Citation
Modeling North Florida Dairy Farm Management Strategies to Alleviate Ecological Impacts under Varying Climatic Conditions: An Interdisciplinary Approach

Material Information

Title:
Modeling North Florida Dairy Farm Management Strategies to Alleviate Ecological Impacts under Varying Climatic Conditions: An Interdisciplinary Approach
Creator:
CABRERA, VICTOR E.
Copyright Date:
2008

Subjects

Subjects / Keywords:
Climate models ( jstor )
Crops ( jstor )
Dairy farming ( jstor )
Farmers ( jstor )
Farms ( jstor )
Forage ( jstor )
Leaching ( jstor )
Manure ( jstor )
Modeling ( jstor )
Pastures ( jstor )
Suwannee River, FL ( local )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Victor E. Cabrera. Permission granted to University of Florida to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
8/7/2004
Resource Identifier:
56813724 ( OCLC )

Downloads

This item has the following downloads:

cabrera_v ( .pdf )

dnfdfm.xls

cabrera_v_Page_110.txt

cabrera_v_Page_007.txt

cabrera_v_Page_215.txt

cabrera_v_Page_269.txt

cabrera_v_Page_030.txt

cabrera_v_Page_092.txt

cabrera_v_Page_067.txt

cabrera_v_Page_151.txt

cabrera_v_Page_182.txt

cabrera_v_Page_091.txt

cabrera_v_Page_114.txt

cabrera_v_Page_051.txt

cabrera_v_Page_065.txt

cabrera_v_Page_273.txt

cabrera_v_Page_049.txt

cabrera_v_Page_179.txt

cabrera_v_Page_259.txt

cabrera_v_Page_162.txt

cabrera_v_Page_154.txt

cabrera_v_Page_237.txt

cabrera_v_Page_276.txt

cabrera_v_Page_080.txt

cabrera_v_Page_283.txt

cabrera_v_Page_077.txt

cabrera_v_Page_213.txt

cabrera_v_Page_203.txt

cabrera_v_Page_148.txt

cabrera_v_Page_285.txt

cabrera_v_Page_035.txt

cabrera_v_Page_202.txt

cabrera_v_Page_158.txt

cabrera_v_Page_264.txt

cabrera_v_Page_255.txt

cabrera_v_Page_166.txt

cabrera_v_Page_192.txt

cabrera_v_Page_211.txt

cabrera_v_Page_093.txt

cabrera_v_Page_218.txt

cabrera_v_Page_029.txt

cabrera_v_Page_261.txt

cabrera_v_Page_144.txt

cabrera_v_Page_251.txt

cabrera_v_Page_295.txt

cabrera_v_Page_089.txt

cabrera_v_Page_105.txt

cabrera_v_Page_021.txt

cabrera_v_Page_298.txt

cabrera_v_Page_299.txt

cabrera_v_Page_050.txt

cabrera_v_Page_277.txt

cabrera_v_Page_103.txt

cabrera_v_Page_191.txt

cabrera_v_Page_128.txt

cabrera_v_Page_140.txt

cabrera_v_Page_156.txt

cabrera_v_Page_281.txt

cabrera_v_Page_108.txt

cabrera_v_Page_244.txt

cabrera_v_Page_268.txt

cabrera_v_Page_038.txt

cabrera_v_Page_020.txt

cabrera_v_Page_242.txt

cabrera_v_Page_278.txt

cabrera_v_Page_064.txt

cabrera_v_Page_141.txt

cabrera_v_Page_153.txt

cabrera_v_Page_279.txt

cabrera_v_Page_008.txt

cabrera_v_Page_113.txt

cabrera_v_Page_296.txt

cabrera_v_Page_172.txt

cabrera_v_Page_224.txt

cabrera_v_Page_180.txt

cabrera_v_Page_152.txt

cabrera_v_Page_266.txt

cabrera_v_Page_241.txt

cabrera_v_Page_119.txt

cabrera_v_Page_189.txt

cabrera_v_Page_293.txt

cabrera_v_Page_037.txt

cabrera_v_Page_115.txt

cabrera_v_Page_072.txt

cabrera_v_Page_233.txt

cabrera_v_Page_145.txt

cabrera_v_Page_157.txt

cabrera_v_Page_245.txt

cabrera_v_Page_149.txt

cabrera_v_Page_112.txt

cabrera_v_Page_045.txt

cabrera_v_Page_005.txt

cabrera_v_Page_274.txt

cabrera_v_Page_014.txt

cabrera_v_Page_216.txt

cabrera_v_Page_002.txt

cabrera_v_Page_198.txt

cabrera_v_Page_196.txt

cabrera_v_Page_250.txt

cabrera_v_Page_016.txt

cabrera_v_pdf.txt

cabrera_v_Page_031.txt

cabrera_v_Page_185.txt

cabrera_v_Page_290.txt

cabrera_v_Page_178.txt

cabrera_v_Page_223.txt

cabrera_v_Page_155.txt

cabrera_v_Page_143.txt

cabrera_v_Page_181.txt

cabrera_v_Page_073.txt

cabrera_v_Page_116.txt

cabrera_v_Page_236.txt

cabrera_v_Page_286.txt

cabrera_v_Page_097.txt

cabrera_v_Page_193.txt

cabrera_v_Page_188.txt

cabrera_v_Page_235.txt

cabrera_v_Page_271.txt

cabrera_v_Page_003.txt

cabrera_v_Page_289.txt

cabrera_v_Page_131.txt

cabrera_v_Page_074.txt

cabrera_v_Page_052.txt

cabrera_v_Page_068.txt

cabrera_v_Page_070.txt

cabrera_v_Page_260.txt

cabrera_v_Page_120.txt

cabrera_v_Page_117.txt

cabrera_v_Page_207.txt

cabrera_v_Page_059.txt

cabrera_v_Page_066.txt

cabrera_v_Page_130.txt

cabrera_v_Page_194.txt

cabrera_v_Page_221.txt

cabrera_v_Page_099.txt

cabrera_v_Page_146.txt

cabrera_v_Page_220.txt

cabrera_v_Page_270.txt

cabrera_v_Page_300.txt

cabrera_v_Page_129.txt

cabrera_v_Page_222.txt

cabrera_v_Page_263.txt

cabrera_v_Page_063.txt

cabrera_v_Page_122.txt

cabrera_v_Page_118.txt

cabrera_v_Page_136.txt

cabrera_v_Page_102.txt

cabrera_v_Page_036.txt

cabrera_v_Page_200.txt

cabrera_v_Page_254.txt

cabrera_v_Page_163.txt

cabrera_v_Page_133.txt

cabrera_v_Page_247.txt

cabrera_v_Page_058.txt

cabrera_v_Page_098.txt

cabrera_v_Page_199.txt

cabrera_v_Page_197.txt

cabrera_v_Page_204.txt

cabrera_v_Page_161.txt

cabrera_v_Page_219.txt

cabrera_v_Page_109.txt

cabrera_v_Page_071.txt

cabrera_v_Page_027.txt

cabrera_v_Page_227.txt

cabrera_v_Page_028.txt

cabrera_v_Page_084.txt

cabrera_v_Page_165.txt

cabrera_v_Page_087.txt

cabrera_v_Page_047.txt

cabrera_v_Page_069.txt

cabrera_v_Page_214.txt

cabrera_v_Page_226.txt

cabrera_v_Page_061.txt

cabrera_v_Page_106.txt

cabrera_v_Page_280.txt

cabrera_v_Page_044.txt

cabrera_v_Page_004.txt

cabrera_v_Page_015.txt

cabrera_v_Page_195.txt

cabrera_v_Page_055.txt

cabrera_v_Page_288.txt

cabrera_v_Page_267.txt

cabrera_v_Page_183.txt

cabrera_v_Page_104.txt

cabrera_v_Page_187.txt

cabrera_v_Page_081.txt

cabrera_v_Page_138.txt

cabrera_v_Page_253.txt

cabrera_v_Page_205.txt

cabrera_v_Page_171.txt

cabrera_v_Page_079.txt

cabrera_v_Page_075.txt

cabrera_v_Page_135.txt

cabrera_v_Page_173.txt

cabrera_v_Page_170.txt

cabrera_v_Page_184.txt

cabrera_v_Page_011.txt

cabrera_v_Page_249.txt

cabrera_v_Page_086.txt

cabrera_v_Page_096.txt

cabrera_v_Page_230.txt

cabrera_v_Page_009.txt

cabrera_v_Page_256.txt

cabrera_v_Page_262.txt

cabrera_v_Page_001.txt

cabrera_v_Page_174.txt

cabrera_v_Page_238.txt

cabrera_v_Page_053.txt

cabrera_v_Page_284.txt

cabrera_v_Page_082.txt

cabrera_v_Page_006.txt

cabrera_v_Page_012.txt

cabrera_v_Page_217.txt

cabrera_v_Page_134.txt

cabrera_v_Page_033.txt

cabrera_v_Page_291.txt

cabrera_v_Page_125.txt

cabrera_v_Page_164.txt

cabrera_v_Page_258.txt

cabrera_v_Page_229.txt

cabrera_v_Page_056.txt

cabrera_v_Page_176.txt

cabrera_v_Page_101.txt

cabrera_v_Page_297.txt

cabrera_v_Page_139.txt

cabrera_v_Page_083.txt

cabrera_v_Page_085.txt

cabrera_v_Page_107.txt

cabrera_v_Page_142.txt

cabrera_v_Page_024.txt

cabrera_v_Page_124.txt

cabrera_v_Page_010.txt

cabrera_v_Page_121.txt

cabrera_v_Page_147.txt

cabrera_v_Page_022.txt

cabrera_v_Page_243.txt

cabrera_v_Page_159.txt

cabrera_v_Page_094.txt

cabrera_v_Page_169.txt

cabrera_v_Page_305.txt

cabrera_v_Page_150.txt

cabrera_v_Page_042.txt

cabrera_v_Page_060.txt

cabrera_v_Page_032.txt

cabrera_v_Page_265.txt

cabrera_v_Page_212.txt

cabrera_v_Page_225.txt

cabrera_v_Page_057.txt

cabrera_v_Page_054.txt

cabrera_v_Page_168.txt

cabrera_v_Page_303.txt

cabrera_v_Page_239.txt

cabrera_v_Page_023.txt

cabrera_v_Page_095.txt

cabrera_v_Page_018.txt

cabrera_v_Page_210.txt

cabrera_v_Page_272.txt

cabrera_v_Page_090.txt

cabrera_v_Page_228.txt

cabrera_v_Page_126.txt

cabrera_v_Page_076.txt

cabrera_v_Page_019.txt

cabrera_v_Page_282.txt

cabrera_v_Page_257.txt

cabrera_v_Page_167.txt

cabrera_v_Page_232.txt

cabrera_v_Page_209.txt

cabrera_v_Page_248.txt

cabrera_v_Page_039.txt

cabrera_v_Page_043.txt

cabrera_v_Page_252.txt

cabrera_v_Page_127.txt

cabrera_v_Page_088.txt

cabrera_v_Page_026.txt

cabrera_v_Page_246.txt

cabrera_v_Page_111.txt

cabrera_v_Page_275.txt

cabrera_v_Page_234.txt

cabrera_v_Page_160.txt

cabrera_v_Page_240.txt

cabrera_v_Page_137.txt

cabrera_v_Page_301.txt

cabrera_v_Page_025.txt

cabrera_v_Page_186.txt

cabrera_v_Page_206.txt

cabrera_v_Page_034.txt

cabrera_v_Page_292.txt

cabrera_v_Page_294.txt

cabrera_v_Page_287.txt

cabrera_v_Page_177.txt

cabrera_v_Page_304.txt

cabrera_v_Page_041.txt

cabrera_v_Page_078.txt

cabrera_v_Page_302.txt

cabrera_v_Page_013.txt

cabrera_v_Page_132.txt

cabrera_v_Page_100.txt

cabrera_v_Page_201.txt

cabrera_v_Page_048.txt

cabrera_v_Page_040.txt

cabrera_v_Page_208.txt

cabrera_v_Page_017.txt

cabrera_v_Page_123.txt

cabrera_v_Page_175.txt

cabrera_v_Page_046.txt

cabrera_v_Page_231.txt

cabrera_v_Page_190.txt

cabrera_v_Page_062.txt


Full Text












MODELING NORTH FLORIDA DAIRY FARM MANAGEMENT STRATEGIES TO
ALLEVIATE ECOLOGICAL IMPACTS UNDER VARYING CLIMATIC
CONDITIONS: AN INTERDISCIPLINARY APPROACH
















By

VICTOR E. CABRERA


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

UNIVERSITY OF FLORIDA


2004

































Copyright 2004

by

Victor E. Cabrera

































To my wife Milagritos















ACKNOWLEDGMENTS

I want to start by recognizing the exceptional interdisciplinary and collaborative

work of my supervisory committee. I would like to express my eternal gratitude and

highest appreciation to Dr. Peter E. Hildebrand, my mentor and adviser, for his

unconditional support, encouragement, and outstanding guidance. My special thanks go

to Dr. James W. Jones for his advice and commitment, acute professional challenging,

and always useful feedback during the entire research process. I am also extremely

grateful to Dr. Robert McSorley for his permanent, detailed, and constructive critique;

Dr. Albert De Vries for his valuable inputs and constant assistance; and Dr. Hugh

Popenoe for his interesting and useful comments.

This study would not have been possible without the collaboration of stakeholders,

those representatives of many organizations who were actively involved in it. I would

like to gratefully recognize the involvement and the unconditional help of Darrell Smith,

coordinator of the Suwannee River Partnership (SRP), who was a major actor in the

networking of people and organizations, and supporting the research from the beginning.

I am also very appreciative of the help and dedication provided by Chris Vann, Clifford

Starling, Russ Giesy, and Marvin Weaver, -University of Florida Cooperative

Extension agents, who were the liaison with the dairy farmers. I also recognize the

valuable information, feedback, and help provided by Hugh Thomas (SRP), Justin Jones

(IFAS REC at Live Oak), W.C. Hart (US Department of Agriculture, Lafayette), Joel

Love (SRP), and David Hornsby (Suwannee River Water Management District).









I would like to particularly thank all the dairy farmers, small, medium, and large,

for providing time, data, and critical feedback during the processes of creation,

calibration and validation of the models created in this study. This study was conducted

with the farmers for the farmers. Their participation was crucial to the successful

advancement and culmination of this research. For obvious reasons their names are

omitted.

Besides members of my supervisory committee, many other professionals from the

University of Florida were very helpful, providing information, networking, and inputs in

this study. I would like to recognize the collaboration of Dr. Lynn E. Sollenberger for

information about forage systems; Dr. Jack Van Horn for information about dairy

nutrient management; Dr. Jerry Kidder and Greg Means for information about soil series;

Dr. Donald Graetz for information, feedback, and networking on the soil/water nitrate

issue, Dr. Sabine Grunwald for providing geospatial data; Dr. Clyde Fraisse for feedback

on crop models and climate components; Dr. Carrol Chambliss for information and

feedback on forage systems, Dr. Ken Boote for information on species parameters of crop

models; Dr. Shrikant Jagtap for meteorological information; Dr. Ramon Litell for help

and inputs in statistical analysis; Dr. Norman E. Breuer for information about north-

Florida forage systems and constructive feedback; Cheryl Porter for help running the crop

models and using the DSSAT software, Stuart Rymph for support in creating and

calibrating forage crop models; Marinela Capanu for support with statistical analyses

using SAS software; Daniel Herrera and Sergio Madrid for touring dairy farms with me;

and Amy Sullivan for constructive feedback. Appreciation is extended to Dr. Gerrit









Hoogenboom from the University of Georgia, for his comments on species files for the

crop models.

I want to express my appreciation to the interdisciplinary Sondeo team for their

enthusiastic and productive work in the early stages of this study. Thanks to Diana

Alvira, Sebastian Galindo, Solomon Haile, Kibiby Mtenga, Alfredo Rios, Nitesh Tripathi,

and Jose Villagomez.

I would like to gratefully acknowledge the support and unconditional help provided

by a private enterprise, the Soil and Water Engineering Technology, Inc. My sincere

thanks to Dr. Del Bottcher, Dr. Barry Jacobson, and Suzanne Kish not only for their

direct feedback and first-hand valuable information, but also for inviting me to participate

on the monthly "manure lunch" that was a great opportunity for networking and

collecting information.

I want to express my thanks to Jesse Wilson, Bill Reck, and Nga Watts from the

Gainesville office of the Natural Resources and Conservation Service (US Department of

Agriculture) for providing feedback, constructive critique, and valuable information for

the study.

I hope I am not missing anyone, but if I do, I want to express my appreciation to all

and each one who in some way or another contributed to this study. Your help has been

highly valued and critical for the final outputs.

This work was supported by a grant from NOAA (Office of Global Programs)

through the South Eastern Climate Consortium (a Regional Integrated Science

Application center) under the direction and guidance of Dr. James W. Jones and Dr. Peter

E. Hildebrand.
















TABLE OF CONTENTS
page

A C K N O W L E D G M E N T S ................................................................................................. iv

LIST OF TA BLE S .................................. ........... ............. ................ .. xii

LIST OF FIGURES ......... ....................... ........ ....... ........... xiv

LIST OF ABBREVIATION S ............................................................ .............. xix

ABSTRACT ........ .............. ............. .. ...... ..................... xxi

CHAPTER

1 IN TR OD U CTION ............................................... .. ......................... ..

1.1 B background ........... ......... .................... .. .. .......................
1.2 The Problem atic Situation....................................................... ............... 1
1.3 Research A rea .............. ............... ............... ......................
1.4 O bjectiv es ....................................................... 6
1.5 H ypothesis.............................................. 8
1.6 Fram ew ork ......... ...... ........ ...... ................ ....................... 8
1 .7 M e th o d s ............. .. ......... .............. .. ..............................................1 1
1 .8 R e le v a n c e ...................... .. .............. .. ............................................1 3

2 NITROGEN POLLUTION IN THE SUWANNEE RIVER WATER
MANAGEMENT DISTRICT: REALITY OR MYTH? ........................................14

2.1 Introduction .................................... 14
2.2 Evidence of N Increase in the Suwannee River Water Management
District ................ ...................... ........ ..................... 16
2.2.1 Assessments at the District Level ........................................ ....16
2.2.2 Assessments at the Watershed Level .......................................... 23
2.2.3 Assessments at the National Level ........................................ ...24
2.2.4 Assessments at the Dairy Farm Level .........................25
2.3 Human and Ecological Concerns because of N Pollution .............................29
2.4 Conclusions and Recommendations ......................................... ............. 31









3 NORTH FLORIDA STAKEHOLDER PERCEPTIONS TOWARD
NUTRIENT POLLUTION, ENVIRONMENTAL REGULATIONS, AND
THE USE OF CLIMATE FORECAST TECHNOLOGY .......................................34

3.1 Introduction .............. ...... .... ..... ..... ............ ................ 34
3.2 M materials and M ethods.......................................................... ............... 36
3 .2 .1 Stu dy A rea .................................................................. 36
3 .2 .2 T h e S on d eo ............................................... ................ 3 6
3.2.3 Interview s......................... ......... ....................... ............... 38
3.2 .4 F ocu s G roups ........................ .. ................................ .... ..... ...... 38
3 .2 .5 A n aly sis.............................................. ................ 3 9
3.3 Results and Discussion .............................. ..................39
3.3.1 Stakeholder Perceptions of the N Issue ........ ................................39
3.3.1.1 Perception of water contamination ..............................................39
3.3.1.2 Reasons to mitigate contamination..............................................42
3.3.1.3 Existing and potential means of mitigation ...............................42
3.3.1.4 Implementation procedures for Best Management Practices
(B M P s) ....................................... ........... ............ 45
3.3.1.5 Im pacts of BM Ps on farm s ................................. ................ 47
3.3.2 Perceptions of Regulations by Dairy Farmers ....................................48
3.3.2.1 Farmers in disagreement with regulatory measurements..............49
3.3.2.2 Farmers who agree with regulations but criticize part of it ..........53
3.3.2.3 Farmers who completely agree with regulations .........................56
3.3.3 Perceptions of Climatic Technology Innovation ..................................57
3.3.3.1 Farmers that would use long-term climatic forecast....................57
3.3.3.2 Farmers who would not use long-term climatic forecast .............62
3.3.3.3 Farmers who are not sure they would use long-term
clim atic forecast ........................... ...... ............... .............. 63
3.4 Conclusions and Recomm endations ..................................... ............... 63

4 A METHOD FOR BUILDING CLIMATE VARIABILITY INTO FARM
MODELS: PARTICIPATORY MODELING OF NORTH FLORIDA
D A IR Y FA R M SY STE M S ............................................................. .....................67

4.1 Introduction ................................................... ..................... ..... ....... 67
4.2 M materials and M ethods.......................................................... ............... 71
4.2.1 M ain Fram ew ork.......................................................... ............... 72
4 .2 .2 T he Sondeo ........................ .. ....................... ...... .. ........... 74
4.2.3 Interview s.............................................................................. 75
4.2.4 Focus G roups ................................ ..... ... ...... .. ... .. ............ 76
4.2.5 Other Stakeholder Interactions .................................... ............... 77
4.2.6 A analysis of Inform ation ...................................................................... 77
4.3 R results and D discussion ............................................................................. 78
4.3.1 Phase One: Initial Developments............... ..........................................78
4.3.2 Phase Tw o: The Sondeo........................................... ............... .... 81
4.3.3 Phase Three: Greater Stakeholder Involvement ...................................84
4.3.4 Phase Four: The Prototype M odel .................................. ............... 88









4.3.5 Phase Five: The Final M odel ........................................ .....................93
4.4 Summary, Conclusions, and Recommendations................... .....................95

5 HOW MUCH MANURE NITROGEN DO NORTH FLORIDA DAIRY
F A R M S P R O D U C E ? ........................................................................ ...................98

5.1 Introduction ........... ........ ........ ...............98
5.2 Materials and Methods ............ ............. ............... 102
5.3 Model Development ............. .... ............................... 103
5.3.1 G general Characteristics ................................................................... 103
5.3.2 Markov-Chains, Culling Rates, Reproduction Rates, and Milk
Production R ates .................................................................. ........... .... 103
5.3.3 M anure N Produced by M ilking Cows............. ................................. 108
5.3.4 Nitrogen and Crude Protein .......... ....... .. .... ........ .. .. ............ 110
5.3.5 M ilking Times and Cattle Breeds .......... .. ............... ................. 111
5.3.6 Manure N Produced by Dry Cows, Young Stock, and Bulls..............113
5.3.7 Manure, Feces, Urine, and Dry Matter Intake Estimations ................14
5.3.8 W ater Utilization for M anure Handling...............................................116
5.3.9 W aste M management Handling Systems ...............................................117
5.4 Results and Discussion .......................... ........ .. ................... 119
5.4.1 The Model (Livestock Dynamic North-Florida Dairy Farm
M odel, LDNFDFM ).......................... ........ ..................119
5.4.2 Model Comparison with Other Widely Used Estimations .................120
5.4.3 Estimation of Manure N Produced in North-Florida Dairy Farm
System s ................................... .......................... ... ...... 122
5.5 Conclusions .................................... ............................. .........128

6 FORAGE SYSTEMS IN NORTH FLORIDA DAIRY FARMS FOR
RECYCLING N UNDER SEASONAL VARIATION ................. ... .................132

6.1 Introduction ...................................................................... ......... 132
6.2 Materials and Methods................................... ..................135
6.2.1 Survey, Focus Groups, and Additional Information.............................135
6.2.2 Location of North-Florida Dairy Farms and Their Soil Series............. 135
6.2.3 Climate Information and El Nifio Southern Oscillation (ENSO)
P h ases ............................................... ........ ..... 13 8
6.2.4 Forage Crop System s ................................................................ ...... 138
6.2.5 M anure N A application ........................................ ....... ............... 142
6.2.6. Crop Sim ulations ........................................................................... 142
6.2.7 A analyses ................................... .......................... ... ...... 143
6.3 Results and Discussion ............................. .... ... ............... 143
6.3.1 Forage Crop Systems in North-Florida Dairies .................................. 143
6.3.1.1 Spring-summer crops .................................... ............... 144
6.3.1.2 Sum m er-fall crops................................... ........................ 144
6.3.1.3 Fall-w inter crops .............................................. ......... ...... 145
6.3.1.4 Sequences of forages................................................................. 145
6.3.1.5 H arvest and processing ............................................................ 146









6.3.2 Calibration and Validation of DSSAT Crop Models..........................146
6.3.2.1 Forage sorghum ........................................ ........ ............... 147
6.3.2 .2 P earl m illet............ .... ............................ ...... ...... . ........... 149
6.3.2.3 Silage corn .......................... .. .. ........ ........ ........ ......151
6.3.2.4 W inter forages........................................................................ 152
6.3.3 Sensitivity Analysis of Forage Crops to Manure N Application..........154
6.3.4 Variability by ENSO Phases.................................. ............... 156
6.3.4.1 Nitrogen leaching in different ENSO phases.............................158
6.3.4.2 Biomass in different EN SO phases.............................................166
6.3.5 Variability by Crop Sequences .................................. ............... 168
6.3.5.1 Nitrogen leaching and crop systems ......................................... 170
6.3.5.2 Biomass and crop systems .................................. ........... 171
6.3.6 Variability by Soil Types ................. ..... ...............173
6.3.6.1 Nitrogen leaching and soil types ......... .....................................176
6.3.6.2 Biomass and soil types..... .................................179
6 .4 C on clu sion s ............................................... ................ 18 1

7 AN INTEGRATED SIMULATION MODEL TO ASSESS ECONOMIC
AND ECOLOGIC IMPACTS OF NORTH FLORIDA DAIRY FARM
S Y S T E M S ........................................................................ 18 4

7.1 Introduction ............. ..... .... .... .............................. .......... 184
7.2 Materials and Methods ........... ...... ......... .................. 185
7.2.1 Conceptual M odel ...... .............................................. .. ............... 185
7.2.2 Computer Implementation ............................. .... ............... 188
7.2.3 Analysis of Synthesized Dairies in North-Florida ............................. 188
7.3 M odel D description .................................... ................. ....... ... 189
7.3.1 Overall Characteristics.................... ...... .......................... 189
7.3.2 C om ponents .................................... ................... ....... 190
7.3.3 The Livestock M odel ........................................ ........................ 191
7.3.4 T he W aste M odel ...................................................................... .. .... 197
7.3.5 The Crop M odels ............................................................................ 198
7.3.6 The Econom ic M odule...................................... ........................ 203
7.3.7 The Optim ization M odule.................................. ....................... 204
7.3.8 The Feasible Adjustments Component............................................... 206
7.4 User Friendly Implementation ..... .................... ...............206
7.4.1 General Characteristics ..... ..................... ...............206
7.4.2 The Livestock M odule ................... .. .... .............. ............... 210
7.4.3 The W aste System M odule ................................ ................ 211
7.4.4 The Soil M odule ........................................................ ............... 2 12
7.4.5 The Forage System s M odule ..................................... ............... ..212
7.4.6 The Climatic M odule ................... .. .... ............. ............... 213
7.4.7 The Econom ic M odule...................................... ........................ 214
7.4.8 The Optim ization M odule.................................. ....................... 217
7.5 Results for Synthesized Farms............................ .... ............... 218
7.5.1 Synthesized Small Dairy Operation............................... ..... ......218
7.5.2 Synthesized Medium Dairy Operation.............................. ...............219









7.5.3 Synthesized Large Dairy Operation....................................................221
7.5.4 Comparison of the Three Synthesized Farms ............ .. .............. 222
7 .6 C on clu sion s.......................................................................... ............... 22 5

8 ECONOMIC AND ECOLOGIC IMPACTS OF NORTH FLORIDA DAIRY
FARM SYSTEMS: TAILORING MITIGATION OF NITROGEN
P O L L U T IO N .............................................................................. .. ................ .. 2 2 7

8.1 Introduction ...................................................................... ......... 227
8.2 M materials and M ethods..................................................................................228
8.2.1 Dynamic North-Florida Dairy Farm Model .............. .... ...............228
8.2.2 A Synthesized North-Florida Dairy Farm..........................................229
8.2.3 Statistical A nalyses .................................... ............................ ......... 229
8.2 .4 G raphical A naly ses ................................................................... ....230
8.2.5 O ptim ization A nalyses...................................... ........................ 232
8.2.6 V alidation w ith Real Farm s ...................................... ............... 232
8.3 R results and D discussion ............................................................................ 233
8.3.1 Statistical A naly ses ................................................................... ... ..233
8.3.1.1 N itrogen leaching ............................. .... ... ............... 233
8 .3 .1 .2 P ro fit ...................................................................................... 2 3 4
8.3.2 Graphical Analyses ....... .......................................................... 235
8.3.2.1 Soil series ................................................... .......... 235
8.3.2.2 Crop systems in sprayfields .......................... ...... ......... 238
8.3.2.3 Pastures and crude protein in the diet (CP)...............................240
8.3.2.4 Confined time (CT) and crude protein in the diet (CP) .............242
8.3.2.5 Isolated effects of single managerial changes in the
Synthesized farm ............................................... ............... 244
8.3.3 Optimization Analysis: Decreasing Environmental Impacts and
Increasing Profitability ............................................. ............... 246
8.3.3.1 The Synthesized dairy farm "as is"...........................................246
8.3.3.2 Optimization of the Synthesized farm ............. .....................247
8.3.3.3 Feasible practices for the Synthesized farm..............................249
8.3.4 Examples of Applications of the DNFDFM to Specific Farms............253
8 .3 .4 .1 F arm O n e .......................................................... ....................2 5 3
8.3.4.2 Farm Tw o ............. ................... ....................................... .. ..254
8.3.4 .3 F arm T hree............. ........................ .......... .... .... ........... 256
8.4 Conclusions and Recomm endations ................................... .................257

9 SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS .........................260

L IST O F R E F E R E N C E S ...................................................................... .....................27 1

B IO G R A PH IC A L SK E T C H ........................................ ............................................283
















LIST OF TABLES


Table pge

4-1. Stakeholder interactions, suggestions and changes made to model....................91

5-1. Dry matter intake, manure, and N excretion by Florida dairy cattle based
on milk production. (lbs cow-1 day-) and "low" and "high" (NRC
standards) crude protein (CP) content of diets ................................................109

5-2. Water utilization by milking cows during months and percentage use
relative to m onth of m axim um use (July)........................................................... 117

5-3. Nitrogen lost in the waste management system: a baseline..............................119

5-4. Five hypothetical north-Florida dairy farms and their characteristics for
m anure N estim nations ................................................ .............................. 122

5-5. Estimated amounts of manure N to sprayfields in five hypothetical
north-Florida dairy farm s............................................. ............................ 125

5-6. Estimated amounts of manure N deposited on pasturelands in five
hypothetical north-Florida dairy farms ......................................................126

5-7. Estimated amounts of manure, DMI, and water use in five hypothetical
north-Florida dairy farm s............................................. ............................ 128

6-1. Soil types, some characteristics, and their sources of information used for
the study ....................................................................... ..........137

6-2. Information sources for calibration and validation of forage crops in
north-Florida dairy farm systems.................................. ..............148

6-3. Coefficients values and coefficients definitions of modified crops in
D SSA T ................................................................ .... ..... ......... 150

6-4. Nitrogen leaching (kg ha-1) and ENSO phases, when all other factors are
averaged ............................................................... ... .... ......... 160

6-5. Biomass (kg ha-1) and ENSO phases, when all other factors are averaged......... 166

6-6. Nitrogen leaching (kg ha-1) and crop sequences, when all other factors are
averaged ............................................................... ... .... ......... 171









6-7. Biomass (kg ha-1) and crop sequences, when all other factors are averaged.......172

6-8. Nitrogen leaching (kg ha-1) and soil types, when all other factors are
av eraged ..................................................... ............................ 178

6-9. Biomass (kg ha-1) and soil types, when all other factors are averaged..............180

7-1. State variables and monthly update of the livestock model..............................194

7-2. D ry m atter intake estim nations ..................................................... ... .......... 195

7-3. M anure excretion estim nations ..................................................... ... .......... 196

7-4. N itrogen excretion estim nations ........................................ ........................ 197

7-5. Nitrogen through waste system estimations ................................ ............... 198

7-6. Codification of ENSO phases, soil types, and forage systems from crop
m o d els ...................................... ................................................ 2 0 0

7-7. Nitrogen applied estimated by the waste model and N applied in crop
m o d els ...................................... ................................................ 2 0 2

7-8. Definition of variables in economic module (US$ cwt-1 milk-1)..........................203

7-9. Monthly milk prices distribution in north-Florida as cumulative
probability (U S$ cw t-1 m ilk-1) ................................................... ........ ....... 204

7-10. C rops in different fields ............................................................ .....................2 18

7-11. Comparison of three synthesized north-Florida farms and their estimated
annual N leaching ...................... ...... .......... ............... .... ....... 224

8-1. Characteristics of a Synthesized north-Florida dairy farm ...............................230

8-2. Combination of management strategies simulated by the DNFDFM for the
statistical an aly sis...................................... ................. ....................... 2 3 1

8-3. Significance of factors and main interactions for N leaching.............................234

8-4. Significance of factors and main interactions for profit .....................................235

8-5. Crops in different fields in the Synthesized farm .............................................246

8-6. Management strategies selected by the optimizer ............................................248

8-7. Feasible crop systems in Synthesized dairy farm ..........................................250















LIST OF FIGURES


Figure pge

1-1. N orth-Florida landscape representation........................................ .....................3

1-2. The Suwannee River Water Management District, the study area........................

2-1. Nitrate-N concentration in the SRW M D .................................... ............... 17

2-2. Nitrate-N concentrations in three springs in the middle Suwannee River
B a sin .............................................................................. 19

2-3. Estimated annual mean of groundwater N load ...............................................20

2-4. Long term records of Nitrate-N for 4 Suwannee springs.................. ............22

2-5. Long term records of Nitrate-N for 3 Lafayette springs.............. ... .............23

2-6. Nitrate-Nitrite and total ammonium plus organic N at different points in
nine dairy farms in the middle Suwannee River Basin........................................26

2-7. Dissolved Nitrate+Nitrite and Ammonium+Organic N in monitoring wells
(3-m deep) in different areas of a dairy farm.............. ...... ...............27

2-8. Monitoring wells with higher than drinking water allowable nitrate levels
(PDWS>10 mg L-1) in the four dairy farms .............................................. 29

3-1. Farmers' perceptions of environmental regulations in north-Florida.................50

3-2. Farmers' perceptions of climate technology adoption........................................58

4-1. Schematic of participatory modeling ...... ......... .....................................72

4-2. M ethodology of participatory modeling .................................... .... ..... .......... 73

4-3. Phases of the participatory modeling for creating the DNFDFM..........................78

5-1. Pregnancy rates of adult cow s................................................... .....................105

5-2. Culling rates of adult cow s ............................................................................ 106

5-3. M ilk production rates of milking cows............... ............................................. 107









5-4. Manure N produced by low-CP diet cows according to milk production ...........111

5-5. M anure excretion based on milk production ....................................................... 114

5-6. Dry matter intake (DMI) based on milk production ................. ................115

5-7. Nitrogen excretion by a "low" protein diet in a 100-cow dairy estimated
by different m models .................. ............................ ........ .. ............ 120

5-8. Nitrogen excretion by a "high" protein diet in a 100-cow dairy estimated
by different m models .................. ............................ ........ .. ............ 121

6-1. Study area, dairy farm s, and soil types ..................................... ............... ..136

6-2. Daily rainfall (1956-1998) in north-Florida for different ENSO phases ...........139

6-3. Distribution of monthly precipitation (1956-1998) in north-Florida for
different EN SO phases............................................... .............................. 140

6-4. Daily precipitation in north-Florida (1956-1998) for different ENSO
phases ............................................................... .... ..... ......... 140

6-5. Daily temperature (1956-1998) in north-Florida for different ENSO phases......141

6-6. Daily solar radiation (1956-1998) in north-Florida for different ENSO
phases ............................................................... .... ..... ......... 141

6-7. Forage systems and their seasonality in north-Florida dairy farms.....................144

6-8. Forage sorghum: Observed and simulated biomass production, Bell,
1996-1998. Individual mean square error (MES) and overall Root Mean
Squ are E rror (R M SE ) ............................................................... ..................... 149

6-9. Forage millet: Observed and simulated biomass production, Bell,
1996-1998. Individual mean square error (MES) and overall Root Mean
S qu are E rror (R M SE ) ............................................................... ..................... 15 1

6-10. Forage millet: Observed and simulated biomass production, Quincy 1992........152

6-11. Silage corn: Observed and simulated biomass production, Bell,
1996-1998. Individual mean square error (MES) and overall Root Mean
S qu are E rror (R M SE ) ............................................................... ..................... 153

6-12. Winter forages: Observed and simulated biomass production, Bell
1997-1998. Individual mean square error (MES) and overall Root Mean
Square E rror (R M SE) ................................................ .............................. 154









6-13. Sensitivity response of N leaching and biomass accumulation to different
rates of manure N applied for simulations in a Kershaw soil series during
the season 1996/1997 .................................... .............. .. ..... .. ........ .. .. 155

6-14. Nitrogen leaching vs. biomass accumulation for different ENSO phases
under two forage systems consisting of bermudagrass and millet-sorghum,
four manure N applications (10, 20, 40, and 80 kg ha-1 mo-1), and two type
of soils (Arredondo-Gainesville-Millhopper and Millhopper-Bonneau) ............157

6-15. Nitrogen leaching vs. ENSO phases relative to absolute averages, when all
other factors are averaged ........................................................ ............. 162

6-16. Probability distribution of N leaching by ENSO phases, when all other
factors are averaged ............................................ .. .... .... ............... 163

6-17. Frequency distribution of N leaching and month of the year by ENSO
phase, when all other factors are averaged .................................. ............... 165

6-18. Biomass accumulation vs. ENSO phases relative to absolute averages,
when all other factors are averaged ........................................ ............... 167

6-19. Nitrogen leaching vs. biomass accumulation for different forage systems
under different ENSO phases (La Nifia, Neutral, and El Niho), two manure
N applications (40 and 80 kg ha-1 mo-1), and soils of type Penney-Kershaw...... 169

6-20. Nitrogen leaching vs. crop sequences relative to absolute averages, when
all other factors are averaged ...................................................... ........ ...... 170

6-21. Biomass accumulation vs. crop sequences relative to absolute averages,
when all other factors are averaged ........................................ ............... 173

6-22. Nitrogen leaching vs. biomass accumulation for different soil types under
different ENSO phases (La Nifa, Neutral, and El Niho), two manure N
applications (40 and 20 kg ha-1 mo-1), and forage system consisting of
corn-sorghum -w inter forage ........................................... ......................... 175

6-23. Nitrogen leaching vs. soil types relative to absolute averages, when all
other factors are averaged ........ ................... ................. ... ............... 177

6-24. Nitrogen leaching vs. soil types relative to absolute averages, when all
other factors are averaged ........ ................... ................. ... ............... 181

7-1. Representation of the dynamic north-Florida dairy farm model ....................191

7-2. Representation of the livestock model .......... .. ............ ........... ........ ........... 192

7-3. Data management of simulation of crop models .............................................201









7-4. The process of optim ization ..................................................................... ..... 205

7-5. Main screen of the computer implementation of The Dynamic
North-Florida Dairy Farm Model (DNFDFM).............................208

7-6. The start message menu control ...............................................209

7-7. The livestock module control.......................... ......... .................... 211

7-8. The w aste m anagem ent control ........................................ ....................... 212

7-9. Dairy farm location and soil type by farm component control..........................213

7-10. Forage systems component control.................. ... .... ................. 214

7-11. The clim atic com ponent control ............................................... ............... 215

7-12. The economy ic com ponent control.......... .................................... ..... ............. 216

7-13. Distribution of milk prices in north-Florida................................................... 16

7-14. The optim ization control ......... ................. ................................. ............... 217

7-15. Monthly estimations for Synthesized Small north-Florida farm .......................220

7-16. Small north-Florida farm, yearly estimated N leaching and profit................221

7-17. Medium north-Florida farm, yearly estimated N leaching and profit..................222

7-18. Large north-Florida farm, yearly estimated N leaching and profit................223

7-19. Overall N leaching from the synthesized farms ......... ................................... 225

8-1. Nitrogen leaching and profit for different soil types in north-Florida:
Estimations for different ENSO phases for common crop systems.....................236

8-2. Nitrogen leaching and profit for different crop systems in north-Florida:
Estimations for different ENSO phases for common pastures and soil
types ............................................................... ..... ..... ........ 239

8-3. N leaching and profit for different pasture crops and crude protein in
north-Florida: Estimations for different ENSO phases for common
sprayfield crops and soil type ........................................ ......................... 241

8-4. N leaching and profit for different confined time and crude protein:
Estimations for different ENSO phases for common crop systems and soils......243

8-5. Impacts of management changes for the Synthesized farm with soils type
4 (Penney-O tela) ...................... .................. ................. .... ....... 245


xvii









8-6. Nitrogen leaching and profit with current practices for a Synthesized
north-F lorida farm ............ ............................................................. .... .... .... .. 247

8-7. Optimized vs. current practice yearly N leaching and profit for a
north-Florida Synthesized farm ........................................ ....................... 249

8-8. "Feasible" yearly N leaching and profit for a Synthesized north-Florida
farm ........................................................ ............... ........... 251

8-9. Optimization of a Synthesized north-Florida farm, scenario One.......................251

8-10. "Second feasible" yearly N leaching and profit for a north-Florida
Synthesized farm.................... ..................... ............ 252

8-11. Farm One: N leaching and profit for different management strategies .............254

8-12. Farm Two: N leaching and profit for different management strategies.............255

8-13. Farm Three: N leaching and profit for different management strategies...........257


xviii
















LIST OF ABBREVIATIONS


AFO

ASAE

AWMFH

BMPs

CAFO

CARES

CP

DBAP

DHIA

DNFDFM

DM

DMI

DRMS

ENSO

EPA

FDH

FFB

FGS

FLDEP

FSR

FSRE

IC

LDNFDMF


Animal Feeding Operations

American Society of Agricultural Engineers

Animal Waste Management Field Handbook

Best Management Practices

Concentrated Animal Feeding Operation

County Alliance for Responsible Environmental Stewardship

Crude Protein

Dairy Business Analysis Project

Dairy Herd Improvement Association

Dynamic North-Florida Dairy Farm Model

Dry Matter

Dry Matter Intake

Dairy Records Management Systems

El Nifio Southern Oscillation

Environmental Protection Agency

Florida Department of Health

Florida Farm Bureau

Florida Geological Service

Florida Department of Environmental Protection

Farming Systems Research

Farming Systems Research and Extension

Initial Conditions for Simulation

Livestock Dynamic North-Florida Dairy Farm Model










LN

MCL

MHB

MIM

MIP

NAC

NAWQA
NH, -Org
No-_N2-
NO3- NO

NOAA

NRC

NRCS

NRM

PDWS

PRA

RHA

RRA

SRP

SRWMD

SWIM

TWG

USDA

USEPA, EPA

USGS

WARN

WATNUTFL


Lactation Number

Maximum Concentration Load

Methemoglobinemia

Months in Milk

Months in Pregnancy

Number of Adult Cows

National Water Quality Assessment

Ammonium + Organic N

Nitrites + Nitrates

National Oceanic and Atmospheric Administration

National Research Council

Natural Resource and Conservation Service

Natural Resource Management

Primary Drinking Water Standard

Participatory Rural Appraisal

Rolling Herd Average

Rapid Rural Appraisal

Suwannee River Partnership

Suwannee River Water Management District

Surface Water Improvement and Management

Technical Working Group

US Department of Agriculture

US Environmental Protection Agency

US Geological Service

Water Assessment Regional Network

Water budget and Nutrient balance for Florida















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

MODELING NORTH FLORIDA DAIRY FARM MANAGEMENT STRATEGIES TO
ALLEVIATE ECOLOGICAL IMPACTS UNDER VARYING CLIMATIC
CONDITIONS: AN INTERDISCIPLINARY APPROACH

By

Victor E. Cabrera

August 2004

Chair: Peter E. Hildebrand
Major Department: Natural Resources and Environment

High levels of nitrogen (N) in water affect human health and ecosystem welfare.

Dairy waste is considered an important contributor to increased water N levels in the

Suwannee River Basin. The main aim of the study was to assess management options that

decrease N leaching in dairy systems under different climatic conditions, while

maintaining profitability. Whole-dairy-farm system simulations were used to study

economic and ecologic sustainability, which would be impractical to test with regular

experiments. Dynamic, event-controlled, and empirical models were coupled with

process-based and decision-making models to represent and test management strategies.

User-friendliness and stakeholder interaction were always pursued. Eight focus groups

and 24 dairy farmer interviews were conducted, and a collaborative user-friendly decision

support system was created. The Dynamic North-Florida Dairy Farm Model (DNFDFM)

uses Markov-chains to simulate probabilistic livestock reproduction, culling, milk

production, and N excretion; process-based crop models to simulate forage production, N









uptake, and N leaching; and linear programming optimization to tailor management

strategies. The DNFDFM is an environmental accountability tool for dairy farmers and

regulatory agencies capable of simulating any individual north-Florida dairy farm.

Results indicated that north-Florida dairy farmers could maintain profitability and

decrease N leaching 9 to 25% by decreasing crude protein in the diet, adjusting confined

time of milking cows, and adjusting sequences of seasonal forages. During "El Niho"

years, when higher rainfall and colder temperatures are experienced, 5% more N leaching

is estimated compared with "Neutral" years; and 13% more compared with dryer and

hotter "La Nifia" years.














CHAPTER 1
INTRODUCTION

1.1 Background

The presence of high levels of N in water is an environmental hazard because it

affects human health and ecosystem welfare. The Suwannee River Basin has received

much attention in recent years because of increased N levels in water bodies. Dairy waste

is thought to be an important factor contributing to this water N pollution.

Dairy farmers are now required to comply with stricter environmental regulations

either under permit or under voluntary incentive-based programs. Dairy farmers are also

aware that environmental issues will be among the greatest challenges they face in the

near future.

Evidence indicates that farms may reduce their total N loads by changing some

management strategies. Improvements in long-term seasonal predictions (6 to 12 months)

such as in El Nifio and La Nifia forecasts, can play an important role in implementing

management strategies that dairy farmers in north-Florida may adopt in order to pursue

economic and ecological sustainability.

1.2 The Problematic Situation

Dairy farming is an important part of Florida's agricultural industry. The Florida

Statistical Service showed that milk and cattle sales from dairies contributed $429 million

directly into the Floridian economy in the year 2001. Florida is the leading dairy state in

the Southeast; it ranks 13th nationally in cash receipts for milk, 15th in milk production

and 15th in number of cows (Bos taurus) (Florida Agricultural Statistics Service, 2003).









According to USDA, there were about 152,000 cows on about 220 dairy farms at the end

of 2002, and more than 30% of the dairy operations (65 operations) and number of cows

(> 45,000 cows) of the State of Florida are located in the Suwannee River Basin. Indeed,

more than 25% of the dairy cows in the state are in four counties in north central Florida

that bound the Suwannee River (Bisoondat et al., 2002): Gilchrist, Lafayette, Suwannee,

and Levy.

This dairy industry, according to Giesy et al. (2002), faces three significant

challenges: production, economic, and environmental. Dairymen perceive that

environmental issues will be the most stressful in the near future (Staples et al., 1997).

Dairies face increased regulation because social pressure, while larger herds attract the

attention of neighbors and activists concerned with odors, flies, and especially potential

leaching of nutrients that might influence water quality (Giesy et al., 2002).

The presence of N in surface water bodies and groundwater aquifers is recognized

as a significant water quality problem in many parts of the world (Fraisse et al., 1996).

The Suwannee River Partnership states that over the last 15 years, nitrate levels in the

middle Suwannee River Basin have been on the increase; and these elevated nitrate levels

can cause health problems in humans as well as negative impacts on water quality

(Suwannee River Partnership, 2004a). Indeed, the Suwannee River Basin has received

much attention in recent years because increased N levels in the groundwater-fed rivers

of the basin could seriously affect the welfare of the ecosystem (Albert, 2002). According

to Katz (2000), N levels have increased from 0.1 to 5 mg L-1 in many springs in the

Suwannee basin over the past 40 years. Pittman et al. (1997) found that nitrate

concentrations in the Suwannee River itself have increased at the rate of 0.02 mg L-1









year1 over the past 20 years. In a 53.1 km river stretch between Dowling Park and

Branford, the nitrate loads increased from 2,300 to 6,000 kg day-1. They also estimated

that 89% of this came from the lower two-thirds, where agriculture is the dominant land

use. Soils in this region are generally deep, well-drained sands (Figure 1-1); and nutrient

management is a major concern (Van Horn et al., 1998).


.4







rg ir in n n n n r r ,
..... "",,,^ W t ..


-W ~ jf -, --. "_ < a _- .








Figure 1-1. North-Florida landscape representation. Source: Katz et al., 1999, p. 1.

As Conway (1990) stated, farms in industrialized countries have become larger in

size, fewer in number, highly mechanized, and highly dependent on external inputs; and

can cause environmental problems by producing a number of wastes that might be

hazardous. Recent evidence that agriculture in general, and animal waste in particular,
may be an important factor in surface and groundwater quality degradation has created a

strong interest in nutrient management research (Katz, 2000, Van Horn, 1997, Fraisse et









al., 1996). Over-applying manure nutrients to land is considered a major cause of nitrates,

converted from manure ammonia sources while in the soil, leaching to groundwater and

contributing to surface runoff of N that contaminates surface waters. Concerns regarding

nutrient losses from manure of large dairy herds to groundwater or surface runoff have

been extremely acute in Florida (Van Horn, 1997). One of the most publicized concerns

is N losses in the form of nitrate into the groundwater through the deep sandy soils of the

Suwannee River Basin (Van Horn et al., 1998). Many soils in the Suwannee River Basin

are excessively drained, deep, sandy, Entisols with low exchange capacity and low

organic matter, which are prone to nitrate leaching (Woodard et al., 2002).

Environmental regulation of livestock waste disposal has become a major public

concern, and much of the focus for this issue in Florida has been on dairies. Dairymen are

now required to develop manure disposal systems in order to comply with Florida

Department of Environmental Protection water quality standards (Twatchtmann, 1990).

This fact has led to considerable research emphasizing N recycling and addressing issues

such as maximum carrying capacity and nutrient uptake by crops (Fraisse et al., 1996).

Dairymen in the Suwannee River Basin have expressed their willingness to

participate in initiatives that promote reduced environmental impacts. In fact, many of

them are already involved in programs such as Better Management Practices (BMPs) to

reduce nitrate overflow promoted by the Suwannee River Partnership (Smith, 2002, Pers.

Comm.). Staples et al. (1997) interviewed 48 dairy farmers in Lafayette, Suwannee,

Columbia, Hamilton, and Gilchrist counties. By far farmers rated their perception of

anticipated costs of complying with upcoming environmental regulations as the top

challenge to successful dairying in the future.









Overflow of nitrates in these dairy systems is affected by adopted management

practices and by environmental conditions. Changing management practices might have a

great impact on overflow amounts. For instance, Albert (2002) found that by reducing the

duration of the irrigation, reducing the fertilizer application rate, and improving the

timing of fertilizer applications, N leaching can be reduced by approximately 50% while

maintaining acceptable potato yields in sandy soils in north-Florida. It is unknown if

similar large reductions are also possible in dairy farms. Many environmental conditions

are also determined by climate factors such as rainfall, temperature, and radiation (which

are variable and uncertain). Dairy operations are affected by changes in climatic

components at many different levels, and in many different ways. Rainfall patterns might

result in very different amounts of nutrient leaching. For example, higher temperature in

winter can promote more biomass accumulation, and increase plant N uptake.

Currently, better-quality long-term seasonal forecasts (6 to 12 months in advance)

are available. The National Oceanic and Atmospheric Administration (NOAA) can

predict-using the El Nifio Southern Oscillation (ENSO) technology-accurate

long-term seasonal climate patterns. These are based on predicting El Niho, La Nifia, or

Neutral years (which determine rainfall and temperature patterns). The ENSO signal in

Florida has been well documented, in terms of its effects on climatic variables (Jones et

al., 2000). Precipitation in Florida is particularly variable, with an excess of over 30% of

the normal seasonal total across much of the state during an El Nifio winter. La Nifia has

the opposite effect, with deficits of 10 to 30% lasting from fall through winter and spring.

Florida and the Gulf Coast can expect to see average temperatures 2 to 30F below normal

during El Nifio years. La Nifia has the opposite effect, with temperatures 20 to 40F above









normal during winter months. La Nifia's effect on temperature is more pronounced in

north-Florida, Alabama, and Mississippi. Solar radiation decreases in association with

higher rainfall during El Nifio years. Notice also that climatic conditions in north-Florida

allow year-round crop production, thus allowing more effluent-waste application,

compared with dairy production systems in temperate areas (Bisoondat et al., 2002).

Proactive use of seasonal prediction can be adopted by dairy farmers. The

opportunity to benefit from climate prediction arises from the intersection of human

vulnerability, climate prediction, and decision capacity (Hansen, 2002a). This benefit

from the information can be used to decrease environmental impacts, while maintaining

profitability in north-Florida dairy operations.

1.3 Research Area

The study was performed inside the Suwannee River Water Management District,

which encompasses all or part of 15 counties in north central Florida (Figure 1-2). The

study was centered in five counties where N pollution is more problematic: Alachua,

Levy, Gilchrist, Lafayette, and Suwannee.

1.4 Objectives

The main aim of the study was to assess and test management options that decrease

nutrient overflow under different climatic conditions in north-Florida dairy operations

while maintaining profitability. To achieve this, it was necessary to

* Thoroughly understand the north-Florida dairy system through farmer interactions
Conceptualize north-Florida dairy farms as systems
Identify dairy farm system components
Recognize interaction among these dairy farm system components

* Create a comprehensive dairy farm model adaptable to north-Florida dairies









* Simulate north-Florida dairy farms under different environmental conditions,
different management strategies, and different climate conditions

* Identify climate-management interactions that result in lower nutrient pollution,
especially less N leaching, while maintaining acceptable profit levels

* Suggest management strategies that individual north-Florida dairies could follow
based on climate forecast

* Based on previous strategies, produce an extension circular with a bullet list of results
to be available on paper and on the Internet for north-Florida dairy farmers and other
stakeholders

* Create a user-friendly computer application with the simulation models that
north-Florida dairy producers and other stakeholders can use for environmental
accountability


Figure 1-2.


The Suwannee River Water Management District, the study area. Source:
Land Use Survey, Suwannee River Water Management District, Florida
Geographic Data Library, 1995. Note: Dots represent dairy farms and
numbers indicate the number of dairy farms by county.









1.5 Hypothesis

North-Florida dairy farms can decrease their environmental impacts and still

maintain profitability by changing management strategies, which include the use of

improved long-term seasonal forecasts. This hypothesis can also be expressed as a

research question: Can north-Florida dairy farms maintain profitability and decrease their

environmental impacts by combining management strategies and seasonal forecasts?

1.6 Framework

Most dairies in Florida manage animals under semi-intensive or intensive systems.

Florida dairies are principally in the business of producing milk: decisions are based on

profit maximization and operation regulations. North-Florida dairies are business

enterprises and have full access to credit opportunities, information, and new

technologies (Adams, 1998). Additionally, dairy farms (which produce both livestock

and crops) are composed of interconnected units; any part of which is affected by and

affects the behavior of other parts, including human management decisions and

strategies.

Since a whole-dairy-farm encompasses environmental, economic, and bio-physical

components, its analysis is best accomplished by a systems approach (that accounts for

the interactions of these components, and that can trace the consequences of an

intervention through the entire system) (Kelly, 1995). While systems can be

conceptualized at any of many scale levels, a systems approach defines a specific scale of

interest and an appropriate boundary of analysis. Analysis in the systems approach is

marked by recognition of the whole system and the interactions within that system rather

than looking only at a system component. A systems approach uses specific techniques

and tools (such as rapid appraisal, pattern analysis, diagrams, and modeling, often in a









multidisciplinary fashion) to identify system boundaries and recognize component

interactions (Kelly, 1995).

The proposed systems approach was Farming Systems Research and Extension

(FSRE) as described by Hildebrand (1990). The "farm household" in livelihood systems

analyses are replaced by "farm management" (Hazell and Norton, 1986) for commercial

or industrial farm dimensions such as north-Florida dairy farms. The farming system is an

arrangement of subsystem components that function as units. These subsystems include

economic subsystems and a number of agroecosystems. The agroecosystems, as

described by Powers and McSorley (2000), include the populations and communities

organized in ecosystems in relationship with the abiotic factors.

To assess what would happen to the whole dairy operation and what would happen

specifically to nitrate flow under different climate-management scenarios, models of

these realities were set up (Thomley and Johnson, 1990). Models allow us to foresee

results, when temporal or material conditions make it impractical to develop field

experiments. Successful planning of an animal waste management system requires the

ability to simulate the impact of waste production, storage, treatment, and utilization on

the water resources; it must address overall nutrient management for the operation

(Fraisse et al., 1996).

Different simulation methods were needed to accurately represent the north-Florida

dairy crop and livestock systems; and to test management strategies that yield the most

desired outcomes. Whole farm scale simulation models include biophysical,

environmental, and economic components and relationships of the dairy operation to

climate variability. Complex systems demand the study of interactions among processes









acting on multiple scales. Thus, these problems require innovative approaches,

combining advances in dynamical theory, stochastic processes, and statistics (Levin,

1991).

Budgeting nutrient flow through the total dairy farm system as conceptualized by

Van Horn et al. (2001, 1998, 1994, 1991); NRCS (2001), Van Horn (1997); and Lanyon

(1994) were adjusted, modified, and used for keeping track of dynamic N movement in

all system components. Van Horn et al. (1998) mentioned the quantitative information

needed on nutrient flow through all segments of the system. They listed the following

critical research questions:

* How much of individual nutrients are excreted by dairy cows?

* How does the manure management system affect nutrient flow and nutrient
recoveries for fertilizer use?

* What is the potential nutrient removal by plants?

* How can a manure nutrient budget be developed?

* What steps help to document environmental accountability?

Agronomic measures of nutrient balance, and tracking of inputs and outputs for

various farm management units, can provide the quantitative basis for management to

better allocate manure to fields, to modify dairy rations, or to develop alternatives to

on-farm manure application (Lanyon, 1994).

Similarly to budgeting nutrients, The Netherlands has implemented the Mineral

Accounting System (MINAS) which focuses on nutrient (N and phosphorus) flows on

individual farms; and taxes farms whose nutrient surplus exceeds a defined limit. MINAS

embodies a new approach to environmental problems caused by agriculture. According to

Ondersteijn et al. (2002) focusing on individual farmers has two major advantages. First,

individuals are considered polluters and are individually held accountable for their









pollution, according to the "polluter pays" principle. Second, individuals have control

over their pollution problem, and will be able to deal with it on an individual level

(instead of being forced to comply with general measures that may be ineffective for their

specific situation). After interviewing 240 farmers during 3 years, they also found large

variation in N and phosphate surpluses within and between land-based farm types in The

Netherlands. They surveyed farmers stated that MINAS provided enormous and often

surprising insight into their management.

Crop models (Jones et al., 2003, Jones et al. 1998) are process-based dynamic

simulation models that allow biophysical and environmental conditions to be translated

into agricultural outcomes (Philips et al., 1998). Crop models can be used to estimate

monthly biomass accumulation and N leaching (by a number of crop and/or pasture

sequences under different climatic conditions, depending on soil conditions, manure

applications, and other managerial choices).

Linear programming (LP) models are optimization models capable of devising

alternative management choices that maximize (minimize) an objective function

according to a set of restrictions (Hazell and Norton, 1986). The LP models have been

widely used for analyzing farming systems.

1.7 Methods

North-Florida dairy farmers were involved in the entire research process. They

provided the base information, and they validated results of the analyses. Few studies to

date have incorporated such first-hand data from interviews and validation interaction

with dairy managers (which are believed extremely important). Other stakeholder

agencies also participated actively as was the case for the Suwannee River Partnership

(that involves 45 agencies in total); the Florida Cooperative Extension Service; and some









private enterprises. These stakeholders provided relevant information and permanent

feedback to the study. In all interactions, a laptop computer was used. There were two

rounds of interviews and focus groups. For the first round, a "prototype" model was used

(created by suggestion of one stakeholder) as a discussion, engagement, and

data-collection tool. In both the interviews and focus groups, the interaction started by

explaining the prototype model and demonstrating its functioning, followed by an

interactive discussion of all specific points. The researcher recorded all suggestions and

information. Some modifications were made during the process to the "prototype" model,

however most of the suggestions were incorporated into a new model, which was then

used in the second round. In the interviews of dairy farmers, an additional 1 to 2 hour

conversation was used to collect information regarding biophysical farm factors,

livestock, livestock feeding program, crops, crop fertilization program, waste

management, perception of environmental regulations, perception of seasonal forecast

use, and other factors. In several occasions after the interview, a tour of the farm

followed. There were 21 interviews and six focus groups during this first round.

A second round of focus groups and interviews was arranged to validate the

functioning of the model. Two focus groups and three interviews confirmed that the

model was performing inside the expected ranges, and that results were realistic. Local

published data from related studies were also used as baseline in the modeling.

An integrated, whole-dairy-farm simulation that links decision making with

biophysical, environmental, and economic processes was developed, the Dynamic

North-Florida Dairy Farm Model (DNFDFM). The simulation approach included a

Dynamic event-controlled simulation model (D, driver) connected to crop models (CM)









and to optimization models (OM). The driver, which had many modules, followed a

dynamic adaptation of the budgeting framework for tracking the flow of N through the

entire system; and for tracking production and additional economical variables, according

to a pre-defined set of management and climatic conditions. Crop models assessed the N

recycled by plants and the amount of N leached according to management, genetic, soil,

and climatic specific conditions. The optimization models consisted of linear

programming models that either minimize N leaching, or maximize profit for specific

farm conditions. The DNFDFM runs monthly for all the processes following

user-friendly menus.

1.8 Relevance

Results are expected to directly help Suwannee River Basin dairy operations by

providing a decision tool that can be used to decrease environmental impacts by reducing

nitrate pollution (currently one of the most important environmental concerns). Results

should also contribute to overall Suwannee River Basin environmental health by reducing

the total N load in the watershed, which has been rapidly increasing.

Innovative frameworks and methodologies used in this research could serve as

baselines for other studies in which interdisciplinary and participative outcomes are

desired. For example a similar approach could be used to address the phosphorus problem

in the Okeechobee area.














CHAPTER 2
NITROGEN POLLUTION IN THE SUWANNEE RIVER WATER MANAGEMENT
DISTRICT: REALITY OR MYTH?

2.1 Introduction

In response to increased environmental concerns, the U.S. Environmental

Protection Agency (USEPA) has increased regulatory controls for concentrated animal

feeding operations (CAFO) under the Clean Water Act. One of the changes is to require

CAFOs that apply manure to land to meet application limits defined by a nutrient

standard (USEPA, 2003). The goal of the nutrient standard is to minimize nutrient loss

from fields receiving manure. Farms meet this standard by applying manure nutrients at a

rate consistent with the agronomic needs of crops receiving manure. Nutrient standards

can be N -or P-based, depending on the nutrient content of the soil. USDA is also

encouraging the voluntary adoption of nutrient standards on all animal feeding operations

(AFO) not subject to USEPA regulation (USDA, NRCS, 1999).

The Suwannee River Partnership1 (SRP) states that over the last two decades,

nitrate levels in the middle Suwannee and Santa Fe river basins have been on the

increase. The SRP is a coalition of state, federal and regional agencies, local

governments, and private industry representatives with 46 partners working together to

reduce nitrate levels in the surfacewaters and groundwater within the basins, or

watersheds. The Suwannee River Basin has received much attention in recent years

because increased N levels in the groundwater-fed rivers of the basin could seriously


1 www.mysuwanneeriver.com/features/suwannee+river+partnership









affect the welfare of the ecosystem (Albert, 2002). The federal safety standard for

nitrate-N in drinking water is 10 mg L-1 (Florida Department of Health, FDH, and Florida

Department of Environmental Protection, FLDEP). According to Hornsby (pers. comm.),

a water quality analyst from the Suwannee River Water Management District (SRWMD)

at Live Oak, nitrate levels in some areas have reached ranges of 20 to 30 mg L1.

Rising nitrate levels in north-Florida have been a growing concern because of the

effect of nutrients on public health and the environment (Van Horn et al., 1998). Hornsby

et al., 2002b alleged that the most likely contributors to contamination are farms, industry

and human settlements in the region. Soils in this region are generally deep, well-drained

sands; and one of the most critical concerns is N loss in the form of nitrate into the

groundwater through these deep sandy soils of the Suwannee River Basin (Van Horn et

al., 1998).

Changes in agricultural land use and the intensification of this land use during the

past 50 years have contributed variable amounts of N to the groundwater system. In

general, four main sources of N to groundwater can be recognized: fertilizers applied to

cropland, animal wastes from dairy and poultry operations, atmospheric deposition, and

septic tank effluents (FLDEP, 2001).

Despite the above arguments, it was found (by talking with local people) that some

stakeholder groups believe that either the problem of N pollution is not as bad as many

people perceive, or the problem does not exist at all. Therefore there is a feeling that the

regulatory agencies are overacting and pushing excessively in protecting the

environment.









Under these circumstances, it is critical to clearly understand the problem and the

concerns. Clarifying the real problem and the real actors would help in promoting

realistic ways of solving it. The objective of this chapter is to discuss the evidence of

increasing levels of N in water bodies in the Suwannee River Basin, explore evidence of

N pollution by dairy farms, and briefly address potential ecological and human

implications.

2.2 Evidence of N Increase in the Suwannee River Water Management District

2.2.1 Assessments at the District Level

The Suwannee River Water Management District (SRWMD) is the water authority

in the study area. The SRWMD deals with quantity and quality of water. Regarding the

quality of the water, it permanently monitors the quality of the ground and surface water,

and every year produces a comprehensive report. The last available report is for the year

2001/2002 (Hornsby et al., 2003).

The SRWMD has developed a comprehensive Water Assessment Regional

Network (WARN) consisting of 97 Floridian aquifer wells for a trend network and 120

wells for a status network, distributed along the water management district. These are

used to assess the groundwater quality in the District. In this last report (Hornsby et al.,

2003) containing results only for the status wells between October 2001 and September

2002, nitrate-N concentrations above a background concentration of 0.05 mgL-1 existed

for all or part of every county in the District. The possible nitrate sources were stated as

fertilizers, animal waste, and human waste. Concentrations are elevated in the Middle

Suwannee River Watershed (Suwannee and Lafayette Counties); and within an area

generally defined along an axis starting in northern Jefferson County and ending in

eastern Alachua County (Figure 2-1). However, the three samples that exceeded the










Maximum Concentration Load (MCL) or the Primary Drinking Water Standard (PDWS)

set at 10 mg L-1, were located in Levy County during May and June 2001 (10.6, 12.0, and

14.2 mg L-1). By comparing overall data with the other four water districts in Florida, the

SRWMD has the largest area of elevated groundwater nitrate-N concentrations in the

State.


Nitrate-nitrogen (mg/L)
is a measure of the amount
0 5to 1 0 of nitrate dissolved in the
ground water expressed in
terms of the amount of
nitrogen in the form of
005 to05 nitrate.


Figure 2-1. Nitrate-N concentration in the SRWMD. Source: Hornsby et al. (2003), p. 22.

Regarding surface water quality, the SRWMD has implemented the Surface Water

Improvement and Management (SWIM) program, a quality-monitoring network

consisting of 67 water chemistry stations. In water year 2002 (October to September),

3,012 tons of nitrate-N were transported to the Gulf of Mexico by the Aucilla, Econfina,

Fenholloway, Steinthatchee, Suwannee, and Wacasassa rivers; the Suwannee river alone









accounted for 98.6% of it. The middle Suwannee River Basin that accounts for only 8.6%

of the Suwannee River area accounted for 29.3% of the total loading; and the middle

Santa Fe River that represents only 5.7% of the Suwannee River Basin accounted for

19.6% of the total loading; those are the N hotspots.

In 2002, of the 35 springs whose water-quality standards were studied, 5 were

classified as poor quality and 15 as fair quality. Of the 26 rivers (or river sections)

identified in the District, 1 was classified as poor and 13 as fair quality.

Using the SRWMD database of monitoring point water quality, graphs for 3

springs in the middle Suwannee River Basin were constructed. Figure 2-2 shows N levels

in 3 central springs (Blue, Ruth/Little, and Troy) with data for the monitoring network of

the SRWMD. There clearly is an increasing trend. In Figure 2-2C additional data from

the US Geological Service (USGS) and from the Florida Geological Service (FGS) for

previous years are presented. Figure 2-2C clearly shows an increase in the N levels in this

surface water body; the line indicates the increasing trend.

In previous water year, 2001, (Hornsby et al., 2002a) the total loading of nitrate-N

going to the Gulf of Mexico was 3067 tons; and 97.8% was accounted for by the

Suwannee River Basin. These numbers suggest a loading decrease from 2001 to 2002.

However, in absolute numbers, loading was higher in 2002, since the discharge of the

Suwannee River increased. The fact of increased loadings year to year is reported to be

true also for many previous years.

In 1999, the FLDEP (Copeland et al., 1999) after sampling 491 wells in the middle

Suwannee River, found that wells located 300 ft from poultry or dairy farms had nitrate

concentrations greater than the background, with a 25% chance of exceeding the MCL.













3.5

S3
--
2.5

S2


1
1.5



S0.5

0--
Lt
0')
Q_

14

12

S10

S8
0
2 6

- 4

2

0--


Q(0 (0 1- 0 0 0 C CC
0') 0') 0') 0') 0 R 0 ? I? 0 I? 0
L >n
o c a -0
U- C O M Z CO )


SRVVrM.


FGS


0( N o r. O00) M N CO LO O (0) 0 0 Y- CDI CO O 0-0 C-- CO -T to 0 M N


|~o >,._ ^ 1 'g-S ^ 5l l '"G | o lo >. |_.,-- .
(.(.0 (0(0)( (U Z)))))) )W 00 M<-


Figure 2-2. Nitrate-N concentrations in three springs in the middle Suwannee River
Basin. A) Blue, B) Ruth/Little, C) Troy. Source: Prepared with data from the
SRWMD, FGS, and USGS.


(0 (0 r 0 0 0 N M 0
-5 c" 0 c o o o
o -


LO
0')
Q0


3.5 -

3
-J
S2.5

= 2

2 1.5

d 1
- USGS


to

. **








The 2001 FLDEP report for the Suwannee River Basin states that 7 lakes (Cherry,

Alligator, Crosby, Rowell, Sampson, Santa Fe, and Hampton), 1 stream segment in Swift

Creek; several segments of the Santa Fe River; the Middle and Lower Suwannee Rivers,

and their estuary are potentially impaired because of excess nutrient concentrations. This

FLDEP water quality report is summarized in Figure 2-3.





















Ground Water Nitrogen Load
in ThousanC Kilograms





Figure 2-3. Estimated annual mean of groundwater N load. Source: FLDEP (2001), p.
81.

A study by the FGS (1992) reports that the proportion of samples of water

exceeding the PDWS for Florida of 10 mgL1 is low (0.6% for superficial system, 0% for

intermediate aquifer system, and 1% for the Floridian aquifer system); however, it
,."un Ii ta .-." ge ^~ad
-nTosn iorm
1 J ."_


















intermediate aquifer system, and 1% for the Floridian aquifer system); however, it









mentioned that for the Suwannee River Basin high nitrate concentrations (up to 140 mg

L-1) in monitoring wells associated with dairy operations have been found.

A cooperative study between the DEP and the USGS measured monthly water

samples of 14 wells in the middle Suwannee River Basin (Katz and Bohlke, 2000). This

study found concentrations of nitrate-N ranging from 0.02 to 22 mgL-1, with large

seasonal variations. The fluctuations were believed to be correlated mostly with

agricultural practices and climate conditions; for example, higher concentrations occurred

in winter when waste or fertilizer was applied. Additionally, the use of N isotope

indicated that the most likely cause of nitrate in 5 wells were dairy and/or poultry

operations (Katz and Bohlke, 2000). This coincided with the fact that these farm

operations were close to the sampling sites.

