Environ. Sci. Technol. 2010, 44, 8435–8440
Impacts of the Canadian Forest Fires on Atmospheric Mercury and Carbonaceous Particles in Northern New York YUNGANG WANG,† JIAOYAN HUANG,‡ TIFFANY J. ZANANSKI,† P H I L I P K . H O P K E , * ,† A N D THOMAS M. HOLSEN‡ Center for Air Resource Engineering and Science, Clarkson University, Potsdam, New York 13699-5708 and Department of Civil and Environmental Engineering, Clarkson University, Potsdam, New York 13699-5710
Received July 21, 2010. Revised manuscript received September 23, 2010. Accepted September 30, 2010.
The impact of Canadian forest fires in Quebec on May 31, 2010 on PM2.5, carbonaceous species, and atmospheric mercury species was observed at three rural sites in northern New York. The results were compared with previous studies during a 2002 Quebec forest fire episode. MODIS satellite images showed transport of forest fire smoke from southern Quebec, Canada to northern New York on May 31, 2010. Backtrajectories were consistent with this regional transport. During the forest fire event, as much as an 18-fold increase in PM2.5 concentration was observed. The concentrations of episode-related OC, EC, BC, UVBC, and their difference (DeltaC), reactive gaseous mercury (RGM), and particle-bound mercury (PBM) were also significantly higher than those under normal conditions, suggesting a high impact of Canadian forest fire emissions on air quality in northern New York. PBM, RGM, and Delta-C are all emitted from forest fires. The correlation coefficient between Delta-C and other carbonaceous species may serve as an indicator of forest fire smoke. Given the marked changes in PBM, it may serve as a more useful tracer of forest fires over distances of several hundred kilometers relative to GEM. However, the Delta-C concentration changes are more readily measured.
1. Introduction Remote wildfires elevate atmospheric carbonaceous particulate matter concentrations and influence global climate (1, 2). They also cause adverse human health effects (3). Forest fire particles were more toxic than equal doses of particles collected from ambient air from the same region during a comparable season (4). Large-scale lightning-ignited forest fires are common in the boreal zone (45°N-75°N) of Canada (5), producing significant air pollution (6). The Quebec forest fire that occurred in early July 2002 resulted in 30-fold and 5-fold increases in ambient PM2.5 mass concentrations in Baltimore, MD (7) and Philadelphia, PA (8), respectively. * Corresponding author phone: 315-268-3861; fax: 315-268-4410; e-mail:
[email protected]. † Center for Air Resource Engineering and Science. ‡ Department of Civil and Environmental Engineering. 10.1021/es1024806
2010 American Chemical Society
Published on Web 10/27/2010
There are three atmospheric mercury (Hg) species: gaseous elemental mercury (GEM), reactive gaseous mercury (RGM), and particle-bound mercury (PBM) (9). GEM comprises up to 95% of the background global atmospheric Hg and has a residence time of 0.5-2 years (10). It can be transported over long distances (11, 12). RGM and PBM can be quickly removed by either dry or wet deposition resulting in short atmospheric lifetimes (days to weeks). Field measurements in Petersham, MA in July 2002 show evidence of long-range transport of gaseous mercury in the smoke plume from a series of boreal forest fires in northern Quebec (13). Hg emissions from forest fires also include resuspension of industrial Hg emissions that were deposited on trees and the soil in the forest. Hg emissions from biomass burning were estimated based on the enhanced ratios between GEM and carbon monoxide (CO) (14, 15). GEM can be oxidized to RGM during combustion. RGM can easily attach to particles to form PBM, particularly at high particle concentrations. However, this study did not explore the influence of RGM and PBM from wildfires. Generally, RGM and PBM contribute less than 5% of total atmospheric mercury. Because of their relative low background concentrations, the concentrations of these two species may be more substantially influenced by forest fires compared to GEM and thus, serve as better indicators of the fire. The two-wavelength Aethalometer (16) was designed to measure the optical absorption of ambient PM at 880 nm (black carbon, BC) and 370 nm (ultraviolet black carbon, UVBC). Certain organic aerosol components of wood combustion particles strongly absorb at 370 nm radiations relative to 880 nm. Allen et al. (17) have suggested that the difference in estimated black carbon mass as measured at these two wavelength (Delta-C ) UVB370 nm - BC880 nm) can serve as an indicator of wood combustion