中文摘要：

准确监测牧区积雪覆盖范围,对有效防灾减灾和牧区畜牧业持续发展具有特别重要的意义.利用积雪产品和气象台站的观测资料,对比分析了北疆地区2002年11月1日-2005年3月31日三个积雪季的AMSR-E 每日雪水当量产品及其与MOD10A1每日积雪分类产品合成的图像MODAE1的积雪识别精度.结果表明,1）AMSR-E每日雪水当量产品的积雪识别率为66.59%,总精度为69.49%;2）利用用户自定义的合成算法计算的合成图像MODAE1,结合了AMSR-E 雪水当量产品不受天气影响和MOD10A1每日积雪产品较高空间分辨率的优点,积雪识别率达76.43%;3）雪深和土地覆盖对合成图像MODAE1的积雪识别率具有重要的影响.在雪深为1-40cm时,合成图像的积雪识别精度随雪深的增加而增大;在雪深为31-40cm 时,积雪识别率可达90.19%;在雪深大于40cm 时,积雪识别率开始下降.在牧区合成图像的积雪识别率可达77.6%,而在开阔的灌丛区积雪识别率略有下降,为72.7%.

英文摘要：

Accurate monitoring of snow-covered areas plays a significant role in dynamic studies and prevention of snow-caused disasters and of sustainable development in pastoral areas. Using snow cover products and climatic station data, the snow classification accuracy was analyzed for AMSR-E daily SWE （snow water equivalent） products and composite images using AMSR-E SWE and MOD10A1 daily snow cover products in the three snow seasons of November 1 to March 31 from 2002 to 2005 in the north of Xinjiang Province. Results suggested that the snow classification accuracy for AMSR-E daily SWE products is 66. 59% and the overall accuracy is 69. 49%. The composite images （denoted MODAE1）, calculated by the user-defined composite algorithm, combined the advantages of AMSR-E SWE products that are not affected by weather and of MOD10A1 daily snow cover products with higher spatial resolution so that the snow classification accuracy reached 76.43%. The snow depth and land cover have an important effect on the accuracy of snow classification for the MODAE1 composite images. The snow classification accuracy for the composite images increased when the snow depth （SD） was between 1 and 40 cm, reached 90. 19% when SD was 31-40 cm; hut started to decrease when SD was more than 40 cm. The snow classification accuracy in the pastoral areas reached 77.6%, but declined to 72.7% in open shrub land areas.