<%BANNER%>

Evapotranspiration Based Irrigation Scheduling for a Tropical Fruit Orchard in South Florida

Permanent Link: http://ufdc.ufl.edu/UFE0025090/00001

Material Information

Title: Evapotranspiration Based Irrigation Scheduling for a Tropical Fruit Orchard in South Florida
Physical Description: 1 online resource (132 p.)
Language: english
Creator: Kisekka, Isaya
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: et, evapotranspiration, irrigation, spatial, water
Agricultural and Biological Engineering -- Dissertations, Academic -- UF
Genre: Agricultural and Biological Engineering thesis, M.E.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: EVAPOTRANSPIRATION BASED IRRIGATION SCHEDULING FOR A TROPICAL FRUIT ORCHARD IN SOUTH FLORIDA The goal of this research was to evaluate the suitability of evapotranspiration (ET)-based irrigation scheduling technologies for agricultural applications, specifically the ability of the technologies to: apply the appropriate amount of water at the appropriate time, accurately estimate reference ET (ETo) and maintain root zone soil water content in an optimum range (close to field capacity). To address this challenge, two studies were conducted with the following overall objectives: 1) evaluate ET-based irrigation water management in a tropical fruit orchard in south Florida and 2) compare various ETo estimation equations and spatial interpolation techniques in south Florida. The study was conducted at the University of Florida Institute of Food and Agricultural Sciences Tropical Research and Education Center (UF/IFAS TREC) in Homestead, FL. The experiment was conducted in an orchard of Arkin carambola. The experiment consisted of four treatments replicated three times and arranged in a completely randomized design (CRD). T1 was based on a real-time ET irrigation schedule, T2 was based on historical ET, T3 was a standard irrigation practice (76 mm/wk) and T4 was a non irrigated treatment. Irrigation was measured using water meters, soil water tension was measured using 15 cm Irrometer tensiometers, stem water potential ( ) was measured using a pressure chamber and leaf gas exchanges parameters were measured using an infrared gas analyzer. ETo sent to the real-time ET controller were recorded daily and compared to onsite estimated ETo. Performance of five ETo estimation equations including: UF IFAS (1984) Penman, South Florida Water Management District simple method, Priestley-Taylor, Turc (1961) and Hargreaves and Samani (1985) were compared against American Society of Civil Engineers-Environmental and Water Resources Institute (ASCE-EWRI) standardized ETo equation. An average of 9.9 mm d-1 of water was applied by T3 which was significantly different (P < 0.0001) from the quantity of water applied by T1 (3.1 mm d-1) and T2 (2.9 mm d-1). T1 saved 68% of irrigation water compared to T3, while T2 saved 70%. T1 average weekly soil water content (?) (29%) was significantly different from ? maintained by T2 (28.1%) and T3 (28.0%). T4 (24%) ? was significantly different (P < 0.0001) from all the other treatments. There were no significant differences in among all treatments. T2 produced the highest mean transpiration (E) of 1.9 mmol m-2 s-1 while T4 (1.71 mmol m-2 s-1) was the lowest. Stomatal conductance to water vapor (gs) followed similar trends as transpiration. Across treatments, there was no significant different in net CO2 assimilation (A) and all treatments averaged 4.7 ?mol m-2 s-1. ETo values sent to the real-time ET controller under estimated onsite ETo by an average of 25%. This was probably due to the fact that remote weather stations did not accurately represent onsite conditions. Results from comparison of ETo estimation equations revealed that the Turc (1961) method had the highest overall performances while Hargreaves had the lowest. There were no substantial differences in ETo surfaces generated using inverse distance weighted averaging and spline interpolation methods.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Isaya Kisekka.
Thesis: Thesis (M.E.)--University of Florida, 2009.
Local: Adviser: Migliaccio, Kati W.
Local: Co-adviser: Dukes, Michael D.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0025090:00001

Permanent Link: http://ufdc.ufl.edu/UFE0025090/00001

Material Information

Title: Evapotranspiration Based Irrigation Scheduling for a Tropical Fruit Orchard in South Florida
Physical Description: 1 online resource (132 p.)
Language: english
Creator: Kisekka, Isaya
Publisher: University of Florida
Place of Publication: Gainesville, Fla.
Publication Date: 2009

Subjects

Subjects / Keywords: et, evapotranspiration, irrigation, spatial, water
Agricultural and Biological Engineering -- Dissertations, Academic -- UF
Genre: Agricultural and Biological Engineering thesis, M.E.
bibliography   ( marcgt )
theses   ( marcgt )
government publication (state, provincial, terriorial, dependent)   ( marcgt )
born-digital   ( sobekcm )
Electronic Thesis or Dissertation

Notes

Abstract: EVAPOTRANSPIRATION BASED IRRIGATION SCHEDULING FOR A TROPICAL FRUIT ORCHARD IN SOUTH FLORIDA The goal of this research was to evaluate the suitability of evapotranspiration (ET)-based irrigation scheduling technologies for agricultural applications, specifically the ability of the technologies to: apply the appropriate amount of water at the appropriate time, accurately estimate reference ET (ETo) and maintain root zone soil water content in an optimum range (close to field capacity). To address this challenge, two studies were conducted with the following overall objectives: 1) evaluate ET-based irrigation water management in a tropical fruit orchard in south Florida and 2) compare various ETo estimation equations and spatial interpolation techniques in south Florida. The study was conducted at the University of Florida Institute of Food and Agricultural Sciences Tropical Research and Education Center (UF/IFAS TREC) in Homestead, FL. The experiment was conducted in an orchard of Arkin carambola. The experiment consisted of four treatments replicated three times and arranged in a completely randomized design (CRD). T1 was based on a real-time ET irrigation schedule, T2 was based on historical ET, T3 was a standard irrigation practice (76 mm/wk) and T4 was a non irrigated treatment. Irrigation was measured using water meters, soil water tension was measured using 15 cm Irrometer tensiometers, stem water potential ( ) was measured using a pressure chamber and leaf gas exchanges parameters were measured using an infrared gas analyzer. ETo sent to the real-time ET controller were recorded daily and compared to onsite estimated ETo. Performance of five ETo estimation equations including: UF IFAS (1984) Penman, South Florida Water Management District simple method, Priestley-Taylor, Turc (1961) and Hargreaves and Samani (1985) were compared against American Society of Civil Engineers-Environmental and Water Resources Institute (ASCE-EWRI) standardized ETo equation. An average of 9.9 mm d-1 of water was applied by T3 which was significantly different (P < 0.0001) from the quantity of water applied by T1 (3.1 mm d-1) and T2 (2.9 mm d-1). T1 saved 68% of irrigation water compared to T3, while T2 saved 70%. T1 average weekly soil water content (?) (29%) was significantly different from ? maintained by T2 (28.1%) and T3 (28.0%). T4 (24%) ? was significantly different (P < 0.0001) from all the other treatments. There were no significant differences in among all treatments. T2 produced the highest mean transpiration (E) of 1.9 mmol m-2 s-1 while T4 (1.71 mmol m-2 s-1) was the lowest. Stomatal conductance to water vapor (gs) followed similar trends as transpiration. Across treatments, there was no significant different in net CO2 assimilation (A) and all treatments averaged 4.7 ?mol m-2 s-1. ETo values sent to the real-time ET controller under estimated onsite ETo by an average of 25%. This was probably due to the fact that remote weather stations did not accurately represent onsite conditions. Results from comparison of ETo estimation equations revealed that the Turc (1961) method had the highest overall performances while Hargreaves had the lowest. There were no substantial differences in ETo surfaces generated using inverse distance weighted averaging and spline interpolation methods.
General Note: In the series University of Florida Digital Collections.
General Note: Includes vita.
Bibliography: Includes bibliographical references.
Source of Description: Description based on online resource; title from PDF title page.
Source of Description: This bibliographic record is available under the Creative Commons CC0 public domain dedication. The University of Florida Libraries, as creator of this bibliographic record, has waived all rights to it worldwide under copyright law, including all related and neighboring rights, to the extent allowed by law.
Statement of Responsibility: by Isaya Kisekka.
Thesis: Thesis (M.E.)--University of Florida, 2009.
Local: Adviser: Migliaccio, Kati W.
Local: Co-adviser: Dukes, Michael D.

Record Information

Source Institution: UFRGP
Rights Management: Applicable rights reserved.
Classification: lcc - LD1780 2009
System ID: UFE0025090:00001


This item has the following downloads:


Full Text

PAGE 1

1 EVAPOTRANSPIRATION BASED IRRIGATION SCHEDULING FOR A TROPICAL FRUIT ORCHARD IN SOUTH FLORIDA By ISAYA KISEKKA A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ENGINEERING UNIVERSITY OF FLORIDA 2009

PAGE 2

2 2009 Isaya Kisekka

PAGE 3

3 To my wife and daughter, R acheal and Esther Kisekka

PAGE 4

4 ACKNOWLEDGMENTS I am very grateful for the unconditional s upport I have continuously received from my parents, siblings, wife and daught er. I would also like to thank everyone at the University of Florida department of agricultura l and biological engineering in Gainesville and at the Tropical Research and Education Center (TREC) Homest ead for the encouragement and support all of you have given me over time. I would like to gi ve special thanks to Dr Migliaccio for her encouragement and patience in gui ding me through every step of the way. Dr Migliaccio, seeing your high level of efficiency in whatever you do has made a positive impact on me. My thanks also go to all my other committ ee members whose continuous counsel enabled this project to become a success. I would like to thank Dr Dukes for being such an inspirational role model. To Dr Schaffer thank you for your continuous encouragement and for availing me some of the equipment I needed for this study. Dr Crane thank you for helping me to clearly articulate water related problems faced by trop ical fruit growers in south Florida, your knowledge and experience of tropical fruits was a big boost to this study. I would also like to thank th e following people for positively contributing to the success of this project: Harry Trafford, Tina Dispenza, Mike Gutierrez, Richar d Carey, Nicholas Kiggundu and Luis Barbquin. Finally, I would like to acknowl edge the organizations that supported this study which included: USDA-CSREES, Toro company Inc. a nd Hydropoint Data systems Inc.

PAGE 5

5 TABLE OF CONTENTS page ACKNOWLEDGMENTS ............................................................................................................... 4LIST OF TABLES ...........................................................................................................................8LIST OF FIGURES .........................................................................................................................9LIST OF ABBREVIATIONS ........................................................................................................1 3ABSTRACT ...................................................................................................................... .............14 CHAPTER 1 INTRODUCTION ................................................................................................................ ..16Rationale ..................................................................................................................... ............16Evapotranspiration Concepts ..................................................................................................17Potential Evapotranspiration ...........................................................................................18Reference Evapotranspiration .........................................................................................19Estimation of Reference Evapotranspiration ...................................................................19ASCE-EWRI standardized ETo equation: Combination based method ...................20UF IFAS (1984) Penman equatio n: Combination based method .............................23SFWMD-SM equation: Ra diation based method .....................................................24Priestley Taylor equation: Radiation based method .................................................25Turc (1961) equation: Radiation based method .......................................................25Hargreaves and Samani (1985) equa tion: Temperature based method ....................26Calculation of Crop Evapotranspiration (ETc) ................................................................27Earlier Work on Comparison of ETo Estimation Equations in Florida ..................................27Spatial Interpolation of Weather Data ....................................................................................28Inverse Distance Weighted Averaging (IDWA) .............................................................29Spline Interpolation .........................................................................................................29Earlier Work on Spatial Interpolation of ETo .........................................................................30Weather Data Quality Control ................................................................................................31Irrigation Scheduling Concepts ..............................................................................................31Evapotranspiration Based Irrig ation Scheduling Devices ......................................................32Signal ET-based irrigation scheduling devices ...............................................................33Historical ET-based irrigation scheduling devices ..........................................................33Standalone (onsite) ET-based ir rigation scheduling devices ...........................................34Water Potential .......................................................................................................................35Effect of Water Stress on Pl ant Physiological Processes .......................................................35Carambola General Characteristics ........................................................................................37Carambola Crop Coefficients .................................................................................................382 EVALUATION OF EVAPOTRANSPIRA TION-BASED IRRIGATION WATER MANAGEMENT IN A TROPICAL FRUI T ORCHAR IN SOUTH FLORIDA ..................39

PAGE 6

6 Introduction .................................................................................................................. ...........39Materials and Methods ...........................................................................................................42Irrigation Treatments .......................................................................................................43Soil Moisture Measurement ............................................................................................46Measurement of Stem Water Potentia l Leaf Gas Exchange and Yield ...........................47Comparison of ETo Estimated Onsite Using Remote and Onsite Weather Data ............48Weather data collection ............................................................................................49Assessing integrity of weather data ..........................................................................49Data Analysis ...................................................................................................................50Results and Discussion ........................................................................................................ ...51Results of Quantity of Irrigation Water Applied .............................................................51Results of Assessment of Root Zone Soil Water Content ...............................................52Effect of ET-Based Irrigation Scheduling on Stem Water Potential ...............................53Effect of ET-Based Irrigation Scheduling on Transpiration (E) .....................................55Effect of ET-Based Irrigation Scheduling on Stomatal Conductance (gs) ......................55Effect of ET-Based Irrigation Scheduling on Net CO2 Assimilation (A) .......................56Results of Yield ...............................................................................................................56Results of Weather Data Quality Control ........................................................................57Onsite Comparison of ETo From Different Sources ........................................................59Conclusion .................................................................................................................... ..........593 COMPARISON OF VARIOUS REFERENCE EVAPOTRANSPIRATION ESTIMATION EQUATIONS AND SPATIAL INTERPOLATION TECHNIQUES IN SOUTH FLORIDA .................................................................................................................76Introduction .................................................................................................................. ...........76Materials and Methods ...........................................................................................................79ETo Estimation Equations ................................................................................................80ASCE-EWRI standardized ETo equation .................................................................81UF IFAS (1984) Penman ..........................................................................................82SFWMD-SM ............................................................................................................82Priestley-Taylor ........................................................................................................83Turc (1961) Equation ...............................................................................................83Hargreaves and Samani (1985) Equation .................................................................84Comparison Statistics ......................................................................................................84Spatial Interpolation of ETo .............................................................................................86Data Analysis ...................................................................................................................87Results and Discussion ........................................................................................................ ...88Comparison of daily ETo Estimations of the Various Equations ....................................88Comparison of Mean Monthly ETo Estimates for the Various Equations ......................90Results of Spatial Interpolation of ETo ............................................................................91Conclusion .................................................................................................................... ..........934 SUMARY AND CONCLUSION .........................................................................................120Objective 1 .....................................................................................................................120Objective 2 .....................................................................................................................121

PAGE 7

7 APPENDIX LIST OF REFERENCES .............................................................................................................123BIOGRAPHICAL SKETCH .......................................................................................................131

PAGE 8

8 LIST OF TABLES Table page 2-1 Site specific parameters entered into th e real-time irrigation sc heduling controller .........622-2 Average monthly ETo and carambola crop coefficients used to calculate irrigation run times per cycle for the historical ET-based treatment .................................................622-3 Dry season (October to April) daily ir rigation application an d percentage water saving. ....................................................................................................................... .........622-4 Average weekly root zone soil wa ter content (%) and suction (kPa) ................................632-5 Carambola physiological parameters and yield data. ........................................................632-6 Average daily ETo estimated at the study site using different sources of weather data and equations .....................................................................................................................643-1 Selected weather stations ................................................................................................. ..953-2 Mean Daily ETo 1 estimates by different equations (mm/day) ...........................................953-3 Performance statistics for daily ETo 1 comparison with ranking in parenthesis .................963-4 Performance statistics for mean monthly ETo 1 comparison ..............................................973-5 Interpolation methods comparisons at Homestead FL FAWN1 weather station ...............98

PAGE 9

9 LIST OF FIGURES Figure page 1-1 Carambola crop coefficient curve developed based on crop phenology ...........................382-1 Experimental plot layout. ................................................................................................. ..652-2 Experimental unit of three carambola trees spaced 4.5 m apart in an orchard of Arkin carambola in Homestead, FL. ............................................................................................652-3 DLJ Water meter ........................................................................................................... .....662-4 Installation of a Toro TIS 615 and Rainbird ESP controllers. ...........................................662-5 A 15 cm Irrometer tensiometer for soil water monitoring. ................................................662-6 Carambola leaf in aluminum foil bag. ...............................................................................662-7 Pressure chamber for stem water potential measurements. ...............................................662-8 CIRAS infrared analyzer for l eaf gas exchange measurements. ........................................662-9 Carambola harvesting and yield measurem ent at an orchard in Homestead, FL. ..............672-10 Downloading weather data from a HOBO w eather station installe d at the study area. .....672-11 Stem water potential ( ) (MPa) of carambola trees under different irrigation scheduling. ................................................................................................................... ......682-12 Transpiration (E) (mmol m-2s-1) of carambola trees under different irrigation scheduling. ................................................................................................................... ......692-13 Stomatal conductance (gs) (mmol m-2s-1) of carambola trees unde r different irrigation scheduling. ................................................................................................................... ......702-14 Net CO2 assimilation (A) ( mol m-2s-1) of carambola trees under different irrigation scheduling. ................................................................................................................... ......712-15 Measured solar radiation (Rs) reported by the Florida Automated Weather Network (FAWN) and calculated clear sky solar radiation (Rso) data for Homestead, FL against time. ................................................................................................................. ......722-16 Measured solar radiation (Rs) and calculated clear sky solar radiation (Rso) for the onsite weather station plotted against time ........................................................................722-17 Minimum temperature (Tmin) versus calculated de w point temperature (Tdew) reported by the Florida Automated Weather Networ k (FAWN) for Homestead, FL compared to the one-to-one line. ........................................................................................................73

PAGE 10

10 2-18 Minimum temperature (Tmin) versus calculated de w point temperature (Tdew) collected from the onsite weather stati on compared to the one-to-one line. .....................732-19 Daily mean of maximum and minimum temperature plotted against daily mean temperature for Florida Automated Weathe r Network (FAWN) data at Homestead, FL ............................................................................................................................ ...........742-20 Daily mean of maximum and minimum temperature plotted against daily mean temperature for the onsite weather stati on compared to the one-to-one line. ....................742-21 Daily maximum and average wind speeds for Florida Automated Weather Network (FAWN) data at Homestead, FL plotted against time. ......................................................752-22 Daily maximum and average wind speeds collected from the onsite weather station. ......753-1 Study area and location of selected weather stations in Broward and Miami-Dade Counties in south Florida ...................................................................................................993-2 Comparison of daily refe rence evapotranspiration (ETo) estimated by the ASCEEWRI standardized equation to ETo estimated by other equations using Homestead, FL weather station data (J anuary 2001 to March 2009). .................................................1003-3 Comparison of daily refe rence evapotranspiration (ETo) estimated by the ASCEEWRI standardized equation to ETo estimated by other equations using Fort Lauderdale, FL weather station da ta (January 2001 to March 2009). .............................1013-4 Comparison of daily refe rence evapotranspiration (ETo) estimated by the (ASCEEWRI standardized equation to ETo estimated by other equations using S331W, FL weather station data (Jan uary 2001 to March 2009). .......................................................1023-5 Comparison of daily refe rence evapotranspiration (ETo) estimated by the ASCEEWRI standardized equation to ETo estimated by other equations using G3ASWX, FL weather station data (J anuary 2001 to March 2009). .................................................1033-6 Comparison of daily refe rence evapotranspiration (ETo) estimated by the ASCEEWRI standardized equation to ETo estimated by other equations using JBTS, FL weather station data (Jan uary 2001 to March 2009). .......................................................1043-7 Comparison of mean monthly ETo estimated using the ASCE-EWRI standardized equation to ETo estimated using five other equations at the Homestead, FL weather station. ...................................................................................................................... ........1053-8 Comparison of mean monthly ETo estimated using the ASCE-EWRI standardized equation to ETo estimated using five other equati ons at the Fort Lauderdale, FL weather station. ................................................................................................................105

PAGE 11

11 3-9 Comparison of mean monthly ETo estimated using the ASCE-EWRI standardized equation to ETo estimated using five other equa tions at the S331W, FL weather station. ...................................................................................................................... ........1063-10 Comparison of mean monthly ETo estimated using the ASCE-EWRI standardized equation to ETo estimated using five other equa tions at the G3ASWX, FL weather station. ...................................................................................................................... ........1063-11 Comparison of mean monthly ETo estimated using the ASCE-EWRI standardized equation to ETo estimated using five other equations at the JBTS, FL weather station. .1073-12 Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for January. .................................................................1083-13 Spline based mean monthly re ference evapotranspiration (ETo) surface for January. ....1083-14 Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for February. ...............................................................1093-15 Spline based mean monthly re ference evapotranspiration (ETo) surface for February. ......1093-16 Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for March. ...................................................................1103-17 Spline based mean monthly re ference evapotranspiration (ETo) surface for March. ......1103-18 Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for April. .....................................................................1113-19 Spline based mean monthly re ference evapotranspiration (ETo) surface for April. ........1113-20 Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for May. ......................................................................1123-21 Spline based mean monthly re ference evapotranspiration (ETo) surface for May ..........1123-22 Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for June. ......................................................................1133-23 Spline based mean monthly re ference evapotranspiration (ETo) surface for June. .........1133-24 Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for July. .......................................................................1143-25 Spline based mean monthly re ference evapotranspiration (ETo) surface for July. ..........1143-26 Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for August. ..................................................................115

PAGE 12

12 3-27 Spline based mean monthly re ference evapotranspiration (ETo) surface for August. .....1153-28 Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for September. .............................................................1163-29 Spline based mean monthly re ference evapotranspiration (ETo) surface for September. .................................................................................................................... ...1163-30 Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for October. .................................................................1173-31 Spline based mean monthly re ference evapotranspiration (ETo) surface for October. ....1173-32 Spline based mean monthly re ference evapotranspiration (ETo) surface for November. ..................................................................................................................... ...1183-33 Spline based mean monthly re ference evapotranspiration (ETo) surface for November. ..................................................................................................................... ...1183-34 Spline based mean monthly re ference evapotranspiration (ETo) surface for December. ..................................................................................................................... ...1193-35 Spline based mean monthly re ference evapotranspiration (ETo) surface for December. ..................................................................................................................... ...119

PAGE 13

13 LIST OF ABBREVIATIONS ASCE American society of civil engineers ASCE-EWRI ASCE-Environmental and water resources institute FAWN Florida automated weather network SFWMD South Florida water management district IDWA Inverse distance weighted averaging ET Evapotranspiration ETo Reference evapotranspiration ETp Potential evapotranspiration ETc Crop evapotranspiration SEE Standard error of estimate RMSD Root mean square difference R2 Coefficient of determination UF University of Florida IFAS Institute of food and agricultural sciences TREC Tropical research and education center CRD Completely randomized design FAO Food and agriculture organization IA Irrigation association Kc Crop coefficient TIN Triangular irregular network GLM General linear model

PAGE 14

14 Abstract of Thesis Presen ted to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Engineering EVAPOTRANSPIRATION BASED IRRIGATION SCHEDULING FOR A TROPICAL FRUIT ORCHARD IN SOUTH FLORIDA By Isaya Kisekka August 2009 Chair: Kati W. Migliaccio Cochair: Michael D. Dukes Major: Agricultural and Biological Engineering The goal of this research was to evaluate th e suitability of evapotranspiration (ET)-based irrigation scheduling technologies for agricultural applications, specifically the ability of the technologies to: apply the appropriate amount of water at the appropriate time, accurately estimate reference ET (ETo) and maintain root zone soil wa ter content in an optimum range (close to field capacity). To address this ch allenge, two studies we re conducted with the following overall objectives: 1) evaluate ET-based irrigation water management in a tropical fruit orchard in south Florid a and 2) compare various ETo estimation equations and spatial interpolation techniques in south Florida. The study was conducted at the University of Florida Institute of Food and Agricultural Sciences Tropical Research and Education Center (UF/IFAS TREC) in Homestead, FL. The experiment was conducted in an orchard of Arki n carambola. The experiment consisted of four treatments replicated three tim es and arranged in a completely randomized design (CRD). T1 was based on a real-time ET irrigation schedule, T2 was based on historical ET, T3 was a standard irrigation practice (76 mm/wk) and T4 was a non irrigated treatment. Irrigation was measured using water meters, soil water tension was measured using 15 cm Irrometer tensiometers, stem

PAGE 15

15 water potential () was measured using a pressure chambe r and leaf gas exchanges parameters were measured using an infrared gas analyzer. ETo sent to the real-time ET controller were recorded daily and compared to onsite estimated ETo. Performance of five ETo estimation equations including: UF IFAS (1984) Penman, Sout h Florida Water Management District simple method, Priestley-Taylor, Turc (1961) and Hargreaves and Samani (1985) were compared against American Society of Civil EngineersEnvironmental and Water Resources Institute (ASCE-EWRI) standardized ETo equation. An average of 9.9 mm d-1 of water was applied by T3 which was significantly different (P<0.0001) from the quantity of water applied by T1 (3.1 mm d-1) and T2 (2.9 mm d-1). T1 saved 68% of irrigation water compared to T3, while T2 saved 70%. T1 average weekly soil water content ( ) (29%) was significantly different from maintained by T2 (28.1%) and T3 (28.0%). T4 (24%) was significantly different (P<0.0001) from all the other treatments. There were no significant differences in among all treatments. T2 produced the highest mean transpiration (E) of 1.9 mmol m-2 s-1 while T4 (1.71 mmol m-2 s-1) was the lowest. Stomatal conductance to water vapor (gs) followed similar trends as transpiration. Across treatments, there was no significant different in net CO2 assimilation (A) and all treatments averaged 4.7 mol m-2 s-1. ETo values sent to the real-time ET controller under estimated onsite ETo by an average of 25%. This was probably due to the fact that remo te weather stations did not accurately represent onsite conditions. Results from comparison of ETo estimation equations revealed that the Turc (1961) method had the highest overall performan ces while Hargreaves had the lowest. There were no substantial differences in ETo surfaces generated using inverse distance weighted averaging and spline interpolation methods.

