The intensification of land use has contributed to the emergence of environmental impacts such as soil loss, silting of water bodies, and reduction of biodiversity, among others. Using models capable of seasonally diagnosing environmental damage is essential in territorial planning and management, demonstrating the spatial distribution of the environment's sensitivity to developing erosion processes and quantitatively valuing soil loss. Thus, assuming a significant relationship exists between the seasonal variation in environmental fragility and the validated estimate of soil loss, reflecting the conservation status of the river basin. Therefore, this work aims to analyze the seasonal Environmental Fragility (EF) from the autumn of 2019 to the summer of 2020 using the soil loss estimate. Data such as slope, erodibility, erosivity, and the normalized difference vegetation index (NDVI) were used to achieve this. Statistical tests were also applied to assess the significance level of the models in the seasonal evaluation and the validation based on ground truth points. The results showed seasonal differentiation in the EF and the soil loss estimation. Spring was the one that resulted in the most extensive area classified as high EF (27%) and with an estimated soil loss of 0.3733 t.ha-1month-3. The summer presented the highest soil loss estimation with an average value of 0.4393 t.ha -1month-3. Autumn (0.07683 t.ha-1 month-3) and winter (0.0569 t.ha-1 month-3) showed the lowest rates of soil loss, and the most prominent areas were classified in the low class of EF, as a result, mainly of the erosivity of the rains. The results indicated by the seasonal models of EF and soil loss were validated through erosion points using spatial statistics tests.

Engineered loess-filled gullies, which are widely distributed across China's Loess Plateau, face significant stability challenges under extreme rainfall conditions. To elucidate the regulatory mechanisms of antecedent rainfall on the erosion and failure processes of such gullies, this study conducted large-scale flume experiments to reveal their phased erosion mechanisms and hydromechanical responses under different antecedent rainfall durations (10, 20, and 30 min). The results indicate that the erosion process features three prominent phases: initial splash erosion, structural reorganization during the intermission period, and runoff-induced gully erosion. Our critical advancement is the identification of antecedent rainfall duration as the primary pre-regulation factor: short-duration (10-20 min) rainfall predominantly induces surface crack networks during the intermission, whereas long-duration (30 min) rainfall directly triggers substantial holistic collapse. These differentiated structural weakening pathways are governed by the duration of antecedent rainfall and fundamentally control the initiation thresholds, progression rates, and channel morphology of subsequent runoff erosion. The long-duration group demonstrated accelerated erosion rates and greater erosion amounts. Concurrent monitoring demonstrated that transient pulse-like increases in pore-water pressure were strongly coupled with localized instability and gully wall failures, verifying the hydromechanical coupling mechanism during the failure process. These results quantitatively demonstrate the critical modulatory role of antecedent rainfall duration in determining erosion patterns in engineered disturbed loess, transcending the prior understanding that emphasized only the contributions of rainfall intensity or runoff. They offer a direct mechanistic basis for explaining the spatiotemporal heterogeneity of erosion and failure observed in field investigations of the engineered fills. The results directly contribute to risk assessments for land reclamation projects on the Loess Plateau, underscoring the importance of incorporating antecedent rainfall history into stability analyses and drainage designs. This study provides essential scientific evidence for advancing the precision of disaster prediction models and enhancing the efficacy of mitigation strategies.

Gullying is one of the problems that cause soil degradation in semi-arid areas and should be predicted to mitigate its damaging effects. Three machine learning models have been employed in this work to map the susceptibility to gully erosion in the N'fis river basin in the Moroccan High Atlas. Utilizing high-resolution images from Google Earth alongside fieldwork data, we digitized 434 gully erosion events to construct the comprehensive inventory map. These data were divided into two groups: training (70%) and test (30%). Based on the literature research and the multicollinearity test, 11 conditioning factors were selected. The receiver operating characteristic (ROC) approach and other statistical measures were used to quantify the model's accuracy. The study findings highlight the significance of drainage density, slope, NDVI, and distance from roads as crucial factors influencing gully erosion in the study area. Among the evaluated machine learning algorithms, the random forest (RF) model exhibited the highest performance, with an area under the curve (AUC) value of 0.932. It was followed by adaptive boosting (AB) with an AUC of 0.902 and gradient-boosted decision trees (GBDT) with an AUC of 0.893. The maps produced reveal that the southern and central regions of the study area have the classes of very high and high gully erosion susceptibility. The outputs of the current study can be used by decision-makers to improve prevention planning and mitigation techniques against gully erosion damage.