Quaternary Deserts and Climatic Change, Alsharhan, Glennie, Whittle & Kendall (eds)© 1998 Balkema, Rotterdam, ISBN9054105976

Late Quaternary climate changes in the Central Namib Desert, Namibia

Klaus HeineDepartment of Geography. University, Regensburg, Germany

ABSTRACT: The hyper-arid central Namib Desert represents a flat gravel plain (Namib Unconformity Surface)that rises from the coast in the west to about 1000 m a.s.!. in the east. In the south the Kuiseb River gorgeseparates the plains with inselbergs from the Namib erg. The aridity along the Namib coast is caused by thecold Benguela Current, which has persisted since Tertiary times. Little is known about late Quaternary climatechanges in the central Namib Desert. Here we refer to investigations on Kuiseb River sediment sequences(Homeb Silt Formation), gypsum soils of the Namib Unconformity Surface, and speleothems of Namib cavesto discuss their paleoclimatic evidence. According to geomorphologic, sedimentologic, paleopedologicevidence and information of speleothem growth in combination with 14C, UrTh, and TL age determinations,the central Namib Desert experienced an arid climate during the late Quaternary. No evidence was found thatthe major late Quaternary climate changes affected the central Namib area. We conclude that the cold BenguelaCurrent was stable off the central Namib coast, thus producing the conditions for the aridity throughout the lateQuaternary. UiTh dates point to some fluctuations in available moisture during and before the isotope stage 5.

I INTRODUCTION

The occurrence of coastal deserts along the westernshore of most continents has attracted the attention ofscientists for a long time. These deserts are ofdifferent natures; some are hyper-arid while othersare semi-arid. In recent years the age and origin ofthese deserts and their Quaternary climatic changeshave been the focus of much research of aninterdisciplinary nature. The Namib Desert ofSouthwestern Africa (Fig. I) is about 1400 km longand between 40 and 120 km wide. It is characterizedby a climate that has been more or less arid forapproximately 40 million years. Marine pollenassemblages in cores off the coast document aridclimatic conditions along the coast ever sincePliocene times. Whether more humid phasesoccurred in the Namib desert and surrounding areasduring the late Quaternary has been discussedrecently by different authors (Heine 1992, 1995,Rust 1989, 1994, Teller et a!. 1990, Geyh 1995). Ananalysis of the paleoclimatic information fromlandforms, palesols, sediments, and speleothems insouthwestern Africa shows that only few dataconcerning the climatic history are useful (Heine1995). Many problems hamper the geochronologicaland paleoclimatical interpretation of the 14C, Ll/Th,and TL dates (Geyh 1995). By comparison, theclimatic development during the Holocene is ratherwell established. But for the Upper Pleistocene, noexact information can be given because of difficultieswith the absolute age determinations. nopaleoclimatic information can be obtained for the time

before 40 ka BP based on 14C dates, and before 100ka BP based on TL dates. All paleoclimatic dataobtained in recent years from calcretes, fossilhorizons, speleothem etc. point to a relatively aridclimate in the central Namib Desert, at least since thelast glacial maximum, this part of the Namib Desertwas not affected by more than normal rainfal!.Fluctuations in precipitation are superimposed, butthese fluctuations did not lead to a decisive change ofthe general climatic regime.

Here we show that in the hyper-arid Namib Desert,in the area between the Kuiseb and Swakop valleysas well as in the area of the Rossing Mountains andthe Tinkas Flats, late Quaternary fluctuations inprecipitation did not affect the desert environment.

2 THE STUDY AREA

The hyper-arid central Namib Desert borders on thecold Atlantic Ocean to the west and a dissectedplateau slope beneath the Escarpment (which has anaverage altitude of 1500 m, rising to 2300 m inplaces) some 160 to 180 km to the east (Lindesay &Tyson 1990) (Fig. 1).

The Swakop and Kuiseb River valleys are deeplyincised into bedrock. The latter separates therelatively flat gravel plains with inselbergs (Rossingmountains) to the north from the coast-parallel lineardunes of the sand sea to the south. The plain of thecentral Namib Desert rises from sea level to about1000 m a.s.!. at the foot of the Great Escarpment.This plain is called the Namib Unconformity Surface.

293

It was formed by long continued erosion across thePrecambrian rocks, thus separating the metamorphicbedrock from all younger deposits (Oilier 1977).

The mean annual rainfall of less than 20 mmlyearalong the coastal belt makes the Namib Desert one ofthe driest areas in the world according to commondefinitions of a desert (Fig. 2). Rainfall increasesfrom the coast towards the interior changing. to asummer rainfall zone on its eastern edge.' Thecombination of cool air masses under the influence ofthe cold Benguela Current, together with the highpressure conditions in the subtropics is mainlyresponsible for the scanty rainfall. Apart fromrainfall, precipitating fog is the second source ofmoisture that sustains life in the Namib. Rainfall overthe central Namib Desert occurs mainly in the formof convective summer storms (Sharon 1981) fromwhich maximum precipitation is received over theExcarpment to the east, while fog precipitation is thedominant moisture source over the western parts ofthe desert adjacent to and inland from the coast(Lancaster et al. 1984, Lindesay & Tyson 1990).

