Variability in the circulation of coastal oceans must ultimately be driven by changes in the meteorological conditions that force currents in the coastal ocean, and by variability in the waters entering the coastal ocean from elsewhere. If a coastal ocean is to be understood and modeled accurately, the external forcing that drives the largest portions of the circulation variability must be observed adequately. Thus, some method of comparing the relative importance of various sources of circulation variability must be developed, so that there is confidence that the most important sources of variability are included in any analysis or modeling. This is done for variability in the time-integrated transport across a section in a coastal ocean.

The relative importance of various sources of circulation variability in the Gulf of Maine (GoM) are then quantified, with an emphasis on variability on timescales longer than tidal or weather-band timescales. In order to concentrate on interannual changes, the seasonal cycle is not included in estimates of variability. It is found that the variability forced by fluctuations in the winds and the volume of water entering from the Scotian Shelf to the GoM produce roughly comparable amounts of circulation variability. However, changes in the density structure of the GoM produce changes in time-integrated transport that are an order of magnitude larger, at least in the central GoM. The changes in the large-scale density gradients are governed by mixing processes in the Gulf and by changes in the water masses entering the GoM from the Scotian Shelf and the Northeast Channel. Unless the heat, freshwater and volume transport of the waters entering the GoM are routinely observed, numerical models will fail to capture much of the variability in the circulation of the Gulf. An analysis is given of the minimal set of observations needed to allow numerical models of the GoM to resolve adequately the true variability in the circulation.