|
Reducing Water, Nutrient, Pesticide, and Carbon Footprints and Costs for Drip-Irrigated Vine Crops with Compact Bed Plasticulture
Vijay Santikari1, Sanjay Shukla1,*, Rajendra P. Sishodia2, Kira M. Hansen3, Gregory S. Hendricks1
Published in Journal of the ASABE 67(2): 501-515 (doi: 10.13031/ja.15813). Copyright 2024 American Society of Agricultural and Biological Engineers.
1 Agricultural & Biological Engineering Department, Southwest Florida Research and Education Center, University of Florida, Immokalee, Florida, USA.
2 Resilient Environment Department, Broward County Environmental Protection and Growth Management, Plantation, Florida, USA.
3 Kimley Horn, Fort Myers, Florida, USA.
* Correspondence: sshukla@ufl.edu
The authors have paid for open access for this article. This work is licensed under a Creative Commons Attribution-NonCommercial-NoDerivatives 4.0 International License https://creative commons.org/licenses/by-nc-nd/4.0/
Submitted for review on 11 September 2023 as manuscript number NRES; approved for publication as a Research Article by Associate Editor Dr. Arun Bawa and Community Editor Dr. Kati Migliaccio of the Natural Resources & Environmental Systems Community of ASABE on 2 January 2024.
Highlights
Abstract. Raised bed plasticulture, used globally for producing high-value fruits and vegetables, is a high input and high intensity production system. We show that improved water and nutrient efficiency with reduced material inputs, costs, and carbon footprint is achieved with the use of taller and narrower plastic-mulched compact beds [61×30 (width × height in cm), 61×10, 46×30, and 41×30] compared to the wider and shorter conventional bed [76×20] used for growing watermelon. The primary motivation behind the compact bed design is to improve economic and environmental sustainability by reducing the bed width, to increase the fraction of the bed that is wetted by drip irrigation and to reduce the area of farmland covered under plastic. Compact beds were evaluated at two locations in South Florida with different hydrologic settings, using an ensemble of field data (inputs, plant health, yield, soil moisture, and nutrient levels) and a field-verified hydrologic model (HYDRUS). No reduction in yields was observed with compact beds, indicating no production risk for the growers. Compact beds conserve water and reduce deep percolation losses by 6 cm compared to conventional beds, thereby improving water productivity by 20%. Compact beds require less fertilizer (10%-50%) to maintain the same nutrient (N-P-K) concentrations as those in conventional beds because of smaller soil volumes. Reductions in water and fertilizer inputs provide the environmental benefit of reducing nutrient leaching to groundwater, while reduced impervious area reduces runoff losses. Compact beds reduced input (fertilizer, plastic mulch, pesticide, and diesel) costs of the system by US$720/ha and greenhouse gas emissions by 1630 kg CO2e/ha (9%) compared to conventional plasticulture. The annual cost savings create a win-win for farmers to increase profits and/or invest in conservation measures such as precision irrigation. Compact beds offer a climate change mitigation and adaptation strategy for the plasticulture system for watermelon and other similar cucurbit crops (e.g., cucumber, pumpkin, and zucchini).
Food demand is projected to increase by 110% by 2050 to meet the needs of the growing population (Tilman et al., 2011). The expansion of global agriculture to meet this demand is likely to exacerbate environmental degradation due to the increased use of inputs such as water, fuel, and fertilizer. Currently, irrigated agriculture accounts for more than 70% of global water withdrawals (FAO, 2020b), and it is a primary cause of groundwater depletion and reduced baseflows in parts of the world. Nearly 3.2 billion people reside in water scarce regions (FAO, 2020b), where agricultural water use competes with personal, industrial, and environmental uses. The use of synthetic fertilizers has altered nitrogen cycling (Vitousek et al., 1997), phosphorus cycling (Cordell et al., 2009; He et al., 2006), and downstream transport.
Excessive nutrient loading of water bodies through leaching and runoff from agricultural fields (Bennett et al., 2001; Delin and Stenberg, 2014; MEA, 2005; Wang et al., 2019; Yang et al., 2008) causes groundwater contamination (Dubrovsky and Hamilton, 2010), eutrophication (Chakraborty et al., 2017; Howarth, 2008), loss of biodiversity, and changes in ecosystem composition. Agriculture accounts for 17% of global greenhouse gas (GHG) emissions (FAO, 2020a), contributing to climate change (IPCC, 2021), which in turn puts additional stress on food production due to the increased frequency of droughts and extreme weather events.
To achieve sustainability, global agriculture must meet the expected increase in demand for food production while minimizing environmental impacts (Foley et al., 2011). Efficiency of water and fertilizer use can be improved by plasticulture, which is the practice of growing crops on raised plastic-covered beds, usually with drip irrigation. It has been mainly employed in high-value vegetable (e.g., tomato) and fruit (e.g., strawberry) crops (Ngouajio et al., 2008; Shennan et al., 2014; Zhang et al., 2017), and it is expanding to other agronomic crops such as cotton and corn in arid areas. Although a high intensity system, it has the benefit of improving quantity and quality of yields due to improved water and nutrient use efficiencies, reduced pressure from soilborne pests (e.g., weeds, nematode) and diseases, and modulated soil temperatures (Lamont, 1996; Lamont, 2005). These benefits come with certain tradeoffs, such as increased microplastics within the soil (Bläsing and Amelung, 2018; van Schothorst et al., 2021), which can be absorbed into plant roots (Li et al., 2020). Plasticulture also increases the fraction of impervious areas within the field, which can result in increased runoff and peak flows and a decrease in field storage (Arnhold et al., 2013; Jaber et al., 2006; Pandey et al., 2007). There is a need to optimize the current plasticulture system to minimize these environmental tradeoffs.
Recent studies have shown that replacing the wider and shorter plastic-covered conventional beds with redesigned taller and narrower compact beds can improve system efficiency and sustainability by reducing inputs (irrigation, fertilizer, plastics, and pesticides) while maintaining plant health and yield of erect crops (tomato, pepper, and eggplant) (Holt and Shukla, 2016; Holt et al., 2017; Holt et al., 2019). Input reductions translate to reduced costs and GHG emissions. Simulations based on a vadose zone model have shown that field runoff and leaching losses were also reduced when compact beds were used for tomato production (Holt et al., 2017, 2019). However, the compact beds have been tested only for erect crops and not for vine crops (e.g., watermelon and other cucurbits) that account for a significant area under plasticulture. Vine crops grow laterally, covering the entire bedded area, and therefore rainfall interception, evaporation, and infiltration differ from those in erect crops. Verification of the production and environmental benefits of compact beds for other crop types and hydrologic settings is necessary for a wider adoption of compact bed plasticulture to sustainably intensify agriculture.
This study evaluates the effect of raised bed geometry on water and nutrient use efficiency, production cost, and carbon footprint for a vine crop grown at commercial farms in two different hydrologic settings. One hydrologic setting comprises well-drained soils with a deep water table, while the other features poorly drained soils, flat topography, and a shallow water table (less than 1 meter below the ground surface). Watermelon was chosen as a representative vine crop due to its large production acreage and economic importance (USDA, 2022). The experiments were conducted at commercial farms because, often, small-scale studies from research farms may not align with the goals of commercial producers (Kravchenko et al., 2017). Small-scale studies are logistically convenient, offer greater control of the experimental setup, and allow for ease of measurements from successive seasons. However, they are not representative of a commercial setting and, therefore, are not scalable. Such a lack of representativeness creates a barrier for the adoption of innovative conservation practices by the industry. To facilitate buy-in from the growers and promote a large-scale adoption, all cultural practices in this study were conducted by the growers, to be consistent with the standard practices at their respective farms.
Materials and Methods
Experimental Sites
Field experiments were conducted at two commercial farms, one in North Florida (Site 1, Levy County) and the other in South Florida (Site 2, Hendry County). Site 1 is located in the Suwanee River Watershed, which is characterized by flat to rolling terrain with an average elevation of 11 m above mean sea level (USDA, 1996). The soil is classified as the Jonesville-Otela-Seaboard complex (USDA, 2019), consisting of well-drained Pleistocene sands (less than 9 m thick) overlying Eocene limestones (USDA, 1996). The average depth to the water table at Site 1 is 8 m (https://nwis.waterdata.usgs.gov). The terrain at Site 2 is flat (< 2% slope) with an average elevation of 5 m (USDA, 2019). The soil is classified as Myakka sand, which is poorly drained. The water table is shallow (< 45 cm deep) due to its proximity to the Everglades, a subtropical, iconic wetland ecosystem in South Florida. The potential for surface runoff at Site 2, which mainly occurs due to saturation excess when the water table reaches the ground surface, is much greater than that at Site 1.
The annual rainfall (20-year average) at both sites is 120 cm (SRWMD, 2020; SFWMD, 2021). Nearly 60% of the annual rainfall occurs in the summer between the months of June and September. Annual averages of daily mean, minimum, and maximum temperatures at Site 1 are 21.0 °C, 15.4 °C, and 28.0 °C, respectively (FAWN, 2021). At Site 2, they are 23.0 °C, 18.2 °C, and 28.8 °C, respectively. The experiment at Site 1 was conducted during the gradually warming spring growing season (February-June 2019), whereas the experiment at Site 2 was conducted during the gradually cooling fall growing season (August-December 2020). Together, the two sites represent contrasting hydrologic and climatic settings and growing environments.
