| Nome: | Descrição: | Tamanho: | Formato: | |
|---|---|---|---|---|
| 1.73 MB | Adobe PDF |
Autores
Orientador(es)
Resumo(s)
O tubarão-azul (Prionace glauca), um predador pelágico-oceânico amplamente distribuído, enfrenta forte exploração apesar da sua importância ecológica, tornando urgente a sua boa gestão e conservação. O turismo de mergulho com tubarões-azuis está em ascensão, adquirindo crescente relevância social e económica nos Açores, o que possibilita uma plataforma informal e extensa de recolha de dados. Utilizando estes dados oportunísticos recolhidos por mergulhadores e cientistas locais, este estudo explora se uma metodologia padronizado e não-invasivo de foto-identificação (photo-ID) pode identificar de forma fiável indivíduos de tubarão ao longo do tempo, tanto dentro como entre anos. Foram analisados dados de 2019–2021 e 2024, focados na Pedra de Sousa, no Monte Submarino Condor e no Banco Princesa Alice, recorrendo a imagens e vídeos oportunísticos recolhidos por investigadores e mergulhadores locais durante expedições científicas e de mergulho com tubarões. As imagens selecionadas foram analisadas com base em três métricas de identificação: (i) morfologia da margem da barbatana dorsal, (ii) padrão dos poros das ampolas de Lorenzini e (iii) características complementares como marcas de mordidas, marcas de artes de pesca ou outros traços visíveis. Além disso, o comprimento total dos tubarões foi medido através de fotogrametria a laser sempre que possível. A metodologia desenvolvida neste estudo permitiu identificar 34 indivíduos únicos, tendo 4 deles sido reavistados em anos diferentes. O estudo sugeriu que a barbatana dorsal é o traço mais fiável a longo prazo entre as três métricas analisadas, sobretudo quando características adicionais estão disponíveis para confirmar a identidade do tubarão ao longo do tempo. O mapeamento manual dos poros das ampolas de Lorenzini revelou-se a métrica menos eficaz para ser aplicada ao material disponível, devido a problemas de visibilidade, elevado consumo de tempo e baixa qualidade das imagens. No entanto, esta característica pode ser de interesse se forem utilizados softwares avançados e sistemas de aprendizagem automática (IA) para o seu mapeamento. O facilmente reconhecível “Scarface” foi usado para validar a metodologia.
A metodologia desenvolvida foi útil para identificar mais 4 indivíduos (todos juvenis ou machos adultos) que regressaram à área de estudo ao longo dos anos. Estes reavistamentos sugerem uma potencial fidelidade ao local, consistente com o conhecimento prévio sobre padrões sazonais de segregação sexual na região. A investigação sublinha o valor da colaboração intersectorial — entre investigadores, operadores turísticos e cientistas cidadãos — para maximizar coletivamente os esforços de investigação, reconhecendo ao mesmo tempo os desafios de utilizar dados oportunísticos. Estas limitações dificultam a identificação de tubarões individuais e devem-se principalmente à baixa qualidade das imagens em alguns registos, bem como à ausência de métricas de referência visíveis em determinados fotogramas. Tal deve-se frequentemente à presença simultânea de um elevado número de tubarões ou a posicionamentos subótimos da câmara em relação ao animal. Foram apresentadas recomendações para melhorar a qualidade dos dados disponíveis para esforços futuros e monitorização a longo prazo, resumidas no documento “Boas Práticas de Recolha de Dados”. Em última análise, o estudo demonstra que um método de foto-ID de baixo custo, replicável e baseado principalmente na morfologia única da barbatana dorsal, mas complementado com outros traços, pode ser utilizado de forma eficaz para identificar tubarões-azuis ao longo do tempo. A integração de outras tecnologias na monitorização do tubarão-azul pode automatizar o processo e melhorar a qualidade dos resultados. Este estudo pode ser usado como ferramenta para o avanço da investigação sobre fidelidade ao local, movimentos animais, residência, estrutura e abundância populacional, potenciais áreas de maternidade, bem como os possíveis impactos de uma indústria de mergulho com tubarões em crescimento.
