JOURNAL OF GEOPHYSICAL
RESEARCH, VOL. 106, NO. D22, PAGES 28,347-28,356, NOVEMBER
27, 2001
Remote sensing of aerosol optical characteristics in sub-Sahel,
West
Africa
G. Pandithurai, •,2 R. T. Pinker, • O. Dubovik, 3,4B. N. Holben, 3 and T. O. Aro s
Abstract. We have determinedthe characteristics
of sub-Sahelianaerosolsfrom a 2-year
record of continuousground-basedmeasurements,made at the University of Ilorin, Ilorin
(08ø19'N,04ø20'E),Nigeria, in cooperationwith the Aerosol Robotic Network.
Observationsof spectralaerosoloptical depthsduring the dustyharmattan seasonindicate
more than a twofoldincrease,when comparedto other seasons.Retrieved columnar
volume size distributionsshowthe existenceof bimodalitywith a dominant coarsemode.
The retrievedsize distributionswere groupedaccordingto different rangesof aerosol
optical depthsto characterizethe aerosolsfor this particular region. Monthly meansof
retrievedsingle-scattering
albedosshowa sharpdecreasefrom -0.95 to -0.85 at 500 nm
from the preharmattanto the harmattanseasonwhen biomassburningis alsopracticed,
increasingthe presenceof absorbingaerosols.On the basisof thesecomprehensive
observations,
we proposeto augmentexistingdesert aerosolmodels,as presentedin the
literature, to better characterizethe dust outbreak seasonin West Africa, which is quite
prolongedand overlapswith the biomassburningseason.
1.
Introduction
It hasbeen observedthat eachyear, betweenNovemberand
March, a large amount of dust is transportedover Nigeria to
the Gulf of Guinea [Kalu, 1979].The cold and dry wind, which
is the main agentfor this dusttransportis known as the "harmattan."During thisseason,the atmosphereoverthisregionis
characterizedby high dust levels,relativelyhigh daytime and
low nighttime temperatures,dry and weatheringvegetation,
high incidenceof forest fires, and poor visibility.The harmattan aerosols have been known to have adverse health affects,
acterizationof harmattan dustwas alsocarried out [Coxet al.,
1982;Adedokunet al., 1989;Afeti and Resch,2000];however,a
complete optical characterizationof the harmattan aerosols
and long-term remote measurementsare almostnonexistent.
Recently,it has been speculatedthat sincethe onsetof the
droughts that started in the 1970s, the Sahel is the major
sourceregionof wind-bornedust,contraryto earlier belief that
it is the Sahara desert [N'TchayiMbourou et al., 1994, 1997].
Nicholson[2000] reported that (1) there has been a steady
increasein the frequencyof occurrenceof dust conditionsat
the surfaceover West Africa sincethe early 1970s,(2) an
increasein dust occurrencehas paralleled a decreasein rainfall, and (3) the lengthof the seasonwith dustyconditionshas
steadilyincreased.
Several studiessuggestthat dust aerosol has a significant
impact on meteorologicalprocesses
[Karyampudiand Carlson,
1988;Chang,1993;Alpert
etal., 1998].Generalcirculation
(GCM)
studies[TegenandMiller, 1998]haveshownthat regionalclimate
cannotbe realisticallysimulatedwithout introducingdustmobi-
suchas causingacute respiratoryinfections,pneumonia,and
bronchitis,as documentedfor northern Nigeria, where the
effect is most severe,byAdefolalu[1984] andAdedokunet al.
[1989].The atmospherichaze and the occasionalearly morning/eveningfogsdue to northwardincursionsof moist southwesterlywindslead to frequentcancellationof flights,resulting
in a heavy lossof revenue [Adefolalu,1984]. Moreover, the
"harmattanhaze" is a threat to safetyof civil aviation.Out of
the 11 air disastersin Nigeria and the Ivory Coast during the lization in the model, and modelshave shownthat the dust has an
periodNovember1969to January2000(http://www.airdisasters.impacton synopticconditionsand thereforeon rainfall [Tegen
of dust
co.uk/Disasters.htm),
10 occurredduring the harmattan sea- and Fung, 1994;1995].Hence long-termmeasurements
characteristics are needed in climate research.
son.The harmattandustepisodeof January30, 2000, and the
In this paper we documentthe annual variation of aerosol
corresponding
opticalparametersasobservedat the Ilorin site
optical characteristics
obtainedfrom 2 yearsof ground-based
are discussed
by Pinkeret al. [2001].The harmattandusthaze
measurements,which were made as part of a NASA Earth
is a well-knownphenomenonand was first investigatedusing
Observing System (EOS) validation program, within the
visibilitydata [Adebayo,1989;McTainsh,1980].Chemicalcharframeworkof the Aerosol Robotic Network (AERONET).
•Department
of Meteorology,
University
of Maryland,CollegePark,
Maryland, USA.
gindianInstituteof TropicalMeteorology,
Pune,India.
3Laboratory
for TerrestrialPhysics,
NASA GoddardSpaceFlight
2.
Experimental Site and Data
The experimentalsite is locatedon the campusof the Universityof Ilorin, Ilorin, Nigeria (08ø19'N,04ø20'E,350 m above
4AlsoatScience
Systems
andApplications,
Inc.,Lanham,
Maryland, mean sealevel (amsl)), at the uppertip of the Guinea savan-
Center, Greenbelt, Maryland, USA.
USA.
5University
of Ilorin,Ilorin,Nigeria.
Copyright2001 by the American GeophysicalUnion.
Paper number 2001JD900234.
0148-0227/01/2001 JD900234509.00
nah zone in the sub-Sahel, under the influence of the annual
oscillationof the Intertropical ConvergenceZone (ITCZ).
During the "dry season"(November-March) the prevailing
northeasterlywind, known as "harmattan,"brings in air containingSaharandustfrom the Chadbasinwhen a large amount
28,347
28,348
PANDITHURAI
Ja•t•a•l
ET AL.: SUB-SAHEL
AEROSOLS
tional Space DevelopmentAgency (NASDA) of Japan,
1 9 99
ADEOS-II
mission.
