Soil detachment is one of the most important processes of soil erosion, as it is of great significance for the prevention and treatment of soil erosion in areas subject to seasonal freeze-thaw. However, most previous studies on the effect of freeze-thaw on soil detachment capacity (SDC) of bare soil, little research on SDC under the effect of freeze-thaw and the root system. This study investigated the effects of freeze-thaw and the root system on soil detachment capacity through hydraulic flume experiments to simulate the soil detachment process of two soil types, sand soil and loessal soil, under four treatments, control, freeze-thaw, root system and freeze-thaw + root system. And a prediction model was developed to calculate SDC under the effect of freeze-thaw and the root system. The results illustrated that the SDC of sand soil was higher than that of loessal soil. The SDC of two soils was reduced and increased by the root system and freeze-thaw, respectively, although the former effect was significant (P < 0.05) whereas the latter was not. The effect of freeze-thaw in combination with the root system showed that the root system contributed the majority of SDC variability (99.95%); therefore, inhibition of SDC by the root system played a leading role. When comparing shear stress, unit energy of the water carrying section and unit stream power, stream power was found to be the hydraulic parameter that best predicted SDC (R2¬ > 0.84). The inclusion of root weight significantly improve the accuracy of the SDC prediction model developed by hydraulic parameters. A general model based on stream power and root weight was developed to quantify SDC and was shown to have a high SDC prediction accuracy for both soils treated by freeze-thaw and the root system [NSE = 0.88，R2 = 0.90].
Groundwater is an important water resource for the ecological restoration and the life and production of the inhabitants of the Loess hilly region. A quantitative exploration of groundwater recharge mechanism and the spatial-temporal variation pattern is the key for an effective evaluation of groundwater resources in this region. Based on a Bayesian model and a Sinusoidal model, the spatial-temporal distribution pattern of the water transmission time and the groundwater recharge ratio were investigated. The results showed that the transmission time of groundwater by precipitation and surface water were 443.16 d and 64.58 d, respectively, which yielded a recharge ratio of 29.22% and 70.78%, respectively. Surface water was the main recharge source of groundwater in this region. The values of the groundwater recharge ratio by precipitation in the rainy season and dry season were 32.83% and 25.60%, respectively. Groundwater recharge mainly occurred during the rainy season. From upstream to downstream of the small watershed, the groundwater recharge ratio by precipitation increased gradually, while the groundwater recharge ratio by surface water decreased. The spatial characteristics of groundwater recharge ratio were all nonetheless not obvious. Groundwater recharge mainly took place in the upstream region of the watershed, while the discharge took place in the downstream area. Groundwater recharge occurred mostly through a combination of piston flow and preferential flow. The findings of this study could provide a reference for the development, the utilisation and the protection of groundwater resources in small watershed in the Loess Hilly region.