Sunday, September 22, 2013

Population exposure to Particulate Matter (PM) pollution: PM10 & PM2.5 yearly and daily limits

Figure 1: Combined rural and urban concentration map of PM10 – 36th highest daily average values, year 2010 (from European air quality maps   ETC/ACM Technical Paper - Jan Horálek & al. March 2013)

The Figure above is showing that around 20% of European population had been exposed in 2010 for more than 35 days to Particulate matter (PM10) daily concentration greater than 50μg/m3 (red or brownish squares): Albany, Cyprus, Greece, Macedonia, Norway and Poland being  particularly concerned. 

Whereas during the same period, around 5% only of the same population had been exposed to yearly average concentration greater than 40

Why such a discrepancy between those two figures? Which either from daily or yearly PM10 concentration is more representative? How can we derive area concentration from specific measure in a station? How is defined the population exposure?

Those are the relevant question concerning PM air pollution that we want to address is the following posts.

PM is referring to a heterogeneous mixture of particles (solid and liquid) suspended in air, also known as atmospheric aerosols. PM mixture size and chemical composition change in time and space, depending on emission sources, atmospheric and weather conditions.     

PM originate partly from human or natural sources, partly from in situ or distance locations. The largest particles of concern are 10 microns or smaller in diameter (PM
10). But the groups of most concern are 2.5 microns or smaller in diameter (PM2.5). The coarse particles are those which size are greater than 2.5 microns and smaller than 10 microns.

This post concentrates on PM pollutant only as currently done by the scientific community, although we are exposed to a complex mixture of pollutants. But additional research is needed to understand and quantify the possible additive, synergetic or antagonistic effects between pollutants which are encountered simultaneously in the ambient air.

Sources of PM and their effects on health, environment and climate

PM sources 

PM in the atmosphere originates from primary particles emitted directly & secondary particles produced as a result of chemical reactions involving so-called PM volatile organic compounds (VOC).
Although the influence of emission sources on PM concentration is greatest near the sources, PM concentrations are influenced by atmospheric transport between countries. In Southeast Asia, haze spell are observed in Malaysia and Singapore generated during the dry season from open fires in Sumatra (see my posts dated 15 July 2013 & 14 Sep 2012). 
Studies have shown that dust aerosols from Asia Northeastern Taklimakan, Gobi & Badain Jaran deserts and the Loess Plateau are transported to vast downwind areas including large portions of China, Korean Peninsula, Japan , North Pacific, North America , and even the Arctic regions (see Studies on a severe dust storm in East Asia by Wiao-Xian Huang & al., 2012) .

Health effects 

Some PM are small enough to pass from the lung into the bloodstream just like oxygen molecules. Then they might cause or aggravate cardiovascular and lung diseases, heart attacks and arrhythmias, affect the central nervous system, the reproductive system and cause cancer. The outcome can be premature death.

Environmental effects

PM affect animals in the same way as humans. Dust aerosol lifted into atmosphere and deposited to oceans can affect global biogeochemical cycles as well as human health. They affect plant growth and ecosystem processes, cause damage to buildings and reduced visibility.

Climate effects

PM climate effect varies depending on particle size and composition: some might lead to net cooling (white PM), while others (brown or black) to warming. Can lead to changed rainfall patterns. Deposition can lead to changes in surface albedo.

PM limit standards for PM10 and PM2.5

The following Figure 2  is summing up important existing standards for both PM10 and PM2.5.  

Limit concentrations may refer to both yearly and daily averages.  

Figure 2: Air quality standards or PM limit values to avoid health issues

A 24-hour exceedance limit applies to the 90.4th percentile (cf. EU-PM10), which means that PM10 may be greater than 50μg/m3 but less than 35 exceedance days (9.6%) over a year. Or that the 36th highest daily average value is smaller than 50μg/m3 (cf. Figure 1 above).
The World Health Organization (WHO) air quality limits, shown in the above table, are stricter than the EU air quality standards. The WHO explains that it is necessary to achieve the lowest concentrations possible, because no threshold for PM has been identified below which no damage to health is observed.

To reduce issues due to the cumulative effect of PM breathing on human health, the total exposure days to a relatively small PM concentration limit is more relevant than the yearly average.

Relation between daily and yearly PM concentration

A log-normal distribution is a variable distribution whose logarithm is normally distributed.  A log-normal distribution law is used to describe random variable taking only positive values in all fields of science such as: latency periods of diseases, species’ abundancy, financial asset values, geological data, fruit and flower rises etc... These distributions follow a geometric Brownian motion with constant drift and volatility.

The measured daily average PM10 concentrations in a specific station during one year is also log-normally  distributed as seen in the following Figure 3 (see “Relation between daily exceedance and yearly averages for PM10”, Joost Wesseling &al., 2011).

Figure 3: An example of an individual concentration distribution of measured daily average PM10 concentrations in 2006 at a rural station in the Netherlands (from Joost Wesseling and alias 2011).

The two parameters μ (location) and σ (scale) determine the shape of the distribution. However, in a large number of cases, a shift (δ) (to the right) is necessary in order to satisfactorily describe the measured concentration distributions, as there is always a certain background concentration present.

Using the assumed log-normal distribution, the fraction of exceedance days having concentrations above 50μg/m3 less than 35 days over one year can be computed by integrating this curve distribution. In the curve Figure 4, this condition is met when Yearly PM10 is 31μg/m3.

If we want to have concentration above 150μg/m3 less than 1 day over 3 years (cf. Figure 1 USA PM10 standards), with same σ and σ that in Figure, this condition is met when Yearly PM10 is 35μg/m3.

Figure 4: Relation between the measured yearly average PM10 concentration and the number of days having average concentrations above 50 μg/m3. The curve labeled   is a lognormal curve with   δ=6 μg/m3 and σ=0.56.

The agreement is quite satisfactory with a ±5% confidence interval. Around 35 exceedance days per year, 5% amounts to an average underestimation of 1.75 days.So we can see that the daily limitation of PM10>50μg/m3 only 35 days per year is more stringent that EU yearly limitation (40μg/m3) but less tight than the WHO limit (20μg/m3) for two reasons: the shape of the lognormal curve and the value of the confidence interval.

For the distribution given above, the EU condition PM10<50μg/m3 up to the 90.4th percentile over 1 year (ou PM> 50μg/m3 only 35 days per year) is roughly equivalent to the USA condition PM10<150μg/m3 up to the 99.9th percentile over 3 years  (or PM10>150μg/m3 only 1 day per 3 years). If we look at the number of days- with log normal distributions having same sigma, delta as in the curve above- but variable yearly PM10: the 2 concentration day EU & US limits need  a yearly  PM10 in the range 31-35μg/m3, with EU limit being lightly more stringent to meet (yearly=31μg/m3) than  US limit (yearly =35μg/m3).