Artificial Intelligence for Climate Action: A Comprehensive Review of Data-Driven Methods, Applications, and Future Research Directions

With the capacity to swiftly and accurately analyse complex environmental data, AI has become a powerful tool in combating climate change, providing unprecedented predictive capabilities and helping decision support systems for both mitigation and adaptation. This review provides a systematic overview of the recent developments using AI methods for climate science, from machine learning to deep learning, optimization, and hybrid modelling techniques. We summarize recent progress across important directions, including extreme weather forecasting, emission monitoring, renewable energy optimization, disaster risk, and climate adaptation planning. Studies have shown that artificial intelligence provides significant improvements in prediction and operational efficiency when it comes to climate-related use cases. However, several challenges endure, including data scarcity in structured and unstructured domains (mostly), algorithmic bias leading to biased models, lack of model interpretability (for black-box assessments), and environmental footprints due to compute-intensive modelling. By synthesizing insights across more than 200 primary studies and high-impact reviews, we highlight pressing research gaps and offer a roadmap for responsible, scalable, and equitable AI deployment while addressing Sustainable Development Goal 13 (Climate Action) and more broadly sustainability goals.