ARTICLE IN PRESS
Atmospheric Environment 38 (2004) 1235–1246
Ambient ammonia and ammonium aerosol across a region of variable ammonia emission density J.T. Walkera,*, Dave R Whitallb, Wayne Robargec, Hans W. Paerld a
Air Pollution Prevention and Control Division, National Risk Management Research Laboratory, US Environmental Protection Agency, E305-2, MD-63, Research Triangle Park, NC 27711, USA b Center for Coastal Monitoring and Assessment, NOAA, Silver Spring, MD 20910, USA c Department of Soil Science, North Carolina State University, Raleigh, NC 27695, USA d University of North Carolina at Chapel Hill, Institute of Marine Sciences, Morehead City, NC 28557, USA Received 7 November 2002; accepted 30 November 2003
Abstract We present 1 year of ambient ammonia (NH3), ammonium (NH+ 4 ), hydrochloric acid (HCl), chloride (Cl ), nitric 2 acid (HNO3), nitrate (NO3 ), nitrous acid (HONO), sulfur dioxide (SO2), and sulfate (SO4 ) concentrations at three sites in the Coastal Plain region of North Carolina. The three sites, Clinton, Kinston, and Morehead City, are located in counties with total NH3 emission densities of 4800, 2280, and 320 kg NH3-N km2 yr1, respectively. Average NH3 concentrations were 5.32, 2.46, and 0.58 mg m3 at Clinton, Kinston, and Morehead City, respectively. Average NH+ 4 2 concentrations were 1.84, 1.25, and 0.91 mg m3, and total concentrations of inorganic (NH+ 4 +NO3 + SO4 +Cl ) particulate matter with aerosol diameters o2:5 mm (PM2.5) were 8.66, 6.35, and 5.31 mg m3 at Clinton, Kinston, and Morehead City, respectively. NH3 concentrations were highest during the summer at all sites, with summer-to-winter concentration ratios of 2.40, 5.70, and 1.70 at Clinton, Kinston, and Morehead City, respectively. NH3 concentrations were higher at night at the Clinton site, during the day at the Kinston site, and day vs. night concentrations were similar at the Morehead City site. NH+ 4 concentrations were highest during the winter at all sites, though this may not be + representative of all years. Average daytime concentrations of NH+ 4 were similar to night values at all sites. NH4 2 2 aerosol was primarily associated with SO4 at all sites, though the degree of SO4 neutralization was highest at Clinton and lowest at Morehead City. NH+ 4 aerosol formation appeared to be acid-gas-limited at the Clinton site during all seasons and during the spring and summer at the Kinston site. This study shows that agricultural NH3 emissions influence local ambient concentrations of NH3 and PM2.5. r 2003 Elsevier Ltd. All rights reserved.
Keywords: Atmospheric nitrogen; Agricultural emissions; Acid gases; PM2.5; Denuders
1. Introduction As the primary gaseous base in the atmosphere, ammonia (NH3) influences the acidity of solid- and aqueous-phase aerosol species, cloud water, and precipitation. Ammonia may be either wet- or drydeposited as a gas, or react with sulfuric (H2SO4), nitric *Corresponding author. Tel.: +1-919-541-2288; fax: +1919-541-7891. E-mail address:
[email protected] (J.T. Walker).
(HNO3), and hydrochloric (HCl) acids to form ammonium sulfate (NH4)2SO4, bisulfate (NH4HSO4), nitrate (NH4NO3), and chloride (NH4Cl) aerosols, thereby contributing to inorganic ambient particulate matter. In some cases, these inorganic species may contribute significantly to total PM2.5 (particulate matter with aerodynamic diameter o2:5 mm), which is typically composed of sulfate (SO2 4 ), nitrate (NO3 ), ammonium + (NH4 ), organic carbon, elemental carbon, hydrogen ions, and an assortment of transition metals (US EPA, 1996). Estimates of the cumulative contribution of
1352-2310/$ - see front matter r 2003 Elsevier Ltd. All rights reserved. doi:10.1016/j.atmosenv.2003.11.027
ARTICLE IN PRESS 1236
J.T. Walker et al. / Atmospheric Environment 38 (2004) 1235–1246
+ SO2 4 , NO3 , and NH4 to total PM2.5 mass at eastern US sites range from 40% to 65% (Malm et al., 1994; US EPA, 1996; Tolocka et al., 2001). As a component of atmospheric nitrogen (N) deposition, NH3 and NH+ 4 also play a role in the cycling of N in terrestrial and aquatic ecosystems. The accumulation of reactive nitrogen, which includes NH3 and NH+ 4 , in environmental reservoirs may have both beneficial and detrimental effects on the biosphere (Galloway and Cowling, 2002). In natural systems where N is the limiting nutrient (Vitousek and Howarth, 1991), atmospherically derived reactive N may have beneficial effects on productivity, including increased photosynthesis (Sievering et al., 2000) and accumulation of inorganic soil N (Padgett et al., 1999). Recent studies also indicate that enhanced N deposition may increase the carbon storage capacity of temperate forests (Sievering, 1999). When N input exceeds system requirements, however, environmental stresses such as soil acidification (Roelofs et al., 1985), forest decline (Nihlgard, 1985), and eutrophication of surface waters (Paerl, 1995; Paerl and Whitall, 1999) may occur. It is, therefore, necessary to characterize the magnitude and spatiotemporal variability of atmospheric NH3 and NH+ 4 concentrations. Globally, domestic animals are the largest source (22 Tg N yr1, 1 Tg=1012 g) of atmospheric NH3, comprising approximately 40% of natural and anthropogenic emissions combined, while synthetic fertilizers and agricultural crops together contribute an additional 12.6 Tg NH3-N yr1 (23% of total emissions) (Bouwman et al., 1997). Within and downwind of agricultural regions, NHx (NHx=NH3+NH+ 4 ) therefore represents a significant fraction of atmospherically derived N entering terrestrial and aquatic systems (Asman and van Jaarsveld, 1992; Whitall and Paerl, 2001). Over the past decade, the Coastal Plain region of North Carolina has experienced a significant increase in agricultural NH3 emissions, owing primarily to growth in swine and poultry populations (Walker et al., 2000a). Beginning with the increase in NH3 emissions, the concentration of NH+ 4 in precipitation has also increased in this part of the state (Walker et al., 2000a, b; Paerl and Whitall, 1999). Only recently, however, have ambient concentrations of NH3 and NH+ 4 been investigated in this region of increased NH3 emissions (Robarge et al., 2002). In this study, we present 1 year of ambient NH3, NH+ 4 , HCl, chloride (Cl ), NO3 , HNO3, nitrous acid 2 (HONO), SO4 , and sulfur dioxide (SO2) concentrations at three sites with widely varying county scale NH3 emissions in the Coastal Plain region of North Carolina.
