Do we know what difference a delay makes?

Abstract

In our original comment [Risbey et al.., 1991] we argued that the work of Schlesinger and Jiang [1991a] is too limited to determine whether or not (as they put it) “the penalty is small for a 10‐year delay in initiating the transition to a regime in which greenhouse‐gas emissions are reduced.” In their reply, Schlesinger and Jiang [1991b] (hereafter S&J) presented their reasons for concluding definitively that the penalty is small. However S&J's discussion of the evidence and literature on climate change and greenhouse warming contains significant omissions and mis‐statements.In dismissing our concern that their model was too simple to evaluate the possibility of abrupt climate change, S&J rely on results from coupled ocean‐atmosphere general circulation models (GCMs), in particular the work of Cubasch et al.. [1991]. Here S&J make two claims, one of which is incorrect and the other questionable. First, they claim that “the coupled atmosphere‐ocean model of Cubasch et al. [1991] does allow the nonlinearities that Risbey et al.. [1991] criticize our simple model for not including.” In fact we explicitly mentioned changes in polar ice caps [Oerlemans and van der Veen, 1984] and release of methane from clathrates [MacDonald, 1990; Bell, 1982], neither of which are included in the model of Cubasch et al.. [1991]. Indeed, none of the published simulations of global warming using coupled ocean‐atmosphere GCMs include these effects. Nor do these models yet include in their enhanced greenhouse simulations many of the possible feedbacks involving the carbon cycle and biosphere [Lashof, 1989; Bacastow and Maier‐Reimer, 1990; Sellers, 1987] that could significantly alter greenhouse gas concentrations and surface properties. The published simulations with these models do allow for some changes in deep ocean circulation and cloud behavior, but there is controversy over whether they correctly represent these processes [Marotzke, 1991; Mitchell, 1989; Cess, 1990]. In addition the coupled models must be arbitrarily tuned (requiring substantial artificial fluxes of heat and moisture) to get the current climate right [Manabe et al.., 1991; Cubasch et al.., 1991]. Their greenhouse change simulations are at least partly constrained by these flux adjustments.