The responses of small mammals to patches regenerating after fire and rainfall in the Simpson Desert,central Australia

Abstract Patch‐burning is frequently advocated as a management tool to enhance the biodiversity and pasture values of spinifex (Triodia) grasslands. In this study we compare the capture rates of small mammals in habitats regenerating shortly after fire (aged 1–5 years) and in long‐unburnt habitats (aged >25 years). To unravel the effects of temporally and spatially variable rainfall on capture rates, the study was replicated at three locations spaced over 50 km apart that experience different rainfall regimes. Ten species of small mammals were captured over the course of the study, between October 1999 and June 2001. Pseudomys desertor showed a strong preference for long‐unburnt habitats. Notomys alexis, Sminthopsis youngsoni and Sminthopsis hirtipes showed some preferences for regenerating habitats, but these were not consistent throughout the study. Factors indicative of temporal and spatial variation in rainfall, time and site had important effects on capture rates. High rainfalls associated with the La Niña phase of the El Niño/Southern Oscillation in 2000 increased seed production and prompted eruptions of rodent species and the carnivorous Dasycercus cristicauda. The greatest numbers of captures were made at the sites that received the highest rainfalls. We conclude that patch‐burning regimes do not benefit small mammals directly, but are likely to increase the resilience of ‘fire‐sensitive’ species that are dependent on dense spinifex by reducing the extent of wildfires.     

Changes in mammal populations in relatively intact landscapes of Kakadu National Park,Northern Territory,Australia

A previous study ( Braithwaite & Muller 1997 ) reported substantial declines in mammal abundance over the period 1986–1993 for a large study area (300 km²) within Kakadu National Park in the tropical savannas of northern Australia. This decline was reported as being a ‘natural’ response to fluctuating groundwater levels, driven by runs of poor wet seasons. We resampled mammals in this area in 1999, following a series of unusually good wet seasons, and examined the prediction that mammal numbers should have recovered. Increases in abundance were evident for four species: the smallest dasyurid (red‐cheeked dunnart Sminthopsis virginiae) and the three smallest rodents (delicate mouse Pseudomys delicatulus, western chestnut mouse Pseudomys nanus and grassland melomys Melomys burtoni). In contrast, the abundance of all mammals combined and that for seven individual mammal species (northern quoll Dasyurus hallucatus, fawn antechinus Antechinus bellus, common brushtail possum Trichosurus vulpecula, northern brown bandicoot Isoodon macrourus, dusky rat Rattus colletti, black‐footed tree‐rat Mesembriomys gouldii and pale field rat Rattus tunneyi) continued to decline. The decline in abundance of these mammal species is consistent with limited observations elsewhere in northern Australia. Although far from conclusive, these observations suggest that the biota of the vast relatively undisturbed tropical savannas can no longer be assumed to be intact nor safe. Further research is needed to test this possible pattern of decline and, if confirmed, to identify and ameliorate the processes contributing to it.     

Acrosomal integrity, viability, and DNA damage of sperm from dasyurid marsupials after freezing or freeze drying

Dasyurids are a diverse group of Australian native carnivores and insectivores that contains several threatened species. Despite successful cryopreservation of sperm from several marsupials, only 3% postthaw motility is reported for dasyurid marsupials. This study examined sperm preservation in the fat-tailed dunnart (Sminthopsis crassicaudata), an experimental model, with supplementary observations on the eastern quoll (Dasyurus viverrinus) and northern quoll (Dasyurus hallucatus). In S. crassicaudata, a toxicity trial demonstrated that incubation with up to 40% glycerol did not reduce sperm viability, suggesting that glycerol is not toxic to dasyurids. On the basis of this finding, S. crassicaudata, D. viverrinus, and D. hallucatus sperm were extended to a final concentration of 20% or 40% glycerol in Tris-citrate fructose and frozen in liquid nitrogen vapor. Postthaw sperm from all three species were nonmotile, and vital staining (SYBR14 and propidium iodide) indicated that sperm were nonviable. However, there was no evidence suggesting disruption of normal gross morphology or loss of acrosomal integrity when assessed by Bryan's staining. After freeze drying, Bryan's staining indicated that approximately 80% of S. crassicaudata sperm had normal acrosomes and no head loss. Despite being nonviable, terminal deoxynucleotidyl transferase–mediated dUTP nick-end labeling showed that S. crassicaudata sperm frozen in 40% glycerol or freeze-dried had no DNA damage compared with fresh controls. This study has described a method for preservation of the dasyurid sperm nuclei, but continued studies are required to achieve viable motile sperm and establish tools for the long-term storage of dasyurid sperm.     

