Updates for the new ismip7 cism source branch

The PR [ESCOMP/CISM/pull/91](https://github.com/ESCOMP/CISM/pull/91) is bringing in a lot of changes for the ISMIP experiments this month. We will need a CISM-Wrapper tag and updates to the namelist for these changes.

Claude went through the PR and came up with these namelist changes to support the new code:

## Summary

PR #91 renames several CISM config-file keys and changes a few defaults as part of a
larger rewrite of basal friction, isostasy, and lateral-melt handling. None of the
renames have a backward-compatible alias, so any CESM namelist/config template that
still uses the old key names will have those settings **silently ignored** (the model
will fall back to the new default rather than erroring). We need to update the CESM-side
config templates and re-verify any experiments that rely on the affected options.

## Renamed keys (old → new)

These need to be updated in all CESM config templates / user_nl files that set them:

- [ ] `coulomb_c_const_hi` → `coulomb_c_hi`
- [ ] `coulomb_c_const_lo` → `coulomb_c_lo`
- [ ] `whichrelaxed` → `which_relaxed`
- [ ] `relaxed_tau` → `tau_relax_const`
- [ ] `lithosphere_period` → `load_update_interval`

## Default value changes (relevant even if the key name is updated correctly)

| Parameter | Old default | New default |
|---|---|---|
| `tau_relax_const` (was `relaxed_tau`) | 4000 yr | 3000 yr |
| `load_update_interval` (was `lithosphere_period`) | 100 yr | 10 yr |
| `coulomb_c_lo` (was `coulomb_c_const_lo`) | 0.10 | 0.05 |

Action: any CESM template currently relying on the *old default* (i.e., not setting
these explicitly) should decide whether to pin the old value explicitly or adopt the
new default going forward.

## Options removed — check for usage in CESM configs

- [ ] `which_ho_babc = 3` ("modified Schoof friction law" / `HO_BABC_MODIFIED_SCHOOF`) — removed entirely.
- [ ] `which_ho_coulomb_c = 4` ("read basin-scale coulomb_c_hi/lo from file" / `HO_COULOMB_C_EXTERNAL_BASIN`) — removed entirely.
- [ ] `ocean_p_timescale` — config key removed with no replacement (the relaxation-toward-target scheme it controlled was removed).
- [ ] `which_ho_flotation_function = 3` ("modified linear, with topg_raised") — removed; the associated `topg_raised` output field is also gone.
- [ ] `coulomb_c_hi`/`coulomb_c_lo` as spatially-varying 2D input fields — no longer supported; these are scalar constants only now.

If any CESM compsets/templates set these, they'll need a config redesign, not just a
key rename.

## New options available (optional follow-up, not blocking)

- `which_lateral_melt` (+ new `[lateral_melt]` section: `melt_rate_const`, `melt_factor`,
`subglacial_discharge_from_ablation`, `thermal_forcing_avg_3d_to_2d`, `ztop_tfavg`, `zbot_tfavg`)
- `schoof_n`, `schoof_gamma`, `schoof_p`
- `inversion_damping_factor`
- `bpmp_threshold`, `haf_threshold` (default 500 m)
- `bmlt_float_init`
- New enumerated values for `which_ho_powerlaw_c` (4), `which_ho_calving_front` (2),
and `which_ho_deltaT_ocn` (4, 5)

## To verify before closing this issue

- [ ] Confirm whether CISM's config parser flags unrecognized keys as an error/warning,
or silently ignores them — this determines how urgently existing CESM templates
need auditing.
- [ ] Grep all CESM CISM config templates and any documentation/user guides for the
five renamed keys above.
- [ ] Re-run any CESM test suites / spin-ups that use `elevation_based_coulomb_c`,
isostasy with a relaxing mantle, or basin-scale Coulomb/powerlaw inversion, since
these are affected by both the renames and some accompanying physics changes in
PR #91 (see PR review discussion for details on the basin-scale inversion mask
changes).
- [ ] Double check the `which_ho_effecpress` option numbering — the printed option
descriptions were reordered in PR #91 and it's not yet confirmed whether the
underlying option values were renumbered to match.

## References
- [ESCOMP/CISM/pull/91](https://github.com/ESCOMP/CISM/pull/91) (full diff reviewed in accompanying code review notes)