Declines in habitat structural complexity have marked ecological outcomes, as currently observed in many of the world’s ecosystems. Coral reefs have provided a model for such changes in marine ecosystems, but our understanding has been centred on corals and fishes at broad spatial scales when metazoan diversity on coral reefs is dominated by small cryptic taxa (herein: ‘cryptofauna’). Given the paucity of studies and high taxonomic complexity of the cryptofauna, both of which limit a priori hypotheses, we asked whether hierarchical structuring theory provides a compelling framework to impose order and quantify pattern. In general terms, we explored whether cryptic communities are sufficiently described by broad seascape parameters or limited by a set of processes operating at their distinctly nested microhabitat scale. To address this theory and gaps in knowledge for the cryptofauna, we characterised community structure in coral rubble, an eroded coral condition where biodiversity proliferates. Rubble was sampled along a depth and exposure gradient at Heron Island on the Great Barrier Reef, Australia, to parameterise environmental and morphological indicators of sessile taxa and motile cryptofauna communities. We employed a hierarchical study framework from microhabitat to seascape scales, which were evaluated using non-structured multivariate analyses and Bayesian structural equation modelling. While the non-structured analyses showed the effects of seascape on the cryptobenthos and its community, this approach overlooked the finer hierarchical patterns in rubble ecology revealed only in the structured model. Seascape parameters (exposure and depth) influenced microhabitat complexity (i.e., rubble branchiness), which determined the cover of sessile organisms on rubble pieces, which shaped the motile cryptofauna community. Rubble is likely to be increasingly prevalent on coral reefs in the Anthropocene and is typically associated with low seascape-level complexity and reduced macrofaunal richness. Parallel with hierarchical structuring theory, we show a similar response operating at the microhabitat scale whereby low rubble complexity (i.e., branchiness) reduces cryptobenthic structure, diversity and size spectra. We expect there may be an initial increase in biodiversity and trophodynamic processes derived from branching rubble, but a delay in ecosystem-scale outcomes if coral, and thus rubble, generation and complexity cannot be sustained in a future ocean.

Overarching aim and design: This study was designed to investigate whether the theory of hierarchical structuring of ecological systems could explain resource aggregation in coral rubble where biodiversity proliferates. Data were analysed across multiple scales of ecological organisation; first at the level of seascape drivers and then at the level of microhabitat.

At the seascape scale, water depth and hydrodynamic exposure represented higher-order environmental limits on rubble patches. Whether these seascape parameters sufficiently explained lower-level entities within rubble was assessed independently using permutational multivariate analysis of variance (PERMANOVA) on (1) rubble piece morphology, (2) the proportional cover of sessile taxa on rubble pieces, and motile cryptofauna (3) density, (4) diversity and (5) size spectra. Separate Excel spreadsheets are attached for each of these five datasets.

To determine the applicability of hierarchical theory to rubble biomes more acutely, the relative influence of seascape drivers (wave exposure and depth) and lower-level microhabitat variables (rubble morphology, sessile taxa, and motile cryptofauna density, diversity and size spectra) on rubble community assembly were interrogated in a Bayesian SEM framework. Explicit consideration of this hierarchical study structure from seascape to microhabitat helped to disentangle the network of relationships that shape the complex rubble cryptobiome where tropical marine biodiversity proliferates and provides a model example of the utility of SEM approaches in ecological research.

Data collection:

Rubble patches were surveyed at Heron Island, situated in the Capricorn Bunker Group in the southern GBR, Australia (23˚26’31” S, 151˚54’50” E), which predominantly experiences moderate south-east winds. Three sites were established along the south-west margin of Heron Reef, where prevailing winds result in a moderate exposure gradient. Six samples of rubble (~500 mL) were taken from two shallow (3–5 m) and two deep (7–11 m) rubble patches at each site, resulting in a total of 72 samples of coral rubble to characterise morphological and biological features across scale.

(1) Rubble piece morphology: Morphometric parameters predicted to be informative in defining rubble were measured for every rubble piece (n = 6–35 per sample). Parameters of individual rubble pieces included (1) widest span (longest length in any direction), (2) mean thickness of primary branch (n = 3 measurements per piece), (3) mean secondary branch length (n = 0–3 measurements per piece; some pieces did not exhibit branching morphology), (4) number of secondary branches > 1 cm (herein: ‘branchiness’), (5) wet and (6) dry weight, (7) wet volume (water displacement volume, mL), and (8) porosity.

(2) Proportional cover of sessile taxa on rubble: Six rubble pieces were taken at random from each sample and photographed to quantify the relative cover of sessile taxa on rubble pieces. Each piece was photographed from two distinct perspectives, as community succession in rubble differs between surface (top) and subsurface (bottom) microhabitats. Proportional cover of major functional groups, including bare surface, encrusting calcifying algae (including crustose coralline algae (CCA) and Peyssonnelia), turf algae, macroalgae (including Dictyota, Lobophora and Halimeda), sponges and ascidians, and bryozoans (erect and encrusting), on each rubble piece was quantified using a gird overlay on each photo. The number of cells occupied by each substrate type was counted. Data attained from both surface and subsurface perspectives were combined to establish a whole-of-piece cover estimate.

Motile cryptofauna (3) density, (4) diversity and (5) size spectra: Contents of each sample were poured into separate trays and conspicuous motile fauna were collected using forceps and spoons. Remaining cryptic fauna were dislodged from rubble pieces using a series of freshwater rinses (~ 1 min) and pressurised water over a 210 µm sieve. All motile organisms were grouped to represent one sample of cryptofauna per rubble sample, preserved in 70% ethanol for later identification. The size (mm), abundance and diversity of motile cryptofauna in each sample were determined. Cryptofauna were identified under a dissecting microscope primarily to the level of family and measured to the nearest 250 µm using a scored dish. Standard measurements for size were used; carapace width for crab-like crustaceans or length for shrimp-like crustaceans, shell length for molluscs (longest distance), diameter for echinoderms with radial symmetry, and length for all types of worms. Density was determined from abundance and the water displacement volume (mL) of each rubble sample. Size spectra were determined for each sample as the linear relationship between log-density and log-size of all individuals, resulting in intercept and slope coefficients for each rubble sample.

Microsoft Excel is required to open the data files. No novel code was used.