A long term and comprehensive study in the Suwannee River Basin developed by

the USGS (Katz et al., 1999) indicated that N levels have increased from 0.1 to 5 mg L-1

in many springs in the Suwannee River Basin over the past 40 years. Several sources of

non-point N (fertilizer, animal wastes, atmospheric deposition, and septic tanks)

increased continuously in Gilchrist and Lafayette counties. Isotope analyses indicated

that the source of N was mixed between organic and inorganic, attributing them to animal

operations and crop fertilizer, respectively, from nearby operations.

This study (Katz et al., 1999) compared long term trends in nitrate concentrations

in selected springs with estimated inputs ofN sources in Lafayette and Suwannee

counties. It found a high correlation between the estimated quantity of fertilizer and the

concentration of N in the springs. It showed for Suwannee a peak in the 1970s, a









decreasing trend until the early 1990s, and finally an increasing trend again after 1993,

while for Lafayette County there was a steady increase since 1960s.

There are historical records for 4 springs in the Suwannee River (Hornsby and

Ceryak, 1999): Charles (1973-97), Little River (1980-97), Running (1980-98), and

Telford (1980-97) (Figure 2-4); and for 3 springs in Lafayette county ((Hornsby and

Ceryak, 1999): Troy (1960-98), Lafayette Blue (1980-97), and Mearson (1980-98)

(Figure 2-5).

3.5


E 25
2-0
S.5 CHARLES SPRING
10 LITTLE RIVER SPRING
RUNNING SPRING
0.5 TELFORD SPRING
0 I I I I I
1970 1975 1980 1985 1990 1995 2000

Figure 2-4. Long term records ofNitrate-N for 4 Suwannee springs. Source: Modified
from Katz et al. (1999), p. 46.

Pitman et al. (1997) reported that nitrate concentrations in the Suwannee River

itself have increased at the rate of 0.02 mg L-1 year- over the past 20 years and that over

a 53-km river stretch between Dowling Park and Branford, the nitrate loads increased

from 2,300 to 6,000 kg day-. Of this total, 89% appeared to come from the lower

two-thirds, where agriculture that includes crops and livestock is the dominant land use.

This study found that nitrate-N concentrations in the 11 measured springs ranged between

1.3 and 8.2 mg L-1, and nitrate-N concentrations for the river of 0.46 mg L-1 in Dowling

Park, 0.46 mg L-1 in Luraville, and 0.86 mg L-1 in Branford; even though the nitrate-N









concentration almost doubled from the upper to the lower part of the river, there was no

noticeable difference among the contributors' springs, suggesting that other sources

might also be affecting the river in the lower part.

3,10 1-

J 2.5 BLUE SPRING.
SLAFAYETTE COUNTY
E 2, MEARSON SPRING -
Z TROY SPRING


Z
0.5

0
1050 %i90 970 $980 990 2000

Figure 2-5. Long term records of Nitrate-N for 3 Lafayette springs. Source: Modified
from Katz et al. (1999), p. 46.

2.2.2 Assessments at the Watershed Level

Asbury and Oaksford (1997) performed a study in 7 rivers in Georgia and Florida;

the Suwannee River was one of them. This study found that the N input ranged from

2,400 kg yr1 km2 in the Withlacoochee river basin to 5,470 kg yr1 km2 in the Altamaha

river basin, with the Suwannee input in the range of 3,939 kg yr- km2. However, when

these numbers were recalculated to in-stream loads, the Suwannee had the highest N

loads (compared with the other 6): 476 kg yr1 km2 (with the range of the other 6 between

152 and 405 kg yr- km2) and discharges in the order of 9,700 kg 103 (with the range of

the others between 720 and 8280 kg 103). Additionally, the ratios between outputs/inputs

of N were calculated for the 7 rivers (range 4.2% to 14.9%) and the Suwannee River was

the second highest with 12.1%.









Ham and Hatzell (1996) realized another comprehensive study of N pollution in

surface waters of Georgia and Florida. This study found, for the 10 major river basins in

the study, that median concentrations were below USEPA guidelines of 10 mg L-1 for

nitrate and 2.1 mg L-1 for ammonia, with only one exception in Swift Creek (3.4 mg L-1

ammonia) in the Suwannee River Basin that was attributed to waste water discharges.

Although in general the nutrient levels were low, long-term increasing trends of nutrients

loads were found in most of the sites. Distribution of results of nitrate-N indicated that

the Suwannee presented the third highest mean but the higher variability with the highest

overall value of more than 70 mg L1.

Ham and Hatzell (1996) also performed a long-term trend for these 10 major river

basins. They developed graph relationships between time and N concentrations. They

found a slope of +0.02 for the Suwannee River Basin, between the years 1971-1991. This

means a 4.61% yearly increase over the baseline of 0.5 mg L-1. This was the fourth

greatest increasing trend regarding mg L-1 year- comparing with the other Basins.

2.2.3 Assessments at the National Level

Berndt et al. (1998) published a USGS regional study of the Georgia-Florida

coastal plain that includes the Suwannee River Basin. This study was part of the National

Water-Quality Assessment (NAWQA) program that includes more than 50 of the

nation's largest river basins and aquifers. This study reports that in the 23 groundwater

samples from the row-crop agricultural area in the upper Suwannee River Basin, 33%

exceeded the drinking water standard. These samples were from aquifers that overlie the

Upper Floridian aquifer, the major source of drinking water for the study area. This study

also found that nitrate concentrations were generally higher in samples from agricultural









areas and shallower wells and N isotope ratios indicated that both fertilizer and animal

wastes were the predominant sources of nitrates in groundwater.

The study ofBerndt et al. (1998) found that the nutrient more frequently found in

groundwater was N, which was mostly affected by land use, geology, soils, and

topography. Especially in the Suwannee River Basin, the problem is aggravated because

it is a karst region with well-drained soils that have a low water-nutrient holding capacity.

Also Berndt et al. (1998) found no direct relation between land uses and N

concentrations. However, they referred to the previous Pitman et al. (1997) study

between Dowling Park and Branford (33-mile Suwannee River reach) including 11

nearby springs that showed much higher N levels in the springs (x = 1.7, range 1.3-8.2

mg L1) than in the river (between 1 and 2 mg L-1): the N levels increased when it moved

to the lower parts. In this case, even though the water flow dynamics have a great effect,

a clear association was found between the predominant agricultural land use and the N

levels; however no specific type of farms could be identified as the sources.

A comparison of the Suwannee River nutrient concentration with the NAWQA

national standards, indicate that overall the levels of N in the streams and in the

groundwater of the Suwannee River Basin are between the 75th percentile and the

median.

2.2.4 Assessments at the Dairy Farm Level

Andrew (1992) did a comprehensive study of nine dairy farms in the middle

Suwannee monitoring waste water lagoons, monitoring wells, and supply wells and their

interactions with the different land uses inside the farms (waste water, intensive areas,

intensive pasture, sprayfields, and low density pastures). This study tracked the N inside







26


the dairy farms after cattle excretion, measuring it at different depths and in different land


uses. The results are quite revealing and they are summarized in Figures 2-6 and 2-7.


230
220
210
200
190
180
170
160
150 -..-.. .................. -
140 -
130 I
1 120 -
E 110
100
90 -
80
70
60 -
50 -
40 -
30 -
20 -
10 -
0 -
N03+NO2 NH3+OrgN N03+NO2 NH3+OrgN N03+NO2 NH3+OrgN N03+NO2 NH3+OrgN
Waste Water Shallow Groundwater Deeper Groundwater Nearby hydrogeologic unit


Figure 2-6. Nitrate-Nitrite and total ammonium plus organic N at different points in nine
dairy farms in the middle Suwannee River Basin. Source: Prepared with data
extracted from Andrew (1992).


Two main forms of N were measured, Nitrite + Nitrate (NO3 NO2 ) and


Ammonium + Organic N (NH4 -Org). In the waste water in the lagoon (effluent) the


concentration of NO- N02 was negligible because most of the N was in form of



NH4 -Org (x =150 mg L-1) as seen in Figure 2-6. Because of fast decomposition and


nitrification N is quickly converted to NO3 N02 forms, which are evidenced in the


monitoring wells (shallow groundwater at 3-m deep), in which the concentrations of


NO NO2 are far higher than the NH4 -Org forms (x = 22 vs. 2.2 mg L1). The trend


continues in the deeper groundwater source wells (12-15 m) where the concentrations of










NO3 NO2 are substantially lower than in the shallow groundwater wells because of


dilution processes, but still high for the water quality standards (x =10 mg L-1). In these

deep wells, the NH4 -Org concentrations are much lower than the waste water, but

slightly higher than the shallow groundwater suggesting that some deep accumulation

may occur.

In the nearby hydrogeologic unit the concentrations of NO3 NO02 are very similar

to those in the deep groundwater wells suggesting that the dairy might be responsible for

such an impact. As expected NH4 -Org forms would be negligible on these components.


3
nlnl.:.r,,Ln l *r.Q g3, Ir

2.5


2


1.5
E




0.5


rea Sprayfield Waste Water Intensive Pasture Low Density
Pasture


Figure 2-7. Dissolved Nitrate+Nitrite and Ammonium+Organic N in monitoring wells
(3-m deep) in different areas of a dairy farm. Source: Prepared with data
extracted from Andrew (1992).

Inside the dairy farms, the collected data demonstrated the movement ofN

downwards at different rates for different parts of the farm. Exploring the data from

shallow wells, the intensive area, where cows deposited manure without regular









collection, is the critical area where the highest concentration of NO3 -NO2 is observed

(x = 26 mg L-1); this is followed by the sprayfields where intense applications of effluent

are performed (x = 23 mg L-1); followed by the wastewater area and the intensive

pasture (both x = 11 mg L-1). The lowest concentration was found in the low density

pastures (x = 5.1 mg L-) but this is still half of the allowable limits of the PDWS. The

NH' -Org concentrations are medium (x = between 0.96 and 2.1 mg L-1, taking into

account that 2.1 mgL1 is the USEPA limit) in these shallow wells; being higher, as

expected, below the wastewater lagoons. This also indicates that there was a downward

infiltration of the wastewater.

In another study from Andrew (1994), four dairy farms were closely monitored

during four years (1990-1993) in the middle Suwannee. In this study, water samples were

evaluated in 51 monitoring wells in these 4 typical dairy farms and 3 nearby springs in

Lafayette and Suwannee counties. Results indicated that

* Water from the monitoring wells had median nitrate exceeding that in water from
wells of the background groundwater quality network sampled by the FLDEP

* Nitrate concentrations near waste water lagoons and intensive livestock areas were
the highest exceeding in the shallow wells the drinking standards of 10 mgL-1

* Concentrations were substantially lower in pastures (less loading of manure) and in
deep wells (dilution and/or denitrification)

* Nitrogen isotope ratios analyses demonstrated that animal waste was the main source
of nitrates in groundwater adjacent to the lagoons and intensive areas, and animal
waste and/or septic tanks were the principal source of nitrate at the three springs

* Nitrogen isotope ratios also demonstrated that synthetic fertilizers or combinations of
synthetic fertilizer, soil N, and livestock wastes were the sources of nitrate in pastures
and sprayfields.










Figure 2-8 shows the average of nitrate concentrations above the 10 mgL-1 for the 4

farms. As reported in the conclusions of Andrew (1994), the wells nearby the intense

areas and the wastewater lagoons are those with higher environmental menace.

Nonetheless, it was noticed with these study data, that 2 of the farms presented wells

below the 10 mgL' for waste water lagoons and 1 farm showed wells of less 10 mgL'1 for

a monitoring well near the intense areas. Also, 2 of the farms presented "high" wells for

pasture fields and 1 farm additionally a "high" well for a sprayfield.


70
60
50 -
S40
E
30
20 -
10
0



FARM


I I- I


I1
w


0 -) 05) ) 0) 0) V 0)5
0) 0)
FARM 2 FR 3 -AR
-pa a, a
"aa aa

FARM 2 FARM 3 FARM 4


Figure 2-8. Monitoring wells with higher than drinking water allowable nitrate levels
(PDWS>10 mg L-1) in the four dairy farms. Source: Prepared with data from
Andrew (1994).

2.3 Human and Ecological Concerns because of N Pollution

High nitrate levels in drinking water can cause methemoglobinemia (MHB) or

"blue baby" syndrome in infants, a condition that diminishes the ability of blood to

absorb oxygen (Vigil et al., 1965; NRC, 1985) because the methemoglobin produced by

nitrates cannot carry oxygen, causing symptoms of oxygen starvation that affects









primarily infants younger than six months, who suffer cyanosis and, in extreme cases,

death (Andrew, 1994).

Nitrate is a precursor of carcinogenic nitrosamines, which are believed to cause

stomach cancer and leukemia (NRC, 1978; Strange and Krupicka, 1984; Stewart, 1990;

Mirvish, 1990; Forman, et al., 1985). Additionally, high levels of nitrates in drinking

water are associated with birth defects (Fan et al., 1987), stress of the cardiovascular

system with extreme artery dilatation (Vrba, 1981), enlargement of the thyroid gland, and

increased sperm mortality (Scragg et al., 1982).

Kross et al. (1992) mention that about 2,000 cases of MHB have been reported

between 1945 and 1990 worldwide with a mortality rate of 10%; however the NRC

(1995) found no studies reporting nitrate-induced MHB since 1990; nitrate induced MBH

seems rather rare now, but these rare cases are due primarily to a lack of reporting

requirements, and secondarily to a lack of physician awareness (Felsot, A. 1998).

Several effects on reproduction, growth and development of animals have also been

noticed because of high N levels, including mortality of larvae of Cutthroat trout,

Chinook salmon and Rainbow trout and mortality of tadpoles in several frog species in

Florida such as Chorus frogs and Leopard frogs (Edwards et al., 2002). In several species

of frogs, reduced activity, weight loss, and deformed digestive tracts also have been

detected (Edwards et al., 2002).

In addition to making water unsafe for humans and many other animals, high

nitrate concentrations can lower water quality in rivers and springs because elevated

concentrations of nitrate in these water bodies can cause eutrophication, which can result

in algal blooms and depletion of oxygen that affects survival and diversity of aquatic









organisms (Katz et al., 1999). Eutrophication is a process in which algae growth is

favored and consequently oxygen is depleted. This problem happens in the surface

water-bodies, which receive inflows from groundwater from springs or diffuse seepage

(Andrew, 1994).

2.4 Conclusions and Recommendations

Several studies have demonstrated that there is a N pollution problem in the

Suwannee River Basin. All studies reviewed for this paper agree on it. Even though the

scopes, methodologies, timeframe, and location differed, all the studies found that the

levels of N in groundwater and surface waters had, at the very minimum, higher N

concentrations than the background. Studies are not conclusive to mention if major

pollution exists in surface water sources or in groundwater sources. Temporal, spatial,

and climatic conditions would have great effect on overall N pollution levels.

When comparing the Suwannee River Basin with other basins in the state, in the

Watershed (including Georgia), and across states, it was found that, overall, the water in

the Suwannee presents lower quality because of N pollution. The Suwannee River Basin

or parts of the Suwannee River Basin are consistently mentioned by the studies as

examples of extreme polluted values of any form of N either in groundwater or surface

water.

Even though, in most of the cases, the N concentrations in the Suwannee River

Basin are below the USEPA standards of 10 mg L-1, several studies that include long

term comparisons, along with hard data collected from monitoring stations, show that

there is a noticeable increasing trend in the N concentrations of the surface and

groundwater bodies that would tend to increase even faster if the land use conditions

persist. Therefore, actions to mitigate the problem are needed.









By exploring hard data generated by monitoring efforts in the Suwannee, the

middle Suwannee (between Suwannee and Lafayette counties) continues being the

hotspot and the area with higher N pollution concerns; nonetheless, there are other areas

and places where the N concentrations are showing as high or even higher than the

middle Suwannee as is the case of areas in Jefferson, Levy, Gilchrist, and Alachua

counties. The current effort to decrease N sources in the middle Suwannee should be

expanded to other areas of concern.

It is reported that the possible nitrate sources are chemical fertilizers, animal

wastes, and human waste. There is no clear cut conclusion as to which one of these

represents the most dangerous source. Dairy operations are many times found as the most

likely source of nitrate-N by several studies either by spatial proximity, temporal

analysis, and/or by the use of isotopes. However, there is no clear estimation of how

much N comes from dairy farms either as a whole or on an individual basis. It is believed

that dairy farms are big contributors because of large amounts of waste produced;

however there are no conclusive reports that estimate and/or compare them with other

sources. Higher efforts are required to distinguish between point vs. non-point sources of

pollution and to make any industry environmentally accountable as a way to face the

problem and their potential solutions.

Two in-depth studies in dairy farms in the study area evidenced a high

environmental impact of these operations regarding N pollution. These studies pointed

out that there is a great variability between dairy operations and their environmental

impacts. An appropriate approach for environmental solutions should be to treat dairy

farms individually regarding their pollution concerns.









Also, these dairy studies found no conclusive answers to what components inside

the operation are the greatest N polluters. These also vary substantially among farms.

Even though only a decade has passed since these studies, the dairy farm systems have

changed substantially, most of them now

* Do not have the "intense area"
* Have a concrete lagoon
* Do not use or use very little synthetic fertilizer

The "intense area" has become part of the "confined area" in a totally concreted

covered space where all the manure is collected and where there is no infiltration. The

manure goes to a concrete lagoon and from there to the sprayfields, in which usually only

waste manure is used as fertilizer. Under this new structure, the sprayfields, primarily,

and the pastures, secondarily, are the most likely zones inside dairy farms with higher N

concentrations and consequently with higher N pollution impacts. Dairy farms as major

actors in the N issue must be studied as a whole system with complete understanding of

all their components under specific environmental and management options affected by

varying climatic conditions.

There is clear evidence that water N pollution affects humans, animals, and

ecosystems. In the study area, even though there are no reported direct or indirect human

effects, there is major evidence of animal and ecosystem effects because increased N

levels.














CHAPTER 3
NORTH FLORIDA STAKEHOLDER PERCEPTIONS TOWARD NUTRIENT
POLLUTION, ENVIRONMENTAL REGULATIONS, AND THE USE OF CLIMATE
FORECAST TECHNOLOGY

3.1 Introduction

Rising nitrate levels in north-Florida's Suwannee River are a growing concern

(Hornsby et al., 2003). Awareness is high given the potential of excess nutrients to affect

the environment and public health (Albert, 2002). It is alleged that the most likely

contributors to contamination are dairy farms, industry, and human settlements in the

region (Andrew, 1994).

The federal safety standard for nitrate-N in drinking water in the United States is 10

mg L-1. In some areas of the Suwannee River Basin, nitrate levels have reached ranges of

20 to 30 mg L-1 (Hornsby, Pers. Comm.). High nitrate levels in drinking water can cause

methemoglobinemia or "blue baby" sickness in infants (NRC, 1985) and other health

problems in humans such as enlargement of the thyroid gland, increased sperm mortality,

and even stomach cancer (NRC, 1978; Strange and Krupicka, 1984; Stewart, 1990;

Mirvish, 1990; Forman, et al., 1985). In addition to making water unsafe for humans and

many other animals, high nitrate concentrations can lower water quality in rivers and

springs, causing eutrophication resulting in algae blooms that consume oxygen needed by

fish and other aquatic animals (Katz et al., 1999).

Dairy farmers are now required to comply with stricter environmental regulations

either under permit or under voluntary incentive-based programs. Dairy farmers are also

aware that environmental challenges will be the greatest problem they will have to face in









the near future. After interviewing 48 dairy farmers located in Lafayette, Suwannee,

Columbia, Hamilton, and Gilchrist counties, Florida, Staples et al. (1997) found that the

perception of anticipated costs of having to comply with environmental regulations was

rated, by far, the top challenge to successful dairying in the future.

Improvements in long-term (6 to 12 month) seasonal climate variability predictions

such as those based on El Nifio and La Nifia forecasts (O'Brien, 1999) can play an

important role in formulating management strategies that dairy farmers in north-Florida

may adopt in order to pursue economic and ecological sustainability. Specialized federal

and state agencies such as the National Oceanic and Atmospheric Administration

(NOAA) predict long term seasonal climate variability fairly accurately. These

predictions are based upon the El Nifio Southern Oscillation (ENSO) phenomenon.

ENSO phases determine El Nifio, La Nifia, or Neutral years, which are related to rainfall

and temperature patterns.

Dairy farmers may benefit from proactive use of seasonal climate variability

prediction. The opportunity to benefit from climate prediction arises from the intersection

of human vulnerability, climate prediction and decision capacity (Hansen, 2002a). This

benefit from the information can be used to decrease environmental impacts while

maintaining profitability in north-Florida dairy operations.

This paper discusses the perspectives, perceptions, and attitudes of stakeholders

regarding the issue of N pollution in the Suwannee River Basin, the current regulations,

and the possibility of incorporating climate variability forecast technology to initiate

mitigating strategies. The principal stakeholders are the dairy farmers. The









methodologies included a Sondeo with stakeholders, 21 interviews with dairy farmers,

and eight focus groups with stakeholders.

Stakeholders' perceptions regarding nutrient overloads, regulations, and the use of

seasonal climate information as a way to mitigate the nitrate problem are critical baseline

data for designing solutions. The information is even more valuable because it is

gathered in a participatory and consensual manner. This study translates the mind-set of

the stakeholders about the N pollution issue and of the dairy farmers regarding

environmental regulations and the use of climate variability forecast technology.

3.2 Materials and Methods

3.2.1 Study Area

The Sondeo was carried out in the Suwannee River Basin that encompasses all of

nine and part of another six north-Florida counties. This represents an area of 7,640 mi2

The area is comprised of 13 river basins and includes a population of 285,000. Within

this area, the principal focus was on the middle Suwannee that includes Lafayette and

Suwannee counties. These two counties are specific areas where the N issue is acute. In

the middle Suwannee River there are hundreds of residential and commercial septic

systems in rural areas. Some 300 row crop and vegetable farms also exist. Finally, 44

dairies with more than 25,000 animals and 150 poultry operations with more than 38

million birds are installed in the area (Suwannee River Partnership, 2003).

Interviews and focus groups were undertaken in Lafayette, Suwannee, Gilchrist,

Levy, and Alachua counties, where the number of dairy farms totals 64.

3.2.2 The Sondeo

A Sondeo was conducted between the months of January and February 2003 by a

multidisciplinary team of eight graduate students from the University of Florida (Alvira









et al., 2003). The Sondeo (Hildebrand, 1981) is a team survey developed to provide

information rapidly and economically about agricultural and rural problems. It is

structured around a series of informal, conversational interviews between the team and

stakeholders. It is a multidisciplinary process from data collection through report writing.

Each team ideally includes members from the social and agricultural sciences, as well as

other disciplines. When members write the final report there is much interdisciplinary

cross fertilization. In the Sondeo, data are shared among the different teams and report

writing is done as a group so that observations are confirmed, debated, and analyzed with

members of the other teams. The results may be quantified or not but the accuracy of the

findings is strengthened by the cross-checking process.

During the Sondeo, 13 stakeholders were interviewed. This sounding out was

undertaken to get a better understanding of the broad issue of N pollution in water and

collect the perception of stakeholders regarding this issue in the study area. Sondeo team

members represented a number of disciplines including agronomy, animal science,

agricultural extension, natural resource management, forestry, and community

development. Two to three-person teams conducted initial transects and stakeholders'

interviews. The interviews were conducted as informal, open-ended conversations. After

a team introduced itself and explained the subject of the research, the stakeholder and the

team conversed about any aspects of the topic that were relevant. This open-ended

approach enabled topics to emerge and be pursued that might have been missed if the

researchers had been constrained by a pre-formulated questionnaire. Despite the

open-endedness, a set of core themes was discussed in almost all interviews. Notes were

not taken. Following each interview, the partners wrote individual notes, compared them,









and then wrote a joint report for the interview. Finally, the joint reports were shared and

thoroughly discussed among all Sondeo members. Each team presented its findings,

which could be clarified, challenged or contrasted with the results of the other teams.

This process of reporting and discussion served as the opportunity to begin noting trends,

gaps in information, sources of initial inferences, and identifying new questions to

pursue.

3.2.3 Interviews

Personal interviews (Bernard, 1995) were undertaken with dairy farmers and/or

dairy farm managers during the fall/winter seasons of 2003/2004. Open-ended interviews

were targeted to a purposeful sample of 21 farmers (30% population). The number of

interviewees was chosen in order to attempt to capture the broad variability represented

in the area in a collaborative manner with the Suwannee River Partnership and the

University of Florida Extension Service agents in the area. A voice recorder was used in

every interview after requesting permission from the interviewee. Interviews included a

series of topics, but emphasized farmers' perceptions regarding the N pollution regulation

issue and the potential use of climate forecasts to address the problem.

3.2.4 Focus Groups

Eight focus groups (Bernard 1995; Stewart and Shamdasani, 1990) with

stakeholders were held between November 2002 and February 2004. Focus groups were

open discussions among 5 to 15 people. Topics focused on perceptions regarding nitrate

pollution, regulation, and seasonal climate forecast innovation technologies as an aid to

mitigate the problem.

Focus group participants ranged from specialists from official agencies such as the

Suwannee River Partnership (SRP), the Suwannee River Water Management District









(SRWMD), the Florida Department of Environmental Protection (FLDEP), the Natural

Resources and Conservation Service (NRCS) from the United States Department of

Agriculture (USDA), the University of Florida IFAS Cooperative Extension Service

(CES), to private sector companies, technicians and farmers. These participants were

typically called to meetings by the Suwannee River Partnership and/or by extension

agents from different counties in the research area in coordination with the researcher.

3.2.5 Analysis

All participatory interactions were recorded, classified, organized, analyzed, and

described. This paper summarizes the results of all the above interactions in a

descriptive/narrative manner. Names of the participants are not revealed to protect the

anonymity of informants.

Emphasis has been placed on a qualitative rather than quantitative description and

analysis of farmer perceptions. The interdisciplinary approach to the complex problem of

nitrate pollution in the Suwannee River Basin requires the use of methods from social

sciences such as applied anthropology and rural sociology for a more holistic view.

3.3 Results and Discussion

3.3.1 Stakeholder Perceptions of the N Issue

3.3.1.1 Perception of water contamination

During a discussion in an initial focus group, one stakeholder, supported by several

participants, challenged two main points: does water quality degradation caused by N

pollution really exist in the Suwannee River; and if the problem exists, how bad is it?

Stakeholders argued that is not easy to know if there is a real problem, and even more

difficult to determine the source of the problem. Additionally, an actual timeline of

events that may have led to the current situation is unknown. Farmers present in the









meeting perceived that the 10 mg L-1 standard ofN concentration in water enforced by

the regulatory agencies is too strict compared to other parts of the world (Europe for

example, has a 40 mgL1 limit). This perceived low threshold is also believed to

exaggerate the real problem. The dairy producers expressed that to their knowledge, there

are no indications of humans being affected by N concentrations in local water. Farmers

also believed that eutrophication of local water bodies is not as evident as media reports

suggest. In their eyes, the appearance of streams and springs is similar today as it has

always been.

Some farmers present at this meeting believed that there is a misconception of the

link between the N issue and dairy farms. Participants indicated that dairy farms in the

study area do not always have overflow problems. Most indicated that in general terms,

large operations have higher chances of overflows. Additionally, many of the participants

indicated that they were not convinced that dairy farms in general are the biggest

polluters.

Conversely to these opinions, the Sondeo found that ten years earlier, an increase of

nitrate concentration was recognized by the Suwannee River Water Management District

in the Suwannee River Middle Basin. The increase in nitrate concentration in the

Suwannee River Middle Basin as a problem was clearly evident to all stakeholders who

participated in the Sondeo. A veterinarian working for a dairy farm expressed his

concern for the pollution, and its potential for driving them out of business in the future.

It seems clear that dairy farming is perceived as the most contaminating activity in the

area. Poultry is perceived to be the second source of contamination. Row crops are

usually perceived as the least contaminating of all food producing activities in the area.









However, one interviewee pointed out that greater leaching might result from fertilizer in

row crops compared to the slower decomposition of manure from dairy operations. This

would occur when precipitation follows an application of fertilizer in row crops, or if

they over-irrigated. Moreover, the informant perceived that all farm activities might

contaminate much less than non-agricultural industrial activities and waste outputs from

concentrated urban dwelling in the area.

Even though the main focus in the discussions had been N loads, phosphorus (P)

was also mentioned as a potential contaminant by participants in the Sondeo. A problem

expressed about both these contaminants is that they may take many years to reach the

groundwater. Nutrient pollution has a long term cumulative effect that results from slow

processes. This makes it much more difficult to pinpoint sources of contamination and

consequently potential ways of mitigation.

The Suwannee River Partnership was formed to determine the sources of N loads in

the basin. The other important and connected objective was to work with local users to

minimize future nutrient loading through voluntary, incentive-based programs. Therefore,

the Partnership works under a non-regulatory precautionary approach. This is envisioned

as being accomplished through monitoring pollution, education, promotion and adoption

of best management practices (BMPs).

The Partnership is focused on the middle river basin. The main activities in that

area i.e., dairy, poultry, beef, and row crop farms are the foci of the Partnership.

Currently, there is little population pressure since the middle river basin is not very

densely urbanized. However, a recent geographic extension of the Partnership to the

southern part of the basin towards the Santa Fe River includes more densely populated









areas. A logical step then would be for the Partnership to focus more deeply on septic

tanks and human dwelling waste outputs.

3.3.1.2 Reasons to mitigate contamination

The main reasons identified by the stakeholders for mitigating water contamination

are to preserve the ecosystem, protect human health, protect fresh water bodies that

provide environmental services, and support economic activities such as tourism.

It is important to note that nobody mentioned that the water quality has an impact

on their herd, poultry, or crop. It is also recognized that pollution in the river basin can

have a negative effect in shellfish production in the Gulf of Mexico.

3.3.1.3 Existing and potential means of mitigation

Stakeholders were proactive about avoiding potential regulations, which did not

seem a good choice for addressing the problem to them. Many farmers might find

themselves in a difficult situation to fulfill imposed regulatory levels. Farmers did not

want to be regulated to have to implement conservation measures. Lake Okeechobee

(South Florida) regulations, which were imposed to reduce the P load level, were not

successful in the sense that 50% of the dairy farms in that area went out of business.

Based on this past misstep, stakeholders worked under a non-regulatory, preventive, or

precautionary approach, through the creation of the Suwannee River Partnership (SRP).

The recognition of rising nitrate concentrations was the leading factor behind the

creation of SRP. The idea of forming a Partnership took years. According to some

stakeholders, just a few people started the partnership in the mid 1990s. More institutions

progressively joined in the following years. In 2003, there were 24 members, and

currently (2004) the number of partners has increased to 45. The long-term goals of the

partnership are to









* Reduce pollutant loading
* Establish interstate watershed management
* Increase number of acres covered by conservation and land management plans
* Assure flow of information
* Secure cooperative long-term funding to maintain a cooperative interstate program
* Improve water conservation between two states (Florida and Georgia).

Several agencies and individuals are part of the Partnership that collaborate to

monitor and to reduce the levels of nitrates in their groundwater and surface waters. For

example, the USDA/ NRCS PL-566 Small Watershed Program for the Middle Suwannee

assists dairy and poultry farmers to improve animal waste management and fertilization

techniques. The NRCS inspects soil types, number of cows and area in pastures, and then

decides the minimum set of measures to be followed. The Florida Farm Bureau (FFB),

that represents farmers' interests, is also involved in developing policies to mitigate

nitrate pollution in the area. The FFB annually publishes a book containing all of its

policies. The Farm Bureau supports technology-based BMPs and their voluntary

implementation. The U.S. Geological Survey (USGS) also has ongoing studies of water

quality within the Suwannee River Basin

The Partnership focuses on finding the most economical and technologically

feasible management techniques (BMPs) available to help farmers and other land users to

satisfy regulatory requirements for protecting public health and the environment.

Through education and outreach programs, the group continues to increase public

awareness of the issues, and encourage citizen and community participation in working

together to find solutions. Other components of the initiative that are contributing to the

Partnership's success include the BMP Quality Assurance Program, County Alliance for

Responsible Environmental Stewardship (CARES), and the On-farm Research Programs









As one component of Partnership activity, several University of Florida

departments and centers, as well as state and federal agencies are cooperating on project

319 entitled "Effectiveness of Best Management Practices (BMPs) for Animal Waste and

Fertilizer Management to Reduce Nutrient Inputs into groundwater in the Suwannee

River Basin." Project 319 addresses BMPs on a dairy, a poultry operation, and a

vegetable farm in Lafayette and Suwannee Counties in the Middle Suwannee River

Basin. Each farm has been instrumented with a groundwater monitoring well network.

Wells and soil profiles are being intensively monitored to determine pre- and post-BMP

nitrate concentrations on each of the farms. An one-year pre-BMP monitoring phase has

been completed on all the farms, and BMPs are now being implemented to address the

elevated nitrate concentrations observed in some of the monitoring wells. The overall

goal of this project is to demonstrate that BMPs can be implemented to reverse the

increasing trend of nitrate concentrations in the groundwater and springs of the Middle

Suwannee River Basin.

The BMP Quality Assurance Program verifies that BMPs are operated and

maintained over a long period of time, and provides assistance to farmers in resolving

problems with the BMPs. The Partnership, in cooperation with Florida Farm Bureau, has

also initiated the CARES Program to certify and recognize farmers participating and

following directives in the Partnership program. Finally, the Partnership has developed a

joint strategy with the SRWMD to monitor groundwater and surface water on a regional

basis to identify trends in water quality over time through an extensive network of

monitoring points.









Technicians from the Partnership visit farms on a regular basis to monitor their N

loads and to help with the implementation of BMPs. One of these technicians pointed out

that many farmers were formerly wasting money buying fertilizer. With new techniques

such as use of a cement pool and mixer in the lagoon to separate solids from liquids, the

farmers are able to recycle their own manure, buy less fertilizer, reduce costs, and reduce

N loads.

In general, farmers who have been applying BMPs feel comfortable with the results

achieved. Most farmers are cooperative and eager to follow SRP directives and the

suggested BMPs. Nevertheless, some practices (such as treatment lagoons or solid

separators), require large investments. Money is required up-front to modify facilities,

train personal and to make other adjustments. In most cases the SRP has funding

available to cost-share initiatives that protect the environment.

The modality of implementation for most BMPs is completely based on voluntary

participation of farmers and aims at helping them avoid stricter regulations. However,

certain activities are already subject to regulatory programs.

3.3.1.4 Implementation procedures for Best Management Practices (BMPs)

BMPs for nitrate reduction in the Suwannee River Basin refer to good stewardship

of land and water management that help farmers reduce nutrient contamination to water

systems. There is general acceptance among stakeholders that such management practices

are sound and positive for the environment. Management practices mainly include

nutrient and manure management, disposal of wastes, irrigation water, technical design

and engineering aspects in farming enterprises such as row crop, dairy, beef cattle and

poultry farms. By large, the concerned stakeholders see BMPs as research-generated

solutions to reduce potential environmental hazards from farm enterprises.









Implementation procedures are specific to each program and farm enterprise. All

programs aim at implementing BMPs that reduce the loading of N contamination to

groundwater while helping farmers to remain profitable. It is worth mentioning that at

present, attaining the twofold objectives is at an early testing period in which cost-sharing

is crucial to ease associated risk for farmers.

Based on the information gathered by the Sondeo, a picture of the step by step

process of BMP implementation in the study area can be drawn. Creating and raising

awareness among public agencies, landowners, agricultural producers and other

stakeholders about existing water quality impacts resulting from human activities is the

first step. It is necessary for recognizing that a new approach in management practices is

needed to address the issue within the area. It is perceived that education on the issues

and solutions, including effective transfer of knowledge and technology, are essential

components of the implementation efforts of the technical working groups in the

Partnership. As part of this process the Partnership conducts meetings for agricultural

producers. During 2001 over 1000 farmers participated in such meetings. Gatherings

included two row crop growers' meetings, a forage field day, five dairy farmers'

meetings, two poultry growers' meetings, a cattlemen's meeting (beef), and a large-scale

growers' meeting.

In the CARES program, growers fill out a self-evaluation regarding their farm

operation to assess environmental practices. One objective of this self-assessment is to

increase farmers' awareness in perceiving their own status.

The next step is that farmers show interest in joining the Partnership. At this stage

agencies in the Partnership help farmers in selection and implementation of BMPs. The









partnership has a Technical Working Group (TWG) in the areas of fertilizer, animal

waste and human waste management. Each TWG assesses sources of nutrients within its

area of expertise, develops and implements a plan of work to fulfill the mission of the

Partnership, and provides updated recommendations for action. BMPs are highly

decentralized and tailored to suit the diversity in size and complexity of farm operations.

Initially, the focus is on management approaches, which are technically viable,

economically feasible, and environmentally friendly. The study team feels that BMPs on

dairy farms are far from being financially feasible. Up-front costs for implementation are

especially hard to recoup. Additionally, BMP on-farm trials are conducted to verify the

technical viability of the recommended practices.

Financial commitment and debt responsibility is required from farmers during

implementation of BMPs. To date, BMP implementation has been very successful. Out

of 144 crop farms, 115 have adopted conservation measures; out of a 35 poultry farms 34

participate; and out of 44 dairy farms, 43 are involved in the SRP. Before the Partnership,

only 3 dairy farms were able to pass environmental quality standards.

After the farmers adopt BMPs that have been promoted by the participating

agencies, a presumption of compliance with state water quality standards can be

established. This places farmers in a much better standing of willingness to cooperate

with environmental stewardship with authorities.

3.3.1.5 Impacts of BMPs on farms

Most stakeholders perceive that farmer participation in BMPs has many economic,

environmental and social benefits. Stakeholders agree that after the initial investment,

dairy, crop and poultry enterprises save money by implementing BMPs since they

recover their own nutrients, which have a value. In the case of dairy farms, the savings









could be important since no fertilizers would be needed for crops. However, stakeholders

perceive that poultry farmers have only modest economic benefits because the processing

of dry litter does not return much income. An additional benefit indicated was that

livestock farmers who implemented BMPs are usually allowed to expand herd or flock

size.

Regarding row crops, stakeholders indicated that the BMPs being promoted for

crops reduce yields. This reduction in crop yields seems to be inevitable because of the

reduction in fertilizer and irrigation contained in the BMP recommendation. Additionally

they would require farmers to spend more time and money on management and

equipment. Both aspects would decrease profitability. A major challenge then, is how to

maintain row crop farmer profit margins while implementing BMPs.

In spite of difficulties, adopting BMPs may help farmers to remain in business in

the long run, which is critical to the Partnership. The initial investment that farmers are

supposed to share may still be a big challenge considering the returns that farmers obtain

as a result of adopting BMPs. But ultimately, middle Suwannee farmers will be able to

choose from one of the three options: receive financial assistance to practice BMPs and

stay regulated under incentive-based programs; receive a certain amount of money and

leave the land; or comply with regulations under permit on their own.

3.3.2 Perceptions of Regulations by Dairy Farmers

During the interviews, dairy farmers expressed their concerns that regulations are

currently stricter than in the past, and that they will become even stricter in the future. At

the same time, several farmers expressed that the manure and all its by-products have a

tremendous commercial potential as a source of energy and nutrients. Potential benefits

from by-product sales could mitigate stronger future regulations.









Sixty-two percent of interviewed farmers participate in incentive-based regulatory

programs. One third of them work under permit, and one farm has no regulation at all.

Incentive-based programs are voluntary commitments between the farmer and the

regulatory agency as promoted by the SRP. Farmers under permit are those that are

enforced by the regulatory agency. Usually medium and small farms (up to 700 cows) are

under incentive-based programs and larger farms are under permit. Also, it can be said

that newer farms are under permit and older farms are more likely to be under

incentive-based programs.

All farmers expressed a clear message. They wanted to make sure the general

public understands that producers are not trying to pollute the environment. Farmers only

seek to earn their living and to do their best to protect the environment.

Eighty-one percent of dairy farmers interviewed agreed or felt comfortable with

current regulations. However, the remaining 19% disagree more or less with the current

environmental regulation framework (Figure 3-1).

The one farm that is not regulated is a small farm that received an invitation, but

does not want to participate in the cost-share program because the owner believes that

they can protect the environment in their own way. Fundamentally, the farmer does not

want to be controlled at all times. He is confident the problem will not increase from

present levels and that his farm will continue functioning as it has done for more than a

half century. However he is aware that if they are forced into compliance with

regulations, he would probably go out of business.

3.3.2.1 Farmers in disagreement with regulatory measurements

One of the farmers in dissension was very disappointed and very upset about the

regulatory agencies, especially with the Florida Department of Environmental Protection










(FLDEP). He felt betrayed in the fact that he was once told that he runs an exemplarily

clean farm and was even publicly recognized for that fact on one occasion. Later, after

activists groups sued the FLDEP, the department requested he make a cumbersome

investment on his own. The farmer stated

Their stuff is not scientific, they listened to what people say and they got
scared. They do not have any money for me, as they have been able to
provide for many other dairies. They have put 18 people out of work on
this farm because I had to reduce the herd size by half. If they place just
one more regulation on me, I will be out of business

The indispensable requirement for scientific studies in the eyes of this farmer is an

interesting and novel point. Often, the opposite view might be expected. Apparently,

the FLDEP in this case is regarded as a politically driven agency that does not necessarily

back up their case with science. A perception of inequitable treatment on the part of the

cost-share program with regard to this farm might be regarded as a strong motive for this

farmer's ill feeling toward the regulation process.


Agree completely
33 %
-I Agree --

81% %-
Criticize part of it
67 % --

N regulations


SAdministrative observations
50 %

S Don't Agree

R Technical observations
50 %


Guarantee of sustainability


Climate conditions not considered
Rapid personnel turnover
Not enforced severely
Are BMPs really good?

Listen more to public and news
Don't treat farmers equally
Too bureaucratic
Too imposed

Are BMPs really good?
Whole farm system understanding
Lack of soil science/scientists


Figure 3-1. Farmers' perceptions of environmental regulations in north-Florida


L









This farmer elaborated further saying that the FLDEP requested he install more

concrete free-stalls to collect manure and more irrigation systems to dispose of the

effluent in the fields. He thinks that is not a correct solution because it would create even

more problems with neighbors concerned with odors. He added

I did not even have a chance to discuss my options. The regulation was
imposed. They told me: do all this or reduce your herd size. I decided to
reduce the herd size because I felt that I could not trust the regulatory
people any longer. It would not be surprising if, after I make all these
investments on my own, after a few years they would come again and ask
me to change all over again because the neighbors are not happy

This farmer expressed a common complaint in processes of this nature. He was left

out of the process and not consulted from the beginning. Furthermore, it is apparent to

the farmer that no matter how hard he tries; the dairy farm is in a vulnerable position with

respect to new regulations in the future. In this farmer's opinion, the interests of the

farmer will always be secondary to those of the "neighbors."

Another complaint regarding the regulations was from one relatively new farmer,

who had to fight for more than two years to get a regulatory permit approved for a small

dairy operation. The farmer considers himself to be over-regulated compared with other

much larger dairies in the area. The farmer also mentioned that the approval process has

been very costly and the farm could not count on any financial aid. He felt that he was

forced to go through the process. This farmer mentioned I have fewer than 400 cows and

they made me install and monitor six wells with all the expenses it implies while there are

farms with thousands of cows who do not have this type of control. This farmer feels that

reports in the news have been particularly prejudicial to his situation and probably one of

the causes of this very strict permit. The other two farmers who do not agree with current

regulations based their perceptions more on technical issues. Both considered that the









suggested BMPs would not be the most appropriate either for the environment, or for the

farms. One important point was the fact that the main BMPs promoted include a flushing

system (dilution of manure to spray it in fields) that, to quote one of the farmers:

Seems to be aimed at losing most of the N to the air instead of trying to
use it as fertilizer. Besides, the promoted system requires a big investment
and is fixed to a physical space. Why do they not promote a system, like
the one in the northern part of the country that uses slurry (with minimum
flushing) and gives the farmer the flexibility to apply it to different plots
or even easily export it off the farm?

The farmer offered a strong critique in this statement of the technical expertise of

those in charge of designing and implementing BMPs. The producer apparently

perceives that the system proposed simply postpones further regulation for the moment

by converting a local negative externality into a diffuse global one. In essence, N is

being spewed into the atmosphere instead recycling it for other uses.

The other important point was the lack of understanding by those creating BMPs of

the dairy farm as a whole system. One of the farmers mentioned that very capable people

exist on the regulatory teams, but there is a lack of personnel with an interdisciplinary

understanding of the dairy system. He has witnessed problems other dairies have had

because the engineered designs did not take into account all the components of the

system and their interactions. Indeed, he is aware that some BMPs are being revisited and

redesigned after several failures have been detected.

These two dairymen think that the soils are the main component in solving the N

issue. They think that farms and regulatory agencies should work more aggressively with

a focus on soils. They both believe there is a lack of soil scientists on the regulatory

teams. The result is that regulatory agencies are not doing a good job protecting the

environment taking into account soil dynamics. One of these farmers added that the









regulatory agencies should clearly distinguish between real data and perception

information regarding water pollution by N

I am not completely sure if a real problem exists. Studies of the trends of
N pollution are not completely conclusive. Nor do I recall were they
conclusive in the case of the P issue in the Okeechobee area, where
regulatory agencies imposed a limit of P run-off discharge lower than
what common rainwater would have

Some farmers hold a less than credible reputation of the FLDEP. This poor

perception dates at least as far back as the years of regulation in the Okeechobee area.

The loss of dairy farm livelihoods in that area was never proven conclusively to be the

result of a tangible problem, at least in their minds.

Another informant also corroborated this last point regarding mistaken regulation

standards. The last two farmers mentioned apparently feel that precaution is not

necessary and that only when science proves without a shadow of a doubt that something

is harmful should regulations be imposed. These farmers apparently did not consider the

notion that it might be too late for action by then.

3.3.2.2 Farmers who agree with regulations but criticize part of it

Among respondents who agree with current regulations, some criticized some

particular aspects related to the regulations. Two farmers mentioned that FLDEP fines

farms without taking into account the climatic conditions that, for example, determine

inevitably high N measurements in lysimeters when wet periods follow dry ones. Another

point was the constant change of personnel in the regulatory agencies that could be

disadvantageous to farmers. This is because over the years, officials become acquainted

with specific farm situations. When new agents are hired, there is a period required for

building-up new relationships and understanding site-specific problems. Besides this,









new, usually young agents, tend to be more environmentally oriented. Often, they fail to

grasp the fact that farmers also need to make a living.

On the other hand there was also one farmer who criticized FLDEP and the

regulations as being relaxed, implying that the regulations should be enforced more

aggressively in order to avoid intervention by the federal Environmental Protection

Agency (EPA). He perceived EPA regulations and enforcement as potentially much

stricter, which would be very difficult for all in the area.

Another criticism of farmers who do agree with the regulations came from one

farmer who believes that a third party plays a role too important in the regulations. This

"third party" is a term used to refer to activist groups that raise public concerns about

environmental issues. In this farmer's opinion, FLDEP takes their opinions into

consideration disproportionately. The argument was that these third parties have nothing

to lose and that the FLDEP listens more to activist groups and the public, and takes

farmers' concerns and situations little into account.

Other respondents corroborated the fact that the news media have a great impact on

regulations and on FLDEP enforcement of regulations. In terms of regulations and

enforcement, the opinion of one journalist could have a greater worth than the opinion of

a number of farmers.

Some other farmers, who in general agree with current regulations, mentioned their

concerns about the scientific basis behind the requested technical actions, which may not

be the most appropriate for specific farms. This is related to the previously mentioned

perception that the regulatory agency is not looking at the overall system. One farmer

expressed: some times it is easy for them to tell you what you have to do. But, usually









they do not know how difficult it is to implement these actions, and what other impacts

could result from implementing the management changes.

Many of the farmers expressed their concerns regarding other non-point potential

sources of pollution. The main argument was centered on challenging the fact that dairy

farms are the most contaminating of all. They believe that this statement still needs to be

proven, and in any case they ask why are not row crop farms, golf courses, human

settlements, among others regulated in the same way as dairy, poultry or pig farms?. One

farmer mentioned: they count every pound of N the cows excrete on this farm but they do

not control the neighbor's farm that uses two to three times more fertilizer than the

recommended rates for vegetable crops.

Another point mentioned by several farmers may have important implications for

regulations. Many people are regularly moving to live in the countryside in Florida.

Often, these newcomers find that they do not like the manure odors until after they are

well established. Most farmers think that newcomers should be aware of what they are

buying into, and expect and accept normal rural odors. However, there are a few farmers

who actually believe that just because farmers were established first is not enough

argument to disregard odor problems of newcomers and they should work trying to

mitigate it. One farmer is convinced that in the future, odor problems, more than any

other issue, will put his farm out of business.

Farmers are also looking for permanence and consistency in the regulations. Some

farmers are concerned about potential changes in the set of regulations that might turn out

to be disastrous for their conditions. Farmers have learned to survive under current

conditions as long as the FLDEP does not change the basic, initial guidelines. This fact is









related to a concern expressed by several other farmers regarding the potential that

regulations might become unreasonable or impossible to implement if they change. This

eventuality is perceived as even more uncertain because farmers understand that the

FLDEP follows directions from the EPA. Farmers feel a total disconnect from the

federal agency.

Almost all interviewed farmers agree that the compliance process with the

regulatory agencies is not a money-making adventure. All are conscious that it costs

money, but they have no choice but to comply.

3.3.2.3 Farmers who completely agree with regulations

At least 33% of the 81% of producers that agree are completely at ease with the

regulations. All of them believe that they are working towards a sustainable way of doing

business. Thus, the regulations are perceived as a guarantee of survival in the long run.

For many of them there is a history of success in north-Florida because of the

cost-sharing program that is allowing them to work together with the research, extension,

and regulatory agencies in searching for regional and site-specific solutions. One farmer

mentioned: the Partnership and the incentive-based program in the Suwannee River Basin

are wonderful things. These will help farmers to be better off in the future. Although

many farmers realize that mistakes have been made in the past, they believe that all

parties are learning from the experience and that everyone together is building expertise

to protect the environment and keep farmers in business.

From all these farms there is a sense of engagement in the program for protecting

the environment that makes them proud and eager to implement environmental friendly

practices even if they do not have an economic return. They believe in general that the

regulations and mostly the participatory approach have been correct. Also, many of them









appreciate the fact that the program is flexible and can be discussed with the regulatory

agencies.

One farmer mentioned that they are trying not only to comply with the current

regulations, but are proactively seeking to exceed these regulations. This producer warns

that he is not necessarily looking for a new set of rules, but rather embracing the

regulations to become better stewards. This farmer predicts that FLDEP will be willing to

face the public for a farmer if the farmer is on the right side at all times. This fact would

be much more important in the future because regulations will become much stricter, not

only in the dairy farm industry, but in all aspects of human life. This farmer feels farmers

will be better prepared for those times thanks to the Partnership's efforts.

3.3.3 Perceptions of Climatic Technology Innovation

Recent advances in atmospheric research have made it possible to predict climate

variability with a relatively high level of probability. Knowledge of climatic conditions

may allow improved seasonal management strategies for dairy production in Florida.

Once again the opinions were divided. Of the 21 farmers interviewed, 14 (67%)

mentioned that they would use improved forecast information, 5 (23%) would not use

this technology, and 2 (10%) were not sure (Figure 3-2).

3.3.3.1 Farmers that would use long-term climatic forecast

All 14 farmers who believe they would use improved long-term climate variation

forecasts say that their main use of the information would be in planning crop systems.

One producer mentioned he already uses ENSO phase information to plan the farm's

crops:

I make decisions every year based on my beliefs of if the seasons will be
dryer/wetter, cooler/hotter. For example, for winter, if I expect an El
Niio, I will plant oats because they do well in moist conditions. If it will










be a La Nifia, I will choose rye because it handles drier conditions much
better. Of course, I will always plant ryegrass and clover, no matter what
the climate prediction

This farmer thinks that forecasts would also be useful for adjusting planting dates

in order to obtain better yields and more N uptake. He has yet to implement this

management strategy. Another farmer also uses long-term forecasts but not based on

ENSO phases. He relies instead on the experiential knowledge of an old, experienced

person who predicts the up-coming seasons in terms of wetter/dryer and colder/warmer.

This in turn determines the crops to be planted. This farmer was very interested in the

ENSO forecast information and tools and inquired as to its availability.


Figure 3-2. Farmers' perceptions of climate technology adoption

In the same manner as the farmer who already uses ENSO forecasts, most of the

farmers believe that the use of the forecast would be in crops. Mainly, the producers

would use information for substituting crops in specific seasons and to change planting

dates. For the winter season, oats and rye were mentioned several times as


Crop systems -- Different crops/vat
100 % -- Planting date,
| Would use
--C7 Area planted
67 %
-- C Drainage system
-- | Irrigation system

Climate Milking cow group
Technolg- Livestock
Technology 36 Breeding date plain

-- | Nutrient mgt 1 Solids collect(
14%

-- | Not sure ]--] Would like to have the forecasts
10 %
Non-flexible and/or satisfied with current system
-- Would not use
23 oo Other more important things to worry about


rieties
s


m
m

rotation
inning

n









interchangeable crops. Most agreed that drier years (La Nifia) are better for rye, and

wetter years (El Nifio) better for oats and/or ryegrass. However, some other producers

mentioned that temperature is the most important variable, and that hotter winters with

less chance of freezes (La Nifia) would be better suited for oats. For the case of freezes,

some farmers agree that if the prediction could address those events, they would be very

useful because that would let them know how much crop production to expect in winter.

Based on this they would adjust crop plans. For example, if there were going to be more

than the typical 20-25 freeze days, winter forages would be affected. Therefore, they

might plant more area in order to produce the required amounts of feed.

Climate forecasts for the spring/summer season would be more beneficial for

planning the planting date for corn. Those who plant corn say that the earlier you plant,

the better the results. If farmers know there will be little risk of late freezes, they could

plant as early as possible, for example in the first week of March.

For the summer/fall season another important decision would be between planting a

second crop of corn or sorghum. Farmers mentioned they would plant a sorghum crop if

they know that drier than normal conditions are expected. Potential changes in choice of

crops refer not only to different species, but also to cultivars within species. Some

varieties have different responses to climate conditions. Farmers mentioned these

possibilities without entering into specifics.

Another group of informants mentioned an additional possible use of climate with

respect to crops. One farmer mentioned they could use the prediction information to alter

the area of planted crops. In wetter conditions area would increase. In drier seasons, less

area would be sown. Another farmer mentioned that climate data would be useful for









planning drainage systems, especially before a wetter summer. A further potential use

mentioned was in planning irrigation systems management. Although these systems are

mostly fixed, farmers could make some adjustments if they know what the seasonal

climate would be like, especially regarding rainfall.

Besides the use of forecasts for crops, another potential use of climate forecasts

would be for the livestock. Three out of the fourteen interviewees who would use the

forecast elaborated on this issue. One farmer mentioned that she could plan her cow

rotations differently. Holstein cows are very sensitive to "heat stress" calculated by an

index that takes temperature and humidity into account. When the index is high, in hot

and humid months, milk production is affected negatively. If seasonal variability toward

hot, humid conditions is known ahead of time they would be able to better plan the

development of cow groups. This farm, like others, has comfort facilities where milking

cows are protected from heat stress, that is, climate effect is diminished. However, on

almost all farms, these milking cows spend part of their time outside. The farmer found it

logical to think of planning ahead of time to provide the most protection time to the most

productive group of cows.

One more decision that could be managed in light of climate forecasts is breeding

time. This would be done by changing the number of cows entering the milking group in

a certain season according to the forecast. For example, if a mild summer were expected,

then the farmer would breed the cows in time to have more milking cows than normal in

production for that summer. Climate knowledge would allow higher than normal

amounts of milk to be produced in summer. Similarly, decisions regarding purchases of

replacements heifers could be improved based on expected climatic conditions.









Typically, farmers buy replacements in fall. If spring or summer were going to be colder,

they could take advantage to buy less expensive replacements before those seasons.

An additional possible use of forecast mentioned for livestock was for the winter

season. Specifically climate forecasts would be used to anticipate colder than normal

winter seasons, which are also detrimental for the cows. In those situations, farmers could

plan ahead of time to have more forage (either planted or purchased) available. Extra

amounts of fiber would help cows stay warmer.

Another two farmers mentioned the potential use of the forecasts for planning

feedstocks. The inventory of feed is a major management task on dairy farms. Therefore,

knowing the climatic conditions ahead of time would be important to preparing

feedstocks (either produced or purchased). For example, one farmer mentioned that if he

would know ahead of time that the next season is going to be dryer, he would plan to

have more forage for his heifers and dry cows. The farmer mentioned additionally that

not only would it be good to plan for feed shortages, but also to plan for excess feed

produced. Too much feed on hand is also a problem. If the predicted conditions for next

seasons are good for forage production, then producers could plan not to buy feed. They

could also sell feed in advance, or decrease acreage planted.

Additional to the use of forecast for crops and/or livestock, three farmers

mentioned that the information could be used for improving N and P management on the

farm. Waste management systems are mostly fixed, so considerable changes to this

system is not an option. Farmers usually change daily management based more on daily

weather than long-term climate, however there are some things that could be planned, as

is the case of for example, of collecting or not collecting solids from the waste









management system. These farmers mentioned that if they know that there will be wetter

conditions, they would prefer to prepare the whole system to not collect solids because

the pile of solids resulting from the collection would be difficult to handle under wetter

conditions. This major change would determine that the effluent will have more solids in

solution. Therefore, the irrigation must have higher pressure and the crops must be

prepared to receive a slightly higher percent of N in the effluent. If they know that the

conditions would be dryer than usual they would prepare to collect all the possible solids

from the waste management system and together with this action they would plan what to

do with these solids, apply to fields of the farm or export out of farm.

3.3.3.2 Farmers who would not use long-term climatic forecast

Farmers that mention they would not use climate forecasts as a technology to help

in decision-making argue that they have relatively "fixed" systems. In their rather

tightly-bound systems, no changes in management need be planned. One farmer

mentioned: I am married to my system operation... it works and I don't want to change

anything that could mess things up. These farmers said that their decisions are based

more on the short term. If they knew there was no rainfall predicted in the following

week, they would cut hay. It is difficult for them to envision how they might change

tasks months in advance. One producer noted: unless something as disastrous as a

hurricane is forecast, I would not do anything different. Another farmer argued: we have

much larger problems to worry about. When we solve those other problems we might

start thinking about incorporating climate forecasts in our planning. Another farmer was

very convinced that:

We are not going to change anything even if we know that a very
abnormal season is coming. Even then, we would continue to do as we









always have, that is, adapt to the weather conditions, making changes to
respond to the weather conditions

Other farmers may share this farmer's skeptical view of the technology. Those

who are naturally averse to change or have not been exposed to new technologies will

exist.

3.3.3.3 Farmers who are not sure they would use long-term climatic forecast

The two farmers in this category had never previously thought about

climate-related management changes. They mentioned that the entire concept was

completely new to them, so they would need time to digest the idea. Initially, however,

they were not sure if they could incorporate climate variability forecasts into their

decision-making process. Both these farmers mentioned that, even though they did not

know if they would use the technology, they would like to be aware of the forecast to

consider it in regard to their performance in the following seasons.

3.4 Conclusions and Recommendations

All stakeholders included in the Sondeo had a general consensus that growing N

levels in the Suwannee River Basin is a real and current problem. However, stakeholders

included in the focus groups, and especially dairy farmers challenged this fact during the

personal interviews. People interviewed during the Sondeo showed more awareness of

the real consequences of increased levels of N in water bodies. Dairy farmers and other

stakeholders involved in the interviews and focus groups, even though they recognized

that the presence of high levels of N is detrimental for the environment, showed a certain

degree of skepticism regarding the gravity of the problem.









Every effort needs to be made to communicate more clearly to the general public

real trends of N in water bodies and their potential consequences. The message must be

conveyed using consistent scientific data.

Most of the informants perceived that dairy farms are the highest polluters.

However, many stakeholders included in the Sondeo, focus groups, and interviews

questioned this popular belief. Dairy farmers in particular, perceive that there is a

misconception of the dairy farm industry and its environmental impact. Further research

should be devoted to investigate the contribution to contamination from each potential

source. It is important that individual operations be held accountable. Point vs. non point

N source pollutions must be further studied. Analyses of long-term monitoring studies

comparing different industries and their environmental impacts must be undertaken.

Farmers are willing to participate in incentive-based programs such as those

promoted by the Suwannee River Partnership. The principal driver behind this

willingness is a desire to avoid being regulated by official agencies. The Suwannee River

Partnership has a very positive perception and support from the stakeholders. Most of the

stakeholders perceive that the promoted BMPs are beneficial to the environment but are

not cost effective to the farms. Some farmers believe that the BMPs could be

re-engineered to better serve their purposes according to specific farm conditions. Further

research and validation of BMPs must be conducted. A guideline describing the steps to

develop BMPs should be created, and an economic feasibility study on BMPs that takes

heterogeneity of farms into account should also be conducted. Extra funds should be

devoted to expanding the impacts of the Partnership.









Farmers feel that environmental regulations are becoming stricter and more

uncertain in time. These regulations represent one of the biggest challenges for the

future. Farmers want to express that they are doing their best to protect the environment.

A vast majority of them agree with current regulations but they have serious observations

to make regarding the technical and administrative procedures. The Partnership is helping

to change farmers' attitudes, but still more communication and outreach work is needed

to gain the complete understanding from farmers.

Two-thirds of the interviewed farmers believe they can use climate variability

forecast proactively. All of them would use it primarily for crop management.

Additionally some could use it for livestock management. Some could use the seasonal

climate variability information for feedstock management, and some could apply this

technology to the management of the waste system.

Further research calls for the inclusion of climate variability forecasts into existing

BMPs experiments. Studies that explore the interaction between climate and nutrient

overflow in dairy farm systems must be promoted. Farmers should be involved from the

early stages of these research processes.

The main barriers for potential adoption of climate forecast technology by farmers

are the "rigidity" of dairy farms and the lack of management resources. Agencies

interested in delivering climate forecasts as an innovation must identify the narrow niches

where they may be used based on the experience of others. The technology should be

translated into clear and rapid management adjustments that farmers could apply without

involving excessive administrative transaction.









The fact that some stakeholders have never before thought of the potential

application of climate forecasts denotes a lack of awareness of available forecast

innovations among the stakeholder community. Step one in any diffusion process, i.e.,

awareness of the existence of a given technology, is missing among dairy farmers in the

study area. Commitment to extension and outreach programs that make this information

available to all networks of stakeholders and especially to dairy farmers is essential to the

future success of incorporating climate-based management into overall schemes to reduce

N loading into the environment.

The participative process described in this study underscores the critical need to

include farmers and other major stakeholders in every stage in order to assess real

perceptions that further will be used for consensual problem-solving. Its relevance is even

higher when complex problems exist as is the case of the interactions of N pollution, its

regulations, and the potential use of climate forecasts as a way of mitigate them.














CHAPTER 4
A METHOD FOR BUILDING CLIMATE VARIABILITY INTO FARM MODELS:
PARTICIPATORY MODELING OF NORTH FLORIDA DAIRY FARM SYSTEMS

4.1 Introduction

Much work has been published on the mathematical description of models and

model outputs. Conversely, little effort has been invested to date to describe a process of

stakeholder interaction or participatory modeling that might be essential for increasing

adoptability of models among end users. Models are often created to simulate real

conditions, predict potential outcomes, and ultimately be a tool used by stakeholders.

Many models have failed in this last requirement because they do not represent the

conditions relevant to stakeholders nor do they comply with requirements of the final

users for simplicity and user-friendliness.

Substantial changes must be incorporated into all phases of model development to

improve end-user adoption. One important aspect required is intensive and effective

participation of all stakeholders (Stoorvogel et al., 2004). This study highlights the use of

stakeholder interaction methodologies-originally developed for use in international rural

development-to first gauge the need for, then refine these management products and

tools. Working with stakeholders in a learning process involving as many feedback loops

as necessary will increase the probability of higher adoption rates.

The traditional system analytic approach emphasizes the importance of quantitative

modeling, rational planning, and economic analysis. This is closely linked with

operational research that looks for optimal solutions for complex problems (Rosenhead,









1989). However, the application of traditional system analytical methods to complex

societal and policy problems has not always been helpful because problems of validation

and because the absence of adequate social scientific theories and empirical data

(Stoorvogel et al., 2004; Jones et al., 1997). One efficient way to elicit up-to-date and

high quality empirical data is through stakeholder interaction (Geurtz and Joldersma,

2001).

During the 1980s, systems analysts realized that much of the understanding of the

problem is generated in the process of model building; consequently, participation of the

stakeholders in the model building process is useful (Geurtz and Joldersma, 2001).

Therefore, the participation of the potential final users should enable the researcher to

enrich the model by including subjective sources of knowledge in addition to the

"objective" knowledge derived from theories and empirical studies (Geurtz and Vennix,

1989; Verburgh, 1994).

Systems approaches at the farm scale must include not only the biophysical

component, but also the social, economic, political environment of the farm together with

a "bottom-up" approach (Jones et al., 1997).

Development in methods of modeling analysis and in practice shows a shift from a

uni-central, analytic, scientific approach to a more multi-central, interactive stakeholder

approach. These shifts are accompanied by a growing interest in participatory styles of

policy making and participatory methods of policy analysis. The application of the

participatory methods generally follows the paths of logic-in-use, which shows the need

for a methodology for participatory policy analysis (Geurtz and Joldersma, 2001).









Participatory Rural Appraisal (PRA) describes a growing family of approaches and

methods to enable local people to share, enhance and analyze their knowledge of life and

conditions, to plan and to act. PRA has sources in activist participatory research,

agro-ecosystem analysis, applied anthropology, field research on farming systems, and

rapid rural appraisal (RRA). Natural resource management (NRM) and agriculture are

fields in which PRA finds applications (Chambers, 1994).

Farming systems research (FSR) advocates for interdisciplinary/participatory

approaches to deal with farm scale simulations (Norman and Collison, 1985).

Nevertheless, some authors have pointed out that FSR is often location specific and

descriptive (Dent and McGregor, 1994; Jones et al., 1997). An alternative approach is to

study the complete regional production system in close interaction with farmers.

Vereijken (1997) developed prototyping as an alternative to the more model-oriented and

farming system approaches. Prototyping procedures have a participatory character of

scientists directly collaborating with farmers.

Simulation-aided discussion about crop management has underpinned advances in

farming systems analysis as a vehicle for management intervention (Keating and

McCown, 2001). Similarly, McCown (2001) and Hammer et al. (2000) both argue that

connectivity and dialogue among key players is essential for achieving relevant and

significant intervention. In this sense, the dialogue around the simulation process could

be as important as the underpinning models, although it is clearly reliant on their

existence and influenced by their reliability (Nelson et al., 2002). Meinke et al. (2000);

Robertson et al. (2000); Keating and McCown (2001); and Meinke et al. (2001) provide

examples of participatory approaches using simulations to drive relevant and significant









results. These approaches are not about science simply providing the answers for

management to practitioners, but rather employing cooperative learning to develop

solutions (Nelson et al., 2002). Direct participatory action research with farmers has

helped establish credibility of models and simulation analyses (Meinke et al., 2001).

According to Archer et al. (2002) if the model simulation software is created for

farmers, extension agents, and/or farm advisers, it should be designed to be user-friendly

and to require minimum data inputs. Also, Archer et al. (2002) argue that it is not only

important that the model be able to predict what is required, but also to make the

predictions quickly and easily, in a readily accessible form, for site-specific conditions, in

real time. Both the participatory and the user-friendly approach were used as the

methodological basis for developing a climate-influenced dairy farm model for

North-Florida. The benefits of involving farmers in research to encourage engagement

and learning for scientists and farmers are well documented (Paine et al., 2002; Onema et

al., 2001).

North-Florida's climate is highly affected by El Nifio Southern Oscillation (ENSO)

(O'Brien, 1999). Since agriculture is the activity most vulnerable to climate variability

(Hansen, 2002b), farmers might benefit from new methods of seasonal climate forecasts

connected to system simulation models to tailor management options regarding

agricultural, economical, and/or environmental outputs. Intensive technologies offer a

broad range of options for adaptive responses to climate information (Hammer et al.,

2000). Dairy production in sub-tropical north-Florida is a very intensive system. The

Southeastern Climate Consortium (formerly Florida Consortium) was created as an

interdisciplinary, multi-institutional partnership involving researchers from 6 universities









in the states of Florida, Georgia and Alabama. The research group found that it is critical

to involve stakeholders actively from the beginning of a process to understand the effects

of climate variability on water and agriculture. Agencies with strong established

relationships and trust with end-users must be engaged to improve the chances of

delivering operational applications of climate information to final users (Jagtap et al.,

2002). One such application to emerge is the dairy model whose development process is

the subject of this paper.

It is not the aim of this paper to describe the mathematics of a dairy farm model.

The main aim is to describe the process of collaborative creation of a regional

north-Florida dairy farm model adaptable to any individual farm. The purpose of the

model is to tailor management strategies for different climatic conditions by estimating

monthly rates of N leaching and profit. An overriding objective of the model as

expressed by stakeholders was to decrease environmental impacts while maintaining

profitability. Specifically, the study sought to engage stakeholders in the process of

creating, calibrating, and validating a whole north-Florida dairy farm model (Figure 4-1).

4.2 Materials and Methods

The study was undertaken in north-Florida's Suwannee River Basin, covering part

or all of Alachua, Gilchrist, Levy, Lafayette, and Suwannee counties. An adaptation of

the emerging approach in Farming Systems Research and Extension presented by

Bastidas, 2001 (Figure 4-2) was used. Interviews and focus group meetings were carried

out on a continuous basis, divided into five phases. Phase One consisted of the initial

developments. Phase Two was about the sounding out or rural appraisal. Phase Three

consisted of greater stakeholder involvement. Phase Four included meetings and

interviews to elicit feedback from stakeholders on the prototype model. And in phase









Five, the tentative final model was taken to stakeholders for the validation, their final

suggestions, and recommendations.


Researchers


Regulatory
Farmers Prototype Agents
[0 Model













Figure 4-1. Schematic of participatory modeling

4.2.1 Main Framework

In Figure 4-2, all stakeholders involved in the issue of N pollution in groundwater

by dairy farms are referred to as "users." Users must be identified and their differences

recognized. Diverse users have specific and diverse needs. These needs must be

identified and considered before starting the modeling/simulation process. PRA methods

are ideal for initiating and maintaining high stakeholder involvement. This can be

characterized as a continuous learning process.

In Figure 4-2, arrows among users, needs, and PRA indicate the iterative and

dynamic process among researchers and stakeholders to determine felt needs that must be

incorporated in the modeling. A following step includes the integration, creation, and

improvement of models. Once again, the arrows indicate the iterative and dynamic









researcher-stakeholder process to improve models in order to make them more amenable

to final users. Inclusion of prototype models during this time is useful.

The process in which adjustments to models are made in order to better represent

reality is called calibration. Because it is a continuous process, calibration moves

seamlessly into validation when fewer and fewer major changes are required. Validation

is the part of the process that is undertaken after researchers and stakeholders agree the

model fits real conditions adequately. At this point testing models against independent

data, i.e., data from stakeholders who had not participated in the creation of the original

models is used.




End products Specific and
Users Diverse
needs




Calibration and
Seating Validaio Participatoryi
Prototype Rural
products Appraisals


Models
Climate \4 Crop
Data models


Figure 4-2. Methodology of participatory modeling. Source: Adapted from Bastidas
(2001), p. 19.

After final adjustments and validation, a final product flexible enough to fit the

diversity of all users is the goal of the interactive process. Because it has been a









participatory product, higher than normal adoption is expected along with higher than

ordinary usefulness.

4.2.2 The Sondeo

A rapid rural appraisal or Sondeo (Hildebrand, 1981) was conducted between the

months of January and February 2003 by a multidisciplinary team of eight graduate

students from the University of Florida (Alvira et al., 2003). During the Sondeo 13

stakeholders were interviewed. This sounding out was undertaken to get a better

understanding of the broad issue of N pollution in water. Another objective was to

understand the dairy farm systems in the Suwannee River Basin and initially identify the

type of modeling products that might be helpful for proactive management in light of

climate variability forecasts.

