Summer and Spring Ozone Profiles Over the North Atlantic from Ozonesonde Measurements S

University of Rhode Island DigitalCommons@URI Graduate School of Oceanography Faculty Graduate School of Oceanography Publications

1996 Summer and spring ozone profiles over the North Atlantic from ozonesonde measurements S. J. Oltmans

H. Levy II

See next page for additional authors

Follow this and additional works at: https://digitalcommons.uri.edu/gsofacpubs

Citation/Publisher Attribution Oltmans, S. J., et al. (1996), Summer and spring ozone profiles over the North Atlantic from ozonesonde measurements, J. Geophys. Res., 101(D22), 29179–29200, doi: 10.1029/96JD01713. Available at: https://doi.org/10.1029/96JD01713

This Article is brought to you for free and open access by the Graduate School of Oceanography at DigitalCommons@URI. It has been accepted for inclusion in Graduate School of Oceanography Faculty Publications by an authorized administrator of DigitalCommons@URI. For more information, please contact [email protected]. Authors S. J. Oltmans, H. Levy II, J. M. Harris, John Merrill, J. L. Moody, J. A. Lathrop, E. Cuevas, M. Trainer, M. S. O'Neill, J. M. Prospero, H. Vömel, and B. J. Johnson

This article is available at DigitalCommons@URI: https://digitalcommons.uri.edu/gsofacpubs/261 JOURNAL OF GEOPHYSICAL RESEARCH, VOL 101, NO. D22, PAGES 29,179-29,200, DECEMBER 20, 1996

Summerand springozone profiles over the North Atlanticfrom ozonesonde measurements

S.J.Oltmans, 1H. Levy11, 2 J.M.Harris, 1J.T. Merrill, 3 J.L.Moody, 4 J.A.Lathrop, 1E. Cuevas,5 M. Trainer,6 M.S. O'Neill,l, 7 J.M.Prospero, 8 H. V6mel, 9 and B.J. Johnson,I, 7

Abstract.Ozone profiles obtained by near-dailyozonesonde observations during campaignsat several sites in theNorth Atlantic are used to constructtime-height cross sectionsof ozoneconcentration through the troposphere. Strong day-to-day ozone variabilityon the scale of synopticmeteorological disturbances is found both in thespring andin thesummer throughout much of thetroposphere. Layers of highozone concentration(-100 ppb)are frequently seen in themiddle and upper troposphere and are invariably associatedwith transport characteristics that strongly support a stratosphericsource for theselayers. Regions of lowozone (<40 ppb) are seen in themiddle and upper troposphere associatedwith higherrelative humidity. The connectionof theseevents with low surface mixingratios suggests that convective processes mix air lowin ozoneup through the troposphere.Vertical layering of ozonemixing ratio, which is seenat all of theobserving locations,is a resultof differingsources of air in thedifferent layers.

Introduction balloon-borne ozonesondes. During the NARE intensive in August 1993, ozone profiles were obtained from Keflavik, The significantrole of ozoneas an atmosphericoxidizing Iceland; Bermuda; and Lages, Azores. In addition, intensive agentand greenhousegas and the importantpart that human series of spring profiles were obtained in April/May 1993 in pollution sources can play in modifying its distribution Bermuda and May 1994 in the Azores. We will also report providestrong motivation for a morecomplete understanding on profiles obtained at Cape Race, Newfoundland, in of ozone'sbehavior in the atmosphere.Because of its impact July/August 1991 and soundingsfrom Tenerife in the Canary on human health and plant life, ozone has been most Islands in August 1993 and May 1994 (see Figure 1 for map intensively observed near the surface. The limitations of of ozonesonde launch locations). The near daily soundings observingozone and other constituentsprimarily at the from several of these campaigns give us the opportunity to surfacehave long beenrecognized. Many componentsof the investigate the processesresponsible for the very strong day- North Atlantic RegionalExperiment (NARE) were designed to-day variability that was regularly seen during the intensive to provide the necessary profile information for profiling protocol. Several events in which high ozone understandingthe chemicalbehavior of the troposphereover amounts were seen in the midtroposphere over Bermuda are the North Atlantic [Fehsenfeldet al., this issue]. analyzed in detail in accompanying papers [Merrill et al., One goal of the vertical profiling measurementprogram this issue; Moody et al., this issue]. In this paper as well as within NARE was to obtain ozone vertical profiles the two accompanying papers, the strong role of synoptic throughoutthe troposphereand lower stratosphereusing scale meteorological events will be noted. The ozone vertical profiles along with the analyzed meteorological fields provide an importanttool for understandingozone INOAA ClimateMonitoring and Diagnostics Laboratory, Boulder, behaviorin the troposphere.There are alsosome significant Colorado. limitationsto this approach. Ozoneprecursor gases are not 2NOAA GeophysicalFluid DynamicsLaboratory, Princeton, New Jersey. measured with the ozone profile. Useful tracers of 3GraduateSchool of Oceanography,University of RhodeIsland, anthropogenic influence such as carbon monoxide are also Narragansett. not obtained from the balloon sounding. Water vapor 4Departmentof EnvironmentalSciences, University of Virginia, amount is measured with the ozone concentration, so this Charlottesville. 5IzafiaBaseline Observatory, Tenerife, Canary Islands, Spain. doesprovide a useful but not alwaysunambiguous indicator 6NOAAAeronomy Laboratory, Boulder, Colorado. of air parcelhistory. In somecases the ozoneprofile and 7CooperativeInstitute for Researchin the EnvironmentalSciences meteorologicalinformation are sufficient to make a strong (CIRES), Universityof Colorado,Boulder. case for a natural source (the stratosphere)for the observed 8RosenstielSchool of Marineand Atmospheric Sciences, University elevated ozone amounts. In other instances a source is not of Miami, Miami, Florida. 9Departmentof Physics,University of Colorado,Boulder. easily identified and, in particular, the anthropogenic contribution may be suggestedby the airflow to the site but cannot be unambiguously identified with this data set alone. Copyright1996 by the AmericanGeophysical Union. The time historyof the ozoneprofiles from the periodsof Paper number 96JDO 1713. near daily soundingsis displayed in the time-height cross 0148-0227/96/96JD-01713509.00 sections. Mean characteristics for each site and season are

29,179 29,180 OLTMANS ET AL.: SUMMER AND SPRING OZONE PROFILES OVER THE NORTH ATLANTIC

km. Temperature,humidity (to -12 km), and pressureare obtainedfrom the accompanyingVaisala radiosonde. Data are sampledevery 8 s. The sensorgives a 1/e responseto a step change in ozone in 20 s, which translates to a vertical :flavik. resolutionof approximately100 m. The normaldata product and the one which will be used here is a 250-m height average.The sondeaccuracy is 10% for troposphericmixing ratiosexcept in the caseof very low mixingratios (<10 ppb) when the accuracymay be degradedto 15%. Most of the launches were done in the morning local time which correspondsto 0900-1500 UT. The trajectoriesused in this work are isentropic10-day Race back trajectories. The trajectory model [Harris and Kahl, 1994] uses gridded analysesproduced by the European _,Szores _ Centre for Medium-Range Weather Forecasts (ECMWF). The trajectoriesdo not pinpointthe origin of an air parcelbut ratherare indicativeof the airflow path underthe assumption of adiabatic displacementof the air parcels. For most of thesesites in the mid-Atlantic, upper air observationsin the surrounding region are sparse so that even close to the Figure 1. Map of the North Atlantic region showing locations where ozonesonde launches were made. trajectorystarting point there is significant uncertaintyin the calculated representationof the air parcel motion. Under certaincircumstances, such as when a synopticscale front is encountered,the flow is probably not well resolved in the used to summarize the numerous soundingsand for comparison of the ozone behavior at the various locations. gridded analysis and the path representedby the trajectory Severalrepresentative ozone "events" are investigatedat the may be suspect. We routinely calculatethe trajectoriesback 10 days, which in some cases gives an unwarranted varioussites in relationto the changesin transportto the sampling location. impressionof our knowledge of the air parcel motion. The 10-day period is chosen primarily for computational uniformity when a large number of casesare being studied. Measurements

