Spatial scales of variation of assemblages of fish in coastal lakes sampled with multi-mesh gillnets: Implications for designing research surveys

A key element in designing surveys to assess populations and assemblages of fish is to understand their spatial and temporal scales of variation. As part of a strategy to develop fishery-independent surveys, a pilot experiment was done in two coastal lakes in south-eastern Australia to examine scales of spatial variation among assemblages of fish caught in multi-mesh gillnets. A hierarchical, nested sampling design measured spatial variation among zones (2–20 km apart), sites within zones (1 km apart) and replicate gillnets at each site (50–100 m apart), in shallow and deep habitats. Assemblages differed between lakes. In both lakes, most species were caught in greater numbers in shallow samples, which had greater proportions of small fish of some species. Spatial variation in uni- and multivariate analyses was greatest at the smallest spatial scale (i.e. among replicates), which was consistent across species, depths and lakes. Variation among samples was also generally greater among sites than zones. It would be unnecessary to include the scale of zones in future sampling and more effort should be placed on sampling replicate nets and sites within lakes. The importance of doing pilot experiments and their general applicability to designing surveys of fisheries resources is illustrated.