Leveraging deep learning and computer vision technologies to enhance management of coastal fisheries in the Pacific region.

This paper presents the design and development of a coastal fisheries monitoring system that harnesses artificial intelligence technologies. Application of the system across the Pacific region promises to revolutionize coastal fisheries management. The program is built on a centralized, cloud-based monitoring system to automate data extraction and analysis processes. The system leverages YoloV4, OpenCV, and ResNet101 to extract information from images of fish and invertebrates collected as part of in-country monitoring programs overseen by national fisheries authorities. As of December 2023, the system has facilitated automated identification of over six hundred nearshore finfish species, and automated length and weight measurements of more than 80,000 specimens across the Pacific. The system integrates other key fisheries monitoring data such as catch rates, fishing locations and habitats, volumes, pricing, and market characteristics. The collection of these metrics supports much needed rapid fishery assessments. The system's co-development with national fisheries authorities and the geographic extent of its application enables capacity development and broader local inclusion of fishing communities in fisheries management. In doing so, the system empowers fishers to work with fisheries authorities to enable data-informed decision-making for more effective adaptive fisheries management. The system overcomes historically entrenched technical and financial barriers in fisheries management in many Pacific island communities.

Functionally diverse reef-fish communities ameliorate coral disease.

Coral reefs, the most diverse of marine ecosystems, currently experience unprecedented levels of degradation. Diseases are now recognized as a major cause of mortality in reef-forming corals and are complicit in phase shifts of reef ecosystems to algal-dominated states worldwide. Even so, factors contributing to disease occurrence, spread, and impact remain poorly understood. Ecosystem resilience has been linked to the conservation of functional diversity, whereas overfishing reduces functional diversity through cascading, top-down effects. Hence, we tested the hypothesis that reefs with trophically diverse reef fish communities have less coral disease than overfished reefs. We surveyed reefs across the central Philippines, including well-managed marine protected areas (MPAs), and found that disease prevalence was significantly negatively correlated with fish taxonomic diversity. Further, MPAs had significantly higher fish diversity and less disease than unprotected areas. We subsequently investigated potential links between coral disease and the trophic components of fish diversity, finding that only the density of coral-feeding chaetodontid butterflyfishes, seldom targeted by fishers, was positively associated with disease prevalence. These previously uncharacterized results are supported by a second large-scale dataset from the Great Barrier Reef. We hypothesize that members of the charismatic reef-fish family Chaetodontidae are major vectors of coral disease by virtue of their trophic specialization on hard corals and their ecological release in overfished areas, particularly outside MPAs.

Geography and island geomorphology shape fish assemblage structure on isolated coral reef systems.

We quantify the relative importance of multi-scale drivers of reef fish assemblage structure on isolated coral reefs at the intersection of the Indian and Indo-Pacific biogeographical provinces. Large (>30 cm), functionally-important and commonly targeted species of fish, were surveyed on the outer reef crest/front at 38 coral reef sites spread across three oceanic coral reef systems (i.e. Christmas Island, Cocos (Keeling) Islands and the Rowley Shoals), in the tropical Indian Ocean (c. 1.126 x 106 km2). The effects of coral cover, exposure, fishing pressure, lagoon size and geographical context, on observed patterns of fish assemblage structure were modelled using Multivariate Regression Trees. Reef fish assemblages were clearly separated in space with geographical location explaining ~53 % of the observed variation. Lagoon size, within each isolated reef system was an equally effective proxy for explaining fish assemblage structure. Among local-scale variables, 'distance from port', a proxy for the influence of fishing, explained 5.2% of total variation and separated the four most isolated reefs from Cocos (Keeling) Island, from reefs with closer boating access. Other factors were not significant. Major divisions in assemblage structure were driven by sister taxa that displayed little geographical overlap between reef systems and low abundances of several species on Christmas Island corresponding to small lagoon habitats. Exclusion of geographical context from the analysis resulted in local processes explaining 47.3% of the variation, highlighting the importance of controlling for spatial correlation to understand the drivers of fish assemblage structure. Our results suggest reef fish assemblage structure on remote coral reef systems in the tropical eastern Indian Ocean reflects a biogeographical legacy of isolation between Indian and Pacific fish faunas and geomorphological variation within the region, more than local fishing pressure or reef condition. Our findings re-emphasise the importance that historical processes play in structuring contemporary biotic communities.

Evidence of stable genetic structure across a remote island archipelago through self-recruitment in a widely dispersed coral reef fish.

We used microsatellite markers to assess the population genetic structure of the scribbled rabbitfish Siganus spinus in the western Pacific. This species is a culturally important food fish in the Mariana Archipelago and subject to high fishing pressure. Our primary hypothesis was to test whether the individuals resident in the southern Mariana Island chain were genetically distinct and hence should be managed as discrete stocks. In addition to spatial sampling of adults, newly-settled individuals were sampled on Guam over four recruitment events to assess the temporal stability of the observed spatial patterns, and evidence of self-recruitment. We found significant genetic structure in S. spinus across the western Pacific, with Bayesian analyses revealing three genetically distinct clusters: the southern Mariana Islands, east Micronesia, and the west Pacific; with the southern Mariana Islands being more strongly differentiated from the rest of the region. Analyses of temporal samples from Guam indicated the southern Mariana cluster was stable over time, with no genetic differentiation between adults versus recruits, or between samples collected across four separate recruitment events spanning 11 months. Subsequent assignment tests indicated seven recruits had self-recruited from within the Southern Mariana Islands population. Our results confirm the relative isolation of the southern Mariana Islands population and highlight how local processes can act to isolate populations that, by virtue of their broad-scale distribution, have been subject to traditionally high gene flows. Our results add to a growing consensus that self-recruitment is a highly significant influence on the population dynamics of tropical reef fish.