Time variation of climate sensitivity in GCMs

Abstract 1 Eectiv e climate sensitivity is often assumed to be constant, but previous studies 2 of General Circulation Model (GCM) simulations have found it varying as the simu- 3 lation progresses. This complicates the tting of simple models to such simulations, 4 as well as having implications for the estimation of climate sensitivity from observa- 5 tions. This study examines the evolution of the feedbacks determining the climate 6 sensitivity in GCMs submitted to the Coupled Model Intercomparison Project. Ap- 7 parent centennial timescale variations of climate sensitivity during stabilisation to 8 a forcing can be considered an artifact of using conventional forcings, allowing only 9 for instantaneous eects and stratosphere-adjustment. If the forcing is adjusted for 10 processes occurring on timescales which are short compared to the climate stabilisa- 11 tion timescale then there is little centennial timescale evolution of climate sensitivity 12 in any of the GCMs. We suggest that much of the apparent variation in climate 13 sensitivity identied in previous studies is actually due to the comparatively fast 14 forcing adjustment. Persistent dierences are found in the strength of the feedbacks 15 between the coupled atmosphere{ocean (AO) versions and their atmosphere{mixed- 16 layer ocean (AML) counterparts, (the latter are often assumed to give the climate 17 sensitivity of the AOGCM). This indicates that the response of the dynamic ocean 18 inuences the atmospheric feedbacks in some GCMs and that the AML model may 19 not provide a good estimate of the equilibrium climate sensitivity for the parallel 20 AO version. The adjustment to the forcing to account for comparatively fast pro- 21 cesses varies in magnitude and sign between GCMs, as well as diering between AO 22 and AML versions of the same model. There is evidence from one AOGCM that 23