Evaluation and Demonstration of Soil Moisture Based On-demand Irrigation Controllers for Vegetable Production (B228) FINAL REPORT

Transcription

1 Evaluation and Demonstration of Soil Moisture Based On-demand Irrigation Controllers for Vegetable Production (B228) SWFWMD Contract 06C UF Project FINAL REPORT By Lincoln Zotarelli 1 Research Scientist Investigator Michael D. Dukes 1 Faculty Investigator Agricultural and Biological Engineering Department 1 Institute of Food and Agricultural Sciences University of Florida Submitted to Southwest Florida Water Management District March 30,

2 Table of Contents 1. Executive Summary Evaluation of soil moisture sensor (SMS) irrigation controllers...5 Pepper fruit yield, plant biomass and water use efficiency...7 Monitoring soil water percolation and nitrate leaching...8 Soil moisture sensor performance, soil moisture and soil water percolation Drainage and nitrogen leaching Pepper biomass accumulation and leaf area Pepper yield and irrigation water use efficiency Use of soil moisture sensor in commercial field Demonstration Farm Overview On farm soil moisture monitoring Practical interpretation of soil moisture for a given field Identification of drawbacks of the use of SMS and future research needs Summary Cited References: Publication production

3 1. Executive Summary Florida is the most important center of production and distribution of vegetables in the southeastern U.S. with 181,000 acres planted in 2006 and a crop value greater than $1.2 billion dollars (USDA, 2008). Among the vegetable crops cultivated in Florida, tomato, bell pepper, strawberry, cucumber are very important economically. Florida is the top water user in the humid region of the U.S., ranking fourth in withdrawal of ground water for public supply in the United States and ranking seventeenth nationally for agricultural water use (Hutson et al., 2004). In 2005, agriculture accounted for 40% of Florida freshwater withdrawals in the state, totaling almost 2.7 billion gallons per day. About 47% of agricultural freshwater withdraws were ground water. As the largest single water use category in Florida, agriculture has been forced to utilize water more efficiently. Even though the adoption of agricultural practices such as the use of polyethylene-mulch and drip irrigation have became very common for vegetable production, there is still room for improvement regarding the irrigation scheduling of vegetable crops. The use of improved irrigation scheduling techniques using soil moisture sensors to monitor soil moisture and control irrigation events has been shown to greatly increase irrigation water use efficiency. With more efficient water use, fertilizer is also retained in the effective root zone longer and growers can attain maximum yields at lower N-fertilizer application rates. As a result, better irrigation scheduling techniques will not only provide substantial water savings but can also greatly reduce potential N-leaching losses and thus minimize water quality impacts. Field experiments conducted between 2006 and 2008 revealed that the use of soil moisture sensors (SMS) to control irrigation resulted in significant reductions in the volume of irrigation applied (up to 50%) without reduction of marketable pepper yield. Pepper yield increased from 4% to 13% for SMS-based treatments compared to fixed time irrigation typically used by growers. 3

4 Additional measurements during the field experiment confirmed that precise irrigation management can greatly enhance N-fertilizer retention in the active root zone, reducing water percolation and NO 3 -N leaching. Fixed irrigation schedules without a realistic evaluation of the actual soil moisture status may result in over-irrigation and N leaching. In the second part of this project, the soil moisture sensor monitoring technology was demonstrated at a commercial vegetable production field. The comprehensive field demonstration has clearly shown the benefits of soil moisture sensor use to manage irrigation of vegetable crops under drip irrigation. The utility of such information allows the irrigation decision maker to continuously monitor the plant water demand in situ to assess critical trends in soil water status. Real-time soil moisture monitoring combined with information obtained from co-located measurements of atmospheric variables to estimate evapotranspiration provides a framework to monitor short and long-term soil moisture storage changes such as during periods of wet or dry conditions. We identified that the implementation of on demand automated irrigation by growers will require further adaptation of the current irrigation systems to the soil moisture sensor technology. Ideally, large areas will require wireless communication between soil moisture sensors and a central computer. This central computer will store and process the changes in soil moisture information according to the plant water requirement and ideal soil moisture threshold defined from the experimental results achieved in this project. Simple algorithms would be able to define the water requirement and automatically start the irrigation pump and/or notify the irrigator by cell phone, text message or any other alert system when and how much irrigate. For example, most of the current drip irrigation systems in the Manasota basin have diesel pumps which are manually turned on, future adaptation of the startup system will be required in order to make the on demand system ready to be used by growers. 4

5 Commercially available soil moisture sensor irrigation control systems do not exist with all of these features. This final report summarizes work performed on this project and reported in the previous task reports and the recommendation for the implementation of the soil moisture sensor devices on a commercial basis. Based on the information from this project, adoption of soil moisture sensors for monitoring of commercial vegetable fields is strongly reccomended since this practice may reduce irrigation water consumption while maintaing crop yield. Guidelines on commercial automatic soil moisture based irrigation controls as a best management practices should be developed for vegetables. The grower guidelines should include number of sensors required and optimum placement relative to varying soil conditions of commercial production. An economic assessment of costs associated with and benefits derived from conversion of irrigation systems in vegetables from seepage to drip irrigation needs to be made to promote water conservation by vegetable growers in Florida. Finally, multiple on-farm demonstrations of this technology will likely be necessary for grower adoption. 2. Evaluation of soil moisture sensor (SMS) irrigation controllers Between 2006 and 2008, field experiments were carried out at the University of Florida, Plant Science Research and Education Unit, near Citra, FL. The field operations were similar to a commercial vegetable production system. Before transplanting, the area was rototilled and raised beds were constructed with 6 ft between bed centers, the soil was fertilized, fumigated and plastic mulched. Irrigation was applied via drip tape and water applied by irrigation and/or fertigation was recorded by positive displacement flowmeters. Green bell pepper transplants variety Brigadier were set in early April of each year. 5

