On foreign standards for air quality reporting
Update: Re-wrote a bit and added screenshot and a few links for clearer picture
Recently the blog The Heart Truths published a post questioning statements made by minister Vivian Balakrishnan that Singapore is probably the only country which reports 3-hourly average PSI with most other countries going by a much longer 24-hour averaging time was accurate. Within hours, the blog was rebutted by commenters as well as the Ministry of Environment and Water Resources (MEWR) Facebook page rebutting the blogger’s claim in a note here (also as a press release here). While Roy, the blogger has since updated his post, some issues bear addressing. Here they are in no particular order.
1. The MEWR Facebook note noted that some of the cited air quality indices for the US, Hong Kong and the UK are based on 24-hour averages, contrary to The Heart Truth’s claim that they provided hourly spot readings.
Having reviewed the links, I believe the verdict on the blogger’s and MEWR’s claim is mixed. Let’s look at some examples I found on real time air quality reporting for other countries.
In Hong Kong’s case it appears that PM10 (RSP) readings are based on 24-hour averages. But what about PM2.5 readings? Some Facebook commenters have pointed to this press release as evidence to show that Hong Kong reports PM2.5 (FSP) hourly readings. Strangely, the table of API subindex pollutants do not include PM2.5 readings. However, PM2.5 readings appear in the hourly pollutant concentration reports here.
What should one make of this? My own interpretation is that PM2.5 pollutant concentration levels, while not (yet?) incorporated in the overall AQI air quality index have been recognised to be sufficiently important to be reported on a spot hourly basis as opposed to a 24-hour rolling average. MEWR on the other hand appears to be correct that PM10 concentrations are reported as a 24-hour rolling average. Oddly, MEWR made no mention of how PM2.5 pollutant concentration levels are reported for Hong Kong. Why?
MEWR’s note cited a technical report by the US EPA (pg 8 here) that PM2.5 and PM10 concentrations are averaged over 24 hours. On this point, MEWR appears to be solid ground especially since the AQI to pollutant concentration calculator explicitly states that PM2.5 and PM10 readings are based on 24-hour averages.
That should settle it right? Not exactly. Given the size of the United States, I was curious to see if state or local level environmental bodies provided spot hourly concentration readings for PM10 or PM2.5 After all federalism requires that each state and local authority are left in charge of micro-managing their own affairs. EPA as a federal agency may not want to mandate that PM10 and PM2.5 pollutant concentrations are reported in spot hourly readings, but each city’s municipal authority may impose that as an additional reporting requirement to give residents a more timely picture.
Let’s take one of America’s largest states (New York) for example. New York state’s Department of Environmental Conservation provides its residents with a real-time hourly spot readings for PM10 and PM2.5 (where available for each monitoring station):
The NYSDEC Air Quality monitoring website allows a real-time view into the ambient air quality database of the New York State Department of Environmental Conservation. In general, data is polled at the top of each hour from each station. It is immediately displayed as it is collected.
For eg. hourly PM2.5 pollutant concentration readings are available for the city of Newburg here (hope the link works). The data provided is regarded too preliminary which necessitates the addition of a disclaimer at the bottom of every page reading:
The information used is the first available data from our air quality monitoring network. The values have not been verified for accuracy or been through the appropriate quality assurance and control validation procedures.
Of course, it isn’t just New York state of course that provides its residents with hourly readings. Another state, Oregon also provides real-time data which comes with the standard disclaimer that it hasn’t been validated as shown here:
What can we make of these couple of examples? Firstly while MEWR is technically and literally correct that the federal environmental agencies do not mandate anything more frequent than a 24-hour averaging time period for PM10 and PM2.5 concentration reporting, states can and do go the extra mile in providing real-time data available on their website. In the case of Singapore, which isn’t exactly a large country like the United States, there is essentially no difference between a federal agency and a state agency. NEA plays both roles. If state-level environmental agencies from Oregon and New York can provide real time spot hourly PM10, PM2.5 pollutant concentration updates, why not NEA?
Just to belabour the above point, let’s throw in another example. This time it’s the Environmental Protection Agency of Victoria state, Australia. The relevant page is here (click data readings to view current pollutant concentration values). Note that it explicitly says that PM10 pollutant concentration values are averaged over 1-hour (yup, not 24-hours or 3-hours) :
This bulletin is updated hourly with information calculated on data readings averaged over 8 hours for carbon monoxide and 1 hour for PM10, ozone, nitrogen dioxide, sulfur dioxide and visibility reduction. Detailed information on the calculation of the index is available below.
And of course, like all instant things it comes with a disclaimer that the data is preliminary and has not been vetted.
One last example. The Canadian province of Ontario also provides real time spot hourly PM2.5 pollutant concentration readings here:
Beautiful ain’t it? If these state-level agencies can provide real time spot hourly PM2.5 and PM10 pollutant concentration readings why can’t NEA do likewise? Food for thought.
It is worth noting that while the concentrations of most air pollutants are measured by a shorter averaging time (like the 1-hour average) for AQI / API calculations, particulate matter (PM) is averaged over a 24-hour period. This is due to the lack of scientific evidence with respect to the exposure-response relationship for PM over a one-hour period (Cairncross et al, 2007).
However, immediately following the quoted text is the following:
As a result, when PM is the dominant pollutant, the AQI / API system is not responsive enough to reflect a sudden surge in the level of PM, because the index is based on its concentrations averaged over the past 24 hours. There is inevitably a time lag between the rise in concentration recorded at the monitoring stations and the rise in AQI / API readings; this time lag will delay the issuance of health advisories for impending air pollution episodes. An example that highlights this problem is presented in Appendix 1.
What does this tell us? Does it not tell us that the study’s expert authors recognised that a 24-hour averaging time would cause the air quality index to lag behind actual pollutant concentrations which is similar to the critics’ argument that a 24-hour averaging period is too slow? It’s amazing how a little context easily debunks the commenter’s argument. Furthermore, Appendix 1 of the report goes on to illustrate how the organisers of a 10 km race had in Nov 2006 decided to let the race go ahead despite hazy skies on account of the API reading in the locality being a mere 52. As a result, 43 people suffered from discomfort with 5 being admitted to hospital for treatment. The study’s authors go on to note that, in complete opposite to the commenter’s position that:
It is logical to assume that the organizers would postpone or cancel the race if they are aware of the rapid rise of the RSP concentrations. However, the use of the hourly API at Yuen Long (which was 52 and representative of the past 24-hour average concentration of RSP) as evidence that the air quality of Yuen Long was acceptable at the time of the race was a mistake and probably contributed to this incident. Instead, the hourly RSP concentration reported at 8am would much better information on the air quality to the organizers.
Which is precisely what supporters of an hourly spot PSI reading are saying. The study as quoted actually supports their position.
An NUS PhD student, Jeremy Chen has done a bit of mathematical modelling and determined that NEA’s 3-hour PSI averaging cannot possibly be based on a simple moving average model because in mathspeak, the error rate (the estimated spot PSI values tend to oscillate) grows with time. He proposes a weighted average model which has been implemented on this website which appears, prima facie to work quite well.