A Bayesian Network For Both Land Subsidence Risk And Soil Contamination Risk Evaluation In Large-Scale Reclaimed Lands Of Shanghai, China

Shanghai has the largest newly reclaimed area in China, and the associated soil contamination risk exists. To evaluate both land subsidence risks and soil heavy metal contamination (HMC) risks in Shanghai's coastal reclaimed regions, a Bayesian Network (BN) was established taking sixteen variables such as average initial void of underlying strata, land reclamation time, thickness of the reclaimed layer, and soil HMC into consideration. In the BN, the ultimate land subsidence since July 1st 2018 and the land subsidence velocity characterized by the land subsidence from July 1st 2018 to January 1st 2019 analytically evaluated in typical reclaimed regions of Shanghai were used to characterize land subsidence. Seven heavy metal elements, i.e., Zn, Cd, As, Ni, Cu, Pb and Cr were included into the HMC evaluation. Influence strength of the BN arcs and node sensitivity were analyzed. The BN analysis shows that the two study zones hold remarkable differences in field basic characteristics, geotechnical parameters and soil HMC distribution. The three land subsidence risk variables hold little correlation with the HMC distribution, but they are both correlated tightly with the basic characteristics of the study points.