A seasonal air transport climatology for Kenya - Wiley Online Library

JOURNAL

OF GEOPHYSICAL

RESEARCH,

VOL. 104, NO. D12, PAGES 14,237-14,244, JUNE 27, 1999

A seasonal air transport climatology for Kenya C. K. Gatebe,•,2P. D. Tyson,3 H. Annegarn,4 S. Piketh,3 and G. Helass Abstract. A climatologyof air transportto and from Kenya hasbeen developedusing kinematictrajectorymodeling.Significantmonthsfor trajectoryanalysishavebeen determinedfrom a classification of synopticcirculationfields.Five-pointback and forward trajectoryclustersto and from Kenyarevealthat the transportcorridorsto Kenya are clearlyboundedand well defined.Air reachingthe countryoriginatesmainlyfrom the Saharanregion and northwesternIndian Ocean of the Arabian Sea in the Northern Hemisphereand from the Madagascanregionof the Indian Oceanin the Southern Hemisphere.Transportfrom eachof thesesourceregionsshowdistinctiveannualcycles related to the northeasterlyAsian monsoonand the southeasterly trade wind maximum over Kenya in May. The Saharantransportin the lower troposphereis at a maximum when the subtropicalhigh over northernAfrica is stronglydevelopedin the borealwinter. Air reachingKenya between700 and 500 hPa is mainlyfrom Saharaand northwestIndian Oceanin the monthsof Januaryand March, which givesway to southwestIndian Ocean flow in May and November.In contrast,air reachingKenya at 400 hPa is mainlyfrom southwestIndian Ocean in Januaryand March, which is replacedby Saharantransportin May and November.Transportof air from Kenyais invariant,both spatiallyand temporally,in the tropicaleasterliesto the CongoBasinand Atlantic Oceanin comparisonto the transportto the country.Recirculationof air has alsobeen observed but on a limited and often local scaleand not to the extent reported in southernAfrica. 1.

Introduction

Many studiesof transportprocesses in the atmosphere,particularlylong-rangetransportand horizontaland verticalmixingof aerosolsandtracegases,havebeenundertakenoverthe last three decades[Pitty,1968;Schiirtz,1980;Prosperoet al., 1981;Perssonet al., 1987; Muhs et al., 1990; Crutzenand Andreae, 1990; Watson et al., 1990; Levine, 1991; Fishman et al.,

1991;Moody et al., 1991;Dorlinget al., 1992;Pickeringet al., 1992,1994;Krishnamurtiet al., 1993;Swapet al., 1993;Franzen et al., 1994]. Most were undertaken in order to understand higher than anticipatedconcentrationof aerosolsand trace gasesin remote environments.Atmospherictransportsover southernAfrica and adjacentoceanshavebeen studiedextensivelyas part of the Transport and AtmosphericChemistry Near the Equatorial Atlantic/SouthernAfrican Fire-AtmosphereResearchInitiative (TRACE A/SAFARI) investigations conductedin the australspringof 1992 [Garstanget al., 1996;Swapet al., 1996]andin subsequent investigations [Tyson etal., 1996a,b, c;Sturmanetal., 1997;TysonandD9tbreton,1998]. The transport of aerosolsand trace gasesover Kenya is currentlyunder investigation. Previousstudiesof atmospheric

transportover East Africa as a whole are of generalmeteorological nature and relate to local and larger-scalemonsoon circulations[Findlater,1968; Fraederick,1972; Kiangi et al., 1981; Okeyo, 1987; Anyamba, 1983, 1990; Mukabana, 1992; Ininda, 1995]. Of particular interest is the work of Findlater [1968]on monthlymeanwind variationsat 11 Kenyanstations at altitudes of 0.9, 1.5, 2.1, and 3 km above mean sea level.

Findlater demonstratedthe gradualwind changefrom northeasterliesin Januaryto easterliesin March, southeasterlies in July, and back to easterliesagain in November.The change wasobservedto occurfirst at low levelsbefore risingto midlevels. The northeasterlymonsoonwas noted to divide into two streamson entering Kenya. One stream curvesto become easterlyand flowsbetweenthe highlandsof Kenya and Ethiopia.The otherbecomesnortherlyandflowsoverthe relatively flat areas to the east of 38øE. Both

of these branches

were

found to coexiston anyone day,but often, one is more strongly developedthan the other. In northwesternKenya, persistent easterlywinds prevail throughoutthe year at all levelsin the lower troposphere. Prior to determiningactualaerosoltransport,it will be useful to establishan air transportclimatologyfor Kenya, as has been done for southernAfrica [Tysonet al., 1996a]. In this paper, sucha climatologyis developedat the synopticscale qnstituteof NuclearScience, Universityof Nairobi,Nairobi. 2Onleaveat Climatology Research GroupandSchonland Research usingLagrangiankinematictrajectorymodeling. Center for Nuclear Sciences,Universityof Witwatersrand,JohannesThe climateof Kenyais dominatedby the annualcycleof the burg, SouthAfrica. latitudinalmigrationof the Hadleycellsand IntertropicalCon3Climatology Research Group,University of Witwatersrand, Johan-

vergenceZone (ITCZ) over the region,followingthe annual movementof the Sun [Newellet al., 1972;Okeyo,1987;Asnani, Witwatersrand,Johannesburg,SouthAfrica. 1993].The ITCZ being a zone of confluenceof northeasterly SBiogeochemistry Department, MaxPlanckInstitutefor Chemistry, and southeasterlytrades is mainly associatedwith rain. The Mainz, Germany. Sun is approximatelyoverheadin East Africa at the end of Copyright1999by the AmericanGeophysical Union. March and againat the end of September,so the heat trough can be expectedto be mosteffectiveabout a month later, i.e., Paper number 1998JD200103. 0148-0227/99/1998JD200103509.00 late April/May and late October/November.The so-called nesburg,South Africa.