6 The experimental design consisted of a complete factorial arrangement of three N-rates and 4 irrigation treatments in 2006, and same three N-rates and 5 irrigation treatments in The N-rate treatments corresponded to 160, 200 and 300 lb ac -1 of N applied as calcium nitrate. In 2008, 5 different irrigation treatments were studied under a single N-rate 200 lb ac -1. In all studied years, the treatments were randomized within blocks with 4 replicates. Weekly N application rates, expressed as a percentage of the total N application, corresponded to 5.5% at weeks 1, 2 and 13; 7.1% at weeks 3, 4 and 12; and 8.9% at weeks 5-11 (Olson et al., 2005). Preplant fertilization corresponded to the application of 100 lb ac -1 of P 2 O 5. Fertilizer was banded and mixed into soil during the bed formation. All nutrients (except phosphorus) were applied via injection in the drip irrigation system (fertigation). Fertilizer sources and rates used were potassium chloride at a rate of 186 lb ac -1 of K 2 O and magnesium sulfate at a rate of 9 lb ac -1 of Mg. Irrigation treatments were regulated by the commercial RS500 soil moisture sensor (SMS) controller manufactured by Acclima, Inc. (Meridian, ID, USA). The RS500 unit controls irrigation application by bypassing time clock initiated irrigation events if soil moisture is at or above a preset threshold of volumetric water content (VWC) at 0.08, 0.10 and 0.12 in 3 in -3, respectively, SS 8, SS 10 and SS 12 in 2006 and In 2007, we introduced the use of double drip irrigation (SD 10 ), which was tested with SMS control and a threshold set at 0.10 in 3 in -3 (Table 1). In the spring of 2008 the SMS treatments tested were preset at VWC of 0.04, 0.08 and 0.12 in 3 in -3, respectively, SS 4, SS 8 and SS 12 ; and with double drip irrigation at VWC of 0.08 in 3 in -3 and 0.12 in 3 in -3, respectively, SD 8 and SD 12 (Table 1). The soil moisture sensor probes were installed at a 45 degree angle between two plants that measured the soil moisture in the top 6 in of the bed. Timed irrigation windows were specified as five possible events per day, starting at 6

7 8:00 am, 10:00 am, 12:00 pm, 2:00 pm, and 4:00 pm for 24 minutes each (2 hr d -1 total, or equivalent to 0.23 in day -1 or 47.2 gal 100ft -1 d -1 ). A reference treatment (TIME) was established, a time-based irrigation treatment with one fixed 2 hr irrigation event per day meant to represent a common grower practice of fixed time based irrigation. Pepper fruit yield, plant biomass and water use efficiency Plots were harvested on 58, 70 and 74 days after transplanting (DAT) in 2006; on 69 and 83 DAT in 2007 and on 62 and 76 DAT in The harvested area consisted of a central 30 ft long region within each plot. Pepper fruits were graded into culls, U.S. Number 2 (medium), U.S. Number 1 (large), and Fancy (extra-large) according to USDA (2005) grading standards for fresh market sweet peppers. Marketable weight was calculated as total harvested weight minus the weight of culls. The number and weight of fruits per grading class were recorded for individual plots. Irrigation water use efficiency (IWUE) expressed in lb in -1, IWUE was calculated by taking the quotient of the marketable yields (lb ac -1 ) and the total applied seasonal irrigation depth (in). Plant biomass accumulation was evaluated by harvesting two representative plants per treatment replicate at transplanting, 23, 34, 49 and 68 DAT. Vegetative and reproductive plant parts were separated. Leaf area was determined for each sample using a LI-300 (Li-cor, Lincoln, Nebraska). Shoot and fruit tissues were dried at 65 o C for subsequent dry weight determination. Photosynthesis was measured at midday (around solar noon) on the youngest sun exposed, fully expanded leaves, using a portable closed gas-exchange photosynthesis system (model LI-6200, LI-COR, Inc. Lincoln, NE) equipped with a ventilated chamber. 7

8 Table 1. Outline and description of irrigation treatments along with threshold volumetric water Treat. codes content (VWC), total number of irrigation events scheduled, allowed to irrigate and skipped. Irrigation description Threshold VWC Max. irrigation Number of irrigation events in 3 in -3 frequency (events d -1 ) Sched. Irrigated Skipped Spring 2006 (N-rates 160, 200 and 300 lb ac -1 ) SS 8 Acclima Digital TDT RS-500 single drip SS SS TIME No soil moisture sensor, daily fixed time irrigation Spring 2007 (N-rates 160, 200 and 300 lb ac -1 ) SS 8 Acclima Digital TDT RS-500 single drip SS SS SD 10 Acclima Digital TDT RS-500 double drip TIME No soil moisture sensor, daily fixed time irrigation Spring 2008 (N-rates 200 lb ac -1 ) SS 4 Acclima Digital TDT RS-500 single drip SS SS SD 8 Acclima Digital TDT RS-500 double drip SD TIME No soil moisture sensor, daily fixed time irrigation Note: all treatments have a maximum daily irrigation volume application volume of 0.23 in or 47.2 gal 100ft -1. Monitoring soil water percolation and nitrate leaching In collaboration with a Florida Department of Agriculture and Consumer Services (FDACS) project #9189, water and nitrate leaching generated by the different irrigation treatments was evaluated. The volumetric water content of the top soil of the production beds was monitored by coupling time domain reflectometry (TDR) probes (CS-616, Campbell Scientific, Inc. Logan, UT, USA) with a data logger (CR-10X, Campbell Scientific, Inc., Logan, UT). Soil moisture probes were placed in the beds at two subsequent soil layers which recorded soil moisture values. The upper probe was inserted at an angle in order to capture soil moisture in the top 10 in of the 8

9 profile and the lower probe was inserted vertically below the upper probe recording soil moisture between 10 in and 22 in. Zero tension drainage lysimeters were located 30 in below the surface of the bed (Zotarelli et al. 2007). The drainage lysimeters were constructed out of 55 gal polyethylene drums that were cut in half lengthwise with a length of 34 in, a diameter of 22 in, and a height of 11 in m (Fig.1). A total of 24 lysimeters were used to evaluate percolated volume and nitrate leaching in all irrigation treatments under N-rate treatments of 200 and 300 lb ac -1 (4 replicates for each treatment). A vacuum pump was used to extract the leachate accumulated at the bottom of the lysimeter. The leachate was removed weekly one day prior to the next fertigation event by applying a partial vacuum (5-6 psi) using 5 gal vacuum bottles for each drainage lysimeter. The use of weekly samplings combined with a partial vacuum allowed for an effective extraction of leachate at the bottom of the drainage lysimeter and the absence of anaerobic conditions. After sampling, soil water in the bottom of the barrel dropped to 15 to 20% VWC and the soil system remained oxygenated between samplings, thereby minimizing denitrification potential. Total leachate volume was determined gravimetrically and subsamples collected from each bottle were analyzed for NO 3 -N and thus total N loading rates could be calculated. Nitrate samples were analysed using an air-segmented automated spectrophotometer (Flow Solution IV, OI Analytical, College Station, TX, USA) coupled with a Cd reduction approach (modified US EPA Method [Jones and Case, 1991]). 9

10 6 ft 38 in Raised bed 22 in Drip 12 in 30 in 11 in Drainage lysimeter 34 in Raised bed 38 in Drip Tape Drainage Lysimeter Pump at 5-6 psi Bottle 5 gal Drainage Lysimeter Fig. 1. Overview of drainage lysimeter details. Soil moisture sensor performance, soil moisture and soil water percolation After plant transplanting, a crop establishment period was characterized by application of similar irrigation volume to all irrigation treatments. This period lasted 14, 12 and 23 days after 10

11 transplanting in 2006, 2007 and 2008, respectively. In the same year order above, the volume of water applied via irrigation corresponded to 2.83, 2.51 and 5.19 in (equivalent to 1,058; 938 and 1,940 gal 100ft -1, respectively). Following this period, irrigation treatments were initiated. The irrigation treatments controlled by SMS were programmed to bypass irrigation if the probe read soil moisture at or above the set threshold at the beginning of a scheduled irrigation cycle (Table 1). During the crop season, programmed irrigation events were skipped which significantly reduced the amount of water applied to soil moisture sensor (SMS) based treatments. The overall percentage (average of three years) of bypassed events for each SMS threshold was 71%; 56%, 45% and 36%, for 0.04, 0.08, 0.10 and 0.12 in 3 in -3, respectively. Accordingly, the overall volume of irrigation increased in the following order SS 4 < SS 8 < SS 10 < SS 12 < TIME, except in 2006 when SS 10 received similar volume of irrigation water as SS 12 and in 2007, when SS 8 and SS 10 had similar volumes applied (Table 2). It was observed that both SMS treatments failed to bypass irrigation events in the beginning of the season in The problem was attributed to cross communication between the TDT sensors, causing each of the irrigation controllers to receive signals from only one of the two wired sensors. Several adjustments were made, but the problem was not solved until each controller was wired to a separate individual irrigation timer. Another important issue encountered during the experimental phase was the location of the probe in the raised bed. Drip irrigation has a source point of irrigation which creates a gradient of soil moisture from the drip emitter and the sides of the raised bed. Therefore, the location of the sensor relative to drip line and plant row plays an important role in the sensing irrigation systems. Even thought the treatments were set at different thresholds, for example, if the SS 8 soil 11

12 moisture probe was placed in a drier spot, it could result in higher irrigation volume applied than the SS 12. The contribution of rainfall to pepper water requirements was not directly considered in the calculations, due to the presence of plastic mulch and the absence of a perched water table, while coarse sandy soils also typically demonstrate very limited lateral flow. Although, it was observed that high intensity precipitation events (> 0.31 in h -1 ) slightly increased the soil water content as measured by TDR. For SMS treatments, there was no increase in number of skipped irrigation events after rainfall. For example, precipitation events of 1.88 and 1.77 in occurring in 2006 and 2007, respectively, showed a slight increase (around 1%) in volumetric soil water content (VWC) in the 0-10 in depth layer (data not shown). 12

13 in depth in depth SS 4 SS 4 Soil volumetric water content Skipped irrigation event Irrigated event F Fertigation F F SS 8 SS 8 SS 12 SS Soil Volumetric Water Content (in 3 in -3 ) F SD 8 SD Soil Volumetric Water Content (in 3 in -3 ) F SS 12 SS 12 F TIME TIME F /01 05/02 05/03 05/04 05/05 05/06 05/07 05/08 05/09 05/10 05/11 05/12 05/01 05/02 05/03 05/04 05/05 05/06 05/07 05/08 05/09 05/10 05/11 05/12 Fig. 2. Example of 11 days of soil volumetric water content measured at 0-10 in (left graphs) and in depth (right graphs) with scheduled irrigation events during pepper vegetative 13

