JOURNAL
OF GEOPHYSICAL
RESEARCH,
VOL. 97, NO. Bll,
PAGES 15,255-15,271, OCTOBER
10, 1992
UpwellingAsthenosphere
BeneathWesternArabia
and Its RegionalImplications
VICTOR
E. CAMP
Department
of Geological
Sciences,
SanDiegoStateUniversity,
SanDiego,California
M. JOHN ROOBOL
Directorate
Generalof MineralResources,
Jiddah,SaudiArabia
Two distinctphases
of continental
magmatism
areevidentin westernArabia. The first,fromabout
30 to 20 Ma,produced
tholeiitic-to-transitional
lavasemplaced
alongNW trends.Thesecond,
fromabout
12 Ma to Recent,producedtransitional-to-strongly-alkalic
lavasemplaced
alongN-S trends.The older
phaseis attributedto passive-mantle
upwelling
duringextension
of the Red SeaBasin,whereasthe
younger
phaseisattributed
to active-mantle
upwelling
butwasfacilitated
byminorcontinental
extension
perpendicular
to platecollision.Theyounger
magmatic
phaseislargelycontemporaneous
witha major
periodof crustalupliftto producethe WestArabianSwellafterabout14 Ma. A varietyof evidence
suggests
thattheWestArabianSwellis thermally
supported
byhot,upwelling
asthenosphere.
In contrast
to thedistinctasymmetry
of upliftandmagmatism
onopposing
sidesof theRedSeaBasin,theseprocesses
weresymmetric
across
a N-Slinemarking
thecentralaxisof theWestArabianSwell.Thisaxiscoincides
with two fundamental features: the Ha'il-Rutbah Arch in the north, and the Makkah-Madinah-Nafud
(MMN) volcanicline in the south. The symmetry
of magmatism
is demonstrated
by petrochemical
evidence
that the MMN harratswerederivedby greaterdegrees
of partialmelting,at shallower
depths,
thanthoseharratslyingto thewestandeastof theMMN line. Thepotentialtemperature
of theasthenosphereis estimated
to be about1436øC
beneath
theMMN lineandabout1354øC
beneaththeflanking,
moreundersaturated
harrats.The sourceof upwelling
is eithera mantleplumecentrallylocatedbeneath
the West ArabianSwellor an elongatedandextendedlobeof hot asthenosphere
emanating
from the
Ethiopian
mantleplume.Convective
flowmayhavebeenchannelled
alonga preexisting,
regional
flexure
in thecontinental
lithosphere
whichconcentrated
hotasthenosphere
beneaththecentralaxisof theAfroArabianDome. Thiscrestof mantleupwelling
underlies
the MMN line in the northandthe Danakil
Depression
and Ethiopianrift systemin the south. It alsopasses
throughthe Red SeaBasinat the
midpointof anunusual,
doublypropagating
rift system
whereaxialseafloor
spreading
began4-5 m.y.ago.
The NW trendof theRed SeaBasinwaswellestablished
by crustalattenuationduringtheoldermagmatic
phase.The Pliocene
invasion
of thisbasinby a N-S zoneof mantleupwelling
hasresultedin seafloor
spreading
parallelto the preexisting
structurealonga rift system
thathascontinued
to propagate
away
from its eccentricmantlesourcein two opposingdirections.
INTRODUCTION
about30 to 20 Ma), they reveala varietyof evidencein
The geographic
association
of uplift,riftingandmagmatism supportof a majorperiodof postriftuplift that beganin the
at the Afar triplejunctionprovidesa visualimpression
consis- mid-Miocene and has continued to the Present.
Fissiontrackdata on crustalapatitesfrom the southeastern
tent with continentalrifting at the crestof a plumegenerated
dome[Cloos,1939;Gass,1970;Burkeand Dewey,1973;Sengor Red Seacoastshowthat the rate of uplift wasacceleratedafter
and Burke,1978]. An activerole for the mantlein the genesis about 14 Ma, althoughit may havebeen initiatedasearly as20
et al., 1989]. An accelerated
periodof uplift in
of the Red Sea and Gulf of Aden rifts, however, has received Ma [Bohannon
little supportfrom the field data. Kinematicmodelssuggest the mid-Mioceneis consistentwith the widespreaddeposition
that theseoceanicrifts were initiatedby tectonicallyinduced of mid-Miocene boulder conglomeratesalong the flanks of
[Schmidtand
extensioninvolving a passivemantle [Cochran,1981, 1983; both the Red Sea and Gulf of Suez depressions
Courtillot et al., 1980; Courtillot, 1982; Bohannon and Eittreirn, Hadley, 1984; Jarrigeet al., 1990] and with the depositionof
1991]. Sucha modelis supported
by a varietyof studiesthat massiveevaporitedepositsin the deeperpartsof the Red Sea
test the relative timing of uplift, rifting, and magmatism graben,indicatinga regressionof the mid-MioceneRed Sea
throughoutthe Red SeaandAfar regions[Schmidtet al., 1983; after about 14 Ma [Gillman, 1968;Bayeret al., 1988].
Here we presentevidencethat this period of uplift has
Schmidtand Hadley, 1984; Almond, 1986a, b; Behre, 1986;
with continenBohannonet al., 1989; Omar et al., 1989; McGuire and Bohan- continuedinto the Holocene,contemporaneous
non, 1989; Bohannonand McGuire, 1990; Jarrigeet al., 1990; tal magmatismrelated to the upwellingof hot asthenosphere
Plaziat et al., 1990; Camp and Roobol, 1991a]. Whereasthese beneathwesternArabia. In examiningthe geologicrecordof
investigations
reveal a lack of evidencefor plume generated westernArabia,we attemptto resolvethevariablecharacterof
asymmetry
eastof
uplift and erosionprecedingthe main period of rifting and continentalmagmatismand its conspicuous
magmatismin the late Oligoceneand early Miocene (from the Red Sea Basin (Figure 1). We attributethe current
physiographyof Arabia and adjacentAfrica to a two-stage
modelinvolvingboth the passive-and active-mantlehypotheses
Copyright1992 by the AmericanGeophysicalUnion.
of Sengorand Burke [1978]. Althoughwe supportthe contention of othersfor passive-mantle
riftingof the Red Sea Basin
Paper number 92JB00943.
0148-0227/92/92JB-00943505.00
betweenabout 30 and 20 Ma [e.g.,McGuire and Bohannon,
15,255
15,256
CAMPAND ROOBOL:UPWELLINGASTHENOSPHERE,
WESTERNARABIA
IAL-RUTBAH
RCH
THE MMN LINE
40 ø
50 ø
HADAN
100
0
200 km
OLDER HARRATS
DIKES
[-20-25
Mal
Fig. 1. Distributionof Cenozoicvolcanicrocksin westernArabiaand
adjacentAfrica. Only the barratlavafieldsof theoldergroup,related
to Red Sea extensionin the Oligo-Miocene,are labeled. Harrats of
the youngergroup,relatedto active-mantle
upwellingin concertwith
plate collision,are labeledin Figure4.
In SaudiArabia, the older rocksare tholeiiticto transition1989],our intentis to demonstrate
that the subsequent
period
of contemporaneous
uplift and magmatism
wascontrolled
by al, and mostwere emplacedalongnorthwesttrendingstrucactive-mantle
upwellingconcentrated
alongan axisthat lies turesparallelto the Red SeaBasin;in contrast,
the younger
eccentric to the Red Sea Basin but coincident with the central
rocksare transitionalto stronglyalkalicand were emplaced
crest of the Afro-Arabian Dome.
alongnortherly
trending
structures
whichdiverge
fromtheRed
Seatrendby about25ø(Figures1 and3). (Thereareonlytwo
TWO STAGES OF CONTINENTAL MAGMATISM
exceptions
to the olderandyoungertrendsin SaudiArabia.
Harrat
Hadan
wasextrudedduringtheolderperiodof magmaThe continentalmagmaticrocksin SaudiArabia can be
dividedinto older and youngergroupswhichdiffer in their tism abovea northtrendingfissurewhichappearsto overlie
overallcomposition
andstructuralsetting(Figure1) [Almond, north trending structuresin the Precambrianbasement,
was the only harratextruded
1986a, b; Coleman and McGuire, 1988; Camp and Roobol, whereasHarrat Nawasif/Boqum
1991a]. As described
below,thesegroupsappearto reflect from a northeasttrendingvent system.) Althoughthe two
they
separatemagmaticphases,contemporaneous
with the two groupshave fundamentallydifferentchemicalsignatures,
tectonicphasesdescribedabove.
Radiometricagedeterminations
from 182localities
in Saudi
Arabiarevealan agerangeof 30-20Ma for the majorityof the
oldergroup(with a peakbetween24 and 21 Ma) andan age
rangefrom 12 Ma to Recentfor theyoungergroup(Figure2).
Sporadicagesbetween20 and12 Ma appearto recorda period
of volumetrically
insignificant
volcanism.The older magmatic
seriesin SaudiArabia is generallycontemporaneous
with the
voluminousTrap Series basaltsin Yemen (41 analysesin
Figure2) andEthiopia[Civettaet al., 1978;Zanettinetal., 1980;
Mohr, 1983; Chiesaet al., 1988;Mohr and Zanettin,1988].
do not form distinctassemblages,
asshownby theiroverlapping
compositions
on the total alkali versussilicadiagram(Figure
3).
The older period of Red Sea extensionis recordedin
numerousnorthwesttrendingnormal faults, dikes, intrusive
complexes,
and basalticvent systemsalongthe entireeastern
marginof the Red SeaBasin(Figure1) [Blank,1977;Stenitzet
al., 1978; Colernanet al., 1983; Pallister,1987; Colemanand
McGuire, 1988;BaMridgeet al., 1991;Sebaiet al., 1991;du Bray
et al., 1991]. Severalrift flanklavafields(harrats)andharrat
remnantsrelatedto thisperiodof extension
are stillpreserved
CAMP AND ROOBOL: UPWELLINGASTHENOSPHERE,WESTERNARABIA
15,257
H UTAY MAH
l
/
•
I
ISHARA
I
HADAN
I NAWASIF
-BOQUM
TUFFIL
BIRK
I
SITA,
BALD,
AND
SHUMAYSI
VOLCANICS
SARAT
SAUDI
INTRUSIONS
I
YEMEN
GRANITES
•
/
I
l
KISHB
KHAYBAR
I RAHAT
l
I
I
a.
I RAHAH
I UWAYRID
LUNAYYIR
//
I KURA
}
I
I
5
10
15
•
AGE
IN
20
25
30
YEMEN
VOLCANICS
35
M.Y.
b.
YOUNGER
0
5
MAGMATIC
PHASE
10
OLDER
15
20
AGE
IN
MAGMATIC
25
PHASE
30
35
M.Y.
Fig. 2. Compilationof radiometricagesfor lava fields,intrusions,and volcanicinterbedseastof the Red Sea in Saudi
Arabia (182 determinations)
and Yemen(41 determinations).For K-Ar agedeterminations
greaterthan5 Ma, only those
with radiogenic
argongreaterthan10% are compiledasa testof reliability.(a) Rangeof determinations
reportedfor individuallavafields,volcanicformations,
andintrusive
groups.(b) Composite
histogram
of reportedageseastof theRedSea.
Data compiledfrom Civettaet at. [1978],Arnoet al. [1980a,b], Colemanet al. [1983],Capaldiet al. [1987],Pallister[1984,
1987],Chiesaet al. [1988],CampandRooOol[1989],du Brayet al. [1991],SeOaiet al. [1991],and Campet al. [1991, 1992].
The compilation
doesnot includeagedatafrom th• northernpartof the Arabiancontinental
basaltprovince,beyondthe
northernlatitudeof the Red Seaat 28øN;nor doesit includethoseagedeterminations
older than 35 Ma notedby Pallister
[1987] andBrownet al. [1989].
immediatelyeastof the Red Seaescarpment.ThesevolumetricallysmallharratsincludeAs Sarat(31-22 Ma [duBrayet al.,
1991]),Hadan(about27.5 Ma [Sebaiet al., 1987,1991]),and
probablythe upliftedHarratsIsharaand Harairahwhichare
not well limitedby K-Ar agedeterminations
(Figures1 and2).
