Bayesian estimates of snow cover area in Eurasia in the 21st century based on the results of calculations with the CMIP6 ensemble of climate models

Based on the results of calculations with the ensemble of global climate models CMIP6, quantitative estimates of changes in the area of snow cover in Eurasia in the 21st century were obtained under scenarios SSP2-4.5 and SSP5-8.5 of anthropogenic impacts using the Bayesian averaging. The contribution (weight) of the models to the overall ensemble estimates was determined by accuracy of reproduction of the long-term average, trend, and interannual variability of the snow cover area in Eurasia by satellite data. The largest inter-model variations in estimates, the most significant of which were calculated for the summer and autumn months, are associated with the description of the trend component and inter-annual variability of the snow cover area of Eurasia, as well as with equally weighted averaging. It is shown that when using Bayesian weights, the uncertainty of snow cover area estimates can be halved compared to the ensemble average with equal model weights. The obtained ensemble estimates of the snow cover area using combined Bayesian weights exceed the corresponding estimates for equally weighted averaging.