According to general circulation models (GCMs) the projected doubling of greenhouse gases by the end of the next century may produce a global temperature increase of 2-5⁰C and a precipitation increase of 7-15%. Regional average temperature changes are predicted to range from -3⁰C to +10⁰C and precipitation changes from -20% to +20% (Schneider and Rosenberg, 1989). A major problem is, however, that predictions of CO₂ - induced climate change by different GCMs do not agree about how regional climates may change (Batie and Shugart, 1989). This makes evaluation of the consequences of greenhouse warming on ecological systems extremely difficult. Yet, significant changes must be expected. Kerr (1988), for example, suggests that for each 1⁰C of warming climate zones and the ecosystems they support may shift 100-150 Km poleward.

If GCMs can not be relied upon to produce accurate portrayals of the Earth's Climate, how can we predict the impact of global warming on the Earths ecosystems? One possibility is to examine conditions during the Holocene Hypsithermal Interval, 8,000-4,000 years ago, the warmest period of postglacial times. During this interval mean annual temperature may have been 1-3⁰C higher in many areas go the temperate zone (Goudie, 1983_p. 125). Another possibility is to examine environments during what may have been the warmest period of the last few hundred thousand years- namely isotope stage 5e of the deep sea record, which lasted from about 125,000 -115,000 BP. At this time sea level stood about 8m above its present level because of the increased melting of the polar ice sheets. Deep ocean sediment, raised coral reef and ice core Theta¹⁸O curves all indicate a period substantially warmer than the Holocene as do studies of carbon dioxide and methane in the polar ice sheets (e.g. Chappell and Shackleton, 1986; Chappellaz et al., 1990; Pillans, 1987).

The age of this exceptionally warm interglacial period has been a significant hindrance to its being used to assess what might happen with increased global warmth. However, in recent years it has been discovered that cave stalactites and stalgmites (speleothems) may contain pollen grains that can be used to reconstruct the vegetation near a cave (e.g. Brook et al, 1987, 1990a, 1990b). Most importantly, cave speleothems can be dated accurately by U-series to 350,000 B.P. they thus provide a way of examining vegetation changes during both the Holocene Hypsisthermal and deep sea isotope stage 5e.