14 development in The dotted line represents date of soil moisture sensor treatments initiation. The soil moisture content was monitored throughout the season at depth layers of 0 to 10 in and 10 to 22 inches by TDR probes. In general, the soil moisture had a noticeable increase after each irrigation event for SMS treatments and TIME (Fig. 2, left side graphs). However, due to the higher number of bypassed irrigation events for the SMS treatment, variations in soil moisture at 0-10 in soil depth layer were not as distinct as for the TIME treatment. The advantage of SMS based irrigation compared to TIME treatment is that the SMSbased system irrigated for short periods of time, in this case, 24 min, and with an interval of at least 2 hours between irrigation events. This irrigation approach results in a relatively small increase in soil moisture in the upper soil layer, and the interval between irrigation events provided time for soil water redistribution, consequently decreasing the volume of percolate in deeper soil layer (Fig. 2 right side graphs). Slight oscillations in soil moisture were observed during the irrigation events in the deep monitored soil layer for higher SMS settings (e.g. SS 10 and SS 12 treatments). On the other hand, the fixed TIME treatment irrigated for a longer time period (2 hr), which resulted in very pronounced soil moisture fluctuations (Fig. 2). These spikes in soil moisture were only temporary, as excess soil moisture that rapidly drained below the root zone in this sandy soil. Soil moisture content returned to field capacity within 12 h. The spikes also indicate that the soil water content as measured by the TDR probes rapidly reaches a point above the soil water holding capacity in the soil upper layer, inducing percolation to deeper soil layers, and explaining the higher percolate values for the TIME treatment compared to the SMS treatments. In fact, similar spikes in soil water content were observed at inches showing appreciable soil water percolation though the soil profile throughout the entire production cycle 14

15 (Fig. 3). In terms of soil water availability to the plants, the TIME treatment initially may provide more favorable growth conditions since the soil remains wetter, thus reducing potential water stress. However, the long term excessive water percolation also increased nitrate leaching and reduced crop N supply and thereby reducing yield for green bell pepper (Fig. 3 and Table 2). Drainage and nitrogen leaching For 2006 and 2007 seasons, water percolation during crop establishment was identical for all treatments; 0.4 and 0.3 inches, respectively. According to the statistical analysis for the postestablishment period, an overall decrease (P 0.05) in soil water percolation was obtained when SMS-based irrigation controlled the water application. The TIME treatment resulted in the highest volume of water percolated below the effective root zone and captured in the lysimeters (Figs. 3 and 4). The volume percolated ranged between 0.7 and 2 inches (Figs. 3 and 4), which corresponded to 10% to 16% of the applied irrigation water. In 2007 season, the volume percolated under SS 10 treatment was 0.6 inches, which corresponded to only 5% of the total irrigation water applied. Similar comparison showed that SS 12 treatment percolated 1.3 inches, which was translated to 11% of the total irrigation water applied. In 2006 and 2007, there was no interaction between irrigation and N-rate treatments for cumulative nitrate loads below root zone. The TIME treatment resulted in the most NO 3 -N leaching. Cumulative NO 3 -N leaching values were ranged between 37 to 70 lb ac -1 for TIME treatment (Fig. 4). The single high volume daily application of the TIME treatment is likely the cause of the appreciable drainage and NO 3 -N leaching below the rootzone compared to the irrigation scheduling. A consistent reduction in NO 3 -N leaching was observed when scheduling irrigation associated to the use of SMS was adopted. These reductions were on the order of 25% 15

16 to 74%, which can be translated to range of 9 to 31 lb ac -1 of N. Independently of the irrigation treatment, the increase in N-rate from 220 to 330 kg ha -1, significantly (P 0.05) increased the NO 3 -N leaching. In 2008, cumulative nitrate leaching at the end of the crop season was significantly higher for TIME 2h with 23 lb NO 3 ac -1 leached. The measured values for SS 4, SS 8, SS 12 and SD 12 ranged from 8 to 12 lb NO 3 ac -1 following the same patterns observed for percolated volume (Fig. 4 right). Excessive soil water percolation and nitrate leaching were clearly associated with overirrigation. Our results clearly show that appropriate irrigation scheduling and matching irrigation amounts with the water holding capacity of the effective root zone thus may provide ways to minimize the incidence of excess nitrogen leaching associated with over-irrigation. Figure 4 is a visual example of the effectiveness of appropriate irrigation scheduling to reduce the volume of water percolated in the soil profile compared to fixed time irrigation. 16

17 Cumulative percolated volume (in) Percolated Volume SS 10 SS 12 TIME Fig.3. Cumulative leachate volume (left) and cumulative NO 3 -N mass leached for different irrigation scheduling and soil moisture sensor irrigation treatments during spring 2006 and Treatment means followed by same letter are not different according to Duncan s Multiple Range Test at P Note: N-rates of 200 and 300 lb ac -1. a b c a a b Cumulative NO 3 -N leached (lb ac -1 ) Days after transplanting N-Rate 200 lb ac -1 SS 10 SS 12 TIME N-Rate 300 lb ac -1 Nitrate Leaching SS 10 SS 12 TIME

