Change in the computation of the scatter index?

[elmedhi]

Hi,
I noticed that @jsmariegaard changed the formula of the scatter index in the skill table. (1d55986).

I have very different results now in some scatter plots, with the same data. For example,
This plot was made in September/October:

This one was made a few days ago:

Could you explain the reason of this modification?

Thank you!

[jsmariegaard]

Hi @elmedhi , it was per request from @daniel-caichac-DHI following this issue: #137 - the work was carried out in the PR #142 - does this provide enough information?

[elmedhi]

Yes, thank you :)

[daniel-caichac-DHI]

@elmedhi the old implementation is not the default anymore but you can get it by using scatter_index2
def scatter_index2(obs: np.ndarray, model: np.ndarray) -> float: