What Does Paleoclimatology Have To Do With Global Warming

TIME (KYR)

FIGURE 9.21 Fluctuations of arboreal pollen (AP) sum at the Funza site, Sabana de Bogotá, Colombia, over the past 1.5 Ma.The former lake of Bogota desiccated at -27 ka B.P., terminating this sequence.The development of a pronounced - 100 ka climatic cycle within the last I Ma is clearly seen (Hooghiemstra and Ran, 1994).

the same time interval) the forest zone was limited to elevations of less than 2000 m, below the intermontane basin. When viewed over very long periods of time (Fig. 9.21) other factors complicate this simple idea (e.g., the establishment of Quercus [which produces a lot of pollen] in the forest at around -263 ka B.P.) and it is then necessary to adjust the AP "threshold values" accordingly. Taking such factors into account, if the upper forest limits are controlled by temperature, AP variations range from -1800 m in glacial times to -3500 m in interglacials and suggest overall temperature changes on the order of -10 °C at Funza (i.e., A1700 m at a moist adiabatic lapse rate of 0.6 °C 100 m1). Whether such changes in vegetation can be interpreted simply in terms of temperature is debatable; studies of vegetation change in other montane environments suggest that the lower concentrations of carbon dioxide in glacial times may have had a particularly strong impact on vegetation at high elevations. Ice-core evidence indicates C02 levels were 80-100 ppm lower at the last glacial maximum compared to pre-industrial levels, so that at elevations above 3500 m, the partial pressure of C02 may have been <130 ppm, preventing C3 plants such as trees from surviving (Street-Perrott, 1994; Street-Perrott et al., 1997, 1998). Lower carbon dioxide levels would cause an increase in stomatal gas exchange, thereby raising transpiration rates and drought stress on plants (Jolly and Haxeltine, 1997). Those C4 plants such as grasses are generally more efficient in utilizing C02 and water, so evidence for lower treelines and more extensive savannas or grassy tundra may have less to do with temperature change than with C02 levels. It is also of interest that the overall late Cenozoic record indicates that a shift in the frequency of climatic change occurred around 1 Ma B.P., from relatively low-amplitude high-frequency variations to higher amplitude fluctuations with a period close to 100 ka (see Fig. 9.21) (Hooghiemstra et al., 1993). However, as the record was dated by tuning it to the marine isotope record, this may not be an entirely independent result.

9.7.3 Amazonia

Although it is commonly assumed that the large numbers of species of flora and fauna in the equatorial lowlands of South America are the result of long periods of stable climate, a number of biogeographical studies have cast doubt on that assumption. In a comprehensive study of different species of birds in and around the Amazon Basin, Haffer (1969, 1974) identified a number of regions he believed had been refugia for groups of birds during drier periods in the past when the extensive tropical forests of today were reduced to discrete forest enclaves separated by savanna vegetation. Forest-dwelling species, which were isolated in this way, differentiated (developed new species) independently from members of the same species, which had been separated into other forest enclaves. Haffer argued that when wetter conditions returned and the forests reoccupied the savanna region, forest-dwelling species also expanded their ranges, coming into contact with other population groups in the intervening areas. In these areas of "secondary contact," hybridization of species took place so that the discrete morphological characteristics of species that had evolved within the forest refugia were no longer obvious.

According to proponents of the refuge hypothesis, the results of these changes can be seen today in contemporary biogeographical distribution patterns. Within the extensive tropical forests, zones of relatively high species diversity (i.e., zones containing extreme concentrations of different plant and animal species) can be identified. These are sometimes referred to as centers of endemism (Brown and Ab'Saber, 1979). Within these zones, individual species may exhibit very uniform morphological characteristics (Vanzolini and Williams, 1970). Such regions are considered by many to be the former forest refuges that served as survival centers for forest-dwellers during drier intervals. Between these centers of endemism, contact areas or "suture zones" are found, characterized by far fewer species than in the refuges and by more diverse morphological characteristics in the population of a particular species.

The refuge hypothesis is controversial. On the one hand there is considerable biogeographical evidence that there are indeed certain regions where species diversity is extraordinarily high. Such regions are generally identified by first mapping the ranges of individual species, then superimposing the distributions, and selecting those areas that exhibit very high levels of species diversity (Haffer, 1982). In this way, studies of rainforest trees, butterflies, and lizards have been undertaken, and all reveal geographically similar core areas to those suggested by Haffer (1974) on the basis of his detailed studies of tropical birds (Vanzolini and Williams, 1970; Vanzolini, 1973; Brown et al, 1974; Prance, 1974, 1982; Brown, 1982). It is interesting that there is also linguistic and ethnographic evidence that points to the existence of similarly distributed forest refuges in prehistoric Amazonia (Migliazza, 1982; Meggers, 1982). The general coincidence of all these regions is quite impressive, considering the range of evidence involved. However, it could be argued that the distribution patterns observed do not reflect former refugia at all, but merely reflect modern ecological units that have evolved together in response to contemporary edaphic and climatic conditions, the uniqueness of which may or may not be immediately obvious (Endler, 1982; Colinvaux, 1996). Similarly, zones of "secondary contact" may simply reflect significant environmental gradients (Benson, 1982).

Clearly, these arguments can only be satisfactorily resolved by well-dated strati-graphic evidence demonstrating that certain areas contained savanna at the same time as other areas (i.e., the postulated refugia) were under forest cover (Livingstone, 1982). So far, there are simply not enough records to resolve this matter unequivocally, but many lines of evidence strongly suggest that the lowland Amazon Basin remained extensively forested throughout the last 40,000 yr (at least). Of particular relevance is a long pollen record from the lowlands of northwestern Brazil, in the heart of the dense tropical rain forest ecosystem (selva). This record is continuous for >40,000 yr and shows no evidence for any savanna phase in this area; arboreal pollen remained at 70-90% of the pollen sum throughout the period (Colinvaux et al., 1996a). Interestingly, the pollen record shows an increase in montane species, such as Podocarpus, in the last glacial period (marine isotope stage 2), so at that time a unique forest assemblage made up of both lowland and montane elements was present. There is no analog for such an assemblage today; it seems to represent a migration of montane plants into a lowland environment that was cooler than today by 5-6 °C (based on a descent of Podocarpus by -800-1000 m and assuming an adiabatic lapse rate of 0.6 °C 100 m"1). Taken on its own, this one site does not provide a totally convincing argument for dismissing the notion of greater aridity and more widespread savanna in glacial times. However, several other studies support the idea of cooler, not drier, conditions, at least in the core of the evergreen tropical forest zone. In eastern Ecuador, a section dating from the last glaciation also shows an unusual mixture of montane and lowland rainforest pollen types, as well as associated macrofossils. In this area, trees such as Alnus and Podocarpus descended -1500 m from their modern range limits (Liu and Colin-vaux, 1985; Bush et al., 1990). Such evidence has also been found in Panama, where oak trees (Quercus) grew 1000 m lower than today in glacial times (Bush and Colinvaux, 1990). In southeastern Brazil, palynological data also suggest cooler, wetter conditions during the Last Glacial Maximum, with temperatures lower by 6-9 °C. Collectively, such evidence provides a compelling argument for an extensive lowland rain forest throughout the last glacial period and into the Holocene, albeit a forest with a quite different composition than that seen in Amazonia today (Colinvaux et al., 1996b). That is not to say that some peripheral areas that currently experience seasonal moisture deficits were not drier (Markgraf, 1989; Van der Hammen and Absy, 1994) but there is currently little stratigraphic evidence to support the idea that most of the Amazon Basin was occupied by savanna in glacial times (Clapperton, 1993a, b). Indeed, recent studies of pollen in marine sediments from off the mouth of the Amazon reveal little change in arboreal pollen percentages throughout the last 100,000 yr; if savanna had been extensive, there ought to be a clear signal in the pollen carried down the Amazon and deposited offshore, but there is not (Haberle, 1997). It therefore seems likely that the centers of endemism noted in so many biogeographical studies reflect a complex of conditions (climatic, topographic, geological, geomorphological, etc.), which have distinguished these regions over long periods of time, even as climate fluctuated from cooler glacial to warmer interglacial conditions.

9.7.4 Equatorial Africa

As in South America, biogeographers have long held the view that the extensive tropical forests of the Congo Basin and adjacent coastal regions in the Gulf of Guinea were formerly more limited in extent, confined to areas where climatic conditions have remained favorable over long periods of time. These refugia are recognized today by the higher number of endemic taxa and rich species diversity compared to other areas (Hamilton 1976; Sosef, 1991; Maley, 1996). Unlike South America, direct evidence in support of the biogeographical arguments is provided by palynological data from lake sediments. In Ghana, a 27,000-yr long record from crater Lake Bosumtwi shows clearly that the semi-deciduous equatorial forests that have occupied this area for most of the Holocene were not present in glacial times (Fig. 9.22). From 19-15 ka B.P., arboreal pollen fell to as low as 5% of the pollen sum (compared to >75% today) and herbaceous plants (grasses and sedges) occupied the area. This is confirmed by 813C in sediment cores; values were -10 to -20%o during the time when levels of Gramineae pollen were high (consistent with the dominance of C4 plants) compared to Holocene values of -28%o, typical of forest

ED 25 50 75 100 %0 25 50 75 100 %0 12.5 25 %0 12.5 25

FIGURE 9.22 Summary of the important pollen variations in Lake Bosumptwi, Ghana, over the last 27,000 yr. Prior to ~I0 ka B.P., grasses (Gramineae) characteristic of open savanna environments, dominated this location and arboreal pollen was low, especially from ~ 19-1 S.5 ka B.R The presence of mountain olive (Olea hochstet-teri) in this lowland environment indicates cooler conditions prevailed during glacial times (Maley, 1996).

ED 25 50 75 100 %0 25 50 75 100 %0 12.5 25 %0 12.5 25

FIGURE 9.22 Summary of the important pollen variations in Lake Bosumptwi, Ghana, over the last 27,000 yr. Prior to ~I0 ka B.P., grasses (Gramineae) characteristic of open savanna environments, dominated this location and arboreal pollen was low, especially from ~ 19-1 S.5 ka B.R The presence of mountain olive (Olea hochstet-teri) in this lowland environment indicates cooler conditions prevailed during glacial times (Maley, 1996).

(C3) plants (Talbot and Johannessen, 1992; Giresse et al., 1994). Certain montane plants (such as mountain olive, Olea hochstetteri) also migrated into the area; today they are found only far to the west, above -1200 m, suggesting temperatures were cooler by 3-4 °C during glacial times. Similar evidence has been reported from sites in West Cameroon, Plateau Bateke in Congo and farther east in Burundi, where data of the modern pollen rain were used to quantify paleotemperature changes (Fig. 9.23) (Maley, 1991; Bonnefille et al., 1992). By piecing together all the palynological and biogeographical evidence, Maley (1996) constructed a map showing the limits of lowland rain forest refugia in equatorial Africa during the last cold, dry period in the region, which corresponds to the Last Glacial Maximum of higher latitudes (Fig. 9.24). This reveals how dramatically different the region was at that time, with savanna and grasslands covering vast areas that are forested today. Conditions changed abruptly at -9500 yr B.P. with the rapid expansion of

5,980 ± 150 6,030 ± 100 6,700 ± 100 12,160 ± 150 13,250 ±250

FIGURE 9.23 Mean annual temperature departures (relative to the present-day temperature of 15.8 °C) derived from pollen in a peat bog at Kashlru, Burundi, East Africa. The site is at 2240 m in the humid montane forest. Paleotemperature estimates were obtained using a network of modern pollen rain samples over a wide area of East Africa, calibrated against modern climatic data from the region. Dashed lines indicate confidence in-terval.The l4C dates are shown on the left (note that the data are plotted linearly with respect to depth, not age). Paleotemperature estimates from the surface samples are affected by human-induced changes in vegetation and hence the modern pollen spectrum provides a temperature estimate warmer than the instrumentally recorded value (Bonnefille et al„ 1992).

FIGURE 9.23 Mean annual temperature departures (relative to the present-day temperature of 15.8 °C) derived from pollen in a peat bog at Kashlru, Burundi, East Africa. The site is at 2240 m in the humid montane forest. Paleotemperature estimates were obtained using a network of modern pollen rain samples over a wide area of East Africa, calibrated against modern climatic data from the region. Dashed lines indicate confidence in-terval.The l4C dates are shown on the left (note that the data are plotted linearly with respect to depth, not age). Paleotemperature estimates from the surface samples are affected by human-induced changes in vegetation and hence the modern pollen spectrum provides a temperature estimate warmer than the instrumentally recorded value (Bonnefille et al„ 1992).

the rainforest to cover an area even larger than today within 2 ka (Maley, 1991). This change is thought to be related to rapid warming of SSTs in the Gulf of Guinea (as a result of reduced upwelling), leading to a longer wet season and higher total rainfall amounts (Maley, 1989a), but dramatic changes also occurred in East Africa around this time (see Fig. 9.23). Jolly and Haxeltine (1997) argue that significant changes in tropical vegetation would have occurred (even at low elevations) regardless of changes in temperature due to the lower C02 levels during glacial times, which favored grasses and sedges (C4 plants) over trees. If this proves to be correct, at least some part of the observed changes may be related to physiological rather than climatic factors (Street-Perrott, 1994).

FIGURE 9.24 Distribution of tropical forest refugia during the last cool, arid phase in equatorial Africa (—20— 16 ka B.P.) and the present day forest limits. Much of the area now occupied by tropical forest was covered by savanna during the Last Glacial Maximum (LGM) (Maley, 1996).

Tadpy

(piiifit Md nounUinSJ

FIGURE 9.24 Distribution of tropical forest refugia during the last cool, arid phase in equatorial Africa (—20— 16 ka B.P.) and the present day forest limits. Much of the area now occupied by tropical forest was covered by savanna during the Last Glacial Maximum (LGM) (Maley, 1996).

At the same time as lowland forest vegetation was expanding in the early Holocene, climatic conditions in sub-Saharan Africa became less arid, enabling the semiarid savanna vegetation belt to extend farther northward (see Section 7.6.3). As a result, the range of large herbivores (such as giraffe, elephant, hippopotamus, and gazelle) was also more extensive and today the bleached bones of these animals provide a mute reminder of the remarkably different climatic conditions that existed there in the early Holocene. Accompanying the animal migration into sub-Saharan Africa were aboriginal hunters who recorded their way of life on magnificent rock paintings and carvings (Lhote, 1959; Monod, 1963; Lajoux, 1963). Today these are found hundreds of kilometers from the nearest permanent settlements.

Of particular significance during the period of more extensive savanna and tropical forests were the much more extensive riverine and lacustrine environments, which effectively provided water connections across the entire sub-Saharan region, from the Nile to Senegal (Beadle, 1974). Fauna of the Lake Chad Basin, for example, provide unequivocal evidence for recent connections, not only with the Niger and Congo basins, but also with the Nile drainage system over 1000 km to the east. Even today, relict populations of animals and plants are found isolated in topographically favorable environments, far removed from their nearest adjacent populations. For example, the Eurasian green frog (Rana ridibunda) has been found in the streams of the Ahaggar mountains at least 1000 km from adjacent population groups. Furthermore, and perhaps most remarkably, a Nile crocodile (Crocodilus niloticus) was found in a pool in the Tassili-N-Ajjer Mountains, separated by vast stretches of desert from major population centers to the east (Seurat, 1934; Beadle, 1974). Such disjunct species bring climatic changes to life.

Subtropical Africa is probably more arid today than it has been during most of the Holocene and was only more arid during the late Wisconsin glacial maximum (see Figs. 7.23-7.25). However, the record is still far from complete in either time or space, so brief periods of relatively moist or even drier conditions (on the order of 102

years) may have occurred. Indeed, lake level evidence does show that abrupt changes in climate (probably related to upwelling) did occur in the late Holocene (e.g., -3500-4000 yr B.P., as recorded by lake-level changes in Lake Bosumptwi) but these oscillations did not persist and no significant change in forest cover resulted (Talbot and Delibrias, 1977; Maley, 1991). Nevertheless, such episodes may have had significant consequences for human populations in the region (Maley, 1989b, 1997).

Was this article helpful?

0 0

Responses

  • swen
    What does paleoclimatology have to do with global warming?
    1 month ago

Post a comment