Comprehensive evaluation of IMERG, ERA5-Land and their fusion products in the hydrological simulation of three karst catchments in Southwest China

Abstract

Study region: The study was carried out in three karst catchments in Southwest China. These catchments, Sancha, Liuzhou and Qianjiang, are located in the middle reaches of the Pearl River Basin with the catchment area of 17067 km², 46166 km², and 134137 km², respectively. Study focus: Satellite or reanalysis precipitation data are potential alternatives that can be used in hydrological models for streamflow forecasting in data-sparse catchments. This study aims to investigate the performance of two widely used precipitation datasets, IMERG and ERA5-Land, as well as their fusion data with rain gauge measurements by the Geographically Weighted Regression method, in three karst catchments of southwest China. New hydrological insights for the region: The results indicate that compared to IMERG, ERA5-Land has a higher correlation coefficient and a better detection rate for daily precipitation with the gauge-based data. However, ERA5-Land overestimates annual precipitation in all three catchments. Merging ERA5-Land and IMERG with rain gauge data can further improve its precipitation detection rate, but does not effectively mitigate its overestimation. Meanwhile, the model calibration through the parameter adjustment can partly compensate for the precipitation error to improve the discharge simulation accuracy, it cannot fully cover the overestimation of ERA5-Land in the two catchments. Therefore, ERA5-Land and its fusion data consistently perform badly in streamflow simulation. In contrast, IMERG demonstrates good agreement with gauge-based precipitation data and shows comparable or even slightly superior performance in streamflow prediction in all catchments. Additionally, the calibrated parameters in the model driven by IMERG closely resemble those driven by gauge-based precipitation data. These results suggest that IMERG can serve as a viable precipitation alternative for hydrological models in discharge simulations in the study region.

Keywords

• IMERG
• ERA5-Land
• GWR
• Hydrological forecasting
• Southwest China