Abstract

Volunteer based monitoring programmes using underwater visual census (UVC)
have been promoted and used as a cost-effective way to gather data on fish. It has been
demonstrated that UVC methods can be subject to biases and errors. These errors and
biases include estimation of the survey area, fish identification, density, and length. New
developments in technology are resulting in increased use of diver-operated stereo video
(stereo-DOV) surveys as a potential alternative to UVC. As stereo-DOV is being
increasingly used for reef monitoring, it is important to understand how this method
compares to standard reef monitoring methods such as UVC. The aim of the current
study was to conduct UVC and stereo-DOV surveys at the same sites, and compare
estimates of fish density, species richness, and fish length. UVC and stereo-DOV surveys
were conducted in Honduras on coral reefs within the Cayos Cochinos Marine Protected
Area (MPA), which allows only artisanal fishing, and Utila, which is open to all types of
fishing. Generalized linear models (GLM) were used to test for an effect of survey
method and protected area status on reef fish community structure. Compared to stereo-
DOV, UVC produced significantly higher estimates of total fish density, species richness,
mean fish length, and predatory and herbivorous fish densities. UVC estimated higher
mean fish lengths in the MPA compared to the non-MPA (significant interaction between
survey method and protection, p<0.001). This may be a result of bias caused by divers
expecting larger fish in the MPA. The results of the current study indicate UVC and
stereo-DOV methods differ significantly in their estimates of tropical reef fish
community variables, and thus should not be used interchangeably. These results have
implications for reef monitoring programmes and marine resource management.