The Sondeo team consisted of graduate students from a number of disciplines

including agronomy, animal science, agricultural extension, natural resource

management, forestry, and community development. Two to three-people teams

conducted transects and stakeholders' interviews. The interviews were conducted as

informal, open-ended conversations. After a team introduced itself and explained the

subject of the research, the stakeholder and the team conversed about any aspects of the

topic that were relevant. This open-ended approach enabled topics to emerge and be

pursued that might have been missed if the researchers had been constrained by a

pre-formulated questionnaire. Despite the open-endedness, a set of core themes was

discussed in almost all interviews. Notes were not taken. Following each interview the

partners wrote individual notes, compared them, and then wrote a joint report of the

interview. Finally, these joint reports were shared and thoroughly discussed among all

Sondeo members. Each team presented its findings, which could be clarified, challenged









or contrasted with the results of the other teams. This process of reporting and discussion

served as the opportunity to begin noting trends, gaps in information and new questions

to be pursued.

4.2.3 Interviews

Personal interviews were undertaken by the researcher with dairy farmers and/or

dairy farm managers during the fall/winter seasons of the years 2003-2004. Open-ended

interviews (Bernard, 1994) were undertaken with a purposeful sample of 24 farmers

(-35% of population) with the main aim of engaging these actors. Specific operation

information was elicited and obtained as feedback in the modeling process. The sample

size and the people to be interviewed were chosen in order to obtain as much of the

existing variability as possible. Selection of participants was done in a participatory

manner with the coordination of the Suwannee River Partnership (SRP) and the

University of Florida Extension offices in the area.

Interviews started with the presentation of either the prototype or the final model on

a laptop computer. A voice recorder was used in every interview after requesting

permission. The first step was to explain the purpose of the visit and then to provide a

general description of the model. This was followed by a detailed explanation of every

component in the model, including those that were as yet not active, and their mechanics.

After discussing similarities and differences between the model and the farmers' specific

production units, a demonstration of the running of the model was presented and

discussed.

Later, dairymen had the opportunity to input their specific data and run the model

for their own specific conditions. Similarities and differences with what farmers expected

were recorded and taken into account. If observations were easy to adjust for, they were









tried "on-the-fly," taking advantage of the user-friendliness of the model's interface. If

specific comments would require more work, they were incorporated in the lab and tested

in later interviews.

Interviews highlighted farmers' reactions to the models. All possible suggestions to

be incorporated in the models were taken into account in successive steps. Two rounds

of interviews were conducted. Twenty-one farmers observed the prototype model in

October-November of 2003. In February-March 2004, the definitive model was used for

validation purposes with 3 farmers.

4.2.4 Focus Groups

Eight focus groups (Bernard 1994; Stewart and Shamdasani, 1990) with

stakeholders were held between November 2002 and February 2004 in the study area.

Focus groups consisted of open discussions among 5 to 15 persons. A stakeholder in

coordination with the researcher called these meetings. The researcher facilitated the

meetings around the different topics of research. The first two focus groups did not

include any model as a basis for discussion. Rather, because the creation of a prototype

model was suggested and encouraged during the second focus group, one was created and

became the focal point for discussion in all groups save the last. In the last focus group,

the final model was discussed.

Focus group participants ranged from specialists from official agencies such as the

Suwannee River Partnership (SRP, www.mysuwanneeriver.com/features/suwannee+

river+partnership), the Suwannee River Water Management District (SRWMD,

http://www.srwmd.state.fl.us), the Florida Department of Environmental Protection

(FLDEP, http://www.dep.state.fl.us), the United States Department of Agriculture

(USDA, http://www.usda.gov) Natural Resources and Conservation Service (NRCS,









http://www.fl.nrcs.usda.gov), the University of Florida IFAS Cooperative Extension

Service (CES, extension.ifas.ufl.edu), to private sector companies, technicians and

farmers. These participants were usually called to a meeting by the Suwannee River

Partnership and/or by extension agents of different counties in the research area.

Focus groups lasted between two and three hours and their main purpose was to

listen to feedback on the modeling process and to actively increase involvement and

diffusion of the research to a more target public.

4.2.5 Other Stakeholder Interactions

The principal researcher approached many other stakeholders between fall 2002

and spring 2004 in order to receive additional inputs into the modeling process. Initially,

the discussion centered on the intended idea of a proposed simulation methodology. In a

second phase, the discussion was based on the prototype model. Finally, interaction with

the clientele was based on the final model and its final tune-ups. These interactions were

meetings with specialists in the issue of the N leaching and the dairy industry and

included private consultants, officials from regulatory agencies, and university faculty,

among many others.

4.2.6 Analysis of Information

All participatory interactions were recorded, classified, organized, analyzed, and

described. The feedback received was included in the modeling process in one way or

another in successive steps. This paper summarizes the results of all the above

interactions in a descriptive/narrative manner. Names of the participants and some

organizations and companies are not revealed to protect the anonymity of interviewees. A

summary of stakeholder interactions can be found in Table 4-1.








78



4.3 Results and Discussion


Results of interactions are described in five phases that detail the iterative process,


step by step, of creating, calibrating, and validating the dynamic north-Florida dairy farm


model, DNFDFM with participation of stakeholders. This process is summarized in


Figure 4-3.




Phase 14nitial Developments

Firml conlctarouild-up of network
*Main* 5ur u eol ,nfliTnatan
*Adjusit reFsearr rarnewc mrnjolecrwves
Collaborllive commimernt of parts

Phase 5-The final model
*Model revistea anr new :ornmonenls of
*Madirk- *ClInmaT component
*Crop models
*Opbmrmtiaon Phase 2-The Sondeo
*Frinal m ini calibrated ana valndaleao -S n t
bltakeholdre Sunannoa Piver Pa tnrhiMp and
tK Cooperam-e Ertension Semnce on Doard
Conceptual.zabon of EMPs
*Wayi ol N pollbJorn rritationr
lord ofmod l Dieild up with staehoiaers
.Cornrnamar,nof parts in furtherF wrk






Phase 4-The prototype model
Phase 3-Grealer stakeholder
*Sewvpml eadbahc.s on prumloype mode L involvement
*Eperts parea consulted about prcloie i Undi .andinq f!Ndair fwsnatsr
*Pnoposed praiirled Ln Icmertti meeal.ng- .Dolein.nBnaclvIoc hr,'c model doinitcns
*Collcorn l im spln decif.c daU *el J Ialcn ol priohtype mrIel
*NRCS and private companion on board


Figure 4-3. Phases of the participatory modeling for creating the DNFDFM


4.3.1 Phase One: Initial Developments


In July 2002, four well-known scientists in Florida, each from a different discipline,


were approached looking for specific information and at the same time ideas about the


original research idea. The scientists were from the areas of forages on dairy farms, dairy


waste management, soils, and crop modeling, respectively. During this first exploration,


confidence was gained in the fact that most of the soil series in Florida are contained in a


database that could be accessible to use in later phases of the research (G. Kidder, pers.




Full Text

PAGE 1

MODELING NORTH FLORIDA DAIRY FA RM MANAGEMENT STRATEGIES TO ALLEVIATE ECOLOGICAL IMPACT S UNDER VARYING CLIMATIC CONDITIONS: AN INTERDISCIPLINARY APPROACH By VICTOR E. CABRERA A DISSERTATION PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY UNIVERSITY OF FLORIDA 2004

PAGE 2

Copyright 2004 by Victor E. Cabrera

PAGE 3

To my wife Milagritos

PAGE 4

ACKNOWLEDGMENTS I want to start by recognizing the exceptional interdisciplinary and collaborative work of my supervisory committee. I would like to express my eternal gratitude and highest appreciation to Dr. Peter E. Hildebrand, my mentor and adviser, for his unconditional support, encouragement, and outstanding guidance. My special thanks go to Dr. James W. Jones for his advice and commitment, acute professional challenging, and always useful feedback during the entire research process. I am also extremely grateful to Dr. Robert McSorley for his permanent, detailed, and constructive critique; Dr. Albert De Vries for his valuable inputs and constant assistance; and Dr. Hugh Popenoe for his interesting and useful comments. This study would not have been possible without the collaboration of stakeholders, those representatives of many organizations who were actively involved in it. I would like to gratefully recognize the involvement and the unconditional help of Darrell Smith, coordinator of the Suwannee River Partnership (SRP), who was a major actor in the networking of people and organizations, and supporting the research from the beginning. I am also very appreciative of the help and dedication provided by Chris Vann, Clifford Starling, Russ Giesy, and Marvin Weaver, University of Florida Cooperative Extension agents, who were the liaison with the dairy farmers. I also recognize the valuable information, feedback, and help provided by Hugh Thomas (SRP), Justin Jones (IFAS REC at Live Oak), W.C. Hart (US Department of Agriculture, Lafayette), Joel Love (SRP), and David Hornsby (Suwannee River Water Management District). iv

PAGE 5

I would like to particularly thank all the dairy farmers, small, medium, and large, for providing time, data, and critical feedback during the processes of creation, calibration and validation of the models created in this study. This study was conducted with the farmers for the farmers. Their participation was crucial to the successful advancement and culmination of this research. For obvious reasons their names are omitted. Besides members of my supervisory committee, many other professionals from the University of Florida were very helpful, providing information, networking, and inputs in this study. I would like to recognize the collaboration of Dr. Lynn E. Sollenberger for information about forage systems; Dr. Jack Van Horn for information about dairy nutrient management; Dr. Jerry Kidder and Greg Means for information about soil series; Dr. Donald Graetz for information, feedback, and networking on the soil/water nitrate issue, Dr. Sabine Grunwald for providing geospatial data; Dr. Clyde Fraisse for feedback on crop models and climate components; Dr. Carrol Chambliss for information and feedback on forage systems, Dr. Ken Boote for information on species parameters of crop models; Dr. Shrikant Jagtap for meteorological information; Dr. Ramon Litell for help and inputs in statistical analysis; Dr. Norman E. Breuer for information about north-Florida forage systems and constructive feedback; Cheryl Porter for help running the crop models and using the DSSAT software, Stuart Rymph for support in creating and calibrating forage crop models; Marinela Capanu for support with statistical analyses using SAS software; Daniel Herrera and Sergio Madrid for touring dairy farms with me; and Amy Sullivan for constructive feedback. Appreciation is extended to Dr. Gerrit v

PAGE 6

Hoogenboom from the University of Georgia, for his comments on species files for the crop models. I want to express my appreciation to the interdisciplinary Sondeo team for their enthusiastic and productive work in the early stages of this study. Thanks to Diana Alvira, Sebastian Galindo, Solomon Haile, Kibiby Mtenga, Alfredo Rios, Nitesh Tripathi, and Jose Villagomez. I would like to gratefully acknowledge the support and unconditional help provided by a private enterprise, the Soil and Water Engineering Technology, Inc. My sincere thanks to Dr. Del Bottcher, Dr. Barry Jacobson, and Suzanne Kish not only for their direct feedback and first-hand valuable information, but also for inviting me to participate on the monthly manure lunch that was a great opportunity for networking and collecting information. I want to express my thanks to Jesse Wilson, Bill Reck, and Nga Watts from the Gainesville office of the Natural Resources and Conservation Service (US Department of Agriculture) for providing feedback, constructive critique, and valuable information for the study. I hope I am not missing anyone, but if I do, I want to express my appreciation to all and each one who in some way or another contributed to this study. Your help has been highly valued and critical for the final outputs. This work was supported by a grant from NOAA (Office of Global Programs) through the South Eastern Climate Consortium (a Regional Integrated Science Application center) under the direction and guidance of Dr. James W. Jones and Dr. Peter E. Hildebrand. vi

PAGE 7

TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES............................................................................................................xii LIST OF FIGURES.........................................................................................................xiv LIST OF ABBREVIATIONS..........................................................................................xix ABSTRACT.....................................................................................................................xxi CHAPTER 1 INTRODUCTION........................................................................................................1 1.1 Background........................................................................................................1 1.2 The Problematic Situation..................................................................................1 1.3 Research Area....................................................................................................6 1.4 Objectives..........................................................................................................6 1.5 Hypothesis..........................................................................................................8 1.6 Framework.........................................................................................................8 1.7 Methods............................................................................................................11 1.8 Relevance.........................................................................................................13 2 NITROGEN POLLUTION IN THE SUWANNEE RIVER WATER MANAGEMENT DISTRICT: REALITY OR MYTH?............................................14 2.1 Introduction......................................................................................................14 2.2 Evidence of N Increase in the Suwannee River Water Management District..............................................................................................................16 2.2.1 Assessments at the District Level...........................................................16 2.2.2 Assessments at the Watershed Level......................................................23 2.2.3 Assessments at the National Level.........................................................24 2.2.4 Assessments at the Dairy Farm Level.....................................................25 2.3 Human and Ecological Concerns because of N Pollution...............................29 2.4 Conclusions and Recommendations................................................................31 vii

PAGE 8

3 NORTH FLORIDA STAKEHOLDER PERCEPTIONS TOWARD NUTRIENT POLLUTION, ENVIRONMENTAL REGULATIONS, AND THE USE OF CLIMATE FORECAST TECHNOLOGY.........................................34 3.1 Introduction......................................................................................................34 3.2 Materials and Methods.....................................................................................36 3.2.1 Study Area..............................................................................................36 3.2.2 The Sondeo.............................................................................................36 3.2.3 Interviews................................................................................................38 3.2.4 Focus Groups..........................................................................................38 3.2.5 Analysis...................................................................................................39 3.3 Results and Discussion....................................................................................39 3.3.1 Stakeholder Perceptions of the N Issue..................................................39 3.3.1.1 Perception of water contamination................................................39 3.3.1.2 Reasons to mitigate contamination................................................42 3.3.1.3 Existing and potential means of mitigation...................................42 3.3.1.4 Implementation procedures for Best Management Practices (BMPs)...........................................................................................45 3.3.1.5 Impacts of BMPs on farms............................................................47 3.3.2 Perceptions of Regulations by Dairy Farmers........................................48 3.3.2.1 Farmers in disagreement with regulatory measurements...............49 3.3.2.2 Farmers who agree with regulations but criticize part of it...........53 3.3.2.3 Farmers who completely agree with regulations...........................56 3.3.3 Perceptions of Climatic Technology Innovation....................................57 3.3.3.1 Farmers that would use long-term climatic forecast......................57 3.3.3.2 Farmers who would not use long-term climatic forecast...............62 3.3.3.3 Farmers who are not sure they would use long-term climatic forecast.............................................................................63 3.4 Conclusions and Recommendations................................................................63 4 A METHOD FOR BUILDING CLIMATE VARIABILITY INTO FARM MODELS: PARTICIPATORY MODELING OF NORTH FLORIDA DAIRY FARM SYSTEMS........................................................................................67 4.1 Introduction......................................................................................................67 4.2 Materials and Methods.....................................................................................71 4.2.1 Main Framework.....................................................................................72 4.2.2 The Sondeo.............................................................................................74 4.2.3 Interviews................................................................................................75 4.2.4 Focus Groups..........................................................................................76 4.2.5 Other Stakeholder Interactions...............................................................77 4.2.6 Analysis of Information..........................................................................77 4.3 Results and Discussion....................................................................................78 4.3.1 Phase One: Initial Developments............................................................78 4.3.2 Phase Two: The Sondeo..........................................................................81 4.3.3 Phase Three: Greater Stakeholder Involvement.....................................84 4.3.4 Phase Four: The Prototype Model..........................................................88 viii

PAGE 9

4.3.5 Phase Five: The Final Model..................................................................93 4.4 Summary, Conclusions, and Recommendations..............................................95 5 HOW MUCH MANURE NITROGEN DO NORTH FLORIDA DAIRY FARMS PRODUCE?.................................................................................................98 5.1 Introduction......................................................................................................98 5.2 Materials and Methods...................................................................................102 5.3 Model Development.......................................................................................103 5.3.1 General Characteristics.........................................................................103 5.3.2 Markov-Chains, Culling Rates, Reproduction Rates, and Milk Production Rates..................................................................................103 5.3.3 Manure N Produced by Milking Cows.................................................108 5.3.4 Nitrogen and Crude Protein..................................................................110 5.3.5 Milking Times and Cattle Breeds.........................................................111 5.3.6 Manure N Produced by Dry Cows, Young Stock, and Bulls................113 5.3.7 Manure, Feces, Urine, and Dry Matter Intake Estimations..................114 5.3.8 Water Utilization for Manure Handling................................................116 5.3.9 Waste Management Handling Systems................................................117 5.4 Results and Discussion..................................................................................119 5.4.1 The Model (Livestock Dynamic North-Florida Dairy Farm Model, LDNFDFM).............................................................................119 5.4.2 Model Comparison with Other Widely Used Estimations...................120 5.4.3 Estimation of Manure N Produced in North-Florida Dairy Farm Systems................................................................................................122 5.5 Conclusions....................................................................................................128 6 FORAGE SYSTEMS IN NORTH FLORIDA DAIRY FARMS FOR RECYCLING N UNDER SEASONAL VARIATION............................................132 6.1 Introduction....................................................................................................132 6.2 Materials and Methods...................................................................................135 6.2.1 Survey, Focus Groups, and Additional Information.............................135 6.2.2 Location of North-Florida Dairy Farms and Their Soil Series.............135 6.2.3 Climate Information and El Nio Southern Oscillation (ENSO) Phases...................................................................................................138 6.2.4 Forage Crop Systems............................................................................138 6.2.5 Manure N Application..........................................................................142 6.2.6. Crop Simulations.................................................................................142 6.2.7 Analyses................................................................................................143 6.3 Results and Discussion..................................................................................143 6.3.1 Forage Crop Systems in North-Florida Dairies....................................143 6.3.1.1 Spring-summer crops...................................................................144 6.3.1.2 Summer-fall crops........................................................................144 6.3.1.3 Fall-winter crops..........................................................................145 6.3.1.4 Sequences of forages....................................................................145 6.3.1.5 Harvest and processing................................................................146 ix

PAGE 10

6.3.2 Calibration and Validation of DSSAT Crop Models............................146 6.3.2.1 Forage sorghum...........................................................................147 6.3.2.2 Pearl millet...................................................................................149 6.3.2.3 Silage corn...................................................................................151 6.3.2.4 Winter forages..............................................................................152 6.3.3 Sensitivity Analysis of Forage Crops to Manure N Application..........154 6.3.4 Variability by ENSO Phases.................................................................156 6.3.4.1 Nitrogen leaching in different ENSO phases...............................158 6.3.4.2 Biomass in different ENSO phases..............................................166 6.3.5 Variability by Crop Sequences.............................................................168 6.3.5.1 Nitrogen leaching and crop systems............................................170 6.3.5.2 Biomass and crop systems...........................................................171 6.3.6 Variability by Soil Types......................................................................173 6.3.6.1 Nitrogen leaching and soil types..................................................176 6.3.6.2 Biomass and soil types.................................................................179 6.4 Conclusions....................................................................................................181 7 AN INTEGRATED SIMULATION MODEL TO ASSESS ECONOMIC AND ECOLOGIC IMPACTS OF NORTH FLORIDA DAIRY FARM SYSTEMS................................................................................................................184 7.1 Introduction....................................................................................................184 7.2 Materials and Methods...................................................................................185 7.2.1 Conceptual Model.................................................................................185 7.2.2 Computer Implementation....................................................................188 7.2.3 Analysis of Synthesized Dairies in North-Florida................................188 7.3 Model Description.........................................................................................189 7.3.1 Overall Characteristics..........................................................................189 7.3.2 Components..........................................................................................190 7.3.3 The Livestock Model............................................................................191 7.3.4 The Waste Model..................................................................................197 7.3.5 The Crop Models..................................................................................198 7.3.6 The Economic Module..........................................................................203 7.3.7 The Optimization Module.....................................................................204 7.3.8 The Feasible Adjustments Component.................................................206 7.4 User Friendly Implementation.......................................................................206 7.4.1 General Characteristics.........................................................................206 7.4.2 The Livestock Module..........................................................................210 7.4.3 The Waste System Module...................................................................211 7.4.4 The Soil Module...................................................................................212 7.4.5 The Forage Systems Module................................................................212 7.4.6 The Climatic Module............................................................................213 7.4.7 The Economic Module..........................................................................214 7.4.8 The Optimization Module.....................................................................217 7.5 Results for Synthesized Farms.......................................................................218 7.5.1 Synthesized Small Dairy Operation......................................................218 7.5.2 Synthesized Medium Dairy Operation..................................................219 x

PAGE 11

7.5.3 Synthesized Large Dairy Operation......................................................221 7.5.4 Comparison of the Three Synthesized Farms.......................................222 7.6 Conclusions....................................................................................................225 8 ECONOMIC AND ECOLOGIC IMPACTS OF NORTH FLORIDA DAIRY FARM SYSTEMS: TAILORING MITIGATION OF NITROGEN POLLUTION............................................................................................................227 8.1 Introduction....................................................................................................227 8.2 Materials and Methods...................................................................................228 8.2.1 Dynamic North-Florida Dairy Farm Model..........................................228 8.2.2 A Synthesized North-Florida Dairy Farm.............................................229 8.2.3 Statistical Analyses...............................................................................229 8.2.4 Graphical Analyses...............................................................................230 8.2.5 Optimization Analyses..........................................................................232 8.2.6 Validation with Real Farms..................................................................232 8.3 Results and Discussion..................................................................................233 8.3.1 Statistical Analyses...............................................................................233 8.3.1.1 Nitrogen leaching.........................................................................233 8.3.1.2 Profit............................................................................................234 8.3.2 Graphical Analyses...............................................................................235 8.3.2.1 Soil series.....................................................................................235 8.3.2.2 Crop systems in sprayfields.........................................................238 8.3.2.3 Pastures and crude protein in the diet (CP)..................................240 8.3.2.4 Confined time (CT) and crude protein in the diet (CP)...............242 8.3.2.5 Isolated effects of single managerial changes in the Synthesized farm..........................................................................244 8.3.3 Optimization Analysis: Decreasing Environmental Impacts and Increasing Profitability........................................................................246 8.3.3.1 The Synthesized dairy farm as is..............................................246 8.3.3.2 Optimization of the Synthesized farm.........................................247 8.3.3.3 Feasible practices for the Synthesized farm.................................249 8.3.4 Examples of Applications of the DNFDFM to Specific Farms............253 8.3.4.1 Farm One.....................................................................................253 8.3.4.2 Farm Two.....................................................................................254 8.3.4.3 Farm Three...................................................................................256 8.4 Conclusions and Recommendations..............................................................257 9 SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS...........................260 LIST OF REFERENCES.................................................................................................271 BIOGRAPHICAL SKETCH...........................................................................................283 xi

PAGE 12

LIST OF TABLES Table page 4-1. Stakeholder interactions, suggestions and changes made to model.......................91 5-1. Dry matter intake, manure, and N excretion by Florida dairy cattle based on milk production. (lbs cow -1 day -1 ) and low and high (NRC standards) crude protein (CP) content of diets.....................................................109 5-2. Water utilization by milking cows during months and percentage use relative to month of maximum use (July)............................................................117 5-3. Nitrogen lost in the waste management system: a baseline.................................119 5-4. Five hypothetical north-Florida dairy farms and their characteristics for manure N estimations..........................................................................................122 5-5. Estimated amounts of manure N to sprayfields in five hypothetical north-Florida dairy farms.....................................................................................125 5-6. Estimated amounts of manure N deposited on pasturelands in five hypothetical north-Florida dairy farms................................................................126 5-7. Estimated amounts of manure, DMI, and water use in five hypothetical north-Florida dairy farms.....................................................................................128 6-1. Soil types, some characteristics, and their sources of information used for the study...............................................................................................................137 6-2. Information sources for calibration and validation of forage crops in north-Florida dairy farm systems.........................................................................148 6-3. Coefficients values and coefficients definitions of modified crops in DSSAT.................................................................................................................150 6-4. Nitrogen leaching (kg ha -1 ) and ENSO phases, when all other factors are averaged...............................................................................................................160 6-5. Biomass (kg ha -1 ) and ENSO phases, when all other factors are averaged.........166 6-6. Nitrogen leaching (kg ha -1 ) and crop sequences, when all other factors are averaged...............................................................................................................171 xii

PAGE 13

6-7. Biomass (kg ha -1 ) and crop sequences, when all other factors are averaged.......172 6-8. Nitrogen leaching (kg ha -1 ) and soil types, when all other factors are averaged...............................................................................................................178 6-9. Biomass (kg ha -1 ) and soil types, when all other factors are averaged................180 7-1. State variables and monthly update of the livestock model.................................194 7-2. Dry matter intake estimations..............................................................................195 7-3. Manure excretion estimations..............................................................................196 7-4. Nitrogen excretion estimations............................................................................197 7-5. Nitrogen through waste system estimations........................................................198 7-6. Codification of ENSO phases, soil types, and forage systems from crop models..................................................................................................................200 7-7. Nitrogen applied estimated by the waste model and N applied in crop models..................................................................................................................202 7-8. Definition of variables in economic module (US$ cwt -1 milk -1 )..........................203 7-9. Monthly milk prices distribution in north-Florida as cumulative probability (US$ cwt -1 milk -1 )..............................................................................204 7-10. Crops in different fields.......................................................................................218 7-11. Comparison of three synthesized north-Florida farms and their estimated annual N leaching................................................................................................224 8-1. Characteristics of a Synthesized north-Florida dairy farm..................................230 8-2. Combination of management strategies simulated by the DNFDFM for the statistical analysis.................................................................................................231 8-3. Significance of factors and main interactions for N leaching..............................234 8-4. Significance of factors and main interactions for profit......................................235 8-5. Crops in different fields in the Synthesized farm................................................246 8-6. Management strategies selected by the optimizer...............................................248 8-7. Feasible crop systems in Synthesized dairy farm................................................250 xiii

PAGE 14

LIST OF FIGURES Figure page 1-1. North-Florida landscape representation...................................................................3 1-2. The Suwannee River Water Management District, the study area..........................7 2-1. Nitrate-N concentration in the SRWMD...............................................................17 2-2. Nitrate-N concentrations in three springs in the middle Suwannee River Basin......................................................................................................................19 2-3. Estimated annual mean of groundwater N load.....................................................20 2-4. Long term records of Nitrate-N for 4 Suwannee springs.......................................22 2-5. Long term records of Nitrate-N for 3 Lafayette springs........................................23 2-6. Nitrate-Nitrite and total ammonium plus organic N at different points in nine dairy farms in the middle Suwannee River Basin..........................................26 2-7. Dissolved Nitrate+Nitrite and Ammonium+Organic N in monitoring wells (3-m deep) in different areas of a dairy farm.........................................................27 2-8. Monitoring wells with higher than drinking water allowable nitrate levels (PDWS>10 mg L -1 ) in the four dairy farms...........................................................29 3-1. Farmers perceptions of environmental regulations in north-Florida....................50 3-2. Farmers perceptions of climate technology adoption...........................................58 4-1. Schematic of participatory modeling.....................................................................72 4-2. Methodology of participatory modeling................................................................73 4-3. Phases of the participatory modeling for creating the DNFDFM..........................78 5-1. Pregnancy rates of adult cows..............................................................................105 5-2. Culling rates of adult cows..................................................................................106 5-3. Milk production rates of milking cows................................................................107 xiv

PAGE 15

5-4. Manure N produced by low-CP diet cows according to milk production...........111 5-5. Manure excretion based on milk production.......................................................114 5-6. Dry matter intake (DMI) based on milk production............................................115 5-7. Nitrogen excretion by a low protein diet in a 100-cow dairy estimated by different models..............................................................................................120 5-8. Nitrogen excretion by a high protein diet in a 100-cow dairy estimated by different models..............................................................................................121 6-1. Study area, dairy farms, and soil types................................................................136 6-2. Daily rainfall (1956-1998) in north-Florida for different ENSO phases.............139 6-3. Distribution of monthly precipitation (1956-1998) in north-Florida for different ENSO phases.........................................................................................140 6-4. Daily precipitation in north-Florida (1956-1998) for different ENSO phases...................................................................................................................140 6-5. Daily temperature (1956-1998) in north-Florida for different ENSO phases......141 6-6. Daily solar radiation (1956-1998) in north-Florida for different ENSO phases...................................................................................................................141 6-7. Forage systems and their seasonality in north-Florida dairy farms.....................144 6-8. Forage sorghum: Observed and simulated biomass production, Bell, 1996-1998. Individual mean square error (MES) and overall Root Mean Square Error (RMSE)..........................................................................................149 6-9. Forage millet: Observed and simulated biomass production, Bell, 1996-1998. Individual mean square error (MES) and overall Root Mean Square Error (RMSE)..........................................................................................151 6-10. Forage millet: Observed and simulated biomass production, Quincy 1992........152 6-11. Silage corn: Observed and simulated biomass production, Bell, 1996-1998. Individual mean square error (MES) and overall Root Mean Square Error (RMSE)..........................................................................................153 6-12. Winter forages: Observed and simulated biomass production, Bell 1997-1998. Individual mean square error (MES) and overall Root Mean Square Error (RMSE)..........................................................................................154 xv

PAGE 16

6-13. Sensitivity response of N leaching and biomass accumulation to different rates of manure N applied for simulations in a Kershaw soil series during the season 1996/1997...........................................................................................155 6-14. Nitrogen leaching vs. biomass accumulation for different ENSO phases under two forage systems consisting of bermudagrass and millet-sorghum, four manure N applications (10, 20, 40, and 80 kg ha -1 mo -1 ), and two type of soils (Arredondo-Gainesville-Millhopper and Millhopper-Bonneau).............157 6-15. Nitrogen leaching vs. ENSO phases relative to absolute averages, when all other factors are averaged....................................................................................162 6-16. Probability distribution of N leaching by ENSO phases, when all other factors are averaged.............................................................................................163 6-17. Frequency distribution of N leaching and month of the year by ENSO phase, when all other factors are averaged..........................................................165 6-18. Biomass accumulation vs. ENSO phases relative to absolute averages, when all other factors are averaged.....................................................................167 6-19. Nitrogen leaching vs. biomass accumulation for different forage systems under different ENSO phases (La Nia, Neutral, and El Nio), two manure N applications (40 and 80 kg ha -1 mo -1 ), and soils of type Penney-Kershaw......169 6-20. Nitrogen leaching vs. crop sequences relative to absolute averages, when all other factors are averaged...............................................................................170 6-21. Biomass accumulation vs. crop sequences relative to absolute averages, when all other factors are averaged.....................................................................173 6-22. Nitrogen leaching vs. biomass accumulation for different soil types under different ENSO phases (La Nia, Neutral, and El Nio), two manure N applications (40 and 20 kg ha -1 mo -1 ), and forage system consisting of corn-sorghum-winter forage................................................................................175 6-23. Nitrogen leaching vs. soil types relative to absolute averages, when all other factors are averaged....................................................................................177 6-24. Nitrogen leaching vs. soil types relative to absolute averages, when all other factors are averaged....................................................................................181 7-1. Representation of the dynamic north-Florida dairy farm model.........................191 7-2. Representation of the livestock model.................................................................192 7-3. Data management of simulation of crop models.................................................201 xvi

PAGE 17

7-4. The process of optimization.................................................................................205 7-5. Main screen of the computer implementation of The Dynamic North-Florida Dairy Farm Model (DNFDFM)....................................................208 7-6. The start message menu control...........................................................................209 7-7. The livestock module control...............................................................................211 7-8. The waste management control...........................................................................212 7-9. Dairy farm location and soil type by farm component control............................213 7-10. Forage systems component control......................................................................214 7-11. The climatic component control..........................................................................215 7-12. The economic component control........................................................................216 7-13. Distribution of milk prices in north-Florida.........................................................216 7-14. The optimization control......................................................................................217 7-15. Monthly estimations for Synthesized Small north-Florida farm.........................220 7-16. Small north-Florida farm, yearly estimated N leaching and profit......................221 7-17. Medium north-Florida farm, yearly estimated N leaching and profit..................222 7-18. Large north-Florida farm, yearly estimated N leaching and profit......................223 7-19. Overall N leaching from the synthesized farms...................................................225 8-1. Nitrogen leaching and profit for different soil types in north-Florida: Estimations for different ENSO phases for common crop systems.....................236 8-2. Nitrogen leaching and profit for different crop systems in north-Florida: Estimations for different ENSO phases for common pastures and soil types.....................................................................................................................239 8-3. N leaching and profit for different pasture crops and crude protein in north-Florida: Estimations for different ENSO phases for common sprayfield crops and soil type..............................................................................241 8-4. N leaching and profit for different confined time and crude protein: Estimations for different ENSO phases for common crop systems and soils......243 8-5. Impacts of management changes for the Synthesized farm with soils type 4 (Penney-Otela)..................................................................................................245 xvii

PAGE 18

8-6. Nitrogen leaching and profit with current practices for a Synthesized north-Florida farm................................................................................................247 8-7. Optimized vs. current practice yearly N leaching and profit for a north-Florida Synthesized farm...........................................................................249 8-8. Feasible yearly N leaching and profit for a Synthesized north-Florida farm......................................................................................................................251 8-9. Optimization of a Synthesized north-Florida farm, scenario One.......................251 8-10. Second feasible yearly N leaching and profit for a north-Florida Synthesized farm..................................................................................................252 8-11. Farm One: N leaching and profit for different management strategies...............254 8-12. Farm Two: N leaching and profit for different management strategies...............255 8-13. Farm Three: N leaching and profit for different management strategies.............257 xviii

PAGE 19

LIST OF ABBREVIATIONS AFO Animal Feeding Operations ASAE American Society of Agricultural Engineers AWMFH Animal Waste Management Field Handbook BMPs Best Management Practices CAFO Concentrated Animal Feeding Operation CARES County Alliance for Responsible Environmental Stewardship CP Crude Protein DBAP Dairy Business Analysis Project DHIA Dairy Herd Improvement Association DNFDFM Dynamic North-Florida Dairy Farm Model DM Dry Matter DMI Dry Matter Intake DRMS Dairy Records Management Systems ENSO El Nio Southern Oscillation EPA Environmental Protection Agency FDH Florida Department of Health FFB Florida Farm Bureau FGS Florida Geological Service FLDEP Florida Department of Environmental Protection FSR Farming Systems Research FSRE Farming Systems Research and Extension IC Initial Conditions for Simulation LDNFDMF Livestock Dynamic North-Florida Dairy Farm Model xix

PAGE 20

LN Lactation Number MCL Maximum Concentration Load MHB Methemoglobinemia MIM Months in Milk MIP Months in Pregnancy NAC Number of Adult Cows NAWQA National Water Quality Assessment 4NH -Org Ammonium + Organic N 3NO 23NO Nitrites + Nitrates NOAA National Oceanic and Atmospheric Administration NRC National Research Council NRCS Natural Resource and Conservation Service NRM Natural Resource Management PDWS Primary Drinking Water Standard PRA Participatory Rural Appraisal RHA Rolling Herd Average RRA Rapid Rural Appraisal SRP Suwannee River Partnership SRWMD Suwannee River Water Management District SWIM Surface Water Improvement and Management TWG Technical Working Group USDA US Department of Agriculture USEPA, EPA US Environmental Protection Agency USGS US Geological Service WARN Water Assessment Regional Network WATNUTFL Water budget and Nutrient balance for Florida xx

PAGE 21

Abstract of Dissertation Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy MODELING NORTH FLORIDA DAIRY FARM MANAGEMENT STRATEGIES TO ALLEVIATE ECOLOGICAL IMPACTS UNDER VARYING CLIMATIC CONDITIONS: AN INTERDISCIPLINARY APPROACH By Victor E. Cabrera August 2004 Chair: Peter E. Hildebrand Major Department: Natural Resources and Environment High levels of nitrogen (N) in water affect human health and ecosystem welfare. Dairy waste is considered an important contributor to increased water N levels in the Suwannee River Basin. The main aim of the study was to assess management options that decrease N leaching in dairy systems under different climatic conditions, while maintaining profitability. Whole-dairy-farm system simulations were used to study economic and ecologic sustainability, which would be impractical to test with regular experiments. Dynamic, event-controlled, and empirical models were coupled with process-based and decision-making models to represent and test management strategies. User-friendliness and stakeholder interaction were always pursued. Eight focus groups and 24 dairy farmer interviews were conducted, and a collaborative user-friendly decision support system was created. The Dynamic North-Florida Dairy Farm Model (DNFDFM) uses Markov-chains to simulate probabilistic livestock reproduction, culling, milk production, and N excretion; process-based crop models to simulate forage production, N xxi

PAGE 22

uptake, and N leaching; and linear programming optimization to tailor management strategies. The DNFDFM is an environmental accountability tool for dairy farmers and regulatory agencies capable of simulating any individual north-Florida dairy farm. Results indicated that north-Florida dairy farmers could maintain profitability and decrease N leaching 9 to 25% by decreasing crude protein in the diet, adjusting confined time of milking cows, and adjusting sequences of seasonal forages. During El Nio years, when higher rainfall and colder temperatures are experienced, 5% more N leaching is estimated compared with Neutral years; and 13% more compared with dryer and hotter La Nia years. xxii

PAGE 23

CHAPTER 1 INTRODUCTION 1.1 Background The presence of high levels of N in water is an environmental hazard because it affects human health and ecosystem welfare. The Suwannee River Basin has received much attention in recent years because of increased N levels in water bodies. Dairy waste is thought to be an important factor contributing to this water N pollution. Dairy farmers are now required to comply with stricter environmental regulations either under permit or under voluntary incentive-based programs. Dairy farmers are also aware that environmental issues will be among the greatest challenges they face in the near future. Evidence indicates that farms may reduce their total N loads by changing some management strategies. Improvements in long-term seasonal predictions (6 to 12 months) such as in El Nio and La Nia forecasts, can play an important role in implementing management strategies that dairy farmers in north-Florida may adopt in order to pursue economic and ecological sustainability. 1.2 The Problematic Situation Dairy farming is an important part of Floridas agricultural industry. The Florida Statistical Service showed that milk and cattle sales from dairies contributed $429 million directly into the Floridian economy in the year 2001. Florida is the leading dairy state in the Southeast; it ranks 13th nationally in cash receipts for milk, 15th in milk production and 15th in number of cows (Bos taurus) (Florida Agricultural Statistics Service, 2003). 1

PAGE 24

2 According to USDA, there were about 152,000 cows on about 220 dairy farms at the end of 2002, and more than 30% of the dairy operations (65 operations) and number of cows (> 45,000 cows) of the State of Florida are located in the Suwannee River Basin. Indeed, more than 25% of the dairy cows in the state are in four counties in north central Florida that bound the Suwannee River (Bisoondat et al., 2002): Gilchrist, Lafayette, Suwannee, and Levy. This dairy industry, according to Giesy et al. (2002), faces three significant challenges: production, economic, and environmental. Dairymen perceive that environmental issues will be the most stressful in the near future (Staples et al., 1997). Dairies face increased regulation because social pressure, while larger herds attract the attention of neighbors and activists concerned with odors, flies, and especially potential leaching of nutrients that might influence water quality (Giesy et al., 2002). The presence of N in surface water bodies and groundwater aquifers is recognized as a significant water quality problem in many parts of the world (Fraisse et al., 1996). The Suwannee River Partnership states that over the last 15 years, nitrate levels in the middle Suwannee River Basin have been on the increase; and these elevated nitrate levels can cause health problems in humans as well as negative impacts on water quality (Suwannee River Partnership, 2004a). Indeed, the Suwannee River Basin has received much attention in recent years because increased N levels in the groundwater-fed rivers of the basin could seriously affect the welfare of the ecosystem (Albert, 2002). According to Katz (2000), N levels have increased from 0.1 to 5 mg L -1 in many springs in the Suwannee basin over the past 40 years. Pittman et al. (1997) found that nitrate concentrations in the Suwannee River itself have increased at the rate of 0.02 mg L -1

PAGE 25

3 year -1 over the past 20 years. In a 53.1 km river stretch between Dowling Park and Branford, the nitrate loads increased from 2,300 to 6,000 kg day -1 They also estimated that 89% of this came from the lower two-thirds, where agriculture is the dominant land use. Soils in this region are generally deep, well-drained sands (Figure 1-1); and nutrient management is a major concern (Van Horn et al., 1998). Figure 1-1. North-Florida landscape representation. Source: Katz et al., 1999, p. 1. As Conway (1990) stated, farms in industrialized countries have become larger in size, fewer in number, highly mechanized, and highly dependent on external inputs; and can cause environmental problems by producing a number of wastes that might be hazardous. Recent evidence that agriculture in general, and animal waste in particular, may be an important factor in surface and groundwater quality degradation has created a strong interest in nutrient management research (Katz, 2000, Van Horn, 1997, Fraisse et

PAGE 26

4 al., 1996). Over-applying manure nutrients to land is considered a major cause of nitrates, converted from manure ammonia sources while in the soil, leaching to groundwater and contributing to surface runoff of N that contaminates surface waters. Concerns regarding nutrient losses from manure of large dairy herds to groundwater or surface runoff have been extremely acute in Florida (Van Horn, 1997). One of the most publicized concerns is N losses in the form of nitrate into the groundwater through the deep sandy soils of the Suwannee River Basin (Van Horn et al., 1998). Many soils in the Suwannee River Basin are excessively drained, deep, sandy, Entisols with low exchange capacity and low organic matter, which are prone to nitrate leaching (Woodard et al., 2002). Environmental regulation of livestock waste disposal has become a major public concern, and much of the focus for this issue in Florida has been on dairies. Dairymen are now required to develop manure disposal systems in order to comply with Florida Department of Environmental Protection water quality standards (Twatchtmann, 1990). This fact has led to considerable research emphasizing N recycling and addressing issues such as maximum carrying capacity and nutrient uptake by crops (Fraisse et al., 1996). Dairymen in the Suwannee River Basin have expressed their willingness to participate in initiatives that promote reduced environmental impacts. In fact, many of them are already involved in programs such as Better Management Practices (BMPs) to reduce nitrate overflow promoted by the Suwannee River Partnership (Smith, 2002, Pers. Comm.). Staples et al. (1997) interviewed 48 dairy farmers in Lafayette, Suwannee, Columbia, Hamilton, and Gilchrist counties. By far farmers rated their perception of anticipated costs of complying with upcoming environmental regulations as the top challenge to successful dairying in the future.

PAGE 27

5 Overflow of nitrates in these dairy systems is affected by adopted management practices and by environmental conditions. Changing management practices might have a great impact on overflow amounts. For instance, Albert (2002) found that by reducing the duration of the irrigation, reducing the fertilizer application rate, and improving the timing of fertilizer applications, N leaching can be reduced by approximately 50% while maintaining acceptable potato yields in sandy soils in north-Florida. It is unknown if similar large reductions are also possible in dairy farms. Many environmental conditions are also determined by climate factors such as rainfall, temperature, and radiation (which are variable and uncertain). Dairy operations are affected by changes in climatic components at many different levels, and in many different ways. Rainfall patterns might result in very different amounts of nutrient leaching. For example, higher temperature in winter can promote more biomass accumulation, and increase plant N uptake. Currently, better-quality long-term seasonal forecasts (6 to 12 months in advance) are available. The National Oceanic and Atmospheric Administration (NOAA) can predictusing the El Nio Southern Oscillation (ENSO) technologyaccurate long-term seasonal climate patterns. These are based on predicting El Nio, La Nia, or Neutral years (which determine rainfall and temperature patterns). The ENSO signal in Florida has been well documented, in terms of its effects on climatic variables (Jones et al., 2000). Precipitation in Florida is particularly variable, with an excess of over 30% of the normal seasonal total across much of the state during an El Nio winter. La Nia has the opposite effect, with deficits of 10 to 30% lasting from fall through winter and spring. Florida and the Gulf Coast can expect to see average temperatures 2 to 3F below normal during El Nio years. La Nia has the opposite effect, with temperatures 2 to 4F above

PAGE 28

6 normal during winter months. La Nia's effect on temperature is more pronounced in north-Florida, Alabama, and Mississippi. Solar radiation decreases in association with higher rainfall during El Nio years. Notice also that climatic conditions in north-Florida allow year-round crop production, thus allowing more effluent-waste application, compared with dairy production systems in temperate areas (Bisoondat et al., 2002). Proactive use of seasonal prediction can be adopted by dairy farmers. The opportunity to benefit from climate prediction arises from the intersection of human vulnerability, climate prediction, and decision capacity (Hansen, 2002a). This benefit from the information can be used to decrease environmental impacts, while maintaining profitability in north-Florida dairy operations. 1.3 Research Area The study was performed inside the Suwannee River Water Management District, which encompasses all or part of 15 counties in north central Florida (Figure 1-2). The study was centered in five counties where N pollution is more problematic: Alachua, Levy, Gilchrist, Lafayette, and Suwannee. 1.4 Objectives The main aim of the study was to assess and test management options that decrease nutrient overflow under different climatic conditions in north-Florida dairy operations while maintaining profitability. To achieve this, it was necessary to Thoroughly understand the north-Florida dairy system through farmer interactions Conceptualize north-Florida dairy farms as systems Identify dairy farm system components Recognize interaction among these dairy farm system components Create a comprehensive dairy farm model adaptable to north-Florida dairies

PAGE 29

7 Simulate north-Florida dairy farms under different environmental conditions, different management strategies, and different climate conditions Identify climate-management interactions that result in lower nutrient pollution, especially less N leaching, while maintaining acceptable profit levels Suggest management strategies that individual north-Florida dairies could follow based on climate forecast Based on previous strategies, produce an extension circular with a bullet list of results to be available on paper and on the Internet for north-Florida dairy farmers and other stakeholders Create a user-friendly computer application with the simulation models that north-Florida dairy producers and other stakeholders can use for environmental accountability Figure 1-2. The Suwannee River Water Management District, the study area. Source: Land Use Survey, Suwannee River Water Management District, Florida Geographic Data Library, 1995. Note: Dots represent dairy farms and numbers indicate the number of dairy farms by county.

PAGE 30

8 1.5 Hypothesis North-Florida dairy farms can decrease their environmental impacts and still maintain profitability by changing management strategies, which include the use of improved long-term seasonal forecasts. This hypothesis can also be expressed as a research question: Can north-Florida dairy farms maintain profitability and decrease their environmental impacts by combining management strategies and seasonal forecasts? 1.6 Framework Most dairies in Florida manage animals under semi-intensive or intensive systems. Florida dairies are principally in the business of producing milk: decisions are based on profit maximization and operation regulations. North-Florida dairies are business enterprises and have full access to credit opportunities, information, and new technologies (Adams, 1998). Additionally, dairy farms (which produce both livestock and crops) are composed of interconnected units; any part of which is affected by and affects the behavior of other parts, including human management decisions and strategies. Since a whole-dairy-farm encompasses environmental, economic, and bio-physical components, its analysis is best accomplished by a systems approach (that accounts for the interactions of these components, and that can trace the consequences of an intervention through the entire system) (Kelly, 1995). While systems can be conceptualized at any of many scale levels, a systems approach defines a specific scale of interest and an appropriate boundary of analysis. Analysis in the systems approach is marked by recognition of the whole system and the interactions within that system rather than looking only at a system component. A systems approach uses specific techniques and tools (such as rapid appraisal, pattern analysis, diagrams, and modeling, often in a

PAGE 31

9 multidisciplinary fashion) to identify system boundaries and recognize component interactions (Kelly, 1995). The proposed systems approach was Farming Systems Research and Extension (FSRE) as described by Hildebrand (1990). The farm household in livelihood systems analyses are replaced by farm management (Hazell and Norton, 1986) for commercial or industrial farm dimensions such as north-Florida dairy farms. The farming system is an arrangement of subsystem components that function as units. These subsystems include economic subsystems and a number of agroecosystems. The agroecosystems, as described by Powers and McSorley (2000), include the populations and communities organized in ecosystems in relationship with the abiotic factors. To assess what would happen to the whole dairy operation and what would happen specifically to nitrate flow under different climate-management scenarios, models of these realities were set up (Thornley and Johnson, 1990). Models allow us to foresee results, when temporal or material conditions make it impractical to develop field experiments. Successful planning of an animal waste management system requires the ability to simulate the impact of waste production, storage, treatment, and utilization on the water resources; it must address overall nutrient management for the operation (Fraisse et al., 1996). Different simulation methods were needed to accurately represent the north-Florida dairy crop and livestock systems; and to test management strategies that yield the most desired outcomes. Whole farm scale simulation models include biophysical, environmental, and economic components and relationships of the dairy operation to climate variability. Complex systems demand the study of interactions among processes

PAGE 32

10 acting on multiple scales. Thus, these problems require innovative approaches, combining advances in dynamical theory, stochastic processes, and statistics (Levin, 1991). Budgeting nutrient flow through the total dairy farm system as conceptualized by Van Horn et al. (2001, 1998, 1994, 1991); NRCS (2001), Van Horn (1997); and Lanyon (1994) were adjusted, modified, and used for keeping track of dynamic N movement in all system components. Van Horn et al. (1998) mentioned the quantitative information needed on nutrient flow through all segments of the system. They listed the following critical research questions: How much of individual nutrients are excreted by dairy cows? How does the manure management system affect nutrient flow and nutrient recoveries for fertilizer use? What is the potential nutrient removal by plants? How can a manure nutrient budget be developed? What steps help to document environmental accountability? Agronomic measures of nutrient balance, and tracking of inputs and outputs for various farm management units, can provide the quantitative basis for management to better allocate manure to fields, to modify dairy rations, or to develop alternatives to on-farm manure application (Lanyon, 1994). Similarly to budgeting nutrients, The Netherlands has implemented the Mineral Accounting System (MINAS) which focuses on nutrient (N and phosphorus) flows on individual farms; and taxes farms whose nutrient surplus exceeds a defined limit. MINAS embodies a new approach to environmental problems caused by agriculture. According to Ondersteijn et al. (2002) focusing on individual farmers has two major advantages. First, individuals are considered polluters and are individually held accountable for their

PAGE 33

11 pollution, according to the polluter pays principle. Second, individuals have control over their pollution problem, and will be able to deal with it on an individual level (instead of being forced to comply with general measures that may be ineffective for their specific situation). After interviewing 240 farmers during 3 years, they also found large variation in N and phosphate surpluses within and between land-based farm types in The Netherlands. They surveyed farmers stated that MINAS provided enormous and often surprising insight into their management. Crop models (Jones et al., 2003, Jones et al. 1998) are process-based dynamic simulation models that allow biophysical and environmental conditions to be translated into agricultural outcomes (Philips et al., 1998). Crop models can be used to estimate monthly biomass accumulation and N leaching (by a number of crop and/or pasture sequences under different climatic conditions, depending on soil conditions, manure applications, and other managerial choices). Linear programming (LP) models are optimization models capable of devising alternative management choices that maximize (minimize) an objective function according to a set of restrictions (Hazell and Norton, 1986). The LP models have been widely used for analyzing farming systems. 1.7 Methods North-Florida dairy farmers were involved in the entire research process. They provided the base information, and they validated results of the analyses. Few studies to date have incorporated such first-hand data from interviews and validation interaction with dairy managers (which are believed extremely important). Other stakeholder agencies also participated actively as was the case for the Suwannee River Partnership (that involves 45 agencies in total); the Florida Cooperative Extension Service; and some

PAGE 34

12 private enterprises. These stakeholders provided relevant information and permanent feedback to the study. In all interactions, a laptop computer was used. There were two rounds of interviews and focus groups. For the first round, a prototype model was used (created by suggestion of one stakeholder) as a discussion, engagement, and data-collection tool. In both the interviews and focus groups, the interaction started by explaining the prototype model and demonstrating its functioning, followed by an interactive discussion of all specific points. The researcher recorded all suggestions and information. Some modifications were made during the process to the prototype model, however most of the suggestions were incorporated into a new model, which was then used in the second round. In the interviews of dairy farmers, an additional 1 to 2 hour conversation was used to collect information regarding biophysical farm factors, livestock, livestock feeding program, crops, crop fertilization program, waste management, perception of environmental regulations, perception of seasonal forecast use, and other factors. In several occasions after the interview, a tour of the farm followed. There were 21 interviews and six focus groups during this first round. A second round of focus groups and interviews was arranged to validate the functioning of the model. Two focus groups and three interviews confirmed that the model was performing inside the expected ranges, and that results were realistic. Local published data from related studies were also used as baseline in the modeling. An integrated, whole-dairy-farm simulation that links decision making with biophysical, environmental, and economic processes was developed, the Dynamic North-Florida Dairy Farm Model (DNFDFM). The simulation approach included a Dynamic event-controlled simulation model (D, driver) connected to crop models (CM)

PAGE 35

13 and to optimization models (OM). The driver, which had many modules, followed a dynamic adaptation of the budgeting framework for tracking the flow of N through the entire system; and for tracking production and additional economical variables, according to a pre-defined set of management and climatic conditions. Crop models assessed the N recycled by plants and the amount of N leached according to management, genetic, soil, and climatic specific conditions. The optimization models consisted of linear programming models that either minimize N leaching, or maximize profit for specific farm conditions. The DNFDFM runs monthly for all the processes following user-friendly menus. 1.8 Relevance Results are expected to directly help Suwannee River Basin dairy operations by providing a decision tool that can be used to decrease environmental impacts by reducing nitrate pollution (currently one of the most important environmental concerns). Results should also contribute to overall Suwannee River Basin environmental health by reducing the total N load in the watershed, which has been rapidly increasing. Innovative frameworks and methodologies used in this research could serve as baselines for other studies in which interdisciplinary and participative outcomes are desired. For example a similar approach could be used to address the phosphorus problem in the Okeechobee area.

PAGE 36

CHAPTER 2 NITROGEN POLLUTION IN THE SUWANNEE RIVER WATER MANAGEMENT DISTRICT: REALITY OR MYTH? 2.1 Introduction In response to increased environmental concerns, the U.S. Environmental Protection Agency (USEPA) has increased regulatory controls for concentrated animal feeding operations (CAFO) under the Clean Water Act. One of the changes is to require CAFOs that apply manure to land to meet application limits defined by a nutrient standard (USEPA, 2003). The goal of the nutrient standard is to minimize nutrient loss from fields receiving manure. Farms meet this standard by applying manure nutrients at a rate consistent with the agronomic needs of crops receiving manure. Nutrient standards can be N -or P-based, depending on the nutrient content of the soil. USDA is also encouraging the voluntary adoption of nutrient standards on all animal feeding operations (AFO) not subject to USEPA regulation (USDA, NRCS, 1999). The Suwannee River Partnership 1 (SRP) states that over the last two decades, nitrate levels in the middle Suwannee and Santa Fe river basins have been on the increase. The SRP is a coalition of state, federal and regional agencies, local governments, and private industry representatives with 46 partners working together to reduce nitrate levels in the surfacewaters and groundwater within the basins, or watersheds. The Suwannee River Basin has received much attention in recent years because increased N levels in the groundwater-fed rivers of the basin could seriously 1 www.mysuwanneeriver.com/features/suwannee+river+partnership 14

PAGE 37

15 affect the welfare of the ecosystem (Albert, 2002). The federal safety standard for nitrate-N in drinking water is 10 mg L -1 (Florida Department of Health, FDH, and Florida Department of Environmental Protection, FLDEP). According to Hornsby (pers. comm.), a water quality analyst from the Suwannee River Water Management District (SRWMD) at Live Oak, nitrate levels in some areas have reached ranges of 20 to 30 mg L -1 Rising nitrate levels in north-Florida have been a growing concern because of the effect of nutrients on public health and the environment (Van Horn et al., 1998). Hornsby et al., 2002b alleged that the most likely contributors to contamination are farms, industry and human settlements in the region. Soils in this region are generally deep, well-drained sands; and one of the most critical concerns is N loss in the form of nitrate into the groundwater through these deep sandy soils of the Suwannee River Basin (Van Horn et al., 1998). Changes in agricultural land use and the intensification of this land use during the past 50 years have contributed variable amounts of N to the groundwater system. In general, four main sources of N to groundwater can be recognized: fertilizers applied to cropland, animal wastes from dairy and poultry operations, atmospheric deposition, and septic tank effluents (FLDEP, 2001). Despite the above arguments, it was found (by talking with local people) that some stakeholder groups believe that either the problem of N pollution is not as bad as many people perceive, or the problem does not exist at all. Therefore there is a feeling that the regulatory agencies are overacting and pushing excessively in protecting the environment.

PAGE 38

16 Under these circumstances, it is critical to clearly understand the problem and the concerns. Clarifying the real problem and the real actors would help in promoting realistic ways of solving it. The objective of this chapter is to discuss the evidence of increasing levels of N in water bodies in the Suwannee River Basin, explore evidence of N pollution by dairy farms, and briefly address potential ecological and human implications. 2.2 Evidence of N Increase in the Suwannee River Water Management District 2.2.1 Assessments at the District Level The Suwannee River Water Management District (SRWMD) is the water authority in the study area. The SRWMD deals with quantity and quality of water. Regarding the quality of the water, it permanently monitors the quality of the ground and surface water, and every year produces a comprehensive report. The last available report is for the year 2001/2002 (Hornsby et al., 2003). The SRWMD has developed a comprehensive Water Assessment Regional Network (WARN) consisting of 97 Floridian aquifer wells for a trend network and 120 wells for a status network, distributed along the water management district. These are used to assess the groundwater quality in the District. In this last report (Hornsby et al., 2003) containing results only for the status wells between October 2001 and September 2002, nitrate-N concentrations above a background concentration of 0.05 mgL -1 existed for all or part of every county in the District. The possible nitrate sources were stated as fertilizers, animal waste, and human waste. Concentrations are elevated in the Middle Suwannee River Watershed (Suwannee and Lafayette Counties); and within an area generally defined along an axis starting in northern Jefferson County and ending in eastern Alachua County (Figure 2-1). However, the three samples that exceeded the

PAGE 39

17 Maximum Concentration Load (MCL) or the Primary Drinking Water Standard (PDWS) set at 10 mg L -1 were located in Levy County during May and June 2001 (10.6, 12.0, and 14.2 mg L -1 ). By comparing overall data with the other four water districts in Florida, the SRWMD has the largest area of elevated groundwater nitrate-N concentrations in the State. Figure 2-1. Nitrate-N concentration in the SRWMD. Source: Hornsby et al. (2003), p. 22. SuwanneeLafayette Gilchrist Alachua Levy Regarding surface water quality, the SRWMD has implemented the Surface Water Improvement and Management (SWIM) program, a quality-monitoring network consisting of 67 water chemistry stations. In water year 2002 (October to September), 3,012 tons of nitrate-N were transported to the Gulf of Mexico by the Aucilla, Econfina, Fenholloway, Steinthatchee, Suwannee, and Wacasassa rivers; the Suwannee river alone

PAGE 40

18 accounted for 98.6% of it. The middle Suwannee River Basin that accounts for only 8.6% of the Suwannee River area accounted for 29.3% of the total loading; and the middle Santa Fe River that represents only 5.7% of the Suwannee River Basin accounted for 19.6% of the total loading; those are the N hotspots. In 2002, of the 35 springs whose water-quality standards were studied, 5 were classified as poor quality and 15 as fair quality. Of the 26 rivers (or river sections) identified in the District, 1 was classified as poor and 13 as fair quality. Using the SRWMD database of monitoring point water quality, graphs for 3 springs in the middle Suwannee River Basin were constructed. Figure 2-2 shows N levels in 3 central springs (Blue, Ruth/Little, and Troy) with data for the monitoring network of the SRWMD. There clearly is an increasing trend. In Figure 2-2C additional data from the US Geological Service (USGS) and from the Florida Geological Service (FGS) for previous years are presented. Figure 2-2C clearly shows an increase in the N levels in this surface water body; the line indicates the increasing trend. In previous water year, 2001, (Hornsby et al., 2002a) the total loading of nitrate-N going to the Gulf of Mexico was 3067 tons; and 97.8% was accounted for by the Suwannee River Basin. These numbers suggest a loading decrease from 2001 to 2002. However, in absolute numbers, loading was higher in 2002, since the discharge of the Suwannee River increased. The fact of increased loadings year to year is reported to be true also for many previous years. In 1999, the FLDEP (Copeland et al., 1999) after sampling 491 wells in the middle Suwannee River, found that wells located 300 ft from poultry or dairy farms had nitrate concentrations greater than the background, with a 25% chance of exceeding the MCL.

PAGE 41

19 00.511.522.533.5Apr-95Feb-96Dec-96Sep-97Jul-98May-99Mar-00Jan-01Nov-01Sep-02Jun-03Apr-04nitrate-nitrogen (mg/L) 02468101214Apr-95Feb-96Dec-96Sep-97Jul-98May-99Mar-00Jan-01Nov-01Sep-02Jun-03Apr-04nitrate-nitrogen (mg/L) 00.511.522.533.5Mar-60Aug-61Dec-62May-64Sep-65Jan-67Jun-68Oct-69Mar-71Jul-72Dec-73Apr-75Aug-76Jan-78May-79Oct-80Feb-82Jul-83Nov-84Mar-86Aug-87Dec-88May-90Sep-91Jan-93Jun-94Oct-95Mar-97Jul-98Dec-99Apr-01Sep-02Jan-04nitrate-nitrogen (mg/L) USGS FGS SRWMD Figure 2-2. Nitrate-N concentrations in three springs in the middle Suwannee River Basin. A) Blue, B) Ruth/Little, C) Troy. Source: Prepared with data from the SRWMD, FGS, and USGS. (A)(B)(C)

PAGE 42

20 The 2001 FLDEP report for the Suwannee River Basin states that 7 lakes (Cherry, Alligator, Crosby, Rowell, Sampson, Santa Fe, and Hampton), 1 stream segment in Swift Creek; several segments of the Santa Fe River; the Middle and Lower Suwannee Rivers, and their estuary are potentially impaired because of excess nutrient concentrations. This FLDEP water quality report is summarized in Figure 2-3. Figure 2-3. Estimated annual mean of groundwater N load. Source: FLDEP (2001), p. 81. A study by the FGS (1992) reports that the proportion of samples of water exceeding the PDWS for Florida of 10 mgL -1 is low (0.6% for superficial system, 0% for intermediate aquifer system, and 1% for the Floridian aquifer system); however, it

PAGE 43

21 mentioned that for the Suwannee River Basin high nitrate concentrations (up to 140 mg L -1 ) in monitoring wells associated with dairy operations have been found. A cooperative study between the DEP and the USGS measured monthly water samples of 14 wells in the middle Suwannee River Basin (Katz and Bohlke, 2000). This study found concentrations of nitrate-N ranging from 0.02 to 22 mgL -1 with large seasonal variations. The fluctuations were believed to be correlated mostly with agricultural practices and climate conditions; for example, higher concentrations occurred in winter when waste or fertilizer was applied. Additionally, the use of N isotope indicated that the most likely cause of nitrate in 5 wells were dairy and/or poultry operations (Katz and Bohlke, 2000). This coincided with the fact that these farm operations were close to the sampling sites. A long term and comprehensive study in the Suwannee River Basin developed by the USGS (Katz et al., 1999) indicated that N levels have increased from 0.1 to 5 mg L -1 in many springs in the Suwannee River Basin over the past 40 years. Several sources of non-point N (fertilizer, animal wastes, atmospheric deposition, and septic tanks) increased continuously in Gilchrist and Lafayette counties. Isotope analyses indicated that the source of N was mixed between organic and inorganic, attributing them to animal operations and crop fertilizer, respectively, from nearby operations. This study (Katz et al., 1999) compared long term trends in nitrate concentrations in selected springs with estimated inputs of N sources in Lafayette and Suwannee counties. It found a high correlation between the estimated quantity of fertilizer and the concentration of N in the springs. It showed for Suwannee a peak in the 1970s, a

PAGE 44

22 decreasing trend until the early 1990s, and finally an increasing trend again after 1993, while for Lafayette County there was a steady increase since 1960s. There are historical records for 4 springs in the Suwannee River (Hornsby and Ceryak, 1999): Charles (1973-97), Little River (1980-97), Running (1980-98), and Telford (1980-97) (Figure 2-4); and for 3 springs in Lafayette county ((Hornsby and Ceryak, 1999): Troy (1960-98), Lafayette Blue (1980-97), and Mearson (1980-98) (Figure 2-5). Figure 2-4. Long term records of Nitrate-N for 4 Suwannee springs. Source: Modified from Katz et al. (1999), p. 46. Pitman et al. (1997) reported that nitrate concentrations in the Suwannee River itself have increased at the rate of 0.02 mg L -1 year -1 over the past 20 years and that over a 53-km river stretch between Dowling Park and Branford, the nitrate loads increased from 2,300 to 6,000 kg day -1 Of this total, 89% appeared to come from the lower two-thirds, where agriculture that includes crops and livestock is the dominant land use. This study found that nitrate-N concentrations in the 11 measured springs ranged between 1.3 and 8.2 mg L -1 and nitrate-N concentrations for the river of 0.46 mg L -1 in Dowling Park, 0.46 mg L -1 in Luraville, and 0.86 mg L -1 in Branford; even though the nitrate-N

PAGE 45

23 concentration almost doubled from the upper to the lower part of the river, there was no noticeable difference among the contributors springs, suggesting that other sources might also be affecting the river in the lower part. Figure 2-5. Long term records of Nitrate-N for 3 Lafayette springs. Source: Modified from Katz et al. (1999), p. 46. 2.2.2 Assessments at the Watershed Level Asbury and Oaksford (1997) performed a study in 7 rivers in Georgia and Florida; the Suwannee River was one of them. This study found that the N input ranged from 2,400 kg yr -1 km 2 in the Withlacoochee river basin to 5,470 kg yr -1 km 2 in the Altamaha river basin, with the Suwannee input in the range of 3,939 kg yr -1 km 2 However, when these numbers were recalculated to in-stream loads, the Suwannee had the highest N loads (compared with the other 6): 476 kg yr -1 km 2 (with the range of the other 6 between 152 and 405 kg yr -1 km 2 ) and discharges in the order of 9,700 kg 10 3 (with the range of the others between 720 and 8280 kg 10 3 ). Additionally, the ratios between outputs/inputs of N were calculated for the 7 rivers (range 4.2% to 14.9%) and the Suwannee River was the second highest with 12.1%.

PAGE 46

24 Ham and Hatzell (1996) realized another comprehensive study of N pollution in surface waters of Georgia and Florida. This study found, for the 10 major river basins in the study, that median concentrations were below USEPA guidelines of 10 mg L -1 for nitrate and 2.1 mg L -1 for ammonia, with only one exception in Swift Creek (3.4 mg L -1 ammonia) in the Suwannee River Basin that was attributed to waste water discharges. Although in general the nutrient levels were low, long-term increasing trends of nutrients loads were found in most of the sites. Distribution of results of nitrate-N indicated that the Suwannee presented the third highest mean but the higher variability with the highest overall value of more than 70 mg L -1 Ham and Hatzell (1996) also performed a long-term trend for these 10 major river basins. They developed graph relationships between time and N concentrations. They found a slope of +0.02 for the Suwannee River Basin, between the years 1971-1991. This means a 4.61% yearly increase over the baseline of 0.5 mg L -1 This was the fourth greatest increasing trend regarding mg L -1 year -1 comparing with the other Basins. 2.2.3 Assessments at the National Level Berndt et al. (1998) published a USGS regional study of the Georgia-Florida coastal plain that includes the Suwannee River Basin. This study was part of the National Water-Quality Assessment (NAWQA) program that includes more than 50 of the nations largest river basins and aquifers. This study reports that in the 23 groundwater samples from the row-crop agricultural area in the upper Suwannee River Basin, 33% exceeded the drinking water standard. These samples were from aquifers that overlie the Upper Floridian aquifer, the major source of drinking water for the study area. This study also found that nitrate concentrations were generally higher in samples from agricultural

PAGE 47

25 areas and shallower wells and N isotope ratios indicated that both fertilizer and animal wastes were the predominant sources of nitrates in groundwater. The study of Berndt et al. (1998) found that the nutrient more frequently found in groundwater was N, which was mostly affected by land use, geology, soils, and topography. Especially in the Suwannee River Basin, the problem is aggravated because it is a karst region with well-drained soils that have a low water-nutrient holding capacity. Also Berndt et al. (1998) found no direct relation between land uses and N concentrations. However, they referred to the previous Pitman et al. (1997) study between Dowling Park and Branford (33-mile Suwannee River reach) including 11 nearby springs that showed much higher N levels in the springs ( x = 1.7, range 1.3-8.2 mg L -1 ) than in the river (between 1 and 2 mg L -1 ): the N levels increased when it moved to the lower parts. In this case, even though the water flow dynamics have a great effect, a clear association was found between the predominant agricultural land use and the N levels; however no specific type of farms could be identified as the sources. A comparison of the Suwannee River nutrient concentration with the NAWQA national standards, indicate that overall the levels of N in the streams and in the groundwater of the Suwannee River Basin are between the 75 th percentile and the median. 2.2.4 Assessments at the Dairy Farm Level Andrew (1992) did a comprehensive study of nine dairy farms in the middle Suwannee monitoring waste water lagoons, monitoring wells, and supply wells and their interactions with the different land uses inside the farms (waste water, intensive areas, intensive pasture, sprayfields, and low density pastures). This study tracked the N inside

PAGE 48

26 the dairy farms after cattle excretion, measuring it at different depths and in different land uses. The results are quite revealing and they are summarized in Figures 2-6 and 2-7. 0102030405060708090100110120130140150160170180190200210220230NO3+NO2NH3+OrgNNO3+NO2NH3+OrgNNO3+NO2NH3+OrgNNO3+NO2NH3+OrgN Waste WaterShallow GroundwaterDeeper GroundwaterNearby hydrogeologic unit mgL-1 NO3+NO2NH3+OrgN Figure 2-6. Nitrate-Nitrite and total ammonium plus organic N at different points in nine dairy farms in the middle Suwannee River Basin. Source: Prepared with data extracted from Andrew (1992). Two main forms of N were measured, Nitrite + Nitrate (-) and Ammonium + Organic N (-Org). In the waste water in the lagoon (effluent) the concentration of 3NO 23NO 4NH23 3NO NO was negligible because most of the N was in form of -Org (=150 mg L 4NH3NO _x -1 ) as seen in Figure 2-6. Because of fast decomposition and nitrification N is quickly converted to forms, which are evidenced in the monitoring wells (shallow groundwater at 3-m deep), in which the concentrations of are far higher than the -Org forms (= 22 vs. 2.2 mg L 3NO4NH 23NO 23NO _x -1 ). The trend continues in the deeper groundwater source wells (12 m) where the concentrations of

PAGE 49

27 3NO are substantially lower than in the shallow groundwater wells because of dilution processes, but still high for the water quality standards ( 23NO x3 =10 mg L -1 ). In these deep wells, the -Org concentrations are much lower than the waste water, but slightly higher than the shallow groundwater suggesting that some deep accumulation may occur. 4NH 051015202530mgL-1 Intensive A In the nearby hydrogeologic unit the concentrations of are very similar to those in the deep groundwater wells suggesting that the dairy might be responsible for such an impact. As expected -Org forms would be negligible on these components. NO 23NO 4NH reaSprayfieldWaste WaterIntensive PastureLow DensityPasture00.511.522.53mgL-1 Ammonium+Organic N NO3+NO2 Figure 2-7. Dissolved Nitrate+Nitrite and Ammonium+Organic N in monitoring wells (3-m deep) in different areas of a dairy farm. Source: Prepared with data extracted from Andrew (1992). Inside the dairy farms, the collected data demonstrated the movement of N downwards at different rates for different parts of the farm. Exploring the data from shallow wells, the intensive area, where cows deposited manure without regular

PAGE 50

28 collection, is the critical area where the highest concentration of is observed ( 3NO 23NO x = 26 mg L -1 ); this is followed by the sprayfields where intense applications of effluent are performed ( x = 23 mg L -1 ); followed by the wastewater area and the intensive pasture (both x = 11 mg L -1 ). The lowest concentration was found in the low density pastures ( x = 5.1 mg L -1 ) but this is still half of the allowable limits of the PDWS. The -Org concentrations are medium ( 4NH x = between 0.96 and 2.1 mg L -1 taking into account that 2.1 mgL -1 is the USEPA limit) in these shallow wells; being higher, as expected, below the wastewater lagoons. This also indicates that there was a downward infiltration of the wastewater. In another study from Andrew (1994), four dairy farms were closely monitored during four years (1990-1993) in the middle Suwannee. In this study, water samples were evaluated in 51 monitoring wells in these 4 typical dairy farms and 3 nearby springs in Lafayette and Suwannee counties. Results indicated that Water from the monitoring wells had median nitrate exceeding that in water from wells of the background groundwater quality network sampled by the FLDEP Nitrate concentrations near waste water lagoons and intensive livestock areas were the highest exceeding in the shallow wells the drinking standards of 10 mgL -1 Concentrations were substantially lower in pastures (less loading of manure) and in deep wells (dilution and/or denitrification) Nitrogen isotope ratios analyses demonstrated that animal waste was the main source of nitrates in groundwater adjacent to the lagoons and intensive areas, and animal waste and/or septic tanks were the principal source of nitrate at the three springs Nitrogen isotope ratios also demonstrated that synthetic fertilizers or combinations of synthetic fertilizer, soil N, and livestock wastes were the sources of nitrate in pastures and sprayfields.

PAGE 51

29 Figure 2-8 shows the average of nitrate concentrations above the 10 mgL -1 for the 4 farms. As reported in the conclusions of Andrew (1994), the wells nearby the intense areas and the wastewater lagoons are those with higher environmental menace. Nonetheless, it was noticed with these study data, that 2 of the farms presented wells below the 10 mgL -1 for waste water lagoons and 1 farm showed wells of less 10 mgL -1 for a monitoring well near the intense areas. Also, 2 of the farms presented high wells for pasture fields and 1 farm additionally a high well for a sprayfield. 01020304050607080IntenseIntenseWastewaterPasturePastureIntenseIntenseWastewaterIntenseSprayfieldPasture FARM 1FARM 2FARM 3FARM 4 mgL-1 Figure 2-8. Monitoring wells with higher than drinking water allowable nitrate levels (PDWS>10 mg L -1 ) in the four dairy farms. Source: Prepared with data from Andrew (1994). 2.3 Human and Ecological Concerns because of N Pollution High nitrate levels in drinking water can cause methemoglobinemia (MHB) or blue baby syndrome in infants, a condition that diminishes the ability of blood to absorb oxygen (Vigil et al., 1965; NRC, 1985) because the methemoglobin produced by nitrates cannot carry oxygen, causing symptoms of oxygen starvation that affects

PAGE 52

30 primarily infants younger than six months, who suffer cyanosis and, in extreme cases, death (Andrew, 1994). Nitrate is a precursor of carcinogenic nitrosamines, which are believed to cause stomach cancer and leukemia (NRC, 1978; Strange and Krupicka, 1984; Stewart, 1990; Mirvish, 1990; Forman, et al., 1985). Additionally, high levels of nitrates in drinking water are associated with birth defects (Fan et al., 1987), stress of the cardiovascular system with extreme artery dilatation (Vrba, 1981), enlargement of the thyroid gland, and increased sperm mortality (Scragg et al., 1982). Kross et al. (1992) mention that about 2,000 cases of MHB have been reported between 1945 and 1990 worldwide with a mortality rate of 10%; however the NRC (1995) found no studies reporting nitrate-induced MHB since 1990; nitrate induced MBH seems rather rare now, but these rare cases are due primarily to a lack of reporting requirements, and secondarily to a lack of physician awareness (Felsot, A. 1998). Several effects on reproduction, growth and development of animals have also been noticed because of high N levels, including mortality of larvae of Cutthroat trout, Chinook salmon and Rainbow trout and mortality of tadpoles in several frog species in Florida such as Chorus frogs and Leopard frogs (Edwards et al., 2002). In several species of frogs, reduced activity, weight loss, and deformed digestive tracts also have been detected (Edwards et al., 2002). In addition to making water unsafe for humans and many other animals, high nitrate concentrations can lower water quality in rivers and springs because elevated concentrations of nitrate in these water bodies can cause eutrophication, which can result in algal blooms and depletion of oxygen that affects survival and diversity of aquatic

PAGE 53

31 organisms (Katz et al., 1999). Eutrophication is a process in which algae growth is favored and consequently oxygen is depleted. This problem happens in the surface water-bodies, which receive inflows from groundwater from springs or diffuse seepage (Andrew, 1994). 2.4 Conclusions and Recommendations Several studies have demonstrated that there is a N pollution problem in the Suwannee River Basin. All studies reviewed for this paper agree on it. Even though the scopes, methodologies, timeframe, and location differed, all the studies found that the levels of N in groundwater and surface waters had, at the very minimum, higher N concentrations than the background. Studies are not conclusive to mention if major pollution exists in surface water sources or in groundwater sources. Temporal, spatial, and climatic conditions would have great effect on overall N pollution levels. When comparing the Suwannee River Basin with other basins in the state, in the Watershed (including Georgia), and across states, it was found that, overall, the water in the Suwannee presents lower quality because of N pollution. The Suwannee River Basin or parts of the Suwannee River Basin are consistently mentioned by the studies as examples of extreme polluted values of any form of N either in groundwater or surface water. Even though, in most of the cases, the N concentrations in the Suwannee River Basin are below the USEPA standards of 10 mg L -1 several studies that include long term comparisons, along with hard data collected from monitoring stations, show that there is a noticeable increasing trend in the N concentrations of the surface and groundwater bodies that would tend to increase even faster if the land use conditions persist. Therefore, actions to mitigate the problem are needed.

PAGE 54

32 By exploring hard data generated by monitoring efforts in the Suwannee, the middle Suwannee (between Suwannee and Lafayette counties) continues being the hotspot and the area with higher N pollution concerns; nonetheless, there are other areas and places where the N concentrations are showing as high or even higher than the middle Suwannee as is the case of areas in Jefferson, Levy, Gilchrist, and Alachua counties. The current effort to decrease N sources in the middle Suwannee should be expanded to other areas of concern. It is reported that the possible nitrate sources are chemical fertilizers, animal wastes, and human waste. There is no clear cut conclusion as to which one of these represents the most dangerous source. Dairy operations are many times found as the most likely source of nitrate-N by several studies either by spatial proximity, temporal analysis, and/or by the use of isotopes. However, there is no clear estimation of how much N comes from dairy farms either as a whole or on an individual basis. It is believed that dairy farms are big contributors because of large amounts of waste produced; however there are no conclusive reports that estimate and/or compare them with other sources. Higher efforts are required to distinguish between point vs. non-point sources of pollution and to make any industry environmentally accountable as a way to face the problem and their potential solutions. Two in-depth studies in dairy farms in the study area evidenced a high environmental impact of these operations regarding N pollution. These studies pointed out that there is a great variability between dairy operations and their environmental impacts. An appropriate approach for environmental solutions should be to treat dairy farms individually regarding their pollution concerns.

PAGE 55

33 Also, these dairy studies found no conclusive answers to what components inside the operation are the greatest N polluters. These also vary substantially among farms. Even though only a decade has passed since these studies, the dairy farm systems have changed substantially, most of them now Do not have the intense area Have a concrete lagoon Do not use or use very little synthetic fertilizer The intense area has become part of the confined area in a totally concreted covered space where all the manure is collected and where there is no infiltration. The manure goes to a concrete lagoon and from there to the sprayfields, in which usually only waste manure is used as fertilizer. Under this new structure, the sprayfields, primarily, and the pastures, secondarily, are the most likely zones inside dairy farms with higher N concentrations and consequently with higher N pollution impacts. Dairy farms as major actors in the N issue must be studied as a whole system with complete understanding of all their components under specific environmental and management options affected by varying climatic conditions. There is clear evidence that water N pollution affects humans, animals, and ecosystems. In the study area, even though there are no reported direct or indirect human effects, there is major evidence of animal and ecosystem effects because increased N levels.

PAGE 56

CHAPTER 3 NORTH FLORIDA STAKEHOLDER PERCEPTIONS TOWARD NUTRIENT POLLUTION, ENVIRONMENTAL REGULATIONS, AND THE USE OF CLIMATE FORECAST TECHNOLOGY 3.1 Introduction Rising nitrate levels in north-Florida's Suwannee River are a growing concern (Hornsby et al., 2003). Awareness is high given the potential of excess nutrients to affect the environment and public health (Albert, 2002). It is alleged that the most likely contributors to contamination are dairy farms, industry, and human settlements in the region (Andrew, 1994). The federal safety standard for nitrate-N in drinking water in the United States is 10 mg L -1 In some areas of the Suwannee River Basin, nitrate levels have reached ranges of 20 to 30 mg L -1 (Hornsby, Pers. Comm.). High nitrate levels in drinking water can cause methemoglobinemia or blue baby sickness in infants (NRC, 1985) and other health problems in humans such as enlargement of the thyroid gland, increased sperm mortality, and even stomach cancer (NRC, 1978; Strange and Krupicka, 1984; Stewart, 1990; Mirvish, 1990; Forman, et al., 1985). In addition to making water unsafe for humans and many other animals, high nitrate concentrations can lower water quality in rivers and springs, causing eutrophication resulting in algae blooms that consume oxygen needed by fish and other aquatic animals (Katz et al., 1999). Dairy farmers are now required to comply with stricter environmental regulations either under permit or under voluntary incentive-based programs. Dairy farmers are also aware that environmental challenges will be the greatest problem they will have to face in 34

PAGE 57

35 the near future. After interviewing 48 dairy farmers located in Lafayette, Suwannee, Columbia, Hamilton, and Gilchrist counties, Florida, Staples et al. (1997) found that the perception of anticipated costs of having to comply with environmental regulations was rated, by far, the top challenge to successful dairying in the future. Improvements in long-term (6 to 12 month) seasonal climate variability predictions such as those based on El Nio and La Nia forecasts (OBrien, 1999) can play an important role in formulating management strategies that dairy farmers in north-Florida may adopt in order to pursue economic and ecological sustainability. Specialized federal and state agencies such as the National Oceanic and Atmospheric Administration (NOAA) predict long term seasonal climate variability fairly accurately. These predictions are based upon the El Nio Southern Oscillation (ENSO) phenomenon. ENSO phases determine El Nio, La Nia, or Neutral years, which are related to rainfall and temperature patterns. Dairy farmers may benefit from proactive use of seasonal climate variability prediction. The opportunity to benefit from climate prediction arises from the intersection of human vulnerability, climate prediction and decision capacity (Hansen, 2002a). This benefit from the information can be used to decrease environmental impacts while maintaining profitability in north-Florida dairy operations. This paper discusses the perspectives, perceptions, and attitudes of stakeholders regarding the issue of N pollution in the Suwannee River Basin, the current regulations, and the possibility of incorporating climate variability forecast technology to initiate mitigating strategies. The principal stakeholders are the dairy farmers. The

PAGE 58

36 methodologies included a Sondeo with stakeholders, 21 interviews with dairy farmers, and eight focus groups with stakeholders. Stakeholders perceptions regarding nutrient overloads, regulations, and the use of seasonal climate information as a way to mitigate the nitrate problem are critical baseline data for designing solutions. The information is even more valuable because it is gathered in a participatory and consensual manner. This study translates the mind-set of the stakeholders about the N pollution issue and of the dairy farmers regarding environmental regulations and the use of climate variability forecast technology. 3.2 Materials and Methods 3.2.1 Study Area The Sondeo was carried out in the Suwannee River Basin that encompasses all of nine and part of another six north-Florida counties. This represents an area of 7,640 mi 2 The area is comprised of 13 river basins and includes a population of 285,000. Within this area, the principal focus was on the middle Suwannee that includes Lafayette and Suwannee counties. These two counties are specific areas where the N issue is acute. In the middle Suwannee River there are hundreds of residential and commercial septic systems in rural areas. Some 300 row crop and vegetable farms also exist. Finally, 44 dairies with more than 25,000 animals and 150 poultry operations with more than 38 million birds are installed in the area (Suwannee River Partnership, 2003). Interviews and focus groups were undertaken in Lafayette, Suwannee, Gilchrist, Levy, and Alachua counties, where the number of dairy farms totals 64. 3.2.2 The Sondeo A Sondeo was conducted between the months of January and February 2003 by a multidisciplinary team of eight graduate students from the University of Florida (Alvira

PAGE 59

37 et al., 2003). The Sondeo (Hildebrand, 1981) is a team survey developed to provide information rapidly and economically about agricultural and rural problems. It is structured around a series of informal, conversational interviews between the team and stakeholders. It is a multidisciplinary process from data collection through report writing. Each team ideally includes members from the social and agricultural sciences, as well as other disciplines. When members write the final report there is much interdisciplinary cross fertilization. In the Sondeo, data are shared among the different teams and report writing is done as a group so that observations are confirmed, debated, and analyzed with members of the other teams. The results may be quantified or not but the accuracy of the findings is strengthened by the cross-checking process. During the Sondeo, 13 stakeholders were interviewed. This sounding out was undertaken to get a better understanding of the broad issue of N pollution in water and collect the perception of stakeholders regarding this issue in the study area. Sondeo team members represented a number of disciplines including agronomy, animal science, agricultural extension, natural resource management, forestry, and community development. Two to three-person teams conducted initial transects and stakeholders interviews. The interviews were conducted as informal, open-ended conversations. After a team introduced itself and explained the subject of the research, the stakeholder and the team conversed about any aspects of the topic that were relevant. This open-ended approach enabled topics to emerge and be pursued that might have been missed if the researchers had been constrained by a pre-formulated questionnaire. Despite the open-endedness, a set of core themes was discussed in almost all interviews. Notes were not taken. Following each interview, the partners wrote individual notes, compared them,

PAGE 60

38 and then wrote a joint report for the interview. Finally, the joint reports were shared and thoroughly discussed among all Sondeo members. Each team presented its findings, which could be clarified, challenged or contrasted with the results of the other teams. This process of reporting and discussion served as the opportunity to begin noting trends, gaps in information, sources of initial inferences, and identifying new questions to pursue. 3.2.3 Interviews Personal interviews (Bernard, 1995) were undertaken with dairy farmers and/or dairy farm managers during the fall/winter seasons of 2003/2004. Open-ended interviews were targeted to a purposeful sample of 21 farmers (30% population). The number of interviewees was chosen in order to attempt to capture the broad variability represented in the area in a collaborative manner with the Suwannee River Partnership and the University of Florida Extension Service agents in the area. A voice recorder was used in every interview after requesting permission from the interviewee. Interviews included a series of topics, but emphasized farmers perceptions regarding the N pollution regulation issue and the potential use of climate forecasts to address the problem. 3.2.4 Focus Groups Eight focus groups (Bernard 1995; Stewart and Shamdasani, 1990) with stakeholders were held between November 2002 and February 2004. Focus groups were open discussions among 5 to 15 people. Topics focused on perceptions regarding nitrate pollution, regulation, and seasonal climate forecast innovation technologies as an aid to mitigate the problem. Focus group participants ranged from specialists from official agencies such as the Suwannee River Partnership (SRP), the Suwannee River Water Management District

PAGE 61

39 (SRWMD), the Florida Department of Environmental Protection (FLDEP), the Natural Resources and Conservation Service (NRCS) from the United States Department of Agriculture (USDA), the University of Florida IFAS Cooperative Extension Service (CES), to private sector companies, technicians and farmers. These participants were typically called to meetings by the Suwannee River Partnership and/or by extension agents from different counties in the research area in coordination with the researcher. 3.2.5 Analysis All participatory interactions were recorded, classified, organized, analyzed, and described. This paper summarizes the results of all the above interactions in a descriptive/narrative manner. Names of the participants are not revealed to protect the anonymity of informants. Emphasis has been placed on a qualitative rather than quantitative description and analysis of farmer perceptions. The interdisciplinary approach to the complex problem of nitrate pollution in the Suwannee River Basin requires the use of methods from social sciences such as applied anthropology and rural sociology for a more holistic view. 3.3 Results and Discussion 3.3.1 Stakeholder Perceptions of the N Issue 3.3.1.1 Perception of water contamination During a discussion in an initial focus group, one stakeholder, supported by several participants, challenged two main points: does water quality degradation caused by N pollution really exist in the Suwannee River; and if the problem exists, how bad is it? Stakeholders argued that is not easy to know if there is a real problem, and even more difficult to determine the source of the problem. Additionally, an actual timeline of events that may have led to the current situation is unknown. Farmers present in the

PAGE 62

40 meeting perceived that the 10 mg L -1 standard of N concentration in water enforced by the regulatory agencies is too strict compared to other parts of the world (Europe for example, has a 40 mgL -1 limit). This perceived low threshold is also believed to exaggerate the real problem. The dairy producers expressed that to their knowledge, there are no indications of humans being affected by N concentrations in local water. Farmers also believed that eutrophication of local water bodies is not as evident as media reports suggest. In their eyes, the appearance of streams and springs is similar today as it has always been. Some farmers present at this meeting believed that there is a misconception of the link between the N issue and dairy farms. Participants indicated that dairy farms in the study area do not always have overflow problems. Most indicated that in general terms, large operations have higher chances of overflows. Additionally, many of the participants indicated that they were not convinced that dairy farms in general are the biggest polluters. Conversely to these opinions, the Sondeo found that ten years earlier, an increase of nitrate concentration was recognized by the Suwannee River Water Management District in the Suwannee River Middle Basin. The increase in nitrate concentration in the Suwannee River Middle Basin as a problem was clearly evident to all stakeholders who participated in the Sondeo. A veterinarian working for a dairy farm expressed his concern for the pollution, and its potential for driving them out of business in the future. It seems clear that dairy farming is perceived as the most contaminating activity in the area. Poultry is perceived to be the second source of contamination. Row crops are usually perceived as the least contaminating of all food producing activities in the area.

PAGE 63

41 However, one interviewee pointed out that greater leaching might result from fertilizer in row crops compared to the slower decomposition of manure from dairy operations. This would occur when precipitation follows an application of fertilizer in row crops, or if they over-irrigated. Moreover, the informant perceived that all farm activities might contaminate much less than non-agricultural industrial activities and waste outputs from concentrated urban dwelling in the area. Even though the main focus in the discussions had been N loads, phosphorus (P) was also mentioned as a potential contaminant by participants in the Sondeo. A problem expressed about both these contaminants is that they may take many years to reach the groundwater. Nutrient pollution has a long term cumulative effect that results from slow processes. This makes it much more difficult to pinpoint sources of contamination and consequently potential ways of mitigation. The Suwannee River Partnership was formed to determine the sources of N loads in the basin. The other important and connected objective was to work with local users to minimize future nutrient loading through voluntary, incentive-based programs. Therefore, the Partnership works under a non-regulatory precautionary approach. This is envisioned as being accomplished through monitoring pollution, education, promotion and adoption of best management practices (BMPs). The Partnership is focused on the middle river basin. The main activities in that area i.e., dairy, poultry, beef, and row crop farms are the foci of the Partnership. Currently, there is little population pressure since the middle river basin is not very densely urbanized. However, a recent geographic extension of the Partnership to the southern part of the basin towards the Santa Fe River includes more densely populated

PAGE 64

42 areas. A logical step then would be for the Partnership to focus more deeply on septic tanks and human dwelling waste outputs. 3.3.1.2 Reasons to mitigate contamination The main reasons identified by the stakeholders for mitigating water contamination are to preserve the ecosystem, protect human health, protect fresh water bodies that provide environmental services, and support economic activities such as tourism. It is important to note that nobody mentioned that the water quality has an impact on their herd, poultry, or crop. It is also recognized that pollution in the river basin can have a negative effect in shellfish production in the Gulf of Mexico. 3.3.1.3 Existing and potential means of mitigation Stakeholders were proactive about avoiding potential regulations, which did not seem a good choice for addressing the problem to them. Many farmers might find themselves in a difficult situation to fulfill imposed regulatory levels. Farmers did not want to be regulated to have to implement conservation measures. Lake Okeechobee (South Florida) regulations, which were imposed to reduce the P load level, were not successful in the sense that 50% of the dairy farms in that area went out of business. Based on this past misstep, stakeholders worked under a non-regulatory, preventive, or precautionary approach, through the creation of the Suwannee River Partnership (SRP). The recognition of rising nitrate concentrations was the leading factor behind the creation of SRP. The idea of forming a Partnership took years. According to some stakeholders, just a few people started the partnership in the mid 1990s. More institutions progressively joined in the following years. In 2003, there were 24 members, and currently (2004) the number of partners has increased to 45. The long-term goals of the partnership are to

PAGE 65

43 Reduce pollutant loading Establish interstate watershed management Increase number of acres covered by conservation and land management plans Assure flow of information Secure cooperative long-term funding to maintain a cooperative interstate program Improve water conservation between two states (Florida and Georgia). Several agencies and individuals are part of the Partnership that collaborate to monitor and to reduce the levels of nitrates in their groundwater and surface waters. For example, the USDA/ NRCS PL-566 Small Watershed Program for the Middle Suwannee assists dairy and poultry farmers to improve animal waste management and fertilization techniques. The NRCS inspects soil types, number of cows and area in pastures, and then decides the minimum set of measures to be followed. The Florida Farm Bureau (FFB), that represents farmers interests, is also involved in developing policies to mitigate nitrate pollution in the area. The FFB annually publishes a book containing all of its policies. The Farm Bureau supports technology-based BMPs and their voluntary implementation. The U.S. Geological Survey (USGS) also has ongoing studies of water quality within the Suwannee River Basin The Partnership focuses on finding the most economical and technologically feasible management techniques (BMPs) available to help farmers and other land users to satisfy regulatory requirements for protecting public health and the environment. Through education and outreach programs, the group continues to increase public awareness of the issues, and encourage citizen and community participation in working together to find solutions. Other components of the initiative that are contributing to the Partnership's success include the BMP Quality Assurance Program, County Alliance for Responsible Environmental Stewardship (CARES), and the On-farm Research Programs

PAGE 66

44 As one component of Partnership activity, several University of Florida departments and centers, as well as state and federal agencies are cooperating on project 319 entitled "Effectiveness of Best Management Practices (BMPs) for Animal Waste and Fertilizer Management to Reduce Nutrient Inputs into groundwater in the Suwannee River Basin." Project 319 addresses BMPs on a dairy, a poultry operation, and a vegetable farm in Lafayette and Suwannee Counties in the Middle Suwannee River Basin. Each farm has been instrumented with a groundwater monitoring well network. Wells and soil profiles are being intensively monitored to determine preand post-BMP nitrate concentrations on each of the farms. An one-year pre-BMP monitoring phase has been completed on all the farms, and BMPs are now being implemented to address the elevated nitrate concentrations observed in some of the monitoring wells. The overall goal of this project is to demonstrate that BMPs can be implemented to reverse the increasing trend of nitrate concentrations in the groundwater and springs of the Middle Suwannee River Basin. The BMP Quality Assurance Program verifies that BMPs are operated and maintained over a long period of time, and provides assistance to farmers in resolving problems with the BMPs. The Partnership, in cooperation with Florida Farm Bureau, has also initiated the CARES Program to certify and recognize farmers participating and following directives in the Partnership program. Finally, the Partnership has developed a joint strategy with the SRWMD to monitor groundwater and surface water on a regional basis to identify trends in water quality over time through an extensive network of monitoring points.

PAGE 67

45 Technicians from the Partnership visit farms on a regular basis to monitor their N loads and to help with the implementation of BMPs. One of these technicians pointed out that many farmers were formerly wasting money buying fertilizer. With new techniques such as use of a cement pool and mixer in the lagoon to separate solids from liquids, the farmers are able to recycle their own manure, buy less fertilizer, reduce costs, and reduce N loads. In general, farmers who have been applying BMPs feel comfortable with the results achieved. Most farmers are cooperative and eager to follow SRP directives and the suggested BMPs. Nevertheless, some practices (such as treatment lagoons or solid separators), require large investments. Money is required up-front to modify facilities, train personal and to make other adjustments. In most cases the SRP has funding available to cost-share initiatives that protect the environment. The modality of implementation for most BMPs is completely based on voluntary participation of farmers and aims at helping them avoid stricter regulations. However, certain activities are already subject to regulatory programs. 3.3.1.4 Implementation procedures for Best Management Practices (BMPs) BMPs for nitrate reduction in the Suwannee River Basin refer to good stewardship of land and water management that help farmers reduce nutrient contamination to water systems. There is general acceptance among stakeholders that such management practices are sound and positive for the environment. Management practices mainly include nutrient and manure management, disposal of wastes, irrigation water, technical design and engineering aspects in farming enterprises such as row crop, dairy, beef cattle and poultry farms. By large, the concerned stakeholders see BMPs as research-generated solutions to reduce potential environmental hazards from farm enterprises.

PAGE 68

46 Implementation procedures are specific to each program and farm enterprise. All programs aim at implementing BMPs that reduce the loading of N contamination to groundwater while helping farmers to remain profitable. It is worth mentioning that at present, attaining the twofold objectives is at an early testing period in which cost-sharing is crucial to ease associated risk for farmers. Based on the information gathered by the Sondeo, a picture of the step by step process of BMP implementation in the study area can be drawn. Creating and raising awareness among public agencies, landowners, agricultural producers and other stakeholders about existing water quality impacts resulting from human activities is the first step. It is necessary for recognizing that a new approach in management practices is needed to address the issue within the area. It is perceived that education on the issues and solutions, including effective transfer of knowledge and technology, are essential components of the implementation efforts of the technical working groups in the Partnership. As part of this process the Partnership conducts meetings for agricultural producers. During 2001 over 1000 farmers participated in such meetings. Gatherings included two row crop growers meetings, a forage field day, five dairy farmers meetings, two poultry growers meetings, a cattlemens meeting (beef), and a large-scale growers meeting. In the CARES program, growers fill out a self-evaluation regarding their farm operation to assess environmental practices. One objective of this self-assessment is to increase farmers awareness in perceiving their own status. The next step is that farmers show interest in joining the Partnership. At this stage agencies in the Partnership help farmers in selection and implementation of BMPs. The

PAGE 69

47 partnership has a Technical Working Group (TWG) in the areas of fertilizer, animal waste and human waste management. Each TWG assesses sources of nutrients within its area of expertise, develops and implements a plan of work to fulfill the mission of the Partnership, and provides updated recommendations for action. BMPs are highly decentralized and tailored to suit the diversity in size and complexity of farm operations. Initially, the focus is on management approaches, which are technically viable, economically feasible, and environmentally friendly. The study team feels that BMPs on dairy farms are far from being financially feasible. Up-front costs for implementation are especially hard to recoup. Additionally, BMP on-farm trials are conducted to verify the technical viability of the recommended practices. Financial commitment and debt responsibility is required from farmers during implementation of BMPs. To date, BMP implementation has been very successful. Out of 144 crop farms, 115 have adopted conservation measures; out of a 35 poultry farms 34 participate; and out of 44 dairy farms, 43 are involved in the SRP. Before the Partnership, only 3 dairy farms were able to pass environmental quality standards. After the farmers adopt BMPs that have been promoted by the participating agencies, a presumption of compliance with state water quality standards can be established. This places farmers in a much better standing of willingness to cooperate with environmental stewardship with authorities. 3.3.1.5 Impacts of BMPs on farms Most stakeholders perceive that farmer participation in BMPs has many economic, environmental and social benefits. Stakeholders agree that after the initial investment, dairy, crop and poultry enterprises save money by implementing BMPs since they recover their own nutrients, which have a value. In the case of dairy farms, the savings

PAGE 70

48 could be important since no fertilizers would be needed for crops. However, stakeholders perceive that poultry farmers have only modest economic benefits because the processing of dry litter does not return much income. An additional benefit indicated was that livestock farmers who implemented BMPs are usually allowed to expand herd or flock size. Regarding row crops, stakeholders indicated that the BMPs being promoted for crops reduce yields. This reduction in crop yields seems to be inevitable because of the reduction in fertilizer and irrigation contained in the BMP recommendation. Additionally they would require farmers to spend more time and money on management and equipment. Both aspects would decrease profitability. A major challenge then, is how to maintain row crop farmer profit margins while implementing BMPs. In spite of difficulties, adopting BMPs may help farmers to remain in business in the long run, which is critical to the Partnership. The initial investment that farmers are supposed to share may still be a big challenge considering the returns that farmers obtain as a result of adopting BMPs. But ultimately, middle Suwannee farmers will be able to choose from one of the three options: receive financial assistance to practice BMPs and stay regulated under incentive-based programs; receive a certain amount of money and leave the land; or comply with regulations under permit on their own. 3.3.2 Perceptions of Regulations by Dairy Farmers During the interviews, dairy farmers expressed their concerns that regulations are currently stricter than in the past, and that they will become even stricter in the future. At the same time, several farmers expressed that the manure and all its by-products have a tremendous commercial potential as a source of energy and nutrients. Potential benefits from by-product sales could mitigate stronger future regulations.

PAGE 71

49 Sixty-two percent of interviewed farmers participate in incentive-based regulatory programs. One third of them work under permit, and one farm has no regulation at all. Incentive-based programs are voluntary commitments between the farmer and the regulatory agency as promoted by the SRP. Farmers under permit are those that are enforced by the regulatory agency. Usually medium and small farms (up to 700 cows) are under incentive-based programs and larger farms are under permit. Also, it can be said that newer farms are under permit and older farms are more likely to be under incentive-based programs. All farmers expressed a clear message. They wanted to make sure the general public understands that producers are not trying to pollute the environment. Farmers only seek to earn their living and to do their best to protect the environment. Eighty-one percent of dairy farmers interviewed agreed or felt comfortable with current regulations. However, the remaining 19% disagree more or less with the current environmental regulation framework (Figure 3-1). The one farm that is not regulated is a small farm that received an invitation, but does not want to participate in the cost-share program because the owner believes that they can protect the environment in their own way. Fundamentally, the farmer does not want to be controlled at all times. He is confident the problem will not increase from present levels and that his farm will continue functioning as it has done for more than a half century. However he is aware that if they are forced into compliance with regulations, he would probably go out of business. 3.3.2.1 Farmers in disagreement with regulatory measurements One of the farmers in dissension was very disappointed and very upset about the regulatory agencies, especially with the Florida Department of Environmental Protection

PAGE 72

50 (FLDEP). He felt betrayed in the fact that he was once told that he runs an exemplarily clean farm and was even publicly recognized for that fact on one occasion. Later, after activists groups sued the FLDEP, the department requested he make a cumbersome investment on his own. The farmer stated Their stuff is not scientific, they listened to what people say and they got scared. They do not have any money for me, as they have been able to provide for many other dairies. They have put 18 people out of work on this farm because I had to reduce the herd size by half. If they place just one more regulation on me, I will be out of business The indispensable requirement for scientific studies in the eyes of this farmer is an interesting and novel point. Often, the opposite view might be expected. Apparently, the FLDEP in this case is regarded as a politically driven agency that does not necessarily back up their case with science. A perception of inequitable treatment on the part of the cost-share program with regard to this farm might be regarded as a strong motive for this farmers ill feeling toward the regulation process. N regulations 81 % B Agree completely Criticize part of it Administrative observations Technical observations Guarantee of sustainability Climate conditions not considered Rapid personnel turnover Not enforced severely Are BMPs really good? Listen more to public and news Dont treat farmers equally Too bureaucratic Too imposed Are BMPs really good? Whole farm system understanding Lack of soil science/scientists Dont Agree Agree 33 % 67 % 50 % 50 % Figure 3-1. Farmers perceptions of environmental regulations in north-Florida

PAGE 73

51 This farmer elaborated further saying that the FLDEP requested he install more concrete free-stalls to collect manure and more irrigation systems to dispose of the effluent in the fields. He thinks that is not a correct solution because it would create even more problems with neighbors concerned with odors. He added I did not even have a chance to discuss my options. The regulation was imposed. They told me: do all this or reduce your herd size. I decided to reduce the herd size because I felt that I could not trust the regulatory people any longer. It would not be surprising if, after I make all these investments on my own, after a few years they would come again and ask me to change all over again because the neighbors are not happy This farmer expressed a common complaint in processes of this nature. He was left out of the process and not consulted from the beginning. Furthermore, it is apparent to the farmer that no matter how hard he tries; the dairy farm is in a vulnerable position with respect to new regulations in the future. In this farmers opinion, the interests of the farmer will always be secondary to those of the neighbors. Another complaint regarding the regulations was from one relatively new farmer, who had to fight for more than two years to get a regulatory permit approved for a small dairy operation. The farmer considers himself to be over-regulated compared with other much larger dairies in the area. The farmer also mentioned that the approval process has been very costly and the farm could not count on any financial aid. He felt that he was forced to go through the process. This farmer mentioned I have fewer than 400 cows and they made me install and monitor six wells with all the expenses it implies while there are farms with thousands of cows who do not have this type of control. This farmer feels that reports in the news have been particularly prejudicial to his situation and probably one of the causes of this very strict permit. The other two farmers who do not agree with current regulations based their perceptions more on technical issues. Both considered that the

PAGE 74

52 suggested BMPs would not be the most appropriate either for the environment, or for the farms. One important point was the fact that the main BMPs promoted include a flushing system (dilution of manure to spray it in fields) that, to quote one of the farmers: Seems to be aimed at losing most of the N to the air instead of trying to use it as fertilizer. Besides, the promoted system requires a big investment and is fixed to a physical space. Why do they not promote a system, like the one in the northern part of the country that uses slurry (with minimum flushing) and gives the farmer the flexibility to apply it to different plots or even easily export it off the farm? The farmer offered a strong critique in this statement of the technical expertise of those in charge of designing and implementing BMPs. The producer apparently perceives that the system proposed simply postpones further regulation for the moment by converting a local negative externality into a diffuse global one. In essence, N is being spewed into the atmosphere instead recycling it for other uses. The other important point was the lack of understanding by those creating BMPs of the dairy farm as a whole system. One of the farmers mentioned that very capable people exist on the regulatory teams, but there is a lack of personnel with an interdisciplinary understanding of the dairy system. He has witnessed problems other dairies have had because the engineered designs did not take into account all the components of the system and their interactions. Indeed, he is aware that some BMPs are being revisited and redesigned after several failures have been detected. These two dairymen think that the soils are the main component in solving the N issue. They think that farms and regulatory agencies should work more aggressively with a focus on soils. They both believe there is a lack of soil scientists on the regulatory teams. The result is that regulatory agencies are not doing a good job protecting the environment taking into account soil dynamics. One of these farmers added that the

PAGE 75

53 regulatory agencies should clearly distinguish between real data and perception information regarding water pollution by N I am not completely sure if a real problem exists. Studies of the trends of N pollution are not completely conclusive. Nor do I recall were they conclusive in the case of the P issue in the Okeechobee area, where regulatory agencies imposed a limit of P run-off discharge lower than what common rainwater would have Some farmers hold a less than credible reputation of the FLDEP. This poor perception dates at least as far back as the years of regulation in the Okeechobee area. The loss of dairy farm livelihoods in that area was never proven conclusively to be the result of a tangible problem, at least in their minds. Another informant also corroborated this last point regarding mistaken regulation standards. The last two farmers mentioned apparently feel that precaution is not necessary and that only when science proves without a shadow of a doubt that something is harmful should regulations be imposed. These farmers apparently did not consider the notion that it might be too late for action by then. 3.3.2.2 Farmers who agree with regulations but criticize part of it Among respondents who agree with current regulations, some criticized some particular aspects related to the regulations. Two farmers mentioned that FLDEP fines farms without taking into account the climatic conditions that, for example, determine inevitably high N measurements in lysimeters when wet periods follow dry ones. Another point was the constant change of personnel in the regulatory agencies that could be disadvantageous to farmers. This is because over the years, officials become acquainted with specific farm situations. When new agents are hired, there is a period required for building-up new relationships and understanding site-specific problems. Besides this,

PAGE 76

54 new, usually young agents, tend to be more environmentally oriented. Often, they fail to grasp the fact that farmers also need to make a living. On the other hand there was also one farmer who criticized FLDEP and the regulations as being relaxed, implying that the regulations should be enforced more aggressively in order to avoid intervention by the federal Environmental Protection Agency (EPA). He perceived EPA regulations and enforcement as potentially much stricter, which would be very difficult for all in the area. Another criticism of farmers who do agree with the regulations came from one farmer who believes that a third party plays a role too important in the regulations. This third party is a term used to refer to activist groups that raise public concerns about environmental issues. In this farmers opinion, FLDEP takes their opinions into consideration disproportionately. The argument was that these third parties have nothing to lose and that the FLDEP listens more to activist groups and the public, and takes farmers concerns and situations little into account. Other respondents corroborated the fact that the news media have a great impact on regulations and on FLDEP enforcement of regulations. In terms of regulations and enforcement, the opinion of one journalist could have a greater worth than the opinion of a number of farmers. Some other farmers, who in general agree with current regulations, mentioned their concerns about the scientific basis behind the requested technical actions, which may not be the most appropriate for specific farms. This is related to the previously mentioned perception that the regulatory agency is not looking at the overall system. One farmer expressed: some times it is easy for them to tell you what you have to do. But, usually

PAGE 77

55 they do not know how difficult it is to implement these actions, and what other impacts could result from implementing the management changes. Many of the farmers expressed their concerns regarding other non-point potential sources of pollution. The main argument was centered on challenging the fact that dairy farms are the most contaminating of all. They believe that this statement still needs to be proven, and in any case they ask why are not row crop farms, golf courses, human settlements, among others regulated in the same way as dairy, poultry or pig farms?. One farmer mentioned: they count every pound of N the cows excrete on this farm but they do not control the neighbors farm that uses two to three times more fertilizer than the recommended rates for vegetable crops. Another point mentioned by several farmers may have important implications for regulations. Many people are regularly moving to live in the countryside in Florida. Often, these newcomers find that they do not like the manure odors until after they are well established. Most farmers think that newcomers should be aware of what they are buying into, and expect and accept normal rural odors. However, there are a few farmers who actually believe that just because farmers were established first is not enough argument to disregard odor problems of newcomers and they should work trying to mitigate it. One farmer is convinced that in the future, odor problems, more than any other issue, will put his farm out of business. Farmers are also looking for permanence and consistency in the regulations. Some farmers are concerned about potential changes in the set of regulations that might turn out to be disastrous for their conditions. Farmers have learned to survive under current conditions as long as the FLDEP does not change the basic, initial guidelines. This fact is

PAGE 78

56 related to a concern expressed by several other farmers regarding the potential that regulations might become unreasonable or impossible to implement if they change. This eventuality is perceived as even more uncertain because farmers understand that the FLDEP follows directions from the EPA. Farmers feel a total disconnect from the federal agency. Almost all interviewed farmers agree that the compliance process with the regulatory agencies is not a money-making adventure. All are conscious that it costs money, but they have no choice but to comply. 3.3.2.3 Farmers who completely agree with regulations At least 33% of the 81% of producers that agree are completely at ease with the regulations. All of them believe that they are working towards a sustainable way of doing business. Thus, the regulations are perceived as a guarantee of survival in the long run. For many of them there is a history of success in north-Florida because of the cost-sharing program that is allowing them to work together with the research, extension, and regulatory agencies in searching for regional and site-specific solutions. One farmer mentioned: the Partnership and the incentive-based program in the Suwannee River Basin are wonderful things. These will help farmers to be better off in the future. Although many farmers realize that mistakes have been made in the past, they believe that all parties are learning from the experience and that everyone together is building expertise to protect the environment and keep farmers in business. From all these farms there is a sense of engagement in the program for protecting the environment that makes them proud and eager to implement environmental friendly practices even if they do not have an economic return. They believe in general that the regulations and mostly the participatory approach have been correct. Also, many of them

PAGE 79

57 appreciate the fact that the program is flexible and can be discussed with the regulatory agencies. One farmer mentioned that they are trying not only to comply with the current regulations, but are proactively seeking to exceed these regulations. This producer warns that he is not necessarily looking for a new set of rules, but rather embracing the regulations to become better stewards. This farmer predicts that FLDEP will be willing to face the public for a farmer if the farmer is on the right side at all times. This fact would be much more important in the future because regulations will become much stricter, not only in the dairy farm industry, but in all aspects of human life. This farmer feels farmers will be better prepared for those times thanks to the Partnerships efforts. 3.3.3 Perceptions of Climatic Technology Innovation Recent advances in atmospheric research have made it possible to predict climate variability with a relatively high level of probability. Knowledge of climatic conditions may allow improved seasonal management strategies for dairy production in Florida. Once again the opinions were divided. Of the 21 farmers interviewed, 14 (67%) mentioned that they would use improved forecast information, 5 (23%) would not use this technology, and 2 (10%) were not sure (Figure 3-2). 3.3.3.1 Farmers that would use long-term climatic forecast All 14 farmers who believe they would use improved long-term climate variation forecasts say that their main use of the information would be in planning crop systems. One producer mentioned he already uses ENSO phase information to plan the farms crops: I make decisions every year based on my beliefs of if the seasons will be dryer/wetter, cooler/hotter. For example, for winter, if I expect an El Nio, I will plant oats because they do well in moist conditions. If it will

PAGE 80

58 be a La Nia, I will choose rye because it handles drier conditions much better. Of course, I will always plant ryegrass and clover, no matter what the climate prediction This farmer thinks that forecasts would also be useful for adjusting planting dates in order to obtain better yields and more N uptake. He has yet to implement this management strategy. Another farmer also uses long-term forecasts but not based on ENSO phases. He relies instead on the experiential knowledge of an old, experienced person who predicts the up-coming seasons in terms of wetter/dryer and colder/warmer. This in turn determines the crops to be planted. This farmer was very interested in the ENSO forecast information and tools and inquired as to its availability. 14 % Do n t 100 % 36 % Climate Technology Would use Would not use 67 % 23 % Crop systems Livestoc k Non-flexible and/or satisfied with current system Other more important things to worry about Different crops/varieties Planting dates Area planted Drainage system Irrigation system Milking cow group rotation Breeding date planning Solids collection Not sure N utrient mg t Would like to have the forecasts Ag r 10 % Figure 3-2. Farmers perceptions of climate technology adoption In the same manner as the farmer who already uses ENSO forecasts, most of the farmers believe that the use of the forecast would be in crops. Mainly, the producers would use information for substituting crops in specific seasons and to change planting dates. For the winter season, oats and rye were mentioned several times as

PAGE 81

59 interchangeable crops. Most agreed that drier years (La Nia) are better for rye, and wetter years (El Nio) better for oats and/or ryegrass. However, some other producers mentioned that temperature is the most important variable, and that hotter winters with less chance of freezes (La Nia) would be better suited for oats. For the case of freezes, some farmers agree that if the prediction could address those events, they would be very useful because that would let them know how much crop production to expect in winter. Based on this they would adjust crop plans. For example, if there were going to be more than the typical 20-25 freeze days, winter forages would be affected. Therefore, they might plant more area in order to produce the required amounts of feed. Climate forecasts for the spring/summer season would be more beneficial for planning the planting date for corn. Those who plant corn say that the earlier you plant, the better the results. If farmers know there will be little risk of late freezes, they could plant as early as possible, for example in the first week of March. For the summer/fall season another important decision would be between planting a second crop of corn or sorghum. Farmers mentioned they would plant a sorghum crop if they know that drier than normal conditions are expected. Potential changes in choice of crops refer not only to different species, but also to cultivars within species. Some varieties have different responses to climate conditions. Farmers mentioned these possibilities without entering into specifics. Another group of informants mentioned an additional possible use of climate with respect to crops. One farmer mentioned they could use the prediction information to alter the area of planted crops. In wetter conditions area would increase. In drier seasons, less area would be sown. Another farmer mentioned that climate data would be useful for

PAGE 82

60 planning drainage systems, especially before a wetter summer. A further potential use mentioned was in planning irrigation systems management. Although these systems are mostly fixed, farmers could make some adjustments if they know what the seasonal climate would be like, especially regarding rainfall. Besides the use of forecasts for crops, another potential use of climate forecasts would be for the livestock. Three out of the fourteen interviewees who would use the forecast elaborated on this issue. One farmer mentioned that she could plan her cow rotations differently. Holstein cows are very sensitive to heat stress calculated by an index that takes temperature and humidity into account. When the index is high, in hot and humid months, milk production is affected negatively. If seasonal variability toward hot, humid conditions is known ahead of time they would be able to better plan the development of cow groups. This farm, like others, has comfort facilities where milking cows are protected from heat stress, that is, climate effect is diminished. However, on almost all farms, these milking cows spend part of their time outside. The farmer found it logical to think of planning ahead of time to provide the most protection time to the most productive group of cows. One more decision that could be managed in light of climate forecasts is breeding time. This would be done by changing the number of cows entering the milking group in a certain season according to the forecast. For example, if a mild summer were expected, then the farmer would breed the cows in time to have more milking cows than normal in production for that summer. Climate knowledge would allow higher than normal amounts of milk to be produced in summer. Similarly, decisions regarding purchases of replacements heifers could be improved based on expected climatic conditions.

PAGE 83

61 Typically, farmers buy replacements in fall. If spring or summer were going to be colder, they could take advantage to buy less expensive replacements before those seasons. An additional possible use of forecast mentioned for livestock was for the winter season. Specifically climate forecasts would be used to anticipate colder than normal winter seasons, which are also detrimental for the cows. In those situations, farmers could plan ahead of time to have more forage (either planted or purchased) available. Extra amounts of fiber would help cows stay warmer. Another two farmers mentioned the potential use of the forecasts for planning feedstocks. The inventory of feed is a major management task on dairy farms. Therefore, knowing the climatic conditions ahead of time would be important to preparing feedstocks (either produced or purchased). For example, one farmer mentioned that if he would know ahead of time that the next season is going to be dryer, he would plan to have more forage for his heifers and dry cows. The farmer mentioned additionally that not only would it be good to plan for feed shortages, but also to plan for excess feed produced. Too much feed on hand is also a problem. If the predicted conditions for next seasons are good for forage production, then producers could plan not to buy feed. They could also sell feed in advance, or decrease acreage planted. Additional to the use of forecast for crops and/or livestock, three farmers mentioned that the information could be used for improving N and P management on the farm. Waste management systems are mostly fixed, so considerable changes to this system is not an option. Farmers usually change daily management based more on daily weather than long-term climate, however there are some things that could be planned, as is the case of for example, of collecting or not collecting solids from the waste

PAGE 84

62 management system. These farmers mentioned that if they know that there will be wetter conditions, they would prefer to prepare the whole system to not collect solids because the pile of solids resulting from the collection would be difficult to handle under wetter conditions. This major change would determine that the effluent will have more solids in solution. Therefore, the irrigation must have higher pressure and the crops must be prepared to receive a slightly higher percent of N in the effluent. If they know that the conditions would be dryer than usual they would prepare to collect all the possible solids from the waste management system and together with this action they would plan what to do with these solids, apply to fields of the farm or export out of farm. 3.3.3.2 Farmers who would not use long-term climatic forecast Farmers that mention they would not use climate forecasts as a technology to help in decision-making argue that they have relatively fixed systems. In their rather tightlybound systems, no changes in management need be planned. One farmer mentioned: I am married to my system operation it works and I dont want to change anything that could mess things up. These farmers said that their decisions are based more on the short term. If they knew there was no rainfall predicted in the following week, they would cut hay. It is difficult for them to envision how they might change tasks months in advance. One producer noted: unless something as disastrous as a hurricane is forecast, I would not do anything different. Another farmer argued: we have much larger problems to worry about. When we solve those other problems we might start thinking about incorporating climate forecasts in our planning. Another farmer was very convinced that: We are not going to change anything even if we know that a very abnormal season is coming. Even then, we would continue to do as we

PAGE 85

63 always have, that is, adapt to the weather conditions, making changes to respond to the weather conditions Other farmers may share this farmers skeptical view of the technology. Those who are naturally averse to change or have not been exposed to new technologies will exist. 3.3.3.3 Farmers who are not sure they would use long-term climatic forecast The two farmers in this category had never previously thought about climate-related management changes. They mentioned that the entire concept was completely new to them, so they would need time to digest the idea. Initially, however, they were not sure if they could incorporate climate variability forecasts into their decision-making process. Both these farmers mentioned that, even though they did not know if they would use the technology, they would like to be aware of the forecast to consider it in regard to their performance in the following seasons. 3.4 Conclusions and Recommendations All stakeholders included in the Sondeo had a general consensus that growing N levels in the Suwannee River Basin is a real and current problem. However, stakeholders included in the focus groups, and especially dairy farmers challenged this fact during the personal interviews. People interviewed during the Sondeo showed more awareness of the real consequences of increased levels of N in water bodies. Dairy farmers and other stakeholders involved in the interviews and focus groups, even though they recognized that the presence of high levels of N is detrimental for the environment, showed a certain degree of skepticism regarding the gravity of the problem.

PAGE 86

64 Every effort needs to be made to communicate more clearly to the general public real trends of N in water bodies and their potential consequences. The message must be conveyed using consistent scientific data. Most of the informants perceived that dairy farms are the highest polluters. However, many stakeholders included in the Sondeo, focus groups, and interviews questioned this popular belief. Dairy farmers in particular, perceive that there is a misconception of the dairy farm industry and its environmental impact. Further research should be devoted to investigate the contribution to contamination from each potential source. It is important that individual operations be held accountable. Point vs. non point N source pollutions must be further studied. Analyses of long-term monitoring studies comparing different industries and their environmental impacts must be undertaken. Farmers are willing to participate in incentive-based programs such as those promoted by the Suwannee River Partnership. The principal driver behind this willingness is a desire to avoid being regulated by official agencies. The Suwannee River Partnership has a very positive perception and support from the stakeholders. Most of the stakeholders perceive that the promoted BMPs are beneficial to the environment but are not cost effective to the farms. Some farmers believe that the BMPs could be re-engineered to better serve their purposes according to specific farm conditions. Further research and validation of BMPs must be conducted. A guideline describing the steps to develop BMPs should be created, and an economic feasibility study on BMPs that takes heterogeneity of farms into account should also be conducted. Extra funds should be devoted to expanding the impacts of the Partnership.

PAGE 87

65 Farmers feel that environmental regulations are becoming stricter and more uncertain in time. These regulations represent one of the biggest challenges for the future. Farmers want to express that they are doing their best to protect the environment. A vast majority of them agree with current regulations but they have serious observations to make regarding the technical and administrative procedures. The Partnership is helping to change farmers attitudes, but still more communication and outreach work is needed to gain the complete understanding from farmers. Two-thirds of the interviewed farmers believe they can use climate variability forecast proactively. All of them would use it primarily for crop management. Additionally some could use it for livestock management. Some could use the seasonal climate variability information for feedstock management, and some could apply this technology to the management of the waste system. Further research calls for the inclusion of climate variability forecasts into existing BMPs experiments. Studies that explore the interaction between climate and nutrient overflow in dairy farm systems must be promoted. Farmers should be involved from the early stages of these research processes. The main barriers for potential adoption of climate forecast technology by farmers are the rigidity of dairy farms and the lack of management resources. Agencies interested in delivering climate forecasts as an innovation must identify the narrow niches where they may be used based on the experience of others. The technology should be translated into clear and rapid management adjustments that farmers could apply without involving excessive administrative transaction.

PAGE 88

66 The fact that some stakeholders have never before thought of the potential application of climate forecasts denotes a lack of awareness of available forecast innovations among the stakeholder community. Step one in any diffusion process, i.e., awareness of the existence of a given technology, is missing among dairy farmers in the study area. Commitment to extension and outreach programs that make this information available to all networks of stakeholders and especially to dairy farmers is essential to the future success of incorporating climate-based management into overall schemes to reduce N loading into the environment. The participative process described in this study underscores the critical need to include farmers and other major stakeholders in every stage in order to assess real perceptions that further will be used for consensual problem-solving. Its relevance is even higher when complex problems exist as is the case of the interactions of N pollution, its regulations, and the potential use of climate forecasts as a way of mitigate them.

PAGE 89

CHAPTER 4 A METHOD FOR BUILDING CLIMATE VARIABILITY INTO FARM MODELS: PARTICIPATORY MODELING OF NORTH FLORIDA DAIRY FARM SYSTEMS 4.1 Introduction Much work has been published on the mathematical description of models and model outputs. Conversely, little effort has been invested to date to describe a process of stakeholder interaction or participatory modeling that might be essential for increasing adoptability of models among end users. Models are often created to simulate real conditions, predict potential outcomes, and ultimately be a tool used by stakeholders. Many models have failed in this last requirement because they do not represent the conditions relevant to stakeholders nor do they comply with requirements of the final users for simplicity and user-friendliness. Substantial changes must be incorporated into all phases of model development to improve end-user adoption. One important aspect required is intensive and effective participation of all stakeholders (Stoorvogel et al., 2004). This study highlights the use of stakeholder interaction methodologiesoriginally developed for use in international rural developmentto first gauge the need for, then refine these management products and tools. Working with stakeholders in a learning process involving as many feedback loops as necessary will increase the probability of higher adoption rates. The traditional system analytic approach emphasizes the importance of quantitative modeling, rational planning, and economic analysis. This is closely linked with operational research that looks for optimal solutions for complex problems (Rosenhead, 67

PAGE 90

68 1989). However, the application of traditional system analytical methods to complex societal and policy problems has not always been helpful because problems of validation and because the absence of adequate social scientific theories and empirical data (Stoorvogel et al., 2004; Jones et al., 1997). One efficient way to elicit up-to-date and high quality empirical data is through stakeholder interaction (Geurtz and Joldersma, 2001). During the 1980s, systems analysts realized that much of the understanding of the problem is generated in the process of model building; consequently, participation of the stakeholders in the model building process is useful (Geurtz and Joldersma, 2001). Therefore, the participation of the potential final users should enable the researcher to enrich the model by including subjective sources of knowledge in addition to the objective knowledge derived from theories and empirical studies (Geurtz and Vennix, 1989; Verburgh, 1994). Systems approaches at the farm scale must include not only the biophysical component, but also the social, economic, political environment of the farm together with a bottom-up approach (Jones et al., 1997). Development in methods of modeling analysis and in practice shows a shift from a uni-central, analytic, scientific approach to a more multi-central, interactive stakeholder approach. These shifts are accompanied by a growing interest in participatory styles of policy making and participatory methods of policy analysis. The application of the participatory methods generally follows the paths of logic-in-use, which shows the need for a methodology for participatory policy analysis (Geurtz and Joldersma, 2001).

PAGE 91

69 Participatory Rural Appraisal (PRA) describes a growing family of approaches and methods to enable local people to share, enhance and analyze their knowledge of life and conditions, to plan and to act. PRA has sources in activist participatory research, agro-ecosystem analysis, applied anthropology, field research on farming systems, and rapid rural appraisal (RRA). Natural resource management (NRM) and agriculture are fields in which PRA finds applications (Chambers, 1994). Farming systems research (FSR) advocates for interdisciplinary/participatory approaches to deal with farm scale simulations (Norman and Collison, 1985). Nevertheless, some authors have pointed out that FSR is often location specific and descriptive (Dent and McGregor, 1994; Jones et al., 1997). An alternative approach is to study the complete regional production system in close interaction with farmers. Vereijken (1997) developed prototyping as an alternative to the more model-oriented and farming system approaches. Prototyping procedures have a participatory character of scientists directly collaborating with farmers. Simulation-aided discussion about crop management has underpinned advances in farming systems analysis as a vehicle for management intervention (Keating and McCown, 2001). Similarly, McCown (2001) and Hammer et al. (2000) both argue that connectivity and dialogue among key players is essential for achieving relevant and significant intervention. In this sense, the dialogue around the simulation process could be as important as the underpinning models, although it is clearly reliant on their existence and influenced by their reliability (Nelson et al., 2002). Meinke et al. (2000); Robertson et al. (2000); Keating and McCown (2001); and Meinke et al. (2001) provide examples of participatory approaches using simulations to drive relevant and significant

PAGE 92

70 results. These approaches are not about science simply providing the answers for management to practitioners, but rather employing cooperative learning to develop solutions (Nelson et al., 2002). Direct participatory action research with farmers has helped establish credibility of models and simulation analyses (Meinke et al., 2001). According to Archer et al. (2002) if the model simulation software is created for farmers, extension agents, and/or farm advisers, it should be designed to be user-friendly and to require minimum data inputs. Also, Archer et al. (2002) argue that it is not only important that the model be able to predict what is required, but also to make the predictions quickly and easily, in a readily accessible form, for site-specific conditions, in real time. Both the participatory and the user-friendly approach were used as the methodological basis for developing a climate-influenced dairy farm model for North-Florida. The benefits of involving farmers in research to encourage engagement and learning for scientists and farmers are well documented (Paine et al., 2002; Onema et al., 2001). North-Floridas climate is highly affected by El Nio Southern Oscillation (ENSO) (OBrien, 1999). Since agriculture is the activity most vulnerable to climate variability (Hansen, 2002b), farmers might benefit from new methods of seasonal climate forecasts connected to system simulation models to tailor management options regarding agricultural, economical, and/or environmental outputs. Intensive technologies offer a broad range of options for adaptive responses to climate information (Hammer et al., 2000). Dairy production in sub-tropical north-Florida is a very intensive system. The Southeastern Climate Consortium (formerly Florida Consortium) was created as an interdisciplinary, multi-institutional partnership involving researchers from 6 universities

PAGE 93

71 in the states of Florida, Georgia and Alabama. The research group found that it is critical to involve stakeholders actively from the beginning of a process to understand the effects of climate variability on water and agriculture. Agencies with strong established relationships and trust with end-users must be engaged to improve the chances of delivering operational applications of climate information to final users (Jagtap et al., 2002). One such application to emerge is the dairy model whose development process is the subject of this paper. It is not the aim of this paper to describe the mathematics of a dairy farm model. The main aim is to describe the process of collaborative creation of a regional north-Florida dairy farm model adaptable to any individual farm. The purpose of the model is to tailor management strategies for different climatic conditions by estimating monthly rates of N leaching and profit. An overriding objective of the model as expressed by stakeholders was to decrease environmental impacts while maintaining profitability. Specifically, the study sought to engage stakeholders in the process of creating, calibrating, and validating a whole north-Florida dairy farm model (Figure 4-1). 4.2 Materials and Methods The study was undertaken in north-Floridas Suwannee River Basin, covering part or all of Alachua, Gilchrist, Levy, Lafayette, and Suwannee counties. An adaptation of the emerging approach in Farming Systems Research and Extension presented by Bastidas, 2001 (Figure 4-2) was used. Interviews and focus group meetings were carried out on a continuous basis, divided into five phases. Phase One consisted of the initial developments. Phase Two was about the sounding out or rural appraisal. Phase Three consisted of greater stakeholder involvement. Phase Four included meetings and interviews to elicit feedback from stakeholders on the prototype model. And in phase

PAGE 94

72 Five, the tentative final model was taken to stakeholders for the validation, their final suggestions, and recommendations. PrototypeModel Farmers Regulatory Agents Extension Agents University Faculty Final Model Researchers Figure 4-1. Schematic of participatory modeling 4.2.1 Main Framework In Figure 4-2, all stakeholders involved in the issue of N pollution in groundwater by dairy farms are referred to as users. Users must be identified and their differences recognized. Diverse users have specific and diverse needs. These needs must be identified and considered before starting the modeling/simulation process. PRA methods are ideal for initiating and maintaining high stakeholder involvement. This can be characterized as a continuous learning process. In Figure 4-2, arrows among users, needs, and PRA indicate the iterative and dynamic process among researchers and stakeholders to determine felt needs that must be incorporated in the modeling. A following step includes the integration, creation, and improvement of models. Once again, the arrows indicate the iterative and dynamic

PAGE 95

73 researcher-stakeholder process to improve models in order to make them more amenable to final users. Inclusion of prototype models during this time is useful. The process in which adjustments to models are made in order to better represent reality is called calibration. Because it is a continuous process, calibration moves seamlessly into validation when fewer and fewer major changes are required. Validation is the part of the process that is undertaken after researchers and stakeholders agree the model fits real conditions adequately. At this point testing models against independent data, i.e., data from stakeholders who had not participated in the creation of the original models is used. Figure 4-2. Methodology of participatory modeling. Source: Adapted from Bastidas (2001), p. 19. After final adjustments and validation, a final product flexible enough to fit the diversity of all users is the goal of the interactive process. Because it has been a

PAGE 96

74 participatory product, higher than normal adoption is expected along with higher than ordinary usefulness. 4.2.2 The Sondeo A rapid rural appraisal or Sondeo (Hildebrand, 1981) was conducted between the months of January and February 2003 by a multidisciplinary team of eight graduate students from the University of Florida (Alvira et al., 2003). During the Sondeo 13 stakeholders were interviewed. This sounding out was undertaken to get a better understanding of the broad issue of N pollution in water. Another objective was to understand the dairy farm systems in the Suwannee River Basin and initially identify the type of modeling products that might be helpful for proactive management in light of climate variability forecasts. The Sondeo team consisted of graduate students from a number of disciplines including agronomy, animal science, agricultural extension, natural resource management, forestry, and community development. Two to three-people teams conducted transects and stakeholders interviews. The interviews were conducted as informal, open-ended conversations. After a team introduced itself and explained the subject of the research, the stakeholder and the team conversed about any aspects of the topic that were relevant. This open-ended approach enabled topics to emerge and be pursued that might have been missed if the researchers had been constrained by a pre-formulated questionnaire. Despite the open-endedness, a set of core themes was discussed in almost all interviews. Notes were not taken. Following each interview the partners wrote individual notes, compared them, and then wrote a joint report of the interview. Finally, these joint reports were shared and thoroughly discussed among all Sondeo members. Each team presented its findings, which could be clarified, challenged

PAGE 97

75 or contrasted with the results of the other teams. This process of reporting and discussion served as the opportunity to begin noting trends, gaps in information and new questions to be pursued. 4.2.3 Interviews Personal interviews were undertaken by the researcher with dairy farmers and/or dairy farm managers during the fall/winter seasons of the years 2003-2004. Open-ended interviews (Bernard, 1994) were undertaken with a purposeful sample of 24 farmers (~35% of population) with the main aim of engaging these actors. Specific operation information was elicited and obtained as feedback in the modeling process. The sample size and the people to be interviewed were chosen in order to obtain as much of the existing variability as possible. Selection of participants was done in a participatory manner with the coordination of the Suwannee River Partnership (SRP) and the University of Florida Extension offices in the area. Interviews started with the presentation of either the prototype or the final model on a laptop computer. A voice recorder was used in every interview after requesting permission. The first step was to explain the purpose of the visit and then to provide a general description of the model. This was followed by a detailed explanation of every component in the model, including those that were as yet not active, and their mechanics. After discussing similarities and differences between the model and the farmers specific production units, a demonstration of the running of the model was presented and discussed. Later, dairymen had the opportunity to input their specific data and run the model for their own specific conditions. Similarities and differences with what farmers expected were recorded and taken into account. If observations were easy to adjust for, they were

PAGE 98

76 tried on-the-fly, taking advantage of the user-friendliness of the models interface. If specific comments would require more work, they were incorporated in the lab and tested in later interviews. Interviews highlighted farmers reactions to the models. All possible suggestions to be incorporated in the models were taken into account in successive steps. Two rounds of interviews were conducted. Twenty-one farmers observed the prototype model in October-November of 2003. In February-March 2004, the definitive model was used for validation purposes with 3 farmers. 4.2.4 Focus Groups Eight focus groups (Bernard 1994; Stewart and Shamdasani, 1990) with stakeholders were held between November 2002 and February 2004 in the study area. Focus groups consisted of open discussions among 5 to 15 persons. A stakeholder in coordination with the researcher called these meetings. The researcher facilitated the meetings around the different topics of research. The first two focus groups did not include any model as a basis for discussion. Rather, because the creation of a prototype model was suggested and encouraged during the second focus group, one was created and became the focal point for discussion in all groups save the last. In the last focus group, the final model was discussed. Focus group participants ranged from specialists from official agencies such as the Suwannee River Partnership (SRP, www.mysuwanneeriver.com/features/suwannee+ river+partnership), the Suwannee River Water Management District (SRWMD, http://www.srwmd.state.fl.us), the Florida Department of Environmental Protection (FLDEP, http://www.dep.state.fl.us), the United States Department of Agriculture (USDA, http://www.usda.gov) Natural Resources and Conservation Service (NRCS,

PAGE 99

77 http://www.fl.nrcs.usda.gov), the University of Florida IFAS Cooperative Extension Service (CES, extension.ifas.ufl.edu), to private sector companies, technicians and farmers. These participants were usually called to a meeting by the Suwannee River Partnership and/or by extension agents of different counties in the research area. Focus groups lasted between two and three hours and their main purpose was to listen to feedback on the modeling process and to actively increase involvement and diffusion of the research to a more target public. 4.2.5 Other Stakeholder Interactions The principal researcher approached many other stakeholders between fall 2002 and spring 2004 in order to receive additional inputs into the modeling process. Initially, the discussion centered on the intended idea of a proposed simulation methodology. In a second phase, the discussion was based on the prototype model. Finally, interaction with the clientele was based on the final model and its final tune-ups. These interactions were meetings with specialists in the issue of the N leaching and the dairy industry and included private consultants, officials from regulatory agencies, and university faculty, among many others. 4.2.6 Analysis of Information All participatory interactions were recorded, classified, organized, analyzed, and described. The feedback received was included in the modeling process in one way or another in successive steps. This paper summarizes the results of all the above interactions in a descriptive/narrative manner. Names of the participants and some organizations and companies are not revealed to protect the anonymity of interviewees. A summary of stakeholder interactions can be found in Table 4-1.

PAGE 100

78 4.3 Results and Discussion Results of interactions are described in five phases that detail the iterative process, step by step, of creating, calibrating, and validating the dynamic north-Florida dairy farm model, DNFDFM with participation of stakeholders. This process is summarized in Figure 4-3. Figure 4-3. Phases of the participatory modeling for creating the DNFDFM 4.3.1 Phase One: Initial Developments In July 2002, four well-known scientists in Florida, each from a different discipline, were approached looking for specific information and at the same time ideas about the original research idea. The scientists were from the areas of forages on dairy farms, dairy waste management, soils, and crop modeling, respectively. During this first exploration, confidence was gained in the fact that most of the soil series in Florida are contained in a database that could be accessible to use in later phases of the research (G. Kidder, pers.

PAGE 101

79 comm.). Also, the current system of crop models contained in the Decision Support System for Agricultural Transfer (DSSAT) (Jones et al., 2003) would be ideal to use in the research for the estimation of N leaching according to climate variability, after major adjustments. Furthermore, it was known of the existence of a growing body of research related to environmental issues and dairy farms in Florida and north-Florida that the study should take advantage of (annual reports on water quality from the SRWMD, basins status report from DEP, water resources investigation series from the USGS, water quality assessments from the EPA, studies on natural resources management from the NRCS, among others). Above all, important contact information with end users and other stakeholders was gained that would be needed in the following phases. In August 2002, a telephone conversation was held with a dairy farm manager to get an initial grasp of the N issue on dairy farms and how it is potentially affected by climatic variability. The conversation was enlightening in capturing the sensitivity of dairy farmers to the issue of N pollution and the need for alternative ways to estimate environmental accountability of dairy farms. This manager set forth the notion that climate variability would have little effect on livestock performance. However climatic variability was thought to have a very high impact on forage field yields. In September 2002, 3 contacts were made with soil scientists working in the study area in a project regarding N pollution from farm activities, including dairy farms (Florida Department of Environmental Protection, 2004). These scientists provided further contacts and references to add to the research. Additionally, suggestions from the contacts led to one-day field visits and participation in the collection of soil and water samples was achieved through the contacts suggestions. This last task was very

PAGE 102

80 important to help understand the methodology being used in current experiments and the implications that results might have on the modeling process. During September and October, stakeholders in the research area were contacted for the first time to get them acquainted with the research project and, most of all, to get them involved in the project. Stakeholders whose farms were located in the heart of the research area, the middle Suwannee River Basin, were very receptive to the idea because environmental degradation from dairy farms is an issue that is very much on the agenda of official agencies. Contacts included members of the SRP and CES agents. During the first contact a verbal idea together with a small paragraph of intention was handed to the stakeholders. Discussions were based on an initial idea. While support for the main idea was expressed, a fuller understanding of the research was still lacking. Nevertheless, the general reaction on the part of stakeholders was let us begin and involve more stakeholders so we can all better understand the project and define further steps. This was the beginning of a very rich feedback loop process with stakeholders. In November 2002 the first focus group was held in the Suwannee River Water Management District (SRWMD) at Live Oak. The SRP called the meeting officially and 6 people attended. Attendees included members of the SRWMD, the SRP, and the CES. This meeting was critical to understand the interests that all sides had in the project and where the project was heading. It was agreed that there were commonalities between the proposed work and current work already underway within the Partnership. Interest existed in supporting the research and benefiting from the outputs. Specifically, the group committed to contacting farmers and convincing them to participate in the research. It was agreed that a purposeful sample of 20 farms (out of 64 in the area) would be enough

PAGE 103

81 to get acquainted with the diversity present in these systems. The stakeholders all agreed that the researcher should prepare a short document explaining the potential outcomes of the research together with a list of topics to be included in interviews. Topics could be handed to farmers ahead of time in order to gain insights even before the interviews were conducted. A time limit of one hour for interviews was established because dairy farmers are really busy people and they dont have much time. The protection of the anonymity of participants was set forth as a must because the nitrate issue is very sensitive. Neither farmers nor their management practices regarding estimations of waste should be identified in the work. Also in November 2002, a visit to a dairy farm and an informal interview with its owner was conducted to gain general insights of the dairy farm systems, the recycling of nutrients, and the potential to benefit from climatic forecasts. This visit was very instructive to understand activities that are dynamically happening in the system and that can be represented by simulation models. The owner highlighted the fact that the farm was following environmentally friendly practices suggested by regulatory agencies. 4.3.2 Phase Two: The Sondeo The very participatory and interdisciplinary approach revealed many characteristics that needed to be incorporated into the participatory modeling process. Background information indicated that ten years ago an increase in nitrate concentration was recognized by the SRWMD in the middle Suwannee River Basin (Hornsby et al., 2002a). Sondeo results found that it was believed that if the problem were not solved, there would be a time when the government would impose regulations. Regulations were perceived as one of the main drivers for farmers to uptake Best Management Practices (BMPs) (Katz and Bohlke, 2000). If regulatory measures had been imposed at that time, some farmers

PAGE 104

82 would not have been able to comply. This could have forced them out of business as happened earlier with some dairy farmers in the Lake Okeechobee area of southern Florida (Alvira et al., 2003). The SRP was formed to determine the sources of N loads in the basin, and to work with local users to minimize future nutrient loading through voluntary, incentive-based programs. The Partnership works under a non-regulatory, preventative approach, through monitoring pollution, education, promotion, and adoption of BMPs (FLDEP, 2001). Informants in the Sondeo generally recognized that the problem of the increase in nitrate concentration in the middle Suwannee River Basin was very evident and that dairy farming was perceived as the most contaminating activity. Even though farmers were aware of the presence of nutrient contamination, they had differing views on the primary source of contamination (point versus non-point source pollution) and potential effects on human health (Alvira et al., 2003). The SRP group is focusing on finding the most economically and technologically feasible management techniques (BMPs) available to help farmers and other land users to satisfy regulatory requirements for protecting public health and the environment. These management practices mainly include nutrient and manure management, irrigation water, technical design and engineering aspects in dairy farming (Suwannee River Partnership, 2004b). By large, the concerned stakeholders see BMPs as research-generated solutions to reduce potential environmental hazards from farm enterprises. From the simulation point of view, modeling would be in essence an opportunity to help these stakeholders to experimentally test BMPs without disturbing the actual dairy farms. Simulation could reproduce farm conditions and test the environmental and economic impacts of intervention, without risking the normal functioning of a farm. Outputs of the

PAGE 105

83 simulations could be used to verify technical viability of BMPs before their implementation by state, federal, and private agencies. Adopting BMPs may help farmers to remain in business, which is critical to stakeholders. However, initial investment in BMPs, even though costs could be shared with official agencies, can be cumbersome (DEP, 2001; Ribaudo et al., 2003). Therefore, it is important that the proposed practices to protect the environment be cost effective taking into account that dairy farmers first will look for survival strategies and then for good stewardship, just as any other enterprise. This last statement would have great implications in the modeling process since the economic outputs of the whole farm would be vital in determining which practices a north-Florida dairy could implement as a technological choice. It was commonly agreed by participants in the Sondeo (Alvira et al., 2003) that the model must have an important economic component and this component must be taken into consideration for potential proposed practices. The Sondeo team had several interesting findings. It concluded that Perceived benefits exist in implementing BMPs However no data exist to support the fact that BMPs are economically and ecologically viable in the long run for farmers Further research and validation of BMPs needs to be conducted A guideline describing the steps to developing BMPs is needed An economic feasibility study on BMPs that takes heterogeneity of farm sizes and types into account also merits research Further research should be devoted to investigate the contribution to contamination from each potential source or to investigate the point-source pollution sources that make each party accountable for their contribution to the overall problem.

