Content area
Full text
ABSTRACT
This paper presents a software tool that enables the identification and automated tracking of oceanic eddies observed with satellite altimetry in user-specified regions throughout the global ocean. As input, the code requires sequential maps of sea level anomalies such as those provided by Archiving, Validation, and Interpretation of Satellite Oceanographic (AVISO) data. Outputs take the form of (i) data files containing eddy properties, including position, radius, amplitude, and azimuthal (geostrophic) speed; and (ii) sequential image maps showing sea surface height maps with active eddy centers and tracks overlaid. The results given are from a demonstration in the Canary Basin region of the northeast Atlantic and are comparable with a published global eddy track database. Some discrepancies between the two datasets include eddy radius magnitude, and the distributions of eddy births and deaths. The discrepancies may be related to differences in the eddy identification methods, and also possibly to differences in the smoothing of the sea surface height maps. The code is written in Python and is made freely available under a GNU license (http://www.imedea.uib.es/users/ emason/py-eddy-tracker/).
1. Introduction
Satellite altimetry has revealed the ubiquity of mesoscale eddies in the global ocean (e.g., Stammer 1997, 1998). Eddies range greatly in shape and size, are often asymmetric, and can have highly variable translational and rotational velocities (McWilliams 2008; Chelton et al. 2011b, hereafter CSS11; Early et al. 2011). Interest in mesoscale eddies arises from their role in the dynamics of the large-scale oceanic circulation; eddies are efficient carriers of mass and its physical, chemical, and biological properties, such that their presence modulates fluxes of heat and momentum and the dynamics of marine ecosystems (Chelton et al. 2011a; Gruber et al. 2011; Stramma et al. 2013).
Recent years have seen the emergence of several automated oceanic eddy tracking algorithms that contribute to knowledge of eddy properties and their variability. The techniques comprise threemainmethods: geometric (e.g., Chaigneau et al. 2008; Nencioli et al. 2010; CSS11); Okubo-Weiss (e.g., Isern-Fontanet et al. 2003; Morrow et al. 2004; Chelton et al. 2007; Ubelmann and Fu 2011); and wavelet (e.g., Doglioli et al. 2007; Rubio et al. 2009); and a comparative analysis of these approaches has been made by Souza et al. (2011). Novel techniques falling outside these methods are a hybrid...