Northern Australia’s vast tropical savannas are highly productive yet fire-prone ecosystems, and it has been suggested that reductions in the frequency of high-intensity fires – through the use of prescribed burning under mild fire weather conditions – could substantially increase the amount of carbon they store, especially in tree biomass. We evaluated this hypothesis by analysing data from 236 long-term vegetation and fire monitoring plots throughout the savannas of northwestern Australia (WA and NT), spanning up to 24 years and a rainfall gradient of 600 to 1700 mm per annum. We used this analysis to develop a process-based model of savanna tree populations to simulate the long-term effects of prescribed burning on live tree biomass. We found that high-intensity fires cause a significant decrease in tree growth and an increase in tree mortality, especially in the smallest and largest trees. Our process-based model shows that by using prescribed burning to reduce the frequency of high-intensity fires, mean tree biomass is likely to increase significantly, around 5–10%, over the long-term. We also show that the moderation of fire regimes is also likely to deliver improvements in key attributes of faunal habitat, especially the availability of large tree hollows. If fire-driven changes in tree biomass, and associated carbon storage, can be incorporated into greenhouse gas accounting methodologies, the viability of savanna fire management projects aimed at generating carbon credits is likely to increase significantly.

Over recent decades, much savanna research has focused on the processes that allow the co-existence of trees and grass. Tree–tree interactions in savannas have received little attention. Northern Australia's mesic savannas are amongst the most fire-prone on Earth, and are dominated by fire-tolerant eucalypts. In recent years, there has been widespread adoption of prescribed burning to generate carbon credits by reducing greenhouse gas emissions, however the long-term effects of prescribed burning on savanna structure are poorly understood. Here we present results from a long-term fire experiment from Northern Australia. This experiment consists of replicated fire treatments in a eucalypt-dominated savanna. We examined the spatial patterns of trees and degree of potential competition among trees. Non-eucalypt trees exhibited a strong clumped spatial pattern, up to a height of 5 m, whereas eucalypts, and non-eucalypts above 5 m height had a random spatial distribution. However, our analysis of size- and distance-dependent competition reveals evidence for strong competitive interactions among trees in the unburnt and low-intensity fire treatments. Young non-eucalypt trees were aggregated in clumps indicating dispersal-dependant establishment of non-eucalypts rather than density-dependent mortality. We also extend our analyses to intra- and inter specific competition of trees at the seedling stage, where grass competition is important. However, our results show the importance of competition and fire to better understand the tree dynamics in the north Australian tropical savannas.

Fire is a major driver of spatial patterns within landscapes and is commonly used as a management tool to promote biodiversity. However, fire is not the only factor affecting biodiversity, with spatial patterns also created by topography, climate, soils, vegetation and habitat structure acting at different scales. There is poor understanding of how spatial patterns in these environmental factors affect species diversity, especially at spatial scales relevant to land management. We investigated how changes in spatial patterns caused by fire and other environmental variables affected mammal species richness. We also tested whether these relationships varied with spatial scale. We sampled ground-dwelling mammals at 187 sites spanning a range of fire histories within the Otway Ranges, Victoria. Bayesian networks were used to assess the influence of spatial patterns in topography, climate, soil, time since fire, vegetation and habitat structure on mammal richness at multiple spatial scales. Mammal richness was most strongly influenced by the configuration of time since fire in the landscape, increasing in areas with both more fire edges and greater contrast in adjacent fire ages. Configuration of elevation and aridity in the landscape had the next greatest effect on species richness. Generally, the observed effects on species richness did not vary greatly with changing scales, although the structure of the Bayesian networks indicated the interactions between these variables do change with scale. These results suggest land managers should aim to use prescribed burning to create areas with high age contrasts and this can be effective at multiple spatial scales.

Australia’s forests and woodlands have been fragmented by agriculture, urbanisation and forestry, and in many fragmented systems, fire is used as a management tool to reduce wildfire risk and enhance biodiversity. Currently, fire managers seek to promote biodiversity by maintaining a range of fire-age classes to suit the requirements of a range of species. However, this approach assumes mammals respond directly to time since fire, without consideration of other factors acting on mammal communities. In this project I investigate the effects of abiotic factors (e.g., climate, topography), landscape structure (the composition and configuration of land-use types), fire history and vegetation structure on ground-dwelling mammal communities in the heathy woodland of western Victoria. I am using data from live trapping and camera trapping to investigate ground-dwelling mammal community composition and species richness, and will present preliminary results from the first of two survey seasons. Results will identify the key drivers of mammal community composition, and help managers better understand how planned burning affects mammal communities.

Planned fire is increasingly recognised as an important tool in conservation. However, fire-regime responses of ground-dwelling mammals are often unclear, and changing fire regimes currently threaten many native species, such as the nationally-endangered southern brown bandicoot (Isoodon obesulus obesulus). Further, the fire regime may interact with other ecological features such as landscape structure (the composition and configuration of land cover types), and responses of ground-dwelling mammals to multiple interacting factors are unknown. We aim to investigate these interactions by surveying mammals in the Mount Lofty Ranges, South Australia, where a fine-scale fire mosaic is embedded within a fragmented landscape. Here, we present preliminary results from camera trapping in 2019, including mammal community composition and southern brown bandicoot distribution in relation to past fire and landscape structure. Later, we will use a landscape genetics approach and fire regime simulation to examine southern brown bandicoot habitat suitability and connectivity in present and possible future landscapes. This project will help identify interacting effects of landscape structure and fire regimes on ground-dwelling mammals and inform ecologically sensitive fire management planning.

Understanding the possible ecological causes of plant rarity and commonness is crucial for conservation in an era of high plant extinction rates. Environmental disturbances such as fire can play a key role in shaping plant abundance and distribution patterns in fire-prone vegetation communities. However, there are few studies examining the contextual drivers of plant rarity and commonness across different vegetation communities under variable fire regimes. As different vegetation communities exhibit variable responses to fire, environmental processes may be better predictors of plant rarity rather than biological traits of a species. We completed vegetation surveys at 86 sites in coastal dry sclerophyll vegetation, in Booderee National Park, Jervis Bay, where fire history records have been maintained for 55 years. We tested for associations between fire regime variables (seasonality, intensity, severity and fire interval length) with patterns of plant rarity and commonness across forest, woodland, and heath vegetation communities. Our findings have important implications for both the research and management of fire regimes and understanding how multiple aspects of fire regimes can influence plant rarity.

Global fire regimes are increasingly altered by land use changes and climate change. Despite decent mechanistic understanding of how inappropriate fire regimes can lead to ecological change and community collapse, there is little guidance for conservation managers on the “ideal” fire characteristics for many ecosystems and species. We used a systematic review methodology to examine the subjects and methods of research suitable for informing community-level fire management. Recommendations rarely target fire regime characteristics other than fire frequency. There is a bias towards vertebrates over invertebrates, and major knowledge gaps include fungal, soil biota and belowground vegetation fire responses. Compared to plant-focused studies, mammal studies produced quantified recommendations twice as often, and mammal and bird studies identified threatened species four times as often. Only 7% of studies assessed both flora and fauna. Only 23% assessed interactions with other disturbance types or variables, despite increasing acknowledgement that fire interacts with many environmental and anthropogenic processes. The most utilised research methods rarely produced quantified management recommendations for biodiversity conservation. Using listed NSW Threatened Ecological Communities (TECs) as a case study, we assessed current evidence for their fire prescriptions. 76% of TEC recovery documents list inappropriate fire regimes as a Key Threatening Process, yet few state whether the risk is from too much fire or not enough, let alone provide quantified guidance on tolerable fire thresholds. There is an urgent need to fill fire ecology knowledge gaps in fire-prone and fire-dependent ecosystems, ensuring that ecological communities are not lost through ineffective or inadequate management.

The severity of a fire is partially influenced by the severity of the fire before it, but how? The leading argument is that the structure of vegetation is altered by the previous fire, depending on its severity. There is some evidence of this, but nothing definitive for the eucalypt forests of south-east Australia. This study aimed to address that lack of evidence. We hypothesised that the shrub cover in dry sclerophyll forests would be higher after a high severity fire than after low severity. We also hypothesised that increased shrub cover would lead to a greater chance of high severity in a subsequent fire. We tested these hypotheses using vegetation surveys through the Greater Blue Mountains surrounding Sydney. 71 field sites were selected to provide a range of times since fire and both high and low severity. A space for time substitution was used to test the effects of shrub structure on subsequent fire severity. Preliminary results suggest that fire severity can explain around 50% of the deviance in shrub cover, with low severity and unburnt sites having similar cover scores and high severity sites having increased shrub cover. This positive relationship supports our hypothesis and is most likely due to mass germination in the soil seedbank due to the heat of a severe fire. Low severity fires are generally less intense, so would not cause such an event. High severity fire drives changes in the fuel structure, which then influences the severity of following fires.

