Search

The inaugural Computer Vision for Marine Scientists workshop was held at the 17th South African Marine Science Symposium, with the primary goal of establishing a community of practice for computer vision (CV) in marine sciences in South Africa. The one-day hybrid event, attended by 97 people, covered the principles of artificial intelligence (AI) techniques required for evaluating video and photographic imagery through presentations, practical demonstrations and interactive discussions. The recordings of the workshop sessions are available online, providing an opportunity to reach marine researchers both regionally and globally. The workshop highlighted that many scientists have begun to incorporate CV and AI into their research activities; however, there is little national coordination and the extent of research is lagging behind international trends. To support image-based AI research in South Africa, it is critical to maintain and expand the network established during the workshop. This would enable a more collaborative and successful approach to incorporating CV technology in the country's marine research initiatives, ultimately leading to ground-breaking discoveries and advancements in the field. [ABSTRACT FROM AUTHOR]

1. Designated using a Statutory Instrument in 2008, Lyme Bay marine-protected area (MPA) is the UK's first and largest example of an ambitious, whole-site approach to management, to recover and protect reef biodiversity. The whole-site approach applies consistent management, in this case excluding bottom towed fishing, across the full 206 km² extent of the MPA, thus protecting a mosaic of reef-associated habitats from regular damage, while still allowing less destructive fishing methods, such as static gear, rod and line, and diving. 2. To assess the effectiveness of this management strategy for mobile taxa and the sustainability for those taxa that continue to be targeted, Exploited and Non-Exploited species' populations were compared inside the MPA, relative to open control sites spanning 11 of the 12 years of designation. baited remote underwater video systems (BRUVs) were deployed annually to assess mobile benthic and demersal fauna. 3. Overall, the number of taxa significantly increased in the MPA relative to the open controls while total abundance increased in both treatments. 4. Exploited fish showed increases in number of taxa (430%) and total abundance (370%) inside the MPA over 11 years. 5. Likewise, but to a lesser degree in the open controls, number of taxa of commercially Exploited fish increased over time, potentially showing 'spillover' effects from the MPA. 6. Non-Exploited fish did not show such changes. Regardless of constituting the majority of the fishery value, highly valuable Exploited invertebrates showed no significant changes over time. 7. Synthesis and applications. The Lyme Bay marine-protected area shows importance of protecting a whole site, comprising mosaics of different benthic habitats, through protection of sessile organisms that contribute to essential fish habitats. This Ecosystem Approach to Fisheries Management can benefit and maintain sustainable fisheries and species of conservation importance. [ABSTRACT FROM AUTHOR]

• Acoustic Complexity Index higher in fished area vs protected area. • Acoustic Complexity Index does not correlate with simultaneous biodiversity indices. • Acoustic Complexity Index covaries with mobile assemblage composition. The soundscape of the marine environment is a relatively understudied area of ecology that has the potential to provide large amounts of information on biodiversity, reproductive behaviour, habitat selection, spawning and predator–prey interactions. Biodiversity is often visually assessed and used as a proxy for ecosystem health. Visual assessment using divers or remote video methods can be expensive, and limited to times of good weather and water visibility. Previous studies have concluded that acoustic measures, such as the Acoustic Complexity Index (ACI), correlate with visual biodiversity estimates and offer an alternative to assess ecosystem health. Here, the ACI measured over 5 years in a Marine Protected Area (MPA) in the UK, Lyme Bay, was analysed alongside another monitoring method, Baited Remote Underwater Video Systems (BRUVs). Two treatments were sampled annually in the summer from 2014 until 2018 with sites inside the MPA, as well as Open Control sites outside of the MPA. Year by year correlations, which have been used elsewhere to test ACI, showed significant correlations with Number of Species and ACI. However, the sign of these correlations changed almost yearly, showing that more in-depth analyses are needed. Multivariate analysis of the benthic assemblage composition (from BRUVs) was carried out by Permutational Multivariate Analysis of Variance (PERMANOVA) using Distance Matrices. Although not consistently correlating with univariate measures, the ACI was significantly interacting with the changing benthic assemblage composition, as it changed over time and protection (Inside vs Outside the MPA). ACI showed potential to allude to shifting benthic communities, yet with no consistency when used alongside univariate measures of diversity. Although it is not without its own disadvantages, and thus should be developed further before implementation, the ACI could potentially reflect more complex changes to the benthos than simply the overall diversity. [ABSTRACT FROM AUTHOR]