Tracking with area_tracker

[chloe-gt]

Hi Antoine,

I'm trying to trace eddies trajectories in my study area with area_tracker.
At first, I run the identification on the ADT with the eddy_identification function.
Every day, the identification results are recorded in 2 netcdf: 1 netcdf for anticyclones and 1 netcdf for cyclones. Then I configured a yaml file. Here it is:
fichier.yaml.zip
Finally, I modified the py_eddy_tracker.poly.vertice_overlap function: hybrid_area=True, min_overlap=0,2
When I run EddyTracking conf.yaml -v DEBUG the following error is displayed:

Please, can you help me correct this error ?

Thanks for your help

[AntSimi]

I think problem come from your yaml settings, you need one space after ':' to separate value from key

[chloe-gt]

Ok thanks it was a syntax problem. Now when I run EddyTracking file.yaml -v DEBUG there is no error :

But no files are saved on SAVE_DIR. Do you think there is another mistake somewhere ?

[AntSimi]

No file, because no tracking :), before to do tracking software check time delta between each detection. In your case it found 2 time step 86400s => 1day and 172800s => 2days. I guess there some detection missing

[chloe-gt]

Thank you very much!

Last question, can you explain to me what is TRACK_DURATION_MIN ? Why do we have to distinguish between long and short trajectory ?

[chloe-gt]

Hello Antoine,

I'm a bit confused about the Area_tracker method.
What is the difference between min_overlap in vertice_overlap and cmin?
What is the overlap rate at which two vortices merge?

Finally, does the overlap between two eddies use the effective contour or the speed contour ?
I haven't found the answer to this question in the code

Thanks

[AntSimi]

Thank you very much!

Last question, can you explain to me what is TRACK_DURATION_MIN ? Why do we have to distinguish between long and short trajectory ?

It's an historical parameter to divide dataset long and short trajectory, short trajectory (less 5/10 days) could be spurious detection .

[AntSimi]

Hello Antoine,

I'm a bit confused about the Area_tracker method. What is the difference between min_overlap in vertice_overlap and cmin?

cmin and min_overlap are different only if you use method hybrid_area vertice_overlap code

What is the overlap rate at which two vortices merge?

I am not sur to understand this question. By default AreaTracker use 20% (cmin parameter) of overlap to associate consecutive observations. But classic tracking doesn't manage splitting/merging events.

Finally, does the overlap between two eddies use the effective contour or the speed contour ? I haven't found the answer to this question in the code

AreaTracker use match function, you could see parameters ìntern=False, so it will use outter/effective contour.