Journal of Geophysical Research: Atmospheres

Estimates of Cl atom concentrations and hydrocarbon kinetic reactivity in surface air at Appledore Island, Maine (USA), during International Consortium for Atmospheric Research on Transport and Transformation/Chemistry of Halogens at the Isles of Shoals

Authors


Abstract

[1] Average hydroxyl radical (OH) to chlorine atom (Cl·) ratios ranging from 45 to 119 were determined from variability-lifetime relationships for selected nonmethane hydrocarbons (NMHC) in surface air from six different transport sectors arriving at Appledore Island, Maine, during July 2004. Multiplying these ratios by an assumed average OH concentration of 2.5 × 106 cm−3 yielded estimates of Cl· concentrations of 2.2 to 5.6 × 104 cm−3. Summed reaction rates of methane and more than 30 abundant NMHCs with OH and Cl· suggest that Cl· reactions increased the kinetic reactivity of hydrocarbons by 16% to 30% over that due to OH alone in air associated with the various transport sectors. Isoprene and other abundant biogenic alkenes were the most important hydrocarbon contributors after methane to overall kinetic reactivity.

1. Introduction

[2] Tropospheric ozone (O3) plays a central role in regulating Earth's environment. Photolysis of O3 in the presence of water vapor produces the OH radical, which is the principal oxidant for many important atmospheric compounds including methane (CH4), other hydrocarbons, carbon monoxide (CO), nitrogen oxides (NOX = NO + NO2), and hydrochlorofluorocarbons. At the Earth's surface, high concentrations of O3 can be toxic to humans and vegetation and it is one of the principal components of smog. In the middle and upper troposphere, O3 is a major greenhouse gas. Until the 1970s it was thought that tropospheric O3 was mainly supplied by transport from the stratosphere and removed by deposition involving reactions with organic materials at Earth's surface. Research since then has shown that tropospheric O3 is in fact strongly affected by chemical production and loss within the troposphere.

[3] Chlorine radical chemistry influences O3 in two ways [e.g., Pszenny et al., 1993]. Some Cl· in marine air reacts directly with O3 to initiate a catalytic destruction sequence:

equation image
equation image
equation image

[4] However, most Cl· reacts with hydrocarbons via hydrogen abstraction to form hydrogen chloride (HCl) vapor. The enhanced supply of odd-hydrogen radicals from hydrocarbon oxidation leads to O3 production in the presence of sufficient NOX. Thus chlorine radical chemistry represents a modest net sink for O3 when the NOX mixing ratio is less than about 20 pmol mol−1 and a net source of O3 when NOX mixing ratios are greater. Recent evidence from the Texas Air Quality Study indicates that chlorine radical chemistry in polluted coastal/urban air leads to significant net O3 production [e.g., Tanaka et al., 2000, 2003; Riemer et al., 2002; Chang et al., 2002].

[5] A major objective of the Chemistry of Halogens at the Isles of Shoals (CHAiOS) component of the International Consortium for Atmospheric Research on Transport and Transformation (ICARTT) studies was to quantify the influences of halogen radicals on O3 production/destruction in polluted air along the New England east coast during summer [Fehsenfeld et al., 2006]. During the ICARTT/CHAiOS intensive sampling period in July and August 2004, a comprehensive suite of chemical and physical characteristics of near-surface air was quantified at an island site a few kilometers offshore from Portsmouth, New Hampshire, USA. In this paper, we estimate average [OH]/[Cl·] ratios from hydrocarbon variability-lifetime relationships and then evaluate the potential for chlorine radical chemistry to affect NMHC kinetic reactivity [Carter, 1991; Seinfeld and Pandis, 1998] due to OH attack. A discussion of the full implications of chlorine chemistry on the local/regional O3 budget will be published separately once simulations of multiphase chemistry with a comprehensive model to be conducted by R. von Glasow (U. East Anglia) are completed.

