METHODS

New Caledonia is surrounded by a barrier reef, which borders one of the largest lagoons in the world (24 000 km2). This lagoon supports diverse populations of fish associated with a variety of habitats, and subject to a wide variety of possible impacts, including fishing and terrestrial run-off. In May 2006, an in situ experiment was conducted on the southwest lagoon, in the highly-protected reserve at Ilot Signal (22º17. 73’S, 166º17. 41’E) and fished area of Récif Larégnère (22º19. 71’S, 166º17. 68’E). Surveys were conducted at three sites inside the reserve (A, B and C) and site D outside (Fig. 1). These sites were chosen to have comparable habitat of fringing coral reef with adjacent soft-sediment areas. At each site, replicate samples (n= 4) were collected using the three techniques to be compared. These were UVS, D-BRUV and HBRUV.

For UVS, sampling was carried out along a 50 x 10 m belt transect (after Samoilys 1997). Commercial fish populations were recorded at species level, including abundance and size of each individual observed. For D-BRUV, we used a system that employs one camera pointed downward towards a bait pot centred on the base of a tripod (Willis et al. 2000, Fig. 2a). The base of the tripod forms a 1.6 m2 quadrat and calibration marks can be used to measure fish seen within the quadrat. This system has been used successfully to monitor populations of commercially important fish inside and outside highly-protected marine reserves in warm and cool temperate areas of New Zealand (Willis and Millar 2005). For H-BRUV, Harvey and Shortis (1996) developed a stereovideo technique using two horizontally mounted video cameras (Fig. 2b) that uses a three dimensional calibration to estimate the size of fish. This system has been used successfully to study reef fish populations in temperate and tropical Western Australia (Watson et al. 2005). In this study,