Deep Sea Drilling Project Initial Reports Supplement to Volumes … · 2007-05-11 · volume of the samples was 5 to 10 cc. The samples were wet sieved using 44 and 63 µm sieves;

8. NEOGENE SAND LAYERS OFF NORTHWEST AFRICA:COMPOSITION AND SOURCE ENVIRONMENT

Michael Sarnthein, Geologisch-Palaeontologisches Institut der Universitaet Kiel, West Germany

ABSTRACTNeogene sand layers were recovered during DSDP Legs 14 and 41

in seven holes at the West African continental margin, between 12°and 35°N. Based on their sedimentary structures, grain sizes, andthe composition of their coarse fractions, three major genetic groupsof sand layers are distinguished: contourites, turbidites originatingfrom fluvial deltas, and turbidites formed by eolian dune sands,both in proximal and distal facies. The source environments arecharacterized by compositional features, which are independent oftransport sorting, such as the abundance of feldspar, of stained andfrosted quartz, or the composition of reworked faunas. Eolian-sandturbidites are the only layers with noticeable open porosity. Theyoccupy a time-transgressive zone which follows a northward plate-tectonic shift from the Pleistocene to the early Miocene of the areabetween 14°-20°N by 6 to 7 degrees of latitude. During thePleistocene, this zone is marked as the synglacial position of Saharadunes advancing towards the shore line at the shelf edge andgenerating turbidity currents at the slope. Fluvial-sand turbiditesoccur further to the south and north. The turbidites at Sites 370 and135 are additionally influenced by the rising Atlas Mountains.

Turbidites and contourites both appear concentrated in a fewstratigraphic horizons: at 23 to 20, ~13, and 3 to 2 m.y.B.P. Thepaleoenvironment of these horizons can be characterized by a lowsea level (indicated by the formation of eolian-sand turbidites itselfand by abundant erosion of glauconite on emerged shelf plains), byvigorous off-shore Trade Winds (moving dunes towards the shoreand forming dust beds on the slope), by strong coastal upwelling(dominance of fish remains and siliceous fossils at the supposedcenter of Trade Winds influence), by temperate temperatures incoastal waters (skeletal sands of "Foramol" group), and byincreased bottom-current activity (contourites). In conclusion, theseconditions resemble Pleistocene glacial events and agree well withevidence from other parts of the oceans.

An early phase of Cape Verde volcanism shows up with pumicegrains in the lower NN 18 Zone.

INTRODUCTION

The composition of turbidites is determined by thenature of the source environment, erosion duringtransport and by sorting processes. This paper providesa summary of information on the composition ofNeogene and Oligocene sandy layers off northwestAfrica, most of which are assumed to be deposited byturbidity currents and related grain flows.' The layerswere studied in 21 cores which were recovered at sevensites during Legs 14 and 41 of the Deep Sea DrillingProject.

This study follows previous ones (Sarnthein andWalger, 1974; Sarnthein and Diester-Haass, 1976) onthe eolian-sand transport off of the western Sahara, andextends models of Holocene and Pleistocene climate

'The term "turbidite" is used in a summarizing sense.

belts into the Tertiary. The turbidites are used as aunique storage of information on: (1) periods of lowsea-stands with a strand line close to the shelf edge, (2)ancient desert environments with an off-shore windregime, and (3) the ancient position of arid and humidclimate belts, which are short lived and shift rapidlyleaving behind only minor and difficult-to-date traceson the continental platforms.

The areal and stratigraphic distribution of cores isshown on Figures 1 and 2, respectively. The upperTertiary was only spot cored because the primaryobjectives of Sites 367 and 370 were to encounter deephorizons. Core recovery was much better at Site 368,but sandy layers rarely occur there. Table 1 summarizesthe length of stratigraphic record represented by thecores (Berger and von Rad, 1972), and the distributionof core lengths at Sites 367, 368, and 370. In regard tothis table, it must be stated that the stratigraphic andpaleoenvironmental interpretations of this report

939

M. SARNTHEIN

CAPE VERDE BASIN

CAPE VERDE ISL|368‰

Figure 1. Physiographic setting of Sites 135, 138 to 140 (Leg 14), 367, 368, and 370 (Leg 41). Note that Site 138 is themost distal of sites and is in line with a channel coming from Site 139. Site 368 is on the Cape Verde Rise. Modifiedbase map from Berger and von Rad (1972).

940

COMPOSITION AND SOURCE ENVIRONMENT OF NEOGENE SAND LAYERS

Q_

LU

LU

2 E•― o

139

135 370

5W1

SW2

3

A

/

/5

\ 1

II I

138

I I

\

\

D I S T A N C E

B E T W E E N S I T E S

500 km

\

\

\

\

\

368

A222

367

mio. yrs.0 NN

10 _

20 _

NNNP

30 _

25

24

I135 370

swi

SW2

B

138 139 140

A2

368 367U 2

3U5

Figure 2. Position (A) and ages (B) of cores containing sand layers. Blank core boxes refer to uncer-tainties in dating.

941

M. SARNTHEIN

TABLE 1Sample Representation of Oligocene to Pleistocene Cores

Percentage of Coresin Length Classes

(m)

Sites 367, 368, 370

Represented Length of StratigraphicRecord (m)

Sites 367 and 370Cores (%)

Site 368Cores (%)

47% >8.015% 4.5-6.023% 2.54.015% 0-2.0

8.5% <IO16.5% 21-308.5% 3140

16.5% 41-5025.0% 51-60

8.5% 81-9016.5% 101-110

64.0% <IO13.5% 11-209.0% 21-30

13.5% 3140

100% (34 cores) 100.0% (12 cores) 100.0 (22 cores)

generally suffer from a paucity of data. Additionally,the interpretations are weakened because turbiditesdisplay insights only into an ephemeral, if notincidental set of environmental conditions.

DISTRIBUTION PATTERN OF SANDY LAYERS

GeneralSite 138 lies most distant from the present shelf edge

(1000 km), whereas Sites 139, 367, and 370 lie closest tothe shelf edge (280, 250, and 120 km distance, respec-tively). The seven sites cover a wide climatic rangebetween the humid northern tropical belt in the south(Site 367: 12°30'N) and the humid Mediterranean beltin the north (Site 135: 35°1O'N, Site 370: 32°50'N). Thisdistribution was shifted some 6° to 7° of latitudesouthward in the early Miocene, according to thereconstructions of plate motions (Francheteau, 1970;Berger and von Rad, 1972).

Each sandy layer which occurs in Oligocene toHolocene cores was sampled and catalogued forthickness, internal fabrics, character of its boundaries,its relations to neighboring facies, and artificialdisturbances.

Unless otherwise stated, the age classification of Leg41 cores was determined by the shipboard paleon-tologists (P. Cepek, D. Johnson, V. Krasheninnikov,and U. Pflaumann) by means of coccolithophorids,radiolarians, and foraminifera. The ages of cores fromLeg 14 are from Hayes, Pimm, et al., 1972. Absoluteages were correlated to the biostratigraphy followingBerggren and van Couvering (1974) and Ryan et al.(1974).

Site 367Site 367 lies southwest of Cape Verde at the eastern

boundary of the Senegal Abyssal plain, off the SaloumRiver mouth, beyond the present influence of the CayarCanyon fan. Two categories of sandy layers occur inCores 1 through 5.

Dark and fine quartzose sandy beds appear in Cores1 and 2 (upper Pleistocene). The beds are thin (Figure3) and frequently dragged into lenticular or wedgeforms by coring disturbance. The sandy layers havesharp contacts with the adjacent nanno marl and darkgray silty clay and show no gradation or internalfabrics. This lack of structures may also be the result ofdrilling disturbance.

Cores 3, 4, and 5 (upper Pliocene) contain medium-sized, reddish-yellow quartzose sand beds which aremassive (Figure 3; up to >l m) and occur at frequentintervals (12 layers in 12.5 m core length). Drillingdisturbance is common and results in a dragging of thesand beds into wedges and lenses, and in the intrusionof flakes from the overlying muds. The beds show nolamination and almost no gradation regardless ofthickness. Heavy mineral layers (1 to 2 cm thick)characterize some of the beds at their base. The uppercontacts are as sharp as the lower ones, but some of thelower contacts are wavy and indicate erosionalprocesses may have preceded deposition.

Site 368Site 368 was drilled on top of the Cape Verde Rise.

Sand layers occur as tiny and sharply defined lenses andinterbeds of fine white sand in the early Miocene ofCores 17 and 18. The layer thickness ranges from 0.1 to1.0 cm. No internal structures were observed. Core 22(middle Eocene to early Miocene) contains a sequenceof coarse-sediment beds a few cm thick which may be ofvolcanic origin. They show an upward gradation intogreenish claystones. The sediments younger than earlyMiocene lack any coarse-sediment layer.

Site 370Site 370 lies in a 650-meter-thick submarine fan off

the Moroccan continental margin. Fan deposition hasbeen active since at least the Paleocene. Hemipelagicsediments between Oligocene (Core 5) and Pliocene(Core 2) contain regular silt and sand layers andindicate a strong terrigenous influence persisted duringthat time. The sand beds occur in narrow intervals (59layers in 18.5-m core length), are generally thin (Figure3), and distinctly graded. Lamination and convolutebedding are common. Most of the lower contacts showthe effects of erosion on reddish-white or greenish-graymarl. The lower part of Core 4 shows a strikingexample of slumping and graded-bed structures. Thesands are predominantly dark-green. Drillingdisturbance is rare and slight.

It is interesting to note that thin layers and nodules ofporcellanites are frequently associated with the top ofgraded layers, the youngest ones occurring in the upperpart of Core 3 (middle Miocene). Lithificationgenerally increases from 200 meters depth (Core 3)downward, a factor which has significantly biased thegrain-size analyses because of the formation of stiffmud and silt lumps. A major hiatus might best explainthe distinct age difference between the two successivecores (Site 370, Cores 3 and 4).

Sites 135 and 138, 139, 140 (Leg 14)The DSDP Leg 14 sites are discussed by Berger and

von Rad (1972). Neogene to Oligocene quartzose sandlayers exist in Cores 135-SW1 and SW2, 138-1 and 2,139-6 and 7, and 140A-1 and 2. Apart from their lesserdegree of staining, most sand beds from Leg 14basically correspond to those of Cores 367-3 to 5 withtheir lack of gradation regardless of thickness, theircoarse, loose appearance, and their massiveness (Figure3). Noticeable signs of lithification were observed onlyin Core 139-6.

942


o

40

0 _

Lü

Z)

σLü

CC

LL

0

40

040

0 _

40

0

368/17*22138/1-2(11 LAYERS)

138/1139/6-7(20 LAYERS)

367/1-2(9 LAYERS)

367/3-5(12 LAYERS)

370/1-5(59 LAYERS)

0-1 1-6 6-12,5 12,5-25 25-50 50-100 >100

cm T H I C K N E S SFigure 3. Frequency distribution of sand-layer thickness.

943

M. SARNTHEIN

DISTRIBUTION OF GRAIN SIZES ANDSAND COMPONENTS

MethodsSamples for coarse-fraction analysis were taken from

the base, the top, and the middle of thick sand layers.Thin layers were sampled only close to their base. Thevolume of the samples was 5 to 10 cc. The samples werewet sieved using 44 and 63 µm sieves; the fine fractionwas collected quantitatively and dried at 60°C. Dryweights of the fine and coarse silt and the sand fractions(>63 µm) were determined. The sand fraction was drysieved into subfractions of-1 to 0, 0 to 1, 1 to 2, 2 to 3,and 3 to 4 Phi grade. The percentage proportions of thesubfractions were evaluated for a median grain-sizevalue of sand.

Two sources of bias must be kept in mind wheninterpreting the grain-size data: (a) It was frequentlyimpossible to completely separate clean samples of thinsand lenses from the surrounding muds, therebygenerating an artificial amount of fine fraction, (b)Increased diagenesis hindered the sieving of stiff muddysands into true grain-size fractions, even after strongtreatment by peroxide. On the other hand, intensivetreatment destroyed debris of brittle pelagic mudcontained as natural particles in the sand layers.

Counts of about 300 to 400 representative grains persand subfraction were made for component analysisusing a stereomicroscope with incident light (mag-nification up to 80×). The grain number is sufficient inorder to obtain a gross picture of the componentdistribution (details in Sarnthein, 1971). Quantitativecomponent analysis was applied to all sandsubfractions in samples from Sites 135, 138, 139, and140A. Samples from Leg 41 sites were counted usingonly two or three dominant "key" subfractions. Theother (mainly the finest) sand subfractions wereexamined only for the abundance of some keycomponents (e.g., mica and stained quartz).

Thirty to 50 components were distinguished in total(Table 2), but the diversity of particles within onesample fraction was usually rather small. The difficultseparation of quartz and feldspar under incident lightwas confined to subfractions >25O µm (rarely >125µm) and was confirmed by spot-check counts of somethin sections and stained quartz-feldspar slides.

Frosted surfaces of quartz and feldspar grains weredetermined using the optical microscope using thecriteria of crescent-shaped pits also occurring onotherwise glossy surfaces. The count of frosted grainswas restricted to grains >250 µm (in rare cases also to>125 µm) because the frosting process by eolian grainsaltation is not clearly developed with smaller sizefractions (Bagnold, 1941). The coarse fraction wasdominated in Site 370 by artificial mud lumps, whichwere difficult to distinguish from fecal pellets andnaturally redeposited particles of pelagic or deltaic mudsediment. Because of this, the composition of mostsamples was estimated only semiquantitatively after aquantitative count of a few random samples had beenmade.

Appendix A summarizes the following results:1) Grain size: (a) percentage of sand fraction (>63

µm) per total sample; (b) percentage of coarse siltfraction (44 to 63 µm) per total sample; (c) median size

of sand fraction; (d) range of median size of quartz plusfeldspar within the sand fraction.

2) Proportions of terrigenous particles, shallow-water biogenous grains, pelagic biogenous components,nonbiogenic shallow-water sediment debris, pelagicsediment debris, fecal pellets.

3) The following calculated ratios (per size fractionsas indicated in the respective columns): (a) % feldsparof total quartz plus feldspar; (b) % stained quartz oftotal quartz; (c) % frosted quartz of total quartz; (d)ratio of quartz plus feldspar to mica.

4) Dominant components in the group "shallow-water biogenous grains."

5) Dominant components in the group "pelagicbiogenous grains."

6) Resulting type of sand layer.

Grain SizeTwo main groups of sand layers are distinguished

based on the abundance of sand fraction (>63 µm)(Figures 4 and 5). Group 1 is composed of clean sandlayers from Cores 139-6 and 7 (55% to 90% sand), Cores367-3 to 5 (75% to 98% sand), and Cores 135-SW1 and 2and 140A-1 and 2. A few samples in group 1 haveartificial mud contamination which reduces the sandpercentages. Group 2 is represented by sands with 5%to 55% coarse fraction which are found in Cores 138-1and 2, Cores 367-1 and 2, Cores 368-17 and 22, andCores 370-2 to 5. These samples may be biased by mudlump contamination. The two groups parallel thegrouping by bed-thickness frequency, however, theyseem not to be related to bed thickness itself.

The abundances of coarse silt (44 to 63 µm) stronglyconfirm this two-part subdivision (Figure 5). Onlyminor amounts of coarse silt occur in the sand layers ofgroup 1 (Cores 367-3 to 5, generally 0.1% to 1.0% silt;Cores of Sites 139 and 140A less than 3.5% to 5.0% silt;and the lower samples from Core 138-1, 2% to 8% silt).On the other hand, sand layers of category 2 arecharacterized by high proportions of coarse silt. Cores367-1 and 2, Cores 368-17 and 22 have 20% to 31% siltand Cores 138-1, top, and 370-2 to 5 have 6% to 28%silt. It must be stressed that the abundance of coarse siltappears to be independent of the position of a samplewithin a sand layer. Summarizing, many sand layers ofgroup 1 show a double-peaked grain-size frequencydistribution, sand layers of group 2 show a simple one.

The median grain diameters of sand layers in group 1generally range between 140 and 225 µm, but a few (lessthan 20%) reach extreme values of up to 290 and downto 110 µm. Median grain sizes of sand layers in group 2vary between 70 and 100 µm. However, frequently theyare strongly biased towards coarser sizes (up to 500 µm)by abundant indestructible lumps. Another factor forerrors in coarse diameters, especially in cores of Site370, is a significant proportion of light sand-sized com-ponents, e.g., planktonic fossils, compared with therather clean quartz sands of group 1. Recalculatedmedian sizes of quartz sand always are smaller than 90µm (Figure 6).

The sand layers of Core 138-1 show a continuousupward decrease in median size (from 113 to 93 µm)(Figure 4). The abundance of sand varies independentlyof this trend and decreases abruptly from 4 to 3 metersbelow the top of the core. A general slight upward

944


TABLE 2Coarse Fraction Components Distinguished in This Study

ClasticTerrigenousComponents

NonbiogenicShallow-Water

Sediment Debris

Shallow-WaterSkeletal Grains(Reworked and

Corroded)

Other BiogenicComponents

(Mostly Pelagic)

PelagicSediment

DebrisMiscellaneousComponents

Quartz unstained

Quartz stained

Quartz unfrosted

Quartz frosted

Quartz angular

Quartz rounded

Feldsparunstained

Feldspar stained

Feldsparunfrosted

Feldspar frosted

Mica

Heavy minerals

Rock fragments

OöidsSuperficial oöidsphosphatic oöids

Faecal pellets,lithified orglauconitic

Beach rock, limo-nitic sand- andsiltstones, partlyrounded and polished

Resedimentedglauconite particles

AlgaeForaminifera(benthos)

Corals

Bryozoa

Bivalves

Gastropods

Barnacles

Echinoids

Plant fibres

Pknktonicforaminifera

Benthonicforaminifera

Diatoms

Radiolaria

Sponges

Bivalves

Gastropods

Pteropoda

Osiracodes

Decapods

Echinoderms

Fish remains

Worm shells

Chert

PorcellaniteLaminatedSiltite

Authigenicautochthonousminerals

Iron crusts

Pumice

Fecal pellets

89 111 140 175 225 µm0 20 40 60 80 100%

138/1

2-

i

89 111 140 175 225 µm0 20 40 60 80 100%

Figure 4a. Core sections showing position of sand layersand grain sizes. Dotted sand layers are muddy and mostlygraded. Coarse signature marks clean ungraded sands.

decrease of sand abundance in Core 139-7 is found aswell, in spite of an increase of bed thickness. Site 367shows only one marked change from sand-layer group1 to 2 between 54 and 17.5 (?12) meters depth, i.e.,between upper Pliocene and middle Pleistocenesediments. Site 370 reveals no general trend except for

139/71

2

Figure 4b. Core sections showing position of sand layersand grain sizes. Dotted sand layers are muddy and mostlygraded. Coarse signature marks clean ungraded sands.

