Fairy Tales on Shaky Ground – Scientific understanding and the Italian court system

On the 6th of April, 2009, a 6.3 magnitude earthquake centered 9.46 km beneath the Abruzzo region in central Italy devastated the city of L’Aquila, ripping the historic heart out of the city and killing 309 people. While the physical scars from this tragedy are fading, cultural aftershocks are still rippling through the scientific community, and reached a peak unexpected by many this week with the conviction of six Italian scientists and a former government official for involuntary manslaughter.

The seven – members of the Italian National Commission for the Forecast and Prevention of Major Risks at the time of the earthquake – were sentenced to 6 years in prison, and ordered to pay court costs and damages amounting to some 7.8 million Euros.

Their conviction was based principally on a number of statements made six days before the damaging tremor, downplaying the likelihood of a major earthquake. Given the manifest impossibility of reliably predicting Earthquake occurrence with our present understanding of seismic processes, the legal precedent for this is presumably drawn from the Brothers Grimm, with imprisonment the prescribed punishment for failing to spin straw into gold.

Earthquakes are caused by the release of energy as fractures propagate through rocks. They are focused in particular regions where creeping deformation deeper in Earth leads to the build up of stress in the more brittle rocks near the surface. Once a fracture has occurred, the crack remains a discontinuity – a weakness – and tends be a site of further failures in the future.

If you live near such an existing fracture – or fault – the likelihood of experiencing an earthquake increases dramatically. This much we can say with confidence. Certainly, in the long view, no-one can claim to be surprised by the damage wrought on L’Aquila, given the city is built on an ancient lake bed known to provide a geological framework that amplifies the local effects of seismic waves, and has been devastated by earthquakes on no fewer than seven historical occasions, in 1315, 1349, 1452, 1501, 1646, 1703, and 1706.

But when will the next Earthquake happen? Now this – the 7.8 million Euro question – is the holy grail of seismic hazard research, and to date, there is no answer. Anybody who tells you differently is probably emailing from Nigeria to offer you ‘the investment opportunity of a lifetime’.

Despite all the work by teams of dedicated and sometimes brilliant researchers over the past century or so, all the collection of data and analyzing of patterns – even for the San Andreas Fault in California, probably the most intensely monitored fault zone in the world – no distinctive and reliable precursor patterns for major earthquakes have ever been recognised.

To hold someone responsible for failing to predict an earthquake on the basis of preceding activity makes all the statistical sense of having your first tip on Melbourne Cup day romp home, doubling down your house on the trifecta in the next race, and then suing your bookie when the horses fail to place.

So how can this miscarriage of justice have occurred? Why weren’t the charges thrown out at the earliest opportunity?

Fundamentally, as I wrote last week (Three Monkeys, Ten Minutes – Scientists and the Importance of Communication Skills – WordPress 18/10/2012):

“Society is complex, and people hold views for all manner of reasons – personal, cultural, logical, or religious, among others. We [as scientists] do not have to share those views, but we do need to appreciate and respect their reality”

When I wrote those words, I hadn’t expected to be confronted by such a glaring (and dark) example of this relationship at work quite so quickly. Scientifically, the question of earthquake prediction doesn’t even get off the ground, but to a broader population un-tutored in statistics and the language of scientific uncertainty, a population here stung by a great tragedy and searching for someone to blame – a sadly common human trait – the Committee’s statements painted them with a target.

People are incredibly good at recognising patterns. This, as much as anything, is the key to our astonishing success as a species. Unfortunately, the flip side to this is that we look for – and expect to see – patterns even when they are not there. This makes us very bad at evaluating the true risk of rare events.

James ‘the Amazing’ Randi is conducting a long term demonstration of this phenomenon. Every morning he writes on a card “I, James Randi, will die today”, which he then dates, signs, and keeps in his pocket in the knowledge that it will one day (may it be far in the future) be a fitting final demonstration of how apparent correlation can be manipulated to lead our minds astray.

Richard Feynmann related a similar story in his memoir “Surely you’re joking, Mr Feynmann” – where he writes of hearing the phone ring in his University dormitory and having a sudden premonition that his grandmother had died. She hadn’t. The phone wasn’t even for Feynmann, and his grandmother continued in rude good health for some time to come. The point is, we have such thoughts and intuitions all the time – for the most part, they don’t turn out to be correct, but occasionally the fates line up. When they do, the glorious pattern-seeking engines that are our brains get a kick of reinforcing dopamine to say ‘job well done’ and we forget about the 999 previous times it hasn’t worked and start to see a correlation.

If you live near a fault line, you will, inevitably, experience earthquakes. Sometimes big ones, often small ones. Sometimes a large one will be preceded by small ones. But usually not. The stochastic patterns – one earthquake here, two the next week, none for six months – have no significance.

There is a real tragedy at L’Aquila, and there are people who should be held to account. But they are not the scientists who gave an accurate representation of the processes at work beneath the town, and the statistical meaninglessness of looking for patterns in the tea leaves of local seismic activity. Rather, the guilty parties – those who should have known better – are the officials and engineers who built structures – schools, gymnasiums, dormitories – in the city that were not designed or constructed to withstand the well known and historically proven earthquake risk.

So, we find ourselves at the end of act 1 in our re-imagining of Rumplestiltskin – the Miller’s daughter has proven unable to spin her straw into gold and the King is about to imprison her in the highest room of the tallest tower. Perhaps in act 2, Uri Geller will step up to the title role and offer to solve her problems by magic in exchange for first authorship on the resulting scientific paper.

Advertisements

2 thoughts on “Fairy Tales on Shaky Ground – Scientific understanding and the Italian court system

  1. Pingback: A Nice Glass of Montepulciano: Risk, lies, and politics in the L’Aquila Earthquake | rocky subjects

  2. Dr Geoff Batt Post author

    Although not exactly a Damascene conversion, my views on the L’Aquila case have evolved significantly since I posted this original essay. For a more nuanced opinion on the L’Aquila trial and its implications, check out my more recent post on the subject – ‘A Nice Glass of Montepulciano: Risk, lies, and politics in the L’Aquila Earthquake’ – WordPress 22/5/2013

    Reply

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out / Change )

Twitter picture

You are commenting using your Twitter account. Log Out / Change )

Facebook photo

You are commenting using your Facebook account. Log Out / Change )

Google+ photo

You are commenting using your Google+ account. Log Out / Change )

Connecting to %s