3 PALEOCLIMATIC INFORMATION FROMCENTRAL NAMIB SITES

3.1 The Homeb Silts

The Kuiseb drainage system rises in the KhomasHighlands to cross the central Namib Desert andreach the coast near Walvis Bay. Near Homeb, in therock-walled canyon of the Kuiseb, some relics ofalluvial terraces are preserved (Fig. 3). The Cenozoicdeposits of the Kuiseb valley have been documentedby Ward (1987) and Ward & Corbett (1990). Thesediments form isolated outcrops hanging from therock walls of tributary valleys. The maximumthickness is more than 25 m. The most elevatedsediments are some 45 m above the present riverbed. Vogel (1982) dates the Homeb Silt Formation tosome 20,000 years BP. The Homeb Silt Formationconsists of eroded remnants of fine-grained alluviumdeposited as flash-flood sediments that werecolonized by successive generations of opportunisticarthropods (Smith et al. 1993). The stacked floodunits each comprise a massive tabular siltstone bedoverlain by a thinner interval of rapidly alternatingsandstone and siltstone with claystone veneers.These flood units are interpreted as havingaccumulated under semi-arid climatic conditions byepisodic back-flooding of the Kuiseb River intoembayments and tributary mouths (Smith et al.1993). Soon after deposition, as the floodwaterssubsided, the sediment was colonized by burrowing,sediment-ingesting organisms, mostly arthropods,that produced a Taenidium ichnofacies. After thefloodwaters had drained, the exposed sediment wascolonized by grasses and burrowed by terrestrialarthropods, probably ants and termites, resulting inan overprint of Termitichnus ichnofacies withassociated pelletal chambers (Smith et al. 1993).Thirteen horizons of rhizocretions and root tubules

occur toward the top of the succession and indicate agradual reduction in the frequency of flooding. Theseimmature calcic paleosols suggest that the climate inthe central Namib Desert was semi-arid ca. 20,000years ago, being wetter and more seasonal than thehyper-arid conditions that prevail today (Smith et al.1993).

Although there seems to be little doubt about theobservations by Smith et al. (1993) that the HomebSilt Formation proves semi-arid climatic conditions,some fundamental questions arise. Previous workershave offered several other explanations of thecharacteristic Homeb sediments (e.g. Scholz 1972,Rust & Wieneke 1980, Marker & Muller 1978,Vogel 1982, Ollier 1977). They were variouslyinterpreted as river end-point accumulations, flooddeposits of an aggrading river, or sedimentsdeposited behind a dune dam. Hovermann (1978)postulates a fluvial origin of the silts, and Heine(1987) describes as their mode of origin flash-floodfluvial sediments that were deposited beside the mainchannel flow in overbank areas (mainly in tributaryvalleys). Thus, the Homeb Silt Formation representsnot only the climatic regime of the central NamibDesert, but also the climatic conditions in the upperreaches of the Kuiseb (interior plateau, GreatEscarpment, eastern Namib).

The age of the Homeb Silt accumuluation is datedby 14C as 23-19 ka BP (Vogel 1982) and by TL asabout 18.3 +/- 3.4 and 17.4 +/- 3.8 ka BP (Eitel1994). Although the age determinations differ withina couple of thousand years, the time of accumulationis centered around 20,000 year BP, and thus isroughly synchronous with the last glacial maximum(LGM).

3.2 Soils of the central Namib Desert

According to Scholz (1972), raw mineral soils,calcrete and gypcrete soils, are the most importantsoil types of the gravel plains of the central NamibDesert. In the vicinity of Gobabeb, north of theKuiseb, most of the soils have a light brown to ochrecolor, and are shallow, with a subsurface horizonwhich may consist of calcrete, gypcrete, or salt.

A detailed investigation of the desert soils withrespect to the meso- and microrelief north of theNamib Research Station of Gobabeb (Walter 1994)yields the distinction of different phases withcharacteristic pedogenic processes. A schematic soilprofile of the central Namib gravel plain is shown inFigure 4.From top to bottom, the following soil horizons are

developed: (I) layer of 2-3 cm of eolian material,with a desert pavement (quartz, rock fragments,feldspar, remnants of calcrete, etc. sandy-gravel),composed of a layer of angular to subroundedgravels one or two stones thick sitting on a mantle offiner stone-free silty sand; (2) gypsum crust; in theupper part characterized by thin, discontinuousgypsum coatings on the undersides of the stones(stage I after the classification of Reheis 1987) andby abundant gypsum pendents under stones and

