Comment on "Using multiple observationally-based constraints to estimate climate sensitivity" by J. D. Annan and J. C. Hargreaves, Geophys. Res. Lett., 2006

Abstract. In their article from 2006, Annan and Hargreaves present a probabilistic estimate of climate sensitivity obtained by using Bayes' theorem to combine information from different sources. In this comment article we critisize two aspects of their reasoning, namely using probability density functions and likelihood functions interchangeably and the assumed independence of evidence from the different sources. The derivation of their result rests on key assumptions, some stated explicitly and some left implicit, which could be unrealistic. Thus their study does not convincingly reduce the large uncertainty of climate sensitivity remaining in previous observationally-based studies.