20. LATE CRETACEOUS AND CENOZOIC PELAGIC DEPOSITS - … · 2007. 6. 5. · LATE CRETACEOUS &...

20. LATE CRETACEOUS AND CENOZOIC PELAGIC DEPOSITS - PALEOENVIRONMENTAND PALEOOCEANOGRAPHY OF THE CENTRAL WESTERN INDIAN OCEAN

Lucien Leclaire, Laboratoire de Geologie du Museum National d'Histoire Naturelle, Paris

ABSTRACTPelagic deposits are the dominant contributors to the sediment

layers that blanket ridges and basaltic sea floor in deep basins.Among pelagic constituents, nannofossils undoubtedly are thepredominant contributor, especially to Cenozoic chalk and ooze.Among Leg 25 sites, the most typical pelagic deposits are foundin the Madagascar Basin (Site 245) including a clayey nannochalk, bearing Fe-Mn oxides, resting directly on the basalt andoverlain successively by a nanno chalk which is interlayered withchert and devitrified volcanic ash and a nearly pure nanno ooze.The lithologic sequence is capped by brown clay. Upper Cenozoicforam sand and foram nanno ooze were found on theMozambique Ridge (Site 249) and particularly on the MadagascarRidge (Site 246). Biogenic silica-rich facies are rare and mainlylocated near the top of the lowest latitude sites.

Descriptions of the major sedimentary facies lead todiscussions and speculations about depositional conditions andhence about the Cenozoic paleoenvironment in the centralwestern Indian Ocean. Important changes in paleoenvironmentalconditions occurred throughout the Tertiary. Thus, increasingcontent in biogenic silica, together with greater foraminiferaamounts in the Somali Basin and on the Davie Ridge, plus theoccurrence of notable quantities of organic matter, are very ükelyconnected to increasing productivity of the overlying watermasses through late Cenozoic times. These facts are interpreted ascorresponding to the gradual onset of the equatorial water masscirculation, whereas the change in redox conditions over theDavie Ridge area is believed to be related to the arrival ofoxygen-depleted waters flowing southward from the Arabian Seaand Gulf of Aden. Formerly, during Eocene and Oligocene times,a great gyral circulation presumably supplied (at least partly) bythe Pacific south equatorial current would have swept thesouthern Indian coast, which was then in the equatorial position,and would have been deflected southwards, thereby winnowingthe Madagascar and Mozambique ridge sediments. The majorcause of such changes in circulation would be the widening of thenorth central part of the Indian Ocean, i.e., the northward driftof India and the Carlsberg Ridge spreading associated with thegradual closing of the north Australian seaway.

Strong western boundary undercurrents supplied withAntarctic Bottom Water, presumably initiated by late Oligoceneto early Miocene times, probably eroded large surfaces of theSomali, Mascarene, Madagascar, and Mozambique basins. Theyare considered to be the primary cause of important hiatuses inthe sedimentary columns that encompass at least most of theOligocene. These deep bottom currents would have been partiallyblocked by the uplift of the Southwest Indian Ridge in middleMiocene time; they still exist today as weak flows in theMozambique, Madagascar, and Mascarene basins.

Paleodepths obtained by "back tracking" the age/depthconstancy curve of the basaltic floor and paleontological datasuggest important changes in the calcium carbonate compensationdepth (CCD) and the associated foram-lysocline through Tertiarytime. The last change is well marked by a relatively shallowlysocline that was followed by a sharp increase in depth in lateMiocene time. This depth increase coincided with an acceleration

481

L. LECLAIRE

in sea-floor spreading and may have originated by volcanogenicsupplies and the weathering of basaltic crust in contact withseawater which lead to the removal of great quantities of calciumfrom fresh basalts.

INTRODUCTION

Objectives

The major objectives of this chapter are the following:(1) to gather data about pelagic sediments in one chapter;(2) to emphasize the pelagic deposit contribution to thesediment columns of ridges and basins; and (3) to test thespreading sea-floor hypothesis with regard to changes inpelagic sedimentation, and thereby to draw attention tomajor changes in paleoenvironmental conditions throughCenozoic time in the central western Indian Ocean. Thefirst two points are restricted more or less to facts, but thethird is open to reasonable interpretations and speculationsbased on these facts.

A major difficulty in this study is the great diversity ofdrilling sites (Figure 1) that are situated on ridges and inbasins each of which really constitutes a particular problem.This diversity certainly limits global correlations of currentchanges and sea-floor spreading phenomena. However, somefacts gathered here give strong evidence for changes incirculation patterns, deepening and shallowing of the seafloor, and large-scale horizontal movements of thecontinents and sea floor. The reader is reminded that manyinterpretations formulated and discussed in this report aresomewhat speculative and need much more checking.However, it is emphasized that the findings from studies ofthis part of the world oceans support the contention thatpelagic sedimentation has been strongly influenced bysea-floor spreading phenomena.

Methods

The single most important data source for the presentaccount is the shipboard Hole Summary Book, especiallythe lithology and paleontology parts which were preparedduring the Leg 25 cruise and later revised on shore.Additional material comes from post-cruise laboratorywork in Paris, mainly calcium carbonate determinations,smear slide analyses, and scanning electron microscopeinvestigations.

PELAGIC DEPOSITS IN THE LITHOLOGIC SEQUENCES

Definition and Classification

Before dealing with specific deposits at the drilling sites,it is necessary to define the meaning of "pelagicsediments." Moreover, because a classification forms thebasis for genetic interpretations, the classification used hereis explained.

"Pelagic" comes from the greek "pelagos" which meanssea. By extension, the general use designates all things

related to the oceanic environment as pelagic; examples arepelagic fauna and pelagic sediments. Pelagic sediments,land-derived or terrigenous deposits, and volcanic sedimentsconstitute the basis for several classifications of marinesediments.

Attempts to classify pelagic sediments have been basedeither on appearances and compositions, or on the origin ofthe components. The later scheme has been applied byArrhenius (1963). Although the origin of most biogenouscomponents is generally obvious, the origin of the majornonbiogenous components has not been established withcertainty in most areas. It appears feasible to distinguishminerals which crystallize in oceanic waters from thosewhich originate, for instance, in basaltic magmas andhydrothermal solutions. Minerals which arise from oceanicwater are described, according to Arrhenius (1963), as"halmeic" (from seawater). The components derived fromexposed continental surfaces are here designated asterrigenous. Part of the terrigenous fraction undergoesslight modifications which are more or less related toion-exchange capacity and to the very long passage of timethrough oceanic waters. This part (mostly clay minerals) ismixed with a major nonaltered mineral fraction and insome places with an authigenic fraction (for example,authigenic clay minerals). Relative proportions of thecomponents generally are unknown and are difficult todetermine. This mixture of components is namedhemihalmeic. Minerals emplaced in the ocean by volcanicactivity are described as volcanogenic. Tests secreted byliving organisms are referred to as biogenous. Biogenous andhalmeic components are truly pelagic in origin. Othercomponents, land derived and volcanogenic, areallochthonous. The classification used here is mostly that ofArrhenius (1963) and is summarized in Table 1.

Because pelagic deposits generally contain a terrigenousfraction, one can expect that it is possible to find depositsintermediate between pelagic and terrigenous. These weredesignated as hemipelagic by Kuenen (1950). Distinction asto whether a deposit is classified as pelagic or hemipelagic isbased on the sedimentation rate. According to Arrhenius(1963), "In the basins accumulating pelagic sediments, theterrigenous deposit rate appears to vary not much morethan one order of magnitude (5X105 to 5X1O4 cm/year)."This is very low when compared to sediment accumulationsin basins fringing the continents. In other words, a rate ofterrigenous deposition not higher than 5 m/m.y. seems tocharacterize and, to some extent, define pelagic deposits.

Volcanic contributions are generally minor in pelagicdeposits, and in most instances they are considered aspelagic components. However, when they are dominant, thesediment can no longer be logically referred to as pelagicbut rather as volcanogenic (or pyroclastic).

482

LATE CRETACEOUS & CENOZOIC PELAGIC DEPOSITS

30° 35° 40°

DSDP LEG 25

TRACK AND SITE LOCATIONS• S.dim.nti and basalt p.n.trot.d• S dim nti only p netrated

A F R I C A

us»

252

35°

35' 40° 45

Figure 1. Index map showing the Leg 25 drilling sites.

50° 55° 60°

483

L. LECLAIRE

TABLE 1Pelagic Deposit Classification(Mostly from Arrhenius, 1963)

Nature of Particles Deposited Components Groups

Foraminifera, nannofossils,sponge spicules, radiolarians

Zeolites, pyrite, Mn-Fe oxides

Long passage time clayminerals

Land-derived detritalparticles

Volcanic glass, shards,devitrified volcanic clays

Biogenous (biogenic)components

Halmeic components

Hemihalmeiccomponents

Terrigenouscomponents

Pyroclastic (volcano-genie components)

Pelagic Deposits at Leg 25 Sites

Using shipboard smear-slide analyses, along withcomplementary smear-slide investigations in the shorelaboratory, a diagram has been constructed (Figure 2).Obviously, such analyses provide only semiquantitativedata, but they are sufficient, in a preliminary attempt, tobroadly outline the relative contribution of pelagicsedimentation to the sediment column at each site. Pelagiccomponents are very important contributors to thesedimentary blanket that was cored in the central westernIndian Ocean.

At Site 239, in the Mascarene Basin, the sedimentcolumn consists of about 50 percent pelagic deposits.Although the top of the lithologic series (subunit IA)consists of minor clayey nanno ooze, subunit IB is mostlyclay-rich nanno ooze and subunit IC is a mixture ofterrigenous detrital deposits and carbonate-rich nanno ooze.In some intervals, part of the clay fraction may beconsidered as pelagic because of the high amount ofmontmorillonite (50%-60%). Below 150 meters, the sedi-mentary column consists predominantly of mixed detritalsilty clay and nanno clay. This sediment can be referred toas hemipelagic (Moore, this volume). From 280 metersdown to the basalt, the major lithology is brown clay whichhas an important volcanic contribution. Nanno ooze isinterlayered with the brown clay, part of which might bedetrital (pelagic turbidites).

