This paper examines trends in the impacts of disasters at the world scale and proposes explanations for them. It is concluded that statistics on natural hazards are questionable, in that changes in recording methods h...This paper examines trends in the impacts of disasters at the world scale and proposes explanations for them. It is concluded that statistics on natural hazards are questionable, in that changes in recording methods have artificially inflated the number of disasters and the seriousness of their impacts. Next, the paper examines the symbolic interpretation of disasters at various scales of analysis in terms of human cultures. It further applies this approach to the analysis of natural hazards using information technology methods such as GIS （Geographic Information Systems）. Returning to the global scale, deficiencies and inequalities in the world disaster relief system are considered. Thereafter, the paper investigates how global economic imbalances are thrown into sharp relief by catastrophes, with examples from the record of past earthquakes, landslides and floods. Finally, it considers the prospects for a major change in the direction of world policy on disasters and concludes that this will only occur if the international financial system is forced to adjust to a ＂super-disaster＂ event, which could possibly be seismic, volcanic or nuclear in origin.展开更多
Let {Tn } be a renewal process in R＋ representing the successive arrival times of some natural events. We studied this process by using a record process approach under the assumption that the interarrival times T,, =...Let {Tn } be a renewal process in R＋ representing the successive arrival times of some natural events. We studied this process by using a record process approach under the assumption that the interarrival times T,, = Tn, - Ta-1, n = 1, 2...are exponentially i.i.d （independent and identically distributed）. The goal is to test that the first observed events are sporadic events. For testing the hypothesis ＂sporadic＂ we used the non-parametric test based on the probability distribution of the statistic of the number of records N, among{Xx }k-1= where Xk = （ΔTk）-1. We showed that it is independent of the cumulative distribution of the observations and that it is exactly calculated for each n. We illustrated this statistic on a simulated trajectory and we compared it with descriptive smoothing methods. We studied an application to a data set as storms in France and US.展开更多
文摘This paper examines trends in the impacts of disasters at the world scale and proposes explanations for them. It is concluded that statistics on natural hazards are questionable, in that changes in recording methods have artificially inflated the number of disasters and the seriousness of their impacts. Next, the paper examines the symbolic interpretation of disasters at various scales of analysis in terms of human cultures. It further applies this approach to the analysis of natural hazards using information technology methods such as GIS （Geographic Information Systems）. Returning to the global scale, deficiencies and inequalities in the world disaster relief system are considered. Thereafter, the paper investigates how global economic imbalances are thrown into sharp relief by catastrophes, with examples from the record of past earthquakes, landslides and floods. Finally, it considers the prospects for a major change in the direction of world policy on disasters and concludes that this will only occur if the international financial system is forced to adjust to a ＂super-disaster＂ event, which could possibly be seismic, volcanic or nuclear in origin.
文摘Let {Tn } be a renewal process in R＋ representing the successive arrival times of some natural events. We studied this process by using a record process approach under the assumption that the interarrival times T,, = Tn, - Ta-1, n = 1, 2...are exponentially i.i.d （independent and identically distributed）. The goal is to test that the first observed events are sporadic events. For testing the hypothesis ＂sporadic＂ we used the non-parametric test based on the probability distribution of the statistic of the number of records N, among{Xx }k-1= where Xk = （ΔTk）-1. We showed that it is independent of the cumulative distribution of the observations and that it is exactly calculated for each n. We illustrated this statistic on a simulated trajectory and we compared it with descriptive smoothing methods. We studied an application to a data set as storms in France and US.
