It is widely acknowledged that many components of the Earth's climate system exhibit nonlinear behavior when forced, with examples ranging from ice sheets to ocean circulations to savanna. These changes, as profound as they are, are not expected to greatly impact global climate, as characterized by values such as the global mean annual surface temperature. This value is usually assumed, and often modeled, to be roughly proportional to the greenhouse gas-induced radiative forcing . In the limit of large (i.e. greater than anthropogenic) forcing, this approximation is expected to break down dramatically. In this talk, I will use a simple conceptual model to argue that this threshold might be closer than one might expect. I will demonstrate how the effect of a temperature-dependent sensitivity can be confused with model error, creating selection bias, with examples drawn from preliminary results of a perturbed physics ensemble. I will also show the importance of this effect for determining the risk of high warming. I will discuss evidence from paleoclimate that can help constrain this temperature dependence. Finally, I will briefly discuss how regional patterns of feedbacks can complicate attempts to diagnose nonlinear sensitivity.
This talk was given as the MCRN Colloquium Webinar on Monday, November 23rd, 10:30am EST, by Jonah Bloch-Johnson (University of Chicago, Department of the Geophysical Sciences).
RENCI, UNC Chapel Hill