BP Neural Network-Based Determination of Soil-Codonopsis under Different Sources Key Factors for Enrichment and Transport of Heavy Metal Elements in the System

In this paper, the Chinese Taoist medicinal herb Codonopsis and its rhizosphere soils from Pingshun County, Shanxi Province, were used as the research objects. Using GIS, descriptive statistics, The BP neural network model constructed by Python, we studied the potential factors to control the enrichment and migration of heavy metal elements in the soil-Codonopsis system under different sources. It was found that the heavy metal elements in the soils of the region were mainly divided into natural background and source of mineral veins, and the content of heavy metal elements in Codonopsis and rhizosphere soils under vein sources was found to be higher than the natural background, while there was no significant change in the corresponding rhizosphere soils, and the biological effectiveness of heavy metal elements in Codonopsis under vein sources was significantly increased, among which Cu, Cd were even more enriched. The bioeffectiveness of heavy metal elements in Codonopsis from mineral veins was significantly increased, among which cu and cd were superenriched. The migration of most heavy metal elements in the natural background was mainly influenced by soil pH, while in the case of mineral vein contamination, the key factors controlling the migration of heavy metal elements were mainly inter-element interactions. Therefore, the subsequent planting of Codonopsis should be far away from the area of vein development and focus on the effect of pH on its quality.