Surface waves modify upper ocean dynamics through Stokes drift related processes. Representation of these processes at either resolved or parameterized scales in an ocean model depends on accurate estimation of Stokes drift profiles. Stokes drift estimated from a discrete wave spectrum is compared to Stokes drift approximations as a monochromatic profile based on bulk surface wave parameters, and to two additional super-exponential functional forms. The impact of these different methods on resolved-scale ocean processes is examined in the context of two test-bed cases of a wave-current coupled system: (1) a shallow water inlet test case and (2) an idealized deep water hurricane case. In case (1), tidal currents can modify the waves and significantly affect Stokes drift computed from the wave spectrum. In both cases, large deviations in ocean current response are produced when the Stokes drift is approximated monochromatically from bulk wave parameters, rather than from integration over the wave spectra. Deep water simulations using the two super-exponential approximations are in better agreement with those estimated from wave spectra than are those using the monochromatic, exponential profile based on bulk wave parameters. In order to represent the impact of Stokes drift at resolved scales, we recommend that for studies of nearshore processes and brief deep water events, like wave-current interactions under storms, the Stokes drift should be calculated from full wave spectra. For long simulations of open ocean dynamics, methods using super-exponential profiles to represent equilibrium wind seas might be sufficient, but appear to be marginally more computationally efficient.