local run aifs
#7
by
agbbruno90
- opened
Hello,
I want to produce ensemble predictions using ERA5 as initial conditions on my local machine. I am having problems on using the ERA5 data from multiple datasets because of the different metadata included in each files. Is there an optimized way to read and use these data as initial condition of the model?
About the ensemble, my understood was to use the ERA5 ensemble members to initialize the model to produce the predictions. Is that correct? Reading a thread here it seems different.
How you can evaluate the spread of the ensemble outputs?