Citation
Crop Coefficients and Water Quality for Watermelon and Bell Pepper under Drip and Seepage Irrigation

Material Information

Title:
Crop Coefficients and Water Quality for Watermelon and Bell Pepper under Drip and Seepage Irrigation
Creator:
SRIVASTAVA, SAURABH
Copyright Date:
2008

Subjects

Subjects / Keywords:
Crops ( jstor )
Groundwater ( jstor )
Irrigation ( jstor )
Irrigation systems ( jstor )
Lysimeters ( jstor )
Peppers ( jstor )
Rain ( jstor )
Soils ( jstor )
Surface runoff ( jstor )
Watermelons ( jstor )
Miami metropolitan area ( local )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Saurabh Srivastava. Permission granted to the University of Florida to digitize, archive and distribute this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
11/30/2005
Resource Identifier:
436098767 ( OCLC )

Downloads

This item is only available as the following downloads:


Full Text

PAGE 1

CROP COEFFICIENTS AND WATER QU ALITY FOR WATERMELON AND BELL PEPPER UNDER DRIP AN D SEEPAGE IRRIGATION By SAURABH SRIVASTAVA 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 2005

PAGE 2

Copyright 2005 by Saurabh Srivastava

PAGE 3

This document is dedicated to my family.

PAGE 4

ACKNOWLEDGMENTS This thesis work would not have been possible without the support of several people whom I wish to thank. First, I would like to thank my supervisory committee chairman, Dr. Sanjay Shukla, for his support, encouragement and guidance throughout my experience in graduate school. I also thank the members of my supervisory committee and Dr. Fouad Jaber for their insightful comments and suggestions for the improvement of this document. Special thanks go to my friend Mr. Chambal Pandey who has helped me immensely during my research work. I am grateful to Mr. Dale Hardin for his invaluable assistance in my research. I also appreciate the assistance of Mr. Roger Mc Gill, the workshop crew and the farm crew at Southwest Florida Research Education Center at Immokalee, FL, for their help during the field work of this research. Finally, I would like to thank the very special people in my life: my mother, father, sister and wife Milan, who have been the pillars of support and encouragement. Without them this would have never been possible and to them I dedicate this work. iv

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS.................................................................................................iv LIST OF TABLES...........................................................................................................viii LIST OF FIGURES.............................................................................................................x ABSTRACT......................................................................................................................xii CHAPTER 1 INTRODUCTION........................................................................................................1 Background...................................................................................................................1 Goals and Objectives....................................................................................................4 2 REVIEW OF LITERATURE.......................................................................................5 Impact of Agricultural Water Management on Water Use...........................................6 Irrigation System...................................................................................................6 Irrigation Scheduling.............................................................................................7 Evapotranspiration.................................................................................................9 Reference evapotranspiration (ETo)...............................................................9 Crop evapotranspiration...............................................................................11 Crop coefficient...................................................................................................12 Estimation of Kc...........................................................................................13 Design considerations for lysimeters...........................................................15 Lysimeter-based crop coefficients................................................................16 Impacts of Agricultural Water Management on Water Quality.................................19 Nitrogen Transport..............................................................................................20 Phosphorus Transport..........................................................................................22 Tools for Quantifying Nutrient Transport...........................................................25 3 CROP COEFFICIENTS AND CROP EVAPOTRANSPIRATION FOR WATERMELON AND BELL PEPPER UNDER DRIP AND SEEPAGE IRRIGATION.............................................................................................................28 Introduction.................................................................................................................28 Material and Methods.................................................................................................30 v

PAGE 6

Study Location.....................................................................................................30 Experimental Design...........................................................................................30 Survey of crop production practices.............................................................31 Lysimeter design..........................................................................................31 Field design..................................................................................................32 Lysimeter Construction and Installation.............................................................35 Monitoring System..............................................................................................38 Irrigation.......................................................................................................39 Drainage and runoff.....................................................................................40 Soil moisture and water table.......................................................................43 Weather parameters......................................................................................45 Fertilizer management..................................................................................45 Crop yield.....................................................................................................45 Crop Production Practices...................................................................................46 Watermelon..................................................................................................46 Pepper...........................................................................................................47 Computation of Rreference Evapotranspiration..................................................47 FAO-Penman-Monteith method...................................................................47 FAOModified Blaney-Criddle method......................................................48 Quantification of Crop Evapotranspiration.........................................................48 Development of Crop Coefficient.......................................................................49 Result and Discussion.................................................................................................51 Bell Pepper..........................................................................................................51 Crop Evapotranspiration (ETc).....................................................................53 Crop Coefficient (Kc)...................................................................................57 Watermelon.........................................................................................................62 Crop Evapotranspiration (ETc).....................................................................64 Crop Coefficient (Kc)...................................................................................66 Summary and Conclusion...........................................................................................70 4 IMPACTS OF DRIP AND SEEPAGE IRRIGATION SYSTEMS ON THE WATER QUALITY FOR VEGETABLE PRODUCTION IN FLORIDA................72 Introduction.................................................................................................................72 Material and Methods.................................................................................................76 Lysimeter System................................................................................................76 Cropping Practice................................................................................................76 Fertilizer Management.........................................................................................76 Nutrient Cycling and Transport Monitoring........................................................78 Sample Collection, Storage and Handling...........................................................79 Soil...............................................................................................................79 Water............................................................................................................80 Plant Tissue..................................................................................................81 Nutrient Balance..................................................................................................81 Resident Soil nutrient...................................................................................82 Irrigation, drainage and runoff discharge.....................................................84 Crop..............................................................................................................84 vi

PAGE 7

Result and Discussion.................................................................................................85 Nitrogen Transport in Spring 2003......................................................................85 Nitrate in soil................................................................................................85 Nitrate in drainage and runoff......................................................................86 Ammonium and dissolved inorganic nitrogen in soil..................................88 Ammonium and DIN in drainage and runoff...............................................89 Nitrogen mass balance.................................................................................91 Nitrogen transport in fall 2003............................................................................92 Nitrate in soil................................................................................................92 Nitrate in groundwater.................................................................................93 Nitrate in drainage and runoff......................................................................95 Ammonium and DIN in soil, groundwater, drainage and runoff.................96 Nitrogen mass balance.................................................................................98 Phosphorus Transport........................................................................................101 Spring 2003................................................................................................101 Fall 2003.....................................................................................................104 Summary and Conclusion.........................................................................................108 5 SUMMARY, CONCLUSIONS AND SUGGESTIONS..........................................109 Summary and Conclusions.......................................................................................109 Suggestions...............................................................................................................111 APPENDIX A WEATHER DATA...................................................................................................113 B WATER INPUTS AND OUTPUTS FOR THE LYSIMETERS.............................116 C WATER TABLE DATA..........................................................................................119 LIST OF REFERENCES.................................................................................................121 BIOGRAPHICAL SKETCH...........................................................................................131 vii

PAGE 8

LIST OF TABLES Table page 2-1 Statistical results for corn crop curves from 11 years of data..................................16 3-1 ETc, ETo and rainfall for bell pepper crop in fall 2003 season.................................57 3-2 Regression coefficients for bi-weekly crop coefficient curve for bell pepper.........61 3-3 Monthly crop coefficients for bell pepper (fall 2003)..............................................61 3-4 ETc, ETo and rainfall for watermelon crop in spring 2004 season...........................65 3-5 Regression coefficients for bi-weekly crop coefficient curve for watermelon........68 3-6 Monthly crop coefficients for watermelon (spring 2004)........................................69 4-1 Crop production schedule for the lysimeters............................................................76 4-2 Crop, rates (kg/ha) and forms of fertilizer application for spring and fall 2003 seasons......................................................................................................................77 4-3 Nitrogen and phosphorus analysis for soil, water and plant components for the lysimeters.................................................................................................................78 4-4 Monitoring and sampling schedule for hydrologic and nutrient input and output to the lysimeters.......................................................................................................79 4-5 Nitrate loading, mean concentration and flow weighted concentration for individual lysimeter drainage and runoff.................................................................88 4-6 DIN loading, mean NH4 concentration and flow weighted NH4 concentration for individual lysimeter drainage and runoff.................................................................90 4-7 Steady-state N mass balance for drip and seepage systems for spring 2003...........91 4-8 Total NO3 and DIN loading in drainage and runoff between groundwater sampling events during fall 2003.............................................................................94 4-9 Total discharge volume (liter) for the lysimeters during study period....................95 viii

PAGE 9

4-10 Total Kjeldhal nitrogen (TKN) concentrations (mg/kg of soil) for soil samples during post-fall 2003 season....................................................................................99 4-11 Steady-state N and transient-state N mass balance for drip and seepage systems for fall 2003............................................................................................................100 4-12 Transient-state N mass balance for the lysimeters during fall 2003......................100 4-13 Total P discharges (drainage and runoff) and mean P concentration in lysimeter drainage and runoff................................................................................................105 4-14 Drainage and runoff P discharge between groundwater sampling events during fall 2003..................................................................................................................107 A-1 Monthly summary of historical weather data and fall 2003 weather data.............114 A-2 Monthly summary of historical weather data and spring 2004 weather data.........115 B-1 Summary of rainfall, irrigation, drainage and runoff during fall 2003 season.......117 B-2 Summary of rainfall, irrigation, drainage and runoff during spring 2004 season..118 ix

PAGE 10

LIST OF FIGURES Figure page 3-1 Study location...........................................................................................................30 3-2 Lysimeter layout for the watermelon crop...............................................................32 3-3 Experimental field layout.........................................................................................34 3-4 Construction details of the lysimeter........................................................................35 3-5 Lysimeters inside the soil pit during installation.....................................................36 3-6 Soil profile inside the lysimeter...............................................................................38 3-7 Schematic of the funnel shaped bottom of the lysimeter.........................................41 3-8 Elevation diagram of the lysimeter..........................................................................42 3-9 Schematic of the water-splitter.................................................................................43 3-10 Instrumentation for the lysimeter.............................................................................44 3-11 Research field during pepper season (fall 2003)......................................................47 3-12 Schematic diagram of the bed-geometry..................................................................49 3-13 Soil water storage in lysimeters during fall 2003.....................................................53 3-14 Bi-weekly ETc for bell pepper during fall 2003.......................................................54 3-15 Bi-weekly ETc for drip irrigated lysimeters for bell pepper (fall 2003)...................55 3-16 Bi-weekly ETc for seepage lysimeters for bell pepper (fall 2003)...........................56 3-17 Bi-weekly crop coefficient for bell pepper using FAO-Penman-Monteith and Blaney-Criddle equations.........................................................................................58 3-18 FAO-Penman Montieth based bi-weekly Kc curve for bell pepper for drip and seepage lysimeters....................................................................................................59 x

PAGE 11

3-19 FAO-Blaney-Criddle based bi-weekly Kc curve for bell pepper for drip and seepage systems........................................................................................................60 3-20 Soil water storage in the lysimeters during spring 2004..........................................63 3-21 Bi-weekly ETc for watermelon during spring 2004.................................................64 3-22 Bi-weekly ETc for drip and seepage lysimeters for watermelon (spring 2004).......65 3-23 Bi-weekly crop coefficient for watermelon using Penman-Monteith and Blaney-Criddle equations......................................................................................................66 3-24 Bi-weekly crop coefficient for watermelon using Penman-Monteith and Blaney-Criddle equations under drip irrigation system........................................................67 3-25 Bi-weekly crop coefficient for watermelon using Penman-Monteith and Blaney-Criddle equations under seepage irrigation system..................................................68 4-1 Nitrate-N level in soil before and after crop season.................................................86 4-2 Average nitrate loading in drainage and runoff in the drip and seepage system during spring 2003 and fall 2003 seasons................................................................87 4-3 Dissolved inorganic nitrogen in the soil before and after crop season.....................89 4-4 Average DIN loading in lysimeter drainage and runoff for the drip and seepage system during spring 2003 and fall 2003 seasons....................................................90 4-5 Nitrate concentration in groundwater (mg/l)............................................................94 4-6 Ammonium concentration (mg/l) in groundwater...................................................96 4-7 Nitrogen crop removal during fall 2003 season.......................................................97 4-8 Extractable soil P status for spring 2003 and fall 2003 seasons.............................102 4-9 Total P loading in lysimeter discharge in spring 2003 and fall 2003 seasons.......103 4-10 Total P removed with harvest crop in fall 2003 season.........................................105 4-11 Total P concentration in groundwater (mg/l) and daily rainfall (mm)...................106 C-1 Water table for the drip and the seepage system and rainfall during fall 2003......119 C-2 Water table for the drip and the seepage system and rainfall during spring 2004.120 xi

PAGE 12

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for the Degree of Master of Engineering CROP COEFFICIENTS AND WATER QUALITY FOR WATERMELON AND BELL PEPPER UNDER DRIP AND SEEPAGE IRRIGATION By Saurabh Srivastava May 2005 Chair: Sanjay Shukla Cochair: Kenneth L. Campbell Major Department: Agricultural and Biological Engineering Agriculture is the single largest user of fresh water in Florida. It has also been recognized as the leading source of pollution of surface and groundwater bodies. Due to concerns of contamination and shortage of water, state agencies in Florida are encouraging the adoption of better water and nutrient management practices. Vegetable production, mainly grown in the southwest region, constitutes a major part of the state’s agriculture. This study was designed to develop local crop coefficients (Kc) for watermelon and bell pepper to aid in better irrigation scheduling, and compare the impacts of the drip and seepage systems on water quality. Six drainage lysimeters (4.87 m 3.65 m 1.37 m) were constructed and installed in a vegetable field at the University of Florida-Institute of Food and Agricultural Sciences (UF/IFAS) research center at Immokalee, Florida. Four lysimeters were irrigated by drip and two by seepage irrigation system. Inputs and outputs for the water xii

PAGE 13

and nutrient (nitrogen and phosphorus) budget were measured individually for each lysimeter. Bi-weekly and monthly Kc for bell pepper (fall 2003) and watermelon (spring 2004) were developed using FAO-Penman Monteith (FAO-PM) and FAO-Blaney Criddle (FAO-BC) based ETo methods. Monthly Kc curves for both ETo methods, conformed better to the conventional crop curves than the biweekly curves. Very high correlation (R2 > 0.99) for watermelon Kc curves compared to bell pepper curves was due to fewer rainfall events during spring 2004. Results indicated that rainfall resulted in increased evaporation losses, introducing uncertainties in the water budget. The use of drip irrigation resulted in 20% and 31% reduction in crop evapotranspiration of bell pepper and watermelon respectively, compared to that under the seepage system Fertigation and lower mass of infiltrating water in the drip system resulted in 40% and 45% lower nitrate and dissolved inorganic nitrogen (N) loading respectively, than in the seepage system. Results also indicated that the drip system had lower denitrification losses than the seepage system. Contrary to N, total phosphorus (P) discharge was 87% higher from the drip system than the seepage system. Higher P discharge from the drip system was likely due to cyclic downward movement of water. Less than 5% of the fertilizer P was lost in drainage and runoff discharge from the lysimeters, which showed that the bulk of the fertilizer P was stored in the soil at the end of the crop season. xiii

PAGE 14

CHAPTER 1 INTRODUCTION Background Florida has been endowed with abundant water resources comprising over 1700 streams and rivers, 7800 fresh water lakes and annual rainfall of 1145-1520 mm (Marella, 1999). However, with a population growth rate of nearly 23% (BEBR-UF, 2001) and blooming economic development, demand for water is increasing continuously. Even with its vast resources, water is in short supply in the state. To make matters worse, contamination from the industrial and the agricultural activities are making the surface and groundwater resources unusable. Conserving water and preserving its quality are two challenges faced by the state. Agriculture is the single largest user of fresh water in the state, accounting for 45% of total fresh water withdrawals in 1995 (Marella, 1999). Vegetable production constitutes a large part of Florida’s agriculture industry. Sub-tropical climate in south Florida makes the area conducive for vegetable production. Main vegetable crops grown in the region are tomato, pepper, watermelon and eggplant. Vegetable production in the state is carried out on highly sandy soils that are characterized by low water holding capacity and organic matter content. Water and nutrients can easily be lost from these soils. Therefore, vegetable production in the state utilizes raised soil beds covered with plastic mulch. These beds help in conserving the soil moisture and reduce nutrient losses. Although Florida receives large amounts of rainfall annually, nearly 70% of the total is received during the non-growing season of June-October. Temporal variability coupled 1

PAGE 15

2 with spatially variable nature of rainfall makes irrigation a necessity for state’s agriculture. While under-application of water can lead to plant stress and increase the salinity of soil, over-application leads to wastage of water and leaching of nutrients from the root zone. Sound irrigation scheduling and use of efficient irrigation systems is the key for optimum plant growth which can also help in conserving the water quantity and quality. To develop an effective irrigation management strategy, it is important to estimate the crop water use. Knowledge of crop coefficient (Kc) is essential for the estimation of water use. It helps in determining the water requirement of the crops according to their growth stage and environmental factors. Kc is the ratio of crop evapotranspiration (ETc) and reference evapotranspiration (ETo). While ETo is estimated by using weather parameters, ETc is measured by conducting lysimeter experiments. Studies have found that Kc for the same crop may vary from place to place based on factors such as climate and soil evaporation (Allen et al., 1998; and Kang et al., 2003). Doorenboss and Pruitt (1977) and Kang et al. (2003) emphasized the need to develop local Kc for accurate estimation of water use, under a specific climatic condition. Studies conducted over the years have developed local Kc for tomato, strawberry (Clark et al., 1996) and blueberries (Haman et al., 1997) under the warm and humid climate of Florida. However, local Kc for crops such as watermelon and bell pepper need to be developed. While Kc is effective for scheduling irrigation for micro-irrigation systems, it may not be useful for seepage irrigation system which is primarily concerned with maintaining a water table to replenish crop water needs.

PAGE 16

3 Seepage system is the most prominent irrigation system used in Florida (Smajstrla and Haman, 1997). Seepage irrigation supplies water to the root zone through capillary movement of water (upflux) from a high water table of 45-60 cm (referenced to the plastic mulch bed). Seepage system has low application efficiency of 30 to 40% (Stanley and Clark, 1991) and requires considerably large amount of water to keep the optimum soil moisture. Considerable part of irrigation in the seepage system is lost to deep percolation, runoff and subs-surface lateral movement of water. Loss of water is accompanied with the loss of nutrients from the root zone. Majority of the nutrients which move away from the root zone eventually find their way into surface or groundwater system. On the other hand, drip system applies water directly to the root zone with high efficiency, thereby minimizing the water loss. Studies have shown that the drip system reduces the water use of tomato by 50% compared to the seepage system in Florida (Pitts and Clark, 1991). Moreover, drip system provides the opportunity to apply fertilizer mixed with irrigation water, on as needed basis through fertigation. Split application of fertilizer compared with a single application, can lower nutrient leaching (Singh and Sekhon, 1976). Through fertigation, drip system can reduce potentially leachable fertilizer in the beds at any given time compared to the seepage system. Therefore, fertigation can potentially reduce the transport of nutrients away from the root zone. Limited research has been undertaken to quantify the benefits of drip system over the seepage system. As a result, adoption of the drip system has been limited in Florida. Unless the benefits of the drip system are validated for a variety of locally grown crops, its acceptance by farming community will be restricted.

PAGE 17

4 Field scale studies have been used in the past to study nutrient transport (Pier and Doerge, 1995; Aceves and Dendooven, 2001; Lamm et al., 2001). However, one of the main sources of uncertainty for a field scale experiment in south Florida is due to the highly interactive surface and groundwater systems. Difficulty in isolating a research plot from the effects of its surrounding fields can result in exchange of dissolved nutrients from one field to another. Lack of definite boundary conditions poses challenge to any attempt of quantifying the effectiveness of irrigation and nutrient best management practices (BMPs). Shortcomings of a field scale experiment can be overcome by the use of lysimeters. Lysimeters are containers used to study the optimization of water management for any crop if they are adequately designed to approximate the physical system (Chow, 1964). Lysimeters provide a direct estimation of ETc (Clark et al., 1996; Steele et al., 1996; Haman et al., 1997; Simon et al, 1998) which is used to develop Kc. Moreover, recent literature shows that lysimeters are increasingly being used in the water quality studies (Prunty and Montgomery, 1991; Martin et al., 1994; and Liaghat and Prasher, 1996). Therefore, a lysimeter is a better tool than field experiment to develop Kc and study the nutrient transport in south Florida. Goals and Objectives The goal of this study was to quantify the differences in ETc and nutrient leaching under drip and seepage system for watermelon and bell pepper for south Florida conditions. The specific objectives were 1. To quantify ETc and develop Kc for the two crops using drip and seepage irrigation systems. 2. To quantify nitrogen and phosphorus leaching under the drip and seepage irrigation systems.

PAGE 18

CHAPTER 2 REVIEW OF LITERATURE Florida is ranked among the top 10 agricultural production states in the nation (Marella, 1999). Fruits, vegetables, field crops, ornamentals and cattle are its main agricultural products. The vegetable production industry in Florida is ranked second after California in terms of acreage of fresh vegetables (United States Department of Agriculture-National Agricultural Service [USDA-NASS], 2003). The agricultural sector relies heavily on natural water resources to meet its demands, accounting for nearly 45% of the total fresh water withdrawals in 1995 (Marella, 1999). However, with a growing demand of water, the resources are beginning to get stressed to a point where withdrawals are more than the recharge. Improper irrigation management and inefficiencies in irrigation systems can result in considerable wastage of water. The problem of water shortage is compounded due to the increased pollution of the resources. United States Environmental Protection Agency (USEPA, 2002) recognized agriculture as the primary cause of pollution of water resources in the USA. Nitrate leaching to groundwater and phosphorus loading in runoff are of special concern in terms of pollution from agriculture. Science based agricultural water management is key to addressing the problems of water shortage and pollution. This study was designed to develop effective irrigation and nutrient best management practices (BMPs) for Florida. Literature pertaining to impact of agricultural water management on water use and water quality was reviewed to support the objectives of this study. 5

PAGE 19

6 Impact of Agricultural Water Management on Water Use Irrigation System Smajstrla and Haman (1997) reported that nearly 43% of the total irrigated land in Florida utilizes seepage irrigation system, while about 30% and 25% of irrigated land is under sprinkler and drip system respectively. The seepage system, which has very low application efficiency of 30 to 40% (Stanley and Clark, 1991), supplies water through open ditches to maintain high water table. The water table supplies moisture in the root zone through upward movement of water. However, sandy soils in Florida have macro pores and tend to lose water very quickly. As a result, they have low plant available water (PAW). PAW is the amount of water between the permanent wilting point (WP) and field capacity (FC). WP is the soil moisture content at which plants can no longer extract water to fulfill their transpiration needs (Michael, 2001). FC is the moisture content held by soil after additional water is drained by gravitational force (Michael, 2001). To achieve proper plant growth, it is imperative to keep the soil moisture content between FC and WP. To avoid moisture stress, continuous supply of water is maintained under seepage system. This increases soil moisture in the non-cropped areas causing high evaporation losses and consequently higher water use. In contrast to the seepage system, the drip irrigation system has a high application efficiency of nearly 85% (Simonne et al., 2003) and can apply water directly to the root zone. Allen et al. (1998) observed that use of drip system under plastic mulch beds can substantially reduce the evaporation (Ea) losses from exposed row middles. Associated with the reduced Ea, there is an increase in transpiration (Tp), caused by the transfer of both sensible and radiative heat from the surface of plastic cover to the adjacent

PAGE 20

7 vegetation. Even with 10-30% increase in Tp, the crop coefficients (Kc) decrease by 10-30%, due to significant (50%) reduction in soil Ea (Allen et al., 1998). Effectiveness of the drip irrigation system over the seepage system was demonstrated by Pitts and Clark (1991). They reported about 50% reduction in the water use of tomato under plastic mulch conditions in Florida, when irrigated by drip irrigation system compared to seepage system. Smajstrla et al. (2000) also reported 36% less water use for potato grown under the drip system compared to the seepage system, without affecting the crop yield. Similar study conducted in India by Singandhupe et al. (2003) reported 31-37% reduction in the water use of tomato using the drip system compared to the gravity flow furrow irrigation. Some studies have even reported that compared to the surface irrigation systems, drip system is better for the plants in terms of their morphological growth (Antony and Singandhupe, 2004) and can result in higher yield (Yohannes and Tadesse, 1998; Cetin and Bilgel, 2002). Irrigation Scheduling While every irrigation system has its advantages and limitations, irrigation scheduling is the most critical factor for the application of right amount of water to the crops. Irrigation scheduling is the decision of frequency, time, and amount of irrigation applied to a crop. Different methods such as hand feel, gravimetric method, water budget, tensiometers and real time soil moisture monitoring (Broner, 2001) are used for scheduling irrigation. While most of these methods are highly labor intensive and need expensive instruments, the hand feel method is the most simple and convenient irrigation scheduling method. But hand feel method is fraught with uncertainties and can easily result in improper irrigation. Hill and Allen (1996) suggested the use of calendar method based on long term weather data for irrigation decision. In a similar study Hill et al.

