Remote Underwater Video Research Articles

First record of the critically endangered great hammerhead shark (Sphyrna mokarran) in its natural habitat in the coast of Paraíba, Northeastern Brazil

The Great Hammerhead Shark, Sphyrna mokarran, is critically endangered around the world. The overwhelming majority of their records are of animals killed, quartered or captured by fishermen, which limit our understanding on its biology and conservation. For the first time, we report on the S. mokarran occurrence in the State of Paraíba, Northeastern Brazil, in its natural habitat. On 21 March 2021 a 2-m sub-adult male was captured by two Baited Remote Underwater Video Surveys (BRUVS) at 55-59 m depth, 30 km off João Pessoa, within the Queimado Shipwreck Marine Protected Area (MPA). Our findings suggest that S. mokarran is rare in the region and might be threatened by unmonitored and unregulated fishing activities. There is an urgent need for effective management and conservation actions for the species in the State of Paraíba. This record has the potential to serve as baseline data for future implementation of non-take zones within the Queimado Shipwreck MPA.

Small marine reserves provide conservation benefits for coastal sharks in southern New Zealand

Abstract With the increasingly imperilled status of shark populations, there is a pressing need to evaluate management solutions. Given the threats posed by fishing, marine reserves (MRs) present a promising option. Ata Whenua (Fiordland) in the South Island of Aotearoa/New Zealand is an ideal location to investigate this phenomenon owing to the presence of several shark species in coastal MRs. One‐hundred and sixty‐seven baited remote underwater video deployments were made in five MRs. A multi‐model inference approach using generalized linear modelling was used to assess the combined effect of the MRs on two trophic groups of coastal sharks. Generalized linear modelling was used to assess the effect of protection on, firstly, the presence of broadnose sevengill sharks (Notorynchus cepedianus), while accounting for variations in environmental variables, and secondly, the combined relative abundance of mesopredatory sharks detected (spiny dogfish, Squalus acanthias; school shark, Galeorhinus galeus; and carpet shark, Cephaloscyllium isabellum). Mesopredators had a higher relative abundance in MRs by a factor of 2.5 and there was a strong significant effect of protection, suggesting that MRs have led to an increase in their abundance or a change in distribution. In contrast, there was no effect of protection detected for sevengill sharks. It is concluded that the relatively small (<40 km2) MRs sampled do not provide conservation benefits for a large, mobile shark, but that they are potentially large enough to offer protection for mesopredatory sharks with smaller home ranges. To be effective for shark conservation, we therefore recommend that MRs need to be appropriately sized for the ranging behaviour of the target species for protection.

Further Insights into Invasion: Field Observations of Behavioural Interactions between an Invasive and Critically Endangered Freshwater Crayfish Using Baited Remote Underwater Video (BRUV).

Competitive behavioural interactions between invasive and native freshwater crayfish are recognised as a key underlying mechanism behind the displacement of natives by invaders. However, in situ investigations into behavioural interactions between invasive and native crayfish are scarce. In Australian freshwater systems, the invasive Cherax destructor has spread into the ranges of many native Euastacus species, including the critically endangered Euastacus dharawalus. Staged contests between the two species in a laboratory setting found E. dharawalus to be the dominant competitor, however, this has yet to be corroborated in situ. Here, we used baited remote underwater video (BRUV) to examine in situ intra- and inter-specific behavioural interactions between E. dharawalus and C. destructor. We sought to evaluate patterns of dominance and differential contest dynamics between the species to provide indications of competition between the two species. We found E. dharawalus to be dominant over C. destructor based on pooled interspecific interaction data and size-grouped interactions where C. destructor was the smaller opponent. Alarmingly, however, when C. destructor was within a 10% size difference the dominance of E. dharawalus was lost, contrasting with the outcomes of the laboratory-staged study. In addition, we report that small C. destructor initiated significantly more contests than larger conspecifics and larger E. dharawalus, a pattern that was not observed in smaller E. dharawalus. Further, intraspecific interactions between C. destructor were significantly longer in duration than intraspecific interactions between E. dharawalus, indicating a willingness to continue fighting. Concerningly, these outcomes point towards inherent and greater aggressiveness in C. destructor relative to E. dharawalus and that only larger E. dharawalus hold a competitive advantage over C. destructor. Therefore, we conclude that C. destructor represents a substantial threat to E. dharawalus through competitive behavioural interactions. Further, due to the disparity between our findings and those produced from laboratory-staged contests, we recommend the use of in situ studies when determining the behavioural impacts of invasive crayfish on natives.

