This item is only available as the following downloads:
1 DEFICIT IRRIGATIO N OF MIXED LANDSCAPES BASED ON TURFGRASS COVERAGE AND REFERENCE EVAPOTRANSPIRATION By SCOTT H SIMPSON A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE UNIVERSITY OF FLORIDA 2012
2 2012 Scott H Simpson
3 To m y w ife Angelia Marie Ellison
4 ACKNOWLEDGMENTS Most of all, I would like to thank Dr. Richard Beeson, my chair, for insisting things be done right, without assumption. I would also like to thank Diane Mealo, my program coordinator, for her guidance and help when I began this quest. I would also like to thank Dr. Gail Hansen De Chapman, my Co c hair, for showing me how landscapes can be d esigned more sustainably. I would like to thank the following list of individuals: Dr. Richard 'Jake' Henny, Dr. Brian Pearson, Dr. Dilma Silva, Steve Toomoth, Ed Tilman, Chris Fooshee, Dr. Robert Stamps, Dr. Charles Guy, Dr. Michael Dukes, and the staff at MREC. Finally, I would like to thank all the other passionate instructors I had the privilege of learning from over the years. I hope I make you all proud. Special thanks goes to the Dean of the College of Agriculture for my part time assistantship This research was supported by the Mid Florida Research and Education Center and funded by the Southwest Florida Water Management District.
5 TABLE OF CONTENTS P age ACKNOWLEDGMENTS ................................ ................................ ................................ .. 4 LIST OF TABLES ................................ ................................ ................................ ............ 7 LIST OF FIGURES ................................ ................................ ................................ .......... 8 LIST OF ABBREVIATIONS ................................ ................................ ........................... 10 ABSTRACT ................................ ................................ ................................ ................... 12 CHAPTER 1 INTRODUCTION ................................ ................................ ................................ ........ 14 2 LITERATURE REVIEW ................................ ................................ .............................. 16 Evapotranspiration ................................ ................................ ................................ .. 16 Definition and Importance ................................ ................................ ................. 16 Reference ET ................................ ................................ ................................ ... 17 Actual ET ................................ ................................ ................................ .......... 17 Calculating ET O ................................ ................................ ................................ ....... 18 Evolution of Penman Monteith ................................ ................................ ......... 18 UN FAO 56 PM with Tolerances ................................ ................................ ...... 18 ET O in Florida ................................ ................................ ................................ ... 19 Determination of ET A ................................ ................................ .............................. 20 Water Balance ................................ ................................ ................................ .. 20 Methods for Measuring ET A ................................ ................................ .............. 20 Water Budget ................................ ................................ ................................ ... 21 Crop Coefficient K C ................................ ................................ ................................ 22 Defining K C Through the Relationship of ET O and ET A ................................ ..... 22 Examples of How K C is Used ................................ ................................ ........... 23 K C Use in Established and Mixed Landscapes (K L ) ................................ .......... 29 Deficit Irrigation ................................ ................................ ................................ ....... 32 Definition and Importance ................................ ................................ ................. 32 Effect on Aesthetics ................................ ................................ .......................... 33 Growth Versus Aesthetic Quality ................................ ................................ ...... 34 Application of K L ................................ ................................ ................................ ...... 35 Residential Landscapes ................................ ................................ ................... 35 Potential of ET Based Irrigation ................................ ................................ ........ 35 3 MATERIALS AND METHODS ................................ ................................ ................... 37 4 RESULTS ................................ ................................ ................................ ................... 51
6 5 DISCUSSION ................................ ................................ ................................ ............. 61 6 CONCLUSION ................................ ................................ ................................ ........... 67 LIST OF REFERENCES ................................ ................................ ............................... 70 BIOGRAPHICAL SKETCH ................................ ................................ ............................ 76
7 LIST OF TABLES Table P age 4 1 Annual irrigation input totals (L) and total water input (L) totals for each deficit level treatment. Total rainfall per lysimet er was 13,410 L ................................ .. 51 4 2 Mean monthly values from 1 June 2010 to 31 May 2011 for daily ET O daily ET A (cm) and K L by irrigation deficit level (DI) ................................ .................... 53 4 3 Visual ratings by season for St. Augustine 'Floratam' turfgrass during the first year of deficit irrigation after plant establishment. ................................ ............... 55
8 LIST OF FIGURES Figure P age 3 1 Drainage lysimeters on the west row after construction. The pipe exiting the bottom was connected to a system designed for leachate collection. The white pipes on the right were for irrigation. ................................ ......................... 37 3 2 Installed drain line prior to covering with rock and sand. ................................ .... 38 3 3 A view of lysimeter project looking from south to north. ................................ ...... 39 3 4 A view of an example of one of the nine lysimeters shortly after transplanting. .. 41 3 5 Freeze damage to turfgrass from winter of 200 9 2010, shown here in mid July 2010. ................................ ................................ ................................ ........... 43 3 6 Measuring device for quantifying leachate from each lysimeter. The upper vessel collects water which drains to the weighing vessel (bottom) through the normally open valve (green and gold). The system is supported by the T frame above the dry well, shown by arrow. ................................ ........................ 44 3 7 Sealed shed which housed the leachate measuring device spring 2010 ........... 45 3 8 Data collection and irrigation control system for the drainage lysimeter project. ................................ ................................ ................................ ................ 46 3 9 Turfgrass visual rating guide used to evalute aesthetic quality of St. Augustine turfgrass. It is based on a one to nine scale, nine being perfect and therefore not shown. ................................ ................................ .................... 49 4 1 Comparison of monthly total rainfall (cm) during the data collection period of 1 June 2010 to 31 May 2011. The onsite weather station was 100 m west of the lysimeter site. The FAWN (Florida Automated Weather Network) average rainfall was based on a 14 yr history of the station within 500 m of th e onsite station. ................................ ................................ ............................... 52 4 2 Comparison of monthly K L by treatment during the data collection period of 1 June 2010 to 31 May 2011. Lines represent different deficit irrigation treatments as a pe rcentage of ET O that was counted toward triggering an irrigation event at >1.9 cm. Each point is the mean of 3 lysimeter replicates. ... 54 4 3 Mean dry mass (g) measurements at each mowing for S t. Augustine 'Floratam' turfgrass during the first year of deficit irrigation after plant establishment. Means represent nine treatment replicates across deficit irrigation treatments. ................................ ................................ ........................... 55
9 4 4 Height (m) measurements of Viburnum odoratissimum during the first year of deficit irrigation after plant establishment. Means represent 9 treatment replicates across deficit irrigation treatments. ................................ ..................... 56 4 5 Projected Canopy Area (PCA) (m 2 ) of Viburnum odoratissimum during the first year of deficit irrigation after plant establishment. Means represent 9 treatment replicates across deficit irrigation treatments. ................................ ..... 57 4 6 Growth Index (GI) (m 3 ) of Viburnum odoratissimum during the first year of deficit irrigation after plant establishment. Means represent 9 treatment replicates across deficit irrigation treatments. ................................ ..................... 57 4 7 Trunk circumference (cm) taken 15 cm above soil level of Magnolia grandiflora during the first year of deficit irrigation after plant establishment. Means represent 9 treatment replicates across deficit i rrigation treatments. ...... 58 4 8 Mean height (m) measurements of Magnolia grandiflora during the first year of deficit irrigation after plant establishment. Each mean represents 3 tree replicatio ns. ................................ ................................ ................................ ........ 59 4 9 PCA (m 2 ) measurements of Magnolia grandiflora during the first year of deficit irrigation after plant establishment. Each mean represents 3 tree replications. ................................ ................................ ................................ ........ 60 5 1 Seasonal shoot and root growth pattern of warm season turfgrass (Turgeon 2002). Each tick on the horizontal represents one month. ................................ 62
10 LIST OF ABBREVIATION S ANOVA Analysis of Variance DI Deficit irrigation level EPA Environmental Protection Agency EST Eastern standard time ET Evapotranspiration ET A Actual Evapotranspiration ET O Reference Evapotranspiration FAWN Florida automated weather network GI Growth index g S S tomata l co nductance IFAS Institute of Food and Agricultural Sciences K C Crop coefficient K L Landscape coefficient LSD Least significant difference NTEP National turfgrass evaluation program PCA Project canopy area PET Potential evapotranspiration PVC Polyvinyl chlor ide SAS Statistical Analysis System St. Saint T SW Transpiration (stomatal water loss) UN FAO56 PM United Nations Food and Agriculture Organization Penman Monteith USDA United States Department of Agriculture USGS United States Geological Survey
11 VAC Volts, alternating current VDC Volts, direct current VPD Vapor pressure deficit WNI Water needs index WUCOLS Water use classification of landscape species
12 Abstract of Thesis Presented to the Graduate School of the University of Florida in Par tial Fulfillment of the Requirements for the Degree of Master of Science DEFICIT IRRIGATION OF MIXED LANDSCAPES BASED ON TURFGRASS COVERAGE AND REFERENCE EVAPOTRANSPIRATION By Scott H S impson May 2012 Chair: Richard C. Beeson, Jr. Co c hair: Gail Hanse n De Chapman Major: Horticultural Science Environmental Horticulture Irrigation of landscapes can be responsible for more than half of the water consumption of residential homeowners. The objectives of the research presented here were to test two hypo theses. First, that irrigation frequency based on turfgrass water needs is sufficient for the irrigation of woody shrubs and trees within a mixed landscape. Second, that warm season St Augustine turfgrass can maintain an aesthetically pleasing appearance at irrigation volumes and frequencies less than predicted by ET O Florida Research an d Education Center Apopka, Flo rida. Lysimeters had a surface area of 13 m 2 each and contained two Viburnum odoratissimum one Magnolia grandiflora magnolia, and 9.7 m 2 of a tam St. Augustine turfgrass Stenotaphrum secundatum Irrigation regimes of 6 0%, 75% and 9 0% of ET O were adhered to throughout the year. Irrigation occurred when the cumulative depth of ET O exceeded 1.90 cm.
13 All magnolias and viburnum hedges displayed aesthetically pleasing quality, independent of DI level throughout the year. Turfgrass qua lity varied among DI levels. All turfgrass plots were rated above the minimum acceptable quality. ET O derived from the UN FAO Penman Monteith equation in Central Florida and still maintain acceptable aesthetic quality. This frequency also maintains acceptable quality of magnolia trees and a typical woody hedge if concurrently irrigated at 72% ET O based on horizontal canopy project area.
14 CHAPTER 1 INTRODUCTION In 1992, t urfgrass accounted for 4.4 million acres of maintained area in Florida, o f which St. Augustine occupied 1.5 million acres (Hodges et al. 1994). T he population of Florida was just under 13 million in 1990, under 16 million in 2000, slighty under 19 mill ion in 2010, and is projected to nearly exceed 22 million by 2020 (Census 2010). The increasing population will likely result in more homes built that require landscape irrigation. Public water use in Florida in 2005 totaled 2.54 billion, gallons/day (US GS 2005). Sixty one percent of this went to residential water use, with sixty four percent of that applied as landscape irrigat ion (Fernald and Purdum 1998). In general, homeowners have a tendency to over irrigate their landscapes (Haley et al. 2007). B ecause irrigation scheduling has historically been based on regular temporal intervals, the same irrigation levels are typically set and applied throughout the seasons with no adjustment for the climatic changes which directly affect landscape plant water needs (Stacia and Dukes 2011). Irrigation with climate based controllers has the potential to save 20% to 50% of the water consumed by irrigation in residential landscapes (Hilaire et al. 2008). Since landscape water needs should be based on maintaining turfgrass, tree, shrub, and ornamental aesthetic appeal rather than maximization of growth or even yield; a deficit of the maximum amount of water required can be used (Pittenger et al. 2001). Research has shown that plant material in nursery production ( Beeson 2006), agricultural settings (Allen et al. 1998), ornamental settings (Scheiber and Beeson 2007), and established landscapes (Sachs 1991) can be maintained at an aesthetically pleasing level with irrigation based on a percentage of reference evapot ranspiration.
