Background Ecological restoration featured with revegetation plays effective role on mitigating soil erosion and related land degradation and disasters. It significantly reduced the surface runoff and sediment generation, enhanced the soil resistance to shear forces, prevent slope collapse and eventually reshaped the morphology of gullies. Therefore, morphological traits of gullies are potential to indicate the impact of ecological restoration, and contribute to assessment of the achievement of soil and water conservation measures. This study focused on the comparative analysis on morphological traits of gullies on the farmland hillslope and the revegetated hillslope in the loess environment, and aimed to reveal the impact of vegetation restoration on the morphological traits of gullies, and reflects the effectiveness of ecological restoration by adopting gully morphological traits as indicators.

Methods A farmland hillslope and a revegetated hillslope with gully system occurred were selected to conduct the comparative analysis of gully morphological traits. High-precision DEM derived from point clouds obtained by using UAV-LiDAR scanning was produced for quantification of the gully morphological traits, including width-height ratio of cross sections and the surface roughness of gully beds. The division of gully systems was implemented based on variation of slope. The Shreve stream ordering method was applied to classify the gully channel network. Subsequently, the point cloud data of cross-sections at gully heads, branch gully, and trunk gully was extracted for gully morphology analysis. Besides, the surface roughness of gully bed was quantified as the deviation of elevation to the mean elevation at a moving window on the high-resolution DEM.

Results Generally, the morphological traits of gullies were apparently different between farmland hillslope and revegetated hillslope. The results indicated that there was noticeable difference in morphological traits between gully systems on farmland hillslope and revegetated hillslope. 1)Average slope gradient of gully walls on farmland hillslope significantly smaller than that on the revegetated hillslope. Peculiarly, gully walls at the gully heads on the revegetated hillslope were steeper than those of gullies on the farmland hillslope (P < 0.001). The average slope gradient of the gully bed is slightly lower on the farmland hillslope than on the revegetated hillslope, while the average slope gradient of the gully slope is similar between the two hillslopes. 2) The width-height ratios of cross-sections at the gully heads and at the branch gullies were significantly different between the two hillslopes (P < 0.001), which is 1.564 on the farmland hillslope and 1.014 on the revegetation hillslope on average. However, the difference in width-height ratio of cross-sections of gullies on the two hillslopes was diminished at the trunks of the gully systems on the two hillslope (P = 0.179), the corresponding width-height ratio of 1.253 and 0.840 on revegetated hillslope and farmland hillslope, respectively. 3) The roughness of the gully bed was higher on the revegetated hillslope than that on the farmland hillslope, with the beds of branch gullies exhibiting greater roughness than that of trunk gullies.

Conclusions This study highlighted the role of revegetation played in mediating gully morphological traits. It was implied that gully morphological traits can be adopted as indicators to assess the effectiveness of ecological restoration on the Loess Plateau. However, this study is limited to the semiarid hilly areas on the Loess Plateau covered by thick loess. The effectiveness of the gully morphological traits as indicators for ecological restoration need to be test further in different environment setting.