The recovered weather archive contains model skin temperature and surface heat flux, but neither currently supplies CALMET's water-temperature calculation. The retained runs use a neutral fallback because they have no overwater temperature observations. Land heat flux is calculated separately inside CALMET. This follows the saved controls and the official CALMET source; the wind-only result files do not save thermal diagnostics.
WRF skin temperature could support a separately labelled model scenario. It is not measured water temperature, and a model water category does not establish open ocean. WRF can calculate skin temperature independently of the archived SST field, which is zero throughout the retained sample. The annual file's initial settings do not prove the June production configuration. WRF ocean model.
There are two reasons to define the scenario before using it. CALMET's prognostic-water option requires acceptable temperature values across the entire input grid, including land. It also applies a legacy MM5 coastal rule that would retain only 125 of this subset's 228 model-water cells. That rule was designed for a different arrangement of model grid points; the CALWRF source places these WRF temperatures and land-use categories at the same locations. Four lake-marked cells survive the filter, so the result cannot be called an ocean-only boundary. A preliminary mapping using the saved coordinates would draw temperatures from up to about 26.6 km away. Seven selections are nearly tied; the actual model selections should be captured before using precise donor counts or distances.
The option also changes near-surface temperature profiles and recalculates surface exchange. WRF's sensible heat flux cannot simply replace CALMET's overwater heat-flux array, which includes moisture's effect on buoyancy. A useful next experiment would preserve the neutral comparison, document the coastal cell selection and temperature assumptions, save thermal diagnostics, and report the resulting changes as a sensitivity—not as validated weather or pollutant dispersion. No model inputs or simulations were changed by this audit.
Audit checks, source provenance, and candidate donor mapping. The candidate mapping is a reconstruction; actual model-selected donors and thermal diagnostics have not been captured.