2. Methods Ambient concentrations of gases and aerosols were measured using the annular denuder system (Ferm,
1979; Allegrini et al., 1987; US EPA, 1997). The configuration used in this study included: two threechannel annular denuder tubes (30 mm OD 242 mm long) arranged in sequence; an aluminum Teflon-coated cyclone (10 l min1, o2:5 mm cutoff); and a 2-stage Teflon filter pack. The first denuder tube was coated with sodium carbonate, which traps HNO3, HONO, HCl, and SO2. The second denuder tube was coated with citric acid to collect NH3. Particles were collected downstream of the denuders in a filter pack that contained a 2 mm Teflon filter followed by a 0.8 mm nylon filter. The Teflon filter essentially traps all of the PM2.5 that passes through the preceding cyclone and denuder tubes. The nylon filter captures NO 3 as HNO3 produced by the dissociation of NH4NO3 on the Teflon filter. The system operated on 12-h day- and night-cycles (0600–1800 hours day-cycle; 1800–0600 h night-cycle), yielding 12-h average concentrations. Ref. US EPA (1997) contains more detailed information on the annular denuder-filter pack method used in this study. Denuder tube extractions were performed in situ and extracted samples were then analyzed at the Analytical Service Laboratory of North Carolina State University’s Department of Soil Science. Standard colorimetric procedures were used to determine soluble NH+ 4 -N and NO 3 -N using a flow injection autoanalyzer. Soluble 2 Cl, NO were determined using ion 2 , and SO4 chromatography. Total uncertainty was determined using paired field observations (n ¼ 90) from the Clinton site (Robarge et al., 2002). For NH3 and SO2, the total uncertainty expressed as a percent coefficient of variation (%CV) was o10: For HONO, HNO3, and HCl, the %CV values were 17.5, 31.0, and 43.0, respectively. For 2 particulate NH+ 4 , SO4 , NO3 , and Cl , %CV values were 13.0, 18.0, 25.0, and 78.0, respectively (Robarge et al., 2002). The sites in this study are located in the Coastal Plain physiographic region of North Carolina. Data were collected at the Clinton Horticultural Crops Research Station, Clinton, Sampson County, NC, from 1 January 2000 to 31 December 2000; Lenoir County Community College, Kinston, Lenoir County, NC, from 16 May 2000 to 31 December 2000; and University of North Carolina, Chapel Hill’s Institute of Marine Sciences, Morehead City, Carteret County, NC, from 17 January 2000 to 19 December 2000. The Kinston and Morehead City sites are approximately 75 km northeast and 155 km east-southeast of the Clinton site, respectively. Data at the Kinston and Morehead City sites were collected on an approximate 3-day cycle—two consecutive 12-h measurements approximately every 3rd day. Data at the Clinton site were collected continuously. To assess the potential effect of sampling frequency on data representativeness, the Clinton data set was reduced to only those days when measurements were taken at Morehead City. Median, as well as 1st and 3rd quartile
ARTICLE IN PRESS J.T. Walker et al. / Atmospheric Environment 38 (2004) 1235–1246
values between the full and reduced Clinton datasets were less than 5% different for both NH3 and NH+ 4 , which both followed an approximately lognormal distribution at each site.
3. Results and discussion 3.1. Ammonia emissions To assess the spatial variability and magnitude of NH3 emissions across eastern North Carolina, county scale emissions were estimated using the Carnegie Mellon University NH3 Inventory Version 2.0 (Strader et al., 2001). Source categories included livestock, fertilizer, domestic and wild animals, humans, mobile, industrial, and publicly owned treatment works (POTWs). Emissions from natural soils and vegetation were not included, due to the large uncertainty in emission factors for these categories. Livestock activity level data (NCDACS, 2002) were updated to 2000 values, and the emission factors proposed by Battye et al. (1994) were used, with the exception of the factor for hogs, which was reduced to 5.36 kg NH3 animal1 yr1 (Asman, 1992). The emission factors proposed by Battye et al. (1994) and Asman (1992) were used for fertilizer. Using this inventory, livestock and fertilizer represent 78% and 10% of total (131,500 ton NH3-N) statewide NH3 emissions, respectively. Detailed information on the activity level data and emission factors for the remaining source categories can be found in the inventory user guide (Strader et al., 2001). Fig. 1 shows county level NH3 emission density for the State of North Carolina. As previously mentioned, livestock and fertilizer collectively account for approximately 90% (115,800 ton NH3-N) of total statewide emissions. For the purposes of this study, the three sites investigated have been assigned the emission density of the county within which the site resides. Emission densities range from 4800 kg NH3-N km2 at the Clinton site to 320 kg NH3-N km2 at Morehead City,
Fig. 1. County scale NH3 emission density for North Carolina along with measurement sites. Livestock activity data represent 2000 levels. All other activity data represent 1996 levels.
1237
while the Kinston site has a corresponding emission density of 2280 kg NH3-N km2. 3.2. Ambient ammonia Fig. 2 and Table 1 show summary statistics for the three sites for all analytes. Ammonia concentrations vary by almost an order of magnitude across sites, and average concentrations are positively correlated with corresponding county level NH3 emission densities. The mean value for the Clinton site (5.3 mg NH3 m3) is comparable to ambient concentrations measured in regions of similar emission densities (Table 2), also primarily from livestock production, in the Netherlands (Buijsman et al., 1998). In their study, Buijsman et al. (1998) characterize emissions of 0–4000 kg NH3N km2 yr1 as low and 4000–16,500 kg NH3N km2 yr1 as moderate. The mean value at the Clinton site is lower than the mean value reported by McCulloch et al. (1998) for September through December 1997 at a nearby commercial swine facility (Table 2). The mean value at the Kinston site (2.46 mg NH3 m3) is similar to the value reported by Buijsman et al. (1998) at the site in their study characterized as having low emissions (Table 2). This value is higher than the values reported by Pryor et al. (2001) at a forested site in southern Indiana considered to be influenced by agriculture, though located in an area of relatively low local emissions (Table 2). The mean value at the Morehead City site (0.58 mg NH3 m3) is within the range of concentrations (mg NH3 m3) recently reported for other non-agricultural US sites and those previously summarized by Langford et al. (1992) (Table 2). The mean concentration at Morehead City is also typical of NH3 concentration data collected in US urban areas (Table 2). Numerous researchers have reported a seasonal cycle in ambient NH3 concentrations (Robarge et al., 2002; Pryor et al., 2001; Lefer et al., 1999; Langford et al.,
Fig. 2. Mean ambient concentrations across sites. Bar extensions represent the upper 95% confidence limit for the mean.
