Interactive comment on “ Increased aridity in southwestern Africa during the last-interglacial warmest periods ” by D . H . Urrego

The article contains an interesting marine pollen sequence but I feel it is my duty as someone who is acknowledged in it, to point out some serious short comings that should be corrected before this paper is published. There are problems with the modern pollen control, a lack of consideration of previous literature in the study region and some unwarranted speculations that result in vague and meaningless interpretation of vegetation changes on the Southern African continent. The pollen sequence without the modern pollen results may be publishable provided but that the interpretation is done very carefully and the limitations of marine pollen spectra in tracking vegetation boundaries on land are realize . The modern pollen survey seems to be flawed and I recommend that it should not be published. I believe too much is read from the pollen


Introduction
Southern Africa is recognised as a climatic (Giorgi, 2006), biodiversity and conservation hotspot (Myers et al., 2000) and as a region of crucial importance for the Figures development of early modern humans (Henshilwood and Dubreuil, 2011).Vegetation and climate changes are potential triggers of cultural innovations.Understanding the processes involved in the evolution of southern African (SAfr) environments is hence relevant both for human and climate sciences.Southern Africa is influenced by tropical and subtropical atmospheric circulation linked to ITCZ shifts, and by both the Indian and the Atlantic Oceans (Tyson and Preston-Whyte, 2000).The water exchange between the two oceans is termed the Agulhas leakage and is documented to vary significantly during glacial vs. interglacials (Peeters et al., 2004).The Agulhas leakage is also suggested as a potential trigger of changes in the meridional overturning circulation (Beal et al., 2011), hence the importance of SAfr for Earth's climate.Climate in southwestern Africa (SWAfr) is also affected by the Benguela upwelling system (BUS), whose strength determines the extent of the coastal Namib Desert (Cowling et al., 1997b).Vegetation and climate changes have been documented for SAfr at orbital and suborbital timescales based on a few marine (Dupont, 2011) and fragmentary terrestrial pollen sequences (Scott et al., 2012;Meadows et al., 2010).Putting together a comprehensive picture of paleoenvironmental conditions in SAfr has been hence controversial.Previous reconstructions point to expansions of the Fynbos biome (Shi et al., 2001) and the winter-rainfall zone during glacial periods (Chase and Meadows, 2007), and to a contracting (Shi et al., 2000) and relatively humid (Scott et al., 2012) Namibian Desert during the Holocene.It is also suggested that savannas expanded southwards during the Holocene climate optimum (Dupont, 2011), and that the SAfr summer-rainfall zone expanded during interglacials due to a strengthening of the walker circulation and an increased dominance of the ITCZ (Tyson, 1999).Contrastingly, significant reductions of austral-summer precipitation in SAfr are suggested to coincide with precession minima both during glacials and interglacials (Partridge et al., 1997), and are independently supported by reductions of grass-fuelled fires in the subcontinent (Daniau et al., 2013).
To disentangle these contrasting hypotheses, we conduct the pollen analysis of marine sequence MD96-2098 collected 200 km off the Namibian coast.In arid Figures

Back Close
Full environments like those of SWAfri, marine sequences can provide continuous records of vegetation change at the regional scale, and help unravel climatic signals obscured by local effects in terrestrial sequences.Furthermore, marine pollen sequences permit to correlate directly changes on land to changes in oceanic circulation and ice dynamics.Marine sequences are therefore pivotal to understanding vegetation and climate dynamics at orbital and suborbital timescales.Additionally, we improve the interpretation of the marine pollen record by performing a pollen-vegetation-climate calibration.
We analyse a north-south transect of 31 pollen samples that are integrated to the marine and terrestrial surface samples compiled by Dupont and Wyputta (2003) and Gajewski et al. (2002), respectively.The present-day pollen-vegetation-climate spatial results are used to infer precipitation variations in SWAfr through time.We correlate these variations with independent climatic markers from the same sample set, along with other regional records for oceanic conditions and global ice dynamics, to reconstruct atmospheric and oceanic configurations around SAfr at orbital timescales.

Modern environmental setting
Climate in SAfr is influenced by oceanic circulation both in the Indian and the Atlantic Oceans and by tropical and subtropical atmospheric circulation (Tyson and Preston-Whyte, 2000).In the Atlantic Ocean, the influence of the cold Benguela current and offshore-wind driven upwelling results in aridity on the adjacent continent (Stuut and Lamy, 2004).In the Indian Ocean, warm waters from the Agulhas current (Beal and Bryden, 1999) and austral-summer heat allow evaporation and relatively high precipitations in southeastern Africa and the interior of the continent (Fig. 1).Austral-summer precipitations are also linked to the position of tropical low pressure systems (i.e.ITCZ) and reduced subtropical high pressure (Tyson and Preston-Whyte, 2000).the southern westerlies that bring precipitation during the austral-winter months (Beal et al., 2011).Biomes of SAfr were initially classified by White (1983), and later revisited and described into seven units by Rutherford (1997).These include the Succulent-Karoo, Nama-Karoo, Desert, savanna, Fynbos, Grassland, and forest (Fig. 1).The Succulent-Karoo receives between 20 and 290 mm yr −1 of which more than 40 % falls during the austral-winter months (Rutherford, 1997).The two most abundant succulent families are Crassulaceae and Mesembryanthemaceae, and non-succulents are Anacardiaceae, Asteraceae, and Fabaceae (Milton et al., 1997).C4 perennial grasses (Poaceae) have relatively low abundance in the Succulent-Karoo (Milton et al., 1997).
The Nama-Karoo receives precipitations from 60 to 400 mm yr −1 falling primarily during the austral summer (Palmer and Hoffman, 1997).Vegetation is characterized as dwarf open shrubland with high abundance of Asteraceae, Poaceae, Aizoaceae, Mesembryanthemacae, Liliaceae and Scrophulariaceae (Palmer and Hoffman, 1997;Milton et al., 1997).The Nama-and Succulent-Karoo are structurally similar but influenced by different seasonal precipitations (Rutherford, 1997).To the northwest, the Nama-Karoo biome transitions into the Desert, where mean annual precipitation can be as low as 20 mm yr −1 (Jurgens et al., 1997).The Desert reaches 300 km inland and its extent is linked to the intensity of BUS (Lutjeharms and Meeuwis, 1987).
High precipitation seasonality and austral-summer rainfall characterize the savanna.
It represents a continuum that includes shrublands, dry forests, lightly-wooded Grasslands, and deciduous forests (Scholes, 1997).At the landscape scale however, the savanna can be subdivided into the fine-and broad-leaved savannas based on moisture conditions and soils (Scholes, 1997).The fine-leaved savanna (Fn-LSav) is associated with arid and fertile environments (between 400 and 800 mm yr −1 ), and the broad-leaved savanna (Bd-LSav) is associated with nutrient-poor and moist environments (up to 1500 mm yr −1 ) (Scholes, 1997).Additionally, in the Fn-LSav fuel load and fire frequency are very low, while the Bd-LSav has high fuel load and fire frequency (Scholes, 1997;Archibald et al., 2010).The Fn-LSav is found to the northeast of the Figures