At Site 240 in the Somali Basin, the lithologic sequenceis mostly detrital (terrigenous sand and silt). However, anotable pelagic sediment contribution occurs at the top ofthe sequence as nanno rad ooze and at the bottom as nannoooze and carbonate-rich nanno ooze. Site 241, drilled insediments which fringe the East African continental margin,displays a lithologic record characterized by two units. Theupper one, from the top down to 470 meters, consistsdominantly of pelagic components (mostly nanno oozemixed with some detrital clay). Certain interlayered beds,apparently of pelagic origin, are redeposited sedimentswhich had been transported by turbidity currents. Belowthis sequence (lower unit), flysch-type detrital sedimentand rock compose the lithologic column. Hence, at leastone-third of the sedimentary sequence, from the upperOligocene to Pleistocene, consists mainly of pelagicdeposits.

On Da vie Ridge, at the northern entrance of theMozambique Channel, the lithologic record obtained at Site242 is characterized by remarkably similar sedimentsthroughout the cored interval. These deposits contain from45 to 70 percent of calcareous biogenic deposits mixedwith an important clay fraction (309^50%). Typically, thiskind of deposit is considered to be hemipelagic with thepelagic part dominant.

Further south, sediment on the bottom and flanks of theZambezi Canyon is devoid of notable pelagic contribution,products of which were either diluted or more likelywinnowed away, as shown by sediments recovered at Sites243 and 244. Similarly, the floor of the Mozambique Basin(Site 248) is covered with a thick terrigenous sequenceexcept at the very top (upper Pliocene-Pleistocene) wherepelagic sediments are dominant and at the bottom wherebrown clay overlies basalt.

The most complete and typical oceanic pelagic sequencedrilled during Leg 25 was at Site 245 in the MadagascarBasin. The whole sedimentary column consists mainly ofnanno ooze and nanno chalk with some brown clay in theupper part. Only a few detrital beds of clay and silt areinterlayered; the dominantly pelagic sequence and brownclay contain a minor detrital silt fraction. Near the bottomof the sequence, part of the nanno chalk contains clay andchert which probably originated through the devitrificationof volcanic components.

At Site 246, on the Madagascar Ridge, a relatively thick(130 m) coarse-grained sequence was cored. Foram sandand ooze accumulated during Miocene, Pliocene, andQuaternary times. These sediments, in spite of some detritalinfluence, can be considered as pure pelagic depositscapping the Madagascar Ridge.

Nanno ooze and nanno chalk comprise about two-thirdsof the whole sedimentary sequence resting upon basalt onthe Mozambique Ridge. The upper part (32 m thick)consists mainly of pure foram-rich nanno ooze whereas thelower part (down to 172 m) contains 10 to 20 percentdetrital clay. Below the hiatus that occurs between middleMiocene and Upper Cretaceous sediments, the majorlithology is foram-bearing clay-rich nanno chalk whichshows a general increase in clay content when compared tothe upper part.

On ridges, on fringes of continental margins, and inbasins, sedimentation in the central western Indian Oceanthroughout Cenozoic time is characterized by major pelagiccontributions. This fact is rather unexpected consideringthe proximity of continental land masses, i.e., East Africaand Madagascar.

Distribution of Pelagic Components

Biogenous Components

Calcareous Nannofossils

Of all the pelagic components, nannofossils are the mostabundant and main contributors to the sediment columns.Not only are they the dominant pelagic component inmany Leg 25 oozes and chalks, but also they are dominantin many sites previously drilled in the world oceans. Since

484

o -

100-

2 0 0 -

3 0 0 -

4 0 0 -

500 -

600-

7 0 0 -

SITES 2390 100%

240100%

•llz~~iBli

llililil

KSs

li

*£&&'//////////

CO OoQ .

LU

LU

oo

o

oLUoLU_ J

-

EAR

L

LEGENDΛ Λ Λ Λ Λ Λ Λ Λ Λ

ΛΛΛΛΛ BASALT

HIATUS

y LOCATION OF CORES IN HOLE

I I I BI OGEN IC COMPONENTS

H U BROWN CLAYS

r ^ n VOLCANOGENICCLAYS

I I TERRIGENOUS SEDIMENTS

n

nwzoNonhβ

wr>ooö

L. LECLAIRE

the beginning of DSDP operations, one leg after the other,coccolithophorids and discoasters have apparently been themost important contributors (volumetrically) to oceandeposits. The contribution of these algae was recognizedmany years ago (Cayeux, 1935; Bramlette, 1958).

In the extreme western Indian Ocean, on theMozambique Ridge (Site 249), abundant nanno oozeaccumulated throughout late Miocene time. Above 100meters depth at Site 249, nannofossils comprise, on theaverage, from 60 to 70 percent of the sediment bulk; above60 meters depth their amount reaches 85 to 90 percent(Figures 2 and 3) and then constitute most of the calciumcarbonate fraction. Figure 4 clearly shows a greater contentin nannofossils mainly from early late Miocene throughmost of late late Miocene. The relative rate of biogenicsedimentation is about 18 m/m.y. which is fairly high whencompared to the absolute rate of sedimentation of 24m/m.y. During the middle Miocene, the nannofossilsedimentation rate was not higher than 5 m/m.y. Below thehiatus between Late Cretaceous and middle Miocene thatbreaks the stratigraphic column, the calcium carbonatecontent decreases abruptly (Girdley, this volume; Table 2)which is connected with a sharp decrease in the nannofossilcontent to 40 to 50 percent. Below the second hiatus,between Cenomanian and late Campanian, they nearlydisappear (Figure 4). Throughout this pelagic sequence,nannofossils are well preserved, which indicates that theCCD was far below the depositional depth.

On Davie Ridge (Site 242), coccolithophorids anddiscoasters comprise about 50 percent of the sediment bulkin the upper Eocene through Pleistocene sediments. Nannocontent increases slightly through late Miocene time whereit reaches 60 percent and then decreases upwards (40%through Plio-Quaternary) (Figures 3 and 4, Table 3). Thenannofossil sedimentation rate reached 20 to 25 m/m.y.during the late Miocene which is surprisingly high comparedto the middle Miocene (2 m/m.y.) and the rate of 2 to10-14 m/m.y. through the Pliocene and Quaternary.Preservation is not as good as at Site 249. Especially inlower Miocene deposits, discoasters are dominant andcoccolithophorids are partly dissolved. It can be inferredthat the CCD was close to the deposition depth at thattime.

Farther north (Site 241), the Neogene and Quaternarysequence, which accumulated on the fringe of the Kenyancontinental margin, contains a significant amount ofnannofossils. Through most of early and middle Miocenetimes, detrital deposits were interlayered with a calcareoussequence composed of about 40 to 60 percent nannofossils.Within the latest Miocene and Pliocene deposits, animportant detrital clay fraction is mixed with nannofossils.Starting in the upper Pliocene to the Holocene, nannos arelargely dominant (50%-70%). The total sedimentation ratefor the Quaternary is about 45 m/m.y. which implies anapparent nanno sedimentation rate of about 20 to 27m/m.y. During the Miocene it approached only 12 to 14

SITE 241

20 40 60 80 100%

100 -

200-

300-

400-

SITE 242

0 20 40 60 80

1 0 0 •

200-

300-

400

100%

. . . • . .

\ \ -– \ \ \ \ N N NL -".*-"NNNNNNNNNN V.\V.'

NNNNNNNNNX N,** J ;*;"N N N N N N N N N N N J & •:

^B

SITE 249

0 20 40 60 80 100%

too -

200-

300

400-

1 1 1 I 1 1 1 1 1

NNNNNNNNNNNNNNW\\'\V'. .'

NNNNNNNNNNNNNNNNNNNHi:-

NNNNNNNNNNNNNNNNNNNNNNN VNT .*-"NNNNNNNNNNNNNNNNNNNNNNNNN V;..NNNNNNNNNNNNNNNNNNNNNNN NJ:. ; V.WNNNNNNNNNNNNNNNNNNNNNNC. •• . ,\NNNNNNNNNNNNNNNNNNNNNNNLL;.;;NNNNNNNNNNNNNNNNNNNNNNNNNF --NNNNNNNNNNNNNNNNNNNNNNNNN T.V*NNNNNNNNNNNNNNNNNNNNNNNNN V. . .NNNNNNNNNNNNNNNNNNNNNN WS^.V*.

NNNNNNNNNNNNNNNNNNNNNNI1. •.:••:•'.•WNNNNNNNNNNNNNNNNNNNN j . .y • '.*.'•NNNNNNNNNNNNN N W N W N N N ] -'.'. y'/. '•

NNNNNNNNNNNNN W W \ \ \ \ \ \ ~^•' \'•\'•

NNNNNNNNNNNNNNNNNNNNW \* < \l l

NNNNNNNNNNNNNNNNNNNNNNNNN h *.*.

N N N N N N N N N N N N N N N N N N N N N N ^ ^ \

NNNNNNNNNNNNNNNNNNNNN^^NW] '/W\W\\\\WWN\WW\W^^, ,'.

SNNNNNNNN WNNNNNNNNNNNN^\\ ,'iNNNNNNNNNNNNNNNNNNNNN\V^Q -,;

NNNNNNNNNNNNNNNNNNNN'1-^^^^^ y

WNWXXNNNNNNNNNNNNN^^•^^^ V

LU

LU

O

s

in

O

LU

o

<

1—

LU

O

>k

o•

al

LU

1—

<

1

Q

1

BIOGENIC SILICA

FORAMINIFERA

NANNOFOSSILS

HIATUS

Figure 3. Relative percentages of three main pelagic constituents: biogenic silica, foraminifera, and nannofossils (semiquanti-tatively approximated from smear-slide analyses).

