An exploratory analysis of forest fine fuel consumption and accumulation using forest inventory data and fire history
Trung H. Nguyen
A B * , Simon Jones A , Karin J. Reinke A and Mariela Soto-Berelov A
A
B
Abstract
Estimating changes in fine fuel loads (FFL) is essential for carbon monitoring and fire management. Field measurements of post-fire fuel response are challenging, leading to reliance on generalised fuel types in operational models.
This study presents a proof-of-concept for estimating fine fuel consumption and accumulation by integrating forest inventory and fire records, aiming to refine fuel dynamics estimates and enhance current practices.
We estimated FFL changes across vertical strata in southeast Australian eucalypt forests, considering burn severity, fire type and forest cover. Fuel consumption was estimated by correlating pre-fire observations with combustion factors defined by burn severity. Fuel accumulation was predicted using modified Olson models with dynamic input parameters.
Wildfires typically occurred in forests with higher FFL and consumed more fuels than prescribed burns. Closed forests experienced greater fuel loss compared with open and woodland forests. Increasing fire severity led to lower decomposition rates and a longer time to reach pre-fire FFL, with denser forests showing higher accumulation rates.
Integrating forest inventory and fire history data offers valuable insights into fuel dynamics, potentially enhancing existing fuel hazard models.
The approach is applicable in regions with mature forest inventories and advanced fire severity mapping.
Keywords: accumulation, burn severity, consumption, eucalypt forests, fine fuel, fine fuel load estimation, forest cover, forest inventory, prescribed burn, wildfire.
Introduction
Fine fuels, defined as those with a diameter of less than 6 mm and primarily comprising leaves, sticks, twigs, bark and grass, exhibit high combustibility and exert a substantial influence on the ignition and propagation of fires (McArthur 1962, 1967; Gould et al. 2011). Therefore, accurate characterisation of fine fuel attributes, including fine fuel loads (FFLs), and monitoring their dynamic changes following fires is imperative (Gould et al. 2011; McCaw et al. 2012). Understanding the dynamics of fine fuels is becoming increasingly important in the context of climate change (Miezïte et al. 2022) and the rising frequency of fire events, especially in fire-prone forests across the Mediterranean Basin, western United States, Russia and southeast Australia (Liu et al. 2010; Matthews et al. 2012; Enright et al. 2015; Whitman et al. 2019; Nolan et al. 2020). Forest fine fuels are consumed by fires, and then gradually reaccumulate over time (Dodge 1972; Birk and Simpson 1980; Raison et al. 1983; McCaw et al. 2002). Consequently, a detailed understanding of these dynamics informs wildfire risk assessments, fire behaviour predictions and prescribed burning planning, facilitating proactive forest management strategies (Dodge 1972; Raison et al. 1983; Fernandes et al. 2006; Hollis et al. 2015; Nolan et al. 2024). Furthermore, this understanding is essential for assessing the impact of wildfires on carbon cycling, as fuel combustion releases carbon dioxide (CO2) into the atmosphere, affecting global carbon budgets and climate feedback mechanisms (Bradstock et al. 2012; Loehman et al. 2014; Bowman et al. 2021b).
Fuel consumption can be accurately measured through direct field measurements, which often involve collecting fuel samples from representative plots before and after a fire (Van Wagner 1968; Brown 1971; Raison et al. 1985; Ottmar 2014; Duff et al. 2017; Nolan et al. 2022; Price et al. 2022; Prichard et al. 2022). Field measurements of fuel consumption have been undertaken in the majority of fire-prone regions across various biomes and geographical locations worldwide (van Leeuwen et al. 2014). Australian forests, recognised for their susceptibility to wildfires (Luke and McArthur 1978; Gill and Zylstra 2005), have also been the focus of numerous field-based investigations over the last few decades (e.g. Raison et al. 1983; Cook 1994; Hurst et al. 1994; McCaw et al. 1997; Carter and Foster 2004; Rossiter-Rachor et al. 2008; Russell-Smith et al. 2009; Hollis et al. 2010, 2011; Meyer et al. 2012; Volkova and Weston 2013; Volkova et al. 2014; Possell et al. 2015; Volkova and Weston 2015; Jenkins et al. 2016; Hollis et al. 2018, 2019; Murphy et al. 2019; Volkova and Weston 2019; Nolan et al. 2022; Price et al. 2022). These studies have often indicated that fine fuels are generally combusted more completely than coarser components but observing the consumption of fine elements can be challenging compared with coarser fuels. Our review of field studies particularly conducted in southeast Australian eucalypt forests (listed in Supplementary Material Table S1) observes the variability of fine fuel consumption rates across different fuel layers and burn severity levels. Typically, lower layers (i.e. surface and near-surface) experience more significant impacts, and more severe fires consume more fuel. Fine fuels at the canopy layer, primarily comprising foliage and twigs from large tree crowns, are generally consumed less than 20% during a low-severity fire (Possell et al. 2015) but 75–100% during a high-severity fire (Volkova et al. 2014; Price et al. 2022). Conversely, a low-severity fire can consume ~45–50% of fine fuels at the elevated layer (i.e. fine components from shrubs, saplings and small trees) (Raison et al. 1983; Possell et al. 2015; Nolan et al. 2022). The consumption rate of near-surface fine fuels (such as herbs, grasses and mosses) and surface fine litter is even more substantial, reaching 60–70% during a low-severity fire and almost complete combustion during a high-severity fire (Hollis et al. 2011; Volkova and Weston 2015; Jenkins et al. 2016; Murphy et al. 2019; Price et al. 2022).
Field measurements of fuel loads, including consumption, can yield accurate results, but they are often difficult to implement owing to their time-consuming nature, logistical and labour requirements, and the complexities of measuring pre- and post-fire fuel loads over extensive areas (Ottmar 2014; van Leeuwen et al. 2014; Duff et al. 2017; Price et al. 2022). Most field studies quantifying fine fuel consumption in southeast Australia were therefore based on the observation of small samples (n <20), leading to high variation in measurement results (Watson 2012), and focused on one or a few specific vegetation types (Supplementary Table S1). Furthermore, the unpredictability of wildfires makes it obviously challenging to plan pre-fire measurements. Indeed, the majority of field measurements of fine fuel consumption were associated with prescribed burns (van Leeuwen et al. 2014). However, prescribed fires may not consistently serve as an accurate substitute for wildfires (Perona and Brebbia 2010; van Leeuwen et al. 2014). Some recent studies in southeast Australian forests, such as Volkova et al. (2014) and Price et al. (2022), have highlighted significant differences in fine fuel consumption between wildfires and prescribed burns, emphasising the need to assess fine fuel changes independently based on these fire types. The dependence on prescribed burns, which are normally at low–medium severity, has also prevented field research on how fuel consumption varies with fire severity or intensity.
Forest inventory data can serve as an alternative approach for quantifying fine fuels and their consumption as they offer regular measurements across broader geographical areas (Keane et al. 2015; Roxburgh et al. 2015; Boisramé et al. 2022). Accurate estimations of FFL across vertical strata, from surface to canopy, can be derived from inventory data using allometric models (Keith et al. 2000; Bi et al. 2004; Nolan et al. 2022). It is, however, a great challenge to directly derive fuel consumption information from forest inventory owing to the unavailability of spatially coincident pre- and post-fire measurements. A possible method for predicting the amount of fuel available for consumption is the Drought Factor model (McArthur 1967), which combines the Keetch–Byram Drought Index with the amount and timing of recent rainfall. Although the fuel moisture content significantly influences fuel consumption, a model based solely on fuel moisture, such as the Drought Factor, cannot adequately predict variations in fuel consumption. Studies indicated that the Drought Factor model may not accurately predict consumption during exceptionally dry or wet periods (McCarthy 2003; McCaw and Hollis 2019). Here, we integrate results from various peer-reviewed measurements of actual fine fuel loss (Supplementary Table S1) with fire severity to define common fuel combustion factors, which then allow us to estimate fuel consumption amounts from forest inventory data. Fuel consumption is highly dependent on fire severity, and within a vegetation type, these metrics are essentially different measures of the same phenomenon (Alexander 1982; Cruz et al. 2022; Price et al. 2022). Several studies have assessed fuel change by burn severity or intensity, and also utilised severity as a framework for explaining variation in fuel consumption rates (e.g. Davies et al. 2013; Volkova and Weston 2015; Collins et al. 2021; Bright et al. 2022; Price et al. 2022; Nolan et al. 2024). Importantly, fire severity can now be robustly mapped and validated using advanced remote sensing techniques (Collins et al. 2020; Gibson et al. 2020), making it feasible to estimate fuel consumption by establishing a relationship between combustion factors and severity (Price et al. 2022).