Core 5 (Oligocene), where sand proportions andmedian sizes decrease and silt proportions generallyincrease from bottom to top (17 samples from 28 sandlayers in 8-meter-thick sediment core). In summary, avertical comparison of the sand layers displays only afew significant overall trends of grain size.

945

M. SARNTHEIN

367/11

2

3

4

5

6

367/21

2

V-

1\

m-•

1-

2

rrr

é

250 µ63 79 99 125 157 200

0 20 40 60 80 100%L _ l I I I I I I 1 I 1 L.

17,5 to

54,0 m

367/31

22\

mi

>63µ

57,5 to

63,5 m

Figure 4c. Core sections showing position of sand layersand grain sizes. Dotted sand layers are muddy and mostlygraded. Coarse signature marks clean ungraded sands.

367M1

2:

m:

73 to150 mdepth

367/5

T 20 40 60 80 100 °/o63 79 99 125 157 200

250µ

Figure 4c. (Continued).

370/2

2i

DO:

—*_^—«

Λ 79 125 200 308 500 urn0 . 20 . 40 . 60 . 80 .iöO°lo>63µ

I IMd >63p

107,0-207,5 m

370/31

2

"—1—•^*

—~ — v

210.5-217,0m

Figure 4d. Core sections showing position of sand layersand grain sizes. Dotted sand layers are muddy and mostlygraded. Coarse signature marks clean ungraded sands.

946


370A

370/51

ED

0 20 40 60 80 100%>63µ79 125 200 308 500 µm

Figure 4d. (Continued).

TERRIGENOUS COMPONENTS

Quartz and FeldsparQuartz and feldspar are the dominant components of

sand layers except for Site 370, where resedimentedglauconite and biogenic particles dominate many sam-ples and where true proportions are partly obscured byan abundance of mud lumps.

Two groups of sand layers are clearly distinguishedby the distribution of median-sand sizes of quartz andfeldspar (Mdq ). There is a correlation in a majority ofcases of the groups with the total sand median sizes

(Figure 6): Fine-sand medians range up to 95 µm(layers in Cores 138-2, 367-1 and 2, 368-17, and 370-2 to5); Coarse medians range from 140 to 290 µm (layers inCores 135-SW1 and 2, 139-6 and 7, 140-1 and 2, and367-3 to 5). There is a gap in MdQ sizes between 95 and115 µm (Figure 6). The gap could be considered tocontinue up to 140 µm because only 5% of all sampleslie between 115 and 140 µm. The only exceptions to thisrule are the sand layers of Site 138 (Core 1), the sitefarthest from Africa. Generally, only quartz sizes withan optimum transport mobility (MdQ = 100 to 120µm; Hjulström, 1939) are absent in the sand layers.

The abundance of medium sand-sized feldspar ishigher than that in the coarse sand-sized grade. Highestratios of feldspar to quartz plus feldspar occur in upperPliocene arkosic sands (Cores 367*3 to 5: >500 µm,10% to 30%; 250 to 500 µm, 17% to 35% feldspar). Theearly Miocene samples of Cores 138-1 and 139-6 and 7are rather mature quartz sands (125 to 500 µm, 10% to23% feldspar; also see Berger and von Rad, 1972). Thegeneral association of feldspar to well-rounded quartzsuggests the feldspars have a continental origin (seesection on facies).

Hematite stains and coatings are abundant on bothquartz and feldspar grains. This feature was termed"desert quartz" by Radczewski (1939) for Holoceneand Pleistocene sediments. More feldspars are stainedthan quartz by some 20% to 40%, and generally thecoarser fractions show staining more than the finerfractions. Most coarse desert quartz is found in upperPliocene sands of Cores 367-3 to 5 (Figure 7). Theabundances of coarse desert quartz are as high as 90%and resemble those of recent dunes (Sarnthein andDiester-Haass, 1976). A smaller but still significantproportion of these grains occur in late Oligocene toearly Miocene sand (sample group C and D). Fine-sandquartz (63 to 125 µm) in upper Miocene, Pliocene, andPleistocene cores is intensively stained, regardless ofgrain-size distribution. Stained fine quartz is generallyrare in lower Miocene and Oligocene cores, except forCores 139-7 and 370-4. The lower Miocene samples ofcategories G and H (Figure 7) are almost unstained.The amount of desert quartz only rarely formscontinuous trends within a single core profile. Forexample, in Core 138-1 the amount of stained quartz(125 to 250 µm) decreases continuously from bottom totop. In Core 139-7 the desert quartz proportionincreases in all sand-size fractions from Sections 6 to 3,and then decreases to Section 1. The fine-sand layers ofSite 370 reveal a continuous increase of the stainedquartz fraction from Oligocene to upper middleMiocene.

The roundness of quartz grains is closely related tograin size, Well-rounded quartz grains are mostdominant in the coarse size fractions (>250 µm) in allsamples, except for those of Cores 135-SW1 and 2.Frosted quartz grains have a distribution patternsimilar to the stained grains and the feldsparpercentages. Frosted quartz grains also indicate acontinental desert origin for the sand layers. Forty-fiveto 70% of the quartz grains >500 µm are frosted inCore 367-4, but the abundance decreases continuouslyto 20% at the top of Core 367-3. Core 139-7 has up to25% frosted quartz grains. For quartz grains >250 µm,

947

M. SARNTHEIN

50

60

*/. f i n e s i l t + c l a y

Figure 5. Gay-coarse silt-sand triangle of sand-layer samples, based on sieve analyses.1 = Cores 367-3 thru 5,2 = 367-1 and 2, 3 = 368-17 and 22, 4 = 370-2 thru 5,5 = 135-SW1 and 2, 6 = 138-1, 7 = 139-7, 8 = 140A-1 and 2.

10

20 .

0 _

138/1

ALL CORES

EXCEPT FOR

138/1

63 ? 1 79 8 β | 99 1 " j 125 ™ 157 1 7 5 200 ™ 25095 115

MEDIAN SAND SIZE OF QUARTZ PLUS FELDSPAR

308

Figure 6. Frequency distribution of median sand size quartzplus feldspar. Note the distinct gap between 95 and 115µm.

the numbers of frosted grains are generally smaller:24% to 43% in Core 367-4, 13% to 32% in Core 367-3,and up to 30% in Core 139-7. The quartz fraction > 125µm has 7.5% to 20% frosted grains in Core 139-7, and5% to 20% in Core 138-1. The bases of sand layers aregenerally enriched in frosted quartz grains with up to10% in abundance.

Mica

The abundance of terrigenous mica is correlated tothe Md because of grain-size sorting on the sea floor(Figure 8). Accordingly, less than 0.01% mica is found

A

B

C .

D

E

FG

J>125367/3-5

-500

139/7,140/1-2

370/2/1-2,370/3/1370/3/2,370/4/3-4

370/5/1-6

138/2,368/17,138/1/2

20 U0 60 80

•/. STAINED QUARTZ + FELDSPAR

100

Figure 7. Abundance of stained quartz plus feldspar in totalquartz plus feldspar, as related to different grain size frac-tions. Generally, the staining increases with increases ingrain size, except for Cores 367-1 and 2. Also the increaseof staining in the fine sands of Site 370 towards the lateNeogene can be interpreted as an increase of influence of(reworked) Sahara dust.

948


10000

1000 _

100

10 _

1 .

© © X*

o Θ

© xo » ©

° xX

Θ 63-125

x 125-250 µm

• 250-500

69 79 99 125 157 200 250 308 µm M D Q

Figure 8. Quartz plus feldspar to mica ratio, related to themedian sand size of quartz plus feldspar. Broad scatteringof data in the ratio range of 400 is due to the grain count300 to 500 where single mica grains have been consideredas traces.

in sand layers of Cores 135-SW1 and 2, 139-7, 140A-1and 2, and 367-3 to 5. One to 4% mica occurs in theremaining samples. The data suggest that a continentalsource did not influence the quartz to mica ratio.

Heavy Minerals

Heavy minerals and iron oxides are a commonconstituent (0.15% to 2.5%) of the very fine sand (63 to125 µm) and coarse-silt fractions of the coarser grainedsample group. Iron oxides frequently are ball-shapedindicating a mechanical rounding by reiterative rollingtransport at shorelines.