486


0 -

100 —

200 —

3 0 0 - •-

4 0 0 -

5 0 0 -

600—

700 —

SITE 249

50 100%0 —

100 —

200 —

^^^"1~j_"'~_i_C" ~

PS- i ^―—^^—

|̂ i--5=:=:=i===:=-:

L U

LU

o

o

s

CRE

ö•Ij

α_

i—

L. LECLAIRE

Sample

TABLE 2Approximate Calculations of Foraminifera, Nannofossil, and

Clay Fraction Contents in the Upper Part of Site 249

Foram Fraction (Weight %)

> lOOµ 50µ


TABLE 3Approximate Calculations of Foraminifera, Nannofossil, and

Clay Fraction Contents of Site 242 Stratigraphic Column

Sample

1,CC

2,CC

3,CC

4-6

4,CC

5,CC

6,CC

7-6

7,CC

8,CC

9,CC

10, CC

11,CC

13, CC

14, CC

16, CC

17, CC

18, CC

Foram Fraction (Weight %)

Φ > 100µ 50µ < Φ < 100µ

10.0

8.0

1.8

0.5

0.5

0.7

0.1

0.3

3.5

0.2

0.1

0.1

0.5

0.2

0.6

1.0

0.1

0.2

7.5

3.0

1.7

1.5

1.02.4

0.4

0.9

3.0

0.6

1.0

0.4

0.5

1.5

1.6

1.4

0.4

0.3

Total

17.5

11.0

3.5

2.0

1.5

3.1

0.5

1.2

5.5

0.8

1.1

0.5

1.0

1.7

2.2

1.6

0.5

0.5

Clay+ NannoFraction

82.5

89.0

96.5

98.0

98.5

97.0

99.5

98.5

93.5

99.0

99.0

99.5

99.0

98.0

98.0

98.4

99.5

99.5

Total

60.5

51.0

54.0

64.0

58.0

62.5

44.5

44.5

50.7

38.0

52.0

51.0

49.0

54.0

58.0

39.5

47.5

40.0

InferredNanno

Fraction

43.0

40.0

51.5

62.0

56.5

59.5

44.0

43.3

44.2

37.0

51.0

50.5

48.0

52.0

55.8

37.9

47.0

39.5

InferredClay

Fraction

39.5

49.0

45.0

36.0

42.0

38.5

55.5

45.2

49.3

62.0

48.0

49.0

51.0

46.0

42.2

60.5

52.5

60.0

and foraminifera, it is obvious that the foram lysocline isquite different from the nanno lysocline and must haveremained well above the CCD.

Black ferromanganoan nanno chalk of Site 245, whichrests upon the basalt, contains a significant amount offoraminifera (about l%-5%). Elsewhere, foraminiferacontents are minor and their contribution is quiterestricted.

Siliceous Components

The presence of a few radiolarians is known in the LateCretaceous clayey nanno ooze of Site 239 and spongespicules appear within the clay/carbonate-rich nanno oozeof middle Miocene age; spicules increase upward throughthe middle and late Miocene and reach 15 percent of thetotal bulk of sediment in Core 6. Radiolarians in lowquantities accompany the sponge spicules. Diatoms werenoticed in the faunal residue of Core 4 and in smear slidesof Cores 2 and 3 (Pliocene). Through late late Miocenetime, sponge spicule content clearly decreases and, as forradiolarians, is restricted to only traces within thePlio-Pleistocene nanno ooze.

Rad ooze interlayered with nanno-bearing silt-rich claycharacterizes the uppermost part of the sedimentarycolumn (Quaternary) in the Somali Basin (Site 240). InCore 1, radiolarian content is 20 to 45 percent. Diatoms arepresent in minor amounts (2%-10%).

Radiolarians in trace amounts occur in the upperMiocene sediment of Site 241 together with rare spongespicules. These radiolarians often are broken and pyritized.Radiolarian and diatom content increases in the Plioceneand Pleistocene strata (Figure 4). The Quaternary sequenceat Site 241 is mostly a rad/diatom/foram-rich or -bearing

nanno ooze. The average combined content in radiolariansand diatoms is about 5 percent.

On Davie Ridge (Site 242), siliceous pelagic componentsinclude radiolarians and sponge spicules in trace amounts.They are localized in the Quaternary deposits.

Radiolarians are relatively abundant in the brown clay(Unit III) and the volcanic silty clay (Unit II) at Site 248 inthe Mozambique Basin. Apparently, they were deposited inPaleocene and early Eocene times. They are absent in theoverlying Miocene deposits.

A few radiolarians and sponge spicules were observed inthe Neocomian strata of the Mozambique Ridge (Site 249).Their amount markedly increases in the upper Albian-lowerCenomanian sequence and then they subsequentlydisappear. Trace amounts of radiolarians also were observedat Site 245 in sieve remains, but are very minor andrestricted to certain levels.

Other Biogenous Components

Tunicate spicules were observed at Site 239 (Cores 1,2,and 3) in fairly high amounts (5%). These spicules werefound in layers that are out of the normal Tunicatebiotope. Moreover, bearing in mind that those spicules arearagonitic, they should have been dissolved (depth of about4970 m). Consequently, those spicules are very likelyredeposited, perhaps from the upper continental slope(Madagascar?).

Small yellow spheres (~200µ), hardly soluble in HC1(dolomite?), were found at Site 239 in Cores 11 and 13.They are thought to be of biogenous origin (recrystallizedpellets?). Some ostracods are disseminated here and there,at the bottom of Hole 249 for instance, but occur only intrace amounts.

489

L. LECLAIRE

Fish remains (teeth and bone fragments) werecommonly found in brown clay at Sites 245 and 239 andalso in the Eocene silty clay of the Somalian continentalslope (Site 241). Such remains were noticed also in thebrown clay resting upon basalt at Site 248.

Carbon-carbonate analyses (Girdley, this volume) showfairly high contents of organic matter in Neocomian andAlbian-Cenomanian strata on the Mozambique Ridge.Throughout 120 meters of claystone and volcanic siltstoneresting upon the basaltic rocks, organic carbon contentsaverage about 1 percent. Peak values of 1.6 percent arecommon. The two sites at low latitudes (240, 241) displaya sedimentary column that has organic carbon in notableamounts (from 0.3%-0.6% with a peak value of 2% at Site241). Some amounts as high as 2 percent typify the flyschsequence of Site 241 which was emplaced mainly in theLate Cretaceous. Origin of this organic matter is possiblyterrigenous. Apart from these deposits, it appears thatorganic carbon content increases progressively upwardsfrom late Miocene to Pleistocene through thepredominantly pelagic deposit of the uppermost part of thesedimentary column at Site 241. This observation issupported by organic carbon analysis of Site 240 sedimentsin which organic carbon content increases from 0.1 to 0.6percent from the upper Miocene to Pleistocene.

Halmeic Components

The most common halmeic components throughout thestratigraphic columns cored in the central western IndianOcean are zeolites, pyrite, and Mn/Fe oxides.

Zeolites

Two major zeolite minerals were found in Leg 25sediments using the methods of smear-slide investigationand X-ray diffraction analysis. There is a discrepancybetween the results obtained by these methods; theclinoptilolite, seen as euhedral to subhedral laths in traceamounts, is not always confirmed by subsequent X-rayanalysis. Such discrepancy has been previously reported byBerger and von Rad (1972) who state that zeolites mustoccur in significant amounts in order to be identified byX-ray diffraction.

Clinoptilolite was observed (by petrographic micro-scope) in the late Cenozoic brown clay of Site 245 and wasalso identified in Eocene nanno ooze (subunit IIIA) where2 to 3 percent occurs (Cook and Zemmels, this volume).Clinoptilolite is also an important component in thevolcanic sands (Cores 10-12) at Site 246 on the MadagascarRidge and was detected in a volcanic clay layer (Cores 12,13) at Site 248. Abundant clinoptilolite helps characterizethe third lithologic unit cored at Site 249. Amounts higherthan 10 percent are relatively frequent at this site reaching38 percent of the 2-20µ fraction in some layers.

Phillipsite is poorly distributed in sediment of thecentral western Indian Ocean. This zeolite was identified inthe Mascarene Basin (Site 239), in lithologic Unit I, whereit comprises from 6 to 15 percent of the 2-20µ fraction. OnMadagascar Ridge (Site 246), phillipsite is associated in highamounts (10 percent) with clinoptilolite, which togetherform zeolite-rich volcanic sands.

Fe-Mn Oxides

Two types of Fe-Mn oxides are distinguished by thenature of sediments in which they are found. The first typegenerally appears as nodules, micronodules, or coatings inpelagic deposits that rest well above the basaltic"basement." An example of this type occurs in lithologicUnit III of Site 242 where an iron manganese-rich chalkoccurs in burrows in which oxides either coat foraminiferatests or are finely disseminated as matrix. Another exampleis furnished at Holes 245 and 245A where brown claydisplays many black streaks, blebs, and patches which arebelieved to be produced by manganese microaggregates. InHole 245A, Core 2, Section 6, a small nodule is 1 cm indiameter. Several similar nodules were seen in the brownclays of Site 239 (in the upper part of Core 15). All in all,however, manganese nodules seem to be very rarethroughout the Leg 25 sediment columns.

The second type of Fe-Mn oxide is exhibited in the basalpart of sedimentary columns resting directly upon basaltic"basement." Amorphous agglomerates of these oxides arefinely disseminated in the sediment. A good example is theferromanganoan nanno chalk of the Madagascar Basin (Site245). Here, Fe-Mn in the matrix gives a brownish-black hueto the deposit. Another example is provided by the red andblackish-red laminations of lithologic Unit III in Hole 248(Paleocene brown clay and brown silty-bearing clay).Iron-manganese-rich sediments at Site 245 are moreextensively studied in a special chapter (Warner andGieskes, this volume).

Pyrite

Pyrite occurs in many different forms throughout thesedimentary columns: as microaggregates and irregulargrains (recorded as opaques in smear-slide analyses); asspherulites with subhedral crystals; as replacement ofsponge spicules and radiolarians; and as fillings of burrowsand foraminifera tests. Some large burrow fillings, severalcentimeters long, were found at Site 242 (Core 2). Thesefillings, under scanning electron microscope investigations,exhibit a peculiar structure (Plate 4). First, the filling is notcompletely achieved and an axial void is left. The first layerof pyrite deposited directly upon the wall of the burrow iscomposed of coalescing microaggregates, whereas pyritecarpeting the inside of the burrow filling shows subhedralto euhedral crystal structure (Plate 4, Figures 3, 4).

Pyrite occurs as sponge spicule and radiolarianreplacements in the white clay-rich nanno ooze of latemiddle and late Miocene age in the Mascarene Basin (Site239). Fillings of foraminifera and individual aggregatesconstitute trace abundances throughout the upper part ofthe sedimentary column, i.e., from middle Miocene up toQuaternary near the East African continental margin (Site241). Pyrite is mainly concentrated in the 2-10µ fraction inamounts higher than 2 percent. Other notable occurrencesof pyrite in Hole 241 are located in Cores 27 (theCretaceous flysch sequence). Throughout the upperMiocene and Quaternary deposits of the Davie Ridge (Site242), trace abundances, mainly of pyrite in the form ofburrow casts, were observed. Abundant black spots and

490


streaks of pyrite are disseminated in lithologic Unit I at thissite. The high concentrations of pyrite in Cretaceousvolcanic siltstone and claystone measured at Site 249 arestriking. In effect, peak peak values as high as 7 percentwere obtained. However, the presence of pyrite in such highamounts was more or less to be expected because of thehigh concentration in organic matter.

Terrigenous and Hemihalmeic Components

Table 4 shows that the minerals quartz, K-feldspar,Plagioclase, and kaolinite, which are mostly of terrigenousorigin, constitute the major fraction of "pelagic" brownclays. If clay minerals also are terrigenous, those depositscould be considered as detrital. However, because of theirvery low rate of sedimentation (less than 5 m/m.y.), thiskind of sediment is referred to as a pure pelagic deposit.Moreover, they generally contain an important halmeicfraction with zeolites and varieties of oxides.

A question which remains, asks if all the clay mineralsare of terrigenous origin. First, the origin ofmontmorillonite is questionable. This clay species isabundantly distributed throughout the lithologic sequences,especially when volcanic influxes are present. It appearsvery often as a product of devitrification of volcanic glass.Consequently, it is not unreasonable to expect animportant part of this montmorillonitic fraction not to beland derived but rather pelagic, resulting from theinteraction between oceanic waters and volcanic products.Palygorskite may be formed from vitreous ash viapalagonite and montmorillonite by precipitation fromMg-rich interstitial solutions or under hydrothermalinfluences (Berger and von Rad, 1972) with silica beingderived from altered silicic ash (Calvert, 1971).Furthermore, according to MacKenzie and Garrels (1966),authigenesis of chlorite and especially of illite seemspossible in oceanic waters. If land-derived clay minerals arenot diagenetically modified to a large extent when enteringoceanic waters, Griffin and Goldberg (1961) suggest thatdegraded illite and chlorite may take up K+ and Mg2+ andchange to more crystalline minerals. Then, it appears thatthe whole clay fraction cannot be considered only as aland-derived product. Part is certainly terrigenous, anotherpart is halmeic; proportions of each part generally areunknown. Finally, the clay mineral fraction is a mixture ofterrigenous and halmeic products and is considered to be anhemihalmeic component.

Volcanogenic Components

Palagonite was identified in Cores 8 and 9 of Hole 239and in subunit IIA together with devitrified volcanicash (believed to be an important contributor to brown clayresting upon the volcanic basement).

Volcanic glass occurs in trace amounts throughout theNeogene hemipelagic sequence off the East Africancontinental slope, and one devitrified volcanic ash bed wasseen in Sample 241-7, CC.

In the Madagascar Basin, numerous devitrified volcanicash layers occur in the nanno chalk sequence. In some ofthem, remains and outlines of partly dissolved volcanic glassfragments were observed.

TABLE 4Mineral Contents (percent) in Selected Brown Clays

(from DSDP X-ray Analysis)

Minerals

Quartz

K-Feldspar

Plagioclase

Kaolinite

Mica

Palygorskite

Gibbsite

Montmorillonite

Site 245a

26.0

16.9

18.6

7.2

19.3

-

1.6

10.4

Site 245Ab

20.0

3.5

17.0

6.4

18.4

26.3

1.0

7.5

Site 239C

17.1

96.4

8.2

15.1

9.6

-

-

23.7

aDepth = 8.2 meters.bDepth= 16.8 meters.cDepth = 159.1 meters.

Volcanic components are abundant in the basallithologic sequence of Site 246. The volcanic breccia andvolcanogenic reworked material that characterize thissequence cannot be considered as a pelagic deposit.Similarly, the origin of gray and olive-black volcanicclaystones which form the lower part of the sedimentarycolumn cored on the Mozambique Ridge (Site 249) isquestionable. Because of the predominant quantity ofpyroclastic material, these deposits are referred to asvolcanogenic deposits (Vallier, this volume).

Preliminary Conclusions on Distribution of PelagicSediments

Pelagic Sediments Compared with Terrigenousand Volcanic Sediments Through Time

It is striking to find out from the foregoing discussionthat all Leg 25 sedimentary columns recovered are cappedpredominantly with relatively pure pelagic deposits. This isindependent of site location.

It appears that throughout late Cenozoic time, pelagicsupplies, when compared to terrigenous ones, progressivelyincrease or definitely dominate, while high sedimentationrates of biogenous deposits were reached. In the SomaliBasin, where detrital and terrigenous sediments aredominant during most of the Late Cretaceous-Paleogeneinterval, pelagic input occurs through Quaternary time atSite 240 and pelagic sedimentation products are abundantfrom early Miocene to Quaternary at Site 241. Even in theMozambique Basin where the Zambezi Canyon debouchesabundant detrital sediments, nanno ooze occurs in thePlio-Pleistocene layers.

Basaltic rocks were reached five times; except at Site249, a relatively thick pelagic sedimentary sequence wasobserved resting directly upon or very close to the basaltic"basement."

Both the Mozambique and Madagascar ridges (Sites 249and 246) are capped with biogenous deposits which overlievolcanoclastic sediments.

Hiatuses mainly occur in pelagic deposits. Either theycut through a predominantly pelagic sedimentary sequenceas at Site 245 or they separate two different lithologies, one

491

L. LECLAIRE

of which is a pelagic deposit (Sites 241, 246, 249).Consequently, it can be inferred that hiatuses areassociated, at least in part, with great changes insedimentation patterns.

Variation of Biogenic Supplies with Time

When compared to nannofossil contents, foraminiferaclearly increase through Cenozoic time and especiallythrough the late Miocene, Pliocene, and Quaternary. Thisphenomenon might be in connection with the increasedpelagic sedimentation during these same times. Morover,siliceous biogenous components, when present in Cenozoicdeposits (Sites 239, 240, 241, and 242), occur chiefly inupper Miocene sequences and then their content increasesthrough Pliocene time (except at Site 239). Notableamounts of sponge spicules and radiolarians were foundwithin the Lower Cretaceous deposits at Site 249, but theydefinitely disappear through middle Cretaceous time andnever occur again in the overlying Cenozoic deposits.

Similarily, contents in organic matter, when present inTertiary deposits, increase through late Cenozoic time(Sites 240 and 241).

Consequently, the first preliminary conclusion that canbe safely drawn is the apparent increasing productivity ofcentral western Indian Ocean waters through late Cenozoictime.

MAIN PELAGIC FACIES-DEPOSITIONALENVIRONMENTS

Nanno Ooze and Chalk

Paleogene Nanno Ooze and Chalk of Madagascar Basin

This calcareous facies, 264 meters thick, constitutes themain sedimentary sequence of Site 245 (see Chapter 7).The average sedimentation rate mostly ranges from 17 to20 m/m.y. but reaches 30 to 35 m/m.y. in the lowerEocene strata. Average as well as peak values are highcompared to other sedimentation rates in regions ofreputed high productivity such as equatorial Pacific beltwhere 15 to 17 m/m.y. were measured for Miocene deposits(Tracey, Sutton, et al., 1970).

X-ray bulk analysis, calcium carbonate content, and HC1residue determinations point to a very low proportion ofquartz and K-feldspar, i.e., to a very weak terrigenousinflux. The main contributor besides nannofossils andforaminifera is undoubtedly montmorillonite. As discussedpreviously, partly dissolved volcanic glass and shard outlineswithin montmorillonite beds testify in favor of a volcanicorigin for this clay mineral. Other volcanic influence issuggested by the presence of clinoptilolite for, according toBiscaye (1965), this mineral is considered to be a volcanicassociated zeolite. Today, kaolinite and gibbsite (fromMadagascar laterite) constitute only a small absolutepercentage of the total sediment around Madagascar. Theprincipal mineral is montmorillonite which is believed to bethe result of the in situ formation from volcanic products inthe Madagascar and Mascarene basins. It is not unreasonableto think that Plagioclase as well might be a volcanicderivative. Consequently, it can be assumed that either theerosional activity was very weak in Madagascar during

Paleocene and early Eocene time, the site was very far fromMadagascar, or the site was isolated from terrigenoussedimentation.

Therefore, pelagic sedimentation, mostly biogenic, waslargely predominant at this site. Because of the high rate ofsedimentation, one could expect that the depositional areawas in a high productivity zone such as an equatorialdivergence during the early Cenozoic. Hays, Saito, et al.(1969); Tracey, Sutton, et al. (1971); van Andel, Heath, etal. (1971); Heezen, MacGregor, et al. (1972); and manyothers extensively studied deposits below the Pacificequatorial belt and Olausson et al. (1971) described thesedimentation below the Somali current. An importantcharacteristic of the sedimentation in high productivityzones can be inferred from those works. Generally, in thesezones, pelagic deposits are not pure nanno ooze but amixture of nannofossils and abundant biogenic silica(diatoms, radiolarians) with, in addition, abundant(sometimes predominant) foraminifera. This is quitedifferent at Site 245 even though the basal part of thesedimentary column bears a certain quantity offoraminifera (l%-5%). Possibly foraminifera and radio-larians might have been dissolved although signs offoraminifera dissolution were not clearly seen. Generally,planktonic foraminifera, when present, are diversified andin an excellent state of preservation; some re crystallizedshells and nannofossils, however, were seen in the lowermiddle Eocene ooze. Moreover, dissolution below highproductivity zone occurs at great depths because the foramlysocline and the CCD are depressed (more than 5000 m iscommon), but Site 245 area deepened only from 2700 to4000 meters which is not deep enough for abundantdissolution. Consequently, the absence or rare occurrencesof foraminifera and radiolarians at Site 245 can hardly beascribed to dissolution. More likely, this phenomenonmight reflect sedimentation below a rather low productivitywater mass.

