USE CASE 2: A better model for wildfire prediction.
How do you know where a wildfire is actually likely to happen? This is what our project’s second use case is about.
The Fire Weather Index (FWI) is the standard tool across Europe, and it’s useful for warning the public. But it ignores fuel and runs at ~9 km resolution. In Catalonia, that means large areas are flagged as high risk despite few historical fires.
So Lobelia developed a new Wildfire Probability Model, trained on real wildfire occurrence and landscape type:
- 100 m resolution instead of ~9 km
- 3–4× more accurate than the FWI at predicting daily wildfires
The result? Road and railway operators can see which kilometres of their network are most exposed today, and how that changes by 2050 under different climate scenarios.
Read the full story here.

Visit the MIRACA Workbench to explore how Lobelia’s Wildfire Probability Model is used to assess the vulnerability of the Catalan transport network under present and future climate change scenarios.



Leave a Reply
Want to join the discussion?Feel free to contribute!