AI for Planetary-Scale Simulation and Decision Intelligence Transition Toward Earth Digital Twins

Earth Digital Twins (EDTs) are emerging as a transformative paradigm to address the growing need for real-time, data-driven understanding of complex Earth systems. This review investigates how advances in artificial intelligence (AI) and Big Data are enabling the transition from static Earth system models to dynamic, interactive platforms capable of simulating “what-if” scenarios for climate, environmental change, and socio-economic systems. By integrating heterogeneous data streams from satellites, IoT networks, and physics-based models, EDTs aim to provide actionable insights for planetary-scale decision-making. This study systematically analyzes the recent literature on EDT architectures and AI-driven methodologies, focusing on core components such as data assimilation frameworks, high-performance computing infrastructures, and modular digital ecosystems. Particular attention is given to AI techniques including deep learning for numerical weather prediction, multimodal data fusion, reinforcement learning for sequential decision processes, and causal inference for human–environment system modeling. The review shows that AI significantly enhances the capabilities of Earth system simulations by enabling fast global-scale predictions, improved uncertainty-aware data assimilation, and adaptive decision-support systems. Emerging EDT frameworks demonstrate the ability to couple environmental processes with socio-economic dynamics, supporting applications such as climate risk assessment, infrastructure optimization, and disaster preparedness. These systems increasingly function as integrated decision-intelligence platforms rather than purely scientific simulation tools. We conclude that Earth Digital Twins represent a critical shift toward AI-driven planetary intelligence, but their realization depends on overcoming challenges related to data interoperability, computational scalability, uncertainty quantification, and ethical governance. Future research should prioritize the development of interoperable architectures, human-centered modeling approaches, and transparent decision frameworks to ensure that EDTs support robust, inclusive, and responsible Earth system management.