PAGE 106

84 The powerful implications mentioned in the above paragraph definitively implied the development of the modeling work in successive steps. First, BMPs require validation. The model or the models would help in this task. Second, the simulation has to be performed in a way that incorporates diversity of farms and allows for individual farm analysis as complete and separate units. 4.3.3 Phase Three: Greater Stakeholder Involvement In February 2003, 3 dairy farms in the study area were visited. From these visits the components of a dairy system in north-Florida to be used in the model took shape. All dairies have a livestock component, forage fields, feed facilities, and some means of disposing of waste produced by confined cows. Some differences exist on specifics, e.g., some farms do not raise heifers or some farms raise more forage crops than others. Nevertheless, it was believed that common components could be found in order to represent all north-Florida diaries in one general model that could be adjusted for any peculiarities to represent specific farms. Farmers noted the only thing dairy farmers have in common is that they all produce milk other than that they can be very different. Also in February 2003, a second focus group with eight stakeholders was held. This meeting was very powerful for the next advances in the research. First, a deeper discussion on the climate component was achieved. These key stakeholders felt that uncertainty and low accuracy in climate (or seasonal) variability forecasts could create problems in the models. It was proposed that a way to handle uncertainty and low skill forecasts was to present results in a probabilistic way. However, a consensus of farmers was that they wanted to have outputs available in as simple a form as possible. It was agreed by all that deterministic results would be the default output but an option to also

PAGE 107

85 see results under other climatic scenarios would be created. If for example an El Nio prediction exists, the model should present the results for an El Nio year. Additionally, an option to present results in case it is not an El Nio year but La Nia or Neutral is necessary. It was argued that not all El Nio years are the same in intensity so there should be a distribution of results instead of just one outcome. Farmers counter-argued to be very careful not to over-complicate the already existing uncertainty. They warned that if it happened, farmers would lose interest. It was during this discussion that that a producer proposed that a probabilistic distribution should instead be applied to milk prices in the economic portion of the model. Farmers clearly understood that there is no true factor that explains price variations and they are in fact, stochastic. Another important point of this second focus group was that the participants better understood the research idea and found similarities between it and the nutrient budgeting approach that the Natural Resource and Conservation Service (NRCS) is using for planning environmental accountability (NRCS, 2001). The NRCS was invited to become involved with the research project and also participate actively in order avoid duplicate efforts. Also in this meeting it was re-affirmed that user-friendliness of the model is extremely important. It is also very important to stakeholders that the model(s) be easily adaptable to different farms because no two dairy farms are the same. Another important decision was taken during this meeting. A stakeholder, supported by all others including the researcher, suggested that a prototype model be created to clearly show what it was intended to accomplish. The idea was to have a prototype model to use as a tool for increasing farmer and other stakeholder involvement.

PAGE 108

86 The model should be constructed in a user-friendly manner so that it could be shaped and molded through further stakeholder interaction. Finally, farmers encouraged continuing with the work, since they were beginning to see how they could benefit from the final product. They mentioned that they could directly test BMPs on their system and that this would be useful to them. Large dairy farms that usually do not participate in cost-share programs might see the advantage of decreasing potential risk of fines from regulations. Additionally, dairy people were eager to see the economic impacts of any potential management changes. At the beginning of March 2003 a meeting with a private consulting company was held. The company worked with some dairies in north-Florida on the N issue. They prepared compliance reports, advised farms on how to decrease environmental concerns, and carried out experiments to gain more insights on the pollution issue. Company professionals quickly understood the aims of the research. The consulting firm expressed its interest in the outputs and was willing to share their expertise and information for the creation of the simulation models. Furthermore they provided other important contacts. Both companies suggested the research might benefit from the incorporation of a hydrological-soil model called GLEAMS that stands for Groundwater Loading Effects of Agricultural Management Systems (Leonard et al., 1987). This would facilitate the estimation of N leaching in the soil. This idea was later aborted because crop models would be used, however the consideration is worthwhile to mention. This meeting with the private consultants highlighted 3 other important points. These were the incorporation of the climate component, the economic module, and contact with another network of stakeholders. The consultants agreed that climatic

PAGE 109

87 variability would affect dairy farm outputs through its effect on crops. According to their experience, even when crops are irrigated, they respond to rainfall patterns and temperature. The incorporation of crop models would be extremely important for taking these responses into account. Private consultants also mentioned that if the research includes a strong economic module, it would be an important addition. Inclusion of this module might be sufficient to gain farmers attention and eventual adoption of the models. Finally it should be noted that the consultants regularly promote and participate in a network of stakeholders on the nutrient issue. An invitation to form part of these meetings was extended. On a monthly basis meetings are held and novel aspects of the nutrient problem are discussed. Also during March 2003, a third focus group was held, this time at the facilities of the NRCS-USDA in Gainesville, FL. A discussion based on the research proposal was opened and specialists were able to input their expertise. Three professionals very knowledgeable on the topic agreed that the research was relevant even though a similar product, The WATNUT software (NRCS, 2001) or Worksheets for Water Budget and Nutrient Management), already existed. They clearly understood that the research was going beyond the WATNUT application because it would have dynamic, economic, and user-friendly components that are not included in WATNUT. The USDA researchers expressed a willingness to work collaboratively in order to obtain a relevant final product. Researchers provided literature and the WATNUT v 2.0 software for use as a baseline on the modeling enterprise. At the end of March 2003, a second meeting with the private consulting company was held. During this meeting the company provided a CD with detailed information of

PAGE 110

88 several years of experiments on a dairy farm in the study area. The CD contained spreadsheets with records of experiments on waste handling and crop systems. This information is not published and was delivered under the consent of the dairy owner. All this information was a very useful reference for the further modeling process. 4.3.4 Phase Four: The Prototype Model In May 2003, a fourth focus group was held in Live Oak, FL with the participation of 6 people. The prototype Dynamic North-Florida Dairy Farm Model (DNFDFM) was presented for the first time. Specialists, extensionists, and technicians participating were impressed with the model performance, considering it was still a prototype. The following adjustments were suggested before the model should be presented to more dairy farmers: Improve the visuals, including more figures and visual aids that would attract farmers attention Use clear labels for the different components and if possible reduce the number of cells to make it easier for farmers to understand Slow down the run so it could easily be tracked by farmers Put money components in green and the N leaching in red to call farmers attention Include number of times cows are being milked in a day It was suggested that after the adjustments the prototype be shown to a list of suggested producers. They would provide rich feedback. In May the model and its documentation were discussed with a specialist at the University of Florida. The main suggestion was that it should be presented at a scientific meeting because the framework could be useful for many people and there could be very interesting feedback that could be incorporated in the final model. In this way, types of stakeholders involved in

PAGE 111

89 providing feedback to models would be expanded beyond farmers, extension agents, government agents and private consultants to the scientific community. In July 2003, the fifth focus group was held in Mayo. Seven people participated, including 4 farmers, 2 extension agents, and 1 technician. The prototype model was presented and explained. A number of unanticipated uses were suggested for the model such as to keep economic records and as an accounting tool. One farmer, especially interested in the work, asked to input his own data into the model to see how well it would perform. After running the model he took a hand-held calculator and started making calculations to compare with the model outputs. The farmer put forth several questions for general discussion. The farmer finally concluded that the model output numbers were reasonably good and that he would like to see more and would be glad to work directly in calibrating and validating the model. Many suggestions were collected at this important meeting. Among them: The model was estimating that the number of replacements should be between 60 and 70% of the adult cow total. The correct range should be between 50 and 60% The model was estimating that the number of freshening cows was about the same the year round, however in reality roughly 60% of births occurs between August and December, so that seasonality should be incorporated Farmers thought that prices received for milk were overestimated in the model since official prices were used. The model should use mailbox prices that farmers would be glad to provide Farmers liked the stochastic component in the prices, but asked if it would be possible to have the option for the user to choose between fixed and stochastic prices for the final model It was realized that not all the options of crop/forage sequences were included in the model. These should be in the final version It was also suggested that the confined time milking cows spend on concrete should be divided by seasons since this changes with conditions through the year

PAGE 112

90 The model required input of specific cow categories before starting. This process was overly time consuming. Farmers wanted to be able to obtain similar results with less input data The model raised all heifers on farm. Farmers suggested the option to not raise or only raise part of the heifers on farm Even though most of the cattle used in the area are Holstein, the model should incorporate an option to handle other breeds In August 2003, the sixth and in September 2003, the seventh focus groups were held at the county extension offices of Suwannee and Lafayette counties, respectively. At the Suwannee meeting a great deal about crop and forage rotations was learned and noted to take into account in the model. At the Lafayette meeting, a group of 5 farmers, a technician and, an extension agent participated. Once again feedback was elicited and collected. The main suggestions were There is a need to clearly include all options of crop/forage rotations in the study area. It was agreed that the stakeholders would make a comprehensive list of practices The prototype model included only sprayfields where effluent was applied, however dairy farms have many pastures where dry cows, heifers and milking cows (when not confined) graze and deposit manure directly and these should also be incorporated in the model The model should maintain an already existing option that lets the user input specific data regarding N content of the manure effluent at different points of the handling system In September 2003, a specialist in climate applications to reduce risk in agriculture from the Southeast Climate Consortium had the opportunity of seeing the prototype model and to input relevant ideas. Input regarding how to incorporate the climatic component using crop models from the DSSAT family (Decision Support System for Agricultural Transfer; Jones et al., 2003) was obtained. These function with daily weather data. In order to incorporate this component in the model, climate would

PAGE 113

91 be analyzed for a long time series (50 years) and results would be classified by ENSO phase (OBrien, 1999). Years would be divided into El Nio, La Nia, and Neutral year categories. An important contribution obtained was that crop models should be pre-run and organized in a database, then incorporated in the dairy farm model. Table 4-1. Stakeholder interactions, suggestions and changes made to model Date Interaction Format Participants Changes Suggested and Incorporated July 2002 Interaction 4 UF Scientists Use soil series data Use DSSAT August 2002 Interview 1 farm manager Climate important for crop yields September 2002 Interview 3 soil scientists Participate in sample collections September-October 2002 Interviews 3 farmers Involve other farmers November 2002 1 st Focus group 6 stakeholders Use 20-farm sample One-hour limit to interviews November 2003 Dairy farm visit 1 farmer Main components of a dairy farm Jan.-Feb. 2003 Sondeo 13 stakeholders Take heterogeneity of farm sizes and types into account. Use model to test economic feasibility of BMPs February 2003 Dairy farm visit 3 owners Diversity stressed February 2003 2 nd Focus group 8 stakeholders Deterministic outputs requested March 2003 Meeting with private consulting company 2 consultants Importance of crop models, economic module and climate component March 2003 3 rd Focus group 3 people WATNUT offered as data March 2003 Private consulting company 2 consultants Waste management and experimental data provided May 2003 4 th Focus group 6 stakeholders Improve visuals. Slow down runs. Use color. Follow herd flow July 2003 5 th Focus group 7 stakeholders Adjust replacement and freshening coefficients. Raise heifers off farm. Reduce prices. Use deterministic prices as well. Include more crop sequences August 2003 6 th Focus group 5 stakeholders Include more crop rotations September 2003 7 th Focus group 5 stakeholders Include all sprayfields. Need a place to input N data September 2003 Meeting 1 Climate specialist Refined use of DSSAT with daily weather data September 2003 Meeting 2 Dairy Specialists Importance of user friendliness October 2003-January 2004 Meetings Farmers Outputs in tables and graphs. All coefficients calibrated February 2004 8 th Focus Group 8 stakeholders Keep model in Excel. Keep model user friendly. Use Markov chain flow model Also in September 2003, a meeting with two dairy specialists was held to discuss the implications of the model and to gain more access with farmers in other counties. During this meeting it was suggested that a component similar to N should be incorporated for P in the final model. After discussion of the scope of the study and the importance of concentrating only on the current problem (N), it was agreed that the

PAGE 114

92 current model could provide a useful framework for P in future research. The fact that user-friendliness had to be maintained at all cost and that the final model would be an important tool for farmers was emphasized. The specialists offered to arrange contacts with stakeholders from other counties, mostly Alachua, Gilchrist, and Levy, all in the Suwannee River Basin. In the first round of interviews with farmers many feedback points already discussed during previous focus groups were repeated. Farmers encouraged continuing with the dynamic simulation in monthly steps instead of weekly or yearly. Also, user-friendliness was again highlighted in all. Farmers and extensionists were surprised that the model could be handled in a program that most of them frequently use, Excel Producers asked that the use of Excel be maintained because that would insure that they themselves would be able to use the model. Two interviewed mentioned that they would use the model as a complement to tools they already possess for their nutrient balance (Van Horn et al., 1998; NRCS, 2001). The opinions regarding the use of climatic information to change management strategies were split. About 67% of interviewees believed that they could proactively use climatic variation information for decreasing N leaching and/or increasing profitability. Farmers who favored the use of climate forecasts were inclined to include it as a user choice option. The interviews were very useful for sensitivity analysis and calibration of the model since all the different conditions observed in these varied farms were run in the model. In the same way, hard data such as number of cows, size of land, type of waste management systems, crop systems, etc. were recorded for future modeling use.

PAGE 115

93 4.3.5 Phase Five: The Final Model A new model was built taking in account all possible previous suggestions of stakeholders and the prototype experience. Four main components were new and all the rest were completely revisited. The new components were The Markov-chain probabilistic simulation of the cow flow The climate forecast interface The crop models, and The optimization module that helps in decision making These new components were included in response to needs expressed during the many meetings. With the incorporation of feedback from all stakeholders, the new model began to take shape. This did not end the participation process, rather, interaction continued as actively as during the first phases. The farmers suggested that a forage crop specialist from the University of Florida should be incorporated in the discussions. A meeting was held and the sequences of forages were completely determined in their final form to be included in the model. It included the list of sequences tillage options, the intercropping options, and the harvesting choices. Between September and November 2003 a new cow-flow model was created with help and feedback from a dairy science professor who was able to validate it for North-Florida conditions. Intense work on climate and crop models started in order to incorporate them into the main shield or interface of the north-Florida dairy farm model. Because the detailed nature intrinsic to this phase work, more specialized people were contacted for feedback at this time. Between October 2003 and January 2004, some 10 meetings were held and several contacts made by email. Even though during this time no model was presented, contacts were already aware on the project behind it and were able to collaborate. Input was provided not only for the specific crop models or climatic components, but also for

PAGE 116

94 the whole simulation approach. During these exchanges feedback loops were incorporated into the model and decisions were made regarding the pre-run simulations of crop models. These are closely linked to the climatic component since they use daily weather. At the end of January 2004, a specialist validated the main results of the crop models before they were incorporated into the main model. In February 2004, the eighth focus group was held at Live Oak with 8 stakeholders. The meeting was a discussion of the final Dynamic North-Florida Dairy Farm Model, which was presented in detail during an one-hour meeting. An one-and-half hour discussion followed. Attendants were encouraged to test the model to its limits in order to find any potential deficiencies that could be corrected. Farmers attempted to confound the model but fortunately the model responded very well. All appreciated the reduced need for initial data to start runs. Farmers also approved of the improved user-friendly interface, and the integration of models in a simple Excel application. An example of a hypothetical farm was run and results were compared using the optimizer. Stakeholders appreciated the possibilities to arrange management options in order to decrease N leaching without decreasing profit. At the end of the meeting farmers asked for and received a copy of the model. The ultimate validation of the DNFDFM was through interaction with 3 additional farmers. Producers were able to compare outputs from the model with their own farms. Farmers agreed that although the output numbers did not correspond exactly to the real data from their farms, the similarity of the outputs to real numbers gave them confidence that the model worked correctly. They expressed their willingness to adopt and use the model in their management strategies.

PAGE 117

95 4.4 Summary, Conclusions, and Recommendations Participatory modeling can enhance the creation of adoptable and adaptable user-friendly models that include environmental, economic and biophysical components. Multiple feedback loops were used during a two-year process of interaction with stakeholders. An iterative process involving creation of a prototype, interaction with stakeholders, improvement of prototype models according to suggestions at meetings was undertaken. Improvements included the incorporation of climate, crop, cow-flow, and optimization. Many adjustments were made to coefficients originally obtained from secondary data. Expert advice was incorporated from specialists in the government, private, extension, university and production sectors, following suggestions made by farmers during focus group meetings. The reality of using secondary information as a starting point and then calibrating and validating in the field has especially strong implications for adoption and long term use of the model. A feeling of ownership was created within the focus groups by listening attentively to all suggestions. An especially strong interaction among farmers and researcher occurred during the calibration and validation process. Models were run with default data and farmers compared outputs with their personal knowledge of farm outcomes. The researcher and farmers were then able to backtrack into the model to adjust coefficients that were deemed to be skewing results. Producers expressed reasonable ranges within which model parameters should be confined. Through this process each farmer felt included in contributing to the success of the final model. Climate is especially relevant as a problem considered by all members during many meetings. In the past, it has been difficult to get extension agents and producers to buy into the idea that management can be tailored in light of climate forecasts based on

PAGE 118

96 ENSO phenomena. By working together on the model and incorporating typical rainfall and temperature data for different ENSO phases, producers were able to understand the effects of seasonal variability on production and N leaching. Understanding the climate-cow-crop-nitrate interaction at the early stages of model development will greatly increase the probability of proactive use of climate information to mitigate negative environmental impacts. After undergoing an arduous but fruitful stakeholder interaction process to develop the North-Florida dairy farm model the following is recommended as a procedure for participatory model development. Initial contacts should be made with local producers to sound out felt needs. If models are useful for addressing the felt needs, a prototype should be developed using secondary data if available. The prototype model should be thoroughly discussed using focus groups or other methods common in Participatory Rural Appraisal. The concept of stakeholder should be understood in its broad context. A stakeholder analysis is useful for learning the relative position and networks that exist. Farmers, private consultants, extension agents, officials of government agencies, university faculty, and the researchers, among others, should be included in the model building process. After adjustments to the prototype, a final model should be developed and taken back to the stakeholders. Researchers should elicit comments on structure and function, parameters, coefficients, and individual modules from stakeholders. In order to insure long-term adoption of the model, good science must be used to respond to specific concerns of stakeholders. When farmers or consultants suggest that crops will be highly affected by climate variability, then modelers should respond by using state of the art crop models such as DSSAT. When producers call for a user

PAGE 119

97 friendly, simple interface, modelers should respond by using visual basic to create an interface that is usable from the farmers point of view. By following these relatively simple steps, which are not always in the mindset of scientists, use of climate-based agriculture and natural resource models can be greatly improved.

PAGE 120

CHAPTER 5 HOW MUCH MANURE NITROGEN DO NORTH FLORIDA DAIRY FARMS PRODUCE? 5.1 Introduction The presence of high levels of N in water is an environmental hazard because it affects human health and ecosystem welfare (Fraisse et al., 1996). The Suwannee River Basin has received much attention in recent years because increased N levels in water bodies (Albert, 2002; Katz et al., 1999; Pitman et al., 1997). Dairies have been identified as an important source of nitrates in the Suwannee River Basin (Andrew, 1994; Katz and DeHan, 1996; Berndt, 1998). It is critical to accurately estimate the amount of manure N production because it directly impacts environmental accountability of dairy farms, and the expense of farmers to comply with regulations. It is also important to perform these estimations seasonally because this determines potential for nutrient recycling, utilization, or loss because climate variability through the year. Other important estimations are those of manure N produced by non confined milking cows, dry cows, young stock, and bulls; interviews showed evidence that these could also be important factors in analyzing entire systems. For Florida, two approaches are being widely used to estimate manure N produced by dairy cows. One is that proposed in several publications by Van Horn et al. (1991, 1994, 1998, 2001) that consists of a sink-source balance, and the other is that presented by the Natural Resource and Conservation Service, NRCS, through an application called WATNUTFL Version 2.0 (NRCS, 2001), that is based on body weight of ruminants. 98

PAGE 121

99 According to Van Horn et al. (1998, 2001) the undigested (or unutilized) amount of N by the digestive system of a cow is the amount that will be present in the manure (feces and urine) of this cow, and undigested N can represent as much as 67% of the ingested N. In other words, the amount of N excreted by a cow will be the difference between the quantity ingested and the quantity of digested or utilized. N, which cows use to produce milk, gain weight, and reproduce. N(excreted) = N(feed) [N(milk) + N(tissue) + N(off spring)] [5-1] This relatively simple approach requires estimations of total amounts of N entering the system and total amounts of N leaving the system. Farmers have detailed records of milk production and they could estimate the quantity of N contained in this milk. They also know the total dry matter intake (DMI) for the whole herd and thus the total amount of N entering the system. But it is difficult to specify what group of cows received what quantity in each season of the year. Nennich et al. (2003) also criticizes this methodology indicating that errors in the estimations may occur since nutrients consumed by animals could vary greatly from what was formulated because cows select feed and the feed mixing process always has errors. The gross yearly differences in feed input and output would also include errors in measuring starting and ending inventories and the individual nutrient content of feedstuffs (Nennich et al., 2003). WATNUTFL, the application created by the NRCS, uses standardized estimations based on live weight of animals (NRCS, 2001). This methodology allows the user to standardize different species of animals by the common factor of body weight and benchmark parameter values. This methodology disregards a very important factor in manure N excretion, i.e., milk production, which many researches have pointed to as the

PAGE 122

100 driving factor in N excretion (Nennich et al., 2003; Van Horn et al., 1998, 2001; Tomlinson et al., 1996). In fact, milk production level, under intensive dairy farm systems, is the single most important factor to take into account regarding N excreted (Nennich et al., 2003). Both Van Horn and NRCS methods use standardized book values (benchmarks) of manure, DMI, water used, basic N excretion, and other parameters from those listed by the Agricultural Waste Management Handbook (AWMFH) of the NRCS (USDA, 1992a, 1992c, 1996), and its AWMFH Florida supplements (USDA, 1992b); several documents from the Department of Environmental Protection (FLDEP) and the Environmental Protection Agency (EPA); those from the American Society of Agricultural Engineers (ASAE, 1990); and other standards like those of digestibility of feed sources and N in diets from the National Research Council (NRC, 2001). These standards are the best sources or book values for which estimations can be supported and these are continuously being reviewed and updated; in some cases, location-specific values are also being released. This present study also bases the values of estimations on these standards. Van Horn et al. (2001) and NRCS (2001) estimate N excretion in whole-year values, ignoring completely the inherent seasonality present in north-Florida dairy farm systems. Dairy cows are highly sensitive to seasons, and in subtropical climate conditions such as north-Florida dairy cattle are negatively affected by high temperature and high humidity. Reproduction and milk production are completely changed by seasons of the year. For example, lowest milk production as well as lowest conception rates occur in summer (De Vries, 2004; personal interviews); consequently, a much higher rate of manure N is expected to be excreted during winter months, not only because milking

PAGE 123

101 cows are more actively producing milk at that time, but also because many more cows are in the high-peak milking production stage because there is a higher proportion of freshening cows in late fall. Manure N produced in mid spring will have more opportunity to be used by rapidly growing plants than that produced in a very cold winter season where less chance of plant utilization would exist and higher amounts of manure N applied in rainy seasons would be more prone to leach. That is why, for example, a study with application of dairy effluents to pastures in New Zealand (Williamson et al., 1998) concluded that effluent applications should be avoided in winter when they detected highest amounts of N leaching by soil lysimeters. The present study included inherent seasonality based on historical Florida dairy herd data (Dairy Herd Improvement Association (DHIA), De Vries, 2004) contained in seasonal conception rates, seasonal culling rates, and seasonal milking production rates. Nennich et al. (2003) suggest, after studying data from several universities in the United Sates for a series of historical records, that separate categories should be used to estimate excretion and nutrient content for calves, replacement heifers and mature cows. Even though previous studies made an effort to include them in balances, this study emphasizes their inclusion for a more accurate estimation of the manure N at the farm level. The objective of this study was to estimate monthly N levels that north-Florida dairy farms deposit onto their farmland. The specific objectives were to perform a survey of dairy farmers to collect information; develop a model that allows final users to have an easy but accurate monthly manure N estimation for specific farms, based solely on number of adult cows (NAC) and milk production (rolling herd average, RHA); compare

PAGE 124

102 the performance of this model with the Van Horn and WATNUT approaches; and assess the monthly N amounts applied to sprayfields and pasturelands by a number of north-Florida dairy farms. 5.2 Materials and Methods The study was conducted in cooperation with dairy operations of the Suwannee River Basin in the counties of Suwannee, Lafayette, Gilchrist, Levy, and Alachua (21' 30' N, 82' 83' W) during the period October, 2002 to December, 2003. Information from 21 dairy farms was collected by a survey. Additional secondary information was obtained from the Dairy Business Analysis Project (DBAP) of Florida (http://www.animal.ufl.edu/dbap), the Dairy Herd Improvement Association (DHIA) (http://www.dhia.org and http://www.animal.ufl.edu/dhia), the Dairy Records Management Systems (DRMS) (http://www.drms.org), and many other published studies. A sample of 21 dairy farms (~35% of the population) to participate in the interviews was obtained in coordination with the Suwannee River Partnership and the University of Florida Extension offices in the area. The sample was intended to cover the variability in north-Florida dairy farm systems with respect to herd size, farm size, manure handling systems, management operations, and environmental conditions. Interviews were arranged to be conversational in nature without an instrument, but with a guideline of topics. Interviews lasted between one and two hours and in several occasions, tours to the farm followed. Conversations were recorded, transcribed, and classified.

PAGE 125

103 5.3 Model Development 5.3.1 General Characteristics A Markov-chain cow-flow model was created to account for the growing, probabilistic reproduction, and probabilistic culling of young stock (calves and heifers) and for adult cows in different lactation cycles (1 to 9). Additionally, the model was set up to use monthly milk production rates according to production curves for Florida. Milk production, culling rates, and reproductive rates were adjusted using real data collected by the southeast DHIA in Florida (De Vries, 2004). The model starts in September of the first year and it run monthly for as many years as desired. The user must input the total number of adults cows (NAC) and the rolling herd average (RHA); however for more customization the user can input other information. Using these initial conditions, the model is capable of self-adjusting to reach a seasonal steady state of cow flow. That is the way north-Florida dairy farms operate, and any planned change to increase or decrease herd size can be accounted for by different model runs. 5.3.2 Markov-Chains, Culling Rates, Reproduction Rates, and Milk Production Rates St-Pierre et al. (1999) indicated that herd milk production through time can be modeled as a discrete stochastic process using finite Markov-chains. Markov-chains are widely used to predict economic outputs (De Vries, 2001; Mourits et al., 1999; Mayer et al., 1996). Cows at time t = 0 are assigned to homogeneous production cells in three-dimensional arrays with coordinates determined by months in milk (MIM = 1 to 32), months in pregnancy (MIP = 0 to 9, where 0 is not pregnant), and lactation number (LN=0 to 9, where 0 are calves and heifers). The processes of aging, pregnancy, involuntary cull, voluntary cull, abortion, dry-off, and freshening from month i-1 to

PAGE 126

104 month i for any month i are accounted for using seasonal probabilities. Estimates of transition probabilities are derived from historical herd data from the databases of the Florida DHIA (De Vries, 2004). Moving animals from one production cell to the next, based on the transitional probability assigned allows the inclusion of the inherent probabilities in the process of reproduction and culling in a dairy operation. Markov-chains represent the cycles that milking cows follow. When a female calf is born, she will be raised and will gain weight for 12 months. When she is 12 months old a pregnancy program starts for her through artificial insemination (AI) or by bulls. The pregnancy rate of heifers is much higher than adult cows: about 60% of the heifers will be pregnant on the first attempt. The estrus cycle in cows lasts 21 days, which is the time needed to know which heifers are pregnant. In the next attempt another 60% of the rest (40%) will become pregnant, etc. If the heifer is not pregnant when she is 24-months old, after 12 attempts, that heifer will be voluntary culled. The proportion of voluntary culling of heifers for not being pregnant is very low. After 9 months of pregnancy the heifer will deliver her first offspring, and she will become a fresh cow, first lactancy, first month, open cow in the milking group. The survey found that not all farmers raise their own replacements needed to maintain the herd size. In those cases, the model can be easily adjusted by a percentage of estimated replacements raised on-farm, with a range of 0% (no replacements at all) to 100%. During the first lactation cycle, Markov-chains divide cows into different categories according to whether or not they are pregnant, how many months they are pregnant, and how many months they have been milking. Similar situations will happen in the following lactation cycles. The number of possible categories in which cows can

PAGE 127

105 stay at any point in time depends on the model dimensions: 10 (lactation stages) x 32 (months growing or in milk) x10 (pregnancy stages) = 3,200. A voluntary waiting period (VWP) of 45 to 60 days will be observed before starting the reproductive program again in this first lactation cow. For adult cows in milking stage, the probability of reproduction is considerably reduced, and it depends on the season. An average value for Florida is 16%, meaning that only 16% of the cows of first lactation (or higher) will get pregnant in an attempt, De Vries (2004). But this varies and the model accounts these variations using DHIA data as seen in Figure 5-1. The pregnancy rate that encompass heat detection and conception rate decreases with months in milk and increases with colder months. 0.100.110.120.130.140.150.16234567891011month in milkpregnancy rate 0.600.700.800.901.001.101.201.301.40SepOctNovDecJanFebMarAprMayJunJulAugIndex B A Figure 5-1. Pregnancy rates of adult cows according to A) month in milk; and B) seasonality index of pregnancy rate by month of the year. Source: Prepared with data from De Vries, 2004. If after 12 months in milking the cow is not pregnant, she will be voluntarily culled from the herd; adult cows have much greater chance of being culled for reproductive problems than heifers. Two months prior to the delivery, the cow is dried off, preparing her for delivery. As soon as she delivers, she enters as a fresh cow, first month, open cow of second lactation. The offspring, if female, will enter the calf group. Calves will

PAGE 128

106 become heifers, heifers will become cows, and cows will age through successive lactation cycles. Fewer cows will arrive at higher lactation cycles because there is always a chance for any cow to be involuntarily culled. The involuntary culling rate depends of cow stage: heifer or adult, months in milk, and the season, as can be seen in Figure 5-2. For example, higher culling rates are expected in warmer months and in the first months of milking. 0.200.250.300.350.400.451234567891011121314151617181920month in milkculling rate 0.700.800.901.001.101.201.30SepOctNovDecJanFebMarAprMayJunJulAugIndex B A Figure 5-2. Culling rates of adult cows according to A) month in milk; and B) seasonality index of culling rate by month of the year. Source: Prepared with data from De Vries, 2004. Adult cows as well as young heifers or calves always have a probability to leave the herd for many reasons. Farmers call involuntary culling as the probability that an animal leaves the herd by an unfortunate event such as death, sickness, injury, etc. and voluntary culling as the probability that the animal becomes unpractical or un-economical to keep in the herd as in the cases of low production or low fertility. There are several studies and controversies about the theory of culling cows; for this study, involuntary culling was adjusted to the DHIA Florida averages (De Vries, 2004) as seen in Figure 5-2. The voluntary culling was based on months after which the cow is not pregnant: for adult cows at 12 months and for heifers at 24 months.

PAGE 129

107 Under this model structure, milk production standards for Florida were installed using information from the DHIA (De Vries, 2004). Since the model bases further calculations on milk production ratios, this was a very detailed component. Cows produce different amounts of milk depending on lactation number, months in milk, and season of the year. Curves show this seasonality according to months in milk for different lactations, Figure 5-3. 1st lactation curves: Jan, Mar, May, Jul, Sep, Nov4050607080JanMarMayJulSepNovJanMarMayJulSepNovJanMarMaylbs/day 2nd lactation curves: Jan, Mar, May, Jul, Sep, Nov4050607080JanMarMayJulSepNovJanMarMayJulSepNovJanMarMaylbs/day A B 3rd+ lactation curves: Jan, Mar, May, Jul, Sep, Nov4050607080JanMarMayJulSepNovJanMarMayJulSepNovJanMarMaylbs/day C Figure 5-3. Milk production rates of milking cows according to A) month in milk and season of the year for cows in first lactation; B) cows in second lactation; and C) cows in third to ninth lactation. Source: Prepared with data from De Vries, 2004. Figure 5-3 shows average rates of milk production for Florida based on lactations, months in milk, and seasonality. These milk production values can be adjusted for every specific farm based on the rolling herd average (RHA, average milk produced by a cow

PAGE 130

108 in the herd over the last 12 months) by running the model and comparing total milk production with the RHA target. The model will self-adjust to reach the RHA for that specific herd with less than 1% of error. 5.3.3 Manure N Produced by Milking Cows Well documented studies such as Van Horn et al. (1991, 1994, 1998, 2001), Tomlinson et al. (1996), Wilkerson et al. (1997), Nennich et al. (2003) have found that excretions are highly variable depending on many things, among them dry matter intake (DMI), nutrient concentration, and digestibility. Different approaches to estimate manure N in Florida have included these factors as well as estimated body weight as predictors (Van Horn et al., 1998; NRCS, 2001). However, all these data are not always available and/or are not totally reliable. This study proposes a simpler way to estimate manure N based only on milk production. Knowing the digestibility (N used in milk, maintenance, and offspring) (Van Horn et al., 1998; Wu et al., 2001), we can estimate the N excretion as a function of the amount of milk produced. A study carried out across the United States (Nennich et al., 2003) found that the single best predictor for N and manure excretion is milk production. Nennich et al. (2003) concluded that milk production drives feed intake and ultimately nutrient excretion values reflect the relationship among feed intake, milk production (nutrient utilization), and nutrient excretion. Crude protein content in the diet also affects the N flow amounts. Wu et al. (2001) mention that 25 to 35% of the N consumed in cows will go with the produced milk, almost entirely the difference (65 to 75%) will be in the manure. Several studies (reported by Van Horn et al., 1998) have found that milk production is correlated with manure production and most importantly with N excretion. Higher manure production in higher milk production months (winter) also has been reported in the direct farmer interviews.

PAGE 131

109 Table 5-1 shows a direct relationship among milk production, DMI, and excretions. High milking cows will consume more dry matter (DM) and excrete more manure and more N. If these high producers consume less DM for any reason, they will produce less milk and consequently less manure and N. Low milking cows will consume less DM and produce less milk compared to high milking cows and consequently excrete less manure and less N. Table 5-1. Dry matter intake, manure, and N excretion by Florida dairy cattle based on milk production. (lbs cow -1 day -1 ) and low and high (NRC standards) crude protein (CP) content of diets N Excretion milk DMI manure low CP high CP 100 55.8 195 0.899 1.030 70 46.3 160 0.727 0.846 50 39.2 125 0.601 0.698 0 25.2 80 0.364 0.439 Sources: Modified from Van Horn et al. (1998), p. 3.;USDA (1992a), p. 8., USDA (1992b), p. 8. Note: NRCS uses these values for estimations. If high and low producers are fed more than required they will not produce more milk but they will tend to gain body weight and will excrete more manure and N. This is not desired by the farmers (found in the interviews) because it is a waste of money and an unhealthy choice for the cows (fat cows have more health problems and higher risks in pregnancy and delivery). Thus, farmers will always try to give only enough feed to achieve optimum performance. Therefore, farmers group cows by production performance, preparing different rations, and giving different feed amounts. It is assumed for the model that the amounts of feed are based on milk production rates and farmers will feed their cows accordingly. Van Horn et al. (1998, 2001) and NRCS (2001) estimate the total N excreted assuming fixed days in lactation stages in continuous identical cycles for all cows. In this study, great effort has been devoted to incorporating milk production by north-Florida

PAGE 132

110 herds using historical data of culling, reproduction, and milk production rates based on seasons. These calculations are used to estimate N excretion. 5.3.4 Nitrogen and Crude Protein Nitrogen is a component of amino acids and ultimately proteins, therefore, protein content of the diet has direct repercussion on quantities of excreted N. Evidence exists that farmers may over-enrich cow diets with proteins; consequently it would be possible to decrease amounts of N overflows in dairy systems by decreasing cow protein intake without decreasing productivity (Van Horn et al., (1998, 2001); Wu et al. (2001); Jonker et al., (2002); Brsting et al. (2003)). Also, studies indicate that the source of N or protein can make a difference in the quantity of excreted N (Wu et al. (2001); Brsting et al. (2003)). Van Horn et al. (1998) present a table with the differences in N excreted by two different ratios of crude protein (CP) in the diet based on the NRC indexes: one that uses high protein to assure that true protein needs are met and the other low that minimizes dietary N intake. For cows producing 100, 70, 50, and 0 lbs of milk the high CP rations were 17.5, 16.4, 15.3, and 12%, respectively and the low diets were 16, 14.8, 13.8, and 11%, respectively. With these two formulations, the milk production remained steady, but the amount of N produced changed. Van Horn et al. (1998) estimate for a year, for a 1400-lb cow, producing 20,000 lbs of milk and consuming 15,099 lbs of dry matter, an N excretion of 234 lbs with the low CP diet, while it would be 273 lbs with the high CP diet. This would mean that a 15% of change in N excretion would be expected by these changes in diets. Additionally, CP in the diet will impact the feeding costs. Wu et al. (2001) conducted a study on dietary manipulation to reduce N excretion by lactating cows in Pennsylvania, concluded that dietary protein can be moderately

PAGE 133

111 manipulated without affecting milk production. They recommended other means of balancing rations for maximal microbial protein production using synthetic limiting amino acids. Strategies to reduce N excretion may include eliminating excess dietary protein, maximizing microbial protein production, and using protected amino acids for reductions in total protein that needs to be fed. Using these estimates, and based on milk production for every cow group in the model, good estimations of DMI, manure, and N can be calculated for high and low CP diets (Figure 5-4.). y = -3E-07x3 + 6E-05x2 + 0.0024x + 0.36400.10.20.30.40.50.60.70.80.91020406080100Milk Production (lbs/day)N excreted (lbs/day) Figure 5-4. Manure N produced by low-CP diet cows according to milk production The function presented in Figure 5-4 can be corrected based on overall CP in the diet (weighted average of crude protein received by the 100, 70, 50 and 0 milking cow groups) that are estimated as 13.947 for a low-CP diet and 15.034 for a high-CP diet (NRC averages). A function with a slope of 0.096 for the average CP diet above 13.937 performs this correction. 5.3.5 Milking Times and Cattle Breeds There is a dilemma regarding the effects of number of milking times per day on milk production. Farmers in north-Florida use two or three times a day milking (2X-3X),

PAGE 134

112 or even, in exceptional cases, four times (4X). Using 3X compared with 2X is reported for Florida to increase milk production (Giesy, R. 1995. Another look at 2x-3x milking. University of Florida Cooperative Extension Service, Unpublished. Dep. of Animal Science, Gainesville, FL.), although these results are not conclusive (Hoekema, 1999). It seems that 3X will produce more milk, but at an elevated production cost, which is why some dairies prefer to reduce 3X to 2X in summer months when overall lower production rates are observed. Direct effects of different milking times on manure N production were not included in the model since it only depends on the quantity of milk produced regardless of the times cows were milked. Another dilemma is variation in N production with cattle breeds. All estimations are based on experiments with Holstein cows, which are dominant in the study area (probably 95% of cows in north-Florida are Holsteins). However, there are a few farms that manage Jersey, Shorthorn, or, more commonly, crosses of these breeds with Holstein. There will be differences between herd breeds and manure N produced. For example Jersey cows are expected to produce less N, but based on previous assumptions, the best indication is the amount of milk produced, which will determine manure, N and DMI. In fact, Knowlton (2000) affirms that Jerseys and Holsteins metabolize nutrients similarly and consequently the amount of manure and N excreted per body unit weight is the same in both breeds; results shown by Knowlton (2000) also demonstrate that the direct relationship between milk production and waste excretion are similar in both breeds. Additionally, using Jonker et al. (2002) assumptions of manure N production by different breeds, including Holstein, Brown Swiss, Ayrshire, Guernsey, Jersey, and

PAGE 135

113 Milking Shorthorn, we can safely estimate manure N production across breeds based solely on milk production. 5.3.6 Manure N Produced by Dry Cows, Young Stock, and Bulls Less information regarding manure N production is available for dry cows, young stock and bulls because they are not usually recognized as environmental hazards. They produce less manure N proportionally, and they are usually kept in open areas under grazing conditions (contrary to confined areas as is the case of milking cows). This study emphasizes the importance of accounting also for these sources of manure N inside the dairy farm system, but through less elaborate estimations. For dry cows, the estimations were based on several references (Van Horn et al. (1998, 2001), NRCS (2001)) that indicate that a dry cow produces 0.364 lbs of N day -1 and this value can be corrected as above, according to CP intake. Estimations for calves and heifers are made using body weight, assuming that a newborn calf weights 90 lbs and gains approximately 40 lbs per month. For this group, book values from Table 4-5 of the AWMFH (USDA, 1992a) for heifers indicate that a heifer of 1,000 lbs produces 0.31 lbs of N (in 85 lbs of manure). With this information as a benchmark, and interpolating by body weight, the model estimates that a heifer of 5 months that weighs 90 + 5 40 = 290 lbs, would produce (290 0.31) / 1000 = 0.0899 lbs of N in (290 85) / 1000 = 24.65 lbs of manure. There are no correction numbers for low/high CP according to NRCS (2001) because their diet is assumed to be standard. For bulls (which are used for reproductive purposes in many dairies) there is not much information about their DMI and/or manure N production. They could be estimated based on their bodyweight and under the parameters of heifer cows in distinct stages. Bulls are usually larger than cows; they weight 10 or 20% more than the 1400-lbs for

PAGE 136

114 average cows. But because they do not produce milk or become pregnant, they require lower feed intake and, consequently their manure N production would be less. It is estimated that bulls consume similarly to dry cows or heifers in relation to their body weight, 25.2 lbs of dry matter/day and produce 80 lbs of manure with 0.364 lbs of N for each 1400 lbs of body weight. Bulls go with the milking cows in order to breed them; their manure joins that of the milking cows. 5.3.7 Manure, Feces, Urine, and Dry Matter Intake Estimations The amount of manure excreted by milking cows can be estimated with the function presented in Figure 5-5 based on values presented by Van Horn et al. (1998), Table 5-1. An estimation for cows producing 20 lbs of milk in a day has been added considering that a milking cow always will produce more manure than a dry cow (more than 80 lbs) and that there will not be milking cows producing less than 20 lbs/day. y = -0.0003x3 + 0.0525x2 1.3708x + 98.750204060801001201401601802002030405060708090100Milk Production (lbs/day)Manure Excretion (lbs/day) Figure 5-5. Manure excretion based on milk production Manure produced by dry cows can be estimated as proposed by Van Horn et al. (1998, 2001) and NRCS (2001) at 80 lbs/cow/day. For heifers, the estimation can be performed similarly to that of N, based on body weight (age) and according to Table 4-5

PAGE 137

115 of the AWMFH (USDA, 1992a). From the total manure, according to Van Horn et al. (1998, 2001), 61.35% and 38.65% are the proportions of feces and urine in the manure, respectively. Dry matter intake (DMI) for milking cows can be estimated based on milk production following the function presented in Figure 5-6, based on Van Horn et al. (1998), Table 5-1. y = -2E-05x3 + 0.0033x2 + 0.1621x + 25.201020304050600204060801Milk Production 00 ( lbs/da y) DMI (lbs/day) Figure 5-6. Dry matter intake (DMI) based on milk production Amounts of DMI by heifers were estimated based on body weights using averages of adult cows. Adult cows, according to Van Horn et al. (1998, 2001) consume on average 15,099 lbs of DMI in a year (1400-lb cows); this is about 3% of their body weight in a day as DM. Interpolating that to calves and heifers, a recent born, 90-lb calf will consume 2.7 lb DM and a 32 month old heifer (40 32 = 1280 lb) will consume as a dry cow = 25.2 lbs. Therefore, for a 16 month old heifer the DMI will be 2.7 + 0.7031 16 = 13.95 lb. DMI for bulls is assumed to be similar to dry cows, 25.2 lbs/day.

PAGE 138

116 5.3.8 Water Utilization for Manure Handling Many components determine the total amount of water (drinking water for cows, cleaning equipment, cooling system, flushing manure, etc.). Van Horn et al. (1994) estimated that the amount of water used for a milking cow in a day could vary from 49 gallons (187 L) for minimum usage (no flushing) to 171 gallons (650 L) in high usage (with flushing). Further, this has a direct relationship with the season of the year: hotter months will require more water than cooler months. Water is critical for the dairy operation; in the case of the north-Florida waste management systems it is also the carrier of the manure and N. All farms studied had some kind of flushing system. Using the above values, an interpolation can be performed to estimate the amount of water used by milking cows, assuming that approximately the use of water in the coldest month will be half of that in hottest month (found from farmers during interviews), and therefore the amount of water used in every month can be expressed as a percentage of the amount used in the hottest month. Amounts of water required by heifers and dry cows will be much less than the milking cows and those will not enter the manure flushing system. These estimations are presented in Table 5-2. Some dairies recycle water from the manure system. Recycle means that they use the flushing water more than once, saving fresh water in the process. Because there are great differences among farms, a user-defined input in the model allows the user to adjust the amount of water they use as a function of the maximum water used in a single month and the percent of water recycled. Water usage for any month will be estimated by multiplying the maximum water used in the hottest month, the percentage of use in a specific month, the number of milking cows, and the percent of water recycled.

PAGE 139

117 Table 5-2. Water utilization by milking cows during months and percentage use relative to month of maximum use (July) Gal cow -1 day -1 % September 146.93 85.90 October 134.87 78.85 November 122.81 71.79 December 110.75 64.74 January 98.68 57.69 February 110.75 64.74 March 122.81 71.79 April 134.87 78.85 May 146.93 85.90 June 158.99 92.95 July* 171.05 100.00 August 158.99 92.95 *Month of maximum water utilization. 5.3.9 Waste Management Handling Systems All visited north-Florida dairy farms handle their waste through a liquid manure system, but there are variants according to how they collect, store, process, transport, and utilize the waste. These variants are important because they determine how much manure is recovered and how much is potentially lost to the air. North-Florida dairy farms have some variants from a usual waste management system, consisting of a water flushing system, a storage pond, and an irrigation system. Additionally they may also have a solid screener and/or a treatment lagoon. The flushing system uses great amounts of water to move the manure (feces and urine) deposited in the concrete; the storage pond receives and stores the liquid manure for variable time; and the irrigation system applies the liquid manure to the crop systems. Both, the solid screener and the treatment lagoon are designed to separate solids from the flushed manure: the first one by mechanical action and the second one by gravity or precipitation. Both, if present, are located before the storage pond.

PAGE 140

118 As soon as the manure is deposited on the concrete, variable rates of ammonia will be lost to the air as volatilization. If solids are screened, the manure that reaches the storage pond (in which it remains a variable time) will be a solution with less than 5% solids. Nitrogen in this liquid manure will be ammonium and organic compounds in solution. Extra amounts of ammonia will still be lost by volatilization to the air during the holding time in the storage pond and during the spraying application in the fields. If solids are not screened, the liquid manure will have greater than 5% solids but it still will be sprayed on the fields using high pressure equipment and an agitator to maintain it liquid. The proportion of solids collected by separators varies greatly among systems; for example a stationary separator can take 20 to 30% of the solids, while a well-designed gravity separator can collect up to 60% of the total solids (Van Horn et al., 1998). However, at least 80% of the N is contained in the liquid solution and is not possible to recover through solids (Tomlinson et al., 1996). If the separation is through a stationary separator, the solids collected will have 72% moisture and from the remaining 28% of DM, only 1 to 1.6% will be N (Van Horn et al., 2001). If it is separated by gravity, the separation has more moisture and is unlikely to be used as bedding, feed or composting; this is usually spread on the land (Van Horn et al., 1994). Even in the most controlled manure systems, loss of N in relatively high quantities will be expected. Between 41 and 50% of the N in manure is in the form of urea or ammonia, which can easily be volatilized (Van Horn et al., 1998). Urea in contact with the ureasa enzyme, present almost everywhere, will convert to ammonia (NH 3 ), and the ammonia as a gas, volatilizes to the air. Even though the fecal form of N, mostly in the

PAGE 141

119 feces, is more stable, it could also be freed up by anaerobic digestion. Alkaline conditions in the collection pathways and lagoons will result in more volatilization, while acidic conditions will decrease volatilization. Once the liquid manure reaches the irrigation system, a last amount of volatilization, proportional to the water evaporation during irrigation can be expected. Amounts vary greatly in different management systems, but Van Horn et al. (1998) presents the result of several University of Florida studies throughout the state (Table 5-3), in which the following indexes of N lost were calculated:` Table 5-3. Nitrogen lost in the waste management system: a baseline % Volatilization before flushing 2.38 Collected solids 6.5 Volatilization in lagoon and or pond 9.21 Sludge accumulation in lagoon and/or pond 5.74* Volatilization during irrigation 11.48 Total N lost during manure handling 35.32 Estimated based on surveyed north-Florida operations. Source: Several University of Florida studies, Van Horn et al. (1998). According to these values, 35.32% of the manure N is lost, and the difference 64.68% reaches the crop fields. Because these great variations, it is important to let the user select his/her system and N indexes. The model allows the user to input these values based on real analyses and reports made to the regulatory agencies. 5.4 Results and Discussion 5.4.1 The Model (Livestock Dynamic North-Florida Dairy Farm Model, LDNFDFM) The main result of this study was the Livestock Dynamic North-Florida Dairy Farm Model. The LDNFDFM is capable of giving accurate manure N estimations solely based on number of adult cows (NAC) and rolling herd average (RHA) of any north-Florida

PAGE 142

120 dairy farm. NAC and RHA are always available in any north-Florida dairy farm and they are highly accurate. An approximation of the amount of crude protein (CP) in the diet, degree of seasonality in operation, confined time, and characteristics of the waste management system can also be manipulated to better mimic specific farm conditions. The LDNFDFM was implemented in Excel using Visual Basic programming as a user-friendly application. 5.4.2 Model Comparison with Other Widely Used Estimations Estimations of Van Horn et al. (1998 and 2001) and WATNUTFL (NRCS, 2001) were compared with estimations from the model. To achieve such comparisons, a hypothetical farm was set up with the following characteristics: the farm had 100 producing cows, which weigh 1400 lbs on average and were confined 100% of the time. A first comparison was performed using low protein diets for cows with a RHA of 18,250 lbs of milk/year. Results are shown in Figure 5-7. 1,7501,8001,8501,9001,9502,0002,0502,100sepoctnovdecjanfebmaraprmayjunjulauglbs of nitrogen VAN HORN NRCS LDNFDFM total=23,400 total=21,973 total=23,281 Figure 5-7. Nitrogen excretion by a low protein diet in a 100-cow dairy estimated by different models

PAGE 143

121 Overall yearly estimation of the LDNFDFM (23,281 lbs of N) was between values estimated by Van Horn (23,400 lbs of N) and NRCS (21,973 lbs of N), but closer to that estimated by Van Horn. Since Van Horn and NRCS calculate values on a yearly basis, they assume steady monthly values as seen in Figure 5-7. However, the DNFDFM shows a more realistic seasonal variation throughout months of the year, following the common seasonality in dairy operations. Another comparison was performed using high protein diets for cows with a yearly milk production of 21,900 lbs. Results are shown in Figure 5-8. 1,9502,0002,0502,1002,1502,2002,2502,300sepoctnovdecjanfebmaraprmayjunjulauglbs of nitrogen VAN HORN NRCS LDNFDFM total=26,371 total=25,723 total=25,550 Figure 5-8. Nitrogen excretion by a high protein diet in a 100-cow dairy estimated by different models Annual estimation by the LDNFDFM (25,723 lbs of N) was between Van Horn (26,371 lbs of N) and NRCS (25,550 lbs of N), but this time estimation was closer to that of the NRCS. Figure 5-8 shows these three estimations and the seasonality present only with the LDNFDFM model.

PAGE 144

122 5.4.3 Estimation of Manure N Produced in North-Florida Dairy Farm Systems In order to keep confidentiality of dairy farms, five hypothetical farms were constructed using data from the farm interviews. Hypothetical farms were created to give real information about the manure N production of north-Florida dairy farms without pointing to specific farms. For every dairy, inputs consisted of a set of information that included the total number of adult cows (NAC), number of bulls, the percentage of replacement animals raised on the farm, the rolling herd average (RHA), the estimated percentage of confined time of milking cows, the general characteristics of the waste management system, and the area of land devoted to sprayfields and pasturelands. This information was used in each case as the initial conditions (IC) inputted in the model before running (Table 5-4). The ratios of N per area (lbs ac -1 ) to be applied to sprayfields and to be deposited in pastureland were estimated for each of these hypothetical farms. Table 5-4. Five hypothetical north-Florida dairy farms and their characteristics for manure N estimations Farm 1 2 3 4 5 NAC, Number of Adult Cows 1,000 150 600 1,500 800 BULLS Head 50 8 65 75 38 HEIFERS % 20% 0% 100% 0% 50% RHA, Rolling Herd Average lbs yr -1 cow -1 16,500 13,000 20,075 16,800 21,000 CONFINED % 62.50% 30.00% 70.00% 40.00% 60.00% WASTE SYSTEM code* 1 2 2 2 1 AREA SPRAYFIELDS 90 130 45 240 120 AREA PASTURELAND ac 440 130** 45 120 250 SPRAYFIELD per COW 0.09 0.87 0.08 0.16 0.15 PASTURELAND per COW 0.44 0.87 0.08 0.08 0.31 TOTAL AREA per COW ac cow -1 0.53 0.87 0.15 0.24 0.46 *1=separate solids; 2=not separate solids. **Are the same sprayfields Farm 1 is a medium-large semi-intensive operation with medium productivity of milk. It has shortage of sprayfields, but more abundant pastureland. Its waste system is efficient and has the facilities and the comfort to manage confined animals. Farm 2 is a

PAGE 145

123 small grazing system where animals spend more of the time outside; the breed is Jersey or Jersey cross that endures heat stress in summer, but has low milk productivity. This farm has plenty of land for its herd, which serves at the same time as sprayfield and pasture (this is unusual). This farm prefers not to raise heifers; it sells all off-spring and buys replacements. Farm 3 is a medium-size intensive operation with high milk productivity; all heifers are raised on-farm and animals spend most of the time confined. This farm puts high pressure on the land not only because of the intensity of the operation, but also because of the shortage of land in both sprayfields and pasturelands. Farm 4 is a mixture between intensive and grazing farm, large in size, medium intensive with medium milk productivity, and low time cows spending over concrete. It has abundant sprayfields, but a shortage of pastureland. It does not raise heifers and the waste system does not include solid collection. Farm 5 is a medium-size intensive system with high milk productivity that keeps cows confined an average time, and seems to have a shortage of sprayfields and pastureland; it raises only half of its heifers and has an advanced waste system that collects solids. For simplicity, the only variations in waste systems accounted for in Table 5-4 are code 1 (they separate the solids), and code 2 (they do not separate solids). Using values previously presented as a baseline in Table 5-3, it was estimated that with code 1 there is 35.32% N lost and with code 2 there is 28.82% N lost through the manure handling system. Specific farms can and should enter their own specific values.

PAGE 146

124 Usually sprayfields receive liquid manure and the pastureland direct excretions of the dry cows, and young stock, and the milking cows and bulls during non-confined time, but there are a few cases where the fields can receive both as is the case of Farm 2, where the sprayfields are also the pasture fields. For all farms a low protein diet was assumed and water was estimated using the ratio 171.05 gallons maximum amount used by a cow per day in the hottest month, and that water is not recycled. Particular farms should enter their own specific data. Tables 5-5, 5-6, and 5-7 present the estimated results for these five hypothetical farms. The first glance indicates that there is great variability among farms. Table 5-5 shows how much manure N the sprayfields receive on a monthly basis according to the characteristics of each of these operations. There is a farm that applies as little as 3 lbs per acre in a month (Farm 2) and a farm that has to apply as much as 136 lbs per acre in a single month (Farm 3). Greater amounts of manure N are sprayed in winter months and lower amounts in summer months, although the amount of variation in seasons also varies according to the characteristics of each farm; some farms have very small variations throughout the year (Farm 2 or Farm 4) while another has greater variations (Farm 3). This difference in seasonal variations is explained by the interaction between natural seasonality of the cow-flow and productivity, and the CT; the lower the CT, the lower the seasonality noticed in the sprayfields because less manure N is going through the waste system. The opposite is also true as in the case of Farms 3 and 5 that have higher seasonality. Amounts of manure N per acre in a whole year are also highly variable among farms. There is a farm that has to apply more than 1,540 lbs of N per acre in a year (Farm

PAGE 147

125 3), while there is a farm that only requires applying 45 lbs of N per acre in a year (Farm 2). All factors in Table 5-4, except percent of heifers raised and land for pastures, affect these amounts. When a higher land per cow ratio, lower RHA, and lower CT the total N per acre per year will be lower as can be seen in Farm 2 and the opposite for Farm 3. The waste system impacts the final amount of manure N reaching the sprayfields since different amounts of N will be lost; in Table 5-4, waste system 1 loses more N than system 2. Therefore, Farms 2, 3, and 4 would decrease amounts of N in sprayfields if they would implement solid collection in the waste system (change from system 2 to system 1). Table 5-5. Estimated amounts of manure N to sprayfields in five hypothetical north-Florida dairy farms Farm 1 2 3 4 5 September 70.22 3.46 119.75 28.08 47.42 October 74.04 3.65 126.08 29.60 49.92 November 73.39 3.62 124.78 29.34 49.41 December 78.53 3.88 133.24 31.39 52.76 January 80.50 3.98 136.41 32.18 54.02 February 75.71 3.74 128.29 30.27 50.81 March 80.73 3.99 136.84 32.27 54.19 April 77.27 3.82 131.09 30.89 51.92 May 78.41 3.87 133.20 31.35 52.76 June 73.24 3.61 124.67 29.28 49.38 July 72.92 3.60 124.34 29.16 49.25 August lbs ac -1 mo -1 71.79 3.54 122.50 28.70 48.51 Total Nitrogen lbs ac -1 yr -1 907 45 1541 363 610 Nitrogen per Cow lbs cow -1 yr -1 82 39 116 58 92 The contribution of each cow to the amount of N applied to sprayfields varies greatly among farms. The model estimates that in Farm 2 the amount of manure N that reaches sprayfields is 39 lbs per cow. This amount is substantially lower than the others and the lowest of all. Farm 2 has the lowest ratio because its RHA is the lowest, reducing the amount of manure N produced for each animal and the confined time of the cows is

PAGE 148

126 also the smallest compared with the others (30%). This low rate of N to sprayfield happens even though this farm does not collect solids from the manure. On the other hand, Farm 3 applies 116 lbs of manure N to sprayfields per each cow. This is substantially higher than the others and the highest of all. Three reasons explain this fact: the high RHA (20,075 lbs cow -1 year -1 the second highest); the high CT (70%, the highest of all); and the waste system does not collect the solids, so more manure N goes to sprayfields. Although the NAC and the RHA have the greatest impact on the overall amount of manure N produced, the waste system and the available sprayfields determine how much manure N reaches these fields. Table 5-6 presents the results of the estimations of the monthly manure N that pastureland would receive in each of the farms. There is also a great difference in the amounts of manure N received among pasture fields. Table 5-6. Estimated amounts of manure N deposited on pasturelands in five hypothetical north-Florida dairy farms Farm 1 2 3 4 5 September 18.96 13.41 140.98 140.90 34.62 October 19.44 13.88 145.06 145.60 35.67 November 18.61 13.44 139.88 140.71 34.40 December 19.03 13.96 143.91 145.75 35.47 January 18.78 13.95 142.88 145.46 35.27 February 17.41 12.98 133.27 135.35 32.86 March 18.61 13.86 142.70 144.47 35.13 April 18.05 13.37 138.20 139.56 34.00 May 18.86 13.84 143.44 144.65 35.30 June 18.53 13.38 139.91 140.10 34.35 July 19.45 13.81 145.72 144.88 35.68 August lbs ac -1 mo -1 19.59 13.83 145.93 145.10 35.75 Total Nitrogen lbs ac -1 yr -1 225 164 1702 1713 418 Nitrogen per Cow lbs cow -1 yr -1 99.14 141.87 127.64 137.00 130.78

PAGE 149

127 Exploring Table 5-6, the first conclusion that can be drawn is that these amounts are not small; in fact, the quantity of manure N applied to the fields through direct excretion could surpass those amounts applied on the sprayfields in some cases (Farms 2, 3, and 4). This fact has profound implications because it discovers a source pollution that should be taken into consideration; current major efforts are addressed mostly to sprayfields and waste systems. Some pastures receive manure N excretions with as little as 13 lbs of N in a month (Farm 2), but others receive excretions with as much as 145 lbs of N in a month (Farms 3 and 4). The quantity of N deposited on pasturelands also has seasonality through the year, but usually different from that of the sprayfields. Higher amounts are expected in summer and fall and lower amounts in winter and spring following the cycle of dry cows and heifers; there would be more cows milking in winter and consequently fewer dry cows on pasture, and more heifers will freshen during fall months, moving them from pastureland to more confined areas. However, as Table 5-6 indicates it varies greatly also because other factors such as the percent of heifers raised on-farm or the confined time of milking cows. The amount of manure N deposited with the excretion in pasturelands varies from 164 lbs of N in a year (Farm 2) to more than 1,700 lbs of N in a year (Farms 3 and 4). Land area in pastures together with the number of heifers, dry cows, and non-confined time of milking cows and bulls determines these amounts. Additionally, the RHA is important depending on how much time milking cows spend on pasture. The amount of N per cow that is deposited in pasture fields depends on the RHA; the number of bulls; the percent of heifers raised on-farm; and the CT. The CT is very

PAGE 150

128 important since it determines how much time milking cows spend in pastures; e.g., Farm 2, that has the lowest CT of all, presented the lowest manure N per cow in sprayfields, however it shows the highest N deposited per cow in pasture fields. Table 5-7 presents additional information estimated by the DNFDFM for the five hypothetical farms regarding the manure produced, dry matter intake, and water use. Table 5-7. Estimated amounts of manure, DMI, and water use in five hypothetical north-Florida dairy farms Farm 1 2 3 4 5 Total Manure Confined 26,333,116 1,663,958 20,889,635 25,570,629 24,397,989 Total Manure Deposited 30,913,417 4,370,231 18,020,830 43,232,565 23,611,177 Total Dry Matter Intake lbs yr -1 17,147,309 2,400,529 11,339,436 25,873,975 15,259,717 Water Use gal yr -1 43,706,188 6,555,930 26,223,723 65,559,310 34,964,950 Table 5-7 summarizes estimations for a whole year. The total manure confined is the wet weight of urine and feces deposited on concrete during the confined time by milking cows and bulls that will reach the waste management system. The manure deposited refers to the amount of feces and urine directly deposited on pasturelands. The total DMI includes all the feedstuff on a dry matter basis for all animals and the water usage is for the confined animals. These estimations are based on livestock characteristics and not on the available land. As in the case of N, estimations of these variables give an idea of the magnitude of variation of these operations regarding the amounts of materials into and out of the system. 5.5 Conclusions A user-friendly model to accurately estimate manure N production by north-Florida dairy farms has been developed. It is called the Livestock Dynamic

PAGE 151

129 North-Florida Dairy Farm Model (LDNFDFM). The LDNFDFM estimates on a monthly basis the amount of N that reaches the sprayfields and the pastureland on a farm with minimal dataset input required. Estimations using LDNFDFM were more easily and quickly performed than with the other two models widely used in Florida (Van Horn et al., 2001; NRCS, 2001). Estimations by the LDNFDFM, although relatively close to the others, are believed to be better because they take into account the real dynamics of dairy cows and additionally include seasonality. The LDNFDFM model requires less information than the others since it does not use body weight or dry matter as parameters of estimation. Comparison of the LDNFDFM with models used in Florida demonstrates that the DNFDFM performs inside the expected ranges but with advantageous characteristics by the inclusion of Florida dairy parameters. This comparison might suggest that previous models either overestimate or underestimate N excretion. Another distinctive characteristic of the LDNFDFM is that this model not only estimates manure N produced by adult cows, but also that produced by young stock and bulls; additionally the estimations not only include amounts of manure N of the waste management system, but also that deposited in pastures. Using the model, estimations for 5 hypothetical north-Florida farms were performed. Results indicated great variability showing that some fields could receive as little as 3 lbs and as much as 145 lbs of manure N per acre in a month according to seasonality of the year, and that these amounts are independent of being a sprayfield or pastureland. Herd size, milk production, and farm size are critical variables affecting these results. The fact that pasture fields can receive large amounts of manure N, found

PAGE 152

130 by simulating these hypothetical farms with the model, has profound implications in policy making and regulations towards decreasing N impacts from dairy farms. After the simulation of the 5 hypothetical farms, it can be concluded that the main factors that lead the N discharge in the fields in a dairy farm are, in order land area/number of animals; milk productivity or intensity of the farm; characteristics of the waste system; confined time; and interactions among previous factors. The number of bulls and the percentage of heifers raised on-farm have lower impacts. Farms with more land available per cow (either in pasture or sprayfields) will have less environmental problems because manure N can be distributed over a higher land area, so the amount of N per unit of area decreases. However the commitment of more land to dairy production has an economic cost. It is also clear that farms with higher ratios of productivity (milk per cow) require more feed, which brings more N that cows intake and excrete. Farms with high productivity have more N in the system and carries a higher environmental impact. Farms have to balance the trade-offs between productivity and environmental impacts. Low milk productivity could be a valid option depending on the compliance costs; it will not be surprising that in the future more farms adopt this alternative if they are able to maintain an acceptable profit level. The waste system and the confined time are closely related since the first determines how much of the confined manure N arrives at the sprayfields, and the latter determines how much manure N pastures receive (plus additions from heifers and dry cows). At the moment, the higher the collection or loss of N from the waste system, the lower the environmental impact. However, this might change in the future when probably every pound of N will have to be accounted, no matter where it goes. The confined time

PAGE 153

131 and the environmental impact vary also according to the available land for sprayfields and pastures; in general, if the sprayfield cow/land ratio is higher than the pasture, it would be environmentally favorable to have a high confined time. User-friendliness of the model was desired and achieved. The DNFDFM works as an Excel file and only requires the initial conditions of the farm to run. It could be recommended to be introduced and compared with traditional methods by farmers and other stakeholders in the issue of N pollution by dairy farms in the Suwannee River Basin.

PAGE 154

CHAPTER 6 FORAGE SYSTEMS IN NORTH FLORIDA DAIRY FARMS FOR RECYCLING N UNDER SEASONAL VARIATION 6.1 Introduction The presence of N in water is an environmental hazard because it affects human health and ecosystem welfare. The Suwannee River Basin has received much attention in recent years because increased N levels in water bodies. Dairy waste may be an important factor contributing to this water N pollution. Dairy farmers are now required to comply with stricter environmental regulations either under permit or under voluntary incentive-based programs. Evidence indicates that the main way farms have to reduce their total N loads is through forage crop systems that are able to recycle an important part of this N produced on the farm. Improvements in climate or seasonal weather predictions (lead times of 6 to 12 months) can play an important role in devising management strategies that dairy farmers in north-Florida could adopt regarding forage crop systems to pursue economic and ecological sustainability. Although exporting manure off the farm is an option, few farms are using this practice and only for a small part of their waste. Evidence from interviews indicates that dairy farmers must to deal with their manure inside the farm gate. Manure is applied to fields through spraying systems and/or through direct animal depositions. Amount of N in this manure is usually high compared with normal agronomic applications. Therefore, efficient crop systems are necessary to uptake as much N as possible and decrease environmental problems. Climatic parameters such as temperature, solar radiation, and 132

PAGE 155

133 rainfall will direct and indirectly affect the crop-soil-environment and interaction with N flow. A few studies have evaluated the manure N uptake and loss in forage systems of dairy farms. In Georgia, Hubbard et al. (1987) found N concentrations between 10 to 50 mg L -1 of NO 3 -N at 2.4 m below the soil surface in a forage system of bermudagrass (Cynodon spp)-ryegrass (Lolium multiflorum) when 530 to 1080 kg ha -1 of N in effluents were applied. Johnson et al. (1991) presented data indicating that bermudagrass uptakes between 107 and 130 kg N ha -1 rye (Secale cereale L.) between 140 and 250 kg N ha -1 and corn (Zea mays L.) between 178 and 237 kg N ha -1 when applied rates of manure N were between 385 and 1,000 kg N ha -1 Vellidis et al. (1993) found N concentrations between 0 and 14 and between 4 and 21 mg L -1 in a system of bermudagrass-rye applied with 400 and 800 kg ha -1 respectively. A follow up of this study, reported by Newton et al. (1995), found that N uptake by corn, bermudagrass, and rye were 86, 143, and 91 kg ha -1 respectively with application of 400 kg N ha -1 and 157, 137, and 169 kg ha -1 respectively, with application of 800 kg N ha -1 For north-Florida conditions, French et al. (1995), presented in Van Horn et al. (1998), found that a crop sequence of perennial peanut (Arachis glabrata Benth.) rye uptakes 430, 470, and 485 kg N ha -1 year -1 receiving 400, 455, and 500 kg N ha -1 year -1 from manure effluent, respectively. The rye forage in this system accumulated 4, 4.7, and 4.5 Mg ha -1 of dry matter containing 60, 80, and 92 kg of N, respectively. Woodard et al. (2002) performed an extensive experiment for four years (1996-2000) in a north-Florida dairy with distinct forage systems under different rates of effluent application (500, 690, and 910 kg N ha -1 year -1 ). They found for these rates of N, the

PAGE 156

134 accumulated dry biomass was 20.5, 20.9, and 21.2 Mg Ha -1 for the bermudagrass, 13.4, 12.6, and 12.8 Mg Ha -1 for the corn, 9.6, 8.7, 9.4 Mg Ha -1 for the sorghum (Sorghum bicolor L.), and 3.4, 3.9, and 4.4 Mg Ha -1 for the rye. The N removal of these forages was 390, 430, and 467 kg N Ha -1 for bermudagrass, 148, 152, 166 kg N Ha -1 for corn, 111, 111, and 120 kg N Ha -1 for sorghum, and 55, 70 and 85 kg N Ha -1 for rye. The rye was in both a bermudagrass-rye system and a corn-sorghum-rye system and it performed similarly in both regarding N uptake and biomass accumulation. For this experiment, lysimeters were installed 1.5 m below the surface and measurements were taken every 14 days: in the bermudagrass-rye system, the NO 3 -N levels never exceeded 10 mg L -1 during the time of the bermudagrass (April-November) but they went above 10, 20 and 30 mg L -1 during the time of the rye (December-March). For the corn-sorghum-rye system, the measurements were much higher: they reach levels of 20 to 40 mg L -1 during corn growth (April-July), 20 to 60 mg L -1 during sorghum growth (August-November), and 30 to 60 mg L -1 during rye growth (December-March). The main aim of the present study was to assess the potential N leaching from forage systems under intensive application of dairy manure in north-Florida. The specific objectives were to estimate The capacity of north-Florida forage systems of accumulate biomass and remove N from the soil The risk of N leaching under different conditions of Climate Crop systems Soils Manure N applications

PAGE 157

135 6.2 Materials and Methods The study was conducted on dairy operations of north-Florida in the Suwannee River Basin, considering their most common forage systems, under the different soils conditions, over a period of 43 years (1956-1998), and with a wide range of manure N application rates by using crop simulation models contained in the Decision Support Systems for Agricultural Transfer, DSSAT v.4.0 Software (Jones et al., 2003). 6.2.1 Survey, Focus Groups, and Additional Information A sample of 21 dairy farms (30% of the population) participated in the interviews. This sample was obtained in cooperation with the Suwannee Partnership and the University of Florida Extension offices in the study area. The sample was intended to cover the variability in north-Florida dairy farm systems with respect to forage systems and their management. Interviews were arranged to be conversational in nature without an instrument, but with a guideline of topics. Interviews lasted between one and two hours and on several occasions, a tour of the farm followed. Conversations were recorded, transcribed, and classified. More general information was obtained by eight focus groups conducted with farmers and other stakeholders such as extension agents, technical people, personnel from government and regulatory agencies, private consultants, among others. Interviews and focus groups were conducted during summer and fall 2003. Additional secondary information was obtained from published studies, dairy farm records, and official records. 6.2.2 Location of North-Florida Dairy Farms and Their Soil Series The study was conducted in dairy operations of Suwannee, Lafayette, Gilchrist, Levy, and Alachua counties in the Suwannee River Basin (21.30 to 30.37 N, and 82.43 to 83.35 W), and included 10 different soil types.

