Temporal variability of fluxes of eolian-transported freshwater diatoms, phytoliths, and pollen grains off Cape Blanc as reflection of land-atmosphere-ocean interactions in northwest

Temporal variability of fluxes of eolian-transported freshwater

diatoms, phytoliths, and pollen grains off Cape Blanc as reflection of

land-atmosphere-ocean interactions in northwest Africa

Oscar E. Romero, Lydie Dupont, and Ulrike WyputtaFachbereich Geowissenschaften, Universitaet Bremen, Bremen, Germany

Susanne JahnsInstitut fur Palynologie und Quartarwissenschaften, Gottingen, Germany

Gerold WeferFachbereich Geowissenschaften, Universitaet Bremen, Bremen, Germany

Received 10 April 2000; revised 4 September 2001; accepted 27 February 2003; published 21 May 2003.

[1] Fluxes of airborne freshwater diatoms (FD), phytoliths (PH), and pollen grains (PO)collected with sediment traps off Cape Blanc, northwest Africa, from 1988 till 1991 arepresented. Both continental rainfall variations and wind mean strength and directionplay a key role in the temporal fluctuations of the fluxes of eolian traces in the pelagicrealm. Drier conditions in Northern Africa in 1987 could have preceded the highlithogenic input and moderate FD flux in 1988. The PH peak in summer 1988 wasprobably caused by increased wind velocity. Wetter rainy seasons of 1988/89 might havepromoted a significant pollen production in summer 1989, and FD in late 1989 andearly 1990, as well as contributed to the reduction of the lithogenic flux in 1989/90.Decreased fluxes of FD, PH and PO, and higher contribution of the 6–11 mm lithogenicfraction in 1991 would mainly reflect minor intensity and decreased amount ofcontinental trade winds. Air-mass backward trajectories confirm that the Saharan AirLayer is predominantly involved in the spring/summer transport. Trade winds play adecisive role in the fall/winter months, but also contribute to the transport during latespring/summer. Origin of wind trajectories does not support a direct relationship betweentransporting wind-layers and material source areas in Northern Africa. High winterfluxes of eolian tracers and high amount of trade winds with continental origin insummer warn against a simplistic interpretation of the seasonal eolian signal preserved inthe sediments off Cape Blanc, and the wind layer involved in its transport. INDEX

TERMS: 3309 Meteorology and Atmospheric Dynamics: Climatology (1620); 4215 Oceanography: General:

Climate and interannual variability (3309); 4516 Oceanography: Physical: Eastern boundary currents; 9305

Information Related to Geographic Region: Africa; KEYWORDS: marine particles, sediment traps, eolian

input, diatoms, pollen, phytoliths, land-atmosphere-ocean interactions, northwest Africa

Citation: Romero, O. E., L. Dupont, U. Wyputta, S. Jahns, and G. Wefer, Temporal variability of fluxes of eolian-transported

freshwater diatoms, phytoliths, and pollen grains off Cape Blanc as reflection of land-atmosphere-ocean interactions in northwest

Africa, J. Geophys. Res., 108(C5), 3153, doi:10.1029/2000JC000375, 2003.

1. Introduction

[2] The most prominent and persistent area of highaerosol distribution all over the world is found in thetropical northeast Atlantic. The dust plume, mainly origi-nating within the arid region of Northern Africa, is a rathercontinuous phenomenon [Schuetz, 1980, 1989], with largeday-to-day and seasonal changes in its location and density[Prospero, 1996; Stegmann and Tindale, 1999]. Dailyfluctuations in dust concentrations are mainly subject to

winds associated with the large-scale meteorological situa-tion, while the longer-term changes are related to climato-logical factors. Dust input can be interpreted as supporting adirect drought-related cause, but other mechanisms couldalso affect the same result [Schuetz, 1989; Prospero, 1990].[3] Several studies of the last 3 decades have greatly

contributed to clarify the teleconnections between NorthernAfrica and the tropical northeast Atlantic. Still, manyquestions remains unanswered (see Prospero [1996] for areview). Most of these studies can be classified into twomain groups: studies that emphasize the role of land-atmosphere interaction in the regional climate [e.g., Littman,1991; Hulme, 1996], and studies that emphasize the role of

JOURNAL OF GEOPHYSICAL RESEARCH, VOL. 108, NO. C5, 3153, doi:10.1029/2000JC000375, 2003

Copyright 2003 by the American Geophysical Union.0148-0227/03/2000JC000375$09.00

22 - 1

ocean-atmosphere interaction over West Africa [e.g., Rowellet al., 1995; Fontaine et al., 1998]. Interdisciplinary studiesfocusing on the connection between sources, transport anddeposition of windblown material in marine sediments arerather scarce. In this regard, trap-collected freshwater dia-toms, phytoliths and pollen can be used as potential sourcemarkers of land-derived material over the tropical northeastAtlantic [Dupont, 1999; Romero et al., 1999a].[4] The survey of the wind-carried material from its source

till its final deposition at the seafloor off northwest Africacontributes to the comprehension of the land-ocean-atmos-phere relationships [Schuetz, 1989]. A straightforward inter-pretation is commonly difficult to obtain because amounts ofterrigenous material are dependent on so many factors suchas aridity, dust mobilization, and wind direction and inten-sity. It has been widely demonstrated that changes in theallochthonous microfossil assemblages in tropical northeastAtlantic sediments provide information on the eolian inputfrom northwest Africa and the wind system involved. There-fore, it documents continental paleoenvironments and bio-stratigraphy during Quaternary times, and contributesignificantly to the reconstruction of paleoclimatic eventsin Northern Africa [e.g., Pokras and Mix, 1985; Gasse et al.,1989, 1990; Pokras, 1991; de Menocal et al., 1993; Dupont,1999]. However, if we wish to interpret the paleoatmosphericcirculation and its relationship with the paleoclimate, it isnecessary to assess the present-day deposition rates of land-