15 The research presented here sought to establish a landscape coefficient based on turf water needs only, yet useful for mixed landscape irrigation scheduling. The coefficient is a correction factor that reduces the amount of water applied to the landscape as a percentage of reference evapotranspiration. By quantifying the relationship between reference evapotranspiration (ET O ) and actual evapotranspiration (ET A ) the goal was to demonstrate that a landscape coefficient (K L ) based on turfgrass water needs only would be sufficient for scheduling irrigation for mixed landscapes. Three levels of deficit irrigation (DI), 60%, 75% and 90% were used to identify the level that maintained acceptable aesthetic quality for the simulated landscapes. In addition this project sought to demonstrate that deficit levels not only could conserve water, but could also limit growth. By limiting unnecessary growth encouraged by excessive irrigation, maintenance costs associated with mowing and trimming would also be reduced. Reduced mowing frequency would also reduce emissions. A 1991 EPA study on non road emissions found that lawn mowers contribute 16,800 to 23,800 tons per year of emissions (EPA 1991).
16 CHAPTER 2 LITERATURE REVIEW Evapotranspiration Definition and I mportance Evapotranspiration (ET) is the term used to describe water loss from a plant system due to a combination of transpiration and evaporation. Plants transpire by uptake of water through roots, where it is transported through the vascular syst em (xylem), to exit through stoma as water vapor. Evaporation occurs not only from the edaphic environment, but also from the cell surfaces inside the leaf before the water of water vapor just above the canopy, which includes transpiration from the leaves plus (Nobel 1999) ET rates are driven by gradients between the atmosphere and a leaf, within the leaf, and the effect of surrounding surfaces. These gradients are affected by atmospheric conditions, plant physiology, and the characteristics of surrounding surfaces. The importance of ET is linked to the proper scheduling of irrigation events in relation to the climatic demands of the region comb ined with the measured water use of the crop. The calculation of ET has its roots in the measurement of evaporation from an open water surface (measured from an open pan or pit) as compared to that of a natural surface (dirt or turf). In 1948, Dr. Howard Penman found that evaporation from a water surface could be correlated to evaporation from a natural surface (Penman 1948) He also observed that for turfgrass, these correlations varied with season and climate. Evaporation from natural surfaces was foun d to be less than that of open water. In his region of England, it was found that winter evaporation from natural surfaces was
17 approximately 60% of the open pan evaporation; and summer evaporation from natural surfaces was approximately 80% of the open pa n evaporation. This process of measurement and calculation came to be accepted as a good model of evapotranspiration for that specific turf in that specific region. Reference ET Evapotranspiration (ET) can be expressed as potential ET (PET), reference E T (ET O ), and actual ET (ET A ). Penman eventually transpired in unit time by a short green crop, completely shading the ground, of uniform (Penman 1948) This would be under ideal growing conditions where the measurements of several climatic factors are used to calculate the maximum evapotranspiration rate of that canopy. PET is calculated by an equation that requires data recorded from an onsite weather station. PET is the basis for ET O and the two are sometimes used interchangeably. More specifically, ET O is the PET that has been calculated from data recorded by a weather station at a specific location, and is used as a reference for that region. ET A is the actual evapotranspiration of a crop. Actual ET ET A can be determined by several methods. In the water balance method, ET A is calculated simply by measuring the water input into a system (rainfall plus irrigation), then subtracting the measured amount of water that is lost from the system. The system can be under controlled conditions such as a greenhouse or open system located outdoors. Systems can range from a single small container to entire watersheds. The measurement of rainfall and irrigation are straight forward measuremen ts. The water that is lost from a system can be measured based on
18 drainage volume or weight, such as in a lysimeter (Howell et al. 1991) Lysimeters are a special subset of the water balance method and usually considered an independent method (Howell et al. 1991) Lysimeters consist of a relatively small vessel in which plants are grown with the water balance determined on a fixed schedule, usually daily. These vessels can range from small 4 cm containers to monolithic sections of soil weighing up to 4.5 4 x 10 5 kg (Schneider et al. 1998) Calculating ET O Evolution of Penman Monteith estimate ET O each taking into consideration a different combination of climatic variables. Howev er, Penman expressed the idea that a specific combination of atmospheric and solar radiation measurements should be used to calculate ET O The Penman equation calculates ET O using daily measurements of temperature, wind speed, relative humidity, and solar radiation providing the most accurate method for daily measurement of ET O (Jones 1984) Penman found that it was the combination of these specific factors that would generate the most accurate estimation of evapotranspiration. Later, Monteith improved t (Monteith 1965) by adding factors such as stomatal conductance (or resistance) and hourly climate measurements. UN FAO 56 PM with T olerances In 1998, stomatal resistance values between day and night were added to fine tun e the Penman Monteith equation (Ventura et al. 1999) These improvements helped to develop the widely used version of the Penman Monteith equation adopted by the Food and Agriculture Organization of the United Nations known as the UN FAO 56 PM.
19 It is cur rently the world standard for calculating ET O based on meteorological data (Irmak et al. 2003a) ET O in Florida Reference evapotranspiration can be calculated by a number of different methods and equations; and is rel evant to the climate for which you wan t data One of the earliest methods for estimating ET was the pan method. By measuring the rate of evaporation from a round basin of specif ic diameter and height, termed coupling that with crop transpiration data, and correcting with a p an coefficient; one can estimate ET O for a specific region (Irmak et al. 2002) However there are several sources of error with this method (Sivyer et al. 1997) and it has not been reported in the United States since the mid es use the Penman Monteith equation or slight variations of it. One or several equations may be more appropriate for either arid, or for humid climates. Florida is considered a humid climate. Since accuracy of the equations varies with region or cli mate, data must be collected and inserted into these equations to compare accuracy and assist with equation selection. Comparison with lysimeter data that yield actual evapotranspiration is helpful (Yoder et al. 2005) Some equations, such as the UN FAO 56 PM, require more detailed climatic data. Others such as the Turc method only need temperature and solar radiation data (Irmak et al. 2003b) However in humid climates, relative humidity plays a big role in ET rates (Carrow 1995) and so should be in cluded in calculations. This mandates use of the UN FAO 56 PM. ET rates for the commonly used warm season St. Augustine turfgrass varieties (Atkins et al. 1991) and Georgia (Carrow 1995) These can be us ed in Florida with a correction factor and applied to ET
20 based irrigation scheduling of turf (Haley et al. 2007) In the Haley study, historical ET rates were used to schedule irrigation for residential landscapes. Residential water usage for landscape i rrigation was isolated and tracked. It was determined that significant savings in irrigation could be realized using ET based scheduling. Yet when compared to concurrent real time climate data, landscape s were still over irrigated, emphasizing the need for real time ET based scheduling. Determination of ET A Water B alance Water balance within the plant relies on the uptake of water and the loss of water vapor through the stomata via transpiration (Kramer and Boyer 1995) Proper water status must be main tained to achieve optimum plant growth. The force pulling in water through roots originates in leaves. This is driven by climatic factors such as relative humidity, temperature, and solar radiation. Xylem cells in the leaf and stem can hold water overni ght, and thus delay the absorption of water in the morning, causing a lag in water uptake despite moist soil. Therefore, soil water content is not a good measure of plant water status after sunrise (Kramer and Boyer 1995) Methods for M easuring ET A Actua l Evapotranspiration (ET A ) can be found by a number of different methods. In measuring forest ET A four prominent methods were explored and compared (Wilson et al. 2001) Sap flow measurements, eddy covariance measurements, catchment water balance measur ements, and soil water budgets were found to correlate on some points, with soil water budgets showing greater potential for inaccuracy. Measuring ET A can also be accomplished with the use of a biophysical model know as a lysimeter
21 (Whitehead and Kellihe r 1991) The differences in lysimeters lay in how they are constructed and how the water usage is measured. Water B udget The hydrological cycle is a closed system that transfers water from the atmosphere to the ground in the form of precipitation. The water is transported in different ways around the earth, and then vaporized back into the atmosphere. Evapotranspiration coupled with precipitation dictates how much water will be available and how a watershed responds to use and precipitation. Water th at is evaporated from a region is typically lost from that region. Evapotranspiration provides insight into the water budget of an ecosystem, quantifies water requirements, and plays a role in determining irrigation regimes (Brutsaert 1982) Water budgets are calculated by measuring the water input into a system minus the water leaving a system. Lysimeters provide a method of precise measure through a closed system to determine water budgets. There are several types of lysimeters which usually fall under the category of either weighing or drainage. Weighing lysimeters measure the soil water balance via differences in mass measurements throughout the day (Howell et al. 1991) The system is irrigated to field capacity, and then weighed periodically throu ghout the day or at a fixed time, such as sunrise. As the day progresses the system loses water due to ET A causing the system to weigh less. These differences in mass provide water loss data required for calculating ET A By subtracting the water out (me asured from the differences in mass) from the water in (precipitation plus irrigation), one can calculate the amount of water loss due to the combination of evaporation and transpiration; or ET A Compared to drainage lysimeters, weighing lysimeters are re latively easy and inexpensive to
22 construct. However, ET A produced by weighing lysimeters not only varies heavily based on the soil characteristics (Kramer and Boyer 1995) but either must be corrected for weight differences resulting from plant growth (Eh ret et al. 2001) or measurements must be taken on daily basis in order to minimize this effect. Drainage lysimeters involve the controlled collection of water through a catchment at the bottom of an enclosed area. This catchment drains to a collection d evice used to measure the amount of water that has percolated through the soil profile. By subtracting the water out (measured from the catchment) from the water in (precipitation plus irrigation), one can calculate the amount of water loss due to the com bination of evaporation and transpiration; or ET A While drainage lysimeters provide the most accurate data (Kramer and Boyer 1995) they require a lot of space, are much more expensive, and are more time consuming to build and maintain. Large drainage l ysimeters are also less accurate over short durations due to the lag time caused by the buffering capacity of the soil. Crop C oefficient K C Defining K C T hrough the R elationship of ET O and ET A Crop coefficients (K C ) are defined as correction factors that a djust reference evapotranspiration according to region and plant species (Jones, 1984). Many K C have been calculated for the different equations that estimate PET or ET O However K C are unique to each method and not interchangeable. The most prevalent e quation in use is the Penman Monteith equation. It has been adapted for use with different types of plant material by the calculation of crop specific K C (Allen et al. 1998) Crop coefficients are calculated as the ratio of ET A /ET O and are unit less. I n practice a crop is grown in a lysimeter, or a very large uni form expanse if using e ddy c orrelation, located in the
23 r egion where the data is needed. ET A can be determined by the water balance method if a lysimeter is employed Soil moisture sensors are also used in this process. ET O is calculated using daily measurements of temperature, wind speed, relative humidity, and solar radiation near the location. Because K C is unitless and ET O is calculated as a depth of water lost, ET A value, if derived from lysimeters, must be converted from volumes or mass to depths of water. In agronomic or grass crops, water loss from the system is divided by the surface area of the lysimeter and sometimes by the leaf area within the lysimeter. For non agronomic and non g rass crops, such as potted plants or container grown nursery crops or ornamental trees, the normalizing volumes or mass to a depth of water becomes more problematic. Since the Penman Monteith equation is based on a reference crop of short grass, the ratio of ET A /ET O corrects the ET rate for the crop of interest. The resulting K C value can be used in models to schedule irrigation (Beeson 2005) This is accomplished by monitoring the environment in a region via a weather station. ET O is then calculated an d multiplied by K C to estimate the amount of water to apply with the goal of achieving maximum growth or crop yield. Examples of H ow K C is U sed By combining the Penman Monteith calculations and lysimeter data, many crop coefficients have been determined a nd used successfully to schedule irrigation for individual agricultural crops (Fereres and Soriano 2007) ; as well as container plant production. A K C of 0.59 was determined for that could be used to schedule irrigation for conta iner production (Beeson 1993) Irrigation models for container production of Ligustrum japonicum based on the relationship between ET O and ET A have been successfully correlated to K C along with the use of canopy projected
24 surface area (CPSA) (Beeson 2005) Modeling of ET A was also successful for A. rubrum (Beeson and Brooks 2006b) providing data useful for calculating acceptable K C values for container production. Stomata l conductance (g S ) in Sweet gum Liquidambar styraciflua were studied concurrently in Utah, Texas, and Florida to determine the relationship between ET O vapor pressure deficits ( VPD ) and water loss through transpiration (T SW ) (Kjelgren et al. 2004) S ubset s of trees grown in each location were shipped overnight to the other locations, with g S and ET A m easured over a two week period. For the first few days after shipping, stomata responded to their previous environment, not the one they were moved to. By the second week stomata responded to the in situ environment. Because the stomata were still responding to their previous environment, K C from western trees shipped to Florida were initially lower than remaining Florida trees and differed from K C calculated the second week after shipping. Conversely, Florida trees shipped to arid clima tes maintained high g S the first few days until accl imatizing to the much higher VPD s and near constant winds. These results suggest that local analysis of transpiration is needed to properly calculate water needs index ( WNI ) K C values, especially between arid and humid climates. Daily crop water use was also examined in relationship to ET O in container grown ornamentals over various climates throughout California (Burger et al. 1987, Schuch and Burger 1997) Water use varied heavily between locations du e to variance in climatic factors such as wind and solar radiation, emphasizing the importance of the use of local, real time climatic data for the calculation of ET O and the importance of measurement of regional ET A values.
