Content uploaded by Paul S. Monks
Author content
All content in this area was uploaded by Paul S. Monks on Mar 27, 2014
Content may be subject to copyright.
JAtmosChem
DOI 10.1007/s10874-012-9224-7
Modelled and measured concentrations of peroxy
radicals and nitrate radical in the U.S. Gulf Coast region
during TexAQS 2006
Roberto Sommariva ·Tim S. Bates ·Daniel Bon ·Daniel M. Brookes ·
Joost A. de Gouw ·Jessica B. Gilman ·Scott C. Herndon ·William C. Kuster ·
BrianM.Lerner·Paul S. Monks ·Hans D. Osthoff ·Alex E. Parker ·
James M. Roberts ·Sara C. Tucker ·Carsten Warneke ·Eric J. Williams ·
Mark S. Zahniser ·Steven S. Brown
Received: 26 August 2011 / Accepted: 26 March 2012
© Springer Science+Business Media B.V. 2012
Abstract Measurements of total peroxy radicals (HO2+RO2) and nitrate radical
(NO3) were made on the NOAA research vessel R/V Brown along the U.S. Gulf
Coast during the TexAQS 2006 field campaign. The measurements were modelled
using a constrained box-model based upon the Master Chemical Mechanism (MCM).
The agreement between modelled and measured HO2+RO2was typically within
∼40% and, in the unpolluted regions, within 30%. The analysis of the model
results suggests that the MCM might underestimate the concentrations of some
acyl peroxy radicals and other small peroxy radicals. The model underestimated the
measurements of NO3by 60–70%, possibly because of rapid heterogeneous uptake
of N2O5. The MCM model results were used to estimate the composition of the
peroxy radical pool and to quantify the role of DMS, isoprene and alkenes in the
formation of RO2in the different regions. The measurements of HO2+RO2and
NO3were also used to calculate the gas-phase budget of NO3and quantify the
R. Sommariva (B)·D. Bon ·J. A. de Gouw ·J. B. Gilman ·W. C. Kuster ·B. M. Lerner ·
H. D. Osthoff ·J. M. Roberts ·S. C. Tucker ·C. Warneke ·E. J. Williams ·S. S. Brown
Earth System Research Laboratory, NOAA, Boulder, CO, USA
e-mail: r.sommariva@uea.ac.uk
R. Sommariva ·D. Bon ·J. A. de Gouw ·J. B. Gilman ·W. C. Kuster ·B. M. Lerner ·
H. D. Osthoff ·S. C. Tucker ·C. Warneke ·E. J. Williams
CIRES, University of Colorado, Boulder, CO, USA
T. S. Bates
Pacific Marine Environmental Laboratory, NOAA, Seattle, WA, USA
D. M. Brookes ·P. S. Monks ·A. E. Parker
Department of Chemistry, University of Leicester, Leicester, UK
S. C. Herndon ·M. S. Zahniser
Aerodyne Research, Inc., Billerica, MA, USA
JAtmosChem
importance of organic peroxy radicals as NO3sinks. RO2accounted, on average,
for 12–28% of the total gas-phase NO3losses in the unpolluted regions and for 1–2%
of the total gas-phase NO3losses in the polluted regions.
Keywords Peroxy radicals ·RO2·Nitrate radical ·NO3·MCM ·TexAQS 2006
1 Introduction
The concentrations and reactivities of radical species, such as OH, NO3and peroxy
radicals, are central to our understanding of atmospheric chemical processes. Peroxy
radicals are intermediates in the oxidation of Volatile Organic Compounds (VOCs),
which is mostly initiated by OH during the day and, in polluted areas, by NO3during
the night. Ozone (O3) and other radicals, such as halogen atoms (e.g., Cl) also con-
tribute to the oxidation of VOCs and to the formation of peroxy radicals. Except for
HO2, peroxy radicals are organic compounds (RO2), whose number and structures
depend on the concentrations of the precursor VOCs and on the fragmentation
patterns created by their reactions with NO, HO2and other organic peroxy radicals.
Typically, the most common organic peroxy radical is CH3O2, but the composition
of the RO2pool can be very complex and it largely depends on the mixture of VOCs
in an air mass, which in turn depends on the history of the air mass itself.
The importance of peroxy radicals is mainly related to the conversion of NO to
NO2, which drives the photochemical formation of ozone in the troposphere, via
the photolysis of NO2(Monks 2005). Additionally, previous studies (Mihelcic et al.
1993; Canosa-Mas et al. 1996; Carslaw et al. 1997; Salisbury et al. 2001; Geyer et al.
2003; Vaughan et al. 2006; Sommariva et al. 2009) have indicated that organic peroxy
radicals interact with the nitrate radical (NO3) and that, under certain conditions,
Present Address:
R. Sommariva
School of Environmental Sciences, University of East Anglia, Norwich, UK
Present Address:
D. Bon
Department of Civil & Environmental Engineering, Washington State University,
Pullman, WA, USA
Present Address:
D. M. Brookes
Air Quality Practice, AEA plc., Harwell, Didcot, UK
Present Address:
H. D. Osthoff
Department of Chemistry, University of Calgary, Calgary, AB, Canada
Present Address:
A. E. Parker
PC2A, Université des Sciences et Technologies de Lille, Lille, France
Present Address:
S. C. Tucker
Ball Aerospace & Technologies Corp., Boulder, CO, USA
JAtmosChem
the RO2+NO3reactions can be significant sinks for NO3, generate OH at night and
decrease night-time loss of NOxspecies, thus affecting the photochemical formation
of ozone at sunrise. The composition of the peroxy radical pool is important to
understand the relationship of peroxy radicals chemistry to the nitrogen, HOx
(OH+HO2) and ozone budgets. In this work, measurements of total peroxy radicals
(HO2+RO2)andNO
3were analyzed using a highly detailed chemical box-model
based upon the Master Chemical Mechanism (MCM, Saunders et al. 2003, Jenkin
et al. 2003).
The measurements were taken during the Texas Air Quality Study (TexAQS)
2006 field campaign onboard the NOAA research vessel R/V Brown. The TexAQS
2006 cruise took place between July 27th and September 11th 2006, with the objective
to study air quality in the U.S. Gulf Coast region and in the Houston, Texas, area.
The R/V Brown sailed from Charleston, South Carolina, to Houston, Texas, along
the Gulf coast, in the Galveston Bay and in the Houston Ship Channel (Fig. 1). A
variety of air masses were sampled during the cruise, ranging from clean marine air
sampled off the coast of Florida and in the Gulf of Mexico to polluted air sampled
in the industrial areas of the Gulf Coast region. The study area (Fig. 1)andits
characteristics have been described in a previous paper (Sommariva et al. 2011),
which also details the observations of HO2+RO2made during the R/V Brown
cruise, as well as in other related papers (Parrish et al. 2009; Gilman et al. 2009;
Tucker et al. 2010).
The main objective of this paper is to assess the agreement between the model and
the measurements of HO2+RO2and NO3as an indicator of our understanding of
radical chemistry under a variety of conditions. Simultaneous in-situ observations
of HO2+RO2and NO3during the cruise make a detailed investigation of the
Fig. 1 Cruise of the R/V Brown during the Texas Air Quality Study 2006. HSC indicates the Houston
Ship Channel
JAtmosChem
interactions between these radicals possible. The explicit description of the chemistry
in the model allowed us to determine the composition of the peroxy radical pool
(i.e., which individual peroxy radicals contributed the most to the total HO2+RO2),
how this was related with the composition of an air mass and how it affected radical
chemistry in the region. The model construction and assumptions are described in
Section 2; the model results are shown and discussed in Sections 3and 4.
2 Methods
2.1 MCM model
The model was built according to the procedure outlined in Carslaw et al.
(1999,2002), Sommariva et al. (2006,2009) using a chemical mechanism taken
from the Master Chemical Mechanism (MCM) version 3.1 (http://mcm.leeds.ac.uk/
MCMv3.1/). The MCM is a quasi-explicit chemical mechanism for tropospheric
chemistry and contains the detailed degradation processes of 135 VOCs, plus a
complete inorganic chemistry mechanism assembled using the IUPAC Gas Kinetic
Data Evaluation (Atkinson et al. 2006). The MCM mechanism protocol is described
in Jenkin et al. (1997,2003), Saunders et al. (2003).
In this work a subset of the MCM v3.1 containing 65 VOCs plus CH4and CO
was used. Dimethyl sulphide (DMS) is not included in version 3.1 of the MCM,
so the DMS oxidation mechanism used in previous studies (Carslaw et al. 1999,
2002; Sommariva et al. 2006,2009) was added to the MCM model. The model was
integrated using an off-line version of AtChem (https://atchem.leeds.ac.uk/webapp/)
and constrained to the measured or estimated values of CO, CH4,H
2,NO,NO
2,
O3,SO
2,H
2O, 65 VOCs, j(O1D), j(NO2), j(NO3), aerosol surface area, temperature,
pressure, latitude and longitude. Description and details of the measured parameters
and of the instruments can be found in Bates et al. (2008) (aerosol surface area),
Parrish et al. (2009)(NO
x,O
3,SO
2,H
2O, photolysis rates), Gilman et al. (2009),
Warneke et al. (2010) (CO, VOCs).
Methane (CH4) and molecular hydrogen (H2) were not measured on the
R/V Brown during the 2006 cruise. Based on average values measured by the NOAA
Global Monitoring Division (http://www.esrl.noaa.gov/gmd/) stations close to the
area of the TexAQS 2006 cruise a constant value of 520 ppb was used for H2;a
constant value of 1800 ppm (open ocean regions) and 1850 ppm (coastal and polluted
regions) was used for CH4. The Gas Chromatography (GC) method used to measure
VOCs (Gilman et al. 2009) could not resolve all the isomers of xylenes and ethyl-
methyl-benzenes: the ratio between m-xylene and p-xylene and the ratio between
1-ethyl-3-methyl-benzene and 1-ethyl-4-methyl-benzene were assumed to be 1:1.
