search

rigoudyg / climaf

CliMAF - a Climate Model Analysis Framework - doc at : http://climaf.readthedocs.org/
Other
18 stars 7 forks source link

Fix seasonal climatologies #150

Closed mickaellalande closed 5 years ago

mickaellalande commented 5 years ago

I was trying to compute the seasonal climatologies by myself and compare it with CLIMAF/CDO and I had different results. I realized that changing the flag -seltimestep to -selmonth works (seltimestep looks like a bug? I don't know, I am not familiar with CDO).

See my notebook on ciclad: /home/mlalande/notebooks/examples/season_clim_example-Copy1.ipynb

senesis commented 5 years ago

The cdo command which represents what Climaf actually does shouldn't be : cdo timmean -seltimestep,'3,4,5' snowc.mon.noaaV2c_.... but rather cdo timmean -seltimestep,'3,4,5' -ymonavg snowc.mon.noaaV2c... This, because function clim_average calls function annual_cycle before doing the seltimestep. So, CliMAF here only miss the accuracy of weighting the months differently according to their number of days

@jservonnat may wish to comment on this analysis, because he wrote that fucntion

mickaellalande commented 5 years ago

Ok my mistake, I didn't see the computation of the annual cycle first. Thanks for the reply!

mickaellalande commented 5 years ago

Is there an easy way to take into account the number of days in a month for monthly climatologies with CDO/CLIMAF (like yearmonmean operator instead of yearmean)? And in your opinion, is this important to take into account for climatologies analysis? It is usually small differences (for example http://xarray.pydata.org/en/stable/examples/monthly-means.html), however I was wondering if this can have certain importance is bias computations or other?