Continental-scale hydrologic modeling using the HLM TETIS

This hydroshare provides the source code utilized for the model runs, calibration, input processing, data analysis and figure creation for the manuscript under review at JAWRA. The abstract of the manuscript is as follows: In this study, we evaluate the performance of the TETIS model structure of the Hillslope-Link Model (HLM), which is a distributed hydrologic model. We explore performance across the contiguous United States (CONUS) at 5046 United States Geological Survey (USGS) streamgages. Specifically, we compare observed daily discharge from 1981 through 2020 with long term continuous simulations from the HLM TETIS. To obtain model parameters across CONUS, we run calibration by partitioning the study area based on 234 HydroSHEDS level 5 basins and calibrating to a single representative location near the outlet of each basin. Next, we utilize the remaining USGS gages for validation. We assess the model performance with the Kling Gupta Efficiency (KGE) and bias. We find the median KGE across CONUS is 0.43, with 80% of the gages above 0 and 43% above 0.5. Furthermore, our results show there is a dependence of the model performance on climate regions, with arid basins performing worse than basins in cold and temperate regions. To improve the model performance, we recalibrate these arid basins and highlight an overall performance improvement. Next, we compare model performance between simulations with different precipitation inputs to examine the robustness of the selected model parameters. Overall, our study highlights the model’s flexibility in performing across regions with different runoff generation mechanisms and provides a basis for future.