AI and eDNA monitoring is on its way to support a sustainable future for fisheries

Fill in the gaps of communication

To unlock the maximum potential of the newest methods, the first half of the TEFIMO project has focused on optimizing sampling procedure and expanding the range of species that can be automatically-detected, including for example benthic macro-fauna.

 “There are many important questions to answer like where in the catch handling process to take the best eDNA samples from and what is the maximum level of catch overlap on the sorting belt for accurate species identification from EM,” says Valentina Melli.

“DTU Aqua has the land- and sea-based facilities necessary to collect unique datasets for addressing these technical challenges and establishing best sampling practices”, which is what the TEFIMO project is contributing to.

An important project focus is also to fill in the gaps of communication with the end-users, e.g. the stakeholders of fisheries monitoring data.

This is why the TEFIMO project has been established as a partnership with the major end-users ICES and DFPO, hereby ensuring that the project's outcomes are aligned with their needs and can be effectively integrated into existing monitoring frameworks.

AI Monitoring methods with a more holistic focus on biodiversity

Another important goal of the TEFIMO project is to support more comprehensive ecosystem based fisheries assessments by ensuring that the new technologies and monitoring methods have a more holistic focus on biodiversity, rather than the traditional one on commercial fish species only.

For example, the current monitoring systems do not take in consideration benthic macro-fauna and flora such as molluscs (whelks) echinoderms (sea urchins), cnidarians (sea pens) and soft corals.

“A more complete picture of what biodiversity is caught in fishing nets is necessary to better assess fisheries impacts, monitor the habitats status and promptly detect changes in the ecosystem,” says Valentina Melli.

 All elements that are critical to pursue an ecosystem-based management approach, in accordance with the objectives of the Marine Strategy Framework Directive (MSFD)

Significant challenges

The lack of resources and effective methods for monitoring different fisheries as well as the status of ecosystems is the background for initiating TEFIMO.

According to the researchers, the monitoring of the ocean faces significant challenges that hinder our ability to make informed and sustainable decisions.

For instance, traditional monitoring methods require manual data collection of species composition and size distribution.

“Such limits are particularly concerning given the speed at which marine species are redistributing and ecosystems are shifting in response to stressors like climate change,” says Valentina Melli.

Working in synergy with EU efforts

The goals of the TEFIMO project perfectly aligns with international efforts going towards fully documented catches in commercial fisheries, such as those pursued by the Horizon EU project Optifish.

In June 2025, the projects joined forces thanks to the coordination efforts of Senior researcher Jordan P. Feekings, who plays a key role in both projects.

The data collection planned for TEFIMO offered the opportunity to attract international partners from OptiFish like EV ILVO and Wageningen Institute to make the most out of the resources available to collect data and work with them at the facilities in Hirtshals.

The collaboration was a great success, securing critical datasets to achieve the deliverables in each project while minimizing animal sampling and favoring exchanges of competences and perspectives across Institutes.

Gold standard datasets

In particular, the intensive week of activities led to the collection of several “Gold Standard datasets” where the entire catch composition (both species and sizes) was measured and each individual was tagged.

This will allow these datasets to be used for several years to test advancements in AI-based detection of species, automated length measurement and discard estimate, both within and beyond the end of TEFIMO.

The datasets are currently being processed, and TEFIMO is now half-way through the project period.