An assessment is made of the ability of CMIP5 models to represent the seasonal biogeochemical cycles over the late twentieth century in the southwest Pacific Ocean. In particular, sea surface temperature (SST), surface chlorophyll a, nitrate, phosphate, silicate, and the depth of the seasonal thermocline, are examined to quantify the physical-biogeochemical capabilities of each model; the result is a ranking estimate enabling model ensemble generation. The better/less ranked ensembles we refer to as inner/outer, respectively. The ensembles then allow less well-observed variables such as iron and vertically integrated primary production to be assessed. The assessment establishes model output confidence limits for setting bounds on future model scenario ecosystem change projections. By the end of the twenty first century under Representative Concentration Pathways (RCP) RCP4.5 and/or RCP8.5, our best estimates suggest that there will be average domain wide increases in SST and surface iron, but average decreases in surface chlorophyll a, nitrate, and phosphate, accompanied by relatively large decreases in the depth of the seasonal thermocline (all changes realized by both ensembles). On the other hand, for surface silicate the inner ensemble suggests general declines, and vice versa for the outer ensemble. For integrated primary production, the ensembles predict declines in subtropical water, but elsewhere generally less significant changes.