2. Methods

2.1. Analytical Methods

[6] Hourly samples were collected in 2-liter electropolished stainless steel canisters (purchased from D.R. Blake, University of California, Irvine) from 2 to 29 July 2004 through an inlet at 43 m ASL on the roof of a WWII-era surveillance tower on Appledore Island, Maine (42°59′13.5″N, 70°23′55.4″W, ∼0.39 km2), located approximately 10 km off the shore of New Hampshire (Figure 1). Canister samples were pressurized to 35 psig using a single-head metal bellows pump (MB-302MOD, Senior Flexonics, Sharon, Massachusetts) and returned to the University of New Hampshire every four days for determination of C2–C10 NMHCs, C1–C2 halocarbons, C1–C5 alkyl nitrates and selected organic sulfur compounds by gas chromatography using a three-GC system equipped with two electron capture detectors (ECDs), two flame ionization detectors (FIDs) and one mass spectrometer (MS). The samples were analyzed by cryotrapping 1772 cm−3 STP (273 K, 1 atm) of air on a glass bead filled loop immersed in liquid nitrogen. After each sample was trapped, the loop was isolated, warmed to 80°C and the sample was injected. Helium carrier gas flushed the contents of the loop and the stream was split into five substreams, each feeding a separate GC column. One 30 m × 0.53 mm I.D., 10 μm film thickness CP-A12O3/Na2SO4 PLOT column (Varian, Inc., Walnut Creek, California), one 60 m × 0.25 mm I.D., 1 μm film thickness OV-1701 column (Ohio Valley, Marietta, Ohio), one 60 m × 0.32 mm I.D., 1.0 μm film thickness DB-1 column (J&W Scientific, Folsom, California), and two 60 m × 0.25 mm I.D., 1.4 μm film thickness OV-624 columns (Ohio Valley, Marietta, Ohio) were used for the trace gas separation. One of the OV-624 columns and the OV-1701 were plumbed into ECDs and used for measuring the halocarbons and alkyl nitrates (not discussed in this paper). The PLOT and DB-1 columns were connected to FIDs and used for the C2–C10 NMHC quantification. The second OV-624 column provided separation for the MS. Electron impact mode for the MS was used for sample ionization along with single ion monitoring. This system provided duplicate measurements for numerous halocarbons and NMHCs. Finally, the gas separation was unique for each of the columns and thus any gases coeluting on one column were usually resolved on another. For the standard analysis protocol, a 1772 cm−3 STP aliquot from one of two working standards was assayed every ninth analysis, thereby quickly drawing attention to any drift or malfunction of the analytical system. The measurement precision for individual halocarbons, hydrocarbons, and alkyl nitrates ranged from 0.1 to 12% depending on the compound and mixing ratio. Further details of the NMHC sampling and analysis are given by Sive et al. [2005], Zhou et al. [2005], and Y. Zhou et al. (Bromoform and dibromomethane measurements in the seacoast region of New Hampshire, 2002–2004, submitted to Journal of Geophysical Research, 2006).

Figure 1.

Map showing (a) the location of the Isles of Shoals on the New Hampshire–Maine, USA border, approximately 10 km offshore, and (b) location of the sampling site on Appledore Island.

2.2. Transport Sector Classification and Local Meteorology

[7] Samples were classified by source region (Figure 2) using back trajectories and the corresponding column and footprint residence time components of retroplumes [Fischer et al., 2006]. Column residence time and footprint residence plots provided by the NOAA Chemical Sciences Division (http://esrl.noaa.gov/csd/ICARTT/analysis/) and HYSPLIT trajectories archived at Plymouth State University (http://pscwx.plymouth.edu/ICARTT/archive.html) were visually inspected to create six regional flow regime classifications. To generate the column residence time and footprint residence plots, the particle dispersion model FLEXPART [Stohl et al., 1998, 2005] was run in backward mode [Stohl et al., 2003; Seibert and Frank, 2004] in order to see where the sampled air masses were potentially exposed to emissions. FLEXPART was driven with ECMWF analyses of 0.36° resolution and accounts for turbulence and deep convection, in addition to the transport by grid-resolved winds. Every 3 hours, 40000 particles were released from the location of the measurement site and followed backward in time for 20 days to calculate a so-called potential emission sensitivity (PES) function, as described by Seibert and Frank [2004] and Stohl et al. [2003]. For the emission distribution, the inventory of Frost et al. [2006] with a base resolution of 4 km, remapped to the 0.25° × 0.333° resolution of the PES output, was used [Frost et al., 2006].

Figure 2.

Idealized depiction of the six transport sectors into which sampled air was classified. Details of the classification scheme are given by Fischer et al. [2006].

[8] The HYSPLIT trajectories [Draxler and Rolph, 2005] were initialized from 10, 500, and 1000 meters above ground level every 6 hours using the North American Mesoscale (NAM) 12 km analysis. Over-water transit times from the coast to Appledore Island were estimated using 36-hour back trajectories calculated from the NOAA ESRL profiler network and ocean buoy data (http://www.etl.noaa.gov/data/profiler_data/processed/ProfilerTrajectory/). Local meteorological data were obtained from National Oceanic and Atmospheric Administration (NOAA) meteorological station IOSN3 on White Island, approximately 2.5 km south of Appledore Island.

2.3. Computational Methods

[9] The basis for the [OH]/[Cl·] ratio estimates is the coherent variability-lifetime trends exhibited by compounds with similar atmospheric source-sink distributions. Both measurements [e.g., Ehhalt et al., 1998; Jobson et al., 1998, 1999, 2004; Millet et al., 2004] and modeling studies [e.g., Hamrud, 1983; Ehhalt et al., 1998; Lenschow and Gurarrie, 2002] indicate that appropriately selected sets of compounds follow power law relationships

equation image

where SlnX is the standard deviation of the natural logarithms of a time series of measured mixing ratios, τ is the local lifetime of the compound (defined below), and A and b are fit parameters. Physical meaning can be attached to parameter A if it is assumed that the measurements represent air parcels with the same initial mixing ratios but varying transit times. Equation (1) can then be rewritten as

equation image

where Ψ is related to some measure of the distribution of transit times from sources [Jobson et al., 1999]. Smaller values of Ψ imply shorter source-receptor transit times. Parameter b ranges between 0 and 1 and can be interpreted as a measure of source-receptor distances [Jobson et al., 1998, 1999]. Smaller values tend to result from observations made close to sources, where variability is driven by source strength variations, while larger values tend to result from observations far from sources and after chemical losses have occurred.

