Conclusions (part 2)
	n Models simulations must have n Accurate mixing and surface fluxes n >year model integrations n Resolved measurements of inflow density n Or models must assimilate measurements of the density field in Gulf of Maine. n Quarterly observations (almost) sufficient.