search

plesager / ece3-postproc

Suite of processing tools for EC-Earth3 output
5 stars 8 forks source link

Cannot produce map of Komolgorov-Smirnov test with missing data #19

Open plesager opened 6 years ago

plesager commented 6 years ago

Part of the code in _/reproducibility/R_scripts/map_diffexperiments.R cannot handle NaNs. Is it something that need to be fixed, or are we happy with the work around that I introduced in 2daa4b5? So far the problem is only with sea ice.

fmassonn commented 6 years ago

I'll let Martin Ménégoz know about this issue and invite him to address it here.

mmenegoz commented 6 years ago

Hi,

Thanks for the changes that you included in the script Philippe. In theory, models outputs should not include NaNs. I would say that we should not exclude them when computing the means. With the original script, any NaNs in the model outputs, and the user will detect them in the maps showing the mean differences.

plesager commented 6 years ago

I moved the discussion about cloning to #24 for easier reference.