Along with estimating fuel consumption, predicting post-fire fine fuel accumulation is crucial for effective fire management and ecological restoration, aiding risk assessment, prescribed burn planning and fire hazard mitigation strategies (Volkova et al. 2014; Cawson et al. 2018; Duff et al. 2019; Burrows et al. 2023; Nolan et al. 2024). Various accumulation models have been developed and utilised to predict fuel recovery in Australian eucalypt forests, employing both linear and non-linear methods (Zazali et al. 2021). One commonly used approach is the Olson (1963) model and its modifications, originally designed for biomass but widely adapted for predicting fuel accumulation. The model is a negative-exponential method describing litter density using a differential equation based on physical processes. The Olson accumulation models are typically derived from either monitoring fuel loads post-fire over time or tracking litterfall and litter decomposition rates. Research conducted in eucalypt forests across southern and eastern Australia generally shows a rapid increase in surface fuel within a few years after fires, followed by a deceleration as the rate of fuel addition aligns with decomposition (Peet 1971; Birk and Simpson 1980; Walker 1981; Raison et al. 1983; Birk and Bridges 1989; O’Connell 1989; McCaw et al. 1996, 2002; Gould et al. 2011; Watson 2012; Dalgleish et al. 2015; Volkova et al. 2019; Neumann et al. 2021; Burrows et al. 2023; Nolan et al. 2024). Studies also demonstrate that the recovery duration varies across fuel layers; for instance, Volkova et al. (2019) found that surface litter can reach a steady-state level approximately 10–12 years after fires, while elevated fine fuels could surpass pre-fire values within a few years.
The original Olson (1963) model simply assumes the reduction of fuels to zero after fire, a scenario often divergent from reality (Nolan et al. 2022; Price et al. 2022). Consequently, it has been modified to incorporate the incomplete loss of fuels by accounting for initial post-fire residues (Birk and Simpson 1980; Raison et al. 1983; Fensham 1992). The determination of initial fuel loads is crucial as it directly influences the pace of fuel accumulation. Post-fire remaining fuel loads depend on fuel consumption and thus are largely connected to burn severity or intensity. In other words, within a vegetation type, fuel accumulation rates may vary depending on fire severity (Eskelson and Monleon 2018; Volkova et al. 2019; Nolan et al. 2024). Yet, operational fuel accumulation curves in Australia currently have not accounted for fire severity (Nolan et al. 2024). Furthermore, fuel recovery is influenced by vegetation characteristics such as tree density, canopy cover and understorey vegetation, which are shaped by local climate and geology (Loudermilk et al. 2022). Thus, incorporating vegetation structural attributes in assessing fuel accumulation can contribute to the enhancement of fire behaviour modelling.
In practice, Australian land managers typically use fuel types derived from spatial vegetation data to estimate landscape-scale fuel conditions. Models such as the Phoenix Rapid-Fire employ modified Olson accumulation curves to predict post-fire fuel accumulation rates based on fuel type and time since the last fire. Although such classification-based models offer simplicity, they may lack the precision required for accurate fuel load mapping (Volkova et al. 2018). These models often rely on generalised accumulation rates that may not fully capture variability across different forest landscapes and disturbance scenarios (Fire Prediction Services 2019). In this study, we introduce a complementary approach by integrating forest inventory data with fire history records, allowing for a more refined estimate of fuel accumulation and potentially enhancing current practices.
This study presents a proof-of-concept for estimating fine fuel consumption and accumulation across forest layers (surface, near-surface, elevated and canopy) using forest inventory data and fire severity records. The specific aims are to (i) estimate fine fuel consumption and post-fire residual amounts by relating observed pre-fire FFL to combustion factors corresponding to burn severity, aggregated from peer-reviewed measurements of actual fine fuel loss in southeast Australian forests (Supplementary Table S1); (ii) characterise patterns of fine fuel changes in Victorian eucalypt forests, southeast Australia, considering fire severity, fire type and forest cover; and (iii) develop fine fuel accumulation curves using a modified version of the Olson model with dynamic input parameters reflective of burn severity, fire type and forest canopy cover variations.
Methods
Study area
The study comprises public forests in Victoria, southeastern Australia (Fig. 1), covering approximately 7.9 million ha and encompassing diverse climatic zones and topography. The northwest region has low elevations (1–100 m) and a semi-arid climate with annual rainfall typically below 250 mm (average summer maximum temperatures from 20 to 35°C). Conversely, the east and southeast regions rise from sea level to over 2000 m, featuring a temperate climate with annual rainfall between 600 and 2200 mm, and average summer maximum temperatures from 15 to 30°C (from http://www.bom.gov.au/climate/maps, accessed 29 July 2024). This diversity creates a mosaic of vegetation types, including eucalypt forests, rainforests, shrublands and grasslands. Eucalypt forests dominate, and feature species such as red stringybark (E. macrorhyncha), brown stringybark (E. baxteri), messmate stringybark (E. obliqua), mountain ash (E. regnans), alpine ash (E. delegatensis) and river red gum (E. camaldulensis) (Bridgewater 1976). In Australia, under the National Forest Inventory, native forests are categorised into three crown cover classes: woodland (20–50%), open forest (50–80%) and closed forest (80–100%) (Walker and Hopkins 1990; Montreal Process Implementation Group for Australia and National Forest Inventory Steering Committee 2023). In Victoria, closed forests typically occur in wet or sheltered eastern and southeastern areas and support a lush understorey. By 2018, closed forests covered 233,000 ha – over half of Australia’s eucalypt closed forests – but this was drastically reduced to just 2000 ha by 2023 mainly due to devastating mega fires between 2018 and 2021 (Montreal Process Implementation Group for Australia and National Forest Inventory Steering Committee 2018, 2023). Eucalypt open forests, comprising approximately 4.6 million ha by 2023, often thrive across the state where annual rainfall exceeds 600 mm and soil phosphate levels are moderate. Woodland forests, covering 2.9 million ha by 2023, are common in drier regions or on poorer soils across the state. Over the last several decades, Victorian forests have endured significant fire impacts, with wildfires burning approximately 6.9 million ha and prescribed burns covering 2.0 million ha (DEECA 2023b). Major recent wildfires in 2003, 2007, 2009 and 2020 have profoundly changed the landscape of the state forests.
Study area across the public forests according to crown cover in 2023 in Victoria, southeast Australia (Australian Bureau of Agricultural and Resource Economics and Sciences 2023), with forest inventory plots and extent of large fires that have occurred in the last 20 years.
Forest inventory data and fine fuel estimation
Forest inventory data were gathered as part of the Victorian Forest Monitoring Program (VFMP), which has been implemented by the State’s land management agency (Department of Energy, Environment and Climate Action or DEECA) since 2011. The VFMP database comprises a network of 859 permanent ground circular plots (0.07 ha), randomly distributed across a systematic state-wide grid (Commissioner for Environmental Sustainability Victoria 2023). In each plot, various measurements were taken for large trees, elevated shrubs, understorey vegetation, coarse woody debris, surface litter and soil (Haywood et al. 2016).
In this study, we extracted all observations from 635 VFMP inventory plots (filtering out unmeasured and non-forest plots) measured between 2011 and 2023, with 524 plots (82.5%) being remeasured in two or three rounds. We computed five above ground fuel variables for each observation: four categorised by vertical layers: canopyFFL, elevatedFFL, near-surfaceFFL and surfaceFFL, and a fifth, totalFFL, representing the sum of these layers. Canopy fine fuels comprise foliage and twigs in the forest’s upper layer, mainly from large tree crowns (diameter at breast height or DBH >10 cm), whereas elevated fine fuels include small components from shrubs, saplings and trees (DBH <10 cm). Near-surface fine fuels consist of non-woody components from understorey vegetation, such as shrubs, herbs, grasses, mosses, lichens and suspended dead material, and surface litter consists of small dead materials, including leaves, bark and small branches (<2.5 cm) found on the forest floor. FFL at the plot level was estimated following the methods outlined in Nguyen et al. (2024).
Fire severity
Burn severity was derived from the Aggregated Fire Severity Classes dataset provided by DEECA (2023a). This dataset contains aggregated historical fire severity classifications for fires occurring in Victoria from 1998 onwards. The burn severity data were derived from various data sources and methods, including remote sensing information ranging from aerial photos to satellite images, as well as field-based investigations. Burn severity was stratified into three levels of low, medium and high, as described in Table 1, based on a Post-fire Burn Classification Schema created by the Victorian Government (Forest Fire Management Victoria 2016). It is worth noting that most severity attributions were accompanied by detailed descriptions of fire cover and impacts, as well as accuracy levels ranging from low to high. Only records with high classification accuracy were selected for further analysis.
| Burn severity | Definition | |
|---|---|---|
| Low | A predominately green forest canopy, with up to 50% scorched canopy crowns or consumed understorey vegetation | |
| Medium | A burn of variable intensity, with greater than 50% scorched forest canopy crowns or consumed understorey vegetation. Up to 50% of forest canopy crowns are consumed | |
| High | Greater than 50% of the forest canopy crowns and understorey vegetation are consumed |
Fine fuel consumption analysis
To estimate fuel consumption, we selected inventory observations that were taken within 3 years prior to the occurrence of a fire disturbance. This approach assumed that FFL in the selected plots remained constant during the period from the measurement date to the fire occurrence date (0–3 years). To ensure the validity of this assumption, we cross-referenced the data with a Landsat-derived forest disturbance dataset (Nguyen et al. 2018) to confirm that no significant disturbances (from fire or logging events) occurred (i) during the time interval between the measurement and fire occurrence, or (ii) within 10 years prior to the measurement. After filtering the data from the original number of observations, a total of 173 observations were identified for the analysis of fuel consumption, each observation corresponding to a single fire event.
It was not possible to directly estimate fuel consumption amounts from inventory observations owing to the lack of spatially coincident post-fire observations. Consequently, we used data from previous peer-reviewed field measurements conducted in southeast Australian eucalypt forests (summarised in Supplementary Table S1) to determine average fine fuel consumption rates corresponding to burn severities, expressed as a fraction of pre-fire values (Table 2). Although some reviewed studies do not spatially overlap with our study area, they focused on similar ecosystems, such as dry sclerophyll forests, making them suitable references. For each observation, the consumed amount was computed by multiplying the pre-fire value by the defined consumption rate, and then post-fire remaining FFL was determined as the difference between the pre-fire and estimated consumption FFL.
| Burn severity | Canopy | Elevated | Near-surface | Surface | |||||
|---|---|---|---|---|---|---|---|---|---|
| n | Mean (s.e.) | n | Mean (s.e.) | n | Mean (s.e.) | n | Mean (s.e.) | ||
| Low | 7 | 0.13 (0.04) | 9 | 0.45 (0.08) | 12 | 0.69 (0.10) | 18 | 0.64 (0.05) | |
| Medium | 2 | 0.20 (0.05) | 3 | 0.49 (0.13) | 3 | 0.80 (0.13) | 4 | 0.72 (0.10) | |
| High | 2 | 0.88 (0.12) | 2 | 0.78 (0.03) | 3 | 1.00 (0.00) | 4 | 0.93 (0.04) | |
Data are derived by aggregating results from previous fuel studies in southeast Australian eucalypt forests (Supplementary Table S1). n represents the number of samples and s.e. denotes the standard error.
Owing to the lack of immediate post-fire inventory observations, we were unable to validate the estimated consumption values. To demonstrate the validity of our analysis, we compared the estimated post-fire FFL (n = 173) against observed values derived from independent inventory observations (n = 23) collected within 1-year post-fire. Given the unpaired nature of the estimated and observed values (i.e. originating from disparate plots), we employed the non-parametric Mann–Whitney U test to conduct the statistical comparison.
Fine fuel consumption was independently evaluated by fire types (prescribed burn and wildfire) and by forest crown cover classes (closed, open and woodland forests). The number of inventory observations used for these analyses is presented in Table 3.
| Burn severity | Fire type | Forest crown cover | ||||
|---|---|---|---|---|---|---|
| Wildfire | Prescribed burn | Closed forest | Open forest | Woodland | ||
| Low | 33 | 35 | 28 | 41 | 12 | |
| Medium | 68 | 17 | 29 | 34 | 9 | |
| High | 20 | – | 8 | 6 | 6 | |
| Total | 121 | 52 | 65 | 81 | 27 | |
Forest crown cover was observed at the time of measurement.
For each stratum, the pre-fire FFL was calculated as the population mean of pre-fire observed values, regardless of burn severity. The mean values of consumed and remaining FFL were both estimated for each burn severity class using the following equation:
where: y̅s is the mean of fuel consumption or remaining associated with burn severity s; n is the sample size of burn severity s; y̅i is the consumption or remaining value of sample i; is the weight of burn severity s, where is the sample mean of pre-fire FFL of burn severity s, and is the population mean of pre-fire FFL. Standard errors of the mean values were calculated, and pairwise comparisons were also conducted using the Wilcoxon rank sum test.
Fine fuel accumulation prediction
As it is often the case that only a proportion of FFL is consumed during a fire event, as defined by burn severity, a modified version of Olson’s model was utilised for investigating the recovery trend of totalFFL and surfaceFFL in this study. The model was proposed by Raison et al. (1983) and takes the form:
where Xt is the FFL at time t since fire; Xss is the steady-state level of FFL in the absence of fire; k is the decomposition constant; and Xi is the initial FFL remaining after a fire, as combustion was not 100%.
A metric called the time to reach 95% (T0.95) of the steady-state level (Xss) was also determined to provide an estimate of the predicted fuel recovery time (Roxburgh et al. 2015) and is given by:
The modified Olson models were developed using 571 VFMP observations conducted after fires. Observations with a time since fire greater than 20 years and those with a high-severity prescribed burn were excluded from the analysis owing to limited sample sizes. Modified Olson curves were developed for each burn severity level, considering fire types and forest cover classes. The number of inventory observations associated with each group is shown in Table 4.
| Burn severity | Fire type | Forest crown cover | ||||
|---|---|---|---|---|---|---|
| Wildfire | Prescribed burn | Closed forest | Open forest | Woodland | ||
| Low | 152 | 151 | 120 | 111 | 40 | |
| Medium | 103 | 60 | 49 | 86 | 37 | |
| High | 105 | – | 26 | 64 | 38 | |
| Total | 360 | 211 | 195 | 261 | 115 | |
Forest crown cover was observed at the time of measurement.
From the selected observations, we defined t as the number of years since fires and Xt as the observed FFL. The steady-state level (Xss) and initial remaining (Xi) FFL were defined as pre-fire and post-fire remaining fuel values, respectively, derived from the previous steps. One exception was that a common value of Xss was calculated and used for both wildfire and prescribed burn models. This common value was determined by taking the mean of pre-wildfire and pre-prescribed burn values, weighted by proportions of the total burn areas (of approximately 6.9 million ha or 77.5% by wildfires and 2.0 million ha or 22.5% by prescribed burns) (DEECA 2023b). The defined parameters (i.e. t, Xt, Xss and Xi) were used to fit the modified Olson model and estimate the decomposition rate (k) along with its standard error. The accuracy of the derived fuel accumulation curves was evaluated by assessing model Residual Standard Errors (RSEs).
Results
Fine fuel consumption
Fig. 2 shows the overall patterns of estimated pre-fire, consumed and post-fire residual FFL across vertical layers in the study area. On average, a fire event could consume 12.9 t ha−1 of totalFFL (or over 60% of the initial 21.3 t ha−1), resulting in a residual amount of 8.4 t ha−1. The majority of surfaceFFL and near-surfaceFFL was consumed by fires, while the upper fuel layers were affected to a lesser extent (Fig. 2b).
The distribution of (a) pre-fire, (b) consumed, and (c) post-fire FFL by fuel layer (n = 173). The red dot is the mean value, black bar is the median value and blue box is the interquartile range.
There was no significant variance between the observed and estimated post-fire residual FFL, as indicated by P values exceeding 0.05 for each of the fuel layers (Fig. 3). Nevertheless, for near-surfaceFFL and surfaceFFL, the observed remaining values often exceeded our estimated values (Fig. 3c, d). This can be due to the fact that field observations often occurred several months after fires, giving FFL some time to recover to a certain level.
Plots burned in wildfires typically had higher pre-fire FFL, and this consequently resulted in greater fuel consumption and post-fire residuals, compared with prescribed burns (Table 5 and Fig. 4). TotalFFL experienced an average reduction of 13.6 t ha−1 (over 62% of pre-fire value) due to wildfires, irrespective of burn severity, in contrast to a decrease of 9.8 t ha−1 (54%) resulting from prescribed burns. A low or medium-severity wildfire typically combusted 8–10 t ha−1 surfaceFFL, whereas a high-severity wildfire had the potential to induce a loss exceeding 12 t ha−1 (Fig. 4d). Both wildfires and prescribed burns caused a significant combustion of near-surfaceFFL, with no discernible statistical difference (P > 0.05, Fig. 4c). Low and medium-severity wildfires consumed approximately half the available elevatedFFL, slightly surpassing prescribed burns, whereas high-severity wildfires could combust all elevatedFFL (Fig. 4b). Likewise, canopyFFL was predominantly impacted by high-severity wildfires, with an average loss of 4.3 t ha−1 (Fig. 4a). A considerable amount of FFL remained after both wildfires and prescribed burns, although this varied across burn severities and fuel layers (Table 5 and Supplementary Table S2).
| Fire type | Pre-fire | Consumption | Post-fire | |||
|---|---|---|---|---|---|---|
| Fuel layer | Mean and s.e. (t ha−1) | Mean and s.e. (t ha−1) | Proportion (%) | Mean and s.e. (t ha−1) | Proportion (%) | |
| Wildfire (n = 121) | ||||||
| CanopyFFL | 4.92 (0.34) | 1.37 (0.14) | 27.8 | 3.55 (0.31) | 72.2 | |
| ElevatedFFL | 1.43 (0.18) | 0.77 (0.10) | 53.8 | 0.66 (0.09) | 46.2 | |
| Near-surfaceFFL | 2.87 (0.24) | 2.29 (0.20) | 79.8 | 0.58 (0.06) | 20.2 | |
| SurfaceFFL | 13.37 (0.77) | 9.70 (0.56) | 72.9 | 3.67 (0.26) | 27.1 | |
| TotalFFL | 21.73 (0.96) | 13.59 (0.66) | 62.2 | 8.14 (0.51) | 37.8 | |
| Prescribed burn (n = 52) | ||||||
| CanopyFFL | 4.25 (0.34) | 0.64 (0.06) | 14.8 | 3.61 (0.29) | 85.2 | |
| ElevatedFFL | 1.01 (0.25) | 0.47 (0.12) | 46.5 | 0.54 (0.13) | 53.5 | |
| Near-surfaceFFL | 2.70 (0.33) | 1.96 (0.24) | 72.6 | 0.74 (0.10) | 27.4 | |
| SurfaceFFL | 11.68 (0.87) | 7.73 (0.56) | 66.6 | 3.95 (0.32) | 33.4 | |
| TotalFFL | 18.01 (1.05) | 9.75 (0.62) | 54.2 | 8.26 (0.49) | 45.8 | |
Average consumption values (t ha−1) for canopyFFL (a), elevatedFFL (b), near-surfaceFFL (c), and surfaceFFL (d) and totalFFL (e), grouped by fire type and burn severity, with standard errors shown as purple bars (see Supplementary Table S2 for numeric data). Significance levels were calculated from the Wilcoxon rank sum tests between wildfire and prescribed burn groups. Note that FFL plot scales are different.
