Probabilities

Saturday morning I was scheduled to fly out to New Delhi through Vienna with Austrian Airlines. As thousands of others over the last days, I have seen my travel plans dashed by the eruption of the Eyjafjallajökull volcano on Iceland.

As one of the others in our group sarcastically noted on Facebook, what are the odds?

Low one is tempted to think. Yet, considering how seismic active many parts of the planet are, it rather seems as if we have been exceptionally lucky in the past. Civil aviation is a relatively new phenomenon and though there are historic examples (such as the two weeks of airspace closure following the 1980 Mount St Helens eruption in Washington), the scale of events this time is obviously unprecedented. But with more than 60 000 flight movements on any given day worldwide we should perhaps not be that surprised that we eventually ran into something like this.

Thinking further about it, I am reminded of an interesting chapter by Milan Ćirković on observation selection effects in a recent book on global catastrophic risks. The basic idea is straightforward enough: past experience may not be the best guide when it comes to estimating the frequency of really big catastrophes (such as major asteroid impacts or bursts of supernova gamma rays) simply because we would not be here doing the estimation if such catastrophes had happened in the past...