First profile measurements of tropospheric BrO - Wiley Online Library

GEOPHYSICAL RESEARCH LETTERS, VOL. 27, NO. 18, PAGES 2921-2924, SEPTEMBER 15, 2000

First Profile Measurements

of Tropospheric BrO

Fitzenberger,R.•, H. BSsch •, C. Camy-Peyret e, M.P. Chipperfield 3, H. Harder•'4'• U. Platt• B.-M Sinnhuber 3 T Wagner • andK. Pfeilsticker 1,• Abstract. TroposphericBrO profiles (about 0.6+0.2 ppt, and 2.0-t-0.8ppt at profile maximum) were mea-

tous in the free troposphere(about 1 - 2ppt, or 1-

sured for the first time.

Thesestudiesreliedon: (1) a comparisonof collocated ground-based, balloon-borne,and satellite-borne(Global Ozone Monitoring Experiment, GOME) BrO ob-

Our measurements

add new

information to recent speculations - based on indirect

evidence- of BrO possiblybeing ubiquitousin the free

2x10•3 molecules/cm 2 in total atmospheric column).

troposphere[Harder.et al., 1998; Friefi et al., 1999; servations,where discrepanciesbetween the integrated Van Roozendael et al., 1999; Pundt et al., 2000]. Our stratosphericBrO profiles (measuredfrom balloons), study relies on a detailed comparison of BrO slant and the total atmosphericBrO column(measuredfrom columndensities(BrO-SCD) measuredin the tropos- the groundand/or satellite)werefound [Harder et al., pherefrom the LPMA/DOAS (Laboratoirede Physi- 1998; Van Roozendael et al., 1999;Pundt et al., 2000], que Mol6culaire et Applications and Differential Op- and (2) a comparisonbetweenmeasuredground-based tical AbsorptionSpectroscopy) gondolaby direct Sun BrO-SCD, and model (SLIMCAT) predicted stratosabsorption,and BrO-SCD valuessubsequentlymeasu- pheric BrO profileswhere the model - accountingfor red in the lowermoststratosphereduring balloonascent. 20 ppt of total stratosphericinorganicbromine [e.g.,

The differencein total atmosphericBrO-SCDs measured in the troposphere and lowermost stratosphere- af-

Schauffler et al., 1998; Wamsley et al., 1998; Harder

ter a suitable correctionfor the changein BrO due to photochemistryand the observationgeometry- is then

mosphericBrO column amount measuredaround local

attributed to tropospheric BrO.

1.

et al., 2000]- significantlyunderpredictedthe total at-

noon[Frie• et al., 1999]. Here, we report on first BrO profiles measured by direct Sun DOAS spectroscopy that were conducted from aboard the azimuth-angle-controlledLP-

MA/DOAS balloonpayload.

Introduction

Recent spectroscopicobservationsof troposphericBrO haveclearlyestablishedthe role inorganicbromineplays

2. Methodology

and Measurements

in the destructionof planetaryboundarylayer (PBL) ozoneduring polar spring. The detectionof PBL BrO

The balloon-borneUV/vis DOAS spectroscopy techni-

[Hausmannand Platt, 1994],aircraft-bornebackscattering UV/vis absorptionmeasurements [McElroyet al., 1999],and/or satellite-borne spectroscopy [Wagnerand Platt, 1998].Also,our grouprecentlyrecordedsizeable amountsof BrO [up to 80 ppt] emittedfrom salt pans locatedat the Dead Sea,Israel [Hebestreitet al., 1999].

2000],and Harderet al. [1998,2000]. The small atmosphericabsorptionof BrO (typical op-

que aims at the profile measurementsof BrO, OC10, - with mixing ratios up to 100 ppt - was based on eit- NO2, 03, 04, IO, and someother gases.More detailsof her ground based long path absorption measurements the techniquecan be foundin Ferlemannet al. [1998, tical densities are •

10-3 for the UV vibrational ab-

sorption bands 4-0 at 354.7 nm, and 5-0 at 348.8 nm of

the A(2•r)•-X(•r) electronic transition)are spectros-

copically monitored by applying the DOAS-technique Speculationsas to whether BrO ubiquitously occurs [Platt, 1994]to directSunspectracollectedwith a Sun in the free tropospherenot just during these speci- tracker mounted on the azimuth-angle-controlledLPfic BrO events - and thus influencesthe chemistry of MA/DOAS balloongondola.Direct Sun tracking- rattropospheric ozone - was prompted by a series of re- her than using a diffuserfor diffusesky and direct Sun cent indirect evidencefor BrO being possiblyubiqui- light collection- is essentialfor a thoroughdirect detection of troposphericBrO, since it establishesa well

definedlight path in the troposphere[e.g., Harder et al., 2000].The Sunspectraare analyzedwithin a pressurized,and thermally stabilised,grating spectrometer x Institut fiir Umweltphysik, Universityof Heidelberg, [Ferleraann et al., 2000].This avoidsunwantedspectral Heidelberg, Germany shiftsand squeezesdue to the changingrefractiveindex 2 LPMA, CNRS, Universit6Pierreet Marie Curie, Paris, causedby changingambient pressures,as well as therFrance 3Schoolof the Environment,Universityof Leeds,Leeds, mally inducedmisalignmentsof the optical set-up. The UK