The ozone vertical profiles in this study were obtained Time-Height Cross Sections from balloon-borne ozonesondes. The ozonesonde consists The time-height cross sections of ozone mixing ratio of an ozonesensor (the electrochemicalconcentration cell), a displayed in Plates 1-6 summarize the profile information noncontaminatingair pump, an electronic interface board, during the periods of intensive observations and dramatize and a model RS-80 Vaisala radiosonde. The electrochemical the kinds of changesseen throughoutthe troposphere. We concentrationcell (ECC) sensor[Komhyr, 1969] is widely first discuss the results from August 1993 when used and has been extensivelytested and comparedwith measurementswere carried out simultaneouslyat Iceland, the other techniques[Komhyr et al., 1995]. The principleof Azores, and Bermuda (Plates 1-3). In Iceland (Plate la) the measurementis the reactionof ozone with potassiumiodide most striking feature is the large changes with time in in solution which producesa small current in the cell. This tropopause height and hence the varying depths to which current is quantitatively proportional to ozone with one large ozone mixing ratios extend. The tropopauseis marked ozone molecule producing a flow of two electrons. This in each cross section with the thick white line. The reaction is not specific to ozone, so that other oxidants and tropopause is determined from the temperature profile reductantsalso produce a response.The primaryinterfering following the standard meteorological convention (the nonozoneoxidant is NO2 which respondswith about 10% the altitude where the average lapse rate first falls below 2øC/km efficiency of ozone. In the atmosphericregime considered for at least 2 km). The dips to lower altitude of high ozone here,the NO2 concentrationis in the sub-ppbrange and thus concentration(e.g., Julian days 239-241 (August 27-29)) do is not a significantinterferent. The primary atmospheric not represent penetration of these large values into the reductantaffecting sensorresponse is SO2 which acts with tropospherebut simply changing tropopauseheight. The equal efficiency but in the opposite direction to ozone. periods of largest ozone mixing ratio in the middle and lower Again SO2 concentrationsare in the sub-ppbrange. The troposphere are all associated with the highest, not the Teflon pump of the sondeis noncontaminatingand doesnot lowest, tropopauseheights. This is seen on Julian days 219, destroyozone. Since ozone is measuredquantitatively by 230, 236, and 243 (August 7, 18, 24, and 31), where ozone the ECC ozonesonde,it is not necessaryto "normalize"the amounts between 40 and 50 ppb extend down to altitudes integratedozone profile amount to an independentmeasure below 2 km. Except for Julian day 243 (August 31), these of column ozone such as the Dobson or Brewer were days of relatively larger ozone concentration at the spectrophotometers.This is important since other column, surface.Although most of the tonguesof large ozonemixing ozone measurements are not available at all of the sites. The ratio stretching from the stratospheredown to the lower small instrumentpackage (-2 kg) is carried aloft on a small troposphereare associatedwith very dry air (e.g., Julian days helium or hydrogen-filled rubber meteorological balloon 236 and 237 (August 24 and 25)), Julian day 230 (August (1200 g) which normally reaches altitudes in excess of 30 18) is not particularly dry (Plate lb). It is also clear that OLTMANSET AL.' SUMMERAND SPRINGOZONE PROFILES OVER THE NORTHATLANTIC 29,181

3.6E-01 Ozone Mixing Ratio over U.S. NAS Keflavik, Iceland 3.BE-01 15 3.0E-01 2.8E-01 a) 2.6E-01 2.4E-01 2.2E-01 2.0E-01 1.8E-01 1.6E-01 1.4E-01 1.2E-01 1.0E-01 9.0E-02 8.0E-02 ?.OE-02 6.0E-02 5.0E-02 4.0E-02 3.0E-02 2.0E-02 1.OE-02 8.0E-03 6.0E-03 4.0E-03 225 230 235 240 245 250 2.5E-03 Julian Day 1993 [ppmv]

Relative Humidity over U.S. NAS Keflavik, Iceland lOO 15 95 b) 90 85

õo

60 55

45 40

35 30

lo ' "• I. ! 5 210 215 220 225 230 235 240 245 250 o Julian Day 1993 [•]

Plate1. Time-heightcross sections at KefiavikAir Base,Iceland, for August1993 for (a) ozonemixing ratio(ppmv) and (b) relativehumidity (%). The tick marksfor theday denote the beginning of theday. The heavytick markdenotes the first dayof themonth. The arrowsalong the loweraxis are at the timeof thesounding. The white, solid line marks the tropopause height based on the temperature profile using the standardmeteorological convention to definethe tropopause. 29,182 OLTMANS ET AL.' SUMMER AND SPRING OZONE PROFILES OVER THE NORTH ATLANTIC

3.6E-01 Ozone Mixing Ratio over Lages, Azores 3.3E-01 15 3.0E-01 2.8E-01 2.6E-01 2.4E-01 2.•E-01 2.0E-01 1.BE-01 1.6E-01 1.4E-01 1.2E-01 1.OE-01 9.0E-02 8.0E-02 7.0E-02 6.0E-02 5.0E-02 4.0E-02 3.0E-02 2.0E-02 1.0E-02 8.0E-03 6.0E-03 o 4.0E-03 21o 215 220 225 2a0 235 240 245 25O 2.5E-03 Julian Day of 1993 [ppmv]

Relative Humidity over Lages, Azores lOO 15 95 9o b) 85

8o 75

5o 45

35 3o 25

to o 5 21o 215 220 225 230 235 240 245 250 o Julian Day of 1993 Plate 2. As Plate 1 but for Lages Air Base, Azores,in August 1993.

relatively smaller ozone amounts(<40 ppb) are found in the In the Azores (Plate 2a) during August, the tropopauseis middle and upper troposphere. The humidity in these usually between 12 and 14 km and larger ozone amounts regions is invariably higher, implying this air had a source (>90 ppb) extend into the upper troposphere (e.g., Julian near the surface. Through the month there is a gradual days 216 and 222 (August 4 and 10)). At Bermuda (Plate 3a) decrease in the near-surface amounts, but this trend does not the tropopause is near 15 km, and there are also several seemto be presentin the midtroposphere. events where ozone amounts over 90 vvb extend into the OLTMANSET AL.: SUMMERAND SPRINGOZONE PROFILES OVER THE NORTHATLANTIC 29,183

3.6E-01 Ozone Mixing Ratio over Bermuda 3.3E-01 15 3.0E-01 2.8E-01 2.6E-01 2.4E-01 2.2E-01 2.0E-01 1.BE-01 1.6E-01 1.4E-01 1.2E-01 1.OE-01 9.0E-02 8.0E-02 7.0E-02 6.0E-02 5.0E-02 4.0E-02 3.0E-02 2.0E-02 1.0E-02 8.0E-03 6.0E-03 4.0E-03 •05 a 10 • 15 aa0 •5 •30 •35 •40 •45 2.5E-03 Julian Day of 1993 [ppmv]

Relative Humidity over Bermuda lOO 15 95 9o

lO •ug 5 205 210 215 220 225 230 235 240 245 o Julian Day of 1993 [•]

Plate 3. As Plate 1 but for BermudaNaval Air Stationin August1993. midtroposphere.Equally apparentat both these locationsis Azores). The regionsof smallerozone mixing ratios are the presenceof regionsof relatively small mixing ratios (30- always associatedwith higher humidity (Plates 2b and 3b). 40 ppb) near the tropopause. In some casesthese features The very strongcorrespondence between high humidityand extend from near the surfaceto near the tropopause(e.g., smallerozone amountsis consistentthrough the monthand around Julian day 227 (August 15) at both Bermuda and the through the entire depth of the troposphere. At both 29,184 OLTMANSET AL.: SUMMERAND SPRING OZONE PROFILES OVER THE NORTH ATLANTIC

3.6E-01 Ozone Mixing Ratio over Cape Race, Newfoundland 3.3E-01 15 3.0E-01 2.8E-01 2.6E-01 a O! _ 2.4E-01 2.2E-01 2.0E-01 1.BE-01 ),,, .'•,.. - 1.6E-01 1.4E-01 1.2E-01 1.OE-01 9.0E-02 8.0E-02 7.0E-02 6.0E-02 5.0E-02 4.0E-02 3.0E-02 2.0E-02 1.OE-02 8.0E-03 6.0E-03 Jul •. • _ •,--•• • • • Aug 4.0E-03 190 195 200 205 210 215 220 2•5 230 2.5E-03 Julian Day of 1991 [ppmv]

Relative Humidity over Cape Race, Newfoundland lOO 15 ' ' ' ' ' I ' ' ' • ' 95 9o b) 85

lo Jul ...... Aug o • ,[ , , , , ! , , , I 5 190 195 200 205 2!0 215 220 225 230 o Julian Day of 1991 [g] Plate 4. As Plate 1 but for Cape Race, Newfoundland, in July and August 1991.

Bermuda and the Azores the two periods, which show this this resultsin net destructionor production [Crutzen, 1988]. patternmost dramatically, occur very near the sametimes The backgroundozone values at Bermuda are smallestin the (Juliandays 227 and242 (August15 and30)) of the month. summer and the vertical depth of this layer reaches 2 km Abundant moisture in the lower troposphere(<2 km) much more consistently than at the other two locations. At during the summer sets the stage for significant ozone the surface,ozone displaysa diurnal signal with an afternoon photochemistry.The locallevels of NOx determinewhether minimum, just as has been observed at the tropical remote OLTMANS ET AL.: SUMMER AND SPRING OZONE PROFILES OVER THE NORTH ATLANTIC 29 185

3.6E-01 Ozone Mixing Ratio over Bermuda 3.3E-01 15 3.0E-01 2.6E-01 .) •.6E-01 Z.4E-01 2.2E-01 2.0E-01 1.BE-01 1.6E-01 1.4E-01 1.2E-01 1.OE-01 9.0E-02 8.0E-02 7.0E-02 6.0E-02 5.0E-02 4.0E-02 3.0E-02 2.0E-02 1.OE-02 8.0E-03 6.0E-03 o 4.0E-03 lOO 105 110 115 120 125 130 135 140 2.5E-03 Julian Day of 1993 [ppmv]

Relative Humidity over Bermuda lOO 15 95 90

b) 85

55 50

20 15

lO o 5 lOO 105 110 115 120 125 130 135 140 o Julian Day of 1993 [•.] Plate •. As Plate 1 but for Bermudain April and May 1993.

locations of Barbados and Samoa [Oltmans and Levy H, awaits the simultaneous measurement of ozone, water, CO, 1992; Oltrearis,1981 ]. This behavior combinedwith the low andNOx in this layer at a locationlike Bermuda. backgroundlevels of surfaceNO x reportedby Dickersonet For Cape Race (Plate 4) the intensiveobservations cover 2 al. [1995] and the simulationsof ozone chemistryreported weeks at the end of July and early August 1991. The upper by Kasibhatlaet al. [this issue],strongly suggests that net troposphereshows that the stronginfluence of the very large destructiondominates. Definitive proof of this hypothesis ozone amounts in the overlaying stratosphereand ozone 29,186 OLTMANSET AL.' SUMMERAND SPRING OZONE PROFILES OVER THE NORTH ATLANTIC

3.6E-01 Ozone Mixing Ratio over Lages, Azores 3.3E-01 15 3.0E-01 2.õE-01 ' ' • , • •t' 2.6E-01 2.4E-01 2.2E-01 2.0E-01 •,,• •"' ':'. % :'.. I 1.BE-01 1.6E-01 1.4E-01 "•;'•'"'. .T..:' •''• 1.2E-01

...... ,. 1.0E-01 9.0E-02 :•. 8.0E-02 7.0E-02 6.0E-02 5.0E-02 4.0E-02 3.0E-02 2.0E-02 1.OE-02 8.0E-03 • . •-,- ' 6.0E-03 o 4.0E-03 2.5E-03 18o 125 130 135 140 145 150 155 160 Julian Day of 1994 [ppmv]

Relative Humidity over Lages, Azores lOO 15 95 90

õo

65 6o

45 4o 35 3o

ß ,

., 15 May '• IO o 5 o 1E0 125 130 135 140 145 150 155 160 Julian Day of 1994 Plate 6. As Plate 1 but for the Azores in May 1994.

amountsof-100 ppb reach into the midtroposphere(Plate Spring profile measurementswere done in Bermuda in 4a). The extension of the large ozone amounts into the April/May 1993 (Plate 5) and the Azoresin May 1994 (Plate midtroposphereis associatedwith both high and low 6). At Bermuda during the second half of May, the tropopauses. Several periods when smaller ozone mixing soundingsare lessfrequent, so that someof the featureslook ratios extend from the surface into the upper tropospherecan more persistentthan they probablyare. In the Azoresthere also be seen. As at other locations, these are associated with are significantlylarger variationsin the tropopauseheight higher relative humidity (Plate 4b). than during the summer. The regions of very large ozone OLTMANSET AL.: SUMMERAND SPRINGOZONE PROFILES OVER THE NORTHATLANTIC 29,187

15 15 a) b) I I I I

, I I

Iceland Cape Race

i I 25 50 75 100 125 150 25 50 75 100 125 150 OZONE MIXING RATIO (ppbv) OZONE MIXING RATIO (ppbv)

15 15 c)

i I I ß

l•l I' t

• - '< 5-

I•l ß I-t Bermuda Azores

25 50 75 100 125 150 25 50 75 100 125 150 OZONE MIXING RATIO (ppbv) OZONE MIXING RATIO (ppbv)

Figure2. Troposphericozone mixing ratios at 1-kmintervals for (a)Iceland, August 1993; (b)Cape Race, July/August1991' (c) Azores, August 1993; (d) Bermuda, August 1993' and (e) Tenerife, August 1993. The solid dot is themean; the bar is the median;the box is the inner50th percentile; and the whiskersare the inner90th percentile. 29,188 OLTMANS ET AL.: SUMMER AND SPRING OZONE PROFILES OVER THE NORTH ATLANTIC

15 I I I At Bermuda the less frequent profiles limit any generalization that can be made from Plate 5 on the frequencyand timing of the events. Two eventsin mid-April I• I ß I and the beginning of May appear to bring stratosphericair into midtroposphericlevels. Just prior to the mid-April I• I ß I I event, strong convectionmixes surfaceair to the tropopause. I I Iol I This particular case is be discussedlater. As at the Azores, surface amounts are significantly higher in spring than I--d Iß I I during the summer[Oltmans and Levy, 1992]. I--• Io I I The picture that emerges from the cross-sectional representationof the intensive ozone profile data in both the I--I oll I springand the summeris the connectionbetween large ozone I I o1• mixing ratiosand dry air in the middle and uppertroposphere with the large ozonevalues in the tropopauseregion. On the HIo H other hand, smaller ozone amounts at these altitudes are associated with high humidity and are often connected vertically with the surface. The timescale of changesfrom large to small ozone mixing ratios or water vapor amountsis « ol • of the order of a few days (somewhatlonger in the summer), which is similar to synoptic scale variations.

