Distribution of Surface Dissolved Oxygen Fronts in China's Marginal Seas

Currently, studies on the spatiotemporal distribution of dissolved oxygen (DO) fronts and their relationships with physical fronts remain limited. This study aims to investigate the spatiotemporal characteristics of DO fronts in China's marginal seas. DO values were first estimated using machine learning (ML) algorithms. Subsequently, DO fronts were identified by integrating gradient‐based and histogram‐based methods. Then, front frequency and average gradient were computed and integrated to analyze the spatiotemporal characteristics of DO fronts. Furthermore, the coupling probabilities between coastal DO fronts and the corresponding fronts of sea surface temperature (SST), sea surface salinity (SSS), and sea surface chlorophyll‐a (SCHL), as well as the correlations in terms of front length and offshore distance, were investigated. Generally, the retrieved root mean square errors and mean absolute percentage error values were mostly below 8 μmol/kg and 2.1%, respectively, across various ML methods, while the correlation coefficients consistently exceeded 0.98. Seasonal analysis revealed that DO fronts were relatively well‐structured in spring. In summer, except for the Korean coastal DO fronts and the Taiwan Strait front, most DO fronts exhibited reduced intensity and stability compared to spring. DO fronts remained weak in both intensity and stability throughout autumn. However, in winter, DO fronts became more well‐defined, with most reaching peak intensity and stability. Additionally, coastal DO front lengths correlated more strongly with SST front lengths than with SSS and SCHL front lengths. Moreover, the offshore distances of DO fronts were mainly influenced by SST fronts during spring and winter, while SCHL fronts had a slightly greater influence in summer and autumn.