EGU26-19404, updated on 14 Mar 2026
https://doi.org/10.5194/egusphere-egu26-19404
EGU General Assembly 2026
© Author(s) 2026. This work is distributed under
the Creative Commons Attribution 4.0 License.
Poster | Wednesday, 06 May, 08:30–10:15 (CEST), Display time Wednesday, 06 May, 08:30–12:30
 
Hall X1, X1.16
Modelling burned area and emissions with deep learning
Seppe Lampe1, Lukas Gudmundsson2, Basil Kraft2, Stijn Hantson3, Emilio Chuvieco4, and Wim Thiery1
Seppe Lampe et al.
  • 1Vrije Universiteit Brussel, Department of Water and Climate, Brussels, Belgium (seppe.lampe@vub.be)
  • 2Institute for Atmospheric and Climate Science, ETH Zürich, Zürich, Switzerland
  • 3School of Sciences and Engineering, Universidad del Rosario, Bogotá, Colombia
  • 4Environmental Remote Sensing Research Group, Universidad de Alcalá, Alcalá de Henares, Spain

Wildfires play a key role in the Earth system by shaping ecosystem dynamics and influencing the carbon cycle and atmospheric composition. Data-driven models have recently emerged as powerful tools for reproducing observed fire activity, particularly burned area, across a range of spatial and temporal scales. The first version of BuRNN (Burned area and emissions modelling through Recurrent Neural Networks) focused solely on burned area and outperformed all process-based fire-coupled DGVMs from ISIMIP over a wide range of spatial, temporal and spatio-temporal skill metrics. Here we present the 2nd version of BuRNN, a data-driven model that now jointly represents burned area and fire-related emissions.

How to cite: Lampe, S., Gudmundsson, L., Kraft, B., Hantson, S., Chuvieco, E., and Thiery, W.: Modelling burned area and emissions with deep learning, EGU General Assembly 2026, Vienna, Austria, 3–8 May 2026, EGU26-19404, https://doi.org/10.5194/egusphere-egu26-19404, 2026.