Issue with tracking about py-eddy-tracker

[mealissa]

Problem solved, Many thanks to you for your assistance.

[AntSimi]

Hi,
May i ask you more information to help you?
Could you share your conf.yaml.
How many detection file have you?
Antoine

[mealissa]

Hi,
Yes of course

PATHS:
  FILES_PATTERN: /home/alissa/home2/TSTECMWF/EULER/ecmwf/out/Anticyclonic_*.nc
  SAVE_DIR: /home/alissa/home2/TSTECMWF/EULER/ecmwf/track

VIRTUAL_LENGTH_MAX: 0
TRACK_DURATION_MIN: 30

I do have 365 detection files. One of the messages I just paid attention to is:

INFO 2020-10-02 10:25:29,003 tracking.track :
                                        b'/home/alissa/home2/TSTECMWF/EULER/ecmwf/out/Anticyclonic_20060110.nc' match with previous state
        DEBUG 2020-10-02 10:25:29,003   tracking.       track :
                                                2286 obs to match
        DEBUG 2020-10-02 10:25:29,360   observation.    solve_simultaneous :
                                                Shape conflict matrix : (151, 151), 329 conflicts
        WARNING 2020-10-02 10:25:29,360 observation.solve_simultaneous :
                                        High number of conflict : 302 (nb_conflict)
        DEBUG 2020-10-02 10:25:29,381   observation.    solve_simultaneous :

[AntSimi]

Hi

Conflicts : when an eddy can be associated to several other
It's normal to have some conflicts, and i don't think it's linked with your problem.

Lot of questions:

1. Did you look at some identification file? you could do with this, maybe display 2 or 3 consecutive identification
2. With which comand did you run identification?
3. Do you have same problem with cyclonic?
4. Do you have same problem with shorter period? (In order to create a minimal example)
5. Is it a global identification?

You error look like if you have no shorter eddy but it's strange (for 30 days).

Antoine

[mealissa]

1. Did you look at some identification file? you could do with this, maybe display 2 or 3 consecutive identification

I will do so and be back to you.

2. With which comand did you run identification?

EddyId path_to_nc_file ssh ugos vgos longitude latitude ./

3. Do you have same problem with cyclonic?

yes same issue with both.

4. Do you have same problem with shorter period? (In order to create a minimal example)

same issue with shorter or longer

5. Is it a global identification?

yes global identification from Model data (surface)

Thank you for your quick answer

[AntSimi]

Is it a daily input file of identification? or you have only one file with all time step?
Could you share ncdump -h of one input file?

[mealissa]

Here are some of the plots (2006-01-05, 2006-02-01, 2006-03-01). I think I get the point. they look all alike. maybe all the identifications are coming from the first day.

yes, I do have one file with all time steps. I will probably have to seperate the "one" file into multi-files and run again.

here is the ncdump output:

ncdump -h ecmwf.nc 
netcdf ecmwf {
dimensions:
        time_counter = UNLIMITED ; // (361 currently)
        lon = 1440 ;
        lat = 720 ;
variables:
        double time_counter(time_counter) ;
                time_counter:standard_name = "time" ;
                time_counter:long_name = "Time axis" ;
                time_counter:units = "seconds since 2006-01-03 00:00:00" ;
                time_counter:calendar = "gregorian" ;
                time_counter:axis = "T" ;
        double lon(lon) ;
                lon:standard_name = "longitude" ;
                lon:long_name = "longitude" ;
                lon:units = "degrees_east" ;
                lon:axis = "X" ;
        double lat(lat) ;
                lat:standard_name = "latitude" ;
                lat:long_name = "latitude" ;
                lat:units = "degrees_north" ;
                lat:axis = "Y" ;
        float sst(time_counter, lat, lon) ;
                sst:standard_name = "Sea Surface temperature" ;
                sst:long_name = "Sea Surface temperature" ;
                sst:units = "C" ;
                sst:_FillValue = -9999.f ;
                sst:missing_value = -9999.f ;
                sst:interval_operation = 1200.f ;
                sst:interval_write = 432000.f ;
                sst:online_operation = "ave(x)" ;
        float ssh(time_counter, lat, lon) ;
                ssh:standard_name = "Sea Surface Height" ;
                ssh:long_name = "Sea Surface Height" ;
                ssh:units = "m" ;
                ssh:_FillValue = -9999.f ;
                ssh:missing_value = -9999.f ;
                ssh:interval_operation = 1200.f ;
                ssh:interval_write = 432000.f ;
                ssh:online_operation = "ave(x)" ;
        float ugos(time_counter, lat, lon) ;
                ugos:standard_name = "Surface zonal Current" ;
                ugos:long_name = "Surface zonal Current" ;
                ugos:units = "m/s" ;
                ugos:_FillValue = -9999.f ;
                ugos:missing_value = -9999.f ;
                ugos:interval_operation = 1200.f ;
                ugos:interval_write = 432000.f ;
                ugos:online_operation = "ave(x)" ;
        float vgos(time_counter, lat, lon) ;
                vgos:standard_name = "Surface meridional Current" ;
                vgos:long_name = "Surface meridional Current" ;
                vgos:units = "m/s" ;
                vgos:_FillValue = -9999.f ;
                vgos:missing_value = -9999.f ;

[AntSimi]

Ok for solution take look at issues #20 and reason are explain in #10 (i forgeted this error )