[10] In polluted air, motor vehicles are the dominant source of many C3 to C8 aliphatic and C6 to C9 aromatic hydrocarbons. The dominant sink for these compounds is usually oxidation initiated by OH. If it is assumed that reaction with OH is the only sink then the local lifetime of a compound is

equation image

where kOH is the compound's OH reaction rate coefficient. Assuming reaction with Cl· to be an additional sink alters (3) to

equation image

[11] Varying [Cl·] relative to [OH] in (4) to minimize the residuals of a fit to (1) yields a “best” estimate of [OH]/[Cl·]. The optimizations were done in two steps. In Excel (Version 11.2.5 for Mac) an approximate maximum correlation of log-transformed values of SlnX and τ was obtained by varying [Cl] manually from an inconsequential value (500 cm−3) to a value higher than can reasonably be expected in polluted MBL air (106 cm−3). For all six transport subsets this yielded a relation with a single maximum in the correlation coefficient for [Cl] between 2 and 6 × 104 cm−3. Excel's “solver” tool was then used to automatically home in on the [Cl] value that maximized the correlation coefficient for each subset. The OH concentration was held fixed at 2.5 × 106 cm−3 on the basis of Warneke et al. [2004] who estimated average [OH] in surface air encountered by NOAA Ship Ronald H. Brown within an approximately 100 km × 100 km area surrounding Appledore Island from mid-July to early August 2002 using a “full” source-sink calculation and, separately, the parameterization of Ehhalt and Rohrer [2000]. Substituting kCl and [Cl] for their ozone analogs in equation (26) of Ehhalt et al. [1998] gives:

equation image

where k′ is a fictitious rate constant determined from the fit of SlnX versus τ. Rearranging (5) yields

equation image

which shows that calculated [Cl] scales directly with assumed [OH]. Jobson et al. [1999] described and used an analogous fitting method to estimate relative radical abundances during polar sunrise at Alert, NWT, Canada.

3. Results and Discussion

3.1. Variability-Lifetime Relationship Calculations

[12] A subset of eight NMHCs (ethyne, propane, i-butane, n-butane, i-pentane, n-pentane, benzene and toluene) was selected with which to construct variability-lifetime relationships and estimate [OH]/[Cl·] as described above. The selection criteria were (1) similar source distributions (specifically, dominated by motor vehicle emissions), (2) small percentage (<3%) of values below detection limit to keep bias in variability small, (3) lifetimes against reaction with 2.5 × 106 OH cm−3 between ∼0.5 and ∼4 days (at 298 K), (4) lifetimes against reaction with a preliminary estimate of ∼4 × 104 Cl· cm−3 [Goldan et al., 2005; Keene et al., 2007] of ∼0.5 day or longer, and (5) negligible reaction rates with oxidants other than OH and Cl·. Criteria 3 and 4 reduce potential confounding effects of compounds with highly variable sources and sinks (e.g., alkenes) and of long-lived compounds whose variability may be affected substantially by multiple sources (e.g., ethane). Rate coefficients used are given in Table 1. Mixing ratios of these eight NMHCs and those of O3 and CO for sample subsets associated with each of the six transport sectors are summarized in Figure 3.

Figure 3.

Box-whisker plot summaries by transport sector of the measured mixing ratios of CO, O3 and the eight NMHCs used in constructing variability-lifetime relationships from which [OH]/[Cl·] ratios were estimated. The central line across each box indicates the median; the top and bottom of each box indicate the 75th and 25th percentile values, respectively; and the upper and lower whisker ends indicate the 90th and 10th percentile values, respectively.

