Snow depth and snow cover over the Tibetan Plateau observed from space in against ERA5: matters of scale

Abstract

There has been a statement that the satellite and reanalyses significantly overestimate snow depths (SDs) as compared with in-situ observations over the Tibetan Plateau (TP). The inconsistency may be partly due to representations related to different spatial resolutions. To further clarify matters of scale, this work estimates and compares the fractional snow cover (FSC) from MODIS (500 m), SDs from Sentinel-1 C-band SAR (1 km), ERA5-Land (9 km) and ERA5 (31 km), together with ground observations from GHCN-D and at 6 newly-established stations in the Namco watershed. SD and FSC from MERRA-2 (50 km) are also discussed in comparison with ERA5s. Results indicate that SD from fine resolution reanalysis has better consistency to in-situ observations over the TP. SD of ERA5-Land matches in-situ measurements better than ERA5 and MERRA-2. Overestimates of SDs in reanalyses are likely for shallow snowpack. However, underestimates are found for deep snow particularly late in the snow season. Improvements are displayed in ERA5-Land, while increased SD is attributed by altitude in fine resolution. Overall SDs of ERA5 and ERA5-Land have similar spatial distributions and annual cycle patterns over the TP, consistent with satellite-based datasets. A notable defect of ERA5s is related to delayed ablation of deep snowpack during spring and early summer. It causes cold temperature biases at surface that may impact on the land-atmosphere interaction. Snow analysis that combines the information of FSC into SD has been demonstrated in MERRA-2 data. Snow analysis has the potential to improve SDs of ERA5s over the TP.