Optimizing hydrologic resiliency in agricultural landscapes is essential for sustaining food production and ecosystem services, and requires a holistic understanding of the hydrologic cycle, including groundwater–surface water interactions. In this study, a large-scale (2056 km²), high-resolution (50–150 m), structurally complex, fully integrated groundwater–surface water model was developed for an agriculturally dominated watershed in the Northern Great Plains. The sensitivity of surface water flow and groundwater conditions was assessed in relation to wetland/depression storage capacity, soil hydraulic conductivity (Ks), field surface roughness (Manning’s n), as well as variability in precipitation and potential evapotranspiration (PET). The model evaluation covered the 2010–2015 period, which encompassed both flood and drought conditions. Results showed that streamflow rates in the watershed were more sensitive to 10% changes in precipitation or PET than to a 50% change in depression storage, a fourfold change in Ks, or a 50% change in n, underscoring the region’s susceptibility to weather variability. Among the landscape parameters tested, increasing depression storage had the greatest effect in reducing peak flood flows compared to increases in Ks or n; however, all three contributed to both flood peak mitigation and enhanced groundwater recharge. When model settings approximated predevelopment (primarily native prairie) conditions, simulated flood peaks were reduced by 22–25%. This study presents a fully integrated groundwater–surface water modeling framework to evaluate hydrologic drivers and the risk mitigation benefits of landscape management practices.