I’ve been following earth science closely for many years through the news and in the literature. One thing I can say confidently is that it is both very human and inhuman. The human side is obvious: geoscientists are some of the most interesting (and interested) people I know. But doing science properly means quashing some of our most human traits as we subject our ideas to cold statistical tests. Case in point: earthquake prediction studies.
Humans have a wonderful ability to spot patterns in seemingly random information. We acquired the knack long ago because it helped us spot zebras in the savanna and tigers in the grass. It’s amusing because it lets us see faces in clouds. In science, our pattern-seeking ability is both a tool and a handicap: it leads us to exciting hunches, but glues us to wrong ideas despite our best efforts. The occurrence of earthquakes is so troubling and mysterious that it takes a great effort to stifle our human reactions and pay attention to what the cold numbers and their unintuitive statistics say. I take as today’s topic the relationship between earthquakes and sunspots.
The sun-Earth interface is an exceptionally energetic region, extending over thousands of kilometers from the upper atmosphere into outer space. It conducts electricity and generates complex magnetic disturbances. It’s as changeable and chaotic as the static heard between radio stations or the shimmering displays of the aurora.
It stands to reason that such a sensitive region would respond to changes from below the Earth’s surface just as it does to changes from above. You would think that the release of earthquake energy, in the electrically conducting deep crust, would have effects on this region too.
But “It stands to reason…” and “You would think that…” are hunches, only the beginning of science. Next comes the inhuman part: holding off on coming to conclusions and running the numbers as carefully as we know how. A trickle of scientific papers have noted apparent relationships between various measures of solar activity and earthquakes. But a recent critique of these papers, published in March in Geophysical Research Letters, noted that all of them fell short in one way or another: they used partial sets of data or focused on one region, for instance. They went beyond “It stands to reason,” but only as far as finding an anomaly, something strangethat is to say, “Hmm, look at this.”
The paper’s authors went on to do a textbook, double-checked study comparing the solar records to the earthquake catalogs, including statistical checks and found nothing that chance alone couldn’t explain just as well. (See my more detailed explanation here.)
However, most of us find it terribly tempting to go from “Hmm, look at this” to “Hey, look at that.” And once someone who seems to know what they’re talking about does that, many more people will take assertion as certainty. For a good example, consider the BBC’s weather blogger Paul Hudson. He’s a trained meteorologist who can explain scientific thinking well. Last month, for instance, you can see him deftly handling a common question among the British public: isn’t our summer weather odd? No, he answers, it’s a phenomenon well-known among his fellow scientists: the North Atlantic Oscillation.
But then go back on his blog to the day of the magnitude-9 Tohoku earthquake in March 2011. Not being a seismologist, Hudson fell back on human habits: he remarked upon recent sunspot activity, cited a few unnamed papers, and quoted a fellow opinionater in support of his hunch. On this subject, he scored no better than the average layman. Indeed, I think that posting his off-the-cuff musings on the very day of that massive earthquake under the headline “Is solar activity to blame for Japan’s massive earthquake?” was not responsible journalism. Fortunately, the Betteridge-Marr law of headlines is our friend: “Any headline which ends in a question mark can be answered by the word ‘no.'”