Denser forests typically had higher pre-fire FFL, leading to greater consumption during fire events (Table 6 and Fig. 5). On average, fires consumed 14.78 t ha−1 (61% of pre-fire amount) of totalFFL in closed forests, leaving a residual of 9.36 t ha−1. Conversely, open and woodland forests experienced comparatively lesser totalFFL consumption, averaging 10.95 and 9.25 t ha−1, respectively. These patterns were generally consistent across the four vertical fuel layers (Table 6) and across burn severity levels (Fig. 5). Results from variance tests indicated a statistically significant difference in fuel consumption among forest cover classes for surfaceFFL and totalFFL (P < 0.05). However, this significance was not consistently observed in the upper layers, particularly between open and woodland forests (Fig. 5). Additionally, there was notable variation in fuel consumption and post-fire values across forest cover classes, as indicated by high standard errors (Table 6).
| Forest cover | Pre-fire | Consumption | Post-fire | |||
|---|---|---|---|---|---|---|
| Fuel layer | Mean and s.e. (t ha−1) | Mean and s.e. (t ha−1) | Proportion (%) | Mean and s.e. (t ha−1) | Proportion (%) | |
| Closed forest (n = 65) | ||||||
| CanopyFFL | 5.36 (0.42) | 1.33 (0.17) | 24.6 | 4.03 (0.37) | 75.4 | |
| ElevatedFFL | 1.82 (0.25) | 0.96 (0.14) | 53.0 | 0.86 (0.12) | 47.0 | |
| Near-surfaceFFL | 3.10 (0.29) | 2.46 (0.24) | 79.4 | 0.64 (0.07) | 20.6 | |
| SurfaceFFL | 15.41 (1.10) | 11.10 (0.82) | 72.1 | 4.31 (0.36) | 27.9 | |
| TotalFFL | 24.14 (1.27) | 14.78 (0.94) | 61.0 | 9.36 (0.61) | 39.0 | |
| Open forest (n = 81) | ||||||
| CanopyFFL | 4.34 (0.38) | 0.92 (0.11) | 21.2 | 3.42 (0.34) | 78.8 | |
| ElevatedFFL | 0.86 (0.13) | 0.43 (0.06) | 50.0 | 0.43 (0.07) | 50.0 | |
| Near-surfaceFFL | 2.81 (0.30) | 2.15 (0.23) | 76.5 | 0.66 (0.08) | 23.5 | |
| SurfaceFFL | 11.58 (0.70) | 7.96 (0.45) | 69.2 | 3.62 (0.27) | 30.8 | |
| TotalFFL | 18.84 (0.93) | 10.95 (0.53) | 58.0 | 7.89 (0.54) | 42.0 | |
| Woodland forest (n = 27) | ||||||
| CanopyFFL | 3.74 (0.38) | 1.13 (0.35) | 30.2 | 2.61 (0.35) | 69.8 | |
| ElevatedFFL | 0.92 (0.37) | 0.49 (0.18) | 53.3 | 0.43 (0.19) | 46.7 | |
| Near-surfaceFFL | 2.00 (0.55) | 1.56 (0.43) | 78.0 | 0.44 (0.12) | 22.0 | |
| SurfaceFFL | 9.36 (0.99) | 6.66 (0.69) | 71.2 | 2.70 (0.41) | 28.8 | |
| TotalFFL | 14.90 (1.19) | 9.25 (0.86) | 62.1 | 5.65 (0.69) | 37.9 | |
Average consumption values (t ha−1) for canopyFFL (a), elevatedFFL (b), near-surfaceFFL (c), and surfaceFFL (d) and totalFFL (e), grouped by forest cover and burn severity, with standard errors shown as purple bars (see Supplementary Table S3 for numeric data). Significance levels were calculated from the Wilcoxon rank sum tests on each pair of the cover classes. Note that FFL plot scales are different.
Fine fuel accumulation
The outcomes from the modified Olson models according to fire type are presented in Table 7 and visualised in Fig. 6. The modelled decomposition rates (k values) ranged from 0.18 to 0.41 for wildfires and from 0.31 to 0.38 for prescribed burns, generally decreasing with higher burn severity. Some k values were estimated with high standard errors, particularly those associated with low burn severity. Furthermore, most derived modified Olson models exhibited a high RSE, typically exceeding 5.0 t ha−1 (Table 7). SurfaceFFL was predicted to reach its steady-state level (12.99 t ha−1) approximately 6.4 years after a low-severity wildfire but over 16 years after a high-severity wildfire. Following prescribed burns, surfaceFFL could return to its pre-fire level in more than 8 years. Similar patterns were observed for totalFFL, though the estimated times were generally shorter.
| Fuel variable | Xss (t ha−1) | Wildfire (n = 360) | Prescribed burn (n = 211) | |||||||
|---|---|---|---|---|---|---|---|---|---|---|
| Severity | Xi (t ha−1) | k (s.e.) | RSE (t ha−1) | T 0.95 (year) | Xi (t ha−1) | k (s.e.) | RSE (t ha−1) | T 0.95 (year) | ||
| SurfaceFFL | ||||||||||
| Low | 12.99 | 4.81 | 0.40 (0.12) | 4.99 | 6.40 | 4.20 | 0.38 (0.09) | 5.44 | 6.80 | |
| Medium | 12.99 | 3.74 | 0.27 (0.05) | 5.38 | 9.67 | 3.27 | 0.32 (0.07) | 4.37 | 8.57 | |
| High | 12.99 | 0.94 | 0.18 (0.03) | 6.86 | 16.44 | |||||
| All | 12.99 | 3.67 | 0.22 (0.03) | 5.93 | 12.27 | 3.95 | 0.32 (0.06) | 6.23 | 8.36 | |
| TotalFFL | ||||||||||
| Low | 20.90 | 10.45 | 0.41 (0.15) | 7.78 | 5.61 | 8.66 | 0.38 (0.08) | 6.29 | 6.44 | |
| Medium | 20.90 | 8.57 | 0.37 (0.09) | 8.07 | 6.60 | 7.23 | 0.33 (0.06) | 5.41 | 7.79 | |
| High | 20.90 | 1.80 | 0.24 (0.05) | 10.2 | 12.13 | |||||
| All | 20.90 | 8.14 | 0.24 (0.03) | 9.22 | 10.59 | 8.26 | 0.31 (0.05) | 7.87 | 8.11 | |
Model scenarios were categorised by fire type and burn severity. The ‘All’ category indicates the average of all burn severity levels combined. Xss is the steady-state level of FFL and Xi is the post-fire remaining FFL. Modelled k values are with their s.e. RSE is the model Residual Standard Error and T0.95 time to reach 95% of Xss.
Fuel accumulation curves for surfaceFFL and totalFFL, grouped by fire type and burn severity. The dashed lines ending with dots indicate the time taken to reach 95% of steady-state levels (T0.95). The ‘All’ category indicates the average of all burn severity levels combined. Note that FFL plot scales are different.
Fine fuel accumulation trends varied across forest cover classes, as depicted in Fig. 7 (see Supplementary Table S4 for the model parameters). Following low or medium-severity fires, fine fuels in closed forests often required slightly longer to reach steady-state levels than in open and woodland forests. However, the estimated time following high-severity fires exceeded 15 years across all forest types, with no remarkable difference. Denser forests had higher steady-state levels, leading to higher fuel accumulation rates in closed and open forests compared with woodland forests. Specifically, ~15 years after a high-severity fire, surfaceFFL was predicted to increase from below 1 t ha−1 to an average of 15.4 t ha−1 in closed forests and 11.6 t ha−1 in open forests, compared with 9.4 t ha−1 in woodland forests. Denser forests exhibited higher modelled k values when burn severity levels were combined, but lower values when burn severities were considered individually (Supplementary Table S4). Similarly to the models by fire type, the models by forest cover also had high RSE values.
Fuel accumulation curves for surfaceFFL and totalFFL, grouped by forest cover and burn severity. The dashed lines ending with dots indicate the time taken to reach 95% of pre-fire FFL (T0.95). See Supplementary Table S4 for the model parameters. The ‘All’ category indicates the average of all burn severity levels combined.
