On a hot day, looking at a thermometer would tell you precisely what the temperature is where you are.

Many seem to think the Pollutant Standards Index (PSI), Singapore's main indicator of air quality, works the same way. They expect it to correlate to the haze they see or smell around them at that moment.

This has given rise to some disquiet about the PSI readings in recent days, as the haze conditions worsened. Some members of the public expressed concerns when the posted PSI values - averaged over the past three or 24 hours - turned out different from what they observed.

Yesterday, the National Environment Agency (NEA) spelt out just how it compiles the PSI and why it does it that way.

It explained that it uses a specific way of collecting air samples over time and measuring the amount of micro-particles in them to determine their quality.

It also requires a wider base of samples, taken from various parts of the island to get a better picture of how polluted the air is across the country.

The NEA said that its methods and equipment to measure pollutants are based on the United States Environmental Protection Agency standards.

The PSI is derived from measurements at 11 ambient air monitoring stations located at places such as Temasek Polytechnic and Nanyang Technological University.

Five key pollutants are measured separately. But the PSI is not a composite index. Instead, it reflects the pollutant of the highest concentration.

But just how does the NEA measure levels of tiny particles in the air like PM2.5 and PM10 and translate them into a PSI reading?

An air sample is drawn into a chamber where particulate matter sticks to a filter;

Then, a beta ray (a form of radiation) is passed through the filter;

The particles on the filter would weaken the ray's intensity, and how dim it gets tells you how much particulate matter there is;

These readings are then averaged over three or 24 hours before they are fed into a standard set of indices, based on the health impacts of the pollutant at different observed levels. This translates the reading into a PSI.

NEA said yesterday that health advisories will continue to be based on the 24-hour PSI and PM2.5 values as those are what there is most scientific evidence for.

It explained that the 24-hour PSI is also the best indicator for the health impact of prolonged exposure to the haze than shorter-term measures, which can fluctuate quite a bit over the course of the day. This was evident yesterday, for example, when Singapore enjoyed blue skies in the afternoon when the haze seemed to clear temporarily.