Average Behavior • Tenerife To look at differences between seasons and locations, we 0 use the box and whisker representationof the median, inner 25 50 75 100 125 150 50th, and inner 90th percentilesto show average behavior OZONE M••G •TIO (ppbv) and dispersionat 1-km incrementsfor each of the periods of observations (Figures 2a-2e). In addition to the four sites Figure 2. (continued) discussedin the previous section, profile data from Tenerife, Canary Islands, for August 1993 and May 1994 will be used amounts(>150 ppb) that dip below-10 km are generallyin to broadenthe scopeof the results. One of the most striking the stratosphere.At Bermuda(Plate 5a) theregions of large features of the summer average profiles (Figure 3) is the ozone mixing ratios below 14 km are all within the troposphere. In theAzores, because there are profiles almost every day 15 duringthe observingperiod, a clearpicture of the frequency ..... Iceland of the variationsat different levels is seen. At most levels, significantmaxima or minimaare seenevery 3-6 days,which X-. -'• AzoresCapeRace ,•". ./••.•'"" is consistentwith synopticscale meteorological influence on © ...... O Bermuda _/ ._/.'ø the changesin ozoneconcentration. Regions of largerozone amounts (>80 ppb) tend to extend downward from the tropopause to about 5-7 km. Below 5 km, down to about 2 km, thereare still significantperiods and regionswith ozone mixingratios >60 ppb. This is the caseespecially during the •- J.•Tenerife1•];•.,• _,, /• • •. o•-- latterhalf of May. The regionswith ozone>60 ppball have low humidity(Plate 6b) of <20%. Boththe drynessof the air ß and the verticalconnection of the large ozonemixing ratio regionwith the tropopausesuggest that the stratosphereplays . a majorrole duringthis time in loadingthe middleand upper tropospherewith ozone. During May there are fewer events in which smaller ozone mixing ratios are found in the middle and upper tropospherecompared to the summer. On Julian days 126, 127, and 137 (May 6, 7, and 17), columns of smallerozone mixing ratio and high humidity are present. The event of Julian day 137 (May 17) is one which extends upward all the way from the surface. After this the troposphereis dominatedalmost exclusively by air with ozonemixing ratios exceeding 60 ppb from the tropopauseto 0 • I I I I ' 2 km. The troposphereappears to be graduallyfilling with 25 50 75 100 125 150 ozonefrom aboveduring the latterhalf of themonth. During this time there are none of the events in which small ozone OZONE MIXING RATIO (ppbv) amountsand high humidity extend upward from the surface Figure 3. Averageozone profiles for eachof the placesand into the troposphere. times shown in Figure 2. OLTMANS ET AL.: SUMMER AND SPRING OZONE PROFILES OVER THE NORTH ATLANTIC 29,189 larger mixing ratios at all levels in the troposphere at Cape 15 Race. At the surface the mean (the solid dot) at all locations ..... Azores (Figure 2) is between 25 and 30 ppb except at Bermuda • ...... o Bermuda where it is slightly smaller (Figure 2d). At Cape Race there • Tenerife are surface events exceeding 50 ppb not seen at the other sites (Figure 2b). There is also a larger discontinuity seen at Cape Race between ozone amountsin the boundarylayer and the next higher layer. On average there is about twice as e•e much ozone in this layer at 1 km then at the surface (Figure 2b). This is consistentwith the effect on air flowing over the North American continent found by Berkowitz et al. [1995]. Tenerife shows a similar feature (Figure 2e) which may result from the influence of the nearby African continent on air flowing to the Canary Islands. During August the ozone mixing ratios above 4 km at the Azores, Bermuda, and Tenerife (Figure 3) are very similar up to the tropopausediffering by 5 ppb or less at each level. Given the high degree of variability seen at each site, this difference of less than 10% in the mean concentration at each level is somewhat unexpected. This may suggest that at these mid-North Atlantic Ocean sites during the summer, the broad influence of the dominant subtropical high-pressure regime also shapesthe ozone distribution over much of the region. It shouldbe noted, however, that each of the sitesis located quite differently relative to the center of this high- 25 50 75 100 125 150 pressuresystem. In the 3 to 7-km region in Iceland (up to OZONE MIXING RATIO (ppbv) near the tropopause),mixing ratios are generallyat least 10% smaller than the other island locations. At all locations there Figure 4. Average ozone profiles for May at the Azores (1994), Bermuda (1993), and Tenerife (1994). is a rapid drop of at least 50% in ozone concentration between 4 km and the surface. At Tenerife (Figure 2e), however, this drop in mixing ratio takes place primarily Azores clearly has less ozone than Bermuda above 3 km, while below 2 km. This is consistent with the results of Schmitt et during the summer, they were nearly identical. Of the three al. [1988] and Prospero et al. [1995], which show elevated island sitesat subtropicallatitudes, the Azoreshas the smallest ozone values persistingthrough the summer months at the mountain station at Izafia located on the island of Tenerife. difference between spring and summer, while Bermuda has significantlymore ozone in the spring. Becausethe tropopause Schmitt et al. [1988] suggest that these enhanced summer is higher in the summer, this means that at Bermuda the spring ozone amountscould be from the transportof ozone resulting and summertropospheric column amountsare quite similar, but from anthropogenicallyproduced emissions in Europe. This at the Azores there is about 50% more ozone in the summer was based on the similarity of the annual ozone cycle at column. Izafia with that at anthropogenically influenced sites in Europe and in contrastto the distinct spring maximum found Case Studies at Mauna Loa, Hawaii. Prospero et al. [1995], however, showed that the large summer concentrations were more A number of individual ozone profiles from both the closelyrelated to tracersof upper troposphericair which did summer and the spring periods are presentedin this section not have a signature of recent continental origin. A very in order to use the transport characteristics, and in some recent set of intensive ozone profiles obtained during June cases other meteorological information, to help understand and July 1995 (E. Cuevas, private communication, 1996) the features in the profile. shows that high ozone mixing ratios dominate the entire Summer events. In the period August 3-4, 1993, there are troposphereduring the early summer and the large values at profiles from each of the three subtropical sites: Bermuda, Izafia at 2.5 km are really an extensionof this. the Azores, and Tenerife. This early August event at The averagebehavior in spring (for these averagesthis is Bermuda is discussedin some detail by Merrill et al. [this primarily May data with a few profiles in late April and early issue]. At each location there is an enhanced ozone layer June)at the subtropicalisland sites is roughlysimilar (Figure4). (>100 ppb) in the middle or upper troposphere(Figures 5a, Tropopauseheights are severalkilometers lower in the spring 6a, 7a). For each location a back trajectory is shown for an than in the summer. At the Azores and Bermuda, there is more end point that arrives in the boundarylayer (Figures 5b, 6b, ozone in the boundary layer in the spring than the summer, 7b) and one that arrives in the enhanced ozone layer (Figures which is consistentwith enhancedphotochemical ozone loss in 5c, 6c, 7c). the summer. At Tenerife, on the other hand, May has lessozone At Bermuda the trajectory near the time of the profile in the 1 to 4-km region than August, reflecting the summer (1200 UT) shows strong cyclonic curvature and descent as enhancementat this altitude. At each site there is a region of the air parcel approachesBermuda at the 8-km level (Figure near constantor slightly increasingmixing ratio from 5 km to 5c). The parcel passes through a region of enhanced near the tropopause.Below 5 km, mixing ratiosdecline with an potential vorticity over the northeast coast of the United even sharperdecline in the lowest kilometer. During May the States 2 days prior to reaching Bermuda [Merrill et al., this 29,190 OLTMANSET AL.: SUMMERAND SPRINGOZONE PROFILES OVER THE NORTHATLANTIC

