An efficient numerical model is used to solve the linear barotropic equations of motion with North Pacific bottom topography and seasonal wind forcing (that is, monthly mean wind stresses with the annual mean removed). The model domain extends from 10¿S to 60¿N and from 120¿E to 100¿W with 1¿¿1¿ resolution. The model-predicted seasonal transport and sea level variations are described and compared with the available data. It is found that at the Tokara Strait, to the south of Japan, the model predicted seasonal transport variations are very similar to those of the observed annual cycle of sea level difference across the Strait, although the amplitude of the model response is less than that suggested by the data. Both model and data indicate enhanced flow of the Kuroshio through the Strait in summer with a minimum in that flow in the winter. This is 6 months out of phase with the seasonal transport variations predicted by the flat-bottomed Sverdrup relation. The model response is very sensitive, however, to the smoothing of the topography, with additional peaks in transport being found in the spring and the fall when the smoothing is increased. This suggests that care in handling the topography is required if a general circulation model is to reproduce this seasonal cycle. Away from the western boundary, the model-predicted transport variations are quite similar to those in a flat-bottomed ocean. The model aslo shows some success at reproducing features of the observed annual cycle of sea level corrected for both atmospheric pressure variations and the steric density effect. An interesting feature of the results is that the model-calculated sea level at the offshore edge of the coastal waveguide is remarkably similar all the way along the North American coast from the Aleutian Islands to California. This shows a peak in model sea level in the summer and a minimum in the winter and indicates that it is the seasonal fluctuations of the subtropical gyre which dominate the model response wven in the northern latitudes.

However, the amplitude of this signal is somewhat less than that observed at the coast which, as previous studies have shown, is strongly influenced by coastal effects. The importance of wind forcing of the coastal waveguide is also apparent in the simple model described here, as is demonstrated by comparing model-calculated sea level at the coast with that at points offshore. ¿ American Geophysical Union 1989