These harrats are the most "alkalic"of the older magmatic
series,althoughnonepossesses
the stronglyalkalinecharacter
of the majorityof the youngerharrats.
The youngermagmaticseriesis composed
entirelyof harrat
lava fields(about85% of the total harratvolume)distributed
1$,258
CAMP AND ROOBOL:UPWELLINGASTHENOSPHERE,
WESTERNARABIA
OLDER
MAGMATIC
YOUNGER
PHASE
I
ALKALINE
ß
•
ß
_
/
•e
.•
/e
ß
ee
e
...
PHASE
I
I
ee
/
• •ee
eem• e ß
ße e• •ek__• ß
. .
ß
...
ß
ß
..
MAGMATIC
ALKALINE
ß
ß
10
.
le
ß
.
ß
.
SUBALKALINE
o
40
...•I ß
50
I
I
60
70
wt.% s•o 2
I
I
I
50
60
70
80
Wt.% SiO2
Fig. 3. Total alkali versussilicadiagramfor 742 analyses
of the older and youngermagmaticphasesfrom Yemen, Saudi
Arabia, and Jordan. Field boundaryis fromIrvineand Baragar[1971]. The oldergroupof 198 analysesincludes14 from
the YemenTraps[Chiesaet al., 1988],122fromdikes,gabbroic-to-granitic
plutons,andgranophyres
(Capaldiet al. [1987],
30 analyses;
Pallister[1987], 11 analyses);
Colemanand McGuire[1988],81 analyses),
41 from Hatrat As Sarat[du Brayet
al., 1991],six from the Sita volcanics[Pallister,1987],six from Hatrat Hadan [Arnoet al., 1980a],and nine from Harrat
Ishara (V.E. Camp and M.J. Roobol,unpublished
data, 1992). The youngergroupof 544 analysesincludes27 from the
RecentYemenvolcanics[Civettaet al., 1980], 19 from Hatrat AI Birk [Arnoet al., 1980b],sixfrom Harrat Nawasif/Boqum
[Arno et al., 1980a], sevenfrom Harrat Kurama (V.E. Camp and M.J. Roobol, unpublisheddata, 1992), 12 from Harrat
Shamah[Barberiet al., 1980],22 from Harrat Kura [Campet al., 1991],108 from Harrat Kishb[Campet al., 1992],and 343
from Harrats Rahat, Khaybar,and Ithnayn (the MMN volcanicline) [Campand Roobol,1989; Camp et al., 1991]. The
youngergroupof harratscan be furtherdividedinto mildlyalkalineandstronglyalkalinelava fieldsas demonstrated
in
Figure 5a.
far beyondthe confinesof the Red Sea,up to 600 km to the
east and as far north as Turkey (Figure 1). Whereasthe
restricted
distribution
and northwest
tism have been contemporaneous
and symmetricacrossa
linear, north-southaxisof mantleupwellingover the past12-14
trends of the older
magmaticrocksare consistentwith a genesisrelated to Red
Searifting,the very wide distributionand northerlytrendsof The West Arabian Swell
the post-12 Ma harrats are incompatiblewith magmatism
An asymmetricdepositionalfaciespattern in the northern
related to Red Sea extension. Gregoryet al. [1982],Almond Red Sealed Bayeret al. [1988] to concludethat the northeast[1986a,b], and Campand Roobol[1991a]notedthat the N-S ern Red Sea marginwasupliftedabovesealevel after 14 Ma,
reorientationof the harrat linear vent systemsmusthavebeen andthat thiswasshortlyfollowedby renewedsubsidence
of the
in responseto a major changein the overall regionalstress Egyptianoffshorearea at about 12 Ma. This age of uplift
regime. Such a changewas recognizedby Sengorand Kidd appearsto be consistent
with the post-14Ma agerecognized
by
[1976],SengorandBurke[1978],andAshwalandBurke[1989], Bohannonet al. [1989]alongthe southeastern
Red Seamargin.
who suggestedthat N-S fissuresbegan to developon the The continuationof Arabianuplift into the Plio-Pleistocene
has
Arabianplate after about12-15Ma in response
to the collision resultedin a topographicasymmetryacrossthe presentRed
of the Arabian and Eurasianplatesalongthe Bitlissuturezone Sea Basin,which is generallycoincidentwith the post-12Ma
of southernTurkey [Hempton,1987]. The orientationof most magmaticasymmetrydiscussed
above. We hereinrefer to this
of the post-12 Ma vent systemsis consistent
with an E-W broad region of uplift and magmatismas the West Arabian
componentof maximumhorizontalextension,
perpendicular
to Swell.
N-S collision.
Although there is a steady south to north decreasein
elevationon both sidesof the Red Sea, mountainpeaksalong
SYMMETRY OF UPLIFT AND MAGMATISM
the Red Seaescarpment
are consistently
higheron the Arabian
Dixon et al. [1989, 1991] attributed the asymmetryof sidethan on the African side [Voggenreiter
et al., 1988;Bohanmagmatismacrossthe Red Sea Basin (Figure 1) to a north nonetal., 1989;Martinezand Cochran,1989;Dixonetal., 1989].
trendingzone of mantleupwellingwhichwaspresentbeneath More significant,however,are the inlandregionswhichare less
westernArabia about 30 million yearsago,but is nowcentered affected by rift shoulderuplift. West of the Red Sea, the
beneath the Red Sea spreadingcenter due to the lateral African inland region is relativelyflat, with an averageelevamigrationof the overlyinglithosphere.Whereaswe agreewith tion of about 500 m. In contrast,the Arabian inland region
the existenceof sucha zone,we think that it is a muchyounger (the WestArabianSwell)is a broad,undulatingtablelandwith
featurewhichwas not involvedin the Oligo-Mioceneperiodof averageelevationsof 1000-1500m [Bohannonet al., 1989],
Red Sea rifting, as they propose[Camp amt Roobol,1991a]. similarto the crestsof the other magmatically
active,nonrifted
Here we examineevidencethat continentaluplift and magma- swellsin North Africa [Crough,1983]. The averagewidth of
CAMPAND ROOBOL:UPWELLINGASTHENOSPHERE,
WESTERNARABIA
the West Arabian Swell is about 1200 km, which is identical to
15,259
that of the HawaiianSwell[yonh•rzon et ai., 1982],andin the
middle range of oceanicswell widths listed by White and
McKenzie[1989].
The West Arabian Swell formsthe northernpart of the
muchlargerAfro-ArabianDome, the dimensions
of whichare
mostrecentlydefinedbyAlmond[1986a](modifiedin Figure
4). Almond [1986a] notes that this broad regionof late
Tertiaryuplift liesfar beyondthe Red Seamarginsandhasan
ArabianSwell. This axiscontainstwo significant
features:the
Makkah-Madinan-•atua(MMN) volcanicline in the south
[Camp et al., 1989] and the Ha'il-Rutbah Arch in the north
[Brown, 1972; Greenwood,1973] (Figure 4). The MMN
volcanicline is definedby north trending,slightlyen echelon
linear vent systemswhich mark the central axesof Harrats
Rahat, Khaybar, and Ithnayn; it extendsover a north-south
distance
of morethan600 km, andit hasbeenthe majorsite
of volcanism
in SaudiArabiaoverthepast10 m.y. The Ha'il-
overall north-south axis which deviates from the Red Sea trend
Rutbah Arch, alignedon strikenorth of the MMN volcanicline
by about25ø, parallelto mostof the post-12Ma harratlinear
vent systems. The southerncrestof this domal anticlinorium
lies along the SW trendingEthiopianRift System,and the
northern crest lies along the central N-S axis of the West
ArabianSwell(Figure4).
In contrastto the asymmetryof late Tertiary uplift and
magmatism
acrossthe Red Sea Basin,there appearsto be a
symmetryof theseprocesses
acrossthe centralaxisof the West
(Figure4), hasapparentlybeenthe siteof severalperiodsof
uplift that precededthe mostrecentperiodof upliftdescribed
above. Late Cretaceous
uplift appearsto havecontinuedinto
the the Early Tertiary, as indicatedby the lack of Eocene
depositionacrossits crest,and by the thickening
of Eocene
strataon both sidesaway from its crest[Powerset al., 1966;
Greenwot•, 1973;RMdleret al., 1986]. At leastsomeuplift
acrossthe arch appearsto have been contemporaneous
with
o,,•Karacalida
I-Rutbah
,,
•.-,.
Arch
rid
Kura
.•Hutaymah
thnayn
•
MMN
LINE
ahat
shb
I 500 krn ß
awasif/
Boqum
AFRO-ARABIAN
DOME
\\ SWELL
Silicic
rift-flank
Afar
volcanics
St ratoid
I l • "••\
RiftSeries
o
l/
/
//
!
\
Series
ETHIOPIAN
-'"•'1
DOME
Fig. 4. Area of the Afro-Arabiandome(hatchedline) and distributionof the post-12Ma volcanicrocksin Arabia and
adjacentAfrica. The Afro-Arabiandomeis modifiedafterAlmond[1986a]to includethe Ha'il-RutbahArch in the north,
whichwasan importantsiteof Tertiaryuplift[Brown,1972;Colemanetal., 1983].Almond[1986a]usesthe sealevelcontour
of the baseof the Mesozoic[Brown,1972] to partlydelineatethe dimensions
of the domein Arabia. The approximate
distributionof the post-12Ma volcanicrocksis fromEyal et al. [1981],Colemanet al. [1983],Mohr[1983],Almond[1986a],
Campand Roobo/[1989],and Campet al. [1991, 1992].
15,260
CAMP AND ROOBOL:UPWELLINGASTHENOSPHERE,
WESTERNARABIA
Miocene basalt extrusionof Harrat Shamah (Figure 4). Zr/Nb values for the more alkalic lavas from Harrat Kura to
Althoughbasaltflowsin Wadi Sirhan,immediatelywestof the the west and Harrat Kishb to the east.
Experimentalstudieson the derivationof basalticmagmas
main body of Harrat Shamah,are interbeddedwith Miocene
shallowmarinesediments,
farthereasttheyrestunconformably at high pressures[O'Hara and Yoder, 1967; Kushiro,1969;
that
upon eroded lower Tertiary marine sedimentsat 700-800 m Walkeret al., 1979;Stolper,1980;Sacketal., 1987]suggest
abovesealevel [Colemanet al., 1983].
thereis a positivecorrelation
betweendepthandalkalicontent.
Thus the flankingmore alkalic harratsmay also have been
StronglyAlkaline Magmatismon OpposingSides
derivedfrom a slightlydeeperasthenospheric
sourcethan the
of the MiMly Alkaline MMN VolcanicLine
lessalkalicMMN harrats. A somewhat
deepersourcefor the
with the rare earth
Of the post-12Ma lava fieldson the West ArabianSwell, flankingharratsappearsto be consistent
light rare earthelement(LREE)/heavyrare
mappablepetrochemicalstratigraphies
have been established data. Increasing
primarybasaltic
for HarratsRahat (circa10 Ma to Recent),Kura (circa10 to earthelement(HREE) ratiosfor cogenetic,
increasein
5 Ma), Khaybar(circa5 Ma to Recent),Ithnayn(circa3 Ma to magmasare generallyattributedto the progressive
Recent),andKishb(circa2 Ma to Recent)[CampandRoobol, the garnet/cpxratio of the sourcewith increasingdepthof
1989, 1991b;Camp et al., 1991, 1992;Roobolatut Camp, 1991a, partialmelting[Kayand Gast,1973;Zielinski,1975]. Greater
1991b]. Althoughthe remainingpost-12Ma harratshavebeen depthsof partial meltingfor the basaniteand AOB lavasthat
studied in less detail, reconnaissance
samplingprovidesan dominateHarratsKura and Kishbmay be indicatedby their
adequatedata base to assesstheir overall compositionand the LREE/HREE ratioswhichare greaterthanthoseof the OTB
trendof fractionation[,,lmoet al., 1980a,b; Barberiet al., 1980; lavasthatdominatetheMMN harrats[CampandRoobol,1989;
Campet al., 1991,1992]. Thesedifferentratiosareunlikelyto
Civettaet al., 1980; Coleman et al., 1983; Coleman and McGuire,
simplyreflectvariationsin crystalfractionation
because(1)
1988].
REE patternscrossin the
The publishedanalyticaldata reveal two distinctmagma manyof the chondrite-normalized
lavatypeshavesimilarMg
series for the post-12 Ma volcanicrocks. One is dearly MREE range,and(2) the analyzed
undersaturatedwith abundant normative nepheline,whereas
the other is transitionalwith both nepheline-normative
and
hypersthene-normative
rock types. The undersaturated
series
is primarily composedof alkali olivine basalt (AOB) and
basanitewhichfractionateto phonolite.The transitional
series
is primarilycomposed
of olivinetransitional
basalt(OTB) and
subordinateAOB whichfractionateto trachyteandcomendite.
The undersaturatedseries containsa significantvolume of
ultramaficxenoliths(generallyentrainedin basanite),whereas
numbers
(Mg # = Mg[Mg+Fe2+]).
Colemanetal. [1983]showthatbasaltanalyses
fromHarrat
Khaybarand Harrat A1 Birk (Figure4) lie alongparalleland
separabletrend linesreflectinglow-pressure
(< 40 km) and
high-pressure
(> 40 km) fractionation
paths,respectively,
on
the diopside-olivine-silica
plot of Walkeret al. [1979]. This
suggests
that the inferredshallowerdepthsof meltingbeneath
the MMN line may correspondwith shallowerdepthsof
fractionation, as well.
We concludethat the MMN harratswere derivedby
greaterdegreesof partialmelting,at shallowerdepths.From
MMN
harrats and for Harrat Kura to the west and Harrat
evidencebelowwe attributethe greaterdegreesof partial
Kishbto the eastprovidescross-sectional
petrogenetic
dataon meltingto a hotter mantlesourcebeneaththe MMN line and
depthof meltingto a coinciding
crestof mantle
the asthenospheric
sourcebeneath the West Arabian Swell. the shallosver
The overall alkalic (undersaturated)
chemicalcomposition
of upwelling,the generalizedsymmetryof which is shownscheHarrats Kura and Kishb are typical of most of the other maticallyin Figure7.
Arabian harrats. In contrast,Harrats Rahat, Khaybar,and
MEET GENERATION
the transitional
series has a distinct lack of xenoliths.
The well-establishedpetrochemicalstratigraphyfor the
Ithnayn are the only post-12Ma harratsin SaudiArabia and
Jordan of the transitionalseries. It is significantthat these
three transitionalharrats,primarilycomposed
of OTB (>60%
of total volume)with subordinateAOB and hawaiite(Figure
5), are restrictedto the centralportion of the West Arabian
Swell alongthe MMN volcanicline.
The overall differencein the chemicalcomposition
of the
MMN harrats is attributed to greater degreesof partial
melting.The degreeof partialmeltingfroma peridotitcsource
should be reflected in the Zr/Nb ratios of the derivative lavas.
Since the distributioncoefficientfor Zr in clinopyroxene
is
about 10 times larger than that for Nb [Dunn and McCallurn,
1982;Watsonand Ryerson,1986],higherZr/Nb ratiosindicate
greater degreesof partial melting from mantle peridotitc
[,'litherret al., 1988]. From 68 representative
harratsamples
from Saudi Arabia, ,'litherret al. [1990] demonstratethat the
largestZr/Nb ratiosare from the MMN line (Harrat Rahat)
and that
the
smallest
ratios
are from
harrats
McKenzieand Bickle [1988] definethe heat contentof the
mantle by its potential temperatureTp, the temperatureit
wouldhaveif broughtto the surfaceadiabatically
withoutmelting. The normalTp of asthenosphere
to producemid-oceanic
ridge basalt (MORB), as calculatedby McKenzieand Bickle
[1988],is 1280øCwith a rangeof _30øC. Usingthistemperature asa referencepoint,we considerthreepossiblemodelsfor
melt generationbeneaththe West ArabianSwell:(1) decompressionalpartial melting of normal temperatureasthenosphere,with Te = 1280øC[e.g., Foucheret al., 1982], (2)
decompressional
partialmeltingof metasomatized
subcontinental lithosphere,with Te < 1280øC[e.g.,Bailey,1987;Bonatti,
1990; Pearceet al., 1990], and (3) partial meltingof hot
asthenosphere,
with Te > 1280øC[e.g.,Whiteand McKenzie,
1989]. Melt generationin all three modelsmusttake place
under adiabatic conditions.
located the
Partial Melting of Asthenosphere
farthestto the west (Uwayrid) and east (Hutaymah)of the Decompressional
with
Te
=
1280øC
MMN line. Althoughthere is little evidencefor a systematic
variation in this reconnaissancedata, it is clear from our data
McKenzie and Bickle [1988] show that the amount of
(Figure6) that the MMN lavashavesignificantly
higherZr/Nb magmaproducedduringcrustalextensiondependslargelyon
and the stretchingfactor of the
ratiosreflectingoverallgreaterdegreesof partialmeltingthan the Tp of the asthenosphere
CAMPAND ROOBOL:UPWELLINGASTHENOSPHERE,
WESTERNARABIA
CREST OF UPWELLEDASTHENOSPHERE'
HARRATSRAHAT,KHAYBAR
AND ITHNAYN }THEMMN LINE}
15
15,261
FLANKSOF UPWELLEDASTHENOSPHERE:
HARRATS K!SHB AND KURA
15
F
xx ßß ß /
ee ß
TRACH
•
• I/BAS_.;.•
.....
•BEN•
5
RHY
5
0
40
50
•
SiO260
I
70
X
0
40
50
SiO260
70
75-
b.
62.79%
52.49
%
38.21%
22.59%
10.67%
3.95%
OTB
AOB
HAW+ BAS
DIFF
6.52
%
2.77%
OTB
AOB
HAW+ BAS
DIFF
Fig.5. Compositional
comparison
of theMMNharrats
lyingabove
thecrestof theupwelling
asthenosphere
andHarrats
KuraandKishblyingabove
theflanks
of upwelling
asthenosphere.
(a) Totalalkaliversus
silica(TAS)diagram
ofLe Bas
etal.[1986].(b)Volume
percent
diagram
forrepresentative
rocktypes,
calculated
fromthepercent
of eachrocktypein
eachunitandtheestimated
volume
of eachunit. Abbreviations
areOTB,olivinetransitional
basalt;
AOB,alkaliolivine
basalt;
HAW,hawaiite;
MUG,mugearite;
BEN,benmoreite,
TRACH,trachyte;
RHY,rhyolite
(comendite);
BAS,basanite;
PHT,phonotephrite;
TPH,tephriphonolite;
PH,phonolite.
Dashed
lineisthefieldboundary
ofb'vine
andBaragar
[1971]
separatingalkalinefrom subalkalinebasalt.
lithosphere
([5,ratioof theinitialto finallithosphere
thick- a basin-and-range
topography
beneathHarratRahathaving
ness).
IV/rite
andMcKenzie
[1989]
demonstrate
thatdecompresmaximumgrabendepthsof about 300 m below the harrat
sionalpartialmeltingof asthenosphere
with Tp = 1280øCsurface[BlankandSadek,1983].Fromthesedata,wecalculate
requiresa stretching
factorbetween2 and5; the formervalue a localized
maximum
stretching
factorof only1.018across
the
istypical
of intracontinental
sedimentary
basins
whichhavenot MMN line.Extension
of evenverythincontinental
lithosphere
subsequently
developed
intooceanbasins,andthe lattervalue cannotgeneratemeltingof normalTpasthenosphere
unlessthe
is roughlythe pointat whichstretched
continental
crustbreaks stretching
factorisgreater
than2 [McKenzie
andBickle,1988].
fully to oceaniccrust.
A stretching
factorof thismagnitude
abovenormalTeasthenoAlthoughstretching
of this magnitude
appears
to be sphereshouldresultin subsidence
of about1.75kin, andeven
consistent
withdecompressional
partialmelting
during
the30- smalldegreesof extensionwouldresultin at leastsomesubsi20 Ma periodof magmatism
andRedSeaextension
[e.g., denceaslongas normalTpasthenosphere
is present
[White
Cochran,
1983;Coleman
andMcGuire,
1988],it is clearly andMcKenzie,1989].Thecombination
of veryminorstretchinconsistent
withthepost-12
Maperiod
of intraplate
volcanisming and significantuplift in westernArabia sincethe midbasedonthelackof evidence
forsignificant
extension.
Buried Miocene
appears
to ruleouta simple
model
of decompressionfissures
beneath
theN-Slinearventsystems
provide
evidenceal partialmeltingof normalTe asthenosphere
by crustal
for negligible
stretching
across
thewidthof theWestArabian extension.
Swellsincethemid-Miocene.
On a morelocalscale,
however,
Uplift and partialmeltingof normalTe asthenosphere
extension
and subsidence
alongthe MMN line are demon- couldconceivably
occurif therewassomeothermechanism,
strated
byhigher
mountain
peaks
to thewestandeastandby besides
extension,
for the adiabatic
riseof asthenosphere
to
1$,262
CAMP AND ROOBOL: UPWELLING ASTHENOSPHERE,WESTERNARABIA
metasomatizedperidotite, however,would be required to
generatethe volume of basalt seen at the surfaceand the
supposed
volumeof basalttrappedat the crust/mantle
boundary. Partial meltingof sucha sourcewould alsorequire an
improbablelinear zone of more fusiblematerialbeneaththe
MMN volcanicline to accountfor the observedsymmetryand
linearityof partial melting. Perhapsthe bestargumentagainst
such a source comes from the two-pyroxeneequilibration
temperaturesfor the mantle xenoliths,between 900ø and
1100øC[Thornberand Pallister,1985;McGuire, 1988a;HenjesKunstet al., 1990;Nasir, 1992],whichare too cool for significant fusionof enrichedperidotite. The evidencesuggests
that
partialmeltingof metasomatized
lithosphere
with Tv < 1280øC
is alsoan unlikelymodelfor melt generationbeneathwestern
200
/
Kura
Phonolite
/
/
Nb
Arabia.
100
Partial Melting of Hot UpwellingAsthenosphere
with Tv > 1280 øC
ß
o
o
I
o
The progressively
higherelevationsof the West Arabian
Swellto the south,towardthe Afar triplejunction,suggest
that
uplift in Arabia may be largelya thermalresponserelatedto
mantleupwellingcenteredin the Afar region. The adiabatic
rise of asthenosphere
with Tp = 1280øCwould require a
greater amountof upwellingto producethe observeduplift
than would the rise of asthenosphere
with T• > 1280øC. As
KISHB LAVAS
MMN
LAVAS
I
500
lOOO
o
Z r ippml
D
Fig.6. Variationin Zr/Nb for theMMN harrats(Rahat,Khaybar,
and
Ithnayn)and the flankingHarratsKura and Kishb. The onlyMMN
::: .;:•::
DESERT
.::.'•
,.:,-)::
.:':.•
: ;.. - :::.:.
harrat sampleswith low Zr/Nb ratios, similar to thosefrom Harrats
Kura and Kishb,are a few high-MgOhawaiiteslocatedat the northern
end of the MMN volcanicline on Harrat Ithnayn(shadedregion).
shallowlevelsnear the crust-mantleboundary. Given a dry
peridotite solidus,McKenzie and Bickle [1988] show that
significantmelting of asthenosphere
with a Tp of 1280øC
probablycannotoccurbelow about50 km. Thiswouldrequire
an unusuallylarge magnitudeof adiabaticupwelling,which
clearly cannot be attributedto crustalextension. Without a
reliablemechanismfor considerable
upwelling,partial melting
of asthenosphere
with Tp = 1280øCappearsto be an unlikely
modelfor magmagenerationbeneathwesternArabia.
Melting
Partial Melting of Metasornatized
Lithosphere
with T• < 1280øC
In contrast to the large degreesof crustal stretching
requiredto melt normal T• asthenosphere,
only smalldegrees
of stretchingmay be requiredto melt metasomatized
mantle
containinglow-temperature
mineralphases.Pearceet al. [1990]
proposesucha modelfor partial meltingbeneaththe Karacalidag lava field on the northernArabian plate.