ABSTRACT: The blue shark (Prionace glauca), a widely distributed pelagic-oceanic predator, faces heavy exploitation despite its ecological importance, making its good management and conservation urgent. Diving tourism with blue sharks is rising, with increasing social and economic relevance in the Azores, which allows for an informal and extensive data collection platform. Using this opportunistic data collected by divers and local scientists, this study explores whether a standardized, non-invasive photo-identification (photo-ID) methodology can reliably identify shark individuals over time, both within and across years. Data from 2019–2021 and 2024, focused on the Pedra de Sousa, Condor Seamount, and Princess Alice Bank, were analyzed, using opportunistic images and videos collected by researchers and local divers during scientific and shark diving expeditions. Selected images were analyzed using three different identification metrics: (i) dorsal fin edge morphology, (ii) ampullae of Lorenzini pore pattern, and (iii) complementary traits such as bite marks, fishing gear marks, or other visible features. Besides, the total length of the sharks was measured using laser photogrammetry when possible. The methodology developed in this study identified 34 unique individuals, with resighting of 4 of them across years. This study suggested that the dorsal fin is the most reliable long-term trait among the three metrics analyzed, especially when additional traits are available to confirm the identity of the shark over time. The ampullae of Lorenzini hand pore mapping proved to be the least effective metric to be applied to available media, due to visibility issues, time consumption, and low image quality. However, this trait can be of interest if advanced software and AI machine learning systems are used to map it. The easily recognizable “Scarface” was used to validate the methodology. The developed methodology helped identify another 4 individuals (all juveniles or adult males) who have returned to the study over the years. These resightings suggest potential site fidelity, which is consistent with the previous knowledge on seasonal sexual segregation patterns in the area. The research highlights the value of cross-sector collaboration—among researchers, tourism operators, and citizen scientists—to collectively maximize research efforts, while acknowledging the challenges of using opportunistic data. These limitations hinder the identification of individual sharks and are mainly due to the low image quality in some media, as well as the absence of visible reference metrics in certain frames. This is often caused by a high number of sharks appearing simultaneously or by suboptimal camera positioning relative to the shark. Recommendations were proposed to improve the quality of the data available for future efforts and long-term monitoring and were summarized into the “Data Collection Best Practices.” Ultimately, the study demonstrates that a low-cost, replicable photo-ID method mainly based on the dorsal fin's unique morphology but complemented with other traits can be effectively used to identify blue sharks over time. Integrating other technologies into blue shark monitoring may automate the process and enhance the quality of the results. This study may be used as a tool to advance research concerning site fidelity, animal movement, residency, population structure and abundance, potential nursery areas, as well as the potential impacts of a growing shark diving industry.
ABSTRACT: The blue shark (Prionace glauca), a widely distributed pelagic-oceanic predator, faces heavy exploitation despite its ecological importance, making its good management and conservation urgent. Diving tourism with blue sharks is rising, with increasing social and economic relevance in the Azores, which allows for an informal and extensive data collection platform. Using this opportunistic data collected by divers and local scientists, this study explores whether a standardized, non-invasive photo-identification (photo-ID) methodology can reliably identify shark individuals over time, both within and across years. Data from 2019–2021 and 2024, focused on the Pedra de Sousa, Condor Seamount, and Princess Alice Bank, were analyzed, using opportunistic images and videos collected by researchers and local divers during scientific and shark diving expeditions. Selected images were analyzed using three different identification metrics: (i) dorsal fin edge morphology, (ii) ampullae of Lorenzini pore pattern, and (iii) complementary traits such as bite marks, fishing gear marks, or other visible features. Besides, the total length of the sharks was measured using laser photogrammetry when possible. The methodology developed in this study identified 34 unique individuals, with resighting of 4 of them across years. This study suggested that the dorsal fin is the most reliable long-term trait among the three metrics analyzed, especially when additional traits are available to confirm the identity of the shark over time. The ampullae of Lorenzini hand pore mapping proved to be the least effective metric to be applied to available media, due to visibility issues, time consumption, and low image quality. However, this trait can be of interest if advanced software and AI machine learning systems are used to map it. The easily recognizable “Scarface” was used to validate the methodology. The developed methodology helped identify another 4 individuals (all juveniles or adult males) who have returned to the study over the years. These resightings suggest potential site fidelity, which is consistent with the previous knowledge on seasonal sexual segregation patterns in the area. The research highlights the value of cross-sector collaboration—among researchers, tourism operators, and citizen scientists—to collectively maximize research efforts, while acknowledging the challenges of using opportunistic data. These limitations hinder the identification of individual sharks and are mainly due to the low image quality in some media, as well as the absence of visible reference metrics in certain frames. This is often caused by a high number of sharks appearing simultaneously or by suboptimal camera positioning relative to the shark. Recommendations were proposed to improve the quality of the data available for future efforts and long-term monitoring and were summarized into the “Data Collection Best Practices.” Ultimately, the study demonstrates that a low-cost, replicable photo-ID method mainly based on the dorsal fin's unique morphology but complemented with other traits can be effectively used to identify blue sharks over time. Integrating other technologies into blue shark monitoring may automate the process and enhance the quality of the results. This study may be used as a tool to advance research concerning site fidelity, animal movement, residency, population structure and abundance, potential nursery areas, as well as the potential impacts of a growing shark diving industry.
Descrição
Palavras-chave
Blue sharks Mid-Atlantic Pelagic-oceanic predator Diving tourism
Contexto Educativo
Citação
HERRERA, Andrea. (2025). "Image-Based Identification of Blue Sharks in the Mid-Atlantic: Evaluating a Standardized Methodology for Long-Term Monitoring". Plentzia: Universidad del País Vasco, 2025, 34 p. Dissertação de Mestrado em Ciências Económicas e Empresariais. Disponível em http://hdl.handle.net/10400.3/8908