From global surveys to local dynamics; quantifying ecological variability in tropical coral reef fish and benthic communities

Abstract

Coral reefs are among the most biodiverse ecosystems on Earth, sustaining essential ecological processes and providing food security, livelihoods, and cultural value to millions of people. Reef fish communities underpin these functions, contributing to herbivory, predation, nutrient cycling, and fisheries production. Yet, accelerating climate-driven disturbances are reshaping coral reef ecosystems worldwide, underscoring the need for robust ecological data to inform adaptive management. Despite decades of research, significant gaps remain in the spatial and temporal coverage of reef monitoring, and many management decisions are based on datasets that are geographically biased, methodologically inconsistent, or statistically limited. These limitations risk obscuring ecological processes and constraining evidence-based strategies for reef conservation. This thesis addresses these challenges by combining a global synthesis of reef fish survey research with an empirical case study of benthic and reef fish assemblages in American Samoa. The first component (Chapters 2 and 3) reviews 309 coral reef fish surveys published between 2010 and 2020, assessing geographic, habitat, and methodological biases in contemporary reef ecology. Survey effort is concentrated in the Indo-Pacific and Tropical Atlantic, with fore reef habitats disproportionately represented. Lagoonal, inshore, and mesophotic reefs remain under-sampled despite their importance for subsistence fisheries and long-term resilience. Belt transects dominate survey methodology, yet few studies explicitly report statistical sampling designs, raising concerns about extrapolating findings to management-relevant scales. Inconsistent metadata reporting further limits reproducibility and constrains opportunities for data integration across studies. Together, these findings highlight limitations in the global evidence base underpinning reef management and emphasise the need for more representative and statistically rigorous monitoring. The second component of the thesis (Chapters 4 and 5) draws on benthic and reef fish monitoring in American Samoa to evaluate the utility of multivariate dispersion (MvD) as a metric of ecological variability. Dispersion-based approaches quantify community heterogeneity, capturing subtle dynamics that may not be detected through static descriptors such as biomass or coral cover. Analysis of monitoring data from 2016 to 2020 revealed substantial benthic shifts in response to disturbance, with pre-disturbance dispersion metrics moderately predictive of subsequent community reorganisation. Similar associations were observed for reef fish assemblages, where initial benthic dispersion, measured at both functional group and coral genus levels, was linked to fish community responses over time. These findings demonstrate that dispersion metrics can provide early-warning indicators of ecological transitions, offering predictive insights beyond conventional monitoring indicators.

This thesis highlights a central challenge in coral reef ecology: global monitoring is often limited by geographic, methodological, and statistical constraints, yet innovative analytical approaches can enhance the interpretive power of local datasets and strengthen ecological inference. While multivariate dispersion does not replace taxonomically detailed monitoring, it serves as a powerful complementary tool, increasing sensitivity to disturbance-driven changes and informing adaptive management. Together, these analyses provide a comprehensive assessment of global reef fish survey practices, demonstrate the constraints of site-based monitoring, and evaluate dispersion-based metrics as tools for detecting and predicting community change. Together, these findings emphasise that strengthening coral reef science requires both structural improvements in global survey practice and methodological innovations at the site level. Expanding habitat coverage, adopting robust statistical designs, and standardising meta data reporting will improve the reliability and scalability of global reef fish datasets. Incorporating dispersion-based indicators can enhance sensitivity to ecological variability, providing managers with tools to anticipate transitions rather than merely reacting to sudden change and documenting decline. By bridging scales from global syntheses to local case studies, this thesis advances the evidence base for adaptive, ecosystem-based management in an era of accelerating environmental change.

Date of Award	10 Jun 2026
Original language	English
Awarding Institution	Bangor University
Supervisor	Gareth Williams (Supervisor) & Adel Heenan (Supervisor)