25 Even when properly spaced, K C for container plants were shown to be considerably higher than those values found for field crops (Schuch and Burger 1997) This may have been because K C was based on container surface area while crop coefficients are based on large ground areas. Alth ough the relationship between ET O and ET A was apparent, the resulting K C values varied heavily not only between locations, but also between plant species and time of year; emphasizing the need for K C to be sensitive to location, specific species, and varia nces throughout the growing season (Garcia Navarro et al. 2004) Because of the difficulty in relating K C calculated for container plants and K C calculated for field crops due to the difference in surface area (Schuch and Burger 1997) these studies for K C values for container grown plants are useful for nursery production but may be of limited use in the landscape. The water use of plants grown in production containers was measured and compared to the water use of plants grown in large lysimeters of fie ld soi l Container plant water use, although overall more than the water use of plants grown in lysimeters, correlated with plants in lysimeters This may prove helpful in grouping plants within the landscape according to water use; which in fact helps i n the landscape design process (Garcia Navarro et al. 2004) In California, many common landscape plants have been assigned recommended ranges of K C and collected under a listing known as WUCOLS (Water Use Classification of Landscape Species) (Pittenger e t al. 2008) Turfgrasses have demonstrated lower ET A when compared to ET O (Jones 1984) indicating the potential for water savings if turf irrigation is managed using ET based scheduling. Cool season and warm season turfgrass responses should be consider ed separately. In general cool season grasses have been shown to have a higher water
26 demand than warm season turfgrasses (Feldhake and Butler 1983) It was also found that management practices and microclimates significantly impacted ET rates of cool se ason grasses (Feldhake and Butler 1983) St. Augustine turfgrass ET rates were determined to have a range from 0. 63 to 0. 96 c m/day in an arid climate (Beard 1985) S oil moisture probes were later used to determine that ET rates for a 'Raleigh' St. Augu stine turfgrass range d from 0. 15 to 0. 56 c m/day in a humid climate (Carrow 1995) ET O calculations using t he pan method were also compared to the Penman method. Using both to calculate K C values, the author noted that the pan method produced different co efficients ( 0 .53 to 0 .89) than the Penman method ( 0 .52 to 1.01). However, both methods indicated the potential to irrigate warm season turfgrass in a humid climate at less than ET O The pan results are also close to those of (Meyer and Gibeault 1987) who found K C values for warm season grasses in general to range from 0 .54 to 0.79 S easonal variation s for ET rates were high and varied among turfgrass species. M onthly averages of K C values and species specific K C values c ould provide more accurate irriga tion scheduling (Carrow 1995) In Nevada (Devitt et al. 1992) and Arizona (Brown et al. 2001) useful monthly K C values were developed for bermudagrass over seeded with ryegrass. In Central Florida, K C values from eddy correlation were found for bahiagras s (Jia et al. 2009) a widely used foliage grass employed as turfgrass due to its apparent drought tolerance mechanisms. K C values spiked upwards of 0 .9 0 during the warmer months and dipped as low as 0 .35 during cool months. Results for warm season turfg rass in the South Florida region indicate that ET A generally occurred below ET O in the mid 1960's (Jones 1984) K C values ranged from
27 0 .85 to 0 .92 year round. This consistency suggests the potential for incorporating a constant K C into irrigation conservi ng scheduling year round for this region. However, while results indicating that a single crop coefficient based on an annual average may be suitable for irrigation scheduling in South Florida, results from other regions differ, such as those from Georgi a (Carrow 1995) the more arid Nevada (Devitt et al. 1992) or Arizona (Brown et al. 2001) In these regions crop coefficients suitable to these climates require seasonal or even monthly adjustment, reinforcing the need to provide regional ET data for ca lculation of crop coefficients. In the humid southeastern and Gulf coast climates, warm season turfgrass ET A rates are significantly lower under humid conditions than reference ET (Carrow 1995) Soil properties are also a significant factor in determining root expansion. Th ese results reinforce the importance of regional ET O (Carrow 1995) In Florida, St. Augustine based irrigation controllers had 20 59% reduction in water use from maximum ET (McCready et al. 2009) Aside fr om these studies, crop coefficient data on warm season turfgrasses in humid climates, specifically i n Central Florida has not been reported (Irmak et al. 2003a) Haley et al. (2007) also ca lled for more work in this area, reinforcing the idea that crop co efficient values for turfgrasses in Florida have not been documented. Two ET irrigation controllers were compared along with soil moisture sensor based irrigation controllers, standard timers with rain sensors, and standard timers alone for turfgrass i rrigation management (McCready et al. 2009) The standard timers alone irrigated on a set schedule of two days per week. The rest were controlled by the data sensors they employed. Turfgrass quality was visually rated using the National
28 Turfgrass Evalua tion Program (NTEP) ET O was derived from a nearby weather station using the Penman method. When ET A was compared between plots, variability was significant. Pre determined K C values, provided by the manufacturer of the ET controllers, were used to corre ct ET O and integrated into the scheduling of irrigation events for the ET controllers. All controllers demonstrated a significant water savings over the standard timer by itself. Coupled with better than acceptable visual ratings, the Toro ET controller had the highest water savings of 62%. However, there were problems with the programming of the other ET controller which produced less than acceptable turf quality ratings. This indicates that differences in controller set up and human error can still aff ect controllers based on ET. This study also demonstrates the direct application of crop coefficients in scheduling irrigation events and the potential for superior water savings compared to other typically used controllers in the residential landscape. In California, four ET controllers were compared for ease of use and accuracy for scheduling irrigation events based on real time climatic data and plant factors (Pittenger et al. 2004) Plant factors are a term used in this study to describe crop coeffic ients for ornamental landscape plant material and used in correcting ET O for residential irrigation scheduling. Results were highly varied, with one controller proving easier to use, while others were commented to require a professional to set it up. Acc uracy was also highly variable with only one providing relatively accurate scheduling, while another grossly over irrigated, and yet another under irrigated. It was concluded that while the use of ET controllers possess the potential for significant water savings, they are still subject to inaccuracies based on design, calculation, and human input. More research involving
29 direct application of these controllers needs to be done to help in bring ing user friendly, reliable, and accurate ET irrigation manage ment into the mainstream usage. K C Use in Established and Mixed L andscapes (K L ) Applying the crop coefficient method to urban landscapes is difficult. Typical urban landscapes are much small er than agricultural field s yet when an urban landscape is viewe d as an entire neighborhood ; it could be likened to an agricultural field. However, there is currently no system for collaboration between neighbors in an urban setting concerning irrigation. Urban landscapes normally contain a mixture of species, as opp osed to a single species. Since each species could have an individual crop coefficient, measurement and calculation of K C becomes more challenging. Finally, both agricultural crops (Allen et al. 1998) and nursery crops (Beeson and Brooks 2006a) are usual ly irrigated in a manner that will maximize yield and minimize time to marketable size. In established urban landscapes though, the goal is not to maximize yield or growth but simply maintain aesthetically appealing and healthy plants in a sustainable lan dscape setting (Sachs et al. 1975) This requires different irrigation models than those used to reach the levels of growth desired in agriculture and nursery production. Therefore concerning landscapes, we must adjust the concept of the crop coefficien t to align with aesthetics and sustainability as opposed to agricultural and production goals. The conc ept of a landscape coefficient ( K L ) was born from this idea. E arly work in California suggested that established landscape plantings can survive at ac ceptable aesthetic levels with irrigation below that generally accepted or typically employed (Sachs et al. 1975) Sachs evaluated shrub and ground cover plantings that were allowed to becom e established over a period of two y ears for the ground covers an d five years for the shrubs (Sachs et al. 1975) Irrigation was
30 performed using flooded trenches and applied at high volume (>9cm) at three levels: bi monthly, monthly, and none at all. He found that many species performed acceptably with bi monthly or even no supplemental irrigation. This may have been due in part to the high water holding capacity of the clay soils in that region. Although a coefficient was not discussed, and irrigation was not based on ET, this was the first published research conce rning reduced irrigation in established landscapes. The results prov ide early evidence that species specific irrigation and plant grouping in the landscape could reduce and normalize irrigation requirements for aesthetic purposes. This would have the added benefit of reducing pruning and fertilization needs. Later Sachs revisited results from one of his mid comparing actual water applied via irrigation to pan evaporation measurements for ET O (Sachs 1991) These hedgerows were established in 1965, then six years later were subjected to irrigation levels at 100% ET O, a n unspecified fraction of ET O and finally zero. Again shrubs not only survived, but pruning of excess growth was minimized because shoot growth has a direct correlation with ir rigation frequency and subsequent soil available water (Sachs 1991) There was discussion that leaf temperature could be used to determine plant water needs and subsequent irrigation requirements, but Sachs concluded that high wind speeds would negate thi s assumption. In 2004, newer research was published on the potential water savings by irrigating established landsca pes based on a percentage of ET O Plots were established with a mixture of plant materials from thirty genera Aesthetic performance was observed at rates of 0.36, 0.18, and 0.0 of ET O (Shaw and Pittenger 2003) Of the 30 genera eight had acceptable performance at 0.0 ET O Thirteen genera
31 demonstrated acceptable aesthetic levels at the 0.18 ET O level. Only two genera Hibiscus and Ligus trum still appeared under irrigated at 0.36 of ET O These results provide additional evidence for genus or species relevant K C values in the use of a mixed landscape coefficient. They also suggest that an accurate K L is achievable if plants are grouped based on water needs, or irrigation is based on the K C of the most water needy plant. The authors stressed the importance of further work in this area in order to clarify the differences between species specific K C and group K C M ixed landscape water us age based on ET was finally addressed in 2010 In pursuit of K L soil moisture sensors and in ground gra vimetric lysimeters with vacuum assisted leachate removal were employed (Pannkuk et al. 2010) ET A was ca lculated from sensor data, with ET O calculate d using the Penman Monteith equation. Landscapes were established in two locations, one in College Station, TX and the other in San Antonio, TX. They consisted of St. Augustine turfgrass only, St. Augustine turfgrass and Shumard Oak tree, Shumard Oak tre e only, native grasses only, and Shumard Oak tree and native grasses. The results were greatly affected by soil salinity levels in College station, as well as by unusually low precipitation amounts that were more than 80% below average in San Antonio duri ng the two year study. Overall native grass landscapes had a low coefficient of 0.3 in College station, while San Antonio's native grass landscapes had a much higher coefficient of 0.61, with a peak of 0.8 in the later part of the year. St. Augustine on ly and oak only had coefficients of 0.34 and 0.21 in College Station, with values of 0.52 and 0.43 in San Antonio respectively. The tree and turfgrass results
