Lagrangian geography of the deep Gulf of Mexico

TL;DR

This study maps the deep Gulf of Mexico's Lagrangian circulation using float trajectories, revealing distinct provinces and connectivity patterns that differ from surface dynamics, with implications for pollutant dispersal and basin ventilation.

Contribution

It introduces a Markov-chain based Lagrangian geography of the deep Gulf of Mexico, highlighting deep circulation provinces and their connectivity, supported by multiple independent data sources.

Findings

Main provinces have ~4.5 and 3.5 years residence times.

Deep Lagrangian geography differs from surface patterns.

Ventilation conduit identified in southeastern Gulf.

Abstract

Using trajectories from acoustically tracked (RAFOS) floats in the Gulf of Mexico, we construct a geography of its Lagrangian circulation within the 1500--2500-m layer. This is done by building a Markov-chain representation of the Lagrangian dynamics. The geography is composed of weakly interacting provinces that constrain the connectivity at depth. The main geography includes two provinces of near equal areas and separated by a roughly meridional boundary. The residence time is about 4.5 (3.5) years in the western (eastern) province. The exchange between these provinces is effected through a slow cyclonic circulation, which is well constrained in the western basin by preservation of $f/H$, where $f$ is the Coriolis parameter and $H$ is depth. Secondary provinces of varied shapes covering smaller areas are identified with residence times ranging from about 0.4 to 1.2 years or so. Except for the main provinces, the deep Lagrangian geography does not resemble the surface Lagrangian geography recently inferred from satellite-tracked drifter trajectories. This implies disparate connectivity characteristics with potential implications for pollutant (e.g., oil) dispersal at the surface and depth. A ventilation conduit through the southeastern corner of the domain from the Caribbean Sea balanced by weak vertical exchange is also inferred. This is supported by the inspection of satellite-tracked profiling (Argo) floats, which, while forming a smaller dataset and having seemingly different water-following characteristics than the RAFOS floats, replicate the main aspects of the Lagrangian geography. Finally, consistency with independent results from a chemical tracer release experiment is found to provide additional support to the results.