A temporal variation and spatial distribution of the snow-covered area (SCA) over the Tibetan Plateau (TP) are analyzed using moderate-resolution imaging spectrometer (MODIS)/Terra 8-day snow cover products (MOD10A2) from 2001 to 2013 and the SCA is compared with in situ snow cover days (SCD) from the meteorological network in the TP. Results show that at monthly levels the minimum SCA occurs in July, followed by August, and the SCA increases rapidly from September, reaching the maximum in March; on average, 2002, 2005, and 2008 are snowy years, whereas 2001, 2003, 2007, and 2010 are less-snow years. Apart from strong seasonal variations, the general trend of interannual snow cover variations from 2001 to 2013 is not obvious, remaining at a relatively stable status. The snow cover over the TP is characterized by uneven geographic distribution. In general, snow is abundant with a long duration in the high mountains while it is less abundant and with a short duration in the vast interior of the TP. The interannual variations of snow cover over the TP from ground-based meteorological stations using SCD are very consistent with MODIS SCA, with a correlation coefficient of 0.80 (P<0.01), indicating that MOD10A2 data have high accuracy to capture and monitor spatiotemporal variations of snow cover over the TP.
In order to reduce the human labor in snow cover monitoring, recent study has been done on modification of the multi-spectral thresholds method which was developed in NSMC in 1996. Based on the analysis of the spectral characteristics of snow, cloud and other types of earth surface with multi-spectral data, an automated processing system with the new thresholds method to distinguish snow and cloud have been set up in NSMC. The devised technique is applied to multi-spectral data from FY-1C and NOAA-16 for mapping snow cover over China during winter season. To assess performance of the modification, the automatically produced snow data sets have been compared with the NOAA operational snow products and validated against in situ land surface observations in China. There is a good consistency between our results, NOAA snow data and ground measurements. The correlation coefficient between the snow cover produced by NSMC and NOAA is about 80%. The results of the comparison show us that the 1.6µm band data is very useful for snow and cloud distinguishing. The new method can reduce the human labor in snow cover monitoring and produce accurate snow cover images in China using FY-1C and NOAA-16 satellite data.