32 were similar to turfgrass only, while the tree and native grass mix was similar to native grass only. The authors explained the ir results by stating that the taller native grasses had more leaf area due to height, and may be opportunistic water users in favorable settings (Pannkuk et al. 2010) This brings up the discrepancy in using agricultural and nursery production ET calculations because the big leaf model may not be as accurate in varying landscape settings where height must also be considered when calculating leaf area available for transpiration. They speculated that a K L of 0.7 may save w ater, but could not support this based on field data. It was also suggested that a seasonal structure of K L would be 0.5 for early in the year, 0.6 for mid year, and end at 0.7 for later in the year (Pannkuk et al. 2010) Additional research was recommend ed to compare aesthetics and irrigation at levels below ET O Deficit Irrigation Definition and I mportance Deficit irrigation can be defined as the practice of irrigating agricultural crops, container plants, or established landscapes at less than 100% ET O It can further be defined as the practice of irrigating to a lower percentage of a known crop coefficient or lower than a known landscape coefficient. A useful percentage can be found by observing individual crop, plant, or landscape performance under pre set deficits of K C or K L Performance can be measured by yield, growth, or aesthetic rating (McCready et al. 2009) D eficit irrigation has been successfully employed in both agricultural production of maize (Kang et al. 2000) and container production of woody ornamentals (Beeson 2006) achieving equival ent or better yields and growth. D eficit irrigation also
33 has the potential to meet the goal of maintaining an aesthetically pleasing landscape while at the same time reducing residential water use. Eff ect on A esthetics Container plants such as Viburnum odoratissimum a commonly used hedge in Florida landscapes, were shown to achieve acceptable growth and maintain acceptable aesthetic quality under deficit irrigation (Beeson 2010) when canopy closure was taken into account. In Florida, coleus subjected to deficit irrigation levels also maintained acceptable aesthetic levels (Scheiber and Beeson 2007) In Colorado, several herbaceous annual ornamentals irrigated based on deficits of ET O provided mixed re sults. Some species such as Impatiens walleriana only did well at 100% ET O while others such as Lobularia maritima and Pelargonium x hortorum did well with 25 to 50% of ET O (Henson et al. 2006) In California, it was demonstrated that ornamentals could b e grown in the landscape and subjected to irrigation levels that could be considered deficits, while still maintaining acceptable aesthetic levels (Sachs et al. 1975, Sachs 1991) Cool season turfgrasses maintained in Colorado under deficit irrigation regim es showed acceptable aesthetic quality up to a 27% reduction of ET O (Feldhake and Butler 1984) However, a significant decline in quality was noted beyond the 27% reduction mark. Although St. Augustine is not known to be a drought tolerant turfgrass, it has shown some tendency toward physiological adaptation to drought stress. In studies of dehydration tolerance, 'Texas common' St. Augustine demonstrated a high dehydration tolerance (Beard 1989) Further, St. Augustine achieved acceptable root growth un der deficit irrigation during the establishment period (Sinclair et al. 2011) This indicates the
34 possibility of a wider range of tolerance for warm season grasses such as St. Augustine, under deficit irrigation regimes, for maintaining acceptable aesthet ic levels. Growth Versus A esthe tic Q uality Woody ornamental trees demonstrated controlled growth with a strong tolerance to deficit irrigation, some even with an increase in quality due to shorter internodes (Cameron et al. 2006) The rate of shoot gro wth is directly correlated to the rate of irrigation. As irrigation frequency increases, so does shoot growth (Stabler and Martin 2000) Although many plants are installed in the landscape for their desirable drought tolerances, they are often overwatere d and subsequently require more frequent pruning. Less frequent irrigation results in reduced shoot growth while maintaining vigor (Sachs et al. 1975) Widespread application of deficit irrigation could result in reduced maintenance requirements by reduc ing the need for frequent pruning. Aesthetic evaluation for turfgrass finds a standard in the National Turfgrass Evaluation Program (NTEP) (Morris and Shearman 2008). This program provides a rating scale of one to nine, nine being perfect. A visual ratin g guide is used, with ratings one through eight pictured. Since 9 is the theoretical ideal, it is not pictured on the rating sheet. At the NTEP website, a rating of six or better is said to be commonly accepted as adequate. However, the authors go furth er to explain that quality ratings differ among turfgrass types, and that a minimum of six for one species may not necessarily be the minimum for another. A study in Florida assigned the minimum acceptable rating for St. Augustine at five, and this was th e value employed in this research (McCready et al. 2009).
35 Application of K L Residential L andscapes Simplification of landscape irrigation is necessary to compensate for poor homeowner management of automated irrigation systems. Maintenance, calibration, accuracy, and seasonal adjustments are just a few areas where property owners fail in managing their irrigation systems (Dukes 2011) There have been numerous attempts at simplifying the process by recommending "deep and infrequent" irrigation, providing depth of irrigation conversions for timers, and incorporating irrigation calculators in system F.A.W.N. provides up to date statewide climatic data to assist homeowners in making seasonal adjustment s to irrigation schedules with limited effort (Dukes 2011) Still, improper scheduling of residential irrigation remains prevalent (Baum 2005) The purpose of determining ET rates and K L values is so that efficient and effective irrigation of suburban la ndscapes can take place. The typical residential landscape contains a mixture of trees, shrubs, ornamentals, and turf. If a mixed landscape can be irrigated based on turf area or mixed landscape area, it would eliminate the need to determine crop coeffic ients of each plant and then combine them. Potential of ET Based I rrigation ET based irrigation controllers have become more available to homeowners in recent years. ET based controllers offer automated irrigation scheduling based on real time climatic d ata. This data can be recorded throughout the region and communicated wirelessly to a controller integrated into the typical residential irrigation system (Stacia and Dukes 2011) In theory, t his type of system provides the climatic and seasonal adjustme nts required for efficient irrigation without the need of homeowner input once
36 installed. When properly installed and managed, ET based irrigation controllers demonstrated a 43% annual water savings compared to other types of controllers (Stacia and Dukes 2011) In San Antonio Texas, the potential for ET based irrigation was examined by recruiting homeowners to participate in a study where landscapes were irrigated based on 100%, 75%, and 50% of ET O for one full year. During this time the participants we re asked to rate the turf based on a one to five scale, one being excellent; five being poor. On average, turf could be managed effectively with the lower coefficients of 0 .75 and 0 .5 0 while staying below the three rating for the summer, and below the tw o rating otherwise (Pope and Fipps 2000) Classes and communication channels were established which encourage d enthusiasm for the project among the participants. There are advancements in ET related technology that have the potential to go beyond what is currently available. Artificial intelligence is being used to produce ET O that can be transmitted to undeveloped regions that cannot produce ET O data by measuring climatic data locally (Adeloye et al. 2011) These types of research and results clearl y indicate the need not only for an interactive program to encourage homeowner participation but also the need for accurate data concerning the irrigation o f landscapes based on the ratio of actual plant water use and real time climatic data.
37 CHAPTER 3 MATERIALS AND METHOD S Eighteen drainage lysimeters were constructed into a hill side such that only the west facing wall was fully exposed (Fig ure 3 1). Lysimeter s were installed in three blocks of three lysimeters in two rows oriented north south. Figure 3 1 Drainage lysimeters on the west row after construction. The pipe exiting the bottom was connected to a system des igned for leachate collection. The white pipes on the right were for irrigation. Photo courtesy of Scott Simpson. Inside dimensions of each lysimeter were 3 .3 m north south and 4.1 m east west for a total surface area of 13 .33 0. 0 5 m 2 The bottom of each was sloped towards the center with a single drain pipe exiting the west wall for drainage c ollection. Inside walls were painted with basement wall waterproof paint ( Seal Krete DampLock, Convenience Products, Auburndale, FL ) in two coats The drainage system consist of a central junction box over the center drain hole with geo textile sock cove red 10 cm corrugated drain pipe extending to diagonal corners (Fig ure 3 2). These were covered with rock,
38 textile cloth, and coarse sand before backfilling with native soil; Apopka fine sand series came from sand loam marine sediment, usually has loam; a n d are siliceous, hyperthermic, grossarenic, and p aleudults (USDA 1989) Lysimeters are 147 cm deep along the outside edge and 155 cm deep in the middle. Figure 3 2 Installed drain line prior to covering with rock and sand. Photo courtesy of Richard Beeson. For this project, only the nine spatially adjacent lysimeters on the north end of the two row s were utilized. After soil was leveled, turfgrass irrigation was installed. This consisted of 1.9 cm polyvinyl chloride (PVC) pipe buried around the inside perimeter of each lysim eter. Pop up spray heads (PRO S 06 10A, Hunter Industries, Inc., San Marcos, CA) were positioned at each corner and in the center along the north and south sides. The northeast pop up was inset 0.9 m from the cor ner to accommodate the shrub bed. Turfgrass and woody plant irrigation were controlled separately using two 24 VAC solenoid valve s (SRV, Hunter Industries, San Marcos, CA), each regulated with
39 a 167 kPa pressure regulator (PMR MF 25, Senninger Irrigation Inc., Clermont, FL). Woody plant irrigation was distributed using 1.9 cm black polyethylene tubing (I.P.S flexible PVC tubing, The Toro Company, Bloomington, MN) to both the tree and shrubs. The tree irrigation employed two 3 0 cm tree stakes with 13.2 L hr 1 nozzles and inverted cone spreaders (Jain Irrig ation Inc., Fresno, CA). Shrub irrigation employed the same spray stake assembly but with four 7.1L hr 1 nozzles. One stake was placed between each plant and along both outside edges. I rrigation val ves were positioned on the outside the west facing wall. One water meter (C700 SF, Elster Amco, Ocala, FL) with an electronic counter (123 counts L 1 ) was installed after each valve and before the pressure regulator In August 2009, the drainage lysimeter s were randomly assigned to treatments. Since two independent experiments were to be conducted concurrently, the site was split spatially based on what was considered to result in the most uniform microclimate per experiment. This experiment was designat ed to occupy the northern six lysimeters of the west row and northern three lysimeters of the east row (Fig ure 3 3 ). Figure 3 3. A v iew of lysimeter project looking from south to north. Photo courtesy of Scott Simpson.
40 The experiment consi sts of nine identical mixed landscape plantings. Each plot Magnolia grandiflora ) planted in the center of eac h lysimeter east west and 1.1 m north of the south wall It was surrounded with an area of 1.1 m 2 of mulch. A shrub hedge of sweet viburnum ( Viburnum odoratissimum ), mulched 1.0 m wide (north south) and 2.0 m long (east west) in the northeast corner completed the layout. This hedge consisted of two plant s. The 9 lysimeters were spatially divided into three replicate blocks. Lysimeters within each block were randomly assigned one of three deficit irrigation (DI) treatments. These consisted of counting 90% of daily ET O towards an accumulated irrigation depth (90% DI), counting 75% of daily ET O tow ards an accumulated irrigation depth (75% DI) or counting 60% of daily ET O toward an accumulated irrigation depth (60% DI). Irrigation was applied to a lysimeter when the cumulative irrigation depth exceeded 1.90 cm. Thus treatments applied 10%, 25% and 40% less irrigation for the 90%, 75% and 60% DI, respectively, than that calculate to replace ET A lost by a well irrigated maintained cool season turfgrass. On 9 September 2009, the magnolias and viburnums were transplanted into each of the lysimeter s according the layout described above (Figure 3 4). The magnolias were approximately 1.8 m tall and 3.8 cm in caliper measured at 15 cm above ground. The y were transplanted from 51 cm Root Control Bags (Root Control, Inc, Stillwater, OK.) Viburnum s we re transplanted from 11.4 L containers. At transplanting, backfill soil was watered in to insure good contact between root balls and the soil. Excess soil was removed from lysimeters. The micro irrigation described above was installed in
41 each lysimeter on 10 September and both species were irrigated to establishment using micro irrigation on alternate days thereafter through late May 2010. After transplanting, tree and shrub dimensions were measured tri weekly and used for aesthetic evaluation Shrub measurements consisted of average width north south (perpendicular to the long axis of the hedge) and east west (parallel to the long axis of the hedge), and average hedge height. Hedges were pruned as needed to maintain maximum dimensions of 1 m north so uth and 2 m east west. Magnolia canopies were measured at the widest width and the width perpendicular to the widest width T ree height to the terminal bud was also measured In addition trunk circumference was measured 15 cm above the soil. Magnolia trees were not pruned. Canopy widths were multiplied to calculate a horizontal projected canopy area (PCA, m 2 ) for each hedge and tree. PCA was multiplied by height to calculate canopy volume (GI,m 3 ) Figure 3 4. A v iew of an example of one of the ni ne lysimeter s shortly after transplant ing Photo courtesy of Scott Simpson.
