Liang, Wenqian , Xu, Ximeng , Zheng, Fenli , Qin, Chao
2025-10-01 CATENA 2025 258(卷), null(期), (null页)
Ephemeral gullies (EGs) are distributed on the farmlands across the world and play a significant role in soil loss and land degradation. However, the complete cyclic evolution process of EG channels, including formation, obliteration by tillage, and reformation under subsequent rainfall, remains underexplored. Thus, this study aims to investigate the EG evolution and morphological changes in the erosion-tillage-erosion cycle on a steep loess slope by using 0.01 m resolution digital elevation models (DEMs) obtained from Terrestrial Laser Scanner. A total of 107 cross-sections across three EGs were extracted, and six morphological indices, including W (EG channel width), D (EG channel depth), A (EG channel cross-sectional area), alpha (the angle from the EG lateral drainage area border to the EG channel border), /i (the angle from the EG channel border to the bottom), and W/D (EG channel width-depth ratio), were obtained. The results revealed that, from the post-rainy season in 2014 to the post-rainy season in 2015, W, D, and A increased by 8%-25%, 20%-40%, and 35%-69%, respectively, while alpha, /i, and W/D exhibited minor changes. EG with larger upslope drainage areas experienced more pronounced evolution, with D, A, and alpha increasing linearly with upslope drainage area across the three EGs, ranging from 0.05 m to 0.49 m, 67 cm2 to 2062 cm2, and 10 degrees to 26 degrees, respectively. A new cross-section morphological index incorporating upslope and lateral flow contributions, alpha & sdot;tan(/i), was introduced, which showed a consistent linear relationship with the product of the upslope drainage area and the square of local slope gradient in EGs, offering a tool for predicting EG cross-sectional morphology. This study provided insights into EG evolution and illustrated the role of tillage practices in EG evolution processes, which can provide a scientific basis for the construction of gully erosion prediction models.