Long term spatio-temporal analyses of snow cover in Central Asia using ERA-Interim and MODIS products

Abstract

In this study, an approach for generating a long term series of snow-cover products from 1979 to 2015 was proposed by integrating the data of ERA-Interim snow-depth and 8-day cloud-free MODIS snow-cover derived by removing cloud from MOD10A2/MYD10A2 product. On the basis of the spatio-temporal analyses and evaluation of snow-cover duration (SCD) during the hydrological year from 1979/1980 to 2014/2015 over Central Asia, the average start and melt date of snow-cover (SCS and SCM, respectively) were estimated using the long term snow-cover product. The results suggested that the snow-cover product derived by this approach is fairly satisfactory with the mean bias error (MBE) of -.55%5.03%. The SCD, SCS and SCM all presented an apparent north-south towards gradient as long as mountainous regions and waterbodies were avoided. The mean SCD over the high-latitude and high-mountainous regions were all beyond 122 days, however, it gradually became shorter with a significant level of a < 0.05 or even a < 0.001 from 1979/1980 to 2014/2015. In contrast, the SCD over low-latitude and low-altitude regions, like Turkmenistan and Uzbekistan, were evidently shorter than the former, but it became significantly longer with a significant level of a < 0.05 or even a < 0.001 in the southwestern, northern Turkmenistan and most of Uzbekistan in the same duration. Notably, most of Turkmenistan and Uzbekistan, where snow-cover usually appeared late and melting out early, even always stay snow-free throughout the year.