Ambient observation methods for E304570, 2023–2025

Checked 5 September 2026. Methods review only: no station statistics or compliance determination. The existing outputs/woodfibre-bc-objectives.md correctly distinguishes the provincial numerical values from the CAAQS, but its broad annual-mean descriptions do not supply all calculation rules. Keep provincial and national calculations separate.

Numerical comparison and year

The provincial factsheet explicitly allows objectives to gauge historical air quality. Preserve gases in ppb when that is the supplied observation unit; its mass conversions are rounded at 25°C/1 atm. The 2030 PM2.5 CAAQS are future values, not the 2025 comparison. B.C. factsheet, 23 December 2025, pp1–3.

Both gas guidance documents explicitly identify 2023–2025 as the first three-year window for the 2025 hourly standards, and 2025 alone for the 2025 annual standards (printed p4/PDF7 in each). Comparisons of individual 2023/2024 annual values with current stricter values must instead be labelled retrospective numerical comparisons; the national annual values then were NO2 17.0 and SO2 5.0 ppb. CCME NO2 §2, CCME SO2 §2.

Provincial PM2.5

The detailed provincial implementation guide specifies:

• Midnight–midnight local standard time; a continuous-monitor day requires at least 18 valid hourly measurements. Divide by available valid hours, never 24 with invented missing-hour zeros.
• A complete year requires at least 75% valid scheduled daily samples in each quarter. Incomplete years with a 98th percentile ≥25 µg/m³ are flagged for reporting, not silently discarded or declared complete.
• Sort valid daily means with ties retained. The ascending index is floor(0.98*N)+1, equivalently descending rank N-floor(0.98*N); no interpolation. The text's word “largest” conflicts with its expressly ascending array; the equivalent rank resolves that wording.
• Report daily means/annual percentiles to whole µg/m³. Retain all precision in intermediate calculations.
• The annual value is the equal-weight average of four quarterly means, reported to 0.1 µg/m³.

Source: June 2009 provincial guide, Appendix IV, printed pp15–16/PDF17–18. Its annual-weighting statement was visually verified. The current provincial PM2.5 page summarizes this as an average of daily values, without the quarterly detail.

Provincial NO2

The November 2021 implementation guide, Appendix I (printed pp8–9/PDF10–11), specifies Pacific Standard Time midnight–midnight days. Daily maxima require ≥18 hours, or at least one available hour >60 ppb. Report daily maxima to 0.1 ppb; retain ties; use descending rank N-floor(0.98*N). Annual percentiles are reported as whole ppb, while intermediate calculations retain all digits. Annual means average all valid hourly values, reported to 0.1 ppb.

For the hourly metric, valid daily maxima must cover ≥75% of the year and ≥60% of each quarter. Incomplete years whose percentile exceeds the objective are retained and flagged. At least two of the three annual percentiles are required. Annual-mean completeness separately requires ≥75% of hourly observations annually and ≥60% each quarter. The provincial appendix supplies no 50%-quarter annual-mean exception. It also does not specify the final three-year rounding step or half-tie convention as explicitly as CCME. Officially ENV-validated observations are required (§5, printed pp6–7). Provincial NO2 implementation guide.

Current national gas rules; SO2 implementation

For NO2 use rank N-floor(98*N/100); SO2 uses N-floor(99*N/100). These are integer order statistics; for 365/366 valid days they select the eighth/fourth highest daily maximum respectively. Average the available annual percentiles across the fixed three-year window; two valid years suffice, divide by two and identify the missing year. NO2 Tables 5-2/5-3 and Appendix B, printed pp12,14,23/PDF15,17,26.

Both gases use the following completeness rules:

Source: SO2 Table 5-3, printed p13/PDF16; the NO2 table is identical except percentile. Retain separate exception flags; “complete” and “accepted under an exception” describe different evidence.

Both guides prescribe hourly observations, selected daily maxima and selected percentiles at 0.1 ppb; annual means at 0.1 ppb; final three-year hourly metrics at whole ppb. The calculated value is truncated one place beyond the target precision, then rounded upward when that remaining digit is ≥5, otherwise downward. Preserve raw precision separately. NO2 Table 5-4, printed p16/PDF19, SO2 Table 5-4, printed p15/PDF18.

The older provincial SO2 guide addresses interim objectives and says they end in 2019, replaced by the CAAQS level and metric (printed p2/PDF4). Its Appendix I's 75 ppb exception and interim percentiles must not be carried into 2025. Provincial SO2 guide, 2017. Current B.C. SO2 reporting methods point to the 2020 CCME calculation guidance.

National PM2.5 is a separate calculation

CCME uses ≥18 hours/day; valid daily averages rounded to 0.1 µg/m³; the same discrete 98th-percentile rank. Annual completeness for both percentile and annual mean requires ≥75% daily coverage annually and ≥60% in each quarter. If the quarterly criterion fails but the percentile exceeds the standard, it is retained with an incomplete-data flag; the wording does not waive the 75% annual criterion. At least two valid years suffice for either three-year metric, with disclosure.

The annual mean averages all valid daily means, not four equally weighted quarters, then rounds to 0.1 µg/m³. The three-year annual result rounds to 0.1; the three-year daily-percentile result rounds to a whole number. Appendix D uses upward rounding for half ties. CCME PM2.5/ozone guide, 2012, §§4.1–4.2, printed pp7–13/PDF17–23 and Appendix D printed p39/PDF49.

Measurement preprocessing: preserve the supplied fields

The provider's saved work/ambient-baseline/metadata/README.txt says the annual archive has undergone QA/QC, except folders explicitly suffixed _Unverified. Thus the column name RAW_VALUE does not itself mean a preliminary live-feed observation. The README does not define RAW_VALUE versus ROUNDED_VALUE, the rounding algorithm, or any hidden validation transforms. The inspected objective guides do not name those database fields. B.C. dataset catalogue. Preserve both and document the actual difference; do not guess which column is the formal input.

For a national gas calculation, Table 5-4 defines the hourly value at 0.1 ppb before selecting maxima/percentiles or computing its annual mean. A reproducible implementation can reconstruct that decimal operation from RAW_VALUE, then check agreement with ROUNDED_VALUE. Keep a separate full-precision descriptive calculation. The provincial PM2.5 and NO2 guides direct retention of intermediate precision but do not explicitly prescribe a raw-hour pre-rounding stage. The national PM2.5 guide explicitly rounds daily values to 0.1 µg/m³; it does not supply a rule for choosing these two provincial hourly fields. These choices must be recorded as method interpretations, not proof of attainment.

Negative values have a specific validation context. The 2020 B.C. Field Sampling Manual's Level 1 check 7 preserves valid subhourly negatives, but for hourly PM2.5 adjusts −3≤x<0 to zero and invalidates x<−3. Valid hourly gases below zero are adjusted to zero, with values below −3 investigated first. This is a provider validation procedure, not permission to overwrite unexpected values in an already verified archive without examining provenance. A residual verified negative would need clarification and a preserved original. Part B1 §7.5, printed p31/PDF34. Root's completed calendar audit reports no negative numeric observations in the selected records; this note did not recalculate that audit.

Candidate dispersion baseline is not an achievement metric

The July 2026 modelling guideline §8.1 recommends the most recent representative year, at least one year, with ≥75% coverage each quarter. Match baseline and model averaging periods. For Level 1, use the measured maximum. For Levels 2/3, hourly baselines use daily-maximum P99 for SO2, daily-maximum P98 for NO2, and screened-hourly P98 for other pollutants; daily baselines use daily-mean P98. Annual baselines use the mean of hourly data over a year, which differs from the provincial PM2.5 objective's quarter-weighted calculation. Screening intermittent episodes requires air-quality/meteorological evidence. An off-site station needs representative meteorology, terrain, surface and emission conditions. Guideline, printed pp70–72/PDF77–79.

Accordingly, label a 2025 E304570 statistic a candidate station-derived baseline, with its field, averaging and screening choices. Its relevance at the flare and around Howe Sound remains a separate question. Do not replace this one-year candidate with the 2023–2025 achievement statistic or claim that a single station supplies a spatial background field.

Implementation decisions and unresolved gates

These are recommendations for the planned analysis, not additional regulatory rules:

1. Confirm E304570's exact station identity, parameter/method, hourly time label and validation status first. A downloaded number is not proof of validated data. Identify whether 00:00 denotes the preceding hour or the first hour beginning that date before assigning days. Use fixed PST/UTC−8 if supported by station metadata, not daylight-saving civil dates. Include leap day: 2024 has 8,784 hours/366 days.
2. Preserve original values/quality flags and a documented valid mask. Do not turn missing or invalid codes into zero, interpolate missing hours, drop valid negatives automatically, combine NOx with NO2, or treat passive monthly SO2 as hourly observations. Investigate the actual archive conventions.
3. Produce four quarter numerators/denominators, annual coverage, valid-day counts, rank N/K, selected dates/values, year inclusion and every exception flag alongside metrics. Use integer rank arithmetic to avoid a floating-point product just below an integer. Apply thresholds before rounding coverage percentages.
4. For BC PM2.5, use the explicit quarter-weighted provincial formula as the declared primary method; also report the pooled-daily alternative because the current webpage is less specific. Define quarterly means from valid daily means and flag that definition as the implementation reading of Appendix IV, not an explicit displayed formula. No current replacement document resolving the wording was found in this bounded review.
5. For provincial NO2, retain unrounded selected percentiles through the three-year calculation and display both unrounded and rounded results. Identify the final half-up whole-ppb operation as borrowed from CCME, since the provincial appendix leaves it unstated. Do not import the national annual-mean exception into a claimed provincial result without clarification. Run national calculations separately using national rounding and thresholds; threshold-dependent exception masks can differ.
6. Keep actual observed concentrations, including wildfire periods, in the main record. Any exceptional-event adjustment needs its own evidence and method; it is not a generic license to delete high observations. Do not describe a station metric as an entire airshed's achievement without reporting-station designation and spatial assessment. The CCME gas guides distinguish calculable station metrics from officially designated reporting stations (§5).

No E304570 data have been scored here. The source PDFs and acquisition hashes are in sources/; provenance updates identify the exact versions examined. The remaining information gaps concern the station archive and the provincial ambiguities identified above, not an assumed default completeness rule.