%0 Dataset
%T Daily Cloud-Free 500 m Snow Cover Fraction Product over High Mountain Asia from MODIS (2000–present)
%J National Cryosphere Desert Data Center
%I National Cryosphere Desert Data Center(www.ncdc.ac.cn)
%U https://www.ncdc.ac.cn/portal/metadata/d9141f2b-16dc-48d4-b054-23b0572b3909
%W NCDC
%R 10.12072/ncdc.db7815.2026
%A Hao Xiaohua
%A Zhao Qin
%A GAO Weiqinag
%A Zhao Zisheng
%K snow cover;snow cover fraction
%X High Mountain Asia (HMA) is one of the areas with the most widespread distribution and most severe changes in snow cover in the Northern Hemisphere. Fractional Snow Cover (FSC) can quantitatively describe the snow cover on a sub-pixel scale. The degree is more suitable to reflect the distribution of snow cover in complex mountainous areas than binary snow cover, and is an important input parameter for snowmelt runoff simulation and climate change prediction in mountainous areas. Aiming at the characteristics of high Asian snow cover, this dataset uses the MODIS reflectance product MOD09GA as the data source, and uses multiple adaptive regression splines selected by classification characteristics.(Multivariate Adaptive Regression Splines (MARS) model LC-MARS develops the MODIS FSC inversion algorithm and prepares FSC products for Asian alpine regions. It then introduces a spatio-temporal multi-step fusion algorithm that combines spatio-temporal filtering, spatio-temporal inverse distance weight interpolation and multi-source data fusion methods.(spatiotemporal multi-step fusion (STMF) achieves complete cloud removal of the product and obtains a high Asian MODIS daily cloud-free 500-meter snow cover area ratio data set. The dataset is stored in the TIF file format. Each TIF file is single-band raster data with a cell range of 0 - 100, which represents the proportion of the snow area (0= no snow, 100= complete snow), and 230= water body. The total Accuracy and Recall of FSC retrieved by the LC-MARS model are 93.4% and 97.1% respectively, the overall RMSE is 0.148, and the MAE is 0.093. The overall accuracy is high. This dataset will be updated in real time based on satellite remote sensing data and algorithm updates, and will be fully open and shared.