We estimated the effects of a temperature control device (TCD) on a suite of thermodynamic and limnological attributes for a large storage reservoir, Shasta Lake, in northern California. Shasta Dam was constructed in 1945 with a fixed-elevation penstock. The TCD was installed in 1997 to improve downstream temperatures for endangered salmonids by releasing epilimnetic waters in the winter/spring and hypolimnetic waters in the summer/fall. We calibrated a two-dimensional hydrodynamic reservoir water quality model, CE-QUAL-W2, and applied a structured design-of-experiment simulation procedure to predict the principal limnological effects of the TCD under a variety of environmental scenarios. Calibration goodness-of-fit ranged from good to poor depending on the constituent simulated, with an R2 of 0.9 for water temperature but 0.3 for phytoplankton. Although the chemical and thermal characteristics of the discharge changed markedly, the reservoir's characteristics remained relatively unchanged. Simulations showed the TCD causing an earlier onset and shorter duration of summer stratification, but no dramatic affect on Shasta's nutrient composition. Peak inreservoir phytoplankton production may begin earlier and be stronger in the fall with the TCD, while outfall phytoplankton concentrations may be much greater in the spring. Many model predictions differed from our a priori expectations that had been shaped by an intensive, but limited-duration, data collection effort. Hydrologic and meteorological variables, most notably reservoir carryover storage at the beginning of the calendar year, influenced model predictions much more strongly than the TCD. Model results indicate that greater control over reservoir limnology and release quality may be gained by carefully managing reservoir volume through the year than with the TCD alone.