The L’Aquila trial

MASSIMO COCCO1, GIOVANNA CULTRERA1*, ALESSANDRO AMATO1,

THOMAS BRAUN2, ANDREA CERASE3, LUCIA MARGHERITI1,

ALESSANDRO BONACCORSO4, MARTINA DEMARTIN1, PAOLO MARCO DE

MARTINI1, FABRIZIO GALADINI1,5, CARLO MELETTI6, CONCETTA NOSTRO1,

FRANCESCA PACOR7, DANIELA PANTOSTI1, SILVIA PONDRELLI8,

FRANCESCA QUARENI8 & MICOL TODESCO8

1Istituto Nazionale di Geofisica e Vulcanologia, Rome, Italy2Istituto Nazionale di Geofisica e Vulcanologia, Arezzo, Italy

3Dipartimento di Comunicazione e Ricerca Sociale, Rome, Italy4Istituto Nazionale di Geofisica e Vulcanologia, Catania, Italy5Istituto Nazionale di Geofisica e Vulcanologia, L’Aquila, Italy

6Istituto Nazionale di Geofisica e Vulcanologia, Pisa, Italy7Istituto Nazionale di Geofisica e Vulcanologia, Milan, Italy

8Istituto Nazionale di Geofisica e Vulcanologia, Bologna, Italy

*Corresponding author (e-mail: Giovanna.cultrera@ingv.it)

Abstract: The first stage of the trial in L’Aquila (Italy) ended with a conviction of seven experts,convened by the head of Civil Protection on 31March 2009, for multiple manslaughter and seriousinjuries. Theywere sentenced to six years in jail, perpetual interdiction from public office and a fineof several million euros to be paid to the victims of the earthquake of 6 April 2009 (moment mag-nitude 6.3) for having caused, by their negligent conduct, the death of 29 persons and the injury ofseveral others. The verdict had a tremendous impact on the scientific community and on the wayscientists deliver their expert opinions to decision makers and society. This paper analyses thescientific argumentations reported in the Verdict Motivations, where scientific data and resultswere largely debated and misused to demonstrate that they should have been considered as atool to predict an impending large earthquake. Moreover, we show that the supposed messageof reassurance was not generated at the experts’ meeting or by the official Istituto Nazionale diGeofisica e Vulcanologia reports. The media had a key role in conveying information during theseismic swarm, contributing to the risk perception. We stress that prevention actions based onseismic hazard knowledge are the best defence against earthquakes.

The events that led to the accusation and convic-tion of seven expert seismologists and engineerswho participated to the so-called meeting of theCommissione Grandi Rischi (High Risks Com-mittee, HRC), six days before the earthquake ofmoment magnitude (Mw) of 6.3 on 6 April 2009 atL’Aquila in Italy, are very complex, preventingthe analysis of individual issues outside the actualtimeline and requiring a synoptic examination.The verdict motivations involve difficult juridicalaspects and arguments connected to the role of sci-entists and risk communication, all issues exceedingthe local (Italian) dimension and attracting broadinterest worldwide.

The seven defendants (Franco Barberi, EnzoBoschi, Michele Calvi, Bernardo De Bernardinis,Mauro Dolce, Claudio Eva and Giulio Selvaggi)

were convicted of the multiple manslaughter of 29victims and multiple serious injuries, caused by‘negligence, imprudence and malpractice’, havingcarried out, during the HRC meeting of 31 March2009, an ‘approximate evaluation of the risks con-nected to the actual seismic activity observed atand around the city of L’Aquila’ (Verdict Motiv-ations 2013). According to the judge, they providedthe National and Regional Civil Protection Depart-ment (Dipartimento della Protezione Civile, DPC),the mayor of L’Aquila and the public with:

‘incomplete, imprecise and contradictory informationon nature, causes, hazard and future development ofthe seismic activity, neglecting thus their duties toevaluate the connected risks and the function of fore-casting, prevention and delivery of clear, correct andcomplete information’.

From: Peppoloni, S. & Di Capua, G. (eds) Geoethics: The Role and Responsibility of Geoscientists.Geological Society, London, Special Publications, 419, http://dx.doi.org/10.1144/SP419.13# 2015 The Author(s). Published by The Geological Society of London. All rights reserved.For permissions: http://www.geolsoc.org.uk/permissions. Publishing disclaimer: www.geolsoc.org.uk/pub_ethics

by guest on February 15, 2015http://sp.lyellcollection.org/Downloaded from

The reassuring message that, according to the judge,was conveyed by the HRC would have inducedpeople not to leave their houses, as they were usedto do by family tradition, after some shocks beforethe deadly Mw 6.3 earthquake (i.e. the behaviourof the defendants caused the death of the victims).

