Challenges for ice age dynamics: a dynamical systems perspective

TL;DR

This paper reviews the slow dynamics of ice ages within the climate system, emphasizing the interactions between various components and the challenges in modeling their complex, multi-scale variability over geological timescales.

Contribution

It introduces the use of stochastic parameterizations in low-resolution models to better capture weather-scale variability and discusses the challenges in understanding interactions between slow and fast climate modes.

Findings

Interactions between slow and fast climate modes are poorly understood.

Stochastic parameterizations can improve modeling of slow climate dynamics.

Challenges include statistical issues and model validation.

Abstract

This chapter is dedicated to the slow dynamics of the climate system, at time scales of one~thousand to one million years. We focus specifically on the phenomenon of ice ages that has characterised the slow evolution of climate over the Quaternary. Ice ages are a form of variability featuring interactions between different large-scale components and processes in the climate system, including ice sheet, deep-ocean and carbon cycle dynamics. This variability is also at least partly controlled by changes in the seasonal and latitudinal incoming solar radiation associated with the combined effects of changes in Earth's orbit shape, precession of equinoxes, and changes in obliquity. A number of possible mechanisms are reviewed in this chapter. We stress that the nature of the interactions between these slow dynamics and faster modes of variability, such as millennium and centennial modes of variability, are still poorly understood. For example, whether the time sequence of ice ages is robustly determined or not by the astronomical forcing is a matter of debate. These questions need to be addressed with a range of models. We propose to use stochastic parameterisations in the lower resolution models (focusing on climate time scales) to account for weather and macro-weather dynamics simulated with higher resolution models. We discuss challenges---including statistical challenges---and possible methods associated with this programme.