LAUSR

Abstract Collapsing gullies involve considerable erosion and extreme landform changes and frequently occur in South China. The amount of rainfall infiltration has a direct influence on the occurrence of collapsing gullies. However, the mechanistic effect of infiltration or rainfall on the development of collapsing gullies still needs to be explored, as it is difficult to spatially describe infiltration amounts of event rainfall using existing models (e.g., SWAT and WEPP). A model based on cellular automata (CA) was developed for the spatial simulation of infiltration amounts after event rainfall by analyzing water allocation within and between the cells of the model. The spatial distributions of infiltration amounts after event rainfalls with differing intensities and durations were then predicted and spatially compared with the distribution of collapsing gullies. The results demonstrate that the model performs with good accuracy, as indicated by the Nash coefficient of 0.63. The variations among different rainfall infiltration amounts demonstrate that both rainfall intensity and duration promote spatial variability in infiltration amounts and reduce regional infiltration efficiency. The spatial comparison between rainfall infiltration amounts and collapsing gullies shows that rainfall infiltration promotes the occurrence of collapsing gullies, but steep topographical conditions are also required to collapse the hillside soil. The comparison also reveals that a lower-intensity rainfall event contributes more to the development of collapsing gullies than a higher-intensity rainfall event, mainly due to the soil absorbing more rainfall under lower intensity, increasing in soil weight to induce collapse. This CA-based model for spatially simulating infiltration amounts provides an improved and effective method for understanding the spatial and temporal variations in rainfall infiltration and to better capture the trends in the occurrence of collapsing gullies from hydrological effects. The model can also be customized for convenient and practical application in other watersheds.