On 11 August, the light over Kuching turned the colour of weak tea and the city topped IQAir's global ranking as the most polluted major city on Earth. The smoke had drifted north from peat fires in West Kalimantan. Measured wildfire aerosol has a number-mode diameter of 212 nanometres — against 61 nm on non-smoke days (Lu, Bhattarai, Samburova & Khlystov, Environmental Science: Atmospheres, 2025). That single number is the reason this smoke is the hardest thing a hepa filter helmet is ever asked to catch, and nobody writing rider advice this month has mentioned it.
Here is the short answer, before the detail: 212 nm sits dead centre of the 0.1–0.3 µm window known as the most-penetrating particle size — the band where filter media reach their minimum efficiency. Haze smoke is not dangerous to a filter because it is the smallest thing in the air. It is dangerous because it is exactly the wrong size.
Key takeaways
- Wildfire smoke has a measured number-mode diameter of 212 nm, against 61 nm on non-smoke days, which places haze particles inside the accumulation mode of roughly 25–400 nm (Lu et al., Environmental Science: Atmospheres, 2025).
- The most-penetrating particle size for fibrous filter media is 0.1–0.3 µm, and EN 1822 rates HEPA media at that size deliberately because it is the worst case the medium will ever face.
- Indonesia had logged more than 154,000 fire hotspots by 11 August 2026, ahead of both 2019 and 2023 at the same point in their seasons, while the national disaster agency's budget fell to its lowest level in fifteen years (Mongabay, 13 August 2026).
- A passive helmet vent is a ram-air device with no stated filtration efficiency at any particle size, and it moves least air exactly when traffic slows and smoke concentration peaks.
Three fire seasons, and one line going the wrong way
2026 is already running ahead of the two worst haze years on record. Counting hotspots at the same point in each season, Indonesia's FireWatch data showed roughly 88,000 by late July 2019 and roughly 58,000 by late July 2023 — against more than 96,000 by 27 July 2026. By 11 August that 2026 figure had passed 154,000, a jump of about 60% in fifteen days (Mongabay, 13 August 2026).
Burned area tells the same story faster. Madani Berkelanjutan's figures rose from 15,474 hectares in June 2026 to 124,040 hectares in July — eightfold in a single month. Indonesia's disaster agency BNPB counted 48,889 hectares across its six priority provinces as of 9 August, with West Kalimantan alone accounting for 28,680 of them, more than half (Mongabay, 13 August 2026).
The consequences crossed the border within days. Tebedu's air pollutant index passed 200 — "very unhealthy" — on 12 August, six schools in Sarawak moved online, and Pontianak suspended classes indefinitely as acute respiratory infection cases climbed (Mongabay, 13 August 2026).
Indonesia fire hotspots, same point in the season
FireWatch / Nusantara Atlas counts, as reported by Mongabay, 13 August 2026.
| Point in season | Hotspots |
|---|---|
| Late July 2019 | 88,000 |
| Late July 2023 | 58,000 |
| 27 July 2026 | more than 96,000 |
| 11 August 2026 | more than 154,000 |
One correction worth making, because it is circulating: this is not a Chiang Mai story. Northern Thailand's burning season peaked on 29 March 2026 and fades with the rains by May; Chiang Mai read an AQI of 48 — "good" — in mid-August. The August smoke is 1,500 km south, and it is larger.
Too big to diffuse, too small to impact
Filters have a worst size, and it is not the smallest one. Fibrous media catch particles by two mechanisms that work at opposite ends of the size range. Below roughly 0.1 µm, Brownian diffusion dominates: tiny particles jitter off course and blunder into fibres. Above roughly 0.3 µm, inertia takes over: larger particles are too heavy to follow the air as it curves around a fibre, so they impact it or are intercepted by it.
Between those two regimes is a trough. Neither mechanism works well, and efficiency reaches its minimum. That is the most-penetrating particle size, conventionally 0.1–0.3 µm, with the wider diffusion-to-impaction gap sometimes given as 0.04–0.4 µm. EN 1822 rates HEPA media at the MPPS on purpose — it is the worst case, so a rating there is a floor rather than a flattering average.
Now put the smoke back in. The measured number-mode diameter of wildfire aerosol is 212 nm, or 0.212 µm, against 61 nm on clear days. Smoke does not make the air's particles smaller; it shifts the whole distribution up into the accumulation mode of roughly 25–400 nm. And 0.212 µm is inside 0.1–0.3 µm. That is the entire argument in one line.
Where filter media are weakest, and where smoke lands
Illustrative curve shape. The 0.1–0.3 µm band and the 212 nm marker are the sourced values.
Two consequences follow for riders, and both are uncomfortable.