294

260

~ Granite inselbergso Metamorphic

(}c rock inselberqs

Iff. Dunes (Namib'/It Sand Sea)

I Cool fog desert

II Alternate fogdessert

lIT Dessert steppe280

Figure 1. Location map. Climatic zones (after Besler 1972).

gypsum crystals scattered through the matrix (stageII after Reheis 1987); in the lower part a gypcretedeveloped, consisting of a continuous gypsum-plugged matrix, with stones and smaller debrisfloating in the gypsum matrix (stage IV after Reheis1987); (3) crust mixed from cemented gypsum andcalcrete with decomposed bedrock; (4) horizon richin CaC03 or decomposed bedrock with freshgypsum minerals formed by crystallization fromdescending soil water; (5) bedrock or calcrete.In the central Namib Desert the pedogenic gypcretes

result from downward movement of gypsum.Gypcretes are found between the coast and the inlandlimit of the influence of frequent fogs. Theimportance of fog carrying dissolved salts isemphasized by many authors (Scholz 1972, Scholz& Beckmann 1971, Walter 1994). The developmentof very thin gypsum crusts only a few millimetersthick takes a period of at least 10,000 years, as isdocumented by stage I-gypcretes on dated alluvialterraces in the Kuiseb valley near Homeb. On theNamib Unconformity Surface, the gypsiferous soilsare found nearly everywhere. Gypcretes occur evenalong slopes and in drainage channels beneath aloose sandy layer of a few to ca. 20 cm depth; theydocument an inactive relief of the NamibUnconformity Surface with regard togeomorphodynamic processes. Active lowering of

1L.°E 180 200 220

]QQ Isohyetvrnrri/a

'">-'. GreatEscarpment

r

1> ,

-\

.~

/0~ioP-o"'? 100 200 300~

krn

n~e

the surface by erosion and denudation has not beenobserved. Therefore, major fluctuations in humidity(rainfall) cannot have occurred during the lateQuaternary. If they had occurred, they would haveremoved the gypcretes that are observed even in thedrainage channels.

3.3 Speleothems of the central Namib Desert

In the area of the Rossing Mountains and the TinkasFlats, caves have developed in dolomites (Fig. 1).As these dolomitic ridges are more resistant todenudation, they surmount the Namib UnconformitySurface by some meters to several decameters. Thecaves show sinter growth. The speleothem formationdepends on local precipitation since surface and/orgroundwater influx is excluded in the dolomite ridges(Heine & Geyh 1984). The caves themselves musthave developed under a comparatively humid climaticregime. Furthermore, the base level, and thereforethe Namib Unconformity Surface, must have been ata higher elevation than today. The geomorphicevidence concerning the formation and developmentofthe Namib Unconformity Surface shows that theseconditions could have prevailed before the LateMiocenelEarly Pliocene calcrete formation on theNamib Unconformity Surface (Besler et al. 1994).

295

15°

(87) 23°

mm

40L"20

oJ 0 ~Okm

15°b (34)- ....••....Swakopmund (3)(~:4~"1

(80) __

-1> ...••••••. • Vogelfederberg

"-,.Rooibanli-', (l84!\\ .

--', Swartbank /\ (31) ~/

\~ ,'<:)(65) -',.Gobabeb Zebra·Pan/ (\,_.u", /"......,. •........ _- ~/ \

Flodden Moor ------~~ 0(36)

••••••••••• Norobeb

•Ganob

23°

Figure 2. a) Mean monthly and mean annual rainfall in mm for longer term stations. b) Mean monthly andmean annual fog precipitation in mm (after Lancaster et al. 1984).

N

NAMIBUNCONFORMITVSURFACE

5

DUNES

• • CAI..CRETE-+ •\,C; 't: • GRAVEl..PAVEMENT (JJ-28000BP) • ·Z·/",'t.~q-2ll· '.' BASEMENTW.~ITH (WCALMETAfv1ORPHICS) TERRACE511- N -ru-:

---I-'---~ • GRAMADULLAS HO,AAC'BSI'IT WITHROUNDED .' .j-.1,/ 1'/"'"' _ IV/L! L KUISEB GRAVELS • /'Tf?15'u:"'1/ '7-'- (2J-79000flP) AT SURFACE, GRAVEL '/ 480

II/ /.f1RY~trj;, .. j:/ ! fkgY!fE!-NTS(LOCAL j~~Jl-lc'tfK (47;~~~;7)BASEMENT t: • GL<:JOS METAMORPHICS) PARTLY GR4VEL " • I I'

(METAMORPHICROCKS) ---tt.L •..m r- TERRACES CEMENTATION~ 1520~I;OBP, CA.33-28000BP

AEOUANSAND ~~"'c N III ~~~I/'rl 1.1" I //1BASAL SCHIST ["?? n [~ FLOODPLAIN/!'CONGLOMERATE ACTUAL·': ',I,,} ./ I'll /.N'!.;- .(- Q2m THICK) RIVIER BED ",.? + i,» . I

. HOMEB SILT BASEMENTBENEATH GLACISSEDIMENTS72m TERRACE(9600BP? )

530mHEIGHT

505

455

430

Figure 3. Schematic cross-section of the Kuiseb River valley near Homeb. The ages are 14C dates (adoptedfrom Vogel 1982).

