Re: "Framing" natural peril probabilities

From: William Wilber
Affiliation:
Address: cen05515@centurytel.net
Date: 04 Sep 2000
Time: 20:56:05

Comments

I recently saw an analysis of celestial body diameter vs. frequency of impact on Earth done over 5 orders of magnitude in diameter and 8 orders of magnitude in time. It was done using a fractal (non-linear) statistical approach.

I think you could model your disasters using such a nonlinear model, perhaps Paretian or Weibull. Perhaps Weibull would be best for earthquakes, since they represent a "breakdown" statistic. Epidemics might be a percolation phenomenon.

Two books on the topic are:

"Fractals, Chaos, and Power Laws" Manfred Schroeder, ed., W. H. Freeman, 1991

"Complexity: Metaphors, Models, and Reality", G. A. Cowan et al, ed., Perseus Books, 1994

Now there are many.