4Schonland ResearchCenterfor NuclearSciences, Universityof

14,237

14,238

GATEBE ET AL.: SEASONAL TRANSPORT CLIMATOLOGY

(a) Northeasterly flow

(b) Equatorialeasterlyflow

(c) Southeasterlyflow

(d) Westerly flow

Figure 1. Major synopticcirculationtypesover Kenya and easternequatorialAfrica: (a) northeasterlyflow associated with the Arabian Ridge of high pressure,(b) equatorialeasterly flow associated with the equatorialIntertropicalConvergenceZone (ITCZ), (c) southeasterly flow associated with a ridgeof highpressureover eastAfrica, (d) westerlyflow associatedwith a ridge of high pressureover the Congo and low pressureor tropical cyclonesin the southwesternIndian Ocean.

cordingto the major synopticfeature(s)associated with each. Four major representativetypeswere identified:northeasterly flow, equatorialeasterlyflow, southeasterly flow, andwesterly flow (Figure 1). Trajectoryflow patternswere then determinedfor specific monthsdominatedby the different circulationtypes.Ten-day backwardand forward trajectoriesstarting at Mount Kenya were calculatedover the 3-yearperiod 1991-1993for January, March, May, and November usingEuropean Center for Medium-RangeWeather Forecasts(ECMWF) operationalanalysesof the 6 hourly,three-dimensional (3-D) wind field data at variouspressurelevels.The ECMWF operationalanalysesof the 3-D wind field are availablegloballyevery 6 hours at 31 levels,of which 13 below 500 hPa are used.Bengtsson [1985a] givesa descriptionof the ECMWF data assimilationsystem. Pickeringet al. [1994, 1996], in a comparisonof trajectory methods and global data sets, find the ECMWF-generated fieldsto yield the bestcurrenttrajectoryresults.Eachtransport pathwayis calculatedover a 10-dayperiodusingthe DMbreton [1996]model.The modelis basedon the principleof Lagrangian advectionwith u, v, and w wind componentsat a given time, and atmosphericlayer and grid usedto computea new downstream(upstream)location of an air parcel at a later (earlier) time. This procedureis then repeatedto producea forwardor backwardtrajectoryover 10 days.Verticalvelocities are determinedfrom nonlinearmode initialization,whichpermits the diabatic as well as the adiabatic processes,which influence vertical motion, to be taken into account.

Daily clustersof five trajectoriesat latitudesand longitudes 1ø apart, with the centroid at Mount Kenya (0.1øS,37.2øE), were calculated for the initial levels of 700, 600, 500, and 400

hPa for the 4 monthsin eachof the years.SinceECMWF data are availableon 2.5ø x 2.5ølatitude-longitudegrid [Bengtsson, 1985b],Mount Kenyais taken as a representativeof the entire country.The merit of Mount Kenya lies in its locationnext to the equator,secondhighestmountainin Africa (5199 m) and an isolated

long-rainseasonlastsfrom roughlyMarch to May and the shortrainsfrom late Octoberto early December[Kiangiet al., 1981]. However, the local variations,becauseof topography and the presenceof the large lakes,introducesuchsignificant modifications

that the seasons cannot be said to follow

the

classical(ITCZ) patternwith enoughcorrelationto make this approacha practicalone [Griffiths,1969].The diurnalvariation of precipitationis largelydeterminedby the mesoscale flows, the synopticscaleflows,and the interactionbetweenthe mesoscaleand the synopticscaleflows[AsnaniandKinuthia,1979; Asnani, 1993]. On a longer timescalethe annualcycleof circulationchangesis modulatedby east-westadjustmentsin the

FOR KENYA

feature

in what

can be considered

as a remote

environmentin central Kenya. It is also in an area that has been known climatologicallyto experiencemonsoonwinds at different times of the year. Matched back and forward trajectorieswere groupedaccordingto similarities.Six major types of transportto Kenyawere identifiedaccordingto sourcearea: from the northwest Indian Ocean, from the southwestIndian

Ocean, anticyclonicSaharantransport,in the Atlantic westerlies, circum Kenyan transport,and localizedtransportwithin Kenya. Likewise,sixmajor transportmodesfrom Kenyawere delimited:equatorialeasterlytransportto the Congoand Atlantic Ocean, recurvedeasterlytransportto the Arabian Penisula,northwesterly transportto the southwestIndian Ocean, zonal Walker circulation associated with E1 Nifio-Southern northeasterlytransportto the northern Indian Ocean,southOscillation(ENSO) [CadetandDiel, 1984;Cadet,1985;Ininda, westerlytransportto the central interior Africa, and return 1987; Nyenzi, 1988; Ogallo, 1988]. Variability of the general circulationto Kenya. circulationpatternson seasonaland longertimescalesimilarly influencestransport[Charney,1975;Charneyand Shukla,1981; 3. Results and Discussion Barnetet al., 1991;Gordonand Hunt, 1991;Asnani,1993]. 3.1.

Synoptic Flow Patterns

Four synopticcirculationtypesdominateairflowoverKenya (Figure1). The firstis the northeasterly flowtype,in whichthe Daily synopticcharts for 1200 UTC at 850 hPa in Kenya 850 hPapressurefield is dominatedby a ridgeof high-pressure over were examinedfor the 5-year period 1971-1975. The period field over the Arabian Penisulaproducingnortheasterlies was chosenbecauseof the availabilityof a wide coverageof Kenya.The easterlyflow type occurswith the equatorialconcontinuousupper air data in Kenyaand the region.Circulation vergenceof northeastand southeasttradewindsoverthe easttypeswere then groupedinto similarpatternsand named ac- ern equatorialregion of Africa. The southeasterly flow type

2.

Data and Methodology

GATEBE ET AL.: SEASONAL TRANSPORT

CLIMATOLOGY

FOR KENYA

14,239

Table 1. Monthly PercentageFrequencyof SynopticCirculationin Kenya

NEF EEF SEF WF

Jan.

Feb.

March

April

May

June

July

Aug.

Sept.

Oct.

Nov.

Dec.

90.3 6.4 2.6 0.6

85.8 10.6 2.1 1.4

54.8 25.1 20.0 0.0

4.7 24.7 70.7 0.0

0.0 0.6 99.4 0.0

0.0 0.7 99.3 0.0

0.6 0.6 98.7 0.0

0.0 1.9 98.1 0.0

0.0 2.0 98.0 0.0

0.0 11.6 88.4 0.0

34.7 35.3 28.7 1.3

72.9 23.9 3.2 0.0

NEF, northeasterlyflow; EF, equatorialeasterlyflow; SEF, southeasterly flow; WF, westerlyflow.

resultsfrom a ridge extendingfrom the southtoward Kenya. The fourth type is the westerlyflow type in whichthe pressure fieldsexhibita ridge from the Congoeastwardand low pressure or tropical cyclonesin the southwesternIndian Ocean. Winds with an easterly component dominate the Kenyan tropics [Anyamba,1990]. However, analysisof the synoptic flow patternsduringthe 5-yearperiod studiedrevealsthat the northeasterlymonsoonsare mostprevalentfrom Decemberto April, with the highestfrequencyof occurrence(90%) during January(Table 1). The southeasterly monsoondominatesfrom April to October,with nearly all the daysfrom May to September,showingat least88% southerly/southeasterly flow.Direct easterlyflow is dominantduringNovember(35%). March is a transitionalmonth between northeasterlyand southeasterly monsoons;November is transitionalbetween southeast and northeastmonsoons.These two monthshave highestrel-

(a) NorthwestIndianOcean

(b) SouthwestIndianOcean

ative frequenciesof easterliesinto Kenya. These resultsconfirm the earlier studiesof Findlater [1968] and Kiangi et al. [1981]. Westerlyflow is observedin January,February,and November but with frequenciesnot exceeding2%. Unstablewesterly currentsare observedoccasionallyat lower and middle levels of the atmosphereover East Africa [Forsdyke,1949, 1960; Thompson,1965;Nakamura, 1968;Harangonzoand Harrison, 1983;Kiangiand Temu, 1984;Daviset al., 1985;Murakami and Sumathipala,1989]. The seasonalcycle of the different flow typeshasbeen usedto definethe optimummonthsto undertake trajectoryanalysisfor determiningair transportcharacteristicsto and from Kenya over 20-dayperiods.

4.

Transport Patterns

Back trajectoryanalysisto show the maximumfrequency, 10-daytransportcorridorsto Mount Kenya revealstransport modesmore complicatedthan the simplecirculationtypology suggests (Figure2). Northeasterlyand southeasterly transport from the northwestern