Several flowstones, stalagmites, stalagtites, dripcurtains, cave popcorn, and other speleothem wereradiocarbon-dated (Heine & Geyh 1984, Heine1991, 1992, Geyh 1995). The resulting ages are allgreater than 25 ka BP. Compact speleothems wereeven older than 35 ka BP. According to theradiocarbon dates no sinter growth occurred after 25ka BP. (Figs 5, 6).

Additionally, several sinter samples from theRossing and Tinkas Caves were dated by Urrh. Theresults differ considerably from the radiocarbon dates(Figs 5, 6). Furthermore, some TL agedeterminations from sands that were blown into theRossing Cave yield minimum ages of> 64.3 +1- 5.2and> 87.5 +1- 6.5 ka BP (Fig. 5).

An interpretation of the 14C, UzTh, and TL ages

296

desert pavementstone - freesilty sand

gypsum crust(stage I - I])

gyp crete (stage IV)

mixed crust(gypcrele and calcrete)

polymorphic mixed crustwith fragmentsof calcrete)

deca mposed calcrete(gypsum in fissures

decomposed bedrock

bedrock

E

~l@.©

Figure 4. Schematic gypcrete profile of the centralNamib Desert. For explanation see text.

from the caves of the central Namib is difficultbecause of methodological problems (Heine 1992,Holmgren et al. 1994, Geyh 1995). The radiocarbondates are very old and therefore probably rejuvenatedby contamination. In no case do they correspondto the Uffh dates and the TL dates. We conclude thatvery old speleothem (older than the last glacialepoch) were moistened by dropping water that

resulted in fog precipitation or in occasionally rainfallin the central Namib Desert. Recrystallization ofcalcite was possible. All photomicrographs (Plates 1-3) of speleothems show very thin calcite layers ontop of the sinter material. In the lower layers, bigger,mosaic and elongated crystals can be observed.Diagenetically altered layers and iron- andmanganese-rich horizons are common in the outerparts of the speleothems. They document hiatuses inthe development of the sinter. Sharp boundariesindicate periods of dry conditions when theenvironment was not conducive to speleothemgrowth and some dissolution of underlying layerstook place. The vague layer boundaries markerosion-free stoppages in sinter formation. Beneaththe sinter surfaces the layers are very thin, thusindicating a slower growth with many interruptions.The results of the age determinations indicate on the

one hand that speleothems were not formed duringthe late Quaternary apart from the very thin layersthat may document minimal moisture fluctuations.Sinter growth cannot be detected during the periodafter ca. 20,000 years BP. On the other hand, theu/Th analyses of some speleothems yieldcharacteristic Eem (stage 5) and, probably, stage 7ages (Heine 1992, Geyh 1995). The speleothemswith these ages consist of very thin layers.Therefore, we conclude that during the lateQuaternary in particular, and during the Quaternaryin general, calcite precipitation was confined to theformation of extremely thin layers. This documents aprolonged period of extreme aridity in the centralNamib Desert, starting in the Miocene. No phaseswith major shifts to more humid conditions can bepostulated.

The discrepancies between the 14C and Uffh ages

14°47'84"E~~ _tca.330masl. 2r31'SS"S _

II

,?,

® //"~1>3621! ;~

I "~'.: •.··'C}S~."'::::::~. .:..'----

\1>4001

?i',, ,

~ Flowstone.•. Stalagmite"m'" Stalagtite__ Popcorn /

~ Sand ®D~;~~ sa~~gr?~ ~o 14C }D .U/Th ages in ka BPgTL

~$

~ 9 8 lp m

?;

/

/@@81>3501 1123-2151 1146-2051

I~IFigure 5. Rossing Cave, Namib Desert, showing the position of the dated samples. 24 14C dates and 16 Uffhdates of speleothems show the problems of dating (see Heine 1992, 1995, Geyh 1995).

297

KOO 639

[

0 ;,:::~:3ySINTER GENERATION 1

.~®1>2661;5cm~~SINTER GENERATION 2

SINTER GENERATION 3

SINTER SAMPLE FROMLOWERMOST CAVE

SAND (WITHORGANICMATERIAU

-425ma.sL2r4S' 25" S15°03' 55" E .... ;. .

. ' ,

o

2m

~ SINTER WITH SAND1;""·1 SAND~ SINTER~ BED ROCK

o 14C} .D U/TH agesmkaB.P

Figure 6. Tinkas Cave, Namib Desert, showing the position of the dated samples (10 14Cdates, 7 UfTh dates).Bottom, plan of the different cave chambers. Top, stratigraphic section through the cave sediments. Thespeleothems show very thin calcite layers on top (sinter generation 1); in lower layers bigger, mosaic andelongated crystals can be observed (sinter generation 2 and 3).

298

('

t'{.;~

, tit!:

Figure 7. Photomicrograph of Namib speleothem, Rossing Cave. Main stalagmite (see Heine & Geyh 1984).Characteristic rhythmically deposited calcite layers. Extremely thin layers in the outer part (A). Elongated bigcrystals in the inner part (C). Diagenetically altered layers and iron- and manganese-rich horizons as boundariesbetween the sets of layers. The hiatuses document weathering horizons during periods of dry climate. Bar = 1mm. In the Namib cave speleothems, growth layers always occur as thin layers in the outer part and thicklayers in the inner part. The outer layers (A) yield a Urrh age of ca. 123 ka BP, the layers (B) a Uz'Th age ofca. 215 ka BP. Radiocarbon ages of the outer layers are ca. 26 to 33 ka BP.

can be explained by the limited growth of the sinterformations, since the samples for radiocarbon datingwere collected from pieces of sinter that comprisesseveral small calcite layers. AMS 14C and massspectrometry U-series age determinations have notyet been processed from the Namibspeleothems. These tasks are in progress.

4 DISCUSSION AND PALEOCLIMATICIMPLICATIONS

Paleoclimatic information from southwest Africansoils and landforms are discussed by Heine (1995).Based on 14C, U/Th and TL ages, partlycontradictory paleoclimatic concepts have beendeveloped (Heine 1991, 1992, 1995, Geyh 1995).The paleoclimatologic information on the availabledates from the central Namib Desert is not yet

299

iiI

definite because of the methods used and theunsuitability of the dated material (Geyh 1995).Nevertheless, a critical evaluation of the Homeb SiltFormation, the Namib gypcretes, and the Namibspeleothems can contribute to a better understandingof the late Quaternary climatic history of the centralNamib Desert.

The Homeb Silts, for the first time, provideevidence for moister conditions in the central NamibDesert. Without any doubt, the description of Smithet al. (1993) document a more humid environmentthan today for the period of the accumulation of theHomeb Silt Formation, namely the last glacialmaximum. Unfortunately, the 14C and TL ages ofthe Homeb Silts can give only a vague criterion forthe true time of sedimentation, at some time or otheraround the LGM. Since there is no other evidence fora markedly wetter phase in the central Namib around20,000 years BP, we cannot exclude that the moistenvironmental conditions at Homeb, documented bythe ichnofacies and the horizons of rhizocretions androot tubules in the sediment sequence, reflectinfluences from the upper reaches of the Kuiseb riversystem. Mineralogical characteristics of the HomebSilts show that these sediments originated in theKhomas Highland (far in the east) (Eitel 1994).Reconstructing the central Namib paleoclimate byreferring to the Homeb Silt Formation, therefore,should not be considered as a reliable method.Investigations on paleosols in Sossus Vlei,Hauchabfontein, and Bullsport (Heine 1993) indicatethat late Quaternary humid phases cannot be traced inthe central Namib Desert, but become more and moreaccentuated towards Northeast and East (NamibianHighland) where more humid phases compared withtodays climatic conditions occurred> 25 to 19 ka BPand about 10 to 8.5 ka BP (Heine 1993).

These conclusions are corroborated byinvestigations of late Quaternary lake deposits in thenorthern Namib Sand Sea south of the Kuiseb valley(Teller et al. 1990). Teller et al. (1990) believe thatthe region remained hyper-arid throughout the lateQuaternary, and ponding and sedimentation in vleisand interdune valleys were related to small increasesin runoff and groundwater recharge from the East.The age of ponding may be oldest in the west (30-25ka BP, west of Gobabeb) and youngest to the east(17.5 to 11.1 ka BP, north of Tsondab VIe i).

Until now there have been no further detailedinvestigations of the Namib Desert soils. Therefore,there is nothing with which to compare the results ofWalter (1994). Little is known about the origin of thecentral Namib gypcretes. Together with Dixon(l994a, b), we recognize the Namib Desert gypsumsoils as "true gypsum" or "croute de nappe" thatoccur as slightly cemented gypsum crystals up to Imm in length commonly developed beneath non-gypsiferous sediment or as desert rose crustconsisting of interlocking lenticular gypsum crystalsranging in size from a few millimeters to 20 em. Thepaleoclimatic interpretation of pedogenic gypcretes isdifficult, since relatively little work has beenundertaken on the downward movement of gypsum,the source of gypsum (which is generally regarded as

aeolian), the time which is necessary to develop acertain abundance of gypsum in characteristic soillayers, and the climatic conditions necessary forgypcrete formation. Furthermore, only little isknown of the relationship between climaticparameters, time, amount of aeolian input of salts,gypsum, etc. on the one hand and the occurrence,morphology, and mineralogy of gypcretes on theother hand. From our investigations of gypcretes(Walter 1994, Heine & Walter 1996), we concludethat in the central Namib Desert there is no evidencefor humid phases during the late Quaternary.