Shallow-Water Sediment Debris

Nonbiogenic sand particles persistently occur whichare characteristic of those formed in very shallow waterenvironments, such as those known from the modernSahel and Sahara shorelines in Senegal and Mauritania(Koopmann et al., in preparation; Schwarz et al., 1975).The principal particles of this group are quartz grainswith superficial, frequently blackish oöid coatings, andbeachrock debris (calcite-cemented quartz sandstone),both of which are characteristic of coastal regions ofsubtropical, arid coasts. Lithified and roundedlimonitic or glauconitic fecal pellets, debris of limoniticsiltites and sandstones also are abundant. The latterfrequently show rounded and polished surfaces as well.Limonitic sediment particles from Holocene sedimentsare common in the semiarid transitional zone.Hematite coatings on desert quartz from this zoneeasily become mobilized in stabilized dunes and wind-dust deposits by temporary fresh-water influence.

Subsequently, the mobilized iron is reprecipitated aslimonite cement. Finally, phosphatic oöids andglauconitic-mudstone particles are included in theshallow-water particle group, because the primaryformation of glauconitic constituents seems to becommonly associated with environments of arid tosemiarid coasts (e.g., Birch, 1973; Lange and Sarnthein,1969). Mature rounding and polishing of the glauconitegrains suggests multiple reworking (e.g., bytransgressions) and placer-like enrichment, such as arecommon at beaches, before the particles finally becomeburied in the deep sea.

The persistent fractions of shallow-water sedimentdebris can serve as reliable tracers for the quartz-feldspar sands to a shoreline origin even if theycontribute generally less than 2% to 3% to the sandlayers. Limonitic siltites and sandstones, as well asglauconite grains, occur in all sand layers, except forSite 368 (Figure 9). Glauconite is most abundant in thenorth, in Core 370-5 (up to 33%); the maximumabundance of limonitic sediment debris is from middlePliocene sediments of Cores 367-4 and 5. Oöids andbeachrock debris are confined to sand layers from thesouthern sites (Sites 138 to 367), except for Cores 367-1and 2. However, no conclusive time-transgressive trendof distribution is found because the abundance of sandlayers is not sufficient.

The fine-grained, silty sand-layer group (Figure 6) ofCores 367-1 and 2 also can be characterized by anextraordinary abundance of fecal pellets and plantfibers. Both can be interpreted as tracers for redeposi-tion of deltaic shelf sediments (Sarnthein, 1971;Sarnthein and Walger, 1973).

Debris from metamorphic rocks contributes anunusually large fraction (20% to 35%) to the sands ofSite 135. These grains may be the product of shallow-water erosional processes. Rare admixtures of meta-morphic rock fragments also occur in Cores 139-6 and7.

Pelagic Sediment DebrisThe pelagic sediment debris consists of angular

grains of chert, porcellanite, microlaminated siltite,nanno chalk, foram-nanno chalk, and limestone. Theircomposition is usually somewhat different from thelocal pelagic fine-fraction background such as found infine-fraction lumps and in the overlying and underlyingsequence. The abundance is generally less than 1% to4% and is randomly distributed. Higher values of 7% to25% are found in the sands of Cores 138-2, 367-5, 368-22, and 370-2 to 5.

The major proportion of pelagic particles isobviously derived from the neighboring continentalslope and might help to decipher the erosion history ofits canyons (Sarnthein and Diester-Haass, 1976).

Authigemic MineralsThe sand-sized authigenic minerals include gypsum,

pyrite, calcite, and quartz (glauconite and phosphateare regarded as displaced shallow-water particles).Crusts of iron oxides are almost impossible todistinguish from contaminations by abrasion ofparticles, from the drill string. Similarly, gypsum

949

M. SARNTHEIN

mio. yrs.0 NN 135 370 138139 UO 367

30

4 0 0

23

22

21

M

-r B

d i s t r i b u t i o n b o u n d a r i e s :

f r a g m e n t s ofm e t a m o r p h i c rock

g l a u c o n i t e p a r t i c l e s

l i m o n i t e s a n d s t o n ea n d s i l t i t e

b e a c h r o c k

n - o o - r s u p e r f i c i a l oö ids

- r - Δ -T- d e l t a - m u d c l a s t s g.a.

Figure 9. Regional and stratigraphic patterns of shallow-water non-biogenous grains. Little hooks at lines point to area ofcomponent abundance. Narrow spacing of hatchure refers to higher abundance.

f r a g m e n t s of g r e e n a l g a e

crystals may have grown after the cores were stored inthe repository because they occur predominantly in theolder cores of Leg 14.

Quartz grains with secondary silica overgr iwth arefound in early Miocene sediments of Cores 139-6 and 7(see Berger and von Rad, 1972). These grains arefrequently associated with calcite cement, which isconfined to these two cores, and represent displacedbeachrock cement and autochthonous cementation ofthe sand layer in Core 139-6. Authigenic silica alsooccurs as void fillings in skeletal grains and as euhedralquartz grains in Cores 367-1, 368-22, and 370-2.Pyritized skeletal grains and voids are common in thefine-grained sand layers, especially in Core 367-2 (up to46%) and in Cores 370-2 to 4.

Volcanogenic Grains

A few olive-brown pumice fragments in Core 367-4,Section 4 are the only volcanogenic components from

the sites studied. The grains are wEll rounded and aremedium and coarse sand-size classes (up to 2.0 mm).The pumice probably marks an early active phase of thecape Verde volcanism (lower NN 18 zone,corresponding to 2 m.y.B.P.) (Hebrard et al., 1969).

Shallow-Water Biogenous Components

In addition to the fragments of shallow-watersediment, the majority of sand layers contains afraction of coarse-grained, benthonic shallow-waterfossils (commonly less than 1% to 5%; Appendix A).Their primary origin is clearly identified by, forinstance, thick shell structure or by the state ofpreservation.2 As an example, many pelecypod shells

2Most of the distinguishing criteria was based on the author'sexperience with shallow-water sediments off northwest Africa andelsewhere (Sarnthein, 1971). The percentage figures on shallow-waterfossils in Appendix A represent minimum values.

950


are coarse and intensely corroded by mechanicalreworking or by onsets of beachrock-like cementationwith other sediment grains. Other shells show aweathered appearance and are perforated by blackishmicroalgal borings filled with goethite or glauconite.They are like "mummies" which are characteristic ofthe modern and sub Holocene sediments along theshorelines and at a few meters water depth off theWestern Sahara desert (Koopmann et al., inpreparation).

Single crystals from the prismatic layer of Pinna sp.are particularly common in the fine-sand layers ofCores 367-1 and 2. The skeletal association as a wholecorresponds to the "foramol" group of Lees (1975)(dominant groups: benthic foraminifers, pelecypods,and barnacles), which is typical of the temperate-watershelf environment. Occasionally, large foraminifera,such as Amphistegina sp. (Cores 367-3, 370-2, 140-1,Section 1), and green-algae fragments (Core 367-5)indicate slightly more favorable temperature conditionsclose to the ones of Lees' "chloralgal" group.

The state of shell preservation was not classifiedquantitatively. However, distinct dissolution patternswere noted on benthonic foraminifera, pelecypods, andbarnacles. Areas such as a damaged (aragonitic)nacreous layer of a bivalve, an etched (calcitic) balanidscutum, or missing keels and holes of (calcitic) tests ofbenthonic foraminifera were scanned for effects ofdissolution. Preservation is generally poor in the sandlayers of Core 139-7, sections 3 to 6, in Core 140A-2(silicified skeletal hardparts only) and in Cores 367-3 to5, with a maximum of dissolution in Core 367-4 (nobivalve shells). The absence of shallow-water fossilsmay be either a primary or a dissolution feature inCores 138-2 and 368-17 and -22. Well-preservedskeletal grains occur in the fine-grained sand layers ofCores 138-1, 367-1 and 2 (e.g., aragonite needles ofPinna sp. from shelf sediments) and Cores 370-2 to 5.There appears a salient difference in the preservationpattern of hard parts taken from fine-grained and fromcoarser-grained sand layers due to a different pore-water circulation.

Pelagic Biogenous Components"Pelagic" biogenic components form a variable

proportion of the sand fraction (Appendix A). Most areof planktonic or nektonic origin. The few benthonicparticles either consist of characteristic deep-waterspecies (e.g., foraminifera) or of forms which do notindicate a formation in a shallow-water environmenteither by shell structure or by shell preservation.

Siliceous and apatitic particles (radiolarians,diatoms, fish remains, and sponge spicules) dominatethe pelagic fossil fraction in many sand layers ofOligocene and early Miocene age (Appendix A, Figure10). Carbonate fossils, such as planktonic andbenthonic foraminifera, ostracodes, and pelecypods,are dominant in the remaining sand layers of variousages. The distribution pattern is difficult to interpretbecause the sample coverage from the sites is poor(Table 1), and because the vast majority of these pelagicfossils are probably reworked from shallower slopeenvironments. However, the Oligocene to early

Miocene and regional concentration of siliceousskeletal grains and fish remains at Sites 138, 139, and368 is striking because it seems to be unrelated to thepaleodepth of the water (Figure 2a). Perhaps thisconcentration of siliceous material is related toupwelling phenomena off western Africa, such asobserved in the Holocene by Diester-Haass (1976, andthis volume).

