This paper investigates the relations between soil organic carbon (SOC) and other physicochemical indices in low hills of Chifeng, Inner Mongolia, and thus to help with soil modification in low hills there. Grey relational analysis and stepwise regression analysis were used to uncover the relations of SOC with total N (TN), pH, available P (AP), available K (AK), and microbial biomass carbon (MBC) under eight land use patterns in a typical small watershed in Chifeng. The eight land uses were Prunus sibirica land, Caragana microphylla land, Populus simonii land, Ulmus pumila land, natural grassland, Pinus Tableulaeformis [Formula: see text] P. sibirica mixed land, secondary forest of natural bush land, and agricultural land. The average SOC contents in the 0–100[Formula: see text]cm layer in the above eight land use patterns are 7.72, 5.32, 7.40, 6.11, 3.14, 10.26, 17.51 and 5.34[Formula: see text]g/kg, respectively, indicating significant differences. The relation coefficients of SOC with TN, pH, MBC, AK, and AP are 0.70, 0.66, 0.63, 0.57 and 0.55, respectively, under all land use patterns. SOC contents are significantly different among these land use patterns, and the relation coefficients of SOC with other physicochemical indices rank in the order as: [Formula: see text]. The optimum regression equation of SOC ([Formula: see text] with TN ([Formula: see text] and pH ([Formula: see text] was built: [Formula: see text], ([Formula: see text]).
Spatial heterogeneity of soil organic carbon in a karst region under different land use patterns