Assessing discard consumption dynamic in shallow coastal environment using underwater video

Food webs are typically studied with respect to predator-prey interactions while often neglecting to address the role or importance of scavenging (i.e. carrion consumption) on their structure and functioning. In some environmental contexts, scavenging can play a key role in supporting populations and communities as exemplified by seabirds feeding on fisheries discards. Yet, a substantial part of fisheries discards sinks in the ocean where the range of consumers is most likely underestimated, owing to a lack of comprehensive understanding of the scavenging ability of marine fauna and local-scale empirical studies. Here, we deployed baited remote underwater video surveys (BRUVs) using discarded fish as baits to study the taxonomic richness of consumers, their interactions and consumption dynamic of discards in shallow coastal fishing grounds subject to discarding (Bay of Bourgneuf, North East Atlantic, France). We show that discards attract and are consumed by a wide taxonomic and trophic diversity of species. Interestingly, only one species observed is considered an obligate scavenger, while scavenging is little known or considered anecdotal in most other identified taxa. Analysis of consumption dynamic and interactions around discards revealed intra and interspecific competition among discard consumers as well as successions between consumers involved in predator-prey interactions. Data also suggests that decapods may facilitate discard consumption by non-piscivorous fish by generating scraps during consumption. We argue that underwater video surveys provide a valuable tool to better assess the occurrence and extent of facultative scavenging behaviors within marine communities and understand its role in food web dynamic.

Comparison of Underwater Visual Census (UVC), Underwater Remote Video (RUV), and Handline Used by Fisheries-Independent Programs to Assess Reef Fish

Fisheries managers rely on independent programs to collect data to design and inform management plans and decisions. These programs aim to implement cost-effective, robust monitoring approaches to assess spatiotemporal changes in fish structure and composition using different methods, which could lead to different conclusions. We compared four methods: two types of underwater visual censuses (UVC) (i.e. transect and roving), remote underwater video (RUV), and fishing using handline (HL) to characterize differences, identify potential complementarities, and estimate relative precision among them. All methods were simultaneously performed on 13 sites in Fajardo, eastern Puerto Rico for a total of 26 visits, and nine sites in La Parguera Natural Reserve (LPNR), western Puerto Rico for a total of 43 visits. Multivariate analysis revealed significant dissimilarities between all methods, with higher species diversity, abundances, and biomass observed in UVC and RUV. Patterns of spatial and temporal variation detected with UVC and RUV were significantly correlated among them, but not with HL. Univariate analysis for commercially important species showed that UVC and RUV methods were more precise and time efficient than HL. Complementarity between UVC and RUV was observed with the latter increasing the diversity observed during surveys. This study showed that UVC and RUV were different, however, they can be used complementarily they were significantly correlated. HL should be used when additional information (e.g. sex and age) is needed. A multi-gear approach with increased use in RUVs is recommended depending on the aims of the study to precisely assess a broader diversity.

Quantifying use of kelp forest habitat by commercially important crustaceans in the United Kingdom

AbstractKelp forests are regarded as important nursery and foraging habitats for commercially important species of finfish and shellfish despite an absence of fishery-independent data in many regions. Here, we conducted targeted surveys at 12 subtidal reefs, distributed across 9° of latitude in the UK, using three complementary techniques (Underwater Visual Census (UVC), Baited Remote Underwater Video (BRUV) and deployment of prawn pots) to quantify the abundance of crustaceans within kelp forests. Commercially important species were recorded at all sites; Cancer pagurus (brown/edible crab) and Necora puber (velvet swimming crab) were the most abundant and commonly observed, although Maja brachydactyla (spider crab), Homarus gammarus (European lobster) and Palaemon serratus (common prawn) were also recorded. The abundance of some species exhibited pronounced regional variability, with higher abundances of C. pagurus within northern regions and, conversely, higher abundances of M. brachydactyla and P. serratus within southern regions. Each sampling technique yielded similar spatial patterns for the most abundant species but had varying sensitivity to some species. Most individuals observed were juvenile or sub-adults, suggesting kelp forests serve as important nursery grounds for commercially and ecologically important crustaceans. Further monitoring efforts, conducted across greater spatiotemporal scales and in different habitat types, are needed to provide a robust baseline against which to detect changes and to inform management and conservation actions.

Fish community assessment of shelf-edge artificial reefs along east-central Florida

Artificial reef development is a popular management tool used to enhance fish stocks, mitigate degradation or loss of natural habitats, and provide additional recreational opportunities. In Florida, artificial reef development is a top priority of the Florida Fish and Wildlife Division of Marine Fisheries Management’s Artificial Reef program; however, despite their popularity and support, our scientific understanding of how artificial reefs ecologically facilitate marine fisheries is limited in certain habitats. This is especially true of the many artificial reefs along Florida’s central east coast, which has received sporadic assessment over the past few decades. The objective of this study was to conduct a fishery-independent assessment of the reef fish communities inhabiting the artificial reef sites off central east coast Florida using a combination of vertical longline sampling (VLL) and baited remote underwater videos (BRUV). Red snapper (Lutjanus campechanus) and vermilion snapper (Rhomboplites aurorubens) had the largest catch-per-unit-effort (CPUE) during VLL sampling, with red snapper CPUE significantly greater on rubble artificial reefs deeper than 40 m and vermilion snapper CPUE significantly greater on artificial reefs shallower than 30 m. A total of 62 fish species were observed on BRUVs, and community analysis indicated that red snapper, vermilion snapper, and tomtate (Haemulon aurolineatum) contributed the most to differences in community composition. Red snapper had the greatest abundance on sites deeper than 40 m, while vermilion snapper and tomtates had greater abundance on sites shallower than 30 m; all three species were more abundant during summer. This study provides the first quantitative assessment of reef fish communities on artificial reefs along central east coast Florida. Further monitoring is encouraged to unravel possible interactive effects of reef type and other environmental factors on fish assemblages, although it is evident that the reefs sampled support multiple managed fisheries species.

Regional patterns in demersal fish assemblages among subsea pipelines and natural habitats across north-west Australia

Regional patterns of fish diversity, abundance, distribution, and assemblage composition are driven by a combination of biotic and abiotic conditions in the marine environment, but these conditions can be altered through anthropogenic activities, such as those associated with oil and gas extraction. The present study utilises data on fish relative abundance and diversity obtained from 1546 baited remote underwater video deployments conducted between 2004 – 2019 in depths of 9 – 170 m across 2000 km of coastline in north-west Australia on natural habitats and subsea pipelines to understand the influence of oil and gas infrastructure on fish assemblages. A total of 450 fish taxa from 56 families was observed, with populations dominated by generalist and invertebrate carnivore taxa. At the regional scale, subsea pipelines had lower diversity (lower taxonomic richness) than natural environments, but possessed a higher abundance of piscivorous and herbivorous fish taxa. Clear patterns in fish assemblage composition were observed in multivariate analyses, reflecting the proximity of oceanic shoals and banks, depth, and to a lesser extent, oil and gas infrastructure. Shallow-water and close to shoals assemblages were characterised by a diversity of site-attached (e.g., wrasses, tuskfish), reef-associated taxa (e.g., emperors). Mesophotic fish assemblages were characterised by commercially important (e.g., goldband snapper), wide-ranging (e.g., sharks) and sand-affiliated (e.g., toadfish, threadfin bream) taxa. Proximity to pipelines and platforms ranked low as predictors in the multivariate analyses suggesting a negligible regional influence of these structures on fish communities in comparison to depth and shoal habitats. Local-scale influences of subsea infrastructure, however, may be important for some fish species (infrastructure vs. immediate surrounds). Our study highlights the influence of abiotic factors on regional-scale patterns in fish assemblage structure across north-west Australia.

Long‐duration remote underwater videos reveal that grazing by fishes is highly variable through time and dominated by non‐indigenous species

AbstractIn the marine environment, fish contribute to key ecological processes such as controlling food‐webs through top‐down impacts, especially on algae. To date, the assessment of fish grazing activity has mostly been performed using short‐term (<1 h) censuses by divers or remote cameras which do not allow estimating the variability of grazing rate within and between days. However, understanding the temporal variation of fish activity and hence contribution of species to ecosystem functioning is of particular interest in the context of biological invasion. Here, using long‐duration remote underwater cameras, we recorded fish abundance and grazing events over three consecutive days in October 2019 in a shallow Mediterranean ecosystem from northern Crete. This novel approach allowed us to assess temporal variation of abundance and grazing activity of the two native (Sarpa salpa and Sparisoma cretense) and the two non‐indigenous fish species (Siganus rivulatus and Siganus luridus). Non‐indigenous Siganus rivulatus was the most common species in the studied coastal habitat, followed by the two native species while the non‐indigenous Siganus luridus was scarce. Overall, the non‐indigenous S. rivulatus and the native S. salpa are responsible for more than 90% of the recorded grazing activity with similar bite rates between the two species. More than 70% of the grazing activity arose in grazing pulses in the afternoon, supporting the diel feeding hypothesis according to which feeding is greater in the afternoon when nutritive quality of macrophytes is the highest. In addition, some of the highest peaks in grazing activity were driven by a few individuals. Hence, surveys of only abundance could not provide accurate estimates of herbivory. Last, Siganus rivulatus presence did not significantly affect grazing activity of the native Sarpa salpa. Our results demonstrate that long‐duration remote underwater videos are a useful tool to accurately assess the contribution of fishes to ecosystem functioning.

Recreational fishing pressure impacts the density and behaviour of the western rock lobster (Panulirus cygnus, George): evidence from small, no-take marine reserves

Abstract It is unknown whether lobster behaviour is sensitive to fishing, as has been previously suggested for fish. This study investigated the impacts of recreational fishing on the wariness of western rock lobster (Panulirus cygnus, George) through comparisons between fished sites and no-take reserves (NTRs) at Rottnest Island. We used visual census to survey the density and size structure of lobster at sites inside and outside three comparable NTRs, and baited remote underwater video to study the behaviour of lobster at comparable sites. The density of legal-sized P. cygnus was twice as high within NTRs (0.34 ± 0.05) than fished sites (0.16 ± 0.03). P. cygnus also spent less time with bait in fished sites (0.42 ± 0.15 min) than NTRs (1.15 ± 0.28 min). This study provides the first evidence that behaviour of decapods, namely P. cygnus, is sensitive to noninjury-related disturbance associated with recreational fishing, with higher wariness in fished areas. Investigation is required into potentially confounding biotic and abiotic factors to further explore the applicability of lobster wariness as an indicator for fishing pressure.

Video surveys of sea snakes in the mesophotic zone shed light on trends in populations

Declines in abundance of sea snakes have been observed on reefs throughout the Indo-Pacific, although the reasons are unknown. To date, surveys have occurred on shallow reefs, despite sea snakes occurring over a large depth range. It is not known if populations of sea snakes in deep habitats have undergone similar declines. To address this, we analysed deep-water video data from a historical hotspot of sea snake diversity, Ashmore Reef, in 2004, 2016, and 2021. We collected 288 hours of video using baited remote underwater videos and a remotely operated vehicle at depths between 13 and 112 m. We observed 80 individuals of seven species with Aipysurus laevis (n = 30), Hydrophis peronii (n = 8), and H. ocellatus (n = 6) being the most abundant. Five of the species (A. duboisii, A. apraefrontalis, H. ocellatus, H. kingii, and Emydocephalus orarius) had not been reported in shallow waters for a decade prior to our study. We found no evidence of a decline in sea snakes across years in deep-water surveys, although abundances were lower than those in early shallow-water surveys. A comparison of BRUVS data from 2004 and 2016 was consistent with the hypothesis that predation by sharks may have contributed to the loss of sea snakes in shallow habitats. Our study highlights the use of underwater video to collect information on sea snakes in the mesophotic zone and also suggests that future monitoring should include these depths in order to capture a more complete representation of habitats occupied.

Observation of juvenile Eastern Blue Groper (Achoerodus viridis) on remnant oyster reefs in New South Wales, Australia.

Observation of juvenile Eastern Blue Groper (Achoerodus viridis) on remnant oyster reefs in New South Wales, Australia.

