The components of the precession of the equinoxes. (a) The precession of the Earth’s axis of rotation. (b) The precession of the Earth’s orbit. (c) The precession of the equinoxes (adapted from Wilson et al., 2000; Maslin et al., 2001). 

Contexts in source publication

Context 1

... This problem of time greatly complicates paleoclimate reconstructions. There are long-term processes, such as tectonic uplift and global climate change, medium-scale changes, such as orbital forcing, short-term changes such as the appearance and disappearance of lakes and changes in vegetation, and extremely short-term changes such as switching on or off of El Ni ñ o-Southern Oscillation (ENSO) as associated with Walker Circulation. In terms of scale, there are processes that affect climate globally (e.g., glacial-interglacial cycles or changes in atmospheric carbon dioxide), regional scale changes such as tectonic uplift and basin formation, and local changes that may involve changes in local vegetation or a single drainage network. Great care must be taken when relating evolution to reconstructed environmental changes that the appropriate temporal and spatial scale is used. For example, changes in one lake may not have af- fected evolution if a population could migrate to another nearby lake. However, if all the lakes in a region dried out at the same time, then this may have had a significant impact on the local population and may have influenced evolution (Trauth et al., 2007) or may have stimulated larger scale migration. Hughes et al. (2007) have used computer models to assess the influence of regional climate change on the patterns of hominin migration to investigate this problem. Long-term climate change seems to be primarily modulated by tectonic changes at the global and local scale (Hay, 1992; Maslin et al., 2001). The first major continental ice sheets formed on Antarctica about 35 million years ago (megaannum, Ma) with the opening of the Tasmania-Antarctic and Drake passages (Kennett, 1996; Huber et al., 2004; Stickley et al., 2004). The resultant ‘Ice House’ climate mode reached a zenith with late Cenozoic global cooling and Northern Hemisphere glaciation. On a regional scale, tectonics provides the physical foundation for lakes, and sometimes create the opportunity for moisture accumulation, thus playing an indirect but key role in faunal distribution. In East Africa long-term climatic change is also controlled by tectonics, with the progressive formation of the East African Rift Valley leading to increased aridity, and an increase in basins suitable for lakes. According to Trauth et al. (2005, 2007) volcanism in East Africa may have started as early as 45 e 33 Ma in the Ethiopian Rift. By 33 Ma it occurred in northern Kenya, although magmatic activity of the central and southern segments of the rift in Kenya and Tanzania did not start until between 15 and 8 Ma (e.g., Bagdasaryan et al., 1973; Crossley and Knight, 1981; McDougall and Watkins, 1988; George et al., 1998; Ebinger et al., 2000). Major faulting in Ethiopia between 20 and 14 Ma was followed by the generation of east-dipping faults in northern Kenya between 12 and 7 Ma, and superseded by normal faulting on the western side of the Central and Southern Kenya Rift between 9 and 6 Ma (Baker et al., 1988; Blisniuk and Strecker, 1990; Ebinger et al., 2000). These early half-grabens were subsequently faulted antithetically between about 5.5 and 3.7 Ma, which generated a full-graben morphology and basins appropriate for lake formation (Baker et al., 1988; Strecker et al., 1990). By 2.6 Ma, the graben was further segmented in the Central Kenya Rift by west dipping faults, creating the 30- km-wide intrarift Kinangop Plateau and the tectonically active 40 km-wide inner rift (Baker et al., 1988; Strecker et al., 1990). In the Tanzanian sector of the rift, sedimentation in isolated basins began at w 5 Ma (Foster et al., 1997). A major phase of rift faulting occurred at 1.2 Ma and produced the present-day rift escarpments (Foster et al., 1997). There is evidence from both soil carbonate (Levin et al., 2004; Wynn, 2004; Quinn et al., 2007; Sikes and Ashley, 2007) and n -alkane carbon isotopes (Feakins et al., 2005) that there was a progressive vegetation shift from C 3 to C 4 plants during the Plio-Pleistocene, correlative with active tec- tonism/faulting. This vegetation shift has been ascribed to increased aridity due to the progressive rifting of East Africa (deMenocal, 2004). Modeling work by Sepulchre et al. (2006) clearly demonstrates the influence of East African regional uplift on rainfall. They see a marked decrease in the predicted rainfall since the wind patterns become less zonal as uplift increases. Hence, as elevation increased, a rain shadow effect occurred, reducing the moisture available for rain on the eastern sides of the mountains/valleys, and producing the strong aridification seen in the paleoenvironmental records (Sepulchre et al., 2006). As well as drying out Africa, the tectonic events described above also produced numerous basins suitable for the formation of lakes. Tectonics were essential for the production of isolated basins in the East African Rift Valley within which lakes could form (see Trauth et al., 2007, their figure 5). The southward propagation of rifting including the formation of faults and the magmatic activity is also reflected in the earliest formation of lake basins in the northern parts of the rift. Many of these lakes formed long after the basins were produced. For example, the geological record of lakes in the Afar, Omo-Turkana, and Baringo-Bogoria basins in the north begins in the middle of the Miocene. The oldest lacustrine se- quences in the central and southern segments of the rift in Kenya and Tanzania are of early Pliocene age (Tiercelin and Lezzar, 2002). These ages are consistent with the general ob- servation that paleolakes in the north of the East African Rift formed earlier than in the south (see Trauth et al., 2007, their figure 5). The shift between river and lake sources of water in East Africa is important when considering hominins, as lakes offer a buffered, more permanent source of water that can sur- vive the oscillation between wet and dry seasons better than streams or other shallow bodies of water. While the late Cenozoic tectonic history of East Africa is well-constrained, that is not the case for southern Africa. There are no long, continuous Plio-Pleistocene records of continental sedimentation. The Tswaing Crater is the only relatively long-lived basin appropriate to accumulation of water and, thus, sediment; however, this basin had alternately wet/ dry cycles and so has limited and likely discontinuous stratigraphic resolution. Although continuously cored, the Tswaing record spans only a short interval ( w 200 ka to present; Partridge et al., 1997). While it offers a useful but brief glimpse into past climate (e.g., Scott, 2002), the Tswaing yields no tectonic information. The region lacks also the volcanic (basalt and ash) record that characterizes East Africa and that permits a radiometri- cally based chronology. Instead, most of the southern African continental evidence is derived from regional geomorphological features associated with tectonic events. Geomorphological features provide key Pliocene climatic data (e.g., terraces of the Vaal River indicating stream rejuvenation; de Wit et al., 2000), but they lack independent age control. Pan deposits, such as those at Florisbad, provide crucial insight into variations in continental climate, but are not continuous and are difficult to date. As in most continental regions, weathering and erosion tend to destroy the evidence, but it is even more difficult for Plio-Pleistocene southern Africa because uplift has led to an extensive erosional surface extending across much of southern Africa (de Wit et al., 2000). Further complicating the attempt to reconstruct continental climate in this region is the fact that unlike East Africa, where lakes dominate, the major repositories for Plio-Pleistocene age sediments in southern Africa are cave sites. Because they have a complex stratigraphy further obscured by quarrying, these caves are not unequivocally dated. Thus, the record of southern African continental climate and inferred tectonic changes come primarily from associations between geomorphological events and global climate changes, observed oceanic changes, and biostratigraphic comparisons between eastern and southern African faunas. For example, uplift of the southeastern and eastern hinterland regions, including East Africa and Zimbabwe, created rain shadow areas to the west of these regions (Tyson and Preston-White, 2000) and is thought to have enhanced aridification. While the timing of this uplift is not well-constrained, they reason it occurred in association with the late Pliocene increase in Northern Hemisphere Glaciation (INHG) based on indicators of northern African aridification. This, of course, raises the problem of circular arguments with no independent measure of tectonic change. There are two vital pieces of information that we do not know concerning regional tectonics in Africa. First is the exact timing and altitude of the uplift. Much excellent work has been done on the timing of key tectonic features, but uplift rates and maximum altitude are still unconstrained, especially for southern Africa. These factors are important because they control local rainfall patterns. So although we know that progressive uplift and rifting has caused East Africa to dry, we do not know precisely when, or at what pace, these changes occurred. The second gap in our knowledge is the effect tectonics had on vegetation, which is crucial knowledge for understanding hominin evolution. For example, was there rainforest in East Africa 10 Ma, before significant doming began? When did the forest fragment? When did grasslands become important? Was there a vegetative corridor between southern Africa and East Africa that may have facilitated dispersal and exchange between populations? At the moment, we have detailed knowledge of vegetation and environmental conditions at sites containing hominin remains for East Africa, and generally less ...

Context 2

... care must be taken when relating evolution to reconstructed environmental changes that the appropriate temporal and spatial scale is used. For example, changes in one lake may not have af- fected evolution if a population could migrate to another nearby lake. However, if all the lakes in a region dried out at the same time, then this may have had a significant impact on the local population and may have influenced evolution (Trauth et al., 2007) or may have stimulated larger scale migration. Hughes et al. (2007) have used computer models to assess the influence of regional climate change on the patterns of hominin migration to investigate this problem. Long-term climate change seems to be primarily modulated by tectonic changes at the global and local scale (Hay, 1992; Maslin et al., 2001). The first major continental ice sheets formed on Antarctica about 35 million years ago (megaannum, Ma) with the opening of the Tasmania-Antarctic and Drake passages (Kennett, 1996; Huber et al., 2004; Stickley et al., 2004). The resultant ‘Ice House’ climate mode reached a zenith with late Cenozoic global cooling and Northern Hemisphere glaciation. On a regional scale, tectonics provides the physical foundation for lakes, and sometimes create the opportunity for moisture accumulation, thus playing an indirect but key role in faunal distribution. In East Africa long-term climatic change is also controlled by tectonics, with the progressive formation of the East African Rift Valley leading to increased aridity, and an increase in basins suitable for lakes. According to Trauth et al. (2005, 2007) volcanism in East Africa may have started as early as 45 e 33 Ma in the Ethiopian Rift. By 33 Ma it occurred in northern Kenya, although magmatic activity of the central and southern segments of the rift in Kenya and Tanzania did not start until between 15 and 8 Ma (e.g., Bagdasaryan et al., 1973; Crossley and Knight, 1981; McDougall and Watkins, 1988; George et al., 1998; Ebinger et al., 2000). Major faulting in Ethiopia between 20 and 14 Ma was followed by the generation of east-dipping faults in northern Kenya between 12 and 7 Ma, and superseded by normal faulting on the western side of the Central and Southern Kenya Rift between 9 and 6 Ma (Baker et al., 1988; Blisniuk and Strecker, 1990; Ebinger et al., 2000). These early half-grabens were subsequently faulted antithetically between about 5.5 and 3.7 Ma, which generated a full-graben morphology and basins appropriate for lake formation (Baker et al., 1988; Strecker et al., 1990). By 2.6 Ma, the graben was further segmented in the Central Kenya Rift by west dipping faults, creating the 30- km-wide intrarift Kinangop Plateau and the tectonically active 40 km-wide inner rift (Baker et al., 1988; Strecker et al., 1990). In the Tanzanian sector of the rift, sedimentation in isolated basins began at w 5 Ma (Foster et al., 1997). A major phase of rift faulting occurred at 1.2 Ma and produced the present-day rift escarpments (Foster et al., 1997). There is evidence from both soil carbonate (Levin et al., 2004; Wynn, 2004; Quinn et al., 2007; Sikes and Ashley, 2007) and n -alkane carbon isotopes (Feakins et al., 2005) that there was a progressive vegetation shift from C 3 to C 4 plants during the Plio-Pleistocene, correlative with active tec- tonism/faulting. This vegetation shift has been ascribed to increased aridity due to the progressive rifting of East Africa (deMenocal, 2004). Modeling work by Sepulchre et al. (2006) clearly demonstrates the influence of East African regional uplift on rainfall. They see a marked decrease in the predicted rainfall since the wind patterns become less zonal as uplift increases. Hence, as elevation increased, a rain shadow effect occurred, reducing the moisture available for rain on the eastern sides of the mountains/valleys, and producing the strong aridification seen in the paleoenvironmental records (Sepulchre et al., 2006). As well as drying out Africa, the tectonic events described above also produced numerous basins suitable for the formation of lakes. Tectonics were essential for the production of isolated basins in the East African Rift Valley within which lakes could form (see Trauth et al., 2007, their figure 5). The southward propagation of rifting including the formation of faults and the magmatic activity is also reflected in the earliest formation of lake basins in the northern parts of the rift. Many of these lakes formed long after the basins were produced. For example, the geological record of lakes in the Afar, Omo-Turkana, and Baringo-Bogoria basins in the north begins in the middle of the Miocene. The oldest lacustrine se- quences in the central and southern segments of the rift in Kenya and Tanzania are of early Pliocene age (Tiercelin and Lezzar, 2002). These ages are consistent with the general ob- servation that paleolakes in the north of the East African Rift formed earlier than in the south (see Trauth et al., 2007, their figure 5). The shift between river and lake sources of water in East Africa is important when considering hominins, as lakes offer a buffered, more permanent source of water that can sur- vive the oscillation between wet and dry seasons better than streams or other shallow bodies of water. While the late Cenozoic tectonic history of East Africa is well-constrained, that is not the case for southern Africa. There are no long, continuous Plio-Pleistocene records of continental sedimentation. The Tswaing Crater is the only relatively long-lived basin appropriate to accumulation of water and, thus, sediment; however, this basin had alternately wet/ dry cycles and so has limited and likely discontinuous stratigraphic resolution. Although continuously cored, the Tswaing record spans only a short interval ( w 200 ka to present; Partridge et al., 1997). While it offers a useful but brief glimpse into past climate (e.g., Scott, 2002), the Tswaing yields no tectonic information. The region lacks also the volcanic (basalt and ash) record that characterizes East Africa and that permits a radiometri- cally based chronology. Instead, most of the southern African continental evidence is derived from regional geomorphological features associated with tectonic events. Geomorphological features provide key Pliocene climatic data (e.g., terraces of the Vaal River indicating stream rejuvenation; de Wit et al., 2000), but they lack independent age control. Pan deposits, such as those at Florisbad, provide crucial insight into variations in continental climate, but are not continuous and are difficult to date. As in most continental regions, weathering and erosion tend to destroy the evidence, but it is even more difficult for Plio-Pleistocene southern Africa because uplift has led to an extensive erosional surface extending across much of southern Africa (de Wit et al., 2000). Further complicating the attempt to reconstruct continental climate in this region is the fact that unlike East Africa, where lakes dominate, the major repositories for Plio-Pleistocene age sediments in southern Africa are cave sites. Because they have a complex stratigraphy further obscured by quarrying, these caves are not unequivocally dated. Thus, the record of southern African continental climate and inferred tectonic changes come primarily from associations between geomorphological events and global climate changes, observed oceanic changes, and biostratigraphic comparisons between eastern and southern African faunas. For example, uplift of the southeastern and eastern hinterland regions, including East Africa and Zimbabwe, created rain shadow areas to the west of these regions (Tyson and Preston-White, 2000) and is thought to have enhanced aridification. While the timing of this uplift is not well-constrained, they reason it occurred in association with the late Pliocene increase in Northern Hemisphere Glaciation (INHG) based on indicators of northern African aridification. This, of course, raises the problem of circular arguments with no independent measure of tectonic change. There are two vital pieces of information that we do not know concerning regional tectonics in Africa. First is the exact timing and altitude of the uplift. Much excellent work has been done on the timing of key tectonic features, but uplift rates and maximum altitude are still unconstrained, especially for southern Africa. These factors are important because they control local rainfall patterns. So although we know that progressive uplift and rifting has caused East Africa to dry, we do not know precisely when, or at what pace, these changes occurred. The second gap in our knowledge is the effect tectonics had on vegetation, which is crucial knowledge for understanding hominin evolution. For example, was there rainforest in East Africa 10 Ma, before significant doming began? When did the forest fragment? When did grasslands become important? Was there a vegetative corridor between southern Africa and East Africa that may have facilitated dispersal and exchange between populations? At the moment, we have detailed knowledge of vegetation and environmental conditions at sites containing hominin remains for East Africa, and generally less information for sites in South Africa. However, these only provide information on the niches that our ancestors were inhabiting, and not the wider environment. A broader perspective is critical and it is being achieved through both climate modeling and paleoenvironmental reconstructions. The modeling paper by Sepulchre et al. (2006) is an excellent approach to the above problems providing reconstruction of climate and vegetation, albeit with a greatly reduced relief over eastern Africa. The next step will be to produce detailed time-slices through the last eight million years with both relief and global climate applied to a regional climate and vegetation model. Another novel approach to link local environmental niches with the wider ...

Context 3

... The second gap in our knowledge is the effect tectonics had on vegetation, which is crucial knowledge for understanding hominin evolution. For example, was there rainforest in East Africa 10 Ma, before significant doming began? When did the forest fragment? When did grasslands become important? Was there a vegetative corridor between southern Africa and East Africa that may have facilitated dispersal and exchange between populations? At the moment, we have detailed knowledge of vegetation and environmental conditions at sites containing hominin remains for East Africa, and generally less information for sites in South Africa. However, these only provide information on the niches that our ancestors were inhabiting, and not the wider environment. A broader perspective is critical and it is being achieved through both climate modeling and paleoenvironmental reconstructions. The modeling paper by Sepulchre et al. (2006) is an excellent approach to the above problems providing reconstruction of climate and vegetation, albeit with a greatly reduced relief over eastern Africa. The next step will be to produce detailed time-slices through the last eight million years with both relief and global climate applied to a regional climate and vegetation model. Another novel approach to link local environmental niches with the wider regional or global climate is employed in this volume by Lee-Thorp et al. (2007), who analyze 13 C/ 12 C ratios in fossil tooth enamel to provide a semi-quanti- tative measure of ‘open’ vs. ‘closed’ habitats. However, this technique is limited to the occurrence of mammalian fossils. On the other hand, the methodology provides a nice opportunity to compare results with morphometric faunal analyses (e.g., Reynolds, 2007). Two further ways to reconstruct past vegetation and, thus, climate are: 1) long continuous lake cores; for example, from Lake Malawi, and 2) the use of coupled climate-vegetation models. Additionally, innovative ap- proaches are being employed to better constrain the age of South African cave deposits and the climatic events contained therein (e.g., this volume, Pickering et al., 2007; Hopley et al., 2007). As the stratigraphy of these important deposits is better understood, it may be possible to derive a tectonic record from these climate archives. The oscillation between glacial and interglacial climates is the most fundamental environmental characteristic of the Quaternary Period and is believed to be primarily forced by changes in the Earth’s orbital parameters (Hays et al., 1976). However, there is not always a direct cause-and-effect relationship between climate cycles and these parameters due to feedback mechanisms internal to the Earth’s climate system. For example, the insolation received at the critical latitude of 65 N during the Last Glacial Maximum (LGM) 18,000 years ago was very similar to today (Laskar, 1990; Berger and Loutre, 1991). The three main orbital parameters d eccentricity, obliquity (tilt), and precession d are each described below and their combined effects discussed. The shape of the Earth’s orbit changes from nearly circular to an ellipse over a period of about 96,000 and 125,000 years, with a long cycle of about 400,000 years. As a result of this change in orbit, the long axis of the ellipse varies in length over time; this is referred to as eccentricity (Fig. 4). Today, the Earth is at its closest (146 million km) to the sun on January 3 rd : this position is known as perihelion. On July 4 th it is at its most distant from the sun (156 million km) at the aphelion. Changes in eccentricity cause only very minor variations, approximately 0.03%, in total annual insolation, but can have significant seasonal effects. If the orbit of the Earth were perfectly circular there would be no seasonal variation in total incoming solar insolation. Hence, the major effect of eccentricity is to modulate the precessional effects (see below). The tilt or obliquity of the Earth’s axis of rotation with re- spect to the plane of its orbit (the plane of the ecliptic) varies between 21.8 and 24.4 over a period of 41,000 years (Fig. 4). It is the tilt of the axis of rotation that gives us seasons; in summer the hemisphere that is tilted towards the Sun is warmer because it receives more than 12 hours of sunlight, and the Sun is higher in the sky. At the same time, in the opposite hemisphere the axis of rotation is tilted away from the Sun and that region is plunged into winter, when it is colder and receives less than 12 hours of sunlight, and the Sun is lower in the sky. This changes the distribution of the incoming solar radiation, and so the larger the obliquity the greater the difference between summer and winter temperatures. There are two components of precession: one relates to its axis of rotation and the other to the elliptical orbit of the Earth. The Earth’s rotational axis wobbles around the whole of the Earth’s orbit; that is, it precesses every 27,000 years (Fig. 5). This is similar to the gyrations of the rotational axis of a spinning toy top, when the central plunger spins but at a much slower rate than the plane of the spinning top. This precession of the axis of rotation causes the position of the Earth relative to the Sun during any particular season to alter. In addition, the whole of the Earth’s orbit precesses (i.e., it swings around in space) with a periodicity of 105,000 years. It is the combination of the different orbital parameters that results in the classically quoted precessional periodicities of 23,000 and 19,000 years. Combining the precession of the axis of rotation (27,000 years) and the precessional changes in Earth’s orbit (105,000 years) produces a period of 23,000 years. Combining the shape of the orbit (i.e., eccentricity, 96,000 years) with the precession of the axis of rotation (27,000 years) results in a period of 19,000 years. These two periodicities combine so that perihelion coincides with the summer season in each hemisphere on average every 21,700 years, resulting in the precession of the equinoxes. Precession has the most significant impact in the tropics, (in contrast to the impact of obliquity at the equator, which is zero). So although obliquity clearly influences high-latitude climate change, which may ultimately influence the tropics, direct effects of insolation in the tropics are due to eccentricity modulated precession. Combining the effects of eccentricity, obliquity, and precession provides the means for calculating the insolation at any latitude back through time (e.g., Milankovitch, 1949; Berger and Loutre, 1991). Changes in insolation may not seem significant until one considers that the maximum difference in solar radiation in the last 600,000 years (Fig. 6) is equivalent to the difference between the amount of summer radiation received today at 65 N and that received at 77 N, over 550 km to the north. Continuing our oversimplification, this would bring the current glacial limit from mid-Norway down to the latitude of Scotland. However, as we have seen, each of the orbital parameters has a different effect with changing latitude. The conventional view of glaciation is that low summer insolation in the temperate Northern Hemisphere allows ice to persist through summer and to build-up on the northern continents (Milankovitch, 1949; and see detailed references in Maslin et al., 2001). However, orbital forcing in itself is insufficient to drive the observed glacial-interglacial variability in climate. Instead, the Earth system amplifies and transforms the changes in insolation at the surface through various feedback mechanisms. As snow and ice accumulate due to initial changes in insolation regime, the ambient environment is modified. This is primarily by an increase in the amount of sunlight reflected back into space by the accumulating ice (i.e., the increased albedo), and with it a reduction in the absorption efficiency of incident solar radiation, and, thus, a suppression of local temperatures. This process promotes the accumulation of more snow and ice and, thus, a further modification of the ambient environment, and so on d the classic ‘ice albedo’ feedback (Fig. 7). Another feedback is triggered when the ice sheets, particularly the Laurentide ice sheet on North America, become large enough to deflect the atmospheric planetary waves. This deflection changes the storm path across the North Atlantic Ocean and prevents Gulf Stream and North Atlantic Drift from penetrating as far north as today. Such surface ocean change, combined with the general increase in melt-water in the Nordic Seas and Atlantic Ocean caused by the presence of large continental ice sheets, ultimately leads to reduction in the production of deep water. This reduction in turn lessens the amount of warm water pulled northwards. All of this leads to increased cooling in the Northern Hemisphere and expansion of the ice sheets (see detailed references in Maslin et al., 2001). However, the action of these and similar ‘physical climate’ feedbacks on variations in solar insolation received at the Earth’s surface is not sufficient to allow a complete accounting of the timing and magnitude of the glacial-interglacial cycles, at least after the Mid-Pleistocene Revolution. There is mount- ing evidence that other feedback mechanisms may be equally, if not more, important in driving the long-term climate system. The role of ‘greenhouse’ gases in the atmosphere that absorb outgoing infrared radiation is critical (e.g., Berger, 1988; Berger and Loutre, 1991; Saltzman et al., 1993; Li et al., 1998). Any reduction in the atmospheric concentration of con- stituents, such as CO 2 , CH 4, and water vapor, will drive a general global cooling (Fig. 7), which in turn furthers glaciation. We already know that these properties varied considerably over the glacial-interglacial cycles. For CO 2 and CH 4 this variability is recorded in air bubbles trapped in polar ...

Context 4

... critical maximum degree of glaciation, such as the LGM. So although eccentricity (and obliquity) determines the envelope of precessional amplitude and, thus, ultimately whether the minimum occurs on the fourth or fifth precessional cycle, the MPR still does not represent the onset of nonlinear amplification of eccentricity as has been suggested by some authors. The MPR had a significant effect on African climate. S ́galen et al. (2007) has concluded that C 4 grasses apparently remained a relatively minor component of African environments from their Miocene re-evolution until the late Pliocene and early Pleistocene. Pedogenic carbonate d 13 C data from ex- isting localities in East Africa suggest that open ecosystems dominated by C 4 grass components emerged only during the MPR (i.e., after 1 Ma). Schefuß et al. (2003) have reconstructed the relative abundance of C 3 and C 4 plants using bio- markers in marine sediments off West Africa. They find that the percentage of C 4 plants follows the enhanced glacial-interglacial cycles after the MPR and is directly influenced by the SSTs off West Africa; a change in plant type would have had a strong impact on faunal distributions. They also find no evidence that the climate of the Congo region is driven by precessional forcing. There is also growing evidence for the formation of large lakes between 1.1 and 0.9 Ma in East Africa, such as in the Olorgesailie Formation, the Naivasha and Elmenteita-Nakuru Basins, and the Afar Basin (Trauth et al., 2005, 2007); the long-term availability of water would have further impacted fauna. The major issue for understanding the influence of global climate transitions on African climate is the lack of high-resolution continental records. This problem is particularly acute for southern Africa. The terrestrial realm is severely restricted in the types of proxies that can be utilized as well as the ability to constrain the ages of the sediments since in many cases the original record is removed through processes associated with sub aerial exposure (e.g., Lowe and Walker, 1984). The continental records that have been published are not generally able to provide the same level of continuous, detailed climate information as oceanic records. At present, only lakes and caves provide continuous sections for East and South Africa. Lakes are present in East Africa, and are a major source of data for this volume (Kingston et al., 2007; Trauth et al., 2007), and Lake Malawi may provide a continuous Plio-Pleistocene record when drilled in the future. Caves are present in southern Africa and, although the cave deposits have yielded abundant specimens, the stratigraphy of the caves is complex (e.g., Pickering et al., 2007) and the paleoenvironmental analyses can be of limited value (e.g., Scott, 2002). Most of these sites were quarried first and analyzed later, thus severely altering stratigraphic control. Finally, in many cases there is limited stratigraphic control and some records are dated based on the assumption of cause/association with global events rather than on an independent chronology. For example, the evidence for increased aridity associated with the INHG, was derived, as outlined in Partridge (1993), from geomorphological and biostratigraphic datasets that do not enjoy independence from one another, and cannot be dated with precision. Hence, well-dated, high-resolution, continuous continental records from Africa are required if we are to fully understand the effects of global climate transitions on all of African climate and mammal evolution. The DWC also provides another interesting twist on African climate; namely, that only once a strong east-west temperature gradient is established in the Pacific Ocean can El Ni ñ o South Oscillation (ENSO) operate. There is strong documen- tary evidence that ENSO has a large influence on the climate of modern East Africa. As natural selection acts on individuals then extreme annual climates driven by ENSO may have had a major impact on human evolution. If so, then we need to understand when ENSO began to influence East Africa and how it has changed through the Plio-Pleistocene. On time scales of more than 100,000 years, rift-related volcano-tectonic processes shaped the landscape of East Africa and profoundly influenced local climate and surface hydrology through the development of relief. Through uplift of the Kenyan and Ethiopian plateaus, changes in orography and associated rain shadow are believed to be the major driving force for increased variability of moisture availability throughout eastern Africa. This increased sensitivity has resulted in a modern Rift Valley that hydrological modelling suggests could support lakes as deep as 150 m with an annual precipitation increase of only 15 e 30% (Bergner et al., 2003). Trauth et al. (2005, 2007) have documented three major Late Cenozoic lake periods in East Africa: 2.7 e 2.5 Ma, 1.9 e 1.7 Ma, and 1.1 e 0.9 Ma (see Trauth et al., 2007, their figure 5). Although preservation of East African lake records prior to 2.7 Ma is patchy, there is limited evidence for lake phases at w 2.95 e 3.20 Ma, w 3.3 e 3.4 Ma, 3.9 e 4.0 Ma, and w 4.3 e 4.7 Ma (Trauth et al., 2007). These lake phases correspond to drops in the East Mediterranean marine dust abundance (Larrasoa ñ a et al., 2003), which is thought to reflect the aridity of the eastern Algerian, Libyan, and western Egyptian lowlands located north of the central Saharan watershed (see Trauth et al., 2007, their figure 6). The lake phases also correspond to an increased occurrence of sapropels in the Mediterranean Sea, which are thought to be caused by increased Nile River discharge (Lourens et al., 2004). The correspondence of the Mediterranean marine records with lake records of East Africa suggest a consistent moisture record for a region encompassing much of central and northern Africa over the last three to five million years. In contrast, these East African wet phases correlate with significant intermediate-term increases in the dust records from ocean sediment cores adjacent to West Africa and Arabia (deMenocal, 1995, 2004). While this at first seems contradic- tory, examination of these data in greater chronological detail demonstrates that both the lake and dust records are responding to precessional forcing and that these records are in-phase (deMenocal, 1995, 2004; Deino et al., 2006; Kingston et al., 2007; Lepre et al., 2007). Hence, the lake records from East Africa and the Indian Ocean dust records document extreme climate variability with precessionally forced wet and dry phases. The late Cenozoic periods of extreme climate variability appear to correlate with maxima in the 400-kyr component of the Earth’s eccentricity cycle. Prior to 2.7 Ma the wet phases appear every 400 kyr (Trauth et al., 2007). After 2.7 Ma, however, the wet phases appear every 800 kyr, with periods of precession forced extreme climate variability at 2.7 e 2.5 Ma, 1.9 e 1.7 Ma, and 1.1 e 0.9 Ma, whereas other periods of eccentricity maxima at w 2.2 Ma, w 1.4 Ma, and w 0.6 Ma are not associated with the alternating formation of large lakes or increased dust. The three late Cenozoic lake phases do, however, correlate with significant global climatic transitions as well as peaks in eccentricity. Hence, after 2.7 Ma global climate changes seem to be required to cause an increased regional climate sensitivity to precessional forced insolation and increased seasonality, which allows either large deep lakes to develop or causes extreme aridity and large dust loads to the adjacent oceans. In contrast, prior to the 2.7 Ma eccentricity maxima alone were sufficient to produce regional sensitivity. It remains to be evaluated whether the long-term drying trend in East Africa or the global cooling trend is responsible for this shift from a simple linear response to long-term eccentricity forcing. The last three major Plio-Pleistocene lake phases all correspond to global climate transitions. The lake phase at 2.7 e 2.5 Ma corresponds to INHG, 1.9 e 1.7 Ma to a DWC, and 1.1 e 0.9 Ma to initiation of the MPR. Each of these global climate transitions was accompanied by reduced North Atlantic Deep Water (NADW) formation (Haug and Tiedemann, 1998) and increased ice rafting from both Greenland and Antarctica (Cowan, 2001; St. John and Krissek, 2002). Ice expansion in both hemispheres would have significantly increased the Pole-Equator thermal gradient, leading to north-south compression of the Intertropical Convergence Zone (ITCZ). A similar effect occurred during the Last Glacial Maximum when a strong compression of the ITCZ is observed both in pa- leo-reconstructions of tropical hydrology (e.g., Peterson et al., 2000; Chiang et al., 2003; Wang et al., 2004) and via climate modelling (Lautenschlager and Herterich, 1990; Bush and Philander, 1999; Bush, 2001). Trauth et al. (2007) have suggested that the compression of the ITCZ is essential to increasing the sensitivity of East Africa to precessional forcing of moisture availability; otherwise moisture is transported north and south away from the Rift Valley. Along the whole length of the rift, without this high-latitude climate control, East Africa cannot receive enough rainfall to fill large deep freshwater lakes during positive precessional periods. After 3 Ma, both global climate forcing and eccentricity maxima are required to generate episodes of extreme precessional forced climate. Hence, tectonics, high latitude, and low-latitude forcing are required to explain the highly variable climate of Africa over the Plio-Pleistocene. The relationship between climate and human evolution seems intuitive and, indeed, environmental factors have been suggested as a driving force in hominin evolution by many authors (See Kingston et al., 2007, for detailed history). As discussed above, Vrba (1985) linked global climate change as a cause of African mammalian evolution by documenting radiations ...

Context 5

... well as drying out Africa, the tectonic events described above also produced numerous basins suitable for the forma- tion of lakes. Tectonics were essential for the production of isolated basins in the East African Rift Valley within which lakes could form (see Trauth et al., 2007, their figure 5). The southward propagation of rifting including the formation of faults and the magmatic activity is also reflected in the earliest formation of lake basins in the northern parts of the rift. ...

Context 6

... oldest lacustrine se- quences in the central and southern segments of the rift in Kenya and Tanzania are of early Pliocene age ( Tiercelin and Lezzar, 2002). These ages are consistent with the general ob- servation that paleolakes in the north of the East African Rift formed earlier than in the south (see Trauth et al., 2007, their figure 5). The shift between river and lake sources of water in East Africa is important when considering hominins, as lakes offer a buffered, more permanent source of water that can sur- vive the oscillation between wet and dry seasons better than streams or other shallow bodies of water. ...

Context 7

... are two components of precession: one relates to its axis of rotation and the other to the elliptical orbit of the Earth. The Earth's rotational axis wobbles around the whole of the Earth's orbit; that is, it precesses every 27,000 years ( Fig. 5). This is similar to the gyrations of the rotational axis of a spin- ning toy top, when the central plunger spins but at a much slower rate than the plane of the spinning top. This precession of the axis of rotation causes the position of the Earth relative to the Sun during any particular season to alter. In addition, the whole of the ...

Context 8

... example, suggest that, although there was a general trend towards more open environments since 3 Ma, the most signif- icant environmental change to open, grassy landscapes oc- curred after 2 Ma rather than between 2.4 and 2.6 Ma as earlier suggested. There is also evidence for large deep lakes occurring in East Africa at about 2 Ma ( Trauth et al., 2005Trauth et al., , 2007, including the Ethiopian Rift, the Afar Basins, the Omo-Turkana Basin, the eastern flank of the southern Kenya Rift, and the Olduvai Gorge (see Trauth et al., 2007, their fig- ure 5). The DWC is also important for tropical Africa as it controls El Ni~ no-South Oscillation (ENSO). ...

Context 9

... et al. (2005Trauth et al. ( , 2007 have documented three major Late Cenozoic lake periods in East Africa: 2.7e2.5 Ma, 1.9e1.7 Ma, and 1.1e0.9 Ma (see Trauth et al., 2007, their figure 5). Although preservation of East African lake records prior to 2.7 Ma is patchy, there is limited evidence for lake phases at w2.95e 3.20 Ma, w3.3e3.4 ...