Identifying the influencing factors controlling the spatial variation of heavy metals in suburban soil using spatial regression models.

Determining the factors that control the spatial variation of heavy metals in suburban soil is important in identifying and preventing pollution sources. Soil intrinsic factors combined with environmental variables can effectively explain the spatial distribution of heavy metals. Compared with classical statistical methods, such as multiple linear regression (MLR) models, spatial regression models that can cope with the spatial dependence of heavy metals have greater potential in establishing an accurate relationship between influencing factors and heavy metals. This study aims to identify the factors that influence the spatial variation of lead (Pb) and cadmium (Cd) in 138 topsoil samples from the suburbs of Wuhan City, China, by using spatial regression models with MLR as the reference. Moran's I values reveal the spatial autocorrelation of Pb and Cd. The spatial lag model (SLM) outperforms MLR and has higher R2 and lower spatial dependence of residuals. The significant coefficients of the spatial lag term in SLMs indicate that the spatial variation of Pb and Cd depends on their surrounding observations. SLM results show that Pb content depends on the distance from the nearest industrial enterprises and suggest that industrial pollution is the main source of Pb. Cd content depends on pH, soil organic matter, and the topographic wetness index, indicating that intrinsic and topographical factors contribute to the spatial variation of Cd. Parent materials and application of phosphorus fertilizer are the most likely sources of Cd. The findings highlight the spatial autocorrelation of heavy metals and the effects of intrinsic factors and environmental variables on the spatial variation of such metals. Moreover, this study reveals the effectiveness of spatial regression models in identifying the influencing factors of heavy metals.