Wildfire is a fundamental disturbance process in many ecological communities, and is critical in maintaining the structure of some plant communities. In the past century, changes in global land use practices have led to changes in fire regimes that have radically altered the composition of many plant communities. As the severe biodiversity impacts of inappropriate fire management regimes are recognized, attempts are being made to manage fires within a more ‘natural’ regime. In this aim, the focus has typically been on determining the fire regime to which the community has adapted. Here we take a subtly different approach and focus on the probability of a patch being burnt. We hypothesize that competing sympatric taxa from different plant functional groups are able to coexist due to the stochasticity of the fire regime, which creates opportunities in both time and space that are exploited differentially by each group. We exploit this situation to find the fire probability at which three sympatric grasses, from different functional groups, are able to co-exist. We do this by parameterizing a spatio-temporal simulation model with the life-history strategies of the three species and then search for the fire frequency and scale at which they are able to coexist when in competition. The simulation gives a clear result that these species only coexist across a very narrow range of fire probabilities centred at 0.2. Conversely, fire scale was found only to be important at very large scales. Our work demonstrates the efficacy of using competing sympatric species with different regeneration niches to determine the probability of fire in any given patch. Estimating this probability allows us to construct an expected historical distribution of fire return intervals for the community; a critical resource for managing fire-driven biodiversity in the face of a growing carbon economy and ongoing climate change.
Citation: Armstrong G, Phillips B (2012) Fire History from Life-History: Determining the Fire Regime that a Plant Community Is Adapted Using Life-Histories. PLoS ONE 7(2): e31544. https://doi.org/10.1371/journal.pone.0031544
Editor: Wendy A. Peer, Purdue University, United States of America
Received: January 23, 2011; Accepted: January 12, 2012; Published: February 21, 2012
Copyright: © 2012 Armstrong, Phillips. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: Financial support was provided by Australian Wildlife Conservancy, Australian Research Council, Holsworth Wildlife Research Endowment, and the Northern Territory Research and Innovation Fund. The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Wildfire is a fundamental disturbance process in many ecological communities across the globe , . Indeed, the fire regime is often a critical factor maintaining the structure of plant communities ,  and changes in a fire regime can radically alter community composition , . Increasingly, however, we seek to manage fire; either for protection of human lives and assets, to maintain biodiversity, or as part of a growing carbon economy , . As the negative biodiversity impacts of inappropriate fire management regimes are increasingly recognized ,  attempts are being made to manage fires within a more ‘natural’ regime. However, it is rarely apparent what, exactly, this means. From an ecological perspective, the critical question in wildfire management is what the fire regime in an area should be: how often, when, and how patchily should an area be burnt so as to maintain or promote current biodiversity values?
Ecologists have sought to answer this question using a variety of techniques that attempt to determine the fire regime to which the community is adapted. One important technique has been the examination of historical records. In Australia, for example, evidence that long-standing, pre-European fire management practices contributed to the demographic structure and geographic range of many vegetation communities , ,  has encouraged research into management that emulates the fire regimes imposed by indigenous Australians. While arguably successful in some cases, this approach has not proven broadly applicable, partly due to a lack of resolution in regard to the identity of vegetation communities which came under human influence , and partly because of a patchy historical record and loss of indigenous knowledge across most of Australia since European settlement , . Analytical techniques have also been employed. For example satellite imagery has become a powerful tool for mapping fire patterns across landscapes . For inferring long-term historical pattern however, it fails simply because satellite imagery does not go far enough back in time. Charcoal deposits can provide longer term data on fire frequency, however they are limited to sites with appropriate depositional conditions and only provide very course frequency resolution inappropriate for ecosystems with very frequent fires, such as savanna which can have a biannual fire return interval . Dendrochronology, the age-dating of trees using tree rings and fire scars also provides quantitative methods for determining fire regimes – but again, the method is not broadly applicable because not all fire prone habitats have tree species with regular cambial growth in which seasonal events are recorded.
A similar technique, applicable to non-woody species, uses knowledge of plant responses to fire (e.g., if the plant is an obligate seeder that only germinates after fire) and data on plant abundance through time to estimate the chronological sequence of fire events over recent history –. This enables the most recent fire intervals to be determined, but again falls far short of making inference about the long-term fire regime to which the community is adapted and is of no use when there is significant germination during the fire interval. , .