4nowwith MeteorologyDepartment,PennstateUniversi-

precision (+ 4%), andaccuracy(-t-12%plus+5 x 10TM molecules/cm 2) ofourstratospheric BrO-SCDmeasure-

Copyright2000 by theAmericanGeophysical Union.

mentsare discussed in detail by Ferlemannet al. [1998, 2000],and Harder et al. [1998,2000]. Eight LPMA/DOAS balloonflightshave alreadybeen successfullyconductedat mid- and high latitudes during all seasons(Table 1). Sincethe detectionof tropospheric BrO essentially relies on a subtraction of

Papernumber2000GL011531. 0094-8276/00/2000GL011531505.00

sequentially measured tropospheric and stratospheric BrO-SCDs during balloonascent- after accountingfor the photochemicalchangein the stratosphericBrO con-

ty, University Park, USA

• Max-Planck-Institutfiir Chemie,Mainz, Germany 6alsowith NOAA, AeronomyLaboratory, Boulder, Co,

USA

2921

Table 1. Compendiumof simultaneously measuredBr0-VCDs by Balloon (LPMA/DOAS), Satellite (GOME), and Ground-Based(GB) No

Date

Location

1.

Nov.

23, 96

Le6n

2.

Feb.

14, 97

Kiruna

3.

Jun.

20, 97

Gap

4.

Mar.

19, 98

LeSn

5.

Aug.

SZA tø ]

Altitude

BrO-VCD

BrO-VCD

(1013rno/ec./crn 2)

SZA [o]

Hfloat [ppt]

74 -- 86.4

12 -- 30.6

2.3 q- 0.5

14.4

8 -- 30

4.0 -F 0.6

81.7

15.6

4- 3

39.8

-

4.3 6.0 6.8 6.4 5.2

65.4

82.5

23.4

15.3

q- 3

12 q- 2

[km]

aboveTropop .... (1013•o•c./c• 2)

(42.6 ø N, 5.7 ø W) -- 88.8

(67.9 ø N, 21.1 ø E) 90 -- 56

-- 31

qqqqq-

0.8(GOME) I(GOME) 0.8(GOME, 2(GB) I(GOME)

Feb.13)

GOME

BrO

above

q- 3

(44.0 ø N, 6.1 ø E) 65 -- 87

5 -- 38.5

2.0 q- 0.5

4.4 q- 0.5(GOME)

45.5

74.8 -- 87

2.2 -- 38.5

3.2 q- 0.5

55.6

94.5

-- 84.7

33.3

82.9

-- 86.2

0.2 -- 28.2

4.8 q- 0.5 -

4.0 3.6 4.0 5.7 6.7 3.3

(42.6 ø N, 5.7 ø W) 19/20,

Kiruna

98 6.

Feb.

(67.9 ø N, 21.1 ø E) 10, 99

Kiruna

-- 27.6

(67.9 ø N,21.1 ø E) 7.

Jun.

25, 99

Gap

8.

Feb.

18, 2000

Kiruna

93.5

-- 50.0

39.1

81.4

-- 89.9

6 -- 30.1

-- 20.5

q- 0 q- 0 q- 0 q- 0 q- 0 q- 0

5(GOME) 8 (GB) 5 (GOME) 7(GOME) 8(GB) 5(GOME)

75

--

17 q- 3 87

56.0 82.1

13 q- 2 13 q- 2

24.3

-

82

17q-2

(44.0 ø N, 6.1 ø E) 3.8 q- 0.5

5.5 q- 0.7(GOME)

(67.9 ø N, 21.1 ø E)

centrationand observationgeometry(seebelow)- the

flight 6 (from Kiruna on Feb. 10, 1999) - with a large

method is only suitableif the resultingdifferenceis lar- Signal-To-Noise ratio (SNR) - is mostsuitablefor direct ger than the combineduncertainties of both measured detectionof the small troposphericBrO concentrations. BrO-SCDs. It is thus necessaryto discussthoroughly TroposphericBrO can also be inferred from the flight the uncertainties

involved in the method.

