Risk Analysis for Fish Diversion Experiments: Pumped Intake Systems

Publisher Website
Google Scholar
Abstract
Such facilities as power‐generating stations, public water systems, and ore‐processing plants draw off large quantities of water from estuaries, coastal seawaters, lakes, and rivers. In turn, large numbers of fishes are often drawn into these pumped intake systems and killed if not otherwise removed or diverted. The large mortalities associated with many intake systems threaten the perpetuation of indigenous stocks. The diversion and removal devices most commonly used for protecting fish life from such risks are presumed to operate on principles shown here to be erroneously conceived. In consequence of these faulty theories, the estimators and experimental designs of standard industry practice seldom reveal the true correlations necessary for improvements in fish conservation systems, nor do the assessments of small‐scale experiments extend with reliability to full‐scale system designs. Such passive devices as angled barrier screens are thought to guide fish in some way into pumped bypassing ducts, but an analysis of existing data supports instead a hypothesis of random encounter whereby the activities of entrapped fish are governed by the probabilistic mechanics associated with random walks and unlike boundary conditions. Experimental designs for separating and assessing time‐dependent risks are developed for the case of competing devices in a given test system. From time‐dependent comparisons between large and small systems (in particular, between a model system and its full‐scale prototype), the net decrease in the probability of fish survival associated with increased system size is shown to be the consequence of increased exposure to the risk of death (increased residence time) rather than the consequence of increases in the unit risk of death itself. Where extensions of small‐scale empirical results to full‐scale system designs are wanted, arithmetic extrapolations yield erroneous results. Because displacement dependence (fish movement and system size) enters the risk analysis, the scaling problem must be resolved instead from a corresponding system of partial differential equations.

This publication has 1 reference indexed in Scilit: