Contours about py-eddy-tracker

[Ashwitaa]

Hi Antoine :)

I did some mistakes earlier and that's why I had different results with different ranges…
thank you :)

I close this issue now…

[AntSimi]

Hi,
In amplitude step you also check if there only one extrema or more if you allow more.

For Okubo Weiss, i never used for eddy detection, only play a little with it:

• Did you use same step in you two range (a) and (b) ?
• How detection are different, which version are better?
• When you compare detection with OW field did you see something incoherent?
• How did you implement eddy detection with OW?

Antoine

[Ashwitaa]

1. yes, I used the same step for both (a) and (b)
   I used the step value of 4e-12
   where, (a) -4e-09 : -4e-12 has 1002 levels, and (b) -4e-10 : -4e-12 has 102 levels

2. It is hard to say which version is better.
   the detection is different in terms of count in various regions (almost everywhere in the ocean)
   I performed numerous tests in different regions and with different time steps to know at which instance what is detected and what is missing with the other. But this doesn't answer me 'why' there are differences as what actual criterion is helping for number of detections.

3. no, I think I don't see anything incoherent
   the structures are fine.
   but why the detections are different from one range to the another I am unable to understand that.

4. In order to implement the eddy detection with OW
   we just asked the algorithm to use the contours of OW instead of SSH

[AntSimi]

Could you share a snapshot of two different version to illustrate difference?