Optimization of Subsurface Eddy Detection with Potential Vorticity from Models in the Arabian Sea

Coherent mesoscale vortices, or eddies, are ubiquitous throughout the world's oceans, including the Arabian Sea. Eddies are typically tracked at the ocean's surface, most often using Sea Surface Height (SSH) or the Okubo Weiss parameter. In this work, we develop a series of algorithms specifically for the detection of subsurface eddies in the Arabian Sea using the outputs of Nucleus for European Modelling of the Ocean (NEMO) model simulations. We achieve this by optimizing each parameter to determine the subsurface eddy detection algorithm that best identifies the vertical extent of an SSH-defined eddy at the surface. The best performing algorithms in this analysis are those that utilize the rescaled isopycnal potential vorticity (PV), which has not been used for automated eddy tracking to this point. We demonstrate that this novel type of algorithm successfully detects eddies that exist in water masses bounded by specific isopycnals, such as Red Sea Water (RSW). These subsurface eddies are categorized in two ways: those visible at the surface and those isolated within the RSW isopycnal layer. We find that 80% of RSW eddies are not visible at the surface. We then display each type of eddies' properties and trajectories, computing subsurface eddy volume transports of RSW into and out of the Gulf of Aden for each monsoon season in weak (2016), normal (2017), and strong (2019) years. We find that while subsurface cyclonic eddy transports are responsible for around 8% of total volume transport at depth into the Gulf of Aden during the winter monsoons, the total volume transports of RSW out of the Gulf of Aden by eddies is otherwise minimal, and transport via anticyclonic eddies is absent. Finally, we discern the three-dimensional structures and dynamics of a specific long-lived, completely subsurface RSW eddy using Lagrangian particle tracking to diagnose the evolution of the PV anomalies that led to its formation. This analysis is made possible by the conserved nature of PV, allowing our new algorithm to investigate these types of eddies and their effects on subsurface water masses more efficiently than any previously developed eddy tracking algorithm.