提出了一种针对被动微波AMSR-E数据的土壤水分反演算法。用改进的积分方程模型(AIEM)模拟表明:在给定的粗糙度条件下,土壤水分和发射率之间有很好的线性关系;18.7GHz与10.7GHz的垂直极化(V)微波指数与土壤水分有很好的关系,而且部分消除了土壤粗糙度的影响,相关系数的平方(R—square)是0.98。因此,建立标准化的微波指数与土壤水分之间的关系是可行的。算法敏感性分析表明,当有降雨时此算法比较敏感。相对于全国农业遥感地面监测东北网点县实测数据,算法平均误差大约是21.5%。此算法低估了土壤水分,用实测数据对反演结果做进一步修正后的误差为7.4%。用AMSR.E数据对2009年2月1日中国主要陆地表面进行了实际反演分析,结果表明反演结果符合实际土壤水分的分布情况,表明算法可行。
This paper proposes an algorithm for retrieving soil moisture by brightness temperature from the data of the advanced microwave scanning radiometer for earth observation systems (AMSR-E). According to the simulation analysis of the AMSR-E passive microwave data using the advanced integrated equation model(AIEM), there is a good linear relationship between emissivity and soil moisture under a given roughness, and the normalized difference between the emissivities of 18.7GHz and 10.7GHz can partly eliminate the influence of roughness, with the R-Square being about 0.98. Thus, establishing the normalized relation between microwave index and soil moisture is feasible. The proposed method avoids the parameter of land surface temperature which is the key parameter for the computation of emissivity. The sensitivity analysis for atmosphere, the main factor for the method, indicates that the method is very good for clear days but is not very well for mining days. The valuation of the algorithm by the ground measurement data shows that the retrieval error of soil moisture is about 21.5% relative to the experimental data. After making a regression revision, the retrieval error of soil moisture is below 7.4%. Finally, the retrieval of the soil moisture from the AMSR-E data for the land of China shows that the distribution trend of soil moisture is consistent with the real world.