How then, do we go about determining the fire regime to which the community is adapted? The metrics of interest are twofold: fire scale (or patchiness), and fire frequency (or fire return interval). Current understanding is, that on a landscape level, fire can promote diversity by providing spatial and temporal variability in habitat across a landscape, creating a range of niches' through time and space . For example, fast growing and maturing species outcompete slower growing species in areas when the fire return interval occurs before the latter mature and set seed . Similarly, species with greater dispersal abilities are at an advantage when fires are uniformly large . Indeed, species can be classified into different functional groups according to their fire regeneration or persistence niche –. Competing species from different functional groups are able to coexist in a single landscape precisely because they utilise the different spatial and temporal opportunities created when fire removes, or at least inhibits, the growth of competitors. The patchiness of fire, thus, creates high species turnover at the local scale while simultaneously enabling stability at the meta-community scale . The long-term fire regime (frequency and patchiness), then, is that which has allowed the long-term persistence of sympatric taxa with differing regeneration niches.
Thus, if we have data on the life history of several plant species that compete for space in the same landscape, we can ask, using an appropriate model, what is the fire regime that allows these taxa to co-exist over time? If these taxa vary in their regeneration niche, we should rapidly converge on a small list of possible fire regimes that will allow long-term persistence. Here we demonstrate this approach, and its utility for inferring the fire regime to which species have co-adapted, using a simple spatio-temporal simulation model. We parameterized the model with life-history data from three congeneric grasses (Triodia spp.), each representing a differing functional group, and then search for the fire regime that allows these three species to co-exist over 200 generations. Rather than modelling fire frequency as a fixed return interval, we treat it as a probabilistic event that affects a patch (and a patch's neighbours if fire scale is large). Treating fire as a probabilistic event allows a distribution of fire return intervals to emerge naturally across the landscape, and it is this distribution which allows competing species to coexist. The models show that the co-existence of these three species can only occur over a very narrow range of fire probabilities but a much broader range of fire scales. This narrow range of probabilities must, therefore, be a key attribute of the long-term historical fire regime of the area.
Each of the three species had an individual response to fire probability and scale as a result of differences in their life histories (Fig. 1). The obligate seeder, T. bitextura, persisted only at fire probabilities below 0.8 at small scales reducing gradually to approximately 0.4 as scale increased. There is only a small plateau of 100% patch occupancy at the lowest fire frequencies and scales. Triodia sp., an obligate seeder, persists at a far greater range of frequencies and scales than T. bitextura with total patch occupancy at all frequencies when scale is small. Both population size and occupancy fall away as scale increases and population size is negatively correlated with fire frequency. Conversely, the facultative resprouter, T. epactia persists across all frequencies above 0.2 at all scales. Below this frequency persistence is marginal, with very low population size, at medium patch occupancy.
Comparison of the individual species responses to fire with those when competition was introduced between pairs (Fig. 2) enabled interspecific interactions to be resolved. T. bitextura showed no differences under both competition scenarios indicating the demography of this species is totally determined by fire rather than competition (i.e., it is the competitively superior species). T. bitextura inhibited both population size and patch occupancy of T. epactia and population size of Triodia sp. at low fire probabilities.
The large range of fire probabilities and scales at which species were able to persist without competition was significantly reduced with competition between pairs and further still when there was competition among all three species (Fig. 3). Visualizing only those cells in which all three species were present (Fig. 4) showed a ridge in mean population diversity centred over fire probability equalling 0.2. The spread, or variability, of the surface about the mean was greater at smaller fire scales, becoming very narrow as the scale increased. When fire is treated as probabilistic, the time since fire over a large number of patches can be described by a negative binomial distribution. When we set the probability of failure in the negative binomial to 0.2 we see a landscape where many patches burnt within the last 5 years (as expected), but with a large number of patches remaining unburnt for considerably longer periods (up to 20 years; Fig. 5).
Our results show that each of the three Triodia species has a very different demographic response to fire probability and scale (Fig. 1). This is in line with our classification of these three species into different functional groups, and reflects differences in their regeneration and persistence niches. The obligate seeder, Triodia sp., is unable to persist at fire probabilities above 0.2 when fires are large (Fig. 1), indicating that both regeneration time and dispersal ability prevent the landscape persistence of this species under these conditions. Triodia bitextura, the obligate resprouter, has a similar response, but appears generally less resilient due to its reduced (i.e., local) dispersal ability. The facultative resprouter, T. epactia on the other hand, has a very different response to the other species (Fig. 1), being able to persist at all fire frequencies and scales, although only at very low densities when fire is absent. The low population size of T. epactia in the absence of fire is probably due to population senescence: the species lacks clonal reproduction, and fire is needed to remove adults and allow seedling germination.
When three-way competition was introduced into the model (Fig. 3) the demographic response of T. bitextura was relatively unchanged compared with the no competition scenario, while there were significant changes for both Triodia sp. and T. epactia. Thus, T. bitextura appears to be the superior competitor, suggesting its demography is totally determined by the fire regime. The other two species, however, are inhibited by competition with T. bitextura and/or each other at low fire frequencies. Importantly, when the response surfaces for all species were combined to show only those regions of fire regime parameter space in which all species coexist (Fig. 4) we see a steep-sided ridge centred over a fire probability of 0.2 with greater spread at lower fire scales compared to larger scales. This indicates fire probability has a greater influence on the demographic structure of this community than fire scale. Given that these three species, from different functional groups, are sympatric over an area of approximately 3000 km2 in a very old and stable landscape, this narrow range of fire probabilities (around 0.2) is that to which these taxa are most likely to have co-adapted in this landscape.
Long term co-persistence amongst these three species in our model is only possible due to the spatio-temporal structure that fire provides in the landscape with fire probability being the critical factor providing this structure. An average fire probability of 0.2 allows a wide range of fire return intervals to exist across the landscape (Fig. 5). Importantly, this result does not concur with previous fire scar analyses in this region, using remote sensed data , in which fire return intervals are typically less than 5 years. Thus, the recent fire frequency across this landscape appears inconsistent with the distribution of fire frequencies we elucidate here and which has enabled the long term persistence of this community. This is a significant observation given the widespread decline of mammals and fire sensitive plant species across northern Australia in recent decades . Inappropriate fire regimes are suspected of playing a major role in these declines, and our results strongly suggest that in our study area, at least, “inappropriate” means “too often”, and perhaps, “too large”.
This study clearly has significant implications for fire management in the study area, where landscape scale fire management is practiced through aerial burning and yearly mapping . Although the consensus among managers in the area is that prior to active management fire occurred too frequently and in too large a patch, our work provides quantitative estimates of how frequent is too frequent, and provides a distribution of fire ages towards which the landscape should move if management is to be effective. Without such quantitative guidance fire management simply remains intuitive. In the same way, our approach is also likely to prove useful for determining the average fire frequency in other plant communities. In making such inference, care will, however, need to be taken to find species from different functional groups that could reasonably be considered to compete for space due to disturbance by fire and not some other niche dimension. This is unlikely to be a problem except in the most changed environments where few original species remain.
Generally, our work demonstrates the efficacy of using competing sympatric species from different regeneration niches' to determine the fire frequency to which they have co-adapted. The approach is, basically, to search for a fire scale and probability that create the correct array of opportunities in space and time for persistence of species from all functional groups. It is this very spatio-temporal habitat structure that enables species from different functional groups to coexist , , , so we can use the fact of coexistence to infer the historical fire regime. We cannot definitively rule out other environmental factors acting as further niche dimensions contributing to the coexistence, or otherwise, of the species in this study, however the evidence that the regeneration niche dominates other potential niche axes is strong in the present case . Further, the species models without interspecific competition demonstrate persistence is limited within definable fire regimes adding weight to the above argument. While we considered fire to be probabilistic in the ‘landscape’, discounting other potential environmental influences on fire behaviour, we feel that any increase in ‘landscape’ heterogeneity would increase the variance of the range of fire probabilities but this would not influence the mean. While this method relies on several independent lines of evidence (life histories of individual species and molecular data regarding population dynamics) it cannot be validated with other data in the way that a charcoal sequence may be correlated to dendrochronological data. However, we feel all methodologies have merit for particular situations and the results presented here are suitable for non-woody ecosystems with intense fire regimes.
The importance of being able to determine the fire regime to which a plant community has adapted cannot be overstated given that the fire regime is a key driver of biodiversity in many systems. This fact, together with the increasing recognition that the global distribution of fire may be altered with climate change –, and the increasing demand to use fire as a carbon emission mitigation strategy, means that careful management of fire is in need of quantification. Here we provide a useful way to create fire management benchmarks for biodiversity.
Study area and species
The three species considered in this study, T. bitextura, T. epactia and Triodia sp nov., have widely differing total ranges but grow in sympatry across an 18×140 km region of the southern Kimberley between 17°30′ and 17°40′S. The climate is influenced by the northern monsoon with a mean of 750 mm rainfall occurring in the summer months from October to April and mean annual temperatures ranging from 32C in the summer to 28C in the winter. The landscape is dominated by the highly eroded sandstone of the King Leopold Ranges with basaltic soils and cobbles in the valleys. Fire is the dominant disturbance in this environment with 30% of the landscape burnt annually and a quantitative analysis of satellite data at MWS, covering the last decade, showed a fire return interval of every 2–3 years (Steve Murphy, 2009 unpubl. data). Active fire management to control the season, extent and spatial complexity of fires has recently been introduced  following a period of uncontrolled extensive wildfires.
The three target species were chosen for the study because they are dominants in this savanna grassland community and are highly competitive ecological generalists which are distinguished by different post-fire regeneration niches, not by their physiological responses to substrate, moisture or community composition , . Each species forms monospecific stands, uncorrelated with edaphic variables, with sharp stand boundaries with little to no intermixing of individuals were different species grow adjacent to each other. Molecular evidence infers they have existed in sympatry in this environment dating back to at least the late Pleistocene . The three study species all rely to differing degrees on fire for regeneration [Armstrong, 2009 unpubl. data]. Triodia bitextura is an obligate resprouter that regenerates after fire through a combination of resprouting adults and production of ramets. A small number of stolons are produced between fires which also develop into free living ramets. Virtually no seed is produced and no evidence has been obtained that seedlings contribute to overall regeneration in this species. Adults of T. epactia, a facultative resprouter, also resprout after fire, but rely on the germination of seedlings for population growth instead of ramets. Triodia sp. an obligate seeder, does not resprout after fire or produce ramets, relying totally on seed germination for population growth. Significant germination only occurs in T. epactia and Triodia sp. after the removal of adult plants during a fire.
The spatially explicit model predicts the persistence of each species individually and in competition with the other species across all fire probabilities and a range of fire scales. Stage structured life-history matrices were constructed based on the sexual (seed) and asexual (resprouting and stolons) fecundity of each species and the transition probabilities of each stage class. Life cycle diagrams are shown for each species and the matrices and initial vectors are shown in the Appendix.
The model ‘landscape’, with a carrying capacity of 1 million individuals, approximately represents a 1.6 km2 landscape containing 256, 100×100 m cells (‘patches’), each with a carrying capacity of 3906 individuals, based loosely on empirical evidence that a 5×5 m quadrat contains a mean number of 10 individuals. For each model run, across successive proportional fire scales,1, 2, 4, 8 and 16, with 16 being a fire across a 16×16 cell area, which is the size of the entire space, the fire probability ranged from 0–1 at 0.1 increments. Fire scale was therefore initially small but doubled with each successive run until it equaled the entire ‘landscape’.
Each model run was initialized with a vector of the pre-fire, adult condition appropriate for each species. A conservative transition probability of 0.9 was set for all adult stages and ramets to account for observed high survival rates in all species. Triodia epactia and T. bitextura also have a probability of 0.9 of living indefinitely beyond their climax stages of 3 and 5 years respectively. The number of populations which actually survive beyond 5 years is expected to be small due to the probability of being burnt before this time. Populations of Triodia sp. on the other hand are observed to senesce at approximately 4 years old  and without fire or dispersal would be expected die out.
We used standard matrix population projection, in which the projection matrix is multiplied by the initial vector, to find the expected population size in each subsequent generation. The resultant population vector was however, modified by three additional (spatial) processes: fire, dispersal and competition. Fire occurred as a stochastic disturbance, random in space. The probability of a fire in any given grid cell in any given generation was determined as the product of two probabilities: the probability of an ignition event, and the conditional probability that an ignition would result in a fire. This latter probability was conditional on whether or not the grid cell had experienced a fire in the previous generation: being set to one where no fire had been experienced, or 0.3 if a fire had been experienced. This is to account for the time needed to establish enough fuel to carry a second fire and is derived by analysing the number of consecutive fires which occurred over 10 years across 100 random points using Landsat imagery (Steve Murphy 2009 unpubl. data). Whether or not a fire actually occurs in a grid cell was then determined by a draw from a binomial distribution for the given frequency.
Dispersal was also modeled as a stochastic process, was scaled by population size and differed between vegetative and seeder species. Vegetative species could only disperse into neighbouring cells (because the amount of seed produced is so negligible that it was considered functionally insignificant), and did so with a probability that scaled with their grid cell population size relative to grid cell carrying capacity. Dispersal distances for Triodia seeds are unknown, however seeds and seedlings are observed to fall and grow around the base of adult plants  suggesting the majority of seeds travel less than a metre. This would concur with observations for other grasses . However, with maximum annual wind speeds occurring in the summer months (www.bom.gov.au/climate/data/2009) when plants are seeding, and the possibility of animal assisted dispersal, there is a possibility of longer dispersal for at least some seeds. It has been demonstrated that for a range of herbaceous species median seed dispersal distance increased linearly with increasing wind speed but exponentially above the 90th percentile of the seed shadow . This means that with increasing wind speeds the majority of seeds will still fall within metres of the adult plant but 10% will disperse much greater distances. For these reasons dispersal was modelled as global in the seeder species (i.e., a dispersing seed could land anywhere on the landscape) but with a small probability of dispersing long distances. In the model, dispersed seeds germinate in the patch in which they land, without the need for fire-induced germination.