To test the sensitivity of the model results to these approximations, the model was
run with changed (±10%) methane and molecular hydrogen concentrations. The
impact on the species of interest was limited: on average, OH changed by <2%,
HO2by <1%, CH3O2and CH3CO3by <6%, C2H5CO3by ∼5%, total organic
peroxy radicals (RO2)by<2% and NO3by <3%. The difference in calculated HO2
and CH3O2was slightly higher (∼8–12%) in clean marine air than in polluted air,
where most of the radical reactivity was controlled by species other than CH4,such
as oxygenated VOCs and alkenes (Gilman et al. 2009). The ratios of xylenes and
JAtmosChem
ethyl-methyl-benzenes isomers were also varied, but the radical concentrations did
not change in a significant way (<1%).
The model also included dry deposition terms for the appropriate species (O3,
NO2,SO
2,HNO
3,H
2O2, HCHO, CH3CHO, alkyl nitrates, organic hydroperoxides,
organic acids, PANs) as in previous studies (Carslaw et al. 1999,2002; Sommariva
et al. 2006,2009) and was constrained to the mixing height determined by the NOAA
High Resolution Doppler Lidar (HRDL, Tucker et al. 2010) which was onboard
the R/V Brown. Heterogeneous uptake of 34 gas-phase species was assumed to be
irreversible and calculated using Eq. 1(Fuchs and Sutugin 1970):
khet =A¯
cγ
4(1)
where Ais the aerosol surface area (cm2cm−3), ¯
cis the mean molecular speed of the
gas molecule (cm s−1)andγis the uptake coefficient. The uptake coefficients were
taken from Atkinson et al. (2006), except for γN2O5which was set to 0.02 (Aldener
et al. 2006)andγHO2: the value of γHO2was set to 0.2based on the work by Thornton
et al. (2008), although this is likely an upper limit value (Thornton and Abbatt 2005).
Recent laboratory studies (Taketani et al. 2008,2009) have reported lower values
(0.07–0.19) on sea-salt and sulphate aerosol at 75% relative humidity, but changing
γHO2to 0.1did not have a significant impact on modelled HO2+RO2. The aerosol
surface area was calculated using the aerosol number and size distributions measured
on the R/V Brown at 60% relative humidity and corrected using a humidity growth
factor (Bates et al. 2008).
The model results were compared to the measurements (Sections 3.1,3.2,4.1)
of total peroxy radicals (HO2+RO2) by PEroxy Radical Chemical Amplification
(PERCA, Sommariva et al. 2011) and of NO3by Cavity Ring-Down Spectroscopy
(CaRDS, Dubé et al. 2006). Both instruments were located in a container on the
forward upper deck of the R/V Brown, about 20 meters above sea level. The
HO2+RO2measurements by PERCA had an overall 2-σuncertainty of 40%with
a detection limit of 2 ppt (=5.0×107molecule cm−3) for 1 min integration time
(Sommariva et al. 2011). The NO3measurements by CaRDS had an overall 2-σ
uncertainty of 25% with a detection limit of 2 ppt (=5.0×107molecule cm−3)for
1 s integration time (Osthoff et al. 2006).
Assessing the uncertanties in the MCM calculation is complex, owing to the very
large number of reactions and kinetic parameters involved: Sommariva et al. (2004)
estimated the uncertainty of OH as 30–40% and of HO2as 25–30% under very
clean unpolluted conditions. No estimate was given for CH3O2or other organic
peroxy radicals, nor for NO3. The MCM model uncertainties are likely greater under
polluted conditions due to the larger number of VOCs involved, for many of which
the kinetic data in the MCM were estimated owing to lack of laboratory experiments
(Jenkin et al. 1997).
2.2 Model results
The model was run for the 30 days of the R/V Brown cruise during TexAQS 2006
(July 30th to September 12th, with a 4 days break on August 18th–22nd); the
model results were filtered to exclude the periods when one or more of the model
constraints were missing (e.g., during calibrations, instrument downtimes, power
JAtmosChem
failures, sampling of the ship’s exhaust) and averaged to have the same frequency
of the observations. The results were also filtered to exclude all concentrations lower
than twice the reported detection limits of the instruments.
The model calculated the concentrations of all the non-constrained species;
modelled OH and HO2/(HO2+RO2) ratio were used in Sommariva et al. (2011)
together with measured HO2+RO2,NO
x,O
3and photolysis rates to calculate the
in-situ photochemical formation of ozone during TexAQS 2006. In the following
sections (Sections 3and 4) the focus will be on HO2, organic peroxy radicals (RO2)
and NO3, with the objectives of: (1) assessing how well the model can reproduce the
observations, (2) determining the composition of the peroxy radicals pool and (3)
investigating the interactions between the nitrate radical and the peroxy radicals.
The modelled results were divided into regions, defined by the location of
the R/V Brown (Fig. 1, see also Sommariva et al. 2011), with different chemical
conditions. Air masses sampled when the R/V Brown was in the Atlantic Ocean and
in the Gulf of Mexico were classified as either Open Ocean or Gulf Coast, depending
on whether they had travelled for a long period of time over the ocean or they
were coming from the continent. The distinction was made using the observations
of 222Rn—a marker of continental influence—taken onboard the R/V Brown (Bates
et al. 2008): air masses with low levels of 222Rn (≤500 mBq/m3) were classified as
Open Ocean, while air masses with higher levels of 222Rn were classified as Gulf
Coast (Sommariva et al. 2011). The Galveston Bay region (Fig. 1) was characterized
by recirculation of air masses from the continent (i.e., more processed air) and
heavy traffic of ships and barges. The Industrial Areas included the Houston Ship
Channel (HSC) and the Matagorda, Freeport and Beaumont harbours with their
concentrations of petrochemical and industrial complexes. Barbours Cut is a shipping
dock South of the Houston Ship Channel (Fig. 1): although the chemical composition
of the air masses sampled in Barbours Cut was not very different from those in
Galveston Bay and in Industrial Areas, these data were analyzed separately, as this
was the location where the R/V Brown spent the longest period of time during the
cruise. The observations and the model results were then averaged to obtain diurnal
(i.e., day-time +night-time) profiles in each region.
3 Peroxy radicals (HO2+RO2)
3.1 Model-measurements comparison for peroxy radicals
The time series of measured and modelled HO2+RO2are shown in Fig. 2and
the average diurnal profiles in different regions are shown in Fig. 3. In general, the
agreement between the model and the measurements was satisfactory throughout
the cruise: the campaign average modelled-to-measured ratio was 0.59, which is
within the uncertainty of the measurements (Sommariva et al. 2011). Previous studies
with similarly constrained models based upon the MCM showed comparable levels
of agreement (typically 30–40% or better) with the measurements for total peroxy
radicals, especially under unpolluted or semi-polluted conditions (Carslaw et al.
1999,1997; Fleming et al. 2006;Emmersonetal.2007). There were, however,
large differences in the model-measurements agreements in different regions and at
different times of the day. Figure 3shows that the model could reproduce the diurnal
JAtmosChem
4x109
3
2
1
0
[HO2+RO2] / molecule cm-3
8/23 8/25 8/27 8/29 8/31 9/2 9/4 9/6 9/8 9/10 9/12
GMT
4x109
3
2
1
0
7/31 8/2 8/4 8/6 8/8 8/10 8/12 8/14 8/16 8/18
measured
modelled
Fig. 2 Modelled and measured HO2+RO2during the TexAQS 2006 cruise of the R/V Brown
2.0x109
1.5
1.0
0.5
0.0
HO2+RO2 (molec cm-3)
00:00 12:00 00:00
GMT
Open Ocean
2.0x109
1.5
1.0
0.5
0.0
HO2+RO2 (molec cm-3)
00:00 12:00 00:00
GMT
Gulf Coast
2.0x109
1.5
1.0
0.5
0.0
HO2+RO2 (molec cm-3)
00:00 12:00 00:00
GMT
Galveston Bay
2.0x109
1.5
1.0
0.5
0.0
HO2+RO2 (molec cm-3)
00:00 12:00 00:00
GMT
Barbours Cut
2.0x109
1.5
1.0
0.5
0.0
HO2+RO2 (molec cm-3)
00:00 12:00 00:00
GMT
Industrial Areas
modelled
measured
Fig. 3 Diurnal profiles of modelled and measured HO2+RO2in different regions during the
TexAQS 2006 cruise of the R/V Brown.Thelines are the averages; the shaded area and the bars
are the 1-σstandard deviation of the model and of the measurements, respectively
JAtmosChem
profile of peroxy radicals, especially in the Open Ocean (e.g., July 30th and 31st as
shown in Fig. 2), although there was large variability and scatter, as illustrated by the
scatter plots in Figs. 4and 5. This was in part due to actual variability in the model
input data, especially in the most polluted areas, and in part related to the fact that
the model constraints were measured at different frequencies (from 1 min to 30 min)
thus forcing the model to interpolate the model constraints.
