We are constantly hearing reports on the nightly news about natural disasters—droughts in Texas, hurricanes in Florida, floods in California, and so on. We often hear that one of these was the “worst...

We are constantly hearing reports on the nightly news about natural disasters—droughts in Texas, hurricanes in Florida, floods in California, and so on. We often hear that one of these was the “worst in over 30 years,” or some such statement. This was especially the case in the last half of 2005, with the terrible hurricanes in the southern United States and the deadly earthquake in Pakistan.

Are natural disasters getting worse these days, or does it just appear so? How might you use simulation to answer this question? Here is one possible approach. Imagine that there are N areas of the country (or the world) that tend to have, to some extent, various types of weather phenomena each year. For example, hurricanes are always a potential problem for Florida.


You might model the severity of the problem for any area in any year by a normally distributed random number with mean 0 and standard deviation 1, where negative values are interpreted as mild years and positive values are interpreted as severe years. (We suggest the normal distribution, but other distributions can be used as well.) Then you could simulate such values for all areas over a period of many years and keep track, say, of whether any of the areas have worse conditions in the current year than they have had in the past 30 years. What might you keep track of? How might you interpret your results?