Line of action:

Marine ecosystems

Status:

En ejecución

Execution date:

2024

End date:

2026

Project web page:

https://www.programapleamar.es/proyectos/proyecto-costtera-2-interaccion-pesquera-y-comportamiento-del-delfin-mular-en-redes-de

The main objective of this project is to study the behaviour of bottlenose dolphins visually and acoustically during their interaction with the fishing nets of trawlers and to determine the degree and type of interaction between dolphins and fishing that occurs in Catalonia.

To this end, the project has different study blocks that include: surveys of artisanal and trawling fishermen, maritime monitoring campaigns (studies of distribution, behavior and photo-id of bottlenose dolphins), installation of cameras and hydrophones in fishing nets and the development of an Artificial Intelligence (AI) algorithm for the analysis of images and sound. improving the protocol established during the COSTtERA project in 2022.

• Surveys of 30% of artisanal and trawling fishermen in Catalonia, to analyse the degree and type of interaction that occurs in the territory and its impacts on both dolphins and fishermen (estimation of accidental catches and estimation of the damage that these interactions can generate for fishermen).
  SUBMON has carried out this type of survey in previous years and the results will be used to analyse the trend of these interactions in the area.
• Maritime outings throughout the year to continue with the study of the bottlenose dolphin population in northern Catalonia carried out by SUBMON since 2017, with special attention to the photo-identification, density, and behaviour of dolphins.

The data obtained will allow the density of animals in the area to be analyzed, the degree of dependence on fishing to be evaluated and the percentage of fishing association and the distribution in the area to be compared between seasons and years. Photographs of each sighting will be taken and included in the entity’s own photo-identification catalog , which includes more than 600 different bottlenose dolphins. During the outings, the surface behavior of the dolphins when they interact with the fishing boats will be recorded. The surface behaviour data obtained will be analysed to establish patterns that are repeated and that can be related to underwater images and acoustic recordings, thus associating surface behaviours with underwater behaviours and sounds.

• Installation of underwater cameras in trawls to study the underwater behaviour of animals when they interact with fishing nets.

Implementation of video cameras in trawl nets to study the feeding behavior of bottlenose dolphins during trawling.
The cameras will be placed on different boats, specifically in those where they have had the most dolphin sightings in recent years and that fish at less than 200 meters, which is the area of presence of the target species.

The installation of the cameras will be carried out following the protocol defined during the COSTtERA project.
They shall be placed along the net, with particular attention to the entrance of the net in order to be able to conclude, among other things, whether the animals enter the trawls in this area and are therefore subject to a higher risk of accidental capture.

• Installation of hydrophones in fishing nets , to record the sounds of the animals and to be able to detect when there are dolphins in the nets. This technique will make it possible to study the acoustic feeding patterns of dolphins, obtain information on the degree of interaction and quantify the interactions that occur.

 

• Development of an Artificial Intelligence algorithm to automatically detect bottlenose dolphin individuals recorded by the hydrophone with the collaboration of Keacoustics and the UPC.

Data exploration and preprocessing will be carried out to evaluate the peculiarities of the recordings. Noise statistics will be extracted, and a review of the type of events and their spatio-temporal statistical characteristics will be carried out, as well as their occurrence. Machine learning models will be developed for the automated detection and classification of sounds of interest (clicks, whistles, impulsive sounds). Acoustic patterns of interest will be identified, which may be associated with feeding episodes.

• Organization of conferences with the fishing sector, to visualize the images recorded, as well as to review the results obtained.