A numerical model described and verified in part 1 of this two-part series (Cooper and Thompson, this issue) is applied to study the sensitivity of hurricane-generated currents on the outer shelf and slope. Numerical experiments are performed in a simple basin with a straight shelf. The sensitivity of the response to changes in storm parameters, direction of storm approach, and topography is quantified. Response is measured in terms of the mixed-layer velocity and depth at sites along the storm track. Results reveal the most important factors are (in decreasing order) wind speed, storm translation speed, direction of storm approach, asymmetry in the wind field, entrainment parameterization, and advection at slower storm translation speeds. Response is largely insensitive (less than 10%) to radius of maximum wind, shelf and slope configuration, bottom friction, atmospheric pressure gradients, and further reductions in the model grid size. For a storm approaching cross shelf, the response is primarily baroclinic (greater than 90%) and only weakly dependent (less than 10%) on the water depth at the site. ¿ American Geophysical Union 1989