Large domainsof metasomatized
mantlemustbe restricted
to the subcontinentallithosphericmantle, well above the
vigorouslyconvectingasthenosphere
[Fitton,1987]. Pressure
estimates
on mantle
xenoliths
show that
the subcontinental
lithospherebeneathwesternArabia occursbetween60 and 40
km; many of thesexenolithsalso providepetrologicevidence
of metasomatism
[McGuire,1988a,b; Thornber,1990;HenjesKunst et al., 1990; Nasir, 1992]. A very large volume of
at
generally lower
temperatures and
greater depths
ß -ISOLATED VOLCANIC
\
CENTERS AND BASALTIC)
OUTCROPS •
VENT
I
DENSITY
50-15vents
/ 100km2
'•..........15-5 vents/100 km2
•
< 5 vents/100 km2
•
Region
ofextrusion
dominated.
by OTB, reflecting greatest
'•
degrees
ofpartialmelting
at •
SAHL
RAKBAH
BASIN
theshallowest
depths
Fig. 7. Map showingthe distributionanddensityof harratventson the
WestArabianSwell,and their proposedspacialrelationship
to mantle
upwelling. The generalizedsymmetryand morphology
of mantle
upwellingare basedon variationsin the relativedepthanddegreeof
partialmeltingacrossthe MMN volcanicline (seetext).
CAMPAND ROOBOL:UPWELLINGASTHENOSPHERE,
WESTERNARABIA
describedabove,it is difficultto envisagea reliablemechanism
'
signitiCant
rise of nu•ma•temperatureastn•nospx•et•
into
the lithosphere.Alternatively,we examinehere evidencefor
heatingand thinningof the lithosphericmantleby the rise of
asthenosphere
havingelevatedTv.
Heatingand thinningof lithospheric
mantle. Geophysical
data providesomeevidencefor elevatedmantle temperatures
beneathwesternArabia. Voggenreiter
et al. [1988] note that
them is an asymmetryin the Bougergravityfield acrossthe
Red Sea Basin, with the Arabian side havingvaluesof-110
mGal and the Egyptian side having values of-80 mGal
[IzzeMine,1982, 1987]. They interpretthesedata as evidence
that the lithosphereon the Arabian side is on averageless
dense and hotter. They concludethat the somewhatasymmetric heat flow profilesof Martinezand Cochran[1989] for
the northern Red Sea may also imply a hotter lithosphere
beneath Arabia, and this appearsto be consistentwith the
evidenceof Bayer et al. [1988] for asymmetricuplift of the
for
Arabian
side of the northern
Red Sea after about 14 Ma.
15,263
would producea partial melt fractionof about3.7%, which is
cmcmauons
c•oseto Rayleighmw
.............
of partial meltingfor the
basaniteaveragecompositionfrom Harrat Kishb. In contrast,
the OTB lavas of the MMN volcanic line were derived by
partial meltingof at least 10%, as determinedfrom Rayleigh
law calculations[Camp arm Roobol, 1989; Camp et al., 1991]
and experimentalinvestigations
on similarrock types[Mysen
and Kushiro,1977;Jaquesand Green,1980]. If we assumea
crestof mantleupwellingbeneaththe MMN line at a depthof
about 60 km, then a temperatureof about 1475øCis required
to produce a melt fraction of 10% [McKenzie,1984]. The
derivationof larger melt fractionswould requireeven greater
mantle temperatures.
The asthenospheretemperaturescalculatedabove equate
to potential temperaturesof about 1354øCbeneath Harrat
Kishb and 1436øCbeneath the MMN volcanicline. Although
these values are meant to be grossapproximations,they lie
significantlyabovethe 1280øCpotentialtemperatureof normal
convectingasthenosphere,
and within the range of elevated
potentialtemperaturesassociated
with mantlehot spots[White
and McKenzie,1989].
Not all geophysicaldata supportthe idea a hotter lithosphere. Surfaceheat flow over westernArabia is coolerthan
The mid-Miocene
to Recent harrats above the West
the world average[Gettings,1981; Gettingset al., 1986] and
(>80,000 km2) but thin and of
projectsto a temperatureof about 400ø-500øCat 40-50 km ArabianSwellare extensive
depth[McGuireand Bohannon,1989]. Thisprojectedtempera- smalltotal volume(probably<10,000km3). Their small
ture range,however,is at variancewith the 900ø-1100øC
range volume is consistent with field evidence for minor crustal
calculatedfrom the mantle xenolith data, at the same depth stretchingsincethe mid-Miocene,even if one considerstheir
temperatures
[Thornberand Pallister,1985;McGuire, 1988a;Henjes-Kunst
et generationto have involvedhigh asthenosphere
al., 1990; Nasir, 1992]. This discrepancysuggeststhat the and a thin lithosphere. For example,giventhe constraintsof
continentallithosphere(the mechanthermalgradientof the uppermantleis not in equilibriumwith a 70-km-thickundisturbed
that at the surface,andMcGuireandBohannon[1989]interpret ical boundarylayerof Parsonsand McKenzie[1978]) abovean
with a potentialtemperatureof 1480øC,McKenthisdisequilibrium
relationshipasevidencethat asthenospheric asthenosphere
upwellingbeneathwesternArabia isyoungerthan20 Ma. This zie atut Bickle [1988] showthat a stretchingfactorof 1.1 will
appearsto be consistent
with apatitefissiontrackdatasuggest- generatea melt thicknessof only 1.5 km; given the same
ing a mid-Mioceneage of uplift [Omaret al., 1989;Bohannon constraints,a stretchingfactorbetween2 and 5 will generate
between 10 and 20 km of melt. The smaller melt thickness is
et al., 1989].
McGuire arm Bohannon [1989] note that the calculated probablystill greater than that generatedbeneaththe MMN
mantle-xenolithtemperaturesare high for their depthbeneath volcanic line. The larger melt thicknessis similar to that
shieldrocksand must thereforerepresenta regionalelevation producedat the highlyextendedintersectionof the Red Sea,
of the mantle temperature• A regionalperturbationof the Gulf of Aden, and Ethiopian rifts near the site of the Afar
et al., 1975].
thermal gradient should result from thinning of the lower mantleplume [Berckhemer
McKenzieand Bickle [1988] demonstrate
that melt generalithosphere,as its mantle portion is convertingto asthenosphere by thermal erosion. A thinned lithosphericmantle tion beneathcontinentsis an importantcontributorto crustal
seemsreasonablefrom thexenolithdata. Pressure-temperatureaccretion, consistentwith evidence that the base of the crust
estimateson xenolithsfrom Harrat Uwayrid led Henjes-Kunst actsas a densitytrap for upwellingbasalticmagmas[Coz,1980;
[1989] to suggestthat the asthenosphere-lithosphere
boundary Stolperand Walker, 1980]. Given the very small stretching
in that regionis at a shallowdepthof about60 lan. McGuire factor, it seemslikely that harrat extrusionis only a small
fractionof the total melt generatedbeneaththe West Arabian
and Bohannon[1989] estimatethe asthenosphere-lithosphere
with petrologicdata indicatingthat
boundarybeneathHarrat Kishb to be at a depthof about75 Swell. This is consistent
km. Both estimatesimply a significantamount of thermal few, if any, of the harrat basalticlavasare primarymeltsbut
erosionof the lower lithosphereby hot upwellingastheno- rather are fractionatedmagmasderived from primary melts
sphere.Altherret al. [1990] cometo a similarconclusion
based which accumulated in reservoirs near the base of the crust
onlow•43Nd/•44Nd
ratiosforArabianharratlavaswhichsuggest [Campand Roobol, 1989; Campet al., 1991]. The crystallithe incorporationof a lithosphericmantle componentinto zation of basalticmagmaat this depth is now reflectedin
igneous-textured
xenoliths entrained in the harrat lavas
asthenosphere-derived
magmas.
Variationsin thedepthand temperature
of upwelling
astheno- [McGuire,1988b].
sphere.The projectionof a bestfit geothermpassingthrough
the Harrat Kishb xenolith data, led McGuire and Bohannon
[1989]to estimatethe temperature
of the asthenosphere
at the
baseof the lithospherebeneathHarrat Kishbto be 1400øCat
depthof about75 km. AlthoughthisPoTestimateis not well
A MODEL FOR THE CONCENTRATION
OF HOT MANTLE
ALONG A LINEAR, MIGRATING AXIS
The contemporaneity
of uplift and magmatismsinceabout
12 Ma (Figure2) togetherwiththe aboveevidence
for elevated
constrained, it is consistentwith model calculationson the mantle temperaturesbeneathwesternArabia are consistent
of Sengorand Burke [1978],
degreeof partialmeltingfor the harratlavas.McKenzie[1984] with the active-mantlehypothesis
showsthat isentropicmeltingat thistemperatureandpressure which attributesthermal uplift and magmatismto the active
15,264
CAMP AND ROOBOL:UPWELLINGASTHENOSPHERE,
WESTERNARABIA
rise of anomalously
hot asthenosphere.
This hypothesis
is only the East African rift zone. We concurwith Bayeret al. [1989]
partly applicable,however,sincemelt generationin Arabiawas and suggestthat the linear morphologyof mantle upwelling
also clearly facilitatedby the developmentof north trending may be at leastpartlycontrolledby directedflow of asthenospherebeneatha preexistingflexurein the continentallithofractureson the upliftedlithospheredueto platecollision.
The West Arabian Swellhasa width andelevation(1200 sphere.
The top of the Precambrianbasementwest of the Afrokm and 1000 m) consistent
with the hotspotmodelof White
and McKenzie[1989]. They showthat swellsof similarsizeare Arabian Dome is essentiallyflat lying or dipsslightlyto the
thermally supportedby narrow plumeswhich are deflected west,but to the eastit dipsby an averageof about10øto form
laterally by the overlyinglithosphereto producemushroom- the basement of the Zagros orogen in the north and the
shapedheadsof anomalously
hot mantlesome1500km across. basementof the Somaliabasinin the south[Bayeret al., 1989].
In its simplestform, however,this modeldoesnot resolve(1) Emplacementof the Zagros and Oman nappesin the Late
with a
the distinctlinearityof volcanismalongthe MMN line or (2) Cretaceousappears to have been contemporaneous
the northwardmigrationof MMN volcanismwhichbeganon majorperiodof upliftalongtheHa'il-Rutbaharch[Greenwo•,
southernHarrat Rahat at about 10 Ma, on Harrat Khaybarat 1973; Hancock et al., 1984; Riddler et al., 1986], so that the
about 5 Ma, and on Harrat Ithnayn at about 3 Ma. Since northernhingeline and easternflank of Afro-ArabianDome
prior to mantleupwellingin the
major rift-related structuresare absent,the linearityof the wasat leastpartiallydeveloped
of the
MMN line cannot be simply attributed to decompressionalTertiary. To the south,the east dippinglithosphere
partial meltingabovea linear rift zone, nor can the linearity Somalibasinappearsto be the resultof Jurassic
downwarping
and volcanicmigrationalongthisvolcanicline be attributedto [Hutchinson
and Engles,1970;Beydoun,1982],so that it was
plate migrationabove a stationaryplume becausethe north- alsopresentbeforeTertiaryupwelling.The deflection
of hot,
ward movementof the Arabian plate is clearlyincompatible risingasthenosphere
againstthis regionallithospheric
flexure
with such a model.
wouldhavemodifiedthe morphology
of mantleupwelling,and
Bayeret al. [1989] suggested
that the asymmetric
distribu- we suggestthat the N-S crest of upwellingnow preserved
tion of volcanic rocks east of the Red Sea could be attributed
beneath the MMN volcanicline in the south,and probably
to the position of a regional hinge zone that occurson the beneaththe Ha'il-RutbahArch in the north,maybe the result
easternside of the Red Sea and continuesto the southalong of such modification.
7'
40 ø
I
25 ø
MMN
Other
Harrats
Harrats
MMN Vent System
0
I
300 km
I
i
Fig. 8. Possiblemodelfor the development
of en echelonvent segments
alongthe MMN line resultingfrom left-lateral
shearimposedby counterclockwise
rotation(torque)of theArabianplate. Modifiedafterthemodelof Ryan[1988]for rift
developmentat Kilauea.
