雅砻江地区无云积雪覆盖产品的制作

A Study on Cloud-free Snow Cover Product in Yalong River Area

MODIS积雪产品在晴空下积雪识别精度很高,但其受云污染导致数据缺失严重。IMS和SWE数据虽为无云产品,但受分辨率的限制积雪监测精度有待提高。以青藏高原东部雅砻江流域及周边地区为例,通过合成MODIS每日积雪覆盖产品、邻近日分析法以及改进的SNOWL判别法对云像素进行重分类,然后用IMS或者SWE无云积雪数据对中间生成的片雪再分类,制作了除云后的逐日无云积雪覆盖产品。再用目视解译法将从HJ-1B卫星影像中提取的积雪覆盖信息作为观测"真值",对无云积雪覆盖产品进行分类精度评估。结果表明:通过算法的改进,提高了该产品与观测数据的积雪一致率和总体分类精度,总体上解决了因云污染导致的数据缺失,IMS和SWE积雪监测精度不足的问题。

The daily MODIS snow cover （SC） product has high accuracy under clear sky,but cloud contami- nation causes serious data defect missing.Snow monitoring accuracy of IMS and SWE needs to be improved due to the lower resolution.So,the Yalong River and its surrounding area in China is taken as an example to produce a cloud-free snow cover product, cloud pixels are reclassified by combining multi-temporal MO- DIS products,the adjacent day method and improved SNOWL algorithm, then partially snow covered area is reclassified by cloud-free snow product .IMS and SWE.Then validation of this product is compared with observation ＂truth＂ extracted from HJ-1B by visual interpretation method.The result shows improved al- gorithm can get a cloud-free product with better snow consistency rate and overall classification accuracy with observations.Overall, problems of data missing caused by cloud contamination and inadequate accuracy of IMS,SWE are solved.