Nighttime isoprene trends at an urban forested site during the 1999

Nighttime isoprene trends at an urban forested site

during the 1999 Southern Oxidant Study

C. A. Stroud,1,2,3 J. M. Roberts,1,2 E. J. Williams,1,2 D. Hereid,1,2 W. M. Angevine,1,2

F. C. Fehsenfeld,1,2 A. Wisthaler,4 A. Hansel,4 M. Martinez-Harder,5 H. Harder,5

W. H. Brune,5 G. Hoenninger,6,7 J. Stutz,6 and A. B. White8

Received 19 June 2001; revised 31 December 2001; accepted 11 February 2002; published 24 August 2002.

[1] Measurements of isoprene and its oxidation products, methacrolein, methyl vinylketone and peroxymethacrylic nitric anhydride, were conducted between 13 June and 14July 1999, at the Cornelia Fort Airpark during the Nashville intensive of the SouthernOxidant Study. Trends in isoprene and its oxidation products showed marked variabilityfrom night-to-night. The reaction between isoprene and the nitrate radical was shown to beimportant to the chemical budget of isoprene and often caused rapid decay of isoprenemixing ratios in the evening. Trends in methacrolein, methyl vinyl ketone, andperoxymethacrylic nitric anhydride were steady during the evening isoprene decay period,consistent with their slow reaction rate with the nitrate radical. For cases when isoprenesustained and even increased in mixing ratio throughout the night, the observed isopreneoxidation rates via the hydroxyl radical, ozone, and the nitrate radical were all small.Sustained isoprene mixing ratios within the nocturnal boundary layer give a uniqueopportunity to capture hydroxyl radical photochemistry at sunrise as isoprene wasobserved to rapidly convert to its first stage oxidation products before vertical mixingsignificantly redistributed chemical species. The observed nighttime isoprene variability aturban, forested sites is related to a complex coupling between nighttime boundary layerdynamics and chemistry. INDEX TERMS: 0345 Atmospheric Composition and Structure: Pollution—

urban and regional (0305); 0365 Atmospheric Composition and Structure: Troposphere—composition and

chemistry; 0317 Atmospheric Composition and Structure: Chemical kinetic and photochemical properties;

KEYWORDS: urban air quality, pollution, isoprene, nitrate radical, oxidant, PTRMS

1. Introduction

[2] Understanding the contribution of biogenic emissionsof isoprene (ISOP) to local and regional air pollution iscritical to enforcing effective air quality control strategies inforested environments [Trainer et al., 1987; Chameideset al., 1988]. Several studies have attempted to quantifythe contribution of ISOP to ozone (O3) formation at forestedsites impacted by nitric oxide (NO) emissions and found,that for certain periods, ISOP photochemistry can dominatethe photochemical production of O3 [Williams et al., 1997;Roberts et al., 1998; Starn et al., 1999a; Roberts et al.,

2002]. The chemical removal of ISOP from the atmosphereis determined by the mixing ratios of various oxidants suchas the hydroxyl radical (OH), O3 and the nitrate radical(NO3). Kinetic and mechanistic studies have been per-formed to determine the rate at which ISOP reacts witheach of these oxidants [Carter and Atkinson, 1996; Stevenset al., 1999]. For typical daytime oxidant levels in an urbanenvironment (0.3 pptv OH, 60 ppbv O3 and 0.5 pptv NO3),ISOP chemical lifetimes are 0.4 hr, 14 hrs, and 30 hrs,respectively. Thus, during the daytime, OH-initiated oxida-tion dominates the chemical removal of ISOP. Productstudies have identified that formaldehyde (HCHO), meth-acrolein (MACR) and methyl vinyl ketone (MVK) are themajor products of OH-initiated ISOP oxidation:

ISOPþOH ! 0:63 HCHOþ 0:32 MVK þ 0:23 MACRþ others

ð1Þ

MACR can further react in the atmosphere to yieldperoxymethacrylic nitric anhydride (MPAN) which, in turn,can thermally decompose or react with OH:

MACRþ OH ! 0:5 MA� RCO3 þ others ð2Þ

MA� RCO3 þ NO2 $ MPAN ð3Þ

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 107, NO. D16, 10.1029/2001JD000959, 2002

1NOAA Aeronomy Laboratory, Boulder, Colorado, USA.2Cooperative Institute for Research in the Environmental Sciences

(CIRES), University of Colorado, Boulder, Colorado, USA.3Now in the Atmospheric Chemistry Division, National Center for

Atmospheric Research, Boulder, Colorado, USA.4Institute for Ion Physics, University of Innsbruck, Innsbruck, Austria.5Department of Meteorology, Pennsylvania State University, University

Park, Pennsylvania, USA.6Department of Atmospheric Sciences, University of California, Los

Angeles, California, USA.7Now at Institut fuer Umweltphysik, University of Heidelberg,

Heidelberg, Germany.8Environmental and Technology Laboratory, Boulder, Colorado, USA.

Copyright 2002 by the American Geophysical Union.0148-0227/02/2001JD000959$09.00

ACH 7 - 1

MA� RCO3 þ NO ! others ð4Þ

MPANþ OH ! others ð5Þ

[3] However, at night, OH drops by more than an orderof magnitude, while NO3 has been observed to increase bymore than an order of magnitude [Platt et al., 1980; Geyeret al., 2001]. The result is that NO3-initiated oxidation candominate the nighttime chemical removal of ISOP. Themixing ratios of NO, NO2, O3, as well as others, are criticalin determining the nighttime abundance of NO3. The NO3

radical is produced directly from the reaction of O3 andNO2:

O3 þ NO2 ! NO3 þ O2 ð6Þ

The NO3 radical reacts rapidly with NO. However, O3 alsoreacts rapidly with NO and the abundance of O3 usuallymaintains NO at very low levels at night:

NO3 þ NO ! 2NO2 ð7Þ

O3 þ NO ! NO2 þ O2 ð8Þ

In the absence of significant NO emissions, the NO3 radicalis available to oxidize reactive hydrocarbons such as ISOP.However, large NO emissions concentrated in a shallownocturnal boundary layer (NBL) can titrate NO3, as well asO3 and other radicals (HO2, RO2, OH), resulting in a surfacelayer with low oxidizing capacity. The NO3 radical can alsoreact with NO2 to form dinitrogen pentoxide (N2O5). BothNO3 and N2O5 can be lost heterogeneously:

NO3 þ NO2 $ N2O5 ð9Þ

NO3 ! Surfaces ð10Þ

N2O5 ! Surfaces ð11Þ

Thus, the nighttime mixing ratios of NOx, are critical indetermining the oxidants available to chemically removeISOP. At low NOx mixing ratios (<0.01 ppbv), radical–radical reactions dominate the chemistry and oxidantmixing ratios (NO3, OH and O3) are low. At intermediateNOx mixing ratios (0.10–10 ppbv), peroxy radical + NOreactions dominate the chemistry and NO3, OH and O3

maximize in concentration. In extremely high NOx

environments (>100 ppbv), oxidant + NOx reactions (i.e.,OH + NO2 ! HNO3, 7 and 8) suppress the levels of NO3,OH and O3 which results in long ISOP chemical lifetimes.[4] An important question regarding nighttime chemistry

is what impact the ISOP reaction with NO3 has on the fateof both ISOP and NOx in forested environments impactedby emissions of NO. Product studies have identified the C5

nitrato carbonyl (RCHO-NO3) as the dominant products ofNO3-initiated ISOP removal:

