High-fidelity CNN-based surrogate modeling for wildfire spread forecasting
Saeed Nematshahi, Rui Fan, Amin Khodaei, Erin Belval, Ali Arabnya · 2026
Read High-fidelity CNN-based surrogate modeling for wildfire spread forecasting by Saeed Nematshahi, Rui Fan, Amin Khodaei, Erin Belval, Ali Arabnya (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13482362.txt.
Description
Nematshahi and co-authors train a neural network in 2026 to forecast wildfire spread almost instantly, replacing physics simulators that take too long when evacuation and power shutoff decisions cannot wait. Their convolutional network, which reads maps the way image software reads pictures, learned from high-resolution fire simulations to predict where a blaze moves next. Speed is the point, since fire crews and utilities plan around forecasts that arrive in time to act.
The surrogate model scored 0.92 on the F1 measure, which balances missed fire against false alarms, while delivering each forecast in a fraction of a second of computing time. The authors aim it at two users at once, electric utilities deciding preemptive safety shutoffs and firefighting teams weighing evacuation calls under time pressure. They frame the trade as a loss of simulation detail in exchange for forecasts fast enough to shape decisions while a fire is young.
Catalog record
| CREATOR | Saeed Nematshahi, Rui Fan, Amin Khodaei, Erin Belval, Ali Arabnya |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 33 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |