OH and HO 2 Chemistry in the urban atmosphere of New York City

Atmospheric Environment 37 (2003) 3639–3651

OH and HO2 Chemistry in the urban atmosphere ofNew York City

Xinrong Rena,*, Hartwig Hardera,b, Monica Martineza,b, Robert L. Leshera,Angelique Oligera, James B. Simpasa, William H. Brunea, James J. Schwabc,

Kenneth L. Demerjianc, Yi Hed, Xianliang Zhoud,e, Honglian Gaoe

aDepartment of Meteorology, Pennsylvania State University, University Park, PA 16802, USAbMax-Planck-Institut f .ur Chemie, D-55116 Mainz, Germany

cAtmospheric Sciences Research Center, University at Albany, State University of New York, Albany, NY 12203, USAdDepartment of Environmental Health and Toxicology, University at Albany, State University of New York, Albany, NY 12222, USA

eWadsworth Center, New York State Department of Health, Albany, NY 12201, USA

Received 6 March 2003; accepted 27 May 2003

Abstract

Observed hydroxyl (OH) and hydroperoxy (HO2) radicals, collectively called HOx, were compared with OH and HO2

calculated by a box model that used the regional atmospheric chemistry mechanism and was constrained to the

ancillary measurements during the PM2.5 Technology Assessment and Characterization Study-New York (PMTACS-

NY) summer 2001 intensive in New York City. The measurements are described in the companion paper, Ren et al.

(HOx concentrations and OH reactivity observations in New York City during PMTACS-NY2001, Atmospheric

Environment, this issue). This comparison enables an investigation of HOx chemistry in this polluted urban

atmosphere. For HO2, the observed concentrations and diurnal variation were usually well reproduced by the model

calculations, with an observed-to-modeled ratio of 1.24, on average, for day and night. For OH, the model was

generally able to match the measured concentrations during daytime with an observed-to-modeled ratio of about 1.10,

but the calculations significantly underestimated OH during nighttime. The budgets of HOx show that its production

was dominated by the photolysis of HONO, accounting forB56% of HOx production on average, during daytime due

to relatively high HONO concentrations, while nighttime HOx production was mainly from the O3 reactions with

alkenes. The OH reactivity measurements agree with the calculations to within 10% for both the composite diurnal

variation and individual days. Calculations indicate that the reactions of OH with NO2, hydrocarbons, CO, NO, and

carbonyls accounted for about 32%, 25%, 12%, 10% and 7% of total OH loss, respectively, in this urban area.

Modeled instantaneous O3 production from HO2 and RO2 reactions with NO was 1507100 ppbv day�1. O3 production

rates from measured HO2ðPðO3ÞHO2obs Þ was greater than modeled HO2ðPðO3Þ

HO2calc Þ at higher values of NO. Average daily

cumulative PðO3ÞHO2obs was B140 ppbv day�1, a factor of 1.5, greater than average daily PðO3Þ

HO2calc :

r 2003 Elsevier Ltd. All rights reserved.

Keywords: Hydroxyl and hydroperoxy radicals; Model comparison; Urban environment; HOx budgets; Ozone production

1. Introduction

Measurements of OH and HO2 were made during the

PM2.5 Technology Assessment and Characterization

Study-New York (PMTACS-NY) intensive field cam-

paign, which occurred during summer 2001 at Queens

College in New York City. Simultaneous measurements

of meteorological variables and other chemicals provide

an opportunity to calculate the OH and HO2 with a

model for comparison to observations. The HOx

measurements are discussed in a companion paper

ARTICLE IN PRESS

AE International – North America

*Corresponding author. Fax: +1-814-865-3663.

E-mail address: [email protected] (X. Ren).

1352-2310/03/$ - see front matter r 2003 Elsevier Ltd. All rights reserved.

doi:10.1016/S1352-2310(03)00459-X

(Ren et al., 2003). Since OH and HO2 have very short

photochemical lifetimes and play important roles in the

atmospheric oxidation processes, comparison of mea-

sured and modeled HOx enables the examination of the

oxidation mechanisms in this urban environment.

Several intensive field campaigns with HOx measure-

ments have occurred at ground-based sites in the past

decade (Mather et al., 1997; Mount and Williams, 1997;

Kanaya et al., 1999, 2001; Creasey et al., 2001; Faloona

et al., 2001; Martinez et al., 2000, 2002; Holland et al.,

2003). Most of these were carried out in relatively clean

environments. Relatively few intensive field campaigns

with HOx measurements have been made in polluted

urban environments. These studies include Hard et al.

(1984, 1986, 1992) in Portland, Oregon; Abram et al.

(2000) near London, UK; George et al. (1999) near Los

Angeles, CA; Martinez et al. (2000) in Nashville, TN;

Martinez et al. (2002) in Houston, TX, and Holland et al.

(2003) near Berlin, Germany. Only the last five studies

had simultaneous measurements of several chemicals

that have strong influence on OH and HO2 and are

necessary to truly constrain the model for comparisons

to measurements.

Two interesting observations have been found in both

Nashville and Houston studies (Martinez et al., 2000,

2002). First, HO2 was greater than expected in air

containing copious NO, thus ozone production did not

decrease as much as expected when NO was greater than

2 ppbv. Second, the OH budget analysis indicated that

additional OH sources were required to balance the OH

sinks. A question remains: Are these observed-to-

modeled differences a result of misunderstood chemis-

try, and thus occurring in all environments, or are they a

result of the unique atmospheric chemical composition

of only certain environments?

Measurements and modeling studies of nighttime

HOx are very limited. George et al. (1999) observed

persistent OH radical at B1� 106 cm�3 at the beginning

of the evening in Los Angeles. Kanaya et al. (1999)

obtained HO2 concentration of about 3 pptv

(B7.5� 107 cm�3) for most of the night in the marine

atmosphere at Oki Island, Japan. In their recent work,

both Creasey et al. (2001) and Faloona et al. (2001)

found that OH radical persisted into the evening hours

after dark in small but significant concentrations at

forested areas in northwestern Greece and in Michigan,

respectively. While Carslaw et al. (2001) found that the

model could reproduce the evening OH in Greece,

Faloona et al. (2001) found that the model could

reproduce the prodigious nighttime OH only if reactions

between O3 and unknown super-reactive alkenes were

included in the model. Elevated nighttime OH and HO2

were frequently observed in Nashville and Houston,

providing strong evidence for an O3+alkene source of

HOx (Martinez et al., 2000, 2002). However, even when

considering the reaction mechanisms involving O3 and

alkenes, the nighttime OH and the HO2/OH ratio could

not be explained. Thus, while significant nighttime HO2

has been observed by several research groups using

different techniques, significant nighttime OH has not.

In this paper, we present a model comparison with the

OH and HO2 observations during the PMTACS-NY

intensive campaign in summer 2001. The purpose of this

work is to substantiate our understanding of the

chemistry controlling HOx levels in this highly polluted

environment. Model comparisons with observations as

well as HOx budgets and photochemical O3 production

are discussed.

2. Ancillary measurements and model description

The site description of the PMTACS-NY 2001

campaign is given in the companion paper (Ren et al.,

2003). During the campaign, the following ancillary data

were continuously measured: O3, CO, SO2, NO, NO2,

CH4, formaldehyde (HCHO), nitrous acid (HONO),

nitric acid (HNO3), speciated nonmethane hydrocar-

bons (NMHCs), some carbonyls, temperature, pressure,

relative humidity (RH), wind speed, wind direction.