PAGE 158

136 There are 64 dairy farms in the study area: 25 in Lafayette, 19 in Suwannee, 7 in Gilchrist, 7 in Levy, and 6 in Alachua. These were located using the land use survey (1995) of the Suwannee River Water Management District, contained in the Florida Geographic Data Library (1995). Figure 6-1. Study area, dairy farms, and soil types

PAGE 159

137 The soil series for each of the farms were located overlaying it to the soil series maps from the Soils Survey Geographic Database (SSURGO) from the Natural Resource Conservation Service (2002). Figure 6-1 and Table 6-1 present the soils of the studied dairy farm systems, which were summarized in 10 soil types. Table 6-1. Soil types, some characteristics, and their sources of information used for the study Drainage 1 CEC 2 pH 3 Type Series County Rate meq 100 g -1 Survey 1 Arredondo-Gainesville-Millhopper Alachua 0.75 6.0 5.9 Thomas et al., 1985 2 Arredondo-Jonesville-Lake Alachua 0.65 5.0 6.3 Thomas et al., 1985 3 Bonneau-Blanton-Eunola Gilchrist 0.80 6.6 5.6 Weatherspoon et al., 1992 4 Penney-Otela Gilchrist, Lafayette 0.80 3.6 4.9 Weatherspoon et al., 1992 5 Penney-Kershaw Gilchrist 0.75 3.2 4.7 Weatherspoon et al., 1992 6 Millhopper-Bonneau Levy 0.85 7.0 5.0 Slabaugh et al., 1996 7 Otela-Jonesville-Seaboard Levy 0.75 5.5 5.3 Slabaugh et al., 1996 8 Blanton(high)-Lakeland Suwannee 0.60 2.7 5.1 Houston, 1965 9 Blanton(low) Suwannee 0.80 7.4 5.3 Houston, 1965 10 Blanton-Ortega-Penny Lafayette 0.75 5.1 5.3 Weatherspoon et al., 1998 Note: Drainage, CEC, and pH are only for the first soil layer. 1 0.60 (well), 0.75 (somewhat excessive), 0.85 (excessive). 2 Cation Exchange Capacity < 3.0 (extremely low), 3.1-5.0 (very low), 5.1-7.0 (low), 7.1-10.0 (medium). 3 < 5.0 (very strongly acid), 5.1-5.5 (strongly acid), 5.6-6.0 (moderately acid), 6.1-6.5 (slightly acid). The 10 soil types are summarized in Table 6-1, where there is a brief characterization of each of them regarding drainage, cation exchange capacity, and pH. For more

PAGE 160

138 information, there is a reference to the specific soil survey publication. Datasets consisting of several layers of information for each type of soil were collected and organized for the study. These specific data were converted to the format needed by the DSSAT v4.0 system, using SBuild software (Uryasev, 2003), where the soil water holding limits were corrected using Saxton et al. (1986). 6.2.3 Climate Information and El Nio Southern Oscillation (ENSO) Phases Daily weather information was selected from Levy (29.42 N, 82.82 W), located in the central south part of the Suwannee River Basin, between the years 1956 and 1998 (Mavromatis et al., 2002) to be representative for the study area. During this time, 11 years were classified as El Nio years, 10 as La Nia years and 23 as neutral years. Each, El Nio, La Nia, or neutral year is defined as beginning in October and running through September of the next calendar year according to the Japan Meteorological Index (JMI) of sea surface temperature (OBrien et al., 1999). Daily weather information of rainfall, minimum and maximum temperature, and solar radiation between October 1 st (day 274 (275 for leap years)) and September 30 th (day 273 (274 for leap years) were selected, classified, and organized for the different ENSO years. Figures 6.2 to 6.6 show these daily values for rainfall, temperature, and solar radiation as an indication of the weather information used for the simulation. The DSSAT v4.0 crop models were fed with daily information of these four climatic parameters during the 43 years (1956-1998). 6.2.4 Forage Crop Systems Part of the study was devoted to identify and understand forage systems used in north-Florida dairy farms, including fodder plans, management practices, and sequences. Then, they were calibrated and validated using the DSSAT v4.0.

PAGE 161

139 Rainfall in LA NIA years051015202530274286298310322334346358416284052647688100112124136148160172184196208220232244256268Day of the Year (DOY)Rainfall (mm)LA Nia years: 55,56,64,67,70,71,73,75,88,98 Rainfall in NEUTRAL years0.005.0010.0015.0020.0025.0030.00274286298310322334346358416284052647688100112124136148160172184196208220232244256268Day of the Year (DOY)Rainfall (mm)Neutral years: 58,59,60,61,62,66,68,74,77,78,79,80,81,83,84,85,89,90,92,93,94,95,96 A B Rainfall in EL NIO years0.005.0010.0015.0020.0025.0030.00274286298310322334346358416284052647688100112124136148160172184196208220232244256268Day of the Year (DOY)Rainfall (mm) 34.1EL Nio years: 57,63,65,69,72,76,82,86,87,91,97 C Figure 6-2. Daily rainfall (1956-1998) in north-Florida for different ENSO phases. A) La Nia; B) Neutral; and C) El Nio years.

PAGE 162

140 Figure 6-3. Distribution of monthly precipitation (1956-1998) in north-Florida for different ENSO phases 051015202530274286298310322334346358416284052647688100112124136148160172184196208220232244256268Day of the Year (DOY)Rainfall (mm) La Nia El Nio Neutral 34.1 Figure 6-4. Daily precipitation in north-Florida (1956-1998) for different ENSO phases

PAGE 163

141 05101520253035274286298310322334346358416284052647688100112124136148160172184196208220232244256268Day of the Year (DOY)Temperature (C) La Nia El Nio Neutral La Nia years: 55,56,64,67,70,71,73,75,88,98MaximumTemperatureEl Nio years: 57,63,65,69,72,76,82,86,87,91,97MinimumTemperatureNeutral years: 58,59,60,61,62,66,68,74,77,78,79,80,81,83,84,85,89,90,92,93,94,95,96 Figure 6-5. Daily temperature (1956-1998) in north-Florida for different ENSO phases 0510152025274286298310322334346358416284052647688100112124136148160172184196208220232244256268Day of the Year (DOY)Solar Radiation (MJ m-2 d-1) La Nia El Nio Neutral Figure 6-6. Daily solar radiation (1956-1998) in north-Florida for different ENSO phases

PAGE 164

142 6.2.5 Manure N Application Dairy farm fields receive a highly variable amount of manure N depending on herd size, land available, and waste management system. Simulating dynamic cow flows by integrating real data from seasonality, culling rates, reproduction rates, and milk production of north-Florida dairies, collected by the Dairy Herd Improvement Association (FL-DHIA) summarized by De Vries (2004); and survey information, in a Markov-chain model (Chapter 5), monthly rates of manure N received by fields were estimated for all variety of north-Florida dairy farms. These rates varied between 20 and 160 kg ha -1 month -1 Considering that sprayfields usually have two applications in a month, four treatments were arranged at 10, 20, 40, and 80 kg ha -1 application -1 in the DSSAT v4.0 system. 6.2.6. Crop Simulations Forage crop simulations were performed using adapted crop models in the Decision Support System for Agricultural Transfer, DSSAT v4.0 (Jones et al., 2003). These crop models are dynamic process models that simulate crop growth and yield in response to management, climate, and soil conditions. They include light interception, photosynthesis, N uptake, soil water balance, evapotranspiration, respiration, leaf area extension, growth of component parts, root growth, senescence, N mobilization, and crop development processes. For the soil C and N components, the Century model (Parton et al., 1979) implemented in the DSSAT by Gijsman et al. (2002) was used. This model estimates soil N balances that include soil and surface organic matter, inorganic N, additions and removals of N, and all the processes that include the N cycle in the soil, such as decomposition, mineralization, N leaching, etc. Bahiagrass (Paspalum notatum) and bermudagrass (Rymph, 2004) are the only forage crops developed for the DSSAT system, therefore, part of the study consisted in

PAGE 165

143 adapting, calibrating, and validating crop models based on the closest existing models. For corn forage, maize was utilized; for forage sorghum, grain sorghum; for pearl millet (Pennisetum glaucum), grain millet; and for winter forage small grains, wheat (Triticum aestivum). These models were utilized after altering their cultivar coefficients. Calibration and validation against local, actual and current studies followed as well as a sensitivity analysis of the N leaching. These procedures are described in detail in the results section. Forage systems were arranged to mimic dairy farm systems in three growing seasons: spring-summer, summer-fall, and winter. All potential forage combinations were run for four N effluent ranges, ten types of soils found in the study area, and for 43 years of daily weather data (1956-1998). Residual organic matter from one crop to the next one was accounted for as well as other management choices that are common in these systems such as extra irrigation and harvests. 6.2.7 Analyses Daily cumulative N leached (kg ha -1 ) and biomass (kg ha -1 ) outputs from the simulations were compiled in monthly rates for the span of the study period (1956-1998). All months were classified according to ENSO phases and results were summarized by the factors incorporated in the simulations: 12 months x 3 ENSO phases x 4 Manure N applications x 10 soil types x 11 forage combinations. 6.3 Results and Discussion 6.3.1 Forage Crop Systems in North-Florida Dairies Forage crop systems in north-Florida dairies can be summarized by seasons of the year. There are three marked seasons in the north-Florida forage calendar. Spring-summer: from late March or early April to mid July; summer-fall: from late July or early August to

PAGE 166

144 early or mid November; and fall-winter: from late November or early December to mid March (Figure 6-7). 6.3.1.1 Spring-summer crops Three crops were reported during this season in dairy farm fields: corn, bermudagrass, and bahiagrass. Sorghum or pearl millet could also be an option in this season, but were not reported. Corn is planted for silage and always as a part of a sequence of crops. Bermudagrass and bahiagrass are used for either hay, haylage, or grazing and they are also usually part of a sequence of crops. It is possible to over plant bermudagrass or bahiagrass with corn in this season. One to three cuttings are expected for the grasses during this season (two are usual), if they are not grazed. MARAPRMAYJUNJULAUGSEPOCTNOVDECJANFEBMARPERENNIAL PEANUT* **FALL WINTERCORNBERMUDA GRASSBAHIA GRASSSORGHUMMILLETCORNRYEOATSWHEATSORGHUM*MILLET*BERMUDA GRASSBAHIA GRASSSPRING SUMMERSUMMER FALLRYEGRASSCLOVER** Figure 6-7. Forage systems and their seasonality in north-Florida dairy farms. Not found in the interviews, but they are possible. ** Not common and not simulated with DSSAT. 6.3.1.2 Summer-fall crops There are more options to grow during summer-fall. Sorghum and millet are very common, although continuation of grasses from the previous season is also common. Some farmers also grow corn for silage in this season or let the bermudagrass or bahiagrass re-grow, if they were over-planted in spring-summer. Another option, mentioned by only one farmer, is the re-growth of perennial peanut forage during this season. Sorghum, millet, perennial peanut, and the grasses can be used either for hay or haylage or for grazing. One

PAGE 167

145 to three cuttings will be expected for the grasses and one to two cuttings for the sorghum and millet if they are not grazed. 6.3.1.3 Fall-winter crops Winter forages are common in the crop sequences of dairy farms in north-Florida. They are usually small grains or ryegrass; small grains often used are rye, oats (Avena sativa) or wheat. Although bermudagrass and bahiagrass are perennial, they are dormant during this season, where they can be over-seeded or multi-cropped with other species. If there are perennial grasses already established in the field, these winter crops are usually non-till over seeded. Small grains are used for hay or haylage, ryegrass is preferred for silage. Winter small grains could be cut one to four times (two cuts are usual), if they are not grazed. Clover (Trifolium spp.) was also mentioned as an option by one farmer, intercropped with several other grasses. 6.3.1.4 Sequences of forages If bermudagrass or bahiagrass is established and allowed to re-grow in the spring-summer, they will continue growing in the summer-fall season, and no other crops will be possible on the same field until fall-winter season. However, if corn is grown in spring-summer over-planted on one of those grasses, the grass will be allowed to re-grow in the summer-fall season. If corn, sorghum, or millet is planted in spring-summer, any summer-fall crop (from Figure 6-7) will be possible to follow. In fall-winter any small grain or ryegrass could follow any summer-fall crop. If perennial grasses are not established, the most common sequence of forages is silage corn in spring-summer followed by sorghum or millet in the summer-fall, and any small grain or ryegrass in the fall-winter. If grasses are established, a very common

PAGE 168

146 sequence is grass-grass-small grain or ryegrass, for the same seasons. Bermudagrass is much more common than bahiagrass and they both are much more common than perennial peanut. Rye, oats, and wheat are very similar forages and farmers use them indistinctly. Ryegrass was indicated as having better-quality forage, but requires more care and time. It is very common to mix these winter forages in the fields. Farmers try to have a crop in the field at all times, with brief windows between the growing seasons. The regulatory agencies strongly recommend this practice in order to ameliorate the risk of N leaching. 6.3.1.5 Harvest and processing With the exception of corn, all other forages have the option of being grazed directly by the animals. All forages can be cut and processed as hay, haylage, or silage. Hay consists of drying directly in the field. Haylage is an intermediate process of silage performed directly in the field by wrapping up bales with plastic. The silage is a more elaborate process that usually takes place in specially designed facilities. Any forage could follow any of these processes after cutting, but corn and ryegrass are much appreciated for silage and the bermudagrass and bahiagrass are usually intended for hay. 6.3.2 Calibration and Validation of DSSAT Crop Models Models must be calibrated and validated using real data; biophysical models such as crop simulation models contained in the DSSAT v4.0 must be tested to make sure that they represent real plant behavior under specific environmental conditions. Pertinent and trustworthy information from different sources was collected for this purpose. Previous studies, farm records, and survey information were used as baseline information to adjust and compare behavior of these forages on north-Florida dairy farms.

PAGE 169

147 Especially designed experiments for this purpose contain the ideal information against which these models can be calibrated and validated. Thanks to very recent studies (Rymph, 2004), the DSSAT v 4.0 has two forage crops adjusted for Florida conditions, the bermudagrass and the bahiagrass. However, experimental information is not available for the other forage crops; indeed, there are no genotypic experimental coefficients available for them. Fortunately, close relatives for most of these forages exist: for silage corn, grain maize; for forage sorghum, grain sorghum; for pearl millet, millet; for winter forages, wheat. Since the main objective in this study was to assess potential N leaching, biomass accumulation (that determines the N uptake) was used as the driver in the calibration and validation processes. Because the main interest in these crops is their potential use as forages, the cultivar coefficients are not as important and therefore they could be manipulated to emulate a desired behavior (G. Hoogenboom, pers. comm.). And, even when some used crop species are not forages per se, biomass and N uptake are equivalent when species produce or do not produce grain. Consequently, models can be simulated ignoring whether or not they would be producing grain (K. Boote, pers. comm.). The most important data sources used for calibration and validation are listed in Table 6-2. 6.3.2.1 Forage sorghum Woodard et al. (2002) found that forage sorghum between early August and early November will produce between 7.3 and 11.1 Mg ha -1 of dry matter and will uptake between 99 and 150 kg N ha -1 under strong N applications of manure effluents. Environmental conditions of the actual experiment were emulated with the DSSAT v4.0. Soil data of the series Kershaw for Gilchrist County were obtained from the Soil Survey Geographic Database (SSURGO) from USDA, NRCS.

PAGE 170

148 Table 6-2. Information sources for calibration and validation of forage crops in north-Florida dairy farm systems Study/Source Location Weather Soil Series Observations Woodard et al. (2002) North-Florida Holsteins Farm, Inc. Bell. 29.73 N, 82.85 W 1996-1998 Kershaw Several forages: corn, sorghum, rye. Manure effluent applied in rates of 500, 690, and 910 kg N ha -1 year -1 Fontaneli et al. (2000) (2001) Forage Field Evaluation Laboratory. Gainesville. 29.08 N, 82.42 W 1996-1997 Sparr Cool season forages: comparing winter multi-crops. Warm season forages: comparing three varieties of millet and two varieties of sorghum Wright et al. (1993) North-Florida Research and Education Center. Quincy. 30.40 N, 84.27 W 1992 Norfolk Millet and sorghum. Different treatments and purposes Johnson et al. Tifton, GA. 31.43 N, 83.89 W Before 1991 Tifton Corn and rye with different rates of Manure effluent applied Survey north-Florida dairy farms Suwannee River Basin Several years Several types Not always precise information, but real and trustworthy Daily weather data of Levy (29.42 N, 82.82 W) compiled by Mavromatis et al. (2002) were used for the study. Manure effluent was set up to be applied as described in the study: two applications every month containing 21, 29, or 38 kg N ha -1 each one, for a total in a year of 500, 690, and 910 kg N ha -1 The sorghum crop received 6 applications with a total of 132, 174, and 228 kg N ha -1 As described in the study, in addition to the water in the liquid manure applied, extra irrigations were realized, so the crop would have most of the water requirements with some minimal stresses. A new cultivar was created, the forage sorghum, by changing some cultivar coefficients. The final cultivar had the coefficients detailed in Table 6-3. With these coefficients, the validation comparison was realized between the simulated and the experimental data as seen in Figure 6-8. Figure 6-8 shows the comparison between simulated versus observed data. Overall values are inside the expected ranges and they follow the same pattern as the observed data.

PAGE 171

149 0.002.004.006.008.0010.0012.00132 N174 N228 N132 N174 N228 N132 N174 N228 N1996199619961997199719971998199819980.400.060.112.410.011.802.250.140.00 Mg/ha Observed Simulated MSE 2.68 RMSE2.68 Figure 6-8. Forage sorghum: Observed and simulated biomass production, Bell, 1996-1998. Individual mean square error (MES) and overall Root Mean Square Error (RMSE) 6.3.2.2 Pearl millet Fontaneli et al. (2001) presented results of a study carried out at the Forage Evaluation Field Laboratory (24 Km NE of Gainesville, 29.08 N, 82.42 W), where three cultivars of pearl millet and two of sorghum were tested at different planting dates. The experiment realized during spring-summer-fall of 1996 and 1997 revealed that no significant differences exist between millet and sorghum regarding biomass accumulation. Millet would be a higher producer, but very slightly. Also, in both sorghum and millet, the planting date greatly affected the biomass accumulation, which varied between less than 5 Mg ha -1 to nearly 8 Mg ha -1 Also, Chambliss (1997) suggests that millet and sorghum would be very similar forage crops under Florida conditions. Under this premises, sorghum data could be used to validate millet considering the lack of detailed experimental information for this forage.

PAGE 172

150 With similar information on environmental conditions of sorghum, a new cultivar was created, the forage millet, by changing some coefficients. The final cultivar had the coefficients shown in Table 6-3. Additionally, the conversion of solar radiation (in the species file) was set at 0.42, which was originally 0.50, in order to diminish a higher than expected growth. Using these coefficients, the validation comparison was realized between the simulated and the experimental data and results can be seen in Figure 6-9. Table 6-3. Coefficients values and coefficients definitions of modified crops in DSSAT Forage P1 P1V P1D P2 P20 P2R P5 G1 G2 G3 PHINT sorghum 500.0 10.5 90.0 540.0 5.0 6.0 44.0 millet 600.0 10.0 90.0 540.0 5.0 6.0 44.0 Corn 200.0 0.52 940.0 620.0 8.5 38.9 winter forage 47.0 64.0 360.0 28.0 25.0 1.3 80.0 P1: Thermal time from seedling emergence to the end of the juvenile phase (expressed in degree days above a base temperature (10 o C sorghum, 8 o C millet and corn) during which the plant is not responsive to changes in photoperiod. P1V: Relative amount that development is slowed for each day of unfulfilled vernalization, assuming that 50 days of vernalization is sufficient for all cultivars P1D: Relative amount that development is slowed when plants are grown in a photoperiod 1 hour shorter than the optimum (which is considered to be 20 hours). P2: Extent to which development (expressed as days) is delayed for each hour increase in photoperiod above the longest photoperiod at which development proceeds at a maximum rate (which is considered to be 12.5 hours). P20: Critical photoperiod or the longest day length (in hours) at which development occurs at a maximum rate. At values higher than P20, the rate of development is reduced. P2R: Extent to which basic development leading to panicle initiation (expressed in degree days) is delayed for each hour increase in photoperiod above P20. P5: Thermal time (degree days above a base temperature (10 o C sorghum, 8 o C millet and corn, 1 o C winter forages) from beginning of grain filling (3-4 days after flowering) to physiological maturity. G1: Scaler for relative leaf size (sorghum and millet); Kernel number per unit weight of stem (less leaf blades and sheaths) plus spike at anthesis (winter forages) G2:Scaler for partitioning of assimilates to the panicle (sorghum, millet, corn); Kernel filling rate under optimum conditions (winter forages) G3: Kernel filling rate during the linear grain filling stage and under optimum conditions (corn); Non-stressed dry weight of a single stem (excluding leaf blades and sheaths) and spike when elongation ceases (winter forages) PHINT: Phylochron interval; the interval in thermal time (degree days) between successive leaf tip appearances Forage millet was also tested using another set of information from Wright et al. (1993) of a study realized in Quincy, Florida (30.40 N, 84.27 W) during 1992 on soils of the series Norfolk. An experiment in DSSAT v4.0 to try to mimic the experimental conditions was set up. Even when this experiment was done outside of the boundaries of the study (soils are very different), a more general comparison could be achieved providing

PAGE 173

151 more support for the previous validation. Experimental data (with a variety of treatments) showed a range of millet production of biomass under these conditions between 12.5 and 14.4 Mg ha -1 Figures 6-9 and 6-10 show the results of the observations and simulations. Overall values are inside the expected ranges and they follow the same pattern as the observed data. 0.002.004.006.008.0010.0012.00132 N174 N228 N132 N174 N228 N132 N174 N228 N1996199619961997199719971998199819980.361.001.216.760.040.816.760.640.16 Mg/ha Observed Simulated MSE17.74 RMSE4.21 Figure 6-9. Forage millet: Observed and simulated biomass production, Bell, 1996-1998. Individual mean square error (MES) and overall Root Mean Square Error (RMSE) 6.3.2.3 Silage corn Woodard et al. (2002) found that corn forage grown between early April and mid July will produce between 10.8 and 15.2 Mg ha -1 of dry matter and will uptake between 119 and 200 kg N ha -1 under strong N applications of manure effluents. Environmental conditions of the actual experiment were emulated as in the case of forage sorghum and millet. Manure effluent was set up to be applied as described in the study: the corn crop received 7 applications with a total of 147, 203, and 266 kg N ha -1

PAGE 174

152 0.002.004.006.008.0010.0012.0014.0016.0025-Jun05-Jul15-Jul25-Jul04-Aug14-Aug24-Aug03-Sep13-SepDate Simulated Observed Figure 6-10. Forage millet: Observed and simulated biomass production, Quincy 1992 A new cultivar was created, the silage corn, on which the calibration was performed. The final cultivar had the coefficients summarized in Table 6-3. Using these coefficients, the validation comparison was realized between the simulated and the experimental data as can be seen in Figure 6-11. Overall values are inside the expected ranges and follow the same pattern. 6.3.2.4 Winter forages Farmers refer to winter forages as a group of forages that include the small grains and ryegrass. Small grains are commonly wheat, rye, and oats; rye is the most common. It is usual that farmers mix these crops in winter with ryegrass. Using information provided during the interviews these forages can be simulated as a single crop. DSSAT v4.0 has the wheat coefficients, which were used as the baseline to simulate winter forages.

PAGE 175

153 0.002.004.006.008.0010.0012.0014.0016.0018.00147 N203 N266 N147 N203 N266 N147 N203 N266 N1996199619961997199719971998199819980.0221.3611.900.0913.488.3116.520.190.02 Mg/ha OBSERVED SIMULATED MSE71.89 RMSE8.48 Figure 6-11. Silage corn: Observed and simulated biomass production, Bell, 1996-1998. Individual mean square error (MES) and overall Root Mean Square Error (RMSE) Woodard et al. (2002) found that rye forage produced between late November and mid March will produce between 2.7 and 5.0 Mg ha -1 of dry matter and will uptake between 46 and 89 kg N ha -1 under strong N applications of manure effluents. Environmental conditions of the actual experiment were emulated with the DSSAT v4.0 as in previous cases. Manure effluent was set up to be applied as described in the study; the small grain forages received 7 applications with a total of 147, 203, and 266 kg N ha -1 Biomass production was compared to the years 1996/97 and 1997/98. A new cultivar was created, the winter forage, upon which the calibration was performed. The final cultivar had coefficients that are shown in Table 6-3. Using these coefficients, the validation comparison was realized between the simulated and the experimental data. Figure 6-12 shows the results of such a comparison. Overall values are inside the expected ranges and follow the same pattern as the observed data.

PAGE 176

154 0.001.002.003.004.005.006.007.00147 N203 N266 N147 N203 N266 N1996/971996/981996/991997/981997/991997/1001.270.220.031.170.000.47 Mg/ha Observed Simulated MSE3.16 RMSE1.78 Figure 6-12. Winter forages: Observed and simulated biomass production, Bell 1997-1998. Individual mean square error (MES) and overall Root Mean Square Error (RMSE) 6.3.3 Sensitivity Analysis of Forage Crops to Manure N Application A sensitivity analysis of biomass accumulation, N uptake, and N leaching was conducted in which extreme values of manure N were used on the crops to be used for this study. Six rates of manure N were tested in all crops 0, 10, 20, 40, 60, 80, and 100 kg per application (2 applications per month), or 0, 240, 480, 960, 1440, 1920, and 2400 kg year -1 Irrigation was arranged as usual in dairy farms and it was tested for 1996/1997 (a Neutral climatic year where results from previous calibration and validation are known) and for Kershaw soils as in the Woodard et al. (2002) experiment. Corn was planted on April 4 th and harvested on July 16 th and received 7 manure N applications. Sorghum was planted on August 14 th and harvested on November 13 th 1996. During this period the crop received 6 manure N applications. Winter forage was planted on December 3 rd and harvested on March 11 th 1997. During this time the crop received 7 manure N applications. Millet was planted on August

PAGE 177

155 14 th and harvested on November 13 th 1997 and received 6 manure N applications. Bahiagrass and bermudagrass were grown from March 30 th to November 16 th 1996 and received 22 manure N applications. 0.002.004.006.008.0010.0012.0014.0016.0018.0020.00070140280420560700MANURE N APPLIED (Kg Ha-1)BIOMASS (Mg Ha-1)01020304050607080N LEACHED (Kg Ha-1) BIOMASS N LEACHED CORN 0.001.002.003.004.005.006.007.008.00060120240360480600MANURE N APPLIED (Kg Ha-1)BIOMASS (Mg Ha-1)0510152025N LEACHED (Kg Ha-1) BIOMASS N LEACHED SORGHUM 0.001.002.003.004.005.006.007.008.00070140280420560700MANURE N APPLIED (Kg Ha-1)BIOMASS (Mg Ha-1)0510152025303540N LEACHED (Kg Ha-1) BIOMASS N LEACHED WINTER FORAGE 0.001.002.003.004.005.006.007.008.009.0010.00060120240360480600MANURE N APPLIED (Kg Ha-1)BIOMASS (Mg Ha-1)0102030405060N LEACHED (Kg Ha-1) BIOMASS N LEACHED MILLET 0.001.002.003.004.005.006.007.008.009.0010.00015030060090012001500MANURE N APPLIED (Kg Ha-1)BIOMASS (Mg Ha-1)0102030405060708090100N LEACHED (Kg Ha-1) BIOMASS N LEACHED BAHIA 0.002.004.006.008.0010.0012.0014.00015030060090012001500MANURE N APPLIED (Kg Ha-1)BIOMASS (Mg Ha-1)-551525354555N LEACHED (Kg Ha-1) BIOMASS N LEACHED BERMUDA F A B C D E Figure 6-13. Sensitivity response of N leaching and biomass accumulation to different rates of manure N applied for simulations in a Kershaw soil series during the season 1996/1997. A) corn; B) sorghum; C) winter forage; D) millet; E) bahiagrass; and F) bermudagrass. As expected, biomass accumulation increases with manure N application, but at a decreasing rate as it reaches a maximum. There is no noticeable effect of toxicity with the high amounts of N. According to the parameters of the soil there is some organic matter

PAGE 178

156 that would provide some N at the beginning, that is why in the zero N application a low but not extremely low biomass and uptake are expected, similarly to the N leached. N leaching increased faster with higher manure N application, consistent for all crops. 6.3.4 Variability by ENSO Phases N leaching increased and biomass accumulation decreased from La Nia years to Neutral and to El Nio years. Lowest N leaching (highest biomass accumulation) was always observed in La Nia years and highest N leaching (lowest biomass accumulation) was always observed in El Nio years. This fact was consistent for different type of soils, for different forage rotations, and for different manure N applications. As an example, a trade-off of N leaching vs. biomass accumulation for two forage systems, two type of soils, and four manure N applications is shown in Figure 6-14. Figure 6-14 shows, for different ENSO phases, a similar increasing trend for both N leaching and biomass accumulation, responding to the increase of manure N application. At the beginning, this N leaching trend was smooth and slow, but it jumped up abruptly when the manure N changed from 40 to 80 kg ha -1 mo -1 Something similar, although less marked happened with the biomass accumulation change from these manure N levels. Different from the sensitivity analysis, it is deduced from these graphs that biomass accumulation had not yet reached a peak with this maximum manure N application. Notice that graphs in Figure 6-14 have different scales for the N leaching in order to show the trend, even though substantial differences in N leaching existed between the soil types and the forage systems. Differences between graph A) and B) are attributed exclusively to the type of soils and differences between graphs A) and C) are exclusively because forage system.

PAGE 179

157 42312413132420022024026028030012,00015,00018,00021,00024,00027,00030,00033,000 LA NIA NEUTRAL EL NIO Biomasskg/year N leachingkg/year Manure N applied kg/ha-mo 1 102 203 404 80 FORAGE SYSTEM: bermudagrass-bermudagrassSOIL SERIE 1: Arredondo-Gainesville-Millhopper FORAGE SYSTEM: bermudagrass-bermudagrassSOIL TYPE 1: Arredondo-Gainesville-Millhopper ABC 29012,00015,00018,00021,00024,00027,00030,00033,000 Biomass kg/year 42312413132427029031033035037039041012,00015,00018,00021,00024,00027,00030,00033,000 Biomass kg/year N leaching kg/year FORAGE SYSTEM: millet-sorghumSOIL SERIE 6: Arredondo-Gainesville-Millhopper FORAGE SYSTEM: sorghum-millet SOIL TYPE 1: Arredondo-Gainesville-Millhopper 132431424231340390440490540 N leaching kg/year FORAGE SYSTEM: bermudagrass-bermudagrassSOIL SERIE 6: Millhopper-Bonneau FORAGE SYSTEM: bermudagrass-bermudagrass SOIL TYPE 6: Millhopper-Bonneau Figure 6-14. Nitrogen leaching vs. biomass accumulation for different ENSO phases under two forage systems consisting of bermudagrass and millet-sorghum, four manure N applications (10, 20, 40, and 80 kg ha -1 mo -1 ), and two type of soils (Arredondo-Gainesville-Millhopper and Millhopper-Bonneau)

PAGE 180

158 The soil type 6 (Millhopper-Bonneau) showed substantially higher N leaching than the soil type 1 (Arredondo-Gainesville-Millhopper). It would be expected that soil type 6 has lower biomass accumulation because more N is being leached and consequently not up-taken by the crops, however higher biomass accumulation was also observed in soil type 6. It seems that low retentive and shallow soils such as soil type 6 provide better forage growth conditions (under these conditions of high nutrients) together with a higher risk of N leaching. A forage system consisting of bermudagrass-bermudagrass leaches substantially less N than sorghum-millet, but sorghum-millet accumulates substantially higher biomass amounts. The rate of N leaching increasing for Neutral years is distinctive from the other two phases when sorghum-millet rotation is present; this is more marked between 40 and 80 kg ha -1 mo -1 Nevertheless, the pattern of highest N leaching during El Nio years and lowest during La Nia years persists. Notice also that the magnitude of change between manure N application treatments in soils type 6, compared with soils type 1 and in rotation bermudagrass-bermudagrass, compared with rotation sorghum-millet varies considerably, indicating that there is interaction among ENSO phases, manure N application, soil types and forage rotations. 6.3.4.1 Nitrogen leaching in different ENSO phases Substantially higher (~10%) N leaching (kg ha -1 year -1 ) was predicted for El Nio (357.25) than for La Nia years (322.29). For Neutral years expected N leaching was between La Nia and El Nio years (342.90), very close to the overall average (341.84), Table 6-4. Winter months, especially January and February, are critical because the N leached in these two months represent more than 50% of the total N leached in a year. Significantly (p<0.05) higher N leaching (Coefficient of Variability, CV) was predicted in

PAGE 181

159 January for El Nio years 127.35 kg ha -1 (21%) when compared to either Neutral 108.9 kg ha -1 (30%) or La Nia years 83.19 kg ha -1 (35%) and also significant (p<0.05) differences were predicted when comparing La Nia with Neutral years in this month. For February, more N leaching (CV) was expected for La Nia years 84.22 (9%) than El Nio years 71.6 (20%), with significant differences (p<0.05). Even though no more significant differences were found in other months, there are substantial differences among ENSO years and different months that can be appreciated in Figure 6-15. Figure 6-15 shows the relative variation with respect to the average of different ENSO phases: El Nio phase presented more overall variability through the year and maximum N leaching in the months of December, January, August, September, and November. The Neutral phase presented maximum N leach amounts during the months of April, May, June, and October; and for La Nia phases more N leach was expected in the months of February, March, and July. The highest N leaching in a month was matched with specific years in order to explain the driving factors. The three highest rates in January in El Nio years (187, 152, and 140 kg ha -1 year -1 ) were found to happen in the years 1964, 1987, and 1970, when the Japan Meteorological Index (JMI) (Center for Ocean-Atmospheric Prediction Studies, 2003) indicated ocean temperature of 0.7, 1.0, and 0.8C, respectively. For Neutral years, the three highest rates were 173, 158, and 156 kg ha -1 year -1 in the years 1979, 1991, and 1994 with JMI of 0.0, 0.4, and 0.2. In La Nia years the highest amounts were substantially lower (108, 107, and 107 kg ha -1 year -1 ) and they happened in 1965, 1972, and 1971, with the JMI of -0.8, -0.7, and -1.11. An attempt to correlate these JMI indexes with the N leached failed and corroborates what previously was found: that there is no significant

PAGE 182

160 relationship of these indexes either with biomass accumulation or N leaching under the study parameters. The same happened when comparing the highest JMI, for example 2.7 or 3.1 for El Nio years that indicated N leaching less than the highest amounts (86 and 117 kg ha year). -1 -1 Table 6-4. Nitrogen leaching (kg ha) and ENSO phases, when all other factors are averaged -1 All Years La Nia Neutral El Nio N Leach CV N Leach CV N Leach CV N Leach CV Dec 10.59 98% 13.15 93% 7.69 15.64 127% Jan 108.42 80% 32% 83.19* 35% 108.90* 30% 127.35* 21% Feb 78.57 24% 84.22* 9% 79.94* 29% 71.60* 20% Mar 20.86 49% 23.54 56% 20.55 42% 19.28 45% Apr 5.81 97% 4.90 78% 6.42 71% 5.45 70% May 12.18 12.39 103% 12.94 59% 10.61 72% Jun 23.22 43% 22.01 45% 25.13 38% 20.69 50% Jul 16.65 44% 18.42 33% 17.09 48% 14.33 37% Aug 17.05 40% 16.99 40% 15.65 38% 19.71 36% Sep 21.22 39% 19.34 44% 20.33 34% 24.50 39% Oct 20.27 33% 19.02 35% 20.79 36% 20.35 21% Nov 7.00 59% 5.11 42% 7.47 58% 7.75 62% Year 341.84 322.29 342.90 357.25 73% Significant differences (p<0.05). CV is coefficient of variation. It was found, as suspected, that N leaching is highly related to precipitation associated with climatic ENSO phases: for most of the months the highest precipitation by ENSO phase presented the highest N leached (Figure 6-3), however it does not happen for February, March, and June. For example, the mean rainfall for January in the study area was 104 mm, and varied for El Nio years (148 mm), for Neutral years (111 mm), and for La Nia years (54 mm) explaining the variation in N leached. Also, matching the highest N leaching with rainfall, we found that the N leached will be highly related to the total amount of rainfall and to the occurrence of extreme rainfall events. For the highest El Nio events the rainfall amounts were 315, 194, and 140 mm; for the highest Neutral years, these were 269, 242, and 239 mm; and for the highest La Nia

PAGE 183

161 years these were 118, 101, 70 mm, respectively, These rainfall amounts were substantially higher than normal, and in the cases when this not happened, as in the third highest El Nio and third highest La Nia, extreme rainfall events occurred which had high amounts of rainfall concentrated in small time periods (e.g., rainfall >20 mm in a day). For February and March, higher leaching in El Nio years would be expected because considerably more precipitation is observed, but more N leaching was predicted for La Nia years. Possible causes would point to the rainfall distribution, other climatic factors such as temperature and solar radiation, and soil N dynamics; during February and March, in La Nia years, the precipitation patterns indicate that the rainfall events were intense (although not as much as El Nio) and infrequent. This could cause the water in the soil to be highly variable (more than in El Nio and Neutral years) because there are drought and flooding periods (different from other years when water is permanent) making the system more unstable and consequently more prone to N leaching. Other possible causes are the higher temperatures and higher solar radiation observed during these months for La Nia phases; and even though both of these factors will induce higher plant growth rates and consequently higher N uptake by plants, it seems that they may increase even more the decomposition and mineralization processes of the N inducing its leaching. Additionally, the dynamics of N in the soil would influence this behavior, meaning that the high amount of N leached in January in El Nio years would leave less N prone to leach the following months and that the relatively lower amount leached in La Nia years during January, leaves more N prone to leach in February. In June, when more N leaching would be expected for La Nia years, substantially more N leaching is predicted for Neutral, followed of El Nio years and the only possible explanation is the rainfall distribution and intensity of the events, because

PAGE 184

162 there are no apparent differences in temperature and solar radiation. Effectively, in Neutral years there are very intense events, just days before and during the month of June. 0.600.700.800.901.001.101.201.301.401.50DECJANFEBMARAPRMAYJUNJULAUGSEPOCTNOVRELAVIVE N LEACHING La Nia Neutral El Nio Figure 6-15. Nitrogen leaching vs. ENSO phases relative to absolute averages, when all other factors are averaged In order to analyze the interactions among N leaching, month of the year and ENSO phase, cumulative probabilities were estimated and plotted (Figure 6-16). In Figure 6-16, the lines are cumulative probabilities of N leaching for the ENSO phases; the Y-axis represents the percent or probability of occurrence. If for example, we would like to know the chance that there will be at most 50 kg ha -1 N leached in January, 50 kg ha -1 in February, and 20 kg ha -1 in all the other months, we would find that: having as little as 50 kg of N leached per ha is very unlikely in the months of January and February; indeed only 25% of the time in La Nia and only 5% of the time in January in Neutral years it would occur. This goal would be reached for February only

PAGE 185

163 10% of the time in El Nio and Neutral phases, respectively. There is no chance for El Nio in January and for La Nia in February that the leaching be less than 50 kg ha -1 For April, August, September, and November the N leaching would always be lower than 20 kg ha -1 independently of the ENSO phase, however for the remaining months the chances would vary depending on climatic conditions. 02040608010002040608 0 Neutral La Nia El Nio DECEMBER % kg/ha 020406080100050100150200 JANUARY % kg/ha 020406080100050100150 FEBRUARY % kg/ha 02040608010002040 60 MARCH % kg/ha 020406080100024 6 A PRIL % kg/ha 02040608010001020 30 MAY % kg/ha 02040608010001020 30 JUNE % kg/ha 02040608010001020 30 JULY % kg/ha 02040608010005101520 A UGUST % kg/ha 0204060801000510 15 SEPTEMBER % kg/ha 0204060801000510 15 OCTOBER % kg/ha 02040608010005 10 NOVEMBER % kg/ha Figure 6-16. Probability distribution of N leaching by ENSO phases, when all other factors are averaged

PAGE 186

164 In December, there is very little chance that the leaching exceeds 20 kg; in La Nia years there would be an 18% chance, in El Nio years a 10% chance, and in Neutral years less than 5%. Similarly for March, the probabilities of having more than 20 kg leached would be 50%, 40%, and 30% for Neutral, La Nia, and El Nio years, respectively. Another way to analyze the interaction among N leaching, month of the year and ENSO phase is through exploring monthly N leaching distributions for El Nio, Neutral, and La Nia years (Figure 6-17). First, for most of the months the distributions were skewed to either the left or the right, or they were far from being normal, with the exception of June and July. For Neutral years, it was consistent to observe, in all months, that its distribution was sparse throughout all the values and it was close to be a normal distribution, meaning that almost any value can be seen in Neutral phases. For December, for any ENSO phase, highest will be the chance that N leaching were between the ranges 3 to 17.1 kg ha -1 however there was a chance that this would be higher than 24 kg ha -1 ; and it would be about 25% for La Nia years, 10% for El Nio years, and 5% for Neutral years. In January where highest amounts of leaching are expected for all phases, in La Nia years the leaching would be at most 112.3 kg ha -1 with a 50% chance of being even lower than 81.9. For neutral years the distribution indicates that most of the time the amount leached would be between 112.3 and 142.6 kg ha -1 and for El Nio years it would have more than 70% chance to be above 142.6 kg ha -1 For February, another high N leaching month, in La Nia years there would be more than 81.5 3 kg ha -1 and a chance of more than 70% that this would be even 99.7 kg ha -1 of N leaching or higher. For El Nio years there is a chance of more than 25% that the leaching would be lower than 63.3 kg ha -1 but about another 25% of chance that it would be 99.7 kg

PAGE 187

165 ha -1 or higher. Neutral ENSO phase for February presents the sparsest distribution; the leaching could be much lower, but it could also be also much higher than La Nia or El Nio years. DECEMBER0204060801003.010.017.124.2More Neutral La Nia El Nio % JANUARY02040608010051.681.9112.3142.6More % FEBRUARY02040608010045.163.381.599.7More % MARCH0204060801008.617.326.034.7More % APRIL0204060801000.11.22.23.2More % MAY0204060801000.94.68.412.2More % JUNE0204060801006.210.915.520.1More % JULY02040608010010.312.915.518.0More % AUGUST0204060801003.76.08.410.7More % SEPTEMBER0204060801002.54.36.28.1More % OCTOBER0204060801003.35.67.910.1More % NOVEMBER0204060801001.43.45.37.3More % Figure 6-17. Frequency distribution of N leaching and month of the year by ENSO phase, when all other factors are averaged

PAGE 188

166 6.3.4.2 Biomass in different ENSO phases Biomass accumulation by forage systems changes by season and crop. The highest accumulations were in spring-summer months (April-July), followed by the summer-fall months (August-November). While in spring, summer and fall a variety of forages were possible, in winter only small grains (rye-ryegrass and relatives) are possible. The biomass accumulation (CV) varied from 86 kg ha -1 (32%) in December to 4,908 kg ha -1 (13%) in June. Biomass accumulation (CV) was higher for La Nia years and lower during El Nio years compared to Neutral, Table 6-5. Table 6-5. Biomass (kg ha -1 ) and ENSO phases, when all other factors are averaged All Years La Nia Neutral El Nio Biomass CV Biomass CV Biomass CV Biomass CV Dec 86 32% 102 30% 77 28% 92 41% Jan 1,367 17% 1,563* 13% 1,319* 14% 1,301* 22% Feb 2,292 10% 2,343* 11% 2,346* 7% 2,151* 9% Mar 1,439 28% 1,329* 23% 1,561* 30% 1,297* 14% Apr 959 35% 1,094 28% 938 32% 883 27% May 4,546 12% 4,691 7% 4,487 16% 4,533 7% Jun 4,908 13% 4,997 9% 4,818 19% 4,997 9% Jul 3,232 13% 3,280 9% 3,173 18% 3,300 8% Aug 1,649 27% 1,660 16% 1,654 20% 1,631 28% Sep 2,784 11% 2,821 10% 2,811 10% 2,704 12% Oct 3,127 12% 3,211 15% 3,060 11% 3,180 9% Nov 811 49% 833 63% 754 51% 899 44% Year 27,202 27,924 26,999 26,968 Higher biomass (kg ha -1 year -1 ) was predicted for La Nia years (27,924) than for Neutral (26,999) or El Nio years (26,968), Table 6-5. Most of the months consequently were predicted to have higher biomass accumulation (CV) in La Nia years, however for February and March the prediction indicates that they would be slighter higher in Neutral years, 2346 (7%), and 1561 (30%). Similarly for June, the accumulation (CV) would be close for La Nia and El Nio years 4,997 (9%), and in July it would be slightly higher for El Nio years, 3,300 (8%), Figure 6-18. This indicates that the accumulation is sensitive to

PAGE 189

167 the ENSO phases but not in the same way as it is with the N leaching. Overall, the crops did not respond to higher rainfall with higher yields; indeed higher biomass accumulation was seen with low or medium rainfall patterns for winter (La Nia and Neutral), for spring-summer (La Nia and El Nio), and also for summer-fall (La Nia). Temperature and radiation also affected the accumulation of biomass. In La Nia years there are consistently higher temperatures and more radiation in the last months and in the first months of a year (fall-winter) coinciding with the higher accumulation rates. 0.800.850.900.951.001.051.101.151.20DECJANFEBMARAPRMAYJUNJULAUGSEPOCTNOVRELAVIVE BIOMAS S La Nia Neutral El Nio Figure 6-18. Biomass accumulation vs. ENSO phases relative to absolute averages, when all other factors are averaged Biomass accumulation in Neutral and El Nio years was significantly (p<0.05) lower than La Nia in January, El Nio events were significantly lower (p<0.05) than Neutral and La Nia years in February and March, respectively, and El Nio events were significantly lower (p<0.05) than Neutral years in March.

PAGE 190

168 6.3.5 Variability by Crop Sequences Crop sequences performed substantially different regarding N leaching and biomass accumulation, Figure 6-19. For example, forage system number 11 (bermudagrass-bermudagrass-winter forage) consistently showed the lowest N leaching rate, but low biomass accumulation. The system bahiagrass-bahiagrass-winter forage showed the lowest biomass accumulation, independently of the ENSO phase and the amount of manure N application rate. Rotations consisting of sorghum-millet and millet-sorghum presented the highest N leaching rates with medium biomass accumulation. Rotations that included corn-corn and corn-millet showed the highest rates of biomass accumulation with medium levels of N leaching. And rotations of corn-millet and sorghum-corn showed a medium N leaching together with a medium biomass accumulation. When corn was followed by bermudagrass and/or bahiagrass, the N leaching was low and the biomass accumulation was medium. As in previous cases, it was consistent that higher N leaching (lower biomass accumulation) occurred in El Nio years and lower N leaching (higher biomass accumulation) in La Nia years. By comparing the two graphs in Figure 6-19, the effect of higher N application (40 to 80 kg ha -1 mo -1 ) can be observed. The measurements moved to the upper-right, indicating that with higher N applications, higher N leaching is expected together with higher biomass accumulation. Three facts deserve mention when comparing graphs A) and B) in Figure 6-19. The bahiagrass-bahiagrass sequence seemed to change drastically with the N application change, showing a higher vulnerability to N leaching under high pressure of N applications; this rotation was one of the lowest N leaching with 40 kg ha -1 mo -1 and it

PAGE 191

169 became the highest with an application of 80 kg -1 ha -1 The bermudagrass-bermudagrass rotation was the most resilient of all to N leaching under higher N applications and the most preventive for N leaching; even though all the other rotations (except bahiagrass-bahiagrass) increased proportionally from Graph A) to B), the bermudagrass rotation increased its N leaching but considerably less than the others. And, the rotation millet-corn responded more than the others with respect to biomass accumulation becoming the highest 55112763918451169374182281493671124029034039044049014,00019,00024,00029,00034,00039,00044,00049,000 LA NIA NEUTRAL EL NIO Biomasskg/ha-yr N leachingkg/ha-yr Crop systems: 1 bahiagrass-bahiagrass2 corn-bermudagrass3 corn-bahiagrass4 corn-millet5 corn-corn6 sorghum-millet7 millet-sorghum8 sorghum-corn9 millet-corn10-corn-sorghum11-bermudagrass-bermudagrass SOIL TYPE 5: Penney-KershawMANURE N: 40 kg/ha-mo SOIL TYPE 5: Penney-Kershaw MANURE N: 40 kg ha-1 mo-1 AB 55112763918451169374182281493671124029034039044049014,00019,00024,00029,00034,00039,00044,00049,000 Biomasskg/ha-yr N leachingkg/ha-yr SOIL TYPE 5: Penney KershawMANURE N: 80 kg/ha-mo SOIL TYPE 5: Penney-Kershaw MANURE N: 80 kg ha-1 mo-1 Figure 6-19. Nitrogen leaching vs. biomass accumulation for different forage systems under different ENSO phases (La Nia, Neutral, and El Nio), two manure N applications (40 and 80 kg ha -1 mo -1 ), and soils of type Penney-Kershaw

PAGE 192

170 in biomass accumulation after the higher manure N application. It seemed that the millet accumulation rate increases faster than the corn with higher N applications. Additionally, the magnitude of the changes (the spatial positions) of the measurements between graphs in Figure 6-19 indicates high interactions among ENSO phases, crop rotations, and N application rates. 6.3.5.1 Nitrogen leaching and crop systems Overall N leaching (kg ha -1 ) in an entire year varied from 281 (bermudagrassbermudagrass) to 397 (sorghum-millet), Table 6-6. 0.000.501.001.502.002.50MARAPRMAYJUNJULAUGSEPOCTNOVRELATIVE N LEACH BE-BE CO-SO MI-CO SO-CO MI-SO SO-MI CO-CO CO-MI CO-BA CO-BE BA-BA Figure 6-20. Nitrogen leaching vs. crop sequences relative to absolute averages, when all other factors are averaged. Note: abbreviations are two first letters of rotations in Table 6-6. N leaching was predicted to be low when the rotation starts in spring with bermudagrass or corn, followed in the summer by bermudagrass, corn or bahiagrass; it was medium if it started with bahiagrass or corn and if they were followed by either bahiagrass, sorghum or millet; and it was high if the sequence started with millet or sorghum and continued with corn, sorghum, or millet (Figure 6-20).

PAGE 193

171 Table 6-6. Nitrogen leaching (kg ha -1 ) and crop sequences, when all other factors are averaged bermudagrass corn millet sorghum millet sorghum bermudagrass sorghum corn corn sorghum millet N Leach CV N Leach CV N Leach CV N Leach CV N Leach CV N Leach CV Mar 19.44 51% 19.57 52% 22.59 45% 23.77 43% 22.59 45% 23.77 43% Apr 1.21 146% 2.10 114% 12.30 50% 13.16 51% 12.30 50% 13.16 51% May 6.13 91% 9.06 77% 18.99 59% 19.10 60% 18.99 59% 19.10 60% Jun 14.77 42% 19.94 41% 29.57 45% 32.51 46% 29.57 45% 32.51 46% Jul 15.50 22% 15.06 42% 14.69 60% 19.52 52% 14.04 63% 18.87 54% Aug 9.83 30% 19.76 42% 15.85 43% 15.85 43% 19.76 42% 19.68 42% Sep 6.43 39% 31.31 43% 17.35 36% 17.35 36% 31.31 43% 33.55 42% Oct 6.52 35% 29.98 30% 17.44 37% 17.44 37% 29.98 30% 31.57 29% Nov 3.12 57% 7.42 71% 9.20 55% 9.20 55% 7.42 71% 6.86 62% Year 281 352 356 365 384 397 Table 6-6. Continued corn corn corn corn bahiagrass corn millet bahiagrass bermudagrass bahiagrass N Leach CV N Leach CV N Leach CV N Leach CV N Leach CV Mar 19.57 52% 19.57 52% 19.57 52% 19.57 52% 19.44 51% Apr 2.10 114% 2.10 114% 2.10 114% 2.10 114% 1.23 146% May 9.06 77% 9.06 77% 9.06 77% 9.06 77% 6.42 93% Jun 19.94 41% 19.94 41% 19.94 41% 19.94 41% 16.74 45% Jul 15.71 41% 15.06 42% 15.94 41% 15.93 41% 22.88 26% Aug 15.85 43% 19.68 42% 15.13 38% 14.69 38% 21.50 33% Sep 17.35 36% 33.55 42% 15.64 37% 14.23 35% 15.42 41% Oct 17.44 37% 31.57 29% 14.24 33% 10.82 30% 15.91 34% Nov 9.20 55% 6.86 62% 6.17 53% 3.93 51% 7.59 56% Year 324 355 315 308 325 6.3.5.2 Biomass and crop systems Overall biomass accumulation (kg ha -1 year -1 ) varied from 14,061 kg ha -1 for bahiagrass-bahiagrass to 39,099 kg ha -1 for the corn-corn system. Also on the high side were the sorghum and millet crops and on the low side bermudagrass, Table 6-7. All crops performed better in the springsummer, when higher accumulation was predicted, June being the single month with the highest biomass accumulated. On a monthly basis, kg ha -1 (CV), the corn substantially outperformed all the others, reaching an

PAGE 194

172 accumulation of 6,700 kg ha -1 (10%) in June or 5,873 kg ha -1 (7%) in September. Millet followed the corn, which could accumulate up to 4,996 kg ha -1 (20%) in the month of June or 4,690 kg ha -1 (20%) in September. A little lower was the sorghum with 4599 kg ha -1 (22%) accumulated in June or 3,529 kg ha -1 (7%) in September. After those crops, bermudagrass and bahiagrass followed. Table 6-7. Biomass (kg ha -1 ) and crop sequences, when all other factors are averaged Bermudagrass Corn Millet Sorghum Millet Sorghum Bermudagrass Sorghum Corn Corn Sorghum Millet Biomass CV Biomass CV Biomass CV Biomass CV Biomass CV Biomass CV Mar 1,435 27% 1,435 27% 1,437 29% 1,447 29% 1,437 29% 1,447 29% Apr 2,078 19% 349 33% 1,441 30% 925 33% 1,441 30% 925 33% May 3,159 10% 6,450 7% 4,788 19% 3,061 18% 4,788 19% 3,061 18% Jun 2,129 6% 6,700 10% 4,996 20% 4,599 22% 4,996 20% 4,599 22% Jul 1,966 6% 3,214 7% 3,565 20% 3,514 19% 3,509 20% 3,458 19% Aug 1,641 7% 111 73% 5,002 8% 5,005 8% 111 26% 33 92% Sep 1,116 15% 3,058 6% 5,809 8% 5,873 9% 3,058 7% 2,685 22% Oct 775 18% 3,529 7% 4,803 10% 4,919 15% 3,529 12% 4,690 20% Nov 229 34% 1,112 26% 540 87% 582 87% 1,112 55% 1,471 51% Year 18,274 29,703 36,125 33,669 27,726 26,114 Table 6-7. Continued Corn Corn Corn Corn Bahiagrass Corn Millet Bahiagrass Bermudagrass Bahiagrass Biomass CV Biomass CV Biomass CV Biomass CV Biomass CV Mar 1,435 27% 1,435 27% 1,435 27% 1,435 27% 1,435 27% Apr 349 33% 349 33% 349 33% 349 33% 1,419 24% May 6,450 7% 6,450 7% 6,450 7% 6,450 7% 2,304 9% Jun 6,700 10% 6,700 10% 6,700 10% 6,700 10% 1,280 7% Jul 4,042 10% 3,214 11% 4,042 10% 4,487 9% 1,139 6% Aug 5,005 8% 30 59% 2,078 7% 3,084 8% 1,014 5% Sep 5,873 9% 2,621 17% 1,065 8% 1,793 8% 868 10% Oct 4,919 15% 4,574 11% 844 9% 1,404 8% 663 15% Nov 582 87% 1,430 46% 288 19% 481 17% 195 30% Year 39,099 30,547 26,994 29,927 14,061 Even though bermudagrass and bahiagrass were in the lower end of biomass accumulation, when they were associated with corn in the spring-summer season, they

PAGE 195

173 could accumulate amounts close to or above the overall average. Another distinction about bahiagrass and bermudagrass was that they maintain a smoother rate of accumulation throughout their growing months. This was different from the other crops that present more variability during the growth season. 0.000.501.001.502.002.503.003.50MARAPRMAYJUNJULAUGSEPOCTNOVRELATIVE BIOMAS S BE-BE CO-SO MI-CO SO-CO MI-SO SO-MI CO-CO CO-MI CO-BA CO-BE BA-BA Figure 6-21. Biomass accumulation vs. crop sequences relative to absolute averages, when all other factors are averaged. Note: abbreviations are two first letters of rotations in Table 6-7. During May June and July corn performed much better than the others; during August, September, and October millet and sorghum accumulated better than the others, together with corn. 6.3.6 Variability by Soil Types In Figure 6-22 it is possible to explore the variability among soil types regarding N leaching and biomass accumulation under different ENSO phases, two manure N

PAGE 196

174 application ratios and with a crop rotation consisting of corn-sorghum. Evidently, the crops in soil type 2 (Arredondo-Jonesville-Lake) leach the least N with a medium biomass accumulation rate and crops in soil type 6 (Millhopper-Bonneau) leach the most N and at the same time accumulate the most biomass. Also, consistently, crops in soil type 4 (Penney-Otela) and soils 10 (Blanton-Ortega-Penny) leach and accumulate medium levels. Crops in soil 3 (Bonneau-Blanton-Eunola), followed by soil type 9 (Blanton (low)) had the lowest biomass accumulation rates with a medium-high level of N leaching. As in previous cases, consistently less N leaching and more biomass accumulation were found in El Nio ENSO phases and the opposite in La Nia years. It is important to notice that the N leaching in soil type 6 is substantially higher than the others together with higher rates of biomass accumulation. These soils have the highest drainage rate and it may be the main reason for these differences, as was previously discussed. Comparing the two graphs A) and B) in Figure 6-22, it can be observed that not much difference is found in N leaching between the two different rates of manure N applied (20 and 40 kg ha -1 mo -1 ), but the difference is marked in the biomass accumulation, indicating that the crops are still up-taking extra amounts of manure N applied. Additionally, the higher amounts of organic matter because of higher manure N applications can help prevent N leaching; this potential explanation would not hold, however, for the case of even higher manure N applications, where much higher N leaching was noticed (Figure 6-22). It can be also observed, by comparing the graphs, that the relative spatial positions of the measurements changes because the interactions of soils, ENSO phases, and manure N

PAGE 197

175 application as can be seen, for example, by the biomass accumulation for soils 6, 4, 10, 9, 8, 5, 4 and 3 in Neutral years, which is lower than the El Nio years for manure N application of 20 kg ha -1 mo -1 compared to the higher application of 40 kg ha -1 mo -1 55276391810456937411082210814936724029034039044049054020,00022,00024,00026,00028,00030,00032,00034,00036,000 LA NIA NEUTRAL EL NIO Soil Series 1 Arredondo-Gainesville-Millhopper2 Arredondo-Jonesville-Lake3 Bonneau-Blanton-Eunola4 Penney-Otela5 Penney-Kershaw6 Millhopper-Bonneau7 Otela-Jonesville-Seaboard8 Blanton(high)-Lakeland9 Blanton(low)10-Blanton-Ortega-Penny Biomasskg/ha-yr N leachingkg/ha-yr ROTATION: corn-sorghumMANURE N: 40 kg/ha-mo SEQUENCE: corn-sorghum MANURE N: 40 kg ha-1 mo-1 AB 55541081936722810147396763941810224029034039044049054020,00022,00024,00026,00028,00030,00032,00034,00036,000 Biomasskg/ha-yr N leachingkg/ha-yr ROTATION corn-sorghumMANURE N: 20 kg/ha-mo SEQUENCE: corn-sorghum MANURE N: 20 kg ha-1 mo-1 Figure 6-22. Nitrogen leaching vs. biomass accumulation for different soil types under different ENSO phases (La Nia, Neutral, and El Nio), two manure N applications (40 and 20 kg ha -1 mo -1 ), and forage system consisting of corn-sorghum-winter forage

PAGE 198

176 6.3.6.1 Nitrogen leaching and soil types N leaching, besides the amount of manure N deposited, depends on the permeability, the water holding capacity, the exchange capacity, the temperature, and the pH of the soil. N leaching is favored when soil conditions present nitrates (), the N compound more readily leached, and low water holding capacity. Nitrates () are more likely to occur in soils with low exchange capacity, moderate or high temperature, and higher pH. Overall, north-Florida soils are highly vulnerable with respect to N leaching because they are very sandy and present low exchange capacity. The pH is an important factor. The pH of these soils varies from slightly acid to very strongly acid. The slightly acid soils are more prone to N leach than the very strongly acids (Table 6-1). Strongly acid or very strongly acid soils (pH 5.5) tend to lose less N because much of the N is converted to ammonium,, which can be adsorbed and stabilized by the exchange complex of the soil. However, soils with higher pH (moderately or slightly acid, pH > 5.5) tend to lose more N because the N can be converted to either ammonia () or to nitrate (). Whilevolatilizes to the air and is not a direct threat to water, is easily leached when soils have low water holding capacity. 3NO3NO3 4NH 3NHNO NO 3NH 3 Additionally, soil temperatures in north-Florida are warm enough to not stop or decrease considerably the processes of mineralization or nitrification of the waste N sources that conduce N leaching. Overall, yearly N leaching (kg ha -1 year -1 ) presented substantial variation for the different soil types found in North-Florida; it ranged from 272 kg ha -1 for soils type 2 to 480 kg ha -1 for soils type 6, Table 6-8. This fact can be mostly attributed to the drainage that is the lowest for soil 2 and the highest for soil 6 (Table 6-1). Exploring more in depth

PAGE 199

177 the N leaching and soils, it can be seen that, comparing soils with the same drainage rate (soils 1, 5, 7 and 10 with 0.75 drainage and soils 3, 4, and 9 with drainage 0.80, Table 6-1), more N leaching occurred on those soils with lower cation exchange capacity and/or in those with higher pH, as would be expected. The months with the lowest rates of N leaching, December and April, are those with the maximum variability. As in previous discussions, January and February presented the highest rates of N leaching independently of the soil type. 0.000.501.001.502.002.50DECJANFEBMARAPRMAYJUNJULAUGSEPOCTNOVRELATIVE N LEAC H SOIL1 SOIL2 SOIL3 SOIL4 SOIL5 SOIL6 SOIL7 SOIL8 SOIL9 SOIL10 6.04 Figure 6-23. Nitrogen leaching vs. soil types relative to absolute averages, when all other factors are averaged. Note: Soil numbers refer to soil types defined in Table 6-1. When crops are initiating their cycle they heavily use nutrients and therefore, there is less available N for leaching. Soil type 3 is an exception that showed a high N leaching rate even during the starting months, probably because of a very poor retentive soil system along with a low depth profile. There was quite a variation among soils in the other

PAGE 200

178 months, but N leaching was the highest during winter in February and March, followed by September and October and next for June and July. Table 6-8. Nitrogen leaching (kg ha -1 ) and soil types, when all other factors are averaged Soil1 Soil2 Soil 3 Soil 4 Soil5 N Leach CV N Leach CV N Leach CV N Leach CV N Leach CV Dec 1.46 489% 1.82 460% 63.98 43% 3.33 370% 14.57 171% Jan 89.04 43% 90.05 37% 128.08 18% 113.35 34% 127.26 20% Feb 84.33 25% 70.94 26% 47.14 28% 84.46 25% 49.57 33% Mar 21.82 54% 16.70 56% 11.57 44% 21.20 52% 11.80 49% Apr 3.53 121% 3.68 118% 13.93 45% 4.96 110% 7.60 78% May 8.73 80% 9.04 79% 16.99 43% 10.87 76% 12.68 49% Jun 18.37 43% 17.42 44% 22.42 42% 20.85 44% 18.51 39% Jul 13.26 45% 11.82 47% 15.85 45% 14.63 47% 12.70 49% Aug 12.82 45% 12.51 45% 24.52 29% 16.36 42% 18.42 30% Sep 17.45 40% 17.08 40% 20.91 38% 20.00 42% 17.27 35% Oct 16.84 33% 15.61 34% 16.83 38% 18.24 36% 14.94 35% Nov 5.59 61% 5.05 62% 6.09 65% 6.02 65% 5.46 59% Year 293 272 388 334 311 Table 6-8. Continued Soil6 Soil7 Soil8 Soil9 Soil10 N Leach CV N Leach CV N Leach CV N Leach CV N Leach CV Dec 3.37 403% 1.43 494% 1.90 436% 6.45 280% 2.00 380% Jan 133.46 33% 86.08 47% 88.15 42% 135.04 26% 97.85 37% Feb 105.05 23% 97.17 21% 83.02 23% 77.71 29% 86.32 22% Mar 27.56 48% 29.83 46% 23.04 50% 19.05 52% 25.55 50% Apr 6.63 121% 3.76 171% 3.56 139% 6.58 97% 4.29 130% May 19.33 77% 10.31 91% 8.02 88% 14.86 60% 11.43 77% Jun 42.04 43% 25.46 44% 16.74 48% 26.86 39% 24.78 42% Jul 31.05 45% 19.07 43% 12.11 48% 17.41 43% 18.71 41% Aug 23.62 43% 14.02 50% 12.84 47% 19.45 36% 15.12 43% Sep 36.03 41% 22.39 43% 16.88 44% 22.64 37% 21.25 38% Oct 38.40 33% 23.88 33% 15.62 36% 20.86 32% 21.89 30% Nov 13.90 60% 8.16 58% 5.11 65% 6.98 58% 8.11 58% Year 480 342 287 374 337 Note: Soil numbers refer to soil types defined in Table 6-1. Monthly variation of N leaching in different soil types (Figure 6-23) showed that with exception of December, April, and August, soil 6 would leach the highest. It is interesting to notice that there was a temporal variation for different soil types and N

PAGE 201

179 leaching as is the case of soil 3 that presented the third highest N leaching rate in January (128.08 kg ha -1 ), but it also presented the lowest N leaching rate in February (47.14 kg ha -1 ). The possible explanation lies in the dynamic of the soils meaning that a high amount of leaching in January would leave less N available for leach in February or vice versa. With the exception of soil 3 that evidently is an outlier, and soil 6 in November, the variation of the mean regarding leaching went from 90% above and below depending of the crops and the season of the year. For the months with more than 10 kg ha -1 mo -1 leached (Table 6-8) there was quite a variation for different soils. Soil 1 was consistently below the average of leaching for all the months except February and March, soil 2 was always 10 to 80 % below the average, and soil 4 was most of the time between 0 and 10 % below the average. 6.3.6.2 Biomass and soil types Overall biomass accumulation (kg ha -1 year -1 ) varied from 24,441 for soils type 3 to 31,297 for soils type 6. Greater biomass accumulation was predicted with soils of type 6 and 7 and less biomass accumulation was observed with soils of type 3 and 9 (Table 6-9). Apparently, the high permeability and low water holding capacity besides promoting higher N leaching also promotes better plant growth and development under the conditions of spray fields in north-Florida dairy farm systems with these high nutrient applications. For example, for soils of type 6 that leached the most, the biomass accumulation was also the highest: for soils of type 2 that had the lowest N leaching, the biomass accumulation was the third lowest (Figure 6-24).

PAGE 202

180 In winter, during February and March soil 6 showed the highest accumulation and soil 3 the lowest. Also during spring-summer soil 6 yielded the most and soil 3 the least. During summer-fall there was less variation among soils. Table 6-9. Biomass (kg ha -1 ) and soil types, when all other factors are averaged Soil1 Soil2 Soil3 Soil4 Soil5 N Leach CV N Leach CV N Leach CV N Leach CV N Leach CV Dec 86 31% 86 31% 86 30% 86 31% 86 31% Jan 1,369 17% 1,313 17% 1,242 15% 1,381 17% 1,441 17% Feb 2,227 10% 2,138 10% 1,987 10% 2,308 10% 2,389 10% Mar 1,401 30% 1,372 29% 1,313 30% 1,449 29% 1,452 31% Apr 977 39% 960 38% 868 34% 973 39% 934 36% May 4,575 11% 4,459 12% 3,875 10% 4,667 11% 4,353 11% Jun 4,842 12% 4,801 13% 4,358 12% 5,017 12% 4,711 12% Jul 3,202 12% 3,170 12% 2,946 12% 3,306 12% 3,111 12% Aug 1,657 31% 1,623 31% 1,532 29% 1,684 31% 1,599 29% Sep 2,778 11% 2,726 11% 2,560 10% 2,855 11% 2,706 10% Oct 3,083 12% 3,043 13% 2,901 12% 3,194 13% 3,020 12% Nov 796 46% 791 46% 773 46% 821 46% 787 46% Year 26,992 26,481 24,441 27,741 26,589 Table 6-9. Continued Soil6 Soil7 Soil8 Soil9 Soil10 N Leach CV N Leach CV N Leach CV N Leach CV N Leach CV Dec 85 31% 86 31% 86 31% 86 31% 86 31% Jan 1,505 20% 1,410 18% 1,336 17% 1,282 16% 1,327 17% Feb 3,007 9% 2,464 10% 2,202 10% 2,034 10% 2,188 10% Mar 1,799 24% 1,545 27% 1,406 29% 1,346 29% 1,396 29% Apr 1,005 42% 988 40% 954 38% 942 37% 961 39% May 5,308 12% 4,857 12% 4,494 11% 4,282 11% 4,544 12% Jun 5,649 12% 5,243 12% 4,865 12% 4,655 13% 4,873 13% Jul 3,602 12% 3,424 12% 3,213 12% 3,092 12% 3,217 12% Aug 1,810 35% 1,729 32% 1,631 30% 1,585 30% 1,645 32% Sep 3,115 12% 2,941 11% 2,759 11% 2,650 11% 2,768 11% Oct 3,534 13% 3,318 13% 3,088 12% 2,955 13% 3,089 13% Nov 879 48% 846 47% 802 46% 782 46% 801 46% Year 31,297 28,850 26,835 25,689 26,895 Note: Soil numbers refer to soil types defined in Table 6-1.

PAGE 203

181 0.850.951.051.151.251.35DECJANFEBMARAPRMAYJUNJULAUGSEPOCTNOVRELATIVE BIOMASS SOIL1 SOIL2 SOIL3 SOIL4 SOIL5 SOIL6 SOIL7 SOIL8 SOIL9 SOIL10 Figure 6-24. Nitrogen leaching vs. soil types relative to absolute averages, when all other factors are averaged. Note: Soil numbers refer to soil types defined in Table 6-1. 6.4 Conclusions North-Florida dairy farm systems use 11 possible crop/pasture sequences for a year in three main crop seasons (spring-summer, summer-fall, and winter). The most common winter forages (WF) are the small grains such as rye, oats, wheat, or ryegrass. The sequences are bermudagrass-bermudagrass-WF; corn-sorghum-WF; millet-corn-WF; sorghum-corn-WF; millet-sorghum-WF; sorghum-millet-WF; corn-corn-WF; corn-millet-WF; corn-bahiagrass-WF; corn-bermudagrass-WF; and bahiagrass-bahiagrass-WF. Although only bermudagrass and bahiagrass have currently available models, all of the above crop/pasture sequences are possible to simulate using the DSSAT v4.0 by adapting crop coefficients from closely related crops: silage corn from maize, forage sorghum from grain sorghum, pearl millet from millet, and winter forages from wheat and calibrating and

PAGE 204

182 validating them against experimental data. This study found crop coefficients to represent north-Florida forage crops. A sensitivity analysis of N leaching and biomass accumulation of the north-Florida forage systems to extreme effluent manure N application (0 to 200 kg N month -1 ) indicated that the rate of N leaching will increase while the rate of biomass accumulation will decrease to zero when the rates of N applied with the manure increase. When the application is approximately 120 kg N month -1 the biomass accumulation reaches a maximum and the leaching of N increases at a higher rate. Higher (lower) N leaching (biomass accumulation) is predicted for El Nio years than Neutral years or La Nia years, attributed mostly to the number and intensity of rainfall events. An attempt to relate the leaching with the Japan Meteorological Index, used to predict the El Nio Southern Oscillation phases, showed little correlation. Winter in general, and specifically January and February are the critical months for N leaching when more than 50% of annual N leaching is predicted. This is the same time that low biomass accumulation is noticed. Biomass accumulation is the highest for corn, millet, and sorghum, followed by bermudagrass and bahiagrass. However, higher stability and consequently less variability is noticed in the systems when they include either of the perennials (bermudagrass or bahiagrass). The best forage systems to prevent N leaching are those that start in spring-summer with bermudagrass or corn; have bermudagrass, bahiagrass or corn in summer; and finish with winter forages. The systems that leach the most are those that include millet and/or sorghum. Bahiagrass could also leach high N in conditions of high manure N application.

