کاربردهای هوش مصنوعی در سازگاری و کاهش تغییرات اقلیمی: مرور نظام‌مند چالش‌ها، شکاف‌های پژوهشی و مسیرهای آینده

Abstract
Objective: Climate change is one of the most pressing challenges of the twenty-first century, requiring innovative and multidisciplinary solutions. This study aims to provide a comprehensive systematic review of the role of artificial intelligence (AI) in addressing climate change, with a particular focus on its practical applications in four key domains: renewable energy, climate prediction, smart agriculture, and natural resource management.
Method: This study was conducted as a systematic literature review following the PRISMA guidelines. A structured search was conducted across major scientific databases, including ScienceDirect, IEEE Xplore, Springer, Frontiers, Nature Climate Change, and official reports from the Intergovernmental Panel on Climate Change (IPCC), covering the period from 2020 to 2026. From an initial pool of 487 publications, 72 studies were selected for final analysis based on predefined criteria of relevance, methodological quality, and publication recency. The extracted data were synthesized and analyzed according to five thematic categories.
Results: The findings indicate that AI has demonstrated significant potential in optimizing renewable energy systems, accurately forecasting extreme climate events, reducing agricultural inputs (up to 30% less fertilizer, 40% lower pesticide use, and 35% lower irrigation water consumption), enabling intelligent natural resource management, and improving carbon stock estimation in ecosystems. Furthermore, advanced deep learning models, such as FourCastNet and GraphCast, have achieved highly accurate weather forecasting while requiring substantially lower computational costs than conventional numerical forecasting models.
Conclusion: The review highlights several strategic priorities for advancing AI-driven climate action, including the development of region-specific AI solutions, increased investment in mitigation-oriented applications, the establishment of ethical and regulatory frameworks to enhance algorithmic transparency, the promotion of physics-informed and data-driven hybrid approaches, stronger interdisciplinary collaboration, and greater consideration of the carbon footprint associated with AI technologies.
 
Extended Abstract
Introduction
Climate change is one of the most significant environmental, economic, and social challenges of the twenty-first century. Rising greenhouse gas emissions have increased global temperatures, intensified extreme weather events, and threatened ecosystems, agriculture, water resources, and human well-being. Addressing these challenges requires innovative solutions that support both climate change adaptation and mitigation.
Artificial Intelligence (AI) has emerged as a promising technology for climate action. Through machine learning, deep learning, computer vision, and optimization techniques, AI enhances climate prediction, precision agriculture, renewable energy management, environmental monitoring, and resource optimization. Recent models, such as GraphCast and FourCastNet, have improved weather forecasting while reducing computational costs. AI has also demonstrated strong potential in smart grids, carbon monitoring, and sustainable agricultural practices.
Despite these advances, several barriers limit the widespread adoption of AI for climate applications. Key challenges include limited data availability and quality, lack of model transparency, algorithmic bias, high energy consumption, implementation costs, insufficient regulatory frameworks, and regional disparities. Moreover, existing research focuses predominantly on climate adaptation, with comparatively less attention given to mitigation.
Given the rapid growth of AI research in climate science, a comprehensive synthesis of current knowledge is needed. Therefore, this systematic review examines AI applications in climate change adaptation and mitigation, identifies major challenges and research gaps, and outlines future directions for supporting sustainable climate action.
 
Method
This study employed a systematic literature review following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) framework. A comprehensive search was conducted using major scientific databases, including ScienceDirect, IEEE Xplore, SpringerLink, Frontiers, Nature Climate Change, and official reports published by the Intergovernmental Panel on Climate Change (IPCC).
The search covered publications published between 2020 and 2025 using combinations of keywords related to Artificial Intelligence, Machine Learning, Climate Change, Climate Adaptation, Climate Mitigation, Renewable Energy, Precision Agriculture, and Environmental Management. After removing duplicate records and screening titles and abstracts, studies were evaluated according to predefined inclusion criteria, including scientific quality, relevance to AI applications in climate change, methodological rigor, and publication recency.
From an initial collection of 487 publications, 72 high-quality studies were selected for detailed analysis. Extracted information included AI techniques, application domains, geographical distribution, implementation challenges, reported outcomes, and research limitations. The selected studies were then categorized into five major thematic areas to facilitate qualitative synthesis and comparative analysis.
 
Results
The reviewed studies demonstrate that Artificial Intelligence has become an essential enabling technology for both climate change adaptation and mitigation across multiple sectors.
One of the most mature application areas is renewable energy management. Machine learning algorithms significantly improve the forecasting of solar and wind energy production, optimize energy storage systems, and enhance smart grid operation. These improvements increase renewable energy integration while reducing operational costs and carbon emissions.
Climate prediction represents another rapidly advancing field. Deep learning models such as GraphCast and FourCastNet have achieved forecasting accuracy comparable to or better than conventional numerical weather prediction systems while requiring substantially lower computational resources. Improved forecasting enhances disaster preparedness and strengthens climate resilience.
Precision agriculture has also benefited considerably from AI technologies. Intelligent irrigation systems, crop monitoring, pest detection, and fertilizer optimization have reduced fertilizer consumption by approximately 30%, pesticide use by nearly 40%, and irrigation water demand by about 35% in several reported case studies. These improvements contribute simultaneously to increased agricultural productivity and reduced environmental impacts.
AI has further demonstrated substantial potential in ecosystem monitoring, carbon accounting, forest management, biodiversity conservation, and environmental risk assessment through satellite image analysis, remote sensing, and automated environmental monitoring systems.
Nevertheless, the review identifies several critical barriers limiting large-scale implementation. The most frequently reported challenges include limited availability of high-quality data, concerns regarding data privacy and ownership, lack of model transparency, algorithmic bias, substantial computational energy consumption, high implementation costs, limited technical expertise, insufficient interdisciplinary collaboration, and the absence of comprehensive regulatory and ethical frameworks. The analysis also reveals an imbalance in current research priorities. Approximately two-thirds of published studies focus primarily on climate adaptation, whereas only a relatively small proportion explicitly addresses AI applications for climate change mitigation. In addition, geographical analysis indicates that Asia and global-scale studies dominate the literature, while Africa and South America remain significantly underrepresented despite their high vulnerability to climate change.
 
Conclusion
Artificial Intelligence is transforming climate science by providing advanced analytical capabilities that improve climate prediction, optimize renewable energy systems, support precision agriculture, and enhance natural resource management. The evidence synthesized in this review demonstrates that AI can contribute significantly to both climate adaptation and mitigation while improving decision-making across multiple sectors.
However, realizing the full potential of AI requires addressing several technical, economic, ethical, and institutional challenges. Improving data accessibility and quality, increasing model transparency and explainability, reducing the environmental footprint of AI systems, strengthening interdisciplinary collaboration, and establishing robust regulatory frameworks should be considered priority research and policy objectives. Future studies should place greater emphasis on AI applications that directly reduce greenhouse gas emissions, develop region-specific solutions for vulnerable countries, and promote equitable access to AI technologies. Addressing these priorities will enable Artificial Intelligence to become a more effective and sustainable tool for accelerating global climate action.