Late Quaternary marine palynology in the NW African setting: biome migrations, trade wind history, and climate change
by
Hooghiemstra, Henry
Institute for Biodiversity and Ecosystem Dynamics, Faculty of Science, Kruislaan 318, 1098 SM Amsterdam, The Netherlands.

Introduction Marine palynology developed in the oil industry where it became a successful biostratigraphic tool (e.g. Muller 1959). With the initiation of the Deep Sea Drilling Project (DSDP) in the 1970’s the potential of palynological analysis was explored in Quaternary sediments. Early DSDP cruises were in the western Atlantic off Eastern USA and pollen in the marine sediments appeared to offer poor results (reference). These caused marine palynologists were seldom involved in subsequent cruises and exploration of the proxy of pollen in marine sediments was hardly explored any more in following DSDP cruises. Only a few colleagues continued the exploration of pollen in marine sediments, such as Linda Heusser in the eastern Pacific off California, and Sander van der Kaars in the area between Indonesia and Australia in particular. Both authors arrived at significant results, such as terrestrial-marine correlations and palaeogeographical maps of last glacial maximum (LGM) coastlines and biome distributions.

With the knowledge of today we can easily understand why analysis of pollen in marine sediments in the Western Atlantic, offshore Eastern USA led to poor results. In eastern USA the distribution of vegetation belts is mainly north south oriented while potential transport systems bringing pollen to the marine sediments are mainly west-east oriented in January and more or less parallel to the coastline in July. There are no zonal wind systems that bring pollen grains to the marine sediments, which could result, into interpretable distribution patterns of fossil pollen grains in marine sediments. As a consequence, the configuration of vegetation belts and potential transport systems in the area of the Eastern USA and the Western Atlantic is not suitable to arrive at meaningful results. Authors of palynological textbooks made premature general conclusions and wrote “Marine deposits are, on the whole, rather disappointing from the point of view of pollen analysis” (Faegri & Iversen 1989) not stimulating further research.

In the area of northwest Africa the environmental setting differs significantly. There is a very distinct system of wind belts (northeast trade winds; southwestern monsoon; and a high elevation jet wind at 21°N), ocean currents (Canary Current and counter currents), and a pronounced absence or presence of river systems depending dryness of climatic conditions. After having understood the operating mechanisms, the pollen signal in the marine sediments offer a wealth of information on the environmental dynamics of the adjacent continent. It is hypothesized that if DSDP had started offshore northwest Africa marine palynology had obtained a significant position in palaeoceanography.

The objective of this paper is to present the mechanisms operating in the northwest African area relevant to marine palynology. Various aspects were earlier published in but not collated into a clear synthesis. We conclude this paper with a set of observations and conclusions that might form a basis to evaluate the potential for marine palynology of other study areas in the world before starting research.

Modern vegetation distribution, wind systems, and ocean currents In northwest Africa the main vegetation belts are east west oriented and have a strong link to north-south gradients in precipitation. From north to south the main vegetation belts are the Mediterranean vegetation, steppe belt, Saharan vegetation, Sahelian vegetation (savannah), Sudan vegetation belt, Guinean vegetation belt, and the tropical rainforest belt reaching the Gulf of Guinea (White 1983).

All these vegetation belts have a distinct floral composition and character plants can be recognised by the pollen grains produced. Therefore, of individual pollen grains in the offshore marine sediments the source area can be identified. Here it should be mentioned that pollen grains identified at the generic level provide a more precise source area than pollen grains identified only up to the family level. Up to as north as the River Senegal, mangrove vegetation occurs along the coast. The main surface wind systems are northwest trade winds and southeast monsoon, meeting at the Intertropical Convergence Zone (ITCZ). The Eastern African Jet (EAJ), also mentioned as the Saharan Air layer (SAL), is operating at 3000 m altitude and transport dust and pollen from the present southern Sahara and Sahel westward; this jet wind is occurs mainly from June to September.

There is a clear seasonality in the position of the ITCZ and the flowering season of the main pollen producing plants in the different vegetation belts. The interpretation of distribution patterns should account for these annual migrations. The average wind patterns of the correct months should match with the reconstructed patterns of pollen distribution in the marine sediments.

The Canary Current and its counter currents at greater depth transport pollen grains that sink through the atmosphere to the sea surface, and enter the water column southward. A model is presented that explains the full route of a pollen grain travelling from its source area to its place of deposition and which mechanisms play a role relevant for the interpretation (Hooghiemstra 1996).

Marine sediments offshore northwest Africa located in the belts of active wind systems that blow from the continent to the Atlantic (trade wind belt, AEJ) contain large amounts of plant debris, pollen, spores, diatoms and other micro- and macrofossils. We easily could reach a relevant pollen sum. At 21°N, in the trajectory of the AEJ, sediments contained such a large component of terrigenous material that pollen in the microscopic slides often were hidden in between the large amount of plant debris. In contrast, the sediments in the Gulf of Guinea were poor in pollen and many pollen samples of an extraordinary volume (15 cm3) did not provide us with a reliable pollen sum. The reason is that humid southeast monsoon winds dominate the Gulf of Guinea. These winds only travelled above oceans and, therefore, do not supply pollen grains. However, the pollen concentration in these Gulf sediments is fluctuating with time and in the last paragraph we explain that these changes do related to variations in the trade wind intensity.

In the literature frequently is claimed that pollen grains of Pinus have a very long residence time in the atmosphere as well as in the water column (Traverse & Ginsburg 1966). Therefore, the signal of Pinus in marine sediments often is considered as not informative. However, the isopoll maps of Pinus for our study area is remarkably informative and consistent. High percentages in the marine sediments near the Iberian Peninsula reflect the pine forests in Spain, Portugal and in the Atlas Mountains. Pinus percentages decrease southwards reflecting pollen grains gradually reach the ocean floor by processes of gravitation. Immediately south of the Canary Islands, pollen percentages are much higher again evidencing that the atmosphere is re-loaded by the pine forest on this archipelago of islands. Finally isopoll percentages decrease gradually to zero near 15°N. In our study area aeolian dispersal of pollen grains of Pinus has much in common with dispersal processes of other pollen taxa.

The distribution map of pollen grains of Chenopodiaceae/Amaranthaceae, elements characteristic of desert and saline environments, in recent surface sediments is very informative. In the north, the source area of these pollen grains is the saline coastal vegetation in Morocco and the northern fringe of the Sahara. The flowering season has a focus from November to April. The northern part of the isopoll map reflects pollen transport by the trade winds; the place where isopoll lines are touching the coastline exactly reflects average wind patterns. In the central and southern Sahara the flowering season is from July to September. The central part of the isopoll map shows highest percentages that reflect pollen collected and transported by the AEJ (note that the flowering season in this area coincides with the period of maximum wind intensity). Offshore the distribution pattern shows a curve into northern direction, reflecting a component of the AEJ that circle around the permanent high-pressure cell above the Sahara into the direction of the Canary Islands. This distribution pattern is also known from maps showing the aeolian dust component in marine sediments. In the southern part of this isopoll map percentages are rapidly decreasing and reflect that air masses are entering the equatorial rain belt where the atmosphere is washed out by tropical rains. The position of the 2% and 0% isopolls reflect the average position of the ITCZ.

Stable versus dynamic wind systems through time When the modern relationship between source area (vegetation), transport systems (wind, water), and distribution pattern of pollen in the marine sediments is understood, we have to explore to which degree these relationships hold with time. In the NW African study area we used the time windows of 9000 radiocarbon years before present (noted as ‘9000 yr BP’) and 18,000 yr BP in a series of marine pollen records that form a transect from Portugal to the Gulf of Guinea. Based on the composition of the pollen spectra from both time windows a palaeo-distribution map was drawn. It was concluded that the trade wind belt had a stable position in time (based on distribution patterns), but fluctuated in intensity (based on pollen influx records). It was concluded that the African Easterly Jet also had a stable position around 21°N (based on the position of maximum isopoll lines). Distribution of the rainforest and adjacent wet forest belts migrated in accordance with the northernmost position of the ITCZ, providing the precipitation on which these vegetation types depend.

Comparison of modern distribution patterns in land and marine data sets Recent pollen rain data from surface soil sediments in coastal northwest Africa between 12°and 21°N, and recent pollen rain data from surface sediments collected on the ocean floor between similar latitudes allowed to compare pollen distribution patterns on land and sea (Lézine & Hooghiemstra 1990). Modern vegetation belts closely correspond to the latitudinal zones that could be recognised in the modern pollen rain from the soil samples. This correspondence was less precise, but still significant, when modern vegetation belts were compared with the modern pollen rain in the samples from the ocean floor. Less pronounced correspondence is obvious as the additional factor of transport in marine sediment samples delutes isopoll lines. The border between Poaceae-dominated savannah vegetation and Chenopodiaceae-dominated Sahara vegetation was very clearly represented in the offshore surface sediments: in real vegetation as well as in marine surface sediments this border lies at 18°N.

The dynamic Saharan-Sahelian boundary Having identified that the transition from Poaceae dominance to Chenopodiaceae dominance in the marine surface sediments reflects the southern border of the Sahara, and having evidenced that the position of the AEJ is stable over time, we have an elegant mechanism to trace migrations of the southern border of the Sahara. When the Sahara is large, and extends far to the south (i.e. the northernmost position of the monsoon front is close to the equator), the AEJ mainly transports Chenopodiaceae pollen grains. This situation was the case e.g. 18,000 yr BP. When the Sahara is small (i.e. the northernmost position of the monsoon front is migrating annually far to the north), the AEJ mainly transports Poaceae pollen grains. This situation was the case from 9000 to 8000 yr BP and reflects the period of the ‘Green Sahara’ when nomadic people with their cattle lived in the present-day inhospitable Central Sahara. Calculating the Poaceae/Chenopodiaceae ratio in ODP site 658 the migrations of the southern border of the Sahara were reconstructed for the past 650,000 yr BP; the maximum migration distance amounts some 9 degrees of latitude (Dupont & Hooghiemstra 1989).

The marine pollen record from Cap Blanc (21°N) clearly documented the extremely dry conditions in present-day Sahel zone: only from 14,000 yr BP onward the Senegal River supplied water to the Atlantic Ocean and from that time onward mangrove vegetation developed at the coast (Hooghiemstra 1988). The LGM was very dry, the Sahara reached far to the south, the West Equatorial rainforest was fragmented and limited to small remnants (glacial forest refuges). Also geomorphological evidence from the valley of the Senegal River showed dry LGM conditions (the Ogolian period), and the Senegal River started to supply water to the Atlantic again around 14,000 yr BP (Michel 1984).

The dynamic Saharan-Mediterranean boundary At latitudes of the northern border of the Sahara marine pollen records are available that represent the last 15,000 yr BP (off Portugal), and the last 130,000 to 250,000 yr BP (near the Canary Islands) (Hooghiemstra et al. 1992). These records clearly show that in the Western Mediterranean interglacial Quercus-dominated forest is replaced during glacial time by Artemisia-dominated steppe vegetation with Pinus wood. The steppe belt, located between the Saharan vegetation in the south and in the north the Mediterranean oak forest during interglacials and Pinus wooded steppe vegetation during glacials, oscillated in width over maximally some 4 degrees of latitude. The Atlas Mountains prevented a larger migration of the northern border of the Sahara.

Notwithstanding the interpretation of marine pollen records in northwest Africa has reached significant accuracy and reliability; reconstructions of environmental change integrate evidence from a large area on the continent. Evidence from terrestrial sites to support the integrated view from marine sediments is needed. Several terrestrial sites from Spain, Morocco, Senegal and Sudan are in support of the dynamic character of the Sahara and the latitudinally migrations of northwest African biomes.

A record of changing intensity of the NW trade winds Seven marine pollen records located on a transect from offshore Portugal (34°N) via the Canary Islands to the Gulf of Guinea (7°N) were transformed into pollen influx records (number of pollen grains. cm-2 year-1) representing the last 140,000 yr BP (Hooghiemstra 1989). Time control was based on radiocarbon dating and the marine oxygen isotope (MIS) stratigraphy. Marine palynology has the advantage that the terrestrial and marine sediment components can be studied in the same core making a direct land-sea comparison possible. High pollen influx values were characteristic of MIS 6, MIS 5d and 5b, MIS 4, and the period of 40,000 to 10,000 yr BP. On a global scale all these intervals reflect relatively cold climatic conditions. During colder conditions the temperature difference between poles and equator is larger which results in more vigorous wind systems. As higher pollen influx values can be interpreted as more efficient pollen transport by more vigorous wind systems, the link between cold episodes and relatively high pollen influx values is obvious. The period from 40,000 yr BP to the Lateglacial-Holocene transition, at 10,000 yr BP, was characterised by vigorous trade winds in particular. Also analysis of sediment fractions confirmed this pollen-based record of changes in the intensity of the northeast trade winds.

Some general conclusions · marine palynology is a very informative addition to the present-day standard proxies used in palaeoceanographic studies in areas where a significant terrigenous input can be expected · pollen supply to offshore sediments by riverine transport leads to coastal point-sources; in case the hinterland of the drainage system contains different vegetation belts it is difficult to separate between them · northwest Africa offers an ideal setting: vegetation belts organised along a north-south gradient, and wind systems parallel (trade winds) and perpendicular (AEJ) on this gradient · the group of Pinus pollen is not necessarily an ‘omni-present and information lacking’ taxon; in the northwest African area transport distance is limited, distribution patterns are meaningful, and point sources (pine forests on the Canary islands) clearly re-load the atmospheric pollen content · in areas with a seasonal climate system, the main flowering season should be identified and compared with the average wind directions of the same period · preferably, a comparison between terrestrial and marine modern pollen rain samples should be made in order to verify to which degree distribution patterns reflect actual vegetation belts · in case time control of sediment cores is sufficient, pollen influx records are very helpful to assess changes in vigour of the transporting wind system · designers of palaeoceanographic research projects are advised to give marine palynological analysis a proper place in their planning when conditions in the study area are favourable. Often it is possible to anticipate which mechanisms potentially operate under conditions different from today.

Acknowledgements This research was carried out at the University of Göttingen during an almost 4 years post-doc position. I thank H.-J. Beug and E. Grüger (both Palynological Institute, Göttingen), M. Sarnthein (Geological Institute, Kiel), G. Tetzlaff (Meteorological Institute, Hannover) for support and cooperation.

Hooghiemstra, H. & Agwu, C.O.C., 1986. Distribution of palynomorphs in marine sediments: a record for seasonal wind patterns over NW Africa and adjacent Atlantic. Geologische Rundschau 75, 81-95.

Hooghiemstra, H., Agwu, C.O.C. & Beug, H.-J., 1986. Pollen and spore distribution in recent marine sediments: a record of NW-African seasonal wind patterns and vegetation belts. 'Meteor'Forschungs Ergebnisse, Reihe C 40, 87-135.

Hooghiemstra, H., 1987. Survey of palynologically analysed deep-sea cores in the northeast Atlantic off northwest Africa. Palaeoecology of Africa 18, 47-52.

Hooghiemstra, H., Bechler, A. & Beug, H.-J., 1987. Isopollen maps for 18,000 years BP of the Atlantic offshore of northwest Africa: evidence for paleowind circulation. Paleoceanography 2, 561-582.

Hooghiemstra, H., 1988. Changes of major wind belts and vegetation zones in NW Africa 20,000-5000 yr BP as deduced from a marine pollen record near Cap Blanc. Review of Palaeobotany and Palynology 55, 101-140.

Hooghiemstra, H., 1988. Palynological records from northwest African marine sediments: a general outline of the interpretation of the pollen signal. Philosophical Transactions Royal Society London B 318, 431-449.

Hooghiemstra, H., 1988. Late Quaternary changes in vegetation and climate in NW Africa: pollen evidence from marine sediments (time-slices and time-series): extended abstract. Palaeoecology of Africa 19, 243-248.

Hooghiemstra, H. & Agwu, C.O.C., 1988. Changes in the vegetation and trade winds in equatorial northwest Africa 140,000-70,000 yr BP as deduced from two marine pollen records. Palaeogeography, Palaeoclimatology, Palaeoecology 66, 173-213.

Hooghiemstra, H., 1989. Variations of the NW African trade wind regime during the last 140,000 years: changes in pollen flux evidenced by marine sediment records. In: Leinen, M. & Sarnthein, M. (eds), Paleoclimatology and paleometeorology: modern and past patterns of global atmospheric transport, pp. 733-770. NATO ASI Series, Kluwer, Dordrecht. ISBN 0-7923-0341-5 Hooghiemstra, H., 1989. Quaternary and Upper Pliocene glaciations and forest development in the tropical Andes: evidence from a long high-resolution pollen record from the sedimentary basin of Bogotá, Colombia. Palaeogeography, Palaeoclimatology, Palaeoecology 72, 11-26.

Dupont, L.M. & Hooghiemstra, H., 1989. The Saharan-Sahelian boundary during the Brunhes chron. Acta Botanica Neerlandica 38, 405-415.

Lézine, A.M. & Hooghiemstra, H., 1990. Land-sea comparisons during the last glacial-interglacial transition: pollen records from West Tropical Africa. Palaeogeography, Palaeoclimatology, Palaeoecology 79, 313-331.

Hooghiemstra, H., 1996. Aspects of Neogene-Quaternary environmental and climatic change in equatorial and Saharan Africa. Palaeoecology of Africa 24, 115-132.

Date received: January 27, 2004