The Georgia Basin-Puget Sound Airshed Characterization Report 2014: chapter 8

8. Particulate Matter

Roxanne Vingarzan (Environment Canada), Robert Kotchenruther (Environmental Protection Agency Region 10), Rebecca Saari, Sarah Hanna and Rita So (Environment Canada)

Particulate Matter (PM) consists of airborne particles in liquid or solid form which can have many different origins and diverse chemical properties. PM is usually characterized by its physical size or diameter, ranging from nanometers (nm) to several tens of micrometers (µm). Of particular interest to this report are particles less than 2.5 µm, which are commonly referred to as PM_2.5 or fine particulate matter.  These fine particles are of special concern for two reasons. First, these particles are able to penetrate deep into the human respiratory system and can be absorbed into the blood where they have been shown to cause biological effects. Second, these particles scatter light very efficiently and therefore play a major role in visibility impairment, a subject addressed in Chapter 9. Some forms of PM are emitted directly into the air (primary PM), including fugitive dust, sea spray, soot from wood smoke, forest fires and fossil fuel combustion. PM can also form in the atmosphere (secondary PM), as a result of reactions involving compounds such as nitrogen oxides (NO_x), sulphur dioxide (SO₂), volatile organic compounds (VOCs) and ammonia (NH₃). This chapter reviews the various constituents that make up PM, the contribution of natural sources and background levels to ambient PM concentrations, the trends and variations in PM levels and identifies areas of concern in the airshed.

8.1 Classification of Ambient PM by Particle Size

As previously mentioned, particulate matter is usually characterized by its physical size and composition. Figure 8.1 illustrates a distribution of fine particulate matter by number count and mass. Particles with a diameter less than 10 μm are classified as PM₁₀, which, for reference, is smaller than a human hair. Particles with diameters between PM₁₀ and PM_2.5 are defined as the coarse fraction. PM_2.5 includes all particles with diameters less than 2.5 μm, also known as fine particles. The smallest particles, ultrafines, have diameters <0.1 um. Ultrafines, which are the most numerous but have the smallest mass and volume, are the subject of recent investigations into health effects.

The residence time of particulate matter in the atmosphere depends on its size. Coarse particles settle out very quickly by gravity. Smaller particles have atmospheric lifetimes on the order of a few days and are eventually removed from the atmosphere by wet deposition in rain, snow or fog, or by dry deposition through gravity or surface turbulence (see Chapter 12, “Atmospheric Deposition and Ecological Effects”).  During their residency in the atmosphere, particles are transported away from their sources and undergo various physical and chemical transformation processes.

 

Nucleation takes place when it is more energetically feasible for a molecule to be in the liquid or solid phase and when there is an insufficient number of particles to scavenge the molecule of interest.  This can occur when hot gases cool and condense, when reactions in the gas phase produce species of lower volatility or species are very hygroscopic.  In each of these cases clusters of molecules lead to the production of a new particle, in a process known as gas-to-particle conversion. Most Aitken nuclei start as primary particles and grow by condensation or by coagulation. Coagulation occurs as a result of collisions between particles to form larger particles. Coagulation effectively increases the mass of particles while depleting smaller particles, resulting in a shift in the aerosol-size distribution toward larger particle sizes. An example of particles which form by gas-phase reactions is the reaction of H₂SO₄ with H2O to form hydrates of sulphuric acid (H₂SO₄·NH₂O), thus transferring gaseous material to the particulate phase (Seinfeld and Pandis, 1998).

Particles may also grow through condensation by coming into contact with reactive vapours which condense onto the surface of the particle to form a larger secondary particle. In the presence of high relative humidity, water vapour will condense onto small particles to form cloud droplets. This aqueous phase enables chemical reactions such as the conversion of SO₂ to SO₄⁼ to take place (NARSTO, 2004).

The volume or mass distribution in Figure 8.1 is dominated by two modes, the accumulation mode (from around 0.1 to around 2 μm) and the coarse mode (from around 2 to around 50 μm). The distribution shows a maximum around 0.6 to 0.8 μm in diameter, with a second peak near 6 μm. Accumulation-mode particles result from primary emissions, formation of secondary sulphate and nitrate aerosols, condensation of organics from the gas phase and coagulation of smaller particles. Particles in the coarse mode are usually produced by mechanical processes, such as wind or erosion (dust, sea salt, pollens, etc.) (NARSTO, 2004).

8.2 Sources and Formation of Secondary PM

The formation and composition of fine particulate matter is related to its emission sources. Primary particles are emitted directly into the atmosphere, whereas secondary particles are formed within the atmosphere.

Primary particulate matter is emitted by sea spray, blowing dust, tree pollens, forest fires, fossil fuel combustion and construction activities. These primary particles are found in both the fine (PM_2.5) and the coarse fractions (PM_2.5 to PM₁₀). Primary particles of natural origin tend to be in the coarse fraction, while anthropogenic primary particles can be either fine or coarse. 

Most secondary particles in the PM_2.5 fraction are formed from gaseous precursors through physiochemical oxidation processes. The gases most commonly involved are SO₂, NO_x, NH₃ and VOCs and the most common oxidants are OH, O₃, and NO₃. This section describes the key sources and formation mechanisms of particulates associated with SO₂, NO_x, NH₃ and VOC precursors.

8.2.1 SO₂

The oxidation of sulphur dioxide (SO₂) to sulphate (SO₄⁼) is an important chemical transformation in the formation of fine particulate matter (NARSTO, 2004). Natural sources of SO₂ include volcanoes, marine bacteria and wetlands, although the dominant source is the combustion of fossil fuels (Makar, 2001). Sulphate is usually created through the chemical reaction of SO₂ with the hydroxyl radical (OH) in air, aqueous oxidation reactions after the dissolution of SO₂ in cloud, fog or rain water, and the oxidation of SO₂ in reactions within the aerosol particles themselves. The first process dominates in the daytime but is negligible at night (NARSTO, 2004). The aqueous-phase oxidation that produces SO₄⁼ in cloud water can occur very quickly, often converting all of the available SO₂. It has been reported that cloud and fog water can be responsible for producing as much as half of the SO₄⁼ (McHenry and Dennis, 1994; Langner and Rodhe, 1991), causing elevated concentrations of sulphate (Pandis et al., 1992). Particles are formed in this aqueous process after the cloud water has evaporated, leaving particulate sulphate in the form of sulphuric acid (H₂SO₄).  Sulphuric acid is very acidic and is one of the main components of acid precipitation.  Ammonia (NH₃), the only basic gas phase species, will readily combine with H₂SO₄ to form ammonium sulphate ((NH₄)₂SO₄). This is the fully neutralized form of the particle; however, if insufficient NH₃ is available, ammonium bisulphate (NH₄HSO₄) is formed, resulting in a particle that is still acidic.

