CDO

Hi all. I have re-gridded the CMIP5 model output from 10 different ESM models to a standard rectangular grid (360x180). I next calculate the ensemble mean (ensmean) for both the historical (1961-1990) and future RCP85 projections for the period 2025-2050. This works mostly fine, but I have some problems with grid cells particularly along land that may contain values in one file while in others only fill_value or landmask is present. Still, ensmean averages across the values for each grid cell regardless if the value is masked in the netcdf file or not. What i would like to do, is to only output values in grid cells where all 10 ESM models have actual values. If not I would like the value to be masked/fill_value. My command for creating the ensemble mean is simply "cdo ensmean *.nc ensemble_average.nc" where the *.nc files contains the ESM model output of the yearly average sea surface temperature value. suggestions for how to solve my problem is highly appreciated.

Thanks! Cheers, Trond