基于地面站点类型代表性的积雪遥感产品精度评价

Based on the Type of Ground Site Representative of Snow Remote Sensing Products Precision Evaluation

地面观测数据是进行遥感产品检验的重要数据源,依据地表类型对地面站点代表性进行了分析,把代表性较好的站点应用到AMSR-E雪深遥感产品的精度评价中。结果表明:森林地区处于明显低估状态;其他地区随着遥感产品像元内优势类型从灌木、草地到裸地的变化,高估状态呈逐渐减小趋势;森林和灌木地区遥感产品的精度比草地与裸地地区的精度要低;当森林混入裸地或灌木混入裸地、草地时,虽然遥感产品和地面观测之间的误差减小了,但是掩盖了森林地区低估的问题和灌木区高估的问题。利用地表类型一致性较好的站点对2010年1月份月平均雪深验证分析,结果表明:遥感产品在较平坦的森林、灌木、裸地以及草地地区的平均偏差分别为-5.56、4.8、1.17、1.16cm。

The ground observation data is an important data source for the validation of remote sensing products validation,based on analysis of the representativeness of the land cove type of ground station,then applied the good representative stations to assessment the accuracy of AMSR-E snow remote sensing products,the results show that：Forest area is significantly underestimated,in the other area with the change of advantage type form shrubs,forest,grassland to bare land in the pixel,the tendency of overestimate is gradually decreasing;the precision of remote sensing products in the forest and shrub area is lower than that of grassland and bare area.When mixed with the bare land in the forest or the bare land and grassland in the shrub,while the error between remote sensing products and ground observation reduced,but that covered the phenomena of underestimation in forest area and overestimation in shrubs area.Analyzing the average snow depth on January 2010with the homogeneous land cover type sites,the average deviation of remote sensing products are-5.56cm,4.8cm,1.17cm,1.16cm respectively in relatively flat forest,shrubs,bare and grass area.