PAGE 21

8 (1984) used computer based scheduling calendars for U.S. and Pakistan. Their model, known as the crop yield and soil water management simulation model (CRPSM), amalgamated the influence of weather conditions on plant growth stage progress and yield as influenced by soil water content. However, both the models discussed above failed to consider the effect of climatic variability on daily water use. Florida has highly variable climatic conditions due to the cloud cover and random rainfall (Smajstrla et al., 1997) and application of calendar method can be questionable as it may result in over or under irrigation on a hot and humid or cold and overcast day respectively. Soil moisture based irrigation scheduling involves using moisture measurement devices to schedule irrigation. Tensiometers and resistance blocks have been used for this purpose. Stenitzer (1996) presented the use of gypsum blocks or granular matrix sensors (GMS) for irrigation scheduling. But, one of the main challenges in using such devices is their installation at locations which are representative of entire field. Moreover, most of these devices have a small sensor (e.g. Watermark GMS = 6 cm; tensiometers = 6 cm) and therefore fail to capture the soil water status of the entire soil profile. One of the most reliable methods of irrigation scheduling is based on soil water status and water budget knows as ET-based irrigation scheduling. This method is based on predicting the crop water use based on crop growth. Crop water use can be defined as the amount of water utilized by a crop during a growing season. It includes the water lost via soil Ea and Tp. Tp constitutes nearly 99% of total water taken up by the plants as they retain less than 1% of the total uptake (Hillel, 1997). The combination of Ea and Tp is termed as evapotranspiration (ET). The ET takes into account factors such as climate, crop, management and environmental conditions (Allen et al., 1998). By combining the

PAGE 22

9 effect of these factors, ET-based irrigation scheduling provides the most comprehensive method of estimating crop water use and scheduling irrigation. Evapotranspiration Ea and Tp are the two most important processes governing removal of water from the land. These processes occur simultaneously, and are hard to be distinguished from each other (Allen et al., 1998). Stanhill (1973) found considerable interaction between the two processes. The term ET was coined to define the total loss of water from an area. While occurring simultaneously, Ea is governed by the availability of water in the topsoil and the fraction of solar radiations reaching soil surface. Amount of solar radiation reaching soil surface varies with the nature of crop canopy. Tp on the other hand is a function of crop canopy and soil water status. Ea has been found to dominate the ET by as much as 100% during early stages of crop growth while Tp contributes to nearly 90% of the ET for a fully matured crop (Allen et al., 1998). Liu et al. (1998) reported that soil Ea constitutes nearly 30% of the total field ET. Similar study by Kang et al. (2003) found that Tp accounted for 67% and 74% of seasonal ET for wheat and maize respectively, grown under semi humid conditions. ET can be classified into two types: reference evapotranspiration (ETo) and crop evapotranspiration (ETc) (Allen et al, 1998). Reference evapotranspiration (ETo) ETo is a representation of the Ea demand of atmosphere, independent of crop growth and management factors (Allen et al., 1998). It is a climatic parameter determined by the weather data. Allen et al. (1994) define ETo as “the rate of ET from a hypothetical reference crop with an assumed crop height of 0.12 m, a fixed surface resistance of 70 sec m-1 and an albedo of 0.23, closely resembling the evapotranspiration from an extensive surface of green grass of uniform height, actively growing, well-watered, and completely

PAGE 23

10 shading the ground.” ETo determines the loss of water from a standardized vegetated surface, which helps in fixing the base value of ET specific to a site. ETo can be measured in-situ by measuring the open water surface evaporation from an evaporation pan. Open water Ea incorporates the effects of temperature, humidity, wind speed and solar radiation. Pan evaporation coupled with the use of a calibrated pan coefficient (Kp) to relate Ea with the standard vegetative surface, can provide good estimates of ETo, provided that soil water is readily available to the crop (James, 1988). Some of the commonly used pans are: Class-A Evaporation pan and Sunken Colorado pan. However, pan evaporation method requires regular maintenance of the evaporation pan and the vegetation around it. Also, unavailability of local pan coefficient can limit the accuracy of ETo estimates. Alternatively ETo can be estimated from meteorological data using empirical and semi-empirical equations. Jacobs and Satti (2001) identified three approaches commonly used for estimating ETo from meteorological data: 1. Temperature based approach: Methods based on this approach use air temperature as the surrogate for the amount of energy available for ETo. Lack of a direct relation between the air temperature and radiation limits the accuracy of these methods. However, local calibration can improve their estimations. Common temperature based methods are McCloud method, Thornthwaite, and Blaney-Criddle methods. 2. Radiation based approach: These methods compute ET based on the amount of radiation energy reaching the surface (Allen et al., 1998). Commonly radiation based methods include Priestly Taylor, Turc and Hargreaves methods. 3. Combination of temperature and radiation based approach: Radiation energy converts water into vapor but the driving force to remove vapor from the evaporating surface is the difference of saturation pressure between air and the evaporating surface. Saturation pressure depends on the relative humidity of the air. Wind velocity also plays a part in ET by replacing saturated air with drier air to keep the process in continuation. Penman (1948) presented a combination approach which includes radiation and aerodynamic terms. This approach requires extensive meteorological data including radiation, air temperature, wind velocity and relative

PAGE 24

11 humidity. Lack of availability of such extensive data can limit its application. Commonly used combination approach methods are based on the Penman 1948 equation with numerous modifications by different researchers over the years. Several studies have been conducted over the years to evaluate the accuracy of different ETo methods. Most of these studies have concluded that Penman-Montieth equation in its different forms provides the best ETo estimates under most conditions (Ventura et al., 1999; Kashyap and Panda, 2001). Therefore, the Food and Agricultural Organization (FAO) recommended FAO-Penman Monteith (FAO-PM) method as the sole standard method for computation of ETo (Allen et al., 1998). FAO-PM can provide accurate ETo estimates for weekly or even hourly periods. However, the Blaney-Criddle method in its different forms is commonly used by the water management districts in Florida for the purpose of water allocations. However, Nichols et al. (2004) noted that the inability of SCS Blaney-Criddle method to account for the processes which govern the Ea rate result in overestimation of ET. Crop evapotranspiration The actual crop water use depends on climatic factors, crop type and crop growth stage. While ETo provides the climatic influence on crop water use, Tp component of water use as affected by crop and management factors is described by ETc. Factors affecting ETc such as ground cover, canopy properties and aerodynamic resistance for a crop are different from the factors affecting reference crop (grass or alfalfa); therefore, ETc significantly differs from ETo. The characteristics that distinguish field crops from the reference crop are integrated into a crop factor or crop coefficient (Kc) (Allen et al., 1998). Kc is used to determine the actual water use for any crop in conjunction with ETo (Equation 2-1).

PAGE 25

12 occETKET Equation 2-1 Crop coefficient The crop coefficient (Kc) is computed as the ratio of reference and crop ET (Equation 2-1). Allen et al. (1998) listed the factors affecting Kc as crop type, crop growth stage, climate and soil evaporation. Kc is most commonly expressed as a function of time (Clark et al., 1996; Haman et al., 1997). However, Kc as a function of time does not take into account environmental and management factors that influence the rate of canopy development (Grattan et al., 1998). Therefore, most researchers have reported Kc as a function of days after transplanting (DAT) which helps to reference Kc on crop development stage (Allen et al., 1998; Tyagi et al., 2000; Kashyap and Panda, 2001; Sepashkah and Andam, 2001). Recent literature also describes Kc as a function of cumulative growing degree days (CGDD) (Steele et al., 1996; and Sepashkah and Andam, 2001) and the canopy cover (Grattan et al., 1998). Accurate prediction of crop water use is the key to develop efficient irrigation management practices making it imperative to develop Kc for a specific crop. Numerous studies have been conducted over the years to develop the Kc for different agricultural crops. Since most of the studies have been specific to one or two crops, Doorenbos and Pruitt (1977) prepared a comprehensive list of Kc for various crops under different climatic conditions by compiling results from different studies. Similar list of Kc was also given by Allen et al. (1998) and Doorenbos and Kassam (1979). Use of standard ETo has made it easier to transfer the Kc from one location to another (Irmak and Haman, 2003). However, Kc for a crop may vary from one place to another depending upon climate, soil, crop type, crop variety, irrigation methods etc (Kang et al., 2003). Thus, for an accurate

PAGE 26

13 estimation of the crop water use, it is imperative to use a local Kc. Doorenbos and Pruitt (1977) and Kang et al. (2003) have emphasized the need for local calibration of Kc under a given climatic conditions. Therefore, the reported values of Kc should be used only in places where local data are not available. In summary, there is a need to develop local Kc for a realistic estimation of water use to better schedule irrigation. Estimation of Kc Brouwe and Heibloem (1986) outlined the steps for development of Kc as: determination of total growing period of the crop, identifying the length of different growth stages, and determination of Kc values for each growth stage. However, Kc cannot be measured directly, but is estimated as a ratio (Equation 2-1). While ETo can be estimated using one of several available methods, ETc can be estimated by a lysimeter study or by Bowen Ratio energy balance method (Gratten et al., 1998). Latter method incorporates the ratio of sensible heat exchange and latent heat exchange. It uses measurements of wind speed and humidity at two levels (Dingman, 1993). However, need of elaborate instrumentation has limited its use to research settings (Dingman, 1993). Alternatively, ETc is measured by growing crops in containers known as lysimeters. A lysimeter is essentially a container which isolates soil and water hydrologically from its surroundings, but still represents the adjoining soil as closely as possible. Lysimeters can be used as a research tool to study plant-water relationships if they are designed adequately to approximate the physical system (Chow, 1964). Lysimeters provide a controlled soil-water or nutrient environment system for precise measurement of water and nutrient use and movement (Chalmers et al., 1992). Aboukhaled et al. (1982) identified two types of lysimeters:

PAGE 27

14 1. Weighing lysimeter: It measures the ETc based on the difference in weight of the lysimeter over a period of time. It uses a balance beam and a counter weight mechanism or load cell to measure the weight of lysimeter. It provides the most accurate measurement of ETc. (McFarland et al., 1983; Sharma, 1985; Meshkat et al., 1999). 2. Non-weighing or drainage lysimeter: These lysimeters are used to estimate ET by computing water balance. Water balance involves measuring all the water inputs and outputs to the lysimeter and the change in soil moisture over a stipulated period of time. These lysimeters a provide viable estimate of ETc for longer periods such as weekly or monthly. Tyagi et al. (2000) used weighing lysimeters to estimate hourly ETc for wheat and sorghum. Kashyap and Panda (2001) measured daily ETc for potato using weighing lysimeters. Several other studies have also utilized weighing lysimeters to estimate ETc (Liu et al., 1998; Tolk et al., 1998 and Zang et al., 2004). However, one of the major challenges in using a weighing lysimeter is the difficulty to accurately measure small change in weight of the lysimeter in relation to the large and heavy soil mass. Therefore, the weighing mechanism needs to be accurately calibrated as small errors can translate into large errors in ET computation. Moreover, these lysimeters are very costly, complex in design and difficult to maintain (Smajstrla, 1985). On the other hand, drainage lysimeters are simpler in design and less costly to construct, as surrounding soil can be used to support the lysimeter (Aboukhaled et al., 1982). In some studies, the drainage lysimeters have been reported to be as accurate as the weighing lysimeters. Oad et al. (1997) studied the accuracy of weighing and drainage lysimeter against a large weighing lysimeter, used as the control. Using six small lysimeters of both types, they found that the weighing lysimeters slightly underestimated ET while the drainage lysimeters overestimated ET. However, they concluded that ET estimates from the two types of lysimeters were not statistically different.

PAGE 28

15 Design considerations for lysimeters One of the most important factors controlling the accuracy of a lysimeter is its size (Gangopadhyaya et al., 1966). Clark and Reddell (1990) noted that the lysimeter surface area and its depth should be large enough to minimize root restrictions. On the effectiveness of miniature lysimeters (10 cm diameter and 10 cm deep), Gangopadhyaya et al. (1966) reported that these lysimeters were “sensitive” but not reliable due to distortions in thermal properties. They concluded that accuracy of lysimeters increase with an increase in their surface area. Boast and Robertson (1982) reported that shallow lysimeters tend to retain more water per unit depth than the actual field and thus introduce a bias by overestimating ET. Yang et al. (2000) reported that groundwater evaporation contributes up to 56% of total ET. Therefore, authors suggested that lysimeters measuring ET should be deep enough to account for soil-water and groundwater exchanges and water table fluctuations. Another debatable topic concerning design of lysimeters is the use of a rain shelter. To avoid unwanted water from entering the lysimeter system via precipitation, rain shelters have been employed at some of the lysimeter sites around the world. By keeping unwanted rainfall away from the system, rain shelters reduce the uncertainty in ET estimation especially, during the times soon after rainfall when extremely wet soil conditions trigger high ET rate. However, their use in field studies also has attracted some criticism. Dugas and Upchurch (1984) reported that the sides of rain shelter can restrict the wind movement under the shelter causing excessive heat. Authors further noted that rain shelter lowered the radiation reaching the plants by 30 40%. Clark and Reddell (1990) noted that permanent rain shelters excessively heated the crop due to improper ventilation.

PAGE 29

16 Lysimeter-based crop coefficients Lysimeters have been successfully used by researchers to measure the ETc and develop Kc for various fruits and vegetables (Clark et al., 1996; and Haman et al., 1997) and field crops (Steele et al., 1996; Simon et al, 1998; Tyagi et al., 2000). Tyagi et al. (2000) estimated Kc for wheat and sorghum for the semi-arid climatic conditions in India by using a weighing lysimeter. Using a lysimeter with effective surface area of 4 m2, Tyagi et al (2000) developed stage-wise Kc for both the crops using four methods of ETo estimation. They also compared their estimated Kc for the two crops with the FAO published Kc (Doorenboss and Pruitt, 1977). Authors reported that their estimated Kc matched with reported values at maturity stage. Kashyap and Panda (2001) used small weighing lysimeters (75 cm diameter; 75 cm depth) to develop Kc for potato for semi-humid climate in India. Using FAO-PM method to develop Kc, they reported differences in their developed Kc and the reported values. Steele et al. (1996) developed mean crop curves for corn as a function of DAT and CGDD based on Jensen and Haise (1963) and modified Penman equation (Allen, 1986) ETo methods. Using 11 years of data from four drainage lysimeters, they developed fifth order crop curves for corn using both ETo methods. Crop curves based on DAT were found to be more accurate with a higher r2 and lower RSME than those based on CGDD. Table 2-1. Statistical results for corn crop curves from 11 years of data. Jensen-Haise (ETo) Penman-Allen (ETo) Kc Statistics DAT based a CGDD based b DAT based CGDD based r2 0.679 0.544 0.622 0.521 Source: Steele et al. (1996); a DAT is days after transplanting; b CGDD is cumulative growing degree days.

PAGE 30

17 The authors explain that lower r2 (Table 2-1) were due to the uncertainties in measurements from the drainage lysimeters. They revealed that the lack of accuracy in determining soil moisture, measured by neutron attenuation method, was the most important source of variability in their study. They noted that lack of monitoring of soil moisture at the bottom 0.3 m region of the lysimeter added uncertainty in the results. Another complicating part of their study was negative Kc for periods when lysimeters were drained after rainfall. Authors did not discuss the reasons for negative Kc, but, they noted that it can be avoided by increasing the time step for estimating ETc to two or more periods (each water balance period in their study was 10 days). Steele et al. (1996) also found that Kc should be referenced to the middle of a time step (t) for periods longer than daily such as weekly, bi-weekly or monthly periods. They noted that referencing Kc to the beginning or end of the growing period can alter the shape, amplitude and position of the crop curve significantly, thereby, reducing its accuracy. Haman et al. (1997) used drainage type lysimeters to study ET and develop Kc for two varieties of young blueberries for Florida. The authors used cylindrical tanks as lysimeters (1.6 m diameter and 1.8 m deep) equipped with porous plates to extract drainage water. The ETc in their study included Tp and Ea from the surface wetted by the irrigation system, but did not include water loss from the grassed alleys. They noted that their computed Kc was different from the standard Kc, but it provided information for actual crop water use. Although Kc for both the varieties followed the same general trend, Kc values for the two varieties were different from each other. Differences in Kc values of the two varieties were attributed to the differences in plant development of the two varieties.

PAGE 31

18 Clark et al. (1996) used drainage lysimeters to compute ETc and develop Kc for drip irrigated strawberry in Florida. They used 16 drainage lysimeters 2.4 m 0.6 m 0.6 m equipped with rain shelters for their study. Since drip irrigation applies water directly to the root zone, actual crop water use can be different from the seepage irrigation system which has high water table and wet row middles. To study differences due to high water table and wet row middles, Clark et al. (1996) used two types of plant arrangements: the first arrangement estimated ETc only from the plants while the second estimated ETc from the plants and the exposed row middles. They reported monthly Kc based on modified Penman (PENET) (Burman et al., 1980), modified Blaney-Criddle (BCRAD) (Shih et al., 1977) and pan evaporation (PANET) (Doorenboss and Pruitt, 1977). Their results indicated that for lysimeters with plant and exposed row middles, ETc and Kc were higher than those only the plants. They estimated that 25-35% of ETc was Ea from exposed row middles. Using linear regression, they observed high r2 for their Kc curves (PENET =0.97, PANET = 0.94, BCRAD = 0.94). They recommended that Kc developed from their study was useful for irrigation scheduling and developing water budgeting procedures for drip irrigated strawberry production in a humid region. Simon et al. (1998) conducted a study to develop local Kc for maize in Trinidad. They used 2 m 2 m 1.2 m drainage lysimeter for three seasons to develop Kc. The effects of dry and wet season (temporal variability of climate) on Kc were also discussed. They found that Kc during a wet season (Kc =1.13 to 1.41) was greater than during a dry season. (Kc = 0.73 to 0.94). They attributed the differences between the wet and dry season Kc to lower ETo during the wet season. Mean Kc for maize was found to be greater

PAGE 32

19 than the reported values by Doorenboss and Pruitt (1977). Therefore, authors stressed on the importance of developing local Kc for accurate irrigation scheduling. Sepaskhah and Andam (2001) used drainage lysimeters to estimate Kc for sesame for semi arid regions of Iran. They developed Kc based on modified Penman-Monteith (Jensen et al., 1990) and FAO-PM, as a function of CGDD and DAT. Authors reported that their observed Kc was different from those given by Doorenboss and Pruitt (1977) and Allen et al. (1998) for similar crops. In a similar study, Lie et al. (2003) used cylindrical drainage lysimeter (diameter = 1 m; depth = 0.8 m) to develop Kc for watermelon and honey dew melons in China using ETo from a pan evaporation. Their reported Kc for watermelon varied from 0.35-2.43. These values were considerably higher than the Kc (0.4-1.0) as reported by Allen et al. (1998). A study by Kang et al. (2003) reported Kc for wheat and maize for semi-humid conditions of northwestern China. They used three 3 m 2 m 2 m drainage lysimeters equipped with rain shelters. An average Kc was developed from 10 years of measured data. Although, their Kc matched well with the Kc given by Doorenboss and Pruitt (1977) during the initial growth period for both the crops, it was higher during the mid and late season. Impacts of Agricultural Water Management on Water Quality Nitrogen (N) and phosphorus (P) are the essential elements for plant growth making application of N and P fertilizer inevitable for crop production. A higher than optimum rate of fertilizer and improper water management practices can result in the movement of N and P into surface and groundwater bodies. Agriculture is the leading source of non point source pollution (NPS) in United States (USEPA, 1997). Water is the main driving force and carrier of NPS pollutant (Zamies and Schultz, 2002). Infiltrating water can carry nutrients vertically downward to the groundwater, while runoff can

PAGE 33

20 transport the nutrients to the surface water. Therefore, water management is the key to minimize movement of nutrients from their point of application into a water body. Soils in Florida have a high sand content. These soils due to low water holding capacity can lose water very rapidly. Therefore, there is always a risk of N and P fertilizer moving out of the root zone. Moreover, rainfall in Florida can easily bring the shallow water table to the surface producing runoff, which can contribute to the NPS. Precise water management is very important under these conditions. Agricultural water management entails proper irrigation scheduling and use of an efficient irrigation system. The importance of irrigation scheduling has been discussed previously. This section discusses the impacts of irrigation systems on nutrient management. To understand the fate of N and P transport in the soil, it is important to develop an understanding of the processes which drive their movement in the soil. Nitrogen Transport N is present in the soil in organic and inorganic forms, with the former as the largest pool of N in the soil. While organic N is slow moving, inorganic N especially nitrate (NO3) does not adsorb to soil particles and readily moves through the soil (Owen, 1990). Processes that govern movement of N in soil are: Mineralization-immobilization: Mineralization is the conversion of organic N into NH4. Immobilization is the opposite of mineralization. These two processes occur simultaneously in soils (Shukla, 2000). Net direction of this transformation is dependent on the ratio of organic carbon to N (C : N). But due to complexities, estimation of net turnover from the mineralization-immobilization process is one of the main challenges while computing the N mass balance (Jansson and Pearson, 1982).

PAGE 34

21 Nitrification: It is the conversion of NH4 into nitrite (NO2) and its further decomposition into nitrate (NO3) (Shukla, 2000). Nitrification is considered important since it converts the relatively immobile NH4 into NO3, which is preferred by the plants but also has the potential to readily move through the soil into groundwater. Denitrification: It is the process of oxidation of NO3 into gaseous N. It is considered an important source of N loss in shallow poorly drained soils. Shukla (2000) noted that quantification of denitrification is fraught with uncertainties due to the difficulties in its measurement and incomplete understanding of the process itself. Fixation: The conversion of atmospheric N into organic form by the leguminous crop and soil micro-organisms is called N fixation. Volatilization: It is the conversion of fertilizer NH4 into gaseous ammonia (NH3). Application of fertilizer under the plastic mulch reduces volatilization losses. The vegetables are traditionally irrigated under seepage irrigation in Florida. One of the major disadvantages of the seepage system is the pre-plant application of fertilizer in the plant beds. Fertilizer N present in the bed is susceptible to the fluctuation of shallow water table due to irrigation and rainfall. Once moved away from the root zone, it will eventually move into the surface or groundwater. Moreover, transportation of N away from the root zone can limit its availability to the plants. Low supply of N in soil affects plant growth, adversely affecting the yield (Syvertsen and Smith, 1996). Alternatively to the seepage system, the drip system has the ability of split applying the fertilizer on an as needed basis through fertigation. Split applications of fertilizer have been found an effective technique to reduce leaching losses and boost crop production. Singh and Sekhon (1976) observed lower nitrate leaching with split application of

PAGE 35

22 fertilizer. Pier and Doerge (1995) studied the N and water interactions of sub-surface trickle irrigation system for watermelon production as an alternate to the furrow irrigation system commonly used in Arizona. They observed a positive water and N interaction with a higher yield from trickle irrigated fields as compared to the conventional furrow irrigated watermelons. Watts and Martin (1981) used a field calibrated model to estimate the impacts of management practices on loss of NO3 from the root zone. They observed reduced NO3 leaching losses for plots with fertilizer application with irrigation water as compared to those with pre-plant application. They found that rainfall increased NO3 leaching loss. These losses were minimal from plots using injected fertilizer compared to plots with pre-plant fertilizer. The investigators reported that under normal seasonal rainfall of 317 mm, 30-35 kg/ha of NO3 loss will occur when 168 kg/ha of N is applied. A study by Phene et al. (1979) reported that fertigation increased the fertilization use efficiency of potato by 200% compared to that under conventional application methods. The use of drip irrigation as compared to the seepage system provides a viable option of minimizing N transport (especially NO3) from the root zone into groundwater. However, no study has been conducted to compare the impact of drip and seepage irrigation systems on the movement of N for vegetable production on sandy soils in Florida. Phosphorus Transport Although P is a macro nutrient for plants, it is a major concern for eutrophication. Compared to N, P is very slow moving and often stays in the soil for many years before being utilized. P losses from an agricultural field are typically below 1 kg P ha-1 yr-1 (Heckrath et al., 1995). While these losses are low with respect to the total P applied as

PAGE 36

23 fertilizer, they can have extensive environmental impacts. To understand the transport of P, it is important to understand the forms of P in soil and processes involved in P transport through the soil. P is present in soil in soluble, labile and fixed form (Larsen, 1967). The fixed P pool can constitute up to 80% of the total P present in the soil (Schachtman et al., 1998). It contains insoluble inorganic phosphate compounds and organic compounds resistant to mineralization. P mineralization converts organic P into inorganic forms. The fixed P is highly immobile and unavailable to the plants. Soluble P, the smallest pool of P in the soil, is the plant available form. It is continuously replenished by the labile P pool. The labile P is the solid form of P which is readily released into soil solution acting as the buffer pool for soluble P (Bushman et al., 1998). Soluble P can be in both inorganic and organic forms. The organic P, mostly considered as immobile, is derived from the plant residue and excreta of above and below ground organism (Haygarth and Jarvis, 1997). Fertilizer and weathering supply the inorganic P in the soil. The inorganic P is easily released into solution and is primarily considered labile (Zaimes and Schultz, 2002). P in soil moves through dissolution processes such as mineralization, adsorption–desorption, solubilization of P from saturated soils, and leaching (Haygarth and Sharpley, 2000). P can also move through physical mechanisms such as soil erosion, displacement and entrapment of soil colloid (Haygarth and Sharpley, 2000). Compared to the dissolution process, the physical processes of P transport are considered macro scale processes (Haygarth and Jarvis, 1997). Different hydrologic processes govern the movement of P in soil. Haygarth and Sharpley (2000) characterized the following pathways through which P is transported in soil.

PAGE 37

24 1. Matrix flow: Vertically downward movement of water through the macro and micro pores. 2. Preferential flow: Vertically downward movement through the cracks, worm holes and borrows in the soil. 3. Overland flow: Movement of water over the soil during rainfall events 4. Interflow: Lateral flow of water due to the slope of agricultural land 5. Drainage: Artificial removal of water to avoid logging conditions. The hydrologic processes discussed above depend upon water management techniques. The overland flow or runoff is the most important pathway for P in agricultural land (Sharpley et al., 1993). Matrix flow and preferential flow are more important in soil with high clay content. P losses through drainage and interflow are lower than those through overland flow (Zaimes and Schultz, 2002). However, Sharpley et al. (1994) reported that sandy spodosols can lose significant P through interflow. Controlled movement of water can minimize transport of P through the soil. Due to water table fluctuations in seepage irrigation, there is greater possibility of migration of P out from the root zone. Moreover, due to higher antecedent soil moisture in the non-cropped areas, the water table can easily move up to the soil surface and produce runoff. As discussed earlier, runoff is considered the major transport process for soil P. Drip irrigation has less water table fluctuations. While different studies have been conducted to quantify the P transport from agricultural land, there is a need to study the movement of P under different irrigation regimes. Moreover, no study has been conducted to compare the effect of cyclic top down movement of water under the drip system with drying and rewetting process under seepage irrigation, on P transport through sandy soils in Florida.

PAGE 38

25 Tools for Quantifying Nutrient Transport Various field studies have been conducted to study the movement of N and P, but these studies have their limitations. Due to the interaction between a research field and its adjoining fields, there is lack of definite boundary conditions. This introduces uncertainty in results from field scale studies. High water table conditions such as those in south Florida are a serious limitation for conducting field study. A lysimeter provides an alternate and effective tool to study the movement of chemicals through the soil profile (Martin et al., 1994). A lysimeter provides definite boundary conditions with absolutely no interaction with the outside soil. Such systems could be used as a model to understand the complex processes in the soil and consequently be used to develop better nutrient management practices for the actual field conditions. Prunty and Montgomery (1991) termed lysimetery as the most sophisticated and precise method to study the movement of chemicals from the root zone to the groundwater. One of the major concerns in lysimeter studies is due to the rebuilding of soil profile to closely approximate the actual conditions. Inability to do so introduces bias in the results and its transferability to model actual field conditions. Martin et al. (1994) identified two common methods of rebuilding lysimeter soil profile. 1. Using disturbed soil: It involves filling the lysimeters layer by layer with disturbed soil. It requires the soil to be removed in layers from the ground and then put back in layers inside the lysimeters. 2. Using un-disturbed soil: It involves digging up the whole soil profile as a monolith block from the ground and placing it inside the lysimeter without affecting the structure of the soil profile. The latter method provides the best representation of actual conditions, but it is labor intensive and costly. The first method is simpler and less costly, but can introduce bias in the results due to disturbing of natural soil. However, Martin et al. (1994) noted

PAGE 39

26 that such errors from using disturbed soil were more pronounced in aggregated and stratified soils and less in soils containing a high sand content. Prunty and Montgomery (1991) employed the use of reconstructed sandy soils in four drainage lysimeters 2.4 m 2.4 m 2.3 m, to study the effect of N fertilizer on the quality of recharge water. They found their results agreeable to the results from Coshocton lysimeter study (Owen, 1990), which used un-disturbed soil. While some lysimeters were buried below the ground surface to allow for tillage operations (Martin et al., 1994), Klocke et al. (1993) used lysimeters whose walls extended above the soil surface. Klocke et al. (1993) discussed design and performance of percolation lysimeters for water quality sampling. They used steel pipes of 0.90 m diameter and 2.4 m deep, as drainage lysimeters. Since native soil in the area of study in Nebraska was silt loam, they used an un-disturbed method of rebuilding the soil profile. Using soil water extractors to remove water from the soil, they used a reservoir at the bottom of the soil monolith to collect the leachate samples. The authors found that high rainfall caused significant variations in the leachate among the lysimeters. Martin et al. (1994) used two drainage lysimeters of 1.52 m 1.22 m 1.83 m to study the effect of reduced N application compared to conventional grower standard on the nitrate leaching under maize production. They observed more than 50% reduction in nitrate leaching from the reduced N application than the grower’s application rate. They concluded that drainage lysimeters can offer an effective tool to evaluate N management strategies for a field condition. In summary, literature review presented in this chapter indicated the need to: 1. Develop local Kc for watermelon and bell pepper to better schedule irrigation.

PAGE 40

27 2. Quantify the impacts of agricultural water management on water use and water quality of vegetable production in Florida.

PAGE 41

CHAPTER 3 CROP COEFFICIENTS AND CROP EVAPOTRANSPIRATION FOR WATERMELON AND BELL PEPPER UNDER DRIP AND SEEPAGE IRRIGATION Introduction Seventy percent of 1,370 mm annual rainfall (Fernald and Patton, 1984) in Florida is received during the summer, which is the non-cropping season in the southern part of the state. Therefore, agriculture in the state has to rely upon irrigation to meet the crop water demands. Proper scheduling of irrigation is not only vital for a good crop production but also helps in saving water by applying only the required amount. One of the most reliable methods of irrigation scheduling is based on soil water status and water budget, also called evapotranspiration (ET)-based irrigation scheduling. Success of ET-based method is dependent on the accurate prediction of crop evapotranspiration (ETc). ETc is calculated using reference evapotranspiration (ETo) and crop coefficient (Kc). The use of a representative Kc is vital for the correct estimation of ETc. A comprehensive list of Kc, for most of the commercial crops, has been provided by Doorenboss and Pruitt (1977) and Allen et al. (1998). But Kc may vary based on local factors such as climate, soil, and irrigation methods (Kang et al., 2003). Due to the influence of local factors on Kc, studies have reported differences between published and locally developed Kc for maize (Simon et al., 1998), sorghum (Tyagi et al., 2000), potato (Kashyap and Panda, 2001), and wheat (Kang et al., 2003). Due to such variations, Jones et al. (1984) recommend the use of locally developed Kc. Local Kc for Florida has been 28

PAGE 42

29 developed for tomato, strawberry (Clark et al., 1996) and blueberries (Haman et al., 1997), however, local Kc for bell pepper and watermelon are not available. While ETc is an estimate of crop water requirement, the amount of irrigation applied in a field is dependent upon the effectiveness of the irrigation system. Seepage irrigation system, despite its low efficiency, is used on majority of vegetable farms in southwest Florida. The reasons for its continued use are low cost, ease of operation, and maintenance. To boost water conservation, vegetable industry in Florida is being encouraged to adopt more efficient micro irrigation systems such as the drip system. Contrary to the seepage system, the drip system can apply water conservatively with efficiency as high as 85% (Simonne et al., 2003). Use of the drip system can lead to significant reduction in water applied for vegetable production compared to the seepage system (Csizinszky, 1980). Comparing drip and seepage system, Pitts and Clark (1991) reported a significant reduction in the water use of tomato using the former, without affecting yield and fruit quality. Although studies have shown that drip system is comparatively more efficient than the seepage system, it is used by only 25% of the growers in Florida (Smajstrla and Haman, 1997). Lack of studies validating the effectiveness of the drip system over the seepage system for different crops in Florida is one of the main reasons which limit the adoption of drip system. To encourage the adoption of the drip system in Florida, more studies are needed, covering a variety of locally grown crops. The specific objectives of this section of the study were: 1. Quantify and compare ETc for bell pepper and watermelon grown under drip and seepage irrigation systems. 2. Development of crop coefficient (Kc) for bell pepper and watermelon grown in Florida.

PAGE 43

30 Material and Methods Study Location This study was conducted at the Southwest Florida Research and Education Center (SWFREC) located in Immokalee, which is in Collier County, Florida. Collier County is one of the five counties together known as southwest Florida (Figure 3-1). The average maximum and minimum temperatures for the region are 29 oC and 17 oC, respectively. Southwest Florida receives an annual rainfall nearly of 1,370 mm. Soils in the area are typically poorly drained, hydric and highly sandy in characteristics. These soils, also known as flatwood soils, have a subsurface spodic horizon which acts as a hard pan to maintain a high water table. Seasonal high water table varies from 15 cm to 45 cm. Figure 3-1. Study location. Experimental Design A set of six drainage lysimeters were used to quantify the ETc and develop Kc for bell pepper and watermelon. To compare the drip and seepage irrigation systems, four of the six lysimeters were irrigated with drip system (designated as D1, D2, D3 and D4) while the remaining two were kept under seepage irrigation system (S1 and S2). A lower number of seepage lysimeters than drip lysimeters was employed due to cost considerations. Vegetables in southwest Florida are grown on raised, pressed soil beds

PAGE 44

31 covered with plastic mulch (Clark et al., 1996) with fixed row-to-row (r-r) and plant-to-plant (p-p) spacing. The r-r and p-p spacings were important factors in designing the size of the lysimeters. To emulate the actual crop management practices, a survey of vegetable production farms was carried out in June-July 2002. Survey of crop production practices A vegetable production survey covering six large vegetable producers in southwest Florida revealed considerable variability in crop production practices. Typical crop rotation in southwest Florida includes tomato or pepper grown in the fall season followed by watermelon, eggplant or tomato during the spring season. While most of the surveyed growers were using seepage irrigation systems, some were using a combination of drip and seepage irrigation. However, even in the latter case seepage irrigation was used as the primary irrigation method. The survey showed that watermelon had the largest r-r spacing among all vegetable crops. The r-r spacing for watermelon varied from 1.8 m to 2.75 m. Since watermelon was one of the two crops to be studied, it was considered as the basis of lysimeter design. Survey further revealed considerable variability in field layouts and other production practices including fertilizer application rates, pesticide use and plant density. Production practices data (e.g. plant density, area) from survey and University of Florida/Institute of Food and Agricultural Sciences (UF/IFAS) recommendation for watermelon (Maynard et al., 2001) were considered as the basis for determining the size of the lysimeters. Lysimeter design The drainage lysimeters designed for this study were 4.87 m 3.65 m 1.37 m (Figure 3-2). The lysimeter was designed to emulate a snapshot of actual field conditions. Each lysimeter had two plant rows (length = 3.65 m) and a seepage ditch in the center.

PAGE 45

32 This ditch was provided for irrigation in the seepage lysimeters. The ditch was also constructed in the drip lysimeters for wetting the soil during bed making. With typical spacing, each lysimeter accommodated six watermelon plants (p-p = 1.2 m; 3 plants/bed) (Figure 3-2) or 40 pepper plants (p-p = 30 cm, 20 plant/bed; 2 plant row/bed). Figure 3-2. Lysimeter layout for the watermelon crop. The dimensions of the lysimeters were designed to: 1) ensure survival of at least three watermelon plant assuming 50% failure due to incidences of disease; 2) capture a snapshot of typical layout of watermelon fields, and therefore, simulate field conditions especially the soil evaporation component; and 3) obtain realistic water use and water quality data. Field design Advection effect has been found to be critical in lysimeter studies in arid regions (Aboukhaled et al., 1982). To minimize advection effect, it is necessary for a lysimeter to be surrounded by a vegetation same as that in the lysimeter, creating a buffer area. A buffer area of 400 times the area of lysimeter was found acceptable for arid climates

PAGE 46

33 (Aboukhaled, 1982). However, smaller buffer areas (e.g. 5 m; Fougerouge, 1966) were acceptable under humid and semi-humid conditions (Thornthwaite and Mather, 1955). To minimize advection effect due to vegetation differences, the field layout (bed height, r-r and p-p spacings) was kept similar to that of the lysimeter. The 0.7 ha experimental field area created a buffer area of 378 times the area of the lysimeters around them. Field was divided into eight blocks of crop rows. Each block had four beds (Figure 3-3). The lysimeters were oriented such that the beds were parallel to its 3.65 m side. Placement of a lysimeter was such that the crop row inside the lysimeter aligned with the crop row on the outside, making continuous crop row. The lysimeters were installed in two groups: for the drip and the seepage system. A group of four lysimeters under the drip system (D1, D2, D3 and D4), was installed in the third block from the north end of the field. The two seepage lysimeters (S1 and S2) were installed in the sixth block from the north end of the field. The lysimeters were spaced at 2.5 m from each other to allow for the installation of instrumentation (Figure 3-3).

PAGE 47

34 Figure 3-3. Experimental field layout.

PAGE 48

35 Lysimeter Construction and Installation Figure 3-4. Construction details of the lysimeter. Low carbon steel sheet (thickness = 3.175 mm) was used to construct the lysimeters. Gas Metal Arc Welding (GMAW) was used for welding together the steel sheets. An angle iron frame, made of 5.08 cm 5.08 cm 0.64 cm mild-steel angle iron bars, was constructed to support the walls of the lysimeter. Steel sheets were welded to the angle iron frame to form the four sides and bottom of the lysimeter. Three additional vertical angle iron bars, spaced at 1.21 m, were used to support the 4.87 m side wall of lysimeter (Figure 3-4). Two vertical bars were used to support the 3.65 m wall. One horizontal bar (0.30 m from the top) was welded on the inside wall of lysimeter. These support bars were used to avoid bending or bowing of lysimeter walls. A screened pipe was welded to the bottom of the lysimeter for drainage. The pipe extended out of the lysimeter through the bottom sheet. To prevent rusting, lysimeters were painted with two coats of multi purpose epoxy paint followed by two coats of anti corrosive black stone guard paint. The bottom of the lysimeter was provided extra protection from rust by painting it with two additional coats of elastomeric cold-tar free paint. After the paint

PAGE 49

36 dried, each lysimeter was tested for leaks by filling them with water. Minor leaks observed during testing were sealed by GMAW followed by painting. The installation of the lysimeters was started in the first week of January 2003 and was completed in 45 days. Two large soil pits (each 1.37 m deep) were excavated in the experimental field, one for each set of drip and seepage lysimeters (Figure 3-5). A backhoe loader was used to remove soil in 15 cm increments. The excavated soil for A and E horizons was stored separately at two different locations to avoid mixing. Figure 3-5. Lysimeters inside the soil pit during installation. A 7.5 cm thick gravel layer was placed in the pit to provide a stable foundation for the lysimeters. Cement blocks (20 cm 30 cm 10 cm) were placed on the gravel layer under each lysimeter leg (four legs on 4.87 m sides; three on 3.65 m sides) (Figure 3-4). These blocks served as the platform for the lysimeters distributing the point load. Laser level was used to ensure that all cement blocks were at the same elevation. However, in some cases the top elevation of a lysimeter was found to be different than the design due to settling of the soil and/or movement of the cement blocks. In such instances, the gravel layer beneath the lysimeter was removed or added to ensure the design elevations. Water was filled in each lysimeter to hold them in place to avoid floating, due to high water table conditions. A wooden frame 5.5 m 4.3 m was constructed around each lysimeter

PAGE 50

37 base using wooden boards (25 cm 5 cm). Flowable cement fill was poured in the wooden frame to fill the open air area between the lysimeter bottom and the gravel layer. The cement foundation provided a solid base for the lysimeter. Two weeks were allowed to ensure proper hardening of the cement. The next step was to rebuild the soil profile in and around the lysimeters. Native soil at the research site, Immokalee fine sand, was used for this purpose. Soil characterization of the native soil in the research field was undertaken prior to the beginning of installation process, which included measuring the depth and bulk density for each soil horizon. The soil horizons observed at the experimental field were typical of the Immokalee fine sand soil: two horizons, A and E. The thickness of the A horizon was 0.30 m while the E horizon was 0.71 m. Bulk density samples for both horizons were also taken at different locations in the experimental field. In the process of rebuilding the soil profile inside the lysimeters, firstly a 20 cm thick layer of coarse sand was placed at the bottom of each lysimeter around the drainage pipe. The layer of coarse sand in the lysimeters served the purpose of trapping sand particles and facilitating drainage (Xu et al., 1998). Since the pipe at the bottom was the only mechanism for drainage in lysimeters, extra protection from clogging was provided through a geo-textile sheet made from woven fabrics of monofilament polypropylene yarn (average mesh size of 0.21 mm) and placed on top of the coarse sand layer. The next step was to rebuild the soil profile inside the lysimeter similar to that in the field. This process was carried in increments of 15 cm with light compaction (Figure 3-6). The E horizon (0.71 m) was rebuilt first by placing it on top of the filter cloth. To bring the bulk density of lysimeter soil similar to that found under natural conditions

PAGE 51

38 lysimeter soil was saturated and then drained. Lysimeter soil was allowed to dry for two days after draining. Process was repeated until the desirable bulk density was achieved. The process was repeated for rebuilding the A horizon (topsoil). Soil in the pit around the lysimeter was also reconstructed at the same time using the similar procedure. Figure 3-6. Soil profile inside the lysimeter. Since the bed construction and the plastic mulch installation equipment were tractor mounted, they could not be used inside the lysimeter. Therefore, plastic mulched beds inside the lysimeters were constructed manually using a wooden mould 1.82 m 0.81 m 0.22 m. This mould was accurately positioned in the lysimeter and filled with soil in increments of 5.0 cm with light compaction. The bed was then covered tightly with the plastic in a manner similar to that achieved by the tractor mounted equipment. Monitoring System Water inputs, outputs (excluding ETc), and storage components (S) were measured to compute the water balance. The governing water balance equation for a drainage lysimeter (Chow, 1964) can be described as: SOutflowInflow Equation 3-1 Rainfall (R) and irrigation (I) are the inflow to a lysimeter while drainage (D), runoff (RO) and crop evapotranspiration (ETc) are outflows. Change in storage is

PAGE 52

39 accounted by the change in soil water storage (SWS) in the entire lysimeter for a given time period. Equation 3-1 can be expanded to estimate the ETc as shown in equation 3-2. SWSRODIRETc Equation 3-1 Irrigation Separate irrigation systems were installed for drip and seepage irrigation in the field and the lysimeters. Drip and seepage irrigation systems for the lysimeters were designed to measure irrigation volumes for individual lysimeters. Both systems were subdivided into separate drip and seepage distribution line for each lysimeter. A flow meter was installed on each distribution lines. Irrigation in each distribution line was controlled using a hydraulic actuator switch installed at the main pump station. Irrigation scheduling of lysimeters was based on the soil moisture in the root zone. It is important to maintain optimum soil moisture for quantifying ETc and Kc. Less than optimum soil moisture can stress the plants and reduce the ETc, while over application can cause an increase in ETc, thereby affecting Kc (Allen et al., 1998). To avoid over application or water stress conditions, irrigation scheduling in all lysimeters was based on soil moisture at 0-10 cm of the plastic mulched bed. Studies have shown that for the purpose of scheduling irrigation, soil moisture should be maintained above 33% of plant available water (PAW) to achieve optimum plant growth (Fares and Alva; Obreza et al., 1997). The PAW is calculated as the difference between the field capacity (FC) and the permanent wilting point (WP) (Fares and Alva, 2000a). However, studies have reported variability in the measured FC and WP for the sandy soils at different locations in Florida, e.g. FC and WP for Candler fine sand was reported as 9% and 1.5% by Fares and Alva (2000a), while Obreza et al. (1997) reported the same as 8.0% and

PAGE 53

40 5.0%. The FC and WP for the native soil (sandy flatwood soil) was reported by Obreza et al. (1997) as 8.7% and 3.5%, respectively. For irrigation scheduling purposes, FC and WP for flatwood soils as reported by Obreza et al. (1997) were used in this study. Due to the variability in FC and WP in the reported data, irrigation scheduling was targeted to keep the soil moisture above 8% in 0-10 cm of the plastic mulched bed. Drainage and runoff The lysimeters were designed to capture drainage from the bottom of the lysimeters. Drainage was facilitated by designing the shape of the lysimeter bottom similar to a funnel (Figure 3-7). The sloped bottom was to ensure that the leachate from the bedded as well as non-bedded areas was well mixed before moving to the drainage pipe at the bottom. The drainage pipe was welded to a solid steel pipe which extended outside the lysimeter. The solid pipe was attached via coupling to a marine grade PVC hose. The PVC hose was connected to the bottom of a drainage sump installed adjacent to the lysimeter (Figure 3-8). The drainage sump was made of hard PVC pipe (diameter = 20 cm). The bottom of the sump was located at the same elevation as the drainage pipe. Water was drained out from the sump by using a pump housed in an instrument box located next to the lysimeter (Figure 3-8). To capture runoff, walls of the lysimeter extended 16 cm above the ground surface. Runoff was collected in two runoff catchments (0.46 m 0.46 m 0.46 m) welded to the lysimeter walls (Figure 3-4). These catchments had adjustable gates which were set to the elevation of soil surface inside the lysimeter. Runoff from these tanks was routed to a sump similar to that used for drainage (Figure 3-8). Water was removed from the sump using a separate pump dedicated to runoff, which was housed alongside the drainage pump in the instrument box. The DC pumps were powered by a 12 V battery, charged

PAGE 54

41 onsite using solar panels, installed on top of the instrument box. Drainage and runoff were measured using separate flow meters housed in the instrument box. Figure 3-7. Schematic of the funnel shaped bottom of the lysimeter. The discharge (drainage and runoff) from the lysimeter needed to be stored onsite for water quality sampling. A 10-year rainfall event (200 mm) was considered to design for storage of discharge. Assuming that water table was at 45 cm from the top of the bed at the time of rainfall (bed height = 20 cm), 25 cm of vertical storage space was left in the soil. If the rainfall occurred soon after an irrigation event with soil moisture at near field capacity (9%), the available storage was 20 mm/10 cm (assuming saturation at 29%). With 25 cm of storage space, only 50 mm of rainfall could be stored from the rainfall (if water table was allowed to rise up to the level of the ground surface). The 150 mm of extra water added by rainfall needed to be discharged. Assuming that 30% of the extra volume of water will contribute to runoff, a 10-year rainfall event will result in nearly 900 liters (l) of runoff and 2100 l of drainage.

PAGE 55

42 Figure 3-8. Elevation diagram of the lysimeter. Large tanks to store the entire drainage and runoff volume could not be installed due to cost and labor considerations. Therefore, to address the large discharge volume, lack of available space and resources to install large tanks for storage, a water distribution instrument, termed here onwards as the water-splitter, was designed and fabricated (Figure 3-9). The water-splitter was designed to split the incoming drainage/runoff from the lysimeters into 10 equal parts. While 1/10 of the volume was stored for sampling, the larger part of discharge was discarded. Stored volume represented a composite sample for the entire discharge volume for water quality comparison. The use of a water-splitter reduced the volume of discharge to be stored by 90% of the total volume. It allowed the use of a smaller capacity (210 l) tank to store the lysimeter discharge (Figure 3-10). Splitting of the incoming flow into 10 equal parts was achieved through a cone with a 10.40 cm disc at the base (Figure 3-9). The cone and disc of the water-splitter were built as one unit, made from a single piece of solid Teflon cylinder. Ten openings of same diameter (2.1 mm) were drilled in the Teflon disc equidistant from each other (Figure 3-9). The disc was mounted inside an enclosure made from PVC casing (height = 30.48 cm, outer diameter = 11.68 cm). The PVC casing was closed at both ends using circular aluminum caps. Four threaded rods were drilled through the aluminum caps to act as the

PAGE 56

43 supporting frame for the PVC casing. Each of these rods had a fly nut attached at the top (Figure 3-9). These four fly nuts also served the purpose of leveling the water-splitters in the field to ensure equal splitting of the incoming volume. Both water-splitters (one each for drainage and runoff) were mounted on a wooden pole and installed between the storage tanks. Pumped outflow from sumps fell directly on top of the cone and was distributed into 10 parts by the openings in the disc. One of the openings on the disc was connected to the 210 l PVC tank through a pipe. Flow from rest of the nine openings was routed out of the splitter and discharged away from the lysimeter using a buried drainage pipe. Each splitter was tested in the laboratory prior to field use. Figure 3-9. Schematic of the water-splitter. Soil moisture and water table Accurate estimation of change in storage (S in Equation 3-1) is critical for a lysimeter study. Weighing lysimeters account for this change by directly measuring the change in weight of the lysimeter itself. In a drainage lysimeter, the change in SWS

PAGE 57

44 accounts for the change in storage. Several methods such as neutron probes (Haman et al., 1997), gypsum block (Simon et al., 1998) and tensiometers (Clark et al., 1996) have been used to measure the SWS in drainage lysimeters. However, lack of accuracy of these devices can introduce uncertainty in the estimation of SWS (Steele et al, 1996). Capacitance based devices can provide accurate estimation of volumetric soil moisture (Fares and Alva, 2000b). Moreover, these devices are insensitive to the changes in fertilizer salt concentration in the soil (Fares and Alva, 2000b). Diviner, a portable capacitance based soil moisture sensor (Sentek PTY Ltd., South Australia) was used for measuring the daily soil moisture in the lysimeters. The soil moisture in a lysimeter was measured at two locations: bedded and non-bedded areas (Figure 3-10). Soil moisture measurements using the Diviner, taken at 10 cm increment, were used to schedule irrigation in the lysimeters. Figure 3-10. Instrumentation for the lysimeter.