ARTICLE IN PRESS J.T. Walker et al. / Atmospheric Environment 38 (2004) 1235–1246
1238
Table 1 Annual summary statistics of 12-h average gas and aerosol concentrations (mg m3) during 2000 Site
Period
Analyte
N
Mean
SD
10th percentile
Median
90th percentile
Clinton
Day
NH3 HNO3 HONO HCl SO2 NH+ 4 NO 3 Cl SO2 4 NH3 HNO3 HONO HCl SO2 NH+ 4 NO 3 Cl SO2 4
283 283 283 283 283 282 282 282 282 283 283 283 283 283 280 280 280 280
3.91 1.14 0.15 0.56 4.87 1.83 2.12 0.04 4.77 6.72 0.50 0.68 0.32 3.11 1.85 2.11 0.07 4.53
2.54 1.12 0.24 0.55 4.49 1.05 2.01 0.06 3.51 5.77 0.59 0.66 0.44 2.97 1.17 2.29 0.14 3.28
0.91 0.09 0.07 0.10 0.61 0.57 0.40 0.02 1.16 1.14 0.09 0.07 0.10 0.57 0.59 0.34 0.02 1.36
3.67 0.78 0.07 0.44 3.77 1.74 1.41 0.02 3.70 5.16 0.34 0.49 0.19 2.16 1.63 1.38 0.02 3.56
6.98 2.72 0.35 1.17 10.15 3.23 4.52 0.10 9.61 13.70 1.16 1.56 0.66 7.03 3.35 4.54 0.18 9.23
NH3 HNO3 HONO HCl SO2 NH+ 4 NO 3 Cl SO2 4 NH3 HNO3 HONO HCl SO2 NH+ 4 NO 3 Cl SO2 4
132 132 132 132 132 132 132 132 132 134 134 134 134 134 132 132 132 132
3.06 0.35 0.27 0.27 3.32 1.30 1.36 0.05 4.05 1.87 0.25 0.56 0.20 1.14 1.20 1.37 0.04 3.34
3.00 0.33 0.31 0.25 4.08 0.77 1.21 0.07 3.08 1.81 0.40 0.60 0.18 2.09 0.83 1.64 0.05 2.46
0.28 0.09 0.07 0.10 0.20 0.31 0.27 0.02 0.99 0.36 0.09 0.07 0.10 0.05 0.32 0.27 0.02 1.00
2.19 0.23 0.07 0.19 1.79 1.20 1.04 0.02 3.38 1.41 0.09 0.40 0.10 0.32 1.04 0.84 0.02 2.80
7.78 0.81 0.74 0.56 8.29 2.30 2.92 0.12 8.00 4.21 0.50 1.38 0.45 2.56 2.11 2.79 0.11 5.94
NH3 HNO3 HONO HCl SO2 NH+ 4 NO 3 Cl SO2 4 NH3 HNO3 HONO HCl SO2 NH+ 4 NO 3 Cl SO2 4
137 137 137 137 137 90 90 90 90 142 142 142 142 142 99 99 99 99
0.58 0.22 0.20 0.26 1.37 0.91 0.73 0.28 3.58 0.59 0.22 0.17 0.40 2.04 0.90 0.83 0.30 3.09
0.58 0.19 0.44 0.20 1.89 0.47 0.55 0.34 2.40 0.54 0.27 0.26 1.62 2.77 0.75 0.86 0.31 2.71
0.06 0.14 0.07 0.10 0.14 0.31 0.18 0.02 0.76 0.06 0.14 0.07 0.10 0.17 0.13 0.18 0.02 0.35
0.41 0.14 0.10 0.20 0.70 0.93 0.58 0.14 3.48 0.47 0.14 0.10 0.21 0.90 0.82 0.58 0.20 2.72
1.42 0.45 0.44 0.51 3.08 1.59 1.51 0.74 6.79 1.16 0.27 0.30 0.49 4.92 1.87 1.64 0.78 6.76
Night
Kinston
Day
Night
Morehead City
Day
Night
N represents the number of 12-h average observations. Observations are grouped by day/night periods.