As suggested by Olausson (1961), coccolithophoridsdominate in low productivity areas while their relativeimportance, on the whole, increases when going from amore fertile to a less fertile region. This idea is supportedby Lisitzin (1971a) and Uschakova (1971) who concludethat areas of purest carbonate accumulation are not in thegreatest production zones. Thus, according to Berger andvon Rad (1972), the high nannofossil proportionsassociated with low foraminiferal content can be consideredas characteristic of deposits below waters of rather lowfertility. This sort of plankton impoverished waters isbelieved to be located in the "arid" zones of the oceans'central water masses which are surrounded by currentgyrals (Be and Tolderlund, 1971), e.g., the central watermasses of the South Atlantic. Usually, those waters arefound today between 15° to 35° latitude and roughlycorrespond to tropical-subtropical belts.

High rates of sedimentation and the associated very lowforam/nannofossil ratio might indicate sorting by bottomcurrents. The great mobility of nannofossils compared toforaminifera is indicated, for instance, by the frequentcontamination of young deposits by older nannofossils. It isvery likely that the abnormal rate of sedimentationmeasured at Site 245 for the lower Eocene ooze results

492


from the reworking of older nannofossils. Such aredeposition would be greatly favored by a surroundingrough topography which is characterized by steep slopeswith resultant unstable deposits.

Thus, it can be tentatively concluded that nanno oozeand chalk of Site 245 were deposited in a somewhatisolated basin below oceanic waters very similar to presenttropical water masses. Isolation of the site area is furthersupported by results from foraminifera studies which pointto a special environment with regard to the well-knownequatorial to tropical faunal assemblages and to classicregions such as the Caribbean and Mediterranean seas andNew Zealand.

Nanno ooze deposition ended abruptly in Core 245-3and is replaced by brown clay mixed with silty clay andinterlayered nanno ooze. Most of the younger lithologicsequence displays some weak turbidite influences.Abundances of land-derived materials such as quartz,K-feldspar, kaolinite, chlorite, hematite, and epidoteindicate considerable terrigenous influxes. Most of theinterlayered nanno ooze beds are redeposited as evidenceby reworked benthonic foraminifera assemblages whichcomprise mixed forms coming from medium shelf depthsand displaced into a bathyal association. Similar deposits ofpelagic turbidites close to or below the CCD has been notedby Bezrukov (1971) in the Indian Ocean. StudyingQuaternary deposits, he observed, at the foot of submarineridges, carbonate sand and ooze down to depths of5500-5600 meters, i.e., considerably deeper than thecritical depth (CCD). '..",..,

In summary, by middle Eocene time, the Site 245 areawas apparently deepening faster than before andsubsequently reached the CCD. At the same time, thephysiographic environment of Site 245 changed. Theheretofore closed area was opened and subsequentlyinvaded by detrital land-derived materials while faunalassemblages were renewed.

Other Paleogene Nanno Ooze and Chalk

In the Mascarene Basin, some clayey nanno ooze ornanno clay layers were found at the bottom of Site 239,resting directly on the basaltic "basement." These minornannofossil facies, among the major facies of brown clay,accumulated in Late Cretaceous and early Paleocene times.They contain a peculiar foraminiferal assemblage which hasbenthic species belonging to the middle or lower part ofcontinental slopes while pelagic forms are absent. It seemsvery likely that such layers were deposited from turbiditycurrents.

Lower Paleocene nanno ooze displays partly dissolvednannofossils and chiefly agglutinant benthic foraminifera.The poverty of this microfauna suggests a bathyal depthwhich was below the foram lysocline and close to the CCD.Compared to the lower Cenozoic sediment sequence of theMadagascar Basin (Site 245) which is typical ofsedimentation over a spreading ridge crest and flank,deposits resting upon the basaltic "basement" at Site 239appear to be somewhat different.

In the Somali Basin, at Site 240, the basal lithologicsequence includes minor nanno ooze and calcareous faciesformed during early Eocene and contain mainly displaced

fauna. These layers are believed to be mostly pelagicturbidites. However, fauna are rich, and well-preservedplanktonic assemblages generally contain guide species.Benthonic species show little diversification and majorspecies apparently represent an impoverished bathyalassemblage.

Miocene Nanno Ooze of Mozambique Ridge

The composition of nanno ooze at Site 249 is illustratedin Figures 2 and 4. The rate of sedimentation during lateMiocene was about 25 m/m.y., which is comparable tosedimentation in the Madagascar Basin through the Eocene.The very minor noncalcareous fraction is mainly clayminerals (montmorillonite). It appears that terrigenoussedimentation was very weak on this ridge through lateMiocene. Consequently, consideration of the sedimentationrate, foraminifera/nannofossils ratio, and foraminiferaassemblages yield significant data about paleoenvironmentalconditions.

First the occurrences of all stratigraphically importantspecies of Miocene planktonic foraminifera andnannofossils indicate that a general worldwide distributionof water masses was present. Faunal assemblage variationsseem to be related only to glacial-induced climaticfluctuations which began as early as early Miocene (Hayes,Frakes, et al., 1973). High content of foraminifera mayindicate both the presence of a high productivity zone andthe winnowing of nannofossils by deep-sea currents.However, the high rate of sedimentation implies rather highproductivity conditions in the overlying water. Presently,higher foraminiferal abundances are found in the majorcurrent gyrals that surround the central water massesdiscussed previously. As an example, Olausson et al. (1971)points to a close connection between Somali Currentmovements and foraminifera concentrations through latePleistocene in sediments of the Somali Basin. Theyconcluded that foram ooze beds are related to theoccurrence of high nutrient levels in the overlying watermasses. One can assume that similar conditions were actingover the Mozambique Ridge by late Miocene times.

Figure 3 shows two striking facts: (1) foraminiferacontent is highest in middle Miocene deposits where thesedimentation rate is lowest (10 m/m.y.) and (2)foraminifera content is the lowest in late Miocene oozewhere the sedimentation rate is the highest (25 m/m.y.).These sharp changes in the foraminifera/nannofossil ratiothrough middle and late Miocene strongly suggest theeffects of current action. It seems very likely that in thefirst case, currents winnowed out the nannofossils, therebyconcentrating foraminifera, whereas in the second case,there was redeposition of previously resuspended nannofos-sils. There currents were rather weak but presumablybecame stronger by Pliocene-Quaternary times which mightaccount for the condensed Pliocene—Quaternary section atthe top of the sedimentary column of Site 249. Currentsare probably still acting, since, according to Saito and Fray(1964), two cores raised from the Mozambique Ridgecontained a mixed assemblage including Late Cretaceous,Eocene, Miocene, and Recent foraminifera. They believethat this sediment was derived from submarine erosion ofsediments cropping out at higher elevations on the

493

L. LECLAIRE

Mozambique Ridge. New data from Harris (1972) stronglysupport this point and will be discussed later.

Late Cenozoic Foram Sand of Madagascar Ridge

Despite the fact that intermittent coring and poorrecovery of the upper part of the sedimentary column atSite 246 did not allow proper studies of the sedimentcomposition, paleoenvironmental reconstruction can beattempted. Conspicuous at this site is the great quantity ofplanktonic foraminifera which constitutes more than 60percent of the sediment bulk on the average. It is possiblethat technical difficulties encountered while drilling thisdeposit came from the sandy texture of the sediment andthe washing may have contributed to further enrichment inforaminifera; however, it seems very likely that the foramsand is the result of original concentration duringsedimentation. Approximate sedimentation rate is about5-6 m/m.y.; it might have been higher during Pliocene andQuaternary. Such a foraminifera accumulation obviouslyimplies both higher productivity and stronger winnowing ofnannofossils than at Site 249. Possibly, currents overMozambique and Madagascar ridges belonged to the samesystem.

Additionally, foraminifera assemblage studies lead to thesupposition that during early Miocene time, the Site 246area was at shallow depths and was affected by atransgression which thereby reworked older sediments (seeChapter 8). Afterwards, during the middle and late Mioceneand Pliocene, the sea deepened and a rich planktonic faunawas deposited. Through the Pleistocene a considerableamount of benthonic fauna from lesser depths was added tothe planktonic fauna, perhaps indicating uplifts in thevicinity. Since the early Miocene transgression, the site areaseems to have undergone strong bottom current action.

Cenozoic Clayey Nanno Ooze and Chalk of theDavie Ridge

Figure 4 shows the importance of the insoluble fractionsin the calcareous sediments of Site 242. The first questionis to know whether this insoluble fraction is mostly pelagicor terrigenous in origin. X-ray analyses (Cook and Zemmels,this volume) yield sufficient data to determine theland-derived origin of most insoluble residues. In effect,reasoning on a carbonate-free basis, quartz and K-feldsparcontents appear to be very high, constituting about 20 to30 percent of the minerals. Moreover, mica contents rangefrom 20 to 40 percent and kaolinite is present also. Thecarbonate-free fraction of this facies appears quite similarto that at Site 241. Another significant point is theconcentration of quartz and K-feldspar in the 2-20µfraction which means that this mineral suite constitutes afine silty fraction. Therefore, this facies can no longer beattributed to true pelagic deposits but has to be referred toas a hemipelagite.

Sedimentation apparently was uninterrupted from lateEocene to Quaternary time at Site 242, which presents anopportunity to trace all the relative variations of pelagicand terrigenous contributions through this time span. First,the upper Eocene and lower Oligocene sequences are thinand, with a sedimentation rate of 5 to 7 m/m.y., are verylikely condensed. Heavy minerals and reworked

foraminifera from lower and middle Eocene strata supportthis point. Displaced fauna are often abundant throughouta large portion of ths sediment column up to Core 16 butare much more disseminated above that level. Lithologystudies show cyclic or oscillatory patterns which appear asincreasing and decreasing insoluble residue contents. Thismay be explained by irregular winnowing and depositionrelated to fairly strong bottom currents. Upwards, thesilt-sized fraction decreases, but the whole terrigenousfraction increases through late Oligocene and early Miocenetimes when the sedimentation rate reaches 15 to 18 m/m.y.This concomitant increase in both the rate of sedimentationand the terrigenous fraction suggests an increasingland-derived supply. It is striking to note that augmentationof terrestrial influx coincides with hiatuses in the SomaliBasin.

The upper Miocene sequence exhibits a thickhemipelagic accumulation,implying an abnormally high rateof sedimentation, about 30 to 40 m/m.y. This high ratecoincides with not only an increase in nannofossilcontributions by one-fifth (Figure 4), but also anaugmentation of terrigenous mud supplies. It is very likelythat this hemipelagic material was emplaced by mechanismslike bottom currents which are able to transport bothnannofossils and terrigenous clays, forming nepheloid layersalong the sea floor (Eittreim et al., 1969).

