A protocol to articulate and quantify uncertainties in climate change detection and attribution

This work develops and describes a formal probabilistic protocol via which the process of identifying lines of evidence for climate change, assessing likely causes for changes in the evidence, and combining the lines of evide nce to make overall attributions of cause to greenhouse gases can be made. This open and detailed model of the detection and attribution process is de signed to identify issues at stake in detection and attribution, and to fac ilitate scrutiny and understanding on this contentious issue in broader com munities. The protocol provides a convenient means to make each of the judg ements in this issue explicit. These judgements are characterized via exper t elicitation techniques in both quantitative and qualitative form. The pro tocol focuses on detecting climate change and attributing causes to the enh anced greenhouse effect rather than on more general anthropogenic change, b ecause the former is more consequential. Major uncertainties identified in the protocol relate to characterization of natural variability for each lin e of evidence, non-greenhouse forcings, and the climate response to forcing . The relative roles of uncertainty in climate sensitivity and climate forc ings are still. unclear in making determinations of attribution. Measures o f attribution need to account for both the amount of signal explained by a postulated cause as well as its associated probability. When combining line s of evidence to form overall measures of attribution, the level of depende nce assumed among lines of evidence is critical. Finally, the protocol high lights the need to reevaluate standards of evidence for attributing greenho use climate change.