The Later Pleistocene

A more spatially representative picture of changes in the distribution and composition of rainforest can be constructed for the later part of the Pleistocene from a greater number of sites that, in addition to those from ODP Sites 1144 and 820, provide continuous or near-continuous palynological records through at least the last glacial cycle (Figure 4.1). The record from core SHI-9014 in the Banda Sea (van der Kaars et al., 2000) provides the most substantial evidence of vegetation change in the rainforest core of Southeast Asia and a framework for examination of variation within the broader area (Figure 4.9). The regional significance of the Banda Sea record is demonstrated by its remarkable similarity to a recent record from core MD98-2175 in the adjacent Aru Sea (van der Kaars, new data). Rainforest is a prominent component of the pollen spectra throughout, although the substantial sclerophyll component—largely Eucalyptus, that would have been derived mainly from the Australian mainland—demonstrates a very broad pollen catchment area. Highest rainforest and pteridophyte values occur during interglacials, indicating that they were much wetter than glacial periods. However, expansion of the Sahul continental shelf during times of low sea level, much of which appears to have been covered largely by grassland (Chivas et al., 2001), would have resulted in excessive Poaceae representation and probable exaggeration of moisture variation through the recorded period. Wet conditions and the maintenance of near-continuous rainforest are certainly evident in some areas—such as highland New Guinea (Walker and Flenley, 1979) and much of the island of Borneo (Anshari et al., 2004; Morley et al., 2004) during the last glacial period, including the Last Glacial Maximum. However, grassland may have disrupted forest growth in more peripheral rainforest (see Section 4.6). The Banda Sea record displays higher values for upper montane taxa during interglacial than glacial periods, and this is surprising considering the abundant evidence from sites in highland parts of the region for much expanded montane vegetation with substantial temperature lowering (see Section 4.6 for discussion of this). This Banda Sea pattern may be a result of an overall reduction in lowland rainforest that is demonstrated to have covered at least parts of the Sunda continental shelf within core rainforest areas during the last glacial period (Morley et

China Isotope Record
Figure 4.9. Selected features of the pollen and charcoal record from Banda Sea core SHI-9014 in relation to the marine isotope record (adapted from van der Kaars et al., 2000). All taxa are expressed as percentages of the rainforest pollen sum.

al., 2004). It may also be the case that the present day terrestrial area of lowland rainforest was little reduced during glacial periods if there was a smaller degree of temperature lowering at low altitudes. One feature of the Banda Sea record that is shared with those from the Aru Sea and Sangkarang-16, offshore Sulawesi, is a dramatic and sustained reduction in pollen of the dominant lowland rainforest family Dipterocarpaceae about 37 kyr. It is possible that the present pattern of representation of the family—that is lower in abundance and diversity in the eastern than western part of the region—is as much the result of this Late Pleistocene event as it is of the historical barrier of Wallace's Line to migration of the Indo-Malaysian flora westwards as generally assumed (Whiffin, 2002). As this Dipterocarpaceae is associated with an increase in charcoal, burning is regarded as the primary cause, and the impact of early people—rather than climate—has been postulated as its major cause (van der Kaars et al., 2000). Similar sustained increases in charcoal recorded in long marine cores from the Sulu Sea (Beaufort et al. 2003) and to the north of New Guinea (Thevenon et al., 2004)—but from a different time, about 52kyr—have been considered as providing support for the human burning hypothesis.

A rare insight into the history of rainforest on the dry margin of rainforest distribution is provided by marine core MD98-2167 in the North Australian Basin, off the coast of the Kimberley Ranges of northwestern Australia (Figure 4.10). Here,

Australia Rainforest Climate Diagram

Figure 4.10. Selected features of the pollen and charcoal record from the North Australian Basin core MD98-2167 (Kershaw et al., in press; van der Kaars, unpublished data) in relation to the marine isotope record of Brad Opdyke (unpublished data). All taxa are expressed as percentages of the total dryland pollen sum.

Figure 4.10. Selected features of the pollen and charcoal record from the North Australian Basin core MD98-2167 (Kershaw et al., in press; van der Kaars, unpublished data) in relation to the marine isotope record of Brad Opdyke (unpublished data). All taxa are expressed as percentages of the total dryland pollen sum.

deciduous vine thickets exist in small pockets surrounded by eucalypt-dominated savanna woodland. The major representative of these vine thickets—that have generally poor pollen dispersal—is considered to be Olea type. It has low but relatively consistent representation through the record—showing little response to inferred changes in rainfall as indicated by broad glacial-interglacial changes in relative abundance of tree and shrub pollen, that derive largely from the Kimberley region—and particularly the pteridophyte spores that must have been derived from the core rainforest area of Indonesia. Burning appears to have increased around 130 kyr, without any notable change to the vegetation structure, apart from some evidence for increased variability in tree and shrub to herb representation. However, major changes occurred around 46 kyr that included the total disappearance of Olea type from the record. This decline in Olea may have been associated with a general further increase in burning that has continued to the present.

4.6 THE LAST GLACIAL MAXIMUM (LGM) TO HOLOCENE

A much greater spread of site records as well as more detailed analysis of sites during this period is available than for previous ones (Figure 4.1), allowing more refined investigation of temporal and spatial patterns both within rainforest and between rainforest and more open vegetation communities. Of particular interest are the extent of rainforest and altitudinal shifts in rainforest communities during the LGM that inform debates on contemporary precipitation and temperature levels.

4.6.1 Last Glacial Maximum

The idea that lowland rainforests might have been replaced by grassy savannas at the LGM is clearly not substantiated from the evidence from longer records, but there is some evidence of savanna expansion. The actual degree and areal expression of this expansion is hotly debated. In some more marginal rainforest areas, savanna vegetation did replace rainforest in part—as around Rawa Danau (van der Kaars et al., 2001) and the Bandung Basin (van der Kaars and Dam, 1995)—or totally—as on the Atherton Tableland in northeastern Australia (Kershaw, 1986)—but increased representation of grasses in coastal sites may have been reflecting more open vegetation on exposed continental shelves or an increased aquatic component. Greatest debate has been over the potential existence of a north-south dry corridor extending through Malaysia and between Sumatra and Borneo during the LGM. At one extreme is the view of Morley (2000, 2002) who considered that rainforest massifs, or refugia, were essentially restricted to southwestern Borneo and the adjacent Sunda Shelf, the western part of Sumatra, and very western tip of Java (see Chapter 1). Major migration of rainforest is implied between glacials and interglacials unless high diversity was conserved in river gallery forests, a situation proposed for northeastern Australia during the last glacial period (Hopkins et al., 1993). Kershaw et al. (2001), on the other hand, see little evidence for such a dry corridor, at least during the last glacial period. Evidence is sparse and their interpretation is based largely on the almost complete dominance of rainforest pollen in submerged peat cores from the Sunda Shelf off southeastern Sumatra (van der Kaars, unpublished data). Although undated, the peat almost certainly derives from the last glacial period rather than any earlier period as it is unlikely to have survived subsequent low sea level stands.

There is general consensus, however, that—in accordance with the reconstruction of Morley (2000)—much of Borneo retained rainforest and that this forest extended over the continental shelf within the South China Sea region. Confirmation of the maintenance of a rainforest cover within inland West Kalimantan is provided by Anshari et al. (2001, 2004), although drier conditions during the later part of the last glacial period are evident, while a marine pollen record from core 17964 in the southern part of the South China Sea (Sun et al., 1999) further substantiates the dominance of a rainforest cover.

In comparison with lowland sites, those from the highlands show the clear maintenance of rainforest through the LGM. Here attention has focused on altitu-dinal changes in representation of rainforest components. There is good pollen evidence from several sites showing movement of montane tree taxa to lower altitudes. An excellent example of this migration is seen in the pollen diagram from the swamp at the edge of Lake di-Atas in Sumatra (Newsome and Flenley, 1988). The site is at 1,535 m a.s.l., and—where the forest around the lake survives—it is dominated by a variety of tropical oak taxa: Lithocarpus, Castanopsis, and Quercus. It is believed that formerly the tree Altingia excelsa (Hamamelidaceae) was abundant also (van Steenis, 1972), but it has been selectively logged. Above 1,800 m the forest changes sharply and becomes dominated by gymnosperms: Dacrycarpus imbricatus, Podocar-pus neriifolius and (in swamps) Dacrydium cf. elatum. Even Pinus merkusii is present, its only natural occurrence in the southern hemisphere. There are also angiosperm trees, the most conspicuous being Symingtonia populnea. The diagram (Figure 4.11) shows that in a phase dated to between c. 18 kyr and c. 12kyr bp (c. 22-14 kcal. yr bp) all those gymnosperms are prominent in the record, only to disappear in the Holocene and be replaced by the distinctive pollen of Altingia (previously rare) and peaks of Quercus and Lithocarpus/Castanopsis. This replacement strongly suggests a climate cooler at the LGM, perhaps by 2°C or more. Interestingly, there is an inversion of radiocarbon dates around 14C kyr bp, which could be explained by lower water tables, permitting erosion of swamp sediments and their redeposition within the core. This depositional event would correlate with the drier lowland Pleistocene climates already mentioned.

Confirmation of these results comes from a site in Java at c. 1,300 m—Situ Bayongbong (Stuijts, 1984)—that is close to the Bandung basin site. Lower altitude sites in Sumatra (Maloney, 1981,1985,1998; Morley, 1982; Maloney and McCormac, 1995) also are supportive. These data bring the records of montane gymnosperms down to 1,100 m at c. 22 kcal. yr bp and do not conflict with the Bandung occurrence of Dacrycarpus pollen at 650 m, and at Rawa Danau at only c. 100 m (van der Kaars et al., 2001). Collectively, therefore, these results support the Bandung estimate (van der Kaars, 1998) of a climate cooler at the LGM by as much as 4°C or 5°C.

The incursion of montane elements into lowland areas could have been even more marked. The site at Kau Bay in Halmahera (Barmawidjaya et al., 1989) is at present-day sea level. This flooded volcanic crater was a freshwater lake when sea level was

Flenley Diagram
Figure 4.11. Pollen diagram from Danau di Atas Swamp, West Sumatra, altitude 1,535m. Values are given as percentages of total dry land pollen. Only selected taxa are shown. After Newsome and Flenley (1988) and Stuijts et al. (1988).

lower by >100m at the LGM. Palynology of this site showed occurrences of Cas-tanopsis/Lithocarpus and Quercus at the LGM. While it is not suggested that these taxa necessarily grew at present sea level, they apparently grew close enough for small amounts of their pollen to enter the record. This record would be consistent with a temperature lowering of c. 6°C at the LGM, even in the lowlands. Similar results have been obtained from lowland sites in West Kalimantan (Anshari et al., 2001, 2004).

There is also evidence from the highest mountains of Indonesia for Pleistocene cooling of as much as 6°C. Leaving aside the evidence from New Guinea, we have evidence of Pleistocene glaciation on Mt. Kinabalu in northern Borneo (Koopmans and Stauffer, 1968) and of deglaciation around 8,000 bp (c. 10kcal. yr bp) at 4,000 m a.s.l. (Flenley and Morley, 1978). Similar evidence (for solifluction at least) is claimed for the slightly lower peak of Gunong Leuser (3,381 m) in northern Sumatra (Beek, 1982).

In New Guinea the best evidence of environmental change comes from upland regions. The site at Sirunki (Walker and Flenley, 1979) appears to cover the last 33 kyr (c. 40kcal.yr bp), at an altitude of 2,500 m a.s.l. The site currently lies in Nothofagus forest (much disturbed), and is some 1,300 m below the altitudinal forest limit at c. 3,800 m a.s.l. Nevertheless the pollen record clearly shows the presence of tropic-alpine herbs (Astelia, Gentiana, Drapetes, etc.) in the Late Pleistocene, when forest pollen values decline to a level consistent with unforested conditions. Similar results were obtained from Lake Inim at 2,550 m by Flenley (1972) (Figure 4.12).

There is, of course, geomorphological evidence of lowered snow lines (U-shaped valleys, moraines, etc.) in the New Guinea mountains. On Mt. Wilhelm, a lowering in the Late Pleistocene of c. 1,000 m is indicated (Loffler, 1972), and there are similar findings from Irian Jaya (Hope and Peterson, 1975). One thousand metres translates into perhaps 6°C cooling, using a modern lapse rate.

Pollen Diagram Climatic Late Glacial

Sornro*n«ysaP*1ritut [jVftW IB ^ HirtlDlKMriiClffiUrt War* itfi bird

MRU CJ » CTS As. 2S5Dm (A30Qft.f Lat STBt Long H33S

Figure 4.12. Pollen diagram from Lake Inim, boreholes C4 and C15, plotted on the same scales. The results are expressed as percentages of pollen of forest types, except in the summary diagram where the total of dry land pollen and spores forms the pollen sum. Only selected taxa shown. After Flenley (1972).

Sornro*n«ysaP*1ritut [jVftW IB ^ HirtlDlKMriiClffiUrt War* itfi bird

MRU CJ » CTS As. 2S5Dm (A30Qft.f Lat STBt Long H33S

Figure 4.12. Pollen diagram from Lake Inim, boreholes C4 and C15, plotted on the same scales. The results are expressed as percentages of pollen of forest types, except in the summary diagram where the total of dry land pollen and spores forms the pollen sum. Only selected taxa shown. After Flenley (1972).

Bare

Tropic-alpine or Sub-afpine Upper Montane Rain Forest Lower Montane Rain Forest Lowland Rain Forest Savanna

Altitudinal Vegetation Limit Upper Montane Forest/ Subalpine Boundary Lower Montane Forest/Upper Montane Forest Boundary Lower Limit of Abundant Gymnosperms in Sumatra Lowland/Lower Montane Forestry Boundary

Was this article helpful?

0 0
Waste Management And Control

Waste Management And Control

Get All The Support And Guidance You Need To Be A Success At Understanding Waste Management. This Book Is One Of The Most Valuable Resources In The World When It Comes To The Truth about Environment, Waste and Landfills.

Get My Free Ebook


Responses

  • MIKE CAMPBELL
    How can mountain barrier affect climate diagram?
    9 years ago

Post a comment