737e1bd
1bf59efae0070d01e2714a62fb22145f23331e17
F20101108_AACGDW frank_j_Page_90.tif
afd0e594604cc776b0186049e1e6fd37
7b25e60d7561179b2559e5b79003e98bb5442500
8078 F20101108_AACFYR frank_j_Page_08thm.jpg
49a0526ba08921e86411cc0174f53c57
580fa9a8c84931fef4ec782bb1a1bc1c207a0d1f
2683 F20101108_AACGFA frank_j_Page_53thm.jpg
6f92b974e923e3efa47766e2c3951f36
6127e00711c2ea47e57e6997efedb0ce132acee0
F20101108_AACFZF frank_j_Page_19.tif
f5fecc818015fb9267b0e6f025293797
ee66373481d5111df1ef104e2017bd339141408e
117545 F20101108_AACGDX frank_j_Page_29.jp2
c81521cc7b2d9ef90684519176be687d
dc1e7a5f1e6783b2a85d4d5c09004f5b083071df
2139 F20101108_AACFYS frank_j_Page_65.txt
4424156ed67e7048e32e80685687d5d8
7c893c5855644e5b2f088c3d121f4f3d88d707e0
6743 F20101108_AACGFB frank_j_Page_35thm.jpg
eff47ed8edf4c397ebf6e4d5ad4168ce
40abbbff9508d07b12f2c10c20b75d353dc60c04
F20101108_AACGEM frank_j_Page_30.tif
3ccbf82c665d49730edc33d7614cd219
123444a3bf92f03f6f2597485794f8b8a27e1571
F20101108_AACFZG frank_j_Page_80.tif
8a12107760956b1e6c86c2cb9be2c100
cd34a487091b5a13bde4d816aa2f57ce7e1c98b4
6790 F20101108_AACGDY frank_j_Page_30thm.jpg
9c4ff89cef4bc1e4917589085e9f6c99
762cc8b0ae2cefe7447608322f53c1cab1f755f6
77739 F20101108_AACGFC frank_j_Page_31.jpg
3403c4c7c54d25f0980ceeee88937027
493a72decf1202c1c9c848d583608c860de23273
42326 F20101108_AACGEN frank_j_Page_75.jpg
b0bd3487e01c3fd04a0b2414727a0ddb
6b2fd8a5502d74a9e2d58597433899eb2c6c7b08
53209 F20101108_AACFZH frank_j_Page_35.pro
30c26bff991798d8d1ed7232ef16a379
e2640419f774aa0c97d7f75b4e831596aead2057
31812 F20101108_AACGDZ frank_j_Page_76.pro
beff56aebfc3d768b1edd149cf1128de
690f18b577885bf958722bc12a85abf7f1a4fc55
F20101108_AACFYT frank_j_Page_57.tif
f61dc691a14d21cb6e4b5aaccedf614f
e9441bb3a90b4e1c9581acf449d4b61421e3b497
114300 F20101108_AACGFD frank_j_Page_21.jp2
31d0e5f0f99bd913a861d8f7aaf7c028
1a5bacf2078baf366ff8c56f159b3e9cd9a32440
53260 F20101108_AACGEO frank_j_Page_16.pro
ae12ab133fbb9d289dc35322f681a69c
11fe7d1c5c42ecc2a1e9a59f0c893a3577c23d61
F20101108_AACFZI frank_j_Page_54.tif
01f9e9a3e410b5d6319d98cdb8eac98a
b166733b2b3aca7aaa2b66ba5d1ceca30839d767
55404 F20101108_AACFYU frank_j_Page_05.pro
ed3bac48e6fad2efb1529443e121182b
0faff29f46efc2416df6c6f65a0ce981b88dd73a
2205 F20101108_AACGFE frank_j_Page_26.txt
ff30ecbc4e76cd7b0e1292600d2b944e
8222b5b2fc3b7efcd66c658ca3c80caa5b51c7e3
1262 F20101108_AACGEP frank_j_Page_53.txt
406a66c7d2c2221b3ba7ca43fdff8817
85b669c068671badf773cfe6859a507f7efab62a
F20101108_AACFZJ frank_j_Page_71.tif
02617267b131df6577ddde8129fe755d
20ca88843fc4b5e337d740dc481617551f3267b5
2646 F20101108_AACFYV frank_j_Page_76thm.jpg
e582a02d2afcb79dc7322ac81c963fd2
1ee56675ea9af3213da55e11584b8b156b7783ac
5258 F20101108_AACGFF frank_j_Page_05thm.jpg
30e3a5ef7ecde689e3017646cfb557be
43d155a1c0424002c8f95a155409ced74459a734
34855 F20101108_AACGEQ frank_j_Page_74.pro
3e5eda3d2e997be2ec3766c9d2ff1d30
8cc0351a0dcf38fa1e9d9155e3b3179564719c6d
20505 F20101108_AACFZK frank_j_Page_13.QC.jpg
94e48e53d94e30f1d848a9a08b7ae84d
53025407e3b36d49abd4f1699e05feb41de07d29
34018 F20101108_AACFYW frank_j_Page_27.pro
62e6857dff81dfff476a9bc98a00a595
c904da7ef0486148e41fb9620b1224fc363fdbad
24501 F20101108_AACGFG frank_j_Page_42.QC.jpg
1eb6d8793557ba7f5c27f9b59baa48b3
2c72c6caa22620839618c806bba8b1ef9783e049
3362 F20101108_AACGER frank_j_Page_90thm.jpg
24a89cfeb77a75d5e5bf58d0470ebcef
402a08b68f7bc7ecf9a985b923ea0f005244c3f0
F20101108_AACFZL frank_j_Page_01.tif
b52eeb19b50598330a6221bf55c79391
cb761b1f8500e1850c0ac2eaf01b510d53b47bda
106755 F20101108_AACFYX frank_j_Page_37.jp2
b0402856d88de4847e686051a74c0897
fc8c8e944bfc30e4b43714c0e06785ece6a72ef5
36256 F20101108_AACGFH frank_j_Page_45.jpg
a1001d06f2ff6983293238a64aebacec
e50855809d95531ede4a9dae53fad6d969373d9a
1051965 F20101108_AACGES frank_j_Page_95.jp2
364aead62a1e295edd82c3770b24a78a
ca85d24cb9e6d1e9878990cdb5fab0a6151f0081
6704 F20101108_AACFZM frank_j_Page_10thm.jpg
cc4875278b6a972c08b4da28b4bfd366
e8d37e537a5aa0aae1eb036b11127ccb87136546
75906 F20101108_AACFYY frank_j_Page_25.jpg
0a3a12392156e1532aa2876e06a9e45a
afaaa6dbcb22f638436f747def52c5dd0172d194
2171 F20101108_AACGFI frank_j_Page_32.txt
8de8e415e4eb229f709c964ba8a89177
355446a95c438881cebeb8f6f31446e4abe8f572
2103 F20101108_AACGET frank_j_Page_39.txt
4a224e75bccc6544d2be405ae5371f1d
64ebe3e40085570fdcdbd1405c1ebd655afc4814
7020 F20101108_AACFZN frank_j_Page_14thm.jpg
52c2d08d628bad672d19c1fdb2b72aa2
efa370fc8cf478133968a652426e2a791503eb95
26582 F20101108_AACFYZ frank_j_Page_93.QC.jpg
2975e433b818d06a4a9f0af7ae1378e8
ff6e89bba767d29398d3fc676886abeeb3eb9f4f
2047 F20101108_AACGFJ frank_j_Page_19.txt
5eafd7a56579ec4fb0f9ab1ef1ab8671
bd131f51c861565a7f05dea88592c8db9692ea42
56413 F20101108_AACGEU frank_j_Page_72.pro
a89df5ce404078986f8eb8e041450acd
7de019ee76ebcb79cc121795edf9ffd413d55c3f
77715 F20101108_AACFZO frank_j_Page_65.jpg
8d31309f264417d52d540778583bb020
cd552bacbd160fdc3fc7c6b725a166251d02de8b
3575 F20101108_AACGFK frank_j_Page_07.txt
5b0bf3a834961c4087229433c7471658
a668c353a9459d74700018f89501529415366b41
4517 F20101108_AACGEV frank_j_Page_73.QC.jpg
dc6ae89def9f7b1738d5d7226e3c9f8b
1b8e7593c22c91f80fe50d99f4c6ddc78cea7afe
109948 F20101108_AACFZP frank_j_Page_41.jp2
f631c9f3f59731466aafd32ff4e1cf1a
2f48cda4147b94dec9a7780828f8fdaf64b9a2ed
F20101108_AACGFL frank_j_Page_67.tif
e9e6c78e040e760d0f46743932023ba6
fc35ec8cdbe8ad20d3721be3a620db7a899f4868
20802 F20101108_AACGEW frank_j_Page_05.QC.jpg
384256f1f8f46ff612342ee5e7ebc660
fb3c47ac8f45c1c9db99ea046a79b8baa69863a6
1988 F20101108_AACFZQ frank_j_Page_11.txt
d13b9f50202e3af45b97cc4012fd0009
35729997e6d44556f7ab56597705180f733c3087
F20101108_AACGFM frank_j_Page_87.tif
a69c116b626cc2e01144368aa834e002
7dea0ba1a05015e2895fb578e9dbda6865812cce
86815 F20101108_AACGEX frank_j_Page_51.jp2
2b00b26419d040cdb15a9993de9a1d0c
d7789e818fce927611a75cfe896c432ba2536272
7031 F20101108_AACFZR frank_j_Page_29thm.jpg
bc623107472a0d8d36b909206f9575e6
60f37b7921b31218fa6c958fe726d41dd4365d48
69440 F20101108_AACGGA frank_j_Page_38.jpg
6065df5ae9c26374a981b5e7ad106a69
4ff2a3d537d4a99cbf1044054188f59e6e0dc5e0
33639 F20101108_AACGEY frank_j_Page_48.pro
a02543fe34c212dafbba545351f024eb
83b5ae8135e425842f3d6ce5bf32e1e7ecf5dd44
128304 F20101108_AACFZS frank_j_Page_07.jpg
920c3419e9908150ec04eedfdacc2e01
71e935b22702a6daf040aa8d6cbc9a5463bf438e
F20101108_AACGGB frank_j_Page_73.tif
e8c121fdce352d99a4f6f2d487d6c7f6
5dadb6bc24c6d6d4f2fe68d7eaaaed3f96db35a8
8602 F20101108_AACGFN frank_j_Page_47.QC.jpg
b658c4c149d3361a945a4093813d2c75
e7ede9019bf84defc5d32e4d9342ce30c0307516
16780 F20101108_AACGEZ frank_j_Page_27.QC.jpg
ce172bb20b46bf6af3eab6578af61196
8308d8cfc4e441ae934a57e01f01d78041d8df11
2676 F20101108_AACFZT frank_j_Page_49thm.jpg
99236ae1e3722f28c55427ed7b23294b
d5c82504ef9fa1338334483c08c9befdf8318df6
7311 F20101108_AACGGC frank_j_Page_52.QC.jpg
bf3bbb457916054600dd00158df14dbe
31309bd3cc3a42556c8e1563ea1aad144630f1f9
4084 F20101108_AACGFO frank_j_Page_96thm.jpg
3894f7685ceb5831b9ea85c013f55b95
da378db23588783f86ace6b732437977feb7c552
25318 F20101108_AACFZU frank_j_Page_65.QC.jpg
3e1f9bbf4c79bd8ee246cd1b4a9cd7ec
db1c1ac980795f4dc3daa20a12baf00368e1e4f6
1919 F20101108_AACGGD frank_j_Page_70.txt
dfe374a0965dba286a3c46c316c86e4d
c88880abf96972d31a1b514d67dfa2d7798c9d77
283 F20101108_AACGFP frank_j_Page_85.txt
4aef781528c9bba65820da71054ccb80
dbd6927698d2c75a916c7c13615178e8702b632b
3107 F20101108_AACFZV frank_j_Page_57thm.jpg
fc33062559aec8b5072dc311c5b940d9
7998039cfe866ec26dbbaa117630b94c27ed2cd4
23555 F20101108_AACGGE frank_j_Page_41.QC.jpg
addba454f6e7f20393441b4b35a195b8
82b1ec09aebd61c575d21303ec1ea8e4a6951876
7581 F20101108_AACFZW frank_j_Page_06thm.jpg
b7c5bc8628ac75feb34909a22cf6bbb4
d8860475855306e4655ce31782ce8df2208c276e
23618 F20101108_AACGGF frank_j_Page_15.QC.jpg
8d3fb9ac16aa299f6c797a3555505d34
3d87c5d87a671f84e34ca33d07df1c7d4844e635
1925 F20101108_AACGFQ frank_j_Page_38.txt
6e1a8c46388ca3c143e265be14d3c3ed
730465dd559199340989c4ad547e579bd18651fb
122514 F20101108_AACFZX frank_j_Page_36.jp2
b2fee890e295586c6944d81d58451954
88f45ab1d8e8b630f2ab7e9366d31e9bd5033e6b
12123 F20101108_AACGGG frank_j_Page_75.QC.jpg
d747c15bf8654a5fa05146f7359e5aac
3ba34e420bb9a46ea615f9a6260b7608206ea4f3
703 F20101108_AACGFR frank_j_Page_60.txt
af993b64c1f3ca745c788b58bbf0607c
a260444964666b018e344bdcd995124a42a03d69
69677 F20101108_AACFZY frank_j_Page_53.jp2
680f96afcfe1414a0af4b2576015a847
32ec84f6a5e3b702b22d4fbb67005cabf54e34ec
4486 F20101108_AACGGH frank_j_Page_43thm.jpg
b3e17b75bb986303d90ade58979fa060
48a99a0ea1d4781f3c3dcd94c26dbdc6a62861d6
1051981 F20101108_AACGFS frank_j_Page_92.jp2
69b4be78f147ec9bdad9249e274de145
b42a90b25613bfdbcfbc24a43608aac6f61bf026
2151 F20101108_AACFZZ frank_j_Page_29.txt
79f859f3a278f0f8630d6f88ae5d82b1
0db0217ead1649d5d671e60f91681d4a69bd0ccc
2260 F20101108_AACGGI frank_j_Page_05.txt
ef88fad0fd429dab7a4f33ed5f194aab
e698c542b25d0eefbbc8a9ad80d73ef95950e830
28172 F20101108_AACGFT frank_j_Page_49.pro
28a56f78e885ea016ca65034222204ba
79085624f3295d3d4238d7f4edcb1d5baa6f0e16
2216 F20101108_AACGGJ frank_j_Page_89thm.jpg
a0b9ae1684ee06f779541d518189d662
61794f19a714656756c7156244e21ece598d227a
34563 F20101108_AACGFU frank_j_Page_55.jpg
152e04c51412fd2c26757fb488af6a7e
92797011078649f5190efae65e1bb69a3230727a
405490 F20101108_AACGGK frank_j_Page_90.jp2
711ffc1c57bf8b14fe0dfd208d028aee
a841d25c5cf1efa1f2c43a934676a79eb45ec4d9
5303 F20101108_AACGFV frank_j_Page_03thm.jpg
4289656152242193587396209fdb44c8
7d9956e94f02904116a8ad0e7c2cee1956f6c678
23754 F20101108_AACGGL frank_j_Page_80.jpg
ae32aed59f0266ceed8bd9eb62841509
c975249da8e1d922375e932170296aa5f9dd198c
2517 F20101108_AACGFW frank_j_Page_52thm.jpg
eb8194dba534e82a43f1a40d454a1038
25868d197c856ed73e7e265cb167aa513a63f2e2
3311 F20101108_AACGHA frank_j_Page_46thm.jpg
a92dc685f6c18b126b70403e450a59b5
10563d6f83f6d8b68df0f69b8cb8a68b6a297d21
80885 F20101108_AACGGM frank_j_Page_03.jp2
ae3ef6c931a4d6fa74ad99644c0127e0
81913afee9d385d06fc4c997c62e66f8c28f8f2b
1945 F20101108_AACGFX frank_j_Page_20.txt
d2abd26853e911306ae002d85bbd1bf3
90c9e21fb5840062c69324397de4e2f9e2d7d8f7
73251 F20101108_AACGHB frank_j_Page_04.pro
6ad7a50e44eb3f45e228f32bef4f553d
1596508fb8d568f0006c83e098b6a0c7489530ba
7014 F20101108_AACGGN frank_j_Page_36thm.jpg
d2717647dca188afd89468ad5cfd023a
d1203c0c81a0e3615f521af08f48781e4c9f58a2
22107 F20101108_AACGFY frank_j_Page_11.QC.jpg
560b08ebc9cfeb86fbc033b391b37d89
2d99b65185956eddc9e05c95d36f6230127497ea
16474 F20101108_AACGHC frank_j_Page_60.pro
41c8fba9718e04bc8e34599d3d21f54e
e6d3771def5501f8157a561c3eebdd8169308574
38512 F20101108_AACGFZ frank_j_Page_51.jpg
c04ab595346a5d6c23f9028cf7a44213
38582c92b19a7fee134f81eec5560ddc951a187a
2967 F20101108_AACGHD frank_j_Page_10.txt
21a12e1b43ed69ac6d6452ab33c68ebf
a596fd52539da59040f5c69616c78ebb7c1d5c81
6747 F20101108_AACGGO frank_j_Page_37thm.jpg
958100559eb8aa1390cc693517433967
93b75b50e5604d0ddfcc73bfc2e61735cdab6ae8
20369 F20101108_AACGHE frank_j_Page_89.jpg
60427f4433a786918a15f9b3ae935a2f
e570b806ad57d8911e5157b59f3db297127871f3
2395 F20101108_AACGGP frank_j_Page_78thm.jpg
a169b4badac48d9629a0915da2098417
2fb0f5a04c61a123e6ea11076d53e2b84725e39c
2076 F20101108_AACGHF frank_j_Page_63.txt
5fc59ed985d05163c30797e99254f090
9d441148c7feb613246e72befd2b63f2e1b6781f
56620 F20101108_AACGGQ frank_j_Page_22.pro
fe314c87a5e80d644f57db561e4f7ba2
215afba6dfc1829d72d854e038b3a4b6f9438d7e
2145 F20101108_AACGHG frank_j_Page_18.txt
542e972241ed70c814f056d06c555a5c
c0ce3fc04e1d5a855ee86f20dc931515aaa94f1e
70818 F20101108_AACGGR frank_j_Page_10.pro
8f81cc40a1ee75cb947c18197f1b683d
6d13048d498e32e2b319e487359e43bdf52ed32f
6992 F20101108_AACGHH frank_j_Page_31thm.jpg
40056b2365e06010256918cedc854a60
291b115f592c8eacf0f8d1872eb09d348aab5251
13363 F20101108_AACGGS frank_j_Page_44.QC.jpg
b6fca0e5aa2727f1baf28d01bc7fffaa
d9070ef143b75f01225a3b74d771e76cf3a2153b
20871 F20101108_AACGHI frank_j_Page_95.QC.jpg
ebb24621f7cfb6425cbaff1fe3419801
e2e6a7c3fc8c233f6b2ceff32d434efb7083db85
F20101108_AACGGT frank_j_Page_92.tif
72064b6f91c0392381f563075114b773
95334b191cf6668e7af7cd7be4ac2cd1f9ac0392
3213 F20101108_AACGHJ frank_j_Page_04.txt
55c5d15a17343088cf018de2e2176a97
ff524a9cc5292acd8b72efa63c6394613932d0a1
6207 F20101108_AACGGU frank_j_Page_09thm.jpg
83b9e9659e114cbe8e2d3cc012cfe8e9
bc3254daa76d1e651fdf34db17cf62d7f7201f08
2765 F20101108_AACGHK frank_j_Page_81thm.jpg
b6785e765a0821bc140eda94c67db5f4
1d890b5e3ce7f0d87d1bb46439f8675497f2a280
24710 F20101108_AACGGV frank_j_Page_16.QC.jpg
bdd99525b486f307f1eea3554e52db77
0b37a0261f1d5ef02551e3658385c57852f60cac
117898 F20101108_AACGHL frank_j_Page_14.jp2
91ae3f391b352e870d00139c7c74a1df
d6de21755f57d27c935c473cfda92124745a3477
77991 F20101108_AACGGW frank_j_Page_29.jpg
77f80dc0076fa8e4d96c172a814ece1f
9bcce7b4f54c532504f95c88ce64fb15a1c32805
45049 F20101108_AACGHM frank_j_Page_95.pro
de3b6ba9220dcc2d4d1e5c0e2117fb45
be742041a203e1ace118c781f596c60d525bd032
F20101108_AACGGX frank_j_Page_10.tif
2ed77aee635b55b84b2aca0458dfc4f5
8b16c8cc88b98854066f34854be3f8c3244436b6
2259 F20101108_AACGIA frank_j_Page_72.txt
1b509c546a2677ec27b3af1323bbeb56
8151b90af03206975aba291a41b204756d297392
7653 F20101108_AACGHN frank_j_Page_07thm.jpg
b7d3af881a87a7bc5d431f6b46f8410f
ccc997e89597f62f6d7c33b6d8d500e5a5f2733c
29167 F20101108_AACGGY frank_j_Page_77.jpg
d0f0d7074c4a1575e8d36c6dd1aa1815
d375ef4c0ffb38e42dbc241bd365d729c1588aa4
50977 F20101108_AACGIB frank_j_Page_27.jpg
8927792aa0dc3a1b630fa4680742a4a4
8da11783b0438d367337bdb4fcb3ef2e301e8596
75438 F20101108_AACGHO frank_j_Page_05.jpg
e817e29455f1f81531585f1e890e291d
1f6a6f0cc5b996dd7d1c97bb86706cc749004ccf
F20101108_AACGGZ frank_j_Page_06.tif
93258194cc22f0096cde211c33b4d965
7cd7ba03651fdac9f04f8ab51f11175177437f3a
F20101108_AACGIC frank_j_Page_14.tif
8198757b64daf281171533195fdbf441
d1d2c9a7822fab05db6fcdd38832381d69a08bee
23790 F20101108_AACGID frank_j_Page_83.QC.jpg
b56297b9695f7079850e6e9477649a76
131b8cd94c0718adf5627aa30524f728b1876b5d
1834 F20101108_AACGHP frank_j_Page_13.txt
301d7f907913595fe63669c28cdfce0b
959e29de8de997d0019c22e98923a6b73d276e9b
23017 F20101108_AACGIE frank_j_Page_52.pro
70aa651a260ad1644d8152b2862e7b7e
b49462086c39635ea676731af669e427b36da226
54345 F20101108_AACGHQ frank_j_Page_65.pro
0e1c86223ae58f252b4791d007bb199d
5d0d153e25f7ef0b86a83ca77381e91be9b56d70
7322 F20101108_AACGIF frank_j_Page_92thm.jpg
cd40a00e4636c143c73bb5b7455583f0
10d878575a3d1e746dd3148945d23cb4a198b49d
7573 F20101108_AACGHR frank_j_Page_01.pro
1c7914cbcb5a183a554c814ef13a7fb1
eddd0fb95286e220f8cba9d1da594db42a6cb457
10574 F20101108_AACGIG frank_j_Page_86.QC.jpg
c354dfeed33193a2faa52f9d1965a54d
b79f549492d3bbce0250e70f24f8b53ad85a079d
6839 F20101108_AACGHS frank_j_Page_21thm.jpg
e431eefc1e4bda9f55285297ad5e4003
187f1ebea33d898e6ae06b4254428778e9444369
2850 F20101108_AACGIH frank_j_Page_45thm.jpg
c37894d1561f3668f2cf9baf80e45b9f
b77fa552867357e326cb26bbd336bcdbeb994780
114077 F20101108_AACGHT frank_j_Page_19.jp2
91fbf7a1dd156764f21ad7725512f39a
4011dc6aad6fd76f8308ee973a94c10bec6d4155
F20101108_AACGII frank_j_Page_66.tif
bc4b6b2104faf809c223e4e9ef8e5892
ad39a18453b5ff6d4390353d030244bb6cdbb7c2
73546 F20101108_AACGHU frank_j_Page_15.jpg
feeacf95e979593fa9c3c6576ff7c2d1
072ad64c0f4e4653fa9c4e9cfee075737ab65f0c
46179 F20101108_AACGIJ frank_j_Page_61.jp2
2358819a56440686993aefecf9c86d61
198873ea7f6a4147d3dbb4ef5e22b6b84d0e7192
26122 F20101108_AACGHV frank_j_Page_68.QC.jpg
6dc5ca0313e2252823c5cb870ff85833
4a4448416d5fb8840f6664b3cf7e3ca8903a0640
84371 F20101108_AACGIK frank_j_Page_07.pro
3f19bd744a64010c681c636ce91776d8
0586c67b9c3ce36344b089a16316c5f02ec6cd40
8548 F20101108_AACGHW frank_j_Page_53.QC.jpg
0d5f75c09e9debcf849b8f1b1584d0b6
d28b216202402ed16f51d0fefe54ab22135cad55
6655 F20101108_AACGIL frank_j_Page_24thm.jpg
e8779d41ecb75ec28cab88c7c4522773
164e97dc4e68e4caf6c3e8589d861ec0c9398993
6941 F20101108_AACGHX frank_j_Page_39thm.jpg
0d7b2b0417cf1d0a2e447e906d1b2e95
1be5506cb65881920d9ec1f4d0443e7b7a0eeceb
F20101108_AACGJA frank_j_Page_56.tif
ee180a3c2d2ff831815d1ac3d67daf87
92f8adf8cbf17471a40ad170e6d5034c8840fd24
1484 F20101108_AACGIM frank_j_Page_55.txt
26f9a22ac4e44b1ad10d0e849aef1fd9
02170a0d218589d49ebf855985bc016d7b0e4ba0
110196 F20101108_AACGHY frank_j_Page_12.jp2
eb2d5ebd81c1271206857e6d78fbc0dd
ab97fd49dd70ab89200ed742ac1b1fa8fc2ddf3a
54961 F20101108_AACGJB frank_j_Page_31.pro
05335a431898625457c76441189faac0
b5df1f2eb75cf9781811534eae4a2a1124483d41
2989 F20101108_AACGIN frank_j_Page_74thm.jpg
fa8f262956d41e61a69624f25aaf934e
282f1271ec9107bd664e5aff3ff03a07c0e8c169
69098 F20101108_AACGHZ frank_j_Page_87.jp2
df15e8214dba21d071f283e60da65765
2aecc373d5b17417f3e5d46afc80c31de33b02b6
F20101108_AACGJC frank_j_Page_11.tif
b9018f951b06504b98848f786aee03f8
161890ad6ebbb768614094d0019a224fdf157053
816 F20101108_AACGIO frank_j_Page_80.txt
665bb1f7b501d4ed43bf23b5b783321f
e8279b19691cd3f0320fd7b84f795a9b466a7838
25324 F20101108_AACGJD frank_j_Page_29.QC.jpg
60470ed6d49eaab310193fc4af9623a0
e190e570d62c0f8bd3fa73884f6f3eafa6d6abaf
54615 F20101108_AACGIP frank_j_Page_29.pro
3d7cb7c4a6b2ad5995f33ad05d9d579a
b0d0c9a2f3d739ad9e4b7295e6b1aeddb8380f6a
1665 F20101108_AACGJE frank_j_Page_51.txt
a743adbcf5ca0a50340f2f9eb7260391
49ffe4d06add4ae7213d54ca7d9b9ba52942ea1f
1993 F20101108_AACGJF frank_j_Page_37.txt
79ddda860b5452edcb5da1798126c641
b9db9ec5b83d38945baf082edb9cbde8025af3bd
32815 F20101108_AACGIQ frank_j_Page_76.jpg
327bd7557ae4e9fb23e1eedbd1260a2d
e2af19ba84c0a522a6ce8a3d921b16b7747099b8
6821 F20101108_AACGJG frank_j_Page_18thm.jpg
5d49da3b94dfd2451692448220d8d1ae
fb488c4201c56fc2c776c6ff2fbf2efb00bacd64
350684 F20101108_AACGIR frank_j.pdf
e61e00873ba6a5027fb472dc84dd119a
c30d939a58a5d3ebd2bdb8752e90406b8be8b3ee
119867 F20101108_AACGJH frank_j_Page_32.jp2
07169d864b080160a821b0bcdb8f9d64
a9cdf96aabf72c452a3499e55aa00ce434c53b30
75961 F20101108_AACGIS frank_j_Page_55.jp2
a409aa3725bd130fa87b76c794cc794a
f963a8cff84567d274805fe1d0703fc45ca80d1d
30493 F20101108_AACGJI frank_j_Page_06.QC.jpg
fd68b5720077081f8908b4ed78c0b254
c8eb7793be3841d444a1b5532fb32e8999d7ea29
47429 F20101108_AACGIT frank_j_Page_88.jpg
cc5da20e0a6ce3049b5aa7445f4e4c2a
f52f6dbfd326f0d48b19ad006ab4aa98674161e0
76685 F20101108_AACGJJ frank_j_Page_18.jpg
892c6875e104992382d24454a3163588
947198c1e155fec5b63fe54899ba58163c95751a
77289 F20101108_AACGIU frank_j_Page_35.jpg
5f472c56e6c7ee8274bc38c698dbd34b
e87f86170192d7aa660c77639f1a456022741a83
1346 F20101108_AACGJK frank_j_Page_27.txt
8729b753627accef9fbfd43d64eec0b7
e91fba769c9a736be7a8419b1c83322628ccb17e
68054 F20101108_AACGIV frank_j_Page_11.jpg
ba9a340e99ffdc71dc0d1cf25526f573
d5e5dea46a399b61cc71dd9ba9ee986e9ca4f3d9
2173 F20101108_AACGJL frank_j_Page_68.txt
81107b78d8725bcc5b369d66a66d89af
91d04791d6ea2548a98f6ed21b9aa4d4e4cb8c30
1051972 F20101108_AACGIW frank_j_Page_94.jp2
994a29c76adf69a8d9341c7a22d1fa65
75d597cdbd9fa351ae22b9d8ba55c6f78048e969
31979 F20101108_AACGKA frank_j_Page_88.pro
445b8c8328979258a56f91ed97ef00f7
15e33907e197b03610c6b157e2dab10fad2e18ac
F20101108_AACGJM frank_j_Page_62.tif
8dccded57fffd5de35d68a6a226a3a50
67f5d09569df05d4a4f35e231e622cbded658881
F20101108_AACGIX frank_j_Page_52.tif
fe6ee5488dddee4595af6e7182d0f226
2f5b17100dfcceb431015d96a72e02228784b67a
7216 F20101108_AACGKB frank_j_Page_28thm.jpg
681792df272580d23a2a2a4ba84d3237
947b12458a0fb01caeabdb1bbd003b68febc3d8f
13466 F20101108_AACGJN frank_j_Page_61.pro
bd9508d685068c39f61332b0f5c598ed
4647c17decd182a04664429e3acbbbcdeb7cba13
74075 F20101108_AACGIY frank_j_Page_54.jp2
d306b1b34f19917304c9631e25b1ee21
d5bfb2b2f293b102fe9e9f544f041889e36edc67
13744 F20101108_AACGKC frank_j_Page_80.pro
41b33c793663b953ed4e70b18277b525
614edfa8ac93b7b30bf6878213686d5dabfeb5df
65038 F20101108_AACGJO frank_j_Page_49.jp2
14e3d2670da41109dab8aa214b774d3e
e7b0a03aa0271ffd846ffe8a7cfa5a6d9e0b4d43
6863 F20101108_AACGIZ frank_j_Page_33thm.jpg
81095dc1c47c6dc38385d336333ee5dc
088bc16e20126fcbb8be968657743b698b4c2810
72064 F20101108_AACGKD frank_j_Page_37.jpg
60ef11253ac9c290c58d41f11c4a9867
261ff5a788b9ce1807692ca8f42870457ff01706
75023 F20101108_AACGJP frank_j_Page_30.jpg
63fb68295743ff7ba5f1d740924ebcc4
f4e15caf4ec9d95bab9a75c3d5aed5b0331f3e76
2995 F20101108_AACGKE frank_j_Page_48thm.jpg
6ce467cfe6f8ee5df67bbc9c6eb2151e
cd0d3beeb1ca5a9caea746fc12d6456e623fa76a
83539 F20101108_AACGJQ frank_j_Page_06.pro
2fd5267dd14ee1759bb741d4c377453a
81a99865569261532d124aa3736bde21c1ce92df
43333 F20101108_AACGKF frank_j_Page_13.pro
61cad59bf8a49c0211407ad2401e4620
8ce07e49a4b29234748753fde4b1f8cd2b4ae76c
26304 F20101108_AACGKG frank_j_Page_36.QC.jpg
1099bccdea38bdc27c724fce916c96bd
22c3ee6aa1496cb2103e81b143bbd01048b14ad2
F20101108_AACGJR frank_j_Page_82.tif
411893d6be0857ce7c40dcba0f13f64b
274046a8a38148877b470aa7442fb4d1945b9908
F20101108_AACGKH frank_j_Page_60.tif
52cb0a5e12166fe0f98f31073907fdbe
2e550b1f3b95babd208a82f2b88a94463fcd4818
106831 F20101108_AACGJS frank_j_Page_40.jp2
dd6e2d5c07c4fc0be9444d58f1b5cab8
857bef722b135167462897ade2fdef0a2e44c883
8489 F20101108_AACGKI frank_j_Page_61.QC.jpg
f3640c8805c590dc1e3d7b9d524ab993
084b30335ce061c1b7ea8855a47397e99c491824
71412 F20101108_AACGJT frank_j_Page_41.jpg
d35c51b5f55af4ab3746276740672d14
9e43304f8f7a3ef2582c56293f7727369a041fdb
115581 F20101108_AACGKJ frank_j_Page_39.jp2
da42fab918b591bc6e1efa133c3d9219
f295e560a31d6316bab29772bd51f9fa1d7f3609
73949 F20101108_AACGJU frank_j_Page_48.jp2
5aed078dc0c8ca4475ee8400b6936a1a
d81b80669b50b6f72d4df9364cdc836008a8241d
2538 F20101108_AACGKK frank_j_Page_77thm.jpg
624796e3af541a0459d9f47ce6e93e20
b4388e3eae0cedfc7387e2f55dc479e818768c9b
52062 F20101108_AACGJV frank_j_Page_19.pro
0e193112951c2c9d5ed48db5c6ff3c0a
e3caab1cced8047f02cb60d4a4b35771beb3ac19
83078 F20101108_AACGKL frank_j_Page_04.jpg
47dedc0760b503bad3d2802878b307fe
e4764a5f3c6cf0cbc0bba52b3a0a47f59017c77e
117831 F20101108_AACGJW frank_j_Page_28.jp2
0886aaa12b3bf1e84d333bcb710d0fb2
776188f056ffb482c1c309a4f0deee9f44f30a0b
27993 F20101108_AACGKM frank_j_Page_10.QC.jpg
5a0083221408d3794ed28d66a86a7dac
b3c2e33793cbdebabd5a3b276f92425548106c83
106552 F20101108_AACGJX frank_j_Page_38.jp2
d86d3475f23c5b034b6fa148113afe13
8b09c612fdcb3de3d260995359c930f3284e2838
F20101108_AACGLA frank_j_Page_58.tif
95c61ebb8fc7871ff8ea14348a777a0a
bbe701778ad9e3d3ff8c2387ee619872dd049a10
49520 F20101108_AACGKN frank_j_Page_37.pro
67d3db8eb1b9692aa8f6a8663eef24f2
ceba7d8e3d38e8faa3987a20e57f5b0b61025833
84157 F20101108_AACGJY frank_j_Page_50.jp2
7b6e5191b2fac4e4c24d319538944517
49f480151f05d74311a82e55148bb9c3fc2d105f
21515 F20101108_AACGLB frank_j_Page_04.QC.jpg
e2972741b32c90479c6f3654df4efbc6
9642ab1646c850d1958ade8e81b18c11af98ce36
115122 F20101108_AACGJZ frank_j_Page_63.jp2
4fbf7a6930e64064d0e93b1fd5cbe6ea
a94d28d192193cf893aeb6dd70e0726137870639
48410 F20101108_AACGLC frank_j_Page_70.pro
015dd10dbf1a3a8cc7ab5c58035faeb1
19bc695ed4988d98bac995f595a069fdab9355ed
4709 F20101108_AACGKO frank_j_Page_27thm.jpg
d823fedd5322ff616d2d0ac8e1a3ffa3
c6a6187df5fe6511ac64e7baebbd9c9edfb4e581
1342 F20101108_AACGLD frank_j_Page_02thm.jpg
ed03b94993b224cdeff7e42fb30f13b1
2ad0d7d810b27f13e219df569589184a443acbb0
2844 F20101108_AACGKP frank_j_Page_60thm.jpg
2d861892d0d5ebbae6ec660c8819cb5a
524b045a56105d12f19c83b76559deb0d224ef13
26550 F20101108_AACGLE frank_j_Page_96.pro
f49e17b7608de4eb2c38ddf8cbfb2f5e
6c0f9123064cc7ff7c0c9e9828183a39bd84fc62
7185 F20101108_AACGKQ frank_j_Page_23thm.jpg
06a2a36b0cc4f2a813d1e96bee06bacb
d628dede8aafc4fa35142583a90862e8caed5686
1051917 F20101108_AACGLF frank_j_Page_10.jp2
4b28e87317a9506921eb26ad39bebd76
cd8d5f5285f4feb9c65f533c128a0f4583fddf13
51442 F20101108_AACGKR frank_j_Page_15.pro
4826698160e59a539116929547a43835
c6efe27f53611f999fc6da5ce19fc72c9658e06d
4262 F20101108_AACGLG frank_j_Page_88thm.jpg
b55440804bbf312627ae66de9306ffd3
672a7e57067597d620e64cb075ab18e9ea2b7dd8
187598 F20101108_AACGLH frank_j_Page_89.jp2
7ab48df4977ff1ac4b67fc0102faa6b9
e3a7e4135a09362bfa27257d4f504826c8478b8c
F20101108_AACGKS frank_j_Page_23.tif
ff5ed5cb0fae28fdc8407e96d09dccbd
b68055145075068248e751cc933beaccad96ae77
1678 F20101108_AACGLI frank_j_Page_73thm.jpg
6f26de2cbbb3ad405580335ae5cabb32
146a8e289930c1c1801193adc944d4ad0f2f2f68
26089 F20101108_AACGKT frank_j_Page_66.QC.jpg
15c2cb03754584dc35e8764bea62527d
ba6801173e607addb6482cf6d29897b0199ff2fb
6530 F20101108_AACGLJ frank_j_Page_15thm.jpg
ce26e35abc7d5a4dc3b1efabde68828f
596060bd9586cd613bf03ced184ba9a46604e290
F20101108_AACGKU frank_j_Page_93.tif
86cdab4b2c22a68ca0c255c50b2a031d
d589e984acbf55dbab699502c6023b523435d2f2
43535 F20101108_AACGLK frank_j_Page_57.pro
114a257104d957aa0cf47d499f5b7463
f13baac31872164085cf86dc02a47b76461cf140
F20101108_AACGKV frank_j_Page_13.tif
c9127daacaad707a9f7aa915227a2426
826dca129d177af5506b2bcd0712abf54ce0a95a
103863 F20101108_AACGLL frank_j_Page_70.jp2
ada13fecaa1b0e5a4a6a38fa40415dde
a3f783205e43143252c157ea637e982b5c19f121
6793 F20101108_AACGKW frank_j_Page_67thm.jpg
2d386882714222ef400c53f2748b962a
fd7515be1331e1debffbdb1846918d1fb409448d
7399 F20101108_AACGMA frank_j_Page_94thm.jpg
dc97d20eb15e90589aa6a48d036b5bec
8be838755425bc622f8e61a94222b5ffea984162
45638 F20101108_AACGLM frank_j_Page_60.jp2
64b5d2d53e6b3ccbe4fcca61c7318313
4e693dd54350e21fcc80cbf626f45dc440da825e
122740 F20101108_AACGKX frank_j_Page_17.jp2
7cb41ad0955196f9ea4c10adf1028f95
4066d4dbdcc010336860b34ab265ccbce9dce5aa
53056 F20101108_AACGMB frank_j_Page_52.jp2
60874e4d4c70ccf803e11e93822ae146
a787fc5618ac08af1f1aa5a6106599d3addb319d
114971 F20101108_AACGLN frank_j_Page_82.jp2
883721c2ee6d6ce476325a23dbff4289
b9ecc118c676e53f05bada898c4dd65f903c9614
7096 F20101108_AACGKY frank_j_Page_42thm.jpg
fbf86c1f84fb2cf0dc84905e042dcc32
8536878c73452250e303575089717758c78c6bcd
1695 F20101108_AACGMC frank_j_Page_46.txt
541eb78ac9e92b11147eec309a9677c3
dc0a20647c704221862707e1bba096ee88ec6b4c
8425398 F20101108_AACGLO frank_j_Page_59.tif
b09155c3fc9454afe31c5d2bf1eff835
36751bd62d08d2d995fcfaf2feab5d5e0d7e7c94
96323 F20101108_AACGKZ frank_j_Page_09.jpg
51d00a44fd6f6828fc4343ca0672c130
50fa69cffea962d828bf2f7866ebd3810f8a17a6
49479 F20101108_AACGMD frank_j_Page_20.pro
276747379452959918c4318973887e47
24410f40f9f6604b37892fdb73218d1f236fa19a
63599 F20101108_AACGLP frank_j_Page_13.jpg
880b18ac46fcb2e822cad6e19d105d0b
a7a7f7d7306d5ca97a7a7e1ede31b9f87a6f196a
2212 F20101108_AACGME frank_j_Page_14.txt
bf1af2cefdfebf1ecb20b0732ef509ad
ef9c59c3f88ca87343394f0afe654bf9ecd67ab0
77998 F20101108_AACGLQ frank_j_Page_68.jpg
ef45f2b09d2a04ec305177a1ca34a35b
c36cd6651710d5bf9179bcaf1ad1786e50dc8610
34726 F20101108_AACGMF frank_j_Page_48.jpg
6cc272a130cfbb5cbecdc2ff321cee5e
e8fd8887ea6e15e6c4eef36e8afb2e42197848e9
88318 F20101108_AACGLR frank_j_Page_93.jpg
4016d6e0e05dd4d7e5955007fc4195af
2af9e015d9ce8e516777a5c418eeae11f919e361
55412 F20101108_AACGMG frank_j_Page_32.pro
490d57804509c36938dd5eef2e37ea0a
1be83b06f7f3279111b86f806fa85ee957ba7cbc
76348 F20101108_AACGLS frank_j_Page_63.jpg
0f2651f72d7b899241c9b167ae32e4fd
0b35acc38e8b4edb49119d3031a1c93a0c2ed2ae
25265604 F20101108_AACGMH frank_j_Page_85.tif
7e0a9257103efe6612bc2cceb3036b7c
749de2756f3b7c488c83414c01b7d9d61b40e6a2
F20101108_AACGMI frank_j_Page_78.tif
1bcbabe085abf15bbf563a48826fedc8
608f6d85bfaea8cda9180ea154a046916cde2ba5
53412 F20101108_AACGLT frank_j_Page_39.pro
9275d3c9613cb61308cde611a38b8289
fa089909138306cacbb15d1af1cb80fe7baf7f38
7500 F20101108_AACGMJ frank_j_Page_85.QC.jpg
6bd062c5d37f338bffbd260e55e044f6
634379ae99d1307645d4ca8c0015c219a251077a
78772 F20101108_AACGLU frank_j_Page_28.jpg
05938e45547d413827ede0c00bcc15a7
b5b2ccfe4e351bb0261150dfb8144075dfd44b9a
7289 F20101108_AACGMK frank_j_Page_89.pro
d1a243e5e6063b81b20bb7bc9ca71acc
50b9af6a73c08a06560d24e7cfc01266540db08b
118150 F20101108_AACGLV frank_j_Page_68.jp2
1c32c44cfd8aeec5483883cdbefeb1fc
4b7a25a5ceb8a0b6cefcfcbc3d667a0f176c7a1e
50334 F20101108_AACGML frank_j_Page_83.pro
4317cf553205016e73e3ca698bd524f7
18a401279a3edebea5d0147dbdd9f84893806c14
117881 F20101108_AACGLW frank_j_Page_26.jp2
bd43c804896a8908a523dc6f5d8048e1
1ae877546f192027cbddbfb4e7cdfe671f0a9f0c
2971 F20101108_AACGMM frank_j_Page_85thm.jpg
1e9a45fb7eaeeb284e87ec9927539c75
e3c5a3453832aed3156112e520f19de962762de0
60404 F20101108_AACGLX frank_j_Page_96.jp2
8d601b110415978e9498f1759d865d87
b76a4fe886ec7db16c34eaedb01fa5433dbb5e9d
5628 F20101108_AACGNA frank_j_Page_02.jp2
b87d9554fbf3e59c1cc5567fb8368607
eef12451426d3bac6106e323cafa278b5a94bed2
3111 F20101108_AACGMN frank_j_Page_02.QC.jpg
48b695a50665ed22dfa4b1a0c9e0caa1
01bb53be037fe6f94c61fc97ee054c840777f66b
110819 F20101108_AACGLY frank_j_Page_67.jp2
1063288d7f12a916e78b6de331b16abe
0f827f75e28aee1276514b4e2cf328cdf795dd07
26335 F20101108_AACGNB frank_j_Page_62.QC.jpg
4149eca0f03088f302170a8084db131d
16fe715215e2eb5ad69e0abd06b0cab5df19a0f7
71123 F20101108_AACGMO frank_j_Page_70.jpg
82a74db09bd059590b67b4a993bc53ac
38e32b20d302d4a6a7578f6491e706a8e7339703
2603 F20101108_AACGLZ frank_j_Page_47thm.jpg
2fc150abdb6592b2917abe50bf826e14
b263666d5e32d779a354628cd314f2412e342b71
F20101108_AACGNC frank_j_Page_30.txt
e6e389a20bdf434d1819d944c93e0faf
c9329523363d3212e9750e3184f94dc4035664f0
1174 F20101108_AACGMP frank_j_Page_47.txt
c1e223433210e4e858df5826d04982ea
e9c18f438522b5ed5e3638db9e5e0f8c9335b8e9
80961 F20101108_AACGND frank_j_Page_72.jpg
e8119f03912b4d44ac173198bbbea6cf
e8c4941ae59487c49d32ab2ee495d1c25c98de99
F20101108_AACGMQ frank_j_Page_88.tif
a6617b7803b60260a6a39a256df87f31
31ecf2b8a78c5d559a4634bd440dc30fecc80ef9
35806 F20101108_AACGNE frank_j_Page_80.jp2
1ff86ec36821952f3be0502d9c460116
25e17c1b00338aa096c6be6c607a114c74939aac
70251 F20101108_AACGMR frank_j_Page_40.jpg
70bc753a4a9ab0a17b0f1e990fb80ff5
a5eb87691af5b0ac9de57ad5cbfd464386e480b1
121877 F20101108_AACGNF frank_j_Page_72.jp2
d795d70192df806229f21b731383715d
9b42a9ef6e7cbee69918bcc1e9711587ef9a2cca
15168 F20101108_AACGMS frank_j_Page_88.QC.jpg
00d0749740b9b9041dc71ff8a403451c
574ad2528659d900e0d25f3e5533d86afc4adab6
115433 F20101108_AACGNG frank_j_Page_25.jp2
d6d783eb76404588da1ce786d50ccfe2
71e188ef16a8ade9b919e6281bdacf40bb51df1e
F20101108_AACGMT frank_j_Page_44.tif
b90710510fbf303af2f5b5c0e8079473
09c3cae9df0a15027130bce2fe5c2d6cbe6b3a51
1021 F20101108_AACGNH frank_j_Page_58.txt
4b8a71b3268a0c11bb64c8eeab597f69
7944665ccf0f9ad58a350b0809236227e0e6cb77
29873 F20101108_AACGNI frank_j_Page_59.jpg
1e40e4c8ce2cf2f54fce6a204bac3806
36820c897324273bfd8a9fa557bb4312763c9f45
F20101108_AACGMU frank_j_Page_91.tif
9a53c80ed346cfd6c1776a68fb2c91b0
e8b1af863a6652f8db63c028ea3b1767f6be6fc4
39769 F20101108_AACGNJ frank_j_Page_46.jpg
59e0e970c6b84a8ed01a7ecf6f841a95
93092372f9ed9948f4fc93cbf0beefbbc60cbba9
25240 F20101108_AACGMV frank_j_Page_25.QC.jpg
c34914c295809df4493ae5eda3dcb2cf
33a71e15b2f2c788e0139ba2fe471ea64359efb7
F20101108_AACGNK frank_j_Page_83.tif
820da41b7ed09f734a645a71d2310b63
3bf76240cd91b47868beeb261c2921286828687d
114027 F20101108_AACGMW frank_j_Page_30.jp2
e31ef47e38993967e556e2cdfba5cf11
5432e994387f1db95262cdca06568b51e720e75c
64855 F20101108_AACGNL frank_j_Page_77.jp2
7deb04208280fbf3ec1f81c8e234158a
bfdaf6140161985ab252180ee85fe6deec2bfc32
25374 F20101108_AACGMX frank_j_Page_14.QC.jpg
f2a5787fc28b89a1fcd1f0db19a35988
7b52ef8c1afad2f5b3d7a0e2543d5569f66f4360
1454 F20101108_AACGOA frank_j_Page_54.txt
6894904279f1cb71d454f487288cd190
9745a576a830022aebaddb438558cb95d58ee592
2010 F20101108_AACGNM frank_j_Page_24.txt
4561d16289166c50f6df0592635750d4
86076107c476d7db87eff8b99ba81c81fc877be0
55040 F20101108_AACGMY frank_j_Page_26.pro
0306809a202fe2a3a73a1dbca077c2ef
8c2113e81e72f02e5350ccff6bfe9d15e37c310a
F20101108_AACGOB frank_j_Page_70.tif
e7e8127ab89aa8e62ede6d46abaa8b4c
48dfa0fc19253ed86e1a07854d9a6dd2398f239d
60219 F20101108_AACGNN frank_j_Page_47.jp2
d60afd1fe60aa3cf817ebbf6e4858473
62a9bde10ae4c77255a41114b5ab3184143bff8a
2691 F20101108_AACGMZ frank_j_Page_59thm.jpg
1a173d77547cef2ffba3d38fd9ae4391
2af48407e3e8d335c90010c2e1c2c544b92e07a4
8736 F20101108_AACGOC frank_j_Page_60.QC.jpg
c06dbd3859032d046efd34f6fa582d62
d4351a76f2dc4c76129e030e6fa0ce4527240a42
40778 F20101108_AACGNO frank_j_Page_51.pro
e761aedf5dc316f556602f3895029e0d
d33b5936b8d5bee3b0c384ce6a4d84ce4ea5859b
23589 F20101108_AACGOD frank_j_Page_12.QC.jpg
6fc8ae55f9fcf646c9abed6ecf31815d
5471fe50fd2efe2f14342bcd4991e96640bfc868
843 F20101108_AACGNP frank_j_Page_81.txt
f1b3d2e9a4b319c493ebfcb4745d98d4
811ff0a1d00398f0d9e0bd5c7cf3e4d58d7627d5
10069 F20101108_AACGOE frank_j_Page_56.QC.jpg
822a43ec2f312a35d9c989c109a85e13
04bc317870ce7d76719143fa5b9be5d689954c33
79917 F20101108_AACGNQ frank_j_Page_62.jpg
e375a936174f4ec5a9297cb694260532
5f5bcb6b3f5ebcbd25f63018ccc4c7a253bf6d56
343772 F20101108_AACGOF frank_j_Page_85.jp2
8723d327e31e07ecdbb279fa0098085c
5d8424f03c83b0c5a6c30113825f64db590b9b16
23619 F20101108_AACGNR frank_j_Page_37.QC.jpg
56fddbb433207fa996a80b68bf412b01
4b60762e13b556b04c4906d38f638476b6aee2b2
7291 F20101108_AACGOG frank_j_Page_81.QC.jpg
4611fe314eaa2616cca20009e6151f0d
1b2fdd67d3b4c0cec7ec96664d79a4304484c8bf
F20101108_AACGNS frank_j_Page_74.tif
17708ee6afe41d1c04b457610f9e7908
1c136c0adc3a6ed1af979146b0e76ee468f87ee8
33816 F20101108_AACGOH frank_j_Page_74.jpg
2222f480f57f3fd295239c726e5cbbff
105a07a48ab6e75da3ffe276fc458103412a5ad9
46961 F20101108_AACGNT frank_j_Page_69.pro
68fd0a36940ef9dc42691aed8f6248a5
65e7d04850ad5b7f7b5def4b3d7cf56a75983c35
F20101108_AACGOI frank_j_Page_23.txt
83cbe1160cb52b83c2e7203cfd5bf7b9
314e5f7b467211d39cdd1af4f1d2389d0a62e979
1875 F20101108_AACGNU frank_j_Page_71.txt
5de97e6996be6f4fb3cd7ff94bb4a56b
775c6703a84f945aaaa285f60ffcac803911373d
21831 F20101108_AACGOJ frank_j_Page_81.jpg
333bb1e89bc72a9cc842d225fb1ccc05
ac2148076c07eeb6288fa897f6f936620e9663b0
37598 F20101108_AACGOK frank_j_Page_56.pro
8ee5fea0cddd22f669197a35e8fd7e72
99973e094e4491a0f28eb20aca81b3f209f26361
7036 F20101108_AACGNV frank_j_Page_93thm.jpg
1a545bc25c70fe20b5bb2b12534aedb9
b7043f87c7213d83bd1523954598a24d05b62f14
6970 F20101108_AACGOL frank_j_Page_22thm.jpg
569bb730ac1c834851381426273946df
df63bba69f63ca3fa7ccbeae21da0d92f82cd787
45810 F20101108_AACGNW frank_j_Page_11.pro
8920cb6cc69453d4a052ed322858062e
de2169f1a512d9fc1afe79461585950db177ff17
23736 F20101108_AACGPA frank_j_Page_69.QC.jpg
1ce4f634ce5bba149b5743c843562a49
c4dc954d1ce4a08beb450428fd5bc277bcca6893
25893 F20101108_AACGOM frank_j_Page_32.QC.jpg
8eed8fc6f51d2c5dfeb5b474ef461be1
be3aa487660e0bdda44e45c420a872bd90ddbad9
F20101108_AACGNX frank_j_Page_09.tif
cc07503e503ee9bbbc0c8b4124a26a6a
cf38fe087e82d1eee3f46a72d2b61b0fe64b6608
3197 F20101108_AACGPB frank_j_Page_80thm.jpg
3c6723570d401f2c90306622437b4eac
c294259312f1278c4b77c4aa343eb0a6b6735f0d
F20101108_AACGON frank_j_Page_20.tif
0cf8a4866d88dfd9acb07681c20e7fab
cf34845de7dc69629aa51d10bd1afdcba3a154a3
7027 F20101108_AACGNY frank_j_Page_72thm.jpg
ab89f1a4a440c7e11bad044f5ebdecad
f31383fd50afd1538cef704160438a316f99bec1
1051978 F20101108_AACGPC frank_j_Page_06.jp2
aea53b5ecd3dca3fb031bafc7f2a8625
298e3059d3c7e986e40fa715071beb0086df3362
1488 F20101108_AACGOO frank_j_Page_03.txt
11a445327bba8b864699783f23fb7dd0
e4870acfb49ad497b791d061b94fe3fccc8425c1
30789 F20101108_AACGNZ frank_j_Page_07.QC.jpg
d7a67b99f31068d2cc976e34f78d4394
51f027cbd14e40854ecec6ef556e0914d0f59e7b
802 F20101108_AACGPD frank_j_Page_79.txt
151c4c86b2676a4795a93874a376bd08
aab735637f7919f777aa2e377bd35c61a23b6338
1628 F20101108_AACGOP frank_j_Page_56.txt
02e8f9e56afa6dbc36ed9da1945cb289
eed5bd75ac8af344f898fac39cb585a9ce773ab9
F20101108_AACGPE frank_j_Page_46.tif
3265d7044559c0574a4eedd17cc207d3
f6876cb3a9e03cf834f252ac11e7ecc366419d00
23319 F20101108_AACGOQ frank_j_Page_20.QC.jpg
ea34d2675a55380b25eecb4bb1d65b3f
1453756c56271a94789e6644f0b0b622223e3f96
6836 F20101108_AACGPF frank_j_Page_26thm.jpg
edfdbdb61f1ec927cbf50298dc4cb1ea
aa5f2383de62d3ff20ad9e0dfa4437a9ffd4948f
F20101108_AACGOR frank_j_Page_48.tif
8d66385dcc2713a1a343268478496d37
8e43cbd783d9b54e20964736b4a5a9e65ac226a7
52844 F20101108_AACGPG frank_j_Page_63.pro
dfe6b8f6e6ae97fcbea048e755a9851a
b0fac8d60bff570d167abc6c7f67710f61658ad1
2807 F20101108_AACGOS frank_j_Page_54thm.jpg
0f057811d10d93387aa7aab2244da375
d8a0b6c10a1b15249f9616cea70c132b704bc553
120452 F20101108_AACGPH frank_j_Page_66.jp2
eef3f684d541ab9a01dfcceb4ffddbac
090f0d399d0ef6a166649733da796eddccbdbe62
3356 F20101108_AACGOT frank_j_Page_50thm.jpg
0e6a28b01e3b31b5767cc8be3acf5410
15a5988cf8d54679c3682b36224fffecbf34426c
1859 F20101108_AACGPI frank_j_Page_69.txt
8f70e316c17d1ae71fd15076fa621437
15fb95f53f1f41056e49b27484088658817ff181
2028 F20101108_AACGOU frank_j_Page_15.txt
ce8cef1390a1c53e6f589815b6d5e0b4
3032e8b472ba8266e4176b727b25b65f4dd3bc31
8954 F20101108_AACGPJ frank_j_Page_59.QC.jpg
37fae1f573b3bf70c15dc0be1fc72888
ed8f2f638dd4f034a00425a5ccc8e0d2ce14e990
44386 F20101108_AACGOV frank_j_Page_79.jp2
ef78ba76bb5963b9e2625cbcb0853387
809513642db10b0cd9d73b5f04c891e8af9105e5
116596 F20101108_AACGPK frank_j_Page_35.jp2
b280f394d61dcb1d7ef75eb647375aa4
745b6e8aa3bd461b98a4bcd74dcde858e9f015e3
8870 F20101108_AACGOW frank_j_Page_49.QC.jpg
eda92a676764cf566aaa5cccf9c72b15
71b7b8242ea05e545b1ff12523c296bc3181e3ce
11458 F20101108_AACGPL frank_j_Page_46.QC.jpg
dc9759e6d213c5e27173f3fa1278c512
1a2d740be3b8b45f766c53b53c5b665f254574f6
94735 F20101108_AACGOX frank_j_Page_92.jpg
c9d5a16c05dceb98564679fc79ecd56b
e57db73f37755943721d55533d16c17d7a60d561
56628 F20101108_AACGQA frank_j_Page_66.pro
75d719faec7a663a8f7c5aaceb9f81f9
a492c90743acd9a06cebe72dd29eccefbefa1c71
7054 F20101108_AACGPM frank_j_Page_68thm.jpg
46f9f3922ea652109fd2c7a5658d21b3
d3895bdff2dc39458d5d6a4a309280c09bf408a3
17525 F20101108_AACGOY frank_j_Page_84.jpg
93a95bc04ea8e35b9c29224cbd9b8e79
3c9e2c5ad0d492690c265505d5d35ae5cb2dc86c
1041 F20101108_AACGQB frank_j_Page_78.txt
4aa8fd5fc6bf7b97c91caf1031fbc723
ef3393039f07124135b599fb353183c0f9ad459c
100 F20101108_AACGPN frank_j_Page_02.txt
2792d346c2c43083ac377af2d5de4aa4
220b107f0db1ce0c48b23d732c2394b90c8ac721
F20101108_AACGOZ frank_j_Page_37.tif
03ac8850679b174f45264bbd1bd2916c
a32d6870bbf2e5bdceb660acb7bef8207fb44fc4
52746 F20101108_AACGQC frank_j_Page_18.pro
683fd2fdad4693a253ebc98551eefe4c
7fa2128506c799d1f89b9e78a873dc04dd5fe2cd
6238 F20101108_AACGPO frank_j_Page_69thm.jpg
e409d76cbe7e02a5602fb13f735ce48b
8c7d1d074d71c65bb3b827a6973dc476567e28cf
F20101108_AACGQD frank_j_Page_42.tif
da702d7c719b3f841a0194cf5b435603
5b94293aa28d7071c4ad45c9e69342ee7f9902ae
28447 F20101108_AACGPP frank_j_Page_47.jpg
9810364b7312391f4a695510d52f39e9
1c1793b5384a879e1f179903deb9a7f6a170ceee
70463 F20101108_AACGQE frank_j_Page_95.jpg
7fbe2260ed0da45375e39d53a309d3e0
5862b79f335cf8b8fd5aa6ee85355c7fa611361f
98568 F20101108_AACGPQ frank_j_Page_94.jpg
23bee034a49fd4a52b9b5d3412c304b2
a95f0f4e4deb3903a4bcf750908039f61a1957bf
113039 F20101108_AACGQF frank_j_Page_42.jp2
ffdddcc34c37f950e5e1bdc0eac8f8aa
679b40a690949bfd933e7a56a03fb52892218596
111668 F20101108_AACGPR frank_j_Page_44.jp2
6779199f44853944db5737fa28c7f14a
a0d995c108a66de2d6c34fd30ec9fbe4dc7db8c6
4331 F20101108_AACGQG frank_j_Page_84.pro
da45feef9bc7ec8dff0734eab5e7f898
ddef8fbb9cc81904316d419aaf64891eec3317ea
3404 F20101108_AACGPS frank_j_Page_91thm.jpg
b9e2e0f3018a973932d07411e496f552
0e792e227a281246f3c8b421d07210d9a59ac70e
43671 F20101108_AACGQH frank_j_Page_96.jpg
54a4e60ca611c25141f0226ad9c44efa
2e267ec58d2c5edb77814e628c3ab6d55af6a2bd
F20101108_AACGPT frank_j_Page_55thm.jpg
9e52d0f6ef634eaa7a1463d12eab581e
9a7daec2b65fd7893a975c84a3764894404df83f
30965 F20101108_AACGQI frank_j_Page_53.pro
c534bec86cde5f14131c19d8cde8992f
1d44bca73abc4bc24da28202e115fe9141324caf
7043 F20101108_AACGPU frank_j_Page_32thm.jpg
f0c0c455275b41afca97c2acace3787b
523f6d6ec010944a77321a4be7ad3e63a97b7f6e
74923 F20101108_AACGQJ frank_j_Page_27.jp2
080308d57e0d288e4fcae6833fb81eb2
3641643406313f74d7e9639793411eb9b963d558
109513 F20101108_AACGPV frank_j_Page_83.jp2
c71cda73e3028f73bea2482d50464255
b0da83e2c3191e9cfc3bb6b442886564ac72f4ea
F20101108_AACGQK frank_j_Page_43.tif
22e76865fcb6f7c30bb8359ac470bca6
8292816a67157d1fcdf8b35de66faf0e4aa911ff
19025 F20101108_AACGPW frank_j_Page_86.pro
2892780f6dc83e7445db27c3478707e0
119195c982df7f33c947d2924bf1f5ec8806e952
34411 F20101108_AACGQL frank_j_Page_08.QC.jpg
d6d4c92fab042506ed99928191f0837f
5aa0773e715187c194d9a36cbdacbf13b9e5840e
2440 F20101108_AACGRA frank_j_Page_58thm.jpg
53cbd94ba91a41b7064a183a81d2f29c
d23becc6c4adcf5c9a71b2b1160f00546af132bb
1448 F20101108_AACGQM frank_j_Page_74.txt
c98076ea5a928123173cac63ab496c6f
39122f0d4ded03bbaea879e3d4fc53900b415653
1090 F20101108_AACGPX frank_j_Page_96.txt
e3f7a5ffc3905ec8b7757fc2e42895fd
e46745db5b89fdd7cb15db56259740bf02208136
1742 F20101108_AACGRB frank_j_Page_87.txt
c9fdc4f338a64620d4556eb04985a706
43c924f06ea4eb355b8993ca801f59d79c48bf63
6748 F20101108_AACGQN frank_j_Page_16thm.jpg
446b8ca49004331ea85bb91317270222
1dfaf0d0fdd5d8889d48419271f6bccd94255201
73046 F20101108_AACGPY frank_j_Page_12.jpg
76aa733f2f6f7337f8bd2f9a260aa02a
d24a60ec9c8faafeaf23acc84fa1cb69d0b1d304
F20101108_AACGRC frank_j_Page_51.tif
a74e0a1f0cd97ee83ed4ebbf94988b3c
9ab1335289a8ac6bacca0359ecc84f08e75a069a
6690 F20101108_AACGQO frank_j_Page_40thm.jpg
78b112560592a18a2f5e9b8054851c7e
69d01a38a3901f5880ce006f9537f267d666d57a
F20101108_AACGPZ frank_j_Page_81.tif
81cb2b7ce30511cf9a1dbe92013ccb20
9f5851bff1d077b8fdae3646e1765b184a341ff8
27052 F20101108_AACGRD frank_j_Page_58.jpg
fad45b05dabb211267d46caba965e00f
c4eb87d9b7b1d94f0ccc2e934853ecb6be3dae5f
15456 F20101108_AACGQP frank_j_Page_87.QC.jpg
39a15a20cfbd22a51972272aa21869f5
b2ed08b0e5c43475207a45daad03c8f5eca733ae
36165 F20101108_AACGRE frank_j_Page_56.jpg
26d82eaf060745952fbd82ff7efdce5f
4eaafc3de35b9e249a4bf61f5506a4f0b2abc69c
55201 F20101108_AACGQQ frank_j_Page_14.pro
7be90055e9fc76109c8c6ad942740240
321e6cf2e4f715dfd853973ac5c60c918eab45b1
7599 F20101108_AACGRF frank_j_Page_34thm.jpg
77d87948a59b3ada1bc2f41099d6460f
7ad3a080d47c53fe194b863558259fac9171e2a7
F20101108_AACGQR frank_j_Page_69.tif
35facf357a0a64a5c506b0c504b7f8ab
9addebec6cb570394ac888cd8cbe78c57787a5d3
25646 F20101108_AACGRG frank_j_Page_28.QC.jpg
5b28ab1fcc160e7b4d3b21f8709bd1fd
3769f5667f29057be54b717724edcdbad06e2dca
79483 F20101108_AACGQS frank_j_Page_22.jpg
c1a5d086e26847756eb251caf4a5907b
54107ef355a0555074fb2f92e9060229e81e5e95
F20101108_AACGRH frank_j_Page_22.txt
2b07d009a78e1aa66c672b234a19604e
a57f6ddd13ee0c3386c1b36b51281ff527a90ee6
50933 F20101108_AACGQT frank_j_Page_41.pro
a6a500b504cde529f980a8af85397a17
9219b547ebd478953ad7dd0e39de9edf26d059f1
13810 F20101108_AACGRI frank_j_Page_81.pro
af1a458c678a9fad6380833389770164
18d9c47c3aaee77d526f58d388dca7599b04e4ec
53134 F20101108_AACGQU frank_j_Page_21.pro
65923c360872717e533c2311b4186fb2
052b49981996d1e282d4251b8cf6505ee768e3f7
27267 F20101108_AACGRJ frank_j_Page_61.jpg
d22dbc91e13da818151218a1dbc42e5b
e60e98ccd96dc30a947d8e6d012f03fbd3a64fc8
74810 F20101108_AACGQV frank_j_Page_21.jpg
2f6d7c388d6590bbca83bc347fcc291a
ce5a96526962c8757c772a46839b3821d735c5d0
6534 F20101108_AACGRK frank_j_Page_41thm.jpg
5e7008ebeb0b5562e498c882fec5aeff
f1e1880ad5b97e1aa43da91a052dbdb30c6e003f
25399 F20101108_AACGQW frank_j_Page_18.QC.jpg
3547d0462d11c3e139e46ae05b5173a6
a95cc029e4a7274482e956da04da5e268060f2bf
25929 F20101108_AACGRL frank_j_Page_31.QC.jpg
e57abec7344b7f9cc29c4c50d417d64a
230900ad1cbc087f99a631efe656e8722fece3b4
24479 F20101108_AACGQX frank_j_Page_78.pro
97d2c1892d5bb9869549dbe7472f5abf
4a8e0dcbb88a165ffbe5cad0bb083f3bd8dacce3
36323 F20101108_AACGRM frank_j_Page_03.pro
f62ad6e4aabe1472a3a7980e518676a9
03aa070530b16670bf79f86645f3f98b0df66df0
50864 F20101108_AACGSA frank_j_Page_24.pro
46f817dd5040a9024b1e600548981d87
82da3e2e0544f8083bc4d4bf29954e1662784c64
4408 F20101108_AACGRN frank_j_Page_87thm.jpg
eb94647e7f06bb6c4bba1fea9682e4e5
3ecc515d0721da695a284f176fef7b723627bc31
F20101108_AACGQY frank_j_Page_08.jp2
1ca935ea2c5d214d675b865990cc3374
121e30d95f17bc960394945f40b03f573ec3de08
3490 F20101108_AACGSB frank_j_Page_75thm.jpg
a5f0374ed070d772653cb53326dad0a4
a56b3013b155dce4ddfee4676c9a81906ba9391b
76247 F20101108_AACGRO frank_j_Page_42.jpg
eccf737e156e1e9fc9e745687f2a4cd2
c5da597a1488f2699a30af392112a54f77929827
27905 F20101108_AACGQZ frank_j_Page_60.jpg
23db8500a7a47231ac5a740902cb2308
6a8eb7431d230f33f0933b5ecc6f9dcf0def59c5
16524 F20101108_AACGSC frank_j_Page_90.pro
b587d53784e93644ef5a7e81396452f4
60d4cc0e30fc6ddb5d40be7ae26ebcb0194e7769
6151 F20101108_AACGRP frank_j_Page_89.QC.jpg
eb1ba843fdff6bbf1307adceb3f31cd1
2a290253ec97c6b1b4e8fd96b2706020123c9c7e
10809 F20101108_AACGSD frank_j_Page_57.QC.jpg
b5c7ab3e80f587697dfa6b5827bc55df
34e0e04e5cd4a3f45d5192e2dd40c6c900fb3c07
F20101108_AACGRQ frank_j_Page_33.tif
1bab7085cc9fbbd2fe037c9c4494f5fa
37d1c9b878a2709f1bd340aa40c1ef6a9b4d1fd4
22875 F20101108_AACGSE frank_j_Page_71.QC.jpg
79f065d53f23e4195a4606bf87869cac
02cc98d453ea86c34546d219f71ecf842b89ea84
1355 F20101108_AACGRR frank_j_Page_76.txt
beea4560bfe02edde0187a5de82f325b
dafdb59ec4a03081cd0d033fa288fd0ae24bb994
F20101108_AACGSF frank_j_Page_16.tif
0fcdf270ce32a7aa4615b7d8dd847b1e
7f5de54d9a81c716661980b6ccd10c691b1460c4
F20101108_AACGRS frank_j_Page_17.tif
7fa11e7d2c92c8011b6419e4fbd352f9
0c9ffe771d7f2dd483ddee17399810e737c5d8f3
60311 F20101108_AACGSG frank_j_Page_92.pro
659d49e305ca79076fbb245fddba8218
2d22ac889ac6cf3a8a31b6e7c9fa0189bb6d07c8
128912 F20101108_AACGRT frank_j_Page_93.jp2
44d4c897b9664b677cc08e2cb24c2b15
0bf38a58a17c1e3f94f712a7b79e8d56a092292c
48715 F20101108_AACGSH frank_j_Page_40.pro
89c408109f2f5bb1c15ee2bd47320e7d
8aecc23e081caeaafd61fa7bbcd24b825cef4a54
69619 F20101108_AACGRU frank_j_Page_88.jp2
19fa698e3dc3cde99dcbfe2451220ce5
6ece80cca4afc390cb01c763b727408ec862e5b6
126089 F20101108_AACGSI frank_j_Page_06.jpg
fde94c6fbcd9799ca0868dc65de0bc5a
7cef0f2132ce1bf3e1d27cf665d39e86788f8746
79581 F20101108_AACGRV frank_j_Page_36.jpg
ebad559349d627c324f8fd47c36fe9af
ec40b34bf7633957558eb22b854fd60658c96d66
101623 F20101108_AACGSJ frank_j_Page_11.jp2
c120545a80df261479aa498e33ca7707
af37b3f5f974ac70a28e39cee80ebf2be6e6b5f5
112327 F20101108_AACGRW frank_j_Page_24.jp2
89d57869825faa383eac6408d9bb88a2
17bdd745a081f2ba4f89488f3822df96d46e8e77
F20101108_AACGSK frank_j_Page_94.tif
7f60f9e24c8d73bf21454ece4a1627d1
623efe74c3b879b351597252462996dcf107778f
1414 F20101108_AACGRX frank_j_Page_45.txt
e1ebe38273acb40f1225a466eb5d97d2
54227911efd492574e1802c11deb04fc4ef8b697
22025 F20101108_AACGSL frank_j_Page_01.jpg
75a97b29c965e61d2dd937b23d3c996a
e9b28f00ebae7f2978645618f9b9582097e0e306
F20101108_AACGRY frank_j_Page_24.tif
dd93ed2753f0b1e7c29bbd3bc4e9e710
abdd859415d59e196eaf48c4093038f464440d30
25435 F20101108_AACGTA frank_j_Page_52.jpg
97fe22c77a01c9d933c954bfa5e972d5
22fa0b96caee76a8edb8f2ab8d2a0d8f61fae0af
110957 F20101108_AACGSM UFE0021869_00001.mets
1cc5f245f24f68cdecab738cb3bcd40b
331a0767fe7817d98e79f81eb3c65d16bd8d5a61
31056 F20101108_AACGTB frank_j_Page_53.jpg
23428d7feceabaec995612911d544d83
85ab61f423631e508dcca8914253b7f46fa577b9
F20101108_AACGRZ frank_j_Page_36.tif
22a0cbcc6528ec9c27b65182abba4842
4fcbb41e72fbe8ef95d647f662e717e37c04b622
40474 F20101108_AACGTC frank_j_Page_57.jpg
4c4e74e5d0fb35af753bba6e6600c63e
53b2b27a2ca0b1b4aa9835333bf3b127b6ea5c05
13353 F20101108_AACGTD frank_j_Page_73.jpg
9db9bfd638bad8a3e71d7f0bbdca4e58
1454dc498451765329a67e15f6cdd6d3f8ef6d30
57102 F20101108_AACGSP frank_j_Page_03.jpg
2a70bc1ae9ffe1c90b8c5f9ebf9f4f1e
52de897ae04d7601d971a272c2f7970f3c6200f0
72673 F20101108_AACGTE frank_j_Page_83.jpg
4973207c8bb1f6add25e41a9e7c3706c
62ef180736394e5e229594a54ff0b512514b246e
74455 F20101108_AACGSQ frank_j_Page_19.jpg
1e37ea46ebd0ccb182743892c4f98b1f
17bd596744f254a515fea3abe6a3a68221166643
41728 F20101108_AACGTF frank_j_Page_91.jpg
e674e3c8b3baf8b4eff5b7335afd0349
89fd26314e0b8143220cf00224fe3fa5f156dbd0
71650 F20101108_AACGSR frank_j_Page_20.jpg
d1dab2af67227bf8dc6e76e050cbfeb0
21e69eee39a7791ff67d1968879495e35fd5c622
1051984 F20101108_AACGTG frank_j_Page_04.jp2
7cab3bfdce8ed09f7d57227441f94025
8210b34978959d062478cf7c4c00fe0fc42e2a92
77614 F20101108_AACGSS frank_j_Page_23.jpg
1798b66a3db999c421ce23643fe5e32c
73dd2757f4ed54b892e17d0600984a829cc0405d
1051950 F20101108_AACGTH frank_j_Page_05.jp2
84dbfe5295ba8358c50911e49e6c4a95
6b71605f83a2c174c64b7e5a3fb0b73e2a754d89
73169 F20101108_AACGST frank_j_Page_24.jpg
d2cce2dd1964fe56c74d45c0642b8525
61679441f256f260f9c00063c2f2983623dfbe0d
96221 F20101108_AACGTI frank_j_Page_13.jp2
54ddd66be3bd8f413aaafee8c79d7aee
18cf5b1f9ece36a4f1594f3a7e03cf1c660a0523
79098 F20101108_AACGSU frank_j_Page_32.jpg
b772b2ba3276801c451212b50092e702
beddcf478a48e58fd7d00e2d456098d3dac6bf7c
109839 F20101108_AACGTJ frank_j_Page_15.jp2
2d534174614393a929f35e2a91904d4c
aaf030a6de4207fe8ee3f8951c8c812469e6ac49
90114 F20101108_AACGSV frank_j_Page_34.jpg
0e0738f46b364b8f953c6e989b94e634
442d587c1ed2bfc52af0dbf1030d5d4d9cd5cea8
115270 F20101108_AACGTK frank_j_Page_18.jp2
550bb5b57aa41980112a8c9fcef3f3f3
a0715fca430a2ca429c6c2c5effe27507ba6cbfd
76332 F20101108_AACGSW frank_j_Page_39.jpg
b538b2ff5b768041a7c997a61bc7806d
c43d3b80ff13a9664c2cbfae9f886b6851176e18
106397 F20101108_AACGTL frank_j_Page_20.jp2
f663fb1208a6ff9f93579e118b1d64bb
56948fd97a6dd90b34d182e141ba1f2286ce1436
48558 F20101108_AACGSX frank_j_Page_43.jpg
812ef345b24077e8afc2829e8f30af7b
294daecbaaa1478a4a6dcb121f0bd13c313b2561
58004 F20101108_AACGUA frank_j_Page_78.jp2
7178d7ef72c29f547c5cf6e6dd712c26
97f6fe2f728e449cbd3e24a7b5c306173c67de85
120704 F20101108_AACGTM frank_j_Page_22.jp2
afd9232d7cc32f4ebbb4a3976fbf714e
289340dc71b9564809813bfe6b3e7cd64eac6397
46828 F20101108_AACGSY frank_j_Page_44.jpg
1ed65dab95731a59ea92040e5b69fc53
eb8562ad2dc8dfde2e1b6c434b762d2234a4a5dc
15531 F20101108_AACGUB frank_j_Page_84.jp2
b79fef51be43295396b4dbf7ba9201b4
a1118343773d6ae6ffc9be2eb5f5f66027d21c69
118292 F20101108_AACGTN frank_j_Page_23.jp2
4b4bd80066dd3767648e5c41147af67a
1fbffa34ff094306355c192e39138b6650b67d89
38449 F20101108_AACGSZ frank_j_Page_50.jpg
65e52409f492eefae049290405bfa318
5f7860df0fc4890e0f1fc2060ff747d29487d2cd
F20101108_AACGUC frank_j_Page_02.tif
df7f24af4500e97392b89bcf2db52c8d
fe2d58df0340903dc080511d0d427b2b1009f9dc
115513 F20101108_AACGTO frank_j_Page_33.jp2
cfff3773b9206ee1031c041d7fb82460
0c2cce21b01d7116ddc3b5b33c80d0a5c5620cbc
F20101108_AACGUD frank_j_Page_04.tif
0e56187aba22736eacfb287730c1d2ca
dd906b821fabd7c18d663d0a8a71cbe5c62b95f7
1051897 F20101108_AACGTP frank_j_Page_34.jp2
6756e4517706b31dfcef5dcfa887d398
adbf4476f592e520e2fb6671e9535f58e75409c0
F20101108_AACGUE frank_j_Page_05.tif
fff613903dc2ceb1e46bf05f04e0cbb9
3c2c8ffee35b5af70d6c1fd77c8e924ebd941b17
70157 F20101108_AACGTQ frank_j_Page_43.jp2
3329f393db8c967a3960953b64d25f22
339b2ada85cfcf68ab89cac39d217dced02708cc
F20101108_AACGUF frank_j_Page_07.tif
0d20e14c4800e1cc3f86e9e9206162db
ac73a928e88bf6942ab17eb244cb45660a8afd5b
84125 F20101108_AACGTR frank_j_Page_56.jp2
026ffddda83234368a9530fd0c0d446a
c2ded198c7e1aaa59b292602de4ee0d5cdea6dfa
F20101108_AACGUG frank_j_Page_12.tif
7f86b0dc59983a0a1e9c964cb845cb78
255f03f1691218bf627e03120b8ab1b1eaa9fabb
559195 F20101108_AACGTS frank_j_Page_59.jp2
112464e0206357981ea2e546f4a99ce0
461725d35b99c195b1a779ffa440d56f300fc5a3