8.2.2 NO_x

Oxides of nitrogen (NO, NO₂) are formed biologically in soil, by lightning, and through reactions with oxygen in high-temperature combustion. The nitrate radical (NO₃), ozone and the OH radical are involved in reactions during the daytime. Although NO_x (NO+NO₂) reacts with ozone to produce the NO₃radical throughout the day, much of it is quickly destroyed via photolysis during the daytime (Makar, 2001). However the lack of photochemistry allows the buildup of NO₃ at night.  This reaction is important in the overnight reduction of ozone concentrations. These reactions alter the concentrations of the three species (NO₃, O₃, OH) that are the main oxidizing agents leading to the formation of aerosols (Makar, 2001). Another important chemical reaction involves the conversion of NO_x (NO+NO₂) to nitric acid (HNO₃). This occurs in reactions between NO₂ and OH during the day to form HNO₃, and between NO₂ and the NO₃ radical to form dinitrogen pentoxide (N2O5) at night. N2O5 undergoes hydrolysis to form HNO₃. HNO₃ equilibrium in the atmosphere between the gaseous and liquid phase is mostly a function of ambient temperature, where warmer temperatures results in more off-gassing of aerosol HNO₃. Nitric acid will react with NH₃, sea salt and dust, although the production of NH₄NO₃ is the most likely outcome when sufficient NH₃ is available (NARSTO, 2004). The latter is also enhanced by lower temperatures and high relative humidity. Particles resulting from the NH₃ pathway are in the fine fraction, whereas particles formed with sea salt or dust can also be found in the coarse fraction. More detailed information on the chemical reactions forming particulate matter and aerosols is provided in Makar (2001).

8.2.3 NH₃

Nitrogen in the form of ammonia (NH₃) reacts with both HNO₃ and H₂SO₄ to produce fine particulate matter. These reactions have been described previously. It is important to point out that the reaction between NH₃ and HNO₃ forming NH₄NO₃ depends on the relative concentrations of NH₃ and SO₄⁼. Generally the available ammonia preferentially reacts with SO₄⁼ first, since it is more acidic than HNO₃ (although it does not always completely neutralize all available SO₄⁼), before any excess becomes available to the NO₃^-neutralization reaction (Blanchard and Hldy, 2003). The nitric acid pathway prefers lower temperatures and higher relative humidities to enhance particle formation (NARSTO, 2004). Ammonia emissions play an important role in the temporal and spatial distribution of fine particulate matter (Barthelmie and Pryor, 1998).

8.2.4 Organics

The organic portion of fine particulate matter is a complex combination of hundreds of different organic compounds (NARSTO, 2004). Primary organic aerosols are emitted directly into the atmosphere from both natural and anthropogenic sources. Compounds that are important to the formation of organic aerosols are gaseous, semi-volatile and non-volatile organic compounds (Makar, 2001). The reactions and processes that lead to the formation of aerosols is complex. To move from the gaseous to the liquid phase, the organic compounds must be either semi- or non-volatile. Gaseous organics can produce products from oxidation reactions that are semi or non-volatile. Hence, many organic compounds may participate in the formation of particles. Particles that are formed in the atmosphere from organics are called secondary organic aerosols (SOAs). The ability of VOCs to form SOAs depends on many of the same conditions required to produce other secondary particles: temperature, relative humidity, concentration of individual VOCs, reactivity, availability of oxidants and surface area, and the volatility of products (NARSTO, 2004).

8.2.5 Natural Sources of PM

Windblown dust is the most common natural source of primary particles. However, sea spray, volcanoes, vegetative pollen and wildfires are also natural sources of primary particles. Organic compounds released from vegetation (biogenics) may be in the form of primary particles or they may react with other compounds to produce SOAs, as described earlier. This process, in which biogenic organic compounds produce fine particles, is very important in areas that are dominated by vegetation and can account for a large portion of the sub-micrometre particles.

Marine environments, volcanoes and wetlands are natural sources of sulphur and other chemical compounds that enter into reactions to form secondary particles. The saltwater areas within Georgia Basin/Puget Sound are examples of a natural source producing sulphur. A study by Norman et al. (2004) found that 30 per cent of the non-sea salt sulphate captured in precipitation over the southern Gulf Islands and up to 13 per cent of the aerosol sulphate in the Lower Fraser Valley (LFV) came from the oxidation of dimethyl sulphide from algal emissions. This estimate is in agreement with the biogenic contribution estimated for the Georgia Basin during peak phytoplankton bloom periods (Sharma et al., 2003). These natural sources provide a source of sulphate that reacts with other air pollutants, NH₃ in particular, to form fine PM.

A portion of ambient PM₁₀ and PM_2.5 present in the Georgia Basin/Puget Sound airshed originates from long-range transport and is comprised of both soil dust and anthropogenic pollution. On average, this contribution is estimated to be in the order of 1.5 - 2 µg/m³ of the annual average PM_2.5 concentrations (McKendry, 2006). Occasionally Asian dust events can cause regional PM episodes. In 1998, an Asian dust event produced PM₁₀ concentrations in excess of 65 μg/m³ over the entire area from late April to early May, when normal levels would be 20 μg/m³(U.S. EPA, 2002). Chilliwack, in the eastern portion of the LFV, recorded PM₁₀concentrations of 120 μg/m³ and PM_2.5 concentrations of 44 μg/m³ (Suzuki and Taylor, 2003) during this period. Large scale long-range transport dust events have covered 15 to 30% of the area of the continental United States and have been shown to increase PM_2.5levels by 8.7±2.3 μg/m³ on average during such events (U.S. EPA, 2002).