There is some doubt in a deep-sea environment inter-pretation for the benthonic foraminifera and ostra-codes that occur in abundance in the sands of Cores367-1 and 2 (Appendix A). For instance, the ostracode-mollusc ratio ranges from 0.5 to 2, the ratio ofbenthonic foraminifera and molluscs between 2 and100. Without knowing their species composition, andonly on the basis of their large proportions in the sedi-ment, both fossil groups generally can be regarded asrather indicative of prodeltaic inner-shelf environments(personal observations at modern Senegal delta;Sarnthein and Walger, 1973). Some features of the sandsamples which support this interpretation aresummarized in the next section.

A GENETIC CLASSIFICATION OF SAND LAYERSDeposition by gravity-driven downslope density

flows and deposition by contour currents parallel to theslope are two basic modes conceivable for the forma-tion of terrigenous sand layers in the deep sea (e.g.,Hollister and Heezen, 1972; Bouma and Hollister,1973; Carter, 1975). To complicate matters, thematerial which comprises a turbidite can be derived at apassive continental margin from either a fluvial or aneolian-dune sediment source and the sediments of bothcan be modified by intercalated long-shore transportphases (Sarnthein and Diester-Haass, 1976). Theturbidite composition is further differentiated by aproximal or distal place of deposition (numerous titlessince Bouma, 1962). Contour currents will also reworkturbidites and create mixed facies types. Finally, duringarid conditions, extremely vigorous off-land storms cansupply coarse material directly to the sea bottomforming fine-sand beds, but these are confined to themost proximal continental margin (Sarnthein andDiester-Haass, 1976).

The sand layers of all cores are tentatively classifiedin Figure 11, Table 3, and Appendix A, into variouscategories using structural, textural, and compositionalfeatures and Holocene analogs from West Africa, suchas are summarized and defined by Fritz and Pilkey(1975) and Sarnthein and Diester-Haass (1976). Eoliandunes and fluvial deltas can be clearly distinguished asa source of turbidites by the different grain-sizefrequency groups and by different abundances offeldspar, frosted and stained quartz, fecal pellets, plantfibers, and faunas. A greater or lesser abundance ofdesert-sand features further differentiates the proximaleolian-sand turbidites of Cores 367-3 to 5 and 139-6 to7. Whether these sediments are a result of differentamounts of reworking by long-shore drift or to adifferent maturity of the desert sands is unknown.

The significant component of angular, fine-gravelfraction in Cores 135-SW1 and 2 excludes a duneorigin, but it does not necessarily argue for a delta

951

M. SARNTHEIN

mio. yrs 135 370 138 139 UO 368 367

f19

10-

2 0 -

NN

NP

3 0 -

20

o-

L\'

* *

dominance of

c a l c a r e o u sf o s s i l s

s i I i c e o u sf o s s i l s

f i s h r ema i n s500 km

f f a e c a l p e l l e t s

Figure 10. Regional and stratigraphic distribution of dominant pelagic biogenous grains.

source. The gravel may be a product of cliff erosionduring an early Miocene transgression in the northernAtlas.

Occasionally, difficulties arise when separating thedistal facies of ungraded eolian-sand turbidites fromcontourites. Contourites in this region should havefiner median grain size, higher quartz-to-feldsparratios, a general absence of shallow-water fossils, andrare desert-quartz features relative to turbidites. Heavyminerals should have a similar distribution in bothsediment types. Only the well-rounded, ball-shapedspecimens of iron oxides argue for a beach derivation ofmany sand beds.

In summary, Table 3 shows a suite of features whichcan be regarded as not significantly influenced by grain-size sorting during transport and as indicative of thesource environment. These include the proportions offeldspar, stained and frosted quartz, the composition ofreworked shallow-water fossils and nonbiogenousgrains, and to a lesser extent the abundance of plantfibers and fecal pellets, and the differences in fabrics.Eolian-sand turbidites lack any gradation structuresbecause of the high degree of sorting of their sourcematerial, i.e., the dune sands.

Of all sand-facies types, the eolian-sand turbidites arethe only layers with high porosity (some 40% to 55%)on the West African continental rise. The importanceof their potential storage capacity for hydrocarbons hasbeen proved by a high initial gas content, which wasdescribed from Core 139-7 by Hayes, Pimm, et al.(1972).

HISTORY OF NEOCENE PALEOENVIRONMENTIN SAND LAYER-SOURCE AREAS

The ModelThe West African continental margin displays a

simple semi-actualistic model on the formation offluvial- and eolian-sand turbidites (Sarnthein andDiester-Haass, 1976), which is extended to the Neogenein Figure 11. Summarizing the model, both thegeneration and the composition of the turbidites arecontrolled by the alternating regimes of glacial andinterglacial stages, which are effective in three majoraspects: (1) Any higher frequency of turbidity currentsis connected to glacial periods with low sea level. Atthese times terrigenous sediments, especially dunesands, can be supplied directly to a shoreline close to

952


i n t e r

latitude of Sites: 35.°20' 32?50'0

30c 26°

g l a c i a l g l a c i a l

posit ion of Sahara dunes

26°23?30' 21?A5' 20° TOO1 U°12?30'

Figure 11. Regional and stratigraphic distribution of different types of sand layers: e.t. = eolian-sand turbidites, f.t. = fluvial-sand turbidites, cont. = contourites. Cores at dotted positions are devoid of sand layers. Hatched oblique lines mark thetime-transgressive shift of latitudes as indicated on top of the figure. Arrows mark phases with major frequency ofturbiditeevents.

the shelf edge and are particularly subject to slidesdown the slope (numerous titles since Fairbridge, 1947;Ericson et al., 1961). (2) The paleoclimatic stagesdetermine the rapidly shifting position of the majorhumid and arid belts. For instance, active desert dunesare confined to zones of < 150, and mostly <50 mmprecipitation. The climate belts, in turn, are reflected bythe different types of fluvial- and eolian-sand turbiditeson the continental rise (Table 3). (3) Eolian-sandturbidites give further evidence of a strong anddominant offshore wind system, mostly Trade Winds,which are responsible for the extensive dune-sandtransport towards the shore. Glacial stages apparentlyresulted in an increase in the vigor of the Trade Windsand in a shift of the whole "Wheel-Round-Latitude"system towards the south by some 6° of latitude (e.g.,Fairbridge, 1965; Elouard et al., 1969; Parkin, 1974;Diester-Haass, 1975). Of course, this information onpaleoclimate can only be extracted from the turbiditesif the course of their transport channels is sufficientlyknown. The residence time of the desert material on the

sea bottom from crossing the shore line to its finaldeposition as a turbidite is probably short because thedune sand once displaced as a turbidity current will notundergo this mode of high energy transport a secondtime. Intermediate depositional phases in canyons arealso rather improbable on a large scale (details inSarnthein and Diester-Haass, in press).

However, a reconstruction of Neogene environmentsat first requires additional information. It requires atleast a crude approach in reconstructing the shift ofgeographic latitude as suggested by lithospheric platemotions. Figure 11 shows a palinspastic reconstructionusing the models of Franchetau (1970). This modelgives an overall northward shift of Africa by some 6° to7° within the past 20 to 22 m.y.

Neogene Phases of Low-Sea Level andSouthward Shift of the Sahara Belt

The eolian-sand turbidites clearly occupy a time-transgressive zone of latitudes (Figure 11), that forms anatural extension of the late Pleistocene synglacial

953

M. SARNTHEIN

TABLE 3

Sand Layer Type

Examples

Bed thickness (cm)Sedimentary structures

% >63µm% 44-63µmGravelsMd grain sizeof quartz plus feldsparin sand fraction (µm)Quartz-feldspar ratio(x:l) (125-250 µm)

(250-500 µm)% stained quartzin total quartz (125-250µm)

(250-500µm)% frosted quartzin total quartz (125-250µm)

(250-500µm)% fecal pellets (125-250µm)

% plant fibers (250-500µm)Shallow-water fossils

Shallow-water nonbiogenousgrains

Preservation of carbonategrainsOther criteria

Contourites

138-1/2, 138-2,368-17, ?367-4/3

<l-6Ungraded sandlenses

<5-30Up to 30No

70-90

High: 5-250-

3-10(generally low,depends on source)

-

0-9-

0-0,8

0-0, 25Very rare

Oöids, roundedDebris

Bad

Little mica

Sandy Turbidites From

River Deltas

367-1 and, 370-2to5,?135-SWl and 2

Mostly < 1-6, up to 35,ungraded and graded,partly laminated,

6-603-31occasionally (S.135)

70-95(Site 135:500-1000)

>70-

2-27

(Site 135: 28-31)

--

8-50(Predominantly con-sisting of silty muds)0,5-5,0Pro-delta faunal comm:thin-shelled mollusks,ostracodes, benthonicforaminifera, echinoidsSemilithified deltaicsilty mud clasts;limo-nitic siltite;glauconitepelletsOptimal

Mica is common

Eolian Dunes(Proximal

Deposition)

139-6 and 7,140A-1 and 2

367-3 to 5

> 1-6, up to 150

Eolian Dunes(Distal

Deposition)

138-1/3 to 6

6-12,5Massive, structureless, some un-graded, unlaminated coarse-tailgrading,35-970-5No

115-300

4,3-9,02,5-9,0

14-68

14-77

10-151243

-

-

34-408-20No

90-115

5-95

10-17

?15

13-19?15

-

Hard-bottom fauna from beachand inner shelf: ]aelecypods, barnacles,large and small foraminifera

Superficial oöids.sandstone debris

Medium to bad

Almost no mica

beachrock, limonite-

Medium to bad

Almost no mica

Sahara-desert belt back into the Tertiary. The zoneincludes the Pliocene sands of Cores 367-3 through 5,which occur on the outside farther south because ofdiverting channel directions on the continental rise.