Bed Geometries
A survey of watermelon growers from 11 counties in Florida revealed that the bed width (at the top) for drip-irrigated watermelon ranged from 61 cm to 91 cm, and the bed height ranged from 17 cm to 25 cm (Shukla et al., 2019). Based on the survey, a typical conventional bed was determined to be 76 cm wide at the top and 20 cm in height. This is also the conventional geometry used for most horticultural crops, including tomatoes and peppers. Three compact bed alternatives that were narrower (widths = 61 cm, 46 cm, and 41 cm) and taller (height = 30 cm) were designed and evaluated against the conventional bed. The naming convention used for the bed geometries in this study refers to the width and height of the beds, e.g., 46×30 refers to a bed that is 46 cm wide at the top and 30 cm in height. Accordingly, the conventional bed is referred to as 76×20. At Site 1, the bed geometry of 61×30 was replaced with 61×10 because the latter was standard at that commercial farm. Therefore, the geometries that were tested at Site 1 were 76×20, 61×10, 46×30, and 41×30; and at Site 2, they were 76×20, 61×30, 46×30, and 41×30.
Implementation and Management Practices
The experimental area at both sites was approximately 0.8 ha. The area was divided into four blocks, which were separated from each other by buffer zones, utility roads, or drainage ditches. In each block, the four bed geometries were laid out based on a randomized complete block design, resulting in four replications for each bed geometry. At Site 1, each geometry-replication had six rows of beds, with a row length of 45-60 m and a row-to-row distance of 244 cm. At Site 2, each geometry-replication had three rows of beds, with a row length of 90 m and a row-to-row distance of 183 cm.
Fertilizer
The experimental sites were tilled, and dry fertilizer was broadcast over the area that was to be bedded. The broadcast rate at Site 1 was 174 kg N/ha, 24 kg P/ha, and 138 kg K/ha. All the rates are scaled to the bedded area, which includes the beds and row middles, but excludes roads, buffer zones, and ditches. Additional fertilizer was applied at Site 1 through fertigation (drip applied fertilizer) at a rate of 71 kg N/ha, 0 kg P/ha, and 59 kg K/ha, so the total nutrient (N-P-K) input was 245-24-197 kg/ha. At Site 2, only dry fertilizer was used at a rate of 303 kg N/ha, 44 kg P/ha, and 391 kg K/ha for all bed geometries. After the fertilizer was broadcast, beds were constructed using a single row tractor-driven bedder with a press pan (bed shaper) of appropriate dimensions for each bed geometry.
Pesticide
One of the benefits of plasticulture is the ability to disinfest the soil with fumigants (gaseous pesticides) to protect the crop against diseases and pests such as soil-borne pathogens, nematodes, and weeds. The most commonly used fumigant is Pic-Clor 60, which is a mixture of chloropicrin (60%) and 1,3-dichloropropene (40%). Pic-Clor 60 was shank-applied at a rate of 140 kg per hectare of bedded area during the bedding of 76×20 geometry at Site 2. The fumigant rates for 61×30, 46×30, and 41×30 at Site 2 were 122, 96, and 92 kg/ha, respectively, as they were varied proportionally with the bottom width of the bed (USEPA, 2020). Fumigant was not used at Site 1 because of the 7-year rotational cropping system with Bahia grass, a nonhost that reduces pest and disease pressures. Fumigant use was necessary at Site 2 due to the cultivation of pepper and other vegetable crops that are susceptible to pests and diseases.
Plastic
A tractor-driven bedder, which forms the beds, was immediately followed by a plastic laying machine that lays the drip tape and plastic mulch. At Site 1, a biodegradable plastic mulch (0.8 mil, black, standard; PolyExpert, Laval, QC, Canada) was used. The width of the plastic was chosen based on the bed geometry, i.e., 168 cm for 76×20, 152 cm for 46×30 and 41×30, and 122 cm for 61×10. A 152 cm wide non-biodegradable plastic (1.1 mil, white) was used for all bed geometries at Site 2. A narrower plastic mulch (137 cm) was originally planned for use over the compact beds (46×30 and 41×30), but it was not available at the time of bedding. A single drip tape (Site 1: 0.36 gph (1.36 lph)/emitter, 30 cm emitter spacing, Netafim, Fresno, CA; Site 2: 0.27 gph (1.02 lph)/emitter, 30 cm emitter spacing, Toro, Riverside, CA) was laid under the plastic mulch.
Planting
Watermelon seedlings were planted 35 days after the bedding at Site 1. Seedless variety (captivation) was planted on 60 cm spacings, and a seeded variety, which served as a source of pollen, was planted for every three seedless plants, 30 cm from the third seedless plant. At Site 2, the seedlings of personal-sized melons were planted 20 days after the bedding on 122 cm spacings.
Irrigation
The determination of the method, timing, and duration of irrigation at each site was left to the respective growers to be consistent with standard industry practices and to facilitate the adoption of compact beds without the extra burden of modifying their irrigation management. Drip irrigation was the only source of water for crops at Site 1. During the first month after planting, the crop was irrigated every two days. After this initial period, irrigation was applied almost daily for the remainder of the season. Each irrigation event lasted from 1 to 5 hours, with the longer durations becoming more common towards the end of the season in May 2019. At Site 2, a combination of drip and subirrigation was used to apply water. Subirrigation, hereafter termed seepage irrigation, involved artificially raising the water table in the field to approximately 30 cm below the bottom of the bed by filling the perimeter ditches around the production field. Drip irrigation was applied twice per week, with each event lasting 1 to 1.5 hours. All beds received the same amount of irrigation volume per plant from the drip lines.
Instrumentation
A single experimental block, containing all four bed geometries, was instrumented at each site immediately after the planting. Time domain reflectometry (TDR) probes (Campbell Scientific Inc., Logan, UT) were used to measure moisture content at the top and bottom of the beds. One probe was installed vertically at the bed top, 10-15 cm from the plant’s root collar. This probe measured the average moisture content within the top 10 cm of soil at the center of the bed. Another probe was installed horizontally, 30 cm below the bed surface, directly beneath the top probe. An additional bottom probe was installed horizontally 20 cm below the bed surface only in 76×20 beds at Site 2. Each set of top and bottom probes at Site 2 was replicated, i.e., one set was installed in the middle bed and one set was in the edge bed (next to the drive row).
Two multi-depth soil moisture probes (Sentek EnviroSCAN, Stepney, Australia) were also installed in the beds at Site 1 to track the water movement vertically. One probe was installed at the center of the bed 10 cm from a plant’s root collar, and the other was installed 7.5 cm from the edge of the bed on the opposite side of the drip tape. The probes measured moisture content at 10 cm intervals up to a depth of 50 cm. The installation configuration of moisture probes was aimed at obtaining measurements from various locations within the bed, both vertically and horizontally (across the bed).
Rainfall was measured using a tipping bucket rain gauge (WaterLog H-340SDI; Xylem YSI, Yellow Springs, OH) installed at the edge of the experimental field. Irrigation volumes were measured using multi-jet flow meters (DLJ75: 5/8 × 3/4"; Daniel L. Jermain Co., Hackensack, NJ). A ground water well installed at the center of the instrumented experimental block at Site 2 was fitted with a pressure transducer (CS451; Campbell Scientific, Logan, UT) to measure depth to the water table.
Suction lysimeters (outer diameter = 5 cm; SoilMoisture Equipment Corp., Goleta, CA) with ceramic cup tips were installed at the center of the bed, 10 cm from the root collar of a plant, to track the nutrient movement. They were used to extract pore water samples from 15 cm and 30 cm beneath the bed surface. Lysimeters for both depths were installed in three replications (i.e., in three blocks) for each of the bed geometries.
Data Collection
Data from the rain gauge, flowmeters, TDR probes, multi-depth sensors, and pressure transducer were stored in the dataloggers (Campbell Scientific, Logan, UT) programmed to collect and store measurements every 15 minutes. Hourly weather data was obtained from two Florida Automated Weather Network (FAWN) stations (Bronson and Clewiston) located approximately 30 km from each site. Pore water samples were collected biweekly and analyzed for Nitrate-Nitrite Nitrogen (NOxN, EPA Method 353.2), Ammoniacal Nitrogen (NH4N, EPA Method 350.1), Total Kjeldahl Nitrogen (TKN, EPA Method 351.2), and Total Phosphorous (TP, EPA Method 365.1). Soil samples were collected at the time of planting and after the last harvest for two depth ranges (0-15 cm and 15-30 cm) to evaluate the effect of bed geometry on initial and residual nutrient levels and mass balance. For each depth range, the entire soil within or under the bed was collected, homogenized, and a subsample was taken. Soil samples were analyzed for NOxN, NH4N, TKN, TP, aluminum (Al), iron (Fe), and potassium (K).
Plant health was assessed through the monthly collection of leaf samples and measurements of the normalized difference vegetation index (NDVI; GreenSeeker, Trimble, Sunnyvale, CA). Leaf tissue samples were analyzed for concentrations of nitrogen (N), phosphorus (P), potassium (K), and metals. Watermelons were harvested twice at Site 1 and once at Site 2. Marketable yield was calculated for each bed geometry replication as the total fruit harvested minus the fruit culled according to the grower-cooperator’s standards and practices. Diploid (seeded pollinators) and triploid (seedless) melons were grouped and weighed separately at Site 1. A randomized subsample of 10-20 melons was weighed individually to determine the weight distribution for each bed geometry replication.