DETECTION OF TURF GRASS STRESS USING GROUND BASED REMOTE SENSING


By

JASON HAMILTON FRANK
















A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE

UNIVERSITY OF FLORIDA

2008





































O 2008 Jason Hamilton Frank









ACKNOWLEDGMENTS

I wish to acknowledge the guidance and direction of Dr. J. Bryan Unruh for all his wisdom

and support during the last two years. It has been a learning experience like none other. Sincere

appreciation also goes to all my committee members, Dr. Barry Brecke, Dr. Laurie Trenholm,

Dr. Won Suk "Daniel" Lee, and Dr. Jasmeet Judge. I appreciate their helping me whenever

asked. Special thanks also go to all the current and past members of the West Florida Research

and Education Center: Dr. Ken Hutto, Raymond Edwards, Rex Lawson, Chris Adkison, Jason

Ging, Chase McKeithen, Phil Moon, and Dr. Darcy Partridge. I appreciate their help and

support. This undertaking could not have been accomplished without them. Bryan Schwartz and

Dr. Jason Dettman-Kruse thank you for all the assistance in SAS: This paper would have never

been written in time without your help.

I appreciate the support and wisdom received from my parents. Their guidance throughout

my life has inspired me to better myself not only through hard work in my career but as person.

I am grateful to the University of Florida for providing me with the most valuable resource

I will ever have, an education and the inspiration to never stop learning and always strive for the

best. I am a proud part of the "Gator Nation." The time I have spent with the University of

Florida will never be forgotten.












TABLE OF CONTENTS

page

ACKNOWLEDGMENTS .............. ...............3.....


LIST OF TABLES ................ ...............6............ ....


LIST OF FIGURES .............. ...............10....


AB S TRAC T ............._. .......... ..............._ 1 1..


CHAPTER


1 LITERATURE REVIEW ................. ...............13...............


Introducti on ................. ...............13......__ ......

Irrigation Importance ........_................. ..........._..........1
Nutrient Importance .................. ...............17.................
Plant Growth Regulator Importance ................. ...............18................
Remote Sensing .............. ...............18....
Vegetative Indices .............. ...............19....
Human Visual Ratings............... ...............20
Irrigation ................. ...............20.................
Nutri ent ................. ...............23.................
Biom ass .............. ...............24....
PGRs ............... .. .... ._ ...............24....
Statistical M ethods .............. ...............25....

Summary ............. ...... ._ ...............26...


2 DETECTION OF LEAF NITROGEN CONCENTRATION AND SOIL
VOLUMETRIC WATER CONTENT USINTG GROUND-BASED REMOTE
SENSING TECHNOLOGY ................. ...............28.......... .....


Introducti on ................. ...............28.................
Materials and Methods .............. ...............34....
Reflectance Measurements ................. ...............3.. 5..............
Statistical Analysis .............. ...............36....
Results and Discussion ................. .. ...............37..
Influence of Irrigation Treatments ................. ...............37................
Nitrogen Effects............... ...............37
Nitrogen X Date Interaction .............. ...............38....
Tissue N Concentration vs. Reflectance............... ...............3
VWC vs. Reflectance .............. ...............39....

Visual Ratings .............. ...............40....
Biomass ................... ...............41..

Crop Circle vs. ASD ................. ...............41................
Conclusions............... ..............4












3 USE OF GROUND-BASED REMOTE SENSING TECHNOLOGY TO ASSESS
BIOMASS PRODUCTION IN RESPONSE TO PLANT GROWTH REGULATOR
APPLICATIONS ................. ...............62.......... ......


Introducti on ................. ...............62.................
Materials and Methods .............. ...............66....
Reflectance Measurements ................. ...............67.......... ......
S tati sti cal Analy si s .............. ...............67....
Results and Discussion .............. ...............68....
B iom ass .............. ............. ..............6
Volumetric Water Content .............. ...............69....
Visual Ratings .............. ...............69....
Color............... ...............69.

Quality ................ ...............70.......... ......
D ensity ...................... ... .. .................7
Reflectance Indices vs. PGR Application rates .......___......... .........___......70
Crop Circle Device ............ ..... .__ ...............70...
ASD Device............... ...............71.
PLS Regression .............. ...............71....
Conclusions............... ..............7


4 CONCLUSIONS .............. ...............82....


APPENDIX


A EXPERIMENTAL LAYOUT FOR DETECTION OF LEAF NITROGEN
CONCENTRATION AND SOIL VOLUMETRIC WATER CONTENT USING
GROUND-BASED REMOTE SENSING TECHNOLOGY ................. .......................84


B DAILY RAINFALL AND ET DATA FOR JAY, FL................ ...............87..


C CUMALITIVE WEEKLY RAINFALL AND ET DATA FOR JAY, FL .............................90


D MOTHLYRAINFALL AND ET AVERAGES FROM 2003-2006 FOR JAY, FL...............91

LIST OF REFERENCES ............ ...............92.....


BIOGRAPHICAL SKETCH .............. ...............96....









LIST OF TABLES


Table page

2-1 Analysis of variance of soil volumetric water content (VWC), percent leaf nitrogen
tissue concentration (%/N), visual ratings: color, quality, density (rated on NTEP 1-9
scale, where 9 is the best rating and 6 is acceptable), and biomass. Normalized
difference vegetation index (NDVI), Leaf area index (LAI), 880 nm, and 650 nm
reflectance were obtained from Crop Circle device. Crop water stress index (CWSI)
was computed from IR thermometer readings. All data was collected from fairway
height hybrid bermudagrass (Cynodon dactylon X C. transvaalensis, Burtt-Davy),
four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25,
50, and 100 kg ha- ), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept.
2007) combined. ............. ...............44.....

2-2 Analysis of variance of individual reflectance bands from ASD reflectance
hyperspectral data with a range of 350-2500 nm and spectral resolution of 1 nm
collected from fairway height hybrid bermudagrass (Cynodon dactylon X C.
transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated
ET values), four N rates (0, 25, 50, and 100 kg ha- ), and five dates (22 June, 12 July,
25 July, 22 Aug., and 4 Sept. 2007) combined. .............. ...............45....

2-3 Analysis of variance of average range computations from 630-690 nm and 2080-
23 50 nm; NDVI, LAI, Stress1 and Stress2 indices; and 605 nm/5 15 nm, 915/975
nm, and 865/725 nm ratios computed from an Analytical Spectral Device (ASD)
with a range of 350-2500nm and spectral resolution of 1 nm collected from fairway
height hybrid bermudagrass (Cynodon dactylon X C. transvaalensis, Burtt-Davy),
four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25,
50, and 100 kg ha- ), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept.
2007) combined. ............. ...............46.....

2-4 Evaluation of means of volumetric water content (VWC), leaf nitrogen tissue
concentration (%N), visual ratings: color, quality, density (rated on NTEP 1-9 scale 9
is best and 6 is acceptable), and biomass (g) collected from fairway height hybrid
bermudagrass (Cynodon dactylon X C. transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation
rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100
kg ha- ), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007)
combined ................. ...............47.................

2-5 Evaluation of means of Normalized difference vegetation index (NDVI), Leaf area
index (LAI), 850 nm, and 650 nm reflectance from Crop Circle device and crop
water stress index (CWSI) computed from IR thermometer readings collected from
fairway height hybrid bermudagrass (Cynodon dactylon X C. transvaalensistr~r~r~r~r~r~ Burtt-
Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N
rates (0, 25, 50, and 100 kg ha- ) and five dates (22 June, 12 July, 25 July, 22 Aug.,
and 4 Sept. 2007) combined............... ...............48










2-6 Evaluation of means of individual band reflectance from the ASD device with a
range of 350-2500 nm and a spectral resolution of 1 nm collected from fairway
height hybrid bermudagrass (Cynodon dactylon X C. transvaalensis, Burtt-Davy),
four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25,
50, and 100 kg ha- ) and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept.
2007) combined. ............. ...............49.....

2-7 Evaluation of means of ASD reflectance readings of average range computations
from 630-690 nm and 2080-23 50 nm, computed indices Normalized Difference
Vegetation Index (NDVI ), Leaf Area Index (LAI) and Stress 1 and Stress2 indices,
and ratios 605/515 nm, 915/975 nm, 865/725 nm taken from an ASD device with a
range of 350-2500 nm and a spectral resolution of 1 nm collected from fairway
height hybrid bermudagrass (Cynodon dactylon X C. transvaalensis, Burtt-Davy),
four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25,
50, and 100 kg ha- ) and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept.
2007) combined. ............. ...............50.....

2-8 Evaluation of means of NDVI and LAI computed from Crop Circle instrument, and
LAI computed from ASD collected from fairway height hybrid bermudagrass
(Cynodon dactylon X C. transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation rates (60, 80,
100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and
five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined, which
had N treatment X date interactions............... ..............5

2-9 Evaluation of Means of leaf N concentration tissue (%N) collected from fairway
height hybrid bermudagrass (Cynodon dactylon X C. transvaalensis, Burtt-Davy),
four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25,
50, and 100 kg ha- ), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept.
2007) combined, which had N treatment X date interactions............... ..............5

2-10 Evaluation of means of individual reflectance wavelengths at 813 nm and 935 nm
collected from fairway height hybrid bermudagrass (Cynodon dactylon X C.
transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated
ET values), four N rates (0, 25, 50, and 100 kg ha- ), and five dates (22 June, 12 July,
25 July, 22 Aug., and 4 Sept. 2007) combined, which had N treatment X date
interactions. .............. ...............53....

2-11 Evaluation of means of ratios 915 nm/975 nm and 865 nm/925 nm computed from
analytical spectral device (ASD) with a range of 350-2500 nm and a spectral
resolution of 1 nm, collected from fairway height hybrid bermudagrass (Cynodon
dactylon X C. transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation rates (60, 80, 100, and
120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha- ), and five dates
(22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined, which had N
treatment X date interactions. ............. ...............54.....










2-12 Coefficient estimates (r2) Of Crop Circle reflectance data Normalized Difference
Vegetation Index (NDVI), Leaf Area Index (LAI), NIR, and Red, and crop water
stress index (CWSI) to soil volumetric water content (VWC) visual color, quality,
density ratings collected from fairway height hybrid bermudagrass (Cynodon
dactylon X C. transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation rates (60, 80, 100, and
120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha- ), and averaged
over all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007). ............. ................55

2-13 Coefficients estimates (r2) Of ASD reflectance readings of individual reflectance
wavelengths taken from an Analytical spectral device (ASD) with a range of 350-
2500 nm and a spectral resolution of 1 nm to %N concentration leaf tissue, soil
volumetric water content (VWC) visual color, quality, density ratings, and biomass
production of fairway height hybrid bermudagrass (Cynodon dactolon X C.
transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated
ET values), four N rates (0, 25, 50, and 100 kg ha- ), and averaged over all dates (22
June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) ................. ............... 56...........

2- 14 Coefficients estimates (r2) Of average range computations from 630-690 nm and
2080-23 50 nm, and computed indices Normalized difference vegetation index
(NDVI), Leaf area index (LAI) and Stress1 and Stress2 indices also taken from an
ASD with a range of 350-2500 nm and a spectral resolution of 1 nm to %N
concentration leaf tissue, soil volumetric water content (VWC) visual color, quality,
density ratings, and biomass production collected from fairway height hybrid
bermudagrass (Cynodon dactylon X C. transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation
rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100
kg ha- ), and averaged over all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept.
2007). ............. ...............57.....

2-15 Partial Least Squares regression on hyperspectral data from ASD device with a range
of 350-2500 nm and a spectral resolution of 1 nm for prediction of soil volumetric
water content (VWC), percent leaf nitrogen concentration (%N), biomass, and visual
quality collected from fairway height hybrid bermudagrass (Cynodon dactylon X C.
transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated
ET values), four N rates (0, 25, 50, and 100 kg ha- ), averaged over all dates (22
June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) ................. ............... 58...........

3-1 Evaluation of means of biomass (g), volumetric water content (VWC), visual ratings:
color, quality, density (rated on NTEP 1-9 scale 9 is best and 6 is acceptable), from
fairway height hybrid bermudagrass (Cynodon dactylon X C. transvaalensis,
'Princess 77) over five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007)
combined. ........._ ...... .. ...............74....

