jingwei-sjtu/OceanVerse

OceanVerse数据集是一个综合性的数据集，旨在解决稀疏海洋观测数据重建的挑战。它集成了自1900年以来的近200万个实测海洋剖面数据点和三组地球系统数值模拟数据。该数据集提供了一个新颖的大型数据集（相比现有数据集大约100×节点），满足MNAR（Missing Not at Random）条件，支持更有效的模型比较、泛化评估和潜在的科学研究架构的进步。数据集和代码库公开发布，以促进AI在海洋领域的研究与发展。

OceanVerse is a comprehensive dataset designed to address the challenge of reconstructing sparse ocean observation data. It integrates nearly 2 million real-world profile data points since 1900 and three sets of Earth system numerical simulation data. The dataset provides a novel large-scale (∼100× nodes vs. existing datasets) that meets the MNAR (Missing Not at Random) condition, supporting more effective model comparison, generalization evaluation, and potential advancement of scientific reconstruction architectures. The dataset and codebase are publicly available to promote further research and development in the field of AI4Ocean.