Advances in data assimilation (DA) methods have greatly improved the accuracy of Earth system predictions. To fuse multi-source data and reconstruct the nonlinear evolution missing from observations, geoscientists are developing future-oriented DA methods. In this paper, we redesign a purely data-driven latent space DA framework (DeepDA) that employs a generative artificial intelligence model to capture the nonlinear evolution in sea surface temperature. Under variational constraints, DeepDA embedded with nonlinear features can effectively fuse heterogeneous data. The results show that DeepDA remains highly stable in capturing and generating nonlinear evolutions even when a large amount of observational information is missing. It can be found that when only 10% of the observation information is available, the error increase of DeepDA does not exceed 40%. Furthermore, DeepDA has been shown to be robust in the fusion of real observations and ensemble simulations. In particular, this paper provides a mechanism analysis of the nonlinear evolution generated by DeepDA from the perspective of physical patterns, which reveals the inherent explainability of our DL model in capturing multi-scale ocean signals.

Using MODIS-Aqua satellite observations, this study analyzes the spatiotemporal distribution characteristics of particulate organic carbon (POC) in China’s marginal seas from 2003 to 2024. The statistical relationships between various marine environmental variables, including sea surface temperature (SST), nutrients, and primary production (PP), and POC concentrations are explored using partial least squares path modeling (PLS-PM). Finally, a box model approach is conducted to assess the POC budget in the study area. The results indicate that the POC concentration in the marginal seas of China generally exhibits a characteristic of being high in spring and low in summer. The highest concentration of POC is observed in the Bohai Sea, followed by the Yellow Sea, and the lowest in the East China Sea, with coastal waters exhibiting higher POC concentrations compared to the central areas. The spatial distribution and seasonal changes in POC are jointly influenced by PP, water mass exchange, resuspended sediments, and terrestrial inputs. Large-scale climate modes show statistical associations with POC concentration in the open waters of China’s marginal seas. PP and respiratory consumption are identified as the predominant input and output fluxes, respectively, in China’s marginal seas. This study enriches the understanding of carbon cycling processes and carbon sink mechanisms in marginal seas.