derived material, in order to establish likely source areas, aswell as to identify the transport agent and its seasonalvariability. The study of eolian tracers collected with sedi-ment traps might help to fulfill some gaps in the cycle ofgenesis in the continent till final deposition in the oceanbottom. The trap site Cape Blanc (�21�N, off Mauritania)allows a good identification of the material from source areasover Northern Africa because it is located within the mainarea of eolian transport over the tropical northeast Atlantic[Schuetz, 1980; Chiapello et al., 1999]. Based on approx-imately 4-year continuous observations (1988–1991), weanalyze the temporal dynamics of trap-collected freshwaterdiatom (FD), phytolith (PH), and pollen (PO) off CapeBlanc, and expand on observations initiated at the samemooring site by Lange et al. [1998], Ratmeyer [1996],Ratmeyer et al. [1999], and Romero et al. [1999a, 1999b].We mainly address the question as to whether seasonalvariation pattern of microorganism and plant remain inputfrom the continent is related to both changes in wind originand strength, and/or the source area in Northern Africa.

2. Atmospheric and Meteorological Conditionsin Northern Africa

[5] Eolian dust transport over northwest Africa and thetropical northeast Atlantic occurs either within the SaharanAir Layer (SAL), or by strong trade winds (Figure 1).

Figure 1. Location of the Cape Blanc (CB) mooring and seasonal pattern of the eolian transport overnorthwest Africa and the tropical northeast Atlantic. Thin arrows represent the low-altitude trade winds(TW, 850 hPa), while thick arrows represent the high-altitude winds within the Saharan Air Layer (SAL,500 and 700 hPa). Solid lines represent the northernmost (August) and southernmost (February) positionsof the Intertropical Convergence Zone (ITCZ). See color version of this figure in the HTML.

22 - 2 ROMERO ET AL.: TEMPORAL VARIABILITY OF EOLIAN SIGNAL TRAPPED OFF NW AFRICA

Blowing at altitudes between 5 and 6 km, the westwardSAL reaches its maximum strength and highest dust loadduring summer, focusing between 17�N and 21�N [Pros-pero, 1990; Chiapello et al., 1995]. In winter, trade windsdominates and are restricted to a shallower layer (approx-imately 1.5 km, 850 hPa) [Chiapello et al., 1995]. TheIntertropical Convergence Zone (ITCZ), a major climaticboundary, separates the northeast trade winds from thesoutheast monsoon (Figure 1). Because of seasonal migra-tions of the adjacent high pressure cells, the ITCZ movesbetween approximately 2�N and 12�N above the easterntropical Atlantic, reaching its northernmost position duringJuly/August.[6] The main climatic feature of northern Africa is an

annual cycling of dry and rainy monsoon seasons, resultingfrom the ITCZ oscillations [Fontaine and Bigot, 1993].Precipitation is distributed in three belts in Northern Africa,approximately south of 20�N and north of 30�N, whereannual rainfall is higher than 200 mm yr�1, while in theSahara the annual rainfall is generally less than 100 mmyr�1 [Nicholson, 1986; Shay-El et al., 1999]. In the northernSahara, the intensity of the summer subtropical high-pres-sure system precludes almost any precipitation during July–September. At this time, to the south of the Sahara, theITCZ has reached its most northerly location. This is thesingle short rainy season for the Sahelian region immedi-ately to the south of the Sahara desert [Ward et al., 1999].The latitudinal positions of the ITCZ appear to be differentduring abnormal dry and wet years over the Sahel [Fontaineet al., 1998]. The unusually wet years are linked to a strongand early northern ITCZ migration over the central equato-rial Atlantic; below average rainfall, it exhibits the reversepattern [Fontaine and Bigot, 1993]. Within the typicalseasonal variation of rainfall, the dust storm maximumoccurs mainly in April, but frequencies are also high inwinter [Littman, 1991]. The dust storm minimum is notreached until October, i.e., two months after the regularrainfall maximum [Lamb and Peppler, 1992].[7] January is the peak season for African savanna burn-

ing north of the equator, when the areas south of 20�Nreceive less than 25 mm of yearly precipitation [Cahoon etal., 1992]. In February and March, precipitation increasesgradually from south to north across the northern Africansavannas, and there is a noticeable drop in fire frequency.By April, savanna burning is at a minimum and the burningzone moves southward. In November–December it isexpected to return to the interior of the savannas north ofthe equator, and spread to the far west later in December.Hence the annual burning cycle of the African savannacontinues [Cahoon et al., 1992].

3. Sources of Freshwater Diatoms, Phytoliths,and Pollen in Northern Africa

[8] Main source areas for windborne FD are a multitudeof depressions through the Sahara and the Sahel. Numerouslake sediment deposits, mostly of Holocene age, are foundin the area around Taoudenni, south of the Hoggar moun-tains, south of the Saharan Atlas, and around Lake Chad(Figure 1) [Gasse et al., 1989]. Highly diversified diatomassemblages are also widespread at large river basins andlacustrine sediments [Gasse et al., 1989]. Grasses are the

main source of PH [Twiss et al., 1969], but they also occurin other plant families [Barboni et al., 1999]. Savanna fireslargely contribute to release PH into the air.[9] Depending on the vegetation, modern PO composi-

tion in dust samples change from north to south [Cour andDuzer, 1980]. Pollen spectra from the Mediterranean forestinclude pine and oak pollen; those of the Mediterranean-Saharan Transitional Steppes and the elevated areas of theAlgerian Plateau include Artemisia, and members of Aster-oideae (tubuliflorous composites) and Chenopodiaceae/Amaranthaceae. Farther south in the Sahara, where POproduction is low, large relative values of Poaceae (grass)and Chenopodiaceae/Amaranthaceae occur. Among thelatter, a number of species are halophytic and grow onsaline desert soils and around saline lakes. AbundantPoaceae together with relatively large amounts of Cyper-aceae (e.g., sedges) are produced in the southern Saharanfringe, the semi-desert grasslands of the Sahel (the SahelAcacia wooded grasslands) and the Sudan savanna zone[Cour and Duzer, 1980; Hooghiemstra et al., 1986; Lezineand Hooghiemstra, 1990]. The main flowering period inthe Mediterranean area and the Mediterranean-SaharanTransition Steppes (in and along the southern fringe ofthe Anti-Atlas and the Saharan Atlas mountains) is in latewinter and spring [Knapp, 1973]. The main floweringperiod in the southern Sahara and the Sahel occurs insummer.

4. Methods

[10] A total of four deep-sea moorings were deployed offCape Blanc (approximately 21�N, 20�W, CB; Figure 1).Details on sampling intervals and trap depths are given inTable 1. The temporal resolution of trap samples is approx-imately 4 weeks between March 1988 and March 1989(CB1), approximately 3 weeks between March 1989 andApril 1991 (CB2 and 3), and 10 days from May toNovember 1991 (CB4). For CB1 and CB2, the classicalcone-shaped traps Mark Vand Mark VI (0.5 m2 and 1.17 m2

collection area, respectively) were used. Kiel SMT 230/23(0.5 m2 collection area) were used for CB3 and CB4.Collection cups were poisoned with HgCl2 before deploy-ment, and NaCl was added to reach a final salinity of 40%.Samples were poisoned again after recovery with HgCl2 andstored at 4�C. The splitting procedure and chemical analy-ses of the <1 mm fraction were carried out at BremenUniversity (described by Fischer and Wefer [1991]).[11] For the study of FD and PH 1/16 and 1/64 splits of

the original sample were used. Samples were prepared for

Table 1. Cape Blanc Traps: Location, Deployment Depths, and

Sampling Duration and Intervals

MooringTrap Type(Opening) Position

WaterDepth, m

TrapDepth, m

SamplingDuration

CB1 Mark V 20�45.30N 3646 2195 3/22/88(0/5 m2) 19�44.50W 3/08/89

CB2 Mark VI 21�08.70N 4092 3502 3/15/89(1/17 m2) 20�41.20W 3/24/90

CB3 Kiel SMT 230 21�08.30N 4094 3557 4/29/90(0/5 m2) 20�40.30W 4/8/91

CB4 Kiel SMT 230 21�08.70N 4108 3562 3/5/91(0/5 m2) 20�41.20W 11/19/91

ROMERO ET AL.: TEMPORAL VARIABILITY OF EOLIAN SIGNAL TRAPPED OFF NW AFRICA 22 - 3

microscopical observation according to the methodologyproposed by Simonsen [1974]. Qualitative and quantitativeFD and PH analyses were carried out on permanent slides ofacid-cleaned material (Mountex mounting medium). Fordiatoms, each individual was identified to the lowest pos-sible taxonomic level. Since few grasses contain PH, whichare unique in shape [Blackman, 1971], they were counted asa group, and a distinction of different morphologies wasattempted based on the morphological classes outlined byBukry [1980]. Counting of replicate slides indicated that theanalytical error for the flux estimates is �15%. For POstudy, 1/64 splits were used. The subsample was treatedwith a mixture of 10% concentrated H2SO4 and 90% aceticacid for 5 min at approximately 90�C. Pollen grains weremounted with glycerin slides. To calculate PO concentra-tions, tablets containing a known number of Lycopodiumspores were added. Pollen was also determined to the lowestpossible taxonomic level.[12] Backward trajectories of air-mass pathways ending

above the CB trap site have been computed for threepressure levels for each day between late February 1988and late November 1991: 500 hPa (Saharan Air Layer,between 6 and 7 km altitude), 700 hPa (approximately 3km), and 850 hPa (approximately 1.5 km). While the 700-and 500-hPa levels represent the area and the upper boun-dary of the SAL [Kubatzki and Claussen, 1998; Diedhiou etal., 1999], the latter marks the upper boundary of the tradewinds [Schuetz, 1980]. These trajectories have been calcu-lated with a global 2-dimensional trajectory isobaric model,a modification of a previously developed model by theNorwegian Meteorological Institute [Eliassen, 1977]. Themodel computes 4-day isobaric backward trajectories fromzonal and meridional components. Wind data were obtainedfrom the European Centre for Medium Range WeatherForecast. The wind velocity (as defined by the trajectoryextension) and days of winds with continental provenancehave been estimated based on the wind trajectories.

5. Results

[13] Land-derived FD, PH and PO were present in allsediment trap samples collected off Cape Blanc from early1988 through late 1991 (Figure 2). The numerical contri-bution of FD was the highest (range �0.4–8 � 104 valvesm�2 d�1), followed by PH and PO (�0.4–17 � 103 bodiesand �0.05–5 � 102 grains m�2 d�1, respectively). A totalof 37 FD species have been identified throughout thesampling period. Aulacoseiragranulata and A. islandicaalways dominated the FD assemblage (>60%), accompa-nied by Stephanodiscus astraea and Hantzschia amphioxys.The PH assemblage was mainly made up by panicoid andchoridoid forms and, as for FD, no particular seasonaldistribution pattern occurred. Regular festucoid PH weresecondary components throughout the four-year samplingperiod. The PO assemblage was variable. Mostly Cyper-aceae and Chenopodiceae/Amaranthaceae contributed to theflux throughout the sampling period, followed by Asteroi-deae and Poaceae. Furthermore, pine and other PO withMediterranean origin were found. A large number ofSahelian taxa was present in low quantities. Mainly Che-nopodiceae/Amaranthaceae and Asteroideae contributed tothe high PO flux values during spring 1988, while the flux

maxima in summer and fall 1989 showed a more diversifiedspectrum.[14] Fluxes had maxima of variable magnitude almost in

each season and showed considerable interannual variation(Figures 2 and 3). Seasonality, as expressed by the occur-rence and magnitude of maxima, was stronger during 1988/1989, and decreased significantly after spring 1990. Corre-lations between fluxes of FD, PH, and lithogenics were ingeneral low (r = 0.36–0.50), while PO paralleled reason-ably well PH and lithogenics (r = 0.59 and 0.69, respec-tively), but it was in turn less correlated with the FD flux(r = 0.47).[15] The highest PO and lithogenic maxima for all the

sampling period were recorded in spring and summer 1988(Figure 2 data lines a and b). Mostly Chenopodiaceae/Amaranthaceae and Asteroideae contributed to the POspring 1988 peak (Figure 3 data line a). The highest amountof winds with continental origin corresponded to the 850-hPa layer in spring 1988 (Figure 2 data line a), mainlycoming from Tamanrasset area and northwest Sahara, asshown by the wind trajectories in Figure 4 data line a. PHreached their highest daily flux in early summer 1988(Figure 2 data line b). A highly diversified PO peak alsooccurred in summer 1988 peak with similar percent con-tributions by Asteroideae, Chenopodiaceae/Amaranthaceae,Cyperaceae and grasses (Figure 3 data line b). As seen inFigure 2 data line b and Figure 4b, the 700- and 500-hPalayers were largely involved in the transport of land-derivedmaterial from the Sahel, while the 850-hPa winds blew fromthe northwesternmost part of Africa during some days inearly summer 1988.[16] Freshwater diatoms were mainly carried by the lower

boundary of the SAL and the trade winds during the early fall1988 maximum (Figure 2 data line c). Wind trajectoriesmainly originated between approximately 17�N–25�N incentral Northern Africa, and in lesser degree in the Sahel(500 hPa, Figure 4c). Freshwater diatoms peaked again inearly winter 1989 (Figure 2 data line d). As seen in Figure 4d,eolian tracers mainly came from an extensive area between18�N–30�N and 5�W in northwest Africa, predominantlytransported by trade winds. The 700-hPa level contributed tothis peak during some days, primarily originated in theSahel-Sudan transition zone.[17] Except for the FD maximum in early winter 1989

and the PO peak in late spring 1989, moderate fluxes werecollected in winter–spring 1989. The PO flux maximum inlate spring 1989 consisted mainly of Chenopodiaceae/Amaranthaceae and Cyperaceae, accompanied by grasses(Figure 3 data line e). Figure 4e shows that during somedays land-derived winds carried material from the Medi-terranean Saharan Domain at 850 hPa or the northernmostSahel at the 500-hPa layer. Lithogenic and PO maximafollowed in midsummer 1989. Grasses and Cyperaceaewere the main PO contributors (Figure 3 data line f). The500- and 700-hPa level pressure were mainly involved intheir transport (see this event in Figure 4f). In late summer1989, PH were transported by the two upper wind layersfrom an area between 20�N–30�N and 0�–5�W.[18] In early fall 1989 a long period of strong FD input

begun which extended approximately 130 days till mid-winter 1990. As seen in Figure 2 data lines g–i, the 500-and 700-hPa layers dominated the transport during the first

22 - 4 ROMERO ET AL.: TEMPORAL VARIABILITY OF EOLIAN SIGNAL TRAPPED OFF NW AFRICA

Figure 2. (top) Time series flux of freshwater diatoms, phytoliths, pollen grains, and lithogenics at theCB trap site from March 1988 to November 1991. (bottom) Wind velocity (dotted line) and days of windsof continental provenance (solid line) for three air pressure levels (500, 700 and 850 hPa). Letters a-l onthe upper part refer to the daily backward air-mass trajectories (see Figures 4a–4l) calculated forsignificant flux maxima. Flux of lithogenics was taken from Ratmeyer [1996] and Ratmeyer et al. [1999].

ROMERO ET AL.: TEMPORAL VARIABILITY OF EOLIAN SIGNAL TRAPPED OFF NW AFRICA 22 - 5

part of this period. Trajectories demonstrate that windsoriginated in a broad area between 10�Wand approximately20�E in Northern Africa, covering a wide range of environ-ments from the Northern Sahel to the Sahara (Figure 4g).The 700- and 850-hPa trajectories were related to the longhigh FD flux period in fall 1989, and came from a wide areabetween 8�N and 25�N in north central Africa (Figure 4h).Trade winds, that originated in the northwesternmost part ofAfrica between 10�E and 15�W, were mainly responsiblefor the transport of land-derived material, with minorcontributions by the 700-hPa layer winds, as depicted inFigure 2 data line i and 4i. All the eolian tracer fluxesstrongly decreased in winter 1990.[19] The highest daily FD flux and second highest PH

maximum were reached in midspring 1990, followed by aperiod of significant FD export until midsummer 1990. The850- and 700-hPa layers were mainly responsible for thetransport of land-derived material in mid spring 1990 peak(see this event in Figure 2 data lines j–k and Figure 4j).Trade winds originated between approximately 13�N–30�Nand 10�E–10�W, while 700-hPa winds mainly came from anarrower area between the Sahel and approximately 5�N.[20] As show in Figure 2 data line k and its associated

trajectories (Figure 4k), the high FD contribution in mid-summer 1990 was linked to the two upper wind layers

which carried material from the Holocene lake deposits at15�N–20�N and the Sahel. A considerable flux decrease ofall the measured parameters followed the high export phasein midsummer 1990. Some minor peaks were observed inspring–summer 1991 (data line l in Figures 2 and 3). Theflux of FD reached similar values to those of summer 1989,and the PH peaks in summer 1991 were almost as high asthose previously recorded in summer 1990. A similarmaximum pattern was observed for PO and lithogenics.The transport of material for the summer 1991 peak mainlyoccurred within high altitude winds (500 and 700 hPa),originated between 5�N–30�N, and 10�W in the Africancontinent (Figure 4l).

6. Discussion

[21] Our analysis on the seasonal flux of three differenteolian tracers off Cape Blanc in the northeast tropicalAtlantic shows that their transport from the African con-tinent into the oceanic realm does not always happensynchronously. As demonstrated by our wind trajectories,a geographically wide area in the northern African Con-tinent is associated with the origin of each eolian tracer fluxmaximum. Large changes in concentration at specific originlocations in short periods of time (days to weeks, our

Figure 3. Time series flux of total pollen grains, and cumulative percentage of the most abundantgroups of pollen at the CB trap site from March 1988 to November 1991. Letters a, b, d and e–h at top ofplot refer to the daily backward trajectories (below Figures 4a–4l) calculated for significant flux maxima.Cumulative percentage is only shown for samples with a significant statistical amount of pollen grains.

22 - 6 ROMERO ET AL.: TEMPORAL VARIABILITY OF EOLIAN SIGNAL TRAPPED OFF NW AFRICA

Figure 4. Daily backward trajectories in three air pressure levels (500, 700, and 850 hPa) correspondingto significant flux maxima at the CB trap site from March 1998 to November 1991 (See Figures 2 and 3).For each flux maxima between 14 and 21 daily trajectories are represented.

ROMERO ET AL.: TEMPORAL VARIABILITY OF EOLIAN SIGNAL TRAPPED OFF NW AFRICA 22 - 7

Figure 4. (continued)

22 - 8 ROMERO ET AL.: TEMPORAL VARIABILITY OF EOLIAN SIGNAL TRAPPED OFF NW AFRICA

Figure 4. (continued)

ROMERO ET AL.: TEMPORAL VARIABILITY OF EOLIAN SIGNAL TRAPPED OFF NW AFRICA 22 - 9

sampling resolution) would significantly affect the sedimenttrap-collected signal. Temporal fluctuations of land-derivedmaterial collected at the CB trap site appear related tochanges of both source-area aridity, and wind path andstrength. Rainfall levels in the 1980s and early 1990s laybelow the long-term average in Northern Africa, especiallyin the Sahel [Tucker et al., 1991; Nicholson, 1994]. Inparticular, 1987 and 1990 experienced a moderate to severedrought, while 1988/89 were wetter than normal years[Ward, 1992; Rowell et al., 1995; Nicholson et al., 1998].On the basis of the correlation between the dustiness of aseason and the rainfall depth of the previous rainy season inNorthern Africa [Lamb and Peppler, 1991; N’Tchayi et al.,1994], we speculate that drier conditions in 1987 may haveincreased the soil erosion, and preceded the high lithogenicinput and the moderate FD flux in 1988. Barbosa et al.[1999] postulated that the strong drought in 1986/1987 mayhave lead to a lower biomass production with a consequentreduction in the number of fires and burned areas in 1987/1988 in northern Africa. The phytolith maximum in summer1988 was probably related with the increased velocity ofboth SAL and trade winds and a higher capacity transport,as can be seen in Figure 2 and its associated trajectories inFigure 4b.[22] The wetter rainy season of 1988/1989 might have

promoted the development of significant amounts of pollenin summer 1989, and, as a possible consequence of thedecrease in soil erosion, caused a reduction in the flux oflithogenics in 1989/1990. Nicholson et al. [1998] observedthat the northern border of the vegetation in the Sahelmoved northward in 1988/1989, as precipitation increased,thus demonstrating that shifts in the desert boundary andrainfall pattern are closely coupled. More-humid conditionsmay have also favored a higher production of FD in thesmall water bodies of the Tamanrasset and the Sahel, latelytransported mainly by trade winds in late 1989 and early1990. While variations in the lithogenic concentrations maybe interpreted as supporting a direct-drought related causefor increased dust, there are other mechanisms that couldaffect the same results. For example, small shifts in thelarge-scale wind system or in the dust sources in northernAfrica could result in very large changes in dust transportand deposition into the ocean [Schuetz, 1989; Prospero,1990, 1996]. The seasonal pattern of the wind regimeaffecting our CB trap site remains almost constant fromyear to year, but the amount of winds with continentalprovenance shows significant differences (Figure 2). Thewind velocity for the three pressure levels analyzeddecreased from 1988 through 1991, and it is conceivablethat the carrying capacity of both trade winds and SALdiminished. Hence, reduced fluxes of FD, PH and PO, andthe higher contribution of the 6–11 mm fraction of litho-genic particles in 1991 [Ratmeyer, 1996] reflect loweredintensity and lesser amount of continental trade winds.[23] It has commonly been assumed that the largest

portion of the wind-transported material deposited off CapeBlanc comes from southern Sahara and the Sahel, being theSAL the main carrier agent [e.g., Pokras and Mix, 1985;Chiapello et al., 1999]. Lower contributions have mostlybeen attributed to trade winds, mainly originating in theAtlas Mountains and the western Saharan coastal deserts[e.g., Pokras and Mix, 1985; Gasse et al., 1989, 1990;

Pokras, 1991]. However, the occurrence of importantamounts of material in the low-level trade winds hasrecently been demonstrated [Bergametti et al., 1989; Chia-pello et al., 1995; Stegmann and Tindale, 1999]. Off CapeBlanc, SAL wind trajectories are predominantly involved inlate spring-summer transport with substantial differences invelocity and amount of days with winds of continentalprovenance are seen (Figure 2). As expected, trade windsare more intense in winter and early spring, but they alsocontribute significantly to the eolian tracer transport insummer, at least during some years. All these observationsconfirm suggestions by Lange et al. [1998] and Romero etal. [1999a] that the CB trap site experiences the influence ofboth summer and winter eolian plumes. Air-mass trajecto-ries ending at the CB site show that source regions ofeolian-transported material deposited off Cape Blanc arewidespread in Northern Africa. Some Saharan and Sahelianregions act as source for dust events observed both in winterand summer. A direct relationship between a particular windlayer and a defined source area in Northern Africa isdifficult to recognize. This conclusion coincides with modelrepresentations developed by Marticorena et al. [1997],preventing to establish a direct link between dust transportand source regions.[24] Highest FD fluxes can reflect increased diatom

productivity in inland waters, but it also is believed thatmainly Holocene paleolakes widespread in northern Africanoperate as sources for windblown diatoms [Gasse et al.,1989, 1990]. Dissolution-resistant species of the diatomAulacoseira, which dominate the FD sediment trap assem-blage off Cape Blanc throughout the four-year sampling, arethe most common components in lacustrine diatom-bearingsediments in deflation areas of Sahara and Sahel, smallwaterbodies in central and northern Sahara, and the LakeChad basin [Gasse et al., 1989, 1990]. Seasonality in theflux of PH forms was very weak. Higher abundance ofpanicoid forms, composed of mixed C4-C3 tall grasses,characterize the grasslands of the Sahel belt and Sudan[Alexandre et al., 1997]. Chloridoid forms, mostly from C4

short grasses, but also produced in small quantities by C3Bambusoideae and Arundinoideae [Alexandre et al., 1997],occur warm and dry conditions of the Sahara. The richnessof the PO assemblage at the CB trap site reflects both thegeographical diversity related to the origin of the windtrajectories, and the wide spectrum of vegetation areas inNorthern Africa. In spite of the pollen diversity, a straight-forward relationship between the dominant pollen groupsand a particular wind layer is difficult to find.[25] Extremely intense eolian events observed at Canary

Islands in spring and summer 1988 [Chiapello et al., 1999]were also recorded southward off Cape Blanc. Thus dra-matic fluctuations in the flux of land-derived material over ashort timescale of a few days to a few weeks, when a dustoutbreak moves from northern Africa into the oceanicrealm, are quickly recorded by deepwater sediment trapsat the CB site. This observation supports the assumptionthat the transfer of particles from the atmosphere down tothe surface sediments underlying the CB trap is rapidenough in order to avoid almost any relevant inferencefrom hemipelagic influxes through lateral advection [Rat-meyer et al., 1999]. The lowered lithogenic flux during thefirst months of 1989 also mirrors a regional phenomenon

22 - 10 ROMERO ET AL.: TEMPORAL VARIABILITY OF EOLIAN SIGNAL TRAPPED OFF NW AFRICA

offshore northwest Africa. In coincidence with the fluxdecrease of our eolian tracers, and the enhanced percentcontribution of the 6–11 mm lithogenic particles, mineraldust concentrations measured at the Canary Islands alsodecreased significantly [Chiapello et al., 1999].[26] Information on the role of different wind systems

operating seasonally in the Atlantic can be of help wheninterpreting the land-derived microorganisms and plantremains preserved in the sedimentary record. Though com-parison between the trap-collected and sediment-preservedcompartments is somehow limited by differences in time-scale, a good correspondence exists between FD and PHassemblages in both compartments at the CB site [Lange etal., 1998; Romero et al., 1999a, 1999b]. Significant fluxesof FD, PH and PO in winter and high amount of trade windswith continental origin in summer warn against a simplisticinterpretation of the seasonal airborne signal preserved inthe sediments off Cape Blanc, and the wind layer involvedin the transport. We demonstrate that enhanced continentalinputs are due to the interplay of several factors. From thecomparison between fluxes and rain interannual variability,it is evident that dryness cannot be the only factor control-ling the eolian input off Cape Blanc. Continental aridity innorthern Africa and changes in wind strength and directionplay a decisive role in the variations of the seasonal patternof microorganism and plant remain input into the tropicalnortheast Atlantic.

[27] Acknowledgments. Thanks to A. Dickson and two anonymousreviewers for helpful reviews of this manuscript. Comments and sugges-tions by I. Chiapello were very useful in preparing the first draft. Thecompetent assistance of the officers and crew of the research vessel R/VMeteor in the recovery of the moorings is gratefully acknowledged.Financial support for this work was provided by the Deutsche Forschungs-gemeinschaft (Sonderforschungsbereich 261 at the Bremen University,contribution 337), and by the Bundesministerium fur Bildung, Wissen-schaft, Forschung und Technologie to the Palynology Department atGottingen University (project 0JKF018/2).

ReferencesAlexandre, A., J.-D. Meunier, A.-M. Lezine, A. Vincens, and D. Schwartz,Phytoliths: Indicators of grassland dynamics during the late Holocene inintertropical Africa, Palaeogeogr. Palaeoclimatol. Palaeoecol., 136,213–229, 1997.

Barboni, D., R. Bonnefille, A. Alexandre, and J.-D. Meunier, Phytoliths aspaleoenvironmental indicators, West Side Middle Awash Valley, Ethio-pia, Palaeogeogr. Palaeoclimatol. Palaeoecol., 152, 87–100, 1999.

Barbosa, P. M., D. Stroppiana, and J.-M. Gregoire, An assessment of ve-getation fire in Africa (1981–1991): Burned areas, burned biomass, andatmospheric emissions, Global Biogeochem. Cycles, 13, 933–950, 1999.

Bergametti, G., L. Gomes, G. Coude-Gaussen, P. Rognon, and M. N. LeCoustumer, African dust observed over Canary Islands: Source-regionidentification and transport pattern for some summer situations, J. Geo-phys. Res., 94, 14,855–14,864, 1989.

Blackman, E., Opaline silica bodies in the range grasses of southernAlberta, Can. J. Bot., 49, 769–781, 1971.

Bukry, D., Opal phytoliths from the tropical western Pacific Ocean, DeepSea Drilling Project Leg 54, in Initial Reports, Deep Sea Drilling Project,vol. 54, edited by D. R. Rosendahl et al., pp. 575–589, U. S. Gov. Print.Off., Washington, D. C., 1980.

Cahoon, D. R., Jr., B. J. Stocks, J. S. Levine, W. R. Coffer III, and K. P.O’Neill, Seasonal distribution of African savanna fires, Nature, 359,812–815, 1992.

Chiapello, I., G. Bergametti, L. Gomes, B. Chatenet, F. Dulac, J. Pimenta,and E. Santos Soares, An additional low layer transport of Sahelian andSaharan dust over the northeastern tropical Atlantic, Geophys. Res. Lett.,22, 3191–3194, 1995.

Chiapello, I., J. M. Prospero, J. R. Herman, and N. C. Hsu, Detection ofmineral dust over the North Atlantic Ocean and Africa with the Nimbus 7TOMS, J. Geophys. Res., 104, 9277–9291, 1999.

Cour, P., and D. Duzer, Action actuelle des courants atmospheriques sur ladissemination des pollens au Sahara et dans les regions climatiques avoi-sinantes le long d’un transect Oran-Abidjan, Mem. Museum Nat. Hist.Nat., Ser. B, 27, 66–72, 1980.

de Menocal, P. B., W. F. Ruddiman, and E. M. Pokras, Influences of high-and low-latitude processes on African terrestrial climate: Pleistocene eo-lian records from equatorial Atlantic Ocean Drilling Program Site 663,Paleoceanography, 8, 209–242, 1993.

Diedhiou, A., S. Janicot, A. Viltard, P. de Felice, and H. Laurent, Easterlywaves regimes and associated convection over west Africa and tropicalAtlantic: Result from the NCEP/NCAR and ECMWF reanalyses, Clim.Dyn., 15, 795–822, 1999.

Dupont, L., Pollen and spores in marine sediments from the east Atlantic-Aview from the ocean into the African Continent, in Use of Proxies inPaleoceanography: Examples From the South Atlantic, edited byG. Fischer and G. Wefer, pp. 523–546, Springer-Verlag, New York, 1999.

Eliassen, A., The trajectory model: A technical description, technical report,70 pp., Norw. Inst. of Air Res., Kjeller, Norway, 1977.

Fischer, G., and G. Wefer, Sampling, preparation and analysis of marineparticulate matter, in Marine Particles: Analysis and Characterization,Geophys. Monogr. Ser., vol. 62, edited by D. C. Hurd and W. Spencer, pp.391–397, AGU, Washington, D. C., 1991.

Fontaine, B., and S. Bigot, West African rainfall deficits and sea surfacetemperatures, Int. J. Climatol., 13, 271–285, 1993.

Fontaine, B., S. Trzaska, and S. Janicot, Evolution of the relationshipbetween near global and Atlantic SST modes and the rainy season inWest Africa: Statistical analysis and sensitivity experiments, Clim. Dyn.,14, 353–368, 1998.

Gasse, F., B. Stabell, E. Fourtanier, and J. van Iperen, Freshwater diatominflux in intertropical Atlantic: Relationships with continental recordsfrom Africa, Quat. Res., 32, 229–243, 1989.

Gasse, F., R. Tehet, A. Durand, E. Gibert, and J.-C. Fontes, The arid-humidtransition in the Sahara and the Sahel during the last deglaciation, Nature,346, 141–146, 1990.

Hooghiemstra, H., C. O. C. Agwu, and H.-J. Beug, Pollen and spore dis-tribution in recent marine sediments: A record of NW-African seasonalwind patterns and vegetation belts, Meteor Forschungsergeb. Reihe C,40, 87–135, 1986.

Hulme, M., Recent climatic change in the world’s drylands, Geophys. Res.Lett., 23, 61–64, 1996.

Knapp, R., Die Vegetation von Afrika, 626 pp., Fischer Verlag, Stuttgart,Germany, 1973.

Kubatzki, C., and M. Claussen, Simulation of the global bio-geophysicalinteractions during the last Glacial Maximum, Clim. Dyn., 14, 461–471,1998.

Lamb, P. J., and R. A. Peppler, West Africa, in Teleconnections LinkingWorldwide Climate Anomalies, edited by M. Glantz, R. W. Katz, andN. Nicholls, pp. 121–189, Cambridge Univ. Press, New York, 1991.

Lamb, P. J., and R. A. Peppler, Further case of tropical Atlantic surfaceatmospheric and oceanic patterns associated with Sub-Saharan drought,J. Clim., 5, 476–488, 1992.

Lange, C. B., O. E. Romero, G. Wefer, and A. J. Gabric, Offshore influenceof coastal upwelling off Mauretania, NWAfrica, as recorded by diatomsin sediment traps at 2195 m water depth, Deep Sea Res., Part I, 45, 985–1013, 1998.

Lezine, A. M., and H. Hooghiemstra, Land-sea comparisons during the lastglacial-interglacial transition: Pollen from west tropical Africa, Palaeo-geogr. Palaeoclimatol. Palaeoecol., 79, 313–331, 1990.

Littman, T., Rainfall, temperature and dust storm anomalies in the AfricanSahel, Geogr. J., 157, 136–160, 1991.

Marticorena, B., G. Bergametti, B. Aumont, Y. Callot, C. N’Doume, andM. Legrand, Modelling the atmospheric dust cycle: 2. Simulation ofSaharan dust sources, J. Geophys. Res., 192, 4387–4404, 1997.

Nicholson, S. E., Sub-Saharan rainfall 1981–1984, J. Clim. Appl. Meteor-ol., 25, 1365–1381, 1986.

Nicholson, S. E., Century-scale series of standardized annual departures ofAfrican rainfall, in Trends (93): Compendium of Data on Global Change,edited by T. A. Boden et al., pp. 952–962, Carbon Dioxide Inf. Anal.Cent., Oak Ridge Natl. Lab., Oak Ridge, Tenn., 1994.

Nicholson, S. E., C. J. Tucker, and M. B. Ba, Desertification, drought, andsurface vegetation: An example from the West African Sahel, Bull. Am.Meteorol. Soc., 79, 815–829, 1998.

N’Tchayi, G. M., J. Bertrand, M. Legrand, and J. Baudet, Temporal andspatial variations of the atmospheric dust loading throughout West Africaover the last thirty years, Ann. Geophys., 12, 265–273, 1994.

Pokras, E. M., Source areas and transport mechanisms for freshwater andbrackish-water diatoms deposited in pelagic sediments of the equatorialAtlantic, Quat. Res., 35, 144–156, 1991.

Pokras, E. M., and A. Mix, Eolian evidence for spatial variability of LateQuaternary climates in tropical Africa, Quat. Res., 24, 137–149, 1985.

ROMERO ET AL.: TEMPORAL VARIABILITY OF EOLIAN SIGNAL TRAPPED OFF NW AFRICA 22 - 11

Prospero, J. M., Mineral-aerosol transport to the North Atlantic and NorthPacific: The impact of African and Asian sources, in The Long-RangeAtmospheric Transport of Natural and Contaminant Substances, NATOASI Ser., vol. 297, edited by A. H. Knapp, pp. 59–86, Kluwer Acad.,Norwell, Mass., 1990.

Prospero, J. M., The atmospheric transport of particles to the ocean, inParticle Flux in the Ocean, edited by V. Itekkot, P. Schaeffer, S. Honjo,and P. Depetris, pp. 19–52, John Wiley, New York, 1996.

Ratmeyer, V., Untersuchung zum Eintrag und Transport lithogener undorganischer patikularer Substanz im ostlichen substropischen Nordatlan-tik, Berichte Fachbereich Geowissenschaften, 85, 154, Univ. Bremen,Bremen, Germany, 1996.

Ratmeyer, V., G. Fischer, and G. Wefer, Lithogenic particle fluxes and grainsize distributions in the deep ocean off northwest Africa: Implications forseasonal changes of eolian dust input and downward transport, Deep SeaRes., Part I, 46, 1289–1337, 1999.

Romero, O. E., C. B. Lange, R. J. Swap, and G. Wefer, Eolian-transportedfreshwater diatoms and phytoliths across the equatorial Atlantic recordtemporal changes in Saharan dust transport patterns, J. Geophys. Res.,104, 3211–3222, 1999a.

Romero, O. E., C. B. Lange, G. Fischer, U. F. Treppke, and G. Wefer,Variability in export production documented by downward fluxes andspecies composition of marine planktonic diatoms: Observations fromthe tropical and equatorial Atlantic, in Use of Proxies in Paleoceanogra-phy: Examples from the South Atlantic, edited by G. Fischer andG. Wefer, pp. 365–392, Springer-Verlag, New York, 1999b.

Rowell, D. P., C. K. Folland, K. Maskell, and M. N. Ward, Variability ofsummer rainfall over tropical North Africa (1906–1992): Observationsand modelling, Q. J. R. Meteorol. Soc., 121, 669–704, 1995.

Schuetz, L., Long range transport of desert dust with special emphasis onthe Sahara, in Aerosols: Anthropogenic and natural, sources and trans-port, edited by T. J. Kneip and P. J. Lioy, Ann. N. Y. Acad. Sci., 338,515–532, 1980.

Schuetz, L., Atmospheric mineral dust-Properties and source markers, inPaleoclimatology and Paleometeorology: Modern and Past Patterns ofGlobal Atmospheric Transport, edited by M. Leinen and M. J. Sarnthein,pp. 339–358, Kluwer Acad., Norwell, Mass., 1989.

Shay-El, Y., P. Alpert, and A. da Silva, Reassessment of the moisture sourceover the Sahara dessert based on NASA reanalysis, J. Geophys. Res.,104, 2015–2030, 1999.

Simonsen, R., The diatom plankton of the Indian Ocean Expedition of RV,Meteor 1964–1965 Meteor. Forschungsergeb., D19, 1–66, 1974.

Stegmann, P. M., and N. W. Tindale, Global distribution of aerosols overthe open ocean as derived from the coastal zone color scanner, GlobalBiogeochem. Cycles, 13, 383–397, 1999.

Tucker, C. J., H. E. Dregne, andW.W. Newcomb, Expansion and contractionof the Sahara Desert from 1980 to 1990, Science, 253, 299–301, 1991.

Twiss, P. C., E. Suess, and R. M. Smith, Morphological classification ofgrass phytoliths, Soil Sci. Soc. Am. Proc., 33, 109–115, 1969.

Ward, N. M., Provisionally corrected surface wind data, worldwide ocean-atmosphere surface fields, and Sahelian rainfall variability, J. Clim., 5,454–475, 1992.

Ward, N. M., P. J. Lamb, D. H. Portis, M. El Hamly, and R. Sebbari,Climate variability in Northern Africa: Understanding droughts in theSahel and the Mahgreb, in Beyond El Nino: Decadal and IntradecadalClimate Variability, edited by A. Navarra, pp. 119–140, Springer-Verlag,New York, 1999.

�����������������������L. Dupont, O. E. Romero, G. Wefer, and U. Wyputta, Fachbereich

Geowissenschaften, Universitaet Bremen, Postfach 330440, 28334 Bremen,Germany. ([email protected]; [email protected]; [email protected]; [email protected])S. Jahns, Institut fur Palynologie und Quartarwissenschaften, Wilhelm-

Weber-Straße 2, 37073 Gottingen, Germany. ([email protected])

22 - 12 ROMERO ET AL.: TEMPORAL VARIABILITY OF EOLIAN SIGNAL TRAPPED OFF NW AFRICA