Ocean mixed layer depth 2000–2020: Estimation assessment and long-term trends

Accurate estimation of mixed layer depth (MLD) is crucial for understanding the upper ocean dynamics as well as their implications on climate and marine ecosystems. While various methods have been proposed to estimate MLD from vertical temperature and/or density profiles, a systematic evaluation across different algorithms remains untapped. In this study, twelve approaches for MLD estimation are implemented and compared based on the Argo profiles collected from 2000 to 2020. Each method is assessed based on the quality index (QI), which measures the uniformity of temperature or density profiles within the boundaries of the estimated MLD. We found the two methods of linear fitting and relative variance exhibit similar QI levels higher than the others. A ranking scheme for that the method with the highest QI for a given profile earns one score is then designed and applied. The linear fitting one obtains the highest and is thus recommended as the most reliable method for geophysical analysis. The global MLD patterns derived using this method are analyzed with extremes being predominantly found in mid-latitudes due to strong seasonal variations. Significant long-term changes are observed, particularly in the southern hemisphere during summer, where the MLD shows a deepening trend over the past 20 years. Results presented here shall provide valuable insights for choosing the most reliable MLD estimation methods and understanding the environmental drivers of MLD variations.