Distribution of number of years between hurricane landfalls in a given area

[Team Ghost]

When people say a hurricane landfall is "due" to occur in a specific area, the implicit assumption being made is that a hurricane not making landfall in a particular area during a given year increases the likelihood of a hurricane making landfall in that same area next year. More formally, if L is a random variable that is the number of years between hurricane landfalls in a given area, L does not follow a memoryless distribution. (Since L is a discrete random variable, this is equivalent to saying L does not follow a geometric distribution.)

I think it would be fruitful to investigate the distribution of years between hurricane landfalls in a given area. Naïvely, for a given area whose probability of experiencing a hurricane landfall in an unspecified year is p, I would expect the distribution of years between hurricane landfalls in that area roughly to follow Geom(p). I think the markedly non-memory nature of El Niño Southern Oscillation, with El Niños being more likely after double-dip La Niñas and La Niñas being more likely after El Niños, could cause the distributions to deviate slightly from the expected Geom(p) because of ENSO's influence on steering patterns; however, I expect broad-scale climate patterns to be fairly minor factors in totally of all influences on storms' steering.

[tolakram]

The 'due' talk is simply gamblers fallacy. In reality, regardless of past events, we don't know the odds and they never change. We don't even know if we're dealing with 5 sided dice or 100 sided dice, and how many dice.

A better question rather than look at distribution is to look at patterns present during a landfall somewhere and then how often those conditions persist. Even with that knowledge though, when a condition is present the opportunity for a storm to form also has to be present. Example: Conditions mimic year X location experienced last landfall but no waves or other points of genesis are available to form a storm. Good luck managing all those variables.