Creating an ensemble with different calendar types
Added by Emre Salkım about 1 year ago
Dear all,
I want to create an ensemble data (with "ensmean" command) in CDO for 9 different regional models each starting and ending on same dates. However as they have different calendar types, when I use the ensmean in CDO it warns me as: "Inconsistent ensemble, processed only the first 7170 timesteps!"
I know that this inconsistency happens due to the model with fewer timesteps. My question is how may I make sure that "ensmean" calculates the ensemble with respect to the model with highest timesteps, while assigning NaN values for the missing data in models with shorter timesteps?
Is there a way to solve my problem in CDO or should I follow a different solution path?
Yours,
Emre
Replies (2)
RE: Creating an ensemble with different calendar types - Added by Estanislao Gavilan about 1 year ago
Hi Emre,
it is not difficult to fix that problem. You need to convert everything to the same calendar. You have two choices. The first method is to remove days adjusting the 360 days calendar. The second method is to add days to match the gregorian calendar. Take a look to this post, you can see both examples explained there.
https://code.mpimet.mpg.de/boards/1/topics/14007?r=14027#message-14027
Kind regards,
Estanislao
RE: Creating an ensemble with different calendar types - Added by Emre Salkım about 1 year ago
Dear Estanislao,
Thank you very much for your answer. “inttime” seems to be the real magic! Never knew there was such an operator. I will spend this Sunday playing with it since I have a 1000 years of data with 9 models 😅
Regards,
Emre