Determining landscape fine fuel moisture content of the Kilmore East ‘Black Saturday’ wildfire using spatially-extended point-based models
Highlights
► The Kilmore East wildfire on Black Saturday was the deadliest in Australia's history. ► Understanding the behaviour of such high intensity bushfires is very important. ► Fine fuel moisture content (FMC) is an important determinant of bushfire behaviour. ► Non-spatial FMC models were extended to study landscape fuel moisture on the day. ► Results suggest fuel moisture across the landscape was critically low at <5%.
Introduction
On 7 February 2009, over 300 wildfires occurred across the State of Victoria, Australia (Teague et al., 2010). This day, known as “Black Saturday”, followed two periods of heatwave conditions in south-eastern Australia in which a number of record maximum temperatures were set (Bureau of Meteorology, 2009). Thirteen of these fires developed into major incidents (Fox and Runnalls, 2009), five of which resulted in 173 fatalities (all on the afternoon and evening of 7 February), the destruction of over 2200 houses and 400,000 ha burnt (Teague et al., 2010), with a total estimated cost of over AU$4 billion (Teague et al., 2010). The Kilmore East fire (Fig. 1), which started just before noon on private property from a broken powerline (Teague et al., 2010) and burnt through mixed farmland, woodland and native forest (predominantly mixed dry/wet eucalyptus), was responsible for 121 (70%) of the fatalities and 32% of the total area burned.
Fires on Black Saturday exhibited extremely intense behaviour and rapid rates of spread, both prior to and following the passage of a cold front across the state (Teague et al., 2010). The spread of the Kilmore East fire, particularly once it entered long unburnt forested areas in broken topography associated with the Hume Range east of Wandong, was characterised by intense spotting–the lofting of firebrands by the convection column of the fire that are then carried downwind of the fire to start new fires (Ellis, 2010). These spotfires developed rapidly, greatly hampering suppression efforts and resulting in extensive non-contiguous spread of multiple fires burning simultaneously, themselves generating firebrands and further spotfires (Tolhurst, 2009; Cruz et al., 2012), that may have contributed to many of the fatalities and house-losses.
Most of the energy of a bushfire comes from dead fine surface fuel, grasses, fallen leaves, twigs and bark <6 mm in diameter (Sullivan et al., 2012). In eucalyptus forests, dead fine surface fuel also provides the bed of fuel in which firebrands ignite and spotfires develop. The moisture content of this fuel controls its combustibility and, along with wind speed, is a primary factor in determining the behaviour (speed and intensity) of bushfires (e.g. Anderson and Rothermel (1965); Burrows (1999); Sullivan (2009)). FMC is determined by short and long term weather patterns (Matthews, 2006) and the depth of fuel (Matthews et al., 2007).
As part of a research project to reconstruct the propagation and behaviour of the Kilmore East fire to understand the pattern and chronology of events of the fire on that day, Cruz et al., 2010, Cruz et al., 2012 found it necessary to determine the range, variation and dynamics of a number of critical elements that determined the behaviour of the fire. These elements included the prevailing weather (wind speed and direction, air temperature, relative humidity), fuels (type, structure, amount, moisture), and topography. Many of these could be obtained from historical records for the day. Fuel moisture content (FMC), however, was not measured for any fuel on Black Saturday.
The prevailing weather conditions and experience (e.g. Luke and McArthur (1978)) tell us dead fine FMC must have been low to cause the observed fire behaviour. Determining just how low, particularly considering the magnitude and duration of the preceding heatwaves, and what the variation in FMC across the landscape was, are important for understanding and interpolating the behaviour and spread of the fire for the reconstruction. It is also important for inferring likely recurrence of conditions in the future, particularly when the occurrence of such extreme weather conditions is expected to become more frequent (IPCC, 2007). To fill the gap in the knowledge of the spatial and temporal variation in fine FMC across the fire area on the day, it was necessary to simulate the FMC using existing models suitable for the fuels involved.
A number of models of dead fine FMC have been developed for Australian fuels (Viney, 1991). Some are based on phenomenological indicators such as observed combustibility (Burrows, 1984), measured indirectly by oven-drying (Matthews, 2010), by changes in chemical or electrical characteristics (Chatto and Tohurst, 1997), or analytically from external parameters (e.g. simple correlations of atmospheric variables (McArthur, 1966, McArthur, 1967; Sneeuwjagt and Peet, 1985), fully physical process-based models (Nelson, 1984; Matthews, 2006), or simplifications of complex models (Matthews et al., 2010; Sharples and McRae, 2011)). As many of these models were constructed for particular conditions (e.g. wetting phase of the afternoon) or are only feasible with direct sampling of fuels, most are not suitable for the purpose of ‘hindcasting’ moisture content, particularly with respect to the effect of extreme weather conditions prior to the period of interest. Furthermore, all current models of moisture content are point-based and thus have no explicit spatial elements that vary with topography, elevation or latitude.