42 The remaining surface area (75%) was covered with St Augustine turfgrass ( Stenotaphrum secundatum [Walt.] Kuntze) (Figure 3 4). Fresh cut sod was delivered from a turfgrass farm near Lake Okeechobee, FL on 24 September 2009. This turfgrass was cut from a sand soil to be compatible with the sand soils in the lysimeters. It was installed on 25 September, and required an aggressive irrigation schedule to faci litate establishment. The turfgrass system irrigated three times per day for 30 minutes to wet both the sod and soil beneath. The turfgrass irrigation schedule was changed to four times per day for 15 min utes on 29 S eptember, and then reduced back to thr ee times per day on 12 October. It was further reduced to twice per day on 15 October, to once dail y on 20 October, to once every two days on 2 November, to every three days on 11 December, and finally every four days on 6 December. Turfgrass in each ly simeter was first fertilized at 453 g N per 93 m 2 with a granular fertilizer ( Vigoro All Purpose Plant Food 10 10 10, Vigoro, Sylacauga, AL) on 30 October 2009 using a 46 cm wide Accugreen drop spreader ( Scott's Marysville, OH) Subsequent t urfgrass fert ilization occurred in 2010 on 7 May and again in 13 July. Fertilization continued throughout the experiment based on original and then more recent Institute of Food and Agricultural Sciences ( IFAS ) residential turfgrass general recommendations (IFAS publi cations SL21 & ENH1089). Pesticides and fungicides were applied as needed to control chinch bugs ( Blissus insularis Barber) and gray leaf spot ( Pyricularia grisea ) respectively. The magnolias and viburnum were first fertilized after transplanting on 26 F ebruary 2010 with a slow release nitrogen granular fertilizer (16 4 8 ProSource One, Agro Distribution, Memphis, TN) corresponding with the first
43 pruning of the viburnum. Initial measurements for both magnolias and viburnum for 2010 occurred on 15 March. H ard freezes occurred each morning from 7 to 10 December 2009, freezing all grass blades Hard freezes occurred again eight of the first twelve days of January 2010. Woody plants were not injured, but there were no green leaves in the turf grass by 12 January. The rest of January through March remained unusually cold, with several more freezes By early July 2010 it was concluded that too much of the turfgrass did not recover from the winter and had to be replaced (Figure 3 5). Dead sod was removed 15 and 16 July 2010. For some lysimeters up to 25% of the turfgrass was replaced, most were around 15%. Turfgrass was replaced on 19 July 2010. Irrigation remained under computer control. To facilitate establishment, the new sod was lightly sprayed wit h a hose by hand twice daily for four weeks when it did not rain. Figure 3 5. Freeze damage to turfgrass from winter of 2009 2010, shown here in mid July 2010. Photo courtesy of Scott Simpson.
44 The system to quantify leachate was sheltered and sealed from the elements as described below. A dry well was installed below the 5 cm drain pipe to support the measuring device and to allow for its greater depth below the drain pipe (Fig ure 3 6). This device consisted of an upper collection vessel constructed from a 15 cm PVC schedule 40 cap and a 15 cm length 15 cm diameter schedule 40 pipe with a ~1.5L volume that drained into a weighing vessel through a normally open valve ( Series 8262, ASCO, Florham Park, NJ ). The weighing vessel was similar to the collection vessel, but was constructed using a 45 cm long PVC pipe with a normally closed valve ( DSVP11 8PX8SFX1, Deltrol Controls, Milwaukee, WI ) at the bottom. It was suspended from a 22.7 kg load cell (SSM AJ 50, Interface Inc, Scottsdale, AZ) has an overflow drain near the top to channel excess water below the drain valve should leachate exceed the capacity of the system. The maximum capacity was 1.5 L per two minute cycle. Figure 3 6. Measuring device for quantifying leachate from each lys imeter. The upper vessel collects water which drains to the weighing vessel (bottom) through the normally open valve (green and gold). The system is supported by the T frame above the dry well, shown by arrow. Photo cour tesy of Scott Simpson.
45 In operation, a data logger (CR1000, Campbell Scie ntific Inc., Logan, UT) measured the leachate volum e every two minute s and activated the syste m when a minimum of 1.001 L had been co llected. Twelve VDC power was shunted to the valves to close the upper valve, and open the drain valve to evacuate the weighing vessel. When water drainage wa s <5 mL per ten sec onds the 12 VDC power wa s turned off and the system is reset. This system has an overall capacity to measure aroun d 12.7 cm of rainfall every twenty four h ou r s assuming the soil volume is at field capacity initially. Each measuring device was enclosed within a structure consisting of a tin roof and cement board sides (Hardie board, James Ha rdie, Mission Viejo, CA) w hich wa s sealed with silicon and expanding foam to exclude rainfall, dust and blowing sand (Fig ure 3 7). Figure 3 7. Sealed shed which house d the leachate measuring device s pring 2010. Photo courtesy of Scott Simpson. Two AM16 32 multiple xers (Campbell Scientific, Inc.) with associated wiring and terminal strips were installed for measuring mass of the weighing vessels. Data logger
46 controlled remote relays (SDM CD16AC, Campbell Scientific, Inc.) were installed for control of the leachate measuring and irrigation valves. Electric al power was supplied by 24 VAC and 12 VDC transformers ( Figure 3 8 ). Figure 3 8. Data collection and irrigation control system for the drainage lysimeter project. Photo courtesy of Richard Beeso n. The lysimeter system was controlled by an original algorithm that achieved operational status on 26 May 2010, during the plant establ ishment phase (R Beeson, pers. c omm). H ighlights of the algorithm are described below. Each lysimeter was treated ind ependently for all operations. The system weighed each weighing vessel every two minutes and the amount drained was added to a running daily total. At 5 am (Eastern Standard Time, EST), the running total was stored, reset to zero and data collection bega n anew. At midnight, cumulative daily rainfall and reference evapotranspiration was calculated using
47 Camp b ell Scient ific Inc., Application Note 4 and transferred from the onsite weather station located in a grassy field ~100 m west of the site to the CR10 00 via a common desktop computer (model W3609, eMachines, Irvine, CA ) The weather station consisted of a LI200X pyranometer (Li Cor Inc, Lincoln, NE), a CS215 temperature and relative humidity probe (Campbell Scientific, Inc.), a Wind sentry set (03001, R.M. Young Co., Traverse City, MI), and a tipping bucket rain gauge (TE525, Texas Electronics, Dallas, TX) connected to a CR10X data logger (Campbell Scientific, Inc.). Should the transfer fail, a backup of 0 .46 cm of ET O was assumed by the algorithm. Da ily ET O was then multiplied by one of three treatment coefficients The control coefficient was 0.9 0 This was the K C established previously, that supported aesthetically pl in Central Florida ( M. Dukes, pers. comm ). The o ther two coefficients were hypothesized to be a moderate reduction in irrigation rate (K L = 0.75) that would likely produce acceptable quality, and a severe reduction (K L = 0.6 0 ) that would likely result in unacceptable quality. Cumulative totals of the a djusted ET O cumulative ET O for a lysimeter exceeded 1.90 cm, the actual cumulative ET O was multipl ied by the turf grass area (10.0 m 2 ) to calculate the volume of irrigation to apply to the turf. Similarly the same actual cumulative ET O was multiplied by the projected canopy area (widest width x width perpendicular, PCA) of the tree and shrubs and by their respective K C 0 for the shrubs). These volumes were summed and applied using the independent woody plant irrigation system. Daily rainfall was subtracted from each cumulative ET O assuming a rooting depth of 30 cm for the turfgrass. For this soil type, this depth would retain only the first 6.25
48 mm of a rainfall event ( Orange Co unty Soil Conservation, (USDA 1989) Consecutive days of rainfall could reduce a cumulative ET O to no lower than minus 0.625 cm. Turfgrass irrigation occurred beginning at 0500 h ou r EST. Woody irrigation was applied beginning at 0700 h ou r EST. Irrigation was delayed until near sunrise to minimize the time turfgrass was wet to reduce incidence of disease, and for better irrigation uniformity due to normally calm or low wind speeds. Turfgrass was mowed with a push mower (GVC 160 Americ an Honda Motor Co. Inc., Alpharetta, GA. ) equipped with twin blades and a discharge bagging system. The blades were sharpened regularly. Turfg rass runners extending outside lysimeter surface area s or into the mulch beds were clipped by hand and included in clipping harvest Turf clippings were harvested at each mowing, kept separate by lysimeter and quantified after being dried to a constant weight at 65 C. After drying the clippings were weighed on a digital scale ( PB5001, Mettler Toledo Inc, Columbu s, OH ) and measurements recorded. Turfgrass was first mowed on 3 December 2009, but not again until 9 April 2010. From there mowing occurred about every 2 weeks until late May. Thereafter it was mowed weekly until October where mowing was reduced to b i weekly intervals through November 2010 The turfgrass did not require mowing in December 2010, January, and most of February 2011. In Late February mowing resumed on monthly basis until May, and then a bi weekly schedule until the end of May, which was the completion of this study. Turfgrass visual ratings were performed on a monthly basis by three people to evaluate the aesthetic quality of the turf. Evaluations were made using NTEP, the
49 National Turfgrass Evaluation Program (Morris and Shearman 200 8) Evaluations are made based on a one to nine scale, each level associated with a picture to compare to the actual turf. On this scale, 1 is dead and nine is perfect. Since nine is the theoretical ideal, it is not pictured, and ratings were taken from a comparison with eight pictures (Figure 3 9). Condition of the turf was matched to one of the numbered pictures and the number was assigned for the rating. Figure 3 9. Turfgrass visual rating guide used to evaluate aesthetic quality of St. Augustine turfgrass. It is based on a one to nine scale, nine being perfect and therefore not shown. Photos courtesy of Michael Dukes.
50 In order to take into account the effects of all seasons, data collection took place for a period of on e full calendar year from 1 June 2010 through 31 May 2011. Visual turf ratings, turf dry mass, and woody growth factors were analyzed Growth factors for Viburnum odoratissimum were collected for projected canopy area (PCA) growth index (GI) and height Growth factors for Magnolia grandiflora were collected for PCA trunk circumference at 15 cm above the soil and height. All data collected was compared to the three deficit irrigation levels of 60%, 75%, and 90% of ET O ET A and ET O were organized in to monthly results. Monthly data was grouped up to the point where there was a lull in input and output volumes that allowed for a reasonable calc ulation of ET A based on lysimeter hysteresis. The next significant input was also taken into account as a br eak point in the data. Because of lysimeter hysteresis drainag e was still occurring several days after a rain or irrig ation event. This A calc ulation until the next sig nificant rain or irrig ation event occurred Ther efore, if the drainage ran over into the next month for a few (less than five) days, it was included i n the current month s calculations. To maintain consistency, the same pattern of going into ea rly days of the next month was used throughout. The data wa s analyzed using SAS version 9.2 (SAS, Inc, Cary, NC) Data analysis for water input and turf ratings were conducted using one way ANOVA. Data analysis for growth data was conducted using repeated measures using split plot and then mean separations using Fisher's protected LSD.
51 CHAPTER 4 RESULTS Rainfall for the year of data collection from the onsite weather station was compared to the 14 year average (Figure 4 1) from the Mid Florida Research and Education Center Apopka FAWN (Florida Automated Weat her Network) site. Total rainfall during the experimental period was not significantly different ( P> 0.05 ) from the 14 year average Annual rainfall at the research site was 104.67 cm. The 14 year FAWN average was 119.49 cm. Yet when rainfall was examin ed on a monthly basis, there were some obvious differences. October 2010 had 0 cm rainfall compared to the average of 8.63 cm while June rainfall was also 8 cm below normal (Figure 4 1) U nusual rainfall spikes occurred during the dry season in January and March of 2011. January rainfall total was 12.34 cm compared to an average of 5.34 cm. March rainfall totaled a much higher 20.73 cm when compared to an average of 8.49 cm. Irrigation input and total water input differed ( P < 0 0001) among DI (Table 4 1). Total irrigation input was defined as water entering a lysimeter by both turf and woody irrigation systems. T otal water input was the sum of total irrigation input and rainfall. For both, the amount of water input into a lysimeter increased with decreasing percentages of deficit irrigation. Table 4 1. Annual irrigation input totals (L) and total water input (L) totals for each deficit level treatment. Total rainfall per lysimeter was 13,410 L Deficit irrigation z Total irrigation i nput y Total water input x 60% 6,333 a w 19,743 a 75% 9,058 b 22,468 b 90% 11,857 c 25,267 c z Percentage of daily ET O summed to trigger an irrigation event when cumulative ET O > 1.9 cm. y Total irrigat ion input is the total amount from woody and turf irrigation systems. x Total water input is the total from rainfall, woody irrigation, and turf irrigation systems. w Values are the mean of 3 lysimeter replications. Means within a column with the same let ter are not different at P =0.05 based on Fisher's Protected LSD.
52 Figure 4 1 Comparison of monthly total rainfall (cm) during the data collection period of 1 June 2010 to 31 May 2011. The onsite weather station wa s 100 m west of the lysimeter site. The FAWN (Florida Automated Weather Network) average rainfall was based on a 14 yr history of the station within 500 m of the onsite station. Monthly mean daily ET A ET O and K L are p resented below (Table 4 2). The 60% DI had a single digit low of 0.09 cm ET A per day in February. ET A was highest at 0.68 cm per day for the 90% DI in June. The annual average daily value for ET O was 0.42 cm. ET O variations also followed seasonal weath er patterns, with lower values in the cooler months of November through February. Monthly K L values varied among treatments depending on the month ( P <0.05). Values ranged from a low in February of 0.27 for 60% DI up to a high of 1.49 in March for 90% DI (Figure 4 2). March had no differences among treatments. For June August January and April there were some slight difference s among treatment s For June and August the 75% DI was lower than the other two, whereas in January and April, K L values for t he 60% DI were lower than the other two. On the other hand, in
53 July September, October, February and May differences among deficit irrigation treatments were considerable The 60 and 75 % DI were similar in November and December However, K L values at 90 % DI were much higher for November and December Table 4 2. Mean monthly values from 1 June 2 010 to 31 May 2011 for daily ET O daily ET A (cm) and K L b y irrigation deficit level (DI) 60% DI z 75% DI 90% DI ET A y K L x ET A K L ET A K L ET O w Jun 10 0.66 v 1.22 0.64 1.14 0.68 1.25 0.55 Jul 10 0.41 0.78 0.34 0.65 0.54 1.03 0.52 Aug 10 0.41 0.86 0.37 0.79 0.42 0.90 0.47 Sep 10 0.41 0.85 0.36 0.83 0.50 1.15 0.45 Oct 10 0.15 0 .46 0.34 0.84 0.29 0.74 0.38 Nov 10 0.20 0.71 0.20 0.71 0.29 1.03 0.28 Dec 10 0.15 0.63 0.15 0.67 0.19 0.79 0.23 Jan 11 0.29 1.19 0.33 1.33 0.31 1.29 0.25 Feb 11 0.09 0.27 0.17 0.52 0.22 0.67 0.34 Mar 11 0.62 1.45 0.61 1.46 0.63 1.49 0.42 Apr 11 0.22 0.38 0.29 0.50 0.38 0.61 0.59 May 11 0.36 0.62 0.44 0.73 0.53 0.94 0.58 Averages: 0.33 0.78 0.35 0.85 0.41 0.99 0.42 z Percentage of daily ET O summed to trigger an irrigation event when cumulative ET O > 1.9 cm. y Actual evapotranspiration of a lysimet er with St. Augustine turfgrass, two Viburnum odoratissimum shrubs, and one Magnolia grandiflora calculated by the difference of water out subtracted from water in. x Landscape coefficient calculated as a ratio of ET A to ET O w Mean monthly values for dai ly ET O v Values are the mean of 3 lysimeter replications. K L values began high in June 2010, and then generally decreased through December. The September 90% DI value was much higher than 60 and 75% DI, while the October 60% DI value was much lower th an 75 and 90% DI. 2011 saw more pronounced differences. January and March were much higher than February and A pril, with peaks upwards of 1.2; w hereas dips to 0.6 or lower occurred in February and April. During the months of April to May K L showed a sl ight increase. Differences among treatments were more pronounced in 2010 than during the w inter to spring
54 seasons of 2011; however oscillations in the K L values varied much more greatly in the spring. Figure 4 2 Comparison of monthly K L by treatment during the data collection period of 1 June 2010 to 31 May 2011. Lines represent different deficit irrigation treatments as a percentage of ET O that was counted toward triggering an irrigation ev ent at >1.9 cm. Each point is the mean of 3 lysimeter replicates. The plots were mowed 24 times over the year (Figure 4 3), with 17 occurring weekly during the peak growing period June through September (Figure 5 1). The plots only needed to be mowed 7 times during the remaining months, of which there was no mowing in December or January. There were no differences among treatments ( P >0.05) in dry mass harvested from the turfgrass at each mowing. Total mean dry
55 mass for the 60% DI was low est at 2.525 kg, while 90 % and 75% DI were 2.919 kg and 2.991 kg respectively However, as could be expected there was an effect of time of year (P<0.0001). Figure 4 3. Mean dry mass (g) measurements at each mowing for St. Augustine 'Floratam' turfgrass during the first year of deficit irrigation after plant establishment. Means represent nine treatment replicates across deficit irrigation treatments. Table 4 3. Visual ratings by seaso n for St. Augustine 'Floratam' turfgrass during the first year of deficit irrigation after plant establishment. Deficit irrigation z Annual y 60% 5.61 a x 75% 5.95 b 90% 6.31c z Percentage of daily ET O summed to trigger an irrigation event when cumulative ET O > 1.9 cm. y Annual mean visual ratings for the entire year 6/1/2010 5/31/2011. x Turf grass visual ratings varied by DI ( Table 4 3; P < 0 .0001 ). Average visual ratings decreased with decreasing DI, and were significantly different among all DI. Although there were differences in quality ratings, annual mean values for all treatment s
56 were still above the minimum aesthetic threshold of 5.0 established for St. Augustine turfgrass (McCready et al. 2009) Viburnum growth was similar among treatments ( P >0.05) for height, PCA, and canopy volume (GI). All three components of growth increa sed during the year ( P < 0.0 001 ) Shrub height increased from 0.7 m to 1.1 m (Figure 4 4), with most increases in height occurring August to September and May to June. Shrub PCA and GI followed similar patterns with peak growth occurring September to Octob er and April to May. Shrub PCA increased from 1.3 m 2 to 1.8 m 2 (Figure 4 5). Shrub GI increased from 2.9 to 6.3 m 3 (Figure 4 6). Irregularities in both the PCA and GI increases were the result of on demand pruning to generally maintain the hedge within dimensions of 1 m width north south and 2 m lengths east west. Figure 4 4 Height (m) measurements of Viburnum odoratissimum during the first year of deficit irrigation after plant establishment. Means represen t 9 treatment replicates across deficit irrigation treatments.
57 Figure 4 5 Projected Canopy Area (PCA) (m 2 ) of Viburnum odoratissimum during the first year of deficit irrigation after plant establishment Means represent 9 treatment replicates across deficit irrigation treatments. Figure 4 6 Growth Index (GI) (m 3 ) of Viburnum odoratissimum during the first year of deficit irrigation after plant estab lishment. Means represent 9 treatment replicates across deficit irrigation treatments.
58 Increases in magnolia trunk circumference were similar among DI ( P >0.05), and increased with time ( P < 0.0 001 ) Mean trunk circumference increased from 15.6 cm to 18.5 cm over the year (Figure 4 7). Circumference increased steadily from June to October, then was quiescent until March, when it began to increase through May. Figur e 4 7 Trunk circumference (cm) taken 15 cm abov e soil level of Magnolia grandiflora during the first year of deficit irrigation after plant establishment. Means represent 9 treatment replicates across deficit irrigation treatments. In contrast, both height and PCA of magnolia v aried among treatments depending on time of the year ( P <0.05). Overall annual mean tree height increased from 2.59 m to 3.17 m (Figure 4 8) At the b eginning of June the 60% DI trees were 5.3 cm taller than the 90% DI, and 12.3 cm taller than the 75% DI. Late J une 2010 through
59 early April 2011 height of 60 % DI trees remained taller than 90 % DI trees, which were slightly taller than the 75% DI trees. This was the same February through March. In April and for the rest of the experimental period, the 60% DI trees still remained taller than the other two, while 90% DI trees also remained taller than 75% DI trees. Figure 4 8 Mean height (m) measurements of Magnolia grandiflora during the first year of deficit ir rigation after plant establishment. Each mean represents 3 tree replications. At the beginning of June the 90% DI tree had only slightly greater PCA than the 60% DI, both of which were nearly 0.15 m 2 greater ( P <0.05) than the 75% DI. For 75% DI magnolias, PCA increased nearly 50% with bud break and branch growth from June to early July (Figure 4 9). The PCA for 75% DI increased the most during this period,
60 expanding from 0.933 m 2 to 1.367 m 2 an increase of 0.433 m 2 The 90% DI trees had a s imilar increase of 0.420 m 2 The increase for 60% DI trees was nearly half as much at 0.240 m 2 All treatments were unchanged through March 2011. Beginning in April 2011, growth increased dramatically with spring bud flush. Figure 4 9 PCA (m 2 ) measurements of Magnolia grandiflora during the first year of deficit irrigation after plant establishment. Each mean represents 3 tree replications.