Landscape partitioning among Triodia spp. (Poaceae) in the fire prone Kimberley,north‐west Australia

The processes which determine the structure of plant communities vary across spatial and temporal scales. Climatic factors are more likely to influence community structure at a regional scale with more transient environmental effects such as disturbance or demographic interactions having a greater influence at local scales. Understanding these differences is important for managing communities at a landscape scale. Triodia spp. grasslands are the most extensive plant community in Australia, covering 1.4 million km², and yet little is known about the processes which structure these communities. We collected data on six sympatric Triodia spp. at the regional, landscape and local scale across the 325 000 ha property, Mornington Wildlife Sanctuary, in the Kimberley region of northern Western Australia to investigate the processes which structure this community. Regionally we looked for correlations between species distributions and substrate or rainfall. At the landscape scale we collected data on substrate, drainage and vegetation type and at the local scale we determined the extent to which individuals form mono‐specific stands both along and across the contour gradient. Only one species, T. aeria, was found to be substrate specific and only T. epactia was restricted to the drier southern end of the property. The other species were not restricted by substrate or rainfall at the regional scale and were found to be habitat generalists at the landscape scale. All species grew in mono‐specific stands with little to no mixing at shared boundaries. However, this pattern broke down when crossing the contour gradient on hillsides. The results suggest rainfall may influence the distribution of some Triodia spp. at a regional scale with interspecific competition, due to differences in post‐fire regeneration niches, structuring the community at the local scale. At the landscape scale community structure appears to be influenced by feedback mechanisms involving differences in the post‐fire regeneration strategies of sympatric species and subsequent competition for establishment microsites.     

Response of native ungulates to drought in semi‐arid Kenyan rangeland

The distribution and abundance of native ungulates were measured on commercially managed, semi‐arid rangeland in central Kenya over a 3‐year period that encompassed severe drought and above‐average rainfall. Native ungulate biomass density averaged 5282 kg km^?2 over the study and was dominated by elephant (Loxodonta africana), impala (Aepyceros melampus) and dik‐dik (Madoqua kirkii). Biomass density of domestic cattle (Bos taurus) averaged 2280 kg km^?2 during the study. Responses of native ungulates to severe drought were variable. Impala densities were similar to or greater than densities for similar habitat in protected areas, and varied from 12 to 16 km^?2 during and following the drought to 24–29 km^?2 following above‐average rainfall. Dik‐dik densities were also greater than densities reported for protected areas and were surprisingly stable throughout the study despite the wide annual fluctuations in rainfall. Elephant migrated out of the region during drought but were present at high densities (2.9–5.2 km^?2) during wet seasons, consistent with telemetry studies emphasizing the importance of Acacia bushland habitat on commercial rangelands for the migratory portion of the Laikipia–Samburu elephant population. Results show that substantial densities of native browsing and mixed‐feeding ungulates can occur on rangeland managed for commercial beef production and suggest that the capacity for ungulates to move over large spatial scales (>100 km²) and to shift distributions in response to locally variable thunderstorms may be important for sustaining these populations.     

Predicting the potential distribution of a riparian invasive plant: the effects of changing climate,flood regimes and land‐use patterns

Climate change is likely to affect plants in multiple ways, but predicting the consequences for habitat suitability requires a process‐based understanding of the interactions. This is at odds with existing approaches that are mostly phenomenological and largely restricted to predicting the effects of changing temperature and rainfall on species distributions at a coarse spatial scale. We examine the multiple effects of climate change, including predicting the effects of altered flood regimes and land‐use change, on the potential distribution of the invasive riparian species lippia (Phyla canescens) across a 26 000 km² catchment in eastern Australia. We determined habitat suitability for lippia by combining process‐understanding of experts and an eco‐physiological bioclimatic model within a Bayesian belief network. The bioclimatic model predicted substantial changes in habitat suitability by 2070 under both a wetter (Echam Mark 3) and drier (Hadley Centre Mark 2) climate change scenario, but only the more likely drier scenario reduced suitability in our test region. The area suitable for lippia was predicted to increase at least threefold with increased flooding under a wet climate scenario, although this would be partially negated by land‐use change to cultivation. The region would become unsuitable to lippia with reduced flooding under a drier scenario irrespective of land‐use changes, although existing populations would persist if grazing persisted. Independent field validation verified model structure and parameterization, and therefore the opinion of experts, but identified site‐scale deficiencies in the available environmental data layers. Model predictions suggest that adaptation options for managing lippia will be greatly reduced under a drying scenario, but identify potential restoration opportunities under either scenario. This work highlights the value of predictive models that incorporate process‐understanding at sufficiently fine spatial resolution to capture the important processes underpinning habitat suitability.     

Absence of adaptive nonshivering thermogenesis in a marsupial,the fat-tailed dunnart (<Emphasis Type="Italic">Sminthopsis crassicaudata</Emphasis>)

The presence of nonshivering thermogenesis in marsupials is controversially debated. Survival of small eutherian species in cold environments is crucially dependent on uncoupling protein 1 (UCP1)-mediated, adaptive nonshivering thermogenesis that is executed in brown adipose tissue. In a small dasyurid marsupial species, the fat-tailed dunnart (Sminthopsis crassicaudata), an orthologue of UCP1 has been recently identified which is upregulated during cold exposure resembling adaptive molecular adjustments of eutherian brown adipose tissue. Here, we tested for a thermogenic function of marsupial brown adipose tissue and UCP1 by evaluating the capacity of nonshivering thermogenesis in cold-acclimated dunnarts. In response to an optimal dosage of noradrenaline, cold-acclimated dunnarts (12°C) showed no additional recruitment of noradrenaline-induced maximal thermogenic capacity in comparison to warm-acclimated dunnarts (24°C). While no differences in body temperature were observed between the acclimation groups, basal metabolic rate was significantly elevated after cold acclimation. Therefore, we suggest that adaptive nonshivering thermogenesis does not occur in this marsupial species despite the cold recruitment of oxidative capacity and UCP1 in the interscapular fat deposit. In conclusion, the ancient UCP orthologue in marsupials does not contribute to the classical nonshivering thermogenesis, and may exhibit a different physiological role.     

The spatial structure of variation in salamander survival,body condition and morphology in a headwater stream network

相似文献   

Congruent morphological and genetic differentiation as a signature of range expansion in a fragmented landscape

Phenotypic differentiation is often interpreted as a result of local adaptation of individuals to their environment. Here, we investigated the skull morphological differentiation in 11 populations of the white‐footed mouse (Peromyscus leucopus). These populations were sampled in an agricultural landscape in the Montérégie region (Québec, Canada), at the northern edge of the distribution of the white‐footed mouse. We found a strong pattern of phenotypic differentiation matching the genetic structure across these populations. Landscape fragmentation and the presence of geographic barriers, in particular north–south oriented rivers, contribute to this differentiation and modulate the pattern of rapid ongoing northward range expansion of the white‐footed mouse in response to climate warming. We conclude that while large rivers and postglacial recolonization routes have shaped the current pattern of distribution and differentiation of white‐footed mouse populations, further local differentiation is occurring, at the scale of the landscape. We posit that the northern expansion of the white‐footed mouse is achieved through successive independent founder events in a fragmented landscape at the northern range edge of the species. The phenotypic differentiation we observe is thus a result of a number of mechanisms operating at different spatial and temporal scales.     

Why are there so many coexisting species of lizards in Australian deserts?

Because Australian skinks of the genus Ctenotus display very high local species richness in arid-zone spinifex grasslands but not in mesic habitats, these lizards have been used as ”model organisms” to ask why ecologically similar taxa coexist under some circumstances but not others. Previous work has involved detailed studies within small areas, and has looked for differences in ecological processes between arid versus mesic habitats. We suggest a radically different explanation for the high species-richness of arid-zone Ctenotus, by shifting attention to a larger spatial scale: the regional species pool. Analyses of the geographic distributions of Ctenotus species confirm that more species coexist at sites in the arid-zone (mean =9.3 species per site) than in other climatic zones (means 2.4–7.6). However, the total number of species occurring within the arid-zone is actually lower, per km² of habitat, than is the case in some other climatic zones. That is, arid-zone Ctenotus show a higher local (alpha) species diversity, but a lower regional (gamma) diversity, than their mesic-habitat congeners. This apparent paradox occurs because most arid-zone species occur over vast areas (mean =1,035,000 km²), whereas congeners from other climatic zones have smaller geographic ranges (200–373,000 km²). The broad distributions of arid-zone taxa reflect the great spatial homogeneity in climatic conditions in this zone. That is, the ”climate spaces” occupied are similar for Ctenotus species from all bioclimatic regions. Thus, a given amount of climatic space translates into a larger geographic distribution (and hence, more sympatry) in the arid-zone than in other areas. In summary, the high number of coexisting Ctenotus species in arid-zone habitats may simply reflect the facts that the arid zone is large (so that many species have evolved therein) and climatically homogeneous (so that any species evolving in that habitat type can disperse very widely, and thus overlap with many other species). Our approach explains much of the variance in local-assemblage species richness from regional to site scales; but explanations invoking biological attributes of the species concerned, the nature of their interactions with other species or with particular resources (such as prey or shelter) may still be significant at microhabitat scales. For lizard communities in Australia, species richness at a site may be determined more by continental biogeography rather than by ecological interactions. Received: 28 June 1999 / Accepted: 14 April 2000     

Climate change alters the trophic niche of a declining apex marine predator

Changes in the world's oceans have altered nutrient flow, and affected the viability of predator populations when prey species become unavailable. These changes are integrated into the tissues of apex predators over space and time and can be quantified using stable isotopes in the inert feathers of historical and contemporary avian specimens. We measured δ¹³C and δ¹⁵N values in Flesh‐footed Shearwaters (Puffinus carneipes) from Western and South Australia from 1936–2011. The Flesh‐footed Shearwaters more than doubled their trophic niche (from 3.91 ± 1.37 ‰² to 10.00 ± 1.79 ‰²), and dropped an entire trophic level in 75 years (predicted δ¹⁵N decreased from +16.9 ‰ to + 13.5 ‰, and δ¹³C from ?16.9 ‰ to ?17.9 ‰) – the largest change in δ¹⁵N yet reported in any marine bird, suggesting a relatively rapid shift in the composition of the Indian Ocean food web, or changes in baseline δ¹³C and δ¹⁵N values. A stronger El Niño‐Southern Oscillation results in a weaker Leeuwin Current in Western Australia, and decreased Flesh‐footed Shearwater δ¹³C and δ¹⁵N. Current climate forecasts predict this trend to continue, leading to increased oceanic ‘tropicalization' and potentially competition between Flesh‐footed Shearwaters and more tropical sympatric species with expanding ranges. Flesh‐footed Shearwater populations are declining, and current conservation measures aimed primarily at bycatch mitigation are not restoring populations. Widespread shifts in foraging, as shown here, may explain some of the reported decline. An improved understanding and ability to mitigate the impacts of global climactic changes is therefore critical to the long‐term sustainability of this declining species.     

