Performance Evaluation of Conceptual and Physically Based Hydrologic Models

摘要

This paper focuses on evaluating the performance of physically based distributed and conceptual models. and assesses their sensitivity to changes in the temporal and spatial sampling of rainfall. The Hydrologic Modeling System (HMS) was selected to represent conceptual hydrologic models. while MIKE-SHE and GSSHA were selected to represent distributed physically based models. This manuscript presents results with MIKE- SHE, while the poster at the conference will include results from GSSHA. The performance evaluation criterion is the overall agreement between observed and predicted hydrographs and the models' ability to predict time and magnitude of peak discharges and runoff volume. Both models were carefully calibrated and validated using numerous storm events for a 21.4 km2 watershed in northern Mississippi. The results indicated that MIKE- SHE captured the peak runoff discharges and total runoff volume better than HMS. However overall, the performance of both models was quite reasonable. To assess the models' requirements for rainfall information. an in-depth investigation of the impact of the spatial and temporal sampling of rainfall on the prediction accuracy of each model was conducted. The study showed that MIKE- SHE was more sensitive to both the spatial and temporal sampling of rainfall than HMS.