Hydrologic Modeling
Soil water content (SWC) distributions within the bed and flux below the bed for each bed geometry at both sites were simulated using an unsaturated zone hydrologic model, HYDRUS 2.05 (Simunek et al., 2012). Measured groundwater depth and weather data were used to set up the model boundary conditions. The model was calibrated and validated using the measured soil moisture.
Model Domain
A two-dimensional vertical plane domain was used to represent the cross-section of the raised bed plasticulture system for watermelon (fig. 1). In the horizontal direction (perpendicular to the row), the center of the domain was at the center of the bed, and the width was the same as the row-to-row distance of 244 cm at Site 1 and 183 cm at Site 2. The bottom boundary of the model was 200 cm below the ground surface at Site 1 and 60 cm below the ground surface at Site 2. The reason for choosing 60 cm at Site 2 was that the measured water table depth was less than 60 cm during the entire season, which is typical of the farms within the Everglades Watershed with poorly drained soil and flat topography (Shukla et al., 2010). Based on field observations, the soil profile at Site 1 was divided into three layers: (1) in-bed layer, (2) below bed layer (0-46 cm below ground surface), and (3) bottom layer (46-200 cm below ground surface). The profile at Site 2 was divided into two layers: (1) in-bed layer and (2) bottom layer (0-60 cm below the ground surface).
Mesh lines were inserted at the bed bottom, 30 cm below the bed, and at maximum rooting depth (60 cm at Site 1 and 30 cm at Site 2) to track the water movement below the beds and for calculating nutrient leaching loads. Observation nodes were setup to match the locations of moisture measurements within the bed. When the sensing volume of an instrument was large (e.g., 3600 cm3 for a TDR probe), a subregion of the same volume (area in 2D) was setup around the instrument to estimate the average moisture content of the region. Soil water retention characteristics were defined using the van Genuchten–Mualem model (Mualem, 1976; Van Genuchten, 1980), assuming no hysteresis. The initial values of the retention curve shape parameters (a, n), residual water content (?r), saturated water content (?s), and saturated hydraulic conductivity (Ks) were estimated from the NRCS database (USDA, 2019) for the observed soil types at each site.
Boundary Conditions
The boundary conditions were set up in 15-minute to one-hour intervals based on the data from which they were derived. A no-flux boundary was imposed along the impervious plastic mulch covering the surface of the bed (fig. 1). The exposed soil surface in the row middles was designated as the atmospheric boundary, which is subjected to rainfall input and evaporation loss. The runoff from plastic into the row middles was scaled appropriately using bed width and row spacing. Site 2 experienced flooding due to a large rainfall event, for which the atmospheric boundary switched to a specified head boundary.
 |
| Figure 1. HYDRUS model domain representing the cross-section of a raised bed for 46×30 geometry at Site 1. Depth of the domain below the ground surface (atmospheric boundary) was 200 cm, and the width was 244 cm. At Site 2, the model domain was similar but narrower (183 cm) and shallower (60 cm deep), with specified groundwater head imposed on the bottom boundary. |
Evaporation from the row middles and transpiration from plants were calculated using the dual crop coefficient methodology (Allen et al., 1998; Allen et al., 2005). First, reference ET was estimated with weather data using RefET software (Allen, 2000). Then, based on the crop stage of watermelon, evaporation and transpiration were calculated using the soil evaporation coefficient (Ke) and the basal crop coefficient (Kcb) (Appendix) derived from the lysimetric studies (Shrestha and Shukla, 2014).
The left and right boundaries of the model domain were assigned no-flux boundary condition (fig. 1) because there were other beds on either side, and therefore the conditions were assumed to be symmetrical about these boundaries. The bottom of the model domain was assigned a free-drainage boundary (deep water table) at Site 1 and a specified head boundary (shallow water table) at Site 2. Ground water level measurements at Site 2 were used to determine the head at the bottom boundary and at the ground surface (during flooding).
The drip emitter was represented as a line source of specified flux (fig. 1) calculated using measured irrigation rates averaged over the length of the drip tape. The plant hole was also represented as a line source (in 2D) for Site 2 during rainfall events, based on the field observation that moisture levels at the top of the bed spiked immediately after the rainfall. Initial soil moisture at Site 1 was assumed to be 12% throughout the model domain, and at Site 2, it was assumed to vary linearly from 15% at the top of the bed to saturation (˜ 40%) at the bottom of the domain (60 cm below ground surface).
Plant Water Uptake
Transpiration calculated using the dual crop coefficient methodology (Allen et al., 2005) corresponds to conditions with little or no water stress. The actual transpiration in HYDRUS was calculated using the Feddes’ model, which considers reduced transpiration under moisture stressed conditions (Feddes et al., 1978). Water used in transpiration was removed from the root zone defined within the bed. Depth of maximum density and maximum depth for watermelon roots were assumed to be 15 cm and 60 cm, respectively, at Site 1 based on literature values for raised bed plasticulture in sandy soils (Miller et al., 2013; NeSmith, 1999). At Site 2, the maximum rooting depth was taken to be 30 cm because it was assumed that saturation from shallow ground water limited the root growth. Maximum root density was assumed to occur at the midpoint between drip emitter and plant, based on the studies conducted on drip irrigated crops in raised bed plasticulture (Miller et al., 2013; Zotarelli et al., 2009).
Evaluation
The time step during the model run varied from one second to one hour, and the model output was generated in 15-minute to one-hour intervals. Models were evaluated using soil moisture data measured by TDR and multi-depth probes within the beds. The soil moisture data used for calibration and validation were collected from different beds from the same experiment. For Site 1, the model was calibrated using the data collected throughout the season from a 46×30 bed and validated using the data collected throughout the season from a 76×20 bed. Similarly, for Site 2, the model was calibrated using the data collected from a 46×30 bed and validated using the data collected from a different 46×30 bed. Using data from multiple growing seasons from the same location, to evaluate the model was not feasible because the growers followed a multiyear rotation cycle with forage grass (pasture) or other vegetable crops. Repeating the study at the same location would have resulted in conditions (initial N-P-K levels, soil hydraulic properties) that were different from the first season. The later seasons would not be representative of the usual growing conditions, and therefore calibration and validation of the model across seasons would be invalid. Model calibration involved finding the optimum values for soil parameters (a, n, ?r, ?s, and Ks) using the built-in inverse parameter estimation tool in HYDRUS. Once the model was evaluated, the same set of soil parameters were used in the models of other bed geometries at that site.
Nutrient Leaching
Water flux was calculated 30 cm below the top of the bed to coincide with the depth at which the pore water quality samples were collected. Advective flux of nutrients at this depth was calculated by multiplying water flux with measured nutrient concentrations for each bed geometry. Water flux was also calculated at the bed bottom and below the root zone to estimate potential nutrient leaching losses to groundwater. Simulation of nutrient cycling and transport was not considered due to a lack of understanding of nutrient (N and P) transformation processes under plastic mulched conditions and the large variability in the measured N and P concentrations.
Trigger Irrigation
A potential benefit of compact beds is that they might require less water than conventional beds because of their narrower bed width. In the highly conductive sandy soils of South Florida, water applied through drip irrigation quickly moves downward through the soil profile. Therefore, a longer duration of irrigation is required for the wider beds to allow sufficient time for water to move laterally and wet the entire width of the bed. Irrigation requirements based on bed width could not be investigated in the field experiments because the same amount of water was applied on all bed geometries. In commercial farms, a single irrigation system can cover a large area, and all the plants in that area receive the same irrigation rate. To evaluate irrigation requirements based on bed width and estimate water savings, trigger irrigation scenarios were simulated using HYDRUS for each bed geometry at Site 1. The irrigation volume used by the grower was considered the reference against which the trigger irrigation volumes were compared. Irrigation was triggered when the soil matric potential at a node 10 cm below the bed surface (Müller et al., 2016) at the center of the bed (fig. 1) fell below -16.5 kPa (SWC = 8.2%). This condition corresponded to 40% depletion (Card and Bauder, 2019) of the available water content (AWC) for the soil at Site 1. Once the condition was met, irrigation was triggered with a half-hour delay to avoid a second triggering before the waterfront reached the node, and it was applied at a rate of 1.24 l/h per emitter (1.67 mm/h per bedded area; observed rate at the farm) for a fixed duration. The duration for each bed geometry was manually calibrated such that the seasonal average SWC at a node 10 cm below the bed surface and 5 cm from the edge of the bed, at the far side of drip tape (fig. 1), was close to 8%. This target SWC was chosen to match the measured value at the same node in 61×10 bed geometry (grower standard) in the field experiment at Site 1. The rationale for choosing the node at the edge of the bed for calibration is that optimum soil moisture needs to be maintained throughout the bed, and the edge that is further from the drip tape tends to be drier than the center of the bed when irrigation is adequate. This rationale also represents the prevailing irrigation management strategy at commercial farms.
Trigger irrigation was not implemented for Site 2 because water was supplied primarily through upflux from ground water. Unlike a drip emitter, ground water is a non-point source that can supply water uniformly throughout the width of the bed. Therefore, the edges of wider beds are not expected to be drier under seepage irrigation.
Statistical Analyses
Statistical analyses were conducted to detect significant differences in yield, plant health, soil moisture, nutrient concentrations in pore water, soil matrix, leaf tissue, and daily leaching losses across bed geometries. The tests used linear mixed model procedures, assuming that the variation across bed geometries is a fixed effect and the variation across blocks is a random effect. The means of each data type from all the bed geometries were then differentiated by the Tukey-Kramer HSD test at the 95% confidence level.
