In recent years, China’s hilly and mountainous areas have faced widespread farmland abandonment. However, research on farmland abandonment and its driving mechanisms in hilly and mountainous regions is limited. This study proposes a transferable methodological framework that integrates Landsat data, Google Earth Engine, a time sliding-window algorithm, and the interpretable XGBoost–Shapley Additive explanation (SHAP) model. The time sliding-window algorithm is used to robustly detect long-term land cover changes across the entire study period. The SHAP quantifies the contributions of key drivers to farmland abandonment, providing transparent insights into the driving mechanisms. Applying this framework, we systematically analyzed the spatiotemporal evolution patterns and driving factors of farmland abandonment in Ji’an City, a typical city located in the hilly and mountainous areas of southern China and ultimately developed a farmland abandonment probability distribution map. The findings demonstrate the following. (1) Methodological validation showed that the random forest classifier achieved a mean overall accuracy (OA) of 91.05% (Kappa = 0.88) and the abandonment maps achieved OA of 91.58% (Kappa = 0.83). (2) Spatiotemporal analysis revealed that farmland area increased by 13.26% over 1990–2023, evolving through three stages: fluctuation (1990–2005), growth (2006–2015), and stability (2016–2023). The abandonment rate showed a long-term decreasing trend, peaking in 1998, whereas the abandoned area reached its minimum in 2007. From a spatial perspective, abandonment was more pronounced in mountainous and hilly regions of the study areas. (3) The XGBoost–SHAP model (R² > 0.85) identified key driving factors, including the potential crop yield, soil properties, mean annual precipitation, population density, and terrain features. By offering an interpretable and transferable monitoring framework, this study not only advances farmland abandonment research in complex terrains but also provides concrete policy implications. The results can guide targeted protection of high-risk abandonment zones, promote sustainable land-use planning, and support adaptive agricultural policies in hilly and mountainous regions.

Soil texture, defined by the proportions of sand, silt, and clay particles in the soil, is one of the most essential physical properties of soil. High-resolution soil texture data can provide critical parameter support for soil hydrological modeling, agricultural production management, and ecosystem assessment. In digital soil mapping, previous studies often predicted the sand, silt, and clay contents in soil and then indirectly calculated soil texture. Currently, approaches that directly map soil texture by classification modeling are gaining increasing attention due to the decreased error from data conversion, but few studies have systematically compared these two methods yet. In this study, we comprehensively assessed the performance of direct and indirect predicting soil texture using four machine learning algorithms (e.g., extreme gradient boosting, random forest, gradient boosting decision tree, and extremely randomized tree) with 190 covariates from the Digital Elevation Model, Sentinel-1/2 satellite images, and classification maps and generated a 10 m resolution soil texture map based on 405 topsoil (0–20 cm) sample data collected in Suichuan County, China. The results showed that compared with indirect predictions, direct predictions improved overall accuracy (OA) by 20.57–44.19% and the Kappa coefficient (Kappa) by 0.220–0.402. Among the models used, the XGB model achieved the highest accuracy (OA: 0.948; Kappa: 0.931) and the lowest uncertainty (confusion index: 0.052). The direct prediction map (nine classes recorded) exhibited more detailed and diverse spatial distribution patterns than the indirect prediction map (six classes recorded), aligning better with the actual environment. Based on accuracy validation and spatial distribution, the performance of the XGB model was best during direct prediction. The Shapley additive explanation from the XGB model revealed that the normalized height and stream power indices were the most significant factors driving the soil texture in the study area. Our results provide a reference for future studies on soil texture mapping using machine learning models.

The concept of ecosystem health (EH), proposed in the 1980s