Thus, the lithology of the eastern flank of Davie Ridgeinvolves the existence of low-energy currents over an areaunaffected by tectonic activity, thereby allowing acontinuous accumulation of sediment without any apparenthiatus.

The Oligocene clayey nanno chalk of Davie Ridge offersthe unique possibility to link early Cenozoic nannofossils'depositional history of the Mascarene Basin to that of lateCenozoic nanno ooze in the Mozambique and Madagascarbasins which have Oligocene hiatuses. This facies shouldpresent an opportunity to trace the evolution ofnannofossils through Cenozoic times in the central westernIndian Ocean. Although it is difficult to approximate thephytoplankton production from lithologic data because ofdissolution and reworking, for instance, a general feelingemerges from the foregoing descriptions. It seems that adecrease in production occurred in the Oligocene and thatmicroflora deposition was considerably reduced from theEocene level. Similarly, the relatively increased Mioceneproduction was again followed by a decline to the present.This fits very well with the conclusions of Tappan (1968)who attributes those fluctuations to contemporaneousphysiographic modifications assuming that extensiveepicontinental seas were at first favorable to phytoplanktongrowth.

Biogenic Silica-Rich Facies

The presence of high concentrations of radiolarians,diatoms, and even sponge spicules in a pelagic deposit areinteresting because they are believed to indicate highnutrient level, i.e., to point to high productivity areas asdiscussed above. According to Lisitzin (1971b), more than70 percent of the silica presently in suspension consists ofdiatoms, but in the equatorial belt radiolarians increase inimportance. In addition, unlike calcium carbonate,

494


apparently no critical depth (CCD) exists for amorphoussilica because siliceous sediments are encountered at alldepths.

Occurrences of Radiolaria and diatoms in the upperMiocene deposits of the Somali Basin are significant (Site241). Increasing productivity, reaching a high level duringthe Pleistocene, is confirmed by the Site 240 lithologicrecord. Further, a few radiolarians and sponge spicules areseen in trace amounts in the Pleistocene deposits of theDavie Ridge. Consequently, one can assume that thepresent equatorial circulation pattern of the central westernIndian Ocean was initiated by late Miocene times and wasgradually enhanced during Pliocene.

The Neogene lithologic sequence of Site 239 apparentlycontradicts the foregoing, for radiolarians do occur in theMiocene, but subsequently they disappear in the Pliocene.These biogenic silica remains are included in a veryparticular facies which might have resulted from reworkingand redeposition of sediments further north. Signs ofreworking are indicated by the presence of displacedbenthic foraminifera.

At other sites, especially in the Neocomian sequence ofthe Mozambique Ridge (Site 249), radiolarians seem to beclosely associated with local volcanism.

Brown Clay

The mineral composition of brown clay has previouslybeen discussed. Brown clay is the end member in a series ofpelagic deposits which accumulate in a particularenvironment characterized by (1) a strong dissolution ofcalcium carbonate, and (2) a low rate of sedimentation ofterrigenous and volcanogenic components, generallyrestricted to clay-sized material. This means that brownclay is deposited below the calcium compensation depthand away from highly productive areas. The CCD is,therefore, the depth of a facies boundary betweencalcareous ooze and brown clay which greatly dependsupon sea-floor topography, deep bottom water circulation,and seawater chemical equilibrium. Thus, a brown clayfacies is a good indicator in paleoenvironmentalreconstructions.

Brown clay forms the end member of the Site 245lithologic sequence which is a good example of asedimentary column resting upon a spreading ridge belowmid-latitudes and is similar to sedimentary columns inmany other DSDP boreholes. At the bottom of Core 245-1,a late middle Miocene age was obtained from anintercalated clayey nanno ooze bed. Possibly, this layer wascaused by the redeposition of nannofossils resuspendedfrom neighboring upper slopes; a Paleocene assemblage wasalso observed in the same core. But, assuming that the timeinterval between initial deposition and redeposition for theyoungest material is short, in other words negligible, thesubsequent computed rate of sedimentation is about 1.5m/m.y. This value is low but values of similar magnitudepreviously have been found, for example 2.5 m/m.y.(Berger and von Rad, 1972). The lower boundary of theSite 245 brown clay facies was placed at 63 meters whereintercalated nanno ooze gives an age of about 40 m.y., i.e.,a total rate of sedimentation of 1.5 m/m.y. This value doesnot warrant a break in the sedimentary column as a hiatus

encompassing an Oligocene-early Miocene interval. Rather,the brown clay interval might be without any hiatus; morelikely, it is condensed. The abnormally high amount ofsilt-sized material, which comprises about 23 to 30 percent,could favor such a hypothesis. Instead of denoting strongterrigenous influx, these high contents in silt-sized materialcould have resulted from the winnowing of the clayfraction. Removal of much of the clay-sized material wouldresult in a concentration of the silt fraction and a loweringof the computed sedimentation rate.

The boundary between Site 245 brown clay and theunderlying lithologic unit is well marked with the upwarddisappearance of intercalated thick nanno ooze beds. Thus,by middle late to late late Eocene, Site 245 was definitivelybelow the CCD, which means either a subsidence of thearea or, less simply, subsidence plus a rise of the CCD.

At Site 239, in the Mascarene Basin, brown clay andbrown silty clay characterize the Paleogene series.Throughout the Paleocene, oozes were intercalated with theclay. The average sedimentation rate, if that hiatus exists,would range from 4 to 5 m/m.y. Terrigenous influx wasimportant, especially during late Paleogene but beds of purebrown clay with very low silt-sized material content alsooccur.

The brown clay at the bottom of Site 248 is not quitesimilar. Presence of hematite and goethite attests toiron oxide influx, the source of which could be theunderlying basalt. This facies is rather similar toiron-enriched basal sediments like those presently formingat the crest of the East Pacific Rise (Cronan et al., 1972).

Minor Facies

Organic Carbon-Rich and Pyritic Facies

At Site 241, concomitant occurrences of organic carbonand pyrite occur in noticeable amounts through middleMiocene to Pleistocene times. Evidently, one main factorresponsible for such simultaneous occurrences is a high rateof sedimentation, about 35 m/m.y. Frequently, in areassituated below high sedimentation rate zones, organicmatter is rapidly buried, thereby favoring the developmentof sulfate-reducing bacteria which in turn initiates areducing environment favoring organic matter preservationand pyrite formation. Another important factor must bethe high organic matter productivity of water massesoverlying the same area which provides organic carbon tothe sediment. Presence of oxygen-depleted bottom water isnot necessary to initiate reducing conditions as long as therate of sedimentation and organic matter productionremain at a high level. This authigenesis of pyrite connectedwith organic matter production supports the hypothesis ofan increasing productivity of Somali Basin waters throughlate Cenozoic time.

The sedimentary column recorded from the east flank ofthe Davie Ridge is characterized by an apparently minordifference between lithologic Units II and III. The lowerone (Unit III) displays brown hues whereas the overlyingUnits II and I are gray colored. These two gray units bearpyrite in trace amounts (burrow casts, streaks, etc.) and thegray hue is certainly closely related to the presence of ironsulfides. The color change occurs in the late early Miocene

495

L. LECLAIRE

and very likely marks a considerable modification in redoxconditions at that moment. Although the sedimentationrate is fairly high during late Cenozoic times, supplies oforganic matter were certainly low as evidenced by lowvalues of organic carbon contents (0.1^0.2% with the 0.2%significantly located in the Pliocene-Pleistocene sequences).Consequently, a more or less oxygen-depleted environmentseems to be necessary and would have occurred by lateearly Miocene time and thence prevail until the latePleistocene. According to Sverdrup et al. (1942),oxygen-depleted water flowing from the Red Sea and theGulf of Aden presently passes over the Site 242 area.Oxygen minima are located at depths of about 1300meters, but low oxygen contents extend down to about2000-2200 meters (2.8-3.2 ml/L). Such oxygen-depletedwaters allow initiation of reducing conditions at thesediment/water interface, especially when the sedimentsurface is heavily burrowed by bottom dwellers as was thecase there. However, the top of Hole 242 is brown andapparently devoid of pyrite, which would mean thatpresent conditions (during Holocene) do not initiate pyriteformation. It is very likely that the redox front movesthrough time according to climatic fluctuations, forinstance, and presently is staying only for a short time justbelow the sediment/seawater interface. Presumably, anincreasing flow of Antarctic Intermediate Waters wouldlower the depth of the oxygen-depleted, more saline waterand so favor the occurrence of reducing conditions at thesediment surface (Sverdrup et al., 1942).

Thus, it seems reasonable to propose that a change inwestern water masses' circulation occurred by late earlyMiocene, which is marked by the color change in thesedimentary column of Site 242 described previously. Thischange would depend most probably on the arrival, at theentrance of the Mozambique Channel, of oxygen-depletedhypersaline deep water from the opening Red Sea and Gulfof Aden and also from the Arabian Sea. Presently thosenorth Indian deep waters flowing south in the westernIndian Ocean, between 1000 and 2000 meters depth, areknown as the second principal source of low-oxygenatedwater which forms the extensive oxygen minimum in theupper circumpolar deep water (Callahan, 1972).

Other organic, carbon-rich pyritic facies like Cretaceousvolcanic siltstone and claystone of the Mozambique Ridge(Site 249) seem related to local conditions characterized byhigh volcanic influxes associated with high benthonicproductivity. The organic matter found in the Cretaceousflysch sequence of Site 241 might mainly originate fromland-derived organic matter that was deposited inoxygen-deficient water.

