Habitat configurations shape the trophic and energetic dynamics of reef fishes in a tropical-temperate transition zone: implications under a warming future

Abstract

Understanding the extent to which species' traits mediate patterns of community assembly is key to predict the effect of natural and anthropogenic disturbances on ecosystem functioning. Here, we apply a trait-based community assembly framework to understand how four different habitat configurations (kelp forests, Sargassum spp. beds, hard corals, and turfs) shape the trophic and energetic dynamics of reef fish assemblages in a tropical-temperate transition zone. Specifically, we tested (i) the degree of trait divergence and convergence in each habitat, (ii) which traits explained variation in species' abundances, and (iii) differences in standing biomass (kg ha^-1), secondary productivity (kg ha^-1 day^-1) and turnover (% day^-1). Fish assemblages in coral and kelp habitats displayed greater evidence of trait convergence, while turf and Sargassum spp. habitats displayed a higher degree of trait divergence, a pattern that was mostly driven by traits related to resource use and thermal affinity. This filtering effect had an imprint on the trophic and energetic dynamics of reef fishes, with turf habitats supporting higher fish biomass and productivity. However, these gains were strongly dependent on trophic guild, with herbivores/detritivores disproportionately contributing to among-habitat differences. Despite these perceived overall gains, turnover was decoupled for fishes that act as conduit of energy to higher trophic levels (i.e. microinvertivores), with coral habitats displaying higher rates of fish biomass replenishment than turf despite their lower productivity. This has important implications for biodiversity conservation and fisheries management, questioning the long-term sustainability of ecological processes and fisheries yields in increasingly altered marine habitats.

Keywords: Community assembly; Ecosystem functions; Ecosystem services; Trait-based ecology; Tropicalisation.

Fig. 1

a Map of the study region (Houtman Abrolhos Islands, Wallaby Group, 28° latitude, ~ 60 km offshore) in the midwest bioregion of Western Australia. Black dots depict the spatial location of study sites. The green polygon delineates the no-take zone. b–d Yearly-averaged trends in b annual, c summer (December, January, February), and d winter (June, July, August) sea surface temperatures (SST, °C) from 1982 to 2020, across study sites. e–j. Photos of transect-level habitat clusters used in the analyses: e corals, f kelp, g Sargassum spp., and h turf

Fig. 2

Relative contribution of over-dispersion (i.e., divergence, red), under-dispersion (i.e., convergence, blue), and randomness (grey) to patterns of reef fish community assembly across habitats. The pattern of standardized effect size functional diversity (SES FD) was investigated under increasing sensitivity to functional entities’ relative abundance: a “q” = 0 (species compositions only), b “q” = 1 (higher weight on common entities’), and c “q” = 2 (higher weight on dominant entities’). Dark red (over-dispersed) and blue (under-dispersed) colours depict assemblages that were significantly different from the null model (outside the 95th percentile of the null distribution). Sample sizes (N) for each habitat category are included above barplots

Fig. 3

Trait–environmental relationships. a Fourth-corner analysis testing the interaction between functional traits and the percent cover (%) of habitat groups to predict the abundance of fish species. Colour intensity denotes the magnitude of the standardised coefficients (β) from GLMs with a LASSO penalty for model selection. Red = positive correlations, blue = negative correlations. b, c Shared co-structure between the percent cover of habitat groups (b), fish traits (c), and fish abundance (d) tables in RLQ analyses. In (d), only species with higher scores in the RLQ axes are magnified for visualization purposes. RLQ scores for trait and environmental variables are shown in Fig. S3, whilst species codes are provided in Table S6

Fig. 4

Predicted (mean ± SE) a standing biomass, b productivity, and c turnover of reef fish trophic guilds, from GLMs, across habitat groups: coral (red), kelp (blue), mixed (cyan), Sargassum spp. (orange), and turf (brown). HerDet herbivores/detritivores, HigCar higher carnivores, SesInv sessile invertivores, MiInv microinvertivores, Plktiv planktivores. For visualization purposes standing biomass is shown on a log scale. Sample sizes for each habitat category are included in the legend