Table 1. Rate Coefficients at 298 K Used for Reactions of CO and Hydrocarbons With OH and Cl Atoms
CompoundkOH, cm−3 molecule−1 s−1ReferencekCl, cm−3 molecule−1 s−1Reference
Carbon monoxide1.50 × 10−13DeMore et al. [1997]3.15 × 10−14DeMore et al. [1997]
Methane6.40 × 10−15Atkinson et al. [2005]1.00 × 10−13Atkinson et al. [2005]
Ethane2.40 × 10−13Atkinson et al. [2005]5.90 × 10−11Atkinson et al. [2005]
Ethene9.00 × 10−12Atkinson et al. [2005]1.10 × 10−10Atkinson et al. [2005]
Ethyne9.00 × 10−13Atkinson [1990]5.20 × 10−11Atkinson et al. [2005]
Propane1.10 × 10−12Atkinson et al. [2005]1.40 × 10−10Atkinson et al. [2005]
Propene3.00 × 10−11Atkinson et al. [2005]2.70 × 10−10Atkinson et al. [2005]
i-butane2.12 × 10−12Atkinson [2003]1.43 × 10−10Atkinson [1997]
n-butane2.36 × 10−12Atkinson [2003]2.18 × 10−10Atkinson [1997]
t-2-butene6.40 × 10−11Atkinson [1997]3.31 × 10−10Ezell et al. [2002]
1-butene3.14 × 10−11Atkinson [1997]3.38 × 10−10Ezell et al. [2002]
c-2-butene5.64 × 10−11Atkinson [1997]3.76 × 10−10Ezell et al. [2002]
i-pentane3.60 × 10−12Atkinson [2003]2.20 × 10−10Atkinson [1997]
n-pentane3.80 × 10−12Atkinson [2003]2.80 × 10−10Atkinson [1997]
2-methyl-2-butene8.69 × 10−11Atkinson [1997]3.95 × 10−10Ezell et al. [2002]
1-pentene3.14 × 10−11Atkinson [1997]3.97 × 10−10Ezell et al. [2002]
Cyclohexane6.97 × 10−12Atkinson [2003]3.50 × 10−10Atkinson [1997]
2-methylpentane5.20 × 10−12Atkinson [2003]2.90 × 10−10Atkinson [1997]
3-methylpentane5.20 × 10−12Atkinson [2003]2.80 × 10−10Atkinson [1997]
n-hexane5.20 × 10−12Atkinson [2003]3.40 × 10−10Atkinson [1997]
n-heptane6.76 × 10−12Atkinson [2003]3.97 × 10−10Ezell et al. [2002]
n-octane8.11 × 10−12Atkinson [2003]4.60 × 10−10Atkinson [1997]
n-nonane9.70 × 10−12Atkinson [2003]4.80 × 10−10Atkinson [1997]
n-decane1.10 × 10−11Atkinson [2003]5.50 × 10−10Atkinson [1997]
Isoprene1.00 × 10−10Atkinson et al. [2005]5.10 × 10−10Finlayson-Pitts et al. [1999]
2,4-dimethylpentane4.80 × 10−12Atkinson [2003]2.90 × 10−10Atkinson [1997]
Methylcyclohexane9.60 × 10−12Atkinson [2003]3.90 × 10−10Atkinson [1997]
2,2,4-trimethylpentane3.34 × 10−12Atkinson [2003]2.60 × 10−10Atkinson [1997]
Benzene1.23 × 10−12Atkinson [1990]4.00 × 10−12Wallington et al. [1988]
Toluene5.96 × 10−12Atkinson [1990]5.90 × 10−11Shi and Bernhard [1997]
1,3,5-trimethylbenzene5.67 × 10−11Atkinson [1990]2.42 × 10−10Wang et al. [2005]
a-pinene5.30 × 10−11Atkinson et al. [2005]5.30 × 10−10Timerghazin and Ariya [2001]
b-pinene7.90 × 10−11Atkinson [1997]5.30 × 10−10Finlayson-Pitts et al. [1999]
Limonene1.70 × 10−10Atkinson [1997]6.40 × 10−10Finlayson-Pitts et al. [1999]

3.2. Interpretation of Relationships

[13] The results of fits to equation (1) are given in Table 2 and illustrated in Figure 4. Two groups of relationships are clearly evident in Figure 4. One group contains the sample subsets associated with the three “westerly” transport sectors (SW urban, midwest and NW) while the other group contains the subsets associated with “easterly” transport sectors. This grouping is reflected in the values of the transit time variability index (Ψ) derived from fit parameter A (Table 2). Less transit time variability for the “westerly” group is consistent with the shorter upwind distances to shore (10 to 50 km) and shorter transit times from major sources of vehicular emissions than for the “easterly” group. The value of A is sensitive to the assumed value of [OH] and the estimated value of [Cl·] so some caution is necessary in interpreting the absolute values of Ψ.

Figure 4.

Plots of SlnX versus τ for the eight selected NMHC and best fit variability-lifetime relationships (solid lines) for the subsets of canister samples representing the six transport sectors defined by Fischer et al. [2006]. Colors represent sectors as follows: red, SW urban; orange, midwest; yellow, NW; green, north-northeast Canada; blue, marine; and violet, south coastal. Dotted lines illustrate 95% confidence intervals for two of the six fits. Numerical fit results are given in Table 2.

Table 2. Results of Fits to Functions of the Form SlnX = Aτ−b by Transport Sector for [OH] = 2.5 × 106 molecules cm−3
Transport SectorNAΔAabΔbaΨ,b hours[OH]/[Cl] From Fit"Best" [Cl],c molecules cm−3
  • a

    95% confidence interval for value in preceding column.

  • b

    Ψ = A1/b × 24; range (in parentheses) corresponds to extremes calculated using A ± ΔA and b ± Δb.

  • c

    Uncertainties are approximate 95% confidence intervals; see text for explanation.