The instrumentusedin the presentstudyto monitoraerosol
propertiesis a CIMEL skyradiometer,of the typeusedin the
AerosolRoboticNetwork(AERONET), whichis a worldwidefederatednetworkoperatedat numerousclimaticallyimportant regionsand led by the NASA GoddardSpaceFlight
Center [Holbenet al., 1998].The instrumentis an automatic
Sun-tracking
skyradiometercapableof measuringboth Sun
and sky radianceat eight spectralchannels.The filters are
centeredat wavelengths340, 380, 440, 500, 670, 870, 940, and
1020nm,locatedin a filterwheel,whichis drivenby a stepper
motor. The observations are transmitted almost in real time via
the EuropeanMeteosat(D. Tanre, private communication,
1997).Detailedinformationon measurement
protocol,radiometricprecision,
calibration
procedures,
andprocessing
methods are describedby Holbenet al. [1998].Data usedin the
presentstudyutilize directsolarand skyradiancein the almucantarfrom April 1998to March2000.The CIMEL skyradiometeropticalheadhasbeenreplacedwith a recalibratedone
in August 1999in order to meet the AERONET maintenance
schedule.
3.
lO > [ms-,]
Figure 1. NCEP reanalysisof surfacewindsoverWest•rica
for JanuaE 1999 and 2000.
of dustparticlesis transportedin a plumedownwindfrom the
sourceregionby the strongnortheasterly
windsmainlyat the
900 and 850 mbar levelsnorth of the ITCZ, taking a southwesterlytrajectoryover Nigeria [Kalu, 1979].In this season,
dustplumescanhavea thickness
of up to 3 km, varystrongly
in aerosolcontent,and reducethe visibilityto lessthan 1 km.
Harmattanwindsare the main mechanism
for dusttransport
and produceseverewinter weather conditionsin West Africa,
known as the "harmattanhaze." During the "wet season"
(April-October), conditionsare typified by moist maritime
southwesterlyflow from the Gulf of Guinea over West Africa.
Figure1 showsthe averagesurfacewindflowpatternsasprovided by the National Center for EnvironmentalPrediction
(NCEP), NumericalWeatherPrediction(NWP) modelanalysisfor January1999and 2000.As evident,the intensityof the
northeasterly
flow in January2000wasmuchstrongerthan a
year before.
Measurements
of aerosolopticaldepthsat thissitestartedin
1987 [Pinkeret al., 1994],measurements
of surfaceradiative
fluxesstartedin 1992[Miskolczi
etal., 1997]andwereupgraded
in May 1995,underthe NASA EOS ValidationProgram.The
siteis beingcontinuously
upgradedto meet the requirements
of theWorldClimateResearch
Programme
(WCRP) Baseline
SurfaceRadiationMeasurement(BSRN) Network[Ohmuraet
al., 1998]. The radiation observationsmade at the site are
archivedat the World RadiationData Center(WRDC) locatedat the Eidgenossische
Technische
Hochschule
(ETH),
Zurich,Switzerland.
Observations
of aerosolopticalproperties
at Ilorin, Nigeria,havebeenupgradedin May 1998,aspart of
the AerosolRoboticNetwork(AERoNET) activity,in support of current and future satellitemissionssuchasthe Clouds
Results
and Discussion
3.1. SpectralAerosol Optical Depth
Timeseriesof aerosolopticaldepthsat 500nm andmonthly
meansat all the wavelengths
are shownin Figures2 and 3a,
respectively.Optical depthsduring Decemberto March for
bothyearsare higherwhencomparedto othermonths.Thisis
due to the dust outbreaks(natural) and biomassburning
events(anthropogenic)
mostcommonduringthe dryseasonat
the observational
site.A strongspectraldependence
in aerosol
opticaldepths(AODs) is alsoevident.Thissuggests
that aerosolsfrom biomassburningare presentduringthe dry season
andaresignificant.
The day-to-day
variabilityof aerosoloptical
depthwasconsiderably
higherduringthe dryseasonthanduring the otherseasons,
dueto dustandbiomass
burningevents.
In thepast,N'TchayiMbourouetal. [1994,1997]usedvisibility
data to examinethe spatialand temporaldistributionof dust
over West Africa. From those studies it was evident that there
hasbeen a steadybuildupof dustover West Africa sincethe
early 1970s, and it is consistentwith the increasedtrend in
African dust as measuredat sitesdownstream[Carlsonand
Prospero,1972].While there are severalsourceareasfor dust
aerosols,
the areaof Bilma(Niger)andFayaLargeau(Chad)
are the main sourceregionsfor dust over Nigeria [Wilson,
o
3.0-
[ 2000
2.5-
1998
1999
0
2.0.
o
0
•1.$
oøo
0
o •
o o ø
•.o. o
• •
0.0 • , ,
1O0
&•o
o o • ••
0 - o- o oo•O••a•
,
o
• , ,
200
,
• , ,
300
o
•_ o
•o
o
•oo
, • , , ,
400
500
600
700
800
Day Number(fromApril 1998)
and Earth RadiantEnergySystem(CERES), the Moderate- Figure 2. Time seriesof dailymeanaerosolopticaldepthsat
ResolutionImagingSpectroradiometer
(MODIS), andtheNa- 500 nm as observedat Ilorin, Nigeria.
PANDITHURAI
ET AL.: SUB-SAHEL
AEROSOLS
28,349
1.6
1.4
•1020
nm
ß870
nm
+380
nm
-•
1.2
ß,e
ßß
670 nm
•
---•---
500 nm
440 nm
X
340 nm
0.6
0.4
0.2
0.0
[
I
'
M98
I
'
J98
I
'
S98
I
'
I
N98
'
J99
I
M99
'
I
'
M99
I
'
J99
I
'
S99
I
N99
'
I
J00
'
I
MOO
Figure3. Month-to-month
variationin spectralaerosolopticaldepths(in sevenspectralintervals)as
observedat Ilorin, Nigeria. duringApril 1998 to March 2000.
1971;Kalu, 1979].In summer,AODs are found to be lower
overNigeriawhendustaerosols
followa trajectoryoversouthernAlgeria,Morocco,andacross
theAtlanticto the Caribbean
Islands[Martin,1975;Prospero,
1999].
In Figure4 a comparison
of the intensities
of directshortwaveradiation,asmeasuredwith an EppleyNormal Incidence
Pyrheliometer
duringtwodifferentdaysin the harmattanseason,is presented.
The top panelis for a high-aerosol
loading
1.5
800.•
700 -
Feb 10, 1999
Clear Sky
.
1.4
600
.
500
1.3
.
400
.
300
1.2
.
•-. 200
E
-
• 100
-,-'
0
._
rh
6
ß
800
•
700
1.2 m__.
Feb
07,1999
o
..c
600
<--Cloudy
Sky
'•
1.0
500
400
300
case(% (500 nm) ---1.3)whenthe maximumnormalincidence
shortwave
radiationwasonly400W m-2. The bottompanelis
for a relativelylow-aerosolloadingcase(% (500 nm) ---0.8)
when normal incident shortwave radiation reached as much as
600 W m -2. This is to illustrate the drastic radiative effects
duringhigh-aerosolloadingdays,which are frequentduring
the harmattan
season.
3.2. Intraseasonal Variation in Spectral Aerosol Optical
Depth and Precipitable Water Vapor
To studythe intraseasonal
variationin spectralaerosoloptical depthsand in precipitablewatervapor,the observations
weregroupedintofour seasons,
asshownin Figure5. There is
a significant
increasein aerosolopticaldepthand strongspectral dependence
duringDecember-March,ascomparedto the
other seasons.Aerosol optical depthsremained almostconstantduringthe other seasons,
but there was an increasein
spectralAODs in 1999-2000ascomparedto 1998-1999.Precipitablewater vapor,as derivedfrom the CIMEL observationsin the spectralchannels870 and 940 nm, alsoshowsa
systematic
seasonal
variation,high duringJune-Augustand
low duringDecember-February.The precipitablewatervapor
asavailablefrom the NCEP reanalysis
datais shownalongwith
the CIMEL-derived values for comparison.SpectralAODs
duringthepreharmattan
season
(SON) dependstrongly
on the
start of the biomassburning of the savannahs.Interannual
variationof AODs suggestthat the lengthof the seasonwith
dustyconditions
is longerin the 1998-1999harmattanascompared to the 1999-2000 harmattanseason.
200
3.3.
100
0.6
øI
6
' I ' I ' I ' I ,oI '
8
10
12
14
16
18
Size Distribution
The spectraldependenceof AOD containsinformation
aboutthe sizeof the particles[Junge,1955;Rangarajan,1972;
Pandithurai et al., 1997; Rerner et al., 1999]. The Angstrom
exponenta, whichis a measureof the sizedistribution,
canbe
Figure 4. Diurnal variationof normal incidenceshortwave obtainedby fittinga powerlaw to the aerosolopticaldepths
radiationand aerosolopticaldepths(at 500 nm) duringhigh- andwavelength[Ecket al., 1999;Reidet al., 1999],asgivenin
Time (Hrs)
and low-aerosolloading days.
the followingexpression:
28,350
PANDITHURAI
ET AL.: SUB-SAHEL AEROSOLS
1.5
7
Precipitablewater vapor
•CIMEL
!r2'272•
340nm 1727-/I380nm
•
•
440nm •
rrmm•
670nm •
•
r:•
550nm
870nm
6
1.0
5
'5'
O.5
2
0.0
'
:
AM98
JJA98
SON98
DJF98-99
MAM99
JJA99
SON99
1
DJF99-00
Season
Figure 5. Intraseasonalvariation of spectralaerosoloptical depthsand precipitablewater vapor, as observedat Ilorin, Nigeria. Precipitablewater was derivedboth from the CIMEL skyradiometerobservations
and from the NCEP reanalysis.
ß
x
Angstrom exponentscomputed from such spectral dependence were grouped into monthly means and are shown
in Figure 6. In general, the exponentis higher when there
is a relative dominance of small particles. The monthly
mean values of the Angstrom exponentshow higher values
during both July-Septemberand November-January.The
largernumberof accumulation
modeparticlesobservedduring
November-Januaryis due to aerosolsfrom biomassburning.
Size distributionswere retrieved using a radiative transfer
algorithmdevelopedbyDubovikandKing[2000].Opticalcharacteristics
of aerosolssuchasopticalthickness,
phasefunction,
and single-scattering
albedoare modeledfrom microstructure
parametersusingthe followingapproximations:
,ext(X)P(0; X) =
•rm
max
Kscat(0,X, m, r)n(r) dr,
in
where Kscatis a scatteringcrosssection,P(0, X) is the phase
function,and n(r) denotesparticlenumbersize distribution.
Aerosolsare assumedto be sphericalparticlesand approximated by Mie functionsderived for sphericaland homogeneousparticleswith a complexrefractiveindexm (,•). Volume
size distributionswere retrieved using a detailed radiative
transfer Sun-skyradiance algorithm by Dubovik and King
[2000].More detailson the procedureand aboutthe accuracy
of the retrievedaerosolopticalpropertiescan be found in the
work of Duboviket al. [2000].
The retrievedvolume size distributionsgenerallyrepresent
bimodal
or trimodal
distributions.
About
135 retrieved
size
distributionswere grouped as a function of aerosoloptical
depth. Figure 7 illustratesthe aerosolsize distributionsas a
function of AOD at 670 nm. A bimodal distribution is evident,
1.4-
and for intermediateoptical depthvalues,a secondarycoarse
modeis present.In general,moderadii decreasefor increasing
opticalthicknessvaluesduringthe harmattanseasondustep-
1.2-
isodes, accumulation mode radii shift from 0.10 to 0.07 /•m,
and coarsemode radii vary from 2.0 to 4.0/•m. Characterizing
particlesas sphericalintroducebiases,namely,the appearance
of an artificiallyhigh mode of smallparticlesin sizessmaller
than 0.1/•m may be an artifact of nonsphericity.
0.6-
0.4-
0.2-
3.4.
Single-ScatteringAlbedo
Single-scattering
albedo(•Oo)of atmosphericaerosolsis an
important parameter for obtainingaccurateestimatesof the
Figure 6. Month-to-month variation in Angstrom exponent Earth's radiativebudget.Proceduresto retrieve •oo are not yet
as observedat Ilorin, Nigeria, during April 1998 to March well established due to the limited information content of
2000.
opticalmeasurements.
The inversionmethodusedin thisstudy
M98
J98
S98
N98
J99
M99
M99
Month/Year
J99
S99
N99
JO0
MOO
PANDITHURAI
ET AL.' SUB-SAHEL
AEROSOLS
1.00
1.4
ß
0.96
1.2
o
•670'
1.0
•
m
--o--
o
•
,, ,•, _ ,,•,
670 nm
--,•-- 870 nm
•IL,,•',•51•
•':', ,,',•', ---•---1020
nm
'•• /
0.92
'•,,'
•-" '.•--,•-•'•',",
Id2, •-,_ U-•-•',' '.'•
0,34
---(:>--0.61
ß'
• 0.88
--.A..0.68
I-.-o.741
i ---o--0.84I
o.8
•
28,351
ß .0'\ '•
'•,
._
i_._0.98I
1_•_1.•91
•
0.84
,•.•-
E
'---'
0.80
13.6
,
0.0
' 0•.2 ' 0'.4 ' 0:6 ' 0•.8
I
,
' 1:2
1.4
Angstrom exponent
Figure 9. Single-scatteringalbedo rooas a function of the
Angstrom exponentfor wavelengths440, 670, 870, and 1020
0.4
nm.
0.2
0.0
,
,
,
, ill,
0.01
,
,
•
0•.1
, 111
,
,
,
'•i
1;
.............
1oo
Radius (;am)
Figure 7. Proposedsizedistributionmodel for differentvalues of aerosoloptical depth.
allowsglobal fitting of spectraland multiangleSun/skyradiance with a simultaneous
search for the size distribution
and
complexrefractiveindex.Accuracyassessments
of the inversion code by Dubovik et al. [2000] indicate that the error in
retrieved roois smallerfor higher optical thickness,-0.03 for
ra (440nm) -> 0.5, andfor solarzenithanglesgreaterthan45ø.
Becauseopticalthicknessis highover Ilorin, Nigeria, throughout the year, and since the recommendedcriteria for solar
zenithangleswere applied,the expectederrorsin the retrieved
rooare low.
In theretrieval
process,
[heaerosol
particles
areassumed
to
be polydispersed
homogeneous
sphereswith the samecomplex
refractive index. Tests have shown that the assumptionsof
homogeneous
sphereswith an effectivecomplexindex of refractionallowboth the accuratefitting of the majorityof actual
observationsof atmosphericradiance and the appropriate
modelingof the main radiativeeffectsof the aerosol(absorption and scattering).The current estimatesof accuracyfor
desertdust or other aerosoldominatedby coarseparticlesare
as follows:dV(r)/ln r, 15-25% for r -> 0.5 ram;tOo(k),0.03;
k(x), S0%;
0.0S.
If the nonsphericityassumptionis not satisfied,it is possible to obtain nonrealistically high fine mode with maximum at r < 0.1 ram. This effect is at a maximum for high
solar zenith angles and at a minimum for low solar zenith
angles(20ø-30ø).
Monthly and seasonalmean valuesof spectralrooare shown
in Figure 8. To increasethe reliability of the retrieval, only
observationsfor which solar zenith angleswere greater than
45øwere used.This resultedin missingvaluesfor somemonths.
The spectraldependenceof rooshowsincreasingvalueswith an
increasein wavelength,and a reversewavelengthdependence
of roocan be noted in the dry season,which is due to the
dominanceof absorbingaerosolsfrom biomassburning. Seasonalmeansof rooindicatelow values(<0.9) during DJF of
1998-1999 and 1999-2000. Lower valuesof a,o duringSON of
1999 are due to the early start of biomass burning in the
vicinityof the observingstation.Differencesbetweenthe spectral dependenceof rooof DJF of 1998-1999 and DJF of 19992000 indicate the dominanceof dust aerosolsduring January
1.60
0.98
1.55
0.94
m 0,92
.•_
.-.
1.50
•
1.45
•
1.40
,_
• 0.90
.• 0.88
._>
--
670 nm
--A--
870 nm
'7
• 1.35
440 nm
---e---
=
I• 1.30
440 nm
---o--
670 nm
---•--
1020 nm
870 nm
)•
1.25
1020
nm
1.2O
M98
J98
S98
N98
J99
M99
M99
J 9
S 9
N 9
J 0
M 0
I
Month/Year
M98
'
I
J98
,
I
S98
'
I
N98
'
I
J99
'
I
'
M99
I
M99
'
i
J99
,
i
S99
,
i
N99
,
I
J00
'
I
MOO
Month
Figure 8. Month-to-month variation in spectral singlescatteringalbedoas observedat Ilorin, Nigeria, duringApril
Figure 10. Month-to-month variation of the real part of the
1998 to March
refractive
2000.
index.
28,352
PANDITHURAI ET AL.: SUB-SAHELAEROSOLS
0.018
1.2
--
0.015
._E 0.012
+AVHRR
(4N,
5E)I
440 nm
.-.-e...
670 nm
-- •---
870 nm
.... v---.1020
--
CIMEL
nm
x
•
0.009
._>
0.4
'• 0.006
0.2
0.003
0.0
0.000
'
i
M98
,
i
J98
•
i
S98
,
i'
N98
,
i
'
J99
i
M99
'
i
M99
'
i
'
J99
i
S99
'
i
N99
'
i
'
J00
'
I
M98
i
'
I
'
J98
I
S98
'
I
N98
'
I
'
J99
MOO
I
'
M99
I
M99
'
I
J99
'
I
S99
'
i
N99
,
I
J00
'
i
MOO
Month
Month
Figure 13. Comparisonof CIMEL-derivedaerosoloptical
Figure 11. Month-to-month
variationof the imaginarypart depthsat 670 nm and the NOAA 14 AVHRR operational
of refractive index.
aerosolopticaldepth[Stoweet al., 1997].
andFebruary2000.Frequency
distribution
of single-scattering
and composition
of the aerosol.The
albedoat all fourwavelengths
indicatesa modalvalueof >0.95 generationmechanisms
maybe represented
by
(mostlyof scattering
aerosols),
but the 2-yearmean(<0.9) bulkopticalpropertiesof the aerosols
indexof refraction
(m = r/real-- inimag)
whichisa
suggests
anabsorbing
typeof aerosols
(biomass
burning).
Vari- a complex
able elementalratios reflect variable mineralogicalcomposi- functionof wavelengthin bothvisibleand infraredregionsof
The concepts
for determining
aerosolrefractive
tion of soils.Dust in the Sahelianregionis characterizedby a the spectrum.
are based
highFe/A1ratio due to abundance
of ferraliticsoilsin the indicesfrom multiangularradiancemeasurements
of theeffectsof refractive
Sahel [Sokolikand Toon, 1999]. In Figure 9 the single- on theprincipleof partialseparation
scatteringalbedois presentedas a functionof the Angstrom index and size distributionon the angular variability of sky
andvonHoyningen-Huene,
1994;Yamasoe
et
exponent.
Asevident,
several
distinct
sections
canbeidentified radiance[Wendish
in thisfigure,startingwithrelativelyhighervaluesof too above al., 1998].The retrievalmethodof DubovikandKing[2000],
0.95 at lowvaluesof the Angstromexponent,characteristic
of usedin the presentstudy,implementsretrievalvia simultadust,and graduallydecreasing
towardlowervaluesof too at neousfitting of radiancesmeasuredin the entire available
highervaluesof the exponent,
characteristic
of a highercon- angularand spectralrange and thus can improveretrieval
A sufficientlyaccurateretrievalof aerosolsinglecentrationof aerosolsfrom biomassburning,indicatingclearly accuracy.
albedoandcomplexindexof refractionarepossible
that dustis muchlessabsorbing
than aerosolsemanatingfrom scattering
onlyfor high-aerosol
AOD. On the basisof a sensitivity
study
the burningof biomass.
[Duboviket al., 2000]it wasfoundthat the AOD at 440 nm
3.5. Complex Refractive Indices
shouldbe largerthan 0.2. This conditionis satisfiedin the
The aerosolrefractiveindicesare determinedby the physical presentstudy.
The retrieved real and imaginaryrefractiveindiceswere
andchemical
properties
of the aerosol,namely,in termsof the
4
100
200
300
400
500
600
700
800
Day Number(FromApril1998)
Figure12. Comparison
of CIMEL-derived
aerosolopticaldepthsat 340nm andTOMS-derived
aerosol
indexfor a 2-yearperiod[Hermanand Celarier,1997].
PANDITHURAI
Table
1.
Sub-Sahelian
Season
Harmattan
Aerosol
*a (440)
0.895
(Nov.-March)
ET AL.: SUB-SAHEL AEROSOLS
28,353
Model a
,a (500)
,• (670)
0.830
0.695
0.59b
0.653c
0.29 f
0.485 e
,• (870)
,• (1020)
0.602
0.558
PWV, cm
2.787
2.76c
1.35d
0.333 g
Nonharmattan
(April-Oct.)
0.476
1.13 e
0.455
0.405
0.431b
0.704
e
0.372
0.363
4.620
2.75d
0.19 f
3.84 e
0.273 g
aSpectralaerosoloptical depthsand precipitablewater vapor information.
bdVtlmeida
[1987],measurements
madeat Zaria,Nigeria,during1981-1982.
CFaizounet al. [1994], measurementsmade at Ouangofitiniduring 1985-1987.
dTuller[1968].
eFaizounet al. [1994], measurementsmade at Bidi during 1987-1989.
fTegenet al. [1997],obtainedfrom monthlymeantotalsof nine individualspecies,
derivedfrom
transportmodels,in a grid cell over Ilorin, Nigeria.
gValuesfrom the Global Aerosol Data Sets(GADS) by Koepkeet al. [1997] for winter (0% RH) and
summer(70% RH), at 500 nm over a 10øN,5øEgrid box.
groupedinto monthlymeansand are shownin Figures10 and
11. Monthly meansof the real part of the refractiveindexshow
a gradual decreaseduring the wet season.Starting from November,there is an increasein thesevalues,peakingin January.The artificialspectraldependenceobservedin the real part
of refractive index is mainly due to nonsphericityerrors, and
therefore onlyvaluesfor 870 and 1020nm shouldbe used.The
frequency distributionof the real refractive indices gives a
mean value of 1.46 at 670 nm, which is in agreement with
refractiveindicesof quartz and silicates[Krekov,1993]. Mineralogical analysisof the harmattan dust indicated that it is
predominantlycomposedof quartz (>70%) followed by microcline,kaolinite, and tracesof mica and halloysite;chemical
analyses
indicateda predominanceof SiO2(>60%) [Adedokun
et al., 1989].
Monthly means of imaginary refractive indices also show
significantseasonalvariationwith highvaluesin the dry season
(more absorbingaerosols,suchas mineral dust and biomass)
and low valuesin the wet season(scattering-typewater soluble). Interannualvariationsshowa significantspectraldependence during the 1999-2000 harmattan season.A statistical
analysisof retrieved imaginaryrefractive indicesover Ilorin,
Nigeria, yieldsa mean value of 0.0065 at 670 nm. Lindbergand
tion of retrieved values from recently launched instruments,
suchas the EOS MODIS on the TERRA mission[Kinget al.,
1992].To comparethe groundmeasurements
with the TOMS
aerosolindex,whichis derivedmainlyfrom the 340 and 380 nm
reflectances,CIMEL-observed aerosoloptical depths at 340
nm were groupedinto dailymeansfor 2 years(from April 1998
to March 2000). TOMS aerosolindexover Ilorin, Nigeria, was
computedby taking weighted means from four neighboring
grids of 1ø spatial resolution. A comparisonof daily mean
AOD at 340 nm and TOMS aerosolindex is shownin Figure
12, illustrating good agreement in the detection of the dust
Laude[1974]measured
F/imag
asa functionof wavelength
for
3.7.
desertdust at visibleand near-infraredwavelengthsby diffuse
reflectance spectroscopy,and their values decreasedfrom
-0.01 at 400 nm to -0.006 at 600 nm. Lindbergand Gillespie
[1977] reportedthat the larger soil particles,thosewith aerodynamicradii >5.5 /•m, showedsignificantlyless absorption
than the smallerparticles,reflectingon differencesin composition. It is evident from our observationsthat the imaginary
refractive indices are found to be low during the dust storm
episodes.The seasonalmeansindicatehigh imaginaryrefractive indicesduringthe dry season,which is dominatedby biomassburningeventsand soildusttransportto the site [Nicholson, 2000;Kalu, 1979].
Desert aerosolshave large variability in their physicaland
opticalproperties,when comparedto other typesof aerosols.
Generally, during the winter season,the aerosolsare considered as desertbackgroundaerosols,and thoseduringthe summer seasonare termed aswind carried [d9tlmeida,1987].This
3.6.
outbreak.
Monthly means of analyzed aerosol optical depths at the
closestocean grid (4øN, 5øE) to Ilorin (08ø19'N, 04ø20'E),
Nigeria, is shown in Figure 13, along with CIMEL-derived
AODs at 670 nm. The NOAA operationallyretrieved AODs
compare well with the nearest ground-basedmeasurements,
being slightlylower, as shouldbe expected.The possibledifferencesmay be due to excesscloud screeningin the NOAA
product, or the difference in distancebetween the two sites.
The much higher valuesobservedduring November 1999 over
Ilorin may be due to local biomassburning effects.
"Sub-Sahelian
Aerosol
is not the case over sub-Sahel.
Table
2.
Sub-Sahelian
Season
Harmattan
Comparison With TOMS and AVHRR Data
Ground-basedremotelysensedaerosoldata over this region
can be usedto evaluateexistingspace-basedaerosolretrievals,
suchas thosefrom TOMS [Hermanand Celarier,1997; Torres
et al., 1998] and AVHRR [Stoweet al., 1997; Nakajima and
Higurashi,1998;Chowdharyet al., 2001] and for future valida-
Dust"
Nonharmattan
Dust mobilization
Aerosol
to0 (440)
Model
Model a
to0 (670)
to0 (870)
0.880
0.887
0.887
0.71b
0.75b
0.72b
0.71 c
0.92 c
0.929
0.932
0.935
0.75b
0.79b
0.83b
aSingle-scattering
albedo.
bd9tlrneida
[1986,1991].
CCarlson
and Cavefly[1977].
occurs over
to0 (1020)
0.889
0.938
28,354
PANDITHURAI
Table
3.
Sub-Sahelian
Aerosol
ET AL.: SUB-SAHEL AEROSOLS
Model a
Ref Index
Season
Harmattan
Ref Index (440)
Ref Index (670)
Ref Index (870)
Real
Real
Real
1.436
Imag
0.0085
1.468
0.01b
Imag
0.0074
Imag
(1020)
Real
Imag
1.475
0.0075
1.470
0.0077
1.457
0.0042
1.465
0.0044
0.004b
0.008 c
0.006 d
0.01d
Nonharmattan
1.525 e
1.405
0.005 e
0.0036
1.434
0.0039
areal and imaginaryrefractiveindices.
bpatterson
et al. [1977].
CCarlson
and Caverly[1977].
dLindberg
andLaude[1974].
eGramset al. [1974].
sub-Sahelduring the winter due to favorablemeteorological related with precipitablewater vapor. The proposedsize distribution model suggests
bimodal/trimodaldistributionwith a
model, characteristicof the Sahelian conditions,could aug- dominant coarse mode. Shift in coarse mode radii and trimoment existingdust models and better fit the conditionsin this dal distributionfor higheropticaldepths(ra (6 70) > 1.0) can
region. Moreover, during the dry winter conditionsin this be noted.The seasonal
variationof the single-scattering
albedo
region,biomassburningis practicedand hasa strongeffect on showslowervaluesduringthe dry season,indicatingthe domthe optical properties of the aerosols.The 2-year record of inance of aerosolsfrom biomassburning. Despite the lower
continuousobservations
of directSunand diffuseskymeasure- accuracyin the real and imaginaryrefractiveindicesretrieval,
ments made at this site servesfor constructinga preliminary the mean valuesof the spectralreal and imaginaryrefractive
sub-Saheliandustmodel that describesmore closelyconditions indicesare closeto valuesreported for silicatesand organic
in thisregionthan the availabledesertaerosolmodelsfoundin species.
conditions, such as the harmattan wind. As such, a dust aerosol
the literature.
The proposedcharacteristics
for the sub-Sahelianaerosols
are presented in Tables 1-3. The spectral aerosol optical
depths,single-scattering
albedo,real, and imaginaryrefractive
indices exhibit strong seasonalvariation and therefore were
groupedasfollows:(1) the dryseason(November-March)and
(b) the wet season(July-September).In Table 1 the spectral
aerosolopticaldepthsand precipitablewatervaporfor dry and
wet seasonsare given and comparedwith information from
other sources.In Table 2 the single-scatteringalbedo at different wavelengthsis illustrated and comparedwith results
from previousstudies[dMlmeida, 1987; Carlsonand Caverly,
1977]. In Table 3 the real and imaginaryrefractiveindicesare
presentedfor different wavelengths,both for the dry and for
the wet seasons.Imaginaryrefractiveindicesshowhighvalues
during the dry season,suggestingan abundanceof absorbing
aerosols.The presentstudyis basedon state of the art observations of accurate Sun and sky measurements,as well as
advancedretrieval techniques[Dubovikand King, 2000]. It is
hoped that the proposed aerosol characteristicswill depict
more accuratelythe conditionsin this unique region.
Acknowledgments.This work was supportedunder NASA EOS
Validationgrant NAG-56464 (D. O' C. Starr, ProjectManager) and
grantNAG-11832 (R. J. Curran,ProjectManager).Thanksare due to
the grantingagency,to the AERONET team,to D. Tanre for real-time
transmissionof the CIMEL data, to J. K. Schemmand S. Nigam, for
providingthe NCEP wind data, to C.-J. Meng, for helpingwith the
wind analysis,to F. Miskolczifor technicalassistance
with the instrumentation,and the Universityof Ilorin stafffor the dailymaintenance
of the skyradiometer.
References
Adebayo,S. I., Trajectoriesof advectedSaharandust in Nigeria,Atmos. Environ., 23, 1581-1589, 1989.
Adedokun, J. A., W. O. Emofurieta, and O. A. Adedeji, Physical,
mineralogicaland chemicalpropertiesof Harmattan dustat Ile-Ife,
Nigeria, Theor.Appl. Climatol.,40, 161-169, 1989.
Adefolalu,D. O., On bioclimatological
aspectsof Harmattandusthaze
in Nigeria,Arch. Meteorol.Geophys.
Bioclimatol.Ser.B, 33(4), 387404, 1984.
Afeti, G. M., and F. J. Resch,Physicalcharacteristics
of Saharandust
near the Gulf of Guinea, Atmos. Environ., 34, 1273-1279, 2000.
Alpert, P., Y. J. Kaufman, Y. Shay-E1,D. Tanre, A. da Silva, S.
Schubert,and J. H. Joseph,Quantificationof dust-forcedheatingof
the lower troposphere,Nature,395, 367-370, 1998.
American GeophysicalUnion, The perfect haze: Scientistslink 1999
4. Summary
U.S. pollution episodeto Midwest aerosolplumes and Kennedy
Optical characteristics
of aerosolsobservedover Ilorin, Niplane crash,Eos Trans.AGU, 81(2), 13-14, 2000.
geria, are distinctlydifferent during the harmattan, as com- Bruegge,C. T., J. E. Conel, R. O. Green, J. S. Margolis, R. G. Holm,
and G. Toon, Water vapor column abundanceretrievalsduring
pared to other seasons.The strong spectraldependencein
FIFE, J. Geophys.Res.,97, 18,759-18,768, 1992.
aerosolopticaldepthsduringthe dry seasonindicatesan abunCarlson,T. N., and R. S. Caverly,Radiativecharacteristics
of Saharan
danceof sub-micron-size
particles.During dust storm events,
dustat solarwavelengths,
J. Geophys.Res.,82, 3141-3152, 1977.
spectral aerosol optical depths were found to have near- Carlson, T. N., and J. M. Prospero,The large scale movement of
Saharanair outbreaksoverthe northernequatorialAtlantic,J.Appl.
neutral extinction,with Angstromexponentsbeingreducedto
Meteorol., 11,283-297, 1972.