3-2 Evaluation of means of Normalized Difference Vegetation Index (NDVI), Leaf Area
Index (LAI) computed from Crop Circle and ASD reflectance data, as well as Stress1
and Stress2 indices computed from ASD reflectance data from fairway height hybrid
bermudagrass (Cynodon dactylon X C. transvaalensis, 'Princess 77) over five dates
(22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined ................. ................75










3-3 Coefficient estimates (r2) Of Crop Circle reflectance data, Normalized Difference
Vegetation Index (NDVI), Leaf Area Index (LAI), 880nm, and 650nm, and, ASD
reflectance data computed into NDVI, LAI, Stress1, and Stress2 indices to
volumetric water content (VWC) visual color quality, density ratings, and biomass
production from fairway height hybrid bermudagrass (Cynodon dactylon X C.
transvaalensistr~r~r~r~r~r~ 'Princess 77) averaged over Hyve dates (22 June, 12 July, 25 July, 22
Aug., and 4 Sept. 2007). ............. ...............76.....

3-4 Coefficient estimates (r2) Of ASD reflectance data at individual wavelengths, as well
as range averages from 630-690nm and 2080-23 50nm to volumetric water content
(VWC) visual color, quality, density ratings, and biomass production from fairway
height hybrid bermudagrass (Cynodon dactylon X C. transvaalensistr~r~r~r~r~r~ 'Princess 77)
averaged over Hyve dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007). All
numbers in column headings are expressed nanometers (nm) ................. ............... .....77

3-5 Coefficient estimates (r2) Of range averages from 630-690 nm and 2080-23 50 nm
and ratios 605/515 nm, 915/975 nm, and 865/725 nm to volumetric water content
(VWC) visual color, quality, density ratings, and biomass production from fairway
height hybrid bermudagrass (Cynodon dactylon X C. transvaalensistr~r~r~r~r~r~ 'Princess 77)
averaged over Hyve dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007). All
numbers in column headings are expressed nanometers (nm) ................. ............... .....78

3-6 Partial Least Squares regression coefficients on hyperspectral data from ASD device
to predict volumetric water content (VWC), biomass, and visual quality (rated on
NTEP 1-9 scale 9 is best and 6 is acceptable) in fairway height bermudagrass
(Cynodon dactylon X C. transvaalensis, 'Princess 77) averaged over five dates (22
June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) ................. ...............79...........

A-1 Description of treatments for Figure A-1 ................. ...............86..............

B-1 Daily Rainfall and ET data for Jay, FL............... ...............87...

C-1 Cumalitive weekly rainfall and et data for Jay, Fl ................ ...............90.............

D-1 MothlyRainfall and ET averages from 2003-2006 for Jay, FL .............. ....................91










LIST OF FIGURES


Figure page

2-1 Partial Least Squares regression on hyperspectral data from ASD device with a range
Of 3 50-2500 nm and a spectral range of 1 nm for prediction of volumetric water
content (VWC) collected from fairway height hybrid bermudagrass (Cynodon
dactylon X C. transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation rates (60, 80, 100, and
120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha- ), averaged over
all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007). ............. ....................59

2-2 Coefficient estimates (r2) Of Crop Circle and ASD reflectance data computed into a
Normalized Difference Vegetation Index (NDVI), to soil volumetric water content
(VWC) collected from fairway height hybrid bermudagrass (Cynodon dactylon X C.
transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated
ET values), four N rates (0, 25, 50, and 100 kg ha- ), and averaged over all dates (22
June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007). ................. ....._.._................6

2-3 Coefficient estimates (r2) of Crop Circle vs. ASD using reflectance data to compute
a Normalized Difference Vegetation Index (NDVI) collected from fairway height
hybrid bermudagrass (Cynodon dactylon X C. transvaalensis, Burtt-Davy), four
irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50,
and 100 kg ha-1), and averaged over all dates (22 June, 12 July, 25 July, 22 Aug.,
and 4 Sept. 2007). ............. ...............61.....

3-1 Biomass measurements collected from fairway height hybrid bermudagrass
(Cynodon dactylon X C. transvaalensis 'Princess 77) treated with trinexapac-ethyl at
0.05, 0.1, 0.2, and 0.4 kg a.i. ha- Plots measured 1.5 X 3.0 m with three
replications. Means for biomass measurements was computed using Fishers
protected LSD (p< 0.05) from data over 5 dates (22 June, 12 July, 25 July, 22 Aug.,
and 4 Sept. 2007) combined............... ...............80

3-2 Biomass measurements collected from fairway height hybrid bermudagrass
(Cynodon dactylon X C. transvaalensis 'Princess 77) treated with flurprimidol at 0.3,
0.6, 1.1, and 2.3 kg a.i. ha- Plots measured 1.5 X 3.0 m with three replications.
Means for biomass measurements was computed using Fishers protected LSD (p<
0.05) from data over 5 dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007)
combined. ........._._.. ...._... ...............81....

A-1 Experimental layout for detection of leaf nitrogen concentration and soil volumetric
water content using ground-based remote sensing technology............... ...............8









Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for the Master of Science

DETECTION OF TURF GRASS STRESS USING GROUND BASED REMOTE SENSING

By

Jason Hamilton Frank

May 2008

Chair: J. Bryan Unruh
Major: Horticulture Sciences

Many turf managers and researchers still use time and labor intensive techniques to

manage and assess turf that originate from decades ago. Many of these management practices

have been proven repeatedly to work in a variety of situations to assess turfgrass stress, but may

be time consuming and inconsistent. However, there are new developments in ways of assessing

and mapping stress that increase the efficiency by which one manages and assesses turf. With

increasing water and other environmental restrictions turf managers and researchers need be

aware of technological advances in their industry.

Two Hield experiments were conducted at the University of Florida, West Florida Research

and Education Center near Jay, FL. The first was the detection of leaf nitrogen concentration

and soil volumetric water content using remote sensing technology. Plots were arranged in a

randomized complete block design and treatments were arranged in a split plot factorial design

with four N levels split across four irrigation regimes with three replications per treatment. The

second experiment used remote sensing technology to assess biomass production created by

applications of plant growth regulators. Plots were arranged in a factorial design observing three

different PGRs, each at four incremental levels with three replications per treatment. Spectral

measurements were taken from two remote sensing devices.









A ground based vehicle mounted optical sensor Crop Circle (Model ACS-210) (Holland

Scientific) which produces its own light to compute a normalized difference vegetation index

(NDVI) which is calculated as (880nm-650nm) / (880nm+650nm). A hand-held hyperspectral

spectroradiometer (FieldSpec Pro; Analytical Spectral Devices, Inc., Boulder, CO) with a

spectral range of 300-2500 nm was also used to collect reflectance data. All reflectance readings

were taken in full sunlight between the hours of 1100 and 1400 Central Standard Time (CST) to

minimize variance caused by solar radiation. Plots were ground-truthed using time domain

reflectometry for soil moisture content (VWC) and clipping samples obtained from plots were

dried, weighed for biomass, and tested for Total Kj eldahl N content. Visual ratings were also

taken.

The results from the nitrogen and irrigation experiment indicate reflectance data best

modeled soil VWC, visual ratings, and some treatment effects from N. Highest correlations for

VWC were achieved from PLS regression on ASD data (r2=0.71) and the Crop Circle device

(r2=0.70). The results from the PGR experiment modeled the biomass reduction expected from

increasing rates of flurprimidol and trinexapac-ethyl compared to the untreated control, but not

ethephon. Similar trends were observed in NDVI and leaf area index (LAI) indices computed

from Crop Circle and ASD instruments, as well as several stress indices computed from the ASD

device.

If remote sensing technology could be used to adequately assess turf parameters, and was

coupled with the global positioning system (GPS) and geographic information systems (GIS)

technologies, irrigations and chemical applications could be applied on a site specific basis. In

this setting, turf managers could potentially reduce inputs thereby reducing cost and

environmental impact.









CHAPTER 1
LITERATURE REVIEW

Introduction

The maj ority of turf managers and researchers still use outdated time and labor intensive

techniques to manage and assess turf. Many of these management practices have been

repeatedly proven to work in a variety of situations to assess turfgrass stress, but may be time

consuming and inconsistent. For example, in the turfgrass research field visual evaluations can

be labor intensive and variable among evaluators (Trenholm et al., 1999). However, there are

new developments in ways of assessing and mapping stress that will increase the efficiency by

which one manages and assesses turf. With increasing water restrictions and other

environmental constraints turf managers and researchers need be aware of technological

advances in their industry. Recent environmental concerns, such as nutrient and pesticide

leaching and runoff, have pressed turf managers to reduce inputs used for turf maintenance (Bell

et al., 2002).

There are a variety of new or improved methods available to detect water and nutrient

stresses before they are visible. The basis of this technology relies heavily on the global

positioning system (GPS), geographic information systems (GIS), and remote sensing

technology. Much of this technology is in the beginning stages, however, the theory for use of

this technology has been proven in other agronomic industries such as food production systems.

This technology has been used as a management tool to maximize crop yield and minimize cost

through site-specific application of inputs by applying nutrients, water and pesticides only where

needed (Lee et al., 1999).










Irrigation Importance

Water is the primary requirement for development and survival of turfgrass and many

other agronomic crops (Turgeon, 2008). Plant water potential is essential for plant growth as

well as most physiological processes the plant must perform (Idso et al., 1981). Therefore, water

availability is one of the most critical factors in plant health (Ripple, 1986). Plants are made up

of cells, which are filled with water to maintain their turgor. However, if turgor is not

maintained the plant begins to wilt (Unruh and Elliott, 1999). Many plants contain 75-85 %

water and begin to die if this percentage decreases below 65 % (Unruh and Elliott, 1999). Water

is taken up by the roots in the soil and is moved throughout the plant and eventually released

through the stomata through a process known as transpiration (Salisbury and Ross, 1992).

Stomata are responsible for gas exchange in the plant and they close when the plant is under a

water deficit or stress (Salisbury and Ross, 1992). This ultimately reduces photosynthesis and

stunts leaf growth (Unruh and Elliott, 1999). Geeske et al. (1997) found that this stunting in

growth is the plant' s defense mechanism to reduce nutrient and water requirements. The

reduction in leaf area serves as a beneficial factor that ultimately reduces transpiration and the

need for increased physiological processes (Geeske et al., 1997).

In addition, transpiration serves many different purposes such as nutrient movement

through the plant and evaporative cooling, also known as evapotranspiration (ET) (Turgeon,

2008). Evapotranspiration is the total water lost by all transpiratory movement of water from

soil, through the plant and ultimately into the atmosphere, where it evaporates (Turgeon, 2008).

Many factors affect this process such as light intensity, humidity, wind velocity and temperature

(Unruh and Elliott, 1999).

Water stress has a direct affect on the rate of photosynthesis. It first causes stomatal

closure, which causes a reduced supply of CO2. Secondly, water stress reduces water potential,









which causes a decrease in the structural integrity of many important elements of the

photosynthetic process (Hopkins, 1999). This in turn will decrease electron transport and

photophosphorolation leading to damage in the thylakoid membrane and adenine triphosphate

(ATP) synthetase protein (Hopkins, 1999). The hormone abscisic acid (ABA) also plays an

important role in the plant' s ability to endure water stress. Under stress, ABA accumulates in the

leaves causing an efflux of potassium from the guard cells, resulting in stomatal closure and

reducing transpiration (Hopkins, 1999).

Plant water stress can occur if a plant receives too much water or if it does not receive

enough. Flooding leads to a lack of oxygen, this leads to decreased respiration, nutrient uptake,

and root functionality, thus reducing photosynthesis (Hopkins, 1999). Drought stress from lack

of water leads to increased solute concentration in the protoplasm, which in turn affects many

physiological processes (Hopkins, 1999). However, mild drought stress can cause the roots of a

plant to grow deeper. This is why irrigation events should be spaced out as long as possible

(Unruh and Elliott, 1999).

Water is an important resource; however, it is not a renewable resource (Barret, 2003).

Ninety-seven percent of all the earth's water is salt water, three percent is fresh water; however,

two-thirds of this is frozen in polar ice caps (Barret, 2003). Therefore, one percent of the Earth' s

fresh water must meet the water needs of manufacturing, mining, agriculture, and turf not met by

rainfall (Barret, 2003). In Florida the average rainfall does not meet the watering requirements

to maintain golf courses at the expectations of players and industry viewers (Unruh and Elliott,

1999). As a result, the irrigation system is the single most important tool available for turfgrass

management to a golf course superintendent or turf manager (Barret, 2003).









This is why research focusing on improving water use management in agriculture is very

important. It has both environmental and economic consequences, especially as water

availability becomes more limited. According to a survey conducted in 2000 by the Golf Course

Superintendents Association of America, the median annual water usage of a golf course in the

United States is 32.49 cm applied to an average of 31.44 ha resulting in 108 million L used

annually (Barret, 2003).

Historically the maj ority of turfgrass managers scheduled irrigation usage based on their

experience and set schedule of time intervals, both of which can result in over watering

(Augustin and Snyder, 1984). Irrigation schedules are often based on calendar dates such as 3

or 7 times a week. Studies have shown that this style of irrigation may promote over watering

and provide too much moisture to the turf (Unruh and Elliott, 1999). This wastes water and

energy, produces poor playing and agronomic conditions, and leads to degradation of the

environment from excessive runoff (Barret, 2003). Tools are needed to monitor turf water status

to aid in decisions that result in more efficient irrigation (Bahrun et al., 2003). Irrigation systems

should be operated so that the addition of water never exceeds that lost by ET in any given area

(Brown et al., 1977). This will increase irrigation efficiency and possibly reduce plant water

stress in those areas.

Researchers and turf managers over the past two decades have investigated many ways to

determine when to irrigate. Many of the techniques developed to assess plant water status are

time consuming, spatially restrictive, and costly (Ripple, 1986). Currently, visual symptoms, ET

rates, and tensiometers are all methods for determining irrigation needs (Turgeon, 2008).

However, when visual symptoms of drought are apparent, it is already too late because the turf is

already stressed. Likewise, ET values are generally not site specific and actual ET can differ










greatly from one location to another. It is assumed that all the turf within a given area is

precisely the same and responds exactly the same to environmental conditions. Conversely,

tensiometers have proven to be an adequate assessment of plant water status, but require periodic

maintenance, and may have to be removed during periods of cold weather (Unruh and Elliott,

1999). In addition, the tensiometer reading is only appropriate in the area adjacent to the

placement of the ceramic tips and do not indicate water status for a large area (Unruh and Elliott,

1999). Therefore, because conservation of this limited resource is critical, it is increasingly

necessary to develop technology to help turf managers allocate irrigation water more efficiently.

Nutrient Importance

Aside from the importance of water to plants, available nutrients also drive growth and

development. Nutrient availability affects many physiological processes the plant must perform

in order to develop properly. Nitrogen (N) is the most limiting nutrient in production of non-

legumous crops (Osborne et al., 2002). It is essential to the building of several important plant

components including proteins, nucleic acids, hormones, and chlorophyll, which harvests light

energy used in photosynthesis (Hopkins, 1999). Phosphorus (P) plays an important role in the

transfer of energy within the plant in the form of compounds like adenosine triphosphate (ATP),

adenosine diphosphate (ADP), and phosphate (Pi). P also is an important factor in root growth

(Hopkins, 1999). Potassium (K) is also a very important nutrient in plant growth and is required

in large amounts. Potassium serves to activate a number of enzymes and regulates stomatal

conductance; a very important process controlling transpiration (Hopkins, 1999). All of these

nutrients, if in deficit of plant requirement, can cause a number of physiological disruptions

which are detrimental to plant growth and development. Once these stresses become visible it

may be too late to correct the damage. Therefore, methods are needed to predict plant nutrient

status that can detect physiological stress before it becomes visible to the naked eye.









Plant Growth Regulator Importance

Plant Growth Regulators (PGRs) have become a very important tool in turfgrass

management. Many turfgrass managers use PGRs to suppress vertical turf growth which

ultimately reduces maintenance costs and improves turf quality. Some data suggests that PGR-

treated turf can produce greater root mass, recover from injury faster, reduce water use rate,

reduce disease incidence, and reduce Poa annua weed populations (Branham, 1997).

A PGR is a substance that adjusts the growth and development of a plant. This is generally

achieved through the inhibition of gibberellic acid (GA), a hormone which is responsible for cell

elongation. Plant growth regulators were first introduced for use on Eine turf in 1987

(VanBibber, 2006). The first of these chemicals were flurprimidol (Trade name Cutless) and

paclobutrazol (Trade name Trimmit), which inhibit GA synthesis at relatively the same point in

the GA biosynthesis pathway, resulting in very similar plant responses. However, paclobutrazol

is the more active compound and uses lower rates to achieve the same response as higher rates of

flurprimidol (Branham, 1997). The newest of the PGRs, trinexapac-ethyl (Trade name Primo),

was released commercially in 1995 and has become the standard for use throughout the industry

(Branham, 1997). It facilitates GA inhibition later in the GA biosynthesis pathway so that GA's

are formed but are not active and serves as precursors for plant biochemical processes. Plant

growth regulators are relatively expensive compared to other chemicals used by turf managers.

More efficient applications of PGRs could potentially reduce the application cost of these

chemicals.

Remote Sensing

Remote sensing can be defined as obtaining information about an obj ect, area, or

phenomenon by analyzing data acquired by a device that is not in contact with that obj ect, area,

or phenomenon (Lillesand and Keifer, 2000). For many years researchers have entertained the









idea of measuring plant stress using remote sensing technology. Plant light interception

influences growth and physiological responses (Salisbury and Ross, 1992). When light is

intercepted by the plant, it is absorbed, transmitted, or reflected (Turgeon, 2008). Therefore, by

quantifying the reflected light, remote sensing technology can help to detect the onset of

turfgrass stress (Ikemura and Leinauer, 2006).

Presently, real world applications of remote sensing technology in turfgrass management

are still in their infancy (Ikemura and Leinauer, 2006). Studies have shown that image analysis

and various remote sensing devices have strong potential to detect a variety of turfgrass stresses

(Ikemura and Leinauer, 2006). Spectral radiometry (Trenholm et al., 2000), and infrared

thermometry (Slack et al., 1981) have been shown to adequately assess light reflectance at

various wavelengths. By differentiating certain wavelength characteristics, insights into growth

and adaptive characteristics and how a plant responds to stress can be seen (Trenholm et al.,

2000). There have been several aspects of remote sensing technology explored to reveal the level

of stress that may be occurring in a plant. Researchers have created a variety of theories utilizing

remote sensing techniques and exploring many parts of the electromagnetic spectrum to describe

various plant stresses.

Vegetative Indices

Researchers have found certain relationships exist between various aspects of the

electromagnetic spectrum and a variety of plant parameters and stress (Bell et al., 2000, 2002a,

2002b, 2004; Trenholm et al., 1999a, 1999b, 2000; Kruse et al., 2005; Fenstemaker-Shaulis et

al., 1997; Bahrum et al., 2003; Xiong et al., 2007). These relationships are formulated into

indices to normalize data for varying illumination conditions (Lillesand and Kiefer, 2003). One

of the most common indices is the Normalized Difference Vegetation Index (NDVI) computed









as: reflectance in the Near Infrared (NIR) region minus reflectance in the Red (R) region divided

by a sum of both (NIR-R / NIR+R) (Rouse et al., 1973) .

Human Visual Ratings

Traditionally, visual observation has been the standard for assessing turf stress for turf

managers. Researchers quantified this observation by creating a scale. The National Turfgrass

Evaluation Program (NTEP) scale is a 1-9 assessment of turf color, quality and density, where 6

is the least acceptable and has been the standard for turf research for many years (Morris, 2007).

However, recent investigations have looked beyond this scale to compare remote sensing

systems with traditional human evaluation. Trenholm et al. (1999) tested four of the current

indices and compared them to visual evaluations. These indices were:

* NDVI computed as (935-661 nm)/ (935+661 nm)
* Leaf area index (LAI) computed as 93 5/661 nm
* Stress 1 computed as 706/760 nm
* Stress 2 computed as 706/813 nm

They concluded that NDVI, LAI, and Stress 2 indices, and individual wavelength reflectance

measurements at 661 nm and 813 nm were comparable to visual evaluations (Trenholm et al.,

1999). Bell et al. (2002b) found strong correlations between NDVI and turf color, and

subsequent work by Bell et al. (2004) concluded that NDVI was a better estimator of chlorophyll

content than visual color evaluation.

Irrigation

Early research involving remote sensing to predict plant water status related plant canopy

temperature to plant water potential (Ehrler et al., 1978). This was derived from the fact that

transpiration plays a maj or role in the plant' s ability to cool itself through ET. As soil water

decreases, evaporative cooling also decreases causing the canopy temperature to rise. This is

what many researchers have used to relate canopy temperature to relative soil water content










(Aston and Van-Bavel, 1972). Stomatal closure of sunlit leaves results in increased leaf

temperature if factors such as wind speed and vapor pressure remain constant (Ehrler et al.,

1978). This temperature difference between canopy and air can be determined in several ways.

Originally, microthermocouples were used, but this only worked on a small scale. However,

advances in infrared thermometry have increased the ability to detect crop water stress through

the utilization of remote sensing devices (Slack et al., 1981). Infrared thermometers registered in

a narrow bandpass (8-14 Clm or 10.5 12.5 Clm) are capable of measuring crop temperatures with

0.5 oC accuracy (Ehrler et al., 1978). They can be used for measuring canopy temperature at

ground level, from aircraft, or satellite. These methods are done in a much timelier manner

compared to old methods such as tensiometers and microthermocouples (Ehrler et al., 1978).

From this relation of canopy temperature to plant water potential, many indices have been

developed to best predict plant water status. Idso et al. (1981) proposed a temperature-based

index as a water stress indicator. This was intended to normalize canopy temperature minus air

temperature, To-Ta, for environmental variability. It was introduced as the Crop Water Stress

Index (CWSI). CWSI produces a species-specific linear relationship between air and canopy

temperature, which is irrespective of other environmental factors except cloud cover (Slack et al.,

1981). CWSI can be described as the "fractional decrease of potential ET" which can be

presented in terms of the canopy temperature minus the air temperature (To-Ta) or

CWSI= (Toc-T~a)_S c-Ta)1


Where Te= canopy temperature, Ta= actual temperature, a= actual, u= upper limit, and 1= lower

limit (Jalali-Farahani et al., 1994)

The CWSI was found to have potential in turf irrigation scheduling (Throssel et al., 1987).

Further investigations of this model spawned numerous theories accounting for several









environmental factors as well. Horst et al. (1989) suggested that canopy temperature minus air

temperature plotted as a function of Vapor Pressure Deficit (VPD), which is the difference in the

amount of moisture in the air and the amount of moisture the air can hold. This may be species

and mowing height dependent and may require further adjustment of the function for each

species under individual environmental regimes (Horst et al., 1989). Jalali-Frahani et al. (1993)

proposed that CWSI of turfgrass was not only a function of VPD but also net radiation from

sunlight. Martin et al. (1994) found that using variables accounting for VPD, wind speed, and

net radiation increased coefficient of determination (r2) ValUeS dramatically across several

varieties of turfgrass. Ehrler et al. (1978) concluded that temperature difference (T,-Ta) in wheat

(Triticum spp., durham 'Produra') is a reliable method for monitoring plant stress from aircraft or

satellite. Frequent measurements were necessary to reduce the variability in the data. Throssel

et al. (1987) found that temperature difference (Te-Ta) appeared to be indicative of water stress

for well-watered Kentucky bluegrass (Poa pratensis L.).

More recent investigations into plant and soil water status rely on spectral reflectance data

measured from plant or turfgrass canopies. Fenstemaker-Shaulis et al. (1997) found a negative

correlation with NDVI and canopy temperature (r2=0.74) and a positive correlation to plant

moisture content (r2=0.90). In addition, they concluded, that mapping NDVI values can also

provide valuable insight into the status of irrigation system uniformity and management

practices. Hutto et al. (2006) found that drought stressed species had a higher amount of

reflectance in the short wave infrared region (SWIR) (1,300-2,500 nm) than that of the non-

stressed control. Ripple (1986) found that reflectance in the 630-690 nm and 2,080-2,350 nm

ranges is strongly correlated with leaf water content. As reflectance decreased in these ranges

there was an inverse linear relationship with leaf relative water content. However, Ripple (1986)









concluded that relationships between spectral reflectance and leaf water content can be direct or

indirect and are wavelength dependent. It is believed that elevated correlations in the 630-690

nm range with spectral reflectance data was due to a reduction in chlorophyll as the leaves dried

and not a true relation to actual plant water status (Ripple, 1986).

Nutrient

To a large extent, much remote sensing work has been done on the assessment of N and

chlorophyll levels. Overall, reflectance is low in the visible range of the electromagnetic

spectrum due to the high absorption by chlorophyll for use in photosynthetic processes

(Salisbury and Ross, 1992 ). For this reason, if there is a particular reflectance in this region this

could correlate to a certain physiological disruption in the plant. It is generally assumed that

there is a close relationship between plant color and chlorophyll content (Bell et al., 2004).

Research has shown reflectance at 550 nm to be the most sensitive assessment of N in many

agronomic crops, which is what the human eyes see as green (Blackmer et al., 1994, 1996, Yoder

and Pettigree-Crosby, 1995). Bell et al. (2002a) found vehicle mounted optical sensors provided

a good assessment of variable N applications. Xiong et al. (2007) found significant seasonal

responses to N treatments.

For other nutrient deficiencies such as P and K, limited research has been conducted to

establish their function in the spectral constituency. Research that has been conducted suggests

that reflectance wavelengths for detecting P concentration in corn (Zea mays L.) are in the NIR

(730 nm and 930 nm) region of the spectrum as well as reflectance from the blue region (440 nm

and 445 nm) (Osborne et al., 2002). This can possibly be attributed to the production of

anthocyanin which is produced in P deficient plants that absorb more green light and reflect

more blue light (Osborne et al., 2002). Research conducted by Kruse et al. (2005) found that

spectral reflectance data can be used as a good estimation of P concentration in creeping










bentgrass (Agrostris stolinifera L.). Reflectance wavelengths for predicting P concentration of

plant tissue were in the blue (480 nm), yellow (565 nm), orange (595 nm), and red (650 nm)

regions with an overall r2 Value Of 0.73. This reflectance response relates to the spectral

characteristic of anthocyanin which increased in concentration under P stress conditions (Kruse

et al., 2005). Hutto et al. (2006) found that potassium deficiency could be detected at

bandwidths from 750 nm to 785 nm.

Biomass

Green plant material has high reflectance in the NIR (700-1100 nm) region, while dead or

dying plant material is just the opposite (Geeske et al., 1997). Yoder and Pettigree-Crosby

(1995) related this correlation in the NIR region to biomass, or energy status of the plant.

Similar results were found by Kruse et al. (2005) on creeping bentgrass, where reflectance in the

green (5 10 nm and 53 5 nm), red (63 5 nm), and NIR (73 5 nm) were significant in predicting

biomass. Starks et al. (2006) found crude protein concentration, biomass production, and crude

protein availability of bermudagrass (Cynodon dactylon L. Pers.) grown in pastures closely

correlated with canopy reflectance ratios of 605/515 nm, 915/975 nm, and 865/725 nm.

Conversely, Fenstemaker-Shaulis et al. (1997) found that NDVI is not a strong predictor of

biomass (r2=0.37). Kruse et al. (2005) also observed no correlation between biomass and NDVI.

Kruse et al. (2006) found implementing Partial Least Squares regression on hyperspectral data to

be an adequate predictor of biomass in creeping bentgrass compared to indices NDVI, LAI,

Stress, and Stress2.

PGRs

Not a large body of literature exists in the area of remote sensing and the effects of PGR

applications. Some image-based applications have been tested in cotton (Gosspypium spp.) to

reduce PGR use for economic reasons (Bethel et al., 2003). Among comparisons of different










image based-application methods (Variable Rate and Site-Specific) to traditional blanket

applications for PGR' s in cotton, it was found that the broadcast applications were on average 19

% more costly than the site-specific method, and 27 % more costly than the variable rate method

(Bethel et al., 2003). These results demonstrate the potential economic benefit of this technology

for future use in other aspects of agriculture to reduce excess chemical runoff and leaching into

the surrounding environment (Bethel et al., 2003).

Statistical Methods

Older work involving remote sensing typically used multispectral data, which is collected

in individual or wide bandwidths and give a coarse representation of the electromagnetic

spectrum (Lillesand and Kiefer, 2003). With the introduction of hyperspectral data, which

typically acquires 200 or more bands in very narrow resolution throughout the visible, NIR, and

SWIR regions, researchers have more data to analyze (Lillesand and Keifer, 1987). These

enhanced techniques for obtaining the data have created the need to explore different statistical

methods to quantify and interpret the data. Traditionally linear regression is the simplest way to

understand variance. However, in interpreting reflectance data it has been found that other

regression techniques better assess plant status (Kruse et al., 2005).

Partial Least Squares (PLS) regression is a method used to predict variables when there are

a large number of factors; typically more than the number of observations (Tobias, 1997). The

use of multiple variables is generally tested through multiple linear regression (MLR) (Tobias,

1997). However, the problem with MLR is when the number of factors becomes too large the

model may become over-fitting (Tobias, 1997). This is observed when a large number of

variables are present but only a few account for most of the variation (Tobias, 1997). Almost

perfect models may be achieved; however they will not be highly predictive of new data.









When prediction is the goal, then PLS regression is a highly useful tool (Tobias, 1997).

Kruse et al. (2006) found PLS regression yielded a stronger relationship to predict N

concentration compared to NDVI in plant leaf tissue indicating its potential use in future

development models. Lee et al. (1999) also found that PLS and Principal Component Analysis

(PCA) prediction models performed similar to each other and both were better than MLR

regression to predict the N status in corn.

Summary

The use of remote sensing technology to predict turfgrass stress could decrease the time

and labor required to evaluate turfgrass health (Osborne et al., 2002). This would not only reduce

the environmental impact of turfgrass management practices, but could reduce maintenance costs

of the various inputs required by turfgrass managers. Current dependence on agricultural

chemicals have heightened many environmental concerns in Florida due to the sandy soils and

heavy rainfall which increases potential for runoff and leaching of chemicals if applied in excess

(Min and Lee, 2003). As stated earlier, real world applications of remote sensing technology for

utilization in turfgrass are still in their infancy. At the present time, an experienced turf

manager' s eye is still better than current computerized systems (Ikemura and Leinauer, 2006).

However, studies have shown that various remote sensing devices have strong potential to detect

a variety of turfgrass stresses (Ikemura and Leinauer, 2006).

Applying remote sensing technology to the field of turfgrass management could possibly

help reduce irrigation and provide more efficient watering and fertilization routines (Ikemura and

Leinauer, 2006; Bell et al., 2004). Combined with GPS, GIS, and variable rate application

technology, there is a potential to apply nutrients as well as irrigation based on plant need rather

than broadcast applications of entire areas (Bell et al., 2004). Currently, most of the remote

sensing research is conducted on agronomic food crops and forested areas. Only limited









research is available in the area of turfgrass management (Trenholm et al., 1999). Further

research is needed to incorporate remote sensing into turfgrass management (Trenholm et al.,

1999). If conditions across the entire golf course were uniform in microclimates, soil types, turf

varieties, pest densities, nutrient status, and irrigation performance there would be little need for

advanced sensor systems to measure the variability of these parameters to better meet the needs

on a locational basis (Stowell and Gelerntner, 2006). However, that is not the case.

Optimistically, Bell et al. (2002a) concludes "If an optical sensing (In reference to remote

sensing) system and software can be economically produced, reasonably priced, and mounted

effectively on normal maintenance equipment, a turf practitioner could save enough money in

fertilizers and pesticides to pay for the equipment. This approach would be useful for reducing

the amount of fertilizers and pesticides needed to adequately manage large turf areas. The use of

optical sensing to determine fertilizer rate before or during application could increase turf

uniformity and possibly, turf health. Sensor maps of large turf areas could be used to signal turf

decline and provide an early warning system for turf mangers."









CHAPTER 2
DETECTION OF LEAF NITROGEN CONCENTRATION AND SOIL VOLUMETRIC
WATER CONTENT USING GROUND-BASED REMOTE SENSING TECHNOLOGY

Introduction

Remote sensing is defined as obtaining information about an obj ect, area, or phenomenon

by analyzing data acquired by a device that is not in contact with that obj ect, area, or

phenomenon (Lillesand and Keifer, 1987). For many years researchers have entertained the idea

of measuring plant stress using remote sensing technology. Plant light interception significantly

influences growth and physiological responses (Salisbury and Ross, 1992). When light is

intercepted by the plant it is absorbed, transmitted, or reflected (Salisbury and Ross, 1992).

Using remote sensing technology to quantify the light that is reflected could help to detect the

onset of turfgrass stress (Ikemura and Leinauer, 2006; Kruse et al., 2005; Hutto et al., 2006;

Trenholm et al., 1999a). Real world applications of remote sensing technology for use in

turfgrass management are still in their infancy; however, studies have shown that image analysis

and various remote sensing devices have strong potential to detect a variety of turfgrass stresses

(Ikemura and Leinauer, 2006).

Recent investigations into remote sensing instrumentation have produced promising results

for turf managers. However, the majority of turf managers and researchers still use outdated

time and labor intensive techniques to manage and assess turf. Nonetheless, there are new

developments in ways of assessing and mapping plant stress that may increase the efficiency by

which we manage and assess turf. The use of remote sensing technology to estimate turfgrass

stress could significantly decrease the time and labor required to assess these levels in traditional

ways, thus reducing cost as well (Osborne et al., 2002).

The many great achievements in U.S. agricultural productivity in the past decades can be

attributed to the use of agricultural chemicals, including fertilizer and pesticides (Lee et al.,










1999). However, the increased dependence on pesticides and fertilizers have heightened many

environmental concerns especially in Florida due to the sandy soils and heavy rainfall, which

increases potential for heavy runoff and leaching of chemicals (Min and Lee, 2003). With

increasing water and other environmental restrictions, turf managers and researchers must be

aware of technological advances in their industry.

Recent concerns have pressed turf managers to reduce nutrient and pesticide inputs used

for turf maintenance (Bell et al., 2002a). In other industries, such as agronomic food production

crops, precision agriculture has been used as a management tool to maximize yield and minimize

cost (Bethel et al., 2003). The basis of this technology relies heavily on the Global Positioning

System (GPS), Geographic Information Systems (GIS), and Variable-Rate Application (VRA)

and remote sensing technology. Using this technology for site-specific management of nutrients,

water, and pesticides could optimize yield while minimizing cost (Lee et al., 1999).

Water is an important resource in turfgrass management; however, it is not a renewable

resource (Barret, 2003). The Earth's water reserves are 97 % salt water and 3 % fresh water. Of

the fresh water reserve, two-thirds is frozen in polar ice caps (Barret, 2003). Therefore, only 1 %

of the Earth' s water is available to meet the water needs of manufacturing, mining, agricultural,

and turf not met by rainfall (Barret, 2003). In addition, the average rainfall in Florida does not

meet the watering requirements to maintain golf course quality at the level expected by players

and industry viewers (Unruh and Elliot, 1999). As a result, the irrigation system is the single

most important tool available to a golf course superintendent or turf manager (Barret, 2003). As

water restrictions increase and pressure rises from environmental groups to restrict water use on

golf courses, it is increasingly imperative that turfgrass managers improve water use efficiency

(Martin et al., 1994).









Aside from the importance of water to plants, available nutrients also drive plant growth

and development. Nutrient availability affects many of the physiological processes a plant must

perform in order to develop properly. Nitrogen (N) is the most limiting nutrient in production of

non-legumous crops (Osborne et al., 2002). It is essential to the building of several important

plant structures including, proteins, nucleic acids, hormones, and most importantly chlorophyll,

which harvests light energy used in photosynthesis (Hopkins, 1999).

Remote sensing technology has been shown to assess stress that may be occurring in a

plant. Researchers have created a variety of theories and indices exploring many parts of the light

spectrum to describe various stresses. These stresses include water, nutrient, soil compaction,

pests, and disease.

One of the early theories involving remote sensing to predict plant water status was the

Crop Water Stress Index (CWSI), which correlated plant water potential with canopy

temperature (Ehrler et al., 1978). More recent investigations rely on spectral reflectance data

measured from plant or turfgrass canopies. Several spectral reflectance bands have been

identified for use in modeling various turfgrass stresses, especially those in the visible (400-700

nm) and near infrared (NIR) regions (700-1,300 nm) of the spectrum, as well as short wave

infrared (SWIR) regions (1,300-2,500 nm) (Hutto et al, 2006). Reflectance from healthy plant

canopies is characteristically low in the visible range of the spectrum due to high absorption by

chlorophyll for use in photosynthetic processes (Trenholm et al., 2000), If there is a particularly

high reflectance in this region, this could correlate to a physiological disruption in the plant

(Geeske et al., 1997).

Green plant material has high reflectance in the NIR range, while dead or dying plant

material has low reflectance (Geeske et al., 1997). Yoder and Pettigree-Crosby (1995) relate this









correlation in the NIR region to biomass, or the energy status of the plant. Similar results were

found by Kruse et al. (2005) on creeping bentgrass (Agrostis stolinifera L.), where reflectance in

the green (5 10 nm and 53 5 nm), red (63 5 nm), and NIR (73 5 nm) were significant in predicting

biomass. Trenholm et al. (1999a) found that wear stress in several varieties of hybrid

bermudagrass (Cynodon dactylon x C. transvalensis Burtt-Davy) and seashore paspalum

(Paspalum vaginatum Swartz) is associated with the specific wavelength at 813 nm. Starke et

al., (2006) found that crude protein concentration, biomass production, and crude protein

availability of bermudagrass (C. dactylon L. Pers.) pastures is closely related with canopy

reflectance ratios of 605/515 nm, 915/975 nm, and 865/725 nm.

Researchers have been exploring the idea that plant water status can be determined by

measuring reflectance in the short wave infrared (SWIR) region (1,300-2,500 nm). Ripple (1986)

found that SWIR reflectance is strongly correlated with leaf water content and that significant

correlations exist between certain spectral bands (630-690 nm and 2,080-2,350 nm ranges), with

SWIR reflectance having the highest correlations. As reflectance decreased in the 630-690 nm

range there was an inverse linear relationship with leaf relative water content. However, Ripple

(1986) concluded that relationships between spectral reflectance and leaf water content can be

direct or indirect and are wavelength dependent. It is believed that elevated correlations with

spectral reflectance data was due to a reduction in chlorophyll as the leaves dried and not a true

relation to actual plant water status (Ripple, 1986). Hutto et al. (2006) also documented that

drought stressed creeping bentgrass had a higher amount of reflectance in the SWIR regions than

that of the non-stressed control.

Several different spectral responses to light to assess indications of plant stress, nutrient

and water status have been tested in the form of indices. One of the most common indices is the









Normalized Difference Vegetation Index (NDVI) which is computed as the reflectance in the

NIR region minus the reflectance in the red region divided by a sum of both (NIR-R)/ (NIR+R).

Bahrum et al. (2003) found that NDVI values from remotely sensed data can be used to assess

turfgrass stress. Fenstemaker-Shaulis et al. (1997) found a negative correlation between NDVI

and canopy temperature (r2=0.74), and a positive correlation between NDVI and plant moisture

content (r=0.90). However, NDVI was not as strong a predictor of biomass (r2=0.37). In

addition, spatial mapping of NDVI values can provide valuable insight into the status of

irrigation system uniformity and management practices (Fenstemaker-Shaulis et al., 1997). Bell

et al. (2002b) found strong correlations between NDVI and turfgrass color ratings. Additionally,

the use of NDVI ratings obtained from vehicle mounted optical sensors were effective for

measuring herbicide damage (Bell et al., 2000), variable N applications (Bell et al., 2002a), and

significant seasonal responses to N treatments (Xiong et al., 2007). It has also been reported that

NDVI is a better estimator of chlorophyll content than visual color evaluation (Bell et al., 2004).

Trenholm et al. (1999) tested four indices to detect stress in turfgrass. These indices were:

* NDVI computed as (935-661 nm)/ (935+661 nm)
* Leaf Area Index (LAI) computed as 93 5/661 nm
* Stress 1 computed as 706/760 nm
* Stress 2 computed as 706/813 nm

They concluded NDVI, IR/R, and Stress 2 indices, as well as individual wavelength

measurements at 661 nm and 813 nm were well correlated to visual ratings.

Older work involving remote sensing typically used multispectral data, which is collected

in individual or wide bandwidths and gave a coarse representation of the electromagnetic

spectrum (Lillesand and Kiefer, 2003). With the introduction of hyperspectral data, which

typically acquires 200 or more wavelength bands in very narrow resolution throughout the

visible, NIR, and MIR regions, researchers have more data to analyze (Lillesand and Keifer,









1987). These enhanced techniques for obtaining the data have created the need to explore

different statistical methods to quantify and interpret the large amount of data produced.

Traditionally, linear regression is the simplest way to understand variance. However in

interpreting reflectance data it has been found that other regression techniques better assess plant

status (Kruse et al., 2005).

Partial Least Squares (PLS) regression is a method used to predict variables when there are

a large number of factors; typically more than the number of observations (Tobias, 1997). The

use of multiple variables is generally tested through Multiple Linear Regression (MLR) (Tobias,

1997). However, the limitation with MLR is when the number of factors becomes too large the

model may become over-fitting (Tobias, 1997). This is observed when a large number of

variables are present but only a few account for most of the variation (Tobias, 1997). Almost

perfect models may be achieved; however they will not be highly predictive of new data.

When prediction is the obj ective then PLS regression is a highly useful tool (Tobias,

1997). Kruse et al. (2006) found PLS regression yielded a stronger relationship to predict N

concentration compared to NDVI in creeping bentgrass indicating its potential use in future

development models. Lee et al. (1999) also found that PLS and Principal Component Analysis

(PCA) prediction models performed similar to each other and both were better than MLR

regression to predict the N status in corn (Zea mays).

Currently, the maj ority of remote sensing research is conducted on agronomic food crops

and forested areas. Only limited research is available in the area of remote sensing and turfgrass

management (Trenholm et al., 1999b). If conditions across the entire golf course were uniform

in microclimates, soil types, turf varieties, pest densities, nutrient status, and irrigation

performance there would be little need for advanced sensor systems to measure the variability of









these parameters to better meet the needs on a locational basis (Stowell and Gelernter, 2006).

However highly variable agronomic conditions create the need for further research to incorporate

the use of remote sensing into turfgrass management (Trenholm et al., 1999b). The obj ective of

this research was to assess the ability of three remote sensing instruments to detect soil moisture

and leaf nitrogen levels of Tifsport bermudagrass turf.

Materials and Methods

The field experiment was conducted at the University of Florida, West Florida Research

and Education Center near Jay, FL. Plots were arranged in a randomized complete block design

and treatments were arranged in a split plot factorial design with four N levels (sub plot) split

across four irrigation regimes (whole plot) with three replications per treatment (Appendix A).

Whole plots were 4.88m x 4.88m in size and subplots within the whole plots were 1.22m x

4.88m in size. Whole plots were spaced 4.88m apart on each side to minimize the impact of drift

from irrigation from plot to plot. Ammonium sulfate was applied on 14 June 2007 and 20 July

2007 at 0, 25, 50, and 100 kg hal with a Gandy drop spreader and then watered in with 0.5 cm of

supplemental irrigation. Irrigation was applied at 60%, 80%, 100%, and 120% of weekly

estimated ET values obtained from an onsite weather station that is part of the Florida

Agricultural Weather Network (http:.//fawn.ifas.u-fl .edu/) when ET was in deficit of 1.27cm or

more. This occurred only on 24 August 2007 (Appendix B-D). Plots were mowed 2 to 3 times

per week at a height of 1.2 cm with a Toro Reelmaster 3100-D. Data were collected five times

(22 June, 12 July, 25 July, 22 Aug., and 4 Sept.) throughout the growing season in 2007. All

measurements were taken within a one hour time period and collected between 1100-1300 hrs.

Plots were rated visually for color, quality, and density based on the standard 1-9 National

Turfgrass Evaluation Program (NTEP) rating scale where 9 is the highest possible rating, 6 is the

lowest acceptable rating, and 1 indicates dead turf. Volumetric water content (VWC) readings









were taken via Time Domain Reflectometry (TDR) with a Field Scout TDR 300 Soil Moisture

meter (Spectrum Technologies, Inc., Plainfield, IL). Clipping tissue samples were immediately

weighed after collection and oven dried at 52 oC for 7d before being weighed again for

determination of biomass production. The oven dried tissue was ground and then analyzed for

Total Kj eldahl N content using the RFA Method, number A303-SO75-01 (Alpkem Corporation,

1990) as described by Mylavarapu and Kennelly (2007).

Reflectance Measurements

Spectral measurements were taken from three remote sensing devices. An infrared

thermometer (Raytek Raynger@ STTw, TTI, Inc.) was used to determine canopy temperature of

the turf and to develop a CWSI calculated as: CWSI= (To-Ta) a- (To-Ta)1/ (To-Ta)u (To-Ta)1,

Where Te= Canopy Temperature, Ta= Actual Temperature, a= actual, u= upper limit, and 1=

lower limit (Jalali-Farahani et al., 1994).

A Crop Circle (Model ACS-210) (Holland Scientific) was fitted onto a Toro Greensmaster

1000 walking greens mower 0.81m above the turf canopy which provided a 0.45 m2 field of

view. The Crop Circle sensor produces its own light source at 650 nm and 880 nm and measures

reflected values to produce an NDVI which is calculated as (880-650 nm)/ (880+650 nm). No

calibration method was used for the sensor since it has the ability to detects it own light source

from incoming ambient sunlight.

A hand-held hyperspectral spectroradiometer (Field Spec Pro; Analytical Spectral Devices,

Inc., Boulder, CO) was also used to collect reflectance data. This radiometer has a spectral range

of 300-2,500 nm with a 1 nm resolution and a 23 o foreoptic. Reflectance readings were collected

from shoulder height on each plot which provided a 0.65 m2 field of view. Radiance values are

expressed in percent reflectance compared to standardization with a white reference value across

the entire spectral range. A white reference was used for calibration purposes at the beginning of









data collection with the ASD unit, and again every 15 minutes depending on the length of time

needed to collect data. Usually only one white reference was needed to take data on all plots.

All reflectance readings were taken in full sunlight between the hours of 1100 and 1400 Central

Standard Time (CST) to minimize variance caused by incoming solar radiation.

Several bands were extracted from ASD reflectance data for comparison based on previous

research. These bands consisted of wavelengths 550 nm (Blackmer et al., 1994; Blackmer et al.,

1996; Yoder and Pettigree-Crosby, 1995); 510 nm, 535 nm, 635 nm, and 735 nm (Kruse et al.,

2005); 545 nm, 755 nm and 935 nm (Starks et al., 2006); and 661 nm and 813 nm (Trenholm et

al., 1999a). Spectral ranges 630-690 nm and 2080-2350 nm were averaged (Ripple 1986) and

ratios 605/515 nm, 915/975 nm, and 865/725 nm were calculated (Starks et al., 2006). Also

NDVI ((880-650 nm)/ (880+650 nm)) and LAI (880/650 nm) indices were calculated for

comparison against Crop Circle NDVI and LAI calculations, as well as Stress1 (706/760 nm)

and Stress2 (706/813 nm) indices (Trenholm et al., 2000).

Statistical Analysis

Analysis of variance (ANOVA) was performed using the PROC GLM method (SAS

Inst., 2003) to compare the differences among N and irrigation treatments at various dates to

visual ratings, Crop Circle readings, IR temperature readings calculated into a CWSI, various

individual reflectance wavelengths quantified from the ASD unit, soil VWC, and leaf N

concentration. Regression analysis was performed using the PROC REG method (SAS Inst.,

2003) to correlate relationships between visual ratings, Crop Circle readings, and IR temperature

readings, various individual reflectance wavelengths quantified from the ASD unit, soil VWC,

and leaf nitrogen concentration. Additionally, due to the large amount of data generated by the

ASD device, ASD data were rendered using PROC PLS (SAS Inst., 2003) for prediction of soil

VWC, leaf N concentration, visual quality ratings, and biomass production. Equations were









validated through a single sample cross-validation. For cross-validation, ten percent of the

sample was omitted for prediction purposes so that the number of factors chosen creates the

minimal predicted residual sum of squares (PRESS). This process was repeated so that every

observation was used exactly once for cross-validation.

Results and Discussion

Influence of Irrigation Treatments

ANOVA (Tables 2-1, 2-2, 2-3) indicated no irrigation treatment effects were significant.

This is likely due to rainfall received during the experiment which eliminated differential soil

VWC. As stated previously in the materials and methods, irrigation treatments were only

applied once during the experiment due to the fact that the weekly cumulative ET calculations

(Appendix C) based on daily rainfall and ET values (Appendix B) were only in deficit of more

than 1.27cm once during the experiment. In review of the average rainfall and ET data from the

last 3 years, the experiment should be run earlier in the year when cumulative ET minus rainfall

differences are not as great (Appendix D).

Nitrogen Effects

ANOVA indicated a large source of variation in bermudagrass response to N treatments.

Leaf tissue N concentration, color, quality, NDVI and LAI from both optical sensing devices, as

well as many individual reflectance wavelengths and reflectance ratios indicated a significant

response to N treatments (Tables 2-1, 2-2, 2-3). This can be attributed to the ability of N to

increase overall turf health since it is the single most important macronutrient to the plant

(Hopkins, 1999).

NDVI and LAI from the Crop Circle device indicated a sampling date effect (Table 2-1).

Additionally, reflectance at wavelength 813 nm, LAI, and 915/975 nm and 865/725 nm ratios

quantified from the ASD device also indicated a sampling date effect (Tables 2-1, 2-2, 2-3).









Research conducted by Xiong et al. (2007) showed seasonal NDVI and LAI effects in response

to N treatments.

There was a significant improvement in visual color between no N applied and N applied

at all levels across all dates (Table 2-4). Visual Quality rating (Table 2-4) and 650 nm

reflectance from the Crop Circle device were both greater at the highest N rate (100 kg ha l)

compared to no N applied (Table 2-5). Similarly, reflectance wavelengths (510 nm, 535 nm, 545

nm, 550 nm, and range average 630-690 nm) (Table 2-6), ratios (605/515 nm, 915/975 nm, and

865/725 nm) (Table 2-7), and indices (NDVI, Stress1, and Stress2) derived from ASD data

detected the high rates of applied N (Table 2-7).

Nitrogen X Date Interaction

It is apparent from the N X Date interaction means that there is a significant sequential

response to increasing N treatments on percent leafN concentration the first three sampling dates

(22 June, 12 July, and 25 July 2007) (Tables 2-8, 2-9, 2-10, and 2-11). NDVI and LAI indices

calculated from the Crop Circle device (Table 2-8) also show this response as well as LAI (Table

2-8), 813 nm and 935 nm (Table 2-10), and 915/975 nm and 865/925 nm ratios (Table 2-11)

calculated from the ASD reflectance data. This response in reflectance data can be attributed to

the ability of the devices to detect overall increases in turf health due to a sequential vigor

response in reaction to varying N treatments. This agrees with the work of Bell et al. (2002a,

2004).

The last two sampling dates (22 Aug. and 4 Sept. 2007) do not show a response to

increasing N treatments for any parameter measured that was significant by date (Tables 2-8, 2-

9, 2-10, and 2-11). This is likely due to a waning N response, since N had not been applied for

33 and 46 days respectively for the final two sampling dates.









Tissue N Concentration vs. Reflectance

Reflectance values in this experiment did not prove to be a good predictor of leaf tissue N

concentration in bermudagrass using linear or PLS regression to model percent N verses

reflectance measurements However, some weak correlations exist (Tables 2-12, 2-13, 2-14, 2-

15). This could potentially be due to the experimental design where N treatments were only

applied twice, one month apart, causing the waning effect mentioned in the N X Date interaction

section. Comparatively, research conducted by Kruse et al. (2005) and Xiong et al. (2007) where

N treatments were applied every two weeks promoted a more consistent response to treatments

across the growing season.

VWC vs. Reflectance

The highest r2 ValUeS for estimation of VWC are derived from PLS regression (r2= 0.72)

calculated from reflectance data from the ASD device (Table 2-15) (Figure 2-1), and NDVI ( 2=

0.71) derived from reflectance data from the Crop Circle device (Table 2-12) (Figure 2-2). Also,

LAI (r2= 0.59), NIR reflectance (r2= 0.61), and RED reflectance (r2= 0.52) derived from Crop

Circle values provide adequate correlation as well (Table 2-12). This corresponds to vegetative

indice values NDVI (r2= 0.57), LAI (r2= 0.55), Stress1 (r2= 0.61), Stress2 (r2= 0.64), and ratios

605/515 nm (r2= 0.66), 915/975 nm (r2= 0.68), 865/725 nm (r2= 0.44) derived from the ASD unit

(Table 2-14). Some correlations at individual bands 755 nm (r2= 0.53), 813 (r2= 0.53), and 935

nm (r2= 0.49) were also notable (Table 2-13).

The high r2 ValUeS for estimation of VWC with remotely sensed data from Crop Circle and

various individual bands, indices, and ratios from the ASD device could be attributed to the

extremely low VWC values recorded on 22 Aug 2007. On this date, cumulative ET calculations

were in deficit 2.39 cm causing a significant decline in VWC. This resulted in a maj or decline in

turf health. In previous research many indices such as NDVI do well in predicting turf stress in









response to various plant parameters and not the actual plant parameter (Trenholm et al., 1999a).

Similarly, the ability of the remote sensing devices to model VWC through individual bands,

indices and ratios, could potentially be attributed to decline in turf health in response to VWC

and not a true predictor of actual VWC.

Furthermore, the high r2 value achieved by PLS regression to model soil VWC (Table 2-

15) (Figure 2-1) could potentially be highly predictive of new data and specific to actual VWC

and not a response to plant stress. This is due to the re-sampling method employed by the PLS

procedure mentioned in the materials and methods.

Visual Ratings

Visual ratings were reasonably predicted from reflectance data from both devices. Visual

assessment of color, quality, and density was best predicted by vegetative indices NDVI and LAI

from both instruments (Table 2-12, 2-14)). Furthermore, raw reflectance values 650 nm and 880

nm from the Crop Circle device achieved reasonable r2 ValUeS (Table 2-12), as well as, 510 nm,

535 nm, 545 nm, 550 nm, 635 nm, and 661 nm individual reflectance wavelengths (Table 2-13),

630-690 nm and 2080-23 50 nm range averages, and 605/5 15 nm, 915/975 nm, and 865/725 nm

ratios (Table 2-14) from the ASD device. Conversely, quality was not as well modeled by PLS

Regression (Table 2-15), compared to the vegetative indices, individual reflectance wavelengths,

and the ratios mentioned above.

The ability of vegetative indices, NDVI and LAI, to model visual assessments could

potentially be attributed to the fact that these indices were formulated to normalize the data.

Since regression functions were run across all dates the normalization of these indices could

potentially smooth out variation in reflectance due to differences in light intensity on various

evaluation dates.









Biomass

Biomass was not modeled well by reflectance data compared to other factors tested.

However, some weak correlations exist (Tables 2-12, 2-13, 2-14, and 2-15). This is similar to

research conducted by Kruse et al., 2005, 2006; Fenstemaker-Shaulis et al., 1997; and Trenholm

et al., 1999b.

The inability to effectively model biomass in this experiment could be attributed to the

methodology used. Nitrogen treatments were only applied twice one month apart. In other

research where biomass was adequately modeled (Kruse et al., 2005, 2006) N treatments were

applied every 2 weeks thus continually promoting greater differences in turf growth.

Crop Circle vs. ASD

The Crop Circle and ASD instruments are two very different remote sensing devices that

differ dramatically in price and use-ability. The ASD device is a highly scientific instrument

with great ability to give a researcher a plethora of reflectance data from a broad region of the

electromagnetic spectrum. However, it is limited in its ability because it requires full sunlight

for use and only collects reflectance data when manually prompted by the user. Nonetheless,

when reflectance data in this magnitude is analyzed using the modern regression techniques, like

PLS regression, the correlations could potentially be very specific to the parameter tested and

highly predictive of new data. Therefore, prediction models could likely be constructed for

specific forms of turfgrass stress.

Conversely, optical sensing devices like the Crop Circle instrument are less complicated to

use but only give a limited amount of information typically in the visible and NIR regions.

However, this data can be used to compute NDVI, LAI, and a variety of other indices to make

the data more usable. Furthermore, they produce their own light source and, thus, can be

mounted on a vehicle or mower for autonomous reflectance measurement collection.









In this experiment VWC was best modeled by both remote sensing devices, Crop Circle

and ASD. NDVI from the Crop Circle (r2=0.71) (Table 2-12) (Figure 2-2) device and PLS

regression from the ASD device (r2=0.72) were the highest correlations achieved (Table 2-15)

(Figure 2-1). NDVI from the ASD device was close in comparison to the Crop Circle (r2=0.57)

to model VWC (Figure 2-2).

Even though both of these devices utilize completely different sources of light they both

calculate NDVI similarly and are close in comparison to model VWC (Figure 2-3). Granted the

prediction of VWC with PLS regression on ASD data could be highly predictive of new data and

specific to actual VWC and not just turf health, the Crop Circle device shows there is some

sensitivity to this parameter.

Conclusions

The results indicate reflectance data best modeled soil VWC, visual ratings, and some

treatment effects from N. Based on research by Ripple (1986), the elevated levels of prediction

for VWC could be due to a reduction in chlorophyll as the leaves dried in response to limiting

soil VWC and not a true relation to actual soil water status. Nonetheless, this research gives

greater insight into the use of remote sensing technology and its place for use by turf managers

and researchers demonstrating the sensitivity of remote sensing instrumentation to various

parameters of turfgrass stress. Previous research found optical sensing measurements using

NDVI could provide a fast and accurate alternative to traditional visual rating systems (Trenholm

et al., 1999b; Bell et al., 2000) as well as assessing turfgrass stress (Trenholm et. al., 1999a,

Fenstemaker-Shaulis et al., 1997). Our results from two different remote sensing instruments are

in agreement with this previous work.

The Crop Circle instrument performed equally well with the ASD device in modeling

stress attributed to reduced soil VWC (Figure 2-2) and was consistent with visual ratings in this










experiment. It is simple to use and is potentially non-restrictive of environmental conditions,

like cloud cover and variation in solar radiation. Currently, devices like this could provide fast

and easy supplemental information to turfgrass managers for assistance in making decisions.

Coupled with GPS and GIS technology, ground based remote sensing devices, like the Crop

Circle, can give turfgrass managers greater insight into managing their turf by potentially

allowing them to more efficiently schedule their irrigation, thereby reducing water consumption,

leaching, and runoff.

Furthermore, if devices like the Crop Circle could be improved to produce and quantify

additional wavelengths these devices could potentially be employed by turfgrass managers due to

their ease of use. However, further investigation into data from devices like the ASD using PLS

regression is still needed to investigate the entire spectral constituency for optimal wavelengths

to predict various turf parameters. Only then can prediction equations be formulated and

incorporated into devices like the Crop Circle that could produce the correct wavelengths to

calculate these prediction equations.





















Visual Ratings -


Table 2-1. Analysis of variance of soil volumetric water content (VWC), percent leaf nitrogen tissue concentration (%N), visual
ratings: color, quality, density (rated on NTEP 1-9 scale, where 9 is the best rating and 6 is acceptable), and biomass.
Normalized difference vegetation index (NDVI), Leaf area index (LAI), 880 nm, and 650 nm reflectance were obtained
from Crop Circle device. Crop water stress index (CWSI) was computed from IR thermometer readings. All data was
collected from fairway height hybrid bermudagrass (Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation
rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha )~, and five dates (22 June, 12
July, 25 July, 22 Aug., and 4 Sept. 2007) combined.


Crop Circlez
880
LAIW nm
.0001 0.0001
.0001 0.0001
.1709 0.0515
.9631 0.9970


1RY

CWSI"
0.2616
0.0001
0.7858
0.9896

0.5505
0.8610
0.7184


Source of
Variation df VWC % N Dry W Color Quality Density
Rep 2 0.0001 0.0044 0.0001 0.0102 0.0048 0.0015
Date 4 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001
Irrigation 3 0.3106 0.8398 0.3815 0.9382 0.5418 0.4083
Dxl 12 0.9981 0.9934 0.9401 0.9968 0.6717 0.9902
Error a 12
Nitrogen 3 0.1447 0.0001 0.1100 0.0001 0.0153 0.1087
IxN 9 0.0004 0.0502 0.3567 0.6267 0.2050 0.0065
DxN 12 0.2988 0.0002 0.4345 0.0595 0.7875 0.5126
error b 36
Probability of greater F ratio (P>F) for Date, Irrigation, and N treatment components
z Model AC S-210, Holland Scientific
SIR thermometer Raytek Raynger@ STTw, TTI, Inc.
x NDVI= (880-650 nm)/(880+650 nm)
LAI= 880/650 nm

Where Te= Canopy Temperature, Ta= Actual Temperature, a= actual, u= upper limit


650
nm
0.0001
0.0001
0.2266
0.9860

0.0001
0.3908
0.0612


NDVIx
0.0001
0.0001
0.2885
0.9921

0.0001
0.8093
0.0029


0
0
0
0


I


0.0001
0.8038
0.0001


0.2223
0.4815
0.0309


:, and 1= lower limit












Table 2-2. Analysis of variance of individual reflectance bands from ASD reflectance hyperspectral data with a range of 3 50-2500 nm
and spectral resolution of 1 nm collected from fairway height hybrid bermudagrass (Cynodon dactylon X C. transvaalensis,
Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha )~,
and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined.
ASDz (nm)
Source DF 510 535 545 550 635 661 735 755 813 935 2132
Rep 2 0.0429 0.0002 0.0001 0.0001 0.0596 0.1303 0.0001 0.0001 0.0001 0.0001 0.0584
Date 4 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001
Irrigatio 3 0.2045 0.4189 0.4541 0.4650 0.2572 0.1908 0.0999 0.1148 0. 1094 0.1134 0.1102
Dxl 12 0.0826 0.2307 0.2667 0.2756 0.0948 0.0659 0.9956 0.9554 0.9698 0.9933 0.0774
error a 12
Nitroge 3 0.0029 0.0002 0.0001 0.0001 0.0039 0.0116 0.9039 0.0275 0.0053 0.0073 0.097
IxN 9 0.5871 0.3050 0.2683 0.2631 0.6254 0.6496 0.0874 0.4093 0.4387 0.4533 0.726
DxN 12 0.4677 0.2312 0.2048 0.2034 0.4701 0.5078 0.4558 0.7430 0.0222 0.1810 0.7521
error b 36
Probability of greater F ratio (P>F) for Date, Irrigation, and N treatment components
z FieldSpec Pro; Analytical Spectral Devices, Inc., Boulder, CO










Table 2-3. Analysis of variance of average range computations from 630-690 nm and 2080-23 50 nm; NDVI, LAI, Stress1 and
Stress2 indices; and 605 nm/515 nm, 915/975 nm, and 865/725 nm ratios computed from an Analytical Spectral Device
(ASD) with a range of 350-2500nm and spectral resolution of 1 nm collected from fairway height hybrid bermudagrass
(Cynodon dactylon X C. transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values),
four N rates (0, 25, 50, and 100 kg ha )~, and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined.
ASDz
Source of 630-690 2080- 605/515 915/975 865/725
Variation DF nm 2350 nm NDVIY LAI x Stress1 Stress2 0 nm nm nm
Rep 2 0.1209 0.0627 0.0313 0.1656 0.0099 0.0545 0.0206 0.0025 0.0018
Date 4 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001
Irrigation 3 0.2010 0.1034 0.1562 0.6883 0.1654 0.4132 0.8121 0.3246 0.6495
Dxl 12 0.0712 0.0768 0.1737 0.9830 0.1296 0.4497 0.8643 0.8901 0.9094
error a 12
Nitrogen 3 0.0105 0.0869 0.0050 0.0001 0.0089 0.0019 0.0001 0.0001 0.0001
IxN 9 0.6521 0.7096 0.8275 0.9484 0.8028 0.9325 0.6612 0.9299 0.9001
DxN 12 0.5014 0.7276 0.3930 0.0008 0.2417 0.3474 0.1934 0.0299 0.0001
Error b 36
Probability of greater F ratio (P>F) for Date, Irrigation, and N treatment components
zFieldSpec Pro; Analytical Spectral Devices, Inc., Boulder, CO
YNDVI= (880-650 nm)/(880+650 nm)
x LAI= 880/650 nm
SStressl= 706/760 nm
a Stress2= 706/813 nm









Table 2-4. Evaluation of means of volumetric water content (VWC), leaf nitrogen tissue concentration (%N), visual ratings: color,
quality, density (rated on NTEP 1-9 scale 9 is best and 6 is acceptable), and biomass (g) collected from fairway height
hybrid bermudagrass (Cynodon dactylon X C. transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation rates (60, 80, 100, and 120%
estimated ET values), four N rates (0, 25, 50, and 100 kg ha- ), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4
Sept. 2007) combined.
Visual Ratings
N Rate VWC % N Color Quality Density biomass (g)
0 kg ha-l 22.04 3.38 7.44b 7.12b 7.01 48.47
25 kg hal 22.07 3.61 7.73a 7.29ab 6.99 55.85
50 kg hal 21.87 3.62 7.76a 7.22ab 7.03 53.23
100 kg hal 22.48 3.99 7.85a 7.40a 7.12 57.05
LSD N.S. -- 0.20 0.21 N.S. N.S.
"--" denotes LSD is not valid due to a significant date by N treatment interaction. Reported separately in Table 2-5 based on
interaction.
All values with a given number for the LSD denotes p < 0.05









Table 2-5. Evaluation of means of Normalized difference vegetation index (NDVI), Leaf area index (LAI), 850 nm, and 650 nm
reflectance from Crop Circle device and crop water stress index (CWSI) computed from IR thermometer readings
collected from fairway height hybrid bermudagrass (Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation
rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-l ), and five dates (22 June, 12
July, 25 July, 22 Aug., and 4 Sept. 2007) combined.
Crop Circle Reflectance Readingsz IR ThermometerY
N Rate NDVIx LAI" 880 nm 650 nm CWSI"
0 kg ha-l 0.59 4.35 1.38 .34a 0.44
25 kg hal 0.61 4.58 1.38 .33a 0.44
50 kg hal 0.60 4.61 1.38 .33a 0.45
100 kg hal 0.62 4.99 1.39 .31b 0.43
LSD -- -- N.S. 0.01 N.S.
"--" denotes LSD is not valid due to a significant date by N treatment interaction. Reported separately in Table 5 based on interaction.
All values with a given number for the LSD denotes p < 0.05
z Model ACS-210, Holland Scientific
SIR thermometer Raytek Raynger@ STTw, TTI, Inc.
P xN\DVI= (880-650 nm)/(880+650 nm)
LAI= 880/650 nm
"CWSI= (To-Ta)a (To-Ta)1 / (To-Ta)u (To-Ta)1
Where Te= Canopy Temperature, Ta= Actual Temperature, a= actual, u= upper limit, and 1= lower limit











Table 2-6. Evaluation of means of individual band reflectance from the ASD device with a range of 3 50-2500 nm and a spectral
resolution of 1 nm collected from fairway height hybrid bermudagrass (Cynodon dactylon X C. transvaalensistr~r~r~r~r~r~ Burtt-
Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha- ) and
five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined.
ASD z (nm)
N Rate 510 535 545 550 635 661 735 755 813 935 2132
0 kg hal 0.054a 0.081a 0.086a 0.088a 0.074a 0.069ab 0.29 0.366a 0.407 0.437 0.183
25 kg hal 0.055a 0.081a 0.086a 0.088a 0.075a 0.071a 0.29 0.363a 0.404 0.433 0.190
50 kg hal 0.055a 0.081a 0.086a 0.088a 0.076a 0.072a 0.29 0.361a 0.402 0.431 0.194
100 kg hal 0.051b 0.077b 0.082b 0.084b 0.068b 0.064b 0.29 0.367a 0.409 0.437 0.178
LSD 0.002 0.002 0.002 0.002 0.005 0.005 N.S. 0.007 N.S. -- N.S.
"--" denotes LSD is not valid due to a significant date by N treatment interaction
z FieldSpec Pro; Analytical Spectral Devices, Inc., Boulder, CO









Table 2-7. Evaluation of means of ASD reflectance readings of average range computations from 630-690 nm and 2080-23 50 nm,
computed indices Normalized Difference Vegetation Index (NDVI ), Leaf Area Index (LAI) and Stress1 and Stress2
indices, and ratios 605/515 nm, 915/975 nm, 865/725 nm taken from an ASD device with a range of 350-2500 nm and a
spectral resolution of 1 nm collected from fairway height hybrid bermudagrass (Cynodon dactylon X C. transvaalensistr~r~r~r~r~r~
Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-l ),
and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined.
Ranne Averane (ASD z) Venetative Indices (ASD z) Ratios (ASD z
605/515 915/975 865/725
N Rate 630-690 nm 2080-23 50 nm NDVI LAI x Stress 1 Stress 2 nm nm nm

0 kg ha-l 0.071a 0.175 0.715ab 7.06 0.367ab 0.788ab 1.02a 0.99 1.88

25 kg hal 0.073a 0.182 0.707b 7.26 0.375a 0.794a 1.03a 0.99 1.89

50 kg hal 0.743a 0.186 0.700b 7.28 0.383a 0.791a 1.03a 0.99 1.90

100 kg hal 0.657b 0.170 0.729a 8.21 0.354b 0.779b 1.03b 1.00 1.97
LSD 0.0053 N.S. 0.018 -- 0.019 0.009 0.01---
"--" denotes LSD is not valid due to a significant date by N treatment interaction
z FieldSpec Pro; Analytical Spectral Devices, Inc., Boulder, CO
YNDVI= (880-650 nm)/(880+650 nm)
x LAI= 880/650 nm
SStressl= 706/760 nm
a Stress2= 706/813 nm











Table 2-8. Evaluation of means of NDVI and LAI computed from Crop Circle instrument, and LAI computed from ASD collected
from fairway height hybrid bermudagrass (Cynodon dactylon X C. transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation rates (60,
80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and five dates (22 June, 12 July, 25
July, 22 Aug., and 4 Sept. 2007) combined, which had N treatment X date interactions.


NDVIz (Crop Circle )
25 kg 50 kg 100 kg
ha-l ha-l ha-
0.73a23 0.73a2 0.75al
0.69b2 0.69b2 0.71b1
0.69b2 0.70abl2 0.71b1
0.42dl2 0.39d2 0.53cl
0.52cl 0.49c1 0.5201


LAIx (Crop Circle)
25 kg 50 kg 100 kg
ha-l ha-l ha-
6.34a23 6.58a2 7.34al
5.49b2 5.50b2 5.99b1
5.41b2 5.69bl2 5.95 67b 1
2.45dl2 2.96cl2 2.41d2
3 .20cl 2.34dl 3.28c1


LAIx (ASD")
25 kg 50 kg
ha-l ha-
11.11a23 11.65a2
8.59b2 8.73bl2
9.29b2 9.32b2
3.39c1 3.1901
3.93cl 3.54cl


0 kg
ha-
0.70a3
0.67b3
0.66b3
0.53cl
0.44dl


0 kg
ha-
5.79a3
5.10b3
4.87b3
2.58dl
3.39c1


0 kg
ha-
10.20a23
8.28b2
8.25b3
3.66c1
4.89c1


100 kg
ha-l
13.08al
9.60b1
10.31b1
3.41cl
4.64d1


Date
6/22
7/12
7/25
8/22
9/4


Means in the same column and category followed by the same letter are not significantly different (LSD; P, 0.05); means in the same
row and category followed by the same number are not significantly different (LSD; P, 0.05).
zNDVI= (880-650 nm)/(880+650 nm)
SModel AC S-210, Holland Scientific
x LAI= 880/650 nm
" Field Spec Pro; Analytical Spectral Devices, Inc., Boulder, CO












Table 2-9. Evaluation of Means of leaf N concentration tissue (%N) collected from fairway height hybrid bermudagrass (Cynodon
dactylon X C. transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates
(0, 25, 50, and 100 kg ha- ), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined, which had N
treatment X date interactions.
%N
Date 0 kg ha-' 25 kg ha-' 50 kg ha-' 100 kg ha-'
6/22 3.51b3 4.02a2 4.19a2 4.66al
7/12 3.24bc3 3.25b2 3.38b2 3.81b1
7/25 3.26bc3 3.91a2 3.91a2 4.59al
8/22 2.90cl 2.91b1 2.75cl 2.83cl
9/4 4.01al 3.98al 3.88al 4.10b1
Means in the same column and category followed by the same letter are not significantly different (LSD; P, 0.05); means in the same
row and category followed by the same number are not significantly different (LSD; P, 0.05).












Table 2-10. Evaluation of means of individual reflectance wavelengths at 813 nm and 935 nm collected from fairway height hybrid
bermudagrass (Cynodon dactylon X C. transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated
ET values), four N rates (0, 25, 50, and 100 kg ha- ), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007)
combined, which had N treatment X date interactions.
813 nm (ASDz) 935 nm (ASDz)
Date 0 kg ha-l 25 kg ha-l 50 kg ha-l 100 kg ha-l 0 kg ha-l 25 kg ha-l 50 kg ha-l 100 kg ha-l
6/22 0.47al 0.47al 0.48al 0.49al 0.49al 0.49al 0.50al 0.51al
7/12 0.43bc1 0.43b1 0.43b1 0.44b1 0.45b1 0.45cl 0.45cl 0.46b1
7/25 0.42cl 0.43b1 0.43b1 0.43b1 0.45b1 0.46bc1 0.46bc1 0.46b1
8/22 0.26dl 0.25cl2 0.24c2 0.24c2 0.30cl 0.29dl2 0.28d2 0.28c2
9/4 0.45ab1 0.44b1 0.43b1 0.45b1 0.49al 0.48ab1 0.47b1 0.49al
Means in the same column and category followed by the same letter are not significantly different (LSD; P, 0.05); means in the same
row and category followed by the same number are not significantly different (LSD; P, 0.05).
z FieldSpec Pro; Analytical Spectral Devices, Inc., Boulder, CO












Table 2-11i. Evaluation of means of ratios 915 nm/975 nm and 865 nm/925 nm computed from analytical spectral device (ASD) with a
range of 350-2500 nm and a spectral resolution of 1 nm, collected from fairway height hybrid bermudagrass (Cynodon
dactylon X C. transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates
(0, 25, 50, and 100 kg ha- ), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined, which had N
treatment X date interactions.


915/975 nm_ (ASDz)
25 kg ha-l 50 kg ha
1.03a2 1.03a2
1.02bl2 1.02bl2
1.02b23 1.02b2
0.95dl2 0.94dl2
0.98c1 0.9801


865/925 nm (ASDz)


0 kg ha-
1.03a2
1.01b2
1.01b3
0.95dl
0.9901


100 kg ha-
1.04al
1.02b1
1.03b1
0.94d2
0.9901


0 kg ha-
2.12a3
2.00b2
1.96b3
1.66c1
1.68c


25 kg ha-
2.20a23
2.02b2
2.03b23
1.63cl2
1.60cl


50 kg ha-' 100 kg ha'
2.23a2 2.35al
2.04b2 2.13b1
2.04b2 2.12b1
1.61cl2 1.63c2
1.57c1 1.63c1
0.05); means in the same


Date
6/22
7/12
7/25
8/22
9/4


Means in the same column and category followed by the same letter are not significantly different (LSD; P,
row and category followed by the same number are not significantly different (LSD; P, 0.05).
z FieldSpec Pro; Analytical Spectral Devices, Inc., Boulder, CO





Parameter NDVIx LAI" 880 nm 650 nm
%N 0.19*** 0.21*** 0.15*** 0.15***
VWC 0.71*** 0.59*** 0.61*** 0.52***
Color 0.59*** 0.57*** 0.39*** 0.61***
Quality 0.70*** 0.67*** 0.50*** 0.71***
Density 0.67*** 0.65*** 0.48*** 0.69***
Biomass 0.26*** 0.33*** 0.14*** 0.34***
*, **,** Significant at the 0.05, 0.01, and 0.001 probability levels, respectively
z MLodel ACS-210, Holland Scientific
SIR thermometer Raytek Raynger@ STTw, TTI, Inc.
xNDVI= (880-650 nm)/(880+650 nm)
LAI= 880/650 nm
"CWSI= (To-Ta~a (To-Ta)1 / (To-Ta)u (To-Ta)I
Where Te= Canopy Temperature, Ta= Actual Temperature, a= actual, u= upper limit, and 1= lower limit


Table 2-12. Coefficient estimates (r2) Of Crop Circle reflectance data Normalized Difference Vegetation Index (NDVI), Leaf Area
Index (LAI), NIR, and Red, and crop water stress index (CWSI) to soil volumetric water content (VWC) visual color,
quality, density ratings collected from fairway height hybrid bermudagrass (Cynodon dactylon X C. transvaalensis, Burtt-
Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha )~, and
averaged over all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007).
Crop, Circlez IR ThermometerY


CWSI"
0.02*
0.02*
0.13***
0.33***
0.38***
0.05**












Table 2-13. Coefficients estimates (r2) Of ASD reflectance readings of individual reflectance wavelengths taken from an Analytical
spectral device (ASD) with a range of 3 50- 2500 nm and a spectral resolution of 1 nm to %N concentration leaf tissue, soil
volumetric water content (VWC) visual color, quality, density ratings, and biomass production of fairway height hybrid
bermudagrass (Cynodon dactolon X C. transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated
ET values), four N rates (0, 25, 50, and 100 kg ha )~, and averaged over all dates (22 June, 12 July, 25 July, 22 Aug., and 4
Sept. 2007).
ASDz Reflectance (nm)

Parameter
510 535 545 550 635 661 735 755 813 935 2132
%/N 0.00 0.02* 0.02* 0.02* 0.01 0.02* 0.25*** 0.29*** 0.29*** 0.28*** 0.02*
VWC 0.08*** 0.01 0.00 0.36 0.19*** 0.23*** 0.38*** 0.53*** 0.52*** 0.49*** 0.29***
Color 0.28*** 0.15*** 0.13*** 0.13*** 0.36*** 0.38*** 0.09*** 0.23*** 0.24*** 0.19*** 0.40***
Quality 0.52*** 0.32*** 0.31*** 0.31*** 0.60*** 0.62*** 0.06*** 0.22*** 0.23*** 0.17*** 0.69***
ulDensity 0.55*** 0.37*** 0.34*** 0.35*** 0.62*** 0.65*** 0.17** 0.18*** 0.19*** 0.13*** 0.71***
Biomass 0.25*** 0.19*** 0.18*** 0.18*** 0.26*** 0.25*** 0.01 0.06*** 0.07*** 0.05** 0.25***
*, **, *** Significant at the 0.05, 0.01, and 0.001 probability levels, respectively
z FieldSpec Pro; Analytical Spectral Devices, Inc., Boulder, CO












Table 2- 14. Coefficients estimates (r2) Of average range computations from 630-690 nm and 2080-23 50 nm, and computed indices
Normalized difference vegetation index (NDVI), Leaf area index (LAI) and Stress1 and Stress2 indices also taken from an
ASD with a range of 3 50-2500 nm and a spectral resolution of 1 nm to %N concentration leaf tissue, soil volumetric water
content (VWC) visual color, quality, density ratings, and biomass production collected from fairway height hybrid
bermudagrass (Cynodon dactylon X C. transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated
ET values), four N rates (0, 25, 50, and 100 kg ha )~, and averaged over all dates (22 June, 12 July, 25 July, 22 Aug., and 4
Sept. 2007).
Range Average (ASDz) Vegetative Indices (ASDz) Ratios (ASDz)
630-690 2080-2350 605/515 915/975 865/725
Parameter nm nm NDVIY LAIx Stress 1" Stress 2" nm nm nm
%/N 0.02 0.02* 0.13*** 0.20** 0.15*** 0.16*** 0.21*** 0.22*** 0.14***
VWC 0.22*** 0.29*** 0.57*** 0.53*** 0.61*** 0.64*** 0.66*** 0.68*** 0.44***
Color 0.38*** 0.40*** 0.53*** 0.55*** 0.53*** 0.55*** 0.45*** 0.52*** 0.51***
Quality 0.62*** 0.69*** 0.71*** 0.66*** 0.67*** 0.69*** 0.46*** 0.61*** 0.67***
Density 0.65*** 0.71*** 0.69*** 0.63*** 0.65*** 0.67*** 0.44*** 0.58*** 0.67***
Biomass (g) 0.26*** 0.25*** 0.26*** 0.35*** 0.24*** 0.27*** 0.15*** 0.24*** 0.67***


*, **, *** Significant at the 0.05, 0.01, and 0.001 probability levels, respectively
z FieldSpec Pro; Analytical Spectral Devices, Inc., Boulder, CO
YNDVI= (880-650 nm)/(880+650 nm)
x LAI= 880/650 nm
" Stressl= 706/760 nm
SStress2= 706/813 nm


ul









Table 2-15. Partial Least Squares regression on hyperspectral data from ASD device with a range of 350-2500 nm and a spectral
resolution of 1 nm for prediction of soil volumetric water content (VWC), percent leaf nitrogen concentration (%N),
biomass, and visual quality collected from fairway height hybrid bermudagrass (Cynodon dactylon X C. transvaalensis,
Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha )~
averaged over all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007).
Calibration No. of Factors* r2 SE n
VWC 8 0.72 0.030 240
% N 7 0.25 0.043 240
Biomass (g) 6 0.31 0.040 240
Visual Quality 7 0.55 0.031 240
*Number of factors required to achieve a minimal Predicted Residual Sum of Squares (PRESS) of prediction for the partial least
squares regression model.
z FieldSpec Pro; Analytical Spectral Devices, Inc., Boulder, CO
















40
ASD *
35 y = 0.7166x + 6.231
U R2 = 0.7083 4 e


20 4
a ~~***4 *~
~ zj *


10 4* *i *

10


0 5 10 15 20 25 30 35 4
-5
Actual VWC


Figure 2-1. Partial Least Squares regression on hyperspectral data from ASD device with a range Of 350-2500 nm and a spectral range
of 1 nm for prediction of volumetric water content (VWC) collected from fairway height hybrid bermudagrass (Cynodon
dactylon X C. transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates
(0, 25, 50, and 100 kg ha )~, averaged over all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007).













0.9

0.8

0.7

0.6

0.5

0.4


ASD
y = 52.74x 15.50
R2 = 0.57


Crop Circle
y = 65.19x 17.46
R2 = 0.71


VWC


Figure 2-2. Coefficient estimates (r2) Of Crop Circle and ASD reflectance data computed into a Normalized Difference Vegetation
Index (NDVI), to soil volumetric water content (VWC) collected from fairway height hybrid bermudagrass (Cynodon
dactylon X C. transvaalensistr~r~r~r~r~r~ Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates
(0, 25, 50, and 100 kg ha )~, and averaged over all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007).

































0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1


ASD


Figure 2-3. Coefficient estimates (r2) of Crop Circle vs. ASD using reflectance data to compute a Normalized Difference Vegetation
Index (NDVI) collected from fairway height hybrid bermudagrass (Cynodon dactylon X C. transvaalensistr~r~r~r~r~r~ Burtt-Davy),
four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and averaged
over all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007).









CHAPTER 3
USE OF GROUND-BASED REMOTE SENSING TECHNOLOGY TO ASSESS BIOMASS
PRODUCTION IN RESPONSE TO PLANT GROWTH REGULATOR APPLICATIONS

Introduction

Remote sensing is defined as obtaining information about an obj ect, area, or phenomenon

by analyzing data acquired by a device that is not in contact with that obj ect, area, or

phenomenon (Lillesand and Keifer, 1987). For many years researchers have entertained the idea

of measuring various plant parameters using remote sensing technology. Plant light interception

significantly influences growth and physiological responses (Salisbury and Ross, 1992). When

light is intercepted by the plant it is absorbed, transmitted, or reflected (Salisbury and Ross,

1992). Using remote sensing technology to quantify the light that is reflected could help to

detect the onset of turfgrass stress (Ikemura and Leinauer, 2006). Real world applications of

remote sensing technology for use in turfgrass management are still in their infancy; however,

studies have shown that image analysis and various remote sensing devices have strong potential

to detect a variety of turfgrass stresses (Ikemura and Leinauer, 2006). Recent investigations into

this technology show that there is potential for turfgrass managers to manage and assess turf

more efficiently. The use of remote sensing technology to estimate turfgrass stress could

significantly decrease the time and labor required to assess these levels in traditional ways, thus

reducing cost as well (Osborne et al., 2002).

The many great achievements in U.S. agricultural productivity can be attributed to the use of

agricultural chemicals including fertilizer and pesticides (Lee et al., 1999). However, the

increased dependence on pesticides and fertilizers have heightened many environmental

concerns especially in Florida due to the sandy soils and heavy rainfall, which increases potential

for heavy runoff and leaching of chemicals, if applied in excess (Min and Lee, 2003). With

increasing water and environmental restrictions, turf managers and researchers must be aware of









technological advances in their industry. These environmental concerns have pressed turf

managers to reduce nutrient and pesticide inputs used for turf maintenance (Bell et al., 2004). In

other industries, such as agronomic food production crops, precision agriculture has been used as

a management tool to maximize yield and minimize cost (Bethel et al., 2003). The basis of this

technology relies heavily on the Global Positioning System (GPS), Geographic Information

Systems (GIS), Variable Rate Application (VRA) and remote sensing technology. Using this

technology to apply nutrients, water, and pesticides only where they are needed can optimize

yield while minimizing cost. (Lee et al., 1999)

Plant growth regulators (PGRs) are used by turf managers to suppress vertical turf

growth, which ultimately reduces maintenance costs due to reduced mowing requirements. A

PGR is a substance which adjusts the growth and development of a plant. This is generally

achieved through the inhibition of the gibberellic acid (GA), a hormone which is responsible for

cell elongation. PGRs were first introduced for use on fine turf in 1987 (VanBibber, 2006).

Some data suggests that PGR-treated turf can produce greater root mass, recover from injury

faster, reduce water use rate, reduce disease incidence, and reduce Poa annua weed populations

(Branham, 1997). The first of these chemicals were flurprimidol (Trade name Cutless) and

paclobutrazol (Trade name Trimmet) which inhibit GA synthesis at relatively the same point in

the GA biosynthesis pathway resulting in very similar plant responses (Branham, 1997).

However, paclobutrazol is a more active compound and requires lower rates to achieve the same

response as higher rates of flurprimidol (Branham, 1997). The newest of the PGR' s, trinexapac-

ethyl (TE) (Trade name Primo), became commercially available in 1995 and has become the

standard for use throughout the turf industry (Branham, 1997). TE facilitates GA inhibition later









in the GA biosynthesis pathway so that GA' s are formed but are not active and serve as

precursors for plant biochemical processes.

Historically PGRs are relatively expensive compared to other chemical applications made

by turf managers. In an effort to reduce the cost of using PGRs, researchers have investigated

new methods to reduce the use rate. Researchers used remote sensing, the global positioning

system (GPS), and geographic information systems (GIS), and variable rate application (VRA)

technology to make site specific applications ofPGRs. These image-based PGR applications

were tested in cotton (Gossypium sp.) to reduce the chemical input of PGRs while also reducing

cost. Among comparisons of different image based-application (Variable Rate and Site-Specific)

methods to traditional broadcast applications for PGRs in cotton, it was found that the broadcast

applications were on average 19 % more costly than the site-specific applications, and 27 %

more costly than applications using variable rate methods. This demonstrates the potential

economic benefit of remote sensing technology for PGR use (Bethel, 2003).

In the turf industry, research has been conducted to assess various parameters of turfgrass

stress using remotely sensed data. Several different indices have been developed that have been

correlated with plant stress. One of the most common indices to assess turfgrass stress is the

Normalized Difference Vegetation Index (NDVI) which is computed as the reflectance from the

Near Infrared (NIR) region minus the reflectance from the Red (R) region divided by a sum of

both (NIR-R )/(NIR+R) (Rouse et al., 1973). Fenstemaker-Shaulis et al. (1997) found a negative

correlation with NDVI and canopy temperature, (r2=0.74) and a positive correlation with NDVI

and plant moisture content (r2=0.90). They also concluded that NDVI was not as a strong

predictor of biomass (r2=0.37). Trenholm et al. (1999b) tested various wavelengths 507 nm, 559

nm, 661 nm, 706 nm, 760 nm, 813 nm, and 935 nm as well as NDVI, IR/R (LAI) computed as









935/661 nm, Stress1 computed as 706/760 nm, and Stress2 computed as 706/813 nm. The

results indicated that reflectance data was not successful in explaining variability in the models

for clipping yield. However, Kruse et al. (2006) found conducting Partial Least Squares

regression on hyperspectral data (400-1,100nm) was an adequate predictor of biomass in

creeping bentgrass (Agrostis stolinifera L.)

With the introduction of hyperspectral data, which typically measures more than 200

wavelengths, there has also been investigation of different statistical methods to quantify and

interpret remotely sensed data. Traditionally, linear regression is the simplest way to understand

variance. However, in interpreting reflectance data, it has been found that other regression

techniques better assess plant status. Partial Least Squares Regression (PLS) is a method used to

predict variables when there are a large number of factors, typically more than the number of

observations (Tobias, 1997). The use of multiple variables is generally tested through Multiple

Linear Regression (MLR) (Tobias, 1997). However, the limitation with MLR is when the

number of factors becomes too large the model may become over-fitting (Tobias, 1997). This is

when it uses a large number of variables but only a few account for most of the variation, also

called latent factors (Tobias, 1997). Almost perfect models may be achieved; however they will

not be highly predictive of new data. When prediction is the obj ective then PLS regression is a

highly useful tool (Tobias, 1997).

Currently, the maj ority of remote sensing research is done on agronomic food crops and

forested areas. There is only limited research available in the area of remote sensing and

turfgrass management (Trenholm et al., 1999b). If conditions across the entire golf course were

uniform in microclimates, soil types, turf varieties, pest densities, nutrient status, and irrigation

performance there would be little need for advanced sensor systems to measure the variability of









these parameters to better meet the needs on a locational basis (Stowell and Gelernter, 2006).

However highly variable agronomic conditions create the need for further research to

incorporate the use of remote sensing into turfgrass management to reduce chemical inputs.

(Trenholm et al., 1999b). The obj ective of this research was to assess the ability of remote

sensing instruments to detect differences in turf growth parameters as influenced by applications

of PGRs on hybrid bermudagrass (Cynodon dactylon X C. transvaalensis~t~t~t~t~t~t~ 'Princess 77) turf.

Materials and Methods

The field experiment was conducted at the University of Florida, West Florida Research

and Education Center, Jay, FL. Plots measured 1.5 m x 3.0 m and treatments were arranged in a

factorial design observing three different PGRs, each at four incremental levels- V/2, 1, 2, and 4

times the labeled rates, with three replications per treatment. Plant growth regulators were

applied on 14 June 2007 and 20 July 2007. Plant growth regulators treatments included

trinexapac-ethyl applied at 0.05, 0.1, 0.2, and 0.4 kg a.i. ha- ethephon applied at 3.8, 7.6, 15.2,

30.4 kg a.i. ha- and flurprimidol applied at 0.3, 0.6, 1.1, and 2.3 kg a.i. ha- Plots were mowed

2 to 3 times per week at a height of 1.2 cm with a Toro Reelmaster 3100-D. Supplemental

irrigation was applied as needed. Data was collected five times throughout the growing season

in 2007 on 22 June, 12 July, 25 July, 22 Aug., and 4 Sept. When data were collected all

measurements were taken within a one hour time period.

Plots were rated visually for color, quality, and density based on the standard 1-9

National Turfgrass Evaluation Program (NTEP) rating scale where 9 is the highest and 6 is the

lowest acceptable rating. Soil volumetric water content readings were taken via Time Domain

Reflectometry with a Fieldscout TDR 300 soil moisture meter (Spectrum Technologies, Inc.).

Clipping samples for biomass determination were collected from a 4.3 m2 area immediately after

reflectance data and visual evaluations were obtained. Clipping tissue samples were weighed









and oven dried at 52 oC for 7d before being weighed again for determination of biomass

production.

Reflectance Measurements

Spectral measurements were taken from two remote sensing devices. A ground based

vehicle mounted optical sensor was used to assess turf health. A Crop Circle (Model ACS-210)

(Holland Scientific) was fitted onto a Toro Greensmaster 1000 walking greens mower 0.81 m

above ground which provided a 0.45 m2 field of view. The Crop Circle sensor produces it own

light source at 650 nm and 880 nm and measures reflected values to produce an NDVI which is

calculated as (880-650 nm)/(880+650 nm). No calibration method was used for the sensor since

it has the ability to detect its own light source from incoming ambient sunlight.

A hand-held hyperspectral spectroradiometer (FieldSpec Pro; Analytical Spectral

Devices, Inc., Boulder, CO) was also used to collect reflectance data. This radiometer has a

spectral range of 300-2,500 nm with a 1 nm resolution and a 23 o foreoptic. Reflectance readings

were collected from shoulder height on each plot which provided a 0.65 m2 field of view.

Radiance values are expressed in percent reflectance compared to standardization with a white

reference value across the entire spectral range. A white reference was used for calibration

purposes at the beginning of data collection with the ASD unit, and again every 15 minutes

depending on the length of time needed to collect data. Usually only one white reference was

needed to take data on all plots. All reflectance readings were taken in full sunlight between the

hours of 1100 and 1400 Central Standard Time (CST) to minimize variance caused by incoming

solar radiation.

Statistical Analysis

Analysis of variance (ANOVA) was performed using the PROC GLM method (SAS

Inst., 2003) to compare the differences among PGR treatments to biomass, soil VWC, visual









ratings, Crop Circle readings (NDVI and LAI) and ASD readings (NDVI, LAI, Stress1, and

Stress2). Regression analysis was performed using the PROC REG method (SAS Inst., 2003) to

correlate relationships between visual ratings, Crop Circle readings, various individual

reflectance wavelengths quantified from the ASD unit, and soil VWC. Additionally, due to the

large amount of data generated by the ASD device, ASD data were rendered using partial least

squares regression PROC PLS (SAS Inst., 2003) to soil VWC, visual quality ratings, and

biomass production. Equations were validated through a single sample cross-validation. For

cross-validation, ten percent of the sample was omitted for prediction purposes so that the

number of factors chosen creates the minimal predicted residual sum of squares (PRESS). This

process was repeated so that every observation was used exactly once for cross-validation.

Results and Discussion

Biomass

ANOVA reveals that there was a significant reduction in biomass in the 0.1 kg a.i. ha-l

and 0.4 kg a.i. ha-l rates of TE, the highest rate of flurprimidol (2.3 kg a.i. ha-l ) and the low rate

of ethephon (30.4 kg a.i. L ha- ) (Table 3-1). Although not always significant, a dose response

trend was observed with TE (Figure 3-1) and flurprimidol (Figure 3-2). This dose response was

not observed with Ethephon due to the fact that ethephon is primarily used for seedhead

suppression as opposed to biomass suppression in turf (Anonymous 2007).

Biomass did not correlate well with any remotely sensed parameter quantified from either

device (Tables 3-3, 3-4, 3-5, and 3-6). This data is in agreement with the research conducted by

Fenstemaker-Shaulis et al. (1997) and Trenholm et al. (1999b) who found remotely sensed data

was not an adequate predictor of biomass. The inability to adequately model biomass reduction

through increasing rates of PGR applications may be attributed to a number of factors. PGRs

have been found to limit vertical growth but increase prostrate growth, thus increasing density









(Branham, 1997). In other research where biomass has been adequately modeled, PGRs were

not used (Kruse et al., 2005, 2006) and biomass was modeled from plots with increasing N rate

applications applied at frequent intervals.

Volumetric Water Content

PGRs applied to bermudagrass in this study had only a slight effect on soil VWC. Only

TE and flurprimidol applied at the lowest rate gave significant differences from the untreated

control (Table 3-1). Slight correlations were observed (Table 3-3) in modeling soil VWC with

NIR reflectance quantified from the Crop Circle device as well as reflectance in the SWIR region

at 2,132 nm (Table 3-4) and the 2,080-2,350 nm range (Table 3-5) quantified from the ASD

device. This is in agreement with research conducted by Ripple (1986) and Hutto et al. (2006).

Visual Ratings

Color

Color rating means (Table 3-1) show an increase in turf color by TE applications at 0.05 and 0. 1

kg a.i. ha- flurprimidol applications at 0.3, 0.6, and 1.1 kg a.i. ha- and ethephon applications at

the low rate compared to the untreated control, however, these increases were not significant

(Table 3-1). This general increase could be attributed to seed head reduction achieved by PGR

applications, since seed heads are generally less green causing a waning in color on the turf.

Additionally, increasing rates of PGRs resulted in a trend towards reduced turf color which could

be caused by a slight injury effect.

Only slight correlations in visual color ratings to remotely sensed data from either device

were achieved (Tables 3-3, 3-4, 3-5, 3-6). The highest correlations were observed in data taken

with the ASD device at wavelengths 661 nm (r2=0.42) and 5 10nm (r2=0.3 5), and ratio 605/5 15

nm (r2=0.41) (Table 3-4 and 3-5).









Quality

Improvements in quality were observed with PGR applications. Although they were not

significant, all rates of TE and flurprimidol, and the 3.8 kg a.i. ha-l rate of ethephon slightly

improved turf quality (Table 3-1). Similar to turf color ratings, only slight correlations existed

between remotely sensed data and visual quality ratings. The highest correlations were achieved

by the ASD device at wavelengths 661 nm (r2=0.53) and 510 nm (r2=0.48), and ratio 915/975 nm

(r2=0.42) (Table 3-4 and 3-5).

Density

Visual density ratings showed a general increase in density from all rates of TE and

flurprimidol, and the low rate of ethephon, however these increases were not significant. Only

slight correlations existed between remotely sensed data and visual density ratings. The highest

correlations were achieved by the ASD device at wavelengths 661 nm (r2=0.53) and 510 nm

(r2=0.48) (table 3-4).

Reflectance Indices vs. PGR Application rates

Data derived from the Crop Circle device show a general increase in NDVI and LAI for

all rates of TE and flurprimidol compared to the untreated, however, only the 0.05 and 0. 1 kg a.i.

ha-l rate of TE and the 0.3, 0.6, and 1.1 kg a.i. ha-l rate of flurprimidol are significant (Table 3-

2). Furthermore, NDVI and LAI derived from ASD data show the same trend, yet only the 2.2

kg ha-l is statistically different from the untreated control (Table 3-2). The general trend for the

STRESS 1 and STRESS 2 indices show similar results, however, STRESS 1 reveals that TE and

flurprimidol treated turf was under less stress than the untreated control (Table 3-2).

Crop Circle Device

Reflectance data obtained from the Crop Circle device did not demonstrate it's ability to

detect biomass differences attributed to PGR applications (Table 3-3). The ability of the device









to correlate with visual human evaluations was weak. However, the NIR (880 nm) reflectance

values from the device slightly modeled changes in soil VWC (r2=0.43) throughout the

experiment (Table 3-3).

The inability of the Crop circle device to adequately detect differences in biomass, visual

ratings, and VWC can be attributed to a number of things previously mentioned. Among these

are the seedhead production of the turf which could have potentially skewed the reflectance data,

and the lack of significant differences among the parameters tested.

ASD Device

Individual wavelengths, computed indices, and various ratios obtained from ASD

reflectance data did not demonstrate the ability to detect differences in biomass attributed to PGR

applications. Some weak correlations exist between visual ratings and NDVI, LAI, Stress1, and

Stress2 indices computed from the raw data generated by the ASD instrument (Table 3-3).

Similar, and sometimes stronger, correlations exist with visual ratings with individual

wavelengths investigated throughout the visible range (Table 3-4). Again, this inability to

adequately model visual ratings, VWC, and biomass could be attributed to the lack of significant

differences in the various parameters of the treated verses the untreated control and the seedhead

production of the turf.

PLS Regression

The use of PLS regression on the hyperspectral data generated from the ASD instrument

did not perform well in predicting differences in biomass created by various applications of

PGRs. Furthermore PLS regression only slightly predicted differences in soil VWC (r2=0.36)

and visual quality (r2=0.26) (Table 3-6). This again is likely due to the lack of significant

differences in soil VWC and visual quality from the untreated control.









Conclusions

This experiment shows the biomass reduction expected from increasing rates of PGRs in

two of the three PGRs applied (Table 3-1) (Figure 3-1, 3-2). Flurprimidol and TE reduced

biomass in response to increasing application rates compared to the untreated control; however,

the results are not always significant (Table 3-1) (Figure 3-1, 3-2). Similar trends were observed

in NDVI and LAI indices computed from Crop Circle and ASD instruments, as well as Stress1

and Stress2 indices computed from the ASD device (Table 3-2). Increases in NDVI and LAI

which model plant health, as well as reductions in Stress1 and Stress2 indices, could be

attributed to suppression of seed head production in addition to the decrease of overall biomass.

This overall decrease in biomass prevented the scalping of the turf in the study area that was

observed in the untreated control and surrounding areas not treated with PGRs.

The inability to adequately model biomass reduction, visual ratings, and VWC through

increasing rates of PGR applications may be attributed to a number of factors already mentioned.

PGRs have been found to limit vertical growth but increase prostrate growth, thus increasing

density (Branham, 1997). A general increase in density was observed in this experiment with

increasing application rates of TE and flurprimidol, however, they were not significantly

different from the untreated control (Table 3-1). The lack of significant differences in density

and all visual ratings affected the ability of the remote sensing devices to detect differences

among theses various parameters. Additionally, this experiment was performed on Princess 77

bermudagrass turf which produces seed heads. These seed heads are white in color and could

potentially skew remotely sensed data compared to a turf that does not produce seed heads.

Currently, turf managers apply PGRs in the form of broadcast applications to reduce

biomass. If remote sensing technology could be used to adequately detect biomass, and was

coupled with GPS and GIS technologies, applications of PGRs could be applied on a site specific









basis. In this setting, turf managers could potentially reduce chemical input thereby reducing the

high cost of broadcast PGR applications.










Table 3-1. Evaluation of means of biomass (g), volumetric water content (VWC), visual ratings: color, quality, density (rated on
NTEP 1-9 scale 9 is best and 6 is acceptable), from fairway height hybrid bermudagrass (Cynodon dactylon X C.
transvaalensistr~r~r~r~r~r~ 'Princess 77) over five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined.
Biomass (g) VWC Color Quality Density
Trinexapac-ethyl
0.05 kg a.i. hal 52.47 a-c 35.73 a 8.00 ab 8.00 ab 7.87 a-c
0.1 kg a.i. ha-l 45.20 bc 34.45 a-c 8.10 ab 8.05 ab 8.00 ab
0.2 kg a.i. ha-l 47.20 a-c 32.87d 7.73 a-c 7.80 a-c 7.93 ab
0.4 kg a.i. ha-l 43.80 bc 34.20 b-d 7.83 a-c 7.98 ab 8.00 ab
Flurprimidol
0.3 kg a.i. ha-l 53.93 a-c 35.85 a 8.10 ab 8.17 a 8.10 ab
0.6 kg a.i. ha-l 49.93 a-c 34.85 a-c 8.19 a 8.11 a 8.17 a
1.1 kg a.i. ha-l 48.13 a-c 35.23 ab 8.07 ab 8.08 a 8.10 ab
2.3 kg a.i. ha-l 36.27 c 34.72 a-c 7.83 ab 7.93 ab 7.83 a-c
Ethephon
3.8 kg a.i. ha-l 45.73 bc 35.21 ab 7.96 ab 7.91 ab 7.80 a-c
7.6 kg a.i. ha-l 56.07 a-c 35.24 ab 7.63 bc 7.58 b-d 7.67 b-d
15.2 kg a.i. hal 66.13 ab 33.88 b-d 7.47 c 7.33 cd 7.40 cd
30.4 kg a.i. hal 66.27 ab 33.37 cd 7.40 c 7.25 d 7.30 d

Untreated 71.60 a 34.01 b-d 7.87 a-c 7.77 a-c 7.73 a-d
Means in the same column followed by the same letter are not significantly different (LSD; P< 0.05)










Table 3 -2. Evaluation of means of Normalized Difference Vegetation Index (NDVI), Leaf Area Index (LAI) computed from Crop
Circle-and ASD reflectance data, as well as Stress1 and Stress2 indices computed from ASD reflectance data from fairway
height hybrid bermudagrass (Cynodon dactylon X C. transvaalensistr~r~r~r~r~r~ 'Princess 77) over five dates (22 June, 12 July, 25
July, 22 Aug., and 4 Sept. 2007) combined.
Crop Circlez ASDY
NDVIx LAI" NDVIx LAI" Stress1v Stress2u
Trinexapac-ethyl
0.05 kg a.i. hal 0.66 ab 5.08 ab 0.76 ab 7.42 a-c 0.38 bc 0.79 cd
0.1 kg a.i. ha-l 0.67 ab 5.13 ab 0.77 ab 7.76 ab 0.36 c 0.77 cd
0.2 kg a.i. ha-l 0.65 a-d 4.90 a-c 0.76 ab 7.43 a-c 0.38 bc 0.78 cd
0.4 kg a.i. ha-l 0.66 a-c 5.00 a-c 0.76 ab 7.46 a-c 0.37 bc 0.78 cd
Flurprimidol
0.3 kg a.i. ha-l 0.67 ab 5.10 ab 0.77 ab 7.51 ab 0.37 bc 0.78 cd
0.6 kg a.i. ha-l 0.66 ab 5.01 ab 0.76 ab 7.34 a-d 0.37 bc 0.78 cd
1.1 kg a.i. ha-l 0.68 a 5.24 a 0.78 a 7.82 a 0.36 c 0.77 d
2.3 kg a.i. ha-l 0.66 a-c 5.03 ab 0.76 ab 7.49 a-c 0.37 bc 0.78 cd
Ethephon
3.8 kg a.i. ha-l 0.65 a-d 4.82 a-c 0.75 a-c 7.16 a-d 0.39 a-c 0.78 b-d
7.6 kg a.i. ha-l 0.64 b-d 4.64 b-d 0.75 a-d 6.96 a-d 0.39 a-c 0.78 a-d
15.2 kg a.i. hal 0.62 de 4.45 dc 0.72 cd 6.50 cd 0.41 a 0.79 ab
30.4 kg a.i. hal 0.60 e 4.23 d 0.72 d 6.39 d 0.42 a 0.80 a

13) Untreated 0.63 c-e 4.64 b-d 0.74 b-d 6.81 b-d 0.40 ab 0.79 a-c
Means in the same column followed by the same letter are not significantly different (LSD; P< 0.05)
z Model ACS-210, Holland Scientific
YFieldSpec Pro; Analytical Spectral Devices, Inc., Boulder, CO
xNDVI= (880-650nm)/(880+650nm)
" LAI= 880/650nm
SStressl= 706/760nm
a Stress2= 706/813nm










Table 3-3. Coefficient estimates (r2) Of Crop Circle reflectance data, Normalized Difference Vegetation Index (NDVI), Leaf Area
Index (LAI), 880nm, and 650nm, and, ASD reflectance data computed into NDVI, LAI, Stress1, and Stress2 indices to
volumetric water content (VWC) visual color quality, density ratings, and biomass production from fairway height hybrid
bermudagrass (Cynodon dactylon X C. transvaalensis, 'Princess 77) averaged over five dates (22 June, 12 July, 25 July, 22
Aug., and 4 Sept. 2007).
Crop Circlez ASDY
r2 NDVIx LAI" 880nm 650nm NDVIx LAI" Stress 1 Stress2u
VWC 0.10*** 0.12*** 0.43*** 0.00 0.11*** 0.10*** 0.09*** 0.10***
Color 0.18*** 0.16*** 0.00 0.28*** 0.23*** 0.22*** 0.19*** 0.21***
Density 0.22*** 0.19*** 0.03* 0.21*** 0.28*** 0.23*** 0.24*** 0.25***
Quality 0.22*** 0.18*** 0.02 0.24*** 0.29*** 0.24*** 0.23*** 0.26***
Biomass 0.06** 0.10*** 0.00 0.09*** 0.03* 0.08*** 0.05* 0.05***
*, **,** Significant at the 0.05, 0.01, and 0.001 probability levels, respectively
z Model ACS-210, Holland Scientific
YFieldSpec Pro; Analytical Spectral Devices, Inc., Boulder, CO
xNDVI= (880-650nm)/(880+650nm)
~" LAI= 880/650nm
SStressl= 706/760nm
a Stress2= 706/813nm












Table 3-4. Coeffieient estimates (r2) Of ASD reflectance data at individual wavelengths, as well as range averages from 630-690nm
and 2080-23 50nm to volumetric water content (VWC) visual color, quality, density ratings, and biomass production from
fairway height hybrid bermudagrass (Cynodon dactylon X C. transvaalensis, 'Princess 77) averaged over Hyve dates (22
June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007). All numbers in column headings are expressed nanometers (nm).
ASD z (nm)
r2 510 535 545 550 661 735 755 813 935 2132
VWC 0.12*** 0.06** 0.05* 0.04* 0.07*** 0.02* 0.04* 0.06* 0.11*** 0.43***
Color 0.35*** 0.25*** 0.25*** 0.25*** 0.42*** 0.05* 0.01 0.01 0.02* 0.11***
Density 0.51*** 0.39*** 0.38*** 0.39*** 0.53*** 0.07* 0.01 0.01 0.02 0.06**
Quality 0.48*** 0.35*** 0.35*** 0.35*** 0.53*** 0.06* 0.01 0.01 0.02 0.08***
Biomass 0.00 0.00 0.00 0.01 0.01 0.02 0.03* 0.04* 0.05* 0.01
*, **, *** Signifieant at the 0.05, 0.01, and 0.001 probability levels, respectively
z FieldSpec Pro; Analytical Spectral Devices, Inc., Boulder, CO













Table 3-5. Coefficient estimates (r2) Of range averages from 630-690 nm and 2080-23 50 nm and ratios 605/5 15 nm, 915/975 nm, and
865/725 nm to volumetric water content (VWC) visual color, quality, density ratings, and biomass production from fairway
height hybrid bermudagrass (Cynodon dactylon X C. transvaalensistr~r~r~r~r~r~ 'Princess 77) averaged over five dates (22 June, 12
July, 25 July, 22 Aug., and 4 Sept. 2007). All numbers in column headings are expressed nanometers (nm).
ASD z (nm)
r2 630-690 2080-2350 605/515 nm 915/975 nm 865/725 nm
VWC 0.19*** 0.40*** 0.12** 0.14** 0.03
Color 0.19*** 0.12*** 0.41*** 0.32*** 0.24***
Density 0.30*** 0.08*** 0.32*** 0.39*** 0.22***
Quality 0.29*** 0.10 0.26*** 0.42*** 0.23***
Biomass 0.00 0.01 0.08* 0.06* 0.13**
*, **, *** Significant at the 0.05, 0.01, and 0.001 probability levels, respectively
z FieldSpec Pro; Analytical Spectral Devices, Inc., Boulder, CO










Table 3-6. Partial Least Squares regression coefficients on hyperspectral data from ASD device to predict volumetric water content
(VWC), biomass, and visual quality (rated on NTEP 1-9 scale 9 is best and 6 is acceptable) in fairway height bermudagrass
(Cynodon dactylon X C. transvaalensis, 'Princess 77) averaged over five dates (22 June, 12 July, 25 July, 22 Aug., and 4
Sept. 2007)
Calibration No. of FactorsY r2 SE n
VWC 5 0.36 0.05 195
Biomass 2 0.03 0.05 195
Visual Quality 8 0.26 0.03 195
SNumber of factors required to achieve a minimal Predicted Residual Sum of Squares (PRESS) of prediction for the partial least
squares regression model.
z Field Spec Pro; Analytical Spectral Devices, Inc., Boulder, CO

















80


70-


60-


S 50 -



E 40
ooo

30-






20




Untreated 0.05 0.1 0.2 0.4

Rate (kg a.i. ha l)



Figure 3-1. Biomass measurements collected from fairway height hybrid bermudagrass (Cynodon dactylon X C. transvaalensis
'Princess 77) treated with trinexapac-ethyl at 0.05, 0.1, 0.2, and 0.4 kg a.i. ha '. Plots measured 1.5 X 3.0 m with three
replications. Means for biomass measurements was computed using Fishers protected LSD (p< 0.05) from data over 5
dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined.

















70


60


50




00 .P
F9 30


20


10



Untreated 0.3 0.6 1.1 2.3
Rate (kg a.i. ha l)


Figure 3-2. Biomass measurements collected from fairway height hybrid bermudagrass (Cynodon dactylon X C. transvaalensis~t~t~t~t~t~t~
'Princess 77) treated with flurprimidol at 0.3, 0.6, 1.1, and 2.3 kg a.i. ha l. Plots measured 1.5 X 3.0 m with three
replications. Means for biomass measurements was computed using Fishers protected LSD (p< 0.05) from data over 5
dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined.









CHAPTER 4
CONCLUSIONS

Currently many turf managers and researchers still use time and labor intensive techniques

to manage and assess turf that originate from decades ago. Many of these management practices

have been proven repeatedly to work in a variety of situations to assess turfgrass stress, but may

be time consuming and inconsistent. However, there are new developments in ways of assessing

and mapping stress that increase the efficiency by which one manages and assesses turf. With

increasing water and other environmental restrictions turf managers and researchers need be

aware of technological advances in their industry.

The research discussed in this thesis gives greater insight into the use of some the

technological advancements available to turfgrass managers demonstrating the sensitivity of

remote sensing instrumentation to various parameters of turfgrass stress. In conducting these

experiments many realizations were made about the usability and future potential for remote

sensing technology in the turfgrass management field. Some recommendations for anyone who

wishes to continue the research referred to in this thesis are as follows:

Irrigation X Nitrogen study. Conduct the experiment in the earlier part of the growing

season, when rainfall is not so prevalent. This will potentially reduce uniformity in soil VWC.

Also, apply N every two weeks as described by Kruse et al. (2005); this will continually promote

differences in growth. In addition, investigate more into the use of PLS regression on

hyperspectral data to formulate prediction models for N and soil VWC levels.

PGR study. Conduct the experiment on turfgrass that does not produce seedheads. In this

experiment seed heads could have potentially skewed remotely sensed data; this will at least take

that factor out. Also, obtain remote sensing data before and after PGR applications to assess the

differences they create. From this data possibly try to make applications based on remotely










sensed data and not just on an incremental scale. In addition, investigate more into the use of

PLS regression on hyperspectral data to formulate prediction models for assessment of the

biomass of the turf.

Crop circle. Test in various light and ambient conditions to assess different turfgrass

parameters and investigate its usability in an assortment of environmental settings. Turfgrass

managers work in all conditions, rain or shine. An assessment of how these devices will work in

varying environmental conditions must be evaluated before they are employed into real world

applications of turfgrass management.

Beyond these experiments. Conduct experiments on larger scale while adding other

technologies such as GPS, GIS, and VRA. This will allow researchers to better understand the

usability of the technology and the logistics of operating it as a turfgrass manager. Remote

sensing in a large scale setting is not very effective without the combination of other

technologies. You need GPS and GIS to georeference remotely sensed data to understand where

the data is coming from. Then by applying VRA technology some sort of economic impact can

be assessed to understand the potential benefit of the use all of these technologies. Remote

sensing is just the beginning of a long road of investigation in helping turfgrass managers reduce

irrigation and chemical use and apply these inputs only where needed.

Currently the use of remote sensing technology in the turfgrass management industry is

still in its infancy. However, the experiments described in this thesis along with other research

that has been referenced show it has potential to assist turfgrass managers to make better

decisions about managing their turf. Presently, this technology is not in a very usable form and

much research still needs to be conducted to understand its place in the field of turfgrass

management.









APPENDIX A
EXPERIMENTAL LAYOUT FOR DETECTION OF LEAF NITROGEN CONCENTRATION
AND SOIL VOLUMETRIC WATER CONTENT USING GROUND-BASED REMOTE
SENSING TECHNOLOGY











101
102
103
104


201
202
203
204


301
302
303
304


401
402
403
404


Figure A-1. Experimental layout for detection of leaf nitrogen concentration and soil volumetric water content using ground-based
remote sensing technology









Table A-1. Description of treatments for Figure A-1


Whole Plot
Irrigation
60% ET
60% ET
60% ET
60% ET
80% ET
80% ET
80% ET
80% ET
100% ET
100% ET
100% ET
100% ET
120% ET
120% ET
120% ET
120% ET


Subplot
Nitrogen
0 kg ha
25 kg ha l
50 kg ha l
100 kg ha l
0 kg ha-
25 kg ha l
50 kg ha l
100 kg ha l
0 kg ha-
25 kg ha l
50 kg ha l
100 kg ha l
0 kg ha-
25 kg ha l
50 kg ha l
100 kg ha l


Rep
2
704
702
703
701
802
803
804
801
503
502
501
504
602
601
603
604


Treatment #
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16


1
101
102
103
104
201
202
203
204
301
302
303
304
401
402
403
404


3
1201
1203
1202
1204
901
903
904
902
1003
1002
1004
1001
1101
1104
1102
1103










APPENDIX B
DAILY RAINFALL AND ET DATA FOR JAY, FL


Table B-1. Daily Rainfall and ET data for Jay, FL


Cumulative
Difference
(cm)
(0.10)
(0.31)
(0.18)
1.57
1.24
3.30
4.09
3.50
2.97
2.51
2.21
1.85
1.60
1.24
1.04
0.61
0.45
(0.08)
(0.51)
(0.82)
(0.26)
0.35
0.33
(0.16)
0.17
(0.36)
(0.84)
(1.32)
(0.92)
(1.37)
(1.71)
0.78
0.56
0.76
0.28
(0.18)
(0.64)
0.43


Rainfall
(cm>a
0.38
0.00
0.51
2.29
0.00
2.36
1.27
0.00
0.00
0.00
0.00
0.00
0.10
0.00
0.25
0.03
0.38
0.00
0.00
0.00
0.89
0.94
0.46
0.00
0.76
0.00
0.00
0.00
0.89
0.00
0.00
2.79
0.10
0.71
0.00
0.00
0.00
1.40


Irrigation
Applied (cm)
0.20b
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.20b
0.00


ET (cm)a
0.48
0.41
0.38
0.53
0.33
0.30
0.48
0.58
0.53
0.46
0.30
0.36
0.36
0.36
0.46
0.46
0.53
0.53
0.43
0.30
0.33
0.33
0.48
0.48
0.43
0.53
0.48
0.48
0.48
0.46
0.33
0.30
0.33
0.51
0.48
0.46
0.46
0.53


Day
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21


June


July










Table B-1. Continued


Cumulative
Difference
(cm)
0.15
3.07
2.89
2.53
2.05
1.54
1.09
3.22
2.79
5.07
4.64
4.34
3.88
3.42
2.99
2.51
2.08
1.72
1.64
1.29
1.87
1.44
0.96
0.50
0.04
(0.39)
(0.49)
(0.92)
(1.40)
(1.89)
(2.39)
(2.78)
(3.16)
(3.56)
2.99
2.94
3.63
3.52
3.12
2.91
2.64


Rainfall
(cm>a
0.53
3.35
0.28
0.00
0.00
0.00
0.00
2.54
0.00
2.62
0.00
0.15
0.00
0.00
0.00
0.00
0.00
0.00
0.41
0.03
0.84
0.00
0.00
0.00
0.00
0.00
0.33
0.00
0.00
0.00
0.00
0.00
0.00
0.05
4.34
0.28
0.89
0.33
0.00
0.08
0.05


Irrigation
Applied (cm)
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
2.54
0.00
0.00
0.00
0.00
0.00
0.00
0.00


Day
23
24
25
26
27
28
29
30
31
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
1


ET (cm)a
0.41
0.43
0.46
0.36
0.48
0.51
0.46
0.41
0.43
0.33
0.43
0.46
0.46
0.46
0.43
0.48
0.43
0.36
0.48
0.38
0.25
0.43
0.48
0.46
0.46
0.43
0.43
0.43
0.48
0.48
0.51
0.38
0.38
0.46
0.33
0.33
0.20
0.43
0.41
0.28
0.33


August





































September










Table B-1. Continued


Cumulative
Difference
(cm)
4.64
5.56


Rainfall
Day (cm>a ET (
3 2.34 0.
4 1.32 0.
aBased on data from http://fawn.ifas.ufl .edu/.
SIrrigation applied to water in fertilizer application


Irrigation
Applied (cm)
0.00
0.00


(cm~a
28
41










APPENDIX C
CUMALITIVE WEEKLY RAINFALL AND ET DATA FOR JAY, FL

Table C-1. Cumalitive weekly rainfall and et data for Jay, FI
Cumulative


Week
Ending On Rainfall (cm)a ET (cm)a
June 15 0.38 0.89
22 6.43 3.15
29 0.38 2.74
July 6 2.67 2.95
20 3.61 2.87
27 5.56 3.07
August 3 5.31 3.02
10 0.41 3.10
17 0.86 2.90
24 0.33 3.10
31 5.97 2.44
aBased on data from http://fawn.ifas.ufl .edu/.
bIrrigation applied to water in fertilizer application


Irrigation
Applied (cm)
0.00
0.00
0.00
0.00
0.20b
0.00
0.00
0.00
0.00
2.54
0.00


Difference (cm)
(0.31)
2.97
0.61
0.33
(0.64)
2.05
4.34
1.64
(0.39)
(3.16)
2.91
















ET (cm)
4.92125
5.5118
8.46455
10.4775
12.5984
12.192
12.5984
12.2047
10.0965
7.28345
4.7625
3.34645


Rainfall-ET (cm)
2.36855
8.1026
3.4544
4.8768
-4.7879
11.89355
11.71575
8.0391
5.9436
-1.59385
7.23265
7.3533


Based on data from http://fawn.ifas.ufl .edu/


January
February
March
April
May
June
July
August
September
October
November
December


APPENDIX D
MOTHLYRAINFALL AND ET AVERAGES FROM 2003-2006 FOR JAY, FL


Table D-1.


MothlyRainfall and ET averages from 2003-2006 for Jay, FL


Rainfall
(cm)
7.2898
13.6144
11.91895
15.3543
7.8105
24.08555
24.31415
20.2438
16.0401
5.6896
11.99515
10.69975


Cumulative Total
Rainfall-ET (cm)
2.36855
10.47115
13.92555
18.80235
14.01445
25.908
37.62375
45.66285
51.60645
50.0126
57.24525
64.59855










LIST OF REFERENCES


Anonymous. 2007. Proxy (specimen label). Bayer Scientific, Triangle Park, NC. U.S. Patent
4,240,819. Date issued: 25 Apr. 2005.

Aston, A.R., and C.H.M. Van Bavel. 1972. Soil Surface Water Depletion and Leaf Temperature.
Agron. J. 64:368-373.

Augustin, B.J., and G.H. Snyder. 1984. Moisture Sensor Controlled Irrigation For Maintaining
Bermudagrass Turf. Agron. J. 76:848-850.

Bahrun, A., V.O. Mogensen, and C.R. Jensen. 2003. Water Stress Detection in Field Grown
Maize by Using Spectral Vegetation Index. Commun. Soil Sci. Plant Anal. 34:65-79.

Barret, J. 2003. Golf Course irrigation: Environmental Design and Management Practices. 413-
417. John Wiley & Sons, Inc., Hoboken, NJ. 413-417.

Bethel, M., T. Gress, S. White, J. Johnson, T. Sheely, B. Roberts, N. Gat, G. Scriven, G.
Hagglund, M. Paggi, and N. Groenenberg. 2003. Image-Based Variable Rate Plant Growth
Regulator Application in Cotton at Sheely Farms in California. 6-10 Jan. 2003. Beltwide
Cotton Conferences, Nashville, TN.

Bell, G.E., B.M. Howell, G.V. Johnson, W.R. Raun, M.L. Stone, and J.B. Solie. 2004. Optical
Sensing of Turfgrass Chlorophyll Content and Tissue Nitrogen. Hortic. Sci. 39(5): 1130-
1132.

Bell, G.E., D.L. Martin, R.M. Kuzmic, M.L. Stone, and J.B. Solie. 2000. Herbicide Tolerance of
Two Cold-Resistant Bermudagrass (Cynodon spp.) Cultivars Determined by Visual
Assessment and Vehicle Mounted Optical Sensing. Weed Tech. 14:635-641.

Bell, G.E., D.L. Martin, M.L. Stone, J.B. Solie, and G.V. Johnson. 2002a. Turf area Mapping
Using Vehicle Mounted Optical Sensors. Crop Sci. 42:648-651.

Bell, G.E., D.L. Martin, S.G. Wiese, D.D. Dobsom, M.W. Smith, M.L. Stone, and J.B. Solie.
2002b. Vehicle- Mounted Optical Sensing An Obj ective Means for Evaluating Turf
Quality. Crop Sci. 42:197-201.

Blackmer, T.M., J.S. Schepers, and G. E. Varvel. 1994. Light reflectance Compared with Other
Nitrogen Stress Measurements in Corn Leaves. Agron. J. 86: 934-938.

Blackmer, T.M., J.S. Schepers, G.E. and Varvel, E.A. Walter-Shea. 1996.Nitrogen Deficiency
Detection Using Reflected Shortwave Radiation From Irrigated Corn Canopies. Agron. J.
88: 1-5.

Branham, Bruce. "Plant Growth Regulators for Fine Turf Use."
http ://www.interactiveturf.com/tips/1 997_02.htm. 24 Oct. 2007.










Brown, K.W., R.L. Duble, and J.C. Thomas. 1977. Influence of Management and Season on Fate
of N Applied to Golf Greens. Agron. J. 69:667-671.

Ehrler, W.L., S.B. Idso, R.D. Jackson, and R.J. Reginato. 1978. Wheat canopy temperature:
Relation to Plant Water Potential. Agron. J. 70: 251-256.

Fenstermaker- Shaulis, L.K., A. Leskys, and D.A. Devitt. 1997. Utilization of Remotely Sensed
Data to Map and Evaluate Turfgrass Stress Associated with Drought. Journal of Turfgrass
Management. 2:65-81.

Geeske, Joel, J.A. Gamon, and C.B. Field. 1997. Production Effieiency in Sunflower: The Role
of Water and Nitrogen Stress. Remote Sens. Environ. 62:176-188.

Hopkins, W. G. 1999. Introduction to Plant Physiology. 2nd ed. John Wiley and Sons Inc. New
York, Ch. 22 The physiology of plants under stress. 453-457.

Horst, G.L, J.C. O'Toole, and K.L. Faver. 1989. Seasonal and Species Variation in Baseline
Functions for Determining Crop Water Stress Indices in Turfgrass. Crop Sci. 29:1227-
1232.

Hutto, K.C., R.L King, J.D. Byrd, and D.R. Shaw. 2006. Implementation of Hyper spectral
Radiometry in Irrigation Management of Creeping Bentgrass Putting Greens. Crop. Sci.
46:1564-1569.

Idso, S.B., R.J. Reginato, D.C Reicosky, and J.L. Hatfield. 1981. Determining Soil Induced Plant
Water Potential Depressions in Alfalfa by Means of Infrared Thermometry. Agron. J. 73:
826-830.

Ikemura, Y., and B. Leinauer. 2006. Remote Sensing Technology Detects Turfgrass Stress.
Golfdom, May, p.63-66.

Jalali-Farahani, H.R., A.D. Matthias, and D.C. Slack. 1993. Crop Water Stress Index Models for
Bermudgrass Turf: A Comparison. Agron. J. 85:1210-1217.

Jalali-Frahani, H.R., D.C. Slack, D.M. Kopec, A.D. Matthias, and P.W. Brown. 1994. Evaluation
of Resistances for Bermudagrass Turf Crop Water Stress Index Models. Agron. J. 86: 574-
581.

Kruse, J.K., N.E. Christians, and M.H. Chaplin. 2005. Remote Sensing of Phosphorus
Deficiencies in Agrostis Stolonifera. Inter. Turfgrass Soc. Res. J. 10:923-928.

Kruse, J.K., N.E. Christians, and M.H. Chaplin. 2006. Remote Sensing of Nitrogen Stress in
Creeping Bentgrass. Agron. J. 98:1640-1645.

Lee, Wonsuk, S.W. Searcy, and T. Kataoka. 1999. Assessing Nitrogen Stress in Corn Varieties
of Varying Color. 18-21 July. ASAE Annual International Meeting. Toronto, Ontario,
Canada.










Lillesand, T.M., and R.W. Keifer. 2000. Remote Sensing and Image Interpretation. 4th ed. John
Wiley & Sons, New York.

Martin, D.L, C.S. Throssel, and D.J. Wehner. 1994. Models for Predicting Lower Limit of The
Canopy-Air Temperature Difference of Two Cool Season Grasses. Crop Sci. 34: 192-198.

Min, M. and W. S. Lee. 2003. Spectral Based Nitrogen Sensing for Citrus. 27-30 July. ASAE
Annual Intemnational Meeting. Las Vegas, NV

Morgan, W.C., J. Letey, S.J. Richards, and N. Valoras. 1966. Physical Soil Amendments, Soil
Compaction, Irrigation, and Wetting Agents in Turfgrass Management Effects on
Compactibility, Water Infiltration Rates, Evapotranspiration, and Number oflIrrigations.
Agron. J. 58:525-527.

Morris, K.N. A Guide to NTEP Turfgrass Ratings. http://www.ntep. org/reports/ratings.htm. 14
Nov. 2007.

Mylavarapu, R.S. and E.D. Kennelly. UF/IFAS Extentsion Soil Testing Laboratory (ESTL)
Analytical Procedures and Training Manual .
http://edis.ifas.ufl .edu/pdffiles/S S/SS3 1200.pdf. 14 Dec. 2007.

Osbome, S.L., J.S. Schepers, D.D. Francis, and M.R. Schlemmer. 2002. Detection of Phosphorus
and Nitrogen Deficiencies in Corn Using Spectral Radiance Measurements. Agron. J.
94:1215-1221.

Ripple, W.J. 1986. Spectral Reflectance Relationships to Leaf Water Stress. Photogramm. Eng.
Rem. Sens. 52:1669-1675

Rouse, J.W., R.H. Haas, J.A. Schell, and D.W. Deering. 1973. Monitoring the vernal
advancement and retroradation (green wave effect) of Natural Vegetation. Prog. Rep. SRC
1978-1, Remote Sensing Center, Texas A&M Univ., College Station, 93p. (NTIS) No.
E73-106393.

Salisbury, F.B. and C.W. Ross. 1992. Plant Philology. 4th ed. Wadsworth Publ. Co. Belmont,
CA.

SAS Institute. 2002-2003. The SAS System for Windows. Release 9.1. SAS Inst., Cary, NC.

Slack, D.C., K.M. Geiser, K.W. Stange, and E.R. Allred. 1981. Irrigation Scheduling in Sub-
Humid Aareas with Infrared Thermometry. p. 116-124. In Proc. Irrigation Scheduling
Conf., Chicago, and IL. 14-15 Dec. 1981. ASAE Annual Intemnational Meeting, St Joseph,
MI.

Starks, P. J., D. Zhao, W.A. Phillips, and S.W. Coleman. 2006. Development of Canopy
Reflectance Algorithms for Real-Time Prediction of Bermudagrass Pasture Biomass and
Nutritive Values. Crop Sci. 46:927-934.










Stowell, L. and W. Gelernter. 2006. Sensing the Future. Golf Course Management. March p. 107-
110.

Throssel, C.S., R.N. Carrow, and G.A. Milliken. 1987. Canopy Temperature Based Irrigation
Scheduling Indices for Kentucky Bluegrass Turf. Crop Sci. 27: 126-131.

Tobias, R.D. 1997. An Introduction to Partial Least Squares Regression (Online). Available at
SAS institute http://ftp. sas.com/techsup/download/technote/ts509.pd 17 April 2007.

Trenholm, L. E., R.N. Carrow, and R.R. Duncan. 1999b. Relationship of Multispectral
Radiometry Data to Qualitative Data in Turfgrass Research. Crop Sci. 39:763-769.

Trenholm, L.E., R.R. Duncan, and R.N. Carrow. 1999a. Wear Tolerance, Shoot Performance,
and Spectral Reflectance of Seashore Paspalum and Bermudagrass. Crop Sci.
39:1147:1152.

Trenhholm, L.E., M.J. Schlossberg, G. Lee, and W. Parks. 2000. An Evaluation of Multispectral
Responses on Selected Turfgrass Species. Int. J. Remote Sensing. 21:709-721.

Turgeon. A.J. 2008. Turfgrass Management. 8th edition. Pearson Education Inc. Upper Saddle
River, NJ.

Unruh, J. and M.L. Elliot. 1999. Best Management Practices for Florida Golf Courses.
University of Florida., Gainesville.

VanBibber, L. 2006. Putting the Numbers to PGR's. http://grounds-
mag.com/chemicals/grounds~maintenance_puttignmespr/ 24 Oct. 2007.

Yoder, B. J. and R.E. Pettigree-Crosby. 1995. Predicting Nitrogen and Chlorophyll Content and
Concentrations From Reflectance Spectra (400-2500) at Leaf and Canopy Scales. Remote
Sens. Environ. 53:199-211.

Xiong, X, G.E. Bell, J.B. Solie, M.W. Smith, and B. Martin. 2007. Bermudagrass Seasonal
Responses to Nitrogen Fertilization and Irrigation Detected Using Optical Sensing. Crop
Sci. 47:1603-1610.









BIOGRAPHICAL SKETCH

Jason Hamilton Frank was born in 1982, in Deland, FL. He lived there until 2001when he

graduated from Deland high school and was accepted to the University of Central Florida. He

spent a year there as a business administration maj or before realizing he wanted to attend the

University of Florida. He transferred to Daytona Beach Community College in August of 2002

and received his A.A. degree in July of 2003. He then transferred to the University of Florida in

August of 2003 where he completed his B.S. in turfgrass science with a minor in business

administration in December of 2005. He then went on to graduate school, also at the University

of Florida, where he will graduate with his M. S. in horticulture sciences with a minor in

agricultural and biological engineering and interdisciplinary concentration in geographic

information sciences in May of 2008. Upon graduation Jason has accepted a 2nd assistant

superintendent position at Royal Poinciana Golf Club in Naples, FL.





PAGE 1

1 DETECTION OF TURFGRASS STRESS US ING GROUND BASED REMOTE SENSING By JASON HAMILTON FRANK A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2008

PAGE 2

2 2008 Jason Hamilton Frank

PAGE 3

3 ACKNOWLEDGMENTS I wish to acknowledge the guidance and directio n of Dr. J. Bryan Unruh for all his wisdom and support during the last two year s. It has been a learning expe rience like none other. Sincere appreciation also goes to all my committee members, Dr. Barry Brecke, Dr. Laurie Trenholm, Dr. Won Suk Daniel Lee, and Dr. Jasmeet Judge I appreciate their helping me whenever asked. Special thanks also go to all the current and past member s of the West Florida Research and Education Center: Dr. Ken Hutto, Raymon d Edwards, Rex Lawson, Chris Adkison, Jason Ging, Chase McKeithen, Phil Moon, and Dr. Darcy Partridge. I appreciate their help and support. This undertaking could not have been accomplished without them. Bryan Schwartz and Dr. Jason Dettman-Kruse thank you for all the assi stance in SAS: This paper would have never been written in time without your help. I appreciate the support and wis dom received from my parent s. Their guidance throughout my life has inspired me to better myself not only through hard work in my career but as person. I am grateful to the University of Florida for providing me with the most valuable resource I will ever have, an education and the inspiration to never stop learning and always strive for the best. I am a proud part of the Gator Nation. The time I have spent wi th the University of Florida will never be forgotten.

PAGE 4

4 TABLE OF CONTENTS page ACKNOWLEDGMENTS...............................................................................................................3 LIST OF TABLES................................................................................................................. ..........6 LIST OF FIGURES................................................................................................................ .......10 ABSTRACT....................................................................................................................... ............11 CHAPTER 1 LITERATURE REVIEW.......................................................................................................13 Introduction................................................................................................................... ..........13 Irrigation Importance.......................................................................................................... ....14 Nutrient Importance............................................................................................................ ....17 Plant Growth Regulator Importance.......................................................................................18 Remote Sensing................................................................................................................. .....18 Vegetative Indices...........................................................................................................19 Human Visual Ratings.....................................................................................................20 Irrigation..................................................................................................................... .....20 Nutrient....................................................................................................................... .....23 Biomass........................................................................................................................ ...24 PGRs........................................................................................................................... .....24 Statistical Methods..........................................................................................................25 Summary........................................................................................................................ .........26 2 DETECTION OF LEAF NITROG EN CONCENTRATION AND SOIL VOLUMETRIC WATER CONTENT USING GROUND-BASED REMOTE SENSING TECHNOLOGY...................................................................................................28 Introduction................................................................................................................... ..........28 Materials and Methods.......................................................................................................... .34 Reflectance Measurements..............................................................................................35 Statistical Analysis..........................................................................................................36 Results and Discussion......................................................................................................... ..37 Influence of Irrigation Treatments...................................................................................37 Nitrogen Effects...............................................................................................................37 Nitrogen X Date Interaction............................................................................................38 Tissue N Concentration vs. Reflectance..........................................................................39 VWC vs. Reflectance......................................................................................................39 Visual Ratings.................................................................................................................40 Biomass........................................................................................................................ ...41 Crop Circle vs. ASD........................................................................................................41 Conclusions.................................................................................................................... .........42

PAGE 5

5 3 USE OF GROUND-BASED REMOTE SENSING TECHNOLOGY TO ASSESS BIOMASS PRODUCTION IN RESPONSE TO PLANT GROWTH REGULATOR APPLICATIONS................................................................................................................... .62 Introduction................................................................................................................... ..........62 Materials and Methods.......................................................................................................... .66 Reflectance Measurements..............................................................................................67 Statistical Analysis..........................................................................................................67 Results and Discussion......................................................................................................... ..68 Biomass........................................................................................................................ ...68 Volumetric Water Content..............................................................................................69 Visual Ratings.................................................................................................................69 Color.........................................................................................................................69 Quality......................................................................................................................70 Density.....................................................................................................................70 Reflectance Indices vs. PGR Application rates...............................................................70 Crop Circle Device..........................................................................................................70 ASD Device.....................................................................................................................71 PLS Regression...............................................................................................................71 Conclusions.................................................................................................................... .........72 4 CONCLUSIONS....................................................................................................................82 APPENDIX A EXPERIMENTAL LAYOUT FOR D ETECTION OF LEAF NITROGEN CONCENTRATION AND SOIL VOLUMETRIC WATER CONTENT USING GROUND-BASED REMOTE SENSING TECHNOLOGY.................................................84 B DAILY RAINFALL AND ET DATA FOR JAY, FL............................................................87 C CUMALITIVE WEEKLY RAINFALL AND ET DATA FOR JAY, FL.............................90 D MOTHLYRAINFALL AND ET AVERAGES FROM 2003-2006 FOR JAY, FL...............91 LIST OF REFERENCES .......................................................................................................... ..92 BIOGRAPHICAL SKETCH.........................................................................................................96

PAGE 6

6 LIST OF TABLES Table page 2-1 Analysis of variance of soil volumetric wa ter content (VWC), percent leaf nitrogen tissue concentration (%N), visual ratings: co lor, quality, density (rated on NTEP 1-9 scale, where 9 is the best rating and 6 is acceptable), and biomass. Normalized difference vegetation index (NDVI), Leaf area index (LAI), 880 nm, and 650 nm reflectance were obtained from Crop Circle device. Crop water stress index (CWSI) was computed from IR thermometer readings. All data was collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rates (60, 80, 100, and 120% es timated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined.................................................................................................................44 2-2 Analysis of variance of individual reflectance bands from ASD reflectance hyperspectral data with a range of 350-2500 nm and spectral resolution of 1 nm collected from fairway he ight hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation ra tes (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined..................................................................45 2-3 Analysis of variance of average ra nge computations from 630 nm and 2080 2350 nm; NDVI LAI Stress1 and Stress2 indices; and 605 nm/515 nm, 915/975 nm, and 865/725 nm ratios computed from an Analytical Spec tral Device (ASD) with a range of 350-2500nm and spectral reso lution of 1 nm collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rates (60, 80, 100, and 120% es timated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined.................................................................................................................46 2-4 Evaluation of means of volumetric wa ter content (VWC), leaf nitrogen tissue concentration (%N), visual ra tings: color, quality, density (rated on NTEP 1-9 scale 9 is best and 6 is acceptable), and biomass (g) collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET va lues), four N rates (0, 25, 50, and 100 kg ha-1), and five dates (22 June, 12 Ju ly, 25 July, 22 Aug., and 4 Sept. 2007) combined....................................................................................................................... .....47 2-5 Evaluation of means of Normali zed difference vegetation index (NDVI ), Leaf area index (LAI ), 850 nm, and 650 nm reflectance from Crop Circle device and crop water stress index (CWSI) computed from IR thermometer readings collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, BurttDavy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1) and five dates (22 June 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined...............................................................................................48

PAGE 7

7 2-6 Evaluation of means of individual band reflectance from the ASD device with a range of 350-2500 nm and a spectral resolu tion of 1 nm collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rates (60, 80, 100, and 120% es timated ET values), four N rates (0, 25, 50, and 100 kg ha-1) and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined.................................................................................................................49 2-7 Evaluation of means of ASD reflectance readings of average range computations from 630 nm and 2080 nm, computed indices Normalized Difference Vegetation Index (NDVI y), Leaf Area Index (LAI) and Stress1 and Stress2 indices, and ratios 605/515 nm, 915/975 nm, 865/725 nm taken from an ASD device with a range of 350-2500 nm and a spectral resolu tion of 1 nm collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rates (60, 80, 100, and 120% es timated ET values), four N rates (0, 25, 50, and 100 kg ha-1) and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined.................................................................................................................50 2-8 Evaluation of means of NDVI and LAI com puted from Crop Circle instrument, and LAI computed from ASD collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis Burtt-Davy), four ir rigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined, which had N treatment X date interactions..................................................................................51 2-9 Evaluation of Means of leaf N concentr ation tissue (%N) coll ected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rates (60, 80, 100, and 120% es timated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined, which had N treatment X date interactions............................................52 2-10 Evaluation of means of individual re flectance wavelengths at 813 nm and 935 nm collected from fairway he ight hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation ra tes (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combin ed, which had N treatment X date interactions................................................................................................................... ......53 2-11 Evaluation of means of ratios 915 nm/ 975 nm and 865 nm/925 nm computed from analytical spectral device (ASD) with a range of 350-2500 nm and a spectral resolution of 1 nm, collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined, which had N treatment X date interactions.............................................................................................54

PAGE 8

8 2-12 Coefficient estimates (r2) of Crop Circle reflectance data Normalized Difference Vegetation Index (NDVI), Leaf Area Index (LAI), NIR, and Red, and crop water stress index (CWSI ) to soil volumetric water conten t (VWC) visual color, quality, density ratings collected from fa irway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and averaged over all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007)...............................55 2-13 Coefficients estimates (r2) of ASD reflectance readings of individual reflectance wavelengths taken from an Analytical sp ectral device (ASD) w ith a range of 3502500 nm and a spectral resolution of 1 nm to %N concentration leaf tissue, soil volumetric water content (VWC) visual colo r, quality, density ratings, and biomass production of fairway height hybrid bermudagrass ( Cynodon dactolon X C. transvaalensis, Burtt-Davy), four irrigation ra tes (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and averaged over all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007)............................................................56 214 Coefficients estimates (r2) of average range computations from 630 nm and 2080 nm, and computed indices Normalized difference vegetation index (NDVI), Leaf area index (LAI) and Stress1 and Stress2 indices also taken from an ASD with a range of 350-2500 nm and a spectral resolution of 1 nm to %N concentration leaf tissue, soil volumetric wa ter content (VWC) visu al color, quality, density ratings, and biomass production co llected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET va lues), four N rates (0, 25, 50, and 100 kg ha-1), and averaged over all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007).......................................................................................................................... ........57 2-15 Partial Least Squares regression on hypers pectral data from ASD device with a range of 350-2500 nm and a spectral resolution of 1 nm for prediction of soil volumetric water content (VWC), percent leaf nitrogen concentration (%N), biomass, and visual quality collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation ra tes (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), averaged over all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007)............................................................58 3-1 Evaluation of means of biomass (g), volum etric water content (VWC), visual ratings: color, quality, density (rated on NTEP 1-9 s cale 9 is best and 6 is acceptable), from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis Princess 77) over five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined....................................................................................................................... .....74 3-2 Evaluation of means of Normalized Di fference Vegetation Index (NDVI), Leaf Area Index (LAI) computed from Crop Circle and ASD reflectance data, as well as Stress1 and Stress2 indices computed from ASD reflectan ce data from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis Princess 77) over five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined.....................................75

PAGE 9

9 3-3 Coefficient estimates (r2) of Crop Circle reflectance data, Normalized Difference Vegetation Index (NDVI), Leaf Area I ndex (LAI), 880nm, and 650nm, and, ASD reflectance data computed into NDVI LAI Stress1, and Stress2 indices to volumetric water content (VWC) visual colo r quality, density ratings, and biomass production from fairway hei ght hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis Princess 77) averaged over five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007).....................................................................................................76 3-4 Coefficient estimates (r2) of ASD reflectance data at in dividual wavelengths, as well as range averages from 630-690nm a nd 2080nm to volumetric water content (VWC) visual color, quality, density rati ngs, and biomass production from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis Princess 77) averaged over five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007). All numbers in column headings ar e expressed nanometers (nm)...........................................77 3-5 Coefficient estimates (r2) of range averages from 630-690 nm and 2080 nm and ratios 605/515 nm, 915/975 nm, and 865/ 725 nm to volumetric water content (VWC) visual color, quality, density rati ngs, and biomass production from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis Princess 77) averaged over five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007). All numbers in column headings ar e expressed nanometers (nm)...........................................78 3-6 Partial Least Squares regression coeffi cients on hyperspectral data from ASD device to predict volumetric water content (VWC), biomass, and visual quality (rated on NTEP 1-9 scale 9 is best and 6 is accep table) in fairway height bermudagrass ( Cynodon dactylon X C. transvaalensis Princess 77) averaged over five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007)............................................................79 A-1 Description of treat ments for Figure A-1...........................................................................86 B-1 Daily Rainfall and ET data for Jay, FL..............................................................................87 C-1 Cumalitive weekly rainfall and et data for Jay, Fl.............................................................90 D-1 MothlyRainfall and ET averages from 2003-2006 for Jay, FL.........................................91

PAGE 10

10 LIST OF FIGURES Figure page 2-1 Partial Least Squares regression on hypers pectral data from ASD device with a range 0f 350-2500 nm and a spectral range of 1 nm for prediction of volumetric water content (VWC) collected from fair way height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), averaged over all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007).......................................59 2-2 Coefficient estimates (r2) of Crop Circle and ASD reflect ance data computed into a Normalized Difference Vegetation Index ( NDVI), to soil volumetric water content (VWC) collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation ra tes (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and averaged over all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007)............................................................60 2-3 Coefficient estimates (r2) of Crop Circle vs. ASD using reflectance data to compute a Normalized Difference Vegetation Index (NDVI) collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C transvaalensis Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estima ted ET values), four N rates (0, 25, 50, and 100 kg ha-1), and averaged over all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007).............................................................................................................. .61 3-1 Biomass measurements collected fr om fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis Princess 77) treated with trinexapac-ethyl at 0.05, 0.1, 0.2, and 0.4 kg a.i. ha-1. Plots measured 1.5 X 3.0 m with three replications. Means for biomass meas urements was computed using Fishers protected LSD (p< 0.05) from data over 5 dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined...............................................................................................80 3-2 Biomass measurements collected fr om fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis Princess 77) treated w ith flurprimidol at 0.3, 0.6, 1.1, and 2.3 kg a.i. ha-1. Plots measured 1.5 X 3.0 m with three replications. Means for biomass measurements was com puted using Fishers protected LSD (p< 0.05) from data over 5 dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined....................................................................................................................... .....81 A-1 Experimental layout for detection of l eaf nitrogen concentration and soil volumetric water content using ground-base d remote sensing technology..........................................85

PAGE 11

11 Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Master of Science DETECTION OF TURFGRASS STRESS US ING GROUND BASED REMOTE SENSING By Jason Hamilton Frank May 2008 Chair: J. Bryan Unruh Major: Horticulture Sciences Many turf managers and researchers still use time and labor inte nsive techniques to manage and assess turf that originate from decad es ago. Many of these management practices have been proven repeatedly to work in a variety of situations to assess turfgrass stress, but may be time consuming and inconsistent. However, th ere are new developments in ways of assessing and mapping stress that increase the efficiency by which one manages an d assesses turf. With increasing water and other environmental restrict ions turf managers and researchers need be aware of technological advan ces in their industry. Two field experiments were conducted at the Univ ersity of Florida, West Florida Research and Education Center near Jay, FL. The first wa s the detection of leaf nitrogen concentration and soil volumetric water content using remote sensing technology. Plots were arranged in a randomized complete block design and treatments we re arranged in a split plot factorial design with four N levels split across four irrigation regi mes with three replications per treatment. The second experiment used remote sensing tec hnology to assess biomass production created by applications of plant growth re gulators. Plots were arranged in a factorial design observing three different PGRs, each at four incremental levels w ith three replications pe r treatment. Spectral measurements were taken from two remote sensing devices.

PAGE 12

12 A ground based vehicle mounted optical sensor Crop Circ le (Model ACS-210) (Holland Scientific) which produces its own light to compute a normali zed difference vegetation index (NDVI) which is calculated as (880nm-650nm) / (880nm+650nm). A hand-held hyperspectral spectroradiometer (FieldSpec Pro; Analytical Spectral Devices, Inc., Boulder, CO) with a spectral range of 300-2500 nm was also used to colle ct reflectance data. A ll reflectance readings were taken in full sunlight between the hours of 1100 and 1400 Central Standard Time (CST) to minimize variance caused by solar radiation. Plots were ground-truthed using time domain reflectometry for soil moisture content (VWC) a nd clipping samples obtained from plots were dried, weighed for biomass, and tested for Total Kj eldahl N content. Visu al ratings were also taken. The results from the nitrogen and irrigation ex periment indicate re flectance data best modeled soil VWC, visual ratings, and some treatme nt effects from N. Highest correlations for VWC were achieved from PLS regression on ASD data (r2=0.71) and the Crop Circle device (r2=0.70). The results from the PGR experiment modeled the biomass reduction expected from increasing rates of flurprimidol and trinexapac-ethyl compared to the untreated control, but not ethephon. Similar trends were observed in NDVI and leaf area index (LAI) indices computed from Crop Circle and ASD instruments, as well as several stress indices computed from the ASD device. If remote sensing technology could be used to adequately assess turf parameters, and was coupled with the global positioning system (GPS ) and geographic information systems (GIS) technologies, irrigations and chemi cal applications could be applie d on a site specific basis. In this setting, turf managers could potentially reduce inputs thereby reducing cost and environmental impact.

PAGE 13

13 CHAPTER 1 LITERATURE REVIEW Introduction The majority of turf managers and researchers still use outdated time and labor intensive techniques to manage and assess turf. Many of these management practices have been repeatedly proven to work in a variety of situat ions to assess turfgrass stress, but may be time consuming and inconsistent. For example, in the turfgrass research field visual evaluations can be labor intensive and variable among evaluators (Trenholm et al., 1999). However, there are new developments in ways of assessing and mapp ing stress that will increase the efficiency by which one manages and assesses turf. With increasing water restrictions and other environmental constraints turf managers and researchers need be aw are of technological advances in their industry. R ecent environmental concerns, such as nutrient and pesticide leaching and runoff, have pressed turf managers to reduce inputs used for turf maintenance (Bell et al., 2002). There are a variety of new or improved met hods available to dete ct water and nutrient stresses before they are visible. The basi s of this technology relie s heavily on the global positioning system (GPS), geographic informa tion systems (GIS), and remote sensing technology. Much of this technol ogy is in the beginning stages, however, the theory for use of this technology has been proven in other agronomic industries such as food production systems. This technology has been used as a management tool to maximize crop yield and minimize cost through site-specific application of inputs by applying nutrients, wa ter and pesticides only where needed (Lee et al., 1999).

PAGE 14

14 Irrigation Importance Water is the primary requirement for devel opment and survival of turfgrass and many other agronomic crops (Turgeon, 200 8). Plant water potential is essential for plant growth as well as most physiological processes the plant must perform (Idso et al., 1981). Therefore, water availability is one of the most critical factors in plant health (Ripple, 1986). Plants are made up of cells, which are filled with water to mainta in their turgor. Howeve r, if turgor is not maintained the plant begins to wilt (Unruh and Elliott, 1999). Many plants contain 75-85 % water and begin to die if this percentage decreases below 65 % (Unruh and Elliott, 1999). Water is taken up by the roots in the soil and is m oved throughout the plant and eventually released through the stomata through a pr ocess known as transpiration (Salisbury and Ross, 1992). Stomata are responsible for gas ex change in the plant and they close when the plant is under a water deficit or stress (Salisbury and Ross, 1992) This ultimately re duces photosynthesis and stunts leaf growth (Unruh and Elliott, 1999). G eeske et al. (1997) found that this stunting in growth is the plants defense mechanism to reduce nutrient and water requirements. The reduction in leaf area serves as a beneficial fact or that ultimately reduces transpiration and the need for increased physiological pr ocesses (Geeske et al., 1997). In addition, transpiration serves many differe nt purposes such as nutrient movement through the plant and ev aporative cooling, also known as evapotranspiration (ET) (Turgeon, 2008). Evapotranspiration is the total water lost by all transpiratory movement of water from soil, through the plant and ultimately into the atmosphere, where it evaporates (Turgeon, 2008). Many factors affect this process such as light intensity, humidity, wind ve locity and temperature (Unruh and Elliott, 1999). Water stress has a direct affect on the rate of photosynthesis. It first causes stomatal closure, which causes a reduced supply of CO2. Secondly, water stress reduces water potential,

PAGE 15

15 which causes a decrease in the structural integrity of many important elements of the photosynthetic process (Hopkins, 1999). This in tu rn will decrease electron transport and photophosphorolation leading to damage in the thylakoid membrane and adenine triphosphate (ATP) synthetase protein (Hopkins, 1999). The hormone abscisic acid (ABA) also plays an important role in the plants ability to endure wa ter stress. Under stress, ABA accumulates in the leaves causing an efflux of potassium from the guard cells, resulting in stomatal closure and reducing transpiration (Hopkins, 1999). Plant water stress can occur if a plant receives too much wa ter or if it does not receive enough. Flooding leads to a lack of oxygen, this l eads to decreased respiration, nutrient uptake, and root functionality, thus reducing photosynt hesis (Hopkins, 1999). Drought stress from lack of water leads to increased solute concentrati on in the protoplasm, which in turn affects many physiological processes (Hopkins, 1999). However, mild drought stress can cause the roots of a plant to grow deeper. This is why irrigation events should be spaced out as long as possible (Unruh and Elliott, 1999). Water is an important resource; however, it is not a renewa ble resource (Barret, 2003). Ninety-seven percent of all the earths water is salt water, three percent is fresh water; however, two-thirds of this is frozen in polar ice caps (B arret, 2003). Therefore, on e percent of the Earths fresh water must meet the water needs of manuf acturing, mining, agriculture, and turf not met by rainfall (Barret, 2003). In Florida the average rainfall does not meet the watering requirements to maintain golf courses at the expectations of players and industry viewers (Unruh and Elliott, 1999). As a result, the irrigation system is the si ngle most important tool available for turfgrass management to a golf course superintende nt or turf manager (Barret, 2003).

PAGE 16

16 This is why research focusing on improving wate r use management in agriculture is very important. It has both environmental and ec onomic consequences, especially as water availability becomes more limited. According to a survey conducted in 2000 by the Golf Course Superintendents Association of Am erica, the median annual water usage of a golf course in the United States is 32.49 cm applied to an average of 31.44 ha resulting in 108 million L used annually (Barret, 2003). Historically the majority of turfgrass mana gers scheduled irrigation usage based on their experience and set schedule of time intervals, both of which can result in over watering (Augustin and Snyder, 1984). Irrigation schedules are often based on calendar dates such as 3 or 7 times a week. Studies have shown that th is style of irrigation may promote over watering and provide too much moisture to the turf (U nruh and Elliott, 1999). This wastes water and energy, produces poor playing and agronomic c onditions, and leads to degradation of the environment from excessive runoff (Barret, 2003). Tools are needed to monitor turf water status to aid in decisions that result in more efficient irrigation (Bahrun et al., 2003). Irrigation systems should be operated so that the addition of water never exceeds that lost by ET in any given area (Brown et al., 1977). This will increase irrigati on efficiency and possibly reduce plant water stress in those areas. Researchers and turf managers over the past two decades have investigated many ways to determine when to irrigate. Many of the techni ques developed to assess plant water status are time consuming, spatially restrictiv e, and costly (Ripple, 1986). Currently, visual symptoms, ET rates, and tensiometers are all methods fo r determining irrigation needs (Turgeon, 2008). However, when visual symptoms of drought are apparent, it is alrea dy too late because the turf is already stressed. Likewise, ET values are gene rally not site specific and actual ET can differ

PAGE 17

17 greatly from one location to anothe r. It is assumed that all the turf within a given area is precisely the same and responds exactly the sa me to environmental conditions. Conversely, tensiometers have proven to be an adequate asse ssment of plant water status, but require periodic maintenance, and may have to be removed duri ng periods of cold weather (Unruh and Elliott, 1999). In addition, the tensiometer reading is only appropriate in the area adjacent to the placement of the ceramic tips and do not indicate water status for a large area (Unruh and Elliott, 1999). Therefore, because conservation of this limited resource is critical, it is increasingly necessary to develop technology to help turf managers allocate ir rigation water more efficiently. Nutrient Importance Aside from the importance of water to plants available nutrients al so drive growth and development. Nutrient availability affects ma ny physiological processes the plant must perform in order to develop properly. Nitrogen (N) is the most limiting nutri ent in production of nonlegumous crops (Osborne et al., 200 2). It is essential to the bu ilding of several important plant components including proteins, nucleic acids, hor mones, and chlorophyll, which harvests light energy used in photosynthesis (Hopkins, 1999). P hosphorus (P) plays an important role in the transfer of energy within the pl ant in the form of compounds like adenosine triphosphate (ATP), adenosine diphosphate (ADP), and phosphate (Pi). P also is an importa nt factor in root growth (Hopkins, 1999). Potassium (K) is also a very impor tant nutrient in plant growth and is required in large amounts. Potassium serves to activ ate a number of enzymes and regulates stomatal conductance; a very important pr ocess controlling tr anspiration (Hopkins, 1999). All of these nutrients, if in deficit of pl ant requirement, can cause a numbe r of physiological disruptions which are detrimental to plant growth and develo pment. Once these stresses become visible it may be too late to correct the damage. Therefor e, methods are needed to predict plant nutrient status that can detect physio logical stress before it becomes visible to the naked eye.

PAGE 18

18 Plant Growth Regulator Importance Plant Growth Regulators (PGRs) have beco me a very important tool in turfgrass management. Many turfgrass managers use PGRs to suppress vertical turf growth which ultimately reduces maintenance costs and improves turf quality. Some data suggests that PGRtreated turf can produce greater root mass, recover from injury faster, reduce water use rate, reduce disease incidence, and reduce Poa annua weed populations (Branham, 1997). A PGR is a substance that adjusts the growth an d development of a plant. This is generally achieved through the inhibition of gibberellic acid (GA), a hormone which is responsible for cell elongation. Plant growth regul ators were first introduced for use on fine turf in 1987 (VanBibber, 2006). The first of these chemicals were flurprimidol (Trade name Cutless) and paclobutrazol (Trade name Trimmit), which inhib it GA synthesis at relatively the same point in the GA biosynthesis pathway, resulting in very similar plant responses. However, paclobutrazol is the more active compound and uses lower rates to achieve the same respons e as higher rates of flurprimidol (Branham, 1997). The newest of the PGRs, trinexapac-ethyl (Trade name Primo), was released commercially in 1995 and has become the standard for use throughout the industry (Branham, 1997). It facilitates GA inhibition la ter in the GA biosynthesi s pathway so that GAs are formed but are not active and serves as prec ursors for plant biochemical processes. Plant growth regulators are relatively expensive compared to other chemicals used by turf managers. More efficient applications of PGRs could potentially reduce the application cost of these chemicals. Remote Sensing Remote sensing can be defined as obtaini ng information about an object, area, or phenomenon by analyzing data acquired by a device that is not in contact with that object, area, or phenomenon (Lillesand and Keifer, 2000). For ma ny years researchers have entertained the

PAGE 19

19 idea of measuring plant stress using remote sensing technology. Plant light interception influences growth and physiological respons es (Salisbury and Ross, 1992). When light is intercepted by the plant, it is absorbed, transmitted, or reflec ted (Turgeon, 2008). Therefore, by quantifying the reflected light, remote sensing technology can help to detect the onset of turfgrass stress (Ikemura and Leinauer, 2006). Presently, real world applications of remote sensing technology in turfgrass management are still in their infancy (Ikemu ra and Leinauer, 2006). Studies have shown that image analysis and various remote sensing devices have strong pote ntial to detect a variet y of turfgrass stresses (Ikemura and Leinauer, 2006). Spectral radiom etry (Trenholm et al., 2000), and infrared thermometry (Slack et al., 1981) have been show n to adequately assess light reflectance at various wavelengths. By differen tiating certain wavelength characteri stics, insights into growth and adaptive characteristics and how a plant responds to stress can be seen (Trenholm et al., 2000). There have been several aspects of remote sensing technology explored to reveal the level of stress that may be occurring in a plant. Researchers have created a variety of theories utilizing remote sensing techniques and exploring many part s of the electromagnetic spectrum to describe various plant stresses. Vegetative Indices Researchers have found certain relationships exist between vari ous aspects of the electromagnetic spectrum and a va riety of plant parameters and stress (Bell et al., 2000, 2002a, 2002b, 2004; Trenholm et al., 1999a, 1999b, 2000; Krus e et al., 2005; Fenstemaker-Shaulis et al., 1997; Bahrum et al., 2003; Xiong et al., 2007). These relationships are formulated into indices to normalize data for varying illumina tion conditions (Lillesand and Kiefer, 2003). One of the most common indices is the Normalized Difference Vegetation Index (NDVI) computed

PAGE 20

20 as: reflectance in the Near Infrared (NIR) regi on minus reflectance in the Red (R) region divided by a sum of both (NIR-R / NI R+R) (Rouse et al., 1973) Human Visual Ratings Traditionally, visual observation has been the standard for assessing turf stress for turf managers. Researchers quantified this observati on by creating a scale. The National Turfgrass Evaluation Program (NTEP) scale is a 1-9 assessment of turf color, quality and density, where 6 is the least acceptable and has been the standard for turf research for many years (Morris, 2007). However, recent investigations have looked beyond this scale to compare remote sensing systems with traditional human ev aluation. Trenholm et al. (1999) tested four of the current indices and compared them to visual evaluations. These indices were: NDVI computed as (935-661 nm)/ (935+661 nm) Leaf area index (LAI) computed as 935/661 nm Stress 1 computed as 706/760 nm Stress 2 computed as 706/813 nm They concluded that NDVI, LAI, and Stress 2 indices, and individual wavelength reflectance measurements at 661 nm and 813 nm were compar able to visual evaluations (Trenholm et al., 1999). Bell et al. (2002b) found strong correlations between NDVI and turf color, and subsequent work by Bell et al. (2004) concluded that NDVI was a better estimator of chlorophyll content than visual color evaluation. Irrigation Early research involving remote sensing to pr edict plant water status related plant canopy temperature to plant water potentia l (Ehrler et al., 1978). This wa s derived from the fact that transpiration plays a major role in the plants ab ility to cool itself through ET. As soil water decreases, evaporative cooling al so decreases causing the canopy te mperature to rise. This is what many researchers have used to relate ca nopy temperature to relative soil water content

PAGE 21

21 (Aston and Van-Bavel, 1972). Stom atal closure of sunlit leaves results in increased leaf temperature if factors such as wind speed and va por pressure remain c onstant (Ehrler et al., 1978). This temperature difference between canopy and air can be determined in several ways. Originally, microthermocouples we re used, but this only worked on a small scale. However, advances in infrared thermometr y have increased the ability to detect crop water stress through the utilization of remote sensing de vices (Slack et al., 1981). Infrar ed thermometers registered in a narrow bandpass (8-14 m or 10.5 12.5 m) are cap able of measuring crop temperatures with 0.5 C accuracy (Ehrler et al., 1978). They can be used for measuring canopy temperature at ground level, from aircraft, or satellite. Th ese methods are done in a much timelier manner compared to old methods such as tensiometers and microthermocouples (Ehrler et al., 1978). From this relation of canopy temperature to plant water potential, many indices have been developed to best predict plant water status. Idso et al. ( 1981) proposed a temperature-based index as a water stress indicator. This was in tended to normalize canopy temperature minus air temperature, Tc-Ta, for environmental variability. It was introduced as the Crop Water Stress Index (CWSI). CWSI produces a species-spec ific linear relationshi p between air and canopy temperature, which is irrespective of other envi ronmental factors except cloud cover (Slack et al., 1981). CWSI can be described as the fractiona l decrease of potential ET which can be presented in terms of the canopy temp erature minus the air temperature (Tc-Ta) or CWSI= (T c -T a ) a (T c -T a ) l (Tc-Ta)u (Tc-Ta)l Where Tc= canopy temperature, Ta= actual temperature, a= actual, u= upper limit, and l= lower limit (Jalali-Farahani et al., 1994) The CWSI was found to have pot ential in turf irrigation scheduling (T hrossel et al., 1987). Further investigations of this model spaw ned numerous theories accounting for several

PAGE 22

22 environmental factors as well. Horst et al. (1989) suggested that ca nopy temperature minus air temperature plotted as a function of Vapor Pressure Deficit (VPD), which is the difference in the amount of moisture in the air a nd the amount of moisture the air can hold. This may be species and mowing height dependent and may require further adjustment of the function for each species under individual environmen tal regimes (Horst et al., 1989). Jalali-Frahani et al. (1993) proposed that CWSI of turfgrass was not only a function of VPD but also net radiation from sunlight. Martin et al. (1994) found that usi ng variables accounting for VPD, wind speed, and net radiation increased coefficient of determination (r2) values dramatically across several varieties of turfgrass. Eh rler et al. (1978) concluded that temperature difference (Tc-Ta) in wheat ( Triticum spp ., durham Produra) is a reliable method for monitoring plant stress from aircraft or satellite. Frequent measurements were necessary to reduce the va riability in the data. Throssel et al. (1987) found that temperature difference (Tc-Ta) appeared to be indicative of water stress for well-watered Kentucky bluegrass ( Poa pratensis L.). More recent investigations into plant and soil water status rely on spectral reflectance data measured from plant or turfgrass canopies. Fe nstemaker-Shaulis et al (1997) found a negative correlation with NDVI and canopy temperature (r2=0.74) and a positive correlation to plant moisture content (r2=0.90). In addition, they concluded, that mapping NDVI values can also provide valuable insight into the status of irrigation system uniformity and management practices. Hutto et al. (2006) found that dr ought stressed species had a higher amount of reflectance in the short wave infrared regi on (SWIR) (1,300-2,500 nm) than that of the nonstressed control. Ripple (1986) found that reflectance in the 6 30 nm and 2,080-2,350 nm ranges is strongly correlated with leaf water content. As refl ectance decreased in these ranges there was an inverse linea r relationship with leaf relative wate r content. However, Ripple (1986)

PAGE 23

23 concluded that relationships betw een spectral reflectance and leaf wa ter content can be direct or indirect and are wavelength depe ndent. It is believed that elevated correlations in the 630 nm range with spectral reflectance data was due to a reduction in chlorophyll as the leaves dried and not a true relation to actual plan t water status (Ripple, 1986). Nutrient To a large extent, much remote sensing work has been done on the assessment of N and chlorophyll levels. Overall, re flectance is low in the visible range of the electromagnetic spectrum due to the high absorption by chlo rophyll for use in photosynthetic processes (Salisbury and Ross, 1992 ). For th is reason, if there is a particular reflectance in this region this could correlate to a certain physio logical disruption in the plant. It is generally assumed that there is a close relationship between plant colo r and chlorophyll conten t (Bell et al., 2004). Research has shown reflectance at 550 nm to be the most sensitive assessment of N in many agronomic crops, which is what the human eyes see as green (Blackmer et al., 1994, 1996, Yoder and Pettigree-Crosby, 1995). Bell et al. (2002a) found vehicle mounted optical sensors provided a good assessment of variable N applications Xiong et al. (2007) found significant seasonal responses to N treatments. For other nutrient deficiencies such as P and K, limited research has been conducted to establish their function in the sp ectral constituency. Research that has been conducted suggests that reflectance wavelengths for detecting P concentration in corn ( Zea mays L.) are in the NIR (730 nm and 930 nm) region of the spectrum as well as reflectance from the blue region (440 nm and 445 nm) (Osborne et al., 2002). This can possibly be attributed to the production of anthocyanin which is produced in P deficient plants that absorb more green light and reflect more blue light (Osborne et al., 2002). Res earch conducted by Kruse et al. (2005) found that spectral reflectance data can be used as a good estimation of P concentration in creeping

PAGE 24

24 bentgrass ( Agrostris stolinifera L.). Reflectance wavelengths fo r predicting P concentration of plant tissue were in the blue (480 nm), yellow (565 nm), orange (595 nm), and red (650 nm) regions with an overall r2 value of 0.73. This reflectance response relates to the spectral characteristic of anthocyanin which increased in concentration under P stress conditions (Kruse et al., 2005). Hutto et al. (2006) found that potassium deficiency could be detected at bandwidths from 750 nm to 785 nm. Biomass Green plant material has high reflectance in the NIR (700 nm) region, while dead or dying plant material is just the opposite (Gees ke et al., 1997). Yoder and Pettigree-Crosby (1995) related this correlation in the NIR region to biomass, or energy status of the plant. Similar results were found by Kruse et al. (2005) on creeping bentgrass, wh ere reflectance in the green (510 nm and 535 nm), red (635 nm), and NIR (735 nm) were significant in predicting biomass. Starks et al. (2006) found crude protein concentrat ion, biomass production, and crude protein availability of bermudagrass ( Cynodon dactylon L. Pers.) grown in pastures closely correlated with canopy reflectance ratios of 605/515 nm, 915/975 nm, and 865/725 nm. Conversely, Fenstemaker-Shaulis et al. (1997) found that NDVI is not a strong predictor of biomass (r2=0.37). Kruse et al. (2005) also observe d no correlation between biomass and NDVI. Kruse et al. (2006) found implementing Partial Leas t Squares regression on hyperspectral data to be an adequate predictor of biomass in cree ping bentgrass compared to indices NDVI, LAI, Stress1, and Stress2. PGRs Not a large body of literature exists in the ar ea of remote sensing and the effects of PGR applications. Some image-based appli cations have been tested in cotton ( Gosspypium spp. ) to reduce PGR use for economic reasons (Bethel et al., 2003). Among compar isons of different

PAGE 25

25 image based-application methods (Variable Rate and Site-Specific) to traditional blanket applications for PGRs in cotton, it was found that the broadcast applications were on average 19 % more costly than the site-spe cific method, and 27 % more costly than the variable rate method (Bethel et al., 2003). These results demonstrate th e potential economic benefit of this technology for future use in other aspects of agriculture to reduce excess chemical runoff and leaching into the surrounding environment (Bethel et al., 2003). Statistical Methods Older work involving remote sensing typically used multispectral data, which is collected in individual or wide bandwidt hs and give a coarse representation of the electromagnetic spectrum (Lillesand and Kiefer, 2003). With the introduction of hyperspectral data, which typically acquires 200 or more bands in very narr ow resolution throughout the visible, NIR, and SWIR regions, researchers have more data to analyze (Lillesand and Keifer, 1987). These enhanced techniques for obtaining th e data have created the need to explore different statistical methods to quantify and interpret the data. Tradit ionally linear regression is the simplest way to understand variance. However, in interpreting re flectance data it has been found that other regression techniques better assess plan t status (Kruse et al., 2005). Partial Least Squares (PLS) regression is a met hod used to predict variables when there are a large number of factors; typi cally more than the number of observations (Tobias, 1997). The use of multiple variables is generally tested through multiple linear regression (MLR) (Tobias, 1997). However, the problem with MLR is when the number of factors becomes too large the model may become over-fitting (Tobias, 1997). This is observed when a large number of variables are present but only a few account for mo st of the variation (T obias, 1997). Almost perfect models may be achieved; however they will not be highly predictive of new data.

PAGE 26

26 When prediction is the goal, then PLS regressi on is a highly useful tool (Tobias, 1997). Kruse et al. (2006) found PLS regression yielded a stronger relationship to predict N concentration compared to NDVI in plant leaf tissue indicating its pot ential use in future development models. Lee et al. (1999) also f ound that PLS and Princi pal Component Analysis (PCA) prediction models performed similar to each other and both were better than MLR regression to predict the N status in corn. Summary The use of remote sensing technology to pred ict turfgrass stress could decrease the time and labor required to evaluate tu rfgrass health (Osborne et al ., 2002). This would not only reduce the environmental impact of turfgrass manageme nt practices, but could reduce maintenance costs of the various inputs required by turfgrass managers. Curre nt dependence on agricultural chemicals have heightened many environmental c oncerns in Florida due to the sandy soils and heavy rainfall which increases potential for runoff and leaching of chemical s if applied in excess (Min and Lee, 2003). As stated earlier, real wo rld applications of remote sensing technology for utilization in turfgrass are sti ll in their infancy. At the present time, an experienced turf managers eye is still better than current co mputerized systems (Ikemura and Leinauer, 2006). However, studies have shown that various remote sensing devices have strong potential to detect a variety of turfgrass stresses (Ikemura and Leinauer, 2006). Applying remote sensing technology to the fi eld of turfgrass management could possibly help reduce irrigation and provide more efficient watering and fertil ization routines (Ikemura and Leinauer, 2006; Bell et al., 2004). Combined with GPS, GIS, and variable rate application technology, there is a potential to apply nutrients as well as irriga tion based on plant need rather than broadcast applications of entire areas (Bell et al., 2004). Currently, most of the remote sensing research is conducte d on agronomic food crops and forested areas. Only limited

PAGE 27

27 research is available in the area of turfgr ass management (Trenholm et al., 1999). Further research is needed to incorporate remote sens ing into turfgrass management (Trenholm et al., 1999). If conditions across the entire golf course we re uniform in microclimates, soil types, turf varieties, pest densities, nutrien t status, and irrigation performance there would be little need for advanced sensor systems to measure the variabilit y of these parameters to better meet the needs on a locational basis (Stowell and Gelerntner, 2006). However, that is not the case. Optimistically, Bell et al. (2002a) concludes If an optical sensing (In reference to remote sensing) system and software can be economi cally produced, reasonabl y priced, and mounted effectively on normal maintenance equipment, a turf practitioner could save enough money in fertilizers and pesticides to pay for the equipmen t. This approach would be useful for reducing the amount of fertilizers and pesticides needed to adequately manage large turf areas. The use of optical sensing to determine fertilizer rate be fore or during applica tion could increase turf uniformity and possibly, turf health. Sensor maps of large turf areas could be used to signal turf decline and provide an early warn ing system for turf mangers.

PAGE 28

28 CHAPTER 2 DETECTION OF LEAF NITROGEN CO NCENTRATION AND SOIL VOLUMETRIC WATER CONTENT USING GROUND-BA SED REMOTE SENSING TECHNOLOGY Introduction Remote sensing is defined as obtaining info rmation about an object, area, or phenomenon by analyzing data acquired by a device that is not in contact with that object, area, or phenomenon (Lillesand and Keifer, 1987). For many years researchers have entertained the idea of measuring plant stress using remote sensi ng technology. Plant light interception significantly influences growth and physiological responses (Salisbury and Ross, 1992). When light is intercepted by the plant it is ab sorbed, transmitted, or reflecte d (Salisbury and Ross, 1992). Using remote sensing technology to quantify the light that is reflected coul d help to detect the onset of turfgrass stress (Ikemura and Leinau er, 2006; Kruse et al., 2005; Hutto et al., 2006; Trenholm et al., 1999a). Real world applicatio ns of remote sensing technology for use in turfgrass management are still in their infancy; however, studies have shown that image analysis and various remote sensing devices have strong pote ntial to detect a variet y of turfgrass stresses (Ikemura and Leinauer, 2006). Recent investigations into remo te sensing instrumentation have produced promising results for turf managers. However, the majority of turf managers and researchers still use outdated time and labor intensive techniques to manage and assess turf. Noneth eless, there are new developments in ways of assessing and mapping plant stress that may increase the efficiency by which we manage and assess turf. The use of remote sensing t echnology to estimate turfgrass stress could significantly decrease the time and labor required to assess these levels in traditional ways, thus reducing cost as well (Osborne et al., 2002). The many great achievements in U.S. agricultural productivity in the past decades can be attributed to the use of agricultural chemicals, including fertilizer and pesticides (Lee et al.,

PAGE 29

29 1999). However, the increased dependence on pest icides and fertilizers have heightened many environmental concerns especially in Florida due to the sandy soils and heavy rainfall, which increases potential for heavy runoff and l eaching of chemicals (Min and Lee, 2003). With increasing water and other environmental restrict ions, turf managers and researchers must be aware of technological advan ces in their industry. Recent concerns have pressed turf managers to reduce nutrient and pesticide inputs used for turf maintenance (Bell et al ., 2002a). In other industries, such as agronomic food production crops, precision agriculture has be en used as a management tool to maximize yield and minimize cost (Bethel et al., 2003). The basis of this technology relies heavily on the Global Positioning System (GPS), Geographic Information Systems (GIS), and Variable-R ate Application (VRA) and remote sensing technology. Using this techno logy for site-specific management of nutrients, water, and pesticides could optimize yiel d while minimizing cost (Lee et al., 1999). Water is an important resource in turfgrass management; however, it is not a renewable resource (Barret, 2003). The Ea rths water reserves are 97 % salt water and 3 % fresh water. Of the fresh water reserve, two-thirds is frozen in polar ice caps (Barret, 2003). Therefore, only 1 % of the Earths water is available to meet the water needs of manufacturing, mining, agricultural, and turf not met by rainfall (Barre t, 2003). In addition, the averag e rainfall in Florida does not meet the watering requirements to maintain golf c ourse quality at the le vel expected by players and industry viewers (Unruh and El liot, 1999). As a result, the ir rigation system is the single most important tool available to a golf course su perintendent or turf ma nager (Barret, 2003). As water restrictions increase and pressure rises fr om environmental groups to restrict water use on golf courses, it is increasingly im perative that turfgrass managers improve water use efficiency (Martin et al., 1994).