particles. Aethalometer measurements made in Philadelphia, PA during the Quebec forest fire event in July, 2002 showed a substantial enhancement in optical absorption at 370 nm relative to 880 nm (8). However, there is no published work on the use of Delta-C values to identify large fire episodes. Sandradewi et al. (18) used differences in light absorption at two other wavelengths to determine the contribution of wood burning and traffic emissions to ambient PM1 concentrations in a small village of Switzerland. In late May 2010, multiple forest fires were ignited in southern Quebec, Canada. Smoke from these fires significantly reduced the visibility in Quebec City, Montreal, and Ottawa. Transport of smoke to the northeastern United States occurred on May 31 with impacts on local air quality. The present study illustrates results of temporal variations of PM2.5, carbonaceous species (elemental carbon (EC), organic carbon (OC), BC and Delta-C), and atmospheric Hg concentrations (GEM, RGM, and PBM) measured at three rural sites in northern New York. These measurements show a relationship between concentration increases and transport from the forest fires. The correlations between Delta-C and other carbonaceous species and between Delta-C and PBM/ RGM were also investigated.
2. Experimental Methods During May-Jun 2010, measurements of atmospheric constituents were being continuously made at three regionally representative rural sites within northern New York. PM2.5 data were collected at the New York State Department of Environmental Conservation (NYSDEC) Whiteface Mountain base (WMB) monitoring site (44°23′35′′N, 73°51′32′′W, elevation ) 610 m) using a 30 °C Tapered Element Oscillating VOL. 44, NO. 22, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8435
FIGURE 1. MODIS satellite image taken on May 31, 2010, 1530 EDT. The red dots represent the locations of forest fires. The yellow dots represent the locations of the three sampling sites in northern New York. Microbalance (TEOM, Rupprecht & Patashnick Co., Inc., Albany, NY). OC and EC concentrations were measured at Clarkson University (CU) (44°39′43′′N, 75°00′02′′W, elevation ) 160 m) using a PM2.5 sharp cut cyclone and a dual system semicontinuous OC/EC analyzer (Sunset Laboratory, Forest Grove, OR). This dual system is a modification of the system described by Bauer et al. (19). In our system, two separate Sunset field OC/EC monitors are used rather than the specially built dual oven in the Bauer et al. unit. The semivolatile analyzer was not functioning properly during this period, so only the particulate OC/EC data were collected. The ambient aerosol was introduced into the system through a short stainless steel tube with no size segregation. A twowavelength Aethalometer (Model AE-42, Magee Scientific, Berkeley, CA) connected to the same PM2.5 sharp cut cyclone was used to measured BC and Delta-C at the CU site. The specific sampling and analysis methods of the R&P TEOM, Sunset Lab OC/EC analyzer and Aethalometer were described elsewhere (8, 20). Delta-C represents a difference in molecular absorbance at the two specified wavelengths (370 and 800 nm in this work). It is not a direct quantitative mass measurement. A recent field study conducted in Rochester, NY in November 8436
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 22, 2010
and December 2009 has shown that Delta-C values were higher by a factor of 3 during the night than the daytime values when wood burning was common (21). Although the method requires further study, the relationships between Delta-C and other measured species can be examined. These relationships may differ between locations and meteorological conditions. Atmospheric Hg data were collected with a Tekran speciation system (Tekran, Toronto, Canada) at the Huntington Forest (HF) site (43°54′36′′N, 74°13′12′′W, elevation ) 530 m) located in the Adirondack Park (a New York State Park Preserve). Ambient air was drawn through an elutriator/ acceleration jet and an impactor to remove particles larger than 2.5 µm. A potassium chloride coated annular denuder was used to capture RGM. PBM was collected through a regeneratable particulate filter. GEM concentrations were monitored at five-minute intervals during sampling mode (2 h). These stages were then desorbed (1 h) to measure system blanks, RGM, and PBM concentrations. This system was thus operated on a 3-h interval. Detailed descriptions of the mercury monitoring approach are given by Choi et al. (22). Meteorological parameters including ambient temperature, relative humidity, precipitation, wind direction, and wind speed were measured at both the WMB and HF sites.
FIGURE 2. Back-trajectories of air parcels ending at the sampling area at a height of 500 m above ground level, from the HYSPLIT model. The hollow squares represent the locations of the three sampling sites. The red triangles represent the locations of forest fires. The black circles represent the areas of high forest fire activities. Identification of the forest fire locations was based on the data available from the Forest Fire Protection Agency of Quebec (SOPFEU, http://sopfeu.poly9.com/). The transport of the forest fire plume to the sampling sites was observed using the National Aeronautics and Space Administration (NASA) moderate resolution imaging spectroradiometer (MODIS) and modeled with the National Oceanic and Atmospheric Administration (NOAA) Hybrid Single-Particle Lagrangian Integrated Trajectory (HYSPLIT) (23). HYSPLIT backward trajectories of 24 h were calculated with Eta Data Assimilation System (EDAS) (40 km grid). Trajectories were calculated from the center of the three sites using a 500 m starting height (22).
3. Results and Discussion There were 55 active forest fires reported and about 65 000 ha of forest was consumed in Quebec, Canada on May 31, 2010. A NASA MODIS satellite image taken on May 31 at 1530 EDT qualitatively indicates a clear smoke plume extending from the forest fires in Quebec to the northern New York (Figure 1). The HYSPLIT backward trajectories at a height of 500 m above ground level (Figure 2) show that fire plumes passed over the three sampling sites on May 31, 2010. The origin of air parcels advecting into the sampling area shifted to the west late in the afternoon of May 31 so that the area was only exposed to the fire aerosol over a limited time span. Figure 3 presents the temporal variations of BC, Delta-C, OC, EC, GEM, RGM, PBM, and PM2.5 concentration profiles measured from May 24 to Jun 7, 2010. The impact of the Canadian forest fires is clearly observed by the elevated concentrations of all pollutants on May 31. Table 1 summarizes the concentrations of all pollutants during the sampling period. The average PM2.5 mass concentration
measured at the WMB site on May 31 (71.5 ( 56.8 µg/m3) was higher by a factor of 18 than values measured under normal condition (May 24-Jun 7, except for May 31) and exceeded the U.S. EPA 24-h National Ambient Air Quality Standards (NAAQS). The maximum hourly average PM2.5 concentration (155.1 µg/m3) during the event is the highest value since April 2005 when data started being recorded at the site. The OC, EC, BC, and Delta-C values measured at the CU site increased by factors ranging from 5 to 50 during the forest fire episode compared to the normal condition. The negative values for Delta-C indicate little or no impacts from wood burning (20). The mean concentrations of episode-related RGM and PBM were statistically higher than those under normal condition. Previous studies of the impact of Canadian forest fires on ambient air pollutants also found elevated concentrations of atmospheric mercury and carbonaceous species during forest fire episodes. Table 2 summarizes the results of studies conducted in Philadelphia, PA and Petersham, MA. During the summer of 2002 in Philadelphia, smoke from a Canadian forest fire was transported to this urban site. The mean PM2.5 value during the fire was ∼3× greater than the mean during the rest of the measurement period. Compared with the normal ambient conditions, the OC and EC concentrations both significantly increased by a factor of 7. The mean concentration of BC was ∼3× higher than the mean observed during normal days. During July 2002, long-range transport of GEM in the forest fire plume was observed in Petersham, MA. The mean GEM concentration during the event was enhanced by 29% above the background level, which is slightly greater than the enhancement of 17% in this study. The relationships between Delta-C and other carbonaceous species during the forest fire episode and under normal conditions were compared in Figure 4. Delta-C values VOL. 44, NO. 22, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8437
FIGURE 3. Temporal variations of BC, Delta-C, OC, EC, GEM, RGM, PBM, and PM2.5 measured at three sampling sites within northern New York during the period of May 24-June 7, 2010.
TABLE 1. Mean and Standard Deviation of PM2.5, OC, EC, BC, Delta-C, GEM, RGM, and PBM Concentrations Measured during May 24-June 7, 2010 pollutant 3
PM2.5 (µg/m ) OC (µg/m3) EC (µg/m3) BC (µg/m3) Delta-C (µg/m3) GEM (ng/m3) RGM (pg/m3) PBM (pg/m3) a
forest fire episode
normal condition
ratio
site
71.5 ( 56.8 (n ) 24) 46.5 ( 30.7 (n ) 10) 5.0 ( 5.9 (n ) 10) 1.0 ( 0.6 (n ) 21) 0.6 ( 0.6 (n ) 21) 1.4 ( 0.1 (n ) 9) 3.1 ( 1.1 (n ) 9) 11.4 ( 3.0 (n ) 9)
3.8 ( 4.0 (n ) 334) 3.3 ( 2.5 (n ) 151) 0.1 ( 0.4 (n ) 151) 0.2 ( 0.1 (n ) 335) -0.01 ( 0.04 (n ) 335) 1.2 ( 0.3 (n ) 101) 1.6 ( 1.5 (n ) 101) 4.3 ( 2.7 (n ) 101)
18.8 14.1 50.1 5.1
WMB CU
1.2a 1.9 2.7
HF
Indicates there is no statistical difference between the two values.
TABLE 2. Mean and Standard Deviation of PM2.5, OC, EC, BC, Delta-C, and GEM Concentrations Measured in Previous Canadian Forest Fire Studies pollutant 3
PM2.5 (µg/m ) OC (µg/m3) EC (µg/m3) BC (µg/m3) Delta-C (µg/m3) GEM (ng/m3)
forest fire episode
normal condition
reference
79.8 ( 36.2 (n ) 82) 30.6 ( 16.2 (n ) 41) 2.9 ( 1.6 (n ) 41) 2.3 ( 1.5 (n ) 82) 1.3 ( 0.5 (n ) 82) 1.8 ( 0.1 (n ) 9)
22.6 ( 17.1 (n ) 205) 4.8 ( 2.1 (n ) 205) 0.4 ( 0.3 (n ) 205) 0.7 ( 0.4 (n ) 205) -0.2 ( 0.2 (n ) 205) 1.4 ( 0.1 (n ) 10)
8
were well correlated with OC, EC, and BC during the forest fire episode (r2 > 0.50, p < 0.01). Under normal conditions, 8438
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 22, 2010
13
no statistically significant correlations were observed between Delta-C value and OC, EC, and BC concentrations (r2 < 0.10).
FIGURE 5. Correlations between Delta-C and PBM during the forest fire episode and under normal conditions.
FIGURE 4. Correlations between Delta-C and other carbonaceous species (a) during the forest fire episode and (b) under normal condition. Negative correlations were found between Delta-C and EC (slope ) -0.23) and between Delta-C and BC (slope ) -0.80). The results indicate that the organic compounds emitted during the fire resulted in significant correlations between Delta-C values and OC, EC, and BC concentrations. Delta-C is essentially independent of the carbonaceous species during normal periods resulting in the poor correlations. Large values of Delta-C appear to provide a good indicator of wood fires. Figure 5 shows the correlations between PBM and DeltaC. The coefficients of determination during the forest fire episode were 0.90 (p < 0.01) with a slope of 10.15. Under normal conditions, the correlation coefficient was essentially zero. Note that the hourly average Delta-C data were merged into 2-h averages to match the PBM data. The enhanced slope and correlation coefficient during the forest fire episode relative to the normal condition indicate that PBM and Delta-C could both be emitted from forest fires and PBM seems to be more sensitive in forest fire smoke compared to Delta-C. Figure 6 presents the correlation between RGM and Delta-C. The correlation coefficients during the forest fire episode and under normal condition were 0.73 (p < 0.01) with a slope of 2.59. Under normal conditions, there was again no correlation. Delta-C can be considered as an indicator of biomass combustion. Forest fires produce large amounts of particulate matter, and RGM can be easily deposited on those particle surface forming PBM. During the normal period, occasional high PBM peaks indicate the impacts from local anthropogenic sources, and the Delta-C value is usually negative or below zero. Little or no correlation between RGM and Delta-C