Back Close
Full Nama-Karoo biome (Fig. 1), known as the Kalahari Highveld transition zone (Cowling and Hilton-Taylor, 2009).Due to the transitional character of the Fn-LSav, some of its outer parts can be included in the Grassland or Nama-Karoo biomes.The composition of the Fn-LSav can be similar to that of the Nama-Karoo, with dominance of C4 grasses (Poaceae) and succulent plants, however it differs in having scattered trees (Cowling et al., 1994).The Bd-LSav is characterized by broad-leaved trees from the Caesalpinaceae and Combretaceae families and an under growing dominated by grasses (Scholes, 1997).
The Grassland biome is dominated by C4 grasses and non-grassy forbs as Anthospermum sp., Lycium sp., Solanum sp. and Pentzia sp.(O'Connor and Bredenkamp, 1997).In the high elevations the biome is dominated by C3 grasses.In the Grasslands, precipitation is highly seasonal with mean annual rainfall ranging between 750 and over 1200 mm, and falling primarily during the austral-summer months (O'Connor and Bredenkamp, 1997) (Fig. 1).
The southernmost part of Africa is dominated by the Fynbos biome, a fire-prone vegetation form dominated by Ericaceous and Asteraceae shrubs, diverse Proteas and Restionaceae herbs (Cowling et al., 1997a).The Fynbos biome is subject to low precipitation seasonality and relatively high annual precipitation (1200 mm per year) concentrated during the austral-winter months (Rutherford, 1997).The coastal forest biome is found in the eastern coast of the subcontinent and often occurs in small patches with high abundance of Podocarpus (Rutherford, 1997).

Surface-sample collection and analysis
A total of 31 surface samples were collected in a transect from Cape Town (Fig. 1) and was designed to cover the four major biomes of SWAfr (Table 1 in the Supplement): Desert, Fynbos, Nama-and Succulent-Karoo.Although we did not conduct a vegetation survey at each sampling site, the surrounding vegetation was determined based on detailed descriptions and maps of SWAfr biomes (Mucina et al., 2007).We were also able to collect one sample in the coastal forest biome.While we did not intend to characterize the pollen spectra from coastal forests, including this sample in the dataset allowed to increase the variability of pollen spectra.Sediment samples were collected from small, probably ephemeral, water-logged depressions where pollen preservation was likely.When we found moss attached to rocks or soil, we collected pitches from several spots within a five meter radius.Water samples were also collected from small streams or rivers when neither water-logged depressions nor moss were found.As a result, our sample set included 12 sediment, 8 moss, and 11 water samples (Table 1 in the Supplement).Surface samples were concentrated down to pellets using a manual and portable centrifuge in the field.Pellets were treated with standard acetolysis in the laboratory (Faegri and Iversen, 1989) and residues were mounted in glycerol and scanned under the microscope at 400 and 1000 magnification.Pollen sums were greater than 300 grains, and spores were quantified but excluded from this total.Four out of 31 surface samples had such low pollen concentration that their spectra were excluded from the analysis (Table 1 in the Supplement).
Pollen identification was aided by the pollen reference collection of the Department of Plant Sciences at University of the Free State, Bloemfontein, South Africa, the African Pollen Database (APD) (http://medias3.mediasfrance.org/pollen), the Universal Pollen Collection (http://www.palyno.org/pollen),and pollen descriptions published by Scott (1982).Pollen grains from the Asteraceae family were grouped into three pollen taxa: Artemisia-type, Stoebe-type and other morphotypes were classified into Asteraceae-other.Artemisia-type likely corresponds to the Asteraceae-low spine reported by other authors (Meadows et al., 2010).Stoebe-type may include pollen from Tarchonanthus sp.due to similarities in their pollen morphology.Pollen Figures

Back Close
Full percentages from more than one morphotype were grouped into the families Acanthaceae, Chenopodiaceae-Amaranthaceae, Crassulaceae, Cyperaceae, Ericaceae, Myrtaceae, Ranunculaceae, Restionaceae, and Solanaceae.Canonical correspondence analysis (CCA) was used to summarize changes in pollen spectra from surface samples and explore their relationship with bioclimatic variables (McCune and Grace, 2002).Four temperature and four precipitation bioclimatic variables were extracted for each sampling point from the WorldClim Global Climate dataset (Hijmans et al., 2005), but only variables with the most constraining powers were preserved.Ordinations were run filtering out aquatic taxa to remove local effects and taxa that only occurred in one sample to avoid overweighting rare taxa.We also used previously published pollen spectra from 150 additional surface samples collected between 22 and 35 • latitude south (African Pollen Database, APD, Gajewski et al., 2002) to assess the pollen-percentage distribution, and potential as indicators of large biomes, of seven abundant pollen taxa over SAfr.These taxa included Artemisia-type, Asteraceae-other, Chenopodiaceae-Amaranthaceae, Poaceae, Podocarpus, Restionaceae, and Stoebe-type.ArcGIS 10 was used to draw iso-lines of pollen percentages by interpolating values from a total of 178 surface samples through the natural neighbour method.Maps of bioclimatic variables were also drawn for comparison.

