Investigating Soil Parameters Effect on Crop Yields and Hydrology at Field Scale in the Southeast US Using the Soil and Water Assessment Tool

Abstract

Over the past forty years, the Apalachicola–Chattahoochee–Flint (ACF) river basin in Alabama, Georgia, and Florida has been the subject of numerous litigation and research regarding water allocation. The state of Georgia’s heavy reliance on the ACF’s water resources for the city of Atlanta water supply and agricultural production has been a partial cause of this conflict between Alabama, Georgia, and Florida. Regional, watershed, and field-scale models have been employed by researchers to better understand the hydrology of this area; however, few studies exist focusing on proper multi-variable calibration and validation that include plant growth of cotton and peanut, surface runoff, soil moisture, and soil nitrate. Cotton and peanut are primary crops in this region and greatly affect the hydrology. In addition, this area is home to many different types of soils. Soil type and morphology can affect crop yields, but how different soils in Georgia effect crop yields in SWAT has yet to be quantified. The first objective of this study was to create, calibrate, and validate a field-scale model using the Soil and Water Assessment Tool (SWAT) of fields at a research station in the Lower Flint River Basin. The research station modeled is the Stripling Irrigation Research Park (SIRP) located in Camilla, Georgia and run by the University of Georgia (UGA). UGA provided all management information needed to create the model, including crop type, fertilizer rates, irrigation amounts, planting dates, harvest dates, and crop yields. Three fields were modeled, which grew corn, peanut, and cotton, respectively, after a winter cover crop of Rye and strip-tilling. Each field contained three duplicate plots with 9 different fertilizer/irrigation treatments and had two plots with berms surrounding the plots to isolate overland flow. Plot specific soil nutrients, soil texture, biomass, yields, LAI for cotton, TKN, surface runoff, and composite runoff nutrient samples were obtained for the growing year 2018. Multivariable calibration and validation for surface runoff, soil moisture, crop biomass, corn and peanut yields, LAI for cotton (yields for cotton were not available), nitrogen uptake by plants, soil nitrate, and nitrate in runoff were conducted in this study. The model performed very good for surface runoff, crop growth, and nitrogen uptake, and fair for soil moisture and nitrate cycling except for soil nitrate in peanuts. Calibration of each variable following runoff gradually improved surface runoff performance. Analysis of nitrogen and water balances over 30 years were also simulated and found nitrate leaching to be very low compared to what is generally expected in this area. However, removing soil moisture and soil nitrate calibration, respectively, resulted in higher leaching values. These results indicate calibrating with fewer variables and higher quality measured data can result in a more properly calibrated model. The second objective of this study was to use a field scale model to determine the effect of soil types in southwestern Georgia on crop yields and soil moisture. A SWAT model previously calibrated for a cotton-cotton-peanut rotation in Tifton, Georgia was used in this study with 30 years of weather data from NLDAS. 24 different types of soils covering over 98% of Region V Soil-Water Conservation District (SWCD) in the STATSGO map were selected and integrated into the model, with Tifton and Orangeburg covering 46% of the area. Soil properties from SSURGO were matched to the STATSGO soils and used in this study, allowing for the diversity of soils to be accounted for while also using a more detailed soils database. A multiple comparison analysis of the different soils was run with the native SSURGO Tifton soil used as the control. When under UGA Checkbook Irrigation, crop yields had little response to the different Georgia soil types tested in this study excepting for one very sandy soil. Overall yields were lower for all Georgia soils investigated without irrigation, but top 305mm of soil will have a larger response to soil parameterization. Soil moisture for the top layer showed much more variation and all soils were statistically significant compared to the control soil. Soil moisture tended to decrease as available water content decreased, clay content decreased, and hydraulic conductivity increased. Future research into individual soil parameters effect on yields and soil moisture is needed to better understand the relationship between crop yields and soil properties in SWAT.