Evaluation of Climate Data from ECMWF Reanalyses over Siberia and the Russian Far East

Reanalysis data provide a complete picture of the past climate by re-running previous forecasts using modern methods and assimilating with observations. In this work three reanalysis datasets from the European Centre for Medium-Range Weather Forecasts (ECMWF), ERA-Interim, ERA5 and ERA5-Land, were validated against data from 29 meteorological stations across Siberia and the Russian Far East. ERA5 offers improved spatial and temporal resolutions compared to ERA-Interim, as well as starting twenty years earlier in 1959 and being continually updated to the present day with little delay. ERA5-Land replays the land component of ERA5 over an improved 9km spatial resolution and the dataset begins in 1950. The validation was conducted at daily, monthly, seasonal and annual timescales for seven climate variables first to 1979, then additionally to 1959 for ERA5 and ERA5-Land. We found that the snow depth values in ERA5 are only assimilated with meteorological station data from 1992 onwards, leading to significant and inhomogeneous overestimations before this time. As ERA5-Land uses the ERA5 values as its boundary conditions, the snow depth values in this dataset are further from the observations. The mean sea level pressure in the reanalyses is closest to those from the meteorological stations. The daily minimum 2-metre air temperature is noticeably weak during the summer months, however on broader timescales the reanalyses perform very well for the minimum, average and maximum temperatures. Total precipitation and wind speed consistently have the lowest correlations at all temporal and spatial resolutions. Despite the increased spatial resolution, we found no improvement to using ERA5-Land over ERA5, however we would recommend using ERA5 over ERA-Interim due to the larger amount of data available.