Marine core description and pollen analysis
Pollen analysis was conducted on the marine core MD96-2098 ( 25• 36 S, 12 • 38 E).
This giant CALYPSO core was collected during the IMAGES II-NAUSICAA cruise at a 2910-m water depth from the L üderitz slope in the Walvis Basin, approximately 500 km northwest of the Orange River mouth (Fig. 1).The sediments of this 32-m long core were composed of calcium carbonates (foraminifera), biogenic silica (nannofossil mud), clays and organic matter (Bertrand et al., 1996).The core was sampled every 10 cm between 450 and 1940 cm (uncorrected depth) for pollen analysis.Additionally, two marine pollen samples from the upper part of core MD96-2098 (5 and 10 cm depth) were Figures

Back Close
Full analysed and compared with the pollen signal of the modern vegetation to evaluate how well marine sediments represent the vegetation of the adjacent landmasses, and to aid interpretation of the pollen record.Sample volumes were estimated by water displacement and pollen concentrations per unit volume were calculated based on a known spike of exotic Lycopodium spores added to each sample.Pollen extraction techniques included treatment with Hydrofluoric and Hydrochloric acids, and sieving through 150 and 10-µm filters.This filtration allowed eliminating small non-palynomorph particles and concentrating pollen grains and spores.An independent test of this protocol showed that the use of a 10-µm sieve had no effect on the pollen composition of marine samples, i.e. comparison of filtered and unfiltered samples confirmed that neither Poaceae nor other taxa were selectively filtered out during pollen preparation and concentration (see http://www.ephe-paleoclimat.com/ephe/Lab%20Facilities.htm for a detailed pollen preparation protocol).To summarize changes in the fossil pollen record over time, detrended correspondence analysis (DCA) was used in tandem with Nonmetric multidimensional scaling (NMDS) as parametric and non-parametric alternatives.Results from the DCA ordination were preferred when NMDS was unable to reach a stable solution after several random starts, and when stress levels were too high to allow a meaningful interpretation (McCune and Grace, 2002).These ordinations were run filtering out taxa that only occurred in one sample to avoid overweighting rare taxa.

Marine core chronology
An age model based on 16 marine isotope events (MIE) from the Cibicidoides wuellerstorfi δ 18 O benthic record of MD96-2098 (Bertrand et al., 2002) and 14 Accelerator Mass Spectrometer radiocarbon ages (AMS 14 C) produced at the Laboratoire de Mesure du Carbone 14 from mixed planktonic foraminifera was established for the core (Table 2 in the Supplement).One single 14 C date showed an age reversal and was therefore excluded from the chronology on the principle of parsimony.AMS 14 C ages were calibrated using the marine09.14ccurve (Hughen et al., 2004) from CALIB REV6.0.0 (Stuiver and Reimer, 1993).We applied a 400-yr global reservoir Introduction

Conclusions References
Tables Figures

Back Close
Full correction factor and a weighted mean Delta R of 157 yr derived from 9 regional reservoir error values from the Marine Reservoir Correction Dataset (Dewar et al., 2012;Southon et al., 2002).MIE ages were derived from LR04 global stack (Lisiecki and Raymo, 2005) and additional sources (Henderson and Slowey, 2000;Drysdale et al., 2007;Waelbroeck et al., 2008;Masson-Delmotte et al., 2010;Sanchez Go ñi and Harrison, 2010) (Fig. 1 in the Supplement).Sample ages were calculated using a linear interpolation between AMS 14 C ages and MIE using the R package PaleoMAS (Correa-Metrio et al., 2010).

Present-day pollen-vegetation-climate relationships in southern Africa
Asteraceae-other, Chenopodiaceae-Amaranthaceae, Poaceae and Restionaceae add up to 80 % of the pollen sums and are found in all surface samples (Fig. 2 in the Supplement).Cyperaceae pollen is also found in all but one of our surface samples.High Cyperaceae percentages probably reflect local abundances of sedges growing around small waterlogged depressions or along rivers as we deliberately targeted sites where pollen preservation was likely.As a result, we excluded Cyperaceae pollen percentages from ordination analyses.
In the surface samples collected in the Desert, Asteraceae-other percentages are up to 20 % and Chenopodiaceae-Amaranthaceae percentages are less than 10 %.Poaceae and Cyperaceae show up to 45 % and between 50 and 60 %, respectively.The source of Poaceae, Asteraceae-other and Chenopodiaceae-Amaranthaceae pollen in the Desert can be perennial grasses growing on dunes (Jurgens et al., 1997).Our results show that pollen spectra from the Desert are more similar to spectra from the Succulent-Karoo than to any other biome (Fig. 2a), suggesting that Desert surface samples receive pollen from Succulent-Karoo transitional patches.Introduction

Conclusions References
Tables Figures

Back Close
Full In the Nama-Karoo surface samples, Poaceae pollen reaches values up to 60 % (Fig. 2 in the Supplement).Asteraceae-other pollen percentages are null in northern samples but increase to 20 % in samples collected near the Succulent-Karoo.Chenopodiaceae-Amaranthaceae pollen percentages are up to 10 %.Tribulus and Acanthaceae pollen are only found in the Nama-Karoo samples and reach up to 12 %, consistent with their abundance in the vegetation source.On the other hand, Crassulaceae and Euphorbia pollen occur at low percentages in samples from Nama-Karoo despite having been described as common in the vegetation (Honig et al., 1992).The pollen spectra from samples collected in the Nama-Karoo form a tight cluster in the DCA ordination and are clearly separated from samples from other biomes (Fig. 2a).
Surface samples collected in the Succulent-Karoo are characterized by pollen percentages of Asteraceae-other up to 20 % and Poaceae between 15 and 20 % (Fig. 2 in the Supplement).This result corresponds with the abundance of species from the Asteraceae family in this biome, and the relatively less abundance of C4 grasses compared to the Nama-Karoo (Cowling and Hilton-Taylor, 2009).Small percentages of Olea and Podocarpus also found in the pollen spectra of the Succulent-Karoo could be the result of long-distance wind transport.The DCA ordination groups the Succulent-Karoo pollen spectra near Desert samples, and this clustering is constrained by Asteraceaeother, Euphorbia and Chenopodiaceae-Amaranthaceae (Fig. 2a).
In composition of pollen spectra from the Fynbos biome in the DCA ordination is clearly distinguished from pollen spectra from other biomes (Fig. 2a).The classification of sample Fy22 near samples from the Nama-Karoo likely results from the relatively high abundance of Poaceae pollen in Fy22 compared to other Fynbos samples (Fig. 2 in the Supplement).
The spatial variability of pollen spectra from biomes in SWAfr is closely associated with precipitation variability and seasonality (Fig. 2b).Desert pollen spectra are linked to precipitation seasonality (PS), and plot in the opposite direction of all other precipitation variables.Similar to Desert samples, the Succulent-Karoo pollen spectra are associated with PS and somewhat linked to PCQ (coldest quarter precipitation), although less than Fynbos pollen spectra.Fynbos pollen spectra are linked to all precipitation variables except PS.Pollen spectra from the Nama-Karoo are correlated with PS, are somewhat linked to warmest quarter precipitation (PWQ) and clustered opposite PCQ.
The four temperature bioclimatic vectors plot in the same ordination quadrant (Fig. 3 in the Supplement), indicating they are correlated.In addition, Nama-Karoo pollen spectra are linked to all of these temperature bioclimatic variables.These results indicate that temperature does not appear to have as much constraining power as precipitation bioclimatic variables when classifying pollen spectra from SWAfr biomes, except in the case of the Nama-Karoo.
Some individual taxa are associated with the clustering of pollen spectra from the SWAfr biomes in the DCA ordination (Fig. 2a), suggesting their potential as indicators of specific biomes.Pollen percentage iso-lines drawn for six of these abundant taxa using our pollen surface samples and those of the APD complement the picture of pollen percentage distribution in the subcontinent (Fig. 3).These six taxa are abundant both in terrestrial and marine pollen sequences (Dupont, 2011;Scott et al., 2012) and can therefore be valuable for the interpretation of fossil pollen records.The pollen iso-lines of Asteraceaeother show 25 % near the transition of the Nama-Karoo and the Grassland, and 40 % in part of the Succulent-Karoo (Fig. 3).Chenopodiaceae-Amaranthaceae pollen percentages are as high as 40 % in the Nama-Karoo and Desert biomes, and are also found up to 40 % in the transition between the Nama-Karoo and the Grasslands.This pollen distribution indicates that Chenopodiaceae-Amaranthaceae and Asteraceae-other high pollen percentages can be characteristic of the Succulent-Karoo, Nama-Karoo and Desert biomes of SWAfr.
High Podocarpus pollen percentages in surface samples from SAfr show a localized pattern in areas with high precipitation, namely coastal forest and in the eastern part of the Fynbos biome (Fig. 3) and in the eastern highlands.Consistently, Gajewski et al. (2002) reports maxima of Podocarpus pollen percentages in African regions where precipitation is at least 1000 mm per year.
Restionaceae plants are found mostly in the Fynbos biome (Cowling et al., 1997b) and its pollen has been used as a Fynbos indicator (Shi et al., 2001).However, the distribution of its pollen in our surface samples is only partly related to the distribution of the Fynbos biome (less than 5 %) (Fig. 3).Up to 4 % of Restionaceae pollen is found in surface samples from the Nama-Karoo, Succulent-Karoo and the Desert (Fig. 3), suggesting that these pollen grains are the result of long-distance transport.Restionaceae are wind pollinated (Honig et al., 1992), hence its pollen is likely transported by wind far from the vegetation source (Fig. 1).Due to this inconsistency between the vegetation source and the spatial distribution of Restionaceae pollen, it would be difficult to discern whether increases in Restionaceae pollen in palaeoenvironmental reconstructions are the result of wind strengthening or due to Fynbos vegetation expansions, without an independent tracer for past wind strength.The distribution of pollen percentages from Artemisia-type and Stoebe-type are concentrated in the Fynbos biome (Fig. 3), and are positively correlated with PCQ (Fig. 2).Pollen signals from Artemisia-type and Stoebe-type, along with other taxa characteristic of the Fynbos vegetation (i.e.Ericaceae, Protea and Passerina, Fig. 2 in the Supplement) might therefore be good tracers for past expansions of the biome.

Poaceae pollen percentage distribution in terrestrial and marine surface samples and its interpretation
Occurrence of Poaceae pollen in all surface samples corresponds to the presence of grass species in virtually all SAfr biomes (Cowling et al., 1997b).However, the spatial distribution of Poaceae pollen percentages appears to be essential information to distinguish major biomes from their pollen spectra.
In the center part of SAfr, Poaceae pollen percentages in terrestrial surface samples up to 60 % in the Nama-Karoo and Fn-LSav, and in the transition between the Fn-LSav and the Grassland (Fig. 4).This suggests an overrepresentation of Poaceae in the pollen rain of the Nama-Karoo biome where grasses can be abundant but are not necessarily dominant.Poaceae is likely to be well represented in other parts of the Fn-LSav, but the paucity of surface samples North of 25 • S within this biome hinder drawing further conclusions.In the Namib Desert where abundance of grasses in the vegetation is low, Poaceae pollen percentages are as high as 30 %, comparable to 20 % reported from hyrax dung (Scott et al., 2004).In the eastern and northeastern part of SAfr, the highest percentages of Poaceae pollen (up to 90 %) are found in the pollen rain of the Bd-LSav.We recognize that the number of terrestrial surface samples available for this calibration is limited, specially for the savanna biome.However, we are able to characterize the pollen spectra from the SWAfr biomes that are most likely recorded in our marine record.
Grazing and agriculture could artificially increase the abundance of Poaceae in surface samples from SAfr, thus complicating the usage of Poaceae pollen to discriminate biomes.We are unable to discard this bias from surface samples extracted from the Figures

Back Close
Full APD.However, we have no record of cultivated Poaceae pollen (i.e.cereals) in the surface samples we analysed.We also collected samples well outside grazing zones.
In marine surface samples along the SWAfr coast, Poaceae pollen percentages are as low as 10 % in samples collected in front of the Bd-LSav at around 15 • S (Fig. 4).Poaceae pollen percentages increase to the South and the highest values (40 %) are found between 20 and 25 • S (Dupont and Wyputta, 2003) and correspond well with the distribution of the Desert and the Fn-LSav on the continent.The Poaceae pollen percentages from the two core-top samples from MD96-2098 dating to the last millennium extend the iso-lines drawn by Dupont and Wyputta (2003) to 25.5 • S, and show that Poaceae pollen percentages are between 30 and 40 % in front of the Desert, Nama-Karoo and Fn-LSav biomes.As Poaceae pollen percentages in Desert surface samples are less than 25 %, high percentages of grass pollen in marine sediments in the SWAfr margin should be interpreted as indicator of the Nama-Karoo and the Fn-LSav, where Poaceae is as high as 60 % in terrestrial surface samples.

Pollen preservation and sources in marine core MD96-2098
Pollen sums excluding spores range from 100 to 240 grains in 141 samples analysed from core MD96-2098, with a mean of 21 taxa per sample.Total pollen concentration changes importantly over the record, ranging between ca.300 and 16 000 grains cm −3 during most of MIS 6, 5 and 3 and increasing up to 48 000 grains cm −3 during MIS 4 (Fig. 4 in the Supplement).The pollen concentrations are relatively low when compared with terrestrial sites, even though BUS facilitates preservation of pollen grains and other organic microfossils at this site (Bertrand et al., 2003).The low net primary productivity that characterizes SWAfr (Imhoff et al., 2004) is partly due to reduced low annual precipitations (Cowling et al., 1997b) and can explain our relatively low pollen concentrations (Fig. 4).
Pollen grains are part of the fine sediment fraction and can be transported by two main vectors: aeolian or fluvial (Hooghiemstra et al., 1986;Heusser and Balsam, 1977).Dupont and Wyputta (2003) modelled present-day wind trajectories for marine core Introduction

Conclusions References
Tables Figures

Back Close
Full locations between 6 and 30 • S along the coastline of SAfr.They suggest aeolian pollen input to the Walvis area ( 23• S) via the south-east trade winds during austral summer, and dominant east-to-west wind directions during the austral fall and winter.These winds transport of pollen and other terrestrial particles from the Namib Desert, southern Namibia and western South Africa and suggest the Namib Desert and the semi-Desert (corresponding to the Nama-Karoo and Succulent-Karoo areas) as the main vegetation sources of pollen in the Walvis area (Dupont and Wyputta, 2003).The authors also infer that south of 25 • S wind directions are predominantly west to east and aeolian terrestrial input very low.Marine site MD96-2098 is located less than a degree south of the area determined by Dupont and Wyputta (2003) to be dominated by terrestrial aeolian input, and given that this threshold was established using only two marine sites located far apart at 23 • 26 (GeoB1710-3) and 29 • 27 (GeoB1722-1), it is difficult to conclude that MD96-2098 only receives wind-transported pollen.Sedimentological analyses of the Orange River delta and plume indicate that fine muds are transported both northwards and southwards (Rogers and Rau, 2006).The authors also infer that the fine fraction of the Orange River plume travels northwards of the mouth aided by oceanic currents.The input of terrestrial particles, including pollen, to marine site MD96-2098 from the Orange River catchment area is hence likely.
The contribution of both aeolian and fluvial processes is likely even during past glacial and interglacial periods.Analyses of aeolian vs river-transported clays from MD96-2098 show that both mechanisms were active and that their contribution varies over time independently of glacial and interglacial conditions (Daniau et al., 2013).Scott et al. (2004) argue that pollen in marine sediments can be the result of longdistance transport by ocean currents, suggesting that pollen assemblages in marine sediments do not reflect accurately past changes in vegetation and climate.However, the highest pollen influx in marine sediments along this margin is not found along the paths of oceanic currents (i.e.Benguela Current), but near the coast and the vegetation source (Dupont et al., 2007).Additionally, analyses of pollen transport vs. source in northwestern Africa show that pollen grains can sink rapidly in the water column Introduction

Conclusions References
Tables Figures

Back Close
Full  (Hooghiemstra et al., 1986) before they can be carried away by ocean currents.As a result, influence of oceanic currents on the composition of pollen assemblages is probably negligible.The marine site MD96-2098 might therefore receive both aeolian and fluvial pollen input from the vegetation located east and southeast of the site.In addition, the pollen composition of two core-top samples from core MD96-2098 are dominated by Poaceae (30 and 40 %), Cyperaceae (20 %) and Chenopodiaceae-Amaranthaceae (20 and 30 %) (Fig. 4 in the Supplement).This composition corresponds well with the pollen spectra from the three major biomes, Desert, Nama-Karoo and Fn-LSav (Figs. 3  and 4) occupying today the adjacent landmasses (Fig. 1a).Pollen percentages from Fynbos are less than 10 %.These results support the assumption that the main pollen source from marine core MD96-2098 is the vegetation east and southeast of the site.

Southwestern Africa vegetation and climatic changes from MIS 6 to 2
The pollen record presented here spans from 24.7 to 190 ka.A log transformation of concentration values results in a curve remarkably similar to that of δ 18 O benthic values (Fig. 5) Poaceae pollen percentages are up to 60 % during MIS 5 and display three percentage peaks that correspond with δ 18 O benthic and precession minima (Fig. 5a).These Poaceae percentage peaks are centred at 125, 107 and 83 ka.Pollen percentage of Acanthaceae, Aizoaceae and Tribulus are low, but consistent with changes in Poaceae pollen percentages (Fig. 4 in the Supplement).The composition of pollen spectra from warm marine substages, MIS 5e, 5c and 5a, is similar to the core-top samples (dating the last millennium).Additionally, Poaceae pollen percentages in these samples are relatively low compared to their maximum during MIS 5e (Fig. 5b and Fn-LSav biomes in SWAfr (Fig. 4).These increases of Nama-Karoo and Fn-LSav during MIS 5e, 5c and 5a likely resulted from expansions in three directions.The Nama-Karoo and Fn-LSav probably expanded to the northwest into the presentday area of the coastal Namib Desert as the intensity of BUS weakened during MIS 5 warm substages.This weakening has been documented through alkenone-based SST from marine core GeoB1711-3 (Kirst et al., 1999) (Fig. 5c), foraminifera-assemblage based SST (Chen et al., 2002) and grain-size end-member modelling (Stuut et al., 2002).Stuut and Lamy (2004) also suggested reduced atmospheric circulation and weakening of trade winds during interglacials compared to glacials, resulting in a reduction of the upwelling.A weakened wind-driven upwelling and the associated relative increase in humidity likely led to a colonization of Desert areas by Nama-Karoo or Fn-LSav (Fig. 6a).Comparable contractions of the Namib Desert linked to increased SSTs and weakening of BUS during the present interglacial are documented by Shi et al. (2000).
To the south, the Nama-Karoo and Fn-LSav likely expanded at the expense of the Succulent-Karoo and Fynbos.Increased Antarctica temperatures recorded during warm substages compared to MIS 1 (EPICA, 2006) would drive the southern westerlies polewards (Ruddiman, 2006), contributing to the ventilation of deep CO 2 -rich waters in the southern Ocean (Toggweiler and Russell, 2008).This mechanism would explain the paralleling trends observed between MIS 5 Nama-Karoo and Fn-LSav expansions in SAfr and the atmospheric CO 2 record (Petit et al., 1999;Bereiter et al., 2012) (Fig. 5).The increased Agulhas leakage documented in the Cape basin record during the last interglacial (Peeters et al., 2004) (Fig. 5c) would also be linked to a southward migration of the subtropical front and the westerlies, reducing austral-winter precipitation over SAfr.Such an atmospheric configuration would in turn favour the development of the Nama-Karoo at the expense of Succulent-Karoo and Fynbos biomes (Fig. 6a).
To the northeast, Nama-Karoo and Fn-LSav likely pushed the limit of Bd-LSav equatorward as austral-summer precipitation decreased (Fig. 5).Austral-summer Figures precipitation reductions in SAfr have been linked to reduced austral-summer insolation in the Pretoria saltpan (Partridge et al., 1997) and to reductions of grassfuelled fires during precession minima reconstructed from MD96-2098(Daniau et al., 2013) (Fig. 5b).Increased northern-hemisphere insolation during MIS5 warm substages would drive the ITCZ northwards while subtropical high pressure cells over the south Atlantic and the Indian Oceans would expand (Fig. 6a) (Ruddiman, 2006).Such changes in the tropical and subtropical pressure systems would allow the expansion of the Nama-Karoo and Fn-LSav to the northeast.
In contrast with the results presented here, previous studies report poleward interglacial expansions of savanna and are based on the increase of Poaceae pollen percentages in marine sediments along the SWAfr coast (Dupont, 2011).This interpretation could be related to two oversights.First, Dupont (2011) omits the possibility that high Poaceae percentages in SWAfr margin sediments could be an indicator of the Fn-LSav and Nama-Karoo on the adjacent continent.Secondly, Dupont (2011) does not differentiate between the broad-leaved and fine-leaved savannas, despite the significant climatic and structural differences between these two subdivisions of the biome.The Bd-LSav is influenced by fire and receives a considerable amount of precipitation during the austral summer (Scholes, 1997).The Fn-LSav is structurally and climatically more similar to the Nama-Karoo biome, as it receives very low austral-summer precipitation and does not burn (Archibald et al., 2010) despite being under a regime of significant precipitation seasonality (Scholes, 1997).Such lack of fires in the Fn-LSav and the Nama-Karoo are probably associated with low biomass and fuel load.If high Poaceae pollen percentages during MIS 5 warm substages in our record were related to expansions of the Bd-LSav and increased summer precipitation, the fire activity should also increase during these substages.Instead, an independent charcoal record from the same marine sequence MD96-2098 (Fig. 5b) documents reductions of grassfuelled fires and a decrease in austral-summer precipitation during MIS-5 precession minima (Daniau et al., 2013).Introduction

Conclusions References
Tables Figures

Back Close
Full Our results suggest that the Bd-LSav retreated equatorwards during MIS 5 precession minima.Accordingly, up to 70 % pollen percentage of Poaceae during MIS 5 compared to 35 % in the core-top samples -bearing in mind that the last millennium is characterized by precession maxima -suggested that the Nama-Karoo and Fn-LSav likely expanded during MIS 5 warm substages and covered a surface area larger than at present.Altogether, such results point to an overall increase in aridity in SWAfr during the warmest periods of the last interglacial.
During glacial isotopic stages, contractions of the Nama-Karoo and Fn-LSav would result from the following atmospheric configuration (Fig. 6b): a southward migration of the ITCZ and the associated South African monsoon (Daniau et al., 2013;Partridge et al., 1997) increasing austral-summer rainfall over SAfr.An intensification of BUS and decreased SST off the Namibian coast (Stuut and Lamy, 2004;Kirst et al., 1999) leading to aridification of coastal areas.Lastly, an equatorward migration of the westerlies increasing austral-winter precipitation and allowing a northward expansion of the winter-rain zone in SAfr (Chase and Meadows, 2007).
The pollen record from MD96-2098 also suggested glacial expansions of Fynbos (Fig. 5b) as pollen percentages of Artemisia-type, Stoebe-type, Passerina and Ericaceae were higher during MIS 6, 4 and 3 than in the core-top samples (Fig. 4 in the Supplement).These results were consistent with glacial northward expansions of Fynbos documented in other palaeoenvironmental records from SAfr (Shi et al., 2000;Dupont et al., 2007).Our record also documented a large peak in Fynbos indicators (mostly Artemisia-type) that coincided with a fast decrease in Poaceae pollen percentages at the MIS 5e/5d transition (c.117 ka), a precession and eccentricity maxima (Laskar, 1990), and an accelerated cooling in Antarctica (EPICA, 2006;Masson-Delmotte et al., 2010).As pollen percentages of Artemisia-type obtained from surface samples were associated with the Fynbos biome and austral-winter precipitation, it cannot be discarded that this resulted from a rapid and short-lived expansion of in the winter-rain zone of SAfr.Transitions MIS 6/5 and 4/3 were characterized by small but rapid increases in Podocarpus, indicating an overall increase in precipitation.Introduction

Conclusions References
Tables Figures

Back Close
Full Such increases in Podocarpus have also been documented in other records from SAfr (Dupont et al., 2011;Dupont, 2011).Finally, the amplitude of millennial-scale vegetation changes increased between ca. 100 and 25 ka and was highlighted by switches from negative to positive DCA1 scores (Fig. 5b) and increased variability of Restionaceae pollen percentages (Fig. 4 in the Supplement).This increased variability could indicate enhanced trade-wind variability or Fynbos vegetation expansions.Other Fynbos indicators did not display such trend, suggesting that Restionaceae variability between 100 and 25 ka were more likely the result of enhanced variability of southeast trade winds.Restionaceae pollen percentage data from a record 2 • N of our marine site also showed increased amplitude of millennial-scale changes (Shi et al., 2001).Grain-size wind strength tracers from the Walvis Ridge also suggested enhanced millennial-scale variability, although only after ca.80 ka (Stuut et al., 2002).Lastly, an analysis of BUS dynamics over the past 190 ka found increased millennial-scale variability of wind strength after ca. 100 ka and the windiest conditions in this zone during MIS 4 and 3 (Stuut et al., 2002).Such millennial-scale atmospheric reorganisations may also have resulted in rapid biome shifts in SWAfr.

Conclusions
The pollen spectra from SWAfr biomes were clearly differentiated based on qualitative and quantitative analyses.Pollen spectra spatial variability was closely associated with precipitation variability and seasonality.Pollen percentage iso-lines drawn for seven abundant taxa complemented the picture of pollen percentage distribution in the subcontinent.Our study showed that the most likely vegetation sources of high Poaceae pollen percentages in terrestrial surface samples from SWAfr were the Nama-Karoo and Fn-LSav.Poaceae pollen percentage distribution in marine surface sediments along the SWAfr coast supported this observation.Introduction

Conclusions References
Tables Figures

Back Close
Full The pollen record from core MD96-2098 documented expansions of the Nama-Karoo and Fn-LSav during MIS 5e, 5c and 5a warm substages.Northwestern expansions of the Nama-Karoo and Fn-LSav into Desert areas likely resulted from the reduction of BUS and a relative increase in humidity.In its eastern boundary, the Nama-Karoo and Fn-LSav likely expanded at the expense of Bd-LSav and Grasslands as a result of increased subtropical high pressure, northward shift of the ITCZ, and reduced australsummer precipitation.Nama-Karoo and Fn-LSav expansions to the southern boundary probably resulted from a southern displacement of the westerlies and the subtropical front, decreasing austral-winter precipitation.Despite the relative increase in humidity associated with a weakened BUS, our results pointed to an overall increase in aridity in SWAfr during the last-interglacial warmest periods.
During glacial isotopic stages MIS 6, 4 and 3 the Fynbos biome likely expanded as the influence of the southern westerlies and austral-winter precipitation increased.Our pollen record also suggested that transitions between isotopic stages and substages were punctuated by short increases in humidity.Increase variability of vegetation changes at millennial timescales around 100 ka was also documented and could be associated with previously identified enhanced variability of the southeastern trade winds.
Interglacial-glacial SAfr biome dynamics were linked to atmospheric and oceanic dynamics resulting from changes in global ice volume and precession at orbital timescales.Atmospheric configurations with westerly winds shifted southwards relative to today have been suggested for other interglacials (Peeters et al., 2004) and are projected for the end of 21st-century under current global warming (Beal et al., 2011).This is likely to reduce austral-winter precipitation over SAfr and favour expansions of the Nama-Karoo at the expense of the winter-rain fed Fynbos and Succulent-Karoo biomes.However, taking the current orbital configuration alone, the Nama-Karoo and Fn-LSav in SAfr might naturally remain relatively reduced for several millennial ahead.Introduction Full  2007) and modified using the savanna classification by Scholes (1997).The broad-leaved savanna distribution includes the Mopane and mixed savannas described by Scholes (1997).Location of the Orange River oceanic currents (grey arrows) and southeastern trade winds (black arrows).Temperature seasonality (b) and annual precipitation in mm yr −1 (c) extracted from the WorldClim dataset (Hijmans et al., 2005).Black dots indicate location of marine core MD96-2098 and numbers indicate the location of surface sample collection points described in Table 1 in , 1997;Mucina et al., 2007).The broad-leaved savanna distribution includes the Mopane and mixed savannas described by Scholes (1997).Iso-lines are plotted based on pollen percentage data from surface samples analysed in this study (red dots) and pollen spectra from other samples previously published and extracted from the African Pollen Database (black dots) (Gajewski et al., 2002)., 1997;Mucina et al., 2007).The broad-leaved savanna distribution includes the Mopane and mixed savannas described by Scholes (1997).Iso-lines are plotted based on pollen percentage data from surface samples analysed in this study (red dots) and pollen spectra from other samples previously published and extracted from the African Pollen Database (black dots) (Gajewski et al., 2002).Poaceae pollen percentage in the marine domain are redrawn from Dupont and Wyputta (2003) and extended beyond latitude 25 • S using two MD96-2098 core-top samples (hatched).Introduction

Conclusions References
Tables Figures

Back Close
Full  (Laskar et al., 2004).(b) Stable Oxygen profile of benthic foraminifera Cibicidoides wuellerstorfi (Bertrand et al., 2002), log-transformed total pollen concentration, detrended correspondence analysis Axis1 scores, pollen percentages of indicator taxa for Nama-Karoo and fine-leaved savanna (Poaceae) and Fynbos (Artemisia-type, Ericaceae, Passerina, Protea, and Stoebe-type) and charcoal concentrations in number of particles per gram (nb g −1 ) all from marine core MD96-2098 and plotted against age in ka (thousands of calibrated/calendar years before present).
Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | As tropical low pressure systems migrate northwards during the austral winter, subtropical high pressure significantly reduces austral-summer precipitation in SAfr.The southwestern part of the continent is influenced by the seasonal migration of the subtropical front and Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | (South Africa) to L üderitz (Namibia) during two weeks of field work in February 2011.The sampling area extended from latitudes 26.5 to 34.5 • S and from longitudes 15 to 23.8 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | surface samples collected within the Fynbos biome, Anacardiaceae, Anthospermum, Artemisia-type, Ericaceae, Passerina, Protea and Stoebe-type reach highest percentages (Fig. 2 in the Supplement).These pollen abundances reflect well the composition of the vegetation source described by Cowling et al. (1997a), as species of Ericaceae, Passerina and Protea are particularly characteristic of the Fynbos biome.Up to 10 % Podocarpus pollen percentages found in the Fynbos likely originate from small forest patches within the area.Poaceae pollen percentages are below 20 % in the Fynbos and in the coastal forest sample.Ilex pollen is found in the sample from the coastal forest at approximately 5 % and Podocarpus pollen is up to 18 %, consistent with their abundance in the vegetation source.Except for sample Fy22, the Introduction Discussion Paper | Discussion Paper | Discussion Paper | For instance, Poaceae, Crassulaceae and Tribulus obtain the highest loadings to classify the pollen spectra from the Nama-Karoo.Asteraceaeother, Chenopodiaceae-Amaranthaceae and Euphorbia obtain high loadings to cluster Suculent-Karoo and Desert samples.Pollen taxa that characterize the Fynbos pollen spectra include Anacardiaceae, Artemisia-type, Ericaceae, Myrica, Passerina, Protea, Stoebe-type.Podocarpus characterizes the pollen spectra from coastal forests and Fynbos biomes.Figures Back Close Full Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | . Relative increases in pollen concentration might indicate an increase of pollen arrival to the site during low sea-level stands when the vegetation source was closest (i.e. during glacial stages).However, this is unlikely because of the width of the Walvis continental Shelf (i.e.steep slope in a few kilometres).An increase of pollen concentration might indicate instead an increase in aeolian and riverine pollen input during glacials, and/or an increase in pollen preservation linked to upwelling enhancement as suggested by Pichevin et al. (2005).Nonetheless, these glacial-interglacial pollen concentration variations have no effect on the pollen record because vegetation changes are reconstructed based on standardized pollen abundances.The axis scores on DCA1 reveal changes in the composition of pollen assemblages that also resemble variations in the δ 18 O benthic record.DCA1 axis scores from MIS 5 and 3 are overall positive in value, while scores from MIS 6 and 4 are negative.A series of large-magnitude changes in DCA1 axis scores (i.e. when adjacent sample scores switch from positive to negative values) are also visible and increase in amplitude after 4340 Discussion Paper | Discussion Paper | Discussion Paper | c. 100 ka.Such changes in DCA Axis1 scores are also observed during MIS 6 but are of lesser magnitude.Changes in DCA axis scores suggest significant changes in vegetation composition from one time step to the next (i.e.adjacent samples).
).During MIS 6 and 4, Poaceae percentages are reduced and co-vary with δ 18 O benthic values.Pollen percentages of Chenopodiaceae-Amaranthaceae and Asteraceae-other are relatively high and increased along with enriched δ 18 O benthic values during MIS 6 and at the end of MIS 4 (Fig. 4 in the Supplement).Cyperaceae pollen percentages vary importantly throughout the record and are as high as 40 % during MIS 4. Fynbos indicators including Artemisa-type and Stoebe-type show relative increases in pollen percentage during MIS 6, 4 and 3 (Fig. 5b).Ericaceae, Passerina, Anthospermum, Cliffortia, and Protea pollen percentages are low but consistent with abundance changes of Artemisiatype and Stoebe-type (Fig. 4 in the Supplement).Pollen percentages of Restionaceae increase after the 105-ka δ 18 O benthic minimum and remain abundant during the rest of MIS 5 through MIS 3, despite a relative decrease during MIS 4. Podocarpus percentages are lower than 10 % but show increases at stage boundaries around 135 ka (MIS 6/5), 100 ka (5c/5b), 75 ka (MIS 5a/4), 60 ka (MIS 4/3), and around 27 ka (MIS 3/2) (Fig. 4 in the Supplement).Based on the observed present-day distribution of Poaceae pollen percentages from marine and terrestrial surface samples, we infer that increases of Poaceae pollen percentages during MIS 5 warm substages correspond to expansions of the Nama-Karoo Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Stuiver, M. and Reimer, P. J.: Extended 14C database and revised CALIB radiocarbon calibration program, Radiocarbon, 35, 215-230, 1993.Stuut, J.-B.W. and Lamy, F.: Climate variability at the southern boundaries of the Namib (southwestern Africa) and Atacama (northern Chile) coastal deserts during the last 120,000 yr, Quaternary Res., 62, 301-309, doi:10.1016/j.yqres.2004.08.001, 2004Discussion Paper | Discussion Paper | Discussion Paper | Fig. 1.(a) Map of biomes of southern Africa based on Mucina et al. (

Fig. 3 .
Fig.3.Bioclimatic variables and pollen percentage iso-lines drawn over biome units of southern Africa (modified fromScholes, 1997;Mucina et al., 2007).The broad-leaved savanna distribution includes the Mopane and mixed savannas described byScholes (1997).Iso-lines are plotted based on pollen percentage data from surface samples analysed in this study (red dots) and pollen spectra from other samples previously published and extracted from the African Pollen Database (black dots)(Gajewski et al., 2002).Numbers and lines represent pollen percentages and are shown for taxa discussed in the text: Asteraceae-other, Chenopodiaceae-Amaranthaceae, Podocarpus, Restionaceae, Artemisia-type, and Stoebe-type.

Fig. 4 .
Fig.3.Bioclimatic variables and pollen percentage iso-lines drawn over biome units of southern Africa (modified fromScholes, 1997;Mucina et al., 2007).The broad-leaved savanna distribution includes the Mopane and mixed savannas described byScholes (1997).Iso-lines are plotted based on pollen percentage data from surface samples analysed in this study (red dots) and pollen spectra from other samples previously published and extracted from the African Pollen Database (black dots)(Gajewski et al., 2002).Numbers and lines represent pollen percentages and are shown for taxa discussed in the text: Asteraceae-other, Chenopodiaceae-Amaranthaceae, Podocarpus, Restionaceae, Artemisia-type, and Stoebe-type.
Discussion Paper | Discussion Paper | Discussion Paper | Fig. 5. Comparison of terrestrial, atmospheric and oceanic markers for southern Africa between 190 and 24.7 ka.(a) Orbital parameters plotted for latitude 25 • 36 S using La2004 Stars on the left correspond to percentage of pollen taxa in two top-core samples dating 530 and 1060 calibrated years before present.(c) Independent climatic records discussed in the text.Stage boundary ages for 3/2, 4/3, and 5/4 from Sanchez Go ñi and Harrison (2010) and 6/5 from Henderson and Slowey (2000).Discussion Paper | Discussion Paper | Discussion Paper | Fig. 6.Schematic and simplified configuration of vegetation, atmospheric, and oceanic systems over southern Africa during (a) the MIS 5 warm substages, and (b) glacial isotopic stages.Rainfall is illustrated as the current configuration of tropical and subtropical convection systems using average austral-summer (a) and austral-winter (b) precipitation data between 1979-1995 from the International Research Institute for Climate Prediction (http://iri.ldeo.columbia.edu).L: tropical low-pressure systems, H: subtropical high-pressure cells, SET: southeast trade winds, SW: southern westerlies, AL: Agulhas leakage.Stars indicate the location of marine records discussed in the text and blue arrows indicate direction of migration of pressure systems.