风云三号微波成像仪积雪参数反演算法初步研究

A Study on Snow Depth Estimating and Snow Water Equivalent Algorithm for FY-3 MWRI

选择新疆地区作为实验区,为风云三号(FY-3)微波成像仪(MWRI)发展中国区域的积雪参数半经验反演算法。使用2003年4个月的新疆地区的台站观测资料和AMSR-E 18.7 GHz,36.5GHz和89 GHz水平和垂直极化亮温作为FY-3 MWRI的模拟数据,在Chang建立的半经验模型的基础上,采用多元线性回归分析,建立一个新算法。用已有方法去除水体、降雨、湿雪、冻土的像元后,用新算法反演了新疆地区的2004年1月的积雪参数,并分别与AMSR-E雪水当量产品和台站观测值进行比较,结果表明新算法在新疆地区优于AMSR-E的反演算法。

FY-3 is the second generation polar orbit meteorological satellite serial of China. There is a radiometer- Microwave Radiometer Imager （MWRI） board on first satellite of FY-3 which will be launched in 2007 in first time. Xinjiang province as a test site was selected to develop the snow parameters algorithm for FY-3 MWRI in china area. Base on Chang＇s semi-empirical model, AMSR-E brightness temperatures at 18.7 GHz,37 GHz and 89 GHz, four-month snow depth and SWE which are observed by meteorology station at Xinjiang in 2003 were used to establish regional algorithm. A monthly records of daily snow depth and SWE in JAN 2004 and AMSR-E SWE products were used to test the new algorithm. The surface water body, wet snow, precipitation and other anomalous scattering signals are screened using established methods. While compare estimated SWE by using new algorithm and SWE from AMSR-E with meteorological stations records respectively, the RMSE are 17.9 mm and 26.4 mm. The result show the new algorithm is better performance than algorithm for AMSR-E over Xinjiang province.