MEPS 254:269-280 (2003)  -  doi:10.3354/meps254269

Movements of reef fishes across marine reserve boundaries: effects of manipulating a density gradient

D. Zeller1,2,*, S. L. Stoute1, G. R. Russ1

1School of Marine Biology and Aquaculture, James Cook University, Townsville, Queensland 4811, Australia
2Present address: Fisheries Centre, 2204 Main Mall, University of British Columbia, Vancouver, British Columbia V6T 1Z4, Canada

ABSTRACT: Marine reserves should display net export of biomass if they are to be of use in fisheries management. We assessed experimentally the export of fish biomass (Œspillover¹) from 2 marine reserves on the Great Barrier Reef. Using the mark-recapture technique with traps, and hook-and-line fishing, movements of serranids, lutjanids and lethrinids were monitored at 2 experimental and 2 control sites. Experimental sites were bisected by reserve boundaries into Œzones¹ open and closed to fishing. Control sites were protected from fishing, with an arbitrary boundary creating 2 equal zones. Between 1997 and 1999, 1374 individuals were externally tagged, and during 1999 densities were manipulated twice. Underwater visual census (UVC) surveys revealed pre-manipulation density gradients between zones that were weak at real boundaries, and stronger at the arbitrary control boundaries. Movements were highly localized, with over 60% of recaptures being made within the 50 x30 m block of initial capture. The mean distances moved differed between the 3 species with the highest recapture rates: Plectropomus leopardus (Serranidae) moved distances >100 m, Lutjanus carponotatus (Lutjanidae) occasionally moved over 50 m, while Cephalopholis cyanostigma (Serranidae) was predominantly caught within the block of initial capture. Movements across real or control boundaries were rare. Twice in 1999, fish densities in zones open to fishing at the experimental sites were reduced by spearfishing (by 61 to 64% and 57 to 83%, respectively). UVC revealed density gradients at the experimental boundaries and no changes at non-manipulated control sites. There was no evidence of experimentally induced directional movements, using traps or hook-and-line sampling; however, 3 species (C. cyanostigma, L. carponotatus, L. quinquelineatus) did show a propensity for increased movements. Detection of spillover requires powerful sampling designs incorporating strong density gradients across large numbers of reserve boundaries.

KEY WORDS: Marine reserves · Spillover · Fisheries management · Coral reefs · Experiment · Movements · Tagging

Full text in pdf format