11 September 2026 · Saved CALMET results for 1 June 2023 · Follow-up to the larger-domain vertical check
The previously identified interior excursion remains about −11 m/s under both tested coordinate conversions. This is a model-implied estimate, not a measured or validated downdraft. The coordinate distinction matters, but it does not explain away this particular large value.
CALMET saves W relative to terrain-following layers. A parcel moving uphill can gain absolute elevation while staying at the same height above terrain; saved W describes change in height above terrain. To estimate vertical motion in ordinary Cartesian coordinates, the vertical contribution of horizontal movement across sloping terrain must also be included. The provider defines the coordinate as height minus terrain elevation. We checked that definition, the continuity equation and the output conventions against the unchanged CALMET 6.5 source. Provider Version 5 theory guide, pp. 2-1 and 2-3 · Provider 6.5 source
The saved horizontal winds and W are at different heights: horizontal winds are at layer centres, while W is at layer tops. Two explicitly labelled estimates address that difference:
Both use the saved model terrain and the source’s centred terrain-slope calculation. They retain CALMET’s local Cartesian approximation for the UTM grid; no extra wind rotation was added. The runs rebuild W from processed horizontal-wind divergence rather than preserve native WRF vertical motion. Exact equations, source spans and assumptions are in the checks and full results.
At the previously selected location near 49.70553° N, 123.13730° W, the 1,500 m face lies 1,500 m above model terrain. The interval is 20:00–21:00 fixed PST (UTC−8), driven by native horizontal wind at its endpoint. Values below are m/s; neither estimate independently validates physical motion.
| Case | Saved terrain-coordinate W | A | B |
|---|---|---|---|
| Terrain, 5 km | -10.921640 | -10.733389 | -10.832249 |
| Terrain, 10 km | -11.166899 | -10.978705 | -11.077473 |
In the 10 km case, the corrections are only +0.188194 m/s for A and +0.089426 m/s for B. The added-area excursion also remains: at the previously identified 2,200 m face in 12:00–13:00, saved W of +12.556588 becomes +12.544014 (A) or +12.713934 (B).
The comparison includes every strict-interior cell of all three expanded runs: 39,601 cells × 24 intervals × nine shared faces, or 8,553,816 values per case. The outermost rows and columns are excluded because their calculation involves boundary values unavailable in the saved file. Complete results separately retain the original-area and added-area populations, every face, and all-ten-face W/A results.
The next table reports the signed value at the largest absolute magnitude in the same nine-face population. Different columns can reach their maxima at different locations.
| Case | Saved W extreme | A extreme | B extreme |
|---|---|---|---|
| Adjustments off | 3.932863 | 10.945280 | 10.986178 |
| Terrain, 5 km | -10.921640 | 11.107415 | 11.089743 |
| Terrain, 10 km | 12.556588 | 12.544014 | 12.713934 |
A coordinate correction can cancel one value and reinforce another. For example, the adjustments-off case’s earlier 3,000 m W maximum falls from +4.509260 to +0.296740 under A, while its B estimate reaches +10.986178 elsewhere. Selected examples do not describe the whole domain.
Mean absolute values over the shared population also increase under both conversions:
| Case | Mean abs W | Mean abs A | Mean abs B |
|---|---|---|---|
| Adjustments off | 0.166562 | 0.512561 | 0.513470 |
| Terrain, 5 km | 0.254918 | 0.583141 | 0.575470 |
| Terrain, 10 km | 0.306565 | 0.622862 | 0.615445 |
A and B differ by about 0.099 m/s at the persistent interior excursion, but their maximum absolute difference across the shared population is 3.017577 m/s in each case. The choice of which horizontal wind to combine with W can matter substantially elsewhere. These are descriptive differences between estimates, not physical-error measures.
All 28,512,720 saved interior W values across the three cases reproduce the source’s continuity arithmetic within the previously declared 0.00005 m/s tolerance. Input hashes, binary records, height and time conventions, terrain equality and interpolation checks passed. A separate scalar parser also confirmed selected extrema using the original binaries.
This establishes coordinate and arithmetic consistency. It does not establish atmospheric realism, validate vertical winds or justify plume transport. A later transport calculation must use a consistent vertical coordinate and interpolation; treating saved W directly as absolute vertical motion would omit the terrain-height change along horizontal motion, while adding that change twice would double-count it. The persistent large excursion still requires physical evaluation. No model settings were changed and no transport calculation was performed.
Complete checks and numerical results · Frozen protocol, methods, full tables and source locators · Original vertical report · Wind timing and observation check