Funding period: 02/08/2021 – 01/04/2022
Funding agency: The project was funded as a part of an “Innovations Connections” research grant in collaboration with the industry partner, RedEye Australia.
Project leader, researchers and collaborators: Dr Chris Gordon (Project leader) Dr Rachael Nolan, Prof Matthias Boer, Dr Anne Griebel, Dr Rachael Nolan acquired the research funding
Project summary: Fire refugia are normally wet locations that experience less wildfire than their surroundings. Due to this, refugia act as “fire breaks” that limit wildfire spread and total burn area. Although rare, wildfires can occur in fire refugia during severe droughts which dry-out vegetation and allow it to burn; however, the degree of landscape dryness required to promote wildfire in refugia is poorly understood. In this project, we determined how eleven variables that differently represent “landscape dryness” impacted wildfire occurrence in forest types that act as fire refugia and non-refugia. We found that vapour pressure deficit, soil moisture and live vegetation moisture content were the best predictors of fire occurrence in both refugia and non-refugia, with vapour pressure deficit the most important. We further found that wildfire was less common in refugia because they experienced a smaller number of days across the year when conditions for wildfire can occur, rather than due to inherent differences in the level of landscape dryness required to promote wildfire. Collectively, our results will be used to increase the accuracy of fire hazard prediction at large spatial scales.
The figures are from a high-level executive summary provided to RedEye. The figures are also part of an academic publication that is currently “in review” in the journal “Geophyscial Research Letters”: Gordon, C.E., Boer, M.M., Griebel, A., Yebra, M., Sturgess, A, Collins, L. & Nolan, R.H. 2022. Fuel moisture moderates wildfire resistance in fire refugia. Geophyscial Research Letter. In Review.
Figure 1. (a) Map of the forest types acting as fire refugia, including rainforests (RF), wet sclerophyll forests with a shrubby (WSF-s) and grassy understory (WSG-g) and non-fire refugia (dry sclerophyll forests; DSF-sg). (b) shows an insert of (a) (for the black rectangle). (c) Shows the study area in east Australia (black rectangle).
Figure 2 to the left. Line plots showing how predicted wildfire probability varies with live fuel moisture content (LFMC-e), soil moisture content (SM-r) and vapour pressure deficit (VPD) within the forest types acting as fire refugia (RF, WSF-s, WSF-g) and non-refugia (DSF-sg; see Figure 1 for forest type acronyms). Only forest types where LFMC-e, SM-r and VPD were important drivers of wildfire occurrence are shown. The plots show that wildfire probability decreases as live fuel moisture and soil moisture content increases; however, it increases as vapour pressure deficit increases. The plots also show similar responses across forest types acting as fire refugia and non-refugia, which implies that similar levels of “landscape dryness” are required to promote wildfire across these forest types.
Figure 3 above. Density histogram showing the percentage of days between 2002–2020 when vapour pressure deficit was above “critical moisture thresholds”, which represent periods with an elevated likelihood of wildfire occurrence. Responses are shown separately for forest types acting as fire refugia (RF, WSF-s, WSF-g) and non-refugia (DSF-sg; see Figure 1 for forest type acronyms). The histogram shows relativized counts of all 30 m pixels within the study area (y-axis) across the range of the percentage of days above the critical moisture thresholds data (x-axis). The dashed lines show median values across the distributions. The plot suggests that the percentage of days above the critical moisture thresholds was lowest in the forest type acting as a persistent fire refugia (rainforests) and highest in the forest type acting as a non-refugia (dry sclerophyll forests).