[A review of multi-scale studies on spatial variation of the lead (Pb) concentration in urban soils].

The accumulation of Pb in urban soils is still apparent in China, and scientific assessment and management of risks from Pb-contaminated soils is necessarily based on contamination levels and extent evaluated accurately. Lead concentration in urban soils has a strong spatial variation and complex spatial structures. Carried out in a single spatial scale, most current investigations cannot comprehensively reveal characteristics of spatial structures, and did not promote more scientific assessment and management of risk. Exploring a new method which can help identify the overall spatial structures is needed. To achieve this aim, this paper firstly investigated the factors linked to the spatial variability of Pb concentration in urban soils, and three major factors were identified: various pollution sources, hierarchical pollution processes and heterogeneous urban landscape. These factors were form a nested hierarchical spatial structure with three spatial levels. Based on the conceptual spatial structure, we proposed a method framework guided by geostatistical theory and focused on linear mixed model (LMM). This proposed framework can divide the nested hierarchical spatial structures of Pb concentration in urban soils into three levels: global trend, random variation with spatial autocorrelation, and outliers. Two recommendations were given to promote the multiple-scale investigation in spatial variation of soil Pb contamination in urban area including: finding more efficient sampling strategy and determining the characteristic scale.