The final stochastic process in our model was competition. Competition assumes that there is limited space in which to grow, and acts to curb population growth in our model. Competition was modeled as a sampling procedure: a survival lottery. When the expected population size of adult plants in a grid cell exceeded the per cell carrying capacity, the final population size of adult plants was determined by a draw from a multinomial distribution with n being set to the grid cell carrying capacity and probabilities scaled with the expected number of adults in each age class. Thus, species with a relatively high number of expected recruits into the adult population would, in the absence of disturbance, come to dominate a grid cell.
The model was run in three configurations; 1) each species individually to determine where each species is able to persist across the range of fire variables, 2) with each combination of species pairs to determine which species compete with each other and 3) with all species combined to determine at which fire frequency and scale the community is at maximum diversity (Shannon-Wiener index) reflecting the greatest probability for coexistence. The mean population size and patch occupancy was recorded for the last 5 of the 200 generations of each run and the mean of these taken over 20 replicate runs. The code was written in R (R Development Core Team 2009) , shown in the appendix, and the Shannon-Wiener index determined with the vegan package (http://cran.r-project.org/, http://vegan.r-forge.r-project.org/). All the results were plotted against fire scales and frequencies on a 3D spline plot using Tibco S-plus . The frequency of patches of particular ages in the ‘landscape’ i.e. time since last fire (measured in years) was plotted for a fire frequency of 0.2 (Fig. 5).
Steve Murphy provided fire scar analysis for MWS. The paper was improved by comments from Stephen Garnett, Lynda Marshall and James Smith. Lynda Marshall and Dave Marshall assisted with the figures.
Conceived and designed the experiments: GA BP. Performed the experiments: GA BP. Analyzed the data: GA BP. Wrote the paper: GA BP. Conceived the project, collected the data, built the life-history matrices, assisted with designing the model and analyzed results: GA. Assisted with designing the model and wrote the code and analyzed results: BP.
- 1. Bond WJ, Woodward FI, Midgley GF (2005) The global distribution of ecosystems in a world without fire. New Phytologist 165: 525–538.WJ BondFI WoodwardGF Midgley2005The global distribution of ecosystems in a world without fire.New Phytologist165525538
- 2. Bond WJ, Keeley JE (2005) Fire as a global ‘herbivore’: the ecology and evolution of flammable ecosystems. Trends in Ecology & Evolution 20: 387–394.WJ BondJE Keeley2005Fire as a global ‘herbivore’: the ecology and evolution of flammable ecosystems.Trends in Ecology & Evolution20387394
- 3. Thuiller W, Slingsby JA, Privett SDJ, Cowling RM (2007) Stochastic Species Turnover and Stable Coexistence in a Species-Rich, Fire-Prone Plant Community. PLoS ONE 2(9): e938.W. ThuillerJA SlingsbySDJ PrivettRM Cowling2007Stochastic Species Turnover and Stable Coexistence in a Species-Rich, Fire-Prone Plant Community.PLoS ONE29e938
- 4. Lehmann CER, Prior LD, Williams RJ, Bowman DMJS (2008) Spatio-temporal trends in tree cover of a tropical mesic savanna are driven by landscape disturbance. Journal of Applied Ecology 45: 1304–1311.CER LehmannLD PriorRJ WilliamsDMJS Bowman2008Spatio-temporal trends in tree cover of a tropical mesic savanna are driven by landscape disturbance.Journal of Applied Ecology4513041311
- 5. Crowley G, Garnett S, Shephard S (2009) Impact of storm-burning on Melaleuca viridiflora invasion of grasslands and grassy woodlands on Cape York Peninsula, Australia. Austral Ecology 34: 196–209.G. CrowleyS. GarnettS. Shephard2009Impact of storm-burning on Melaleuca viridiflora invasion of grasslands and grassy woodlands on Cape York Peninsula, Australia.Austral Ecology34196209
- 6. Russell-Smith J, Yates CP, Brock C, Westcott VC (2009) Fire regimes and interval-sensitive vegetation in semiarid Gregory National Park, northern Australia. Australian Journal of Botany 58: 300–317.J. Russell-SmithCP YatesC. BrockVC Westcott2009Fire regimes and interval-sensitive vegetation in semiarid Gregory National Park, northern Australia.Australian Journal of Botany58300317