Day-time peroxy radicals The best agreement between the model and the measure-
ments was in the Open Ocean and in the Gulf Coast during day-time (Fig. 3). In the
Open Ocean, in the middle of the day, the model overestimated the measurements
by <30%, on average. During the first four modelled days (July 30th to August 2nd,
Fig. 2) the ship sampled air masses from the central Atlantic Ocean and the middle
of the Caribbean Sea, which were the cleanest conditions encountered during the
cruise and can be considered representative of background oceanic air. In the Gulf
Coast, the model underestimated the measurements by 15–30%, in the middle of the
day. In both areas the agreement was better in the early morning and in the late
evening (within 20% or better). Conversely, in Galveston Bay and Barbours Cut,
the model underestimated the measurements during the day by 30–40% and ∼40%,
on average (Fig. 3); in more polluted conditions, such as in the Industrial Areas, the
model underestimated the measurements by up to 55%, although in this region it is
2.5x109
2.0
1.5
1.0
0.5
0.0
HO2+RO2(modelled)
2.0x109
1.00.0
HO2+RO2(measured)
Open Ocean
500
400
300
200
100
0
222Rn (mBq/m3)
2.5x109
2.0
1.5
1.0
0.5
0.0
HO2+RO2(modelled)
2.0x109
1.00.0
HO2+RO2(measured)
Gulf Coast
10
8
6
4
2
0
C2H4 (ppb) 2.5x109
2.0
1.5
1.0
0.5
0.0
HO2+RO2(modelled)
2.0x109
1.00.0
HO2+RO2(measured)
Galveston Bay
20
15
10
5
0
HCHO (ppb)
2.5x109
2.0
1.5
1.0
0.5
0.0
HO2+RO2(modelled)
2.0x109
1.00.0
HO2+RO2(measured)
Barbours Cut
40
30
20
10
0
NO2 (ppb) 2.5x109
2.0
1.5
1.0
0.5
0.0
HO2+RO2(modelled)
2.0x109
1.00.0
HO2+RO2(measured)
Industrial Areas
12
10
8
6
4
2
0
CH3CHO (ppb)
Fig. 4 Day-time modelled vs. measured HO2+RO2during the TexAQS 2006 cruise of the
R/V Brown.Theblack line is the 1:1 line and the data are color-coded with different parameters
in each region
JAtmosChem
5x109
4
3
2
1
0
HO2+RO2(modelled)
5x109
43210
HO2+RO2(measured)
Open Ocean
100
80
60
40
20
0
O3 (ppb)
5x109
4
3
2
1
0
HO2+RO2(modelled)
5x109
43210
HO2+RO2(measured)
Gulf Coast
5x109
4
3
2
1
0
HO2+RO2(modelled)
5x109
43210
HO2+RO2(measured)
Barbours Cut
5x109
4
3
2
1
0
HO2+RO2(modelled)
5x109
43210
HO2+RO2(measured)
Industrial Areas
Fig. 5 Night-time modelled vs. measured HO2+RO2during the TexAQS 2006 cruise of the
R/V Brown.Theblack line is the 1:1 line and the data are color-coded with O3concentrations
more difficult to assess the level of agreement due to rapidly changing conditions and
large variability.
Figure 4shows scatter plots of day-time modelled vs. measured HO2+RO2,
color-coded with selected parameters to analyze the deviations from linearity. In
the Open Ocean the model overestimated the measurements at higher 222Rn counts,
indicating that some air masses of continental origin were also sampled in this region;
in the Gulf Coast, the model tended to underestimate the measurements at higher
concentrations of alkenes and other pollution markers, such as NO2(Fig. 4). These
events likely corresponded to specific emission sources, such as industries on the
coast, other ships or oil extraction platforms in the Gulf. In Galveston Bay, Barbours
Cut and in the Industrial Areas the model underestimated the measurements at
higher concentrations of oxygenated VOCs, PANs, NO2(Fig. 4). The analysis of
HO2+RO2and VOCs observations presented in Gilman et al. (2009), Sommariva
et al. (2011) indicated that oxygenated VOCs in these areas were connected with
aged air masses that were recirculated over Galveston Bay for several hours before
coming back towards Houston.
In general, the model was more reliable under cleaner conditions in all regions.
The dependence of the model-measurements disagreement on oxygenated VOCs
and other secondary oxidation products in the polluted regions suggests that the
discrepancy might be related to the description of carbonyl photochemistry in the
MCM. One aspect of this problem is the formation of acyl peroxy radicals (e.g.,
JAtmosChem
CH3CO3,C
2H5CO3,C
3H7CO3) from aldehydes and ketones, which will be discussed
in more detail in Section 3.2.
Night-time peroxy radicals The R/V Brown did not spend much time in Galveston
Bay at night, so this region will not be considered when discussing night-time
chemistry. The agreement between modelled and measured HO2+RO2during
night-time is more difficult to assess, owing to larger variability and scatter in both the
model and the measurements (Figs. 2and 5). In Barbours Cut and in the Industrial
Areas the agreement was typically within 40–50%, although sometimes it was within
20%(Fig.3). In these areas the conditions often varied rapidly as plumes of NOx
and VOCs from industries and port operations affected the observations. However,
as explained earlier, the model is not able to reproduce this high frequency variability
because it does not have all the information that would be required; this was more
of an issue at night, because during the day ozone (which was measured at high
frequency) was a major driver for radical chemistry even under polluted conditions
(Sommariva et al. 2011). On the other hand, the average night-time profiles (Fig. 3)
showed reasonably good agreement (25–30%) under clean conditions (Open Ocean
and Gulf Coast).
The analysis of the night-time scatter plots did not show any parameters clearly
associated with the model-measurements disagreement. A weak correlation with
NO2and O3was apparent in the Industrial Areas and, less clearly, in Barbours Cut
(Fig. 5): the model tended to overestimate the measurements at lower concentrations
of O3and NO2and underestimate at higher concentrations of O3and NO2.Since
ozonolysis reactions of VOCs were major sources of night-time peroxy radicals
(Section 3.3), this might point to problems in the treatment of these reactions
in the MCM, particularly with regard to their efficiency as radical sources. It is
however difficult to identify any particular reaction or group of reactions, because the
correlation was weak and no specific VOC was likewise correlated with the model-
measurements disagreement.
Additional VOCs A key factor that determines the agreement between the model
and the measurements is the degree to which all of the RO2precursors were included
in the model. Some of the VOCs measured on the R/V Brown which are not included
in the MCM occasionally accounted for a significant fraction of the OH reactivity
during the R/V Brown cruise (e.g., vinyl acetate). There were likely also unmeasured
VOCs, especially when the ship was in the more polluted Industrial Areas: the peroxy
radicals formed by these VOCs were also not included in the MCM model results.
During the cruise, the GC-MS detected a peak corresponding to 2-methyl-2-
butene and to acrolein. The two species could not be separated, therefore both
VOCs were excluded from the model contraints. In order to evaluate the impact of
additional VOCs on the concentration of RO2, a constraint was added to the model
corresponding to a species with the same reactivity of 2-methyl-2-butene and the
concentration determined by the co-eluted peak (average =30 ppt; maximum =
670 ppt). The results showed that the modelled concentration of organic peroxy
radicals increased by <5%, on average. However, RO2could increase by 20–
25% (with a maximum of 35–40%) when large plumes of 2-methyl-2-butene +
acrolein were sampled. The results show that modelled HO2+RO2was sensitive
to concentrated plumes of reactive VOCs from specific sources, but less sensitive
JAtmosChem
to the background levels, even if they were influenced by mixing and dilution of
similar plumes. These effects lead to the large variability observed in Figs. 2and 3
for Barbours Cut and the Industrial Areas.
3.2 Peroxy radicals and PANs
It is difficult to assess the accuracy of the MCM model in calculating the concen-
trations of individual organic peroxy radicals, because there are no observations
available. The only measurement available is the sum of peroxy radicals (HO2+
RO2) and it is possibile that even if the modelled sum agreed reasonably well
with the measured sum (Section 3.1), the individual species were not correctly
represented. In fact, previous comparisons of peroxy radicals using MCM-based
models have showed varying levels of agreement with the observations of HO2;for
example, the modelled concentration of HO2in the marine boundary layer often
overestimated the observations (Carslaw et al. 1999,2002; Sommariva et al. 2004,
2006) and sometimes underestimated them or showed good agreement (Carslaw
et al. 1999;Emmersonetal.2007; Whalley et al. 2010). This suggests that the MCM
often underestimated and sometimes overestimated RO2. In addition, recent reports
that some of the HO2measurements (by laser-induced fluorescence) suffer from
interference under polluted conditions (Fuchs et al. 2011) indicate that the agreement
between modelled and measured HO2might be worse, except on the occasions when
the model underestimated HO2.
The analysis of the model results discussed in Section 3.1 indicated that, in the
more polluted regions (Galveston Bay, Barbours Cut, Industrial Areas), at high
concentrations of oxygenated VOCs, the model underestimated the HO2+RO2
measurements; this might indicate that the model underestimated the concentrations
of acyl peroxy radicals formed by the oxidation of oxygenated VOCs. Each acyl
peroxy radical is the precursor of a single PAN species (e.g., CH3CO3for PAN and
C2H5CO3for PPN). The model can be used to calculate a realistic concentration
of PAN species under certain conditions: if transport is not a dominant factor in
determining the concentrations of PANs and if their background concentrations can
be neglected. At the average (15:00–21:00 GMT, 10:00–16:00 Local Time) conditions
in Barbours Cut during the R/V Brown cruise (NO =9.3×1010 molecule cm−3,
NO2=1.5×1011 molecule cm−3, Temperature =303 K) the net lifetime of PAN—
i.e., the lifetime calculated accounting for the reformation of PAN in the presence
of NO2—is ∼45 min, comparable to the interval between two consecutive inputs of
VOCs in the model (determined by the sampling frequency of the GC instrument).
The observations reported by Roberts et al. (2003) at LaPorte, Texas, approximately
10 km West of Barbours Cut (Fig. 1), clearly showed that PAN was photochemi-
cally formed in this area, with concentrations decreasing to below the instrument’s
detection limit during the night (evidence of very low background values). Under
these conditions, the agreement between modelled and measured PANs can be used
as diagnostic tools to qualitatively assess how well the MCM model calculates the
concentration of some individual RO2.
Modelled and measured PAN and PPN average profiles during day-time in
Barbours Cut are shown in Fig. 6. The model underestimated PAN in the morning by
30–40%, but overestimated it by 25% in the afternoon (after 19:00 GMT, 14:00 Local
Time). On the other hand, the model agreed reasonably well with measurements of
JAtmosChem
5x10
-6
4
3
2
1
0
j
CH3CHO
(s
-1
)
15:00 18:00 21:00
GMT
1.0x10
10
0.8
0.6
0.4
0.2
0.0
PAN (molec cm
-3
)
Barbours Cut
5x10
-6
4
3
2
1
0
j
CH3CHO
(s
-1
)
15:00 18:00 21:00
GMT
2.0x10
9
1.5
1.0
0.5
0.0
PPN (molec cm
-3
)
Barbours Cut
photol. rate
modelled
(MCM)
modelled
(IUPAC)
measured
Fig. 6 Average day-time profiles of modelled and measured PAN, PPN in Barbours Cut during the
TexAQS 2006 cruise of the R/V Brown
PPN in the morning (within 20%), but overestimated measured PPN by approxi-
mately a factor of 2 in the afternoon (Fig. 6). The uncertainty in the observations was
15%(Section2.1). This analysis requires that the source and the sink terms of PANs
are well constrained in the model. The MCM v3.1 uses the same kinetic parameters
for both PAN and PPN (Jenkin et al. 1997; Saunders et al. 2003). While the PAN
kinetic data are consistent with the IUPAC recommendations (Atkinson et al.
