利用半经验模型的土壤水分反演方法

Study on retrieval method of bare soil moisture based on semi-empirical Dubois model

针对土壤水分反演经验模型使用范围受限、而理论模型相对复杂的问题,该文利用积分方程模型及经验相关长度定标模型,分析均方根高度、地表相关长度、土壤体积含水量对雷达后向散射系数组合的影响;结合RADARSAT-2全极化C波段雷达数据和野外实测数据,构建研究区均方根高度反演模型;将该模型与Dubois模型、介电模型进行结合以反演表层土壤含水量,并对反演值进行线性校正,最终实现裸土区表层土壤水分反演。实验结果表明,HH、VV极化下土壤水分实测值与反演值绝对误差在0.06内的点数分别达到总点数的78%和68%,证明该方法具有一定的可靠性和实用性,可为地表均方根高度、土壤水分反演提供方法和借鉴。

Combined with RADARSAT-2C-Band full polarized radar data and field surveying data,a model to retrieve surface RMS height of the research area was constructed based on integral equation model and empirical correlation length calibration model,which were used to analyze the effect of each input parameter on the combination of radar backscattering coefficients.Then the model was combined with Dubois model and dielectric model to retrieve surface soil moisture in bare soil areas,and linear error correction formula was taken to adjust the final retrieval result.The experiment indicated that 68% of the points retrieved for HH polarization and 78% of those for VV polarization respectively had an absolute error less than 0.06.The result proved this method to be reliable and practical,which provided a way to retrieve surface RMS height and volumetric soil moisture.