A tidally and cross-sectionally averaged model based on the temporal evolution of the quasi-steady Hansen and Rattray equations is applied to simulate the salinity distribution and vertical exchange flow along the Hudson River estuary. The model achieves high skill at hindcasting salinity and residual velocity variation during a 110-day period in 2004 covering a wide range of river discharges and tidal forcing. The approach is based on an existing model framework that has been modified to improve model skill relative to observations. The external forcing has been modified to capture meteorological time-scale variability in salinity, stratification, and residual velocity due to sea level fluctuations at the open boundary and along-estuary wind stress. To reflect changes in vertical mixing due to stratification, the vertical mixing coefficients have been modified to use the bottom boundary layer height rather than the water depth as an effective mixing length scale. The boundary layer parameterization depends on the tidal amplitude and the local baroclinic pressure gradient through the longitudinal Richardson number, and improves the model response to spring–neap variability in tidal amplitude during periods of high river discharge. Finally, steady-state model solutions are evaluated for both the Hudson River and northern San Francisco Bay over a range of forcing conditions. Agreement between the model and scaling of equilibrium salinity intrusions lends confidence that the approach is transferable to other estuaries, despite significant differences in bathymetry. Discrepancies between the model results and observations at high river discharge are indicative of limits at which the formulation begins to fail, and where an alternative approach that captures two-layer dynamics would be more appropriate.