Vegetation development

in West Africa

of the biosphere shift

during late Miocene

to (early) Pliocene

Dissertation for the

Doctoral Degree in Natural Sciences

(Dr. rer. nat.)

in the Faculty of

Geosciences

at the

University of Bremen

submitted by

Sebastian Hötzel

September 2013

-Erklärung/Declaration-

Name: Sebastian Hötzel

Anschrift: Elsasser Straße 24, 28211 Bremen

Hiermit erkläre ich, dass ich

1. die Arbeit ohne unerlaubte fremde Hilfe angefertigt habe,

2. keine anderen als die von mir angegebenen Quellen und Hilfsmittel benutzt habe und

3. die den benutzten Werken wörtlich oder inhaltlich entnommenen Stellen als solche kenntlich

gemacht habe.

Bremen, den 26. September

_______________________

II

-Gutachter-

Prof. Dr. Gerold Wefer

Prof. Dr. Heiko Pälike

III

The world is the geologist's great puzzle-box; he stands before it like the child to whom the

separate pieces of his puzzle remain a mystery till he detects their relation and sees where they

fit, and then his fragments grow at once into a connected picture beneath his hand.

Louis Agassiz

IV

Table of Contents

ABSTRACT ....................................................................................................................VII

ZUSAMMENFASSUNG................................................................................................. IX

CHAPTER I: INTRODUCTION ..................................................................................... 1

1.1 Ecosystem savannah...................................................................................................................................1 1.2 C3 and C4 grasses ......................................................................................................................................2 1.3 Evolution of grasses and grasslands ............................................................................................................3 1.4 Motivation and objectives............................................................................................................................4 1.5 Present-day conditions ................................................................................................................................5

1.5.1 Climate and vegetation of south-west Africa........................................................................................5 1.5.2 Oceanic conditions of the south-east Atlantic.......................................................................................6

1.6 Climate during the late Miocene to Pliocene ................................................................................................8 1.7 Material and methods ................................................................................................................................9

1.7.1 Palynological preparation....................................................................................................................9 1.7.2 Terrestrial palynomorphs..................................................................................................................10 1.7.3 Marine palynomorphs ......................................................................................................................10

1.8 Thesis outline ...........................................................................................................................................10 1.9 References.................................................................................................................................................11

CHAPTER II: THE ROLE OF FIRE IN MIOCENE TO PLIOCENE C4

GRASSLAND AND ECOSYSTEM EVOLUTION ........................................................ 17

2.1 Abstract ..................................................................................................................................................17 2.2 Main Text...............................................................................................................................................18 2.3 Methods summary....................................................................................................................................23 2.4 References.................................................................................................................................................23 2.5 Acknowledgments ....................................................................................................................................25 2.6 Author Contributions ..............................................................................................................................25 2.7 Supplementary Information ......................................................................................................................26

2.7.1 Material ..........................................................................................................................................26 2.7.2 Charcoal and charred particles..........................................................................................................26 2.7.3 Interpretation of the pollen record......................................................................................................26 2.7.4 Indications for aridification...............................................................................................................28 2.7.5 δ13C of plant-wax alkanes in C3 and C4 plants...............................................................................29 2.7.6 Seasonality, aridity, grasses and fire..................................................................................................30 2.7.7 References.........................................................................................................................................30

CHAPTER III: MIOCENE-PLIOCENE VEGETATION CHANGE IN SOUTH-

WESTERN AFRICA (ODP SITE 1081, OFFSHORE NAMIBIA).................................33

3.1 Abstract ..................................................................................................................................................33 3.2 Introduction .............................................................................................................................................34 3.3 Modern regional climate and south-western African vegetation...................................................................35 3.4 Methods and material ..............................................................................................................................36 3.5 Results.....................................................................................................................................................37

3.5.1 Pollen percentages and pollen flux rates.............................................................................................37

V

3.5.2 Endmember results ..........................................................................................................................38 3.6 Discussion ...............................................................................................................................................41

3.6.1 Transport and source of pollen and spores and the desiccation of Lake Cunene..................................41 3.6.2 Vegetation development....................................................................................................................42 3.6.3 Steps in aridification ........................................................................................................................43

3.7 Conclusions..............................................................................................................................................44 3.8 Acknowledgements...................................................................................................................................45 3.9 References ................................................................................................................................................45

CHAPTER IV: MIOCENE-PLIOCENE STEPWISE INTENSIFICATION OF THE

BENGUELA UPWELLING OVER THE WALVIS RIDGE OFF NAMIBIA.............. 51

4.1 Abstract ..................................................................................................................................................51 4.2 Introduction .............................................................................................................................................52

4.2.1 The Benguela Upwelling System.......................................................................................................52 4.2.2 The Angola-Benguela Front ............................................................................................................53

4.3 Material and methods ..............................................................................................................................54 4.4 Results.....................................................................................................................................................55 4.5 Discussion ...............................................................................................................................................58

4.5.1 Dinoflagellates as proxies for environmental conditions .....................................................................58 4.5.2 The Angola-Benguela Front ............................................................................................................58 4.5.3 Upwelling intensification and meridional pressure gradient................................................................59 4.5.4 Exceptional conditions related to the Mediterranean Salinity Crisis..................................................60 4.5.5 Upwelling intensification and river supply.........................................................................................63

4.6 Summary and conclusions ........................................................................................................................64 4.7 Acknowledgements...................................................................................................................................64 4.8 References ................................................................................................................................................64

CHAPTER V: SYNTHESIS AND OUTLOOK.............................................................. 73

5.1 Aridification favoured the savannah expansion.........................................................................................73 5.2 Fire and its role in the savannah expansion..............................................................................................73 5.3 Aridification and its link to the Benguela Upwelling ................................................................................74 5.4 Outlook...................................................................................................................................................75 5.5 References ................................................................................................................................................75

ACKNOWLEDGEMENTS............................................................................................. 77

VI

Abstract

Aridification is a major aspect of the climatic trend in Africa during the late Neogene. Main

causes of the aridification are changes in the atmospheric circulation as part of the steepening

meridional pressure gradient. The trend is affecting biomes largely by shifting, shrinking and

expanding vegetation belts. During the Miocene to Pliocene especially the tropical to subtropical

savannah grasslands (C4 grasslands) expanded. Triggers of the C4 grassland expansion are not

well understood and several mechanisms have been proposed explaining it such as decreasing

CO2 levels, aridification, increased seasonality and fire.

The thesis focuses on pollen, spores and microscopic charcoal records from a sediment core

retrieved at Ocean Drilling Program Site 1081 offshore Namibia. The aim is to investigate the

Miocene to Pliocene vegetation change of south-west Africa with emphasis on the expansion of

the savannah grassland and the mechanisms behind it. Additionally, dinoflagellate cysts from the

same sediments are analysed to study the link between continental and oceanic conditions.

In the first part (Chapter II) the focus is on the savannah grassland expansion. It was thought

that decreasing CO2 concentrations favoured globally the expansion of grasses during the Mio-

/Pliocene. However, it was shown that the CO2 concentrations of the Miocene to Pliocene were

not changing significantly. Moreover, the expansion of grasslands occurred asynchronously on

different continents enhancing the plausibility of regional to local drivers on each continent over

one global trigger. The general competitiveness of grasses at higher temperatures and their

adaptations to aridity and the occurrence of frequent fires support the hypothesis of aridification

and fire as a driver of savannah expansion. The pollen and microscopic charcoal record in

combination with carbon stable isotopes of plant waxes revealed that the savannah in south-

west Africa expanded during an aridification period starting at around 8.3 Ma. Furthermore an

extreme fire regime from 7.1 until 5.8 Ma favoured the C4 grasses, which nowadays dominate

tropical grasslands. As aridification continued shrubland and desert vegetation expanded and the

grasslands became sparser during the Pliocene. In the remaining savannahs C4 grasses increased

and dominated the C3 grasses.

In the second part (Chapter III) the general vegetation change in south-west Africa is studied.

Using an unmixing model we show that the pollen and spore record revealed three steps of

vegetation changes, which we interpret as the result of a continuous aridification process. A

relative wet phase with higher representation of Cyperaceae, mountain and woodland vegetation

is exchanged by a dry one with strongly increased abundances of desert and semi-desert

VII

indicators via a transition phase in which mainly grasses expanded. The results indicate ongoing

aridification which is probably caused by the steepening meridional pressure gradient. In the last

phase pollen from aquatic vegetation started to appear around 5 Ma, which we link to a

relocation of the lower course of the Cunene River to its modern outlet in the Atlantic Ocean.

Redirection of the Cunene River toward the Atlantic would have deprived the palaeolake

Cunene of an important source of fresh-water ultimately resulting in the desiccation of the lake

and the formation of the Etosha Pan.

The last part of the study (Chapter IV) focuses on the dinoflagellate cysts and the oceanic

conditions along the Namibian coast. Changes in the assemblages of dinoflagellate cysts suggest

a stepwise intensification of the Benguela Upwelling. During the late Miocene a higher influence

of the southward flowing warmer Angola Current associated with a more southwards located

Angola-Benguela front is likely caused by a weaker meridional pressure gradient. During the

latest Miocene to earliest Pliocene the meridional pressure gradient steepened and the upwelling

intensified. An intermediate period from 6.2 to 5.5 Ma is shown by the dominance of

Habibacysta tectata, cysts of a cool-tolerant dinoflagellate known from the northern Atlantic,

indicating non-reoccurring oceanic conditions, i.e. a change in the quality of the upwelled

waters, which occurred during the Messinian salinity crisis. From 4.4 Ma on a stronger upwelling

with well mixed nutrient waters is indicated.

VIII

Zusammenfassung

Die Ausbreitung von Aridität ist der übergeordnete klimatische Trend im späten Neogen.

Dieser wurde verursacht durch Änderungen der atmosphärischen Zirkulation, die im

Zusammenhang mit dem sich intensivierenden meridionalen Luftdruckgradienten stehen. Diese

Entwicklung beeinflusst auch Biome, die gezwungen werden sich zu verschieben, sich

zusammenzuziehen oder sich auszubreiten. Während des Miozäns und Pliozäns expandierten

besonders die tropischen und subtropischen C4 Grassavannen. Mehrere Mechanismen wurden

vorgeschlagen um diese Ausbreitung zu erklären. Beispiele hierfür sind sinkende atmosphärische

CO2 Konzentrationen, Aridität, intensive Saisonalität und Feuer.

Ziel der vorliegenden Arbeit ist die Rekonstruktion der miozänen-pliozänen

Vegetationsveränderung in Südwest-Afrikas. Um diese zu erstellen, konzentriert sich die

Untersuchung auf Pollen, Sporen und mikroskopische Holzkohle aus Sedimenten eines

Bohrkerns des Ocean Drilling Program Site 1081 vor Namibia. Im Besonderen soll die

Ausbreitung der Savanne und die dahinter stehenden Verursacher festgestellt werden. Ferner

wurden Zysten von Dinoflagellaten analysiert, die ebenfalls in den Sedimenten enthalten waren,

um die Wechselwirkungen von ozeanischen und kontinentalen Bedingungen zu untersuchen.

Der erste Abschnitt (Kapitel II) behandelt die Ausbreitung des Savannen-Graslandes. Es wurde

angenommen, dass sinkende CO2 Konzentrationen global die Ausbreitung der

Graslandschaften während des Miozäns und Pliozäns antrieben. Jedoch stellte sich heraus, dass

sich die Konzentrationen während dieser Zeit nicht ausschlaggebend veränderten. Zusätzlich

breiteten sich die Savannen auf den Kontinenten nicht zeitgleich aus, so dass regionale bis lokale

Gründe für die Ausbreitungen wahrscheinlicher sind als globale. Wegen der guten Anpassung

von Gräsern an hohe Temperaturen, Feuer und Trockenheit werden als Hypothesen zur

Graslandausbreitung besonders die Austrocknung und Feuer als fundamentale Treiber genannt.

Die untersuchten Pollen und mikroskopischen Holzkohlenfragmente wurden mit einem stabilen

Kohlenstoff Datensatz von Pflanzenwachsen aus denselben Sedimenten verglichen. Die

Ergebnisse zeigen, dass die Savanne ab 8,3 Ma begann sich auszubreiten. Außerdem bevorzugte

ein extremes Feuerklima zwischen 7,1 und 5,8 Ma speziell die Ausbreitung von C4 Gräsern, die

heute in tropisch bis subtropischen Graslandschaften dominieren. Unter anhaltender

Intensivierung der Aridität breitete sich daraufhin Buschvegetation und Wüstenvegetation aus,

wohingegen die Graslandschaft sich ausdünnte. Hier zogen sich offensichtlich besonders C3

Gräser zurück, so dass der relative Anteil von C4 Gräsern in der Savannengraslandschaft

zunahm.

IX

X

Im zweiten Teil (Kapitel III) wird der generelle Vegetationswechsel im Mio-/Pliozän in

Südwest-Afrika untersucht. Es wurde ein „unmixing“ Modell, unter Verwendung der Pollen und

Sporen Daten angewendet. Dieses zeigte, dass eine kontinuierliche Aridifikation stattfand, die

durch drei Stufen repräsentiert wird. Eine relativ feuchte Phase mit hohen Anteilen von

Cyperaceae, Berg- und Waldvegetation wird durch eine trockene Vegetation, mit vermehrtem

Auftreten von Wüsten und Halbwüsten Indikatoren, ausgetauscht. Dazwischen liegt eine

Übergangsphase, in der sich hauptsächlich die Savannengräser ausbreiteten. Die Ergebnisse

zeigen eine anhaltende Aridifikation, die wahrscheinlich durch eine Intensivierung des

meridionalen Luftdruckgradienten verursacht wurde. Darüber hinaus treten Pollen von

aquatischer Vegetation ab 5 Ma auf, die eine Verlaufsänderung des Nieder-Kunene zur

modernen Mündung in den Atlantik andeuten. Diese Verlaufsänderung verursachte den Verlust

eines wichtigen Zulaufes von Wasser in den Kunene Paläosee. Dieser trocknete daraufhin aus

und entwickelte sich zu der heutigen Etosha-Pfanne.

Der dritte Teil der Arbeit (Kapitel IV) behandelt die Dinoflagellatenzysten und die

rekonstruierten ozeanischen Bedingungen entlang der namibischen Küste. Veränderungen der

Dinoflagellaten-Zysten-Vergesellschaftungen deuten auf eine stufenweise Intensivierung des

Benguela Auftriebgebietes hin. Während des späten Miozäns ist ein stärkerer Einfluss des

Angola-Stroms wahrscheinlich, der durch eine südlichere Position der Angola-Benguela Front

verursacht wurde. Diese südlichere Position ist vermutlich die Konsequenz eins schwächeren

meridionalen Luftdruckgradientes. Während des spätesten Miozän und frühen Pliozän

intensivierte sich dieser Gradient jedoch, so dass auch der Auftrieb zunahm. Eine

Übergangsperiode, die dominiert ist von Habibacysta tectata, der Zyste einer kaltwasser-toleranten

Dinoflagellate, auch bekannt aus dem Nordatlantik, deutet auf veränderte ozeanische

Bedingungen hin. Diese beeinflusste die Qualität des aufgetriebenen Wassers. Die veränderten

Bedingungen waren zeitgleich mit der Messinischen Salinitätskrise. Von 4,4 Ma an deuten die

Zysten-Vergesellschaftungen einen stärkeren Auftrieb und gut durchmischte nährstoffreiche

Oberflächenwässer an.

Chapter I

Chapter I: Introduction

The climate of the earth changed in its

history several times from a greenhouse to

an icehouse climate and vice versa. The last

50 million years (Ma) were affected by

general global cooling, from a greenhouse

climate of the Eocene times to the

Pleistocene glacial/interglacial climate

(Zachos et al., 2001, 2008). The main driver

during the cooling was the growth of ice

shields on Antarctica and later on the

northern Hemisphere. But besides that

several further mechanisms are important

for the cooling of the climate also

influencing the growth of ice sheets, e.g.

deep ocean circulation, global carbon

cycling, changing oceanic currents and

tectonic uplifts (e.g. Flower & Kennett,

1994; Zachos et al., 2001; Prange & Schulz,

2004). During the late Neogene, from 20 to

2 Ma, the East Antarctic Ice shield started to

grow again after a warmer period, the mid-

Miocene climatic optimum (Flower &

Kennett, 1994). The growing ice sheets

affected the climate by changing

atmospheric circulations (e.g. Hadley cell

circulation) and its intensities by increasing

the meridional thermal gradients (Flower &

Kennett, 1994). In general, wind systems,

especially the trade winds become more

intense affecting precipitation patterns

towards aridification especially in the

subtropics. Hence, this climate change had

an intense impact on the biosphere shifting

biomes (Pound et al., 2012). The

aridification forced complete biomes to

retreat or allowing them to expand and to

evolve. During the mid-Miocene humid and

warm conditions prevailed over subtropics

and even in higher latitudes such as central

Europe (e.g. Herold et al., 2011). From the

late Miocene to Pliocene, deserts such as the

Atacama Desert and the Namib Desert

established (van Zinderen Bakker, 1975; Van

Zinderen Bakker, 1984; Houston & Hartley,

2003). Furthermore, it is suggested that the

general drier climate of the late Neogene

favoured grasslands which expanded during

the late Miocene to Pliocene especially in

tropical and subtropical areas (Retallack,

2001; Edwards et al., 2010; Strömberg,

2011). Also in other parts of Africa the

aridification took place and introduced the

present biomes (Senut et al., 2009). Beside

the Namib Desert and the Sahara Desert the

aridification might have caused “the most

dramatic example of biome assembly in the

geological record”, the rise of C4 grassland

(Edwards et al., 2010).

1.1 Ecosystem savannah

Today grasses (Poaceae) inhabits all major

landmasses and they dominate biomes

covering in total up to 40 % of the earth’s

surface (Strömberg, 2011) (Figure 1.1).

1

Introduction

Considering Africa, the savannahs are the

most dominant biome and inhabit tropical

to subtropical areas. Within these biomes

they form the natural habitat of specialized

grazing and browsing animals, especially

megaherbivores like Elephants, buffalos or

giraffes (Uno et al., 2011).

Figure 1.1: World map with vegetation zones dominated by grasses (Rommerskirchen et al., 2006). Grey one indicating tropical grasslands which are mainly dominated by C4 grasses. Black areas indicate temperate grasslands with mainly C3 grasses.

Additionally, these herbivores serve as food

source for predators which hide well in high

grasses. The grassland savannahs especially

of Africa are therefore a valuable ecosystem.

Furthermore, grasslands are a unique

environment closely linked to fires. Grasses

are adapted to fires as a result of their fast

re-growth after such disturbance. Fires

control the tree cover and hence the

boundary between grasslands and

woodlands or forests by re-occurring

annually or biannually (Bond et al., 2003;

Bond, 2008; Furley et al., 2008). Therefore

fires enable grasslands to prevail in areas

were it is usually wet enough for trees.

However, grasslands have also some

feedbacks to the fire regime, e.g. by

accumulating biomass which when dried

provide as excellent fuel (Beerling &

Osborne, 2006). This again creates a positive

feedback leading to higher tree mortality.

1.2 C3 and C4 grasses

Principally, there are two different types of

photosynthesis, C3 and C4. The C4 pathway

occurs in 18 different plant families

including many subfamilies of the Poaceae.

However, most C4 plants are grasses (Sage et

al., 1999a). In both physiological pathways

the plants use the same enzyme (Rubisco)

and the same biochemical cycle for carbon

fixation (Calvin-Benson Cycle). However,

the C4 way has several carbon-acid products

before whereof the first product has four

carbons. As an anatomical feature in the

leaves almost every C4 plant has a kranz

anatomy which acts as a CO2 pump

increasing the efficiency under lower

atmospheric CO2 conditions and/or higher

temperatures (Ehleringer et al., 1997). This

pump allows the plant under temperature

stress to maintain energy winning while

2

Chapter I

doing photosynthesis and not switching to

photorespiration as C3 plants do (Tipple &

Pagani, 2007). A further advantage is a

higher water-use-efficiency as compared to

C3 plants under higher temperatures. C4

grasses need only half as much water per

unit carbon assimilation as C3 grasses under

25°C (Hatch, 1987). In summary C4 plants

have advantages under higher temperature,

aridity, high light conditions and lowered

CO2 concentrations (Sage et al., 1999b).

As a by-product of the CO2 accumulation

within the C4 photosynthesis the plant takes

up a higher amount of heavier C isotopes,

e.g. 13C as other grasses or even other plants

like most trees and has therefore a special

isotopic value (O’Leary, 1988).

1.3 Evolution of grasses and grasslands

The oldest known Poaceae fossils are

phytoliths which are silica bodies on the

leaves of many grasses. These silica bodies

were found in sediments in India with an

age around 66 Ma (latest Cretaceous)

(Prasad et al., 2011). Based on these findings

a phylogenetic analysis suggests an origin of

the Poaceae between 129 and 107 Ma

(Prasad et al., 2011). So that all lineages of

Poaceae were already present before

Gondwana broke apart explaining a today

world-wide distribution of all sub-families

(Prasad et al., 2005). The oldest isotopic

evidence for C4 photosynthesis in grasses

was found in a special analysis of isolated

grass pollen from the earliest Oligocene

(Urban et al., 2010). True fossil evidence

showing the typical C4 kranz anatomy is

described from Miocene sediments

(Thomasson et al., 1986). However, the

grasses expanded not before the late

Miocene, and it seemed that they established

between the late Miocene to Pliocene on all

continents (Cerling et al., 1997; Jacobs et al.,

1999). Decreasing atmospheric CO2

concentrations were thought to have

favoured the C4 grasses and enabled the rise

of the grasslands by crossing a CO2-

threshold (500 parts per million by volume

(ppmv) at high growing-season

temperatures) (Cerling et al., 1997).

Reconstructions of palaeo-CO2 levels,

however, revealed that the threshold at 500

ppmv was already crossed in the Oligocene

(Figure 1.2) and might be linked to the

origin of C4 photosynthesis in general

(Pagani et al., 2005; Zachos et al., 2008).

Furthermore, these reconstructions showed

that the CO2 level was from the Middle

Miocene on always at comparable levels as

before the industrial revolution. Instead of

decreasing CO2 concentrations the level

even increased slightly at the end of the

Miocene which would have been a

disadvantage for plants using C4

photosynthesis. Additionally, an increasing

amount of studies about the rise of

grasslands revealed that the expansions on

the different continents and in different

regions were not synchronous (Edwards et

al., 2010; Strömberg, 2011). Hence, the

direct driver for the expansion could not

have been a global mechanism and the

trigger must be sought on regional to local

scales.

3

Introduction

Figure 1.2: Reconstruction of atmospheric CO2 concentration since the Eocene (after Zachos et al., 2008) showing that the atmospheric CO2 threshold (500 ppmv) was crossed at the end of the Oligocene, long before the Savannah grassland expansion of the late Miocene. Additionally, it shows a small increase of CO2 concentrations at the end of the late Miocene.

1.4 Motivation and objectives

Since it became clear that the reasons for the

savannah grassland expansions are not

linked to CO2 the drivers must be of a more

regional to local origin. However, the

general cooling and drying of the globe

within the Neogene aridification suited the

grasses. Aridity is thought to be the

immediate driver for the rise of grasses

including a development of well pronounced

seasons (Jacobs et al., 1999; Osborne, 2008;

Strömberg, 2011; Scheiter et al., 2012). An

enhanced seasonality enables during wet

summers the build-up of biomass (e.g.

grasses) which can act as fuel for possible

fires whereas the flammability of this fuel

increases during dry winters (Daniau et al.,

2013). This climate suits grasses, especially,

because they tolerate a long drying phase

and can re-grow quickly when rains return.

Additionally, since the present day boundary

between woodlands and grasslands is

defined by fires in many parts of the globe,

fire is also thought to be a substantial driver

for the expansion of grasslands fitting in the

general idea of an increased seasonality

(Jacobs et al., 1999; Keeley & Rundel, 2005;

Beerling & Osborne, 2006; Tipple & Pagani,

2007; Osborne, 2008; Edwards et al., 2010;

Scheiter et al., 2012). However, data

showing the link between savannah

grassland expansion and fire were not

reported so far. To test these hypotheses

this study focuses on vegetation changes in

south-west Africa based on palynological

investigations. The savannah grassland

expansion of south-west Africa has not been

investigated in detail, so that the present

work will close a gap of vegetation history of

Africa. In detail, the thesis focuses on three

hypotheses:

I. Aridification took place during the late

Miocene to Pliocene in south-west Africa

favouring the savannah grassland expansion

II. Furthermore, fire was involved in the grass

expansion

III. Regional aridification is linked to the

development of the Benguela Upwelling

4

Chapter I

1.5 Present-day conditions

1.5.1 Climate and vegetation of south-

west Africa

In Africa the climate range is wide, from

tropical rainforest in the heart of the

continent to the deserts and arid vegetation

of e.g. the Sahara Desert and the south

western parts of the continent (Figure 1.3).

Figure 1.3: Modern vegetation cover of Africa representing climatic zones. Green colours represent forests, red to orange colours represent woodlands and shrublands, grasslands are in yellow colours, purple are agricultural used areas and in white are deserts (Mayaux et al., 2004)

The distribution of those climate zones are

linked to atmospheric circulations and

oceanic interactions. For south-west Africa

the South Atlantic Anticyclone (high

pressure zone) is the most prominent

feature influencing the entire sub-continent.

The high is located at 27°S 10°W during

austral winters whereas it moves south to

32°S 5°W during austral summers

(Petterson & Stramma, 1991) (Figure 1.4).

During the year a tropical low, Inter

Tropical Convergence Zone (ITCZ) moves

north and south. Towards its southern

boundary it forms the Congo-Air Boundary

(CAB) (Figure 1.4 and 1.5). This boundary

divides the sources of air masses. To the

west and north of the CAB the air masses

derive mostly from the Atlantic whereas the

air masses to the east are mainly derived

from the Indian Ocean because the strong

trade winds hinder the Atlantic air masses

to penetrate further east (Gasse et al.,

2008). Similar patterns are represented in

the source of precipitation (Gimeno et al.,

2010) (for precipitation over southern

Africa see Figure 1.5). East of the CAB the

5

Introduction

source of precipitation is the Indian Ocean

and to the north and west the source is the

Atlantic. The strong south-east trade winds,

however, have a drying effect, so that the

coast receives little rain.

Figure 1. 4: Atmospheric circulations over Africa in austral winter and austral summer (Nicholson, 1996; Gasse et al., 2008). Note position shifts of the Inter Tropical Convergence Zone (ITCZ), the Congo-Air boundary (CAB).

Closely linked to the amount of rainfall is

the distribution of vegetation. Towards

northern parts of southern Africa, e.g.

Angola, the precipitation is with over 800

mm per year high allowing woodlands and

forest to grow (White, 1983; Giess, 1998).

In more southern parts the precipitation

decreases from the east to the west. Eastern

parts of south-west Africa receive 250 to

500 mm per year. These areas are parts of

the Kalahari savannah, often also called

Kalahari Desert. Here the grass cover can

be relatively dense, especially after rain. Up

to 40 % of the area can be covered by a

closed canopy (White, 1983). In wide areas

of south-west Africa, e.g. Namibia the C4

contribution of the grassland is between 90

and 100 % (Sage et al., 1999b; Still &

Powell, 2010). With westwards decreasing

amounts of rain, the vegetation is getting

sparser and more desert-like. In areas

receiving between 100 and 250 mm per year

the Karoo shrubland prevails (White, 1983).

This vegetation is dominated by shrubs and

bushy plants, e.g. Compositae (or

Asteraceae), Fabaceae or Acanthaceae

(White, 1983; Cowling et al., 1998; Giess,

1998). Along the coast the aridity is most

intense forming the Namib Desert. In the

Namib the annual precipitation is in most

parts less than 100 mm per year and is

often so low that germination is impossible

for most plants (Southgate et al., 1996) so

that in some parts of the Namib vegetation

is completely absent (White, 1983).

1.5.2 Oceanic conditions of the south-

east Atlantic

The South Atlantic Anticyclone is

influencing the surface currents in the

south-east Atlantic by directing the

Benguela Current northwards along the

south-west African coast (Figure 1.5). The

Benguela Current is fed by the cold

6

Chapter I

eastward flowing Antarctic Circumpolar

current and the retroflection of the Aguhlas

Current from the Indian Ocean. At 28°S

the Benguela current splits into two

currents, the northwards flowing Benguela

Coastal Current (BCC) and the

continuation of the Benguela Current (BC)

which follows the atmospheric circulation

of the SAA and turns westward towards

South America (Figure 1.5).

Figure 1. 5: Map of location of ODP site 1081A (grey dot) in the south east Atlantic. The mean annual precipitation of southern Africa (Nicholson, 2000) with the austral summer (1020 mbar in red) and austral winter (1024 mbar in yellow) position of the South Atlantic Anticyclone (SAA) (Petterson & Stramma, 1991) and the Congo-Air-Boundary (Leroux, 1983) with the average summer position in red and the average winter position in yellow. Major Ocean features are Antarctic Circumpolar Current (ACC), Agulhas Current (AgC), Benguela Coastal Current (BCC), Benguela Current (BC), Angola Current (AC), the Angola Benguela Front (ABF) (Petterson & Stramma, 1991). Bathymetry is indicated by fine grey lines.

The BCC is furthermore linked to the trade

winds which via Ekman-transport induce

the Benguela Upwelling by pushing the

surface waters offshore so that subsurface

waters wells up (Lutjeharms & Meeuwis,

1987). These waters are derived from the

Antarctic Intermediate Water which is cold

and nutrient-rich. When these waters reach

the photic zone biomass production

increases significantly. Generally, the

Benguela Upwelling is divided into 8

upwelling zones with different intensities

and season maxima (Lutjeharms & Meeuwis,

1987). From these cells the winds push the

upwelled water up to 600 km offshore

producing filaments that mix with the south

Atlantic surface waters. These mixed waters

have still higher nutrient contents and lower

SSTs (Lutjeharms & Meeuwis, 1987;

Lutjeharms & Stockton, 1987; Summerhayes

et al., 1995).

At around 15°S the BCC meets the

southward flowing warm and nutrient-poor

Angola Current (Figure 1.5) at the so-called

7

Introduction

Angola-Benguela Front (ABF). The contact

zone is composed of several fronts arranged

into two zones, a northern and a southern

one (Kostianoy & Lutjeharms, 1999).

Depending on the meridional pressure

gradient determined by the strength and

position of the SAA, the southern frontal

zone moves north or south. When the

atmospheric pressure gradient is steep the

southern frontal zone moves north and the

ABF becomes narrower and sharper.

However, when the SAA is exceptional

weak the ABF shifts further south and

allows the warm Angola Current to

penetrate southwards as far as 24°S

(Meeuwis & Lutjeharms, 1990). In this case

the precipitation on the continent increases

(Hirst & Hastenrath, 1983; Nicholson &

Entekhabi, 1987; Rouault, 2003) which

happens on inter-annual timescales similar

to the El Niño (Shannon et al., 1986).

1.6 Climate during the late Miocene to

Pliocene

During the Miocene to Pliocene the climatic

conditions were different. Glaciations were

unipolar with expanded ice shields only on

Antarctica which caused a northwards

position of all climatic zones relative to

today, especially on the northern hemisphere

(Flohn, 1981) and a weak meridional

atmospheric pressure gradient. The still

existing tropical sea gateways involved a

different ocean circulation. Especially the

closing of the Central American Seaway

influenced the deepwater circulation from

the late Miocene onwards, e.g. by stepwise

increasing the north Atlantic deepwater

formation (Billups, 2002). This pre-setting

together with slightly higher CO2 levels than

today (around 400 ppmv) (Bartoli et al.,

2011) enabled a general warmer and wetter

climate during the late Miocene compared to

today (Knorr et al., 2011). Since only the

southern pole had extended ice shields the

meridional pressure/thermal gradient was

flat so that the winds were weak, too

(Flower & Kennett, 1994). However, during

the late Miocene the ice sheets on east

Antarctica expanded again increasing the

meridional pressure gradients and

intensifying the winds (Flower & Kennett,

1994). Especially the strengthened trade

winds had an impact on the ocean by

inducing e.g. the Benguela Upwelling during

the late Miocene (Siesser, 1980;

Rommerskirchen et al., 2011) which had

further influenced the precipitation on the

continent. The upwelling together with the

final closing of the Central American Seaway

changed the ocean circulation in the

Pliocene and introduced the modern cross-

equatorial heat transfer from the southern

hemisphere to the northern (Prange &

Schulz, 2004; Salzmann et al., 2011).

Vegetation and biome models suggest that

due to the general warmer and wetter

climate of the Miocene many nowadays arid

regions were then covered with shrublands

or woodlands (Pound et al., 2011). In Africa

the vegetation of the Miocene is also

indicating wetter conditions, however, with

the intensification of the meridional pressure

gradient the aridification started in south-

west Africa, e.g. the Namib desert (Van

Zinderen Bakker, 1984; Segalen et al., 2002).

8

Chapter I

The desertification progressed later in other

parts of Africa, e.g. the Sahara which is

thought to be arid since 6-7 Ma and east

Africa which also shows an arid phase

between 7 and 6 Ma (Bonnefille, 2010). A

Niger delta pollen record showed that

grasslands begun to spread in West Africa in

the late Miocene probably also caused by

aridification (Morley & Richards, 1993).

1.7 Material and methods

The material for this study is sampled of a

core retrieved at Ocean Drilling Program

Site 1081A (19°37.1818’S 11°19.1598’E)

(Figure 1.5) in 4794.1 m water depth (Berger

et al., 1998). The site is located around 160

km offshore Namibia on the Walvis ridge.

The sediment is composed of olive-gray

clayely nannofossil ooze and olive-gray to

black clays (Berger et al., 2002). The age

model is based on biostratigraphy, magnetic

reversals and magnetic susceptibility

resulting in sedimentation rates between 2

and 5 cm/ka (Berger et al., 2002).

Sedimentation at the site, on the one hand,

is influenced by filaments derived from the

Benguela Upwelling leading to increased

productivity of e.g. phytoplankton. The high

productivity creates good conditions for

fossilization of a diverse microfauna and

flora (e.g. cysts of dinoflagellates). On the

other hand the easterly winds transport dust,

pollen and spores as well as microscopic

plant fragments to the site which are buried

in the sediments.

Palynomorphs (e.g. pollen, spores and

dinoflagellate cysts) are made of organic

compounds such as sporopollenin or

dinosporin which are resistant in non-

oxidating environments. Since the high

productivity at the site reduces the oxygen

content in the sediments, the preservation of

microfossils, e.g. palynomorphs, is enhanced

(Wefer et al., 1998).

1.7.1 Palynological preparation

For the study a standard palynological

preparation was followed (e.g. Traverse,

2007). At first each sample is measured with

the water displacement method to specify

the amount of sediment (usually around 5

cm3) which can be used later to calculate

flux rates or accumulation rates of pollen

and dinocysts. Additionally, each sample

received a known number of indicator

spores (Lycopodium clavatum). For the actual

chemical treatment the sample was treated

with ~5 % HCl to dissolve calcareous

material and treated with ~20 % HF to

dissolve silcates. The residuals were not

exposed to KOH to avoid dissolving of

sensitive organic-walled dinoflagellate cysts.

After neutralisation the residuals were sieved

under ultrasonic treatment using a 8 µm

screen effectively holding back particles

bigger than 10-15 µm. The sieved and

cleaned residuals were stored in water. For

further investigation some drops of the

residual were placed with glycerol on a slide

to allow examination with a light

transmitting microscope using

magnifications of 400x, 600x and 1000x. For

each sample at least 300 pollen and spores

and 300 dinoflagellates were counted.

Besides the pollen and dinoflagellate cysts

also the number of charred particles (e.g.

9

Introduction

microscopic charcoal) was counted. In the

same slides the numbers of indicator spores

were counted to calculate the concentration

of the palynomorphs per ml sediment. The

accumulation or flux rates of palynomorphs

was then calculated by multiplying the

concentration with the sedimentation rate

after Berger et al. (2002).

1.7.2 Terrestrial palynomorphs

Terrestrial derived palynomorphs can be

transported via rivers or by wind to oceanic

sites. Spores derive from ancient plants

such as bryophytes and ferns whereas pollen

comes from angiosperms (flowering plants)

and gymnosperms. Pollen and spores are

very diverse and allow determining their

botanical origin at least to family level.

Although each plant species produces

different amounts of pollen grains they can

be used to reconstruct vegetation changes

and to distinguish the source of origin. In

the south-east Atlantic the buried pollen and

spores were in most parts transported by

winds. Dupont & Wyputta (2003) showed

that the sediments along the Walvis ridge

contain pollen and spores transported via

easterly winds from the adjacent desert, semi

desert and savannah. Hence, the sediment

core 1081A can be used to describe changes

in the vegetation of the continent.

1.7.3 Marine palynomorphs

Every dinoflagellate has ecological

boundaries in terms of temperature, salinity

or nutrients. For recent dinoflagellates

laboratory and in-situ measurements can

reveal those boundaries which enables us to

use their fossil counterparts to reconstruct

(palaeo-)environmental conditions (Dale et

al., 2002; Marret & Zonneveld, 2003;

Holzwarth et al., 2007; Zonneveld et al.,

2013). However, older sediments contain

cysts of extinct species with unknown

ecological boundaries. For these species

palaeo-distributions are the only tool to

narrow down the environmental range of

species. For the Miocene many modern

dinoflagellates were already present allowing

comparisons with modern values.

1.8 Thesis outline

In the following chapters I present the core

of the thesis which is formed by three

manuscripts accepted by or prepared for

submission to peer-reviewed journals.

Chapter II: The Role of Fire in Miocene

to Pliocene C4 Grassland and Ecosystem

Evolution (in press at Nature

Geoscience)

Sebastian Hoetzel, Lydie Dupont, Enno Schefuß,

Florian Rommerskirchen, Gerold Wefer

This study shows that fire was an important

trigger in the establishment of the modern-

like C4 grassland in the Kalahari. A

combination of a pollen, spore and

microscopic charcoal record and a

comparison with a carbon isotopic data set

revealed that C4 grasses were favoured

during times of intensified fire activity.

Therefore this studies answers hypothesis I

and II.

Chapter III: Miocene-Pliocene

Vegetation change in south-western

10

Chapter I

Africa (ODP Site 1081, offshore

Namibia) (ready for submission to

Palaeogeography, Palaeoclimatology,

Palaeoecology)

Sebastian Hoetzel, Lydie Dupont, Gerold Wefer

The third study focuses on the details of the

vegetation change during the Miocene to

Pliocene. An endmember unmixing model

was used to describe the vegetation change

which resulted in three different phases: A

relatively wet one with high representations

of Cyperaceae, a dry one with high more

desert and semi-desert plants, and an

intermediate phase when especially

grasslands expanded. These three phases

represent ongoing aridification. Additionally,

the study describes a sudden occurrence of

aquatic vegetation indicators at around 5 Ma

which is linked to a change of the course of

the Lower Cunene. This redirection of the

river might have been the main cause of the

desiccation of the palaeolake Cunene on

land into the present-day Etosha pan. This

manuscript focuses on hypothesis I.

Chapter IV: Miocene-Pliocene Stepwise

Intensification of the Benguela

Upwelling over the Walvis Ridge off

Namibia (ready for submission)

Sebastian Hoetzel, Lydie Dupont, Fabienne

Marret, Gerold Wefer

The second manuscript focuses on the

development of the Benguela Upwelling by

describing community changes of

dinoflagellate cyst assemblages. The results

show that during the late Miocene the

Angola-Benguela front was located further

southwards and that intensification of the

meridional pressure gradient let the front

migrate northwards. As a consequence the

upwelling intensified. During the Messinian

Salinity crisis the quality of the upwelled

water changed and enabled unique

dinoflagellate communities to bloom.

During the Pliocene upwelling intensified

again and influence of freshwater input is

suggested. This manuscript focuses on

hypothesis III.

1.9 References

Bartoli G., Hönisch B., & Zeebe R.E. (2011) Atmospheric CO2 decline during the Pliocene intensification of Northern Hemisphere glaciations. Paleoceanography, 26, .

Beerling D.J. & Osborne C.P. (2006) The origin of the savanna biome. Global Change Biology, 12, 2023–2031.

Berger W.H., Lange C.B., & Wefer G. (2002) Upwelling History of the Benguela-Namibia System: A Synthesis of Leg 175 Results. Proceedings of the Ocean Drilling Program, Scientific Results (ed. by G. Wefer, W.H. Berger, and C. Richter), pp. 1–103.

Berger W.H., Wefer G., Richter C., Lange C.B., Giraudeau J., Hermelin O., & Shipboard Scientific Party (1998) The Angola-Benguela Upwelling System: Paleoceanographic Synthesis of Shipboard Results from Leg 175. Proceedings of the Ocean Drilling Program, Initial Reports (ed. by G. Wefer, W.H. Berger, C. Richter, and E. Al.), pp. 505–531.

Billups K. (2002) Late Miocene through early Pliocene deep water circulation and climate change viewed from the sub-Antarctic South Atlantic. Palaeogeography, Palaeoclimatology, Palaeoecology, 185, 287–307.

11

Introduction

Bond W.J. (2008) What Limits Trees in C4 Grasslands and Savannas? Annual Review of Ecology, Evolution, and Systematics, 39, 641–659.

Bond W.J., Midgley G.F., & Woodward F.I. (2003) What controls South African vegetation — climate or fire? South African Journal Of Botany, 69, 79–91.

Bonnefille R. (2010) Cenozoic vegetation, climate changes and hominid evolution in tropical Africa. Global and Planetary Change, 72, 390–411.

Cerling T.E., Harris J.M., Macfadden B.J., Leakey M.G., Quadek J., Eisenmann V., & Ehleringer J.R. (1997) Global vegetation change through the Miocene / Pliocene boundary. Nature, 153–158.

Cowling R.M., Rundel P.W., Desmet P.G., & Esler K.J. (1998) Extraordinary High Regional-Scale Plant Diversity in Southern African Arid Lands: Subcontinental and Global Caomparisons. Diversity and Distributions, 4, 27–36.

Dale B., Dale A.L., & Jansen J.H.F. (2002) Dinoflagellate cysts as environmental indicators in surface sediments from the Congo deep-sea fan and adjacent regions. Palaeogeography, Palaeoclimatology, Palaeoecology, 185, 309–338.

Daniau A.-L., Sánchez Goñi M.F., Martinez P., Urrego D.H., Bout-Roumazeilles V., Desprat S., & Marlon J.R. (2013) Orbital-scale climate forcing of grassland burning in southern Africa. Proceedings of the National Academy of Sciences of the United States of America, 1–5.

Dupont L.M. & Wyputta U. (2003) Reconstructing pathways of aeolian pollen transport to the marine sediments along the coastline of SW Africa. Quaternary Science Reviews, 22, 157–174.

Edwards E.J., Osborne C.P., Strömberg C. a E., Smith S. a, Bond W.J., Christin P.-

A., Cousins A.B., Duvall M.R., Fox D.L., Freckleton R.P., Ghannoum O., Hartwell J., Huang Y., Janis C.M., Keeley J.E., Kellogg E. a, Knapp A.K., Leakey A.D.B., Nelson D.M., Saarela J.M., Sage R.F., Sala O.E., Salamin N., Still C.J., & Tipple B. (2010) The Origins of C4 Grasslands: Integrating Evolutionary and Ecosystem Science. Science, 328, 587–91.

Ehleringer J.R., Cerling T.E., & Helliker B.R. (1997) C4 photosynthesis, atmospheric CO2, and climate. Oecologia, 112, 285–299.

Flohn H. (1981) A hemispheric circulation asymmetry during Late Tertiar. Geologische Rundschau, 70, 725–736.

Flower B.P. & Kennett J.P. (1994) The middle Miocence climatic transition: East Antarctic ice sheet development, depp ocean circulation and global carbon cycling. Palaeogeography, Palaeoclimatology, Palaeoecology, 108, 537–555.

Furley P.A., Rees R.M., Ryan C.M., & Saiz G. (2008) Savanna burning and the assessment of long-term fire experiments with particular reference to Zimbabwe. Progress in Physical Geography, 32, 611–634.

Gasse F., Chalié F., Vincens A., Williams M. a. J., & Williamson D. (2008) Climatic patterns in equatorial and southern Africa from 30,000 to 10,000 years ago reconstructed from terrestrial and near-shore proxy data. Quaternary Science Reviews, 27, 2316–2340.

Giess W. (1998) A preliminary Vegetation Map of South West Africa. Dinteria, 4, 5–112.

Gimeno L., Drumond A., Nieto R., Trigo R.M., & Stohl A. (2010) On the origin of continental precipitation. Geophysical Research Letters, 37, 1–7.

Hatch M.D. (1987) C4 photosynthesis: a unique blend of modified biochemistry, anatomy and

12

Chapter I

ultrastructure. Biochimica et Biopohysica Acta, 895, 81–106.

Herold N., Huber M., & Müller R.D. (2011) Modeling the Miocene Climatic Optimum. Part I: Land and Atmosphere. Journal of Climate, 24, 6353–6372.

Hirst A.C. & Hastenrath S. (1983) Atmosphere-Ocean Mechanisms of Climate Anomalies in the Angola-Tropical Atlantic Sector. Journal of Physical Oceanography, 13, 1146–1157.

Holzwarth U., Esper O., & Zonneveld K. (2007) Distribution of organic-walled dinoflagellate cysts in shelf surface sediments of the Benguela upwelling system in relationship to environmental conditions. Marine Micropaleontology, 64, 91–119.

Houston J. & Hartley A.J. (2003) The central Andean west-slope rainshadow and its potential contribution to the origin of hyper-aridity in the Atacama Desert. International Journal of Climatology, 23, 1453–1464.

Jacobs B.F., Kingston J.D., & Jacobs L.L. (1999) The Origin of Grass-Dominated Ecosystems. Annals of the Missouri Botanical Garden, 86, 590–643.

Keeley J.E. & Rundel P.W. (2005) Fire and the Miocene expansion of C4 grasslands. Ecology Letters, 8, 683–690.

Knorr G., Butzin M., Micheels A., & Lohmann G. (2011) A warm Miocene climate at low atmospheric CO2 levels. Geophysical Research Letters, 38, n/a–n/a.

Kostianoy A.G. & Lutjeharms J.R.E. (1999) Atmospheric effects in the Angola-Benguela frontal zone. Journal of Geophysical Research, 104, 20963–20970.

Leroux M. (1983) Le climat de L’Afrique tropicale (Texte and atlas). Champion, Paris.

Lutjeharms J.R.E. & Meeuwis J.M. (1987) The extent and variability of South-

East Atlantic upwelling. South African Journal of Marine Science, 5, 51–62.

Lutjeharms J.R.E. & Stockton P.L. (1987) Kinematics of the upwelling front off southern Africa. South African Journal of Marine Science, 5, 35–49.

Marret F. & Zonneveld K.A.F. (2003) Atlas of modern organic-walled dinoflagellate cyst distribution. Review of Palaeobotany and Palynology, 125, 1–200.

Mayaux P., Bartholomé E., Fritz S., & Belward A. (2004) A new land-cover map of Africa for the year 2000. Journal of Biogeography, 31, 861–877.

Meeuwis J.M. & Lutjeharms J.R.E. (1990) Surface thermal characteristics of the Angola-Benguela front. South African Journal of Marine Science, 9, 261–279.

Morley R. & Richards K. (1993) Gramineae cuticle: a key indicator of Late Cenozoic climatic change in the Niger Delta. Review of Palaeobotany and Palynology, 77, 119–127.

Nicholson S. (2000) The nature of rainfall variability over Africa on time scales of decades to millenia. Global and Planetary Change, 26, 137–158.

Nicholson S.E. (1996) A Review of Climate Dynamics and Climate Variability in Eastern Africa. The Limnology, Climatology, and Paleoclimatology of the East African Lakes (ed. by T.C. Johnson and E.O. Odada), pp. 25–56. Gordon and Breach Publisher,

Nicholson S.E. & Entekhabi D. (1987) Rainfall Variabillity in Equatorial and Southern Africa: Relationships with Sea Surface Temperatures along the Southwestern Coast of Africa. Journal of Climate and Applied Meteorology, 26, 561–578.

O’Leary M. (1988) Carbon Isotopes Photosynthesis Fractionation techniques may reveal new aspects of

13

Introduction

carbon dynamics in plants. BioScience, 38, 328–336.

Osborne C.P. (2008) Atmosphere, ecology and evolution: what drove the Miocene expansion of C4 grasslands? The Journal of Ecology, 96, 35–45.

Pagani M., Zachos J.C., Freeman K.H., Tipple B., & Bohaty S. (2005) Marked decline in atmospheric carbon dioxide concentrations during the Paleogene. Science, 309, 600–3.

Petterson R.G. & Stramma L. (1991) Upper-level circulation in the South Atlantic Ocean. Progress in Oceanography, 26, 1–73.

Pound M.J., Haywood A.M., Salzmann U., & Riding J.B. (2012) Global vegetation dynamics and latitudinal temperature gradients during the Mid to Late Miocene (15.97–5.33Ma). Earth-Science Reviews, 112, 1–22.

Pound M.J., Haywood A.M., Salzmann U., Riding J.B., Lunt D.J., & Hunter S.J. (2011) A Tortonian (Late Miocene, 11.61–7.25Ma) global vegetation reconstruction. Palaeogeography, Palaeoclimatology, Palaeoecology, 300, 29–45.

Prange M. & Schulz M. (2004) A coastal upwelling seesaw in the Atlantic Ocean as a result of the closure of the Central American Seaway. Geophysical Research Letters, 31, L17207.

Prasad V., Strömberg C.A.E., Alimohammadian H., & Sahni A. (2005) Dinosaur coprolites and the early evolution of grasses and grazers. Science, 310, 1177–80.

Prasad V., Strömberg C.A.E., Leaché A.D., Samant B., Patnaik R., Tang L., Mohabey D.M., Ge S., & Sahni A. (2011) Late Cretaceous origin of the rice tribe provides evidence for early diversification in Poaceae. Nature communications, 2, 480.

Retallack G.J. (2001) Cenozoic Expansion of Grasslands and Climatic Cooling. The Journal of Geology, 109, 407–426.

Rommerskirchen F., Condon T., Mollenhauer G., Dupont L., & Schefuß E. (2011) Miocene to Pliocene development of surface and subsurface temperatures in the Benguela Current system. Paleoceanography, 26, 1–15.

Rommerskirchen F., Plader A., Eglinton G., Chikaraishi Y., & Rullkötter J. (2006) Chemotaxonomic significance of distribution and stable carbon isotopic composition of long-chain alkanes and alkan-1-ols in C4 grass waxes. Organic Geochemistry, 37, 1303–1332.

Rouault M. (2003) South East tropical Atlantic warm events and southern African rainfall. Geophysical Research Letters, 30, 8009.

Sage R.F., Li M., & Monson R.K. (1999a) The Taxonomic Distribution of C4 Photosynthesis. C4 Plant Biology (ed. by R.F. Sage and R.K. Monson), pp. 551–584. Academic Press, San Diego.

Sage R.F., Wedin D.A., & Li M. (1999b) The Biogeography of C4 Photosynthesis: Patterns and Controlling Factors. C4 Plant Biology (ed. by R.F. Sage and R.K. Monson), pp. 313–374. Academic Press, San Diego.

Salzmann U., Williams M., Haywood A.M., Johnson A.L. a., Kender S., & Zalasiewicz J. (2011) Climate and environment of a Pliocene warm world. Palaeogeography, Palaeoclimatology, Palaeoecology, 309, 1–8.

Scheiter S., Higgins S.I., Osborne C.P., Bradshaw C., Lunt D., Ripley B.S., Taylor L.L., & Beerling D.J. (2012) Fire and fire-adapted vegetation promoted C4 expansion in the late Miocene. The New phytologist, 195, 653–666.

Segalen L., Renard M., Pickford M., Senut B., Cojan I., Le Callonnec L., & Rognon P. (2002) Environmental and

14

Chapter I

climatic evolution of the Namib Desert since the Middle Miocene: the contribution of carbon isotope ratios in ratite eggshells. Comptes Rendus Geoscience, 334, 917–924.

Senut B., Pickford M., & Ségalen L. (2009) Neogene desertification of Africa. Comptes Rendus Geoscience, 341, 591–602.

Shannon L.V., Boyd A.J., Brundrit G.B., & Taunton-Clark J. (1986) On the existence of an EI Niño-type phenomenon in the Benguela System. Journal of Marine Research, 44, 495–520.

Siesser W.G. (1980) Late Miocene Origin of the Benguela Upswelling System off Northern Namibia. Science, 208, 283–5.

Southgate R.I., Masters P., & Seely M.K. (1996) Precipitation and biomass changes in the Namib Desert dune ecosystem. Journal of Arid Environments, 33, 267–280.

Still C.J. & Powell R.L. (2010) Continental-Scale Distributions of Vegetation Stable Carbon Isotope Ratios. Isoscapes: Understanding Movement, Pattern, and Process on Earth Through Isotope Mapping (ed. by J.B. West, G.J. Bowen, T.E. Dawson, and K.P. Tu), pp. 179–193. Springer Netherlands, Dordrecht.

Strömberg C.A.E. (2011) Evolution of Grasses and Grassland Ecosystems. Annual Review of Earth and Planetary Sciences, 39, 517–544.

Summerhayes C.P., Kroon D., Rosell-Melé A., Jordan R.W., Schrader H.-J., Hearn R., Villanueva J., Grimalt J.O., & Eglinton G. (1995) Variability in the Benguela Current upwelling system over the past. Progress in Oceanography, 35, 207–251.

Thomasson J.R., Nelson M.E., & Zakrzewski R.J. (1986) A fossil grass (gramineae: chloridoideae) from the miocene with kranz anatomy. Science, 233, 876–8.

Tipple B.J. & Pagani M. (2007) The Early Origins of Terrestrial C4 Photosynthesis. Annual Review of Earth and Planetary Sciences, 35, 435–461.

Traverse A. (2007) Paleopalynology. Springer Netherlands, Dordrecht.

Uno K.T., Cerling T.E., Harris J.M., Kunimatsu Y., Leakey M.G., Nakatsukasa M., & Nakaya H. (2011) Late Miocene to Pliocene carbon isotope record of differential diet change among East African herbivores. Proceedings of the National Academy of Sciences of the United States of America, 108, 6509–14.

Urban M.A., Nelson D.M., Jimenez-Moreno G., Chateauneuf J.-J., Pearson A., & Hu F.S. (2010) Isotopic evidence of C4 grasses in southwestern Europe during the Early Oligocene-Middle Miocene. Geology, 38, 1091–1094.

Wefer G., Berger W.H., Richter C., Adams D.D., Anderson L.D., Andreasen D.J., Brüchert V., Cambray H., Christensen B.A., Frost G.M., Giraudeau J., Gorgas T.J., Hermelin J.O.R., Jansen J.H.F., Lange C.B., Laser B., Lin H.-L., Maslin M., Meyers P.A., Motoyama I., Murray R.W., Perez M.E., Pufahl P.K., Spiess V., Vidal L., Wigley R., & Yamazaki T. (1998) Site 1081. Proceedings of the Ocean Drilling Program, Initial Reports, 175, 223–727.

White F. (1983) The Vegetation of Africa, Natural Resources Research. UNESCO, Paris.

Zachos J., Pagani M., Sloan L., Thomas E., & Billups K. (2001) Trends, rhythms, and aberrations in global climate 65 Ma to present. Science, 292, 686–93.

Zachos J.C., Dickens G.R., & Zeebe R.E. (2008) An early Cenozoic perspective on greenhouse warming and carbon-cycle dynamics. Nature, 451, 279–83.

Van Zinderen Bakker E.M. (1975) The Origin and Palaeoenvironment of the

15

Introduction

16

Namib Desert Biome. Journal of Biogeography, 2, 65–73.

Van Zinderen Bakker E.M. (1984) Palynological evidence for Late Cenozoic arid conditions along the Namibia coast from Holes 532 and 530A, Leg 75. Deep Sea Drilling Project Initial Reports of the Deep Sea Drilling Project, 75, 763–768.

Zonneveld K.A.F., Marret F., Versteegh G.J.M., Bogus K., Bonnet S., Bouimetarhan I., Crouch E., de Vernal A., Elshanawany R., Edwards L., Esper O., Forke S., Grøsfjeld K., Henry M., Holzwarth U., Kielt J.-F., Kim S.-Y., Ladouceur S., Ledu D., Chen L., Limoges A., Londeix L., Lu S.-H., Mahmoud M.S., Marino G., Matsouka K., Matthiessen J., Mildenhal D.C., Mudie P., Neil H.L., Pospelova V., Qi Y., Radi T., Richerol T., Rochon A., Sangiorgi F., Solignac S., Turon J.-L., Verleye T., Wang Y., Wang Z., & Young M. (2013) Atlas of modern dinoflagellate cyst distribution based on 2405 datapoints. Review of Palaeobotany and Palynology, 191, 1–197.

Chapter II

Chapter II: The Role of Fire in Miocene to

Pliocene C4 Grassland and Ecosystem Evolution

Sebastian Hoetzel, Lydie Dupont, Enno Schefuß, Florian Rommerskirchen*,

Gerold Wefer

MARUM – Center for Marine Environmental Sciences, University of Bremen, Germany

*now at: Intertek Food Services GmbH, Bremen, Germany

2.1 Abstract

Modern savannah grasslands, in which grasses using the C4-photosynthetic pathway dominate,

became globally established through the late Miocene to Pliocene (between 8 and 3 million years

ago) (Edwards et al., 2010). Although this tropical savannah vegetation covers up to 40 % of the

land surface (Strömberg, 2011), the particular processes driving Mio-/Pliocene C4 grass

expansion are still obscure. Several mechanisms have been proposed including decreasing

atmospheric CO2 levels (Cerling et al., 1997), increased aridity, and enhanced rainfall seasonality

(Sage, 2001; Edwards et al., 2010; Strömberg, 2011). Recent studies suggest, however, that the

role of fire in transforming ecosystems is underestimated (Keeley & Rundel, 2005; Beerling &

Osborne, 2006; Osborne, 2008; Edwards et al., 2010). Here we demonstrate that grasses have

expanded since 8 million years ago (Ma) in conjunction with increased aridity and enhanced fire

activity in south-western Africa. This finding is based on pollen and microscopic charcoal

records that are complemented by stable-carbon-isotope analyses of plant waxes derived from

marine sediments of ODP Site 1081 offshore Namibia. The increased occurrence of fire

favoured C4 grass expansion and led to the establishment of modern-like C4 grasslands. During

the Pliocene, C4 plants expanded further due to continuing aridification leading to the

establishment of desert and semi-desert conditions. Our results confirm that ecological

disturbance by fire provided an essential feedback mechanism in the establishment of tropical

C4 grasslands.

17

Fire and the Miocene to Pliocene C4 Grassland Evolution

2.2 Main Text

Eighty percent of the modern warm-

temperate to tropical savannah grassland’s

annual primary biomass production is by

plants using the C4 photosynthetic pathway

(Sage, 2001). This type of photosynthesis

evolved multiple times in different grass

lineages starting around 30 Ma (Pagani et al.,

2005; Vicentini et al., 2008; Urban et al.,

2010), but large-scale expansion of C4 plants

did not occur until the Late Miocene (c. 8

Ma) (Edwards et al., 2010; Strömberg, 2011)

and was previously thought to be

synchronous on all continents (Cerling et al.,

1997). Initially, decreasing atmospheric

pCO2 levels during the Neogene were

assumed to be the main driver because C4

plants are favoured over C3 plants under low

atmospheric pCO2 (Cerling et al., 1997;

Ehleringer et al., 1997). Atmospheric CO2

concentrations, however, had probably

already dropped below the critical threshold

as early as the Oligocene (Pagani et al.,

2005), i.e., long before the C4 plant

expansion. Recent studies on the expansion

of C4 grasslands in various regions have

revealed globally asynchronous expansions

(Edwards et al., 2010; Strömberg, 2011),

which were often preceded by an expansion

of C3 grasslands (Edwards et al., 2010), for

example, in North America and China.

Hence, other environmental factors, like

changes in the amount and seasonality of

precipitation and disturbance by fires, have

been suggested as potential triggers for the

C4 grass expansion (Keeley & Rundel, 2005;

Beerling & Osborne, 2006; Osborne, 2008;

Edwards et al., 2010). The fire hypothesis

has been recently strengthened by vegetation

modelling (Scheiter et al., 2012) but data to

validate the role of fire are still missing. To

understand the origin of the southern

African C4 grass savannah we conducted a

palynological and biogeochemical study of

sediments of Ocean Drilling Program

(ODP) Site 1081 (Fig. 2.1) located 160 km

off the Namibian coast (Supplementary

Information), at the latitude of the modern

grassland-woodland transition. Along the

coast, desert and semi-desert prevail (White,

1983) as a result of intense aridity induced

by upwelling of cold waters along the coast

driven by the strong south-easterly winds

(Dupont et al., 2005). Consequently, pollen,

plant particles, microscopic charcoal and

plant waxes are transported to the site

mainly by south-easterly winds (Dupont &

Wyputta, 2003) from areas located farther

inland , such as the Kalahari. The Kalahari

dry savannah today is a dry summer rain

area (precipitation of less than 500 mm; in

Namibia usually between 150 and 250 mm

per year (White, 1983)) and consists of

grasslands comprising up to 90 % C4 plants

(Still & Powell, 2010). In the Kalahari

burning occurs regularly at the end of the

dry period when the vegetation is driest

whereas the coastal desert does not provide

enough fuel for fires (Fig. 2.1).

18

Chapter II

Figure 2.1: Modern vegetation map(White, 1983) of the study area with the location of ODP Site 1081. In blue are oceanic features, arrows represent Benguela Current and Benguela Coastal Current, blue filled area represents today’s location of the Benguela coastal upwelling. Dashed grey line shows bathymetry of the 4000 m depth line. Narrow black arrows indicate prevailing surface winds during austral spring (based on long-term mean data from 2000 to 2010 from the Twentieth Century Reanalysis Project, NOAA). Fire symbols represent large fires from October to November 2011 based on Global Fire Maps (NASA Earth Data).

We focused on pollen types from grasses

and semi-desert and desert plants found in

ODP Site 1081 to trace the evolution of arid

ecosystems. We also counted microscopic

charred particles reflecting past fire activities

(Supplementary Information). In addition,

we measured the stable carbon isotopic

compositions (δ13C) of long-chain n-alkanes

derived from the protective wax coating of

leaf-bearing terrestrial plants (Eglinton &

Hamilton, 1967) to estimate the relative C4

plant contribution. Due to their enhanced

carbon fixation mechanism, C4 plants

produce wax lipids with heavier δ13C values

(Vogts et al., 2009). In general, the δ13C

values of the n-C31 and n-C33 alkanes (δ13C n-

C31, δ13C n-C33) for C3 plants from rainforest

and savannah vegetation average around -36

‰ VPDB while those for C4 grasses are

around -22 ‰ VPDB (Vogts et al., 2009).

Today, the n-C31 alkane is present in all

major vegetation types of south-western

Africa, whereas the n-C33 alkane is

predominantly derived from savannah herbs

and C4 grasses in particular (Vogts et al.,

2009).

19

Fire and the Miocene to Pliocene C4 Grassland Evolution

Figure 2.2: Miocene-Pliocene records from ODP Site 1081. a, Grass pollen percentages considering the total pollen and spore sum. Shading indicates the 95 % confidence interval. b, Stable carbon isotopes (δ13C) of plant wax (n-C31-alkanes in green, n-C33-alkanes in yellow) in per mil Pee Dee Belemnite (‰ VPDB) with two sigma error bars. c, ratio of charred particles over the sum of charred particles, spores and pollen. d, Pollen percentages of semi-desert / desert vegetation considering the sum of pollen and spores excluding Poaceae. Shading as in a. e, Sea surface temperatures (°C) calculated from the alkenone unsaturation index UK’37

20

Chapter II

The record of the grass pollen (Fig. 2.2a)

ranges between 30 and 45 % from 9 until 8

Ma and steeply rises from ~ 30 % at 8.4 Ma

to ~ 70 % at 6.8 Ma in the Late Miocene,

indicating that grasses expanded to form

extensive grasslands in the Kalahari between

8 and 7 Ma. The δ13C values of n-C31 (Fig

2.2b) increase from less than -29 ‰ before

7.1 Ma to -27 ‰ at 6 Ma and reach -26 ‰

by 5 Ma indicating that the increase of C4

plants in the vegetation took place over a

longer period. For the δ13C of n-C33, the

trend is stronger starting at around -30 ‰

and reaching -26 ‰ already at 6 Ma

representing a significant change in the

composition of the vegetation to a greater

proportion of C4 grasses (see Supplementary

Information). The stable isotope

compositions of plant waxes before 7 Ma

indicate that C4 plants represented a minor

share of the vegetation. After 6 Ma, C4

plants became a dominant component of the

vegetation and a composition similar to the

modern situation was established, which

corresponds to δ13C changes in fossil ratite

eggshells from the Namib region (Ségalen et

al., 2006). Also in East Africa, grasslands

expanded before the spread of C4 grasses

(Feakins et al., 2013). The grassland

expansion in south-western Africa is

accompanied by higher ratios of charred

particles from 7.1 to 5.8 Ma (Fig. 2.2c),

reflecting enhanced fire activity likely caused

by the expanded grasslands providing more

combustible material. Additionally, a

pronounced dry season was necessary to dry

out the vegetation, making it more

susceptible to fire (see Supplementary

Information) (Osborne, 2008). This period

of enhanced fire activity was followed by a

drier period starting at 5.8 Ma indicated by a

rise of desert and semi-desert vegetation (Fig

2.2d) and a drop in Poaceae pollen and

charcoal. The lowered charcoal abundance is

a result of an expansion of desert and semi-

desert vegetation, which decreased the fuel

supply. The grass vegetation expanded again

briefly at 4.8 Ma, as indicated by grass-pollen

values of up to 81.5 % together with an

increase in the charcoal ratio. Afterwards,

the grass-pollen curve fluctuates strongly,

but generally shows a decreasing trend and

finally stabilises at 3.3 Ma at values of

around 60 %. This, together with the

increasing desert and semi-desert

representation, indicates expanded deserts

and semi-deserts with reduced fire potential

due to further increased aridity. The δ13C

values, however, continued to increase and

the relative abundance of Poaceae pollen

decreased after 4.8 Ma, which we interpret

to reflect a shift towards a greater

proportion of C4 grasses within the grass

community.

The initial grassland expansion in south-

western Africa and especially the expansion

of the desert and semi-desert vegetation

over the entire record correspond to the

onset and intensification of the Benguela

Upwelling (Rommerskirchen et al., 2011).

Decreasing sea-surface temperatures

offshore Namibia (Fig 2.2e) were driven by

intensified upwelling caused by strengthened

trade winds, leading to aridification of the

coastal region (Dupont et al., 2005). A

21

Fire and the Miocene to Pliocene C4 Grassland Evolution

further impact on the regional climate had

the expansion of the grasslands itself by

reducing the regional rainfall due to lower

evapotranspiration compared to more tree-

and shrub-spotted landscapes (Beerling &

Osborne, 2006). Furthermore, grasslands are

less efficient in moisture recycling and hence

the expansion of grasslands intensified the

dry season. Thus, the expansion of

grasslands created a positive feedback of

increased aridification, promoting higher fire

activity as long as there was enough fuel to

burn (Beerling & Osborne, 2006).

Today, fires have a strong influence on the

composition of modern African vegetation,

e.g., by shaping the border between

woodlands and savannahs (Furley et al.,

2008). Through multiple positive feedbacks

of droughts and fires, burning accelerates

the loss of tree cover and promotes C4

grassland expansion (Beerling & Osborne,

2006), as grasses are morphologically well

adapted to fires and rebound much faster

than trees (Bond, 2008). Vegetation

modelling suggests that the exclusion of fire

in the Kalahari savannah would increase tree

cover by up to 80 % (Bond et al., 2003). A

further ecological impact of fire would have

been of an evolutionary nature, affecting

selection processes on the composition of

the grasses, e.g. by favouring different grass

lineages containing C4 grasses. The grass

lineage of Aristidoideae, an almost

exclusively C4 clade, is most diverse and its

members most widely distributed in those

areas of South Africa where a distinct dry

season prevails and the rate of disturbance is

high (Visser et al., 2012). Members of the

tribe Andropogoneae (also C4) mostly occur

in areas that are frequently burnt (Visser et

al., 2012). Today, members of both clades

are common elements in Namibian grass

floras (Klaassen & Craven, 2003)

Our findings thus suggest that fires were

selecting and favouring ecologically better

adapted C4 grasses over C3 grasses, so that

the high-C4 plant fraction in the vegetation

could proliferate. Aridification favoured first

the expansion of grasslands and later the

coastal desert and semi-desert vegetation.

The grass expansion created massive

accumulations of fuel for fires. An

intensified fire regime was established

because of a generally drier climate, an

enhanced dry season, and a higher fuel

supply. We propose that after fire triggered

the initial establishment of C4 plants,

providing essential conditions for the

dominance of C4 grasses, the continuing

aridification led to a retreat of grasslands in

which the C3 grasses disproportionally

declined, leading to a relative increase in C4

plants. Driven by the enhancement of the

Benguela Upwelling the vegetation changed

to sparse grassland and desert vegetation,

starving the fire of its fuel. The decrease of

grasslands combined with the expansion of

desert and semi-desert vegetation

dramatically reduced the fire occurrence in

the Pliocene. Aridification and the

corresponding greater fire activity probably

also affected the vegetation in other regions

where savannahs expanded during the

Miocene. A palynological record from the

Niger deep-sea delta shows that grass pollen

22

Chapter II

and microscopic charcoal concurrently

increased 7 Ma ago (Morley & Richards,

1993), indicating that fire also played a

pivotal role in the development of grass-

dominated mesic savannahs in West Africa.

We thus infer that an intensified fire regime

caused by aridification triggered the

establishment of C4 plants in African

grasslands and created preconditions for the

domination of C4 grasses in savannah

grasslands, which had an irreversible effect

on the evolution of African ecosystems.

2.3 Methods summary

For palynological analyses the sediment was

treated with 5 % HCl and 20 % HF and all

particles smaller than 10 – 15 µm were

removed by ultrasonic sieving. The residuals

were mounted on a microscope slide and

analysed at a magnification of 400. For each

sample at least 300 pollen and spores were

counted, and additionally all charred

particles. The statistical error (95 %

confidence intervals) was calculated after

Maher (1971).

For organic-geochemical analyses, sediment

samples were extracted by accelerated

solvent extraction using

dichloromethane/methanol (9:1, three times

for 5 min., 70 bar, 100°C). The total lipid

extracts were separated by Al2O3 column

chromatography using

hexane/dichloromethane (9:1) to elute an

apolar fraction and

dichloromethane/methanol (1:1) to elute a

polar fraction. The polar fraction, containing

alkenones for UK’37 analyses, was saponified

using 0.5 M KOH in methanol and re-

extracted with n-hexane after addition of

distilled water. Polar compounds were

converted to trimethylsilyl-ether derivatives

before UK’37 analysis by gas chromatography

with flame ionisation detection (GC-FID).

SST estimates were calculated using the

equation of Müller et al. (1998), and errors

calculated using duplicate analyses. Apolar

fractions were purified by elution over Ag-

impregnated SiO2-colums to remove

unsaturated compounds. Compound-

specific stable carbon isotope compositions

were analysed on a ThermoFisher Trace GC

Ultra coupled via a combustion reactor to a

Finnigan MAT 252 mass-spectrometer.

Isotope values were measured against

calibrated CO2 reference gas. δ13C values are

reported in ‰ VPDB. An externally

calibrated n-alkane standard was measured

every 6 runs. Averaged absolute deviations

from known δ13C values were consistently <

0.5 ‰ VPDB. All samples were run at least

in duplicate with a reproducibility of < 0.5

‰ VPDB. Precision and accuracy of the

squalane internal standard were 0.5 and 0.0

‰ VPDB, respectively.

2.4 References

Beerling D.J. & Osborne C.P. (2006) The origin of the savanna biome. Global Change Biology, 12, 2023–2031.

Bond W.J. (2008) What Limits Trees in C4 Grasslands and Savannas? Annual Review of Ecology, Evolution, and Systematics, 39, 641–659.

23

Fire and the Miocene to Pliocene C4 Grassland Evolution

Bond W.J., Midgley G.F., & Woodward F.I. (2003) What controls South African vegetation — climate or fire? South African Journal Of Botany, 69, 79–91.

Cerling T.E., Harris J.M., Macfadden B.J., Leakey M.G., Quadek J., Eisenmann V., & Ehleringer J.R. (1997) Global vegetation change through the Miocene / Pliocene boundary. Nature, 153–158.

Dupont L.M., Donner B., Vidal L., Pérez E.M., & Wefer G. (2005) Linking desert evolution and coastal upwelling: Pliocene climate change in Namibia. Geology, 33, 461.

Dupont L.M. & Wyputta U. (2003) Reconstructing pathways of aeolian pollen transport to the marine sediments along the coastline of SW Africa. Quaternary Science Reviews, 22, 157–174.

Edwards E.J., Osborne C.P., Strömberg C. a E., Smith S. a, Bond W.J., Christin P.-A., Cousins A.B., Duvall M.R., Fox D.L., Freckleton R.P., Ghannoum O., Hartwell J., Huang Y., Janis C.M., Keeley J.E., Kellogg E. a, Knapp A.K., Leakey A.D.B., Nelson D.M., Saarela J.M., Sage R.F., Sala O.E., Salamin N., Still C.J., & Tipple B. (2010) The Origins of C4 Grasslands: Integrating Evolutionary and Ecosystem Science. Science, 328, 587–91.

Eglinton G. & Hamilton R.J. (1967) Leaf Epicuticular Waxes. Science, 165, 1322–1335.

Ehleringer J.R., Cerling T.E., & Helliker B.R. (1997) C4 photosynthesis, atmospheric CO2, and climate. Oecologia, 112, 285–299.

Feakins S.J., Levin N.E., Liddy H.M., Sieracki a., Eglinton T.I., & Bonnefille R. (2013) Northeast African vegetation change over 12 m.y. Geology, 41, 295–298.

Furley P.A., Rees R.M., Ryan C.M., & Saiz G. (2008) Savanna burning and the

assessment of long-term fire experiments with particular reference to Zimbabwe. Progress in Physical Geography, 32, 611–634.

Keeley J.E. & Rundel P.W. (2005) Fire and the Miocene expansion of C4 grasslands. Ecology Letters, 8, 683–690.

Klaassen E.S. & Craven P. (2003) Checklist of grasses in Namibia. Southern African Botanical Diversity Network No. 20. SABONET, Pretoria & Windhoek.

Maher L.J.J. (1971) Nomograms for computing 0.95 confidence limits of pollen data. Review of Palaeobotany and Palynology, 13, 85–93.

Morley R. & Richards K. (1993) Gramineae cuticle: a key indicator of Late Cenozoic climatic change in the Niger Delta. Review of Palaeobotany and Palynology, 77, 119–127.

Müller P.J., Kirst G., Ruhland G., von Storch I., & Rosell-Melé A. (1998) Calibration of the alkenone paleotemperature index UK’37 based on core-tops from the eastern South Atlantic and the global ocean (60°N-60°S). Geochimica et Cosmochimica Acta, 62, 1757–1772.

Osborne C.P. (2008) Atmosphere, ecology and evolution: what drove the Miocene expansion of C4 grasslands? The Journal of Ecology, 96, 35–45.

Pagani M., Zachos J.C., Freeman K.H., Tipple B., & Bohaty S. (2005) Marked decline in atmospheric carbon dioxide concentrations during the Paleogene. Science, 309, 600–3.

Rommerskirchen F., Condon T., Mollenhauer G., Dupont L., & Schefuß E. (2011) Miocene to Pliocene development of surface and subsurface temperatures in the Benguela Current system. Paleoceanography, 26, 1–15.

Sage R.F. (2001) Environmental and Evolutionary Preconditions for the Origin and Diversification of the C4

24

Chapter II

25

Photosynthetic Syndrome. Plant Biology, 3, 202–213.

Scheiter S., Higgins S.I., Osborne C.P., Bradshaw C., Lunt D., Ripley B.S., Taylor L.L., & Beerling D.J. (2012) Fire and fire-adapted vegetation promoted C4 expansion in the late Miocene. The New phytologist, 195, 653–666.

Ségalen L., Renard M., Lee-Thorp J.A., Emmanuel L., Le Callonnec L., de Rafélis M., Senut B., Pickford M., & Melice J.-L. (2006) Neogene climate change and emergence of C4 grasses in the Namib, southwestern Africa, as reflected in ratite 13C and 18O. Earth and Planetary Science Letters, 244, 725–734.

Still C.J. & Powell R.L. (2010) Continental-Scale Distributions of Vegetation Stable Carbon Isotope Ratios. Isoscapes: Understanding Movement, Pattern, and Process on Earth Through Isotope Mapping (ed. by J.B. West, G.J. Bowen, T.E. Dawson, and K.P. Tu), pp. 179–193. Springer Netherlands, Dordrecht.

Strömberg C.A.E. (2011) Evolution of Grasses and Grassland Ecosystems. Annual Review of Earth and Planetary Sciences, 39, 517–544.

Urban M.A., Nelson D.M., Jimenez-Moreno G., Chateauneuf J.-J., Pearson A., & Hu F.S. (2010) Isotopic evidence of C4 grasses in southwestern Europe during the Early Oligocene-Middle Miocene. Geology, 38, 1091–1094.

Vicentini A., Barber J.C., Aliscioni S.S., Giussani L.M., & Kellogg E.A. (2008) The age of the grasses and clusters of origins of C4 photosynthesis. Global Change Biology, 14, 2963–2977.

Visser V., Woodward F.I., Freckleton R.P., & Osborne C.P. (2012) Environmental factors determining the phylogenetic structure of C4 grass communities. Journal of Biogeography, 39, 232–246.

Vogts A., Moossen H., Rommerskirchen F., & Rullkötter J. (2009) Distribution patterns and stable carbon isotopic composition of alkanes and alkan-1-ols from plant waxes of African rain forest and savanna C3 species. Organic Geochemistry, 40, 1037–1054.

White F. (1983) The Vegetation of Africa, Natural Resources Research. UNESCO, Paris.

2.5 Acknowledgments

This work was supported by the DFG

Research Center/Cluster of Excellence

‘MARUM - The Ocean in the Earth System’

and ‘GLOMAR – Bremen International

Graduate School for Marine Sciences’. F.R.

was supported by the Deutsche

Forschungsgemeinschaft (Sche903/6) within

the research unit ’Understanding Cenozoic

Climate Cooling: The Role of the Hydrology

Cycle, the Carbon Cycle, and Vegetation

Changes‘ (FOR 1070). 20th Century

Reanalysis data were provided by the

NOAA/OAR/ESRL PSD, Boulder,

Colorado, USA, from their Web site at

http://www.esrl.noaa.gov/psd. Global fire

map data were provided by NASA Earth

Data from their Web site at

https://earthdata.nasa.gov/.

2.6 Author Contributions

Palynological data analysis was carried out

by S.H. and biogeochemical data by F.R.;

Interpretation was carried out by S.H., L.D.,

E.S., F.R. and G.W. Correspondence and

requests for material should be addressed to

S.H.

Fire and the Miocene to Pliocene C4 Grassland Evolution

2.7 Supplementary Information

2.7.1 Material

The sampled 452 m-long core was retrieved

in a water depth of 794.1 m on the Walvis

ridge (19°37.1818’S 11°19.1598’E) and is

composed of olive-gray clayey nannofossil

ooze and olive-gray to black clays (Berger et

al., 2002). The age model is based on

biostratigraphy, magnetic reversals and

magnetic susceptibility (Berger et al., 2002).

Due to a narrow shelf and steep slope the

distance to the coast was not largely affected

by sea level changes. Microscopic charred

particles, pollen and spores as well as plant

waxes were transported predominantly by

winds. Today only one larger river exists in

the vicinity. The Cunene River was very

likely not flowing into the Atlantic until the

Pliocene but feeding a huge lake, the Cunene

lake (Hipondoka, 2005).

2.7.2 Charcoal and charred particles

Microscopic charcoal and charred particles

are produced when plants are burnt. During

fires microscopic particles can be elevated

within a heat plume reaching altitudes up to

5 km (Palmer & Northcutt, 1975) and

transported via wind over large areas

(Verardo & Ruddiman, 1996). Numbers of

charred particles can therefore be used to

reconstruct past fire activities (Tinner & Hu,

2003). We considered opaque rectangular

shaped organic particles bigger than 10 µm

(< 50 µm), e.g. true charcoal (burnt wood)

and other charred plant particles. We suggest

a similar source area for the recorded

charcoal as for the pollen and spores. Today

fires occur in north and east Namibia

especially during austral-spring (September

to November) (Silva, 2003) [NASA Earth

DATA]. We expressed charred particles as

ratios of those over the sum of charred

particles, spores and pollen. Because both

the pollen and spores and the charred

particles are wind-blown, effects of wind

strength via sorting are minimised.

2.7.3 Interpretation of the pollen record

Although grasses occur in almost all

vegetation types in SW Africa (White, 1983)

the main source area of Poaceae (grass)

pollen for sediments of ODP 1081 is the

Kalahari savannah (Dupont & Wyputta,

2003). Changes in the palaeo-record of grass

pollen, therefore, represent primarily

changes in the expansion of the Kalahari

savannah. Poaceae pollen percentages of the

early part of the record are comparable to

modern values (Figure 2.3) suggesting largely

expanded grasslands during the latest

Miocene. Prior to 8 Ma the record suggests a

patchy open grassland possibly mixed with

Cyperaceae and differing thus from the

modern vegetation.

26

Chapter II

Figure 2.3: Relative abundance of Poaceae pollen in surface sediments along the African west coast(Dupont, 2011). Modern vegetation is represented by yellow colours for open vegetations and grasslands, brown and orange for wood and bushlands, violet for agricultural used areas and green for forest vegetation(Mayaux et al., 2004).

Almost all the investigated sediment samples

are dominated by grass pollen which is

related to the wind pollination mechanism

of grasses producing large amounts of well

dispersible pollen grains. Because Poaceae

are over-represented in the pollen record

percentages of other pollen groups are based

on the total pollen and spore counts

excluding Poaceae pollen.

Typical origins of pollen of the desert and

semi-desert elements are vegetation types

along the coast of Namibia (White, 1983;

Giess, 1998). The coastal desert vegetation is

represented by pollen of Amaranthaceae

(including the subfamily Chenopodiaceae)

(Figure 2.4d). In the group of semi-desert

taxa pollen of Euphorbia, Acacia spp.,

Blepharis, Monechma, Nyctaginaceae and Riccia

are included (Giess, 1998). Pollen of the

desert / semi-desert group is always present

(except for two samples at 7.96 and 7.8 Ma)

but percentages increase only in the younger

parts of the record. At 6.3 Ma and 6.1 Ma

values of up to 11 % are recorded. From 5.8

Ma onward the pollen record of this group

is steadily increasing reaching 23 % (of total

pollen without Poaceae pollen) at 3.1 Ma

suggesting hyper-arid conditions along the

coast, which is coherent with the

interpretation that desert / semi-desert

vegetation expanded at the expense of

savannah grasses and shrubs in the Pliocene.

The desert / semi-desert vegetation would

have produced less fuel to feed fires. The

rise of δ13C alkanes during the Pliocene

indicating further expansion of C4 vegetation

27

Fire and the Miocene to Pliocene C4 Grassland Evolution

may be attributed to an expansion of

drought-adapted C4-species from the

Amaranthaceae, which is indicated by an

increase of Amaranthaceae pollen at 3.2 Ma

(Figure 2.4d) corresponding with a δ13C

maximum.

Another significant pollen group including

Petalidium, Ruellia, Acanthaceae spp.,

Piliostigma, Colophospermum mopane, Brachystegia

and Lycopodium represents the woodland.

Pollen from woodlands is present in the

entire profile but fluctuations are minor and

usually in the range from 5 to 10 % (Figure

2.4a). We suggest that they mainly originated

from north-east Namibia or Angola and

indicate a constant background signal.

Cyperaceae (sedges) are mainly but not

exclusively growing in wetter areas such as

riversides and along lakes. The initial pollen

percentages are relatively high and increasing

until 8.3 Ma with maximum values of 29 %

(twice as high as in subsurface sediments

(Dupont & Wyputta, 2003)), decline to low

values until 4.6 Ma, and rise again to the end

of the record (Figure 2.4b). The high

Cyperaceae pollen percentages of the early

part are linked to wetter conditions. For the

period after 4.6 Ma we suggest a different

origin probably related to a change in the

course of the Cunene River. Prior to 4.6 Ma

the Cunene River flowed into the Cunene

Lake but was re-routed to the Atlantic

Ocean afterwards. Discharge of the Cunene

River into the Atlantic ceased the water flow

into the Cunene Lake causing the onset of

its desiccation leading to the development of

the Etosha Pan (Hipondoka, 2005). The

additional pollen source area enclosed by the

Cunene River is indicated by the occurrence

of pollen of aquatic indicators such as

Nymphaeaceae and Typha in the marine

sediments, which are missing in the older

parts of the record.

2.7.4 Indications for aridification

The percentages of Asteraceae pollen are

increasing until 6 Ma (Figure 2.4c) which

suggest increasingly dry conditions, because

many Asteraceae shrubs and annual herbs

are even better adapted to water stress than

grasses. The charcoal record also indicates

aridification because the charcoal particles

are derived from fires which more

frequently occur with enhanced aridity (see

below for additional factors influencing

biomass burning). Additionally, the

transport of microscopic charcoal is

favoured under dry conditions because

winds can more easily pick up dry dust

(including microscopic charcoal). Further

indication of aridification during the first

period until 6 Ma is the decrease of

Cyperaceae pollen percentages during this

interval.

Over the studied period desert and semi-

desert pollen percentages are increasing

(Figure 2.4), in particular during the Pliocene

when Asteraceae pollen abundances decline,

which we interpret as an expansion of

deserts and semi-deserts caused by

intensified aridification along the coast.

28

Chapter II

Figure 1.4: Percentages of pollen of a woodland taxa, b Cyperaceae, c Asteraceae, and d Amaranthaceae. Percentage calculation is based on the total counts of pollen and spores excluding Poaceae pollen.

2.7.5 δ13C of plant-wax alkanes in C3 and

C4 plants

Relatively large ranges of δ13C values of n-

alkanes are detected in plants under different

environmental conditions, in particular for

C3 plants during water stress (Diefendorf et

al., 2010). During water stress the plants

close their stomata or develop less stomata

to increase the water use efficiency and

hence use more 13C so that the δ13C will

become less negative (Farquhar et al., 1989).

As each plant species has its environmental

boundaries (e.g. humidity) under which it

can exists and reproduce, its δ13C values can

be influenced by water stress on the short

term resulting in a certain range of δ13C. We

argue that under lasting non-favourable

conditions, however, the plants cannot

survive, so that changes in δ13C on longer

time-scales will reflect changes in vegetation

composition and water stress might play

only a subordinate role in affecting δ13C

values.

In the current vegetation of south-western

Africa, the n-C31 alkane is present in all main

vegetation types, whereas the n-C33 is

predominantly present in savannah plants

occurring in dry climates and, in particular,

C4 grasses (Rommerskirchen et al., 2006;

Vogts et al., 2009). The δ13C values of n-C31

and n-C33 of C3 plants range between -30 and

-42 ‰ whereas the δ13C values of C4 grasses

29

Fire and the Miocene to Pliocene C4 Grassland Evolution

are between -26 and -18 ‰

(Rommerskirchen et al., 2006; Vogts et al.,

2009). An expansion of C4 grasses would

therefore show a stronger increase in the

δ13C of n-C33 than for the n-C31 which

indeed is observed during the first period of

enhanced fire occurrence (Figure 2.2).

However, due to the large ranges in δ13C

values of plants of both photosynthetic

pathways absolute estimates of C3/C4 plant

ratios are difficult to constrain and

associated with large errors. Therefore, we

refrain from making quantitative estimates.

2.7.6 Seasonality, aridity, grasses and fire

A pronounced climatic seasonality as well as

massive fuel are requirements for an intense

fire regime (Osborne, 2008). Wet austral

summers lead to accumulation of fuel by

providing favourable conditions for grass

expansion. An increase of rainfall by 200

mm per year (during the growing season)

has been inferred to double the grass

production (Daniau et al., 2013). Long

enough dry winters increase the vegetation’s

flammability (Daniau et al., 2013). In

contrast, during dry summers less biomass

will accumulate and during short dry seasons

the vegetation will not so easily ignite. Both

conditions lead to a less intense fire regime.

Hence, we suggest that fire promoting

conditions existed in the period of enhanced

charcoal abundance and expansion of C4

grasses. During the Pliocene the desert and

semi-desert expanded on the cost of

grasslands which is most probably caused by

a decrease in overall rainfall and potentially a

shift to a less fire promoting climate. A close

connection of fire and C4 grasslands in sub-

Saharan Africa has already been

demonstrated for glacials of the past one

million years (Bird & Cali, 1998).

2.7.7 References

Berger W.H., Lange C.B., & Wefer G. (2002) Upwelling History of the Benguela-Namibia System: A Synthesis of Leg 175 Results. Proceedings of the Ocean Drilling Program, Scientific Results (ed. by G. Wefer, W.H. Berger, and C. Richter), pp. 1–103.

Bird M.I. & Cali J.A. (1998) A million-year record of fire in sub-Saharan Africa. Nature, 394, 767–769.

Daniau A.-L., Sánchez Goñi M.F., Martinez P., Urrego D.H., Bout-Roumazeilles V., Desprat S., & Marlon J.R. (2013) Orbital-scale climate forcing of grassland burning in southern Africa. Proceedings of the National Academy of Sciences of the United States of America, 1–5.

Diefendorf A.F., Mueller K.E., Wing S.L., Koch P.L., & Freeman K.H. (2010) Global patterns in leaf 13C discrimination and implications for studies of past and future climate. Proceedings of the National Academy of Sciences, 107, 5738–5743.

Dupont L. (2011) Orbital scale vegetation change in Africa. Quaternary Science Reviews, 30, 3589–3602.

Dupont L.M. & Wyputta U. (2003) Reconstructing pathways of aeolian pollen transport to the marine sediments along the coastline of SW Africa. Quaternary Science Reviews, 22, 157–174.

Farquhar G.D., Ehleringer J.R., & Hubick K.T. (1989) Carbon Isotope Discrimination and Photosynthesis. Annual Review of Plant Physiology and Plant Molecular Biology, 40, 503–537.

30

Chapter II

31

Giess W. (1998) A preliminary Vegetation Map of South West Africa. Dinteria, 4, 5–112.

Hipondoka M. (2005) The Development and Evolution of Etosha Pan, Namibia. Julius-Maximilian-Universität Würzburg,

Mayaux P., Bartholomé E., Fritz S., & Belward A. (2004) A new land-cover map of Africa for the year 2000. Journal of Biogeography, 31, 861–877.

Osborne C.P. (2008) Atmosphere, ecology and evolution: what drove the Miocene expansion of C(4) grasslands? The Journal of Ecology, 96, 35–45.

Palmer T.Y. & Northcutt L.I. (1975) Convection Columns Above Large Experimental Fires. Fire Technology, 11, 111–118.

Rommerskirchen F., Plader A., Eglinton G., Chikaraishi Y., & Rullkötter J. (2006) Chemotaxonomic significance of distribution and stable carbon isotopic composition of long-chain alkanes and alkan-1-ols in C4 grass waxes. Organic Geochemistry, 37, 1303–1332.

Silva J.M.N. (2003) An estimate of the area burned in southern Africa during the 2000 dry season using SPOT-VEGETATION satellite data. Journal of Geophysical Research, 108, 1–11.

Tinner W. & Hu F.S. (2003) Size parameters , size-class distribution and area-number relationship of microscopic charcoal: relevance for fire reconstruction. The Holocene, 13, 499–505.

Verardo D.J. & Ruddiman W.F. (1996) Late Pleistocene charcoal in tropical Atlantic deep-sea sediments: Climatic and geochemical significance. Geology, 24, 855–857.

Vogts A., Moossen H., Rommerskirchen F., & Rullkötter J. (2009) Distribution patterns and stable carbon isotopic composition of alkanes and alkan-1-ols from plant waxes of African rain forest

and savanna C3 species. Organic Geochemistry, 40, 1037–1054.

White F. (1983) The Vegetation of Africa, Natural Resources Research. UNESCO, Paris.

Chapter III

Chapter III: Miocene-Pliocene Vegetation

change in south-western Africa (ODP Site 1081,

offshore Namibia)

S. Hoetzel1, L. Dupont1 & G. Wefer1

1Marum – Center for Marine Environmental Sciences, Bremen, Germany

3.1 Abstract

Aridification is an important aspect of Late Neogene climate change in south-western Africa

probably caused by modifications in the atmospheric circulation in relation to the initiation and

intensification of the Benguela Upwelling System due to globally steepening of the meridional

pressure gradient. Intensification of the meridional pressure gradients influenced the climate

intensively which had then an impact on the vegetation. However, vegetation changes of south-

western Africa from the Miocene to Pliocene have not yet been reported and only indirectly

investigated. Here, we present a pollen record of marine core ODP 1081A retrieved 160 km

offshore Namibia covering the time between 9 and 2.7 Ma. Using an endmember unmixing

model we distinguished three vegetation phases: a relative wet phase showing higher

representations of Cyperaceae, a dry one with a strong representation of desert and semi-desert

plants, and a transition phase when especially grasses expanded. The three phases indicate

ongoing aridification probably caused by intensified meridional pressure gradients and shifts of

the Congo-Air Boundary. Additionally, aquatic vegetation indicators appear in our pollen record

from around 5 Ma on, which we attribute to a relocation of the lower course of the Cunene

River to its modern outlet in the Atlantic Ocean. Redirection of the Cunene River toward the

Atlantic would have deprived the palaeolake Cunene of an important source of fresh-water

ultimately resulting in desiccation of the lake and the formation of the Etosha Pan.

33

Miocene-Pliocene Vegetation change

3.2 Introduction

The vegetation distribution in south-western

Africa is linked to climate, in particular to

rainfall, which depends on the atmospheric

circulation. Major features of the

atmospheric circulation in south-western

Africa are the Congo-Air boundary (CAB)

(Fig. 3.1) and the strong south-east trade

winds blowing along the coast (Nicholson,

2000; Gasse et al., 2008) that are responsible

for the very dry environment of the Namib

Desert in combination with the cold sea

surface temperatures of the Benguela

Upwelling System (Petterson & Stramma,

1991). Initiation of upwelling, implying the

increase of the trade winds, is dated in the

late Miocene between 10 to 15 Ma (Siesser,

1980; Diester-Haass et al., 1990; Heinrich et

al., 2011; Rommerskirchen et al., 2011).

Increase of the trade winds is attributed to a

steepening of the meridional pressure

gradient in association with global cooling

during the late Neogene, when the climate

shifted from the warm and humid

conditions of the mid-Miocene to drier and

cooler ones of the Plio/Pleistocene. Driven

by the growth of the ice sheets on East

Antarctica (Wright et al., 1992; Zachos et al.,

2001; Billups & Schrag, 2002), the southern

polar front moved northwards strengthening

the meridional pressure gradient and the

pressure systems on the Southern

Hemisphere (Flower & Kennett, 1994;

Zachos et al., 2001). In detail, the South

Atlantic Anticyclone got stronger and so did

the trade winds along the Namibian coast

hindering air masses to penetrate into the

continent. The late Neogene aridification

started in the Namib region from which it

spread over the African continent (Senut et

al., 2009). In spite of the importance of the

region, changes of climate and vegetation in

south-western Africa are poorly investigated.

Van Zinderen Bakker (1984) describes one

palynological sample of Late Miocene age

which might have been originated from a

dry, very open vegetation in Namibia coeval

with palms growing in the Cape Region

(Van Zinderen Bakker, 1984). Most

reconstructions of south-western African

climate are based on sedimentological data.

They are usually limited to the development

of the Namib desert where fossil dunes

(aeolianites) including ratite egg shells and

other macro and micro faunal fossils were

dated between 16 and 17 Ma (Pickford et al.,

1995; Senut et al., 1998, 2009). Fossil fauna

sites in the Namibian Sperrgebiet (SW

Namibia) indicate a mid-Miocene vegetation

of woodland or open forest in an area that is

extremely arid today (Senut et al., 2009).

Stable carbon (δ13C) and stable oxygen

(δ18O) isotope compositions of fossil ratite

eggshells showed that the vegetation in

Namibia changed and a C4 grassy vegetation

established through the late Miocene to

Pliocene underlining aridification (Ségalen et

al., 2006; Dupont et al., 2013). In the greater

Cape Flora Region (southernmost Namibia

and west South Africa) more summer-

drought and dry-adapted vegetation types

gradually replaced tropical species during the

late Miocene (Dupont et al., 2011). Wetter

conditions for the late Miocene compared to

those of the Pliocene are also suggested by

modelling studies (Francois et al., 2006).

34

Chapter III

So far, a detailed vegetation reconstruction

of the Late Miocene during the aridification

in south-western Africa is missing. Here, we

reconstruct the Miocene to Pliocene

vegetation of northern Namibia based on

the pollen and spore content of marine

sediments from Ocean Drilling Program Site

1081 Hole A, retrieved offshore Namibia.

We discuss the vegetation change in south-

western Africa during the late Neogene and

link these to the global cooling and

aridification.

Figure 3.1: Map of Africa with the major vegetation groups (after White, 1983). Average position of the Congo-Air Boundary in January and July after (Tyson & Preston-Whyte, 2000).

3.3 Modern regional climate and south-

western African vegetation

Modern south-western Africa is

climatologically controlled by aridity which

is reflected in the vegetation (Fig. 3.1). There

are two major gradients of decreasing annual

rainfall, one directed north to south and the

other east to west (White, 1983; Gasse et al.,

2008). Along the coast precipitation is

lowest and seldom reaches 100 mm per year

causing the aridity of the Namib Desert. The

Namib is characterized by extremely sparse

and contracted vegetation, which consists of

isolated plants. In some areas vegetation is

completely absent (White, 1983). Inland

parts of south-western Africa receive

between 250 and 500 mm rainfall per year.

Here, wooded grassland of the Kalahari

35

Miocene-Pliocene Vegetation change

savanna occurs. Between the Kalahari and

the Namib lies the Karoo shrubland (Nama

Karoo) receiving precipitation between 100

and 250 mm. Typically, the landscape in the

Nama Karoo is dotted with shrubs or small

bushy trees of families such as Compositae

(or Asteraceae), Fabaceae and Acanthaceae

(White, 1983; Cowling et al., 1998). Most of

the precipitation in south-western Africa

falls in austral summer, whereas in the South

African Cape Region including the succulent

Karoo the rain falls mainly during austral

winter (Tyson & Preston-Whyte, 2000). The

winter rain area in Namibia is very rich in

endemic species, many of them belonging to

Aizoaceae, Geraniaceae, and other families

(Cowling et al., 1998). In northern Namibia

the Zambezian woodland grows in a climate

with very pronounced seasonality and

precipitation ranging from 500 to more than

1000 mm per year (White, 1983). The

woodland is defined as an open stand of

trees forming a canopy of at least 40 %.

Between these wooded areas the ground is

covered by grasses, herbs or shrubs. The

vegetation is linked to rainfall which itself is

linked to the atmospheric circulation. In the

north humidity derives mainly from the

Atlantic, whereas further southwards in the

summer rain areas of south-western Africa

including the Kalahari savanna, air masses

derived from the Indian Ocean transported

by the Easterlies are the major source of

precipitation (Gimeno et al., 2010). Along

the coast the strong south easterly trade

winds hinder air masses to penetrate over

the continent (Gasse et al., 2008). Where

Easterlies and Atlantic derived air masses

meet the Congo-Air boundary (CAB) is

formed which is considered a southern

branch of the Inter Tropical Convergence

Zone (ITCZ). During austral summers the

CAB is located around 20-15°S whereas

during the winter it is located at 9-6°S

(Leroux, 1983). The position of the CAB is

determined by the global thermal gradient

and the strength and position of the Hadley

cell (Nicholson, 2000). Therefore it is very

likely that the position of the CAB changed

in the geological past.

3.4 Methods and material

The material was sampled from cores

retrieved at Ocean Drilling Program Site

1081 Hole A (19°37.1818’S 11°19.1598’E)

(Fig. 3.1) in 794.1 m water depth. The site is

located 160 km off the Namibian coast and

is influenced by upwelling and easterly winds

(Berger et al., 1998). The sediment is

composed of olive-gray clayely nannofossil

ooze and olive-gray to black clays (Berger et

al., 2002). The age model is based on

biostratigraphy, magnetic reversals and

magnetic susceptibility resulting in

sedimentation rates between 2 and 5 cm/ka

(Berger et al., 2002).

Seventy-one samples were taken with ages

between 9 and 2.7 Ma and prepared for

palynological investigation. Volume was

measured using water displacement. For

decalcification diluted HCl (5%) was used

and afterwards the material was stored in

diluted HF (~20%) for two days to remove

silicates. Before the mineral dissolving step,

two tablets with a known number of

Lycopodium spores were added to determine

36

Chapter III

the pollen and spore concentration per ml.

The residuals were sieved under ultrasonic

treatment to remove particles smaller than

10-15 µm. The cleaned residuals were stored

in water and mounted in glycerol for

investigation under a light microscope using

magnifications of 400x and 1000x. For each

sample at least 300 pollen grains and spores

were identified and counted. For pollen

identification Scott (1982), Ybert (1979),

Bonnefille and Riollet (1980), Maley (1970),

the African Pollen Database

http://apd.sedoo.fr and the reference

collection of African pollen grains of the

Department of Palynology and Climate

Dynamics of the University of Göttingen

was used.

For calculating the pollen grain flux rates the

pollen and spore concentration per ml was

multiplied by the sedimentation rates of

Berger et al. (2002).

We carried out a multivariate analysis using

an endmember model unmixing procedure

(Weltje, 1997). We used a version of the

unmixer algorithm programmed in

MATLAB by Dave Heslop in 2008. The

program is specifically designed to analyse

percentage data.

3.5 Results

3.5.1 Pollen percentages and pollen flux

rates

The identified pollen and spores were

grouped into vegetation categories (Table 1).

Based on the total number of pollen and

spores the percentages were calculated and

plotted in Figures 3.2-3.5. Additionally,

pollen flux rates are plotted in Figure 3.2.

Figure 3.2: Overview of the ratio of endmembers, the relative abundances of vegetation groups in a cumulative diagram, and the pollen grain flux ratios (logarithmic scale) over the studied record.

37

Miocene-Pliocene Vegetation change

Pollen flux rates are generally high but

somewhat lower in the lower part of the

record before 5.3 Ma ranging from 3419 to

19,615 grains per cm² per ka. After 5.3 Ma

the pollen flux becomes higher with a

minimum of 10,288 and a maximum of

62,587 grains per cm² per ka.

152 pollen and spore types have been

identified, of which 73 types occurred in

more than five samples (Table 1, at the end

of this chapter). We grouped the pollen taxa

into a) desert, b) shrubland, c) Compositae,

d) Poaceae, e) aquatic vegetation, f)

mountain vegetation, g) woodland

vegetation and h) Cyperaceae (Table 1)

based on Giess (1998), Scott (1982), and

White (1983). Most abundant are pollen

from the families of Poaceae, Cyperaceae

and Compositae. The desert group includes

Tribulus and species of the Amaranthaceae

(inclusive Chenopodiaceae).

The pollen record is dominated by Poaceae

pollen. However, the period before 8-7 Ma

is characterized by lower relative values for

Poaceae (less than 45 %) and relatively high

representations of Cyperaceae (up to 20 %),

woodlands (around 10 %), mountain

vegetation (reaching 10 %) and desert

vegetation (around 5 %).

Between 7.5 and 4.5 Ma the Poaceae pollen

abundances are high up to 82.7 %

fluctuating between 60 and over 80 %.

Especially the representation of Cyperaceae,

woodlands and mountain vegetation

decrease or are very low. From 5.3 Ma on

representatives of the aquatic vegetation are

present. After 4.5 Ma woodland is again

slightly better represented with pollen

percentages between 4 and 8 % instead of

less than 4 %. Contemporaneously,

Cyperaceae pollen is reappearing again

reaching 8 %. Additionally, percentages of

the shrubland (maximum at 8 % at 3.3 Ma)

and the desert indicators (Amaranthaceae)

scoring in EM1 (maximum at 8 % at 3.1 Ma)

are increasing. After 4 Ma the Poaceae

pollen percentages fluctuate around 60 %.

3.5.2 Endmember results

We used a model with three principal

components or endmembers (randomly

ordered and named EM1, EM2 and EM3)

explaining over 98 % of the variance (r² =

0.983). Calculation of endmembers was

stopped at 1000 iterations resulting in a

convexity at termination of -1.92. Only

pollen and spore types occurring in at least

five samples were considered. In Table 1, we

listed pollen and spore types and coefficients

of determination for each component (rc2).

Sixty-four pollen and spore types had

significant scores (rc2 > 0.055) using a

confidence level of 0.95.

Hence, the pollen record can be summarised

into the three Endmembers (Fig. 3.2).

Basically there is a shift of the dominant

contribution from the EM2 (9-7.5 Ma) via

EM3 (7.5-5.3 Ma) to EM1 (from 5.3 Ma on).

The contribution of EM3 disappears after

3.5 Ma. Although the majority of pollen

does not score on one endmember only, the

three endmembers differ clearly in their

mixtures of pollen and spore types, which

can be interpreted in terms of vegetation

(Table 1). In most cases the pollen types

38

Chapter III

classified in a vegetation group score on one

endmember.

EM1 is mainly dominated by Poaceae

pollen. Additionally, most representatives of

shrubland and aquatic vegetation score

highest on EM1 (Figure 3.3).

Figure 3.3: Relative abundances of pollen groups of which the individual pollen types have high scores on Endmember 1. ‘Desert’ comprises pollen of Amaranthaceae and Merremia, ‘Shrubland’ pollen of Blepharis-type, Justica-type, Euphorbia, Acacia, Neuradaceae and spores of Riccia, and ‘Aquatic’ pollen of Nymphaea and Typha. Upper panel shows the ratio of EM1 compared to the other EMs.

Furthermore, Amaranthaceae score on

EM1 (Table 1 and Fig. 3.3).

Complementing, the Compositae Pentzia-,

Berkehya- and Artemisia-types score also high

on EM1. The total contribution of EM1 on

the total ratio is relatively low prior to 8 Ma

and its peaks increases to 0.4 until around 6

Ma (Fig. 3.3). From 6 Ma on, the

contribution rises to 0.8 at around 4.6 Ma

and fluctuates then between 0.7 and 1.

Cyperaceae and most pollen types of the

groups woodland, mountain vegetation, and

desert vegetation score high on EM2.

Additionally, the Compositae pollen types

Cotula, Dicoma, Gerbera, Vernonia and

undetermined liguliflorae types score high

on this endmember. EM2 is more

prominent in the early parts of the record

with contributions of 0.6 to1 around until 8

Ma. From 8 to 7 Ma the contribution of

EM2 decreases at levels below 0.4 and

reaches zero between 6.4 and 6.2 Ma

Afterwards the contribution stays below 0.3

and is again zero from 4.9 until 4.5 Ma.

Most vegetation groups have one or a few

pollen or spore types scoring on EM3. The

only category almost exclusively scoring on

EM3 is the group of spores. Additionally,

Dombeya (Malvaceae) and other Malvaceae,

Proteaceae spp., Phyllanthus

(Euphorbiaceae), Commiphora (Burseaceae),

the Compositae pollen type Senecio and

undetermined tubuliflorae types, have high

scores on EM3. Poaceae pollen scores

39

Miocene-Pliocene Vegetation change

almost as high on EM3 as on EM1. Until

7.7 Ma and after 4.3 Ma the contribution of

EM3 is usually below 0.2 except for two

spikes reaching over 0.3 (at 9 and 8.7 Ma).

Between 7.7 and 4.4 Ma the percentages of

EM3 fluctuate and reach highest

contributions (close to 1) between 6.4 and

6.1 Ma.

Figure 3.4: Relative abundances of pollen groups of which the individual pollen types have high scores on Endmember 2. ‘Woodlands’ comprise pollen of Peltalidium, Ruellia, other Acanthaceae, trichothomosulcate-type, Brachystegia, Colophospermum mopane, Detarium, Grewia, ‘Mountain’ spores of Phaeoceros, Lycopodium, Mohria-type and pollen of Stoebe-type, Ericaceae, Olea, Podocarpus, Passerina and other Thymelaeaceae, ‘Compositae’ pollen of Cotula-type, Dicoma-type, Gerbera-type, liguliflorae type, Vernonia-type, and ‘Desert’ pollen of Aizoaceae, Monsonia, Welwitschia, Tribulus. Upper panel shows the ratio of EM2 compared to the other EMs.

40

Chapter III

Figure 3.5: Relative abundances of pollen groups of which the individual pollen types have have high

scores on Endmember 3. ‘Compositae’ comprise pollen Senecio-type and tubuliflorae type, ‘Spores’

comprise Osmunda-type, Selaginella-type, spore-type 2 bacculate, spore-type 3 leiotrilete, ‘Woodlands’

comprise pollen of Dombeya and Malvaceae spp., and ‘Shrubland savanna’ pollen of Commiphora,

Phyllanthus and Proteaceae. Upper panel shows the ratio of EM3 compared to the other endmembers.

3.6 Discussion

3.6.1 Transport and source of pollen and

spores and the desiccation of Lake

Cunene

The transport of pollen and spores to ocean

sites is either fluvial or aeolian. To ODP Site

1081 pollen is mainly carried by easterly

winds. The aeolian pollen sources are along

the coast and in the hinterland covering

areas of the Namib Desert, Nama Karoo

and the savanna (Dupont & Wyputta, 2003).

Sedimentological studies showed that the

south-east trade winds and seasonal winds

from the north-eastern sector have been

present since the late Miocene (Ségalen,

2004).

Geochemical studies have shown that the

perennial Cunene River also influences the

sedimentation on the Walvis ridge (Bremner

& Willis, 1993). Shi et al. (1998) suggested

significant fluvial input from the Cunene

River at a site located further north based on

the occurrence of pollen originating from

the Angolan mountains. Thus, a minor part

of our pollen transport is probably fluvial.

However, it is suggested that the Cunene

River did not reach the Ocean until the

Pliocene but was feeding a palaeolake, called

Lake Cunene (Hipondoka, 2005).

The modern Etosha pan (Fig. 3.1) is

described as a relic of the palaeolake Lake

Cunene (also called Lake Etosha)

(Hipondoka, 2005). The pan is the

41

Miocene-Pliocene Vegetation change

topographical centre of the Owambo Basin

which has sediments deposited mainly by

the Cubango and Cunene Rivers. Both rivers

dominated the sedimentation in the basin

since the early Neogene (Dill et al., 2012). A

micropalaeontological study revealed that a

freshwater lake, Lake Cunene, existed since

16 Ma (Dill et al., 2012) and a

sedimentological-palaeontological study

showed that the lake existed until late

Miocene or early Pliocene and then started

to desiccate (Miller et al., 2010). The

desiccation is attributed to a redirection of

the lower course of the Cunene River in

addition to aridification of the region

(Hipondoka, 2005). It is assumed that the

river relocated and connected Lake Cunene

with the Atlantic Ocean by the end of the

Miocene or the beginning of the Pliocene.

Our pollen record supports the idea that the

Cunene River started to discharge into the

Atlantic by the beginning of the Pliocene.

An increase in the pollen flux ratio at 5.3 Ma

is probably the result of additional fluvial

discharge of pollen. Since 5.1 Ma pollen of

Nymphaeaceae and Typha (Fig. 3.5) regularly

appears in our record together with an

increase in the representation of Cyperaceae

(sedges) and woodland. We interpret the

occurrence of pollen from aquatics like

Nymphaeaceae and Typha as the effect of

drainage of a freshwater body, e.g. a lake.

The desiccation of the lake might also have

led to higher representations of halophytes

like some Amaranthaceae species. The

presence of aquatic indicators in the pollen

record of ODP Site 1081 marks the start of

fresh-water discharge into the Atlantic by

the Cunene River and the desiccation of

Lake Cunene shortly before 5 Ma.

3.6.2 Vegetation development

The pollen assemblages over the entire

record suggest open vegetation based on the

high Poaceae pollen percentages and

relatively low number of pollen from trees.

However, prior to 7 Ma representations of

woodland are higher indicating wetter

conditions. Another important component

during this period is afromontane vegetation

including Passerina and Ericaceae. The

mountain vegetation might have been

relative wide-spread in Angola until 8 Ma.

Furthermore, Cyperaceae pollen, originating

from sedges usually growing in swampy

environments, is very abundant in the period

before 7 Ma and almost twice as common

than today (Dupont & Wyputta, 2003). In

our interpretation it reflects open vegetation

including small ponds, or a landscape with

small rivers. Paradoxically, the Tribulus

record, an indicator of the Namib Desert

(Fig. 3.4, main contributor in the desert

vegetation group) is also very high during

the period. Although Tribulus is a typical

desert plant, it grows in areas with

intermittently access to water such as river

banks of ephemeral rivers and Tribulus may

form flower carpets immediately after

rainfall (Scott, 1982; White, 1983). The

occurrence of this desert indicator

corroborates a mid to late Miocene age of

the Namib Desert (Senut et al., 2009).

Between 8 and 7 Ma the representation of

savanna vegetation increased as indicated by

the records of Poaceae, Commiphora,

42

Chapter III

Phyllanthus and Protea (Fig. 3.5). In terms of

pollen percentages, the latter taxa are

underrepresented, because their pollen is

dispersed by animals rather than wind. The

period from the latest Miocene to the early

Pliocene spanned the major expansion of

savanna grasslands in many parts of the

world (Cerling et al., 1997; Tipple & Pagani,

2007; Strömberg, 2011). Here, in south-

western Africa the savanna started to expand

at around 8.3 Ma as indicated by a rise in

Poaceae representation (Figs. 3.2, 3.3, 3.5).

The maximum values of Poaceae pollen of

more than 80 % (at 6 and 4.9 Ma) even

exceed maximum values recorded for the

Holocene (Dupont et al., 2007) suggesting

that during the Miocene savanna grasslands

expanded further and occurred more widely

than today. We propose that the climate in

terms of precipitation and the length of the

wet season favoured the expansion of

grasses, but also promoted fires. This is

indicated by the charcoal record of ODP

Site 1081 (same material as used in this

study) representing increased fire activity

between 7.1 and 5.8 Ma (Hoetzel et al. 2013

Nature Geoscience). It is possible that

barring fire the amount of rainfall in the

Miocene would have been sufficient to

enable a denser growth of trees and shrubs.

The fire regime might have favoured grasses

and suppressed woody vegetation as

indicated by vegetation models of the

Kalahari in which the tree cover increased

up to 80 % if fire is excluded (Bond et al.,

2003). The expansion of grasslands and the

shift to more savanna is consistent with

increased δ13C values of fossil ratite

eggshells indicating more C4 plants in the

vegetation (Ségalen et al., 2006).

Decreasing relative abundances of Poaceae

pollen indicates that grasslands remained

important but might have become less dense

after 5 Ma. Between 4 and 3 Ma elements

from desert vegetation (halophytic

Amaranthaceae) and shrubland became

more abundant suggesting increasingly drier

conditions.

3.6.3 Steps in aridification

The vegetation indicates that relative humid

conditions still prevailed during the Late

Miocene, which is conform to global climate

reconstructions and modelling (Pound et al.,

2011, 2012). A reason for a warm climate

might have been slightly higher CO2 values

(Bartoli et al., 2011; Pound et al., 2012).

Wetter conditions could be explained by

generally warmer sea surface temperatures

of the Southern Ocean prior to the closing

of the South American Sea-way, as models

suggest (Prange & Schulz, 2004; Lunt et al.,

2007). Furthermore, an expanded warm

pool would be associated with flat

meridional temperature gradients, a slower

Hadley circulation, and permanent El-Niño-

like conditions (Brierley et al., 2009 and

references therein). Together the warmer

oceans and the weaker upwelling fit into a

scheme with weak meridional pressure

gradients explaining the wet conditions in

south-western Africa during the Miocene.

The coastal desert developed in close

relation to the onset of upwelling in the

adjacent ocean 10 to 15 Ma ago (van

Zinderen Bakker, 1975; Siesser, 1980).

43

Miocene-Pliocene Vegetation change

During the initiation of the upwelling the

meridional thermal gradient was low so that

pressure systems and winds were weaker

than today. The Congo-Air boundary might

have been less well developed or less strong

allowing more humidity to penetrate into the

continent from the Atlantic (Dupont et al.,

2013), which would explain the relatively

wet vegetation of the hinterland (indicated

by higher values for Cyperaceae and EM2)

in combination with a desert strip along the

coast.

The endmember model indicates three

phases of vegetation development

represented by EM2, EM3, and EM1,

respectively. Based on the scores of the

pollen types on each of the endmembers we

interpret the shift in vegetation as the result

of aridification, beginning before 8 Ma and

continuing after 5 Ma. The increasing aridity

is manifested by the still wide-spread

grasslands starting to get less dense together

with an expansion of shrubland vegetation

around 4 Ma. Further aridification is

suggested by the expansion of halophytic

desert plants around 3.3 Ma (Fig. 3.3).

The increasing aridity is probably caused by

atmospheric strengthening of the pressure

systems, e.g. steepening of the meridional

thermal/pressure gradient. The drivers for

the steepening are twofold. Firstly, the

pressure zones and the ITCZ might have

been pushed northwards (on the southern

hemisphere) as a consequence of ice sheet

expansions on Antarctica (Billups, 2002),

similar to the northward movement during

austral winters. Alternatively the ITCZ

contracted, as a result of compression of

atmospheric circulation due to decreasing

global temperatures (Collins et al., 2010).

Secondly, the Atlantic Meridional

Overturning circulation intensified at the

end of the Miocene (Poore et al., 2006;

Butzin et al., 2011) which led to a cooling,

especially of the Southern Hemisphere by a

strong heat transport from the Southern to

the Northern Hemisphere (Prange & Schulz,

2004). Both led to an intensification of the

trade winds (including upwelling), a stronger

CAB and a more restricted transport of

humidity from the Atlantic. Intensification

of upwelling and therefore of trade winds

since 10 Ma has recently been demonstrated

by a study of ODP Site 1085

(Rommerskirchen et al., 2011). With the

strengthening of the trade winds and the

decrease of coastal SSTs, the major source

of precipitation shifted from the Atlantic to

the Indian Ocean leading to less

precipitation over south-western Africa

(Dupont et al., 2013). The resulting

aridification would have favoured first the

expansion of savanna and later the increase

of desert and shrubland vegetation.

3.7 Conclusions

Miocene to Pliocene samples from the ODP

Site1081 have been investigated for their

pollen and spore content to reconstruct

vegetation changes. The pollen record

shows a continuing aridification manifested

in three different steps represented by three

endmembers. The earlier and wetter period

runs from the beginning of the record until

ca. 8 Ma and is characterised by stronger

representations of swamp, woodland and

44

Chapter III

mountain vegetation. The period might

represent weakly developed meridional

pressure gradients and, therefore, weak

atmospheric conditions. During the

following period (8 – 5 Ma) savanna

grasslands expanded in Namibia indicating

increasingly arid conditions. Probably the

meridional thermal gradient got steeper and

wind systems more intense. The record is

complemented by the third period (5-2.7

Ma) representing further aridification. This

aridification is indicated by some reduction

in grassy vegetation cover joined by an

expansion of shrubland and desert

vegetation.

Apart from the vegetation development, the

record indicates that a rerouting of the lower

course of the Cunene River to the Atlantic

Ocean draining the palaeolake Cunene took

place shortly before 5 Ma by the sudden

appearance of Typha and Nymphaeaceae

pollen in combination with generally higher

pollen flux ratios.

3.8 Acknowledgements

This work was supported by the DFG

Research Center/Cluster of Excellence

‘MARUM - The Ocean in the Earth System’

and ‘GLOMAR – Bremen International

Graduate School for Marine Sciences’. For

laboratory support we thank Marta Witek.

3.9 References

Bartoli G., Hönisch B., & Zeebe R.E. (2011) Atmospheric CO2 decline during the Pliocene intensification of Northern Hemisphere glaciations. Paleoceanography, 26, .

Berger W.H., Lange C.B., & Wefer G. (2002) Upwelling History of the Benguela-Namibia System: A Synthesis of Leg 175 Results. Proceedings of the Ocean Drilling Program, Scientific Results (ed. by G. Wefer, W.H. Berger, and C. Richter), pp. 1–103.

Berger W.H., Wefer G., Richter C., Lange C.B., Giraudeau J., Hermelin O., & Shipboard Scientific Party (1998) The Angola-Benguela Upwelling System: Paleoceanographic Synthesis of Shipboard Results from Leg 175. Proceedings of the Ocean Drilling Program, Initial Reports (ed. by G. Wefer, W.H. Berger, C. Richter, and E. Al.), pp. 505–531.

Billups K. (2002) Late Miocene through early Pliocene deep water circulation and climate change viewed from the sub-Antarctic South Atlantic. Palaeogeography, Palaeoclimatology, Palaeoecology, 185, 287–307.

Billups K. & Schrag D.P. (2002) Paleotemperatures and ice volume of the past 27 Myr revisited with paired Mg/Ca and 18O/16O measurements on benthic foraminifera. Paleoceanography, 17, 1–11.

Bond W.J., Midgley G.F., & Woodward F.I. (2003) What controls South African vegetation — climate or fire? South African Journal Of Botany, 69, 79–91.

Bonnefille R. & Riollet G. (1980) Pollens des savanes d’Afrique orientale. CNRS Editions, Paris.

Bremner J.M. & Willis J.P. (1993) Mineralogy and geochemistry of the clay fraction of sediments from the Namibian continental margin and the adjacent hinterland. Marine Geology, 115, 85–116.

Brierley C.M., Fedorov A. V, Liu Z., Herbert T.D., Lawrence K.T., & LaRiviere J.P. (2009) Greatly Expanded Tropical Warm Pool and Weakened Hadley Circulation in the Early Pliocene. Science, 323, 1714–1718.

45

Miocene-Pliocene Vegetation change

Butzin M., Lohmann G., & Bickert T. (2011) Miocene ocean circulation inferred from marine carbon cycle modeling combined with benthic isotope records. Paleoceanography, 26, .

Cerling T.E., Harris J.M., Macfadden B.J., Leakey M.G., Quadek J., Eisenmann V., & Ehleringer J.R. (1997) Global vegetation change through the Miocene / Pliocene boundary. Nature, 153–158.

Collins J.A., Schefuß E., Heslop D., Mulitza S., Prange M., Zabel M., Tjallingii R., Dokken T.M., Huang E., Mackensen A., Schulz M., Tian J., Zarriess M., & Wefer G. (2010) Interhemispheric symmetry of the tropical African rainbelt over the past 23,000 years. Nature Geoscience, 4, 42–45.

Cowling R.M., Rundel P.W., Desmet P.G., & Esler K.J. (1998) Extraordinary High Regional-Scale Plant Diversity in Southern African Arid Lands: Subcontinental and Global Caomparisons. Diversity and Distributions, 4, 27–36.

Diester-Haass L., Meyers P.A., & Rothe P. (1990) Miocene history of the Benguela Current and Antarctic ice volumes: Evidence from rhythmic sedimentation and current growth across the Walvis ridge (Deep Sea Drilling project sites 362 and 532). Paleoceanography, 5, 685–707.

Dill H.G., Dohrmann R., Fenner J., Gersdorf U., Kringel R., Lindenmaier F., Ludwig R., Miller R.M., Nick A., Noell U., & Walzer A. (2012) Groundwater for the North of Namibia Project. Kalahari Research Project pp. 1–50. BGR,

Dupont L.M., Behling H., Jahns S., Marret F., & Kim J.-H. (2007) Variability in glacial and Holocene marine pollen records offshore from west southern Africa. Vegetation History and Archaeobotany, 16, 87–100.

Dupont L.M., Linder H.P., Rommerskirchen F., & Schefuß E. (2011) Climate-driven rampant speciation of the Cape flora. Journal of Biogeography, 38, 1059–1068.

Dupont L.M., Rommerskirchen F., Mollenhauer G., & Schefuß E. (2013) Miocene to Pliocene changes in South African hydrology and vegetation in relation to the expansion of C4 plants. Earth and Planetary Science Letters, 375, 408–417.

Dupont L.M. & Wyputta U. (2003) Reconstructing pathways of aeolian pollen transport to the marine sediments along the coastline of SW Africa. Quaternary Science Reviews, 22, 157–174.

Flower B.P. & Kennett J.P. (1994) The middle Miocence climatic transition: East Antarctic ice sheet development, depp ocean circulation and global carbon cycling. Palaeogeography, Palaeoclimatology, Palaeoecology, 108, 537–555.

Francois L., Ghislain M., Otto D., & Micheels A. (2006) Late Miocene vegetation reconstruction with the CARAIB model. Palaeogeography, Palaeoclimatology, Palaeoecology, 238, 302–320.

Gasse F., Chalié F., Vincens A., Williams M. a. J., & Williamson D. (2008) Climatic patterns in equatorial and southern Africa from 30,000 to 10,000 years ago reconstructed from terrestrial and near-shore proxy data. Quaternary Science Reviews, 27, 2316–2340.

Giess W. (1998) A preliminary Vegetation Map of South West Africa. Dinteria, 4, 5–112.

Gimeno L., Drumond A., Nieto R., Trigo R.M., & Stohl A. (2010) On the origin of continental precipitation. Geophysical Research Letters, 37, 1–7.

Heinrich S., Zonneveld K.A.F., Bickert T., & Willems H. (2011) The Benguela upwelling related to the Miocene

46

Chapter III

cooling events and the development of the Antarctic Circumpolar Current: Evidence from calcareous dinoflagellate cysts. Paleoceanography, 26, PA3209.

Hipondoka M. (2005) The Development and Evolution of Etosha Pan, Namibia. Julius-Maximilian-Universität Würzburg,

Leroux M. (1983) Le climat de L’Afrique tropicale (Texte and atlas). Champion, Paris.

Lunt D.J., Valdes P.J., Haywood A., & Rutt I.C. (2007) Closure of the Panama Seaway during the Pliocene: implications for climate and Northern Hemisphere glaciation. Climate Dynamics, 30, 1–18.

Maley J. (1970) Contributions à l’étude du Bassin tchadien: Atlas de pollens du Tchad. Bulletin du Jardin botanique national de Belgique,

Miller R.M., Pickford M., & Senut B. (2010) The Geology, Palaeontology and Evolution of the Etosha Pan, Namibia: Implications for terminal Kalahari depositions. South African Journal of Geology, 113, 307–334.

Nicholson S. (2000) The nature of rainfall variability over Africa on time scales of decades to millenia. Global and Planetary Change, 26, 137–158.

Petterson R.G. & Stramma L. (1991) Upper-level circulation in the South Atlantic Ocean. Progress in Oceanography, 26, 1–73.

Pickford M., Senut B., & Dauphin Y. (1995) Biostratigraphy of the Tsondab sandstone (Namibia) based on gigantic avian eggshells. Geobios, 28, 85–98.

Poore H.R., Samworth R., White N.J., Jones S.M., & McCave I.N. (2006) Neogene overflow of Northern Component Water at the Greenland-Scotland Ridge. Geochemistry Geophysics Geosystems, 7, Q06010.

Pound M.J., Haywood A.M., Salzmann U., & Riding J.B. (2012) Global vegetation dynamics and latitudinal temperature gradients during the Mid to Late Miocene (15.97–5.33Ma). Earth-Science Reviews, 112, 1–22.

Pound M.J., Haywood A.M., Salzmann U., Riding J.B., Lunt D.J., & Hunter S.J. (2011) A Tortonian (Late Miocene, 11.61–7.25Ma) global vegetation reconstruction. Palaeogeography, Palaeoclimatology, Palaeoecology, 300, 29–45.

Prange M. & Schulz M. (2004) A coastal upwelling seesaw in the Atlantic Ocean as a result of the closure of the Central American Seaway. Geophysical Research Letters, 31, L17207.

Rommerskirchen F., Condon T., Mollenhauer G., Dupont L., & Schefuß E. (2011) Miocene to Pliocene development of surface and subsurface temperatures in the Benguela Current system. Paleoceanography, 26, 1–15.

Scott L. (1982) Late Quaternary Fossil Pollen grains from the Transvaal, South Africa. Review of Palaeobotany and Palynology, 36, 241–278.

Ségalen L. (2004) Reconstitution of dune morphologies and palaeowind regimes in the Proto-Namib since the Miocene. Bulletin de la Société Géologique de France, 175, 537–546.

Ségalen L., Renard M., Lee-Thorp J.A., Emmanuel L., Le Callonnec L., de Rafélis M., Senut B., Pickford M., & Melice J.-L. (2006) Neogene climate change and emergence of C4 grasses in the Namib, southwestern Africa, as reflected in ratite 13C and 18O. Earth and Planetary Science Letters, 244, 725–734.

Senut B., Dauphin Y., & Pickford M. (1998) Nouveaux restes aviens du Néogène de la Sperrgebiet: complément à la biostratigraphie avienne des éolianites du désert de Namib. C. R. Acad. Sci. Paris, Sec. IIa, 327, 639–644.

47

Miocene-Pliocene Vegetation change

48

Senut B., Pickford M., & Ségalen L. (2009) Neogene desertification of Africa. Comptes Rendus Geoscience, 341, 591–602.

Shi N., Dupont L.M., Beug H.-J., & Schneider R. (1998) Vegetation and climate changes during the last 21000 years in S.W. Africa based on a marine pollen record. Vegetation History and Archaeobotany, 7, 127–140.

Siesser W.G. (1980) Late Miocene Origin of the Benguela Upswelling System off Northern Namibia. Science, 208, 283–5.

Strömberg C.A.E. (2011) Evolution of Grasses and Grassland Ecosystems. Annual Review of Earth and Planetary Sciences, 39, 517–544.

Tipple B.J. & Pagani M. (2007) The Early Origins of Terrestrial C4 Photosynthesis. Annual Review of Earth and Planetary Sciences, 35, 435–461.

Tyson P.D. & Preston-Whyte R.A. (2000) The Weather and Climate of Southern Africa. Oxford University Press South Africa, Cape Town.

Volk O.H. (1984) Pflanzenvergesellschaftungen mit Riccia-Arten in Südwestafrika (Namibia). Vegetatio, 55, 57–64.

Weltje G.J. (1997) End-member modeling of compositional data: Numerical-statistical algorithms for solving the

explicit mixing problem. Mathematical Geology, 29, 503–549.

White F. (1983) The Vegetation of Africa, Natural Resources Research. UNESCO, Paris.

Wright D., Miller G., & Fairbanks G. (1992) Early and Middle Miocene stable isotopes: Implications for Deepwater circulation and climate. Paleoceanography, 7, 357–389.

Ybert J.P. (1979) Altas des pollens de Côte d’Ivoire. Initiations Docu-mentation Technique 40, ORSTOM, Paris.

Zachos J., Pagani M., Sloan L., Thomas E., & Billups K. (2001) Trends, rhythms, and aberrations in global climate 65 Ma to present. Science, 292, 686–93.

Van Zinderen Bakker E.M. (1975) The Origin and Palaeoenvironment of the Namib Desert Biome. Journal of Biogeography, 2, 65–73.

Van Zinderen Bakker E.M. (1984) Palynological evidence for Late Cenozoic arid conditions along the Namibia coast from Holes 532 and 530A, Leg 75. Deep Sea Drilling Project Initial Reports of the Deep Sea Drilling Project, 75, 763–768.

Chapter III

Table 1: Identified pollen and spore types sorted into vegetation biomes of southern Africa after (White, 1983; Giess, 1998) and their score to the endmember (in percent). Riccia is sorted into semi-desert and shrubland savanna (Volk, 1984). In bold are the high scores. r² is the coefficient of determination. Family names after www.ThePlantList.org.

Family Pollen/Spore Type EM1 EM2 EM3 r2 Poaceae spp. 71.5 40.6 68.6 0.72

Cyperaceae spp. 3.92 9.07 0 0.46

Compositae Compositae Pentzia-type 0.4 0 0.29 0.13 Compositae Cotula-type 0 0.22 0.2 0.12 Compositae Dicoma-type 0.09 1.16 0.57 0.3 Compositae Gerbera-type 0.02 0.08 0 0.09 Compositae liguliflorae type 5.16 14.3 5.83 0.59 Compositae Vernonia-type 0.61 1.04 0.65 0.07 Compositae Senecio-type 0 0.28 0.63 0.31 Compositae tubuliflorae type 1.04 1.24 3.15 0.16

Woodland Acanthaceae spp. 0.62 1.57 0.71 0.23 Acanthaceae Peltalidium 0.1 1.68 0.28 0.47 Acanthaceae Ruellia 0 0.67 0.2 0.41

cf. Arecaceae trichothomosulcate-type (Dupont et al., 2011)

0 0.51 0.14 0.27

Leguminosae Brachystegia 0.08 0.65 0.11 0.1 Leguminosae Colophospermum mopane 1.63 1.89 0.93 0.13 Leguminosae Detarium 0 0.51 0.14 0.24 Malvaceae Grewia 0.1 0.34 0.07 0.11 Arecaceae Borassus 0.28 0.03 0.12 0.1 Malvaceae Dombeya 0 0.07 0.21 0.19 Malvaceae spp. 0.19 0.05 0.3 0.05

Desert vegetation Aizoaceae spp. 0.29 0.31 0.13 0.04 Geraniaceae Monsonia 0.02 0.56 0.34 0.13 Welwitschiaceae Welwitschia 0.01 0.1 0 0.14 Zygophyllaceae Tribulus 1.24 4.93 1.94 0.53 Amaranthaceae Chenopodiaceae spp. 1.64 0.26 1.17 0.12 Convolvulaceae Merremia 0.16 0 0 0.17

Mountain vegetation Compositae Artemisia-type 0.72 0.23 0.23 0.12 Compositae Berkehya-type 0.2 0 0.04 0.12 Anthocerotaceae Phaeoceros 0 0.21 0.06 0.11 Compositae Stoebe-type 0 0.4 0.04 0.13 Ericaceae spp. 0.24 0.78 0.66 0.08 Lycopodiaceae Lycopodium 0.89 1.76 1.28 0.1 Oleaceae spp. 0 0.11 0.02 0.06 Podocarpaceae Podocarpus 0 2.82 2.72 0.36 Schizaeaceae Mohria 0 0.64 0.21 0.4

49

Miocene-Pliocene Vegetation change

50

Thymelaeaceae Passerina 0.03 0.59 0 0.24 Thymelaeaceae spp. 0.68 1.13 0.04 0.14

Shrubland Acanthaceae Blepharis-type 0.1 0 0 0.16 Acanthaceae Justica-type 0.5 0.02 0.1 0.31 Euphorbiaceae Euphorbia 0.89 0 0 0.39 Leguminosae Acacia 0.88 0.53 0.29 0.09 Nyctaginaceae spp. 0.25 0.01 0 0.11 Ricciaceae Riccia 0.29 0 0.16 0.09 Neuradaceae spp. 0.01 0.58 0.22 0.29 Burseraceae Commiphora 0.03 0 0.18 0.12 Euphorbiaceae Phyllanthus 0 0.03 0.1 0.09 Proteaceae spp. 0 0.02 0.11 0.23

Aquatic vegetation Nymphaeaceae Nymphaea 0.23 0 0 0.22 Typhaceae Typha 0.52 0 0.15 0.17 Polygonaceae Polygonum 0 0.51 0.23 0.23

other spores Osmundaceae Osmunda-type 0 0.69 0.85 0.31 Selaginellaceae Selaginella-type 0.07 0.02 0.16 0.06 varia spore sp. 2 bacculate 0 0 0.16 0.37 varia spore sp. 3 leiotrilete, thick labra 0 0.14 0.43 0.28 Polypodiaceae spp. 0.08 0.15 0 0.11

Unsorted Convolvulaceae Anseia 0 0.32 0.19 0.06 Convolvulaceae Evolvulus 0 0.07 0.03 0.08 Malvaceae Abutilon 0.01 0.22 0 0.13 Rubiaceae Oxyanthus-type 0.01 0.14 0.02 0.06 Casuarinaceae Casuarina 0.12 0.04 0 0.05 Liliaceae spp. 0.05 0 0.05 0.05 Meliaceae spp. 0.37 0.15 0.08 0.08 Euphorbiaceae spp. 0.07 0.01 0.21 0.07

Not significant Acanthaceae Barleria 0.13 0.05 0.22 0.04 Arecaceae spp. 0.03 0.03 0 0.02 Compositae Ganzania-type 1.5 2.1 2.39 0.04 Myricaceae Myrica 0.04 0.12 0.04 0.02 Pteridaceae Pteris 0.07 0.02 0 0.04 Restionaceae spp. 0.77 0.43 0.66 0.04 Rosaceae Cliffortia 0.06 0.01 0.01 0.03 Rubiaceae Gardenia 0.12 0.09 0 0.04 varia tricolpate perforate pollen grain 0.01 0.07 0.05 0.01 spores indetermined 0.94 0.74 0.53 0.04

Chapter IV

Chapter IV: Miocene-Pliocene Stepwise

Intensification of the Benguela Upwelling over

the Walvis Ridge off Namibia

Sebastian Hoetzel1, Lydie Dupont1, Fabienne Marret 2& Gerold Wefer1

1 MARUM - Center for Marine Environmental Sciences, University of Bremen, Bremen,

Germany

2 School of Environmental Sciences, University of Liverpool, UK

4.1 Abstract

Upwelling is a significant part of the thermohaline circulation of the oceans controlling largely

the transport of cold waters to the surface and therefore influence global climate. The Benguela

Upwelling System (BUS) is one of the major upwelling areas in the world, however previous

reconstructions of the BUS mainly focused on the onset and intensification in southern and

central parts but changes of the northern part have been rarely investigated. We hereby present

an organic-walled dinoflagellate cyst record of ODP Site 1081 from the Late Miocene to the

Pliocene and use it to reconstruct and discuss the upwelling history on the Walvis Ridge with a

special focus on the movement of the Angola-Benguela Front (ABF). We show that during the

Late Miocene the Angola Current flowed over the Walvis Ridge more frequently than today

because the ABF was probably further southward located as a result of a weaker meridional

pressure gradient. A possible strengthening of the meridional pressure gradient during the latest

Miocene to early Pliocene intensified the upwelling along the coast and the upwelling’s filaments

over the Walvis Ridge. An intermediate period from 6.2 to 5.5 Ma is shown by the dominance

of Habibacysta tectata, cysts of a cool-tolerant dinoflagellate known from the northern Atlantic,

indicating changing oceanic conditions contemporaneous with the Messinian salinity crisis.

From 4.4 Ma on a stronger upwelling signal is recorded with well-mixed nutrient-rich waters.

Our results show a stepwise intensification of the BUS over a long period and a northward

migration of the ABF.

51

Stepwise Intensification of the Benguela Upwelling

4.2 Introduction

Today the Benguela Upwelling System

(BUS) is one of the major upwelling areas in

the world. Along the south-western coast of

Africa, cold and nutrient-rich waters well up

and provide best opportunities for high

primary production. The onset of this

system is supposed to be in the early Late

Miocene (~10-15 Ma) (Siesser, 1980;

Diester-Haass et al., 1990; Heinrich et al.,

2011; Rommerskirchen et al., 2011) during a

phase of global cooling. This cooling is part

of the transition of a warm and humid

variable climate to cooler and drier

conditions of the late Neogene driven by

changes of the Antarctic ice sheets (Wright

et al., 1992; Zachos et al., 2001; Billups &

Schrag, 2002).

Following the mid Miocene climatic

optimum, the re-establishment of major ice

sheets in East Antarctica at around 10 Ma

led to a steeper meridional temperature

gradient which strengthened climatic zones

and intensified pressure systems and ocean

circulations (Flower & Kennett, 1994;

Zachos et al., 2001). As a result, the South

Atlantic Anticyclone (SAA) became stronger

including also the trade winds which led to

increased productivity especially at around

6.5 Ma (Diester-Haass et al., 2002). The

further growth of the ice sheets led to a

northward migration of the polar front

causing also a shift of the BUS (Diester-

Haass, 1988; Diester-Haass et al., 1990,

1992). However, many studies focused

either in the southern and central parts of

the Benguela upwelling or only on the ocean

side of the Walvis Ridge (Siesser, 1980;

Diester-Haass et al., 1992; Marlow et al.,

2000; Udeze & Obohikuenobe, 2005;

Heinrich et al., 2011; Rommerskirchen et al.,

2011) and neglected the movement of the

boundary to the north, the Angola-Benguela

Front (ABF).

Here we present organic-walled

dinoflagellate cyst data covering the late

Miocene to Pliocene (9 Ma to 2.7 Ma). Our

focus of investigation is the Walvis Ridge

which divides the Cape Basin and the

Benguela Current from the Angola Basin

and the warm waters of the Angola Current.

Studies on dinoflagellate cyst distribution in

surface sediments along the West African

coast (e.g. Dale et al., 2002; Marret &

Zonneveld, 2003; Zonneveld et al., 2013)

allow good estimations of relative changes

of oceanic conditions, e.g. in terms of

temperatures, salinities and nutrients. In this

study we aim to describe the changes of the

northern parts of the BUS during times of

global cooling and discuss its forcing

mechanisms and influences.

4.2.1 The Benguela Upwelling System

Over the South East Atlantic the semi-

permanent high pressure system, the SAA, is

the dominating atmospheric feature located

at around 32°S 5°W (austral summer) and

27°S 10°W (austral winter) (Petterson &

Stramma, 1991). The SAA is also affecting

the Benguela Current (BC) which flows

northwards along the south-western coast of

Africa. The BC is derived from the eastward

flowing cold Antarctic Circumpolar Current

(ACC), which is linked to the polar front

and taking up parts of the warm Agulhas

52

Chapter IV

Current (AgC), which is flowing from the

Indian Ocean around South Africa (Fig. 4.1).

At around 28°S the BC is dividing into two

separate currents. One is following the

rotation of the SAA by turning west crossing

the Atlantic. The second current is flowing

further northwards along the south-western

African coast forming the Benguela Coastal

Current (BCC). Additionally, the BCC is

linked to the south-easterly trade winds. The

trade winds push the coastal surface waters

via Ekman-transport off shore inducing the

upwelling of cold and nutrient-rich

subsurface waters (e.g. Lutjeharms &

Meeuwis, 1987). The upwelled nutrients

allow a vast production of phytoplankton in

the photic zone. Eight upwelling cells have

been identified, of which the one west of

Lüderitz Bay is the strongest persisting all

year long resulting in the coldest surface

waters of the BUS (Lutjeharms & Meeuwis,

1987). Up to 600 km off shore cold

upwelled waters from the coastal upwelling

areas mix with surface waters and form

nutrient-rich filaments (Lutjeharms &

Meeuwis, 1987; Lutjeharms & Stockton,

1987; Summerhayes et al., 1995). Water of

these filaments has clearly enhanced nutrient

values, lower temperatures and increased

primary production.

Figure 4. 1: Map with location of studied ODP Site 1081 and discussed ODP Site 1085 within the South East Atlantic showing the mean annual temperatures in colours (Ocean Data View, U.S. NODC World Ocean Atlas 2009), oceanic features in black and grey (AC: Angola current, ABF: Angola-Benguela front, BC: Benguela current, BCC: Benguela Coastal current, ACC: Antarctic Circumpolar current, AgC: Agulhas current) and main atmospheric features in white dashed lines (SAA: South Atlantic Anticyclone, 1020 mbar; L(low): 1006 mbar). Mean atmospheric sea level pressure cells of January after (Petterson & Stramma, 1991).

4.2.2 The Angola-Benguela Front

The BCC meets the southward-flowing

warm and nutrient-poor Angola Current

just north of the Walvis Ridge. Together

they form the ABF which lays today

between 14°S and 16°S (Meeuwis &

Lutjeharms, 1990), depending on the

season. In detail, it is not a single front but

53

Stepwise Intensification of the Benguela Upwelling

a couple of fronts arranged in two frontal

zones, a northern and a southern, whereas

the latter one is the one with a higher

influence on the overall ABF’s

characteristics (Kostianoy & Lutjeharms,

1999). Like the coastal upwelling, the

position of the front depends on the SAA;

when the meridional pressure gradient is

high, the southern front of the ABF is

located further north and the ABF is

narrower and sharper (Kostianoy &

Lutjeharms, 1999). Under weakened SAA

conditions and resulting weakened winds,

the ABF can be located further to the south

(Richter et al., 2010) so that warm nutrient-

poor water of the Angola current can

penetrate further southward, up to 24°S

along the Namibian coast (Meeuwis &

Lutjeharms, 1990). As a result, the

precipitation on the adjacent coast is

enhanced (Hirst & Hastenrath, 1983;

Nicholson & Entekhabi, 1987) and can be

increased farther inland (Rouault, 2003).

The phenomenon occurs on inter-annual

timescales and is called Benguela Niño for

its similarity to the El Niño Southern

Oscillation (Shannon et al., 1986).

4.3 Material and methods

The sampled sediment core was retrieved at

the Ocean Drilling Program Site 1081

(19°37.1818’S 11°19.1598’E) in 794.1 m

water depth. The site is located on the

Walvis Ridge 160 km off the Namibian

coast and is today influenced by filaments

of the BUS. The sediment is composed of

olive-gray clayey nannofossil ooze and

olive-gray to black clays (Berger et al.,

2002b). Sedimentation rates were calculated

between 2 and 5 cm/ka using an age model

based on biostratigraphy, magnetic reversals

and magnetic susceptibility (Berger et al.,

2002b).

Seventy-one samples were taken with ages

between 9 and 2.7 Ma and prepared for

palynological investigation. The volume of

each sample was estimated by water

displacement. For decalcification diluted

cold HCl (~5%) was used and afterwards

the material was treated with cold HF

(~20%) for two days to remove silicates.

Prior to the HCl-treatment Lycopodium

tablets were added. The residuals were

sieved under ultrasonic treatment removing

particles smaller than 10-15 µm. The

cleaned residuals were stored in water and

mounted in glycerol for investigation under

a light microscope using magnifications of

400x, 600x and 1000x. For each sample at

least 300 dinoflagellate cysts were identified

and counted.

Identification of dinoflagellates was done

after de Verteuil & Norris, 1992; Marret &

Zonneveld, 2003; De Schepper & Head,

2009; Schreck et al., 2012. Brigantedinium

spp. includes all round, brown smooth

dinoflagellate cysts. Because

Batiacasphaera micropapillata has a strong

morphological overlap with B. minuta and

both occur in the samples we adopted the

nomenclature of de Schepper & Head

(2008) and Schreck et al. (2012) in treating

both species as one complex.

Each species was sorted ecologically after

its assumed metabolism mechanism

(autotroph and heterotroph, Tab. 2, at the

54

Chapter IV

end of this chapter) and a

heterotroph/autotroph ratio (H/A) was

calculated. Percentages were calculated on

the total number of dinoflagellate cysts

counted. For calculation of the 95 %

confidence intervals of the relative

abundances the equation of (Maher, 1971)

was used. Accumulation rates of

dinoflagellate cysts were calculated by

multiplying the sedimentation rates (Berger

et al., 2002b) with the cyst concentrations

per ml, which was calculated based on the

known number of Lycopodium spores added

in the form of tablets.

4.4 Results

A total of 36 species were identified

whereof the Brigantedinium spp. group is the

most dominant one. In Figures 3.2 and 3.3

the percentages and accumulation rates of

the most abundant species are shown.

Figure 3. 2: Relative and absolute occurrences of selected dinoflagellate cysts within zonation scheme (grey boxes) used in the text. The percentages are shown by lines, shadings represent 95% confidence intervals. Accumulation rates are shown by the filled-in graph.

55

Stepwise Intensification of the Benguela Upwelling

Figure 4. 3: See captures Fig. 4.2

In general, both relative and absolute trends

show similarity for each species. The record

has been visually divided into 5 zones. one I

runs from the start of the record at 9 Ma to

7.8 Ma. This zone is mainly characterized by

cysts of the Batiacasphaera micropapillata

complex with values mostly around 20 %

reaching a maximum of 50 %. Impagidinium

paradoxum is well represented (generally over

10 %) especially in the beginning with a peak

over 20 %. Lingulodinium machaerophorum is

present in all samples of Zone I, with two

exceptions, reaches values up to 20 % at

around 8.5 Ma, and decreases afterwards to

a minor representation in the assemblages.

Cysts of Nematosphaeropsis labyrinthus reach up

to maxima of 15 %. Impagidinium sp. 2 (De

Schepper & Head, 2009) cysts are also

present and make up around 5 % of the

assemblage and have a maximum of 20 % at

around 8.3 Ma. Selenopemphix quanta is only

marginally represented in Zone I.

Brigantedinium spp. have generally low values

(around 5 %) that rise at the end of Zone I

to around 15 %. Almost completely lacking

in this interval are Operculodinium centrocarpum

cysts.

Zone II (7.8 – 6.2 Ma) is marked by a

decline of cysts abundance of the B.

micropapillata complex to less than 10 % (and

56

Chapter IV

less than 5 % around 7 Ma). Additionally the

absolute values are decreasing from around

3000 cysts per cm² per ka to less than 1000.

Values for Brigantedinium spp. and S. quanta

cysts are increased in this zone but show

decreasing trends towards the end of the

Zone II at 6.5 Ma. Also H. tectata decreases

from around 20 % to marginal occurrence at

6.5 Ma. However, an increasing trend shows

the representation of B. hirsuta from low

values of less than 10 % to more than 30 %.

As in Zone I L. machaerophorum is mostly

represented with less than 10 % but peaks at

~7.5 Ma with 63 % of the assemblage.

Again O. centrocarpum is lacking.

The next part, Zones III and IV, starts at 6.2

Ma and ends at 4.4 Ma and is characterized

by H. tectata cysts in Zone III until 5.8 Ma

and B. hirsuta and Brigantedinium spp. in Zone

IV. H. tectata dominates the samples in the

first period and reaches values between 40

and 50 %. Additionally, the S. quanta curve

increases again until 5 Ma. Spiniferites spp.

and B. hirsuta values are also increasing

between 6 and 5 Ma. Brigantedinium has low

relative abundances until 5.2 but reaches 40

% later in Zone IV. Not represented is L.

machaeorophorum. The B. micropapillata

complex is better represented in the period

from 5.2 until 4.2 Ma with maxima around

25 %. At around 4.8 Ma Impagidinium sp. 2 is

well represented and peaks twice at values

between 20 and 30 %.

The last Zone (V) is defined by the

representation of O. centrocarpum which is

rising from a few percent at the start of the

interval to 62 % at around 7.5 Ma. Towards

the end Brigantedinium spp. and Spiniferites

spp. cysts are increasing on the costs of O.

centrocarpum.

In Figure 4.4 the accumulation rates of

summed dinoflagellate cysts are plotted

against age. The accumulation of cysts range

between 6.600 and 113.000 per cm² per ka

showing two periods with maximum rates at

8-6.5 Ma and 5.4-3.5 Ma.

Figure 4. 4: Accumulation rates of all dinoflagellate cysts. Zonation follows Fig. 4.2.

57

Stepwise Intensification of the Benguela Upwelling

4.5 Discussion

4.5.1 Dinoflagellates as proxies for

environmental conditions

Off the 2377 modern motile species listed

worldwide (Gómez, 2012) only between 10

and 20 % of them produce during their

lifecycle a cyst that preserved well in the

sediments (Head, 1996). The organic-walled

cysts found in recent sediments around the

world show a distribution that is related to

sea-surface gradients in temperature, salinity,

nutrients and sea-ice cover (de Vernal et al.,

1997; Zonneveld et al., 2001, 2013; Dale et

al., 2002; Marret & Zonneveld, 2003;

Holzwarth et al., 2007). Therefore, it has

been possible to use them as a tool for

reconstructing surface waters conditions, e.g.

nutrients and temperature, enabling past

upwelling intensity variations. The

dissolution of sensitive species such as

Brigantedinium can not entirely be dismissed

here but the relative high abundance of this

taxa overall in the record, as well as the high

total concentration values (Fig. 4.4) suggest a

minor effect. Therefore, we are confident to

use the heterotrophic versus autotrophic

ratio (H/A ratio, Fig. 4.5). The H/A ratio

indicates upwelling (Lewis et al., 1990) based

on the fact that the autotrophic

dinoflagellates are in strong competition to

diatoms whereas the heterotrophic feed on

diatoms and hence, a high H/A ratio reflects

higher production. In our record the ratio

(Fig. 4.5) is in good concordance to the total

dinoflagellate accumulation (Fig. 4.4). Our

results indicate four different regimes

(discussed below), a weak upwelling regime

with a stronger influence of warm waters

from the Angola Current (Zone I), an

increase of upwelling linked to the

intensification of the meridional pressure

gradient (Zone II), a period with exceptional

conditions during the Mediterranean Salinity

Crisis affecting the Atlantic overturning

circulation (Zone III and IV) and a further

intensification of the upwelling together

with the effect of the Cunene River

discharge (Zone V).

4.5.2 The Angola-Benguela Front

ODP site 1081 is located close to the

modern position of the ABF, which offers

opportunities to record changes in its

position since the Miocene. The

composition of the dinoflagellate cyst

assemblages shows affinity to upwelling

conditions over the entire record through

nutrient indicating cyst species such as

Brigantedinium spp., S. quanta, L.

machaerophorum which is in accordance to an

onset of the BUS before 10 Ma (Siesser,

1980; Diester-Haass et al., 1990; Heinrich et

al., 2011; Rommerskirchen et al., 2011).

However, fluctuations in accumulation rates

and relative abundances, such as the

presence or absence of L. machaeophorum and

O. centrocarpum, indicate significant variability

in the upwelling during the studied period.

Open oceanic species,

Impagidinium paradoxum (Dale et al., 2002;

Zonneveld et al., 2013) and probably

Impagidinium sp. 2 are stronger represented in

Zone I and II (9 - 6.2 Ma) indicating warmer

and less nutrient-rich environments. In this

zone the B. micropapillata complex is well

represented. Unfortunately, the ecological

58

Chapter IV

59

background of the B. micropapillata complex

is uncertain. However, Zegarra & Helenes

(2011) suggest this complex to be an

indicator for warm nutrient-poor waters

because it is often recorded together with

other warm-water indicators. Hence, we

consider B. micropapillata as an indicator of

relative warm nutrient-poor conditions.

These conditions during the early part of the

record might be interpreted as weaker

upwelling and stronger influence of warm

waters from the AC compared to later

periods.

Prior to 8 Ma, sub-Antarctic δ18O values

indicate slightly warmer temperatures and

less intense glaciations in Antarctica (Billups,

2002). The polar front was, therefore,

weaker and situated more southwards and so

was the SAA. Additionally, glaciations in the

Northern Hemisphere were not yet

extensive so that the meteorological equator

might have been located further north than

today creating a weak meridional

temperature and pressure gradient in the

Southern Hemisphere (Flohn, 1981). A weak

meridional temperature gradient would shift

the ABF southwards and allow the AC to

penetrate southwards over the Walvis Ridge

which is suggested by the presence of L.

machaerophorum.

Today L. machaerophorum is not present in the

BUS but in the area of the AC (Holzwarth et

al., 2007). Additionally L. machaeorphorum is

an indicator for stratified warm nutrient-rich

waters, e.g. after upwelling relaxation

(Marret & Zonneveld, 2003; Zonneveld et

al., 2013; and references therein) suggesting

that periods with stratified water conditions

were more frequent and/or longer than

today. In addition, today I. paradoxum and N.

labyrinthus are more frequent in waters

around the ABF (Dale et al., 2002;

Zonneveld et al., 2013) and may indicate

here a relation with the ABF. Today, the

ABF shifts southwards on inter-annual

timescales during Benguela Niños. Under

these special conditions, the waters over the

Walvis Ridge receive a contribution of the

Angola Current (Petterson & Stramma,

1991; Florenchie et al., 2004; Mohrholz et

al., 2004; Richter et al., 2010). It is possible

that these Benguela Niño events were more

common during Miocene times when the

SAA and the trade winds were weaker.

Richter et al. (2010) showed that sea-surface

temperatures along the south-western

African coast respond quickly to changes in

the meridional wind anomalies. In detail,

they showed that a weakening of the

subtropical anticyclone led to warmer

temperatures especially at 20°S.

4.5.3 Upwelling intensification and

meridional pressure gradient

Zone II (7.8 – 6.2 Ma) shows higher overall

dinoflagellate cyst accumulation rates

indicating higher productivity as a result of

stronger upwelling (Fig. 4.4). Further

support of the intensified upwelling is given

by increased heterotrophic dinoflagellate

cyst occurrences, such as Selenopemphix and

Brigantedinium (Zonneveld et al., 2001),

which is also shown in the increased H/A

ratio (Fig. 4.5).

Stepwise Intensification of the Benguela Upwelling

Figure 4. 5: Ratio between heterotroph and autotroph dinoflagellate cysts. Zonation follows Fig. 4.2.

In general, a gradual intensification of the

BUS is suggested by other authors, e.g.

Rommerskirchen et al. (2011) who showed

for ODP Site 1085 a continuously sea-

surface cooling until 6 Ma. However, at the

Walvis Ridge during this interval, periods of

weaker upwelling and stronger influence of

warmer waters may have occurred allowing

the presence and even the dominance at 7.5

Ma of L. machaerophorum. Again, it is likely

that the ABF was weaker and located

further to the south and that the influence

of the AC was stronger due to more

frequently occurring Benguela Niño

conditions. However, the influence of the

AC seems to disappear towards the end of

Zone II indicated by the decrease and low

occurrences of L. machaeorophorum and

oceanic species such as I. paradoxum and the

B. micropapillata complex. In contrast to the

position of the ABF during the previous

warm period, a northward shift and

intensification of the ABF with the

restriction of the AC as well as open ocean

conditions can explain this change.

The growth of Antarctic ice sheets or the

further global cooling (Billups, 2002) could

have caused an increased meridional

temperature and pressure gradient which

resulted in stronger currents and winds.

The strengthened winds, especially the

trade winds would lead to an aridification

of the south-western coast of Africa since

warm and humid air from the Atlantic

would be hindered to reach large areas of

Namibia, like it is today. The expansion of

the grass savanna that took place between

the Miocene and Pliocene, is often linked to

aridification due to the water use efficiency

of C4 savanna grasses (Pagani et al., 2005;

Ségalen et al., 2006). Additionally Dupont

et al. (2013) discuss, based on pollen data

and deuterium values of plant waxes from

ODP Site 1085, that the aridification is

caused by a decrease of precipitation

derived from the Atlantic and that the main

source area of precipitation changed to the

Indian Ocean, which is the case today

(Gimeno et al., 2010). Further support for

the northward shift of the ABF is given by

Diester-Haass et al. (1992) who argue that

the BUS migrated northwards during the

Miocene.

4.5.4 Exceptional conditions related to

the Mediterranean Salinity Crisis

Zone III (6.2 - 5.5 Ma) is characterized by

the dominance of H. tectata which is known

60

Chapter IV

to be a cool-water tolerant species mostly

of the Pliocene and Pleistocene of the

North Atlantic region (Head, 1994;

Versteegh, 1997; Louwye et al., 2004) but

which is also recorded in the Miocene

(Head et al., 1989) and the Mediterranean

(Jimenez-Moreno et al., 2006). De Schepper

et al. (2011) showed that if H. tectata

exceeded 30 % of the dinoflagellate cyst

assemblage, then the reconstructed Mg/Ca

sea-surface temperatures were between 10

and 15 °C. Although H. tectata has not yet

been described from the Southern Atlantic,

it probably represents strong surface water

cooling and therefore intensified upwelling.

According to Vidal et al. (2002), the

sedimentation rate at ODP Site 1085 is

increased during that period (6.1 – 5.8 Ma)

which they correlate to the occurrences of

strong glacial events. As a consequence,

meridional pressure gradients steepened

and upwelling further intensified.

Additionally, S. quanta is also increased

during that period, indicating a higher

nutrient supply supporting the

interpretation of increased upwelling.

Although the H/A ratio is lower than

before it still suggests upwelling. Just before

(end of Zone II, 6.8-6.4 Ma) and directly

after the maximum of H. tectata (Zone III-

IV, 5.6-5.2 Ma), the occurrence of B. hirsuta

increased. However, the ecological

background of B. hirsuta is not well known.

Both, H. tectata and B. hirsuta have not been

yet described in the BUS area but they

alternatively dominate the assemblages

from 6.8 to 5.2 Ma, representing unique

conditions which did not exist before or

after that period.

The representations of H. tectata and B.

hirsuta fall in a period of global oceanic and

climate changes and occur

contemporaneously to the initiation of the

North Atlantic Deepwater (NADW)

formation, the desiccation of the

Mediterranean Sea, and the loss of contact

between the deep waters of the Pacific and

Atlantic Oceans by the decreasing sill depth

of the Central American Seaway (CAS).

According to Billups (2002), the closure of

the CAS had already influenced the ocean

currents between 6.6 and 6 Ma by

increasing the Atlantic overturning

circulation. The data sets show a cooling of

southern ocean upper circum-polar and

intermediate waters during the latest

Miocene. The influence of the closing of

the Panama isthmus on the Atlantic

overturning circulation is furthermore

underlined by models such as the one used

by Butzin et al. (2011) who conclude that

the NADW formation began when CAS

had shoaled to a few hundred meters

during the Late Miocene. Poore et al.

(2006) calculated the proportion of and

describe a dramatic increase of NADW

formation as well as a stronger Atlantic

overturning circulation for the period

between 6 and 2 Ma.

Additionally, shortly before 6.2 Ma, the

Mediterranean begun to desiccate, creating

salty and dense waters flowing into the

Atlantic and contributing to NADW

formation (Pérez-Asensio et al., 2012).

Hence, the overturning was further

61

Stepwise Intensification of the Benguela Upwelling

increased and might have changed the

overall oceanic conditions or at least might

have had an impact on the upwelled waters

of the BUS allowing the increased

occurrence of B. hirsuta. Berger et al.

(2002a) argued that the quality of the

coastal upwelling waters is linked to the

deep water circulation and the strength of

the NADW. Increasing NADW brings,

until a critical point, more silica from the

Northern Ocean into the Southern Ocean

(fire-hose effect) and so changes the

chemistry of the upwelled waters at the

BUS (e.g. enabling the so-called Matuyama

Diatom Maximum). Changes of the silica

content might have affected dinoflagellates

indirectly via ecological competition with

algae or other microorganisms or via

changes in food supply and sources. The

latter one is important for the heterotrophic

species since they feed on diatoms which

are highly depending on silica contents.

Furthermore, Berger et al. (2002a) argued

that the change of subsurface waters

through time was caused by vertical mixing

of the NADW around Antarctica

preventing higher opal concentrations off

South Africa and so explaining the low

diatom concentrations in the BUS during

the Pleistocene. The quality of upwelled

waters is therefore variable and changes in

correspondence to the NADW, for

instance, by varying the silica content.

For the period from 6-5.5 Ma,

Rommerskirchen et al. (2011) described

Tex’86 temperature estimates from ODP

Site 1085 indicating warming of subsurface

waters. They concluded that the subsurface

water warming resulted from a downward

mixing of heat caused by the cessation of

the Mediterranean outflow which weakened

the Atlantic overturning circulation and

reduced the NADW formation. This period

of warm subsurface waters at Site 1085

coincides with the increased representations

of H. tectata so that the downward mixing

might have had also affected the quality of

upwelled waters at the ODP Site 1081

creating special and non-recurring

conditions.

A further change of the quality of the

upwelled waters could be caused by a

poleward undercurrent flowing south from

the Angola dome along the African margin

transporting silica-rich, phosphate-rich and

oxygen depleted waters (Berger et al., 1998)

and increasing the fertile thermocline. A

strengthening of this undercurrent and a

higher silica content of the waters

representing a mixing of the intermediate

water and this poleward undercurrent is

indicated by a radiolarian peak from 5.8 Ma

until 5.25 Ma at ODP site 1085 (Diester-

Haass et al., 2002). The peak in radiolarian

could also explain a missing peak in the

H/A ratio since radiolarian might have

competed with heterotrophic

dinoflagellates.

Despite the warmer subsurface waters,

further upwelling and cooling of the sea-

surface in the southern BUS during that

period is recorded by Rommerskirchen et

al. (2011) at ODP Site 1085. After 5.5 Ma,

the Gibraltar strait opened and again very

salty and dense water re-intensified the

NADW formation and the overturning.

62

Chapter IV

The intensified overturning might have had

an impact on the Hadley circulation and

enforced trade winds so that the upwelling

got stronger indicated by a sharp increase

of Brigantedinium spp. at around 5.3 Ma

which is in correspondence to increased

sedimentation rates at ODP Site 1085

(Diester-Haass et al., 2002). Additionally,

the circulation overturning could have been

enhanced because of relatively high sea

level (Haq et al., 1987). The representation

of the open oceanic species B. micropapillata

and Impagidinium sp. 2, however, suggest

that the upwelling was intermittently less

intense between 5.3 to 4.4 Ma. In between,

the upwelling intensity must have been

strong since the accumulation rates show

the highest values of the record during that

period (Fig. 4.4). Additionally, a strongly

increased primary production is also

indicated by the high H/A ratio (Fig. 4.5).

4.5.5 Upwelling intensification and river

supply

The presence and dominance of O.

centrocarpum in the last zone (from 4.4 Ma

on) indicate nutrient-rich and well mixed

waters representing conditions adjacent to

strong upwelling and/or river outflow

(Dale et al., 2002). Although O. centrocarpum

is a cosmopolitan species, it occurs in the

South East Atlantic in vicinity to the

upwelling area (Dale et al., 2002; Marret &

Zonneveld, 2003; Holzwarth et al., 2007;

Zonneveld et al., 2013). It is, additionally,

represented in turbulent and well mixed

waters at the boundary of coastal and

oceanic waters (Wall et al., 1977; Dale et al.,

2002). L. machaeorophorum is, however,

completely absent, indicating that the partly

warm stratified conditions of the Miocene

have been completely replaced by stronger

upwelling, better mixing, and cooler

conditions. The increase of Brigantedinium

spp. abundance and the still high H/A

ratios underline nutrient rich conditions of

strong upwelling.

During that time (4.4 Ma), further closure

of the CAS and the emergence of the

Panama Isthmus proceeded and led to the

intensification of the overturning

circulation and of the deep waters (Haug &

Tiedemann, 1998) changing the

intermediate waters of the BUS. From 3.2

Ma on, after warming of the Northern

Hemisphere, the intensification of the

Northern Hemisphere glaciation and the

introduction to a bi-polar icehouse started

causing a southward-shift of the Inter-

Tropical Convergence Zone (Billups et al.,

1999). The southward shift caused an

intensified meridional pressure gradient and

therefore an intensification of the trade-

winds as well as linked upwellings. The

dinoflagellate cyst flux decreases while the

H/A ratio increases indicating again a

change of the quality of upwelled waters.

That is in correspondence to a shift of

diatom assemblages at further south located

parts of the BUS during that time (Marlow

et al., 2000). These changes can be linked to

a stronger influence of Antarctic

Intermediate Water.

The increase of Spiniferites at the end of the

record might be linked to more coastal

conditions and river input (Dale et al.,

63

Stepwise Intensification of the Benguela Upwelling

2002). An increased river discharge could

also explain relative high occurrences of O.

centrocarpum. The desiccation and outflow of

the Cunene lake through the Cunene river

is supposed to have started in the early

Pliocene (Hipondoka, 2005).

4.6 Summary and conclusions

A record of organic-walled dinoflagellate

cysts was used to reconstruct upwelling

conditions over the Walvis Ridge from the

Late Miocene to the early Pliocene in

respect to its northern boundary the ABF.

Upwelling related conditions existed over

the entire studied period. A weak pressure

system and a further south located ABF

resulted in more frequently occurring

Benguela Niño-conditions, which were

recorded for the time before 7.8 Ma. The

meridional pressure gradient steepened

afterwards, with a northward migration of

the ABF. The Benguela Niño-conditions

were especially manifested by the

occurrence of L. machaerophorum, a species

blooming in warm stratified nutrient-rich

waters in particular after upwelling

relaxation. The intensification of the

upwelling is shown by decreases of warm

water indicating taxa and increases of cold

and nutrient-rich indicators. Production

was high until around 7 Ma and the portion

of heterotrophic species was enhanced.

Between 6.8 and 5.2 Ma, B. hirsuta and H.

tectata were abundant during a period with

exceptional oceanic conditions related to

the Messinian Salinity Crisis. B. hirsuta

occurred during times with enhanced

NADW production and stronger Atlantic

overturning circulation whereas H. tectata

occurred during a time of reduced NADW

production due to desiccation of the

Mediterranean Basin. The shoaling of the

CAS and later the intensification of the

Northern Hemisphere glaciations enhanced

the NADW production, the Hadley

circulation and the upwelling in the

Benguela area producing well mixed

conditions over the Walvis Ridge from 4.4

Ma on. Remote influence of fresh water

after the start of Cunene River discharge

into the South Atlantic is indicated by the

first occurrence of O. centrocarpum.

4.7 Acknowledgements

This work was supported by the DFG

Research Center/Cluster of Excellence

‘MARUM - The Ocean in the Earth

System’ and ‘GLOMAR – Bremen

International Graduate School for Marine

Sciences’. The authors thank Michael

Schreck and Jens Matthiesen for helping to

identify unknown dinoflagellates. Marta

Witek is thanked for laboratory support.

4.8 References

Berger W.H., Lange C.B., & Pérez M.E. (2002a) The early Matuyama Diatom Maximum off SW Africa: a conceptual model. Marine Geology, 180, 105–116.

Berger W.H., Lange C.B., & Wefer G. (2002b) Upwelling History of the Benguela-Namibia System: A Synthesis of Leg 175 Results. Proceedings of the Ocean Drilling Program, Scientific Results (ed. by G. Wefer, W.H. Berger, and C. Richter), pp. 1–103.

64

Chapter IV

Berger W.H., Wefer G., Richter C., Lange C.B., Giraudeau J., Hermelin O., & Shipboard Scientific Party (1998) The Angola-Benguela Upwelling System: Paleoceanographic Synthesis of Shipboard Results from Leg 175. Proceedings of the Ocean Drilling Program, Initial Reports (ed. by G. Wefer, W.H. Berger, C. Richter, and E. Al.), pp. 505–531.

Billups K. (2002) Late Miocene through early Pliocene deep water circulation and climate change viewed from the sub-Antarctic South Atlantic. Palaeogeography, Palaeoclimatology, Palaeoecology, 185, 287–307.

Billups K., Ravelo A.C., Zachos J.C., & Norris R.D. (1999) Link between oceanic heat transport, thermohaline circulation, and the Intertropical Convergence Zone in the early Pliocene Atlantic. Geology, 27, 319–322.

Billups K. & Schrag D.P. (2002) Paleotemperatures and ice volume of the past 27 Myr revisited with paired Mg/Ca and 18O/16O measurements on benthic foraminifera. Paleoceanography, 17, 1–11.

Butzin M., Lohmann G., & Bickert T. (2011) Miocene ocean circulation inferred from marine carbon cycle modeling combined with benthic isotope records. Paleoceanography, 26, .

Dale B., Dale A.L., & Jansen J.H.F. (2002) Dinoflagellate cysts as environmental indicators in surface sediments from the Congo deep-sea fan and adjacent regions. Palaeogeography, Palaeoclimatology, Palaeoecology, 185, 309–338.

Diester-Haass L. (1988) Sea level changes, carbonate dissolution and history of the Benguela Current in the Oligocene-Miocene off Southwest Africa (DSDP Site 362, Leg 40). Marine Geology, 79, 213–242.

Diester-Haass L., Meyers P.A., & Rothe P. (1990) Miocene history of the Benguela Current and Antarctic ice volumes: Evidence from rhythmic sedimentation and current growth across the Walvis ridge (Deep Sea Drilling project sites 362 and 532). Paleoceanography, 5, 685–707.

Diester-Haass L., Meyers P.A., & Rothe P. (1992) The Benguela Current and associated upwelling on the southwest African Margin: a synthesis of the Neogene-Quaternary sedimentary record at DSDP sites 362 and 532. Geological Society, London, Special Publications, 64, 331–342.

Diester-Haass L., Meyers P.A., & Vidal L. (2002) The late Miocene onset of high productivity in the Benguela Current upwelling system as part of a global pattern. Marine Geology, 180, 87–103.

Dupont L.M., Rommerskirchen F., Mollenhauer G., & Schefuß E. (2013) Miocene to Pliocene changes in South African hydrology and vegetation in relation to the expansion of C4 plants. Earth and Planetary Science Letters, 375, 408–417.

Flohn H. (1981) A hemispheric circulation asymmetry during Late Tertiar. Geologische Rundschau, 70, 725–736.

Florenchie P., Reason C.J.C., Lutjeharms J.R.E., Rouault M., Roy C., & Masson S. (2004) Evolution of Interannual Warm and Cold Events in the Southeast Atlantic Ocean. Journal of Climate, 17, 2318–2334.

Flower B.P. & Kennett J.P. (1994) The middle Miocence climatic transition: East Antarctic ice sheet development, depp ocean circulation and global carbon cycling. Palaeogeography, Palaeoclimatology, Palaeoecology, 108, 537–555.

Gimeno L., Drumond A., Nieto R., Trigo R.M., & Stohl A. (2010) On the origin of continental precipitation. Geophysical Research Letters, 37, 1–7.

65

Stepwise Intensification of the Benguela Upwelling

Gómez F. (2012) A checklist and classification of living Dinoflagellates (Dinoflagellata, Alveolata). CICIMAR Oceánides, 27, 65–140.

Haq B.U., Hardenbol J., & Vail P.R. (1987) Chronology of Fluctuating Sea Levels since the Triassic. Science, 235, 1156–67.

Haug G.H. & Tiedemann R. (1998) Effect of the formation of the Isthmus of Panama on Atlantic Ocean thermohaline circulation. Nature, 393, 673–676.

Head M.J. (1994) Morphology and Paleoenvironmnental Significance of the Cenozoic Dinoflagellate Genera Tectatodinium and Habibacysta. Micropaleontology, 40, 289–321.

Head M.J. (1996) Modern dinoflagellate cysts and their biological affinities. Palynology: principles and applications (ed. by J. Jansonius and D.C. McGregor), pp. 1197–1248. American Association of Stratigraphic Palynologists Foundation, Salt Lake City.

Head M.J., Norris G., & Mudie P.J. (1989) Palynology and Dinocyst Stratigraphy of the Miocene in ODP Leg 105, Hole 645E, Baffin Bay. Proceedings of the Ocean Drilling Program, Scientific Results, 105, 467–514.

Heinrich S., Zonneveld K.A.F., Bickert T., & Willems H. (2011) The Benguela upwelling related to the Miocene cooling events and the development of the Antarctic Circumpolar Current: Evidence from calcareous dinoflagellate cysts. Paleoceanography, 26, PA3209.

Hipondoka M. (2005) The Development and Evolution of Etosha Pan, Namibia. Julius-Maximilian-Universität Würzburg,

Hirst A.C. & Hastenrath S. (1983) Atmosphere-Ocean Mechanisms of Climate Anomalies in the Angola-Tropical Atlantic Sector. Journal of Physical Oceanography, 13, 1146–1157.

Holzwarth U., Esper O., & Zonneveld K. (2007) Distribution of organic-walled dinoflagellate cysts in shelf surface sediments of the Benguela upwelling system in relationship to environmental conditions. Marine Micropaleontology, 64, 91–119.

Jimenez-Moreno G., Head M.J., & Harzhauser M. (2006) Early and Middle Miocene dinoflagellate cyst stratigraphy of the Central Paratethys, Central Europe. Journal of Micropalaeontology, 25, 113–139.

Kostianoy A.G. & Lutjeharms J.R.E. (1999) Atmospheric effects in the Angola-Benguela frontal zone. Journal of Geophysical Research, 104, 20963–20970.

Lewis J., Dodge J.D., & Powell A.J. (1990) 18 . QUATERNARY DINOFLAGELLATE CYSTS FROM THE UPWELLING SYSTEM OFFSHORE PERU, HOLE686B, ODP LEG 112. Proceedings of the Ocean Drilling Program, Scientific Results (ed. by E. Suess and R. von Huene), pp. 323–328.

Louwye S., Head M.J., & De Schepper S. (2004) Dinoflagellate cyst stratigraphy and palaeoecology of the Pliocene in northern Belgium, southern North Sea Basin. Geological Magazine, 141, 353–378.

Lutjeharms J.R.E. & Meeuwis J.M. (1987) The extent and variability of South-East Atlantic upwelling. South African Journal of Marine Science, 5, 51–62.

Lutjeharms J.R.E. & Stockton P.L. (1987) Kinematics of the upwelling front off southern Africa. South African Journal of Marine Science, 5, 35–49.

Maher L.J.J. (1971) Nomograms for computing 0.95 confidence limits of pollen data. Review of Palaeobotany and Palynology, 13, 85–93.

Marlow J.R., Lange C.B., Wefer G., & Rosell-Mele A. (2000) Upwelling intensification as part of the Pliocene-

66

Chapter IV

Pleistocene climate transition. Science, 290, 2288–91.

Marret F. & Zonneveld K.A.F. (2003) Atlas of modern organic-walled dinoflagellate cyst distribution. Review of Palaeobotany and Palynology, 125, 1–200.

Meeuwis J.M. & Lutjeharms J.R.E. (1990) Surface thermal characteristics of the Angola-Benguela front. South African Journal of Marine Science, 9, 261–279.

Mohrholz V., Schmidt M., Lutjeharms J.R.E., & John H.-C. (2004) Space–time behaviour of the Angola–Benguela Frontal Zone during the Benguela Niño of April 1999. International Journal of Remote Sensing, 25, 1337–1340.

Nicholson S.E. & Entekhabi D. (1987) Rainfall Variabillity in Equatorial and Southern Africa: Relationships with Sea Surface Temperatures along the Southwestern Coast of Africa. Journal of Climate and Applied Meteorology, 26, 561–578.

Pagani M., Zachos J.C., Freeman K.H., Tipple B., & Bohaty S. (2005) Marked decline in atmospheric carbon dioxide concentrations during the Paleogene. Science, 309, 600–3.

Pérez-Asensio J.N., Aguirre J., Schmiedl G., & Civis J. (2012) Impact of restriction of the Atlantic-Mediterranean gateway on the Mediterranean Outflow Water and eastern Atlantic circulation during the Messinian. Paleoceanography, 27, PA3222.

Petterson R.G. & Stramma L. (1991) Upper-level circulation in the South Atlantic Ocean. Progress in Oceanography, 26, 1–73.

Poore H.R., Samworth R., White N.J., Jones S.M., & McCave I.N. (2006) Neogene overflow of Northern Component Water at the Greenland-Scotland Ridge. Geochemistry Geophysics Geosystems, 7, Q06010.

Richter I., Behera S.K., Masumoto Y., Taguchi B., Komori N., & Yamagata T. (2010) On the triggering of Benguela Niños: Remote equatorial versus local influences. Geophysical Research Letters, 37, L20604.

Rommerskirchen F., Condon T., Mollenhauer G., Dupont L., & Schefuß E. (2011) Miocene to Pliocene development of surface and subsurface temperatures in the Benguela Current system. Paleoceanography, 26, 1–15.

Rouault M. (2003) South East tropical Atlantic warm events and southern African rainfall. Geophysical Research Letters, 30, 8009.

De Schepper S., Fischer E.I., Groeneveld J., Head M.J., & Matthiessen J. (2011) Deciphering the palaeoecology of Late Pliocene and Early Pleistocene dinoflagellate cysts. Palaeogeography, Palaeoclimatology, Palaeoecology, 309, 17–32.

De Schepper S. & Head M.J. (2008) Age calibration of dinoflagellate cyst and acritarch events in the Pliocene – Pleistocene of the eastern North Atlantic (DSDP Hole 610A). Stratigraphy, 5, 137–161.

De Schepper S. & Head M.J. (2009) Pliocene and Pleistocene Dinoflagellate Cyst and Acritarch Zonation of DSDP Hole 610a, Eastern North Atlantic. Palynology, 33, 179–218.

Schreck M., Matthiessen J., & Head M.J. (2012) A magnetostratigraphic calibration of Middle Miocene through Pliocene dinoflagellate cyst and acritarch events in the Iceland Sea (Ocean Drilling Program Hole 907A). Review of Palaeobotany and Palynology, 187, 66–94.

Ségalen L., Renard M., Lee-Thorp J.A., Emmanuel L., Le Callonnec L., de Rafélis M., Senut B., Pickford M., & Melice J.-L. (2006) Neogene climate

67

Stepwise Intensification of the Benguela Upwelling

change and emergence of C4 grasses in the Namib, southwestern Africa, as reflected in ratite 13C and 18O. Earth and Planetary Science Letters, 244, 725–734.

Shannon L.V., Boyd A.J., Brundrit G.B., & Taunton-Clark J. (1986) On the existence of an EI Niño-type phenomenon in the Benguela System. Journal of Marine Research, 44, 495–520.

Siesser W.G. (1980) Late Miocene Origin of the Benguela Upswelling System off Northern Namibia. Science, 208, 283–5.

Summerhayes C.P., Kroon D., Rosell-Melé A., Jordan R.W., Schrader H.-J., Hearn R., Villanueva J., Grimalt J.O., & Eglinton G. (1995) Variability in the Benguela Current upwelling system over the past. Progress in Oceanography, 35, 207–251.

Udeze C. & Obohikuenobe F. (2005) Neogene palaeoceanographic and palaeoclimatic events inferred from palynological data: Cape Basin off South Africa, ODP Leg 175. Palaeogeography, Palaeoclimatology, Palaeoecology, 219, 199–223.

De Vernal A., Rochon A., Turon J.-L., & Matthiessen J. (1997) Organic-walled dinoflagellate cysts: palynological tracers of sea-surface conditions in middle to high latitude marine environments. Geobios, 30, 905–920.

Versteegh G.J.M. (1997) The onset of major Northern Hemisphere glaciations and their impact on dinoflagellate cysts and acritarchs from the Singa section, Calabria (southern Italy) and DSDP Holes 607/607A (North Atlantic). Marine Micropaleontology, 30, 319–343.

De Verteuil L. & Norris G. (1992) Miocene Protoperidiniacean Dinoflagellate Cysts from the Maryland and Virginia Coastal Plain. Neogene and Quaternary Dinoflagellate Cysts and Acritarchs (ed. by M.J. Head and J.H. Wrenn), pp. 391–

430. American Association of Stratigraphic Palynologists Foundation, Dallas.

Vidal L., Bickert T., Wefer G., & Röhl U. (2002) Late Miocene stable isotope stratigraphy of SE Atlantic ODP Site 1085: Relation to Messinian events. Marine Geology, 180, 71–85.

Wall D., Dale B., Lohmann G.P., & Smith W.K. (1977) The environmental and climatic distribution of dinofalgellate cysts in modern marine sediments from regions in the North and South Atlantic Oceans and adjacent seas. Marine Micropaleontology, 2, 121–200.

Wright D., Miller G., & Fairbanks G. (1992) Early and Middle Miocene stable isotopes: Implications for Deepwater circulation and climate. Paleoceanography, 7, 357–389.

Zachos J., Pagani M., Sloan L., Thomas E., & Billups K. (2001) Trends, rhythms, and aberrations in global climate 65 Ma to present. Science, 292, 686–93.

Zegarra M. & Helenes J. (2011) Changes in Miocene through Pleistocene dinoflagellates from the Eastern Equatorial Pacific (ODP Site 1039), in relation to primary productivity. Marine Micropaleontology, 81, 107–121.

Zonneveld K.A.., P. Hoek R., & Brinkhuis H. (2001) Geographical distributions of organic-walled dinoflagellate cysts in surficial sediments of the Benguela upwelling region and their relationship to upper ocean conditions. Progress in Oceanography, 48, 25–72.

Zonneveld K.A.F., Marret F., Versteegh G.J.M., Bogus K., Bonnet S., Bouimetarhan I., Crouch E., de Vernal A., Elshanawany R., Edwards L., Esper O., Forke S., Grøsfjeld K., Henry M., Holzwarth U., Kielt J.-F., Kim S.-Y., Ladouceur S., Ledu D., Chen L., Limoges A., Londeix L., Lu S.-H., Mahmoud M.S., Marino G., Matsouka K., Matthiessen J.,

68

Chapter IV

Mildenhal D.C., Mudie P., Neil H.L., Pospelova V., Qi Y., Radi T., Richerol T., Rochon A., Sangiorgi F., Solignac S., Turon J.-L., Verleye T., Wang Y., Wang Z., & Young M. (2013) Atlas of modern dinoflagellate cyst distribution based on 2405 datapoints. Review of Palaeobotany and Palynology, 191, 1–197.

69

Stepwise Intensification of the Benguela Upwelling

70

Table 2: Recognised species ordered after metabolism.

Autotrophic species

Impagidinium sp. 2 (De Schepper &

Head, 2009)

aculeatum

paradoxum

sphaericum

patulum

strialatum

Operculodinium spp.

centrocarpum

israelianum

janduchenei

Phentapharsodinium dalei

Lingulodinium machaerophorum

Spiniferites spp.

ramosus

mirabilis

cf mirabilis

membranaceus

Acomosphaera spp

Nematosphaeropsis labyrinthus

Batiacasphaera micropapillata

hirsuta

Tuberculodinium vancampoae

Pyxidinopsis reticulata

Habibacysta tectata

Ataxodinium zevenboomii

confusum

Chapter IV

Heterotrophic species

Brigantedinium spp.

Selenopemphix quanta

neproides

brevispinosa

brevispinosa conspica

armageddonensis

Trinovantedinium glorianum

papulum

fergnomatum

applanatum

Lejeunecysta spp

oliva

sabrina

Quinquecuspis concreta

Unknown, further species

Sumatradinium soucouyanatae

71

Chapter V

Chapter V: Synthesis and Outlook

The main aim of the thesis was to

investigate the Miocene to Pliocene

vegetation change of south-west Africa.

Furthermore the thesis focused on the

mechanisms causing the savannah grassland

expansion. The stated hypotheses were

addressed and answered:

5.1 Aridification favoured the savannah

expansion

The first and the second manuscript

(Chapter II and Chapter III) focused on the

vegetation development of south-west

Africa from the late Miocene to Pliocene (9 -

2.7 Ma). The pollen and spore content

suggests a progressive aridification

represented by vegetation change. At the

beginning of record wet conditions prevailed

indicated by Cyperaceae (sedges) pollen and

a wider expansion of mountain and

woodland vegetation. Wet vegetation types

were successively replaced by drier ones.

From 8.3 Ma on the grasses expanded and

the wet indicators decreased. Since grasses

are generally better adapted to drier

conditions the savannah was able to expand.

However, advanced aridification led the

grassland to become sparser after 4 Ma,

when shrubland and desert vegetation

expanded. It seems that especially C3 grasses

retreated from the savannah because the

overall Poaceae pollen percentages decrease

while at the same time the stable carbon

isotope composition of plant wax (δ13Cwax)

indicates increased C4 contributions.

The aridification is a trend recognizable on

the entire African continent (Senut et al.,

2009) fitting the global cooling trend during

the Neogene. A steeper meridional

pressure/thermal gradient could have been

the driver for these changes by intensifying

the atmospheric circulation (Flower &

Kennett, 1994). As a consequence the trade

winds strengthened and the CAB might

have shifted with the result that intensified

trade winds would have hindered humid air

masses from the Atlantic to reach further

inland into the continent, which led to drier

conditions in Namibia. In South Africa the

climate changed from a summer to a winter

rain climate (Dupont et al., 2011) during the

same period.

5.2 Fire and its role in the savannah

expansion

The charcoal record shown in the first

manuscript suggests that a regime of

intensified fires took place from 7.1 – 5.8

Ma. Intensification of fires is most probably

the result of the aridification in combination

with enhanced seasonality. During the

humid season grasses would grow and built

up fuel which flammability would increase

during the dry period. From 7.1 – 5.8 Ma

the balance between wet and dry periods

obviously was excellent for promoting fires,

73

Synthesis and Outlook

whereby a sufficient amount of rainfall was

essential. Daniau et al. (2013) showed for the

last 170.000 years that charcoal increased in

south-west Africa during wetter stages and

that a moderately wet climate enabled higher

fire activities. We conclude, therefore, that

the conditions between 7.1 – 5.8 Ma were

significantly wetter than today but drier than

before the fire period. As rainfall continued

to decrease the grasses retreated and the

extreme fire regime ended.

During the fire regime a shift in the isotopic

carbon composition is recorded which is

best explained by a change in the

contribution of C4 plants to the total

vegetation. The data suggest a selecting

process of C4 grasses over C3 grasses by the

fires. That fire influences the selection of

grasses has been shown in South Africa

where one lineage consisting almost

exclusively of C4 grasses is very abundant

and diverse in areas with frequent fires

(Visser et al., 2012). The fire regime thus had

an irreversible impact on the south-west

African biomes by establishing C4 grasses

which benefitted afterwards from better

competitiveness under arid conditions (e.g.

higher water-use efficiency) (Hatch, 1987)

and reduced atmospheric CO2 levels

(Ehleringer et al., 1991).

5.3 Aridification and its link to the

Benguela Upwelling

In the third manuscript (Chapter IV) the

dinoflagellate communities suggest changes

in the ocean circulation and upwelling

intensity. A weak pressure gradient is

suggested for the period prior to 7.8 Ma. A

stronger influence of the Angola Current,

comparable with Benguela Niño conditions,

occurred more regularly in that time period.

From 7.8 Ma on, the influence of the

Angola Current decreased most probably

because of an intensifying meridional

pressure gradient. Resulting stronger trade

winds also enhanced the upwelling which is

recorded as an increase of dinoflagellate cyst

accumulation in general and that of

heterotrophic species in particular. Within

the period from 6.8 to 5.2 Ma the quality of

the upwelled water changed which is related

to the Messinian Salinity crisis. The varying

amount of outflow from the Mediterranean

Sea and the closing of the Central American

Seaway influenced the North Atlantic

deepwater formation (Billups, 2002; Pérez-

Asensio et al., 2012) and probably

influenced the source water of the Benguela

Upwelling System as well. The continuation

of the shoaling of Panama and the

intensification of the Hadley Circulation

further increased the upwelling and

therefore the productivity after 4.4 Ma.

These stepwise changes correlate well to the

development of the vegetation on the

continent. In both cases the data suggest a

steepening of the meridional pressure

gradient resulting into a northwards shift of

the Angola-Benguela Front reflected by the

dinoflagellate cyst assemblage changes and a

decrease of humidity on the continent

reflected by the vegetation changes.

74

Chapter V

5.4 Outlook

The suggested role of fire as a driver of the

C4 grassland expansion is to be further

tested in other regions were they dominate

today such as East Africa and South

America. The combination of pollen,

microscopic charcoal and stable isotope

analysis might also reveal the role of fire in

the expansion of East African savannah,

which is of special interest regarding

hominine evolution e.g. the development of

bipedalism (Cerling et al., 2011). A pollen

and charcoal record from sediments of the

Niger Delta (Morley & Richards, 1993)

already suggested that there are links

between fire and grassland expansion in

West Africa. A combination of stable

isotope analysis and pollen analysis might

reveal insights in the contribution of C4

grasses and reveal if they are here favoured

by fires, too. In north-east Africa a study of

pollen and carbon isotope data shows that

the savannah grassland expansion in that

region is independent from the C4 grass

contribution (Feakins et al., 2013). Hence, a

re-analysis of these data considering the

microscopic charcoal might resolve a similar

pattern in relation to fire as for the current

study.

The results of this thesis provide more

details about the meridional

thermal/pressure gradient during the Late

Miocene and Early Pliocene in south-west

Africa, which region has been sparsely

covered so far and could, therefore, help to

improve earth system models of the Mio-

/Pliocene. Additionally, since fire is an

important driver in grassland expansion it

should be considered in vegetation models,

especially because it might give indications

about seasonality contrasts.

5.5 References

Billups K. (2002) Late Miocene through early Pliocene deep water circulation and climate change viewed from the sub-Antarctic South Atlantic. Palaeogeography, Palaeoclimatology, Palaeoecology, 185, 287–307.

Cerling T.E., Wynn J.G., Andanje S. a., Bird M.I., Korir D.K., Levin N.E., Mace W., Macharia A.N., Quade J., & Remien C.H. (2011) Woody cover and hominin environments in the past 6 million years. Nature, 476, 51–56.

Daniau A.-L., Sánchez Goñi M.F., Martinez P., Urrego D.H., Bout-Roumazeilles V., Desprat S., & Marlon J.R. (2013) Orbital-scale climate forcing of grassland burning in southern Africa. Proceedings of the National Academy of Sciences of the United States of America, 1–5.

Dupont L.M., Linder H.P., Rommerskirchen F., & Schefuß E. (2011) Climate-driven rampant speciation of the Cape flora. Journal of Biogeography, 38, 1059–1068.

Ehleringer J.R., Sage R.F., Flanagan L.B., & Pearcy R.W. (1991) Climate change and the evolution of C4 photosynthesis. Trends in ecology & evolution, 6, 95–9.

Feakins S.J., Levin N.E., Liddy H.M., Sieracki a., Eglinton T.I., & Bonnefille R. (2013) Northeast African vegetation change over 12 m.y. Geology, 41, 295–298.

Flower B.P. & Kennett J.P. (1994) The middle Miocence climatic transition: East Antarctic ice sheet development, depp ocean circulation and global carbon cycling. Palaeogeography, Palaeoclimatology, Palaeoecology, 108, 537–555.

75

Synthesis and Outlook

76

Hatch M.D. (1987) C4 photosynthesis: a unique blend of modified biochemistry, anatomy and ultrastructure. Biochimica et Biopohysica Acta, 895, 81–106.

Morley R. & Richards K. (1993) Gramineae cuticle: a key indicator of Late Cenozoic climatic change in the Niger Delta. Review of Palaeobotany and Palynology, 77, 119–127.

Pérez-Asensio J.N., Aguirre J., Schmiedl G., & Civis J. (2012) Impact of restriction of the Atlantic-Mediterranean gateway on the Mediterranean Outflow Water and eastern Atlantic circulation during the Messinian. Paleoceanography, 27, PA3222.

Senut B., Pickford M., & Ségalen L. (2009) Neogene desertification of Africa. Comptes Rendus Geoscience, 341, 591–602.

Visser V., Woodward F.I., Freckleton R.P., & Osborne C.P. (2012) Environmental factors determining the phylogenetic structure of C4 grass communities. Journal of Biogeography, 39, 232–246.

Acknowledgements 

Acknowledgements

First of all, my acknowledgement goes to

Prof. Wefer giving me the chance to work at

MARUM and supporting my studies at every

state of my thesis.

My deepest gratitude to Lydie Dupont who

might be the best supervisor; Thank you for

your always fast replies, your extremely

valuable comments as well as for your

support and your cheering up smiles.

Furthermore I thank Enno Schefuß, you

pushed the idea to publish the first

manuscript in nature geoscience. I thank you

for sharing your experiences in publishing

such “small papers”. In total I would like to

thank my thesis committee (Prof. Wefer,

Lydie Dupont, Enno Schefuß and Matthias

Prange) for directing me through the three

years without difficulties.

I thank GLOMAR and its team for several

fundings enabling me to gather valuable

experiences on various conferences and

workshops. Furthermore GLOMAR

supported me and funded me for my

research stay in Liverpool so that I could

study the dinoflagellate cysts of my samples.

Fabienne Marret I thank for inviting me

over to Liverpool for my research stay

which brought my project miles further.

Thank you for sharing your knowledge

about dinoflagellates.

Jens Matthiessen and Michael Schreck I

would like to thank for clarifying taxonomic

troubleshooting dinoflagellates.

For laboratory work I thank very much

Marta Witek.

My former and recent office mates and

office mates of my “second” office I thank

for help, discussions and distractions (only

to increase creativity). Anna and Anna,

Cyril, Enquing, Francesca, James, Ines,

Rodrigo, Rony and William thank you all for

making my days colourful. Additionally, I

thank my further friends at the university

supporting me through the years; Aline,

Alice, Bevis, Eva, Hendrik, Jana, Katja,

Mariem, Stefan, Yann, Zsuzsanna and for

give me who I might have forgot.

Further thanks go to Christian Breuer and

Ulrike Gerdes for giving me the feeling that

I felt welcome in Bremen from the very first

day.

Coming to the end I thank my parents for

years of support during my studies, helping

to get started in Bremen and also for

worrying (although not necessary).

My last thanks go to Sarah Klingenberg for

being my balance and equilibrium. Thank

you.

With the end of a thesis a period of time

ends… It was a wonderful one!

Thank you all!

Sebastian Hötzel

77