Perhaps the most important contribution of paleoclimate analysis to the ongoing debate concerning our future in the “greenhouse” is that associated with the issue of rapid climate change. The ultra-low-frequency variability derivative of the 100 kyr cycle of the late Pleistocene is now a rather well understood consequence of the nonlinear response of the system to orbital insolation forcing (e.g., see Tarasov and Peltier, 1997, 1999 for detailed discussion and Berger, Chapter 8 in this volume), but no similar degree of understanding may be claimed for the millennium and shorter-period variability that is evident in the δ180 atmospheric temperature proxy from the Summit, Greenland, ice cores during the interval of time known as Oxygen Isotope Stage 3. This is a critical gap, especially inasmuch as it has been suggested that these Dansgaard-Oeschger (D-O) oscillations might also be induced under the enhanced greenhouse conditions that we expect will be characteristic of the future.

The only explicit theory of the D-O oscillation in the current literature is that based on a highly reduced model of the global thermo-haline circulation (Sakai and Peltier, 1995, 1996, 1997, 1999), in which it is shown that the THC should “fibrillate” in response to application of a supercritical freshwater forcing applied to the high latitudes of the North Atlantic basin. In this model the D-O oscillation occurs through a simple Hopf bifurcation when the buoyancy flux is somewhat weaker than would be required to arrest the deep circulation completely.