Over recent years, research within FLARE’s Landscape Flammability theme has been uncovering how microclimate influences vegetation flammability and fire behaviour. A key finding has been the importance of in-forest vapour pressure deficit (VPD) — a measure of the drying power of the atmosphere beneath the forest canopy.
Understanding how atmospheric conditions influence fuel moisture and ignitability could provide new ways to predict where and when vegetation is likely to burn. We are now building on this research to develop high-resolution tools that could help fire practitioners plan prescribed burns and better anticipate wildfire risk.
What is VPD?
Vapour Pressure Deficit (VPD) represents the drying power of the atmosphere. As VPD increases, the atmosphere can remove moisture more rapidly from both live and dead vegetation, increasing fuel drying. VPD can be calculated from temperature and relative humidity.
Firefighters are generally more familiar with relative humidity (RH) as an indicator of atmospheric dryness. However, RH alone can be misleading because the drying power of the atmosphere changes substantially with temperature. VPD incorporates both temperature and humidity into a single, physically meaningful measure of atmospheric drying.
VPD and fire activity
High VPD has long been recognised as an important driver of fire activity because dry atmospheric conditions promote fuel drying. Previous research has linked VPD with fire ignition, the number of fires, overnight fire activity and area burned.
Our research is taking this further by examining how VPD operates within forests, where local conditions can differ substantially from those measured by standard weather stations in the open.
VPD as a key predictor of ignitability
Quantifying when forests will ignite — their ignitability — is an important first step towards understanding forest flammability. We conducted field-based ignition experiments across three fire seasons in different eucalypt forests, testing how ignition varied with fuel moisture, atmospheric dryness and forest characteristics.
We found that near-surface fuel moisture was the strongest predictor of ignitability, with in-forest VPD driving variation in fuel moisture. The conditions required for ignition also varied between forest types. Dense, productive forests in wetter environments required drier fuel and atmospheric conditions to ignite than more open forests in drier environments.
This provides a basis for predicting how ignitability varies across a landscape. Days when ignitability is consistently high may be more conducive to the development of large wildfires, while days with greater spatial variation in ignitability may provide more suitable conditions for prescribed burning.
Researchers testing the ignitability of forest fuels under a range of weather conditions.
VPD and the development of large forest fires
The spatial variation in VPD can also influence how large fires develop. Mesic forests are often among the slowest forest types to dry and can act as barriers to fire spread when conditions are relatively moist.
We analysed 35 large forest fires in Victoria to identify the factors influencing whether active fires spread through mesic forest. VPD was the strongest predictor of fire occurrence, with the likelihood of fire spreading through mesic forest increasing by 65% as VPD increased from 2.5 to 7 kPa.
Topography and disturbance history also influenced fire occurrence, but their effects were substantially smaller than the influence of VPD, particularly under very high VPD. This suggests that as atmospheric conditions become increasingly dry, the capacity of wetter forests to act as barriers to fire spread may weaken.
Conceptual model of the key factors influencing the probability that an active wildfire will propagate through mesic forest. Variables in italics were identified as important. Values indicate the range of conditions needed for pr > 0.7 of burning in mesic forest with all other variables held at their mean.
Bringing forest microclimate into fire planning
Conditions beneath a forest canopy can differ substantially from those measured in open environments. Topography, canopy cover and other landscape features create local microclimates that influence how quickly vegetation dries.
Standard weather products generally represent open conditions and therefore do not capture this variation. Our research has shown that in-forest VPD is a better predictor of ignitability than VPD measured in open conditions.
We have developed a model that predicts in-forest VPD at a 30 m resolution, allowing atmospheric drying conditions to be represented across forested landscapes. We are now testing how this information can be applied to planned burning.
Downscaling 5 km gridded Open VPD to predicted 30 m in-forest VPD for 1st March 2019. Modelling is for a densely forested landscape in the Yarra Ranges surrounding the township of Powelltown.
Predicting burnability within prescribed burns
The probability that an area will burn — its burnability — varies across a landscape and changes with weather conditions. Being able to predict these patterns could help practitioners anticipate how fire will behave within a prescribed burn.
We combined high-resolution in-forest VPD with topography, soil moisture and fuel management zones to develop a 30 m spatial model of burnability. The model identifies areas within a burn block that are more or less likely to burn, with areas predicted to have a higher probability of burning generally aligning with areas that subsequently burned.
This approach could help practitioners better anticipate variation in burn outcomes and understand how changing weather and landscape conditions influence the effectiveness of prescribed burning.
Predicting fuel moisture
Fuel moisture is a critical consideration when deciding when and where to conduct a planned burn. Being able to predict fuel moisture across a landscape several days ahead could improve planning and help practitioners identify suitable burning conditions.
We combined in-forest VPD with soil moisture to develop a 30 m model of dead fuel moisture across dry and wet eucalypt forests and different fuel strata, including suspended fuels, surface litter and profile litter.
The resulting VPD–soil moisture model outperforms existing operational fuel moisture models, providing more precise, location-specific predictions and reducing the tendency of current models to underestimate fuel moisture.
From microclimate to fire planning
Together, this research is building a new approach to understanding and predicting forest flammability. By linking atmospheric dryness, forest microclimate, fuel moisture and ignitability, we are developing tools that can represent how the likelihood of burning varies across landscapes and through time.
Our goal is to translate this knowledge into better information for planning prescribed burns, anticipating wildfire risk and understanding how forests may respond to a changing climate.