干旱区生态环境知识资源中心

For the estimation of snow water resources in China, the passive microwave remote sensing data (SSM/I) have been used to retrieve the snow depth information. Due to the overestimation of snow depth in West China (especially, the Tibet Plateau regions), the global retrieval algorithm of snow depth from SSM/I data is modified ground observations in West China region. Both of the global and modified algorithms adopt the negative gradient of brightness temperature at 19 and 37GHz frequencies at horizontal polarization that were generated the snow layer's scattering feature. However, other surface features such as the precipitation, cold desert, frozen ground, may have similar scattering signature to snow cover. Grody's decision tree containing various filters, is used to separate the scattering signature of snow cover from other scattering signature. For the snow cover pixels after the filters technique, we retrieved the snow depth by global and modified algorithms in East and West China, respectively. The retrieval results have been assessed by the MODIS snow products (MODISC1 snow products) with two methods of accuracy assessment, i.e. the overall accuracy and Kappa analysis based on the error matrix. The two snow products agree very well based on the overall accuracy and Kappa coefficients.' Finally, the operational scheme has been performed to obtain the snow depth datasets in China in a complete hydrological year.