Fractal features of soil particle size distribution under different land-use patterns in the alluvial fans of collapsing gullies in the hilly granitic region of southern China

Collapsing gullies are among the most severe soil erosion problems in the tropical and subtropical areas of southern China. However, few studies have examined the relationship of soil particle size distribution (PSD) changes with land-use patterns in the alluvial fans of collapsing gullies. Recently, the fractal method has been applied to estimate soil structure and has proven to be an effective tool in analyzing soil properties and their relationships with other eco-environmental factors. In this study, the soil fractal dimension (D), physico-chemical properties and their relationship with different land-use patterns in alluvial fans were investigated in an experiment that involved seven collapsing gully areas in seven counties of southern China. Our results demonstrated that different land-use patterns of alluvial fans had a significant effect on soil physico-chemical properties. Compared to grasslands and woodlands, farmlands and orchards generally contained more fine soil particles (silt and clay) and fewer coarse particles, whereas significant differences were found in the fractal dimension of soil PSD in different land-use patterns. Specifically, the soil fractal dimension was lower in grasslands and higher in orchards relative to that of other land-use patterns. The average soil fractal dimension of grasslands had a value that was 0.08 lower than that of orchards. Bulk density was lower but porosity was higher in farmlands and orchards. Saturated moisture content was lower in woodlands and grasslands, but saturated hydraulic conductivity was higher in all four land-use patterns. Additionally, the fractal dimension had significant linear relationships with the silt, clay and sand contents and soil properties and exhibited a positive correlation with the clay (R(2) = 0.976, P<0.001), silt (R(2) = 0.578, P<0.01), organic carbon (R(2) = 0.777, P<0.001) and saturated water (R(2) = 0.639, P<0.01) contents but a negative correlation with gravel content (R(2) = 0.494, P<0.01), coarse sand content (R(2) = 0.623, P<0.01) and saturated hydraulic conductivity (R(2) = 0.788, P<0.001). However, the fractal dimension exhibited no significant correlation with pH, bulk density or total porosity. Furthermore, the second-degree polynomial equation was found to be more adequate for describing the correlations between soil fractal dimension and particle size distribution. The results of this study demonstrate that a fractal dimension analysis of soil particle size distribution is a useful method for the quantitative description of different land-use patterns in the alluvial fans of collapsing gullies in southern China.