The sand layers of all remaining cores and epochs areidentified as of mainly fluvial origin.

Turbiαites ana comourites appear to De concentratedonly in certain horizons: in the early Oligocene, theearliest Miocene, the middle to late Miocene boundary,and the late Pliocene and Pleistocene. This seems tohold true for all sites except for 370, even when theshortage of core recovery is considered. The persistentturbidites of Site 370 are probably the result of a specialhinterland situation with the rising High AtlasMountains.

The distribution of turbidites (Figure 11) appearssignificant with respect to the above model. Theturbidites south of 20°N are interpreted as a repeatedcoincidence of low sea level, strong off-shore TradeWinds moving dunes to the shoreline, and a generalsouthward thrust of the Paleo-Sahara, such as during

the glacials. Such short-term climatic events would dateat approximately 23 to 20, 13 to 12, and 3 to 2, andperhaps also at 38 to ?34 m.y.B.P. The composition ofthe Pleistocene turbidites (fluvial sand) from Cores 367-1 and 2 characterizes the source area as a synglacialenvironment along a desert margin analogous to themodern Senegal River mouth.

The hypothesis of several drastic climatic deteriora-tions in the Neogene is supported by details of the sand-layer composition. The high proportion of glauconiteparticles in most turbidites can be related to intensiveerosion of older shelf sediments, which occurred bestduring periods of low sea level and exposed shelves.Further, laminated coarse siltites are a commoncomponent of pelagic sediment debris. They are inter-preted as erosional remains of typical wind-dust beds atthe continental margin and point to a contemporane-ous or slightly earlier period of dust storms. Inaddition, siliceous fossils and fish remains are thedominant biogenous component in both early Mioceneand early Oligocene contourites and eolian-sand

954


turbidites. These might indicate an increased coastalupwelling activity related to the increased off-landwinds, using the analogy with Pleistocene glacials(Diester-Haass, 1976). These results agree well withdata from Schrader and Diester-Haass (this volume)from Site 369. The concept of increased upwellingcausing cool temperatures along the continental marginalso agrees with the composition of the reworkedshallow-water fossils. This group represents atemperate-latitude association of the "foramol" type(Lees, 1975), which is analogous to those of the modernenvironment off West Africa which is influenced byupwelling. Finally, the early Miocene contourites ofSite 368 mark a period of increased bottom-currentactivity. They may have originated from vigorousintrusions of Antarctic Bottom Water such as werecommon during Pleistocene climatic deteriorations(Gardner, 1975).

The proposed arid, cool-temperature, low sea levelphases cannot be very well substantiated by evidencefrom land geology. Deserts, in general, and especiallysuch short-time glacial-climatic events, can rarely bedated geologically (average duration of Pleistoceneglacial stages 10 to 20 thousand years). Further,regressions result in erosion and leave behind fewtraces, mainly hiatuses. Most of the late Neogene wasformed as "continental terminal" devoid of anyfeatures amenable to precise dating or records ofenvironmental history. Nevertheless, extensiveunconformities and associated phases of sea-levellowering were reported from several parts of northwestAfrica (Ratschiller, 1970; Dillon and Sougy, 1974; Lietzand Schmincke, 1975). These unconformities date at 27to 20, 6.5 and 2.5, possibly also at 40 to 38 m.y.B.P.which correlate with the ages derived from the results ofthis study.

Neogene desert sandstones are rarely encountered.One such deposit is the upper Miocene of SouthernMorocco in the El Aaiun Basin, in a northerly"interglacial" position (Ratschiller, 1970).

Finally, east-west barriers against the north-southmixing of mammalians (e.g., antilopes) have been notedin North Africa by paleontologists, especially before 20m.y.B.P., and possibly at 10 and 6 m.y.B.P. (CharlesSmart, personal communication). These barriers mayhave been the proposed Paleo-Sahara belts.

Widespread evidence of the proposed Neogeneevents can be found in the marine record if one acceptsa global control of climatic changes. Berger andWinterer (1974) and van Andel (1975) report two mainfluctuations of the carbonate compensation depth(CCD) taking place at approximately 27 to 20 and 13m.y.B.P. occurring simultaneously in the differentocean basins. Similarly, Schlanger and Douglas (1974)note well-documented seismic reflectors in large partsof the equatorial Pacific Basin which are related to thedegree of carbonate cementation, and are dated invarious DSDP cores as 26 to 21, 14 to 12, 6 to 5, and 3m.y.B.P. old. They tentatively correlate the horizonswith major glacio-eustatic phases of the growingAntarctic and Arctic glaciations. This seems to berather confirmed by recent oxygen-isotope curves fromsouthern Pacific DSDP sites (Shackleton and Kennett,

1974). They show three phases of surface-watertemperature minima dated as Oligocene to lower earlyMiocene, early late Miocene, and middle Pliocene. Theevents are also correlated with ice-cap formation onAntarctica using estimates of bottom water tempera-tures. Bottom waters are thought to be responsible forthe generation of cold Antarctic Bottom Waterintruding the mid Atlantic. Summarizing the lateNeogene climatic events, Berggren and van Couvering(1974) emphasize the significance of a cooling phase at3.0 m.y.B.P., when the first ice-rafted sedimentsappeared in the North Atlantic. Figure 11 shows noturbidites in response to the worldwide "Messinianevent" of 6 to 5 m.y.B.P. (Ryan et al., 1974). This ispossibly due to an unusually spotty core coverage inthis short stratigraphic section.

Still, many of these correlations might be ratherhypothetical and incidental, even though they show anumber of impressive agreements. The findings on apaleo-Sahara belt during stages of low sea level,increased off-shore Trade Winds, and upwelling addnew inputs to discussions of Neogene climatic history.However, a full judgment of the evidence andinterpretations in this paper is only meaningful whenincluded with the related results from the continuouspelagic sections of Sites 366 and 369 (Diester-Haass,this volume).

ACKNOWLEDGMENTS

This investigation was supported by stimulating discussionsand suggestions by a great number of people. I am particu-larly grateful to E. Seibold, Kiel, to many members of theCLIMAP group at the Lamont-Doherty GeologicalObservatory, and to L. Diester-Haass, Kiel. Sincere thanksare extended to E. Walger and M. Barton, who criticallyreviewed the manuscript. The author received valuabletechnical assistance from M.R. de Voe, J. Mienert, A.Scholtis, and B. Salomon.

Much of this work was carried out during a one-year visitat the hospitable Lamont-Doherty Geological Observatory,N.Y., under the generous funding of the DeutscheForschungsgemeinschaft (Grant Sa 207/4). Finally, I wish toacknowledge the shipboard party for making Leg 41 samplesavailable for this shore-lab study.

REFERENCESBagnold, A., 1941. The physics of blown sand and desert

dunes: New York (Methuen and Co.), p. 265.Berger, W.H. and von Rad, U., 1972. 26. Cretaceous and

Cenozoic sediments from the Atlantic Ocean. In Hayes,D.E., Pimm, A.C., et al., Initial Reports of the Deep SeaDrilling Project, Volume 14: Washington (U.S. Govern-ment Printing Office), p. 787-954.

Berger, W.H. and Winterer, E.L., 1974. Plate stratigraphyand fluctuating carbonate line. In Hsü, K.J. and Jenkyns,H.C. (Eds.), Pelagic sediments on land and under sea: Intl.Assoc. Sedimentol. Spec. Publ. 1, p. 11-48.

Berggren, W.A. and van Couvering, J.A., 1974. The lateNeogene: Dev. Palaeont. Strat. 2: New York (Elsevier).

Birch, G.F., 1973. Phosphorite deposits off the southwesternCape coast and their relationship to the unconsolidatedsediment: Prog. Rept. 1973, S. African Nat. Comm.Ocean. Res. Mar. Geol. Progr. Univ. of Cape Town, Dept.of Geol., v. 6, p. 16-23.

955

M. SARNTHEIN

Bouma, A.H., 1962. Sedimentology of some flysch deposits.A graphic approach to facies interpretation: New York(Elsevier), p. 168.

Bouma, A.H. and Hollister, CD., 1973. Deep ocean basinsedimentation: SEPM Pacific Sect. Short Course,Turbidites and deep-water sedimentation, p. 79-118.

Carter, R.M., 1975. A discussion and classification ofsubaquaeous mass-transport with particular application tograin-flow, slurry-flow, and fluxoturbidites: Earth-Sci.Rev., v. 11, p. 145-178.