Discussion
Given the time and resource-intensive nature of collecting fuel monitoring data, more efficient methods for assessing post-fire fuel responses are needed. This exploratory study proposed a proof-of-concept for estimating fine fuel consumption and accumulation following fires by incorporating forest inventory data and fire records. Leveraging existing inventory data not only streamlines the data acquisition process but also expands the spatial and temporal scope, thereby enhancing the capability to monitor fine fuel dynamics across heterogeneous landscapes. With a larger number of observations from inventory data, we were able to conduct an analysis on a much larger scale compared with direct field measurements (e.g. Hollis et al. 2011; Volkova and Weston 2013; Volkova et al. 2014; Possell et al. 2015; Volkova and Weston 2015). Specifically, we examined FFL changes across vertical layers in relation to burn severity, fire type and forest crown cover. The use of inventory data also enabled us to analyse both fuel consumption and post-fire accumulation processes. The proposed proof-of-concept therefore offers a practical and systematic approach for characterising fine fuel change, which can be applicable to other regions, considering the mature development of national forest inventory programs (Nesha et al. 2021) and advanced remote sensing-based fire severity mapping worldwide (Chuvieco et al. 2020; Kurbanov et al. 2022).
The fuel accumulation models developed in the present study can complement existing classification-based models, such as those in Phoenix Rapid-Fire, by providing a more dynamic and localised approach to fuel load prediction. Whereas classification-based models typically rely on generalised fuel-type classifications, our method incorporates variations in burn severity, fire type and forest cover, offering a more accurate estimate of fuel accumulation. By integrating these dynamic parameters into existing models, our approach can enhance the reliability of fuel hazard predictions, especially in heterogeneous landscapes where forest structure and fire impacts may vary significantly. Although this study used inventory plot data, the findings could be spatially extrapolated using remote sensing or interpolation techniques, enabling land managers to generate more precise fuel change maps over larger areas. Future research should explore integrating these methods into existing mapping frameworks to advance landscape-scale fire management.
The exploratory analysis in this study was facilitated by findings from previous peer-reviewed field observations that measured fine fuel loss during fires in southeast Australian eucalypt forests (Supplementary Table S1). Unfortunately, there was no guarantee that the collected data are representative of all forest types across the study area. Additionally, inconsistencies in how these studies defined and measured fuel data posed another challenge when harmonising the data. We acknowledge that any errors or uncertainties in the initial fuel measurements could impact the accuracy of our estimates, potentially affecting the estimation of carbon emissions from fires (Ottmar 2014). Furthermore, we were unable to directly validate the estimated consumption amounts owing to the lack of immediate post-fire observations spatially aligned with pre-fire inventory data. Nonetheless, results from the comparison of estimated and observed post-fire residual FFL (Fig. 3) suggest the validity of our analysis to some extent. Additionally, our findings are in agreement with other studies conducted in Victorian eucalypt forests (e.g. Volkova and Weston 2013; Volkova et al. 2014; Possell et al. 2015). Taken together, these results suggest it should be acceptable to use the estimated post-fire residuals as inputs for the modified Olson accumulation models. It is important to note that our results are estimated values, and thus further research is necessary to verify inventory-based fine fuel change analysis. Implementing ad hoc post-fire immediate observations at permanent inventory plots could facilitate verification processes, thereby augmenting the efficacy of forest inventory programs in monitoring forest fuel dynamics.
Another limitation of our approach is the assumption that the defined average consumption rates (Table 2) remain constant across different forest types and biophysical conditions. Although existing field studies provide valuable data, it is recognised that fuel consumption rates can vary depending on factors such as fuel moisture, climate conditions and fuel arrangement between and within forests (Hollis et al. 2011; Matthews et al. 2014; Yebra et al. 2018; Prichard et al. 2022). Therefore, it is important to further quantify actual consumption rates for different biophysical conditions, vegetation types and fire scenarios, which could be a focus for future studies.
Variation of fine fuel consumption by fire type and forest cover
Our results indicate that wildfires tended to exhibit higher pre-fire FFL and consumed higher amounts of fine fuel across all four layers as compared with prescribed burns (Table 5). It has been widely demonstrated that wildfires, typically triggered under more extreme conditions (Bowman et al. 2021a), have a greater potential to consume a larger amount of fuel. These findings align with previous research comparing fuel consumption between wildfires and prescribed burns not only in Australian eucalypt forests (van Leeuwen et al. 2014; Volkova et al. 2014; Price et al. 2022) but also in other ecosystems (Davies et al. 2016; Brodie et al. 2024). In Victorian eucalypt forests, a field-based study by Volkova et al. (2014) reported that a wildfire in southeast Victoria consumed an average of 12.55 t ha−1 of surface litter, whereas a fuel reduction burn consumed only 4.89 t ha−1 (compared with 10.53 and 6.54 t ha−1 respectively in our results, Table 5). Our results also show that prescribed burns primarily affect surface and near-surface fuels, aligning with fire management goals to reduce severe wildfires (Fernandes and Botelho 2003; Volkova et al. 2014; Hislop et al. 2020; Collins et al. 2023; Brodie et al. 2024), whereas medium and high-severity wildfires normally impact all fuel layers. Recognising these substantial differences is beneficial for facilitating mitigation activities, as future prescribed burning for fuel hazard reduction may need to target areas with higher surfaceFFL and near-surfaceFFL and/or drier conditions to facilitate greater surface fuel reduction.
The influence of vegetation structure on fuel consumption during fires has been previously studied (Davies et al. 2016; Levine et al. 2020; Loudermilk et al. 2022). These studies are in agreement with our study, which identified distinct patterns of fine fuel changes across different forest cover classes (Table 6). Denser forest structures, characterised by a more compact and diverse arrangement of trees and understories, typically have higher productivity, leading to increased FFL and greater fuel consumption during fires (Thomas et al. 2014; Volkova and Weston 2015). Quantifying these differences is essential for optimising fire mitigation strategies, such as targeting fuel reduction treatments, and for accurately contributing to the national carbon accounting system by predicting carbon emissions and sequestration potential. As expected, our results also indicated that a considerable amount of FFL remained post-fire (~38 and 46% of pre-fire totalFFL for wildfire and prescribed burns, respectively), depending on burn severity and pre-fire amounts. These estimations are crucial for evaluating the effectiveness of fuel reduction treatments and predicting post-fire fuel accumulation and future fire risks.
Dynamic post-fire fine fuel accumulation
Results from the modified Olson’s models indicate the influence of fire characteristics (burn severity and fire type) and forest density on post-fire fine fuel accumulation. Fire severities and types cause varied impacts on post-fire fuel residual amounts and biophysical factors driving recovery, and thus influencing fuel accumulation process. Our findings suggest that more severe fires, especially wildfires, generally extend the time needed for fuels to reach pre-fire levels, consistent with other research (Eskelson and Monleon 2018; Hislop et al. 2019; Etchells et al. 2020; Nolan et al. 2024). Nonetheless, surface fine litter was predicted to reach a steady-state level more than 8 years following prescribed burns, highlighting their efficiency in fire management. Fuel accumulation trends also varied across forest cover classes, with denser forests in favourable climates and soil conditions supporting higher accumulation rates (Mori et al. 2013; Thomas et al. 2014; Dalgleish et al. 2015; Newnham et al. 2017; Fry et al. 2018; Loudermilk et al. 2022). In addition, surfaceFFL tended to accumulate at slower rates when compared with totalFFL, which reached a steady-state more rapidly owing to quicker recovery of near-surface and understorey layers (McCaw et al. 1996). Previous studies have shown that elevated and lower canopy fuels can produce high levels, sometimes exceeding double the pre-fire fuel load several years after fires (Specht 1981; McCaw et al. 1996; Price and Gordon 2016; Volkova et al. 2019). Understanding these dynamic fuel accumulation scenarios can aid in optimising the timing and location of future fuel reduction burns.
Our modelled decomposition rates (k values) varied across different model scenarios but remained within a realistic range of values from 0.09 to 0.91, based on available literature (Newnham et al. 2017). We found that the litter decomposition rates were lower following wildfires compared with prescribed burns (k = 0.22 versus 0.32, respectively, Table 7). These results align with previous studies; for example, Roxburgh et al. (2015) reported k values of 0.319 for wildfire and 0.416 for prescribed burn based on a review of 20 studies conducted in Victoria, Australia. Decomposition rates also decreased with higher burn severity (Table 7 and Supplementary Table S4), as severe fires often destroy microbial communities, deplete essential nutrients and cause significant physical and chemical changes that inhibit decomposition. Our results indicate that, when burn severity was not considered, denser forests tended to have higher decomposition rates due to more favourable climate and soil conditions (Thomas et al. 2014; Newnham et al. 2017). However, this trend did not hold when burn severity was factored in, suggesting that pre- and post-fire fuel loads, as well as the degree of recovery or accumulated load, also play a significant role in determining decomposition rates.
Similarly to other studies predicting fuel accumulation (e.g. McCaw et al. 2002; Gould et al. 2011; Watson 2012; Dalgleish et al. 2015; Roxburgh et al. 2015; Volkova et al. 2019), our modified Olson models were derived with large standard errors related to both modelled parameters (i.e. k values) and fitted values. This can be attributed to several factors, including spatial heterogeneity, the absence of temporal data points and the assumption of input parameters (i.e. initial and steady-state values). Though efforts were made to reduce spatial heterogeneity by pooling observations by forest cover and burn severity, significant variations in forest structures and fuel characteristics across pooled plots may have resulted in effects beyond what the modified Olson model, as a single equation, could accommodate (Tolhurst and Kelly 2003). Furthermore, the Olson model critically relies on long-term fuel observations. As such, the absence of some temporal data points, especially in pools with fewer observations, could significantly impact model outcomes (McCaw et al. 2002; Zazali et al. 2021). The accuracy of our modified Olson models also depended on the precision of estimated initial and steady-state FFL values, as the key input parameters. We assumed that forests with the same cover class shared a common fine fuel steady-state amount, without considering burn severity. This simplification may not account for the actual variability of forest structures and fire impacts within a specific forest cover type, leading to high model errors. Incorporating more precise and comprehensive data can reduce uncertainties in Olson’s accumulation models, and thus it is essential to further measure actual fuel changes across different biophysical conditions, vegetation types and fire scenarios.
