Formation and Transport of the South Atlantic Subtropical Mode Water in Eddy‐Permitting Observations

We present in this study a detecting algorithm that enables to identify both the ocean surface mixed layer depth (MLD) and mode water from the Argo profiling data. This algorithm proposes a detection based on the calculation of the first and second derivatives for each temperature (or density) profile. Low gradients reveal segments of water mass homogeneity, while extreme values of second derivatives indicate the precise depths at which the gradients sharply change. Specifically applied to detect the South Atlantic Subtropical Mode Water (SASTMW), this algorithm leads to a redefinition of the three mode water types discussed in Sato and Polito (2014; https://doi.org/10.1002/2013JC009438) as well as new insights into the origins of these water masses and their dynamics. In particular, we point out that only one of the SASTMW varieties originates from the Brazil‐Malvinas Confluence in the western boundary, whereas the other two are related to the Agulhas Leakage that shapes the Indo‐Atlantic water mass exchanges. As both regions of the SASTMW formation are characterized by intense eddy kinetic energy, the role of mesoscale eddies in the SASTMW formation and transport is also investigated by co‐locating Argo profiles with eddies objectively identified from satellite altimetry maps. It suggests that anticyclonic eddies correspond to a larger number and thicker layers of SASTMW than their cyclonic counterparts. In the Cape Basin, where mode waters are found correlated with the main paths of Agulhas Rings, a potential route of subduction is also proposed. Plain Language Summary Mode waters are essential components in the upper ocean, of which the spatial and temporal variability significantly contributes to the large‐scale climate dynamics. This contribution can be inferred from the weakly stratified characteristics of mode waters that are formed due to winter surface cooling. Regarding the homogeneity of both surface mixed layers and mode waters, we developed a general statistical algorithm to detect their presence from different observational profiling platforms. Argo floats are one of these platforms that drift up and down to retrieve temperature and salinity profiles. Therefore in this study, one attempt is to apply the detecting algorithm to Argo profiles in the South Atlantic subtropics. The novelty of this diagnostic algorithm lies in a precise inspection that results from the searching of variations in each individual profile, instead of implementing uni