Spatial prediction of soil heavy metal cadmium content based on geographically optimal similarity
Based on the law of geographic similarity,40 soil samples collected in the field were used to construct a geographically optimal similarity(GOS)model by combining environmental auxiliary variables to predict the heavy metal cadmium content and its spatial distribution in the study area,and the prediction results were compared and analyzed with those of partial least squares regression(PLSR),random forest(RF)and universal kriging(UK)models.The results show that the mean cadmium content of soil samples in the study area(0.432 mg/kg)is greater than the background value,close to the moderate pollution level(pollution index of 2.18),and the regional soil ecology is under some threat.The GOS prediction results have a coefficient of determination of 0.668,and the root-mean-square error and the mean absolute error are 0.096 and 0.080,which are the best among the four prediction models.The spatial prediction results of the GOS show that the regional content of the heavy metal Cd decreases from the northeast to southwest,and the high values are distributed along rivers and in areas with intensive human activities,reflecting that human activities are the main factors leading to soil heavy metal differentiation in the study area.
geographical similaritysoil propertiesheavy metal Cdspatial prediction
国家科技期刊平台
NSTL
万方数据