Conclusion
This study presents a proof-of-concept for estimating fine fuel consumption and accumulation by integrating forest inventory data and fire severity records across large areas, aiming to refine fuel dynamics estimates and enhance current practices. We analysed fine fuel changes across vertical strata, from surface to canopy, in southeast Australian eucalypt forests. Our results reveal that wildfires typically occurred in forests with higher FFL and thus consumed greater fuel amounts compared with prescribed burns. Distinct patterns of fine fuel consumption were also observed across different forest cover classes, with denser forests experiencing greater fuel loss during fires. Fine fuel accumulation was predicted using a modified Olson model with dynamic input parameters considering burn severity, fire type and forest cover. The results indicate that fire characteristics (severity and type) and forest density significantly impact the post-fire fuel accumulation process, with more severe fires generally leading to lower decomposition rates and longer time to reach pre-fire levels. Surface fine litter is predicted to reach a steady-state level in more than 8 years after low and medium-severity fires, but over 16 years following high-severity fires. Fuel accumulation trends also varied by forest cover classes, with denser forests in favourable biophysical conditions supporting higher accumulation rates. Although exploratory, our study offers a comprehensive understanding of fine fuel dynamics that is beneficial for optimising mitigation activities such as fuel reduction burns. Further research is needed to quantify actual changes in FFL under various biophysical conditions, vegetation types and fire scenarios, and to verify inventory-based fuel change analyses. The proposed framework offers a practical approach that is applicable to other regions, leveraging mature national forest inventory programs and advanced remote sensing-based fire severity mapping technologies.
Declaration of funding
This work was supported by the SmartSat Cooperative Research Centre (CRC), whose activities are funded by the Australian Government’s CRC Program.
Acknowledgements
We express our gratitude to the Victorian Forest Monitoring Program (VFMP) team at the Department of Energy, Environment and Climate Action (DEECA) for providing us access to the VFMP inventory database. We also thank the two reviewers for their great comments and suggestions on earlier drafts of this paper.
References
Alexander ME (1982) Calculating and interpreting forest fire intensities. Canadian Journal of Botany 60(4), 349-357.
| Crossref | Google Scholar |
Australian Bureau of Agricultural and Resource Economics and Sciences (2023) ‘Forests of Australia.’ (Australian Bureau of Agricultural and Resource Economics and Sciences: Canberra, Australia) Available at www.doi.org/10.25814/6cay-a361
Bi H, Turner J, Lambert M (2004) Additive biomass equations for native Eucalypt forest trees of temperate Australia. Trees 18(4), 467-479.
| Crossref | Google Scholar |
Birk EM, Bridges R (1989) Recurrent fires and fuel accumulation in even-aged blackbutt (Eucalyptus pilularis) forests. Forest Ecology and Management 29(1–2), 59-79.
| Crossref | Google Scholar |
Birk EM, Simpson R (1980) Steady state and the continuous input model of litter accumulation and decompostion in Australian Eucalypt forests. Ecology 61(3), 481-485.
| Crossref | Google Scholar |
Boisramé GFS, Brown TJ, Bachelet DM (2022) Trends in western USA fire fuels using historical data and modeling. Fire Ecology 18(1), 34.
| Crossref | Google Scholar |
Bowman D, Williamson GJ, Gibson RK, Bradstock RA, Keenan RJ (2021a) The severity and extent of the Australia 2019–20 Eucalyptus forest fires are not the legacy of forest management. Nature Ecology and Evolution 5(7), 1003-1010.
| Crossref | Google Scholar |
Bowman D, Williamson GJ, Price OF, Ndalila MN, Bradstock RA (2021b) Australian forests, megafires and the risk of dwindling carbon stocks. Plant, Cell & Environment 44(2), 347-355.
| Crossref | Google Scholar |
Bradstock RA, Boer MM, Cary GJ, Price OF, Williams RJ, Barrett D, Cook G, Gill AM, Hutley LBW, Keith H, Maier SW, Meyer M, Roxburgh SH, Russell-Smith J (2012) Modelling the potential for prescribed burning to mitigate carbon emissions from wildfires in fire-prone forests of Australia. International Journal of Wildland Fire 21(6), 629-639.
| Crossref | Google Scholar |
Bridgewater P (1976) Geographical elements in the flora of Victoria. Australian Journal of Botany 24(2), 249-264.
| Crossref | Google Scholar |
Bright BC, Hudak AT, McCarley TR, Spannuth A, Sánchez-López N, Ottmar RD, Soja AJ (2022) Multitemporal lidar captures heterogeneity in fuel loads and consumption on the Kaibab Plateau. Fire Ecology 18(1), 18.
| Crossref | Google Scholar | PubMed |
Brodie EG, Knapp EE, Brooks WR, Drury SA, Ritchie MW (2024) Forest thinning and prescribed burning treatments reduce wildfire severity and buffer the impacts of severe fire weather. Fire Ecology 20(1), 20.
| Crossref | Google Scholar |
Brown JK (1971) A planar intersect method for sampling fuel volume and surface area. Forest Science 17(1), 96-102.
| Google Scholar |
Burrows N, Wills A, Densmore V (2023) Fuel weight and understorey hazard dynamics in mature karri (Eucalyptus diversicolor) forests in southwest Western Australia. Australian Forestry 86(2), 68-81.
| Crossref | Google Scholar |
Carter MC, Foster CD (2004) Prescribed burning and productivity in southern pine forests: a review. Forest Ecology and Management 191(1–3), 93-109.
| Crossref | Google Scholar |
Cawson JG, Duff TJ, Swan MH, Penman TD (2018) Wildfire in wet sclerophyll forests: the interplay between disturbances and fuel dynamics. Ecosphere 9(5), 18.
| Crossref | Google Scholar |
Chuvieco E, Aguado I, Salas J, García M, Yebra M, Oliva P (2020) Satellite remote sensing contributions to wildland fire science and management. Current Forestry Reports 6(2), 81-96.
| Crossref | Google Scholar |
Collins L, McCarthy G, Mellor A, Newell G, Smith L (2020) Training data requirements for fire severity mapping using Landsat imagery and random forest. Remote Sensing of Environment 245, 14.
| Crossref | Google Scholar |
Collins L, Hunter A, McColl-Gausden S, Penman TD, Zylstra P (2021) The effect of antecedent fire severity on reburn severity and fuel structure in a resprouting Eucalypt forest in Victoria, Australia. Forests 12(4), 17.
| Crossref | Google Scholar |
Collins L, Trouvé R, Baker PJ, Cirulus B, Nitschke CR, Nolan RH, Smith L, Penman TD (2023) Fuel reduction burning reduces wildfire severity during extreme fire events in south-eastern Australia. Journal of Environmental Management 343(14)), 118171.
| Crossref | Google Scholar | PubMed |
Cook G (1994) The fate of nutrients during fires in a tropical savanna. Australian Journal of Ecology 19(4), 359-365.
| Crossref | Google Scholar |
Cruz MG, Alexander ME, Fernandes PM (2022) Evidence for lack of a fuel effect on forest and shrubland fire rates of spread under elevated fire danger conditions: implications for modelling and management. International Journal of Wildland Fire 31(5), 471-479.
| Crossref | Google Scholar |
Dalgleish SA, van Etten EJB, Stock WD, Knuckey C (2015) Fuel dynamics and vegetation recovery after fire in a semiarid Australian shrubland. International Journal of Wildland Fire 24(5), 613-623.
| Crossref | Google Scholar |
Davies GM, Gray A, Rein G, Legg CJ (2013) Peat consumption and carbon loss due to smouldering wildfire in a temperate peatland. Forest Ecology and Management 308, 169-177.
| Crossref | Google Scholar |
Davies GM, Domènech R, Gray A, Johnson PCD (2016) Vegetation structure and fire weather influence variation in burn severity and fuel consumption during peatland wildfires. Biogeosciences 13(2), 389-398.
| Crossref | Google Scholar |
DEECA (2023a) Aggregated fire severity classes from 1998 onward (https://datashare.maps.vic.gov.au/). Available at https://datashare.maps.vic.gov.au/search?md=ab51cc60-38f9-5a40-acfd-08b6a79514f9 [verified 14 February 2023]
DEECA (2023b) Fire history records of fires across Victoria (https://datashare.maps.vic.gov.au/). Available at https://discover.data.vic.gov.au/dataset/fire-history-records-of-fires-across-victoria1 [verified 14 February 2023]
Dodge M (1972) Forest fuel accumulation - A growing problem. Science 177(4044), 139-142.
| Crossref | Google Scholar | PubMed |
Duff T, Keane R, Penman T, Tolhurst K (2017) Revisiting wildland fire fuel quantification methods: the challenge of understanding a dynamic, biotic entity. Forests 8(9), 18.
