Ground water in the 3,458-square-mile lower Susquehanna River basin occupies secondary openings in bedrock. The distribution of openings is a function of lithology, depth, and topography. Local flow systems account for most of the total ground-water flow. Average annual recharge for the lower basin is 1,857 million gallons per day, most of which discharges to streams. The water table is a subdued replica of land surface; its depth varies with topography but is generally 20 to 70 feet below land surface. Ground water circulates to depths of 500 to 600 feet below the water table. A digital model of regional, unconfined groundwater flow was developed and used to evaluate the ground-water resources of the lower basin. On the basis of lithologic and hydrologic differences, the area was subdivided into 21 hydrogeologic units, each with different hydrologic characteristics. Each unit was divided into two layers to take into account decreasing secondary permeability with depth. A finite-difference grid with square blocks approximately 1 mile on a side was used. The model was calibrated under steady-state and transient conditions. In the steady-state calibration, the model-generated results were compared with estimated water-table altitudes and estimated base flows. In the transient calibration, the model-generated results were compared with observed changes in water-table altitude from November 1, 1980, through April 22, 1981. Hydraulic conductivity increases from hilltops to Valley bottoms. The average hydraulic conductivity for carbonate units is about 21 feet per day, which is an order of magnitude greater than the corresponding averages for Paleozoic sedimentary, Triassic sedimentary, and crystalline units. The Cumberland Valley carbonate rocks have the greatest average hydraulic conductivity-about 174 feet per day in valley bottoms. The average gaining-stream leakage coefficient for all carbonate units is about 16 feet per day, which is two orders of magnitude greater than the corresponding averages for the other lithologies. The Cumberland Valley carbonate rocks have the greatest gaining-stream leakage coefficient--about 43 feet per day. The specific yields are 0.035, 0.020, 0.020, and 0.007 for the carbonate, Paleozoic sedimentary, crystalline, and Triassic sedimentary units, respectively. The calibrated model was used to simulate the effects of a ground-water withdrawal of 1 inch per year on water-table altitudes and average annual base flows in the modeled area. The overall effect is least for the carbonate units and greatest for the Triassic sedimentary units. The model also was used to simulate a standardized potential yield for each unit by assuming that the maximum acceptable consequence of a hypothetical withdrawal scheme is an ultimate 50-percent reduction in average annual base flow. Based on this, the potential yield for the modeled area is 891 million gallons per day. The Cumberland Valley carbonate rocks have the greatest potential yield--0.47 million gallons per day per square mile. The carbonate units have the greatest average potential yield, followed by the Paleozoic sedimentary, crystalline, and Triassic sedimentary units. About 90 percent of the eventual decline in water-table altitudes and the eventual reduction in average annual base flows occurs within 5 years of the implementation of the hypothetical withdrawal scheme. Nearly all of the ground water withdrawn is derived from reduced discharge to streams. The calibrated model can be used to estimate the impacts of ground-water development schemes on regional ground-water levels and base flows of streams, it cannot be used to simulate local cones of depression or local base-flow changes. The reliability of the model is a function of its approximation of the physical characteristics of the ground-water flow system, the two calibrations, various simplifying assumptions, and the lack of calibration under ground-water withdrawal conditions, it ca