HONO and HNO3 were measured by aqueous-phase

scrubbing and HPLC analysis (Huang et al., 2002). The

other measurements were made by the New York State

Department of Environmental Conservation from one

of their New York City PAMS (Photochemical Assess-

ment Monitoring Station) sites (US Environmental

Protection Agency (USEPA), 1994). HCHO was

scrubbed from ambient air into water, then mixed with

the reagents 2,4-pentanedione and ammonium acetate to

form stoichiometrically the product derivative 3,5-

diacetyl 1,4-dihydrolutidine followed by fluorescence

detection. Carbonyls including aldehydes from C2 to C5

and several ketones (including acetone, 2-butanone and

methacrolein) were measured by HPLC technique.

Speciated NMHCs were measured on every third day

with a 24-h integration time at the Queens College site,

while they were measured on-line with 1-h integration

times on the campus of Queensborough Community

College (QCC), which is located B2.5mile to the east of

the Queens College site. By comparing the speciated

NMHC measurements at the Queens College site and at

the QCC site (averaging data for the whole day), a ratio

was calculated for each chemical in the NMHCs. These

ratios were then used to scale the QCC NMHC data for

the days when NMHCs were measured at the Queens

College site. The same ratios were assumed for the day

before and the day after, when NMHCs were not

measured at the Queens College site. The ratios were

generally close to 1.0 and were between 0.5 and 2.0 for

most chemicals, suggesting that both sites had similar

speciated NMHC concentrations. These scaled 1-h

NMHC measurements were then used in the model.

ARTICLE IN PRESSX. Ren et al. / Atmospheric Environment 37 (2003) 3639–36513640

The measured chemicals that are important to HOx,

along with the detection methods, time resolution and

detection limits, are given in Table 1.

The data were reported at different time intervals for

different measurements. O3, NO, NO2, CO, SO2, and

CH4 were averaged over 1min. HONO, HNO3 and

HCHO were average over 10min. In the analysis, all

these data were average to 10-min time intervals. The

HONO and HNO3 data from 13 to 17 July were not

available, so average diurnal variation of the whole

campaign was used for these 5 days. Other missing data

were obtained by linear interpolation. Since NMHCs

were only measured in 1-h integrated samples, their

concentrations were assumed to remain constant during

the 1-h period. Sensitivity studies show that it matters

little to HOx whether an interpolation or constant values

are used. Carbonyls except HCHO were measured only

every third day with a 3-h integrated sampling period

(8 samples for a day). Their concentrations were also

assumed to remain constant during the sample period.

This assumption introduced maximum uncertainties of

4% for modeled OH and 7% for modeled HO2. H2 was

fixed at 500 ppbv for the whole campaign. Hydrogen

peroxide and organic hydroperoxides were not measured

and were calculated by the model.

For model calculations, the Regional Atmospheric

Chemistry Mechanism (RACM) (Stockwell et al., 1997)

was used to calculate the OH and HO2 concentrations.

Kinetic rate coefficients were updated using the results

by DeMore et al. (1997) and Sander et al. (2000). The

measured VOCs were categorized as suggested by

Stockwell et al. (1997) (Table 2). Since a-pinene,d-limonene and other such alkenes were not measured

during this campaign, reactions relevant to these

chemicals were not included in the model. Reactions of

O3 with alkenes have been largely revised to represent

latest radical yields suggested by recent experiments

(Paulson et al., 1999; Rickard et al., 1999; Fenske et al.,

2000). Heterogeneous reactions of SO3 and N2O5 were

included in the model, although no dry deposition

processes were. Dry deposition can be ignored if OH

and HO2 are in steady state. The assumption of steady

state certainly applies to OH, which had a lifetime

shorter than 0.1 s; it should usually apply to HO2 except

in very clean, rarely encountered conditions, when its

lifetime becomes a few tens of seconds.

During the field campaign, photolysis frequencies

were not measured directly. However, a Yankee ultra-

violet multifilter rotating shadowband radiometer (UV-

MFRSR) was used to measure the global, direct, and

diffuse UV components of solar irradiance at seven

different wavelength bands near 299, 305, 311, 317, 324,

332, 367 nm, each with a 2 nm effective bandwidth. All

photolysis frequencies (J values) used in the model were

calculated from following expression (Jenkin et al.,

1997):

Ji ¼ Li cosðwÞMi exp½�Ni secðwÞ�; ð1Þ

where w is the solar zenith angle and Li; Mi; Ni are

chemical-specific parameters. These parameters were

derived via a fitting procedure for each chemical

originally developed by Hough (1988). As described by

Grenfell et al. (1999), the parameters in this model are

suitable for a higher-latitude region with a clear sky and

higher ozone column density (345 Dobson units).

During the campaign, the average total ozone column

density over the site was 332 Dobson units, as measured

by total ozone mapping spectrometer-earth probe

satellite sensor (data available at http: //toms.gsfc.

nasa.gov/teacher/ozone overhead.html). The solar spec-

tral irradiance over the range 300–320 nm would be

similar to that with 345 Dobson units of column ozone.

Therefore expression (1) and the parameters by Jenkin

et al. (1997) were used to calculate the frequencies for

the 23 photolysis reactions in the RACM model. The

midday J values agreed to within 5–10% with those

calculated with the TUV radiative transfer model

(http://www.acd.ucar.edu/TUV). The field experiment

log and the UV radiation profile measured on 3 July

indicated that 3 July was a clean day with a clear sky.

Thus the ratio of calculated photolysis frequency to UV

radiation intensity for each photolytic chemical on this

ARTICLE IN PRESS

Table 1

Measured chemicals important to HOx, along with the detection methods, time resolution and detection limits

Chemicals Techniques/instruments Time resolution Detection limits

O3 UV absorption, TECO 49 1min 2 ppbv

NO, NO2 Chemiluminescence, TECO 42 1min 0.5 ppbv

NOy Chemiluminescence, modified TECO 42 1min 0.05 ppbv

CO IR correlation, modified TECO 48 1min 0.1 ppmv

SO2 Pulsed UV fluorescence, TECO 43 1min 0.5 ppbv

HONO Scrubbed derivatization HPLC 10min 5 pptv

HNO3 Scrubbed derivatization HPLC 10min 20 pptv

Hydrocarbons C2–C12 Auto GC/MS 1h —

HCHO Fluorescence assay, Alpha Omega 10min 0.1 ppbv

Carbonyls HPLC 3h —

X. Ren et al. / Atmospheric Environment 37 (2003) 3639–3651 3641

day was calculated and used as reference to correct for

cloud effects. The photolysis frequencies on other days

were then scaled to the UV radiation intensities. Since

the UV-MFRSR detector has a cosine light dependence

with incidence angle greater than 80�, the uncertainties

in the photolysis frequencies were about a factor of 2–3

at solar zenith angles greater than 80�.

The model was run with the FACSIMILE software

(UES Software Inc.). The model used the 10-min

average values of O3, NO, NO2, CO, SO2, categorized

VOCs, water vapor, temperature, pressure, and calcu-

lated photolysis frequencies as constrained input para-

meters. It then calculated OH and HO2 as well as other

reactive chemicals such as NO3 and organic peroxy

radicals (RO2). The uncertainty in this RACM model

was estimated to be745% for OH and770% for HO2,

with 2s confidence. These uncertainties are based on the

combined uncertainties of the kinetic rate coefficients

(DeMore et al., 1997; Sander et al., 2000; Stockwell et al.,

1997), the measured chemical concentrations, and the

calculated photolysis frequencies, as estimated with a

Monte Carlo approach (as in Carslaw et al., 1999).

3. Model results and comparison with observations

In situ OH and HO2 were measured by the Penn State

laser-induced fluorescence (LIF) instrument. As de-

scribed in the companion paper, the detection limits

were about 3� 105 cm�3 for OH and 2.5� 106 cm�3

(0.1 pptv) for HO2, with 2s confidence and a 1-min

integration time. The measurements were made from the

end of June to the beginning of August 2001. The OH

reactivity was measured at the same time, as discussed in

the companion paper. The overlap of all necessary

ancillary measurements allowed model comparisons

to be made with measurements between 10 July and

2 August.

3.1. OH comparison

As shown in Fig. 1, the measurements and model

exhibit similar diurnal and day-to-day variations, with

maxima in the early afternoon and minima at night. The

agreement is good especially for the last part of the

campaign, although modeled values are often lower than

measured values in early evening and during nighttime.

Good agreement was obtained for daytime, which is

defined as period when O3 photolysis frequency J(O1D)

was greater than 0.1� 10�5 s�1, corresponding the

period from B7:00 to B18:30 (Eastern Daylight Time

(EDT)). The observed-to-modeled ratio was about 1.10

on average; the correlation coefficient, R2; was 0.65.

Unlike most of the study, the period between 16 and 19

July has measured OH substantially larger than modeled

OH. Because no instrument effects have been found that

can explain the discrepancy, more investigation will be

required to determine its cause. With a few exceptions,

the differences between the measurements and model are

less than the 2s uncertainties in the measurements and

model.

The median daytime OH and its variation are

captured by the model (Fig. 2(a)). Both the measure-

ments and the model show that the peak occurred about

2 h after local solar noon, which was at B12:55 EDT.

The delay was caused by the high morning NOx, which

acted as an effective sink for HOx by the reaction

ARTICLE IN PRESS

Table 2

Categorization of measured chemicals used in the RACM model

Chemicals Classified chemicals measured in the campaign

ETH Ethane

HC3 Ethyne, propane, i-butane, n-butane, 2,2-dimethylbutane, 2,3,4-trimethylpentane

HC5 n-Pentane, i-pentane, cyclopentane, n-hexane, 2,3-dimethylbutane, 2-methylpentane, 3-methylpentane,

2,3-dimethylbutane, 2,3-dimethylpentane, 2,4-dimethylpentane

HC8 Methylcyclopentane, cyclohexane, 2-methylhexane, 2-methylhexane, 3-methylhexane, n-heptane, methylcyclohexane,

2-methylheptane, 3-methylheptane, n-octane, n-nonane, n-decane, n-undecane, n-dodecane

ETE Ethene

OLT Propene, 1-butene, 1-pentene, 1-hexene

OLI trans-2-butene, cis-2-butene, trans-2-pentene, cis-2-pentene

ISO Isoprene

TOL Toluene, benzene, ethylbenzene, isopropylbenzene, n-propylbenzene

XYL o; m; p-xylene, m; p-diethylbenzene, 1,3,5-trimethylbenzene, 1,2,3-trimethylbenzene, 1,2,4-trimethylbenzene, styrene

HCHO Formaldehyde

ALD Acetaldehyde, propionaldehyde, n-butyraldehyde, valeraldehyde, hexanal, crotonaldehyde, benzaldehyde,

m-tolualdehyde

KET Acetone, 2-butanone

MACR Methacrolein, acrolein

X. Ren et al. / Atmospheric Environment 37 (2003) 3639–36513642

OH+NO2+M-HNO3+M, as was observed in pre-

vious studies (Creasey et al., 2001; Kanaya et al., 2001).

The ratio of measured-to-modeled OH depends on

both NO and the HOx production rate, P(HOx), as

shown in Fig. 3(a). The ratios are plotted as a function

of NO for four different ranges of P(HOx): o0.5� 107,

0.5–2� 107, 2–4� 107, >4� 107 cm�3 s�1. Each line

includes the bin-averaged values in the same log(NO)

intervals. For P(HOx)o0.5� 107 cm�3 s�1, the mea-

sured OH is B10 times larger than the modeled OH.

However, for increasingly larger P(HOx), the measured-

to-modeled ratio becomes essentially 1. For each range

in P(HOx), especially the highest, the measured-to-

modeled OH ratio is roughly independent of NO for NO

between 1 and 50 ppbv.

3.2. HO2 comparison

As shown in Fig. 4, the measured and modeled HO2

exhibit the similar diurnal and day-to-day profiles, just

as was found for OH. The model did overestimate HO2

during daytime on some days, such as 15 and 21 July.

Fig. 2(b) shows the median diurnal variations of

measured and calculated HO2. As can be seen from this

figure, measured and modeled HO2 agree very well

during midday, while the model trends to underestimate

the measurements in the morning when NO was usually

high. On average, the model slightly underestimated the

observation with a measured-to-modeled ratio of about

1.24 on average and a correlation coefficient, R2; of 0.71.Just as for OH, the maximum values of both measured

ARTICLE IN PRESS

Jul 10 Jul 11 Jul 12 Jul 13 Jul 14 Jul 15 Jul 16 Jul 170

5

10

15x 10

OH

(cm

3 )O

H (

cm3 )

OH

(cm

3 )

Jul 18 Jul 19 Jul 20 Jul 21 Jul 22 Jul 23 Jul 24 Jul 250

5

10

15x 106

6

Jul 26 Jul 27 Jul 28 Jul 29 Jul 30 Jul 31 Aug 1 Aug 20

5

10

15x 106

Date of 2001 (EDT)

Fig. 1. Comparison of the measured OH concentrations (open circles) and model calculations (solid line) from 10 July to 2 August

2001. The data shown are 10-min averages.

0.0

2.0x106

4.0x106

6.0x106

8.0x106

00:00 03:00 06:00 09:00 12:00 15:00 18:00 21:000.0

5.0x107

1.0x108

1.5x108

(b)

(a)

OH

(cm

-3)

00:00

HO

2 (cm

-3)

Time of Day (EDT)

Fig. 2. Median diurnal variations of measured and calculated

OH (a) and HO2 (b). The measurements are indicated by small

open circles and calculations are shown by solid line. All data

are in 10-min averages.

X. Ren et al. / Atmospheric Environment 37 (2003) 3639–3651 3643

and calculated HO2 appeared after midday, usually

between 14:00 and 15:00 (EDT).

The ratio of measured-to-modeled HO2 also depends

on both NO and P(HOx), as shown in Fig. 4(b). As in

the figure for the OH ratios, the HO2 ratios are plotted

as a function of NO for four different ranges of P(HOx).

Similar to the OH ratios, the measured-to-modeled HO2

ratio is higher when P(HOx) is lower. Unlike the OH

ratios, the measured-to-modeled HO2 ratios are close to

1 for NO less than 1 ppbv, but increase for increasing

NO for all P(HOx) ranges. This behavior has been seen

in other environments, including the upper troposphere

(Folkins et al., 1997; Tan et al., 2001a; Faloona et al.,

2000) and other urban areas (Martinez et al., 2002,

ARTICLE IN PRESS

10-1

100

101

102

10-1

100

101

102

NO (ppbv)

HO

2 ob

s./H

O2

mod

el r

atio

10-1

100

101

102

10-1

100

101

102

NO (ppbv)

OH

obs

./OH

mod

el r

atio

(a) (b)

Fig. 3. Ratios of observations to calculations for OH (a) and HO2 (b) as a function of measured NO. Data are binned by same

log (NO) intervals. The points are linked by four lines according to different production rates of HOx (P(HOx)). Ranges for P(HOx) are

indicated by different symbols: squares (o0.5� 107 cm�3 s�1, linked by dashed line); circles (0.5–2� 107 cm�3 s�1, linked by dotted

line); crosses (2–4� 107 cm�3 s�1, linked by dash-dotted line); diamonds (>4� 107 cm�3 s�1, linked by solid line). The horizontal

dotted lines show the ratios of unity.

Jul 10 Jul 11 Jul 12 Jul 13 Jul 14 Jul 15 Jul 16 Jul 170

1

2

3x 10

8

HO

2 (c

m3 )

HO

2 (c

m3 )

HO

2 (c

m3 )

Jul 18 Jul 19 Jul 20 Jul 21 Jul 22 Jul 23 Jul 24 Jul 250

1

2

3

4

5x 108

Jul 26 Jul 27 Jul 28 Jul 29 Jul 30 Jul 31 Aug 1 Aug 20

1

2

3x 108

Date of 2001 (EDT)

Fig. 4. Comparison of the measured HO2 (open circles) and modeled HO2 (solid line) for 10 July to 2 August 2001. All data are

resolved by 10min.

X. Ren et al. / Atmospheric Environment 37 (2003) 3639–36513644

2003). The reason for HO2 behavior is not clear, but it

may be due to additional unknown HOx sources, such as

VOCs, that increase as NO increases or to a short-

coming in the understanding of HOx–NOx chemistry

(Faloona et al., 2000).

3.3. HO2/OH ratio comparison

The HO2/OH ratio tests the understanding of the HOx

photochemistry because the cycling between OH and

HO2 is very fast and the photochemical equilibrium

among OH and HO2 is closely tied to the interconver-

sion of NO to NO2 in the troposphere (Stevens et al.,

1997). Both the measured and modeled HO2/OH ratios

decrease with increasing NO (Fig. 5). This decrease

occurs because NO shifts HOx into OH by reacting with

HO2. However, when NO is lower than 1 ppbv, the

modeled HO2/OH ratios are significantly higher than the

measured. The agreement of measured and modeled

HO2 to OH ratios is good when NO is around a few

ppbv. The slope of measured HO2/OH as a function of

NO is significantly less than the modeled slope. This

difference is consistent with measured OH being greater

than modeled OH at low NO, while measured HO2 is

much greater than modeled HO2 at high NO.

3.4. Nighttime HOx comparison

Recent studies show that in the planetary boundary

layer two oxidation pathways can contribute to night-

time HOx: (1) O3 can react with alkenes to produce a

significant amount of OH and HO2 (Donahue et al.,

1998; Kroll et al., 2001; Salisbury et al., 2001), and/or (2)

NO3 can produce HOx directly via reaction with HCHO

or indirectly after conversion of the RO2 that is initially

produced by VOCs+NO3 (Bey et al., 2001). These

processes become more important for the nighttime

HOx production because daytime HOx photolytic

sources vanish at night.

As can be seen from Fig. 4 and Fig. 2(b), the modeled

HO2 agrees reasonably well with the measurements

during nighttime, with a measured-to-modeled ratio of

1.28, which is within the combined uncertainties of

measured and modeled HO2. However, for nighttime

OH, the model significantly underpredicted the mea-

surements by a factor of 6 on average. This difference

indicates that the RACM mechanism fails to capture the

processes that create nighttime OH in this urban

environment.

The measured nighttime OH and HO2 have an O3-

dependence in this polluted atmosphere. The linear fits

for the nighttime observations of OH, HO2 as a function

of O3 are given by the equations:

OH ðpptvÞ ¼ 0:0034O3 ðppbvÞ þ 0:0004;

R2 ¼ 0:86; ð2Þ

HO2 ðpptvÞ ¼ 0:072O3 ðppbvÞ þ 0:088;

R2 ¼ 0:65: ð3Þ

These dependences are 24 and 15 times larger than

those observed in the laboratory during O3 interference

tests for OH and HO2, respectively (Ren et al., manu-

script in preparation). A similar dependence was

observed during the SOS campaign in Nashville, 1999

(Martinez et al., 2003), although relationship was

described by the expressions: OH (pptv)=0.0012 O3

(ppbv)+0.013 and HO2 (pptv)=0.18 O3 (ppbv)+2.2.

That these dependences are so much larger than those

observed in the laboratory provides evidence that the

measured OH was not from dark reactions of O3 that

produced OH inside the instrument, such as has been

seen by others (Holland et al., 1995, 2003). Other

interferences such as photolysis of HONO, HCHO,

H2O2 by laser beam were found to be insignificant (Ren

et al., 2003). Thus, we have confidence that the observed

nighttime HOx signals were from ambient HOx.

3.5. HOx Budgets

The model can be used to interpret the observed HOx

behavior and to determine the sources and sinks of OH

and HO2. Based on model calculations, the average

diurnal variations of OH and HO2 budgets for 10 July to

2 August are shown in Fig. 6. Fig. 6(a) and (b) show the

production and loss for OH, while Fig. 6(c) and (d) show

the production and loss for HO2.

Fig. 6(a) indicates that, apart from HO2+NO-OH+NO2, the most important OH source was daytime

photolysis of HONO. HONO formation rate from

ARTICLE IN PRESS

0.1 1 10 100

0.1

1

10

100 modeled HO2/OH measured HO2/OH

HO

2 to

OH

rat

io

NO (ppbv)

Fig. 5. Relationship between HO2/OH ratios and NO concen-

trations. Data are binned by the same log(NO) intervals. The

NO-dependences of measured (circles) and modeled (squares)

HO2/OH ratios have different slopes. The error bars indicate

the statistical variation (1s) of the ratios included in each point.

X. Ren et al. / Atmospheric Environment 37 (2003) 3639–3651 3645

reaction of OH and NO was subtracted from total

HONO photolysis rate to reflect the net OH production

from HONO photolysis. Even in the afternoon this

source was much greater than OH production from the

reaction O(1D)+H2O. HONO was relatively high in this

polluted environment through the whole campaign; it

was on average 1.4 ppbv in the morning and 0.4–

0.5 ppbv in the afternoon. This observation of OH

production is quite different from those obtained in

relatively clean environments (Kanaya et al., 2001;

Creasey et al., 2001; Tan et al., 2001b) and in other

polluted environments (George et al., 1999; Platt et al.,

2002; Martinez et al., 2002), where photolysis of O3

dominated primary OH production in the late morning

and afternoon. At night in New York City, O3 reactions

with alkenes were the dominant OH source after

HO2+NO-OH+NO2.

OH loss was mainly determined by it reactions with

NO2 and VOCs through the whole campaign. Reactions

with CO and NO were the next two most important OH

loss processes.

For HO2, the production processes in the order of

importance were RO2+NO reactions, OH reactions

with CO and HCHO, photolysis of HCHO, and O3

reactions with alkenes. The loss of HO2 was dominated

by its reaction with NO. Thus in Fig. 6(d), the two lines

of total HO2 loss and HO2 loss from reaction HO2+NO

almost overlap.

It is also useful to investigate the net HOx production

and loss, since the balance between HOx production and

loss indicates a good understanding of HOx sources and

sinks. The total HOx production includes the processes:

the UV-B O3 photolysis followed by the reaction of

O(1D) with H2O, the net HONO photolysis, the HCHO

photolysis (the radical-produced pathway), and the O3

reactions with alkenes. HOx loss includes the processes:

the OH reaction with NO2, and the reactions between

OH, HO2 and RO2. Calculated HOx production was

dominated by photolysis of HONO during daytime, but

was mainly from O3 reactions with alkenes during

nighttime. Photolysis of HCHO was not as important to

the HOx production as the photolysis of HONO because

the HCHO concentrations were not particularly high

(only about 1.9 ppbv on average during daytime). On

average, the daily HOx production rate was

34.7 ppbv day�1, in which the photolysis of HONO,

photolysis of O3, O3 reactions with alkenes, and

photolysis of HCHO accounted for about 56%, 13%,

10%, and 8%, respectively. For HOx loss, the clearly

dominant process was the OH reaction with NO2, which

was 99.3% of the HOx loss. The self-reactions between

OH, HO2, RO2 contributed little to HOx loss, less than

1% for day and night.

3.6. OH reactivity

The OH reactivity, measured by a instrument named

total OH loss rate measurement (TOHLM) (Ren et al.,

2003), can be compared to calculated OH reactivity,

which was found by taking the sum of the products of

the measured reactants with OH and their rate

coefficients for reactions with OH. The measured and

ARTICLE IN PRESS

0:00 6:00 12:00 18:00 0:0010

4

105

106

107

108

P(O

H)

(cm

-3 s

-1)

P(OH) sumHO2+NOHONO+hvO(1D)+H2OO3+alkenes

0:00 6:00 12:00 18:00 0:00105

106

107

108

L(OH) sumOH+VOCsOH+NO2OH+COOH+NO

0:00 6:00 12:00 18:00 0:00104

105

106

107

108

P(O

H )

(cm

-3 s

-1)

2

Time of day (EDT)

P(HO2) sumRO2+NOOH+COOH+HCHOHCHO+hvO3+alkenes

0:00 6:00 12:00 18:00 0:00106

107

108

L(H

O2)

(cm

-3 s

-1)

L(H

O)

(cm

-3 s

-1)

Time of day (EDT)

L(HO2) sumHO2+NO

(a)

(c)

(b)

(d)

Fig. 6. The modeled average diurnal variations of the production and loss for OH and HO2: (a) OH production, (b) OH loss, (c) HO2

production, and (d) HO2 loss. Different symbols in each plot indicate different sources and sinks as shown in the legends.

X. Ren et al. / Atmospheric Environment 37 (2003) 3639–36513646

calculated OH reactivity agreed to within 10% (Fig. 7),

which is better than we should expect when we consider

the uncertainties associated both with the measurement

and the calculation. This agreement is in light of the

measurement of large number of previously unmeasured

hydrocarbons in an urban environment by Lewis et al.

(2000) and the greater measured-than-calculated OH

reactivity observed in Nashville (Kovacs et al., 2003).

Maximum OH reactivity was found in the morning due

to high chemical concentrations during the morning

rush hour. OH lifetime was about 0.05 s in this polluted

environment. As shown in Fig. 7, the calculations are a

little lower than the measurements, but are within the

uncertainty of the measurements of B715% (Kovacs

and Brune, 2001). Calculations also indicate that the

reactions of OH with NO2, hydrocarbons, CO, NO, and

carbonyls account for about 32%, 25%, 12%, 10% and

7%, respectively, of the average total OH reactivity.

Because OH has such a short lifetime, OH production

and loss should always come into balance in less than a

second. Since the OH loss is the product of the OH

reactivity and the OH concentration, our measurements

of OH and OH reactivity give a measurement of the OH

loss. A comparison of the calculated OH production

with the OH loss then tests for the presence of additional

OH sources. The OH production was calculated as the

sum of the known processes: the HO2 reaction with NO;

the UV-B O3 photolysis followed by reaction of O(1D)

with H2O; the HONO photolysis; and the O3 reactions

with alkenes. This OH production is slightly different

from that in the HOx budget calculation because

measured and not modeled HO2 was used. The

measured OH loss was similar to or greater than the

OH production except in the morning from 6:00am to

noon (Fig. 8).

During nighttime, the total OH loss and OH

production rates were, on average, 1.33� 107 and

4.5� 106 cm�3 s�1, respectively. To balance the OH

production and loss rates, an additional OH production

rate of 8.8� 106 cm�3 s�1 is required. As mentioned

above, calculated nighttime OH came primarily from O3

reactions with alkenes. This missing OH production rate

would be possible from the ozonolysis of reactive

biogenic terpenes such as a-pinene, b-pinene, d-limo-nene, which were not measured during this campaign.

The ozonolysis of these unsaturated terpenes is capable

of significant OH yields. In a forest area, Faloona et al.

(2001) reproduced the elevated OH levels during night-

time in their model by including the ozonolysis of

extremely reactive biogenic terpenoids. Another possi-

bility is the reactions of unmeasured anthropogenic

chemicals with O3 or NO3.

Because the measured and calculated OH reactivity

agree so well at night, the OH loss is greater than the

OH production because of the larger-than-expected

measured OH. However, if this OH is due to reactions

between O3 and unmeasured alkenes, we would expect

the measured OH reactivity to be greater than the

calculated OH reactivity. A possible solution to this

discrepancy is that the ratio of OH production by ozone

and alkenes to the OH loss due to all resulting VOCs is

greater than laboratory measurements suggest.

Between 6:00 and 12:00 EDT, the average calculated

OH production exceeded the average measured OH loss

ARTICLE IN PRESS

0:00 6:00 12:00 18:00 0:000

10

20

30

40

50

60

Time of day (EDT)

OH

rea

ctiv

ity (

s-1 )

Fig. 7. Diurnal variations of measured OH reactivity from

TOHLM (solid squares) and calculated OH reactivity (circles),

calculated for 1-h intervals. The gray points show all measured

OH reactivity. Error bars on squares indicate the absolute

uncertainties of the hourly averages.

0:00 6:00 12:00 18:00 0:00-5

0

5

10

15

20

Time of day (EDT)

P(O

H),

L(O

H)

(10

7 cm

-3 s

-1)

Fig. 8. Diurnal variations of measured OH production and loss

rates (in cm�3 s�1). The gray points are the measured OH loss

rates determined from the product of measured OH reactivity

and measured OH concentrations. One-hour median values for

the measured OH loss rate (solid squares) and calculated OH

production (circles) are similar, but have noticeable differences.

Error bars on squares indicate the absolute uncertainties of the

hourly average measured values.

X. Ren et al. / Atmospheric Environment 37 (2003) 3639–3651 3647

by (1–4)� 107 cm�3 s�1. OH production exceeding OH

loss is difficult to understand because the OH loss was

measured directly. One possibility is that the HO2

measurement was too high in the morning when the

NO was high, since the reaction HO2+NO dominated

OH production. However, no laboratory or field tests

have shown a problem with the HO2 measurement at

higher NO. A second and more likely possibility is that

we have under-corrected the OH decays in the OH

reactivity measurements for the reaction HO2+NO-OH+NO2 (Ren et al., 2003). In this case, the OH

reactivity was actually higher than was measured, and

the increase in OH loss rate will bring the production

and loss rates closer. Accurate correction of the OH

decays for NO requires accurate measurements of NO

and HO2/OH. While NO measurements were frequent in

New York City, measurements of HO2/OH inside

TOHLM were not, especially at morning rush hour.

The difference between the OH production and the OH

loss during morning rush hour correlated well with

ambient NO both in Nashville study (Martinez et al.,

2003) and in this study, indicating that the OH reactivity

may indeed be under-corrected for high NO.

3.7. Instantaneous O3 production

The instantaneous photochemical O3 production

during daytime can be calculated from expression:

PðO3Þ ¼ RðHO2 þNOÞ þX

RiðRO2i þNOÞ: ð4Þ

In New York City, modeled P(O3) peaked in the mid-

afternoon (Fig. 9). The average value reached

B20 ppbv h�1, but on a few individual days, was as

high as 50 ppbv h�1. This diurnal behavior is quite

different from the diurnal profile of O3 production rate

from measured HO2 reaction with NO, designated as

PðO3ÞHO2obs (Fig. 9), and from other urban areas in which

P(O3) peaks in mid-morning (Martinez et al., 2003;

Frost et al., 2002). The measured O3 followed the P(O3)

very well, with a peak that was about 1 h behind that of

P(O3). Based on the model calculations, the daily O3

production was 1507100 ppbv day�1, in which about

90780 ppbv day�1 was attributed to O3 production

from modeled HO2 reaction with NO, designated as

PðO3ÞHO2calc ; while P(O3) from RO2 was about 40% of total

P(O3) from HO2 and RO2.

Ozone production, like HO2, depends on NO and

P(HOx). O3 production is expected to become maximum

when NO is 1–2 ppbv, just as OH does and decreases as

NO increases. Higher P(HOx) leads to higher O3

production. This observed behavior agrees qualitatively

with expectations (Trainer et al., 2000). However,

because measured HO2 does not decrease as much as

expected as a function of NO (Fig. 3b), O3 production

determined from measured HO2, designated as

PðO3ÞHO2obs ; does not decrease as much as expected

(Fig. 10a). The PðO3ÞHO2obs to PðO3Þ

HO2calc ratio increases as

well (Fig. 10c). The cumulative difference between

PðO3ÞHO2obs and PðO3Þ

HO2calc results in an ozone difference

of 50 ppbv day�1. Similar results were observed in the

SOS99 Nashville study (Martinez et al., 2003). If we

assume that PðO3ÞHO2obs is 60% of total ozone production,

as in the model, then PðO3ÞHO2obs suggests that the actual

total ozone production was 230 ppbv day�1, a factor of

1.5 times the modeled P(O3). This difference between

ARTICLE IN PRESS

0:00 6:00 12:00 18:00 0:000

10

20

30

40

50

60

Time of day (EDT)

P(O

) (

ppbv

hr-

1 )3

0

10

20

30

40

50

60

O3

(pp

bv)

Fig. 9. Composite modeled diurnal variation of daytime total

O3 production, P(O3), calculated from modeled RO2 and HO2.

The gray small dots are all 10-min average values. The thick

gray line shows the 30-min binned average P(O3). Squares are

the O3 production from modeled HO2, PðO3ÞHO2calc and solid

diamonds are the O3 production from measured HO2,

PðO3ÞHO2obs : The open circles linked by a solid line are the

measured O3 (in ppbv) with 30-min averages.

1

10

1

10

1 10 100

1

10

(a)

(b)

(c)

NO (ppbv)

Fig. 10. (a) Ozone production from measured HO2, PðO3ÞHO2obs ;

(b) ozone production from modeled HO2, PðO3ÞHO2calc ; and

(c) ratio of PðO3ÞHO2obs to PðO3Þ

HO2calc vs. NO and P(HOx) for

all observations. Data are binned by log(NO). Ranges for

P(HOx) are indicated by different symbols: triangles

(>4� 107 cm�3 s�1); circles (1.5�4� 107 cm�3 s�1); squares

(o1.5� 107 cm�3 s�1). Gray dots are all 10-min data.

X. Ren et al. / Atmospheric Environment 37 (2003) 3639–36513648

measured and modeled P(O3) may be difficult to detect

in the ozone measurements.

4. Summary

The RACM model used here is able to simulate the

measured HO2, measured daytime OH, and measured

OH reactivity to within the 1s measurement uncertainty

for New York City in summer 2001. The measured

concentrations and diurnal variation of HO2 were

usually well reproduced by the model calculations for

day and night. For OH, the model could generally

match the measured concentrations during daytime. The

ratios of calculations to measurements for both OH and

HO2 are close to 1 when HOx production was high.

Agreement is less good when either P(HOx) is low or for

HO2, when NO is high. Measured and modeled HO2/

OH ratios agree reasonably when NO was around a few

ppbv.

The model could not reproduce the nighttime OH

levels, with modeled OH being significantly less than

measured OH. Several possibilities might explain this

difference. (1) A missing chemical or reaction mechan-

ism may convert HO2 to OH with a reduction potential

similar to NO (Tan et al., 2001b). (2) One or more of the

ancillary measurements may have a problem. (3)

Processes important to OH production are not included

in the current model.

The budget analysis shows that photolysis of HONO

dominated the HOx production during daytime. This

HONO dominance is a quite different result from HOx

sources in similar environments and the clean atmo-

sphere. At night, the O3 reaction with alkenes was the

main source of HOx. The OH lifetime was short in this

polluted area, B50ms on average. Comparison of total

OH production and loss rates suggests that

8.8� 106 cm�3 s�1 of OH production is required to

balance the measured OH loss; this additional produc-

tion might come from the ozonolysis of unmeasured

VOCs. Instantaneous O3 production is determined by

the concentrations of NOx and the production of HOx.

PðO3ÞHO2obs did not decrease as much as expected when

NO was greater than 2–3 ppbv while PðO3ÞHO2calc did,

resulting in 50 ppbv day�1 cumulative difference be-

tween PðO3ÞHO2obs and PðO3Þ

HO2calc :

This comparison between measured and modeled

HOx and OH reactivity is encouraging. However, the

discrepancies found here and in other environments, as

well as the differences among various instruments and

field studies, do point to the need for additional

investigation of HOx sources, the rapid cycling between

OH and HO2, and the instrument behavior. These

investigations will become more essential as more and

more polluted environments are explored and the

uncertainties of HOx-measuring instruments are reduced

by comparisons and technical improvements.

