The link between ‘fire mosaics’ and persistence of animal species is part of a prominent ecological/land management paradigm. This paradigm deals largely with the effects of fire on animals on the basis of individual events. The universality of the paradigm can be questioned on a variety of grounds, a major deficiency being the inability to deal with quantitative effects of recurrent fire (the fire regime). A conceptual model of fire-related habitat elements is proposed for exploration of a continuum of species/habitat/landscape/fire regime combinations. This approach predicts that the dependence of species on fire-mediated habitat heterogeneity will be highly variable and strongly context-dependent. A spatially explicit simulation model was used to examine the persistence of malleefowl (Leipoa ocellata) in a specific landscape/habitat context where dependence on fire-mosaics should be high. Results suggest that persistence of L. ocellata populations will be dependent on intervention using small patchy fires but that there is an optimum rate of intervention. Results were sensitive to spatial pattern of prescribed fire, landscape type (topography) and probability of wildfire. Underlying effects of the fire-interval distribution (the ‘invisible’ mosaic) on plant species and habitat account for these results. A management emphasis on species/landscape context and awareness of the ‘invisible’ mosaic is advocated.

The probability of large-fire (≥1000 ha) ignition days, in the Sydney region, was examined using historical records. Relative influences of the ambient and drought components of the Forest Fire Danger Index (FFDI) on large fire ignition probability were explored using Bayesian logistic regression. The preferred models for two areas (Blue Mountains and Central Coast) were composed of the sum of FFDI (Drought Factor, DF = 1) (ambient component) and DF as predictors. Both drought and ambient weather positively affected the chance of large fire ignitions, with large fires more probable on the Central Coast than in the Blue Mountains. The preferred, additive combination of drought and ambient weather had a marked threshold effect on large-fire ignition and total area burned in both localities. This may be due to a landscape-scale increase in the connectivity of available fuel at high values of the index. Higher probability of large fires on the Central Coast may be due to more subdued terrain or higher population density and ignitions. Climate scenarios for 2050 yielded predictions of a 20–84% increase in potential large-fire ignitions days, using the preferred model.

Soil temperatures were measured during 11 experimental fires in semi-arid mallee shrublands in central NSW. Sensors were placed at depths from 1-10 cm beneath the soil surface in three fuel types; litter beneath Eucalyptus shrubs, live hummocks of the grass Triodia irritans and litter beneath shrubs oi Acacia species. Weights of these fuels per unit area were determined.Maximum soil temperature and its duration were related to fuel type and depth. Mean weights oi Eucalyptus and Triodia fuels were similar (0.35 kg m'-), while there was less Acacia fuel (0.1 kg m"-). Highest maximum temperatures were registered under Eucalyptus litter (e.g. 140°Cat 2 cm). Maximum temperatures under Triodia And Acacia litter were similar (e.g. 60-70°C at 2 cm). Durations were examined in two temperature classes (60-120 and > 120°C) chosen to represent threshold for stimulation of germination and mortality, respectively, of soil-stored seeds. Temperatures between 60 and 120°C were recorded only between 0-2 cm soil depth for Acacia and Triodia (one exception at 4 cm). No temperatures >120°C were recorded for these fuel types. Temperatures between 60 and 120°C were recorded to 5 cm depth under Eucalyptus fuels while putative lethal temperatures for seeds occurred occasionally at 0-2 cm depth.The results indicated greatest potential for stimulation of germination and death of buried seeds under Eucalyptus fuels, although the level of variability of temperature was highest under Eucalyptus fuels. Despite similar fuel loads, differences between temperatures under Eucalyptus and Triodia fuels reflected the influence of the depth of the fuel bed, with Triodia hummocks constituting a deep fuel bed and Eucalyptus litter a shallow fuel bed.

Question: To what extent do low flammability fuel traits enhance the survival and persistence of fire‐sensitive (slowing‐growing, non‐serotinous, non‐resprouting) dominant trees in highly flammable landscapes, under varying fire‐weather conditions? Location: Mixed forests co‐dominated by flammable Eucalyptus species and fire‐sensitive Callitris glaucophylla in Pilliga State Forest, southeast Australia. Methods: The influence of vegetation composition (relative abundance of Callitris and flammable Eucalyptus) on fire intensity and survival of fire‐sensitive Callitris was assessed across gradients of Callitris abundance in mixed Eucalyptus–Callitris forests that burned under low‐moderate and extreme fire‐weather conditions. Results: In areas that burned under low‐moderate fire‐weather conditions, as Callitris abundance increased, fire intensity declined and Callitris survival increased (46%). By comparison, in extreme fire‐weather conditions, lower fire intensity at higher levels of Callitris abundance, was not sufficient to increase Callitris survival (4%). Callitris survival was also positively related to trunk diameter. Ground fuel type, but not biomass, varied with vegetation composition. Conclusions: These results demonstrate that flammable feedbacks, mediated by low flammability fuel traits of dominant trees, can provide an important mechanism for enhancing the survival and persistence of slow‐growing, non‐serotinous, non‐resprouting, fire‐killed trees in highly flammable landscapes. By modifying vegetation and fuel structure, patches of fire‐sensitive Callitris reduce fire intensity, and thereby reduce Callitris mortality, enhancing population persistence. However, this feedback loop is insufficient to ensure Callitris survival under extreme fire‐weather conditions, when fire intensity is greater. After burning, stands remain vulnerable to future fires, until trees grow large enough to modify fuel levels and reduce stand flammability.

