Shaddy et al. (2026) Generative Algorithms for Wildfire Progression Reconstruction from Multi-Modal Satellite Active Fire Measurements and Terrain Height

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Remote Sensing
• Year: 2026
• Date: 2026-01-10
• Authors: Bryan Shaddy, Brianna Binder, Agnimitra Dasgupta, Hanqiao Qin, James Haley, A. Farguell, Kyle Hilburn, Derek V. Mallia, Adam K. Kochanski, Jan Mandel, Assad A. Oberai
• DOI: 10.3390/rs18020227

Research Groups

Not explicitly mentioned in the provided text.

Short Summary

This study develops and validates a conditional Wasserstein Generative Adversarial Network (cWGAN) to estimate wildfire progression history from satellite and terrain data, achieving an average Sørensen–Dice coefficient of 0.81 against observed perimeters.

Objective

• To develop an approach for estimating wildfire progression history from VIIRS active fire measurements, GOES-derived ignition times, and terrain height data, addressing the limitations of traditional coupled atmosphere-wildfire models.

Study Configuration

• Spatial Scale: Pacific US wildfires (regional scale).
• Temporal Scale: Multi-day simulations of historic wildfires.

Methodology and Data

• Models used: WRF-SFIRE (coupled atmosphere–wildfire model), Conditional Wasserstein Generative Adversarial Network (cWGAN).
• Data sources: VIIRS active fire measurements (satellite), GOES-derived ignition times (satellite), terrain height data, high-resolution perimeters measured via aircraft (validation data).

Main Results

• The cWGAN approach successfully generates fire progression estimates consistent with WRF-SFIRE solutions.
• Validated on five Pacific US wildfires, the approach achieved an average Sørensen–Dice coefficient of 0.81 when compared against high-resolution aircraft-measured perimeters.
• Terrain data was found to have an increased contribution to fire progression estimates when other measurements were uninformative.

Contributions

• Introduces a novel cWGAN-based data assimilation technique for estimating wildfire progression history.
• Provides a measurement-based assessment tool for wildfire state, improving upon the limitations of traditional coupled atmosphere-wildfire models.
• Demonstrates the utility of integrating satellite observations and terrain data for more accurate wildfire progression estimates.

Funding

Not explicitly mentioned in the provided text.

Citation

@article{Shaddy2026Generative,
  author = {Shaddy, Bryan and Binder, Brianna and Dasgupta, Agnimitra and Qin, Hanqiao and Haley, James and Farguell, A. and Hilburn, Kyle and Mallia, Derek V. and Kochanski, Adam K. and Mandel, Jan and Oberai, Assad A.},
  title = {Generative Algorithms for Wildfire Progression Reconstruction from Multi-Modal Satellite Active Fire Measurements and Terrain Height},
  journal = {Remote Sensing},
  year = {2026},
  doi = {10.3390/rs18020227},
  url = {https://doi.org/10.3390/rs18020227}
}