PAGE 205

183 There is great variability in the interaction of soil types and months of the year regarding N leaching and biomass accumulation. Water holding capacity, pH and permeability of soils are believed to hold most of the causes for these differences. Soils with very low permeability and higher pH will facilitate the leaching of N, which also will vary throughout the year, and at the same it will determine higher biomass accumulation. There is more biomass accumulation for soils with higher N leaching rates because in these intense systems there will be little stress because lack of N, other nutrients, and/or water because these fields have irrigation systems that deliver the manure effluent and artificial irrigation as needed. The use of crop simulations were critical in recognizing trends, interactions, and identifying absolute values of N leaching and biomass accumulation under different conditions of climate, manure N applications, soil types, and forage systems. It would be impossible to design and conduct a real experiment of these magnitudes.

PAGE 206

CHAPTER 7 AN INTEGRATED SIMULATION MODEL TO ASSESS ECONOMIC AND ECOLOGIC IMPACTS OF NORTH FLORIDA DAIRY FARM SYSTEMS 7.1 Introduction Environmental concerns for dairy systems, N leaching specifically, for north-Florida are recognized as the most important factors for successful business in the future (Staples et al., 1997; Fraisse et al., 1996; Van Horn et al., 1998). Because the increasing concern regarding N levels in the Suwannee River Basin, north-Florida dairy farms are being more closely scrutinized by regulatory agencies (Twatchtmann, 1990), and effective tools to estimate point-source pollution of specific operations as well as potential ways to mitigate the problem inside the farm gate is an imperative need. The complexity of dairy farms in north-Florida justifies the creation of a whole-farm model, integrating several modeling approaches, in order to better analyze these systems (Herrero, 2000). Participatory and interdisciplinary methodologies can be pursued along with model creation, calibration, and validation in order to have better acceptance by final users (Geurtz and Joldersma, 2001). The objective of this paper is to describe the EXCEL-based Dynamic North-Florida Dairy Farm Model (DNFDFM). The DNFDFM is a simple event-controlled, dynamic model coupled to crop and optimization models created to assess N leaching from north-Florida dairy farm systems and the economic impacts of reducing it. The model responds to specific environmental (climate and soils) and managerial characteristics of dairies (total number of cows, milk production, crop rotations, waste system, etc.) and 184

PAGE 207

185 can be used to study the economic and ecologic sustainability of these systems. The driver of the DNFDFM is a dynamic adaptation of the framework balance and/or budgeting of nutrients in dairy farms, commonly used in Florida (Van Horn et al. (1991, 1994, 1998, 2001) and NRCS (2001) (Chapters 1 and 5). The DNFDFM incorporates Markov-chain probabilistic simulations of cow-flows (Chapter 5), crop simulation models (Jones et al., 2003) (Chapter 6) for historical climatic years for El Nio Southern Oscillation (OBrien, 1999) (Chapter 6), automated linear programming optimization of managerial options (Hazell and Norton, 1986), participatory modeling (Geurtz and Joldersma, 2001) (Chapter 4), and user-friendliness. 7.2 Materials and Methods 7.2.1 Conceptual Model Most dairies in Florida manage animals under semi-intensive or intensive systems and are essentially businesses producing milk: decisions are based on profit maximization and operation regulations (Adams, 1998). North-Florida dairies are business enterprises and have full access to credit opportunities, information, and new technologies (Adams, 1998). Additionally, dairy farms, which produce both livestock and crops, are composed of interconnected units; any part of which is affected by and affects the behavior of other parts, including human management decisions and strategies (Adams, 1998). Since a whole-dairy-farm encompasses environmental, economic, and bio-physical components, its analysis is best served by a systems approach that accounts for the interactions of these components, and that can trace the consequences of an intervention through the entire system (Kelly, 1995). While systems can be conceptualized at any of many scale levels, a systems approach defines a specific scale of interest and an appropriate boundary of analysis; analysis in the systems approach is marked by

PAGE 208

186 recognition of the whole system and the interactions within that system rather than looking only at a system component (Kelly, 1995). A systems approach employs specific techniques and tools, such as rapid appraisal, pattern analysis, diagrams, and modeling, often in a multidisciplinary fashion, to identify system boundaries and recognize component interactions (Kelly, 1995). In order to assess what would happen with the whole dairy operation and what would happen specifically with nitrate flow under different climate-management scenarios, models of these realities were set up (Thornley and Johnson, 1990). Models allow foreseeing results when temporal or material conditions make it impractical to develop field experiments. The successful planning of an animal waste management system requires the ability to simulate the impact of waste production, storage, treatment, and utilization on the water resources; it must address overall nutrient management for the operation (Fraisse et al., 1996). The systems approach used is Farming Systems Research and Extension (FSRE) as described by Hildebrand (1990). The farm household in livelihood systems analyses is replaced by farm management (Hazell and Norton, 1986) for commercial or industrial farm dimensions such as north-Florida dairy farms. The farming system is composed of an arrangement of subsystem components that function as units. These subsystems include biophysical, environmental, and economic components as well as relationships of the dairy operation to climate variability. These agroecosystems, as described by Powers and McSorley (2000), include the populations and communities in relationship with the abiotic factors organized in ecosystems.

PAGE 209

187 The Van Horn et al. (1991, 1994, 1998, 2001), NRCS (2001), and Lanyon (1994) approaches were adjusted, modified, and used for keeping track of dynamic N movement in all system components. Van Horn et al. (1998) mentioned the quantitative information needed on nutrient flow through all segments of the system. They list the following critical research questions: How much of individual nutrients are excreted by dairy cows? How does the manure management system affect nutrient flow and nutrient recoveries for fertilizer use? What is the potential nutrient removal by plants? How can a manure nutrient budget be developed? What steps help to document environmental accountability? Agronomic measures of nutrient balance and tracking of inputs and outputs for various farm management units can provide the quantitative basis for management to better allocate manure to fields, to modify dairy rations, or to develop alternatives to on-farm manure application (Lanyon, 1994). For instance, identifying where pollutants are being directly discharged and taking measures to reduce these discharges are more easily accomplished than controlling the indirect sources or non-point source pollutants (Environmental Literacy Council, 2002). Indeed, historically, scientists have been able to track single, or "point" sources of pollution, but tracking multiple, or non-point sources has been difficult (Ondersteijn, 2002). Crop models (Jones et al., 2003, Jones et al. 1998) are process-based dynamic simulation models that allow translating biophysical and environmental conditions into agricultural outcomes (Philips et al., 1998). Crop models can be used to estimate monthly biomass accumulation and N leaching by a number of crop and/or pasture sequences

PAGE 210

188 under different climatic conditions, depending on soil conditions, manure applications, and other managerial choices. Linear programming models (LP) are optimization models that can devise alternative management choices that maximize (minimize) an objective function according to a set of restrictions (Hazell and Norton, 1986). LP models have been widely used in analysis of farming systems, beginning about the time of the Heady and Candler book in 1958. Different simulation methods were needed to represent accurately the north-Florida dairy crop and livestock systems, and test management strategies that yield the most desired outcomes. Complex systems demand the study of interactions among processes acting on multiple scales, and consequently new and innovative approaches are needed to deal with these problems, coupling advances in dynamical theory, stochastic processes, and statistics (Levin, 1991). 7.2.2 Computer Implementation The model, Dynamic North-Florida Dairy Farm Model (DNFDFM) was developed in Excel using embedded Visual Basic in order to produce a user-friendly final product for farmers, Extension services, and regulatory agencies. The crop models were previously run by using the Decision Support System for Agricultural Transfer, DSSAT v4.0 (Jones et al., 2003) and the results were implemented to be reproduced in the DNFDFM in Excel. The optimization component is automatically set up and solved taking advantage of the solver tool that ships with Excel. 7.2.3 Analysis of Synthesized Dairies in North-Florida The model was run for 3 synthesized farms; synthesized farms were created using real data. Because of the sensitivity of the topic, synthesized farms have been created

PAGE 211

189 as examples for small, medium, and large operations. These synthesized farms combine many real characteristics of dairies in the study area without pointing out any specific farm information for which the DNFDFM can be run to perform relevant analyses. Outputs of simulating these farms are presented in the results sections. Statistical, optimizations, and other analyses of results are presented in the next chapter (Chapter 8). 7.3 Model Description 7.3.1 Overall Characteristics The Dynamic North-Florida Dairy Farm Model (DNFDFM) is an integrated, whole-dairy-farm simulation that links decision making with biophysical, environmental, and economic processes. The main objective of the DNFDFM is to predict monthly N leaching and profit in response to environmental (climate and soils) and managerial characteristics of farms (livestock management, waste management, and crops management). The DNFDFM can be used to study the economic and ecologic sustainability of these systems, as well as to diagnose sources of problems and search for feasible alternative management practices. The simulation approach includes a dynamic event-controlled simulation model connected to crop models and to optimization models. The driver, that has the livestock and waste modules, follows a dynamic adaptation of the budgeting framework of tracking the flow of N through the entire system along with production and additional economic variables, according to a pre-defined set of management and climatic conditions. Crop models assess the N recycled by plants and the amount of N leached according to management, genetic, soil, and climatic specific conditions. The optimization model consists of a linear programming model that either minimizes N leaching or maximizes profit for specific farm conditions. The DNFDFM incorporates Markov-chain

PAGE 212

190 probabilistic simulation of cow-flows and crop simulation for historical climatic years (El Nio Southern Oscillation), automated optimization of managerial options, participatory modeling, and user-friendliness. 7.3.2 Components The main components of the DNFDFM are displayed in Figure 7-1. The livestock model simulates the cow-flow and manure N excretion. The waste model receives inputs from the livestock model and simulates the manure N flow through the waste handling system. The crop models receive inputs from the waste model to simulate N leaching and biomass accumulation in the crop fields. The livestock, waste, and crop models run in parallel dynamically in monthly steps. Each one of them is a function of a set of management practices. Crop models are additionally sensitive to climate conditions and soil characteristics. The outputs of crop simulations are measured in monthly estimations of overall N leaching and biomass accumulation in a dairy farm. An economic component goes across the livestock, waste, and crop models to estimate monthly overall profit. The optimizer is a linear programming model that solves a matrix built with outputs of several simulations. The optimizer changes management scenarios and controls multiples simulation runs. The optimizer that is an automated tool that triggers the linear programming modeling, minimizes N leaching or maximizes profit under a set of restrictions. Suggested management strategies by the optimizer can be compared with feasible farm practices in consultation with the farm manager to find the feasible adjustments in

PAGE 213

191 each particular case in an iterative process of changing input managements and re-run the simulation. Figure 7-1. Representation of the dynamic north-Florida dairy farm model Following are the specifics of each one of the DNFDFM components regarding model creation and implementation. For more details other chapters can be also reviewed. Chapter 1 deals with overall characteristics of models, the north-Florida dairy farm characteristics, and the conceptual theory in the modeling. Chapter 5 describes in detail the parameterizations of the livestock and waste models. Chapter 6 is comprehensive in the simulation of forage systems using crop models. 7.3.3 The Livestock Model The livestock model simulates the cow-flow (C ijk ); milk production (Mk); water consumed (Wc); manure excreted in concentrated areas (Mc), in pasture (Mp), and in the

PAGE 214

192 waste system (Mw); and N excreted in concentrated areas (Nc), in pasture (Np), and in the waste system (Nw) (Figure 7-2). Additionally, the livestock model estimates the dry matter intake (DMI) and the N inputted in the feed (Nf). Required inputs as initial conditions are: total adult cows (TAC), total number of bulls (TNB), percent of heifers raised (PHR), rolling herd average (RHA), percent of seasonality (POS), amount of crude protein (ACP), percent of confined time (PCT), time spend in concentrated areas (TCA), maximum water usage (MWU), and the amount of water recycled (Rc). In Table 7-1, i is the months in milk after delivery (1-18) for adult cows or the months of age after birth (1-32) for heifers; j is the pregnancy months (0-9); k is the lactation cycle (0-9), and m is the month of the year. Figure 7-2. Representation of the livestock model

PAGE 215

193 imkX Y, and represent three dimensional matrices that include the parameterization of the culling rates, reproduction rates, and the milk production rates, respectively, according to north-Florida dairy farms conditions (De Vries, 2004). A fourth dimension common to these matrices is the seasonality level. This seasonality level goes between 0% (minimum variation of milk production throughout the year) and 100% (maximum variation of milk production in seasons of the year). By default the seasonality level is 100% because that is the way most north-Florida dairy farms operate. Milk production is additionally adjusted to match the inputted RHA, after the model starts to run. imk imkM imkM Equation [7-1] simulates the cows that do not become pregnant, while equation [7-2] represents cows that become pregnant in a determined month. Equation [7-3] updates cows that were already pregnant and Equation [7-4] deals with cows that are freshening (having calves) and entering a superior lactation cycle. Every month there are cows that are involuntary culled, but additionally, if cows are not pregnant after 12 months in milk or heifers are not pregnant when 24 months of age, they are voluntarily culled. Milking cows are adult cows that produce milk, including cows of first lactation or higher and between 0 and 7 months pregnant, Equation [7-5]; dry cows are adult cows of first lactation or higher, between 8 and 9 months pregnant, Equation [7-6]. Heifers are young livestock that include all animals in lactation 0 plus the female off-spring that are 50% of all born. The number of heifers is adjusted by the PHR, Equation [7-7]. imkX Milk production is estimated by the productivity rates of milking cow groups,, and the number of cows at each stage of productivity. Milk production is adjusted by the imkM

PAGE 216

194 Table 7-1. State variables and monthly update of the livestock model Name State Variable Units t+1 # Cow category ijkC Heads imkimkjkiYXC 1)1( imkimkkjiYXC)1)(1( 1)1)(1( jimkkjiXC 93211)1(9jimkikiXC [7-1] [7-2] [7-3] [7-4] Milking cows )91)(70(iC Heads 1817091iijkjkC [7-5] Dry cows )91)(97(iC Heads 1819891iijkjkC [7-6] Heifers )0(ijC Heads 2/93211)1(9321090jimkikiiijjXCPHRC [7-7] Milk Mk lbs month -1 mimkiijkjkWMC.1817091 [7-8] Water consumption Wc Gallons month -1 mmiijkjkWRcRWMWUC1817191 [7-9] Dry matter intake DMI lbs month -1 BHDMDMIDMIDMIDMI [7-10] N inputted in feed Nf lbs month -1 BHDMNfNfNfNf [7-11] Manure waste system Mw lbs month -1 ))((mBMPCTMM [7-12] Manure conc. areas Mc lbs month -1 ))(1)((TCAPCTMMmBM [7-13] Manure pasture Mp lbs month -1 DHmBMMMTCAPCTMM )1)(1)(( [7-14] Nitrogen waste system Nw lbs month -1 ))((mBMPCTNN [7-15] Nitrogen conc. areas Nc lbs month -1 ))(1)((TCAPCTNNmBM [7-16] Nitrogen pasture Np lbs month -1 DHmBMNNTCAPCTNN )1)(1)(( [7-17]

PAGE 217

195 number of days in each month of the yearW, including a differentiation between regular and leap years, Equation [7-8]. m The livestock model requires a time for self adjustment, in which it runs in the background to find a steady state of the cow-flow and the objective inputted RHA. Equation [7-9] estimates the monthly water usage based on the number of milking cows, the MWU, the ratio of monthly usage,, and the amount of water recycled, Rc. This is adjusted to the number of days in each month,W. mRW m The dry matter intake is the sum of the independent DMI of the milking cows ( M ), the dry cows ( D ), the heifers ( H ), and the bulls ( B ) as noted in Equation [7-10]. The specifics can be seen in equations [7-17] through [7-20] in Table 7-2. Functions in Table 7-2 use book values parameterized for north-Florida conditions. The partition of the manure in different parts of the farm is estimated using Equation [7-12] for manure that goes through the waste system, Equation [7-13] for manure deposited in concentrated areas, and Equation [7-14] for manure deposited in pastures. Table 7-2. Dry matter intake estimations Variable Units t+1 # MDMI DDMI HDMI BDMI lbs month -1 18135237091)(2)(3.3)(16.2.25imijkimkimkimkjkWCMMM 9198912.25kmijkjkWC miijjWPHRCi321090)70.0(7.2 ))()(2.25(mWB [7-17] [7-18] [7-19] [7-20]

PAGE 218

196 Notice that manure from milking cows and bulls ( M B ) can be deposited in any of these compartments, however manure from heifers and dry cows ( H D ) is only deposited in pastures. The PCT can be constant across months or can be inputted as a function of the month of the year (PCT m ). The functions of waste excretion by different group of cows can be observed in Equations [7-21] through [7-24] in Table 7-3, which are based on book values and milk production levels. Table 7-3. Manure excretion estimations Variable Units t+1 # MM DM HM BM lbs month -1 1813327091)(3)(05.)(37.175.98imijkimkimkimkjkWCMMM 181989180imijkjkWC miijjWPHRCi321090)4.3(5.76 ))()(80(mWB [7-21] [7-22] [7-23] [7-24] The excretion of N in different parts of the farm is estimated using Equations [7-15] through [7-17] in Table 7-1 and the specifics are detailed in Equations [7-25] through [7-28] in Table 7-4. The partition of N excreted that goes to the waste system is expressed in Equation [7-15], the N excreted that goes to the concentrated areas is shown in Equation [7-16], and the N excreted in pastures is accounted for by Equation [7-17]. Following the same pathway as the manure, N manure from milking cows and bulls ( M B ) can be deposited in any of these compartments, however N manure from heifers and dry cows ( H D ) is only deposited in pastures. Functions shown in Equations [7-25],

PAGE 219

197 [7-26], [7-27], and [7-28] are based on book values for Florida and parameterizations of N excreted based on milk production. The N inputted in the feed, Equation [7-11], is estimated similarly as the sum of specific functions of animal groups represented in equations [7-25] to [7-28] in Table 7-4, but in this case each one divided by the constant of undigested N in feed of 0.67 (Van Horn et al., 1998). Table 7-4. Nitrogen excretion estimations Variable Units t+1 # MN DN HN BN lbs month -1 181372537091)(3)(6)(4.236.imijkimkimkimkjkWCMMM 181989136.imijkjkWC miijjWPHRCi32102290)2.1(8.2 ))()(364(.mWB [7-25] [7-26] [7-27] [7-28] Additionally, the N inputted in feed (Nf), Equation [7-11] and the N excreted by the animals (Nw, Nc, Nw), Equations [7-15], [7-16], and [7-17] are corrected by a factor of the amount of crude protein in the diet low (13.9370%) or high (15.0342%) according to NRC standards by Equation [7-29] and Equation [7-30]. MN = + (ACP-13.9370)(.0956) MN MN DN =+( ACP-13.9370)(.0956) DN DN [7-29] [7-30] 7.3.4 The Waste Model The waste model follows the flow of the manure N in the concentrated areas, the pasture, and the waste system after its excretion and before its utilization. Essentially, the waste model balances the manure N that reaches the crop fields after loss in different

PAGE 220

198 parts of the system: volatilization before flushing (VBF), solids collected (SC), volatilization in lagoon (VL), volatilization in storage pond (VP), sedimentation in sludge (SS), volatilization during irrigation (VI), and volatilization in the soil after being applied (VS). These are percentage rates required as initial conditions to be inputted before running the model. Table 7-5. Nitrogen through waste system estimations Variable Units t+1 # fMw fNw fMc f Nc fMp fNp lbs month -1 SSSCMw VSVISSVPVLSCVBFMw Mc VSNc Mp VSNp [7-31] [7-32] [7-33] [7-34] [7-35] [7-36] In Table 7-5, Equations [7-31] through [7-36] estimate the manure and manure N at the final point of utilization (f) for the waste system, the concentrated area, and the pastures. This model estimates the manure and manure N movement in successive steps and results in the model can be explored at any of the points of this continuous flow. Notice that the amount of manure for concentrated areas and pasture remains the same as that estimated in the livestock model. 7.3.5 The Crop Models The crop models from the Decision Support Systems for Agricultural Transfer (DSSAT v4.0, Jones et al., 2003) use daily weather information of temperature, irradiation, and precipitation along with soil characteristics, manure effluent applied, and other management choices to estimate monthly biomass accumulation and N leaching for

PAGE 221

199 specific crop sequences that are common in north-Florida dairy farm systems. Dynamic, process-based crop simulation models quantify interactions between environmental physical conditions of soils and climate, genetic physiological conditions of crops, and management conditions (Boote et al., 1996) and they are being widely used to experiment with in situations when limitations of time, space, or finances do not allow a field experiment. Crop models simulate light interception, leaf-level photosynthesis, N uptake, soil water balance, evapotranspiration, respiration, leaf area growth, growth of component parts, root growth, senescence, N mobilization, carbohydrate dynamics, and crop development processes. For the soil component, the Century model (Parton et al., 1979) implemented in the DSSAT by Gijsman et al. (2002) was used, which estimates soil N balances that include soil and surface organic matter, inorganic N, additions and removals of N, and all the processes that include the N cycle in the soil such as decomposition, mineralization, N leaching, or denitrification. Specific data for each of the soil types were converted to the DSSAT v4.0 system using SBuild software (Uryasev, 2003), where the drained upper limit values were corrected using Saxton et al., (1986). Forage crops were adapted, calibrated, and validated for north-Florida dairy farm conditions and all potential forage combinations were run for all N effluent ranges, 10 types of soils found in the study area (Table 7-6), and for 43 years of daily weather data (1956-1998). Residual organic matter from one crop to the next was accounted for as well as other management choices that are common in these systems such as extra irrigation and harvests.

PAGE 222

200 Daily cumulative N leached (lbs acre -1 ) and biomass (lbs acre -1 ) outputs from the simulations were compiled in monthly rates for the span of the study period (1956-1998). All months were classified according to ENSO phases and the main factors incorporated in the simulations (Figure 7-3). Table 7-6. Codification of ENSO phases, soil types, and forage systems from crop models e, ENSO phase f, Forage Systems 1 La Nia 1 Bermudagrass Bermudagrass 2 Neutral 2 Corn Sorghum 3 El Nio 3 Millet Corn s, Soil Types 4 Sorghum Corn 1 Arredondo-Gainesville-Millhopper 5 Millet Sorghum 2 Arredondo-Jonesville-Lake 6 Sorghum Millet 3 Bonneau-Blanton-Eunola 7 Corn Corn 4 Penney-Otela 8 Corn Millet 5 Penney-Kershaw 9 Corn Bahiagrass 6 Millhopper-Bonneau 10 Corn Bermudagrass 7 Otela-Jonesville-Seaboard 11 Bahiagrass Bahiagrass 8 Blanton(high)-Lakeland 9 Blanton(low)10 Blanton-Ortega-Penny In order to couple the crop models with the livestock and waste models two matrices were created with outputs of the crop models, one for N leaching () and the other for biomass accumulation () (lbs acre mesfnNl mesfnBa -1 month -1 ). Each of these matrices has 5 dimensions: month of the year m, ENSO phase e, soil type s, forage system f, and manure N applied n. The values for e, s, and f are inputs (that the user enters according to his/her own farm conditions) while m and n are estimated (calculated by the by the simulation). Coefficients of the N leaching () and for biomass accumulation () (lbs acre mesfnNl mesfnBa -1 month -1 ) are coded according to Table 7-6 and Table 7-7.

PAGE 223

201 Figure 7-3. Data management of simulation of crop models Estimations of N leaching () and biomass accumulation () are summarized in Equations [7-37] and [7-38]. and at the farm level are the sum of or in the pasture fields t plus the or in the sprayfields q. To estimate these in each pasture or spray field with a determined forage system f, a coefficient from the matrix or is obtained according to m, e, s, f, and n and multiplied by the area of the field, , respectively. Nl Ba NlBa Ba Nl BamesfnNl Nl mesfnBatPFa qSFa

PAGE 224

202 N leaching Nl lbs month -1 farm-1 QqqmesfnFfTttmesfnFfSFaNlPFaNl1111 [7-37] Biomass accumulation Ba lbs month -1 farm -1 QqqmesfnFfTttmesfnFfSFaBaPFaBa1111 [7-38] For the pasture fields, a correction is performed based on the fact that these fields receive direct deposits of manure and they are not irrigated. By using local knowledge and farm registers in the area, it was estimated that the N leaching is reduced 25% and biomass accumulation 40% in pasture fields, compared to sprayfields. The parameter n that represents the manure N applied to the fields that is estimated by the waste model does not necessarily match the amounts used in the experiments that were run with the crop models contained in the matrices and They are interpolated between the points in which the crop models were run, which are coded as n in Table 7-7. mesfnNl mesfnBanNa Equation [7-39] represents a general interpolation to find the in sprayfields, which can be modified to perform the same estimations for pastures replacing with or for Ba replacing with Nl fNw fNp Nl Ba Table 7-7. Nitrogen applied estimated by the waste model and N applied in crop models fNw or fNp n nNa < 8.9218 1 0 8.9218-17.8436 2 8.9218 17.8436-35.6872 3 17.8436 35.6872-71.3743 4 35.6872 71.3743 5 71.3743 fnfnnnmesffnmesfnfmesfnNwNaNwNaNaNlNwNlNaNwNl11)1()1(/ [7-39]

PAGE 225

203 7.3.6 The Economic Module This module estimates the overall profit (US$ month -1 ) of the simulated farm, based on a profitability ratio of milk productivity. Table 7-8 defines the variables involved in the calculations, all of them are inputs. The profit is estimated as the difference between the revenues less the expenses, Equation [7-40]. Equations [7-41] and [7-42] that calculate the R and E include the value of the biomass (Ba) and N (Nl) as a revenue and an expense, respectively. Table 7-8. Definition of variables in economic module (US$ cwt -1 milk -1 ) Revenues R Cow Sales Rc Heifer Sales Rh Gain on purchased livestock Rg Crop sales Rf Other revenues Ro Expenses E Personnel Ep Feed purchased Es Milk marketing Em Crop expenses Ef Other expenses Eo Other Inputs Milk price Mp N value Nv (US$ cwt -1 ) Crop value Cv (US$ cwt -1 ) ER [7-40] BaCvRoRfRgRhRcMpMkR [7-41] NlNvEoEfEmEsEfEpE [7-42] The value of the milk sold as revenue is expressed as the multiplication of the milk produced by its market price. More than 90% of the revenue on north-Florida dairy farms comes from the milk sold; therefore the model is highly sensitive to this factor. The model can run with a fixed milk priceinputted or by using a stochastic milk price Mpthat is estimated by a stochastic milk price generator included with the Mkf Mp oMp

PAGE 226

204 economic module. The stochastic functionresembles the distribution found in monthly prices of milk in the last 10 years in north-Florida, which can be seen in Table 7-9. fMp rmMp The stochastic price function generates a random number r (0 < r < 1) which together with the month of the year is used to predict the stochastic price as shown in Equation [7-43]. fMp Table 7-9. Monthly milk prices distribution in north-Florida as cumulative probability (US$ cwt -1 milk 1 ) Cumulative Probability Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 0.10 14.80 14.40 14.40 14.10 14.50 15.10 14.80 14.70 14.70 14.40 14.80 13.50 0.20 14.90 14.80 14.80 14.30 15.10 15.20 14.90 15.10 15.10 14.90 15.70 14.70 0.30 15.20 14.90 14.80 15.00 15.30 15.40 15.00 15.40 15.20 15.20 16.00 16.00 0.40 16.30 15.60 15.70 15.00 15.30 15.70 15.70 15.60 15.30 15.50 16.00 16.20 0.50 16.40 15.90 15.80 15.60 15.40 15.80 15.70 15.60 16.10 16.10 16.20 16.20 0.60 16.50 16.20 16.30 15.60 15.50 15.80 15.90 16.20 16.70 17.30 18.00 16.40 0.70 16.70 16.30 16.50 16.20 16.20 15.80 16.10 16.30 17.90 19.30 18.70 16.60 0.80 17.00 16.90 16.60 16.30 16.50 16.80 16.20 18.30 19.40 19.80 20.00 17.90 0.90 17.40 17.70 17.30 17.00 16.70 17.40 18.30 18.90 19.80 20.20 20.10 18.90 1.00 20.30 20.30 19.30 17.00 17.70 18.40 18.60 19.00 20.00 20.30 20.70 21.30 Source: Prepared with data from Florida Agricultural Statistics Service, 2003. fMp [7-43] 7.3.7 The Optimization Module The optimization process is conceptualized in Figure 7-4 and starts by selecting management options that a dairy farm potentially could change. The next step defines the levels of each management practice to be tested. A further task combines all levels of any management practice with all levels of all management practices, forming the scenarios. The simulation module is run for each of these scenarios and monthly results of profit and N leaching are organized in a matrix and then optimized for different ENSO phases using a linear programming model.

PAGE 227

205 Figure 7-4. The process of optimization The optimization module is a mathematical linear programming model that maximizes profit () or minimizes N leaching () of multiple scenario simulation runs s under restrictions of at most average or at least average Nl Nl respectively, as seen in following Equations [7-44], [7-45] and [7-46]. SsmmsmsXCNl1121,minmax, [7-44] subject to __121NlAXmsmms __121msmmsAX [7-45] and 0msX [7-46] msCA is the profit or the N leaching per unit of activity in a month m and in a scenario s. are the technical coefficients obtained by the simulation of multiple scenarios. msX ms __Nl

PAGE 228

206 and are the average of all scenarios selected in the optimization process. Optimization is repeated independently for each ENSO phase: La Nia, Neutral, and El Nio years, so results and suggested practices are a function of predicted climate conditions. __ The final results of the optimization process are maximum levels of profit or minimum levels of N leaching at the farm level together with the combination of scenarios suggested as management practices that the farm could implement in order to decrease N leaching and/or increase profit, under different climate forecast conditions. 7.3.8 The Feasible Adjustments Component The feasible adjustments refer to the feasible practices a dairy farm could implement pursuing an objective (i.e., decrease N leaching). It is a user-friendly concept of taking advantage of the simulation and the optimization components of the DNFDFM. It is not automated and requires interaction with the farmer or farm manager. After the optimization module proposes a set of management practices to minimize N leaching with a defined level of profitability, the farmer can decide which practices and their levels it would be possible to implement. The DNFDFM simulation module is then run again, several times as needed, to find an environmental/profitable level for a farm that is superior to the current situation, by implementing only those practices that are feasible in the eyes of the farmer. 7.4 User Friendly Implementation 7.4.1 General Characteristics The model was developed in Excel using embedded Visual Basic in order to produce a user-friendly final product for farmers, extension services, and regulatory agencies. Users only need to open a file of Excel and the model is ready to work.

PAGE 229

207 Excel software is one of the most widely used computer programs. The DNFDFM takes advantage of this fact and delivers an application that will be easier to adopt for final users as an Excel file (xls extension). The DNFDFM does not change any of the normal capabilities of a normal Excel software; it only locks the codes and some cell contents in order to protect the application from unintended changes, but leaves intact all the other Excel capabilities, so the users can utilize the Excel as usual when they want to make spreadsheet calculations, printouts, or take any other action along with the model. 2 The main components of the DNFDFM simulation package are the livestock module, the waste management handling system, the location control, the crop controls, the climate controls, the economic components, and the optimization module. For user-friendliness all modules and their connections are graphically represented in the starting spreadsheet of Excel (Figure 7-5). A button inserted in this spreadsheet triggers a user-friendly menu that guides the user through the use of the application. This menu first appears showing a tab called start with a message welcoming the user to the system, briefly explaining the purpose of the application, and suggesting the help file be read before starting the runs. This welcoming tab also gives the contact information to obtain support and the date that will be used for the printouts. Both, the welcome page and the start spreadsheet have a button to trigger the help file. The user-friendly menu (Figure 7-6) has eight tabs (top) and a common section (bottom). Besides the welcoming start tab, there are tabs for: livestock, N, soil, crop, climate, economics, and optimize; in each of which the user can navigate for reviewing, 2 The model can be downloaded by clicking in the link: DNFDFM.xls (3.44 MB). It requires at least Windows XP and Office XP.

PAGE 230

208 inputting and/or selecting specific data. The common section consists of a drop-box menu, buttons of view selection, and a button called run. The drop-box menu allows the user quickly to customize small, medium, or large farms to represent a typical north-Florida farm of any of these sizes; and the view buttons allow the user select between seeing graphs or numbers in real time of simulation. These global options are useful for analysis purposes when several runs are intended as is the case of optimizations, but also are useful to use as templates to begin the specific customization of each farm. The run button triggers a simulation run of the DNFDFM. The livestock tab deals with the cow-flow and milk production variables; the N tab is about the waste management system and water usage; the soil tab deals with the farm Figure 7-5. Main screen of the computer implementation of The Dynamic North-Florida Dairy Farm Model (DNFDFM)

PAGE 231

209 location and the type of soils; the crop section defines all the farm fields, their size and their crop sequences. The climate module allow the user define the number of years of simulation together with the climatic years (El Nio, La Nia, or Neutral climatic years, according to ENSO); the economic section allows the user to personalize the revenues and expenses for specific farms; and the optimization tab can be used to define and trigger the optimization task. Figure 7-6. The start message menu control

PAGE 232

210 For all modules, there are default values to start with as a help to the user for introducing information. Additionally, there are other buttons, figures, check boxes and other objects to help the user fill in customized information. After the initial information is inputted, the user can hit the run button and the model will start simulating; a time lag will proceed to let the model self adjust and then the results will be seen in real time by numbers and/or graphs changing in the screen. After running, the user can review the results by moving a scrollbar changing the timeline of the run. Also the user can review the database generated from the results. 7.4.2 The Livestock Module The livestock model could represent the dynamics of the cows and the seasonality of milk production and manure N of a north-Florida dairy farm very well by only inputting the total number of adults cows and the rolling herd average; however for more customization and for other estimations the user can input other layers of information such as the total number of bulls, the percent of heifers raised on the farm, the percent of seasonality, the amount of crude protein in diet, the percent of confined time, and the percent of time milking cows spend in concentrated areas (Figure 7-7). For the confined time, there is an additional button that allows the user to enter monthly variations of confined time if desired. The livestock module communicates directly with the feed component to retrieve information about the dry matter intake and the protein amounts and moves its outputs to the modules that handle the waste in the concentrated areas, the pasture fields, and the waste management system.

PAGE 233

211 Figure 7-7. The livestock module control 7.4.3 The Waste System Module This module allows the user to customize and define specific farm parameters about the N lost in the waste system and each one of its components and the water utilization by the farm (Figure 7-8). Numbers vary greatly in different management systems. For example Van Horn et al. (1998) presents an average system for Florida in which there is an estimated total loss (total N lost, TNL) of 28.81% of N through the waste system with specific characteristic in all components. These values are used as default, but ultimately the users must change them to better mimic their known specific conditions.

PAGE 234

212 Figure 7-8. The waste management control 7.4.4 The Soil Module In order to run the crop systems, the user first selects the location of the farm either by clicking in a map or by selecting from a menu (Figure 7-9). Through this soil control, the user locates the farm, which in turn determines the type of soils in which the forage crops will be grown. The all soils option means the average results across all ten different soil types. 7.4.5 The Forage Systems Module The crop systems are defined in a user-friendly menu located in the crop tab (Figure 7-10). The forage system control allows the user to input data in up to 10 fields,

PAGE 235

213 which detail field size, the type of field (sprayfield or pasture), and the sequence of forage crops. Figure 7-9. Dairy farm location and soil type by farm component control 7.4.6 The Climatic Module The climatic module defines the forecasted climatic conditions under which the forage systems will grow. It is a user choice based on a classification of climatic years based on El Nio Southern Oscillation (ENSO) in El Nio, La Nia, and Neutral years. The climatic component allows the user to select the starting year of the simulation and run for a selected number of years, Figure 7-11. -yr Average option means the average climatic conditions for 43 years (1956-1998) of crop model simulations. Graphs that

PAGE 236

214 describe the climatic conditions under different ENSO phases are available for the user to review after clicking a button located in this tab. After selecting the run timeline, the user can select the forecasted climatic years. The concept is to devise management options that would yield more desirable outputs according to ENSO phases. Figure 7-10. Forage systems component control 7.4.7 The Economic Module The economic module requires current farm economic data inputted at the beginning (Figure 7-12). Profitability is estimated every month and is sensitive to the manure N recycled on-farm and the amount of biomass produced as feed inside the farm gate. This module estimates a monthly balance of revenues less expenses based on

PAGE 237

215 detailed information inputted by the user. As a guide for the user, it shows at the beginning an average for Florida farms published by the DBAP (De Vries et al., 2000); the user must personalize these values. Figure 7-11. The climatic component control The user has the option to fix or use stochastic milk prices. The stochastic milk price is a monthly random function based on 10 years of information (1993-2002); a graph that describes these monthly distributions is available for the user to review by hitting a button located together with the stochastic choice, and is shown in Figure 7-13. Additionally, the economic menu has the option of inputting the estimated value of N as fertilizer and the value of the dry matter produced on the farm.

PAGE 238

216 Figure 7-12. The economic component control Figure 7-13. Distribution of milk prices in north-Florida

PAGE 239

217 7.4.8 The Optimization Module The optimization function is triggered by a button called optimization located in the tab called optimize (Figure 7-14). Figure 7-14. The optimization control The optimization module builds and solves a linear programming matrix with the monthly results of N leaching along with the profitability of several runs. When finished, it shows N leaching minimized for average levels of profit. Results are presented comparing N leaching and profitability by current practices and by optimum practices and the suggested changes in order to reach these levels.

PAGE 240

218 7.5 Results for Synthesized Farms 7.5.1 Synthesized Small Dairy Operation This farm has 400 adult cows, 16 bulls, and raises 100% of its heifers. The rolling herd average is 17,000 lbs yr -1 the milk production is 100% seasonal, and the amount of crude protein in the diet is high. The milking cows spend 80% of their time confined. The total N lost through the waste management system is 28.81% and it is estimated that there is an extra 30% N volatilized from the soil when applied; the volatilization by direct deposition in soils is estimated at 40%. The small dairy farm has 70 acres of sprayfields and 80 acres of pasture fields with diverse crops as indicated in Table 7-10. Table 7-10. Crops in different fields Field Area (Ac) Type Spring Summer Winter 1 10 SPRAYFIELD CORN SORGHUM RYE 2 20 SPRAYFIELD CORN MILLET RYEGRASS 3 10 SPRAYFIELD SORGHUM MILLET WHEAT 4 20 SPRAYFIELD BAHIAGRASS BAHIAGRASS OATS 5 10 SPRAYFIELD MILLET SORGHUM RYE 6 20 PASTURELAND BAHIAGRASS BAHIAGRASS RYE 7 10 PASTURELAND BERMUDAGRASS BERMUDAGRASS RYEGRASS 8 20 PASTURELAND BAHIAGRASS BAHIAGRASS WHEAT 9 10 PASTURELAND BERMUDAGRASS BERMUDAGRASS OATS 10 20 PASTURELAND BAHIAGRASS BAHIAGRASS RYE The DNFDFM simulation was run for the above conditions, using the average soil type and the overall net income at 0.89 US$ cwt -1 milk produced -1 The simulation started in October 2004 and followed three consecutive years with 2004/2005 being a La Nia year, 2005/2006 a Neutral year, and 2006/2007 an El Nio year. Results contained in Figure 7-15 indicate that the overall N leaching in a year would be the lowest for La Nia years (13,528 lbs or 90.2 lbs per acre), followed by Neutral years (7% higher) and by El Nio years (13% higher).

PAGE 241

219 On a monthly basis, the N leaching varies from 150 lbs in April for La Nia years to 6,033 lbs in January for El Nio years with January and February the months with the highest leaching rates (Figure 7-15A). Accumulated biomass increases, as expected, towards the summer months (Figure 7-15B) when higher temperatures and rainfall determine greater plant growth, and profitability is highest between the months of April to July (maximum in May for La Nia years) because the higher N recycled and/or higher biomass accumulation. Figure 7-16 shows the N leaching and profit variability on a yearly basis for different ENSO phases. Always the overall N leaching is greater in El Nio years than Neutral years and Neutral years will present more leaching than La Nia years. The profitability is inversely related to the N leaching, La Nia years have the highest profitability and El Nio years present the lowest (Fig. 7-16). 7.5.2 Synthesized Medium Dairy Operation This farm has 800 adult cows, 30 bulls, and raises 100% of its heifers. The rolling herd average is 19,000 lbs yr -1 the milk production is 100% seasonal, the amount of crude protein in the diet is high and the milking cows spend 80% of their time confined. The total N lost through the waste management system is 35.31% and it is estimated that there is an extra 30% N volatilized from the soil when applied; the volatilization by direct deposition on soils is estimated in 40%. The medium dairy farm has 163 acres of sprayfields and 150 acres of pasture with the same crop combinations and the same proportions as indicated in Table 7-10. This arrangement was run under the same soils and profitability conditions of the previous Small farm.

PAGE 242

220 01,0002,0003,0004,0005,0006,0007,000OCTNOVDECJANFEBMARAPRMAYJUNJULAUGSEPLBS N LEACHED La Nia Neutral El Nio 050,000100,000150,000200,000250,000300,000350,000400,000OCTNOVDECJANFEBMARAPRMAYJUNJULAUGSEPLBS BIOMASS ACCUMULATED La Nia Neutral El Nio 3,0003,5004,0004,5005,0005,5006,0006,5007,000OCTNOVDECJANFEBMARAPRMAYJUNJULAUGSEPUS$ PROFIT La Nia Neutral El Nio A B C Figure 7-15. Monthly estimations for Synthesized Small north-Florida farm. A) N leaching. B) Biomass accumulation. C) profit.

PAGE 243

221 12,50013,00013,50014,00014,50015,00015,500La NiaNeutralEl NioLBS N LEACHED65,40065,60065,80066,00066,20066,40066,600US$ PROFIT PROFIT Figure 7-16. Small north-Florida farm, yearly estimated N leaching and profit Results for the different ENSO phases on a yearly basis are presented in Figure 7-17. Overall N leaching varied from 25,096 to 28,389 lbs yr -1 for La Nia and El Nio years, respectively. 7.5.3 Synthesized Large Dairy Operation This farm has 3,000 adult cows, 120 bulls, and raises 100% of its heifers. The rolling herd average is 20,000 lbs yr -1 the milk production is 100% seasonal, the amount of crude protein in the diet is high (NRC standard), and the milking cows spend 80% of their time confined. The total N lost through the waste management system is 40% and it is estimated that there is an extra 30% N volatilized from soil when applied; the volatilization by direct deposition in soils is estimated at 40%.

PAGE 244

222 The large dairy farm has 642 acres of sprayfields and 1090 acres of pasture with the same crop combinations and the same proportions as indicated in Table 7-10. This arrangement was run under the same conditions of the previous small and medium farms of soils and profitability. 23,00024,00025,00026,00027,00028,00029,000La NiaNeutralEl NioLBS N LEACHED147,000147,500148,000148,500149,000149,500US$ PROFIT PROFIT Figure 7-17. Medium north-Florida farm, yearly estimated N leaching and profit Results for different ENSO phases regarding N leaching and profitability are shown in Figure 7-18 The N leaching is estimated in 81,741 lbs yr -1 and the profitability is estimated as US$ 597,397 for a Neutral year. 7.5.4 Comparison of the Three Synthesized Farms It is important to understand that dairy farms in north-Florida are highly variable and a comparison of them is only an indication of their potential impacts with these synthesized farms; the estimations should be conducted on a specific farm basis. Also, the comparison should be performed on a whole-dairy-farm basis, since they must be

PAGE 245

223 accountable for their total impacts rather than per acre or per head of livestock; however having these estimations as well as the absolute values are believed to be illustrative. 70,00072,00074,00076,00078,00080,00082,00084,00086,00088,000La NiaNeutralEl NioLBS N LEACHED594,000595,000596,000597,000598,000599,000600,000601,000602,000603,000604,000US$ PROFIT PROFIT Figure 7-18. Large north-Florida farm, yearly estimated N leaching and profit Table 7-11 presents the main management differences among the 3 synthesized farms, the cow/land ratios, and the N leached estimated for the overall farm and the ratios of N leached per cow and per acre of land. The overall environmental impact from N leaching of a large farm is much higher (~80,000 lbs yr -1 ) than the medium (~26,000 lbs yr -1 ), and the medium than the small farm (~14,000 lbs yr -1 ) (Figure 7-19). Because the crop rotations were held proportionally equivalent in the three farms, the ENSO impacts are proportionally similar in the three, with the N leaching highest in El Nio years and lowest in La Nia years. From Table 7-11, the number of head per unit of land is higher on the small farm and lower on the large farm and because of this, the amounts of N leaching when

PAGE 246

224 measured per unit of area or per unit of animal are higher for the small farm than the large farm; it is estimated that a cow in the small farm would leach ~36 lbs N yr -1 while a cow in the large farm would leach ~27 lbs N yr -1 When the comparison is switched to N leached per acre, the differences are even more pronounced, the estimated N leached per acre in the small farm is ~96 lbs N yr -1 while this estimation is ~47 lbs N yr -1 for the large farm. Table 7-11. Comparison of three synthesized north-Florida farms and their estimated annual N leaching Small Medium Large TAC head 400 800 3000 BULLS head 16 30 120 RHA lbs yr -1 17,000 19,000 20,000 Total N Lost Waste System (TNL) percent 28.81 35.31 40 SPRAYFIELDS acre 70 163 642 PASTURE acre 80 150 1090 TOTAL LAND AREA acre 150 313 1732 COWS PER ACRE SPRAYFIELDS head 5.71 4.91 4.67 COWS PER ACRE PASTURE head 5.00 5.33 2.75 TOTAL COWS PER ACRE head 2.67 2.56 1.73 La Nia lbs 13,528 25,096 76,318 Neutral lbs 14,429 26,774 81,471 El Nio TOTAL N LEACHED lbs 15,275 28,389 86,319 La Nia lbs 33.82 31.37 25.44 Neutral lbs 36.07 33.47 27.16 El Nio N LEACHED PER COW lbs 38.19 35.49 28.77 La Nia lbs 193.25 153.96 118.88 Neutral lbs 206.13 164.26 126.90 El Nio N LEACHED PER ACRE SPRAYFIELD lbs 218.21 174.16 134.45 La Nia lbs 169.09 167.31 70.02 Neutral lbs 180.36 178.49 74.74 El Nio N LEACHED PER ACRE PASTURE lbs 190.94 189.26 79.19 La Nia lbs 90.18 80.18 44.06 Neutral lbs 96.19 85.54 47.04 El Nio N LEACHED PER ACRE OF TOTAL LAND lbs 101.83 90.70 49.84

PAGE 247

225 010,00020,00030,00040,00050,00060,00070,00080,00090,000100,000SMALLMEDIUMLARGE La Nia Neutral El Nio Figure 7-19. Overall N leaching from the synthesized farms 7.6 Conclusions The present study demonstrates the feasibility of integrating complex models in a user-friendly application in an interdisciplinary approach that includes socio-economic, bio-physical, and environmental components, the Dynamic North-Florida Dairy Farm Model (DNFDFM). The DNFDFM is capable of assessing N leaching from dairy farm systems and the economic impacts of reducing it, with minimum data inputs through the use of simple commands. Different from previous approaches, the DNFDFM estimates seasonal annual and inter-annual climatic variations responding to environmental and managerial dairy characteristics and can be used to study the economic and ecologic sustainability of dairy farms in north-Florida. Specifically, the DNFDFM is a decision

PAGE 248

226 support system that includes livestock, waste system, crop models, economic, climatic, soil, and optimization components. This paper describes all the DNFDFM components, the details among their interactions, and the specific equations that define their dynamic simulation. From running the simulation part of the DNFDFM for 3 synthesized north-Florida dairy farms, it was found that great differences exist among systems regarding environmental impacts and profitability. Across all of them, higher N leaching occurs in El Nio than Neutral and in Neutral more than La Nia years; the N leaching varies between 13,000 and 86,000 lbs N year -1 farm -1 25 and 33 lbs N year -1 cow -1 and 44 and 102 lbs N year -1 acre -1 Since the model and all its components are well documented and are based on previous models created and/or used by regulatory agencies in Florida, the DNFDFM could serve as an environmental accountability tool for farmers and regulatory agencies as well as an educational tool for Extension.

PAGE 249

CHAPTER 8 ECONOMIC AND ECOLOGIC IMPACTS OF NORTH FLORIDA DAIRY FARM SYSTEMS: TAILORING MITIGATION OF NITROGEN POLLUTION 8.1 Introduction The presence of high levels of N in water is an environmental hazard because it affects human health and ecosystem welfare (Fraisse et al., 1996). The Suwannee River Basin has received much attention in recent years because increased N levels in water bodies (Albert, 2002; Katz et al., 1999; Pitman et al., 1997) and dairies have been identified as an important source of nitrates in the Basin (Andrew, 1994; Katz and DeHan, 1996; Berndt, 1998). It is critical to keep individual dairy farms environmentally accountable (Ondersteijn et al., 2002), to distinguish between non-point and point sources of pollution (Andrew, 1992), and to test management options that may decrease environmental impacts of dairy farms (Suwannee River Partnership, 200 3 ). The objective of this paper is twofold Present and discuss the ecologic and economic impacts of the north-Florida dairy farm systems under different management strategies based on the results of simulating them by using the Dynamic North-Florida Dairy Farm Model (DNFDFM) Test combinations of management options that may decrease N leaching in specific dairy farm systems while maintaining profitability First, a statistical analysis of the main sources of variation based on the results from simulating multiple management strategies in a Synthesized north-Florida dairy farm system is presented. Second, the sources of variation in N leaching and profitability of the Synthesized dairy farm system are graphically discussed. Third, an analysis of 227

PAGE 250

228 optimization that includes reduced N leaching and potential increased profit for the Synthesized farm is presented. And fourth, some cases of real-life applications of the DNFDFM on north-Florida dairy farms are discussed. Any north-Florida farm is a unique production unit in several aspects. It is recommended that the DNFDFM be used separately for each specific case. Results of this paper are indicative of the usefulness of the model in practical applications. 8.2 Materials and Methods 8.2.1 Dynamic North-Florida Dairy Farm Model The DNFDFM as described in Chapters 5 and 7 was used as a diagnostic and analytic tool to simulate N leaching and profit under existing conditions and then to optimize practices under different farm scenarios. The DNFDFM is a simple event-controlled, mechanistic, dynamic model coupled to crop and optimization models created to assess N leaching from north-Florida dairy farm systems and the economic impacts resulting from reducing it. The DNFDFM responds to dairy-specific environmental (climate and soils) and managerial characteristics (total number of cows, milk production, crop land, crop rotations, waste system, etc.) The driver of the DNFDFM is a dynamic adaptation of the framework balance and/or budgeting of nutrients in dairy farms, commonly used in Florida: Van Horn et al. (1991, 1994, 1998, 2001); NRCS (2001). The DNFDFM incorporates Markov-chain probabilistic simulation of cow-flows and crop simulation for historical climatic years (El Nio Southern Oscillation), over a period of 43 years (1956-1998); automated optimization of managerial options for different ENSO phases; participatory modelling; and user-friendliness.

PAGE 251

229 The DNFDFM has both simulation and optimization functions. For effects of analyses, simulation refers to a dynamic run of the model for a year and optimization means the process of solving a linear programming optimization matrix built by the results of multiple simulation runs that combine chosen dairy management scenarios. 8.2.2 A Synthesized North-Florida Dairy Farm In order to study the sources of variation in N leaching and profitability as related to managerial options in a whole north-Florida dairy farm while maintaining confidentiality of specific farms, a Synthesized farm was created using real data. This Synthesized farm combines many real characteristics of dairies in the study area for which the DNFDFM can be run to perform relevant analyses. This farm has 400 total adult cows (TAC), 16 bulls, and raises 100% of its heifers. The rolling herd average (RHA) is 17,000 lbs cow -1 yr -1 and milk production is 100% seasonal. The waste management system includes flushing-lagoon-spraying and has a total N loss (TNL) of 28.81%. When effluent is applied to spray fields, an estimated additional 30% of N is volatilized. When cows directly deposit manure in pastures, an estimated 40% of N is lost through volatilization before it becomes available for the crops. The farm has 70 acres of sprayfields and 80 acres of pasture divided into several fields. There is an overall net income of US$ 0.89 cwt milk -1 (Table 8-1). The Synthesized farm was used for the statistical, graphical, and optimization analyses from sections 2.3, 2.4, and 2.5; while real farm operations were used for analyses in section 2.6. 8.2.3 Statistical Analyses The DNFDFM was run for the Synthesized farm and various potential environmental and managerial options as scenarios that could impact N leaching and

PAGE 252

230 profitability. Scenarios were created by combinations of 10 north-Florida dairy farm soil types, three ENSO phases (La Nia, Neutral, and El Nio climatic years), two levels of confined time (CT, 80% and 60%), 2 combinations of forage sequences in pasture, 10 combinations of forage sequences in sprayfields, and 2 levels of protein in the diet (low and high) (Table 8-2). Table 8-1. Characteristics of a Synthesized north-Florida dairy farm TAC head 400 Bulls head 16 Heifers percent 100% RHA lbs year -1 17,000 Seasonality percent 100% TNL percent 28.81% N Volatilized Sprayfields percent 30% N Volatilized Pasture percent 40% Sprayfields Acre 70 Pasture Acre 80 Net Income US$ cwt -1 milk 0.89 The number of combinations for these scenarios was: 12 (months) x 10 (soils) x 3 (ENSO phases) x 2 (CT) x 2 (Pasture) x 10 (Sprayfields) x 2 (CP) = 28,800. Results were measured as monthly estimations of N leaching and profit. Statistical analyses were performed based on the 28,800 combinations by fitting mixed effects models (Litell et al., 1996) using the PROC MIXED procedure in SAS (SAS institute, 1992). All factors were analyzed as well as their interactions with a 0.05 level of significance. 8.2.4 Graphical Analyses Graphical analysis was used to visualize and help diagnose the effects of the main factors in the trade-off of N leaching and profit. A series of graphs with selected factor levels was prepared to show the combined effects of these factors with emphasis on

PAGE 253

231 ENSO climatic phases and Soil series and crop systems in sprayfields; pastures and crude protein in the diet; confined time and crude protein in the diet Table 8-2. Combination of management strategies simulated by the DNFDFM for the statistical analysis Soil Types 1 Arredondo-Gainesville-Millhopper 2 Arredondo-Jonesville-Lake 3 Bonneau-Blanton-Eunola 4 Penney-Otela 5 Penney-Kershaw 6 Millhopper-Bonneau 7 Otela-Jonesville-Seaboard 8 Blanton(high)-Lakeland 9 Blanton(low) 10 Blanton-Ortega-Penny ENSO Phases 1 La Nia 2 Neutral 3 El Nio Confined Time (CT) 1 80% 2 60% Forages Pasture (PASTURE) 1 bahiagrass-bahiagrass-winter forage 2 bermudagrass-bermudagrass-winter forage Forages Sprayfields (FIELDS) 1 Bahiagrass-Bahiagrass-Winter Forage 2 Corn-Bermudagrass-Winter Forage 3 Corn-Bahiagrass-Winter Forage 4 Corn-Millet-Winter Forage 5 Corn-Corn-Winter Forage 6 Millet-Sorghum-Winter Forage 7 Sorghum-Corn-Winter Forage 8 Millet-Corn-Winter Forage 9 Corn-Sorghum-Winter Forage 10 Bermudagrass-Bermudagrass-Winter Forage Crude Protein (CP) 1 Low 2 High In order to compare complete sets of management options, data were compiled in yearly outputs regarding N leaching and profit as the total sum of the individual months from October of one year to September of the next year. The information was summarized in a database from which queries were used to retrieve specific information in order to prepare meaningful figures.

PAGE 254

232 8.2.5 Optimization Analyses This section analyzed potential management strategies the Synthesized farm might implement to decrease environmental impacts while maintaining profitability. After the farm was simulated, the optimization tool from the DNFDFM was used to minimize (maximize) N leaching (profit) in the Synthesized farm. The Synthesized farm, with an initial set of crop sequences in pastures and sprayfields, was optimized by changing crude protein (CP), confined time (CT), crop sequences in pastures, and crop sequences in sprayfields. These options were selected for the optimization because they are the practical ways north-Florida dairy farms can change their management choices in the short or medium run. A comparison of current farm practices with practices suggested by the optimizer was performed. This comparison is designed to help the farm management decide on feasible modifications environmentally superior to the initial conditions. This set of feasible practices was used again to simulate the farm with the DNFDFM and compare these results with those of the original farm practices. 8.2.6 Validation with Real Farms The simulation and optimization process described above was then performed individually with three real farms in north-Florida. Each owner was visited and the process of simulation and optimization, with discussion and analysis of feasible modifications was achieved with each one of them. Each farmer collaborated with the process by inputting detailed data on their operations to feed the model. After filling in the appropriate information, a simulation run was performed to show the performance of the model and validate the results with the

PAGE 255

233 reports and knowledge from the farmers. This on-site validation was crucial to gain confidence of the farmers in the performance of the model. Once the validation was concluded, the optimization tool was explained and used to optimize each farm situation. Followed, the results of the optimization were discussed with the farmers. From this discussion, a feasible set of management choices was selected and the DNFDFM simulator was run again in order to obtain the final results. For anonymity purposes, the results are presented in absolute numbers with only the required details on the characteristics of each one of the farms. The main objective of this section, besides validation, was to demonstrate the utility of the DNFDFM for practical applications in real-life situations. 8.3 Results and Discussion 8.3.1 Statistical Analyses 8.3.1.1 Nitrogen leaching Results of the significance of the factors and their main interactions regarding the estimation of N leaching are presented in Table 8-3. The statistical model has an R-square equal to 0.9989 and is significant at the 0.01% probability level There are significant differences for all main factors: soils, month of the year, ENSO phase, confined time, pasture, sprayfields, and crude protein. For example, ENSO phases and their associated climatic patterns determine significantly different amounts of N leaching. Some interactions among factors were also statistically significant as was the case of interactions among month and ENSO phases, sprayfield rotations, soils, crude protein levels, confined time levels, and pasture rotations; interactions among pasture rotations and sprayfields rotations, crude protein levels, and confined time levels; interaction

PAGE 256

234 among sprayfield rotations and confined time; and interactions among month, pasture rotations, and crude protein levels. This means that there are statistically significant differences among these interaction combinations as is the case for example of month of the year and soil types, in which, the N leaching among types of soils are statistical different in each month of the year. Table 8-3. Significance of factors and main interactions for N leaching Source DF Type I SS Mean Square Pr > F Soil 9 287.36 31.93 111352 <.0001 Month 11 6100.83 554.62 1934266 <.0001 ENSO 2 3.48 1.74 6075 <.0001 CT 1 10.89 10.89 37966 <.0001 Pasture 1 19.11 19.11 66638 <.0001 Sprayfields 9 119.98 13.33 46492 <.0001 CP 1 26.76 26.76 93310 <.0001 Month*ENSO 22 99.32 4.51 15745 <.0001 Month*Sprayfields 99 311.10 3.14 10959 <.0001 Pasture*Sprayfields 9 0.63 0.07 243 <.0001 SOIL*Month 99 738.79 7.46 26026 <.0001 CT*Sprayfields 9 2.38 0.26 923 <.0001 Pasture*CP 1 0.00 0.00 14 0.0002 CT*Pasture 1 1.86 1.86 6485 <.0001 Month*CP 11 0.14 0.01 45 <.0001 Month*CT 11 5.00 0.45 1585 <.0001 Month*Pasture 11 24.57 2.23 7790 <.0001 Month*Pasture*CP 11 0.01 0.00 1.96 0.0281 F Value Interactions among month of the year with sprayfield or pasture rotations for the months December, January, and February were not significant; this was because the crop rotation in winter months is the same, as explained in the model description. The potential combinations of other factors not presented in Table 8-3 were not statistically significant at the 0.05 level. 8.3.1.2 Profit Similar statistical analyses were performed for the profit variable (Table 8-4). This model has an R-square of 0.9941 and is significant at the 0.01% probability level.

PAGE 257

235 Table 8-4. Significance of factors and main interactions for profit Source DF Type I SS Mean Square F Value Pr > F 9 287.36 31.93 111352 <.0001 Month 11 6100.83 554.62 1934266 <.0001 ENSO 2 3.48 1.74 6075 <.0001 CT 10.89 10.89 37966 <.0001 Pasture 1 19.11 19.11 66638 <.0001 Sprayfields 9 119.98 13.33 46492 <.0001 CP 1 26.76 26.76 93310 <.0001 Month*ENSO 22 99.32 4.51 15745 <.0001 Month*Sprayfields 99 311.10 3.14 10959 <.0001 Pasture*Sprayfields 9 0.63 0.07 243 <.0001 Soil*Month 99 738.79 7.46 26026 <.0001 CT*Sprayfields 9 2.38 0.26 923 <.0001 Pasture*CP 1 0.00 0.00 14 0.0002 CT*Pasture 1 1.86 1.86 6485 <.0001 Month*CP 11 0.14 0.01 45 <.0001 Month*CT 11 5.00 0.45 1585 Soil 1<.0001 Significant effect exist in the estimated profit for all main factors and for the interactions among month and ENSO phases, sprayfields rotations, and soils; soil and ENSO phases and sprayfield rotations; and for the interaction among soils, months and ENSO phases. All the individual interactions shown in Table 8-4 were highly significant (P<0.001), with the exception of the months of December, January, and February for the sprayfield and pasture rotation interaction, again because the crop rotation in winter does not change. 8.3.2 Graphical Analyses 8.3.2.1 Soil series Figure 8-1 illustrates N leaching and profitability for soil types in north-Florida dairy farms. Soils of type 6 (Millhopper-Bonneau) are those that leach the most N, but with a medium-low profit level. Soils of type 3 (Bonneau-Blanton-Eunola) are the second highest in N leaching and those with the lowest net return.

PAGE 258

236 552763918104569374101822810493678,00010,00012,00014,00016,00018,00020,00061,00062,00063,00064,00065,00066,00067,00068,00069,00070,00071,000 LA NIA NEUTRAL EL NIO Soil Series 1 Arredondo-Gainesville-Millhopper2 Arredondo-Jonesville-Lake3 Bonneau-Blanton-Eunola4 Penney-Otela5 Penney-Kershaw6 Millhopper-Bonneau7 Otela-Jonesville-Seaboard8 Blanton(high)-Lakeland9 Blanton(low)10-Blanton-Ortega-Penny Profit US$/year N leachinglbs/year CT = 80%; CP = HIGHPASTURE = BERMUDAGRASSSPRAYFIELD = BERMUDAGRASS 6620,000 LA NIA NEUTRAL EL NIO 55541081936722811047397394108210,00012,00014,00016,00018,000 8,00061,00062,00063,00064,00065,00066,00067,00068,00069,00070,00071,000 Profit US$/year 555410819367228110473967639410828,00010,00012,00014,00016,00018,00020,00061,00062,00063,00064,00065,00066,00067,00068,00069,00070,00071,000 LA NIA NEUTRAL EL NIO Profit US$/year N l eac hi nglbs/year CT = 80%; CP = HIGHPASTURE = BAHIAGRASSSPRAYFIELD = BERMUDAGRASS CT=80%; CP=HIGH PASTURE=BAHIAGRASS SPRAYFIELD=BERMUDAGRASS C A B N l eac hi nglbs/year CT = 80%; CP = HIGHPASTURE = BERMUDAGRASSSPRAYFIELD = CORN-SORGHUM CT=80%; CP=HIGH PASTURE=BERMUDAGRASS SPRAYFIELD=BERMUDAGRASS CT=80%; CP=HIGH PASTURE=BERMUDAGRASS SPRAYFIELD=CORN-SORGHUM Figure 8-1. Nitrogen leaching and profit for different soil types in north-Florida: Estimations for different ENSO phases for common crop systems

PAGE 259

237 These soils (6 and 3) are very sandy and with very low water holding capacity; soil type 6 is the shallowest and because of that the highest N leaching. This, however, is in certain measure favorable for plant growth and because of that the profitability is not as low as with soil type 3. On the other hand, soils of type 2 (Arredondo-Jonesville-Lake) has the least N leaching of all, followed by soils type 1, 8 and 10 (Arredondo-Gainesville-Millhopper, Blanton (high)-Lakeland, and Blanton-Ortega-Penny), which also have the greatest N uptake and biomass accumulation expressed in farm profitability. Also in Figure 8-1 the marked differences among ENSO phases can be observed; there will always be higher N leaching and lower profitability for El Nio years than Neutral years and in Neutral years more N leaching and less profitability than La Nia years. By exploring the circles with the same number (same soil type) in Figure 8-1, it can be noticed that there is greater difference between La Nia and Neutral years than El Nio and Neutral years, meaning that N leaching and profit are substantially superior to the normal or neutral years during La Nia years. It can also be observed that the distance among different ENSO phases (Figure 8-1 A, B, and C) varies slightly under different scenarios of changing pasture crop systems and/or crop systems in the sprayfields. By using crop systems that include bermudagrass in spring and summer in both pastures and sprayfields, the lowest N leaching is observed (Figure 8-1A). The inclusion of a corn-sorghum rotation in the sprayfields would increase the profitability, but it would also increase N leaching (Figure 8-1B). If the change includes bahiagrass in the pasture, this would increase slightly the N leaching and decrease the profit (Figure 8-1C).

PAGE 260

238 8.3.2.2 Crop systems in sprayfields Figure 8-2 displays the relationships among 10 different crop systems in the sprayfields. Evidently, rotation 1 (bermudagrass-bermudagrass) outperforms with the least N leaching system with a medium-low profitability. It is followed by crop system 9 (corn-bermudagrass) which has medium-to-high profitability. Rotations 8 (corn-bahiagrass), 6 (corn-corn) have low-medium N leaching and medium-high profitability. Crop system 10 that includes bahiagrass-bahiagrass has low-medium N leaching, but the lowest profitability of all. On the other hand, system 5 (millet-sorghum), followed by crop systems 7 (corn-millet), and 2 (corn-sorghum) presents the highest levels of N leaching with medium-high profit. The most profitable crop system is rotation 3 (millet-corn) that has a medium level of N leaching. As in the previous case, higher N leaching and lower profitability are observed for El Nio, followed by Neutral years. La Nia years presents the lowest N leaching and the highest profit. Even though the differences among ENSO phases across scenarios seem to be similar, there are slight differences that are attributed to other factors such as crops in pastures and soil types (Figure 8-2 A, B, and C). It is also noticeable that the relative differences among ENSO phases vary among crop systems: i.e., there is less difference in crop system 6 between El Nio and Neutral years compared with crop systems 10 and 1. Confirming the results from Figure 8-1, there is less N leaching with soils type 2 (Figure 8-2C), more with soils type 6 (Figure 8-2B), and medium with soils type 4 (Figure 8-2A). Additionally, bermudagrass in pasture (Figure 8-2C) results in lower N leaching than bahiagrass (Figure 8-2 A, B). With soils type 2 and bermudagrass in pastures, a slightly higher profitability is observed (Figure 8-2C).

PAGE 261

239 5527639181045693741108221081493678,00010,00012,00014,00016,00018,00020,00022,00061,00062,00063,00064,00065,00066,00067,00068,00069,00070,00071,00072,000 LA NIA NEUTRAL EL NIO Profit US$/year N leachinglbs/year Crop systems: 1 bermudagrass-bermudagrass2 corn-sorghum3 millet-corn4 sorghum-corn5 millet-sorghum6 corn-corn7 corn-millet8 corn-bahiagrass9 corn-bermudagrass10-bahiagrass-bahiagrass CT = 80%; CP = HIGHPASTURE = BAHIAGRASSSOIL SERIE 4: PENNEY-OTELA 5552473734220,00022,000 LA NIA NEUTRAL EL NIO 410819367281019669181010,00012,00014,00016,00018,000 8,00061,00062,00063,00064,00065,00066,00067,00068,00069,00070,00071,00072,000 Profit US$/year 5527639181045693741108221081493678,00010,00012,00014,00016,00018,00020,00022,00061,00062,00063,00064,00065,00066,00067,00068,00069,00070,00071,00072,000 LA NIA NEUTRAL EL NIO Profit US$/year N l eac hi nglbs/year CT = 80%; CP = HIGHPASTURE = BERMUDAGRASSSOIL SERIE 2: ARREDONDO-JONESVILLE-LAKE CT=80%; CP=HIGH PASTURE=BERMUDAGRASS SOIL SERIE 2=ARREDONDO-JONESVILLE-LAKE N l eac hi nglbs/year CT = 80%; CP = HIGHPASTURE = BAHIAGRASSSOIL SERIE 6: MILLHOPPER-BONNEAU ABC CT=80%; CP=HIGH PASTURE=BAHIAGRASS SOIL SERIE 4=PENNEY-OTELA CT=80%; CP=HIGH PASTURE=BAHIAGRASS SOIL SERIE 6=MILLHOPPER-BONNEAU Figure 8-2. Nitrogen leaching and profit for different crop systems in north-Florida: Estimations for different ENSO phases for common pastures and soil types

PAGE 262

240 8.3.2.3 Pastures and crude protein in the diet (CP) Figure 8-3 shows four management options of the combinations of two pasture crop systems (bermudagrass and bahiagrass) and two CP choices (low and high). Consistently, option 4 (bermudagrass and low CP) has the lowest N leaching and the highest profit; conversely, option 1 (bahiagrass and high CP) has the highest N leaching and the lowest profit. The CP in the diet has greater effect on N leaching than the pasture crop as evidenced in the fact that option 2 (bahiagrass and low CP) presents less N leaching than option 3 (bermudagrass and high CP). However the pasture crop has greater effect than the CP over profitability: option 3 results in higher profit levels than option 2. As in previous cases, the lowest N leaching and the highest profit are observed in La Nia years; Neutral and El Nio years that are closer together present higher N leaching and lower profitability. Having bahiagrass or corn-corn in the sprayfield would leach N very similarly, with corn-corn being slightly lower (Figure 8-3 A, C); however, the profitability changes substantially with these two choices. For example, with bahiagrass the estimated profitability for option 3 in a Neutral year would be lower than US$ 63,000 while it would be more than US$ 66,000 when corn-corn is the system in the sprayfield for this Synthesized farm. The millet-sorghum rotation in the sprayfield would leach substantially higher amounts than the other sprayfield options (bahiagrass and corn-corn) and it would present profitability slightly lower than the corn-corn rotation (Figure 8-3B).

PAGE 263

241 1324314242318,00010,00012,00014,00016,00018,00020,00061,00062,00063,00064,00065,00066,00067,00068,00069,00070,00071,000 LA NIA NEUTRAL EL NIO Profit US$/year N leachinglbs/year Management options: 1 Pasture = bahiagrass CP = high2 Pasture = bahiagrass CP = low3 Pasture = bermudagrass CP = high 4 Pasture = bermudagrass CP = low CT = 80%SPRAYFIELD = BAHIAGRASSSOIL SERIE 9: BLANTON (low) 20,000 LA NIA NEUTRAL EL NIO 42312413132410,00012,00014,00016,00018,000 8,00061,00062,00063,00064,00065,00066,00067,00068,00069,00070,00071,000 Profit US$/year 1324314242318,00010,00012,00014,00016,00018,00020,00061,00062,00063,00064,00065,00066,00067,00068,00069,00070,00071,000 LA NIA NEUTRAL EL NIO Profit US$/year N leachinglbs/year CT = 80%SPRAYFIELD = CORN-CORNSOIL SERIE 9: BLANTON (low) CT=80% SPRAYFIELD=CORN-CORN SOIL SERIE 9=BLANTON (LOW) CT = 80%SPRAYFIELD = MILLET-SORGHUMSOIL SERIE 9: BLANTON (low) ABC CT=80% SPRAYFIELD=BAHIAGRASS SOIL SERIE 9=BLANTON (LOW) CT=80% SPRAYFIELD=MILLET-SORGHUM SOIL SERIE 9=BLANTON (LOW) N leachinglbs/year Figure 8-3. N leaching and profit for different pasture crops and crude protein in north-Florida: Estimations for different ENSO phases for common sprayfield crops and soil type

PAGE 264

242 8.3.2.4 Confined time (CT) and crude protein in the diet (CP) Figure 8-4 displays four management options of the combinations of two possible CT options (80% and 60%) and two CP choices (low and high). Option 4 (CT=60% and CP=low) leaches the least and has the second highest profit and option 1 (CT=80% and CP=high) leaches the most and has the lowest profit. The effect of the CP in the diet is higher than the 20% CT decrease (from 80 to 60%) in the N leaching because option 3 (CT=80% and CP=low) leaches less N than option 2 (CT=60% and CP=high). However, option 3 presents the highest profitability of all and option 2 has the second lowest profit. By comparing graphs A, B, and C in Figure 8-4 it can be noticed that the lowest N leaching is for the option of having bermudagrass in the pasture and be located in a soil type 5 (Penney-Kershaw). This is closely followed by the option of changing the pasture to bahiagrass, indicating that the effect of reducing N leaching of bermudagrass over bahiagrass in the pasture is not as marked as happens in the sprayfields. The option that includes bermudagrass (Figure 8-4C) has more profit. With the same crop systems, soil type 5 shows substantially greater N leaching and decrease in profit compared with soil type 3. N leaching is the lowest and profit is the highest when La Nia ENSO phases occur, however this is more marked for soil type 5 than soil type 3. Also, the highest N leaching and the lowest profitability in El Nio years is more marked with soil type 3 (Figure 8-4 B, C). By exploring relative positions among points in Figure 8-4 A, B, and C, it can be observed that, depending of the management options, there are slight differences in the spatial distributions of the circles.