TEMPERATURE AND FROSTPOINT (C) issue]. In contrast,the low-level flow (Figure 5b) has -90 -60 -30 0 30 remained south of 20øN for most of its transit over the ocean. 15 Ozoneamounts are about20 ppbnear 1 km in the profile, a) andthe stronggradient below 3 km suggeststhat significant ozone is lost in the low levels. On August3 in the Azores(Figure 6a) a peakis seenat 7.5 km and a somewhatstronger one at 10 km. The trajectoryreaching the Azores at 8 km (Figure 6c) shows strong descent 6 days earlier over the East Coast of the UnitedStates and passesnear Bermuda about 4 daysprior. Bermuda Thispath indicates that the sameevent that brought the high August 3, 1993 ozone seen at Bermuda at the end of July (Plate 3a) also reachedthe Azores. The low-level flow (Figure6b) on this 1310 UT day appearsto comearound the northernedge of the high- pressuresystem after being over the ocean for at least 10 days. The low-altitudeamounts are not as low as Bermuda, however, where the trajectory stays much farther to the south. At Tenerifethe ozonepeak on August4 is near9 km (Figure 7a), which is well within the troposphere. The Ozone trajectorynear the time of the soundingpasses a bit eastof the Azores (Figure 7c) about 2 days before reaching Air Temp. Tenerife. The peak at Tenerifemay be part of the sameevent Frostpoint and flow that were responsiblefor the 10-km peaksin the o Azoresearlier. At 1 km over Tenerifethe airflow (Figure 0 25 50 75 100 125 150 7b) also appearsto bring air that reachesback to the North OZONE MIXING RATIO (p@bv) Americancontinent but with a much different path than at higherlevels. This air parcelpath from higherlatitude likely Figure 5. (a) Ozone mixing ratio, temperature,and frost- contributesto the 20 ppb larger amountsat 1 km at Tenerife point temperatureprofiles at Bermudaon August3, 1993. compared to the other two sites. (b) Trajectoriesfor Bermudaon August3 arrivingat 1 km (a level in the boundarylayer). (c) Same as Figure 5b but At each of these locationsthe middle to high-elevation arriving at 8 km (a level near the center of the enhanced ozone peaks are related to flow back toward the North ozone layer). Americancontinent. Limitationsin this type of trajectory

Bermuda•_C•;•• • • •1 '• August3,1993 /•• ' 60øN ß .•.. . 60øN...... •...... <.•...•..

30øN 30øN

_ 00 Dashed12 UT : •fi • 130• 100• 70• 40• 10• 170• 140• 110øW 80øW 50øW

3 4 5 6 7 8 9 0 1 2 3 4 5 6 7 8 9 10 Days from Bermuda Days from Bermuda

Figure 5. (continued) OLTMANSET AL.: SUMMERAND SPRINGOZONE PROFILES OVER THE NORTHATLANTIC 29,191

TEMPERATURE AND FROSTPOINT (C) that contributes to this in the midtroposphere is shown in -90 -60 -30 0 30 Figure 8. At 7 km on July 23, 1991, the ozone mixing ratio 15 exceeds 150 ppb (Figure 8a). The previous day a layer was centered2 km lower with a peak value of-95 ppb (see Plate a) 4a). By July 24 the profile (Figure 8c) has relaxed back to about 75 ppb from 2 to 12 km with a couple of narrow layers Azores with lower ozone content. The trajectories for July 23 and 24 (Figures 8b and 8d) show that flow by midday on July 23 August 3, 1993 is coming from the north and is associatedwith a large low- •10 1110 UT pressure system north of Baffin Island. The air parcels descend over 2 km in the 2 days before arriving at Cape Race. By July 24 the flow reaching Cape Race has become much more zonal, passing over the Pacific Ocean 3 days Ozone back. The high ozone peak on July 23 is also very dry Air Temp. (Figure 8a) with the relative humidity at all levels between5 Frostpoint and 8 km on July 23 at less than 5%. Strong cyclonic < 5 activity is not located directly over Cape Race but well upstreamin northern Canada. Although we do not have an ozone soundingfrom the location where the transfer of ozone from the stratosphere into the troposphere could actually occur, the pattern is similar to the August 3 casefor Bermuda (also discussedby Merrill et al., [this issue]). The rapidity with which changes can occur in the ozone profile over a 0 station is demonstratedin this example. The soundingsare 25 50 75 100 125 150 only 18 hours apart, but in this time the ozone mixing ratio OZONE MIXING RATIO (ppbv) goes from over 150 ppb to under 50 ppb at 7 km. This is Figure 6. (a) Similar to Figure5 but for the Azoreson consistent with the rapid change in the air parcel path as August3, 1993, and with trajectoryarrival at (b) 1 km and representedby the trajectories. (c) 7 kin. From each of these summer cases from four different locations it is clear that events which bring ozone from the analysisdo not let us unequivocallytie featuresat one station upper troposphere and stratosphere are relatively common, with measurementsat sitesupwind, but they do suggestthat and that these events have peak ozone amounts as large as high ozone seen over Bermuda may be related to similar those seen in the spring. These events contribute features seen a few days later in the eastern subtropical substantially to the ozone budget of the troposphere during Atlantic. this season. At Cape Race where these events are the most The higher averagetropospheric ozone amountsat Cape frequent the tropospheric ozone mixing ratios are also the Race were noted in a previous section. One type of event highest.

••• •l August3,1993 •-••• • 7 /•A•zoresf• c) 60øN 60øN

30øN 30øN • • 381mb • Solid 00 UT 331K • DashedI•UT• • 120øW 90øW 60øW 30øW 0øW 120øW 90• 60• 30• 0•

i , , i i i i , :: 0 1 2 3 4 5 6 7 8 9 10 3 4 5 6 7 8 9 10 Days from the Azores Days from the Azores

Figure 6. (continued) 29,192 OLTMANS ET AL.: SUMMER AND SPRING OZONE PROFILES OVER THE NORTH ATLANTIC

TEMPERATURE AND FROSTPOINT (C) at each of the three subtropicalisland locationsare used to -90 -60 -30 0 30 illustratethe flow patternsassociated with theseprofile 15 characteristics. a) The ozone profile of June 1, 1994, in the Azores showsa maximumof over 100 ppbbetween 5.5 and6 km (Figure9a). The humidity in this layer is under 10%, and in the cross section(Plate 6a), the ozonemaximum shows up asa "bull's- eye"which is a commonfeature of this representationof the Tenerife data. The trajectoriesfor this layer (Figure9c) and the August 4, 1993 surroundinglevels at 2.5 km and 7.5 km (Figures9b, 9d) showthat the flow is dramaticallydifferent in eachlayer. 1200 UT For the trajectoryending within the enhanced ozone layer at 5.5 km, air travels from northern Canada and descendsmore than 2 kin. This is quite similar to the trajectoryseen at Cape Race for the summer case discussedearlier and follows Ozone muchthe samepath (Figure 8b). At 2.5 km (Figure 9b) the Air Temp. air parcel remains over the Atlantic Ocean for most of the 10 Frostpoint daysrepresented by the trajectory. At 7.5 km (Figure 9c) and 9.5 km (not shown) the flow representedby the trajectories is basically zonal, extending back over the Pacific Ocean. This example,and many more like it which we have examined,shows that the layered structureof ozone (and water vapor) resultsfrom varyingtransport at different levels from different source (and sink) regions. The air 25 50 75 100 125 150 parcelswith enhancedozone (-100 ppb or greater) almost OZONE MIXING RATIO (ppbv) alwayshave very low humidity, come from higher latitudes Figure7. (a) Similarto Figure5 butfor Tenerifeon August and altitudes, and descend significantly in altitude. In 4, 1993, andwith trajectoryarrival at (b) 1 km and(c) 9 km. addition,the generallylow tropopauseheights in the region northof 60øN,through which many of the trajectoriespass, contributeto the potential for significanttransfer of ozone Spring Events. During the late spring (May) ozone rich stratosphericair into the troposphereas observedin mixing ratios in the middle and lower troposphereare these layers. somewhathigher than during the summer(compare Figures 4 A seriesof five profilesat Bermudaduring April 19-23, and 5). The upper troposphereoften has mixing ratios 1993 (Plate 3a), demonstratethe dramatic day-to-day greater than 100 ppb. Prominent layers are common in the changesthat occur in the troposphereand how various midtroposphereand lower troposphericamounts of 50 ppb meteorological factors influence the ozone distribution over a are relatively common. Examplesfrom April, May, and June particularsite. On April 19 the ozonemixing ratio smoothly

