Development of a Comprehensive Watershed Model Applied to Study Stream Yield under Drought Conditions

Abstract We developed a model code to simulate a watershed's hydrology and the hydraulic response of an interconnected stream‐aquifer system, and applied the model code to the Lower Republican River Basin in Kansas. The model code links two well‐known computer programs: MODFLOW (modular 3‐D flow model), which simulates ground water flow and stream‐aquifer interaction; and SWAT (soil water assessment tool), a soil water budget simulator for an agricultural watershed. SWAT represents a basin as a collection of subbasins in terms of soil, land use, and weather data, and simulates each subbasin on a daily basis to determine runoff, percolation, evaporation, irrigation, pond seepage, and crop growth. Because SWAT applies a lumped hydrologic model to each sub‐basin, spatial heterogeneities with respect to factors such as soil type and land use are not resolved geographically, but can instead be represented statistically. For the Republican River Basin model, each combination of six soil types and three land uses, referred to as a hydrologic response unit (HRU), was simulated with a separate execution of SWAT. A spatially weighted average was then taken over these results for each hydrologic flux and time step by a separate program, SWBAVG. We wrote a package for MODFLOW to associate each subbasin with a subset of aquifer grid cells and stream reaches, and to distribute the hydrologic fluxes given for each subbasin by SWAT and SWBAVG over MODFLOW's stream‐aquifer grid to represent tributary flow, surface and ground water diversions, ground water recharge, and evapotranspiration from ground water. The Lower Republican River Basin model was calibrated with respect to measured ground water levels, streamflow, and reported irrigation water use. The model was used to examine the relative contributions of stream yield components and the impact on stream yield and base flow of administrative measures to restrict irrigation water use during droughts. Model results indicate that tributary flow is the dominant component of stream yield and that reduction of irrigation water use produces a corresponding increase in base flow and stream yield. However, the increase in stream yield resulting from reduced water use does not appear to be of sufficient magnitude to restore minimum desirable streamflows.