ARTICLE IN PRESS J.T. Walker et al. / Atmospheric Environment 38 (2004) 1235–1246
1239
Table 2 Ambient NH3 concentrations summarized by study Concentration (mg m3)
Land use
Comment
3.0 11.0 10.48 0.65–1.2 0.26 0.34 0.02 0.62–1.47 0.29 0.22 0.21 0.16 0.21 0.23 0.63 4.75 0.38–1.49 0.63–0.72 1.18
Agricultural Agricultural Agricultural Agricultural Agricultural Non-agricultural Non-agricultural Non-agricultural Non-agricultural Non-agricultural Non-agricultural Non-agricultural Non-agricultural Non-agricultural Non-agricultural Non-agricultural Urban Urban Urban
Low NH3 emissionsa Moderate NH3 emissionsa Fallb Springc Winterc High elevation, summer and falld High elevation, summere High elevation, summerf High elevation, summerg Coastal, summerg Forest, summerg Foresth Wetland, summerg Wetland, summerg Desert, summerg Grassland, summerg Pittsburgh, PA; summeri Research Triangle Park, NC; fallj Vinton, VA; summerk
a
Buijsman et al. (1998). McCulloch et al. (1998). c Pryor et al. (2001). d Tarnay et al. (2001). e Rattray and Sievering (2001). f Aneja et al. (1998). g Langford et al. (1992). h Lefer et al. (1999). i McCurdy et al. (1999). j Sickles et al. (1990). k Leaderer et al. (1999). b
1992), with maximum concentrations occurring during warm months. This seasonal cycle in ambient concentrations is in agreement with the temperature dependence of aqueous-phase partitioning between NH3 and NH+ 4 , as well as the equilibrium between aqueous- and gas-phase NH3 as predicted by Henry’s law, which results in increasing ammonia emissions from animal manure, soils, and vegetation with increasing temperature (Asman et al., 1998; Ni, 1999). This model has been confirmed for natural and agricultural systems (Langford et al., 1992; Aneja et al., 2000; Ni et al., 2000; Harper et al., 2000; Roelle and Aneja, 2002). Table 3 shows seasonal median ambient concentrations of NH3 across the three sites in this study. Concentrations reach a maximum during warm months at all sites. Due to the lower number of winter observations at Kinston and Morehead City relative to the other seasons, seasonality is illustrated in Table 3 by comparing ratios of summer to fall median concentrations. The lowest ratio of summer to fall concentrations (2.06) is observed at the non-agricultural Morehead City site. The maximum value of this ratio (3.16) is observed
at the Kinston site. To inspect the pronounced seasonality at the Kinston site more closely, we examined the relationship between 12-h average wind direction and NH3 concentration. Winds at the Kinston site were primarily from the the southwest during spring and summer and from the north-northeast during fall and winter (SCONC, 2002), which is consistent with the climatology of the region. NH3 concentrations were higher within all wind quadrants during spring and summer. The highly amplified seasonality at the Kinston site may result primarily from increased emission rates of local agricultural NH3 sources during spring and summer, though transport from downwind sources to the southwest (Fig. 1) during these seasons appears to influence concentrations as well. At Morehead City, winds were primarily from the southwest during spring and summer and from the north-northeast during fall and winter (SCONC, 2002). Researchers have also noted a diurnal pattern in NH3 concentrations. Typically, concentrations peak in the afternoon before decreasing to a minimum in the early morning (Langford et al., 1992). It should be noted that
ARTICLE IN PRESS J.T. Walker et al. / Atmospheric Environment 38 (2004) 1235–1246
1240
Table 3 3 Seasonal median NH3 and NH+ 4 concentrations (mg m ), with number of observations in parenthesis, along with the ratio of summer to fall median concentrations Clinton
Kinston
Morehead city
NH3
Winter Spring Summer Fall Summer/fall
2.59 4.62 6.18 2.85 2.17
(155) (115) (166) (130)
0.49 3.93 2.72 0.86 3.16
(24) (40) (106) (96)
0.33 0.40 0.72 0.35 2.06
(44) (64) (87) (84)
NH+ 4
Winter Spring Summer Fall Summer/fall
1.88 1.69 1.69 1.45 1.17
(155) (115) (166) (130)
1.56 1.14 0.98 1.23 0.80
(24) (40) (106) (96)
0.95 0.95 0.40 0.39 1.03
(44) (64) (87) (84)
Table 4 3 Day vs. night median NH3 and NH+ 4 concentrations (mg m ) during fall and summer Season
Site
Perioda
N
NH3
NH+ 4
Summer
Clinton
D N D N D N
83 83 42 45 53 53
5.06b 9.27 0.70 0.73 3.59b 1.65
1.70 1.68 0.93 0.79 1.11 0.89
D N D N D N
65 65 42 42 48 48
2.88 2.77 0.33 0.49 0.86 0.83
1.62 1.40 0.62 0.34 1.26 1.17
IMS Kinston
Fall
Clinton IMS Kinston
a
D=Day, N=Night. Indicates a statistically significant difference between day and night median concentrations at pp0:01 (based on a non-parametric two-sample median test (SAS, 1999)). b
the sites in the studies summarized by Langford et al. (1992) experienced relatively low NH3 concentrations. This pattern is thought to be a function of higher temperatures, and subsequent elevated NH3 emissions, during the day. Buijsman et al. (1998) observed the same pattern at sites in low NH3 emission areas and the opposite pattern at sites within high emission areas. Table 4 shows the difference in average day (0600–1800) vs. night (1800–0600) concentrations. At the Clinton site, concentrations are considerably higher at night. This is consistent with the pattern observed by Buijsman et al. (1998) in high NH3 emission areas, where higher concentrations at night are likely the result of accumulation and inefficient vertical mixing within a relatively shallow boundary layer. Also, the difference between day and night concentrations is larger during the spring and summer, suggesting an influence of higher emissions
during warm months. Concentrations are higher during the day at the Kinston site, which is the pattern observed by Buijsman et al. (1998) at sites within low NH3 emission areas. Buijsman et al. (1998) suggested that the transport of NH3 from downwind sources was the cause of higher daytime concentrations, while dry deposition and conversion to aerosol exceeded the contribution from transport at night, resulting in lower concentrations. In this study, the difference between average day and night concentrations is greatest during spring and summer, when winds are primarily from the southwest. As suggested by Buijsman et al. (1998), dry deposition of NH3 at night may exceed the contribution from transport, resulting in higher concentrations during the day. The lack of diurnal variability in NH+ 4 concentrations at the Kinston site suggests that conversion of NH3 to aerosol is not responsible for the diurnal NH3 pattern
ARTICLE IN PRESS J.T. Walker et al. / Atmospheric Environment 38 (2004) 1235–1246