CAMPAND ROOBOL:UPWELLINGASTHENOSPHERE,
WESTERNARABIA
38 ø
36 ø
40 ø
15,265
42 ø
Rotation
of Danakil
Block
.?.y from
original
N-S
orientation (Le Pichon and Franch-
22 ø"
eteau, 1978)
I
\x?\\
\
\\? \
I
Proposed
axisof mantle
I
upwelling
20 ø
Harrat
volcanic
Known
crest
rocks
\
ß.
A
of mantle
..
upwelling associated with
18ø
18ø
t???!iiiii?ii!iiiiii!iiiiiiiiiiiii?
ili?!?ii]
tholeiitic-to-transitional
volcanism in oceanic areas
and transitional-to-mildlyalkalic
volcanism
nental
areas
Unknown
16 ø
16øI
in conti-
basement
over-
I lain
byCenozoic
basalt
or
sediments
Pre-Cenozoic
continental
basement
14 ø
•:• Arabian
plate
+
+ETHIOPIAN+
+ PLATEAU + +
• ++
++
++
•African
plate
12 ø
AGE OF OCEANIC
CRUST
A --
<lMa
'
B-•:::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::::
1- 2 Ma
C
2-3 Ma
D -•======================================================::=========
3- 4 M a
10 ø
+
+
E -
'
4-5Ma
Fig.9. Relationship
of proposed
axisof mantleupwelling
to riftingandvolcanism
in theDanakilDepression
andto seafloor
spreadingin the southernRed Sea. Separationof the Arabianand Africanplateshasnot beendemonstrated
in the
southern
RedSeasouthof 15øN,sothattheplateboundary
in thenorthernAfar regionis unclear[e.g.,Barberiand l/aret,
1977;Le PichonandFrancheteau,
1978].In describing
therotatedDanakilblockaspartof theattenuated
Arabianplate,
Cochran[1983]considers
theDanakilDepression
to be a moreappropriate
plateboundary.Ageof knownoceaniccrust
derivedby axialseafloorspreading
is fromthepaleomagnetic
dataof Roeset[1975]andCochran[1981,1983].
We proposetwo possiblemodelsfor the hot asthenosphereexample,the interpretationof similaren echelonventsegments
source:(1) upwellingof a mantle plume centrallylocated from Kilauea[Ryan,1988]).
beneath the West Arabian Swell or (2) convectiveflow
The principal problem with the secondmodel for the
emanatingfrom the Ethiopianmantleplumeto producean asthenosphere
sourceis that it requiresa northwardasymmetry
elongatedandextendedlobeof hot asthenosphere
beneaththe of convectiveflow away from the Ethiopianplume over a
West Arabian Swell. The similaruplift agesin Africa and distanceof about2000 km to accountfor uplift andvolcanism
Arabia would seemto supportthe secondmodel,and convec- on the West Arabian Swell. This distance is about twice the
tive flow away from the Afar regionwould be a convenient radii of typicalplumeheadsbeneathoceanicswells[Whiteand
mechanismto explainthe northwardmigrationof volcanism McKenzie,1989]. The White and McKenzie model, however,
alongthe MMN line asit is funnelledbeneaththe preexisting is basedon oceanicexampleswhich form by the circular
lithospheric
flexuredescribed
above. An alternativeexplana- deflectionof mantleplumesbeneathhorizontal,homogeneous
tion for this migration,consistentwith both models,is to oceaniclithosphere.It is unlikelythat the suchcircularshapes
attributeit to thenorthward
progression
of fissuredevelopmentwould be maintained beneath heterogeneous,regionally
and Gibson,
resultingfrom torque imposedon the lithosphereby the deformedcontinentallithosphere[e.g.,Thompson
counterclockwise
rotationof the Arabianplatedueto seafloor 1991], and for this reason,we do not entirelyrule out the
spreadingin the Gulf of Aden. Suchrotationis evidentfrom lateral migrationof asthenosphere
over suchgreat distances.
the paleomagnetic
data of Tarling[1970] and from the left- On the other hand,it is more difficultto explainhow such
lateralsenseof shearalongthe MMN line as indicatedby its migrationcouldoccurwell beyondthe confinesof the Red Sea
right-stepping
en echelonvent segments
(Figure8) (seefor Basinwhichshouldhavebeena naturalchannelfor upwelling
15,266
CAMPAND ROOBOL:UPWELLINGASTHENOSPHERE,
WESTERNARABIA
asthenosphere.We are thereforeinclinedto favor a separate however,have been controversial[cf., Cochran,1983; Girdler
the West Arabian Swell. In context of
and Southren,1987;Le Pichonand Gaulier,1988]. Le Pichon
the East African Rift system,sucha sourcecouldbe considered and Gaulier [1988] arguethat oceanfloor accretionbeganin
the northernmostin a seriesof mantle plumes,analogousto the Red Sea at about 12-13 Ma, contemporaneous
with that in
thosebeneaththe Kenyanand EthiopianDomes.
the Gulf of Aden [Cochran,1981]. This age appearsto be
consistent
with the initiationof uplift and harratvolcanismin
A COMPARISON WITH ETHIOPIA
westernArabia, as describedabove,and may reflectthe arrival
Both the West Arabian Swell and the EthiopianDome of a thermal anomalybeneaththe entire Afro-ArabianDome.
asit appears
appear to be underlain by hot, upwellingmantle; these two If thisthermalanomalyhasa linearmorphology,
mantle source beneath
regionsdiffer, however,in some importantrespects. The
to have beneath the West Arabian Swell, then its intersection
>3000-m-high crestof the Ethiopian Dome has been the site
with
the Red
Sea Basin should be evident
in the seafloor-
of dramaticrifting and prolificeruptionsof silicicignimbrite spreadinghistory.
We can ignore the debate on the initiation of seafloor
[Mohrand Zanettin,1988]. In markedcontrast,the MMN line
by examiningonly the last4-5 m.y.of knownocean
occurs at lower elevations,in a region of minor normal spreading,
faulting, and sporadicsilicic volcanismrelated to crystal floor accretion. Althoughrecentintrusionof basalticmagma
fractionationprocesses
[Campand Roobol,1989;Campet al., is evident at severalsites along the entire length of the Red
1991]. It seemsclear that the thermal structureof the litho- Sea axial depression[e.g., Cochran, 1983; Bonatti, 1985;
sphere differs beneath these two regions. This does not Cochranet al., 1991], the continuousgenerationof oceanic
necessarilymean that Ethiopia is underlain by a hotter crustalongthe axisover the last 5 m.y. can only be positively
asthenosphere,however, since variations in lithospheric demonstratedin the southernpart of the Red Sea [Roeser,
thicknesses,
at constantasthenosphere
Tp, canalsoaccountfor 1975;Cochran,1983](Figure9). Le Pichonand Gaulier[1988]
the observeddifferences.We suggestthat Ethiopiais under- argue for continuousseafloor spreadingin the central and
lain by a thinner subcontinentalmantle and that the volumi- northern Red Sea, but this cannot be confirmed by the
noussiliciceruptionsnear the rift reflectcrustalanatexiswhere paleomagneticdata.
Both the natureof the crustandthe styleof extension
prior
the mantlelithosphere
hasbeentotally,or nearly,consumed
by
thermalerosion. The smallerdegreesof uplift associated
with to 5 Ma are irrelevant to the model we describe below; the
the West Arabian Swell,togetherwith the lackof evidencefor model rests, however,on the assumptionthat the magnetic
byRoeser[1975]andCochran[1983]
majorrifting and crustalmeltingalongthe MMN line,suggest anomalypatterndescribed
a thickerlithosphere,possiblyabovean equallyhot astheno- approximatesthe distributionof basalt generatedby axial
spherethat has not yet penetratedthe crust-mantle
boundary. seafloorspreadingin the southernRed Sea sinceabout5 Ma.
Counillot [1982] demonstratedthat this period of seafloor
spreadingis unusualin that it has involvedrift propagationin
REGIONAL
IMPLICATIONS
two opposingdirections(to the northsvest
and southeast).He
The hingeline recognized
by Bayeret al. [1989]is equiva- pointsout that the southeastdirectionof rift propagationis
lent to the Cenozoicaxisthe Afro-ArabianDome (Figure4).
inconsistent
with a simple asthenospheric
sourceemanating
In Arabia this axis coincides with the Ha'il-Rutbah arch and
from a largemantleplumecenteredat the Afar triplejunction.
the MMN volcanic line; in Africa it coincideswith the high
At first glance,there appearsto be no compellingreasonwhy
central region of the Ethiopian Dome and the southwest
axial seafloor spreadingshouldbe restrictedlargely to the
trendingaxisof the EthiopianRift System.Boththe Arabian southernportionof the Red Seasinceabout5 Ma, nor is there
and African regionsappearto have similaruplift agesrelated
an adequateexplanationfor the peculiarstyleof rift propagato the upsveiling
of hot asthenosphere
concentrated
alongthis tion in two opposingdirections.
centralaxis. Althoughit is not clearwhetheror not the West
The sourcefor a doublypropagatingrift systemhas to lie
Arabian Swell and the Ethiopian Dome are underlainby beneaththe axialplane,betweenthe two propagating
rift ends.
separate mantle plumes, a genetic link between these two When the crestof asthenospheric
upwellingbeneaththe MMN
regionsmaybe apparentalongthe connecting,
middleportion line is projectedacrossthe Red Sea into the Danakil Depresof the axis. Here, the N-S strike line is the focus of two
sion, its intersection with the Red Sea axis occurs at the
additionalregionsof well-knownmantleupwelling(Figure9):
initiationsite of axialseafloorspreading4-5 m.y. ago (Figure
(1) it is coincident
with the N-S axisof the magmatically
active 9). We suggestthat seafloorspreadingmay havebegunwith
Danakil Depressionand (2) it cuts acrossthe southernRed the invasionof thishot line into the previouslyattenuatedRed
Sea Basinat the site of axial seafloorspreading4-5 m.y. ago.
The Red Sea
Sea crust,which resultedin rift propagationin two opposing
directions,away from the mantle sourcebut parallel to the
preexistingbasin structure. In a departurefrom conventional
models,we proposethat the main asthenospheric
sourcefor
Red Sea volcanismmay lie eccentricto its axisand coincident
The Danakil Depressionis separatedfrom the MMN
volcanicline by the Red SeaBasin. After a 10-m.y.periodof
passiveextensionrelatedto the oldermagmatic
phase,the Red with a north-south thermal line beneath the crest of the AfroSea Basin was well establishedalong a northwesttrend by ArabianDome (Figure9).
about 20 Ma. Below,we examinethe proposedeffect of the
Unlike the middle section of the Red Sea, which is characyoungermagmaticphase,whena zoneof N-S mantleupwelling terizedby normalMORB (NMORB), the southernsectionis
wassuperimposed
discordantly
acrossthe Red Seatrendafter characterizedby transitionalMORB (TMORB) and AOB
about 14 Ma.
similarin composition
to oceanicislandbasalt[Altherret al.,
There is a consensus
of opinionthat the Red Sea magnetic 1988; Barrat et al., 1990]. Barrat et al. [1990] attributethis
lineamentsrecognizedby Roeser[1975] reflect oceaniccrust differenceto the mixingof depletedMORB mantlewith an
producedby axial seafloorspreadingover the past 4-5 m.y. enrichedhotspotcomponent
in the southernRed Sea region.
Opinionson the generationof oceaniccrustbeforethis time, A comparison
of incompatible
elementabundances
for near-
CAMPAND ROOBOL:UPWELLINGASTHENOSPHERE,
WESTERNARABIA
lOO
15,267
APPROXIMATE
POTENTIAL
TEMPERATURE
IøCI
1500
ø••
•,•.
ø [...........
1450 ø
LU
OTB
4949
0
North
10
I 500 km
Rb
Ba
K
Nb
Sr
P
Zr
Ti
I
Y
Fig. 10. A comparison
of trace-element
abundances
for near-primary
AOB sampleZI0 (Mg # = 65.6) from the southernRed Seaand two
of the mostprimitivebasaltsfrom the MMN line (Mg # = 66.1 and
67.6 for samplesOTB 4949 and AOB 4947, respectively).Traceelementabundances
are normalizedto the primordialmantleof Wood
et al. [1979]. Data for sampleZ10 are fromAltherret al. [1988].
primm3talkalic basaltsfrom the southernRed Sea and the
MMN line (Figure 10) appearsto indicatesimilar mantle
sourceregionsalongthisproposedzoneof N-S upwelling.