PAGE 58

45 A water table monitoring well was installed and fitted with a pressure transducer know as the Levelogger (Solinst, 2002) in each of the lysimeters (Figure 3-10) to collect water table data. Weekly water table data were also collected manually as backup. Weather parameters Weather data including rainfall, air temperature (at 2 m), wind speed (at 10 m), relative humidity and solar radiation were collected at the UF/IFAS Florida Automated Weather Network (FAWN) weather station located at SWFREC, Immokalee. This weather station was located 50 m north of experimental field (Figure 3-11). FAWN weather data ( http://fawn.ifas.ufl.edu ) were used to estimate ETo in this study. Fertilizer management The N-P-K fertilization for the lysimeters was based on UF/IFAS fertilizer recommendations (Maynard et al., 2001). All six lysimeters received the same amount of fertilizer. However, time of application of fertilizer was different for drip and seepage lysimeters. All the fertilizer in seepage lysimeters was applied in the bed before planting. Fertilization of drip lysimeters included application of 25% of total fertilizer requirement as pre-plant in the bed. Remaining fertilizer in these lysimeters was applied through weekly fertigation. Fertilizer rate and application schedule for the rest of the field was kept the same as that of drip lysimeters. Crop yield Yield data were collected from the lysimeters as well as the outside field for each season. To compare the lysimeter yield with rest of the field, six check plots were established (Figure 3-3). Each check plot had the same number of plants as any lysimeter. Harvesting in the lysimeters and the rest of the field was done at the same time. Only

PAGE 59

46 fruits of marketable quality were harvested to compute the yield. Yield data collection included fruit count and weight. Crop Production Practices Watermelon The spring 2003 watermelon crop was transplanted on 03/10/2003. N-P-K fertilizer was added at 168-112-168 kg/ha in the lysimeter and the field. First transplants showed symptoms of a fungal disease caused by Pythium spp. As a result, the lysimeters were replanted. However, successive transplanting failed within one week of transplanting. Fifth transplants drenched in recommended preventive chemical Rodomil Gold 4 EC (Maynard et al., 2001) survived till the 6th week after transplantation. The crop became infected with Fusarium wilt caused by Fusarium oxysporum during the 6th week and damaged the entire crop by the 8th week. Crop failure did not allow for a full season of data to be collected. Due to lack of full season of data, Kc for spring 2003 was not developed. To avoid the occurrence of fungal disease in the spring 2004 season, preventive fumigant (K-pam HL, application rate = 250 l/ha) was applied in the lysimeters prior to planting. The rest of the experimental field was fumigated with Telone, which was added to soil at the time of bed preparation. Watermelon was planted on 02/24/2004 in spring 2004. The crop showed signs of disease in some parts of the field during the 2nd week after planting. To avoid spreading of the disease, watermelon was replanted on 03/08/2004. N-P-K fertilizer was applied at 168-0-168 kg/ha during the spring 2004. The crop showed signs of a disease known as “vine decline” during the 11th week after transplanting. Spread of the disease in the research field ended the crop season after 81

PAGE 60

47 days after transplanting (DAT). The season was limited to two harvests on 5/25/2004 and 05/28/2004 Pepper Prior to planting in the fall 2003, lysimeters and the surrounding experimental field were fumigated with K-pam HL and Telone respectively. Pepper was planted on 09/10/2004. N-P-K fertilizer was applied at 224-168-224 kg/ha. Crop was harvested three times during the season on 11/07/2003, 12/01/2003 and 12/19/2003. Figure 3-11. Research field during pepper season (fall 2003). Computation of Rreference Evapotranspiration For the purpose of developing Kc curves, two different ETo methods were used: FAO Penman-Monteith method (FAO-PM) and FAO-modified Blaney-Criddle method (FAO-BC). FAO-Penman-Monteith method The FAO-PM (Allen et al., 1998) is the standard method of ETo estimation. Allen et al. (1998) described the methodology of estimating ETo using FAO-PM (equation 3-3).

PAGE 61

48 )34.01()()273(900)(408.022ueeuTGRETasno Equation 3-3 where Rn is the net radiation at the crop surface [MJ m-2 day-1], G is the soil heat flux density [MJ m-2 day-1], T is the mean daily air temperature at 2 m height [C], u2 is the wind speed at 2 m height [m s-1], es is the saturation vapor pressure [kPa], ea is the actual vapor pressure [kPa], es-ea is the saturation vapor pressure deficit [kPa], is the slope vapor pressure curve [kPa C-1], is the psychometric constant [kPa C-1]. FAOModified Blaney-Criddle method Blaney-Criddle method is commonly used by water management districts in Florida for the purpose of water allocations. FAO-BC (Doorenboss and Pruitt, 1977) was used in this study (Doorenboss and Pruitt, 1977) Jensen et al. (1990) have observed that FAO-BC provided better estimates than the previous versions of the original Blaney-Criddle method. FAO-BC method estimates ETo as shown in equation 3-4 to 3-7. bfaETo Equation 3-4 41.1)(0043.0minNnRHa Equation 3-5 )1060.0())/(1060.0()066.0)/(07.1()1041.0(82.0min3min2min2ddURHNnRHUNnRHb Equation 3-6 )13.846.0(Tpf Equation 3-7 where RHmin is the minimum relative humidity [%] n is the actual daily sunshine hours [h] N is the maximum possible daily sunshine hours [h] P is monthly percentage of daytime hours T is the mean temperature [C] Ud is the day time wind speed [m s-1] Quantification of Crop Evapotranspiration Bi-weekly and monthly water balance was computed for each lysimeter using equation 3-2. Daily volumes of water inputs and outputs to the lysimeter were totaled to

PAGE 62

49 compute the bi-weekly and monthly input and output volumes (expressed in mm). Since the daily soil moisture readings () were taken at 2 PM, the start and end time for the water balance was 2 PM on the first day and 1:59 PM on the last day of balance, respectively. The SWS was computed using daily soil moisture reading () obtained using the Diviner. SWS were computed for every 10 cm increment up to depth of 50 cm for bedded area and up to a depth of 40 cm for non-bedded area (Equation 3-8). To estimate SWS in the bed, bed geometry was divided into three sections (Figure 3-12). 210ZZStorageWaterSoil Equation 3-8 Where Z2 = 50 cm (bedded area) 40 cm (non-bedded area) Figure 3-12. Schematic diagram of the bed-geometry. Development of Crop Coefficient The bi-weekly and monthly Kc were developed for bell pepper and watermelon using ETo estimates from FAO-PM and FAO-BC methods. The Kc was calculated using equation 3-9 occETETK Equation 3-9 To compute Kc based on crop development stage, it is important to establish the length of different crop growth stages. Allen et al. (1998) divided the crop cycle into four

PAGE 63

50 stages: initial stage (marked with about 10% of plant cover), development stage (marked with the growth of plant from 10% to 100% canopy cover), maturity stage (from 100% plant cover to fruit maturity) and later stage (from maturity to harvesting). To establish the length of crop stages, the following sources from literature were used. According to Maynard et al. (2001), in Florida the initial growth stage of bell pepper and watermelon extends to about two weeks. The two crops reach maturity in about 60 DAT. Similar observation regarding the crop stages of watermelon and bell pepper was also made by Allen et al. (1998). ETc and Kc were computed for bell pepper and watermelon using fall 2003 and spring 2004 season data respectively. Kc for both the crops was developed as a function of DAT, by plotting the computed Kc values against the DAT. The Kc was referenced to middle of the water balance period (e.g. the 7th day for bi-weekly time step) for the purpose of developing a regression equation also known as the “crop curve”. A third degree polynomial regression equation was fitted to bi-weekly and monthly Kc (equation 3-10). Since two irrigation systems were used in the study, separate Kc curves were developed for the drip and seepage irrigation systems. The average Kc for the crop was calculated by averaging the Kc for the drip and seepage lysimeters. 3),(42),(3),(2),(1),(][][][][][XCXCXCCKbababababac Equation 3-10 Where X is the DAT a is the method of irrigation (1 = average; 2 = drip; 3 = seepage) b is the method of ETo estimation (1 = ETo-FAOPM; 2 = ETo-FAOBC) C1, C2, C3 and C4 are the regression coefficients

PAGE 64

51 Result and Discussion Bell Pepper Post-season analysis of fall 2003 data showed possible leakage in D2. To avoid uncertainty in ETc estimates for D2, data from this lysimeter were not included in this study. Table A-1 shows the historical weather data and average weather conditions during fall 2003. Total rainfall during the season was similar to the historical average. Maximum rainfall in the season was during September 2003 (21% higher than the historical average for this month). Mean temperatures during fall 2003 (September to December) were relatively cooler than historical average temperatures. During the fall 2003, daily ETo was highest during September. ETo demand decreased from September to December. As shown in Table A-1, total ETo estimated by FAO-BC was slightly higher than estimates using FAO-PM method. The drip lysimeters produced 22,300 kg/ha of marketable yield of pepper which was similar to that from the seepage lysimeters (22,700 kg/ha). The average yield from the lysimeters was less than the 31,100 kg/ha average yield for Florida during 2002-03 (USDA-NASS, 2004). As shown in Table B-1, irrigation was highly dependent on the rainfall distribution. Reduction in irrigation in all the lysimeters during 49-62 DAT and 91-100 DAT was due to 68.1 mm and 39.6 mm rainfall respectively. Similar irrigation was applied in the three drip lysimeters throughout fall 2003. However, the two seepage lysimeters show marked difference in irrigation during 77-100 DAT. During this time, the average soil moisture in 0-10 cm (in plastic mulch bed) in the S1 was about 2.5 mm less than that in the S2. Therefore higher irrigation was applied in S1. The crop water requirement increased considerably as the crop approached maturity (60 DAT). Increased water demand was met with more water being applied in the

PAGE 65

52 seepage lysimeters to maintain optimum soil moisture. Soil moisture in the top 10 cm soil (in plastic mulch bed) in drip lysimeters were maintained within the optimum range without a considerable increase in irrigation. Overall, drip lysimeters used 31% less irrigation compared to the seepage lysimeters. Drainage and runoff data were collected on an event basis. The highest drainage during fall 2003 occurred during 0-20 DAT as a result of 104.3 mm rainfall. Lysimeters also showed considerable drainage during 49-62 DAT in response to a 68.1 mm rainfall event. Overall, drainage was similar in both the systems; however, there were differences in the drainage from individual lysimeters. Runoff occurred less frequently than drainage in the lysimeters because of the nature of flatwood soils. Due to high hydraulic conductivity of these soils, most of the rainfall infiltrates into the soil as soon as it falls. Runoff occurs only when the water table reaches or is close to the ground surface. Table B-1 shows that runoff was produced during 0-20 DAT in response to 104.3 mm rainfall event. Runoff during the 49-62 DAT was a result of a series of rainfall events producing 68.1 mm rainfall. As shown in Table C-1, the rainfall events which produced runoff resulted in very high water table conditions in both the systems thereby producing runoff. Overall the seepage lysimeters produced slightly higher (14%) runoff volume than the drip lysimeters. Higher soil moisture in non-bedded areas in the seepage system (Figure 3-13) reduced the available SWS in the seepage system resulting in slightly higher runoff than the drip system. The SWS in the seepage lysimeters remained higher than that in the drip lysimeters for most part of the season. However, Figure 3-13 shows that SWS was similar among the drip and the seepage replications. Lateral movement of water in seepage lysimeters

PAGE 66

53 from the ditch to the bed results in higher wetted area as compared to a small wetted area in the drip lysimeters. Furthermore, higher water table in the seepage lysimeters also contributed to higher SWS. Figure C-1 shows that the water table in seepage lysimeter was higher than that in drip lysimeters for most times during the season. Similar water table was observed in the two systems during the times of rainfall. Highest water table differences between the two systems were observed during 63-90 DAT (November 10th-December 10th) during the period of low rainfall. During this time, water table was lower than the remaining season in the drip system but the soil moisture in the 0-10 cm was maintained in optimum range due to application of water to the root zone. 60.0070.0080.0090.00100.00110.00120.00130.00140.00150.000102030405060708090100Days after transplantingSoil Storage (mm of water) D1 D3 D4 S1 S2 Planting Date september 10, 2003 Figure 3-13. Soil water storage in lysimeters during fall 2003. Crop Evapotranspiration (ETc) As shown in Figure 3-14, there was considerable difference (> 20 mm) between actual ET (ETc) and potential ET (ETo-FAO-PM) during the initial stage of crop

PAGE 67

54 development. During the initial stage, ETc was primarily due to surface evaporation (Ea), while transpiration (Tp) contributed only a small part of the ETc. As the season progressed, with increasing Tp, the ETc approached ETo-FAO-PM. Peak ETc was reached during 48-62 DAT, when average ETc of bell pepper was equivalent to ETo-FAOPM. Timing of the peak ETc was consistent with Maynard et al (2001) and Allen et al (1998), who noted crop maturity for bell pepper about 60 DAT. Figure 3-14 showed that seepage system always had higher ETc compared to the drip system with the exception of 90-100 DAT. ETc differences were small during initial stages and increased during later stages. 010203040506001020304050607080901Days after Transplanting (DAT)Crop evapotranspiration (ETc) in mm 00 ETo-PM Drip Seepage Average Figure 3-14. Bi-weekly ETc for bell pepper during fall 2003. Similar trend of ETc was observed for all the three drip lysimeters (Figure 3-15). However, ETc values for D1 were slightly different compared to D3 and D4. All three lysimeters reached peak ETc during 48-62 DAT, followed by decline during 63-76 DAT. While peak ETc during 48-62 DAT could be due to the crop reaching 100% plant cover, increased Ea losses due to 69 mm rainfall may have amplified ETc (Allen et al., 1998). Decline in ETc during 63-76 DAT was likely the effect of rainfall and drainage also

PAGE 68

55 reported by Steele et al (1996). During 48-62 DAT, 69 mm rainfall event produced 34.3 mm drainage from the lysimeters. Figure C-1 shows considerable decline in water table due to the drainage in both the systems. 01020304050600102030405060708090Days after transplanting (DAT) (ETc) Crop Evapotranspiration (mm) 100 D1 D3 D4 ETo-PM Figure 3-15. Bi-weekly ETc for drip irrigated lysimeters for bell pepper (fall 2003). Compared to the drip system replications, differences in ETc between the two seepage replications were higher (Figure 3-16). Higher differences in seepage replications were mainly due to differences in SWS in S1 and S2 (Figure 3-13). Drip system had a smaller wetted area compared to the lateral movement of water in seepage system resulting in more wetted area, particularly in the non-bedded area. Higher wetted area could cause variability in Ea from non-bedded area. Moreover, Figure C-1 shows higher fluctuations in water table due to irrigation in the seepage system as compared to the drip system. The fluctuations in water table caused further variability in Ea from the non-bedded areas in the seepage system. The SWS for each lysimeter were computed based on soil moisture measurements at two locations. Due to lesser variations in Ea, SWS computed for the drip lysimeters were more representative of the actual SWS than

PAGE 69

56 those for the seepage lysimeters. Better representation of SWS in drip lysimeters than in seepage lysimeters explains lesser differences between the replications of the former. 01020304050600102030405060708090Days after Transplanting (DAT)(ETc) Crop Evapotranspiration (mm) 100 S1 S2 ETo-PM Figure 3-16. Bi-weekly ETc for seepage lysimeters for bell pepper (fall 2003). Comparison between the two systems shows higher ETc demand in the seepage system. Higher ETc in seepage system was due to following reason: firstly, Ea losses were higher at the time of irrigation in the seepage system, due to open water evaporation from the seepage ditch. These losses were minimized in the drip system by using emitters covered under plastic mulch. Secondly, higher SWS in non-bedded areas in the seepage system, as explained above, caused higher soil Ea losses compared to the losses under drip system. Clark et al. (1996) also reported that Ea losses from exposed row middles increased ETc. The total ETc for drip and seepage irrigation systems is shown in Table 3-1. Overall, the drip system showed nearly 20% reduction of ETc compared to the seepage system. It should be noted that in actual field conditions these differences will be much larger due to increased deep percolation losses from the seepage system.

PAGE 70

57 Table 3-1. ETc, ETo and rainfall for bell pepper crop in fall 2003 season. ETc (mm) Average Drip Seepage ETo-FAO-PM (mm) Rainfall (mm) 216.7 200.5 240.8 291.8 226.20 Crop Coefficient (Kc) The Kc curve for bell pepper developed in this study (Figure 3-17) followed the trend of a classic Kc curve, where Kc is small at the beginning of the season and increases as the plant grows until it reaches a maximum value at crop maturity (Allen et al., 1998; Simon et al., 1998). The initial Kc during 0-20 DAT (Kc > 0.40) for bell pepper developed in this study was higher than those reported in literature. Higher initial Kc is an artifact of the production practices in southwest Florida. As stated before, due to higher water table at the time of transplanting, Ea losses are higher leading to higher ETc and as a result higher Kc. Moreover, rainfall during that time may have further increased the Ea. High Kc was obtained during 48-62 DAT followed by a decline during 62-76 DAT (Figure 3-17).The fluctuations were likely the effect of rainfall and drainage (Steele et al., 1996). However, high Kc during 48-62 DAT could be due to crop maturity or increased Ea due to rainfall. Allen et al. (1998) reported higher than normal Kc due to increased soil Ea if the soil remained wet from rainfall or irrigation. Simon et al (1998) have also reported that for maize grown in sandy soils, Kc for the wet season was greater than the Kc during dry season. The highest Kc (Kc FAO-PM = 1.18; Kc FAO-BC = 0.97) value for bell pepper was computed at the end of crop cycles (90-100 DAT). But it should be noted that 39.6 mm of rainfall during 90-100 DAT may have caused higher than normal Kc. Highest Kc for a crop is expected when it attains maturity, which is around 60 DAT for bell pepper in Florida (Maynard et al., 2001). Unless crop maturity was delayed by nearly 30 days, peak Kc during 90-100 DAT was due to high Ea losses. Although there are chances

PAGE 71

58 that Kc during 48-62 DAT was amplified due to rainfall effect, it can be considered as the peak Kc for bell pepper due to crop maturity. However, due to influence of rainfall and drainage, it is not possible to separate the effect of crop maturity and increased Ea on bell pepper Kc. FAO-PM : y = 2 0-06 x3 0.0004 x2 + 0.0244x + 0.1855R2 = 0.79FAO-BC : y = 3 10-06 x3 0.0004 x2 + 0.0214x + 0.2939R2 = 0.600.000.200.400.600.801.001.201.400714212835424956637077849198105Days after transplanting (DAT)Crop Coefficient (Kc) Kc FAO-PM Kc FAO-BC Poly. (Kc FAO-PM) Poly. (Kc FAO-BC) Figure 3-17. Bi-weekly crop coefficient for bell pepper using FAO-Penman-Monteith and Blaney-Criddle equations. As shown in Figure 3-17, FAO-PM based Kc were consistently higher than FAO-BC based Kc, with the exception of first two weeks. On average, FAO-PM based Kc was found to be 7.2% higher than FAO-BC based Kc. ETo equations based differences in Kc were reported by several studies (and Steele et al., 1996; Simon et al., 1998; Tyagi et al., 2000; Sepaskhah and Andam, 2001). The crop curves were developed by fitting a third order polynomial curve (Table 3-2). Similar equations have been used in the past (Elliott et al., 1988; and Sepaskhah and Andam, 2000) to describe Kc models.

PAGE 72

59 Drip : y = 310-06x3 0.0005 x2 + 0.0261 x + 0.1532R2 = 0.81Seepage : y = 5 10-07x3 0.0002 x2 + 0.0217 x + 0.2338R2 = 0.790.000.200.400.600.801.001.201.400714212835424956637077849198105Days after transplanting (DAT)Crop Coefficient (Kc) Drip Seepage Poly. (Drip ) Poly. (Seepage) Figure 3-18. FAO-Penman Montieth based bi-weekly Kc curve for bell pepper for drip and seepage lysimeters. Figure 3-18 shows the differences in Kc between the drip and seepage systems. The seepage Kc curve follows similar trend as the drip Kc curve, but the former consistently remains higher than the latter with exception of 90-100 DAT. As discussed, there was increased Ea during 90-100 DAT. During this time, all drip lysimeters show similar Kc (D1 = 1.38; D3 = 1.25; D4 = 1.19) while there was differences between Kc computed for S1 (1.35) and S2 (0.75). Difference between S1 and S2 was likely due to the inability to accurately capture variability in SWS in the seepage lysimeters. Overall, higher Kc for seepage lysimeters as compared to the drip lysimeters was expected due to higher Ea in the former. Third order equation obtained high coefficient of determination (R2) for both drip and seepage systems (Table 3-2). The Kc curve for seepage system plateaued around 60 DAT (Figure 3-18) which was the expected maturity time for bell pepper (Maynard et

PAGE 73

60 al., 2001). Literature (Allen et al., 1998; Simon et al., 1998) showed that Kc after attaining its peak value, either was maintained (shown by plateau in Kc curve) or was reduced slightly due to management factors. Drip : y = 4 10-06x3 0.0005 x2 + 0.0225 x + 0.2632R2 = 0.68Seepage : y = 1 10-06x3 0.0003 x2 + 0.0198 x + 0.3399R2 = 0.510.000.200.400.600.801.001.201.400714212835424956637077849198105Days after transplanting (DAT)Crop Coefficient (Kc) Drip Seepage Poly. (Drip) Poly. (Seepage) Figure 3-19. FAO-Blaney-Criddle based bi-weekly Kc curve for bell pepper for drip and seepage systems. Bi-weekly Kc developed using FAO-BC (Figure 3-19) and FAO-PM (Figure 3-18) follows similar trends. Lower R2 values for FAO-BC based Kc curve (R2drip = 0.68; R2seepage = 0.51) were obtained compared to those for FAO-PM based Kc curves. The R2 values obtained for this study were within the range of the R2 reported in other lysimeter studies. Steele et al. (1996) generated fifth order polynomial Kc equations which provided the best R2 (0.52 to 0.68) for their data. Low R2 suggests that there are sources of variation in the results which could not been explained. As noted earlier, inaccuracies in estimation of SWS was one of the biggest sources of uncertainty in the results.

PAGE 74

61 Table 3-2. Regression coefficients for bi-weekly crop coefficient curve for bell pepper. ETo-PM ETo-BC Coefficient Average Drip Seepage Average Drip Seepage C1 0.185 0.153 0.239 0.294 0.263 0.339 C2 0.0244 0.0261 0.0217 0.0215 0.0225 0.0198 C3 -0.0004 -0.0005 -0.0002 -0.0004 -0.0005 -0.0003 C4 2 10-6 3 10-6 5 10-7 3 10-6 4 10-6 1 10-6 R2 0.79 0.81 0.79 0.60 0.68 0.51 To reduce the temporal variation due to rainfall and drainage, monthly Kc was also developed. Monthly Kc values (Table 3-3) were more consistent with the classic Kc curve than bi-weekly Kc curves. Monthly Kc values showed peak value during the third month (referenced to 75 DAT). The fluctuations observed in bi-weekly Kc were reduced in monthly computations. Similar to bi-weekly Kc, seepage lysimeters had higher monthly Kc compared to the drip lysimeters. Also, there were differences in Kc due to the ETo method. For the purpose of comparison, Kc developed for bell pepper was compared to the reported values by Allen et al (1998) was used. They have reported Kc in three stages: initial, mid and late. Monthly Kc developed in this study was comparable to the Kc given by Allen et al (1998). The peak Kc for bell pepper at crop maturity was reported as 1.05, although the authors did not illustrate the ETo method used to develop their Kc. The peak monthly Kc for bell pepper in this study using FAO-PM (1.01) was slightly less than that provided by Allen et al. (1998) but peak Kc for using FAO-BC was about 19% less than the reported Kc. Table 3-3. Monthly crop coefficients for bell pepper (fall 2003). Kc FAO Penman Monteith Kc FAO Blaney Criddle DAT Average Drip Seepage Average Drip Seepage 15 0.50 0.47 0.55 0.54 0.50 0.59 45 0.80 0.71 0.93 0.73 0.65 0.85 75 1.01 0.97 1.07 0.86 0.83 0.91

PAGE 75

62 Watermelon Spring 2004 was a relatively dry season receiving only 30% of the historical rainfall (Table A-2). Mean air temperatures were at least 1oC cooler than the historical mean temperatures during March and April 2004. The daily ET rate increased from March to May. The highest ET was observed during May 2004. Seasonal ETo estimates by FAO-BC were about 18% less than those by FAO-PM method (Table A-2). The drip lysimeters produced 20,300 kg/ha marketable yield, which was about 17% higher than the 17,000 kg/ha marketable yield obtained from the seepage lysimeters. The average yield of watermelon during 2002-03 was nearly 37,600 kg/ha (USDA-NASS, 2004). The marketable yield collected from the lysimeters did not include the fruits damaged due to disease pressure, which was the main reason for low watermelon yield from the lysimeters during spring 2004. The average irrigation amounts applied in drip and seepage lysimeters were similar during the first two weeks after transplanting (Table B-2). From the third week onwards (15 DAT) irrigation applied to seepage lysimeters always exceeded the irrigation in drip lysimeters, with the exception of 29-42 DAT. During 29-42 DAT, irrigation requirement in all the lysimeters was reduced due to 34.5 mm rainfall. The irrigation demand increased considerably after 42 DAT in the seepage lysimeters due to two reasons. Firstly, watermelon crop approached maturity, and secondly, ETo demand was very high during May 2004. Drip lysimeters were able to maintain the soil moisture above 8% without considerable increase in irrigation. Overall, drip lysimeters required 59% less irrigation than the seepage lysimeters (Table B-2). Spring 2004 witnessed three considerable rainfall events. However, there was only a single drainage event during 29-42 DAT (Table B-2). Due to higher resident water table

PAGE 76

63 in seepage lysimeters than the drip lysimeters, SWS in the former was less than the latter. Thus, comparatively higher drainage was obtained from the seepage lysimeters. The isolated runoff event in S2 (6.45 mm) was the result of a leak through the adjustable plate on the runoff gates. The SWS in seepage lysimeters were slightly higher than that in drip lysimeters (Figure 3-20). Sharp decline in SWS during May 2004 was more prominent in drip lysimeters than in seepage lysimeters. Figure C-2 shows continuous water table decline in the drip system after May 4th 2004 (61 DAT). The seepage system on the other hand did not show similar decline in water table due to higher irrigation. As a result the non-bedded areas in these lysimeters remained wetter that those in drip lysimeters and consequently the SWS in seepage lysimeter did not show sharp decline. 0.0020.0040.0060.0080.00100.00120.00140.00160.00051015202530354045505560657075808590Day of YearSoil Water Storage (mm) D1 D3 D4 S1 S2 Planting date March 8th 2004 March 2004April 2004May 2004 Figure 3-20. Soil water storage in the lysimeters during spring 2004.