PAGE 265

243 4231241313248,00010,00012,00014,00016,00018,00020,00061,00062,00063,00064,00065,00066,00067,00068,00069,00070,00071,000 LA NIA NEUTRAL EL NIO Profit US$/year N leachinglbs/year Management options: 1 CT = 80% CP = high2 CT = 60% CP = high3 CT = 80% CP = low 4 -CT = 60% CP = low PASTURE = BAHIAGRASSSPRAYFIELD = CORN-MILLETSOIL SERIE 5: PENNEY-KERSHAW 20,000 LA NIA NEUTRAL EL NIO 13243142423110,00012,00014,00016,00018,000 8,00061,00062,00063,00064,00065,00066,00067,00068,00069,00070,00071,000 Profit US$/year 1324314242318,00010,00012,00014,00016,00018,00020,00061,00062,00063,00064,00065,00066,00067,00068,00069,00070,00071,000 LA NIA NEUTRAL EL NIO Profit US$/year N leachinglbs/year PASTURE = BERMUDAGRASSSPRAYFIELD = CORN-MILLETSOIL SERIE 3: BONNEAU-BLANTON-EUNOLA PASTURE=BERMUDAGRASS SPRAYFIELD=CORN-MMILLET SOIL SERIE 3=BONNEAU-BLANTON-EUNOLA N leachinglbs/year PASTURE = BERMUDAGRASSSPRAYFIELD = CORN-MILLETSOIL SERIE 5: PENNEY-KERSHAW ABC PASTURE=BAHIAGRASS SPRAYFIELD=CORN-MMILLET SOIL SERIE 5=PENNEY-KERSHAW PASTURE=BERMUDAGRASS SPRAYFIELD=CORN-MMILLET SOIL SERIE 5=PENNEY-KERSHAW Figure 8-4. N leaching and profit for different confined time and crude protein: Estimations for different ENSO phases for common crop systems and soils

PAGE 266

244 This means that the management options not only change the overall N leaching and profit, but also they have interactions with the climate, the CT and the CP. For example, management option 1 that always has the most leaching, is more different between El Nio and Neutral years and less different between Neutral and La Nia years in Graph C compared with Graphs A and/or B, In Graphs B and C, it presents the lowest profitability. However in Graph A, it appears to be the second lowest in profit. 8.3.2.5 Isolated effects of single managerial changes in the Synthesized farm Figure 8-5 displays the impacts of four management changes, one at a time, regarding A) sprayfield crop system, B) pasture crop system, C) CT, and D) CP for the Synthesized farm. The initial conditions were set up to be sprayfield rotation = millet-sorghum, pasture = bahiagrass, CT = 80%, and CP = high. All the tested changes in Figure 8.5: A) sprayfield crop system = bermudagrass, B) pasture crop = bermudagrass, C) CT = 60%, and D) CP = low decrease the overall N leaching. However this decreasing of N leaching is, by far, greater when the change is in the sprayfield rotation indicating the high sensitivity of the farm to the crop systems installed on sprayfields. The second highest decrease in N leaching occurs when the CP is decreased from high to low levels. The third highest decrease is by changing the CT (from 80 to 60%). The least decrease in N leaching happens when the pasture system is changed from bahiagrass to bermudagrass, indicating that dairy systems may not be as sensitive to the pastures as to the other factors. The effect of these changes in the overall profit varies. For the case of maximum N leaching decrease (sprayfield sorghum-millet to bermudagrass), there is also a maximum decrease in the profitability of the farm. For the change in the CT, there is no variation in

PAGE 267

10,00011,00012,00013,00014,00015,00016,00064,00064,50065,00065,50066,00066,50067,00067,50068,00068,500 LA NIA NEUTRAL EL NIO Profit US$/year N leachinglbs/year CT = 80%; CP = HIGHPASTURE = BAHIAGRASSSPRAYFIELD = MILLET-SORGHUM CT = 80%; CP = HIGHPASTURE = BAHIAGRASSSPRAYFIELD = BERMUDAGRASS 10,00011,00012,00013,00014,00015,00016,00064,00064,50065,00065,50066,00066,50067,00067,50068,00068,500 LA NIA NEUTRAL EL NIO Profit US$/year N leachinglbs/year CT = 80%; CP = HIGHPASTURE = BAHIAGRASSSPRAYFIELD = MILLET-SORGHUM CT = 80%; CP = HIGHPASTURE = BERMUDAGRASS SPRAYFIELD = MILLET-SORGHUM 10,00011,00012,00013,00014,00015,00016,00064,00064,50065,00065,50066,00066,50067,00067,50068,00068,500 LA NIA NEUTRAL EL NIO Profit US$/year N leachinglbs/year CT = 80%; CP = HIGHPASTURE = BAHIAGRASSSPRAYFIELD = MILLET-SORGHUM CT = 60% ; CP = HIGHPASTURE = BAHIAGRASSSPRAYFIELD = MILLET-SORGHUM 10,00011,00012,00013,00014,00015,00016,00064,00064,50065,00065,50066,00066,50067,00067,50068,00068,500 LA NIA NEUTRAL EL NIO Profit US$/year N leachinglbs/year CT = 80%; CP = LOW PASTURE = BAHIAGRASSSPRAYFIELD = MILLET-SORGHUM CT = 80%; CP = HIGHPASTURE = BAHIAGRASSSPRAYFIELD = MILLET-SORGHUM CT=80%; CP=HIGHPASTURE=BAHIAGRASS SPRAYFIELD=SORGHUM CT=80%; CP=LOW PASTURE=BAHIAGRASS SPRAYFIELD=SORGHUM 245 A B C D CT=80%; CP=HIGHPASTURE=BAHIAGRASS SPRAYFIELD=SORGHUM CT=80%; CP=HIGH PASTURE=BAHIAGRASS SPRAYFIELD=BERMUDAGRASS CT=80%; CP=HIGHPASTURE=BAHIAGRASS SPRAYFIELD=SORGHUM CT=80%; CP=HIGH PASTURE=BERMUDAGRASS SPRAYFIELD=SORGHUM CT=80%; CP=HIGHPASTURE=BAHIAGRASS SPRAYFIELD=SORGHUM CT=60% ; CP=HIGHPASTURE=BAHIAGRASS SPRAYFIELD=SORGHUM Figure 8-5. Impacts of management changes for the Synthesized farm with soils type 4 (Penney-Otela). A) Sprayfield crop system. B) Pasture crop system. C) Confined time (CT). D) Crude protein (CP).

PAGE 268

246 the profitability level. And for the other options (change in pasture and CP) there is a slight increase in farm profitability. Even though in Figure 8-5 the changes in N leaching and profitability seem to be of similar magnitudes, they are quite different in terms of proportions. As a percent, the decrease in N leaching is five or six times greater than the change in profitability. This has great connotation for the potential changes farmers could adopt in order to pursue environmental and economic sustainability. 8.3.3 Optimization Analysis: Decreasing Environmental Impacts and Increasing Profitability 8.3.3.1 The Synthesized dairy farm as is The Synthesized north-Florida dairy farm was simulated with the following management practices: crude protein in the diet high, CT 80%, and the forage crops described in Table 8-5. This is a common scenario for many north-Florida dairy farms. Table 8-5. Crops in different fields in the Synthesized farm Field Area (Ac) Type Spring Summer Winter 1 10 SPRAYFIELD CORN SORGHUM RYE 2 20 SPRAYFIELD CORN MILLET RYEGRASS 3 10 SPRAYFIELD SORGHUM MILLET WHEAT 4 20 SPRAYFIELD BAHIAGRASS BAHIAGRASS OATS 5 10 SPRAYFIELD MILLET SORGHUM RYE 6 20 PASTURELAND BAHIAGRASS BAHIAGRASS RYE 7 10 PASTURELAND BERMUDAGRASS BERMUDAGRASS RYEGRASS 8 20 PASTURELAND BAHIAGRASS BAHIAGRASS WHEAT 9 10 PASTURELAND BERMUDAGRASS BERMUDAGRASS OATS 10 20 PASTURELAND BAHIAGRASS BAHIAGRASS RYE The yearly results for profit and N leaching are summarized in Figure 8-6. There is lower N leaching in La Nia years than Neutral years (6.67%) and Neutral years have lower N leaching than El Nio years (5.86%). El Nio years present 12.92% more N leaching than La Nia years (Figure 8-6). Conversely, profit is lower in El Nio years

PAGE 269

247 than Neutral years (1.47%) and Neutral years present lower profit than La Nia years (0.45%). El Nio years result in 1.47% less profit than in La Nia years. N leaching varies from 13,528 lbs for the whole farm for a La Nia year to 15,275 lbs for an El Nio year. An estimated 14,429 lbs of N are leached in a Neutral year. Profit varies from US$ 65,480 for an El Nio year to US$ 66,455 in a La Nia year. An intermediate figure of US$ 65,773 was obtained for a Neutral year. 10,00011,00012,00013,00014,00015,00016,000La NiaNeutralEl NioLBS N LEACHED65,40065,90066,40066,90067,40067,90068,40068,900US$ PROFIT PROFIT Figure 8-6. Nitrogen leaching and profit with current practices for a Synthesized north-Florida farm 8.3.3.2 Optimization of the Synthesized farm The linear programming optimizer included in the DNFDFM included combinations of the following management options: confined time (CT), crude protein (CP), pasture forage rotations, and sprayfield forage sequences. The optimizer found a set of management practices with better characteristics of N leaching and profit when compared with the baseline situation of the farm. The selected management options are

PAGE 270

248 summarized in Table 8-6 and results comparing with the original situation are presented in Figure 8-7. The management strategies selected by the optimizer were a low CP level; a variable level of CT, 60% of CT for cows that will use 59% of the pasture and 80% of CT for cows that use 41% of the pasture; a bermudagrass-bermudagrass-winter forage sequence for all pasture land; variable areas of forages for sprayfields for different ENSO phases; as seen in Table 8-6. With these combinations, the optimizer estimated that N leaching would vary from 10,146 lbs year -1 for a La Nia year to 11,497 lbs year -1 for an El Nio year. An intermediate figure of 10,838 lbs year -1 was the outcome for a Neutral year (Figure 8-7). Table 8-6. Management strategies selected by the optimizer CP low CT 60% 59% of the pasture 80% 41% of the pasture Pasture bermudagrass-bermudagrass-winter forage Sprayfields Acres La Nia 22.72 corn-corn-winter forage 15.44 bermudagrass-bermudagrass-winter forage 13.51 corn-bermudagrass-winter forage 9.57 millet-corn-winter forage 8.75 corn-bermudagrass-winter forage Neutral 20.90 corn-corn-winter forage 16.74 bermudagrass-bermudagrass-winter forage 12.22 corn-bermudagrass-winter forage 11.22 millet-corn-winter forage 8.93 corn-bermudagrass-winter forage El Nio 21.52 corn-corn-winter forage 15.97 bermudagrass-bermudagrass-winter forage 12.99 corn-bermudagrass-winter forage 11.20 millet-corn-winter forage 8.33 corn-bermudagrass-winter forage

PAGE 271

249 Comparing these values with the previous results of the farm as is, substantial variations are noticed. N leaching could be decreased up to 25% and profit could still be increased by approximately 3.15%, in all ENSO phases. 8.3.3.3 Feasible practices for the Synthesized farm In reality, farmers cannot easily change all practices proposed by the optimizer. For instance, the slight changes in areas of the sprayfields may not be feasible. However, dairy farmers would be able to change some or part of the proposed practices. The optimization results serve as the guidelines for making such changes, taking into account that these optimization results would present better environmental outputs. By cross-referencing results achieved by the optimizer with the proposed practices, changes that are feasible for farmers can be determined. Current Practice Current Practice Current Practice Optimized Optimized Optimized Profit (Current Practice) Profit (Optimized)10,00011,00012,00013,00014,00015,00016,000La NiaNeutralEl NioLBS N LEACHED65,40065,90066,40066,90067,40067,90068,40068,900US$ PROFIT Figure 8-7. Optimized vs. current practice yearly N leaching and profit for a north-Florida Synthesized farm

PAGE 272

250 Scenario One. The optimization proposes a decrease in CP in the diet from high to low. This is a possible change for a dairy farmer to make. Next, the optimization proposes to switch the CT between 80 and 60% for different group of cows. It is believed that it would be difficult for the farmer to implement this specific change, so CT levels would remain at 80%. Finally, the optimization proposed a series of crops for different ENSO phases. It is assumed that the farmer can rather easily implement similar, though not identical, changes to the farm management practices. Feasible crop systems for the farm are shown in Table 8-7. This implements bermudagrass instead of bahiagrass in pastures and sprayfields, and includes more in corn than sorghum and/or millet in sprayfields. With these feasible combinations the DNFDFM simulation was re-run. Results estimated that the N leaching would vary from 10,410 lbs year -1 for a La Nia year to 11,820 lbs year -1 for an El Nio year. A Neutral year would result in N leaching of 11,130 lbs year -1 (Figures 8-8 and 8-9). Table 8-7. Feasible crop systems in Synthesized dairy farm Field Area (Ac) Type Spring Summer Winter 1 23 SPRAYFIELD CORN CORN RYE 2 15 SPRAYFIELD BERMUDAGRASS BERMUDAGRASS RYEGRASS 3 14 SPRAYFIELD CORN BERMUDAGRASS WHEAT 4 19 SPRAYFIELD CORN CORN OATS 5 80 PASTURE BERMUDAGRASS BERMUDAGRASS OATS Comparing these N leaching values with the original, farm as is, it is noticed that N leaching could be greatly reduced by using a feasible set of practices instead of the as is practices. By implementing these practices, N leaching would be reduced in the order of approximately 23% in all ENSO phases, and the profit could still increase by 2.5%.

PAGE 273

251 10,00011,00012,00013,00014,00015,00016,000La NiaNeutralEl NioLBS N LEACHED65,40065,90066,40066,90067,40067,90068,40068,900US$ PROFIT PROFIT Figure 8-8. Feasible yearly N leaching and profit for a Synthes ized north-Florida farm Figure 8-9. Optimization of a Synthesized north-Florida farm, scenario One. A) Current, optimized, and feasible practices and B) percent of N leaching and profit variation from current to feasible practices. Scenaould be resistant to change the CP in the diet. The DNFDFM was run maintaining the A B rio two. Another potential scenario simulated the case where the farmer w

PAGE 274

252 CP in. In the the high mode and holding all the other management choices as scenario Onethis scenario the major changes between the original as is farm and the proposed management strategy lies in the crop systems in the pastures and sprayfields. Crop systems were based on those proposed by the optimization runs, in the same way asprevious feasible practices that are stated in Table 8-7. 15,00016,000 10,00011,00012,00013,00014,000La NiaNeutralEl NioLBS N LEACHED65,40065,90066,40066,90067,40067,90068,40068,900US$ PROFIT PROFIT Figure 8-10. Second feasible yearly N leaching and profit for a north-Florida Synthesized farm Results of this simulation demonstrate that the farmer would still be able to decrease N lely and marginally increase the profitability of the farm, even g es, small changes can aching substantial though the farmer does not want to change the CP in diet. Expected N leachinvaried from 11,195 lbs year-1 in a La Nia year to 13,549 lbs year-1 in an El Nio year. N leaching was 12,723 lbs year-1 for a Neutral year. Compared with the original practices, the overall decrease in N leaching would be more than 11% and profit would increase slightly more than 2% fo r all ENSO phases (Figure 8-10). Although the best management practices reached with the optimizer and the feasibility of the farm worked similarly for the different ENSO phas

PAGE 275

253 be stir, raised on-farm, 00 lbs year-1 cow-1. The amount of CP was high. This is mostly a grazin N Nia years to 1,802 in Neutral years to 1,879 in El Nicrease y changing some manage ll implemented for better results according to different ENSO phases. Howevethese specific analyses should be performed for each farm on an individual basis, because site-specific interactions of the livestock, waste management systems, and the environmental parameters can have a great impact on results. 8.3.4 Examples of Applications of the DNFDFM to Specific Farms 8.3.4.1 Farm One This is a small farm with about 400 TAC, 10 bulls, 0% of heifers and a RHA of 15,0 g farm with only 37% CT. The estimated N lost from the waste system was 30.81%and, it has soil type 2. This farm ha s 2 sprayfields; one of 40 acres and the other of 80, but in both of them the crop rotations are the same: sorghum-sorghum-winter forage. As pasture, this farm has 120 acres of bahiagrass-winter forage rotation. The farmer mentioned that at the moment the net income for the farm was US$ 0.00 cwt-1 milk-1. That is, the farm was just breaking even. A simulation run for this farm with its current conditions indicated that estim at edleaching per year varied from 1,668 in La o years. Estimated net profits varied from US$ 5,868 to 5,848 to 5,839, respectively (Figure 8-11). Though a small amount of profit was indicated in the simulation, it satisfied the farmer that the model was very good. The solution presented by the optimizer estimated that N leaching could deapproximately 20% and profit could increase approximately 5% b ment practices. The optimizer suggested decreasing the CP level from high to low; maintaining confined time at 37%; changing the pasture field rotations to

PAGE 276

254 bermudagrass-winter forage; and introducing millet, bermudagrass, and corn in difields and seasons in the sprayfields. After seeing these results, the fa fferent rmer mentioned that probably not all the proposed change s could be implemented. The farmer said that it would be worthwhile to try the proposed changes in the pasture and sprayfields. So the simulation model was run againfor only these proposed feasible changes. The feasible practices showed that for this farm, N leaching could be decreased up to 9%, while profit could still be increased approximately 3% (Figure 8-11). Figure 8-11. Farm One: N leaching and profit for different management strategies s a medium-sized farm with 600 TAC, 22,400 lbs cow-1 of RHA, no bulls, 1 8.3.4.2 Farm Two Farm Two i 00% of heifers raised on-farm, high protein in the diet, and 100% CT. This farmhas a 35.31% estimated loss of N during waste handling, soil type 1, 155 acres of

PAGE 277

255 sprayfields with sorghum-millet-winter forage, and 300 acres of pasture with bahiagrass-bahiagrass-winter forage, and an estimated net income of US$ 0.80milk-1. cwt-1 For this second farm, the DNFDFM simulation estimated that N leaching would be 13,9d to 14, 14,954, and 15,466 lbs N year-1 for La Nia, Neutral, and El Nio years, respectively. Using the optimizer, it was determined that these rates could be decrease9,633, 10,375, and 10643, respectively (Figure 8-12). This would represent an overall decrease in N leaching of more than 30%. Regarding profit, the optimization found thatwith this combination of factors, profit would have a small increase of about 0.5%. Figure 8-12. Farm Two: N leaching and profit for different management strategies change Also, The optimization proposed changes in cropping patterns. One management was to substitute sorghum for corn in the spring-summer season. Another proposed adjustment was to substitute bahiagrass for bermudagrass in the pastures.

PAGE 278

256 the optimizer proposed adjustment in CP and CT. The farmer said that it would be possible to change the diet and some of the crops, but that confined time was fixed.A re-run of the model with these feasible parameters showed that N leachin g could b n, operation in the study area with the following characteristics: 3,000 s at this farm would release 92,008, 98,059 and 104,326e 89, y. patterns, CT, and CP levels in the diet. The farmer confirmed that this farm would be e reduced by changing crops and crude protein in this farm. The solution found that this farm could decrease N leaching up to 20% by changing only feasible practices.However, in this specific farm, the farmer would give up 2% of estimated profits by implementing these management changes. The farmer felt comfortable with that optiostating that a 2% decrease in net income would not be threatening to the farms survival; however, decreasing N lost by 20% would be a plus for sustainability. 8.3.4.3 Farm Three This is a large TAC, 20,000 lbs cow-1 of RHA, 80% CT, 120 bulls, 100% of heifers raised and a CP high. It has 450 acres of sprayfields with corn-sorghum-winter forages, and 800 acreof pasture with bahiagrass-bahiagrass-winter forage. The net income was estimated by the farmer as US$ 1.12 cwt-1 milk-1. Simulation results s h owed t h lbs of N year-1 for La Nia, Neutral, and El Nio years, respectively (Figur8-13). The optimizer f o und that N leaching could be decreased to 67,531, 73,771, and77,562 lbs of N year-1 for La Nia, Neutral, and El Nio years, respectively. This approximately 25% decrease in N leaching could be performed while still slightly increasing profit (<1%) from US$ 585,018 to 590,336, from US$ 580,270 to 585,3and from US$ 578,375 to 583,880, for La Nia, Neutral, and El Nio years, respectivelThe optimization found that these levels could be achieved by changing cropping

PAGE 279

257 willingld and capable to change to the suggested practices. Therefore, in this case, the optimized solution was equal to the feasible solution. This big dairy operation woube able to decrease its N pollution by 25% and still slightly increase profitability. Figure 8-13. Farm Three: N leaching and profit for different management strategies The proposed and feasible changes included millet and bermudagrass in the the summ N leaching and profit are significantly sensitive to parameters of all main fonsequently, the perfoh sprayfields in the spring season, and replace sorghum with corn and bermudagrass in er season; replace the bahiagrass with bermudagrass in the pasture; reduce the CTto 68%; and change the CP to low. 8.4 Conclusions and Recommendations Statistical analyses showed that actors and several combinations of them. C rmance of the DNFDFM is highly sensitive to specific dairy farm parameters. Estimations of N leaching and profitability of the DNFDFM change substantially wit

PAGE 280

258 different farms; therefore the analyses should be performed on an individual basis wspecific farms. Graphical representations of different farm management options regarding N leaching and pro ith fit indicated consistently that N exactly the opposite happens for El Nio years There are marked differences for soil types, with soil series of the types Millhopper, ies Penny less prone to leach N Decreasing the amount of crude protein in the diet impacts by decreasing the N Decreasing the confined time for a Synthesized farm decreases the N leaching and it Individual management changes in the Synthesized farm showed that there is higher north-Florida dairy farm demonstrated the power of the model in proposing ecologically and eions. For the 3 farms itrogen leaching is the lowest and profit the highest for La Nia ENSO phases; Bonneau, Blanton, Eunola more prone to leach N and have less profit, and soil serof the types Arredondo, Jonesville, Lake, Ortega Crop systems that include bermudagrass, corn, and bahiagrass leach less N and crop systems that include millet and sorghum tend to leach more N leaching and increasing profitability could also slightly decrease the profitability. sensitivity of the N leaching to changes in sprayfields and lower sensitivity to changes in pastures The optimization analysis using the DNFDFM with data of a synthesized conomically sound alternatives to specific dairy farms. For the case of a north-Florida Synthesized farm, the optimizer was able to combine factors to decrease up to 25% the N leaching and still increase profitability more than 3%. The applications of the DNFDFM in real farms, besides serving as a validation of the model, confirmed the usefulness of the model in practical applicat in which it was tested, the optimizer was able to find practices better than current practices that would have lower environmental impacts and still maintain profitability

PAGE 281

259 levels. By using the concept of feasible practices, which are those practices proposed by the optimizer that the farmer feels are feasible to implement, it was possible to reducthe N leaching by 9%, 20%, and 25% in a small, a medium, and a large farm, respectively, in north-Florida. These farms would reduce their N leaching while maintaining their profitability levels. The optimization analyses found that some practices would work across north-Florida dairy farm operations in e reducing N leaching. These are to change the crude proteiasture n to low standards and promote bermudagrass instead of bahiagrass in pand sprayfields. The other crop systems, the confined time, and other potential management choices would present different outputs to specific dairy farm conditions.

PAGE 282

CHAPTER 9 SUMMARY, CONCLUSIONS, AND RECOMMENDATIONS The Suwannee River Basin has received much attention in recent years and more so in recent months because increased N levels in water bodies. Dairy waste is an important factor contributing to this water N pollution, but dairies may reduce their total N loads by adjusting management strategies. Improvements in climate variation predictions (lead times of 6 to 12 months) can play an important role in devising management strategies that dairy farmers could adopt to pursue economic and ecological sustainability. The main aim of this dissertation was to test management options that decrease nutrient overflow in dairy systems under different climatic conditions while maintaining profitability through the creation and use of simulation and optimization models. Whole-dairy-farm system simulation was used to test alternative management strategies that would not be practical to test under regular experimental designs. Dynamic, event-controlled, process-based, and empirical models were coupled with optimization models to represent north-Florida dairy farms. User friendliness of models and high interaction with stakeholders were always pursued. The Dynamic North-Florida Dairy Farm Model (DNFDFM) was intended to be an environmental accountability tool for dairy farmers and regulatory agencies capable of simulating and optimizing any individual north-Florida dairy farm. Economic and ecological impacts were studied along with the effects of seasonal or climate variability for different dairy farm types. General results indicated that north-Florida dairy farms could substantially decrease N leaching and maintain profitability levels by adjusting crop sequences in sprayfields, 260

PAGE 283

261 crop rotations in pastures, crude protein in the diet, and confined time of milking cows. Climatic conditions greatly affect N leaching: during El Nio years, when higher rainfall and colder temperatures are experienced, higher N leaching is estimated than in La Nia years, when lower rainfall and hotter temperatures are observed; in Neutral years, medium N leaching is registered. Farmers could vary crop proportions to better prepare for these different climatic forecasts. In Chapter 2, an intense literature review was used to explore the hypotheses that An N pollution problem exists in the Suwannee River Basin Dairy farms are major contributors to this problem Results clearly indicated that not only does the problem of N pollution exist, but also there is an increasing trend of N concentrations in ground and surface waters. Although the N levels are currently below the regulatory standards (<10 mg L -1 ), they are always above the background, and so measures of mitigation are needed to avoid uncontrollable situations in the near future. Dairy farms are indeed major contributors to the problem through manure applications that exceed the carrying capacity of the soils and crops. However, substantial differences exist among dairy farms because of temporal, spatial, and climatic conditions. It is highly recommended, therefore, to assess point sources of pollution and keep individual dairy farms environmentally accountable. Components inside dairy farms and their interactions also vary greatly in their environmental impacts and these need to be taken into consideration when analyzing complete dairy farm systems. In Chapter 3 a Sondeo, a Survey, and several Focus Groups were used as instruments to capture the Suwannee River Basin stakeholders perceptions regarding Water contamination Current environmental regulations Potential adoption of climate forecast innovations

PAGE 284

262 It was found that most of the stakeholders perceived that there is an environmental problem of water pollution that needs to be addressed. However, some farmers, technicians, and extensionists challenged this fact and showed skepticism of the gravity of the problem. Major efforts are needed to clearly communicate to all groups of stakeholders, especially to the skeptical ones about the real problem and its potential consequences. Most of the farmers agree with current environmental regulations and many of them are very thankful to have the opportunity to participate in incentive-based programs. Nonetheless, most of them also believe that regulations will be the major challenge they will have to face in the future. Some farmers are in disagreement with current regulations because of technical and/or administrative aspects. Therefore, regulatory agencies should work collaboratively with all stakeholders groups to test the economic and ecologic feasibility of promoted best management practices, BMPs. Two-thirds of dairy farmers would use climate forecast products in crop decisions for selecting forage crops, changing planting dates, and/or planning areas of forage; some of them would additionally benefit from this climate forecast use in livestock, feedstock, and waste management. The main barriers to climate forecast adoption is the lack of flexibility of north-Florida dairy farms and the management resources required for their implementation. Consequently, forecast innovations must fit narrow niches of application and require realistic management efforts in order to be adopted. Further research calls for the inclusion of climate variability forecasts into BMP experiments. Commitment in extension and outreach programs to deliver climate forecasts together with their regular products is also required to diffuse these climate innovation products.

PAGE 285

263 Chapter 4 describes the collaborative process of building the Dynamic North-Florida Dairy Farm Model (DNFDFM) with the stakeholders. Multiple feedback loops were used during a two-year iterative process that included initial developments, creation of a prototype model, and calibration and validation of a final model that simulates monthly N leaching and cash flows on north-Florida dairy farm systems affected by a climatic component determined by the El Nio Southern Oscillation (ENSO) phenomenon. Participatory modeling can enhance the creation of adoptable and adaptable user-friendly models that include environmental, economic and biophysical components. These component models were embedded in the DNFDFM. By providing farmers, policy makers, and other stakeholders with a more holistic view of current practices, common ground among them was more easily identifiable and collaboration was therefore fostered. By having stakeholders participate in the creation and development of the models, there is increased chance that the final product will better represent the reality and interests of stakeholders, and consequently a greater chance that the model will be adopted by interested users. Chapter 5 tested the hypothesis that there could be better ways to estimate N excretion from dairy cows than those commonly used. This chapter described and analyzed the creation of a new model, the north-Florida livestock model, which is the driver of the DNFDFM. This model introduced dynamic components, parameterization of north-Florida herd environmental and management conditions, minimum requirement of initial data, and calibration with farmers participation. Results of this model regarding N received by land (lbs acre -1 ) were compared with those yielded by the two most commonly used approaches in Florida. This comparison suggested that previous models have either overestimated or

PAGE 286

264 underestimated N excretion and did not account for seasonality. This new model is believed to be more accurate because it captures more realistic and relevant characteristics of north-Florida dairy operations. Results showed great variability in manure N production due principally to season of the year, number of cows, milk production levels, and size of farm. Values indicated that some fields may receive as little as 3 lbs N acre -1 month -1 and other fields may receive as much as 145 lbs N acre -1 month -1 Chapter 6 assessed the potential N leaching of north-Florida dairy forage systems under the premise that these crop systems are the main means of recycling the always abundant manure N produced as a consequence of milk production. Findings were summarized for incorporation into the DNFDFM. Eight focus groups with stakeholders and 21 interviews with dairy farmers were performed in order to collect information about the forage systems and their parameters; additional information was collected from published sources, farm records, and official regulatory agencies. It was found that the common forage systems in north-Florida dairy farms include a spring/summer forage (corn, sorghum, millet, bermudagrass, or bahiagrass), a summer/fall forage (corn, sorghum, millet, bermudagrass, or bahiagrass), and a fall/winter forage (rye, oats, wheat, or ryegrass). Crop models contained in the Decision Support System for Agricultural Transfer, DSSAT v4.0, were used to predict monthly N leaching and biomass accumulation in these forage crop systems under different parameters of Crop sequences Manure N received Seasonal climatic conditions Soil characteristics Other variable management specifications

PAGE 287

265 After creating, calibrating, and validating the common forage systems for north-Florida, these crop models were run for 43 years of daily weather data (1956-1998). Results indicated that higher N leaching and lower biomass accumulation are predicted for El Nio followed by Neutral years; La Nia years showed the lowest N leaching and the highest biomass accumulation. Winter in general, and January and February, specifically, are the critical months with the highest N leaching. The best forage systems to prevent N leaching are those that start in spring-summer with bermudagrass or corn; have bermudagrass, bahiagrass or corn in summer; and finish with winter forages. Systems that leach the most N are those that include millet and/or sorghum. Water holding capacity and pH among soil types may explain most of the great variability found in N leaching. Chapter 7 describes the DNFDFM, a simple, event-controlled, mechanistic, dynamic model that couples livestock, crop and optimization models created to assess N leaching from entire north-Florida dairy farm systems and the economic impacts resulting from reducing it. The model responds to dairy-specific environmental (climate and soils) and specific characteristics of Herd management Waste handling system Crops in sprayfields Crops in pastures Economic/financial situation The DNFDFM is used to simulate an existing system, optimize management options to diagnose sources of problems and suggest improvements in the system, and then simulate the modified system to compare it with the existing system. The model could also be used for pre-testing BMPs for specific farms and also for designing new dairy farms.

PAGE 288

266 The driver of the livestock and crop portions of the DNFDFM is a dynamic adaptation of the framework balance and/or budgeting of nutrients in dairy farms, commonly used in Florida. The DNFDFM incorporates Markov-chain probabilistic simulation of cow-flows and crop simulation for historical climatic years (El Nio Southern Oscillation). It also includes automated optimization of managerial options, participatory modelling, and user-friendliness. This chapter discussed the model structure, its components, its functioning, and its limitations. The DNFDFM should be used on an individual farm basis with the specific environmental and managerial conditions of each dairy farm. As an indication, outputs of running the model for some representative, but synthesized north-Florida dairy farms indicated that a small dairy farm would leach between 34 and 38 lbs N year -1 cow -1 ; while a medium farm would leach between 31 and 35 lbs N year -1 cow -1 ; and a large operation would leach between 25 and 29 lbs N year -1 cow -1 for La Nia and El Nio years, respectively. Chapter 8 presents and discusses results of simulating and optimizing synthesized and real dairy farms in north-Florida using the DNFDFM. The chapter included statistical, graphical, optimization, and application analyses. Results indicated that there is significant and great variability among soil types, i.e. soil type 6 (Millhopper-Bonneau) leaches the most N and soil type 2 (Arredondo-Jonesville-Lake) leaches the least N. A crop system that included bermudagrass-bermudagrass had the least N leaching and those that had sorghum-millet leached the most N. Higher crude protein in the diet or 20% higher confined time of milking cows created significantly higher amounts of N leaching.

PAGE 289

267 The optimization analysis using the DNFDFM with data from a Synthesized north-Florida dairy farm demonstrated the capability of the model in proposing ecologically and economically sound alternatives for specific dairy farms. The optimizer was able to combine factors to decrease N leaching up to 25% and still increase profitability by 3%. The applications of the DNFDFM in real farms, besides serving as a validation of the model, confirmed its usefulness in practical and real applications. For the 3 farms, in which it was validated, the optimizer was able to find better than current practices that would have lowered environmental impacts and still maintained profitability levels. By using the concept of feasible practices, which are those proposed by the optimizer that the farmer feels can be implemented, it was possible to reduce by 9%, 20%, and 25% the N leaching in a small, a medium, and a large farm in north-Florida. The optimization analyses found that some practices would be especially useful to north-Florida dairy farm operations in reducing N leaching. These are to decrease the crude protein in the diet and promote bermudagrass instead of bahiagrass in pastures and sprayfields. The other crop systems, the confined time and other potential management choices would present different outputs to specific dairy farm conditions. The DNFDFM can help farmers and other major stakeholders in making complicated decisions in optimizing both profitability and sustainability. Some practical, real-life envisioned applications are listed below. The DNFDFM has been conceived and developed with farmers, consequently it is expected that most of the benefits of its direct use return to these dairy farmers:

PAGE 290

268 The DNFDFM can be an important tool devising management strategies dairy farms could follow according to predicted long-term climate conditions The DNFDFM could help farmers maintain or gain compliance while maintaining profitability levels It can help in the design and testing of best management practices (BMPs) at the farm level. The virtual testing of these practices would decrease the risk of failure, save money, and decrease the time of implementation and re-design The DNFDFM could be used as a tool in the decision of design and implement a new dairy operation. Similarly, the model could be helpful in the decision of locating new or moving farms The model could be especially important in the reporting process dairy farms need to do to the official regulatory agencies. Dairy farmers could use it as a tool of individual environmental accountability The DNFDFM would allow farmers to demonstrate their stewardship towards environmentally sound practices and account for individual point-source pollution The model could also be used as an administrative tool, inside the dairy farm, for record keeping and/or planning administrative activities, since it computes economic outputs, along with all major activities The utility of the DNFDFM would be diminished if regulatory agencies would not accept and/or not use the model. It is envisioned that the regulatory agencies would find relevant uses for it: The DNFDFM could be used by the regulatory agencies to assess all the points expressed above, and additionally The model used on several farms could help regulatory agencies devise regional environmental practices according to forecasted climate conditions Regulatory agencies would be able to assess the economic impacts of proposed BMPs and the need of having financial help for farmers in order to comply The model could be important as an accountability tool for the regulatory agencies to demonstrate to activist groups and the general public the effectiveness of the regulations and the proposed practices

PAGE 291

269 It is also believed that activist groups (i.e., environmental groups) and the public in general would appreciate the model and its use: These groups could use it for most of the above uses, and additionally These groups may be interested in assessing potential regional impacts of the dairy industry according to climate conditions It would give them a more holistic view of the dairy farm enterprise and its components, and realistic options towards decreasing environmental impacts It is also envisioned that the model would serve the scientific community in some meaningful ways: It may help scientists assess trends/relationships between N pollution in water and estimated dairy farm releases by the DNFDFM Scientists could follow parts of the methods/frameworks used in the conception, design, implementation, and use of the DNFDFM to be applied to other studies The DNFDFM could still further be improved in order to better serve its purposes. Modifications, without losing user-friendliness that would improve its functionality are listed below. The livestock and waste components would be improved by the inclusion of farm-specific culling/reproduction rates Specific details inside the livestock component could still be adjusted for major accuracy. Some examples are Include an exponential function of calves/heifers gain of weight by using exponential parameters instead of the current constant rate. This incorporation would only have very slight changes in the pasture component and negligible overall impacts in the farm The general dietary protein function included following the NRC standards could be broken down to include information about the use of total crude protein, rumen protected amino acids, non-protein N, and ruminally degradable and undegraded protein. This incorporation would require parameterize experimental data and a higher level of inputs for running the model

PAGE 292

270 The DNFDFM should be capable of handling an increase/decrease in herd size and/or field size inside a run in order to simulate the desire or the real conditions of expansion/reduction of a farm and perform analyses under those conditions The model would benefit from higher customization of the crop model results. It would imply increasing the number of factors used to run the crop models and implement the results in the model. Especially important would be to: Include daily weather information for other meteorological stations in the study area. Include direct manure deposition in pastures Seasonality of crops according to climate conditions Seasonality of crop for harvesting Specific winter forages The model would improve with the inclusion of more detailed economic factors such as fixed costs and the initial costs as well as the value of the infrastructure and their depreciation and the specific costs of compliance The DNFDFM would benefit by including stochastic results according to climate ENSO phases. It would imply reassembling the distribution of crop model results for different climate years into a probabilistic function The optimization process could be faster if adjustment variables would be kept from one simulation to the next one Optimization could be improved if more decisions could be chosen. Also optimization could be improved if non-linear optimization methods and multi-objective approaches are used.

PAGE 293

LIST OF REFERENCES Adams, A.L. 1998. Dietary fat effects on rumen fermentation, milk production, and reproduction of dairy cattle, and economic implications for dairy production. Ph.D. thesis, University of Florida. Gainesville, FL. Albert, M. 2002. Monitoring and modeling the fate and transport of nitrate in the vadose zone beneath a Suwannee River Basin vegetable farm. M.S. thesis, University of Florida. Gainesville, FL. Alvira, D., Cabrera, V.E., Galindo, S., Haile, S., Mtenga, K., Rios, F., Tripathi, N., Villagomez, J.A. 2003. Nutrient management in the Suwannee River Basin: perceptions of stakeholders. University of Florida, Food and Resource Economics Department, Staff Paper Series, SP 04-2. Gainesville, FL. American Society of Agricultural Engineers. 1990. Manure production characteristics. Engineering practices subcommittee, ASAE Agric. Sanit. Waste Manag. Comm. ASAE Standard D384.1, Am. Soc. Agric. Eng., St. Joseph, MI. Andrew, W.J. 1994. Nitrate in ground water and spring water near four dairy farms in north Florida 1990-1993. US Geological Survey, Water-Resources Investigations Report 94-4162, 63 p. Tallahassee, FL. Andrew, W.J. 1992. Reconnaissance of water quality at nine dairy farms in north Florida, 1990-1991. US Geological Survey, Water-Resources Investigations Report 92-4058, 39 p. Tallahassee, FL. Archer, D., Eklund, J., Walsh, M., Forcella, F. 2002. WEEDEM: A user-friendly software package for predicting annual ryegrass and wild radish emergence. In: Spafford, J.H., Dodd, J., Moore, J.H. (eds.), WEEDS: "threats now & forever? p. 252-253. In Proc. 13 th Australian Weeds Conf., Perth, WA, September 8-13, 2002. Department of Primary Industries, Water and Environment, Government of Tasmania. Hobart, Tasmania, Australia. Asbury, C.E., Oaksford, E.T.1997. A comparison of drainage basin inputs with in stream nutrient loads for seven rivers in Georgia and Florida, 1986-90 US Geological Survey Water Resources-Investigations Report 97-4006, 8 p. Tallahassee, FL. Bastidas, E. 2001. Assessing potential response to changes in the livelihood system of diverse, limited-resource farm households in Carchi, Ecuador: modeling livelihood strategies using participatory methods and linear programming. Ph.D. thesis, University of Florida. Gainesville, FL. 271

PAGE 294

272 Bernard, H.R., 1994. Research Methods in Anthropology: Qualitative and quantitative approaches. Sage. Newbury Park, CA. Berndt, M.P., Hatzell, H.H., Crabdall, C.A., Turtora, M., Pitman, J.R., Oaksbord. 1998. Water quality in the Georgia-Florida coastal plain, Georgia and Florida, 1992-1996. Circ. 1151. U.S. Geological Survey, Reston, VA. Bisoondat, M., Woodard, K.R., Sollenberger, L.E., French, E.C., Portier, K.M., Graetz, D.A., Prine, G.M., Van Horn, H.H. 2002. Dairy effluent effects on herbage yield and nutritive value of forage cropping systems. Agron. J. 94, 043-1049. Boote, K.J., Jones, J.W., Hoogenboom, G. 1998. Simulation of crop growth: CROPGRO model. In Peart R.M., Curry R.B. (eds.) Agricultural systems modeling and simulation. Marcel Dekker, New York. Brsting, C.F., Kristensen, T., Misciattelli, L., Hvelplund, T., Weisbjerg, M.R. 2003. Reducing N surplus from dairy farms. Effects of feeding and management. Livest. Prod. Sci. 83,165-178 Center for Ocean-Atmospheric Prediction Studies. 2003. ENSO index according to JMA SSTA (1868-present). COAPS-Florida State University. Accessed January, 2004 ftp://www.coaps.fsu.edu/pub/JMA_SST_Index/ Chambers, R. 1994. The origins and practice of participatory rural appraisal. World Dev. 22, 953-959. Chambliss, C.G. (ed.). 1997. Florida forage handbook. Florida Coop. Extension Service, IFAS, Electronic Data Information Source, University of Florida. Gainesville, FL. Available at: http://edis.ifas.ufl.edu/pdffiles/AG/AG17000.pdf Conway, G. R. 1990. Agroecosystems. In: Jones J.G.W., Street, P.R. (eds.). Systems theory applied to agriculture and the food chain. Elsevier Applied Science, New York. Copeland, R., David, J., Hansard, P. 1999. Nitrate source impacts on private drinking well waters in northwest Lafayette County, Florida. p. 1-6, Ambient Monitoring Section, FL Department of Environmental Protection. Ambient Newsletter 3, 3, August, 1999. Available at: http://www.dep.state.fl.us/water/monitoring/docs/news/news9908.pdf. Dent, B.J., McGregor, M.J., 1994. Rural and farming systems analysis. A European perspective. Oxford Univ. Press, Oxford, UK. Department of Environmental Protection (DEP). 2001. Florida water conservation initiative. Agricultural working group. A report of the Florida Department of Environmental Protection. Orlando, FL. De Vries, A. 2004. Economic value of late replacement when cow performance is seasonal. J. Dairy Sci. In Press.

PAGE 295

273 De Vries, A. 2001. Statistical process control charts applied to dairy herd reproduction. Ph.D. thesis. University of Minnesota, Saint Paul, MN. De Vries, A., Giesy, R., Ely, L., Webb, D., Andreasen, A., Broadus, B., Miller, P., Seabright, T, Vann, C., de Araujo, A. 2000. Dairy business analysis project: 2000 financial summary. University of Florida, Institute of Food and Agricultural Sciences, AN 135. Gainesville, FL. Edwards, T.M., Barbeau, T.R., Guillette, L.J. Jr. 2003. Effects of nitrate contamination on reproduction growth, and development of animals associated with Floridas Springs. In Proc. Florida Springs Conference. Gainesville, FL. Feb. 5-7, 2003. Florida Center for Library Automation, Gainesville, FL. Environmental Literacy Council. 2002. Non-point source pollution. The Environmental Literacy Council, Washington DC. Accessed December, 2003 http://www.enviroliteracy.org/article.php/413.html. Fan, A.M., Wilhite, C.C., Book, S.A. 1987. Evaluation of the nitrate drinking water standard with reference to infant methemoglobinemia and potential reproductive toxicity. Regulatory Toxicology and Pharmacology 7, 135-137. Felsot. A. 1998. Re-examining the Link between Nitrates and "Blue-Baby" Syndrome: A Necessary First Step For Managing Ground Water Quality to Protect Public Health. Cooperative Extension, Washing State University. Accessed April, 2004, http://www.aenews.wsu.edu/Oct98AENews/aenewsoctober98.htm#anchor545063. Florida Agricultural Statistics Service. 2003. Livestock, dairy and poultry page. U.S. Department of Agriculture. Accessed March, 2004, http://www.nass.usda.gov/fl Florida Department of Environmental Protection (FLDEP). 2004. EPA 319 Grant Program. Accessed, May, 2004 http://www.dep.state.fl.us/water/nonpoint/319h.htm. Florida Department of Environmental Protection (FLDEP). 2001. Basin status report: Suwannee. Division of Water Resource Management. Bureau of Water Management, Tallahassee, FL. Available at: http://www.dep.state.fl.us/water/basin411/suwannee/ status/SUWANEE.pdf Florida Geographic Data Library. 1995. Data download. Accessed October, 2002 ftp://ftp1.fgdl.org/pub/county/ Florida Geological Survey, Department of Natural Resources. 1992. Floridas ground water quality monitoring program background hydrogeochemistry. Florida Geological Survey Special Publication No 34. Tallahassee, FL. Fontaneli, R.S., Sollenberger, L.E., Staples, C.R. 2001. Yield, yield distribution and nutritive value of intensively managed warm-season annual grasses. Agron. J. 93, 1257-1262.

PAGE 296

274 Fontaneli, R.S., Sollenberger, L.E., Staples, C.R. 2000. Seeding date effects on yield and nutritive value of cool-season annual forage mixtures. Soil Sci. Soc. Florida Proc. 59, 60-67. Forman, D., Al-Dabbagh, S., Doll, R. 1985. Nitrates, nitrites, and gastric cancer in Great Britain. Nature 313, 620-625. Fraisse, C.W., Campbell, K.L., Jones, J.W., Boggess, W.G. 1996. GIDM: A GIS-based model for dairy waste management analysis. In Proc. Symposium on GIS and Water Resources, Ft. Lauderdale, Sep 22-26, 1996. American Water Resources Association, Middleburg, VA. Geurtz, J.L.A., Joldersma, F. 2001. Methodology for participatory policy analysis. European Journal of Operational Research 128, 300-310. Geurtz, J.L.A., Vennix, J. 1989. Verkenningen in Beleidsanalyse: Theorie en Praktijk van Modelbouw en Simulatie. Kercke-bosch, Zeist. Giesy, R., de Vries, A., Bray, D. 2002. Florida dairy farm situation 2002. Cooperative Extension Service, IFAS, University of Florida. Gainesville, FL. Gijsman, A.J., Hoogenbomm, G., Parton, W.J., Kerridge, P.C. 2002. Modifying DSSAT crop models for low-input agricultural systems using soil organic matter-residue module from CENTURY. Agron. J. 94, 462-474. Ham, L.K., Hazell, H.H. 1996. Analysis of nutrients in the surface waters of the Georgia-Florida coastal plain study unit, 1970-1991. US Geological Survey, Water-Resources Investigations Report 96-4037, 67 p. Tallahassee, FL. Hammer, G.L., Nicholls, N., Mitchell, C. 2000. Applications of Seasonal Climate Forecasting in Agricultural and Natural Ecosystems: The Australian Experience. In: Hammer G.L., Nicholls, N. and Mitchell, C. (eds.). Applications of seasonal climate forecasting in agricultural and natural ecosystems the Australian Experience, Kluwer Academic, The Netherlands. Hansen, J.W. 2002a. Applying seasonal climate prediction to agricultural production. Agr. Syst. 74, 305-307. Hansen, J. W. 2002b. Realizing the potential benefits of climate prediction to agriculture: issues, approaches, challenges. Agr. Syst. 74, 309-330. Hazell P.B.R., Norton, D. 1986. Mathematical programming for economic analysis in agriculture. Macmillan, New York. Herrero, M., Fawcett, R.H., Silviera, V., Busqu J., Bernus, A., Dent, J.B. 2000. Modelling the growth and utilisation of kikuyo grass (Pennisetum clandestinum) under grazing. 1. Model definition and parametrisation. Agr. Syst. 65, 73-97.

PAGE 297

275 Hildebrand, P.E. 1990. Farming systems research-extension. p. 169-180. In: Jones J.G.W., Street, P.R. (eds.). Systems theory applied to agriculture and the food chain. Elsevier Applied Science, New York. Hildebrand, P.E. 1981. Combining disciplines in rapid appraisal: The Sondeo approach. Agri. Admin. 8, 423-432. Hornsby, D., Ceryak, J., Zwanka, W. 2003. Groundwater quality report. Department of Water Resources, Suwannee River Water Management District. Live Oak, FL. Hornsby, D., Ceryak, J., Zwanka, W. 2002b. Groundwater quality report. Department of Water Resources, Suwannee River Water Management District. Live Oak, FL. Hornsby, D., Ceryak, R., 1999, Springs of the Suwannee River Basin in Florida: Suwannee River Water Management District Report WR99-02. Live Oak, FL. Hornsby, D., Mattson, R., Mirti, T. 2002a. Surface water quality and biological annual report. Department of Water Resources, Suwannee River Water Management District. Live Oak, FL. Houston, T. B., Hazen, M.W., Mathews, T.C., Brown, G.A. 1965. Soil Survey of Suwannee County, Florida. USDA Natural Resour. Conserv. Serv., Washington, DC. Hubbard, R.K., Thomas, D.L., Leonard R.A., Butler, J.L. 1987. Surface runoff and shallow ground water quality as affected by center pivot applied dairy cattle waste. Trans. ASAE 30, 430-437. Jagtap, S.S., Jones, J.W., Hildebrand, P., Letson, D., O'Brien, J. J., Podest, G., Zierden, D., Zazueta, F. 2002. Responding to stakeholders demands for climate information: from research to applications in Florida. Agr. Syst. 74, 415-430. Johnson, J.C., Newton, G.L., Butler, J.L. 1991. Recycling liquid dairy cattle waste to sustain annual triple crop production of forages. p. 41-50. In Proc. 28 th Annual Florida Dairy Production Conf., Gainesville, FL. 9-10 Apr. 1991. Dairy Sci. Dep., CES, IFAS, Gainesville, FL. Jones J.W., Hansen, J.W., Royce, F.S., Messina, C.D. 2000. Potential benefits of climate forecasting to agriculture. Agr. Ecosyst. Environ. 82, 169-184. Jones, J.W., Hoogenboom, G., Porter, C.H., Boote, K.J., Batchelor, W.D., Hunt. L.A., Wilkens, P.W., Singh, U., Gijsman, A.J., Ritchie, J.T. 2003. The DSSAT cropping system model. Eur. J. Agron. 18, 235-265. Jones J.W., Thornton, P.K., Hansen, J.W. 1997. Opportunities for systems approaches at the farm scale. p1-18. In Teng et al., (ed.) Applications for systems approaches at the farm and regional levels. Syst. Approaches for Sustainable Agric. Dev. 5. Kluwer Acad. Publ., Dordrecht, The Netherlands.

PAGE 298

276 Jones, J.W., Tsuji, G.Y., Hoogenboom, G., Hunt, L.A., Thornton, P.K., Wilkens, P.W., Imamura, D.T., Bowen, W.T., Singh, U. 1998. Decision support system for Agrotechnology transfer: DSSAT v3. p. 155-177. In Tsuji, G.Y., Hoogenboom, G., Thornton, P.K. (eds.). Understanding options for agricultural production. Kluwer Acad. Publ., Dordrecht, The Netherlands. Jonker, J.S., Kohn, R.A., High, J. 2002. Dairy herd management practices that impact N utilization efficiency. J. Dairy Sci. 85, 1218 Katz, B. 2000. Sources of nitrate contamination of spring waters, Suwannee River basin, Florida. In Proc. Florida Springs Conference, Gainesville, FL. Feb. 9-10, 2000. Florida Center for Library Automation, Gainesville, FL. Katz, B., Hornsby, D., Bohlke, F., Mokray, M. 1999. Sources and chronology of nitrate contamination in spring waters, Suwannee River Basin, Florida. US Geological Survey, Water-Resources Investigations Report 99-4252, 54 p. Tallahassee, FL. Katz, B.G., Bohlke, J.K. 2000. Monthly variability and possible sources of nitrate in ground water beneath mixed agricultural land use, Suwannee and Lafayette counties, Florida. US Geological Survey Water-Resources Investigation Report 00-4219, 28 p. Tallahassee, FL. Katz, B.G., DeHan, R.S. 1996. The Suwannee River basin pilot study: Issues for watershed management in Florida. Fact Sheet FS-080-96. U.S. Geological Survey. Reston, VA. Keating, B.A., McCown, R.L. 2001. Advances in farming systems analyses and intervention. Agr. Syst. 70, 555. Kelly, T.C. 1995. A bioeconomic systems approach to sustainability analysis at the farm level. Ph.D. thesis, University of Florida. Gainesville, FL. Knowlton, K.F. 2000. Are Jerseys part of the answer? The Jersey Journal 47, 81. Kross, B.C., Ayebo, A., Fuortes, L.J. 1992. Am. Fam. Physician 46, 183-188. Lanyon, L.E. 1994. Dairy manure and plant nutrient management issues affecting water quality and the dairy industry. J. Dairy Sci.77, 1999-2007. Leonard, R.A., Knisel, W.G., Still, D.A. 1987. GLEAMS: Groundwater Loading Effects of Agricultural Management Systems. Trans., Amer. Soc. of Agric. Engrs. 30,1403-1418. Levin, S.A. 1991. The mathematics of complex systems. p. 215-226. In: Mooney, H.A., Medina, E., Schindler, D.W., Schulze, E.D., Walker, B.H. (eds.). Ecosystem experiments, scope 45. John Wiley & Sons, West Sussex, England. Litell, R.C., Milliken, G.A., Strouop, W.W., Wolfinger, R.D. 1996. SAS system for mixed models. SAS Inst., Cary, NC.

PAGE 299

277 Mavromatis, T., Jagtap, S.S., Jones, J.W. 2002. El Nio-Southern Oscillation effects on peanut yield and N leaching. Climate Res. 22, 129-140. Mayer, D.G., Belward J.A., Burrage, K. 1996. Use of advanced techniques to optimize a multi-dimensional dairy model. Agr. Syst. 50, 239-253. Meinke, H. Baethgen, W.E., Carberry, P.S., Donatelli, M, Hammer, G.L., Selvaraju, R., Stockle, C. 2001. Increasing profits and reducing risks in crop production using participative systems simulation approaches. Agr. Syst. 70, 493. Meinke, H., Hochman, Z. 2000. Using seasonal climate forecasts to manage dryland crops in northern Australia experiences from the 1997/98 seasons. pp. 149. In: Hammer G.L., Nicholls, N. and Mitchell, C. (eds.) Applications of seasonal climate forecasting in agricultural and natural ecosystems The Australian Experience, Kluwer Academic, The Netherlands. Mirvish, S.S. 1990. The significance of human health of nitrate, nitrite, and N-nitroso compounds. In Nitrate Contamination: Exposure, Consequence, and Control. NATO ASI Series, Vol. G30. Bogardi et al. (Eds.). Springer Verlag, Berlin. Mourits, M.C.M., Huirne, R. B.M., Dijkhuizen, A.A., Kristensen, A.R., Galligan, D.T. 1999. Economic optimization of dairy heifer management decisions. Agr. Syst. 61, 17-31. National Research Council (NRC). 2001. Nutrient Requirement of dairy cattle (7 th Ed.). National Academy Press, Washington, DC. National Research Council (NRC). 1995. Nitrate and nitrite in drinking water. National Academy Press, Washington, DC. National Research Council (NRC). 1985. The health effects of nitrate, nitrite, and N-nitroso compounds: National Academy of Sciences, Washington, DC. National Research Council (NRC). 1978. The health effects of nitrate, nitrite, and N-nitroso compounds: National Academy of Sciences, Washington, DC. Natural Resource Conservation Service (NRCS). 2002. Soil survey geographic (SSURGO) database. Accessed December, 2002 http://www.ncgc.nrcs.usda.gov/branch/ssb/ products/ssurgo/ Natural Resource Conservation Service (NRCS). 2001. Water budget and nutrient balance worksheet, WATNUTFL Vers. 2.0. Part 650, Engineering Field Handbook. 210-VI-EFH. Gainesville, FL. Nelson, R.A., Holzworth, D. P., Hammer, G. L., Haymand, P. T. 2002. Infusing the use of seasonal climate forecasting into crop management practice in North East Australia using discussion support software. Agricultural Systems 74, 393-414.

PAGE 300

278 Nennich, T., Harrison, J.H., Meyer, D., Weiss, W.P., Heinrichs, A.J., Kincaid, R.L., Powers, W.J., Koelsch, R.K., Wright, P.E. 2003. Development of standard methods of estimate manure production and nutrient characteristics from dairy cattle. ASAE 701P1203, 263-268. Newton, G.L., Johnson, J.C., Davis, J.G., Vellidis, G., Hubbard, R.K., Lowrance, R. 1995. Nutrient recoveries from varied year round applications of liquid dairy manure on sprayfields. p. 113-122. In Proc. 32 nd Annual Florida Dairy Production Conf., Gainesville, FL. April, 1995. Dairy Sci. Dep., CES, IFAS. Gainesville, FL. Norman, D., Collison, M. 1985. Farming systems research in theory and practice. pp16-30. In: Remenyi J.V. (ed.) Agricultural systems research for developing countries. Australian Centre for International Agricultural Research Proceedings No. 11, ACIAR, Canberra, Australia. OBrien, J.J., Zierden, D.F., Legler, D., Hansen, J.W., Jones, J.W., Smajstrla, A.G., Podest, G., Letson, D. 1999. El Nio, La Nia and Floridas Climate: Effects on Agriculture and Forestry. The Florida Consortium: The Florida State Univ., the Univ. of Florida, University of Miami. Tallahassee, FL. Ondersteijn C.J.M., Beldmanb, A.C.G., Daatselaar, C.H.G., Giesen, G.W.J., Huirne, R.B.M. 2002. The Dutch mineral accounting system and the European nitrate directive: implications for N and P management and farm performance. Agr. Ecosyst. Environ. 92, 283-296. Onema, J., Koskamp, G.J., Galama, P.J. 2001. Guiding commercial pilot farms to bridge the gap between experimental and commercial dairy farms: the project cows and opportunities. Neth. J. Agr. Sci. 49(2-3), 277-296. Paine, M., Weatherly, J., Crawford, A., Champion, S. 2002. Farmlets as learning platforms: a national approach to farming systems research, development and extension in the Australian dairy sector. p. 729-738. In Proc. Fifth IFSA European Symposium, Florence, Italy 8-11 April 2002. International Farming Systems Association, European Group, Dip. di Economia Agraria e delle Risorce Territoriali, Facolta di Agraria, Universita degli Studi di Firenze, Florence, Italy. Parton, W.J., Schimel, D.S., Cole, C.V., Ojima, D.S., 1979. Analysis of factors controlling soil organic matter levels in Great Plains grasslands. Soil Sci. Soc. Am. J. 51, 1173-1179. Phillips, J.G., Cane, M.A., Rosenzweig, C. 1998. ENSO seasonal rainfall patterns and simulated maize yield variability in Zimbabwe. Agric. For. Meteorol. 90, 39-50. Pittman, J.R., Hatzell, H.H., Oaksford, E.T. 1997. Spring contributions to water quantity and nitrate loads in the Suwannee River during base flow in July 1995. U.S. Geological Survey Water-Resources Investigations Report 97-4152. 12p. Tallahassee, FL. Powers, L.E., McSorley, R. 2000. Ecological principles of agriculture. Delmar, Albany, New York.

PAGE 301

279 Ribaudo, M., Gollehon, N., Aillery, M., Kaplan, J., Johansson, R., Agapoff, J., Christensen, L., Breneman, V., Peters, M. 2003. Manure management for water quality costs to animal feeding operations of applying manure nutrients to land. Agricultural Economic Report 824. Resource Economics Division. Economic Research Service. U.S. Department of Agriculture. Washington, DC. Robertson, M.J., Carberry, P.S., Lucy, M. 2000. Evaluation of a new cropping option using a participatory approach with on-farm monitoring and simulation: a case study of spring sown mungbeans. Aust. J. Agr. Res. 51, 1. Rosenhead, J. (Ed.), 1989. Rational analysis for a problematic world: problem structuring methods for complexity, uncertainty and conflict. Wiley, Chichester, UK. Rymph, S., Boote, K.J. 2004. Modifying the Code of CropGro to Predict Growth of a Perennial Tropical Forage Grass. p. 66-67. In Proc. Biological Systems Simulation Conference, Gainesville, FL. 8-10 March 2004. University of Florida, IFAS, Gainesville, FL. St-Pierre, N.R., Thraen, C.S., Jones, L.R. 1999. Forecasting herd structure, milk production and their variances for production risk management. Extension Research. Special Circ. 169-99. The Ohio State University. Columbus, OH. SAS Institute. 1992. SAS/STAT software: Changes and enhancements. Release 6.07. SAS Tch. Rep. SAS Inst., Cary, NC. Saxton, K.E. Rawls, W.J., Romberger, J.S., Papendick R.I. 1986. Estimating generalized soil-water characteristics from texture. Soil Sci. Soc. Amer. J. 50(4), 1031-1036. Scragg, P.K.R., Dorsch, M.M., McMichael, A.J., Baghurst, P.A. 1982. Birth defects and household water supply. Med. J. Australia 2, 577-579. Slabaugh, J.D., Jones, A.O., Puckett, W.E., Schuster, J.N. 1996. Soil survey of Levy County, Florida. USDA Natural Resour. Conserv. Serv., Washington, DC. Staples C.R., Tervola, R.S., French, E.C. 1997. Forage Production Practices by Dairy Producers in the Suwannee Valley. Cooperative Extension Service, IFAS, University of Florida. Gainesville, FL. Stewart, D.W., Shamdasani, P.N. 1990. Focus groups: theory and practice. Sage. London. Stewart, J.C. 1990. Drinking water hazards: how to know if there are toxic chemicals in your water and what to do if there are. Envirographics Publishing. Hiram, OH. Stoorvogel, J.J., Bouma, J., Orlich, R.A. 2004. Participatory research for systems analysis: prototyping for a Costa Rican banana plantation. Agron. J. 96(2), 323-336. Strange, M., Krupicka, L. 1984. Farming and cancer, in Its not all sunshine and fresh air. Center for Rural Affairs, Walthill, NE. p. 41-53.

PAGE 302

280 Suwannee River Partnership. 2004a. Suwannee River Partnership Background. Suwannee River Water Management District. Accessed May, 2004, http://www.srwmd.state.fl.us/ features/suwannee+river+partnership/background.htm Suwannee River Partnership. 2004b. Suwannee River Partnership Background. Suwannee River Water Management District. Accessed May, 2004, http://www.srwmd.state.fl.us/ resources/22x11.pdf Suwannee River Partnership. 2003. Suwannee River Partnership Programs and Research. Suwannee River Water Management District. Accessed February, 2003, http://www.srwmd.state.fl.us/features/suwannee+river+partnership/programs+and + research.htm Thomas, B.P., Cummings, E., Wittstruck, W.H., 1985. Soil survey of Alachua County, Florida. USDA Natural Resour. Conserv. Serv., Washington, DC. Thornley, J.H.M., Johnson, I.R.. 1990. A mathematical approach to plant and crop physiology. Clarendon Press. Oxford. Tomlinson, A.P., Powers, W.J., Van Horn, H.H., Nordstedt, R.A., Wilcox, C.J. 1996. Dietary protein effects in N excretion and manure characteristics of lactating cows. Trans. ASAE 39(4),1441-1448. Twatchtmann, D. 1990. Regulation of dairies in Suwannee River basin. Memorandum. Florida Department of Environmental Regulation, Tallahassee, FL. Uryasev, O. Gijsman, A.J., Jones, J.W., Hoogenboom, G. 2003. SBUILD Create/edit soil input files for evaluation and application on crop simulation models for DSSAT v4. Agricultural and Biological Engineering Department, University of Florida. Gainesville, FL. US Environmental Protection Agency (USEPA). 2003. National Pollutant Discharge Elimination System Permit Regulation and Effluent Limitations Guidelines and Standards for Concentrated Animal Feeding Operations; Final Rule. Federal Register. 68(29), 7175-7274. Feb. US Department of Agriculture (USDA), Natural Resources Conservation Service. 1999. Nutrient Management Policy, NRCS General Manual, Title 190, Part 402. Available at: http://www.nrcs.usda.gov/technical/ECS/nutrient/gm-190. US Department of Agriculure (USDA). Natural Resources Conservation Service. 1996. Agricultural waste management field handbook, Chapter 11: Waste utilization. Soil Conservation Service. 210-VI-NEH-651.11. Available at: http://www.info.usda.gov/ CED/ftp/ CED/ neh651-ch11.pdf

PAGE 303

281 US Department of Agriculure (USDA). Natural Resources Conservation Service. 1992a. Agricultural waste management field handbook, Chapter 4: Agricultural waste characteristics. Soil Conservation Service. 210-VI-NEH-651.04. Available at: http://www.info.usda.gov/CED/ftp/CED/neh651-ch4.pdf US Department of Agriculure (USDA). Natural Resources Conservation Service. 1992b. Agricultural waste management field handbook, Chapter 4: Agricultural waste characteristics. Amendment FL-9, FL651.1004(f). Soil Conservation Service. 210-VI-NEH-651.04-FL-9, FL651.1004(f) US Department of Agriculure (USDA). Natural Resources Conservation Service. 1992c. Agricultural waste management field handbook, Chapter 6: Role of plants in waste management. Soil Conservation Service. 210-VI-NEH-651.06. Available at: http://www.info.usda.gov/CED/ftp/CED/neh651-ch6.pdf. Van Horn, H.H. 1997. Manure Issues: Identifying nutrient overload, odor research report. Department of Dairy and Poultry Sciences, Univ. Florida. Gainesville, FL. Van Horn, H.H., Newton, G.L., Kidder, G., French, E.C., Nordstedt, R.A. 2001. Managing dairy manure accountability: worksheets for nutrient budgeting. Coop. Ext. Serv. Circ. 1196, Univ. Florida. Gainesville, FL. Van Horn, H.H., Newton, G.L., Nordstedt, R.A., French, E.C., Kidder, G.K., Graetz, D.A., Chambliss, C.G. 1998. Dairy manure management: Strategies for recycling nutrients to recover fertilizer value and avoid environmental pollution. Florida Coop. Ext. Serv. Circ. 1016. Univ. Florida. Gainesville, FL. Van Horn. H. H., Nordstedt, R.A., Bottcher, A.V., Hanlon, E.A., Graetz, D.A., Chambliss, C.F. 1991. Dairy manure management: strategies for recycling nutrients to recover fertilizer value and avoid environ-mental pollution. Florida Coop. Ext. Serv. Circ. 1016, Univ. Florida, Gainesville, FL. Van Horn, H.H., Wilkie, A.C.,Powers,W.J., Nordstedt, R.A. 1994. Components of dairy manure management systems. J. Dairy Sci. 77, 2008-2030. Vellidis, G., Davis, J. Lowrance, R., Williams, R. 1993. Soil water nutrient concentrations in the vadose zone of a liquid dairy manure land application site. Paper no. 93-2549. In 1993 Int. Winter Meeting, Chicago. 14-17 Dec. 1993. Am. Soc. Agric. Eng., St. Joseph, MI. Verburgh, L.D., 1994. Participative policy modeling, applied at the health care insurance industry. Ph.D. thesis. Benda, Nijmegen, The Netherlands. Vereijken, P. 1997. A methodological way of prototyping integrated and ecological arable farming systems (I/EAFS) in interaction with pilot farms. Eur. J. Agron. 7, 235-250. Vigil, J., Warburton, S., Haynes, W.S., Kaiser, L.R. 1965. Nitrates in municipal water supply cause methemoglobinemia in infant: Public Health Report, v. 80, no 12, p 112-119.

PAGE 304

282 Vrba, J. 1981. General aspects of ground-water nonpoint pollution. In Van Duijvenbooden, W. (eds.). Quality of groundwater. Proceedings of an international symposium. Studies in Environmental Science 17. London, U.K. Weatherspoon, R.L., Anderson, K., Anzalone, W., Bednarek, W., Chibirka, J., Cummings, E., Dahl, R., French, C., Jakel, D., Johnson, W.R., Neilson, R.W., Shurtliff, D. 1998. Soil survey of Lafayette County, Florida. USDA Natural Resour. Conserv. Serv., Washington, DC. Weatherspoon, R.L., Cummings, E., Wittstruck, W.H. 1992. Soil survey of Gilchrist County, Florida. USDA Natural Resour. Conserv. Serv., Washington, DC. Wilkerson, V.A., Glenn, P.B., McLeod, K.R. 1997. Energy and N balance in lactating cows fed diets containing dry or high moisture corn in either rolled or ground form. J. Dairy Sci. 80, 2487-2496. Williamson, J.C., Taylor, M.Z.D., Torrens, R.S., Vojvodic-Vukovic, M. 1998. Reducing N leaching from dairy farm effluent-irrigated pasture using dicyanamide: a lysimeters study. Agr. Ecosyst. Environ. 69, 81-88. Woodard, K.R., French, E.C., Sweat, L.A., Graetz, D.A., Sollenberger, L.E., Macoon, B., Portier, K. M., Wade, B.L., Rymph, S.J., Prine, G.M., Van Horn, H.H. 2002. N removal and nitrate leaching for forage systems receiving dairy effluent. J. Environ. Qual. 31, 1980-1992. Wright, D.L., Teare, D.L.., Rhoads, I.D., Sprenkel, R.K. 1993. Pearl millet production in a no-tillage system. North Florida research and education center, Quincy. Florida Agric. Exp. Stat. Journal Series No: R-03054. North Florida Research and Education Center, Quincy, FL. Wu, Z., Tozer, P.R., Groff E.B. 2001. Dietary manipulation of to reduce phosphorus and N excretion by lactating dairy cows. Department of Dairy and Animal Science. Pennsylvania State University. University Park, PA.

PAGE 305

BIOGRAPHICAL SKETCH Victor E. Cabrera was born on February 9, 1969, in Cusco, Peru. He received his B.S. and Engineering degrees in agronomy in 1993, from La Molina Agrarian University, in Lima. Victor did a postgraduate course in management of agricultural schools in Spain and France, in 1995. Victor received his M.S. degree in agricultural education and communication from the University of Florida, in 1999. Currently, Victor is finishing his Doctoral degree in interdisciplinary ecology in the School of Natural Resources and Environment at the University of Florida. He will graduate in August 7, 2004. Victor E. Cabrera has worked as faculty, extension agent, extension programmer, consultant, manager, and research assistant. His website is http://plaza.ufl.edu/vcabrera 283