Production Costs and Carbon Footprint
The differences in production cost and carbon footprint between the bed geometries were estimated based only on the inputs that differed, i.e., plastic mulch, fumigant, and water (trigger irrigation simulations) between bed geometries. The costs for fumigant and plastic mulches (standard and biodegradable) of various widths were obtained from the manufacturers. The disposal cost (dumpster contract) of plastic for conventional bed geometry was obtained from enterprise budgets for the production of fruits and vegetables in Florida (Wade et al., 2020a,b). Disposal costs for other bed geometries were estimated to be proportional to the weight of the plastic. Fuel usage for pumping for drip irrigation was determined to be 73 liters of diesel per million liters of water, based on personal communication with the grower-cooperator. The price of diesel was assumed to be $0.95/l based on the national average in the USA in October 2021. The costs for fertilizers were also obtained for the same timeframe.
Greenhouse gas emissions associated with fertilizer, plastic, fumigant, and diesel (for irrigation) were calculated based on the estimates for plasticulture in Florida (Jones et al., 2012). The study by Jones et al. (2012) presents detailed estimates of GHG emissions from various inputs for plasticulture, farm operations, field losses, and irrigation types. Out of these, only the emissions pertaining to the inputs that varied between the bed geometries were selected for this study. The emissions due to fertilizer and fumigant usage included activities such as manufacturing, transportation, and storage. The emissions due to the burning of the plastic were excluded because the plastic at Site 2 was disposed of in a landfill. The emissions due to biodegradable plastic used at Site 1 were assumed to be 50% of those from regular plastic (Álvarez-Chávez et al., 2012). To estimate the percentage change in emissions, total system emissions were considered based on the respective irrigation types (Jones et al., 2012) at Site 1 (drip) and Site 2 (sub-irrigation).
The same width (152 cm) of regular plastic was applied on all bed geometries at Site 2 due to the unavailability of narrower plastic rolls with the same specifications (thickness). Plastic mulch is generally available in 15 cm increments, and 137 cm wide plastic would have been sufficient for the two compact beds (46×30 and 41×30). Therefore, plastic widths of 152 cm for 76×20 and 61×30, and 137 cm for 46×30 and 41×30 were assumed for estimating costs and GHG emissions at Site 2. Similarly, the actual plastic widths used at Site 1 were found to be more than necessary for all the bed geometries except for 61×10. To be consistent with the calculations, plastic widths of 152 cm for 76×20, 122 cm for 61×10, and 137 cm for 46×30 and 41×30 were assumed to be sufficient at Site 1.
Scale-up
In the USA, watermelon is grown over an area of 40,000 ha annually, out of which Florida accounts for 10,700 ha and the Southern USA accounts for 31,400 ha (USDA, 2022). The input costs and GHG emissions were scaled up to estimate the benefit of the adoption of compact bed system at the state and Southern USA levels, where the growing conditions are expected to be similar to those in this study. For the scale-up, the differences in inputs such as fertilizer, plastic, fumigant, and diesel (irrigation) between the conventional bed (76×20) and the most compact bed (41×30) were considered. The differences for fertilizer were estimated based on the observed nutrient concentrations in the soil matrix, and the differences for diesel were estimated based on simulated irrigation requirements. It was assumed that 75% of the watermelon acreage uses plasticulture, 25% uses fumigant, 80% uses drip irrigation, and 100% uses fertilizer. This assumption is based on a survey of Florida watermelon growers, which indicated that almost 95% of the farmers used plasticulture with drip irrigation and 25% of the growers used fumigant (Shukla et al., 2019). Our farm visits and interactions with growers, allied industry and university researchers, extension agents, and professionals outside Florida also indicated that all the watermelon growers used plasticulture and drip irrigation. The percentages were adjusted to accommodate some unknown growers, especially those who farm small areas, that may use open field production without drip irrigation. Crop productivity, calculated as the watermelon yield per unit of input, was estimated for N, P, K, plastic, fumigant, and diesel to quantify the system efficiencies of conventional and compact beds.
Results and Discussion
Hydrology
The two sites experienced different weather and hydrologic conditions during their respective growing seasons. Site 1 received 44 cm of rainfall over the entire growing season (Mar-Jun) of 104 days. Rainfall of 29 cm during the month of June was high compared to the 20-year average of 17 cm at the site. Site 2 received 39 cm of rainfall over the growing season (Sep-Nov) of 69 days. Rainfall of 28 cm during the month of November was unusually high compared to the 20-year average of 4 cm at Site 2. This was mainly due to a large event on 8 and 9 November, when a record 21 cm of rainfall occurred within 32 hours. The event caused extensive flooding in the row middles for 26 hours, and in some cases, the peak flood water levels reached the top of the beds.
Approximately 378 liters (50.9 cm per bedded area) of irrigation was applied per drip emitter during the growing season at Site 1. At Site 2, 19 liters of irrigation (3.4 cm per bedded area) were applied per drip emitter. The lower irrigation volume at Site 2 was mainly because the upflux from the shallow groundwater (seasonal average depth to water table = 28 cm) also contributed to the soil moisture within the rootzone.
Soil Moisture
Seasonal averages of soil moisture content at various depths within the beds at Site 1 ranged from 7% to 16%. The daily average moisture content within the top 10 cm, measured using TDR probes near the center of the bed, was higher in the wider beds (table 1). However, average moisture near the edge of the bed, measured using multi-depth sensors, was numerically higher in the compact beds, confirming the improved wetting of the bed.
| Table 1. Seasonal average volumetric moisture content (%) at various depths near the center (measured by TDR probe; 0-10 cm) and the edge (measured by multi-depth sensor; 15-45 cm) of the beds at Site 1. | ||||||
| Bed Geometry (width × height in cm) | Center | Edge | ||||
| 0-10 cm | 15 cm | 25 cm | 35 cm | 45 cm | ||
| 76×20 | 12.9 | 10.6 | 11.4 | 11.6 | 8.7 | |
| 61×10 | 10.8 | 11.0 | 12.5 | 8.5 | 6.8 | |
| 46×30 | 10.0 | 10.8 | 14.5 | 15.7 | 10.3 | |
| 41×30 | 10.0 | 14.5 | 14.1 | 12.2 | 10.2 | |
At Site 1, soil moisture at all depths responded to irrigation events, indicating percolation losses of water (and nutrients) below the beds. This finding, along with the observed moisture values (table 1) being generally higher than field capacity (10%) (USDA, 2019) in the top 30 cm, suggests that there is an opportunity to reduce the irrigation volumes, especially in 46×30 and 41×30 compact beds, because they can maintain greater moisture levels and avoid reaching the wilting point (5%) near the edges than the wider beds.
Seasonal average soil moisture content at Site 2, measured using TDR probes, within the top 10 cm at the center of the beds ranged from 10% to 15%. Average moisture 20 to 30 cm below the top of the beds ranged from 24% to 33%, which is much higher than the field capacity (13%) and close to saturation (36%). There was no clear relationship between moisture content and bed geometry at Site 2 due to the influence of upflux from the water table. Moisture at the top of the beds spiked immediately following rainfall events, which indicates rain or runoff from plastic entering the beds through the plant holes. Flooding of the field on 8 and 9 November completely saturated the beds for nearly 2 days. Overall, the moisture levels at Site 2 were unusually high due to the shallow water table.
Plant Health and Yield
Nutrient (N, P, K) concentrations in leaf samples and NDVI values were not statistically different, at the a = 0.05 level, among the bed geometries at both sites. This indicates that bed geometry did not significantly affect plant health or nutrient uptake.
Although the total marketable yields at Site 1 differed numerically by as much as 9% between the conventional and compact bed geometries (fig. 2), the differences were statistically not significant at the a = 0.05 level (p > 0.8). The yield differences between the bed geometries were also not significant within the triploid (p > 0.4) and diploid (p > 0.6) categories and within individual harvests at Site 1.
The numerical differences in total marketable yields between the conventional and compact bed geometries at Site 2 were as much as 17% (fig. 2). However, these differences were statistically not significant at the a = 0.05 level (p > 0.6) because of the high variability observed among bed geometry replications (across different blocks). Yields from Sites 1 and 2 are not comparable because the plant spacings, number of harvests (two for Site 1, one for Site 2), and watermelon varieties (larger melons at Site 1, personal melons at Site 2) differed between the sites. Moreover, saturation stress 20 days before harvest, caused by prolonged flooding at Site 2, adversely affected the yield. Overall, the results indicate that the changes in bed geometry had no significant effect on production and the associated economic returns.
 |
| Figure 2. Average total marketable yields (metric tons/hectare) with 95% confidence intervals for watermelons from each bed geometry at Sites 1 and 2. Melons at Site 1 included triploid (seeded) and diploid (seedless) captivation varieties that were harvested twice. Personal sized melons were grown at Site 2 and were harvested once. |
Nutrients
Soil Solution
Nutrient concentrations in pore water showed no trend with bed geometry at either site. The differences in concentrations between the bed geometries were also statistically insignificant (at a = 0.05) for most of the sampling events (7 at each site). In only a few cases at Site 1, significant differences were observed, such as in two sampling events conducted 60 and 74 days after planting, concentrations of NH4-N in soil solution 30 cm beneath the bed surface in 76×20 were significantly (p = 0.01) greater than those in all other bed geometries (61×10, 46×30, 41×30). NH4N in soil solution 15 cm beneath the bed surface in 61×10 was significantly (p = 0.07) greater than those in all other bed geometries (76×20, 46×30, 41×30) 60 days after planting. For a sampling event 14 days after planting, the TP concentration 15 cm beneath the bed surface in 61×10 was significantly (p = 0.06) greater than that in other bed geometries. For the same event, there was moderate evidence (p = 0.06) that the TKN concentration 30 cm beneath the bed surface in 76×20 was significantly greater than other bed geometries. The lack of consistency in these results is probably due to the high variability in sample volumes and measured concentrations between treatment replications. Nonuniform and irregular solution volume from the sampler is a common problem due to inherent soil heterogeneity that affects soil moisture and nutrient retention (Cochran, 1970; Litaor, 1988).
