After the execution of the COSTERA and COSTERA-2 projects, SUBMON promotes COSTERA-3 as a new phase of consolidation and progress aimed at deepening the knowledge of the bottlenose dolphin and its interaction with trawling in the north of Catalonia. COSTERA-3 continues the study of the spatial distribution, behaviour and evolution of the degree of dependence of fishing activity over time. However, the project not only maintains the lines of monitoring developed in previous phases, but also incorporates new approaches aimed at better understanding the species.

The main novelties include the study of body condition using non-invasive techniques, as well as the development of predictive models capable of estimating the population response to different scenarios of fishing effort and environmental variability. This approach makes it possible to anticipate possible impacts derived from changes in the availability of resources and support decision-making in terms of management.

The project integrates advanced technological tools – drones, underwater cameras (including 360º), hydrophones and artificial intelligence – which allow a more complete and accurate view of the interaction between dolphins and fishing activity.

In addition, COSTERA-3 reinforces the active participation of the fishing sector, incorporating its knowledge of the marine environment and promoting a collaborative approach. In this way, the project lays the foundations for adaptive management based on scientific evidence, aimed at making the conservation of the bottlenose dolphin compatible with the sustainability of fishing activity.

Methodology

1. Boat departures

Distribution throughout the year: monitoring the presence and distribution of the bottlenose dolphin during different times of the year to identify areas of preferential use, seasonal changes and relationship with fishing activity.

Photo-identification (Photo-ID): photographic record of the individuals observed for their individual identification and monitoring over time. This tool allows estimating area loyalty, residence patterns and group composition.

Surface behaviour: collection of standardised data on surface behaviour (feeding, movement, socialisation, surveillance, evasion, etc.) to detect repetitive patterns and contexts of interaction.

Drones with LiDAR: non-invasive behavior and biometrics: during outings, drones equipped with LiDAR technology are used to document surface behavior and obtain biometric measurements. This allows body condition (size, body condition, proportions) to be assessed, health indicators to be explored, and, when possible, to detect signs compatible with gestational status.

2. Shipments on trawlers

Installation of underwater cameras: installation of cameras in trawl nets following the protocol developed in the COSTtERA and COSTtERA-2 projects to analyse the behaviour of dolphins during their interaction with fishing activity. As a technological improvement, 360º cameras are incorporated to significantly expand the field of view compared to previous studies. This more comprehensive coverage reduces blind spots and allows for better interpretation of the position, trajectories and maneuvers of animals around the net, increasing accuracy in risk assessment.

Acoustics, installation of hydrophones in the networks: this system will allow the presence of individuals interacting in the networks to be detected, thus expanding the monitoring capacity in conditions of low visibility or operational limitation. The technique will allow further study of the acoustic patterns of dolphins during interaction, obtain information on the degree of interaction and quantify the interactions that occur.

3. Integration and modeling of ecological, fisheries, and environmental variables

Development of models to estimate the population response of bottlenose dolphins in different scenarios.

4. Artificial intelligence

The project optimizes artificial intelligence algorithms for the automatic detection of bottlenose dolphins in acoustic recordings and underwater videos obtained in interaction with trawling, giving continuity to the advances developed in COSTtERA-2.

Machine vision models will be improved using new convolutional neural network (CNN) architectures. In addition, the combined analysis of acoustic data and underwater images is integrated, synchronizing both sources to detect behavioral patterns linked to specific acoustic signals.

Collaboration with the fishing sector

As in previous projects, the participation of the fishing sector in this initiative is essential for the understanding of dolphin interactions. For this reason, their support will be available during the different phases of the project and conferences will be organized with the fishing sector, to visualize the images recorded, as well as to review the results obtained.

INTEMARES complementary project.