Evaluation of Routed-Runoff from Land Surface Models and Reanalyses Using Observed Streamflow in Chinese River Basins

Previous studies have demonstrated that offline land surface models (LSMs) and global hydrological models (GHMs) can reasonably reproduce streamflow in large river basins. Global reanalyses supply fine spatiotemporal runoff estimates, but they are not fully intercompared and evaluated in China. This study assesses the routed-runoff from five offline LSM/GHM runs (VIC-CN05.1, CLM-CFSR, CLM-ERAI, CLM-MERRA, and CLM-NCEP) and three reanalysis datasets (ERAI/Land, JRA55, and MERRA-2) against the gauged streamflow (26 stations) in major Chinese river basins during 1980–2008. The Catchment-based Macro-scale Floodplain model (CaMa-Flood) is employed to route those runoff datasets to the hydrological stations. Four statistical quantities, including the correlation coefficient (R), standard deviation (STD), Nash–Sutcliffe efficiency coefficient (NSE), and relative error (RE), along with a ranking method, are used to quantify the quality of those products. The results show that the spatial patterns of both modeled and observed streamflow in summer are similar, but their magnitudes are different. Except for MERRA-2, the other products can reproduce well the interannual variability of streamflow in both the Yangtze and Yellow River basins. All products generally underestimate the magnitude and variance of monthly streamflow, while VIC-CN05.1 and JRA55 are closer to observations compared to other products. The correlation coefficients for all products are overall larger than 0.61, with the highest value (0.85) from VIC-CN05.1. In addition to CLM-MERRA, MERRA-2, and CLM-NCEP with relatively small precipitation, other products can simulate peak flow well with positive NSEs up to 0.41 (ERAI/Land). Considerable uncertainties exist among the eight products at the Yellow River outlet, which might be because the LSMs ignore frequent human activities. Based on the above statistics, performances of the eight runoff products are ranked in descending order as follows: VIC-CN05.1, ERAI/Land, JRA55, CLM-CFSR, CLM-ERAI, MERRA-2, CLM-MERRA, and CLM-NCEP, which provides a reference for flood/hydrological drought warning and hydroclimatic research in the future.