HR: 15:10h
AN: NG12A-07 INVITED     [PDF]
TI: Tails of Natural Hazards
AU: * Malamud, B D
EM: bruce@malamud.com
AF: Environmental Monitoring and Modeling Research Group, Department of Geography, King's College London, Strand, London, WC2R 2LS United Kingdom
AB: There is increasing evidence that many natural hazards satisfy power-law frequency-size statistics. Examples include earthquakes, volcanic eruptions, landslides, snow avalanches, forest and wildfires, meteorite impacts, and possibly floods. Although power-law (fat-tail) distributions are commonly associated with the frequency-size distribution of earthquakes, the frequency-size statistics of many other natural hazards are presently associated (e.g. by government agencies and reinsurance companies) with distributions that are more thin-tailed. The occurrence risk for large and very-large events using power-law frequency-size distributions is often much more conservative, with a greater chance of a large event occurring in a given period of time, compared to thinner tail distributions. One potential explanation for the frequent occurrence of power-law (fractal) frequency-size distributions among natural hazards lies in cellular-automata models, and their association with self-organized criticality and inverse cascades. The power-law behavior of the sandpile cellular-automata model has been associated by some with landslides, the forest-fire model with actual forest fires, and the slider-block model with earthquakes. A relatively simple inverse-cascade of metastable regions can explain the behavior of both models and the actual natural hazards. Metastable regions grow by coalescence and are lost in `avalanches'. However, the losses are dominated by the largest events and have little influence on the inverse cascade of metastable region coalescence. This inverse cascade of metastable regions is self-similar and the number-area statistics are power-law. Although the theoretical explanations are still being debated, the increasing evidence for power-law statistics means that government agencies and reinsurance companies should include this much more conservative frequency-size distribution when calculating the occurrence risk of large natural hazards.
DE: 3200 MATHEMATICAL GEOPHYSICS (New field)
DE: 3250 Fractals and multifractals
DE: 6309 Decision making under uncertainty
SC: Nonlinear Geophysics [NG]
MN: 2003 Fall Meeting