, 2005), induced by a dynamical adjustment of the oceanic circulation. Large changes of opposite sign in some cases

between these studies are presumably due to different models, parametrization and experimental set-up. Polar regions have also been shown to be affected by these biophysical feedbacks (Gnanadesikan and Anderson, 2009, Lengaigne et al., 2009, Patara et al., 2012 and Wetzel et al., 2006): the surface warming in summer resulting from the spring bloom triggering a reduction of sea-ice thickness and concentration. Manizza (2005) demonstrates that biophysical feedbacks prominently enhances the amplitude of the seasonal cycle of Sea Surface Temperature (SST) and mixed layer depth at the mid and high latitude oceans. This study aims at assessing the respective influence of the physical parameterization changes from OPA8 to NEMOv3.2 along see more with the implementation of the interactive biogeochemical module in the coupled system on the mean climate state. Various aspects of the North Atlantic climate variability has been studied in both versions of the model and were shown to be very similar: the atmospheric variability (Msadek and Frankignoul, 2008Gastineau et al., 2012), multidecadal SST variability (Msadek and Frankignoul, 2008 and Persechino et al., 2012Marini and Frankignoul, 2013) air–sea interactions (Gastineau

and Frankignoul, 2011). Bi-decadal energy peak in the North Atlantic is present in both versions (Born and Mignot, 2011 and Escudier

et Anticancer Compound Library al., 2012), although with different mechanisms, as well as in piCtrl_noBio. Extensive comparison of CMIP3 and CMIP5 variability patterns in the Pacific shows that both versions correlate very well with observations (Lengaigne, pers. com.). They are also fairly similar in terms of El Niño-Southern Oscillations characteristics (Bellenger et al., 2013). Section 2 describes the model configurations and the experiments used for this purpose. Section 3 analyses a series of sensitivity tests with ocean-only simulations while coupled models are analysed in Sections 4 and 5. The effect of implementing the biogeochemical module is firstly analysed separately, as it appears to be very important and sometimes contradictory with previous studies. Conclusions are given in Section Celecoxib 6. This study focuses on the outcomes of two sets of simulations, the first one using ocean simulations forced by atmospheric reanalyses while the other ones are coupled to other components of the IPSL earth system model. All simulations use the global Océan Parallèlisé (OPA) ocean general circulation model (OGCM, Madec et al., 1999). This model solves the primitive equations on the Arakawa C grid, with a second order centred finite difference scheme. It assumes the Boussinesq and hydrostatic approximations, the incompressibility hypothesis, and uses a free-surface formulation (Roullet and Madec, 2000).