Enabling Efficient Use of Earth Observations for Hydrologic Model Evaluation

Modelers often create multiple scenarios of a hydrologic model to evaluate the effects of climate change and watershed management decisions on model outputs, test a hypothesis, understand the implications of an improved input parameter or source code, and above all, validate the accuracy of model calibrations. Spatially distributed Earth Observations (EO) offer opportunities for evaluating watershed processes at different spatial scales as opposed to evaluating only streamflow at a few locations. Yet, making EO data ready for model evaluation remains challenging due the difficulty of getting these data to interoperate with the model structure. This becomes even more labor and time-intensive when multiple model scenarios are involved. Currently, there is no open-source tool that can aid this EO-based evaluation process. To fill this gap, we are developing a python-based tool for the Soil and Water Assessment Tool (SWAT) which, for any user-provided dataset of soil moisture, evapotranspiration, and vegetation index with varying spatial and temporal resolution, automatically makes the datasets interoperable with SWAT’s geodatabase structure and simulation time-step and subsequently performs comparison of model outputs at subwatershed scales. A user-friendly graphical user-interface offers a reproducible workflow with numerical and visual aids comprising of timeseries plotting, multi-variable statistical calculations, and spatial statistical mapping analysis to evaluate any SWAT model regardless of watershed size and location. We envision this tool to become a one-stop platform for facilitating widespread use of EO in hydrologic model evaluation and analysis.