Allowed usage of multi_model_statistics on single cubes/products

[schlunma]

Description

This PR allows the usage of our multi_model_statistics and ensemble_statistics preprocessors on single cubes/products.

Closes #1211

---

Before you get started

• ☝ Create an issue to discuss what you are going to do

Checklist

It is the responsibility of the author to make sure the pull request is ready to review. The icons indicate whether the item will be subject to the 🛠 Technical or 🧪 Scientific review.

• 🧪 The new functionality is relevant and scientifically sound
• 🛠 This pull request has a descriptive title and labels
• 🛠 Code is written according to the code quality guidelines
• 🧪 and 🛠 Documentation is available
• 🛠 Unit tests have been added
• 🛠 Changes are backward compatible
• 🛠 Any changed dependencies have been added or removed correctly
• 🛠 The list of authors is up to date
• 🛠 All checks below this pull request were successful

---

To help with the number pull requests:

• 🙏 We kindly ask you to review two other open pull requests in this repository

[codecov]

Codecov Report

Merging #1849 (82a97f4) into main (31328ed) will increase coverage by 0.00%.
The diff coverage is 100.00%.

@@           Coverage Diff           @@
##             main    #1849   +/-   ##
=======================================
  Coverage   91.65%   91.65%           
=======================================
  Files         232      232           
  Lines       11555    11557    +2     
=======================================
+ Hits        10591    10593    +2     
  Misses        964      964           

Impacted Files	Coverage Δ
esmvalcore/preprocessor/_multimodel.py	97.37% <100.00%> (+0.02%)	⬆️

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

[schlunma]

Open question for the @ESMValGroup/esmvaltool-coreteam:

In the current state of the PR, using std_dev on a single cube returns a fully-masked array. The reason for this is the default usage of ddof=1 by iris.analysis.STD_DEV which leads to a division by zero.

This definitely makes sense, but one could also argue that the result should be an array with zeros here (see #1211). How should we deal with this? Iris allows using ddof=0 as kwargs in cube.collapsed, but currently it's not possible to pass arbitrary kwargs to multi_model_statistics. An option could be to add an additional operator that uses ddof=0.

[bouweandela]

Are there more iris analysis operators that support arguments? If yes, it might be nice to add support for that in the multimodel statistics function. Maybe best done in a separate pull request?

[schlunma]

Yeah, lots of operators support arguments, see https://scitools-iris.readthedocs.io/en/latest/generated/api/iris/analysis.html. Examples are mdtol, or the many arguments of iris.analysis.PERCENTILE (which are partly handled at the moment with the syntax pXX.YY for the operator).

I opened an issue here: #1851

[schlunma]

This is ready for review now

[valeriupredoi]

great, cheers @schlunma 🍺