61 CHAPTER 5 DI SCUSSION With a relatively warm fall in 2 009, the turfgrass had begun to grow after installation in September H ard freezes occurred each morning from 7 to 10 December 2009, freezing all grass blades. Hard freezes occurred again eight of the first twelve days of January 2010. Woody plants were not injured, but there were no green leaves in the turf grass by 12 January. The rest of January through March remained unusually cold, with several more freezes. Bud break on the viburnum (Beeson 2004) and magnolias (Beeson 1991) were several weeks lat er than normal Turfgrass exhibited signs of life in late March 2010. Neither the viburnum nor the magnolia had exhibited any shoot growth since transplanting in September 2009 until this point due to normal post transplanting allocation to root growth ( Scheiber et al. 2007) and onset of winter dormancy Magnolia shoot growth did not begin until mid April 2010. Despite the unusually cold winter 2009 2010, t h ere were no woody plant fatalities during the one year period of this experiment. Turfgrass dise ase and pest management was req uired to maintain healthy plots during the year. G aps in fine tuning these requirements combined with the unusually cold winter, followed by a hot dry spring result ed in the need for the repl acement of some turfgrass For the remaining ten months no problems were experienced even though rainfall was frequently below average Irrigation at all DI levels provided acceptable growth for all plant material in the mixed landscape. Although annual rainfall was similar to the 14 year average, monthly deviations greatly influence irrigation needs and leaching below the root systems. With no rainfall in October 2010, 100% of water needs wer e provided by irrigation. At this time, the
62 effect of the different DI levels became evid ent. The 60% DI required an average of 4.29 cm of supplemental irrigation in October while both the 75% and 90% DI required over 9 cm of supplemental irrigation for that period I f the 14 year average rainfall of 8.63 cm had occurred, then almost no sup plemental irrigation would have been needed. If historical rainfall data had been used (Haley et al. 2007) instead of real time weather data (McCready et al. 2009) then the turf plots would have been severely under watered. These findings underscore the need for real time weather data when calculating landscape water needs. Figure 5 1 Seasonal shoot and root growth pattern of warm season turfgrass (Turgeon 2002) Each tick on the horizontal represents one month. Turfgrass growth is moder ated by temperature, and typically occurs with in a range of 40 to 105 F (Beard 1989) T herefore, t urfgrass growth cycles follow seasonal temperatures. St. Augustine is a warm season C4 turfgrass. Warm season growth cycles increase in the March/April ran ge and decline in the September/October range (Figure 5 1) Warm season grasses typically experience max imum growth when
63 daytime temperatures are between 80 an d 95 F (Christians 2011) Even when rainfall is high during periods of low temperatures (Figure 4 1), turfgrass growth does not respond with an increase in rate (Figure 4 3). Based on this response irrigation should follow the same seasonal conditions that moderate turfgrass growth January and March K L values were outliers. These values were disproportionately higher than seasonal trends set October through December, and February. January and March both had higher than average rainfall (Figure 4 1). Low rainfall in November and December resulted in conditions that allowed the soil to dehydra te. Apopka fine sand is somewhat coarse and water can percolate through quickly if the soil has been well irrigated or there has been consistent rainfall. T he water holding capacity of soil that is allowed to dry out is much greater, and would take much longer to reach field capacity and begin draining (Kramer and Boyer 1995) This would cause the dehydrated soil to retain more wa ter during rain and irrigation reducing drainage and providing more plant available water. The reduced drainage would make t he ET A appear higher, since ET A was calculated as the difference between water input into a lysimeter and the volume recovered from drainage. ET O during these months was also lower due low sun angles and cooler temperatures. Since K L is calculated as a r atio of ET A /ET O this would cause higher values for K L The h igh K L value for the September 90% DI treatment resulted from a high average daily ET A value for 90% DI in September of 0.50 cm compared to 0.41 cm for 60% DI and 0.36 cm for 75% (Table 4 2). A sharp decline in mean turfgrass dry mass harvest also occurred during this period (Figure 4 3). Mean dry mass harvest for August were 843 g for 60% DI, 933 g for 75% DI, and 1030 g for 90 % DI; compared to
64 means for September that were 275 g for 60% DI, 4 21 g for 75% DI, and 402 g for 90 % DI. In September warm season turf grass is nearing the end of the peak growing season ( Figure 5 1 ), and growth slows. V isual turfgrass ratings were higher overall for the 90% DI, but there was no time x DI interaction ( P >0.05). Therefore, there is no statistical evidence that factors evolving from dry mass and visual ratings could provide an explanation for higher ET A in September. The opposite effect occurred in October with K L for 60% DI. The K L value was much lowe r than 75% and 90% DI because of low ET A Daily average ET A for October was 0.15 cm for 60% DI compared to 0.34 cm for 75% DI and 0.29 cm for 90% DI (Table 4 2). Turf dry mass for October was low for all treatments with 60% DI at 71 g, 113 g for 75% DI, and 99 g for 90% DI (Figure 4 3). M ean turf visual ratings were lower overal l for the 6 0% DI Again there is also no statistical evidence that factors evolving from dry mass and visual ratings could provide an explanation for the lower ET A in October. B etween April and September, turfgrass dry mass measurements had peaks just after periods of heavy rainfall (Figure 4 1), and dry mass harvest was generally greater during these months (Figure 4 3) However, when rainfall was higher than average in January and March, the turfgrass did not respond with increased dry mass measurements. This response follows the growth pattern of warm season turfgrass (Figure 5 1). This response also indicates that residential landscape irrigation applied above turf water ne eds is wasted (Haley et al. 2007) because turf water needs are typically very low in January and March (Table 4 2)
65 With equivalent rainfall among treatments (Figure 4 1) differences in growth among DI can be attributed to the effects of the irrigatio n frequency. The 60% DI turfgrass had the lowest dry mass for the year, while the 75% and 90% DI both had about 400 more grams more dry mass for the period. Extra irrigation given to the 75% and 90% DI plots encouraged unnecessary growth, to gain only mo dest increases in visual rating scores (Table 4 3) at a cost of nearly twice the irrigation (Table 4 1) to meet the increased DI. Turf visual ratings varied by DI level, with ratings for the 60% DI declining below that of the other two levels. Even then the average score was 5.61 (Table 4 3). This was still above the minimally acceptable level of 5.0 established for St. Augustine turfgrass assessments in a residential setting(McCready et al. 2009) A score of 4.3 in May for the 60% DI was the only valu e below this level. Considering the scores overall, the ratings achieved a high of 6.6. With a potential maximum of 9.0 on the scale that was used, ratings generally remain ed just above acceptable. It is possible that the ratings could improve if the sys tem was observed for a longer period of time, giving the turf more time to produce the dense stoloniferous spread within the turf area that St. Augustine is known for. Woody plants had consistent growth throughout the year, but for the most part did not v ary among treatments with the only exceptions being tree height (Figure 4 8) and tree PCA (Figure 4 9) For these variables tree PCA increase from June to May was greater for 60% DI at 0.32 m 2 than the 0.25 m 2 for 75% DI; and tree height increase was eq uivalent for 60 and 90% DI at 0.46 m 2 and higher than the 0.40 m 2 increase for 75% DI. Both variables demonstrate d responses w h ere less irrigation
66 resulted in greater tree growth over the year The physiological reaction to water stress of many woody pl ants is to increase root growth and slow top growth. T hen when water becomes avail able, th ey can produce faster top growth (Gilman 1990) T hose irrigated more frequently had more top growth (Scheiber and Beeson 2007) and this was also shown in turfg r ass (Sinclair et al. 2011) At lowe st DI, plants were likely encouraged to allocate more resources extending roots to find available water. Root mass was not quantified in this experiment But the behavior described could be attributed to higher root mas s. The concept of deep and infrequent watering encouraging deep root penetration was established by Sachs et al. (1975). When es tablished landscape plants were trench irrigated with > 8 cm of wat er, the soil profile was saturated several feet deep Altho ugh these plots were irrigated much less frequently they still maintained acceptable aesthetic levels. Similar results are reported here for the 60% DI. It appears magnolias in the 60% DI were prepared to take better advantage of available water than 75 % DI, producing more top growth over the year. When you combine the increased water holding capacity of dehydrated soil with the possibility of increased root growth from water stressed plants, you have the opportunity for periods of modestly greater shoo t and canopy growth exemplified by the 60% DI in this experiment The apparent adaptability of woody plants and turfgrasses to stress levels induced by water budgeting could lead to more work that would determine how to encourage this kind of behavior in the landscape It is possible that these differences in top growth could expand over time, and data for a longer period could prove useful.
67 CHAPTER 6 CONCLUSION Turfgrass irrigation based on ET O required less frequent mowing than is typical within th e landscape maintenance industry. Typically, a lawn service will contract with the homeowner for weekly mowing March/April through August/September; and then bi weekly mowing for the remaining winter months. Some clients insist on weekly visits year roun d (personal experience, S. Simpson). This typically results in a lawn being mowed 39 to 45 times or more per year. For this study, mowing totaled 24 occurrences annually, needing bi weekly or less mowing during the dry spring months, needing to be mowed only 7 times in the months where warm season turf growth is minimal ( Figure 5 1), and with no mowing in December, January, and most of February. This is nearly a 50% reduction in the need to run the mower when compared to typical industry practices. Wa rm season turfgrass simply does not grow in the cold temperatures of winter (Christians 2011) Spring in Florida is usually a period of low rainfall (Figure 4 1) and rising temperatures. Often during the summer, there are extended periods with no rainfal l. Extensive supplemental irrigation and spring fertilization take place during these periods These practices work together to stimulate unnecessary spring growth. If irrigation were reduced through the use of real time ET data and landscape coefficien ts, not only would there be a tremendous savings in water use, but also a reduction in fertilizer application, the potential for reduced groundwater contamination, and the environmental benefits of reduced mowing. Reduced maintenance requirements associa ted with less frequent irrigation has more potential than simply cost savings to the homeowner, but also to reduce fuel consumption and subsequent
68 emissions. Since one hour of push mower use equals 50 miles in a typical car (EPA 1991) potential emissions savings statewide become as big a factor as water savings. ET A levels demonstrated increasing values as irrigation levels increased (Table 4 2). Higher ET A values for the turfgrass and plant material indicate a propensity for the plant material to use m ore water as it becomes available. Since all plant material remained above minimum aesthetic rating requirements ( Table 4 3), this indicates that the turfgrass and woody plants adapted to each DI while demonstrating the acceptable aesthetics deemed necess ary for residential landscapes Based on these findings, we could irrigate a mixed landscape at 60% DI and still have turf that looks as good as if we had irrigated at 90% DI. This adaptability is crucial to the success of a landscape facing water stress and should be encouraged through the manipulation of cultural practices including irrigation frequency. Aside from the outliers discussed above, K L results closely followed seasonal demands and indicated the potential for usefulness in water conservation when used to schedule mixed landscape irrigation based only on turf water needs. ET based irrigation controllers could use these K L values to accurately correct ET O and provide irrigation frequently enough to maintain acceptable aesthetic levels in resid ential landscapes and reduce unnecessary watering. Other studies also concluded that St. Augustine turfgrass could be irrigated and maintained at 60% ET O (Dukes 2007) using soil moisture se nsors and ET controllers. The results presented here confirm that not only is irrigation possible at the deficit levels tested, but suggest lower deficits may also demonstrate acceptable aesthetic levels. Weather data and ET A observations over a
69 longer period may be useful when testing lower deficit levels in order to give the turfgrass time to spread and fully extend its roots. Given the va riation in climate and rainfall, data collection over several years would be useful. More exploration into the physiology of root response of turf and woody reaction s to stress a nd potential avenues into how to encourage those stress adaptations could also provide insight into methods of conditioning these plants to handle less frequent irrigation. The effects of soil water holding capacity of dry soils on large scale lysimetry u sed in deficit irrigation research must also be further investigated so that these effects can be accounted for in ET A observations and subsequent K L calculation
70 LIST OF REFERENCES Adeloye, A. J., Rustum, R. and Kariyama, I. D. (2011) 'Neural computing modeling of the reference crop evapotranspiration', Environmental Modelling & Software ; 1364 8152 Allen, R. G., Pereira, L. S., Raes, D. and Smith, M. (1998) 'Crop evapotranspiration Guidelines for computing crop water requirements FAO Irrigation and dra inage paper 56', FAO, Rome, 300, 6541. Atkins, C., Green, R., Sifers, S. and Beard, J. (1991) 'Evapotranspiration rates and growth characteristics of ten St. Augustinegrass genotypes', HortScience, 26(12), 1488 1491. Baum, M. C. (2005) Residential irriga tion water use in the central Florida ridge unpublished thesis University of Florida. Beard, J. (1985) 'An assessment of water use by turfgrasses', Turfgrass water conservation. Cooper. Ext Publ, 21405. Beard, J. (1989) Turfgrass water stress: drought r esistance components, physiological mechanisms, and species genotype diversity, Keynote Address, 6th International Turfgrass Research Conference, Tokyo, translated pp 23 28. Beeson, R. (1991) 'Scheduling woody plants for production and harvest', HortTec hnology, 1(1), 30 35. Beeson, R. (1993) Relationship of Potential Evapotranspiration and Actual Evapotranspiration of Rhododendron sp.Formosa', translated by Florida State Horticultural Society 274 274. Beeson, R. (2006) 'Relationship of plant growth and actual evapotranspiration to irrigation frequency based on management allowed deficits for container nursery stock', Journal of the American society for horticultural science, 131(1), 140 148. Beeson, R. and Brooks, J. (2006a) Evaluation of a model based on reference crop evapotranspiration (ETo) for precision irrigation using overhead sprinklers during nursery production of Ligustrum japonica, translated by 85 90. Beeson, R. and Brooks, J. (2006b) Modeling Actual Evapotranspiration of Acer rubrum from a Rooted Cutting to an 8 m Tall Tree in a Humid Climate, translated by 91 97.
