Background The saturated hydraulic conductivity (K_s) can characterize the strength of soil water conductivity, which is an important indicator of soil water conductivity and is closely related to eco-hydrology and soil erosion. Artificial shrub planting on slopes is an essential measure for controlling soil and water loss and restoring fragile ecosystems in the loess hilly areas, but the effect on K_s remains unclear. Soil transfer functions are constructed by integrating easily measurable or obtainable soil indices to estimate K_s accurately.

Methods This study focused on the sloping land planted with Caragana korshinskii for 15, 25, and 35 years and a control wasteland with wild grass in the loess hilly areas of Ningxia. Three random sample plots were set up along the horizontal slope surface, each with three replicates. The characteristics of K_s variation were investigated and its main soil influencing factors at five slope positions, including top-slope, shoulder-slope, up-slope, mid-slope and low-slope were studied, from 0 to 100 cm (0-40 cm with 10 cm intervals, 40-100 cm with 20 cm intervals). Six common soil transfer functions, Cosby1, Cosby2, Weynants, Saxton, Wang, and WÖsten, were used to fit K_s.

Results The K_s range was 6.81-60.77 mm/h, and C. korshinskii planting duration and slope position had significant effects (P < 0.05). With the increase of C. korshinskii planting years, K_s first increased (15 years of planting) and then decreased in the 0-40 cm soil layer, while K_s continuously increased in the 40-100 cm soil layer. The highest K_s was at the slope bottom, followed by the slope top and middle, and the lowest at the slope shoulder and upper slope. The simulated results of the six selected soil transfer function models indicated that, except for the WÖsten model considering soil layer depth, the predicted values of other models were all lower than the measured values. Correlation analysis showed that K_s was closely related to other physical and chemical properties except capillary porosity. The influence of soil properties on K_s was mainly closely related to the formation and stability of soil structure. Path analysis identified organic carbon, saturated water content, and bulk density as the key soil factors affecting K_s on the slope, and the constructed soil transfer function model based on these factors can predict K_s changes more accurately.

Conclusions In summary, long-term C. korshinskii planting on slopes may generally increase K_s, but this change is jointly affected by the sea buckthorn planting duration and slope position. The K_s soil transfer function constructed in this study may provide a reference for simulating and predicting K_s on slopes in the loess hilly areas.