2006), the PPN rate coefficients are different. The rate coefficient for the formation
reaction is 11% lower and the rate coefficient for the decomposition reaction is
20% higher than those used in the MCM v3.1 (Seefeld and Kerr 1997; Kirchner
et al. 1999; Atkinson et al. 2006): using the recommended rate coefficients instead
of the MCM rate coefficients resulted in lower (∼26%) modelled concentrations
of PPN (Fig. 6), slightly improving the agreement with the measurements in the
afternoon. Given the limitations of the modelling approach, the agreement between
the model and the measurements can be considered satisfactory for PAN, but less
so for PPN, suggesting that the MCM underestimates the concentration of C2H5CO3
(and possibly of other acyl peroxy radicals) by as much as a factor of two, especially
in the morning and in the central part of the day (i.e., before 14:00).
3.3 Composition of the peroxy radicals pool
The explicit treatment of chemistry in the MCM allows the calculation of the concen-
trations of individual organic peroxy radicals (RO2), which could not be measured
during TexAQS 2006. The relative importance of each peroxy radical in an air
mass depends on the VOC composition of the air mass. Figure 7shows the average
modelled fraction of total peroxy radicals constituted by HO2,CH
3O2,CH
3SCH2O2
and selected organic peroxy radicals in the five regions of the R/V Brown cruise
(Fig. 1). In the following discussion, only the most abundant peroxy radicals have
been considered: the sum of the selected peroxy radicals accounted for at least
85% of modelled HO2+RO2during day-time. The remaining fraction of modelled
HO2+RO2consisted of a large number of organic peroxy radicals, mostly derived
by long-chain VOCs and secondary oxidation products, each accounting for a small
percentage of the total; this fraction was proportionally smaller in the more polluted
JAtmosChem
1.0
0.8
0.6
0.4
0.2
0.0
RO2/(HO2+RO2)
00:00 12:00 00:00
GMT
night day
Open Ocean
1.0
0.8
0.6
0.4
0.2
0.0
RO2/(HO2+RO2)
00:00 12:00 00:00
GMT
night day
Gulf Coast
1.0
0.8
0.6
0.4
0.2
0.0
RO2/(HO2+RO2)
00:00 12:00 00:00
GMT
night day
Galveston Bay
1.0
0.8
0.6
0.4
0.2
0.0
RO2/(HO2+RO2)
00:00 12:00 00:00
GMT
night day
Barbours Cut
1.0
0.8
0.6
0.4
0.2
0.0
RO2/(HO2+RO2)
00:00 12:00 00:00
GMT
night day
Industrial Areas
HO2
CH3O2
CH3SCH2O2
ISOPO2
ALKAO2
ALKEO2
CARBO2
AROMO2
Fig. 7 Average modelled RO2/(HO2+RO2) ratios in different regions during the TexAQS 2006
cruise of the R/V Brown.SeeAppendix for the RO2codenames
regions (Fig. 7), where continuous emissions of primary VOCs of industrial origin
controlled the concentrations of the peroxy radicals precursors.
The selected organic peroxy radicals were divided into five classes (ISOPO2,
ALKAO2, ALKEO2, CARBO2, AROMO2), roughly corresponding to the func-
tionality of their main VOC precursors (biogenics, alkanes, alkenes, carbonyls,
aromatics). The classification is not straightforward because most peroxy radicals
have more than one VOC precursor and often these have different functional groups:
additionally, small RO2are often formed from the fragmentation of longer carbon
chains (Sommariva et al. 2008,2011). The peroxy radicals included in the five
classes (ISOPO2, ALKAO2, ALKEO2, CARBO2, AROMO2) and the lists of their
corresponding precursors are given in Appendix.
During the day, HO2was always the dominant peroxy radical, accounting for
∼50% of the total in all regions; the single most important organic peroxy radical
during the day was always CH3O2, which accounted for ∼30% of the total in the
Open Ocean, 20–25% in the Gulf Coast and ∼15% in the more polluted areas
around Houston, Texas (Fig. 7). Day-time radical chemistry under clean conditions
is dominated by the reactivity of CO and CH4, the main precursors of HO2and
CH3O2, which explains why in the Open Ocean they accounted together for ∼80%
of HO2+RO2. Oxygenated VOCs (HCHO) and light alkenes (ethene, propene),
which were the most important OH sinks in the Open Ocean besides CO and CH4
(Gilman et al. 2009), are also efficient sources of HO2and CH3O2; moreover, the
JAtmosChem
third and the fourth most important contributions (∼8% and ∼7%, respectively) to
the RO2pool were from the other organic RO2formed from oxygenates and light
alkenes (CARBO2 and ALKEO2, Appendix,Fig.7).
By contrast, under more polluted conditions (Galveston Bay, Barbours Cut
and Industrial Areas) the contribution of methyl peroxy radical to the RO2pool
decreased, while the contributions from VOCs (mostly alkenes and oxygenates)
increased up to 35%oftotalHO
2+RO2(Fig. 7). Gilman et al. (2009) estimated
that most of the reactivity of OH in the Industrial Areas was due to the oxidation
of alkenes and other highly reactive species (e.g., di-alkenes) and of light (C2-C4)
alkanes (corresponding to ALKAO2 and ALKEO2 peroxy radicals, respectively, in
Fig. 7). The contribution from aromatics was almost always negligible, except for
occasional plumes, but there were significant contributions (7–9%) from isoprene-
related peroxy radicals in Barbours Cut and in the Industrial Areas. Isoprene and
other biogenic VOC were also present at night and contributed significantly to the
RO2pool (20–30%) and to the reactivity of the nitrate radical (Section 4.2)inthese
regions. Gilman et al. (2009) concluded, on the basis of the observed diurnal profile,
that isoprene in the Houston area was mostly biogenic in origin and partly from
industrial sources, but there is evidence (Kuster et al. 2004; Stutz et al. 2010)that
night-time isoprene concentrations were largely related to industrial emissions.
At night HO2accounted for only 10% in the Open Ocean, but up to 20–30% in the
Industrial Areas (Fig. 7). On the other hand the fraction of HO2+RO2constituted
by CH3O2remained approximately the same as during the day in all regions (Fig. 7).
In general, in the Industrial Areas there was less variability between day and night,
as far as the composition of the RO2pool is concerned, because industrial emissions
of VOCs were continuous and did not follow a diurnal cycle. In the Open Ocean
and, to a lesser extent, in the Gulf Coast, CH3SCH2O2—the peroxy radical formed
by the reaction of DMS with NO3– contributed to a significant fraction to HO2+
RO2: 15–20% and ∼10%, respectively. However, DMS oxidation did not contribute
significantly to day-time peroxy radicals, in accord with the findings of Gilman et al.
(2009) that DMS was not an importat reactant for OH during the R/V Brown cruise.
A discussion about the formation of peroxy radicals and, hence, ozone from
different classes of VOCs in the Houston area has been presented in Sommariva et al.
(2011). In that paper it was shown how photochemical formation of ozone was linked
to different classes of VOCs, depending on the sources in each region. Combining
that analysis and the speciation of peroxy radicals obtained with the MCM, the
individual VOCs and peroxy radicals that most contributed to photochemical ozone
formation can be identified. For example, in Barbours Cut, Sommariva et al. (2011)
found that the highest ozone formation rates were related to E-SE air masses rich
in oxygenated and biogenic VOCs, followed by E-NE and W-SW air masses rich
in alkenes: this is consistent with the analysis shown in Fig. 7, which shows that,
besides HO2and CH3O2, the most important contributors to the RO2pool were
the ISOPO2 and CARBO2 peroxy radicals, which were mostly derived from the
oxidation of isoprene, carbonyls and long-chain alkenes (Appendix).
3.4 Case study: Jacinto Point
The complexity of the RO2pool composition is well illustrated by the events during
the night of September 7th, when the R/V Brown was in Jacinto Point, on the North
JAtmosChem
side of the Houston Ship Channel entrance (Fig. 1). The maximum HO2+RO2
concentrations observed during TexAQS 2006 were measured during this night: 134
ppt and 123 ppt (=3.35 ×109and 3.1×109molecule cm−3) around 9:30 GMT (4:30
Local Time) and a second peak of 82 ppt (2.05 ×109molecule cm−3) about two hours
later. These peaks were clearly correlated with changes in the local wind direction
(Fig. 8), indicating that plumes emitted from nearby industries were being sampled.
The composition of these plumes was a complex mixture of VOCs, including
alkanes (up to 450 ppb of n-butane) and a range of light alkenes, some of which are
1.0
0.8
0.6
0.4
0.2
0.0
RO2/(HO2+RO2)
00:00
9/7/2006
12:00 00:00
9/8/2006
GMT
night dayday
HO2
CH3O2
CH3SCH2O2
ISOPO2
ALKAO2
ALKEO2
CARBO2
AROMO2
4x109
3
2
1
0
HO2+RO2
(molec cm-3)
00:00
9/7/2006
12:00 00:00
9/8/2006
GMT
300
200
100
0
Wind Direction (deg)
107
109
1011
1013
1015
NO (molec cm-3)
60
40
20
0
VOCs (ppb)
HO2+RO2 Wind Direction NO ethene propene
2-methyl-1-Butene 1,3-butadiene methylpropene 1-butene
Fig. 8 Modelled RO2/(HO2+RO2)ratios(top) and measured concentrations of selected species
(bottom) during the night of September 7th at Jacinto Point, near the entrance of the Houston Ship
Channel (HSC in Fig. 1). See Appendix for the RO2codenames
JAtmosChem
showninFig.8. The concentration of NO between 9:00–12:00 GMT (4:00–7:00 Local
Time) ranged between 800 ppt (2.0×1010 molecule cm−3) and 38 ppb (9.5×1011
molecule cm−3), which excludes any significant contribution to VOCs oxidation by
the nitrate radical, since NO3is efficiently destroyed by reaction with NO. This
means that the formation of the peroxy radicals observed in Jacinto Point was related
to ozonolysis reactions (Fig. 9) and therefore to the alkenes in the sampled plumes.
The most abundant alkenes in the two plumes were ethene and propene (14 ppb
and 9 ppb, respectively, in the 9:30 plume, and up to 55 ppb each in the 11:30
plume); based on the local wind direction, the earlier plume came from the South
and the later plume came from East-Northeast indicating that they were from two
different industrial sources (Fig. 8). The composition was also different: the 9:30
plume contained lower concentrations of ethene and propene and up to 3 ppb of
1,3-butadiene, while the 11:30 plume contained higher concentrations of 1-butene
and methylpropene (up to 23.5 ppb and 11.3 ppb, respectively) and less than 300 ppt
of 1,3-butadiene.
The composition of the peroxy radicals pool (Fig. 8) clearly changed throughout
the night of September 7th, as different air masses were being sampled. For example,
HO2accounted for about 12%ofHO
2+RO2at 9:00 GMT, but 4–5% during the 9:30
plume and ∼47% during the 11:30 plume. The MCM analysis also showed that the
first plume was composed almost equally of ALKAO2 and ALKEO2 peroxy radicals
(20–30%), while the second plume was richer in ALKEO2 peroxy radicals (∼37%).
The difference between the two plumes can be explained from the breakdown of
the individual radical species: with the exception of 1-butadiene, the concentrations
of most alkenes, and especially of ethene and propene, were much higher in the
Fig. 9 Formation of radicals from the ozonolysis of alkenes (1-butene in this example) in the MCM
v3.1 (Saunders et al. 2003). Peroxy radicals are highlighted in blue,OHinred; the numbers are the
branching ratios
JAtmosChem
11:30 plume than in the 9:30 plume. Since ozonolysis of alkenes promptly forms HO2
(Fig. 9), this explains why HO2was a much larger contributor to the HO2+RO2
pool in the 11:30 plume. However, in the presence of such high concentrations of NO
(Fig. 8), HO2reacted to form OH, which could readily react with the alkenes to form
OH-additioned peroxy radicals (e.g., HOCH2CH2O2, HYPROPO2, IPROPOLO2,
etc...). The peroxy radicals formed by addition of OH reactions are included in
ALKEO2 (Appendix), while the peroxy radicals formed by ozonolysis reactions
(e.g., C2H5O2, NC3H7O2, IC3H7O2, etc...) are included in ALKAO2 (Appendix),
which explains why ALKEO2 was proportionally a larger contributor to HO2+RO2
in the second plume than in the first plume. Night-time formation of OH (Fig. 9)
was occurring in both plumes, but the concentration of OH was very different: the
MCM model calculated <1×105molecule cm−3during the first event and >3×106
molecule cm−3during the second event, so the formation of RO2by addition of OH
to alkenes (ALKEO2) was a much larger factor in the latter.
4 Nitrate radical (NO3)
4.1 Model-measurements comparison for nitrate radical
The MCM model was used to calculate the concentrations of NO3and N2O5during
the R/V Brown cruise. The NO3–N2O5chemical system is schematically represented
in Eqs. 2–5, where X is a generic sink for NO3(e.g., reactions with NO, VOCs and
photolysis) and Y is a generic sink for N2O5(e.g., aerosol uptake, hydrolysis):
NO2+O3→NO3+O2(2)
NO3+NO2N2O5(3)
NO3+X→products (4)
N2O5+Y→products (5)
The model is designed for fast reacting species, such as OH and peroxy radicals,
and it does not include vertical or horizontal transport. As such, the model results can
be considered reliable only under conditions where the reactivity of NO3is rapid, but
are less reliable where reactivity is slow. A model calculation based on integration
from one observed data point to the next in a time series may generate error if the
lifetime of the calculated species (NO3in this case) is long with respect to the model
time steps. In this case, changes due to air mass shifts (i.e., transport) will not be
properly represented. Such errors should lead to over and under-predictions with
equal probability if there are random changes in the air mass between successive
measurements due to transport or to the movement of the ship. NO3levels controlled
by industrial emissions of highly reactive VOCs or by marine emissions of DMS can
be considered suitable for modelling with a zero-dimensional box-model.
The concentrations of NO3(and N2O5) were calculated for 20 nights of the
R/V Brown cruise, mostly during the second half of each leg when the ship spent
more time in semi-polluted and polluted areas than in the clean marine boundary
layer. The full comparison between the model results and the measurements is shown
JAtmosChem
in Fig. 10 as a time series and the average night-time profiles of NO3for each region
areshowninFig.11. The measured concentrations were often below the detection
limit (5.0×107molecule cm−3) of the CaRDS instrument under very clean and very
polluted conditions. For example, on the nights of August 13th, 14th and 15th, the
R/V Brown was in Barbours Cut and sampled air masses influenced by local sources
with concentrations of NO of about 10 ppt, with peaks of ppb level: on those nights
NO3was below the instrumental detection limit, due to the fast reaction with NO,
which was well reproduced by the model (Fig. 10).
4x10
9
3
2
1
0
[NO
3
] / molecule cm
-3
8/25 8/29
9/2 9/6
9/10
GMT
8x10
8
6
4
2
0
8/2 8/6
8/10 8/14 8/18
measured
modelled
1.5x10
10
1.0
0.5
0.0
[N
2
O
5
] / molecule cm
-3
8/25 8/29
9/2 9/6
9/10
GMT
8x10
8
6
4
2
0
8/2 8/6
8/10 8/14 8/18
measured
modelled
Fig. 10 Modelled and measured NO3,N
2O5during the TexAQS 2006 cruise of the R/V Brown.For
clarity, the two parts of the cruise are plotted on different scales
JAtmosChem
2.5x109
2.0
1.5
1.0
0.5
0.0
NO3 (molec cm-3)
00:00 06:00 12:00
GMT
Open Ocean
modelled
measured
2.5x109
2.0
1.5
1.0
0.5
0.0
NO3 (molec cm-3)
00:00 06:00 12:00
GMT
Gulf Coast
2.5x109
2.0
1.5
1.0
0.5
0.0
NO3 (molec cm-3)
00:00 06:00 12:00
GMT
Barbours Cut
Fig. 11 Night-time profiles of modelled and measured NO3in different regions during the TexAQS
2006 cruise of the R/V Brown.Thelines are the averages; the shaded area and the bars are the 1-σ
standard deviation of the model and of the measurements, respectively
In general, the model underestimated the observations during the entire
R/V Brown cruise. The average modelled-to-measured ratios were different in each
region (Fig. 11): ∼0.3 in the Open Ocean, ∼0.4 in the Gulf Coast and in Barbours
Cut. Data in Galveston Bay and Industrial Areas were too sparse to provide a
meaningful statistics. The agreement between the model and the measurements was
similar for N2O5, with a campaign average modelled-to-measured ratio of ∼0.4.
These results are in contrast with a previous study which used a similar modelling
approach (Sommariva et al. 2009): during the NEAQS 2004 cruise, the model
overestimated the measurements by 30–50%, on average. During NEAQS 2004, the
concentrations of NO3were typically a factor of 6–8 smaller and the NO3production
rates (k2×[O3]×[NO2]+k−3×[N2O5]) were typically a factor of 10–12 smaller
than during TexAQS 2006. The comparison of the two studies suggests that under
conditions of larger NO3production rate the model tends to underestimate NO3:the
reason for this difference is otherwise difficult to assess.
Additional VOCs In order to evaluate the impact of additional VOCs on the
concentration of NO3–N2O5system (Eq. 4), a constraint was added to the model,
corresponding to a species with the same reactivity of 2-methyl-2-butene and
the concentration determined by the co-eluted chromatographic peak 2-methyl-2-
butene +acrolein (Section 3.1). The model results showed a decrease in the modelled
concentrations of NO3(and N2O5)of∼10%, on average, and up to 40% during the
largest plume of 2-methyl-2-butene event. The addition of a highly reactive species
could therefore improve agreement on the occasions when the model overestimated
the measurements (Fig. 10). The concentration attributed to 2-methyl-2-butene was
about 20 ppt on that night, but it suggests that other unmeasured VOCs might
have been present. It must be noted, however, that emissions of VOCs were often
coincidental with emissions of NO in the proximity of industrial activities and/or
ship plumes. Since even low concentrations of NO would deplete the concentration
of NO3, the presence of unmeasured/unknown VOCs could explain the model-
measurement discrepancies only under low NO concentrations.
JAtmosChem
N2O5uptake One of the key parameters that control the concentration of NO3
is the uptake coefficient of N2O5(γN2O5,Eq.5) on aerosol. Several studies have
demonstrated the variability of γN2O5in the lower atmosphere, due to the variability
in the composition of aerosol, ambient relative humidity and temperature. In this
work, the base model used a constant γN2O5=0.02 (Section 2.1, Aldener et al. 2006).
In order to test the sensitivity of the calculated concentrations of NO3to aerosol
uptake under the conditions of TexAQS 2006, the model was run with γN2O5=0.001
(Brown et al. 2006)andwithγN2O5=0. The results showed that the concentration
of NO3increased, on average, by about 17% and 18%, respectively, due to slower
or lack of removal of N2O5via aerosol uptake (Fig. 12). Since the uptake rate is
linearly dependent on the aerosol surface area (A,Eq.1), the model showed a similar
sensitivity to the aerosol surface area: varying Aby a factor of 2 resulted in a change
of about ±9% in the modelled concentration of NO3.
A slower uptake rate for N2O5(because of lower γN2O5and/or aerosol surface
area) would improve the agreement between the model and the measurements.
Suppression of γN2O5has been correlated to organic-rich particles (Brown et al. 2006);
however, submicron particles measured during TexAQS 2006 when the R/V Brown
was near the polluted areas or during periods of continental outflow were largely
inorganic (only 22–36% of organic matter, Bates et al. 2008). It must also be noted
that a γN2O5value of the order of 0.001 is inconsistent with ClNO2production from
heterogeneous uptake of N2O5, which was observed during the R/V Brown cruise
(Osthoff et al. 2008).