This paper describes the method of extending two point-based models of dead fine FMC (one for grassy fuels, one for forest litter fuels) to a spatial context across the landscape and presents the FMC results for the afternoon of Black Saturday for the region of the Kilmore East fire. Due to the lack of actual observations of FMC on the day, measurement of fuel moisture at two locations in unburnt fuel within the final burned area soon after the fire is used to assess the accuracy of the model. The implementation is then used to investigate the impact of the heatwave antecedent weather conditions prior to Black Saturday on the fine FMC. Suggestions for further development of such landscape scale FMC modelling are also given.
Section snippets
Antecedent conditions
The 2008/09 fire season in south-eastern Australia was preceded by a number of years of significant rainfall deficit and increased average daily temperatures (Bureau of Meteorology, 2009). For 12 years preceding Black Saturday, much of the state of Victoria had received the lowest rainfall on record (1900–2009). The fire season (generally from October to March (Luke and McArthur, 1978; Long, 2006)) commenced in late 2008 with spring rainfall totals below long-term averages for most districts.
Landscape categorisation
The area of interest for this study was defined as a rectangle surrounding and including the burnt area of the fire at midnight on the evening of 7 February (see Fig. 1). Data on topography, fire history, and vegetation classification within this area were imported into a GIS to enable spatial analysis. These data were then categorised for three classes of fuel type (two forest and one grass), four classes of elevation, aspect and slope (Table 1) to produce a manageable yet suitably detailed
Single-point FMC, no topography
Fig. 5 shows the modelled hourly FMC in dry woodland, wet forest litter and grasslands for the period 6–8 February 2009. These results are for flat topography (i.e. no slope or aspect) at 500 m elevation. The forest FMC is presented as the profile moisture content (i.e. the whole depth of the litter fuel), calculated as the mean of all five layers within the Matthews model, and the top or surface of the litter fuel, calculated as the mean of the upper two layers in the model.
On the day of the
Discussion
Results of this FMC modelling exercise suggest that on the afternoon of 7 February 2009 extraordinarily low (<5%) fine FMC extended across the vast majority of the fire area. Generally, fires in eucalypt forests become severe and extremely dangerous when FMC values fall below 10%, with crowning likely and most firebrands able to start spotfires (Sullivan et al., 2012). At moisture contents less than 5%, fire behaviour in typical bushfire fuels is expected to be highly erratic and of extremely
Conclusions
Two non-spatially explicit FMC models were used to investigate the spatial and temporal variation in FMC across the area of the Kilmore East fire, a major wildfire that occurred on the afternoon of 7 February 2009 in Victoria, Australia. Model results suggest that the FMC across the landscape during the run of this fire was uniformly less than 5%, a value that is commonly associated with rapid growth of fire and extreme and erratic fire behaviour (McArthur, 1967; Sullivan et al., 2012).
The
Acknowledgements
We would like to thank the following people for providing assistance during this project: Tivi Theiveyanathan for net undercanopy radiation data, Anders Siggins and Richard Hurley for GIS spatial analysis and mapping, Jim Gould and Miguel Cruz for vegetation measurements, and Miguel, Matt Plucinski and Jason Sharples for constructive comments on the manuscript. We acknowledge the valuable input to the improvement of the paper by three anonymous referees. We would also like to acknowledge the
References (40)
- et al.
Influence of moisture and wind upon the characteristics of free-burning fires
Symposium (International) on Combustion
(1965) - et al.
Anatomy of a catastrophic wildfire: the Black Saturday Kilmore East fire in Victoria, Australia
Forest Ecology and Management
(2012) - et al.
Modeling topographic influences on fuel moisture and fire danger in complex terrain to improve wildland fire management decision support
Forest Ecology and Management
(2011) - et al.
Behind the flaming zone: predicting woody fuel consumption in eucalypt forest fires in southern Australia
Forest Ecology and Management
(2011) Stochastic space-time weather models from ground-based data
Agricultural and Forest Meteorology
(1995)Grassland fire management in future climate
Advances in Agronomy
(2010)Meteorological Aspects of the 7 February 2009 Victorian Fires, an Overview
(2009)Describing Forest Fires in Western Australia: a Guide for Fire Managers
(1984)Fire behaviour in jarrah forest fuels: 2. Field experiments
CALMScience
(1999)- et al.
The Development and Testing of the Wiltronics T-H Fine Fuel Moisture Meter
(1997)