| Crossref | Google Scholar |
Duff TJ, Cawson JG, Penman TD (2019) Determining burnability: predicting completion rates and coverage of prescribed burns for fuel management. Forest Ecology and Management 433, 431-440.
| Crossref | Google Scholar |
Enright NJ, Fontaine JB, Bowman DM, Bradstock RA, Williams RJ (2015) Interval squeeze: altered fire regimes and demographic responses interact to threaten woody species persistence as climate changes. Frontiers in Ecology and the Environment 13(5), 265-272.
| Crossref | Google Scholar |
Eskelson BN, Monleon VJ (2018) Post-fire surface fuel dynamics in California forests across three burn severity classes. International Journal of Wildland Fire 27(2), 114-124.
| Crossref | Google Scholar |
Etchells H, O’Donnell AJ, Lachlan McCaw W, Grierson PF (2020) Fire severity impacts on tree mortality and post-fire recruitment in tall Eucalypt forests of southwest Australia. Forest Ecology and Management 459, 117850.
| Crossref | Google Scholar |
Fensham RJ (1992) The management implications of fine fuel dynamics in bushlands surrounding Hobart, Tasmania. Journal of Environmental Management 36(4), 301-320.
| Crossref | Google Scholar |
Fernandes P, Luz A, Loureiro C, Ferreira-Godinho P, Botelho Hn (2006) Fuel modelling and fire hazard assessment based on data from the Portuguese National Forest Inventory. Forest Ecology and Management 234(1), S229.
| Crossref | Google Scholar |
Fernandes PM, Botelho HS (2003) A review of prescribed burning effectiveness in fire hazard reduction. International Journal of Wildland Fire 12(2), 117-128.
| Crossref | Google Scholar |
Forest Fire Management Victoria (2016) Post-fire burn classification. Available at https://emap.help.ffm.vic.gov.au/wp-content/uploads/sites/9/2020/04/Bushfire-Management-Manual-3.5.7.1-GUI-Post-Fire-Burn-Classification.pdf [verified 14 February 2023]
Fry DL, Stevens JT, Potter AT, Collins BM, Stephens SL (2018) Surface fuel accumulation and decomposition in old-growth pine-mixed conifer forests, northwestern Mexico. Fire Ecology 14(2), 15.
| Crossref | Google Scholar |
Gibson R, Danaher T, Hehir W, Collins L (2020) A remote sensing approach to mapping fire severity in south-eastern Australia using Sentinel 2 and Random Forest. Remote Sensing of Environment 240, 111702.
| Crossref | Google Scholar |
Gill AM, Zylstra P (2005) Flammability of Australian forests. Australian Forestry 68(2), 87-93.
| Crossref | Google Scholar |
Gould JS, McCaw WL, Phillip Cheney N (2011) Quantifying fine fuel dynamics and structure in dry Eucalypt forest (Eucalyptus marginata) in Western Australia for fire management. Forest Ecology and Management 262(3), 531-546.
| Crossref | Google Scholar |
Haywood A, Mellor A, Stone C (2016) A strategic forest inventory for public land in Victoria, Australia. Forest Ecology and Management 367, 86-96.
| Crossref | Google Scholar |
Hislop S, Jones S, Soto-Berelov M, Skidmore A, Haywood A, Nguyen TH (2019) High fire disturbance in forests leads to longer recovery, but varies by forest type. Remote Sensing in Ecology and Conservation 5(4), 376-388.
| Crossref | Google Scholar |
Hislop S, Stone C, Haywood A, Skidmore A (2020) The effectiveness of fuel reduction burning for wildfire mitigation in sclerophyll forests. Australian Forestry 83(4), 255-264.
| Crossref | Google Scholar |
Hollis J, Matthews S, Ottmar R, Prichard S, Slijepcevic A, Burrows N, Ward B, Tolhurst K, Anderson W, Gould J (2010) Testing woody fuel consumption models for application in Australian southern Eucalypt forest fires. Forest Ecology and Management 260(6), 948-964.
| Crossref | Google Scholar |
Hollis JJ, Matthews S, Anderson WR, Cruz MG, Burrows ND (2011) Behind the flaming zone: predicting woody fuel consumption in eucalypt forest fires in southern Australia. Forest Ecology and Management 261(11), 2049-2067.
| Crossref | Google Scholar |
Hollis JJ, Gould JS, Cruz MG, Lachlan McCaw W (2015) Framework for an Australian fuel classification to support bushfire management. Australian Forestry 78(1), 1-17.
| Crossref | Google Scholar |
Hollis JJ, McCaw WL, Cruz MG (2018) The effect of woody fuel characteristics on fuel ignition and consumption: a case study from a Eucalypt forest in south-west Western Australia. International Journal of Wildland Fire 27(5), 363-375.
| Crossref | Google Scholar |
Hollis JJ, McCaw WL, Cruz MG (2019) The effect of woody fuel characteristics on fuel ignition and consumption: a case study from a Eucalypt forest in south-west Western Australia. International Journal of Wildland Fire 27(5), 363-375.
| Crossref | Google Scholar |
Hurst DF, Griffith DW, Carras JN, Williams DJ, Fraser PJ (1994) Measurements of trace gases emitted by Australian savanna fires during the 1990 dry season. Journal of Atmospheric Chemistry 18, 33-56.
| Crossref | Google Scholar |
Jenkins ME, Bell TL, Poon LF, Aponte C, Adams MA (2016) Production of pyrogenic carbon during planned fires in forests of East Gippsland, Victoria. Forest Ecology and Management 373, 9-16.
| Crossref | Google Scholar |
Keane RE, Loehman R, Clark J, Smithwick EA, Miller C (2015) Exploring interactions among multiple disturbance agents in forest landscapes: simulating effects of fire, beetles, and disease under climate change. In ‘Simulation modeling of forest landscape disturbances’. (Eds A Perera, B Sturtevant, L Buse) pp. 201–231. (Springer: Cham, Switzerland)
Kurbanov E, Vorobev O, Lezhnin S, Sha J, Wang J, Li X, Cole J, Dergunov D, Wang Y (2022) Remote sensing of forest burnt area, burn severity, and post-fire recovery: a review. Remote Sensing 14(19), 35.
| Crossref | Google Scholar |
Levine JI, Collins BM, York RA, Foster DE, Fry DL, Stephens SL (2020) Forest stand and site characteristics influence fuel consumption in repeat prescribed burns. International Journal of Wildland Fire 29(2), 148-159.
| Crossref | Google Scholar |
Liu Y, Stanturf J, Goodrick S (2010) Trends in global wildfire potential in a changing climate. Forest Ecology and Management 259(4), 685-697.
| Crossref | Google Scholar |
Loehman RA, Reinhardt E, Riley KL (2014) Wildland fire emissions, carbon, and climate: seeing the forest and the trees – A cross-scale assessment of wildfire and carbon dynamics in fire-prone, forested ecosystems. Forest Ecology and Management 317, 9-19.
| Crossref | Google Scholar |
Loudermilk EL, O’Brien JJ, Goodrick SL, Linn RR, Skowronski NS, Hiers JK (2022) Vegetation’s influence on fire behavior goes beyond just being fuel. Fire Ecology 18(1), 9.
| Crossref | Google Scholar |
Matthews S, Sullivan AL, Watson P, Williams RJ (2012) Climate change, fuel and fire behaviour in a Eucalypt forest. Global Change Biology 18(10), 3212-3223.
| Crossref | Google Scholar | PubMed |
Matthews S, Loveday S, Hurley R, Roxburgh S (2014) Modelling fuel moisture across the landscape for fire management. In ‘4th Fire Behavior and Fuels Conference’. pp. 411–412. (St Petersburg, Russia). https://www.iawfonline.org/wp-content/uploads/2018/02/4th_Fuels_Conference_Proceedings_USA-Russia_updated_5.28.2015.pdf
McCaw WL, Hollis J (2019) Evaluating fuel and soil moisture indices as a guide to fuel availability in tall open forests. In ‘Proceedings of International Association of Wildland Fire 6th Fuels and Fire Behaviour Conference’, 2019, Sydney, Australia. Available at https://library.dbca.wa.gov.au/static/FullTextFiles/072626.pdf
McCaw WL, Neal JE, Smith RH (1996) Fuel accumulation following prescribed burning in young even-aged stands of Karri (Eucalyptus diversicolor). Australian Forestry 59(4), 171-177.
| Crossref | Google Scholar |
McCaw WL, Smith R, Neal J (1997) Prescribed burning of thinning slash in regrowth stands of Karri (Eucalyptus diversicolor). 1. Fire characteristics, fuel consumption and tree damage. International Journal of Wildland Fire 7(1), 29-40.
| Crossref | Google Scholar |
McCaw WL, Neal JE, Smith RH (2002) Stand characteristics and fuel accumulation in a sequence of even-aged Karri (Eucalyptus diversicolor) stands in south-west Western Australia. Forest Ecology and Management 158(1), 263-271.
| Crossref | Google Scholar |
McCaw WL, Gould JS, Phillip Cheney N, Ellis PFM, Anderson WR (2012) Changes in behaviour of fire in dry Eucalypt forest as fuel increases with age. Forest Ecology and Management 271, 170-181.
