Prior to the early 1990's, ground-water in the Kona area, which is in the western part of the island of Hawaii, was withdrawn from wells located within about 3 mi from the coast where water levels were less than 10 feet above sea level. In 1990, exploratory drilling in the uplands east of the existing coastal wells first revealed the presence of high water levels (greater than 40 feet above sea level) in the Kona area. Measured water levels from 16 wells indicate that high water levels exist in a zone parallel to and inland of the Kona coast, between Kalaoa and Honaunau. Available hydrologic and geophysical evidence is generally consistent with the concept that the high ground-water levels are associated with a buried dike complex. A two-dimensional (areal), steady-state, freshwater-saltwater, sharp-interface ground-water flow model was developed for the Kona area of the island of Hawaii, to enhance the understanding of (1) the distribution of aquifer hydraulic properties, (2) the conceptual framework of the ground-water flow system, and (3) the regional effects of ground-water withdrawals on water levels and coastal discharge. The model uses the finite-difference code SHARP. To estimate the hydraulic characteristics, average recharge, withdrawals, and water-level conditions for the period 1991-93 were simulated. The following horizontal hydraulic-conductivity values were estimated: (1) 7,500 feet per day for the dike-free volcanic rocks of Hualalai and Mauna Loa, (2) 0.1 feet per day for the buried dike complex of Hualalai, (3) 10 feet per day for the northern marginal dike zone (north of Kalaoa), and (4) 0.5 feet per day for the southern marginal dike zone between Palani Junction and Holualoa. The coastal leakance was estimated to be 0.05 feet per day per foot. Measured water levels indicate that ground water generally flows from inland areas to the coast. Model results are in general agreement with the limited set of measured water levels in the Kona area. Model results indicate, however, that water levels do not strictly increase in an inland direction and that a ground-water divide exists within the buried dike complex. Data are not available, however, to verify model results in the area near and inland of the model-calculated ground-water divide. Three simulations to determine the effects of proposed withdrawals from the high water-level area on coastal discharge and water levels, relative to model-calculated, steady-state coastal discharge and water levels for 1997 withdrawal rates, show that the effects are widespread. During 1997, the total withdrawal of ground water from the high water-level area between Palani Junction and Holualoa was about 1 million gallons per day. Model results indicate that it may not be possible to withdraw 25.6 million gallons per day of freshwater from this area between Palani Junction and Holualoa, but that it may be possible to withdraw between 5 to 8 million gallons per day from the same area. For a proposed withdrawal rate of 5.0 million gallons per day uniformly distributed to 12 sites between Palani Junction and Holualoa, the model-calculated drawdown of 0.01 foot or more extends about 9 miles north-northwest and about 7 miles south of the proposed well sites. In all scenarios, freshwater coastal discharge is reduced by an amount equal to the additional freshwater withdrawal. Additional data needed to improve the understanding of the ground-water flow system in the Kona area include: (1) a wider spatial distribution and longer temporal distribution of water levels, (2) improved information about the subsurface geology, (3) independent estimates of hydraulic conductivity, (4) improved recharge estimates, and (5) information about the vertical distribution of salinity in ground water.