This article aims to analyse the scientific argu-mentations reported in the Verdict Motivationsand their use in the trial held in the L’Aquila court.We do not interpret the trial as a prosecution to sci-ence, as claimed by some, but show how scienceentered the prosecution, the trial and the verdict.We then discuss the alleged ‘act of tranquilliza-tion’, that caused the death of 29 persons, togetherwith the official reports to DPC from the IstitutoNazionale di Geofisica e Vulcanologia (INGV)(two defendants, Enzo Boschi and Giulio Selvaggi,were in 2009 the president of INGV and the direc-tor of the National Earthquake Centre of INGV,respectively) and the numerous declarations issuedby various experts in the period before the mainshock on 6 April 2009 and reported by the press

and mass media. We conclude with a short analysison the consequences of the L’Aquila trial for thescientific community and the impact on risk com-munication to the population in the case of naturaldisasters.

State of knowledge

Numerous studies performed in the decades preced-ing the 2009 L’Aquila earthquake yielded scienti-fic results that fostered a clear understanding ofthe seismic hazard of the Abruzzo region. Geo-physical and geological investigations (Galadini &Galli 2000; Boncio et al. 2004; DISS WorkingGroup 2010 among many others) allowed the identi-fication of the major active and seismogenic faults.Analysis of historical seismicity (Baratta 1901;Cavasino 1935; Boschi et al. 2000; Rovida et al.2011) contributed to the comprehension of the seis-mogenic potential of active faults as well as to theearthquake recurrence in the region (Fig. 1).

Fig. 1. Historical seismicity in central Italy for earthquakes with MW ≥ 5; red squares are the earthquakes, labelsindicate events with MW ≥ 6 (Rovida et al. 2011).

M. COCCO ET AL.

by guest on February 15, 2015http://sp.lyellcollection.org/Downloaded from

In 2004, all relevant studies to assess seismichazard were used by scientists to release a newseismic hazard map of Italy. This scientific achieve-ment (the new Italian Seismic Hazard map (Stucchiet al. 2011)), which represents the findings of thecollaborative work of a team of experts officiallyengaged to this task by INGV, became a nationallaw through a legal ordinance published in 2006(PCM 3519/2006; http://zonesismiche.mi.ingv.it).It is still the reference for the Italian buildingcode (Decree of the Ministry of Infrastructure inGazette n. 29 of 4 February 2008) and the seismic

classification. Today, as well as before 31 March2009, the Italian Seismic Hazard Map (Fig. 2) rep-resents a fundamental instrument for the preventionand mitigation of seismic risk. It is one of the waysin which scientists had conveyed their researchproducts to decision and policy makers, includingnational and regional civil defence agencies andlocal authorities.

The seismic hazard map clearly shows thatL’Aquila is one of the most hazardous areas inItaly. This information was accessible to the pub-lic through institutional websites and was delivered

Fig. 2. Italian Seismic Hazard Map (Italian Seismic Hazard map, PCM 3519/2006; http://zonesismiche.mi.ingv.it) ofmaximum acceleration value having 10% of exceedance in 50 years (in g). The inset is a zoom of this map for the regionsurrounding L’Aquila.

THE L’AQUILA TRIAL

by guest on February 15, 2015http://sp.lyellcollection.org/Downloaded from

to decision makers and involved stakeholdersthrough a national law and a legal ordinance. Infor-mation concerning the high seismic hazard of theregion was also conveyed to the public throughspecific dissemination activities (La Longa et al.2012), although this was certainly insufficient tocreate the necessary preparedness and awarenessabout seismic risk. The vulnerability assessmentof public and strategic buildings in the town ofL’Aquila was also known by the local authorities,since it was the object of a specific study publishedby the DPC in 1999 (the so-called ‘Barberi Report’;Dipartimento della Protezione Civile 1999). TheAbruzzo region was one of those described in theBarberi Report, which was sent to all involvedItalian municipalities soon after its publicationincluding those local authorities invited to attendthe HRC meeting on 31 March 2009. All this infor-mation was shared within the scientific communityand, more importantly, it was delivered to decisionmakers (Civil Protection officers, mayors, etc.),who had all the necessary knowledge and time tostart preventive actions in the (ten) years precedingthe 2009 L’Aquila earthquake.

Another missing piece to complete this overviewconcerns the temporal evolution of the seismicsequences. Several studies have described that mosthistorical strong earthquakes were not precededby seismic sequences. A few reports on foreshockactivity have been found for the 1461 earthquake(MW 6.4; Baratta 1901), whereas there was probablynot foreshock activity before the MW 6.7, 1703event (Rovida et al. 2011), apart from the intenseseismicity in the area near Norcia (located about50 km north of L’Aquila). Even the devastatingearthquake of 1915 in the Fucino area south ofL’Aquila was not preceded by foreshocks (Cavasino1935). In addition, it is well-known that several seis-mic sequences, some of which had similar charac-teristics to the one preceding the 2009 earthquake,had occurred in the Abruzzo region in the pastwithout generating a strong earthquake (see Amato& Ciaccio 2012, for a review; similar results havebeen reported for the nineteenth century in Galadini2013). This information was already known withinthe scientific community and G. Selvaggi, one ofthe convicted seismologists, showed at the HRCmeeting the map of previous seismicity, includingthe one produced in 1985.

