Soil erosion is largely driven by climate change and land use dynamics. The objective of this study is to assess the dynamic variation in erosion under the combined effects of precipitation and land use change in the Tigrigra watershed, located in the mountainous region of the Middle Atlas. The RUSLE (Revised Universal Soil Loss Equation) model is used in the methodological approach to estimate soil loss based on various parameters such as precipitation, soil, topography, land cover, and conservation practices. Geographic Information Systems (GIS) and remote sensing tools are essential for applying this method. In addition, the CA-Markov model (cellular automata), which models and predicts land use changes over time, is used to project future land cover scenarios that influence soil erosion dynamics. The research focuses on four previous periods (1991–2000, 2001–2010, 2011–2015, and 2016–2023), as well as a future period (2024–2050), considering two climate scenarios, RCP 2.6 and RCP 4.5. Precipitation data from local weather stations and the CMIP5 climate model were used to calculate the R factor (precipitation erosivity). Land cover analysis was performed using Landsat satellite images (30 m resolution) integrated into the CA-Markov model to calculate the C factor (land cover management). The results show that erosion has gradually decreased over both past and future periods, mainly due to variations in precipitation and vegetation cover. It should be noted that the period from 1991–2000 to 2016–2023 shows higher erosion compared to the future periods, with a maximum value of 17.83 t/ha/year recorded between 1991 and 2000. For the future period 2024–2050, a continuous decrease in erosion is observed under both scenarios, with an average value of 15.30 t/ha/year for the RCP2.6 scenario and 15.86 t/ha/year for the RCP4.5 scenario, with erosion remaining slightly higher under RCP4.5. Overall, erosion decreases across both historical (1991–2023) and projected (2024–2050) periods due to reduced rainfall erosivity. The northern part of the basin is particularly prone to erosion due to the low vegetation cover. The results indicate that areas susceptible to erosion require conservation measures to reduce soil loss. Implementing sustainable agricultural practices is crucial for maintaining long-term soil health and preventing degradation. However, some limitations of the study, such as the lack of data on conservation practices and daily precipitation, might affect the overall robustness of the findings.

This study proposes an innovative framework that combines hydroclimatic and agro-environmental predictors, including nitrate concentration and NDVI, with climatic factors such as Rainfall, temperature, and evapotranspiration to forecast piezometric level variations in the Saïss Basin, Morocco. Eight Machine Learning (ML) models were benchmarked, and feature importance was assessed using Shapley Additive exPlanations (SHAP) to ensure model transparency and interpretability. In parallel, the PELT algorithm was applied to detect structural change points, while Sen’s slope estimator and the Mann–Kendall test were used to quantify long-term trends. The Extra Trees (ET) model achieved the best performance (R² = 0.92), with Rainfall emerging as the most influential predictor, followed by nitrate concentration, confirming the added value of hydrochemical indicators for groundwater forecasting. Change-point analysis revealed significant declines during the 1980s and 1990s, followed by lower-amplitude fluctuations since the late 2000s. Projections toward 2050 suggest partial stabilization in the central part of the basin under favorable recharge conditions, whereas persistent declines are expected to continue in peripheral areas subjected to sustained groundwater abstraction pressure. These findings provide a robust and transferable decision-support tool for the sustainable management of groundwater resources in semi-arid agricultural area.