Soil Matrix
Concentrations of nutrients (N, P, K) in the soil matrix at Site 2 at planting were generally higher in the compact beds compared to the conventional bed (76×20) (fig. 3). A similar trend was also observed for Site 1. However, the differences between the geometries were significant (a = 0.05) only in some cases, such as the concentration of K within the top 15 cm of the bed at planting, which was significantly (p = 0.006) greater in 61×10 than all other geometries at Site 1. At Site 2, concentrations of NH4N, TKN, and K in the top 15 cm of soil at planting were significantly (p < 0.01) greater in 46×30 and 41×30 than in 76×20 and 61×30. The concentration of NOxN between depths 15-30 cm beneath the bed surface at planting was significantly (p = 0.03) greater in 46×30 than in 61×30. Post-harvest TKN concentration in the top 15 cm was significantly (p = 0.04) greater in 46×30 than in 76×20. These significant differences and the overall trend (fig. 3) of higher concentrations in compact beds were because the same mass of fertilizer was applied to all bed geometries, but compact beds have a smaller soil volume within the bed. Average N-P-K concentrations between the sites in the most compact bed (41×30) were 21%, 10%, and 47%, respectively, higher than 76×20 at the time of planting. Since no differences in yields were observed, these results indicate an opportunity for reducing fertilizer input in the compact beds by (10%-30%) to improve productivity and sustainability, while reducing the input cost.
 |
| Figure 3. Concentrations of (a) Total N, (b) Total P, and (c) Total K in soil matrix of all bed geometries at Site 2 for two depth ranges (0-15 cm and 15-30 cm) at planting. |
A comparison of the actual fertilizer rates applied at the sites (N-P-K kg/ha at Site 1: 245-24-197; Site 2: 303-44-391) with the recommended rates (168, 0-49, 0-111 kg/ha), by the University of Florida (Liu et al., 2022), shows that the application exceeded the recommendation. The reason for overapplication is that growers generally perceive the current recommendations to be insufficient for optimum yield, and therefore reducing the rates is considered a production risk. Higher nutrient concentrations in compact beds provides growers an opportunity to reduce the rates at least by as much as the relative difference in concentrations between 41×30 and 76×20 (fig. 3). Farms in North America that use raised bed plasticulture generally apply 10%-50% of the N and K fertilizer and 100% of the P fertilizer in dry form within the bed, while the rest of the N and K fertilizer is applied as liquid through drip irrigation (fertigation). In farms with highly conductive sandy soils (e.g., Florida), growers can reduce the dry fertilizer as a conservative measure if they were to adopt compact beds.
Water and Nutrient Losses
The HYDRUS-2D model predictions of measured soil moisture data were deemed “satisfactory” to “very good” (Moriasi et al., 2007) based on Nash-Sutcliffe Efficiency values, which ranged from 0.76 to 0.82 during model calibration and from 0.68 to 0.82 during model validation. The calibrated values of soil hydraulic parameters were within the acceptable range for sandy soils (Carsel and Parrish, 1988), and they are presented in the Appendix. Simulation results showed that soil moisture at Site 1 was mainly affected by drip irrigation, whereas at Site 2, it was mainly driven by rainfall entering the bed through the plant hole and by upflux from shallow groundwater.
Water Flux at Bed Bottom
The net seasonal water flux across the bed bottom was larger in the shorter beds compared to the taller beds at Site 1. The flux (positive downward), normalized to the bedded area, was 17% and 50% larger in the shorter conventional (76×20) and grower standard (61×10) beds, respectively, compared to the taller compact beds (46×30, 41×30) (table 2). This was mainly because shorter beds have (1) shorter travel times to the bed bottom, and (2) lesser water uptake within the bed due to a larger portion of roots extending below the bed bottom than the taller compact beds.
The seasonal flux across the bed bottom at Site 2 was similar across bed geometries (table 2), as it was also affected by the upflux from the shallow groundwater. For example, the larger downward flux on some days from the shorter conventional bed (76×20) was counterbalanced by the larger upward flux into the bed from groundwater on other days due to the bed being wider. The net downward flux at Site 2 was less than that at Site 1 because of the smaller drip irrigation volume and upward flux from groundwater. It should also be noted that the flux during flooding at Site 2 (Nov 8 and 9, 2021) was excluded from seasonal flux calculation because of high uncertainty in surface and ground water interactions and unaccounted flows of floodwater entering the field from outside the experimental area.
| Table 2. Seasonal net downward water flux (cm), scaled to the bedded area, at the bottom of the bed and below the rootzone (60 cm for Site 1 and 30 cm for Site 2) for all bed geometries at both sites. Season length was 104 days at Site 1 and 69 days at Site 2. | |||||
| Bed Geometry | Site 1 | Site 2 | |||
| At Bed Bottom | Below Rootzone | At Bed Bottom | Below Rootzone | ||
| 76×20 | 34.6 | 60.5 | 8.8 | 17.9 | |
| 61×10 | 44.2 | 58.9 | - | - | |
| 61×30 | - | - | 8.7 | 16.9 | |
| 46×30 | 29.3 | 58.6 | 9.0 | 16.1 | |
| 41×30 | 29.6 | 58.5 | 8.8 | 15.5 | |
Flux below the rootzone was greater than the flux at bed bottom (table 2) because the former was calculated for the entire width of the model domain (fig. 1) and includes infiltration from rainfall and evaporation in the row-middle (non-mulched), whereas the latter only covered the bottom width of the plastic mulched bed. Flux below the root zone increased with bed width, and it was 3% and 15% larger in the widest bed (76×20) compared to the narrowest bed (41×30) at Sites 1 and 2, respectively (table 2). This was mainly because narrower beds have a greater area of exposed soil between the beds (row-middles), which causes greater evaporation from the soil compared to the wider beds.
Results show that the depth to water table and the method of irrigation can profoundly affect the nature of water and nutrient transport from a raised bed plasticulture system. Greater downward water flux from systems that depend mainly on drip irrigation can also potentially cause greater advective leaching of nutrients to the groundwater. However, in sub-irrigation systems with highly saturated soils and shallow water table, nutrient loss can also occur due to dispersive transport within the soil or advective transport in surface runoff.
Nutrient Leaching
Daily advective nutrient flux 30 cm below the bed surface on the days when biweekly soil solution samples were collected was statistically similar (at a = 0.05 level) across the bed geometries at Site 1. The only exception was that NH4N flux was significantly (p = 0.01) greater in 76×20 than all other bed geometries at 60 and 74 days after planting. Because the rootzone at Site 1 was assumed to be 60 cm deep, the advective flux at 30 cm depth does not include total nutrient leaving the rootzone. However, it is a close approximation to the potential nutrient loss because 85% of the watermelon roots are expected to occur within 30 cm (Miller et al., 2013). At Site 2, daily TP flux below rootzone (30 cm) was significantly (p = 0.02) greater in 76×20 than the compact beds (46×30 and 41×30) 37 days after planting. Daily NOxN flux was significantly (p = 0.01) greater in 61×30 than all other bed geometries 55 days after planting. No other significant differences (at a = 0.05 level) were detected at Site 2. Although there was some evidence of reduced nutrient leaching from the compact beds at both sites, the results were not consistent enough to draw firm conclusions.
Average seasonal advective flux of dissolved inorganic nitrogen (NOxN + NH4N), 30 cm below the bed surface, was 111 kg/ha at Site 1 and 29 kg/ha at Site 2. When compared to the total N applied at the two sites (Site 1 = 245 kg/ha, Site 2 = 303 kg/ha), the estimated leaching losses are almost 45% and 10% of the total input. The smaller loss for Site 2 is probably an underestimation because advective leaching to groundwater is not the dominant mechanism of nitrogen loss. Other mechanisms such as dispersion, denitrification (Shukla et al., 2011), and surface runoff could be important because of high soil moisture due to the shallow water table. Overall, the results indicate that there is an opportunity to reduce both irrigation and fertilizer inputs at both sites, and the growers may be more incentivized to do so given the prevailing high prices of fertilizer.