(Table 4). At the Morehead City site, there is little difference between average day and night concentrations (Table 4). Lefer et al. (1999) also observed a lack of significant diurnal variability at the Harvard Forest site. This pattern may be typical of sites that experience minimal impact from local NH3 sources. 3.3. Ambient ammonium aerosol Fig. 2 shows that ambient NH+ 4 aerosol concentrations vary across sites and are also positively correlated with corresponding NH3 emission densities. As a data screening process, NH+ 4 values in this study were not considered for analysis if the molar concentrations of Cl and NO 3 individually or cumulatively exceeded the molar concentration of NH+ 4 . This reduced the original number of NH+ 4 values by o1% at Clinton and Kinston and 28% at Morehead City. The larger percentage at Morehead City is primarily due to occurrences of Cl>NH+ 4 , which may indicate the presence of NaCl aerosol in the PM2.5 size range. The range of mean concentrations observed in this study (0.91–1.84 mg m3) is similar to the range of concentrations (0.89– 1.80 mg m3) measured at Clean Air Status and Trends Network (CASTNet) sites in the southeast US during 2000. CASTNet is a national air monitoring program under which weekly average NH+ 4 aerosol concentrations are measured by the filterpack method (US EPA, 2002; Sickles et al., 1999). At the four CASTNet sites that were operational in North Carolina throughout 2000, annual mean NH+ concentrations, calculated 4 from weekly average concentrations, ranged from 0.93 to 1.72 mg m3 (CASTNet, 2002). The annual mean concentration measured at Clinton (1.84 mg m3) exceeds the highest CASTNet value by approximately 7%. It should be noted that none of these sites are located in counties with high NH3 emission densities. The annual mean concentration measured at Morehead City (0.91 mg m3) in this study and the Beaufort CASTNet site (0.93 mg m3), which is o10 km away, differ by 2%. Sulfur dioxide and HNO3 (Fig. 2), the major precursors to NH+ 4 formation, show the same pattern of concentrations as NH+ 4 , decreasing from the Clinton site to the Morehead City site. The National Emission Trends (NET) Database was used to examine the relative differences in county level SO2 and NOx emissions across the three sites (US EPA, 1999). The NET database is released every 3 years, with 1999 representing the latest available data. Higher SO2 concentrations at the Clinton site are likely due to the location of several point sources within 5–10 km to the west-southwest. While total 1999 county scale emissions are highest for the Kinston site, total county emissions from point sources are dominated by a single facility approximately 15 km to the northeast of the site. The remaining SO2 point sources within 15 km of the
1241
Kinston site are confined to the west-northwest, and therefore do not significantly influence SO2 concentrations at the measurement site. The spatial distributions of NOx sources within the three counties are similar to SO2, and total NOx emissions are highest in Sampson County where the Clinton site is located. This likely contributes to higher observed concentrations of HNO3 at Clinton relative to the other sites. Since SO2 and NOx are precursors to H2SO4 and HNO3, respectively, both SO2 and NO 4 3 concentrations should display similar spatial trends across sites. NH+ 4 , which is associated 2 with NO 3 and SO4 in aerosol form, also follows the same pattern. Similarly to NH3, concentrations of NH+ 4 have been shown to vary by season in both ambient air and precipitation, with higher values occurring during warm months (Robarge et al., 2002; Whitall and Paerl, 2001; Walker et al., 2000a; Sickles, 1999). Table 3 indicates that the seasonal NH+ 4 concentrations observed in this study reach a minimum during warm months, contrary to the results of other researchers (Sickles, 1999). To check the representativeness of this observed seasonality, North Carolina CASTNet data were analyzed across the period 1994–1999. At all four sites, concentrations typically achieved a maximum in spring or summer. In 2000, however, the seasonal cycle across sites is marked by above average fall concentrations and below average summer concentrations. This is coincident with higher than normal fall and lower than normal summer CASTNet values of SO2 4 during 2000. This comparison indicates that the seasonality observed in this study may not be representative of the general seasonality in NH+ 4 concentrations. As shown in Table 4, NH+ 4 did not exhibit significant diurnal variability at any of the sites. Consistent with the diurnal pattern in 2 NH+ 4 , average day vs. night SO4 and NO3 concentrations at individual sites were not significantly different at the 99% confidence level. Annual mean inorganic PM2.5 concentrations, which + include the cumulative mass of SO2 4 , NO3 , NH4 , and 3 Cl , were 8.66, 6.35, and 5.31 mg m at Clinton, Kinston, and Morehead City, respectively. Fig. 3 shows that total mass is less variable across seasons at the Clinton site and most variable at the Morehead City site. With respect to percentage of total mass, SO2 is the 4 most important species on average across all sites. Nitrate contributes more significantly to total mass during colder months when SO2 oxidation rates are reduced in response to lower concentrations of oxidants such as OH. Table 5 shows the molar ratios of the individual constituents of NH+ 4 aerosol. On average, the + ratio of NO to NH (E0.30) is consistent across sites. 3 4 Very little Cl is present at Clinton and Kinston, while E17% of NH+ at the 4 is associated with Cl coastal Morehead City site (Table 5). The Cl values should be interpreted cautiously, however, due to the
ARTICLE IN PRESS 1242
J.T. Walker et al. / Atmospheric Environment 38 (2004) 1235–1246
potential low bias in concentrations caused by dissociation of NH4Cl from the primary Teflon collection filter. In this study, the data have been corrected for dissociation of NH4NO3 from the primary filter by quantification of HNO3 on the backup nylon filter. On average, this correction accounts for approximately 60% of measured NO 3 at all sites. The backup filter was not analyzed for HCl, which is a dissociation product of NH4Cl, and subsequently no correction has been applied to Cl aerosol concentrations. Evidence suggests that, under conditions suitable for NH4NO3 dissociation from the primary Teflon filter, dissociation of NH4Cl may also occur (Matsumoto and Okita, 1998). Assuming that Cl is primarily in the coarse aerosol fraction, this should not impose a significant bias on the NH+ 4 concentrations reported in this study. The dry deposition velocity of NH3 is relatively large compared to NH+ 4 , thus the spatial influence of NH3 emissions depends on the partitioning of NH3 between the gas and aerosol phases. As shown in Fig. 4, the
Fig. 3. Seasonal speciated inorganic PM2.5 across sites. W=winter, Sp=spring, Su=summer, F=fall.