Animals can shape their environment through ecosystem engineering, and this can occur through many different mechanisms. Key among these is through moderation of abiotic processes such as fire. Yet, with the exception of a few well-studied examples, the effects of animals on fire regimes have been long overlooked. We review evidence of the effects of animals on fire, and show how animals can affect fire behaviour by modifying the amount, structure or condition of fuel, or more rarely, through effects on other fire drivers such as landscape structure and fire ignition patterns. Some effects are readily observed and quantified; for example, digging animals can reduce surface fuel biomass through litter burial and turnover, while grazing can reduce fuel amount through consuming grass biomass. More subtle effects on fuel properties, such as animal tracks creating fine-scale fuel heterogeneity, or herbivore damage reducing fuel moisture, are also common, but the consequences of such changes for fire behaviour remain largely unquantified. When accumulated across time, space and animal taxa, even these more subtle effects will likely have important implications for fire behaviour and therefore fire regimes. A combination of manipulative experiments, landscape studies and multi-scale fire models will be necessary to fully understand the consequences of changes in animal populations for landscape fire.

Much research on post-fire resprouting has been performed in woody plants. We examined resprouting after fire in 52 Australian native grasses and found that: 1) C4 grasses survive fire better than C3 grasses and, 2) survival increased with leaf dry matter content (LDMC). The importance of the photosynthetic pathway in post-fire survival can be explained because C4 plants can fix more carbon under warm and sunny environments than C3 grasses. One unanswered question of these results is that if the capacity to efficiently fix carbon was the clue for fire response in grasses, we would expect that photosynthetic pathway (and LDMC) would be a good predictor not only for surviving, but also for the strength of resprouting; this was not the case. Here, we propose an alternative hypotheses for the increased post-fire resprouting in grasses: post-fire resprouting depends on the location of the bud bank. Specifically, we predict that bud location explains resprouting in grasses better than photosynthetic pathway and LDMC. We found survival and resprouting depends not only in having a C4 photosynthesis pathway, but also (and more importantly) on bud location. Post-fire resprouting response was lower for grasses with stolons, intermediate for those that resprout from the crown, and highest for the species with rhizomes. The results support the hypothesis that the location of the bud bank is an important factor determining fire response in grasses. Determining the position of buds is necessary to better understand the underlying mechanisms of response to disturbance in plants.

Temperate grassy ecosystems in southern Australia have been decimated since European occupation, and the few remaining remnants are mostly degraded and fragmented. Changes in disturbance regimes had dramatic and rapid effects—the removal of fire and introduction of livestock grazing resulted in the local extinction of grazing sensitive and fire-dependent species. Hence, many of southern Australia’s grassy ecosystems have remained unburnt for decades, and have been heavily degraded by other disturbances in the interim (e.g., livestock grazing). Recently, fire has been re-introduced into long unburnt landscapes to restore both biodiversity and Indigenous connection to country. This provides a rare opportunity to examine if the return of cultural burning can restore plant diversity. We explored plant community responses to the re-introduction of cultural fire in central Victoria. We sampled burnt and unburnt plots in four grassy ecosystems in central Victoria. Plots were monitored immediately after the fire and at monthly intervals thereafter. At each monitoring time, data were collected on bare ground, litter, vegetation cover, structure (using the golf ball method), soil moisture and light availability to determine the main drivers of plant diversity. Plant species richness and composition were sampled in spring. Resources (soil moisture and light) and structural variables (openness, biomass) differed in burnt and unburnt plots through time. However, species richness did not differ between burnt and unburnt plots. The lack of plant responses to the re-introduction of fire suggests ecosystems are relatively stable and fire-dependent species may be lost and require active restoration through seed addition.

Epicormic and basal resprouting promote tree survival and persistence in fire-prone regions worldwide, and these resprouting traits characterise the widespread dry eucalypt forests of south-eastern Australia. As such, these forests are often assumed to be reliable carbon stores because the dominant eucalypts are thought to be rarely outrightly killed by wildfire. However, little is known about limits to resprouting effectiveness when severe wildfires increase in frequency, and similarly there has been little examination of how carbon stocks in these forests may change after short-interval wildfires. We examined the effects of one and two high-severity wildfires within six years on relationships between tree size (stem diameter) and resprouting (epicormic and/or basal), and on carbon stocks in fire-tolerant eucalypt forests in Victoria. Our study indicated that short-interval wildfires increased tree ‘escape size’ – the stem size required to survive fire - and eroded resprouting success, particularly of middle-sized trees, which were too large for basal resprouting but too small for epicormic recovery. After controlling for site aridity, total carbon stocks and aboveground biomass were significantly decreased at single-burn sites, and further reduced at double-burned sites, relative to unburned sites. On average, the percentage of aboveground carbon stocks in live biomass decreased from 76% at unburned sites, to 62% at single-burned, and 47% at double-burned sites, contributing to increases in standing deadwood carbon stocks from 11% at unburned sites to 42% at double-burned sites. These findings portend structural, demographic and carbon storage challenges for even the most fire-tolerant forests under emerging fire regimes.