Question about the default forcing data on Github

[Yeo-jeong]

Hi everyone, I recently used the three default forcing data (COREII JRA NCAR_bulk) provided on Github and did some test run. And I have few questions about these data. Hope someone may have some idea about these problems. thanks.
1. Accoring to the data description page on Github. CICE-Consortium/CICE The gx3 forcing data includes ones year(1997) of data. But I run the model and the output filename is like 'iceh.1998-01.nc' and also a initial condition file 'iceh_ic.1998-01-01-00000.nc'. I think there is a conflict since you can't generate the results of year 1998 with the forcing data from 1997 .right?
2. The test run with COREII data. Accroding to the description page, COREII and JRA55 all have five years forcing data from 2005-2009, and to my knowledge, COREII is a climatolgical dataset but the output filename is like 'iceh.2017-01.nc'. So I am a bit confused why the forcing data of 2005-2009 generates the year 2017 result.
3. I tried the JRA55 -2008 forcing data and run the model. But the model only generates 11 output files from 'iceh.2008-01.nc' to 'iceh.2008-11.nc' and lack the file of December. I checked the runlog and confirmed the model run successfully. And the last 'idate' is 20081230. So I do not know why the file of last month failed to generate.

figure 1 is the output with gx3 1997 forcing data
figure 2 is the output with gx1 COREII forcing data
figure 3 is the output with gx1 JRA55 2008 forcing data
figure 4 is the the runlog of the last day of year 2008

 

[david_hebert@nrlssc_navy_mil]

Hello Yeo-Jeong,

In regard to your item (3) with the run with JRA55 forcing, does it work properly if you run with year 2007? Also, do you have use_leap_years=.true.? 2008 is a leap year, and there are 8 more records in the JRA55 2008 forcing than the 2007 forcing. I'm wondering if the issue is occuring to the run not doing the last day due to the leap day, thus not completing December to write an output file.

Thanks,
David

Hi everyone, I recently used the three default forcing data (COREII JRA NCAR_bulk) provided on Github and did some test run. And I have few questions about these data. Hope someone may have some idea about these problems. thanks.
1. Accoring to the data description page on Github. CICE-Consortium/CICE The gx3 forcing data includes ones year(1997) of data. But I run the model and the output filename is like 'iceh.1998-01.nc' and also a initial condition file 'iceh_ic.1998-01-01-00000.nc'. I think there is a conflict since you can't generate the results of year 1998 with the forcing data from 1997 .right?
2. The test run with COREII data. Accroding to the description page, COREII and JRA55 all have five years forcing data from 2005-2009, and to my knowledge, COREII is a climatolgical dataset but the output filename is like 'iceh.2017-01.nc'. So I am a bit confused why the forcing data of 2005-2009 generates the year 2017 result.
3. I tried the JRA55 -2008 forcing data and run the model. But the model only generates 11 output files from 'iceh.2008-01.nc' to 'iceh.2008-11.nc' and lack the file of December. I checked the runlog and confirmed the model run successfully. And the last 'idate' is 20081230. So I do not know why the file of last month failed to generate.

figure 1 is the output with gx3 1997 forcing data
figure 2 is the output with gx1 COREII forcing data
figure 3 is the output with gx1 JRA55 2008 forcing data
figure 4 is the the runlog of the last day of year 2008

 

[dbailey]

For the CORE forcing, you should also check the year_init, fyear_init, and use_restart_time flags. Make sure the first two are set to what you want and then set use_restart_time to .false.

Dave

 

[eclare@lanl_gov]

Hi Yeo-Jeong,
Dave's recommendation above is correct, for your first two questions, and David's hypothesis about the leap-day in 2008 is also a good place to start for the third question. The simulation (and output) dates do not need to match the forcing dates -- that is why the namelist file includes both year_init and fyear_init. Also, the date in the restart file can be used to determine the simulation dates, but it doesn't have to -- that is what the use_restart_time flag is for. If you are using the default settings for all of these, then we should change the defaults so that this is not so confusing for users, or at least document it better. Thanks for pointing it out.
Elizabeth

 

[wangxiong]

Hi David
Thank you for your suggestion. I double-checked the ice_in file and runlog and it turns out I already set the "use_leap_years" to True and the last output data in runlog is 20081231 instead of 20081230 (JRA55 2008 run). And, according to your advice, I re-run the model with 2009 JRA55 forcing data. This time all 12 netcdf files were output. So I think your hypothesis about leap-year is right. I am going to change the npt (from 8760 to 8784 ) to see if all 12 month data can be output.

Thanks
Yeo-Jeong