The scooter in front of me at the Kerobokan lights has its exhaust at roughly the height of my chin. It is 08:10, nothing has moved for two minutes, and there are perhaps forty more machines idling in the queue ahead. On nine measured mornings on these exact roads, PM2.5 at breathing height ran between 12 and 99 µg/m³ (Dharmayasa et al., Jurnal Kesehatan Lingkungan, 30 June 2026). Almost every rider I speak to assumes that on a bad morning they are breathing the Kalimantan fires they keep reading about. They are not. The evidence says the smoke is being made right here, by us, and there is one day a year that proves it beyond argument.
The short version
- When Bali shuts down completely for Nyepi, its Day of Silence, airborne particulate in urban areas falls by 73–78% and in suburban areas by about 59% — a regional smoke plume blown in from another island would not observe a Hindu holiday (Nuraini et al., Forum Geografi 33(2), 2019).
- Measured at breathing height on Denpasar and Canggu roads in the 2024 dry season, PM2.5 averaged 24.63 µg/m³ at 08:00 — above the WHO 24-hour guideline of 15 µg/m³ — and peaked at 99 µg/m³ (Dharmayasa et al., Jurnal Kesehatan Lingkungan, 2026).
- During the Southwest Monsoon, winds over the southern ASEAN region blow from the southeast to south, and the documented 2026 haze pathway ran West Kalimantan to Sarawak — northward, away from Bali (ASEAN Specialised Meteorological Centre).
Is Bali haze the same smoke as the Borneo fires?
No. In almost every case, what a rider smells in south Bali is local combustion — traffic and open burning — not transboundary haze from Sumatra or Kalimantan. Indonesia is currently having its worst fire season since 2019, and Bali is conspicuously absent from the emergency.
The scale of that emergency is not in doubt. By 11 August 2026 Indonesia had recorded roughly 154,000 fire hotspots, up about 60% in a fortnight from 96,000 on 27 July, against 88,000 by late July in the catastrophic 2019 season. By 25 August the hotspots were concentrated in South Sumatra, West Kalimantan and Central Kalimantan, with haze pushing into Malaysia and Brunei.
Read those reports closely and one thing stands out: Bali is not mentioned in them. Not as an affected area, not as a monitored one. That is not an oversight. The ASEAN Specialised Meteorological Centre's own review of the July 2026 haze situation records that winds across the southern ASEAN region blew mainly from the southeast to south, and traces the transboundary pathway from West Kalimantan into Sarawak — that is, northward. Kalimantan and Sumatra sit north-west of Bali. On the prevailing monsoon flow, Bali is upwind of the fires, not downwind of them.
One honest nuance: the ASMC did record persistent hotspot activity in the Lesser Sunda Islands — Bali's own island chain — in late July. Those are local dry-season fires. That is this article's argument, not a counter-example to it.
What happens to Bali's air when the whole island stops?
Once a year, Bali runs an experiment no research budget could fund. On Nyepi, the Balinese Day of Silence, the entire island shuts for 24 hours: no vehicles on the roads, no fires lit, no flights, the airport closed.
Researchers from BMKG, Indonesia's national meteorological agency, measured what happened to the air. At two urban sites (Badung and Ubung) and one suburban site (Singaraja), they found that during the Day of Silence, total suspended particulate fell by 73–78% in urban areas and by about 59% in the suburbs, relative to ordinary days either side. Their conclusion, in their own words, is that human activities highly influence the air quality.
A note on precision, because it matters: that study measured TSP — total suspended particulate, meaning all particle sizes — not PM2.5 specifically. It is evidence about where the particles come from, not about their size. The PM2.5 evidence comes from a separate dataset, below. Keeping those two jobs separate is what makes the argument hold.
And the logic is hard to escape. If three-quarters of the airborne particulate over urban Bali disappears the moment Balinese people stop riding and stop burning for a single day, then three-quarters of it was never blown in from anywhere. Smoke that travelled from Kalimantan, hundreds of kilometres away, does not pause for a Hindu holiday. The island's particulate load is, overwhelmingly, made on the island.
What is more, 2026 has made that local source bigger rather than smaller. Bali's main landfill, TPA Suwung, stopped accepting organic waste — roughly 65% of the island's stream — on 1 April 2026, and was ordered to close entirely from 1 August. The waste-to-energy plant intended to replace it is not due to be operational until 2028. Bali generates about 3,400 tonnes of waste a day, of which around 71% was already going unmanaged (SIPSN data via Bali Post; Kompas, August 2025). In the months since the organic ban, residents and local outlets have reported a marked rise in household and roadside burning — trash fires on empty plots, smoke complaints across the south (The Jakarta Post, 26 April 2026). Nobody has formally measured that increase, so I am not going to attach a number to it. But the direction is not in dispute, and every gram of it is produced on the island.
How much PM2.5 is actually on the road you ride?
This is where the measurements get uncomfortably specific. A team publishing on 30 June 2026 put SPS30 sensors at breathing height across five stations — Dauh Puri Kaja, Dauh Puri Kauh, Kerobokan, Tibubeneng and Jimbaran — and sampled at 08:00, 13:00 and 18:00 through the April-to-July 2024 dry season. Those are the roads riders actually use, measured at the height air actually enters a helmet.
The morning average across all stations was 24.63 µg/m³, against a WHO 24-hour guideline of 15 µg/m³. The highest single reading was 99 µg/m³, at Dauh Puri Kauh on 18 June 2024. Kerobokan did not drop below 58 µg/m³ on any of the nine mornings in the series. Indonesia's own national daily limit is 55.5 µg/m³ — close to four times the WHO guideline — so "within the national standard" is a good deal less reassuring than it sounds.
Nine mornings on Denpasar and Canggu roads, at breathing height
PM2.5, µg/m³, measured 08:00, April–July 2024. All values sourced — nothing here is modelled.
- Five-station mean (measured)
- Dauh Puri Kauh, street level (measured)
- Regulatory reference lines
| Date | Five-station mean | Dauh Puri Kauh |
|---|---|---|
| 12 May | 35.40 | 30 |
| 14 May | 32.14 | 14 |
| 16 May | 42.96 | 62 |
| 22 May | 34.40 | 47 |
| 24 May | 53.58 | 93 |
| 29 May | 54.24 | 92 |
| 30 May | 67.56 | 90 |
| 18 Jun | 48.88 | 99 |
| 21 Jun | 37.88 | 55 |
On nine mornings in Denpasar and Canggu, PM2.5 at breathing height ran from 14 to 99 µg/m³, with no haze from Kalimantan involved — the five-station morning mean was 24.63 µg/m³ across the full season (Dharmayasa et al., 2026).
Why the timing of the peak is the real proof
Here is the detail the general coverage misses, and it is the one an engineer notices first.
Transboundary haze produces a flat, multi-day elevation. A smoke plume that has travelled a thousand kilometres has been stirred and diluted the whole way; it arrives as a persistent grey lid that does not care what time it is. What the Denpasar dataset shows instead is a hard 08:00 spike that decays by 13:00.
That shape is the signature of local combustion trapped under a shallow overnight boundary layer. Air cools and settles overnight, forming a lid a few hundred metres up. Everything burned under that lid — every cold-start scooter, every roadside fire — accumulates instead of dispersing. Then the sun comes up, the inversion breaks, and the trapped load mixes away by midday.
The morning peak is not smoke that travelled. It is smoke that never left. And it peaks at precisely the hour riders commute.
Two kinds of bad air, two different shapes
Illustrative — depicts the mechanism, not a measured dataset.
A sharp 08:00 peak that decays by midday is the signature of locally produced combustion trapped under an overnight inversion, whereas transboundary haze arrives as a flat multi-day elevation — the measured Denpasar readings show the former (Dharmayasa et al., 2026).
I wrote about what a genuine transboundary airshed looks like in the Borneo peat-smoke piece and in the Singapore haze contingency guide. Those places really are downwind of somebody else's fire. Bali is a different problem wearing the same word.
Why your helmet vents do nothing about it
Passive helmet vents are ram-air devices. There is no fan in them. They convert forward motion into airflow, which means their output scales with road speed — and collapses toward nothing as speed does.
South Bali peak-hour traffic runs below 20 km/h, and through parts of Canggu and Seminyak it drops to not much more than walking pace. A 5 km trip that takes 10–15 minutes off-peak can stretch past an hour at peak. So the rider is sitting still, inside the 08:00 concentration peak, with the only mechanism that moves air through the helmet switched off by the traffic itself.
Nor is the vent mesh doing any filtering. It is sized to stop gravel and insects — PM2.5 is about one-thirtieth the width of a human hair and goes through untouched. Closing the visor changes the draught, not the dose. (I went through that in detail in the piece on why AQI apps understate what riders breathe.)
That gap is the reason Easi Breezi is an active system rather than a better-shaped vent. It draws from the vehicle's own 12 V supply instead of from forward motion, so it keeps pulling air through its H11 HEPA element (≥95% of PM2.5) while the bike is stationary — which is exactly the condition in which passive vents deliver nothing at all. Patent pending, IP67-rated.
The number that is actually rising is not the concentration
Here is the projection, and I want to be plain about what it is: this is our own synthesis from sourced inputs, not a published forecast. No credible agency publishes a Bali-specific PM2.5 projection to 2031, and I am not going to invent one.
Dose is concentration multiplied by time multiplied by breathing rate. The Denpasar data says concentration is stable-but-high. It is the time term that is compounding:
- Bali registered 4,756,197 motorcycles in 2025 — about 1.07 per resident — and the fleet has been growing roughly 5.2% a year, around five times the rate of the population (BPS Bali; World Resources Institute Indonesia, July 2026). In Denpasar it is close to two motorcycles per resident.
- Foreign arrivals hit an all-time record of 6,948,754 in 2025, up 9.72% on the year, having been effectively zero in 2021.
- The only capacity relief in the pipeline is Phase 1 of the Bali subway, Airport–Kuta–Canggu, targeted for 2028 — one corridor, against island-wide growth.
Put together: peak-hour conditions trend toward becoming the default rather than the exception. On present trends a Canggu rider in 2031 inhales two to three times the 2026 dose on the identical route, in air of essentially the same quality, because the journey takes that much longer. The stakes are not abstract — every additional 10 µg/m³ of sustained PM2.5 costs roughly 0.98 years of life expectancy, and particulate pollution already takes about 1.3 years off the average Indonesian (AQLI, University of Chicago, 2025).
One caveat worth stating: Bali has very little regulatory-grade monitoring, which is why this article leans on peer-reviewed sensor campaigns rather than an official annual series. The widely-quoted 19.4 µg/m³ Denpasar figure is 2019 vintage, not current. Thin data is part of the story.
Frequently asked questions
Is the haze in Bali from the Kalimantan fires?
Almost never. On the Southwest Monsoon flow the wind runs from the southeast to south, and the documented 2026 smoke pathway went from West Kalimantan north into Sarawak — the opposite direction from Bali. What you smell in south Bali is local traffic and open burning.
Why does Bali smell smoky in the dry season, then?
Local reasons. Open waste and agricultural burning increases in the dry months; the Lesser Sunda Islands had their own hotspot activity in late July 2026; and since TPA Suwung stopped taking organic waste in April 2026, household and roadside burning has been widely reported to have risen. All of that is on Bali's own island chain, not hundreds of kilometres upwind.
What time of day is Bali's air worst for riding?
The morning commute. Measured readings peak at 08:00 and fall by 13:00, because overnight cooling traps combustion under a shallow inversion that the sun does not break until midday.
Is Bali's air quality actually dangerous, or just unpleasant?
The 08:00 average of 24.63 µg/m³ is above the WHO 24-hour guideline of 15 µg/m³, and individual street readings reached 99 µg/m³. Indonesia's national limit of 55.5 µg/m³ is close to four times the WHO figure, so passing the national standard is a low bar.
Does closing my visor protect me?
No. Visors and vent mesh are sized for gravel and insects; PM2.5 passes through. Closing the visor changes how much draught you feel, not how much particulate you inhale.
Ready to breathe cleaner on every ride?
Bali's air is not waiting for the subway, and it is not a problem that gets solved by riding faster — because the worst air is precisely where you cannot. The Easi Breezi unit is built for the standstill: an active, vehicle-powered H11 HEPA system that keeps filtering when your vents have stopped working. Pre-orders are $199.
If you want the wider picture, start with the rider air pollution guide, or read the earlier Bali post on what this year's fires meant for riders.
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.