Abstract:
Currently, the assessment of hydro-meteorological forecast products primarily emphasizes qualitative comparisons of forecast performance, with limited emphasis on quantitative evaluations of product usability. Furthermore, existing evaluation metrics often fail to account for the inherent uncertainties associated with forecasts across different lead times. Therefore, the development of a usability evaluation index has become an urgent priority. This study integrates conventional accuracy evaluation indicators—such as the Threat Score , Mean Absolute Relative Error, and Nash-Sutcliffe Efficiency coefficient—with considerations of forecast uncertainty across varying lead times, to propose a comprehensive usability evaluation index. A case study was conducted in the lower reaches of the Jinsha River Basin, where both forecast accuracy and product usability were systematically analyzed. Results indicate that rainfall forecasts for the Jinsha River Basin, as well as inflow and cumulative inflow volume forecasts for the Wudongde Reservoir within 1- to 5-day lead times exhibit high usability. Additionally, flood process forecasts within 1-to 2-day lead times show high usability, while those within the 3-day forecast period are moderately usable. The research results provide technical support for the coordinated operation of reservoir groups in the Jinsha River Basin, and the proposed usability
evaluation index offers a reference for evaluating the usability of hydro-meteorological forecast products.