ISOPþ NO3 ! 0:8RCHO� NO3þ others ð12Þ

Starn et al. [1998b] postulated that the C5 nitrato carbonylproducts may be a significant fraction of odd nitrogen(NOy) and represent an efficient mechanism for NOx

removal at night. Gas-phase removal of NO3 will competewith heterogeneous loss of NO3 at night [Dentener andCrutzen, 1993; Jacob, 2000; Martinez et al., 2000; Geyer etal., 2001]. Understanding the mechanisms of NOx andreactive carbon removal from the nighttime atmosphere isimportant because it impacts the ‘‘fuel’’ available for thenext day’s photochemical production of O3.[5] During the Nashville intensive of the 1999 Southern

Oxidant Study, we conducted measurements of ISOP and itsoxidation products, MACR, MVK, MPAN. The goal of thispaper is to identify which oxidants control the nighttimeremoval of ISOP at an urban, forested site, specificallyfocusing on the day-to-night and night-to-day transitions.To address this question we first examine time series data(ISOP, MACR, MPAN, O3, NOx) to identify nocturnalpatterns and then single out specific periods for moredetailed study (e.g., evenings when ISOP decayed rapidly,nights when ISOP increased, early mornings when ISOPdecayed rapidly). For specific case study periods, we reportsimultaneous measurements of OH, O3 and NO3 so that thecontribution of each oxidant to ISOP chemical removalcould be determined. Meteorological measurements of windspeed, wind direction, water vapor and boundary layerheights were also conducted to understand nighttime boun-dary layer dynamics.

2. Site Description and Experimental Methods

2.1. Site Description

[6] The chemical measurements were conducted between13 June and 14 July, 1999, in sub-urban Nashville, Tennes-see, at the Cornelia Fort Air Park (CFA). The site was in apasture field along the Cumberland River basin with thevegetation being primarily small stands of deciduous trees.The Air Park is approximately 8–9 km northeast of down-town and just west of the Cumberland River. The country-side surrounding Nashville is heavily forested withhardwood deciduous trees and to a less extent conifers.The gas-phase and general meteorological measurementswere made from two 10 m towers. The profiler instrumentswere operated several hundred meters north of the samplingtowers within the Air Park.

2.2. Measurement Descriptions

[7] Gas chromatographic (GC) measurements of ISOP,MACR, MVK and other C2-C5 volatile organic compounds(VOCs) were made hourly using methods described byStroud et al. [2001], and are only briefly summarized here.Measurements were made using an automated instrumentbased on solid sorbent preconcentration, cryofocussing, GCseparation and flame ionization detection. Calibration stand-ards in the parts per trillion (by volume) range were generatedby dynamic dilution of primary gas standards at the samplingsite. The primary standards were prepared gravimetrically byestablished techniques in Acculife-treated aluminum cylin-ders [Apel et al., 1998]. Detection limits for the C4 carbonylcompounds and ISOP were 0.05 ppbv and 0.03 ppbv,respectively, for a 400 scc air sample. GC measurements ofthe PAN compounds, including MPAN, were made every15 min with an electron capture detector as described byRoberts et al. [2002]. Detection limits were 5 pptv for a 2 cm3

sample loop injection. Proton-transfer reaction mass spec-

ACH 7 - 2 STROUD ET AL.: NIGHTTIME ISOPRENE CHEMISTRY AT AN URBAN SITE

trometry (PTR-MS) was also used to perform fast-responsemeasurements of ISOP, MACR+MVK and MPAN asdescribed by Hansel and Wisthaler [2000]. The GC andPTR-MS measurements of ISOP, MACR+MVK and MPANshared common standards throughout the study.[8] Fast response measurements of O3 were performed by

O3-NO chemiluminescence and calibrated based upon a UVabsorption instrument [Williams et al., 1998]. Measure-ments of NOx and NOy were also performed by NO-O3

chemiluminscence with a xenon lamp for NO2 conversionto NO and a molybdenum (Mo) converter for NOy speciesconversion to NO [Williams et al., 1998]. The OH data weremeasured in a low pressure cell atop the 10 m tower bylaser-induced fluorescence at 308 nm as described byMather et al. [1997]. OH detection limits were estimatedat 0.04 pptv for a 1 min. average. Measurements of NO3

were performed using differential optical absorption spec-trometry (DOAS) [Stutz and Platt, 1997] with severaldifferent long paths operated near CFA. The upper pathmeasurement was made between the Gaylord Building justeast of the Cumberland River and the State Laboratory onHeart Lane. The State Laboratory is located �4 km north-east of CFA resulting in a beam path length of 5232 m. Themiddle path was operated between the Gaylord Entertain-ment Building and CFA with a path length of 1367 m. Theupper and middle paths were 35–72 m and 2–35 m abovethe ground, respectively. Detection limits varied with atmos-pheric conditions in the range 1.2–13.9 pptv. To addresshow representative the DOAS sampled air masses were tothe tower measurements, a comparison of the DOAS andtower NO2 data was performed. A linear best-fit analysis ofthe tower NO2 measurements versus middle path DOASNO2 measurements for the evening hours between 17:00and 21:00 CST (Central Standard Time) yielded a slope of0.82 and a correlation coefficient R2 = 0.80, suggestingsimilar air masses were sampled with the middle path andthe tower. After 09:00 CST and throughout the night, stronginversions frequently formed at CFA resulting in conditionsof stratification in the NBL. Since the upper path measure-ments were acquired from an altitude considerably higherthan the tower and averaged over a longer path, there aredifficulties in comparing the upper path measurements withthe tower measurements; thus, care must be taken ininterpreting the upper path NO3 data. However, for theperiods of rapid ISOP decay near sunset, the upper pathDOAS and tower NO2 measurements were observed atsimilar mixing ratios suggesting that it is possible to usethe long path NO3 data to study ISOP oxidation in the earlyevening time period.[9] Wind profiler (915 MHz) and SODAR measurements

of boundary layer height were made in support ofthe chemical measurements. The wind profiler observesthe mixing height as a peak in radar reflectivity due to thehumidity gradient at the boundary layer top. This measure-ment has been extensively validated [Cohn and Angevine,2000; Grimsdell and Angevine, 1998; Angevine et al., 1994].

3. Results and Discussion

3.1. General

[10] Figures 1 and 2 show the chemical measurements ofISOP, MACR, MPAN, O3 and NOx along with ISOP loss

rates via reaction with OH, NO3 and O3 for selected periodsbetween 18–28 June 1999, and 2–13 July 1999. Theinteresting feature about the ISOP measurements is the lackof a consistent diurnal pattern. There are numerous nights(19 June, 23 June, 6 July, 9 July, 11–13 July) when ISOPmixing ratios were observed to decrease below instrumentaldetection limits (<30 pptv). However, there are also severalnights when ISOP maintained daytime mixing ratios andeven increased in mixing ratio throughout the night (22 June,25 June, 6 July, and 10 July). This is in contrast to otherstudies which have showed a consistent diurnal profile withincreased ISOP mixing ratios during the daylight hours and adecline in the evening after sunset [Montzka et al., 1993;Goldan et al., 1995; Biesenthal et al., 1997; Hurst et al.,2001]. These other studies were all performed at isolatedforested sites. Since ISOP emissions are a strong function ofphotosynthetically active irradiance and temperature [Feh-senfeld et al., 1992], one might expect the highest mixingratios during the daytime. However, there have also beenstudies reported where ISOP shows marked diurnal varia-bility, similar to the observations here, and were attributed toa complex combination of dynamics and chemistry (Youthsite during SOS 1995 as reported by Starn et al. [1998b]. Incomparing the Youth site with the CFA site, the one strikingsimilarity is the proximity of both studies to an urban centerand associated high levels of NOx. The Youth site isconsidered suburban (32 km southeast of downtown Nash-ville) and consistently encountered evening/nighttime NOx

levels between 1–10 ppbv. Figures 1 and 2 show that NOx

levels were even higher at the CFA site with severalevenings/nights with NOx over 50 ppbv (22 June; 5–7 July).These high NOx nights generally correspond with nightswhen ISOP reached its highest mixing ratios. As shown inFigures 1 and 2, these high NOx nights also correspond totimes when O3 levels were very low due to titration of the O3

by fresh urban emissions of NO in a shallow NBL. Theevening of 5 July and night of 6 July stand out as the ISOPmixing ratio increased throughout the night to �6 ppbv justbefore dawn. There are also periods in Figures 1 and 2 whenISOP mixing ratios peaked near sunset and then sharplydeclined (27 June, 3 July, and 5 July) with a time scale ofonly a couple of hours. The early evening ISOP maximum iscaused by continuing emissions of ISOP into a shallow NBLwhich grows from the surface upward. However, the mete-orological and/or chemical conditions resulting in the rapidISOP decay are still poorly characterized [Starn et al.,1998a, 1998b; Hurst et al., 2001].[11] Other ISOP oxidation products, namely MACR and

MPAN, are also shown in Figures 1 and 2. MACR mixingratios show remarkable variability from night-to-night,similar to ISOP. In general, for nights when ISOP sustainedhigh mixing ratios or even increased, MACR was alsoobserved at higher than average mixing ratios. For theevenings when ISOP decayed rapidly after sunset, MACRdid not show a consistent pattern. There are evenings whenMACR increased (22 June, 7 July) and evenings whenMACR decreased along with ISOP (27 June and 8 July). Incontrast, MPAN shows a much more consistent diurnalpattern with the highest levels observed around noon. SinceOH-initiated MACR oxidation rates peak at midday hours,these observations are consistent with our view of MPANas a local indicator of fast ISOP photochemistry [Nouaime

STROUD ET AL.: NIGHTTIME ISOPRENE CHEMISTRY AT AN URBAN SITE ACH 7 - 3

et al., 1998]. However, in looking at the nighttime MPANtrends there appears to be an anti-correlation betweenMPAN and its precursors MACR and ISOP. For theevenings when ISOP and MACR were observed at unusu-ally high mixing ratios, the MPAN mixing ratios werebelow detection limit (<5 pptv). As shown in Figures 1and 2, these nights all correspond to very low O3 levels andhigh NOx levels (22 June, 5–7 July). Conversely, on nightswhen ISOP decayed to below detection limit, the MPANmixing ratios were higher than the MPAN nighttime aver-age (23 June, 8 July, 12 July).

[12] Table 1 summarizes the statistics for the nighttimeGC measurements of ISOP and its oxidation products(21:00–06:00 CST). The large variability in the nighttimeISOP data can be seen in the wide range between ISOPminimum (<0.03 ppbv) and maximum (6.6 ppbv) mixingratio and the large 1s standard deviation about the mean,0.48 ± 0.86 ppbv. ISOP’s oxidation products show lessnighttime variability likely due, in part, to their longerchemical lifetimes in the NBL.[13] Throughout the study, there were a wide range of

nighttime NO mixing ratios (between 1 pptv and 80 ppbv).

Figure 1. The top panel presents the GC measurements of ISOP, MACR, MPAN for a selected periodbetween 18 and 28 June 1999. The middle panel shows the NO3-, O3- and OH-initiated ISOP loss ratesfrom the upper path measurement of NO3 and the tower measurements of OH and O3. The bottom panelpresents the measurements of O3 and NOx.


The highest NO3 mixing ratios (between 40 and 100 pptv)were observed with NO mixing ratios in the range between1 pptv and 100 pptv. Nighttime NO3 mixing ratios wereconsistently observed at their lowest mixing ratios (<20pptv) when NO was measured above 1 ppbv. Thus, theobserved NO3 dependence on NO is consistent with ourunderstanding of nighttime chemistry. Figures 1 and 2 showthe instantaneous ISOP loss rates (hr�1) by reaction withNO3, as well as OH and O3. The rate coefficients (kNO3,kOH, kO3) used to calculate ISOP loss rates were taken fromStevens et al. [1999] and Carter and Atkinson [1996] at298 K. The ISOP loss rate directly by O3 reaction (kO3[O3])

did not play a critical role at any time throughout the study.The ISOP loss rate via OH (kOH[OH]) shows a consistentdiurnal pattern with peak oxidation rates at noon and theslowest rates at night. Conversely, the ISOP loss rate viaNO3 (kNO3[NO3]) shows larger ISOP removal rates for theevening and nighttime periods. However, the ISOP loss ratevia NO3 is highly variable from night-to-night. The strikingfeature of the ISOP loss rate data is that there are nightswhen the ISOP oxidation rates by NO3 are comparable tothe peak daytime ISOP oxidation rates by OH. In comparingthe ISOP and NO3 nighttime measurements, there appearsto be a negative correlation as evenings with fast ISOP

Figure 2. The top panel presents the GC measurements of ISOP, MACR, MPAN for a selected periodbetween 2 and 13 July 1999. The middle panel shows the NO3-, O3- and OH-initiated ISOP loss ratesfrom the upper path measurement of NO3 and the tower measurements of OH and O3. The bottom panelpresents the measurements of O3 and NOx.


decays and very low ISOP mixing ratios are consistent withthe highest nighttime NO3 mixing ratios (23 June, 8–9 July,12–13 July). For the evenings when ISOP decayed rapidly,the start of the decays varied from night-to-night with timesas early as 17:00 CST and some as late at 21:00 CST. In anattempt to understand the important chemical removalprocesses during these times, a diurnal average was calcu-lated for ISOP and the ISOP loss rates via OH and NO3.Figure 3 shows the results of the diurnal calculation for thehours between 16:00 and 23:00 CST. There appears to be across-over point at �19:00 CST where early ISOP decaytimes are dominated by OH-initiated removal and laterdecay times are dominated by NO3-initiated removal. How-ever, caution should be used in applying these guidelines toeach night, as the uncertainty limits (1s standard deviationsrepresenting night-to-night variability) are quite large.These measurements are consistent with the results of Hurstet al. [2001], who, at times, observed two different decaytime periods to the evening ISOP decline and attributed thisto the different removal rates by OH early and NO3 later inthe decay. An interesting feature of these diurnal profiles isthat the early evening period coincides with minimum ISOPloss rates, an effect that might contribute to the higher ISOPobserved at this time.[14] Clearly, the trends in ISOP and its oxidation products

are influenced by a wide range of biological, chemical anddynamical processes. The impact of urban fluxes dramati-cally alters the chemistry of a region on short time scaleswhich makes diurnal averaging of data difficult to interpret.Since the trends in these chemical species are so widelyvariable from night-to-night, we will focus the remainder ofthe paper on describing, in detail, several intriguing casestudies which are unique to urban, forested environmentsunder moderate to high NOx conditions. Section 3.2describes several ISOP evening decays under intermediateNOx and high NO3 conditions. Section 3.3 details theevening/early morning of 5–6 July, a high NOx and lowNO3 nighttime period when ISOP increased in concentra-tion. Section 3.4 describes a period of rapid ISOP photo-chemistry on the morning of 6 July after sunrise but beforevertical mixing significantly redistributed the chemicalspecies. These selected case studies represent conditionswhere we believe chemical rather than meteorologicalprocesses dominate the ISOP trends.

3.2. Cases Studies: Rapid ISOP Decays During theEvening

3.2.1. Evening/Nighttime of 22–23 June 1999[15] Figure 4 presents the ISOP, MACR, MPAN, NO,

NO2, O3 and ISOP loss rates by OH and NO3 for the periodof ISOP decay starting at 17:00 CST on 22 June 1999. ISOP

decreased from 0.67 ppbv to near detection limits in 3 hrsand remained at very low mixing ratios throughout the nightuntil sunrise. During the ISOP decay, MACR slowlyincreased and maintained relatively high mixing ratiosthroughout the night. MPAN mixing ratios fell slightlyduring the ISOP decay, however, remained at relativelyhigh mixing ratios throughout the night. To understand theimportant chemical removal processes for ISOP during thisperiod, we compare kOH[OH] and kNO3[NO3]. Initially, OHis the dominant oxidant for ISOP between 17:00 and 18:30CST, however after sunset at 18:30 CST, the upper pathmeasurement of NO3 increased rapidly to mixing ratios of80 pptv at 22:30 CST. Even after the peak NO3 mixingratio, NO3 maintained high levels of �50 pptv throughoutthe night. Using the rate coefficient for the reaction betweenISOP and NO3 of 6.85 � 10�13 cm3molec�1sec�1 at 298 K[Carter and Atkinson, 1996] and an NO3 mixing ratio of 50pptv, we calculate an ISOP chemical lifetime of only 20min. Thus, nighttime NO3 chemistry likely explains theextremely low mixing ratios of ISOP observed throughoutthe entire night. The rate coefficient for the reaction betweenMACR + NO3 has been measured at 4.75 � 10�15

Figure 3. Diurnal average for ISOP and ISOP loss rates viaOH and NO3 for the evening hours between 16:00 and 23:00CST. Calculations are included for both the upper and middlepath NO3 measurements. Standard deviation bars (1s) areincluded and represent evening-to-evening variability.

Table 1. Statistics for the Nighttime GC Measurements of

Isoprene and its Oxidation Products (21:00–06:00 CST)

Species Maximum(ppbv)

Minimum(ppbv)

Mean(ppbv)

StandardDeviation(ppbv)

Median(ppbv)

Points

ISOP 6.6 0.030a 0.48 0.86 0.20 270MACR 2.2 0.092 0.51 0.35 0.42 270MVK 1.7 0.11 0.56 0.28 0.51 270MPAN 0.17 0.005a 0.040 0.025 0.038 1066

aDetection limit.


cm3molec�1sec�1 [Kwok et al., 1996]. At 50 pptv NO3, wecalculate a MACR chemical lifetime via NO3 removal of1.9 days. Similarly, the MPAN + NO3 rate coefficient hasbeen measured recently at 1.45 � 10�16 cm3molec�1sec�1

[Canosa-Mas et al., 1999]. This corresponds to a chemicallifetime of 64 days at 50 pptv NO3. These long chemicallifetimes for MACR and MPAN likely explain their persis-tence throughout the night. The small change in MACR andMVK mixing ratio throughout the ISOP decay is alsoconsistent with their small yield from the reaction of

NO3 + ISOP (3.5% each) as measured by Kwok et al.[1996]. Figure 4 also shows the mixing ratios of NO, NO2

and O3 on the evening of 23 June 1999. We were in anintermediate NOx environment (8–35 ppbv) with low NOmixing ratios (<1 pptv) suggesting we were in the regimefor optimum NO3 mixing ratios and optimum ISOP lossrates via reaction with NO3.[16] Measurements of water vapor, wind speed, wind

direction and NBL height were used to understand thedaytime/nighttime transition in the boundary layer. Water

Figure 4. The top panel presents the GC measurements of ISOP, MACR and MPAN for 22–23 June1999, between 17:00 and 06:00 CST. The middle panel shows the measurements of O3, NO and NO3.The bottom panel presents the OH- and NO3-initiated ISOP loss rates from the tower measurement of OHand upper path measurement of NO3. Vertical bars represent estimated 1s errors.


vapor was observed to increase starting at 16:00 CSTconsistent with continuing surface emissions of water vaporinto a less convective boundary layer. A thunderstorm gustfront passed over CFA at 18:10 CST. The cooler air behindthe gust front help set up a low-level jet at an altitude of 100m. After 18:20 CST, the water vapor trace lost its high-frequency fluctuations suggesting the formation of a low-level inversion with a stratified surface layer below. TheSODAR measurements after the gust front show that theNBL inversion varied in height between 100 and 200 m.Winds were generally from the southeast direction at boththe start and end of the ISOP decay time period; however,wind directions shifted to the northwest for a short periodassociated with the gust front. Wind directions were variable(100�–250�) throughout the night following the eveningISOP decay.[17] One possible explanation for the observed ISOP

decay is a shift in wind direction to an ISOP-poor regionassociated with the gust front and variable winds thereafter.However, other ISOP products (such as MACR, MVK andMPAN) did not show large variability before and after thegust front suggesting similar chemical history for the airbefore and after the gust front. In addition, with relativelyhigh and variable winds throughout the night, if advectionfrom ISOP-poor regions were causing the decay process,then an oscillation of ISOP, MACR and MVK mixing ratioswould have been expected throughout the night as winddirection fluctuated. Thus, it seems unlikely that advectioncan explain the observed ISOP decay. However, the night-time dynamics may still couple with chemistry in animportant way at the CFA site. The height of the NBLmay play a critical role in determining the amount ofdilution for emitted gases such as NO and thus play acritical role in the O3 and NOx balance at the surface.Evenings with generally high NBL heights result in greaterdilution of urban NO emissions compared to evening withlower NBL heights. As a result, evening with relatively highand constant NBL heights throughout the ISOP decay, suchas 22 June, also maintain significant O3 mixing ratios. Thetitration of the NO by O3 results in very low NO mixingratios within the NBL and favorable conditions for NO3

production. In this way, the height of the NBL affects thedilution of urban NO emissions and the eventual oxidizingcapacity of the NBL. In summary, on this night, OHoxidation early on and then NO3 oxidation later appearsresponsible for the chemical loss of ISOP during theobserved decay.3.2.2. Evening/Nighttime of 7–8 July 1999[18] Figure 5 presents the ISOP, MACR, MPAN, NO,

