[MITgcm-support] error tutorial_cfc_offline, uVeltave file not existing when nTimeSteps increased

[Céline Heuzé]

Hello,
I'm starting with MITgcm by running the tutorial_cfc_offline case. All goes well, until I increase nTimeSteps to 30 or larger; then I get the error:
(PID.TID 0000.0001) *** ERROR ***  MDS_READ_FIELD: filename: ../input/input_off/uVeltave.0004248120 , ../input/input_off/uVeltave.0004248120.001.001.data
(PID.TID 0000.0001) *** ERROR ***  MDS_READ_FIELD: Files DO not exist

There was a 2011 thread suggesting to add an offset of 30 timesteps (15 days) to the data.off file; that changed nothing.
Also testing various offlineIter0 only changed the iteration of the uVeltave non-existing file, seemingly to 720 + offlineIter0.

I test with the default settings in data.h, only nTimeSteps increased:
nIter0  = 4269600,
nTimeSteps = 30,
deltaTmom = 900.0,
deltaTtracer= 43200.0,
deltaTClock = 43200.0,
abEps = 0.1,
pChkptFreq=3110400000.,
dumpFreq=  31104000.,
taveFreq=  31104000.,
monitorFreq= 2592000.,
periodicExternalForcing=.TRUE.,
externForcingPeriod=2592000.,
externForcingCycle=31104000.,

Settings in data.off, with an example of how I set the offset to -30 timesteps x deltaT:
  offlineIter0=4248000,
  offlineTimeOffset=-1296000.,
  deltaToffline=43200.,
  offlineForcingPeriod=2592000.,
  offlineForcingCycle=31104000.,


Ultimately, I want to run this test case for several years.
Thanks in advance for any suggestion.

Regards,
Céline
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mitgcm.org/pipermail/mitgcm-support/attachments/20170322/34bd8892/attachment.htm>