Aspects of the demography of Callitris preissii subsp. verrucosa (Cunn. ex. Endl.) J.Garden populations were studied in semi-arid mallee communities situated on sandy soils in central New South Wales. Rates of survival and fecundity were estimated in a range of populations of differing age (year of origin or last fire: 1920, 1957, 1972, 1974 and 1985) that were monitored between 1987 and 1997. Rates of survival of juveniles and adults were high (>0.99 p.a.), although juveniles less than 5 years of age survived at lower rates. Estimated seedbank at the time of first tagging (no. of closed cones per plant) followed an increasing trend with plant age. Density, overstorey, topography and within-age-class site factors had significant effects on stored seedbanks in particular age-classes. There was high interannual variation (1989 v. 1990) in the size of crops of new cones. Trends in seedbank with plant age and seed release rates from tagged cones indicated that the species exhibits strong serotiny. Measurements of litter weight and depth beneath plants from 1920 and 1957 age-classes, indicated relatively low fuel loadings (about 0.2�kg m–2). Bark was sufficiently thick in c. 60-year-old plants to render stems resistant to low-intensity fires. Thus, it is predicted that the probability of propagation of fire in populations >50 years old may decline and that survival of fire in this age range may increase. Given patterns of survivorship, fecundity and seedbank accumulation, populations may decline when fires are relatively frequent (<15-year interval). At longer intervals there is potential for population densities to remain stable or increase, although the nature of variability in sizes of individual seed crops may have a predominant influence on dynamics. The combination of strong serotiny and negative feedback effects of plant populations on their flammability is a paradox, given the likelihood that seedling establishment is tightly keyed to fire.

No abstract

Factors affecting the survival of post-fire germinants in mallee communities, in central western New South Wales, were examined. Experiments compared the relative effects of native and introduced herbivores (kangaroos, goats, rabbits), after small-and large-scale fires (20-50 and > 10 000 ha, respectively), with particular emphasis on edge effects, seedling clustering, topography and eucalypt canopy presence. The experiments (1985)(1986)(1987)(1988)(1989)(1990)(1991)(1992)(1993)(1994)(1995)(1996)(1997) focused on common understorey species Acacia rigens Cunn. ex Don, A. wilhelmiana F.Muell. and Triodia scariosa N.T.Burb. subsp. scariosa, in mallee dominated by Eucalyptus species. Following a large fire (1985), high spring rainfall and rabbit grazing on A. rigens only, survival of Acacia species and T. scariosa remained relatively high 4 years later (60-70%). After small burns (1987, 1988), low spring rainfall and grazing by rabbits and kangaroos, survival of Acacia species declined to between 0 and 30% of the germinants by the second summer. In most cases, local extinction had occurred within 8 years. After small burns (1988, 1989) and low spring rainfall, the survival of T. scariosa declined to between 0 and 35% of germinants by the second summer (effect of grazing unknown). No consistent effect of edge, topography and eucalypt canopy was found. Survival of clustered Acacia seedlings was between 10 and 20% lower than unclustered seedlings. Given the high frequency of low rainfall and its interaction with grazing, prescribed burning of mallee for wildfire control and nature conservation may require the local elimination of rabbits and a reduction in kangaroo numbers, especially in the first spring and summer following seedling germination.

Soil seedbanks play a key role in the post-fire recruitment of many plant species. Seedbank diversity can be influenced by spatial variability (e.g. geographic location), environmental variability (e.g. soils) and temporal disturbance heterogeneity (e.g. time since fire, TSF) across the landscape. Unlike for aboveground vegetation, relationships between these factors and soil seedbank diversity remain largely unknown. Partitioning the influence of spatial and environmental variability from that of TSF, and explaining how these factors interact with seedbank diversity, will assist conservation managers in their application of prescribed burning. We germinated soil seedbank samples from sites ranging from 1 to 75 years since fire in a heathy-woodland ecosystem across the Otway Ranges in Victoria, Australia. We also measured spatial and environmental variability across sites to partition the influence of these variables and TSF on propagules available for recruitment. We found weak positive relationships between seedbank richness and TSF; however, these relationships varied across the landscape. We found composition did not change considerably over time, suggesting, in this ecosystem, pre-fire age is not strongly influencing propagules available for recruitment post-fire. Our results suggest that spatial and environmental variability influence seedbank composition more than TSF.