TwinTrack - Hackathon 2025

Understanding where fish go and the impact of human activities is crucial for marine biodiversity protection and regulatory compliance, such as under the Marine Strategy Framework Directive (MSFD).

The TwinTrack project aims to combine two types of fish tracking data (archival and acoustic telemetry) enriched with information from the Fisheries Atlas and Zooplankton EOV vLabs of BlueCloud, to develop a prototype digital twin for fish movement in the North Sea on the EDITO platform.

Our approach leverages predictive models of fish movement based on Long Short-Term Memory (LSTM) modelling, a machine-learning method well suited for time-series predictions.By integrating habitat, temperature, and prey fields with tracking data, LSTM allows us to generate realistic simulations of fish paths and explore their responses to future environmental scenarios. Our application will focus on skates (Raja clavata, Raja brachyura, Raja montagui).

Ultimately the application built here will be expanded to additional species such as Atlantic cod (Gadus morhua), European eel (Anguilla anguilla), and bluefin tuna (Thunnus Thynnus), for which high-quality datasets exist.

To maximize accessibility and stakeholder engagement, we will embed these models in interactive visualization tools. A web-based viewer will allow users to explore predicted fish movements alongside human pressures such as offshore wind farms. This digital twin prototype represents a first step toward dynamic, data-driven tools that link ecological science with management needs in the North Sea.