Simulation models coupling physics to biological processes in the ocean are central to many current programs. Ocean physical models have approached a high level of sophistication; the physical relationships are canonical, and modern computational technology for fluid mechanics has advanced steadily for two generations or more. The complexity of biological processes in the ocean presents enormous difficulties beyond physics. There is a recognizable mode of operation wherein ‘complete’ physics is coupled to reduced-complexity biology; and simulations are typically chosen to fit field problems and available data. The upshot of this situation is a great diversity in what is possible in ‘replicating observations’, and even more importantly, in assimilating them into simulations and creating forecast systems.