Skip to main content
OSTI.GOVU.S. Department of Energy Office of Scientific and Technical Information
U.S. Department of Energy
Office of Scientific and Technical Information
 

Recurrent Convolutional Deep Neural Networks for Modeling Time-Resolved Wildfire Spread Behavior

Journal Article · · Fire Technology
 [1];  [2];  [3];  [4]
  1. Google Research, Mountain View, CA (United States); SLAC
  2. Stanford Univ., CA (United States)
  3. Google Research, Mountain View, CA (United States)
  4. Google Research, Mountain View, CA (United States); Stanford Univ., CA (United States); SLAC National Accelerator Laboratory (SLAC), Menlo Park, CA (United States)
+ Show Author Affiliations
The increasing incidence and severity of wildfires underscores the necessity of accurately predicting their behavior. While high-fidelity models derived from first principles offer physical accuracy, they are too computationally expensive for use in real-time fire response. Low-fidelity models sacrifice some physical accuracy and generalizability via the integration of empirical measurements, but enable real-time simulations for operational use in fire response. Machine learning techniques have demonstrated the ability to bridge these objectives by learning first-principles physics while achieving computational speedups. While deep learning approaches have demonstrated the ability to predict wildfire propagation over large time periods, time-resolved fire-spread predictions are needed for active fire management. Here, in this work, we evaluate the ability of deep learning approaches in accurately modeling the time-resolved dynamics of wildfires. We use an autoregressive process in which a convolutional recurrent deep learning model makes predictions that propagate a wildfire over 15 min increments. We apply the model to four simulated datasets of increasing complexity, containing both field fires with homogeneous fuel distribution as well as real-world topologies sampled from the California region of the United States. We show that even after 100 autoregressive predictions representing more than 24 h of simulated fire spread, the resulting models generate stable and realistic propagation dynamics, achieving a Jaccard score between 0.89 and 0.94 when predicting the resulting fire scar. The inference time of the deep learning models are examined and compared, and directions for future work are discussed.
Research Organization:
SLAC National Accelerator Laboratory (SLAC), Menlo Park, CA (United States)
Sponsoring Organization:
USDOE Office of Science (SC)
Grant/Contract Number:
AC02-76SF00515
OSTI ID:
1998774
Journal Information:
Fire Technology, Journal Name: Fire Technology Journal Issue: 6 Vol. 59; ISSN 0015-2684
Publisher:
SpringerCopyright Statement
Country of Publication:
United States
Language:
English

References (22)

Wildland Fire Spread Modeling Using Convolutional Neural Networks journal March 2019
A multiphase formulation for fire propagation in heterogeneous combustible media journal March 1998
Climate relationships with increasing wildfire in the southwestern US from 1984 to 2015 journal March 2020
Combustion machine learning: Principles, progress and prospects journal July 2022
Deep learning journal May 2015
A physics-based approach to modelling grassland fires journal January 2007
Wildland surface fire spread modelling, 1990 - 2007. 2: Empirical and quasi-empirical models journal January 2009
Wildland surface fire spread modelling, 1990 - 2007. 1: Physical and quasi-physical models journal January 2009
Wildland surface fire spread modelling, 1990 - 2007. 3: Simulation and mathematical analogue models journal January 2009
LANDFIRE: a nationally consistent vegetation, wildland fire, and fuel assessment journal January 2009
A comparison of three approaches for simulating fine-scale surface winds in support of wildland fire management. Part I. Model formulation and comparison against measurements journal January 2014
Representation and evaluation of wildfire propagation simulations journal January 2014
Machine learning–accelerated computational fluid dynamics journal May 2021
A Prescription for Ecological Disaster journal April 1994
Deep Residual Learning for Image Recognition conference June 2016
Warming and Earlier Spring Increase Western U.S. Forest Wildfire Activity journal August 2006
A review of a new generation of wildfire–atmosphere modeling journal June 2019
A review of machine learning applications in wildfire science and management journal December 2020
Mortality due to Vegetation Fire–Originated PM 2.5 Exposure in Europe—Assessment for the Years 2005 and 2008 journal January 2017
A Spatio-Temporal Neural Network Forecasting Approach for Emulation of Firefront Models conference September 2022
Standard fire behavior fuel models: a comprehensive set for use with Rothermel's surface fire spread model report January 2005
Global fire emissions estimates during 1997–2016 journal January 2017

Figures / Tables (16)

Table 1(p. 4)table Table 1
Figure 1(p. 6)figure Figure 1
Figure 2(p. 8)figure Figure 2
Figure 3(p. 8)figure Figure 3
Figure 4(p. 10)figure Figure 4
Figure 5(p. 14)figure Figure 5
Figure 6(p. 15)figure Figure 6
Figure 7(p. 16)figure Figure 7
Figure 8(p. 17)figure Figure 8
Figure 10(p. 18)figure Figure 10
Figure 9(p. 18)figure Figure 9
Figure 11(p. 19)figure Figure 11
Figure 12(p. 19)figure Figure 12
Figure 13(p. 19)figure Figure 13
Table 2(p. 20)table Table 2
Figure 14(p. 23)figure Figure 14

Similar Records

Machine learning and artificial intelligence for wildfire prediction
Technical Report · Thu Apr 15 00:00:00 EDT 2021 · OSTI ID:1769739
A high-resolution large-eddy simulation framework for wildland fire predictions using TensorFlow
Journal Article · Tue Oct 17 20:00:00 EDT 2023 · International Journal of Wildland Fire · OSTI ID:2511288
Building a machine learning surrogate model for wildfire activities within a global Earth system model
Journal Article · Mon Mar 07 19:00:00 EST 2022 · Geoscientific Model Development (Online) · OSTI ID:1853837

Related Subjects

54 ENVIRONMENTAL SCIENCES
deep neural networks
machine learning
recurrent neural networks
wildland fires