Water Savings with Trigger Irrigation
Trigger irrigation simulations show that narrower beds (46×30 and 41×30) at Site 1 needed up to 41% less water than that applied at the farm to maintain the same average moisture level (SWC = 8%) as the grower standard (61×10) near the edge of the bed (table 3). Trigger irrigation by itself reduced water input by 36% in 61×10 when compared to the actual irrigation applied with manual control. Bed geometry 41×30 required 17% (6 cm) less water compared to 76×20 (table 3).
| Table 3. Comparison of water budgets for all bed geometries under simulated trigger irrigation with the water budget for the grower standard (61×10) bed under actual irrigation at Site 1. Positive values indicate input, and negative values indicate loss from the system. | ||||||
| Flux/Bedded Area (cm) | Actual Irrigation | Trigger Irrigation | ||||
| 61×10 | 76×20 | 61×10 | 46×30 | 41×30 | ||
| Rainfall | 44.2 | 44.2 | 44.2 | 44.2 | 44.2 | |
| Irrigation | 50.9 | 35.9 | 32.6 | 31.2 | 29.8 | |
| Transpiration | -27.3 | -27.3 | -27.3 | -27.3 | -27.3 | |
| Evaporation | -9.0 | -7.9 | -9.0 | -9.4 | -9.5 | |
| Loss below rootzone | -58.9 | -46.9 | -42.1 | -40.0 | -38.7 | |
The plant water uptake was not affected by the irrigation volume because the moisture levels in all bed geometries were more than adequate. Plant uptake was equal to the potential transpiration of 27.3 cm per bedded area (table 3), which occurs when the soil moisture is unlimited. Evaporation losses increased with the width of the row middles, which increased as the bed width decreased. However, the absolute differences in evaporation between bed geometries were small because the magnitude of evaporation was small compared to the other components of the water budget (table 3). Water losses below the rootzone decreased with bed width, and the differences in losses between bed geometries were close to the differences in irrigation volumes. This indicates that any water applied beyond the ET demand in drip irrigated plasticulture is lost through deep percolation to groundwater. Reduction in water loss and the potential for reduction in fertilizer can translate to a reduction in nutrient leaching to groundwater from the compact beds.
Economics and Carbon Footprint
The costs and greenhouse gas emissions, due to the inputs that varied among the bed geometries, generally decreased with bed width (tables 4 and 5). The input costs will be reduced by up to $100/ha at Site 1 and $520/ha at Site 2, per growing season, when growers switch from conventional to the most compact bed (tables 4 and 5). The cost reductions at Site 1 were mainly due to reductions in the width of the plastic used, and at Site 2, they were mainly due to the reductions in fumigant.
The greenhouse gas emissions can be reduced by up to 380 kg CO2e/ha at Site 1 and 1200 kg CO2e/ha at Site 2, per growing season, by adopting the most compact bed (tables 4 and 5). These reductions were 1.7% and 7.6% of the total emissions from the production systems, with specific irrigation types present at Sites 1 (drip) and 2 (subirrigation and drip), respectively. The GHG emissions for the two irrigation types were taken from Jones et al. (2012).
Differences in cost and carbon footprint in tables 4 and 5 are conservative as they are based only on directly quantifiable input differences in this study. Additional reductions in costs and emissions are possible due to a reduction in fertilizer, which was not included in the calculations because it was not varied by bed geometry. At some farms, the plastic is burned at the end of the season instead of being disposed of in a landfill, and this will further increase the GHG emissions by 11%. Furthermore, the risk of hidden costs such as yield losses due to flooding and associated diseases is lower with the taller compact beds.
| Table 4. Seasonal input costs and carbon footprints for the inputs that varied between the bed geometries at Site 1. | |||||||||
| 76×20 | 61×10 | 46×30 | 41×30 | ||||||
| Inputs | Biodegradable plastic width (cm) | 152 | 122 | 137 | 137 | ||||
| Diesel for trigger irrigation (l/ha) | 262 | 238 | 228 | 217 | |||||
| Costs ($/ha) | Biodegradable plastic | 1369 | 1223 | 1302 | 1302 | ||||
| Diesel | 249 | 226 | 216 | 207 | |||||
| Total | 1618 | 1449 | 1518 | 1509 | |||||
| GHG Emissions (kg CO2e/ha) | Biodegradable plastic | 2684 | 2154 | 2419 | 2419 | ||||
| Diesel | 701 | 637 | 609 | 582 | |||||
| Total | 3385 | 2791 | 3028 | 3001 | |||||
Table 5. Seasonal input costs and carbon footprints for the inputs that varied between the bed geometries at Site 2. | |||||||||
| 76×20 | 61×30 | 46×30 | 41×30 | ||||||
| Inputs | Plastic width (cm) | 152 | 152 | 137 | 137 | ||||
| Fumigant (kg/ha) | 140 | 122 | 96 | 92 | |||||
| Costs ($/ha) | Plastic | 741 | 741 | 667 | 667 | ||||
| Fumigant | 1282 | 1118 | 882 | 841 | |||||
| Disposal | 90 | 90 | 81 | 81 | |||||
| Total | 2113 | 1949 | 1630 | 1589 | |||||
| GHG Emissions (kg CO2e/ha) | Plastic | 5367 | 5367 | 4830 | 4830 | ||||
| Fumigant | 2005 | 1746 | 1374 | 1313 | |||||
| Total | 7372 | 7113 | 6204 | 6143 | |||||
Scale-Up
For upscaling the potential benefits to the Florida state and Southern USA (SUSA), the differences in plastic, fumigant, and diesel (irrigation) between the conventional bed (76×20) and the most compact bed (41×30) (tables 4 and 5) were used. Fertilizer reduction was also included in the upscaling based on the observed nutrient concentrations in the soil matrix between 76×20 and 41×30 (fig. 3). Fertilizer input was reduced for the compact bed (smaller bed volume) such that it has the same initial concentration at planting as the conventional bed (larger bed volume). This estimate is conservative considering that the current fertilizer application rates are much higher than those recommended by the University of Florida (Liu et al., 2022) for optimum production. The cost savings due to fertilizer reduction were $160/ha, and the associated reductions in GHG emissions were 280 kg CO2e/ha.
Total input cost savings per season, when switched from 76×20 to 41×30, amount to $720/ha, and the total reductions in GHG emissions amount to 1630 kg CO2e/ha. The GHG emissions for the plasticulture system are only available for tomato (Jones et al., 2012), which has higher inputs (e.g., pesticides and plastic) than watermelon. The estimated GHG reductions in this study are 9% of the total emissions from tomato plasticulture system averaged over all irrigation types (Jones et al., 2012). Therefore, the reductions for the watermelon plasticulture system are expected to be greater than 9%. When upscaled for economic benefit, the annual input cost savings add up to $3.8 million in Florida and $11.3 million in the Southern USA for watermelon (fig. 4). Fertilizer was the largest contributor to the savings, and given that the prices have increased significantly since 2021, the current and future savings are expected to be even greater. The reductions in emissions add up to 10 kiloton CO2e in Florida and 30 kiloton CO2e in the Southern USA for watermelon (fig. 4). Projected water savings that can be achieved by adopting compact beds with trigger irrigation for growing watermelon in Florida and the Southern USA are 5 and 15 billion liters (10,000 ha-cm), respectively. This saved water can irrigate an additional area of 1,700 ha in Florida and 5,100 ha in the Southern USA with a similar production system, or it can be diverted to agriculture or other sectors, especially in the water scarce regions.
 |
| Figure 4. Estimated savings in input costs (US$) and greenhouse gas (GHG) emissions (kilotons CO2e) from adopting compact beds (41×30) in place of conventional beds (76×20) in Florida and the Southern USA (SUSA). |
Watermelon productivities for all the inputs were greater for the compact bed (fig. 5), indicating an improvement in system efficiency over the conventional bed. The greatest improvement was seen in fumigant use (52%). Cost savings due to the greater efficiency of compact beds can add up to $72,000 annually for a typical medium-scale (100 ha) farm. These savings can help farmers cover the costs of modifying bedding and plastic laying machines to make compact beds and to invest in other conservation practices such as precision irrigation management, chemigation, and slow-release fertilizers.
Conclusions
Switching from wider and shorter conventional beds to narrower and taller compact beds in raised bed plasticulture for watermelon can reduce water and nutrient losses, input costs, and greenhouse gas emissions. Compact beds were evaluated at commercial farms in Florida in contrasting growing conditions such as seasons (spring vs. fall), irrigation types (drip vs. sub-irrigation), hydrology (well drained vs. poorly drained), and watermelon variety (large vs. personal size). Results show that growers are not expected to be at an economic disadvantage if they adopt the compact beds because there is no adverse effect on the yield. The adoption of compact beds can be a win-win for farmers and the environment because compact beds allow for reducing inputs and costs while also reducing nutrient leaching and GHG emissions.
 |
| Figure 5. Percent improvement in watermelon productivity (kg watermelon / unit of input) for various inputs when switched from the conventional (76×20) to the most compact (41×30) bed. The baseline productivity values of each input for 76×20 are shown in parentheses. |
Reductions in irrigation volume by up to 6 cm (17%) and loss of water below the root zone by 8 cm provide both water savings and water quality benefits. Watermelon producers can potentially save 15 billion liters of water if compact beds are adopted in the Southern USA. Nutrient concentrations in the soil matrix were consistently greater in the compact beds compared to the conventional beds due to the smaller soil volumes for the former. This can incentivize the growers to reduce fertilizer input by 10%-30% without risking yields. Reductions in water and fertilizer input are likely to reduce the leaching and runoff of nutrients to waterbodies.
Inputs such as fertilizer, plastic mulch, fumigant, and diesel (irrigation) generally decrease with the width of the bed. Therefore, switching to the most compact bed (41×30) from the conventional bed (76×20) can lower input costs by $720/ha and GHG emissions by 1630 kg CO2e/ha. These differences are expected to be larger when other benefits, such as reductions in flooding and the associated saturation stress and diseases, are considered for taller compact beds. The cost savings can help growers cover the cost of adoption machinery and invest in other conservation practices such as irrigation automation and slow-release fertilizers. The industry can be incentivized to adopt the compact bed system, as evidenced by the adoption by two of the top ten tomato producers at all their farms in USA. One of the lessons learned from the wider adoption of compact bed technology in the USA for tomatoes is to conduct experiments on commercial farms to make it relevant and scalable.
Compact beds offer a promising climate change mitigation and adaptation strategy for raised bed plasticulture. By reducing the input requirements, compact beds can conservatively achieve at least a 9% reduction in GHG emissions, which is 20% of the 43% reduction goal by 2030 set by the Intergovernmental Panel on Climate Change (IPCC, 2022). Compact beds can improve the resiliency of plasticulture in dealing with extreme weather resulting from climate change. They can lower the risk of flood damage to the crop because the beds are taller, and they can help conserve water in drought prone areas because of improved water use efficiency. Future studies targeting crop-soil-weather specific compact bed designs, and field-verification of bed-specific irrigation and fertilizer management practices on commercial farms, are needed to expand their adoption in the USA and other parts of the world. Similar studies are also needed for agronomic crops, especially in parts of the world with an arid climate and limited water availability, such as China, where plasticulture is expanding for its water conservation benefit.
Acknowledgments
Funding for this study was provided by the Florida Department of Agriculture and Consumer Services, Southwest Florida Water Management District, and the Institute of Food and Agricultural Sciences, University of Florida. The authors acknowledge the help of the two Florida grower-cooperators who provided land, time, input, personnel, and encouragement for the research. Without their help, this study could not have been completed.
Appendix
| Table A1. Calibrated soil hydraulic parameter values for soil layers at sites 1 and 2. Values that differed between soil layers were shown as a range. | |||
| Parameter | Description | Site 1 | Site 2 |
| a | van Genuchten model parameter | 0.02 | 0.09 ? 0.13 |
| n | van Genuchten model parameter | 2.5 ? 2.6 | 1.3 ? 2.3 |
| ?r | Residual water content (%) | 2.0 | 3.9 ? 5.0 |
| ?s | Saturated water content (%) | 37 ? 40 | 37 ? 38 |
| Ks | Saturated hydraulic conductivity (cm/h) | 20 | 10 |
| Table A2. Dual crop coefficients used to calculate evaporation and transpiration in the model for various growth stages of watermelon. | |||
| Days After Transplant | Basal Crop Coefficient (Kcb) | Soil Evaporation Coefficient (Ke) | Crop Coefficient (Kc) |
| 0-13 | 0.05 | 0.56 | 0.61 |
| 14-27 | 0.22 | 0.55 | 0.77 |
| 28-41 | 0.41 | 0.31 | 0.72 |
| 42-55 | 1.01 | 0.05 | 1.06 |
| 56-69 | 0.93 | 0.05 | 0.98 |
| 70-Harvest | 0.66 | 0.05 | 0.71 |
References
Allen, R. G. (2000). Ref-ET: Reference evapotranspiration calculation software for FAO and ASCE standardized equations. University of Idaho.
Allen, R. G., Pereira, L. S., Raes, D., & Smith, M. (1998). Crop evapotranspiration: Guidelines for computing crop water requirements. Irrigation and Drainage Paper No. 56. Rome, Italy: United Nations FAO.
Allen, R. G., Pereira, L. S., Smith, M., Raes, D., & Wright, J. L. (2005). FAO-56 dual crop coefficient method for estimating evaporation from soil and application extensions. J. Irrig. Drain. Eng., 131(1), 2-13. https://doi.org/10.1061/(ASCE)0733-9437(2005)131:1(2)
Álvarez-Chávez, C. R., Edwards, S., Moure-Eraso, R., & Geiser, K. (2012). Sustainability of bio-based plastics: General comparative analysis and recommendations for improvement. J. Cleaner Prod., 23(1), 47-56. https://doi.org/10.1016/j.jclepro.2011.10.003
Arnhold, S., Ruidisch, M., Bartsch, S., Shope, C. L., & Huwe, B. (2013). Simulation of runoff patterns and soil erosion on mountainous farmland with and without plastic-covered ridge-furrow cultivation in South Korea. Trans. ASABE, 56(2), 667-679. https://doi.org/10.13031/2013.42671
Bennett, E. M., Carpenter, S. R., & Caraco, N. F. (2001). Human impact on erodable phosphorus and eutrophication: A global perspective: Increasing accumulation of phosphorus in soil threatens rivers, lakes, and coastal oceans with eutrophication. Bioscience, 51(3), 227-234. https://doi.org/10.1641/0006-3568(2001)051[0227:Hioepa]2.0.Co;2
Bläsing, M., & Amelung, W. (2018). Plastics in soil: Analytical methods and possible sources. Sci. Total Environ., 612, 422-435. https://doi.org/10.1016/j.scitotenv.2017.08.086
Card, A., & Bauder, T. (2019). Determining irrigation run times with drip tape on specialty crops. Fact sheet 5.623. Colorado State University Extension.
Carsel, R. F., & Parrish, R. S. (1988). Developing joint probability distributions of soil water retention characteristics. Water Resour. Res., 24(5), 755-769. https://doi.org/10.1029/WR024i005p00755
Chakraborty, S., Tiwari, P. K., Sasmal, S. K., Misra, A. K., & Chattopadhyay, J. (2017). Effects of fertilizers used in agricultural fields on algal blooms. Eur. Phys. J. Spec. Top., 226(9), 2119-2133. https://doi.org/10.1140/epjst/e2017-70031-7
Cochran, P. H., Marion, G. M., & Leaf, A. L. (1970). Variations in tension lysimeter leachate volumes. Soil Sci. Soc. Am. J., 34(2), 309-311. https://doi.org/10.2136/sssaj1970.03615995003400020035x
Cordell, D., Drangert, J.-O., & White, S. (2009). The story of phosphorus: Global food security and food for thought. Global Environ. Change, 19(2), 292-305. https://doi.org/10.1016/j.gloenvcha.2008.10.009
Delin, S., & Stenberg, M. (2014). Effect of nitrogen fertilization on nitrate leaching in relation to grain yield response on loamy sand in Sweden. Eur. J. Agron., 52, Part B, 291-296. https://doi.org/10.1016/j.eja.2013.08.007
Dubrovsky, N. M., & Hamilton, P. A. (2010). Nutrients in the nation’s streams and groundwater: National findings and implications. Fact Sheet 2010-3078. USGS. https://doi.org/10.3133/fs20103078
FAO. (2020a). Emissions due to agriculture: Global, regional and country trends 2000-2018. Rome, Italy: United Nations FAO.
FAO. (2020b). The State of Food and Agriculture: Overcoming water challenges in agriculture. Rome, Italy: United Nations FAO. https://doi.org/10.4060/cb1447en
FAWN. (2021). Florida Automated Weather Network. University of Florida Institute of Food and Agricultural Sciences. Retrieved from https://fawn.ifas.ufl.edu/data/reports/
Feddes, R. A., Kowalik, P. J., & Zaradny, H. (1978). Simulation of field water use and crop yield. New York, NY: John Wiley & Sons.
Foley, J. A., Ramankutty, N., Brauman, K. A., Cassidy, E. S., Gerber, J. S., Johnston, M.,... Zaks, D. P. (2011). Solutions for a cultivated planet. Nature, 478(7369), 337-342. https://doi.org/10.1038/nature10452
He, Z. L., Zhang, M. K., Stoffella, P. J., Yang, X. E., & Banks, D. J. (2006). Phosphorus concentrations and loads in runoff water under crop production. Soil Sci. Soc. Am. J., 70(5), 1807-1816. https://doi.org/10.2136/sssaj2005.0204
Holt, N., & Shukla, S. (2016). Transforming the plasticulture production system through novel bed geometry design. Trans. ASABE, 59(3), 993-1003. https://doi.org/10.13031/trans.59.11408
Holt, N., Shukla, S., Hochmuth, G., Muñoz-Carpena, R., & Ozores-Hampton, M. (2017). Transforming the food-water-energy-land-economic nexus of plasticulture production through compact bed geometries. Adv. Water Resour., 110, 515-527. https://doi.org/10.1016/j.advwatres.2017.04.023
Holt, N., Sishodia, R. P., Shukla, S., & Hansen, K. M. (2019). Improved water and economic sustainability with low-input compact bed plasticulture and precision irrigation. J. Irrig. Drain. Eng., 145(7), 04019013. https://doi.org/10.1061/(ASCE)IR.1943-4774.0001397
Howarth, R. W. (2008). Coastal nitrogen pollution: A review of sources and trends globally and regionally. Harmful Algae, 8(1), 14-20. https://doi.org/10.1016/j.hal.2008.08.015
IPCC. (2021). Climate Change 2021: The Physical Science Basis. Contribution of Working Group I to the Sixth Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge, United Kingdom and New York: Cambridge University Press. https://doi.org/10.1017/9781009157896
IPCC. (2022). Intergovernmental Panel on Climate Change Press Release. Retrieved from https://www.ipcc.ch/2022/04/04/ipcc-ar6-wgiii-pressrelease/
Jaber, F. H., Shukla, S., & Srivastava, S. (2006). Recharge, upflux and water table response for shallow water table conditions in Southwest Florida. Hydrol. Process., 20(9), 1895-1907. https://doi.org/10.1002/hyp.5951
Jones, C. D., Fraisse, C. W., & Ozores-Hampton, M. (2012). Quantification of greenhouse gas emissions from open field-grown Florida tomato production. Agric. Syst., 113, 64-72. https://doi.org/10.1016/j.agsy.2012.07.007
Kravchenko, A. N., Snapp, S. S., & Robertson, G. P. (2017). Field-scale experiments reveal persistent yield gaps in low-input and organic cropping systems. Proc. Natl. Acad. Sci., 114(5), 926-931. https://doi.org/10.1073/pnas.1612311114
Lamont, W. J. (1996). What are the components of a plasticulture vegetable system? HortTechnology, 6(3), 150-154. https://doi.org/10.21273/horttech.6.3.150
Lamont, W. J. (2005). Plastics: Modifying the microclimate for the production of vegetable crops. HortTechnology, 15(3), 477-481. https://doi.org/10.21273/horttech.15.3.0477
Li, L., Luo, Y., Li, R., Zhou, Q., Peijnenburg, W. J., Yin, N.,... Zhang, Y. (2020). Effective uptake of submicrometre plastics by crop plants via a crack-entry mode. Nat. Sustain., 3(11), 929-937. https://doi.org/10.1038/s41893-020-0567-9
Litaor, M. I. (1988). Review of soil solution samplers. Water Resour. Res., 24(5), 727-733. https://doi.org/10.1029/WR024i005p00727
Liu, G., Simonne, E. H., Morgan, K. T., Hochmuth, G. J., Agehara, S., Mylavarapu, R., & Williams, P. (2022). Chapter 2. Fertilizer management for vegetable production in Florida: VPH ch. 2, CV296. In P. Dittmar, S. Agehara, & N. S. Dufault (Eds.), Vegetable production handbook of Florida (Vol. 2022). https://doi.org/10.32473/edis-cv296-2022
MEA (Millennium Ecosystem Assessment). (2005). Ecosystems and human well-being: Synthesis. Washington, DC: Island Press.