NO2, O3 and ISOP loss rates by OH and NO3 for the periodof ISOP decay starting at 19:00 CST on 7 July 1999. TheISOP decayed exponentially from a mixing ratio of 0.86ppbv with an observed 1/e time constant of 0.9 hrs. Duringthe ISOP decay, the middle path measurement of NO3 rosesharply peaking at 54 pptv. The ISOP loss rates by OH andNO3 suggest that NO3 reaction was the dominant chemicalremoval term for ISOP throughout most of the ISOP decay.This is due to the later start time for the ISOP decaycompared to the previous 22–23 June case study. Fromthe NO3 and OH measurements during the ISOP decay, wecalculate an ISOP chemical lifetime of 1.3 hr at the start ofthe decay and 0.25 hr by the end of the decay. These short

chemical lifetimes are consistent with the time scales for theobserved ISOP decay.[19] Figure 5 also presents the measurements of MACR

and MPAN for the evening and night of 7–8 July 1999.Both the MACR and MPAN increased in mixing ratiothroughout the ISOP decay and then, in general, decreasedin mixing ratio throughout the night. The observed levels ofMACR were in the range 0.4–0.8 ppbv, relatively high andsimilar to the 22–23 June case study. The observed levels ofMPAN were in the range 50–100 pptv, again relatively highand similar to the 22–23 June case study. The sustainedlevels of MACR and MPAN throughout the night likelyresult from their slower rate of reaction with NO3. Figure 5also shows the balance between NO, NO2 and O3 for theevening and night of 7–8 July 1999. NO remained at lowmixing ratios (<10 pptv) throughout the night. NO2 gen-erally increased throughout the evening (10–40 ppbv) andremained in this range throughout the night. O3 decreasedfrom 60 to 2 ppbv throughout the night, but at no periodwas completely titrated by NO emissions. These conditionsare optimum for rapid NO3 formation and efficient NO3-initiated ISOP removal.[20] Winds remained steady from the northeast direction

(10�–30�) throughout the isoprene decay time period whilewind speeds generally decreased from 3 m/s to less than 1m/s. The local wind measurements at CFA are consistentwith back-trajectory analysis (HYSPLIT4 (HYbrid Single-Particle Lagrangian Integrated Trajectory) Model, 1997,available at http://www.arl.noaa.gov/ready/hysplit4.html,NOAA Air Resources Laboratory, Silver Spring, Maryland)which show air parcels arriving at CFA from the norththroughout the ISOP decay period. Steady wind directionsand the fact that MACR and MPAN show little change inmixing ratio as ISOP decays rapidly supports NO3 chem-istry rather than advection of an air mass of different history.The SODAR measurements show significant wind shearand vertical mixing within the NBL on this evening. NBLheights varied between 100–200 m, similar to the 22–23June case study. Wind shear in the NBL will help to de-stratify high levels of NO in the surface layer, thus enablinghigh levels of NO3 to form near the surface.

3.3. Case Studies: Sustained High Nighttime ISOPMixing Ratios

[21] The evening of 5 July 1999, and early morning of 6July 1999, was an unusual period because ISOP wasobserved to increase by an order of magnitude from atypical daytime value of 0.5 ppbv to nighttime mixingratios in the range 2.5–6.6 ppbv. Figure 6 shows themeasurements of ISOP, MACR+MVK, MPAN, NO, NO2,O3, ISOP loss rates via OH and NO3, water vapor, andsurface wind direction and wind speed for this period.Figure 6 also presents the fast response PTR-MS measure-ments for ISOP and MACR+MVK. The consistencybetween these two vastly different measurement techniquesfor ISOP and MACR+MVK confirms the anomalously highISOP mixing ratios observed on this night were real. Thesum of the MACR+MVK mixing ratios was also highduring this night with mixing ratios between 1.8–4.0 ppbv.Conversely, MPAN was below detection limit throughoutthe night. To understand these chemical trends, it is useful tolook at the balance between O3 and NOx. In contrast to the


previous two case studies (22–23 June, 7–8 July), theevening and night of 5–6 July was impacted by largeemissions of NO. On the evening of 5 July, O3 wascompletely titrated by NO at 22:00 CST and NO increasedin mixing ratio throughout the night to levels between 40–80 ppbv at sunrise. O3, OH and NO3 were all suppressedthroughout the night under these high NO conditions andhigh NO/NO2 ratios. Figure 6 also presents the observedinstantaneous ISOP loss rates via OH and NO3. Throughoutthe evening and night, kNO3[NO3] and kOH[OH] remainedlow. The average of kNO3[NO3] from the middle pathDOAS measurement was 0.3 ± 0.4 hr�1 between the hours

of 19:00 and 01:00 CST resulting in a less efficientchemical removal of ISOP compared to the previous casestudies. The large NO/NO2 ratios also result in an efficientloss mechanism for MPAN, as the thermal equilibriumbetween MPAN and the peroxyacyl radical (R3) is contin-uously perturbed by the rapid peroxyacyl radical reactionwith NO (R4). For example, between 21:30 and 23:15 CST,the NO/NO2 ratio increased from 2.7 � 10�3 to 1.0. Thiscorresponds to a decrease in MPAN lifetime from 180 hrs to1.5 hrs. This large change in lifetime likely explains theobserved decrease in MPAN mixing ratio from 68 pptv tobelow detection limit (<5 pptv) in a matter of 2 hours.

Figure 5. The top panel presents the GC measurements of ISOP, MACR and MPAN for 7–8 July 1999,between 19:00 and 06:00 CST. The middle panel shows the measurements of O3, NO and NO3. Thebottom panel presents the OH- and NO3-initiated ISOP loss rates from the tower measurements of OHand middle path measurements of NO3. Vertical bars represent estimated 1s errors.


[22] The meteorological data for this evening support theformation of a shallow stable NBL just before sunset. Thewater vapor measurements can be used to give an indicationof the collapse of the daytime mixed layer, as a reduction invertical mixing results in the water vapor trace losing its

high frequency perturbations. At CFA, on the evening of 5July, this transition occurred at 19:00 CST. The winddirection also shifted and wind speeds dropped to verylow levels (<0.25 m/s). The water vapor trend did begin toincrease at 17:30 CST, suggesting the boundary layer was

Figure 6. The top panel presents the GC and PTR-MS measurements of ISOP and MACR+MVK, aswell as the GC measurements of MPAN for 5–6 July 1999, between 14:00 and 12:00 CST. The secondpanel shows the measurements of O3, NO and NO3. The third panel presents the OH- and NO3-initiatedISOP loss rates from the tower measurement of OH and both upper and middle path measurements ofNO3. Vertical bars represent estimated 1s errors. The bottom panel presents the surface measurements ofwind direction, wind speed and water vapor. Wind direction data are presented only for wind speedsgreater than 1 m/s.


still convective between 17:30 and 19:00 CST but themixing heights were decreasing. The ISOP measurementsbetween 17:00 and 20:00 CST are consistent with the theorythat surface emissions were being concentrated in a stablesurface layer as ISOP increased from 1.1 to 4.0 ppbv. TheSODAR measurements also show the formation of a shal-low NBL with a height of 75 m. The SODAR data alsoindicate that the NBL remained stable throughout the night.This is confirmed by the O3 measurements which did notshow any spikes that might be indicative of vertical mixing.Back-trajectory analysis for this period suggests air massesoriginated from northwest of Nashville, an area rich in ISOPemissions. The combined analysis of the measurementssuggests that the sustained high ISOP mixing ratiosthroughout the night resulted from the formation of ashallow, stable NBL with surface air parcels originating inISOP-rich regions outside Nashville. On transport to thesite, air parcels were enriched in NO by urban emissionsresulting in a surface layer with low oxidizing capacity andlong ISOP chemical lifetimes.