SW urban820.5440.0480.340.164.1 (0.5–11)534.7 ± 0.12 × 104
Midwest690.5540.0680.200.201.2 (∼0–7.3)1192.1 ± 0.060 × 104
NW1200.5890.0620.230.172.3 (∼0–8.2)455.6 ± 0.15 × 104
N_NE_Canada1680.9510.0950.370.1921 (10–31)643.9 ± 0.13 × 104
Marine780.9180.1340.540.3120 (8.3–30)773.3 ± 0.083 × 104
S Coastal1230.8260.0850.410.2015 (5.8–21)1122.2 ± 0.087 × 104

[14] The values of parameter b vary with transport sector in a different but somewhat consistent manner as that described above for parameter A (Table 2). The three smallest values, suggesting relative proximity to sources of the selected NMHCs, are associated with the three “westerly” transport sectors. The value for the SW urban sector is larger than those for the midwest and NW sectors, possibly because of influence of intense emissions from New York City and other Eastern Seaboard metropolitan areas. Relationships for the two “easterly” sectors that cover the relatively heavily populated New England coastline have b values similar to that for the SW urban sector. The marine sector relationship shows the largest b value, consistent with relatively long source-receptor distances and transit over ocean areas where emissions of the selected NMHCs are much less than emissions over land.

3.3. Estimates of [OH]/[Cl·] Ratios and Cl· Concentrations

[15] Also given in Table 2 are the average [OH]/[Cl·] ratios estimated from the power law fits and the average Cl· concentrations calculated from those ratios assuming [OH] = 2.5 × 106 molecules cm−3. Because [OH] is assumed constant and [OH]/[Cl·] is also constant the only source of uncertainty in calculated [Cl·] derives from uncertainties in the rate constants used for the 16 reactions involved. This uncertainty has a potential systematic component due to the assumption of atmospheric conditions of 298 K, 1 atm and a nonsystematic component due to variability in laboratory rate constant determinations. Using 285 K, 0.8 atm (an approximation for average conditions at 2 km altitude) increases calculated [Cl·] by 6% to 13% depending on transport sector. To estimate uncertainties due to variability in laboratory rate constant determinations, optimizations were repeated 25 times for each transport sector subset, each time with all 16 rate constants varied randomly within the uncertainty ranges given for them in the respective source references. Tabulated uncertainties in [Cl·] are approximate 95% confidence limits based on these 25 optimization runs.

[16] The [Cl·] do not group in a straightforward manner like the fit parameters. Two possible explanations are (1) [Cl·] was high in SW urban sector air because average local surface wind speeds were more than twice those for other sectors and sea salt aerosol concentrations and fluxes were correspondingly greater [Keene et al., 2007] and (2) [OH] was lower than 2.5 × 106 molecules cm−3 in NW sector air, which tended to be cooler and dryer than air from other sectors. Modeling is in progress that may eventually allow resolution of this incongruity.

[17] Published estimates of [Cl·] for various locations and times are presented in Table 3. Because direct measurement of Cl· in ambient air is not yet technologically possible, all existing estimates are based on measurements of hypothesized Cl· precursors or reactant species in combination with a variety of modeling techniques ranging in complexity from straightforward hydrocarbon ratio methods [e.g., Jobson et al., 1994] to detailed multiphase chemistry models [e.g., Sander and Crutzen, 1996]. The [Cl·] concentrations of a few × 104 cm−3 estimated in this study are well within the range of previous estimates, which spans more than four orders of magnitude. No two published studies have employed exactly the same method/model combination, rendering quantitative comparisons questionable, even within the subgroups of studies identified in Table 3.

Table 3. Estimates of Tropospheric [Cl·] From Various Studies
[Cl·], 104 cm−3Location, TimeReference; Notes
Inferred From MBL NMHC Measurements
<1California coast, springParrish et al. [1992]
1–10Alert, NWT, sunriseJobson et al. [1994]
3–7SE N. Atlantic, summerWingenter et al. [1996]
4–76equatorial N. Pacific, boreal autumnSingh et al. [1996a]
<0.7N. Atlantic/Med., springRudolph et al. [1997]
∼0.7Southern Ocean, summerWingenter et al. [1999]
2–6Appledore Island, summerthis study
 
Inferred From Global Budgets of C2Cl4
<0.1–1global MBLRudolph et al. [1996]
<0.5–3global MBL (daytime)Singh et al. [1996b]
 
Inferred From Models
0.1global troposphereSingh and Kasting [1988]; HCl (1 nmol mol−1) + OH
10–100N. Atlantic, summerKeene et al. [1990]; Cl2 from aerosol Cl deficit data
<1–10coastal Florida (Miami), winterPszenny et al. [1993]; HCl and other inorganic Cl gas data
<0.1–1polluted MBLSander and Crutzen [1996]; MOCCA
1–10remote MBLVogt et al. [1996]; MOCCA
0.7/1.8remote/polluted MBLKeene et al. [1998]; MOCCA
∼0.2Oahu, Hawaii, SeptemberPszenny et al. [2004]; MOCCA

[18] We are aware of only one previous estimate of [OH]/[Cl·]. Using a technique analogous to the one employed in the present study in which a fixed concentration of Cl atoms was assumed, Jobson et al. [1999] estimated relative abundances of [OH]:[Cl·]:[Br·] of ∼17:1:3300 in surface air at an ice floe location near Alert NWT during polar sunrise. Previous work at Alert had strongly suggested an important role of Cl and Br chemistry during low-ozone events observed around the time of polar sunrise [Jobson et al., 1994].

