Which global reanalysis dataset has better representativeness in snow cover on the Tibetan Plateau?

Abstract. The extensive snow cover across the Tibetan Plateau (TP) has a major influence on the climate and water supply for over 1 billion downstream inhabitants. However, an adequate evaluation of variability in the snow cover fraction (SCF) over the TP simulated by multiple reanalysis datasets has yet to be undertaken. In this study, we used the Snow Property Inversion from Remote Sensing (SPIReS) SCF dataset for the water years (WYs) 2001–2017 to evaluate the capabilities of eight reanalysis datasets (HMASR, MERRA2, ERA5, ERA5L, JRA55, CFSR, CRAL, and GLDAS) in simulating the spatial and temporal variability in SCF in the TP. CFSR, GLDAS, CRAL, and HMASR are good in simulating the spatial pattern of climatological SCF, with lower bias and higher correlation and Taylor skill score (SS). By contrast, ERA5L, JRA55, and ERA5 have a relatively good performance in terms of SCF annual trends among eight reanalysis datasets. The biases in SCF simulations across reanalysis datasets are influenced by a combination of meteorological forcings, including snowfall and temperature, as well as by the SCF parameterization methods. However, the primary influencing factors vary among the reanalysis datasets. Additionally, averaging multiple reanalysis datasets can enhance the spatiotemporal accuracy of SCF simulations, but this enhancement effect does not consistently increase with the number of reanalysis datasets used.