Quantifying the effects of root, soil properties and hydraulic characteristics on soil detachment of naturally restored grassland on the Loess Plateau

Plant roots, soil properties, and hydraulic characteristics can significantly influence soil detachment by overland flow. However, research quantifying the effects of those various factors on soil detachment capacity ( Dc ) of naturally restored grasslands on the Loess Plateau of China is limited. Moreover, the impact of soil surface electrochemical properties ( EPS ) on Dc has received little attention. Therefore, 135 soil samples were collected from three typical experimental areas on the Loess Plateau and subjected to flow scouring in a hydraulic flume under five shear stresses (combinations of a 15° slope and flow discharges of 0.1, 0.2, 0.3, 0.4, and 0.6 L/s). The measurements were taken for plant roots and soil properties, including basic soil physicochemical properties ( BPS ) and EPS . The results indicated that the mean Dc at different sampling sites ranged from 1.11 to 203.25 g/(m 2 ·s). Dc had a significant exponential correlation with root length density ( RLD ), root surface area density, the sand content, the content of water-stable aggregate > 0.25 mm, and soil surface potential ( φ 0 ) (coefficient of determination ( R 2 ) ranged from 0.315 to 0.771; p < 0.01); Dc was significantly related via a power function to the clay content ( Clay ), soil organic matter ( OM ), the contents of aluminum oxide (Al 2 O 3 ) and iron (Ⅲ) oxide (Fe 2 O 3 ), soil cation exchange capacity, soil specific surface area, flow charge, shear stress, stream power ( w ), and flow velocity ( R 2 ranged from 0.636 to 0.986; p < 0.01). Dc demonstrated significant linear relations with soil bulk density, soil exchangeable sodium percentage, soil surface charge density ( σ 0 ), and soil surface electric field ( R 2 ranged from 0.476 to 0.706; p < 0.01); and Dc was significantly logarithmically related to the silt content ( R 2 = 0.598; p < 0.01). In addition, the ranking of the influence degrees of different factors on the Dc variation was as follows: hydraulic characteristics > EPS > basic soil physical properties > basic soil chemical properties > plant roots. Among them, RLD , Clay , OM , σ 0 , and w were the indicators with the highest explanatory rates among the aforementioned different influencing factors in sequence. The path analysis further indicated that EPS and hydraulic characteristics were the significant factors affecting Dc , with the direct effect playing a dominant role. These findings enhance the understanding of soil erosion mechanisms and contribute to the development of soil erosion models for the Loess Plateau.