3.4. NMHC Kinetic Reactivity With OH and Cl·

[19] Incremental reactivity (IR) is an index of the influence of a volatile organic compound (VOC) on ozone concentration [Carter, 1991; Seinfeld and Pandis, 1998]. The IR of a VOC is defined as the change in ozone concentration for a given change in concentration of the compound:

equation image

[20] IR is the product of kinetic reactivity, which is the production of peroxy radicals from initial attack by OH (or other oxidants):

equation image

and mechanistic reactivity, which includes NO to NO2 conversion and other reactions involved in O3 formation and destruction:

equation image
equation image
equation image
equation image
equation image

and so on.

[21] The kinetic reactivity of compound X by reactions with OH and Cl· is given by

equation image

[22] Table 4a and 4b gives kinetic reactivity calculated from the median mixing ratios of the more than 30 NMHCs measured in the canister samples for each transport sector. Values for CO are from the AIRMAP database (airmap.unh.edu). Methane is assumed present at a constant mixing ratio of 1.8 ppmv.

Table 4a. Percentage Contributions of OH and Cl Reactions to Kinetic Reactivities of Compounds Calculated Using the Median Mixing Ratios of the Respective Compounds in the Subsets of Samples Associated With the SW Urban, Midwest, and NW Transport Sectorsa
CompoundSW UrbanMidwestNW
%OH%ClTotal%OH%ClTotal%OH%ClTotal
  • a

    Unit is molecules cm−3 s−1. Methane mixing ratio was assumed to be 1.8 ppmv and hydroxyl radical concentration was fixed at 2.5 × 106 cm−3 for all subsets. Chlorine atom concentrations were taken as the “best” values listed in Table 2 for the respective subsets.

