Learning from L’Aquila – and why Canterbury is different

by Dr Mark Quigley, Senior Lecturer in Tectonics and Geomorphology at the University of Canterbury and the 2011 Prime Minister’s Science Communicator’s Prize winer

Six scientists and one government official have been found guilty of manslaughter and sentenced to six years in prison for how they assessed and communicated risk prior to the magnitude 6.3 L’Aquila earthquake that killed 309 people on April 6, 2009.

The verdict was not due to their failure to ‘predict’ the earthquake, which most scientists generally agree is not possible with our current knowledge of precursory phenomena.

Rather, the prosecutor reasoned that ‘inadequate’ risk assessment and scientifically incorrect messages were given in public statements prior to the earthquake that ultimately contributed to a higher death toll when the earthquake eventuated. After months of small (magnitude 3.5 to 4.1) earthquakes prior to the L’Aquila earthquake, Bernardo De Bernardinis (then vice-president of Italy’s Civil Protection department) stated that “the scientific community tells me there is no danger because there is an ongoing discharge of energy…” with the inference that these small earthquakes were reducing the possibility of a major quake. It is unclear how representative this statement was of the group. Certainly a statement from one member of the group at this time, Enzo Boschi, then-president of Italy’s National Institute of Geophysics and Volcanology (INGV) in Rome, was both well-balanced and informative, “It is unlikely that an earthquake like the one in 1703 [a devastating earthquake that previously hit L’Aquila] could occur in the short term, but the possibility cannot be totally excluded.”

Many seismologists consider that an increase in the frequency of small to medium sized earthquakes increases the chances of a large earthquake based on long-established fundamental relationships in seismology. However this effect is relatively small in terms of an absolute probability and does not improve the game of earthquake prediction. For instance, in the past 60 years in Italy, only six of 26 major earthquakes have been preceded by foreshocks and many earthquake swarms have occurred without subsequent large earthquakes. Italian scientists concluded that a medium-sized shock in a swarm forecasts a major event within several days only about 2% of the time. If they had issued a specific warning that a major earthquake was coming in L’Aquila prior to this event, they would have had a 98% chance of being wrong.

The L’Aquila earthquake occurred on a previously identified and well-monitored fault zone in an area of elevated historical seismicity that was recognized as one of Italy’s most seismically dangerous regions. From that perspective, the L’Aquila earthquake was no surprise, and the possibility of an earthquake of its magnitude occurring in this region following months of seismicity should not have been publically dismissed. However, seismic hazard maps of this area were publically available prior to the L’Aquila quake, and the major earthquake in 1703 was well recorded in human history.

A clear lesson here is that the general public should be made aware of all possible scenarios within an earthquake sequence, regardless of how small the absolute probability of certain scenarios may seem. The prison sentence for these scientists will seem overly harsh to most of us, given the highly-stressful and complex scientific, societal, emotional, and political environment that develops during an earthquake sequence. It is unclear to me how representative the public statements were of the scientist’s true views, and speaking through political figures and the press always leaves open the possibilities for incomplete representation. Given the long-recognized seismic hazard of this region, one also wonders whether appropriate building codes were applied and enforced, and whether these scientists would have been ‘off-the-hook’ if buildings had coped better with the seismic shaking of L’Aquila’s quake.

The consequences of this indictment on the scientific community remain to be seen, but a clear lesson is that the public needs to be made aware of ‘low-probability, high-consequence’ events regardless of how unpopular and/or distressing these sentiments might be. That said, it takes a ton of courage for scientists to speak openly about low probability scenarios, particularly if these comments are used to accuse scientists of scare-mongering, and/or have detrimental impacts on earthquake recovery, such as decreasing investor and re-insurer confidence and increasing stress levels of local residents.
Applicability to the Canterbury earthquake sequence and beyond

Earthquake scientists around New Zealand took a long, hard look in the mirror after the February 22, 2011 Christchurch earthquake that claimed 185 lives. Once the sorrow of the event subsided to a level where I could refocus, I went through media interviews and public talks in spare moments where I was not involved in the scientific response to this earthquake. It was clear that the public were made aware that a magnitude 6 aftershock was possible by scientists throughout New Zealand; that potential scenarios could include a shallow earthquake in the region east of the Greendale Fault, and that aftershocks beneath Banks Peninsula suggested elevated crustal stresses in that area were being partially accommodated by slip on northeast oriented faults.

In an article written on September 8th, 2010, I stated that “My optimistic guess is that we are unlikely to get an aftershock as big as a Mw 6” based on aftershock data from what I felt were similar earthquake sequences in Haiti and Mexico, but finished this with, “[but] we could get a bigger one months from now”. In retrospect I would have liked to have the former statement back, but in truth this was an example of locally-based optimism at a time of heightened public anxiety. I do feel that scientists have a right to voice well-grounded hypotheses, just as they have a ‘right-to-be-wrong’, providing the justification for said hypotheses and the range of possibilities are publically presented. In truth, the last decade has thrown up many seismic surprises, not least the magnitude 9.0 Tohoku earthquake in Japan that was preceded by a magnitude 7.2 ‘foreshock’, affirming that we still have much to learn about earthquake behaviour on our planet.

Having been through a catastrophic earthquake sequence beneath one of its major population centres, the New Zealand earthquake science community is better placed than before to answer the needs of the public. Data sharing with the general public as quickly as possible via all media avenues is being increasing recognized as an obligatory responsibility by many practicing scientists, but significant barriers to this process remain. Science does not move at the pace of the media, and science that requires substantial peer review may be less interesting to the broader public by the time it has undergone a lengthy review process. Furthermore, in order to make money from expensive journal subscriptions, many publishers do not allow the authors of scientific articles to disseminate their original work publically. This is inappropriate in a post-disaster environment in my view, where the affected public deserve the right to freely scrutinize the raw data that has so often been obtained with public funding.

Looking forward, additional improvements can be made in the way we communicate earthquake science. Published, publically available statements of ‘absolute probabilities’ such as ‘there is a 9% chance of a magnitude 6 to 6.4 earthquake occurring in the Canterbury aftershock region in the next 12 months’ should be contextualized against the ‘probability increase relative to pre-mainshock probabilities’ with statements such as ‘this probability is 100 times greater than the annual probability of a magnitude 6 to 6.4 earthquake occurring in this region prior to the Darfield earthquake’.

In this way, the public can understand that while the absolute chance of a major earthquake is low, it is relatively high compared to the way it was before. Better integration of fault geometries and stress modelling will allow these earthquake forecast models to be improved. Better understanding of the way seismic energy is soaked up on its way through the crust will provide important information to building codes and assessment of potential damage in future earthquakes. The scientific community is thankful for this platform in which to deliver our results to the public.