PAGE 77

64 Crop Evapotranspiration (ETc) Figure 3-21 shows considerable differences (> 45 mm) between actual ET (ETc) and potential ET (ETo-FAO-PM) during initial crop stage (0-14 DAT). ETc increased as crop approached maturity. The Peak ETc during 56-70 DAT was consistent with Maynard et al. (2001) who suggested that watermelon attains maturity around 60 DAT in Florida. As shown in Figure 3-21, ETc for seepage system was consistently higher than that for drip system with exception of 28-42 DAT. As discussed in bell pepper section, higher Ea losses in seepage lysimeters resulted in higher ETc compared to that for the drip system. Similar ETc during 28-42 DAT in drip and seepage system suggest similar Ea losses from the two systems due to rainfall. Similar water table in the two systems during 28-42 DAT confirms similar Ea losses from the two systems (Figure C-2). 0102030405060708090100071421283542495663707784Days after transplanting (DAT)Crop evapotranspiration (ETc) (mm) ETo-FAO-PM Average ETc Drip ETc Seepage ETc Figure 3-21. Bi-weekly ETc for watermelon during spring 2004. The ETc in drip lysimeters shows a decline during 70-80 DAT. Allen et al (1998) noted that post-maturity decline in ETc was due to aging of leaves. However, part of decline in ETc during 70-80 DAT could also be a result of disease (vine decline) which

PAGE 78

65 lowered the Tp rate. Progression of the disease in the field was such that drip lysimeters were affected before the seepage lysimeters. Therefore, seepage lysimeters do not show sharp decline as the drip lysimeters (Figure 3-22). 020406080100120071421283542495663707784Days after transplanting (DAT)Crop evapotranpiration (ETc) (mm) D1 D3 D4 S1 S2 ETo-FAO-PM Figure 3-22. Bi-weekly ETc for drip and seepage lysimeters for watermelon (spring 2004). Comparatively, differences among replications of seepage system were higher than those among the drip replications (Figure 3-22). As explained earlier, higher differences among seepage lysimeters were primarily due to uncertainty in accurately representing SWS in these lysimeters. Figure C-2 shows greater water table fluctuations in the seepage system than the drip system, causing further variability in Ea in the former. Overall, ETc for the drip lysimeter was nearly 31% less than that from seepage lysimeters (Table 3-4). Table 3-4. ETc, ETo and rainfall for watermelon crop in spring 2004 season. ETc (mm) Average Drip Seepage ETo-PM (mm) Rainfall (mm) 287 242 353 384 74

PAGE 79

66 Crop Coefficient (Kc) The Kc curve developed for watermelon follows the trend of a classic Kc curve found in literature. Kc increased steadily with time reaching its peak during 56-70 DAT. The peak Kc was consistent with the maturity of watermelon suggested by Maynard et al. (2001). Decline in Kc during 70-80 DAT was likely the result of reduced Tp due to disease pressure. FAOPM : y = -8 10-06x3 + 0.0008 x2 + 0.0007 x + 0.1637R2 = 0.99FAO-BC : y = -1 10-05x3 + 0.0012 x2 0.006 x + 0.2051R2 = 0.990.000.200.400.600.801.001.201.401.600102030405060708Days after transplanting (DAT)Crop coefficient 0 FAO-PM FAO-BC Poly. (FAO-PM) Poly. (FAO-BC) Figure 3-23. Bi-weekly crop coefficient for watermelon using Penman-Monteith and Blaney-Criddle equations. Similar to bell pepper, there were differences in watermelon Kc derived using FAO-PM and FAO-BC based ETo. Figure 3-23 shows similar initial Kc for FAO-PM and FAO-BC based Kc. However, the two curves deviate from each other as the season progresses. The FAO-BC based Kc curve shows 28% higher peak Kc value compared to FAO-PM based Kc curve (Figure 3-23). Similar observations have been made in other lysimeter studies (Steele et al., 1996; Simon et al., 1998; Tyagi et al., 2000; and Sepaskhah and

PAGE 80

67 Andam, 2001). Kashyap and Panda (2001) reported nearly 10% difference between Kc developed using FAO-PM and FAO-BC. For this study, the highest differences between Kc estimated by the two ETo methods were observed in May 2004 (Figure 3-23). Despite differences in the Kc, third order polynomial curves for the two methods show very high R2 (0.99 for both ETo methods). FAO-BC : y = -1 10-05x3 + 0.0013 x2 0.0085 x + 0.1572R2 = 0.99FAO-PM : y = 1 10-05x3 + 0.001 x2 0.0055 x + 0.1405R2 = 0.980.000.200.400.600.801.001.201.40010203040506070809Days after Transplanting (DAT)Crop Coefficient (Kc) 0 Drip-BC Drip-PM Regression Eqn BC Regression Eqn PM Figure 3-24. Bi-weekly crop coefficient for watermelon using Penman-Monteith and Blaney-Criddle equations under drip irrigation system. Figure 3-24 and 3-25 shows bi-weekly Kc for watermelon under drip and seepage irrigation systems respectively. Kc for the seepage system was consistently higher than that of the drip system for both ETo methods. As discussed earlier, higher Kc is a reflection of water management practice which causes higher Ea losses in the seepage system. Peak Kc for drip lysimeters was considerably less (FAO-BC = 1.24; and FAO-PM = 0.98) than those for seepage lysimeter (FAO-BC = 1.67; FAO-PM = 1.29). The R2 values for both the systems were very high (Table 3-5).

PAGE 81

68 FAO-BC : y = 8 10-06x3 + 0.0007 x2 + 0.0089 x + 0.2089R2 = 0.97FAO-PM ; y = 4 10-06x3 + 0.0003 x2 + 0.0136 x + 0.1712R2 = 0.960.000.200.400.600.801.001.201.401.601.80010203040506070809Days after Transplanting (DAT)Crop Coefficient (Kc) 0 Seepage-BC Seepage-PM Regression Eqn BC Regression Eqn PM Figure 3-25. Bi-weekly crop coefficient for watermelon using Penman-Monteith and Blaney-Criddle equations under seepage irrigation system. Higher R2 values for watermelon crop curves (Table 3-5) were in contrast to the lower R2 values obtained for bell pepper curves (Table 3-2). Better curve fitting obtained for spring 2004 data was probably due to the lack of uncertainty introduced by rainfall. The watermelon season (spring 2004) had very few rainfall events compared to bell pepper season (fall 2003). As a result, the fluctuations in Kc curve due to increased Ea were minimized. Table 3-5. Regression coefficients for bi-weekly crop coefficient curve for watermelon. ETo-PM ETo-BC Regression Coefficient Average Drip Seepage Average Drip Seepage C1 0.1637 0.1405 0.1712 0.2051 0.1572 0.2089 C2 0.0007 -0.0055 0.0136 -0.006 -0.0085 0.0089 C3 0.0008 0.001 0.0003 .00012 0.0013 0.0007 C4 -8 10-6 -1 10-5 -4 10-6 -1 10-5 -1 10-5 -8 10-6 R2 0.99 0.98 0.96 0.99 0.99 0.97

PAGE 82

69 Monthly Kc for watermelon (Table 3-6) also follows the classic Kc curve (Allen et al., 1998). Peak monthly Kc (FAO-PM = 1.01; FAO-BC = 1.28) for watermelon was computed for 60-80 DAT. The peak Kc (1.00) for watermelon reported by Allen et al. (1998) was almost identical to the average peak Kc for watermelon developed using FAO-PM (1.01) in this study. However, peak Kc for drip irrigated watermelon in this study was 18% less than the reported Kc Allen et al. (1998) while it was 28% higher for seepage irrigated watermelons. Table 3-6. Monthly crop coefficients for watermelon (spring 2004). Kc – FAO Penman Monteith Kc – FAO Blaney Criddle DAT Average Drip Seepage Average Drip Seepage 16 0.36 0.27 0.48 0.40 0.30 0.54 45 0.93 0.84 1.06 1.19 1.07 1.36 70 1.01 0.82 1.28 1.28 1.04 1.63

PAGE 83

70 Summary and Conclusion A study was undertaken to quantify and compare the ETc and Kc for vegetable crops under seepage and drip irrigation systems. Six drainage lysimeters were designed and installed at the UF/IFAS research center at Immokalee, Florida. Four lysimeters were irrigated with drip system, while two were under seepage irrigation. Bell pepper was grown in fall 2003, and watermelon was grown in spring 2004. The water balance equation was solved for ETc for each lysimeter, and Kc was developed for both the crops using FAO-PM and FAO-BC ETo methods. The bi-weekly Kc for bell pepper using the FAO-PM method varied from 0.50 during the initial stage to 1.18 at maturity, and from 0.51 to 1.07 for the FAO-BC method. Third order polynomial equations were fitted to the computed Kc to develop Kc curves for bell pepper with respect to DAT. Higher R2 (0.79) was obtained for FAO-PM based Kc curve as compared to the FAO-BC (R2 = 0.60) based Kc curve. The monthly Kc conformed to the classic Kc curve better than the bi-weekly Kc curve. The monthly FAO-PM peak Kc values for bell pepper from this study were similar to that reported by Allen et al. (1998). The comparison of bell pepper ETc for the two irrigation systems showed the drip ETc to be 20% less than the seepage system. The bi-weekly Kc curve for watermelon conformed better to the classic Kc curve as compared to the bi-weekly Kc curve for bell pepper. The R2 for Kc curves for watermelon were very high (> 0.99). Better results for the watermelon season was primarily due to fewer rainfall events in spring 2004, which reduced the uncertainties in water balance due to increased Ea. The FAO-PM based bi-weekly Kc for watermelon varied from 0.20 during the initial stages to 1.09 at maturity. The FAO-BC based Kc values varied from 0.21 to 1.41. The peak monthly Kc for watermelon using the FAO-PM method was

PAGE 84

71 almost identical to the reported value by Allen et al. (1998). ETc of watermelon was reduced by 31% by using drip irrigation compared to seepage irrigation. This study provides the Kc for watermelon and bell pepper which are the two commonly grown vegetable crops in Florida. The Kc values can be useful for scheduling irrigation. Since one season of data were used to develop the Kc for watermelon and bell pepper, the temporal variability due to climatic conditions could not be addressed in this study. The study should be continued for two more years to provide better Kc for the two crops. This study shows the effectiveness of drip system in terms of reducing ETc over the seepage system for bell pepper and watermelon production in Florida.

PAGE 85

CHAPTER 4 IMPACTS OF DRIP AND SEEPAGE IRRIGATION SYSTEMS ON THE WATER QUALITY FOR VEGETABLE PRODUCTION IN FLORIDA Introduction In a report submitted to the US Congress by the US Environmental Protection Agency (USEPA, 1994), agriculture was recognized as the primary source of water pollution. Agriculture contributes to pollution of water bodies in the form of sediments, pesticides and nutrients. The majority of pollutants from agriculture are transported to the water bodies from diffuse sources that do not have an obvious point of entry into a water body (Ongley, 1996), known as non-point source pollution (NPS). Water resources in all the states in the USA, including Florida, have been impacted by the NPS pollution. Growing concern over the pollution of water bodies in the country led to the enactment of the Clean Water Act, in 1977. This act provides the basis for regulating the discharge of pollutants into a water body within the United States. To provide a framework for addressing the growing NPS pollution and maintain water quality standards for water resources, a comprehensive Total Maximum Daily Load (TMDL) program for Florida was developed under section 303 (d) of the Clean Water Act. The TMDL program sets water quality standards by prescribing the maximum allocations for a pollutant load which can be allowed into a water body. To achieve the TMDL criterion in Florida, state agencies are encouraging the agricultural community to develop and adopt Best Management Practices (BMPs) to reduce nutrient loading from agricultural fields. 72

PAGE 86

73 Vegetable production is one of the most important parts of Florida’s agriculture. Vegetable production in south Florida, the main vegetable production area in the state, is carried out on sandy soils also known as flatwood soils. These soils are poorly drained and have low water holding capacity. To conserve moisture, the crops are produced on raised, pressed soil beds covered with plastic mulch (Clark et al., 1996). A large fraction of flatwood soils have a spodic layer, about 90-120 cm below ground surface. The spodic layer has higher clay and silt content than the soil above, and is cemented with Fe, Al, and organic matter. Due to the higher clay content and binding of its constituent materials, the spodic layer has low conductivity and, therefore, perches the water table. As a result, the water table varies from 45-60 cm below the soil surface. A high water table provides an opportunity for upward movement of water (upflux). Taking advantage of the high water table and shallow root zone of vegetable plants, vegetable farms in south Florida have traditionally been irrigated by seepage irrigation system that provides moisture in the root zone through upflux. A typical nutrient management practice for vegetable production under seepage irrigation involves application of inorganic fertilizer pre-plant in the plastic mulch beds. Inorganic fertilizer in the beds is exposed to leaching due to the fluctuations in the water table. Once out of the root zone, part of the fertilizer makes its way to surface and groundwater bodies. Loss of fertilizer from the root zone deprives plants of the required nutrients, which can adversely affect crop yield and farm income. In contrast to a seepage system, the drip irrigation system applies water precisely into the root zone and facilitates fertigation. Fertigation is the application of liquid fertilizer through irrigation water. Through fertigation, fertilizer is applied in smaller doses based on weekly crop nutrient

PAGE 87

74 requirement. Fertigation reduces the amount of fertilizer present in the soil at any given time during the growing season, thereby reducing the amount of potentially leachable fertilizer. Although fertigation reduces the potentially leachable fertilizer, water is the carrier of the pollutants from the agricultural fields into the water bodies. Therefore, sound water management practices are essential for controlling nutrient transport to surface and groundwater. To develop effective water management practices, it is important to understand the dynamics of major pollutants in soil and water. Nitrogen (N) and phosphorus (P) have been recognized by USEPA as the primary agricultural pollutants. Inorganic fertilizer is the source of P and N that are the macro nutrients required for plant growth. However, due to the increased use of fertilizer, N and P contamination of water bodies is of growing concern. P is the principal cause of eutrophication of lakes and reservoirs while increased N levels can cause toxicity to the aquatic life and the consumers of the affected water. N in the soil is present in organic and inorganic forms. However, it can be taken up by plants only in inorganic forms: nitrate (NO3) or ammonium (NH4). Bulk of inorganic N in the root zone comes from the inorganic fertilizers while mineralization of soil organic N and N-fixation contribute a small fraction of total inorganic N. NO3 is the most preferred form of N by the plants, but it is also the most mobile form of N in soil. High NO3 level in groundwater is a serious health hazard causing brain damage and even fatality in extreme cases (USEPA, 1985). Use of fertilizer P has doubled in the last 35 years (Terry et al., 1996). P can be lost from an agricultural field via runoff, erosion and leaching. Studies have found that P from an agricultural field is lost mostly through runoff and erosion, while leaching does

PAGE 88

75 not contribute greatly to the overall loss of P due to the ability of soil particles to fix P from soluble into immobile P. The immobile form of P is not available for plant use and may be lost via one of the hydrologic processes described above. However, due to the nature of flatwood soils, P loss through drainage can be important for Florida. Nutrient transport from an agricultural field to surface and groundwater is affected by water and nutrient management practices. Studies by Bogle et al. (1989), Csizinszky (1980), and Pitts and Clarke (1991) have compared the water quantity benefits of drip irrigation system against traditionally used seepage irrigation system. While the benefits of drip irrigation on water quality are apparent, they have not been quantified for vegetable production on flatwood soils in southwest Florida. High water table conditions combined with spatial variability in soil hydraulic properties create uncertainty for quantifying the impacts of water and nutrient management practices at field scale. Drainage lysimeters are considered an alternate tool to study the downward movement of chemicals in the soil profile. Liaghat and Prasher (1996) used drainage lysimeters to study the movement of pesticide in the soil profile while Klocke et al. (1993) used drainage lysimeters for water quality sampling. Drainage lysimeters were also used to study nitrogen uptake and leaching losses from citrus trees (Syvertsen and Smith, 1996). Similarly, Prunty and Montgomery (1991) and Martin et al. (1994) used drainage lysimeters to study the movement of N through the soil. Watts and Martin (1981) concluded that drainage lysimeters offer an effective tool to evaluate N management strategies for field conditions. The objective of this study was to quantify the impact of seepage and drip irrigation systems on N and P transport in the soil and to the groundwater.

PAGE 89

76 Material and Methods Lysimeter System Six drainage lysimeters were used in this study to investigate the water quality impacts of seepage and drip irrigation systems. Each lysimeter used in this study was 4.87 m 3.65 m 1.35 m. The design, construction and installation of the lysimeters are discussed in Chapter 3. For the purpose of comparison of the two irrigation systems, two of the four drip lysimeters (D1 and D4) were randomly selected and compared with the seepage lysimeters (S1 and S2). Cropping Practice Watermelon was grown in spring 2003 and 2004 cropping season, while bell pepper was grown in fall 2003. Summer is usually the non-cropping season in south Florida. Cow pea was grown as cover crop during the summer 2003 season to facilitate uptake of residual N and P and improve soil fertility. Table 4-1 describes the planting and harvesting schedule for the cropping seasons during the study period. Results from spring 2003 and fall 2003 seasons were used for. Spring 2003 watermelon crop was not harvested due to disease problem which resulted in pre-mature failure of the crop. Bell pepper was harvested three times in the fall 2003. Table 4-1. Crop production schedule for the lysimeters. Crop Season Crop Planting Date Harvesting (Date; DAT a ) First Second Third Spring 2003 Watermelon b 4/4/2003 Fall 2003 Bell Pepper 09/10/03 11/07/03 12/01/03 12/19/03 a Days after transplanting; b Crop was not harvested due to crop failure Fertilizer Management The N-P-K fertilizer application was based on University of Florida/Institute of Food and Agriculture Sciences (UF/IFAS) fertilizer recommendations for watermelon

PAGE 90

77 and bell pepper (Maynard et al., 2001). Table 4-2 describes the fertilizer application schedule and rates under the two irrigation systems. The extra fertilizer (N and K) in seepage lysimeters during spring 2003 (Table 4-2) was a result of failed transplanting early in the season. To match the amount of fertilizer applied with fertigation in drip lysimeters during the failed transplanting, an equivalent amount of fertilizer was applied in seepage lysimeters following the survival of plants from the fifth transplanting. Table 4-2. Crop, rates (kg/ha) and forms of fertilizer application for spring and fall 2003 seasons. Drip Seepage Pre-plant Pre-plant Crop Season Bottom Top Fertigationd Bottom Top Extra e N 45.36 a 95.76 45.36 122.64 18.48 P 112 b 112 Watermelon Spring 2003 K 45.36 95.76 45.26 122.64 18.48 N 56 a 168 56 168 P 168 b 168 Pepper Fall 2003 K 56 c 168 56 168 a Ammonium Nitrate (NH4NO3, composition 33-0-0); b Triple Super phosphate (P2O5, composition 0-46-0); c Muriate of Potash (K2O, composition 0-0-60); d Liquid Fertilizer (composition 8-0-8); e Mixture of NH4NO3 and K2O N Fertilizer application was carried out in two parts. The first part included applying dry fertilizer prior to making the soil bed (bottom mix). This application was common for both irrigation systems. The remainder part of fertilizer application in the seepage system involved placing the fertilizer in a band on top of the soil bed (top mix). Remaining fertilizer in the drip system was applied through weekly fertigation (Table 4-2). Fertilizer application for the surrounding experimental field followed the method that was used for drip lysimeters. Fertilizer P was applied as bottom mix in the beds to all the lysimeters and the field.

PAGE 91

78 Nutrient Cycling and Transport Monitoring To develop N and P balances under the two irrigation systems, soil, water and plant samples were analyzed for N and P species shown in Table 4-3. The inorganic N species nitrate-nitrogen (NO3-N) and ammonium-nitrogen (NH4-N) present in crop, water and soil were quantified in this study. The NOx-N in Table 4-3 is the sum of NO3-N and nitrite (NO2-N), which is referred as NO3 in this text. To compute N removed by crop harvest, fruit samples were analyzed for Total Kjeldahl Nitrogen (TKN). Table 4-3. Nitrogen and phosphorus analysis for soil, water and plant components for the lysimeters. Nitrogen Phosphorus Soil Water Plant Soil Water Plant NOx-N a NOx-N TKN c TP d TP e TP e NH4-N b NH4-N a Nitrate-Nitrogen; b Ammonium-Nitrogen; c Total Kjeldahl Nitrogen; d Total available phosphorus (Mechlich-1); e Total available phosphorus (Inductively coupled plasma (ICP) method). Similar to N, P is present in the soil in organic and inorganic forms. However, P is taken up by plants in the soluble form. Soluble P may be present in either organic or inorganic form. The amount of P in plant available form (soluble P) present in the soil was quantified by using Mehlich-1 procedure. Mehlich-1 extracts P from aluminum, iron and calcium phosphates present in soil (Pierzynski, 2000). Concentration of P in drainage and runoff was analyzed using inductively coupled plasma (ICP) spectrophotometry. ICP method was also utilized for determining P in the harvested crop. To quantify water and nutrient (N and P) balances for each lysimeter, water and nutrient (N and P) inputs and outputs to each lysimeter were measured independently. Water and nutrient components measured in this study and their sampling schedule are summarized in Table 4-4. Mechanism for monitoring water balance components such as

PAGE 92

79 irrigation, drainage, runoff, soil moisture and rainfall is discussed in Chapter 3. The monitoring schedule for N and P inputs and outputs (Table 4-4) is discussed below. Table 4-4. Monitoring and sampling schedule for hydrologic and nutrient input and output to the lysimeters. Time based Event based Monitoring System Components Pre-plant Post-harvest Daily Monthly Irrigation Drainage Runoff Soil moisture Hydrologic Rainfall Soil Fertilizer Ground Water Irrigation Drainage Runoff Nutrient Crops Sample Collection, Storage and Handling Soil Soil samples were taken Pre-planting and at post-season to compute the resident soil N and P. Pre-planting soil samples were taken for 0-30 cm and 30-60 cm soil depths before making beds in lysimeters. Post-season soil sampling included sampling both bedded and non-bedded areas of lysimeter. Soil samples from the bed were taken at 20 cm interval up to a total depth of 60 cm from the top of bed. Non-bedded samples were taken at 20 cm interval to a depth of 40 cm from the top of soil surface. Since average bed height was nearly 20 cm, 0-20 cm and 20-40 cm sample from the non-bedded area corresponded to 20-40 cm and 40-60 cm sample from the bedded area respectively. To obtain a composite soil sample for each depth, six sub-samples from six locations distributed within each lysimeter were collected. These sub-samples for a

PAGE 93

80 specific depth (e.g. 0-20 cm) were mixed well to obtain composite sample. Samples for bedded and non-bedded areas were analyzed separately. After collection, the samples were dried and stored as per the Florida Department of Environmental Protection (FDEP) protocol (FDEP, 2001) until analyzed. Soil samples were sent to the Analytical Research Laboratory (ARL) at Gainesville for analysis. Water As outlined in Table 4-4, water samples from the lysimeters were collected at fixed time intervals and on event basis. Drainage and/or runoff after a rain event was pumped out of the lysimeter using the discharge mechanism as described in Chapter 3. Only a part (1/10th) of drainage and runoff was stored for water quality samples (Chapter 3). Drainage and runoff discharge stored in the tanks was well stirred before collecting 125 ml sample (using peristaltic pump). After collecting the samples, the storage tanks were emptied and cleaned as per the FDEP protocols for cleaning water quality equipment. Monthly groundwater samples were collected using groundwater monitoring well installed in each lysimeter (Figure 3-10). Standard Operating Procedure (SOP) for groundwater sampling developed by FDEP (FDEP, 2001) was followed for collecting groundwater samples. Irrigation was collected monthly (Table 4-3). After collection, all the water samples were brought to the water resources laboratory at SWFREC, UF/IFAS, Immokalee, FL, for further processing, storage and shipping. At the laboratory, the samples were transferred from the 125 ml field sampling bottles to 20 ml scintillation vials. Each sample was divided into three different 20 ml scintillation vials for N and P tests (Table 4-3) based on the amount of sample required for each individual test. The samples were prepared in accordance with the SOP for each

PAGE 94

81 designated test and stored at 4oC (FDEP, 2001) before being shipped to ARL, Gainesville, FL, for analyses. Plant Tissue Fruit samples were collected and analyzed to quantify the amount of N and P removed from the lysimeters by each harvest. Harvested fruits were stored in separate containers for individual lysimeter. Five fruits were randomly selected from the harvest of each lysimeter to obtain a representative sample. For ease of drying, fruit samples were cut into smaller pieces and dried. Dried fruit samples were grounded and shipped to the ARL, Gainesville, FL, for N and P analysis. Nutrient Balance For each lysimeter, the inorganic N mass balance (steady-state) and the organic N balance (transient-state) was computed for each season. Steady-state N mass balance is based on the assumption that organic N in the soil is in steady-state (i.e. no net change in soil TDN). Transient-state mass balance shows the total N loss, which includes the conversions from organic N into inorganic N (Shukla, 2000). The N mass balance was computed as shown by Equation 4-1. To estimate the steady-state and transient-state N mass balance, DIN (dissolved inorganic N) and TDN (total dissolved N) in soil and water were computed using Equations 4-2 and 4-3. Mass balance was not computed for P, as Mehlich-1 test used in this study to evaluate soil P status is only an index of soluble P present in the soil and does not estimate the total soluble P in the soil. outinBALNNN Equation 4-1 Where, Nbal is the seasonal steady-state N balance; Nin and Nout are the input and output for the lysimeter.

