黑龙江流域MODIS和AMSR-E雪盖产品精度分析

摘要

以黑龙江流域气象站数据作为"真值",检验了2002-2007年黑龙江流域MODIS积雪产品和AMSR-E数据识别积雪的精度,分析了云以及海拔高度对积雪识别的影响.结果表明,MOD10A2、MOD10C2、AMSR-E识别积雪的平均精度分别为75.4%、88.7%和88.9%,云仍然是影响积雪识别精度的主要原因,MOD10C2的积雪识别结果接近于气象站的真实值,AMSR-E高估积雪覆盖面积.此外通过软分类方法得出的MOD10A2和MOD10C2的积雪覆盖面积要低于硬分类得出的积雪覆盖面积,软分类法得到的结果更接近真实值.%Using meteorological site data in Heilongjiang drainage basin as the ‘ truth valves' , we evaluated the snow identification accuracies of MODIS snow products and AMSR-E data from 2002 ～ 2007 and analyzed the influence of cloud and altitude on the snow identification accuracy. The results showed that MOD10A2,MOD10C2,and AMSR-E snow identification accuracies are 75.4％ , 88.7％ , and 88.9％ , respectively. The existence of cloud is the main factor affecting snow identification accuracy. MOD1OC2 snow identification results are close to the true values of meteorological stations and AMSR-E tends to overestimate the snow cover area. Besides,the results indicate that snow cover areas derived from the soft classification of MOD10A2 and MOD10C2 are smaller less than that derived from the hard classification,and the results obtained from the soft classification are closer to the true values.