1 IRRIGATION APPLICATION AND NITROGEN LEACHING FROM WARM SEASON TURFGRASS USING SMART IRRIGATION CONTROLLERS AND DEVELOPMENT OF AN INTERACTIVE IRRIGATION TOOL By NICOLE ASHLEY DOBBS A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF ENGINEERING UNIVERSITY OF FLORIDA 2012
2 2012 Nicole Ashley Dobbs
3 To God, I dedicate all of my work
4 ACKNOWLEDGMENTS I foremost thank G od, from whom all blessings come. I thank my parents for all of the love, support and encouragement that they have given me throughout my entire life for sending me to Catholic school, instilling in me moral values, nurturing my faith, and always being t here for me along every step of my journey I thank my brother, Collin, for his love and comic relief. I am thankful for my advisor, Dr. Kati Migliaccio, for her guidance and support for all aspects of my research. She is an excellent advisor. She has hel ped me to be better organized and allowed me to manage my own project, while still always being available for advice and helpful input. Her attention to detail and critiques have refined my research and writing skills. I am thankful the support of other fa culty members who willingly served on my committee: Dr. Michael Dukes, Dr. Kelly Morgan, and Dr. Yuncong Li. I am thankful for their support in helping me to improve my research methods as well as writing. I am sincerely grateful to the companies, including Hunter Industries Inc. and Carlos Victoria and ValleyCrest, who contributed supplies to implement the field study. This research was also supported by University of Florida Institute of Food and Agricultural Sciences Tropical Research and Education Cent er (UF IFAS TREC ) UF Agricultural and Biological Engineering ( ABE) Department, and Miami Dade Water and Sewer Department. I am thankful for the support of many individuals at UF IFAS TREC who helped to make the field study a success including (but not li mited to) Tina Dispenza, Michael Gutierrez, Isaya Kisekka, Teresa Pissarra, David Li, Mary McCready, Jesus Lomeli Jorge Vergel, Jr., Manny Soto Nicholas Hughes Guiqin Yu (her help with sample
5 analysis), Manuel Sacramento Vitola, Letty Almanza, and Felip e Minoletti. The field study would not have been possible without their help. I am also thankful to the support of my many friends at UF IFAS TREC who have encouraged, helped, and inspired me throughout my studies including Xiaodan Mo, Jiebin Guo, Megha Kalsi, Justine Tatto, Vivek Jha, Garima Kakkar, Shimei Zheng, Kaileigh Calhoun, Anne Costa, Maria Angelica Sanclemente, Alina Campbell, Tanh Nguyen, Bernardo Navarrete, Renato Galdiano, Octavio Menocal, and Daniel Irick.
6 TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................. 4 LIST OF TABLES ............................................................................................................ 8 LIST OF FIGURES ........................................................................................................ 10 LIST OF ABBREVIATIONS ........................................................................................... 13 ABSTRACT ................................................................................................................... 14 CHAPTER 1 INTRODUCTION .................................................................................................... 16 Rationale ................................................................................................................. 16 Research Goal ........................................................................................................ 19 Application .............................................................................................................. 20 The Fl orida Lawn and Soil Conditions .................................................................... 20 Soil Water Terminology ........................................................................................... 22 Soil Water Balance ................................................................................................. 23 Evapotranspiration ........................................................................................... 24 Infiltration and Runoff ....................................................................................... 25 Soil Properties ........................................................................................................ 28 Nutrients ................................................................................................................. 29 Nitrogen ............................................................................................................ 30 Nutrient Leaching ............................................................................................. 30 Previous Studies ..................................................................................................... 31 Irrigation Technologies ..................................................................................... 31 Nutrient Leaching from Turfgrass ..................................................................... 34 Irrigation Equipment and Design ............................................................................. 37 Traditional Irrigation Equipment ........................................................................ 37 Rain Sensors .................................................................................................... 37 Soil Water Sensors ........................................................................................... 38 Neutron moderation ................................................................................... 38 Dielectric methods ..................................................................................... 39 Tensiometric methods ................................................................................ 39 Evapotranspiration Controllers ......................................................................... 40 2 EVALUATION OF WATER APPLICATION BY fOUR IRRIGATION SYSTEMS AND ASSOCIATED NUTRIENT LEACHING .......................................................... 46 Background ............................................................................................................. 46 Materials and Methods ............................................................................................ 51 Lysimeter Design and Installation ..................................................................... 52
7 Irrigation Treatments ........................................................................................ 52 Uniformity Testing ............................................................................................ 54 Turfgrass Quality Evaluation ............................................................................ 55 Fertilizer Application ......................................................................................... 55 Nitrogen Analysis ............................................................................................. 56 Water Balance .................................................................................................. 57 Data Analysis ................................................................................................... 57 Results a nd Discussion ........................................................................................... 58 Results of Quantity of Irrigation Water Applied ................................................. 58 Results of Water Leached ................................................................................ 59 Results of Nutrients Leached ........................................................................... 60 NO3N ........................................................................................................ 60 NH4N ......................................................................................................... 61 Turfgrass Quality Evaluation Results ............................................................... 62 Tissue and Soil Results .................................................................................... 62 Discussion ........................................................................................................ 62 3 INTERACTIVE TOOL FOR SIMULATING Irrigation technologies IN A VIRTUAL TURFGRASS SYSTEM .......................................................................................... 75 Background ............................................................................................................. 75 Methods .................................................................................................................. 79 Model Development ......................................................................................... 79 Irrigation Technologies ..................................................................................... 82 Model output ..................................................................................................... 83 Assumptions ..................................................................................................... 83 Model Validation ............................................................................................... 84 Graphi cal User Interface (GUI) ......................................................................... 85 Results and Discussion ........................................................................................... 86 Model Development ......................................................................................... 86 Graphical User Interface................................................................................... 94 Defaults ...................................................................................................... 95 4 SUMMARY AND CONCLUSION .......................................................................... 111 Objective 1 ............................................................................................................ 111 Objective 2 ............................................................................................................ 112 APPENDIX A RESULTS OF MODELING VALIDATION ............................................................. 114 B MODEL MANUAL ................................................................................................. 120 LIST OF REFERENCES ............................................................................................. 128 BIOGRAPHICAL SKETCH .......................................................................................... 139
8 LIST OF TABLES Table page 1 1 Methods used for calculating ET. ....................................................................... 42 1 2 Studies on water saving irrigation technologies. ................................................. 44 1 3 Sprinkler types and approximate water application rates (Michael Dukes, University of Florida, personal communication, 9 May 2011). ............................ 45 2 1 Evapotranspiration (ET) controller input parameters used. ................................ 67 2 2 Crop coefficient (Kc) values used by the Rain Bird ESP SMT ET controller for warm season turfgrass ....................................................................................... 67 2 3 Irrigation system quality ratings for evaluating distribution uniformity (DU) ........ 67 2 4 Comparison of medians, means, and standard deviations of water depths applied during three timebased application rates among the four treatments: time based (T1), evapotranspiration controller (T2), rain sensor (T3), and soil water sensor (T4) ................................................................................................ 67 2 5 Compariso n of average cumulative quantities*, medians, means, and standard deviations among the four treatments: timebased (T1), evapotranspiration controller (T2), rain sensor (T3), and soil water sensor (T4) for water depths applied and leached during the study period ........................... 68 2 6 Comparison of water depths leached among the four treatments: timebased (T1), evapotranspiration controller (T2), rain sensor (T3), and soil water sensor (T4) for the thr ee timebased application rates .................................................... 69 2 7 Comparison of nitrate and ammonium percent leached for two fertilizer applications (noted as 1 and 2) among the four treatments: timebased (T1), evapotranspiration c ontroller (T2), rain sensor (T3), and soil water sensor (T4) .. 69 2 8 Comparison of nitrate and ammonium concentrations for two fertilizer applications (noted as 1 and 2) among the four treatments: time based (T1), evapotranspiration controller (T2), rain sensor (T3), and soil water sensor (T4) .. 6 9 2 9 Comparison of nitrogen loads among treatments for two fertilizer applications (noted as 1 and 2) among the four treatments: timebased (T1), evapotranspiration controller (T2), rain sensor (T3), and soil water sensor (T4). 70 2 10 Initial and final ratings of turfgrass quality ........................................................... 70 2 11 Total percent nitrogen and carbon in the turfgrass tissue at the beginning of the experiment a nd at the end of the experiment ............................................... 70
9 2 1 2 Total percent nitrogen and carbon in the soil at the beginning of the experiment and at the end of the experiment ..................................................... 71 2 13 Water savings of three technology based irrigation systems co mpared to time based irrigation ........................................................................................... 71 2 14 Comparison of leachate load results to the results of similar studies on turfgrass ............................................................................................................. 72 3 1 E xample input for ET Controller .......................................................................... 96 3 2 Monthly crop coefficient (Kc) values for determining ETa .................................... 96 3 3 Field capacities and wilting points for various soil types ..................................... 96 3 4 CN(II) values used by the model ........................................................................ 96 3 5 Default values for the GUI and model ................................................................. 97 3 6 Average absolute differences over irrigation events between measured and model predicted values of irrigation and percolation. .......................................... 97
10 LIST OF FIGURES Figure page 1 1 Toro expanding disk rain sensor with adjustable setting .................................. 45 2 1 Lysimeter for collecting leachate below root zone. ............................................. 73 2 2 Time based irrigation and percolation during wet season (2 October 2011 to 5 November 2011), showing that leachate was collected only after heavy rainfall. ................................................................................................................ 73 2 3 Time based irrigation and percolation during dry season after irrigation runtime was increased to apply 3.2 cm per event (22 January 2012 to 16 February 2012), showing that leachate was collected after each irrigation event and even greater amounts of leac hate following rainfall. .......................... 74 3 1 Average measured data versus model predicted values for the timebased irrigation during the wet season (2 October through 5 November 2011) for application of appr oximately 1.27 cm per irrigation event. .................................. 98 3 2 Average measured data versus model predicted values for the timebased irrigation during the dry season (6 November through 17 December 2011) for application of approximately 1.27 cm per irrigation event ................................... 98 3 3 Average measured data versus model predicted values for the timebased irrigation during the dry season (18 December 2011 through 21 January 2012) for application of approximately 1.91 cm per irrigation event (changed to 3.18 cm on 18 January). ................................................................................. 99 3 4 Average measured data versus model predicted values for the timebased irri gation during the dry season (21 January through 18 February 2012) for application of approximately 3.18 cm per irrigation event. .................................. 99 3 5 Average measured data versus model predicted values for th e rain sensor based irrigation during the wet season (2 October through 5 November 2012) for scheduled application of approximately 1.27 cm per irrigation event and rain sensor setting of 12 mm. ........................................................................... 100 3 6 Average measured data versus model predicted values for the rain sensor based irrigation during the dry season (13 November 2011 through 21 January 2012) for scheduled application of approximately 1.27 cm per irrigation event and rain sensor setting of 3 mm. .............................................. 101 3 7 Average measured data versus model predicted values for the rain sensor based irrigation during the dry season (22 January through 18 February 2012) for scheduled application o f approximately 1.91 cm per irrigation event and setting of 3 mm. ......................................................................................... 101
11 3 8 Average measured data versus model predicted values for the soil water sensor based irrigation during the wet season (2 O ctober through 5 November 2012) for scheduled application of approximately 1.27 cm per irrigation event and threshold setting of 0.70. ................................................... 102 3 9 Average measured data versus model predicted values for the soil water sensor based irrigation during the dry season (6 November through 17 December 2012) for scheduled application of approximately 1.27 cm per irrigation event and threshold setting of 0.70. ................................................... 102 3 10 Average measured data versus model predicted values for the soil water sensor based irrigation during the dry season (18 December 2011 through 21 January 2012) for scheduled application of approximately 1.27 cm per irrigation event and threshold setting of 0.70. ................................................... 103 3 11 Average measured data versus model predicted values for the soil water sensor based irrigation during the dry season (22 January through 18 February 2012) for scheduled application of approximately 1.27 cm per irrigation event and threshold setting of 0.70. ................................................... 103 3 12 Average measured data versus two sets of model predicted values for the ET controller b ased irrigation during the wet season ........................................ 104 3 13 Average measured data versus two sets of model predicted values for the ET controller based irrigation during the dry season ........................................ 105 3 14 Comparison of ETa of the ET controller used in the field study and ETa as predicted by the model using FAWN data and crop coefficient from Romero and Dukes (2011). ............................................................................................ 106 3 15 Screenshot of rooting depth entry in metric units. ............................................. 106 3 16 Screenshot of soil type selection and description. ............................................ 106 3 17 Screenshot of irrigated area entry .................................................................... 107 3 18 Screenshot of possible irrigation system selection and description of each technology. ....................................................................................................... 107 3 19 Screenshot of selection of timebased irrigation system with a rain sensor with default rain sensor setting ......................................................................... 107 3 20 Screenshot of selection of timebased irrigation system with a soil water sensor with default threshold setting. ............................................................... 107 3 21 Screenshot of automatically generated irrigation days, based on Miami Dade County zip code and street number. ................................................................. 108
12 3 22 Screenshot of irrigation day(s) selection for zip codes entered outside of Miami Dade County. ......................................................................................... 108 3 23 Screenshot of first entry method for applied i rrigation amount ......................... 108 3 24 Screenshot of second entry method for applied irrigation amount: selection of irrigation system type and runtime entry. .......................................................... 109 3 25 Screenshot of weekly email report. ................................................................... 110
13 LIST OF ABBREVIATION S EPA United States Environmental Protection Agency ET Evapotranspiration ETA Actual evapotranspiration ETo Reference evapotranspiration FAWN Florida Automated Weather Network FC Field capacity FDEP Florida Department of Environmental Protection I Irrigation IA Irrigation Association IFAS Institute of Food and Agricultural Sciences KC Crop coefficient MAD Management Allowable Depletion N Nitroge n Q Runoff PERC Percolation R Rainfall RS Rain s ensor SRN Slow release nitrogen SWS Soil w ater sensor TDR Time domain reflectometry TDT Time domain transmission TMDL Total maximum daily l oad TREC Tropical Research and Education Center UF University of Flor ida USGS United States Geological Survey
14 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 IRRIGATION APPLICATION AND NITROG EN LEACHING FROM WARM SEASON TURFGRASS USING SMART IRRIGATION CONTROLLERS AND DEVELOPMENT OF AN INTERACTIVE IRRIGATION TOOL By Nicole Ashley Dobbs August 2012 Chair: Kati Migliaccio Major: Agricultural and Biological Engineering The research goals were to evaluate the ability of irrigation technologies to reduce nitrogen leaching from turfgrass and to simulate irrigation, leaching, and potential water savings associated with these irrigation systems as compared to a time based system To achieve these g oals, there was : (1) a field study in south Florida comparing water quantity applied and water quantity and quality (NO3N and NH4N ) leached by four different irrigation systems: time based, timebased with a rain sensor (RS), time based with a soil wat er sensor (SWS) and an evapotranspiration ( ET ) controller and (2) development of a simple soil water balance (SWB) model to simulate a turfgrass system irrigated by one of four irrigation systems, which was developed into the Interactive Irrigation Tool T he field study indicated that the three technology based irrigation treatments applied significantly less water than the time based system. Nitrogen concentrations and loads were relatively low among all treatments ra nging from an average of 6.80E 03 kg NO3N ha1 for the ET controller up to 0.053 kg NO3N ha1 for rain sensor based irrigation; a nd 1.72E 03 kg NH4 ha1 for the ET controller up to 9.84E 03 kg NH4 ha1 for
15 the time based irrigation over nearly 5 months The model, validated using the field data, simulated the SWB including irrigation, rainfall, ET, runoff, and percolation for turfgrass irrigated by one of the four irrigation technologies. Average absolute differences between measured and predicted values ranged from 0.37 cm to 1.62 cm for percolation and 0.0 cm to 0.28 cm for irrigation depths
16 CHAPTER 1 INTRODUCTION Rationale Florida is ranked first in the Southeast and third in the nation for employment related to natural resources and leisure and hospitality, both industries closely tie d to the states water resources. It follows that water conservation is a central focus of various public agencies of Florida ( e g water management districts, Florida Department of Environmental Protection [ F DEP], United States Geological Survey [USGS]). Florida has 11,158 square kilometers ( 4,308 square miles ) of water; 1,927 km ( 1,197 miles ) of coastline; 3,664 km ( 2,276 miles ) of tidal shoreline; over 17,700 km ( 11,000 miles ) of rivers, streams, and waterways; about 7,700 lakes greater than 4 hectares ( 10 acres ) ; and 27 first magnitude springs (more than any other state) ( S tate of F lorida.com 2011 ). Florida is one of the top travel destinations in the world, which can be attributed primarily to its attractive water resources, making water quality necessary not only for ecological benefits, but also to ensure Floridas economy. With the many demands on the water resources of Florida from recreational, public water supply, agriculture, and the natural ecosystem and unpredictable weather patterns, it is nec essary to employ conservative management techniques in order to meet and sustain water supply needs. For several years, a major focus of FDEP and the United States Environmental Protection Agency (EPA) has been the reduction of nutrient pollution which is partially from nonpoint sources where fertilizer and water are applied in excess. There are regulations in place to protect designated uses of water bodies from both point and nonpoint sources based on total maximum daily loads (TMDLs); however, nutrient pollution continues to degrade water quality. From 2008 to 2010, identified nutrient impaired waterways in Florida increased from about 1,610 to 3,058 km of impaired
17 streams and rivers and 141,640 to 152,972 ha of impaired lakes. The FDEP and EPA are conti nuing to develop numeric nutrient criteria for Floridas waters. These criteria establish allowable limits for nutrient loads and/or concentrations (i.e., nitrogen [N] and phosphorous [P]) in waterways to improve water quality and protect public health, a quatic life and the long term recreational uses of Floridas waters. Following a lawsuit from the Florida Wildlife Federation in 2008, EPA set a rule in the fall of 2010 for water quality standards for inland waters in Florida, outside of south Florida. T he proposal of a second rule for estuaries, coastal waters and flowing waters in south Florida was set to be released in March 2012. In response to the rules set by the EPA, FDEP petitioned in April 2011 to have these rules repealed and to develop its own set of nutrient criteria. Given that states have the primary responsibility to maintain the quality of their waters, EPA may withdraw these rules if FDEP can develop their own protective and scientifically sound numeric standards. A final rule by FDEP was adopted in January 2012, to be followed by legislative ratification (EPA, 2012). The various uses of freshwater include public supply, domestic, irrigation, livestock, aquaculture, industry, mining, and thermoelectric power. In a report by Kenny et al. ( 2009), irrigation was defined as agricultural and recreational (e.g. golf courses, parks), but not residential. Public water supply, as defined by Kenny et al. (2009), may include a variety of uses such as outdoor domestic use, such as residential irrigati on. This study indicated that in Florida, public water supply was 37% of the total freshwater use. Public water supply accounts for the majority of groundwater withdrawals (52% ), followed by 34% for agricultural and recreational irrigation. Due to the high rate of groundwater withdrawal and the interaction of groundwater with surface pollutants,
18 decreasing groundwater levels and the contamination of groundwater and surface water is a concern across Florida. Irrigation is often required to maintain turf qual ity during the growing season. Automatic irrigation systems for turf and ornamental landscapes have been shown to apply more water than manual irrigation or hoseend sprinklers (Maye r et al., 1999) Surveys conducted by Whitcomb (2005) of 7,200 singlefamily homes across Florida (16 water utilities) revealed that 53% of irrigation systems are inground with automatic timers, 13 % are inground manual, 25% are hoseend, and 10% do not irrigate. Some states, including Florida, require by statute water saving devices such as rain sensors (RSs) to be added to automatic irrigation systems (s.373.62 Florida Statutes, 1991). Still, this is not strictly enforced as the study by Whitcomb (2005) found that only 25% of single family homes with automatic irrigation systems across Florida have a RS and only 1% has a soil water sensor (SWS) The demand for public water supply is predict ed to continue to increase ( Marella, 2009). In addition to the increasing number of visitors to Florida, there has been a 17.6% increase in population from 2000 to 2010, almost twice the national average (U.S. Census Bureau, 2011). This trend in population is expected to continue, but may be limited by the supply of freshwater. Freshwater use in Florida is expected to increase by 7.6 billion liters (2 billion gallons) per day by 2025 (FDEP, 2010). Additionally, u npredictable changes in weather patterns must be considered in developing a comprehensive water conservation s trategy. This strategy should include a means of conserving urban water use, given that the majority (approximately 80% ) of the U.S. population resides in urban areas (U.S. Census Bureau, 2010).
19 Conservative water practices are necessary to allocate water supplies considering all water users and to plan for future water demands One sector where potential water savings exist is with landscape irrigation users. This research project was conceived based on the statewide need to gain knowledge on the relatio nship between irrigation management and nutrient leaching and for better online tools to engage the public in water conservation efforts related to landscape irrigation. Nutrient leaching and runoff from turf and landscape may be considered more critical t han in agricultural areas due to the proximity to drinking water supply (Waller, 2007). There are several recommendations given by the University of Florida (UF) Institute of Food and Agricultural Sciences (IFAS) on proper irrigation and fertilization, but compliance by property owners regarding these recommendations or statutes (e.g. mandatory RS s with automatic irrigation systems) remains a challenge. Recent studies in Florida indicate that residential properties can greatly reduce water applied by using soilwater based or evapotranspiration (ET) based irrigation scheduling devices (Dukes and Haley, 2009; Davis and Dukes, 2010; McCready and Migliaccio, 2011). Other studies have shown that soil moisture sensing devices can reduce nutrient leaching compared to conventional automatic irrigation (Augustin and Snyder, 1984; Snyder et al., 1984; Pathan et al., 2007). Research Goal The goal of this project is to compare water application and nutrient leaching among different turf irrigation technologies in a field setting and to develop a simple model that predicts water losses based on the irrigation method: timebased, timebased with a RS SWS or ET controller. This model will be used to provide a more detailed alter n ative to the Urban Irrigation Scheduler on the Florida Automated
20 Weather Network (FAWN) website so that users will be able to evaluate their own or alternative irrigation schemes with respect to water stress and losses. Application This research will contribute to water conservation efforts by o ffering a comparison of the water volumes applied among the various irrigation technologies, applicable to various turf landscapes in a residential, commercial, or public setting. The study will also provide an estimate of nitrate and ammonium leaching und er the various irrigation treatments. In addition to optimizing irrigation scheduling, this information would improve the ability to quantify nutrient contributions from nonpoint sources such as residential land, aiding in the development of a more comprehensive nutrient management plan. The online tool can assist in the management of water applied to turfgrass. The model will use real time weather data and user inputs to estimate water stress and losses for a particular lawn. This model could be used by irrigation professionals, home owners, or students interested in modeling these technologies and comparing their benefits. The Florida Lawn and Soil Conditions There are over 2 million hectares (5 million acres) of home lawns in Florida, with over 400,000 hectares (1 million acres) professionally managed (Trenholm and Unruh, 2005). There is a variety of warm season turfgrass, but the three most common species are bahiagrass, bermudagrass, and St. Augustinegrass. There are three maintenance levels for lawns: basic, moderate, and high. Low maintenance turf may be fertilized only two to three times per year, with or without irrigation, and mowed as needed. Highmaintenance turf is fertilized up to five times annually according to IFAS
21 recommendations, with fr equent irrigation during dry weather, increased mowing, and increased pest and disease problems. Most homeowners prefer to install an automatic irrigation system for convenience. The downside to this automation is that systems tend to over water, resulting in decreased turf quality and potentially loss of applied nutrients (Trenholm and Unruh, 2005) Fertilization recommendations vary by region in Florida: north, central, and south; generally increasing application amount from north to south. Mowing frequency depends on turfgrass variety and desired level of maintenance, but can be minimized by reducing the quantity of water and fertilizer applied. The UF IFAS has published much information related to proper lawn management, freely available through sources such as the UF IFAS Extension Electronic Data Information Source (EDIS), Florida Friendly LandscapingTM, and Florida Automated Weather Network (FAWN). On average, most of Florida receives at least 152 cm (60 in) of rainfall annually, with the majority occurring between June and October (Tre nholm and Unruh, 2005) The typical field capacity for Florida soils is 1/12 cm/cm (1 cm of water to the top 12 cm of soil). The typical rooting depth for turfgrass is about 10 to 15 cm (4 to 6 in). Based on this information, the current UF IFAS recommend ation is to apply 1.3 to 1.9 cm (0.5 to 0.75 in) of water when the lawn show signs of wilt in order to wet the rooting zone and slightly below (Trenholm et al., 2009) UF/IFAS irrigation recommendations for turfgrass in Florida are provided in several publications, however, Dukes (2008) summarized these recommendations as follows: (1) I rriga ting "deep and infrequently" (1.3 to 1.9 cm) for wilting turf in a sandy soil where vert ical root growth is not limited; (2) Irrigation frequency and run times are recommended based on irrigation application rate, month of the year, and different climate areas within t he state.
22 N itrogen application recommendations do not require soil testing, but are based on crop needs as found in literature (Kidder et al., 1998) The recommended fertilizer application rates are provided in Tables 12 and 1 3 for north, central, and south Florida for three common varieties of turfgrass (Kidder et al., 1998). Water and fertilizer needs vary by location and season, but some general recommendations for the lawn are given by Trenholm et al. (2009) IFAS recommends that soluble N application not exceed 24.4 kg ha1 (0.5 lb per 1000 square feet). Slow release N (SRN) sources are preferred and application rates will depend on the N release rate. A range of N appl ication rates is given by Trenholm et al. (2002) and Trenholm and Unruh ( 2005 ) based on variations in the desired maintenance level, localized environmental conditions, and irrigation schemes. Soil Water Terminology Water is mainly absorbed by turfgrass t hrough the roots and lost through transpiration. The quantity of water absorbed is a function of root depth, root number, amount of available water in root zone, root extension rate, transpiration rate, and soil temperature. The amount of available water i n the root zone is controlled by precipitation rate, irrigation rate, water table depth, soil properties, and salt concentration of the soil. Water absorption rate may be restricted by excessive N fertilization, overwatering, and acidic, compacted soil (Beard, 1973) Available water is the total amount of water available for a plant to absorb based on the root zone (RZ) and the water holding capacity (WHC). WHC is the amount of water that is available for plant uptake, given as a ratio of the depth of water avai lable for uptake for a given depth of soil; this is the difference between field capacity (FC) and permanent wilting point (PWP), shown in E quation 13 FC is the upper limit of water
23 storage given as depth of water per depth of soil. PWP is the lower limi t, or the point at which the plant can no longer absorb water from the soil. WHC = FC PWP (1 1 ) AW = WHC x RZ (1 2 ) To maintain plant quality in an efficient manner, irrigation should occur when a certain percentage of the AW, or the m anagement allowa ble depletion (MAD) is reached, typically taken as 0.5 (Allen et al., 1998) RAW = MAD x AW (1 3 ) where RAW is the readily available water, or represents the lower boundary of an ideal operating range for irrigation based on soil water content. The soil water depletion depends on plant charact eristics, soil characteristics, and atmospheric conditions. By keeping track of the inputs and outputs of the system using a SWB irrigation can be scheduled so that losses are minimized. Soil Water Balance The SWB is shown in Eq uation 1 4 with all quantities in units of length. This balance consists of water applied, ET runoff, and drainage below the root zone. It provides a way to check irrigation scheduling efficiency so that the optimal amount is being applied to maintain w ater for plant growth. Q D ET I R SW ( 1 4 ) W is change in soil water storage, R is rainfall, I is irrigation, D is drainage, and Q is runoff.
24 Evapotranspiration ET is the combination of water loss by evaporation from the soil and plant and by t ranspiration from the plant. Energy from the sun may be used for ET, reflected, or transferred by conduction. By transpiration, water flows from the roots to plant leaves to transport nutrients and maintain a healthy temperature range for growth. Transpiration increases with an increased vapor pressure gradient as the internal leaf vapor pressure increases with increasing leaf temperature. ET can account for 80 to 85% of soil water loss (Harrold and Dreibelbis, 1951) ET varies seasonally, diurnally, and hourly. Typically, maximum rates occur when solar radiation is greatest while there is a significant decrease during months with lower rates of solar radiation (Harrold and Dreibelbi s, 1951; USDA, 1955; Richards and Weeks, 1963; Pruitt, 1964; Horn, 1965; Ekern, 1966) ET rate is affected by temperature, light duration, atmospheric vapor pressure, wind, water absorption rate, and soil moisture tension. Atmospheric water vapor content and wind speed tend to have a significant influence on ET rate (Beard, 1973) Transpiration increases with the influence of wind which lowers external vapor pressure. Transpiration may stop temporarily if stomata close during periods of extreme high temperature. ET is also reduced in the shade due to decreased reception of solar energy. Transpiration rate is affected by water absorption rate which is a function of soil water availability, soil temperature, and the extent and quality of the roots. Soil water leve l varies with incoming water, soil properties, and plant uptake. Transpiration is reduced when there is a soil water deficit which causes a decrease in the internal plant vapor pressure, reducing the vapor pressure gradient with the external atmosphere (Eagleman and Decker, 1965; USDA, 1955) Transpiration rate varies among turfgrass
25 species due to differences in r ooting depth, root density, root shoot ratio, total leaf surface area, cuticle thickness, osmotic pressure of cells in the leaf, leaf morphology, leaf orientation, internal leaf structure, differences in stomata, and leaf folding capability (Beard 1973) Pesticides and fertilizers may also influence transpiration rate (Wilson and Runnels 1931; Runnels and Wilson 1934; Slayter and Bierhuizen, 1964) ET can be calculated by a variety of methods ( Table 11 ) ET for a specific plant (or ETc) is calculated as (Dukes et al., 2009) : o ET c K c ET (1 5 ) where ETc is crop ET and Kc is the crop coefficient. Crop coefficients depend on specific crop, management practices, and location. ETc may also be referred to as ETa or actual ET. Infiltration and Runoff There are four common me thods of calculating runoff: Hortons equation, Philips equation, GreenAmpt method, and Soil Conservation Service ( SCS ) method. The first three methods involve calculating infiltration. Runoff is then calculated as the difference between rain intensity and infiltration capacity ( i f ). Hortons and Philips equations were developed from approximate solutions of Richards equation, the governing equation for unsteady unsaturated flow in a porous medium (Eq uation 1 6 ). (1 6 ) K is hydraulic conductivity. K z D z t
26 Hortons (1933, 1939) equation can be derived from Richards equation by assuming K and D are constants independent of soil moisture content, reducing E quation 16 to 2 2dz D t (1 7 ) From Equation 1 7 soil water content can be found as a function of time and depth (Eagleson, 1970; Raudkivi, 1979). Philips (1957, 1969) equation assumes that K and D vary with moisture content and sol ves fo r cumulative infiltration, i.e. F(t), as Kt St t F 5 0 ) ( (1 8 ) where S is sorptivity. The Green Ampt method has an exact analytical solution derived from a physical theory, using the principles of continuity and momentum (Green and Ampt, 1911) It Kt t F t F ) ( 1 ln ) ( (1 9 ) The SCS method (1972) was developed as a simple means to calculate runoff without calculating infiltration. The depth of runoff Pe is assumed to be less than precipitation depth P and the water that is retained Fa is less than or equal to a maximum potential S. There is an initial amount of rainfall for which no runoff occurs, Ia
27 so that potential runoff is P Ia. The main assumption is that the ratios of the two actual and two potential quantities are equal as shown below a e aI P P S F (1 1 0 ) Fa is depth of water retained in watershed, S is potential maximum retention, Pe is runof f depth, P is precipitation depth, and Ia is initial abstraction before ponding. All depths are in inches. Based on continuity, a F a I e P P (1 1 1 ) enabling the solution of Pe in E quation 11 0 After much study, an empirical relationship was dev eloped: S Ia2 0 (1 1 2 ) The final equation to calculate runoff is: S P S P Pe8 0 ) 2 0 (2 (1 1 3 ) Curve numbers (CNs) were derived from plotting P versus Pe for many watersheds. S CN 10 1000 (1 1 4 ) where S is in inches. CNs are known for four hydrologic soil groups and various land uses, whic h can be used to solve for S in E quation 114 and, subsequently, Pe in E quation 11 3
28 The Chemicals, Runoff, and Erosion from Agricultural Management Systems (CREAMS) model combined modified the CN method by relating S to soil water content (Knisel, 1980) : UL SM UL mx S S (1 1 5 ) where UL is the upper limit of soil water storage of FC and SM is the soil water content in the root zone. The maximum value of S, Smx (cm), is calculated using the antecedent moisture condition (AMC) I (dry): 4 25 2540 I CN mx S ( 1 1 6 ) where CNI is the CN for AMC I (dry). CNI ca n be calculated by Equation 11 7 as: 3 ) ( 0001177 0 2 ) ( 01379 0 ) ( 348 1 91 16 II CN II CN II CN I CN ( 1 1 7 ) where CNII is the CN for AMC II (normal) (Knisel, 1980) Soil Properties Soil properties affect certain components of the SWB na mely the drainage, storage, infiltration, and runoff components, and therefore affect irrigation rates. Physical properties that influence infiltration, retention, and water movement include soil texture, depth, structure, and porosity. Texture refers to s oil particle size, while structure is the arrangement of soil particles into larger aggregates. The four major components of soil which will determine these physical properties are mineral s, organic matter, water, and air. V olumetric soil water content var ies with soil texture (Zotarelli et al., 2010) The preferred textures for turfgrass are loamy sands, course sandy loams, and loams. Modifying the structure such as adding course aggregate such as sand can
29 improve soil conditions for turfgrass by increasing aeration and water movement in places of heavy traffic. Percolation, the downward movement of water through the soil, is affected by the size, number, and continuity of pores; hydration of the pores; and resistance of entrapped air. Percolation rates aff ect infiltration. Percolation is relatively fast in course sands and slow in clays. Desirable characteristics for high water retention include fine soil texture, good structure, and high organic matter (Beard, 1973) Nutrients Turfgrass requires N a nd P in greater amounts than other essential macronutrients necessary for optimum growth. N itrogen is the most heavily required nutrient by turfgrass and so N is often applied at a higher rate and more frequently than other nutrients. The quantity applied depends on a variety of factors including: turfgrass variety, quality desired, soil type, and the amount of water applied which varies by region. For example, bahiagrass requires small amounts of N for optimal growth, compared to bermudagrass and St. Augustinegrass w hich cannot survive most sandy Florida soils without added N (Trenholm and Unruh, 2005). There are two main forms of N fertilizer: soluble and slow release (SRN). There are advantages and disadvantages for each form, but typically a mixture of the two is preferred, with a higher SRN content. Soluble N may burn turfgrass if applied in excess of the recommended 24.4 kg ha1 (0.5 lb/1000 ft2). Excess water application and heavy rainfall events may leach soluble N below the root zone. SRN offer the benefits of minimizing losses by leaching and an extended response, but turfgrass is considered to be an efficient N absorbing ground cover if the turfgrass is properly maintained and has been fertilized at the UF IFAS recommended rate and frequency (Trenholm and Unruh, 2005).
30 Nitrogen N itrogen is the most common limiting nutrient for turfgrass. N itrogen provides dark green leaf color, shoot density, and tolerance to nutrient and pest stresses. Excessive N application may result in increased growth and water demand, i ncreased disease susceptibility, and reduced high temperature tolerance (Havlin et al., 2005) N itrogen is transformed between inorganic and organic forms within the soilplant atmosphere system, converting N2 to plant available N. The N cycle consists of inputs, outputs, and cycling within the soil. The inputs include fertilizers; outputs include plant uptake, denitrification, volatilization, leaching, and NH4 + fixati on; cycling processes within the soil include immobilization, mineralization, and nitrification. The N cycle consists of the following steps: (1) N2 from the atmosphere and the inorganic N sources on the ground surface are added to the soil, (2) organic N is mineralized into NH4 +, (3) NH4 + is converted to NO3 -, (4) NO3 and NH4 + are taken up by plants, (5) NO3 lost through leaching, (6) some NO3 converted to N2 and N oxides, and (7) some NH4 + converted to NH3 (Havlin et al., 2005) .Turfgrass will take up N as NO3 and NH4 +, but NO3 is the preferred form (Blackmer, 2000). Nutrient Leaching Nutrient leaching is a concern in Florida because of the sandy soils with low f ield capacity and low nutrient retention, and relatively high water table. In order to maintain turf quality with Florida soils, a fertilization regimen is required. Turfgrass can uptake N efficiently, but excess N may be lost through runoff or leaching. A ccording to a study by Easton and Petrovic (2004) nutrient leaching is a function of soil infiltration rate, fertilizer source, time, shoot density, and antecedent soil moisture. Leaching is assumed to occur whenever the soil moisture of the root zone exceeds field capacity (Smith and
31 Will iams, 1980). Soil pH also has an effect on leaching Plant nutrients leach more rapidly at soil pH values less than 5.0 than in the range between 5.0 and 7.5 (T renholm and Unruh, 2005) N poses a threat to water sources as a risk to human health and the ecosystem. Once applied to the soil, N can be transported following heavy rain or irrigation either by runoff to surface water or leaching. N in runoff may enter surface water bodies and contribute to increased rates of eutrophication. Once leached, N may enter the groundwater, possibly being conveyed to public drinking water supply or discharged by springs. The public drinking water standard for NO3N is 10 mg L1 (EPA, 2011). Previous Studies Irrigation T echnologies Several studies have been conducted in Florida on water savings associated with RSs SWSs and ET controllers compared to conventional timebased irrigation scheduling (Davis et al., 2007; Haley and D ukes, 2007; Shedd et al., 2007; Cardenas Laihalcar et al., 2008; Dukes et al., 2008; Davis and Dukes, 2008, 2010; Davis et al., 2009; McCready et al., 2009; Cardenas Laihalcar et al., 2010; Cardenas Laihalcar and Dukes 2010; McCready and Dukes, 2011; Hale y and Dukes, 2012). McCready et al. (2009) tes ted the ability of SWS controllers, ET controllers, an d RSs to conserve water while maintaining quality of St. Augustinegrass. There was a 730% reduction in water use for RS treatments ; 0 74% for SW S treatment s; and 2562 % reduction for ET treatments. A 40% reduced irrigation schedule coupled with a RS resulted in water savings of 3653% A study comprised of 58 homes with preexisting automatic time based irrigation systems in Pal m Harbor, Florida showed that SW S system controllers were able to
32 reduce irrigation by 65 % corresponding to a maximum reduction of 992 L d1, compared to conventional timebased irrigation ( Haley and Dukes, 2012). The four treatments in this study were (1) automatic timer with a bypas s SW S control system, (2) automatic timer with RS and educational materials, (3) automatic timer with RS (4) automatic timer only or time based (typical for the region); each treatment was replicated at least three times. Homeowners ultimately had control over the operation of their irrigation system. Data collected over 26 months included irrigation application, frequency, quarterly turf quality ratings, and weather data. Despite dry weather conditions, the SWS controllers still reduced the number of mont hly irrigation events, averaging two irrig ation events per month versus 4 to 5 events for timebased irrigation systems. Treatment 2, which provided homeowners with educational materials on recommended time clock settings, showed initial water savings simi lar to the SWS, but was ultimately not significantly different from the timebased. The RSs did not provide significant water savings, possibly due to the dry weather conditions. Cardenas Lailhacar et al. (2008) results from experimental study plots (not residenti al land) under wet weather conditions revealed that timebased irrigation systems with a RS at a 6 mm threshold applied 34% less water than those without one. Four brands of SWS controllers at three frequency settings were tested. It was concluded that SWS based irrigation offers a significantly more efficient means of irrigation compared to timebased, with water savings of up to 92 % while maintaining turfgrass quality. Of the three frequencies of SWS based treatments tested (1, 2, and 7 days per week), the 7 day/week frequency applied the least amount of water, indicating that highfrequency irrigation offers a means of conserving water during a rainy season for sandy soil. Due to the favorable weather and environmental conditions, even the
33 nonirrigated c ontrol plots maintained acceptable quality standards. Regarding the inconsistent results among the various brands tested, it was noted that performance of individual sensors depends greatly on threshold settings and burial depth. Pathan et al. (2007) in Western Australia evaluated the capability of a SWS to red uce water and nutrient leaching in turfgrass, compared to conventional scheduled irrigation. Results showed that volumetric water applied was reduced by 25% for the turfgrass irrigated with a SWS Leaching was also lower with the SWS with only 4% of appli ed water draining from SWS plots and 16% from conventionally irrigated plots. Overall, nutrient leaching was minimal, but was 5.5 times lower in the SWS plots. Overall appearance remained acceptable under both conditions, but according to certain metrics ( e.g. clippings, color, and osmotic potential of leaf sap), the SWS plots were considered to have had mild drought stress. A study in Gainesville, Florida on RS accuracy by Meeks et al. (2012) tested the accuracy of seven model and set point combinations to interrupt irrigation events based on threshold settings. Results showed no single trend for all RSs but showed much variability in terms of accuracy over time. A higher threshold setting did correspond with lower accuracy for the same brand. There was also improved accuracy over time. One brand, Toro TWRS, showed 95% accuracy at the beginning and 98 % three years later. Some brands showed decreased accuracy over time or after changing threshold settings. Reducing the threshold of one brand from 25 to 6 mm reduced accuracy from 27 to 98 % The average weighted accuracy of another brand, set at 3 and 13 mm, over the same time period decreased from 84 to 59 % Due to low costs and maintenance
34 requirements, RSs are still considered a useful water saving device, but should not be used wher e high accuracy and precision are required. Nutrient L eaching from T urfgrass Bowman et al. (2002) compared N leaching and uptake among six warm season turfgrasses: St. Augustinegrass, centipedegrass, Meyer zoysiagrass, Emerald zoysiagrass, Tifway bermudagrass, and common bermudagrass. Grasses were established in sand filled columns in a greenhouse and ammonium nitrate (340 0) was applied at 50 kg N ha1 on seven dates. In order to promote leaching of ~50% and measurable differences among the turfgrass species, columns were irrigated three times per week (410 cm of water per week). Leachate samples were collected following each irrigation event and analyzed for NO3N and NH4N. Leachi ng losses ranged from 48100 % for NO3N and 4 to 16 % for NH4N. Results showed that St. Augustine and bermudagrass (common and hybrid) were most efficient at reducing leaching. Initially, NO3N and NH4N leached readily, until the root system developed. With the exception of Meyer zoysiagrass following one of the applications, all volumeweighted NO3N concentrations were below the 10 mg N L1 threshold. The N uptake was measured in clippings, roots, and shoots. Root length density at various soil depths was also measured at the conclusion of the experiment. At soil depths >30 cm, St. Augustinegrass and bermudagrasses, the species w hich produced the least leaching losses, had significantly higher root length density than the other grasses, indicating that turfgrass species and rooting depth significantly affect NO3N leaching and N uptake. Augustin and Snyder (1984) tested the abili ty of tensiometer soil tension sensing devices to save water and reduce leaching of soluble N and SRN on bermudagrass, grown in Pompano fine sand in south Florida over a four year period. Maintenance
35 levels were similar to those of golf course fairways in Florida. There were two irrigation schemes, with six replicates each: (1) daily irrigation equivalent to the historical maximum daily ET amount and (2) irrigation when a sufficient deficit was detected by the tensiometer. Leaching fractions (LF) were calculated as R I ET R I LF (1 18) where I is irrigation and R is rainfall. During the wet season, water savings of 4295% were obtained from the tensiometer irrigation, compared to daily irrigation treatment; leachi ng fractions averaged 0.55 and 0.02 for daily and sensor treatments, respectively. During the dry season, water savings of 4795 % were obtained from the tensiometer irrigation compared to daily irrigation. T he leaching fraction was also greater for the daily irrigation treatment. T ensiometer i rrigation applied only 26% of the total water applied by the daily irrigation treatment and also improved turf appearance. Results showed that tensiometer controllers reduced N leaching from all N sources. The greatest N leaching (2256% of that applied) occurred with the treatment of NH4NO3 at 5 g m2 month1 and daily irrigation. SWS controller and fertigation (applying the NH4NO3 with irrigation simultaneously) produced the least amount of leaching (<1 to 6% of that applied). SRN fertilization treatment, sulfur coated urea, combined with tensiometer based irrigation resulted in the second least amount of leaching (0.311.2% of that applied). Soluble N fertilizer application in conjunction with the tensiometer irrigation treatment produced statistically eq uivalent turf appearance to the SRN fertilizer and daily irrigation treatment, indicating that fertilizer costs can be reduced with the SWS irrigation treatment.
36 Erickson et al. (2001) compared N losses by runoff and leaching of a St. Augustinegrass and a mixed species landscape. A RS was connected to the irrigation timer. They found that N losses in runoff were minimal for both and that losses by leaching were surprisingly greater for the mixedspecies landscape at 48.3 kg ha1 and 4.1 kg ha1 for the St. Augustinegrass. R esults indicated that properly established St. Augustinegrass is not a significant contributor to N pollution. Petrovic (1990) prov ided a review of several studies at a variety of locations on the fate of N applied to turfgrass (i.e. plant uptake, atmospheric loss, soil storage, leaching, and runoff). He reported that fertilizer uptake by turfgrass was a function of N release rate, N application rate, and grass species. Uptake ranged from 5 to 74% of applied N. Leaching was affected by fertilizer management practices (rate and timing), N source, soil texture, rainfall, and irrigation. Nitrate leaching was highly variable among studies, mostly less than 10% of that applied, but as high as 80 % The percent that leached was particularly high for the studies where fertil izer was applied at rates greater than the normal rate. Atmospheric loss by volatilization and denitrification ranged from 0 to 93 % of applied N. Volatilization can be reduced by irrigating following N application. Significant denitrification was found to occur only on finetextured, saturated, warm soils. Although few studies had been conducted on N concentration in runoff at the time of the Petrovic (1990) review, concentrations of N in runoff were rarely found to be above the 10 mg L1 drinking water standard. The source of N and timing of application affects the percent leached as shown in a study by Easton and Petrovic ( 2004) on turfgrass in sandy soils. During a dry and normal year of precipitation, there was little difference between the soluble and SR N
37 source, with the soluble leaching being 0.9% to 5% of that applied and the SRN being 0.5 to 7.4% of that applied. Duri ng a wetter than normal year, there was a significant difference in the percentage leached between the two N sources, wi th the SRN leaching 12 to 29 % of that applied and the slow release only 2 to 7 % Fertilizer application timing for this particular regio n (northeast U.S.) was found to result in the least amount of leaching in late fall as compared to the end of the growing season. Nutrient leachate reduction by turfgrass was related to overall plant growth and shoot density. More soluble synthetic organic fertilizers resulted in greater N loss (Ea ston and Petrovic, 2004) Irrigation Equipment and Design A traditional automatic irrigation system consists of sprinkler heads, valves, controllers, and the appropriate pipes and fittings. Three types of water saving irrigation technologies are discussed: RSs SWSs and ET controllers. Traditional Irrigation E quipment There are several types of sprinkler heads with various flow patterns, volumetric rates, and spray distance. The two basic flow patterns are stream and spray. A spray is a continuous flow of water and a stream is a rotating jet of water which can reach a larger area. Valves, typically operated electrically, are used to release the water coming from a main line to the sprinkler head at the appropriate time. Controllers are electrical timer s that send a signal to the valve to open or close to allow an irrigation event to occur for a specified length of time. Rain Sensors Rain sensors ( Figure 1 1 ) are the simplest device to add to a conventional timebased irrigation system of the three water saving technologies. The most common type
38 of RS and the one used in this study, is the sensor with an expansion disk. When rainfall enters the device, an expanding cork disk triggers a pressure valve which will then stop scheduled irrigation if a prese t amount of rainfall occurs. The most recent version of Florida statute 373.62 requires that all automatic irrigation systems have a RS or some other water saving irrigation contr oller (e.g. SWS or ET controller). Cardenas Lailhacar and Dukes (2008) recommend setting the lowest threshold possible (3 mm or 1/8 in on most devic es), or no higher than 6 mm (0.25 in), in order to achieve greater water savings. Soil Water Sensors There are direct and indirect methods to measure soil wat er content. Direct methods, including thermogravimetric and thermovolumetric, require soil sampling and laboratory analysis. Direct methods do not provide continuous monitoring of soil water content at the same location. Indirect methods, classified as either volumetric or tensiometric, estimate soil water content through a calibrated relationship with some other known variable. A discussion of different indirect methods follows. Neutron moderation Neutrons emitted from a radioactive source (within the probe) collide with particles of the same mass such as hydrogen ions. Since the source of hydrogen ions in most soils is water, the density of neutrons around the probe is proportional to the soil water volume fraction. Soil water is obtained from the devic e based on a linear calibration between the number of neutrons and the soil water content measured from nearby soil samples ( Muoz Carpena, 2004).
