ILIAD
Coastal Crete: a Digital Twin of the Ocean for near real-time oil spill detection
A Digital Twin of the Ocean for near real-time oil spill detection.
The project
ILIAD, funded by the Horizon Europe programme under grant agreement No 101037643, is a Research and Innovation project aimed at developing interoperable digital twins of the ocean, virtual representations that integrate Earth observation, modelling and digital infrastructures to provide predictions of future developments “at sea”. The project enhances existing data and promotes the use of innovative technologies through a marketplace and community engagement.
MEEO’s role
Within ILIAD, MEEO co-developed Coastal Crete, the Cretan Sea Digital Twin, together with the Foundation for Research and Technology – Hellas (FORTH). Coastal Crete provides advanced, high-resolution forecasting services, enhanced by the integration of Sentinel data, real-time in-situ observations and Machine Learning algorithms, that allow the simulation of predictive (what-if) scenarios.
MEEO acts as a technical partner, contributing to the Oil Spill Response Ocean Twin, one of the thematic twins developed under the Cretan Sea Digital Twin, which provides early detection of marine oil spills and accurate, short-term operational forecasting of spill trajectories to support immediate response to pollution events.
MEEO's technical contribution is structured as a three-stage pipeline:
- Preprocessing: Sentinel-1 SAR imagery is prepared for automatic analysis (selecting VV polarization), reducing noise, converting and normalising the data, and enhancing contrast. Output: a clean radar image, ready for automatic analysis.
- AI workflow: the processed image is analysed to identify the spill, combining object detection (FCOS), semantic segmentation (U-Net) and traditional methods (adaptive thresholding). Output: a binary map of the spill, giving its extent and position.
- Forecasting and dynamic modelling: the detected spill feeds the MEDSLIK-II dispersion model, a particle-tracking model coupled to high-resolution numerical weather (WRF), hydrodynamic (NEMO) and sea-state (WAVEWATCH III) models for the Cretan Sea; its results are compared against real Sentinel-1 data. Output: a forecast of the spill's future trajectory.
Beyond this detection-to-forecast chain, the Digital Twin also delivers an operational picture for the deployment of response measures and offers oil spill forecasting as an on-demand (what-if) tool, applicable to the global ocean. Results are visualised on GeoMachine, which turns the model output into an interactive map.
- 3
- Stages, from satellite detection to trajectory forecast