A SINGLE-FIELD SWAT MODEL APPROACH TO EVALUATE IMPROVEMENTS IN PROCESS REPRESENTATION FOR PERCHED WATER TABLE DEPTH SIMULATION AND SOIL TEMPERATURE DYNAMICS

Abstract

While the Midwest has fertile soil ideal for agricultural production, in many cases these soils were historically wetlands that have been artificially drained. A soil profile with a seasonally high perched water table is not suitable for growing most crops. Subsurface (“tile”) drainage helps lower the soil water table and increase crop yield by moving water out of the farm field through subsurface waterways. However, tile drains are also a direct conduit for nutrients to leave the field. Nutrient runoff contributes to harmful algal blooms and dead zones, which threaten ecosystems and human health. The Soil and Water Assessment Tool (SWAT) is a hydrologic model used to simulate the effects of land use, topography, soil type, and climate on hydrology and nutrient runoff. Several states rely on SWAT to calculate total maximum daily loads (TMDLs) and guide progress toward water quality goals. However, previous studies have identified areas for improvement in SWAT’s ability to simulate cold-season processes. Research has also indicated that the calculation of the perched water table depth is not fully tied to soil moisture by layer or tile flow. To address this, (1) a single-field SWAT model was developed, (2) a physically based soil temperature equation created by Qi et al. (2016a,b; 2019) was integrated into the SWAT source code, (3) a new method of calculating perched water table depth by using soil water by layer above field capacity was added to the SWAT source code, and (4) the single-field SWAT model was calibrated to evaluate model improvements. Improvements to soil temperature simulation by Qi et al. (2016a,b; 2019) were found to improve both soil temperature simulation (RSR < 0.5) and some subsurface hydrology parameters when compared to measured data. Statistical results for the perched water table improvements to the SWAT source code were mixed for hydrologic variable outputs. However, this thesis is the first step in improving the simulation of the perched water table, and further improvements are needed.