Issues with Restarting ICEPACK Standalone v1.4.0 and Setting Initial Temperature Conditions

[Ajay Bhadran]

Hello everyone,

I am currently using the ICEPACK standalone driver (version 1.4.0) and am facing some challenges with model restart and initialization, particularly concerning initial temperature conditions.

1. Restart Issue
I attempted to restart the model using the restart option specified in the `icepack_in` file. For your reference, I have attached the following files:
- `icepack_in`
- settings file
- log file

Unfortunately, the model stops with an error during the restart process. I would greatly appreciate guidance on the correct procedure to enable and run a restart in ICEPACK standalone.

2. Initial Condition Problem
As an alternative approach, I tried to manually set the initial conditions, but I am uncertain if I did this correctly. Currently, I define the initial ice thickness and snow depth directly in `icedrv_init.F90` (around lines 1420 and 1429). This method works for initializing thickness; however, I am unable to set:
- the exact surface temperature, or
- the initial internal ice temperature profile.

Consequently, the simulated surface temperature is higher than observed, which leads to a bias that propagates throughout the simulation, causing the final model output to deviate significantly from actual observations.

Due to my limited understanding, I am modifying `icedrv_init` directly, but I am not confident that this is the recommended approach. I would appreciate advice on the proper way to set initial surface and internal temperature conditions in ICEPACK standalone.

Thank you very much for your time and assistance.

Best regards,
Ajay

 

[dbailey]

I don't think the SST is saved to the restart, so I believe this might be a time interpolation issue with reading in the dataset. Let me look into this further.

 

[Ajay Bhadran]

I don't think the SST is saved to the restart, so I believe this might be a time interpolation issue with reading in the dataset. Let me look into this further.

Thanks for looking into this.

 

[dbailey]

Ah, since you are running with the internal slab ocean model (SOM), then SST is actually on the restart. Looking at your log file:

Using restart dump=iced.2022-03-14-00000.nc


(icedrv_system_abort) ABORTED:
(icedrv_system_abort) called from /home/ajaybhadran/Icepack1.4.0_internal_temp/configuration/driver/icedrv_restart.F90
(icedrv_system_abort) line number 319
(icedrv_system_abort) string = (restartfile)Couldnt open netcdf file

Is this file in the restart directory? Was it not complete? That is, can you do an ncump -h on the netcdf file?

 

[MarekM]

I have basically the same issue/error message when trying to start a run from a restart file.

I'm using Icepack 1.5.3 and I have one case to create the restart file (here called case-spinup) and then try to use this file in case-from-restart to restart the simulation.

In the runlog i have basically the exact same error message (just the line reference is slightly moved):

(icedrv_system_abort) ABORTED:
(icedrv_system_abort) called from .../Icepack-Icepack1.5.3/configuration/driver/icedrv_restart.F90
(icedrv_system_abort) line number 326
(icedrv_system_abort) string = (restartfile)Couldnt open netcdf file

The file is in the restart directory, which I have set as follows:

restart = .true.
restart_dir = '../../runs/case-spinup/restart/'
restart_format = 'nc'

If I am in the terminal in the folder of the case and then cd to the path from the restart_dir then I get to the right folder (the same error message also occurs when I use the full path). I have attached the ncdump view of the file at the end. I'm not really sure where to look right now and would appreciate any help. Thanks!

ncdump -h iced.2015-01-02-00000.nc


netcdf iced.2015-01-02-00000 {


dimensions:


ni = 4 ;


ncat = 60 ;


variables:


double aicen(ncat, ni) ;


double vicen(ncat, ni) ;


double vsnon(ncat, ni) ;


double Tsfcn(ncat, ni) ;


double sice001(ncat, ni) ;


double sice002(ncat, ni) ;


double sice003(ncat, ni) ;


double sice004(ncat, ni) ;


double sice005(ncat, ni) ;


double sice006(ncat, ni) ;


double sice007(ncat, ni) ;


double qice001(ncat, ni) ;


double qice002(ncat, ni) ;


double qice003(ncat, ni) ;


double qice004(ncat, ni) ;


double qice005(ncat, ni) ;


double qice006(ncat, ni) ;


double qice007(ncat, ni) ;


double qsno001(ncat, ni) ;


double scale_factor(ni) ;


double swvdr(ni) ;


double swvdf(ni) ;


double swidr(ni) ;


double swidf(ni) ;


double sst(ni) ;


double frzmlt(ni) ;


double alvl(ncat, ni) ;


double vlvl(ncat, ni) ;


double apnd(ncat, ni) ;


double hpnd(ncat, ni) ;


double ipnd(ncat, ni) ;


double fsnow(ni) ;


double dhsn(ncat, ni) ;


double ffracn(ncat, ni) ;





// global attributes:


:istep1 = 24 ;


:time = 86400. ;


:time_forc = 0. ;


:nyr = 2015 ;


:month = 1 ;


:mday = 2 ;


:sec = 0 ;


}