3.4. Case Studies: Early Morning Signature ofOH-Photochemistry on 6 July 1999

[23] Two hours after sunrise, between 06:45 and 08:00,ISOP mixing ratios were observed to sharply drop with acorresponding increase in MACR+MVK. Figure 7 showsthe PTR-MS measurements of ISOP and MACR+MVK forthe 30 min period starting at 06:47 CST. ISOP decreasedfrom 4.9 to 3.2 ppbv while MACR+MVK increased from4.2 to 5.2 ppbv. A 0-D chemical box model was used toinvestigate the possible OH-initiated chemical conversion ofISOP to its first-stage oxidation products, MACR andMVK. The OH measurements were used to constraintime-dependent pseudo-first order rate coefficients for theOH reactions with ISOP, MACR and MVK. The traces inFigure 7 are the results of the chemical box model (inte-grated with Facsimile) initialized with mixing ratios at

06:47 CST. Since the model required speciation for MACRand MVK and the PTR-MS measured only the sum,speciation was determined from the GC measurements ofMVK and MACR at 07:00 CST and applied to the initialPTR-MS sum at 06:47 CST. The general consistencybetween the ISOP decay and the model suggests that onthis morning OH was an important oxidant converting ISOPto its first-stage products, MACR and MVK.[24] The MVK/MACR ratio has been modeled previ-

ously and a daytime ratio of 2.0 is expected from OH-initiated photochemistry [Montzka et al., 1993]. The GCmeasurements of MVK and MACR between 07:00 and08:00 CST increased in mixing ratio by 0.52 ppbv and 0.26ppbv, respectively. Thus, the ratio of the MVK mixing ratiochange to the MACR mixing ratio change was 2.0,identical to that expected from OH-kinetics, and furthersupports that the production of MVK and MACR duringthis early morning time period was driven by OH-photo-chemistry.[25] Interpreting the trends in chemical species based

purely on chemistry is complicated during this period dueto the growth of the surface layer. Water vapor can again beused as a tracer of boundary layer dynamics. At sunrise,surface evaporation results in water vapor being concen-trated in the surface layer; however as the boundary layergrows, drier air is entrained from the residual layer. Thecompeting factors result in a surface water vapor maximumand then decline. Turbulent mixing as the boundary layergrows also results in the water vapor trace exhibiting highfrequency perturbations. The water vapor measurements atCFA on the morning of 6 July maximize at 06:30 CST andbegin to show increased high frequency perturbationsshortly after. Thus, it does appear from the water vapordata that the NBL had started to grow during the PTR-MStrends. These observations are consistent with the SODARdata which also show increased vertical mixing starting at06:30 CST. Between 06:45 and 7:15 CST, the SODARmeasurements suggest the surface layer increased in eleva-tion from 125 to 275 m.[26] The O3 measurements can also be used to yield

insight into the breakup of the NBL. O3 slowly increasedjust after sunrise in response to increased NO2 photolysisaccording to the photostationary state between O3, NO andNO2. O3 gradually increased until 07:00 CST at which pointthe rate of change increased. This is likely due to theincreased importance of vertical mixing after 07:00 CST.Vertical mixing continued to contribute to the O3 increaseuntil 10:00 CST when the boundary layer approached mid-day elevations. The small increase in O3 between sunriseand 07:00 CST suggests that vertical mixing played only aminor contribution to the PTR-MS ISOP trend.[27] The GC measurements of MACR+MVK increased to

a maximum mixing ratio of 5 ppbv at 08:00 CST and thendeclined to typical daytime values between 09:00 and 10:00CST. These observations suggest that the mixing ratios ofMACR+MVK would have needed to be higher than 5 ppbvin the residual layer to produce the observed time series byentrainment. This possibility is highly unlikely as the resid-ual layer is composed of air from the previous days con-vective boundary layer. Typical afternoon MACR+MVKmixing ratios were 0.8 ppbv. The maximum afternoonMACR+MVK mixing ratio for the entire campaign, under

Figure 7. PTR-MS measurements and 0-D box modelresults of ISOP and MACR+MVK for a 30 min periodstarting at 06:47 CST on 5 July 1999.


a variety of wind directions, was 2.0 ppbv. The nighttimechemical processing of ISOP in the residual layer is apossibility, however, calculations with a constant 2 ppbvISOP and 60 ppbv O3 over 10 hrs suggests that ozonolysis ofISOP in the residual layer is too slow to account forsignificant production of MACR+MVK (1.4 ppbv). Thus,it seems unlikely that vertical mixing would cause the firststage oxidation products to increase in concentration.[28] Between 09:00 and 10:00 CST, MACR and MVK

decreased sharply; a trend expected when boundary layergrowth entrains air from the residual layer. Other carbonyls,such as acetaldehyde and propanal, also showed an increasebetween 07:00 and 09:00 CST, presumably due to theoxidation of reactive anthropogenic hydrocarbon precur-sors, then a decrease after 09:00 CST. CO also decreasedsharply between 09:00 and 10:00 CST. These chemicaltrends are supported by the wind profiler measurements ofboundary layer height. Boundary layer heights increasedmarginally between 07:00 and 08:30 CST (320–420 m) andthen significantly between 08:30 and 10:00 CST (420–1310 m) as the inversion broke and the convective boundarylayer grew rapidly. These observations suggest that verticalmixing only marginally affected the chemical trendsbetween 06:45 and 8:00 CST and that OH-initiated oxida-tion dominated the ISOP trend between 06:47 and 07:15CST.

3.5. Overall Trends in Nighttime ISOP

[29] Table 2 summarizes cases when ISOP decayed in theevening to near detection limits. Table 3 summarizes caseswhen ISOP sustained high ISOP mixing ratios throughoutthe night. By tabulating the nighttime data, we can begin to

look for the factors that distinguish these two regimes. Thedistinctive features of the ISOP decay cases are the sim-ilarities in the NO mixing ratio (<0.083 ppbv), O3 mixingratio (28–67 ppbv) and NO2 mixing ratio (11–20 ppbv).The large [O3][NO2] products result in large NO3-initiatedISOP oxidation rates (1.2–3.0 hr�1). In contrast, the sus-tained nighttime ISOP cases are characterized by low O3

mixing ratios (<22 ppbv) and low ISOP loss rates via OHand NO3. Differences in urban emissions, NBL dynamicsand wind direction from night-to-night affect the observedbalance between NO, NO2 and O3 and thus set up theconditions for the two regimes.

