SOURCES, COMPOSITION AND HEALTH IMPACTS OF AIRBORNE PARTICULATE MATTER

Roy M. HarrisonUniversity of Birmingham

and UK National Centre for Atmospheric Science

United Kingdom

CONTENT

A few definitions Sources of particles Chemical composition, and what it tells us Implications for human health

PRIMARY AND SECONDARY PARTICLES

Primary particles are those emitted directly from a source

Secondary particles are those formed in the atmosphere from chemical reactionso sulphates from sulphur dioxide oxidationo nitrates from oxidation of oxides of nitrogeno secondary organic aerosol from oxidation of

volatile organic compounds

DEFINITION OF TERMINOLOGYAirborne particles are measured as:

o PM10 – approximates to particles less than 10 micrometres diameter, measured by weighing (µg m-3)

o PM2.5 – less than 2.5 micrometres diameter (known as fine particles) (µg m-3)

o PM10 minus PM2.5 – coarse particles (µg m-3)

o Nanoparticles (ultrafine) – less than 0.1 µm diameter, usually evaluated by counting (# cm-3)

Total PM2.5 emissions (kt), 1990-2009, 2015 and 2020

PM10-Overall (BCCS)

Iron-rich Dusts13.4%

NaCl9.3%

NH4NO3/NaNO318.5%

Carbonaceous Material31.7%

Other3.7%

(NH4)2SO416.0%

Calcium Salts7.4%

PM2.5-Overall (BCCS)

NaCl4.0%

Iron-rich Dusts5.9% Calcium Salts

2.5%

NH4NO3/NaNO321.2%

Carbonaceous Material37.3%

Other5.2%

(NH4)2SO424.0%

PM10 - Episode (BCCS)

NaCl2.6%

Calcium Salts3.8%

Iron-rich Dusts14.3%

Carbonaceous Material22.0%

(NH4)2SO418.2%

NH4NO3/NaNO339.0%

PM2.5-Episode (BCCS)

NaCl1.7%

Calcium Salts1.0%

Iron-rich Dusts6.3%

Carbonaceous Material22.6%

(NH4)2SO422.7%

NH4NO3/NaNO345.8%

(a)

(b)

RECEPTOR MODELLING Use of air quality data to infer the sources responsible for measured

pollution levels (opposite of dispersion modelling!)

Receptor modelling of airborne particles depends upon an assumption of mass conservation

Ci =

where Ci = airborne concentration of component, i= mass fraction of component i in particles from source, j= mass of particles from source j in an air sample

Analysis of many air samples for multiple chemical components is necessary

TYPES OF RECEPTOR MODELLING OF PARTICULATE MATTERThere are two main types• Chemical Mass Balance

- Requires only one air sample, although better results are obtained with more

- Requires knowledge of chemical composition of particles from each source ( )

- Varies for all chemical components to obtain best fit to mass conservation equation

• Multivariate Statistical- Principal Component Analysis widely used, but Positive Matrix

Factorization (PMF) has advantages and is more frequently utilised- Requires no advance knowledge of source chemical composition- Requires many separate samples, and identifies temporal

correlations of components (e.g. Na and Cl in sea salt) in a multidimensional space.

RECEPTOR MODELLING WITH CMB MODEL

Uses organic molecular markers and trace elements to apportion the carbonaceous component of PM2.5

Source apportionment of the entire PM2.5 is conducted using the Pragmatic Mass Closure Model

Results have been processed for winter air samples collected at LNK and HAR

NK Site During ClearfLo (1)

Daily PM2.5 Source Contribution Estimates with Secondary Biogenic Components at NK

Daily PM2.5 Source Contribution Estimates at HAR

0.0

5.0

10.0

15.0

20.0

25.0

30.0

35.0

Jan-

12Ja

n-13

Jan-

14Ja

n-15

Jan-

16Ja

n-17

Jan-

18Ja

n-19

Jan-

20Ja

n-21

Jan-

22Ja

n-23

Jan-

24Ja

n-25

Jan-

26Ja

n-27

Jan-

28Ja

n-29

Jan-

30Ja

n-31

Feb-

01Fe

b-02

Feb-

03Fe

b-04

Feb-

05Fe

b-06

Feb-

07Fe

b-08

Ove

rall

PM2.

5 C

ontr

ibut

ion

( mg

m -3

)

HAR

Other

AmmoniumNitrateAmmoniumSulphateSeaSalt

Other OM

Total Traffic

FoodCookingCoalCombustionDust/Soil

Natural Gas

Woodsmoke

VegetativeDetritusMeasuredPM2.5

SELECTED MEAN CONTRIBUTION TO PM2.5 MASS, µg m-3 (%)

How do we know that particles cause adverse health effects?

How large are those effects?

Air Pollution EpidemiologyCurrent understanding of the effects of air pollution on humanpopulations is mainly from:

• time series studies, which relate day-to-day changes in air pollutant concentrations to mortality and hospital admissions. Such studies inform us about the effects of short-term (24-hour) exposures to particles and other air pollutants.

• cohort studies, which relate long-term air pollutant concentrations in different cities to rates of death and causes of death (disease type).

Results of the Cohort StudiesThere have been few cohort studies of air pollution effects. Twohave been especially comprehensive and influential.

• The Harvard Six Cities Study, which showed a relationship between average concentrations of particulate matter and death rates (controlled for other risk factors) in six US cities. The study examined effects of both PM10 and PM2.5 and found a closer relationship with mortality for PM2.5.

