The interaction of ice sheets with climate models possessing different sensitivities is tested with the aid of a vertically integrated ice sheet model with isostatic bedrock adjustment. The ice sheet model has two horizontal dimensions and a coarse geography/topography over a domain encompassing the area north of 40¿N. Two-coupled ice sheet-climate model experiments are designed. The first experiment is started with today's surface conditions and continuous extreme low summer insolation orbital conditions are switched on. The second experiment is started with boundary conditions of the last glacial maximum (LGM) and a continuous extreme high summer insolation is applied. Control runs with current orbital conditions are compared. For each of the two experiments, coupled runs are performed twice: (1) with a climate component possessing a moderate sensitivity (β=1.8¿C) and (2) with a high sensitivity (β=4.0¿C) to 1% changes in solar constant, as developed in Part 1 of this study. The moderate-sensitivity coupled run cannot explain the onset of ice age in the first experiment, while the high-sensitivity coupled run utterly fails to reproduce the decay of the LGM ice sheet in the second experiment. A great significance is attached to the geographical distribution as resolved in two horizontal dimensions by our ice sheet model. The land portion susceptible to glaciation onset is extremely small, while the land portion crucial for deglaciation is much larger. The lower continentality of North America is found to be a promoting factor for glaciation but an impeding factor for deglaciation.

The feedback between the ice sheets and climate component is well simulated when the coupled system is exposed to variations in the solar constant (-2.5%/+5% have simulated creation/destruction of a large ice sheet with the moderate-sensitivity coupled run), resulting in a steady drift of climate and ice sheet extent for each successive run of the coupled components. Insolation changes of limited duration are found to have a durable effect, shifting the ice sheet-climate system into a new mode. Evidence for the global synchronism of ice age terminations (with a slight lag of the northern, more massive ice) and for the existence of ~100-Kyr cyles in nonglacial eras leads us to suspect that there might be a missing mechanism for the translation of weak orbital forcing into global equivalent insolation changes, but we cannot rule out genuine solar constant variations of purely internal changes in the global radiation budge as the primary cause for the major glacial-interglacial transitions.