Miller, G., Khalilian, A., Adelberg, J. W., Farahani, H. J., Hassell, R. L., & Wells, C. E. (2013). Grafted watermelon root length density and distribution under different soil moisture treatments. HortScience, 48(8), 1021-1026. https://doi.org/10.21273/hortsci.48.8.1021
Moriasi, D. N., Arnold, J. G., Van Liew, M. W., Bingner, R. L., Harmel, R. D., & Veith, T. L. (2007). Model evaluation guidelines for systematic quantification of accuracy in watershed simulations. Trans. ASABE, 50(3), 885-900. https://doi.org/10.13031/2013.23153
Mualem, Y. (1976). A new model for predicting the hydraulic conductivity of unsaturated porous media. Water Resour. Res., 12(3), 513-522. https://doi.org/10.1029/WR012i003p00513
Müller, T., Ranquet Bouleau, C., & Perona, P. (2016). Optimizing drip irrigation for eggplant crops in semi-arid zones using evolving thresholds. Agric. Water Manag., 177, 54-65. https://doi.org/10.1016/j.agwat.2016.06.019
NeSmith, D. S. (1999). Root distribution and yield of direct seeded and transplanted watermelon. J. Am. Soc. Hortic. Sci., 124(5), 458-461. https://doi.org/10.21273/jashs.124.5.458
Ngouajio, M., Wang, G., & Goldy, R. G. (2008). Timing of drip irrigation initiation affects irrigation water use efficiency and yield of bell pepper under plastic mulch. HortTechnology, 18(3), 397-402. https://doi.org/10.21273/horttech.18.3.397
Pandey, C., Shukla, S., & Obreza, T. A. (2007). Development and evaluation of soil moisture-based seepage irrigation management for water use and quality. J. Irrig. Drain. Eng., 133(5), 435-443. https://doi.org/10.1061/(ASCE)0733-9437(2007)133:5(435)
SFWMD. (2021). DBHYDRO (Environmental Data). South Florida Water Management District. sfwmd.gov
Shennan, C., Muramoto, J., Lamers, J., Mazzola, M., Rosskopf, E. N., Kokalis-Burelle, N.,... Kobara, Y. (2014). Anaerobic soil disinfestation for soil borne disease control in strawberry and vegetable systems: Current knowledge and future directions. In ISHS Acta Horticulturae 1044: VIII International Symposium on Chemical and Non-Chemical Soil and Substrate Disinfestation (1044 ed., pp. 165-175). Leuven, Belgium: International Society for Horticultural Science (ISHS. https://doi.org/10.17660/ActaHortic.2014.1044.20
Shrestha, N. K., & Shukla, S. (2014). Basal crop coefficients for vine and erect crops with plastic mulch in a sub-tropical region. Agric. Water Manag., 143, 29-37. https://doi.org/10.1016/j.agwat.2014.05.011
Shukla, S., Boman, B. J., Ebel, R. C., Roberts, P. D., & Hanlon, E. A. (2010). Reducing unavoidable nutrient losses from Florida’s horticultural crops. HortTechnology, 20(1), 52-66. https://doi.org/10.21273/horttech.20.1.52
Shukla, S., Goswami, D., Graham, W. D., Hodges, A. W., Christman, M. C., & Knowles, J. M. (2011). Water quality effectiveness of ditch fencing and culvert crossing in the Lake Okeechobee Basin, Southern Florida, USA. Ecol. Eng., 37(8), 1158-1163. https://doi.org/10.1016/j.ecoleng.2011.02.013
Shukla, S., Hendricks, G., Hansen, K., & Sishodia, R. (2019). Design and evaluation of compact bed geometries for drip-irrigated watermelon. Survey report. Tallahasse, FL: Florida Department of Agriculture & Consumer Services.
Šimunek, J., Van Genuchten, M. T., & Šejna, M. (2012). The HYDRUS software package for simulating the two-and three-dimensional movement of water, heat, and multiple solutes in variably-saturated porous media. Technical manual, version 2.0.
SRWMD. (2020). Water Data Portal, Rainfall Station Interactive Map. Suwannee River Water Management District. Retrieved from http://www.mysuwanneeriver.org/portal/rainfall.htm
Tilman, D., Balzer, C., Hill, J., & Befort, B. L. (2011). Global food demand and the sustainable intensification of agriculture. Proc. Natl. Acad. Sci., 108(50), 20260-20264. https://doi.org/10.1073/pnas.1116437108
USDA. (1996). Soil survey of Levy County, Florida. United States Department of Agriculture, Natural Resources Conservation Service. Retrieved from https://archive.org/details/usda-soil-survey-of-levy-county-florida-1996
USDA. (2019). Web Soil Survey. Natural Resources Conservation Service, United States Department of Agriculture. Retrieved from https://websoilsurvey.sc.egov.usda.gov/App/HomePage.htm
USDA. (2022). Vegetables 2021 Summary. National Agricultural Statistics Service. United States Department of Agriculture. Retrieved from https://downloads.usda.library.cornell.edu/usda-esmis/files/02870v86p/zs25zc490/9593vz15q/vegean22.pdf
USEPA. (2020). Pesticide product label, Pic-Clor 60. Washington, DC: USEPA. Retrieved from https://www3.epa.gov/pesticides/chem_search/ppls/008536-00008-20200528.pdf
van Genuchten, M. T. (1980). A closed-form equation for predicting the hydraulic conductivity of unsaturated soils. Soil Sci. Soc. Am. J., 44(5), 892-898. https://doi.org/10.2136/sssaj1980.03615995004400050002x
van Schothorst, B., Beriot, N., Huerta Lwanga, E., & Geissen, V. (2021). Sources of light density microplastic related to two agricultural practices: The use of compost and plastic mulch. Environments, 8(4), 36. https://doi.org/10.3390/environments8040036
Vitousek, P. M., Aber, J. D., Howarth, R. W., Likens, G. E., Matson, P. A., Schindler, D. W.,... Tilman, D. G. (1997). Human alteration of the global nitrogen cycle: Sources and consequences. Ecol. Appl., 7(3), 737-750. https://doi.org/10.1890/1051-0761(1997)007[0737:HAOTGN]2.0.CO;2
Wade, T., Hyman, B., & McAvoy, E. (2020a). Constructing a Southwest Florida bell peppers enterprise budget: FE1088, 11/2020. EDIS, 2020(6). https://doi.org/10.32473/edis-fe1088-2020
Wade, T., Hyman, B., McAvoy, E., & Vansickle, J. (2020b). Constructing a Southwest Florida tomato enterprise budget: FE1087, 11/2020. EDIS, 2020(6). https://doi.org/10.32473/edis-fe1087-2020
Wang, Y., Ying, H., Yin, Y., Zheng, H., & Cui, Z. (2019). Estimating soil nitrate leaching of nitrogen fertilizer from global meta-analysis. Sci. Total Environ., 657, 96-102. https://doi.org/10.1016/j.scitotenv.2018.12.029
Yang, Y., He, Z., Stoffella, P. J., Yang, X., Graetz, D. A., & Morris, D. (2008). Leaching behavior of phosphorus in sandy soils amended with organic material. Soil Sci., 173(4), 257-266. https://doi.org/10.1097/SS.0b013e31816d1edf
Zhang, Y.-L., Wang, F.-X., Shock, C. C., Yang, K.-J., Kang, S.-Z., Qin, J.-T., & Li, S.-E. (2017). Effects of plastic mulch on the radiative and thermal conditions and potato growth under drip irrigation in arid Northwest China. Soil Tillage Res., 172, 1-11. https://doi.org/10.1016/j.still.2017.04.010
Zotarelli, L., Scholberg, J. M., Dukes, M. D., Muñoz-Carpena, R., & Icerman, J. (2009). Tomato yield, biomass accumulation, root distribution and irrigation water use efficiency on a sandy soil, as affected by nitrogen rate and irrigation scheduling. Agric. Water Manag., 96(1), 23-34. https://doi.org/10.1016/j.agwat.2008.06.007