The deep reservoir model with one-dimensional assumptions can be applied to a reservoir or lake where the principal variation of flow characteristics is in the vertical direction. Among the models evaluated, the MIT deep reservoir model appears to be most easily used and to give results most compatible with the measured temperatures. The temperature predicted is strongly dependent upon the magnitude of the absorption coefficient of water, and the diffusion coefficient. However, our sensitivity analysis shows that an absorption coefficient of about 0.75/m and a diffusion coefficient of 15 to 20 times molecular diffusion are appropriate choices for the seven TVA reservoirs studied. The determination of whether or not a reservoir model depends on the Densimetric Froude number. However, the representativeness of the result is not solely dependent upon the Densimetric Froude number. By the use of a fitted curve to the measured temperatures, it was possible to determine the maximum standard error of estimate for the predicted outlet level temperature, 1.6C. Temperatures on individual days may exceed these values and they surely are exceeded at other depths in the reservoir. These limits are suggested as the limit of accuracy of these types of models.

"EPA-660/3-75-038"--Cover "June 1975"--Cover "Vanderbilt University." "Grant no. R-800613; program element 1BA032; ROAP 21AJH/Task 12." "Project Officer: Bruce Tichenor." "Pacific Northwest Environmental Research Laboratory, National Environmental Research Center." "This report was submitted in fulfillment of Grant R-800613 by Vanderbilt University, Nashville, Tennessee, under the sponsorship of the Environmental Protection Agency. Work was completed as of June 1975"--Page ii Includes bibliographical references. Sponsered by the Environmental Protection Agency