Fish Attraction Devices (FADs) and experimental designs

Abstract

There is widespread use of fish attraction devices (FADs) in commercial fisheries and research. Investigations on the utility of FADs to catch fishes, and factors influencing fishes associated with FADs, require careful consideration of experimental designs. The development of appropriate models, from observations and the literature, should be developed before hypotheses can be tested with robust sampling designs. Robust sampling designs may only be possible if investigators have some role in the planning stage of deploying FADs. If the objective of the study is to determine the influence of FADs on assemblages of fishes, then experimenters need to consider that a `FAD-effect´ (=impact) cannot be demonstrated without controls. Some preliminary studies may be required to determine the spatial extent of a FAD-effect before suitable sites can be chosen for controls. Other controls may also be necessary, depending on the method used to estimate numbers of fishes (e.g. controls for disturbance). Recent advances in sampling designs that are applicable to impact studies are discussed. Beyond-BACI (Before After Control Impact) and MBACI (Multiple BACI) designs are recommended because they cater for temporal and spatial variation in the abundance of organisms, which is generally great for pelagic fishes. The utility of orthogonal sampling designs is emphasised as a means of elucidating the influence of multiple factors and, importantly, interactions between them. Further, nested analyses are suggested to deal with multiple temporal and/or spatial time scales in sampling designs. The independence of replicate FADs should also be considered. Problems of independence include: FADs that are connected, thus providing potential routes of movement of associated fishes; temporal dependence where the number of fish at a time influences the number at the next time due to fish becoming residents; and the fact that the proximity of other FADs may influence numbers of fishes. Solutions to these problems are suggested. The recommendations in this paper are relevant to investigations on pelagic fish of all sizes, with and without FADs.