Past Climate Variability through Europe and Africa

This volume contains a series of papers that collectively summarise evidence for climate change in Africa and Europe over approximately the last 200,000 years. It is a product of PEP III, the pole-equator-pole transect through Europe and Africa, that has been defined by the IGBP-PAGES project for palaeoclimate study along with PEP II (Asia-Australasia) and PEP I, the Americas (Fig. 1). The PEP transects focus on two time intervals, time-stream 1, the Holocene (with an emphasis on the last 2000 years), and time-stream 2, the last two glacial cycles. The principal long-term objectives of PEP III are (i) to understand how and why climate has varied in the past along the transect; (ii) to assess how climate change and variability has affected natural ecosystems and human society in the past; and (iii) to provide a basis both for developing and testing climate models that are needed to forecast climate change in the future. The specific papers contained in this volume were presented at a conference on “Climate Variability through Europe and Africa” held in Aix-en-Provence in August, 2001. The volume presents an attempt to bring together in a coherent way our present understanding of past climate change along the PEP III transect, providing a basis for the future work needed to make progress towards these objectives.

This publication also complements other related syntheses that overview the scientific achievements of the PAGES community over the last decade, notably for the PAGES programme as a whole (Alverson et al. 2003), and for the PEP III sister transect, PEP I (Markgraf 2001) and PEP II (Dodson et al. 2004).

The PEP III Transect (Gasse and Battarbee, this volume) spans an immense range of environmental and cultural diversity. It includes regions that have been the cradle of Old World civilisations dating back many millennia, it is home to some of the most advanced and favoured societies on the planet, some of the least hospitable environments and some of the most vulnerable and economically impoverished peoples on earth. Documenting and understanding the ways in which climate has varied across the whole length of the Transect and teasing out the past interactions between climate and human welfare pose immense challenges to the scientific community. Moreover, the challenges encompass themes that are of outstanding practical importance for our future as well as part of our compulsive fascination with the past.

The challenges require us to make full use of the instrumental record of recent climate variability and this is the concern of the first part of this chapter. But that alone does not suffice.Valuable though it is, the brief instrumental record fails to encompass the full range of natural climate variability in both time and space that has characterised the last few thousand years. In order to provide a record of that variability, it is necessary to turn to a whole range of ‘archives’, both documentary and environmental, and to analyse them using methods that can capture datable and decipherable signals of past variability. For this, we rely largely on what are usually termed ‘proxies’. After outlining the scope for instrumentally based climate reconstructions across the whole length of the Transect, the present chapter seeks to give a brief introduction to the archives and proxies that provide a basis for reconstructing past environmental and climatic variability within the PEP III domain.

One of the most exciting questions in palaeoclimatology is the study of the complex interactions between the different components of the climate system in order to understand how climate changes occur at Milankovitch as well as at millennial and centennial time-scales. The primary objective of this paper is to place the PEP III transect palaeo-data within a global climate context in relation to oceanic climate variability during the last glacial. To take ocean-continent interactions into account is essential to develop our understanding of past climate change.

Recently, several studies have brought to light the role of the tropical/subtropical areas in climate changes. Continental and oceanic low-latitude records have documented the high frequency climate variability during the last glacial period. The inferred continental climate conditions in the sub-tropics might provide evidence for the contribution of atmospheric processes associated with rapid climate variability.

Another key question concerns the forcing mechanisms for rapid climate change. One of the most discussed hypotheses relies on ice sheet-ocean-atmosphere interactions driving millennial scale climate variations during the last glacial period (Alley et al. 1999; Ganopolski and Rahmstorf 2001). As already mentioned, ocean-atmosphere coupling is very sensitive to the freshwater flux into the North Atlantic. Modelling of the cryosphere shows that the internal dynamics of the northern hemisphere ice-sheets during periods of ice-building could lead to major reorganisation of the global climate (MacAyeal 1993). Likewise internal oscillations of the ocean-atmosphere system should also be considered as a potential candidate for climate forcing (Winton 1993; Paul and Schulz 2002). On the other hand, studies on rapid climate variability during the Younger Dryas and the Holocene point to an external forcing by revealing a possible connection between solar activity and global climate (Renssen et al. 2000; this volume; Bond et al. 2001). However, knowledge about the importance of the different potential triggers is poor, as climate components have different responses in space and time (and most probably during glacial and interglacial periods).

Therefore, in this paper we concentrate on the interactions between the different climate sub-systems.Areviewof rapid climate variability during the last glacial period mostly based on marine sediment records with a focus on recent studies based on sub-tropical records is attempted. By compiling these high quality data, we discuss what can be learned about millennial climate changes and connections.

Terrestrial records for the Middle Pleistocene of Southern Africa are sparse and mostly fragmentary because the landscape was largely devoid of suitable depositories. Fortunately, one continuous lacustrine sequence, that of the Tswaing impact crater near Pretoria, South Africa (Fig. 1), spans the period from ~200 kyr to the present. Recent analyses of marine cores off the west coast of Southern Africa have, in addition, vouchsafed important proxy evidence of palaeo-wind regimes, periods of aridity and terrestrial flora.A generalised, but largely consistent, picture of palaeoenvironmental fluctuations during Isotope Stage 6 can therefore be drawn.

Southern Africa’s unique mid-latitude oceanic position invites broad comparisons with palaeoclimatic records across the PEP III transect including long-distance thermohaline circulation teleconnections. Atmospheric and oceanic circulation systems around Southern Africa (Fig. 1) are linked (Lutjeharms et al. 2001). They interact to influence distribution of biomes including prominence of C₄ and C₃ grasses in the summer-rain and winter rain regions respectively (Vogel et al. 1978; Cowling et al. 1997) (Fig. 2). Climate is dominated by two systems, the westerlies and easterlies (Fig. 1) and shifts in these systems undoubtedly affected the climate history during the Holocene. The generally moderating effect of the oceans, in particular the warm Agulhas western boundary current on the east coast (Lutjeharms et al. 2001), and the semi-arid nature of the region suggests that moisture rather than temperature changes is the more important climate parameter, at least on the Holocene time scale. Furthermore, the cold Benguela upwelling zone and its associated atmospheric circulation system in the South Atlantic is strongly linked to coastal aridity on the west coast.

The large lakes of the East African Rift Valley provide a magnificent array of sedimentary basins that are actively recording the modern climate dynamics of tropical East Africa. The basins are perhaps 10 million or more years old (Cohen et al. 1993), and their deepest reaches have archived continuous records of past climate change that are unparalleled anywhere else in the tropics in terms of longevity coupled with resolution (Johnson 1993). These lakes are highly sensitive to climate variability, and their sediments carry rich signals of past environmental dynamics, inscribed in the assemblages of microfossils, the abundance and isotopic composition of endogenic minerals, bulk sediment texture and structure, and organic geochemistry.

The International Decade for the East African Lakes (IDEAL) is a programme that has helped to steer and coordinate palaeoclimate studies on the rift lakes since the programme’s inception in 1993 (Johnson 1993). The IDEAL community of scientists from North America, Europe, and Africa is investigating not only the climatic history of the large rift lakes, but modern processes within their basins as well. The “decade” of IDEAL officially started in 1995 with its first field programme on Lake Victoria (Lehman 1998). Since then IDEAL scientists have continued to carry on investigations on Lakes Edward, Malawi and Tanganyika, as well as on smaller lakes in East Africa.

The IDEAL palaeoclimate community is presently turning its focus to Lake Malawi, where a major drilling programme is planned for the last two months of 2004. Lake Malawi is the southern-most of the rift lakes, and it appears to be exceptionally promising as a target for long, continuous records of past climate change. It has responded somewhat differently from lakes farther north in Africa to global climate forcing associated with orbital (Milankovitch) cycles of insolation. In contrast to Lake Tanganyika and the lakes of North Africa, Lake Malawi experienced relatively dry conditions during much of the early Holocene (Finney et al. 1996; Johnson 1996). The history of Lake Malawi derived from drilling operations therefore may provide important insight into how the southern tropics of Africa have responded differently from the northern tropics on time-scales ranging from millennia to decades.

This paper presents recent data derived from piston cores recovered from the north basin of Lake Malawi. It focuses on the abundance and mass accumulation rate of biogenic silica (BSi MAR) measured in the cores, supplemented by some trace-element data and the species composition of the diatom assemblages. These results imply linkages between the climate dynamics of tropical Southern Africa and extratropical regions as far away as Greenland.

Tropical Africa is at the geographical heart of the PEP III transect and forms part of the heat engine which drives the meridional circulation of the atmosphere. The tropics are therefore central to studies of climate change not only in the equatorial belt but also in sub-tropical regions (Yin and Battisti 2001), and may even lead some high latitude climate changes (Henderson and Slowey 2000). Despite this pivotal role, the tropics have historically been the poor relation of temperate regions in palaeoenvironmental research. Here we synthesise newpalaeoclimate information derived primarily from forest and savanna regions of tropical Africa. These recent studies have helped to close some empirical gaps in data coverage within the region and at the same time have led to a new conceptual understanding of past tropical climates. The latter has been achieved through the rigorous application of classical methods as well as the development of several new proxies of environmental change.

We focus our discussion on several key palaeoclimate issues that fall within the scope of PAGES time stream 2 (cf. Gasse and Battarbee (this volume)). Although this is defined as spanning the last two glacial cycles, we will place greatest emphasis on the events since the penultimate glacial (marine isotope stage 6, hereafter MIS 6) and especially on the last 30,000 calendar years BP (abbreviated below to kyr) (broadly MIS 1-3). At these long time-scales, orbital-forcing factors, modified by earth-surface feedback mechanisms, ought to be dominant and be recognisable in the biotic and hydrological systems. We will then examine the evidence for abrupt events at century to millennial scales, which are prominent features in many climate reconstructions since the Last Glacial Maximum (LGM). Climate forcing at these time-scales is tightly coupled to oceanic changes and so we will extend our discussion to a number of offshore sites. Numerous terrestrial archives are now being explored including some excellent high-resolution speleothem records (Holmgren et al. 2001), but to limit our study we will confine our discussion to lakes and the major river systems (Fig. 1).

Few proxies are directly related to the fundamental climate variables of precipitation and temperature. Even in the tropics where shifts from wet to dry conditions commonly outweigh temperature changes, it is difficult to tease these variables apart. Moreover, the part played by atmospheric gases such as CO₂ in terrestrial environmental change is now also recognised (Street-Perrott 1994; Jolly and Haxeltine 1997; Street-Perrott et al. 1997). Identifying the specific contribution of the different variables is of primary concern to tropical palaeoenvironmental studies and has promoted the development of several novel approaches to environmental reconstruction.

Holocene climate in high-latitude regions of the world has been relatively stable compared to glacial climates. In contrast, tropical Africa and other low-latitude continental regions were marked by a succession of millennium-scale wet and dry episodes, separated by rather abrupt transitions (Gasse and Van Campo 1994; Lamb et al. 1995; Gasse 2000). These continent-wide fluctuations in the balance of rainfall and evaporation must somehow have resulted from large-scale variation in the position or intensity of large-scale tropical monsoon systems, but their relationship to Holocene climate variability in extra-tropical regions (e.g., Bond et al. (2001)) and the likely mechanisms of external climate forcing are only just beginning to be revealed (Gupta et al. 2003; Hoelzmann et al., this volume).

Compared to this marked hydrological instability of African climate during the early and middle Holocene, the last 2000 years have commonly been thought of as rather stable and uneventful. This idea can be traced back to the first reviews of late-Quaternary vegetation and lake-level change in Africa (Butzer et al. 1972; Livingstone 1975; 1980; Hamilton 1982), which in their focus on the prominent late-Glacial and early-Holocene events found littleworth mentioning in the late Holocene. Lowtime resolution and poor age control meant that the last 2000 years were typically represented by just a few data points floating on an interpolated section of the time line. In addition, evidence for 20th-century landscape disturbancewas often missing because soft surface muds had not been recovered, or had been discarded. This lack of a reference frame for signatures of pre-modern human impact, together with the assumption of relative climatic stability, helped perpetuate among palaeoecologists, archaeologists, and geomorphologists the paradigm that most evidence for vegetation and landscape change in tropical Africa younger than 2000 years is due to human activity (Taylor 1990; Jolly et al. 1997; Eriksson 1998).

The Horn of Africa extends from the Sahara desert eastwards to the Red Sea, the Gulf of Aden and the Indian Ocean. Due to its complex volcano-tectonic evolution over the past 15 million years (Mohr 1971), the region is characterised by considerable changes in elevation within short distances. The Ethiopian highlands, which rise to altitudes exceeding 4000m above sea level, form an extensive uplifted plateau, delimited by pronounced escarpments on both east and west. In Southern Ethiopia the 1000 km-wide uplifted volcanic province is divided asymmetrically into northwestern and southeastern plateaux by the Main Ethiopian Rift which runs SSW-NNE and represents the northern end of the continental East African Rift. Eastward, the continental Afar Rift (Ethiopia and Republic of Djibouti) sits astride the Gulf of Aden—Red Sea sea-floor spreading axis, with several areas below sea level (e.g.,–155mat Lake Asal). These marked altitudinal gradients result in a wide variety of climates and environments, making the region particularly suitable for investigating past environmental change. The region is subject to the seasonal migration of the Inter-Tropical Convergence Zone (ITCZ) and is very sensitive to monsoon variability. Several studies have revealed climate variability from millennial to inter-annual timescales during the Late Quaternary (Gasse 2000).

This paper provides a detailed overview of this region complementary to those of Hoelzmann et al. (this volume) and Barker et al. (this volume). It reviews Late Quaternary palaeoenvironmental and palaeoclimatic changes, as recorded from various archives (lakes, high altitude peat-bogs, glacial landforms, palaeosols), and proxies (geomorphology, sedimentology, geochemistry, pollen, diatoms). The chronological framework of most records is based on radiocarbon ages (¹⁴C yr BP). Radiocarbon ages are converted into calendar estimates (cal. yr) using the CALIB 4.0 program (Stuiver et al. 1998) back to 18,000 ¹⁴C yr BP, or the polynomial equations established by Bard (1998) which allow calendar estimates back to 36,000 ¹⁴C yr BP (41,516 cal. yr). Ages are first given in the time-scale provided in the original literature (¹⁴C kyr BP or cal. kyr), then the calibrated scale is indicated in brackets.

The pollen data available for the Atlantic equatorial African region indicate that during the Late Quaternary, major climate changes occurred and caused important modifications in tropical lowland rain-forest in terms of composition and distribution (Maley 1991; Elenga et al. 1991; 1994; Farrera et al. 1999; Vincens et al. 1999). Nevertheless, despite a wealth of proxy evidence from various disciplines, there are very few terrestrial sites dated continuously since the Last Glacial Maximum (LGM) to the Late Holocene in this area (e.g., Lanfranchi and Schwartz (1990), Servant and Servant-Vildary (2000)). The oldest radiocarbon-dated sites are from Lake Barombi Mbo in Cameroon (Brenac 1988; Maley 1991; Maley and Brenac 1998a) and the Ngamakala Pond in the Congo (Elenga et al. 1994), although most Holocene sequences from this area start during the mid-Holocene (Vincens et al. 1999). This paper reviews the results available for some key sites, allowing regional correlations for major Late Quaternary periods.We have focused on the vegetation changes and the climate related to these changes. Numerous radiocarbon dates allowprecise palaeoenvironmental reconstructions, especially since the mid-Holocene to be made. In this paper all ¹⁴C dates are uncalibrated.

A palynological study of an 11m terrestrial core from the Badagry area in coastal South- Western Nigeria was undertaken as part of an international project: “The Dahomey Gap: vegetation history and archaeobotany of the forest/savanna boundary in Bénin and South- Western Nigeria.” The project is the joint effort of a team comprising an archaeobotanist (German), a botanist (Béninois), and four palynologists (Béninoise, German, Moroccan, and Nigerian). The Dahomey Gap is an unusual portion of the Guinean-Congolian forest zone, comprising a mosaic of savanna and the drier type lowland rain-forest. This Gap today stretches from the western-most part of Southern Nigeria through Bénin (formerly Dahomey) and Togo to the eastern-most part of Ghana. It effectively partitions the forest zone into western and eastern blocks (Plate 2).

As reviewed by Hamilton (1976), Martin (1990) and Maley (1996), biogeographers, studying the African tropical forest biota, plants, mammals, birds, insects, butterflies and molluscs, were the first to recognise centres of endemism in West and Central Africa. These centres, which “showa considerably higher species diversity than surrounding forest areas” (Martin 1990) are believed to have been refuges during the arid phases of the Pleistocene when the rain-forest was drastically reduced and fragmented. The issue of exactly where the refuges were is as yet not completely resolved. In the most current reconstruction, based on the work of several authors, these refuges during the peak of the last major arid phase are believed to have been in Liberia, Southwest Ghana and Southeast Côte d’Ivoire in the west, and South Cameroon, Gabon and Congo-Zaïre basin in the east.

The PEP III Arid to Subarid Belt includes the largest hot desert in the world, the Sahara- Arabian desert and the Sahel zone. The region of interest extends south of the Atlas Mountains and south and east of the Mediterranean Sea to approximately 10 °N and shows a broadly zonal pattern with a varying seasonal distribution of precipitation. In the north (ca. 20–23 °N), rainfall results from the southward displacement of the midlatitude westerlies during winter whereas the south is governed by seasonal northward migration of the Intertropical Convergence Zone (ITCZ). Contraction and expansion phases of these presently semi-arid to hyper-arid desert areas result from significant changes in local precipitation. Palaeoenvironmental records from Northern Africa (north of 10 °N) and the surrounding seas document long-term changes in the magnitude and extent of the African monsoon in response to orbitally-forced changes in insolation. However, marine records as well as terrestrial palaeohydrological indicators (e.g., lakes, speleothems, rivers, pollen and charcoal) show that there have been changes in the hydrological cycle superimposed on the long-term waxing and waning of the monsoon which cannot be explained exclusively by changes in insolation. These fluctuations in space, time and magnitude were on a regional to continental scale.

Here, we review available data on near-surface palaeohydrological indicators and vegetational changes in arid North Africa and the Arabian Peninsula as well as changes in the intensity of the South Asian Monsoon identified from marine sediments of the Arabian Sea. A comparison of regional environmental changes can clarify relations between the environment and changes in the Earth’s climate system. Each data-set is initially presented independently because they represent heteregeneous records from different regions and time periods and thereby emphasise their potential to provide evidence of continental chronostratigraphic palaeoenvironmental changes. Data-sets of lake status and vegetational change are complementary as they strongly reflect hydrological variation. Deep-sea sediments from the Arabian Sea were used to generate continuous records of oceanic upwelling, continental humidity, and dust and river discharge, that are closely related to palaeoenvironmental changes on the surrounding continents.After presenting the individual data-sets we compare the palaeoclimatic reconstructions derived from the different types of evidence.

At present the El Niño system is one of the primary causes of inter-annual climatic variability, not only in tropical regions but also on a global level (Ropelewsky and Halpert 1987; Philander 1989; Kiladis and Diaz 1989; Glynn 1990). A considerable effort is dedicated to the understanding of the El Niño-Southern Oscillation (ENSO) system and its interactions with other kinds of climatic variations, at the decadal/interdecadal (e.g., the Pacific Decadal Oscillation) and centennial (Little Ice Age, MedievalWarm Period) scales (Anderson et al. 1992; Bradley and Jones 1992; Allan et al. 1997; Diaz and Markgraf 2000). Under societal and political pressure, several international research programs aim to decipher the relationships that link ENSO and the present global warming of the planet (Ropelewesky 1992; Cane et al. 1997;Timmermann et al. 1999). These scientific challenges and questions require that long series of climate variations, much longer than those available from instrumental data alone, be compiled and studied. A renewed interest for the climate evolution, at local, regional and global scales, during the last few centuries, is thus justified by the need to understand better the climatic particularities of the Little Ice Age (LIA) and the transition period to the modern situation (19th and 20th centuries) (e.g., Grove (1988), Bradley and Jones (1992, 1993), Allan and d’Arrigo (1999), Allan (2000), Verschuren (this volume), Scott and Lee-Thorp (this volume)). These studies should help to determine how much of the recent global warming results from anthropogenic activities as against a natural post-LIA evolution. Deciphering the relationships between the global state of the planet and the El Niño system (frequency and intensity of individual events, teleconnection links) requires us to document as precisely as possible former climate variations and ENSO manifestations, during warm and cold periods, such as the MedievalWarm Period and the LIA respectively, and to compare them to the present (instrumentally controlled) climate variability (Bradley and Jones 1992; 1993; Diaz and Pulwarty 1994; Jones et al. 2001).

Droughts and wet periods in recent earth history have a clear link with human habitation and migration notably in the Middle East and in Africa. Water availability in the form of perennial rivers, lakes and springs dictated the first settlement patterns. Reference is found in the first written records to dramatic climatic changes affecting water availability (Lambert and Millard 1969) and a fascinating challenge exists to reconstruct and to relate the hydrological records to the early historical and archaeological evidence (Issar 1990; Hassan 1997) as well as the rapidly growing body of data on palaeoclimate.

Groundwater is emerging as an archive at both low and mid-latitudes of past climatic and hydrological change, which may be used alongside other proxy data. Indirect evidence of the palaeohydrology in the late Pleistocene and Holocene, has been deduced from various sources especially lake sediments (Fontes and Gasse 1991; Gasse 2000; Hoelzmann et al. 2000; this volume) and speleothems (Bar-Matthews et al. 1997). In contrast to other archives such as ice cores or tree rings, which contain high-resolution information, data available in large groundwater bodies are of low resolution (typically ±1000 yr). This is due to the advection or dispersion of any climatic input signal in the water body. Many groundwater data are obtained from pumped samples where sample intervals may extend over tens of metres. Nevertheless, specific indications of palaeo-temperature, air mass circulation and vegetation history may be retained in a range of different chemical and isotopic signals (Fontes et al. 1993a; Stute and Schlosser 1993), notably in confined aquifers where sequential changes may be recorded along flow lines, or in the stratification of phreatic aquifers. Dated groundwaters are important since these contain the direct evidence of prolonged wet episodes. Even the absence of dated waters over a specific time interval may indicate periods of drought (Sonntag et al. 1978). The correlation between groundwater records and aeolian deposition in semi-arid/arid regions (Stokes et al. 1998; Swezy 2001) can also provide complimentary evidence of wet and dry intervals.

Moisture in the unsaturated zone may under favourable circumstances also contain records of past environments and climate at decadal to millennial scale resolution, contained mainly as variations in salinity and in stable isotope enrichments (Edmunds and Tyler 2002). Such records are found in porous media in areas of low moisture flux, notably beneath modern arid or semi-arid areas. The resolution of unsaturated (vadose) zone records will depend on the dispersion of the signal (Cook et al. 1992) but decadal scale records may be retained, as in West Africa, over several hundred years (Edmunds and Gaye 1997), or at the millennial scale over the late Pleistocene (Tyler et al. 1996). Some of the classical studies of hydrogeological systems that contain palaeo-waters have been conducted in Northern Africa and in Western Europe over the past three decades. In this paper the evidence contained in unsaturated zone profiles and especially in phreatic and confined aquifers in semi-arid and arid areas of Northern Africa is used to demonstrate the current possibilities in using groundwater archives. The results from Northern Africa are then compared with the evidence found in palaeowaters in Europe and elsewhere along the PEP III transect.

In this paper, we present a comprehensive synthesis of our recent studies on the past hydrological changes of the Mediterranean Sea (Kallel et al. 1997a; 1997b; 2000) and their relation with climatic change, which occurred in the Mediterranean area during the last two climatic cycles. In addition, we have generated a new detailed climatic record of the Holocene in a Western Mediterranean Sea core to determine the timing of the Holocene wet period.We therefore estimated variations of sea-surface temperature and salinity in the Mediterranean Sea during the last 200,000 years. Four cores recovered in the major basins were used. We first show that Mediterranean sediments record both changes observed in the North Atlantic temperature and salinity and in the local Mediterranean fresh-water budget by using a new data of a well-dated core recovered in the Tyrrhenian Sea. We then reconstruct a map of Mediterranean Sea surface salinity (SSS) at the time of the last sapropel (middle Holocene). These data, combined with the sea-surface temperature (SST) estimates, are used to propose a mode of formation for the sapropel. Finally, we have tested the validity of this model for five other sapropels, which occurred during the last 200,000 years.

The Mediterranean region is characterised by two features that make it especially interesting for palaeoclimatic studies over the last 250,000 years. On the one hand it has yielded a number of long terrestrial records, in some cases spanning with continuity more than one interglacial-glacial cycle. These complement the Mediterranean Sea records, providing a useful measure for comparison with global climatic signals. On the other hand, the palaeoclimatic evidence shows considerable complexity and sometimes conflicting features. These are likely to be a reflection of the large degree of heterogeneity that also at present characterises the geography and climate of the region.

The Mediterranean basin, which lies between 30° and 46 °N, is the largest area of the world to experience a climate of summer drought, winter rain of cyclonic origin and a mean annual temperature of 15±5 °C (Köppen type Cs). Its flora is distinctive and adapted to both periodic desiccation and burning (Allen 2001). The region also has an exceptionally long and rich history of human use and abuse, stretching back to the advent of Neolithic farming in Southwest Asia at the start of the Holocene. The complex history of culturalenvironmental relations around the Mediterranean “Lake” can create serious difficulties in distinguishing climate change from human impact in many proxy-data records (Bottema et al. 1990; Grove and Rackham 2001; Roberts 2002). In particular, once complex societies emerged during the Bronze Age between 5000 and 3000 BP, vegetation disturbance starts to become clearly visible in pollen diagrams. In compensation, the region offers a wealth of written archival and archaeological records back to before 2500 BP.

In this chapter we are concerned with PEP III time-stream 1 (cf. Gasse and Battarbee (this volume)) for the Mediterranean sector and we summarise recent progress in addressing the following key objectives:

The best available global climate information covering the last few thousand years is that obtained from the marine record, which averages worldwide effects of temperature and ice volume change, and is only marginally influenced by the localised climate changes that occur in areas less than continental in size. Consequently one of the key questions in palaeoclimate study is to understand climate changes occurring on land, and the sea-land relationships. Continental palaeoclimates have been studied using a variety of approaches such as geomorphological analyses, lake levels, archaeological studies, and pollen studies (e.g., Street and Grove (1976), Cerling et al. (1989), Gasse et al. (1990), Rossignol-Strick (1995), Weiss (2000), Magri et al. (this volume), Roberts et al. (this volume)). Recently there has been an increased interest in using the oxygen and carbon isotopic composition of speleothems as a climate proxy, because these isotopic compositions provide climate information, and U - series isotope dates give accurate chronological information. The study of the isotopic composition of speleothems has been performed throughout the world in different climatic zones in the northern hemisphere, the tropics, the Mediterranean, and the southern hemisphere (e.g., Dorale et al. (1992, 1998), Holmgren et al. (1995), Ayliffe et al. (1998), Burns et al. (1998, 2001), Bar-Matthews et al. (1999, 2000), Denniston et al. (1999), Frumkin et al. (1999), Lauritzen and Lundberg (1999), McDermott et al. (1999), Williams et al. (1999), Musgrove et al. (2001)).

In this paper we integrate studies of the present-day parameters such as average cave and air temperature, rainfall amount and its isotopic composition, the isotopic composition of the sea-surface source, and of the host rock and soil, in order to understand how the isotopic composition of speleothems recorded palaeoclimate conditions during the last 185 kyr.

Key criteria when constructing regional summaries of the sequence and patterns of climate change during the last interglacial-glacial cycle are: (i) well-dated, high resolution palaeoenvironmental data-sets based on multi-disciplinary investigations; (ii) the construction of sophisticated databases able to store, and facilitate analysis of, such complex data-sets; and (iii) the development of closer links between the ‘palaeo-data’and global climate modelling communities, in order to test ideas about the nature and causes of abrupt climate changes. In this review we examine the nature of the palaeo-environmental data generated by studies conducted within the Mid-Latitude Europe sector of PEP III. The review is restricted to Time Stream 2 only (Gasse and Battarbee, this volume).

In this chapter we outline the nature and range of the records, and the clarity and timing of the climatic signal. After brief consideration of the background record of change furnished by the ice and ocean-core records we deal with two of the oldest terrestrial sources of proxy climate records from our transect area, peat and pollen data, and then consider annuallyresolved records from lakes, tree rings and speleothems, before summarising changes from the Alps. We highlight evidence from key sites with high quality proxy records, rather than attempting to synthesise a Europe-wide picture, which would be premature, and needs further refinement of site chronologies. Existing compilations of palaeoecological data for Europe, such as Berglund et al. (1996), illustrate the size of such a task.

In this chapter special reference will be made to the south-eastern sector of the Scandinavian Ice Sheet, which extends from Finland to the Valdai-Vologda area of the Russian Plain, into which region the ice advanced during the LateWeichselian. This is the area where much new data have been collected in recent years from terrestrial sequences, and where the history of glacier fluctuations during the Valdai glaciation can now be reconstructed in considerably more detail than hitherto (Lunkka et al. 2001; Saarnisto and Saarinen 2001). Data obtained from the SE sector of the Scandinavian Ice Sheet provide more precise dates on the growth and decay of ice better than are available from other parts of Eurasia. The aim of the present chapter is: (i) to describe the fluctuations of Eurasian glaciers since the last (i.e., Eemian/Mikulino) interglacial and assess their influence on palaeohydrology; (ii) to summarise biostratigraphical data available for the last interglacial/glacial cycle in that area; and (iii) to compare these data with other proxy data on climate.

To improve our understanding of climate variability on decadal to centennial time-scales, it is crucial to use a hierarchy of climate models in addition to palaeoclimate reconstructions based on proxy data. Climate models give a physically consistent overview of the global climate on all time-scales. They are useful tools in palaeoclimatology, since: (i) they can be used to test hypotheses that have been inferred from palaeo-data; and (ii) they can provide plausible explanations of observed phenomena (e.g., Isarin and Renssen (1999), Kohfeld and Harrison (2000)). In recent years, considerable progress in palaeoclimate modelling has been made with the extensive use of models that consider the coupling of the different components of the climate system (atmosphere, ocean, sea-ice, vegetation). The aim of this paper is to inform the palaeo-data community on recent developments in palaeoclimate modelling, with special reference to the Holocene climate. In the first section, different model types and experiments are discussed, together with a short overview of Holocene climate modelling studies and differences between models and palaeo-data. In the second section, three important issues are further illustrated by discussing in detail three studies that use state-of-the-art models.

In this paper, we will present preliminary results from the working group on coupled experiments. We first describe the boundary conditions and the models (section 2) and some aspects of the simulated mid-Holocene changes (section 3) as shown by both the basic PMIP experiments and the coupled OAGCM experiments. In sections 4 and 5, we present the data-sets and the methodology used for the model-data comparisons. We then focus on the evaluation of the coupled simulations over Northern Africa (section 6) and compare them with the basic PMIPAGCMsimulations. Finally we discuss the implications of these results for the future.

In this paper we provide an example of the alternative approach, based on the Banyoro people in the humid, Interlacustrine region of Western Uganda known historically as Kitara, the region roughly centred on Mubende and encompassing much of what was regarded as Bunyoro, Toro, Nkore, and perhaps western parts of Buganda in the late nineteenth century (Sutton 1993). Kitara has the advantage of being a locus for a range of relevant, multi-disciplinary information, sources of which include rich oral histories, archaeological excavations at a comparatively advanced state and sediment-based, palaeoenvironmental studies.

Africa’s environment is closely linked with its climate, so that climatic constraints have been a major force in the development of vegetation, soils, agriculture and general livelihood (Nicholson 2001). The African continent, one of the most vulnerable regions to climate change, is subject to frequent droughts and famine. These events reflect the large range of climatic variability that envelops mean trends in the major climatic parameters such as temperature and precipitation. Africa will experience the effects of the human-induced changes in climate, but much work remains to be done in trying to isolate those aspects of African climate variability that are natural from those that are related to human influences (Hulme et al. 2001). However, the current economic decline caused by high rate of population growth, inefficient resource use, weak institutional capacity, inadequate human resources, low levels of investment and savings and a general decline in income and living standards is expected to impair Africa’s capacity to respond effectively to disruptions emanating from climate change (Ottichilo et al. 1991). Nonetheless, effective responses to climate change impacts can only evolve from an understanding of the driving forces of climate change and from effective prediction of the future state, not only on seasonal or annual time-scales, but also on inter-annual and decadal timescales. The short instrumental record in Africa (mainly from the late 1880s to present) does not provide an adequately long time-series to capture and understand fully the range of natural climate variability, nor the frequency and intensity of unique events such as El Niño. It is an impediment to the recognition and understanding of the complex workings and interactions of long-term (decadal, inter-decadal or centennial scale) features of the climate system. For this, we have to turn to the palaeo-records archived within the continent.

In each of the six major study regions constituting the PEP III Europe-Africa transect, the principal trend of Holocene climate change reflects the global climate forcing that is exerted by variation in the seasonal distribution of solar insolation received at the Earth surface, due to precession of the Earth’s orbit around the Sun. In the Northern Hemisphere, summer insolation was above-average between about 15 and 5 kyr (kyr here denotes 103 calendar years ago), peaked at 10 kyr, and is currently near its minimum; in the Southern Hemisphere, summer insolation has been above average for the past 5 kyr, and is currently near its peak. However, the regional expression of this forcing is modulated by a multitude of amplification, damping, and feedback processes involving all four components of the climate system: atmosphere, ocean, continents, and cryosphere. Consequently, regional histories of Holocene temperature and rainfall change often deviate significantly from the perfect hemispheric anti-phasing which orbital forcing would predict. In addition, superimposed on the long-term climate trends attributed to orbital forcing are various modes of Holocene climate variability operating at inter-annual to millennial time scales. Some of these can be linked to other external forcing mechanisms, such as volcanic eruptions and variations in the radiation output of the Sun, but others appear to result from poorly understood periodicity in the internal dynamics of the climate system. These processes have generated tremendous complexity in Holocene climate history at both regional and continental scales, so that the challenge to document this complexity in enough detail to elucidate the exact mechanisms involved constantly strains the possibilities of available climate-reconstruction methods. This Time Stream 1 synthesis chapter serves two purposes. First, a summary of the main patterns of Holocene climate change across the PEP III transect aims to draw attention to the (real or apparent) synchrony and time-lags between distinct climatic anomalies in the different regions, in the hope that it may help generate the sequence of speculation, modelling, and more detailed reconstruction efforts which typically yield exciting new insights. Second, it aims to draw attention to some of the more persistent problems in highresolution climate reconstruction, in the hope that efforts to address these problems face-on may lead to an improved methodology of climate reconstruction which, when applied to both new and existing data, will advance our understanding of exactly how the world’s climate system operates.

The PEP III Europe-Africa transect extends from the arctic fringes of NW Eurasia to South Africa. It encompasses the presently temperate sector of mid-latitude Europe, the Mediterranean region, the arid and semi-arid lands of the Sahara, Sahel and the Arabian Peninsula, and the inter-tropical belt of Africa. The palaeoenvironmental evidence available from these regions, which has been summarised in earlier chapters of this volume and which collectively spans the last 250,000 years, clearly bears the stamp of long-term global climate forcing induced by variations in solar insolation. External forcing is ultimately the reason why the Eurasian continental ice sheets waxed and waned repeatedly during the late Quaternary, and why the southerly limit of permafrost migrated southwards across mid-latitude Europe, periodically becoming degraded during warmer episodes. At the same time, pronounced fluctuations in atmospheric and soil moisture have affected the Mediterranean, desert and Sahel regions, while there is abundant evidence from every sector of the PEP III transect for marked migrations of the principal vegetation belts, as well as for other major environmental changes, that are also considered to reflect long-term climate forcing. It is only in the last decade or so, however, that the full complexity of the history of climate changes during the last interglacial-glacial cycle, and their environmental impacts in continental Europe and Africa, have begun to be recognised. The discovery of evidence for the abrupt Dansgaard-Oeschger (D-O) and Heinrich (H) climatic oscillations in Greenland ice-core (Johnsen et al. 1992) and North Atlantic (Bond et al. 1993) records, have prompted a re-examination of the continental record. This, together with a number of technical improvements in field and laboratory equipment, greater access to sites in remote and difficult terrain, diversification in the range of available palaeoecological and geochronological tools, and closer inter-disciplinary collaboration, have led to a more penetrating examination of the field evidence, which has progressed the science considerably. We can now see that the stratigraphical record is much more complex than appreciated hitherto, and more detailed and refined models of past climatic and environmental models are beginning to emerge. There is, for example, a growing body of evidence which suggests that D-O and H events had significant impacts on the environment of Europe and Africa, as well as on the Mediterranean Sea.