On Reducing the Uncertainty of High Climate Sensitivity

Abstract:

Physical modelling and observational studies sometimes suggest there is a nonnegligible probability that the Earth's equilibrium climate sensitivity is greater than 4.5ºC, the upper end of the IPCC range of likely values. Such worst case scenarios warrant special scrutiny, and there has been much discussion of the methods, and even the feasibility, of reducing our uncertainty of their likelihood. We present a brief review of high sensitivity outcomes and what gives them such high values, highlighting which mechanisms suggest reasonable future pathways for the Earth. In particular, we highlight the role of low cloud feedbacks. We illustrate that the assumption of linearity often used in analysis of climate feedbacks becomes increasingly likely to break down for high sensitivity Earths, and show in a simple thought experiment how the true climate sensitivity can become decorrelated from the one predicted by linear feedbacks as these feedbacks become strongly positive, a condition necessary for high sensitivity. Knowledge of the nonlinearity of these feedbacks is essential to reducing uncertainty of the highest warming scenarios, with implications for the design of the methodologies used to study them.