Resource partitioning in sympatric arctic-breeding geese: summer habitat use, spatial and dietary overlap of Barnacle and Pink-footed Geese in Svalbard

The spatial, habitat and dietary overlap of two breeding goose species was studied in Sassendalen, Svalbard, in summer 2003 based on abundance within 500 × 500‐m grid squares and faecal diet analyses during pre‐breeding, nesting and post‐hatching periods. More than half of all Pink‐footed Geese Anser brachyrhynchus occurred in the absence of Barnacle Geese Branta leucopsis during nesting and post‐hatching periods compared to c. 20% when concentrated by pre‐breeding snow cover. In contrast, only 5% of Barnacle Geese were observed in the absence of Pink‐footed Geese pre‐breeding, 15% during nesting, and 35% post‐hatching. Among six defined habitat types, Barnacle Geese resorted more to ‘upland’ habitats during pre‐breeding and nesting and to lowland lakes post‐hatching when compared to Pink‐footed Geese. Although Pink‐footed Geese showed less change in seasonal habitat preference, many shifted to the river valley bottom post‐hatching, giving access to open water (predator avoidance) and lush green vegetation (foraging for goslings). The smallest extent of distributional overlap between the two species occurred post‐hatching, but each species was also highly restricted by snow cover during pre‐nesting. The greatest extent of overlap in distribution and diet occurred during incubation, when large dietary variation between different breeding valleys reflected local food availability around nests (probably a result of nest‐site preference rather than food selection per se). Whether this means that increased interactions within and between the two goose species with future increases in local density are most likely to be manifest at this stage of the summer is impossible to determine without knowledge of available food resources and manipulative experiments. More detailed investigations of the effects of foraging by both species on plant structure, quality and community composition are necessary to predict likely outcomes of future changes in population densities of both species.     

The relationship between multi‐scale habitat heterogeneity and farmland bird abundance

It has recently been proposed that losses in farmland habitat heterogeneity may have been a primary driver of the profound declines exhibited by many farmland bird species in recent decades. However, it has yet to be demonstrated which facets of heterogeneity and what spatial scales are most important for birds. Here we analyse the relationship between abundance and features of landscape heterogeneity at three spatial scales (1, 9 and 25 km²) for 32 bird species commonly associated with farmland. Heterogeneity was quantified using three contrasting indices reflecting 1) the spatial mixing of land uses, 2) variation in field sizes and 3) the density of field boundaries. The spatial mixing of land‐uses explained, on average, the most variation in, and was most likely to be positively associated with, abundance at all spatial scales. The majority of species (66–75%, depending on the spatial scale) were more common in heterogeneous landscapes overall; however, migrants, those under a high level of conservation concern and farmland specialist species tended to be less abundant in more heterogeneous landscapes at all scales. Ground‐nesting species were also more likely to be found in more homogeneous habitats than non‐ground‐nesters, but only at the finest spatial scale. Relationships between abundance and heterogeneity were generally consistent across spatial scales; however, species of high conservation concern had more variable associations compared with other species. These results highlight a potential role for farmland habitat heterogeneity in determining the abundance of many farmland species but suggest that population responses to an increase in heterogeneity would not be unanimously positive and would probably have negative impacts on some species, notably those that are already threatened.     