Black Ferromanganoan Facies

This particular facies is described in detail elsewhere(Warner and Gieskes, this volume). Consequently, onlyenvironmental implications are discussed here and arerestricted to Site 245. At Site 245 the carbonate-freefraction of the black ferromanganoan clayey nanno chalk isessentially montmorillonite with a brown to olive chalkmatrix of amorphous matter containing mainly iron andmanganese oxides. Very likely, the high amount ofmontmorillonite is related to volcanic ash devitrification

processes similar to other montmorillonite accumulations.The importance of volcanic influences in the formation ofthis type of sediment is suspected by many authors(Bostrom and Peterson, 1969; Bostrom, Peterson, et al.,1969) who showed that low Al/Al+Mn+Fe ratios are foundin sediments resting on active ridges where a source shouldexist for volcanic emanations mixing with bottom waters asthey debouch. The very close geographic proximitybetween active ridges and the low Al/Al+Mn+Fe ratios inpelagic sediments has been well demonstrated. According tovon der Borch et al. (1971b), the reasoning thatferromanganoan deposits are related to ridge crestvolcanism is supported by the fact that iron-rich facies arepresent as a basal facies overlying basalt at all sites where ithas been encountered. In effect, this basal sediment is quitesimilar to the surface deposit settling down over the crest ofthe East Pacific Rise. Thus, at Site 245, the basal iron-richfacies probably was formed when the basement was at thecrest of the rise. The ensuing gradual removal of the sitearea from the ridge is well marked at Site 245 for the highmontmorillonite content decreases through time. Strongvolcanic influences are believed to last 12 m.y. at leastwhich possibly means that volcanic ash might have beenprojected far away from the ridge crest. Several hypotheseswere proposed to account for origin of iron-rich depositssuch as submarine hydrothermal exhalations (Bostrom andPeterson, 1969) and leaching of nearly extruded basalt(Cronan et al., 1972).

The striking fact at Site 245 is that the ferromanganoanmatrix, montmorillonitic phase, and nannofossils seem tobe syngenetic. Fe/Mn deposits do not appear as coatingsbut as amorphous agglomerates of very fine grains (Plate 3,Figures 5 and 6) formed and deposited nearlysimultaneously with nannofossils. This indicates a very fastformation of Fe/Mn oxides. Taxieff (1972) noticed thatdeep-water eruptions do not produce large pyroclasts suchas bombs, lapilli, or pumice because hot lava ejecta isprogressively fragmented during travel through water intosilt-sized volcanic glass (hyaloclasts) and subsequentlydispersed. This high fragmentation brings a large surfacearea of highly reactive material in contact with water whichgreatly favors chemical reactions. One of them would behalmyrolysis of palagonitic material with release ofinsoluble Fe-Mn oxide, calcium, and silica. Such reactionsshould be very fast and might produce a sort of rain ofinsoluble Fe-Mn particles which slowly settle whileaccreting more Fe-Mn ions and mixing with nannofossils.Simultaneously, a calcium-enriched environment wouldfavor the precipitation of calcite as rhombs or asovergrowths.

Evidently, apart from any speculations, ferromanganoandeposits characterize, to some extent, the most active ridgesand according to Bostrom et al. (1969), those with highheat flow and associated high spreading rate. The ferroman-ganoan content in the basal sediments at Site 245 implies asimilar origin.

Silicified Chalk and Chert-Lithification of Chalk

Chert and silicified chalk were found only in theMadagascar Basin, at Site 245 in subunit IIIB (see Chapter7). Problems of chert formation and origin have been

496


previously well studied mainly by Rex (1970), Rex andMurray (1970), Calvert (1971), Mattson and Pessagno(1971), von der Borch et al. (1971a), Berger and von Rad(1972), Wise et al. (1972), and Weaver and Wise (1972).The purpose of this section is not to review these worksbut to emphasize one special point characterizing Leg 25chert. The late Paleocene and early Eocene chert-bearinglithologic unit is practically devoid of any siliceous tests butis rich in montmorillonite. This point could support thecontention of Gibson and Towe (1971) that volcanicderivatives, rather than purely biogenous production, maybe an important source of silica for chert.

Fresh broken surfaces of chert and mainly of a silicifedchalk directly in contact with a brown chert bed aredepicted from scanning electron microscope pictures,shown in Plates 1 and 2. Spherulites consisting of thinblade-like crystals, radiating in disorder out of the spheres,are seen filling foraminifera test cavities. These spherulitesand agglomerates of spheres are identical to those describedby Wise and Hsu (1971) and identified as cristobalite.Further, X-ray analysis of a silicified chalk (Core 9, 312 m)denotes the presence of predominant cristobalite withmontmorillonite and tridgmite. In Plate 1, Figures 1 and 2,molds of recrystallized calcite crystals of a foraminiferatest, removed by the fraction section, are clearly visible, aswell as the cristobalite blades. It appears that in some placesspheres are connected both to each other and to theforaminiferal chamber wall by a network of thin filaments(Plate 1, Figures 5 and 6; Plate 2, Figure 1). One figureseems to be typical of the boundary between silicified chalk(with spherulites) and massive chert, where all featuresdisappear except for ghosts of foraminifera (Plate 2, Figure5). Spherulites are often replaced by sheaves of cristobaliteblades more or less interlaced (Plate 2, Figures 2, 3,4) andare seen either as fillings of foraminifera tests or in pores ofthe chalk.

What appears in a conspicuous manner is thewell-marked difference between silicified chalk and the truechert stringer. Within the former, nannofossils, authigeniccalcite rhombs, overgrown test fragments, and cristobalitespherulites are clearly visible. Within the latter, all thosefeatures disappear and are replaced by mainly homogeneousmaterial where ghosts of fossils can be seen. Obviously,there are two different steps in the silicification. Possibly,the occurrence of a filament network indicates thebeginning of the filling of foraminiferal chambers, the firststep in "chertification."

The origin of silica remains speculative. An absence ofsiliceous fossils—unless they have all been dissolved which isdifficult to believe—implies another source. High content inmontmorillonite of the chalk and abundance of devitrifiedvolcanic ash in the underlying lithologic unit suggest that apredominant part of silica comes from the devitrification ofvolcanic glass. According to Gibson and Towe (1971), theopal-cristobalite-montmorillonite suite is indicative ofaltered volcanic glass. Silica could pass into solution ininterstitial waters and migrate over distances of centimetersor meters (Weaver and Wise, 1972) and even more until ahigh-porosity layer, a foram-rich layer for instance, isreached. There, mobilized silica might reprecipitate ascristobalite blades with prisms of foraminifera tests actingas a catalyst and support.

So, independently of chert, silica and especiallyvolcanically derived silica might play an important part inthe lithification of nanno ooze. As assessed by Wise andHsu (1971), lithification of deep-sea chalk could have beenaccomplished by chemical precipitation of two types ofcement: (1) cristobalite as spherules or sheaves filling largepores; and (2) calcite occurring as secondary overgrowthsand euhedral crystals. From scanning electron microscopeexaminations, it clearly appears that what is designated as"micarb" is composed of overgrown nannofossil plates, butnot always in major amounts. Micarb could also be calciteprisms of desegregated foraminifera tests (near thelysocline?) more or less recrystallized or overgrown.

DEPOSITIONAL PALEOENVIRONMENT-CHANGES THROUGH THE CENOZOIC

Late Cenozoic Changes in Productivity at the Onset of theEquatorial Water Mass Circulation

As discussed previously, increasing content in siliceousfossil, foraminifera, and organic carbon contents in bottomdeposits indicates increasing biogenic production of theoverlying water masses. Presently, the equatorial zone isknown to be one of the regions of high organic productionwhich is related to up welling induced by divergences.

According to Olausson (1971), in the three majoroceans, the south equatorial divergence apparently is themost important because the rate of production is high.Today, it lies at about 8° to 12°S in the central westernIndian Ocean. Despite the fact that it is still not very welldocumented, one can expect that upwellings are notsteadily developed and hence not quite comparable to thosein the Atlantic and Pacific oceans. In effect, the equatorialIndian Ocean circulation pattern is strongly altered by asemiannually reversing monsoon regime in relation to thevicinity of continental masses, particularly Southeast Asiaand the Indian subcontinent. Consequently, whereas theeast-west-flowing south equatorial current is permanentbecause the Southeast Trade Winds are permanent, thenorth equatorial gyre is strongly modified in winter by thesouthwest monsoonal wind which induces the SomaliCurrent. If the location of India is presently very importantregarding the development of the monsoonal regime in thecentral western Indian Ocean, its paleoposition is evidentlya major point too. As long as India was at south latitudes orbelow the equator, the presence of an equatorial circulationsystem was not allowed to develop. According to theMcKenzie and Sclater (1971) reconstruction, the south partof India and Ceylon were still south of the equator byOligocene times (Figure 5). It is most likely that anacceleration of spreading in the north central Indian Oceanby early to mid-Miocene times led to a northward drift ofIndia. This movement might be correlated to the uplift ofthe Himalayas (Laughton, McKenzie, and Sclater, 1972).The first signs of equatorial biogenic production are seen inupper Miocene deposits of the Somali Basin. Consequently,it can be inferred that the onset of the equatorialcirculation pattern was initiated as soon as India wassufficiently beyond the equator. India is believed to havecrossed the equator in the Miocene (McElhinny, 1970).

497

L. LECLAIRE

LEGEND

A SOUTH AFRICAB MOZAMBIQUEC TANZUNIAD KENYAE HORN OF AFRICAF, BROKEN RIDGEF

2KERGUELEN

G CROZETH SEYCHELLESJ NINETYEAST RIDGEK CARLSBERG RIDGEL MASCARENE RIDGEM MALDIVE-LACADIVE ISLANDSN COMORO ISLANDS

• HELOSIEoVAUCLUSEx EPONGEα MAJUNGA BASINb MORANDAVA BASINd LOURENCO MARQUESg ZAMBEZI RIVERi DURBAN

s ZANZIBAR ISLANDt PEMBA ISLANDuMAFIA ISLANDwBEIRAMAGNETIC ANOMALIES

A LEG 25 SITESxxx STATE BOUNDARIES

Figure 5. Early Oligocene reconstruction of the Indian Ocean (36 m.y.) (according to McKenzie and Sclater, 1971).

Simultaneously, Gulf of Aden was created and theArabian Sea widened as Carlsberg Ridge was spreading, anddepending upon climatic conditions, warm and salinewaters probably were formed in these seas with a lowoxygen content. Initiation of their circulation (to thesouth) would have taken place by early middle Miocene andmight be the cause of the redox change over the DavieRidge area. It has been noted previously that these

oxygen-depleted waters are known presently as the northIndian deep waters which still flow southward. Thiscirculation through the Mozambique Channel seems to beblocked today at 20°S. Tchernia et al. (1951) pointed outthat the influence of Arabian waters is limited to 20° S inthe Mozambique Channel because of the hydrological andthermal barrier which obstructs the propagation of theAntarctic Intermediate Waters toward the north and the

498


Arabian Sea waters toward the south. This barrier, due tostrong contrast between the two water masses, causes astagnation of the deep waters over Davie Ridge and if thisexisted in the past, might have contributed to oxygenimpoverishment of bottom water near Site 242.