Oceanographic influences on reef fish assemblages along the Great Barrier Reef

Along the Great Barrier Reef, multi-year fish and oceanographic data were integrated to determine the shelf-scale relationship between oceanographic and environmental variables with reef fish assemblage structure. We used data from long-term monitoring of reef fishes sourced from underwater visual surveys and baited remote underwater video stations, combined with a regional hydrodynamic model, multi-year mooring observations, and surface drifting buoy trajectories. Multivariate analyses revealed distinct latitudinal splits in reef fish assemblages, which aligned spatially with the lagoonal branch of the East Australian Current. This lagoonal current flows onto the continental shelf in the central Great Barrier Reef (∼18-19°S), moves poleward within the lagoon, and exits out the Capricorn Channel (∼23°S). We used oceanographic data to characterize the lagoonal current and its seasonal variability, showing that the current reached its maximum strength during the peak reef fish spawning period in austral spring and summer. The shelf-scale structure in reef fish assemblages was best explained by the orientation angle from reef fish sites to the lagoonal current. The results indicate that the prevailing oceanographic patterns may enhance or restrict larval dispersal and contribute to maintaining distinct mesoscale reef fish communities. These findings provide insights for managing marine populations, and we consider how climate projections could influence large-scale ocean currents and environmental gradients and impact broad-scale reef fish community structure.

Mesophotic Reefs of the Largest Brazilian Coastal Protected Area: Mapping, Characterization and Biodiversity

Mesophotic reefs are poorly known worldwide despite their great ecological relevance and management importance for coral reef conservation strategies. To aid in filling this gap, we conducted a pioneering, large-scale survey, covering a total of around 315 miles in length, in the largest Brazilian coastal Marine Protected Area (MPA) Costa dos Corais. From the digital bathymetry model (30 to 50 m depth) generated by a multibeam echo sounder, we selected areas of greater geomorphological diversity for a detailed investigative expedition of mesophotic ecosystems. Various sampling techniques were used: single-beam echo sounders for detailing the relief, a remotely operated underwater vehicle (ROV) for habitat type investigation, baited remote underwater video (BRUV) for collecting images of the fish community, and scuba diving to perform transects describing the benthic and fish community. We analyzed reef environments from 20 to 68 m deep. As a result, we present the mapping and geomorphological characterization of two compartments of mesophotic reefs at 21–45 m depth and an image library of mesophotic ecosystems with the species description and indications of whether it is a new record in the region. Biodiversity data were collected covering eight species of sponges, with greater abundance for Ircinia spp., Aplysina spp., and Xestospongia muta; eight from corals, mainly Siderastrea spp. And Montastrea cavernosa; and 68 species of reef fish, with the Labridae family (including Scarinae—11 species) being the richest. Our results demonstrate the importance of mesophotic reefs for MPA Costa dos Corais reef biodiversity and, with that, the need to protect these areas through the application of local conservation strategies, such as the creation of “no-take zones”.

Removal of bottom-towed fishing from whole-site Marine Protected Areas promotes mobile species biodiversity

Marine habitats are being altered by anthropogenic pressures, influencing the diversity and distribution of species. Marine Protected Areas (MPAs) are increasingly used as spatial management tools to mitigate these human impacts on marine systems, but levels of protection vary. MPAs that exclude bottom-towed fishing activities from all habitats within their boundaries while still allowing static gear fishing, taking a “whole-site approach”, have shown benefits in terms of increasing biodiversity and biomass on reef habitat. In this study we use Baited Remote Underwater Videos (BRUVs) to quantify differences in mobile species assemblage composition, diversity and abundance on mixed sediment habitats inside and outside three MPAs adopting this whole-site approach within Jersey's territorial waters. Greater numbers of taxa were recorded within all three MPAs compared to nearby unprotected areas, with an average of 4.9 ± 1.8 taxa observed per deployment inside the MPAs compared to 3.9 ± 1.7 taxa outside the MPAs. No significant difference was observed in the overall abundance. The proportion of presence of two indicator species (Labridae Spp. and juvenile bream Spondyliosoma cantharus) was greater within the MPAs than the nearby unprotected areas. IUCN threatened and vulnerable shark species were observed on more BRUVs inside the MPAs, while IUCN threatened and vulnerable ray species were present on fewer BRUVs in the MPAs, but both occurred in too few numbers to statistically assess. Not all species responded positively to the MPAs, with scavenging species such as spider crab (Maja brachydactyla) recorded in lower numbers in two out of the three MPAs compared to the nearby unprotected areas. BRUV surveys that have been used to survey MPAs that exclude bottom-towed fishing in Europe are sparse and have primarily focussed on reef habitat. This study provides the first account of how this whole-site approach for mixed sediment habitats in three MPAs affects mobile species. These results evidence the merit of an ecosystem-based approach to MPA management for species of both conservation and commercial importance.

Fisher–shark interactions: A loss of support for the Maldives shark sanctuary from reef fishers whose livelihoods are affected by shark depredation

AbstractTargets for protecting predatory species often fail to consider the human costs of conservation. Human–wildlife interactions can increase following conservation action and present a major ecological and socioeconomic challenge. Using semistructured interviews (n = 103), participatory mapping (n = 57) and Baited Remote Underwater Video Stations (50 h) we investigated fisher‐shark interactions in one of the world's principal shark sanctuaries. Seventy‐three percent of respondents reported an increase in shark depredation postsanctuary implementation. Fisher‐reported losses due to shark depredation varied significantly between fisheries and were disproportionately high for reef fishers (>21% of daily vessel earnings). This is attributed to extensive spatial overlap (55%–78%) between reef fishing activity and ecologically validated shark hotspots. We show significant correlations between perceptions of depredation and support for shark sanctuary regulations. Findings demonstrate the need to consider fisher–shark interactions in current and future conservation planning and suggest that management of depredation must be sensitive to diverging perceptions among fisher groups.

Checklist of coral reef fishes of Darvel Bay, Sabah, Malaysian Coral Triangle, with a note on the biodiversity and community structure.

The Darvel Bay is a large semi-enclosed bay with spectacular natural land and seascape. The inward side of the Bay has only been recently known to be an important foraging ground for the endangered, threatened and protected (ETP) elasmobranch species, such as the Whale Shark and mobulid rays. Following a recent scientific expedition, we present a checklist of the coral reef fishes of Darvel Bay. A note on the biodiversity and community structure is presented, based on our analysis using diversity indices, univariate and multivariate approaches. Seven natural coral reefs comprising two fringing reefs and five patch reefs, were surveyed at 10 m depth using underwater visual census (UVC) and baited remote underwater video station (BRUVS) methods. A diverse list of 66 species of reef fishes from 17 families is recorded. However, this is overwhelmingly dominated by the small-sized omnivorous damselfish, family Pomacentridae (62%; N = 1485 individuals). Species richness and abundance were observed to increase at sites surveyed furthest from the coast within the Bay. Significantly distinct reef fish assemblages were observed between three priori groups, based on proximity to shore (ANOSIM, R = 0.65, p < 0.05). SIMPER analysis further revealed that 22 species of the total reef fish species recorded drive 76% dissimilarities between the groups. The pattern of the reef fish communities observed, reflected as a logseries distribution model, is that commonly found in disturbed habitats or habitats characterised by restricted resources in a community, where the dominant species takes up a high proportion of available resources. The ecological indices (Shannon-Wiener Diversity Index, 2.05; Simpson Index of Diversity, 0.79; Simpson Dominance Index, 0.20; and Pielou's Evenness Index, 0.43), all reflect the relatively low diversity and uneven species distribution of the reef fish community. We conclude that the present status of the coral reef fish community dominating Darvel Bay as having undergone a rapid shift in structure following intense and rampant fishing pressure, as reported by the media.

Accelerating Species Recognition and Labelling of Fish From Underwater Video With Machine-Assisted Deep Learning

Machine-assisted object detection and classification of fish species from Baited Remote Underwater Video Station (BRUVS) surveys using deep learning algorithms presents an opportunity for optimising analysis time and rapid reporting of marine ecosystem statuses. Training object detection algorithms for BRUVS analysis presents significant challenges: the model requires training datasets with bounding boxes already applied identifying the location of all fish individuals in a scene, and it requires training datasets identifying species with labels. In both cases, substantial volumes of data are required and this is currently a manual, labour-intensive process, resulting in a paucity of the labelled data currently required for training object detection models for species detection. Here, we present a “machine-assisted” approach for i) a generalised model to automate the application of bounding boxes to any underwater environment containing fish and ii) fish detection and classification to species identification level, up to 12 target species. A catch-all “fish” classification is applied to fish individuals that remain unidentified due to a lack of available training and validation data. Machine-assisted bounding box annotation was shown to detect and label fish on out-of-sample datasets with a recall between 0.70 and 0.89 and automated labelling of 12 targeted species with an F1 score of 0.79. On average, 12% of fish were given a bounding box with species labels and 88% of fish were located and given a fish label and identified for manual labelling. Taking a combined, machine-assisted approach presents a significant advancement towards the applied use of deep learning for fish species detection in fish analysis and workflows and has potential for future fish ecologist uptake if integrated into video analysis software. Manual labelling and classification effort is still required, and a community effort to address the limitation presented by a severe paucity of training data would improve automation accuracy and encourage increased uptake.

Predatory fish exploitation and relative abundance in a data‐poor region from the Caribbean coast of Colombia, inferred from artisanal fishery interview surveys and baited remote underwater video systems

Abstract Artisanal fisheries are socially and economically important along the Caribbean coast of Colombia. These fisheries remain poorly characterized, making it difficult to estimate their potential impact, especially on non‐targeted catch such as elasmobranchs. A rapid assessment framework was used to investigate the exploitation, use, and relative abundance of elasmobranchs and predatory teleosts in a region along the Caribbean coast of Colombia. In‐person structured interview surveys (n = 188) were conducted during the autumn of 2016 at eight fishing towns around Cartagena and the islands of the Natural National Park (NNP) Corales del Rosario and San Bernardo. Baited remote underwater video systems (BRUVS) (n = 200) deployed at four reefs were used to assess the relative abundance of elasmobranchs and other predatory teleosts. Fishers reported capturing eight shark and five ray taxa. Although 99% of fishers indicated that they did not target sharks, 83% reported keeping the catches to sell, eat, or both. Similarly, 81% of fishers did not target rays, but 90% reported retaining catches. Most fishers (84%) perceived a decline of sharks in the coastal waters since they started fishing, compared with 41% reporting batoid declines. BRUVS data support the perceptions of elasmobranch declines. Elasmobranch abundances were very low and no difference was detected between protected and unprotected reefs. Results were similar for teleost taxa, but there were higher abundances inside one marine protected area (MPA) that had high enforcement. Taken together, the data suggest considerable degradation of not only elasmobranch populations but also predatory fish populations of coral reef habitats in a region of the Colombian Caribbean. Furthermore, artisanal fishers continue to exploit coral reef resources inside MPAs unless there is strong enforcement. Enforcement can be difficult with limited resources, and therefore community‐based conservation methods may be more effective. Engagement with the local fishing community may help to reduce existing tensions and improve compliance.

Insights into elasmobranch composition, abundance, and distribution in the Bocas del Toro Archipelago, Panama using fisheries-independent monitoring

There are very little data on Central America and the Western Caribbean elasmobranch populations. Here we present the first results of standardized fisheries-independent monitoring of sharks and rays in Panama's Bocas del Toro Archipelago over four consecutive years. Using baited remote underwater videos (BRUVs), underwater visual census (UVC), and scientific longline, we provide baseline data on the composition, relative abundance, distribution, and demographics of seven shark species and six rays observed during the study. Observations differed among methods and habitats, with BRUVs (n = 148) more effective for assessing shark diversity and abundance and UVC (n = 108) more effective for assessing ray abundance. Nurse sharks (Ginglymostoma cirratum) and southern stingrays (Hypanus americanus) were the most commonly observed species overall (frequency of occurrence = 30.4 and 18.9%, respectively). Depth, temperature, and protected area status significantly affected the relative abundance of sharks and rays based on UVC and BRUV data. The use of scientific longlines (n = 55) highlighted the lagoon habitat as a potential nursery site for blacktip sharks (Carcharhinus limbatus), with juvenile (average total length = 113 cm) individuals representing 58% of longline captures. The results provide a current baseline for elasmobranch communities that can further aid in the conservation and management of these species nationally and regionally.