Centre for Environmental Risk Management of Bushfires, Institute for Conservation Biology and Environmental Management, University of Wollongong, Wollongong, New South Wales, Australia.
PLoS One. 2013 Aug 29;8(8):e73421. doi: 10.1371/journal.pone.0073421. eCollection 2013.
Previous investigations into the factors associated with house loss in wildfires have focused on the house construction and its immediate environment (e.g. gardens). Here, we examine how nearby native forest and other houses can influence house loss. Specifically, we used a sample of 3500 houses affected by the Victorian bushfires of February 7th 2009 to explore how the amount of forest, proportion of forest burned by crown fire and the number of nearby houses affected house loss and how far from the house this influence was exerted. These fires were the most destructive in Australian history and so represent the extreme of fire risk. Using generalized linear modeling we found that the probability of house loss increased with forest extent and the proportion burnt by crown fire and this relationship was strongest for forest measured 1 km from the houses. Houses were more likely to be destroyed if there were other houses within 50 m and if they were on a slope. A model containing these variables predicted house loss with 72% accuracy. Our findings have three important implications: i) management to change the occurrence of crown fire will be effective in reducing house loss; ii) this management may be required up to 1 km away from houses in some situations (a much larger zone than is currently used); iii) high density of houses may increase risk of loss. Given the potentially large width of this management zone and the hazard from nearby houses, it may be more sensible to concentrate on modification of buildings to reduce their vulnerability.
先前的研究主要集中在房屋结构及其周边环境(如花园)与野火中房屋损毁之间的关系上。在此,我们研究了附近原生林和其他房屋如何影响房屋损毁。具体而言,我们利用 2009 年 2 月 7 日维多利亚丛林大火中 3500 间受灾房屋的数据,探索了森林面积、树冠火燃烧的森林比例以及附近房屋数量如何影响房屋损毁,以及这种影响的距离有多远。这些火灾是澳大利亚历史上最具破坏性的,代表了火灾风险的极端情况。通过广义线性建模,我们发现房屋损毁的概率随着森林面积和树冠火燃烧比例的增加而增加,这种关系在距离房屋 1 公里处的森林中最为强烈。如果 50 米内有其他房屋,或者房屋位于斜坡上,房屋更有可能被摧毁。包含这些变量的模型可以以 72%的准确率预测房屋损毁。我们的研究结果有三个重要意义:i)管理措施以改变树冠火的发生将有效减少房屋损毁;ii)在某些情况下,这种管理措施可能需要在距离房屋 1 公里的范围内实施(比目前使用的范围大得多);iii)高密度的房屋可能会增加损失的风险。鉴于这种管理区域可能具有较大的宽度以及附近房屋的危害,集中精力对建筑物进行改造以降低其脆弱性可能更为明智。