4. Summary

[30] This paper investigated ISOP trends after sunsetwhen emissions of ISOP were small and the chemicalbudget was dominated by loss via OH and/or NO3. In thisurban, forested environment, ISOP and its first stage oxi-dation products, MACR and MVK, showed marked varia-bility in their evening and nighttime diurnal trends. This isin contrast to remote, forested environments where ISOPhas a consistent diurnal pattern with a daytime maximumand a rapid decay in the evening.[31] For cases when ISOP decayed rapidly in the evening

and remained low throughout the night, observations of OHand NO3 suggest rapid ISOP oxidation rates. OH wasobserved to be an important oxidant for ISOP throughoutthe day and for periods early in the evening (17:00–19:00CST). On average, NO3 was observed to be the dominantoxidant for ISOP after 19:00 CST and throughout the night.For periods when the NO3-initiated ISOP loss rate was

Table 2. Evenings When Isoprene Decayed Significantly to Near or at Detection Limits

Evening/Night

Time(CST)

� [ISOP](ppbv)

�[MACR](ppbv)

Average O3

(ppbv)AverageNO2

(ppbv)

AverageNO

(ppbv)

kNO3[NO3](hr�1)a

kOH[OH](hr�1)a

Wind Direction (�) andSpeed (m/s)

6/22–6/23 17:00–21:00 0.67–<0.04 0.26–0.38 67 15 0.083 0.53–3.3 1.5 1.8–0.55 1.1 Variable between150–300� with avg.speeds of 1 m/s

6/27–6/28 20:00–24:00 3.7–0.072 1.6–0.24 28 11 <0.01 0.76–2.2 1.2 Steady at 0.54 No data7/7–7/8 19:00–22:00 0.84–<0.04 0.42–0.50 49 20 0.080 0.95–3.6 3.0 0.81–<0.3 0.46 Steady from 10–30�

with winds speedsdecreasing

7/8–7/9 18:00–23:00 1.7–0.15 0.83–0.47 59 12 0.020 Variable 2.6 0.86–0.43 0.78 No data7/12–7/13 21:00–22:00 0.66–<0.04 0.38–0.40 56 13 <0.01 2.2–2.3 2.3 0.45–0.47 0.52 Steady from 0� with

avg. speeds of 0.7 m/saRange and average are included for time period.

Table 3. Evenings When Isoprene Increased Significantly

Evening/Night

Time(CST)

�[ISOP](ppbv)

�[MACR](ppbv)

AverageO3

(ppbv)AverageNO2

(ppbv)

AverageNO

(ppbv)

kNO3[NO3](hr�1)a

kOH[OH](hr�1)a

Wind Direction (�)and Speed (m/s)

6/24–6/25 19:00–22:00 0.75–1.1 0.41–0.53 15 23 0.019 Steady at 0.38 <0.3 Variable between200–300� with avg.speed of 1 m/s

7/5–7/6 21:00–06:00 2.9–5.6 0.8–2.0 0.50 23 29 0.70–0.56 0.49 <0.3 Variable and light withavg. speeds <0.25 m/s

7/9–7/10 21:00–02:00 0.35–0.80 0.48–1.0 22 9 <0.01 0.40–0.07 0.30 0.45–0.41 0.55 Steady from 200� at anavg. speed of 3 m/s

aRange and average are included for time period.


greater than 1.5 hr�1, ISOP typically decayed to belowdetection limits. These nights also sustained relatively highMACR, MVK and MPAN mixing ratios consistent withtheir slow reaction rates with NO3. The measurements ofISOP, NO3 and OH described here support the conclusionsof Starn et al. [1998a, 1998b]. The importance of NO3 as anighttime oxidant for ISOP appears to be a characteristic ofurban forested sites where moderate NOx conditions existand the balance of the NOx is in the form of NO2.[32] For cases when ISOP sustained high mixing ratios

throughout the night, the observed ISOP oxidation rates viaOH and NO3 were low. These nights were generallycharacterized by high NO/NO2 ratios resulting in decreasedMPAN lifetimes and low MPAN mixing ratios. In an urbanforested environment, it appears the availability of NO3

depends critically on the balance between NO, NO2 and O3,which in turns depends on the spatial heterogeneity of NOsources as well as wind direction and NBL dynamics. In thisway, the nighttime ISOP trends at urban forested sitesdepends critically on a coupling between chemistry anddynamics. The evening of 6 July represents a case where ashallow NBL combined with wind directions from ISOPand urban emission sources results in the complete titrationof O3 and the build up of NO. This effectively reduces theoxidizing capacity of the nighttime boundary layer so thatreactive hydrocarbons can sustain or even build up inconcentration throughout the night. As observed on themorning of 6 July, this build up of reactive hydrocarbonssets up an interesting scenario just after sunrise. A rapiddecrease in reactive hydrocarbon mixing ratios wasobserved with a corresponding increase in carbonyl mixingratios. The trends in ISOP, MACR and MVK are consistentwith OH-initiated processing just after sunrise but beforemixing significantly redistributed the chemical species ver-tically. Future work should investigate how this earlymorning chemical processing of concentrated VOCs mayaffect peak afternoon O3 levels in urban environments.

[33] Acknowledgments. This work was part of the Southern Oxi-dants Study (SOS), a collaborative university, government, and privateindustry study to improve scientific understanding of the accumulation andeffects of photochemical oxidants. Parts of this work were supported byNOAA through the Health of the Atmosphere Initiative, constitutingNOAA’s contribution to SOS. The authors would like to thank thereviewers, D. Parrish and M. Trainer, for helpful suggestions in thepreparation of the manuscript.

ReferencesAngevine, W. M., A. B. White, and S. K. Avery, Boundary layer depth andentrainment zone characterization with a boundary layer profiler, Bound-ary Layer Meteorol., 68, 375–385, 1994.

Apel, E. C., J. G. Calvert, J. P. Greenberg, D. Riemer, R. Zika, T. E.Kleindienst, W. A. Lonneman, K. Fung, and E. Fujita, Generation andvalidation of oxygenated volatile organic carbon standards for the 1995Southern Oxidants Study Nashville Intensive, J. Geophys. Res., 103,22,281–22,294, 1998.

Biesenthal, T. A., Q. Wu, P. B. Shepson, H. A. Wiebe, K. G. Anlauf, and G.I. Mackay, A study of relationships between isoprene, its oxidation pro-ducts, and ozone, in the Lower Fraser Valley, BC, Atmos. Environ., 98,2049–2058, 1997.

Canosa-Mas, C. E., M. D. King, D. E. Shallcross, and R. P. Wayne, Kineticinvestigation of the reaction between the NO3 radical and peroxy-methacrylic nitric anhydride (MPAN), Phys. Chem. Chem. Phys., 1(10),2411–2414, 1999.

Carter, W. P. L., and R. Atkinson, Development and evaluation of a detailedmechanism for the atmospheric reactions of isoprene and NOx, Int. J.Chem. Kinet., 28, 497–530, 1996.

Chameides, W. L., R. W. Lindsay, J. Richardson, and C. S. Kiang, The roleof biogenic hydrocarbons in urban photochemical smog: Atlanta as a casestudy, Science, 241, 1473–1475, 1988.

