I have downloaded 20 years of temperature reanalysis data from CMEMS. They represent daily temperature in 141 levels, in Mediterranean, between 1993 and 2012. It is an xarray dataset with this characteristics dataset I have also downloaded the climatology offered by CMEMS for the same period and same place. It is also an xarray with this characteristics climatology dataset Now I need to compute anomalies of temperature, which usually are data - climatology. Usually I use this code
ds_clm=ds_clm.rename({'time': 'month'})
ds_temp_ano = ds_temp.groupby('time.month') - ds_clm
The problem is that this code is not working and it is killing my kernel.
I have tried with
ds_clm_broadcasted = xr.broadcast(ds_clm.thetao_avg, ds_temp.thetao)[0]
where thetao is my temperature variable, but this is also killing my kernel. I have tried with a simple
ds_temp_ano = ds_temp.thetao - ds_temp_clm
but it is subtracting clm only to day 16 of each month only in the first year. I have no other ideas how to compute this.