Carbon monoxide99.6%0.4%1.85E + 0699.8%0.2%1.91E + 0699.5%0.5%1.66E + 06
Methane77.3%22.7%9.17E + 0588.3%11.7%8.03E + 0574.1%25.9%9.58E + 05
Isoprene91.3%8.7%1.52E + 0595.9%4.1%4.79E + 0589.7%10.3%3.64E + 05
Ethene81.3%18.7%1.52E + 0590.6%9.4%2.40E + 0586.9%13.1%2.43E + 05
Propane29.5%70.5%1.26E + 0548.2%51.8%1.76E + 0526.0%74.0%1.89E + 05
Ethane17.8%82.2%1.10E + 0594.6%5.4%1.43E + 0581.7%18.3%1.62E + 05
Propene85.6%14.4%1.09E + 0592.9%7.1%1.41E + 0578.5%21.5%1.61E + 05
i-pentane46.6%53.4%7.83E + 0496.9%3.1%1.02E + 0583.2%16.8%1.20E + 05
Toluene84.3%15.7%6.54E + 0492.3%7.7%9.86E + 0415.4%84.6%9.67E + 04
Limonene93.4%6.6%5.32E + 0432.5%67.5%6.58E + 0492.2%7.8%7.61E + 04
b-pinene88.8%11.2%5.31E + 0456.2%43.8%6.30E + 0481.9%18.1%7.27E + 04
n-butane36.6%63.4%4.52E + 0494.7%5.3%5.65E + 0487.1%12.9%6.11E + 04
Ethyne48.0%52.0%4.14E + 0492.2%7.8%5.41E + 0432.6%67.4%5.60E + 04
n-pentane42.0%58.0%3.99E + 0461.7%38.3%4.62E + 0443.6%56.4%4.28E + 04
α-pinene84.2%15.8%3.62E + 0467.2%32.8%3.83E + 0437.7%62.3%4.26E + 04
2-methylpentane48.9%51.1%2.91E + 0468.0%32.0%3.60E + 0444.5%55.5%3.37E + 04
n-hexane44.9%55.1%2.45E + 0463.7%36.3%3.11E + 0490.8%9.2%2.89E + 04
1,3,5-trimethylbenzene92.6%7.4%2.11E + 0496.5%3.5%3.00E + 0445.3%54.7%2.88E + 04
1-butene83.2%16.8%2.08E + 0464.4%35.6%2.96E + 0440.6%59.4%2.53E + 04
i-butane44.1%55.9%2.06E + 0491.7%8.3%2.85E + 0436.4%63.6%2.34E + 04
3-methylpentane49.7%50.3%2.02E + 0468.8%31.2%2.54E + 0439.8%60.2%2.21E + 04
2-methyl-2-butene92.1%7.9%1.86E + 0470.3%29.7%2.21E + 0480.6%19.4%2.15E + 04
n-heptane47.6%52.4%1.58E + 0490.4%9.6%1.89E + 0489.6%10.4%2.02E + 04
c-2-butene88.9%11.1%1.41E + 0466.9%33.1%1.72E + 0491.3%8.7%1.86E + 04
n-decane51.6%48.4%1.31E + 0496.3%3.7%1.67E + 0447.2%52.8%1.74E + 04
2,2,4-trimethylpentane40.6%59.4%1.27E + 0460.3%39.7%1.60E + 0487.0%13.0%1.58E + 04
1-pentene80.8%19.2%1.26E + 0494.7%5.3%1.32E + 0477.9%22.1%1.54E + 04
t-2-butene91.2%8.8%1.17E + 0495.8%4.2%1.32E + 0443.2%56.8%1.52E + 04
Cyclohexane51.5%48.5%1.07E + 0474.5%25.5%1.28E + 0447.1%52.9%1.08E + 04
n-nonane51.9%48.1%9.73E + 0370.2%29.8%1.21E + 0452.4%47.6%1.04E + 04
Methylcyclohexane56.7%43.3%9.58E + 0370.5%29.5%1.07E + 0447.4%52.6%9.45E + 03
n-octane48.4%51.6%8.25E + 0367.6%32.4%8.94E + 0344.0%56.0%8.16E + 03
Benzene94.2%5.8%7.28E + 0397.3%2.7%8.83E + 0342.5%57.5%6.96E + 03
2,4-dimethylpentane46.9%53.1%4.10E + 0366.2%33.8%6.63E + 0393.2%6.8%6.76E + 03
Sum HC69.5%30.5%2.26E + 0684.4%15.6%2.86E + 0670.7%29.3%2.98E + 06
Sum HC + CO83.0%17.0%4.11E + 0690.6%9.4%4.77E + 0681.0%19.0%4.64E + 06
Table 4b. Same as Table 4a, Except for the N_NE_Can, Marine, and S Coastal Transport Sectors
CompoundN_NE_CanMarineS Coastal
%OH%ClTotal%OH%ClTotal%OH%ClTotal
Carbon monoxide99.7%0.3%1.31E + 0699.7%0.3%1.27E + 0699.8%0.2%1.35E + 06
Methane80.4%19.6%8.83E + 0583.0%17.0%8.54E + 0587.8%12.2%8.08E + 05
Isoprene90.5%9.5%1.31E + 0595.3%4.7%1.34E + 0595.6%4.4%8.37E + 04
Ethene92.6%7.4%1.14E + 0523.7%76.3%5.76E + 0490.2%9.8%6.51E + 04
Propane94.4%5.6%9.53E + 0491.9%8.1%5.24E + 0446.8%53.2%6.03E + 04
Ethane33.4%66.6%7.25E + 0493.7%6.3%4.20E + 0496.8%3.2%5.30E + 04
Propene84.0%16.0%6.99E + 0437.5%62.5%3.28E + 0492.6%7.4%5.11E + 04
i-pentane20.6%79.4%5.68E + 0489.5%10.5%3.18E + 0494.4%5.6%4.36E + 04
Toluene87.7%12.3%5.42E + 0486.2%13.8%3.18E + 0431.3%68.7%4.07E + 04
Limonene51.1%48.9%3.57E + 0494.4%5.6%2.66E + 0491.9%8.1%2.43E + 04
b-pinene90.6%9.4%2.79E + 0492.0%8.0%2.48E + 0464.7%35.3%2.19E + 04
n-butane86.6%13.4%2.55E + 0488.4%11.6%1.81E + 0491.8%8.2%2.10E + 04
Ethyne93.4%6.6%2.52E + 0457.0%43.0%1.55E + 0496.1%3.9%2.00E + 04
n-pentane93.7%6.3%2.18E + 0494.7%5.3%1.44E + 0466.0%34.0%1.81E + 04
α-pinene40.9%59.1%2.17E + 0485.8%14.2%1.23E + 0454.8%45.2%1.48E + 04
2-methylpentane52.5%47.5%1.84E + 0493.7%6.3%1.18E + 0496.3%3.7%1.45E + 04
n-hexane92.5%7.5%1.66E + 0487.7%12.3%9.26E + 0394.4%5.6%1.38E + 04
1,3,5-trimethylbenzene46.5%53.5%1.58E + 0465.3%34.7%8.42E + 0391.2%8.8%1.13E + 04
1-butene90.6%9.4%1.57E + 0445.3%54.7%8.02E + 0395.6%4.4%1.11E + 04
i-butane54.3%45.7%1.37E + 0457.4%42.6%7.57E + 0360.3%39.7%1.01E + 04
3-methylpentane83.5%16.5%1.32E + 0460.5%39.5%7.28E + 0389.9%10.1%9.68E + 03
2-methyl-2-butene53.4%46.6%1.31E + 0455.6%44.4%6.12E + 0366.8%33.2%8.39E + 03
n-heptane85.6%14.4%1.15E + 0488.5%11.5%5.68E + 0367.6%32.4%8.06E + 03
c-2-butene45.1%54.9%9.21E + 0360.7%39.3%5.21E + 0363.2%36.8%7.10E + 03
n-decane49.4%50.6%8.87E + 0357.8%42.2%5.09E + 0362.4%37.6%6.48E + 03
2,2,4-trimethylpentane48.7%51.3%8.84E + 0355.9%44.1%4.39E + 0369.2%30.8%6.27E + 03
1-pentene56.1%43.9%7.97E + 0353.9%46.1%3.80E + 0359.0%41.0%4.90E + 03
t-2-butene52.1%47.9%7.91E + 0350.9%49.1%3.77E + 0365.6%34.4%4.63E + 03
Cyclohexane61.1%38.9%7.30E + 0353.1%46.9%3.74E + 0373.4%26.6%4.47E + 03
n-nonane53.0%47.0%7.16E + 0395.9%4.1%3.59E + 0369.4%30.6%4.13E + 03
Methylcyclohexane56.0%44.0%6.90E + 0358.7%41.3%3.44E + 0397.2%2.8%4.09E + 03
n-octane56.4%43.6%5.46E + 0349.5%50.5%3.40E + 0369.1%30.9%3.29E + 03
Benzene51.4%48.6%4.71E + 0360.4%39.6%3.34E + 0366.4%33.6%3.23E + 03
2,4-dimethylpentane95.2%4.8%3.24E + 0356.6%43.4%3.16E + 0365.0%35.0%1.98E + 03
Sum HC77.0%23.0%1.83E + 0680.7%19.3%1.46E + 0684.2%15.8%1.46E + 06
Sum HC + CO86.4%13.6%3.14E + 0689.6%10.4%2.72E + 0691.7%8.3%2.81E + 06

