Wang X.,Sun Yat Sen University |
Zheng H.,Sun Yat Sen University |
Chen Y.,Xinjiang Institute of Ecology and Geography |
Liu H.,Heilongjiang Climate Data Center |
And 4 more authors.
Journal of Applied Remote Sensing | Year: 2014
Cloud contamination is one of the major barriers for wider applications of MODIS snow cover products. This study presents a cloud-removal approach, through multiday backward replacements based on Terra and Aqua daily MODIS snow cover products (MOD10A1 and MYD10A1), to generate a series of daily cloud-free snow cover products for advanced applications (MODMYD-MC). The products are evaluated using in situ snow depth data measured during 2000 to 2010 at 53 weather stations in the Heilongjiang Province, northeast China. The results show that the annual mean cloud covers of MOD10A1, MYD10A1, MODMYD-DC (the daily combination of MOD10A1 and MYD10A1), and MODMYD-MC are 50%, 54%, 35%, and 0%, mean snow covers are 6%, 6%, 10%, and 19%, and their mean agreements of snow cover mapping are 42%, 40%, 51%, and 91%, respectively. The snow-covered days (SCDs) derived from MODMYD-MC are also in good agreement (91%) with those obtained from in situ observations. The MODMYD-MC snow cover images are then used to investigate the detailed variation of snow cover in the XiaoXing 'AnLing watershed. The snow-covered area in the watershed has an increasing trend in the recent decade, with the minimum present in the 2002 (hydrologic year) and the maximum present in 2010. The plains with lower elevation show shorter SCD but larger interannual variations than in the mountainous areas. This study indicates that MODMYD-MC can be applied to monitor the spatiotemporal variations of snow cover in northeast China and elsewhere in the world. © The Authors. Published by SPIE.