Closed etiennesky closed 6 years ago
I am also concerned with the reduced output cases. If a file is missing (but shouldn't), the code will continue: we have to be clear that is ok to have that file missing. Something along those lines in pseudo code:
if file missing then
if not do_3d_var then OK else raise error
Hi @plesager about the reduced output, the missing files are usually not very important for core ECMean diags (such as sea ice in atm-only simulation, which IMHO does not make any sense at all!).
I can adapt the scripts to check for do_clear_flux and do_3d and see what is left.
About the reduced output: ok, we just need to be sure that they are not needed for ecmean.
Hi @plesager I have cleaned up the do_clear_flux and do_3d_vars code added them to the config files.
About the reduced output, I have tried to keep changes at minimum but this is the easiest I can see it. If a file should exists and doesn't, global_mean.sh will ignore it and if the files are required for diagnostics then ECMean will complain.
If you really want to I can replace do_clear_flux with reduced_output and remove some variables from the varlists insteadof the various [ -f ] tests.
this merge requests fixes some minor issues and adds support for "reduced" output which do not contains variables for clear-sky fluxes. missing fields in ECMean output are set as N/A.