C波段SAR山区积雪面积提取研究

Snow Cover Area Retrieval Using C-band SAR in Mountain Areas

合成孔径雷达(SAR)不仅具有穿云透雾、全天候观测地表的能力,而且可穿透地表覆盖一定深度获取地表覆盖物内部特征信息。利用2011年10景ENVISAT-ASAR可变极化模式精细图像(ASA_APP_1P)数据,分析比较了黑河上游祁连山冰沟流域不同时段积雪SAR后向散射特性,应用同期的MODIS积雪面积产品确定研究区积雪的累积和消融背景信息。研究表明:由于融雪期积雪含水量上升,SAR图像后向散射系数相比干雪或无雪图像明显降低,经过分析认为广泛应用的-3dB阈值会明显低估湿雪覆盖范围,-2dB阈值更适合该地区湿雪面积参数提取。山区积雪融化过程中低海拔区域积雪融化而高海拔山区积雪仍可能为干雪,在提取湿雪像元的基础上,根据Sigmoid函数阈值获取的像元湿雪百分比及DEM信息来提取干雪像元,最终获取整个流域积雪面积信息。通过与Landsat ETM+图像积雪面积分类结果进行比较,总体精度达到78%。积雪累积和消融背景信息的分析表明:误差主要源于流域东北部与西北部低海拔区域积雪快速消融。

Spaceborne Synthetic Aperture Radar （SAR） sensors not only has the capabilities of penetrating clouds and providing day and night remote sensing data, but also has the capability of penetrating snow cover to retrieve subsurface information. In this paper, 10 scenes of ENVSAT-ASAR images were acquired with multiple incidence angles, which were used to characterize the seasonal variations of snow covered landscapes in the upper reaches of the Heihe River Basin from 10 April, 2011 to 15 September, 2011. Be- sides, 12 MODIS images were used to provide background information,as well as support the monitoring of spatial and temporal variability of snow cover. During snow melt period,the presence of liquid water within the snowpack,which increased the snow dielectric constant,resulting in an appreciable decrease of back- scatter coefficients in comparison with dry snow and snow-free soil surface. According to analysis of snow cover spatial and temporal variability of study area,the mostly common used --3 dB threshold clearly un- derestimates snow cover in our study area and the --2 dB threshold was applied to retrieve the wet snow cover. On the basis of the wet snow cover map,in association with DEM data and wet snow cover percent image which derived from Sigmoid function, the dry snow cover areas were mapped. The final snow map- ping results were compared with the classification image of ETM＋ ,and the total accuracy reached 78%. The spatial and temporal variability of snow cover indicated that the main errors were caused by snow rap- idly melting process in northeast and northwest low altitude area from 10 April,2011 to 13 April,2011.