Comparing visual counts of coral reef fish: Implications of transect width and species selection

Abstract

We compared fish count data derived simultaneously from visual censuses of transects of different width for a wide range of species on 3 coral reefs within the Great Barrier Reef, Australia. The relationships of count data from non-pomacentrid families between transacts 10 and 5 m wide and pomacentrid genera between transacts 2 and 1 m wide were found to be strongly linear, with correlation coefficients of 0.94 and 0.98 respectively. The strength of both relationships was not significantly compromised by differences among selected taxa or benthic habitats. As such, comparison of data collected from different transect widths was considered feasible after application of conversion factors derived from the linear relationships. We suggest that the use of experimentally derived conversion factors may be applicable for other researchers faced with problems of data comparison across studies where different transect dimensions have been used. Density estimates from wider transacts were 22 to 26 % less than estimates from narrower transacts for the majority of taxa. The precision (coefficient of variation) of density estimates did not vary significantly between transact dimensions, however precision varied greatly between species, genera and families, with values ranging from 0.18 to 0.85. Transact width and the behaviour, habitat specificity and number of target fish taxon are implicated as important contributors to variation in the accuracy and precision of visual census data. We re-stress the need for methodological pilot studies as precursors to visual census studies of reef fish. Australian Institute of Marine Science.