hello! I am trying to use some restart files that were recently generated and there are some issues I’ve not seen before. After the model has completed its run to the final time step (seen in work/atmosphere/log/matmxx.pe00000.log), I get this in accessom2.err.
Error in accessom2_deinit: atm and ice models are out of sync.
atm end date: 2000-02-01T00:00:00.000
ice end date: 2020-02-01T00:00:00.000
My accessom2.nml dates and forcing dates align, otherwise it wouldn’t run. The ice restart file names are also okay, the file names of the restart are below:
[qo9901@gadi-login-09 1deg_restart_test]$ ls /home/561/qo9901/access-om2/1deg_esm1p5_vgrid_fric_mo_i2o/archive/restart019/ice/
grid.nc i2o.nc iced.2000-01-01-00000.nc ice.restart_file kmt.nc monthly_sstsss.nc o2i.nc sicemass.nc u_star.nc
I will note that for the iced.2000-01-01-00000.nc file, these are the global attributes, which are line up with the run I did originally. This is the only place I see where the year ~20 comes up.
// global attributes:
:istep1 = 116880 ;
:time = 631152000. ;
:time_forc = 0. ;
:nyr = 21 ;
:month = 1 ;
:mday = 1 ;
:sec = 0 ;
}
The ice.restart_fileis also the year 2000 as expected, so it looks like a cice inbuilt problem, where nyr is not being changed? Not sure if I’ve missed something, but want to check before manually rewriting the attrs, or if there is a long term solution if restarting from other experiments. (I have tried with and without using payu clone --restart <dir>, and also changing runtype = 'continue' in cice_in.nml but no luck)
Exp dir: /home/561/qo9901/access-om2/1deg_restart_test
Restart dir: /home/561/qo9901/access-om2/1deg_esm1p5_vgrid_fric_mo_i2o/archive/restart019/
Thank you! and tagging some who might be able to answer this with eyes closed @aidan, @anton