Topography-soil relationships in a hilly evergreen broadleaf forest in subtropical China

Purpose: Topography-soil relationships usually vary with climate, vegetation type, degree of human disturbance, type of parent material, and the scale being studied. In this paper, we studied the topography-soil relationship in a hilly forest in subtropical China. Materials and methods: The influence of topography on soil properties (soil moisture, organic carbon (C), total nitrogen (N) and total phosphorus contents, C:N ratio, and pH) was evaluated using a recursive partitioning conditional inference tree (CIT) as well as a multiple linear regression (MLR) method. Results and discussion: The CIT models generally performed better than MLR in describing the topography-soil relationships. Topographic parameters chosen by the CIT models, which indicate the mechanisms at play for the spatial variation of the soil properties, varied with the soil property of concern. The soil moisture, organic C, and total N models contained only primary terrain attributes, the soil C:N ratio and pH models contained both primary and secondary terrain attributes, while the total phosphorus model contained mostly secondary terrain attributes. Conclusions: The CIT method worked well for exploring the topography-soil relationships in the studied undisturbed hilly forest. We conclude that (1) soil moisture, organic C, and total N were strongly affected by location-specific topographic features such as gravitational potential, the amount of precipitation, temperature, and vegetation type; (2) total phosphorus was affected by catchment-related hydrological activities and soil C:N ratio; and (3) pH was affected by location-specific topographic features and catchment-related hydrological activities.