A Vector Geometry–Based Eddy Detection Algorithm and Its Application to a High-Resolution Numerical Model Product and High-Frequency Radar Surface Velocities in the Southern California Bight

Abstract

Automated eddy detection methods are fundamental tools to analyze eddy activity from the large datasets derived from satellite measurements and numerical model simulations. Existing methods are either based on the distribution of physical parameters usually computed from velocity derivatives or on the geometry of velocity streamlines around minima or maxima of sea level anomaly. A new algorithm was developed based exclusively on the geometry of the velocity vectors. Four constraints characterizing the spatial distribution of the velocity vectors around eddy centers were derived from the general features associated with velocity fields in the presence of eddies. The grid points in the domain for which these four constraints are satisfied are detected as eddy centers. Eddy sizes are computed from closed contours of the streamfunction field, and eddy tracks are retrieved by comparing the distribution of eddy centers at successive time steps. The results were validated against manually derived eddy fields. Two parameters in the algorithm can be modified by the users to optimize its performance. The algorithm is applied to both a high-resolution model product and high-frequency radar surface velocity fields in the Southern California Bight.