Acknowledgements

The authors thank all other participants in the

PMTACS-NY2001 field campaign for use of their data

in the model and Kenneth Demerjian for asking us to

participate in PMTACS-NY2001. Terry Shirley and

Jennifer Adams, who were supported by NSF Research

Experience for Undergraduate grants, did a terrific job

in the field measurements. Two anonymous reviewers

are acknowledged for providing insightful comments

and suggestions. This work was supported by NSF

(ATM-9974335 and ATM-0209972), the New York

State Energy Research and Development Authority

(NYSERDA) (contract #4918ERTERES99), the US

Environmental Protection Agency (EPA) (cooperative

agreement #R828060010), and New York State Depart-

ment of Environmental Conservation (NYS DEC)

(contract #C004210). Although the research described

in this article has been funded in part by the US

Environmental Protection Agency, it has not been

subjected to the Agency’s required peer and policy

review and therefore does not necessary reflect the views

of the Agency and no official endorsement should be

inferred.

References

Abram, J., Creasey, D.J., Heard, D.E., Lee, J.D., Pilling, M.J.,

2000. Hydroxyl radical and ozone measurements in England

during the solar eclipse of 11 August 1999. Geophysical

Research Letters 27, 3437–3440.

Bey, I., Aumont, B., Toupance, G., 2001. A modeling study of

the nighttime radical chemistry in the lower continental

troposphere 1. development of a detailed chemical mechan-

ism including nighttime chemistry. Journal of Geophysical

Research 106, 9959–9990.

Carslaw, N., Jacoba, P.J., Pilling, M.J., 1999. Modeling OH,

HO2, and RO2 radicals in the marine boundary layer, 2.

Mechanism reduction and uncertainty analysis. Journal of

Geophysical Research 104, 30257–30273.

Carslaw, N., Creasey, D.J., Harrison, D., Heard, D.E., Hunter,

M.C., Jacobs, P.J., Jenkin, M.E., Lee, J.D., Lewis, A.C.,

Pilling, M.J., Saunders, S.M., Seakins, P.W., 2001. OH and

HO2 radical chemistry in a forested region of northwestern

Greece. Atmospheric Environment 35, 4725–4737.

Creasey, D.J., Heard, D.E., Lee, J.D., 2001. OH and HO2

measurements in a forested region of northwestern Greece.

Atmospheric Environment 35, 4713–4724.

DeMore, W.B., Sander, S.P., Golden, D.M., Hampson, R.F.,

Kurylo, M.J., Howard, C.J., Ravishankara, A.R., Kolb,

C.E., Molina, M.J., 1997. Chemical kinetics and photo-

chemical data for use in stratospheric modeling, evaluation

ARTICLE IN PRESSX. Ren et al. / Atmospheric Environment 37 (2003) 3639–3651 3649

no. 12. JPL Publication 97-4, NASA Jet Propulsion

Laboratory, Pasadena, CA.

Donahue, N.M., Kroll, J.H., Anderson, J.G., 1998. Direct

observation of OH production from the ozonolysis of

olefins. Geophysical Research Letters 25, 59–62.

Faloona, I., Tan, D., Brune, W.H., Jaegl!e, L., Jacob, D.J.,

Kondo, Y., Koike, M., Chatfield, R., Pueschel, R., Ferry,

G., Sachse, G., Vay, S., Anderson, B., Hannon, J.,

Fuelberg, H., 2000. Observations of HOx and its relation-

ship with NOx in the upper troposphere during SONEX.

Journal of Geophysical Research 105, 3771–3783.

Faloona, I., Tan, D., Brune, W., Hurst, J., Barket Jr., D.,

Couch, T.L., Shepson, P., Apel, E., Riemer, D., Thornber-

ry, T., Carroll, M.A., Sillman, S., Keeler, G.J., Sagady, J.,

Hooper, D., Paterson, K., 2001. Nighttime observations of

anomalously high levels of hydroxyl radicals above a

deciduous forest canopy. Journal of Geophysical Research

106, 24315–24333.

Fenske, J.D., Hasson, A.S., Paulson, S.E., Kuwata, K.T., Ho,

A., Houk, K.N., 2000. The pressure dependence of the OH

radical yield from ozone–alkene reactions. Journal of

Physical Chemistry A 104, 7821–7833.

Folkins, I., Wennberg, P.O., Hanisco, T.F., Anderson, J.G.,

Salawitch, R.J., 1997. OH, HO2, and NO in two biomass

burning plumes: sources of HOx and implications for

ozone production. Geophysical Research Letters 24,

3185–3188.

Frost, G.J., Williams, E., Hereid, D., Jobson, B.T., Kuster, W.,

Roberts, J., Trainer, M., Fehsenfeld, F.C., Martinez, M.,

Harder, H., Brune, W., Di Carlo, P., Hall, S., Shetter, R.,

Karl, T., Apel, E., Riemer, D., Geyer, A., Stutz, J.,

Baumann, K., 2002. Model-measurement comparison of

[OH], [HO2], and P(O3) in Houston. Abstract A21F-11,

AGU Fall Meeting, EOS Transactions.

George, L.A., Hard, T.M., O’Brien, R.J., 1999. Measurement

of free radicals OH and HO2 in Los Angeles smog. Journal

of Geophysical Research 104, 11643–11655.

Grenfell, J.L., Savage, N.H., Harrison, R.M., Penkett, S.A.,

Forberich, O., Comes, F.J., Clemitshaw, K.C., Burgess,

R.A., C!ardenas, L.M., Davison, B., McFadyen, G.G., 1999.

Tropospheric box-modelling and analytical studies of the

hydroxyl (OH) radical and related species: comparison with

observations. Journal of Atmospheric Chemistry 33,

183–214.

Hard, T.M., O’Brien, R.J., Chan, C.Y., Mehrabzadeh, A.A.,

1984. Tropospheric free radical determination by FAGE.

Environmental Science and Technology 18, 768–777.

Hard, T.M., Chan, C.Y., Mehrabzadeh, A.A., Pan, W.H.,

O’Brien, R.J., 1986. Diurnal cycle of tropospheric OH.

Nature 322, 617–620.

Hard, T.M., Chan, C.Y., Mehrabzadeh, A.A., O’Brien, R.J.,

1992. Diurnal HO2 cycles at clean air and urban sites in the

troposphere. Journal of Geophysical Research 97,

9785–9794.

Holland, F., Hessling, M., Hofzumahaus, A., 1995. In situ

measurement of tropospheric OH radicals by laser-induced

fluorescence—a description of the KFA instrument. Journal

of Atmospheric Science 52, 3393–3401.

Holland, F., Hofzumahaus, A., Sch.afer, J., Kraus, A., P.atz, H.,

2003. Measurements of OH and HO2 radical concentrations

and photolysis frequencies during BERLIOZ. Journal of

Geophysical Research 108, 8246, doi:10.1029/

2001JD001393.

Hough, A.M., 1988. The calculation of photolysis rates for use

in global tropospheric modelling studies. United Kingdom

Atomic Energy Authority (UKAEA), Harwell Report

AERE H 13259.

Huang, G., Zhou, X., Deng, G., Qiao, H., Civerolo, K., 2002.

Measurements of atmospheric nitrous and nitric acid.

Atmospheric Environment 36, 2225–2235.

Jenkin, M.E., Saunders, S.M., Pilling, M.J., 1997. The tropo-

spheric degradation of volatile organic compounds: a

protocol for mechanism development. Atmospheric Envir-

onment 31, 81–104.

Kanaya, Y., Sadanaga, Y., Matsumoto, J., Sharma, U.K.,

Hirokawa, J., Kajii, Y., Akimoto, H., 1999. Nighttime

observation of the HO2 radical by a LIF instrument at Oki

island, Japan, and its possible origins. Geophysical Re-

search Letters 26, 2179–2182.