o Tenerife t 4, 199 60øN ...i.-. 60øN

30øN ...... •'"9•03 40øN20ONt"•...i•..!..•,,'s• :• ..i.-.'.11•326rnb.., I '•L'•x'-'x 300• I I DashedSolid 00lr2 UT.•UT 333• K 120øW90øW 60øW 30øW 0øW 120øW 80øW 40øW 0øW

3 4 5 6 7 8 9 10 3 4 5 6 7 8 9 Days from Tenerife Days from Tenerife Figure 7. (continued) OLTMANSET AL.' SUMMERAND SPRINGOZONE PROFILES OVER THE NORTHATLANTIC 29,193

TEMPERATURE AND FROSTPOINT (C) TEMPERATURE AND FROSTPOINT (C) -90 -60 -30 0 30 -90 -60 -30 0 30 15 15 Cape Race Cape Race c) July24, 1991 July 23, 1991 1810 UT 1200 UT

Ozone Air Temp. Frostpoint

25 50 75 100 125 150 25 50 75 100 125 150 OZONE MIXING RATIO (ppbv) OZONE MIXING RATIO (ppbv)

60øN 60øN

30øN 30øN

Solid 00UT i'7'••,• I Solid 00UT Dashed12UT i •) !• Dashed12UT 150øW 120øW 90øW 60øW 150øW 120øW 90øW 60øW

I I 0 1 2 3 4 5 6 7 0 I 2 3 4 5 Daysfrom CapeRace Daysfrom CapeRace Figure8. (a) Similarto Figure 5 butfor Cape Race on July 23, 1991and (c) forCape Race on July 24, 1991,with trajectoryarrival at 7 km for (b) July23, and(d) July24. 29,194 OLTMANS ET AL.' SUMMER AND SPRING OZONE PROFILES OVER THE NORTH ATLANTIC

ß .:

15•, i 8,D "•,1 379 SL•'•SC, E•8 . E•

Plate 7. GOES 7 infrared satelliteimages for (a) April 20, 1993, and (b) April 21, 1993. The reddercolors indicate the colder temperaturesand thus the higher cloud tops. OLTMANS ET AL.: SUMMER AND SPRINGOZONE PROFILESOVER THE NORTH ATLANTIC 29,195

TEMPERATURE AND FROSTPOINT (C) -90 -60 -30 0 30 I c b! ! June1,1994 15 i .12UT ' a) Azores June 1, 1994 0930 UT

Ozone 60øN Air Temp. Frostpoint

30øN_ - ...... i..... 5;7mb",2' ...... :.' / '"•%• ' 312 KC• 120 øW 90 øW 60øW 3 0øW 0øW

25 50 75 100 125 150 OZONE MIXING RATIO (ppbv) 0 1 2 3 4 5 Days from the Azores .. •)•'• •/•: June1;1994 . d)•)•.'"'" Y(•! Ju"ne1•, 1994

60øN 60øN ••

x / • • 2.5km ' 7.5'• .' 30øN • • '." 758mb.... 30øN•...... • ...... '•..... :½52•b' '•• 299•K I 319K 120øW 90øW 60øW 30øW 0øW 120øW 90 øW 60øW 30øW 0øW

I 0 I 2 3 4 5 6 7 0 1 2 3 4 5 Days from the Azores Days from the Azores Figure9. (a) Similarto Figure 5 butfor the Azores on June 1, 1994,and with trajectory arrival at (b) 2.5km, (c) 5.5 km, and (d) 7.5 km. 29,196 OLTMANS ET AL.' SUMMER AND SPRING OZONE PROFILESOVER THE NORTH ATLANTIC

TEMPERATURE AND FROSTPOINT (C) TEMPERATURE AND FROSTPOINT (C) -90 -60 -30 0 30 -90 -60 -30 0 30 15 15 Bermuda Bermuda c) a) April 20, 1993 April 21, 1993 1325 UT 1340 UT

10 10

A Ozone ¾ Air Temp. I Frostpoint

0 25 50 75 100 125 150 0 25 50 75 100 125 150 OZONE MIXING RATIO (ppbv) OZONE MIXING RATIO (ppbv)

60øN

30øN

170øW

i ! i I I I I 3 4 5 6 7 3 4 5 6 7 8 9 Days from Bermuda Days from Bermuda Figure 10. (a) Similarto Figure5 but for Bermudaon April 20, 1993 and (c) April 21, 1993, with trajectory arrivalat (b) 6 km for April 20 and (d) 9 km for April 21.

increasesfrom 50 ppb at the surfaceto about80 ppbnear the complex over Bermuda and the surrounding waters with tropopause. This relatively "unperturbed"profile looks cloud top temperaturesof about -60øC which are equivalent similar to the average profile for this season at Bermuda to the temperaturesat the height of the tropopause(Figure (Figure4). On April 20, Bermudais locatedin a regionof 10a). The ozone mixing ratio is extremely constantat 45 ppb intenselocal convection in associationwith a deepeninglow- from the surface to the tropopause and the atmosphereis pressure center at the surface. A GOES7 enhanced infrared nearly saturated over much of the same altitude range satellite image (Plate 7a) shows a mesoscale convective (Figure 10a). In the midtroposphere (6 km) the trajectory OLTMANS ET AL.: SUMMER AND SPRING OZONE PROFILESOVER THE NORTH ATLANTIC 29,197

TEMPERATURE AND FROSTPOINT (C) By the following day (April 21) the low center over -90 -60 -30 0 30 Bermuda had developed and moved in retrograde to the north and west of Bermuda, in conjunction with an upper level low 15 which was almost cut off at 500 hPa. The satellite image for a) April 21 (Plate 7b) shows Bermuda was in a cloud free Bermuda region. At Bermuda there is a 3-km-thick layer in the upper troposphere with ozone mixing ratios of -250 ppb April 23, 1993 (Figure 10c) and the mixing ratio is over 100 ppb down to 8 1350 UT km giving a 7-km-thick layer of enhanced ozone. The trajectory arriving at 9 km over Bermuda shows strong Ozone cyclonic curvature and a drop of over 2 km between the East Air Temp. Coast of the United States and Bermuda (Figure 10d) in the Frostpoint region where the upper level trough deepened. The region of subsidence associated with this upper level feature was upwind (to the north and west) of the mesoscale convection present on April 20. On April 22 the layer of large ozone mixing ratio was only 3 km thick and peak values are-140 ppb (see Plate 3a). By April 23, troposphericmixing ratios (Figure 11a) are mostly less than 60 ppb and values under 25 ppb are present near 11km. This region of small ozone amounts is relatively moist (RH>30% compared to <20% below and <10% above this layer). There is an important change in the trajectory at 11 km between April 22 and 23 that provides evidence for the source of the small ozone 0 25 50 75 100 125 150 amountsat 11 km. On April 22 the trajectory to Bermuda at OZONE MIXING RATIO (ppbv) 1200 UT (Figure l lb) stays north of about 25øN. On April 23, however, the air parcel comes from south of 20øN to Figure 11. (a) Similar to Figure 5 but for Bermudaon April Bermuda(Figure 1lc) in just 1 day. Equally important, there 23, 1993, with trajectoryarrival at 1 lkm for (b) April 22 and is a bank of convective clouds across southern Mexico (the (c) April 23. northern edge of this band can be seen in Plate 7) in close proximityto the air parcelpath as definedby the trajectory. approachesBermuda from the southafter passingover the The air parcel also passes over southern Florida where Caribbean and northern Mexico (Figure 10b) a few days anotherarea of convectiveactivity is presenton April 22. earlier. The soundingwas made in a fairly heavy rain The small ozonemixing ratio (-20 ppb) at 11 km suggests shower, which is another indication that active convection that the air would have to come from the subtropicalor was present. tropical marine boundary layer especially since some

