A Graph Neural Network for Global Daily Fire Radiative Power Prediction at Medium-Range Lead Times
Abstract
Skillful prediction of biomass-burning activity several days in advance is important for air-quality forecasting and aerosol prediction. Two operational constraints motivate this work. First, the GBBEPx satellite fire radiative power (FRP) product used to initialize NOAA's GEFS-Aerosols is available with about a 1.5-day latency, so each forecast cycle relies on the most recently available, but already outdated, fire observations. Second, these fire inputs are then held fixed throughout the subsequent 5-day operational forecast, or 7 days in the GSL experimental system, effectively assuming no evolution in fire activity. We develop a data-driven model that predicts global FRP one to seven days ahead from the most recent available observations. The model adapts a spatiotemporal graph neural network using reanalysis meteorology, land-cover and vegetation information, recent fire history, and GBBEPx FRP as the training target. It is trained on 2020-2022 data and evaluated for 2023-2024. The model reproduces the global seasonal cycle and substantially outperforms persistence. At 0.1∘ resolution, mean squared error is reduced by 32% at one-day lead and 43% at seven days in 2023, and by 24% and 40% in 2024. At 1∘ resolution, the critical success index ranges from 0.32 to 0.60. Detection skill declines only modestly with lead time, whereas intensity skill degrades more rapidly. Large fires are detected reliably, but their radiative power is systematically underestimated. These results demonstrate useful predictability of fire activity several days ahead and identify intensity calibration and small-fire placement as the main remaining challenges before predicted FRP can support operational aerosol forecasts.
Related
- A neural network-based Universal Thermal Climate Index for reliable global thermal-stress classification across extreme weather
- Summer seasonal precipitation prediction in the middle and lower reaches of the Yangtze River using neural network approaches
- Biogeochemistry-Informed Neural Network (BINN v1.0) for improving accuracy of model prediction and scientific understanding of soil organic carbon storage
- Global mineral constraints on dust shortwave radiative effects
- Global Harmonization of Life Cycle Assessments for Small- and Medium-Scale Wastewater Treatment
BibTeX
@misc{zhang2026graph,
title = {{A Graph Neural Network for Global Daily Fire Radiative Power Prediction at Medium-Range Lead Times}},
author = {Li Zhang and Jun Wang and Isidora Jankov and Yongxin Liu and Gonzalo A. Ferrada and Ravan Ahmadov and Ligia Bernardet and Haonan Chen and Shobha Kondragunta},
howpublished = {arXiv},
year = {2026},
month = oct,
doi = {10.48550/arxiv.2610.11022},
eprint = {2610.11022},
archivePrefix = {arXiv},
url = {https://arxiv.org/abs/2610.11022},
}
RIS
TY - UNPB
TI - A Graph Neural Network for Global Daily Fire Radiative Power Prediction at Medium-Range Lead Times
AU - Li Zhang
AU - Jun Wang
AU - Isidora Jankov
AU - Yongxin Liu
AU - Gonzalo A. Ferrada
AU - Ravan Ahmadov
AU - Ligia Bernardet
AU - Haonan Chen
AU - Shobha Kondragunta
PY - 2026
DA - 2026/10/08
PB - arXiv
DO - 10.48550/arxiv.2610.11022
UR - https://arxiv.org/abs/2610.11022
AB - Skillful prediction of biomass-burning activity several days in advance is important for air-quality forecasting and aerosol prediction. Two operational constraints motivate this work. First, the GBBEPx satellite fire radiative power (FRP) product used to initialize NOAA's GEFS-Aerosols is available with about a 1.5-day latency, so each forecast cycle relies on the most recently available, but already outdated, fire observations. Second, these fire inputs are then held fixed throughout the subsequent 5-day operational forecast, or 7 days in the GSL experimental system, effectively assuming no evolution in fire activity. We develop a data-driven model that predicts global FRP one to seven days ahead from the most recent available observations. The model adapts a spatiotemporal graph neural network using reanalysis meteorology, land-cover and vegetation information, recent fire history, and GBBEPx FRP as the training target. It is trained on 2020-2022 data and evaluated for 2023-2024. The model reproduces the global seasonal cycle and substantially outperforms persistence. At 0.1^∘ resolution, mean squared error is reduced by 32% at one-day lead and 43% at seven days in 2023, and by 24% and 40% in 2024. At 1^∘ resolution, the critical success index ranges from 0.32 to 0.60. Detection skill declines only modestly with lead time, whereas intensity skill degrades more rapidly. Large fires are detected reliably, but their radiative power is systematically underestimated. These results demonstrate useful predictability of fire activity several days ahead and identify intensity calibration and small-fire placement as the main remaining challenges before predicted FRP can support operational aerosol forecasts.
ER -