Coupled surface-groundwater modeling of water stress-triggered irrigation and climate change effects on maize production in Southwestern Kansas

Sustained groundwater depletion in the Ogallala Aquifer threatens the long-term viability of irrigated agriculture in the U.S. High Plains. This study evaluates deficit irrigation as an adaptation strategy using the coupled surface–groundwater model SWAT + gwflow in the Sunset Memorial watershed of southwestern Kansas. The model was calibrated and validated for the period 2001–2024 using streamflow and groundwater observations and then applied to five water-stress-triggered irrigation practices IP1 to IP5 (Ws < 0.9, 0.8, 0.7, 0.6, and 0.5). Model performance was moderate for streamflow (NSE = 0.51; PBIAS = 28.59) and strong for both groundwater heads (NSE = 0.86; PBIAS = 0.25) and maize yield (PBIAS = 1.37%; Willmott agreement index = 0.85). Across the five irrigation practices, irrigation demand declined by 18.5% for alfalfa and 21.4% for maize, while maize yield showed no significant temporal trend. Groundwater heads, however, declined significantly in all irrigation practices, with median Sen’s slopes ranging from −0.64 to −0.53 m year⁻¹ , indicating persistent and regional aquifer depletion. The statistical analysis revealed that irrigation practice 3, in which irrigation was triggered when plant water stress falls below 0.7, maintains crop yield and reduces irrigation by 10.7%, leading to a 7–9 cm year −1 reduction in groundwater decline. Future climate simulations under SSP126, SSP245, and SSP585 showed significant yield declines (0.01–0.04 tons ha −1 year −1 ) and continued groundwater depletion (0.4–0.34 m year −1 ), although moderate irrigation reduced losses compared with higher irrigation levels. The areas with groundwater decline greater than 0.5 m year −1 account for 33.1%, 18.9%, and 14.2% in IP1, IP3 and IP5under SSP245, respectively, which increase to 36.0%, 21.4%, and 17% under SSP585. The identified water-stress irrigation threshold (Ws<0.7) offers an actionable scheduling trigger for producers, while the spatial vulnerability mapping provides regions that need urgent intervention, underscoring that irrigation scheduling improvements must be paired with broader policy and agronomic measures to achieve long-term aquifer sustainability.