# Air quality & pollution — gaps & caveats (SEA-11)

**As of:** 2026-09-06  
**Pack:** `topics/air_quality/scoreboard.csv`

## Hard gaps (empty cells — do not invent)

| Gap | Countries | Why |
|-----|-----------|-----|
| `fossil_fuel_energy_pct` | **All 11** | WDI `EG.USE.COMM.FO.ZS` returns **0** for 2016–2023 after last plausible non-zero (~2014/2015; Singapore even shows a negative 2015 artefact). **Do not ship zeros.** Use `fossil_fuel_elec_pct` as the citeable fossil-dependence companion. |
| City / station AQI | All | **Out of scope** — national exposure/emissions only; no invented city AQI time series. |

No other locked metrics are empty in this freeze.

## Stale / vintage badges

| Indicator | Country | Year | Note |
|-----------|---------|------|------|
| PM2.5 mean exposure | All | **2023** | IHME GBD 2023 via WDI |
| PM2.5 % above WHO | All | **†2017** | Series ends 2017; WDI threshold = **10 μg/m³** (AQG 2005) |
| CO₂ pc / Mt | All | **2024** | EDGAR via WDI |
| Fossil electricity % | **TLS** | **†2021** | Oldest fossil-elec cell |
| Fossil electricity % | BRN, KHM, LAO, MYS, MMR, PHL, VNM | **2022** | |
| Fossil electricity % | IDN, SGP, THA | **2023** | Newest fossil-elec cells |

**Mandatory dashboard badge:** **mixed vintages** — especially †2017 WHO exceedance and TLS †2021 fossil electricity.

## WHO guideline caveat (critical for writers)

1. WDI `EN.ATM.PM25.MC.ZS` measures population living where annual mean PM2.5 **> 10 μg/m³** (WHO **AQG 2005**).
2. WHO **AQG 2021** annual PM2.5 guideline is **5 μg/m³**. Under that standard, **even Brunei (7.6 in 2023) exceeds** the current guideline.
3. Do **not** claim “meets WHO guideline” for any SEA-11 country using the 2021 AQG, and do **not** present †2017 exceedance shares as 2021-AQG exceedance.

## Source consistency choices

1. **PM2.5 mean** — WDI `EN.ATM.PM25.MC.M3` latest non-null (**2023**).
2. **PM2.5 above WHO %** — WDI `EN.ATM.PM25.MC.ZS` latest non-null (**2017**), including Brunei **0** (genuine under-10 share that year).
3. **CO₂ per capita / total** — WDI EDGAR successors `EN.GHG.CO2.PC.CE.AR5` / `EN.GHG.CO2.MT.CE.AR5` (**2024**). Archived `EN.ATM.CO2E.*` ignored.
4. **Fossil electricity** — WDI `EG.ELC.FOSL.ZS` latest non-null; badge country year; same series as energy pack.
5. **Fossil TFEC %** — **GAP** for all (broken zeros).

## Caveats for readers / writers

- **Brunei** has the cleanest PM2.5 mean (**7.6**) and was the only SEA-11 with **0%** above the AQG-2005 10 μg/m³ line in 2017 — but CO₂ per capita is the **highest** (**20.2**).
- **Myanmar** has the worst PM2.5 mean (**35.1**).
- **Indonesia** dominates **scale** CO₂ (**812 Mt** in 2024); per-capita story is different (IDN **2.87** vs MYS **8.32** / SGP **9.75**).
- **Philippines** PM2.5 (**22.2**) is mid among MY/TH/VN/ID peers; CO₂ pc (**1.50**) is the **lowest** of that peer five.
- Singapore PM2.5 mean (**15.1**) with †2017 WHO% **100** is consistent (everyone above 10 does not require a very high national mean).
- Empty ≠ zero for TFEC; never zero-fill fossil energy %.

## Optional indicators deferred

- Direct WHO Ambient Air Quality Database city/station tables (not national series)
- EDGAR raw downloads outside WDI republication
- Black carbon / NO₂ / O₃ national exposure
- Production- vs consumption-based CO₂
- LULUCF-inclusive GHG totals
- Real-time AQI widgets

## Reading rule

Badge †2017 on every WHO-exceedance chart. State the **10 μg/m³** threshold in footnotes. Never invent post-freeze city AQI into this pack. Prefer EDGAR label on CO₂ footnotes.