Computationally efficient watershed-scale hydrological modeling: integrating HYDRUS 1D and KINEROS2 for coupled surface-subsurface analysis

Authors: Meles, Menberu; CHEN, LIN - University Of California, Riverside; Unkrich, Carl; AJAMI, HOORI - University Of California, Riverside; Bradford, Scott; SIMUNEK, JIRI - University Of California, Riverside; Goodrich, David - Dave

Submitted to: Journal of Hydrology
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 6/20/2024
Publication Date: 7/3/2024
Citation: Meles, M.B., Chen, L., Unkrich, C.L., Ajami, H., Bradford, S.A., Simunek, J., Goodrich, D.C. 2024. Computationally efficient watershed-scale hydrological modeling: integrating HYDRUS 1D and KINEROS2 for coupled surface-subsurface analysis. Journal of Hydrology. 640. https://doi.org/10.1016/j.jhydrol.2024.131621.
DOI: https://doi.org/10.1016/j.jhydrol.2024.131621

Interpretive Summary: We have developed a novel and efficient modeling framework to couple HYDRUS-1D and KINEROS2 (K2) for simulating watershed flow processes. This framework extends the capability of a hillslope-scale coupled H1D-K2 model to the watershed scale, employing cascades of connected rectangular planes and channel elements for surface flow and 1D soil profiles for representing unsaturated flow processes. Integration of the models includes boundary condition switching to account for surface ponding and water exchange between domains, alongside dynamic time-stepping and dimensionality reduction techniques. Benchmark simulations demonstrate the performance and efficiency of the new watershed model, with calibrated hydrographs and water balance components showing excellent agreement with observed data from the Walnut Gulch Experimental Watershed.

Technical Abstract: Watershed flow processes consist of partitioning, movement, storage, and redistribution of water fluxes in space and time. However, the integrated modeling of these processes is challenging due to computational burden, extensive data requirements, and/or reliance on simplifying assumptions. This study introduces a novel and computationally efficient modeling framework that leverages two state-of-the-art process-based models: HYDRUS-1D (H1D) for unsaturated flow and KINEROS2 (K2) for overland flow. The framework extends a hillslope-scale coupled H1D-K2 model to simulate watershed-scale processes, where H1D replaces the three-parameter Parlange's infiltration equation in the event-based K2 model. Boundary condition switching is employed to account for surface ponding and water exchange between the two model domains. The structure of the coupled watershed-scale H1D-K2 model consists of a cascade of connected rectangular planes, channel elements, and 1D soil profiles to simulate 1D overland flow, infiltration, unsaturated zone flow, and recharge. Computational efficiency relative to HYDRUS-2D is achieved through dynamic time-stepping approach and dimensionality reduction. The performance and efficiency of the new watershed model are demonstrated using benchmark watershed and/or hillslope simulations. The calibrated hydrographs and water balance components using Walnut Gulch Experimental Watershed data showed an excellent agreement with observed data.