fraction of NH3 in the gas phase follows a seasonal trend across sites. At the Clinton site, NH3 is primarily in the gas phase during all seasons, indicating a system in which NHx is strongly influenced by local NH3 sources and dry deposition of NH3 is likely a significant component of NHx deposition. NH+ 4 present in this system is contributed by long-range transport and local formation. At the Kinston site, the fraction of NH3 in the gas phase exceeds 0.5 during spring and summer. This represents a more dynamic NHx system when viewed within the context of seasonal variability. During the spring and summer, when Kinston is downwind of sources to the southwest (Fig. 1) and higher temperatures increase local agricultural NH3 emissions, the NHx system is similar to Clinton, though also includes a contribution from non-local NH3 sources. During colder months when meteorology acts to reduce local NH3 emissions and the influence of more distant sources to the southwest, NHx shifts toward a predominance of NH+ 4 . In this case, the NHx system is strongly influenced by transport of NH+ 4 and NHx deposition is likely dominated by dry and wet deposition of NH+ 4 . The majority of NH3 exists as NH+ 4 aerosol during all seasons at the Morehead City site, indicating an NHx system influenced by transport of NH+ 4 and one in which the relative contributions of NH3 and NH+ 4 to NHx deposition may be approximately equal on an annual basis. Another important characteristic of NHx is the degree of SO2 4 aerosol neutralization by NH3. This neutralization influences the overall acidity of fine particles, as well as the radiative properties of SO2 4 aerosol. Boucher and Anderson (1995) show that, at 80% relative humidity, the climate forcing efficiency of (NH4)2SO4 (125 W g1 1 SO2 SO2 4 ) is greater than NH4HSO4 (104 W g 4 ). In 2 the eastern US, molar ratios of NH+ to SO 4 4 typically reach a maximum near 2 during winter when acidic gas
Table 5 Molar ratios of aerosol components calculated from 12-h average concentrations Site
Ratioa
N
Mean
SD
Clinton
+ NO 3 /NH4 Cl/NH+ 4 2b NH+ 4 /SO4
562 562 562
0.33 0.02 1.46
0.19 0.05 0.46
Kinston
+ NO 3 /NH4 Cl/NH+ 4 2b NH+ 4 /SO4
264 264 264
0.33 0.03 1.23
0.19 0.04 0.41
Morehead City
+ NO 3 /NH4 Cl/NH+ 4 2b NH+ 4 /SO4
189 189 189
0.30 0.17 0.86
0.19 0.16 0.47
a b
Ratios of molar (nmol m3) concentrations. + Molar concentrations of Cl and NO 3 have been subtracted from the NH4 concentration.
ARTICLE IN PRESS J.T. Walker et al. / Atmospheric Environment 38 (2004) 1235–1246
Fig. 4. Seasonal ratio of NH3 to NH3+NH+ 4 across sites. Bar extensions represent the upper 95% confidence limit for the mean. W=winter, Sp=spring, Su=summer, F=fall.
2 Fig. 5. Ratio of NH+ 4 to SO4 across sites. The 2:1 reference line represents complete SO2 4 neutralization.
species are less abundant, indicating complete neutralization of H2SO4 and a predominance of (NH4)2SO4 aerosol (Sickles, 1999). In this study, the average molar ratio of NH+ 4 (minus molar equivalents of NO3 and 2 Cl ) to SO4 varies from 1.46 at the Clinton site to 0.86 at the Morehead City site. Since a ratio of 2 indicates that all of the aerosol is present as (NH4)2SO4, this signifies differing degrees of aerosol neutralization across sites, as illustrated by the differing slopes of the + lines relating molar concentrations of SO2 4 and NH4 in Fig. 5. At all sites, the ratio reaches a minimum during warm months and a maximum during colder months. At the Clinton site, the value approaches 2 during the winter, suggesting that SO2 4 is completely neutralized. Interestingly, this ratio is o2 at Clinton on average, though NH3 concentrations exceed NH+ 4 concentrations
1243
Fig. 6. Seasonal excess as: Excess NH3 ¼ NH3 defined ½NH3 ðgÞþ NHþ 2 SO2 4 ðaÞ NO3 ðaÞ ½HNO3 ðgÞ 4 ðaÞ ½HClðgÞ Cl ðaÞ across sites. Bar extensions represent the upper 95% confidence limit for the mean. W=winter, Sp=spring, Su=summer, and F=fall.
during all seasons. This may suggest that the NH+ 4 / SO2 4 system is not in equilibrium at the Clinton site or that the ratio, as determined in this study, is biased low by an overestimate of NO 3 and Cl , or an under+ 2 estimate of NH4 . In general, SO4 aerosol is more completely neutralized at the Clinton site, which has the highest concentrations of NH3, relative to Kinston and Morehead City. The variability of SO2 neutralization across sites 4 implies that the relative importance of gas precursors (NH3, H2SO4, HNO3, and HCl) to inorganic aerosol formation may also differ across sites. Using the approach of Ansari and Pandis (1998) it is possible to examine the budget of NH+ aerosol species and 4 precursors in order to determine the sensitivity of aerosol formation to acid gas and NH3 variability. In this case, excess NH3, which represents the quantity of NH3 that would remain after complete neutralization of available aerosol and acid gases, is defined as: Excess NH3 ¼ ½NH3 ðgÞ þ NHþ 4 ðaÞ 2 2 SO4 ðaÞ NO 3 ðaÞ ½HNO3 ðgÞ ½HClðgÞ Cl ðaÞ : ð1Þ Fig. 6 shows seasonal excess NH3 across sites. At the Clinton site, excess NH3 exists during all seasons, indicating that inorganic PM2.5 formation is more sensitive to variability in HNO3, HCl, and acidic SO2 4 than NH3. At the Morehead City site, variability in any of the aerosol precursors may influence inorganic PM2.5
ARTICLE IN PRESS 1244
J.T. Walker et al. / Atmospheric Environment 38 (2004) 1235–1246
levels. The Kinston site is similar to Clinton during warmer months when NH3 concentrations are highest and similar to Morehead City during the fall and winter. Though this analysis greatly simplifies the complex relationship between inorganic PM2.5 and its precursors, it does illustrate potential sitewise differences in the sensitivity of aerosol concentrations to SO2 and NOx vs. NH3 emission reduction strategies. 4. Conclusions This study shows that agricultural NH3 emissions influence local concentrations of NH3 and PM2.5. At the three sites investigated in this study, ambient levels of NH3 and inorganic PM2.5 exhibit a positive correlation with local NH3 emission density. At the coastal Morehead City site, annual concentrations of NH3, NH+ 4 , and inorganic PM2.5 are 90%, 51%, and 40% lower, respectively, than at the agricultural Clinton site. As we have shown, elevated levels of NH3 have an important influence on the basic characteristics of inorganic PM2.5. The fractionation of NH3 between the gas and aerosol phases is drastically different across sites and across seasons within sites, with NH3 primarily in the gas phase during all seasons where local NH3 emissions are high. This is evidence that the characteristics of atmospheric nitrogen deposition may also change drastically with season and across a relatively small spatial scale in response to agricultural NH3 emissions. The paucity of NH3 dry deposition data in the US makes this hypothesis difficult to confirm. Our results show that NH+ 4 aerosol concentrations are highest at the site with highest local NH3 emissions. While NH3 concentrations may influence the inorganic component of PM2.5, additional work is needed to determine if agricultural NH3 emissions may contribute to exceedances of PM2.5 standards in some locations. With respect to PM2.5 control strategies, our data indicates that areas with high NH3 emissions may be more sensitive to SO2 and NOx, rather than NH3, emission reductions. These findings illustrate the connection between agriculture and PM2.5 formation and the potential importance of agricultural NH3 emissions to particulate formation in rural areas. Acknowledgements We appreciate the support of the North Carolina Division of Air Quality, North Carolina Pork Council, National Pork Producers Council, and North Carolina Water Resources Research Institute. Also, we sincerely appreciate the field support of Lynette Mathis (North Carolina State University), Mark Barnes (North Carolina State University), and Brad Hendrickson (UNCCH IMS) throughout the course of the project.
