Terrifying tremors

31% have the events have occurred in the last 7% or so of the time represented. That's odd.

@Harpua24

That's probably because modern earthquake measurement technology has improved so much, as well as a surge in human population living in previously unsettled areas.

While the disaster is enormous, the good news in this is the relatively light property damage and casualties due to the quake itself. Obviously no building can withstand a tsunami, but Japan has put a lot of effort and money into earthquake-resistant civil engineering, and it paid off today. I have yet to see a picture of a building that collapsed due to the shaking, despite severe damage to the contents.

@D.Sherman
Don'y rush to conclusions just yet. We know from our experience at Kobe that the full extent of the damage won't become clear until 48 hours after the quake itself. I am already hearing reports of entire coastal communities washed away, and bodies washed up to 10km inshore...
Latest estimates already put more than 1000 missing, with more than 200 confirmed dead, and these estimates are still going up by the hour...

Since 1900, there has been a mega-quake somewhere in the world about once every 7 years... since 2004, there has been a mega-quake roughly once a year. Something is bubbling deep inside the Earth.

In the for-what-its-worth department:

I've experienced a couple in the 5 to 6.5 range, which were big enough to wake me up and have me running for the exits. And, I imagine I was on the far periphery of the impact areas.

So, given that Richter is on a log-scale, those listed are serious big-a-- quakes.

My comment might appear off the subject but an interesting fact that comes out of this graph is the correlation between GDP+level of development and the occurrence of earthquakes within a country. Indeed, Chile,Indonesia and Russia have encountered the most intensive earthquakes at repeated times. Could The Economist actually measure the impact of such disasters on the economies of theses countries (infrastructures, industries, costs of rebuilding but also the level of investments etc..) and try to draw some inferences?

@Harpua24

must be global warming.

@Harpua24 - perhaps it's cyclical. The 1980s and 90s are conspicuous by their absence. But a couple occur in 1920s-30s; in 1950s-60s and then 2000s.

I could be wrong, but I think the Japanese earthquake was upgraded to 9.1

Sometime in the 90s, The Economist printed a story about overdue movement in the plates that caused the Tokyo earthquake in the 1920s. Was this quake a movement between those plates, or is there another one lurking out there?

I have attempted to improve on your graph by showing the magnitude scale, comparing with Haiti and New Zealand and introducing the number of deaths of each earthquake. See it on Twitter.

Why don't you use a trigonometric scale since the levels are based on that?

The big earthquake in Valdivia was also one of the longest ones. It last for 10 minutes and killed 2,000 people.

The Japan earthquake was upgraded to 9,1 Richter.

It's trivial, but nowadays we're using the Moment Magnitude Scale (MMS) rather than the Richter scale which so many people (including journalists) are still quoting.

http://en.wikipedia.org/wiki/Moment_magnitude_scale

Regarding an apparent increase in earthquake frequency, listen to the clicks on a Geiger counter some time. There are spells with few or no clicks, and spells with clusters of clicks. Our brains are wired to search for patterns, since that is vital to many aspects of survival. The Geiger counter clicks seem to have a pattern or come in clusters, but in fact, nuclear decay is an absolutely completely random process. The time at which a given atom of a radioiosotope "decides" to decay has absolutely nothing to do with when any other atom near it decayed or with how long that atom has been in existence.

Likewise, a completely random rate of occurrence of large earthquakes tends to look to our pattern-sensitive brains as if "there's something else going on here", when in fact it's purely random.

One must keep in mind that these 8.5+ magintude events are extremely rare in the overall picture of earthquake frequency, so that apparent clustering in time is not surprising nor does it suggest a fundamental change in earthquake dynamics. The statistics of extremely rare events are not smoothly distributed.