Melting of hot asthenosphere
at the site of rifting should
normally result in a thickenedcrust from magmaticunderplating due to the increasein melt production[White and
McKenzie,1989]. Seismicrefractionstudies,however,show
LATE TERTIARY SEAFLOOR
SPREADING
IN PREVIOUSLY
ATTENUATED OLIGO-MIOCENE
CRUST
Fig. 11. Modelshowing
theproposed
relationship
of theAfro-Arabian
dometo potentialasthenosphere
temperatures
at thebaseof thelithoBanks in the southernRed Sea is not greater (and perhaps sphere.The modelisbasedontherelativetemperature
gradientacross
muchless)thanabout10 km [Bohannon
andEittreim,1991],as the MMN line in the north and acrossthe Ethiopianplumein the
comparedto the 7-km thicknessof averageoceaniccrust south. The approximateabsoluteT•, of the mantleis taken from the
[McKenzieand Bickle, 1988]. The lack of a substantially overallhotspotmodelof Whiteand McKenzie[1989]and from the
temperature
estimates
discussed
in text (seesectionon melt generathickenedcrustin the southernRed Searegionmayindicatea tion). It is not knownwhetherthe sourceof upwelling
beneaththe
somewhatcoolerasthenosphere
(perhapsTp _<1380øC)than WestArabianSwellis froma separate
mantleplume,asdepicted,
or
flowawayfromthe Ethiopianmantleplumealonga
that presentbeneath the Ethiopian Dome in the south and fromconvective
that
the crustal
beneath the MMN
thickness beneath
the Dahlak
volcanic line in the north.
and Farsan
This would tend
preexistinglithosphericflexure.
to supportthe idea of a separatemantle plume beneaththe
West Arabian
Swell.
Discussion
The Red Sea and Gulf of Aden basinsare clearlythe two
mostprominentregionsof shallowasthenospheric
upwellingin
northeast Africa
and Arabia.
These two rift arms were well
developedby 20 Ma as the result of tectonicallyinduced
passiverifting [Cochran,1981, 1983] associated
with the older
periodof tholeiiticto transitionalmagmatism
(Figures2 and
3). We suggest
that a third,lessprominentregionof asthenosphericupwellingdevelopedlater, beneaththe centralcrestof
but then it turns to the southwestalongthe rifted crestof the
EthiopianDome (Figure4). At its intersection
acrosseachof
theseprovincesthere is petrologicaland structuralevidence
that this axis is coincidentwith a zone of upwellingasthenosphere. The symmetnyof mantleupwellingon opposingsides
of this thermal line is evident from (1) lavas derived by
decreasingdegreesof partial melting,at increasingdepths,west
andeastof the MMN volcanicline, (2) the top of the Precambrian basementwhichdecreases
in elevationon opposingsides
of both the Ha'il-Rutbah Arch and the generalcrestof the
the Afro-Arabian Dome, where the active rise of hot astheno- Afro-ArabianDome,and(3) thedoublypropagating
rift system
spherehasresultedin uplift associated
with the youngerperiod on opposingsidesof the N-S line at its intersectionwith the
of transitionalto stronglyalkalicmagmatism.
Red Sea axis. The morphologyand temperaturegradientof
The central axis of the Afro-Arabian Dome extends across
mantleupwellingalongthiscentralaxismaybe similarto that
severaldistinctprovinces. It maintainsa north-southorienta- portrayed in Figure 11, althoughthe N-S extensionof hot
into westernArabia may or may not involvea
tion from the Ha'il-RutbahArch throughthe MMN volcanic asthenosphere
line andacrossthe Red SeaBasininto the DanakilDepression, separatemantle plume.
15,268
CAMPAND ROOBOL:UPWELLINGASTHENOSPHERE,
WESTERNARABIA
REFERENCES
The youngerperiodof magmatismin northeastAfrica and
adjacentArabia has been dominatedby two separatecompo- Almond, D.C., The relation of Mesozoic-Cainozoicvolcanismto tectonnentsof mantleupwelling:(1) deepupwellingof elevatedTv
ics in the Afro-Arabian dome, J. Volcanol. Geotherrn.Res.,28, 225-
asthenospherebeneath the central crest of the Afro-Arabian
246, 1986a.
Dome, and (2) shallowupwellingof normalTv asthenosphere Almond, D.C., Geological evolution of the Afro-Arabian dome,
Tectonophysics,
131, 301-332, 1986b.
beneath the Gulf of Aden, and much less conspicuously, Altherr, R., F. Henjes-Kunst,H. Puchelt,and A. Baumann,Volcanic
beneaththe central and northernaxial planeof the Red Sea
activityin the Red Sea axial trough--Evidence
for a large mantle
Basin [Cochran,1983; Bonatti, 1985; Cochranet al., 1991].
diapir,Tectonophysics,
150, 121-133,1988.
versus
Asthenospheric
upwellingbeneaththe southernpart of the Red Altherr, R., F. Henjes-Kunst,and A. Baumann,Asthenosphere
lithosphere
as
possible
sources
for
basaltic
magmas
erupted
during
Sea axial plane appearsto be a mixture of thesetwo mantle
formationof the Red Sea:Constraintsfrom Sr, Pb and Nd isotopes,
components
[Barratet al., 1990].
Earth Planet. Sci. Lett., 96, 269-286, 1990.
The post-14 Ma transitionalto stronglyalkalic volcanic Arno, V., et al., Recentvolcanism
withintheArabianplate:Preliminary
data from Harrats Hadan and Nawasif-AlBuqum,in Geodynamic
rocksof the EthiopianRift, the Danakil Depression,and the
Evolutionof the Afro-ArabianRift System,Publ. 47, pp. 629-643,
West ArabianSwellwere generatedby a deep,enrichedsource
beneath the crest of the Afro-Arabian Dome; in contrast,the
partly contemporaneous
tholeiiticrocksin the Gulf of Aden
and Afar regionswere generatedby a shallow,depletedsource
related to seafloorspreading. Whereasthe staticEthiopian
mantleplume is rooted in the deep mantle,the independent,
convectingasthenosphere
beneath the Gulf of Aden has
migratedinto the Afar regiondueto the westward
propagation
of the Sheba ridge spreadingcenter [Courtillotet al., 1980;
Cochtz•n,1981;Courtillot,1982].
The presentconfigurationof the Afar triple junctionis
largely a consequence
of the post-14Ma superimposition
of
this two-componentasthenospheric
system (the younger
magmaticphase)uponthe previously
developed
Red Seaand
Gulf of Adenbasins(the oldermagmatic
phase).Of the three
Academia Nationale dei Lincei, Rome, 1980a.
Arno, V., et al., Recentbasicvolcanismalongthe Red Seacoast: the
AI Birk volcanicfield in SaudiArabia,in Geodynamic
evolution
of
the Afro-Arabian rift system,Publ. 47, pp. 644-654, Academia
Nationale dei Lincei, Rome, 1980b.
Ashwal,L. D., and K. Burke,African lithosphericstructure,volcanism,
and topography,
Earth Planet.Sci.Lett.,96, 8-14, 1989.
Bailey, D.K., Mantle metasomatism--Perspective
and prospect,in
AlkalineIgneous
Rocks,R. Geol. Soc.Spec.Publ.30, editedby J.G.
Fitton and B.G.J. Upton, pp. 1-13, BlackwellScientific,Boston,
Mass., 1987.
Baldridge,W.S., Y. Eyal,Y. Bartov,G. Steinitz,andM. Eyal,Miocene
magmatismof Sinairelatedto the openingof the Red Sea,Tectonophysics,197, 181-201,1991.
Barberi,F., andJ. Varet, Volcanismof Afar: Small-scale
platetectonic
implications,Geol.Soc.Am. Bull., 88, 1251-1266,1977.
BarberiF., G. Capaldi,P. Gasperini,G. Marinelli,R. Santacroce,
R.
Scandone,M. Treuil, and J. Varet, Recent basalticvolcanismof
rift armscomprisingthe triple junction,the Red Sea and Gulf
Jordanand its implications
on the geodynamic
historyof the Dead
of Aden rifts were well developedby 20 Ma as the resultof
Sea shearzone,in Geodynamic
Evolutionof theAfro-ArabianRift
passive-mantle
rifting [Cochran,1981, 1983]; in contrast,the
System,
Publ.47, pp.667-683,AcademiaNationaledei Lincei,Rome,
1980.
Ethiopianrift did not developuntil about10 Ma [Kazminand
Berhe,1978;Kazminet al., 1980],shortlyafterthe initiationof Barrat,J.-A., B.M. Jahn,J.-L. Joron,B. Auvray,andH. Hamdi,Mantle
heterogeneityin northeasternAfrica: Evidencefrom Nd isotopic
significantcrustaluplift alongthe crestof the Afro-Arabian
compositions
andhygromagmaphile
elementgeochemistry
of basaltic
Dome •llrnomt, 1986a,b;Bohannonet al., 1989]. The Ethioprocksfrom the Gulf of Tadjouraand southernRed Sea regions,
Earth Planet. Sci. Lett., 101, 233-247, 1990.
ian rift appearsto be the only arm of the Afar triplejunction
which can unambiguously
be consideredthe sole productof Bayer,H.-J., H. Hotzl, A.R. Jado,B. Roscher,and W. Voggenreiter,
Sedimentaryand structuralevolutionof the northwestArabianRed
active-mantleupwelling,contemporaneous
with active-mantle
Sea margin,Tectonophysics,
153, 137-151,1988.
upwellingbeneaththe MMN line on the WestArabianSwell. Bayer, H.-J., Z. El-Isa, H. Hotzl, J. Mechie, C. Prodehl,and G.
Saffarini,Largetectonicand lithospheric
structures
of the Red Sea
In our model,three limbsof upwellingasthenosphere
are
connectedat the Ethiopian mantle plume (Figure 11).
Although two coincidewith the Ethiopianand Gulf of Aden
rifts, the third limb of upwellingdivergesfrom the Oligo-
MioceneRed Sea Basin(Figure11). The apparentshallow
region,J. Afr. Earth Sci.,8, 565-588,1989.
Berckenhemer,H., B. Baier, H. Bartelsen,A. Behle, H. Burkhardt,H.
Gebrande,J. Makris, H. Mezel, H. Miller, and R., Vees, Deep
seismicsoundingsin the Afar region and on the highlandof
Ethiopia,in Afar Depression
of Ethiopia,pp. 89-107,Schweizerbart,
Stuttgart, 1975.
asthenosphere
beneaththe centraland northernRed SeaBasin
constraints
on theevolution
hasbeen attributedto a passivemantlefillingthe voidsof an Berhe,S.M., Geologicandgeochronologic
of the Red Sea-Gulfof Aden andAfar Depression,
J.Afr. EarthSci.,
extendedcrust [Cochran,1983; Cochranet al., 1991]. We
5, 101-117, 1986.
attribute seafloorspreadingin the southernRed Sea to the Beydoun,Z.R., The Gulf of Aden and northwestArabianSea,in The
Plioceneinvasionof thisextendedcrustby the north-south
limb
OceanBasinsand Margins,vol. 6, The Indian Ocean,editedby
A.E.M. Nairn and F.G. Stehle,pp. 253-313,Plenum,New York,
of risingasthenosphere
describedabove. On this basis,we
1982.
suggestthat the southernRed Sea spreadingaxismay be a
Blank, H.R., Aeromagneticand geologicstudyof Tertiary dikesand
uniqueexampleof a doublypropagating
rift systemeccentric
related structureson the Arabian marginof the Red Sea, in Red
to its zone of active-mantleupwelling.
Sea Research, 1970-1975: Saudi Arabian Dir. Gen. Min. Res. Bull.,
G1-G18, 1977. (U.S. GeologicalSurvey,Saudi Arabian Project
Report 210, Reston,Va., 1977)
of the 1976aeromagnetAcknowledgments.
Thisstudyisfoundedupona projectto mapthe Blank,H.R., andH.S. Sadek,Spectralanalysis
ic surveyof Harrat Rahat, Kingdomof SaudiArabia, U.S. Geol.