- 7. Flannigan MD, Krawchuk MA, de Groot WJ, Wotton BM, Gowman LM (2009) Implications of changing climate for global wildland fire. International Journal of Wildland Fire 18: 483–507.MD FlanniganMA KrawchukWJ de GrootBM WottonLM Gowman2009Implications of changing climate for global wildland fire.International Journal of Wildland Fire18483507
- 8. Strassburg BBN, Kelly A, Balmford A, Davies RG, Gibbs HK, et al. (2010) Global congruence of carbon storage and biodiversity in terrestrial ecosystems. Conservation Letters 3: 98–105.BBN StrassburgA. KellyA. BalmfordRG DaviesHK Gibbs2010Global congruence of carbon storage and biodiversity in terrestrial ecosystems.Conservation Letters398105
- 9. Keith DA (1996) Fire-driven extinction of plant populations: a synthesis of theory and review of evidence from Australian vegetation. Proceedings of the Linnean Society of NSW 116: 37–78.DA Keith1996Fire-driven extinction of plant populations: a synthesis of theory and review of evidence from Australian vegetation.Proceedings of the Linnean Society of NSW1163778
- 10. Bunk S (2004) World on Fire. PLoS Biol 2(2): e54.S. Bunk2004World on Fire.PLoS Biol22e54
- 11. Russell-Smith J, Lucas D, Gapindi M, Gunbunuka B, Kapirigi N, et al. (1997) Aboriginal Resource Utilization and Fire Management Practice in Western Arnhem Land, Monsoonal Northern Australia: Notes for Prehistory, Lessons for the Future. Human Ecology 25: 159–195.J. Russell-SmithD. LucasM. GapindiB. GunbunukaN. Kapirigi1997Aboriginal Resource Utilization and Fire Management Practice in Western Arnhem Land, Monsoonal Northern Australia: Notes for Prehistory, Lessons for the Future.Human Ecology25159195
- 12. Bowman DMJS (1998) The impact of Aboriginal landscape burning on the Australian biota. New Phytologist 140: 385–410.DMJS Bowman1998The impact of Aboriginal landscape burning on the Australian biota.New Phytologist140385410
- 13. Vigilante T, Bowman DMJS (2004) Effects of individual fire events on the flower production of fruit-bearing tree species, with reference to Aboriginal people's management and use, at Kalumburu, North Kimberley, Australia. Australian Journal of Botany 52: 405–416.T. VigilanteDMJS Bowman2004Effects of individual fire events on the flower production of fruit-bearing tree species, with reference to Aboriginal people's management and use, at Kalumburu, North Kimberley, Australia.Australian Journal of Botany52405416
- 14. Enright NJ, Thomas I (2008) Pre-European fire regimes in Australian ecosystems. Geography Compass 2: 979–1011.NJ EnrightI. Thomas2008Pre-European fire regimes in Australian ecosystems.Geography Compass29791011
- 15. Fensham R (1997) Aboriginal fire regimes in Queensland, Australia: analysis of the explorers' record. Journal of Biogeography 24: 11–22.R. Fensham1997Aboriginal fire regimes in Queensland, Australia: analysis of the explorers' record.Journal of Biogeography241122
- 16. Preece N (2002) Aboriginal fires in monssonal Australia from historical accounts. Journal of Biogeography 29: 321–336.N. Preece2002Aboriginal fires in monssonal Australia from historical accounts.Journal of Biogeography29321336
- 17. Archibald S, Scholes RJ, Roy DP, Roberts G, Boschettl L (2010) Southern African fire regimes as revealed by remote sensing. International Journal of Wildland Fire 19: 861–878.S. ArchibaldRJ ScholesDP RoyG. RobertsL. Boschettl2010Southern African fire regimes as revealed by remote sensing.International Journal of Wildland Fire19861878
- 18. Higuera PE, Whitlock C, Gage JA (2011) Linking tree-ring and sediment-charcoal records to reconstruct fire occurrence and area burned in subalpine forests of Yellowstone National Park, USA. Holocene 21: 327–341.PE HigueraC. WhitlockJA Gage2011Linking tree-ring and sediment-charcoal records to reconstruct fire occurrence and area burned in subalpine forests of Yellowstone National Park, USA.Holocene21327341
- 19. Niklasson M, Granstrom A (2000) The numbers and sizes of fires: long-term spatially explicit fire history in a Swedish boreal landscape. Ecology 81: 1484–1499.M. NiklassonA. Granstrom2000The numbers and sizes of fires: long-term spatially explicit fire history in a Swedish boreal landscape.Ecology8114841499
