TY - JOUR
T1 - A generalised model of rainforest vulnerability to fire in eastern Australia
AU - Gordon, Christopher E.
AU - Khanal, Shiva
AU - Nolan, Rachael H.
AU - Peacock, Ross J.
AU - Boer, Matthias M.
PY - 2025/10
Y1 - 2025/10
N2 - Fuel load/structure, fuel dryness, ignition frequency and meteorological condition (“fire weather”) are universal constraints governing fire activity globally. However, our understanding of how these factors control fire activity in rainforests is limited, especially at landscape scales. We used random forest models to identify key correlates of fire occurrence during Australia's 2019/20 fire season in ecologically important world heritage Gondwanan rainforests of northern New South Wales. Our analyses focused on identifying factors facilitating the occurrence of both fire (burned vs unburned), and high severity fire (high severity tree canopy vs low severity understory fire) across four major rainforest types. More than 90 % of rainforest area burned experienced low severity fire, meaning that our models of fire occurrence proxied low severity responses. Variables associated with fuel dryness (soil moisture, negative) and fire weather (wind speed, positive parabolic; relative humidity, negative) were the most important correlates of fire occurrence. However, variables associated with fuel dryness (live fuel moisture content, negative; soil moisture, negative) and fuel load/structure (canopy height, negative; tree cover, positive) were the most important correlates of high severity fire occurrence. Key correlates of fire occurrence differed between cool temperate and other rainforest types. Our results suggest that key drivers of rainforest wildfire differ between areas burned by low and high severity fire and highlight important environmental correlates useful for future fire behaviour prediction and risk planning. This information will be crucial for management interventions aiming to retain fire-sensitive rainforests, in-light of projected increases in fire activity.
AB - Fuel load/structure, fuel dryness, ignition frequency and meteorological condition (“fire weather”) are universal constraints governing fire activity globally. However, our understanding of how these factors control fire activity in rainforests is limited, especially at landscape scales. We used random forest models to identify key correlates of fire occurrence during Australia's 2019/20 fire season in ecologically important world heritage Gondwanan rainforests of northern New South Wales. Our analyses focused on identifying factors facilitating the occurrence of both fire (burned vs unburned), and high severity fire (high severity tree canopy vs low severity understory fire) across four major rainforest types. More than 90 % of rainforest area burned experienced low severity fire, meaning that our models of fire occurrence proxied low severity responses. Variables associated with fuel dryness (soil moisture, negative) and fire weather (wind speed, positive parabolic; relative humidity, negative) were the most important correlates of fire occurrence. However, variables associated with fuel dryness (live fuel moisture content, negative; soil moisture, negative) and fuel load/structure (canopy height, negative; tree cover, positive) were the most important correlates of high severity fire occurrence. Key correlates of fire occurrence differed between cool temperate and other rainforest types. Our results suggest that key drivers of rainforest wildfire differ between areas burned by low and high severity fire and highlight important environmental correlates useful for future fire behaviour prediction and risk planning. This information will be crucial for management interventions aiming to retain fire-sensitive rainforests, in-light of projected increases in fire activity.
KW - Conservation
KW - Fire risk
KW - Fuel moisture
KW - Fuel structure
KW - Gondwanan rainforest
KW - High severity fire
KW - LiDAR
KW - Random forest model
UR - http://www.scopus.com/inward/record.url?scp=105012943373&partnerID=8YFLogxK
U2 - 10.1016/j.jenvman.2025.126799
DO - 10.1016/j.jenvman.2025.126799
M3 - Article
C2 - 40812103
AN - SCOPUS:105012943373
SN - 0301-4797
VL - 393
JO - Journal of Environmental Management
JF - Journal of Environmental Management
M1 - 126799
ER -