Using reanalyses for studying Eurasian snow cover and its relationship with circulation variability

TitleUsing reanalyses for studying Eurasian snow cover and its relationship with circulation variability
Publication TypeProceedings
Date Published06/2013
EventECMWF-WWRP/THORPEX workshop on polar prediction
AuthorsBrun, E, Peings, Y, Vionnet, V, Boone, A, Decharme, B, Douville, H, Karbou, F, Morin, S
Event Series/CollectionECMWF-WWRP/THORPEX Workshop
Conference LocationECMWF
Keywordspolar prediction, THORPEX, workshop, WWRP

In order to investigate the relationships between the climate variability and the state of snow cover in Northern Eurasia, we used two different sources of simulations to reconstruct the historical snowpack conditions. The first one is based on a stand-alone simulation with the detailed snowpack model Crocus. Forcing data were extracted from ERA-interim reanalyses. Without assimilating any local snow-related observation, the simulated snow depth, snow water equivalent (SWE) and density have been successfully compared with local observations from more than 1000 monitoring sites. The simulations show a very small bias as well as high correlations. In terms of SWE, the overall performance of the simulations is very similar to the accuracy of the satellite gridded snow product GlobSnow which, in contrast, assimilates local snow depth observations.

The second approach directly uses the snow cover from the meteorological reanalysis 20CR which covers the whole past century. A detailed comparison with more than 1000000 daily in-situ snow depth observations shows that the simulated snow cover in October and November exhibits a very high and steady performance throughout the XXth. Century. 20CR performance for detecting the presence of snow on the ground is even higher than the performance of the NOAA satellite-based snow cover product. This snow reconstruction was used to investigate the relationship between the onset of snow cover over Northern Eurasia in Fall and the Arctic Oscillation (AO) and North Atlantic Oscillation (NAO) conditions which occurred the following Winter. It shows that the high correlation which is observed for the last decades has not been steady over the whole Century.

PDF icon Download (955.64 KB)