The seasonally averaged zonal momentum equation tendencies at 140W are studied in a high-resolution primitive equation ocean general circulation model simulation of the tropical Pacific. The model experiment, forced by climatological monthly average wind-stress, reproduces well the observed boreal springtime eastward surge of the normally westward surface flow, as well as many features of the acceleration and deceleration between the surface and 200 m between January and October. We present each of the zonal momentum equation tendency terms for the depth range 0-160 m, but our discussion focuses on the behavior of the boreal springtime near-surface flow, perhaps the most distinctive feature of the seasonal cycle. The eastward surface surge in boreal spring depends crucially on the springtime weakening of the otherwise westward tendency from tropical instability waves (TIWs). The TIW effects, together with the eastward tendency from the seasonal weakening of the easterly wind-stress, drive the eastward surface current surge. Although the 'negative viscosity' effect of the TIWs is small in the annual mean, as we have previously shown, its seasonal variation is necessary to the surface flow reversal and eastward surge in this model. A series of experiments, each with weaker TIWs than its predecessor, shows a progressive weakening and eventual absence of springtime eastward surface flow, supporting the above analysis. The seasonal zonal velocity accelerations and decelerations are small compared with the terms in the zonal momentum equation; these terms must be known to an accuracy of at least 10 cm s-1 month-1 (2-5% of the largest terms) if a meaningful budget is to be obtained. This is a strong constraint that must be planned for in future observational studies. We find that nonlinear terms are O(1) in the vertically-integrated balance as well as the local balance, in contrast with some recent observational estimates. Extrapolated velocity errors, neglected terms, data processing assumptions, and crude finite-differencing in the observational studies may account for the differences, as appeared to be the case in the annual mean balances. The model dynamical balances cannot be reproduced if the methods used to analyze observational data are applied to the model output fields. Very near-surface currents must be measured rather than extrapolated if the ocean shear is similar to that of the model flows.
All Science Journal Classification (ASJC) codes