What Malaysia publishes
The Air Pollutant Index Management System publishes a composite API figure hourly, free and without registration, from 68 monitoring stations across the country. That number is updated every hour and is driven by PM2.5 during a haze episode. It is genuinely useful and its publication, after years of government resistance, was an unambiguous gain — HAZE-OC II says so and means it.
What APIMS does not publish alongside that composite number is the raw PM2.5 concentration in micrograms per cubic metre as a separately accessible figure. What you see is an index value — a number on a 0-to-500 scale — not the underlying concentration in the unit that WHO guidelines, commercial apps, and Singapore’s own system use. There is, as one recent Malaysian guide puts it, no reliable mental arithmetic that converts an API into a µg/m³ figure, because the index compresses a curve and may be driven by a pollutant other than PM2.5 entirely on a clear day.
What Singapore publishes
Singapore’s National Environment Agency publishes two things simultaneously during haze: the 24-hour Pollutant Standards Index, which carries the same rolling-average lag as Malaysia’s API, and a separate 1-hour PM2.5 concentration in µg/m³, updated in real time. NEA’s own guidance is explicit about which to use for which purpose: the 1-hour figure for decisions about right now, the 24-hour PSI for planning ahead. The two numbers answer different questions, and NEA says so plainly rather than leaving the public to reconcile conflicting app readings on their own.
Malaysia has been measuring PM2.5 at its monitoring stations since before 2018, when the DOE first began incorporating it into the public API. The hourly concentration data therefore exists. It is not classified — the OSA battle is over and was won. It is simply not published as a standalone figure for the public to use.
What two researchers said about it during this episode
On 26 August 2026, as the haze worsened across the peninsula and Sarawak, Free Malaysia Today asked two atmospheric experts the direct question. Associate Professor Dr Helena Varkkey of Universiti Malaya’s Department of International and Strategic Studies said that Singapore reports both hourly and 24-hour average indices, and that it would be great if Malaysia provided this data too, so that people could make more accurate real-time decisions — specifically mentioning whether it is safe right now to go for a walk outside. Dr Jayaprakash Murulitharan, Malaysia’s first Cambridge-trained atmospheric scientist in the field and currently a senior principal assistant secretary at GovTech Malaysia, made the same call independently: the current practice of reporting based on a 24-hour average should be reviewed, because haze conditions can change rapidly with wind direction and weather patterns, making the average less representative of the prevailing air quality at any given time.
Varkkey made the same argument in The Star in October 2023, co-authored with Associate Professor Dr Matthew Ashfold of the University of Nottingham Malaysia, writing that Malaysia’s school closure and outdoor-activity decisions are typically based on the API of the previous 24 hours, which may not reflect the current haze situation. That piece was written during the 2023 haze season. It appears in this episode’s reporting without having produced any change in the intervening three years.
Why the lag has a clinical cost
This is not a data aesthetics argument. The 24-hour averaging lag maps onto a health timescale that is shorter than it implies.
The US Environmental Protection Agency, reviewing the literature, states that PM2.5 exposure over a few hours can trigger cardiovascular events — heart attacks, arrhythmias, acute heart failure — in people with existing cardiac disease. The operative word is hours, not days. Emergency ambulance dispatch data from Japanese cities shows significant increases in respiratory and all-cause presentations at lag 0 — same-day — with effects concentrated in the first 24 to 48 hours of exposure. A 2026 study in the American Journal of Respiratory and Critical Care Medicine found that 18% of hours occurring on days officially classified as “Unhealthy for Sensitive Groups” actually exceeded that category’s own concentration threshold — meaning the daily average labelled a day as moderately dangerous while individual hours within it were measurably worse. For schoolchildren, acute reductions in lung function are measurable at 1-day moving average concentrations; the effect is real within the school day, not spread across the week.
The practical consequence is specific. A school principal in Kuching deciding at 7am whether to halt outdoor activities is reading yesterday’s average. A parent of an asthmatic child checking the APIMS app before sports day is reading a number that may be masking a spike that developed after midnight. A GP advising a patient with unstable angina about whether to walk to the market has no official real-time PM2.5 figure to cite — only an index number that compresses time and dilutes peaks.
Singapore’s system exists precisely because its designers recognised this gap. Its dual publication is not a technical flourish; it is an acknowledgement that the 24-hour average and the real-time concentration answer different clinical questions, and that both questions deserve an official answer.
The history that explains the gap
This argument is not new. I first made it here in 1999.
In August of that year, during the controversy over the government’s decision to withhold the API entirely, Clouding the Issue made the clinical case in plain terms: “With the API being available, the ‘picture’ would be clearer and thus, accurate. Parents, elderly, the sick (respiratory problems) and medical doctors would benefit and could make meaningful decisions.” Singapore was already cited then as the counter-example — publishing its PSI hourly while Malaysia published nothing.
The same argument appeared again in A.P.I: Time to Put It Out! in June 2004, and again in Release AP-I in July 2005, both during further haze episodes in which the index was still being withheld. The 2005 column ended with a postscript that was not ironic: “Just when I completed this article and sent it to the editor, the API was released! At last someone had the courage to undo what should not have been done in the first place.”
That is the precedent. The argument was made in print, across multiple episodes, by multiple people, and it eventually produced a result. The API was released. APIMS now publishes hourly readings from 68 stations, free and without registration. That gain is real and was not inevitable.
What followed was the same pattern, one step delayed. Malaysia continued reporting its index based on PM10 — the coarser, less hazardous particulate — for years after it had begun measuring PM2.5 internally. When Wong Chen, then MP for Kelana Jaya, pointed this out in 2015 — noting that Singapore had already switched its index to PM2.5 the previous year — the deputy minister of Natural Resources and Environment defended the older standard. PM2.5 was incorporated into the public API on 16 August 2018, four years after Singapore, in the same year as Thailand.
The pattern across nearly three decades is consistent: Malaysia measures the data, withholds it or delays its public form, defends the position when challenged, and eventually moves — after sustained argument combined with a counter-example that makes the existing position untenable. The hourly PM2.5 concentration in µg/m³ is the current item in that sequence. The measurement exists. The counter-example is Singapore, where it has been separately published since 2016. What is missing, as it has been missing at each prior step, is the decision to publish.
The closing question from that 1999 column — written after the worst haze episode Malaysia had seen, asking whether we would be better prepared the next time around — was borrowed from a 1997 Berita MMA piece by Dr Krishna Gopal Rampal, then of the Department of Community Health, Universiti Kebangsaan Malaysia, now Professor Emeritus at the University of Cyberjaya, written during the haze that year.
Twenty-seven years on, the question is narrower but structurally the same.
One ask — the same one, restated
Publish the hourly PM2.5 concentration in µg/m³ alongside the composite API, for every station, updated in real time, during this episode — not after a review, not pending a new system, but now, from data the DOE already holds.
Singapore’s NEA shows what this looks like in practice: a 24-hour PSI for planning, a separate 1-hour PM2.5 figure for deciding whether to step outside now. Both have been published since June 2016. The MyJAS EQMS app already exists. This requires a new field, not a new platform.
The 24-hour average has its place — school closure policy, longer-term planning, epidemiological tracking — and should remain. This is not an argument to replace it. It is an argument to add the figure that answers the question the 24-hour average cannot: what is the air doing right now, in the unit that matters clinically, for the patient deciding whether to walk to the market.
This argument has been made from this column in 1999, in 2004, and in 2005. The previous version of it was eventually won. The API was released. APIMS exists. The question is whether the next step takes another decade, or whether someone finds the courage this season.
The 1999 column closed with a question borrowed from a 1997 Berita MMA piece by Professor Emeritus Dr Krishna Gopal Rampal, then of the Department of Community Health, Universiti Kebangsaan Malaysia, asking whether we would be better prepared the next time around. The answer then was “Are we?”
It still is.