Ramblings about what I encounter within the realm of the geosciences, as well as the occasional rant about nonsense.

25 March 2011

I liked his books...

A few days ago, I heard that Simon Winchester (he of "The Map that Changed the World", "Krakatoa", and "Crack in the Edge of the Word" fame) had published an editorial in Newsweek regarding the recent Japan Earthquake. The tweet that brought it to my attention was not the most glowing review of an editorial I have ever seen, so I was curious about what Winchester was saying.

At the end, I was not overly impressed. The primary thesis of Winchester's editorial deals with earthquake clustering. Earthquake clustering is the concept that a large earthquake associated with one tectonic plate will cause other large earthquakes on the same plate (or elsewhere on the planet). However, he provides no data to back up this assertion. He only uses cherry-picked anecdotal evidence (which is not data, it may be the start of a hypothesis, but it isn't data). With nothing else to support his argument, it is just unconvincing.

If that is all the essay said, that would be fine. However, Winchester uses this concept to argue that this means there will be an earthquake along the West Coast of North America in the very near future. This is just sensationalism, and it has been covered thoroughly by several other geoblogs (notably Highly Allocthonous and Geotripper) and CNN. I recommend looking through them, since I won't be going into much analysis about this.

Last night, Winchester responded to his critics. The Daily Beast article is rather weak in my opinion. It just presents more cherry picked anecdotes and then claims it is obvious to everybody except for geophysicists (and I'd assume geologists too). However, this is a foolish argument to make. Science is not necessarily intuitive.

If science was always intuitive, Aristotle and Plato would still be the top dogs as far as understanding the natural world go. Heavy objects will fall proportionally faster than light objects. There would only be four elements (Earth, Wind, Fire, Water) and a fifth "essence" making up the celestial spheres (the origin of the word "quintessential"). The Earth would also be the center of the universe and the continents would not be moving about the planet.

These were all blatantly obvious to the majority of our population, but when people actually started to collect evidence, it was realized that a counter intuitive explanation is what made sense in light of the data collected. This is why scientists go to great lengths to collect data and build massive datasets to demonstrate the validity of a hypothesis.

He also justifies his argument by saying "All that can be said with certainty is that they are more likely to break tomorrow than yesterday". This justification is meaningless as it is true of any natural phenomenon given enough time. With the same confidence as Winchester is displaying with regards to earthquake prediction, I can claim that a giant asteroid is going to collide with Earth. I wish I could take credit for this concept, but it was already advocated by Hsu and his paper "Catastrophic Extinctions and the Inevitability of the Improbable (1989)" (a brilliant, must-read paper).

I will conclude by saying one thing that has bothered me about this from the start. Where is the data to support this? That is an easy one. There are large stores of data available to individuals who wish to study earthquakes. Hell you can follow earthquakes as they happen via twitter (@quakenotices) and practically any mobile device these days. More to the point, it is all aggregated within the USGS and freely available to the public. So why did Winchester not provide any data supporting his point?

We have amassed plenty of data, all of it begging to be analyzed. There is no reason to think that Winchester is wrong advocating earthquake clustering (unless somebody has done the analysis and I just don't know about it). However, without the data to back up his hypothesis, he is not going to convince anybody within the scientific community.

I still like his books, but this is just sloppy thinking. Even worse, it is irresponsible.

Works Cited
Hsu, K. J., 1989, Catastrophic Extinctions and the Inevitability of the Improbable: Journal of the Geological Society, v. 146, no. 5, p. 749-754.

Edit: Even Better, just read this note on Facebook by Dr. Rowe which is the emails sent back and forth between herself and Simon Winchester regarding his articles. It is very good stuff. She also clarified a few things I had questions about myself (most notably, hadn't this already been thought of?). Thanks to Callan and Brian for tweeting this.

1 comment:

BrianR said...

you're absolutely right about the Hsu paper ... another one that goes very nicely w/ it is the Gretener (1967?) paper called 'Significance of the Rare Event'.


All the Latin on this page is from my vague recollections from High School. There are mistakes in the text. I just was trying to get the point across

Between Los Alamos,NM and White Rock, NM

Between Los Alamos,NM and White Rock, NM
The photo of the travertine spring was taken in the small opening in the center of the image.

Lectio Liber