PAGE 95

82 34NONHDIN Equation 4-2 3NOTKNTDN Equation 4-3 Where DIN and TDN are in kg/ha); NH4 is ammonium-nitrogen (kg/ha); and NO3 is nitrate-nitrogen (kg/ha); TKN is Total Kjeldahl Nitrogen (kg/ha). Total input and output were calculated using the following equations: fixiwfPinNNNNN Equation 4-4 glcrodoutNNNNN Equation 4-5 Where Np is rainfall N (kg/ha); Nf is the total inorganic fertilizer applied (kg/ha); Niw is irrigation water N (kg/ha); Nfix is the N fixation; Nd is drainage water N (kg/ha); Nro is runoff N (kg/ha); Nc is the N removed by crop harvest (kg/ha); and Ngl is the volatilization losses (kg/ha). For simplicity, Nfix was assumed to be similar for all the lysimeters. Due to low ammonia volatilization under plastic mulch condition, Ngl was assumed to be negligible. The unaccounted N (Nl) in the steady-state mass balance included losses through denitrification assuming no immobilization-mineralization turnover. Nl for the transient-state mass balance estimated denitrification losses including inter-conversions of organic and inorganic N. For both N mass balances, Nl was computed using equation 4-6. The change in soil storage (S) was computed using the resident soil N. SNNball Equation 4-6 Resident Soil nutrient The resident soil N and P before and after each season were estimated as follows: 01.0)()()/()/(3mdmkgsoilofkgmgMhaKgNutrientSoils Equation 4-7

PAGE 96

83 Where M is the nutrient concentration in soil (mg/kg of soil); s is the dry bulk density of soil (kg m-3); and d is the depth of soil layer (m). The average soil bulk densities for the A and E horizons computed for the lysimeters were used to estimate the resident soil N and P (Equation 4-7). Since the soil samples were taken only up to a depth of 60 cm for the 101 cm soil profile, nutrient level in the remaining 41 cm of soil below the sampling depth needed to be quantified to compute the total N and P storage in the lysimeters. N and P stored in the lower 41 cm of soil was quantified using groundwater concentration of the sample taken close to the soil sampling date. Volumetric water content for 60 70 cm measured from the Diviner was used to compute the soil N and P status as shown in equation 4-8. )()()/()/(haAFlVlmgChakgStatusNutrientSoilw Equation 4-8 Where C is groundwater concentration (mg/l); Vw is volume of water (l); A is the lysimeter area (ha); and F is the conversion factor to convert from mg/ha to kg/ha. However, groundwater sample collection started from the fall 2003 season. Therefore, N and P storage in lower 41 cm were computed using the soils data for the periods when groundwater sample was not taken. For computing resident N and P using soils data, it was assumed that N and P in E horizon was uniformly distributed throughout the profile. Therefore, the N and P concentration measured in the lowest soil sampling depth (30-60 cm) from the top of the ground surface was applied to entire 71 cm of E horizon. N and P storage in the coarse sand layer below 101 cm of soil profile was not considered in the balance.

PAGE 97

84 Irrigation, drainage and runoff discharge The N and P discharge in drainage and runoff and in irrigation for an individual event was computed using equation 4-9. )()()/()/()(haAlQlmgChakgLLoadingPandN Equation 4-9 Where C is the concentration of DIN or Total P; Q is the irrigation, drainage or runoff volume; and A is lysimeter area. The total loading for the entire season was estimated by Equation 4-10. niiLhakgLoadingPandNSeasonal1)/( Equation 4-10 Where L is the event based nutrient loading (kg/ha); i is the days after transplanting (DAT); and n is total number of days in the crop growing season. Crop The N and P content in the dried fruit samples were used for estimating N and P removed from the lysimeter system via crop harvest (Equation 4-11). Moisture content in bell pepper was assumed to be 90% of the total weight of the crop. 1.0)/( YMhakgremovalcropNutrient Equation 4-11 Where M is the total N or P content in fruits (mg/kg of dried fruits); Y is the yield (kg/ha); and 0.1 is the weight of the dry fruits (moisture content of fruits = 90%).

PAGE 98

85 Result and Discussion Nitrogen Transport in Spring 2003 Nitrate in soil Figure 4-1 shows the soil nitrate (NO3) in the 101cm soil profile (above the coarse sand) of the lysimeters. Soil NO3 levels in all the lysimeters before planting in the spring 2003 season were low. Low NO3 levels were mainly because the soil used in the lysimeters had not been farmed and did not receive any N fertilizer for many years prior to spring 2003. A series of wetting and drying of lysimeter soil during rebuilding of soil profile may have resulted in further loss of resident inorganic N from the lysimeter soil. Figure 4-1 shows a net gain in soil NO3 in spring 2003 season for all lysimeters. This increase was more prominent in the seepage lysimeters than the drip. On average, seepage lysimeters had 47 kg/ha more soil NO3 than drip lysimeters at the end of season. The differences in soil NO3 between the two systems could be due to 46 kg/ha more inorganic fertilizer N (NH4NO3) applied to seepage lysimeters compared to drip lysimeters during spring 2003 (Table 4-2). Higher fertilizer application for seepage system was due to pre-mature failure of watermelon crop. The 46 kg/ha additional fertilizer N applied to seepage lysimeters contained 35.6 kg/ha of NO3-N. While higher fertilization may substantiate the overall increase in soil NO3 in seepage lysimeters compared to drip lysimeters, it should be noted that soil NO3 is an ever changing concept depending upon contributions from nitrification and mineralization processes. As shown in Figure 4-1, the difference between the resident soil NO3 in drip and seepage lysimeters was more than the additional NO3 applied in the former, therefore, resident soil NO3 at the end of spring 2003 was not only due to unused fertilizer in the soil, but also included contributions from nitrification and mineralization processes.

PAGE 99

86 050100150200250Nitrate N in soil (0-101 cm) (kg/ha) D1 D4 S1 S2 Average Drip Average Seepage D195354200 D49271184 S197924137 S213942114 Average Drip94032142 Average Seepage118713126Spring 03-pre seasonspring 03-post seasonFall 03 pre seasonFall 03 post season Figure 4-1. Nitrate-N level in soil before and after crop season. Nitrate in drainage and runoff NO3 is a very mobile form of N and can easily leach through sandy soils to reach the groundwater (Watts and Martin, 1981). Therefore, application of more fertilizer NO3 could result in higher NO3 losses through drainage and runoff in seepage lysimeters. Figure 4-2 exhibits higher NO3 discharge with drainage and runoff from the seepage lysimeters compared to that from the drip lysimeters during spring 2003 season. Runoff loading was small during the season, constituting less than 2.5% of the total NO3 loading in drainage and runoff. Therefore, a sum of drainage and runoff loadings which was mainly drainage loadings, is shown in Figure 4-2. Hallberg and Keeney (1993) list the primary factors responsible for NO3 leaching as potentially leachable N (PLN), mass of infiltrating water and hydraulic conductivity. It can be safely assumed that hydraulic conductivity of the soil in all lysimeters would be similar. Therefore, main driving factors for the movement of NO3 in the soil were PLN and mass of infiltrating water.

PAGE 100

87 Main reason of higher PLN in seepage lysimeter compared to drip lysimeters was that all the fertilizer in the former was applied in a single application (pre-plant) in contrast to multiple applications (fertigation) for the latter. In addition to pre-plant application, PLN in seepage lysimeters was further increased with the addition of 35.6 kg/ha NO3 through extra fertilizer. Moreover, mass of infiltrating water in the seepage lysimeters was considerably higher than that under drip lysimeters. During spring 2003, 43% less irrigation was applied in drip the system than the seepage system. As a result, NO3 discharge in drainage and runoff from drip system was 52% less than those from the seepage system (Figure 4-2). 0102030405060708090100Nitrate-N in lysimeter discharge (kg/ha) Spring 03Fall 03TOTALCrop season Average drip system Average seepage system Figure 4-2. Average nitrate loading in drainage and runoff in the drip and seepage system during spring 2003 and fall 2003 seasons. Higher NO3 discharge is a function of discharge volume and concentration. As shown in Table 4-5, higher NO3 discharge in drainage and runoff from the seepage system compared to the drip system, was due to higher mean NO3 concentration from the former. Higher NO3 concentration reveals that more NO3 moved to the groundwater in the seepage lysimeters than the drip lysimeters.

PAGE 101

88 Table 4-5. Nitrate loading, mean concentration and flow weighted concentration for individual lysimeter drainage and runoff. NO3 loading (kg/ha) Mean concentration (mg/l) Flow weighted concentration (mg/l) Lysimeter Spring 03 Fall 03 Spring 03 Fall 03 Spring 03 Fall 03 D1 8.01 37.15 3.53 32.25 3.68 28.85 D4 12.92 58.03 6.03 29.46 5.94 34.03 S1 27.73 56.08 10.63 50.12 12.32 39.91 S2 16.18 93.87 8.10 40.65 9.90 50.72 Average drip 10.5 47.6 4.8 30.9 4.8 31.4 Average seepage 22.0 75.0 9.4 45.4 11.1 45.3 Ammonium and dissolved inorganic nitrogen in soil DIN (NO3 + NH4) was examined to better understand the exchanges from various N processes for computing the total unaccounted N from the lysimeters. On average, soil NH4 contributed nearly 27% of DIN in the lysimeters prior to spring 2003 season, but it was reduced to only 6% at the end of season (Figure 4-1 and 4-3). The biggest source of NH4 in the soil was inorganic fertilizer (ammonium nitrate), adding about 42 kg/ha and 31 kg/ha NH4 in seepage and drip lysimeters respectively. However, soil NH4 level shows an average increase of 1 kg/ha in all lysimeters at the end of the season. At the same time, soil NO3 increased by nearly 54 kg/ha on average in the lysimeters. Therefore, bulk of NH4 from the fertilizer was either lost through runoff and/or drainage or it was converted to soil NO3 by nitrification process or organic N via immobilization process. Immobilization and mineralization occur simultaneously in nature and the direction of transformation is governed by the ratio of organic carbon and N (C/N) (Shukla, 2000). If the ratio is less than 20, mineralization exceeds immobilization (Mulvaney et al., 1993). C/N ratio in all lysimeters was less than 20 in spring 2003. Therefore, mineralization may have further contributed to NH4 in the soil. Despite contributions from mineralization, marginal increase in NH4 at the end of season suggests that the bulk

PAGE 102

89 of NH4 was either lost through drainage and runoff or was converted to NO3 via nitrification. 050100150200250Dissolved Inorganic N in Soil (0-101 cm) (kg/ha ) D1 D4 S1 S2 Average Drip Average Seepage D1135758209 D4123018101 S1128427170 S21710014142 Average Drip124338155 Average Seepage159221156Spring 03-pre seasonspring 03-post seasonFall 03 pre seasonFall 03 post season Figure 4-3. Dissolved inorganic nitrogen in the soil before and after crop season. Ammonium and DIN in drainage and runoff During the spring 2003, NH4 contributed only about 10% of the DIN loading in drainage and runoff from the drip lysimeters and 12% from the seepage lysimeters (Table 4-6). Mean NH4 concentration in drainage and runoff (Table 4-6) was also less (< 2mg/l) compared to the mean NO3 concentration in drainage and runoff (Table 4-5). Therefore, bulk of NH4 (from fertilization and mineralization) was converted to highly mobile soil NO3. Comparison of the two systems (Table 4-6) shows that DIN discharge in drainage and runoff from the drip system was 53% less than that from seepage system. Higher DIN leaching from the latter was due to higher NO3 and NH4 concentrations (Table 4-5 and Table 4-6) suggesting higher leaching losses in the system.

PAGE 103

90 Table 4-6. DIN loading, mean NH4 concentration and flow weighted NH4 concentration for individual lysimeter drainage and runoff. DIN loading (kg/ha) Mean NH4 concentration (mg/l) Flow weighted NH4 concentration (mg/l) Lysimeter Spring 03 Fall 03 Spring 03 Fall 03 Spring 03 Fall 03 D1 9.33 44.52 0.65 5.43 0.60 5.72 D4 13.90 89.22 0.48 15.26 0.45 18.29 S1 31.76 78.84 1.47 20.97 1.79 16.20 S2 18.51 155.69 1.25 27.64 1.43 33.41 Average drip 11.65 66.90 0.56 10.34 0.52 12.00 Average seepage 25.15 117.30 1.36 24.30 1.61 24.80 Another reason for the difference in DIN loading from the two systems may be higher plant N uptake in drip lysimeters than the seepage system. Phene et al. (1979) reported that fertigation can increase fertilizer use efficiency by 200%. Higher N uptake in the drip system will also result in lower PLN in soil, and consequently lower the DIN leaching. To better understand the difference in DIN loading from the two systems, steady-state mass balance was computed for the four lysimeters. 020406080100120140160Dissolved inorganic N in lysimeter discharge (kg/ha) Spring 03Fall 03TOTALCrop season Average drip system Average seepage system Figure 4-4. Average DIN loading in lysimeter drainage and runoff for the drip and seepage system during spring 2003 and fall 2003 seasons.

PAGE 104

91 Nitrogen mass balance Table 4-7 shows the steady-state balance for the two systems for spring 2003. NBAL in the balance indicates amount of N retained in the soil after accounting for losses through hydrologic processes and crop removal. Nl indicates unaccounted inorganic N losses which cannot be directly computed (through denitrification, volatilization and immobilization-mineralization processes). Plastic mulch beds reduce volatilization losses to minimal. Therefore, Nl in this balance would primarily include losses through denitrification and immobilization mineralization turnover. Table 4-7. Steady-state N mass balance for drip and seepage systems for spring 2003. Parameters Spring 2003 (kg/ha) Drip Seepage Soil (N s-pre) 12.33 14.562 Inputs Fertilizer (Nf) 141.12 186.48 Rain (Np) 1.67 1.67 Irrigation (Niw) 1.69 3.00 Fixation (Nfix) a NIN 144.48 191.15 Outputs Drainage + Runoff (Ndl) 11.61 25.14 Crop Removal (Nc)b Gaseous losses (Ngl) c Negligible Negligible NOUT 11.61 25.14 NBALd 132.87 166.01 Soil (N s-post) 43.39 91.74 Soil Storage (S in) 31.06 77.23 Nl e 101.81 88.78 a Assuming similar N-fixation in each lysimeter; b No crop removal due to crop failure; c Assuming negligible gaseous losses from the plastic covered plant beds; d NBAL = NIN NOUT; e Nl = N BAL S Despite higher leaching losses from seepage system, unaccounted N (Nl) from the seepage system was 12.8 kg/ha lower than that from drip system. While 46 kg/ha more fertilizer N was applied to the seepage lysimeters compared to fertilization in drip lysimeters, soil DIN in the former at the end of season shows 46 kg/ha higher DIN in soil

PAGE 105

92 than the latter. Therefore, higher soil DIN can negate the extra fertilizer applied to the seepage system which suggests higher unaccounted N (Nl) from the drip system. Before concluding that the drip system had higher Nl than seepage system, it is important to consider some of the uncertainties in the mass balance shown in Table 4-7. Firstly, N uptake is considered as accounted loss from the system in this balance, primarily since the post-spring 2003 soil sampling done soon after crop failure did not allow enough time for decomposition of N in the plants. Therefore, N tied to the crop cannot be accounted in this balance. Secondly, variable mineralization rate (organic N to inorganic N) can cause uncertainty in the balance. As discussed previously, mineralization rate dependent upon C/N ratio. Pre-plant application of fertilizer N in seepage system may have resulted in higher mineralization rate than that in the drip system. The steady-state mass balance assumes no net contribution of N from organic to inorganic form. Therefore higher mineralization rate in seepage system would result in lower the DIN loss as the contributions from organic N cannot be accounted for in the steady-state mass balance. Nitrogen transport in fall 2003 Nitrate in soil Pre-fall 2003 season data (Figure 4-1) shows that there was loss of NO3 in the entire soil profile in all lysimeters during summer 2003. The decline was likely the impact of planting cow pea in all lysimeters during the season. Cow pea is a source for fixing atmospheric N in the soil. It can also convert resident inorganic soil N into organic soil N through plant uptake. Figure 4-1 shows that at the time of soil sampling prior to fall 2003, there were considerable differences in soil NO3 among the lysimeters. Overall, seepage system lost considerably more soil NO3 than drip system during summer 2003.

PAGE 106

93 However, these differences were not due to the irrigation treatment since the lysimeters were not irrigated during summer 2003. Figure 4-1 shows net gain in soil NO3 at the end of fall 2003 for both the systems. While soil NO3 gain for the two seepage lysimeters was similar, there was difference in the change in soil NO3 between the two drip lysimeters. Soil NO3 gain in D1 (143 kg/ha) was almost twice that in D4 (73 kg/ha) (Figure 4-1). Higher gain in D1 could be partly due to the lower NO3 loading through drainage and runoff in D1 as compared to D4 (Table 4-5). Lower drainage losses resulted in higher retention of soil NO3 in D1 compared to D4. Even though there was variability in the drip system, average residual soil NO3 in drip system was higher than that in the seepage system at the end of fall 2003. To better explain the differences and understand the movement of NO3 in the two systems, it is important to look into the NO3 concentrations in groundwater. Nitrate in groundwater Time series plot of NO3 concentration in groundwater (Figure 4-5) shows a sharp increase in groundwater NO3 concentrations in seepage lysimeters shortly after planting. In contrast to the seepage lysimeters, D1 and D4 show a more gradual rise in groundwater NO3 concentration. The sharp increase in seepage lysimeters was likely due to the application of 224 kg/ha of inorganic fertilizer N pre-planting in contrast to only 56 kg/ha of fertilizer N applied to drip lysimeters. Therefore, PLN in seepage lysimeters was much higher than that in the drip lysimeters. Rainfall shortly after fertilization (Sept 15th 2003) flushed out considerable NO3 from the root zone in all lysimeters. But due to lower PLN, NO3 leaching in drip lysimeters was less compared to that in S1 and S2.

PAGE 107

94 0204060801001201401601802007/238/129/19/2110/1110/3111/2012/1012/301/192/82/283/19TimeNO3-N concentration (mg/l)0102030405060RainfallRainfall (mm) Rain (mm) D1 D4 S1 S2 Planting Harvesting Figure 4-5. Nitrate concentration in groundwater (mg/l). Between the two seepage lysimeters, S2 shows comparatively higher groundwater NO3 peak in October 2003 (Figure 4-5). Higher NO3 leaching signifies higher leaching, as S2 shows nearly five times more NO3 discharge in drainage and runoff than S1 during September October 2003 (Table 4-8). Following peak groundwater NO3 concentration during October 2003, seepage lysimeters show gradual decline in concentrations (Figure 4-5), which was likely due to the dilution effect from rainfall. However, concentration in S1 and S2 still remain high (> 90 mg/l). More rainfall during 11/02/2003 11/06/2003, caused more leaching of NO3 into the groundwater in seepage lysimeters. Table 4-8. Total NO3 and DIN loading in drainage and runoff between groundwater sampling events during fall 2003. Period Nitrate Loading (kg/ha) DIN loading (kg/ha) D1 D4 S1 S2 D1 D4 S1 S2 Sept Oct 6.3 23.8 8.0 44.73 10.2 38.5 13 75.4 Oct Nov 24.8 18.5 31.4 37.3 27.9 30.7 45.4 63.9 Nov Dec 6.0 15.5 18.8 12.0 6.4 19.9 23.4 17.4

PAGE 108

95 In contrast to seepage, gradual increase in drip lysimeters is indicative of relatively lower movement of NO3 through soil profile in the drip lysimeters. However, increase in groundwater NO3 concentration in D4 was more gradual than that in D1. As shown in Figure 4-1, D1 had the highest resident soil NO3 amongst all four lysimeters before the beginning of fall 2003 season. Therefore, higher peak groundwater NO3 in D1 compared to D4 could have the effect of higher residual soil NO3 in the former. Despite the difference between replications, data shows that at the time of groundwater sampling in November (nearly 70 days after transplanting), the seepage system had lost nearly 28% of fertilizer N in drainage compared to only 16% loss in drip system. These losses further signify the role of fertigation in reducing NO3 leaching. Nitrate in drainage and runoff Due to a greater movement of NO3 to groundwater in seepage system than drip system, latter shows 37% reduction in NO3 discharge in drainage and runoff than the former (Table 4-5). Table 4-9shows that seepage lysimeters had slightly higher discharge (drainage + runoff) volume than drip lysimeters. But, considerably higher NO3 concentration (Table 4-5) in seepage system was responsible for higher NO3 discharge. Combining spring and fall 2003, drip lysimeters show a 40% reduction in NO3 discharge in drainage and runoff compared with seepage lysimeters for both seasons. Table 4-9. Total discharge volume (liter) for the lysimeters during study period. Lysimeters D1 D4 S1 S2 Average Drip Average Seepage Spring 2003 3821 3820 3955 2874 3821 3414 Fall 2003 2263 2997 2469 3252 2630 2861

PAGE 109

96 Ammonium and DIN in soil, groundwater, drainage and runoff Post-harvest soil NH4 levels in all lysimeters were higher in fall 2003 than those in post-spring 2003 (Figure 4-1 and 4-3). This was likely due to the contributions from organic N into NH4 from cover crop (cow pea) residue through mineralization process. Figure 4-3 shows similar change in DIN soil storage in S1 and S2 during fall 2003. However, D1 and D4 show considerable differences in change in DIN storage. 0204060801001201401601802007/238/129/19/2110/1110/3111/2012/1012/301/192/82/283/19TimeNH4 concentration (mg/l)0102030405060RainfallRainfall (mm) Rain (mm) D1 D4 S1 S2 Planting Harvesting Figure 4-6. Ammonium concentration (mg/l) in groundwater. As shown in Figure 4-6 the NH4 groundwater concentrations in seepage lysimeters were considerably higher than those in drip lysimeters. A higher groundwater concentration signifies higher leaching of NH4 in seepage lysimeters. Apart from the factors explained earlier responsible for higher leaching in seepage lysimeter, there could have also been a difference in mineralization rate between drip and seepage system.

