Berry, L. E., Lindenmayer, D. B., & Driscoll, D. A. (2014). Large unburnt areas, not small unburnt patches are needed to conserve avian diversity in fire‐prone landscapes. Journal of Applied Ecology.
1. Mitigating the impacts of large-scale fires on biodiversity is becoming increasingly important as their frequency increases. In response, fire managers have engaged with the concept that retaining small unburnt residual areas of vegetation within extensively burnt landscapes may facilitate biodiversity conservation. However, it remains uncertain how the size and isolation of these unburnt residuals influence faunal distributions, persistence and recovery following fire.
2. We used a replicated observation study to test bird responses to the size and isolation of unburnt residuals in a Mallee woodland area recently burnt by a 28,000-ha wildfire in southern Australia. The scale of our study provided a rare opportunity to consider the responses of large mobile organisms to fire-induced habitat fragmentation. Within five replicated spatial blocks, we crossed two levels of isolation with large (5–7ha) and small (1–3ha) unburnt patches and matrix sites burnt five years previously. We compared these site types to six continuous (non-fragmented) unburnt sites. We surveyed each site on eight occasions.
3. Most birds occurred more frequently in unburnt habitat beyond the extent of the fire. Bird responses to the availability and spatial distribution of unburnt remnants within the fire were largely influenced by their ability to use the recently burnt matrix. Occurrence of five species was higher in unburnt residuals when more of the landscape within 500 m was burnt.
4. A fire refuge effect may be likely for two competitive species that occurred more frequently in unburnt residuals than in the burnt matrix or continuous unburnt habitat. For the weebill, re-colonisation following fire was likely to occur gradually over-time from ex-situ sources.
5. Synthesis and applications. To maintain avian diversity in fire-prone landscapes, our results suggest a need to shift management focus from creating networks of small unburnt patches, towards preserving large, intact areas of habitat. However, five species common to the burnt matrix preferentially selected residual patches when unburnt resources were locally scarce. Therefore, to benefit birds, land managers should limit the extent of applied burns and use narrow burns. When planning large burns, practitioners should consider that a number of species will remain absent from the landscape for several decades.