39 Dielectric methods Soil water content is estimated by measuring soil bulk permittivity, or dielectric constan t, Kab, shown below. 22 L t c Kat o b (1 19) where co is the speed of light in a vacuum, tt is the travel time of electromagnetic pulse, and L is length of the probe (Evett, 2007). Soil permittivity consists of the relative contribution of the various soil components (i. e. soil, water, air) and is dominated by water which has the highest dielectric constant (Kaw = 81). For most soils with a water content less than 50% the dielectric constant and volumetric water content (VWC) are often related by the equation established by Topp et al. (1980) shown below 3 6 2 4 2 210 3 4 10 5 5 29 2 10 3 5b b bKa Ka Ka VWC (1 20) Within the dielectric methods, there is time domain reflectometry (TDR), frequency domain (FD), amplitude domain reflectometry (ADR), phase transmission, and time domain transmission (TDT). Tensiometric methods Soil water matric potential is estimated by means of a porous medium in contact with the soil. Soil water moves into the medium in relatively wet soil or out of the medium in dry soil creating a tension within the tube that relates to matric potential of the soil.
40 Evapotranspiration Controllers There are two general categories of ET controllers: (1) SWB based and (2) nonSWB based ( Dukes, 2012) The SWB based controllers calculate a SWB continuously; whereas, the nonSWB controllers adjust run times relative to historical peak ET based on real time onsite measured weather conditions (i.e. temperature and/or solar radiation) (Dukes, 2012) The SWB based controller maintains a record of the SWB by recording water applied and estimating ET losses. ET is estimated by a standard equation which uses a combination of the following environmental parameters: temperature, solar radiation, relative humidity, and wind speed. Depending on the available parameters, there are a number of equations to calculate ET. T hree types of ET controllers described by Dukes et al. (2009) : s ignal based, standalone, and addon The signal based and standalone ET controllers measure ET at designated intervals and calculate optimal irrigation runtimes based on weather conditions and preset factors such as plant type, soil type, and percent shade. For signal based, ET is calculated from data from remote weather stations Either the weather data or the calculated ET is sent wirelessly to the controller (Dukes, 2012). A di sadvantage of signal based, particularly with sporadic weather patterns in pla ces such as Florida, is that the weather at the weather station from which it receives data may not be the same weather conditions at the actual site. A RS is recommended to be used with signal based ET controllers. The standalone (or onsite) ET controlle rs adjust irrigation according to weather conditions measured continuously onsite either by a SWB or replacement of ET (Dukes, 2012) The add on devices cannot calculate irrigation runtimes, but will only bypass a scheduled irrigation event based on the depth of irrigation input, rainfall, and ET, similar to a RS Some controllers have an alternative
41 source of ET from programmed historical data or will default to the last broadcasted ET value. Historicalbased is another type of controller which uses a pr e programmed ET curve for a particular region, modified by an onsite weather sensor (i.e. temperature and/or solar radiation) (Dukes, 2012).
42 Table 11 M ethods used for calculating ET. Equation Name Reference Equation American Society of Civil Engi neers (ASCE EWRI) standardized ETo equation Allen et al. ( 2005 ) ETO = reference ET (mm d1) Rn = net radiation (MJ m2 d1) G = heat flux (MJ m2 d1) U2 = wind speed (m s1) T = temperature (C) 1) 1) es = saturation vapor pressure (kPa) ea = actual vapor pressure (kPa) Cn = constant (900) Cd = constant (0.34) UF IFAS (1984) Penman equation* Jones et al. ( 1984) EToIFAS = potential ET from vegetated surface (mm d1) pressure temperature curve for air (mb C1) 1) Rs = total incoming solar radiation (cal cm2 d1) Boltzmann constant (11.71 x 108 cal cm2 d1 K1) T = mean daily air temperature (K) ea = vapor pressure of air (mb) R so = total daily cloudless sky radiation (cal cm 2 d 1 ) ) 2 1 ( 2 ) ( 273 ) ( 408 0 u d C u a e s e T n C G n R O ET 42 0 42 1 ) 08 0 5 0 ( 4 ) 1 ( so R s R d e T s R oIFAS ET ) 2 0062 0 5 0 )( ( 263 0 u d e a e
43 Table 11. Continued. Equation Name Reference Equation n (cal cm 1 mm 1 ) u2 = wind speed at 2 m height (km d1) e d = vapor pressure at dew point temperature (mb) SFWMD simple method Abtew and Obeysekera ( 1995 ) ETp = potential ET from vegetated surface (mm d1) K1 = regression coefficient (0.53) RS = total incoming solar radiation (MJ m2 d1) 1 ) Priestley Taylor Priestley and Taylor (1972); Sumner and Jacobs (2004) temperature curve (kPa C1) Rn = net radiation (MJ m2 d1) G = soil heat flux (MJ m 2 d 1 ) Turc ( 1961 ) Rs = daily solar radiation (cal cm2 d1) T = daily mean air temperature (C) Hargreaves and Samani ( 1985 ) Ra = extraterrestrial radiation (MJ m2 d1) Tmin = minimum daily air tempera ture (C) Tmax = maximum daily air temperature (C) T = daily mean air temperature (C) *Used by FAWN S R K P ET 1 ) ( G n R p ET ) 50 ( 15 013 0 s R T T p ET 17.8) + ( 0.5 ) min max ( a 0.0023 T T T R p ET
44 Table 12 Studies on water saving irrigation technologies. Irrigation Technology Weather Conditions Water Savings Reference Rain sensor Normal up t o 34% Cardenas Lailhacar et al. (2008) Normal 7 30% McCready et al. (2009) Dry 13 24% Cardenas Lailhacar et al. (2010) 19% Haley and Dukes (2007) Dry 14% Haley and Dukes (2012 ) Soil moisture sensor Normal 69 92% Cardenas Lailhacar et al. (2008) Dry 16 83% Cardenas Lailhacar et al. (2010) Dry 11 53% McCready et al. (2009) 25% Pathan et al. (2007 ) 73% Qualls et al. ( 2001 ) 74% (42 95%) Augustin and Snyder, (1984 ) Dry 65% Haley and Dukes (2012) 42 95% Snyder et al. ( 1984) 40% Horst and Peterson (1990) ET controller Dry to normal 25 62% McCready et al. (2009 ) 20 60% Davis et al. (2007)
45 Table 13 Spr inkler types and approximate water application rates (Michael Dukes, University of Florida, personal communication, 9 May 2011). Sprinkler Type Application Rate (in/hr) (cm/hr) Fixed spray head 1.5 3.9 Gear driven rotary 0.5 1.3 Impact 0.5 1.3 Fi gure 11 Toro expanding disk rain sensor with adjustable setting.
46 CHAPTER 2 EVALUATION OF WATER APPLICATION BY FOUR IRRIGATION SYSTEMS AND ASSOCIATED NUTRIENT LEACHING Background A major environmental focus at the national and state levels has been th e reduction of nutrient pollution in waterways. Point source pollution can be successfully managed by establishing enforceable effluent limits. N onpoint source pollution from anthropogenically managed lands is more difficult to mitigate and continues to threaten the balance of aquatic ecosystems worldwide (Duda, 1993). The diffuse nature of n onpoint source pollution suggests that a set of management practices, rather than effluent limits, are a more realistic abatement method to reduce n onpoint source pollu tion (Loehr, 1974). Despite regulations to protect water bodies from point and nonpoint sources based on total maximum daily loads (TMDLs) in the US, excessive nutrients (primarily from n onpoint sources) continue to degrade water quality. From 2008 to 2010, identified nutrient impaired waterways in Florida nearly doubled from about 1,600 to 3,000 km of streams and rivers and from 141,640 to 152,972 ha of lakes. The United States Environmental Protection Agency (EPA) and Florida Department of Environmental P rotection (FDEP) are developing numeric nutrient criteria for Floridas waters to protect these resources from nonpoint and point pollution. These criteria establish allowable limits for nutrient loads and/or concentrations (i.e., nitrogen and phosphorous) in waterways to improve water quality and protect public health, aquatic life and the long term recreational uses of Floridas waters (EPA, 2011). Following a lawsuit from the Florida Wildlife Federation in 2008, EPA set a rule in the fall of 2010 (effective July 2012) for water quality standards for inland waters in Florida, outside of south Florida.
47 The proposal of a second rule for estuaries, coastal waters and flowing waters in south Florida was set to be released in May 2012. In response to the rules set by the EPA, FDEP petitioned in April 2011 to have these rules repealed and to develop its own set of nutrient criteria. Given that states have the primary responsibility to maintain the quality of their waters, EPA may withdraw these rules if FDEP ca n develop their own protective and scientifically sound numeric standards (EPA, 2011). In January 2012, there was adoption of a final rule by FDEP, to be followed by legislative ratification (EPA, 2012). Once nutrient criteria have been established water bodies will be evaluated and designated as impaired or not impaired based on this standard. If impaired, nutrient sources will need to be quantified in order to allocate loads for establishing a TMDL. One source of nutrient nonpoint source pollution is fertilizers. Fertilizers are used to maintain turf and landscapes aesthetic qualities and are also used in agriculture to increase crop yields. Excessive nutrients in leachate or runoff may lead to increased eutrophication or conditions that alter natural ecosystem communities affecting freshwater supply and ecosystem balance. Nutrient leaching and runoff from turf and landscape may be considered more critical than in agricultural areas due to the proximity to drinking water supply (Waller, 2007). Nitrogen ( N) is a common nutrient in fertilizers and has been shown to be susceptible to leaching and runoff due to its solubility in water, particularly the nitrate ( NO3) form (Havlin et al., 2005). Nitrogen leaching is considered a primary means of N loss in humid climates and under irrigated conditions (Havlin et al., 2005). Some factors that influence NO3 leaching include: rate, tim ing source, and method of fertilization;
48 plant uptake; soil profile characteristics which affect percolation; and quantity, pattern, and time of rainfall and/or irrigation (Havlin et al., 2005). Certain forms of N such as NO3 and NH4, are also commonly limiting nutrients in ecosystems and a potential health hazard if NO3N is over 10 mg/L (EPA, 2012). While irrigation and fertilizati on are often required to maintain turf quality due to insufficient rainfall and soil nutrient deficiencies, excessive irrigation (in addition to heavy rainfall) can lead to increased runoff or leaching of nutrients (Barton and Colmer, 2006). An average of 2.95 1010 L (7.8 billion gallons) of water is used each day in the US for outdoor purposes, including irrigation, which constitutes 30% of the total daily water consumption in the US ( EPA, 2010). Even with proper fertilizer application, excessive water application may result in nutrient leaching (Barton and Colmer, 2006). Irrigation is often applied in excess when using an automatic irrigation system due to the set and forget mentality associated with these systems, which may apply 47% more water than nonautomatic techniques (Mayer et al., 1999). Over watering also results in decreased turf quality and potentially loss of applied nutrients (Trenholm and Unruh, 2005). Conservation water practices are necessary to ensure all water users have adequate supply and future water demands can be met There are several recommen dations given by the University of Florida (UF) Institute of Food and Agricultural Sciences (IFAS) on proper irrigation and fertilization, but compliance by property owners to follow these r ecommendations or statutes (e.g. mandatory rain sensors with automatic irrigation systems) remains a challenge. Studies in Florida indicate that residential properties can greatly reduce water applied by using soil water
49 based or weather based irrigation s cheduling devices (Dukes and Haley, 2009; Davis and Dukes, 2010; McCready and Migliaccio, 2011). Other studies have shown that soil water sensing devices can reduce nutrient leaching compared to conventional automatic irrigation (Augustin and Snyder, 1984; Snyder et al., 1984; Pathan et al., 2007). It follows that the other weather based technologies (e.g., evapotranspiration [ET] based) should also reduce percolation and nutrient leaching. Soil water sensors provide an indirect estimate of the soil water c ontent in situ. One common technology for regulating irrigation by estimating soil water content is time domain transmission (TDT). This technology measures the oneway (transmission) travel time of an electromagnetic pulse along a metal probe within a dielectric medium (i.e. soil, water, and air). The velocity is a function of soil permittivity, a, which is then related to the soil water content by calibration. Calibration is used to determine the field capacity (FC). Based on site FC, scheduled irrigation will be prevented if a certain percentage of FC exists or be applied if FC is below a set percentage of FC is reached (Muoz Carpena et al., 2005; Evett, 2007). Evapotranspirationbased controllers schedule irrigation using an estimation of ET and, typic ally, other sitespecific factors (i.e., soil type, plant type, etc.). There are two general categories of ET controllers: (1) soil water balance (SWB) based and (2) non soil water balance (nonSWB) based (Dukes, 2012). The SWB based controllers calculate a soil water balance continuously; whereas, the nonSWB controllers adjust run times relative to historical peak ET based on real time onsite measured weather conditions (i.e. temperature and/or solar radiation) (Dukes, 2012). Three basic ET controller ty pes are (1) signal based, (2) onsite weather measurement, and (3) historical
50 ET based (Dukes, 2009). For SWB based controllers, t he required amount of irrigation is calculated based on the amount of water necessary to replace water lost through ET from a w ater budget of irrigation and rainfall gains and ET losses ( using either real time or historical weather data). Controller type (1) receives weather data from a local weather station (typically for a service fee) and typically uses more real time weather d ata to calculate ET (i.e. wind speed, relative humidity, etc.) (Dukes et al., 2009). Controller type (2) has a small on site weather station and uses real time data (typically temperature and rainfall) and may also use historical data to calculate ET (Dukes, 2009). Controller type (3) is preprogrammed with regionspecific water use curves (which may be improved with additional onsite weather sensors) to predict irrigation requirements (Dukes, 2009). Depending on the manufacturer there are a variety of ot her factors that can be entered or selected including ZIP code, soil type, sprinkler type (with an assumed application rate and efficiency factor) slope, plant type, shade factor, and plant maturity to estimate the soil water budget. Another component that is used to improve the efficiency of automatic irrigation is a rain sensor (RS). Rain sensors are external devices that can be added to an automatic irrigation timer to control irrigation application by preventing a scheduled irrigation event if a certai n depth of rainfall has occurred, as detected by the device. There are various types of RSs including those based on water weight, electrical conductivity of water, or expansion disks which proportionally expand with the absorption of water (Dukes and Haman, 2002a). Thus, we combined irrigation technologies that have successfully been applied to improve irrigation practices with monitoring equipment to compare the potential leaching
51 of N and water losses by typical automated irrigation systems to that of the technology based irrigation systems. This knowledge will assist water managers with load estimations and allocations, if and when TMDLs are needed. This was accomplished by comparing (1) applied water quantities, (2) leachate quantities, and (3) leacha te N loads and concentrations among three technology based irrigation sys tems and a traditional automatic time based irrigation system. Materials and Methods The field study site was at the UF IFAS Tropical Research and Education Center (TREC) in Homestead Florida. The study was conducted during the end of the wet season and part of the dry season (31 August 2011 [29 September 2011 for leachate volumes] through 16 February 2012). The experiment was conducted on previously established bahiagrass, with an ar ea of 429.2 m2 (4620 ft2). The area was divided into 16 square plots, each 20.9 m2 (225 ft2), and 0.61 m (2 ft) buffer areas between plots. The experiment consisted of four treatments (T1 through T4), arranged in a randomized block design Each treatment w as replicated four times for water application and three times for leachate collection. In each plot, there were four quarter circle pop up irrigation heads with MP Rotator nozzles (Hunter Industries, Inc., San Marcos, CA) with an application rate of (0.5 in h1), one in each corner Leachate was collected twice a week from lysimeters following each scheduled irrigation event, in 250 mL plastic bottles using a small peristaltic pump (TAT pumps, Logan, OH) with 1.9 cm (0.75 in) outer diameter plastic tubing connected to a 12 V battery Leachate volumes were measured using a graduated cylinder. Water application volumes were monitored with DLJ multi jet water meters (Daniel L. Jerman Co., Hackensack, NJ) and recorded manually after each irrigation event.
52 Lysi meter Design and Installation Lysimeters were designed to capture up to a 3.81 cm (1.5 in) rainfall event. The height of 58.4 cm (23 in) water holding tube was selected based on diameter of the top catchment area (17.1 cm or 6.75 in) and the 3.8 cm (1.5 in) diameter to collect the maximum rainfall depth. Tubing was connected, one to collect water and the other to allow air to enter to reduce pressure inside. Screening was attached to the bottom of the catchment to minimize debris entering the water holding tube. Sand was placed in the catchment area to prevent finer particles from entering the water holding tube. The lysimeters were buried below the root zone (approximately 30 cm) in 12 plots, with three replicates for each treatment. Irrigation Treatments Treatment 1 (T1), or time based, consisted of irrigation controlled by a digital timer (Hunter Pro C Conventional 9 Zone Outdoor Model, Hunter Industries, Inc., San Marcos, CA), set to irrigate biweekly on Sunday and Thursday (Miami Dade County water rest rictions), from 02:00 to 03:00 to apply 1.3 cm (0.5 in). The runtime was increased to 1.5 h on 17 of December 2011 to apply 1.9 cm (0.75 in). The runtime was increased a second time to 2.5 h on 18 January 2012 to apply 3.2 cm (1.25 in). The runtime of the time based treatment was in creased twice to promote irrigation induced leaching (otherwise, leachate was only collected following relatively heavy rainfall events). The first two irrigation rates were within UF IFAS recommendations, which were not great enough to cause leaching from irrigation; however, the last rate was beyond the recommended rate, which caused leaching to occur following irrigation. Treatment 2 (T2) consisted of irrigation controlled by an ET controller (Rain Bird ESPSMT, Rain Bird Inc. Tucson, AZ). The ET controller was programmed by inputting
53 parameters including: ZIP code, soil type, sprinkler type, slope, plant type, shade factor, plant maturity (T able 2 1). The ET controller initiated irrigation based on an accumulated water requir ement, as determined by the digital record of real time weather data, including ETo and effective rainfall. Effective rainfall wa s determined by the onsite tipping bucket, which measured rainfall amount and intensity, and soil type input. The ET wa s calcu lated by an undisclosed equation using real time temperature and historical wind speed and solar radiation data for the designated ZIP code. Irrigation occur red when the management allowable depletion (MAD) of 50% wa s reached, as determined by the controll er monitored soil water balance. Based on inputs, the controller determine d the proper length of time and frequency to irrigate in order to prevent runoff (Q). Watering days were programmed as Sunday and Thursday (given Miami Dade County watering restricti ons) Allowed watering time was programmed as 1.5 h so that sufficient irrigation could occur if required, given that it operates on a cycle and soak principle to minimize Q. The controller default crop coefficient ( Kc) values are from the Irrigation Ass ociation (T able 22 ) (IA, 2008). The runtime and weather data were recorded manually from the ET controller after each irrigation event (although a digital record of the last 30 days was available). Treatment 3 (T3) consisted of a RS with an internal water absorbing expansion disk (Toro, Bloomington, MN). The RS treatment was controlled by a separate timer, as this technology will either apply the scheduled amount or completely bypass irrigation if the rainfall threshold has occurred, as detected by the sensor. Irrigation was set to occur on Sunday and Thursday, 03:00 to 04:00. At first, the rainfall threshold was manually set as 12 mm (22 September through 8 November 2011), but was replaced
54 and set at 3 mm on 9 November 2011. There were also two runtimes for T3: 1.0 h to apply 1.3 cm (0.5 in) per event (31 August 2011 to 22 January 2012) and 1.5 h to apply 1.9 cm (0.75 in) per event (23 January through 16 February 2012). Treatment 4 (T4) used a SWS (Baseline WaterTec S100, Baseline, Inc., 2009) to manage irr igation. T1 and T4 we re operated on the same controller, but at separate times to avoid system pressure losses. SWS treatments were scheduled for 04:00 to 05:00 on Sunday and Thursday, but irrigation was applied to a manufacturedprogrammed percenta ge (i.e 70%) of FC or bypassed if this minimum level of soil water was detected by the sensor. The SWS devices were installed in four plots. In each of the four plots, one sensor was buried in a corner opposite of the lysimeter, 58 cm (2 3 in) below the ground surface according to the Baseline WaterTec S100 Installation Manual. The sensor was installed horizontally to prevent pooling on the wide surface area of the sensor (Baseline, Inc., 2009). Each sensor had a digital controller that was wired into the automatic irrigation system and also connected to the buried sensor. Automatic calibration was performed by saturating the soil surrounding the sensor and using the 24hour automatic calibration feature on the controller. The Baseline SWS device uses TDT technol ogy and determines sufficient water content based on field capacity, permanent wilting point, and a manufacturer programmed threshold setting. Uniformity Testing Uniformity of water application was tested according to American Society of Agricultural Engin eers (ASAE) standard, American National Standards Institute ( ANSI ) /ASAE S436.1 (ASAE, 2003) prior to the beginning of the experiment (14, 15, and 16 September 2011). Uniformity tests were performed during the evening and early morning hours to avoid evapor ative losses. Each treatment zone was tested at separate
55 times (consistent with the irrigation schedule) to avoid system pressure losses. Twenty five catchments (GladTM 947 mL disposable containers) were placed in an evenly spaced grid pattern in each plot with the outermost catchments being 0.3 m (1 ft) inset from the plot boundaries (distinguished by sprinkler head location). The irrigation runtime was two hours. The wind speed was recorded every 15 minutes with a handheld digital anemometer to ensure that wind speed did not exceed 5 m s1, which may affect the results At the end of the test period, collected water volumes were measured with a graduated cylinder. The opening diameter of the catchments was 13.3 cm (5.25 in). Distribution uniformit y (DU) w as calculated with Equation 21. tot DU lq DU lq DU ( 2 1 ) Turfgrass Quality Evaluation Turfgrass quality was evaluated according to the National Turfgrass Evaluation Program guidelines for the following characteristics: genetic color, turfgrass densit y, percent living ground cover, and texture. Quality was based on a visual rating scale of one (worst) to nine (best). For genetic color, or inherent genotype color, one is light green and nine is dark green. Turfgrass density is a visual estimate of living turfgrass per unit area. Percent living ground cover is an estimate of surface area that is covered by the originally planted species. Turfgrass texture is an estimate of leaf width, with one being course and 9 being fine (Morris and Shearman, 1998). Fer tilizer Application All 16 plots were fertilized according to UF IFAS recommendation for slow release N fertilizer, 4.8 g N m2 (1 lb N per 1000 ft2) (Trenholm and Unruh, 2005). LESCO Professional Turf 262 11 (N P K) fertilizer was used with 26% total N, 1.15%
56 ammonical N and 24.85% Urea N (LESCO, Cleveland, OH). Given this composition, the total amount applied to the 16plot study area was 8.8 kg (19.3 lbs). The urea N was 6.5% slowly available urea N from Sulfur coated urea. Other components included 2% phosphate, 11% soluble potash, 0.50% magnesium total, 4% sulfur, 0.10% copper total, 3% iron total, 0.80% manganese total, 0.20% zinc total. The maximum chloride was 8.25%. F ertilizer was applied using a Scotts Standard broadcast spreader by pushing the spreader at a constant pace in a serpentine pattern at the setting of 5.5, as recommended on the LESCO fertilizer bag. The first fertilizer application was on 7 September 2011 and the second application was 18 January 2012. Nitrogen Analysis L eachate colle cted from lysimeters was filtered with Whatman 42 filter paper into 20 mL vials, frozen, thawed, and analyzed for NO3N and NH4N concentrations using a SEAL AQ2 discrete analyzer (SEAL Analytical, Inc., Mequon, WI). Nitrate was determined spectrophotometr ically by reducing NO3 to NO2 using a cadmium coil and the determining the NO2 concentration b y USEPA method 353.2 ( EPA, 1993 a ). Ammonia and ammonium ion were determined spectrophotometrically by USEPA method 350.1 Rev. 2.0 ( EPA, 1993 b ). The detection l imits of NO3N and NH4N were 0.021 and 0.046 mg/L, respectively. For concentrations less than these detection limits, the concentration was taken as half of the limit (i.e., the minimal NO3N and NH4N concentrations were 0.01 and 0.023 mg/L, respectively ). Initial (before fertilizer application) and final soil and tissue samples were also collected and analyzed. The tissue and soil samples were analyzed for total N and C using the elemental analyzer vario MAX CNS (Elementar Analysensysteme, Hanau, Germany ). This automatic instrument uses catalytic tube combustion under oxygen supply and high temperatures.
