Dataset: Geange_2010 Larger Heterospecifics Habitat Availability

Three transects within Moorea’s northern lagoon were established. Each transect was approximately 800 m in length, and ran from shore to the barrier reef crest. Along each transect, five approximately equally spaced sites were identified (90 (SE = 3), 180 (SE = 7), 289 (SE = 7), 409 (SE = 7) and 508 (SE = 11) m from the reef crest). Within each site, 16 patch reefs (hereafter reefs) of similar size (n = 240 reefs: average surface area = 7.98 m2, SE = 0.24; average height = 0.83 m, SE = 0.01) were haphazardly selected. For each reef, and a one-meter halo surrounding the reef, percent cover of 16 substrate categories was visually estimated: Porites massive; Porites rus; Montipora spp.; Pocillopora spp.; Acropora spp.; other live coral; Turbinaria spp.; Dictyota spp.; Halimeda spp.; turf (a mixed filamentous red algal assemblage consisting predominately of Polysiphonia spp.); sponge; bare; other; sand; coral rubble; and pavement). For a subset of 45 reefs, the accuracy of visual estimates of percent cover against Fixed Point Contact (FPC) estimates of percent cover (a commonly accepted standard; Floyd & Anderson 1987) was evaluated. Visual estimates of percent cover were 84.8% accurate relative to FPC estimates and include better estimates of rare habitat than FPC (Meese & Tomich 1992). Habitat surveys were conducted between 7 March and 29 April 2005. On each reef and the surrounding one-meter halo, the identity of all juvenile labroids was recorded and their Standard Length (SL; to the nearest mm) visually estimated. The substrate directly below each individual at the time it was first observed was also recorded. Juveniles were defined as individuals less than 25mm SL. The accuracy of visual estimates of SL made in the field were evaluated by capturing a subset of fish (n = 298 from 8 different species) and comparing visual estimates of SL for these individuals to SL measured in the lab with callipers. Based on mean absolute error (MAE), visual estimation of SL was 94.7% accurate. Fish surveys were conducted between 0800 and 1600 hours (peak activity time for diurnal benthic fishes: Galzin 1987) from 27 May to 6 June 2005.