PAGE 16

16 CHAPTER 1 INTRODUCTION Rationale Several global water scarcity studies have concluded that more than two thirds of the worlds population will live in watersheds affected by water stress by 2025 as populations increase, peoples life styles ch ange and extreme weather patter ns associated with climate change increase (Raskin et al. 1997; Shiklo manov 1998; Alcamo et al. 1997, 2000; Rijsberman 2004). According to the United Stat es Geological Survey (USGS 2007) water scarcity is already felt in several parts of the country with water disputes within and across states spreading to the Midwest, east and south United States. Water availa bility issues are also occurring in Florida, particularly the highly develope d southeastern coast. Populati on is projected to reach over 3 million people in Miami-Dade County by the year 2030 representing a 24% increase from 2006 (Miami-Dade County Department of Planning and Zoning 2004). Increas ed water demand for environmental restoration projec ts in the Everglades and prol onged drought (similar to that experienced in 2006-2007 resulting in a rainfall deficit of more 300 mm across south Florida) (Neidrauer et al. 1999; South Fl orida Water Management District 2008), all combine to exert pressure on the regions fresh water resources. Agricultural irrigation is the largest user of fresh water in Florida accounting for 45% of the states fresh water withdrawals (Marella 200 8). In Miami-Dade County, agricultural selfsupplied withdrawals of groundwater (from mainly the Biscayne aquifer) constitute over 89% of all fresh water supplies for agricu ltural production (Meralla 1999). To ensure sustainability of Miami-Dades agricultural sector, growers agree on the need to conserve water (Muoz-Carpena et al. 2003), thereby also reducing agroch emical leaching and increasing energy saving (Migliaccio 2007; Zotarelli et al 2009). Evapotranspiration (ET)-b ased irrigation scheduling that

PAGE 17

17 is being applied in landscape ir rigation in Florida has shown wate r savings of over 40% (Davis et al. 2007). If well adapted, this irrigation schedul ing technology has the po tential to result in lower waster use volumes in agriculture. Research on ET-based irrigation scheduling in agriculture, specifically for tropical fruit orchards in south Florida, were conducted to evaluate the technology s suitability in saving irrigation water used in tropical fruit orchards without nega tively affecting physiological processes of fruit trees and accurately estimatin g location specific reference evapotranspiration (ETo). The following two chapters will discuss in detail the studies that were conducted: 1) Evaluation of ET based irrigation water management in a tropical fruit orchard in south Florida and 2) Comparison of various ETo estimation equations and spatia l interpolation techniques in south Florida. Evapotranspiration Concepts ET is defined as the process through which water is lost from the earths surface by evaporation (from soil and plant su rfaces) and plant transpiration (A llen et al. 1998; Allen et al. 2005). ET is a mass transfer as well as an energy tr ansfer process in which energy is lost to the outgoing water (Burt et al. 2005). ET is affected by a number of factors including weather parameters (i.e., solar radiation, temperature, relative humidity and wind speed); crop factors (i.e., crop type, stage of growth, variety and pl ant density) and management and environmental factors (i.e., soil fertility, salin ity, pest and disease control) (Allen et al. 1998). Direct measurement of ET is expensive and difficult, requiring well trained personnel to collect accurate measurements. ET may be measured by c onducting a soil water balance with lysimeters or by conducting an energy balan ce with microclimatological met hods such as the Bowen ratio and eddy covariance systems. Lysimeters measure ET by isolating a soil-plant system from its environment and evaluating the components of the water balance by measuring incoming and

PAGE 18

18 outgoing water fluxes. The Bowen ratio and e ddy covariance methods require accurate measurement of vapor pressure, temperature a nd wind speed at differe nt heights above the evapotranspiring surface (Allen et al. 1998). Recently, remote sens ing techniques have been used to estimate ET (Courault et al. 2005). Remote se nsing methods match ET to latent heat of vaporization and estimate ET by assessing surface energy balances throug h a number of surface properties including: albe do, leaf area index, vegetation cove r and surface temperature. Using this data, ET is estimated from a quantity called the evaporative fraction (EF), defined as the ratio of the latent heat fluxes to the availabl e energy for convective fluxes. Assuming EF to be constant during the daytime, ET is estimated as the product of the EF and the net radiation (Courault et al. 2005; Tang et al. 2008). Potential Evapotranspiration The concept of potential evapotranspiration (ETp) was first introduced in the 1940s by H.L. Penman and is defined as the amount of water tr anspired in a given time by a short green crop, completely shading the ground, with uniform height and having adequate water in the soil profile (Penman 1948). The ETP concept was introduced to eliminate crop specific changes in the ET processes. Penman derived the ETP (Equation 1-1) based on the en ergy balance approach at the transpiring surface. ETp is only influenced by meteorological processes. The problem with the ETP definition was that many crops could fit the description of a short green well watered crop. For purposes of uniformity, the choice of the green crop became a problem. a n PER ET / (1-1) Where Ea is defined as follows: 2006.05.0 263.0 u ee Eda a (1-2)

PAGE 19

19 pET represents potential evapotranspiration from a vegetated surface [mm d-1], is slope of vapor pressure-temperature curve for air [mb C-1], is psychrometric constant 0.66 mb C-1, is albedo or canopy refl ection coefficient, nR is net radiation [cal cm-2 d-1], ae is vapor pressure of air [mb], is latent heat of vaporization [cal cm-1 mm-1] about 58 cal cm-2 at 29oC, 2uis wind speed at 2 m height [km d-1] and de vapor pressure at dew point temperature [mb] (Jones et al. 1984). Reference Evapotranspiration In the 1970s the concept of ETo was developed as a replacement for the term ETP. The reason for the replacement was ETo was more practical and clear ly defined the vegetative characteristics of the reference surface called the hypothetical crop (A llen et al. 1998; Allen et al. 2005). Reference evapotranspiration is defined as the rate at which readily available soil water is lost through the processes of ev aporation and transpiration from a specified vegetative surface (Jensen et al. 1990). More quantitatively ETo refers to the rate of ET from a hypothetical crop surface, defined as a uniform surface of dense, actively growing vegetation having a specified height and surface resistance, not short of soil wa ter and representing a distance of at least 100 m of similar vegetation (Allen et al. 2005). It is worth noting that the concept of ETo was introduced to study the evaporative demand of the atmosphere independent of crop type, stage of development and management practices (Allen et al. 1998). Therefore, ETo is a parameter expressing the evaporative index of the atmosphe re and is computed from weather data and atmospheric parameters. Estimation of Reference Evapotranspiration ETo can be computed from mete orological data such as solar radiation, temperature, relative humidity and wind speed. ETo can also be estimated using an evaporation pan in which

PAGE 20

20 ETo is given as the product of the pan coefficien t and the amount of water evaporated from the pan in a given time. The most common methods for estimating ETo are the meteorological based methods, which can be categorized into temper ature-based methods, ra diation-based methods and combination-type methods. Several studies have established that combination type equations, more specifically the Penman Mont eith (PM) equation, yiel d the most accurate estimate of ETo across all climatic conditions (Katul et al. 1992; Allen et al. 1998; Wright et al. 2000; Itenfisu et al. 2003). The suitabili ty of other methods for estimating ETo vary with climate probably due to the fact that they were developed for specific climatic conditions. There are many meteorological base d methods used to estimate ETo, but for purposes of this study, six were selected: two combination ty pe equations, three radi ation based equations and one temperature based equation. These include: 1) American Society of Civil Engineers (ASCE-EWRI) standardized ETo equation, 2) University of Flor ida (UF) Institute of Flood and Agricultural Sciences (IFAS) (1984) Penman equation, 3) South Flor ida Water Management District (SFWMD) simple method, 4) Priestley-Taylor, 5) Turc (1961) and 6) Hargreaves and Samani (1985) equation. The most widely accepte d of all the equations for calculation of ETo is the ASCE-EWRI standardized ETo equation which is based on the ASCE-PM equation. It has received endorsements from the Irrigati on Association (IA), ASCE and the Food and Agricultural Organization (FAO) of th e United Nations (Allen et al. 2005). ASCE-EWRI standardized ETo equation: Combination based method In late 1990s there arose a need to have a standardized equation for calculating ETo that would enable transfer of crop coefficients and reduce confusion as to which equation to use in estimating ETo, since many equations had been devel oped and published for this purpose. Collaboration between the Irrigation Associat ion (IA) and ASCE on the formulation of a standardized ETo estimation equation led to the devel opment of the ASCE-EWRI standardized

PAGE 21

21 ETo equation (Equation 1-3) which is now used as the standard for calculating evaporative demand (Allen et al. 1998; Allen et al. 2005) The ASCE-EWRI standardized equation was derived from the ASCE-PM equation by fixing the crop height h=0.12 m for short reference grass (ETo) and h=0.5 m for tall reference crop or alfalfa (ETr). The constants in the equation depend on the type of reference surface, calculati on time step and the time of day (night or day) (Allen et al. 2005). 2 21 273 408.0 UC eeU T C GR ETd as n n o (1-3) Variables in the equation are expressed as follows. 23.237 3.237 27.17 exp2503 T T T (1-4) 2min maxTeTe eo o s (1-5) 3.273 27.17 exp6108.0 T T Teo (1-6) 2 100 100min max max minRH Te RH Te eo o a (1-7) s nsR R 1 (1-8) 2 14.034.04 min 4 max K K a cd nlTT e fR (1-9) 35.035.1 so s cdR R f (1-10) a soRzx R510275.0 (1-11)

PAGE 22

22 s srsc adGR sincoscossinsin 24 (1-12) J dr365 2 cos033.01 (1-13) 39.1 365 2 sin409.0 J (1-14) tantan arccos s (1-15) 42.58.67 87.42 w zzIn uU (1-16) Px310665.0 (1-17) OET represents reference evapotranspiration [mm d-1],nR is net radiation at the crop surface [MJ m-2 d-1], G is soil heat flux density [MJ m-2 d-1], T is mean daily air temperature at 2 m height [C], minT is minimum daily air temperature at 2 m height [C], maxT is maximum daily air temperature at 2 m height [C], min KT is minimum daily air temperature at 2 m height [K], max KT is maximum daily air temperature at 2 m height [K], 2U is wind speed at 2 m height [m s-1], zu is wind speed at wz m height [m s-1], se is saturation vapor pressure [kPa], ae is actual vapor pressure [kPa], asee is saturation vapor pressure deficit [kPa], is slope of vapor pressuretemperature curve [kPa C-1], is psychrometric constant [kPa C-1], P is measured mean atmospheric pressure [kPa], is albedo or canopy refl ection coefficient (0.23), is StefanBoltzmann constant 4.901 x 10-9 MJ K-4 m-2 d-1, fcd is cloudiness function(0.05 fcd 1.0), soR is calculated clear-sky radiation [MJ m-2 d-1],aR is extraterrestrial radiation [MJ m-2 d-1], z is station elevation above sea level [m], rd is inverse relative dist ance factor for the earth-sun, is solar declination [rad], is latitude [rad], is sunset hour angle [rad], J is Julian day, scG is solar

PAGE 23

23 constant 4.92 MJ m-2hr-1, sR is incoming solar radiation [MJ m-2 d-1], nlR is net outgoing longwave radiation [MJ m-2 d-1], nsR is net short-wave radiation [MJ m-2 d-1], RH is relative humidity [%], and Teo is saturation vapor pressure function [kPa]. In Florida a short reference crop is commonly used in the estimation of ETo since alfalfa is not produced on a large scale (Irmak and Hama n 2003). Therefore, considering a 24 hour time step constantsnCand dC values are 900 and 0.34, respectively (Allen et al. 2005). UF IFAS (1984) Penman equati on: Combination based method The UF IFAS (1984) Penman equation is de rived from an energy balance at the soil surface and is based on four parameters: 1) net ra diation, 2) air temperature, 3) wind speed and 4) vapor pressure deficit. The UF IFAS (1984) Penman equation (Equation 1-18) is currently being used to predict ETo for the Florida Automated Weather Network (FAWN) (Jones et al. 1984): /42.042.108.056.0 14 so s d s oIFASR R e TR ET 20062.05.0(263.0 u eeda (1-18) Variables are defined as follows: avgT 055.059.59 (1-19) ]0000342.0)8072.000738.0(05904.0[8639.337 T (1-10) ]001316.0)488.1(000019.0)8072.000738.0[(8639.33)(8 T T Te (1-21) 2min maxee ea (1-22) bs nRR R 1 (1-23)

PAGE 24

24 42.0/42.108.056.04 sos d bRRe TR (1-24) so sRS R 61.035.0 (1-25) 2min maxTT Tavg (1-26) 273 avgTT (1-27) 2.0 22 Z uuz (1-28) oIFASET represents potential evapotranspiration from a vegetated surface [mm d-1], is slope of vapor pressure-temperature curve for air [mb C-1], is psychrometric constant 0.66 mb C-1, is albedo or canopy refl ection coefficient, sR is total incoming solar radiation [cal cm-2 d-1], is Stefan-Boltzmann constant 11.71 x 10-8 cal cm-2d-1K-1, T is mean daily air temperature [K], avgT is average temperature [oC], ae is vapor pressure of air [mb], soR is total daily cloudless sky radiation [cal cm-2 d-1], S is percent sunshine hours, is latent heat of vaporization [cal cm-1 mm1] about 58 cal cm-2 at 29oC, 2uis wind speed at 2 m height [km d-1], zu wind speed at height z [km d-1], z is height of wind measurement [m] and de vapor pressure at dew point temperature [mb] (Jones et al. 1984). SFWMD-SM equation: Radiation based method A simple ETp estimation equation for south Florid a was developed at the SFWMD and calibrated using three year (1993 to 1995) of lysimeter data colle cted at the Everglade nutrient removal project constructed wetla nd. This equation only requires solar radiation (Equation 1-29) (Abtew et al. 1995): s pR KET1 (1-29)

PAGE 25

25 ETp represents potential evapotranspira tion from a vegetated surface [mm d-1], sR is total incoming solar radiation [MJ m-2 d-1], is latent heat of vaporization [MJ kg-1] and 1K is a regression coefficient (0.53). Priestley Taylor equation: Radiation based method The Priestly Taylor equation is a simplifica tion of the Penman combination equation with only radiation as the input and is valid under conditions of minimum advection and for areas with low moisture stress (Pri estley and Taylor 1972; Vija y and Frevert 2002). The ET under potential conditions using the Priestly Taylor method from a vegetated surface is expressed as (Equation 1-30) (Priestley and Taylor 1972; Sumner and Jacobs 2004): GR ETn P (1-30) ETP represents potential evapotranspira tion from a vegetated surface [mm d-1], is slope of vapor pressure-temperature curve for air [kPa C-1], is psychrometric constant [kPa C-1], is a constant determined empirically; for humid areas it is equal to 1.26, nR is net radiation [MJ m-2 d-1], G is soil heat flux [MJ m-2 d-1] and is latent heat of vaporization [MJ kg-1]. The net radiation, slope of the vapor pressure-temperatu re curve, latent heat of vaporization and psychrometric constant are evaluated using the ASCE-EWRI standardized ETo equations procedures and the soil heat flux is assumed to be 0 for daily ca lculation time steps based on the recommendations of Allen et al. (2005). Turc (1961) equation: Radiation based method The Turc (1961) method for estimating ETP is based on solar radia tion and temperature as the inputs. Several studies have shown that Turc provides similar results as lysimeter data and combination equations like the Penman and Penman Monteith (Jensen et al. 1990; Abtew et al.

PAGE 26

26 1996a; Kashyap and Panda, 2001; Jacobs and Sa tti, 2001; Nandagiri and Kovoor, 2006). The Turc method for sub humid climates with relative humidity greater than 50% is calculated as (Equation 1-31): )500239.0( 15 013.0 s pR T T ET (1-31) ETP represents potential evapotranspira tion from a vegetated surface [mm d-1], T is daily mean air temperature [C] and sR daily solar radiation [cal cm-2 d-1]. Hargreaves and Samani (1985) equation: Temperature based method The Hargreaves and Samani (1985) equation wa s developed for dry areas in the western United States and is the only temperature based equation that has produced results with some global validity; many other temper ature based equations require local calibration (Allen et al., 1998). This method is described by Hargre aves and Samani (1985) (Equation 1-32): 8.17 0023.05.0 min max TTTR ETa P (1-32) ETp represents potential evapotranspira tion from a vegetated surface [mm d-1], T is daily mean air temperature [C], minT is minimum daily air temperature at 2 m height [C], maxT is maximum daily air temperatur e at 2 m height [C] andaR is extraterrestrial radiation, values of aR for all latitudes can be obtained from the Smithsonian meteorological tables [MJ m-2 d-1], aR is a function of latitude, date and time of day and is converted to units of depth of water per day using Equation (1-33) in which represents latent of va porization of water [MJ kg-1] and w is density of water [kg m-3]. wdaymMJ Radiation daymm Radiation]//[ ]/[2 (1-33)

PAGE 27

27 Calculation of Crop Evapotranspiration (ETc) Crop evapotranspiration, or sometimes referred to as actual evapotranspiration, is calculated as the product of the estimated ETo or ETp (for purposes of this study ETp is treated as equivalent to ETo) of a specified location and the crop coefficient (Kc) of a particular crop (Equation 1-34). The Kc integrates the following four main characteristics that distinguish the crop from the reference grass: 1) crop height which influences the aerodynamic resistance, 2) albedo which influences the net ra diation of the surface, 3) canopy resistance or the resistance of the crop to vapor transfer that influences surface resistance an d 4) evaporation from the soil particularly bare soil (Allen et al. 1998). cocKETET (1-34) Earlier Work on Comparison of ETo Estimation Equations in Florida There have been earlier studies conducte d in Florida to identify the best ETo estimation equations in comparison to lysimeter data (S tephens and Stewart 1963 and Abtew et al. 1996a) and to identify which simpler methods produced results closet to the recommended Penman Monteith (Jacobs and Satti 2001). Using lysime ter data from St. Augustine grass in Fort Lauderdale, Stephens and Stew art (1963) compared nine ETP equations under south Florida climatic conditions and noted that combination based methods like the Penman and modified Penman were most accurate. In situations of low availability of me teorological data, they recommended the modified Blaney-Cradle e quation. Jacobs and Satti (2001) evaluated performance of various ETo methods in north and central Florid a and they noted that temperature based equations underestimated ETo during the winter and overestimated ETo during the summer months. Performance of the Hargreaves met hod exceeded all the other temperature based equations, especially in the summer months. Jac obs and Satti (2001) also noted that radiation

PAGE 28

28 based ETo estimation methods produced better results than the temperature based methods but still under estimated ETo during the winter compared to the ASCE-PM equation. The Turc method under estimated ETo during the summer but captured the overall patterns very well. Performance of radiation based equations vari ed with location. Combin ation equations were reasonably consistent throughout th e year especially the FAO Pe nman Monteith equation (also known as FAO-56) (which is very similar to the ASCE PM). Using long term water balances for 36 forest watersheds, Lu et al. (2005) compared performance of six potential evapotranspiration estimation equations in southeastern United Stat es and made the following conclusions: 1) ETP from different methods were significantly diffe rent and the differen ces were greater under temperature based compared to radiation based equations and 2) under conditions of limited data availability, Priestley-Ta ylor, Turc and Hamon ETP methods were recommended. Hamon is a temperature based method that only requires day time lengths (hour s) and temperature (oC) (Hamon 1963). Spatial Interpolation of Weather Data Weather data are collected at meteorological stations that re present a specific location and therefore provide data correspond ing to weather at that single location. Spatia l interpolation could be used to estimate the values of weat her parameters at locations in-between weather stations so that discrete data points in space can be translated into a complete surface area of values (Hartkamp et al. 1999). A simple interpolat ion method used in evaluating spatial rainfall data is the Thiessen polygon method in which the properties of the un-sampled point are predicted using the nearest sampled point. For more accurate analysis, more sophisticated methods can be applied, such as Inverse Distan ce Weighted Averaging (IDWA), Spline, Kriging and Co-Kriging (Hutchinson 1991; Hutchinson an d Corbett 1995; Hartkamp et al. 1999). However, for purpose of this study only the first two approaches will be discussed.

PAGE 29

29 Inverse Distance Weighted Averaging (IDWA) The IDWA method estimates weather parameters for un-sampled locations using a linear combination of values from the sampled locations Weighting of sampled locations is a function of the distance from the un-sampled locati on and the sampled location. The underlying assumption in this method is that the values closest to the un-sampled location are more representative of the un-sampled location, thus sampled locations near the un-sampled site have larger weights than those further away. The ID WA has been used in interpolating climatic parameters like surface air temperature and preci pitation (Stallings et al. 1992; Hartkamp et al. 1999). The choice of the weights in the IDWA me thod (Equation 1-35) aff ects the interpolation output. For example, higher weights result in values close to the nearest neighbor method. iixyxy (1-35) i represents the weight assigned to a sampled location (weights are assigned based on the inverse of the distance between th e sampled and un sampled point) ixy is a value of the variable measured at the sampled location and xy predicated value of the variable at the unsampled location. The sum of the weights should equal to one. Spline Interpolation Spline interpolation is a form of data inte rpolation in which the function performing the interpolation is a piecewise pol ynomial called a spline. The splin e function generates a smooth and continuous curve across all the sampled data points. Splines are preferred over ordinary polynomials because they minimize interpolati on errors by minimizing the difference between the predicted data and the sampled data range. Th e most commonly used type of spline is cubic spline which fits piecewise polynomials across al l the data points (McKinley and Levine 1998):

PAGE 30

30 n n nxxxifxs xxxifxs xxxifxs xS1 1 3 2 2 2 1 1)( )( )( )( (1-36) )(xS, represents a piecewise function that interpolates all data poi nts and is continuous over the interval x1 to xn and is represents the thir d degree polynomialiiiiiii idxxcxxbxxaxs )()()()(3 (1-37) i =1, 2, 3n-1 n is the number of data points and ai, bi, ci and di are coefficients of the spline that bend the curve to pass through each data po int without erratic behavior or break in continuity In the generation of a spline, the first and second derivativ es of Equation (1-37) must be continuous on the interval [x1, xn]. Since the piecewise function in terpolates all the data points the value of the un sampled point yi at point xi is: iiyxS ) ( (1-38) Earlier Work on Spatial Interpolation of ETo McVicar et al. (2007) produced 100 m resolution spatially distributed surfaces of monthly ETo for the Loess plateau in China. Using a spline model with a linear sub-model based on elevation, McVicar et al. (2007) spatially interpolated maximum and minimum air temperatures, wind speed and vapor pressure and modeled solar ra diation while taking into account the effects of topographic change. They concluded that to pography, especially aspe ct, had a significant effect on the resultant ETo, which had also been reported by Chuanyan et al. (2004). Another study involving spatia lly interpolating ETo was completed for the Yangtze river catchment in China by Xu et al. (2006). Xu et al. (2006) did not consider the effects of changing elevation on meteorological parameters and used the calculate then interpolate method. In this method, ETo is estimated at points with measur ed meteorological parameters before it is spatially interpolated.

PAGE 31

31 In the south Florida Water Management Model (SFWMM), the SFWMD uses the Triangular Irregular Network (TIN ) linear interpolation model to generate spatially distributed surfaces of wet marsh ETP for south Florida at a spatial resolution of 3.2 x 3.2 km (Irizarry-Ortiz 2003). Topographical effects were not considered in this study due to the relatively flat topography of south Florida with many places less than 10 m above sea level. Weather Data Quality Control Quality control of solar radiation (Rs) was performed by plotting Rs and Rso (clear sky solar radiation) over time (Allen et al. 1998) For a well calibrated and properly operating pyranometer, Rso should plot as an upper envelope of the measured daily solar radiation. For properly collected relative humidity data, the de w temperature and minimum temperature should be similar the majority of the time. For good qual ity temperature data, daily average temperature and the mean of the daily maximum and minimum temperatures should be almost the same with maximum differences of about 3oC, if there are no changes in air mass (Allen et al. 2005). The best wind speed data should have average wind speed readings of higher than 1 m s-1 at a measurement height of 2 m. Irrigation Scheduling Concepts Irrigation scheduling is the process through whic h water lost through actual or crop ET is replaced to maintain a desired soil moisture conten t. In general, water requirements of plants are determined from a balance of water inputs and ou tputs of the root zone. The major inputs to the root zone include rainfall, ir rigation and capillary contributi ons. The main outputs include ET, runoff and percolation. All the abov e parameters are summarized in what is referred to as the root zone soil water balance (Equation 1-39): RDSCIPETc (1-39)

PAGE 32

32 ETc represents actual or crop evapotranspiration, P is precipitation, I is net irrigation (gross irrigation which is the actual am ount of irrigation water applie d includes both the net irrigation and some additional water to account for irrigation system inefficiencies), C is capillary contribution, S is change in soil water storage, D is drainage and R is runoff (all variables with units of depth) (Burt et al. 2005). For irrigation sc hedules designed to maintain specific root zone soil moisture content, deep percolation and runo ff tend to be zero. Thus, a reduced form of the root zone soil water balance can be us ed to calculate irrigation needs: PETIc (1-40) When managing for specific soil water content, irrigation should be applied when a certain percentage of the available water is deplete d. This percentage is called the management allowable depletion (MAD) and depends on management requirements. However, in the absence of specific information, a MAD of 50% is normally used in Florida (Dukes et al. 2009) where: rWP FCZ TAW )(1000 (1-41) TAWxMAD RAW (1-42) TAWis total available water [mm], FC is water content at field capacity [m3 m-3], WP is water content at wilting point [m3 m-3] and rZ is rooting depth [m]. Coupled with TAW is the readily available soil water (RAW). RAW is the fraction of TAW that is readily available for crop use and can be depleted with out stress to the crop. Evapotranspiration Based Irrigation Scheduling Devices ET-based irrigation scheduling de vices use weather data (i.e., solar radiation temperature, humidity, wind speed and rainfall) in conjunc tion with plant char acteristics (i.e., crop coefficient), irrigation system characteristics (app lication rate and efficiency) and site specific conditions (i.e., latitude, soils, sl ope and shade) to schedule irriga tion (US Department of Interior

PAGE 33

33 and Bureau of Reclamation 2007; Dukes 2009). Depending on the ET-based irrigation scheduling method used, all or some of the above site specific paramete rs are incorporated. ETbased irrigation scheduling methods are classified based on the way weather data used in estimating the ETo is received. There are generally three types of ET-based irrigation scheduling: 1) signal based ET, 2) historical based ET and 3) onsite measurement or stand alone ET-based irrigation scheduling. There are many commercially available irrigation control systems whose mode of operation is based on the principle of replacement of water lost through ET and some are equipped with rainfall sensors to account for the effective contribution of rainfall. Signal ET-based irrigation scheduling devices Signal ET-based irrigation sc heduling devices receive ETo data via wireless signals from a company that uses meteorological data collected from public weather stations or a network of private weather stations located near the irrigation site to estimate ETo (Davis et al. 2007). Most companies (e.g. HydroPoint Data Systems Inc., Petaluma, CA and Rain Bird Inc., Azusa, CA) calculate ETo using the ASCE-EWRI standardized ET equation (Allen et al. 2005; US Department of Interior and Bureau of Reclam ation, 2007). The number of weather stations used to estimate ETo for a given irrigation site depends on the weather data service providers. For signal ET-based irrigation schedules, there is norma lly a service charge associated with receiving the signal that ranges from $15 to $50 per year (Davis et al. 2007). Information specific to the crop, such as the crop coefficient, can either be entered by the user or pr eset by the manufacturer. Signal-based approaches are currently applied in lawn and ornamental landscape sites. Historical ET-based irrigat ion scheduling devices Historical ET-based irrigation scheduling is base d on the idea that irrigation is adjusted to replace the amount of water hist orically observed to be lost from the system by ET for the location of interest. Typically, a user would determine monthly ETo rates and adjust the irrigation

PAGE 34

34 schedule to reflect this. Historic ET estimation is not as accurate as real-time (as expected) because of variability in weather patterns that occur daily and annually. However, there are attachments like temperature and so lar radiation sensors that can be added to the historical ETbased irrigation scheduling de vices to modify daily ETo estimations (Dukes 2009). Examples of commercially available historical ET based irri gation scheduling devices include: Alex-Tronix Enercon Plus (Alex-Tronix Inc., Fresno, CA) which uses stored historical climatic data and an onsite temperatures sensor to adjust daily ETo estimates, Aqua Conserve ET-6 Controller Series (Aqua Conserve Inc., Riverside, CA) which uses an onsite temperature sensor and stored daily historical ETo to schedule irrigation and Hydrosaver ETIC Controller (Hydr osaver Inc., Signal Hill, CA) which uses an ET scheduling engine that is based on only historical ETo (US Department of Interior and Bureau of Reclama tion 2007). Historical weat her data can also be used to estimate mean monthly ETo that is used to calculate ETc and entered manually into an automatic irrigation timer. For example, a su rvey conducted among tropical fruit growers in Miami Dade County revealed th at 50% of respondents used automatic irrigation timers and monitored weather data to guide their irriga tion scheduling decision making (Muoz-Carpena et al. 2003). Standalone (onsite) ET-based irrigation scheduling devices Standalone ET-based ir rigation scheduling uses onsite sens ors to measure a limited number of weather parameters (e.g., temperature and solar radiation) to estimate ETo as opposed to a complete weather station. The measured weather parameters are transmitted to the ET scheduling device using wires or wireless t echnology. The irrigation scheduling device then uses a computer to calculate reference ETo (US Department of Interior and Bureau of Reclamation 2007). Although the irrigation scheduling devices may record weather data every 15 min, the ETo used in scheduling irrigation is usually the daily average (Davis et al. 2007) Generally, standalone

PAGE 35

35 weather stations do not use the most wi dely accepted ASCE-EWRI standardized ETo equation (Allen et al. 2005) to calculate ETo due to the limited number of weather parameters collected; most use other simpler temperature or solar radiation dependent methods such as the Hargreaves equation to estimate ETo. Water Potential Water potential refers to the total potential energy of water in a system and is essential in determining the flow of water (pure water has th e highest water potential and water flows from high to low water potential) (Jones et al. 1985). The four components that constitute water potential include: gravit ational water potential gthat depends on the location of water in the gravitational field (for plants th is components is usually small and often o ignored), matrix water potentialmthat depends on the forces holding water (e.g. in the soil matrix), osmotic potential othat depends on concentration of solute s in solution and pressure potential pthat depends on the hydrostatic pressure (Jones et al. 1985; Taiz and Zieger 1998) Water potential is used as an indicator of plant water stat us. A more negative water potential indicates lower water status (i.e. signs of water stress) while less negative water potential indicates high water status (i.e. well watered plant). Effect of Water Stress on Pl ant Physiological Processes Stem water potential is the free energy of wa ter in the xylem and can be measured using pressure chambers (Scholander et al. 1965; Taiz and Zieger 1998) The pressure chamber operate on the principle that the external pressure required to recover the water column in the xylem before the leaf was removed from the plant is eq ual to but opposite in sign to the xylem (stem) water potential (Scholander et al. 1965; Taiz and Zi eger 1998). In the xylem m and otend to very below low compared top (e.g. typically o > -0.1 MPa compared to a typical midday

PAGE 36

36 xylem tension of -1 MPa
PAGE 37

37 Carambola General Characteristics Carambola also called star fruit is a member of the Oxalidacea family (Nakasone and Paull 1998). The exact origin of carambola is not known, but it is believed to have originated from Southeast Asia and was introduced to Florida a century ago (Mossler and Neshiem 2002). In their natural tropical environment carambola tr ees produce fruit all year, given the minimum presence of environmental stress factors like wind th at causes leaves to fall off the trees and cold temperatures that limit physiological processes such as plant water uptake (Nuez-Elisea and Crane 2000). Carambola has a smooth short tree tr uck with an evergreen canopy with branches very close to the base of the trunk. Leaves can reach 20 cm in length and are arranged in an alternate manner with young leaves taking on a bron ze-red color and the matu re leaves dark to pale green. Carambola fruit is a fleshy berry wi th a star shape in cross-section and 5-6 longitudinal ribs of 5-12 cm long (Nakasone and Paull 1998). The fruit is considered ripe when it turns green to gold or yellow to orange. In Flor ida and Hawaii, fruit is mature 60-65 days from fruit set. Oxalic and malic acid levels decrease wh ile sugar levels increase as the fruit ripens, but fruit sugar content does not increas e after the fruit is picked (N unez-Elisea and Crane 2000). This might explain why the sweetest fruits are those that ripen on the tree. Carambola is most productive in tropical and subtropical low lands w ith well drained soils and uniformly distributed rainfa ll. In Florida the commonly gr own cultivars of carambola are arkin and golden star (Simonne et al. 2004). Caram bola is planted at spacing of 4 to 6 m between plants and 6 to 7 m between rows. In southeast Florida there are two major bloom periods (April to May and September to October) consequently with two harvest seasons (August to September and December to February) (Nuez-Elisea and Crane 2000).

PAGE 38

38 Carambola Crop Coefficients There is very little information in the literature for carambola crop coefficients. Crane (1993) recommends application of 2. 54 to 5.1 cm of water after seven to 10 days of little or no rainfall. Based on carambola phenology, the crop coeffi cient curve (Figure 1-1) is being used to guide irrigation scheduling in south Florida (Crane, 2008 personal communication). Figure 1-1. Carambola crop coefficient cu rve developed based on crop phenology 0.9 0.95 1 1.05 1.1 1.15 1.2 1.25 NovDecJanFebMarAprMayJunJulAugSepOctCrop coefficient (K c )Defoliation Harvesting Flowering &fruit set Harvesting Flower bud dev Flower bud dev

PAGE 39

39 CHAPTER 2 EVALUATION OF EVAPOTRANSPIRA TION-BASED IRRIGATION WATER MANAGEMENT IN A TROPICAL FRUIT ORCHAR IN SOUTH FLORIDA Introduction Many global scale water scarcity assessment studies have concluded that up to two thirds of the worlds population will live in water st ressed watersheds by the year 2025 (Shiklomanov 1998; Rijsberman 2004; Alcamo et al. 2007). This trend in fresh water availability has been attributed to increasing populati on, raising incomes, water pollu tion and the negative impacts of global climate change (Raskin et al. 1998; Rijsberman 2004; Arnell 2004, Alcamo et al. 2000, 2007). Gallopin and Rijsberman (2000) noted that future trends in water demand will heavily depend on success of current water c onservation initiatives and life st yles of future generations. According to the United States Geological Surv ey (USGS 2007) water scarcity is already felt in several parts of the country with water disputes within and across states spreading to the Midwest, east and south United States. Competiti on for fresh water among different users is also occurring in Florida especially in the southern part of the st ate due to this regions high population growth rate, environmental restorati on projects in the Ever glades and prolonged droughts (similar to that experienced in the 20062007 that resulted in a ra infall deficit of more 300 mm across south Florida) (Nei drauer et al. 1999; SFWMD, 2008) Despite this situation of growing water scarcity and uncertainty of future water supplies, irrigated agriculture still remains the number one consumer of freshwater in Flor ida accounting for 45% of the states fresh water withdrawals (Marella, 2008). In south Florida, agricultural self supply withdrawals from the Biscayne aquifer constitute over 89% of all the water used for agricultural production (Marella, 2008). This underscores the fact that sustainability of south Florid as agricultural sector heavily depends on availability, and thus, co nservation of this water resource.

PAGE 40

40 One of the major agricultural subsectors in Miami-Dade County is the tropical fruit industry, encompassing over 6,475 ha with an annual crop value of $137 million (Crane, 2003). To ensure sustainability of this subsector, gr owers agree on the need to conserve water (MuozCarpena et al. 2003). The desire to conserve water by the growers has developed out of the need for more sustainable producti on that increases yield, reduces agrochemical leaching to groundwater and saves energy (Migliaccio, 2007; Zotarelli et al. 2009). Although soil water content based irrigation sche duling technologies (e.g. soil wa ter sensors, tensiometers, capacitance probes) have shown water savings of over 90% (Migliacci o, 2007), there has been slow adoption of these technologies in agricult ure. The slow adoption of these technologies by growers has been associated with high initial co st of some soil water c ontent monitoring devices like capacitance and neutron probes and the conti nual maintenance requirements associated with simpler soil water monitoring de vices like tensiomete rs (Al-Yahyai et al. 2003). Advancements in irrigation technology in the la st 20 years has led to the deve lopment of evapotranspiration (ET)-based irrigation scheduling devices which are designed to apply irrigation water that replaces that lost through crop ET. ET-based irrigation scheduling de vices are irrigation control sy stems that use weather data (i.e., solar radiation temperatur e, humidity, wind speed and rainfa ll) in conjunction with plant characteristics (i.e., crop coefficient), irrigation system character istics (e.g. application rate and uniformity) and site specific c onditions (i.e., latitude, soils, ground slope and shade) to schedule irrigation (US Department of Interior and Bureau of Reclamation, 2007; Dukes, 2009). ET-based irrigation scheduling devices are cl assified based on the way weather data is used in generating the irrigation schedules is received. The thr ee main types of ET-based irrigation scheduling devices are: 1) signal ba sed ET, 2) historical based ET and 3) onsite or stand alone ET-based

PAGE 41

41 irrigation scheduling controllers. There are ma ny commercially available irrigation control systems but all are operated based on the principl e of estimating water losses to ET and replacing these losses with irrigati on water. In addition, some are equipped with rainfall sensors to account for the effective rainfall. Recent studies in Florida have shown that ET-based irrigation scheduling devices provide a practical tool for irrigation scheduling at mi nimum cost and labor re quirement for Florida ornamental landscapes, with water savings excee ding 40% (Davis et al. 2007). There have also been studies conducted in western St ates, mainly in California to ev aluate the suitability of these irrigation scheduling devices to save water for re sidential irrigation applications. Hunt et al. (2001) reported that ET-based irrigation schedu ling devices produced water saving of up to 24% in the Irvine water district, Orange County, Calif ornia. In addition, the Irrigation Association (IA) developed a protocol for determining the e ffectiveness of an irrigation schedule generated by ET-based irrigation scheduling devices. The pr otocol evaluates the ability of ET-based irrigation scheduling devices to ad equately and efficiently irrigate a landscape by using a virtual landscape subjected to a representative climate (Irrigation Association 20 06). To the authors knowledge, there is no information available on a standard protocol to evaluate ET-based irrigation scheduling devices wh en applied in agricultural settings. However, automatic ETbased irrigation scheduling devi ces are being tested by avoca do growers in California in collaboration with a private weather data service provider (Hydropoint Data Systems Inc., Petaluma, CA) (California Farm Bureau Federation, 2008). Little information is available in scientif ic literature on the suitability of ET-based irrigation scheduling technologies for agriculture. Specifically, information is not available evaluating the technologys ability to: accurately apply irrigation water that replaces that lost

PAGE 42

42 through crop ET, maintain root zone soil water content in an optimum range (close to field capacity) reduce irrigation water volumes app lied and save water without limiting optimum plant growth. To address this problem, a study wa s conducted in a tropical fruit orchard in south Florida with the following objectives: 1) evaluate the amount of irrigati on water applied by realtime and historical ET-based irrigation schedulin g methods in a carambola orchard compared to the typical irrigation practice for ca rambola, 2) evaluate weekly ch ange in root zone volumetric soil water content in response to ET based ir rigation scheduling, 3) evaluate carambola physiological responses (stem water potential, leaf gas exchange and yield) to ET-based irrigation scheduling and 4) compare ETo estimated at the study site using weather data measured onsite to ETo estimated for the same study site using w eather data from remote weather stations. Materials and Methods The study was conducted at the University of Florida Institute of Food and Agricultural Sciences Tropical Research and Education Center (UF/IFAS TREC) in Homestead, FL. The experiment was conducted in a 15 year old orchard of Arkin Carambola grafted on open pollinated Golden Star rootstock and planted in very gravely loam Krome soil of south Florida. Using a Geo XT 2005 series GPS, the coordinates at the center of the orchard were established as 25o30.420N and 80o30.492W. The experiment consisted of four treatments (T1 through T4) arranged in a completely randomized design ( CRD). Each treatment was replicated three times making a total of 12 experime ntal units (Figure 2-1). Each re plicate consisted of three trees spaced 4.5 m apart (Figure 2-2). Fe rtilizer application, pest management and weed control were uniformly applied to all the treatments base d on UF/IFAS-TREC Tropi cal Fruit Specialist recommendations (Crane, 1994). The trees were irrigated using micro sprinklers (Maxijets Special max-12 Fill-in 300o spray; Maxijet, Dundee, FL). The sprinklers had a flow rate of 89 L hr-1 and wetted diameter of 2 m at a pressure of 137 kPa. Irrigation contro l valves consisted of

PAGE 43

43 2.54 cm electric control valves. Irrigation schedu ling was based on real-time ET, historical ET and a typical irrigation practice for commercial carambola production of 76 mm per week (Lim, 1996; Crane, 2008, Personal communication). The study was conducted during the dry season (24 October 2008 to 30 April 2009). Irrigation Treatments Treatment one (T1) was based on a real-time irrigation sc hedule that was fully controlled through an automatic ET-based irrigation schedul ing device (Toro Intelli-Sense TIS 612, Toro Company Inc., Riverside, CA). The real -time ET-based treatment received daily ETo data via one way radio technology from a private weat her data service provi der (HydroPoint Data Systems Inc., Petaluma, CA). According to HydroPoint Data Systems the received ETo data were estimated using the American Society of Civil Engineers Environment and Water Resources Institute (ASCE-EWRI) standardized ETo equation (Equation 2-1). 2 21 273 408.0 UC eeU T C GR ETd as n n o (2-1) OET represents reference evapotranspiration [mm d-1], nR is net radiation at the crop surface [MJ m-2 d-1], G is soil heat flux density [MJ m-2 d-1], T is mean daily air temperature at 2 m height [C], 2U is wind speed at 2 m height [m s-1], se is saturation vapor pressure [kPa], ae is actual vapor pressure [kPa], asee is saturation vapor pressure deficit [kPa], is slope of vapor pressure-temperature curve [kPa C-1] and is psychrometric constant [kPa C-1]. In Florida, a short reference crop is commonly used in the estimation of ETo since alfalfa is not produced on a large scale (Irmak and Haman, 2003), therefor e considering a 24 hour time step constantsnCand dCare 900 and 0.34, respectively (Allen et al. 2005).

PAGE 44

44 After initial programming, the i rrigation scheduling device for T1 generated an approximate eight week irriga tion schedule that was adjusted daily based on weather data received and the site specific information progr ammed into the irrigation scheduling device. The site specific information entered into the irriga tion scheduling device included: soil type, plant type and crop coefficient, depth of the root zo ne, irrigation type and sy stem application rate, topography and sun or shade conditions (Table 2-1). The soil type options provided were limited and did not include an option that represented the study site soil (i .e. Krome very gravelly loam soil type). Among the options provided, sandy soil was the closest in terms of physical characteristics to the Krome soil and was used in this study. The crop co efficient was adjusted monthly based on plant phenology (Crane, Person al communication 2008). Daily and weekly average ETo data were manually recorded from the To ro irrigation controller daily between 1700 to 1800 hrs. The Toro real-time irrigation cont roller was not connected to a rain sensor. However, in case of a rainfall event the real-tim e ET-based treatment received a rain pause signal from HydroPoint Data Systems to pause irriga tion. The maximum rain pause duration was set at 24 hrs. T1 was restricted to irriga te between 0200 to 0500 hrs any day of the week. The total amount of water applied per treatment was meas ured using in DLJ water meters (Daniel L Jerman, Co.,Hackensack., NJ) and was recorded weekly (Figure 2-3) for all treatments. Treatment two (T2) was based on mean monthly historical ETo derived from 11 years of weather data from 1998 to 2008. The weather data were obtained from the Florida Automated Weather Network (FAWN) weather station located at TREC. Using mean monthly ETo provided by FAWN (which were estimated using the IFAS Penman method (Equation 2-2) (Jones et al. 1984)) and carambola crop coefficients ( Kc), the actual crop evapotranspiration (ETc) was calculated (Equation 2-3) and was used to derive irrigation runtimes for each month (Table 2-2).

PAGE 45

45 /42.042.108.056.0 14 so s d s oIFASR R e TR ET 20062.05.0(263.0 u eeda (2-2) cocKETET (2-3) oIFASET represents potential evapotranspiration from a vegetated surface [mm d-1], is slope of vapor pressure-temperature curve for air [mb C-1], is psychrometric constant 0.66 mb C-1, sR is total incoming solar radiation [cal cm-2 d-1], is Stefan-Boltzmann constant 11.71 x 10-8 cal cm-2 d-1K-1, T is mean daily air temperature [K], ae vapor pressure of air [mb], soR is total daily cloudless sky radiation [cal cm-2 d-1], is latent heat of vaporization [cal cm-1 mm-1] about 58 cal cm-2 at 29oC, 2uis wind speed at 2 m height [km d-1] and de vapor pressure at dew point temperature [mb] (for practical purposes in Florida dew point temperature is equal minimum temperature) (Jones et al. 1984). T2 irrigation was controlled us ing a typical irrigation timer (Rain Bird ESP-12 LXPlus, Rain Bird Inc., Azusa, CA); the timer was set to irrigate daily st arting at 0400 hrs. T2 was also subject to rainfall bypass as the controller was connected to a rain sensor (Toro Model 53769, Toro Company Inc., Riverside., CA) set to bypass irrigation if rainfall exceeded 6 mm. Figure 24 illustrates the installation of th e two irrigation controllers for T1 and T2 at the experimental site. Treatment three (T3) was a typical irriga tion practice for commerc ial carambola production in south Florida. T3 irrigation rate was 76 mm of water applied over 7 to 10 days (Crane 2008, personal communication). The irriga tion cycles and run times were controlled through the same irrigation timer as T2. Three irrigation cycles of two hr s duration each pe r week occurring on

PAGE 46

46 Monday, Wednesday and Friday represented a typi cal irrigation schedule. Daily irrigation run times were adjusted for rainfall events exceeding 6 mm using a rain sensor (Toro Model 53769, Toro Company Inc., Riverside, CA). The timer schedule for T3 was set to irrigate between 0500 to 1100 hrs. Treatment four (T4) was a non irrigated control treatment. Soil Moisture Measurement In each of the 12 experimental units, tensio meters (Irrometer Company, Inc., Riverside, CA) were installed to a depth of 15 cm at a distance of 91 cm along the North-South trench of the tree (Figure 2-5). Installation of the tensiometers involved making a hole slightly bigger in diameter compared to the tensiometer to a dept h of 19 cm. This was followed by preparation of slurry using sieved Krome soil and water. Half the mixture was poured to the hole before the tensiometer was inserted and the rest after the tensiometer was placed in the hole to ensure good contact between the ceramic cup and the soil (N uez-Elisea et al. 2001). The tensiometers were used to track root zone weekly soil water content throughout the st udy period (October 2008 to April 2009) and tensiometer readin gs were manually recorded ev ery Tuesday at 1700 hrs. To ensure efficient operation of the tensiometers in the field, accumulated air bubbles in the tensiometers were removed using a poking tube and a pressure pump on a weekly basis. Tension values ( cm H2O) were converted to volumetric soil water content ( ) using the van Genuchten equation (Equation 2-4) and curving f itting parameters provide d by Al-Yahyai et al. (2006). The parameter values included (21.48) n (1.09) and m (0.07) as well as saturated s (0.47) and residual water content r (0.01), values were developed by Al-Yahyai et al. (2006) using neutron probe measured volumetric soil wa ter content data at 20 cm for Krome soils under field conditions. rrs m n 1 (2-4)

PAGE 47

47 Measurement of Stem Water Potentia l Leaf Gas Exchange and Yield Stem water potential (s) was measured using a pressure chamber (Soil Moisture Equipment Corp., Santa Barbara, CA) by randomly selecting three recently matured terminal leaves from one tree in each experimental unit. The selected leaves while still on the tree were covered in an aluminum foil bag (Figure 2-6) fo r about 90 min to enable the stem water potential to attain equilibrium with leaf water potential (L ) (Shackel et al. 1997; Naor 2000; Al-Yahyai et al. 2005). Leaves were removed from the tr ee using razor blades. I mmediately after leaf removal, the leaves were placed in an ice box a nd taken to the laboratory for analysis (Figure 27). Measurements of stem water potentia l were done between 0600 and 0700 hours on a biweekly basis from November 2008 to March 2009. Net carbon dioxide assimilation (A), st omatal conductance of water vapor (gs) and transpiration (E) were measured bi-weekly between 1100 to 1400 hrs from November 2008 to April 2009 by randomly selecting three recently ma tured terminal leaves from one tree in each experimental unit. The number of leaves selected per tree was based on earlier studies by AlYahyai et al. (2005) on carambola physiological responses to soil moisture depletion in Krome soils. All leaf gas exchange measurements were made with a portable infrared gas analyzer (CIRAS II; PP Systems Ltd., Hert s, UK) (Figure 2-8). For leaf gas exchange measurements, photosynthetic photon flux (PPF) in the leaf cuvette was prov ided by a halogen lamp and maintained at 1,000 mol m-2 s-1 (Marler et al. 1 994). The reference CO2 in the chamber was 375 mol m-2 s-1 and had a flow rate of 200 ml min-1 for all leaf gas exchange measurements. Harvesting was conducted on 10 December 2008. Fruits on all the trees in each experimental unit were handpicked and placed in plastic buckets. A weighing scale was then

PAGE 48

48 used to measure the weight of the fruit harves ted from each tree (Figure. 2-9). Harvesting was conducted when majority of the fruit on all the trees had turned from green to golden yellow. Comparison of ETo Estimated Onsite Using Remote and Onsite Weather Data ETo estimated using weather data measured onsite was compared to ETo estimated for the same site using data from remote weather stati ons. Remote weather data refers to weather data collected at a location other than with in the experiment study site boundary. ETo from remote weather stations included: 1) ETo estimated using weather data fr om the closest public weather station which is the FAWN weather station located approximately 200 m from the study site and 2) ETo sent to the real-time ET-based irrigation scheduling device from Hydr oPoint data Systems (HydroPoint Data Systems Inc., Petaluma, CA), that uses weather data from a number of weather stations located close to the installation site of the irrigation scheduling device. The specific number of weather stations used by HydroPoint Data Systems to estimated ETo for the study area was not known to the author at th e time of the study. According to HydroPoint Data Systems, the ETo sent to the controller was estimated using the ASCE-EWRI standardized ETo equation (US Department of Interior and Bureau of Reclama tion, 2007) and delivered to the controller using paging technology. The ETo values generated using FAWN data were calculated using the ASCE-EWRI standardized equation and ETo obtained from FAWN website ( http://fawn.ifas.ufl.edu/ ) were estimated using the IFAS Penman equation (Jones et al. 1984). Onsite ETo was estimated using the ASCE-EWRI standardized ETo equation based on a short reference crop having a height of 0.12 m and a daily su rface resistance of 70 s m-1 (Allen et al. 2005). Four different ETo sources and estimation methods were considered: ETo data received by the real-time ET-based irrigation schedule (R1), ETo estimated using onsite measured weather data and the ASCE-EWRI standardized ETo (R2), ETo determined using FAWN data and the

PAGE 49

49 ASCE-EWRI standardized ETo equation (R3) and ETo determined using FAWN data and the IFAS Penman equation (R4). The IFAS Penman equation was in cluded as it is the method used to calculate ETo values posted on the FAWN Web site. Weather data collection Onsite weather data was collected using a weather station (HOBO weather station, Onset Computer Corporation., Pocasset, MA) located within the boundaries of the study area. Every effort was made to site the weather station at a location within th e orchard that best approximated the recommended reference conditio ns (Allen et al. 1998) (Figure 210). Data collected from the onsite weather station included: solar radiation, maximum a nd minimum air temperatures, relative humidity and wind speed. Weather data were collected at 15 min intervals from the onsite weather station and transformed to a daily average time series. Weather data was also collected from a FAWN weather station located at TREC-Homes tead within 200 m of the study site. At the FAWN station, all th e sensors were similar to thos e at the onsite weather station except the ultrasonic wind speed sensor that wa s installed at 10 m. Data collected from the FAWN weather station included: solar radiation, maximum a nd minimum air temperatures, relative humidity and wind speed. Daily averages of all the above weat her parameters were downloaded from the FAWN Web site ( http://fawn.ifas.ufl.edu/ ). Assessing integrity of weather data All the weather data collected were subjected to data quality control assessments proposed by Allen et al. (1998) and A llen et al. (2005). Quality c ontrol of solar radiation (Rs) was performed by plotting Rs and Rso (clear sky solar radiation) ov er time. For a well calibrated and properly operating pyranom eter, the computed Rso using Equations 2-5 (A llen et al. 1998) should plot as an upper envelope of th e measured daily solar radiation.

PAGE 50

50 a soRR 76.0 (2-5) Relative humidity data were assessed by plot ting computed dew point temperature against time (Allen et al. 1998). Dew point temperature should remain somewhat constant if the air mass is stable and advection of dry air from outside does not occur over a 24 hr period. In addition, daily minimum temperature was also plotted against time and compared to dew point temperature. For properly collected relative humidity data, the dew temperature and minimum temperature should be similar a majority of the time. Temperature data were assesse d by plotting daily average temp eratures computed from 15 min interval values and daily averages computed from daily maximum and minimum temperatures and plotting them over time. For good quality temperature data, the two should be almost the same with maximum differences of about 3oC, if there are no changes in air mass (Allen et al. 2005; Da vis et al. 2008). Assessment of the integrity of wind speed data is difficult but was achieved in this study by identifying consistently low wind speed recordings from measured wind speed data. Typically, consistent low wind speed recordings of 0.5 m s-1 or less would mean extremely calm conditions or faulty anemometer or ultrasoni c wind sensor (Allen et al. 2005). The best wind speed data should have average wind sp eed readings of greater than 1 m s-1 at a measurement height of 2 m. Data Analysis The data collected in this study included total irrigation water volume applied, daily ETo estimated onsite and from remote weather stations and tensiomete r reading for estimation of root zone soil water content. Daily water applicatio ns were used for sta tistical comparisons among treatments. SAS statistical software (SAS Institu te., Cary, NC) was used for all the statistical

PAGE 51

51 analysis using the General Lin ear Model (GLM). The confidence interval was assumed to be 95% and means separation was completed using the Duncan Multiple range test or WallerDuncan K-Ratio Test. Results and Discussion Results of Quantity of Irrigation Water Applied Evaluation of results for the quantity of irriga tion water applied indicated that an average of 9.9 mm of irrigation water wa s applied by the typical irriga tion practice per day which was significantly different (P<0.0001) from the quantit y of water applied by the ET based treatments (T1 and T2). The real-time ET-based treatment (T1) applied an average of 3.1 mm per day while the historical ET-based treatment (T2) applied 2.9 mm per day. However, there was no significant difference between the quanti ty of irrigation water applied by T1 and T2 treatments (Table 2-3). Results indicated that T2 had the greatest water saving (70%) compared to the typical irrigation practice. The real-time ET-based treatment (T1) applied 68% less water than the typical irrigation practice. The wa ter savings by ET-based treatments can be attributed to the fact that daily changes in we ather conditions through ETo were considered as opposed to T3 that applied water based on a set schedule. Cumulatively the typical ir rigation practice applied the greatest amount of water 1,892 mm, while T2 applied the least amount of water 560 mm (Table 2-3). The quantity of irrigation water applied by all the treatments was occasiona lly reduced when a scheduled irrigation would be bypassed as result of a signal from HydroPoint Data Systems or the rain sensor to bypass irrigation due to a precipita tion event of more than 6 mm in the case of T2 and T3. For T1, HydroPoint Data Systems remotely sent a signa l to pause irrigation for a maximum duration of 24 hrs if a rain event occurred.

PAGE 52

52 Results of Assessment of Root Zone Soil Water Content The average weekly root zone soil water content of T1 was significant different from that in treatments T2, T3 and T4 (P<0.0001) (Table 2-4). T1 maintained the highest average weekly soil water content at 29% (volumetric soil water c ontent). The average weekly soil water content maintained by the historical ET-base d irrigation scheduling treatment (T2) and the typical irrigation practice were not significantly different at the 95% level. T2 and T3 maintained volumetric soil water contents of 28.1% and 28.0 %, respectively. The non irrigated treatment (T4) maintained the lowest average weekly ro ot zone soil moisture content at 24%. Water retention characteristics of Krome cal careous very gravely loam soils can be explained in part by the soil part icle size distribution that is 51 % course particles and 49% loam particles (Al-Yahyai et al. 2006; Muoz-Carpena et al. 2009). The nature of particle size distribution in Krome soils results in rapid water depletion in the macropore fraction followed by slow water release from the micropore fraction. De spite the large volume of water applied by the typical irrigation practice (T3) compared to the ET-based treatments (T1 and T2), any excess water above the fill capacity of 34% volumetric wate r content (field capacity) could not be stored in the root zone and hence, did not substantially raise the average weekly root zone volumetric water content for T3. Despite the fact that the amount of irrigation applied by treatments T1 and T2 were not significantly different (95%) level, there were significant differences in average weekly root zone soil water content maintained by these two treatments. The original purpose for measuring soil suction was to assess how ET contro llers were able to maintain root zone soil water content. Results suggested that ET contro llers maintained root zone soil water content within an acceptable level. However, results also indicated that irrigati on scheduling time might have influenced results as the time between ir rigation and tensiometer reading collection for T3 was greater than that for the ET based treatments (T1 and T2). Thus, measuring soil water content

PAGE 53

53 on an daily time step could have improved accuracy of tracking changes in root zone soil water content in relationship to treatments. High frequency irriga tion cycles in T1, T2 and T3 contributed to the average soil suction being below 10 kPa during the study period. T4 had an average weekly volumetric soil water content of 24% at a suction of over 40 kPa. The small change in soil water content observed between T3 and T4 was associated with large changes in soil suction (Table 2-5) as expected considering the soil water characteristic curve for Krome soil. This characteristic of Krome soils is advantageous for low volume, high frequency ir rigation because for soil suction above 10 kPa, a small change in volumetric soil water conten t results in large reductions in soil suction. Effect of ET-Based Irrigation Sche duling on Stem Water Potential As shown in Figure 2-11, stem water potential (s ) remained above -1.0 MPa, averaging -0.8 MPa in all treatments which is greater than the lethal leaf water potential for carambola (-2.9 MPa) (Ismail et al. 1994). Although real-time a nd historical ET-based irrigation schedules resulted in a 68% and a 70% reduction in wate r volumes applied, respec tively, (Table 2-3) compared to the typical irrigation practice there were no significant differences in s (P < 0.05) among the ET-based treatments, typical irrigation practice or the non-irri gated treatment (Table 2-5). Particularly interesting to note was the fact that the non-ir rigated treatment did not receive irrigation for five months and only 61 mm of rainfall during the study period but had a water potential indicative of non stressed carambola. The average water potentia l recorded of -0.8 MPa among all treatments was similar to the -0.8 MP a that George et al (2002) reported for nonwater stressed carambola. In a previous study in the same orchard, Al-Yahyai et al. (2005) observed that in Krome very gravely loam so il at a soil water depletion of 0 to 30%, the s remained above -1.0 MP and concluded that th ere was no significant correlation between soil

PAGE 54

54 water depletion of 0 to 30% and stem water potential for carambola. They attributed their observations to frequent precipita tion and capillary rise due to the shallow water table. Although capillary contribution from the shallow groundwater table (2 m below the soil surface) could have partly contributed to the wa ter needs of the plant in the present study, we could not attribute the lack of significant differences in s among treatments to frequent precipitation because during the study it was very dry with only 2 pr ecipitation events over 10 mm. The two events occurred on 23 October 2008 and 17 March 2009 and where 28 and 10 mm, respectively. There were an additional 19 precipitation ev ents with amounts less than 7 mm each. In the present study, stem water potential observa tions could probably be attributed to the ability of carambola trees to regulate their water potentia l through osmotic adjustment by increasing their levels of pro line, increasing leaf abscissi on and reducing their stomatal conductance (gs) (Razi et al. 1992; Ismail et al. 1994, 1996). In this study, we also observed that trees reduced leaf surface area by changing leaf orientation from a horizontal to vertical position towards the end of the dry season (March to Ap ril); this was followed by rapid leaf abscission although the later could have been triggered by wind stress. Other re searchers have observed that carambola leaves begin to wilt when water supply to the plant is cut off for several days and leaf water potential reaches -2.0 MP (Salakpetch et al. 1990; Ismail et al. 1994; Al-Yahyai et al. 2005). Another explanation for our observations could be the low temp eratures that were experienced in south Florida during the study period that may have resulted in reduced transpiration and thus limited plant water uptak e. There were a total of 54 days with soil temperatures (at a depth of 10 cm) below 15oC and among these 5 days had temperatures below 10oC. George et al. (2002) obser ved that leaf water potential of 1-year-old carambola trees in

PAGE 55

55 containers became more negative for trees wi th root zone temperatures of 5 and 10oC compared to those with root zone te mperatures between 20 and 25oC. Effect of ET-Based Irrigation Scheduling on Transpiration (E) T2 produced the highest m ean E of 1.9 mmol m-2 s-1 followed by T3 (1.88 mmol m-2 s-1), T1 (1.87 mmol m-2 s-1) and T4 (1.71 mmol m-2 s-1) as shown in (Figure 2-12). E from treatment T2 was significantly different from E of treatment T4 while T1, T2 and T3 were not significant difference. There were also no signifi cant differences in E among treatments T1, T3 and T4 (Table 2-5). Al-Yahyai et al. (2005) observed simila r ranges 1.7 to 2 mmol m-2 s-1for E for carambola with leaf water potent ial higher than -1.0 MPa and conc luded that E decreased with decrease in sfollowing a curvilinear relationship. Razi et al. (1992) resu lts from a greenhouse carambola study suggest that E was only affected when s was below -0.85 MPa. The lack of significant differences in samong ET-based, typical irrigati on practice and non-irrigated treatments in this study suggest s that the ability of carambola trees to regulate their water potential enabled the trees not to experience substantial water stress during the study period. Effect of ET-Based Irrigation Sche duling on Stomatal Conductance (gs) Stomatal conductance to water vapor (gs) followed similar trends as E (Figure 2-13). T2 produced the highest mean gs of 63.85 mmol m-2 s-1 followed by T3 61.44 mmol m-2 s-1 and finally, T1 60.18 mmol m-2 s-1 and T4 56.38 mmol m-2 s-1, respectively. There was a significant difference in gs between T2 and T4, but no significant difference between treatments T1, T2 and T3. There were also no significant differences between T1, T3 and T4 (Table 2-5). However, the mean gs for all the treatments (at a s of -0.8 MPa) were lower than the average of gs of 80 mmol m-2 s-1 at the same water potential for caram bola under field conditions that was earlier observed by Al-Yahyai et al. (2005) for a carambola orchard in south Florida. This reduction in

PAGE 56

56 gs could probably explain carambola response to water deficits through an avoidance mechanism in which the tree minimizes water loss through increased stom atal control (Neuhas, 2003). In a similar study in Malaysia, Razi et al. (1992) observed that gs of young carambola plants decreased with decrease in leaf water potential. Besides irrigation water applied, the measured carambola physiological processes (stem water potential and leaf gas exchange) could have been more influenced by the trees ability to mitigate itself from the effects of water stress and the potential supply of soil water fr om the shallow groundwater table. Effect of ET-Based Irrigation Scheduling on Net CO2 Assimilation (A) There was no significant differe nt in A among treatments and the mean of all treatments was 4.7 mol m-2 s-1 (Table 2-5). The range of A values obt ained in this study were in agreement with earlier mean values (4.8 mol m-2 s-1) reported for field carambola at a leaf water potential of -0.8 MPa (Al-Yahyai et al. 2005 ). Despite lack of significant difference in A means among all treatments, there was considerable variability in A values within each treatment (Figure 2-14) probably due to inherent geneti c variability among trees. Razi et al. (1992) noted that A only began to decrease when s was less than -0.8 MPa. In a si milar study, Al-Yahyai et al. (2005) noted that A for potted carambola plants began to decrease when s was less than -1.0 MPa, the differences in s threshold values between these tw o studies could have been caused by differences in soil conditions. The lack of signi ficant differences in A am ong treatments in this present study could be attr ibuted to the fact that s was high throughout the study period. Results of Yield There were no significant di fferences in yield among all treatments (Table 2-5). T1 produced the highest average yield per tree of 157 kg tree-1, while the non irrigated treatment produced the lowest yield per tree of 74 kg tree-1. However, although there were no significant

PAGE 57

57 differences in yields among treatments, ther e was a lot of variability within and across treatments. This variability has been attributed to the inherent variabil ity in genetic potential between the trees and the variability in the golde n start rootstock that was used in establishing the orchard. Results of Weather Data Quality Control From the plots of solar radiation (Rs) and clear sky so lar radiation (Rso) against time, we observed that equations 2-5 predicated Rso fairly well throughout the study period. For the FAWN weather station, Rso plotted as an upper envelope of Rs from October to November 2008 but Rs did not fit the Rso curve from late November 2008 to early January 2009 (Figure 2-15). This was probably due to high cloudy cover during th is period that led to a reduced transmission index. From February to early March 2009, Rs values plotted with Rso values against time resulted in a good curve fit, probably due to in creased atmospheric clearance. Based on the above analysis in which plotted Rso was able to plot as an upper envelope of Rs, the radiation data measured at the FAWN weather station pass ed the quality assessment test to be used for computation of ETo. For the onsite weather station, equa tion 2-5 was used to predict the Rso curve that plotted as an upper envelope of Rs (Figure 2-16). However, for the entire study period Rs did not perfectly fit the Rso curve but followed the same pattern as the radiation recorded at the FAWN weather station. Slight ly lower recordings of measur ed solar radiation by the onsite weather station could be attribut ed to inadequate calibration of the pyranometer. In addition, the onsite pyranometer malfunctioned from 7 January 2009 to 13 January 2009. It was repaired by cleaning with de-ionized water and Kim wipes. In summary, since Rso plotted as an upper of Rs throughout the study period, the sola r radiation measured onsite passed the radiation quality assessment test.

PAGE 58

58 Relative humidity data were verified by calculating dew point temperature (Tdew) and plotting it against mini mum daily temperature (Tmin). For well measured relative humidity data, dew point temperature should be as close as possible to the minimum temperature. For the FAWN data, a plot of Tdew versus Tmin fit closely to the one-to-one line, especially between 10oC and 23oC (Figure 2-17). A similar trend was observed for data from the onsite weather station (Figure 2-18). Changes in Tdew could be attributed to change s in the air mass during the study period; south Florida expe rienced frequent cold front s during the study period. Temperature data were checked for integrity by plotting daily averages temperatures calculated as a mean over a 24 hr period agains t the mean of the daily maximum and minimum temperatures. These two means should be similar or the same for accurately measured temperature data as was evident in the collect ed datasets (Figures 2-19 and Figure 2-20). Wind speeds from both weather stations were assessed for integrity by plotting daily average wind speed and maximum wind spee d. The average wind speed from the FAWN weather station was consistently above 1 m s-1, averaging 2.2 m s-1 with an average maximum speed of 6.9 m s-1 (Figure 2-21). However, the onsite weat her station recorded consistently lower wind speed averaging 0.9 m/s with maximum wind speeds averaged 2.6 m s-1 (Figure 2-22). The wind speed measured by the onsite weather station did not pass the quality test and could not be used in the calculation of ETo. The low wind speeds recorded by the onsite weather station were due to the wind brakes planted around the experi mental site (i.e. carambola orchard) and the height of the canopy. To overcome this limitation, wind speed data at from the FAWN station located close to the study site was used in the estimation of ETo for (R2).

PAGE 59

59 Onsite Comparison of ETo From Different Sources Mean ETo for R1 of 2.7 mm day-1 during the study period (02 November 2008 to 03 March 2009) was significantly different from all other estimated ETo at the study site. Mean ETo for R2 during this period was 3.2 mm day-1 meaning R1under estimated R2 by 15%. However, ETo for R2 and ETo for R3 were not significantly different, onl y differing by 0.3%. This could probably be attributed to the relatively short distan ce between the study site and the FAWN weather station located at TREC. In a ddition, using wind speed data from the FAWN weather station at TREC to calculate ETo for R2 could also have improved ETo estimation for R2. The fact that R1, R2 and R3 were estimated using the same ETo estimation equation (ASCE-EWRI standardized ETo equation) could also have infl uenced the results. Mean daily ETo for R4 during the study period was 1.8 mm day-1 and was significantly different from all the other estimated ETo at the study site. ETo for R4 under estimate ETo for R2 by 43% and under estimated ETo for R3 by 40% despite the fact the source of weather data for R3 and R4 was the same. One of the reasons that could explain the large difference is R4 was estimated using a different ETo estimation equation (i.e. UF IFAS (1984) Penman). Conclusion Results from this study revealed that both si gnal based and historic al ET-based irrigation scheduling technologies can ade quately supply the water needs of the plant by replacing water which was lost through ET. In addition, the study revealed that substantial quantities of water were saved by the ET-based irrigation schedules co mpared to the typical irrigation practice (68 and 70% less irrigation water was applied by th e real-time and historical ET-based treatments respectively compared to the ty pical irrigation practice for co mmercial carambola production in south Florida). It is interesting to note that although the ET-ba sed treatments applied less water compared to the typical irriga tion practice, the real-time ET-based treatment maintained the

PAGE 60

60 highest average weekly root zone soil water content, while the soil water content in the historical ET-based treatment and the typical irrigation prac tice were not significantly different. However, all the treatments except the non ir rigated treatment were able to maintain soil water content in an optimum range (i.e. close to field capacity of 34% volumetric soil water content). Despite the differences in quantity of irrigation water applied by T1, T2 and T3, there were no significant differences insamong treatments. All treatments were measured to have a swhich was greater than that of water st ressed carambola (-2.9 MPa). T2 produced the highest E (1.9 mmol m-2 s-1) among all treatments and was significan tly different from E of treatment T4 but was not significantly different from E of T1and T3. Trends in gs were similar to transpiration trends with T2 having the highest gs (64 mmol m-2 s-1) while T4 had the lowest gs (56 mmol m-2 s-1). There were no significant differences in A among tr eatments. There were also no significant differences in yield among treatments. ETo from remote weather stations (i.e. ETo sent to the Toro controller) R1 underestimated onsite ETo for R2 by 25%. However, for purposes of this study wind speed data from the closest weather station to the study s ite was used in calculating ETo for R2 in order to eliminate the effect of canopy height and wind brakes plan ted around the orchard on wind speed. There were no significant differences in ETo of R2 and R3. ETo for R4 under estimated ETo for R2 and R3 by 43% and 40%, respectively. To further improve the performance of ET-based irrigation scheduling technologies, more work should be done to develop crop coefficients for commercial agriculture crops and improving a ccuracy of location specific ETo estimation. More research should be conducted to compare the performance of stand alone or onsite ET-based irrigation schedules to signal based ET-based irrigation schedules under agricu ltural settings. Overall realtime and historical ET-based irrigation scheduling tr eatments saved water, maintained root zone

PAGE 61

61 soil water content close to field capacity, did not appear to affect carambola physiological processes and ETo estimates from remote weather stations were generally repr esentative of the study area. Based on the results of this study, adopting this irrigati on scheduling technology could provide growers with water and energy savings contri buting to reduced costs of production. In addition, adoption of ET-based irrigation scheduling in tropical fruit production could lead to reduced nutrient leaching, hen ce conserving the quality of south Floridas groundwater resources.

PAGE 62

62 Table 2-1. Site specific parameters entered into the real-time irrigation scheduling controller Site specific parameters Input data Units Soil type Sandy N/A Precipitation rate 28 mm/hr Crop coefficient Varies N/A Slope % Factor 0-5% N/A Sprinkler Efficiency 95% N /A Root depth 15 mm Sprinkler type Spray Head N/A Microclimate sunny all day N/A % Usable rainfall 100% N/A Sprinkler location on slope All parts of slope N/A *Not applicable Table 2-2. Average monthly ETo and carambola crop coefficients us ed to calculate irrigation run times per cycle for the historical ET-based treatment Month Monthly average ETo (mm) Kc Irrigation runtime/ cycle (minutes) October 2.96 1.2 7.5 November 2.34 1.1 5.5 December 1.98 1.1 4.6 January 2.12 1.0 4.5 February 2.84 1.0 6.0 March 3.49 1.15 8.5 April 4.29 1.2 10.9 Table 2-3. Dry season (October to April) daily irrigation application and percentage water saving. Treatment Description Average irrigation water applied (mm/day) Cumulative irrigation water applied (mm) Irrigation water saving compared to T3 (%) T1 Real-time ET-based schedule 3.1b 589 68% T2 Historical ET-based schedule 2.9b 560 70% T3 Typical irrigation practice* 9.9a 1892 -Numbers with different letters in the third colu mn indicate differences at the 95% confidence level using the Waller-Duncan grouping.

PAGE 63

63 Table 2-4. Average weekly root zone soil water content (%) and suction (kPa) Treatment Description Average weekly soil water content (%) Average weekly soil suction (kPa) T1 Real-time ET-based schedule 29.24a 3 T2 Historical ET-based schedule 28.07b 5 T3 T4 Typical irrigation practice Non irrigated 27.96b 24.15c 6 40 Numbers with different letters in the third colu mn indicate differences at the 95% confidence level using the Duncan Multiple Range test grouping. Table 2-5. Carambola physiological parameters and yield data. Treatment (MPa) E (mmol m-2s-1) gs (mmol m-2s-1) A ( mol. m-2s-1) Yield (Kg/tree) T1 -0.8a 1.9ab 60.2ab 4.7a 157a T2 -0.8a 2.0a 63.9a 4.7a 102a T3 T4 -0.8a -0.8a 1.9ab 1.7b 61.4ab 56.4b 4.7a 4.7a 126a 74a T1 is real-time evapotranspiration (ET)-based irrigation scheduling. T2 is historical ET-based irrigation scheduling. T3 is typical irrigation practice for commerc ial carambola production in south Florida. T4 is non-irrigated treatment. is stem water potential. E is transpiration. gs is stomatal conductance to water vapor. A is net CO2 assimilation. Numbers with different letters in indicate differences at the 95% confidence level using the Duncan Multiple Range test grouping.

PAGE 64

64 Table 2-6. Average daily ETo estimated at the study site using different sources of weather data and equations Treatment Description No of observations Average daily ETo (mm/day) R1 Real-time ETo 119 2.7b R2 Onsite ETo 119 3.2a R3 R4 FAWN-ASCEEWRI ETo FAWN-UF IFAS Penman ETo 84 119 3.1a 1.8c R1 is ETo estimated using remote real-time weathe r data from Hydropoint Data Systems and ASCE-EWRI standardized equation. R2 is ETo estimated using weather data from a weat her installed at the study site and ASCEEWRI standardized equation. R3 is ETo estimated using weather data from a FAWN weather station at the Tropical Research and Education Center (TREC) Homestead, FL and ASCE-EWRI standardized equation. R4 is ETo estimated using weather data from a FAWN weather stati on at TREC and the University of Florida Institute of Food and Agricultural Sciences (UF-IFAS) (1984) Penman equation. ASCE-EWRI is American Societ y of Civil Engineers Enviro nmental and Water Resources Institute. FAWN is Florida Automated Weather Network. Numbers with different letters in the fourth co lumn indicate differences at the 95% confidence level using the Duncan Multiple Range test grouping.

PAGE 65

Figure 2 Figure 2 1. Experim e irrigation irrigation 2. Experim e carambol a e ntal plot la y schedule, T practice for e ntal unit o f a in Homest y out, where 2 is a histori carambola i f three cara m ead, FL. 65 T1 is real-ti m cal ETb ase i n south Flo r m bola trees s p m e Evapotr a d irrigation s r ida and T4 i p aced 4.5 m a nspiration ( E s chedule, T 3 i s a non irri g apart in an o E T)-based 3 is typical g ated treatm e o rchard of A e nt. A rkin

PAGE 66

66 Figure 2-3. DLJ Water meter Figure 2-4. Installation of a Toro TIS 615 and RainbirdESPcontrollers. Figure 2-6. Carambola leaf in aluminum foil bag. Figure 2-5. A 15 cm Irrometer tensiometer for soil water monitoring. Figure 2-8. CIRAS infrared analyzer for leaf gas exchange measurements. Figure 2-7. Pressure chamber for stem water potential measurements.

PAGE 67

67 Figure 2-9. Carambola harvesting and yield meas urement at an orchard in Homestead, FL. Figure 2-10. Downloading weather data from a HOB O weather station instal led at the study area.

PAGE 68

68 Dec Jan Feb Mar Apr Stem Water Potential ( (MPa) 0 2 4 6 8 10 12 T1 Dec Jan Feb Mar Apr Stem Water Potential ( (MPa) 0 2 4 6 8 10 12 T2 Dec Jan Feb Mar Apr Stem Water Potential ( (MPa) 0 2 4 6 8 10 12 T3 Dec Jan Feb Mar Apr Stem Water Potential ( (MPa) 0 2 4 6 8 10 12 T4 Figure 2-11. Stem water potential ( ) (MPa) of carambola trees under different irrigation scheduling during the study period October 2008 to April 2009. A) T1 is real-time ET-based irrigation schedule, B) T2 is historical ET-based irrigation schedule, B) T3 is typical irrigation practice and D) T4 is a non irrigated treatment. A B C D

PAGE 69

69 Nov Dec Jan Feb Mar Apr Transpiration (E) (mmol m-2s-1) 1.2 1.4 1.6 1.8 2.0 2.2 2.4 T1 Nov Dec Jan Feb Mar Apr Transpiration (E) (mmol m-2s-1) 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 T2 Nov Dec Jan Feb Mar Apr Transpiration (E) (mmol m-2s-1) 1.2 1.4 1.6 1.8 2.0 2.2 2.4 T3 Nov Dec Jan Feb Mar Apr Transpiration (E) (mmol m-2s-1) 0.8 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 T4 Figure 2-12. Transpiration (E) (mmol m-2s-1) of carambola trees under different irrigation scheduling during the study period October 2008 to April 2009. A) T1 is real-time ET-based irrigation schedule, B) T2 is historical ET-based irrigation schedule, C) T3 is typical irrigation practice and D) T4 is anon irrigated treatment. A C B D

PAGE 70

70 Nov Dec Jan Feb Mar Apr Stomatal conductance (gs) (mmol m-2s-1) 30 40 50 60 70 80 90 T1 Nov Dec Jan Feb Mar Apr Stomatal conductance (gs) (mmol m-2s-1) 40 45 50 55 60 65 70 75 T2 Nov Dec Jan Feb Mar Apr Stomatal conductance (gs) (mmol m-2s-1) 30 40 50 60 70 80 90 T3 Nov Dec Jan Feb Mar Apr Stomatal conductance (gs) (mmol m-2s-1) 20 40 60 80 100 120 T4 Figure 2-13. Stomatal conductance (gs) (mmol m-2s-1) of carambola trees under different irrigation scheduling during the study period October 2008 to April 2009. A) T1 is real-time ET-based irrigation schedule, B) T2 is historical ET-based irrigation schedule, C) T3 is typical irrigation practice and D) T4 is a non-irrigated treatment. A C B D

PAGE 71

71 Nov Dec Jan Feb Mar Apr Net CO2 assimilation (A) mol m-2s-1) 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 T1 Nov Dec Jan Feb Mar Apr Net CO2 assimilation (A) ( mol m-2s-1) 3.0 3.5 4.0 4.5 5.0 5.5 6.0 6.5 7.0 T2 Nov Dec Jan Feb Mar Apr Net CO2 assimilation (A) ( mol m-2s-1) 2 3 4 5 6 7 8 T3 Nov Dec Jan Feb Mar Apr Net CO2 assimilation (A) ( mol m-2s-1) 1 2 3 4 5 6 7 8 9 T4 Figure 2-14. Net CO2 assimilation (A) ( mol m-2s-1) of carambola trees unde r different irrigation scheduling during the study period October 2008 to April 2009. A) T1 is real-time ET-based irrigation schedule, B) T2 is historical ET-based irrigation schedule, C) T3 is typical irrigation practice and D) T4 is a non-irrigated treatment. B C A D

PAGE 72

72 Figure 2-15. Measured solar radiation (Rs) reported by the Florida Automated Weather Network (FAWN) and calculated clear sky solar radiation (Rso) data for Homestead, FL against time. Figure 2-16. Measured solar radiation (Rs) and calculated clear sky solar radiation (Rso) for the onsite weather station plotted against time 0 5 10 15 20 25 29-Sep-088-Nov-0818-Dec-0827-Jan-098-Mar-09Radiation (MJ/m2day) Rso Rs 0.76 Ra 0 5 10 15 20 25 19-Oct-0828-Nov-087-Jan-0916-Feb-09Radiation (MJ/m2/day) Rs Rso 0.76 Ra

PAGE 73

73 Figure 2-17. Minimum temperature (Tmin) versus calculated de w point temperature (Tdew) reported by the Florida Automated Weathe r Network (FAWN) for Homestead, FL compared to the one-to-one line. Figure 2-18. Minimum temperature (Tmin) versus calculated de w point temperature (Tdew) collected from the onsite weather stat ion compared to the one-to-one line. R = 0.8486 1 = 1 0 5 10 15 20 25 30 35 40 45 01020304050Minimum temperature (oC)Dew point temperature (oC) R = 0.72671 =1 0 5 10 15 20 25 30 35 05101520253035Minimum temperature (oC)Dew temperature (oC)

PAGE 74

74 Figure 2-19. Daily mean of maximum and mini mum temperature plotted against daily mean temperature for Florida Automated Weathe r Network (FAWN) data at Homestead, FL Figure 2-20. Daily mean of maximum and mini mum temperature plotted against daily mean temperature for the onsite weather stati on compared to the one-to-one line. R = 0.9576 1 = 1 0 5 10 15 20 25 30 35 05101520253035Daily Mean of Max and Min Tempearature (oC)Average Temperature over 24 hrs (oC) R = 0.9365 1 = 1 0 5 10 15 20 25 30 35 05101520253035Daily Mean of Max and Min Tempearture (oC)Average Temperature over 24 hrs (oC)

PAGE 75

75 Figure 2-21. Daily maximum and average wind speeds for Florida Automated Weather Network (FAWN) data at Homestead, FL plotted against time. Figure 2-22. Daily maximum and average wind speed s collected from the onsite weather station. 0 2 4 6 8 10 12 14 16 29-Sep-088-Nov-0818-Dec-0827-Jan-098-Mar-0917-Apr-09Wind speed2 m (m/s) Avg wind speed Max wind speed 0 2 4 6 8 10 12 14 16 29-Sep-088-Nov-0818-Dec-0827-Jan-098-Mar-09Windspeed 2m (m/s) Max wind speed Avg wind speed

PAGE 76

76 CHAPTER 3 COMPARISON OF VARIOUS REFERENCE EVAPOTRANSPIRATION ESTIMATION EQUATIONS AND SPATIAL INTERPOLATION TECHNIQUES IN SOUTH FLORIDA Introduction The success of any evapotranspiration based (E T-based) irrigation management plan relies to a large extent on the ability to accurately estimate reference ET (ETo). ETo is defined as the rate of evapotranspiration from a specified vege tative or reference surfac e. The reference surface is a hypothetical grass or alfalf a reference crop that resembles an extensive well watered green grass surface with uniform height, actively growi ng and completely shading the ground (Allen et al. 1998). Due to the complexities associ ated with direct measurement of ETo, several temperature-based, radiation-based and combinati on-type equations have been formulated for estimation of ETo. Several studies have been conducte d to identify the most suitable ETo estimation equations for Florida climatic conditions (Stephens and Stewart 1963; Abtew et al. 1996a; Jacobs and Satti 2001). Stephens and Stew art (1963) compared ni ne potential ET (ETP) estimation equations to lysimeter data from St. Augustine grass under south Florida climatic conditions and noted that combination based me thods like the Penman and modified Penman were most accurate. Abtew et al. 1996a using thr ee years of lysimeter data noted that radiation based equations were the most suitable simpler ETo estimation equations for Florida. Jacobs and Satti (2001) evaluated perf ormance of several ETo estimation equations against the American Society of Civil Engineers (ASCE)-Penman Monte ith (PM) for north central Florida and noted that combination based methods like modified Penman and FAO-56 PM gave the best results. Although combination based equations have prov ed their ability to accurately estimate ETo under different climatic conditions, their wide scale ut ilization has been limited due to the large number of input data required for their evaluation (Katul et al. 1992; Allen et al. 1998; Wright et al. 2000; Itenfisu et al. 2003). Altern atively, estimation accuracy of ETo equations that use a limited

PAGE 77

77 set of weather parameters varies with climatic conditions (Allen et al. 1998; Nandagiri and Kovor 2006). Recent advances in irrigation technology ha ve led to the development of ET-based irrigation scheduling controllers ma inly designed for landscape ir rigation applications but with potential to save water in an ag ricultural setting. Res earch conducted in north Florida and Orange County in California have indicated that ET-based irrigation sc heduling controllers can reduce irrigation by up to 40% (Hunt et al. 2001; Da vis et al. 2007). Similarly, an ongoing study in South Florida to evaluate ET-based irrigati on technology in a tropica l fruit orchard has demonstrated that substantial amounts of irriga tion water can be saved using this technology. ET-based irrigation scheduling methods are classified based on the way weather data is used in estimating the ETo is received. There are generally three types of ET-based irrigation scheduling controllers: 1) signal based ET, 2) historical based ET and 3) onsite measurement or stand alone ET-based irrigation scheduling. However, there remains concerns a bout the ability of the signal based technology to estimate site specific ETo for agriculture applications at locations where ETo is generated with sparsely distributed weathe r station data. Onsite or standalone ET-based irrigation scheduling controllers partly address this problem but the simpler ETo estimation equations used in their irrigation scheduling algorithms need local validation. Thus, there is a need to id entify the most appropriate ETo estimation equations for purposes of improving performance of onsite ET-ba sed irrigation scheduling controllers and for improving ETo estimation from weather stations having li mited availability of weather data. The objectives of this study were to: 1) Evaluate and compare five ETo estimation methods that cover combination based (University of Florida Institute of Food and Agricultural Sciences (UF IFAS) 1984 Penman), radiation based (South Florida Water Management Simple Method (SFWMD-

PAGE 78

78 SM), Turc (1961) and Priestley Taylor) and temperature based equations (Hargreaves) to the ASCE-EWRI (Environmental and Water Resources Institute) standardized method considering daily and monthly time steps and 2) Evaluate two spatial interpolati on methods for developing spatially distributed ETo surfaces for south Florida.