18 4.0 Percolated Volume 30.0 Nitrate Leaching Percolated volume (in) Plant establishment (5.19 in applied) SS 4 SS 8 SS 12 SD 12 Time 2h a b bc bc c N-NO 3 (lb ac -1 ) Plant establishment (5.19 in applied) a b bc c Days after transplanting Fig.4. Cumulative leachate volume (left) and cumulative NO 3 -N mass leached for different irrigation scheduling and soil moisture sensor irrigation treatments during spring Treatment means followed by same letter are not different according to Duncan s Multiple Range Test at P Note: N-rate of 200 lb ac

19 8 in 8 in 16 in 25 in 16 in +35 in Fig. 4. Demonstration of the effectiveness of soil moisture sensor based irrigation systems in enhancing nutrient retention for soil moisture sensor irrigation (top row) due to small frequent irrigation events compared to fixed time irrigation schedule with single daily large irrigation events (bottom row) applying dye through the fertigation drip lines. Soil moisture sensor irrigation (top row) and fixed time schedule irrigation (bottom row) at after 24 h (left), after 3 d (center) and after 7 d (right) of the injection of dye. (Photos: L. Zotarelli). 19

20 The fixed TIME irrigation treatment resulted in the highest volume of water percolated below the effective root zone and captured in the lysimeters (Fig. 3). The volume percolated for TIME treatment ranged between 1.25 and 3.15 inches which corresponded to 10% to 23% of the applied irrigation water. A significant reduction in soil water percolation was achieved by using SMS to control irrigation in all studied years, regardless of SMS threshold. Figure 3 shows an example of the use of SMS on soil water percolation control. In spring 2008, the overall leachate amounts since transplanting were 0.94, 0.86, 1.18, 1.06 and 3.14 in, for SS 4, SS 8, SS 12, SD 12 and TIME 2h, respectively. However, most of the water percolation occurred during the establishment period, except for TIME 2h, which consistently percolated about inches wk -1 (Fig. 3 left). Similarly to the percolated soil water volume, the TIME treatment resulted in the most NO 3 -N leaching (Fig. 3, right). Cumulative NO 3 -N leaching values were ranged between 22 and 42 lb ac -1 for TIME treatment. The single high volume daily application of the TIME treatment is likely the cause of the appreciable drainage and NO 3 -N leaching below the rootzone compared to the SMS irrigation scheduling. A consistent reduction in NO 3 -N leaching was observed when scheduling irrigation with SMS systems was adopted. These reductions were on the order of 25% to 74%, which can be translated to range of 9 to 31 lb ac -1 of less N being lost by leaching. Pepper biomass accumulation and leaf area Shoot dry biomass accumulation ranged between 720 to 750 lb ac -1 in 2006 and 970 to 1,161 lb ac -1 in There was no significant (P 0.05) difference between SMS-based treatments and TIME treatments, neither between N-rates. In 2008, overall shoot (leaves and stem) dry weight ranged between 1,250 and 1,700 lb ac -1 with no differences between treatments (Fig. 5 bottom). A reduced leaf area was observed for SS 4 and SS 8 treatments, while other treatments showed a 20

21 leaf area above 527 in 2 plant -1. Interestingly, the leaf area for TIME 2h treatment reached a plateau around 50 DAT, indicating earlier plant maturity, which may be associated to the excessive leaching and lack of nitrogen, compared to the other treatments (Fig. 5 top). On the other hand, the reduction in leaf area for treatments SS 4 and SS 8 was related to the soil water availability and irrigation water distribution. Besides the reduced number of irrigation events for SS 4 and SS 8 compared to the other SMS treatments, the amount of water applied to these two treatments was not enough to wet soil between 6-12 inches and VWC decreased below 10% several days after the treatments started (Fig. 2 right column). 21

22 Leaf area (in 2 plant -1 ) Plant biomass (lb ac -1 ) Single drip Soil Moist. Sensor Plant establishment period SS 4 SS 8 SS 12 SS 4 SS 8 SS 12 a b Double drip Soil Moist. Sensor SD 8 SD 12 Days after transplanting a a Single Drip Fixed Time TIME ab Days after transplanting Fig.5. Leaf area (top row) and cumulative plant biomass (bottom row) of pepper as affected by different irrigation scheduling and soil moisture sensor irrigation treatments during spring SD 8 SD 12 TIME

23 Pepper yield and irrigation water use efficiency The overall marketable yield for green bell pepper ranged between 12,990 to 15,270 lb ac -1 in 2006; 21,255 to 26,525 lb ac -1 in 2007; and 23,578 and 37,688 lb ac -1 in Except in 2006, when unfavorable environmental conditions occurred, bell pepper yield obtained in these experiments were in the range of those reported in the literature for sandy soils in Florida (Dukes et al. 2003; Maynard and Santos 2007; Simonne et al. 2006). The lower yield in 2006 compared to 2007 and 2008 was attributed to the effect air temperature on plant development and flowering. Low night time temperatures were shown to have a considerable effect on flower morphology and functioning, larger flowers, with swollen ovaries and shorter styles in comparison with flowers grown under higher temperature conditions. This effect of low temperatures has a direct effect on pepper production by decreasing the total number of pollen grains formed and by reducing their viability and germination capacity. A detailed analysis of measured air temperature during the entire crop cycle revealed that in 2006, pepper plants were exposed to temperatures below 57.2 ºF during 311 hours, while in 2007 and 2008, the cumulative hours with low temperatures (<57.2 ºF) were 181 and 185 hours, respectively. In addition, temperatures below 57.2 ºF occurred during the entire plant development and reproduction stages in 2006, while in 2007, low temperatures occurred throughout the season for short periods of time, however, between 49 and 63 DAT (peak of flowering stage) there was no occurrence of low temperatures. The use of soil moisture sensor irrigation control significantly affected the irrigation water use efficiency (IWUE) (Table 2). The treatment ranking for IWUE was as follows: SS 4 > SS 8 >SS 10 > SS 12 > TIME. The TIME treatment had a lowest IWUE values (< 2.2 lb in -1 ) due to the high irrigation rates applied. In 2006, reduced yields associated to the high volume of 23