This behaviour is common to many regions inItaly. The number and features of the numerous seis-mic sequences occurring every year are very highand heterogeneous. Arcoraci and colleagues deter-mined that 70% of the earthquakes occurring inItaly cluster in sequences or swarms (the remaining30% pertain to the so-called ‘background seismi-city’) and reported at least 127 sequences occurredin Italy in the period 2008 to 2010 (Arcoraci et al.

2011) (Fig. 3). If we look at the time distributionof these sequences, we see clearly that they lastfrom few days to several months and their magni-tude distribution in time is highly variable. We cantherefore conclude that the past and present stateof knowledge would have not allowed us to iden-tify a ‘peculiar character’ for the 2009 seismicsequence up to 31 March, in contrast with theVerdict Motivations.

The meeting of the national High Risks

Committee

A seismic sequence had been occurring sinceJanuary 2009,withmany small earthquakes ofmaxi-mum magnitude of 2.9 on 11 March (Fig. 4). At thattime, two seismic sequences were simultaneouslystriking the surroundings of L’Aquila and Sulmonatowns (Abruzzo region), creating panic since earth-quakes were often felt. According to standard algor-ithms (Reasenberg 1985), the space–time evolutionof the seismicity in 2008 and 2009 shows an abruptchange in the number of seismic events after mid-January 2009, with a maximum magnitude of 2.9within 30 km from L’Aquila (Fig. 5). The seismicactivity in the previous 12 months is well describedas ‘background seismicity’, which is peculiar ofactive tectonic regions and may include sporadic-ally felt earthquakes. Background seismicity is iden-tified by a constant slope of the curve of cumulativenumber of earthquakes (Fig. 5) and small statisticalvariations (in both number and magnitude) are com-mon features of all the active seismogenic areas.

The reaction of the population was influenced bythe amateur predictor, Giampaolo Giuliani, a tech-nician of the former IFSI-Torino (National Institutefor Astrophysics, Italy), who claimed to be able topredict earthquakes in the area on the basis of hishomemade instruments. During the seismic activity,he made contradictory forecasts for either an immi-nent large earthquake or the end of the seismicactivity, raising public order problems. These warn-ings were never precise enough to constitute anaccurate prediction and were subsequently shownto be unsound; at the time, however, they causedpublic alarm. The reaction of the population, andthe lack of preparedness and awareness of seismichazard, demonstrate that there had not been ade-quate prevention and mitigation initiatives in theprevious years by the local authorities to increasethe resilience of society to seismic hazard.

G. Giuliani used his homemade instruments todetect radon variations and interpreted his obser-vation as the evidence for impending earthquakes;this activity was not institutional and he stillcarries it out within his private foundation. On29 March, after a shock of a Richter magnitude

M. COCCO ET AL.

by guest on February 15, 2015http://sp.lyellcollection.org/Downloaded from

Fig. 3. Main seismic sequences that occurred in Italy in 2008–2010 (modified from Arcoraci et al. 2011). Almost allsequences (black squares) occur along the Apennine chain where most of the tectonic deformation was released.

THE L’AQUILA TRIAL

by guest on February 15, 2015http://sp.lyellcollection.org/Downloaded from

(ML) of 3.9, he predicted a large earthquake inthe town of Sulmona (Fig. 4); different from whatwas predicted, the following day a ML 4.0 shockoccurred in L’Aquila, more than 50 km north-westof Sulmona (Fig. 4), increasing the panic in thepopulation.

The head of the national DPC, Guido Bertolaso,then decided to convene in L’Aquila a meeting ofthe HRC experts on the following day, 31 March.The HRC is an advisory body of the NationalCivil Protection Department which has the duty ofevaluating the risk associated with situations suchas earthquakes, volcanic eruptions and nuclear acci-dents, providing expert opinions to DPC, HRCevaluations are based on scientific data provided

by experts and research institutes like INGV (theDPC ‘competence centre’ for earthquakes).