These results suggest that additional loss terms for N2O5, e.g., the gas-phase
hydrolysis which has been suggested as potential sink by laboratory experiments
(Menteletal.1996; Wahner et al. 1998) would result in a larger discrepancy between
the model and the measurements. This is in accord with some of the previous studies
of the nitrate radical chemistry (Aldener et al. 2006; Brown et al. 2006), although
it is in contrast with others (Ambrose et al. 2007; Sommariva et al. 2009). The
model response to changes in the uptake coefficient of N2O5indicate relatively low
sensitivity to N2O5sinks, suggesting that the concentrations of the nitrate radical
during TexAQS 2006 were controlled by NO3sinks instead.
3.0x10
9
2.5
2.0
1.5
1.0
0.5
0.0
modelled (
γ
N2O5
= 0.001)
2.5x10
9
0.0
modelled (
γ
N2O5
= 0.02)
slope = 1.17
r = 0.98
4x10
9
3
2
1
0
modelled (
γ
N2O5
= 0.02)
4x10
9
3210
measured
slope = 0.22
r = 0.53
4x10
9
3
2
1
0
modelled (
γ
N2O5
= 0.001)
4x10
9
3210
measured
slope = 0.25
r = 0.49
Fig. 12 Left: comparison of modelled NO3calculated with γN2O5=0.02 and with γN2O5=0.001.
Middle: modelled vs measured NO3with γN2O5=0.02.Right: modelled vs measured NO3with
γN2O5=0.001.Theblack lines are the 1:1 lines and the red lines are the linear fits
JAtmosChem
4.2 Nitrate radical budget and interactions with RO2
The nitrate radical has only one well known source, the reaction between NO2
and O3(Eq. 2). The sinks, however, are more uncertain: besides the equilibrium
reaction with NO2to form N2O5,NO
3reacts with a large number of VOCs, with
peroxy radicals and, if present, with NO. The relative importance of each sink term
determines the impact of NO3chemistry on the composition of the troposphere. For
example, reactions with some VOCs form organic nitrates, which act as reservoirs of
nitrogen and have been linked to the formation of secondary organic aerosol (Brown
et al. 2009); the reaction with DMS can be the most important process controlling the
oxidation of this species (Osthoff et al. 2009) and the reactions with peroxy radicals
can generate night-time OH radicals (Vaughan et al. 2006). The measurements of
NO3, VOCs and HO2+RO2made during TexAQS 2006 were used to calculate the
first-order loss rate of NO3during the R/V Brown cruise. The results—averaged and
divided according to the location of the ship—are shown in Fig. 13a, b. Data from
Galveston Bay were neglected in the following analysis because too few data were
taken in this region during the night.
Alkanes and aromatics were negligible contributors to the gas-phase removal
of NO3and oxygenated VOCs contributed up to 2% only in the Gulf Coast.
The major sinks for NO3in all regions were alkenes, biogenic VOCs, DMS and
peroxy radicals (Fig. 13b). In the Open Ocean and in the Gulf Coast, the most
important NO3sink was DMS (79%and38%, respectively). This is consistent with
the modelled composition of the peroxy radicals pool during the night discussed
earlier (Fig. 7,Section3.3), which showed a significant presence of CH3SCH2O2.
DMS was a significant sink also in the more polluted areas, although it was much
less important than alkenes (19–26%) and biogenic VOCs (50–58%). As explained
above (Section 3.3), biogenic VOCs also have industrial origin and this was likely the
case in Barbours Cut and the Industrial Areas at night-time (Stutz et al. 2010), since
emissions from vegetation typically follow a diurnal cycle.
Peroxy radicals were important contributors to the gas-phase removal of NO3,
especially in the less polluted areas. In the more polluted areas (Barbours Cut and
Industrial Areas) the contribution of peroxy radicals was comparable to those of
aromatics and more important than alkanes and oxygenated VOCs (1–2%). Since
HO2was typically a minor component of the peroxy radicals pool during the night
(Fig. 7) this loss process can mostly be attributed to the reaction of NO3with organic
peroxy radicals. In the Open Ocean, peroxy radicals were the second most important
NO3sink after DMS (12%) and in the Gulf Coast they were the third most important
NO3sink after DMS and alkenes (28%).
The numbers shown in Fig. 13b can be compared with those reported by Brown
et al. (2011), derived by three flights around Houston, Texas, in October 2006. In air
masses advected from the Houston Ship Channel, biogenic and oxygenated VOCs
contributed ∼15% and ∼18%, respectively, to the gas-phase loss of NO3.Thisis
very different from the budget calculated from the ship data when the R/V Brown
was in Barbours Cut or in the Industrial Areas (Fig. 13b) which show much smaller
contribution from oxygenated VOCs (∼1%) and much higher contribution from
biogenic VOC (50–58%). The R/V Brown data were closer to the source than the
aircraft data, so much of the biogenics were oxidized during transport and secondary
products (such as oxygenates) were more important a few miles downwind of the
JAtmosChem
10-7
10-6
10-5
10-4
10-3
10-2
s-1
Alkanes
Alkenes
Aromatics
DMS
Biogenics
Oxygenates
HO2+RO2
Open Ocean
Gulf Coast
Barbours Cut
Industrial Areas
DMS
(a)
(b)
Fig. 13 Average measured first-order loss terms for NO3between 01:00 and 11:00 GMT (20:00–
06:00 local time) in different regions during the TexAQS 2006 cruise of the R/V Brown:aabsolute
values in s−1,brelative values in %
emission sources. Brown et al. (2011) calculated that peroxy radicals contributed 4–
5% to the total loss of NO3; this is larger than 1–2% observed in Barbours Cut and
in the Industrial Areas (Fig. 13b), which are near the emission sources and where
primary VOCs concentrations were high. However, it is much smaller than 12–28%
observed on the R/V Brown in the Open Ocean and in the Gulf Coast (Fig. 13b):
both of these are downwind regions, as were those sampled by the aircraft in Brown
et al. (2011), but peroxy radicals were measured on the R/V Brown while they were
estimated on the aircraft (based on PANs observations).
JAtmosChem
The importance of RO2+NO3reactions as sinks for the nitrate radical has
been noted previously: for example, the model study by Sommariva et al. (2009)
determined that RO2accounted on average for ∼20% of the NO3losses in a semi-
polluted environment; the observations reported here clearly support this view and
the model results were consistent with the previous study. The implications of organic
peroxy radicals being a major sink for NO3are yet unclear, mostly because the kinetic
database of this type of reaction is limited (Canosa-Mas et al. 1996; Vaughan et al.
2006). These reactions are known to form NO2and HO2, which can be a source of
ozone at sunrise and a source of OH at night, respectively. For example, in the case
of methyl peroxy radical:
CH3O2+NO3→CH3O+NO2→HCHO +HO2+NO2(6)
The formation of OH via these night-time reactions occurs largely via the reaction
of HO2with O3, as NO concentration must be very low (otherwise it would titrate
NO3). In turn, the oxidation of VOCs will form other RO2, starting a positive
feedback on the concentrations of radical species that can significantly enhance
the oxidation of VOCs at night, as long as NO3production is sustained. On the
other hand, the sequence of reactions in Eq. 6decreases the night-time removal
of NOx(via NO3+NO2N2O5aerosol
−−−−→ HNO3and NO3+VOCs −→ HNO3reac-
tions, followed by HNO3deposition) resulting in more NO2being available at sunrise
to form ozone. It is, however, not possible to quantify this effect with a box-model in
which both NO2and O3were constrained to the observations (Section 2.1).
5 Summary
During the TexAQS 2006 cruise of the R/V Brown observations of total peroxy
radicals (HO2+RO2) and nitrate radical (NO3) were taken in the Atlantic Ocean,
in the Gulf of Mexico and in the industrial regions around Houston, Texas. These
measurements were compared with the results of a box-model based upon the Master
Chemical Mechanism (MCM) and constrained to the chemical and physical parame-
ters measured onboard the R/V Brown. The model could reproduce the observations
of HO2+RO2during the day to within ∼40%, on average. In the unpolluted regions,
the agreement was better, both during day-time (15–30%) and during night-time
(25–30%). The analysis of the model results suggests that the model underestimated
the measurements particularly in aged air masses, e.g., at high levels of oxygenated
VOCs, and that the MCM might underestimate the concentrations of some acyl
peroxy radicals and, possibly, of other short-chain RO2.
The information included in the MCM was used to estimate the relative impor-
tance of individual peroxy radicals in various regions of the R/V Brown cruise and,
in particular, at Jacinto Point, a location near the Houston Ship Channel where the
highest concentrations of HO2+RO2were observed (134 ppt) during the night. The
break-down of the RO2pool indicated that HO2constituted ∼50% of the peroxy
radical pool during the day and between 10%and30% during the night. During the
night, isoprene—which was in part of industrial origin—contributed up to 30%tothe
RO2pool in the Industrial Areas, and DMS <20% in the unpolluted regions.
The model consistently underestimated the measurements of NO3, especially
in the Open Ocean: the typical model-to-measurements ratio was 0.3–0.4. The
JAtmosChem
agreement between the model and the measurements could be slightly improved
by using a lower uptake coefficient for N2O5on sub-micron aerosol, although
this cannot be justified during TexAQS 2006 where submicron aerosol was mostly
inorganic. The nitrate radical budget—calculated from the measurements—indicated
that DMS, alkenes and RO2were the most important NO3sinks in the Open Ocean
and in the Gulf Coast, while alkenes and biogenics were the most important NO3
sinks in the polluted regions. The peroxy radicals accounted, on average, for 12–28%
of the total gas-phase NO3sinks in the clean and semi-polluted regions, consistent
with a previous study. In the polluted regions, the peroxy radicals accounted, on
average, for 1–2% of the total gas-phase NO3sinks. The quantification of the
NO3–RO2interactions was possible because the two parameters were measured
simultaneously on the R/V Brown during the TexAQS 2006 cruise.
