A series of projections of climate change were applied to a watershed model of the Delaware River basin to identify sources of uncertainty in predicting effects of climate change on drought in the basin as defined by New York City reservoir contents. The watershed model is a calibrated, monthly time-step water-balance model that incorporates the operation of reservoirs and diversion canals, and accounts for all inflows to and outflows from the basin at several key nodes. The model assesses the effects of projected climate change on reservoir contents by calculating the frequency with which the basin enters drought conditions under a range of climate-change conditions. Two primary sources of uncertainty that affect predictions of drought frequency in the Delaware River basin were considered: (1) uncertainty in the amount of change in mean air temperature and precipitation, and (2) uncertainty in the effects of natural climate variability on future temperature and precipitation. Model results indicate that changes in drought frequency in the Delaware River basin are highly sensitive to changes in mean precipitation; therefore, the uncertainty associated with predictions of future precipitation has a large effect on the prediction of future drought frequency in the basin.