Scale dependence in desert plant diversity

The relationship between the number of species and the area sampled is one of the oldest and best-documented patterns in community ecology. An equation of the form S = cA ^z describing more precisely the species–area relation for plant species in smaller area is proposed as a result of intensive examination of species presence. Several study and field data from a wild range of plant and animal taxa suggest that the slope, z, of a graph of the logarithm of species richness against the logarithm of area is not a constant to the grassland or woodland community. We collected replicated and randomized plant data at 6 spatial scales from 1 m² to 1 km² in the desert region of northwest China to identify the scale dependence in desert plant biodiversity. The results showed that the slope of the log–log plot varied systematically with spatial scale. The value of z was high (0.37) at small scales from 1 to 10 m² and it decreased with increased spatial scale subsequently. When spatial scales varied from 1 m² to 1 km², the value of z varied from 0.37 to 0.035, suggesting that desert plant diversity has strongly scale-dependence at the small scales (less than 100 m²). The result is different from the research of grassland and woodland communities.     

Molecular evidence for the pattern and timing of cladogenesis in dasyurid marsupials

Recent molecular studies have provided estimates of phylogeny for nearly all living and recently extinct species in the Order Dasyuromorphia, the dominant clade of insectivorous‐carnivorous marsupials in Australasia. We review these studies along with morphology‐based ones, and present an analysis of all cytochrome b, 12S rRNA, and protamine Pl gene sequences available. In light of these results, we provide a revised suprageneric classification and assess the implications of molecular and paleontological data for dasyurid cladogenesis. Molecular results divide extant dasyurids (Dasyuridae) into four major clades apart from the numbat (Myrmecobiidae) and thylacines (Thylacinidae). We recognize these clades as tribes Dasyurini (Dasyurus, Phascolosorex, and allied genera) and Phascogalini (Antechinus, Murexia, Phascogale) in the Subfamily Dasyurinae, and tribes Sminthopsini (Sminthopsis, Ningaui, Antechinomys) and Planigalini (Planigale) in the Subfamily Sminthopsinae. Each tribe shows a basal radiation of lineages corresponding to genera or species groups. Our results concur with the most recent previous synthesis of dasyurid phylogeny in many respects, but subsumption of New Guinean ‘phascolosoricines’ and ‘muricines’ within Dasyurini and Phascogalini, respectively, constitute significant differences. In particular, the sister‐pairing of ‘phascolosoricines’ with a Dasyurus‐Sarcophilus clade implied by molecular data is difficult to reconcile with anatomy. Divergence rates of mitochondrial sequences are calibrated approximately by comparing thylacine‐to‐dasyurid distances with the age of the oldest thylacinid (Badjcinus, latest Oligocene). Estimated cladogenic dates suggest that extant subfamilies shared a common ancestor around 24 Mya and that major radiations began late in the mid‐Miocene, consistent with the results of previous paleontological studies. The late‐middle and late Miocene corresponds to an episode of faunal turnover in Australian marsupials (including the decline of thylacinid and bandicoot genera, as well as the rise of dasyurids) and to a time when uplift of the New Guinean highlands accelerated the transition from rainforest to drier habitats. Our findings are consistent with the hypothesis that continent‐wide climate changes modulated macroevolution across these independent marsupial clades.     

What factors influence the diversity of saproxylic beetles? A multiscaled study from a spruce forest in southern Norway

The diversity in different groups of obligate saproxylic beetles was related to ecological variables at three levels of spatial scale in mature spruce-dominated forest. The variables were connected to: (i) decaying wood, (ii) wood-inhabiting fungi, (iii) the level of disturbance, (iv) landscape ecology, and (v) vegetational structure. Several strong relationships were found at medium (1 km²) and large scales (4 km²), while only weak relationships were found at a small scale (0.16 ha; 1 ha=10⁴ m²). This may be explained by the local variations in habitat parameters and the high mobilities of many beetle species. Factors connected to decaying wood and wood-inhabiting fungi were clearly the most important factors at all scale levels. In particular, the variables diversity of dead tree parts, number of dead trees of large diameter and number of polypore fungi species increased the species richness of many groups and increased the abundance of many species. Eight species were absent below a certain density of decaying wood per 1 or 4 km². Former extensive cutting was a negative factor at large scale, probably because of decreasing recolonization with increasing distance to the source habitats. Thinning reduced the diversity of species associated with birch. The development of guidelines favouring the diversity of saproxylic beetles are discussed below.     

