Closed casadoj closed 8 months ago
I have replaced 'approach'
by 'model'
this line in 4_confusion_matrix.ipynb:
total_exc = xr.Dataset({
'1_deterministic_+_1_probabilistic': deterministic_probabilistic,
'model_mean': model_mean,
'member_weighted': member_weighted,
'brier_weighted': brier_weighted,
}).to_array(dim='model')
Fixed. The only issue is that, now, the confusion matrices in COMB that were computed previously (with the dimension approach
) will not work in the notebook that computes skill.
In the COMB (combination of models) experiment, the xarray.Datasets have a dimension called
approach
with the different combination methods. If I renamed this dimension asmodel
, it would have the same name as in the experiment NWP, which would ease the notebook computing skill.So far, I had to create an object called
model_dim
to specify the name of this dimension: