Coastal New England pilot study to determine fossil and biogenic formaldehyde source contributions using radiocarbon



[1] Compound specific radiocarbon analyses of atmospheric formaldehyde are reported as fraction modern (Fm) for a limited number of winter and summer air samples collected in coastal southern New England in 2007. The 11 of 13 samples with Fm < 0.2 were collected under the influence of the semipermanent Bermuda high-pressure system with transport from the Washington, D. C., to New York City urban corridor. The two samples with Fm > 0.2 (max ∼ 0.35) were collected on days with strong northwesterly flow and the least urban impact. The Fm data were combined with VOC observations from the Rhode Island Department of Environmental Management, estimates of oxygenated VOC (OVOC), and back trajectories to interpret the relative contributions of biogenic and fossil carbon sources. It is argued that CH2O sources were dominated by pollutant VOCs and OVOCs from upwind coastal cities as opposed to more local biogenic VOCs at the times of sample collection.

1. Introduction

[2] Formaldehyde (CH2O) is an intermediate product in the photochemical oxidation of volatile organic compounds (VOCs). Its abundance and chemistry relate not only to VOC oxidation, but also to odd-hydrogen chemistry (e.g., HO, HO2, HOx), ozone (O3) production, and global carbon monoxide budgets [e.g., Finlayson-Pitts and Pitts, 2000]. By extension of its impact on O3 and HOx, CH2O affects the processing of nearly all compounds in the troposphere; sulfur dioxide to sulfuric acid, nitric oxide to nitric acid, VOCs to carbon monoxide and on to carbon dioxide. Surface O3 precursors are a combination of natural and anthropogenic nitrogen oxides (NOx) and VOCs and one of many O3 control strategies is focused on reducing anthropogenic reactive VOC (RVOC) emissions, thereby, in effect reducing anthropogenic CH2O as a radical producing source of ozone. The efficacy of this strategy depends upon the relative mix of biogenic and anthropogenic RVOC emissions [e.g., Trainer et al., 1987; Carter, 1994; Russell et al., 1995; Derwent et al., 2007; Luecken and Mebust, 2008].

[3] Biogenic and anthropogenic VOC emission inventories are under continued development and assessment. For example, CH2O column measurements from space-based sensors (GOME, SCIAMACHY, OMI) have been used to constrain bottom-up biogenic isoprene emission estimates [e.g., Palmer et al., 2006; Fu et al., 2007; Millet et al., 2008; Dufour et al., 2009; Stavrakou et al., 2009]. Top-down correction factors to bottom-up inventories can range over a factor of 2 and vary spatially throughout the season. The accuracy of these findings depends on the chemical transport model (CTM) assumptions used to define the relationship between VOC emissions and CH2O column amounts. Depending on the chemical mechanism employed and the simulated NOx concentration, first-day CH2O yields can range from 0.2 to 0.6 moles of CH2O per atom-C of isoprene [Dufour et al., 2009; Palmer et al., 2006]. Warneke et al. [2007], Buzcu-Guven and Fraser [2008] and Mendoza-Dominguez and Russell [2001] provide examples of similar CTM measurement constraints on anthropogenic VOC emissions for the New York, Houston and Atlanta metropolitan centers, respectively, with adjustment factors again on the order of 2 to 3.

[4] Alternatively, the relative contribution of fossil and biogenic sources of RVOCs to CH2O can be experimentally evaluated by measuring the radiocarbon (14C) content of ambient CH2O [Tanner et al., 1996; Shen et al., 2009]. In this paper, the radiocarbon content is expressed as fraction modern, Fm, as defined by McNichol and Aluwihare [2007, and references therein]. Fraction modern is the 14C/12C ratio in a sample relative to the ratio in a “modern” standard where the modern standard represents a theoretical wood sample with a fixed 13C/12C ratio (δ13C = −25‰) in 1950. Fraction modern is calculated from two NIST standards (NIST SRM 4990B and NIST SRM 4990C). Normalizing to a constant δ13C is necessary to remove the effects of isotopic fractionation on the 14C/12C ratio and setting 1950 as the reference point for modern removes the impact of nuclear bomb testing and nuclear power production on age determinations. Weapons testing increased the atmospheric inventory of radiocarbon and resulted in fraction modern values of up to 2.0 in atmospheric carbon dioxide in the 1960s. This signal is mixing into the oceans and biosphere and, since 1985, the value has decreased from 1.2 to its present-day value of 1.05. The radiocarbon content of atmospheric methane has also been modified by weapons testing and fossil fuel burning but to a different extent than CO2; the present-day fraction modern value of methane (FmCH4) is 1.3. Fraction modern should not be confused with biogenic fraction [Klouda et al., 1996] where measured Fm values are normalized to the fraction modern of contemporary atmospheric carbon dioxide under the assumption that the source of carbon for biogenic matter is the contemporary atmosphere. We make this correction using the present-day value of 1.05 but caution that the complexity of carbon sources for formaldehyde makes the appropriate choice difficult. However, given the error of our measurements and the closeness of the Fm value in contemporary atmosphere to 1.0, the correction is relatively small.

[5] Radiocarbon analysis has been used in numerous studies to determine the fraction of fossil and modern carbon sources of trace gases and atmospheric particles. The radiocarbon results for VOCs and carbonyl compounds are of specific interest. The Fm values cited below have been rounded to the nearest tenth percentile. Klouda et al. [1996, 2002] found that only 10–20% of VOCs were derived from modern sources in Atlanta, Georgia, during summer, although they state their sample timing might have missed the diel peak in biogenic VOCs, and the VOCs were 0–40% modern carbon in September 1997 at Azusa, California. Larsen et al. [2001, and references therein] showed carbonyl compounds contained about 30–85% modern carbon at semiremote sites in Italy and Portugal, and about 60–90% modern carbon at a rural area in Germany during autumn. Kato et al. [2008] found acetaldehyde inside an unoccupied home located in a suburb of Tokyo was 50–70% modern carbon. Tanner et al. [1996] reported 80–90% modern carbon in CH2O at Chebogue Point, Nova Scotia, in summer 1993.

[6] Ambient 14CH2O is expected to vary widely, as well, and especially in Rhode Island (RI) given the rural-to-urban character of the State and the air masses transported to it. Warneke et al. [2004] and de Gouw et al. [2005] illustrated the extent of urbanization in New England and its impact on VOC and oxygenated VOC (OVOC) sources and chemistry. They also showed RI as a “hot spot” of biogenic isoprene emission. Further, several major metropolitan areas lie 24 h upwind of southern New England including New York City (∼19 million people, ∼ 220 km away), Baltimore-Washington–Northern Virginia (∼8 million, 550 km), Philadelphia–Southern New Jersey (∼5 million, ∼350 km), Montreal (∼4 million people, 475 km), and Boston (∼4 million, 100 km) and transported air masses from these urban centers to RI are expected to be depleted in biogenic RVOCs [e.g., Warneke et al., 2004; Sommariva et al., 2008], to have large fossil VOC mixing ratios and thus to have small biogenic contributions to CH2O. de Gouw et al. [2005] further showed a suite of OVOCs in the Boston-NYC plume to have estimated biogenic contributions less than 30%.

[7] The results of a pilot study exploring ambient 14CH2O at a southern New England coastal site are reported and interpreted here. The objectives of the study were the development of a compound specific radiocarbon analysis method (CSRA) for CH2O and to answer the questions: What are the Fm values in ambient CH2O impacting the site? Are CH2O sources dominated by biological emissions or fossil pollutants? Does the source mix vary seasonally with a winter maximum in fossil contributions and a summer maximum in biogenic contributions?

2. Methods

2.1. Site Description

[8] In the pilot study, a limited number of CH2O samples were collected on the roof of the CACS building (∼15 m above ground) at the University of Rhode Island's Narragansett Bay Campus (NBC) and their 14C content measured. The locations of the University of Rhode Island's Narragansett Bay (NBC) and Alton Jones (AJC) campuses are shown on the map in Figure 1. AJC is located in the town of West Greenwich, RI, and hosts a Rhode Island Department of Environmental Management (RIDEM) air-monitoring site. The AJC site is referred to as West Greenwich in the manuscript. A second RIDEM sample air-monitoring site at the Francis School in East Providence, RI (EP) is also indicated.

Figure 1.

Map showing the locations of the Narragansett Bay Campus (NBC) sample site in Narragansett, RI, and RIDEM air monitoring sites in East Providence and West Greenwich (image covers 69°45′W–74°15′W and 40°15′N–43°40′N). East Providence (EP) is an urban site often downwind of the Providence metropolitan area. The Alton Jones Campus (AJC) is a semirural site in West Greenwich. The red line indicates the coastal climate zone specified for Connecticut, Massachusetts, and Maine and interpolated through New Hampshire and Rhode Island. It roughly corresponds with the inland extent of the coastal sea breeze. The inset details the NBC, CACS sample location, and surrounding vegetation and housing density (CACS is located at 71°25′24″W, 41°29′26″N). The CH2O sampling site is about 30 km southeast of the AJC site, about 40 km south of EP and in the interpolated coastal zone. Base images from Google Earth 2008© showing state boundaries and major highways.

[9] The NBC 14CH2O hi-vol sampler was located on the southwest corner of the CACS building roof, approximately 15 m off the ground and 40 m MSL. CACS is approximately 250 m from Narragansett Bay to the east and 65 m from a woodlot to the west. The general area outside of the NBC is a mix of high-density suburban homes (0.4 km south) and small businesses (0.3–1 km southwest to west). Boston Neck Rd (rte 1A), a main secondary road, runs north and south ∼1 km to the west. The inset on Figure 1 shows housing, vegetation and roadways near NBC.

[10] Sample dates and times are listed in Table 1. Sample days were selected to be nonprecipitating and mostly sunny to avoid CH2O wet deposition, to maximize the likelihood of high CH2O mixing ratios, and, in summer, to correspond with meteorological conditions conducive to surface ozone production. Average daytime temperature was ∼5°C for the winter samples and ∼23°C for the summer (July) samples.

Table 1. Ambient Air CH2O Faction Modern at Narragansett
Sample IdentificationDate 2007Start (EST)Sample Time (hours)Air Volume (m3)Thiaza (μgC)AMS Carbonb (μgC)Fm CH2OcBiogenic Fraction CH2Od
  • a

    Total amount of thiazolidine collected, in μgC, estimated before preparative capillary gas chromatography.

  • b

    Total amount of carbon converted to graphite.

  • c

    Fraction modern (Fm) carbon ± propagated errors.

  • d

    Biogenic carbon fraction based on Fmbiol = 1.05.

OS-604526 Feb10008.96177.442.50.11 ± 0.010.10
OS-6045026 Mar070012.05570.648.50.16 ± 0.010.15
OS-6045427 Mar070012.34966.244.20.10 ± 0.010.10
OS-6045529 Mar070012.65258.830.40.35 ± 0.050.33
OS-6045631 Mar070011.95445.222.40.13 ± 0.020.12
OS-623411 Jul070011.946101.638.90.28 ± 0.020.27
OS-620393 Jul070011.947127.251.6<0.05<0.05
OS-624477 Jul070013.25087.845.50.10 ± 0.010.10
OS-623438 Jul070012.24785.035.00.16 ± 0.010.15
OS-6203710 Jul080012.145116.559.4<0.05<0.05
OS-6244812 Jul080012.14599.543.70.12 ± 0.010.11
OS-6244517 Jul080012.14498.546.40.09 ± 0.010.09
OS-6203824 Jul100012.044114.256.30.11 ± 0.010.10

2.2. CH2O Instrumental Method

[11] The 14CH2O was determined using the CSRA method described by Shen et al. [2009] and reported as fraction modern, Fm. Briefly, ambient CH2O was collected on sodium bisulfite coated glass-fiber filters using a high-volume air sampler. The exposed NaHSO3-coated filter was rolled within a piece of precombusted aluminum foil, folded, and stored in a capped glass bottle at 4°C until analyzed. The exposed filter was extracted with water and CH2O was derivatized to thiazolidine using cysteamine. The thiazolidine was extracted into dichloromethane and separated from other organic compounds present in the extract by an automated preparative capillary gas chromatography system and collected as a phosphate salt. The thiazolidine salt carbon was combusted to CO2 and subsequently converted to graphite using established procedures for AMS analysis at National Ocean Sciences Accelerator Mass Spectrometry Facility, Woods Hole Oceanographic Institution [McNichol et al., 1992; Pearson et al., 1998]. The Fm limit of detection was <0.05 [Shen et al., 2009]. Radiocarbon results are reported as Fm as defined by McNichol and Aluwihare [2007].

2.3. RIDEM VOC and Ozone Measurements

[12] VOC composition and CH2O source contributions were investigated using RIDEM VOC measurements made at two sites, West Greenwich and East Providence, representing rural and urban RI, on days surrounding the CH2O sampling dates. Site locations relative to NBC are shown in Figure 1. Monthly mean VOC mixing ratios were used for this analysis since the RIDEM data are from 24 h integrated samples collected at 6 day intervals and their sample days did not correspond with CH2O sample days. The two VOC mixes are detailed in Table 2 for July and shown in Figure 2 for February, March and July. VOCs attributed to anthropogenic-fossil emissions were in greater abundance than biogenic VOCs in all three months. There was little or no isoprene in the air during February and March and it increased in July as expected.

Figure 2.

Mean monthly VOC mixing ratios, nmol/mol, for February, March, and July 2007 using RIDEM VOC data from (top) West Greenwich and (middle) East Providence, and (bottom) a summer synthetic urban plume. There are two columns for each month. The left column of a pair shows anthropogenic VOCs, and the right column of a pair shows biogenic isoprene. Only July (summer) had biogenic mixing ratios large enough to be seen above the axis. Anthropogenic VOCs include 23 selected compounds indicated by a separate color (blues and greens are alkanes; oranges and reds are alkenes).

Table 2. Reaction Rate Coefficients, Initial CH2O Yields, and Selected VOC Mixing Ratiosa
CompoundkOH (×1012, cm3 molec−1 s−1)ϕ (CH2O yield per molec)West Greenwich (pmol/mol)East Providence (pmol/mol)Northeast U.S. Urban Plume (pmol/mol)
Anthropogenic VOCs
2, 2-Dimethylbutane2.340.30228
1, 3-butadiene1511018NA
Biogenic VOC
Methane, ∼70% bio0.006411,800,0001,800,0001,800,000
2-Methyl-1, 3-butadiene1000.71099251250

[13] Hourly ozone mixing ratio data were obtained from RIDEM for Narragansett (200 m north of CACS), West Greenwich, and East Providence RI. Site locations are given in Figure 1.

2.4. CH2O Source Model and Estimated Fm

[14] Estimates of VOC contributions to CH2O production, Pi, are shown in Figure 3. Pi was calculated following the work of Lee et al. [1998] and Sumner et al. [2001],

equation image

using representative VOC concentrations ([VOC]i), their respective HO reaction rate coefficients (ki), their corresponding molar CH2O yield (ϕi), and HO concentration ([HO]). RIDEM VOC measurements and the literature were used below to develop several VOC case study mixes. Table 2 lists the selected VOCs, their reaction parameters, and case study mixing ratios. A 12 h average [HO] of 1.6 × 106 molec cm−3 [Seinfeld and Pandis, 1998] was assumed and while the CH2O production rate from each species is proportional to the assumed [HO], their relative contributions are independent of the chosen [HO]. The production estimates, Pi, were combined to calculate an estimated CH2O Fm,

equation image

where the bis represent the fraction of CH2O from various sources and given by

equation image
equation image
equation image

Since Fmfossil ≡ 0, equation (2) is reduced to

equation image

image ≈ 1.3 and Fmisoprene ≈ 1.05 based upon work by Lassey et al. [2007] and the value of contemporary photosynthetic material [e.g., Hsueh et al., 2007] and experiments showing biogenic isoprene is produced from contemporary photosynthetic material [e.g., Sharkey and Yeh, 2001]. In 2007, if all atmospheric formaldehyde was derived without fractionation from contemporary photosynthetic material (e.g., isoprene), atmospheric methane, or fossil carbon (e.g., coal or petroleum), the image would have been 1.05, 1.3 or 0.0, respectively.

Figure 3.

Initial CH2O production rates, nmol/mol h−1, for summer VOC case studies. Each case shows three columns: anthropogenic VOCs (blue to red), isoprene (black), and methane (gray). Anthropogenic VOC contributions are from 23 selected compounds indicated by the color legend (blues and greens are alkanes; oranges and reds are alkenes). VOC mixes are developed and discussed within the text.

2.5. Air Mass Trajectories

[15] Air mass histories (transport) were examined at the regional-to-national scale using meteorological back trajectories calculated using HYSPLIT (R. R. Draxler and G. D. Rolph, HYSPLIT—HYbrid Single-Particle Lagrangian Integrated Trajectory Model, 2003, accessed via NOAA ARL READY Web site, The back trajectory calculations used the EDAS meteorological data set with 3 h 40 km resolution. A set of seven 48 h back trajectories was run for each CH2O sample. The first back trajectory in a set was initiated at a sample's end time and the others initiated at 2 h intervals back in time. Each back trajectory was geographically started at 71.4231°W, 41.4901°N and 300 m altitude. Another set of trajectories were begun 10 km south of this location and a third set ∼10 km west; these were done as a catastrophe check on the meteorological data set and resulting trajectory calculation. All three sets showed the same flow paths.

3. Results and Discussion

3.1. The 14CH2O Observed

[16] The CH2O Fm, biogenic fraction and ancillary data are listed in Table 1. The late winter Fm values ranged from 0.10 ± 0.01 to 0.35 ± 0.05 and those for summer varied from <0.05 to 0.28 ± 0.02. Four of five winter samples contained less than 20% modern carbon. Seven of eight summer samples also contained less than 20% modern carbon. Only 2 of 13 samples contained more than 20% modern carbon: 28% and 35% on 1 July and 29 March, respectively. CH2O derived solely from contemporary biogenic materials should have a Fm > 1.00 because of the current enhancement in atmospheric 14C due to nuclear bomb testing and nuclear power generation relative to the prenuclear standard reference material used. CH2O from purely fossil carbon origin should contain no detectable 14C and have Fm = 0.00. The airflow on the two days with the highest Fm, 29 March and 1 July, was from the northwest. The other days were characterized by airflow from the southeast to west. The results indicated CH2O precursors were dominated by fossil VOC sources on all sample days during winter and summer at NBC. Further, an increased proportion of biogenically derived CH2O was not observed in summer compared to winter, contrary to expectation. The initial expectation is reexamined and alternative explanations are explored below.

3.2. The 14CH2O Source Attribution: Local VOCs

[17] Estimated initial CH2O production rates, nmol/mol h−1, for West Greenwich and East Providence are shown in Figure 3 for July following Lee et al. [1998] and Sumner et al. [2001]. Three columns are shown for each VOC case studied. The left, middle and right columns for a case show initial CH2O production rates from anthropogenic VOCs, isoprene, and methane, respectively. The contribution from α-pinene was calculated, was much smaller than isoprene and is not shown. The contribution from methane was constant at 0.066 nmol/mol h−1. Currently, 70% of atmospheric methane is estimated to be from biological sources [Lassey et al., 2007] and its contribution to CH2O is distributed proportionately.

[18] The calculations (not shown) indicated fossil VOCs were the major sources of CH2O during the winter season at the RIDEM sites, as observed. The biogenic methane and isoprene fraction was predicted to contribute about 10% to CH2O production in February, about 30% in March and this was mostly due to methane. In July, the biogenic isoprene and methane fractions were estimated to contribute about 80% of the CH2O production at West Greenwich and 60% at East Providence (Figure 3). Hence, a large summer Fm value or a large biogenic contribution to CH2O was predicted but not observed.

3.3. The 14CH2O Source Attribution: Transport and Regional Urban VOCs

[19] The combined effects of transport and reactivity suggest the VOC concentrations measured at the inland RIDEM sites were not representative of the VOC mix at the NBC sampling site. Our sample day selection criteria preferentially chose days when prevailing southwesterly winds were predicted with likely transport over water. A regional sea breeze reinforces this flow in daylight hours in summer [e.g., Fisher, 1960; Angevine et al., 2004]. In contrast with the NBC site the inland RIDEM sites are beyond the landward penetration of the sea breeze front. The pattern of hourly surface ozone shown in Figure 4 illustrates this contrast. The inland sites, especially West Greenwich, show depletion ascribed to titration through surface deposition and reactions with NO and reactive VOC emissions under a nocturnal inversion. Angevine et al. [2004] and Darby et al. [2007] similarly attributed variations in O3 in coastal New Hampshire to transport effects.

Figure 4.

Hourly average O3 for Narragansett, West Greenwich, and East Providence, RI. Month/day and Sundays are indicated along the x axis and hour of day (EST) is indicated along the y axis. Colors denote O3 mixing ratios ranging from light pink (<5 nmol/mol) to dark red (>120 nmol/mol). The green-to-yellow transition indicates 75 nmol/mol. Missing data are indicated by white. The arrows along the top of Figure 4 (bottom) show 14CH2O sample days at the Narragansett Bay Campus. The double arrows along the top of Figure 4 (middle) show RIDEM VOC sample days. Bold red boxes show periods over which the 8 h average ozone standard of 75 nmol/mol was exceeded at each site.

[20] We examined summer transport patterns using back trajectories. Figure 5 illustrates the spatial domain enclosed by the 7 trajectories for each sample in July, truncated at 12 h. The photochemical lifetime of CH2O is <3 h in daylight at 41°N in July. Consequently, if the CH2O concentration is constant, >95% of the CH2O sampled would have been produced in the previous 12 h; or >95% of the CH2O present 12 h upstream will have been lost by the time of sampling. The trajectory envelopes indicate sampling of coastal air masses, including the New York City urban plume, with the exception of one sample. The 1 July 2007 sample has the least urban influence of all of the July samples, as the transport path lies well west of the Boston metropolitan area, and Montreal is 15–18 h upwind. The 1 July sample has the second-highest Fm value. The February and March trajectories showed similar behavior but with longer paths owing to stronger winds. The 29 March case has a trajectory envelope nearly identical with the 1 July case; this sample has the highest Fm value.

Figure 5.

Air mass envelopes defined by 12 h back trajectories initiated every 2 h from sample start to end times. Dates 1, 3, 7, 8, 10, 12, 17, and 24 July are indicated by the colors green, cyan, blue, magenta, red, white, orange, and yellow, respectively. With the exception of 1 July, each envelope included a significant coastal air component and the New York City metropolitan area, highlighted by the black ellipse. Sample site location is shown by the orange star labeled CACS. Image covers approximately 69°W–76°W, 38°N–44°N. Base image from Google Earth 2008©.

[21] Two additional VOC mixes were developed to better reflect transport and upwind VOC sources, an urban plume case and a coastal inflow case. The northeast urban plume case was developed following Sommariva et al. [2008] with one exception: isoprene was significantly reduced from their value of 1740 pmol/mol to 250 pmol/mol. Isoprene concentrations at West Greenwich in July 2007 were less than the values used by Sommariva et al. [2008] for their urban plume case study. Further, whole-air-sample VOC measurements collected on the NOAA WP-3 in ICARTT-2004 near New York City and within the near-field NYC urban plume contained less than 250 pmol/mol isoprene when two or more of ethane, propane, ethene or propene were in excess of 2000, 1000, 500 and 100 pmol/mol, respectively (E. Atlas, University of Miami, personal communication, 2009). The later mixing ratios were in keeping with the Sommariva urban plume case and East Providence urban data. Hence, isoprene was set at 250 pmol/mol for the urban plume case mix whose VOC mixing ratios are listed in Table 2.

