Hi, I've used version 2 many times, but now I'm trying to get spun up on using version 3 with some climate model data on an unstructured grid (E3SM). I'm having trouble creating the vdc file. I get an empty file, but there's no error that something went wrong. It seems that the documentation hasn't been updated to reflect the new command line tools, so maybe I'm just using the wrong commands.
Here are the commands I'm using (shortened for readability):
cfvdccreate -force -vars CLDLIQ:LHFLX:TMQ test_case.cam.h2.nc test_case.cam.h1.nc test.vdc
cf2vdc -vars CLDLIQ:LHFLX:TMQ test_case.cam.h2.nc test_case.cam.h1.nc test.vdc
the output is just this:
Copying variable ilev
Time step 0
Copying variable lev
Time step 0
Copying variable CLDLIQ
Copying variable LHFLX
Copying variable TMQ
but when do an ncdump on the file there's no mention of the variables.
So what am I doing wrong?
Vapore doesn't have much support for unstructured grids yet (the MPAS grid is the only exeption) . Are you sure that is what you have? I'm suprised that the command didn't generate an error. In any event the variables will only be in the test.vdc file if they are under a certain size. Typically they would be found in a sub-directory under "test_data". Lastly, just an FYI that you should only go through the step of creating a .vdc file if your data are large in the spatial domain (e.g. high resolution). In most cases you are better off simply importing the data directly into vaporgui without creating a .vdc file. Hope this helps.
The data I'm working with is roughly 25km resolution output from E3SM (similar to CAM-SE), so it's on an unstructured cubed-sphere grid. But it's also a spectral element grid, so I was already a little skeptical that it would work right out of the box.
The variables are large (48k columns x 72 levels) so the size might be the reason I didn't get an error, but I got the same result even when I just focused on using the 2D variables.
Should I use the same commands if I remap to a lat-lon grid?
VAPOR3 doesn't understant the CAM SE grid. You'll need to convert to lat-lon. You should be able to use the same commands. However, you might want to try importing the data directly as a NetCDF CF file and see if the performance is adequate.