Multiple Regression Analysis for Forecasting Critical Fish Influxes at Power Station Intakes

(1) Certain coastal power stations in Britain are subject to ingresses of fish which may jeopardize plant operation or which may cause temporary local depletion of fish populations. (2) There have been attempts in the United States and in Europe to develop costly and often ineffectual exclusion devices. As an alternative, it is proposed that, for sites where the problem is intermittent, statistical models should be developed which allow critical levels of fish ingress to be predicted and the necessary reduction in cooling water flow determined. (3) The procedure for developing a multiple regression model is described and illustrated using fish impingement data collected from Fawley Power Station. Over 60% of the overall variation was accounted for in terms of a small number of easily measured variables (cooling water flow, water temperature, tide height, wind speed). In other research work, values of up to 80% have been quoted. With refinements of the model formulation higher values should be possible and the approach then would have practical application.