Rename `*_ICE` diagnostic fields to `*_GLC`

Some confusion arose recently connected to the `_ICE` suffix on CTSM history variables. For most history fields, `SOMETHING_ICE` means that this is the SOMETHING field but only over glacier landunits. However, there are a few – such as `QRUNOFF_ICE_TO_COUPLER` – where `ICE` means that this is an ice as opposed to liquid term. So in this case, this is ice runoff averaged over all landunits.

I proposed, and LIWG members supported, that we change the suffix that denotes that a field is only averaged over glacier landunits, using `_GLC` rather than `_ICE` for this purpose.

This will be an easy change that will only affect a handful of history fields (hence I'm giving it the "simple bfb" label). However, it will impact the ability to compare master with CMIP6/ISMIP6 runs, so I'd suggest holding off on this until we're past the point that people will frequently want to do such comparisons with master – e.g., holding off until after water isotopes are in place (hence I'm giving it the "blocked: answer changing" label). (And yes, I recognize that it seems weird to have both of these labels, but both seem appropriate in this case.)