Expert Assessment of Uncertainties in Detection and Attribution of Climate Change

Abstract

The problem of detection of climate change and attribution of causes of change has been formalized as a series of discrete probability judgements in an expert elicitation protocol. Here results are presented from the protocolfor 19 experts, highlighting areas of convergence and divergence among experts. There is broad agreement among the experts that the global mean surface air temperature, vertical pattern of temperature change, geographical pattern of temperature change, and changes in diurnal temperature are the important lines of evidence for climate change detectionand attribution. For the global mean and vertical pattern lines of evidence, the majority of experts (90%) reject the null hypothesis (no climate change) at the 5% significance level, thereby lending strong support to detection of climatechange. For these lines of evidence the median probability of detection at the 5% significance level across experts exceeds 0.9. For the geographical pattern and diurnal cycle lines of evidence, there is far less agreement andfewer than half the experts support detection at even the 10% level of significance. On attribution there is a broad consensus that greenhouse forcing is responsible for about half the warming in global mean temperature in the past century. This result is fairly robust to uncertainties assessed in the relevant forcings by this set of experts.For the other lines of evidence, greenhouse forcing makes smaller fractional contributions with more spread among expertassessments. The near consensus of the experts on detection of climate change and attribution to greenhouse gases rests on the evidence of change in global mean surface air temperature. For the other lines of evidence, there is eithersignificant expert disagreement on detection (the geographical pattern and diurnal cycle), or attribution of change is predominantly to causes other than greenhouse gas forcing (the vertical pattern).