Kanaya, Y., Sadanaga, Y., Nakamura, K., Akimoto, H., 2001.

Behavior of OH and HO2 radicals during the observations

at a remote island of Okinawa (ORION99) field campaign

1. Observation using a laser-induced fluorescence

instrument. Journal of Geophysical Research 106,

24197–24208.

Kovacs, T.A., Brune, W.H., 2001. Total OH loss rate

measurement. Journal of Atmospheric Chemistry 39,

105–122.

Kovacs, T.A., Brune, W.H., Harder, H., Martinez, M., Simpas,

J.B., Frost, G.J., Williams, E., Jobson, T., Stroud, C.,

Young, V., Fried, A., Wert, B., 2003. Direct measurements

of urban OH reactivity during Nashville SOS in summer

1999. Journal of Environmental Monitoring 5, 68–74,

doi:10.1039/b204339d.

Kroll, J.H., Clarke, J.S., Donahue, N.M., Anderson, J.G.,

Demerjian, K.L., 2001. Mechanism of HOx formation in the

gas-phase ozone–alkene reaction. 1. Direct pressure-depen-

dent measurements of prompt OH yields. Journal of

Physical Chemistry A 105, 1554–1560.

Lewis, A.C., Carslaw, N., Marriott, P.J., Kinghorn, R.M.,

Morrison, P., Lee, A.L., Bartle, K.D., Pilling, M.J., 2000. A

large pool of ozone-forming carbon compounds in urban

atmospheres. Nature 405, 778–781.

Martinez, M., Harder, H., Simpas, J.B., Brune, W.H.,

Williams, E.J., Hall, S.E., Shetter, R.E., 2000. OH and

HO variations during summertime Nashville southern

oxidant study in 1999. Abstract A71E–10, AGU Fall

Meeting, San Francisco, CA, EOS Transactions.

Martinez, M., Harder, H., Brune, W., Di Carlo, P., Williams,

E., Hereid, D., Jobson, T., Kuster, W., Roberts, J., Trainer,

M., Fehsenfeld, F.C., Hall, S., Shetter, R., Apel, E., Riemer,

D., Geyer, A., 2002. The behavior of the hydroxyl and

hydroperoxyl radicals during TexAQS2000. Abstract

A12D-0174, AGU Fall Meeting, San Francisco, CA, EOS

Transactions.

Martinez, M., Harder, H., Kovacs, T.A., Simpas, J.B., Bassis,

J., Lesher, R., Brune, W.H., Williams, E.J., Stroud, C.A.,

Frost, G., Hall, S.R., Shetter, R.E., Wert, B., Fried, A.,

Alicke, B., Stutz, J., 2003. OH and HO2 concentrations,

production and loss rates during the southern oxidant study

in Nashville, TN, summer 1999. Journal of Geophysical

Research, submitted for publication.

ARTICLE IN PRESSX. Ren et al. / Atmospheric Environment 37 (2003) 3639–36513650

Mather, J.H., Stevens, P.S., Brune, W.H., 1997. OH and HO2

measurements using laser-induced fluorescence. Journal of

Geophysical Research 102, 6427–6436.

Mount, G.H., Williams, E.J., 1997. An overview of the

tropospheric OH photochemistry experiment, Fritz Peak,

Idaho Hill, Colorado, fall 1993. Journal of Geophysical

Research 102, 6171–6186.

Paulson, S.E., Chung, M.Y., Hasson, A., 1999. OH radical

formation from the gas-phase reaction of ozone with

terminal alkenes and the relationship between structure

and mechanism. Journal of Physical Chemistry A 103,

8125–8138.

Platt, U., Alicke, B., Dubois, R., Geyer, A., Hofzumahaus, A.,

Holland, F., Martinez, M., Mihelcic, D., Klupfel, T.,

Lohrmann, B., Patz, W., Perner, D., Rohrer, F., Schafer,

J., Stutz, J., 2002. Free radicals and fast photochemistry

during BERLIOZ. Journal of Atmospheric Chemistry 42,

359–394.

Ren, X., Harder, H., Martinez, M., Lesher, R.L., Oliger, A.,

Shirley, T., Adams, J., Simpas, J.B., Brune, W.H., 2003.

HOx concentrations and OH reactivity observations in

New York City during PMTACS-NY2001, Atmospheric

Environment, this issue, X-ref: doi:10.1016/S1352-2310(03)

00460-6.

Ren, X., Harder, H., Martinez, M., Faloona, I., Tan, D.,

Lesher, R.L., Di Carlo, P., Simpas, J.B., Brune, W.H.

Interference testing for atmospheric HOx measurements by

laser-induced fluorescence, in preparation.

Rickard, A.R., Johnson, D., McGill, C.D., Marston, G., 1999.

OH yields in the gas-phase reactions of ozone with alkenes.

Journal of Physical Chemistry A 103, 7656–7664.

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., 2001.

The production of peroxy radicals at night via reactions of

ozone and the nitrate radical in the marine boundary layer.

Journal of Geophysical Research 106, 12669–12688.

Sander, S.P., Friedl, R.R., DeMore, W.B., Golden, D.M.,

Kurylo, M.J., Hampson, R.F., Huie, R.E., Moortgat, C.K.,

Ravishankara, A.R., Kolb, C.E., Molina, M.J., 2000.

Chemical kinetics and photochemical data for use in

stratospheric modeling. JPL Publication 00–3, NASA Jet

Propulsion Laboratory, Pasadena, CA.

Stevens, P.S., Mather, J.H., Brune, W.H., 1997. HO2/OH and

RO2/HO2 ratios during the Tropospheric OH photochem-

istry experiment: measurement and theory. Journal of

Geophysical Research 102, 6379–6391.

Stockwell, W.R., Kirchner, F., Kuhn, M., 1997. A new

mechanism for regional atmospheric chemistry modeling.

Journal of Geophysical Research 102, 25847–25879.

Tan, D., Faloona, I., Simpas, J.B., Brune, W., Olson, J.,

Crawford, J., Avery, M., Sachse, G., Vay, S., Sandholm, S.,

Guan, H.-W., Vaughn, T., Mastromarino, J., Heikes, B.,

Snow, J., Podolske, J., Singh, H., 2001a. OH and HO2 in the

tropical pacific: results from PEM-Tropics B. Journal of

Geophysical Research 106, 32667–32681.

Tan, D., Faloona, I., Simpas, J.B., Brune, W., Shepson, P.B.,

Couch, T.L., Sumner, A.L., Carroll, M.A., Thornberry, T.,

Apel, E., Riemer, D., Stockwell, W., 2001b. HOx budgets in

a deciduous forest: results from the prophet summer

campaign. Journal of Geophysical Research 106,

24407–24427.

Trainer, M., Parrish, D.D., Goladan, P.D., Roberts, J.,

Fehsenfeld, F.C., 2000. Review of observation-based

analysis of the regional factors influencing ozone concentra-

tions. Atmospheric Environment 34, 2045–2061.

US Environmental Protection Agency, 1994. Photochemical

Assessment Monitoring Stations (PAMS) Implementation

Manual. EPA-454/B-93/051. Research Triangle Park, NC,

USA.

ARTICLE IN PRESSX. Ren et al. / Atmospheric Environment 37 (2003) 3639–3651 3651