GWR模型下农用地土壤镍空间分布预测

Prediction of Soil Nickel Spatial Distribution in Agricultural Soil Under GWR Model

传统农用地土壤分析方法耗时耗力,利用光谱及各类其他因子进行重金属浓度快速反演的方法因为其高效、快速、低成本的优点受到越来越多研究者的青睐。文章试图探索利用光谱信息与土壤镍含量信息构建地理加权回归(geographic weighted regression,GWR)反演模型,并实现对农业土壤镍空间分布的预测。首先,以栾川县石宝沟农用地中镍的实测含量为目标变量,综合利用Landsat-8波段反射率、样品采样点的空间位置及地形信息等作为变量,使用相关性分析及逐步回归的方法,选择3个变量,即采样点与厂区的最短距离、ln(band2/band3)、band3-band5,作为解释变量;其次,利用地理加权回归的方式进行建模,模型决定系数达到0.64;然后,用测试样本点进行模型验证,Acc值达到96.51%,可见所建立的模型能够较好地拟合农用地土壤中镍含量;最后,对整个研究区域内农用地进行反演,并对其空间分布进行评价。

Traditional methods for soil heavy metal content analysis in agricultural soil are time-consuming and labor-consuming.Spectral inversion method has been favored by more and more researchers because of its advantages of high efficiency,fast speed and low cost.Thus,this study attempts to use spectral information and soil nickel content information to build GWR inversion model,and finally achieve the prediction of the spatial distribution of agricultural soil nickel.Firstly,the measured content of nickel in the agricultural land of Shibaogou in Luanchuan county was taken as the dependent variable.With the method of correlation analysis and stepwise regression,three variables were selected as explanatory variables,including the shortest distance between the sampling point and the plant area,the natural log of band2 over band3,band3 minus band5.After that,these variables would be used for geographically weighted regression modeling.The result showed that the coefficient of determination reached 0.64 and the accuracy value of the test set samples reached 96.51%,indicating that the established model performed excellent in fitting the nickel content of agricultural land soil.Eventually,the model was used to predict and evaluate the spatial distribution of nickel in agricultural soil in the whole study area.