Acknowledgements We thank the crew of the NOAA R/V Brown for their contribution to the
field work. Thanks to M.J. Pilling and C.J. Martin for assistance in setting up the MCM model and
to A. Bonzanini for help in assembling the appendix. This work was funded in part by NOAA’s Air
Quality and Atmospheric Chemistry and Climate Programs and in part by the Texas Commission on
Environmental Quality (TCEQ).
Appendix: RO2precursors in MCM v3.1
JAtmosChem
JAtmosChem
JAtmosChem
JAtmosChem
JAtmosChem
JAtmosChem
References
Aldener, M., Brown, S.S., Stark, H., Williams, E.J., Lerner, B.M., Kuster, W.C., Goldan, P.D., Quinn,
P.K., Bates, T.S., Fehsenfeld, F.C., Ravishankara, A.R.: Reactivity and loss mechanisms of NO3
and N2O5in a polluted marine environment: results from in situ measurements during New Eng-
land Air Quality Study 2002. J. Geophys. Res. 111, D23S73 (2006). doi:10.1029/2006JD007252
Ambrose, J.L., Mao, H., Mayne, H.R., Stutz, J., Talbot, R., Sive, B.C.: Nighttime nitrate radi-
cal chemistry at Appledore Island, Maine during the 2004 International Consortium for At-
mospheric Research on Transport and Transformation. J. Geophys. Res. 112, D21302 (2007).
doi:10.1029/2007JD008756
Atkinson, R., Cox, R.A., Crowley, J.N., Hampson, R.F., Hynes, R.G., Jenkin, M.E., Kerr, J.A., Rossi,
M.J., Troe, J.: Summary of evaluated kinetic and photochemical data for atmospheric chemistry.
Tech. rep., IUPAC Subcommittee on Gas Kinetic Data Evaluation for Atmospheric Chemistry
(2006). http://www.iupac-kinetic.ch.cam.ac.uk/
Bates, T.S., Quinn, P.K., Coffman, D., Schulz, K., Covert, D.S., Johnson, J.E., Williams, E.J., Lerner,
B.M., Angevine, W.M., Tucker, S.C., Brewer, W.A., Stohl, A.: Boundary layer aerosol chemistry
during TexAQS/GoMACCS 2006; insights into aerosol sources and transformation processes. J.
Geophys. Res. 113, D00F01 (2008). doi:10.1029/2008JD010023
Brown, S.S., Ryerson, T.B., Wollny, A.G., Brock, C.A., Peltier, R., Sullivan, A.P., Weber, R.J.,
Holloway, J.S., Dubé, W.P., Trainer, M., Meagher, J.F., Fehsenfeld, F.C., Ravishankara, A.R.:
Variability in nocturnal nitrogen oxide processing and its role in regional air quality. Science 311,
67–70 (2006)
Brown, S.S., de Gouw, J.A., Warneke, C., Ryerson, T.B., Dubé, W.P., Atlas, E., Weber, R.J.,
Peltier, R.E., Neuman, J.A., Roberts, J.M., Swanson, A., Flocke, F., McKeen, S.A., Brioude, J.,
Sommariva, R., Trainer, M., Fehsenfeld, F.C., Ravishankara, A.R.: Nocturnal isoprene oxidation
over the Northeast United States in summer and its impact on reactive nitrogen partition-
ing and secondary organic aerosol. Atmos. Chem. Phys. 9(9), 3027–3042 (2009). www.atmos-
chem-phys.net/9/3027/2009/
Brown, S.S., Dubé, W.P., Peischl, J., Ryerson, T.B., Atlas, E., Warneke, C., de Gouw, J.A.,
te Lintel Hekkert, S., Brock, C.A., Flocke, F., Trainer, M., Parrish, D.D., Feshenfeld, F.C.,
Ravishankara, A.R.: Budgets for nocturnal VOC oxidation by nitrate radicals aloft during the
2006 Texas Air Quality Study. J. Geophys. Res. 116, D24305 (2011). doi:10.1029/2011JD016544
Canosa-Mas, C.E., King, M.D., Lopez, R., Percival, C.J., Wayne, R.P., Shallcross, D.E., Pyle, J.A.,
Daële, V.: Is the reaction between CH3C(O)O2and NO3important in the night-time tro-
posphere? Faraday Trans. 92, 2211–2222 (1996). doi:10.1039/FT9969202211
Carslaw, N., Carpenter, L.J., Plane, J.M.C., Allan, B.J., Burgess, R.A., Clemitshaw, K.C., Coe, H.,
Penkett, S.A.: Simultaneous observations of nitrate and peroxy radicals in the marine boundary
layer. J. Geophys. Res. 102(D15), 18,917–18,933 (1997)
Carslaw, N., Creasey, D.J., Heard, D.E., Lewis, A.C., McQuaid, J.B., Pilling, M.J., Monks, P.S.,
Bandy, B.J., Penkett, S.A.: Modeling OH, HO2,andRO
2radicals in the marine bound-
ary layer—1. Model construction and comparison with field measurements. J. Geophys. Res.
104(D23), 30,241–30,255 (1999)
Carslaw, N., Creasey, D.J., Heard, D.E., Jacobs, P.J., Lee, J.D., Lewis, A.C., McQuaid, J.B., Pilling,
M.J., Bauguitte, S., Penkett, S.A., Monks, P.S., Salisbury, G.: Eastern Atlantic Spring Exper-
iment 1997 (EASE97)—2. Comparisons of model concentrations of OH, HO2,andRO
2with
measurements. J. Geophys. Res. 107(D14), 4190 (2002). doi:10.1029/2001JD001568
Dubé, W.P., Brown, S.S., Osthoff, H.D., Nunley, M.R., Ciciora, S.J., Paris, M.W., McLaughlin, R.J.,
Ravishankara, A.R.: Aircraft instrument for simultaneous, in situ measurement of NO3and
N2O5via pulsed cavity ring-down spectroscopy. Rev. Sci. Instrum. 77, 034101 (2006)
Emmerson, K.M., Carslaw, N., Carslaw, D.C., Lee, J.D., McFiggans, G., Bloss, W.J., Gravestock,
T., Heard, D.E., Hopkins, J., Ingham, T, Pilling, M.J., Smith, S.C., Jacob, M, Monks, P.S.: Free
radical modelling studies during the UK TORCH campaign in summer 2003. Atmos. Chem.
Phys. 7, 167–181 (2007)
Fleming, Z.L., Monks, P.S., Rickard, A.R., Heard, D.E., Bloss, W.J., Seakins, P.W., Still, T.J.,
Sommariva, R., Pilling, M.J., Morgan, R., Green, T.J., Brough, N., Mills, G.P., Penkett, S.A.,
Lewis, A.C., Lee, J.D., Saiz-Lopez, A., Plane, J.M.C.: Peroxy radical chemistry and the control
of ozone photochemistry at Mace Head, Ireland during the summer of 2002. Atmos. Chem. Phys.
6(8), 2193–2214 (2006) www.atmos-chem-phys.net/6/2193/2006/
JAtmosChem
Fuchs, H., Bohn, B., Hofzumahaus, A., Holland, F., Lu, K.D., Nehr, S., Rohrer, F., Wahner, A.: De-
tection of HO2by laser-induced fluorescence: calibration and interferences from RO2radicals.
Atmos. Meas. Tech. 4, 1209–1225 (2011)
Fuchs, N.A., Sutugin, A.G.: Highly Dispersed Aerosols. Ann Arbor Science, Ann Arbor, MI, USA
(1970)
Geyer, A., Bächmann, K., Hofzumahaus, A., Holland, F., Konrad, S., Klüpfel, T., Pätz, H.W., Perner,
D., Mihelcic, D., Schäfer, H.J., Volz-Thomas, A., Platt, U.: Nighttime formation of peroxy
and hydroxyl radicals during the BERLIOZ campaign: observations and modeling studies. J.
Geophys. Res. 108(D4), 8249 (2003). doi:10.1029/2001JD000656
Gilman, J.B., Kuster, W.C., Goldan, P.D., Herndon, S.C., Zahniser, M.S., Tucker, S.C., Brewer,
W.A., Lerner, B.M., Williams, E.J., Harley, R.A., Fehsenfeld, F.C., Warneke, C., de Gouw, J.A.:
Measurements of volatile organic compounds during the 2006 TexAQS/GoMACCS campaign:
industrial influences, regional characteristics, and diurnal dependencies of the OH reactivity. J.
Geophys. Res. 114, D00F06 (2009). doi:10.1029/2008JD011525
Jenkin, M.E., Saunders, S.M., Pilling, M.J.: The tropospheric degradation of volatile organic com-
pounds: a protocol for mechanism development. Atmos. Environ. 31(1), 81–104 (1997)
Jenkin, M.E., Saunders, S.M., Wagner, V., Pilling, M.J.: Protocol for the development of the Master
Chemical Mechanism, MCM v3 (Part B): tropospheric degradation of aromatic volatile organic
compounds. Atmos. Chem. Phys. 3(1), 181–193 (2003)
Kirchner, F., Mayer-Figge, A., Zabel, F., Becker, K.H.: Thermal stability of peroxynitrates. Int. J.
Chem. Kinet. 31, 127–144 (1999)
Kuster, W.C., Jobson, B.T., Karl, T., Riemer, D., Apel, E., Goldan, P.D., Fehsenfeld, F.C.: Inter-
comparison of volatile organic carbon measurement techniques and data at La Porte during the
TexAQS2000 Air Quality Study. Environ. Sci. Technol. 38, 221–228 (2004)
Mentel, T.F., Bleilebens, D., Wahner, A.: A study of nighttime nitrogen oxide oxidation in a large
reaction chamber—the fate of NO2,N
2O5,HNO
3,andO
3at different humidities. Atmos.
