Learning about the Tail of the Probability Density Function for Equilibrium Climate Sensitivity

University Of Illinois At Urbana-Champaign, Urbana IL

Investigators

Abstract

The importance of human-induced climate change depends critically on the equilibrium climate sensitivity (referred to as climate sensitivity hereafter, i.e., the change in global surface air temperature resulting from a doubling of the pre industrial atmospheric carbon dioxide concentration). The Intergovernmental Panel on Climate Change (IPCC) historically assessed the climate sensitivity at a range from 1.5 to 4.5 degrees Celsius. The PIs document that there is a significant likelihood that the climate sensitivity lies outside this range, with probability density functions (pdfs) that have a thick upper tail. The most recent IPCC report presents a number of pdfs for the sensitivity, all of which have a thick upper tail. This research focuses on the thick upper tail. Testing whether or not the real climate system is located in this thick upper tail is not just a tantalizing scientific puzzle, but can also provide large economic value, as it can inform the design of improved climate policies. A recent economic analysis shows that the thick upper tail of current climate sensitivity estimates can dominate the ranking of proposed climate-change strategies in terms of their benefit/cost ratio. The intuition behind this is that a decision maker may give considerable weight to the very costly outcomes of different climate change strategies that are associated with the cases where the thick upper tail of current climate sensitivity estimates is revealed to be the true state of the system. The PIs will use their simple climate model and updated observational data to test five hypotheses: (1) The inclusion of the additional near-surface air temperature observations since 1998 will reduce the 90% confidence interval of climate sensitivity; (2) The inclusion of non-sulfate aerosol radiative forcing will increase the 90% confidence interval of climate sensitivity, this because the resulting negative radiative forcing will be greater than that for sulfate aerosol alone; (3) The inclusion of oceanic heat-content data will reduce the 90% confidence interval of climate sensitivity; (4) The combined effect of these three datasets will be a reduction of the 90% confidence interval of climate sensitivity; (5) The uncertainty in climate sensitivity due to climatic noise can be reduced by learning about climate sensitivity overtime. In addition, the PIs will examine how climate variability and paleoclimatic information influence the estimation of the pdf of climate sensitivity. The Bayesian averaging method that they will use will allow them to learn about different radiative forcing models and incorporate their associated hindcast abilities into the estimation of the distribution of climate sensitivity. The intellectual merit of the research is that it will produce improved estimates of climate sensitivity from both the Frequency and Bayesian points of view using additional observations: updated surface air temperatures, ocean heat content, paleoclimatic data, and radiative forcing from non-sulfate aerosols. This knowledge will inform decision-makers about the magnitude and rate of reductions in the anthropogenic emission of greenhouse gases that are needed to "...prevent dangerous anthropogenic interference with the climate system", as required by Article 2 of the UN Framework Convention on Climate Change. This is also the most important broader impact of the research. The second broader impact will be the use of the results by other scientists to study the impacts and policy implications of human-induced climatic change. The third broader impact will be public outreach and communication. The PI will continue this via his lectures to the public, business leaders and policymakers.