[22] A southern New England coastal onshore flow VOC case study was defined using the West Greenwich measurements and the whole-air-sample measurements from ICARTT-2004 over Rhode Island Sound, Block Island Sound, Long Island Sound and the coastal North Atlantic south of Long Island, NY (E. Atlas, University of Miami, personal communication, 2009). The “Coastal” over-water VOC mixing ratios were near those listed in Table 2 for the West Greenwich rural site with the exception of isoprene and α-pinene, which were both < 1 pmol/mol. VOC and OVOC plume studies over New England support using reduced isoprene values in the over-water VOC case. Warneke et al. [2004], Goldan et al. [2004], de Gouw et al. [2005] and Sommariva et al. [2008] have argued biogenic VOCs (isoprene and monoterpenes) are oxidized during both day and nighttime hours by HO and nitrogen trioxide (NO3), respectively, whereas, the anthropogenic VOCs are oxidized mostly during daylight hours by HO. The oxidation of VOCs by HO is expected to lead to the formation of CH2O and the oxidation of isoprene and monoterpenes by NO3 is expected to yield organic nitrates [e.g., Tuazon et al., 1999; Atkinson and Arey, 2003] and so is not expected to yield significant amounts of CH2O. Rhode Island is approximately 10 h to 1 day transport time from New York City (200 km) or Washington, D. C. (500 km), respectively, and urban polluted air masses transported over water are expected to have their biogenic VOCs depleted relative to the longer-lived anthropogenic VOCs. The consequences of day/night oxidation, biogenic/anthropogenic VOC emission patterns and transport path over water/land should have a marked impact on O3 production [Warneke et al., 2004] and CH2O sources. Hence, the defined “Coastal” inflow VOC case has the anthropogenic VOC mixing ratios of the West Greenwich case listed in Table 2 but without isoprene or α-pinene.

[23] The CH2O source contribution results for the additional urban plume and the urban plume over water cases are shown on Figure 3. Both cases indicated fossil VOCs contribute ∼60% of the CH2O. While these two cases developed Fms indicating larger fossil contributions than the original East Providence or West Greenwich cases, their biogenic VOC contributions were still higher by about 0.2 to 0.3 in Fm than those observed and additional fossil sources of CH2O were sought.

3.4. The 14CH2O Source Attribution: Regional Urban VOCs and OVOCs

[24] The inclusion of oxygenated volatile organic compounds (OVOCs), organic nitrates and organic peroxy nitrates as precursors to CH2O formation could further help bridge the gap between that expected from the CH2O source analyses and the summer Fm measurements. Shen [2008] analyzed near surface CH2O sources using the same point-model method and data as presented by Fried et al. [2008] for high-altitude INTEX-NA data and found the OVOC group (methanol, acetaldehyde, acetone), together with peroxy compounds (PAN and peracetic acid), contributed 10–40% of the total CH2O production at altitudes <1 km over eastern North America. An analysis of VOCs and OVOCs in urban plumes over the Gulf of Maine by de Gouw et al. [2005] showed acetaldehyde, propanal, acetone, methyl ethyl ketone, methanol, peroxy propyl nitrate, and peroxy isopropyl nitrate to each have biogenic source contributions below 20%. They also showed formic and acetic acids and peroxy acetyl nitrate to each have biogenic source contributions of 20–30% and only the isoprene derived product, peroxy methacryloyl nitrate, had a biogenic source contribution (86%) larger than >30%. From these two studies, the urban OVOC contribution to CH2O is presumed to have a predominantly fossil source character, which would increase the fossil source component in the urban plume cases and would lower the expected Fm for urban air by another 0.1 to 0.3 from the concomitant reductions in bisoprene and image

[25] The combination of metropolitan plume VOCs and OVOCs with over water transport brought the calculated fossil contribution into plausible agreement with the Fm observed at NBC. However, this analysis is by no means complete. Other OVOC components not considered such as methyl vinyl ketone or methacrolyn, products of isoprene oxidation and potential secondary biogenic sources of CH2O away from the primary emission location [e.g., de Gouw et al., 2005; Millet et al., 2006], would increase the disagreement. Unaccounted for near-site local CH2O and HRVOC emissions, (e.g., propane tanks, diesel traffic, light industry) would decrease the disagreement. The data to quantitatively evaluate these factors are unavailable.