[23] For aromatics and some alkenes the kinetic reactivity with Cl· is small or negligible compared to that with OH. However, for CH4 and most aliphatic alkanes kinetic reactivity with Cl· rivals or surpasses that with OH. Overall, Cl· attack is estimated to account for 16–30% of hydrocarbon kinetic reactivity with these two oxidants depending on transport sector. The differences are due to the relative amounts of the various NMHCs and the estimated [Cl·] in each sector. Including CO in the sums reduces the contribution of Cl· by a third to a half, but Cl· still accounts for 8–20% of the combined kinetic reactivity.

[24] The NMHCs are ranked in Table 5 according to their contributions to combined kinetic reactivity with OH and Cl·. Six of the top ten in overall rank are alkenes with predominantly biogenic sources, reflecting the importance of natural emissions in fueling photochemical O3 production in air that bathes coastal New England during summer. To evaluate incremental reactivity and the effects of chlorine chemistry on the O3 budget and related photochemistry in the coastal New England atmosphere requires simulations with a comprehensive multiphase photochemical model. Appropriate simulations are planned by R. von Glasow (U. East Anglia) as part of a more general modeling effort of halogen chemistry during CHAiOS. The results of that effort will be published separately.

Table 5. NMHCs Ranked Reactivity With OH and Cl Atoms Combined, by Transport Sector
CompoundRank
SW UrbanMidwestNWN_NE_CanMarineS CoastalAverage
Isoprene1112411.7
Propane3334533.5
Ethene2255723.8
b-pinene9421364.2
Limonene8693145.2
Propene5567655.7
Ethane4976275.8
a-pinene13114810109.3
i-pentane688920910.0
Toluene77101021810.5
Ethyne11131214111212.2
n-butane10101113171312.3
2-methyl-2-butene2023151191114.8
1,3,5-trimethylbenzene16162212121415.3
n-pentane12121316261816.2
2-mepentane14141420232017.5
1-butene17182021151617.8
cis-2-butene2225241781518.5
n-hexane15171723252219.8
t-2-butene26262115141719.8
1-pentene25212519131920.3
3-mepentane19191618292120.3
i-butane18151924272321.0
n-decane23202325192422.3
2,2,4-trimethylpentane24241822302523.8
n-heptane21222626322625.5
Methylcyclohexane29272827162725.7
n-nonane28292930222827.7
n-octane30303028183127.8
Cyclohexane27282729313028.7
2,4-dimethylpentane32323131243230.3
Benzene31313232282930.5

4. Summary

[25] Estimated chlorine atom concentrations derived from NMHC variability-lifetime relationships were used to evaluate the impact of chlorine radical chemistry on hydrocarbon degradation in surface air over coastal New England during summer 2004. Results suggest that Cl· attack increases NMHC kinetic reactivity by 16% to 30% over that due to OH attack in air masses with various transport histories. Isoprene and other abundant biogenic alkenes are the most important contributors to overall NMHC kinetic reactivity.

Acknowledgments

[26] We thank James Morin, former director, and the staff of the Shoals Marine Laboratory for outstanding logistical support during the field campaign. Jesse Ambrose, Karl Haase, Carsten Nielsen, Patrick Veres, and Yong Zhou assisted with sampling. William Keene, Lynn Russell, Jochen Stutz, and Roland von Glasow collaborated in the CHAiOS effort. We thank Paul Goldan and David Parrish of the NOAA Earth System Research Laboratory, Chemical Sciences Division, for sharing their ICARTT results and for discussion that stimulated this analysis. One anonymous reviewer's comments led to marked improvements in the manuscript. Principal financial support was provided by the National Science Foundation through award ATM-0401622. Additional support was provided by the Office of Oceanic and Atmospheric Research at NOAA under grants NA04OAR4600154 and NA05OAR4601080. This is contribution 135 to the Shoals Marine Laboratory.

Ancillary