5 (Kiruna on Aug. 19, 1998), however,with a smaller

Unfortunatelyonlytwo LPMA/DOAS balloonflightsso far performed are suitable for a searchfor tropospheric

BrO. First, sunriseoccultationmeasurements(flights 3, 7, secondpart of flight 5) are conductedfrom the balloon flying in the mid- and upper stratosphere,and thus detection of troposphericBrO is not possible.Se-

cond, strongtroposphericshearwind (flights 1, 2, and 4), technicalproblemsand/or clouds(flight8) prevented a proper Sun pointing during the troposphericpart of the ascent.Third, the solarzenithangle(SZA) - and thus the air massfactor under which the comparably small troposphericand much larger stratosphericBrO columns are observed - are too small for some of the

SNR than for flight 6. Also the ascentmeasurementsof flight 5 were partly affectedby remnants of one of the auxiliary balloonsthat unfortunately engulfedparts of the balloongondola,and thus eventuallyobscuringthe telescope'sview of the solar disc. For flight 6, Figure 1 comparesthe measuredand modeled BrO-SCD as a functionof flight time. As the balloon ascendsthrough the troposphereinto the lower stratosphere,the BrO-SCDs remain nearly constant.Further on, the measuredBrO-SCDs decreasestrongly,particularly when the balloon passesthe maximum height of the BrO concentrationprofile and then ascendsfurther to the balloonfloat altitude (at 28- 29 km, and

flights(1 and 4) to separatethe smalltropospheric from SZA=86.2ø). During solar occultation,the BrO-SCD the larger stratosphericcolumn. Therefore, we feel that

5-Jll

I 4,)

•

•

'. I ,'

I. ' ,I ' . I

o

'

valuesincreaseagain, becausethe line of sight through the absorbingBrO layer becomeslonger. The modeled BrO-SCDs are obtained by integrating photochemically modeledstratosphericBrO concentrationsalong the line of sightof our observation(for detailsseeFigure 1

in Harder et al., [2000]).The modeledSZA-dependent

,

BrO concentration

ø o%•% %••o

field is taken from the SLIMCAT

3D-CTM modelfor the closestgrid point to our measurement[Chipperfield, 1999].Clearly,the measuredand modeled BrO-SCDs compareexcellentlyfor the obser-

"•

'

E

.,-.

Troposphere Stratosphere •'

model to correctly predict lowermoststratosphericBrO concentrationsfor that flight, the remainder of the gap may possibly point to tropospheric BrO. Since a forthcoming paper - involving a comparisonof total stratospheric bromine inferred from our BrO measurement,

5

O

SZA=. Trol: p

SZA=86.2

0•,' ' ' ' ' ' ' ' ' ' '4' i

11:3012:0012:3013:0013:3014:001

'30

Time [ UT] Figure

vation after 12:30 UT (above20 km), while an increasinggap opensbetweenmeasuredand modeledBrOSCD with decreasingatmosphericheight. While a part of the gap is due to a shortcomingof the SLIMCAT

1. Comparison of measured and mode-

led BrO-SCDs for the balloon flight from Kiruna (67.9øN,21.1øE) on Feb. 10, 1999 as a functionof flight time. The modeled curves are simulated line of

sightBrO-SCDsfor (a) stratospheric BrO predictedby SLIMCAT (curve1), (b) the stratospheric BrO profile as measuredby our instrument without photochemical correction(curve2), and (c) as (b) but including the photochemical correction(curve3) (for detailssee text).

anda GC/MS analysisof total organicbrominein stratosphericair samplestaken closeto our flight [Pfeilstickeret al., submittedto GRL, 2000]- will address the model'sunderpredictionof lower stratosphericBrO in detail, we focushere on the troposphericcontribution to this gap. A zoom of the low altitude intercomparison is also shownin Figure 1. All three simulations(curves1, 2, and 3) describethe amount of BrO-SCD that wouldhave been measured if there was only stratosphericBrO present. Curve 1 is a SLIMCAT simulation of stratosphericBrO for our flight. Curve 2 is calculatedfrom the later on measuredstratosphericBrO profile including no photochemicalcorrections.It has to be noted that assuming no photochemicalrelated changein stratospheric BrO during the courseof the measurementsslightly underpredictsthe BrO-SCD, mainly becauseduring

balloonascentsomeof the stratosphericBrO (•.4 %)

