Citation

Penman TD, McColl-Gausden SC, Cirulis BA, Kultaev D, Ababei DA, Bennett LT. J. Environ. Manage. 2021; 301: 113789.

Copyright

(Copyright © 2021, Elsevier Publishing)

DOI

10.1016/j.jenvman.2021.113789

PMID

34592661

Abstract

Wildfire extent and their impacts are increasing around the world. Fire management agencies use fire behaviour simulation models operationally (during a wildfire event) or strategically for risk assessment and treatment. These models provide agencies with increased knowledge of fire potential to improve identification of the best strategies for reducing risk. One of the greatest areas of uncertainty in fire simulations is the data relating to fuel, which are usually based on simplified response trajectories with time since fire within vegetation communities. There is a clear need to better predict relevant fuel variables across landscapes to reduce uncertainties in fire simulations. In this study, we compare the performance of fuel hazard models based on environmental variables (environmental model) with those currently implemented based on a negative exponential relationship with time since fire (NEGEXP) using the state of Victoria in south-eastern Australia as an environmentally diverse case study. The models predicted similar broadscale patterns in fuel hazard but with considerable regional variation. The NEGEXP model was less accurate than the environmental model, which had 41-47% accuracy on an independent data set cf. 24-35% for NEGEXP. Model differences resulted in significant differences in the extent and spatial location of predicted fires with NEGEXP consistently predicting larger fires. Fuel is made up of the live and dead components of vegetation, both of which are influenced by a range of environmental factors. As our study highlights, ignoring environmental factors in simple fuel models based on broad vegetation types (like NEGEXP) will likely compromise the predictive accuracy of fire behaviour models. Only when environmental factors are accounted for can we more accurately predict fuels across landscapes and thereby improve the accuracy of fire behaviour predictions and the estimation of fire risks.

Language: en

Keywords

Fire behaviour; Fire prediction; Fire risk; Forest fuel; Vegetation; Wildfire