基于AMSR-E数据的青海省雪深遥感监测模型及其精度评价

基于2002-2008年10至翌年3月6个积雪季的AMSR-E被动微波亮温数据,以青海省为研究区,结合该区43个气象台站实测雪深等气象数据,系统分析雪深模型建立的影响因子,提出研究区样本筛选的方法.对18GHz和36 GHz波段的水平、垂直极化方式的亮温差和实测雪深值进行回归分析,建立了青海省基于AMSR-E亮温数据的...

Monitoring Model and Accuracy Evaluation of Snow Depth in Qinghai Province Based on AMSR-E Data

The factors affecting snow depth were systematically analyzed based on a developed snow depth model and the data of AMSR-E passive microwave brightness temperature and snow depth measured by 43 meteorological stations in Qinghai Province during 6 snow seasons from October to next March from 2002 to 2008, and then the approaches of screening samples in the study area were put forward. The regression analyses of horizontal and vertical polarization brightness temperature difference data of 18 GHz and 36 GHz bands and snow depth values were carried out. The equation was , and the correlation coefficient was 0.65. Analyzed results of the accuracy reveal that the absolute error of 80% validated samples between the simulated and measured snow depth values was 2 cm or less, and that of 92% was within 3 cm. When the snow depth was ≤ 8 cm, the accuracy of the simulated results was high, especially the average error was only 0.83 cm and the RMSE value was 1.77 cm when the snow depth varied in a range of 6-8 cm. When the snow depth was thicker than 8 cm, the error of the simulated snow cover depth was high. Basically, the inverted results could reflect the change trend of snow cover depth in Qinghai Province. However, the accuracy of the simulated snow cover depth was low in some regions.