Factorial kriging and stepwise regression approach to identify environmental factors influencing spatial multi-scale variability of heavy metals in soils

Abstract

The knowledge about spatial variations of heavy metals in soils and their relationships with environmental factors is important for human impact assessment and soil management. Surface soils from Rizhao city, Eastern China with rapid urbanization and industrialization were analyzed for six key heavy metals and characterized by parent material and land use using GIS-based data. Factorial kriging analysis and stepwise multiple regression were applied to examine the scale-dependent relationships among heavy metals and to identify environmental factors affecting spatial variability at each spatial scale. Linear model of coregionalization fitting showed that spatial multi-scale variation of heavy metals in soils consisted of nugget effect, an exponential structure with the range of 12km (short-range scale), as well as a spherical structure with the range of 36km (long-range scale). The short-range variation of Cd, Pb and Zn were controlled by land use, with higher values in urban areas as well as cultivated land in mountain area, and were related to human influence; while parent material dominated the long structure variations of these elements. Spatial variations of Cr and Ni were associated with natural geochemical sources at short- and long-range scales. At both two scales, Hg dominated by land use, corresponded well to spatial distributions of urban areas, and was attributed to anthropic emissions and atmosphere deposition.