ASAR GM后散时间序列数据估算黑河上游地表土壤水分

利用2008~2011年的ENVISAT/ASAR全球监测模式(GM)数据,采用时间序列变化检测算法,估算地表相对土壤水分,并利用Van Genuchten方法将相对土壤水分转换为绝对土壤水分,最终获得研究区内的土壤体积含水量。利用阿柔冻融观测站2008~2011年10cm土壤水分数据验证,均方根误差为0.11cm3/cm3;利用八宝河流域无线传感器网络的36个WATERNET节点2013~2014年的4cm体积含水量月均值进行空间分布的间接比较检验,估算月均值的均方根误差在0.03~0.11cm3/cm3的节点有19个,在0.11~0.16cm3/cm3的节点有15个,大于0.16cm3/cm3的有2个。另外考虑遥感数据和算法(暂未考虑土壤容重、土壤残余含水量的不确定性)对估算结果的影响,体积含水量最大估计误差范围为0.03~0.12cm3/cm3,研究区内91.77%的像元小于0.06cm3/cm3。

Estimation of High|resolution Soil Moisture by Using ENVISAT/ASAR Global Mode Backscattering in the Upper Reaches of Heihe River Basin

The change detection method is adopted to estimate the relative soil moisture by using ENVISAT/ASAR global mode data with 1 km resolution in the upper reaches of Heihe River Basin.Then the relative soil moisture is convert to absolute soil moisture by the Van Genuchten formula based on the soil bulk.The comparison analysis and validation during the period from 2008 to 2011 by using the 10 cm observations at A’rou freeze/thaw observation station,show that the root mean square error (RMSE) of the estimated volumetric soil moisture from ASAR is 0.11 cm³/cm³.The indirect validation in the spatial domain by using the mean values of observations at 36 WATERNET nodes,shows that there are 19 nodes which RMSE range from 0.03 cm³/cm³ to 0.11 cm³/cm³,there are 15 nodes which RMSE range from0.11 cm³/cm³ to 0.16 cm³/cm³,and there are 2 nodes which RMSE range from 0.16 cm³/cm³ to 0.19 cm³/cm³.The theoretical maximum estimation errors of volumetric soil moisture range from 0.03 cm³/cm³ to 0.12 cm³/cm³,and that 91.77% of grids is below 0.06 cm³/cm³,the result also shows that the perform of this algorithm at flat topography is better than the mountains.