• The American Cancer Society Study showed a relationship between PM2.5 and death rates for cardiovascular disease, respiratory disease and lung cancer in 50 US cities.

Adjusted Mortality Relative Risk (RR) Associated with a 10 µg/m3 Change in Fine Particles Measuring Less than 2.5 µm in Diameter

Adjusted RR (95% CI)* Cause of Mortality 1979-1983 1999-2000 Average All-cause 1.04 (1.01-1.08) 1.06 (1.02-1.10) 1.06 (1.02-1.11)

Cardiopulmonary 1.06 (1.02-1.10) 1.08 (1.02-1.14) 1.09 (1.03-1.16)

Lung cancer 1.08 (1.01-1.16) 1.13 (1.04-1.22) 1.14 (1.04-1.23)

All other causes 1.01 (0.97-1.05) 1.01 (0.97-1.06) 1.01 (0.95-1.06)

*Estimated and adjusted based on the baseline random-effects Cox proportional hazards model, controlling for age, sex, race, smoking, education, marital status, body mass, alcohol consumption, occupational exposure, and diet. CI indicates confidence interval. Source: Pope et al. (2002)

ESTIMATE OF EFFECT OF CHRONIC PARTICULATE MATTER EXPOSURE UPON MORTALITY

“The current (2008) burden of anthropogenic particulate matter is, with some simplifying assumptions, an effect on mortality in 2008 equivalent to nearly 29,000 deaths in the UK at typical ages and an associated loss of total population life of 340,000 life-years. The burden can also be represented as a loss of life expectancy from birth of approximately 6 months”.

Health Protection Agency (2010)

CHRONIC EFFECTS(LONG-TERM EXPOSURE)

Burden of disease attributable to 20 leading risk factors in 2010, expressed as a percentage of global disability-adjusted life-years for both sexes

What do I mean by ultrafine particles?

It is particles with one dimension < 100 nm, but either

o measured by mass (usually referring to an aerodynamic diameter) – PM0.1

o measured by number (usually referring to a mobility diameter)

PARTICLE SIZEDISTRIBUTIONMEASURED IN BIRMINGHAM

Particle Number, Surface Area and Mass

Measuring:Particle number reflects particles < 100

nanometres primarily

Particle surface area reflects mainly particles of 50-1000 nm

Particle mass reflects particles of > 100 nanometres (usually to 2.5 µm or 10 µm)

Complete Dataset UNECE-EuropeTransport sector contributes about 75% to PN emissions

Source: Hugo Denier van der Gon

Spatial distribution

Source: Hugo Denier van der Gon

SHORT TERM (ACUTE) EFFECTS- Elucidated through the “time series” epidemiological studies

which relate day-to-day changes in air pollutant concentrations with changes in daily mortality and hospital admissions, frequently with a lag of up to three days between cause and effect.

- Department of Health Committee on the Medical Effects of Air Pollutants (COMEAP) has made quantification estimates of mortality and hospital admissions resultant from exposure to PM10, sulphur dioxide, nitrogen dioxide and ozone.

- Biggest uncertainty is attached to who is dying prematurely and the extent of loss of life expectancy.

Richard Atkinson and Ross Anderson (St. George’s Hospital Medical School) used case-crossover time series methodology to estimate the percentage increase in a given health outcome corresponding to the inter-quartile range (75%ile minus 25%ile) of concentration for several particle metrics

London Epidemiological Study

Urban Ambient Particle Metrics and Health: A Time Series Analysis, R.W. Atkinson, G.W. Fuller, H.R. Anderson, R.M. Harrison and B. Armstrong, Epidemiology, 21, 501-511 (2010).

Respiratory Mortality (lag 2)(Graph shows % change between 25%ile

and 75%ile concentration and 95% CI)-2

02

46

8%

(95%

CI)

PNCNO3 CL

SO4 BS

GR PM10

NK PM10

PM10

GR PM25

PM25

GR PM10

2.5PRI10

NONP10

NONP2.5

NONP102.5

-20

24

68

% (9

5% C

I)

PNCNO3 CL

SO4 BS

GR PM10

NK PM10

PM10

GR PM25

PM25

GR PM10

2.5PRI10

NONP10

NONP2.5

NONP102.5

Cardiovascular Mortality (lag 1)(Graph shows % change between 25%ile

and 75%ile concentration and 95% CI)

-20

24

6%

(95%

CI)

PNCNO3 CL

SO4 BS

GR PM10

NK PM10

PM10

GR PM25

PM25

GR PM10

2.5PRI10

NONP10

NONP2.5

NONP102.5

-20

24

6%

(95%

CI)

PNCNO3 CL

SO4 BS

GR PM10

NK PM10

PM10

GR PM25

PM25

GR PM10

2.5PRI10

NONP10

NONP2.5

NONP102.5

CONCLUSIONS Airborne particles in the UK have complex chemical

composition reflecting a mix of primary and secondary sources.

Poorly quantified and little regulated sources contribute significant to particulate matter mass.

Fuel combustion sources dominate the ultrafine particle fraction, with sulphur content being crucial.

Airborne particle exposure has a major impact upon morbidity and mortality in the UK.

There is some, limited, evidence that ultrafine particles are primarily responsible for effects upon cardiovascular health outcomes, with larger particles causing respiratory effects.

THANK YOU