• b) •-• April22, 1993 tl c) •-4 -'-April23, 1993 60ON(• ^_ /12UT• 60øN

238mb•,'X,•• • • 30øN N.• • .•7mb Bermuda , • • 170øW 14'0øW1i0øW 80ow 50øw

3 4 5 6 7 8 0 1 2 3 4 5 6 7 Days from Bermuda Days from Bermuda Figure11. (continued) 29,198 OLTMANS ET AL.: SUMMER AND SPRING OZONE PROFILESOVER THE NORTH ATLANTIC entrainment of air with larger mixing ratios is likely. For There are also some significant differences between lower example, at Barbados, which is at 13øN, the average surface and upper troposphericbehavior. At the lowest levels in the ozone during the last half of April 1993 was about 18 ppb troposphere(<2 km) during the summer, ozone amounts are with amounts varying between 16 and 20 ppb. We generally low (<30 ppb) and the stronggradient below -4 km hypothesizethat the small ozone mixing ratio at 11 km over is evidence that ozone is being destroyed, probably Bermuda is a remnant of air mixed from low in the photochemically. At middle and upper tropospheric troposphere,likely from a convectiveevent earlier in the air altitudes, on the other hand, summer ozone amounts are parcel history. Thus it appearsthat both peaks(as seenin the similar or only slightly less than those in spring. June Azores profile) and relative minima, as seen here, can In a rare instance in which an ozone profile was obtained be caused by abrupt changes of air parcel source with within a precipitating convective cell, we saw the mixing of altitude. This sequence of profiles over a 4-day period air with relatively smaller, near-surfaceozone concentrations shows the strong modulating effect that meteorological all the way to the tropopause. At each location studied,both changes play in determining the ozone distribution in the in the summer and in the spring, the time-height cross troposphere. sectionsshow features that appear to extend vertically from At Tenerife a profile on May 25, 1994 (Figure 12a), has the surface to near the tropopausein which ozone amounts ozone mixing ratios over 100 ppb from 7 km to the are small and the humidity high (see, for example, Julian day tropopauseat--13 km. Below 5 km, ozone drops markedly 241 in both Plate 2 and Plate 3). These events last 2-3 days from-75 ppb to <20 ppb at the surface. A series of and seem to be more coherent in their vertical structure than trajectories on May 25 from 11 km to 1 km, in 2 km the larger ozone concentrationevents that extend downward increments(Figure 12b-12g), demonstratesthe consistencyin from above. Such features (small ozone mixing ratios and the behavior with altitude of the flow regimes in explaining high humidity through much of the troposphere)could occur the structureof the ozone profile. At the lowest two levels not only by mixing immediately above the stationbut also by (Figures 12f and 12g), air parcels reaching Tenerife have transport of air that had been mixed earlier by convective spentthe entire previous 10 days over open ocean(keeping activity. in mind the limitations of trajectories out to this length of For all of the sites studied here, except Iceland, the time) and have likely experiencedsignificant photochemical summer also has numerous events where ozone mixing ratios ozone destruction as discussed earlier. At the 5 to 11-km attain 75 ppb deep into the troposphereand mixing ratios levels there is a coherent pattern where flow around a deep >100 ppb are seen fairly often between 7 km and the troughover the easternUnited Statesreaches Tenerife in 3-5 tropopause. These high ozone events usually extend days. As with the other casespresented here and two other downward from the tropopause region. There are also more detailed studies[Moody et al., this issue;Merrill et al., numerous local maxima of high ozone in the cross section this issue], the dramatically elevated ozone amountsin the middle and upper troposphereare associatedwith strong cyclonicactivity that likely producesexchange between the TEMPERATURE AND FROSTPOINT (C) stratosphereand the troposphere. -90 -60 -30 0 30 15 Discussion and Conclusions Tenerife a) From an examination of several sets of intensive ozone May 25, 1994 1200 [IT profiles obtained over the past several years and the application of air parcel trajectory information, we have developeda basic picture of the way in which meteorological conditions (systems) influence the tropospheric ozone distribution over the North Atlantic Ocean. Several of the important patterns derived for surface ozone behavior, primarily at Bermuda [Oltmans and Levy, 1992; Moody et Ozone al., 1995], can be extendedto other parts of the troposphere. As was seen in the surface observations, larger ozone Air Temp. amounts (in most cases in several kilometer-thick layers) in Frostpoint the middle and upper tropospherewere associatedwith dry air moving from higher latitudes and altitudes to the observing site. Although none of the trajectories we examined show air coming all the way from the tropopause to the lower troposphere,the example from May 25, 1994, at Tenerife (Figure 12b-12e) shows that air parcels at progressively lower levels have come from layers several kilometers above which in turn have received air from layers even higher in elevation. At the time of an individual ozone 25 50 75 100 125 150 profile the transportabove the station is often not so coherent as in the May 25 example at Tenerife. Under the circum- OZONE MIXING RATIO (ppbv) stancesof a less coherent flow pattern over the station, only Figure 12. (a) Similar to Figure 5 but for Tenerife on May the flow in a restricted layer may be arriving from a region 25, 1994, and with trajectory arrival at (b) 11 km, (c) 9 km, where active exchange with layers above has taken place. (d) 7 km, (e) 5 km, (f) 3 km, and (g) 1 km. OLTMANS ET AL.: SUMMER AND SPRINGOZONE PROFILESOVER THE NORTH ATLANTIC 29,199

••'-'• •'•• % ? Tenerife o Solid 00UT/ May 25, 1994 Dashed 12 UT 60øN •• 60øN 40ON . .• 40øN/ ...... 20øN"'..ii•, ..•..• ...i.....i... I 20øN • 120 øW 80 øW 40 øW 0øW 120 øW 80 øW 40 øW 0øW

• 0 1 2 3 4 5 6 7 8 9 10 ::: 0 1 2 3 4 5 •; 7 8 9 1'0 Daysfrom Tenerife Daysfrom Tenerife d.,!,...... '"'*•""••,• • • •f/•t ' e)/•• • •--• 60øNii •• 60øN 40ON 40øN [ '...... ß...... 20ON 20øN

120 øW 80 øW 40 øW 0øW 120 øW 80 øW 40 øW 0øW

ß• 0 1 2 3 4 5 6 • 8 9 10 = 0 I 2 3 4 5 6 7 8 9 10 • Daysfrom Tenerife Daysfrom Tenerife

60ON 0øS....•..... !..:•...... •1111'."'"'"'f• iii!f. 42001•rt _.•. "'••'•-= 120øW 80øW 40øW 0øW 120øW 80øW 40øW 0øW

• 0 I 2 • 4 5 !• 7 8 • 10 = 0 I 2 3 4 5 6 7 8 9 10 Days from Tenetire Days from Tenetire