0.3. Precipitablewater vapor,whichwas retrievedusingdirect
Chang,C.-B., Impact of desertenvironmenton the genesisof African
Sun measurementsat 940 nm [Bmeggeet al., 1992], exhibits
wave disturbances,J. Atmos. Sci., 50, 2137-2145, 1993.
significantseasonalvariation and agreeswell with NCEP re- Cox, R. E., M. A. Mazurek, and B. R. T. Simoneit,Lipids in Harmattan aerosolsof Nigeria, Nature,296, 848-849, 1982.
analysisvalues.Seasonalmeansof AODs are negativelycot-
PANDITHURAI
ET AL.: SUB-SAHEL AEROSOLS
28,355
Chowdhary,J., B. Cairns,M. Mishchenko,and L. Travis, Retrieval of
propertiesover the oceanusingpolarizationas well as intensityof
reflected sunlight,J. Geophys.Res., 102, 16,989-17,013, 1997.
aerosolpropertiesoverthe oceanusingmultispectralandmultiangle
photopolarimetricmeasurements
from the ResearchScanningPo- Mishchenko,M. I., I. V. Geogdzhayev,B. Cairns,W. B. Rossow,and
larimeter, Geophys.Res.Lett., 28, 243-246, 2001.
A. A. Lacis,Aerosol retrievalsover the oceanby use of channels1
d'Almeida, G. A., A model for Saharandust transport,J. Clim. Appl.
and 2 AVHRR data: Sensitivityanalysisand preliminary results,
Meteorol., 24, 903-916, 1986.
Appl. Opt., 38, 7325-7341, 1999.
d'Almeida, G. A., On the variability of desert aerosolradiative char- Miskolczi, F., T. O. Aro, M. Iziomon, and R. T. Pinker, Surface
acteristics,
J. Geophys.Res.,92, 3017-3026, 1987.
radiativefluxesin Sub-SahelAfrica, J. Appl. Meteorol.,36, 521-530,
1997.
d'Almeida,G. A., P. Koepke,and E. P. Shettle,Atmospheric
Aerosols,
Nakajima, T., and A. Higurashi,A useof two-channelradiancesfor an
Global Climatologyand Radiative Characteristics,561 pp., A.
Deepak, Hampton, Va., 1991.
aerosolcharacterizationfrom space,Geophys.Res.Lett., 25, 38153818, 1998.
Dubovik, O., and M.D. King, A flexibleinversionalgorithmfor retrieval of aerosol optical propertiesfrom Sun and sky radiance Nicholson,S., Land surfaceprocessesand Sahel climate, Rev. Geomeasurements,J. Geophys.Res., 105, 20,673-20,696, 2000.
phys.,38, 117-139, 2000.
Dubovik, O., A. Smirnov,B. N. Holben, M.D. King, Y. J. Kaufman, N'Tchayi Mbourou, G. M., J. Bertrand, M. Legrand, and J. Baudet,
Temporal and spatial variationsof the atmosphericdust loading
T. F. Eck, and I. Slutsker,Accuracyassessment
of aerosoloptical
properties retrieval from AERONET Sun and sky radiance meathroughoutWest Africa overthe lastthirty years,Ann. Geophys.,12,
265-273, 1994.
surements,J. Geophys.Res.,105, 9791-9806, 2000.
Eck, T. F., B. N. Holben, J. S. Reid, O. Dubovik, A. Smirnov, N. T.
N'Tchayi Mbourou, G. M., J. Bertrand, and S. E. Nicholson, The
diurnal and seasonalcyclesof wind-borne dust over Africa north of
O'Neill, I. Slutsker,and S. Kinne, Wavelengthdependenceof the
the equator,J. Appl. Meteorol.,36, 868-882, 1997.
opticaldepthof biomassburning,urban,and desertdustaerosols,J.
Ohmura, A., et al., Bull. Am. Meteorol. Soc., 79, 2115-2136, 1998.
Geophys.Res., 104, 31,333-31,349, 1999.
Faizoun, C. A., A. Podaire, and G. Dedieu, Monitoring of Sahelian Pandithurai,G., P. C. S. Devara, P. E. Raj, and S. Sharma,Retrieval
aerosoland atmosphericwater vapor content characteristics
from
of aerosolsizeindexfrom high-resolution
spectroradiometer
observations, Aerosol Sci. Technol., 26, 154-162, 1997.
sun-photometermeasurements,J. Appl. Meteorol., 33, 1291-1304,
1994.
Patterson,E. M., D. A. Gillette, and B. H. Stockton,Complexindexof
Grams, G. W., I. H. Blifford Jr., D. A. Gillette, and P. B. Russell,
refractionbetween300 and 700 nm for Saharanaerosols,J. Geophys.
Res., 82, 3153-3160, 1977.
Complexindex of refractionof airbornesoil particles,J. Appl. Meteorol., 13, 459-471, 1974.
Haywood,J. M., V. Ramaswamy,
and B. J. Soden,Troposphericaerosolclimateforcingin clear-skysatelliteobservations
overthe oceans,
Science,283, 1299-1303, 1999.
Herman, J. R., andE. Celarier,Earth'ssurfacereflectivityclimatology
at 340-380 nm from TOMS data, J. Geophys.Res., 102, 28,00328,012, 1997.
Holben, B. N., et al., AERONET--A
federated instrument network
and data archive for aerosol characterization, Remote Sens.Environ.,
66, 1-16, 1998.
Houghton, J. T., L. G. M. Filho, B. A. Callandar, N. Harris, A. Kat-
ternberg,and K. Maskell (Eds.), IPCC 95, climate change,in The
Scienceof Climate Change,572 pp., CambridgeUniv. Press,New
York, 1996.
Junge,C. E., The size distributionand aging of natural aerosolsas
determinedfrom electricaland optical data on the atmosphere,J.
Meteorol., 12, 13-25, 1955.
Kalu, A. E., The African dustplume: Its characteristics
and propagation acrossWest Africa in winter, in Saharan Dust, edited by C.
Morales,pp. 95-118, 1979.
Karyampudi,V. M., and T. N. Carlson,Analysisand numericalsimulations of the Saharan air layer and its impact on easterlywave
Pinker, R. T., G. Idemudia, and T. O. Aro, Characteristic aerosol
optical depthsduringthe Harmattan seasonin sub-SaharaAfrica,
Geophys.Res.Lett., 21, 1099-1103, 1994.
Pinker, R. T., G. Pandithurai, B. N. Holben, O. Dubovik, and T. O.
