Presentation

Our century is marked by a race against time to limit global warming to 1.5 °C above pre-industrial levels, as agreed under the Paris Agreement by 192 Parties in December 2015. In this context, the structuring of regional green hydrogen value chains is a major strategic challenge, particularly for the island territories of the South-West Indian Ocean (SWIO), whose abundant yet intermittent renewable resources hold considerable — and largely untapped — potential.

Accurately assessing this potential, however, runs up against a central bottleneck: the effective accessibility of the climate and energy data required to characterise renewable resources. Where such data exist, they are scattered across institutions, produced according to heterogeneous standards and stored in disparate formats, which hinders their discovery, interpretation and reuse. The island territories concerned also suffer from observation gaps and incomplete time series, limiting the reliability of analyses. To these challenges are added the scientific hurdles of climate downscaling: computational constraints, stochasticity of deep learning models, the capacity to generalise beyond the training period, and the need to preserve the physical consistency of high-resolution reconstructed fields.

To overcome these obstacles, an approach combining data engineering, climate modelling and machine learning is now being mobilised. U-Net neural network architectures, coupled with ingestion, normalisation and interoperable distribution pipelines (OPeNDAP, THREDDS), make it possible to transform a fragmented set of resources into a coherent, well-documented and easily exploitable ecosystem. These approaches pave the way for a reliable assessment of green hydrogen production potential and for the development of decarbonisation scenarios tailored to territorial realities.

The LIGHTEN-IO project — Leadership Initiative for Green Hydrogen Transition Energy Network in the Indian Ocean —, scientifically led by Prof. Béatrice Morel, is part of this dynamic. It aims to structure a regional network of stakeholders around green hydrogen in the SWIO area, to generate reference regionalised climate and energy data, to assess green hydrogen production potential across the target territories, and to build sectoral decarbonisation scenarios. The project brings together an international consortium including the Université des Mascareignes, the University of Mauritius, IST-T Antananarivo, the University of Nairobi, the Université des Comores, Forschungszentrum Jülich, Météo-France and the Seychelles Meteorological Authority.

The LIGHTEN-IO project is funded with a total budget of €598,939.78 under the INTERREG VI Indian Ocean 2021-2027 programme (Action Sheet 1.3), with a €509,098.80 (85%) contribution from the ERDF and an €89,840.98 (15%) matching contribution from the Région Réunion. Europe is committed in Réunion through ERDF funding.

This project strengthens local skills at the interface of climate sciences, energy and data sciences, and contributes to the structuring of a regional green hydrogen value chain serving the energy transition and the resilience of island territories in the South-West Indian Ocean.