The CF conventions state that ['scale_factor_at_projection_origin' is a [required attribute for Stereographic grid-mapping](https://cfconventions.org/cf-conventions/cf-conventions.html#_stereographic) But our code is [completely ignoring this](https://github.com/SciTools/iris/blob/a338ca530d5220330f136e58a70d502f62f52e6a/lib/iris/fileformats/_pyke_rules/fc_rules_cf.krb#L1479). ( Note that the comment ["this is checked elsewhere"](https://github.com/SciTools/iris/blob/a338ca530d5220330f136e58a70d502f62f52e6a/lib/iris/fileformats/_pyke_rules/fc_rules_cf.krb#L1496) is totally untrue ! It couldn't be really, as this is the only place that we will inspect the file variable ) By ignoring this, we are mis-representing any input where this is different from 1.0. Suggested action: **We should either support this, or at least check it and raise an error/warning when our loaded version is incorrect.** Note also: when saving, we currently [warn that the result is not exact + don't save anything](https://github.com/SciTools/iris/blob/a338ca530d5220330f136e58a70d502f62f52e6a/lib/iris/fileformats/netcdf.py#L2130). We could fix that too, but at least it is not "wrong".