The Major Components of PM_2.5 at Remote Sites across the United States

Author:  Robert Eldred, CNL, University of California, Davis (raeldred@ucdavis.edu)

Date:  February 1998   

ABSTRACT

The IMPROVE (Interagency Monitoring of Protected Visual Environments) network measures the concentrations of the major components of PM_2.5 mass at 70 sites throughout the United States:  sulfate, nitrate, organics, elemental carbon (EC), and soil.  The calculation of these components is discussed.  The only components that inter-correlate are organics and EC, with EC being 10% to 18% of organics.  The spatial patterns of the annual concentrations are discussed.  Sulfate is much higher in the East than in the West.  Organics and EC shows less variation between regions.  In the East , the means increase from north to south.  In the West, they are highest in the Northwest and California.  Nitrate is generally higher in the East, but is highest at sites near western urban areas.  Soil tends to increase from north to south with little east-west difference.  The relative contributions of the five components are examined.  In the East, sulfate accounts for 60% of the total and organics plus EC 30%.  In the Northwest, organics plus EC account for 50% and sulfate 25%.  In the Southwest, sulfate, organics, and soil are approximately equal.  The seasonal variations are discussed briefly.  Comparisons are also made of two urban sites (Washington DC and Seattle) with nearby sites at Class I areas.  Of all the components, sulfate shows the least variation from remote to urban.

INTRODUCTION

            The IMPROVE (Interagency Monitoring of Protected Visual Environments) network was established in 1987 to assist the Environmental Protection Agency (EPA) and the Federal Land Managers in protecting Class I visibility areas from increased visibility impairment.^{1, 2}  The network monitors visibility parameters of scattering and extinction and several aerosol parameters:  PM_2.5 and PM₁₀ mass, and the major components and trace elements of PM_2.5.  The purposes of the network include establishing background concentrations and monitoring trends.  Participants in the network include the National Park Service, Forest Service, and Fish and Wildlife Service, Bureau of Land Management, Department of Energy, Tahoe Regional Planning Agency, interstate agencies, and several states. 

            In 1998, EPA will begin the PM_2.5 Federal Reference Method (FRM) network to determine compliance toward a health-based mass standard.  Associated with this will be a chemical speciation network at some of the sites to provide information on the composition of the PM_2.5 mass.  This network will use collection and analysis protocols similar to those used in IMPROVE.  The IMPROVE network will be used to establish regional background concentrations and transport to facilitate understanding of the chemical speciation data of the FRM network.

            This paper focuses on the five major components of PM_2.5 mass:  sulfate, nitrate, organics, elemental carbon, and soil.  The data is based on samples collected at 64 sites during the annual period from March 1996 to February 1997.  The seasonal analysis separates the time into spring (March-May), summer (June-August), fall (September-November), and winter (December-February). 

SAMPLE COLLECTION AND ANALYSIS

            The sample collection and analysis are described in detail in a companion paper in this volume.³  PM_2.5 particles are collected on Teflon, nylon, and quartz filters every Wednesday and Saturday.  The sulfate and soil components are based on the x-ray analysis of the Teflon filter.  The nitrate component is based on the Ion chromatography analysis of the nylon filter.  The nylon filter has a carbonate denuder to ensure that only particulate nitrate is collected.  Sulfate is also measured on the nylon filter; the concentrations are equivalent to the sulfur on the Teflon.  The Teflon sulfur is used because some sites do not have nylon filters.  The dual measurements provides a valuable quality control of the data.  The organics and elemental carbon components are based on Thermal/Optical Reflectance measurement of carbon on the quartz filter.  The distinction between organic and elemental carbon is whether oxygen is needed to produce CO₂.  A correction for the pyrolysis of organic carbon to elemental carbon during analysis is made based on the optical reflectance of the sample.

The 64 sampling sites with complete data records for the annual period from March 1996 to February 1997 are shown in Figure 1.  Of these, 7 sites collected particles on only Teflon filters.  Two sites are in large urban areas (Washington DC and Seattle), while a third in in a small urban area (South Lake Tahoe).

CALCULATION OF MAJOR COMPONENTS

Sulfate

            The sulfate component assumes that all sulfate is present as (NH₄)₂SO₄.  In this paper, it is calculated from the sulfur measured on the Teflon filter (sulfate = 4.125 S), although it could be calculated from the sulfate measured on the nylon filter (sulfate = 1.375 SO₄).  In the East, especially in summer, the sample may be primarily (NH₄)HSO₄ or even H₂SO₄.²  The dry weights of these are slightly lower than for (NH₄)₂SO₄.  Although sulfate is very hygroscopic, the calculated component does not include the attached water. 

            The associated water can be estimated using the aerosol growth curves for pure sulfate salts as a function of relative humidity.^{4, 5}  If the concern is the ambient mass at the time of sampling, the ambient RH is appropriate.  If the concern is the gravimetric mass measurement, the laboratory RH is used.  One problem is that real sulfate is probably not a pure salt, but could have a coating of organic that might modify the growth.  A second problem is that the form of sulfate is known only if ammonium ion is also measured; this is important since the growth curves for (NH₄)₂SO₄, (NH₄)HSO₄, and H₂SO₄ are different.  Fortunately, sulfate is predominantly (NH₄)₂SO₄ except during the summer at eastern sites.  With (NH₄)₂SO₄, there is no significant growth below 80% RH.  Thus, there would be little added water for (NH₄)₂SO₄ as long as the particles never exceed 80% RH in the atmosphere or between collection and analysis.  At an ambient RH of 90%, the added water would be about 3 times the dry ammonium sulfate mass for any of the three sulfate forms. A sample collected above 80% RH will lose most of the water when brought down to a laboratory RH of 45%.  The remaining water would be about 25% of 4.125*S for (NH₄)₂SO₄ or (NH₄)HSO₄, and 50% for H₂SO₄.  Thus, for sample in which 4.125*S is 75% of the gravimetric mass, the ambient mass at 90% RH would be about 3 time the gravimetric mass for either of the three forms. 

Soil

            This component is basically the sum of the soil-derived elements plus oxides.  The details are based the composition of average sediment.⁶  The primary minerals in sediment are Al₂O₃, SiO₂, K₂O, CaO, FeO, Fe₂O₃, and TiO₂.  Iron is assumed to be equally split between FeO and Fe₂O₃.  In samples with biomass combustion emissions, a large fraction of the potassium may be non-soil.  For average sediment, the ratio of K/Fe is 0.6.  (This ratio was also obtained in coarse particle samples in the National Park Service network from 1982-85.)  The K contribution is based on this ratio rather than on the measured concentration.  These elements account for 86% of the mass of average sediment.  The final expression for soil is given in Equation 1.

            soil = ( 1.89 Al + 2.14 Si + 1.40 Ca + 1.67 Ti + 2.08 Fe ) / 0.86                                 (1)

            Comparisons of reconstructed and gravimetric mass for samples with high soil indicate that the equation is reasonable.

Nitrate

            Nitrate is assumed to be present as (NH₄)NO₃, or nitrate = 1.29 * NO₃, using the measurement on the nylon filter.  In many cases, a large fraction of the NO₃ may be lost from the Teflon filter during sampling.  Therefore, caution must be exercised in including nitrate in the reconstructed mass when comparing with Teflon gravimetric mass.

Organics

            This paper uses a multiplicative factor of 1.4, which assumes that an organic molecule is 71% carbon by weight.  To determine this factor, complete organic analyses would be needed for a wide variety of environments.  This has not yet been accomplished.  An estimate of this factor can be obtained by comparing reconstructed and gravimetric mass for samples from selected environments.  The hygroscopicity of sulfate does not allow this comparison to be done with high RH and predominantly sulfate samples, such as in the eastern United States in summer.  Sites with high nitrate must also be avoided because of uncertainty in the volatilization of nitrate.  Marine environments must be avoided because of the uncertainty in the molecular compounds of sulfate and sodium.  Figure 2 makes a comparison for all western IMPROVE sites without high nitrate or marine aerosols collected in summer 1996.  The reconstructed mass here does not include nitrate.  All of the samples with high mass are dominated by OC, presumably associated with emissions from fires.  With an organic factor of 1.4 (left side) the slope is 0.86.  Increasing the factor by 15% to 1.61 increases the slope to 0.96, while increasing the factor by 20% to 1.68 increases the slope to 1.00.  A slope of 0.96 may be the most reasonable, since it allows a little room for nitrate and any water associated with organics.  The reconstructed mass with a factor of 1.61 is shown in the right plot.  The conclusion is that the factor for samples in non-urban sites could perhaps be increased by about 15% from 1.4 to 1.6.  A factor greater than 1.68 is not warranted by the data. 

Elemental Carbon

            For pure elemental carbon, the mass of the component would be equal to the measured EC.  Because of the analytical definition by TOR, EC may contain some non-carbon elements.  No correction is made for this.

RESULTS AND DISCUSSION

Organics and Elemental Carbon

            The only components that inter-correlate are organics and EC.  Figure 3 shows a comparison for summer 1996.  The right side is for the major urban sites of Washington DC and Seattle, while the left side is for all other sites.  For the remote sites, the correlation (r²) is 0.76, with one half of the points lying between EC/organics ratios of 0.10 and 0.17.  The high correlation suggests that much of the organics and EC have a common origin.  The correlation (r²) for any other pairs of components is less than 0.20.  For the urban areas of Washington DC and Seattle, there is still a good correlation (0.52), but the relative amount of EC is much higher.  This suggests that there are additional sources of EC in urban areas, such as diesel combustion.

Annual Regional Patterns of Components

            Figures 4-7 show the mean annual concentrations of sulfate, carbonaceous (organics plus EC), nitrate, and sulfate for the non-urban sites for the period from March 1996 to February 1997.  All concentrations are in µg/m³.  The volumes are all local, so that the concentrations are not normalized to sea level.  The elevations range from sea level to 3400m.  To convert from 2000m to sea level, the concentrations would be multiplied by 1.26.

            For sulfate (Figure 4), there is a major difference between the East and the West, with mean concentrations throughout Appalachia of 6-7 µg/m³ compared to 0.4-0.6 µg/m³ in parts of the Northwest.  The concentrations in the Northeast and southern Florida are about one/half those in Appalachia.  Many of the sites in Oregon, Idaho, Northern California and Nevada have concentrations that are as low as that at Denali, in central Alaska.  The highest concentrations in the West are in the Southwest, with a maximum at Big Bend.  The absence of sites (and Class I areas) in the central part of the United States does not allow a complete mapping of sulfate from the IMPROVE network alone. 

            Figure 5 shows the annual concentrations for the sum of organics and EC.  (Because of the high correlation between organics and EC, it is possible to combine the two into a single category.  Separate plots provide no additional information.)  There is much less variation than for sulfate, ranging from 1.0-1.5 µg/m³ in arid portions of the West to 3-5 µg/m³ in much of the East and at Sequoia.  The Lone Peak site in northern Utah shows elevated carbon compared to its neighbors.  This site is impacted by sources in the Salt Lake/Provo basin.  The carbon concentrations at Denali are less than any in the contiguous United States. 

            The mean annual concentrations of nitrate are shown in Figure 6.  The West is characterized by very low background levels (0.2-0.3 µg/m³) but with several sites with much higher concentrations, all near urban areas.  The sites with elevated nitrate and the nearby urban areas (in parenthesis) include San Gorgonio (Los Angeles region ), Sequoia and Yosemite (San Joaquin Valley), Point Reyes and Pinnacles (San Francisco area), Columbia River (Portland), and Lone Peak (Salt Lake/Provo).  The nitrate concentrations in the East are generally higher than in the West. 

            The concentrations for soil (Figure 7) generally show a slight increase from north to south, but no major east/west difference.  This trend is definitely seen in the East, with concentrations ranging from 0.2 µg/m³ in the Northeast to 1.3 µg/m³ in southern Florida.  The West shows a more varied pattern.  The lowest concentration are at the coastal sites in California and Washington (0.2 µg/m³).  This is the same as Denali, Haleakala and the Northeast.  The highest concentrations for 1996 were Sequoia (2.8 µg/m³), but are associated with major construction in the area of the site.  The highest concentration without construction is at Guadalupe Mountains in West Texas (2.0 µg/m³).  Based on prior years, the fine soil concentrations at the Virgins Islands site is the highest in the network.  The annual 1996 data is not shown because the spring quarter was lost following a hurricane.  The source of this soil appears to be the Sahara, and occurs throughout the spring and summer.⁷  The Saharan dust reaches the sites in the contiguous states only during summer.  During these episodes, the fine soil concentrations in the Southeast and in Appalachia are higher than any observed in western IMPROVE sites.  The Saharan dust episodes have been seen as far west as Big Bend, in West Texas. 

Comparison of the Components

            In order to facilitate discussion, the non-urban sites will be divided into 8 regional groupings, as shown in Figure 8.  Denali (AK) and Sequoia are not included in any grouping.  The composition at Denali is similar to the average of the Northwest sites, but the concentrations are lower.  Sequoia was excluded because of extensive road construction over much of the year.  The grouping West Coast includes the three California coastal sites.  At these sites, NaCl is a significant portion of the PM_2.5 mass.  NaCl is larger than any of the five categories at Point Reyes and Redwoods.  San Gorgonio Wilderness, east of the Los Angeles Basin, is treated as a separate region.  Appalachia includes two sites not actually in Appalachia, Brigantine (NJ) and Upper Buffalo (AR). 

            Figure 9 shows the annual concentrations of the components in each of the regions.  The main difference between the Northwest and the Southwest is the relative concentrations of sulfate and organics: organics is larger in the Northwest, while sulfate is larger in the Southwest.  The Southwest also has more soil.  The three West Coast sites are characterized by higher nitrate and lower soils than the average in the West.  Based on the S/Na ratio for dissolved salts in sea water,⁸, and the observed Na concentrations at these sites, approximately 25% of the sulfur at Point Reyes and Redwoods and 9% at Pinnacles are marine in origin.  (The fraction of marine sulfur is 6-8% at Everglades, the Columbia River sites, and Seattle.  It is about 3% at the other eastern coastal sites.)   San Gorgonio is dominated by nitrate.  The three eastern sites are dominated by sulfate; this is about 60% of the total. 

            The pie diagram (Figure 10) emphasizes the composition of the aerosol..  The three eastern regions have similar composition, and are combined into a single diagram.  In the East, sulfate accounts for 58% of the total, and organics an additional 26%.  EC, nitrate, and soil are each about 5%.  There is much more variation in composition in the West, so that it is not reasonable to calculate an average West.  The largest component in the Northwest is organic, with organics plus EC accounting for about one-half of the total.  Sulfate contribute an additional 25% and soil 17%.  Note that increasing the organic factor by 15% from 1.4 to 1.61 would increase the organics contribution in the Northwest from 44% to 48%.  In the Southwest, the largest component is sulfate, but it is only slightly larger than organics or soil.  PM_2.5 soil accounts for one-quarter of the mass budget.  At San Gorgonio, near Los Angeles, nitrate is the largest component, accounting for 38% of the total. 

Seasonal Variation of the Components at Remote and Urban Sites

            Figure 11 shows the seasonal mean concentrations for each of the 8 regions.  In general, sulfur is highest in summer.  A seasonal comparison for selenium does not show a similar summer increase.⁹  Since selenium generally correlates with sulfur, especially in the East, and is emitted along with SO₂ from coal-fired power plants and some other sources, this suggests that the summer increase is not associated with better summer transport.  The most likely explanation for the summer high is higher conversion rates from SO₂ to sulfate.  Organics and soil also tend to be highest in summer.  At San Gorgonio, nitrate is highest in spring and summer. 

            IMPROVE samplers operated in two major urban areas in 1996, Washington DC and Seattle.  Figure 12 shows the seasonal concentrations of the components at these urban sites and at nearby Class I sites.  The urban sites are at a lower elevation than the Class I sites.  Washington DC is at sea level, Seattle at 100m, Mt Rainier at 400m, and the others at 1100m.  To eliminate the effect of the expansion of a given parcel of air with altitude, all concentrations are normalized to sea level. 

            For the Washington DC set, there is no difference in the concentrations of sulfate between any of the sites for any season; overall the urban area is within 5% of the Class I mean.  The urban/Class I ratio for organics is 1.5 for the year, varying from 1.1 in summer to 2 in fall and winter.  Nitrate and EC are always higher in the urban site, with ratios of 2 to 3, except in winter for nitrate where it is nearly 5.  Soil is also higher at the urban site, with an annual ratio of 1.7.  Comparing summer and winter at the Washington DC site, the winter concentrations of sulfate and soil are one-half those in summer, organics and EC are the same, but nitrate increases by a factor of 3.  The net result is a 30% decrease in the total of the five components.  The corresponding decease at the Class I sites is 60%.  The conclusion is that is that the total particle concentration is higher at the urban site, especially in winter, with the increase primarily associated with nitrate.

            In the Seattle comparison, the total concentration of the components are always higher at the urban site than in the Class I sites, with ratios ranging from 1.4 in summer, 2.5 in spring and fall, and 4 in winter.  All components are higher in Seattle than in the Class I sites for every season, especially EC and nitrate.  The smallest difference is for sulfate in summer, which differ by only 25%.  The largest difference is in winter, when all components at Seattle are 3 to 5 times the average of the Class I sites.  Comparing summer and winter at the Seattle site, the total of the components remain about the same.  A sharp decrease from summer to winter for sulfate and nitrate are canceled by an increase in organics and EC.  A possible source of the increased winter organics is residential woodburning.  At the Class I sites, all components, including organics, decrease sharply from summer to winter. 

CONCLUSIONS

            For the non-urban sites, EC correlates with OC, with an EC/organics ratio of about 0.14.  Thus for non-urban sites, most EC and OC have a common origin.

            For the eastern Class I sites, the mass budget is dominated by sulfate, which contributes about 60% of the total of the five components.  The contribution would be much larger if the associated water were included, because of the high ambient relative humidity at most eastern sites in summer, when the sulfate concentrations are highest.  Organics and EC combined contribute an additional 30%, or slightly more than one-half of the sulfate.

            For the western Class I sites, organics, sulfate, nitrate, and soil all contribute to the mass budget, with the relative composition depending on the region.  In the Northwest, organics are the largest component, accounting for nearly one-half of the mass budget.  In the Southwest, sulfate is slightly larger than organics and soil.  For the average western site, soil is the third largest component, at about 20% of the total.  Nitrate can be a major component at sites near western urban areas, especially at San Gorgonio, near Los Angeles, where nitrate is the largest component. 

            Comparing the urban site in Washington DC with nearby Class I sites, shows similar concentrations in summer, but higher urban concentrations the rest of the year.  The increase is associated with all components except sulfate.  In winter, the aerosol at Washington DC has more organics plus EC than sulfate.  Comparing the urban site in Seattle with nearby Class I sites, shows higher urban concentrations every season, with the greatest difference in fall and winter.  At the nearby Class I sites, the total concentrations are highest in summer, dropping in half in fall and winter.  At Seattle, the highest concentrations occur in fall and winter.  The organics and EC components are the major contributors to this fall and winter rise.  Thus, for both Washington DC and Seattle, the major difference between urban and Class I aerosols occurs in winter, with organics and EC playing key roles.

REFERENCES

1.   Joseph, D.B.; Metsa, J.; Malm, W.C.; Pitchford, M.L.  "Plans for IMPROVE: A federal program to monitor visibility in class I areas," In Visibility Protection: Research and Policy Aspects, P.S. Bhardwaja, Editor, APCA, Pittsburgh, PA, 1990, pp. 113-125.

2.   Malm, W.C.; Sisler, J.F.; Huffman, D.; Eldred, R.A.; Cahill, T.A. "Spatial and seasonal trends in particle concentration and optical extinction in the United States," J. of Geophysical Research 1994, 99, 1347-1370.

3.   Eldred, R.A., Feeney, P.J.; Wakabayashi, P.K.; Chow, J.C.; Hardison, E.  “Methodology for Chemical Speciation Measurements in the IMPROVE Network,” this volume.

4.   Tang, I.N.; Wong, W.T.; Munkelwitz, H.R. “The relative importance pf atmospheric sulfates and nitrates in visibility reduction,” Atmos. Environ. 1994, 28, 3061-3071.

5.   Tang, I.N.; Munkelwitz, H.R. “Aerosol growth studies-III, ammonium bisulfate aerosols in a moist atmosphere,” J. Aerosol Sci. 1977, 8, 321-330.

6.   Carmichael, R.S. CRC Practical Handbook of Physical Properties of Rocks and Minerals; (Table 58) CRC Press: Boca Raton, FL, 1989.

7.   Perry, Kevin, Cahill, T.A., Eldred, R. A., Dutcher, D.D, Gill, T.E.  Long-range transport of North African dust to the eastern United States.  J. Geophysical Res., 1996, 102, 11225-11238.

8.   Eldred, R.A. “Comparison of selenium and sulfur at remote sites throughout the United States,”  J. Air & Waste Manage. Assoc. 1997, 47, 204-211.

9.   Kennish, M.J. Practical Handbook of Marine Science; CRC Press, Boca Raton, FL., 1994.

Figure 1.  IMPROVE sampling sites with complete record from March 1996 to February 1997.  The data from the sites with partial speciation is used only for sulfur and soil.

Figure 2.  Comparison of reconstructed and gravimetric mass for western sites with minimal impacts from nitrate and marine air, for summer 1996.  The reconstructed mass does not include nitrate or water.  The left plot assumes a usual organic factor of 1.4, while the right side has a factor of 1.61, which is 15% higher.  The correlation (r²) is 0.95 in both cases.

Figure 3.  Comparison of organics (1.4 OC) and elemental carbon, in µg/m³, for samples collected in summer 1996.  The right plot is for two urban sites, Seattle and Washington DC, while the left plot is for all other sites.  One half of the points for non-urban sites lie between the two lines.

Figure 4.  Annual mean concentration of sulfur, expressed as (NH₄)₂SO₄ (4.125 S), in µg/m³, for samples collected from March 1996 to February 1997. 

Figure 5.  Annual mean concentration of total carbonaceous mass, equal to (1.4 OC+EC), in µg/m³, for samples collected from March 1996 to February 1997.

Figure 6.  Annual mean concentration of (NH₄)NO₃ (1.29 NO₃), in µg/m³, for samples collected from March 1996 to February 1997.

Figure 7.  Annual mean concentration of soil, equal to the sum of the soil-derived elements plus oxides, in µg/m³, for samples collected from March 1996 to February 1997.

Figure 8.  Separation of sites into 8 regions. 

Figure 9.  Annual concentrations of the five components for the 8 regions shown in Figure 8.

Figure 10.  Annual concentrations of the five components for four regions.  East is the average of Northeast, Appalachia, and Southeast. 

Figure 11.  Seasonal concentrations of the five components for the 8 regions shown in Figure 8.

Figure 12.  Seasonal concentrations of the five components at Washington DC and Seattle and nearby class I sites.  Snoqualmie represents Alpine Wilderness.  All concentrations are normalized to sea level.