Probabilistic risk assessment for wildfires

被引:42
作者
Brillinger, D. R.
Preisler, H. K.
Benoit, J. W.
机构
[1] Univ Calif Berkeley, Dept Stat, Berkeley, CA 94720 USA
[2] US Forest Serv, USDA, Pacific SW Res Stn, Albany, CA 94710 USA
[3] US Forest Serv, USDA, Pacific SW Res Stn, Riverside, CA 92507 USA
关键词
biased sampling; false discovery rate; forest fires; generalized mixed model; penalized quasi-likelihood; risk;
D O I
10.1002/env.768
中图分类号
X [环境科学、安全科学];
学科分类号
08 ; 0830 ;
摘要
Forest fires are an important societal problem. They cause extensive damage and substantial funds are spent preparing for and fighting them. This work develops a stochastic model useful for probabilistic risk assessment, specifically to estimate chances of fires at a future time given explanatory variables. Questions of interest include: Are random effects needed in the risk model? and if yes, How is the analysis to be implemented? An exploratory data analysis approach is taken using both fixed and random effects models for data concerning the Federal Lands in the state of California during the period 2000-2003. Published in 2006 by John Wiley & Sons, Ltd.
引用
收藏
页码:623 / 633
页数:11
相关论文
共 13 条