基于Radarsat-2 SAR数据反演定西裸露地表土壤水分

摘要

利用Radarsat-2 SAR数据和定西地区野外土钻法及WET仪器观测的土壤水分数据，分析了同极化后向散射系数与不同土层深度土壤水分之间的关系，采用交叉极化（VV/VH）组合模型反演土壤水分并进行对比验证。结果表明：水平、垂直同极化后向散射系数均与10～20 cm土壤含水量相关性最好，相关系数R均为0．74；受地表粗糙度和土壤质地等影响，同极化后向散射系数与0～10 cm土壤水分相关性均较低。交叉极化组合模型的反演值与10～20 cm实测土壤水分相关性较高，R值达0．75，而与0～10 cm和20～30 cm实测值的相关性较低（R值分别为0．47和0．52），但均通过α＝0．05的显著性检验；WET仪器实测0～6 cm土壤水分经校正后与反演值的相关系数为0．46（通过α＝0．01的显著性检验），校正后的结果有效提高了WET仪器测量精度。交叉极化组合模型可用于裸露地表土壤水分的反演，更适用于提取10～20 cm土壤含水量信息。%In this study,the relationships between the homology polarization backscattering coefficient and the soil moisture at different depths were analyzed based on the Radarsat-2 SAR data and the soil moisture observational data by using soil-drilling method and the WET sensor in Dingxi area,and the soil moisture from SAR was retrieved by using the cross polarization (VV/VH)combination model and validated with the measurement value.The results displayed that the horizontal and vertical polarization backscattering coef-ficients had the best correlation with soil water content at 10-20 cm depth,and their correlation coefficients R were all 0.74.Howev-er,they were not significantly correlated to soil water content at 0-10 cm depth due to the influence of surface roughness and soil tex-ture.The soil moisture from SAR simulated by the cross polarization combination model was significantly correlated to the measured value at 10-20 cm depth,and the correlation coefficient R was 0.75,while it was lower correlated with the measured soil moisture at 0-10 cm and 20-30 cm depths,and the correlation coefficients were 0.47 and 0.52,respectively,which only passed 0.05 signifi-cance test.In additional,the correlation coefficient between corrected soil moisture measured by WET sensor at 0-6 cm depth and the simulation value was 0.46 (passed 0.01 significance test),and the calibration result improved effectively the measuring accuracy by WET sensor.So the cross polarization combination model could be used to retrieve the bare soil moisture,especially for 10 -20 cm depth soil moisture.