Prediction of soil erosion control ecosystem service using machine learning based on the ANN model in Asia

The impact of climate change on soil erosion control is one of the main consequences of climate change on ecosystem services, as it increases the risk of soil erosion. Soil erosion control is a critical ecosystem service that plays a key role in regulating agricultural sustainability, preventing land degradation, and ensuring food security. However, it is highly influenced by climatic and environmental changes. This study aims to predict the effects of climatic and environmental variables on soil erosion control ecosystem services using an artificial neural network model in Asia. Two datasets were collected: soil erosion control data derived from satellite imagery and climatic and environmental variables, including DEM, Normalized Difference Vegetation Index, Slope, Air Temperature, Land Surface Temperature, Soil Moisture, Temperature Condition Index, Vegetation Condition Index, Vegetation Health Index, and Precipitation. The findings showed that the R 2 for the prediction model is 0.98, suggesting that artificial neural networks can predict 98% of the variations in soil erosion control ecosystem service based on climatic and environmental variables. Sensitivity Analysis results revealed that soil moisture (at 70 cm depth) and vegetation health index significantly influence the model. This approach can highlight the potential of Machine Learning Algorithms for predicting various Ecosystem Services and acts as a future research blueprint. Its practical implications support the development of sustainable land management strategies in Asia. • The ANN model predicted 98% of SEC-ES changes using factors like soil moisture, VHI, and NDVI. • Soil moisture at 70 cm depth had the highest impact on the model, while air temperature had the least. • Machine learning in predicting ESs provides an effective approach to sustainable land management in erosion-prone areas.