Diester-Haass, L., 1975. Late Quaternary sedimentation andclimate in the East Atlantic between Senegal and CapeVerde Islands: "Meteor" Forsch.-Ergebn. C, v. 20, p. 1-32.

, 1976. Sediments as indicators of upwelling: III. Int.Symp. Upwelling Ecosystems Kiel, Proc: Heidelberg(Springer-Verlag).

Dillon, W.P. and Sougy, J.M.A., 1974. Geology of WestAfrica and Canary and Cape Verde Islands. In Nairn,A.E.M. and Stehli, F.G. (Eds.), The ocean basins andmargins, The North Atlantic: New York (Plenum Press),v. 2, p. 315-390.

Elouard, P., Faure, H., and Hebrard, L., 1969. Quaternairedu littoral Mauretanien entre Nouakchott et Port-Etienne:Assoc. Sénég. Quatern. Ouest-Afr. Bull., v. 23, p. 15-24.

Ericson, D.B., Ewing, M., Wollin, G., and Heezen, B.C.,1961. Atlantic deep-sea sediment cores: Geol. Soc. Am.Bull., v. 72, p. 193-286.

Fairbridge, R.W., 1947. Coarse sediments on the edge ofcontinental shelf: Am. J. Sci., v. 245, p. 146-153.

, 1965. Eiszeitklima in Nordafrika: Geol. Rdsch.,v. 54, p. 399-414.

Franchetau, J., 1970. Paleomagnetism and plate tectonics:SIO Reference 70-30, p. 345.

Fritz, St.J. and Pilkey, O.H., 1975. Distinguishing bottomand turbidity current coarse layers on the continental rise:J. Sediment. Petrol., v. 45, p. 57-62.

Gardner, J.V., 1975. Late Pleistocene carbonate dissolutioncycles in the eastern equatorial Atlantic. In Sliter, W., Be,A.W.H., and Berger, W. (Eds.), Dissolution of Deep-SeaCarbonates: Cushman Found. Foram. Res., Spec.Publ. 13, p. 129-141.

Hayes, D., Pimm, A.C., et al., 1972. Initial Reports of theDeep Sea Drilling Project, Volume 14: Washington (U.S.Government Printing Office).

Hebrard, L., Faure, H., and Elouard, P., 1969. Age absolu duvolcanisme quaternaire de Dakar (Senegal): Assoc. Sénég.Quatern. Ouest-Afr. Bull., v. 22, p. 15-19.

Hjulström, F., 1939/1955. Transportation of detritus bymoving water. In Trask, P.D. (Ed.), Recent marine sedi-ments: Am. Assoc. Petrol. Geol. Symp., p. 5-31.

Hollister, CD. and Heezen, B.C., 1972. Geologic effects ofocean-bottom currents: western North Atlantic. InGordon and Breach (Eds.), Studies in physical oceanog-raphy: v. 2, p. 37-66.

Koopmann, B., Lees, A., Piessens, P., and Sarnthein, M., inpreparation. Recent silty marls and skeletal sands of theBaie du Lévrier, Mauritania.

Lange, H. and Sarnthein, M., 1970. Glaukonitkörner inrezenten Sedimenten des Persischen Golfs: Geol. Rdsch.,v. 60, p. 256-264.

Lees, A., 1975. Possible influence of salinity and temperatureon modern shelf carbonate sedimentation: Mar. Geol.,v. 19, p. 159-198.

Lietz, J. and Schmincke, H.U., 1975. Miocene-Pliocene sealevel changes and volcanic phases on Gran Canada(Canary Islands) in the light of new K-Ar ages: Paleo-geography, v. 18, p. 213-239.

Parkin, D.W., 1974. Trade-wind during the glacial cycles:Roy. Soc. London Proc, Ser. A, v. 337, p. 73-100.

Radczewski, O., 1939. Eolian deposits in marine sediments. InTrask, P.D. (Ed.), Recent marine sediments: Am. Assoc.Petrol. Geol. Symp., p. 496-502.

Ratschiller, L.K., 1970. Lithostratigraphy of the NorthernSpanish Sahara: Mem. Museo Tridentino Sci. Nat., v. 18,p. 9-84.

Ryan, W.B.F., Cita, M.B., Dreyfus-Rawson, M., Burckle,L.H., and Saito, T., 1974. A paleomagnetic assignment ofNeogene stage boundaries and the development ofisochronous datum planes between the Mediterranean, thePacific and Indian Oceans in order to investigate theresponse of the world ocean to the Mediterranean "salinitycrisis": Riv. Ital. Paleontol., v. 80, p. 631-688.

Sarnthein, M., 1971. Oberflachensedimente im PersischenGolf und Golf von Oman. II. Quantitative Kom-ponentenanalyse der Grobfraktion: "Meteor" Forsch.-Ergebn. C, v. 5, p. 1-113.

Sarnthein, M. and Diester-Haass, L., in press. Eolian-sandturbidites: J. Sediment. Petrol.

Sarnthein, M. and Walger, E., 1973. Classification of modernmarl sediments in the Persian Gulf by factor analysis. InPurser, B. (Ed.), Persian Gulf symposium: New York(Springer-Verlag), p. 81-98.

, 1974. Der aolische Sandstrom aus der W-Saharazur Atlantikküste: Geol. Rdsch., v. 63, p. 1065-1087.

Schlanger, S.O. and Douglas, R.G., 1974. The pelagic ooze-chalk-limestone transition and its implications for marinestratigraphy: Int. Assoc. Sediment. Spec. Publ., v. 1,p. 117-148.

Schwarz, H.-U., Einsele, G., and Herm, D., 1975. Quartz-sandy, grazing contoured stromatolites from coastalembayments of Mauritania, West Africa: Sedimentology,v. 22, p. 539-562.

Shackleton, N.J. and Kennett, J.P., 1974. Paleotemperaturehistory of the Cenozoic and the initiation of Antarcticglaciation: Oxygen and carbon isotope analysis in DSDPSites 277, 279, and 281. In Kennett, J.P., Houtz, R.E., etal., Initial Reports of the Deep Sea Drilling Project,Volume 29: Washington (U.S. Government PrintingOffice), p. 743-755.

Van Andel, T., 1975. Mesocoic/Cenocoic calcite compensa-tion depth and the global distribution of calcareoussediments: Earth Planet. Sci. Lett., v. 26, p. 187-194.

APPENDIX ASummary of Coarse Fraction Analysis

(detailed explanation see also text)

Note: biog., biogenous; comp., component; qu., quartz; sed.,sediment; w., water;A, arenaceous foraminifera; B, barnacles; Br, bryozoans; D,decapoda; Dt, diatoms; E, echinoids; F, benthonic foraminifera; Fi,fish debris; G, gastropods; O, ostracodes; P, pelecypods; PI,planktonic foraminifera; R, radiolarians; S, sponges;cont., contourite; e.t., eolian-sand turbidite; f.t., fluvial-sandturbidite. Changes of grain-size fraction of data as indicated incolumns.

956

M. SARNTHEIN

Sample(Interval in cm)

135-SW1-1135-SW1-2135-SW2138-1-2, 10138-1-2, 133-135138-1-3, 106-108138-14,83-85138-1-5,50-52138-1-6,50-52138-1-6, 50-52138-1-6, 120138-2-1,2-3138-2-1,80-81138-2-1, 139-141139-7-1,20-21139-7-1,84-85139-7-1, 147-148139-7-2,45139-7-2,96-98139-7-2, 133-134139-7-3,38-39139-7-3,58-59139-7-3, 146-147139-74,48139-74, 88-89139-74,94-95139-7-5,49139-7-5, 87-88139-7-6,4244139-7-6,119-120139-7-6, 140-141140A-1-1, 84-86140A-2-6, 127-128

367-1-2,98-100367-1-2, 120-122367-1-3,100-101367-1-3, 126-128367-1-5,135-138367-2-1,85-87367-2-1, 131-133

367-3-1,2367-3-1,4243367-3-1,80367-3-1, 107-108367-3-2, 7367-3-2, 35367-3-2, 98-99367-3-2,113-1153674-1,29-303674-1,82-833674-1, 1103674-1, 123-1243674-1,1293674-1,1313674-3, 10136744, 57-5836744,64-6536744,933674-5,62-63367-5-1,4367-5-1,62367-5-1,98-99367-5-1, 119368-17-5, 138-142368-22-1, 17-18368-22-2, 12-57

370-2-2, 35

Λ

95.445.024.7

2.68.3

36.133.738.4

5.839.637.929.718.114.853.757.783.320.067.046.166.549.976.661.284.269.770.387.567.089.779.629.870.4

37.853.153.644.639.139.813.3

37.087.787.895.078.486.382.291.483.879.575.676.180.992.648.594.189.947.535.697.497.797.495.822.322.611.3

29.0

E

•tf

S§

0.47.72.37.6

16.820.414.419.1

1.98.16.6

4.44.52.61.83.35.12.42.72.00.11.92.52.41.83.21.81.91.02.0

30.725.130.021.922.326.629.1

4.31.00.70.20.70.50.60.40.80.70.80.91.00.62.60.30.30.80.90.90.10.30.6

29.020.6

5.4

4.3

a.

-αc

-aS

700975560

9297

101103103110111113

797175

141157139136225120150160167165170167167166155170180175230

75909072958783

160141165285167175175187160230180280215165117190165145145115130141125

89165115

260

'εa.O

•β

800

51080839092

100104104110797175

141157139136225120150160167165170167167166155170180167245

907678

859083

16014116528516717517518716023018028021516511719016514514511513014112589

165115

260

TerrigenousComponents

©©

©in<N

52.223.491.462.542.920.753.253.339.279.427.7

0.4

99.298.599.899.899.699.799.499.399.698.298.799.399.499.483.899.399.843.797.0

1.30.4

0.5

95.579.688.798.094.496.695.199.088.692.495.295.687.873.166.798.898.694.585.610.830.690.441.1

72.5

1.2

a.oln<N

64.210.117.269.161.789.495.096.197.797.198.345.9

8.64.7

96.198.699.398.899.498.799.198.898.999.197.099.097.196.896.398.698.665.696.4

7.120.527.6

20.822.3

5.9(500-

1000 µm)71.430.760.095.077.088.788.899.272.692.992.990.578.7

98.697.178.665.0

100.0

0.3

Shallow-WaterFossils

a.ooin

©in

0.7

0.7

2.94

51.31.2

0.2

1.04.32.51.2

0.30.90.60.52.51.61.30.61.00.6

7.2

7.617.13.63.5

a.

in

0.51.4

0.21.60.2

0.3

0.3

0.50.3

0.20.2

28.42.3

2.83.61.0

0.82.90.2

(500-1000 µm)

21.611.3

3.51.92.33.62.54.51.03.90.63.2

0.7

a.

o

ó

4.98.91.8

25.042.644.815.620.0

8.72.9

12.0100.084.871.7

0.3

0.30.4

0.90.5

0.20.3

0.23.70.9

93.773.586.0

9.462.010.1

2.912.4

6.3

1.81.21.90.38.83.51.32.46.9

20.633.3

0.20.62.2

11.971.841.4

4.645.5

42.8

PelagicFossils

a.o

3.36.6

12.54.6

10.32.20.80.60.90.40.6

50.873.371.1

2.81.0

0.50.40.80.5

0.5

0.80.41.71.00.90.60.93.8

71.248.348.7

29.432.325.2

(500-1000 µm)

31.815.50.61.92.90.90.8

22.33.30.92.83.5

100.0

0.30.7

33.3100.083.3

85.7

56.1

% NonbiogenicShallow-Water

Sed

a.oi / i

σ

CJ

0.5

14.313.818.826.7

1.2

0.42.2

0.2

0.2

0.20.3

0.20.20.41.0

0.2

3.50.5

11.880.4

0.3

0.3

0.6

0.30.30.30.36.6

0.20.3

0.30.41.80.50.7

1.0

14.7

. Debris

a.oir ,

in<N

0.2

3.75.93.41.91.90.20.80.20.32.53.00.2

0.50.7

0.2

0.50.31.40.60.61.00.90.30.30.2

1.11.82.3

0.46.6

59.6

(500-1000 µm)

2.30.5

1.31.01.42.4

1.91.82.22.2

0.32.2

14.3

% PelagicSed.

a.oin

©

41.766.8

6.012.5

20.712.5

52.214.759.0

12.526.2

0.81.30.2

0.70.21.50.5

0.4

14.90.7

0.80.9

0.3

1.04.8

0.73.52.5

3.51.21.90.30.41.91.91.24.04.1

0.63.41.29.49.10.88.7

25.5

41.1

Debris

a.o<N

l/->

<N

31.883.168.822.220.9

2.51.71.10.70.80.60.2

12.420.6

0.30.30.2

0.30.20.9

0.60.5

0.21.02.00.4

1.71.0

1.70.3

1.70.8

(500-1000 µm)

13.713.91.9

18.05.25.20.40.62.90.61.9

12.5

21.41.6

16.7

14.3

29.3

958


% FecalPellets

(UI

a.ooò<N

0.6

1.2

1.2

0.2

0.2

2.824.913.6

90.125.2

4.6

0.2

0.5

'εa.

o>/•>

fN

fN

0.3

0.40.80.3

0.80.3

0.2

16.125.419.8

48.633.5

8.1

% Feldspar

ε=L

OOupol/•>

<N

23.325.516.6

17.3

14.917.917.1

9.517.411.017.210.020.814.017.321.322.021.122.822.4

31.735.538.831.526.525.034.126.327.028.520.324.825.128.133.325.629.717.620.910.020.120.832.2

16.0(63-125

µm)1.9

"εa.

Ot o

>Λ

~

35.630.025.015.911.812.712.512.110.016.2

9.99.5

25.711.810.510.418.512.313.114.014.511.416.212.910.713.715.316.618.715.715.6

(500-1000)25.012.030.719.519.518.324.222.212.310.112.5

9.711.0

19.714.418.243.3

% StainedQuartz

?a.

ooó

31.631.428.1

(50.0)(25.0)(14.3)(28.5)( 4.8)(10.0)

15.715.519.518.621.120.520.724.124.721.023.215.514.520.320.519.516.522.517.1

(63-125µm)33.530.030.8

33.523.023.0(250-500)68.563.965.262.065.564.666.661.553.665.869.563.270.053.750.075.867.352.576.968.956.563.170.0

3.8(63-125

µm)22.0

?a.cI/OCN

i/nr•i

1̂

19.117.027.8

7.49.4

10.712.413.511.215.517.010.4

3.74.67.8

11.910.314.313.518.213.223.912.418.212.113.2

9.17.7

19.48.99.0

19.110.4

20.027.026.0

8.020.013.3(500-1000)66.672.682.382.084.177.369.876.075.081.082.376.480.0

84.087.187.6

100.0

100.0

% FrostedQuartz

ia.ooó

7.28.66.3

(16.7)(28.5)(43.0)(14.3)(15.0)

14.812.117.421.427.624.919.319.719.228.820.514.815.713.214.513.411.923.524.5

31.119.113.028.426.434.232.426.342.937.538.941.225.524.2

24.240.924.131.042.214.824.118.3

19.0

?a.oLO

1.8

9.56.6

14.914.119.415.412.816.14.17.40.09.8

17.414.314.914.612.011.511.420.712.516.713.011.9

7.79.99.1

12.216.119.9

(500-1000 µm)

33.318.219.028.742.847.848.043.468.056.855.548.358.0

44.052.267.676.6

100.0

Quartz + Feldsp.Mica

εa.i/i

" ^

o

872

434

153

450

282486628235

434208325

665083

757454

9000

6500

16200

44009300

10000

8000

3000

9000

106

ε~a.

©m<N

uo<N

1722

142113445210493

2811.5

270

638572

719.526

2022

6

52004900

3800

620050

1

α %

IIC wo u

( j •g, •>.

e <s s.5 oε ~-Λa 3Q £

E/P/•

-/-/-P/F/G

-/-/-P/P/G/F

p/•

P/

F/P/

P/F/F/F/P/DF/Pl/

F/F/P/B/EP/F/

P/F/P/F/DP/F/

P/P/F/

P/

F/P/P/F/BP/F/BF/B/P

B/P/P/G/P/F/BF/P/BF/P/BP/B/FF/B/PG/B/PB/Br/FP/E/

F/P/

P/Br/EP/Br/FP/F/DP/G/B

G/

g

d.

<siC O

Q CL,

Pl/G/FPl/F/RPl/F/R

Fi/Dt/Fi/F/PFi/P/PlFi/P/OR/O/FiR/F/EFi/R/PF/R/SR/S/FiR/S/Fi

R/Fi/SDt/F/RDt/R/SP/Dt/R

F/Dt/R/FiDt/Fi/PDt/R/FiR/Dt/R/P/Fi

R/Dt/P/FR/Dt/PlR/Dt/FR/F/DtR/Dt/FDt/F/DR/Dt/F

F/Pl/B

R/

F/Pl/OPI/F/PF/Pl/P

Pl/F/FiPl/F/OPl/F/S

F/Pl/GPl/F/EPI/F/Pl/F/EPl/F/FiF/Pl/PPl/F/Pl/Fi/FPl/F/AF/Pl/PPl/F/Pl/F/Pl/F/APl/F/PPI/PI/F/Pl/F/

PI/ A/Pl/F/Pl/P/FP/Pl/FP/F/PlPl/P/F

Fi/

Fi/

Pl/F/A

T3

eCΛ

o(U

f

f.t.f.t.f.t.

cont.cont.e.t.e.t.e.t.e.t.e.t.e.t.cont.cont.cont.e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.

f.t.f.t.f.t.

f.t.f.t.f.t.

e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.cont.?e.t.e.t.e.t.e.t.e.t.e.t.e.t.e.t.

cont.cont.cont.

f.t.

959

Deep Sea Drilling Project Initial Reports Supplement to Volumes … · 2007-05-11 · volume of the samples was 5 to 10 cc. The samples were wet sieved using 44 and 63 µm sieves;

Documents