A two-dimensional computer ground-water model was constructed of the Manati-Vega Baja area to improve the understanding of the unconfined upper aquifer within the North Coast Province of Puerto Rico. The modeled area covers approximately 79 square miles within the municipios of Manati and Vega Baja and small portions of Vega Alta and Barceloneta. Steady-state two-dimensional ground-water simulations were correlated to conditions prior to construction of the Laguna Tortuguero outlet channel in 1940 and calibrated to the observed potentiometric surface in March 1995. At the regional scale, the unconfined Upper North Coast Limestone aquifer is a diffuse ground-water flow system through the Aguada and Aymamon limestone units. The calibrated model input parameters for aquifer recharge varied from 2 inches per year in coastal areas to 18 inches per year in the upland areas south of Manati and Vega Baja. The calibrated transmissivity values ranged from less than 500 feet squared per day in the upland areas near the southern boundary to 70,000 feet squared per day in the areas west of Vega Baja. Increased ground-water withdrawals from 1.0 cubic foot per second for 1940 conditions to 26.3 cubic feet per second in 1995, has reduced the natural ground-water discharge to springs and wetland areas, and induced additional recharge from the rivers. The most important regional drainage feature is Laguna Tortuguero, which is the major ground-water discharge body for the upper aquifer, and has a drainage area of approximately 17 square miles. The discharge to the sea from Laguna Tortuguero through the outlet channel has been measured on a bi-monthly basis since 1974. The outflow represents a combination of ground- and surface-water discharge over the drainage area. Hydrologic conditions, prior to construction of the Laguna Tortuguero outlet channel in 1943, can be considered natural conditions with minimal ground-water pumpage (1.0 cubic foot per second), and heads in the lagoon were 2.4 feet higher. The model was calibrated to March 1995 conditions during a dry period of minimal aquifer recharge and relatively constant water levels in the upper aquifer. For the steady-state 1995 model simulation, however, ground-water pumpage had been increased to 26.3 cubic foot per second, due to increased demand for public water supply, the heads at 0.9 feet, and the outflow to the sea at Laguna Tortuguero had been lowered considerably. Simulated ground-water inflow for 1940 hydrologic conditions included 35.9 cubic feet per second from areal recharge, contributions from streamflow along the southern boundary of 1.6 cubic feet per second, and streamflow infiltration to the upper aquifer of 4.2 cubic feet per second. Simulated ground-water outflow for 1940 hydrologic conditions are discharge to springs of 17.4 cubic feet per second, total ground-water withdrawals of 1.0 cubic feet per second, and aquifer contribution to streamflow or wetland areas of 23.4 cubic feet per second. Simulated ground-water inflow for hydrologic conditions of March 1995 include d contributions from streamflow along the southern boundary of 1.6 cubic feet per second, areal recharge of 35.9 cubic feet per second, and streamflow infiltration to the upper aquifer of 11 cubic feet per second. Simulated ground-water outflow for hydrologic conditions of March 1995 are ground-water withdrawals of 26.3 cubic feet per second, discharge from springs of 7.3 cubic feet per second, and aquifer contribution to streamflow or wetland areas of 14 .9 cubic feet per second. The overall ground-water budget increased from 41.8 cubic feet per second for 1940 conditions to 48.6 cubic feet per second for the hydrologic conditions of March 1995. The increase in ground-water budget is a direct result of increased ground-water withdrawals, which induced greater streamflow infiltration. Simulated ground-water flux to Laguna Tortuguero for 1940 conditions was 11 cubic feet per second, which drop