HEINRICH-TYPE GLACIAL SURGES IN A LOW-ORDER DYNAMICAL CLIMATE MODEL

Recent studies suggest the occurrence of sporadic episodes during whic h the ice streams that discharge ice sheets become enormously active, producing large numbers of icebergs (reflected in North Atlantic sea c ores as ''Heinrich events'') and possibly causing the partial collapse of the ice sheets. To simulate the mechanism of internal thermo-hydro dynamical instability implied by such behavior in the context of a mor e general paleoclimate dynamics model (PDM), we introduce a new slidin g-catastrophe function that can account for ice-sheet surges. In parti cular, using simple scaling estimates derived from the equations of mo tion and thermo-conductivity for ice flow, we express this function in terms of the thickness, density, viscosity, heat-capacity, and heat-c onductivity of ice. Analysis of the properties of this function sugges ts that these Heinrich-type instability events might be of three possi ble kinds: the first type of event occurs in periods of glacial maximu m when temperature conditions on the ice surface are extremely cold, b ut internal friction within bottom boundary layer is also at its maxim um and is strong enough to melt ice and cause its surge. The second ty pe of event may happen during an interglacial, when the ice thickness is small but relatively warm climatic conditions on the upper surface of ice can be easily advected with the flow of ice to the bottom where even a small additional heating due to friction may cause melting. Th e third and, perhaps, most interesting type of event is one that may o ccur during ice sheet growth; in this period particles of ice reaching the bottom still ''remember'' the warm temperature conditions of the previous interglacial and additional heating due to increasing frictio n associated with the growing ice sheet may again cause melting. To th e extent that the upper glacier surface temperature depends on atmosph eric carbon dioxide concentration, this third case introduces the inte resting possibility that earlier CO2 concentrations may be as importan t for the present-day climate as its current value. We present results of numerical experiments demonstrating how these three kinds of insta bility can originate and interact with other components of the global climate system to produce variations of the Heinrich-event type. In pa rticular, according to our model the climate system seems more vulnera ble to surges during the penultimate interglacial period than in the p resent one, which may contribute to an explanation of the recent resul ts of the Greenland Ice Core Project.