24 irrigation applied for all treatments were responsible for the lower IWUE values (<3.94 lb in -1, Table 2). It is important to point out that high irrigation rates as applied for TIME did not increase yield, conversely, the use of scheduling irrigation by using SMS allowed application of less water, divided in five possible irrigation events per day (low volume, high frequency), which resulted in higher IWUE values. While TIME treatment had a single irrigation event (high volume, low frequency), which promotes excessive water percolation. Table 2. Irrigation treatments effects on marketable pepper fruit yield, irrigation water application, and irrigation water use efficiency (IWUE) for pepper, spring 2006, 2007 and Mkt. Yield (lb ac -1 ) Irrig. (inches) IWUE 2 (kg frt m -3 ) Water savings (%) Spring 2006 SS 8 15,272 a a 51 % SS 10 13,039 a b 8 % SS 12 13,129 a b 12 % TIME 14,468 a cb - Spring 2007 SS 8 26,525 a a 45 % SS 10 24,560 ab b 42 % SS 12 21,256 b c 16 % SD 10 24,828 ab c 3 % TIME 21,970 b c - Spring 2008 SS 4 23,578 c a 76 % SS 8 27,418 bc ab 66 % SS 12 31,616 ab ab 60 % SD 8 26,972 c ab 67 % SD 12 38,228 a bc 41 % TIME 29,472 bc d - 1 Yield and water use efficiency means followed by the same letter in the column do not differ (P>0.05) by Duncan s Multiple Range test. 2 Calculated based on total yields excluding the irrigation volume during the establishment phase. 24

25 Relative pepper yield (%) Fixed Time Reference SMS 0.12 in 3 in -3 SMS 0.10 in 3 in -3 RY% = *WS *WS 2 R 2 = 0.99 SMS 0.08 in 3 in -3 SMS 0.04 in 3 in Relative water savings (%) Fig. 6. Relative marketable yield of bell pepper with increase of relative water savings using soil moisture sensors (SMS) to control irrigation in different volumetric water (VWC) content settings. A relationship between irrigation water saving and relative pepper marketable yield were established using combined results from three years of scheduling irrigation using SMS compared to a fixed irrigation. Reduction in irrigation water application and its relative contribution to pepper yield was estimated by a quadratic regression (Fig. 6). The response of pepper yield increased when irrigation water application was reduced. The yield plateau was reached at 25-30% of irrigation water reduction, which was obtained when SMS were set at 0.12 in 3 in -3, which was slightly above soil field capacity for the experiment site. After reaching the 25

26 plateau, the yield response was reduced as water savings increased (lower setting of SMS), indicating that relative water savings higher than 55% may result in plants under water stress followed by yield reduction. Because of the high demand for and the importance of water to the plants, anytime that water becomes limiting, photosynthesis rate is reduced, as well as plant growth and yield. This was clearly demonstrated by the drier irrigation treatments with a threshold below the soil field capacity point. On the other hand, due to the low soil water retention capacity, additional irrigation water application above the soil field capacity will result in excessive water percolation and nutrient leaching, although this practice will not affect the water supplied to the plant, it will negatively affect the plant mineral nutrition and nitrate leaching. In fact, pepper photosynthetic rates along the crop season decreased with the increase of irrigation water saving (Fig. 7), there was no significant differences in photosynthetic rate between TIME and SMS treatments set near the soil field capacity point (0.12 in 3 in -3 ), showing that the potential photosynthetic rate. Soil moisture sensor settings at 0.08 in 3 in -3 and lower, resulted in lower photosynthetic rate and yield, which was associated to soil water deficit (Figs. 6 and 7). The experimental approach allowed us to analyze both extremes of the irrigation management, from over-irrigation (TIME) to sub-irrigation condition (SS 4 ) and determine an ideal soil moisture target level (safe irrigation zone) that maximize fruit yield with the least amount of irrigation water application, and consequently reduce leaching resulting in loss of water and nutrients. 26

27 30 20 Photosynthesis (mmol m -2 s -1 ) Relative pepper photosyntetic rate (%) Fixed Time Reference SMS 0.12 in 3 in -3 PR% = *WS R 2 = 0.82 SMS 0.08 in 3 in -3 SMS 0.04 in 3 in -3 SS 4 SS 8 SS 12 SD 8 SD 12 Treatments TIME Relative water savings (%) Fig.7. Season average of pepper photosynthesis rate as affected by different irrigation scheduling and soil moisture sensor irrigation treatments during spring 2008 (left) and relative photosynthetic rate of bell pepper with increase of relative water savings using soil moisture sensors (SMS) to control irrigation in different volumetric water (VWC) content settings (right). Bars indicate standard error. 27

28 3. Use of soil moisture sensor in commercial field The introduction of soil moisture sensors to a commercial vegetable field created a unique opportunity to evaluate the performance, applicability and potential use of SMS by a grower on irrigation management at a production scale. In fact, during the project several questions were raised as well as ideas to better adapt SMS to the grower conditions. Demonstration Farm Overview The Weirs-Turner Farms LLC cultivates vegetable crops in five different farms located in Manatee County. In the fall of 2008, the area planted in vegetables included 270 acres of peppers, 200 acres of tomatoes, 180 acres of cucumbers and 125 acres of squash. Crop management was typical of vegetable production in the Manasota basin with drip irrigation, plastic mulch, and raised beds (Fig. 8). Fig. 8. Manasota Basin map. The red arrows indicate the location of the farms. (Map source: SWFWMD, The irrigation system of each farm was equipped with a diesel pump, which was manually switched on. The decision of when and how much irrigate was supported by several 28

