A Modified Model for Quantitative Heavy Metal Source Apportionment and Pollution Pathway Identification

Abstract

Current source apportionment models have successfully identified emission sources and quantified their contributions. However, when being utilized for heavy metal source apportion in soil, their accuracy needs to be improved, regarding migration patterns. Therefore, this work intended to improve the pre-existing principal component analysis and multiple linear regression with distance (PCA-MLRD) model to effectively locate pollution pathways (traffic emissions, irrigation water, atmospheric depositions, etc.) and achieve a more precise quantification. The dataset of soil heavy metals was collected from a typical area in the Chang-Zhu-Tan region, Hunan, China in 2021. The identification of the contribution of soil parent material was accomplished through enrichment factors and crustal reference elements. Meanwhile, the anthropogenic emission was identified with principal component analysis and GeoDetector. GeoDetector was used to accurately point to the pollution source from a spatial differentiation perspective. Subsequently, the pollution pathways linked to the identified sources were determined. Non-metal manufacturing factories were found to be significant anthropogenic sources of local soil contamination, mainly through rivers and atmospheric deposition. Furthermore, the influence of irrigation water on heavy metals showed a more pronounced effect within a distance of 1000 m, became weaker after that, and then gradually disappeared. This model may offer improved technical guidance for practical production and the management of soil heavy metal contamination.

Keywords

• heavy metals
• source apportionment
• receptor model
• grouped principal components
• GeoDetector