issues with outputFreq : trajectories seem to increase when outputFreq increase

[Camille Mazoyer]

Dear all,

I am using CMS to compute trajectories of particles inside the Toulon bay (France) using oceanographic models. And I am experiencing an issue about the setting of the "outputFreq" parameter in the runconf.list file. Depending on its value, the output trajectories for the particles are drastically different for a same simulation with a constant duration.

When I increase the outputFreq value (without changing the duration of the simulation), the trajectory increase in the same time as if the particles "moved faster". I was expected that the outputFreq does not change the main trajectory.

Please, find attached some illutrations and below the parameters of the simulations.

Does anyone knows if I missed something somewhere ?

Thanks a lot for your help

Model used:

NEMO model

Output frequency : 1h

Horizontal resolution : 400m

(I also tested another model from the code MITgcm and I obtained the same issue)

Simulation :

Particles : 4 particles released

Start time : 4th of February 2018

Duration : 172800 seconds (=48h)

releaseFile :

1   6.2  42.9  0.6  1  2018  02  04  0
1   6.3  42.9  0.6  1  2018  02  04  0
1   6.4  42.9  0.6  1  2018  02  04  0
1   6.5  42.9  0.6  1  2018  02  04  0

runconf.list :

Fig1: outputFreq=3600

&runconf
!====================================================================!
nnests                    = 1
timeMax                 = 172800           !in seconds
timestep                 = 1800               !in seconds
outputFreq             = 3600               !in seconds
releaseFilename    = "releaseFile"
!====================================================================!

Fig2: outputFreq=7200

&runconf
!====================================================================!
nnests                    = 1
timeMax                 = 172800           !in seconds
timestep                 = 1800               !in seconds
outputFreq             = 7200               !in seconds
releaseFilename    = "releaseFile"
!====================================================================!

Fig3: outputFreq=14400

&runconf
!====================================================================!
nnests                    = 1
timeMax                 = 172800           !in seconds
timestep                 = 1800               !in seconds
outputFreq             = 14400             !in seconds
releaseFilename    = "releaseFile"
!====================================================================!

Regards,

[Nicolas Le Corre]

Hi Camille,

I've never seen that before. I did a quick test on my side and did not notice any problem. Could you please send the raw 3 traj_files from CMS.

Is it possible that you plotted the first X rows of each traj_file without considering the different timestep of each file?  

Cheers,

Nicolas

[Camille Mazoyer]

Hi Nicolas,

Thank you for your reply.

Here is a link with all my files for a MITgcm configuration: https://drive.google.com/open?id=1OQv_MDIBCNMqJh9IJqGsaQviL0Tg6tl6 ( the configuration above with NEMO output is not in my computer, I will ask my colleague)

I am checking my matlab file and I think I have a bug inside. I will confirm you this tomorrow morning (it's the night here). But I think you are right, I may have plotted wrong positions.

Thanks a lot for your help,

Camille

[Camille Mazoyer]

Ok, we can close the post. I found the bug into my own matlab routine.

Thanks a lot for your help,

Camille