References Allegrini, I., De Santis, F., Di Palo, V., Febo, A., Perrino, C., Possanzini, M., Liberti, A., 1987. Annular denuder method for sampling reactive gases and aerosols in the atmosphere. The Science of the Total Environment 67, 1–16. Aneja, V.P., Chauhan, J.P., Walker, J.T., 2000. Characterization of atmospheric ammonia emissions from swine waste storage and treatment lagoons. Journal of Geophysical Research 105, 11535–11545. Aneja, V.P., Murthy, A.B., Battye, W., Battye, R., Benjey, W.G., 1998. Analysis of ammonia and aerosol concentrations and deposition near the free troposphere at Mt. Mitchell, NC, USA. Atmospheric Environment 32, 353–358. Ansari, A.S., Pandis, S., 1998. Response of inorganic PM to precursor concentrations. Environmental Science and Technology 32, 2706–2714. Asman, W.A.H., 1992. Ammonia emission in Europe: updated emission and emission variations, Report No. 228471008, National Institute of Public Health and Environmental Protection (RIVM), Bilthoven, The Netherlands. Asman, W.A.H., van Jaarsveld, J.A., 1992. A variableresolution transport model applied for NHx for Europe. Atmospheric Environment 26A, 445–464. Asman, W.A.H., Sutton, M.A., Schjorring, J.K., 1998. Ammonia: emission, atmospheric transport, and deposition. New Phytologist 139, 27–48. Battye, R., Battye, W., Overcash, C., Fudge, S., 1994. Development and selection of ammonia emission factors. EPA-600/R-94-190 (NTIS PB95-123915), US EPA, Atmospheric Research and Exposure Assessment Laboratory, Research Triangle Park, NC. Boucher, O., Anderson, T.L., 1995. General circulation model assessment of the sensitivity of direct climate forcing by anthropogenic sulfate aerosols to aerosol size and chemistry. Journal of Geophysical Research 100, 26117–26134. Bouwman, A.F., Lee, D.S., Asman, W.A.H., Dentener, F.J., van der Hoek, K.W., Olivier, J.G.J., 1997. A global highresolution emission inventory for ammonia. Global Biogeochemical Cycles 11, 561–587. Buijsman, E., Aben, J.M.M., van Elzakker, B.G., Mennen, M.G., 1998. An automatic atmospheric ammonia network in The Netherlands: setup and results. Atmospheric Environment 32, 317–324. CASTNet, 2002. Clean air status and trends network. Data accessed 3/27/02. http://www.epa.gov/castnet/data.html. Ferm, M., 1979. Method for determination of atmospheric ammonia. Atmospheric Environment 13, 1385–1393. Galloway, J.N., Cowling, E.B., 2002. Reactive nitrogen and the world: 200 years of change. Ambio 31, 64–71. Harper, L.A., Sharpe, R.R., Parkin, T.B., 2000. Gaseous nitrogen emissions from anaerobic swine lagoons: ammonia, nitrous oxide, and dinitrogen gas. Journal of Environmental Quality 29, 1356–1465. Langford, A.O., Fehsenfeld, F.C., Zachariassen, J., Schimel, D.S., 1992. Gaseous ammonia fluxes and background concentrations in terrestrial ecosystems of the United States. Global Biogeochemical Cycles 6, 459–483. Leaderer, B.P., Naeher, L., Jankun, T., Balenger, K., Holford, T.R., Toth, C., Sullivan, J., Wolfson, J.M., Koutrakis, P.,
ARTICLE IN PRESS J.T. Walker et al. / Atmospheric Environment 38 (2004) 1235–1246 1999. Indoor, outdoor, and regional summer and winter + + concentrations of PM10, PM2.5, SO2 4 , H , NH4 , NO3 , and nitrous acid in homes with and without kerosene space heaters. Environmental Health Perspectives 107, 223–231. Lefer, B.L., Talbot, R.W., Munger, J.W., 1999. Nitric acid and ammonia at a rural northeastern US site. Journal of Geophysical Research 104, 1645–1661. Malm, W.C., Sisler, J.F., Huffman, D., Eldred, R.A., Cahill, T.S., 1994. Spatial and seasonal trends in particle concentration and optical extinction in the United States. Journal of Geophysical Research 99, 1347–1370. Matsumoto, M., Okita, T., 1998. Long-term measurements of atmospheric gaseous and aerosol species using an annular denuder system in Nara, Japan. Atmospheric Environment 32, 1419–1425. McCulloch, R.B., Few, G.S., Murray Jr., G.C., Aneja, V.P., 1998. Analysis of ammonia, ammonium aerosols and acid gases in the atmosphere at a commercial hog farm in eastern North Carolina, USA. Environmental Pollution 102, 263–268. McCurdy, T., Zelenka, M.P., Lawrence, P.M., Houston, R.M., Burton, R., 1999. Acid aerosols in the Pittsburgh metropolitan area. Atmospheric Environment 33, 5133–5145. NCDACS, 2002. Agricultural Statistics Division, North Carolina Department of Agriculture and Consumer Services, P.O. Box 27767, Raleigh, NC. http://www.ncagr.com/stats/ index.htm. Ni, J., 1999. Mechanistic models of ammonia release from liquid manure: a review. Journal of Agricultural Engineering Research 72, 1–17. Ni, J., Heber, A.J., Dielh, C.D., Lim, T.T., 2000. Ammonia, hydrogen sulphide and carbon dioxide release from pig manure in under-floor deep pits. Journal of Agricultural Engineering Research 77, 53–66. Nihlgard, B., 1985. The ammonium hypothesis—an additional explanation to the forest dieback in Europe. Ambio 14, 2–8. Padgett, P., Allen, E., Bytnerowicz, A., Minisch, R., 1999. Changes in soil