- 20. Wood SW, Hua Q, Allen KJ, Bowman DMJS (2010) Age and growth of a fire prone Tasmanian temperate old-growth forest stand dominated by Eucalyptus regnans, the world's tallest angiosperm. Forest Ecology and Management 260: 438–447.SW WoodQ. HuaKJ AllenDMJS Bowman2010Age and growth of a fire prone Tasmanian temperate old-growth forest stand dominated by Eucalyptus regnans, the world's tallest angiosperm.Forest Ecology and Management260438447
- 21. O'Dowd DJ, Gill AM (1984) Predator satiation and site alteration following fire: mass reproduction of Alpine Ash (Eucalyptus delegatensis) in south eastern Australia. Ecology 65: 1052–1066.DJ O'DowdAM Gill1984Predator satiation and site alteration following fire: mass reproduction of Alpine Ash (Eucalyptus delegatensis) in south eastern Australia.Ecology6510521066
- 22. Noble J, Gitay H (1996) A functional classification for predicting the dynamics of landscapes. Journal of Vegetation Science 7: 329–336.J. NobleH. Gitay1996A functional classification for predicting the dynamics of landscapes.Journal of Vegetation Science7329336
- 23. Lavorel S, Garnier E (2002) Predicting changes in community composition and ecosystem functioning from plant traits: revisiting the Holy Grail. Functional Ecology 16: 545–556.S. LavorelE. Garnier2002Predicting changes in community composition and ecosystem functioning from plant traits: revisiting the Holy Grail.Functional Ecology16545556
- 24. Keith DA, Holman L, Rodoreda S, Lemmon J, Bedward M (2007) Plant functional types can predict decade-scale changes in fire-prone vegetation. Journal of Ecology 95: 1324–1337.DA KeithL. HolmanS. RodoredaJ. LemmonM. Bedward2007Plant functional types can predict decade-scale changes in fire-prone vegetation.Journal of Ecology9513241337
- 25. Ooi MKJ, Whelan RJ, Auld TD (2006) Persistence of obligate-seeding species at the population scale: effects of fire intensity, fire patchiness and long fire-free intervals. International Journal of Wildland Fire 15: 261–269.MKJ OoiRJ WhelanTD Auld2006Persistence of obligate-seeding species at the population scale: effects of fire intensity, fire patchiness and long fire-free intervals.International Journal of Wildland Fire15261269
- 26. Duren OC, Muir PM (2010) Does fuels management accomplish restoration in southwest Oregon, USA, chaparal? Insights from age structure. Fire Ecology 6: 76–95.OC DurenPM Muir2010Does fuels management accomplish restoration in southwest Oregon, USA, chaparal? Insights from age structure.Fire Ecology67695
- 27. Tilman D (2004) Niche tradeoffs, neutrality and community structure: A stochastic theory of resource competition, invasion and community assembly. PNAS 101: 19854–10861.D. Tilman2004Niche tradeoffs, neutrality and community structure: A stochastic theory of resource competition, invasion and community assembly.PNAS1011985410861
- 28. Grubb PJ (1977) The maintenance of species richness in plant communities: the importance of the regeneration niche. Biological Review 52: 107–145.PJ Grubb1977The maintenance of species richness in plant communities: the importance of the regeneration niche.Biological Review52107145
- 29. Bond WJ, Midgley JJ (2001) Ecology of sprouting in woody plants: the persistence niche. Trends in Ecology & Evolution 16: 45–51.WJ BondJJ Midgley2001Ecology of sprouting in woody plants: the persistence niche.Trends in Ecology & Evolution164551
- 30. Pausas JG, Bradstock RA, Keith DA, Keeley JE, The GFN (2004) Plant Functional Traits in Relation to Fire in Crown-Fire Ecosystems. Ecology 85: 1085–1100.JG PausasRA BradstockDA KeithJE KeeleyGFN The2004Plant Functional Traits in Relation to Fire in Crown-Fire Ecosystems.Ecology8510851100
- 31. Russell-Smith J, Yates C, Edwards A, Allen GE, Cook GD, et al. (2003) Contemporary fire regimes of northern Australia, 1997–2001: change since Aboriginal occupancy, challenges for sustainable management. International Journal of Wildland Fire 12: 283–97.J. Russell-SmithC. YatesA. EdwardsGE AllenGD Cook2003Contemporary fire regimes of northern Australia, 1997–2001: change since Aboriginal occupancy, challenges for sustainable management.International Journal of Wildland Fire1228397
- 32. Yates CP, Edwards AC, Russell-Smith J (2008) Big fires and their ecological impacts in Australian savannas: size and frequency matters. International Journal of Wildland Fire 17: 768–81.CP YatesAC EdwardsJ. Russell-Smith2008Big fires and their ecological impacts in Australian savannas: size and frequency matters.International Journal of Wildland Fire1776881