71 Beeson, R. C. (2004) 'Response of Individual Plant Actual Evapotranspiration to Canopy Closure: A Three Dimensional Model Derived from Viburnum odoratissimum', HortScience, 39( 4), 857 857. Beeson, R. C. (2005) 'Modeling irrigation requirements for landscape ornamentals', HortTechnology, 15(1), 18 22. Beeson, R. C. (2010) 'Response of evapotranspiration of Viburnum odoratissimum to canopy closure and the implications for water conservation during production and in landscapes', HortScience, 45(3), 359 364. Brown, P. W., Mancino, C. F., Young, M. H., Thompson, T. L., Wierenga, P. J. and Kopec, D. M. (2001) 'Penman Monteith crop coefficients for use with desert turf systems', Crop Sci, 41(4), 1197 1206. Brutsaert, W. (1982) Evaporation into the atmosphere: Theory, history, and applications, 1st ed., Springer. Burger, D., Hartin, J., Hodel, D., Lukaszewski, T., Tjosvold, S. and Wagner, S. (1987) 'Water use in California's ornamenta l nurseries', California Agriculture, 41(9), 7 8. Cameron, R., Harrison Murray, R., Atkinson, C. and Judd, H. (2006) 'Regulated deficit irrigation: a means to control growth in woody ornamentals', Journal of horticultural science & biotechnology, 81(3), 4 35 443. Carrow, R. N. (1995) 'Drought resistance aspects of turfgrasses in the southeast: evapotranspiration and crop coefficients', Crop science, 35(6), 1685 1690. Census, U. S. (2010) 'Resident Population Data', Population Density Florida [online], ava ilable: http://2010.census.gov/2010census/data/apportionment dens text.php [accessed 31 January 2012]. Christians, N. (2011) Fundamentals of turfgrass management, 3rd ed., Wiley. Devitt, D., Morris, R. and Bowman, D. (1992) 'Evapotranspiration, crop coefficients, and leaching fractions of irrigated desert turfgrass systems', Agron. J, 84, 717 723. Dukes, M. D. (2007) 'Turfgrass irrigation requirements simulation in F lorida', 28th Annual Irrigation Show, San Diego, CA Dukes, M. D. (2011) 'Summary of IFAS turf and landscape irrigation recommendations'. IFAS publication #AE436 [online] available: http://edis. ifas.ufl.edu/document_ae436 [accessed 17 December 2011].
72 Ehret, D. L., Lau, A., Bittman, S., Lin, W. and Shelford, T. (2001) 'Automated monitoring of greenhouse crops', Agronomie, 21(4), 403 414. EPA, U. (1991) 'Nonroad Engine and Vehicle Emission Study Report', [online], available: http://www.epa.gov/otaq/regs/nonroad/nrstudy.pdf [accessed 10 February 2012]. Feldhake, C. and Butler, R. (1983) 'Turfgrass Evapotranspiration. I. Factor s Influencing Rate in Urban Environments1', Agronomy Journal, 75(5), 824. Feldhake, C. and Butler, R. (1984) 'Turfgrass Evapotranspiration. 11. Responses to Deficit Irrigation1', Agronomy journal, 76(1), 85. Fereres, E. and Soriano, M. A. (2007) 'Deficit irrigation for reducing agricultural water use', Journal of Experimental Botany, 58(2), 147. Fernald, E. and Purdum, E. (1998) 'Water resource atlas', Institute of Public Affairs, Florida State Univ., Tallahassee, Fla 114 119. Garcia Navarro, M. C., Ev ans, R. Y. and Montserrat, R. S. (2004) 'Estimation of relative water use among ornamental landscape species', Scientia horticulturae, 99(2), 163 174. Gilman, E. F. (1990) 'Tree root growth and development. II. Response to culture, management and planting ', J. Environ. Hort, 8(4), 220 227. Haley, M. B., Dukes, M. D. and Miller, G. L. (2007) 'Residential irrigation water use in Central Florida', Journal of Irrigation and Drainage Engineering, 133, 427. Henson, D. Y., Newman, S. E. and Hartley, D. E. (2006 ) 'Performance of selected herbaceous annual ornamentals grown at decreasing levels of irrigation', HortScience, 41(6), 1481. Hilaire, R. S., Arnold, M. A., Wilkerson, D. C., Devitt, D. A., Hurd, B. H., Lesikar, B. J., Lohr, V. I., Martin, C. A., McDonald, G. V. and Morris, R. L. (2008) 'Efficient water use in residential urban landscapes', HortScience, 43(7), 2081 2092. Hodges, A. W., Haydu, J., Van Blokland, P. and Bell, A. (1994) Contribution of the Turfgrass Industry to Florida's Economy, 1991 92: A Va lue Added Approach, Food and Resource Economics Dept., Florida Agricultural Experiment Stations, Institute of Food and Agricultural Sciecnes, University of Florida. Howell, T. A., Schneider, A. D. and Jensen, M. E. (1991) History of lysimeter design and u se for evapotranspiration measurements, translated by 1 9.
73 Irmak, S., Allen, R. and Whitty, E. (2003a) 'Daily grass and alfalfa reference evapotranspiration estimates and alfalfa to grass evapotranspiration ratios in Florida', Journal of irrigation and d rainage engineering, 129, 360. Irmak, S., Haman, D. and Jones, J. (2002) 'Evaluation of Class A pan coefficients for estimating reference evapotranspiration in humid location', Journal of irrigation and drainage engineering, 128, 153. Irmak, S., Irmak, A ., Allen, R. and Jones, J. (2003b) 'Solar and net radiation based equations to estimate reference evapotranspiration in humid climates', Journal of irrigation and drainage engineering, 129, 336. Jia, X., Dukes, M. D. and Jacobs, J. M. (2009) 'Bahiagrass c rop coefficients from eddy correlation measurements in Central Florida', Irrigation Science, 28(1), 5 15. Jones, J. W. (1984) 'Estimated and measured evapotranspiration for Florida climate, crops, and soils', Bulletin/University of Florida. Agricultural E xperiment Station (USA) Kang, S., Shi, W. and Zhang, J. (2000) 'An improved water use efficiency for maize grown under regulated deficit irrigation', Field crops research, 67(3), 207 214. Kjelgren, R., Montague, T. and Beeson, R. (2004) 'Water Use and S tomatal Behavior of Sweetgum (Liquidambar styraciflua) Relative to Reference Evapotranspiration in Three Contrasting Regions', HortScience, 39(4), 764 764. Kramer, P. J. and Boyer, J. S. (1995) Water relations of plants and soils 1st ed., Academic Press. McCready, M., Dukes, M. and Miller, G. (2009) 'Water conservation potential of smart irrigation controllers on St. Augustinegrass', Agricultural Water Management, 96(11), 1623 1632. Meyer, J. and Gibeault, V. (1987) 'Turfgrass performance when under irr igated', Applied agricultural research, 2. Monteith, J. (1965) Evaporation and environment, translated by 205 234. Morris, K. N. and Shearman, R. (2008) 'NTEP turfgrass evaluation guidelines', National Turfgrass Evaluation Program, Beltsville, MD) [onli ne], available: http://www. ntep.org/pdf/ratings. pdf 21 [accessed 16 January 2012] Nobel, P. S. (1999) Physicochemical and environmental plant physiology, 4th ed., Academic Pr ess
74 Pannkuk, T. R., W hite, R. H., Steinke, K., Aitkenhead Peterson, J. A., Chalmers, D. R. and Thomas, J. C. (2010) 'Landscape coefficients for single and mixed species landscapes', HortScience, 45(10), 1529 1533. Penman, H. L. (1948) 'Natural evaporation from open water, bar e soil and grass', Proceedings of the Royal Society of London. Series A. Mathematical and Physical Sciences, 193(1032), 120. ittenger, D., Henry, M. and Shaw, D. (2008) 'Water needs of landscape plants', Proc. UCRT Turfgrass and Landscape Research Field Day Pittenger, D. R., Shaw, D. A., Hodel, D. R. and Holt, D. B. (2001) 'Responses of landscape groundcovers to minimum irrigation', Journal of Environmental Horticulture, 19(2), 78 84. Pittenger, D. R., Shaw, D. A. and Richie, W. E. (2004) 'Evaluation o f weather sensing landscape irrigation controllers', Retrieved August, 3, 2006. Pope, C. and Fipps, G. (2000) Potential e vapotran spiration for irrigation water management in urban landscapes t he San Antonio Experience translated by Amer Society of Agri cultural, 474. Sachs, R. (1991) 'Stress adapted landscapes save water, escape injury in drought', California agriculture, 45(6), 19 21. Sachs, R., Kretchun, T. and Mock, T. (1975) 'Minimum irrigation requirements for landscape plants', J. Amer. Soc. Hort Sci, 100, 499 502. Scheiber, S., Gilman, E., Paz, M. and Moore, K. (2007) 'Irrigation affects landscape establishment of burford holly, pittosporum, and sweet viburnum', HortScience, 42(2), 344 348. Scheiber, S. M. and Beeson, R. C. (2007) 'Landscape g rowth and aesthetic quality of coleus managed with irrigation deficits', Horttechnology, 17(4), 561 566. Schneider, A., Howell, T., Moustafa, A., Evett, S. and Abou Zeid, W. (1998) 'A simplified weighing lysimeter for monolithic or reconstructed soils'. Schuch, U. K. and Burger, D. W. (1997) 'Water use and crop coefficients of woody ornamentals in containers', Journal of the American Society for Horticultural Science, 122(5), 727 734. Shaw, D. and Pittenger, D. (2003) Performance of landscape ornamentals given irrigation treatments based on reference evapotranspiration, translated by 607 614.
75 Sinclair, T. R., Schreffler, A., Wherley, B. and Dukes, M. D. (2011) 'Irrigation Frequency and Amount Effect on Root Extension during Sod Establishment of Warm sea son Grasses', HortScience, 46(8), 1202 1205. Sivyer, D., Harris, J. R., Persaud, N. and Appleton, B. (1997) 'Evaluation of a pan evaporation model for estimating post planting street tree irrigation requirements', Journal of Arboriculture, 23, 250 256. S tabler, L. B. and Martin, C. A. (2000) 'Irrigation regimens differentially affect growth and water use efficiency of two southwest landscape plants', Journal of Environmental Horticulture, 18(2), 66 70. Stacia, L. and Dukes, M. D. (2011) 'Irrigation Using ET Controllers: A Summary of Research Performed by the University of Florida'. [online], available: http://elibrary.asabe.org/azdez.asp?JID=5&AID=35890&CID=ndci2010&T =2 [accessed 31 October 2011]. Turgeon, A. (2002) Turfgrass management 6th ed., Prentice Hall. USDA (1989) 'Soil Survey of Orange County, Florida', [online], [online] available: http://soils.usda.gov/survey/online_surveys/florida/orange/Orange.pdf [accessed 10 February 2012]. USGS (2005) 'Water Use in Florida, 2005 and Trends 1950 2005', Water Use in Florida [online], available: http://pubs.usgs.gov/fs/2008/3080/ [accessed 31 January 2012]. Ventura, F., Spano, D., Duce, P. and Snyder, R. (1999) 'An evaluation of common evapotranspiration equations', Irrigation Science, 18(4), 163 170. Whitehead, D. and Kelliher, F. (1991) 'Modeling the water balance of a small Pinus radiata catchment', Tree Physiology, 9(1 2), 17. Wilson, K. B., Hanson, P. J., Mulholland, P. J., Baldocchi, D. D. and Wullschleger, S. D. (2001) 'A comparison of methods for determining forest evapo transpiration and its components: sap flow, soil water budget, eddy covariance and catchment water balance', Agricultural and Forest Meteorology, 106(2), 153 168. Yoder, R., Odhiambo, L. and Wright, W. (2005) 'Evaluation of methods for estimating daily re ference crop evapotranspiration at a site in the humid southeast United States', Applied engineering in agriculture, 21(2), 197 202.
76 BIOGRAPHICAL SKETCH Scott Simpson received his Master of Science degree from the University of Florida in spring 2012. H is major was environmental horticulture. Previously in 1992, he received his Bachelor of Arts degree fro m University of Central Florida Scott was required to return to community college to complete biology I and II and chemistry I and II prerequisites in fall 2007. In fall 2008, he was admitted to University of Florida where he pursued his post baccalaureate studies in l andscape m anagement. Upon completion of these requirements, he was accepted into the masters p rogram in fall 2009. During this time, h e also started and operated a landscape management business, eventually expanding into lands cape design and installation. Scott now specialize s in the retrof it of existing landscapes with n ative, Florida friendly, and water wise plant material.