Improving SWAT auto-irrigation functions for simulating agricultural irrigation management using long-term lysimeter field data

Abstract

Decreasing groundwater availability in the Texas High Plains has resulted in the widespread adoption of management allowed depletion (MAD) irrigation scheduling. Modeling of such practices and their effects on water balance components can be a cost-effective and time-saving alternative to field-based research. However, studies have identified deficiencies in the auto-irrigation algorithms in the Soil and Water Assessment Tool (SWAT) including the continuation of irrigation during the non-growing season and an inability to simulate growth stage-specific irrigation. Consequently, new and representative auto-irrigation algorithms were developed using 1) a uniform, single season MAD and 2) a growth stage-specific MAD with options for seasonal growth stage partitioning based on scheduled date and accumulated heat units. Comparisons with observed data from an irrigated lysimeter field showed improved model performance for simulations of irrigation amount and frequency and actual evapotranspiration. Minimal differences in leaf area index and yield were observed with the non-water stressed management.