Clastic cave sediments and speleothems areproviding important new information about pastclimates in arid and semi-arid southern Africa (Brook1995, Brook et al. 1990, Burney et al. 1994,Holmgren et al. 1994, Holmgren et al. 1995, Heine1991, 1992, Heine & Geyh 1984, Geyh 1995).Unfortunately, no investigations were carried out inthe Namib caves apart from the work by Heine(1992), Heine & Geyh (1984), and Geyh (1995).For the interior of southern Africa, despite the datingproblems (see Heine 1991, Holmgreen et al. 1994,Geyh 1995), some conclusions can be reachedconcerning the nature of late Quaternary changes orthe timing of climatic fluctuations. Based onspeleothem investigations, the Namibian highlandsand the Kalahari Basin (Botswana) experiencedwetter conditions in late glacial times (ca. 25-12 kaBP) and more arid conditions in the early Holocene(10-7 ka BP). In the western parts of semi-aridsouthern Africa (Etosha area to the Tsauchab valley),the opposite was the case during the early Holocene(Heine 1995) when conditions became more humid.However, the Ll/Th dates from speleothems showthat the last period with extensive sinter growthoccurred during isotope stage 7 (Brook et al. 1990,Heine 1992) and not during the late Quaternary.

Although studies of the paleoclimate signalcontained within speleothem growth are few, someimplications can be reported: (1) Speleothemgrowth rates were highest in pre-Quaternarytimes in the Namib caves. (2) Only extremely smallgrowth rates occurred during the Quaternary in theNamib caves, and after ca. 125 ka BP nospeleothem growth was observed. (3) In the arid tosemi-arid interior of southern Africa, speleothemgrowth rates decreased after isotope stage 7 as well.(4) Climatic fluctuations are documented byspeleothem formation during different lateQuaternary periods. (5) During and before isotopestage 7, conditions for speleothem growth were morefavorable than during later phases with sinterdevelopment. This applies to the Namib caves aswell as to the interior caves. (6) Since isotope stage7, the range of moisture fluctuations seems to havedecreased. At the same time, the magnitude ofclimatic changes decreases from East to West. As aresult. the central Namib Desert did not experiencefluctuations in precipitation during the lateQuaternary, whereas the interior of southern Africawas effected by late glacial changes in precipitation.(7) The early Holocene of the Namib Desert seems tobe out of phase with regard to the humidity

300

Figure 8. Photomicrograph of Namib speleothem, Rossing Cave, flowstone intercalated with aeolian sand (seeHeine & Geyh 1984, Heine 1992). A UfTh age of ca. 185 ka BP was processed from the calcite layers. A setof several of such flowstone layers and sand layers documents alternating deposition of sinter and aeolian sand.This record indicates changes in available moisture during a period dated around 146 to 205 ka BP (see Fig. 5).Bar = 1 mm.

fluctuations in the interior of southwestern Africa.Thus, the late Quaternary climatic development of

the central Namib Desert reflects the situation of thesurface oceanography in the South Atlantic Ocean,which is characterized by a cyclonic gyre circulationincluding the northwest-directed Benguela Current,the eastward South Equatorial Counter Current, andthe Angola Current which flows southward along theAngola Margin. During the late Quaternary, thesurface oceanography off southwestern Africa(central Namib area) was more or less stable.Geochronological evidence suggests a prolonged

period of aridity during the late Quaternary that wasinfluenced by neither the large climatic changes northe fast and abrupt climate variability recorded inGreenland ice cores (Broecker 1994, Zahn 1994,Bond 1995) or in long terrestrial sediment sequences(e.g. Hooghiemstra 1989). The part of the globalocean circulation that influences the BenguelaCurrent off the central Namib coast may not havebeen altered during the late Quaternary. Sudden(recurring) changes in the operation of the BenguelaCurrent did not occur.

301

Figure 9. Photomicrograph of Namib speleothem, Tinkas Cave, "sinter-with-sand" stalagtite (see Heine 1992and Fig. 6). Cross-section with canals (c) and growth rings (a). A weathering horizon (w) developed on thesurface and along the canals. The weathering crust consists of CaC03, blown-in minerals (quartz, mica etc.)and colloidal Si02 with iron. The weathering horizon documents that there was no speleothem formationduring the late Quaternary. UfTh ages from speleothems are older than ca. 220 ka BP. Bar = I mrn.

