Evaluation of heavy metal pollution in the sediment of Poyang Lake based on stochastic geo-accumulation model (SGM)

The uncertainties introduced by sampling errors, measurement errors, and sediment heterogeneity in the evaluation of heavy metal pollution in the sediment of Poyang Lake are inevitable. The conventional geo-accumulation index (IGeo) cannot overcome these uncertainties. Thus, a stochastic geo-accumulation model (SGM) is established to solve this problem. In the SGM, the distribution of the heavy metal's concentration is stimulated by maximum entropy principle. Then, a membership vector is designed to quantify the pollution condition of each pollutant. Finally, a synthetic membership vector is generated to represent the comprehensive situation of heavy metal pollution in the sediment. SGM is applied in the evaluation of heavy metal pollution in four wetlands of Poyang Lake. Results show that (i) the SGM exhibits better capabilities in uncertainty analysis, risk recognition, and comprehensive pollution evaluation than the conventional IGeo and Hakanson index (HI) models. (ii) The pollution grade of heavy metals in the sediment of Longkou Wetland is "moderately contaminated," whereas the pollution categories in Nanjishan, Wucheng, and Baishazhouwetlands are "uncontaminated to moderately contaminated." (iii) Copper and lead are the key risk factors in Poyang Lake. (c) 2018 Elsevier B.V. All rights reserved.