MEPS 198:249-260 (2000)  -  doi:10.3354/meps198249

Detection of spatial variability in relative density of fishes: comparison of visual census, angling, and baited underwater video

Trevor J. Willis1,*, Russell B. Millar2, Russell C. Babcock1

1Leigh Marine Laboratory University of Auckland, PO Box 349, Warkworth, New Zealand
2Department of Statistics, University of Auckland, Private Bag 92019, Auckland, New Zealand

ABSTRACT: The ability to make accurate estimates of fish relative abundance is the basis of both ecological and environmental effects studies, and flawed sampling methods may give misleading results even in otherwise well-designed surveys. This paper compares surveys of snapper Pagrus auratus (Sparidae) and blue cod Parapercis colias (Pinguipedidae) conducted using 3 methods (underwater visual census, experimental angling, and baited underwater video) inside and outside the Cape Rodney-Okakari Point marine reserve in northeastern New Zealand. Angling and baited video consistently detected adult P. auratus at protected and fished sites, providing estimates of 36.7 and 39.2 times greater density of fishable P. auratus within the reserve, respectively. Visual surveys provided the least reliable measure of density of P. auratus, with adults only detected at the reserve centre where fish have been habituated to divers by hand-feeding. Measures of the size structure of P. auratus were consistent between angling and video, but mean size was significantly smaller using visual census methods. Relative density of P. colias was similar for all 3 methods, but angling estimated larger mean size, probably due to hook selectivity against smaller fish. The study indicates that methodological standardisation across all species is not always appropriate for environmental effects studies, and that different survey methods should be considered according to the biology and behaviour of the species of interest.

KEY WORDS: Angling · Baited underwater video · Exploited fishes · Behaviour · Log-linear model · Marine reserves · Recovery · Survey bias · Visual census

Full text in pdf format