PAGE 79

79 Materials and Methods This study was conducted in Miami-Dade and Broward Counties in southeastern Florida. This area is characterized by a sub humid climate. Five weather stations were selected as having adequate data for the analyses conducted in this study (Figure 3-1, Table 31). The stations were selected based on availability of complete time series datasets for solar radiation, temperature, wind speed and relative humidity. A nother criteria used in the sele ction of weather stations was their relative distances from the coastline in order to have an even spatial distribution of the monitoring networks. Weather data were obtai ned from the SFWMD environmental database that stores historical and real -time hydrologic, meteorologic, hydrogeologic and water quality data for south Florida called DBHydro database (http://my.sfwmd.gov/dbhydroplsql/show_dbkey_inf o.main_menu) and from the Florida Automated Weather Network (FAWN) Web site (h ttp://fawn.ifas.ufl.edu/). Although the weather stations recorded solar radiat ion, temperature, wind speed and relative humidity at 15 minute intervals, only daily averages were used in this study. Air temperature at SFWMD weather station si tes was measured using Campbell Scientific Models 107 and 108 (Campbell Scie ntific Inc., Logan, UT) temp erature probes. Relative humidity was measured using the HMP45C Temperature and Relative Humidity probe (Campbell Scientific Inc., Logan, UT) that contai ns a Platinum Resistance Temperature (PRT) detector and a Vaisala HUMICAP 180 capacitive relative humidity sensor. The HMP45C has a measurement range of 0-100% and accuracy of up to 2-3 at 20oC. Barometric pressure at the SFWMD sites was measured using PTA-427 Ba rometric Pressure Transducer with a measurement range of 80-106 kPa and accuracy of .05 kPa. Solar radiation was measured using a LI-COR LI200S pyranometer (LI-COR, Inc., Lincoln, NE) with a t ypical sensitivity of 0.2 kW m-2 and absolute error or accuracy of up to % under natural day light conditions and

PAGE 80

80 wind speed was measured using Vaisala WS425 u ltrasonic wind sensor (Vaisala Inc., Boston, MA) with a measurement range of 0-64 m s-1 and accuracy of up to depending on the wind speed (Pathak 2008). Air temperature was measured at the FAWN weather stations using a CS107 sensor thermistor (Campbell Scientific Inc., Logan, UT ) installed at 0.6 m above the soil surface that has a measurement range of -24o to 48oC and overall accuracy of 0.4. Relative humidity (RH) was detected using a CSL Temperature and Rela tive Humidity Probe (Campbell Scientific Inc., Logan, UT) installed at 2 m, with a measuremen t range of 0-100% and accuracy of up to 0.4 depending on temperature. The wind speed was measured using Vaisala WS425 ultrasonic wind sensor (Vaisala Inc., Boston, MA) wi th a measurement range of 0-64 m s-1 and accuracy of up to depending on the wind speed. Solar radiati on was measured using an arm mounted LI-COR LI200S pyranometer (LI-COR, Inc., Lincoln, NE ) with a typical sens itivity of 0.2 kW m-2 and absolute error or accuracy of up to 3% under natural day light conditions. Barometric pressure, another important atmospheric parameter in estimation of ETo, was measured using PT101B barometric pressure sensor (Vai sala Inc., Boston, MA) (FAWN, 2009). The data available for the selected five weathe r stations was reduced to consider data with common time spans, thus the data series used in this analysis was fr om January 2001 to March 2009. Integrity checks were performed using proce dures described by Allen et al. (1998) on all the weather parameters (i.e., solar radiation, temp erature, wind speed and relative humidity) from each weather station. Further scr eening included eliminating all th e days that had missing solar radiation and temperature data. ETo Estimation Equations The ETo estimation equations included in this comparative analysis are the UF IFAS (1984) Penman, the SFWMD-SM, the PriestleyTaylor, the Turc and the Hargreaves. The

PAGE 81

81 ASCE-EWRI standardized ETo was used as the bench mark fo r comparison in this study given its physical basis and ability to accurately estimate ETo under a wide range of climatic conditions as reported by earlier investigator s (Wright et al. 2000; Itenfisu et al. 2003; Allen et al. 2005; Kovor 2006). Computational procedures and suppor ting equations for each of these methods are outlined in Equations 3-1 to Equation 3-6. For all ETo calculations the reference crop was assumed to be a short green grass with surface resistance of 70 s m-1 (Allen et al., 2000) and all the computations were performed using Microsoft Excel. ASCE-EWRI standardized ETo equation The ASCE-EWRI standardized ETo equation is based on the ASCE PM, the standardization simplified the ASCE PM by cons idering two reference surfaces: a short grass and a tall reference surf ace (alfalfa). The constants in th e equation depend on the type of reference surface, calculation time step and the time of day (ni ght or day) (Allen et al. 2005). 2 21 273 408.0 UC eeU T C GR ETd as n n o (3-1) OET represents reference evapotranspiration [mm d-1],nR is net radiation at the crop surface [MJ m-2 d-1], G is soil heat flux density [MJ m-2 d-1], T is mean daily air temperature at 2 m height [C], 2U is wind speed at 2 m height [m s-1], se is saturation vapor pressure [kPa], ae is actual vapor pressure [kPa], asee is saturation vapor pressure deficit [kPa], is slope of vapor pressure-temperature curve [kPa C-1] and is psychrometric constant [kPa C-1]. In Florida, a short reference crop is commonly used in the estimation of ETo since alfalfa is not produced on a large scale (Irmak and Haman 2003). Therefore, considering a 24 hour time step constantsnCand dCwill take on values of 900 and 0.34, respectively (Allen et al. 2005).

PAGE 82

82 UF IFAS (1984) Penman The UF IFAS (1984) Penman equation is a combination based method and the four climatic input parameters needed are: 1) net ra diation, 2) air temperatur e, 3) wind speed and 4) vapor pressure deficit. The UF IFAS (1984) Penm an equation is expressed as (Jones et al. 1984): /42.042.108.056.0 14 so s d s oIFASR R e TR ET 20062.05.0(263.0 u eeda (3-2) oIFASET represents potential evapotranspiration from a vegetated surface [mm d-1], is slope of vapor pressure-temperature curve for air [mb C-1], is psychrometric constant (0.66 mb C-1), sR is total incoming solar radiation [cal cm-2 d-1], is Stefan-Boltzmann constant (11.71 x 10-8 cal cm-2 d-1 K-1), T is mean daily air temperature [K], ae vapor pressure of air [mb], soR is total daily cloudless sky radiation [cal cm-2 d-1], is latent heat of vaporization (58 cal cm-2 at 29oC), 2uis wind speed at 2 m height [km d-1] and de vapor pressure at dew poi nt temperature [mb] (for practical purposes in Florida dew point temperatur e is equal minimum temperature) (Jones et al. 1984). SFWMD-SM A simple ETp estimation equation for south Florid a was developed at the SFWMD and calibrated using three year (1993 to 1995) of lysimeter data from wetland plants under inundated condition. This equation only requires solar radi ation as its input and is mathematically expressed as (Abtew et al. 1995): s pR KET1 (3-3)

PAGE 83

83 ETp represents potential evapotranspira tion from a vegetated surface [mm d-1], sR is total incoming solar radiation [MJ m-2 d-1], is latent heat of vaporization [MJ Kg-1] and 1K is a coefficient (0.53). Priestley-Taylor The Priestly Taylor equation (E quation 3-4) is a simplification of the Penman combination equation with only radiation as the main input. The equation is valid under minimum advection conditions and it is used in areas of low moistu re stress. The evapotra nspiration under potential conditions from a vegetated surface is expr essed as (Priestley and Taylor 1972): GR ETn P (3-4) ETP represents potential evapotranspira tion from a vegetated surface [mm d-1], is slope of vapor pressure temperature curve for air [kPa C-1], is psychrometric constant [kPa C-1], is a constant determined empirically and for humid areas it equal to 1.26, nR net radiation [MJ m2 d-1], G is soil heat flux [MJ m-2 d-1] and is latent heat of vaporization [MJ kg-1]. The net radiation, slope of the vapor pressure-temperatu re curve, latent heat of vaporization and psychrometric constant are evaluated using the ASCE-EWRI standardized ETo equations procedures and the soil heat flux is assumed to be 0 for daily ca lculation time steps based on the recommendations of Allen et al. (2005). Turc (1961) Equation The Turc (1961) method for estimating ETP is based on solar radia tion and temperatures as the main inputs. Several studies have shown that Turc provides similar results as lysimeter data and combination equations like the Penman and Pe nman Monteith (Jesen et al. 1990; Abtew et al. 1995; Jacobs and Satti 2001; Nandagiri and Kovoor 2006; Trajovi and Stojni 2007). The

PAGE 84

84 Turc equation for sub humid climate with relative humidity greater than 50% is expressed as (Turc 1961): )500239.0( 15 013.0 s pR T T ET (3-5) ETP represents potential evapotranspira tion from a vegetated surface [mm d-1], T is daily mean air temperature [C] and sR daily solar radiation [cal cm-2 d-1]. Hargreaves and Samani (1985) Equation The Hargreaves and Samani (1985) equation de veloped for dry areas in the western United States is the only temperature ba sed equation that has produced resu lts with some global validity, many other temperature based equations require local calibration (Allen et al. 1998). The equation is expressed as (H argreaves and Samani 1985): 8.17 0023.05.0 min max TTTR ETa P (3-6) ETp represents potential evapotranspira tion from a vegetated surface [mm d-1], T is daily mean air temperature [C], MinT is minimum daily air temperature at 2 m height [C], MaxT is maximum daily air temperatur e at 2 m height [C] andaR is extraterrestrial solar radiation [MJ m2 d-1]. Comparison Statistics For each of the weather stati ons, the difference between ETo estimated by the ASCEEWRI standardized ETo equation and each of the other five methods was evaluated using the Standard Error of Estimate (SEE) statistic, th e standard deviation of the ratio of the ETo estimated by any ETo estimation equation to the ETo estimated by the ASCE-EWRI standardized method (S), the coefficient of determination (R2) and the Root Mean Square Difference (RMSD). ASCE-EWRI standardized ETo values were considered the measured or bench mark values and

PAGE 85

85 estimates from each of the other five equations were considered the model predicted values (Nandagri and Kovooor 2006; Temesgen et al., 2005). Statistical comparisons were made separately for daily and monthly time steps. The standard error of the estimate is a meas ure of the accuracy of predictions made with the assistance of a regression line. It estimates the square root of the average of the squared deviations from the regression line. 2 )(2 n PO SEEn i ii (3-7) A good model is one that results in a low value of the SEE between the observed quantity (O) and the predicated quantity (P).Where n represents the total number of ob servations. A ratio of ETo estimated by each of the five ETo equations to the ETo estimated by ASCE-EWRI standardized equation was calcula ted. The standard deviation S (Equation 3-8) of the ratio for each ETo estimation equation was used as a comparative statistic. 2 n YY Sn i i (3-8) Where S is sample standard deviation, iY represents each data point, Y is the value of the mean and n is the sample size. The coefficient of determination (R2) Equation (3-9) was used as a measure of the reliability of the linear relationship between observed (O) and predicted (P) values. Values of R2 range between 0 to 1, values close to 1 indicate excellent reliabi lity while values close to 0 indicate poor reliability of the linear relationship.

PAGE 86

86 ])()(][)()([ )(2 2 2 2 2 i i i i ii iiPPnOOn POPOn R (3-9) The RMSD (Equation 3-10) was used to measure the accuracy of each ETo estimation equation in predicating the ETo values estimated by the ASCE -EWRI standardized equation. 5.0 2 n PO RMSDn i ii (3-10) The ETo estimation equations were ranked on the ba sis of the SEE, Standard deviation of the ratio of ETo estimated by each method to the ETo estimated by the ASCE-EWRI standardized method, R2and the RMSD. In order to determine the best method, an overall rank was calculated as the average of the ranks from the above four performance statistics since each statistic represented a different as pect of the equation. Spatial Interpolation of ETo Data collected at weather stations provide poi nt estimations of weat her conditions, in order to determine weather conditions at other locations between weather statio ns data interpolation procedures are needed. ETo surfaces were generated using the interpolation to raster sub routine in ESRI ArcGIS and a geo-referenced event laye r that was created using point weather data in ArcGIS. The two approaches used in the interpol ation were (1) the Inverse Distance Weighted Average (IDWA) and (2) the spline method. The IDWA method estimates weather parameters for unknown locations using a linear combination of values from the sampled locations. Weighting of sampled locations is only a func tion of distance from the un-sampled location. The underlying assumption in this method is that the values closest to the un-sampled location are more representative of the un-sampled location (Stallings et al. 1992; Hartkamp et al. 1999).

PAGE 87

87 Spline interpolation, on th e other hand is a form of data in terpolation in which the function performing the interpolation is a piecewise polynomial called a spline. The spline function generates a smooth and continuous curve across all the sampled data points. Splines are preferred over ordinary polynomials as interpolants because they minimize interpolation errors by staying close to the range of the sample d data (McKinley and Levine, 1998). Data Analysis For comparison of ETo estimations on a daily and monthly time steps, all the data from the five weather stations were us ed in the estimations of ETo by the various equations. However, for the evaluation of spatial interpol ation techniques, only data from f our weather stations were used to generate the ETo surfaces. Weather data from the FAWN weather station at TREC Homestead were set aside for validation of the ETo surfaces.

PAGE 88

88 Results and Discussion Comparison of daily ETo Estimations of the Various Equations Mean daily ETo estimates calculated by averaging result s over the entire period of record for each weather station are provided in Ta ble 3-2. The UF IFAS (1984) Penman equation produced the highest estimated mean daily ETo at all the five weather stations followed by the Hargreaves equation. On average the UF IFAS (1984) Penman equation over estimated daily ETo at all weather stations by an average of 23%. Walter et al. (2 001) also observed that the 1964 Penman ETo equation over estimated ETo calculated with the ASCE-PM by an average of 7% while the FAO-24 version of the Penman ETo equation over estimated ASCE-PM ETo by an average of 20%. The mean daily values for the other equations were similar to those of the ASCE-EWRI ETo standardized equation. Results of the UF IFAS (1984) Penman equation are of special interest because this e quation is still popular in Florid a. In comparison to the ASCEEWRI standardized equation, the UF IFAS (1984) Penman equation over estimated ETo as shown by the high RMSD values (Tables 3-3), but the equation was able to follow the same pattern as the ASCE-EWRI standardized e quation as shown by the high values of R2 (Table 3-3 and Figure 3-2 to Figure 3-6), this probably is du e to the fact that this method accounts for wind speed which the other temperature and radiation based equations did not. Based on the overall rank calculated as the average of the ranks from the SEE, S, R2 and the RMSD (Table 3-3 and Figure 3-2 to Figure 36) the overall performance of the UF IFAS (1984) equation was satisfactory. UF IFAS ( 1984) Penman ranked second overall at the Homestead and G3ASWX weather st ations and third overall at th e Fort Lauderdale, S331W and JBTS weather stations. Jacobs and Satti (2001) ob served the same results when they compared performance of several combination, radiation and temperature based ETo estimation equations in North and Central Florida.

PAGE 89

89 At the Homestead site, the highest ranking equation was Turc followed by the UF IFAS (1984) Penman and the SFWMD-SM both with an overall rank of 2.5 followed by Priestley Taylor and the Hargreaves equations in the fourth and fifth positions, respectively. Hargreaves had the lowest ranking at this location due to the fact that it consistently over estimated ETo. This is likely the result of the incl usion of only temperature and extr aterrestrial radiation by the Hargreaves equation. The exclusion of solar radiation is critical since it has been shown to be a driving factor of ET processes in humid and sub-hum id climates (Allen et al. 1998). At the other locations, the Turc (1961) equation produced the highest overall rank while Hargreaves had the lowest ranking. Despite the simplicity and mini mum data requirements of the Turc (1961) equation, its daily ETo estimates seem to correlate well with those of the more computationally intensive ASCE-EWRI standardized ETo equation. Results of this study re-enforces the work of earlier investigators who observe d that the Turc method produced similar results as lysimeter data and combination based equa tions like the Penman and Penman Monteith in humid climates (Jesen et al. 1990; Abtew et al. 1995; J acobs and Satti 2001; Nandagiri and Kovoor 2006; Trajovi and Stojni 2007). The performance of the Turc (1961) method could be explained by the basis of its formulation, as it was develope d for western European conditions that closely resemble those of south Florida (i.e. rela tive humidity greater than 50%) (Turc 1961) and accounts for radiation and temperature which are the main parameters influencing ET processes in sub-humid climates. These findings are particul arly interesting in th at if the Turc (1961) equation is incorporated into the irrigation scheduling algorithms of standalone ET-based irrigation scheduling devices that depend on mainly solar radiation and temperature sensors to adjust their daily ETo estimates, this could pr obably result in increased scheduling efficiency of these irrigation scheduling devices in sub humid climates.

PAGE 90

90 Another ETo estimation equation that consistent ly produced satisfactory results in comparison to the ASCE-EWRI standardized ETo equation at all the five sites was the Priestley Taylor, which had the second best overall rank at the all the sites apart from the Homestead site. However, although Priestley Taylor produced the second best overall ranking, the equation performed well in the wint er but over estimated ETo in the summer. As shown in the scatter plots Figure 3-2 to Figure 3-6, Prie stley Taylor tended to overest imate ASCE-EWRI standardized equation at higher values of ETo. The SFWMD-SM gave generally satisfactory results but its performance could be improved by introduci ng a temperature term in the equation. Comparison of Mean Monthly ETo Estimates for the Various Equations Graphical illustrations of the differences in mean monthly ETo estimated by the five methods are shown in Figure 3-7 to Figure 3-11. Seasonal variations in the performance of the methods are evident. It is also clear that UF IFAS (1984) Pe nman method over estimates mean monthly ETo consistently throughout the year although it fo llows a similar pattern to that of the ASCE-EWRI standardized method at all the weat her station sites. The Hargreaves equation consistently overestimated ASCE-EWRI standa rdized equation estimates during the summer months except for the JBTS station where it under estimated ASCE-EWR I standardized method estimates throughout the year. Th is could probably have been caused by lower temperatures experienced at the JBTS site due to its close proximity to the coastline. The Priestly-Taylor method also tended to overestimate ASCE-EWRI standardized method esti mates especially in the summer months. Turc and SFWMD-SM method ETo estimates are slightly lower than ASCE standardized ETo estimates in summer but their results correlated well in wi nter months. This probably is due to the fact that Turc does not account for wind speed. Trajovi and Stojni (2007) demonstrated that addi ng a wind correction f actor to the Turc equation improved its prediction of ETo estimated by the FAO-56 PM. While under estimation of ETo in summer by

PAGE 91

91 the SFWMD-SM could be attributed to the fact the equation does not account for temperature which is a strong driving force for ET processe s in summer. Detailed performance comparisons for mean monthly ETo estimations by the five methods relative to the ASCE-EWRI standardized equation estimates are summarized in Table 3-4. The difference between the ETo estimates by the various methods and the ASCE standardized method ETo estimates was reduced for the monthly time step compared to the daily time step. In addition, monthly time step evalua tions indicated that th ere was no considerable variation in the performance of methods at the different locati ons. The Turc (1961) method still resulted in the highest overall rank at all the lo cations except for the JBTS station were the UF IFAS (1984) method had the highest overall rank (Table 3-4). All the methods produced coefficients of determination of greater than 0.9 except for the Hargreaves method at Homestead and G3ASWX and UF IFAS (1984) at Fort Lauderdale. All the equations yielded lower than 0.35 mm day-1 SEE at all the weather sites indicating an improvement in ETo estimation from daily ETo estimates. This improvement is probably due to a reduction in vari ability of the input weather variables due to the averagin g process. This improvement in ETo estimation by considering longer time steps was ob served by other investigators (J esen et al. 1990; Abtew et al. 1996a; Jacobs and Satti 2001). Based on the performa nce results of monthly time steps, simpler methods like SFWMD-SM, Priestley-Taylor and Turc can be used to estimate ETo on a monthly or annual time steps in south Florida. Results of Spatial Interpolation of ETo ETo estimated by the ASCE-EWRI standardized method from the four weather stations (excluding the FAWN Homestead weather station used for validation purposes) was spatially interpolated using GIS-base d Inverse Distance Weighted Average (IDWA) and spline interpolation techniques Mean monthly daily ETo surfaces were generated (Figures 3-12 to

PAGE 92

92 Figure 3-35). Spatially distributed ETo results indicate an ETo trend that increased with distance from the coastline. Jones et al. (1984) noted that in Florida areas within 5 to 9 km of the coast line could potentially have a different cloud cover from the interior of the state. This argument could be used to explain the increase in ETo more inland as cloud cover would influence solar radiation. Another possible reason could be the low temperatures experienced along the coast due to sea breezes especially in summer. The Homestead weather station was randomly se lected to validate sp atially distributed ETo surfaces. Results suggest that the SEE of the IDWA (0.24 mm day-1) and spline (0.24 mm day-1) methods was lower than the SEE obtaine d based on the principle of using ETo from the closest weather station S331W (0.28 mm day-1). Comparison of ETo for FAWN Homestead weather station and ETo from S331W weather station pro duced an RMSD of 0.68 mm day-1, which is higher than that obtained by th e spline method of 0.59 mm day-1 but lower than IWDA RMSD of 0.74 mm day-1 (Table 3-5). R2 (0.93) from the spline interpolation method was higher than that for IWDA (0.92) and that based on the principle of using weather data from the closest weather station (0.92). Given the coarse distribution of weather stations in Miami-Dade County, by only knowing GPS coordinates of a given location, ETo surfaces from spatial interpolation provide a cheap and easy method for obtaining spatially distributed ETo data. The effect of change in land surface elevation on ETo interpolation was not cons idered due to the relati vely flat topography of south Florida (i.e. most areas in south Florida are less than 10 m above sea level). The results from the interpolation of ETo reveals that using this techniqu e can result in improved estimation of ETo between weather stations, which could result in improved irrigation scheduling. Hence, increasing water use efficiency and energy sa ving. Overall the ther e were no substantial differences in ETo surfaces generated by the IDWA a nd the spline interpolation methods.

PAGE 93

93 Conclusion Comparison of five ETo estimation equations against the ASCE-EWRI standardized equation was completed using four compar ative statistics (RMSD, SEE, S and R2). In summary, Turc (1961) method produced the highest overall rank while the Hargreaves produced the lowest overall rank. Apart from the Hargreaves e quation that consistently over estimated ETo especially in summer, the four other equations produced satisfactory results. The results of the SFWMDSM also correlated well with those of the ASCE-EWRI standardized ETo equation but the method could be improved by introduction of a temperature term in the equation. The PriestlyTaylor tended to overestimate AS CE-EWRI standardized equation ETo estimates during the summer but performed well during winter months. The UF IFAS (1984) Penman equation over estimated ASCE-EWRI standardized ETo estimates throughout the year but was able to follow the same pattern as the ASCE-EWRI standardized ETo equation. The ability of the various ETo estimation equations to predict ASCE-EWRI standardized ETo estimates improved when the evaluation time step was increased from daily to monthly. Finally, the Turc (1961) performed well against the ASCE-EWRI standardized ETo equation but the method could further be improved by introduction of a wind correction factor especially in summer. For areas in south Florida were the only available weather parameters are solar radiation and air temperature, Turc (1961) would be the most appropriate simple equation to use. There was no substantial difference in the results of the two spatial interpolation techniques that were compared. However, the spline ETo surface resulted in slightly better estimates of ETo compared to the closest weather station a pproach. The findings of this study sugge st that performance of onsite or stand alone ET-based irri gation scheduling controllers would be improved by inclusion of the Turc (1961) equation in their ir rigation scheduling algorithms for sub humid and subtropical climatic regions. Spatial interpolation of ETo results in further improvement in ETo estimation

PAGE 94

94 between weather stations compared to using the ETo from the closest weather station. Therefore, if adopted by the growers across the state, this technique could result in improved estimation of how much irrigation water to apply hen ce saving the growers water and energy.

PAGE 95

95 Table 3-1. Selected weather stations Station County Latitude (decimal degrees) Longitude (decimal degrees) Agency Data period Homestead Miami-Dade 25.510 -80.498 FAWN1 1997 to 2009 Fort Lauderdale Broward 26.087 -80.242 FAWN 2001 to 2009 G3ASWX Miami-Dade 25.852 -80.766 SFWMD2 2000 to 2009 S331W Miami-Dade 25.611 -80.509 SFWMD 1994 to 2009 JBTS Miami-Dade 25.225 -80.54 SFWMD 1991 to 2009 1 Florida Automated Weather Network 2South Florida Water Management District Table 3-2. Mean Daily ETo 1 estimates by different equations (mm/day) Station ASCE Standardized2 UF IFAS3PENMAN SFWMDSM4 PT5 TUR6 HARG7 Homestead 3.32 4.61 3.46 3.55 3.39 4.51 Fort Lauderdale 3.40 4.65 3.47 3.58 3.45 4.12 G3ASWX 4.15 4.99 3.76 3.74 3.70 4.25 S331W 3.71 4.74 3.64 3.60 3.55 4.48 JBTS 3.79 4.94 4.22 4.24 4.15 3.33 1Reference evapotranspiration 2American Society of Civil Engineers standardized ETo Equation 3 University of Florida Institute of Food and Agricultural Sciences 4South Florida Water Management District-Simple Method Equation 5 Priestley-Taylor Equation 6Turc Equation 7Hargreaves Equation

PAGE 96

96 Table 3-3. Performance statistics for daily ETo 1 comparison with ranking in parenthesis Station Method SEE2 (mm/day) S3 (mm/day) (R2)4 RMSD5 OVERALL Homestead UF IFAS6 SFWMD7 PT8 TUR9 HARG10 0.346 (2) 0.409(3) 0.422(4) 0.294(1) 0.829(5) 0.189(3) 0.180(1) 0.188(2) 0.199(4) 0.932(5) 0.944(1) 0.896(4) 0.907(3) 0.937(2) 0.304(5) 1.394(4) 0.443(2) 0.554(3) 0.307(1) 1.479(5) 2.5(2) 2.5(2) 3.0(4) 2.0(1) 5.0(5) Fort Lauderdale UF IFAS SFWMD PT TUR HARG 0.356(1) 0.492(4) 0.458(3) 0.358(2) 0.928(5) 0.140(1) 0.206(4) 0.176(3) 0.163(2) 0.644(5) 0.948(1) 0.848(4) 0.896(3) 0.905(2) 0.146(5) 1.395(5) 0.505(2) 0.556(3) 0.374(1) 1.271(4) 2.0(3) 3.5(4) 1.8(2) 1.7(1) 4.8(5) G3ASWX UF IFAS SFWMD PT TUR HARG 0.456(3) 0.476(4) 0.436(2) 0.361(1) 0.792(5) 0.136(3) 0.142(4) 0.132(2) 0.106(1) 0.236(5) 0.882(2) 0.819(4) 0.857(3) 0.887(1) 0.506(5) 0.978(5) 0.617(3) 0.614(2) 0.583(1) 0.830(4) 3.3(3) 3.8(4) 2.3(2) 1.0(1) 4.8(5) S331W UF IFAS SFWMD PT TUR HARG 0.432(2) 0.478(4) 0.456(3) 0.355(1) 0.829(5) 0.147(3) 0.156(4) 0.142(2) 0.118(1) 0.410(5) 0.905(1) 0.843(4) 0.874(3) 0.901(2) 0.425(5) 1.176(5) 0.485(2) 0.495(3) 0.395(1) 1.152(4) 2.8(2) 3.5(4) 2.8(2) 1.3(1) 4.8(5) JBTS UF IFAS SFWMD PT TUR HARG 0.519(3) 0.564(4) 0.411(2) 0.403(1) 0.842(5) 0.194(3) 0.216(4) 0.145(1) 0.156(2) 0.400(5) 0.887(3) 0.834(4) 0.928(1) 0.908(2) 0.303(5) 1.641(4) 0.500(2) 0.502(3) 0.289(1) 1.671(5) 3.3(3) 3.5(4) 1.8(2) 1.5(1) 5.0(5) 1 Reference evapotranspiration 2 Standard Error of Estimate 3 Standard deviation of the ratio of ETo estimated by given equation ETo estimate by ASCEEWRI 4 Coefficient of determination 5 Root Mean Square Difference 6 University of Florida Institute of Food and Agricultural Sciences ETo Equation 7 South Florida Water Manageme nt District-Simple Method ETo Equation 8 Priestley-Taylor ETo Equation 9 Turc ETo Equation 10 Hargreaves ETo Equation

PAGE 97

97 Table 3-4. Performance statistics for mean monthly ETo 1 comparison Station Method SEE2 (mm/day) S3 (mm/day) (R2)4 RMSD5 OVERALL Homestead UF IFAS6 SFWMD7 PT8 TUR9 HARG10 0.148(2) 0.148(2) 0.259(4) 0.043(1) 0.419(5) 0.070(3) 0.024(1) 0.126(5) 0.032(2) 0.116(4) 0.980(2) 0.970(3) 0.957(4) 0.997(1) 0.889(5) 1.233(4) 0.184(2) 0.483(3) 0.114(1) 1.330(5) 2.8(3) 2.0(2) 4.0(4) 1.3(1) 4.8(5) Fort Lauderdale UF IFAS SFWMD PT TUR HARG 0.090(2) 0.175(3) 0.225(4) 0.056(1) 0.384(5) 0.103(5) 0.086(2) 0.099(3) 0.046(1) 0.100(4) 0.992(2) 0.959(4) 0.969(3) 0.997(1) 0.890(5) 1.188(5) 0.224(2) 0.409(3) 0.137(1) 0.876(4) 3.5(4) 2.8(2) 3.3(3) 1.0(1) 4.5(5) G3ASWX UF IFAS SFWMD PT TUR HARG 0.322(4) 0.218(2) 0.257(3) 0.116(1) 0.357(5) 0.086(3) 0.053(2) 0.121(5) 0.029(1) 0.104(4) 0.895(5) 0.918(3) 0.953(2) 0.979(1) 0.913(4) 0.676(5) 0.325(2) 0.463(3) 0.307(1) 0.490(4) 4.3(4) 2.3(2) 3.3(3) 1.0(1) 4.3(5) S331W UF IFAS SFWMD PT TUR HARG 0.186(3) 0.183(2) 0.278(4) 0.110(1) 0.411(5) 0.087(3) 0.064(2) 0.115(5) 0.037(1) 0.105(4) 0.966(2) 0.950(3) 0.950(3) 0.984(1) 0.902(5) 1.122(5) 0.220(2) 0.345(3) 0.171(1) 1.049(4) 3.3(3) 2.3(2) 3.8(4) 1.0(1) 4.5(5) JBTS UF IFAS SFWMD PT TUR HARG 0.183(1) 0.253(3) 0.345(5) 0.270(4) 0.216(2) 0.121(2) 0.219(5) 0.108(1) 0.194(4) 0.164(3) 0.979(1) 0.922(2) 0.914(4) 0.919(3) 0.909(5) 0.993(4) 0.953(3) 1.079(5) 0.920(2) 0.497(1) 2.0(1) 3.3(3) 3.8(5) 3.3(3) 2.8(2) 1 Reference evapotranspiration 2 Standard Error of Estimate 3 Standard deviation of the ratio of ETo estimated by given equation ETo estimate by ASCEEWRI 4 Coefficient of determination 5 Root Mean Square Difference 6 University of Florida Institute of Food and Agricultural Sciences ETo Equation 7 South Florida Water Manageme nt District-Simple Method ETo Equation 8 Priestley-Taylor ETo Equation 9 Turc ETo Equation 10 Hargreaves ETo Equation

PAGE 98

98 Table 3-5. Interpolation methods co mparisons at Homestead FL FAWN1 weather station ETo estimation technique Method SEE4 (mm/day) S5 (mm/day)(R2)6 RMSD7 IDWA2 ASCEEWRI3 0.24 0.10 0.92 0.74 Spline ASCEEWRI 0.24 0.10 0.93 0.59 S331W ASCEEWRI 0.28 0.09 0.92 0.68 1 Florida Automated Weather Network 2 Inverse Distance Weighted Averaging 3American Society of Civil Engineers Environmen tal and Water Resources Institute standardized reference evapotranspiration (ETo) Equation 4 Standard Error of Estimate 5Standard deviation of the ratio of the estimate of ETo from IDWA, spline or the closest weather station (S331W) to ETo for the Homestead, FL weather station 5Coefficient of determination 6Root Mean Square Difference

PAGE 99

99 Figure 3-1. Study area and location of selected weather stations in Broward and Miami-Dade Counties in south Florida

PAGE 100

100 Figure 3-2. Comparison of daily re ference evapotranspiration (ETo) estimated by the American Society of Civil Engineers-Environmenta l and Water Resources Institute (ASCEEWRI) standardized equation to ETo estimated by other equa tions using Homestead, FL weather station data (J anuary 2001 to March 2009). y = 1.3857x R = 0.94440 1 2 3 4 5 6 7 8 9 10 02468UF IFAS (1984) ETo(mm/day)Standardized ASCE ETo(mm/day) y = 1.0487x R = 0.89620 1 2 3 4 5 6 7 8 02468SFWMD ETo(mm/day)Standardized ASCE ETo(mm/day) y = 1.0902x R = 0.90260 1 2 3 4 5 6 7 8 02468Priestley Taylor ETo(mm/day))Standardized ASCE ETo(mm/day) y = 1.0242x R = 0.93720 1 2 3 4 5 6 7 8 02468Turc ETo(mm/day)Standardized ASCE ETo(mm/day) y = 1.3051x R = 0.30380 1 2 3 4 5 6 7 8 9 10 02468Hargreaves ETo(mm/day)ASCE Standardized ETo(mm/day)

PAGE 101

101 Figure 3-3. Comparison of daily re ference evapotranspiration (ETo) estimated by the American Society of Civil Engineers-Environmenta l and Water Resources Institute (ASCEEWRI) standardized equation to ETo estimated by other equations using Fort Lauderdale, FL weather station da ta (January 2001 to March 2009). y = 1.378x R = 0.94760 1 2 3 4 5 6 7 8 9 02468UF IFAS (1984) ETo(mm/day)Standardized ASCE ETo(mm/day) y = 1.0281x R = 0.8480 1 2 3 4 5 6 7 8 02468SFWMD ETo(mm/day)Standardized ASCE ETo(mm/day) y = 1.0806x R = 0.89620 1 2 3 4 5 6 7 8 02468Priestley Taylor ETo(mm/day)Standardized ASCE ETo(mm/day) y = 1.0196x R = 0.9045 0 1 2 3 4 5 6 7 02468Turc ETo(mm/day)Standardized ASCE ETo(mm/day) y = 1.1658x R = 0.1420 1 2 3 4 5 6 7 8 02468Hargreaves ETo(mm/day)Standardized ASCE ETo (mm/day)

PAGE 102

102 Figure 3-4. Comparison of daily re ference evapotranspiration (ETo) estimated by the American Society of Civil Engineers-Environmenta l and Water Resources Institute (ASCEEWRI) standardized equation to ETo estimated by other equations using S331W, FL weather station data (Jan uary 2001 to March 2009). y = 1.2805x R = 0.90450 1 2 3 4 5 6 7 8 9 10 02468UF IFAS (1984) ETo(mm/day)Standardized ASCE ETo(mm/day) y = 0.9798x R = 0.84270 1 2 3 4 5 6 7 8 02468SFWMD ETo(mm/day)Standardized ASCE ETo(mm/day) y = 0.9867x R = 0.87390 1 2 3 4 5 6 7 8 9 02468Priestley Taylor ETo(mm/day)Standardized ASCE ETo(mm/day) y = 0.9561x R = 0.90070 1 2 3 4 5 6 7 8 02468Turc ETo(mm/day)Standardized ASCE ETo(mm/day) y = 1.1718x R = 0.42480 1 2 3 4 5 6 7 8 9 02468Hargreaves ETo(mm/day)ASCE Standardized ETo(mm/day)

PAGE 103

103 Figure 3-5. Comparison of daily re ference evapotranspiration (ETo) estimated by the American Society of Civil Engineers-Environmenta l and Water Resources Institute (ASCEEWRI) standardized equation to ETo estimated by other eq uations using G3ASWX, FL weather station data (J anuary 2001 to March 2009). y = 1.2011x R = 0.88210 1 2 3 4 5 6 7 8 9 10 02468UF IFAS (1984) ETo(mm/day)Standardized ASCE ETo(mm/day) y = 0.9088x R = 0.81970 1 2 3 4 5 6 7 02468SFWMD ETo(mm/day)Standardized ASCE ETo(mm/day) y = 0.9168x R = 0.85710 1 2 3 4 5 6 7 8 02468Priestley Taylor ETo(mm/day)Standardized ASCE ETo(mm/day) y = 0.8935x R = 0.88660 1 2 3 4 5 6 7 02468Turc ETo(mm/day)Standardized ASCE ETo (mm/day) y = 1.0095x R = 0.50630 1 2 3 4 5 6 7 8 02468Hargreaves ETo(mm/day)Standardized ASCE ETo (mm/day)

PAGE 104

104 Figure 3-6. Comparison of daily re ference evapotranspiration (ETo) estimated by the American Society of Civil Engineers-Environmenta l and Water Resources Institute (ASCEEWRI) standardized equation to ETo estimated by other equations using JBTS, FL weather station data (Jan uary 2001 to March 2009). y = 1.29x R = 0.88660 2 4 6 8 10 12 14 16 024681012UF IFAS (1984) ETo(mm/day)Standardized ASCE ETo(mm/day) y = 1.102x R = 0.8340 2 4 6 8 10 12 14 024681012SFWMD ETo(mm/day)Standardized ASCE ETo(mm/day) y = 1.1349x R = 0.92760 2 4 6 8 10 12 14 16 024681012Priestle Taylor ETo(mm/day)Standardized ASCE ETo(mm/day) y = 1.0856x R = 0.9080 2 4 6 8 10 12 14 024681012Turc ETo(mm/day)Standardized ASCE ETo(mm/day) y = 0.8207x R = 0.3030 2 4 6 8 10 024681012Hargreaves ETo(mm/day)Standardized ASCE ETo(mm/day)

PAGE 105

105 Figure 3-7. Comparison of mean monthly ETo estimated using the American Society of Civil EngineersEnvironmental and Water Resources Institute (ASCE-EWRI) standardized equation to ETo estimated using five other equations at the Homestead, FL weather station. Figure 3-8. Comparison of mean monthly ETo estimated using the American Society of Civil Engineers-Environmental and Water Resources Institute (ASCE-EWRI) standardized equation to ETo estimated using five other equati ons at the Fort Lauderdale, FL weather station. 0 1 2 3 4 5 6 7 JANFEBMARAPRMAYJUNJULAUGSEPOCTNOVDECETo(mm/day) ASCE IFAS SFWMD PT TURC HARG 0 1 2 3 4 5 6 7 JANFEBMARAPRMAYJUNJULAUGSEPOCTNOVDECETo(mm/day) ASEC IFAS SFWMD PT TURC HARG

PAGE 106

106 Figure 3-9. Comparison of mean monthly ETo estimated using the American Society of Civil Engineers-Environmental and Water Resources Institute (ASCE-EWRI) standardized equation to ETo estimated using five other equa tions at the S331W, FL weather station. Figure 3-10. Comparison of mean monthly ETo estimated using the American Society of Civil Engineers-Environmental and Water Resources Institute (ASCE-EWRI) standardized equation to ETo estimated using five other equa tions at the G3ASWX, FL weather station. 0 1 2 3 4 5 6 7 JANFEBMARAPRMAYJUNJULAUGSEPOCTNOVDECETo(mm/day) ASCE SFWMD PT TURC HARG IFAS 0 1 2 3 4 5 6 7 JANFEBMARAPRMAYJUNJULAUGSEPOCTNOVDECETo(mm/day) ASCE IFAS SFWMD PT TURC HARG

PAGE 107

107 Figure 3-11. Comparison of mean monthly ETo estimated using the American Society of Civil Engineers-Environmental and Water Resources Institute (ASCE-EWRI) standardized equation to ETo estimated using five other equations at the JBTS, FL weather station. 0 1 2 3 4 5 6 7 JANFEBMARAPRMAYJUNJULAUGSEPOCTNOVDECETo(mm/day) ASCE IFAS SFWMD PT TURC HARG

PAGE 108

108 Figure 3-12. Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for January. Figure 3-13. Spline based mean monthl y reference evapotranspiration (ETo) surface for January.

PAGE 109

109 Figure 3-14. Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for February. Figure 3-15. Spline based mean monthl y reference evapotranspiration (ETo) surface for February.

PAGE 110

110 Figure 3-16. Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for March. Figure 3-17. Spline based mean monthl y reference evapotranspiration (ETo) surface for March.

PAGE 111

111 Figure 3-18. Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for April. Figure 3-19. Spline based mean monthl y reference evapotranspiration (ETo) surface for April.

PAGE 112

112 Figure 3-20. Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for May. Figure 3-21. Spline based mean monthl y reference evapotranspiration (ETo) surface for May

PAGE 113

113 Figure 3-22. Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for June. Figure 3-23. Spline based mean monthl y reference evapotranspiration (ETo) surface for June.

PAGE 114

114 Figure 3-24. Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for July. Figure 3-25. Spline based mean monthl y reference evapotranspiration (ETo) surface for July.

PAGE 115

115 Figure 3-26. Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for August. Figure 3-27. Spline based mean monthl y reference evapotranspiration (ETo) surface for August.

PAGE 116

116 Figure 3-28. Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for September. Figure 3-29. Spline based mean monthl y reference evapotranspiration (ETo) surface for September.

PAGE 117

117 Figure 3-30. Inverse Distance Weighted Average (IDWA) based mean monthly reference evapotranspiration (ETo) surface for October. Figure 3-31. Spline based mean monthl y reference evapotranspiration (ETo) surface for October.

PAGE 118

118 Figure 3-32. Spline based mean monthl y reference evapotranspiration (ETo) surface for November. Figure 3-33. Spline based mean monthl y reference evapotranspiration (ETo) surface for November.

PAGE 119

119 Figure 3-34. Spline based mean monthl y reference evapotranspiration (ETo) surface for December. Figure 3-35. Spline based mean monthl y reference evapotranspiration (ETo) surface for December.

PAGE 120

120 CHAPTER 4 SUMARY AND CONCLUSION The goal of this research was to evaluate a nd generate information on the suitability of evapotranspiration (ET)-based irrigation scheduling technologies for agricultural applications, specifically the technologies abili ty to estimate ET losses and apply water volumes to maintain a desired soil water content (close to field ca pacity) for optimum crop growth and to reduce volumes of irrigation water app lied. To address this challenge, the following objectives were completed: 1) evaluation of ET-based irrigation wa ter management in a tropical fruit orchard in south Florida and 2) comparis on of various reference ET (ETo) estimation equations and spatial interpolation techniques in south Florida. Objective 1 ET-based irrigation scheduling technologies adequately s upplied the water needs of 15 year old Arkin carambola trees by replacing wate r that was daily lost through crop ET. In addition, the study revealed that substantial qua ntities of water were saved by the ET-based irrigation scheduling treatments compared to th e typical irrigation pract ice (68% and 70% less irrigation water was applied by th e real-time and historical ET-based treatments, respectively, compared to the typical irrigation practice for commercial carambola production in south Florida). ET-based irrigation schedul ing methods also demonstrated th eir ability to maintain root zone soil water content below 10 kPa (close to field capacity of 34% volumetric water content). Evaluation of carambola physiological responses to real-time and historical ET-based irrigation scheduling, revealed that this t ype of irrigation water management did not appear to negatively affect stem water potential and leaf gas exchange which were us ed in the study as indicators of plant water stress. Yield data was also collected and results indicated no significant differences among treatments. ETo estimated using data from remote weather stations under estimated ETo

PAGE 121

121 estimated using onsite weather data probably due to the fact that weat her conditions at the remote weather stations did not accurately represent the weathe r conditions at the study site. Objective 2 The Turc (1961) radiation based ETo equation produced the best overall performance in predicting ETo estimated by the American Society of Civil Engineers-Environmental and Water Resources Institute (ASCE-EWRI) standardized equation. The Hargreaves method, the only temperature based ETo equation included in the study, had the lowest performance against the ASCE-EWRI standardized equation which was used as the bench mark. The results of the UF IFAS (1984) Penman and Priestley Taylor equations were satisfactory, although the two equations tended to overestimate ASCE-EWRI ETo especially in summer. The results of the South Florida Water Management District-Sim ple Method (SFWMD-SM) correlated well with the ETo estimated by ASCE-EWRI st andardized equation, although the equation tended to slightly under ETo in winter. The surprisingly good pe rformance of the simple Turc (1961) equation that only requires radi ation and temperature as input s provides an opportunity for improving irrigation scheduling algorithms for onsite or standalone ET-base d controllers that are usually equipped with radiation and temperature sensor for agricultural applications. There was no substantial difference in the ETo surfaces generated using inverse distance weighted averaging and spline interpolation techniques. Spline ETo surfaces slightly improved ETo estimation between weather stations compared to using ETo from the closest weather station. ET-based irrigation scheduling technologies provide a practical way of increasing water conservation in agriculture without negatively a ffecting the physiological development of the plant and with minimum inconvenience to the grow er. Results of the study demonstrate that over 50% of irrigation water could be saved in tropical fruit orchar ds by adopting ET-based irrigation scheduling technologies. Increased water conservation in agriculture would also translate into

PAGE 122

122 reduced agro-chemical leaching and increased availa bility of water for other uses. However, their remain two main challenges that need additional research in order to improve performance of ET-based irrigation scheduling tech nologies in agriculture: 1) mo re work should be done to develop crop coefficient for the major commerci al crops and 2) additi onally studies should be conducted to evaluate the perfor mance of onsite or stand alone ET-based irrigation technologies in agriculture settings, since these technol ogy could accurately estimate location specific ETo using solar radiation and temperature se nsors coupled with the Turc equation.

