Estimates of the relative density of fishes
form the basis of many marine ecological studies as well as the assessment of
effects of fishing or pollution. Plasticity in the behavioural response of
large reef fishes to SCUBA divers means that commonly used underwater visual
census (UVC) techniques do not always provide reliable estimates of relative
density. The paper describes the system configuration, deployment methods,
testing and use of a remotely deployed baited underwater video (BUV) system
for the survey of carnivorous reef fishes (snapper,
Pagrus auratus and blue
cod,Parapercis colias) in marine reserves of northern
New Zealand. Concurrent UVC and BUV surveys inside and outside a marine
reserve showed that, whereas UVC detected few snapper in either area
(resulting in little confidence in statistically significant results), BUV
demonstrated significant differences in relative density. Conversely, blue cod
were found to occur at significantly higher densities within the reserve by
UVC, but not by BUV. The provision of accurate estimates of fish size (<20
mm error) from video footage also illustrated differences in size structure
between protected and fished populations. The data suggest that a combination
of survey techniques is likely to be necessary where multispecies assemblages
are being assessed.