Impact of submesoscales on surface material distribution in a gulf of Mexico mesoscale eddy

Abstract

Understanding material distribution at the ocean’s surface is important for a number of applications, in particular for buoyant pollutants such as oil spills. The main tools to estimate surface flows are satellite altimeters, as well as data-assimilative ocean general circulation models (OGCMs). Current-generation altimeter products rely on the geostrophic approximation to derive surface currents. Recent modeling and experimental work revealed existence of ageostrophic submesoscale motions within the upper ocean boundary layer.The next frontier is how submesoscales influence transport pathways in the upper ocean, which is a multi-scale problem involving the interaction of submesoscale and mesoscale coherent structures. Here we focus on a mesoscale eddy that exhibits submesoscale fluctuations along its rim. The high-resolution OCGM fields are then treated with two filters. A Lanczos filter is applied to velocity fields to remove the kinetic energy over the submesoscales. Then a Gaussian filter is used for the modeled sea surface height to simulate a geostrophic velocity field that would be available from gridded satellite altimeter data. Lagrangian Coherent Structures (LCS) are then generated from full-resolution and filtered fields to compare Lagrangian characteristics corresponding to different realizations of the surface velocity fields.It is found that while mesoscale currents exert a general control over the pathways of the tracer initially launched in the mesoscale eddy, there is a leak across the mesoscale transport barriers, induced by submesoscale motions. This leak is quantified as 20% of the tracer when using the submesoscale filter over one month of advection, while it increases to 50% using the geostrophic velocity field. We conclude that LCS computed from mesoscale surface velocity fields can be considered as a good first-order proxy, but the leakage of material across them in the presence of submesoscales can be significant.