Ocean digital twin technology integrates achievements from ocean observation, artificial intelligence, advanced simulation, and supercomputing to achieve a real, high-resolution, multi-dimensional, and interactive real-time virtual representation of the ocean. By digitally creating virtual models of physical entities and using data to simulate their behavior in the real environment, digital twins enhance physical entities' capabilities through interactive feedback between the virtual and physical worlds, data fusion analysis, and iterative optimization of decision-making. As a technology that integrates models, data, and intelligence, digital twins span the entire product lifecycle, connecting the physical world with the information world to provide more real-time, efficient, and intelligent services.

Ocean digital twins combine advancements in communications, artificial intelligence, computer graphics, and other fields to realize the mapping between ocean twins and physical ocean entities. By constructing marine virtual reality and fostering efficient dynamic interactions between humans and twins, this technology meets the needs of multi-level users and various service application scenarios, achieving the goals of "prophetic foresight" and "shared wisdom." Therefore, one key question is: how to construct a framework for ocean digital twin that integrates data, models, visualizations, and application services?

Data and models are the foundation of digital twins, and ocean data is characterized by its multi-source heterogeneity. Therefore, the accuracy of numerical simulation is essential for establishing digital twin oceans. Effectively utilizing existing ocean observation resources, combined with advanced numerical models, to optimally reconstruct multi-space-time scale and multi-dimensional information of the ocean, and to accurately map the complex physical entity of the ocean in the digital space, is the fundamental requirement for constructing digital twin oceans. Therefore, another key question is: how to build efficient  ocean datalake and intelligent numerical models?