29 tensiometers installed across the fields. According to the farmer, there were two tensiometers per irrigation block. Tensiometer data were acquired manually by two employees. Routinely during the vegetable season, farm workers would drive around the farms and write down the tensiometer readings before 9 am. The soil moisture info provided by tensiometers for each irrigation field was then analyzed manually by the irrigation manager and the decision of irrigation was then made. In July 2008, a field test and demonstration of soil moisture sensors was initiated at selected pepper area of 41 acres. The area was divided in 3 irrigation zones. The soil of the demonstration area is classified as Spodosol, which is a sandy mineral soil, low in organic matter and natural fertility in the surface layer. The area slope was 1% and the water table is found between 24 and 36 inches depth depending of the location in the landscape. The pepper variety was HM The transplanting occurred on September 15. Green bell pepper was transplanted in twin staggered rows at 16 inches within row spacing and 10 inches between plants in the row. The area was roto-tilled and raised beds (8 inches) were constructed in a North-South arrangement with 6 feet between bed centers in July Fertilizer was incorporated into the bed at rate of 160 lb N ac -1 ; 120 lb P ac -1 and 320 lb K ac -1. The beds were fumigated after placement of drip irrigation and plastic mulch. On farm soil moisture monitoring On 17 September 2008, four wireless data recorders wired to four soil moisture sensors each (SD12 Acclima, Inc.) were installed in 41 acres area planted with green bell peppers (Zones 5, 6 and 7, Figs. 9-10). These commercial products were similar probes for soil moisture sensing as used in earlier tasks of this study and measure electrical conductivity and temperature along with soil moisture content. Unlike previous plot work in this project, soil moisture, 29

30 electrical conductivity, and temperature data were measured and acquired by a data recorder measured by the probes. The soil moisture probes were not configured to automatically start the irrigation as in plot work. These probes and controllers are not readily adapted to irrigation controlled by diesel pumps that need to be started manually. These devices are capable of providing soil moisture data from surface and subsurface soil depth layer for the farmer to make irrigation decisions. The subsurface soil moisture provides information about the fluctuation of water table level, which is frequently related to the rainfall and potential of occurrence of pepper root diseases. A D C N B Fig. 9. Aerial photo of the SK Kenny s Fall Field zones 5, 6 and 7. The pattern area represents the 41 acres and the letters indicate the position of soil moisture sensor data recorders. White arrows indicate descending slope. Photo source: Web Soil Survey, NRCS. 30

31 The sensor data recorders were distributed in the field in order to monitor the soil moisture and EC along the irrigated beds (North-South) and along the manifold (West-East). The 41 acre area was irrigated by a main manifold (irrigation water source) located at the north side of the field. The field is divided by the farmer into 3 irrigation zones with one flush valve per zone (Fig. 10). Manifold 2,170 ft Zone A D 980 ft Drip Lines C N B Flush valves Sensor Data recorder (Acclima SD12) Manifold Drip lines Fig. 10. Irrigation zones 4,5 and 6 and irrigation scheme and distribution of soil moisture sensor data recorders in the demonstration pepper field. The arrows indicate the water flow. Soil moisture and electrical conductivity (EC) was monitored at 0-6; 6-12; and inches depth layers. The soil moisture and EC data was be recorded every 15 minutes for each sensor. The data was retrieved monthly. The soil moisture nodes were placed in different points of the topography. In this case, node A was located in the highest point of the landscape, followed by node D, C and B (lowest point). At the installation of the soil moisture sensors, a perched water table at different depths 31

32 for each node location was observed. The depth of the water table ranged between 35 at node A location (Fig. 11) and 18 at node C location (Fig. 12). The level of the water table was monitored by probes installed at and depth layers. Detailed oscillations in soil moisture during fall and winter of 2008 can be observed in the figures 11 and 12. After each irrigation event there was a noticeable increase in soil moisture content in the 0-6 depth. The degree to which the soil moisture content increases is dependent on the duration of the irrigation event. For example, short irrigation run times result in a relatively small increase in soil moisture, consequently decreasing the volume of percolate (water moving below the root zone). Alternatively, when the irrigation occurred for a longer time period, a relatively larger increase in soil moisture was observed, and consequently increasing the soil moisture in deeper depths, especially at 6-12 depth layer, and eventually to the deeper soil layers. This spike in soil moisture is temporary, as the irrigation water rapidly drains, ultimately bringing the soil moisture content back to where it was before the event in a relatively short period of time. This rapid spike in soil water content indicates that the soil water content as measured by the soil moisture probes rapidly reaches a point above the soil water holding capacity and the water starts to percolate down to deeper soil layers. Excessive water percolation may result in nutrient leaching and reduced yield. In other words, the length of irrigation event could be reduced, or an irrigation event could be skipped to avoid deep percolation. 32

33 Fig. 11. Soil moisture content measured at depth layers of 0-6, 6-18, 12-24, by Data Recorder A during pepper Fall season of

34 Fig. 12. Soil moisture content measured at depth layers of 0-6, 6-18, 12-24, by Data Recorder C during pepper Fall season of