FIGURE 6. Correlations between Delta-C and RGM during the forest fire episode and under normal conditions. was expected under normal condition compared to forest fire episode. The positive slope and enhanced correlation coefficient during the forest fire relative to the normal period suggest that PBM/RGM and Delta-C could both be emitted from forest fires. GEM generally represents the largest portion of Hg emitted from forest fires. However, its atmospheric concentration is 100-1000× higher than PBM and RGM and relatively constant compared to the other two species during the forest fire episode. GEM cannot be easily removed from the atmosphere by precipitation. Its behavior is quite different from PBM, RGM, and Delta-C. The reactive mercury that is released in a fire can easily attach to the high concentration of particulate matter that is simultaneously released. This process results in a larger correlation coefficient between Delta-C and PBM than that between Delta-C and RGM. This work illustrates the significance of forest fires and the meteorological conditions on the downwind impacts of atmospheric mercury and carbonaceous species. The correlation coefficient between Delta-C and other carbonaceous species may serve as an indicator of forest fire smoke. Given the marked changes in PBM, it may serve as a more useful tracer of forest fires over distances of several hundred kilometers relative to GEM. However, the Delta-C concentration changes are more readily measured.
Acknowledgments This work was supported by the United States Environmental Protection Agency (U.S. EPA), the New York State Energy VOL. 44, NO. 22, 2010 / ENVIRONMENTAL SCIENCE & TECHNOLOGY
9
8439
Research and Development Authority (NYSERDA), the U.S. EPA Atmospheric Clean Air Markets Division and NADP Hg Monitoring Network, and the New York State Department of Environmental Conservation (NYSDEC). Although the research described in this work has been partly funded by the U.S. EPA, it has not been subjected to the agency’s required peer and policy review and therefore, does not necessarily reflect the views of the agency and no official endorsement should be inferred. The authors gratefully acknowledge the NOAA Air Resources Laboratory (ARL) for the provision of the HYSPLIT transport and dispersion model and/or the READY website (http://www.arl.noaa.gov/ready.html) used in this work.
Literature Cited (1) Crutzen, P. J.; Andreae, M. O. Biomass burning in the tropics: impact on atmospheric chemistry and biogeochemical cycles. Science 1990, 250, 1669–1678. (2) Lavous, D.; Liousse, C.; Cachier, H.; Stocks, B. J.; Goldammer, J. G. Modeling of carbonaceous particles emitted by boreal and temperate wildfires at northern latitudes. J. Geophys. Res. 2000, 105, 26, 87126, 890. (3) Kunzli, N.; Avol, E.; Wu, J.; Gauderman, W. J.; Rappaport, E.; Millstein, J.; Bennion, J.; McConnell, R.; Gilliland, F. D.; Berhane, K.; Lurmann, F.; Winer, A.; Peters, J. M. Health effects of the 2003 Southern California wildfires on children. Am. J. Respir. Crit. Care Med. 2006, 174, 1221–1228. (4) Wegesser, T. C.; Pinkerton, K. E.; Last, J. A. California wildfires of 2008: Coarse and fine particulate matter toxicity. Environ. Health. Persp. 2009, 117, 893–897. (5) Conny, J. M.; Slater, J. F. Black carbon and organic carbon in aerosol particles from crown fires in the Canadian boreal forest. J. Geophys. Res 2002, 107, 4116, doi: 10.1029/2001JD001528. (6) Wotawa, G.; Trainer, M. The influence of Canadian forest fires on pollutant concentrations in the United States. Science 2000, 288, 324–328. (7) Sapkota, A.; Symons, J. M.; Kleissl, J.; Wang, L.; Parlange, M. B.; Ondov, J.; Breysse, P. N.; Diette, G. B.; Eggleston, P. A.; Buckley, T. J. Impact of the 2002 Canadian forest fires on particulate matter air quality in Baltimore City. Environ. Sci. Technol. 2005, 39, 24–32. (8) Jeong, C.-H.; Hopke, P. K.; Kim, E.; Lee, D.-W. The comparison between thermal-optical transmittance elemental carbon and Aethalometer black carbon measured at multiple monitoring sites. Atmos. Environ. 2004, 38, 5193–5204. (9) Schroeder, W. H.; Munthe, J. Atmospheric mercury - An overview. Atmos. Environ. 1998, 32 (5), 809–822.
8440
9
ENVIRONMENTAL SCIENCE & TECHNOLOGY / VOL. 44, NO. 22, 2010
(10) Munthe, J.; McElroy, W. J. Some aqueous reactions of potential importance in the atmospheric chemistry of mercury. Atmos. Environ. 1992, 26A, 553–557. (11) Bullock, O. R. Current methods and research strategies for modeling atmospheric mercury. Fuel Process. Technol. 2000, 65-66, 459–471. (12) Obrist, D.; Hallar, A. G.; McCubbin, I.; Stephens, B. B.; Rahn, T. Atmospheric mercury concentrations at Storm Peak Laboratory in the Rocky Mountains: Evidence for long-range transport from Asia, boundary layer contributions, and plant mercury uptake. Atmos. Environ. 2008, 42, 7579–7589. (13) Sigler, J. M.; Lee, X.; Munger, W. Emission and long-range transport of gaseous mercury from a large-scale Canadian boreal forest fire. Environ. Sci. Technol. 2003, 37, 4343–4347. (14) Wiedinmyer, C.; Friedli, H. Mercury emission estimates from fires: an initial inventory for the United States. Environ. Sci. Technol. 2007, 41, 8092–8098. (15) Friedli, H. R.; Arellano, A. F.; Cinnirella, S.; Pirrone, M. Initial estimates of mercury emissions to the atmosphere from global biomass burning. Environ. Sci. Technol. 2009, 43, 3507–3513. (16) Hansen, A. D.; Rosen, H.; Novakov, T. The aethalometer - An instrument for the real time measurements of optical absorption by aerosol particles. Sci. Total Environ. 1984, 36, 191–196. (17) Allen, G. A.; Babich, P.; Poirot, R. L. Evaluation of a new approach for real time assessment of wood smoke PM. In Proceedings of the Regional and Global Perspectives on Haze: Causes, Consequences and Controversies - Visibility Specialty Conference; Air & Waste Management Association: Asheville, NC, 2004. (18) Sandradewi, J.; Prévôt, A. S. H.; Szidat, S.; Perron, N.; Alfarra, M. R.; Lanz, V. A.; Weingartner, E.; Baltensperger, U. Using aerosol light absorption measurements for the quantitative determination of wood burning and traffic emission contributions to particulate matter. Environ. Sci. Technol. 2008, 42, 3316– 3323. (19) Bauer, J. J.; Yu, X. Y.; Cary, R.; Laulainen, N.; Berkowitz, C. Characterization of the Sunset semi-continuous carbon aerosol analyzer. J. Air. Waste. Manag. Assoc. 2009, 59, 826–833. (20) Grover, B. D.; Kleinman, M.; Eatough, N. L.; Eatough, D. J.; Cary, R. A.; Hopke, P. K.; Wilson, W. E. Measurement of fine particulate matter nonvolatile and semi-volatile organic material with the Sunset laboratory carbon aerosol monitor. J. Air. Waste. Manag. Assoc. 2008, 58, 72–77. (21) Wang, Y.; Hopke, P. K.; Rattigan, O. V.; Zhu, Y. Characterization of ambient black carbon and wood burning particles in two urban areas. Environ. Pollut. 2010, Submitted. (22) Choi, H.-D.; Holsen, T. M.; Hopke, P. K. Atmospheric mercury (Hg) in the Adirondacks: concentrations and sources. Environ. Sci. Technol. 2008, 42 (15), 5644–5653. (23) Draxler, R. R.; Hess, G. D. HYSPLIT 4 USER’s Guide. In NOAA Technical Memorandum ERL ARL-230. Washington, DC, 2005.
ES1024806