Saudi Arabian harrats as part of the work programof the Saudi
ArabianDirectorateGeneralof MineralResources
(DGMR), Jiddah.
Surv.OpenFile Rep.,USGS-OF-03-67,29 pp., 1983.
of passive
The work is publishedwith the approvalof the AssistantDeputy Bohannon,R.G., and S.L. Eittreim,Tectonicdevelopment
continental marginsof southern and central Red Sea with a
Ministerdesignate
for SurveyandExplorationat theDGMR, Mohamcomparison
to WilkesLand,Antarctica,Tectonophysics,
198,129-154,
medTawfiq,whowe thankfor continuedencouragement
andsupport.
1991.
We thank Bob Bohannonand Anne McGuire for thoughtfulreviews
Res.,95, 5165and KevinBurkeand PeterHooperfor providinguswithconstructive Bohannon,R.G., andA.V. McGuire,Reply,J. Geophys.
5170, !990.
commentson earlier manuscriptdrafts.
CAMPANDROOBOL:UPWELLING
ASTHENOSPHERE,
WESTERN
ARABIA
Bohannon,R.G., C.W. Naeser,D.L. Schmidt,and R.A. Zimmermann,
The timingof uplift,volcanism,
and rifting peripheralto the Red
Sea:A casefor passiverifting?,J. Geophys.
Res.,94, 1683-1701,1989.
Bonatti,E., Punctiforminitiationof seafloor spreadingin the Red Sea
duringtransitionfroma continental
to an oceanicrift, Nature,316,
33-37, 1985.
Bonatti,E., Not so hot "hotspots"in the oceanicmantle,Science•
250,
107-111, 1990.
Brown, G.F., Tectonic map of the Arabian Peninsula,Dir. Gen. of
Miner. Resour.,Arabian PeninsulaMap AP-2, scale 1:4,000,000,
Jiddah, Saudi Arabia, 1972.
Brown,G.F., D.C. Schmidt,and C. Huffman,Geologyof the Arabian
PeninsulaShieldarea of westernSaudiArabia, U.S. Geol.Sur.Prof.
Pap., 560-A, 188 pp., 1989.
Burke, K., and J.F. Dewey, Plume generatedtriple junctions:Key
indicatorsin applyingplatetectonics
to old rocks,J. Geol.,81, 406433, 1973.
Camp,V.E., and M.J. Roobol,The Arabiancontinentalalkalibasalt
province,part I, Evolutionof Harrat Rahat, Kingdomof Saudi
15,269
Crough,S.T., Rifts and swells,geophysical
constraintson causality,
Tecto;.,ophysics,
94, 23-37, !983.
Dixon,T.H., E.R. Ivins,and B.J. Franklin,Topographicand volcanic
asymmetryaroundthe Red Sea:Constraintson rift models,Tectonics, 8, 1193-1216, 1989.
Dixon, T.H., E.R. Ivins, and B.J. Franklin, Reply, Tectonics,
10, 653655, 1991.
du Bray, E.A., D.B. Stoeser,and E.H. McKee, Age and petrologyof
the Tertiary As Sarat volcanic field, southwesternSaudi Arabia,
Tectonophysics,
198, 155-180, 1991.
Dunn, T., and I.S. McCallum,The partitioningof Zr and Nb between
diopsideandmeltsin the systemdiopside-albite-anorthite,
Geochim.
Cosmochim.Acta, 46, 623-629, 1982.
Fitton, J.G., The Cameroonline, West Africa:A comparison
between
oceanicand continentalalkaline volcanism,in Alkaline Igneous
Roc'ks,R. Geol. Soc. Spec. Publ. 30, edited by J.G. Fitton and
G.G.L. Upton,pp.273-291,BlackwellScientific,Boston,Mass.,1987.
Foucher, J.-P., X. Le Pichon, and J.-C. Sibuet, The ocean-continent
transition in the uniform lithosphere stretchingmodel: Role of
partialmeltingin the mantle, Philos.Trans.R. Soc.London,Ser.A,
Arabia, Geol. Soc.Am. Bull., 101, 71-95, 1989.
305, 27-43, 1982.
Camp,V.E., andM.J. Roobol,Commenton 'Topographic
andvolcanic
asymmetry
aroundthe Red Sea:constraints
on Red Seamodels"
by Gass, I.G., Tectonic and magmatic evolution of the Afro-Arabian
T.H. Dixon, E.R. Ivins, and J.F. Brenda, Tectonics,10, 649-652,
1991a.
Camp,V.E., and M.J. Roobol,Geologicmap of the CenozoicHarrat
Rahat lava field, Kingdomof SaudiArabia, Geosci.Map GM-130,
with text, scale1:250,000,101 pp.,SaudiArabianDeputyMinist.for
Miner. Resour.,Jiddah, 1991b.
dome,in African Magmatismand Tectonics,
editedby T.E. Clifford,
and I.G. Gass,pp. 285-300, Oliver and Boyd, Edinburgh,1970.
Gettings,M.E., A heat flow profile acrossthe ArabianShieldand Red
Sea, Eos Trans.AGU, 62, 407, 1981.
Gettings,M.E., H.R. Blank, W.D. Mooney, and J.H. Healey, Crustal
structureof southwesternSaudiArabia, J. Geophys.
Res.,91, 6491-
6512, 1986.
Camp, V.E., M.J. Roobol, and P.R. Hooper, Intraplate alkalic
volcanismand magmaticprocesses
alongthe 600-km-longMakkah- Gillman, M., Primary resultsof a geologicaland geophysical
reconMadinah-Nafud volcanic line, western Saudi Arabia, in General
naissanceof the Jizancoastalplain in SaudiArabia, in Proceedings,
pp. 189-212,Societyof PetroleumGeologists,
SaudiArabianSection,
Assemblyof the International Associationof Volcanologyand
SecondRegional Technical Symposium,Dhahran, Saudi Arabia,
Chemistryof theEarth'sInterior,Continental
Magmatism
Abstracts,
N.M. Bur. Mines Geol. Bull., 131, 49, 1989.
Proceedings,1968.
Camp,V.E., M.J. Roobol, and P.R. Hooper,The Arabiancontinental Girdler, R.W., and T.C. Southren, Structure and evolution of the
northern Red Sea, Nature, 330, 716-721, 1987.
alkali basalt province, part II, Evolution of Harrats Khaybar,
Ithnayn,and Kura, Kingdomof SaudiArabia,Geol.Soc.Am. Bull., Greenwood,W.R., The Ha'il arch--a key to the deformationof the
103, 363-391, 1991.
Arabian Shield during evolutionof the Red Sea rift, SaudiArabian
Camp,V.E., M.J. Roobol,and P.R. Hooper,The Arabiancontinental
Dir. Gen. Miner. Resour.Bull. 7, 5,1973.
alkalibasaltprovince,part III, Evolutionof Harrat Kishb,Kingdom Gregory, R.T., R.G. Coleman, and G.F. Brown, Cenozoicvolcanic
of Saudi Arabia, Geol. Soc.Am. Bull., 104, 379-396, 1992.
rocksof SaudiArabia: Evidencefrom the continentfor a two-stage
Capaldi,G., P. Chiesa,G. Manetti, G. Orsi, and G. Poli, Tertiary
openingof the Red Sea, Geol. Soc.Am. Abstr.Programs,14, 502,
anorogenicgranitesof the westernborder of the Yemen Plateau,
1982.
Lithos, 20, 433-444, 1987.
Chiesa,S., L. Civetta, M. De Fino, L. La Volpe, and G. Orsi, The
Yemen trap series:genesisand evolutionof a continentalflood
basaltprovince,J. Volcanol.Geotherrn.
Res.,36, 337-350,1988.
Civetta,L., L. LaVolpe, and L. Lirer, K-Ar agesof the YemenPlateau,
Hancock,P.L., A. AI Kadhi, and N.A. Sha'at,Regionaljoint setsin the
Arabian platformas indicatorsof intraplateprocesses,
Tectonics,
3,
27-43, 1984.
Hempton, M.R., Constraintson Arabian plate motionand extensional
historyof the Red Sea, Tectonics,6, 687-705, 1987.
J. Volcanol. Geotherm.Res., 4, 307-314, 1978.
Henjes-Kunst,F., Mantle xenolithsin westernArabia andtheir bearing
Civetta, L., M. De Fino, L. La Volpe, and L. Lirer, Recentvolcanism
on asthenosphere-lithosphere
dynamicsduringformationof the Red
of North Yemen:Structuralandgeneticimplications,in Geodynamic
Sea Rift--A review, Spannungund Spannungsumwandlung
in der
evolutionof the Afro-Arabianrift syston,Publ. 47, pp. 617-628,
lithosphare,
Symposium
on theAfro-ArabianRift System,pp. 49-50,
Academia Nationale dei Lincei, Rome, 1980.
KarlsruheUniversity,Karlsruhe,Germany,1989.
Cloos,H., Hebung-Spaltung-Volcanismus,
Geol.Rundsch.,30, 405-427, Henjes-Kunst,F., R. Altherr, andA. Baumann,Evolutionandcomposi1939.
tion of the lithosphericmantle underneath the western Arabian
Cochran,J.R., The Gulf of Aden: Structureand evolutionof a young
peninsula:Constraintsfrom Sr-Nd isotopesystematics
of mantle
oceanbasin and continentalmargin,J. Geophys.Res.,86, 263-287,
xenoliths,Cont•ib. Mineral. Petrol., 105, 460-472, 1990.
1981.
R.W., andG.G. Engels,Tectonicsignificance
andregional
Cochran,J.R., A modelfor development
of Red Sea,AAPG Bull., 67, Hutchinson,
geologyand evaporiticlithofaciesin northeasternEthiopia,Philos.
41-69, 1983.
Trans.R. Soc. London, Ser.A, 267, 313-338, 1970.
Cochran, J.R., J.-M. Gaulier, and X. Le Pichon,Crustal structureand
Irvine, T.N., and W.R. Baragar,A guideto the chemicalclassification
the mechanism of extension in the northern Red Sea: constraints
of the commonvolcanicrocks,Can. J. Earth Sci., 8, 523-548, 1971.
from gravityanomalies,Tectonics,
10, 1018-1037,1991.
Coleman,R.G., andA.V. McGuire, Magmasystems
relatedto the Red Izzeldin, A.Y., On the structureand evolutiono[ tl•e Red Sea, Ph.D.
dissertation,
Univ. LouisPasteur,163 pp.,Strasbourg,
France,1982.
Seaopening,Tectonophysics,
150, 77-100, 1988.
Coleman, R.G., R.T. Gregory, and G.F. Brown, Cenozoicvolcanic Izzeldin,A.Y., Seismic,gravityand magneticsurveysin the centralpart
of the Red Sea: Their interpretationand implicationsfor the
roc'ksof SaudiArabia, U.S. Geol. Surv.OpenFile Rep., USGS-OFstructureandevolutionof the Red Sea,Tectonophysics,
143, 269-306,
03-93, 82 pp., 1983.
1987.
Courtillot, V., Propagatingrifts and continental breakup, TectonJaques,A.L., andD.H. Green,Anhydrous
meltingof peridotiteat 0-15
ophysics,
1, 239-250, 1982.
kbpressure
andgenesis
of tholeiiticbasalts,
Contrib.
Mineral.Petrol.,
Courtillot, V., A. Galdeano,and J.L. LeMouet, Propagationof an
73, 287-310, 1980.
accretingplate boundary:A discussion
of new aeromagnetic
data in
the Gulf of TadjourahandsouthernAfar, Earth Planet.Sci.Lett.,47 Jarrige,J-J.,P.O. Destevou,P.F. Burollet,C. Montenat,P. Prat,J.-P.
Richert, and J.-P. Thiriet, The multistagetectonicevolutionof the
144-160, 1980.
Cox, K.G., A model for flood basalt vulcanism,J. Petrol., 21, 629-650,
1980.
Gulf
of Suez and northern
Red
Sea continental
observations,Tectonics,9, 441-465, 1990.
rift
from
field
15,270
CAMPANDROOBOL:
UPWELLING
ASTHENOSPHERE,
WESTERN
ARABIA
Kay,R.W.,andP.W.Gast,The rareearthcontentandoriginof alkalirich basalts,J. Geol., 81,653-682, 1973.
