HYDRO-XM: A computationally efficient, externally coupled, and flexible watershed model

Authors: Meles, Menberu; RAHMAN, A.T.M. SAKIUR - University Of California, Riverside; ŠIMUNEK, JIRÍ - University Of California, Riverside; AJAMI, HOORI - University Of California, Riverside; Unkrich, Carl; ACERO TRIANA, JUAN - University Of California, Riverside; GOULDING, JASON - University Of California, Riverside; Bradford, Scott

Submitted to: Journal of Hydrology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/29/2026
Publication Date: 7/2/2026
Citation: Meles, M.B., Rahman, A., Šimunek, J., Ajami, H., Unkrich, C.L., Acero Triana, J.S., Goulding, J., Bradford, S.A. 2026. HYDRO-XM: A computationally efficient, externally coupled, and flexible watershed model. Journal of Hydrology. 677(C). Article 135954. https://doi.org/10.1016/j.jhydrol.2026.135954.
DOI: https://doi.org/10.1016/j.jhydrol.2026.135954

Interpretive Summary: Watershed systems are inherently complex, with water moving and interacting across surface and subsurface domains in ways that vary over space and time. Accurately simulating these dynamics is critical for understanding water availability, managing aquifers, and predicting the impacts of land-use and climate change. However, existing models often face trade-offs between computational efficiency and process detail, making integrated, watershed-scale simulations difficult. This study introduces a computationally efficient, modular framework that bridges this gap by coupling three specialized, open-source models: KINEROS2 for overland flow, HYDRUS-1D for unsaturated flow, and MODFLOW for groundwater. By linking these models at the hillslope scale and extending to entire watersheds, the framework captures the dynamic exchange of water between surface and subsurface systems, including infiltration, soil moisture redistribution, and recharge to groundwater. Key innovations—such as boundary condition switching, dynamic time-stepping, and dimensionality reduction—enable stable, continuous simulation of complex hydrological processes without excessive computational cost. Benchmark tests show that the framework accurately reproduces streamflow, runoff, groundwater dynamics, and soil moisture patterns, yielding results comparable to both the ParFlow model and observed field data. This integrated tool provides researchers and water managers with a robust platform for evaluating managed aquifer recharge strategies, assessing land-use impacts, and examining the effects of climate variability at watershed scales. It demonstrates that high-resolution, process-based modeling can be achieved in a computationally efficient, scalable framework, offering new opportunities for sustainable water management and watershed planning.

Technical Abstract: Numerical simulation of watershed processes involves modeling the spatial and temporal dynamics of water movement, storage, and redistribution across surface and subsurface domains. Integrating these processes is challenging due to computational intensity, extensive data requirements, and the complexity of fluxes at surface–subsurface interfaces. We present a computationally efficient, modular framework that couples three state-of-the-art, open-source models: KINEROS2 (K2) for overland flow, HYDRUS-1D (H1D) for variably saturated flow, and MODFLOW for saturated groundwater flow. The framework begins with hillslope-scale three-way coupling and scales to entire watersheds using interconnected planar surfaces and channels. Each planar element is associated with a HYDRUS column, enabling dynamic exchange of excess rainfall with K2 and time-varying recharge and unsaturated zone fluxes with MODFLOW at each time step. K2 solves the one-dimensional kinematic wave and advection–dispersion equations, HYDRUS simulates unsaturated flow via the Richards equation including solute and reactive transport, and MODFLOW handles saturated groundwater flow. Key innovations include boundary condition switching, dynamic time-stepping, and dimensionality reduction allow stable, continuous simulation of overland flow, infiltration, soil moisture redistribution, and groundwater recharge. The framework replaces K2’s Parlange infiltration model with a HYDRUS-based mechanism and connects HYDRUS bottom fluxes directly to MODFLOW via source/sink terms. Benchmark simulations demonstrate accurate reproduction of streamflow, runoff, groundwater dynamics, and soil moisture distribution, with results comparable to the ParFlow integrated process-based model and observed data. This framework provides a robust tool for evaluating managed aquifer recharge strategies, land-use impacts, and climate variability effects at watershed scales.