PAGE 110

97 Figure C-1 shows greater water table fluctuations in the seepage system than the drip system during fall 2003. Water table fluctuations resulted in cycles of drying and wetting in the seepage system which may have contributed to higher mineralization rate (Agarwal et al., 1971). NH4 contributed nearly 33% of DIN loading from the lysimeter during fall 2003 (Figure 4-4) while fertilizer contained about 29% NH4, therefore, mineralization increased the NH4 level in soil leading to more leaching of NH4. 0102030405060Nitrogen (kg/ha) D1D4S1S2DripSeepageLysimeters 1st Harvest 2nd Harvest 3rd Harvest Figure 4-7. Nitrogen crop removal during fall 2003 season. Table 4-6 shows the drip system had 54% less NH4 discharge than seepage during fall 2003. Overall, the drip system had 43% reduction in DIN discharge in drainage and runoff (Table 4-6) compared to seepage system. For spring and fall 2003 season combined, drip lysimeters showed 44% reduction in DIN discharge compared with seepage lysimeters. Reduced DIN leaching and discharge under drip system was largely the effect of fertigation. N mass balance was computed to determine the overall losses from the two systems during fall 2003.To compute the N mass balance crop removal of N

PAGE 111

98 was quantified. Figure 4-7 shows the removal of nitrogen with the fruit harvest at the end of the fall 2003 season. While there were considerable differences between N removed from individual lysimeters, average N removed from the two systems was similar. Similar N removal suggest that overall, the plant uptake was similar in the two systems. Nitrogen mass balance Table 4-11 shows the steady-state mass balance for fall 2003. Due to higher DIN discharge, seepage system shows 48 kg/ha less NBAL than that in drip system. Contrary to NBAL, Nl in drip system show 28.5 kg/ha losses, while seepage system shows a net gain of DIN (Table 4-11). Overall, compared to spring 2003, overall unaccounted N in fall 2003 was considerably less. However, there could be uncertainties in steady-state mass balance mainly due to the conversion of organic N (from cow pea) into DIN through mineralization. The addition of DIN depends upon the mineralization rate of soil, which varies based on several local factors, discussed previously. Watts et al. (1991) computed 193 kg/ha mineralization on sandy soil. Variable rate of mineralization in the lysimeters can result in uncertainties in the balance. Therefore, to account for the N added through mineralization, a transient-state N mass balance was also computed for fall 2003. To compute the transient-state N mass balance only the change in organic N in top 30 cm of soil was considered. It was assumed that organic N was in steady-state conditions below 30 cm. This assumption was based on the fact that cover crop when ploughed into the soil remained mostly in top 30 cm of the soil. Therefore, bulk of organic matter was added only in the top 30 cm of soil. TKN analysis results (0-30cm) from post and pre-fall 2003 season were used. The post-season analysis of soil TKN result showed all lysimeters had similar TKN values for 0-20 cm and 20-40 cm. However, the TKN for 0-20 cm in S1 was nearly 178%

PAGE 112

99 higher than the average TKN in the other three lysimeters, while 20-40 cm show similar TKN in all four lysimeters. Significantly, high TKN for 0-20 cm in S1 could have been due to high spatial variability of organic N in the soil. Considering 0-20 cm TKN concentration was an outlier, average TKN of the remaining samples in S1 was used to compute organic N for 0-20 cm in S1. Table 4-10. Total Kjeldhal nitrogen (TKN) concentrations (mg/kg of soil) for soil samples during post-fall 2003 season. Lysimeter 0-20 cm 20-40 cm D1 240.37 267.56 D4 254.04 230.01 S1 720.90 278.46 S2 285.48 235.48 The transient-state N mass balance for fall 2003 shows that seepage lysimeters had net loss of 61.19 kg/ha N, while, drip lysimeter had a net gain of 79.69 kg/ha (Table 4-11). Net gain (negative balance) shows the uncertainties in the computations of mass balance. To better explain differences in transient-state mass balance, it is important to look into the variability in each system. Table 4-12 provides the Nl from each lysimeter. As shown in Table 4-12 three lysimeter had net losses, but D4 shows negative losses. High spatial variability in soil may have resulted in the unusual behavior of D4. Transient-state N losses computed in Table 4-12 illustrate the denitrification losses under the two systems for sandy soils in Florida. Denitrification primarily occurs under low supply of oxygen (Meisinger and Randall, 1991) which is controlled by soil water content (Follett et al., 1991). Denitrification was higher in seepage system due to higher wetted area than that in drip system. Smaller wetted area in drip system may have limited the denitrification losses only during wetting caused from rainfall.

PAGE 113

100 Table 4-11. Steady-state N and transient-state N mass balance for drip and seepage systems for fall 2003. Parameters Steady-state balance (kg/ha) Transient-state balance (kg/ha) Drip Seepage Drip Seepage Soil (N s-pre) 37.72 20.54 807.18 983.23 Inputs Fertilizer (Nf) 224.00 224.00 224.00 224.00 Rain (Np) 1.67 1.67 1.67 1.67 Irrigation (Niw) 1.56 3.00 1.56 3.00 Fixation (Nfix) a NIN 227.23 228.67 227.23 228.67 Outputs Drainage + Runoff (Ndl) 67.45 118.06 74.05 153.52 Crop Removal (Nc) 13.80 13.21 13.80 13.21 NOUT 81.24 131.27 87.85 166.73 (NBAL) c 145.98 97.40 139.38 61.94 Soil (N s-post) 155.16 155.70 1026.25 983.98 Soil Storage (S in) 117.44 135.16 219.07 0.75 Nl d 28.54 -37.76 -79.69 61.19 a Assuming similar N-fixation in each lysimeter; b Assuming negligible gaseous losses from the plastic covered plant beds; c NBAL = NIN NOUT; d Nl = N BAL S Large variability in results could have been due to combination of following reasons. 1. Variability in oxygen content in the soil due to variable soil moisture mainly in non-bedded areas resulted in different aerobic and anaerobic environments. 2. Differences in mineralization rate in each lysimeter. 3. Errors in computations due to extrapolation of point soil samples to compute the overall nutrient status in the soil While there was variability, overall, the drip system had comparatively lesser unaccounted transient-state N than the seepage system, which was largely the effect of fertigation and lower mass of infiltrating water. Table 4-12. Transient-state N mass balance for the lysimeters during fall 2003. Parameters Transient-state balance (kg/ha) D1 D4 S1 S2 Nl a 37.76 -197.15 38.07 84.31 a Nl = N BAL S

PAGE 114

101 Phosphorus Transport Spring 2003 The primary difference between the comparison of N and P loading for the two irrigation systems is that while N was split applied as fertigation in the drip lysimeters, all the P fertilizer was applied pre-planting as bottom mix to all the lysimeters (Table 4-2). Due to same application method, the differences in P transport will solely be the result of differences in the movement of water under the two systems. Melich-1 extractable P in top 101 cm of soil is termed as soil P in this discussion. Results from pre-spring 2003 (Figure 4-8) show that there were differences in soil P before fertilizer application in the lysimeters. Drip system on an average had 71 kg/ha higher soil P compared to the seepage system. However, at the end of spring 2003, results show 207 kg/ha more soil P in seepage than the drip system. Moreover, there were differences among the lysimeters in level of soil P at the end of spring 2003. All the lysimeters except D4 show a net gain in soil P during the season. S1 and S2 show an increase of 286 and 277 kg/ha respectively while D1 show an increase of 103 kg/ha. At the time of planting, 112 kg/ha of fertilizer P was applied to all lysimeters (Table 4-2). Fertilizer P is highly soluble and available for plant uptake (Busman, 1998). Moisture starts dissolving fertilizer P soon after fertilization. As soluble P moves away from the fertilizer particle, it reacts with constituents present in the soil such as calcium, magnesium, iron and aluminum. Soluble P which forms complex compounds with these constituents becomes unavailable to plant uptake, adding to the fixed P pool in the soil. However, the migration of soluble P is very slow and depends on the movement of water (Busman, 1998). Cyclic saturation of the root zone in the drip system provided an opportunity for accelerated movement of soluble P as opposed to that in the seepage

PAGE 115

102 system. Figure C-1 shows greater water table variations in the seepage lysimeters as compared to the drip lysimeters, in response to the irrigation, which resulted in the cycles of drying and wetting in the former. It is possible that movement of water under the drip system facilitated soluble P to associate with constituents in the soil as compared to that in the seepage system. However, the increase in soil P, which was more than the fertilizer P applied, in seepage lysimeters can be a result of high spatial variability. 0100200300400500600700Extractable soil-P (kg/ha) D1 D4 S1 S2 Average Drip Average Seepage D1245348233233 D4364265228265 S1299585194248 S2166443258370 Average Drip304307231249 Average Seepage233514226309Spring 03-pre seasonspring 03-post seasonFall 03 pre seasonFall 03 post season Figure 4-8. Extractable soil P status for spring 2003 and fall 2003 seasons. Difference in the net gain of soil P between the two systems could be due to differences in the drainage and runoff discharge of P. Figure 4-9 shows that there was slight difference in discharge of P in drainage and runoff between the two systems. But the total P discharge from the lysimeters was extremely low (< 1 kg/ha) during spring 2003 season. Therefore bulk of fertilizer P applied to the lysimeters did not leave the

PAGE 116

103 lysimeters but was still present in the soil at the end of spring 2003. While seepage lysimeter showed very high soil P (Figure 4-8) at the end of season, drip lysimeters showed only slight increase. Therefore, bulk of fertilizer P in drip system may have converted into non-labile P in the soil. Some sources of variability in the computation of soil P before and after spring 2003 were: firstly, variability due to extrapolation of point soil sampling to compute the P in the entire soil P. Secondly, soil sampling was only done to a depth of 60 cm. Therefore, P in lower 41 cm was extrapolated from the P concentration in 30-60 cm soil sample. The variability in the vertical distribution of soil P can also result in differences in the computed soil P. Thirdly, there was some mixing of the P (fixed pool) rich spodic material in the E horizon of lysimeter soil during the rebuilding of soil profile in the lysimeters. The spodic material may act as the source of P. However, differences in water regimes in the drip and seepage system may have resulted in differences in the release of P from the spodic material in the two systems. 0123456Total P in lysimeter discharge (kg/ha) Spring 03Fall 03TOTALCrop season Average drip system Average seepage system Figure 4-9. Total P loading in lysimeter discharge in spring 2003 and fall 2003 seasons.

PAGE 117

104 Overall, drainage data (Figure 4-8) suggested that downward movement of soluble P was more prevalent in the drip lysimeters than the seepage lysimeters resulting in 42% higher drainage discharge of P in the former during spring 2003. On the other hand, drip showed 31% reduction in loading through runoff compared to seepage lysimeters. Fall 2003 Figure 4-8 shows differences in soil P (available P) in the lysimeters before and after the fall 2003 season. However, the differences in available P between the lysimeters were considerably reduced compared to the differences at the end of spring 2003. The resident soil P at the end of fall 2003 in D4 show an increase of 37 kg/ha while D1 did not show increase (Figure 4-8). Resident P in S1 increased by 54 kg/ha and by 112 kg/ha in S2. However, changes in soil P in the two systems can be explained by difference in leaching and soil P storage especially below 60cm of soil. During fall 2003, 168 kg/ha of fertilizer P was added to all the lysimeters. Total loading through drainage and discharge was less than 5 kg/ha from the lysimeters during the season. It suggests that the bulk of fertilizer P should still be present in the soil. While there was accumulation of soluble P in the soil at the end of fall 2003, the accumulation does not account for the total fertilizer applied. There is a possibility that the P may have converted into immobile P and is no longer in available form. Lower increase of resident P in D4 compared to S1 and S2 could be due to variable rate of conversion of soluble P into immobile P.

PAGE 118

105 0246810121416Phosphorus (kg/ha) D1D4S1S2DripSeepageLysimeters 1st Harvest 2nd Harvest 3rd Harvest Figure 4-10. Total P removed with harvest crop in fall 2003 season. Similar to spring 2003, drainage P discharge in fall 2003 from drip systems was nearly twice that from the seepage lysimeters. On the other hand, runoff P discharge from the two systems was similar (Table 4-13). Higher P discharge in drainage in fall 2003 further confirms the presence of downward movement in drip system. However, P crop uptake was similar in drip and seepage system (Figure 4-10) Table 4-13. Total P discharges (drainage and runoff) and mean P concentration in lysimeter drainage and runoff. P loading (kg/ha) Mean P concentration (mg/l) Lysimeter Spring 03 Fall 03 Spring 03 Fall 03 D1 1.03 0.61 0.46 0.23 D4 0.94 9.06 0.35 4.12 S1 0.79 1.40 0.32 0.80 S2 0.73 3.28 0.33 1.63 Average Drip 0.98 4.84 Average Seepage 0.76 2.34 The primary reason for low P loading through drainage and runoff was the very low mobility of P which resulted in low P concentrations (Table 4-13). The mean concentration of P in lysimeter discharge (drainage + runoff) was less than 0.5 mg/l in all

PAGE 119

106 the lysimeters during spring 2003. It increased during fall 2003, but still remained less than 5 mg/l. Drainage and runoff concentrations in this study were higher than Campbell et al (1985) who reported P concentration of less than 2 mg/l in drainage and runoff for soils in Florida. However, they reported results were from a field study which can have higher dilution in the non-cropped areas as opposed to the closed lysimeter system. Results from this study show that P does not easily move through the soil and the potential for P to leach into groundwater is low. Soil particles adsorb the soluble P as the infiltrating water passes through the soil profile into groundwater. 0123456789107/23/20039/11/200310/31/200312/20/20032/8/20043/29/2004TimeP concentration (mg/l)0102030405060RainfallRainfall (mm) Rain (mm) D1 D4 S1 S2 Figure 4-11. Total P concentration in groundwater (mg/l) and daily rainfall (mm). Figure 4-11 show the time series plot of groundwater concentration of total P during fall 2003. Data in Figure 4-11 show low concentrations (< 0.2 mg/l) in all the lysimeters during in August 2003. After the application of fertilizer at the time of planting in fall 2003 (Table 4-2) total P in groundwater shows an increasing trend in all lysimeters. Lysimeter D4 shows a sharp increase in the groundwater P concentration

PAGE 120

107 during October and November 2003 while other lysimeters shows more gradual trend. As explained earlier, the migration of P in soluble form away from the fertilizer is dependent upon the movement of water. Therefore, it is possible that rainfall during September and October 2003 pushed a plume of soluble P from the root zone into the ground water. The timing of peak P concentration soon after rainfall events is similar to that seen in the emergence of groundwater concentration peak for NO3 and NH4, smaller peaks signifying low mobility of P. Table 4-14 shows drainage P discharge from D4 was considerably high compared to other lysimeters between groundwater sampling events in September and October. Therefore, it was the effect of rainfall which resulted in leaching of soluble P from the root zone into the groundwater. While groundwater concentration falls in D4 after October 2003, other lysimeters continue to show an increase. Table 4-14. Drainage and runoff P discharge between groundwater sampling events during fall 2003. Period Drainage Loading (kg/ha) Runoff loading (kg/ha) D1 D4 S1 S2 D1 D4 S1 S2 9/10 -10/19 0.3084 5.1836 0.5335 1.3858 0.0507 0.1634 0.0891 0.0603 10/19 11/12 0.2057 2.5862 0.4324 1.1862 0.0038 0.0029 0.0199 0.0439 11/12 12/20 0.0387 1.1266 0.4115 0.6512 0.0032 0.0009 0.0197 The main conclusion which drawn from this study is that bulk of applied P does not easily moves through the soil adding to the resident soil P. Results also showed that drip system had comparatively higher P loading in drainage than the seepage system. These results were consistent with Zhang et al. (2002) who indicated that sandy soils were at high risk of P leaching due to their low sorption capacity. Top down movement of water in drip system was likely the reason of higher leaching of P than in seepage system.

PAGE 121

108 Summary and Conclusion Four drainage lysimeters were used to compare the water quality impacts of drip and seepage irrigation systems. Hydrologic and nutrient (N and P) input and output data from spring 2003(watermelon) and fall 2003 (bell pepper) seasons were used in this study. Effects of fertigation (drip irrigation) versus pre-plant (seepage irrigation) N application were investigated in this study. All the fertilizer P was applied pre-planting in the beds in all lysimeters. N mass balance was computed to study N dynamics under the two systems. The effects of different water regimes in the two systems on the movement of available P were also quantified. Based of the results from this study following conclusions can be drawn: 1. Drip irrigation system reduced the NO3-N discharge in drainage and runoff by nearly 40% compared to those from the seepage system. DIN discharge in drainage and runoff was reduced by 44% in the drip system. 2. Lower DIN discharge from the drip system was mainly due to lower PLN (effect of fertigation) and lower mass of infiltrating water (effect of irrigation practice) in these lysimeters compared to the seepage lysimeters. 3. The transient-state N mass balance indicates that drip system had lower denitrification losses than the seepage system 4. In contrast to DIN discharge, P discharge in drainage and runoff was only a small fraction of total P applied as fertilizer. Comparatively, P discharges (drainage + runoff) from the drip system were 87% higher than the seepage system. 5. Bulk of fertilizer P converted into non-labile P adding to the resident soil P. 6. There was no difference in N and P plant uptake between the two systems.

PAGE 122

CHAPTER 5 SUMMARY, CONCLUSIONS AND SUGGESTIONS Summary and Conclusions This study was conducted at the University of Florida-Institute of Food and Agricultural Sciences (UF/IFAS) research center at Immokalee, Florida, to develop local crop coefficients (Kc) for watermelon and bell pepper for Florida. The study also quantified the impacts of drip and seepage irrigation systems on crop evapotranspiration (ETc) and water quality for vegetable production in Florida. Six drainage lysimeters (4.87 m 3.65 m 1.37 m) were constructed and installed in a vegetable field (area = 0.70 ha). The soil profile inside the lysimeters was rebuilt using the native soil, Immokalee fine sand (A horizon = 0.30 m; E horizon = 0.71 m) by light compaction followed by wetting and drainage cycles. Four lysimeters were irrigated by drip and two by a seepage irrigation system. Irrigation, drainage and runoff were measured individually for each lysimeter. Drainage was collected through a horizontal well screen installed at the bottom of lysimeter. Runoff was captured by the walls of the lysimeters extending above the ground surface and collected in two catchments attached to the lysimeter wall. Soil moisture was measured at two locations in each lysimeter using a frequency domain reflectometry (FDR) device. Irrigation scheduling was based on daily soil moisture measurements. Irrigation scheduling was not based on water table measurements for seepage irrigated lysimeters, which is practiced by some growers in south Florida. Daily weather data to estimate reference evapotranspiration (ETo) were collected at the UF/IFAS Florida 109

PAGE 123

110 Automated Weather Network (FAWN) weather station located at SWFREC, Immokalee. Bi-weekly and monthly water balances were computed for each lysimeter to compute the ETc and develop Kc. Watermelon was planted in spring 2003 and 2004; bell pepper was planted in fall 2003. N-P-K fertilizer application rates were based on the UF/IFAS recommendations. Soil samples were collected before and after each season to compute the resident soil nitrogen (N) and phosphorus (P). Drainage and runoff sample were collected on event basis while, groundwater was collected monthly. Fruits were sampled to compute the removal of N and P with the harvested crop. Bi-weekly Kc for bell pepper (fall 2003) using FAO-Penman-Montieth (FAO-PM) varied from 0.50 (initial) to 1.18 (maturity) and 0.51 to 1.07 for FAO-BC based curve. Third order polynomial equations, used to describe the Kc curve, yielded high correlation (R2 FAO-PM = 0.79; R2 FAO-BC = 0.60). Monthly Kc curves followed the conventional crop curve more closely than the bi-weekly curves. Peak monthly Kc for bell pepper (1.01) developed in this study was similar to the reported value (1.05) by Allen et al. (1998). The ETc for bell pepper was reduced by 20% under the drip system compared to that in seepage system. Bi-weekly Kc for watermelon (spring 2004) using FAO-PM varied from 0.20 to 1.09 and from 0.21 to 1.41 for FAO-BC. Very high correlation (R2 > 0.99) for watermelon compared to those for bell pepper was mainly due to the effect of fewer rainfall events during the spring 2004. Rainfall introduces uncertainty in water balance due to increased evaporation. The ETc of watermelon was reduced by 31% by using the drip system compared to that under the seepage system.

PAGE 124

111 N discharges from the drip system were considerably smaller than those from the seepage system. Overall, drip system reduced nitrate (NO3-N) loading by 40% and DIN loading by 45 %. Reduced losses from the drip system were the effect of fertigation and lower mass on infiltrating water compared to the seepage system. TDN mass balance shows that drip system had lower denitrification losses than the seepage system. Only a small fraction (< 5 kg/ha) of fertilizer P was lost in drainage and runoff. Bulk of unused fertilizer was converted into non-labile P adding to the resident soil status. In contrast to runoff, P discharge through drainage was higher. P concentration in drainage remained low (< 1 mg/l) for most part of the season, but, occasionally reached as high as 4.12 mg/l. Higher P loading in drainage was likely because of low sorption capacity of the soil. Comparatively, P discharge was 87 % higher from drip system than the seepage system. Higher P loading from drip system was likely due to cyclic downward movement of water. It should be noted that the use of lysimeter in this study provides a closed system, which is one of the major differences between this study and an actual field condition. The ET and the nutrient leaching for the same crop reported in this study may be different from an actual vegetable farm condition due to differences in water management, soil, and hydrologic factors. Suggestions This study does not address the temporal variation due to rainfall, as only one season was used to compute water use and Kc for both crops. Abukhalaed et al. (1982) recommended data collection for a minimum of three seasons for such studies. Spatial variability in N and P added uncertainty in quantifying the N and P storage in the lysimeter soil. However, results from this study open avenues for furthering current

PAGE 125

112 research to obtain high quality data for developing better water and nutrient management practices for southwest Florida. Based on the results from this study, following suggestions may be helpful for future research work 1. Study should be continued to minimum of three seasons for computing ETc, develop Kc and better understand the N and P dynamics. 2. The higher P discharge seen in drainage resulting from the use of drip system compared to seepage system in sandy soils warrants further study.

PAGE 126

APPENDIX A WEATHER DATA

PAGE 127

114Table A-1. Monthly summary of historical weather data and fall 2003 weather data. Month Fall 2003 Season Historical data a Air temperature (oC) d Total ETo h (in mm/day) Tmean b (oC) Rain c (mm) Tmean Tmax Tmin Rs e (MJ m-2 d-1) U2 f (ms1) Rain g (mm) FAO-PM FAO-BC September 27.14 166.4 25.8 33.5820.38 17.02 1.45 201.5 4.0 3.5 October 24.64 69.3 24.3 32.5 14.28 17.19 1.45 21.4 3.6 3.8 November 21.56 58.2 21.4 30.546.68 13.74 1.66 52.7 2.8 3.2 December 18.75 45.2 15.7 27.94-0.2 12.79 1.66 50.9 2.3 2.5 Total i 339.1 326.5 322.7 336.0 a 30-yr historical data for Immokalee, b Historical mean air temperature for Immokalee; c Historical average rainfall for Immokalee; d Air temperature (fall 2003); e Daily incoming solar radiation; f Wind speed at 2 m height; g Rainfall (fall 2003); h Average daily ETo; i in mm.

PAGE 128

115of historical weather data and spring 2004 weather data Fall 2003 Season Table A-2. Monthly summary Historical data a Air temperature (oC) d Total ETo h (in mm/day) Month Tmean b (oC) Rain c (mm) Tmean Tmax Tmin Rs e (MJ m-2 d-1) U2 f (ms1) Rain g (mm) FAO-PM FAO-BC March 20.56 75.69 19.7 30.156.87 19.87 1.87 2.0 4.40 3.94 April 22.22 61.46 20.85 32.156.71 22.46 1.87 57.1 4.72 3.82 May 24.92 110.23 24.49 35.9 13.63 23.50 1.87 15.8 5.26 4.14 Total i 247.38 73.5 390.53 321.14 a 30-yr historical data for Immokalee; b Historical mean air temperature for Immokalee; c Historical average rainfall for Immokalee; d Air temperature (spring 2004); e Daily incoming solar radiation; f Wind speed at 2 m height; g Rainfall (spring 2004); h Average daily ETo; i in mm.

PAGE 129

APPENDIX B OEIMER WATER INPUTS AND OUTPU TS F R TH LYS ET S

PAGE 130

117Table B-1. Summary of rainfall, irrigation, drainage and runoff during fall 2003 season. Rain Irrigation (mm) Dr ainage (mm) Runoff (mm) DAT a (mm) D1 D3 D4 S1 S2 D1 D3 D4 S1 S2 D1 D3 D4 S1 S2 0-20 104.3 12.0 8.3 8.2 16.2 18.0 71.8 79.2 73.7 66.0 78.6 11.0 15.0 14.3 15.4 8.6 21-34 1.0 16.1 19.7 20.3 20.6 21.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 35-48 1.1 25.5 26.5 28.7 40.6 38.8 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 49-62 68.1 8.9 10.1 9.9 18.3 18.2 32.8 35.8 35.0 30.3 37.5 2.2 2.4 1.0 2.8 7.3 63-76 2.6 19.3 20.0 25.1 33.1 27.8 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 77-90 9.6 23.0 24.4 25.0 43.0 36.2 0.0 2.7 4.5 9.9 9.0 0.0 0.0 0.0 0.0 0.0 91-100 39.6 4.7 5.5 5.7 3.1 0.0 6.3 13.8 15.9 12.6 16.4 1.6 0.7 0.4 0.0 3.7 TOTAL 226 109 114 123 175 160 111 131 129 118 141 14.8 18.1 15.7 18.2 19.6 a DAT is days after transplanting

PAGE 131

118n. Runoff (mm) Table B-2. Summary of rainfall, irrigation, drainage and runoff during spring 2004 seaso Rain Ir Draina rigation (mm) ge (mm) DAT (mm D1 D3 S2 D1 D3 D4 D1 D3 D4 S1 a ) D4 S1 S1 S2 S2 0-14 20.2 18.2 18.6 25.8 25.6 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.5 15-28 0.4 25.4 26.2 21.9 51.7 45.8 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 6.5 29-42 14.8 14.5 12.5 16.2 16.2 5.3 5.3 4.4 6.8 6.6 0.0 0.0 0.0 0.0 0.0 43-56 37.1 24.5 22.8 20.7 49.8 47.4 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 21.8 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 34.5 57-70 71-81 TOTAL 20.7 13.4 116 18.8 10.1 103 73.1 69.9 286 70.2 63.4 269 0.0 74 10.9 118 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 5.3 5.3 4.4 6.8 6.6 0.0 0.0 0.0 0.0 6.5 a DAT is days after transplanting

PAGE 132

119 WATER TABLE DATA APPENDIX C -120 -110 -100 -90 -80 -70 -60 -50 -40 -30 -20 -10 0 9/139//25/110/19111/1211/1811//30612/1DateW Tablm)0 10 20 30 40 50 60Rain (mm) 1991010/7 1310/10/250/31611/ 241112/12/12 8 ater e (c Rainfall Drip Seep Figure C-1. Water table for the drip and the seepage system and rainfall during fall 2003.

PAGE 133

120 -120-110-100-80-60-40-2003/1333/24/2/74/275/25/75/125/17DateWatr table (m)0102030405060Rain (mm) -90 -70e -50c -30 -10 3/8 3/18/2384 4/124/174/22 Rall ainf Dr ip See p Figure C-2. Water table for the drip and the seepage system and rainfall during spring 2004.

PAGE 134

121 LIST OF REFERENCES Aboukhaled, A., A. Alfamith. 1982. Lysimeters. FAO Irrigation and drainage paper 39. Rome, Italy: Food and Agriculture Organization of United Nations. Aceves, M. B., and L. Dendooven. 2001. Nitrogen, carbon and phosphorus mineralization in soils from semi arid highlands of central Mexico amended with mineralization as affected by drying-wetting cycle. Soil Science Society of America proceedings 35: 96-100. an for all seasons. Journal of Irrigation and Drainage Engineering 112(4): 348-368. . G., L. S. Per, Daes, and M. Smith. 1998. Crop evapotranspiration. Guidelines for computing crop water requirements. FAO Irrigation and Drainage paper 56. Romlture Organization of United Nations. ith, A. Perrier, and L. S. Pereira. 1994. An update for the definition of reference evapotranspiration. International Commission on Irrigation and Drainage Bulletin 43(2): 1-34. pact of drip and surface irrigation on growth yield and WU (Capsicumral WaManagement 65: 121-132. ic and Business Research, UFlorida population: Census summary 2000. Gainesville, FL: Bureau of Economic and Business Research, University of Florida. A micro-lysimeter method for determining evaporation fromescription and laboratory evaluation. Soil Science Society of America Journal 46: 689-696. furrow irrigation on plastic mulched and bare soil for tomato production. Journal of American Society of Horticultural Sciences 114(1): 40-43. Agarwal, A. S., B. R. SiAllen, R. G. 1986. A PenmAllAllen, R.G., M. SmAntony, E., and R. B. Singandhupe. 2004. ImBureau of EconomBoast, C. W., and T. M. Robertson. 1982. Bogle, C.R., T.K. Hartz, and C. Nuntoez. ngh, and Y. Kanehiro. 1971. Soil nitrogen and carbon en, R eira. R e, Italy: Food and Agricu E of capsicum annum L.). Agricultu ter niversity of Florida [BEBR-UF]. 2001. bare soil: d 1989. Com parison of subsurface trickle and tannery sludge. Bioresource Technology 77: 121-130. ro, and M. S

PAGE 135

122 Broner, I. 2001. Irrigation scheduling. In Crop Series: Irrigation, Colorado State University Cooperative Extension. Fort Co llins, CO: Colorado State University, Available at http://www.ext.colostate.edu/pubs/crops/pubcrop . Accessed on June 21, 2004. Brouwe, C., and M. Heibloem. 1986. Irrig ation water needs. Irrigation water management: Training Manual No. 3. Ro me, Italy: Food and Agriculture Organization of United Nations. Burman, R. D., P. R. Nixon, J. L. Wright, a nd W. O. Pruitt. 1980. Water requirements. In Design and Operation of Farm s. Ed. M. E. Jensen. St. Jo Michigan: American Society of Agricultural Engineering [ASAE]. Busman, L., J. Lamb, G. Randall, G. Rehm, and M. Schmitt. 1998. Nature of P in soils. FO-06795-GO. St. Paul, MN: University Capbell, K. L., J. S. Rogers, and D. potato production. Transactions of ASAE 28(6): 1798-1801. Cetin, O., and L. Bilgel. 2002. Effects of different irriga tion methods on shedding and yield of cotton. Agricultural Wa ter Management 54(1): 1-15. Chalmers, D. J., P. K. Andrews, K. M. Harris, and E. A. Cameron. Performance of drainage lysimeters for the evaluation of water use by Asian Pears. HortScience 27(3): 263-265. Chow, V. T. 1964. Handbook of Applied Hydrology. New York, NY: McGraw-Hill Book Co. croclimate modifications of a permanent rain shelte red lysimeter system. Transactions of ASAE 33(6): 1813-1822. Clark, G.A., E.E. Albergts, C.D. Stanley, A.G. Smajstrla, and F.S. Zazueta. 1996. Water requirements and crop coefficients of drip-irrigated strawberry plants. vegetable crops with seepage and drip irrigation systems. Florid a Scientist 43(4): 285-292. Doorenbos, J., and A. H. Kassam 1979. Yield response to water: FAO Irrigation and Nations. Irrigation System seph, of Minnesota, Extension Service. m R. Hensel. 1985. Drainage water quality from Clark, G.A., and D.L. Reddell. 1990. C onstruction details and m i Transactions of AS AE 39(3): 905-913. Csizinszky, A.A. 1980. Yield and water use of Dingm an, S. L. 1993. Physical Hydrology. U pper Saddle River, NJ: Prentice Hall. Drainage paper 33. Rome, Italy: Food and Agriculture Organization of United

PAGE 136

123 Doorenbos, J., and W. O. Pruitt. 1977. Crop water requirements. FAO Irrigation and Drainage paper 24. Rome, Italy: Food and Agriculture Organization of United Nations. Dugas, W. A., and D. R. Upchurch. 1984. Micr oclimate of a rainfall shelter. Agronomy Journal 76(6): 867-871. Fares, A., and A. K. Alva. 2000a. Evaluation of capacitance probes for optimal irrigation onitoring in an entisol profile. Irrigation Science 19: 57-64. Fares, A., and A. K. Alva. 2000b. Soil water components based on capacitance probes in of America Journal 64: 311-318. Florida Department of Environmental Protection [FDEP]. 200 1. Department of Environmental Protection Standard Oper ating Procedures for Field Activities. ental Assessment Section. Florida Department of Environmental Protection. Tallahassee, FL. Available at te.f l.us/labs/qa/sops.htm . Accessed on June 26, 2004. Fernald, E. A., and D. J. Patton. 1984. Wate r resource atlas of Fl orida. Tallahassee. Florida: Florida State University. Follett, R. F, D. R. Keeney, and R. M. Cruse. 1991. Managing Nitrogen for Groundwater Quality and Farm Profitability. Madison, WI: Soil Science Society of America Inc. Fougerouge, J. 1966. Quelques problems de bioclimatologie en Guyanne Francaise. L’Agron. Tropicale 3: 291-346. Gangopadhyaya, M., G. E. Harbeck, T. J. Nordenson, M. H. Omar, and V. A. Uryvaev. 1996. Measurement and estimation of eva poration and evapotranspiration. World te No. 83, WMO No. 201TP.105: World Meteorological Organization. . e.ng, R. L. Snyder, J. J. Carroll, and W. George. 1998. Crop coefficient: The key to im crop yield. Irrigation Journal July/August Hallberg, G. R., and D. R. Keeney. 1993. Nitr ate. In Regional Ground-Water Quality, Ed. Haman, D. Z., R. T. Pritchard, A. G. Sm a, F. S. Zazueta, and P. M. Lyrene. 1997. Evapotranspiration and crop coefficients for young blueberries in Florida. Applied Engineering in Agri culture 13(2): 209-216. inology for Phosphorus transport. Journal of Environmental Quality 29:5. of citrus through soil m oisture m sandy soil. S oil Science Society Bureau of Laboratories, Environm http://www.dep.sta Meteorological Organization. Tech. No Grattan, S. R., W Bowrs, A Dro proving ajstrl 10-1 1998. W. M. Alley. NY: Von Nostrand Reinhold. Haygarth, P.M., and A.N. Sharpley. 2000. Term

PAGE 137

124 Haygarth, P.M., and S.C. Jarvis. 1997. Soil derived phosphorus in surface runoff from grazed grassland lysimeters. Water Resources Research 31:140-248. Heckrath, G., P.C. Brookes, P.R. Poulton, and K.W.T. Goulding. 1995. Phosphorus leaching from soils containing different concentrations in the Broadbalk experiment. Journal of Enviro. 24:904-910. Hill, R.W., R.J. Hanks, and Wright, J.L. 1984. Crop Yield Models Adapted to Irrigation Scheduling Programs. Final Report USDA-ARS Cooperative Research No. 58-9AHZ-9-440. Utah Agricultural Experiment Station Research Report No. 99. Logan, UT: Utah State University. Hill, R. W., and R. G. Allen. 1996. Simple irrigation calendar: A foundation for water management. In: Irrigation scheduling: from theory to practice, proceedings ICID/FAO workshop, September 1995, Water Report No. 8, FAO. Rome, Italy: Food and Agriculture Organization of United Nations. Hillel, D. 1997. Chapter 3: Improving water use efficiency. In Small-Scale Irrigation for Arid Zones. Principles and options. FAODevelopment series-2. FAO, Land and Water Division. Rome, Italy: Food and Agriculture Organization of United Nations. Irmak, S., and D. Z. Haman. 2003. Evapotranspiration: potential or reference? Agricultural Engineering, Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences. Gainesville, FL: University of Florida. Jacobs, J. M., and S. R. Satti. 2001. Evaluation of reference evapotranspiration methodologies and AFSIRS crop water use simulation model. Division of water supply management. Palatka, FL.: St. John’s River Water Management District. James, L. G. 1988. Principles of Farm Irrigation System and Design. Ed. 1993. New Jansson, S. L., and J. Pearsson. 1982. Mineralization and immobilization of soil nitrogen. In. Nitrogen in Agricultural Soils, Ed. F. J. Stevenson. Agronomy Monograph No. 22. Wisconsin, WI: American Society of Agronomy. Jensen, M. E., and H. R. Haise. 1963. Estimating evapotranspiration from solar radiation. Journal of Irrigation and Drainage Division of American Society of Civil Engineering [ASCE] 89(IR4): 15-41. Jensen, M.E., R.D. Burman, and R.G. Allen. 1990. Evapotranspiration and irrigation water requirements Manual and Reports on Engineering Practice No. 70, New York, NY: ASCE. nmental Quality York, NY: John Wiley and Sons.

PAGE 138

125 Jones, J. W., L. H. Allen, S. F. Shih, J. S. Rogers, L.C. Hammond, A. G. Smajstrla, and J. D. Martsolf. 1984. Estimated and Measured Evapotranspiration for Florida Climate, Crops, and Soils. Institute of Food and Agricultural Sciences, Bulletin 840. Gainesville, FL: University of Florida. Kang, S., B. Gu, T. Du, and J. Zhang. 2003. Crop coefficient and ratio of transpiration to evapotranspiration of winter wheat and maize in a semi humid region. Agricultural Water Management 59 (1): 239-254. Kashyap, P. S., and R. K. Panda. 2001. Evaluation of evapotranspiration methods and development of crop coefficients for potato crop in a sub-humid region. Agricultural Water Management 50: 9-25. Klocke, N.L., R.W. Todd, G.W. Hergert, D.G. Watts, and A.M. Parkhurst. 1993. Design, installation and performance of percolation lysimeters for water quality sampling. Transactions of ASAE 36(2): 429-435. Lamm, F. R., T. P. Trooien, H. L. Manges, and H. D. Sunderman. 2001. Nitrogen fertilization for subsurface drip irrigated corn. Transactions of ASAE. 44(3): 533-542. Larsen, S. 1967. Soil phosphorus. Advances in Agronomy 19:151-210. depth effects on pesticide residues in drainage water. Transactions of ASAE 39(5): 1731-1738. Lie, T., J. Xiao. G. Li, Y. Yu, Z. Liu, and J. Wang. 2003. Estimating crop coefficients of drip irrigated watermelons and honeydew melons from pan evaporation. ASAE Paper No.032247. St. Joseph, Michigan: ASAE. Liu, C., Z. Xiying, and Y. Maozheng (1998). Determination of daily evaporation and evapotranspiration of winter wheat field by large-scale weighing lysimeter and micro lysimeter. Journal of Hydraulic Engineering 10. Marella, R. L. 1999. Water withdrawals, use, discharge and trends in Florida 1995. Water resources investigations report 99-4002. Tallahassee, FL: United States Geological Society [USGS]. Martin, E. C., T. L. London, J. T. Ritchie, and A. Werner. 1994. Use of drainage lysimeters to evaluate nitrogen and irrigation management strategies to minimize nitrate leaching in maize production. Transactions of ASAE 37(1): 79-83. Maynard, D.N., G.J. Hochmuth, C.S. Vavrina, W.M. Stall, T.A. Kucharek, S.E. Webb, T.G. Taylor, and S.A. Smith 2001. Cucurbit production in Florida. In Vegetable Production Guide for Florida, ed. D.N. Maynard and S. Olson, Ch. 27, 151-178. Gainesville, FL: Institute of Food and Agricultural Sciences, University of Florida. Liaghat, A., and S. O. Prasher. 1996. A lysimeter study of grass cover and water table

PAGE 139

126 Mc Farland, M. J., J. W. Worthington, and J. S. Newman. 1983. Design, installation and operation of a twin weighing lysimeter for fruit trees. Transactions of ASAE 26: 1717-1721. Meisinger, J. J., and G. W. Randall. 1991. Estimating nitrogen budgets for soil crop systems. In. Managing Nitrogen for Groundwater Quality and Farm Profitability. 85-124. Ed. R. F. Follett, D. R. Keeney, and R. M. Cruse. Madison, WI: Soil Meshknd tation for assessing evapotranspiration from a large undisturbed soil Michaeent . of Illinois. e, and on of evaporation estimation methods for a Oad., RObreza, T. A., D. J. Pitts, L. R. Parsons, T. A., Wheaton, and K. T. Morgan. 1997. Soil cts citrus irrigation scheduling strategy. Ongley and taly: Food and Agriculture Organization of United Owen, tal Quality 19: 131-135. Phene, C. J., J. L. Fouss, and D. C. Sander. 1979. Water nutrient herbicide management Pier, J. trickle irrigated watermelon. Soil Science Society of America Journal 59: 145-150. Science Society of America, Inc. at, M., R. C. Warner, and L. R. Walton. 1999. Lysimeter design, construction, ainstrumen monolith. Applied Engineering in Agriculture 15(4): 303-308. l, A. M. 2001. Irrigation: Theory and Practice. New Delhi, India: Vikas Publishing House. Mulvaney, R. L., F. Azam, and F. W. Simmons. 1993. Immobilization of differnitrogen fertilizers. Illinois Fertilizer Conference Proceeding. January, 1993Urbana-Champagne. IL: University Nichols, J., W. Eichinger, D. I. Cooper, J. H. Prueger, L. E. Hipps, C. M. U. NealA. S. Bawazir. 2004. Comparis riparian area. IIHR Technical report No. 436. IIHR-Hydroscience & Engineering, College of Engineering, Iowa City, IA: University of Iowa. ., K. Lusk, and T. Podmore. 1997. Consumptive use and return flows in urban lawn water use. Journal of Irrigation and Drainage Engineering 123(1): 62-69. water holding characteristic affe Proceedings of Florida State Horticulture Society. 110: 36-39. , E. D. 1996. Control of water pollution from agriculture. FAO IrrigationDrainage paper 55. Rome, I Nations. L. B. 1990. Nitrate-nitrogen concentrations in percolate from lysimeters planted to a legume-grass mixture. Journal of Environmen Penman, H. L. 1948. Natural evaporation from open water, bare soil and grass. Proceedings of the Royal Society. Series-A. 193:120-145. of potatoes with trickle irrigation. American Potato Journal. 56: 51-59. W., and T. A. Doerge. 1995. Nitrogen and water interactions in

PAGE 140

127 Pierzynski, G.M. 2000.Methods of phosphorus analysis for soils, sediments, residwaters. June 2000. Southern Cooperative Series Bulletin No. 396. SERA-IEG-17Manhattan, KS: Kansas State University. ual and . Pitts, D.J., and G.A. Clark. 1991. Comparison of drip irrigation to sub-irrigation for ulture 7(2): Pote, D.H., T.C. Daniel, D.J. Nichols, P.A. Moore, Jr., D.M. Miller, and D.R. Edwards. Prunty, L., and B. R. Montgomery. 1991. Lysimeter study of nitrogen fertilizer and n 81. Sharpley, A.N., S.C. Chapra, R. Wedephol, J.T. Sims, T.C. Daniel, and K.R. Reddy. 1994. Managing agricultural phosphorus for protection of surface waters: issues Sharpley, A.N., T.C. Daniel, and D.R. Edwards. 1993. Phosphorus movement in the Shih, Sshoe, and G. Kidder. 1977. Water management for sugarcane production in Florida Everglades. In proceedings of International Shukla, S. 2000.Impacts of best management practices on nitrogen discharge from a c Institute and State University. Simon, C. M., E. I. Ekwue, F. A.Gumbs, C. V. Narayan. 1998. Evapotranspiration and tomato production in southwest Florida. Applied Engineering in Agric117-184. 1999. Seasonal and soil-drying effects on runoff phosphorus relationships to soil phosphorus. Soil Science Society of America Journal 63:1006. irrigation rates on quality of recharge water and corn yield. Journal of Environmental Quality 20: 373-380. Schachtman, D. P., R. J. Reid, and S. M. Ayling. 1998. Phosphorus uptake by plants:from soil to cell. Plant Physiology 116: 447-453. Sepaskhah, A. R., and Andam, M. 2001. Crop coefficient of sesame in a semi arid regioof I.R. Iran. Agricultural Water Management 49(1): 51-63. Sharma, M. L. 1985. Estimating evapotranspiration. Advances in Irrigation 3: 213-2 and options. Journal of Environmental Quality 23:437-451. landscape. Journal Prod. Agriculture 6:492-500. . F., D. L. Myhre, J. W. Mi Society of Sugarcane Technologists, 16th Congress (Sao Paulo, Brazil) 2: 995-1010. virgina coastal plan watershed. PhD diss. Blacksburg, VA.: Department of Biological Systems Engineering, Virginia Polytechni crop coefficients of irrigated maize (Zea mays L.) in Trinidad. Tropical Agriculture 75(3):342-346.

PAGE 141

128 Simonne, E. H., M. D. Dukes, and D. Z. Haman. 2003. Principles and practices of irrigation management for vegetables. Chapter 8. Horticultural Science Department, Florida Cooperative Extens ion Service, Institute of Food and Agricultural Sciences, AE260. Gainesville, FL: University of Florida. Singandhupe, R. B., G.G.S.N. Rao, N.G. Patil, and P.S. Brahmanand. 2003. Fertigation studies and irrigation scheduling in drip irrigation system in tomato crop Singh, B., and G. S. Sekhon. 1976. Some measures of reducing leaching loss of nitrates beyond potential rooting zone. I. Proper coordination of nitrogen splitting with Smajstrla, A. G. 1985. A field lysimeter system for crop water use and water stress edings Smajstr, and F. S. Zazueta. Engineering, Florida Cooperative Extension Service, Institute of Food and Smajstrla, A.G., and D.Z. Haman. 1997. Irrigated acreage in Florida. A summary through da, g ral and Biological Engineering, Florida Smajstr ontrol and potato production. Applied Engineering Solinstolinst.com. Accessed on December 09, 2002. Salts in Ecosystems, 207-220. Ed. Stanleyorida proceedings 50:6-8. . Transactions of ASAE 39(2): 931-936. (Lycopersicon esculentum L.). European Journal of Agronomy 19: 327-340. water management. Plant Science 44: 193-200. studies in humid regions. Soil and Crop Science Society of Florida Proce44: 53-59. la, A.G., B.J. Bowman, D. Z. Haman, F. T. Izuno, D. J. Pitts1997. Basic irrigation scheduling in Florida. Agricultural and Biological Agricultural Sciences, Bulletin 249. Gainesville, FL: University of Florida. 1998. Agricultural Engineering, Florida Cooperative Extension Service, Institute of Food and Agricultural Sciences, CIR 1220. Gainesville, FL: University of Flori Smajstrla, A. G., F. S. Zazueta, G. A. Clark, and D. J. Pitts. 2000. Irrigation schedulinwith evaporation pans. Agricultu Cooperative Extension Service, Institute of Food and Agricultural Sciences, Bulletin 254. Gainesville, FL: University of Florida. la, A. G ., S. J. Locascio, D. P. Weingartner, and D. R. Hensel. 2000. Subsurfacedrip irrigation for water table c in Agriculture 16(3): 225-229. . 2002. Levelogger instructions. Solinst Canada Ltd. Ontario, Canada. Available at www.s Stanhill, G. 1973. Evaporation, transpiration and evapotranspiration: a case for ockham’s razor. In: Physical Aspects of Soil Water and A. Hadas. NY: Springer-Verlag. , C.D., and G.A. Clark. 1991. Water table management using micro irrigation tubing. Soil and Crop Sciences Society of Fl Steele, D. D., A. H. Sajid, and L. D. Prunty. 1996. New corn evapotranspiration crop curves for southeastern North Dakota

PAGE 142

129 Stenitzer, E. 1996. Irrigation scheduling with gypsum blocks in Austria. In: Irrigation scheduling: from theory to practice, proceedings ICID/FAO workshop , September 1995, Water Report No. 8. Rome, Italy: Food and Agricultural Organization. Syverts losses ate. Journal of American Society of Horticultural Sciences 121(1): 57-62. Terry, DThornthwaite, C. W., and Mather J. R. 1955. The water balance. Laboratory of Tolk, J): 447-454 United States Department of Agriculture-National Agricultural Statistics Services d States USDA-NASS. 2004. Florida Agricultural Facts 2004. Washington, D.C.: United States United USEPA. 1994. National Water Quality Inventory. 1992 Report to Congress. EPA-841-R-USEPA. 1997. Nonpoint source pollution: thter quality problem. USEPA. 2002. National WVentura, F., D. Spano, P. Duce, and R. L. Snyder. 1999. An evaluation of common evapotranspiration equations. Irrigation Science 18: 163-170. en, J. P., and M. L. Smith. 1996. Nitrogen uptake efficiency and leachingfrom lysimeter grown citrus tress fertilized at three nitrogen r . L., P. Z. Yu, and H, S. Spenser. 1996. Commercial fertilizers 1995. Assoc. Amer. Plant Con. Off., Lexington KY: 41. Climatology, Centerton, NJ: John Hopkins University. . A., T. A. Howell, and S. R. Evett. 1998. Evapotranspiration and yield of corn grown on three high plan soils. Agronomy Journal 90(4 Tyagi, N. K., D. K. Sharma, and S. K. Luthra. 2000. Evapotranspiration and crop coefficients of wheat and sorghum. Journal of Irrigation and Drainage Engineering 126 (4): 215-222. [USDA-NASS]. 2003. Agricultural statistics. 2003. Orlando, FL.: UniteDepartment of Agriculture-National Agricultural Statistics Services. Department of Agriculture-National Agricultural Statistics Services. States Environmental Protection Agency [USEPA]. 1985. Nitrate/ nitrite health advisory (draft). Office of Water. Washington, D.C.:U.S. Environmental Protection Agency. 94-001. Office of Water. Washington, D.C.: U.S. Environmental Protection Agency. e nation’s largest wa Pointer No. 1. EPA 841-F-96-004A. USEPA. Office of Water. Washington, D.C.: U.S. Environmental Protection Agency. ater Quality Inventory. 2000 Report to Congress. EPA-841-F02-003. Office of Water. Washington, D.C.: U.S. Environmental Protection Agency.

PAGE 143

130 Watts, D. G., and D. L. Martin. 1981. Effects of water and nitrogen management on nitrate leaching loss from sands. Transactions of ASAE 24: 911-916. D. G., G. W. Hergert, and J. T. Nichols. 1991 Watts, . Nitrogen leaching losses from irrigated orchard grass on sandy soils. Journal of Environmental Quality 20: 355-Xu, X.,ion of evapotranspiration in the desert area using lysimeters. Communications in Soil Science and Plant Analysis Yang, J., B. Li, and S. Liu. 2000. A large weighing lysimeter for evapotranspiration and soil water and groundwater exchange studies. Hydrological Processes 14: 1887-Yohannrip and furrow irrigation and plant spacing on yield of tomato at Dire Dawa, Ethiopia. Agricultural Water Zaimes, G. N., and R. C. Schultz. 2002. Phosphorus in agricultural watersheds. A literature review. Department of Forestry. Ames, IA: Iowa State University. Zang, Y., Q. Yu, C. Liu, J. Jiang and X. Zhang. 2004. Estimation of winter wheat Zhang, M. K., Z. L. He, D. V. Calvert, P. J. Stoffella, Y. C. Li, and E. M. Lamb. 2002. 362. R. Zhang, X. Xue, and M. Zhao. 1998. Determinat 29(1&2): 1-13. 1897. es, F., and T. Tadesse. 1997. Effect of d Management 35(1): 201-207. evapotranspiration under water stress with two semi-empirical approaches. Agronomy Journal 96: 159-168. Release potential of phosphorus in florida sandy soils in relation to phosphorus fractions and adsorption capacity. Journal of Environmental Science and Health A37(5): 793.

PAGE 144

BIOGRAPHICAL SKETCH Saurabh Srivastava was born in the town of Orai, India. He was awarded a Bachel University, India, in 1999. He enrolled in the Master of Engineering degree program in the Agricultural and Biological Engineering Department at the University of Florida in e water qproduction in Florida. He is Manag or of Technology degree with honors in agricultural engineering from Allahabad August 2001. His research at the University of Florida was focused on addressing thuantity and quality issues related to agricultural currently working as a water resource engineer with the South Florida Water ement District for ADA Engineering Inc. 131