Kazmin,V.G., and S.M. Berhe, Geologyand development
of the
Nazref area,Rep. 100, EthiopianInst. of Geol.Sun,.,AddisAbaba,
1978.
Kazmin, V.G., S.M. Berhe, M. Nicoletti, and C. Petrucciani,Evolution
of the northernpart of the Ethiopianrift, in Geodynamic
Evolution
Roeser,H.A., A detailedmagneticsurveyof the southernRed Sea,
Geol. Jahrb.,ReiheD, 13, 131-153, 1975.
Roobol,M.J., and V.E. Camp,Geologyof the Cenozoiclavafield of
HarratsKhaybar,Ithnayn,andKura,Geosci.
MapGM-131,1:250,000
scale,with text,76 pp.,SaudiArabianDir. Gen of Miner.Resour.,
Jiddah, 1991a.
Roobol,M.J. and V.E. Camp,Geologyof the Cenozoiclavafieldof
of the Afro-ArabianRift System,Publ. 47, pp. 275-292,Academia
Harrat Kishb,Geosci.Map GM-132, 1:250,000scale,with text,63
Nazionale dei Lincei, Rome, 1980.
pp.,SaudiArabianDir. Gen. of Miner. Resour.,Jiddah,1991b.
Kushiro,I., The systemforsterite-diopside-silica
with andwithoutwater
at highpressures,
Am. J. Sci.,267A, 269-294,1969.
Le Bas, M.J., R.W. Le Maitre, A. Streckeisen,and B. Zanettin, A
chemical
classification
of volcanic
rocks based on the total alkali-
silicadiagram,J. Petrol.,27, 745-750, 1986.
Le Pichon,X., andJ. Francheteau,A plate tectonicanalysisof the Red
Sea-Gulf of Aden area, Tectonophysics,
46, 369-406,1978.
Le Pichon, X., and J.M. Gaulier, The rotation of Arabia and the
Levant fault system,Tectonophysics,
153, 271-294,1988.
Martinez, F., and J.R. Cochran, Geothermal measurementsin the
northern Red Sea: Implicationsfor lithosphericthermal structure
and mode of extensionduring continentalrifting J. Geophys.Res.,
94, 12,239-12,265, 1989.
McGuire, A.V., Petrologyof mantlexenolithsfrom Harrat al Kishb:
The mantle beneath western Saudi Arabia, J. Petrol., 29, 73-92,
1988a.
Ryan,M.P., The mechanics
and three-dimensional
internalstructureof
activemagmatic
systems:
Kilaueavolcano,
Hawaii,J. Geophys.
Res.,
93, 4213-4248, 1988.
Sack,R.O.,D. Walker,andI.S.E.Carmichael,
Experimental
petrology
of alkaliclavas:constraints
on cotecticsof multiplesaturationin
naturalbasicliquids,Contrib.Mineral.Petrol.,96, 1-23, 1987.
Schmidt,D.L., and D.G. Hadley,Stratigraphy
of the MioceneBaid
formation,southernRed Sea coastalplain, Kingdomof Saudi
Arabia,Tech.Rec.USGS-TR-04-23,
46 pp.,SaudiArabianDeputy
Minist. for Miner. Resour.,Jiddah, 1984.
Schmidt,D.L., D.G. Hadley, and G.F. Brown, Middle Tertiary
continentalrifting and evolutionof the Red Sea in southwestern
SaudiArabia, U.S. Geol.Surv.OpenFile Rep.,USGS-OF-03-6,
56
pp., 1983.
Sebai,A., G. Feraud,G. Giannerini,R. Campredon,
andH. Bertrand,
39Ar-4øAr
datingon Cenozoic
maficvolcanics
of theArabianPlate
McGuire, A.V., The mantlebeneaththe Red Sea margin:Xenoliths
associated
with earlyRed Seaopening,TerraCognita,7, 201, 1987.
from westernSaudi Arabia, Tectonophysics,
150, 101-119,1988b.
McGuire, A.V., and R.G. Bohannon,Timing of mantle upwelling: Sebai, A., V. Zumbo, G. Feraud, H. Bertrand, A.G. Hussain,G.
Giannerini,and R. Campredon,
4øAr/39Ar
datingof alkalineand
Evidencefor a passiveorigin for the Red Searift, J. Geophys.
Res.,
94, 1677-1682, 1989.
McKenzie,D., The generation
andcompaction
of partiallymoltenrock,
tholeiiticmagmatism
of SaudiArabiarelatedto the earlyRed Sea
Rifting,Earth Planet.Sci.Lett., 104, 473-487,1991.
Sengot,
A.M.C.,andK. Burke,Relativetimingof riftingandvolcanism
on Earth and its tectonicimplications,
Geophys.
Res.Lett.,5, 419McKenzie,D., andM.J. Bickle,The volumeandcomposition
of melt
421, 1978.
generatedby extensionof the lithosphere,
J. Petrol.,29, 625-679,
J. Petrol.,25, 713-765, 1984.
1988.
Sengot, A.M.C., and W.S.F. Kidd, Post-collisionaltectonicsof the
Turkish-Iranianplateauand a comparison
with Tibet, TectonoMohr,P.A.,Ethiopianfloodbasaltprovince,
Nature,303,577-584,1983.
physics,55, 361-376, 1976.
Mohr, P.A.,andB. Zanettin,The EthiopianFloodBasaltProvince,
in
of
Continental
FloodBasalts,editedby J.D. Macdougal,
pp. 63-110, Steinitz,G., Y. Bartov,andJ.C.Hunziker,K-Ar agedetermination
someMiocene-Pliocene
basaltsin Israel:Their significance
to the
Kluwer Academic,Boston,Mass., 1988.
of the Rift Valley,Geol.Mag.,115,329-340,1978.
Mysen,B.O., and I. Kushiro,Compositional
variations
of coexisting tectonics
for mid-ocean
ridgebasalts:
Preliminary
phases
withdegreeof meltingof peridotite
in theuppermantle,
Am. Stopler,E., A phasediagram
Mineral., 62, 843-865, 1977.
resultsandimplications
forpetrogenesis,
Contrib.
Mineral.Petrol.,74,
7-12, 1980.
Nasir,S., The lithosphere
beneaththe northwestern
part of the
composition
Arabianplate(Jordan):Evidence
fromxenoliths
andgeophysics,Stopler,E., andD. Walker,Melt densityandtheaverage
of basalt,Contrib.Mineral.Petrol.,74, 7-12, 1980.
Tectonophysics,
201, 357-370, 1992.
andthe originof the RedSeaandthe
O'Hara, M.J., and H.S. Yoder, Formation and fractionationof basic Tarling,D.H., Paleomagnetism
Gulf of Aden,Philos.Trans.R. Soc.London,Ser.A, 267, 219-226,
magmasat high pressures,
Scot.J. Geol.,3, 67-117,1967.
1970.
Omar, G.I., M.S. Steckler,W.R. Buck,and B.P Kohn,Fission-track
R.N., and S.A. Gibson,Subcontinental
mantleplumes,
analysis
of basement
apatitesat thewesternmarginof theGulf of Thompson,
andpre-existing
thinspots,
J. Geol.Soc.,London,148,973Suez,Egypt:Evidencefor synchroneity
of uplift and subsidence, hotspots
Earth Planet.Sci.Lett., 94, 316-328, 1989.
977, 1991.
withpetrologic
Pallister,
J.S.,Reconnaissance
geologic
mapof theHarratHutaymah Thornber,C.R., Geologicmapof HarratHutaymah,
quadrangle,
sheet26/42A,Kingdomof SaudiArabia,U.S.Geol.Surv.
OpenFile Rep.USGS-OF-04-46,
77 pp., 1984.
Pallister,
J.S.,Magmatichistoryof Red Searifting:Perspective
from
the centralSaudiArabiancoastalplain,Geol.Soc.Am. Bull.,98,
400-417, 1987.
Parsons,B., and D. McKenzie, Mantle convectionand the thermal
structureof plates,J. Geophys.
Res.,83, 4485-4496,1978.
classification
anddistribution
of ultramafic
inclusions,
SaudiArabia,
U.S. Geol. Surv.Mis. Field Stud.Map, MF-2129, 1:100,000-scale,
1990.
Thornber, C.R., and J.S. Pallister, Mantle xenolithsfrom northern
SaudiArabia,Eos Trans.AGU, 66, 393, 1985.
Voggenreiter,
W., H. Hotzl, and J. Mechie,Low-angle
detachment
originfor the Red SeaRift System?,
Tectonophysics,
150, 51-75,
1988.
Pearce,J.A., J.F. Bender,S.E. De Long,W.S.G. Kidd, P.J. Low, Y.
Guner, F. Saroglu,Y. Yilmaz, S. Moorbath,and J.G. Mitchell, von Herzon,R.P., R.S. Detrick,S.T. Crough,D. Epp, and U. Fehn,
Thermal origin of the Hawaiian swell: Heat flow evidenceand
Genesisof collisionvolcanismin easternAnatolia,Turkey,J.
thermalmodels,J. Geophys.,
Res.,87, 6711-6723,1982.
Volcanol.Geothertn.
Res.,44, 189-229,1990.
Plaziat,J.-C.,C. Montenat,F. Orszag-Sperber,
E. Philobbos,
andB.H. Walker,D., T. Shibata,andS.E.Delong,Abyssal
tholelites
fromthe
Purser,
Geodynamic
significance
ofcontinental
sedimentation
during
initiationof theNW RedSearift (Egypt),J.Aft.EarthSci.,10,355360, 1990.
Powers,R.W., L.F. Ramirez,C.D. Redmond,and E.L. Elberg,Jr.,
Geologyof the ArabianPeninsula,
sedimentary
geologyof Saudi
Oceanographer
FractureZone, II, Phaseequilibriaand mixing,
Contrib.Mineral.Petrol.,7, 111-125,1979.
Watson,E.B., and F.J. Ryerson,Partitioning
of zirconiumbetween
clinopyroxene
andmagmatic
liquidsof intermediate
composition,
Geochbn.Cos•nochim.
Acta, 50, 2523-2526,1986.
Arabia, U.S. Geol.Surv.Prof.Pap.,560-d, 1966.
White,R.,andD. McKenzie,
Magmatism
at riftzones:
Thegeneration
Riddler,G.P.,M. vanEck,N.C.Aspinall,
J.J.McHugh,T.W.H.,Parker,
of volcanic
continental
margins
andfloodbasalts,
J. Geophys.
Res.,
A.M., Farasani,and S.M. Dini, An assessment
of the phosphate
94, 7685-7729, 1989.
resourcepotentialof theSirhan-Turayf
region,Tech,Rec.RF-TR-06- Wood,D.A.,J.L.Joron,
M. Treuil,M.J.Norry,andI. Tarney,
Elemen2, withl:500,000-scale
coloredgeologic
map,194pp.,SaudiArabian
tal andSr isotopevariationsin basiclavasfromIcelandandthe surDeputy Minist. for Miner. Resour.,Jiddah,1986.
roundingoceanfloor,Cont•ib.Mineral.Petrol.,70, 319-339,1979.
CAMP AND ROOBOL:UPWELLINGASTHENOSPHERE,
WESTERNARABIA
Zanettin, B., E. Justin-Visentin, M. Nicoletti, and E.M. Piccirillo,
Correlationsamong Ethiopian volcanicformationswith special
referencesto the chronological
and stratigraphical
problemsof the
Trap Series, in GeodynamicEvolutionof the Afro-ArabianRift
System,
Publ.47, pp.231-252,AcademiaNationaledeiLincei,Rome,
V.E. Camp, Departmentof GeologicalSciences,San Diego State
University.,
c/o 4757 EdgewareRoad, SanDiego,CA_92116.
M.J. Roobol, Saudi Arabian Directorate General of Mineral
Resources,Box 345, Jiddah, 21191, Saudi Arabia.
1980.
Zielinski, R.A., Trace element evaluation of a suite of rocks from
Reunion Island, Indian Ocean, Geochim.Cosmochim.Acta, 39, 713734, 1975.
15,271
(Received July 10, 1991;
revisedApril 23, 1992;
acceptedApril 23, 1992.)