and southwestern

Indian

Ocean corre-

spondsto their seasonalmonsoonairflow co.unterparts(Figures2a and 2b). Likewise,anticyclonicSaharantransportis a prominentfeature (Figure 2c), and a westerlytransportmode is evident(Figure 2d). Near-closedanticyclonictransportoccursaroundKenya at different spatialscales(Figures2e and 2f). The frequenciesof transportmodesare givenin Table 2 for different monthsand levelsin the atmosphere. NorthwestIndian Ocean and anticyclonicSaharantransport appearsto be dominantat all levelsin the monthsof January and March whennortheasterlies and northerliesare mostprevalent. It is at thesetimesthat Kenya is under stronginfluence of the Arabian and Azores semipermanentanticyclonesof the Northern Hemisphere.Southwesterntransportis most dominant in May in line with the period when southeasterlies are dominant in Kenya. Atlantic westerly transport is an upper tropospherephenomenonin January,March, and November. !el Circum Kenya (f) Within Kenyan The two anticycloniccirculationsare evidentlyabsentin the mediumlevelsin the month of Novemberand only appearat 400 hPa. During this month, Kenya is under the influenceof both the northeasterliesand the southeasterliesconverging over the region. Air transport away from Kenya, as determinedfrom clustered forwardtrajectoryanalysis,revealssixmajor modes(Figure 3). The equatorialeasterlyto the Atlantic Oceanand the recurvedeasterlyto the Arabian Penisulaappeardominantin Figure 2. Predominant atmospherictransport modes con- January,March, andNovember.This impliesthat the air transveyingair to Kenyaas determinedby backtrajectorymodeling port to Kenya either as northerly, northeasterly,or easterly endingat Mount Kenya (triangle): (a) from the northwest Indian Ocean,(b) from the southwest Indian Ocean,(c) anti- over the sameperiod exitsmainly to the Atlantic and Saudi cyclonicSaharantransport,(d) in the Atlantic westerlies,(e) Arabia. During the month of May air transport, which is circum Kenyan transport, and (f) localizedwithin Kenyan mainly from the southernwesternIndian Ocean, exits mostly transport. to the Atlantic Ocean.Return circulationto Kenya,thoughnot

14,240

GATEBE ET AL.: SEASONAL TRANSPORT CLIMATOLOGY

Table 2. PercentageFrequencyof Air Transport to Kenya 700 hPa

600 hPa

500 hPa

400 hPa

FOR KENYA

at 700 hPa). Transportto the southwestIndian Ocean in the regionof Madagascarincreasessignificantlyat this time of the year.

The annual variation of transportin the three streams,accountingfor the largest proportion of the total transport,is striking(Figure 8). The transportto Kenya showsmore variability than that awayfrom the country.In the latter case,the transportto the CongoBasinregionand to the Atlantic Ocean is relatively invariant throughout the year, being greatest

January Northwest Indian Ocean Southwest Indian Ocean

44 0

55 4

61 10

20 50

AnticyclonicSaharan Atlantic westerly Circum Kenyan Within Kenyan

47 0 0 9

34 0 0 7

22 0 7 0

10 12 8 0

Northwest Indian Ocean Southwest Indian Ocean

March 49 23

49 14

25 22

37 25

AnticyclonicSaharan Atlantic westerly Circum Kenyan Within Kenyan

28 0 0 0

36 0 0 1

47 0 5 1

29 6 3 0

Northwest Indian Ocean Southwest Indian Ocean

4 96

25 70

42 35

48 32

0 0 0 0

1 0 0 4

18 0 0 5

15 0 0 5

(--•70%) in Januaryand least (--•40%) in November.In contrast,transportto Kenyafrom the Madagascarregionis greatest (>90%) in May and least(zero) in January. Integrating the lower-tropospheretransport, from 700 to 500 hPa, allowsthe transportfieldsto be further simplifiedto advantage(Figure9). The distinctiveseasonalcharacteristic of transportto Kenya from the Saharanand Madagascanregions is clear;the relative constancyof transportfrom the northwest Indian Ocean region is likewiseapparent.In the integrated transportfrom Kenya it is only the tropical easterlyflow from the country to the Congo Basin and Atlantic Ocean that is important.

May

AnticyclonicSaharan Atlantic westerly Circum Kenyan Within Kenyan

Measurements

November

Northwest Indian Ocean Southwest Indian Ocean

AnticyclonicSaharan Atlantic westerly Circum Kenyan Within Kenyan

44 50

66 33

63 28

6 0 0 0

1 0 0 0

7 2 0 0

of the annual

variation

of the constituent

elementsof the aerosolloading at the remote Mount Kenya site are currentlybeing made. Given knowledgeof air trans-

49 26

16 5 1 3

(a) Equatorialeasterly

(b) Recurvedeasterly

frequent,is nonetheless interesting(Figure3f). Frequenciesof occurrence of the differenttransportmodesare givenin Table 3.

5.

Dominant Transport Fields

The flow of air into and out of Kenya in four different months and levels is shownin Figures 4-7. Analysisof the combinedflow duringthe monthof January(Figure4) reveals that at the 700 hPa level, most air reachingKenya flowsanticyclonicallyfrom the Saharanregion (56%). With increasing height,the proportiondropsto 29% at 500 hPa and to 18% at 400 hPa. In contrast,the proportioncomingin from the northeast Indian Ocean rises from 44% at 700 hPa to a maximum

of

61% at 500 hPa.At all levels,air beingtransportedfrom Kenya is at maximumtoward the Atlantic Ocean. The greatesttransport is at 700 hPa (69%). The secondmostimportanttransport stream awayfrom Kenya is that directedtoward the Arabian Penisula(16%). In March (Figure5) a similarpictureis maintained,exceptthat in lowerlevelsthe transportfrom southwest Indian Ocean area increases,as that from the Saharanregion beginsto decrease.Transportfrom Kenya remainspredominantly toward the Congo Basin and the Atlantic Ocean. In contrast,duringMay the transportpatterns,particularlyin the lowertroposphere,are markedlydifferent.Transportfrom the Saharanregion is absentor substantiallydiminished,whereas that in the southeasttrades from the Southern Hemisphere increasessignificantly to reach96% at 700 hPa (Figure 6). In November(Figure7), transportof air to Kenyais mainlyfrom the Madagascanregionand from Kenya to the Atlantic Ocean. At the same time, outward transport toward the Arabian Desert increasessignificantly. Anticyclonicflow remainsmaximizedin the tropicaleasterliesto the Congoand beyondat all levelsandto the greatestextentin the lowertroposphere(42%