57 Water Balance In addition to comparing significant differences among treatments, an overall soil water balance was estimated for each irrigation treatment considering cumulative dept hs of irrigation (I) and percolation (Perc), averaged across replicates, and rainfall (R) and reference ET (ETo) ( Equation 2 2 ). The average water depths applied and leached for each treatment were summed over the entire study period. Daily r ainfall and ETo were obtained from the FAWN Homestead weather station and summed over the study period. Runoff was not measured and is considered negligible in our system due to high infiltration rates with a mean asymptotic infiltration capacity of 36 cm h1 according to Chin (2007) and flat topography (slope = 0) Perc I o ET R SW ( 2 2 ) Data Analysis Irrigation water volumes applied and leached per plot were measured and converted to depths. Data were analyzed for significant differences (p treatments for water applied, leachate depths, NO3N and NH4N concentrations, and NO3N and NH4N loads by performing either or a oneway analysis of variance (ANOVA) (parametric) for normally distributed data or Kruskal Wallis one way ANOVA on ranks (nonparametric) for data that failed normality. For the ANOVA on ranks, an H statistic and P value were given by ranking all observations in ascending order and comparing the average value of the ranks for each treatment. Pairwise mult iple comparisons were performed using Tukeys test. Data analyses for water quantities applied and leached and were divided into three time periods according to the application rates per event of the time based treatment (T1): (1) 1.3 cm (0.5 in) from 22
58 S eptember through 16 December 2011; (2) 1.9 cm (0.75 in) from 17 December 2011 through 17 January 2012; and (3) 3.2 cm (1.25 in) from 18 January through 16 February. Data analyses for N concentrations and loads were divided into two periods: after the first fertilizer application (22 September 2011 to 18 January 2012) and after the second application (19 January to February 16 2012). For times when no leachate was collected from the lysimeter, the NO3N and NH4N concentrations and loads were considered to b e 0. Results and Discussion Results of Quantity of Irrigation Water Applied The normality test failed for data collected from all three timebased (T1) application rates, so data were analyzed by a oneway ANOVA on ranks using the Kruskal Wallis test. For the three different T1 application rates, water depths applied were significantly different (P<0.001). Water depth applied by T1 was always greater and significantly different (P<0.05) from each of the three technology based treatments (T2, T3, and T4). Ad ditionally, T2, T3, and T4 were compared over the entire study period. These data were not normally distributed, but based on the ANOVA on ranks, the medians of these treatments were significantly different (P <0.001). Results showed that T3 was significan tly different (P<0.05) from T2 and T4, as well as T2 being significantly different from T4. Cumulatively, T1 applied the greatest amount of water 68.2 cm (replicate summation average), while T2 applied the least amount of water 17.3 cm (Table 25) Results indicated that T2 had the greatest water savings (70, 69, and 83% for T1 application rates 1, 2, and 3, respectively). During T1 application rate 3, T4 also resulted in 83% water savings, although the T4 savings were 62 and 60% for T1 application rates 1 and 2, respectively. The RS based treatment (T3) resulted in the
59 least water savings of 9% for the 12 mm rain sensor setting (31 August to 9 November 2011) and averaging 26% savings (1653% during the different application rates) for the 3 mm setting (10 N ovember 2011 to 16 February 2012). The average application depth of T1 for application rate 1 was 1.1 cm, 1.7 cm for application rate 2, and 2.9 cm for application rate 3 (Table 24) Over the entire study period, the average application depths of ET based (T2) and SWS based irrigation (T4) were 0.4 cm and 0.5 c m, respectively. The RS based treatment (T3) applied an average of 1.0 cm per irrigation event from 22 September 2011 through 22 January 2012 (although equipment was replaced and threshold changed fr om 12 to 6 mm on 9 November 2011). After the scheduled watering window for T3 was increased to 1.5 h (23 January through 16 February 2012), T3 applied an average of 1.4 cm per event. Over the entire study period (22 September to 16 February 2012), cumulati ve rainfall was 55.4 cm and cumulative reference ET was 30.8 cm. Averaging the three replicates for each treatment, the average total leachate collected from T1, T2, T3, and T4 was 22.1 cm, 14.1 cm, 14.4 cm, and 15.2 cm, respectively (T able 2 5). Results o f Water Leached The normality test failed for all three time periods (T1 application rates 1 through 3) and a nonparametric test was performed for each leachate data set. For application rate 1 (29 September to 17 December 2011), the leachate depths were not significantly different. For application rate 2, the differences in median values among treatments were significantly different (P < 0.001); additionally T2 and T3 were significantly different from T1. For application rate 3, the differences in median v alues among treatments were significantly different (P < 0.001). During application rate 3; T2, T3, and T4 leachate depths were significantly different from T1 ( Table 2 6). The average leachate depth of T1
60 for application rate 1 was 0.5 cm, 0.1 cm for appli cation rate 2, and 1.1 cm for application rate 3 ( Table 2 6). Although the irrigation was increased for application rate 2, there was the least amount of rainfall during this period; correspondingly, leachate depth was the lowest during this time. Total ra infall depths during application rates 1, 2, and 3 were 36.7, 0.03, and 6.9 cm, respectively. Table 2 6 provides the mean leachate depths for the individual timebased application rates; over the entire study period, the average leachate depth of T2 and T4 were 0.4 cm and 0.3 cm, respectively. For both irrigation runtimes of T3 (1.0 h from 31 August 2011 to 22 January 2012 and 1.5 h from 23 January through 16 February 2012), the average leachate depth was 0.4 cm. Results of Nutrients Leached NO3N The medians, means, and standard deviations of the concentrations and loads of NO3N and NH4N for two fertilizer applications for T1 through T4 are given in Table s 2 7 and 2 8. F or all treatments over the entire study period, the average loads were 2.44E 02 kg NO3N ha1 and 5.39E 03 kg NH4 ha1. After the first fertilizer application, differences among treatment median NO3N concentrations and loads were not significantly different (P = 0.299 and P = 0.819 for concentrations and loads, respectively) based on a one way ANOVA on ranks ( Tables 2 7 and 2 8). When irrigation was increased passed the recommended amount along with the second fertilizer application, concentrations and loads were significantly different (P<0.001 for both). Additionally, after irrigation was increased passed the recommended amount along with the second fertilizer application, T1 NO3N concentrations and loads were significantly greater than each of the technology based treatments (P<0.05) ( T ables 2 7 and 2 8). Over the entire study period, t he average NO3N concentrations of T1, T2, T3,
61 and T4 were 1.37, 0.24, 2.36, and 0.24 mg L1, respectively ; the average loads were 0.24, 0.08, 0.70, and 0.08 mg, respectively The average NO3N concentrations of T3 after the first and second fertilizer applications were 0.68 and 4.05 mg L1 ( average lo ads of 0.12 and 1.3 mg ) respectively. The maximum concentrations of NO3N for T1, T2, T3, and T4 were15.8, 15.6, 93.9, and 15.1 mg L1, respectively which exceed the drinking water standard. Maximum concentr ations of T1 and T3 were obtained after the second fertilizer application (and corresponding irrigation increase beyond UF/IFAS recommendation) while the maximum concentrations of T2 and T4 were obtained after the first application. NH4N Similar to the NO3N results, differences among treatment median NH4 concentrations and loads were not significantly different (P = 0.086 and P = 0.528 for concentrations and loads, respectively) after the first fertilizer application based on a oneway ANOVA on ranks ( T ables 2 8 and 2 9 ). When irrigation was increased passed the recommended amount along with the second fertilizer application, concentrations and loads were significantly different (P<0.001 for both). Additionally, after irrigation was increased above the r ecommended amount along with the second fertilizer application, T1 NH4 concentrations and loads were significantly greater from each of the technology based treatments (P<0.05) ( Tables 2 7 and 2 8). Over the entire study period, the average NH4 concentrati ons of T1, T2, T3, and T4 were 1.08, 0.07, 0.06, and 0.19 mg L1; the average loads were 0.12, 0.01, 0.01, and 0.04 mg per event respectively. The maximum concentrations of NH4 for T1, T2, T3, and T4 were 9.21, 5.95, 4.85, and 6.26 mg L1, respectively, a ll obtained after the first fertilizer application. The percentage of
62 NO3N and NH4N leached of that applied is presented in Table 27, showing that, on average, a higher percentage of the ammonical form was leached. Turfgrass Quality Evaluation Results Results from the initial and final turfgrass evaluations indicated that the overall quality of the turfgrass for each treatment either remained the same or slightly improved over the duration of the experiment ( Table 2 10). No differences were observed among treatments suggesting that no treatment was water stressed. This is supported by the water balance which showed that irrigation and rainfall exceeded ET demands. Tissue and Soil Results The results of the initial (before fertilization) and final turfgrass tissue and soil total percent N and total percent C revealed that overall, averaging all plots, total percent N in the tissue increased from 1.89% to 2.17% ( Table 2 11). Total percent N in the soil increased from an average of 0.20% to 0.25% ( Table 2 12 ). Discussion As expected the technology based irrigation systems, T2 through T4, resulted in significant water savings compared to the time based control. This has been demonstrated before in similar field studies (Davis et al., 2007; Haley and Dukes, 2007; Shedd et al., 2007; Cardenas Laihalcar et al., 2008; Dukes et al., 2008; Dukes and Haley, 2009; McCready et al., 2009; Cardenas Laihalcar et al., 2010; Cardenas Laihalcar and Dukes, 2010; Davis and Dukes, 2010; McCready and Dukes 2011). Table 2 13 comp ares water savings of the technology based irrigation systems in this study and similar studies. The variance of applied water among replicates was lowest for the ET controller, followed by SWS, RS and time based ( Table 2 14). The greater variability among the SWS replicates is expected given that irrigation was controlled by
63 a separate sensor in each replicate plot and the soil water content would not be exactly consistent among replicates. The SWS data were random, there was no particular trend. The water savings results of the ET controller we re relatively high compared to other studies ( Table 2 13), and still the turfgrass quality was maintained throughout the study. This is partly due to comparison to an elevated timebased irrigation rate. Considering the cumulative water balance (except runoff) over the study period for the ET controller, there was still an average surplus of 27.8 cm. The greater savings may be due to different weather conditions from other studies which caused the ET controller to apply less water or because of the crop coefficient which are higher for this area and therefore would have result ed in greater irrigation. Other studies indicated that SWS based irrigation also reduces nutrient leaching (Augustin and Snyder, 1984; Snyder et al., 1984; Pathan et al., 2007). Pathan et al. (2007) found that leachate was only 4% of applied water draining from SMS plots versus 16% from conventionally irrigated plots. Leachate median quantities were significantly different among treatments during the second and third timebased application rates. Additionally T2, T3, and T4 leachate depths were significantly less from T1 during the third application rate. When the average leachate depths were summed over the entire study period, the timebased irr igation resulted in the greatest quantity of leachate (22.1 cm); whereas the ET based irrigation resulted in the least amount of leachate (14.1 cm), which is a 36% reduction in leachate compared to the traditional time based irrigation. The SWS based irrigation yielded 15.2 cm leachate (31% less than time based) and the rain sensor based yielded 14.4 cm leachate (35% less than time based) over the entire study period. The rain sensor based irrigation resulted in
64 slightly less leachate than the SWS, but th e cumulative leachate depths of all three technology based systems were relatively close. The slight differences in the leachate summations may be due to inconsistencies in lysimeter design among the treatments which may have enabled one lysimeter to collect and pump water more efficiently than another, which also affected the numbers of samples being collected from each treatment. The lysimeters were pumped a total of 114 days; however, total leachate samples from T1, T2, T3, T4 were 53, 14, 26, 28 samples, respectively. The greater number of leachate samples collected from the SWS compared to the rain sensor likely biased the results for the cumulative leachate depth. According to Petrovic (1990), l eaching is affected by fertilizer management practices (rat e and timing), N source, soil texture, rainfall, and irrigation. Given that UF IFAS recommended amounts of irrigation water and fertilizer were applied, the resulting loads were similar to the values obtained by other studies which also applied moderate am ounts of fertilizer and irrigated at rates equal to or less than potential ET (Table 2 15) (Barton and Colmer, 2006). Even when time based irrigation was increased beyond the recommended level and a second fertilizer application, the highest average NO3N and NH4N concentrations among all treatments were still substantially below the 10 mg/L drinking water standard, 1.43 mg NH4 L1 (time based) and 4.05 mg NO3N L1 (RS), or loads of 0.75 kg NH4 ha1 (time based) and 5.03 kg NO3N ha1 (RS). For the majori ty of the study period, which took place during the dry season, leaching from all treatments seemed to be driven by heavy rainfall rather than irrigation ( Figure 2 2). This was reasonable given that irrigation was not applied excessively, except during the last four weeks when timebased irrigation was increased beyond the
65 UF IFAS recommended level. For excessive application, leachate was collected from the T1 replicates after every irrigation event ( Figure 2 3). Still, rainfall further increased the leacha te volume collected ( Figure 2 3). A possible explanation for the low leachate volume collected may be due to preferential flow outside of the catchment area of the lysimeters and the inflow rate limitations of the lysimeters; therefore, the true amount of leachate may not have been collected During the lysimeter installation, the top layer of soil was removed and replaced so that normal drainage patterns were disturbed and preferential flow pathways may have been created which drain outside of the lysimete r catchment area. A portion of applied water is usually intercepted in discontinuous macropores (internal catchments) within the soil. Booltink and Bouma (1991) demonstrated that 42% of the volume of applied water was affected by these internal catchments. Preferential flow may occur in, and even be enhanced by, unsaturated conditions, which corresponds to when most differences occurred between observed and simulated values. There are several examples of this occurrence, as cited by Nimmo (2012). Preferenti al flow is typically considered as nonequilibrium flow, given that, infiltrating water does not have sufficient time to equilibrate with resident water (Jarvis, 1998). Due to this nonequilibrium, water has been found to flow in thin films and that preferential flow stops suddenly when water is no longer applied, implying that unsaturated adjacent matrix material readily absorbs the water (Cey and Ruldoph, 2009). Rimon et al. (2011) also found different parts of a pore wetted while other portions remained dry. Antecedent moisture conditions do not necessarily affect preferential flow, as Stumpp and Maloszewski (2010) concluded that preferential flow can also occur during highintensity
66 rainfall even if the soil is initially dry. Structural continuity of the soil matrix also affects drainage patterns. Smettem et al. (1994) stated that structural continuity substantially increases soil drainage, more so than the antecedent water content. Wetter antecedent moisture conditions may not always guarantee faster flow (Nimmo, 2012). Furthermore, preferential flow can occur within different parts of a pore. Considering all of these supporting facts, it is likely that preferential flow had a significant impact on the outcome of field results and that the full amount of leachate was not captured. It is possible that applied water was conveyed horizontally through preferential pathways away from the lysimeter catchment. Future lysimeter designs should include ways to account for preferential flow in both dry and wet c onditions.
67 Table 21. Evapotranspiration ( ET ) controller input parameters used. Parameter Site specific input Z IP code 33031 Soil type Loamy sand Sprinkler type Rotary nozzle Slope 0 2% Plant type Grass lawn > warm season variety Shade factor Full sun Plant maturity Established Table 2 2 Crop coefficient (Kc) values used by the Rain Bird ESP SMT ET controller for warm season turfgrass (IA, 2008). K c Jan 0.52 Feb 0.64 Mar 0.70 Apr 0.73 May 0.73 Jun 0.71 Jul 0.69 Aug 0.67 Sep 0.64 Oc t 0.60 Nov 0.57 Dec 0.53 *Bermudagrass values used for all warm season turfgrass species. Table 2 3 Irrigation system quality ratings for evaluating distribution uniformity (DU) (IA, 2005 a ). Rating DU Fail < 0.40 Poor 0.40 0.49 Fair 0.50 0.54 Good 0.55 0.64 Very Good 0.65 0.74 Excellent 0.75 and above Table 2 4 Comparison of medians, means, and standard deviations of water depths applied during three timebased application rates among the four treatments: time based (T1), evapotranspi ration controller (T2), rain sensor (T3), and soil water sensor (T4). Rate 1 Rate 2 Rate 3 T1 Median cm 1.1a 1.7a 2.9a Mean cm 1.1 1.7 2.9 Standard Deviation cm 0.1 0.1 0.2
68 Table 24. Continued. Rate 1 Rate 2 Rate 3 T2 Median cm 0.4b 0 .5b 0.6b Mean cm 0.3 0.5 0.5 Standard Deviation cm 0.3 0.1 0.3 T3 Median cm 1.0c 1.1c 1.5c Mean cm 0.9 1.1 1.4 Standard Deviation cm 0.4 0.1 0.5 T4 Median cm 0.0b 1.1b,c 0.0b Mean cm 0.4 0.7 0.5 Standard Deviation cm 0.5 0.6 0.6 *Treatment m edian depths with different letters were significantly different (P<0.05) by column (application rate) using Tukey test. Table 2 5 Comparison of average cumulative quantities*, medians, means, and standard deviations among the four treatments: timebase d (T1), evapotranspiration controller (T2), rain sensor (T3), and soil water sensor (T4) for water depths applied and leached during the study period (22 September 2011 through 16 February 2012). b Water a Irrigation a Percolation Balance Cumulativ e cm 68.2 22.1 +70.7 T1 Median** cm 0.0a Mean** cm 0.6 Standard Deviation** cm 0.9 Cumulative cm 17.3 14.1 +27.8 T2 Median cm 0.5a 0.0b Mean cm 0.4 0.4 Standard Deviation cm 0.3 0.8 Cumulative cm 44.7 14.4 +54.9 T3 Median cm 1.0b 0.0b Mean cm 1.0 0.4 Standard Deviation cm 0.4 0.8 Cumulative cm 19.7 15.2 +29.1 T4 Median cm 0.0c 0.0b Mean cm 0.5 0.4 Standard Deviation cm 0.6 0.8 *Replicate values were averaged and summed. **For T1, there were three different applic ation rates, so the median, mean, and standard deviation are presented in a separate table. aTreatment median depths with different letters were significantly different (P<0.05) by column based on Tukey test. bCalculated as summation of rainfall and irrigation, minus ET and percolation over study period.
