Abstract: With the development of remote sensing (RS) technology, the RS data have become as the indispensable materials in dynamic monitoring of snow cover, and they play an important role in monitoring snow coverage and analyzing snow disasters. In this paper, the snow coverage retrieved from Landsat TM was taken as the ground truth, the fractional snow coverage (FRA) retrieved from MOD10A1 was validated using Landsat TM image, and an improved segmentation model was proposed for retrieving the FRA by using the MODIS ground reflectivity products (MOD09GA) in some parts of the Qinghai-Tibet Plateau. The results are as follows: ① The average snow coverage retrieved from MOD10A1FRA was 29.16% in our study, but it was 22.45% in TM, and the correlation coefficient of FRA between those two datasets was 0.74; ② The FRA model improved with the MOD09GA product had much better accuracy in SCA acquisition, the average retrieved snow coverage was 21.70%, and the correlation coefficient was 0.85;③ Compared with the “ground truth”, the performance of both the standard error of the improved fractional model and the mean absolute error was better than the MOD10A1/FRA product, which could be used to improve the accuracy of monitoring snow coverage in the study area.  

Key words: font-family: 宋体, mso-bidi-font-family: 宋体, mso-font-kerning: 1.0pt, mso-ansi-language: EN-US, mso-fareast-language: ZH-CN, MODIS')">mso-bidi-language: AR-SA">MODIS, TM image, fractional snow coverage, standard model, validation, Qinghai-Tibet Plateau