Figure 12. (continued) 29,200 OLTMANS ET AL.: SUMMER AND SPRING OZONE PROFILES OVER THE NORTH ATLANTIC

that appear to be a result of varying air parcel origin with Browell, E. V., M. A. Fenn, C. F. Butler, W. B. Grant, R. C. Harriss, altitude that gives a layered structureto individual profiles. and M. C. Shipham, Ozone and aerosol distributions in the summertimetroposphere over Canada, J. Geophys.Res., 99, 1739- These layers of high ozone are also regions of lower relative 1755, 1994. humidity in the profile. Trajectories calculated to arrive at Crutzen,P. J., Troposphericozone: An overview,in Tropospheric the elevation of the peak ozone concentration show Ozone,edited by I.S.A. Isaksen,pp. 3-32, D. Reidel,Norwell, Mass., distinctive characteristicsin the transportpath that strongly 1988. implicate the upper troposphere and stratosphereas the Dickerson,R. R., B. G. Doddridge,P. Kelley, and K. P. Rhoads,Large- scalepollution of the atmosphereover the remote Atlantic Ocean: source of the elevated ozone concentrationlayer. These Evidencefrom Bermuda, J. Geophys.Res., Tee, 8945-8952,1995. characteristicsare strong descent of air parcels prior to Fehsenfeld,F. C., A. Volz-Thomas, S. Penkett, M. Trainer, and D. D. arrival at the altitude of the elevated concentrationlayer, a Parish,North Atlantic RegionalExperiment (NARE) 1993 summer trajectory path that extends back to the region of an upper intensive:Foreword, J. Geophys.Res., this issue. Harris,J. M., and J. D. W. Kahl, Analysisof 10-dayisentropic flow level trough or cutoff low (typically associatedwith regions patternsfor Barrow, Alaska: 1985-1992, J. Geophys.Res., 99, of high isentropic potential vorticity [see Moody et al. this 25,845-25,855, 1994. issue; Merrill et al., this issue]), and trajectories that often Kasibhatla,P.S., H. Levy II, A. Klonecki, and W. L. Chameides,A come from higher latitudes, often north of 60øN, where three-dimensionalview of the large-scaletropospheric ozone tropopauseheight is generally lower. distributionover the North Atlantic Ocean during summer, J. Geophys.Res., this issue. There are many profiles or portions of a profile when Komhyr,W. D., Electrochemicalcells for gasanalysis, Ann. Geophys., ozone is not greatly elevated above the median value. We 25, 203-210, 1969. have not systematically looked at the trajectories for every Komhyr,W. D., R. A. Bames,G. B. Brothers,J. A. Lathrop,and D. P. ozone profile throughout the troposphericheight range. We Opperman, Electrochemical concentration cell ozonesonde performanceevaluation during STOIC 1989, J. Geophys.Res., Tee, have looked at some of these cases, however, and find that 9231-9244, 1995. the flow is more zonal, the altitude changesof the air parcel Merrill, J. T., J. L. Moody, S. J. Oltmans,and H. Levy II, Meteorological are not so marked, nor do they show strong descent. Ozone analysisof troposphericozone profiles at Bermuda,J. Geophys.Res., amounts of 50-75 ppb are often associated with these this issue. Moody, J. L., S. J. Oltmans,H. Levy II, and J. T. Merrill, Transport conditions through much of the troposphere above 3 km. climatology of tropospheric ozone: Bermuda, 1988-1991, J. The source of this "background" concentration cannot be so Geophys.Res., Tee, 7179-7194, 1995. readily assessed from the analysis carried out here, but Moody, J. L., J. C. Davenport,J. T. Merrill, S. J. Oltmans,D. D. Parish, several points can be made. First, cases of elevated ozone J. S. Holloway, H. Levy II, G. L. Forbes, M. Trainer, and M. Buhr, layers are common both in the spring and in the summer. Mechanismfor transporting0 3 overthe western North Atlantic: A casestudy for August24-29, 1993, J. Geophys.Res., this issue. Such layers make a significant contribution to the "average" Oltmans,S. J., Surfaceozone measurementsin clean air, J. Geophys. profile ozone levels. It is likely that as these 2 to 4 km-thick Res., 86, 1174-1180, 1981. layers are mixed horizontally and vertically as the air is Oltmans,S. J., and H. Levy II, Seasonalcycle of surfaceozone over the transported over time and space that they are an important western North Atlantic, Nature, 358, 382-384, 1992. Oltmans, S. J., and H. Levy II, Surface ozone measurementsfrom a contribution to the "background" condition [Browell et al., globalnetwork, Atmos. Environ., 28, 9-24, 1994. 1992, 1994]. Events in which higher ozone from the Prospero,J. M., R. Schmitt,E. Cuevas,D. L. Savoie,W. C. Graustein, boundary layer (<2 km) could be traced to higher levels were K. K. Turekian, A. Volz-Thomas, A. Diaz, S. J. Oltmans, and H. not found. There were, however, events where lower ozone Levy II, Temporalvariability of summertimeozone and aerosolsin amounts in the middle and upper troposphereappeared to be the free troposphereover the eastemNorth Atlantic, Geophys.Res. Lett., 22, 2925-2928, 1995. linked to the lower boundary layer amounts, probably Schmitt,R., B. Schreiber,and I. Levin, Effects of long-rangetransport through convective mixing. We plan to carry out additional on atmospherictrace constituentsat the Baseline Station Tenerife intensive sounding campaigns at Bermuda, the Azores, (CanaryIslands), J. Atmos.Chem., T, 335-351, 1988. Tenerife, and Newfoundland. The focus in these intensives will be to make profile measurementsat near daily frequency J.M. Harris,B.J. Johnson,J.A. Lathrop,S.J. Oltmans,and M.S. at all of the sites simultaneously. In addition, aircraft O'Neill,NOAA ClimateMonitoring and Diagnostics Laboratory, 325 campaigns,in which several tracers and key ozone precursors Broadway,R/E/CG1, Boulder,CO 80303. are to be measured, will be carried out between the eastern H. Levy,NOAA GeophysicalFluid Dynamics Laboratory, Princeton, NJ 08542. United States and Bermuda in the spring 1996. The strong J.T.Merrill, Graduate School of Oceanography,University of Rhode constraintsplaced on the ozone distribution by the profile Island,Narragansett, R102882. measurements and the additional planned chemical J.L. Moody,Department of EnvironmentalSciences, University of constituent measurements should lead to further progress in Virginia, Charlottesville,VA 22903. E. Cuevas,Izafia Baseline Observatory, Tenerife, Canary Islands, understanding the controlling mechanisms in the ozone Spain. budget over the North Atlantic. M. Trainer,NOAA AeronomyLaboratory, Boulder, CO 80303. M.S. O'Neill, Cooperative Institute for Research in the References EnvironmentalSciences (CIRES), University of Colorado,Boulder, CO 80309. Berkowitz,C. M., K. M. Busness,E.G. Chapman,J. M. Thorp,and R. J.M. Prospero,Rosenstiel School of Marine and Atmospheric D. Saylot,Observations of depleted ozone within the boundary, layer Sciences,University of Miami, Miami, FL 33149. of thewestern North Atlantic, J. Geophys.Res., lee, 11,483-11,496, H. V6mel, Dept. of Physics,University of Colorado,Boulder, CO 1995. 80309. Browell,E. V., C. F. Butler,S. A. Kooi, M. A. Fenn,R. C. Harriss,and G. L. Gregory,Large-scale variability of ozoneand aerosols in the summertimeArctic and subarctic troposphere, J. Geophys.Res., 97, (ReceivedJune 19, 1995; revisedApril 29, 1996; 16,433-16,450, 1992. acceptedMay 15, 1996.)