Unrated media have no answer here. A bandana, a surgical mask or a generic "pollution" face cover carries no stated efficiency at the most-penetrating particle size — often no stated efficiency at all. The honest answer to "what does it stop at 212 nm?" is that nobody has measured it, which is not the same as "most of it".
A passive vent has no rating at any size. A helmet vent is a hole with a screen to keep insects out. It is also a ram-air device: it needs forward motion to move air at all. In a haze episode, visibility drops and traffic slows, so vent pressure collapses at exactly the moment concentration peaks. Stop at a junction and you are sitting in a static pocket of 212 nm smoke with no flow whatsoever.
Twice the bad seasons: the next five years
The driver behind this season has not peaked yet. NOAA's Climate Prediction Center puts an 81% chance on this El Niño reaching "very strong" intensity between October and December 2026 — strengthening after the August–September haze peak has passed (CPC ENSO advisory, via Mongabay, 13 August 2026).
Longer term, Cai et al., Nature Climate Change, 2014 found across 20 climate models that extreme El Niño events roughly double in frequency under greenhouse warming, from about one in twenty years to about one in ten.
What that means for a rider is a planning statement rather than a forecast. Projection, and its arithmetic stated openly: if extreme El Niño moves from roughly one-in-twenty to roughly one-in-ten years, then a rider commuting daily in the Borneo–Sumatra corridor between 2026 and 2031 should expect at least one full August–September season of smoke in that band inside the window — rather than treating 2026 as a freak year. That is the doubling from Cai et al. applied to a five-year horizon; it is a synthesis, not a published figure.
Two things make that planning assumption harder to dodge. Indonesia's 2024 fire area was 376,000 hectares, down 68% on 2023 (Katadata Databoks, 2010–2024 series) — the quiet years are genuinely quiet, which is exactly why "last year was fine" is worthless as a guide. And preparedness is moving the wrong way: BNPB's 2026 budget is Rp 491 billion, roughly $27.5 million, its lowest in fifteen years and down from Rp 2.01 trillion in 2025 (Mongabay, 13 August 2026).
What actually moves air through rated media
The problem in the section above is pressure, not filtration. The moment you push rated media into a passive vent, you kill the airflow the vent depends on — in our own system modelling the filter occupies about a quarter of the fan footprint and consumes roughly 96% of available static pressure. An unpowered filter in a vent is mostly a plug.
Easi Breezi drives air through the filter with a vehicle-powered fan rather than relying on forward motion, so flow at a red light matches flow at speed — which is the condition a haze episode actually creates. The media is H11, ≥95% at PM2.5, the unit is rated IP67, and the system is Patent Pending. The filter is a consumable you swap. We have not measured our own efficiency at the most-penetrating particle size, so we do not claim one.
Frequently asked questions
Do motorcycle helmets have air filters?
Almost none do. Standard helmet vents are unfiltered openings with a mesh screen for insects, carrying no stated filtration efficiency at any particle size. Filtration in a helmet is currently an aftermarket or purpose-built addition, not a stock feature.
Does a HEPA filter actually help against haze smoke?
Rated media are the only thing in this conversation with a measured efficiency, which puts them well ahead of unrated face coverings. The caveat is that the most-penetrating particle size is where any medium performs worst, and 212 nm smoke sits in that band — so ratings matter, and airflow to drive particles through the medium matters just as much.
Is closing my visor enough during a haze episode?
No. A closed visor slows exchange but does not seal, and a helmet is not a sealed enclosure. It also traps heat and fogs, which is why riders reopen it within minutes. We covered the visor question in more detail in our post on visors and air pollution.
Which filter grade should I be looking at?
That is a longer answer than this post has room for. We compared the grades directly in HEPA filter for a motorcycle helmet: H11 vs H13 vs carbon.
Are haze particles bigger or smaller than traffic particles?
Bigger, and counterintuitively that is worse for filtration. Traffic ultrafines cluster far below 100 nm, where diffusion capture is strong; smoke shifts the mode up to around 212 nm, into the trough. We wrote about the traffic end of that distribution in ultrafine particles and motorcycle riders.
Ready to breathe cleaner on every ride?
August and September are the two months Indonesia's disaster agency flags as highest-risk, and this season is already ahead of 2019 and 2023. If you ride through them, the question is not which face covering to buy — it is whether anything in your helmet has a rated efficiency at all, and whether air is being driven through it when traffic stops. Pre-order the Easi Breezi unit at $199.
Written by Ash — mechanical engineer and founder of Easi Breezi, building an active HEPA filtration system for motorcycle helmets (patent pending). Based between Hong Kong and Bali, riding daily in the traffic this blog writes about.