8.2.6 Background PM

Background levels of particulate matter describe the portion of the PM from sources that are not within the control of local jurisdictions. McKendry (2006) reviewed studies of background PM relevant to British Columbia and identified the following major candidate sources:

• Wind-blown local crustal material. The most well-known are the winter outflow events in the eastern Lower Fraser Valley when strong winds (8-10 m/s) mobilize riverbed materials (McKendry, 2000).
• Local forest fires (predominantly during summer) that affect mainly interior regions of B.C. but can have an impact on coastal air quality and neighbouring areas (including Washington State).
• Trans-Pacific transport of dust (McKendry et al. 2001) arising from severe spring dust storms in Asia. Due to the long distance from source, the mean particle diameter for this material is approximately 2.5 μm and can thus contribute significantly to PM_2.5 (Husar et al. 2001).

The impact of episodic events of these sources on background PM levels are summarized in Table 8.1. The contribution of long-range transport will be explored further in Chapter 11,”Transboundary Transport”.

Notes:

¹ McKendry et al, 2001; Husar et al. 2001
² U.S. EPA (2002) suggest an increase of 8.7± 2.3 μg/m³ during dust events, with mean maximum dust contributions of 19.7± 8.4 μg/m³

 

Several studies have estimated PM background levels in regions within or near the Georgia Basin/Puget Sound airshed. Dann et al.(2011) reported mean and median values of 3.6 and 3.2 µg/m³, respectively, for the Saturna Island regional site in the Georgia Basin. This was based on an analysis of air parcel trajectory clusters associated with the lowest 6-hour average PM_2.5concentrations at the 95^th percentile between 2003 and 2006. McKendry (2006) estimated a mean background concentration in air masses arriving in British Columbia from north Pacific trajectories to be 1.5 - 2 µg/m³. Also on the Pacific west coast, Jaffe et al. (2005) estimated a combined marine/Asian PM_2.5 background of 1.5 µg/m³ (median) and 2.1 µg/m³ (mean) at Crater Lake, Oregon. Vingarzan (2004) reported a western North American background range of 1-4 µg/m³.  Long-term temporal trends in background PM are not available due to lack of data.

8.3 Annual, Seasonal and Diurnal Concentrations of Ambient PM_2.5

8.3.1 PM_2.5 Concentrations and Temporal Cycles

The annual average PM_2.5 mass concentrations in rural portions of the Georgia Basin/Puget Sound airshed vary from 2 to 4 μg/m³. Near the major urban centres of Seattle, Victoria and Vancouver, PM_2.5 annual averages are 7 to 8 μg/m³. Annual average PM_2.5 TEOM-derived concentrations in the Lower Fraser Valley range between 5-6 μg/m³ (NAPS, 2006). 
In the Lower Fraser Valley, summertime PM_2.5 episodes sometimes occur during stagnant meteorological conditions conducive to secondary aerosol formation.  In addition, transported smoke from forest fires in British Columbia and elsewhere can also contribute to locally observed PM_2.5 episodes.

Cold season PM_2.5 episodes occur in the fall or winter months when emissions from wood burning and space heating add to the existing PM load, coupled under stagnant meteorological conditions and lower mixing heights. Elevated PM_2.5 levels at some locations in the LFV, including Vancouver Airport and Pitt Meadows can be caused by wood smoke in nearby residential areas (Larson et al., 2007). Similarly, peak average annual concentrations (98^th percentile) in Seattle and Tacoma are usually observed in the fall and winter, reaching levels of 28-31 μg/m³ in Seattle and 43-46 μg/m³ in Tacoma.

During episodic conditions, 24 hour maxima typically exceed 15 μg/m³ and occasionally exceed 20 μg/m³ in the Lower Fraser Valley (Brook et al., 2011). Although rare, there have been occasional episodes where PM_2.5 concentrations have reached extreme levels. The most recent such event occurred at the time of the Burns Bog fire in the summer of 2005. On September 19, a blanket of smoke covered the LFV, causing hourly PM_2.5 concentrations to exceed 274 μg/m³ at Burnaby South and 24 hour average PM_2.5concentrations to exceed 40 μg/m³ in some locations. As a result of forest fires, summertime smog, winter inversions and localized fireworks activity PM_2.5 levels occasionally exceed Metro Vancouver’s 24 hour objective of 25 μg/m³ (Metro Vancouver, 2010).

Weekly variation in PM_2.5 in the Georgia Basin is generally small (Suzuki and Taylor, 2003).  The same is true of daily PM_2.5, although several sites show evening peaks, and some show traffic-related peaks.  Suzuki and Taylor (2003) found a diurnal pattern consisting of small peaks in the morning and evening, and mass concentrations varying only a few μg/m³ from the mean daily value. More recently, for the year 2006, Brook et al. (2011) reported a morning rush hour peak and a more pronounced evening peak at the Kitsilano site, which is located in a dense urban area near Vancouver’s downtown. The morning PM_2.5 peak was not observed at the more rural sites of Hope and Langley. The shallow nocturnal boundary layer kept average hourly PM_2.5 levels elevated overnight at all sites by concentrating local emissions (Brook et al., 2011).

8.4 Ambient PM_2.5 Levels Compared to Standards and Objectives

Ambient standards for PM_2.5 provide a regulatory baseline for assessing ambient concentrations. PM_2.5 concentrations in the Georgia Basin/Puget Sound are usually below national objectives and standards in most regions, with a notable exception in the Puget Sound.

Figure 8.2 shows that Tacoma PM_2.5 has been exceeding the 24-hour NAAQS of 35 μg/m³ for the entire period of record. For Marysville, PM_2.5 exceeded the 24-hour NAAQS but has declined and reached borderline attainment since 2005. Regional average levels in Pierce County also exceed this standard; its highest 98^th percentile 24-hour PM_2.5 levels were over 41 μg/m³ between 2005 and 2008 (U.S. EPA, 2010). On December 14, 2009, EPA formally designated the Tacoma-Pierce County as a non-attainment area for the 24-hour PM_2.5 NAAQS (WA DOE, 2010a).

 

Notes: 3 year running means of highest 98^th percentile from daily data (AIR Station 53-061-1007 for Marysville)
(AIRS Station 53-033-0080 for Seattle) (AIRS Station 53-053-0029 for Tacoma)

Other areas in the Georgia Basin/Puget Sound remain well below their respective national PM_2.5 standards. Figure 8.2 shows Seattle to be in attainment of the U.S. National Ambient Air Quality Standard. Similarly, PM_2.5 levels at four sites in the Lower Fraser Valley (Kitsilano, Pitt Meadows, Abbotsford, and Chilliwack) are well below the Canada-Wide Standard (Figure 8.3).

8.5 PM_2.5 Trends

National-scale PM_2.5 monitoring efforts are relatively recent in both countries, so the period of record is too short to develop long-term trends at many sites. Trend analysis is further complicated by changes to PM monitoring methods, namely the addition of dryers to TEOMs. Nonetheless, PM_2.5 concentrations at most sites appear to be on a slow decline (Figure 8.2 and Figure 8.3), with the exception of Tacoma. At Mount Rainier in the Puget Sound (not shown), WA DOE (2002) found a statistically significant decreasing trend in PM_2.5of 0.32 μg/m³ per year from 1988-1999. Similarly, 24-hour average PM_2.5 data from Rocky Point in the Georgia Basin (not shown) indicate a statistically significant (p<0.05) decreasing trend of 0.42 μg/m³per year between 1993 and 2004. Recent analyses of particle backscatter (a surrogate for PM_2.5 measurements) in the Lower Fraser Valley also indicate statistically significant declining trends since 2002 (So and Vingarzan, 2010).

Notes: 3 year running means of highest 98^th percentile based on 24-hour average data (NAPS Stations, west to east: 100118, 101202,101004, 101101)
Pre-2004 data was collected without a dryer and is thus shown separately.

8.6 Compositional Analysis of fine PM

Understanding the composition of PM is the key to understanding its impacts on human health and the environment. Particulate matter can carry toxins and metals and variably affect the climate, ecology and visibility. Thus, particle composition can be as important as concentration when it comes to understanding and assessing ambient PM_2.5impacts.

In Figure 8.4 , the relative concentrations of ammonium, sodium, sulphate and nitrate in PM_2.5 are shown along with the measured particle size for two sites in the WISE (note that particle size decreases from left to right). Sodium nitrate dominates the coarse fraction (>2.5 um), whereas ammonium sulphate and, to a lesser degree, ammonium nitrate appear in the finer fraction <2.5 um). This is primarily due the formation of coarse mode sodium nitrate following the reaction of nitric acid and sea salt (NaCl).

 

Notes: (1) Ten-stage micro-orifice uniform deposit impactors (MOUDI-110) were deployed at Langley and Sumas Mountain. Each stage has a successively smaller cut-point diameter to enable particle size classification. (2) Ion balance is a means of checking the accuracy of analytical results. Departures from a ratio of anions to cations of 1.0 is a measure of error in the analysis or a missing ion.

A compositional analysis of PM_2.5 in the Georgia Basin/Puget Sound airshed is depicted in Figure 8.5. Organic carbon, ammonium nitrate and ammonium sulphate are the major components at all sites. To compare the four sites, which range from the remote Mount Rainier to urban Seattle, first consider the range in annual average PM_2.5 concentrations. Seattle has the highest annual PM_2.5 level (7.2 μg/m³), and Mount Rainier has half the level of any other site (3.1 μg/m³). Urban-residential Burnaby and rural Abbotsford have commensurate levels, which are similar to but less than those of Seattle. 

There are some notable differences among the composition of PM_2.5 at these sites. Ammonium sulphate is most important at the Seattle site, possibly due to higher local industrial and marine emissions. Mt. Rainier has a high sulphate component as well, which is likely attributable to transported sulphate and precursors; however, its total sulphate levels are much lower than Seattle’s, on a mass basis (0.8 μg/m³ compared to 1.6 μg/m³).

The total mass of sulphate at the Burnaby and Abbotsford sites are very similar however due to the presence of high concentrations of NH₄, due to intense agricultural activity, the Abbotsford site has considerably higher nitrate present.  High ammonia levels enhance ammonium nitrate formation, which in this airshed is limited by the availability of oxidized nitrogen species, not ammonia.

 

Notes for Puget Sound RTM:

Average for all days from 2004-2008
ANO₃ = ammonium nitrate = 1.29*NO₃
ASO₄ = ammonium sulphate = 1.375*SO₄
EC = elemental carbon
OMC = organic matter = 1.8*OC (for rural locations) or 1.4*OC (for urban locations)
Sea Salt = sodium chloride = 1.8*(Chloride)
Soil = 2.2*Al + 2.49*Si + 1.63*Ca + 2.42*Fe + 1.94*Ti
Trace = sum of all other trace constituents

8.6.1 Seasonal Variations in PM_2.5 Composition

The annual average compositions shown in Figure 8.5 also vary seasonally. These variations are shown in Figure 8.6 and Figure 8.7for the Puget Sound and the Georgia Basin, respectively. The seasonal patterns in the urban and rural sites in Figure 8.6 differ noticeably, both with regard to total mass and composition. Total mass peaks in the fall and winter for urban locations like Seattle, reflecting the increased frequency of reduced mixing depths and stagnant air episodes coupled with higher levels of direct PMemissions from wood smoke. Residential wood smoke is an important source of PM_2.5in Seattle (responsible for 51% of King County’s emissions in 2005), and its emissions are known to peak during the colder months (WA DOE, 2007). Burnaby South, in the Georgia Basin, is an urban-residential site showing a similar but less pronounced pattern with PM_2.5 concentrations peaking in autumn and winter (Figure 8.7). Conversely, rural Mt. Rainier has higher PM_2.5 concentrations in summer, likely reflecting biomass burning (e.g. forest fire plumes) and SOA production from oxidation of biogenic VOCs.

Abbotsford, a site located in an agricultural area of the LFV, differs in its seasonal pattern from the urban and remote sites already mentioned. At Abbotsford, daily PM_2.5 concentrations are lowest in the winter, are higher in spring and reach a maximum in the fall with a notable increase in the organic matter fraction (Figure 8.7). Note that this site experiences distinctly different wind patterns in the winter, when winds are mostly northeasterly and pass across an upwind residential area. In the summer, winds are southwesterly and blow from the direction of agricultural fields and Whatcom County in Washington State.

Among the four sites shown in Figure 8.6 and Figure 8.7, several seasonal patterns in PM composition are apparent. Sulphate concentrations peak in the summer due to higher rates of photochemical conversion of SO₂ to sulphate, increased biogenic sulphur emissions and increased boating/cruise vessel activities. At the more urban sites of Burnaby and Seattle, maximum nitrate concentrations occur in winter because nitrate deposition to particles is favoured under colder temperatures and higher humidity. At Abbotsford, nitrate deposition to particles increases in the summer months, likely due to higher ammonia availability from large emissions during the summer from agriculture.

 

Notes: Based on 24-hour average concentrations from 1-in-3 day sampling for the years 2004-2008. IMPROVE station for Seattle (53-033-0080) and Mt. Rainier (53-053-0014). See notes for Figure 8.5 for reconstruction method and definition of constituents.

Notes: Based on 24-hour average concentrations from 1-in-3 day sampling for the years 2003-2008. NAPS station #100119 (Burnaby) and #101004 (Abbotsford). See notes for Figure 8.5 for reconstruction method and definition of constituents.

The seasonal variations in PM_2.5 concentrations and composition shown in Figure 8.6 and Figure 8.7 are merely demonstrative and are not representative of the entire Georgia Basin/Puget Sound airshed. Seasonal patterns are likely to vary throughout the airshed depending on nearby sources and can be better understood for a given site through PMsource appointment.

8.7 PM Source Apportionment

Once the chemical composition of particulate matter has been analysed, the relative importance of the various sources involved in its formation can be determined through statistical methods such as Positive Matrix Factorization (PMF), Principal Component Analysis (PCA) and the Conditional Probability Function (CPF).  

Several recent applications of PMF and CPF have been undertaken using speciated PM_2.5 data collected by speciation samplers from NAPS and IMPROVE stations in the Georgia Basin/Puget Sound airshed. This includes two sites in the Georgia Basin, Burnaby South and Abbotsford (So et al., 2010) and two in the Puget Sound, Tacoma (WA DOE, 2010b), and Mount Rainier (Rose, 2006).  Burnaby South, an urban-residential site in Metro Vancouver, is impacted by traffic, mixed industry and marine emissions.  Abbotsford, a semi-rural site in central Lower Fraser Valley, is impacted by significant ammonia emissions, vehicle traffic and growing urbanization. Mount Rainier is a remote site in the Puget Sound, impacted by the transport of anthropogenic sulphur emissions, biogenic VOCs and biomass burning. Tacoma is an urban residential site neighbouring a major freeway and is influenced by downtown Tacoma, the Port of Tacoma, an Air Force base, an Army Base and several major industries (WA DOE, 2010b).

8.7.1 PM Sources in the Georgia Basin

Figure 8.8 and Figure 8.9 present PMF source factor contributions to PM_2.5 mass for both summer and winter seasons for Burnaby South and Abbotsford, respectively (So et al., 2010). Dominant sources contributing to PM_2.5 are motor vehicles, nitrate (linked to mixed combustion, but dominated here by transportation and agriculture) and sulphate (industrial, shipping and marine biogenic sources). In the Greater Vancouver area, SO₂ emissions from marine traffic and oil refining contribute to particulate sulphate. 

Source contributions vary by season at both sites. Wood burning, secondary nitrate, and vehicle emissions are more important in the winter, while fuel oil and marine biogenic influences are more important in the summer. Note that the marine oil combustion fraction (driven by the 3:1 Ni/V ratio tracer) is larger in the summer, during the cruise ship season. Secondary sulphate is higher during the summer due to a combination of increased SO₂oxidation and biogenic emissions. Marine biogenic emissions in the form of dimethyl sulphide are also significant during the spring and summer, contributing as much as 26% of total sulphur emissions in the airshed during algal bloom periods (Sharma et al., 2003). Most of the nitrate in the urban core comes from oxidation of NO_x from vehicle emissions, marine and industrial sources. Secondary nitrate contributions are lower in the summer due to preferential uptake on the particle of ammonium sulphate over ammonium nitrate, the uptake of nitric acid on coarse mode sea-salt particles and ammonium nitrate aerosol equilibrium being favoured under cold temperatures and high humidity.

 

Notes: Based on PMF analysis with EPA PMF 3.0 of 24-hour average concentrations from 1-in-3 day sampling for the years 2003-2008. NAPS station #100119. (So et al., 2010)

Notes: Based on PMF analysis with EPA PMF 3.0 of 24-hour average concentrations from 1-in-3 day sampling for the years 2003-2008. NAPS station #101004. (So et al., 2010).

At Abbotsford (Figure 8.9), secondary sulphate and oil combustion are dominant in the summer, and secondary nitrate and vehicle emissions dominate in the winter.  Contributions from wood burning and crustal elements are also more important in the winter. Note that sea salt is overestimated at both sites due to factor separation uncertainties.

A source analysis of speciated data from Abbotsford for the period April 1994 to May 1995 estimated mobile sources at 39% (diesel 3%; gasoline 36%), secondary nitrate at 29% and secondary sulphate at 13% (Ostermann, 2002). For the more recent period of 2003-2008, So et al. (2010) estimated vehicle emissions at 21%, secondary nitrate at 24% and secondary sulphate at 12%. Methodological differences and uncertainties must be considered when comparing Ostermann’s results to So et al. (2010). Nonetheless, their results are similar, and the decrease in the relative contribution from vehicles in 2003-2008 (So et al., 2010) compared to 1994-1995 (Ostermann, 2002) is in line with the decrease in smog-forming emissions from vehicles as shown in Figure 5.9 (Chapter 5, “Emissions”).

8.7.2 PM Sources in Tacoma in the Puget Sound

Tacoma

WA DOE (2010b) performed a PMF study for Tacoma using similar methods to So et al. (2010). The data were collected from January 2006 to May 2009 using the Speciation Trends Network (STN) sampling protocols in South Tacoma, Washington. The source apportionment analysis was conducted using Positive Matrix Factorization (PMF) (Paatero and Tapper, 1994; Paatero, 1997) and the Conditional Probability Function (CPF) (Ashbaugh et al., 1985).

The sources contributing to PM_2.5 at the Tacoma site, from most to least significant, were as follows: wood smoke (45%), secondary particles (25%), motor vehicles (13%), industrial emissions (6%), fugitive dust (4%), fresh and processed sea salt (4%), fireworks (2%), and oil combustion and ships (1%). Wood smoke was the single most important source of PM_2.5 in Tacoma, contributing about half of the PM_2.5 on most days in the winter, as well as on days when PM_2.5 levels were the highest. Since outdoor burning is prohibited near the site, the smoke origin is likely from residential wood stoves. 

Pronounced seasonal variations were observed for wood smoke, fireworks, secondary nitrate, motor vehicles, industrial emissions and fresh sea salt. Most of the fugitive dust and fireworks contributions occurred in the summer, while most of the wood smoke contributions occurred in the winter. Average contributions from motor vehicles and industry were highest in the winter probably due to the low atmospheric mixing heights. The increased number of cold-starts may also be a factor in the winter.

Although the annual average contributions from wood smoke and local industry appear to have declined each year since 2006, a similar decline was not observed with other sources or with the annual 98^th percentile contributions from wood smoke.

Mount Rainier

At Mount Rainier, the dominant components of PM_2.5 are organics and sulphate (Figure 8.5), which Rose (2006) attributed largely to biomass and secondary sulphate, respectively. Rose’s analysis compared the periods of 1991-1995 and 2000-2003 to explore multi-year trends and seasonal variations in PM sources at the remote site. The compositional analysis of the average and 90 percentile daily PM_2.5 from Mount Rainier for the 1991-1995 and 2000-2003 time periods is presented in Figure 8.10. Biomass (organic carbon) contributed the largest amount of fine particulates for both time periods (35%-40% of the reconstructed total mass). This biomass contribution was likely composed of two main sources: (1) local or transported biomass smoke (from outdoor burning, wood-stoves, wild fires and other sources); and (2) secondary organic aerosols (SOAs) produced by oxidation of biogenic volatile organic compounds (VOCs) (Rose, 2006) (WA DOE, 2002). Between the two time periods, average biomass concentrations at Mt. Rainier decreased by 42% and the concentrations from secondary sulphate source, contributing the second highest amount of fine particulate (~30% of the reconstructed total mass), decreased by 37%. A possible reason could be that emission controls were added to a large power plant upwind from Mount Rainier National Park in 2001, which lead to a significant reduction in sulfur dioxide emissions, from 87,756 tons in 1999 to 3,355 tons by 2005 (Clow and Campbell, 2008).

Seasonal trends were identified in the secondary sulphate, secondary nitrate, soil and marine sources. The months of the year during which each source made its highest contribution to PM_2.5 at Mount Rainier are shown in Table 8.2.

 

Notes: Based on PMF analysis with EPA PMF 3.0 of mean 24-hour average concentrations, and 90^th percentile 24-hour concentrations, from 1-in-3 day sampling for the years 1991-1995 and 2000-2003. IMPROVE station (53-053-0014). (Rose, 2006).

 

Seattle

A source apportionment study by Wu et al. (2007) using PM_2.5 and VOC data from Beacon Hill in Seattle indicated that wood burning (24%-31%), secondary sulphate (20%-24%) and secondary nitrate (15%-20%) were the main contributors to PM_2.5. Figure 8.11 shows the source contribution estimates from the Multi-linear Engine (ME-2) models with PM_2.5 data only (left) and with both PM_2.5and VOCs data (right) The study concluded that adding the VOC data to the speciated PM_2.5 data in source apportionment modelling resulted in more accurate source contribution estimates for combustion related sources.

 

Notes: Based on PMF analysis with Multi-linear Engine (ME-2) of mean 24-hour average concentrations, from 1-in-3 day sampling for the years 2000-2004. The left stacked bar indicates results generated by the PM_2.5 data only model with 100% constraint. The right stacked bar indicates results generated from the combined PM_2.5 and VOCs data model with 100% constraint. IMPROVE station (53-033-0080). (Adapted from Wu et al., 2007).

8.7.3 PM Composition and Source Apportionment of the Lower Fraser Valley during the Pacific 2001 Air Quality Study

The Pacific 2001 Air Quality Study was undertaken to characterize the physical and chemical properties of particulate matter in the Canadian Lower Fraser Valley (LFV) and to provide detailed data for air quality model development (Vingarzan and Li, 2006). Five main ground sites were used during the August 13-31 study period, as illustrated in Figure 8.12. They included: Cassiar Tunnel in urban Vancouver, Slocan, an urban-suburban site in Vancouver; Langley, a rural site in the south-central LFV, Sumas, a semi-rural site in the eastern LFV, and Golden Ears, an elevated site on the north shore mountains. A large number of ground-level measurements were collected during the study, as well as LIDAR scans, vertical profiling, radiosonde and aircraft measurements.

 

Several chemical regimes were found to be influencing air quality in the WISE region. At Cassiar Tunnel, ambient aerosols were strongly dominated by gasoline and diesel emissions, the latter evidenced by elevated black carbon levels. At Slocan, the composition of ambient aerosols indicated an emission-controlled regime dominated by organics, of both anthropogenic and biogenic origin. Anthropogenic emissions were strongly influenced by vehicles and fugitive emissions of petroleum fuels. Plumes from ships operating around the area or at berth were observed to impact the city on a regular basis (Lu et al., 2006). At the rural site of Langley, the PM_2.5 regime was controlled by secondary aerosol formation. Here the average particle composition was dominated by inorganics, with contributions from sea salt, sulphate, nitrogen species and aged organics. Elevated ammonia and biomass burning markers reflected the influence of agricultural sources. At the semi-rural site of Sumas, the particle mass composition indicated a mixed emission-formation regime dominated by urban, agricultural and biogenic sources. At the forested site of Golden Ears, the particle composition was dominated by biogenic organics. The aerosol chemical composition at this site was impacted by anthropogenic sources due to upslope flow, while biogenically-derived secondary organic aerosols (SOA) were transported into the WISE by downslope flow.

Particle composition studies at many of the sites indicated that organic carbon comprised approximately half of the PM1.0 mass, with the rest being comprised by inorganic species. Sulphate dominated the inorganic fraction of the fine particle mode at all sites. Sea salt chemistry was found to be important in the formation of coarse-mode sodium nitrate aerosols. The fine particle fraction was dominated by ammonium sulphate aerosols, with occasional peaks in ammonium nitrate. Overall, elemental carbon comprised a minor part of the fine particle fraction. Isoprene and monoterpenes were the dominant biogenic hydrocarbons detected, and their presence at all sites indicated the broad influence of biogenic organics throughout the LFV.

Source analysis using isotopic methods suggested that the WISE was influenced by a well dispersed source of sulphate, consistent with gasoline, diesel and gas combustion. Biogenic contributions were found to be significant, with approximately 30% of the sulphate in PM_2.5 aerosols estimated to be derived from the oxidation of dimethyl sulphide (DMS). Organic aerosols were found to have significant contributions from terrestrial (plant material) sources during the summer months and from marine, biogenic and fossil fuel sources during the spring months. Biomass burning generally had a small, but measurable impact on the composition of ambient aerosols in the LFV. The importance of biogenics and fuel combustion (especially by cars) as determined by the Pacific 2001 Study in the WISE is similar to Wu’s (2007) findings for Seattle.

Particle formation and growth was observed to occur via several mechanisms.  Nucleation was observed on cleaner days while, on more polluted days condensation and coagulation dominated particle growth. Ozone episodes were associated with an increase in both particle size and mass. Sulphate growth events were associated with photochemical production from advected SO₂ from industrial sources in north-west Washington State. Within the City of Vancouver, sulphate peaks were observed to be associated with ship plume emissions. Fine particle nitrate events were associated with advection of HNO₃ into the eastern WISE reacting with local sources of ammonia. Organic particle growth events were associated with stagnant conditions and advection of air over urban Vancouver. The extent of processing of air masses was found to increase in a west to east direction, although photochemically aged pollutants from the area where the Strait of Juan de Fuca converges with the Strait of Georgia could reverse this gradient.  Both biogenic and anthropogenic species were observed to be chemically reactive and contribute to secondary aerosol formation. 

8.8 Chapter Summary

Fine particulate matter (PM_2.5) is a health concern because it is in a size range that is easily inhaled by the lungs. Elevated levels of PM_2.5 are also responsible for visibility degradation. The principal contaminants responsible for the secondary formation of PM_2.5 include NO_x, SO₂, NH₃ and organic carbon. In the atmosphere, these contaminants are transformed by chemical and physical processes to ammonium nitrate, ammonium sulphate and organic compounds. Elemental carbon (from combustion of organic material), sea salt, and soil also contribute to fine particulate matter.

Concentrations of PM_2.5 in the Georgia Basin/Puget Sound airshed can either peak in fall or winter (due to heating and wood combustion emissions and reduced dispersion under lower boundary layer heights) or in the late summer (due to enhanced photochemical conditions).  Daily cycles include rush-hour peaks at urban sites and elevated levels in the evening and overnight, associated with low mixing heights. PM_2.5 concentrations in the airshed are typically below national objectives and standards, with the notable exception of the Tacoma-Pierce County nonattainment area in Washington State. Outside the non-attainment area, PM_2.5 shows evidence of a slight decline, although data records are too short to calculate robust trends.

Although PM_2.5 composition varies by location, organic carbon was found to be the dominant component in the airshed followed by ammonium sulphate and ammonium nitrate. Contributions from sulphate peak in the summer due to increased biogenic sulphur emissions and higher rates of photochemical conversion of sulphur dioxide to sulphate. In contrast, maximum nitrate contributions occur at different times of the year in different locations due to local chemistry, including availability of ammonia. Where ammonia is abundant, ammonium nitrate production is favoured during the agriculturally active seasons.

Key sources contributing to ambient PM_2.5 in the Canadian Lower Fraser Valley are motor vehicles, marine transport, agriculture and industry. Spatial differences include a stronger influence from marine transportation in the western part of the airshed. Seasonal influences include wood burning, secondary nitrate, and vehicle emissions which are more important in the winter, and increased cruise vessel activities and biogenic influences which are more important in the summer. In Seattle, wood burning, secondary sulphate and secondary nitrate are the three main contributors to PM_2.5. At Tacoma, wood burning, secondary particles and motor vehicles dominate the PM_2.5 composition. At Mount Rainier, biomass from natural and anthropogenic combustion sources contributes the largest amount to fine particulate composition.  Between the early 1990s and the early 2000s, average biomass concentrations at Mt. Rainier decreased by 42% and the concentrations from secondary sulphate sources decreased by 37%, the latter reflecting implementation of industrial emission controls during this period.

In summary, PM_2.5 levels in the Georgia Basin/Puget Sound airshed generally comply with national standards and regional objectives, but episodes of elevated particulate matter continue to affect certain areas. In most areas, ambient PM_2.5concentrations have been declining slightly over the past decade. Future PM levels will depend on changes in background levels, climate and local and transported emissions.

8.9 References

Ashbaugh, L.L., Malm, W.C., and Sadeh, W.Z.,1985. A Residence Time Probability Analysis of Sulfur Concentrations at Grand Canyon National Park. Atmospheric Environment 19: 1263-1270.

Barthelmie, R.J. and Pryor, S.C., 1998. Implications of ammonia emissions for fine aerosol formation and visibility impairment - a case study from the Lower Fraser Valley, British Columbia. Atmospheric Environment 32 (3): 345-353.

Blanchard, C. and Hldy, G., 2003. Effects of changes in sulfate, ammonia, and nitric acid on particulate nitrate concentrations in the southeastern United States. Air & Waste Management Association 53: 283-290.

Brook, J.R., Farrell, C., Friesen, K., Langley, L., Kellerhals, M., Morneau, G., Reid, N., Vingarzan, R., Waugh, D., and Wiens, B., 2011. "Chapter 7. Air quality at the regional and local scale: The what, where, why and how of concentration variations". In:Canadian Smog Science Assessment. Volume 1. Atmospheric Science and Environmental Effects. Environment Canada and Health Canada.(Executive summary) (Full report available upon request) Environment Canada, Science and Technology Branch, 4905 Dufferin St, Downsview, Ontario, M3H 5T4

Clow, D.W., and Campbell, D.H., 2008. Atmospheric deposition and surface-water chemistry in Mount Rainier and North Cascades National Parks, U.S.A., water years 2000 and 2005-2006: U.S. Geological Survey Scientific Investigations Report 2008-5152, 37 p.

Dann, T., Vingarzan, R., Chan, E., Vet, B., Brook, J., Martinelango, K., Shaw, M., Dabek, E., Wang, D., Herod, D., Mignacca, D., Anlauf, K., Graham, M., O’Brien, J., 2011. “Chapter 3, Ambient Measurements and Observations”. In: Canadian Smog Science Assessment. Volume 1. Atmospheric Science and Environmental Effects. Environment Canada and Health Canada.(Executive summary) (Full report available upon request) Environment Canada, Science and Technology Branch, 4905 Dufferin St, Downsview, Ontario, M3H 5T4

Husar R., et al., (28 coauthors) 2001. The Asian Dust Events of April 1998. Journal of Geophysical Research106 (D16): 18,317-18,330, doi:10.1029/2000JD900788.

Langner, J., Rodhe, H., 1991. A global three dimensional model for the tropospheric sulfur cycle. Journal of Atmospheric Chemistry 13, 225-263.

Larson, T., Su, J., Baribeau, A.M., Buzzelli, M., Setton, E., and Michael Brauer, 2007. A Spatial Model of Urban Winter Woodsmoke Concentrations. Environmental Science & Technology 41 (7): 2429-2436.

Makar, P.A., 2001. Processes linking NO_x, SO₂, NH₃, and VOCs to secondary particle formation, Chapter 2: Precursor Contributions to Ambient Fine Particulate Matter in Canada. Meteorological Service of Canada. ISBN: 0-662-30650-3. May 2001, 237 pp.

McHenry, J.N., Dennis, R.L., 1994. The relative importance of oxidation pathways and clouds to atmospheric ambient sulfate production as predicted by the Regional Acid Deposition Model. Journal of Applied Meteorology 33: 890-905.

McKendry, I.G., 2000. PM₁₀ Levels in the Lower Fraser Valley, British Columbia, Canada: An Overview of Spatiotemporal Variations and Meteorological Controls. Journal of Air & Waste Management Association 50: 443-445.

McKendry, I.G., J.P. Hacker, R. Stull, S. Sakiyama, D. Mignacca and K. Reid, 2001. Long-range Transport of Asian Dust to the Lower Fraser Valley, British Columbia, Canada. Journal of Geophysical Research 106: 18 361-18 370.

McKendry, I. G., 2006.  Background concentrations of PM_2.5 and Ozone in British Columbia, Canada. University of British Columbia report prepared for the British Columbia Ministry of the Environment, Vancouver, B.C.

NAPS (National Air Pollution Surveillance Network), 2006. NAPS Data Archive. Environment Canada, Environmental Protection Service, Ottawa, Ontario, K1A 0H3.

NARSTO, 2004. Particulate Matter Science for Policy Makers: A NARSTO Assessment. P. McMurry, M. Shepherd, and J. Vickery, eds. Cambridge University Press, Cambridge, England. ISBN 0 52 184287 5.

Norman, A., Belzer W., and Barrie, L., 2004. Insights into the biogenic contribution to total sulphate in aerosol and precipitation in the Fraser Valley afforded by isotopes of sulphur and oxygen. Journal of Geophysical Research D. 109, D05311, doi:10.1029/2002JD003072.

Ostermann, K., 2002. Source apportionment of particulate matter by Positive Matrix Factorisation in the Lower Fraser Valley of British Columbia. Prepared for the B.C. Ministry of Water, Lands and Air Protection, Nov. 2002, 89 pp.

Paatero, P. and Tapper, U., 1994. Positive Matrix Factorization: A Non-Negative Factor Model with Optimal Utilization of Error Estimates of Data Values. Environmetrics 5, 111-126.

Paatero, P., 1997. Least Squares Formulation of Robust Non-Negative Factor Analysis. Chemometrics and Intelligent Laboratory Systems 37, 23-35.

Pandis, S.N., Seinfeld, J.H., Pilinis, C., 1992. Heterogeneous Sulfate Production in an Urban Fog. Atmospheric Environment 26: 2509 - 2522.

Rose, K., 2006. Source Allocation and Visibility Impairment in Two Class I Areas with Positive Matrix Factorization. Report for U.S. Environmental Protection Agency, Region 10.

Seinfeld J. H. and Pandis S. N., 1998. Atmospheric Chemistry and Physics: From Air Pollution to Climate Change, 1st edition, J. Wiley, New York.

Sharma, S., Vingarzan, R., Barrie, L.A., Norman, A., Sirois, A., Henry, M., and di Cenzo, C., 2003. Concentrations of dimethyl sulphide in the Strait of Georgia and its impact on the atmospheric sulphur budget of the Canadian West Coast. Journal of Geophysical Research D. 108, D15, 4459.

So, R., Vingarzan, R. and Meyn, S. 2010. Source Apportionment of PM_2.5 at Burnaby South and Abbotsford. Environment Canada. Environment Canada, Meteorological Services, #201-401 Burrard, Vancouver, BC, V6S 3S5.

Suzuki, N.M. and B. Taylor, 2003. Particulate matter in BC: a report on PM₁₀ and PM_2.5 mass concentrations up to 2000. Co-produced by British Columbia Ministry of Water, Land and Air Protection, and the Pacific and Yukon Region of Environment Canada. ISBN 0-7726-4904-9. October 20, 2003, 118 pp. (Accessed: July 6, 2010)

U.S. EPA (Environmental Protection Agency), 2002. Air Quality Criteria for Particulate Matter. Volume 1. Third External Review (April 2002). National Center for Environmental Assessment - RTP Office, Office of Research and Development, U.S. Environmental Protection Agency, Research Triangle Park, N.C.EPA/600/P-99/002aC

U.S. EPA (Environmental Protection Agency), 2010. “Air Quality Monitoring Information.” (Accessed: July 6, 2010)

Vingarzan, R., 2004. Ambient Particulate Matter Concentrations in Canada and Background Levels. Environment Canada, Environmental Conservation Branch, Aquatic and Atmospheric Sciences Division, #201-401 Burrard Street, Vancouver, British Columbia, V6C 3S5

Vingarzan R. and Li, S-M., 2006. The Pacific 2001 Air Quality Study - synthesis of findings and policy implications. Atmospheric Environment 40: 2637-2649.

WA DOE (Washington State Department of Ecology), 2002. Review of the WA State Visibility Protection State Implementation Plan. Washington State Department of Ecology Air Quality Program. November, 2002. (Accessed: May 27, 2010)

WA DOE (Washington State Department of Ecology), 2007. Washington State Base Year 2005 County Inventories. WA DOE internal document; personal communication via Sally Otterson.

WA DOE (Washington State Department of Ecology), 2010a. Nonattainment Areas in Washington. (Accessed: July 15, 2010).

WA DOE (Washington State Department of Ecology), 2010b. Sources of Fine Particles in the Wapato Hills-Puyallup River Valley PM_2.5 Nonattainment Area. Air Quality Program Publication 10-02-009. (Accessed: July 15, 2010).

Wu, C., Larson, T., Wu, S, Williamson, J., Westberg, H., Liu, L-J., 2007. “Source apportionment of PM_2.5 and selected hazardous air pollutants in Seattle”, Science of The Total Environment 386: 42-52.