Environ. 30(23), 4007–4020 (1996)
Mihelcic, D., Klemp, D., Müsgen, P., Pätz, H.W., Volz-Thomas, A.: Simultaneous measurements of
peroxy and nitrate radicals at Schauinsland. J. Atmos. Chem. 16, 313–335 (1993)
Monks, P.S.: Gas-phase radical chemistry in the troposphere. Chem. Soc. Rev. 34, 376–395 (2005)
Osthoff, H.D., Sommariva, R., Baynard, T., Pettersson, A., Williams, E.J., Lerner, B.M., Roberts,
J.M., Stark, H., Goldan, P.D., Kuster, W.C., Bates, T.S., Coffman, D., Ravishankara, A.R.,
Brown, S.S.: Observation of daytime N2O5in the marine boundary layer during New England
Air Quality Study-Intercontinental Transport and Chemical Transformation 2004. J. Geophys.
Res. 111, D23S14 (2006). doi:10.1029/2006JD007593
Osthoff, H.D., Roberts, J.M., Ravishankara, A.R., Williams, E.J., Lerner, B.M., Sommariva, R.,
Bates, T.S., Coffman, D., Quinn, P.K., Dibb, J.E., Stark, H., Burkholder, J.B., Talukdar, R.K.,
Meagher, J., Fehsenfeld, F.C., Brown, S.S.: High levels of nitryl chloride in the polluted subtrop-
ical marine boundary layer. Nature Geosci. 1, 324–328 (2008). doi:10.1038/ngeo177
Osthoff, H.D., Bates, T.S., Johnson, J.E., Kuster, W.C., Goldan, P.D., Sommariva, R., Williams,
E.J., Lerner, B.M., Warneke, C., de Gouw, J.A., Pettersson, A., Baynard, T., Meagher, J.F.,
Fehsenfeld, F.C., Ravishankara, A.R., Brown, S.S.: Regional variation of the dimethyl sulfide
oxidation mechanism in the summertime marine boundary layer in the Gulf of Maine. J. Geo-
phys. Res. 114, D07301 (2009). doi:10.1029/2008JD010990
Parrish, D.D., Allen, D.T., Bates, T.S., Estes, M., Fehsenfeld, F.C., Feingold, G., Ferrare, R.,
Hardesty, R.M., Meagher, J.F., Nielsen-Gammon, J.W., Pierce, R.B., Ryerson, T.B., Seinfeld,
J.H., Williams, E.J.: Overview of the Second Texas Air Quality Study (TexAQS II) and the Gulf
of Mexico Atmospheric Composition and Climate Study (GoMACCS). J. Geophys. Res. 114,
D00F13 (2009). doi:10.1029/2009JD011842
Roberts, J.M., Jobson, B.T., Kuster, W.C., Goldan, P.D., Murphy, P., Williams, E., Frost, G., Riemer,
D., Apel, E., Stroud, C., Wiedinmyer, C., Fehsenfeld, F.C.: An examination of the chemistry of
peroxycarboxylic nitric anhydrides and related volatile organic compounds during Texas Air
Quality Study 2000 using ground-based measurements. J. Geophys. Res. 108(D16), 4495 (2003).
doi:10.1029/2003JD003383
Salisbury, G., Rickard, A.R., Monks, P.S., Allan, B.J., Bauguitte, S., Penkett, S.A., Carslaw, N.,
Lewis, A.C., Creasey, D.J., Heard, D.E., Jacobs, P.J., Lee, J.D.: Production of peroxy radicals
at night via reactions of ozone and the nitrate radical in the marine boundary layer. J. Geophys.
Res. 106(D12), 12,669–12,687 (2001)
Saunders, S.M., Jenkin, M.E., Derwent, R.G., Pilling, M.J.: Protocol for the development of the
Master Chemical Mechanism, MCM v3 (Part A): tropospheric degradation of non-aromatic
volatile organic compounds. Atmos. Chem. Phys. 3(1), 161–180 (2003)
JAtmosChem
Seefeld, S., Kerr, J.A.: Kinetics of the reactions of propionylperoxy radicals with NO and NO2:per-
oxypropionyl nitrate formation under laboratory conditions related to the troposphere. Environ.
Sci. Technol. 31, 2949–2953 (1997)
Sommariva, R., Haggerstone, A.L., Carpenter, L.J., Carslaw, N., Creasey, D.J., Heard, D.E., Lee,
J.D., Lewis, A.C., Pilling, M.J., Zádor, J.: OH and HO2chemistry in clean marine air during
SOAPEX-2. Atmos. Chem. Phys. 4(3), 839–856 (2004). www.atmos-chem-phys.net/4/839/2004/
Sommariva, R., Bloss, W.J., Brough, N., Carslaw, N., Flynn, M., Haggerstone, A.L., Heard, D.E.,
Hopkins, J.R., Lee, J.D., Lewis, A.C., McFiggans, G., Monks, P.S., Penkett, S.A., Pilling, M.J.,
Plane, J.M.C., Read, K.A., Saiz-Lopez, A., Rickard, A.R., Williams, P.I.: OH and HO2chemistry
during NAMBLEX: roles of oxygenates, halogen oxides and heterogeneous uptake. Atmos.
Chem. Phys. 6(4), 1135–1153 (2006). www.atmos-chem-phys.net/6/1135/2006/
Sommariva, R., Trainer, M., de Gouw, J.A., Roberts, J.M., Warneke, C., Atlas, E., Flocke, F.,
Goldan, P.D., Kuster, W.C., Swanson, A.L., Fehsenfeld, F.C.: A study of organic nitrates forma-
tion in an urban plume using a Master Chemical Mechanism. Atmos. Environ. 42(23), 5771–5786
(2008). doi:10.1016/j.atmosenv.2007.12.031
Sommariva, R., Osthoff, H.D., Brown, S.S., Bates, T.S., Baynard, T., Coffman, D., de Gouw, J.A.,
Goldan, P.D., Kuster, W.C., Lerner, B.M., Stark, H., Warneke, C., Williams, E.J., Fehsenfeld,
F.C., Ravishankara, A.R., Trainer, M.: Radicals in the marine boundary layer during NEAQS
2004: a model study of day-time and night-time sources and sinks. Atmos. Chem. Phys. 9(9),
3075–3093 (2009). www.atmos-chem-phys.net/9/3075/2009/
Sommariva, R., Brown, S.S., Roberts, J.M., Brookes, D.M., Parker, A.E., Monks, P.S., Bates,
T.S., Bon, D., de Gouw, J.A., Frost, G.J., Gilman, J.B., Goldan, P.D., Herndon, S.C., Kuster,
W.C., Lerner, B.M., Osthoff, H.D., Tucker, S.C., Warneke, C., Williams, E.J., Zahniser, M.S.:
Ozone production in remote oceanic and industrial areas derived from ship based measure-
ments of peroxy radicals during TexAQS 2006. Atmos. Chem. Phys. 11(6), 2471–2485 (2011).
www.atmos-chem-phys.net/11/2471/2011/
Stutz, J., Wong, K.W., Lawrence, L., Ziemba, L., Flynn, J.H., Rappenglück, B., Lefer, B.: Nocturnal
NO3radical chemistry in Houston, TX. Atmos. Environ. 44(33), 4099–4106 (2010)
Taketani, F., Kanaya, Y., Akimoto, H.: Kinetics of heterogeneous reactions of HO2radical at
ambient concentration levels with (NH4)2SO4and NaCl aerosol particles. J. Phys. Chem., A
112(11), 2370–2377 (2008)
Taketani, F., Kanaya, Y., Akimoto, H.: Heterogeneous loss of HO2by KCl, synthetic sea salt and
natural seawater aerosol particles. Atmos. Environ. 43 1660–1665 (2009)
Thornton, J., Abbatt, J.P.D.: Measurements of HO2uptake to aqueous aerosol: mass ac-
commodation coefficients and net reactive loss. J. Geophys. Res. 110, D08309 (2005).
doi:10.1029/2004JD005402
Thornton, J.A., Jaeglé, L., McNeill, V.F.: Assessing known pathways for HO2loss in aqueous
atmospheric aerosols: regional and global impacts on tropospheric oxidants. J. Geophys. Res.
113, D05303 (2008). doi:10.1029/2007JD009236
Tucker, S.C., Banta, R.M., Langford, A.O., Senff, C.J., Brewer, W.A., Williams, E.J., Lerner, B.M.,
Osthoff, H., Hardesty, R.M.: Relationships of coastal nocturnal boundary layer winds and tur-
bulence to Houston ozone concentrations during TexAQS 2006. J. Geophys. Res. 115, D10304
(2010). doi:10.1029/2009JD013169
Vaughan, S., Canosa-Mas, C.E., Pfrang, C., Shallcross, D.E., Watson, L., Wayne, R.P.: Kinetic studies
of reactions of the nitrate radical (NO3) with peroxy radicals (RO2): an indirect source of OH at
night? Phys. Chem. Chem. Phys. 8, 3749–3760 (2006)
Wahner, A., Mentel, T.F., Sohn, M.: Gas-phase reaction of N2O5with water vapor: importance of
heterogeneous hydrolysis of N2O5and surface desorption of HNO3in a large teflon chamber.
Geophys. Res. Lett. 25(12):2169–2172 (1998)
Warneke, C., de Gouw, J.A., Negro, L.D., Brioude, J., McKeen, S., Stark, H., Kuster, W.C., Goldan,
P.D., Trainer, M., Fehsenfeld, F.C., Wiedinmyer, C., Guenther, A.B., Hansel, A., Wisthaler, A.,
Atlas, E., Holloway, J.S., Ryerson, T.B., Peischl, J., Huey, L.G., Hanks, A.T.C.: Biogenic emis-
sion measurement and inventories determination of biogenic emissions in the eastern United
States and Texas and comparison with biogenic emission inventories. J. Geophys. Res. 115,
D00F18 (2010). doi:10.1029/2009JD012445
Whalley, L.K., Furneaux, K.L., Goddard, A., Lee, J.D., Mahajan, A., Oetjen, H., Read, K.A.,
Kaaden, N., Carpenter, L.J., Lewis, A.C., Plane, J.M.C., Saltzman, E.S., Wiedensohler, A.,
Heard, D.E.: The chemistry of OH and HO2radicals in the boundary layer over the tropical
Atlantic Ocean. Atmos. Chem. Phys. 10, 1555–1576 (2010)