Simulation and Prediction of PCBs in Seven Regions Using BETR-GLOBAL and Random Forest

An integrated model approaching to combining the BETR-GLOBAL model with a Random Forest method was developed in this research. Firstly, the BETR-GLOBAL model was employed to simulate the spatial and temporal distribution of PCBs concentrations in seven representative regions from 2001 to 2010. These simulation results were used to train a Random Forest regression model to predict the trend of PCBs concentrations from 2031 to 2040. In addition, Spearman’s rank correlation coefficient was applied to evaluate the relationships between PCBs pollution levels across different regions. The results show that the future trends in PCBs concentrations could effectively predicted from this method and correlations between regional pollution levels were revealed. Moreover, that PCBs are likely to persist and accumulate in the environment throughout 2031-2040 was found in the research, reflecting a spatiotemporal lag in their declines. A scientific reference may be served from the findings of this research and data supporting the control of PCBs pollution and the formulation of regional environmental policies were provided in this article.