I talked with Sean Davis over the phone a bit today. I think that, by leaving all the default domain files in place, and changing only SSTICE_DATA_FILENAME, he's gotten to the point where this works for him and is the same as what AMP usually does.Now, it may be the case that there's something different that AMP could be doing, but I think that Sean D.'s immediate issue is resolved now.
-
The www2.cesm.ucar.edu website is currently down for maintenance
You are using an out of date browser. It may not display this or other websites correctly.
You should upgrade or use an alternative browser.
You should upgrade or use an alternative browser.
More problems using a B compset output to force an F compset
- Thread starter sean_m_davis@noaa_gov
- Start date
sean_m_davis@noaa_gov
Member
I'm still confused. Just to be sure I'm understanding the terminology correctly -- The gx1v6 grid is 384x320 (i.e., the grid that the AICE/SST variables are output from the B_ compset run)? When I create a case with the f19_g16 resolution, the out-of-box SSTICE_DATA_FILENAME file is not on 384x320 resolution, but rather on a 1x1 degree grid. My assumption is that if I were to run at this resolution, I would need to convert my B_ compset output to a 1x1 deg grid from 384x320. Or are you saying that I DON'T need to convert my ICE/SST input data if I use f19_g16? Sorry to be dense here...
sean_m_davis@noaa_gov
Member
I'm still confused. Just to be sure I'm understanding the terminology correctly -- The gx1v6 grid is 384x320 (i.e., the grid that the AICE/SST variables are output from the B_ compset run)? When I create a case with the f19_g16 resolution, the out-of-box SSTICE_DATA_FILENAME file is not on 384x320 resolution, but rather on a 1x1 degree grid. My assumption is that if I were to run at this resolution, I would need to convert my B_ compset output to a 1x1 deg grid from 384x320. Or are you saying that I DON'T need to convert my ICE/SST input data if I use f19_g16? Sorry to be dense here...
If you run with f19_g16, you do not need to interpolate/remap your data. The data models are built to interpolate from a standard 1x1 degree grid to any grid, but not the reverse. This is why the out of the box dataset used is 1x1 degree. However, since your data are already on the gx1v6 (g16 for short), then no interpolation/regridding is required.Dave
If you run with f19_g16, you do not need to interpolate/remap your data. The data models are built to interpolate from a standard 1x1 degree grid to any grid, but not the reverse. This is why the out of the box dataset used is 1x1 degree. However, since your data are already on the gx1v6 (g16 for short), then no interpolation/regridding is required.Dave
sean_m_davis@noaa_gov
Member
If I input the g16 data, what do I use for the SSTICE_GRID_FILENAME ? The out-of-box default is 1x1. It looks like this file might work though (https://svn-ccsm-inputdata.cgd.ucar.edu/trunk/inputdata/ocn/docn7/domain.ocn.gx1v6.090206.nc). Or does it not matter?
sean_m_davis@noaa_gov
Member
If I input the g16 data, what do I use for the SSTICE_GRID_FILENAME ? The out-of-box default is 1x1. It looks like this file might work though (https://svn-ccsm-inputdata.cgd.ucar.edu/trunk/inputdata/ocn/docn7/domain.ocn.gx1v6.090206.nc). Or does it not matter?
sean_m_davis@noaa_gov
Member
OK, using that domain file worked. The model is running now!
However, I'm getting a weird result whereby my input sea ice doesn't seem to match my output sea ice very well. There is some weird business going on with masking and the sea ice values around the coastlines. To make sure this wasn't an issue with my current run, I used an out-of-box F_ compset run. The attached plot illustrates the problem.Plotted are the input ice_cov for January from sst_HadOIBl_bc_1.9x2.5_clim_c061031.nc, as well as the "pre-diddled" value (pre-bcgen). I also plot the output january ICEFRAC variable from the h0 file, and the difference between the output and the (pre-diddled) input. As you can see, there are some differences. I'm seeing very similar differences with the run I'm doing where I specify the sea ice myself. I should not that in all these plots, I've masked out land area using the "mask" variable in domain.camocn.1.9x2.5_gx1v6_090403.nc. Is there an explanation for these differences? Thanks!Sean
However, I'm getting a weird result whereby my input sea ice doesn't seem to match my output sea ice very well. There is some weird business going on with masking and the sea ice values around the coastlines. To make sure this wasn't an issue with my current run, I used an out-of-box F_ compset run. The attached plot illustrates the problem.Plotted are the input ice_cov for January from sst_HadOIBl_bc_1.9x2.5_clim_c061031.nc, as well as the "pre-diddled" value (pre-bcgen). I also plot the output january ICEFRAC variable from the h0 file, and the difference between the output and the (pre-diddled) input. As you can see, there are some differences. I'm seeing very similar differences with the run I'm doing where I specify the sea ice myself. I should not that in all these plots, I've masked out land area using the "mask" variable in domain.camocn.1.9x2.5_gx1v6_090403.nc. Is there an explanation for these differences? Thanks!Sean
sean_m_davis@noaa_gov
Member
OK, using that domain file worked. The model is running now!
However, I'm getting a weird result whereby my input sea ice doesn't seem to match my output sea ice very well. There is some weird business going on with masking and the sea ice values around the coastlines. To make sure this wasn't an issue with my current run, I used an out-of-box F_ compset run. The attached plot illustrates the problem.Plotted are the input ice_cov for January from sst_HadOIBl_bc_1.9x2.5_clim_c061031.nc, as well as the "pre-diddled" value (pre-bcgen). I also plot the output january ICEFRAC variable from the h0 file, and the difference between the output and the (pre-diddled) input. As you can see, there are some differences. I'm seeing very similar differences with the run I'm doing where I specify the sea ice myself. I should not that in all these plots, I've masked out land area using the "mask" variable in domain.camocn.1.9x2.5_gx1v6_090403.nc. Is there an explanation for these differences? Thanks!Sean
However, I'm getting a weird result whereby my input sea ice doesn't seem to match my output sea ice very well. There is some weird business going on with masking and the sea ice values around the coastlines. To make sure this wasn't an issue with my current run, I used an out-of-box F_ compset run. The attached plot illustrates the problem.Plotted are the input ice_cov for January from sst_HadOIBl_bc_1.9x2.5_clim_c061031.nc, as well as the "pre-diddled" value (pre-bcgen). I also plot the output january ICEFRAC variable from the h0 file, and the difference between the output and the (pre-diddled) input. As you can see, there are some differences. I'm seeing very similar differences with the run I'm doing where I specify the sea ice myself. I should not that in all these plots, I've masked out land area using the "mask" variable in domain.camocn.1.9x2.5_gx1v6_090403.nc. Is there an explanation for these differences? Thanks!Sean
andrew_kren@colorado_edu
Member
Dave,What do you mean by modifying the user_nl_cice and docn_streams.txt files? What do I put in them if I am using a sstice_ts.nc ocean file that I created from a B run? I am using the f19_g16 resolution as well.
andrew_kren@colorado_edu
Member
Dave,What do you mean by modifying the user_nl_cice and docn_streams.txt files? What do I put in them if I am using a sstice_ts.nc ocean file that I created from a B run? I am using the f19_g16 resolution as well.