Wildfires, or bushfires, are not uniform in severity. Variation in topography, fuel loads and weather conditions result in patchiness: some areas are burnt intensely, others are burnt lightly, while others may escape burning altogether. The patchiness, or heterogeneity, created by fire has profound implications for forest ecosystems. It has been particularly well studied in Yellowstone National Park, USA, following an intense wildfire in 1988 that burnt a third of the national park during drought conditions. Spatial variation in fire severity created a complex post-fire mosaic of burnt and unburnt forest, varying abundance of dead timber, and fine-scale variation in understorey and microbial communities – that will shape the forest ecosystem for decades to come.
A potentially important role of unburnt patches is to serve as refuges for fire-sensitive plants and animals, assisting them to survive during the fire event, persist in the burned landscape after the fire, and recolonise the burned landscape. But how many unburnt patches remain? Few studies have examined this issue. In boreal forests of Quebec, Canada, for example, a study of 33 large wildfires found that on average 10% of the fire area remained unburnt, with this varying from 2-22% among fires.
What about bushfires in southern Australia? Are unburnt patches common? Where are they located? What determines whether a patch remains unburnt? These are some of the questions examined in a recent study led by Steve Leonard (La Trobe University), of unburnt forest patches within the boundary of the Kilmore-Murrindindi fire complex, part of the devastating Black Saturday bushfires in Victoria in February 2009.
Within the area of the fire boundary, some 250,000 ha in total, 84 unburnt patches of 1 ha or greater were identified. They ranged up to 306 ha in size, with an average of 27.1 ha. Collectively, however, they made up less than 1% of the overall study area! Unburnt patches were rare indeed in this fire.
For the overall fire area, a strong predictor of unburnt forest patches was topography. Sites within deep gullies were less likely to burn, especially in wet eucalypt forest and cool temperate rainforest vegetation. In contrast, for that part of the study area with dry eucalypt forest, the time since last fire was a strong predictor of unburnt patches. Areas that had been burned within the previous 3.5 years were more likely to be spared from the bushfire.
A further important influence was the proportion of the surrounding area that experienced high severity fire that burned tree crowns. The higher the intensity of a fire as it moves through an area, the less likely that patches of forest will remain unburnt. Notably, no unburnt patches remained within the forests burnt during the extreme weather conditions when the fire started, prior to a wind shift later in the evening. Kilometre after kilometre of forest was entirely burned, including some stands where fuel reduction burns had occurred in the previous year.
Are such unburnt patches important for the persistence plants and animals of this forest system? That’s a topic for further posts and the theme of a research project underway in our group.