First of all, although it is clear that the oceanic adjustment requires several hundreds of years, this figure illustrates that all simulations approaches an equilibrium state after 300 years. The latter is nevertheless not reached and may require thousands of years, as it was necessary in CM5_piCtrl. CM5_piStart ends to a globally colder state for the upper ocean than CM5_RETRO, but it is still warmer than the corresponding CM5_piCtrl, further suggesting that the model is not fully equilibrated. The Antarctic
circumpolar current (ACC) at Drake Passage is stronger in CM5_piStart than in CM5_RETRO while the magnitudes of the AMOC maximum are similar. These dynamical adjustments will be analysed in Section 5 by comparing the last 92 years of CM5_piStart and CM5_RETRO (yrs 400–491 of these experiments). Given that the two simulations start from the OSI-906 concentration same initial conditions, comparing these relatively short simulations still gives insight in the changes of simulated
mean climate. To evaluate the effect of the interactive chlorophyll concentration variations related to the inclusion of the biogeochemical component PISCES in IPSL-CM5A as compared to IPSL-CM4 (Section 2a), we performed a sensitivity experiment (called CM5_piCtrl_NoBio) with a set up identical to CM5A_piCtrl, except for the chlorophyll concentration within the ocean which was fixed in time GSI-IX datasheet and space to its AZD9291 mw global mean value of 0.05 mg/m3 (see Table 1). This value is assumed to be representative to a globally-averaged
surface chlorophyll concentration estimated from satellite measurement. This set up aims at evaluating how chlorophyll bio-optical properties impact the ocean thermal structure and circulation. This simulation differs from CM4_piCtrl through the atmospheric and oceanic parameterizations, the atmospheric resolution, but also from the treatment of light penetration into the ocean: the simple 2-waveband scheme assumed for the downward irradiance in IPSL-CM4 is replaced by the RGB formulation described above in both CM5_piCtrl and CM5_piCtrl_noBio simulations. CM5_piCtrl_noBio was run for 350 years, starting from year 1800 in CM5_piCtrl. Differences between these two simulations are described in Section 4. Note that all coupled simulations were run under constant pre-industrial boundary conditions. Furthermore, no specific tuning of the model in general and of the atmosphere in particular was done when plugging the different versions of the oceanic and biogeochemistry model. The tuning is thus identical to CM5_piCtrl. As displayed on Fig. 1, CM5_piCtrl_noBio (green curve) stabilizes to a warmer global upper ocean state than CM5_piCtrl.