inorganic nitrogen as related to atmospheric nitrogenous pollutants in southern California. Atmospheric Environment 33, 769–781. Paerl, H.W., 1995. Coastal eutrophication in relation to atmospheric nitrogen deposition: current perspectives. Ophelia 41, 237–259. Paerl, H.W., Whitall, D.R., 1999. Anthropogenically derived atmospheric nitrogen deposition, marine eutrophication and harmful algal bloom expansion: is there a link? Ambio 28, 307–311. Pryor, S.C., Barthelmie, R.J., Sorenson, L.L., Jensen, B., 2001. Ammonia concentrations and fluxes over a forest in the midwestern USA. Atmospheric Environment 35, 5645–5656. Rattray, G., Sievering, H., 2001. Dry deposition of ammonia, nitric acid, ammonium, and nitrate to alpine tundra at Niwot Ridge, Colorado. Atmospheric Environment 35, 1105–1109. Robarge, W.P., Walker, J.T., McCulloch, R.B., Murray, G., 2002. Atmospheric concentrations of ammonia and ammonium at an agricultural site in the southeast United States. Atmospheric Environment 36, 1661–1674. Roelle, P.A., Aneja, V.P., 2002. Characterization of ammonia emissions from soils in the upper coastal plain, North Carolina. Atmospheric Environment 36, 1087–1097.
1245
Roelofs, J.G.M., Kempers, A.J., Houdijk, A.L.F.M., Jansen, J., 1985. The effect of airborne ammonium sulphate on Pinus nigra var. maritima in the The Netherlands. Plant and Soil 84, 45–56. SAS Institute Inc., 1999. SAS Version 8.0 System Help. SAS Institute Inc., Cary, NC. SCONC, 2002. State Climate Office of North Carolina, College of Physical and Mathematical Sciences, North Carolina State University, Raleigh, NC 27695. Data provided 3/27/ 02. http://www.nc-climate.ncsu.edu/. Sickles II, J.E., 1999. A summary of airborne concentrations of sulfur- and nitrogen-containing pollutants in the northeastern United States. Journal of the Air and Waste Management Association 49, 882–893. Sickles II, J.E., Hodson, L.L., Vorburger, L.M., 1999. Evaluation of filter pack for long-duration sampling in ambient air. Atmospheric Environment 33, 2187–2202. Sickles II, J.E., Hodson, L.L., McClenny, W.A., Paur, R.J., Ellestad, T.G., Mulik, J.D., Anlauf, K.G., Wiebe, H.A., Mackay, G.I., Schiff, H.I., Bubacz, D.K., 1990. Field comparison of methods for the measurement of gaseous and particulate contributors to acidic dry deposition. Atmospheric Environment 24A, 155–165. Sievering, H., 1999. Nitrogen deposition and carbon sequestration. Nature 400, 629–630. Sievering, H., Fernandez, I., Lee, J., Hom, J., Rustad, L., 2000. Forest canopy uptake of atmospheric nitrogen deposition at eastern US conifer sites: carbon storage implications? Global Biogeochemical Cycles 14, 1153–1160. Strader, R., Anderson, N., Davidson, C., 2001. User Guide— CMU NH3 Inventory Version 2.0. Carnegie Mellon University, Pittsburgh, PA. http://www.envinst.cmu.edu/ nh3/. Tarnay, L., Gertler, A.W., Blank, R.R., Taylor Jr., G.E., 2001. Preliminary measurements of summer nitric acid and ammonia concentrations in the Lake Tahoe Basin air-shed: Implications for dry deposition of atmospheric nitrogen. Environmental Pollution 113, 145–153. Tolocka, M.P., Solomon, P.A., Mitchell, W., Norris, G.A., Gemmill, D.B., Wiener, R.W., Vanderpool, R.W., Homolya, J.B., Rice, J., 2001. East versus West in the US: chemical characteristics of PM2.5 during the winter of 1999. Aerosol Science and Technology 34, 88–96. US EPA, 2002. Clean Air Status and Trends Network (CASTNet) 2000 Annual Report, Office of Air and Radiation, Research Triangle Park, NC. http://www.epa. gov/castnet/library/annual00.html. US EPA, 1999. National Emission Trends (NET) Database, Office of Air Quality Planning and Standards. http:// www.epa.gov/air/data/index.html. US EPA, 1997. Determination of reactive acidic and basic gases and strong acidity of fine particles (o2.5 m). Compendium of Methods for the Determination of Inorganic Compounds in Ambient Air, Method IO-42. EPA-625/R-96/010a, Center for Environmental Research Information, Office of Research and Development, Cincinnati, OH. US EPA, 1996. Air quality criteria for particulate matter. EPA/ 600/P-95/001aF, National Center for Environmental Assessment, Office of Research and Development, Research Triangle Park, NC.
ARTICLE IN PRESS 1246
J.T. Walker et al. / Atmospheric Environment 38 (2004) 1235–1246
Vitousek, P.M., Howarth, R.W., 1991. Nitrogen limitation on land and in the sea: how can it occur? Biogeochemistry 13, 87–115. Walker, J.T., Nelson, D., Aneja, V.P., 2000a. Trends in ammonium concentration in precipitation and atmospheric ammonia emissions at a Coastal Plain site in North Carolina, USA. Environmental Science and Technology 34, 3527–3534.
Walker, J.T., Aneja, V.P., Dickey, D., 2000b. Atmospheric transport and wet deposition of ammonium in North Carolina, USA. Atmospheric Environment 34, 3407–3418. Whitall, D.R., Paerl, H.W., 2001. Spatiotemporal variability of wet atmospheric nitrogen deposition to the Neuse River Estuary, NC. Journal of Environmental Quality 30, 1508–1515.