PAGE 123

123 LIST OF REFERENCES Abtew, W. (1996). Evapotranspiration measuremen ts and modeling for three Wetland Systems in South Florida. Journal of American Water Resources Association. 32 (3) 465-473. Abtew, W., and J. Obeysekera. 1995. Lysimete r study of evapotranspiration of Cattails and comparison of three estimation methods. Tr ansactions of the ASAE, 38(1): 121-129. Alcamo, J., P. Doll, F. Kaspar, and S. Sieber t .1997. Global Change and Global Scenarios of water Use and Availability: An Applicati on of WaterGAP 1.0. University of Kassel, CESR, Kassel. Alcamo, J., T. Henrichs, and T. Rosch .2000. World Water in 2025: Global modeling and scenario analysis. World Water Scenarios Analyses, World Water Council, Marseille. Allen, R.G., I.A. Walter, R. Elliot, T. Howell, D. Itenfisu, and M. Jensen. 2005. The ASCE Standardized Reference Evapotranspiration E quation. American Society of Civil Engineers Environmental and Water Resource Ins titute (ASCE-EWRI). Available at: http://www.kimberly.uidaho.edu/water/asceewri/ascestzdetmain2005.pdf Accessed 7 December 2008. Allen, R.G., L.S. Pereira, D. Raes, and M. Smith. 1998. Crop evapotranspiration guidelines for computing crop water requirements. FAO Irrigation and drainage paper 56, FAO, Rome, Italy. Al-Yahyai, R.., B. Schaffer and F. S. Davi es. 2003. Monitoring soil moisture content for irrigation scheduling in Carambola orchar d in a gravely limestone stone. In Proc Florida State Hort. Soc. 116:37-41. Al-Yahyai, R., B. Schaffer and F.S. Davies. 2005. Physiological response of carambola to soil water depletion. HortScience 40(7): 2145-2150. Al-Yahyai, R., B. Schaffer, F.S. Davies, and J.H. Crane. 2005. Four levels of soil water depletion minimally affect carambola phenological cycles. Hort Technology 15 (3): 623-630. Al-Yahyai, R., B. Schaffer, F.S. Davies, a nd R. Munoz-Carpena. 2006. Characterization of soil water retention of a very gravelly loam soil varied with determination method. Soil Science 171:85-93. Burrough, P.A., and R.A. McDonnell. 1998. Principl es of Geographical Information Systems. New York, Oxford University Press. Burt, C.M., A. J. Mutziger, R. G. Allen, and T.A. Howell. 2005. Evaporation research: Review and Interpretation. Journal of Irrigation a nd Drainage Engineering ASCE 131(1): 37-58. Cardenas-Lailhacar, B., M.D. Dukes, and G.L. Miller. 2005. Sensor-Based Control of Irrigation in Bermudagrass. ASABE Paper No. 052180 St. Joseph, MI.: ASABE.

PAGE 124

124 Chandra S. Pathak. 2008. South Florida Envi ronmental Report-Appendix 2-1: Hydrological Monitoring Network of the South Florida Wate r Management District. West Palm Beach, FL. Available at: https://my.sfwmd.gov/portal/page?_pag eid=2714,14424186&_dad=portal&_schema=POR TAL. Accessed February 15 2009. Christine Souza. 2008 Irrigation technology: Smart water solutions fo r state's farmers. California Farm Bureau Federation. Available at: http://www.cfbf.com/agalert/AgAle rtStory.cfm?ID=1087&ck=A26398DCA6F47B49876C BAFFBC9954F9 Accessed November 5 2008. Chuanyan, Z., N. Zhongren, and F. Zhaodong, 2004. GI S-assisted spatially distributed modeling of the potential evapotranspiration in semi-arid climate of the Chinese Loess Plateau. Journal of Arid Environments. 58: 387. Courault, D., B. Seguin, and A. Olioso. 2005. Revi ew on estimation of evapotranspiration from remote sensing data: From empirical to numerical modeling approaches. Springer Irrigation and Drainage Systems 19: 223. Crane, J.H. 1993. Commercialization of Carambola, Atemoya, and Other Tropical Fruits in South Florida. In: New crops, 448-460. J. Janick and J.E. Simon,. Wiley, New York. Crane, J.H. 1994. The Carambola (Star fruit) Fact Sheet HS-12. IFAS, Gainesville. Crane, J.H., 2007. Carambola growing in the Florida home landscape. EDIS Fact sheet HS12. Available at: http://edis.ifas.ufl. edu/pdffiles/MG/MG26900.pdf Accessed November 8 2008. Davis, S., M. D. Dukes, V. Sudeep, and G. L. Miller. 2007. Evaluation and Demonstration of Evapotranspiration-Based Irrigation Cont rollers. In Proc ASCE EWRI World Environmental & Water Resources Congress, Tampa Florida. Doorenbos, J and W. O Pruitt. 1977. Guidelin es for predicting crop water requirements. Irrigation and Drainage Paper 24, 2nd ed., Food and Agriculture Organization of the United Nations, Rome. Dukes, M. D., L. Zotarelli., J. M. Schol berg and R. Muoz-Carpe na. 2006. Irrigation and Nitrogen Best Management Practices Under Drip Irrigated Vegetable ProductionProceedings ASCE EWRI World Wate r and Environmental Resource Congress Omaha NE. Dukes, M. D., M. L. Shedd, and S. L. Davis. 2009. Smart Irrigation Controllers: Operation of Evapotranspiration-Based Controll ers. EDIS AE 446. Available at: http://edis.ifas.ufl.edu/AE446 Accessed November 20 2008. George, H.L., F.S. Davies, J.H. Crane, and B. Schaffer. 2002. Root temperature effects on Arkin carambola (Averrhoa carambola L.) trees: I. Leaf gas exchange and water relations. Sci. Hort. 96:53-65.

PAGE 125

125 Hamon, W.R., 1963. Computa tion of Direct Runoff Amounts from Storm Rainfall. Int. Asso. Sci, Hydrol. Pub. 63:52-62. Available at: http://www.cig.ensmp.fr/~iahs/redbooks/a063/063006.pdf Accessed March 5 2009. Hargreaves, G.H. and Z.A. Samani. 1985. Reference Crop Evapotranspiration From Temperature. Applie Eng in Agric. 1(2):96-99. Hartkamp, A.D., K. De Beurs, A. Stein and J.W. White. 1999. Interpolation Techniques for Climate Variables. NRG-GIS Series 9901. Mexico, D.F.: CIMMYT. Available at: http://www.cimmyt.org/Res earch/NRG/pdf/NRGGIS%2099_01.pdf Accessed March 25 2009. Hunt, T., Lessick, D., Berg, J and Wiedmann, J. (2001). Residential weather-based irrigation scheduling: Evidence from the Irvine ET Cont roller study. Irvine Ra nch Water District, Irvine, CA. Available at: http://www.irrigation.or g/swat/images/irvine.pdf Accessed March 27 2009. Hutchinson, M.F. 1991. The application of thin pl ate smoothing splines to continent-wide data assimilation. BMRC Research Report Seri es. Melbourne, Australia, Bureau of Meteorology 27:104-113. Hutchinson, M.F and J.D. Corbett. 1995. Spatial in terpolation of climate data using thin plate smoothing splines. In: Coordination and harm onization of databases and software for agroclimatic applications. Agroclimatol ogy Working paper Series, no. 13. Rome: FAO. Hutchinson, M.F and P.E. Gessler. 1994. Splines more than justa smooth interpolator. Geoderma 62, 45. IA. (2006b). Smart Water Application T echnology (SWAT) Performance Report: WeatherTRAK. Irrigation Association, Falls Church, VA. Available at: http://www.irrigation.org/ gov/pdf/SWAT_PerformanceSummary_WeatherTRAK_2006-19.pdf Accessed 29 March, 2009. Irizarry-Ortiz, M. M. 2003. Selected Methodol ogy for Long-Term (1965-2000) Solar Radiation and Potential Evapotranspiration Estim ation for the SFWMM2000 Update. SFWMD Memorandum RES 17-06. Irmak, S and Haman, D.Z. (2003). Evapotranspiratio n: Potential or reference? ABE 343, Institute of Food and Agricultural Sciences. University of Florida, Gainesville, FL. Available at: http://edis.ifas.ufl.edu/AE256 Accessed 29 February, 2009. Ismail, M. R., and K.M. Noor. 1996. Growth, water relations and physiological processes of star fruit (Averrhoa Carambola) plants under root growth restriction. Sci. Hort. 66: 5 1-58. Ismail, M. R., S.W Burrage, H. Tamizi, and M. A. Aziz. 1994. Growth, plant water relations, photosynthesis rate, and accumulation of prolin e in young carambola plants in relation to water stress. Sci. Hort. 66: 101-114.

PAGE 126

126 Itenfisu, D., R. L. Elliott, R. G. Allen, a nd I. A. Walter. 2003. Comparison of Reference Evapotranspiration Calculations as Part of the ASCE Standardization Effort. Journal of Irrigation and Drainage Engineering-ASCE, 129: 440. Jacobs, J.M and S.R. Satti. 2001. Evaluation of reference evapotranspiration methodologies and AFSIRS crop water use simulation model, Final Report, St. Johns River Water Management District, Palatka, FL, April 2001. Available at: http://www.sjrwmd.com/technicalreports/pdfs/SP/SJ2001-SP8.pdf Accessed 29 November 2009. Jensen, M.E., R.D. Burman, and R.G. Allen. 1990. Evapotranspiration and Irrigation Water Requirements. ASCE Man. and Rep. on Engineering Pract. No. 70, New York, 332 p. Jonathan Crane. 2003. New Plants for Florid a: Tropical Fruit. EDIS Circular 1440 Jones J. W., L.H Allen, S. F. Shih, J. S. Roge rs, L. C. Hammond, A. G. Smajstrala, and J. D. Martsolf. 1984. Estimated and Measured Evapot ranspiration for Florida Climate crops and soils. IFAS technical Bulletin 840. Jones, H.G., A.N Lakso, and J.P. Syvertsen. 198 5. Physiological control of water status in temperate and subtropical fru it trees. Hort. Rev. 7:301-344 Kashyap, P. S., and R. K. Panda. 2001. Evaluation of evapotranspiration estimation methods and development of crop-coefficients for pot ato crop in a sub-humid region. Elsevier Agricultural Water Mana gement (50) 1: 9-25. Katul, G. G., and R. H. Cuenca. 1992. Analysis of Evaporative Flux Data for Various Climates Journal of Irrigation and Drainage Engineering-ASCE (118) 4: 601-618. Lim, T.K. 1996. Carambola growing and mark eting. Agnote D32, Agdex No: 238/10, ISSN No: 0157-8243 Lu, Jianbiao., G. Sun, S. G. McNulty, and D.M. Amatya. 2005. A comparison of six potential evapotranspiration methods for regional use in the southeastern United States. Paper No. 03175 of the Journal of the American Wate r Resources Association (JAWRA) Pages 621632. Available at: http://www3.interscience.wiley.com/cgibin/fulltext/118664972/PDFSTART Accessed 10 March 2009. Marella, R.L., 1999. Water Withdrawals, Use, Di scharge, and Trends in Florida, 1995 WaterResources Investigations Report 99-4002. US Geological Survey, Reston VA. Marella, R.L., 2008. Water use in Florida 2005, and trends 1950-2005 USGS Fact sheet 20083080 U.S. Geological Survey Florida Integrated Science Center (FISC) Orlando, FL 32826. Available at: http://pubs.usg s.gov/fs/2008/3080/ Accessed 30 April 2009. Marler, T.E., B. Schaffer, and J.H. Crane. 1994. Developmental light level affects growth, morphology, and leaf physiology of young carambola trees. J. Am. Soc. Hort. Sci. 119: 711.

PAGE 127

127 Mary, S., M. D. Dukes, and G. L. Miller. 2007. Evaluation of Evapotranspiration and Soil Moisture-based Irrigation Control on Turfgrass, Proceedings ASCE EWRI World Environmental & Water Resources Congress, Tampa, FL. McVicar, T R., T. G. Van Niel, L.T. Li, M. F. Hutchinson, X. Mu, and Z. Liu. 2007. Spatially distributing monthly reference evapotrans piration and pan eva poration considering topographic influences. Elsevi er. Journal of Hydrology 338:196. Migliaccio, K.W. 2007. Sustainability of Agriculture in Miam i-Dade County: Considering Water Supply. EDIS publication ABE 380. Available at: http://trec.ifas.ufl.edu/kwm/f iles/pdf/publications/edis_ABE380.pdf Accessed 17 July 2008. Migliaccio, K.W., and Y. Li. 2009. South Irrigation Scheduling for Tropical Fruit Groves in South. EDIS publication Fact Sheet TR001. Available at: http://edis.ifas.ufl.edu/TR001 Accessed 1 May 2009. Migliaccio, K.W., J. Crane, E. Evans, B. Sc haffer, Y. Li, and R. Muoz-Carpena. 2006. South Florida Tropical Fruit Grower Perspectives : Water Conservation Management Practices. EDIS publication ABE 368. Available at: http://edis.ifas.ufl.edu/AE397 Accessed 17 October 2008. Mossler, M. A. and O.N. Nesh eim. 2002. Florida crop pest management profile Carambola. EDIS CIR 1416. Muoz-Carpena, R., J.H. Crane, I.D. Glenn a nd C.F. Balerdi. 2003. Tropical Fruit Growers' Water Use and Conservation Practices in Miami-Dade County. EDIS publication ABE 345. Muoz-Carpena, R.,Y. Li and T. Olczyk. 2002. A lternatives of Low Cost Soil Moisture Monitoring Devices for Vegetable Producti on in South Miami-Dade County. EDIS publication ABE 333. Available at: http://edis.ifas.ufl.edu/AE230 Accessed 25 February 2009. Nakasone, H.Y and R.E. Paull. 1998. Crop pr oduction science in horticulture: Tropical fruits.CAB International 1998. Nandagiri, L., and G. M. Kovoor. 2006. Performan ce evaluation of reference evapotranspiration equations across a range of Indian climat es. Journal of Irrigation and Drainage Engineering-ASCE (132) 3: 238-249. Naor, A. 2000. Midday stem water potential as a plant water stress i ndicator for irrigation scheduling in fruit trees. Proc. 3rd IS on Irr igation Hort. Crops. Acta Hort 537, ISHS 2000. Neidrauer, C., L. Brion, L. Cadavid, K. Tar boton and J. Barnes. 1999. What are Southeastern Florida's Water Needs? South Florida Water Ma nagement District. Poster presented, at the South Florida Restoration Science Forum.

PAGE 128

128 Neuhaus, A. 2003. Managing irrigation for yield and fruit quality in avocado. Ph.D. Thesis, School of Plant Biology, University of Western Australia, Perth, Australia. Nuez-Elisea, and J.Crane. 2000. Selective pruni ng and crop removal increase early-season fruit production of Carambola. Elsevier Scie nce Scientia Hortic ulturae 86: 115-126. Nuez-Elisea, R., B. Schaffer, M. Zekri, S. K. OHair and J. H. Crane. 2001. In situ soilwater characteristic curves for tropical fruit or chards in trenched calcareous soil. Hort Technology. 11: 65. Penman, H.L. 1948. Natural evaporation from open water, bare soil and grass. Proceedings of the Royal Society of London, A193: 120-146. Pilar, B., P. Sanchez-de-Miguel, A. Centeno, P. Junquera, R. Linares, J. R. Lissarrague. 2007. Water relations between leaf water potential, photosynthesi s and agronomic vine response as a tool for establishing thresholds in ir rigation scheduling. Elsevier Science Scientia Horticulturae 114: 151-158. Pimentel D., J. Houser, E. Preiss, O. White, H. Fang, L. Mesnick, T. Barsky, S. Tariche, J. Schreck, and S. Alpert.1997. Wate r Resources: Agriculture, th e Environment, and Society. BioScience, American Institute of Biol ogical Sciences, Vol. 47, No. 2, pp. 97-106. Priestley, C.H.B., and Taylor, R.J., 1972. On the assessment of surface heat flux and evaporation using large-scale parameters. Monthly Weather Review 100: 81. Raskin, P., P. Gleick, P. Kirshen, G. Pontius, and K. Strzepek. 1997. Wa ter Futures: Assessment of long range patterns and prospects. (Stockholm Environment Institute, Stockholm). Razi, M.I., M. Awang and S. Razlan. 1992. Effe ct of water stress on growth and physiology processes of Averrahoa caram bola. Acta. Hort. 321: 505-509. Rijsberman, FR. 2004. Water Scarcity: Fact or Fiction? New directions for a diverse planet. Proceedings of the 4th International Crop Science Congress. Brisbane, Australia. Salakpetch, S., D.W. Turner, and B. Dell. 1990. The flower ing of carambola (Averrhoa carambola L.) is more strongly influenced by cultivar and water stress than by diurnal temperature and photoperiod. Sci. Hort. 43:83-94. Shackel, K.A. H. Ahmadi, W. Biasi, R. Buchne r, D. Goldhamer, S. Gurusinghe, J. Hasey, D. Kester, B. Krueger, B. Lampinen, G. McGourty, W. Micke, E. Mitcham, B. Olson, K. Pelletrau, H. Philips, D. Ramos, L. Schwankl, S. Sibbett, R. Snyder, S. Southwick, M. Stevenson, M. Thorpe, S. Weinbaum and J. Y eager. 1997. Plant water stat us as an index of irrigation need in deciduous fr uit trees. Hort Technology. 23. Shackel, K. A., B. Lampinen, S. Sibbett and W. Olson. 2000a. The relation of midday stem water potential to the growth and physiology of fr uit trees under water limited conditions. Acta Hort. 537:425-430.

PAGE 129

129 Shiklomanov, I. A .1998. World Water Resources: An Appraisal for the 21st Century. IHP Report. (UNESCO, Paris). Simonne, A., L. B. Bobroff, A. Cooper, S. Poirier, M. Murphy, M. Oswald, and C. Procise. 2004. South Florida Tropicals: Carambola. EDIS. Fact Sheet FCS 8520. Stallings, C., R.L. Huffman, S. Khorram and Z. Guo. 1992. Linking Gleams and GIS. ASAE Paper 92 3613. St. Joseph, Michigan. Stephens, J.C., and Stewart, E.H. 1963.A Comparison of Procedures for Computing Evaporation and Evapotranspiration. A.S.H. Committ ee for Evaporation. (62) 1: 123-133. Sumner, D M.,. and J. M. Jacobs. 2004. Utility of PenmanMonteith, PriestleyTaylor, reference evapotranspiration, and pan eva poration methods to estimate pasture evapotranspiration. Elsevier Journal of Hydrology 30: 81. Taiz, L. and Zeiger, E. 1998. Plant Physio logy. 2nd ed. Sinauer Associates, Sunderland, Massachusetts, USA. Tang, Q., S. Peterson, R. Cuenca, Y. Hagimoto, D. P. Lettenmaier. 2008. Satellite-based realtime estimation of crop water Consumption. Journal of geophysical research, Vol. XXXX, DOI:10.1029 Temesgen, B., S. Eching, B. Davidof and K. Frame. 2005. Comparison of Some Reference Evapotranspiration Equations for California, Journal of Irrigation and Drainage Engineering-ASCE,( 131)1: 73. Trajkovi S and V. Stojni 2007. Effect of wind speed on accuracy of Turc method in a humid climate. Facta Universitatis. Architectu re and Civil Engineering (5) 2: 107. Turc, L., 1961. Evaluation des besoins en eau dirrigation, vapotrans piration potentielle, formulation simplifi et mise jour. Ann. Agron., 12: 13-49. US Department of interior Bureau of R eclamation. 2007. Weather and Soil moisture based landscape irrigation scheduling devices, tec hnical review report 2nd Edition. Lower Colorado region, southern Carlif ornia office. Available at: http://www.usbr.gov/waterconservation/docs/SmartController.pdf Accessed 16 August 2008. USGS Circular 1309. 2007. Facing Tomorrows Cha llengesU.S. Geological Survey Science in the Decade 2007. Available at: http://www.usbr.gov/waterconservation/docs/SmartController.pdf Accessed 25 March2009. Vijay P. S. and D. K. Frevert. 2002. Mathematic al models of large watershed hydrology Water Resources Publications. LLC.

PAGE 130

130 Walter, I. A., R. G. Allen, M. E. Jensen, R. L. Elliott, R. H. Cuenca, S. Eching, M. J. Hattendorf, T. A. Howell, D. Itenfisu, D. L. Martin, B. Mecham, R. L. Snyder, T. L. Spofford, P.W. Brown, and J. L. Wright. 2002. The ASCE st andardized reference evapotranspiration equation draft report. Appendices A F. Envi ronmental and Water Resources Institute of the American Society of Civil Engineers. Wright, J.L., R.G. Allen and T.A. Howell. 2000. Conversion between evapotranspiration references and methods. 251-259, Proc., 4th De cennial National Irrigation Symposium, Phoenix, AZ, ASAE, St. Joseph, MI. Xu, C., Gong, L. Jiang, T. Chen, and D. Singh. 2006. Analysis of patial di stribution and temporal trend of reference evapotrans piration and pan evaporation in Changjiang (Yangtze River) catchment. Journal of Hydrology 327: 81. Zotarelli, L., J.M. Scholberg, and M. Dukes, R. Muoz-Carpena, and J. Icerman. 2004. Tomato yield, biomass accumulation, root distribution and irrigation water use efficiency on a sandy soil, as affected by nitrogen rate a nd irrigation scheduling. Agricultural Water Management 96:23. Elsevier Science.

PAGE 131

131 BIOGRAPHICAL SKETCH Kisekka Isaya began his education at Re v. John Foundation primary school. Towards the end of his primary level educat ion he moved to St. James prim ary school located in Kampala city, Uganda. In 1992, he started his secondary sc hool education at St. Jo sephs senior secondary school Naggalama. At Naggalama, he develope d a love for mathematics and physics. Great encouragement from his mother, friends and teacher s, together with hard work helped Isaya to excel at the Uganda Certificate of Education (U CE) national exam. This was a turning point in his life, because of his good UCE score he was admitted in one of the oldest and top ranking secondary school in Uganda called Namilyango College. At Namilyango College, he majored in physics, chemistry and mathematics. In 1998, he took the Uganda Adva nced Certificate of Education (UACE) national examination, through gr eat support from his family and teachers he excelled again in the UACE exam. This enabled him to win a govern ment of Uganda scholarship to study bachelors degree in Agricultural Engi neering of Makerere University Kampala, Uganda. After his undergraduate degree in 2002, he worked as an irrigation engineer with Balton (U) Ltd, a place where he learned a great deal about the engineering profession. Three years later, he joined the National Agricultural Rese arch Organization (NARO) of Uganda. He has also received several professional trainings in irrigatio n and drainage engineeri ng in several countries including Israel, Egypt and Kenya. In 2007, he decided to join graduate school and he was lucky to be accepted at the University of Florida department were his majoring in Agricultural and Biological engineering under the superv ision of Dr. Migliaccio K. White.

PAGE 132

EVAPOTRANSPIRATION BASED IRRIGATION SCHEDULING FOR A TROPICAL FRUIT ORCHARD IN SOUTH FLORIDA Isaya Kisekka 305-246-7001 Agricultural and Biological Engineering Dr Migliaccio K. White Masters of Engineering August 2009 Demand for fresh water is growing in Florid a due to high mainly high population growth rate. Conserving water in agricu lture in considered as one ways of increasing fresh water availability for other designated uses, since agricultures uses the largest amount of water in the state. Advances in irrigation scheduling t echnologies have led to the development of Evapotranspiration (ET) controlle rs. These are devices that use weather data and site specific conditions to schedule irrigati on. A study was conducted to evalua te their ability to reduce volumes of water used in irri gation without negatively affecti ng optimum crop growth. Results indicate the technology has the po tential to save substantial am ounts of water in tropical fruit orchards without negatively affecting trees. The st udy also revealed that ef ficiency of irrigation scheduling could be improved especially for ons ite ET controllers if a simple equation called Turc (1961) is used in estimating refere nce evapotranspiration for south Florida.