Keeping it simple: the value of an irreducibly simple climate model

Abstract

Richardson et al. (Sci Bull, 2015. doi:10.1007/s11434-015-0806-z) suggest that the irreducibly simple climate model described in Monckton of Brenchley et al. (Sci Bull 60:122–135, 2015. doi:10.1007/s11434-014-0699-2) was not validated against observations, relying instead on synthetic test data based on underestimated global warming, illogical parameter choice and near-instantaneous response at odds with ocean warming and other observations. However, the simple model, informed by its authors’ choice of parameters, usually hindcasts observed temperature change more closely than the general-circulation models, and finds high climate sensitivity implausible. With IPCC’s choice of parameters, the model is further validated in that it duly replicates IPCC’s sensitivity interval. Also, fast climate system response is consistent with near-zero or net-negative temperature feedback. Given the large uncertainties in the initial conditions and evolutionary processes determinative of climate sensitivity, subject to obvious caveats a simple sensitivity-focused model need not, and the present model does not, exhibit significantly less predictive skill than the general-circulation models.