A field comparison of baited remote underwater video stations (BRUVS) and prawn (shrimp) trawls was made to assess the sampling options for describing patterns of fish biodiversity in the lagoonal waters of the Great Barrier Reef Marine Park. Replicated comparisons were made during the day and night on trawl grounds with different biophysical characteristics. Each trawl was made close and parallel to a long-shore set of five BRUVS set at regular intervals along one nautical mile. For each species, the sum of the maximum number of fish sighted on BRUVS at any one time (SigmaMaxN(1, ..., 5)) was compared with the number of fish caught in trawls (N). The two techniques recorded significantly different components of the fish fauna on the trawl grounds. Trawls caught mainly small (less than or equal to 300 mm), sedentary or cryptic, demersal species-such as flatfishes, apogonids, synodontids, triglids and callionymids. The BRUVS recorded more larger, mobile species from a much wider size range of families, including large elasmobranchs, more fusiform pelagic species (such as carangids and scombrids), and numerous eels. Species accumulation curves were parallel and very similar in shape for both techniques, but were separated by about 11 extra species in trawls. Fifty-two species in 17 families were caught only in trawls and 38 species in 15 families were recorded only by BRUVS. The occurrence of 38 small mobile species in 21 families was common to both techniques, but most of these showed marked differences in relative abundance. Trawls recorded higher species richness at all sites, and at night, but the average number of species and individuals recorded per transect were only about 26% and 19% lower for the BRUVS. The BRUVS consistently recorded more species during the day than trawls, and are limited most by water clarity. Multivariate analyses showed that both techniques indicated the presence of six fish assemblages based on day and night in three locations, despite sampling quite different components of the fauna. The six assemblages accounted for 78.9% of the variation in the BRUVS data compared to 68.9% of the variation in the trawl data, and there was a lower error rate in discriminating the six assemblages in the BRUVS data. BRUVS are non-extractive and can be used on complex topographies and in sensitive habitats in most or all zones of the Marine Park, but models of bait plumes are needed to convert fish sightings to density estimates. They may have a particular role in studies of larger, rarer elasmobranchs and teleosts of special conservation interest. Trawls can be used in any level of water clarity and provide direct estimates of fish density, particularly those smaller teleosts most vulnerable to the effects of trawling. The complementary use of prawn trawls and BRUVS would enable a more comprehensive assessment of teleost and elasmobranch diversity in the variety of inter-reef seabed topographies and management zones comprising the Marine Park. Crown Copyright (C) 2003 Published by Elsevier B.V. All rights reserved.
