Identifying factors of trace metal concentrations in urban residential soil at different geographical scales using an optimal parameters-based geographical detector model

Abstract

Australia’s national citizen program VegeSafe has collected > 22,000 urban residential soils and analyzed them for trace element concentration. After eight years of operation, the large quantity of data accumulated by the program has enabled unparalleled and robust factor investigation of trace element contamination (such as Pb) in soils to better understand potential exposure risks. This analysis aims to explore anthropogenic and natural factors influencing the spatial pattern of soil Pb concentration at different geographical scales (from 0.1 to 4 km). We interpolated VegeSafe soil Pb data within the Sydney greater metropolitan area, combined with other socio-economic and environmental data, and applied an optimal parameters-based geographical detector model (OPGD) that can identify the optimal spatial scale for spatial heterogeneity analysis. Results show that relative impacts of different factors are varied with the change of spatial scale. Anthropogenic factors (percentage of old homes, road density, percentage of industrial area, etc.) were major contributors to soil Pb concentration with increasing impacts as the scale increased. Natural factors (soil pH values, soil type, elevation, etc.) played a secondary role with stable impacts on the scale of 0.1 to 2 km and increasing impacts on the scale of 2 to 4 km. Most factors reached the maximum impact when the spatial scale was 4 km, selected as the optimal spatial scale for the subsequent factor exploration. By incorporating the process of spatial scale optimization, the VegeSafe data plus OPGD modelling framework provide a more accurate understanding of trace element contamination in urban residential soil.