Artificial intelligence for restoring land: A review of land degradation mapping, climate change prediction and ecosystem service valuation

Land degradation, climate change, and declining ecosystem services are increasingly interconnected challenges that threaten ecological integrity, agricultural productivity, and livelihood security across vulnerable landscapes. At the same time, advances in artificial intelligence (AI), remote sensing, geospatial analytics, and large environmental datasets are reshaping the way these challenges can be assessed, predicted, and managed. This review synthesises peer-reviewed literature published between 2010 and March 2026 on AI-based applications across three closely linked domains: land degradation mapping, climate change prediction, and ecosystem service valuation. It examines the use of machine learning, deep learning, hybrid geospatial models, and multi-source data integration for detecting degradation hotspots, assessing desertification and vegetation stress, forecasting climate-linked risks such as drought, heat stress, and fire, and estimating ecosystem services, including carbon sequestration, soil retention, water regulation, and forage productivity. The reviewed literature shows that AI-based approaches increasingly combine long-term satellite archives, climate datasets, terrain variables, soil attributes, field observations, and socio-environmental data; for example, some studies have used multi-temporal Landsat records of up to 35 years for soil degradation assessment, while others apply AI models to drought, evapotranspiration, wildfire risk, and ecosystem service mapping. The review further evaluates the strengths and limitations of existing approaches in terms of scale, data dependency, transferability, uncertainty, and interpretability. It concludes that AI can substantially strengthen restoration science by linking degradation diagnosis, climate-risk prediction, and ecosystem service assessment within a unified decision-support framework. Therefore, AI should be viewed not merely as a technical tool for environmental analysis but as a strategic enabler for restoration prioritisation, ecosystem resilience, and evidence-based environmental governance in climate-vulnerable landscapes.