We report the effects of an extensive (>7000 km2), high-intensity late-dry-season fire in the central Kimberley, Western Australia, on the species richness and abundance of mammals, reptiles and birds. Five weeks after the fire we surveyed 12 sites (six burnt, six unburnt); each pair of sites was closely matched for soil type and vegetation. The species richness and abundance of mammals and reptiles was greater at unburnt sites, especially for mammals (with a 4-fold difference in abundance between burnt and unburnt sites). There was an indication that reptiles immigrated into unburnt patches, but mammals did not. There were also species-specific responses to the fire: Rattus tunneyi and Pseudomys nanus were much more abundant in unburnt sites, whereas Pseudomys delicatulus was caught in equal numbers at burnt and unburnt sites. Diurnal reptiles were more abundant at unburnt sites, but nocturnal reptiles were equally common at burnt and unburnt sites. Avian species richness and overall abundance was similar between burnt and unburnt patches, although a few species showed preferences for one state or the other. The overall high trapping success for mammals (18% across all sites; 28% in unburnt patches) contrasts with the well documented mammal collapse in parts of northern Australia and seems paradoxical given that our study area has experienced the same increase in fire frequency and extent that is often blamed for species collapse. However, our study area has fewer pressures from other sources, including grazing by large herbivores, suggesting that the effects of these pressures, and their interaction with fire, may have been underestimated in previous studies.

Context. Mismanaged fire is one of the main processes threatening biodiversity in northern Australian savannas. While the suite of species adversely affected by fire has been quite well documented, empirical studies that seek to identify the ecological mechanisms that underpin these declines are rare. This dearth of mechanistic knowledge is hindering the refinement and development of conservation management practices and policy, such as early dry-season prescribed burning programmes. Aims. Our aim was to describe why red-backed fairy-wrens (RBWs; Malurus melanocephalus) decline after fire, as shown by previous studies. Methods. We examined the effects of early and late dry-season fire (and an unburnt control) on the mortality, dispersal, physiology and breeding of colour-marked RBWs. We also tested whether the early fire area became an important refuge after the late fire. Key results. Neither fire treatment caused a detectable increase in mortality. Individuals relocated short distances to unburnt habitat following both fires. Some individuals used the early fire area after the late fire, but only if they were already living nearby and most also relied on adjacent unburnt habitat. Blood-parasite infection and pectoral-muscle attrition were unaffected by the treatments. The early fire caused loss of body mass at a time when the control population gained mass after breeding, whereas the late fire had no effect on mass. Blood haemoglobin concentrations in birds affected by the late fire deviated from the levels observed in the control population, presumably because of disruption of behaviours associated with the onset of the breeding season. Mayfield estimations of daily egg-survival probability showed no difference among the treatments, whereas the daily probability of nestling survival was significantly lower in the late fire area. Pairs affected by the late fire also had a shorter breeding season and none attempted multiple broods after earlier successes. Conclusions. Our results suggest that the reproductive output of RBWs is low following late dry-season fire, and that this is the main mechanism that explains their decline. Implications. This study provides support for the value of using early dry-season prescribed burning to limit the effects of late dry-season wildfire. However, there are important caveats to this, such as the retention of near-by unburnt habitat, and the consideration of variable regeneration rates for different vegetation types.

Fire is an integral part of savanna ecology and changes in fire patterns are linked to biodiversity loss in savannas worldwide. In Australia, changed fire regimes are implicated in the contemporary declines of small mammals, riparian species, obligate-seeding plants and grass seed-eating birds. Translating this knowledge into management to recover threatened species has proved elusive. We report here on a landscape-scale experiment carried out by the Australian Wildlife Conservancy (AWC) on Mornington Wildlife Sanctuary in northwest Australia. The experiment was designed to understand the response of a key savanna bird guild to fire, and to use that information to manage fire with the aim of recovering a threatened species population. We compared condition indices among three seed-eating bird species–one endangered (Gouldian finch) and two non-threatened (long-tailed finch and double-barred finch)—from two large areas (> 2,830 km2) with initial contrasting fire regimes (‘extreme’: frequent, extensive, intense fire; versus ‘benign’: less frequent, smaller, lower intensity fires). Populations of all three species living with the extreme fire regime had condition indices that differed from their counterparts living with the benign fire regime, including higher haematocrit levels in some seasons (suggesting higher levels of activity required to find food), different seasonal haematocrit profiles, higher fat scores in the early wet season (suggesting greater food uncertainty), and then lower muscle scores later in the wet season (suggesting prolonged food deprivation). Gouldian finches also showed seasonally increasing stress hormone concentrations with the extreme fire regime. Cumulatively, these patterns indicated greater nutritional stress over many months for seed-eating birds exposed to extreme fire regimes. We tested these relationships by monitoring finch condition over the following years, as AWC implemented fire management to produce the ‘benign’ fire regime throughout the property. The condition indices of finch populations originally living with the extreme fire regime shifted to resemble those of their counterparts living with the benign fire regime. This research supports the hypothesis that fire regimes affect food resources for savanna seed-eating birds, with this impact mediated through a range of grass species utilised by the birds over different seasons, and that fire management can effectively moderate that impact. This work provides a rare example of applied research supporting the recovery of a population of a threatened species.

No abstract