One sky, three verdicts: why Singapore says unhealthy and Malaysia says moderate

Johor Bahru and Singapore are about 25 kilometres apart and share their air. On a day when fine particulate sits at 70 micrograms per cubic metre, Singapore's PSI reads about 116 and calls it Unhealthy. Malaysia's API reads 96 and calls it Moderate, under an advisory column that says no restriction on outdoor activity.

Neither reading is wrong. They are different scales laid over the same measurement, and one of them is anchored to a target that expired.

The same air, three answers

All these indices work the same way. They take the pollutant concentration, find which band it falls in, and interpolate linearly to an index number. The bands are the only thing that differs. Running 70 µg/m³ of 24-hour PM2.5 through three of them:

A 65-point spread on one number. The meaningful gap is not in the arithmetic but in the public advice it triggers. One government's scale tells people to reduce outdoor exertion while its neighbour advises no restriction.

Where the difference comes from

Almost all of the difference comes down to one number — the concentration each index defines as its 100 mark, the top of the second band.

An index of 100 is not a measurement. It is a decision about where "acceptable" stops, and the three governments have placed it more than twice as far apart as the numbers alone suggest.

The part that is checkable, and awkward

Malaysia set a national ambient air quality standard for 24-hour PM2.5 in 2020. That standard is 35 µg/m³.

Its air quality index reaches 100 — the top of Moderate, the last band before the advisory changes — at 75.5 µg/m³.

75.5 ÷ 35 = 2.16

⚠️ The index therefore reports air as acceptable at a concentration 2.16 times the country's own published standard for that pollutant. Singapore's PSI has the same mismatch in the same direction, but smaller: its 100 mark of 55 µg/m³ is 1.57 times a 35 standard.

The 75.5 figure is not arbitrary either. It is the World Health Organization's Interim Target 1 for 24-hour PM2.5 — the first rung on a ladder intended to be climbed. The index is anchored to a waypoint rather than to a destination, and to the waypoint that the country's own 2020 standard already moved past.

One number moved, and the rest did not

Malaysia's other band edges — 12.0, 150.4, 250.4, 350.4, 500.4 — are the United States' pre-2024 AQI breakpoints, to the decimal place. The table was adopted wholesale.

Every anchor except one. The 100 line was moved from the American 35.4 to 75.5, and nothing else was.

That is worth stating carefully, because it cuts both ways. The divergence appears to be a deliberate, isolated decision rather than a different scientific framework — somebody chose that number on purpose. And the rest of the scale was inherited from a country that has since revised its own: the United States tightened its 100 mark to 35.4 in 2024, so Malaysia's table is a copy of a superseded table with one edit.

What this does not mean

It does not mean Malaysia's air is being misreported. The underlying concentration is measured the same way and published; anyone can read the µg/m³ figure and ignore the index entirely.

It does not mean an index should equal the health standard, either. Indices exist to trigger public advisories, and a government may reasonably decide that the point at which it tells millions of people to stay indoors is not the same as the point at which long-term exposure becomes a concern. Setting those two numbers apart is a policy choice, and a defensible one.

What it does mean is that the category words — "Moderate", "Unhealthy" — are not comparable across a border. "Moderate" in one jurisdiction can be double the concentration of "Moderate" in another. During a haze episode, when both countries are reporting on the same smoke, that is the difference between two headlines about the same sky.

What to do with it

Read the concentration, not the category. Every one of these agencies publishes the underlying µg/m³ alongside the index, and it is the only figure that means the same thing on both sides of the Causeway.

If you are deciding whether to run, or whether to keep a child indoors, use one scale consistently rather than whichever app you happen to open — and know which one it is. An app showing 96 and an app showing 161 for the same hour are not disagreeing about the air.

And when comparing a reading against a health guideline, compare concentrations. An index of 100 does not mean "at the limit"; it means "at whatever this particular table calls 100", which is 35.4 in one country and 75.5 in another.

Where this comes from, and what will date it

The breakpoint tables are each agency's own published calculation documents: Singapore's environment agency for the PSI, Malaysia's Department of Environment for the API, and the US federal regulations for the AQI, in the form effective from 2024. Malaysia's 35 µg/m³ standard is from its 2020 ambient air quality standard.

Every index figure above was computed here from those breakpoints by linear interpolation, not quoted: at 70 µg/m³ the PSI works out at 116.5, the API at 95.7 and the AQI at 161.2. ⚠️ The PSI publishes as 116 or 117 depending on whether the agency rounds or truncates, which we have not established — it does not affect the category, and the finding is the category.

⚠️ Two limits. We have compared published breakpoint tables, not monitoring-station performance, siting or calibration, any of which could make two nearby readings differ for ordinary reasons. And 70 µg/m³ is a worked value chosen because it separates the categories cleanly; at very low or very high concentrations the three scales converge and the disagreement disappears.

This dates the moment any of the three revises its table, and Malaysia's is the one under most pressure — its own 2020 standard already points at a lower number, and the source table it copied has since been tightened. If it adopts a 35.4 anchor the gap closes entirely. The durable point survives that. An index is a chosen scale, and the word printed beside it is a policy decision dressed as a measurement.