PAGE 30

30 Aside from the importance of water to plants, available nutrients also drive plant growth and development. Nutrient availability affect s many of the physiological processes a plant must perform in order to develop properly. Nitrogen (N) is the most limiting nutrient in production of non-legumous crops (Osborne et al ., 2002). It is essential to th e building of several important plant structures including, prot eins, nucleic acids, hormones, a nd most importantly chlorophyll, which harvests light energy used in photosynthesis (Hopkins, 1999). Remote sensing technology has been shown to assess stress that may be occurring in a plant. Researchers have created a variety of theo ries and indices explorin g many parts of the light spectrum to describe various stresses. These stresses include water, nutrient, soil compaction, pests, and disease. One of the early theories involving remote se nsing to predict plant water status was the Crop Water Stress Index (CWSI), which corr elated plant water potential with canopy temperature (Ehrler et al., 1978). More recent investigations rely on spectral reflectance data measured from plant or turfgrass canopies. Several spectral reflectance bands have been identified for use in modeling various turfgrass stresses, especially thos e in the visible (400-700 nm) and near infrared (NIR) regions (700-1,300 nm ) of the spectrum, as well as short wave infrared (SWIR) regions (1,300-2,5 00 nm) (Hutto et al, 2006). Reflectance from healthy plant canopies is characteristically lo w in the visible range of the spectrum due to high absorption by chlorophyll for use in phot osynthetic processes (Trenholm et al., 2000), If there is a particularly high reflectance in this region, this could correlate to a physio logical disruption in the plant (Geeske et al., 1997). Green plant material has high reflectance in the NIR range while dead or dying plant material has low reflectance (Geeske et al., 1997). Yoder and Pettigree-Crosby (1995) relate this

PAGE 31

31 correlation in the NIR region to biomass, or the en ergy status of the plant. Similar results were found by Kruse et al. (2005) on creeping bentgrass ( Agrostis stolinifera L. ), where reflectance in the green (510 nm and 535 nm), red (635 nm), a nd NIR (735 nm) were significant in predicting biomass. Trenholm et al. (1999a) found that wear stress in several varieties of hybrid bermudagrass ( Cynodon dactylon x C. transvalensis Burtt-Davy) and seashore paspalum ( Paspalum vaginatum Swartz) is associated w ith the specific wavelengt h at 813 nm. Starke et al., (2006) found that crude protein concentr ation, biomass production, and crude protein availability of bermudagrass ( C. dactylon L. Pers.) pastures is closely related with canopy reflectance ratios of 605/515 nm, 915/975 nm, and 865/725 nm. Researchers have been exploring the idea that plant water status can be determined by measuring reflectance in the short wave infr ared (SWIR) region (1,300-2,500 nm). Ripple (1986) found that SWIR reflectance is strongly correlated with leaf water content and that significant correlations exist between certain spectral bands (630 nm and 2,080-2,350 nm ranges), with SWIR reflectance having the highe st correlations. As reflectan ce decreased in the 630-690 nm range there was an inverse linear relationship with leaf relative water content. However, Ripple (1986) concluded that relationships between spectral reflectance and leaf water content can be direct or indirect and ar e wavelength dependent. It is believed that elevated correlations with spectral reflectance data was due to a reduction in chlorophyll as the leaves dried and not a true relation to actual plant water st atus (Ripple, 1986). Hutto et al. (2006) also documented that drought stressed creeping bentgras s had a higher amount of reflectan ce in the SWIR regions than that of the non-stressed control. Several different spectral respons es to light to assess indicatio ns of plant stress, nutrient and water status have been tested in the form of indices. One of the most common indices is the

PAGE 32

32 Normalized Difference Vegetation Index (NDVI) which is computed as the reflectance in the NIR region minus the reflectance in the red region divided by a su m of both (NIR -R)/ (NIR+R). Bahrum et al. (2003) found that NDVI values from remotely sensed data can be used to assess turfgrass stress. Fenstemaker-Shaulis et al (1997) found a negative correlation between NDVI and canopy temperature (r2=0.74), and a positive correlation be tween NDVI and plant moisture content (r2=0.90). However, NDVI was not as strong a predictor of biomass (r2=0.37). In addition, spatial mapping of NDVI values can pr ovide valuable insight into the status of irrigation system uniformity and management pract ices (Fenstemaker-Shau lis et al., 1997). Bell et al. (2002b) found strong correlations between NDVI and turfgr ass color ratings. Additionally, the use of NDVI ratings obtaine d from vehicle mounted optical sensors were effective for measuring herbicide damage (Bell et al., 2000), variable N applica tions (Bell et al., 2002a), and significant seasonal responses to N treatments (Xiong et al., 2007). It has also been reported that NDVI is a better estimator of chlo rophyll content than visual colo r evaluation (Bell et al., 2004). Trenholm et al. (1999) tested four indices to detect stress in tu rfgrass. These indices were: NDVI computed as (935-661 nm)/ (935+661 nm) Leaf Area Index (LAI) computed as 935/661 nm Stress 1 computed as 706/760 nm Stress 2 computed as 706/813 nm They concluded NDVI, IR/R, and Stress 2 i ndices, as well as i ndividual wavelength measurements at 661 nm and 813 nm were well correlated to visual ratings. Older work involving remote sensing typically used multispectral data, which is collected in individual or wide bandwidt hs and gave a coarse repres entation of the electromagnetic spectrum (Lillesand and Kiefer, 2003). With the introduction of hyperspectral data, which typically acquires 200 or more wavelength band s in very narrow resolution throughout the visible, NIR, and MIR regions, researchers have more data to analy ze (Lillesand and Keifer,

PAGE 33

33 1987). These enhanced techniques for obtaining th e data have created the need to explore different statistical methods to quantify and interpret the large amount of data produced. Traditionally, linear regression is the simplest way to understand variance. However in interpreting reflectance data it has been found that other regre ssion techniques better assess plant status (Kruse et al., 2005). Partial Least Squares (PLS) regression is a met hod used to predict variables when there are a large number of factors; typi cally more than the number of observations (Tobias, 1997). The use of multiple variables is generally tested through Multiple Linear Regression (MLR) (Tobias, 1997). However, the limitation with MLR is when the number of factors becomes too large the model may become over-fitting (Tobias, 1997). This is observed when a large number of variables are present but only a few account for mo st of the variation (T obias, 1997). Almost perfect models may be achieved; however they will not be highly predictive of new data. When prediction is the objective then PLS re gression is a highly us eful tool (Tobias, 1997). Kruse et al. (2006) found PLS regression yielded a str onger relationship to predict N concentration compared to NDVI in creeping bent grass indicating its pote ntial use in future development models. Lee et al. (1999) also fo und that PLS and Princi pal Component Analysis (PCA) prediction models performed similar to each other and both were better than MLR regression to predict the N status in corn ( Zea mays ). Currently, the majority of remote sensing research is conducted on agronomic food crops and forested areas. Only limited research is ava ilable in the area of remote sensing and turfgrass management (Trenholm et al., 1999b). If conditions across the entire golf course were uniform in microclimates, soil types, turf varieties, pest densities, nutrient status, and irrigation performance there would be little need for advanced sensor systems to measure the variability of

PAGE 34

34 these parameters to better meet the needs on a locational basis (Stowell and Gelernter, 2006). However highly variable agronomic conditions create the need for further research to incorporate the use of remote sensing into turfgrass management (Trenhol m et al., 1999b). The objective of this research was to assess the ability of three remote sensing instruments to detect soil moisture and leaf nitrogen levels of Tifsport bermudagrass turf. Materials and Methods The field experiment was conducted at the Univ ersity of Florida, West Florida Research and Education Center near Jay, FL. Plots were arranged in a randomized complete block design and treatments were arranged in a split plot factorial design with four N levels (sub plot) split across four irrigation regimes (w hole plot) with three replicati ons per treatment (Appendix A). Whole plots were 4.88m x 4.88m in size and subp lots within the whol e plots were 1.22m x 4.88m in size. Whole plots were spaced 4.88m apar t on each side to minimi ze the impact of drift from irrigation from plot to plot. Ammonium sulfate was applied on 14 June 2007 and 20 July 2007 at 0, 25, 50, and 100 kg ha-1 with a Gandy drop spreader and th en watered in with 0.5 cm of supplemental irrigation. Irrigation was appl ied at 60%, 80%, 100%, and 120% of weekly estimated ET values obtained from an onsite we ather station that is part of the Florida Agricultural Weather Network (http://fawn.ifas.ufl.edu/ ) when ET was in deficit of 1.27cm or more. This occurred only on 24 August 2007 (Appendix B-D). Plots were mowed 2 to 3 times per week at a height of 1.2 cm with a Toro R eelmaster 3100-D. Data were collected five times (22 June, 12 July, 25 July, 22 Aug., and 4 Sept .) throughout the growi ng season in 2007. All measurements were taken within a one hour time period and collected between 1100-1300 hrs. Plots were rated visually for color, quality, and density based on the standard 1-9 National Turfgrass Evaluation Program (NTE P) rating scale where 9 is the highest possible rating, 6 is the lowest acceptable rating, and 1 indicates dead turf. Volumetric water content (VWC) readings

PAGE 35

35 were taken via Time Domain Reflectometry (T DR) with a FieldScout TDR 300 Soil Moisture meter (Spectrum Technologies, Inc ., Plainfield, IL). Clipping tissue samples were immediately weighed after collection and oven dried at 52 C for 7d before being weighed again for determination of biomass production. The oven dried tissue was ground and then analyzed for Total Kjeldahl N content using the RFA Me thod, number A303-S075-01 (Alpkem Corporation, 1990) as described by Mylavarapu and Kennelly (2007). Reflectance Measurements Spectral measurements were taken from three remote sensing devices. An infrared thermometer (Raytek Raynger ST, TTI, Inc.) wa s used to determine canopy temperature of the turf and to develop a CWSI calculated as: CWSI= (Tc-Ta) a (Tc-Ta) l / (Tc-Ta) u (Tc-Ta) l, Where Tc= Canopy Temperature, Ta= Actual Temperature, a= act ual, u= upper limit, and l= lower limit (Jalali-Farahani et al., 1994). A Crop Circle (Model ACS-210) (Holland Scien tific) was fitted onto a Toro Greensmaster 1000 walking greens mower 0.81m above the turf canopy which provided a 0.45 m2 field of view. The Crop Circle sensor produces its own light source at 650 nm and 880 nm and measures reflected values to produce an NDVI which is calculated as (880-650 nm)/ (880+650 nm). No calibration method was used for the sensor since it has the ability to de tects it own light source from incoming ambient sunlight. A hand-held hyperspectral spectroradiometer (F ieldSpec Pro; Analytical Spectral Devices, Inc., Boulder, CO) was also used to collect reflec tance data. This radiometer has a spectral range of 300-2,500 nm with a 1 nm resolution and a 23 foreoptic. Reflectance re adings were collected from shoulder height on each plot which provided a 0.65 m2 field of view. Radiance values are expressed in percent reflectance compared to st andardization with a white reference value across the entire spectral range. A white reference was used for calibration purposes at the beginning of

PAGE 36

36 data collection with the ASD unit, and again every 15 minutes depending on the length of time needed to collect data. Usually only one white re ference was needed to take data on all plots. All reflectance readings were taken in full sunlight be tween the hours of 1100 and 1400 Central Standard Time (CST) to minimize vari ance caused by incoming solar radiation. Several bands were extracted from ASD refl ectance data for comparison based on previous research. These bands consisted of wavelengths 550 nm (Blackmer et al., 1994; Blackmer et al., 1996; Yoder and Pettigree-Crosby, 1995); 510 nm 535 nm, 635 nm, and 735 nm (Kruse et al., 2005); 545 nm, 755 nm and 935 nm (Starks et al., 2006); and 661 nm and 813 nm (Trenholm et al., 1999a). Spectral ranges 630-690 nm and 2080-2350 nm were averaged (Ripple 1986) and ratios 605/515 nm, 915/975 nm, and 865/725 nm were calculated (Starks et al., 2006). Also NDVI ((880-650 nm)/ (880+650 nm)) and LAI (8 80/650 nm) indices were calculated for comparison against Crop Circle NDVI and LAI ca lculations, as well as Stress1 (706/760 nm) and Stress2 (706/813 nm) indices (T renholm et al., 2000). Statistical Analysis Analysis of variance (ANOVA) was perf ormed using the PROC GLM method (SAS Inst., 2003) to compare the differences among N a nd irrigation treatments at various dates to visual ratings, Crop Circle read ings, IR temperature readings calculated into a CWSI, various individual reflectance wavelengt hs quantified from the ASD unit, soil VWC, and leaf N concentration. Regression analysis was perf ormed using the PROC REG method (SAS Inst., 2003) to correlate relationships between visual ra tings, Crop Circle readi ngs, and IR temperature readings, various individual reflectance wavele ngths quantified from the ASD unit, soil VWC, and leaf nitrogen concentration. Additionally, du e to the large amount of data generated by the ASD device, ASD data were rendered using PROC PLS (SAS Inst., 2003) for prediction of soil VWC, leaf N concentration, vi sual quality ratings, and biom ass production. Equations were

PAGE 37

37 validated through a single sample cross-validatio n. For cross-validation, ten percent of the sample was omitted for prediction purposes so th at the number of factors chosen creates the minimal predicted residual sum of squares (PRESS) This process was repeated so that every observation was used exactly once for cross-validation. Results and Discussion Influence of Irrigation Treatments ANOVA (Tables 2-1, 2-2, 2-3) indi cated no irrigation treatment effects were significant. This is likely due to rainfall received during the experiment which eliminated differential soil VWC. As stated previously in the material s and methods, irrigation treatments were only applied once during the experiment due to the fa ct that the weekly cumulative ET calculations (Appendix C) based on daily rainfall and ET values (Appendix B) we re only in deficit of more than 1.27cm once during the experiment. In review of the average rainfall and ET data from the last 3 years, the experiment s hould be run earlier in the year when cumulative ET minus rainfall differences are not as great (Appendix D). Nitrogen Effects ANOVA indicated a large source of variation in bermudagrass response to N treatments. Leaf tissue N concentration, color, quality, NDVI and LAI from both optical sensing devices, as well as many individual reflectan ce wavelengths and reflectance ratios indicated a significant response to N treatments (Tables 21, 2-2, 2-3). This can be attri buted to the ability of N to increase overall turf health si nce it is the single most impor tant macronutrient to the plant (Hopkins, 1999). NDVI and LAI from the Crop Circ le device indicated a sampling date effect (Table 2-1). Additionally, reflectance at wavelength 813 nm, LAI, and 915/975 nm and 865/725 nm ratios quantified from the ASD device also indicated a sampling date effect (Tables 2-1, 2-2, 2-3).

PAGE 38

38 Research conducted by Xiong et al (2007) showed seasonal NDVI and LAI effects in response to N treatments. There was a significant improvement in visual color between no N applied and N applied at all levels across all dates (Table 2-4). Visual Quality rating (Table 2-4) and 650 nm reflectance from the Crop Circle device were bo th greater at the highe st N rate (100 kg ha-1) compared to no N applied (Table 2-5). Simila rly, reflectance wavelengths (510 nm, 535 nm, 545 nm, 550 nm, and range average 630-690 nm) (Tab le 2-6), ratios (605/515 nm, 915/975 nm, and 865/725 nm) (Table 2-7), and indices (NDVI, St ress1, and Stress2) derived from ASD data detected the high rates of applied N (Table 2-7). Nitrogen X Date Interaction It is apparent from the N X Date interaction means that there is a significant sequential response to increasing N treatments on percent leaf N concentration the first three sampling dates (22 June, 12 July, and 25 July 2007) (Tables 2-8, 2-9, 2-10, and 2-11). NDVI and LAI indices calculated from the Crop Circle device (Table 2-8) also show this response as well as LAI (Table 2-8), 813 nm and 935 nm (Table 2-10), and 915 /975 nm and 865/925 nm ratios (Table 2-11) calculated from the ASD reflectance data. This response in reflectan ce data can be attributed to the ability of the devices to detect overall incr eases in turf health due to a sequential vigor response in reaction to varying N treatments. This agrees with the work of Bell et al. (2002a, 2004). The last two sampling dates (22 Aug. a nd 4 Sept. 2007) do not show a response to increasing N treatments for any parameter measur ed that was significant by date (Tables 2-8, 29, 2-10, and 2-11). This is likely due to a waning N response, since N had not been applied for 33 and 46 days respectively for the final two sampling dates.

PAGE 39

39 Tissue N Concentration vs. Reflectance Reflectance values in this experiment did not prove to be a good predictor of leaf tissue N concentration in bermudagrass using linear or PLS regression to model percent N verses reflectance measurements However, some weak correlations exist (T ables 2-12, 2-13, 2-14, 215). This could potentially be due to the expe rimental design where N treatments were only applied twice, one month apart, causing the waning effect mentioned in the N X Date interaction section. Comparatively, research conducted by Kr use et al. (2005) and Xi ong et al. (2007) where N treatments were applied every two weeks promot ed a more consistent response to treatments across the growing season. VWC vs. Reflectance The highest r2 values for estimation of VWC ar e derived from PLS regression (r2= 0.72) calculated from reflectance data from the ASD device (Table 2-15) (F igure 2-1), and NDVI (r2= 0.71) derived from reflectance data from the Crop Ci rcle device (Table 2-12) (Figure 2-2). Also, LAI (r2= 0.59), NIR reflectance (r2= 0.61), and RED reflectance (r2= 0.52) derived from Crop Circle values provide adequate correlation as well (Table 2-12). This corresponds to vegetative indice values NDVI (r2= 0.57), LAI (r2= 0.55), Stress1 (r2= 0.61), Stress2 (r2= 0.64), and ratios 605/515 nm (r2= 0.66), 915/975 nm (r2= 0.68), 865/725 nm (r2= 0.44) derived from the ASD unit (Table 2-14). Some correlations at individual bands 755 nm (r2= 0.53), 813 (r2= 0.53), and 935 nm (r2= 0.49) were also notable (Table 2-13). The high r2 values for estimation of VWC with remo tely sensed data from Crop Circle and various individual bands, indices, and ratios from the ASD devi ce could be attributed to the extremely low VWC values record ed on 22 Aug 2007. On this date, cumulative ET calculations were in deficit 2.39 cm causing a significant decline in VWC. This resulted in a major decline in turf health. In previous research many indices such as NDVI do well in predicting turf stress in

PAGE 40

40 response to various plant parameters and not th e actual plant parameter (Trenholm et al., 1999a). Similarly, the ability of the remote sensing devices to model VWC through individual bands, indices and ratios, could potentially be attributed to decline in turf health in response to VWC and not a true predictor of actual VWC. Furthermore, the high r2 value achieved by PLS regressi on to model soil VWC (Table 215) (Figure 2-1) could potentia lly be highly predictive of new data and specific to actual VWC and not a response to plant stress. This is due to the re-sampling method employed by the PLS procedure mentioned in the materials and methods. Visual Ratings Visual ratings were reasonably predicted from reflectance data from both devices. Visual assessment of color, quality, and density was best predicted by vegetative indices NDVI and LAI from both instruments (Table 2-12, 2-14)). Furt hermore, raw reflectance values 650 nm and 880 nm from the Crop Circle device achieved reasonable r2 values (Table 2-12), as well as, 510 nm, 535 nm, 545 nm, 550 nm, 635 nm, and 661 nm individual reflectance wavelengths (Table 2-13), 630-690 nm and 2080-2350 nm range averages, and 605/515 nm, 915/975 nm, and 865/725 nm ratios (Table 2-14) from the ASD device. C onversely, quality was not as well modeled by PLS Regression (Table 2-15), compared to the vege tative indices, individua l reflectance wavelengths, and the ratios mentioned above. The ability of vegetative indices, NDVI and LAI, to model visual assessments could potentially be attributed to the fact that these indices were formulated to normalize the data. Since regression functions were run across all dates the normaliz ation of these indices could potentially smooth out variation in reflectance due to differences in light intensity on various evaluation dates.

PAGE 41

41 Biomass Biomass was not modeled well by reflectance da ta compared to other factors tested. However, some weak correlations exist (Tables 2-12, 2-13, 2-14, and 2-15). This is similar to research conducted by Kruse et al., 2005, 2006; Fenstemaker-Shaulis et al., 1997; and Trenholm et al., 1999b. The inability to effectively model biomass in this experiment could be attributed to the methodology used. Nitrogen treatments were only applied twice one month apart. In other research where biomass was adequately modele d (Kruse et al., 2005, 2006) N treatments were applied every 2 weeks thus continually promo ting greater differences in turf growth. Crop Circle vs. ASD The Crop Circle and ASD instruments are two ve ry different remote sensing devices that differ dramatically in price and use-ability. Th e ASD device is a highly scientific instrument with great ability to give a rese archer a plethora of reflectance data from a broad region of the electromagnetic spectrum. However, it is limite d in its ability because it requires full sunlight for use and only collects reflectance data when manually prompted by the user. Nonetheless, when reflectance data in this magnitude is analyz ed using the modern regression techniques, like PLS regression, the correlations could potentially be very specific to the parameter tested and highly predictive of new data. Therefore, pred iction models could likely be constructed for specific forms of turfgrass stress. Conversely, optical sensing devices like the Crop Circle instrume nt are less complicated to use but only give a limited amount of informati on typically in the visi ble and NIR regions. However, this data can be used to compute NDVI, LAI, and a variety of other indices to make the data more usable. Furthermore, they produce their own light s ource and, thus, can be mounted on a vehicle or mower for autonomous reflectance measurement collection.

PAGE 42

42 In this experiment VWC was best modeled by both remote sensing devices, Crop Circle and ASD. NDVI from the Crop Circle (r2=0.71) (Table 2-12) (Figure 2-2) device and PLS regression from the ASD device (r2=0.72) were the highest correla tions achieved (Table 2-15) (Figure 2-1). NDVI from the ASD device was close in comp arison to the Crop Circle (r2=0.57) to model VWC (Figure 2-2). Even though both of these device s utilize completely different sources of light they both calculate NDVI similarly and are close in comp arison to model VWC (Fi gure 2-3). Granted the prediction of VWC with PLS regr ession on ASD data could be highl y predictive of new data and specific to actual VWC and not just turf health, the Crop Circ le device shows there is some sensitivity to this parameter. Conclusions The results indicate reflectance data best m odeled soil VWC, visual ratings, and some treatment effects from N. Based on research by Ripple (1986), the elevated levels of prediction for VWC could be due to a reduction in chlorop hyll as the leaves dried in response to limiting soil VWC and not a true relation to actual soil water status. None theless, this research gives greater insight into the use of remote sensi ng technology and its place for use by turf managers and researchers demonstrating the sensitivity of remote sensing instrumentation to various parameters of turfgrass stress. Previous re search found optical sens ing measurements using NDVI could provide a fast and accurate alternativ e to traditional visual rating systems (Trenholm et al., 1999b; Bell et al., 2000) as well as assessing turfgra ss stress (Trenholm et. al., 1999a, Fenstemaker-Shaulis et al., 1997). Our results from two different remote sensing instruments are in agreement with this previous work. The Crop Circle instrument performed equa lly well with the ASD device in modeling stress attributed to reduced soil VWC (Figure 2-2) and was consistent with visual ratings in this

PAGE 43

43 experiment. It is simple to use and is poten tially non-restrictive of environmental conditions, like cloud cover and variation in so lar radiation. Currently, device s like this could provide fast and easy supplemental information to turfgrass managers for assistance in making decisions. Coupled with GPS and GIS technology, ground ba sed remote sensing devices, like the Crop Circle, can give turfgrass mana gers greater insight into mana ging their turf by potentially allowing them to more efficien tly schedule their irrigation, ther eby reducing water consumption, leaching, and runoff. Furthermore, if devices like the Crop Circ le could be improved to produce and quantify additional wavelengths these devices could potentia lly be employed by turfgrass managers due to their ease of use. However, further investigatio n into data from devices like the ASD using PLS regression is still needed to investigate the entire spectral constituency for optimal wavelengths to predict various turf parame ters. Only then can predicti on equations be formulated and incorporated into devices like the Crop Circle that could produce the correct wavelengths to calculate these prediction equations.

PAGE 44

44Table 2-1. Analysis of variance of soil volumetric water content (VWC), percent leaf nitrogen tissue concentration (%N), visual ratings: color, quality, density (rated on NTEP 1-9 scale, where 9 is the best rating and 6 is acceptable), and biomass. Normalized difference vegetation index (NDVI), Leaf area index (L AI), 880 nm, and 650 nm reflectance were obtained from Crop Circle device. Crop water stress i ndex (CWSI) was computed from IR thermometer readings. All data was collected from fairway he ight hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET valu es), four N rates (0, 25, 50, and 100 kg ha-1), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined. _____Visual Ratings______ _________Crop Circlez_________ __I Ry__ Source of Variation df VWC % N Dry W Color Quality Density NDVIx LAIw 880 nm 650 nm CWSIv Rep 2 0.0001 0.0044 0.0001 0.0102 0.0048 0.0015 0.0001 0.0001 0.0001 0.0001 0.2616 Date 4 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 Irrigation 3 0.3106 0.8398 0.3815 0.9382 0.5418 0.4083 0.2885 0.1709 0.0515 0.2266 0.7858 DxI 12 0.9981 0.9934 0.9401 0.9968 0.6717 0.9902 0.9921 0.9631 0.9970 0.9860 0.9896 error a 12 Nitrogen 3 0.1447 0.0001 0.1100 0.0001 0.0153 0.1087 0.0001 0.0001 0.2223 0.0001 0.5505 IxN 9 0.0004 0.0502 0.3567 0.6267 0.2050 0.0065 0.8093 0.8038 0.4815 0.3908 0.8610 DxN 12 0.2988 0.0002 0.4345 0.0595 0.7875 0.5126 0.0029 0.0001 0.0309 0.0612 0.7184 error b 36 Probability of greater F ratio (P>F) for Date, Irrigation, and N treatment components z Model ACS-210, Holland Scientific y IR thermometer Raytek Raynger ST, TTI, Inc. x NDVI= (880-650 nm)/(880+650 nm) w LAI= 880/650 nm v CWSI= (T c -T a ) a (T c -T a ) l / (T c -T a ) u (T c -T a ) l Where Tc= Canopy Temperature, Ta= Actual Temperature, a= actual, u= upper limit, and l= lower limit

PAGE 45

45 Table 2-2. Analysis of variance of indi vidual reflectance bands from ASD reflectan ce hyperspectral data with a range of 350-250 0 nm and spectral resolution of 1 nm collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rates (60, 80, 100, and 120% es timated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and five dates (22 June, 12 July, 25 Ju ly, 22 Aug., and 4 Sept. 2007) combined. _____________________________________ASDz (nm)___________________________________________ Source DF 510 535 545 550 635 661 735 755 813 935 2132 Rep 2 0.0429 0.0002 0.0001 0.0001 0.0596 0.1303 0.0001 0.0001 0.0001 0.0001 0.0584 Date 4 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 Irrigatio 3 0.2045 0.4189 0.4541 0.4650 0.2572 0.1908 0.0999 0.1148 0.1094 0.1134 0.1102 DxI 12 0.0826 0.2307 0.2667 0.2756 0.0948 0.0659 0.9956 0.9554 0.9698 0.9933 0.0774 error a 12 Nitroge 3 0.0029 0.0002 0.0001 0.0001 0.0039 0.0116 0.9039 0.0275 0.0053 0.0073 0.097 IxN 9 0.5871 0.3050 0.2683 0.2631 0.6254 0.6496 0.0874 0.4093 0.4387 0.4533 0.726 DxN 12 0.4677 0.2312 0.2048 0.2034 0.4701 0.5078 0.4558 0.7430 0.0222 0.1810 0.7521 error b 36 Probability of greater F ratio (P>F) for Date, Irrigation, and N treatment components z FieldSpec Pro; Analytical Spec tral Devices, Inc., Boulder, CO

PAGE 46

46 Table 2-3. Analysis of variance of average range computations from 630690 nm and 2080 nm; NDVI LAI Stress1 and Stress2 indices; and 605 nm/515 nm, 915/975 nm, and 865/725 nm ratio s computed from an Analytical Spectral Device (ASD) with a range of 350-2500nm and spectral resolution of 1 nm collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irriga tion rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined. ________________________________________ASDz____________________________________________ Source of Variation DF 630-690 nm 20802350 nm NDVI y LAI x Stress1 v Stress2 u 605/515 nm 915/975 nm 865/725 nm Rep 2 0.1209 0.0627 0.0313 0.1656 0.0099 0.0545 0.0206 0.0025 0.0018 Date 4 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 0.0001 Irrigation 3 0.2010 0.1034 0.1562 0.6883 0.1654 0.4132 0.8121 0.3246 0.6495 DxI 12 0.0712 0.0768 0.1737 0.9830 0.1296 0.4497 0.8643 0.8901 0.9094 error a 12 Nitrogen 3 0.0105 0.0869 0.0050 0.0001 0.0089 0.0019 0.0001 0.0001 0.0001 IxN 9 0.6521 0.7096 0.8275 0.9484 0.8028 0.9325 0.6612 0.9299 0.9001 DxN 12 0.5014 0.7276 0.3930 0.0008 0.2417 0.3474 0.1934 0.0299 0.0001 error b 36 Probability of greater F ratio (P>F) for Date, Irrigation, and N treatment components z FieldSpec Pro; Analytical Spec tral Devices, Inc., Boulder, CO y NDVI= (880-650 nm)/(880+650 nm) x LAI= 880/650 nm v Stress1= 706/760 nm u Stress2= 706/813 nm

PAGE 47

47 Table 2-4. Evaluation of means of volumetric water content (VWC), leaf nitrogen tissue concentrati on (%N), visual ratings: col or, quality, density (rated on NTEP 1-9 scal e 9 is best and 6 is acceptable), and bi omass (g) collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irriga tion rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined. _____________Visual Ratings______________ N Rate VWC % N Color Quality Density biomass (g) 0 kg ha-1 22.04 3.38 7.44b 7.12b 7.01 48.47 25 kg ha-1 22.07 3.61 7.73a 7.29ab 6.99 55.85 50 kg ha-1 21.87 3.62 7.76a 7.22ab 7.03 53.23 100 kg ha-1 22.48 3.99 7.85a 7.40a 7.12 57.05 LSD N.S. -0.20 0.21 N.S. N.S. "--" denotes LSD is not valid due to a significant date by N treatment interaction. Reporte d separately in Table 2-5 based on interaction. All values with a given number for the LSD denotes p < 0.05

PAGE 48

48 Table 2-5. Evaluation of means of Norm alized difference vegetation index (NDVI ), Leaf area index (LAI ), 850 nm, and 650 nm reflectance from Crop Circle device and crop water st ress index (CWSI) computed from IR thermometer readings collected from fairway he ight hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET valu es), four N rates (0, 25, 50, and 100 kg ha-1) and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined. ____________Crop Circle Reflectance Readingsz______________ IR Thermometery N Rate NDVIx LAIw880 nm 650 nm CWSIv 0 kg ha-1 0.59 4.35 1.38 .34a 0.44 25 kg ha-1 0.61 4.58 1.38 .33a 0.44 50 kg ha-1 0.60 4.61 1.38 .33a 0.45 100 kg ha-1 0.62 4.99 1.39 .31b 0.43 LSD --N.S. 0.01 N.S. "--" denotes LSD is not valid due to a significant date by N treatment interaction. Reported separately in Table 5 based on in teraction. All values with a given number for the LSD denotes p < 0.05 z Model ACS-210, Holland Scientific y IR thermometer Raytek Raynger ST, TTI, Inc. x NDVI= (880-650 nm)/(880+650 nm) w LAI= 880/650 nm v CWSI= (Tc-Ta)a (Tc-Ta)l / (Tc-Ta)u (Tc-Ta)l Where Tc= Canopy Temperature, Ta= Actual Temperature, a= actual, u= upper limit, and l= lower limit

PAGE 49

49 Table 2-6. Evaluation of means of individual band reflectance from the ASD device with a range of 350-2500 nm and a spectral resolution of 1 nm collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, BurttDavy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1) and five dates (22 June, 12 July, 25 July 22 Aug., and 4 Sept. 2007) combined. _____________________________________________ASD z (nm)_________________________________________ N Rate 510 535 545 550 635 661 735 755 813 935 2132 0 kg ha-1 0.054a 0.081a 0.086a 0.088a 0.074a 0.069ab 0.29 0.366a 0.407 0.437 0.183 25 kg ha-1 0.055a 0.081a 0.086a 0.088a 0.075a 0.071a 0.29 0.363a 0.404 0.433 0.190 50 kg ha-1 0.055a 0.081a 0.086a 0.088a 0.076a 0.072a 0.29 0.361a 0.402 0.431 0.194 100 kg ha-1 0.051b 0.077b 0.082b 0.084b 0.068b 0.064b 0.29 0.367a 0.409 0.437 0.178 LSD 0.002 0.002 0.002 0.002 0.005 0.005 N.S. 0.007 N.S. -N.S. "--" denotes LSD is not valid due to a significant date by N treatment interaction z FieldSpec Pro; Analytical Spec tral Devices, Inc., Boulder, CO

PAGE 50

50 Table 2-7. Evaluation of means of ASD reflectance readings of average range computations from 630 nm and 2080 nm, computed indices Normalized Difference Vegetation Index (NDVI y), Leaf Area Index (LAI) and Stress1 and Stress2 indices, and ratios 605/515 nm, 915/975 nm, 865/725 nm take n from an ASD device with a range of 350-2500 nm and a spectral resolution of 1 nm collected fr om fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rates (60, 80, 100, and 120% es timated ET values), four N rates (0, 25, 50, and 100 kg ha-1) and five dates (22 June, 12 July, 25 Ju ly, 22 Aug., and 4 Sept. 2007) combined. Range Average (ASD z) Vegetative Indices (ASD z) Ratios (ASD z) N Rate 630-690 nm 2080-2350 nmNDVI y LAI x Stress 1 v Stress 2 u 605/515 nm 915/975 nm 865/725 nm 0 kg ha-1 0.071a 0.175 0.715ab7.06 0.367ab 0.788ab 1.02a 0.99 1.88 25 kg ha-1 0.073a 0.182 0.707b 7.26 0.375a 0.794a 1.03a 0.99 1.89 50 kg ha-1 0.743a 0.186 0.700b 7.28 0.383a 0.791a 1.03a 0.99 1.90 100 kg ha-1 0.657b 0.170 0.729a 8.21 0.354b 0.779b 1.03b 1.00 1.97 LSD 0.0053 N.S. 0.018 -0.019 0.009 0.01 --"--" denotes LSD is not valid due to a significant date by N treatment interaction z FieldSpec Pro; Analytical Spec tral Devices, Inc., Boulder, CO y NDVI= (880-650 nm)/(880+650 nm) x LAI= 880/650 nm v Stress1= 706/760 nm u Stress2= 706/813 nm

PAGE 51

51 Table 2-8. Evaluation of means of NDVI and LAI computed from Crop Circle instrument, and LAI computed from ASD collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), an d five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined, which had N treatment X date interactions. NDVIz (Crop Circley) LAIx (Crop Circle) LAIx (ASDw) Date 0 kg ha-1 25 kg ha-1 50 kg ha-1 100 kg ha-1 0 kg ha-1 25 kg ha-1 50 kg ha-1 100 kg ha-1 0 kg ha-1 25 kg ha-1 50 kg ha-1 100 kg ha-1 6/22 0.70a3 0.73a23 0.73a2 0.75a1 5.79a3 6.34a23 6.58a2 7.34a1 10.20a23 11.11a23 11.65a2 13.08a1 7/12 0.67b3 0.69b2 0.69b2 0.71b1 5.10b3 5.49b2 5.50b2 5.99b1 8.28b2 8.59b2 8.73b12 9.60b1 7/25 0.66b3 0.69b2 0.70ab12 0.71b1 4.87b3 5.41b2 5.69b12 5.9567b1 8.25b3 9.29b2 9.32b2 10.31b1 8/22 0.53c1 0.42d12 0.39d2 0.53c1 2.58d1 2.45d12 2.96c12 2.41d2 3.66c1 3.39c1 3.19c1 3.41c1 9/4 0.44d1 0.52c1 0.49c1 0.52c1 3.39c1 3.20c1 2.34d1 3.28c1 4.89c1 3.93c1 3.54c1 4.64d1 Means in the same column and category fo llowed by the same letter are not significantly different (LSD; P, 0.05); means in the same row and category followed by the same number are not significantly different (LSD; P, 0.05). z NDVI= (880-650 nm)/(880+650 nm) y Model ACS-210, Holland Scientific x LAI= 880/650 nm w FieldSpec Pro; Analytical Spec tral Devices, Inc., Boulder, CO

PAGE 52

52 Table 2-9. Evaluation of Means of leaf N concentration tissue (%N) collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined, which had N treatment X date interactions. __________________________________ %N _________________________________________ Date 0 kg ha-1 25 kg ha-1 50 kg ha-1 100 kg ha-1 6/22 3.51b3 4.02a2 4.19a2 4.66a1 7/12 3.24bc3 3.25b2 3.38b2 3.81b1 7/25 3.26bc3 3.91a2 3.91a2 4.59a1 8/22 2.90c1 2.91b1 2.75c1 2.83c1 9/4 4.01a1 3.98a1 3.88a1 4.10b1 Means in the same column and category fo llowed by the same letter are not significantly different (LSD; P, 0.05); means in the same row and category followed by the same number are not significantly different (LSD; P, 0.05).

PAGE 53

53 Table 2-10. Evaluation of means of individua l reflectance wavelengths at 813 nm and 935 nm collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rate s (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and five dates (22 June, 12 Ju ly, 25 July, 22 Aug., and 4 Sept. 2007) combined, which had N treatment X date interactions. ___________________813 nm (ASDz)________________ __________________935 nm (ASDz)__________________ Date 0 kg ha-1 25 kg ha-1 50 kg ha-1 100 kg ha-1 0 kg ha-1 25 kg ha-1 50 kg ha-1 100 kg ha-1 6/22 0.47a1 0.47a1 0.48a1 0.49a1 0.49a1 0.49a1 0.50a1 0.51a1 7/12 0.43bc1 0.43b1 0.43b1 0.44b1 0.45b1 0.45c1 0.45c1 0.46b1 7/25 0.42c1 0.43b1 0.43b1 0.43b1 0.45b1 0.46bc1 0.46bc1 0.46b1 8/22 0.26d1 0.25c12 0.24c2 0.24c2 0.30c1 0.29d12 0.28d2 0.28c2 9/4 0.45ab1 0.44b1 0.43b1 0.45b1 0.49a1 0.48ab1 0.47b1 0.49a1 Means in the same column and category fo llowed by the same letter are not significantly different (LSD; P, 0.05); means in the same row and category followed by the same number are not significantly different (LSD; P, 0.05). z FieldSpec Pro; Analytical Spectr al Devices, Inc., Boulder, CO

PAGE 54

54 Table 2-11. Evaluation of means of ratios 915 nm/975 nm and 865 nm/925 nm computed from analytical spectral device (ASD) with a range of 350-2500 nm and a spectral resolution of 1 nm, collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined, which had N treatment X date interactions. 915/975 nm _(ASDz) 865/925 nm_(ASDz) Date 0 kg ha-1 25 kg ha-1 50 kg ha-1 100 kg ha-1 0 kg ha-1 25 kg ha-1 50 kg ha-1 100 kg ha-1 6/22 1.03a2 1.03a2 1.03a2 1.04a1 2.12a3 2.20a23 2.23a2 2.35a1 7/12 1.01b2 1.02b12 1.02b12 1.02b1 2.00b2 2.02b2 2.04b2 2.13b1 7/25 1.01b3 1.02b23 1.02b2 1.03b1 1.96b3 2.03b23 2.04b2 2.12b1 8/22 0.95d1 0.95d12 0.94d12 0.94d2 1.66c1 1.63c12 1.61c12 1.63c2 9/4 0.99c1 0.98c1 0.98c1 0.99c1 1.68c 1.60c1 1.57c1 1.63c1 Means in the same column and category fo llowed by the same letter are not significantly different (LSD; P, 0.05); means in the same row and category followed by the same number are not significantly different (LSD; P, 0.05). z FieldSpec Pro; Analytical Spectr al Devices, Inc., Boulder, CO

PAGE 55

55 Table 2-12. Coefficient estimates (r2) of Crop Circle reflectance data Normalized Difference Vegetation I ndex (NDVI), Leaf Area Index (LAI), NIR, and Red, and crop water stress index (CWSI ) to soil volumetric water content (VWC) visual color, quality, density ratings collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, BurttDavy), four irrigation rates (60, 80, 100, and 120% esti mated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and averaged over all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007). _________________________Crop Circlez___________________________ IR Thermometery Parameter NDVIx LAIw 880 nm 650 nm CWSIv %N 0.19*** 0.21*** 0.15*** 0.15*** 0.02* VWC 0.71*** 0.59*** 0.61*** 0.52*** 0.02* Color 0.59*** 0.57*** 0.39*** 0.61*** 0.13*** Quality 0.70*** 0.67*** 0.50*** 0.71*** 0.33*** Density 0.67*** 0.65*** 0.48*** 0.69*** 0.38*** Biomass 0.26*** 0.33*** 0.14*** 0.34*** 0.05** *, **, *** Significant at the 0.05, 0.01, and 0.001 probability levels, respectively z Model ACS-210, Holland Scientific y IR thermometer Raytek Raynger ST, TTI, Inc. x NDVI= (880-650 nm)/(880+650 nm) w LAI= 880/650 nm v CWSI= (Tc-Ta)a (Tc-Ta)l / (Tc-Ta)u (Tc-Ta)l Where Tc= Canopy Temperature, Ta= Actual Temperature, a= actual, u= upper limit, and l= lower limit

PAGE 56

56 Table 2-13. Coefficients estimates (r2) of ASD reflectance readings of individual reflectance wavelengt hs taken from an Analytical spectral device (ASD) with a range of 3502500 nm and a spectral resolution of 1 nm to %N concentration leaf tissue, soil volumetric water content (VWC) visual co lor, quality, density ratings, and biom ass production of fair way height hybrid bermudagrass ( Cynodon dactolon X C. transvaalensis, Burtt-Davy), four irrigation rate s (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and averaged over all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007). ASDz Reflectance (nm) Parameter 510 535 545 550 635 661 735 755 813 935 2132 %N 0.00 0.02* 0.02* 0.02* 0.01 0.02* 0.25*** 0.29*** 0.29*** 0.28*** 0.02* VWC 0.08*** 0.01 0.00 0.36 0.19*** 0.23*** 0.38*** 0.53*** 0.52*** 0.49*** 0.29*** Color 0.28*** 0.15*** 0.13*** 0.13*** 0.36*** 0.38*** 0.09*** 0.23*** 0.24*** 0.19*** 0.40*** Quality 0.52*** 0.32*** 0.31*** 0.31*** 0.60*** 0.62*** 0.06*** 0.22*** 0.23*** 0.17*** 0.69*** Density 0.55*** 0.37*** 0.34*** 0.35*** 0.62*** 0.65*** 0.17** 0.18*** 0.19*** 0.13*** 0.71*** Biomass 0.25*** 0.19*** 0.18*** 0.18*** 0.26*** 0.25*** 0.01 0.06*** 0.07*** 0.05** 0.25*** *, **, *** Significant at the 0.05, 0.01, and 0.001 probability levels, respectively z FieldSpec Pro; Analytical Spectr al Devices, Inc., Boulder, CO

PAGE 57

57 Table 214. Coefficients estimates (r2) of average range computations from 630 nm and 2080 nm, and computed indices Normalized difference vegetation index (NDV I), Leaf area index (LAI) and Stress1 a nd Stress2 indices also taken from an ASD with a range of 350-2500 nm and a spectra l resolution of 1 nm to %N concentra tion leaf tissue, so il volumetric water content (VWC) visual color, quality, density ratings, and biomass production co llected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rate s (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and averaged over all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007). Range Average (ASDz) Vegetative Indices (ASDz) Ratios (ASDz) Parameter 630-690 nm 2080-2350 nm NDVIy LAIx Stress 1w Stress 2v 605/515 nm 915/975 nm 865/725 nm %N 0.02 0.02* 0.13*** 0.20** 0.15*** 0.16*** 0.21*** 0.22*** 0.14*** VWC 0.22*** 0.29*** 0.57*** 0.53*** 0.61*** 0.64*** 0.66*** 0.68*** 0.44*** Color 0.38*** 0.40*** 0.53*** 0.55*** 0.53*** 0.55*** 0.45*** 0.52*** 0.51*** Quality 0.62*** 0.69*** 0.71*** 0.66*** 0.67*** 0.69*** 0.46*** 0.61*** 0.67*** Density 0.65*** 0.71*** 0.69*** 0.63*** 0.65*** 0.67*** 0.44*** 0.58*** 0.67*** Biomass (g) 0.26*** 0.25*** 0.26*** 0.35*** 0.24*** 0.27*** 0.15*** 0.24*** 0.67*** *, **, *** Significant at the 0.05, 0.01, and 0.001 probability levels, respectively z FieldSpec Pro; Analytical Spec tral Devices, Inc., Boulder, CO y NDVI= (880-650 nm)/(880+650 nm) x LAI= 880/650 nm w Stress1= 706/760 nm v Stress2= 706/813 nm

PAGE 58

58 Table 2-15. Partial Least Squares regressi on on hyperspectral data from ASD device with a range of 350-2500 nm and a spectral resolution of 1 nm for prediction of soil volumetric water co ntent (VWC), percent leaf n itrogen concentration (%N), biomass, and visual quality collected fr om fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rates (60, 80, 100, and 120% es timated ET values), four N rates (0, 25, 50, and 100 kg ha-1), averaged over all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007). Calibration No. of Factors* r2 SE n VWC 8 0.72 0.030 240 % N 7 0.25 0.043 240 Biomass (g) 6 0.31 0.040 240 Visual Quality 7 0.55 0.031 240 *Number of factors required to achieve a minimal Predicted Resi dual Sum of Squares (PRESS) of prediction for the partial least squares regression model. z FieldSpec Pro; Analytical Spec tral Devices, Inc., Boulder, CO

PAGE 59

59 ASD y = 0.7166x + 6.231 R2 = 0.7083 -5 0 5 10 15 20 25 30 35 40 45 0510152025303540 Actual VWCPredicted VWC Figure 2-1. Partial Least Squares regression on hyperspectral data from ASD device with a range 0f 350-2500 nm and a spectral r ange of 1 nm for prediction of volumetric water content (VWC ) collected from fairway he ight hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), averaged over all dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007).

PAGE 60

60 Crop Circle y = 65.19x 17.46 R2 = 0.71 ASD y = 52.74x 15.50 R2 = 0.57 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 0510152025303540 VWCNDVI Figure 2-2. Coefficient estimates (r2) of Crop Circle and ASD reflectance data com puted into a Normalized Difference Vegetation Index (NDVI), to soil volumetric water content (VWC) collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis, Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and averaged over all da tes (22 June, 12 July, 25 Ju ly, 22 Aug., and 4 Sept. 2007).

PAGE 61

61 y = 0.8478x + 0.0025 R2 = 0.88 0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1 00.10.20.30.40.50.60.70.80.91 ASDCrop Circle Figure 2-3. Coefficient estimates (r2) of Crop Circle vs. ASD using reflectance data to compute a Normalized Difference Vegetat ion Index (NDVI) collected from fair way height hybrid bermudagrass ( Cynodon dactylon X C transvaalensis Burtt-Davy), four irrigation rates (60, 80, 100, and 120% estimated ET values), four N rates (0, 25, 50, and 100 kg ha-1), and averaged over all dates (22 June, 12 Jul y, 25 July, 22 Aug., and 4 Sept. 2007).

PAGE 62

62 CHAPTER 3 USE OF GROUND-BASED REMOTE SENSIN G TECHNOLOGY TO ASSESS BIOMASS PRODUCTION IN RESPONSE TO PLANT GROWTH REGULATOR APPLICATIONS Introduction Remote sensing is defined as obtaining info rmation about an object, area, or phenomenon by analyzing data acquired by a device that is not in contact with that object, area, or phenomenon (Lillesand and Keifer, 1987). For many years researchers have entertained the idea of measuring various plant parameters using remo te sensing technology. Plant light interception significantly influences growth and physiologi cal responses (Salisbury and Ross, 1992). When light is intercepted by the plant it is absorbe d, transmitted, or reflected (Salisbury and Ross, 1992). Using remote sensing technology to quantify the light that is refl ected could help to detect the onset of turfgrass st ress (Ikemura and Leinauer, 2006). Real world applications of remote sensing technology for use in turfgrass ma nagement are still in their infancy; however, studies have shown that image analysis and variou s remote sensing devices have strong potential to detect a variety of turfgrass stresses (Ikemur a and Leinauer, 2006). Recent investigations into this technology show that there is potential for turfgrass managers to manage and assess turf more efficiently. The use of remote sensi ng technology to estimate turfgrass stress could significantly decrease the time and labor required to assess these levels in traditional ways, thus reducing cost as well (Osborne et al., 2002). The many great achievements in U.S. agricultural productivity can be attributed to the use of agricultural chemicals in cluding fertilizer and pesticides (Lee et al., 1999). However, the increased dependence on pesticides and fer tilizers have heightened many environmental concerns especially in Florida due to the sandy soils and heavy ra infall, which increases potential for heavy runoff and leaching of chemicals, if applied in excess (Min and Lee, 2003). With increasing water and environmental restrictions, turf managers and researchers must be aware of

PAGE 63

63 technological advances in their industry. Thes e environmental concerns have pressed turf managers to reduce nutrient and pesticide inputs used for turf maintenance (Bell et al., 2004). In other industries, such as agr onomic food production crops, precision agriculture has been used as a management tool to maximize yield and minimize cost (Bethel et al., 2003) The basis of this technology relies heavily on the Global Positi oning System (GPS), Geographic Information Systems (GIS), Variable Rate Application (V RA) and remote sensing technology. Using this technology to apply nutrients, wate r, and pesticides only where they are needed can optimize yield while minimizing cost. (Lee et al., 1999) Plant growth regulators (PGRs) are used by turf managers to s uppress vertical turf growth, which ultimately reduces maintenance costs due to reduced mowing requirments. A PGR is a substance which adjusts the growth and development of a plant. This is generally achieved through the inhibition of the gibberellic acid (GA), a horm one which is responsible for cell elongation. PGRs were first introduced for use on fine turf in 1987 (VanBibber, 2006). Some data suggests that PGR-treat ed turf can produce greater r oot mass, recover from injury faster, reduce water use rate, reduce disease incidence, and reduce Poa annua weed populations (Branham, 1997). The first of these chemicals were flurprimidol (Trade name Cutless) and paclobutrazol (Trade name Trimmet) which inhib it GA synthesis at relativ ely the same point in the GA biosynthesis pathway resulting in very similar plant responses (Branham, 1997). However, paclobutrazol is a more active compound and requires lower rates to achieve the same response as higher rates of flur primidol (Branham, 1997). The newe st of the PGRs, trinexapacethyl (TE) (Trade name Primo), became commer cially available in 199 5 and has become the standard for use throughout the turf industry (B ranham, 1997). TE facilitates GA inhibition later

PAGE 64

64 in the GA biosynthesis pathway so that GAs are formed but are not active and serve as precursors for plant biochemical processes. Historically PGRs are relatively expensive comp ared to other chemical applications made by turf managers. In an effort to reduce the co st of using PGRs, resear chers have investigated new methods to reduce the use rate. Research ers used remote sensi ng, the global positioning system (GPS), and geographic information system s (GIS), and variable rate application (VRA) technology to make site specific applications of PGRs. These image-based PGR applications were tested in cotton ( Gossypium sp.) to reduce the chemical i nput of PGRs while also reducing cost. Among comparisons of different image base d-application (Variable Ra te and Site-Specific) methods to traditional broadcast applications for PGRs in cott on, it was found that the broadcast applications were on average 19 % more costly than the site-specific applications, and 27 % more costly than applications using variable rate methods. Th is demonstrates the potential economic benefit of remote sensing t echnology for PGR use (Bethel, 2003). In the turf industry, research has been conducted to assess vari ous parameters of turfgrass stress using remotely sensed data. Several differe nt indices have been developed that have been correlated with plant stress. One of the most common indices to assess turfgrass stress is the Normalized Difference Vegetation Index (NDVI) whic h is computed as the reflectance from the Near Infrared (NIR) region minus the reflectance from the Red (R) region divided by a sum of both (NIR-R )/(NIR+R) (Rouse et al., 1973). Fenstemaker-Shaulis et al. (1997) found a negative correlation with NDVI and canopy temperature, (r2=0.74) and a positive correlation with NDVI and plant moisture content (r2=0.90). They also concluded that NDVI was not as a strong predictor of biomass (r2=0.37). Trenholm et al. (1999b) test ed various wavelengths 507 nm, 559 nm, 661 nm, 706 nm, 760 nm, 813 nm, and 935 nm as well as NDVI, IR/R (LAI) computed as

PAGE 65

65 935/661 nm, Stress1 computed as 706/760 nm, an d Stress2 computed as 706/813 nm. The results indicated that reflectance data was not successful in expl aining variability in the models for clipping yield. However, Kruse et al (2006) found conducting Pa rtial Least Squares regression on hyperspectral data (400,100nm) wa s an adequate predictor of biomass in creeping bentgrass ( Agrostis stolinifera L.) With the introduction of hyperspectral data, which typically measures more than 200 wavelengths, there has also been investigation of different statistical methods to quantify and interpret remotely sensed data. Traditionally, linear regression is the simplest way to understand variance. However, in interpreting reflectance data, it has been found that other regression techniques better assess plant stat us. Partial Least Squares Regre ssion (PLS) is a method used to predict variables when there are a large number of factors, typically more than the number of observations (Tobias, 1997). The use of multiple variables is generally tested through Multiple Linear Regression (MLR) (Tobias, 1997). Howe ver, the limitation with MLR is when the number of factors becomes too large the model may become over-fitting (Tobias, 1997). This is when it uses a large number of va riables but only a few account fo r most of the variation, also called latent factors (Tobias, 1997) Almost perfect models may be achieved; however they will not be highly predictive of new data. When pr ediction is the objective then PLS regression is a highly useful tool (Tobias, 1997). Currently, the majority of remote sensing research is done on agronomic food crops and forested areas. There is onl y limited research available in the area of remote sensing and turfgrass management (Trenholm et al., 1999b). If conditions across the entire golf course were uniform in microclimates, soil type s, turf varieties, pest densiti es, nutrient status, and irrigation performance there would be little need for advanced sensor systems to measure the variability of

PAGE 66

66 these parameters to better meet the needs on a locational basis (Stowell and Gelernter, 2006). However highly variable agronomic conditions create the need for further research to incorporate the use of remote se nsing into turfgrass management to reduce chemical inputs. (Trenholm et al., 1999b). The objective of this re search was to assess th e ability of remote sensing instruments to detect diffe rences in turf growth parameters as influenced by applications of PGRs on hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis Princess 77) turf. Materials and Methods The field experiment was conducted at the Univ ersity of Florida, West Florida Research and Education Center, Jay, FL. Plots measured 1.5 m x 3.0 m and treatments were arranged in a factorial design observing three different PGRs, each at four incremental levels, 1, 2, and 4 times the labeled rates, with three replications per treatment. Plant growth regulators were applied on 14 June 2007 and 20 July 2007. Plan t growth regulators treatments included trinexapac-ethyl applied at 0.05, 0.1, 0.2, and 0.4 kg a.i. ha-1, ethephon applied at 3.8, 7.6, 15.2, 30.4 kg a.i. ha-1, and flurprimidol applied at 0.3, 0.6, 1.1, and 2.3 kg a.i. ha-1. Plots were mowed 2 to 3 times per week at a height of 1.2 cm with a Toro Reelmast er 3100-D. Supplemental irrigation was applied as needed. Data was co llected five times throughout the growing season in 2007 on 22 June, 12 July, 25 July, 22 Aug., and 4 Sept. When data were collected all measurements were taken within a one hour time period. Plots were rated visually for color, qual ity, and density based on the standard 1-9 National Turfgrass Evaluation Prog ram (NTEP) rating scale where 9 is the highest and 6 is the lowest acceptable rating. Soil volumetric water content readings were taken via Time Domain Reflectometry with a Fieldscout TDR 300 soil mo isture meter (Spectrum Technologies, Inc.). Clipping samples for biomass determin ation were collected from a 4.3 m2 area immediately after reflectance data and visual evaluations were obtained. Clipping tissue samples were weighed

PAGE 67

67 and oven dried at 52 C for 7d before being weighed agai n for determination of biomass production. Reflectance Measurements Spectral measurements were taken from two remote sensing devices. A ground based vehicle mounted optical sensor was used to asse ss turf health. A Crop Circle (Model ACS-210) (Holland Scientific) wa s fitted onto a Toro Greensmaste r 1000 walking greens mower 0.81 m above ground which provided a 0.45 m2 field of view. The Crop Circ le sensor produces it own light source at 650 nm and 880 nm and measures re flected values to produce an NDVI which is calculated as (880-650 nm)/(880+650 nm). No calib ration method was used for the sensor since it has the ability to detect its own light source from incoming ambient sunlight. A hand-held hyperspectral spectroradiomete r (FieldSpec Pro; Analytical Spectral Devices, Inc., Boulder, CO) was also used to co llect reflectance data. This radiometer has a spectral range of 300-2,500 nm with a 1 nm resolution and a 23 foreoptic. Reflectance readings were collected from shoulder height on each plot which provided a 0.65 m2 field of view. Radiance values are expressed in percent reflect ance compared to standardization with a white reference value across the entire spectral range. A white reference was used for calibration purposes at the beginning of data collection with the ASD unit, and again every 15 minutes depending on the length of time needed to coll ect data. Usually only one white reference was needed to take data on all plots. All reflectance readings were taken in full sunlight between the hours of 1100 and 1400 Central Standard Time (C ST) to minimize variance caused by incoming solar radiation. Statistical Analysis Analysis of variance (ANOVA) was perf ormed using the PROC GLM method (SAS Inst., 2003) to compare the differences among PG R treatments to biomass, soil VWC, visual

PAGE 68

68 ratings, Crop Circle readings (NDVI and LAI) and ASD readi ngs (NDVI, LAI, Stress1, and Stress2). Regression analysis was performed using the PROC REG method (SAS Inst., 2003) to correlate relationships between visual ratings Crop Circle readings, various individual reflectance wavelengths quantified from the ASD unit, and soil VWC. Additionally, due to the large amount of data generated by the ASD device, ASD data were rendered using partial least squares regression PROC PLS (SAS Inst., 2003) to soil VWC, visual quality ratings, and biomass production. Equations were validated th rough a single sample cross-validation. For cross-validation, ten percent of the sample wa s omitted for prediction purposes so that the number of factors chosen create s the minimal predicted residual sum of squares (PRESS). This process was repeated so that every observat ion was used exactly on ce for cross-validation. Results and Discussion Biomass ANOVA reveals that there was a significant reduction in biomass in the 0.1 kg a.i. ha-1 and 0.4 kg a.i. ha-1 rates of TE, the highest rate of flurprimidol (2.3 kg a.i. ha-1 ) and the low rate of ethephon (30.4 kg a.i. L ha-1) (Table 3-1). Although not alwa ys significant, a dose response trend was observed with TE (Figur e 3-1) and flurprimidol (Figur e 3-2). This dose response was not observed with Ethephon due to the fact th at ethephon is primarily used for seedhead suppression as opposed to biomass suppr ession in turf (Anonymous 2007). Biomass did not correlate well with any remote ly sensed parameter quantified from either device (Tables 3-3, 3-4, 3-5, and 3-6) This data is in agreement with the research conducted by Fenstemaker-Shaulis et al. (1997) and Trenholm et al. (1999b) who found re motely sensed data was not an adequate predictor of biomass. Th e inability to adequately model biomass reduction through increasing rates of PGR app lications may be attributed to a number of factors. PGRs have been found to limit vertical growth but incr ease prostrate growth, thus increasing density

PAGE 69

69 (Branham, 1997). In other research where biom ass has been adequately modeled, PGRs were not used (Kruse et al., 2005, 2006) and biomass was modeled from plots with increasing N rate applications applied at frequent intervals. Volumetric Water Content PGRs applied to bermudagrass in this study ha d only a slight effect on soil VWC. Only TE and flurprimidol applied at th e lowest rate gave significant differences from the untreated control (Table 3-1). Slight co rrelations were obser ved (Table 3-3) in modeling soil VWC with NIR reflectance quantified from the Crop Circle de vice as well as reflectance in the SWIR region at 2,132 nm (Table 3-4) and the 2,080-2,350 nm ra nge (Table 3-5) quantified from the ASD device. This is in agreement with research cond ucted by Ripple (1986) and Hutto et al. (2006). Visual Ratings Color Color rating means (Table 3-1) s how an increase in turf color by TE applications at 0.05 and 0.1 kg a.i. ha-1, flurprimidol applications at 0.3, 0.6, and 1.1 kg a.i. ha-1, and ethephon applications at the low rate compared to the untreated control, however, these increases were not significant (Table 3-1). This general increase could be at tributed to seed head reduction achieved by PGR applications, since seed heads are generally le ss green causing a waning in color on the turf. Additionally, increasing rates of PG Rs resulted in a trend towards reduced turf color which could be caused by a slight injury effect. Only slight correlations in visu al color ratings to remotely se nsed data from either device were achieved (Tables 3-3, 3-4, 3-5, 3-6). The highe st correlations were observed in data taken with the ASD device at wavelengths 661 nm (r2=0.42) and 510nm (r2=0.35), and ratio 605/515 nm (r2=0.41) (Table 3-4 and 3-5).

PAGE 70

70 Quality Improvements in quality were observed with PGR applications. A lthough they were not significant, all rates of TE and fl urprimidol, and the 3.8 kg a.i. ha-1 rate of ethephon slightly improved turf quality (Table 3-1). Similar to turf color ratings, only slight correlations existed between remotely sensed data a nd visual quality ratings. The hi ghest correlations were achieved by the ASD device at wavelengths 661 nm (r2=0.53) and 510 nm (r2=0.48), and ratio 915/975 nm (r2=0.42) (Table 3-4 and 3-5). Density Visual density ratings showed a general increase in density from all rates of TE and flurprimidol, and the low rate of ethephon, however these increases were not significant. Only slight correlations existed between remotely sens ed data and visual dens ity ratings. The highest correlations were achieved by the AS D device at wavelengths 661 nm (r2=0.53) and 510 nm (r2=0.48) (table 3-4). Reflectance Indices vs. PGR Application rates Data derived from the Crop Circle device s how a general increase in NDVI and LAI for all rates of TE and flurprimidol compared to the untreated, ho wever, only the 0.05 and 0.1 kg a.i. ha-1 rate of TE and the 0.3, 0.6, and 1.1 kg a.i. ha-1 rate of flurprimidol are significant (Table 32). Furthermore, NDVI and LAI derived from ASD data show the same trend, yet only the 2.2 kg ha-1 is statistically different from the untreated control (Table 3-2). Th e general trend for the STRESS 1 and STRESS 2 indices show similar re sults, however, STRESS 1 reveals that TE and flurprimidol treated turf was under less stre ss than the untreated control (Table 3-2). Crop Circle Device Reflectance data obtained from the Crop Circle device did not demonstrate its ability to detect biomass differences attributed to PGR applic ations (Table 3-3). Th e ability of the device

PAGE 71

71 to correlate with visual huma n evaluations was weak. Howeve r, the NIR (880 nm) reflectance values from the device slightly modeled changes in soil VWC (r2=0.43) throughout the experiment (Table 3-3). The inability of the Crop circle device to adeq uately detect differences in biomass, visual ratings, and VWC can be attributed to a number of things previously mentioned. Among these are the seedhead production of the turf which could have potentially skewed the reflectance data, and the lack of significant differe nces among the parameters tested. ASD Device Individual wavelengths, computed indices and various ratios obtained from ASD reflectance data did not demonstrate the ability to detect differences in biomass attributed to PGR applications. Some weak correlations exist betw een visual ratings and NDVI, LAI, Stress1, and Stress2 indices computed from the raw data generated by the ASD instrument (Table 3-3). Similar, and sometimes stronger, correlations exist with visual ratings with individual wavelengths investigated throughout the visible range (Table 3-4) Again, this inability to adequately model visual ratings, VWC, and biomass could be attributed to the lack of significant differences in the various parameters of the trea ted verses the untreated control and the seedhead production of the turf. PLS Regression The use of PLS regression on the hyperspectral data generated from the ASD instrument did not perform well in predicting differences in biomass created by various applications of PGRs. Furthermore PLS regression only slig htly predicted differences in soil VWC (r2=0.36) and visual quality (r2=0.26) (Table 3-6). This again is likely due to the lack of significant differences in soil VWC and visual qu ality from the untreated control.

PAGE 72

72 Conclusions This experiment shows the biomass reduction ex pected from increasing rates of PGRs in two of the three PGRs applied (Table 3-1) (Figure 3-1, 3-2). Flurprimidol and TE reduced biomass in response to increasi ng application rates compared to the untreated control; however, the results are not always signifi cant (Table 3-1) (Figure 3-1, 3-2) Similar trends were observed in NDVI and LAI indices computed from Crop Circ le and ASD instruments, as well as Stress1 and Stress2 indices computed from the ASD de vice (Table 3-2). Increases in NDVI and LAI which model plant health, as well as reductions in Stress1 and Stress2 indices, could be attributed to suppression of seed head production in addition to th e decrease of overall biomass. This overall decrease in biomass prevented the scalping of the tu rf in the study area that was observed in the untreated c ontrol and surrounding areas no t treated with PGRs. The inability to adequately model biom ass reduction, visual ra tings, and VWC through increasing rates of PGR applications may be attributed to a number of factors already mentioned. PGRs have been found to limit vertical growth but increase prostrate growth, thus increasing density (Branham, 1997). A general increase in de nsity was observed in this experiment with increasing application rates of TE and flurpr imidol, however, they were not significantly different from the untreated cont rol (Table 3-1). The lack of significant differences in density and all visual ratings affected the ability of th e remote sensing devices to detect differences among theses various parameters. Additionally, this experiment was performed on Princess 77 bermudagrass turf which produces seed heads. These seed heads are white in color and could potentially skew remotely sensed data compared to a turf that does not produce seed heads. Currently, turf managers apply PGRs in the form of broadcast applications to reduce biomass. If remote sensing technology could be used to adequately detect biomass, and was coupled with GPS and GIS technologi es, applications of PGRs coul d be applied on a site specific

PAGE 73

73 basis. In this setting, turf ma nagers could potentially reduce chem ical input thereby reducing the high cost of broadcast PGR applications.

PAGE 74

74 Table 3-1. Evaluation of means of biomass (g), volumetric water content (VWC), visual ratings: color, qual ity, density (rated o n NTEP 1-9 scale 9 is best a nd 6 is acceptable), from fairwa y height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis Princess 77) over five date s (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined. Biomass (g) VWC Color Quality Density Trinexapac-ethyl 0.05 kg a.i. ha-1 52.47 a-c 35.73 a 8.00 ab 8.00 ab 7.87 a-c 0.1 kg a.i. ha-1 45.20 bc 34.45 a-c 8.10 ab 8.05 ab 8.00 ab 0.2 kg a.i. ha-1 47.20 a-c 32.87d 7.73 a-c 7.80 a-c 7.93 ab 0.4 kg a.i. ha-1 43.80 bc 34.20 b-d 7.83 a-c 7.98 ab 8.00 ab Flurprimidol 0.3 kg a.i. ha-1 53.93 a-c 35.85 a 8.10 ab 8.17 a 8.10 ab 0.6 kg a.i. ha-1 49.93 a-c 34.85 a-c 8.19 a 8.11 a 8.17 a 1.1 kg a.i. ha-1 48.13 a-c 35.23 ab 8.07 ab 8.08 a 8.10 ab 2.3 kg a.i. ha-1 36.27 c 34.72 a-c 7.83 ab 7.93 ab 7.83 a-c Ethephon 3.8 kg a.i. ha-1 45.73 bc 35.21 ab 7.96 ab 7.91 ab 7.80 a-c 7.6 kg a.i. ha-1 56.07 a-c 35.24 ab 7.63 bc 7.58 b-d 7.67 b-d 15.2 kg a.i. ha-1 66.13 ab 33.88 b-d 7.47 c 7.33 cd 7.40 cd 30.4 kg a.i. ha-1 66.27 ab 33.37 cd 7.40 c 7.25 d 7.30 d Untreated 71.60 a 34.01 b-d 7.87 a-c 7.77 a-c 7.73 a-d Means in the same column followed by the same lett er are not significantly different (LSD; P< 0.05)

PAGE 75

75 Table 3-2. Evaluation of means of Normalized Difference Vegeta tion Index (NDVI), Leaf Area In dex (LAI) computed from Crop Circle and ASD reflectance data, as well as Stress1 and Stress2 indices computed from ASD re flectance data from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis Princess 77) over five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined. ______Crop Circlez_____ _______________________ASDy__________________________ NDVIx LAIw NDVIx LAIw Stress1v Stress2u Trinexapac-ethyl 0.05 kg a.i. ha-1 0.66 ab 5.08 ab 0.76 ab 7.42 a-c 0.38 bc 0.79 cd 0.1 kg a.i. ha-1 0.67 ab 5.13 ab 0.77 ab 7.76 ab 0.36 c 0.77 cd 0.2 kg a.i. ha-1 0.65 a-d 4.90 a-c 0.76 ab 7.43 a-c 0.38 bc 0.78 cd 0.4 kg a.i. ha-1 0.66 a-c 5.00 a-c 0.76 ab 7.46 a-c 0.37 bc 0.78 cd Flurprimidol 0.3 kg a.i. ha-1 0.67 ab 5.10 ab 0.77 ab 7.51 ab 0.37 bc 0.78 cd 0.6 kg a.i. ha-1 0.66 ab 5.01 ab 0.76 ab 7.34 a-d 0.37 bc 0.78 cd 1.1 kg a.i. ha-1 0.68 a 5.24 a 0.78 a 7.82 a 0.36 c 0.77 d 2.3 kg a.i. ha-1 0.66 a-c 5.03 ab 0.76 ab 7.49 a-c 0.37 bc 0.78 cd Ethephon 3.8 kg a.i. ha-1 0.65 a-d 4.82 a-c 0.75 a-c 7.16 a-d 0.39 a-c 0.78 b-d 7.6 kg a.i. ha-1 0.64 b-d 4.64 b-d 0.75 a-d 6.96 a-d 0.39 a-c 0.78 a-d 15.2 kg a.i. ha-1 0.62 de 4.45 dc 0.72 cd 6.50 cd 0.41 a 0.79 ab 30.4 kg a.i. ha-1 0.60 e 4.23 d 0.72 d 6.39 d 0.42 a 0.80 a 13) Untreated 0.63 c-e 4.64 b-d 0.74 b-d 6.81 b-d 0.40 ab 0.79 a-c Means in the same column followed by the same le tter are not significantly different (LSD; P< 0.05) z Model ACS-210, Holland Scientific y FieldSpec Pro; Analytical Spectr al Devices, Inc., Boulder, CO x NDVI= (880-650nm)/(880+650nm) w LAI= 880/650nm v Stress1= 706/760nm u Stress2= 706/813nm

PAGE 76

76 Table 3-3. Coefficient estimates (r2) of Crop Circle reflectance data, Normalized Differen ce Vegetation Index (NDVI), Leaf Area Index (LAI), 880nm, and 650nm, and, ASD reflectance data computed into NDVI LAI Stress1, and Stress2 indices to volumetric water content (VWC) visual co lor quality, density ratings, and bioma ss production from fair way height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis Princess 77) averaged over five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007). Crop Circlez ASDy r2 NDVIx LAIw 880nm 650nm NDVIx LAIw Stress1v Stress2u VWC 0.10*** 0.12*** 0.43*** 0.00 0.11*** 0.10*** 0.09*** 0.10*** Color 0.18*** 0.16*** 0.00 0.28*** 0.23*** 0.22*** 0.19*** 0.21*** Density 0.22*** 0.19*** 0.03* 0.21*** 0.28*** 0.23*** 0.24*** 0.25*** Quality 0.22*** 0.18*** 0.02 0.24*** 0.29*** 0.24*** 0.23*** 0.26*** Biomass 0.06** 0.10*** 0.00 0.09*** 0.03* 0.08*** 0.05* 0.05*** *, **, *** Significant at the 0.05, 0.01, and 0.001 probability levels, respectively z Model ACS-210, Holland Scientific y FieldSpec Pro; Analytical Spectr al Devices, Inc., Boulder, CO x NDVI= (880-650nm)/(880+650nm) w LAI= 880/650nm v Stress1= 706/760nm u Stress2= 706/813nm

PAGE 77

77 Table 3-4. Coefficient estimates (r2) of ASD reflectance data at in dividual wavelengths, as well as range averages from 630-690nm and 2080nm to volumetric water content (VWC) visual colo r, quality, density ratings, and biomass production from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis Princess 77) averaged over five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007). All numbe rs in column headings are expressed nanometers (nm). ___________________________________________ASD z (nm)____________________________________________ r2 510 535 545 550 661 735 755 813 935 2132 VWC 0.12*** 0.06** 0.05* 0.04* 0.07*** 0.02* 0.04* 0.06* 0.11*** 0.43*** Color 0.35*** 0.25*** 0.25*** 0.25*** 0.42*** 0.05* 0.01 0.01 0.02* 0.11*** Density 0.51*** 0.39*** 0.38*** 0.39*** 0.53*** 0.07* 0.01 0.01 0.02 0.06** Quality 0.48*** 0.35*** 0.35*** 0.35*** 0.53*** 0.06* 0.01 0.01 0.02 0.08*** Biomass 0.00 0.00 0.00 0.01 0.01 0.02 0.03* 0.04* 0.05* 0.01 *, **, *** Significant at the 0.05, 0.01, and 0.001 probability levels, respectively z FieldSpec Pro; Analytical Spectr al Devices, Inc., Boulder, CO

PAGE 78

78 Table 3-5. Coefficient estimates (r2) of range averages from 630-690 nm and 2080 nm and ratios 605/515 nm, 915/975 nm, and 865/725 nm to volumetric water content (VWC ) visual color, quality, density rati ngs, and biomass production from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis Princess 77) averaged ove r five dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007). All numbers in column headings are expressed nanometers (nm). ___________________________________ASD z (nm)______________________________________ r2 630-690 2080-2350 605/515 nm 915/975 nm 865/725 nm VWC 0.19*** 0.40*** 0.12** 0.14** 0.03 Color 0.19*** 0.12*** 0.41*** 0.32*** 0.24*** Density 0.30*** 0.08*** 0.32*** 0.39*** 0.22*** Quality 0.29*** 0.10 0.26*** 0.42*** 0.23*** Biomass 0.00 0.01 0.08* 0.06* 0.13** *, **, *** Significant at the 0.05, 0.01, and 0.001 probability levels, respectively z FieldSpec Pro; Analytical Spectr al Devices, Inc., Boulder, CO

PAGE 79

79 Table 3-6. Partial Least Squares regression co efficients on hyperspectral data from ASD device to predict volumetric water content (VWC), biomass, and visual quality (rated on NTEP 1-9 scale 9 is best and 6 is acceptable) in fairway height bermudagrass ( Cynodon dactylon X C. transvaalensis Princess 77) averaged over five date s (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007). Calibration No. of Factorsy r2 SE n VWC 5 0.36 0.05 195 Biomass 2 0.03 0.05 195 Visual Quality 8 0.26 0.03 195 y Number of factors required to achieve a mi nimal Predicted Residual Sum of Squares (P RESS) of prediction for the partial least squares regression model. z FieldSpec Pro; Analytical Spectr al Devices, Inc., Boulder, CO

PAGE 80

80 Figure 3-1. Biomass measurements collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis Princess 77) treated with trinexapac -ethyl at 0.05, 0.1, 0.2, and 0.4 kg a.i. ha-1. Plots measured 1.5 X 3.0 m with three replications. Means for biomass measurements was comput ed using Fishers protected LSD (p< 0.05) from data over 5 dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined. 0 10 20 30 40 50 60 70 80 Untreated 0.05 0.1 0.2 0.4Rate (kg a.i. ha1 ) Biomass (g)

PAGE 81

81 Figure 3-2. Biomass measurements collected from fairway height hybrid bermudagrass ( Cynodon dactylon X C. transvaalensis Princess 77) treated with flurprimidol at 0.3, 0.6, 1.1, and 2.3 kg a.i. ha-1. Plots measured 1.5 X 3.0 m with three replications. Means for biomass measurements was comput ed using Fishers protected LSD (p< 0.05) from data over 5 dates (22 June, 12 July, 25 July, 22 Aug., and 4 Sept. 2007) combined. 0 10 20 30 40 50 60 70 80 Untreated 0.30.61.12.3 Rate (kg a.i. ha-1) Biomass (g)

PAGE 82

82 CHAPTER 4 CONCLUSIONS Currently many turf managers and researchers still use time and labor intensive techniques to manage and assess turf that originate from d ecades ago. Many of these management practices have been proven repeatedly to work in a variety of situations to assess turfgrass stress, but may be time consuming and inconsistent. However, th ere are new developments in ways of assessing and mapping stress that increase the efficiency by which one manages an d assesses turf. With increasing water and other environmental restrict ions turf managers and researchers need be aware of technological advan ces in their industry. The research discussed in this thesis give s greater insight into the use of some the technological advancements available to turfgr ass managers demonstrat ing the sensitivity of remote sensing instrumentation to various parame ters of turfgrass stress. In conducting these experiments many realizations we re made about the usability a nd future potential for remote sensing technology in the turfgr ass management field. Some recommendations for anyone who wishes to continue the research referr ed to in this thesis are as follows: Irrigation X Nitrogen study. Conduct the experiment in th e earlier part of the growing season, when rainfall is not so prevalent. This will potentially reduce un iformity in soil VWC. Also, apply N every two weeks as described by Kr use et al. (2005); this will continually promote differences in growth. In addition, investig ate more into the use of PLS regression on hyperspectral data to formulate predic tion models for N and soil VWC levels. PGR study. Conduct the experiment on turfgrass that does not produce seedheads. In this experiment seed heads could have potentially skewed remotely sensed data; this will at least take that factor out. Also, obtain remote sensing da ta before and after PGR a pplications to assess the differences they create. From this data possi bly try to make applications based on remotely

PAGE 83

83 sensed data and not just on an in cremental scale. In addition, i nvestigate more into the use of PLS regression on hyperspectral data to formul ate prediction models for assessment of the biomass of the turf. Crop circle. Test in various light and ambient conditions to assess different turfgrass parameters and investigate its usability in an assortment of environmental settings. Turfgrass managers work in all conditions, rain or shine. An assessment of how these devices will work in varying environmental conditions must be evalua ted before they are employed into real world applications of turfgrass management. Beyond these experiments. Conduct experiments on larger scale while adding other technologies such as GPS, GIS, and VRA. Th is will allow researcher s to better understand the usability of the technology and th e logistics of operating it as a turfgrass manager. Remote sensing in a large scale setti ng is not very eff ective without the co mbination of other technologies. You need GPS and GIS to georefer ence remotely sensed data to understand where the data is coming from. Then by applying VR A technology some sort of economic impact can be assessed to understand the potential benefit of the use all of these technologies. Remote sensing is just the beginning of a long road of investigation in helping turfgrass managers reduce irrigation and chemical use and appl y these inputs only where needed. Currently the use of remote sensing technol ogy in the turfgrass ma nagement industry is still in its infancy. However, the experiments desc ribed in this thesis al ong with other research that has been referenced show it has potential to assist turfgrass managers to make better decisions about managing their turf Presently, this technology is not in a very usable form and much research still needs to be conducted to understand its place in the field of turfgrass management.

PAGE 84

84 APPENDIX A EXPERIMENTAL LAYOUT FOR DETECTION OF LEAF NITROGEN CONCENTRATION AND SOIL VOLUMETRIC WATER CON TENT USING GROUND-BASED REMOTE SENSING TECHNOLOGY

PAGE 85

85 101 201 301 401 102 202 302 402 103 203 303 403 104 204 304 404 501 601 701 801 502 602 702 802 503 603 703 803 504 604 704 804 901 1001 1101 1201 902 1002 1102 1202 903 1003 1103 1203 904 1004 1104 1204 Figure A-1. Experimental layout for detecti on of leaf nitrogen concentration and soil volumetric water content using ground-bas ed remote sensing technology

PAGE 86

86 Table A-1. Description of treatments for Figure A-1 Whole Plot Subplot Rep Treatment # Irrigation Nitrogen 1 2 3 1 60% ET 0 kg ha-1 101 704 1201 2 60% ET 25 kg ha-1 102 702 1203 3 60% ET 50 kg ha-1 103 703 1202 4 60% ET 100 kg ha-1 104 701 1204 5 80% ET 0 kg ha-1 201 802 901 6 80% ET 25 kg ha-1 202 803 903 7 80% ET 50 kg ha-1 203 804 904 8 80% ET 100 kg ha-1 204 801 902 9 100% ET 0 kg ha-1 301 503 1003 10 100% ET 25 kg ha-1 302 502 1002 11 100% ET 50 kg ha-1 303 501 1004 12 100% ET 100 kg ha-1 304 504 1001 13 120% ET 0 kg ha-1 401 602 1101 14 120% ET 25 kg ha-1 402 601 1104 15 120% ET 50 kg ha-1 403 603 1102 16 120% ET 100 kg ha-1 404 604 1103

PAGE 87

87 APPENDIX B DAILY RAINFALL AND ET DATA FOR JAY, FL Table B-1. Daily Rainfall and ET data for Jay, FL Day Rainfall (cm)a ET (cm)a Cumulative Difference (cm) Irrigation Applied (cm) June 14 0.38 0.48 (0.10) 0.20b 15 0.00 0.41 (0.31) 0.00 16 0.51 0.38 (0.18) 0.00 17 2.29 0.53 1.57 0.00 18 0.00 0.33 1.24 0.00 19 2.36 0.30 3.30 0.00 20 1.27 0.48 4.09 0.00 21 0.00 0.58 3.50 0.00 22 0.00 0.53 2.97 0.00 23 0.00 0.46 2.51 0.00 24 0.00 0.30 2.21 0.00 25 0.00 0.36 1.85 0.00 26 0.10 0.36 1.60 0.00 27 0.00 0.36 1.24 0.00 28 0.25 0.46 1.04 0.00 29 0.03 0.46 0.61 0.00 30 0.38 0.53 0.45 0.00 July 1 0.00 0.53 (0.08) 0.00 2 0.00 0.43 (0.51) 0.00 3 0.00 0.30 (0.82) 0.00 4 0.89 0.33 (0.26) 0.00 5 0.94 0.33 0.35 0.00 6 0.46 0.48 0.33 0.00 7 0.00 0.48 (0.16) 0.00 8 0.76 0.43 0.17 0.00 9 0.00 0.53 (0.36) 0.00 10 0.00 0.48 (0.84) 0.00 11 0.00 0.48 (1.32) 0.00 12 0.89 0.48 (0.92) 0.00 13 0.00 0.46 (1.37) 0.00 14 0.00 0.33 (1.71) 0.00 15 2.79 0.30 0.78 0.00 16 0.10 0.33 0.56 0.00 17 0.71 0.51 0.76 0.00 18 0.00 0.48 0.28 0.00 19 0.00 0.46 (0.18) 0.00 20 0.00 0.46 (0.64) 0.20b 21 1.40 0.53 0.43 0.00

PAGE 88

88 Table B-1. Continued Day Rainfall (cm)a ET (cm)a Cumulative Difference (cm) Irrigation Applied (cm) 23 0.53 0.41 0.15 0.00 24 3.35 0.43 3.07 0.00 25 0.28 0.46 2.89 0.00 26 0.00 0.36 2.53 0.00 27 0.00 0.48 2.05 0.00 28 0.00 0.51 1.54 0.00 29 0.00 0.46 1.09 0.00 30 2.54 0.41 3.22 0.00 31 0.00 0.43 2.79 0.00 August 1 2.62 0.33 5.07 0.00 2 0.00 0.43 4.64 0.00 3 0.15 0.46 4.34 0.00 4 0.00 0.46 3.88 0.00 5 0.00 0.46 3.42 0.00 6 0.00 0.43 2.99 0.00 7 0.00 0.48 2.51 0.00 8 0.00 0.43 2.08 0.00 9 0.00 0.36 1.72 0.00 10 0.41 0.48 1.64 0.00 11 0.03 0.38 1.29 0.00 12 0.84 0.25 1.87 0.00 13 0.00 0.43 1.44 0.00 14 0.00 0.48 0.96 0.00 15 0.00 0.46 0.50 0.00 16 0.00 0.46 0.04 0.00 17 0.00 0.43 (0.39) 0.00 18 0.33 0.43 (0.49) 0.00 19 0.00 0.43 (0.92) 0.00 20 0.00 0.48 (1.40) 0.00 21 0.00 0.48 (1.89) 0.00 22 0.00 0.51 (2.39) 0.00 23 0.00 0.38 (2.78) 0.00 24 0.00 0.38 (3.16) 0.00 25 0.05 0.46 (3.56) 2.54 26 4.34 0.33 2.99 0.00 27 0.28 0.33 2.94 0.00 28 0.89 0.20 3.63 0.00 29 0.33 0.43 3.52 0.00 30 0.00 0.41 3.12 0.00 31 0.08 0.28 2.91 0.00 September 1 0.05 0.33 2.64 0.00

PAGE 89

89 Table B-1. Continued Day Rainfall (cm)a ET (cm)a Cumulative Difference (cm) Irrigation Applied (cm) 3 2.34 0.28 4.64 0.00 4 1.32 0.41 5.56 0.00 a Based on data from http://fawn.ifas.ufl.edu/ b Irrigation applied to water in fertilizer application

PAGE 90

90 APPENDIX C CUMALITIVE WEEKLY RAINFALL AND ET DATA FOR JAY, FL Table C-1. Cumalitive weekly rain fall and et data for Jay, Fl ______________Cumulative________________ Week Ending On Rainfall (cm)a ET (cm)a Difference (cm) Irrigation Applied (cm) June 15 0.38 0.89 (0.31) 0.00 22 6.43 3.15 2.97 0.00 29 0.38 2.74 0.61 0.00 July 6 2.67 2.95 0.33 0.00 20 3.61 2.87 (0.64) 0.20b 27 5.56 3.07 2.05 0.00 August 3 5.31 3.02 4.34 0.00 10 0.41 3.10 1.64 0.00 17 0.86 2.90 (0.39) 0.00 24 0.33 3.10 (3.16) 2.54 31 5.97 2.44 2.91 0.00 a Based on data from http://fawn.ifas.ufl.edu/ b Irrigation applied to water in fertilizer application

PAGE 91

91 APPENDIX D MOTHLYRAINFALL AND ET AVERAGES FROM 2003-2006 FOR JAY, FL Table D-1. MothlyRainfall and ET averages from 2003-2006 for Jay, FL Rainfall (cm) ET (cm) Rainfall-ET (cm) Cumulative Total Rainfall-ET (cm) January 7.2898 4.92125 2.36855 2.36855 February 13.6144 5.5118 8.1026 10.47115 March 11.91895 8.46455 3.4544 13.92555 April 15.3543 10.4775 4.8768 18.80235 May 7.8105 12.5984 -4.7879 14.01445 June 24.08555 12.192 11.89355 25.908 July 24.31415 12.5984 11.71575 37.62375 August 20.2438 12.2047 8.0391 45.66285 September 16.0401 10.0965 5.9436 51.60645 October 5.6896 7.28345 -1.59385 50.0126 November 11.99515 4.7625 7.23265 57.24525 December 10.69975 3.34645 7.3533 64.59855 Based on data from http://fawn.ifas.ufl.edu/

PAGE 92

92 LIST OF REFERENCES Anonymous. 2007. Proxy (specimen label). Bayer Scie ntific, Triangle Park, NC. U.S. Patent 4,240,819. Date issued: 25 Apr. 2005. Aston, A.R., and C.H.M. Van Bavel. 1972. Soil Su rface Water Depletion and Leaf Temperature. Agron. J. 64:368-373. Augustin, B.J., and G.H. Snyder. 1984. Moisture Se nsor Controlled Irri gation For Maintaining Bermudagrass Turf. Agron. J. 76:848-850. Bahrun, A., V.O. Mogensen, and C.R. Jensen. 2003. Water Stress Detec tion in Field Grown Maize by Using Spectral Vegetation Index. Co mmun. Soil Sci. Plant Anal. 34:65-79. Barret, J. 2003. Golf Course irrigation: Envi ronmental Design and Management Practices. 413417. John Wiley & Sons, Inc., Hoboken, NJ. 413-417. Bethel, M., T. Gress, S. White, J. Johnson, T. Sheely, B. Roberts, N. Gat, G. Scriven, G. Hagglund, M. Paggi, and N. Groenenberg. 2003. Im age-Based Variable Rate Plant Growth Regulator Application in Cotton at Sheely Farms in California. 6-10 Jan. 2003. Beltwide Cotton Conferences, Nashville, TN. Bell, G.E., B.M. Howell, G.V. Johnson, W.R. Raun, M.L. Stone, and J.B. Solie. 2004. Optical Sensing of Turfgrass Chlorophyll Content and Tissue Nitrogen. Hortic. Sci. 39(5):11301132. Bell, G.E., D.L. Martin, R.M. Kuzmic, M.L. Stone, and J.B. Solie. 2000. Herbicide Tolerance of Two Cold-Resistant Bermudagrass ( Cynodon spp.) Cultivars Determined by Visual Assessment and Vehicle Mounted Optic al Sensing. Weed Tech. 14:635-641. Bell, G.E., D.L. Martin, M.L. Stone, J.B. Solie, and G.V. Johnson. 2002a. Turf area Mapping Using Vehicle Mounted Optical Sensors. Crop Sci. 42:648-651. Bell, G.E., D.L. Martin, S.G. Wiese, D.D. D obsom, M.W. Smith, M.L. Stone, and J.B. Solie. 2002b. VehicleMounted Optical Sensing An Objective Means for Evaluating Turf Quality. Crop Sci. 42:197-201. Blackmer, T.M., J.S. Schepers, and G. E. Varv el. 1994. Light reflectan ce Compared with Other Nitrogen Stress Measurements in Co rn Leaves. Agron. J. 86: 934-938. Blackmer, T.M., J.S. Schepers, G.E. and Varv el, E.A. Walter-Shea. 1996.Nitrogen Deficiency Detection Using Reflected Shortwave Radiat ion From Irrigated Corn Canopies. Agron. J. 88: 1-5. Branham, Bruce. Plant Growth Regulators for Fine Turf Use. http://www.interactiveturf.com/tips/1997_02.htm 24 Oct. 2007.

PAGE 93

93 Brown, K.W., R.L. Duble, and J.C. Thomas. 1977. Influence of Management and Season on Fate of N Applied to Golf Greens. Agron. J. 69:667-671. Ehrler, W.L., S.B. Idso, R.D. Jackson, and R.J. Reginato. 1978. Wheat canopy temperature: Relation to Plant Water Pote ntial. Agron. J. 70: 251-256. Fenstermaker-Shaulis, L.K., A. Leskys, and D.A. Devitt. 1997. Utilization of Remotely Sensed Data to Map and Evaluate Turfgrass Stress A ssociated with Drought. Journal of Turfgrass Management. 2:65-81. Geeske, Joel, J.A. Gamon, and C.B. Field. 1997. Pr oduction Efficiency in Sunflower: The Role of Water and Nitrogen Stress. Remote Sens. Environ. 62:176-188. Hopkins, W. G. 1999. Introduction to Plant Physiology. 2nd ed. John Wiley and Sons Inc. New York, Ch. 22 The physiology of plants under stress. 453-457. Horst, G.L, J.C. OToole, and K.L. Faver. 1989. Seasonal and Species Variation in Baseline Functions for Determining Crop Water Stress Indices in Turfgrass. Crop Sci. 29:12271232. Hutto, K.C., R.L King, J.D. Byrd, and D.R. Shaw. 2006. Implementation of Hyper spectral Radiometry in Irrigation Management of Creeping Bentgrass Putting Greens. Crop. Sci. 46:1564-1569. Idso, S.B., R.J. Reginato, D.C Reicosky, and J.L. Hatfield. 1981. Determin ing Soil Induced Plant Water Potential Depressions in Alfalfa by M eans of Infrared Thermometry. Agron. J. 73: 826-830. Ikemura, Y., and B. Leinauer. 2006. Remote Se nsing Technology Detects Turfgrass Stress. Golfdom, May, p.63-66. Jalali-Farahani, H.R., A.D. Matthias, and D.C. Slack. 1993. Crop Water Stress Index Models for Bermudgrass Turf: A Comparison. Agron. J. 85:1210-1217. Jalali-Frahani, H.R., D.C. Slack, D.M. Kopec, A.D. Matthias, and P.W. Brown. 1994. Evaluation of Resistances for Bermudagrass Turf Crop Wa ter Stress Index Models. Agron. J. 86: 574581. Kruse, J.K., N.E. Christians, and M.H. Chaplin. 2005. Remote Sensing of Phosphorus Deficiencies in Agrostis Stolonifera. Inter. Turfgra ss Soc. Res. J. 10:923-928. Kruse, J.K., N.E. Christians, and M.H. Chap lin. 2006. Remote Sensing of Nitrogen Stress in Creeping Bentgrass. Agron. J. 98:1640-1645. Lee, Wonsuk, S.W. Searcy, and T. Kataoka. 1999 Assessing Nitrogen Stress in Corn Varieties of Varying Color. 18-21 July. ASAE A nnual International Meeting. Toronto, Ontario, Canada.

PAGE 94

94 Lillesand, T.M., and R.W. Keifer. 2000. Remote Sensing and Image Interpretation. 4th ed. John Wiley & Sons, New York. Martin, D.L, C.S. Throssel, and D.J. Wehner. 1994. Models for Predicting Lower Limit of The Canopy-Air Temperature Difference of Two C ool Season Grasses. Crop Sci. 34: 192-198. Min, M. and W.S. Lee. 2003. Spectral Based Nitrogen Sensing for Citrus. 27-30 July. ASAE Annual International Meeting. Las Vegas, NV Morgan, W.C., J. Letey, S.J. Richards, and N. Valoras. 1966. Physical Soil Amendments, Soil Compaction, Irrigation, and Wetting Agents in Turfgrass Management Effects on Compactibility, Water Infiltration Rates, Eva potranspiration, and Number of Irrigations. Agron. J. 58:525-527. Morris, K.N. A Guide to NTEP Turfgrass Ratings. http://www.ntep.org/reports/ratings.htm. 14 Nov. 2007 Mylavarapu, R.S. and E.D. Kennelly. UF/IFA S Extentsion Soil Testi ng Laboratory (ESTL) Analytical Procedures and Training Manual1. http://edis.ifas.ufl.edu/pdffiles/SS/SS31200.pdf 14 Dec. 2007. Osborne, S.L., J.S. Schepers, D.D. Francis, a nd M.R. Schlemmer. 2002. De tection of Phosphorus and Nitrogen Deficiencies in Corn Using Spectral Radiance Measurements. Agron. J. 94:1215-1221. Ripple, W.J. 1986. Spectral Reflectance Relationsh ips to Leaf Water Stress. Photogramm. Eng. Rem. Sens. 52:1669-1675 Rouse, J.W., R.H. Haas, J.A. Schell, and D.W. Deering. 1973. Monitoring the vernal advancement and retroradation (green wave effect) of Natural Vegetation. Prog. Rep. SRC 1978-1, Remote Sensing Center, Texas A&M Univ., College Station, 93p. (NTIS) No. E73-106393. Salisbury, F.B. and C.W. Ross. 1992. Plant Philology. 4th ed. Wadsworth Publ. Co. Belmont, CA. SAS Institute. 2002-2003. The SAS System for Windows. Release 9.1. SAS Inst., Cary, NC. Slack, D.C., K.M. Geiser, K.W. Stange, and E.R. Allred. 1981. Irrigation Scheduling in SubHumid Aareas with Infrared Thermometr y. p. 116-124. In Proc. Irrigation Scheduling Conf., Chicago, and IL. 14-15 Dec. 1981. ASAE Annual International Meeting, St Joseph, MI. Starks, P. J., D. Zhao, W.A. Phillips, a nd S.W. Coleman. 2006. Development of Canopy Reflectance Algorithms for Real-Time Predicti on of Bermudagrass Pasture Biomass and Nutritive Values. Crop Sci. 46:927-934.

PAGE 95

95 Stowell, L. and W. Gelernter. 2006. Sensing th e Future. Golf Course Management. March p.107110. Throssel, C.S., R.N. Carrow, and G.A. Milliken. 1987. Canopy Temperature Based Irrigation Scheduling Indices for Kentucky Bl uegrass Turf. Crop Sci. 27:126-131. Tobias, R.D. 1997. An Introduction to Partial Leas t Squares Regression (Online). Available at SAS institute http://ftp.sas.com/techsup/download/technote/ts509.pdf 17 April 2007. Trenholm, L. E., R.N. Carrow, and R.R. Duncan. 1999b. Relationship of Multispectral Radiometry Data to Qualitative Data in Turfgrass Research. Crop Sci. 39:763-769. Trenholm, L.E., R.R. Duncan, and R.N. Carrow. 1999a. Wear Tolerance, Shoot Performance, and Spectral Reflectance of Seashore Pa spalum and Bermudagrass. Crop Sci. 39:1147:1152. Trenhholm, L.E., M.J. Schlossberg, G. Lee, a nd W. Parks. 2000. An Evaluation of Multispectral Responses on Selected Turfgrass Specie s. Int. J. Remote Sensing. 21:709-721. Turgeon. A.J. 2008. Turfgrass Management. 8th edition. Pearson Education Inc. Upper Saddle River, NJ. Unruh, J. and M.L. Elliot. 1999. Best Manageme nt Practices for Florida Golf Courses. University of Florida., Gainesville. VanBibber, L. 2006. Putting the Numbers to PGR's. http://groundsmag.com/chemicals/grounds_maintenance_putting_numbers_pgrs/. 24 Oct. 2007 Yoder, B. J. and R.E. Pettigree-Crosby. 1995. Predicting Nitrogen and Chlorophyll Content and Concentrations From Reflectance Spectra ( 400-2500) at Leaf and Canopy Scales. Remote Sens. Environ. 53:199-211. Xiong, X, G.E. Bell, J.B. Solie, M.W. Smith, and B. Martin. 2007. Bermudagrass Seasonal Responses to Nitrogen Fertilization and Irrigat ion Detected Using Op tical Sensing. Crop Sci. 47:1603-1610.

PAGE 96

96 BIOGRAPHICAL SKETCH Jason Hamilton Frank was born in 1982, in Deland, FL. He lived there until 2001when he graduated from Deland high school and was accepted to the University of Central Florida. He spent a year there as a business administration ma jor before realizing he wanted to attend the University of Florida. He transferred to Daytona Beach Community College in August of 2002 and received his A.A. degree in Ju ly of 2003. He then transferred to the University of Florida in August of 2003 where he completed his B.S. in turfgrass science with a minor in business administration in December of 2005. He then went on to graduate school, also at the University of Florida, where he will gradua te with his M.S. in horticultu re sciences with a minor in agricultural and biologic al engineering and in terdisciplinary concen tration in geographic information sciences in May of 2008. Upon graduation Jason has accepted a 2nd assistant superintendent position at Royal Po inciana Golf Club in Naples, FL.