| Crossref | Google Scholar |
Meyer C, Cook GD, Reisen F, Smith T, Tattaris M, Russell‐Smith J, Maier S, Yates C, Wooster M (2012) Direct measurements of the seasonality of emission factors from savanna fires in northern Australia. Journal of Geophysical Research: Atmospheres 117(D20), D20305.
| Crossref | Google Scholar |
Miezïte LE, Ameztegui A, De Cáceres M, Coll L, Morán-Ordóñez A, Vega-García C, Rodrigues M (2022) Trajectories of wildfire behavior under climate change. Can forest management mitigate the increasing hazard? Journal of Environmental Management 322, 116134.
| Crossref | Google Scholar | PubMed |
Mori S, Itoh A, Nanami S, Tan S, Chong L, Yamakura T (2013) Effect of wood density and water permeability on wood decomposition rates of 32 Bornean rainforest trees. Journal of Plant Ecology 7(4), 356-363.
| Crossref | Google Scholar |
Murphy BP, Prior LD, Cochrane MA, Williamson GJ, Bowman D (2019) Biomass consumption by surface fires across Earth’s most fire prone continent. Global Change Biology 25(1), 254-268.
| Crossref | Google Scholar | PubMed |
Nesha MK, Herold M, De Sy V, Duchelle AE, Martius C, Branthomme A, Garzuglia M, Jonsson O, Pekkarinen A (2021) An assessment of data sources, data quality and changes in national forest monitoring capacities in the Global Forest Resources Assessment 2005–2020. Environmental Research Letters 16(5), 054029.
| Crossref | Google Scholar |
Neumann M, Turner J, Lewis T, McCaw L, Cook G, Adams MA (2021) Dynamics of necromass in woody Australian ecosystems. Ecosphere 12(8), 15.
| Crossref | Google Scholar |
Nguyen TH, Jones SD, Soto-Berelov M, Haywood A, Hislop S (2018) A spatial and temporal analysis of forest dynamics using Landsat time-series. Remote Sensing of Environment 217, 461-475.
| Crossref | Google Scholar |
Nguyen TH, Jones S, Reinke KJ, Soto-Berelov M (2024) Estimating fine fuel loads in Eucalypt forests using forest inventory data and a modelling approach. Forest Ecology and Management 561, 10.
| Crossref | Google Scholar |
Nolan RH, Boer MM, Collins L, Resco de Dios V, Clarke H, Jenkins M, Kenny B, Bradstock RA (2020) Causes and consequences of eastern Australia’s 2019–20 season of mega-fires. Global Change Biology 26(3), 1039-1041.
| Crossref | Google Scholar | PubMed |
Nolan RH, Price OF, Samson SA, Jenkins ME, Rahmani S, Boer MM (2022) Framework for assessing live fine fuel loads and biomass consumption during fire. Forest Ecology and Management 504, 9.
| Crossref | Google Scholar |
Nolan RH, Gibson RK, Cirulis B, Holyland B, Samson SA, Jenkins M, Penman T, Boer MM (2024) Incorporating burn heterogeneity with fuel load estimates may improve fire behaviour predictions in south-east Australian Eucalypt forest. International Journal of Wildland Fire 33(3), 14.
| Crossref | Google Scholar |
O’Connell AM (1989) Nutrient accumulation in and release from the litter layer of Karri (Eucalyptus diversicolor) forests of southwestern Australia. Forest Ecology and Management 26(2), 95-111.
| Crossref | Google Scholar |
Olson JS (1963) Energy storage and the balance of producers and decomposers in ecological systems. Ecology 44(2), 322-331.
| Crossref | Google Scholar |
Ottmar RD (2014) Wildland fire emissions, carbon, and climate: modeling fuel consumption. Forest Ecology and Management 317, 41-50.
| Crossref | Google Scholar |
Peet GB (1971) Litter accumulation in Jarrah and Karri forests. Australian Forestry 35(4), 258-262.
| Crossref | Google Scholar |
Possell M, Jenkins M, Bell TL, Adams MA (2015) Emissions from prescribed fires in temperate forest in south-east Australia: implications for carbon accounting. Biogeosciences 12(1), 257-268.
| Crossref | Google Scholar |
Price OF, Gordon CE (2016) The potential for LiDAR technology to map fire fuel hazard over large areas of Australian forest. Journal of Environmental Management 181, 663-673.
| Crossref | Google Scholar | PubMed |
Price OH, Nolan RH, Samson SA (2022) Fuel consumption rates in resprouting Eucalypt forest during hazard reduction burns, cultural burns and wildfires. Forest Ecology and Management 505, 10.
| Crossref | Google Scholar |
Prichard SJ, Rowell EM, Hudak AT, Keane RE, Loudermilk EL, Lutes DC, Ottmar RD, Chappell LM, Hall JA, Hornsby BS (2022) Fuels and consumption. In ‘Wildland fire smoke in the United States: a scientific assessment’. (Eds DL Peterson, SM McCaffrey, T Patel-Weynand) pp. 11–49. (Springer International Publishing: Cham, Switzerland) 10.1007/978-3-030-87045-4_2
Raison R, Woods P, Khanna P (1983) Dynamics of fine fuels in recurrently burnt Eucalypt forests. Australian Forestry 46(4), 294-302.
| Crossref | Google Scholar |
Raison RJ, Khanna PK, Woods PV (1985) Transfer of elements to the atmosphere during low-intensity prescribed fires in three Australian subalpine Eucalypt forests. Canadian Journal of Forest Research 15(4), 657-664.
| Crossref | Google Scholar |
Rossiter-Rachor N, Setterfield S, Douglas M, Hutley L, Cook G (2008) Andropogon gayanus (gamba grass) invasion increases fire-mediated nitrogen losses in the tropical savannas of northern Australia. Ecosystems 11, 77-88.
| Crossref | Google Scholar |
Russell-Smith J, Murphy BP, Meyer CM, Cook GD, Maier S, Edwards AC, Schatz J, Brocklehurst P (2009) Improving estimates of savanna burning emissions for greenhouse accounting in northern Australia: limitations, challenges, applications. International Journal of Wildland Fire 18(1), 1-18.
| Crossref | Google Scholar |
Thomas PB, Watson PJ, Bradstock RA, Penman TD, Price OF (2014) Modelling surface fine fuel dynamics across climate gradients in Eucalypt forests of south-eastern Australia. Ecography 37(9), 827-837.
| Crossref | Google Scholar |
van Leeuwen TT, van der Werf GR, Hoffmann AA, Detmers RG, Rücker G, French NHF, Archibald S, Carvalho Jr JA, Cook GD, de Groot WJ, Hély C, Kasischke ES, Kloster S, McCarty JL, Pettinari ML, Savadogo P, Alvarado EC, Boschetti L, Manuri S, Meyer CP, Siegert F, Trollope LA, Trollope WSW (2014) Biomass burning fuel consumption rates: a field measurement database. Biogeosciences 11(24), 7305-7329.
| Crossref | Google Scholar |
Van Wagner C (1968) The line intersect method in forest fuel sampling. Forest Science 14(1), 20-26.
| Google Scholar |
Volkova L, Weston C (2013) Redistribution and emission of forest carbon by planned burning in Eucalyptus obliqua (L. Hérit.) forest of south-eastern Australia. Forest Ecology and Management 304, 383-390.
| Crossref | Google Scholar |
Volkova L, Weston CJ (2015) Carbon loss from planned fires in southeastern Australian dry Eucalyptus forests. Forest Ecology and Management 336, 91-98.
| Crossref | Google Scholar |
Volkova L, Weston CJ (2019) Effect of thinning and burning fuel reduction treatments on forest carbon and bushfire fuel hazard in Eucalyptus sieberi forests of south-eastern Australia. Science of The Total Environment 694, 133708.
| Crossref | Google Scholar | PubMed |
Volkova L, Meyer CPM, Murphy S, Fairman T, Reisen F, Weston C (2014) Fuel reduction burning mitigates wildfire effects on forest carbon and greenhouse gas emission. International Journal of Wildland Fire 23(6), 771-780.
| Crossref | Google Scholar |
Volkova L, Meyer CPM, Haverd V, Weston CJ (2018) A data–model fusion methodology for mapping bushfire fuels for smoke emissions forecasting in forested landscapes of south-eastern Australia. Journal of Environmental Management 222, 21-29.
| Crossref | Google Scholar | PubMed |
Volkova L, Weiss Aparicio AG, Weston CJ (2019) Fire intensity effects on post-fire fuel recovery in Eucalyptus open forests of south-eastern Australia. Science of The Total Environment 670, 328-336.
| Crossref | Google Scholar | PubMed |
Whitman E, Parisien M-A, Thompson DK, Flannigan MD (2019) Short-interval wildfire and drought overwhelm boreal forest resilience. Scientific Reports 9(1), 18796.
| Crossref | Google Scholar | PubMed |
Yebra M, Quan X, Riaño D, Rozas Larraondo P, van Dijk AIJM, Cary GJ (2018) A fuel moisture content and flammability monitoring methodology for continental Australia based on optical remote sensing. Remote Sensing of Environment 212, 260-272.
| Crossref | Google Scholar |
Zazali HH, Towers IN, Sharples JJ (2021) A critical review of fuel accumulation models used in Australian fire management. International Journal of Wildland Fire 30(1), 42-56.
| Crossref | Google Scholar |
