基于遥感的土壤电阻率估算研究

以我国长江中下游的南京及其周边15个市（县）为例,采用野外测定、室内分析与遥感反演相结合的方法,开展了土壤电阻率估算研究.选用影响土壤电阻率的土壤水分、土壤温度、土壤可溶盐总量与土壤阳离子交换量等4个主要因子,遥感反演土壤水分和温度空间分布,以获取估算土壤电阻率所需要的主要参数;采用偏最小二乘二次多项式（PLSQM）模型对不同地表覆盖类型下的土壤电阻率进行估算,PLSQM估算模型的估算值与实测值的相关系数达到0.85,平均相对误差（MRE）为19.02%,均方根误差（RMSE）为7.79.结果表明,草地、农田、林地3种不同地表覆盖类型下土壤电阻率有明显差异,PLSQM模型实现了较高估算精度,具有较好的应用潜力.

Studies on estimation of soil resistivity based on remote sensing

Soil resistivity of Nanjing and its surrounding 15 cities/counties in the middle and lower reaches of Yan- gtze River is estimated, using an integrated method of field survey, laboratory chemical analysis and remote sensing retrieval. Four main influencing factors of soil resistivity, including soil moisture, soil temperature, soil soluble salt content and cation exchange capacity （CEC） ,were chosen to be the main factors of the estimation model. Spatial distribution of soil moisture and soil temperature were retrieved from MODIS data. A partial least squares quadratic model （PLSQM） is established to estimate soil resistivity under different land cover types. The correlation coeffi- cient between estimated and observed soil resistivity values is 0.85,with mean relative error （MRE） being 19.02% and root mean square error （RMSE） being 7.79. Soil resistivity, estimated by PLSQM method, differs obviously un- der land cover of grass, crop, or forest. The proposed PLSQM method can estimate soil resistivity with high accuracy, thus has a good application potential.