5 CONCLUSIONS The UfTh dates, as well as TL dates from cave dunesands, show that chronostratigraphies based on 14Cdates have to be revised. Radiocarbon ages ofspeleothems seem not to reflect the time of thespeleothem growth, but the time of contamination or afinite age with no realistic reference to any morehumid period. The last significant pre-Holocenepluvial phase with speleothem growth in the centralNamib Desert probably occurred during oxygenisotope stage 7 (ca. 200-220 ka BP). It seems likelythat climate in the arid parts of southwestern Africaexperienced a more humid phase around and after

In the hyper-arid central Namib Desert, in the areanorth of the Namib Sand Sea, flash-flood sedimentsof the Kuiseb River, gypcretes of the NamibUnconformity Surface, and speleothems of caveshave been found to be an excellent repository ofpaleoclimatic data for terrestrial environments. Inparticular, the very presence of speleothem relicsindicate more moisture at the time of formation. Both,14C and UfTh age determinations were used to datethe speleothems and, hence, the climatic conditions.

302

200 ka BP. As the TL chronostratigraphies ofpaleodunes of the Etosha area in northern Namibiatogether with paleosol evidence show no indicationof a significantly more humid phase after 120 ka BP(Heine 1992), the results from the central NamibDesert prove continued aridity along the Namibiancoast between Walvis Bay and Swakopmund.

ACKNOWLEDGMENTS

This research was supported by grants of theDeutsche Forschungsgemeinschaft. Sincere thanks isalso extended to Almut, Jan, Henrike and CarolineHeine, Gisela Weigmann, Mebus A. Geyh, andmany other persons for their help in the field. MebusA. Geyh provided radiocarbon and UlTh dates.Gernot Endlicher made thin sections andphotomicrographs. I am grateful to the numerouspeople who have rendered information and who havediscussed material with me. K.W. Glennie isacknowledged for the valuable comments made to themanuscript.

REFERENCES

Besler, H. 1972. Kli maverhaltni sse undkJimamorphologische Zonierung der zentralenNamib (Siidwestafrika). Stuttgarter Geogr. Studien83:1-209.

Besler, H., Bliimel, W.D., Heine, K, Huser, K.,Leser, H. & Rust, U. 1994. Geomorphogeneseund Palaoklima Namibias. Die Erde 125:139-165.

Bond, G.e. 1995. Climate and the conveyor. Nature377 (6548):383-384.

Broecker, W.S. 1994. Massive iceberg discharges astriggers for global climate change. Nature 372:621-622.

Brook, G.A. 1995. Quaternary environments in theAfrican and Indian Ocean tropical zone: evidencefrom cave sediments in Somalia, Zaire, Botswana,Namibia and Madagascar (abstract). Presented atthe Southeast Asia Conference, The InternationalAssociation of Geomorphologists, Singapore /8-23 June 1995: 34. Singapore:National University.

Brook, G.A., Burney, D.A. & Cowart, LB. 1990.Desert paleoenvironmental data from cavespeleothems with examples from the Chihuahuan,Somili-Chalbi, and Kalahari deserts. Palaeogeogr.Palaeoclimatol. Palaeoecol. 76: 311-329.

Burney, D.A., Brook, G.A. & Cowart, lB. 1994.A Holocene pollen record for the Kalahari Desertof Botswana from a U-series dated speleothem.The Holocene 4(3):225-232.

Dixon, J.e. 1994a. Aridic soils, patterned ground,and desert pavements. In A.D. Abrahams & AJ.Parsons (eds), Geomorphology of DesertEnvironments: 64-81. London:Chapman and Hall.

Dixon, J.e. 1994b. Duricrusts. In A.D. Abrahams& A.1. Parsons (eds), Geomorphology of DesertEnvironments:82-105. London:Chapman and Hall.

Eitel, B. 1994. Kalkreiche Decksedimente undKalkkrustengenerationen in Namibia: Zur Frage

der Herkunft und Mobilisierung desCalciumcarbonats. Stuttgarter Geogr. Studien123:1-193.

Geyh, M.A. 1995. Geochronologische Aspektepalaohydrologischer und palaoklirnatischerBefunde in Namibia. Geomethodica (= Verof].20. BGC) 20:75-99.

Heine, K. 1987. Jungpleistozane f1uvialeGeomorphodynamik in der Namib, SiidwestafrikalNamibia. Z. Geomorph. Suppl. -Bd. 66: 113-134.

Heine, K 1991. Palaoklirna und Reliefentwicklungder Namibwiiste im iiberregionalen Vergleich.Geomethodica (=Verof!. 16. BGC) 16:53-92.

Heine, K 1992. On the ages of humid LateQuaternary phases in southern African arid areas(Namibia, Botswana). Palaeoecology of Africa23: 149-164.

Heine, K. 1993. Zum Alter jungquartarerFeuchtphasen im ariden und semiaridensiidwestlichen Afrika. Wiirzburger geogr. Arbeiten87:149-162.

Heine, K 1995. Palaoklirnatische Informationen aussiidwestafrikanischen Boden undOberflachenforrnen: Methodische Uberlegungen,Geomethodica (= Verof!. 20. BGC) 20:27-74.

Heine, K. & Geyh, M.A. 1984. Radiocarbon datingof speleothems from the Rossing Cave, NamibDesert, and palaeoclimatic implications. In J.e.Vogel (ed), Late Cainozoic Palaeoclimates of theSouthern Hemisphere: 465-470. Rotterdam:Balkema.

Heine, K. & Walter, R. 1996. DieGipskrustenboden der zentralen Namib (Namibia)und ihr palaokf imatischer Aussagewert.Petennanns Geographische Mitteilungen 140:237-253.

Holmgren, K., Karlen, W. & Shaw, P.A. 1995.Paleoclimatic Significance of the stable IsotopicComposition and Petrology of a Late PleistoceneStalagmite from Botswana. Quaternary Research43:320-328.

Holmgren, K., Lauritzen, S.-E. & Possnert, G.1994. 230Th/234U and 14C dating of latePleistocene stalagmite in Zobatse II Cave,Botswana. Quaternary Geochronology (QuaternaryScience Review) 13: 111-119.

Hooghiemstra, H. 1989. Quaternary and upper-Pliocene glaciations and forest development in thetropical Andes: evidence from a long high-resolution pollen record from the sedimentaryBasin of Bogot·, Colombia. Palaeogeogr.Palaeoclimatol. Palaeoecol. 72: 11-26.

Hoverrnann, J. 1978. Formen und Formung derPranamib (Flachen-Narnib). z. Geomorph. 30:55-73.

Lancaster, J., Lancaster, N. & Seely, M.K 1984.Climate of the central Namib Desert. Madoqua 14:5-61.

Lindesay, J.A. & Tyson, P.D. 1990. Climate andnear-surface airflow over the central Namib. InM.K. Seely (ed), Namib ecology: 25 years ofNamib research:27-37. Pretoria:Transvaal MuseumMonograph 7.

Marker, M.E. & Miiller, D. 1978. Relict vlei silt of

303

:

the middle Kuiseb valley, South West Africa.Madoqua 11(2):151-162.

Ollier, C.D. 1977. Outline geological andgeomorphic history of the central Namib Desert.Madoqua 10 (3):207-212.

Reheis, M.C. 1987. Cypsic soils on the Kanealluvial fans, Big Horn County, Wyoming.(USGS Bull. 1590 C). Washington DeUSGovernment Printing Office.

Rust, U. 1989. (Palaoj-Klirna und Relief: DasReliefgeftige der sildwestafrikanischen

amibwiiste (Kunene bis 27° s.B.) Miinchn e rCeograph. B7:1-158.

Rust, U. 1994. Skizzierung eigener Beitrage zurRelief- und PaHioklimaforschung in einemafrikanischen Trockenraum - Namibia. Mitt. geogr.Ces. Munchen 79:167-177.

Rust, U. & Wieneke, F. 1980. A re-investigation ofsome aspects of the evolution of the Kuiseb Rivervalley upstream of Gobabeb, South West Africa.Madoqua 12(3):163-173.

Scholz, H. 1972. The soils of the central NamibDesert with special reference to the soils of thevicinity of Gobabeb. Madoqua 11( 1):33-51.

Scholz, H. & Beckmann, H. 1971. DasVersalzungsproblem der Bewasserungsboden imUnterlauf des Swakops in Sildwestafrika. Ztschr.Kulturtechnik u. Flurbereinigung 15:367-375.

Sharon, D. 1981. The distribution in space of localrainfall in the Namib desert. J. Climatology 1:69-75.

Smith, R.M.H., Mason, T.R. & Ward, J.D. 1993.Flash-flood sediments and ichnofacies of the LatePleistocene Homeb Silts, Kuiseb River, Namibia.Sed. Ceol. 85:579-599.

Teller, J.T., Rutter, N. & Lancaster, N. 1990.Sedimentology and Paleohydrology of LateQuaternary Lake Deposits in the Northern NamibSand Sea, Namibia. Quaternary Science Reviews9: 343-364.

Vogel, J.e. 1982. The age of the Kuiseb River siltterrace at Homeb. Palaeoecology o.lAfrica 15:201-209.

Walter, R. 1994. Beziehungen zwischenKrustenbden (Gips, Kalk) und der Morphologieder Flachennamib. Unpublished M.Sc. Thesis,University of Regensburg (Physical Geography).

Ward, J.D. 1987. The Cenozoic succession in theKuiseb Valley, central Namib Desert. Mem. Geol.Surv. Namibia 9:1-123.

Ward, J.D. & Corbett, I. 1990. Towards an age forthe Namib. In M.K. Seely (ed), Namib Ecology:25 years of Namib research: 17 - 26.Pretoria:Transvaal Museum Monograph 7.

304