69 Table 2 6 Comparison of water depths leached among the four treatments: timebased (T1), evapotranspiration controller (T2), rain sensor (T3), and soil water sensor (T4) for the three timebased applicat ion rates. Rate 1 Rate 2 z Rate 3 z T1 Median cm 0.0 0.1a 0.8a Mean cm 0.5 0.1 1.1 Standard Deviation cm 1.0 0.1 0.8 T2 Median cm 0.0 0.0b,c 0.0b Mean cm 0.5 0.0 0.3 Standard Deviation cm 0.9 0.0 0.7 T3 Median cm 0.0 0.0b,c 0.0b Mean cm 0.5 0 .0 0.3 Standard Deviation cm 0.9 0.0 0.7 T4 Median cm 0.0 0.0a,c 0.0b Mean cm 0.5 0.1 0.3 Standard Deviation cm 0.9 0.2 0.7 zTreatment median depths with different letters were significantly different (P<0.05) by column (application rate) based on Tukey test. Table 2 7 Comparison of nit rate and ammonium percent leached for two fertilizer applications (noted as 1 and 2) among the four treatments: timebased (T1), evapotranspiration controller (T2), rain sensor (T3), and soil water sensor (T4). NO 3 N (1) NO 3 N z (2) NH 4 N (1) NH 4 N z (2) % % % % T 1 1.47 2.43 14.6 24.7 T 2 3.13 0.01 6.81 0.20 T 3 2.26 7.69 7.95 0.24 T 4 0.53 0.75 32.2 0.22 Average 1.84 2.72 15.4 6.34 Table 2 8 Comparison of nitrate and ammonium concentrations for two fertilizer applications (noted as 1 and 2) among the four treatments: timebased (T1), evapotranspiration controller (T2), rain sensor (T3), and soil water sensor (T4). NO 3 N (1) NO 3 N z (2) NH 4 N (1) NH 4 N z (2) T1 Median mg/L 0.0 0 0.30a 0.31 0.02a Mean mg/L 0 .46 2.29 0.73 1.43 Standard Deviation mg/L 2.19 4.08 1.90 2.23 T2 Median mg/L 0.00 0.0 0 b 0.10 0.00b Mean mg/L 0.48 0.0 1 0. 14 0.01 Standard Deviation mg/L 2.25 0.0 2 0.65 0.01 T3 Median mg/L 0.00 0.0 0 b 0.02 0.00b Mean mg/L 0.68 4.05 0.10 0.01 Sta ndard Deviation mg/L 3.48 18.21 0.52 0.01
70 Table 28. Continued. NO 3 N (1) NO 3 N z (2) NH 4 N (1) NH 4 N z (2) T4 Median mg/L 0.00 0.0 0 b 0.54 0.00b Mean mg/L 0.26 0.23 0.38 0.01 Standard Deviation mg/L 1.59 1.1 7 1.14 0.01 zTreatment median depths wit h different letters were significantly different (P<0.05) by column (applicat ion rate) based on Tukey test Table 2 9 Comparison of nitrogen loads among treatments for two fertilizer applications (noted as 1 and 2) among the four treatments: timebased (T1), evapotranspiration controller (T2), rain sensor (T3), and soil water sensor (T4). NO 3 N (1) NO 3 N (2) NH 4 N z (1) NH 4 N (2) T1 Median mg 0.0 0 0.02a 0.00 0.01 a Mean mg 0.08 0.41 0.03 0.21 Standard Deviation mg 0.80 0.86 0.10 0.39 T2 Median mg 0.00 0.00b 0.00 0.00 b Mean mg 0. 1 6 0.0 0 0.02 0.00 Standard Deviation mg 0.95 0.00 0.06 0.00 T3 Median mg 0.00 0.00b 0.00 0.0 0b Mean mg 0.12 1.28 0.02 0.00 Standard Deviation mg 0.6 9 6.32 0.08 0.00 T4 Median mg 0.00 0.00b 0.00 0.0 0b Mean mg 0. 0 3 0. 13 0.08 0.00 Standard Deviation mg 0. 17 0.64 0.28 0.00 zTreatment median depths with different letters were significantly different (P<0.05) by column (applicat ion rate) based on Tukey test Table 210. Initial and final ratings of turfgrass qualit y, based on scale of 1 to 9. Initial Final T 1 6 6 T 2 5 6 T 3 6 6 T 4 5 6 Table 2 11. Total percent nitrogen and carbon in the turfgrass tissue at the beginning of the experiment (before fertilizer applied) and at the end of the experiment. Total % N Total % C Treatment Initial Final Initial Final 1 1 2.2 2.3 45.0 41.9 1 2 2.0 1.8 44.2 42.5 1 3 1.8 2.0 44.2 42.5 1 4 2.0 2.0 44.8 42.4 2 1 2.0 2.3 45.2 42.6
71 Table 211. Continued. Total % N Total % C Treatment Initial Final Initial Final 2 2 2.1 2.5 45.4 42.9 2 3 2.1 2.0 45.1 42.1 2 4 1.9 1.9 44.8 42.9 3 1 2.0 2.5 44.9 42.8 3 2 2.0 2.5 44.7 42.2 3 3 2.1 2.1 45.5 42.7 3 4 1.9 2.5 45.5 42.5 4 1 1.3 2.1 45.0 42.6 4 2 1.6 2.1 44.8 42.7 4 3 1.6 1.6 44.7 44.7 Average 1.9 2.2 44.9 42.7 Table 2 12. Total percent nitrogen and carbon in the soil at the beginning of the experiment (before fertilizer applied) and at the end of the experiment. Total % N Total % C Treatment Initial Final Initial Final 1 1 0.21 0.23 6.12 5.73 1 2 N/A 0.3 5 N/A 8.06 1 3 0.27 0.16 7.62 3.88 1 4 0.14 0.29 3.09 5.36 2 1 0.35 0.26 6.22 5.80 2 2 0.23 0.26 5.70 7.29 2 3 0.21 0.19 6.36 4.95 2 4 0.24 0.21 6.28 4.86 3 1 0.27 0.35 5.67 8.01 3 2 0.17 0.34 3.30 7.80 3 3 0.18 0.28 3.72 6.22 3 4 N/A 0.19 N/A 4. 62 4 1 0.20 0.15 4.66 3.51 4 2 0.27 0.12 6.64 3.13 4 3 0.16 0.32 3.65 6.07 4 4 0.20 0.27 5.62 4.30 Average 0.20 0.25 4.67 5.60 Table 213. Water savings of three technology based irrigation systems compared to time based irrigation. Water Savings (%) Resultsa Similar Studies Reference Rain sensor 16 53 7 30 McCready et al. (2009) 1324 Cardenas Lailhacar et al. (2010) 14* Haley and Dukes (2012)
72 Table 213. Continued. Water Savings (%) Resultsa Similar Studies Reference 19 Haley and Dukes (2007) 34 Cardenas Lailhacar et al. (2008) Soil water sensor 60 83 11 53 McCready et al. (2009) 1683 Cardenas Lailhacar et al. (2010) 25 Pathan et al. (2007) 40 Horst and Peterson (1990) 42 95 Augustin and Snyder, (1984) 42 95 Snyder et al. ( 1984) 65 Haley and Dukes (2012) 6992 Cardenas Lailhacar et al. (2008) 73 Qualls et al. (2001) Evapotranspiration controller 6983 2563 McCready et al. (2009) 20 60 Davis et al. (2007) aRange of average savings for three different timebased application rates. *Not statistically significant. Table 214. Comparison of leachate load results to the results of similar studies on turfgrass. NH4Na kg ha -1 yr -1 NO3Na kg ha -1 yr -1 Other Studiesb kg N ha -1 yr -1 Reference Time based 6.34 c 2.96 c 2.79 (0) d Morton et al. (1988) e 13.65 (97) 31.94 (244) 1.3 1.4 (0) Gold et al. (1990) f 4.1 (300) Erikson et al. (2001) g Rain s ensor 16.2 0.61 Soil water sensor 2.08 2.44 1.88 (0) Morton et al. (1988) 3.04 (97) 4.87 (244) E vapotranspiration controller 5.09 0.53 aFertilizer application rate equal to 249 kg ha-1 yr-1. The study period less than 1 year was converted from days to a fraction of year for comparison purposes. bOther studies may include other forms of N and also have variable irrigation and fertilization rates.
73 cThree different timebased irrigation rates over study period (2.5, 3.8, and 6.4 cm wk-1). dFertilizer application rate. eIrrigation = 3.75 cm wk-1. fIrrigation rate not given. gI rrigation = 2.0 cm wk-1. Figure 2 1 Lysimeter for collecting leachate below root zone. Figure 22. Time based irrigation and percolation during wet season (2 October 2011 to 5 November 2011), showing that leachate was col lected only after heavy rainfall.
74 Figure 2 3 Time based irrigation and percolation during dry season after irrigation runtime was increased to apply 3.2 cm per event (22 January 2012 to 16 February 2012), showing that leachate was collected after each irrigation event and even greater amounts of leachate following rainfall.
75 CHAPTER 3 INTERACTIVE TOOL FOR SIMULATING IRRIGATION TECHNOLOGIES IN A VIRTUAL TURFGRASS SY STEM Background Turfgrass is a substantial land cover in the United States (US) encompass ing approximately 20 million ha with a value of US$40 billion which exceeds the combined value of corn and soybean (IA, 2005a). In fact, Milesi et al. (2005) calculated that turfgrass covers an area three times larger than any other irrigated crop. This is not su rprising considering that 80% of households, or 85 million households, participate in outdoor lawn or garden activities (IA, 2011). Turfgrass growth occurs during warm months when rainfall may not meet plant requirements resulting in water stress. Since aesthetics and health are primary factors for many turfgrass owners, irrigation systems are often implemented to supplement plant water needs. The typical lawn consumes an additional 37,854 L (10,000 gal) of water beyond rainwater each year (USEPA, 2 010). Irrigating turfgrass has evolved from manual hoseand sprinkler systems to automated irrigation systems. The downside of automation is that systems tend to over water, resulting in decreased turf quality and potentially loss of applied nutrients (Tr enholm and Unruh, 2005). Mayer et al. (1999) states that 47% more water is used by automated irrigation systems than nonautomated systems (or manual systems) for landscape irrigation due to a set and forget mentality. Thus, it is not surprising that US outdoor water use can account for up to 70% of household water use during summer months, with a 32% annual usage average (IA, 2011).
76 According to Milesi et al. (2005), Florida ranks second (behind Texas) in the 48 conterminous states in estimated turfgras s area, with 1.2 million ha of turfgrass. Over 400,000 ha of home lawns in Florida are professionally managed (Trenholm and Unruh, 2005). Many residential properties in Florida have automatic inground irrigation systems to maintain green lawns. However, automatic irrigation systems are particularly inefficient in Florida due to the wet and dry seasonal climate, the intensity of rainfall events, and the low water storage capacity of most Florida soils. This is supported by Milesi et al. (2005) who simulated the differences between two irrigation methods, (1) a set application of 2.54 cm of water per week and (2) irrigation based on potential evapotranspiration ( ET ) and rainfall. Their results indicated that the second method would result in greater water application in the West (up to an additional 72 cm/yr in the Southwest) and less in the Southeast (decreased by as much as 38 cm/yr in southern Florida) (Milesi et al., 2005). The inefficiency of automatic irrigation systems has led to the development of tech nologies that use rain sensors (RS s) soil water sensors (SWS s), and/or ET controllers to improve irrigation effectiveness. These irrigation technologies have been branded by the Irrigation Association (IA) as Smart Controllers (IA, 2007). Several studies have been conducted in Florida on the water savings associated with these Smart Controllers compared to conventional timebased irrigation scheduling (Davis et al., 2007; Haley and Dukes, 2007; Shedd et al., 2007; Cardenas Laihalcar et al., 2008; Dukes e t al., 2008; Dukes and Haley, 2009; McCready et al., 2009; Cardenas Laihalcar et al., 2010; Cardenas Laihalcar and Dukes, 2010; Davis and Dukes, 2010; McCready and Dukes 2011). The concept behind Smart Controllers is to irrigate based on the
77 soil water deficit, by either sensing soil water content or estimating ET, so that the soil water content (SWC) is restored to a certain percentage of field capacity (FC). Soil water sensors provide an indirect estimate of the SWC in situ. One common technology for regulating irrigation by estimating soil water content is time domain transmission (TDT). This technology measures the oneway (transmission) travel time of an electromagnetic pulse along a metal probe within a dielectric medium (i.e. soil, water, and air ). The velocity is a function of soil permittivity, a, which is then related to the soil water content by calibration. Calibration is used to determine the FC. Based on site FC, scheduled irrigation will be prevented if a certain percentage of FC exists or be applied until the set percentage of FC is reached (Muoz Carpena et al., 2005; Evett, 2007). Evapotranspirationbased controllers schedule irrigation based on an estimation of ET and, typically, other sitespecific factors (i.e., soil type, plant type, etc.). There are two general classifications of ET controllers: (1) soil water balance (SWB) based and (2) nonSWB. The three basic ET controller types are (1) signal based, (2) onsite weather measurement, and (3) historical ET based (Dukes, 2009). For S WB based controllers, t he required amount of irrigation to be applied during scheduled times is calculated based on the amount of water necessary to replace water lost through ET from a water budget of irrigation and rainfall gains and ET losses (based on either real time or historical weather data). Controller type (1) receives weather data from a local weather station (typical ly for a service fee) and uses an equation to calculate ET based on additional real time weather factors (i.e. wind speed, relative humidity, etc.) (Dukes et al., 2009). Controller type (2) has a small onsite weather station and uses real time
78 data (typically temperature and rainfall) and historic data to calculate ET (Dukes, 2009). Controller type (3) is preprogrammed with regionspecific water use curves (which may be improved with additional onsite weather sensors) to predict irrigation requirements (nonSWB based) (Dukes, 2009). Depending on the manufacturer there are a variety of other factors that can be entered or selected i ncluding ZIP code, soil type, sprinkler type, slope, plant type, shade factor, and plant maturity to evaluate the soil water budget (T able 3 1). Another component that is used to improve the efficiency of automatic irrigation is a RS Rain sensors are ext ernal devices that can be added to an automatic irrigation timer to control irrigation application by preventing a scheduled irrigation event if a certain depth of rainfall has occurred, as detected by the device. There are various types of RSs including t hose based on water weight, electrical conductivity of water, or expansion disks which proportionally expand with absorption of water (Dukes and Haman, 2002a). The SWSs, ET controllers, and RSs are used to improve the estimation of irrigation needs based, in some regards, on the concept of a soil water balance (Allen et al., 1998): i DP i c ET i CR i net I i RO P i r D i r D ) ( 1 ( 3 1 ) where Dr (mm) is depletion of water from root zone, i is the current day, i 1 is the previous day, P (mm) is daily precipitation, RO (mm) is runo ff, Inet (mm) is net irrigation depth, CR (mm) is capillary rise from the groundwater table, ETc (mm) is crop ET and DP (mm) is deep PERC (Allen et al., 1998). Depending on the technology used, different components of Eq 1. are considered. The SWS controls irrigation solely on a Dr
79 field measurement. The ET controller estimates Dr using input data and P and ET estimates to calculate I. Rain sensors use P to bypass I. Thus, there is potential to simulate how these technologies might work based on this relati onship and user input. Currently there is no simple virtual simulation of this concept where Floridians can evaluate different irrigation technologies before investing in modifications of their irrigation system. The objectives were to (1) develop a simple model to simulate the soil water balance based on sitespecific soil characteristics, irrigation schedule, real time weather data, and irrigation device and (2) provide homeowners and landscape professionals with an interactive tool to evaluate and improv e irrigation scheduling. Methods Model Development A simple, onedimension soil water balance model was designed to simulate water movement in the root zone of turfgrass in a typical Florida lawn. We included processes of irrigation (I), rainfall (R), ET, runoff (Q), infiltration (F), and percolation (PERC). The model simulates a daily timestep with all variable units evaluated as depth (i.e., cm). The SWC was calculated based on water gains (i.e., R and I ) and losses (i.e., ET, Q, PERC) to the initial SWC quantity in one soil layer (root zone) as follows: i PERC i Q i a ET i I i R SWC i SWC 0 ( 3 2 ) where i is the current day and ETa (cm) is actual ET. Actual ET was calculated with reference ET (ETo, cm) from the Florida Automated Weather Network (FAWN) and a crop coeff icient (Kc) which is estimated with monthly variance for Florida ( Equation 3 3 Table 3 2). o ET c K a ET ( 3 3 )
80 The initial SWC was assumed to be 75% of FC for the first day of simulation ( Equation 3 4). The incoming water was determined based on real time rainfall (R) from FAWN and irrigation (I). Soil water content was calculated based on the initial SWC, root depth (RD), field capacity (FC) and wilting point (WP) ( Table 3 3), real time weather data (R and ET) from FAWN, I (amount and technolog y), Q, and PERC. Irrigation depth was directly input into the model for each day. RD FC SWC 75 0 0 ( 3 4 ) The SWC was limited between the bounds of FC and WP so that SWC was assumed to equal FC if the value was greater than FC or assumed to equal WP if the value was less than WP. Field capacity and WP values vary with soil texture class ( Table 3 3). The six soil texture classes shown in Table 3 3 were included in the model. Runoff (cm) was calculated by the SCS curve number (CN) method, as used in the CREAMS model (Knisel, 1980). This involved calculating intermediate variables including maximum storage capacity (Smx, cm) and the soil storage capacity (S, cm) ( Equations 3 5 and 3 6). An input value of irrigated area (ha or acre) was used to determine the CN(II) for a particular lot size ( Table 3 4). These CN(II) values were associated with a normal antecedent moisture condition (AMC) or AMC II. An empirical formula was used to convert CN(II) to CN(I) ( Equation 3 6) to obtain Smx ( Equation 3 5). Runof f ( Equation 3 8) occurred if combined R and I were greater than 20% of S ( Equation 3 7); otherwise Q does not occur. Infiltration was calculated as the remaining incoming water (R, I) minus Q ( Equation 3 9).
81 4 25 2540 I CN mx S ( 3 5 ) where CNI is the CN for AMC I (dry). 3 ) ( 0001177 0 2 ) ( 01379 0 ) ( 348 1 91 16 II CN II CN II CN I CN ( 3 6 ) where CNII is the CN for AMC II (normal). RD FC m SWC RD FC mx S S ( 3 7 ) where SWCm, modified SWC, was an intermediate calculation of SWC to ensure that SWC was greater than or equal to WP. S i I i R S i I i R i Q 8 0 2 2 0 ( 3 8 ) i Q i R i I i F ( 3 9 ) Percolation (cm) was calculated as the summation of SWC and infiltration ( F ) exceeding FC ( Equation 3 10). If this summation was less than FC, then no PERC occurred. There was a verification step to ensure that if PERC occurred, SWC would equal FC at the beginning of the next day; if PERC does not occur, SWC was calculated as shown in E quation 3 11. ) ( RD FC i SWC i F i PERC ( 3 10) a ET i F SWC i SWC 0 ( 3 11) User inputs or default values were required to simulate the s oil water balance. Real time rainfall (cm) and ET (cm) from the FAWN database were used based on user input ZIP code values. The user also input RD (cm), soil type, irrigated area (ha), irrigation days, and irrigation depth (cm) per event. Additional inputs were required for
82 two of the simulated irrigation technologies including a rain sensor setting (if using the rain sensor) and a SWS threshold value (if using the SWS technology). Irrigation Technologies The four tec hnolog ies simulated by the model: (1) time based irrigation, (2) time based with RS (3) time based with SWS, and (4 ) ET controller. The primary difference in model simulation among the technologies was the irrigation application depth. This was simulated by requiring specific input parameters for each technology in addition to a scheduled irrigation depth (required for all options). The irrigation depth affects total incoming water. For all systems, real time weather data and the calculations of SWC, Q, F, and PERC remained the same. For tech nology (1), the rain sensor setting was input as this is an adjustable feature. Rain sensor settings are used to set a threshold for bypassing irrigation where a greater setting theoretically corresponds to a greater depth of water needed to bypass irrigat ion. Thus, if rainfall is less than the rain sensor setting for a specific day, I would occur as scheduled on that day. For technology (2), the setting for the SWS was also an adjustable input. This setting is used as a threshold to apply irrigation until an acceptable level of soil water (certain percentage of FC) is reached or to bypass irrigation if the acceptable level of soil water already exists. If SWC was greater than the threshold setting, the only incoming water was R; otherwise incoming water was equal to the summation of R and I. For technology (3), there is no specific setting for the controller; irrigation depth was equal to the minimum of cumulative ET (ETC) or FC on allowed irrigation days. The ETC was determined from the nearest FAWN station using real time weather data and a monthly Kc ( Equation 3 12). For the incoming water
83 calculation, the minimum value of either I or FC was added to R to ensure that incoming water from I did not exceed FC. 1 i ETC i R i a ET i ETC ( 3 12) Model output The mo del output included daily and weekly water volume applied (I + R), water volume not used by the turfgrass (Q + PERC), depths of water leached and lost to runoff, and water stressed days. Volumes of water applied were determined using the irrigation depth applied and the irrigated area input. Water volumes not used by the landscape were calculated by adding PERC and Q depths and multiplying by the irrigated area input. Percent of water leached or runoff was calculated as 100 % R I PERC PERC ( 3 13) 100 % R I Q Q ( 3 14) Water stress was calculated using the management allowable depletion (MAD) concept. The MAD was considered to be 50% of the available water (AW). Available water is the amount of water stored in the root zone that is available to t he plant ( Equation 3 15). According to Allen et al. (1998), a MAD value between 40% and 60% is typical for turfgrass irrigation. RD WP FC AW ) ( ( 3 15) Thus, water stress occurred for a day if the SWC was less than MAD or 50% of AW. Assumptions Sev eral assumptions were made for the model: the initial SWC was assumed to be 75% of FC; monthly crop coefficients were assumed based on Romero and Dukes
84 (2011); the CN approach was used to estimate Q and F (Knisel, 1980); if PERC occurred, SWC was set equal to FC for the beginning of the next day; SWC was not allowed to go above FC or below the WP; R and ET were from the nearest FAWN station; water stress was based on a MAD value equal to 50% of AW; the SWS threshold was assumed to be 70% of FC; the ET contr oller was assumed to have an onsite R measuring device; and FAWN R and ET were used for simulating the ET controller values. Model Validation C omputational accuracy was tested using a Microsoft Excel model and a program written in Java to mathematically s imulate the model. The Java program was used to create the graphical user interface (GUI), hosted by Google App Engine (Jie Fan, FAWN, personal communication, 7 March 2012). The GUI generated output in comma separated files so that equal input values in both Excel and the GUI could be compared for various scenarios. Simulation accuracy was completed using the measured data set from chapter 2. The measured water depths applied and water depths leached were compared to the simulated values of irrigation depths and PERC. Simulated Q values were not compared to measured values, as these values were not measured. For the ET controller, two methods of validation were used: (1) FAWN ETo and Kc values from Romero and Dukes (2011 ) as used for the other technologies and (2) ETo from the onsite ET controller and Kc values from IA (2008). The ET controller, similar to the model, monitors the soil water deficit based on rainfall and ET; however the ET controller and the model used different Kc values. By simulating soil water balance by these two methods, the influence of Kc values on model output was obtained.
85 Graphical User Interface (GUI) The Interactive Irrigation Tool GUI was developed to be simple and engaging for the user (i.e. property owner or irrigation profes sional). The following inputs were required by the user: unit system (English or metric), RD (input value or default was 30 cm), soil type (select one of six types; default was sand), irrigated area (input value or default is 0.10 ha), irrigation system (s elect one of four choices; default was timebased scheduler), ZIP code, irrigation depth (for all technologies except ET controller). Irrigation depth was input by the user or based on a system runtime (min) and an irrigation system selection of microirrig ation (1.27 cm/h), fixed head (3.8 cm/h), gear driven head (1.27 cm/h), or impact head (1.27 cm/h). The soil type was used to determine FC and WP. The ZIP code served two purposes. It linked the model to the closest FAWN station for real time weather data and it was used to implement water restrictions for Miami Dade County. Only Miami Dade County water restrictions were implemented as an example case; current funding limited further statewide implementation of restrictions in the model. Because of the na ture of Miami Dade County restrictions, a house number was requested to determine watering days for these ZIP codes. Watering days vary by odd/even house numbers in Miami Dade County with even numbered houses restricted to Sunday and Thursday and odd numbered houses restricted to Wednesday and Saturday (Miami Dade County, 2012). For ZIP codes outside of Miami Dade County, the user selected irrigation days. All inputs on the GUI were assigned default values except for ZIP code and street number ( Table 3 5). A definition or description was provided for some of the required inputs that were not considered to be common knowledge. These were visible to the user by scrolling over question mark icons located adjacent to the required input.
86 The user was also given the option to subscribe to weekly updates on the virtual lawn via email. Results and Discussion Model Development The computational accuracy of the model was successfully simulated by generating equal output in Microsoft Excel and a Java written program. T he model simulation showed that irrigation and PERC of timebased and time based with a RS was consistently higher than SWS and ET controller based irrigation. This finding was consistent with other reported measured data (Davis et al., 2007; Haley and Duk es, 2007; Shedd et al., 2007; Cardenas Laihalcar et al., 2008; Dukes et al., 2008; Dukes and Haley, 2009; McCready et al., 2009; Cardenas Laihalcar et al., 2010; Cardenas Laihalcar and Dukes, 2010; Davis and Dukes, 2010; McCready and Dukes 2011). The mode l output has been validated with measured data from chapter 2 (2 October 2011 to 16 February 2012). When validating the model with the measured data from 20week long field study (chapter 2), the soil type was sand with corresponding FC and WP values of 0. 08 and 0.02, respectively. The RD was 15 cm, which is typical for southern Miami Dade County. The corresponding monthly Kc was used for estimating ETa. Each plot was approximately 20.9 m2, which was used to select the appropriate CN and calculate Smx. Vari ous irrigation depths were tested in the model, as the runtimes of the time based and rain sensor treatments were changed during the study (i.e. 1.27 cm from 2 October 2011 to 16 December 2011, 1.91 cm from 17 December 2011 to 17 January 2012, and 3.18 cm from 18 January to 16 February 2012 for timebased; 1.27 cm from 2 October 2011 to 22 January 2012 and 1.91 cm from 23 January to 16
87 February 2012 for rain sensor). Figures 3 1 through 3 13 show the differences between observed data from the field study an d the model predicted values for various irrigation depths during the wet and dry seasons. From the daily predicted I and PERC values, weekly summations were also provided by the model output. The weekly summation values of I and PERC were compared to the weekly measured values during the same time span. Table 3 6 measured (x) and predicted (y) values calculated for each of the four irrigation technologies during the wet and dry seasons of the entire field study period (2 October 2011 through 16 February 2012). n n i y x 1 ( 3 16) where n is the total number of values in either the wet or dry season. Predicted and measured R data, from FAWN, were equal for all irrigation technologies, except the ET Controller. The ET controller in the field study was equipped with a tipping bucket to measure rainfall. For each technology, results were presented as the range of percent absolute differences between predicted and measured values during the wet and dry seasons. Measured values of PERC differed greatly between the wet and dry seasons due to the differences in weather conditions, particularly rainfall intensity. Thus, there was a greater absolute percent difference between predicted and measured values during the dry season. The tim e based irrigation results (Figures 3 1 to 3 4 ) showed that the model simulated I accurately and always predicted PERC to be greater than measured. Simulated PERC ranged from 0.37 cm to 1.50 cm (1.21 cm average) greater than the
88 measured values during the wet season (Figure 3 1 ). During the dry season, simulated PERC ranged from 0.69 cm to 3.45 cm (2.03 cm average) greater than measured values ( Figure 3 2 to Figure 3 4 ). For all 20 weeks, PERC was predicted to occur; for seven out of the 20 weeks, PERC was measured as 0.0. Percolation was not measured during two weeks of the dry season, 18 through 31 December 2011, when PERC was predicted as 2.25 cm and 2.32 cm, respectively. The RS setting was set in the model to match the setting used in the field study. The initial setting in the field study was 12 mm (31 August 2011 through 8 November 2011). The RS was replaced and set at 3 mm setting on 9 November 2011 and remained at that setting for the remainder of the study period). For the simulation of time based irrigation with a SWS, a threshold value of 0.70 was used. For the RS based simulation, I was closely predicted (0.3 cm average absolute difference between predicted and measured values), considering both wet and dry seasons, different RS settings (1.2 cm from 2 October to 8 November 2011 and 0.30 cm from 9 November 2011 to 16 February 2012), and different scheduled irrigation depths (1.27 cm from 2 October 2011 to 22 January 2012 and 1.91 cm from 23 January to 16 February 2012) ( Figure 3 5 to Figure 3 7 ). During the wet season, I was usually predic ted as measured, but was under predicted twice by 0.96 cm ( Figure 3 5 ). For one week of the dry season, predicted I was 0.96 cm less than measured I and for another week I was not predicted, when I (1.67 cm) actually occurred. Similar to the timebased irrigation, PERC was predicted to occur during each of the 20 weeks; however, PERC was not measured during 12 of the 20 weeks. Predicted PERC was always greater than
89 measured values, ranging from 0.16 cm to 1.83 cm greater during the wet season and 0.15 cm to 2.49 cm greater during the dry season ( Figure 3 5 to Figure 3 7 ). For the SWS based technology, I was predicted exactly by the model during the wet season and closely during the dry (0.1 cm average absolute difference between predicted and measured values ). For three of the 15 dry season weeks, I predictions were less than measured values only twice (up to 0.91 cm), and never greater than measured values ( Figure 3 9 to Figure 3 11). Predicted PERC during the wet season was equal to or greater than measured, up to 1.84 cm greater ( Figure 3 8 ). During the dry season, predicted PERC was usually greater than and sometimes equal to measured values, ranging from 0.24 cm less than up to 1.60 cm greater than measured values ( Figure 3 9 to Figure 3 11). The values w ere only equal during three weeks when both predicted and measured PERC were 0.0 cm. For the ET controller, two sets of predicted values were generated using two different sets of ET values: (1) FAWN ETo and Kc values from Romero and Dukes (2011) as used for the other technologies and (2) ETo from the on site ET controller and Kc values from IA (2008). The second set of predicted values was closer than the first set to the measured values, particularly the predicted I values. During the wet season, the fir st set of predicted I values was usually higher than measured, up to 0.91 cm greater than (average absolute difference = 0.26 cm). During the dry season, the first set of predicted I values was usually lower than measured values, ranging from 0.56 cm below measured values to 0.19 cm greater than (average absolute difference = 0.25 cm). The second set of predicted I values was closer to measured values, with average absolute differences of 0.13 cm and 0.18 cm, during wet and dry seasons, respectively.
90 Predic ted PERC values improved for the second set of predictions only during the dry season; however during the wet season, the average absolute difference was greater for the second set of predictions ( Table 3 6 ). For the ET controller based irrigation, one ex planation for the differences between measured and predicted values may be the way that ET was calculated by the ET controller in the field ( Figure 3 14). The Rain Bird ESPSMT controller was equipped with a tipping bucket that measured both total amount and intensity of rainfall, or effective rainfall, and a temperature sensor to calculate ET; whereas the model uses the R and ET amounts from FAWN. Although ETo values were relatively close for FAWN and the ET controller, the model used a monthly Kc from Romero and Dukes (2011) to calculate ETa; whereas, the ET controller used default values which were Kc values from IA (2008). Additionally, the ET controller in the field uses an efficiency factor to adjust irrigation runtime, given that the actual application of water by an irrigation system will not equal the net irrigation requirement (IA, 2005b) To improve the accuracy of the model, which usually underestimated irrigation, an irrigation efficiency factor may be added. An efficiency factor of 60% is ass umed in Dukes and Haman (2002b) to determine gross irrigation requirement. In the model, if the ETC was greater than FC, I was equal to FC; when, I should equal, at most, a percent of FC. Additionally, Kc values used in the model should be changed to valu es appropriate for the study area to mimic ET controller calculations. Among the four options, the timebased irrigation resulted in the highest overall absolute difference (1.62 cm) between predicted and measured PERC values; whereas, the first method of predicting PERC values for the ET controller resulted in the least
91 overall absolute difference (0.37 cm). This is likely due to the correlation between a pplied water (I and R) and PERC For the ET based irrigation, PERC was usually equal to 0, unless it rained. Since most measured data were during the dry season, both measured and predicted PERC for the ET based irrigation usually equaled 0. There may have been more variation between measured and predicted values if there were more measured data during the wet season. During the dry season, timebased irrigation usually produced measurable PERC although these values were always lower than predicted; therefore, the overall absolute difference for the time based was higher than the other technology based irrig ation options. For all technologies, model predicted PERC was usually greater than measured data. This greater model estimation of PERC may be due to preferential flow outside of the catchment area of the lysimeters (as described in chapter 2) and the inf low rate limitations of the lysimeters; therefore, the true amount of PERC water was not captured in the field study. Future models and lysimeter designs should include ways to account for preferential flow in both dry and wet conditions. Another source of error in the model may have been the Q calculation. There is some evidence that the original SCS CN method overestimates Q (Jain et al., 2005; Carlesso et al. 2011; Xiao et al., 2011). Ponce and Hawkins (1996) attribute CN variability to spatial and temporal variability of R, measured data quality, antecedent R and soil moisture content variability. Jain et al. (2005) evaluated the effectiveness of the original CN method, along with version modified by Mishra and Singh (2002) and other variants. They concl uded that the original CN method is better suited for high runoff producing agricultural watersheds, rather than pasture/rangeland and sandy soils.
92 Mishra and Singh (2002) modified the original CN method to more accurately account for antecedent soil moist ure (M) (Equations 3 17 through 3 20). This modified version considered the 5day antecedent rainfall, P5, and assumed the watershed to be dry for five days prior to the rain event. I S P I S I S P M 8 0 5 ) 2 0 5 ( ( 3 17) where SI was the potential maximum retention for dry antecedent m oisture conditions. Given that they propose SI = S + M, S P S S M 5 4 2 64 0 2 1 5 0 ( 3 18) which was used to modify the original equation. M S M F a I P Q ( 3 19) which resulted in S M a I P M a I P a I P Q ) )( ( ( 3 20) Another method to calculate Q was developed by the Irrigation Association: IR PR ASA Rt 60 (max) ( 3 2 1 ) where Rt(max) is the maximum runtime allowable before Q occurs, ASA is allowable surface accumulation, PR is precipitation rate, IR is the soil intake rate (IA, 2008). Accumulated time in excess of Rt(max) is converted to water depth and considered as Q ( IA, 2008). This method is not as reliable as the SCS CN method or its variations. When tested for a typical scenario, the Rt(max) was negative, given that IR was greater than PR.
93 For the Q methods proposed by Mishra and Singh (2002) and Michel et al. (2005), antecedent soil moisture conditions are considered; however, the initial soil moisture content was assumed as 75% of FC for this simple model. The Mishra and Singh (MS) method was tested using FAWN data and compared to the Q and PERC results from the or iginal model (CREAMS version) and the original SCS CN method for time based irrigation. Results showed that R is the greatest influencing factor on Q and PERC, over I and P5 for all three methods. For weeks with higher R, the CREAMS CN method yielded the l owest PERC and either the highest or second highest Q, after the MS method. When R was equal to 0, all methods yielded equal amounts of Q (0 cm) and PERC (higher with higher I). The CREAMS CN method usually yielded Q values higher than or equal to and PERC values less than or equal to the values of the other methods. Runoff contributes to total water losses from the soil (Q, PERC, ETa); however, PERC likely accounts for the majority of losses from turfgrass, given the high infiltration rates of sandy soils and low slopes of most land in Florida. This was demonstrated by all variations of the CN method, but particularly by the original SCS CN method which always yielded PERC values higher than Q. The variance among the Q and PERC values of the three different methods was relatively low. Therefore, the original SCS CN method (most simple method) is sufficient for this simple model, which is not concerned with high temporal resolution, but being used to attain weekly estimates of Q and PERC. The source of R data (Homestead FAWN weather station) was the same for the measured and predicted R values due to the close proximity of the field study site used for the model validation to the weather station (~0.20 km). However, for other users the
94 accuracy of R data may be another source of error due to the spatial variability of rainfall in Florida and the limited number of FAWN weather stations (36 stations across Florida). This value may be improved by obtaining rainfall data from other sources across Florida or allowing the input of user obtained rainfall data from an onsite rainfall measuring device. Graphical User Interface The GUI was successfully launched online 1 March 2012. The tool can be found at http://fawn.ifas.ufl.edu/tools/interactive_irrigation_tool/. Because the tool was built on Google technology, an acti ve Google account is required. The user was prompted for a Google username and password after selecting the link. Once signed in, the user was directed to the GUI. Figures 3 15 to 3 24 show screenshots of the various entries required in the GUI including units, soil characteristics, irrigation technology used, irrigation schedule, and irrigation amount (unless ET controller is selected). Weekly emails were sent to users selecting this option ( Figure 3 25). The boxes with descriptions or definitions of certain terms appear only when the user scrolls over the question mark icon. Some entry boxes appeared depending on previous input values or technology selection. For all irrigation technologies options except ET controller, irrigation amount applied per event was entered as either a depth per event or based on the selection of irrigation system type (i.e. microirrigation [1.27 cm/h], fixed head [3.8 cm/h], gear driven head [1.27 cm/h], or impact head [1.27 cm/h]) and a system runtime entered. If the Timebased plus rain sensor was selected, the RS setting was also required. If the Timebased plus soil moisture sensor was selected, the threshold value was also required.
95 After the user submitted all of the required values, an email was sent to the user with information including the amount (volume) and percentage by which the irrigation system overwatered and the number of days for which the turfgrass did not receive enough water ( Figure 3 25) Based on these values, the lawn was ranked on a scale of one to five, one representing an efficient irrigation system and five inefficient. The scale also corresponded to a color of the virtual lawn (dark green [rating equal to one] to light brown [rating equal to five]) The email also displayed the distance between the users ZIP code and the FAWN station. Defaults All GUI inputs have default values except for ZIP code and house number. These were based upon literature values, typical Florida conditions and UF IFAS reco mmendations. The default unit system is English. The default root zone depth was 30 cm, based on studies by Doss et al. (1960) who found 76% of the root zone of five warm season forage species (including bahiagrass and bermudagrasses) to be in the upper 30 cm of soil and Peacock and Dudeck (1985) who found 81% of the root mass of St. Augustinegrass to be in the upper 30 cm of soil. For timebased irrigation, the default irrigation depth is 1.27 cm, applied on Sunday and Thursday, based on Miami Dade County watering restrictions. For the RS the default rainfall depth setting chosen was 1.27 cm. For the SWS, the default threshold setting is 70% of FC. Both of these default values are based on the UF IFAS recommendation to apply 1.27 to 1.91 cm (0.5 to 0.75 in) of water or 50% to 75% of FC per irrigation event when soil is dry, given that Florida soils typically hold 2.54 cm (1 in) of water in the top 30 cm (12 in) of soil (Trenholm and Unruh, 2005).
96 Table 3 1 Example input for ET Controller (Rain Bird, Tucs on, AZ). Parameter Input Z IP code 33031 Soil type Loamy sand Sprinkler type Rotary nozzle Slope 0 2% Plant type Grass lawn warm season Shade factor Full sun Plant maturity Established Table 3 2 Monthly crop coefficient (Kc) values for determining ETa (Romero and Dukes, 2011). Month K c Jan 0.71 Feb 0.79 Mar 0.78 Apr 0.86 May 0.99 Jun 0.86 Jul 0.86 Aug 0.90 Sep 0.87 Oct 0.86 Nov Dec 0.84 0.71 Table 3 3 Field capacities and wilting points for various soil types (aRawls et al., 1982; bZotarelli et al., 2010). Soil type Porosity a (m3/m3) Bulk Density* (g/cm 3 ) FCb (cm/cm) WPb (cm/cm) Sand 0.44 1.48 0.08 0.02 Sandy loam 0.45 1.45 0.16 0.06 Loam 0.46 1.43 0.26 0.08 Silt loam 0.50 1.32 0.31 0.10 Clay loam 0.46 1.43 0.34 0.14 Clay 0.4 8 1.37 0.37 0.16 b = (1 n)*2.65, where 2.65 is normal particle density (g/cm3). Table 34. CN(II) values used by the model (Chow et al., 1988) Lot size Lot size (acre) Curve numbers for hydrologic soil group (ha) A B C D 0. 0 5 or less 1/8 or less 77 85 90 92
97 Table 34. Continued. Lot size Lot size (acre) Curve numbers for hydrologic soil group (ha) A B C D 0.10 61 75 83 87 0.13 0.33 57 72 81 86 0.20 54 70 80 85 0.40 1 51 68 79 84 0.80 2 46 65 77 82 Table 3 5 Default values for the GUI and model. Parameter description Value Reference Root depth 30 cm Doss et al., 1960; Peacock and Dude ck, 1985 Soil type Sand Trenholm and Unruh, 2005 Field capacity 0.08 cm/cm Trenholm and Unruh, 2005 Wilting point 0.02 cm/cm Trenholm and Unruh, 2005 Irrigated area 0.10 ha (0.25 ac) Chow et al., 1988 Irrigation schedule Sun/Thur Miami Dade County, 2 012 Irrigation application 1.27 cm/event Trenholm and Unruh, 2005 Initial soil water content 75% of FC Rain sensor setting 12 mm Trenholm and Unruh, 2005 SWS threshold 70% FC Trenholm and Unruh, 2005 Table 3 6 Average absolute differences over ir rigation events between measured and model predicted values of irrigation and percolation. Technology Average Absolute I Difference (cm) Average Absolute PERC Difference (cm) Wet Season Dry Season Wet Season Dry Season Time based (TB) 0.0 0.0 1.21 2.03 TB + RS 0.39 0.18 0.99 1.18 TB + SWS 0.0 0.10 0.52 0.55 ET controller 1 0.26 0.25 0.60 0.14 ET controller 2 0.13 0.18 0.86 0.13 Overall average 0.16 0.14 0.84 0.81 1FAWN ETo and Kc values from Romero and Dukes (2011) 2ETo from the onsite ET controller and Kc values from IA (2008)
98 Figure 3 1 Average measured data versus model predicted values for the timebased irrigation during the wet season (2 October through 5 November 2011) for application of approximately 1.27 cm per irrigation event. Figur e 3 2 Average measured data versus model predicted values for the timebased irrigation during the dry season (6 November through 17 December 2011) for application of approximately 1.27 cm per irrigation event.
99 Figure 3 3 Average measured data versus model predicted values for the time based irrigation during the dry season (18 December 2011 through 21 January 2012) for application of approximately 1.91 cm per irrigation event (changed to 3.18 cm on 18 January). Figure 3 4 Average measured data versus model predicted values for the timebased irrigation during the dry season (21 January through 18 February 2012) for application of approximately 3.18 cm per irrigation event.
100 Figure 3 5 Average measured data versus model predicted values for the rain sensor based irrigation during the wet season (2 October through 5 November 2012) for scheduled application of approximately 1.27 cm per irrigation event and rain sensor setting of 12 mm.
101 Figure 3 6 Average measured data versus model predicted val ues for the rain sensor based irrigation during the dry season (13 November 2011 through 21 January 2012) for scheduled application of approximately 1.27 cm per irrigation event and rain sensor setting of 3 mm. Figure 3 7 Average measured data versus model predicted values for the rain sensor based irrigation during the dry season (22 January through 18 February 2012) for scheduled application of approximately 1.91 cm per irrigation event and setting of 3 m m.
102 Figure 3 8 Average measured data versus model predicted values for the soil water sensor based irrigation during the wet season (2 October through 5 November 2012) for scheduled application of approximately 1.27 cm per irrigation event and threshold setting of 0.70. Figure 3 9 Average meas ured data versus model predicted values for the soil water sensor based irrigation during the dry season (6 November through 17 December 2012) for scheduled application of approximately 1.27 cm per irrigation event and threshold setting of 0.70.
103 Figure 3 10. Average measured data versus model predicted values for the soil water sensor based irrigation during the dry season (18 December 2011 through 21 January 2012) for scheduled application of approximately 1.27 cm per irrigation event and threshold setting of 0.70. Figure 3 1 1 Average measured data versus model predicted values for the soil water sensor based irrigation during the dry season (22 January through 18 February 2012) for scheduled application of approximately 1.27 cm per irrigation event and threshold setting of 0.70.
104 Figure 3 12. Average measured data versus two sets of model predicted values for the ET controller based irrigation during the wet season (2 October through 5 November 2012). Predicted1 was using FAWN ETo and Kc values from Romero and Dukes (2011) and Predicted2 was using ETo from the on site ET controller and Kc values from IA (2008).
105 Figure 3 13. Average measured data versus two sets of model predicted values for the ET controller based irrigation during the dry s eason (6 November through 18 February 2012). Predicted1 was using FAWN ETo and Kc values from Romero and Dukes (2011) and Predicted2 was using ETo from the on site ET controller and Kc values from IA (2008).
106 Figure 3 14. Comparison of ETa of the ET c ontroller used in the field study and ETa as predicted by the model using FAWN data and crop coefficient from Romero and Dukes (2011) Figure 3 15. Screenshot of rooting depth entry in metric units. Figure 3 16. Screenshot of soil type selection and description.
107 Figure 3 17. Screenshot of irrigated area entry (English units). Figure 3 18. Screenshot of possible irrigation system selection and description of each technology. Figure 3 19. Screenshot of selection of timebased irrigation s ystem with a rain sensor with default rain sensor setting (metric units). Figure 3 2 0 Screenshot of selection of timebased irrigation system with a soil water sensor with default threshold setting.
108 Figure 3 2 1 Screenshot of automatically generat ed irrigation days, based on Miami Dade County zip code and street number. Figure 3 2 2 Screenshot of irrigation day(s) selection for zip codes entered outside of Miami Dade County. Figure 3 2 3 Screenshot of first entry method for applied irrigation amount (metric units).
109 Figure 3 2 4 Screenshot of second entry method for applied irrigation amount: selection of irrigation system type and runtime entry.
110 Figure 3 2 5 Screenshot of weekly email report.
111 CHAPTER 4 SUMMARY AND CONCLUSI ON The goa l of this research was to (1) compare irrigation quantity applied and nitrogen (NO3N and NH4N ) leached from three technology based irrigation treatments (i.e. time based with a rain sensor [RS] soil water sensor [SWS] or evapotranspiration [ ET ] control ler) to a traditional time based irrigation and (2) to develop a simple model that predicts water losses from residential turfgrass based on the irrigation technology used. The model led to the development of the Interactive Irrigation Tool on the Florida Automated Weather Network (FAWN) website which enables users to evaluate various irrigation schemes with respect to water stress and losses. Objective 1 Sixteen turfgrass plots were fertilized with commercial slow release nitrogen fertilizer and irrigated by either timeb ased or one of three irrigation systems: time based with a RS time based with a SWS, or ET controller over nearly a five month period (31 August 2011 to 16 February 2012) The time based irrigation was always greater and significantly diff erent (P<0.05) from each of the technology based systems. Of the three technology based irrigation systems, the ET based irrigation resulted in the greatest amount of water savings ( up to 83%), compared to timebased irrigation, applied at three different depths (1.3, 1.9, and 3.2 cm) The ET based irrigation applied the least amount of water (17.3 cm) and leached the least (14.1 cm), while still maintaining acceptable turfgrass quality. Leachate quantity measured and leachate water quality was analyzed for NO3N and NH4N concentrations among the different treatments. Due to preferential flow, limited inflow rate of the lysimeters, and inconsistencies in design among the lysimeters, leachate volumes were lower than
112 expected (given the amount of water applied). The NO3N and NH4N concentrations and loads were very low compared to other studies, with an overall average among all treatments of 2.44E 02 kg NO3N ha1 and 5.39E 03 kg NH4N ha1 over the study period. Objective 2 The model successfully simulated irrigation and percolation on a daily time step for time based irrigation and the three technology based irrigation systems. Using the 20week field study data, the range of differences between measured and predicted values were calculated for all irrigat ion options. Cons idering all irrigation systems, differences were greater for percolation values, with absolute average differences ranging from 0.37 cm to 1.62 cm, compared to differ ences of 0.0 cm to 0.28 cm for irrigation. Irrigation was predicted exact ly as measured for the timebased irrigation and closely for the other technologies. T he SWS based irrigation yielded the least difference between measured and predicted irrigation values which suggests the model sufficiently mimicked soil water status and the soil water sensors operated as expected. The RS based irrigation yielded the greatest difference between predict ed and measured irrigation, which is expected as the model only considered the amount of rainfall on a day without any simulation of the actual RS dry out time. Thus, the RS system was not simulated as accurately as the other technologies. The ET based irrigation resulted in the least difference between percolation values; whereas, the timebased irrigation resulted in the greatest differenc e between percolation values. Smart irrigation technologies provide more precise amounts of required irrigation compared to timebased irrigation which meet the plant water needs, without over or under watering, by considering weather and/or soil conditi ons including rainfall, ET, soil
113 water content, and field capacity. Results indicate that ET based irrigation can save up to 70% compared to the applying the UF IFAS irrigation recommendation of 1 cm to 2 cm per irrigation event, or up to 83% if irrigation is applied beyond the recommended amount at 3 cm. The SWS based irrigation saved up to 62% compared to the UF IFAS recommended rates and 83% compared to the elevated rate. The rain sensor based irrigation resulted in water savings of 16% 53% during the di fferent application rates. Ways to reduce irrigation water applied is a major component of overall water conservation efforts, given that irrigation can account for up to 70% of outdoor water use during summer months (IA, 2011) Except for when irrigation was applied beyond the recommended rate, percolation was primarily driven by heavy rainfall rather than irrigation. Given that the NO3N and NH4N concentrations were well below the drinking water standard, nitrogen leaching was not a significant pathway of nitrogen loss from turfgrass in this study. Although there are minimal practices that can reduce leaching during the wet season, additional field data should be collected during this season using UF IFAS fertilizer recommendations and the technology based irrigation systems in order to better understand the system and how to optimally manage turfgrass under these seasonal conditions. The model and interactive tool provide a means for property managers to test the various irrigation schemes before investing in modifications to their irrigation system. Additional versions of the model and interactive tool may include more input parameters, higher spatial resolution rainfall data, and more output information as a learning tool for students.
114 APPENDIX A RESULT S OF MODELING VALIDA TION Table A 1 Measured values, predicted values, and di fferences for the timebased and rain sensor based irrigation during the wet season. Time Based Rain Sensor Measured Predicted Difference Measured Predicted Diff erence Week cm cm cm cm cm cm 10/2/11 10/8/11 1 Irr 2.2 2.2 0.0 2.0 2.0 0.0 PERC 0.0 0.4 0.4 0.0 0.2 0.2 Q 0.0 0.0 10/9/11 10/15/11 2 Irr 2.2 2.2 0.0 2.0 1.0 1.0 PERC 2.7 3.7 1.0 2.6 3.1 0.5 Q 12.4 12.3 10/16/11 1 0/22/11 3 Irr 2.1 2.1 0.0 1.0 0.0 1.0 PERC 3.1 6.0 2.9 3.2 5.1 1.8 Q 5.9 4.9 10/23/11 10/29/11 4 Irr 2.1 2.1 0.0 2.9 2.9 0.0 PERC 0.0 0.3 0.3 0.0 1.1 1.1 Q 0.0 0.0 10/30/11 11/5/11 5 Irr 2.2 2.2 0.0 1.0 1.0 0.0 PERC 2.3 3.8 1.5 1.8 3.2 1.4 Q 0.6 0.1
115 Table A 2 Measured values, predicted values, and differences for the soil water sens or based irrigation and evapotranspirationbased irrigation during the wet season. Soil Water Sensor Evapot ranspiration Controller Measured Predicted Difference Measured Predicted 1 Predicted 2 Difference 1 Difference 2 Week cm cm cm cm cm cm cm cm 10/2/11 10/8/11 1 Irr 1.1 1.1 0.0 1.2 2.1 1.7 0.9 0.5 PERC 0.0 0.2 0.2 0.0 0.3 0.1 0.3 0.1 Q 0.0 0.0 0.0 10/9/11 10/15/11 2 Irr 1.0 1.0 0.0 0.7 0.8 0.6 0.1 0.1 PERC 2.6 2.6 0.1 2.5 2.4 2.9 0.1 0.4 Q 12.3 12.3 11.6 10/16/11 10/22/11 3 Irr 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 PERC 3.2 5.1 1.8 3.3 5.1 6.2 1.8 2.9 Q 4.9 4 .9 3.5 10/23/11 10/29/11 4 Irr 0.0 0.0 0.0 1.3 1.5 1.4 0.3 0.1 PERC 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 Q 0.0 0.0 0.0 10/30/11 11/5/11 5 Irr 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 PERC 2.0 2.5 0.5 1.6 2.5 2.6 0.9 0.9 Q 0.1 0.1 0.1 1FAWN ETo and Kc values from Romero and Dukes (2011) 2ETo from the onsite ET controller and Kc values from IA (2008)
116 Table A 3 Measured values, predicted values, and differences for the timebased and rain sensor base d irrigation during the dry season. Time Based Rain Sensor Measured Predicted Difference Measured Predicted Difference Week cm cm cm cm cm cm 11/6/11 11/12/11 6 Irr 2.17 2.17 0.00 1.98 1.98 0.00 PERC 0 0.86 0.84 0 0.67 0.66 Q 0.00 0.00 11/13/11 11/19/11 7 Irr 2.11 2.11 0.00 1.92 0.96 0.96 PERC 0 1.04 0.97 0 0.15 0.15 Q 0.00 0.00 11/20/11 11/26/11 8 Irr 2.11 2.11 0.00 1.93 1.93 0.00 PERC 1.51 3.36 0.85 1.44 3.19 0.79 Q 0.68 0.68 11/27/11 12/3/11 9 Irr 2.08 2.08 0 .00 1.92 1.92 0.00 PERC 0 0.69 0.66 0 0.53 0.51 Q 0.00 0.00 12/4/11 12/10/11 10 Irr 2.04 2.04 0.00 1.91 1.91 0.00 PERC 0 1.16 1.06 0 1.03 0.96 Q 0.00 0.00 12/11/11 12/17/11 11 Irr 2.35 2.35 0.00 1.23 1.23 0.00 PERC 0 1. 94 1.54 0 0.87 0.67 Q 0.06 0.01 12/18/11 12/24/11 12 Irr 3.37 3.37 0.00 2.11 2.11 0.00 PERC 2.25 1.01 Q 0.02 0.00 12/25/11 12/31/11 13 Irr 3.47 3.47 0.00 2.18 2.18 0.00 PERC 2.32 1.05
117 Table A 3. Continued. Time Based Rain Sensor Measured Predicted Difference Measured Predicted Difference Week cm cm cm cm cm cm 12/25/11 12/31/11 13 Q 0.02 0.00 1/1/12 1/7/12 14 Irr 3.48 3.48 0.00 2.20 2.20 0.00 PERC 0.47 2.45 1.37 0.00 1.26 1.08 Q 0.10 0.00 1/8/12 1/14/12 15 Irr 3.47 3.47 0.00 2.19 2.19 0.00 PERC 0.07 2.42 1.82 0.00 1.17 1.05 Q 0.03 0.00 1/15/12 1/21/12 16 Irr 4.56 4.56 0.00 2.20 2.20 0.00 PERC 0.32 3.24 1.99 0.00 1.15 1.04 Q 0.27 0.00 1/22/12 1 /28/12 17 Irr 5.67 5.67 0.00 2.75 2.75 0.00 PERC 1.52 4.26 1.52 0.00 1.56 1.33 Q 0.21 0.00 1/29/12 2/4/12 18 Irr 5.85 5.85 0.00 3.35 3.35 0.00 PERC 1.17 4.35 2.01 0.00 1.98 1.73 Q 0.13 0.00 2/5/12 2/11/12 19 Irr 5.69 5.69 0 .00 1.67 0.00 1.67 PERC 3.26 6.72 0.88 1.98 2.72 0.14 Q 2.31 0.62 2/12/12 2/18/12 20 Irr 5.74 5.74 0.00 3.31 3.31 0.00 PERC 2.82 5.02 0.55 0.93 3.42 1.53 Q 1.38 0.55
118 Table A 4 Measured values, predicted values, and diff erences for the soil water sensor based irrigation and evapotranspirationbased irrigation during the dry season. Soil Water Sensor Evapotranspiration Controller Measured Predicted Difference Measured Predicted 1 Predicted 2 Difference 1 Dif ference 2 Week cm cm cm cm cm cm cm cm 11/6/11 11/12/11 6 Irr 2.2 2.20 0.00 1.31 1.01 0.85 1.31 0.46 PERC 0 0.89 0.87 0 0.00 0.00 0.00 0.00 Q 0.00 0.00 0.00 11/13/11 11/19/11 7 Irr 2.14 2.14 0.00 1.16 0.77 0.66 0.38 0.50 PER C 0 1.07 1.00 0 0.00 0.00 0.00 0.00 Q 0.00 0.00 0.00 11/20/11 11/26/11 8 Irr 0 0.00 0.00 0.85 0.77 0.69 0.08 0.16 PERC 0.8 2.02 0.38 2.05 2.14 2.06 0.75 0.83 Q 0.68 0.68 0.68 11/27/11 12/3/11 9 Irr 1.61 1.61 0.00 0. 9 1.09 1.02 0.19 0.12 PERC 0 0.22 0.22 0 0.00 0.00 0.00 0.00 Q 0.00 0.00 0.00 12/4/11 12/10/11 10 Irr 1.06 0.55 0.51 0.79 0.58 0.61 0.21 0.18 PERC 0 0.12 0.12 0 0.00 0.00 0.00 0.00 Q 0.00 0.00 0.00 12/11/11 12/17/ 11 11 Irr 0 0.00 0.00 0 0.00 0.00 0.00 0.00 PERC 0 0.00 0.00 0 0.00 0.00 0.00 0.00 Q 0.00 0.00 0.00 12/18/11 12/24/11 12 Irr 1.95 1.05 0.91 1.35 0.79 0.90 0.56 0.45 PERC 0.42 0.00 0.00 Q 0.00 0.00 0.00 12/2 5/11 12/31/11 13 Irr 1.72 1.72 0.00 1.03 0.83 0.93 0.20 0.10 PERC 0.59 0.00 0.00 Q 0.00 0.00 0.00
119 Table A 4. Continued. Soil Water Sensor Evapotranspiration Controller Measured Predicted Difference Measured Predicted 1 Predicted 2 Difference 1 Difference 2 Week cm cm cm cm cm cm cm cm 1/1/12 1/7/12 14 Irr 1.15 1.15 0.00 0.95 0.63 0.84 0.32 0.11 PERC 0.00 0.22 0.22 0.00 0.00 0.00 0.00 0.00 Q 0.00 0.00 0.00 1/8/12 1/14/12 15 Irr 0.82 0.82 0.00 1.06 0 .72 0.95 0.34 0.11 PERC 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Q 0.00 0.00 0.00 1/15/12 1/21/12 16 Irr 0.83 0.83 0.00 0.99 0.75 0.90 0.24 0.09 PERC 0.24 0.00 0.24 0.00 0.00 0.00 0.00 0.00 Q 0.00 0.00 0.00 1/22/12 1/28/12 17 Irr 2.03 2.03 0.00 1.16 0.89 0.97 0.27 0.20 PERC 0.00 0.84 0.78 0.00 0.00 0.00 0.00 0.00 Q 0.00 0.00 0.00 1/29/12 2/4/12 18 Irr 1.17 1.17 0.00 1.14 1.07 1.08 0.08 0.07 PERC 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 Q 0.00 0.00 0.00 2/5/12 2/11/12 19 Irr 0.86 0.86 0.00 0.64 0.15 0.22 0.49 0.42 PERC 1.33 2.93 0.46 2.06 2.82 2.85 0.24 0.20 Q 1.27 1.27 1.12 2/12/12 2/18/12 20 Irr 0.00 0.00 0.00 0.82 0.96 0.91 0.14 0.09 PERC 0.73 1.46 0.29 0.53 1.52 1.47 0.56 0.51 Q 0.17 0.17 0.16 1FAWN ETo and Kc values from Romero and Dukes (2011) 2ETo from the onsite ET controller and Kc values from IA (2008)
120 APPENDIX B MODEL MANUAL FAWN Irrigation Tool Model Manual The concept behi nd this tool is to provide users with a virtual lawn environment where they can test different irrigation strategies. With minimal user input, the tool provides an assessment of potential water stressed days, water volumes applied, tips on improving irriga tion, percent of nitrogen and water leached, and a visual graphic of the quality of the turf. The tool provides users with the ability to test smart technology and estimate the potential savings with their implementation. We believe this tool will be useful for home owners as well as irrigation contractors who are trying to optimize automated irrigation systems and save money. C onversions 1 acre = 0.404685642 ha 1 kg = 2.20462262 lbs 1 square foot = 2.29568411 105 acre User inputs Irrigated turf area (area) [ac or ha] Technology used for irrigation Rain sensor (RSS) o Setting provided by user; default is 1.27 cm ET controller Soil moisture sensor (aka soil water sensor) o User also can provide a threshold setting or a default of 0.7 will be used. Time based schedule (default) Zip code Irrigation schedule Option 1: if zip code is in Miami Dade County the street number will be requested (odd numbers irrigate on Wed & Sat and even numbers irrigate on Sun & Thurs; irrigation amount 1.27 cm); if zip code is not in Miami Dade County a default will be offered for option 1 of irrigating Sun & Thur. Option 2: input days per week, input irrigation amount as cm or in of water Option 3: input days per week, input timer setting (how many minutes), select irrigation t ype from pictures (the model will then calculate the cm of water) Soil characteristics
121 Rooting depth (RD) (cm or in) Soil type (see Table B 1) Fertilization Composition (N P K) Amount applied (mass/area) Date applied User also has the option to select IFA S fertilizer recommendations Table B 1. Field capacities and wilting points for various soil types (aRawls et al., 1982; bZotarelli et al., 2010). Soil type Porosity a (m3/ m3) Bulk Density* (g/cm 3 ) FCb (cm/cm) WPb (cm/cm) Hydrologic soil group S and 0.44 1.48 0.08 0.02 A S andy loam 0.45 1.45 0.16 0.06 A L oam 0.46 1.43 0.26 0.08 B S ilt loam 0.50 1.32 0.31 0.10 B C lay loam 0.46 1.43 0.34 0.14 C C lay 0.48 1.37 0.37 0.16 D FC is field capacity ; WP is wilting point. *Calculated using porosity values and equationb = (1 n)*2.65, where 2.65 is normal particle density (g/cm3). Hydrology This section describes the hydrology module of the model. This component includes a simple water balance that is calculated on a daily timestep using Florida Automated Weather Network (FAWN) data and user inputs that describe the system. The model will initiate based on user or default inputs so that day 1 has an irrigation event. As an example, this means that if the user inputs irrigation to start on the next Wednesday the model will not start until that day. Initial values and/or defaults The initial soil water content (SWC) [cm] is: RD FC SWCo* 75 0 (1) For the beginning of the first day, SWCi=SWCo The first irrigation will occur during day one. Evapot ranspiration
122 This section describes how evapotranspiration (ET) is calculated by the tool using crop coefficients and FAWN data. Crop coefficients (KC) for calculating actual evapotranspiration (ETa) for each month were derived from field experiments on t urfgrass (Romero and Dukes, 2011 ; Table B 2). Table B 2. Crop coefficient values for determining ETa K c values Month Panhandle/North Florida Central/Southwest Florida ** South Florida *** Jan Feb Mar Apr May Jun Jul Aug Sep Oct Nov Dec 0.35 0.35 0.55 0.8 0 0.90 0.75 0.70 0.70 0.75 0.70 0.60 0.45 0.45 0.45 0.65 0.80 0.90 0.75 0.70 0.70 0.75 0.70 0.60 0.45 0.71 0.79 0.78 0.86 0.99 0.86 0.86 0.90 0.87 0.86 0.84 0.71 Reference Jia et al., 2009 Davis and Dukes, 2010 Jia et al., 2009 Mobile, Tallahassee, Gainesville, and Jacksonville. ** Daytona, Orlando, and Tampa (including all locations in Southwest Florida). *** West Palm, Fort Myers, Miami, and Key West. ETa will be determined by using daily FAWN tabulated reference ET (ETo) and the crop coefficie nt (Kc) from Table 1 as: c oK ET ETa (2) Water balance calculation The SWC for each irrigation treatment is determined using the following processes. A. Time Based Irrigation Step 1: D etermine incoming water for day Irrigation is determined f rom user inputs or default values. The default is that irrigation (I) occurs Sunday and Thursday, each event being 1.27 cm. Rainfall (R) is from FAWN. i i iI R WB (3) WB, R, I are all in cm units. WB represents water inputs.
123 Step 2: Calcul ate runoff (Q) A modified version of SWC (i.e., SWCm) is used in the Q calculation due to double counting of ET otherwise. If SWCiETai > WP*RD, then SWCm = SWCiETai Else SWCm = WP*RD The curve number (CN) method is used as described by Chow et al. ( 1988) Lot size is from user input. Table B 3. CN II values used by the model (Chow et al ., 1988) Lot size (acres) Curve numbers II for hydrologic soil group A B C D 1/8 or less 77 85 90 92 61 75 83 87 0.33 57 72 81 86 54 70 80 85 1 51 68 79 84 2 46 65 77 82 4 25 2540 I CN mx S (4) where CNI is the CN for AMC I (dry). 3 ) ( 0001177 0 2 ) ( 01379 0 ) ( 348 1 91 16 II CN II CN II CN I CN (5) where CNII is the CN for AMC II (normal). RD FC m SWC RD FC mx S S (6) where SWCm, modified SWC, was an intermediate calculation of SWC to ensure that SWC was greater than or equal to WP S i I i R S i I i R i Q 8 0 2 2 0 (7)
124 where Smx [ cm ] is the maximum storage capacity of the soil and S [ cm ] is the storage capacity of the soil. The hydrologic soil groups correspond to different soil types as noted in Table 1. A check is performed for Q to det ermine the validity of the runoff equation used. Check for Q: If 0< 0.2*S and 0.2*S 0 i i iQ WB F (10) Check if (Fi +SWCm) < (FC*RD) PERCi = 0 Else PERCi = Fi + SWC m (FC*RD) Else 0 iF PERCi = 0 (11) Step 4: Calculate SWC for i+1 Check the SWC. If percolation is greater than 0, the SWC for i+1 is reset to field capacity or SWCi+1 = RD*FC (12) If percolation (PERCi) < 0, then SWCi+1 = SWCi + Fi ETai (1 3 )
125 B. Irrigation Technology Option 1: Rain Sensor A rainfall setting (RSS) of 1.27 cm was selected for the rain sensors as a default. The user has the option to input an alternative value. Step 1: Determine incoming water f or day If Ri > RSSi WBi = Ri (1 4 ) Else WBi = Ri + Ii ( 15) Steps 24 same for time based C. Irrigation Technology Option 2: Soil Water Sensor Step 1: Determine incoming water f or day Soil water sensors allow for threshold settings. The default setting for this model is 0.7 of the field capacity. Threshold setting (TH) = 0.70 (or value entered by user). Step 1 only occurs on irrigation days. If SWCi > FC*RD*TH Ii = 0 (16) Else Ii = schedule (17) This will depend on the schedule provided for the sensor as user input or a default will be 2 times a week for 1.27 cm. Steps 24 same as for time based. D. Irrigation Technology Option 3: ET Controller Step 1: Determine incoming water for day. Calculate cumulative ET (ETC)
126 If ETCi 1 + E T ai Ri < 0 ETCi = 0 (18) Else ETCi = E T ai Ri+ ETCi 1 (19) Irrigation Check (IC) (on/off) If IC scheduled (yes or no) I = Min (ETCi or FC*RD) (20 ) ETCi = 0 ( 21) Else (no irrigation) I = 0 ETCi = ETai Ri+ ETCi 1 ( 22) Steps 24 same as for timebased. Model Assumptions Water balance assumptions Crop coefficients were assumed based on Romero and Dukes (2009). The Curve Number approach was used to estimate runoff (Q) and infiltration (F) ( Chow et al. 1988). If percolation (PERC) occurred, SWC was set to field capacity (FC) for the beginning of the next day. SWC was not allowed to go below the wilting point (WP). Rainfall (R), evapotranspiration (ET), and temperature (T) are from the nearest FAWN station. Water stress (WS) is based on the management allowable depletion (MAD) value of 50% and the available wat er (AW) The rainfall sensor is assumed to be set at 1.27 cm. The soil water sensor threshold is considered to be 70% of FC; i.e., if SWC is above 0.7*FC irrigation is bypassed. ET controller is assumed to have an onsite rain sensor; FAWN ET was used in sim ulating the ET controller values. Table B 4 Default values used in the model. Parameter description Symbol Value Reference Field capacity FC 1 in/ft Lot size LotSize 1/8 ac Chow et al. 1988
127 Table B 4 Continued. Parameter description Symbol Value Reference Root depth RD 30 cm Doss et al. 1960; Peacock and Dudeck 1985 Irrigation schedule Sun/Thur, 1.27 cm Miami Dade County restriction Soil moisture based threshold TH 0.7 Rain sensor setting RSS 1.27 cm Initial soil water content SWC o 75% of FC Table B 5 S ymbols or abbreviations and definitions. Symbol Definition SWC Soil water content ETa Actual evapotranspiration ETr Reference evapotranspiration Kc Crop coefficient WB Water balance R Rainfall I Irrigation S Storage capacity of the soil Smx M aximum possible retention of water during a storm SWCm Soil water content, modified for runoff and percolation Q Runoff F Infiltration PERC Percolation FC Field capacity WP Wilting point RD Rooting depth RSS Rain sensor setting TH Threshold setting for soil moisture sensor ETC Cumulative evapotranspiration
128 LIST OF REFERENCES Abtew, W. and J. Obeysekera. 1995. Lysimeter study of evapotranspiration of Cattails and comparison of three estimation methods. Transactions of the ASAE 38(1) : 121129. Allen, R.G., L.S. Pereira, D. Raes, and M. Smith. 1998. Crop evapotranspiration: guidelines for computing crop requirements. Irrigation and Drainage Paper No 56, FAO, Rome, Italy. Allen, R.G., I.A. Walter, R. Elliot, T. Howell, D. Iten fisu, and M. Jensen. 2005. The ASCE standardized reference evapotranspiration equation. American Society of Civil Engineers Environmental and Water Resource Institute (ASCE EWRI). Available at: http://www.kimberly.uidaho.edu/water/asceewri/ascestzdetmai n 2005.pdf Accessed 20 June 2011 Augustin, B.J. and G.H. Snyder. 1984. Moisture sensor control led irrigation for maintaining bermudagrass turf. Agronomy Journal 76 : 848 850. American Society of Agricultural Engineers (ASAE). 2003. Test Procedure for determining the uniformity of water distribution of center pivot and lateral move irrigation machines equipped with spray or sprinkler nozzles. ANSI/ASAE S436.1, St. Joseph, Mich. Barton, L. and T.D. Colmer. 2006. Irrigation and fertiliser strategies for minimising nitrogen leaching from turfgrass. Agric. Water Mgmt 80: 160 175. Baseline. 2009. WaterTecTM S100 Installation Manual Boise, ID. Baseline, Inc. Beard, J.B. 1973. Turfgrass Science and Culture. Englewood Cliffs, N.J.: Prentice Hall, Inc. Blackmer, A. M. 2000. Bioavailability of nitrogen. In Handbook of Soil Science, D 3:D 18. Boca Raton, Fla.: CRC Press LLC. Booltink, H.W.G and J. Bouma. 1991. Physical and morphological characterization of bypass flow in a well saturated clay soil. Soil Sci. Society of America Journal 55(5):12491254. Bowman, D.C., C.T. Cherney, and T. W. Rufty. 2002. Fate and Transport of Nitrogen Applied to Six Warm Season Turfgrasses. Crop Sci. 42(3) : 833841. Cardenas Lailhacar, B. and M.D. Dukes. 2008. Expanding disk rain sensor performance and potential savings. J. Irrig. and Drainage Eng. 134(1):67 73.
129 Cardenas Lailhacar, B., M.D. Dukes, and G. L. Miller. 2008. Sensor bas ed automation of irrigation on bermudagrass during wet weather conditions. J. Irrig. And Drainage Eng 134 (2) : 120128. Cardenas Lailhacar, B., M.D. Dukes, and G.L. Miller. 2010. Sensor based automation of irrigation on bermudagrass during dry weather conditions J. Irrig. and Drainage Eng 136 (3) : 161 223. Carlesso, R., R.B. Spohr, F.L.F Eltz, and C.H. Flores. 2011. Runoff estimation in southern Brazil based on Smith's modified model and the Curve Number method. Agric. Water Mgmt 98(6) : 10201026 Cey, E.E. and D.L Rudolph. 2009. Field study of macropore flow processes using infiltrat ion of a dye tracer in partially saturated soils. Hydrol. Process. 23: 17861779. Chin, D. 2008. Phenomenological models of hydrologic processes in south Florida J. of Hydrology 349:230243 Chow, V. T., D.R Maidment, and L.W. Mays. 1988. Applied Hydrol ogy McGraw Hill Higher Education. Davis, S. L and M D Dukes. 2010. Irrigation scheduling performance by evapotranspirationbased controllers. Agric. Water Mgmt 98(1): 1928. Davis, S. L., M.D. Dukes, and G.L. Miller. 2009. Landscape irrigation by e vapotranspirationbased irrigation controllers under dry conditions in s outhwest Florida. Agric. Water Mgmt 96(12):18281836. Davis, S.L., M.D. Dukes, S. Vyapari, and G.L. Miller. 2007. Evaluation and demonstration of evapotranspirationbased irrigation controllers. In Proceedings ASCE EWRI World Environmental and Water Resources Congress Tampa, Fl. Doss, B.D., D.A. Ashley, and O.L. Bennett. 1960. Effect of soil moisture regime on root distribution of warm season forage species. Agron. J. 52(10) : 569 572. Duda, A.M. 1993. Addressing nonpoint sources of water pollution must become an international priority. Water Sci. and Tech. 28(35) : 1 11. Dukes, M.D. 2008. Summary of IFAS Turf and Landscape Irrigation Recommendations AE436. Institute of Food a nd Agricultural Sciences. Universit y of Florida, Gainesville, FL Available at: http://edis.ifas.ufl.edu/ae436 Accessed 20 Sept 2011.
130 Dukes, M. D. 2009. Smart Irrigation Controllers: What Makes a n Irrigation Controller Smart? AE442. Institute of Food and Agricultural Sciences. Universit y of Florida, Gainesville, FL. Available at: http://edis.ifas.ufl.edu/ ae442 Accessed 29 June 2011. Dukes, M.D. 201 2 Water conservation potential of landscape irrigation smart controllers. Trans. of the ASABE 55(2) : 563 569 Dukes, M.D., and M.B. Haley. 2009. Evaluation of soil moisturebased ondemand irrigation controllers, phase II. Final Project Report for Southwest Florida Water Management District. Gainesville, FL: University of Florida. Available at: http://abe.ufl.edu/mdukes/pdf/publications/SMS/SMS%20Phase%20II%20Final% 20Report%201217 09.pdf Accessed 10 June 2011. Dukes, M.D., B. Cardenas Lailhacar, and G.L. Miller. 2008. Evaluation of soil moisturebased ondemand irrigation controllers, phase I. Fin al Project Report for Southwest Florida Water Management District. Agricultural and Biological Engineering Department, University of Florida, Gainesville, FL. Available at: http://abe.ufl.edu/mdukes/pdf/publications/SMS/SMS Phase I Final Report85 08.p df Accessed 12 July 2011. Dukes, M.D and D. Haman. 2002a. Residential Irrigation System Rainfall Shutoff Devices ABE325, Institute of Food and Agricultural Sciences. Universit y of Florida, Gainesville, FL. Available at: http://edis.ifas.ufl.edu/ae221. A ccessed 19 Mar 2012. Dukes, M.D and D. Haman. 2002b. Operation of residential ir rigation controllers. CIR1421, Institute of Food and Agricultural Sciences. Universit y of Florida, Gainesville, FL. Available at: http://edis.ifas.ufl.edu/ae220 Accessed 5 March 2012. Dukes, M. D., M. Shedd, and S.L. Davis. 2009. Smart Irrigation Controllers: Operation of EvapotranspirationBased Controllers. AE446. Institute of Food and Agricultural Sciences. Universit y of Florida, Gainesville, FL. Available at: http://edis.ifas.ufl.edu/ae446. Accessed 30 June 2011. Eagleman, J. R. and W. L. Decker. 1965. The role of soil moisture in evapotranspiration. Agron. J. 57 : 626 629. Eagleson, P.S. 1970. Dynamic Hydrology McGraw Hill, New York. Easton, Z.M. and A.M. Petrovic 2004. Fertilizer Source Effect on Ground and Surface Water Quality in Drainage from Turfgrass. J. of Environ. Quality 33(2) : 645655. Ekern, P. C. 1966. Evapotranspiration by bermudagrass sod, Cynodon dactylon L. Pers ., in Hawaii. Agron. J. 58 : 387 390.
131 EPA ( U.S. Env ironmental Protection Agency). 1993a. Methods for determination of inorganic substances in environmental samples. (EPA 600/R -93/100), Method 353.2. Revision 2. EPA ( U.S. Env ironmental Protection Agency). 1993b. Methods for determination of inorganic substances in environmental samples. (EPA 600/R -93/100), Method 350.1. Revision 2. EPA ( U.S. Env ironmental Protection Agency). 2010. Conserving w ater. Washington, D.C.: United States Environmental P rotection Agency. Available at: http://www .epa.gov/greenhomes/ConserveWater.htm. Accessed 3 April 2012. EPA ( U.S. Env ironmental Protection Agency). 2011. Federal Water Quality Standards for the State of Florida. Washington, D.C.: United States Environmental P rotection Agency. Available at: http ://water.epa.gov/lawsregs/rulesregs/florida _index.cfm. Accessed 4 October 2011. EPA ( U.S. Env ironmental Protection Agency). 2012. National primary drinking water regulations. Washington, D.C.: United States Environmental P rotection Agency. Available at : http://water.epa.gov/drink/contaminants/index.cfm#Inorganic A ccessed 9 May 2012. Erickson, J. E. J.L. Cisar, G.H. Snyder, and J. C. Volin. 2005. Phosphorous and potassium leaching under contrasting residential landscape models established on a sandy soil. Crop Sci. 45: 546 552. Erickson, J.E., J.L Cisar, J.C Volin, and G.H. Snyder. 2001. Comparing nitrogen runoff and leaching between newly established St. Augustinegrass turf and an alternative residential landscape. Crop Sci 41: 1889 1895. Evett, S.R. 2007. Chapter 2: Soil water and monitoring technology. In Irrigation of Agricultural Crops 25 84. R.J. Lascano and R.E. Sojka, eds. M adison, Wisc. American Soc. of Agron., Inc., Crop Sci. Soc. of America, Inc., and Soil Sci. Soc. of America, Inc. FDEP. 2010. Sustaining our water resources, annual report on regional water s upply planning. Florida Department of Environmental Protection. Tallahassee, Fla. Gold, A.J., W.R. DeRagon, W.M. Sullivan, and J.L. Lemunyon. 19 90. Nitrate nitrogen losses t o groundwater from rural and suburban land uses. J. Soil Water Conserv 45: 305 310. Green, W.H. and G.A. Ampt. 1911. Studies on soil physics, part I, the flow of air and water through soils. J. Agric. Sci. 4(1) : 1 24. Haley, M.B. and M. D. Dukes. 20 07. E valuation of sensor based residential irrigation water application. ASABE Paper No. 072251. St. Joseph, Mich.: ASABE.
132 Haley, M.B. and M. D. Dukes. 2012 Validation of landscape irrigation reduction with soil moisture sensor irrigation controllers. J. Ir rig. and Drainage Eng. 138(2) : 135144. Harrold, L.L. and F.R. Dreibelbis. 1951. Agricultural hydrology as evaluated by monolith lysimeters. USDA Technical Bulletin No. 1050 : 1 149. Havlin, J.L., J.D. Beaton, S. L. Tisdale, and W.L. Nelson. 2005. Soil Fertility and Fertilizers An Introduction to Nutrient Management 7th ed. Upper Saddle River, New Jersey: Pearson Prentice Hall. Hargreaves, G.H. and Z.A. Samani. 1985. Reference crop evapotranspiration f rom t emperature. Applied Eng. in Agric. 1(2):9699. Hor n, G.C. 1965. Excess water and turfgrass management. I. Florida Turfgrass Association Bulle tin 12(2): 1 3. Horst, G.L. and J. Peterson. 1990. Turfgrass irrigation scheduling and quality with soil moisture sensors. In Visions of the future, 82 87. ASAE Pub l. 04 90. ASAE, St. Joseph, Mich. Horton, R. E. 1933. The role of infiltration in the hydrologic cycle. Trans. Am. Geophys. Union 14 : 446 460. Horton, R.E. 1939. Analysis of runoff plat experiments with varying infiltration capacity. Trans. Am. Geophys. Union 20 : 693 711. Irrigation Association [IA]. 2005a. Turf and landscape best management practices. Available at : www.irrigation.org/gov/pdf/IA_BMP_APRIL_2005.pdf Accessed 27 May 2006. Irrig. Assoc., Falls Church, VA. Irrigation Association [IA]. 2005b. Landscape Irrigation Sc heduling and Water Management. Irrigation Association Water Management Committee, Falls Church, VA. Irrigation Association [IA]. 2007. Definition of a Smart Controller. The Irrigation Association, Falls Church, VA. Availabl e at: http://www.irrigation.org/gov/pdf/ Definition_ Smart_Controller.pdf. Accessed 17 February 2009. Irrigation Association [IA]. 2008. Smart Water Application Technologies (SWAT), Climatologically Based Controllers, Eighth Testing Protocol. SWAT Committ ee, Falls Church, VA.
133 Irrigation Association [IA]. 2011. Smart Water Application T echnologies Brings Residential Irrigation into the 21st Century. Available at: http://www.irrigation.org/ SWAT/Background.aspx Accessed 2 Mar 2011. Irrig. Assoc., Falls Church, VA. Jain, M.K., S. K. Mishra, and V P. Singh. 2005. Evaluation of AMC dependent SCS CNbased models using watershed characteristics. Water Resources Mgm t 20(4): 531 552. Jarvis N. 1998. Modeling the impact of preferential flow on nonpoint source pollution. In Physical non equilibrium in soils -Modeling and Application, Selim H.M., Ma L. (eds.) Ann Arbor Press: Chelsea, MI; 195 221. Jia, X., M. D Dukes, and J.M. Jacobs. 2009. Bahiagrass crop coefficients from eddy correlation measurements in Centr al Florida. Irrig. Sci : 8(1) :5 15. Jones J. W., L.H Allen, S.F. Shih, J.S. Rogers, L.C. Hammond, A.G. Smajstrala, and J. D. Martsolf. 1984. Estimated and measured evapotranspiration for Florida c limate crops and soils. IFAS T echnical Bulletin 840. Kenny, J F., N.L. Barber, S.S. Hutson K.S. Linsey, J.K. Lovelace, and M.A. Maupin. 2009. Estimated use of water in the United States in 2005: U.S. Geological Survey Circular 1344. Kidder, G., E.A. Hanlon, T.H. Yeager, and G. L. Miller. 1998. IFAS Standardized Fertilization Recommendations for Environmental Horticulture Crops. SL141. Soil and Water Science Department, Florida Cooperative Extension Service, IFAS, UF. Available at : http://edis.ifas.ufl.edu/cn011#FOOTNOTE_1. Accessed 30 June 2011. Knisel, W. G. (Editor), 1980. CREAMS: A Field Scale Model for Chemicals, Runoff, and Erosion from Agricultural Management Systems. Conservation Research Report No. 26, U.S. Department of Agriculture, Washington, D.C., 643 pp. Available online at: http://www.tifton.uga.e du/sewrl/Gleams/creams2.pdf Accessed 30 June 2011. Loehr, R.C. 1974. Characteristics and comparative magnitude of nonpoint sources. Journal (Water Pollution Control Federation) 46(8) : 18491872. Mayer, P.W., W.B. DeOreo, E.M. Opitz, J.C. Kiefer, W.Y. Da vis, B. Dziegielewski, and J. O. Nelson. 1999. Residential End Uses of Water. Denver, Colorado. AWWA Research Foundation and American Water Works Association. Marella, R.L. 2009. Water withdrawals, use, and trends in Florida, 2005: U.S. Geological Survey Scientific Investigations Report, 20095125.
134 McCready, M.S. and M.D. Dukes. 2011. Landscape irrigation scheduling efficiency and adequacy by various control technologies. Agric Water Mgm t 98(4) : 697704. McCready, M. and K. W. Migliaccio. 2011. Impleme nting smart contr ollers for lawn and ornamental irrigation in the real world. Florida Water Resources J 63(10) : 1820. McCready, M. S., M.D. Dukes, and G.L. Miller. 2009. Water conservation potential of smart irrigation controllers on St. Augustinegr ass. Agric Water Mgm t 96(11) : 16231632. Meeks, L., M.D. Dukes, K.W. Migliaccio, and Cardenas Lailhacar, B. 2012. Long t erm expanding disk rain sensor accuracy J. Irrig. Drainage Eng. 138:1620. Miami Dade County. 2012. W ater Restrictions. Available at : http://www.miamidade.gov/ conse rvation/water_restrictions.asp. Accessed 5 Mar 2012. Miami Dade County, FL. Michel, C., V. Andrassian, and C. Perrin. 2005. Soil Conservation Service Curve Number method: How to mend a wrong soil moisture accounting procedure?. Water Resour ces Res. 41 W02011. Milesi, C., S. Running, C. Elvidge, J. Dietz, B. Tuttle, and R. Nema ni. 2005. Mapping and modeling the biogeochemical cycling of turf grasses in the United States. Environ. Mgm t 36(3) : 426 438. Mishra, S.K and V.P. Singh. 2002. Chapter 13: SCS CNbased hydrologic simulation package. In Mathematical Models in Small Watershed Hydrology 391 464. V.P. Singh and D.K. Frevert. eds. Littleton, Colorado: Water Resources Publications. Migliaccio, K.W. and N. Do bbs. 2011. FAWN Irrigation Tool Model Manual. Morris, K.N. and R.C. Shearman.1998. NTEP turfgrass evaluat ion guidelines. NTEP Turfgrass Evaluation Workshop. Beltsville, Maryland, pp. 1 5. Morton, T.G., A.J Gold, and W.M Sullivan 1988. Influence of ov e rwatering and fertilization on nitrogen losses from home lawns. J. Environ. Quality 17 : 124 130. Muoz Carpena, R. 2004. Field devices for monitoring soil water content. BUL343. Agricultural and Biological Engineering Department, Florida Cooperative Ext ension Service, IFAS, UF. Available at : http://edis.ifas.ufl.edu/ae266. Accessed 18 June 2011. Muoz Carpena, R., A. Ritter, and D.D. Bosch. 2005. Field methods for monitoring soil water status. J. Alvarez Benedi, R. Muoz Carpena (Eds.), Soil Water Sol ute Process C haracterization, CRC Press, Boca Raton (2005), pp. 167 195.
135 Nimmo, J.R. 2012. Preferential flow occurs in unsaturated conditions. Hydrol. Process 26: 786 789. Qualls, R.J., J.M. Scott, and W. B. DeOreo. 2001. Soil moisture sensors for urban landscape irrigation: effectiveness and reliability. Journal of American Water Resources Association 37: 547559. Pathan, S.M, L. Barton, and T.D. Colmer. 2007. Evaluation of a soil moisture sensor to reduce water and nutrient leaching in turfgrass. Aust ralian Journal of Experimental Agriculture 47: 215222. Peacock, C.H. and A.E. Dudeck. 1985. Effect of irrigation interval on St. Augustinegrass rooting. Agronomy Journal 77(5) : 813815. Petrovic, M. A. 1990. The fate of nitrogenous fertilizers applied to turfgrass. J. of Environmental Quality 19(1) : 1 14. Philip, J.R. 1957. The theory of infiltration: 1. The infiltration equation and its solution. Soil Sci. 83(5) : 345 357. Philip, J.R. 1969. Theory of infiltration in Advances in Hydroscience V.T. Chow (e d.). Volume 5: 215 296. Ponce, V.M. and R.H Hawkins. 1996. Runoff curve number: Has it reached maturity?. J. Hydrol. Eng. ASCE 1 (1), 11 19. Priestley, C.H.B. and Taylor, R.J. 1972. On the assessment of surface heat flux and evaporation using largescale parameters. Monthly Weather Review 100: 81 92. Pruitt, W. O. 1964. Evapotranspirationa guide to irrigation. California Turfgrass Culture 14(4) : 27 32. Raudkivi, A.J. 1979. Hydrology Pergamon Press, Oxford. Rawls, W.J., D.L. Brakenseik, and K.E. Saxton. 1 982. Estimation of soil water properties. Trans. ASAE 25:13161320. Richards, S.J. and L.V. Weeks. 1963. Evapotranspiration for turf measured with automatic irrigation equipment. California Agriculture 17(7) : 1213. Rimon, Y., R. Nativ, and O. Dahan. 2011 Physical and chemical evidence for porescale dual domain flow in the vadose zone. Vadose Zone Journal 10(1) : 322331.
136 Romero, C. and Dukes, M.D. 2011. Net Irrigation Requirements for Florida Turfgrass lawns: Part 3 Theoretical Irrigation R equirements AE482 Institute of Food and Agricultural Sciences. Universit y of Florida, Gainesville, FL. Available at: http://edis.ifas.ufl.edu/ae482. Accessed 20 September 2011. Runnels, H. A. and J.D. Wilson. 1934. The influence of certain spray materials, herbicides, and other compounds on the desiccation of plant tissue. Ohio Experimental Experiment Station Bimonthly Bulletin 19(166) : 104 109. Shedd, M., M.D. Dukes, and G.L. Miller. 2007. Evaluation of evapotranspiration and soil moisturebased irrigation control on turfgrass. In Proceedings ASCE EWRI World Environmental and Water Resources Congress Tampa, Fla. Slayter, R.O. and J.F. Bierhuizen. 1964. The effect of several foliar sprays on transpiration and water use efficiency of cotton plants. Agricultural M eteorology 1 : 4253. Smettem, K.R J., C. Kirkby, and D.J. Chittleborough. 1994. Hydrologic response of undisturbed soil cores to simulated rainfall. Australian J of Soil Res 32(6) : 11751187. Smith, R.E. and J.R. Williams. 1980. Simulation of the surface hydrology. In: Knisel W (ed) CREAMS: A field scale model for chemicals, runoff and erosion from agricultural management systems. US Dep Agric Conserv Res Report 26:60 8 Snyder, G. H., B.J. Augustin, and J.M. Davidson. 1984. Moisture sensor controller irrigation for reducing N leaching in bermudagrass turf. Agron. J. 76(6) : 964 969. Soil Conservation Service (SCS) 1972. Hydrology. In National Engineering Handbook Washington, D.C.: U.S. Government Printing Office. State of Florida.com. 2011. Florida Quic k Facts. Available at : http://www.stateofflorida. com/portal/desktopdefault.aspx?tabid=95. Accessed 12 July 2011. Stumpp, C. and P. Maloszewski. 2010. Quantification of preferential flow and flow heterogeneities in an unsaturated soil planted with different crops using the environmental isotope 18O. Journal of Hydrology 394(34) : 407415. Sumner, D. M. and J. M. Jacobs. 2004. Utility of Penman Monteith, Priestley Taylor, reference evapotranspiration, and pan evaporation methods to estimate pasture evapotranspiration. Journal of Hydrology 30 : 81 104. Topp, G.C., J.L. Davis, and A.P. Annan. 1980. Electromagnetic determination of soil water content: measurements in coaxial transmission lines. Water Resour. Res. 16:574582.
137 Trenholm, L.E. and J. B. Unruh. 2005. Florida Lawn Handbook Third. Ga inesville: University Press of Florida. Trenholm, L.E., E.F. Gilman, G. Denny, and J. B. Unruh. 2002 Fertilization and Irrigation Needs for Florida Lawns and Landscapes. ENH860. Environmental Horticulture Department, Florida Cooperative Extension Service, IFAS, UF. Available at : http://edis.ifas.ufl.edu/ep110. Accessed July 12 2011. Turc, L. 1961. Evaluation des besoins en eau dirrigation, vapotranspiration potentielle, formulation simplifi et mise j our. Ann. Agron., 12 : 13 49. UF/IFAS Analytical Services Laboratories. Nitrogen and Phosphorous Cycle. Gainesville, Fla. Available at: http://nutrients.ifas.ufl.edu/Cycle.asp. Accessed 4 October 2011. U.S. Census Bureau. 2010. 2010 Census urban and rural classification and urban area c riteria Washington, D.C.: United States Census Bureau. Available at : http://www.census.gov/geo/www/ua/2010urbanruralclass.html. Accessed 20 June 2011. U.S. Census Bureau. 2011. State and County QuickFacts: Florida. Washington, D.C.: United States Census Bureau. Available at : http://quickfacts.census.gov/qfd/ states/12000.html Accessed 20 June 2011. USDA. 1955. Evaluation of agricultural hydrology by monolith lysimeters. USDA Technical Bulletin No. 1179: 1 166. Waller P. 2007. Turf and Landscape Irrigation. In Irrigation of Agricultural Crops 2nd ed., 337361. Madison, Wisc.: American Society of Agronomy, Crop Science Society of America, Soil Science Society of America. Whitcomb, J. B. 2005. Florida Water Rates Eval uation of singlefamily homes. Southwest Florida Water Management District. Available at: http://www.swfwmd.state.fl.us/documents/reports/water_rate_report.pdf Accessed 5 October 2011. Wilson, J.D. and H. A. Runnels. 1931. Bordeaux mixture as a factor i ncreasing drought injury. Phytopathology 21: 729 738. Xiao, B., Q.H Wang, J. Fan, F.P. Han, and Q.H Dai. 2011. Application of the SCS CN model to runoff estimation in a small watershed with high spatial heterogeneity. Pedosphere 21(6) : 738 749.
138 Zota r elli, L., M. D. Dukes, and K. T. Morgan. 2010. Interpretati on of soil moisture content to determine soil field capacity and avoid over irrigating s andy soils using soil moisture sensors. AE460. Agricultural and Biological Engineering Department, Florida Cooperative Extension Service, IFAS, UF. Available at : http://edis.ifas.ufl.edu/ae460. Accessed 17 June 2011.
139 BIOGRAPHICAL SKETCH Nicole Dobbs attended primary schools St. Stephens Catholic School in Bradshaw, Maryland and St. Josephs Catholic School in Dallastown, Pennsylvania. She attended secondary school York Catholic High School in York, Pennsylvania, where she graduated as salutatorian of her class. Beginning in primary school, she had an interest in exploring natural environments. This interes t was nurtured by her parents outside of school. In school, she took a particular interest in science classes and participated in extracurricular sciencebased activities, includi ng an academic team that studied different aspects of local wildlife biology, forest ecology, and current environmental issues. T he appreciation for the beauty of the natural world, combined with the passion for understanding the natural systems and wanting to conserve natural resources greatly influenced her decision to study envi ronmental engineering during her undergraduate studies at the University of Delaware in Newark, Delaware. After obtaining her b achelors, she moved to Florida, first to intern with the South Florida Water Management District, and then begin graduate studies at the University of Florida in the Agricultural and Biological Engineering Department.