Reported are diverse efforts to verify, or test the accuracy of,a numerical model for thermal pollution predictions. The model,developed by Prych (1972) at SMHI, is an integral analysis ofa buoyant, turbulent surface jet. The program. is slightly revised,and effects of the changes are presented.

Numerical model results are compared directly with ten prototype situations: eight sets of data from a Swedish power station on the coast of the Baltic Sea, and data from two power plants on Lake Michigan in the United States. A few direct comparisons are made with a laboratory model of the Swedish plant and with laboratory data for momentumless surface jets. To minimize coastal effects, a scheme is presented for interfacing then umerical model with undistorted laboratory model results.

The discharge situation is described nondimensionally, and numerical experiments are made to determine the sensitivities of jet temperatures to input variables. Results are compared with phenomenological and analytical models. In addition to the main verification effort, investigations are made concerning methods of field data collection.

Conclusions are presented for direct application and for relative or comparative use of the numerical model.