Thus, a change in the circulation system of the centralwestern Indian Ocean waters very likely took place by earlymiddle Miocene and gave rise to the present circulationpattern. Is it possible to find any signs of the previouscirculation system? It is tempting to assume that theMozambique vortical circulation (Harris, 1972) was notinitiated by Pliocene times or even by late Oligocene, butearlier. In effect, a big hiatus encompasses about 40 m.y.(early and middle Cenozoic) at Site 249. This is by far thegreatest hiatus, which suggests either a stronger action ofbottom currents, or more likely, a longer duration. Frakesand Kemp (1972) postulated the existence of a westernprolongation of the South Equatorial Current of the PacificOcean into the Indian Ocean during Eocene and Oligocenetimes through a north Australian seaway (Figure 6). Thiscurrent, deflected southward by India to the eastMadagascar and the southeast African coast, would havehad the same effect as the present Indian Ocean SouthEquatorial Current which is a vortical circulation patternover Mozambique Ridge and Mozambique Basin (AgulhasCurrent).

The great oceanic gyre postulated by Frakes and Kemp(1972) would have passed north of the Site 245 area,contouring waters similar to the present Central Water Masses.

Mid-Cenozoic Hiatuses and Deep Bottom WaterCirculation

Introduction

Important breaks in the sedimentary columns weredetected in Leg 25 cores. Most are true hiatuses, as at Sites241 and 249, but some may be restriced to condensedsequences as discussed before for Site 245. Commentsabout hiatuses are made elsewhere (Sigal, this volume;Kent, this volume). Figure 2 indicates that hiatuses involvemainly pelagic deposits or, at least, are marked by clearchanges in sedimentation. This is the reason the importantproblem of hiatuses is shortly discussed in this chapter.Reinterpreted hiatuses and condensed series are depictedtogether with some major events of the Cenozoic IndianOcean history (Figure 7).

First of all, one must bear in mind that hiatuses are"asymmetrical." According to Pimm and Hayes (1972), thechange from sedimentation to nondeposition or erosionmay represent the reaching of an erosional level which doesnot necessarily correspond to the beginning of an associatedphysical event. On the contrary, it is the end of the hiatuswhich dates the end of the physical event action. Followingthis reasoning, most attention has been given to thetermination of Leg 25 hiatuses (Figure 7).

It is striking to note that mid-Cenozoic hiatuses end bylate Oligocene-early Miocene time in eight differentsedimentary columns raised from the central western IndianOcean. The earliest end is late Oligocene in age at Sites 241

paleotemperature

(20°) mean annualt empera ture ( a i r )

calculated paleotemperature

Figure 6. Mid-Oligocene global circulation patterns. Note the counter-clockwise gyral in the Indian Ocean, in the prolongmentof the South Equatorial Current coming from the Pacific Ocean. Madagascar Island is mislocated (according to Frakes andKemp, 1972).

499

L. LECLAIRE

Sites 240A 245A0 " I F7777I

249 248 246 241 239 1 2 3 4 5 6 7

20 -

40 H

>>ε60 -

0)cr>


of Rodriguez Island, and on the Madagascar Ridge. Otherevidences of current action are found on the eastern andwestern flank of the mid-Indian Ridge. Ewing et al. (1969)noticed that the southwest branch of the mid-Indian Ridgehas a well-defined, wide band barren of sediment over thecrest and emphasized the concept of current control ofsedimentation.

Although very little is known of present bottom-watercirculation, Antarctic Bottom Water is believed to penetratedeep areas in the southwestern part of the Indian Oceanaccording to bottom-water potential temperature studies.Le Pichon (1960) thought that: "a deep current toward thenorth, having the characteristic of a western boundarycurrent is deflected and weakened by the complex ridgesystem." Flows of Atlantic Deep Water, strengthened by asimilar flow of Antarctic Bottom Water, apparently wouldenter the Mozambique Basin. It is striking that in the samearea Ewing et al. (1968), have shown sedimentary featurestermed "giant ripples" evidencing the strong influences ofbottom currents. Le Pichon (1960) further postulatedanother similar western boundary current flowingnorthward and adjacent to Madagascar through theMacarene Basin. The existence of this current has beenrecently confirmed by Warren (1971). This current is wideand by 23° S, it exists as a northward flow immediatelyadjacent to Madagascar and a less pronounced one in thecentral part of the basin near longitudes 57°-62°E. This is aconfirmation of Le Pichon's suggestion. Other investiga-tions (Leclaire and Duplessy, Osiris cruise onboard MarionDufresne, 1973, unpublished results) support the existenceof this current. An erosional surface was seen at the top ofa pilot core raised from the South Mascarene Basin.Hydrological measurements suggest a passage of AntarcticBottom Water through the Southwest Indian Ridge, southof Rodriguez Island (A. Tanguy, personal communication).

Harris (1972) has found that mixing of great watermasses coming partly from the equatorial circulationsystem, and more explicitly from the South EquatorialCurrent, gives rise to a vortex pattern with giant vorticesover Mozambique and Madagascar ridges, probably down to1000 to 2000 meters depth. No doubt, Recent andPliocene-Pleistocene winnowings seen at Sites 249 and 246are closely related to this circulation as is thePliocene-Quaternary condensed series (or hiatus) at Site249. Other probably effects of this circulation pattern wereseen by Saito and Fray (1966) who reported the presenceof mixed assemblages including Upper Cretaceous, Eocene,Miocene, and Recent foraminifera found in two cores fromthe flank of Mozambique Ridge and assumed erosion bycurrents of outcropping at high elevations on the ridge. It isworthwhile to note in passing that Eocene sediments seemto be present at places on the Mozambique Ridge whereasthey are absent at Site 249, which strongly suggests bottomerosion.

Summarizing, two main circulation patterns presentlyexist in the central western Indian Ocean: (1) a deepbottom circulation pattern related to Antarctic BottomWater flowing northwards, and (2) a surface to subsurfacecurrent pattern more or less derived from the equatorialsystem.

The fact that mid-Cenozoic hiatuses are found in regionswhere those currents have been described cannot be

coincidental. Hiatuses at Sites 240 and 241 are apparentexceptions, but no data are available concerning a deepbottom current in the Somali Basin. However, aprolongment of the western boundary current fromMascarene Basin into Somali Basin is probable. Thus, thehypothesis that the same deep bottom circulation wasacting in the past must be discussed.

The present source of Antarctic Bottom Water is mainlythe WeddellSea (Jacob et al., 1970) and is closely related toAntarctic ice-sheet influence which cools the shelf waters.According to Frakes and Kemp (1972), the warm and wetconditions which apparently characterized latitudes beyond45° in the late Eocene were altered during the Oligocene toconditions more like the present, with cool surface waters.Glomar Challenger boreholes from Wilkes Land AbyssalPlain (Hayes, Frakes, et al., 1973) indicate cooltemperatures over this area in the late Oligocene-earlyMiocene. Consequently, formation of cold water in theWeddell Sea by Oligocene times is quite probable even ifthe Antarctic ice sheet was not fully developed. Followingthis reasoning, Antarctic Bottom Water, flowing northward,may have existed by Oligocene time and may have supplieda western boundary undercurrent in the western IndianOcean. If hiatuses are ascribed to those currents, aconsiderable weakening of their vigor from mid-lateMiocene until present time must have occurred. Thisassumption is apparently in opposition with the fact thatSouth Atlantic hiatuses (Berger, 1972) show a considerableincrease of the Antarctic Bottom Water flow related to thegrowth of the Antarctic ice sheet and the onset ofglaciation. In fact, the solution is contained in a remark ofLe Pichon (1960) cited before, i.e., that the westernboundary current has been deflected and weakened by acomplex ridge in middle Miocene time as well as presently.This complex ridge system was the same as today, thesouthwest branch of the mid-Indian Oceanic Ridge whichthereby suggests a strong uplift of the ridge beginning in thelate Oligocene-early Miocene. This point is stronglysupported by Sclater and Harrison (1971) who postulatedthe formation (or reactivation) of the southwest branch 20million years ago and can be correlated to the generalMiocene transgression over the East African coast (Kent,this volume; Sigal, this volume).

Consequently, deep western Indian Basin hiatuses arebelieved to be related to western boundary current actioninitiated by late Oligocene-early Miocene time. Thesecurrents stopped and hiatuses ended gradually when upliftof the south west branch took place. This complex ridgesystem acted and is still acting as a barrier, partly blockingand considerably weakening western boundary currents.

Presence of hiatuses at Sites 249 and 246 cannot besimilarly explained and the fact that they ended at aboutthe same time as the others seems like a coincidence.Obviously, those hiatuses may have been originated fromthe vortical circulation of the Agulhas-Mozambique systemdescribed before. But this circulation is connected with theequatorial system which supposedly started by middle-lateMiocene, i.e., when hiatuses ended. Apparently, there is adilemma here. Moreover, during Eocene and Oligocenetimes, an east-west water mass drift may have replaced thepresent South Equatorial Current. Additionally, Frakes andKemp (1972), who determined global paleoclimates from a

501

L. LECLAIRE

reconstruction of past positions of continents, drew a greatcounter-clockwise gyral (Figure 6) originating from theSouth Equatorial Pacific Water entering the Indian Oceanthrough a north Australian seaway (as mentioned before)during Eocene and Oligocene. Both flows may have actedsimilar to the present South Equatorial Current and wereperhaps even stronger, which possibly gave rise to a similarvortical circulation pattern over Mozambique andMadagascar ridges where strong reworking of older materialwas seen in the upper Cenozoic pelagic deposits. The end ofhiatuses seems to correspond to changes in the circulationpattern and to the onset of the equatorial current system.

Older hiatuses encompassing Late Cretaceo

20. LATE CRETACEOUS AND CENOZOIC PELAGIC DEPOSITS - … · 2007. 6. 5. · LATE CRETACEOUS &...

Documents

Transcript of 20. LATE CRETACEOUS AND CENOZOIC PELAGIC DEPOSITS - … · 2007. 6. 5. · LATE CRETACEOUS &...