# Current application VOC evidence

5 November 2026. Research audit; no Explorer inputs or public health classifications changed.

## Finding and decision

The application supports more specific chemical research than the public notice alone: it publishes gas-stream compositions and selected individual VOC/PAH screening concentrations. It does not justify treating the Explorer's aggregate VOC concentration as benzene, a particular toxic mixture, or a directly convertible ppm value. Retain the reported aggregate rates, with their source attribution. Do not create a benzene plume by multiplying total VOC by feed-gas mole fraction.

## Source trail

The primary application is the October 2026 package for application 455458:
https://woodfibrelng.ca/wp-content/uploads/2026/10/Woodfibre-LNG-Project_WDA-Air-Emissions-Permit-Application-Package.pdf

Its retained SHA256 is `3e48a58676379cfca88dce6003d139b332d79f37f0ffead146b5661f7cadf536`. Page numbers below are physical PDF pages; printed appendix pages are also given where useful. Text extracts are retained beside this report. Tables B.1, B.2 and D.1 were checked against rendered pages, not text extraction alone.

PDF22 and PDF24 explicitly attribute HP/LP flare VOC, CO and NOx estimates to EPA AP-42 section13.5 (2018). The cited primary source is retained as `epa-ap42-industrial-flares-2018.pdf`:
https://www.epa.gov/sites/default/files/2020-10/documents/13.5_industrial_flares.pdf

EPA Table13.5-2 distinguishes its VOC factor (0.66 lb/million Btu) from CO (0.31), with net heating value as the basis. Both factors have a poorly representative rating. The table concerns well-operated assisted flares and is not a site measurement or an unburned-vent factor. Separate Table13.5-1 gives a methane-equivalent THC factor; it must not be substituted for VOC. All six published HP/LP combustion-case VOC:CO ratios are consistent, within rounding, with 0.66:0.31 (see `factor-ratio-check.csv`). This supports identifying the apparent factor pair; it does not reproduce the complete emissions workbook or establish the individual emitted species. The bounded audit does not establish a site-specific methane inventory or the precise species accounting behind the vent VOC total.

## Individual compounds already screened by the applicant

DMP Appendix B, PDF243-245 (B-1 to B-3), expressly covers final-design **normal and stabilizing operations**. Table B.1 excludes baseline; B.2 adds **EA background**, not newly measured 2026 background. These are applicant-reported maximum predictions, not our calculations, measurements or named-community values. Their source/time/receptor details have not been reconstructed here.

For benzene, B.1 reports maximum one-hour concentrations of 0.0140 and 0.0407 micrograms/m3 for normal and stabilizing operations respectively, versus 0.660 for EA normal operations. B.2 reports 2.53 and 2.56 micrograms/m3 with EA background of 2.52. The table's comparator is an Alberta objective of 30 micrograms/m3; this audit records the application's comparator, not an independently updated BC health benchmark. Annual values and exact reported precision are retained in `benzene-screening.csv`.

These results narrow a previous information gap for normal/stabilizing operations. They do not establish benzene exposure from the distinct process-upset or thermal-oxidizer-trip cases. They cannot be assigned to Brackendale or another community without the underlying receptor results. Do not generalize the application's statement that results are generally lower: its n-hexane row increases from 0.060 to 2.20/2.24 micrograms/m3, while remaining below its stated comparator.

## What gas composition adds, and what is still needed

Table D.1, PDF258 (D-1), gives mole fractions for fuel gas, BOG, feed gas and acid/flash/condensate-storage streams. It reports benzene, toluene and xylene among other constituents. For the high-CO2 design case, benzene mole fractions are 0.00002 (acid), 0.000031 (flash) and 0.00155 (storage); feed gas is 0.00003. These are stream compositions, not ambient concentrations or combustion exhaust speciation. Printed zeroes indicate the reported precision and do not establish exact absence.

PDF194/Table6.5 separates 7,641 sm3/h diverted vent gas from 1,204 sm3/h normal HP gases. PDF255 (C-9) identifies component flows of 7,264 acid, 358 flash and 19.2 storage sm3/h; their rounded sum is 7,641.2. PDF60 describes diverted streams as unburned while pilots/purge operate. Consequently the 8,845 sm3/h total must not be assigned one feed-gas composition or one destruction efficiency.

Before deriving a species-specific vent rate, resolve the standard-volume and moisture basis. PDF5 and PDF22 specify 20 degrees C and 1 atmosphere (PDF5 also says dry), whereas Appendix C uses 23.6449 m3/kmol in its sulphur conversions. Under the ideal-gas relation at 101.325 kPa, that molar volume corresponds to approximately 15 degrees C; 20 degrees C gives approximately 24.055 m3/kmol. D.1 also contains water mole fractions (including 0.06482 for design acid gas). This is a conversion-basis question, not evidence that either published total should be silently corrected. Obtain the original stream worksheet, dry/wet convention and conversion conditions, plus species-specific vent and combustion emission calculations. No new benzene release rate has been adopted from these rounded values.

## Concrete next research step

Request the existing species emissions workbook underlying the six combustion cases, vent VOC 11.75 g/s, and Appendix B tables, retaining component definitions (including methane/ethane treatment), factors, heating values, destruction/formation assumptions, gas standard conditions, water correction and source IDs. Retrieve receptor-keyed species results for the cases actually evaluated. If upset/vent species modelling was not performed, establish a documented species source schedule before extending the public map.

Coverage: targeted reading/search within the first 379 application pages, the listed source tables, and cited EPA section13.5. This is not an exhaustive review of all 4,149 application pages. No communications were sent. No new scientific model performance claim is made.