35 During the soil moisture monitoring period, situations were observed when proper irrigation management was performed and alternatively situations when the irrigation could have been skipped or the volume of irrigation water reduced. For example, Zone 4 (East side), the period of 11/1 until 12/3 (Fig. 11) might be considered as an example of the proper irrigation management, which no excessive water percolation from irrigation was detected (increase of volumetric soil water content in depth). This period was characterized by irrigation events that did not increase the soil water content at the soil layer depth. During this period there was no irrigation, and it was observed that the level of the water table decreased below 18 inches. In addition, during some periods with large rainfall events (specifically for 10/7; 10/23; 1/13), irrigation events were skipped by the farmer and they were resumed when soil moisture levels decreased to critical levels. Conversely, there were some periods that the irrigation could have been safely skipped, or the irrigation event length could have been reduced. For example, on the days of 10/25; 10/28; 12/12 and 12/24 (Fig. 11) when over-irrigation clearly occurred, evidenced by the spikes in moisture content in the deeper soil layers. The depth of irrigation should consider the depth of the plant root system, in other words, it is not necessary to irrigate the entire soil profile to maximize the plant water uptake. In our previous work, we have shown that about 80% of the root of pepper and tomato was found in the 0-12 depth (Zotarelli et al., 2009) and it is from that particular depth that most of the water and nutrients are taken up by vegetable plants. It was also observed that soil moisture sensor probes installed at and of Node B and C (Fig. 11) showed a higher soil moisture content than the probe located at upper layers. In fact, during the installation of the probes, a perched water table level was observed frequently below 20 depth, indicating that the probes below 24 depth would be in the perched 35

36 water table. Values of soil moisture content above 0.30 in 3 /in 3 is an indication that soil layer is close to the saturation. All monitored points of the irrigation zones showed soil saturated conditions at 30 depth (Figs. 11 and 12). One of the main concerns of the grower was high water table conditions that can reach close enough to the root zone to have a direct influence on the vigor and productivity of vegetable on raised beds. Another negative effect of the high water table for long periods, which is common during the in fall due to the hurricane season, is the high occurrence of root diseases. On the other hand, during high water table periods, the irrigation events can be dramatically reduced. In early December of 2008, there was a slight increase of the water table level observed by the increasing of volumetric water content values in the soil layer of depth. This fact can be explained by the increase of rainfall in that period. The influence of rainfall events on the soil moisture content can be noticed in this period as a result of similar spikes of soil moisture at depths of 0-6, 6-12 and 12-24, simultaneously. Again, the irrigation could be discontinued in the days after the rainfall until the soil moisture reaches some critical level when irrigation should be resumed. However, the grower rarely suspended irrigation practices as a result of rainfall events of this magnitude or lower. It is important to note that the grower was not fully relying in the soil moisture sensor data as the exclusive decision support to irrigate or not. He stated that he was frequently checking the sensor data to understand oscillations in water table and how his irrigation management affected the soil moisture content. Therefore, more than one season of work will be necessary to demonstrate and prove that the farmer can manage an entire field based on few measurement points. 36

37 In our interaction with the farmer, we also realized that he preferentially liked the graphical soil moisture data rather than tables of data. Automated and reliable data acquisition and automatic transfer to a database seems to be essential for a practical operation of the system. In addition, a simple interface where the user can consult past and current soil moisture information from different points on the farm is required. These characteristics seemed to be essential to the adoption of the SMS system by vegetable growers. Practical interpretation of soil moisture for a given field An important point to be considered for the successful adoption of the soil moisture sensor technology it the practical interpretation of the soil moisture patterns and understand how irrigation practices/rainfall can affect the soil water content. Once installed, soil moisture sensors can be successfully used to monitor volumetric water content and guide irrigation management. However a correct interpretation of the soil moisture readings is very important to assure a proper irrigation management and avoid over irrigation. There is a very simple way to interpret and evaluate soil moisture characteristics using SMS when soil moisture trends are available in graphical format to the grower. Figure 13 shows volumetric soil water content at depth of 0-6 inches measured by a capacitance sensor during a period of four days. There were two irrigation events on 11/06 and 11/09 at 7:00 am. For the soil field capacity point determination, we intentionally applied an irrigation depth that resulted in saturation of studied soil depth layer, in this particular case 0-6 inches. 37

38 Volumetric soil water content (in 3 /in 3 ) Irrigation event 64 gal/100ft (0.17 in) slope of drainage and extraction lines = soil field capacity ( ) Volumetric soil water content at 0-6" depth ( ) Irrigation event ( ) Rate of water use and drainage ( ) Day period ( ) Night period /06 0:00 11/06 6:00 11/06 12:00 11/06 18:00 11/07 0:00 11/07 6:00 11/07 12:00 11/07 18:00 11/08 0:00 11/08 6:00 11/08 12:00 11/08 18:00 11/09 0:00 11/09 6:00 11/09 12:00 Fig.8. Example of practical determination of soil field capacity for sandy soil after irrigation event. The depth of irrigation applied was 0.17 inches (34 gal 100 ft -1 ). After both irrigation events, there was an expected increase in soil moisture content. The degree to which the soil moisture content increases, however, is dependent upon volume of irrigation, which is normally set by the duration of irrigation event. For plastic mulched drip irrigation in sandy soils, long time period of irrigation results in a relatively large increase in soil moisture in the area below the drip emitter. In Figure 8, the spike in soil moisture was temporary, as the irrigation water rapidly drained down beyond the 6 inch zone (11/06, between 9:00 am and 7:00 pm). This rapid 38