The fractal characteristics of soil particle-size distribution (PSD) constitute an important soil physical property, and fractal models of soil PSD are increasingly used to describe the effects of vegetation on the improvement of soil-related properties. Based on the fractal theory, this paper selected four typical vegetation restoration types (Quercus acutissima, QAC; Robinia pseudoacacia, RPL; Pinus densiflora, PDS; QAC × PDS) in the Taiyi mountainous area as the research object, and the single-fractal dimension (D) and multi-fractal parameters of PSD and its correlation with soil-associated properties were studied. The results show that (1) QAC × PDS reduced the heterogeneity of soil sand distribution, which also increased the range and concentration of soil PSD in the dense area. Soil clay and silt contents showed QAC × PDS > RPL > QAC > PDS. QAC × PDS significantly increased clay and silt content in 0–20 cm soil. (2) D varies among different vegetation restoration types, which was QAC × PDS > RPL > QAC > PDS, and the D of 0–20 cm soil was greater than that of 20–40 cm. For the multi-fractal parameters of PSD, the capacity dimension (D0), information dimension (D1), correlation dimension (D2) and D1/D0 in 0–20 cm soil of different vegetation restoration types showed significant differences. (3) D and multi-fractal parameters were significantly positively correlated with clay and silt contents, which were significantly negatively correlated with sand contents. (4) QAC × PDS had the smallest soil bulk density and largest porosity. Fractal dimension was negatively correlated with soil bulk density and positively correlated with soil total porosity and capillary porosity. These results indicate that the soil fractal dimension can well characterize the vegetation improvement on soil structure and properties in the earthy-rocky mountain areas of northern China.

Subsurface runoff is central to sloping land hydrology research. To explore the mechanism of nitrogen (N) and phosphorus (P) loss in surface and subsurface runoff in mountain forestland, we selected the typical tree species Quercus acutissima Carr. and Robinia pseudoacacia L. as our research objects and carried out simulated rainfall experiments to study the N and P loss characteristics of surface and subsurface runoff on different areas of slope land. Our results show that total N and total phosphorus P concentrations in surface runoff were initially high but tended to be stable as rainfall continued. Surface runoff accounted for most losses of both N and P, with a greater proportion found for forested land than for wasteland. Surface losses of N ranged from 70.8–88.8%, and those of P ranged from 82.2–87.5% of total losses in forest and wasteland, respectively. Our results show that the main pathway for N and P loss was surface runoff. However, N and P loss in subsurface runoff cannot be ignored in research on sloping land hydrology and the control by forestland of nonpoint source pollution in mountain forestland. Our research enriches sloping land hydrology research and provides a basis for the effective control of runoff, N, and P loss in forestland of the mountainous region of China and similar areas around the world.