Showing posts with label probability. Show all posts
Showing posts with label probability. Show all posts

Friday, January 29, 2010

NY Times article on the Haiti quake

The New York Times had the expected interesting article on the Haiti earthquake in its Tuesday science supplement.

I won't copy the article here, but some of the interesting points:
  • proximity to the boundaries of the Caribbean plate means natural hazards are expected
  • there are two major parallel faults - the Enriquillo and the Septentrional - crossing Haiti in an east-west direction
  • the earthquake rupture may actually increase the stress and earthquake probability on adjacent segments of the Enriquillo fault
  • the aftershock sequence has been rather active
  • there are NO permanent seismograph stations in Haiti
I'd say the article is slightly misleading, although not really wrong, in one respect.  It is said that one might have expected a failure on the Septentrional fault, rather than the one that occurred on the Enriquillo. An analogy is that if one were to flip a coin twice, one might "expect" the outcome to be one head and one tails, rather than two heads (or two tails). Indeed, there would be a 50% change of getting one head and one tails, and a 25% chance of getting either two heads, or two tails.  Ok, but it would neither be rare nor completely unexpected to get one of the other outcomes. So much of science, including earthquake forecasting, has to do with probabilities, but that is not a concept most people are very comfortable with.

The graphic below, from the article, is quite good. Makes me glad that when I worked in western Jamaica in 1999, on the westward extension of the same fault that was responsible for the Haiti quake, all was seismically quiet, although a major quake hit the area in 1957.

Thursday, October 8, 2009

Three magnitude 7's in one day

According to the list of earthquakes from the USGS, there were three magnitude 7 quakes on Oct. 7.

These all occurred near Vanuatu, so are probably not independent. But if we take the annual global magnitude statistics, there are typically 17 quakes per year with magnitudes above 7. What are the chances of getting three independent magnitude 7's on the same day?

The chance of getting one is 17/365 = 4.6%. I think (correct my joint probabilities, please) that the chances of three in one day would then be this probability cubed, or 1 chnce in 10,000,

If you are more energetic than I am, go back and check when was the last time this happened.

Or, would it make more sense to simply consider the probability of a magnitude 8 quake, which would release more energy than three magnitude 7's anyway?

Image from http://ocw.mit.edu