Figure 3. Predominant atmospherictransport modes conveying air from Kenya as determinedby forward trajectory modelingstartingfrom Mount Kenya(triangle):(a) equatorial easterlytransportto the Congoand Atlantic Ocean, (b) recurvedeasterlytransportto the Arabian Penisula,(c) northwesterlytransportto the southwestIndian Ocean, (d) northeasterly transport to the northern Indian Ocean, (e) southwesterly transportto the centralinterior Africa, and (f) return circulationto Kenya.

GATEBE ET AL.: SEASONAL TRANSPORT

Table

3.

MARCH (a) 700 hPa

PercentageFrequencyof Air Transportfrom

Kenya 700 hPa

600 hPa

CLIMATOLOGY

500 hPa

400 hPa

61 24 5 0 7

66 24 1 0 6

59 30 0 0 9

3

3

2

67 26 0 0 6

60 22 6 10 0

55 15 8 10 4

53 23 3 6 1

1

2

8

14

48 0 37 8 7

43 2 36 8 11

26 0 35 8 31

37 10 26 6 21

0

0

0

0

42 31 4 6 10

43 35 9 5 8

27 30 14 2 26

28 11 27 25 3

7

0

1

6

FOR KENYA

14,241

(b) 600 hPa

January

Equatorialeasterly Recurvedeasterly Northwesterly Northeasterly Southwesterly Return circulation

69 16 4 0 7 4

(c) 500 hPa

March

Equatorialeasterly Recurvedeasterly Northwesterly Northeasterly Southwesterly Return circulation

o,

d) 400 hPa 23%

May

Equatorialeasterly Recurvedeasterly Northwesterly Northeasterly Southwesterly Return circulation

Figure 5. March air transportto andfrom Kenyaat 700,600, 500, and 400 hPa.

November

Equatorialeasterly Recurvedeasterly Northwesterly Northeasterly Southwesterly Return circulation

tropicallarge-scale regionaltransportof air, and the matter containedtherein,over easternequatorialAfrica. 6.

Conclusion

A climatology of air transportto and from Kenyahasbeen developedusingkinematictrajectorymodeling.Significant portclimatology andthesourceandsinkareasof air reaching monthsfor trajectoryanalysishavebeen determinedfrom a of synoptic circulationfields.Five-pointbackand and movingawayfrom the site, it shouldprovepossibleto classification quantifyaerosoltransportin the Kenyanregion.More impor- forwardtrajectoryclustersto and from Kenyarevealthat the tantly,it will allowestimates of cross-equatorial hemispheric transportcorridors,or plumes,to Kenyaare clearlybounded transfers of aerosols to be made for the first time over contiand well defined.Air reachingthe countryoriginatesmainly nentalAfrica.The simpletechniqueusedin thispaperpresents from the Saharanregionand northwesternIndian Ocean of a furthermeansof understanding the regionalmanifestationof the Arabian Sea in the Northern Hemisphereand from the the generalcirculationof the atmosphere andresultantinter- Madagascanregion of the Indian Ocean in the Southern JANUARY (a) 700hPa

May (b)600hPa

(a) 700 hPa

(b) 600hPa

41%

Figure 4. Januaryair transportto and from Kenya at 700,

Figure 6. May air transportto and from Kenyaat 700, 600,

600, 500, and 400 hPa.

500, and 400 hPa.

14,242

GATEBE ET AL.: SEASONAL TRANSPORT

NOVEMBER (a) 700 hPa

CLIMATOLOGY

a) JANUARY

FOR KENYA

,,

(b) MARCH

(b) 600 hPa

ß

42%

43%

65%•• ,, I

(d) NOVEMBER

(c) MAY

(c) 500 hPa

(d) 400 hPa

i Figure 9. Integrated700-500 hPa air transportto and from Figure 7. Novemberair transportto and from Kenya at 700,

Kenya.

600, 500, and 400 hPa.

Hemisphere.Transport from each of these sourceregions showsdistinctiveannual cyclesrelated to the northeasterly Asian monsoonand the southeasterlytrade wind maximum over Kenya in May. The Saharantransportin the lower troposphereis at a maximumwhen the subtropicalhigh over

100a)ToKenya

90

/'••

80

/ ]

70

50

• '

Anticyclone SaharanRecirculation

•

Jar

•

h

Acknowledgments.The first author completedthe analyseson whichthispaperis basedwhileon START andIPPSfellowships at the Universityof Witwatersrand,Johannesburg, workingin the Climate

100

•J Mar

May

Nov

. b) FromKenya

90-

-.. •,

80-

of air has also been observed but on a limited

--.•. Northwest Indian Oceanandoftenlocalscaleandnotto the extentreportedin southern Southwest Indian OceanAfrica [Tysonet al., 1996b].

lO

o-/

northernAfrica is stronglydevelopedin the borealwinter.Air reachingKenyabetween700 and 500 hPa is mainlyfrom Sahara and northwestIndian Ocean in the monthsof January and March, whichgivesway to southwestIndian Oceanflow in May andNovember.In contrast,air reachingKenyaat 400 hPa is mainlyfrom southwestIndian Oceanin JanuaryandMarch, whichis replacedby Saharantransportin May andNovember. Transportof air from Kenya is invariant,both spatiallyand temporallyin the tropicaleasterliesto the CongoBasinand AtlanticOceanin comparison to the transportto the country.

70 60

% 50

Research Group and SchonlandResearch Center for Nuclear Sciences.The authorsare gratefulto the followinginstitutionsfor their contributionsto the study:the IGBP/WCRP/IHDP START; International Programmein PhysicalSciences(IPPS), Sweden;Tertiary Education Support Program of Eskom and Foundation for Research THRIP Programme,SouthAfrica, GermanyAcademicExchangeProgramme(DAAD), DeansCommitteeof the Universityof Nairobi,and KenyaMeteorologicalDepartment.Many thanksto MichaelGarstang Equatorial easterly transport for hiscomments on an initial draft of the paper.The authorswishto toCongoandAtlantic thankWendyJob of the ClimatologyResearchGroup,Universityof Recurved eastedy Witwatersrand,and Kiremia Josephof the GeographyDepartment, Northwesterly University of Nairobi, who prepared the figures, and Kollikho Wamukotaof the KenyaMeteorologicalDepartment,who helpedin collectingdata on synopticcharts.We thank the anonymousreferees for their suggestions, whichhavehelpedto improvethe paper.

40 30

References

20

Anyamba,E. K., On the monthlymeanlowertropospheric circulation and anomalouscirculationduringthe 1961/62floodsin EastAfrica, M.Sc. thesis,Dep. of Meteorol., Univ. of Nairobi, Kenya,1983. Anyamba,E. K., A diagnosticstudyof low frequencyoscillationsin tropical outgoinglongwaveradiation, Ph.D. dissertation,Univ. of

10

____.

0 Jan

Mar

I

May

/

Nov

Figure 8. Annual variation of transport at 700 hPa in the three most significantstreamsconveyingair to and from Kenya.

Calif., Davis, 1990.

Anyamba,E. K., andP.M. R. Kiangi,Mean motionfield in eastAfrica at the level of the east African low level jet core,Arch. Meteorol. Geophys.Biøklimatøl"Ser.B, 36, 29-41, 1985.

GATEBE ET AL.: SEASONAL TRANSPORT

CLIMATOLOGY

FOR KENYA

14,243

Asnani,G. C., TropicalMeteorology, vol. 1, edited by G. C. Asnani, 603 Kiangi,P.M. R., M. M. Kavishe,and J. K. Patnaik,Someaspectsof the mean troposphericmotion in East Africa during the long rain seapp., Pune Press,Pune, India, 1993. son,KenyaJ. Sci. Tech.Ser.A, 2, 91-103, 1981. Asnani,G. C., and J. H. Kinuthia, Diurnal variation of precipitationin east Africa, Tech.Res.Rep. 8/79, pp. 1-58, Kenya Meteorol. Dep., Krishnamurti,T. N., H. E. Fuelberg,M. C. Sinha,D. Oosterhof.,E. L. Nairobi, 1979. Bensman,and V. B. Kumai, The meteorologicalenvironmentof the Barnett,T. P., M. Latif, E. Kirk, and E. Roeckner,On ENSO Physics, troposphericozone maximum over the tropical South Atlantic J. Clim., 4, 487-515, 1991. Ocean,J. Geophys.Res.,98, 10,621-10,641,1993. Levine, J. S., Global biomass,in Atmospheric,Climateand Biospheric Bengtsson,L., Medium-range forecastingraTheexperienceof ECMWF, Bull. Am. Meteorol. Soc., 66, 1133-1146, 1985a. Implications,edited by J. S. Levine, 551 pp., MIT Press,Cambridge, Mass., 1991. Bengtsson, L., Medium rangeforecastingat ECMWF, Adv. Geophys., 238, 3-56, 1985b. Moody, J. L., A. A. P. Pszenny,A. Grandy, W. C. Keene, J. N. Cadet, D. L., The southernoscillationover the Indian Ocean, J. Clim., Galloway,and G. Polian,Precipitationcompositionand its variabil5, 189-212, 1985. ity in the southernIndian Ocean: Amsterdam Island, 1980-1987, J. Cadet, D. L., and B.C. Diel, Interannual variability of surfacefields Geophys.Res., 96, 20,769-20,786, 1991. over Indian Ocean during recent decades,Mon. WeatherRev., 112, Muhs, D. R., C. A. Bush, and K. C. Stewart, Geochemical evidence of 1921-1935, 1984. saharandustparentdustmaterialfor soilsdevelopedin quarternary limestonesof Carribean and western Atlantic Islands, Q. Res., 33, Charney,J. G., Dynamicsof desertsand droughtin the Sahel,Q. J. R. Meteorol. Soc., 101, 193-203, 1975.

157-177, 1990.

Charney,J. G., and J. Shukla, Predictabilityof MonsoonDynamics, Mukabana,J. R., Numerical simulationof the influenceof the largescalemonsoonflow on the weather patterns over Kenya using a edited by Lighthill and Pearce,CambridgeUniv. Press,New York, 1981. three dimensionallimited area model,251 pp., Ph.D. thesis,Dep. of Meteorol., Univ. of Nairobi, Kenya, 1992. Crutzen,P. J., and M. O. Andreae, Biomassburningin the tropics: Impact on atmosphericchemistryand biogeochemical cycles,Sci- Murakami, T., and W. L. Sumathipala,Westerly bursts during the ence, 250, 1669-1678, 1990.

D'Abreton, P. C., Langragiankinematicand isentropictrajectorymodels for aerosoland trace gastransportstudiesin southernAfrica, S.

Afr. J. Sci.Lett.,92, 157-160•1996. Davis, T. D., C. E. Vincent, and A. K. C. Beresford, July-August rainfall in west-centralKenya,J. Clim., 5, 17-33, 1985. Dhonneur, G., Nouvell Approache des Realite Meteorologiquesde lMfrique Occidentaleet Centrale,ASECNA, vol. 1 (358 pp.), vol. 2 (470 pp.), Univ. de Dakar, Senegal,1974. Dorling, S. R., T. D. Davies, and C. E. Pierce, Cluster analysis:A techniquefor estimatingthe synopticmeteorological controlson air and precipitation chemistry--Results from Eskdalemuir, South Scotland,Atmos. Environ., 26, 2583-2602, 1992. Findlater, J., The month to month variation of mean winds at low level over easternAfrica, Tech.Memo. 12, 27 pp., KenyaMeteorol. Dep., Nairobi, 1968.

Fishman,J., K. Fakhruzzaman,B. Cros,and D. Mganga,Identification of widespreadpollutionin the southernhemispherededucedfrom satelliteanalyses,Science,252, 1693-1696, 1991. Forsdyke,A. G., Weather forecastingin the tropical regions,Great Britain MeteorologicalOffice,Geophys. Mem., X(82), 32, 1949. Forsdyke,A. G., Synopticmodelsof the tropics,in Proceedings at the Symposiumon TropicalMeteorologyin Africa, pp. 14-23, Manitalp Found., Nairobi, 1960.

1982/83 ENSO, J. Clim., 2, 71-85, 1989.

Nakamura,K., Equatorialwesterliesover East Africa and their climatological significance,Geogr. Rep. 3, pp. 43-61, Tokyo Metrop. Univ., 1968.

Newell, R. E., J. W. Kidson,D. G. Vincent, and G. J. Boer (eds.),The GeneralCirculationof the TropicalAtmosphereand InteractionsWith Extratropical Latitudes,vol. 1,258 pp., MIT Press,Cambridge,Mass., 1972.

Nyenzi, B. S., Mechanismof East African rainfall variability,Ph.D. dissertation,Florida State Univ., Tallahassee, 1988.

Ogallo,L. J., Relationshipbetweenseasonalrainfall in EastAfrica and the southern oscillation,J. Clim., 8, 34-43, 1988.

Okeyo,A. E., The influenceof Lake Victoria on the convectiveactivitiesover Kenyahighlands,J. Meteorol.Soc.Jpn.,65, 689-695, 1987. Persson,C., H. Rodhe, and L. E. DeGeer, The Chornobyl'accident;a meteorologicalanalysisof how radionuclidesreachedand were depositedin Sweden,Ambio, 16, 20-31, 1987. Pickering,K. E., A.M. Thompson,D. P. McNamara,M. R. Schoeberl,, L. R. Lait, P. A. Newman, C. O. Justice, and J. D. Kendal, A trajectory modeling investigationof the biomassburning tropical ozone relationship,in Proc. Quad. Ozone Symp.,NASA Goddard SpaceFlight Center, Greenbelt,Md., 1992. Pickering, K. E., A.M. Thompson, D. P. McNamara, and M. R. Schoeberl,An intercoparisonof isentropictrajectoriesover the

south Atlantic, Mon. WeatherRev., 122, 864-879, 1994. Fraedrick, K., A simple climatologicalmodel of the dynamicsand energeticsof the nocturnal circulationat Lake Victoria, Q. J. R. Pickering,K. E., A.M. Thompson,D. P. McNamara,M. R. Schoeberl, Meteorol. Soc., 98, 332-335, 1972. R. O. Loring, M. V. Watson, K. Fakhruzzaman,and A. S. Bachmeier, TRACE-A trajectoryintercomparison,1, Effectsof different Franzen, L. G., M. Hjelmroos, P. Kallberg, E. Borstrom-Lunden, input analyses,J. Geophys.Res.,101, 23,909-23,925, 1996. S. Juntto,andA. Sarolainen,A casestudyof long-rangetransportof soil, pollen and stable organic compounds,Atmos. Environ., 28, Pitty,A. F., Particlesizeof the saharandustwhichfell in Britainin July 3587-3604, 1994.

Garstang,M., P. D. Tyson,R. Swap,M. Edwards,P. Kallberg,andJ. A. Lindesay,Horizontal and vertical transport of air over southern Africa, J. Geophys.Res.,101, 23,721-23,736,1996. Gordon, H. B., and B. G. Hunt, Droughts, floods and sea surface temperatureanomalies,A modelingapproach,J. Clim., 11,347-367, 1991.

Griffiths, J. F., Climates of Africa, in East Africa: It's People and Resources, editedby W. Morgan, chap.9, pp. 109-117, Oxford Univ. Press, New York, 1969.

Harangozo, S., and M. S. J. Harrison, On the use of synopticdata in indicatingthe pressureof cloudbandsover southernAfrica, S. Afr. J. Sci., 79, 413-422, 1983.

Ininda, J. M., Spatial and temporalcharacteristics of droughtin eastern and southernAfrica, M.Sc. thesis,Dep. of Meteorol., Univ. of Nairobi, Kenya, 1987. Ininda, J. M., Numerical simulation of the influence of SST anomalies

on the East African seasonalrainfall, Ph.D. thesis,Dep. of Meteorol., Univ. of Nairobi, Kenya, 1995. Kiangi, P.M. R., and J. J. Temu, Equatorialwesterliesin Kenya, are they alwaysrain laden?,in Proceedings of the WMO RegionalScientific Conferenceon GATE, WAMEX and TropicalMeteorology,pp. 144-146, World Meteorol. Organ., Geneva, 1984.

1968, Nature, 220, 364-365, 1968.

Prospero,J. M., R. A. Glaccum,and R. T. Nees, Atmospherictransport of soil dust from Africa to SouthAmerica, Nature, 289, 570572, 1981.

Schiirtz,L., Long rangetransportof desertdustwith specialemphasis on the Sahara,in Aerosols: Anthropogenic and Natural, Sourcesand Transport,edited by T. J. Kneip and P. J. Lioy, pp. 515-532, 338, Annals of the New York Academy of Sciences,1980. Sturman,A. P., P. D. Tyson,and P. D'Abreton, A preliminarystudyof the transportof air from Africa and Australiato New Zealand,J. R. Soc.New Zealand,27(4), 485-498, 1997. Swap,R. J., M. Garstang,S. Greco,R. Talbot, •/nd P. Kallberg,Saharan dust in the Amazon Basin, Tellus, 44, 133-149, 1993.

Swap,R. J., M. Garstang,S. A. Macko, P. D. Tyson,W.• Maenhaut, P. Artaxo, P. Kallberg, and R. Talbot, The long range transportof southernAfrican aerosolsto the tropicalsouthAtlantic,J. Geophys. Res., 101, 23,777-23,791, 1996.

Thompson,B. W., The Climate of Africa, 132pp., Oxford Univ. Press, New York, 1965.

Tyson, P. D., and P. C. D'Abreton, Transport and recirculationof aerosolsoff southernAfrica: Macroscaleplume structure,Atmos. Environ., 32, 1511-1524, 1998.

Tyson,P. D., M. Garstang,R. Swap,P. Kallberg,and M. Edwards,An

14,244

GATEBE ET AL.: SEASONAL TRANSPORT

air transportclimatorologyfor subtropicalsouthernAfrica,J. Clim., 16, 265-291, 1996a.

Tyson,P. D., M. Garstang,and R. Swap,Large-scalerecirculationof air over southernAfrica, J. Appl. Meteorol.,35, 2218-2236, 1996b. Tyson,P. D., M. Garstang,A.M. Thompson,P. C. D'Abreton, R. D. Diab, and E. V. Browell,Horizontal andverticaltransportof ozone over south central southern Africa, J. GeophysRes., 102, 10,62310,635, 1996c.

CLIMATOLOGY

FOR KENYA

C. K. Gatebe, Institute of Nuclear Science,Universityof Nairobi, P.O. Box 30197, Nairobi, Kenya. (e-mail: charles.gatebe@lion. meteo.go.ke) G. Helas,Max PlanckInstitutefor Chemistry,Biogeochemistry Department, P.O. Box 3060, D-55020 Mainz, Germany. P. D. Tysonand S. Piketh,ClimatologyResearchGroup,University of Witwatersrand,Johannesburg,P.O. Wits, 2050, South Africa.

Watson, C. E., H. Fishman, and H. J. Reichle, The significanceof biomassburningas a sourceof carbonmonoxideand ozone in the southernhemispheretropics:A satelliteanalysis,J. Geophys.Res., 95, 16,433-16,450, 1990.

H. Annegarn, SchonlandResearch Center for Nuclear Sciences, University of Witwatersrand,Johannesburg,P.O. Wits, 2050, South (ReceivedJuly20, 1998;revisedNovember9, 1998; Africa. acceptedNovember20, 1998.)