Aro, A dustoutbreakepisodein sub-SahelWest Africa, J. Geophys.
Res., in press,2001.
Prospero,J. M., Long-termmeasurementsof the transportof African
mineral dust to the southeasternUnited States:Implicationsfor
regional air quality,J. Geophys.Res., 104, 15,917-15,927,1999.
Rangarajan,S., Wavelengthexponentfor hazescatteringin the tropics
asdeterminedby photoelectricphotometers,Tellus,24, 51-59, 1972.
Reid, J. S., T. F. Eck, S. A. Christopher,P. V. Hobbs, and B. N.
Holben, Use of the Angstromexponentto estimatethe variabilityof
optical and physicalpropertiesof agingsmokeparticlesin Brazil, J.
Geophys.Res., 104, 27,473-27,489, 1999.
Remer, L. A., Y. J. Kaufman, and B. N. Holben, Interannual variation
of ambient
aerosol characteristics
on the east coast of the United
States,J. Geophys.Res., 104, 2223-2231, 1999.
Shettle, E. P., Optical and radiative properties of a desert aerosol
model, in Proceedings
of the Symposiumon Radiation in theAtmosphere,edited by G. Fiocco,pp. 74-77, A. Deepak, Hampton, Va.,
1984.
disturbances,J. Atmos. Sci., 45, 3102-3136, 1988.
Sokolik,I. N., and O. B. Toon, Incorporationof mineralogicalcomKaryampudi,V. M., et al., Validation of the Saharan dust plume
positioninto modelsof the radiativepropertiesof mineral aerosol
from UV to IR wavelengths,
J. Geophys.Res.,104, 9423-9444, 1999.
conceptualmodel using lidar, Meteosat and ECMWF data, Bull.
Am. Meteorol. Soc., 80, 1045-1075, 1999.
Stowe,L. L., A.M. Ignatov,and R. R. Singh,Development,validation,
and potential enhancementsto the second-generation
operational
Kaufman,Y. J., Satellitesensingof aerosolabsorption,J. Geophys.
Res., 92, 4307-4317, 1987.
aerosolproductat the National EnvironmentalSatellite,Data, and
Kaufman,Y. J., and R. S. Fraser, The effect of smokeparticleson
Information Serviceof the National Oceanicand AtmosphericAdcloudsand climate forcing, Science,277, 1636-1639, 1997.
ministration,J. Geophys.Res., 102, 16,923-16,934, 1997.
King, M.D., Y. J. Kaufman, W. P. Menzel, and D. Tanre, Remote Tegen, I., and I. Fung, Modeling of mineral dust in the atmosphere:
Sensingof cloud,aerosolandwatervaporpropertiesfrom the ModSources,transport, and optical thickness,J. Geophys.Res., 99,
22,897-22,914, 1994.
erate Resolution Imaging Spectrometer(MODIS), IEEE Trans.
Geosci.Remote Sens.,30, 2-27, 1992.
Tegen,I., and I. Fung,Contributionto the atmosphericmineralaerosol load from land surface modification, J. Geophys.Res., 100,
Koepke,P., M. Hess,I. Schult,and E. P. Shettle,Global AerosolData
18,707-18,726, 1995.
Set (GADS), MPI Meteorol.Rep. 243, 44 pp., Max Planck Inst.,
Hamburg, Germany, 1997.
Tegen, I., and R. Miller, A general circulation model study on the
interannual variability of soil dust aerosol,J. Geophys.Res., 103,
Krekov, G. M., Models of atmosphericaerosols,in AerosolEffectson
25,975-25,995, 1998.
Climate,edited by S. G. Jennings,pp. 9-64, 1993.
Lindberg,J. D., andJ. B. Gillespie,Relationships
betweenparticlesize Tegen, I., P. Hollrig, M. Chin, I. Fung, D. Jacob, and J. Penner,
and imaginaryrefractiveindex in atmosphericdust,Appl. Opt., 16,
Contributionof different aerosolspeciesto the global aerosolex2628-2630, 1977.
tinctionopticalthickness:
Estimatesfrom modelresults,J. Geophys.
Res., 102, 23,895-23,915, 1997.
Lindberg,J. D., and L. A. Laude, Measurementof the absorption
Torres, O., P. K. Bhartia, J. R. Herman, Z. Ahmad, and J. Gleason,
coefficientof atmosphericdust,Appl. Opt., 13, 1923-1927,1974.
McTainsh,G., Harmattandustdepositionin northernNigeria,Nature,
Derivation of aerosol properties from satellite measurementsof
286, 587-588, 1980.
backscatteredultraviolet radiation: Theoretical basis,J. Geophys.
Mishchenko, M. I., and L. D. Travis, Satellite retrieval of aerosol
Res., 103, 17,099-17,110, 1998.
28,356
PANDITHURAI
ET AL.: SUB-SAHEL AEROSOLS
Tuller, S. E., World distributionof mean monthlyand annualprecipitable water, Mon. WeatherRev., 96, 785-797, 1968.
Wendish, M., and W. von Hoyningen-Huene,Possibilityof refractive
index determination of atmosphericaerosol particles by groundbased solar extinction and scatteringmeasurements,,4tmos.Environ., 28, 785-792, 1994.
Westphal, D. L., O. B. Toon, and T. N. Carlson,A case study of
mobilizationand transportof Saharandust,J.,4tmos.Sci.,45, 2145-
Holben, and P. Artaxo, Retrieval of the real part of the refractive
index of smoke particles from Sun/sky measurementsduring
SCAR-B, J. Geophys.Res., 103, 31,893-31,902, 1998.
T. O. Aro, Universityof Ilorin, Ilorin, Nigeria.
O. Dubovik and B. N. Holben, Laboratoryfor Terrestrial Physics,
NASA Goddard Space Flight Center, Code 923, Greenbelt, MD
20771.
G. Pandithuraiand R. T. Pinker, Departmentof Meteorology,UniWilson, I. G., Desert sandflow basinsand a model for the develop- versity of Maryland, College Park, MD 20742, USA. (pinker@
ment of ERGS, Geograph.J., 137, 180-199, 1971.
atmos.umd.edu)
Wilson,R., and J. Spengler(Eds.),Particlesin OurAir: Concentrations
and HealthEffects,254 pp., Harvard Univ. Press,Cambridge,Mass.,
1996.
(ReceivedAugust31, 2000;revisedMarch 29, 2001;
Yamasoe, M. A., Y. J. Kaufman, O. Dubovik, L. A. Remer, B. N.
acceptedApril 4, 2001.)
2175, 1988.