Validation of MERRA and ERA-Interim Snow Depth Quantities in Eurasia

Abstract:

Snow cover plays an important role in the energy balance of the ground surface, especially in frozen ground areas. It acts as an insulator, slowing the transfer of heat between the air and the land. Therefore, the ability to quantify snow depth is important for investigating its impact on the degradation of frozen lands due to climate change and other factors. Eurasia contains a significant amount of frozen ground, but point measurements of snow depth are sparsely distributed across the continent, making a comprehensive study of its effect difficult. Reanalysis products provide data between these point measurements with a continuous grid of best-guess quantities of snow depth. However, the quality of reanalysis snow depth products is uncertain. This project quantifies the ability of two reanalysis products, MERRA and ERA-Interim, to capture snow depth variability at 1,259 stations across Eurasia on a monthly scale between 1979 and 2009. Various statistics, including root mean square error, mean absolute error, correlations, and trends are computed to quantify each reanalysis product’s performance. Statistics show that reanalysis performs reasonably well, as noted by median correlation coefficients at 0.6 or higher for most months. Correlations are lower in the transitional months of May, September, and October. The range in monthly snow depth trends is larger in the observational data, but reanalysis trends match fairly well throughout the year outside of February and March, when observed trends are positive, while reanalysis indicates negative trends. Relative errors between the observational and reanalysis datasets range between ~15-30% during months with significant snowfall, while the transitional month of October is again an outlier. Results suggest that reanalysis snow depth can be a viable substitute for observations in the data sparse regions of high-latitude Eurasia.