An average number of Earthquakes is the result of some years being above average and some below.
If there are 10 major quakes per year on average, one would not predict 10 major quakes this year.
The past years might go like this example:
5, 15, 2, 18, 11, 9, 8,12, 14, 6 and average out to 10 major quakes per year, but having 14 major quakes would be normal… Not average, but normal. And in no year listed do we see 10 quakes.
Stats are funny like that.
So how far from the average (the mean?) does the number of strong earthquakes per year typically wander? If the average is ten a year for example, how rare is a year with thirty? Is the number of earthquakes as random as thermal noise, or does it follow some distribution? Does the number of quakes tend to flatten out over a long enough period of measurement (decades? centuries?) or will there still be notable spikes?
There’s not much deviation on the number of strong earthquakes per year (this observation on the local level was sort of the gist of Richter and Gutenberg’s original work that lead to the scale, and I would presume it is applicable to the global situation), but there can be plenty of variation on where those strong earthquakes occur. So assuming that it isn’t just entirely a confirmation bias issue, it’s quite possible that we’ve just had a run of bad luck lately and a lot of those strong earthquakes have happened to occur near populated areas.
Here’s what happened. The year started off with an unusually devastating earthquake in Haiti, which received worldwide, round-the-clock attention from the national media for over a month. As a result, you are more likely to notice all the other earthquakes which have occurred since then, and the media is more likely to spend more time focusing on them.
In other words, it’s a clear-cut case of confirmation bias.
Yep. And the Haiti 'quake was big but wasn’t that big, it only devastated the country because they have such weak structures with no standards or building codes. If the same 7.0 had hit, say, Japan, it would be just another day of the week and you would have never heard about it.
Well, they had a 7.0 centered at Okinawa just over a month ago that didn’t cause any damage at all. Other factors besides the magnitude, like how far below the surface the earthquake occurs can affect the damage caused as well.
Earthquakes can trigger other earthquakes. When they happen in the same place as the original quake, they’re called aftershocks. But we’ve learned in recent years that large earthquakes can trigger other earthquakes even in distant places.
Of course, just because that can happen doesn’t mean that that’s what’s happening now. These earthquakes could all be happening coincidentally at the same time.
A meteorologist wouldn’t necessarily know more about earthquakes than any random person off the street. If a geologist or seismologist says something like this, I’ll sit up and pay attention.
If they are truly random events, it’s a Poisson process - the number occurring in a given interval should be Poisson distributed, the times between them Erlang distributed. “Random” events or “random” spacial distributions are far lumpier than the average persons intuition suggests they ought to be. If you take a piece of paper and ask somebody to simulate a random distribution of 100 dots on it by hand, they will nearly always produce something far too uniform. Expecting “random” to be more uniform than it is is at the root of a lot of misconceptions about the significance of a spate of random occurrences.
Here’s somebody’s simulation of the Poisson distribution if you wish to play with it. Set the expected number of occurrences per trial to 10, and see what you get: