I have encountered several ways to detect ocean eddies: Okubo-Weiss, Vector Geometry, winding angle, from SST fronts, etc. However I don't know which one to use. Does a method differ in detecting eddies with latitude or region? I also find that validation of some methods require experts looking at the SLA and geostrophic velocity maps but it has also high uncertainties.
Regardless of the method, an important thing is to consider the methods uncertainty in your results. Indicators such as Success of Detection Rate (SDR, identifying those eddies that exist), and Excess of Detection Rate (EDR, identifying as an eddy something that is not) should be part of your results. Some methods publish those (or others) indicators in their description, others don't. For an example, see Fig. 2 of Chaigneau et al (2008), where their method has a similar SDR as using Okubo-Weiss method (~90%), but much lower (but still ~20%) EDR.
My suggestion is to use a method of your choice and then do a manual identification of 10-20 sample images, to validate the code in your results by visually identifying the eddies. The more colleagues that can help your with this, the better.
REF: Chaigneau, A., Gizolme, A., & Grados, C. (2008). Mesoscale eddies off Peru in altimeter records: Identification algorithms and eddy spatio-temporal patterns. Progress in Oceanography, 79(2-4), 106-119.
- Badges
- Science topic
- Similar topics
- Geoscience
- Remote Sensing