Digital twin integration with generative AI and foundation models for real-time precision agriculture and crop resilience

Digital Twins (DTs) are emerging as a foundational technology in precision agriculture, enabling real-time virtualization, prediction, and autonomous optimization of crop systems. This review synthesizes current advances in DT applications across three major frontiers: precision nutrient management, dynamic water use efficiency, and proactive stress and resilience management. We highlight how hybrid modelling frameworks combining mechanistic crop models, machine learning, deep reinforcement learning, and high-throughput phenotyping enhance predictive capability and operational decision-making. The paper also examines the critical roles of Explainable AI, federated learning, and edge cloud architectures in improving transparency, data sovereignty, and scalability. Persistent challenges remain in data standardization, interoperability, rural digital infrastructure, and workforce capacity. Future progress depends on integrating foundation models, biologically constrained simulations, and privacy-preserving learning protocols into next-generation DT systems. Together, these developments position DTs as a transformative engine for sustainable, efficient, and climate-resilient agriculture.