A local evaluation of the individual state‐space to scale up Bayesian spatial capture–recapture

Spatial capture–recapture models (SCR) are used to estimate animal density and to investigate a range of problems in spatial ecology that cannot be addressed with traditional nonspatial methods. Bayesian approaches in particular offer tremendous flexibility for SCR modeling. Increasingly, SCR data are being collected over very large spatial extents making analysis computational intensive, sometimes prohibitively so. To mitigate the computational burden of large‐scale SCR models, we developed an improved formulation of the Bayesian SCR model that uses local evaluation of the individual state‐space (LESS). Based on prior knowledge about a species’ home range size, we created square evaluation windows that restrict the spatial domain in which an individual's detection probability (detector window) and activity center location (AC window) are estimated. We used simulations and empirical data analyses to assess the performance and bias of SCR with LESS. LESS produced unbiased estimates of SCR parameters when the AC window width was ≥5σ (σ: the scale parameter of the half‐normal detection function), and when the detector window extended beyond the edge of the AC window by 2σ. Importantly, LESS considerably decreased the computation time needed for fitting SCR models. In our simulations, LESS increased the computation speed of SCR models up to 57‐fold. We demonstrate the power of this new approach by mapping the density of an elusive large carnivore—the wolverine (Gulo gulo)—with an unprecedented resolution and across the species’ entire range in Norway (> 200,000 km²). Our approach helps overcome a major computational obstacle to population and landscape‐level SCR analyses. The LESS implementation in a Bayesian framework makes the customization and fitting of SCR accessible for practitioners working at scales that are relevant for conservation and management.     

Using generalized autoregressive error models to understand fire–vegetation–soil feedbacks in a mulga–spinifex landscape mosaic

Aim To develop a new modelling approach for spatially autocorrelated non‐normal data, and apply it to a case study of the role that fire–vegetation–soil feedbacks play in maintaining boundaries between fire‐sensitive and fire‐promoted plant communities. Location A mulga (Acacia aneura) shrubland–spinifex (Triodia spp.) grassland mosaic, central Australia. Methods Autoregressive error models were extended to non‐normal data by incorporating neighbourhood values of the response and predictor variables into generalized nonlinear models. These models were used to examine the environmental correlates of three response variables: mulga cover; fire frequency in areas free of mulga; and the presence of mulga banding. Mulga cover and mulga banding were assessed visually by overlaying 4477 × 1 km² grid cells on both Landsat 7 ETM+ and very high resolution imagery. Fire frequency was estimated from an existing fire history for central Australia, based on remotely sensed fire scars. Results The autoregressive error models explained 27%, 47% and 57% of the null deviance of mulga cover, fire frequency and mulga banding, respectively, with 12%, 15% and 24% of the null deviance being explained by environmental variables alone. These models accounted for virtually all residual spatial autocorrelation. While there was a clear negative relationship between mulga cover and fire frequency, there was little evidence that mulga was being restricted to parts of the landscape with inherently low fire frequencies. Mulga was most abundant at very low slope angles and on red earths, both of which are likely to reflect high site productivity, while fire frequency was not clearly affected by slope angle and was also relatively high on red earths. Main conclusions The modelling approach we have developed provides a much needed way of analysing spatially autocorrelated non‐normal data and can be easily incorporated into an information‐theoretic modelling framework. Using this approach, we provide evidence that mulga and spinifex have a highly antagonistic relationship. In more productive parts of the landscape, mulga suppresses spinifex and fire, while in less productive parts of the landscape, fire and spinifex suppress mulga, leading to the remarkable abruptness of mulga–spinifex boundaries that are maintained via fire–vegetation–soil feedbacks.