Establishing a spatial grouping base for surface soil properties along urban–rural gradient—A case study in Nanjing, China

Abstract

Conventional classification systems based on vegetation and land use are frequently used to characterize or describe urban soils to determine the influence of urbanization on soils. In this study, the sensitivity of different grouping methods in reflecting soil variations along an urban–rural gradient was compared. The objective of this study was to determine the most sensitive grouping system in depicting and explaining variations of soil attributes around an urban area. Grouping methods, including urban–rural division, in situ vegetation type, land use types in different scales and numerical clustering, were compared for both single soil attributes and “soil set” defined by multiple variables. The result shows urbanization has a strong impact on many soil properties, especially that of gravel content, sand content, pH, phosphorus and soil compaction. In terms of the variations of soil attributes, in situ vegetation type is the most sensitive in comparison with local land use types and district-viewed land use types. In other words, soil properties in this study are not sensitive to coarser spatial resolution. Therefore, it's hard to interpret the spatial variation of urban soil by regular methods using natural soil-landscape paradigm. Furthermore, vegetation would best proxy the delineation of single attribute of urban soils. Numerical clusters effectively reflect the land use types and their change during urbanization. All clusters were interpreted as different sets with practical meanings: soil in abandoned greenbelt, soil in ill-managed greenbelt, soil in new vegetable land, extreme urban conditioned soil, soil in well-managed greenbelt, soil in highly mellowed vegetable land, soil in common urban–peri-urban greenbelt and weak-urban-impacted soil. They can be used as bases for soil regionalization in urban and peri-urban environment.