7 September 2026 · Follow-up to the first regional wind calculation · Historical weather, 1 June 2023
Both terrain blocking and slope flows change the calculated wind at Brackendale. Two additional full-day CALMET runs separate these processes while retaining the weather and other settings from the first regional study. This comparison identifies what drives the calculation; observations are still needed to establish which settings reproduce local weather.
Terrain blocking diverts flow around obstacles under suitable atmospheric conditions. Slope flows represent winds generated by heating and cooling along terrain. CALMET parameterizes these processes; a 200 m grid does not resolve every local airflow. B.C. dispersion modelling guideline, Table 6.2
The table compares each setting with both adjustments switched off. It uses all 24 intervals at the existing place-label reference, at a nominal wind height of 10 m above model terrain.
| Adjustment included | Mean horizontal wind-vector difference | Largest interval difference |
|---|---|---|
| Blocking alone | 0.30 m/s | 3.11 m/s |
| Slope flows alone | 0.48 m/s | 1.70 m/s |
| Both | 0.77 m/s | 3.08 m/s |
Vector differences include direction and speed. The largest values can occur at different times, so the maxima should not be added. These are model-setting differences, without observations or pollutant releases. The two new runs each completed 24 intervals and passed record, time and finite-value checks. All reference-point intervals
The new cases and the existing both-on case use a 10 km terrain search radius. The retained both-off reference uses 5 km; source inspection shows that the searched terrain extrema do not feed the disabled blocking and slope-flow routines in that reference. This equivalence is based on the code path, without an extra both-off numerical run. Other assumptions, including the missing-sea-temperature fallback, remain common.
CALMET stores a vertical component relative to its terrain-following coordinates. Its reported magnitude is not automatically an Earth-vertical downdraft, and is not a measurement over a home. With the selected settings, the model obtains this component by integrating horizontal divergence through its layers. Provider model source: DIAGNO, DIVCEL, TOPOF2 and OUTHR
Reconstructing that equation from the saved horizontal winds agrees with all 7,056,720 tested interior values across the original three cases; the maximum difference is below 0.000002 m/s. This explains the stored numbers. It does not establish physical accuracy.
At the previously identified maximum in the both-on case—20:00 fixed PST, 1,500 m above model terrain, near 49.70553° N, 123.13730° W—the terrain-following component is −11.17 m/s. The same location and interval gives −10.32 m/s with blocking alone, −0.83 m/s with slope flows alone, and +0.02 m/s with neither. Blocking produces most of this particular excursion. This finding does not justify switching blocking off to obtain smaller values. Stability, ridge representation and resolution need evaluation against evidence.
Source inspection confirms that the model searches only within its supplied terrain grid. The Brackendale reference cell is about 3.1 km from the grid’s outer boundary, so both its 5 km and 10 km terrain searches are cut short. This does not prove a missing ridge changes the result; a larger-domain calculation must test that.
Across the current grid, 2,601 of 10,201 cells have full 5 km search support, and only the central cell has full 10 km support. The current radius comparison therefore combines a change in search distance with incomplete surrounding terrain. It is not a reliable uncertainty range for a fully supported valley model.
Adding 10 km of terrain on every side would supply the nominal 10 km searches for the original display grid. That is a geometric minimum, not proof of adequate weather or pollutant-transport boundaries. The shareable map can retain its requested viewing area while the underlying calculation uses a larger region.
| Work | What it resolves |
|---|---|
| Extend terrain, surface inputs and regional weather beyond the displayed square; compare the original inner area | Whether omitted ridges, valleys or boundaries affect community winds |
| Test horizontal and vertical resolution with declared settings and common comparison locations | Whether the calculated wind changes materially with numerical spacing |
| Recover usable marine temperatures and the native WRF land-use definitions | The explicit neutral air–water temperature assumption and shoreline classification |
| Obtain qualified local surface winds and vertical wind/temperature profiles during inflow, outflow, inversions and transitions | Whether the blocking, slope-flow and mixing calculations reproduce observed valley behaviour |
| Link a current flare design to emission rates, composition and event duration | Conversion from weather research to pollutant concentrations and exposure comparisons |
No new chemical concentrations or community risk colours follow from this wind experiment. The existing interactive explorer remains illustrative.
Settings, results, equation checks and domain-support calculation · Vertical profiles at the three references