Closed nhsavage closed 1 year ago
Once this happens I suggest we consider bumping up PyPRECIS to v3.0 to reflect these breaking changes
v3.3 includes the fix needed to correctly read RegCM data over Africa https://scitools-iris.readthedocs.io/en/v3.3.x/whatsnew/3.3.html
note that once this is completed, there is no longer any need for the attempt at chunking in 5.2. Calculating percentiles (which doesn't actually work properly) because as of https://github.com/SciTools/iris/pull/3901 . There should be a follow on ticket created to do this.
There were a significant number of breaking changes that came in at Iris v3.0 so there will be some minor rewriting neede to update the notebooks. It may make sense to delay this until an Iris realease includes this PR so that we can more easily load RegCM data over Africa (and perhaps other domains with Mercator)
This also implies creating a new conda environment and updating the lock file