Distinguishing Between Insolation‐Driven and Phase‐Locked 100‐Kyr Ice Age Scenarios Using Example Models

Abstract Glacial‐interglacial oscillations exhibit a periodicity of approximately 100 Kyr during the late Pleistocene. Insolation variations are understood to play a vital role in these ice ages, yet their exact effect is still unknown; the 100 Kyr ice ages may be explained in two different ways. They could be purely insolation‐driven, such that ice ages are a consequence of insolation variations and would not have existed without these variations. Or, ice ages may be self‐sustained oscillations, where they would have existed even without insolation variations. We develop several observable measures that are used to differentiate between the two scenarios and can help to determine which one is more likely based on the observed proxy record. We demonstrate these analyses using two representative models. First, we find that the self‐sustained model best fits the ice volume proxy record for the full 800‐Kyr time period. Next, the same model also shows a 100 Kyr peak consistent with observations, yet the insolation‐driven model exhibits a dominant 400 Kyr spectral peak inconsistent with observations. Our third measure indicates that midpoints in ice volume during terminations do not always occur during the same phase of insolation in both observations and the self‐sustained scenario, whereas they do in the insolation‐driven scenario. While some of these results suggest that the self‐sustained ice ages are more consistent with the observed record, they rely on simple representations of the two scenarios. To draw robust conclusions, a broader class of models should be tested using this method of producing observable differences.