The U.S. Geological Survey has developed several ground-water models in support of an investigation of ground-water contamination being conducted by the Army National Guard Bureau at Camp Edwards, Massachusetts Military Reservation on western Cape Cod, Massachusetts. Regional and subregional steady-state models and regional transient models were used to (1) improve understanding of the hydrologic system, (2) simulate advective transport of contaminants, (3) delineate recharge areas to municipal wells, and (4) evaluate how model discretization and time-varying recharge affect simulation results. A water-table mound dominates ground-water-flow patterns. Near the top of the mound, which is within Camp Edwards, hydraulic gradients are nearly vertically downward and horizontal gradients are small. In downgradient areas that are further from the top of the water-table mound, the ratio of horizontal to vertical gradients is larger and horizontal flow predominates. The steady-state regional model adequately simulates advective transport in some areas of the aquifer; however, simulation of ground-water flow in areas with local hydrologic boundaries, such as ponds, requires more finely discretized subregional models. Subregional models also are needed to delineate recharge areas to municipal wells that are inadequately represented in the regional model or are near other pumped wells. Long-term changes in recharge rates affect hydraulic heads in the aquifer and shift the position of the top of the water-table mound. Hydraulic-gradient directions do not change over time in downgradient areas, whereas they do change substantially with temporal changes in recharge near the top of the water-table mound. The assumption of steady-state hydraulic conditions is valid in downgradient area, where advective transport paths change little over time. In areas closer to the top of the water-table mound, advective transport paths change as a function of time, transient and steady-state paths do not coincide, and the assumption of steady-state conditions is not valid. The simulation results indicate that several modeling tools are needed to adequately simulate ground-water flow at the site and that the utility of a model varies according to hydrologic conditions in the specific areas of interest.