- 33. Legge S, Murphy S, Kingswood R, Maher B, Swan D (2011) EcoFire: restoring the biodiversity values of the Kimberley region by managing fire. Ecological Management & Restoration 12: 84–92.S. LeggeS. MurphyR. KingswoodB. MaherD. Swan2011EcoFire: restoring the biodiversity values of the Kimberley region by managing fire.Ecological Management & Restoration128492
- 34. Groeneveld J, Enright NJ, Lamont BB, Wissel C (2002) A spatial model of coexistence among three Banksia species along a topographic gradient in fire-prone shrublands. Journal of Ecology 90: 762–774.J. GroeneveldNJ EnrightBB LamontC. Wissel2002A spatial model of coexistence among three Banksia species along a topographic gradient in fire-prone shrublands.Journal of Ecology90762774
- 35. Higgins SI, Flores O, Schurr FM (2008) Costs of persistence and the spread of competing seeders and sprouters. Journal of Ecology 96: 679–686.SI HigginsO. FloresFM Schurr2008Costs of persistence and the spread of competing seeders and sprouters.Journal of Ecology96679686
- 36. Armstrong G, Garnett S (2011) Landscape partitioning among Triodia spp. (Poaceae) in the fire prone Kimberley, north-west Australia. Austral Ecology 36: 849–857.G. ArmstrongS. Garnett2011Landscape partitioning among Triodia spp. (Poaceae) in the fire prone Kimberley, north-west Australia.Austral Ecology36849857
- 37. Krawchuk MA, Moritz MA, Parisien MA, Van Dorn J, Hayhoe K (2009) Global Pyrogeography: the Current and Future Distribution of Wildfire. PLoS ONE 4(4): e5102.MA KrawchukMA MoritzMA ParisienJ. Van DornK. Hayhoe2009Global Pyrogeography: the Current and Future Distribution of Wildfire.PLoS ONE44e5102
- 38. Thomas CD, Cameron A, Green RE, Bakkenes M, Beaumont LJ, et al. (2004) Extinction risk from climate change. Nature 427: 145–148.CD ThomasA. CameronRE GreenM. BakkenesLJ Beaumont2004Extinction risk from climate change.Nature427145148
- 39. Yates CJ, Elith J, Latimer AM, Le Maitre D, Midgley GF, et al. (2010) Projecting climate change impacts on species distributions in megadiverse South African Cape and Southwest Australian Floristic Regions: Opportunities and challenges. Austral Ecology 35: 374–391.CJ YatesJ. ElithAM LatimerD. Le MaitreGF Midgley2010Projecting climate change impacts on species distributions in megadiverse South African Cape and Southwest Australian Floristic Regions: Opportunities and challenges.Austral Ecology35374391
- 40. Armstrong G, Legge S (2011) The post-fire response of an obligate seeding Triodia species (Poaceae) in the fire-prone Kimberley, north-west Australia. International Journal of Wildland Fire 20: 974–981.G. ArmstrongS. Legge2011The post-fire response of an obligate seeding Triodia species (Poaceae) in the fire-prone Kimberley, north-west Australia.International Journal of Wildland Fire20974981
- 41. Armstrong G (2011) Evidence for the equal resilience of Triodia spp. (Poaceae), from different functional groups, to frequent fire dating back to the late Pleistocene. Heredity 107: 558–564.G. Armstrong2011Evidence for the equal resilience of Triodia spp. (Poaceae), from different functional groups, to frequent fire dating back to the late Pleistocene.Heredity107558564
- 42. Westoby M, Rice B, Griffin G, Friedel M (1988) The soil seed bank of Triodia basedowii in relation to time since fire. Australian Journal Ecology 13: 161–169.M. WestobyB. RiceG. GriffinM. Friedel1988The soil seed bank of Triodia basedowii in relation to time since fire.Australian Journal Ecology13161169
- 43. Ernst WHO, Veenendaal EM, Kebakile MM (1992) Possibilities for dispersal in annual and perennial grasses in a savanna in Botswana. Plant Ecology 102: 1–11.WHO ErnstEM VeenendaalMM Kebakile1992Possibilities for dispersal in annual and perennial grasses in a savanna in Botswana.Plant Ecology102111
- 44. Jongejans E, Schippers P (2009) Modeling seed dispersal by wind in herbaceous species. Oikos 87: 362–372.E. JongejansP. Schippers2009Modeling seed dispersal by wind in herbaceous species.Oikos87362372
- 45. R-Development-Core-Team (2010) R: A language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. R-Development-Core-Team2010R: A language and environment for statistical computingVienna, AustriaR Foundation for Statistical Computing
- 46. S-plus (2009) TIBCO Software Inc. Palo Alto CA USA. S-plus2009TIBCO Software Inc.Palo Alto CA USA