A literature survey was conducted to evaluate the significance of selective withdrawal models available for impounded water. Field investigations were made at several lakes and Lake Livingston to substantiate these models and to determine the changes in water quality of the Trinity River as Lake Livingston fills. The Koh-Debler and Bohan-Grace models were used to compare data from the literature with observed results from Lake Travis. The various inputs of data and the effect of varying these inputs upon the predicted water outlet temperature were investigated. It was concluded that a broader range of tests must be run before either model can be applied with confidence to Southwestern United States impoundments. During the tests for water quality at Lake Livingston it was found that a high influx of oxygen demanding substances and nutrients enter the reservoir via the Trinity River. Turnover of water in the reservoir also affected water quality. Selective withdrawal did not appear to be feasible for water depths less than 40 feet but the data indicated the selective withdrawal programs at greater depths will be successful. (WRSIC abstract)