Quantitative analysis of the influencing factors of soil detachment capacity under vegetation restoration in loess hilly and gully regions

Vegetation restoration has effectively controlled the soil detachment in the loess hilly gully region. However, the little attention has been received on the influence of the soil surface electrochemical properties on the soil detachment capacity. It is unclear on the interactions and contribution rates of the various influencing factors on the soil detachment capacity. Therefore, this study aims to quantify the influencing factors on the soil detachment capacity in the loess hilly gully region. Eight typical sites were selected with the different dominant species in the Zhifanggou watershed, according to the vegetation distribution. The soil flume scouring method was carried out on the primary trough. The flow rates were set at 0.05, 0.10, 0.15, 0.20, 0.25, and 0.30 L/s, with a slope of 15°. The soil detachment capacity of each vegetation site was then measured under different flow rates. Finally, Pearson correlation analysis, random forest, and partial least squares equation were used to quantify the relationships and pathways among the plant roots, soil properties, hydraulics, and soil detachment capacity. The results showed that there were the significant differences in the root parameters, soil properties, and soil detachment capacity among different vegetation sites in the experimental area (P<0.05). The soil detachment capacity of the different vegetation sites was ranged from 19.20 to 315.42 g/(m²·s), due to the variations of the plant roots, soil properties, and hydraulics. Pearson correlation analysis indicated that the soil detachment capacity was significantly negatively correlated with the root length density, root surface area density, and root volume density (P<0.01); There were the significant correlations with the contents of the clay, silt, and sand among the basic soil physical properties, soil organic matter and soil total nitrogen among the basic soil chemical properties, soil cation exchange capacity, soil surface potential, soil surface electric field, soil specific surface area, and soil surface charge density among the soil surface electrochemical properties (P<0.01 or P<0.05); It was also significantly positively correlated with the flow charge, velocity, shear stress, and stream power among the hydraulics (P<0.01). In addition, the random forest analysis presented that the contribution rate of the influencing factors to the variation in soil detachment capacity decreased in the order of hydraulics, root characteristics, soil surface electrochemical properties, basic soil physical properties, and basic soil chemical properties. Among them, the root length density (14.66%), the content of silt (8.09%), soil cation exchange capacity (8.18%), and stream power (7.79%) were the main influencing factors on the soil detachment capacity, whose contribution rates all reached above 7.5%. The partial least squares structural equation showed that there was the significant influence on the pathways of the plant roots and hydraulics on the soil detachment (P<0.05), with the path coefficients of -0.57 and 0.49, respectively. These findings can greatly contribute to the initiation and development of the soil erosion in the loess hilly gully region. The soil erosion can be better predicted in the Loess Plateau.