Size structured reef ecosystems : linking size spectrum theory to global field observations of reef fauna from copepods to sharks

Across aquatic systems, the body size of an organism is often more important than its species identity in determining how it interacts with its predators, competitors, and habitat. The relationship between body size and abundance is often described by a linear function on the log-log scale (the size spectrum). The size spectrum provides a very simple way to represent an ecological community comprised of potentially many species, life stages and individuals with just two parameters, a slope, and an intercept. The size spectrum slope, for example, can inform us about how energy moves through the system (e.g., feeding behaviours), and deviations from expected slope values can inform us about disturbances (e.g., fishing impacts). To determine 'deviations', we must first characterise a baseline size spectrum, vital to its use as an ecological indicator of reef health. The study of community size spectra requires individual-level body size and abundance data. Empirical studies of size spectra therefore rarely span multiple taxonomic groups or broad spatial scales, prohibiting more general conclusions about the ecosystem. The Reef Life Survey (RLS) is a global-scale citizen-science program surveying the marine life of both tropical and temperate reefs. The data from this program provides an incomparable resource to test various size spectra theories in reef ecosystems and forms a core basis for the research presented in this thesis. The primary goal of the thesis was to extend our knowledge on the size structuring of reef communities across taxonomic groups and scales, by combining these global empirical data with novel analytical approaches. Three discrete aims were: 1) Describe the empirical size spectrum of reefs globally, including both fishes and large mobile invertebrates. 2) Investigate the cause of the commonly observed 'dip' in abundance of the small fishes in reef size spectra, which is often assumed to be a result of under-sampling of small reef fishes. 3) Determine how the size structure of reef communities relates to abundance and the number of species, three elements of biodiversity rarely considered together. The first analytical chapter of this thesis addresses the issue of reef size spectra studies focusing on a single taxonomic group and therefore potentially missing large components of the energy pathway. The study estimated invertebrate body size data based on asymptotic length and combines it with fish body size data to provide the first global-scale estimates of reef size spectra that extend beyond just the fishes. The study highlights the importance of including invertebrates in reef size spectra, develops a method for estimating invertebrate body size in other data-poor situations, and provides a baseline size spectrum for reefs. The second analytical chapter addresses the issue of reef size spectra studies ignoring small fishes. Observed reef size spectra including small to medium sized fishes tend to be unimodal on the log-log scale. It has been common practice to remove small-bodied individuals less than the modal body size and fit a linear model to the descending limb‚ÄövÑvp. This practice has been justified by the potential of under-sampling these small individuals. The study tests this theory of under-sampling by extending the body size range of the size spectrum to incorporate the smallest sized reef fauna ‚Äö- epifaunal individuals down to 0.125 mm in body size. This study provides evidence for this abundance 'dip' being a true feature of an underlying nonlinear size spectrum. This study extended the linearity of reef size spectra investigated on reefs to span the entire range of size classes of consumers, for the first time, to my knowledge. Outcomes have important implications for the justification of the removal of the smaller-bodied individuals, and therefore for the estimation of the size spectrum slope. The final analytical chapter uses size spectra to address a broad question regarding the complex inter-dependencies between body size, abundance, and species richness. The study applies, and further develops, a method originally developed for estimating community size spectra from protist species size distributions to investigate three important macroecological relationships; 1) the abundance size spectrum, 2) the species-richness size spectrum (how species richness varies with body size), and 3) the combination of these; the proposed linear (on the log-log scale) relationship between species richness and abundance within size classes. The study also provides a methodology to accurately reconstruct these three relationships in situations with minimal body size data (e.g., from only the species abundance and an estimate of asymptotic species size), providing a pipeline by which future studies can investigate these relationships using datasets that lack equivalent size detail to the RLS data used here. The combined outcomes of these three analytical chapters include enhanced understanding of energy flow and size structure of reef ecosystems. In particular, they confirm alignment with theoretical expectations when much of the full ecosystem size spectrum is covered, rather than removing portions of it to confirm theory or extrapolating from a single taxonomic group. The study also provides a means to better progress use of size spectra as ecological indicators of reef health. By describing both the linear and nonlinear aspects of the empirical size spectrum of reefs, we are now better positioned to identify the environmental, ecological, and anthropogenic drivers of variation and deviation in the size spectrum. This PhD project, which benefitted from a highly detailed dataset to generate and test methods, now provides a set of approaches that can be applied by ecologists in other fields, under less data-rich situations.