Evaluating Multiple Heavy Metal Pollutants in Soil by Artificial Neural Network: A Case Study in Baotou, China.

Abstract: Assessment of heavy metal risk and spatial distribution for special area are fundamental to the study of soil environmental quality. The paper provides a case of artificial neural network using for the soil quality assessment, which does focus on eight heavy metal pollutants, Hg, Cd, Pb, Cr, As, Zn, Cu and Ni. Firstly, the paper introduces the Back Propagation Neural Network theory using in the soil evaluation of heavy metals pollutants. Secondly, the authors designed the model with 8 input parameters. Thirdly, the model was simulated and experiments were done extensively using MATLAB neural network toolbox. In the study area City Baotou and its suburbs, we totally collected 220 soil samples from measuring points. From the GIS analysis, we find that the pollution is mainly caused by human industrial activities. And the enrichment of heavy metals in topsoil is very obvious in industrial areas and regions near both sides of the Kundulun River. It is concluded that the model of BP neural network is a practical and simple method for the soil quality assessment. The model can take advantages of simple network structure, fast convergence rate and well generality for soil heavy metal pollution evaluation. [Copyright &y& Elsevier]