FITZENBERGER

ET AL.' FIRST PROFILE MEASUREMENT

OF TROPOSPHERIC

BRO

2923

hasalreadyreactedinto its nighttimereservoir species The measured total BrO-VCDs are taken from three (believed to be mainlyBrC1for this flight).In order sources:(1) the 4 combinedBrO measurements-taken to accountfor this photochemical lossin stratospheric from the ground prior to the launch- result in a BrOBrO, simulation 3 was carried out. Here we corrected VCD of (5.7+0.6)xl0 •3 molecules/cm 2 forSZA=82.8ø,

our measuredstratosphericBrO profile with the model-

(2) GOME measured almost the same total atmopredictedlossof stratosphericBrO during our measu- sphericBrO-VCD of (5.7+0.7)x10TMmolecules/cm 2 rements. The corrections are -•4 % for the total atmosfor a similar SZA (- 82.2ø) as for our balloon phericcolumn(seealsoFigure2 for the model-predicted measurement,and (3) our ground-basedinstrument changein stratospheric BrO for the neargroundandthe installed at Kiruna/Institutet f6r Rymdfysik (IRF), lowermoststratosphericobservations at SZA = 82.8ø, roughly 50 km from the balloon launch site, meaand 83.3ø, respectively). sureda BrO-VCD of (6.7+0.8)x10TMmolecules/cm 2 To consolidatethis result, we conducteda statistical at a SZA = 82.2ø. Combining all three measuretest of the difference in measured and modeled BrOments, then the combined atmospheric BrO-VCDs

SCDsfor the tropospheric observations (difference of ((5.9+0.4)x10•3 molecules/cm 2) aswellaseachof the the measuredvaluesto curve3). The test clearlyshows individual total BrO-VCDs are significantlylarger than

that, the combinedmeasuredtroposphericBrO-SCDs the integrated stratosphericBrO profile. Finally, our are on a 97.5 % significance level larger than the pre- integrated tropospheric profile results in a BrO codicted values.Therefore, we concludethat some BrO lumnamountof (0.5-}-0.2)x1013molecules/cm 2, a vawas present in the troposphere. lue being in reasonableagreementwith the above difAnother test for the presenceof troposphericBrO re- ferencein the combinedatmosphericBrO-VCD ((0.2 lies on a comparisonof measuredtotal atmospheric 2.1)x10•3 molecules/cm2)• taking into accountall BrO verticalcolumndensity(BrO-VCD), andthe inte- to errors. It has to be noted that our tropospheric BrOgrated stratosphericprofile. Our integratedstratospheVCD closelycorrespondsto the troposphericBrO-VCD

ric BrO profileyields(4.8+0.5)x10TMmolecules/cm 2. that is required to bring the measured total atmosphe-

ric and modeledstratosphericBrO-VCD into agreement for a comparisonat the same location, and season,ho-

25

wever,for 1995 [Friefi et al. 1999].Finally, inspecting

0'I' ' ,O I'I'.

E 15

ß

.

10

•

Tropopause Kiruna 98. TropepauseKiruna99

the inferred profile obtained for our balloon flight, the tropospheric BrO maximizes at around 2-3 km altitu-

de (0.6+0.2 ppt), and it is closeto zero at the local tropepause(Figure 2). In our search for tropospheric BrO, we also inspected closerflight 5. As before,the BrO-SCD valuesin the troposphereare larger than if there was only BrO in the

stratosphere (not shown).For this flight- despitethe

smaller SNR for the measured BrO-SCDs because of

smallerSZAs and henceline of sightair masses- weak shearwindsallowedus to recordoccasionally spectra from 2.2 km altitudeupwards.The inferredBrO profile

(Figure2) looksverysimilarto the winterprofile(flight 6), however,with larger BrO mixing ratios at profile maximum(2+0.Sppt)than detectedfor flight6. As before, comparingthe total vertical column amounts- for

the whole atmosphere,stratosphere,and troposphere - yieldsthe followingresults(4.4+0.5), (3.2+0.5), and

(1.2+0.4)10 TMmolecules/cm 2, respectively. A compari;

_--

o

0

--o,

,

sonof the valuesevidencesthat troposphericBrO was alsopresentduring flight 5.

,

ß

Kiruna Summer

o

Kiruna Winter ,

,

1998

Table1 summarizes ourfindings forthe8 LPMA/DOAS flightsalreadyconducted.The inter comparison reveals that tropospheric BrO - amounting to (0.6to 3.7)x l0 •3

molecules/cm 2, or 0.4 pptvto 2.3 pptv- if uniformly

1999 ,

Concentration [10 molecule/era a] Figure 2. BrO profilesfor the balloon flight from Kiruna on August 19/20, 1998 (dark squares),and Feb. 10, 1999 (open circle). The measuredstratospheric BrO profile (open circle) is an averageover 82.8ø