Floods and erosion are natural hazards that present a substantial risks to both human and ecological systems, particularly in coastal regions. Flooding occurs when water inundates typically dry areas, causing widespread damage, while erosion gradually depletes soil and rock through processes driven by water and wind. This study proposes an innovative approach that integrates Deep Neural Decision Forest (DeepNDF), Feedforward Neural Network (FNN), autoencoders, and Bidirectional Recurrent Neural Networks (Bi-RNN) models for flood prediction, enhanced through transfer learning for erosion mapping in coastal environments. Utilizing multi-source datasets from the United States Geological Survey (USGS), Climate Hazards Group InfraRed Precipitation with Station (CHIRPS), the National Algerian Institute of Cartography, and Sentinel-2 imagery, the key conditioning factors using Geographic Information System (GIS) were generated. The conditioning factors included elevation, slope, flow direction, curvature, distance from rivers, distance from roads, hillshade, topographic wetness index (TWI), stream power index (SPI), geology, and land use/land cover (LULC), as well as rainfall. To ensure the modeling reliability, the performance was rigorously evaluated using multiple statistical metrics, including the Area Under the Curve-Receiver Operating Characteristic (AUC-ROC), Precision, Recall, and F1 Score. The DeepNDF model achieved the highest performance for flood prediction with an AUC-ROC of 0.97, Precision of 0.93, Recall of 0.92, and an F1 Score of 0.925, while the transfer learning approach significantly improved erosion prediction, reaching an AUC-ROC of 0.92, Precision of 0.90, Recall of 0.92, and an F1 Score of 0.91. The analysis indicated that flood risks predominantly affected rangeland (18.68%) and bare ground (20.48%), while cropland was found to face the highest erosion risk, affecting approximately 3,471 km2. This research advances predictive modelling in hydrology and environmental science, providing valuable insights for disaster mitigation and resilience planning strategies in coastal areas.

In the context of climate change, rainstorm events are becoming increasingly frequent. In particular, on the Loess Plateau, heavy rainstorms are the primary cause of soil erosion. This study investigated and analysed different types of soil erosion hotspots and influencing factors in small watersheds under different rainstorm events in different areas of the Loess Plateau. The results indicate that the erosion intensities of rills, gullies, landslides and collapses ranged from 13600-46244, 1982-772201, 1163-172153 t km-2 and 1867-94985 t km-2, respectively. Newly constructed terraces exhibited an erosion intensity 1.6 times greater than that of old terraces, while terraces constructed before the rainy season in the current year exhibited an erosion damage intensity 2.6 times greater than that of terraces constructed after the rainy season in the previous year. In addition, under rainstorm conditions, landslides represented the most severe type of erosion in the watersheds, with the maximum amount of erosion accounting for more than 90 % of the total erosion amount, followed by gully or collapse erosion, with the collapse of terrace risers as the main contributor. Slope cultivation land, unpaved roads, terrace risers, and valley slopes below the gully shoulder line were identified as erosion hotspot areas. Rainstorm erosion was significantly influenced by the land use type and slope, which explained 14.2 %-41.5 % and 9.7 %-15.1 %, respectively, of the total variance in erosion intensity. We suggest that soil erosion prevention and control efforts on the Loess Plateau should focus on landslides on valley slopes below gully shoulder lines, followed by gullies on unpaved roads and the collapse of terraced fields. Drainage ditches and water cellars should be constructed above the gully shoulder line and on the inside of roads and terraces, thereby reducing erosion. Our research is crucial for optimizing and adjusting watershed management measures and preventing rainstorm erosion disasters.