[Post-Processing Issue]: CESM2 Calendar Type Inconsistency

[willow-the-pillow]

Summary

Some of the CESM2 experiments have their calendar encoding as noleap (the historical runs) whereas some have their calendar encoding as 365_day (the rest of the experiments). In practice they're the same, but for consistency, we converted our CESM2 files to use noleap, as it seems to be more often used in the other models.

Models Impacted

CESM2

Which experiments were impacted?

All tier 1 transient experiments, All tier 2 transient experiments

Ensemble Members

All members

Relevant log output

Calendar Check Err Output: 
Comparing majority opinion hurs_day_CESM2_ssp126_r11i1p1f1_gn_20450101-20541231.nc with b.e21.BHISTsmbb.f09_g17.LE2-1091.005.cam.h1.U10.19900101-19991231.nc
365_day vs noleap

Details

To check the calendar type of an xarray dataset ds, you can print ds.time.encoding['calendar'].

Solution (if you already have one)

This is an example of changing a single dataset to have the correct encoding.