Feedbacks, climate sensitivity, and the limits of linear models

Abstract:

The term “feedback” is used ubiquitously in climate research, but implies varied meanings in

different contexts. From a specific process that locally affects a quantity, to a formal framework

that attempts to determine a global response to a forcing, researchers use this term to separate,

simplify, and quantify parts of the complex Earth system.

We combine large (>120 member) ensemble GCM and EMIC step forcing simulations over a broad range of forcing levels with a historical and educational

perspective to organize existing ideas around feedbacks and linear forcing-feedback models.

With a new method overcoming internal variability and initial condition problems we quantify the non-constancy of the climate feedback parameter.

Our results suggest a strong state‐ and forcing‐dependency of feedbacks, which is not considered appropriately in many studies.

A non‐constant feedback factor likely explains some of the differences in estimates of equilibrium climate sensitivity from

different methods and types of data. We discuss implications for the definition of the forcing term and its various adjustments.

Clarifying the value and applicability of the linear forcing feedback framework and a better quantification of feedbacks on various timescales and spatial

scales remains a high priority in order to better understand past and predict future changes in the climate system.