Detecting conservation benefits of marine reserves on remote reefs of the northern GBR.

The Great Barrier Reef Marine Park (GBRMP) is the largest network of marine reserves in the world, yet little is known of the efficacy of no-fishing zones in the relatively lightly-exploited remote parts of the system (i.e., northern regions). Here, we find that the detection of reserve effects is challenging and that heterogeneity in benthic habitat composition, specifically branching coral cover, is one of the strongest driving forces of fish assemblages. As expected, the biomass of targeted fish species was generally greater (up to 5-fold) in no-take zones than in fished zones, but we found no differences between the two forms of no-take zone: 'no-take' versus 'no-entry'. Strong effects of zoning were detected in the remote Far-North inshore reefs and more central outer reefs, but surprisingly fishing effects were absent in the less remote southern locations. Moreover, the biomass of highly targeted species was nearly 2-fold greater in fished areas of the Far-North than in any reserve (no-take or no-entry) further south. Despite high spatial variability in fish biomass, our results suggest that fishing pressure is greater in southern areas and that poaching within reserves may be common. Our results also suggest that fishers 'fish the line' as stock sizes in exploited areas decreased near larger no-take zones. Interestingly, an analysis of zoning effects on small, non-targeted fishes appeared to suggest a top-down effect from mesopredators, but was instead explained by variability in benthic composition. Thus, we demonstrate the importance of including appropriate covariates when testing for evidence of trophic cascades and reserve successes or failures.