In Australia, catch-per-unit-effort (CPUE) is the only index of relative abundance in the majority of fisheries; many smaller, data-poor fisheries have none at all. For CPUE to be even partially successful as an index of relative abundance it obviously needs to have some relationship with a stock’s abundance. This need not necessarily be linear but as abundance goes up and down so too should CPUE. Because there can be many influences on CPUE other than a stock’s biomass, statistical standardization is now becoming the accepted practice. Such analyses of CPUE tend to focus on extracting the up and down trends through time in the hope of improving our understanding of a stock’s dynamics. Instead of examining trends, here we will conduct an empirical exploration of the properties of CPUE across a number of fisheries. Gaining an understanding of these properties provides insights into how best to analyze CPUE without expecting more than it can give.
The reporting of CPUE in logbooks usually involves estimating the catch of each species caught and the effort expended to catch them. Mistakes can be made at many stages of the process leading to outliers, errors, and misleading data and estimates of catch and effort are often rounded off rather than measured precisely. Not all such potential problems can be solved by judicious selection of data but awareness, acknowledgement, and documentation of their existence means that analyses can proceed without unreasonable expectations. We examine some of the main issues with the use of CPUE, compare some of the major methods of statistical standardization, and conclude by attempting to list some generalized guidelines for how best to use CPUE. In the end we should obtain something useful for fisheries management rather than either a pig’s pinna or a silken bag.