5 September 2026 · Research methods
Canadian elevation data can now pass through the retained AERMAP terrain processor with explicit coordinate handling and independently checked ground interpolation. This is a preparation step for site modelling. It supplies no new wind field or concentration estimate, and the public plume map is unchanged.
The original NRCan CDEM 092G raster uses NAD83(CSRS) horizontal coordinates and CGVD28 elevations. AERMAP 24142 does not recognize the native EPSG4617 horizontal code. We used the pinned EPSG1946 horizontal operation to create a WGS84 raster, preserving CGVD28 heights in metres. The operation is approximate and time independent; its stated 1 m accuracy does not establish surveyed or epoch-specific alignment.
Two terrain extents, fixed before calculation, enclose nine reference and numerical-control points with 0.05° and 0.10° margins. The grids contain 1,290 × 1,899 and 1,770 × 2,379 samples at 0.75 arc-second spacing. Pam Rocks is included as a technical reference beyond the community map's 20 km view. It does not enlarge that map or define a plume boundary.
An independent reconstruction frozen before reading the produced rasters or AERMAP results passed 1,516 checks with zero discrepancies. This is an AI-assisted numerical review, not qualified site-model approval.
Both columns below are absolute elevations in CGVD28 metres. The hill value is AERMAP's terrain-search parameter, not a plume height or height above local ground. The Woodfibre location is an approximate operations-control-tower proxy, not a surveyed flare tip. The four centre/fraction points are synthetic interpolation controls.
| Reference or control | Ground | Hill parameter |
|---|---|---|
| Woodfibre operations-control-tower proxy | 8.79 | 2586.48 |
| Squamish Elementary E304570 | 6.33 | 2586.48 |
| Squamish Airport | 59.41 | 2586.48 |
| Pam Rocks | 7.36 | 1774.44 |
| NW computational terrain check | 1099.12 | 2586.48 |
| shore numerical centre control | 8.95 | 2586.48 |
| shore numerical fraction control | 8.47 | 2586.48 |
| slope numerical centre control | 1090.04 | 2586.48 |
| slope numerical fraction control | 1088.34 | 2586.48 |
Complete comparison table · Check summary and provenance hashes
The CDEM metadata describes ground or reflective surfaces and a vertical-accuracy range of 3–41 m. It is not a current construction-site survey. Regridding and then interpolating at a receptor uses different support from directly sampling the original raster; the largest difference in this fixture is about 1.22 m. This difference is retained in the table rather than treated as a processor error.
No vertical-datum conversion was performed. Future source bases, terrain and receptors need a common documented vertical reference; release height above ground is a separate input. This check covers interior, valid four-node interpolation. It does not certify missing-data or boundary behavior generally.
The unchanged hill values from one extent extension do not demonstrate final search-domain adequacy. Hill values were recorded, without an independent full hill-search reconstruction. Model selection for the fjord, current source geometry, local weather, receptor coverage and numerical convergence remain separate decisions.
The audit retains the initially rejected fixed-format controls and the correction to a TIFF-tag verification step. Neither correction changed the frozen coordinates, extents, transformation or numeric terrain samples. No concentration calculation was run.