3.5. The 14CH2O Source Attribution: Prior Work

[26] In the only other 14CH2O study, Tanner et al. [1996] reported ∼87% modern carbon for samples collected at Chebogue Point, Nova Scotia, in the summer during NARE-1993. They concluded that the majority of CH2O measured was derived from biogenic precursors. Their interpretation was based upon air samples collected at 10 m above the surface for 24 to 48 h. Analyses of meteorological conditions and air mass composition from both ground and airborne measurements indicated surface air at their site was frequently isolated from polluted air aloft as the relatively warm polluted air from the northeast United States and Canada was advected over relatively cold North Atlantic waters upwind of Chebogue Point. This led to a strong temperature inversion isolating the pollutants from the surface site [e.g., Tanner et al., 1996]. The CH2O concentrations observed (0.3–1.5 nmol/mol) by Tanner et al. were in the same range as airborne observations of CH2O made over this region below 500 m altitude during INTEX-NA 2004 [Shen, 2008; Fried et al., 2008]. From the analysis by Shen [2008], 60 to 70% of the CH2O in maritime air near Nova Scotia was attributed to the oxidation of methane and methyl hydroperoxide (mainly derived from methane). The relatively high 14C content of CH4 (Fm = 1.3 [Lassey et al., 2007]), together with the source analysis by Shen [2008] indicated a mix of 60–70% methane and 30–40% anthropogenic VOC would yield CH2O with an Fm of 0.8 to 0.9 and consistent with that observed by Tanner et al. [1996]. It should be noted that Millet et al. [2006] suggested long-range transport of biogenic OVOCs as the source of the biogenic CH2O at Chebogue Point and these sources would be in addition to that from methane. The Tanner et al. study and the one presented here emphasize the need to carefully consider transport, as well as, local chemistry in the interpretation of CH2O radiocarbon observations.

4. Conclusions and Future Work

[27] Gaseous CH2O samples were collected at Narragansett, RI during winter and summer 2007 and their radiocarbon content determined. The results show a significantly large fraction of fossil carbon in collected CH2O, 11 of 13 samples had Fm < 0.2, and indicated the precursors of CH2O were dominated by fossil sources at the NBC site during each sample's collection period. The winter-spring Fm values were in agreement with those expected based upon a CH2O source contribution analysis. The summer Fm values were considerably lower than first expected from this same analysis using VOC observations from rural and urban Rhode Island locations. It is suggested that pollutants from upwind coastal metropolitan areas were transported to the sample site, a local sea breeze circulation reinforced this pollutant transport flow regime and that urban fossil VOCs and OVOCs could have dominated the CH2O source chemistry at NBC. Additional CH2O radiocarbon measurements and more sophisticated chemical transport modeling of atmospheric CH2O in a variety of environments, e.g., urban, semiurban, and rural areas, are needed to confirm this interpretation of CH2O sources and to further develop strategies to regulate its precursor VOCs and OVOCs and potential contributions to surface O3.


fraction of CH2O production from a specific source.


fraction modern.


highly reactive volatile organic compound.


chemical species or process index.


chemical reaction rate coefficient.


oxygenated volatile organic compound.


production rate from species i.


volatile organic compound.


reactive volatile organic compound.


effective CH2O yield from VOC reaction with HO.

[ ]

molecular concentration.


[28] This research was supported by a graduate student internship program at WHOI National Ocean Sciences Accelerator Mass Spectrometry Facility (NSF OCE-9807266) and by NASA project NNG04GB38G. The Rhode Island Department of Environmental Management is thanked for providing their VOC and O3 measurements. The authors gratefully acknowledge the NOAA Air Resources Laboratory (ARL) for the provision of the HYSPLIT transport and dispersion model and/or READY Web site ( used in this publication. Elliot Atlas, University of Miami, is thanked for the use of his whole-air-sample VOC data from ICARTT-2004, and we thank Joost de Gouw, NOAA ESSL, for discussion on urban plume chemistry over the northeastern United States.