Cohn, S. A., and W. M. Angevine, Boundary layer height and entrainmentzone thickness measured by lidars and wind profiling radars, J. Appl.Meteorol., 39, 1233–1247, 2000.

Dentener, F. J., and P. J. Crutzen, Reaction of N2O5 on tropospheric aero-sols: Impact on the global distributions of NOx, O3 and OH, J. Geophys.Res., 98, 7149–7163, 1993.

Fehsenfeld, F. C., et al., Emissions of volatile organic compounds fromvegetation and the implications for atmospheric chemistry, Global Bio-geochem. Cycles, 6(4), 389–430, 1992.

Geyer, A., B. Alicke, S. Konrad, T. Schmitz, J. Stutz, and U. Platt, Chem-istry and oxidation capacity of the nitrate radical in the continentalboundary layer near Berlin, J. Geophys. Res., 106, 8013–8025, 2001.

Goldan, P. D., W. C. Kuster, F. C. Fehsenfeld, and S. A. Montzka, Hydro-carbon measurements in the southeastern United States: The Rural Oxi-dants in the Southern Environment (ROSE) Program 1990, J. Geophys.Res., 100, 25,945–25,963, 1995.

Grimsdell, A. W., and W. M. Angevine, Convective boundary layer heightmeasured with wind profilers and compared to cloud base, J. Atmos.Oceanic Technol., 15, 1332–1339, 1998.

Hansel, A., and A. Wisthaler, A method for real-time detection of PAN,PPN and MPAN in ambient air, Geophys. Res. Lett., 27, 895–898, 2000.

Hurst, J. M., et al., Investigation of nighttime decay of isoprene, J. Geo-phys. Res., 106, 24,335–24,346, 2001.

Jacob, D., Heterogeneous chemistry and tropospheric ozone, Atmos. Envir-on., 34, 2131–2159, 2000.

Kwok, E. S. C., S. M. Aschmann, J. Arey, and R. Atkinson, Productformation from the reaction of the NO3 radical with isoprene and rateconstants for the reactions of methacrolein and methyl vinyl ketone withthe NO3 radical, Int. J. Chem. Kinet., 28, 925–934, 1996.

Martinez, M. D., D. Perner, E.-M. Hackenthal, S. Kulzer, and L. Schutz,Measurements of NO3 at Helgoland during the NORDEX campaign inOctober 1996, J. Geophys. Res., 105, 22,685–22,695, 2000.

Mather, J. H., P. S. Stevens, and W. H. Brune, OH and HO2 measurementsusing laser-indused fluorescence, J. Geophys. Res., 102, 6427–6436,1997.

Montzka, S. A., M. Trainer, P. D. Goldan, W. C. Kuster, and F. C. Feh-senfeld, Isoprene and its oxidation products, methyl vinyl ketone andmethacrolein, in the rural troposphere, J. Geophys. Res., 98, 1101–1111,1993.

Nouaime, G., S. B. Bertman, C. Seaver, D. Elyea, H. Huang, P. B. Shepson,T. K. Starn, D. D. Riemer, R. G. Zika, and K. Olszyna, Sequentialoxidation products from tropospheric isoprene chemistry: MACR andMPAN at a NOx-rich forest environment in the southeastern UnitedStates, J. Geophys. Res., 103, 22,463–22,471, 1998.

Platt, U., D. Perner, G. W. Harris, A. M. Winer, and J. M. Pitts, Detection ofNO3 in the polluted troposphere by differential optical absorption, Geo-phys. Res. Lett., 7, 89–92, 1980.

Roberts, J. M., et al., Measurements of PAN, PPN, and MPAN made duringthe 1994 and 1995 Nashville intensives of the southern oxidant study:Implications for regional ozone production from biogenic hydrocarbons,J. Geophys. Res., 103, 22,473–22,490, 1998.

Roberts, J. M., F. Flocke, C. A. Stroud, D. Hereid, E. J. Williams, F. C.Fehsenfeld, W. H. Brune, M. Martinez Harder, and H. Harder, Ground-based measurements of PANs during the 1999 Southern Oxidant StudyNashville intensive, J. Geophys. Res., 107, 10.1029/2001JD000947, inpress, 2002.

Starn, T. K., P. B. Shepson, S. B. Bertman, J. S. White, B. G. Splawn, D. D.Riemer, R. G. Zika, and K. Olszyna, Observations of isoprene chemistryand its role in ozone production at a semirural site during the 1995Southern Oxidants Study, J. Geophys. Res., 103, 22,425–22,435, 1998a.

Starn, T. K., P. B. Shepson, D. D. Riemer, R. G. Zika, and K. Olzyna,Nighttime isoprene chemistry at an urban-impacted forest site, J. Geo-phys. Res., 103, 22,437–22,447, 1998b.

Stevens, P., D. L’Esperance, B. Chuong, and G. Martin, Measurements ofthe kinetics of the OH-initiated oxidation of isoprene: Radical propaga-tion in the OH + Isoprene + O2 + NO reaction system, Int. J. Kinet., 31,637–643, 1999.

Stroud, C. A., et al., Isoprene and its oxidation products, methacrolein andmethyl vinyl ketone, at an urban forested site during the 1999 SouthernOxidant Study, J. Geophys. Res., 106, 8035–8046, 2001.

Stutz, J., and U. Platt, Improving long-path differential optical absorptionspectroscopy with a quartz-fiber mode mixer, Appl. Opt., 36, 1105–1115,1997.

Trainer, M., E. J. Williams, D. D. Parrish, M. P. Buhr, E. J. Allwine, H. H.Westberg, F. C. Fehsenfeld, and S. C. Liu, Models and observations ofthe impact of natural hydrocarbons on rural ozone, Nature, 329, 705–707, 1987.


Williams, E. J., et al., Intercomparison of ground-based NOy measurementtechniques, J. Geophy. Res., 103, 22,261–22,281, 1998.

Williams, J., et al., Regional ozone from biogenic hydrocarbons deducedfrom airborne measurements of PAN, PPN and MPAN, Geophys. Res.Lett., 24, 1099–1102, 1997.

��W. M. Angevine, F. C. Fehsenfeld, D. Hereid, J. M. Roberts, and E. J.

Williams, NOAA Aeronomy Laboratory, 325 Broadway, Boulder, CO80303, USA. ([email protected]; [email protected]; [email protected]; [email protected]; [email protected])W. H. Brune, H. Harder, and M. Martinez-Harder, Department of

Meteorology, Pennsylvania State University, University Park, PA 16802,USA. ([email protected]; [email protected]; [email protected])A. Hansel and A. Wisthaler, Institute for Ion Physics, University of

Innsbruck, A-6020, Innsbruck, Austria. ([email protected]; [email protected])G. Hoenninger, Institut fuer Umweltphysik, University of Heidelberg, D-

69117, Heidelberg, Germany. ([email protected])C. A. Stroud, Atmospheric Chemistry Division, National Center for

Atmospheric Research, Boulder, CO 80305, USA. ([email protected])J. Stutz, Department of Atmospheric Sciences, University of California,

Los Angeles, CA 90095, USA. ( [email protected])A. B. White, Environmental and Technology Laboratory, 325 Broadway,

Boulder, CO 80303, USA. ([email protected])


Documents

Nighttime isoprene trends at an urban forested site during the 1999