HYCOM was run with KPP mixing and relaxation boundary conditions to produce the “best” possible simulation. Many of the observed differences between the simulated and observed fields can be attributed at least in part to the lack of river runoff and sea ice in the model, not to mention forcing errors and the poor estimate of observed mixed layer thickness obtained from the smooth Levitus climatology. Errors in the simulated dynamical fields are partly related to the errors in the simulated thermodynamical fields.