The experts were charged, then, with providingan ‘accurate analysis of the scientific and civil pro-tection aspects of the recent seismic sequence occur-ring in the Abruzzo area’ (Convocational Letter2009). As the judge claimed in the Verdict Motiv-ations, ‘a well-founded evaluation of risk predic-tion’ was not explicitly requested. In addition, thevolcanologist F. Barberi, the vice-president of theHRC, stated at the beginning of the meeting thatthe committee should have performed an objectiveevaluation of the ongoing seismicity also in termsof what could be forecasted, and that it shouldhave discussed and provided indications on the

Fig. 4. Seismicity localized in the Abruzzo region up to 31 March 2009. Yellow and green stars are the 2009 seismicsequences; black dots and stars are older earthquakes. The figure is taken from the Istituto Nazionale di Geofisica eVulcanologia (INGV) report discussed at the High Risks Committee meeting.

0.0

1.0

2.0

3.0

4.0

0

200

400

600

Mag

nitu

de

cum

ulat

ive

num

ber

Fig. 5. Cumulative number and magnitude of earthquakes in the region of L’Aquila from 1 January 2008 to 29March 2009; selection includes epicentres within 30 km of the town. For clarification, the Verdict motivations claimthat the sequence began in June 2008, whereas it started in January 2009 (courtesy of F. Mele).

M. COCCO ET AL.

by guest on February 15, 2015http://sp.lyellcollection.org/Downloaded from

alarms that people were suffering (Meeting draftreport 2009).

Several authorities attended the meeting, each ofthem having different roles and duties: governmentofficials (national and regional DPC, the Mayor ofL’Aquila, etc.) and scientists appointed as mem-bers of the HRC or invited without voting rights.In particular, the INGV seismologist and directorof the INGV National Earthquake Centre (GiulioSelvaggi) was appointed by INGV to accompanyhis president, Enzo Boschi (an HRC member), toreport on the seismicity recorded and localized bythe INGV seismic network.

The minutes of the meeting show that the expertswere neither alarmist nor reassuring, presentingthe hazard map of the area and debating on whatthe scientific community knew at that time (Minutesreport 2009). The report explains that seismicsequences frequently occur without leading to adamaging shock and that the earthquakes are notpredictable deterministically. However, the occur-rence of a damaging earthquake in the L’Aquilaregion, such as the one of 1703, was considered unli-kely in the short term but not ‘excluded in absoluteterms’, as suggested by the seismic history of theregion and by the presence of many active faultscapable of causing damaging earthquakes.

As part of their suggestion for DPC actions, theexperts underlined that, nowadays, it is impossibleto predict earthquakes ‘but their effects can be miti-gated and so it should be appropriate to make pre-vention (resistant buildings)’ (E. Boschi, Meetingdraft report 2009) and that:

‘the only way to be protected against earthquakes is tostrengthen and improve the constructions and theirability to withstand the earthquake. Another importantaspect to be treated for the purposes of civil protectionis to improve the level of preparedness to handle anemergency earthquake’ (F. Barberi, Minutes report2009).

No minutes of the meeting were available before 6April and therefore no-one could have read thembefore the night of the earthquake.

The press conference, at the end of the HRCmeeting, was convened by the vice-head of thenational DPC, Bernardo De Bernardinis, and theregional DPChead officer, Daniela Stati. Its contentswere not agreed during the HRC meeting. Seve-ral participants attended the press conference inwhich the seismologists were not invited to partici-pate. Nobody knows exactly what was said becausethere is no audio recording or a press release.

In the interviews that followed (e.g., the one ofF. Barberi to the Abruzzo24ore.tv channel), theconcept of the impossibility to predict earthquakeswas repeated and no reassuring message wasexpressed. However, local media (Abruzzo24ore.tv

channel) reported that the presence of the ongoingseismic swarm was positive due to the dischargeof energy. Some went on broadcasting the interviewof the vice-head of national DPC (B. De Bernardi-nis), given before the HRC meeting, where hedeclared that the situation was positive: ‘the scienti-fic community keeps on telling me that the situ-ation is favourable and that there is a discharge ofenergy’ due to the continuous microseismic activityobserved in the weeks before 6 April. According tothis (scientifically wrong) idea, the occurrence ofmany small seismic events would reduce the accu-mulated energy on the fault, thus preventing theoccurrence of a strong earthquake. He claimed helearned the story of the ‘discharge of energy’ fromnewspaper articles or press releases of researchers’interviews, without either considering the officialINGV reports released in February and March(INGV official reports 2009), which had completelydifferent contents, or consulting the scientific com-munity directly. The release of energy in an earth-quake of magnitude 6 is about a thousand timesthe energy released by a magnitude 4 and a milliontimes the energy released by a magnitude 2, soclearly the occurrence of a small-magnitude seismicsequence does not release the energy of a potentiallarge earthquake and thus preventing it. Neithercan it be stated that the low-magnitude seismicityrecorded until 31March could have been consideredas an unequivocal precursor of an imminent large-magnitude event.

Five days after the HRC meeting, two earth-quakes of magnitude .3 occurred in the nightbetween 5 and 6 April, and few hours later thedestructive main shock struck the town of L’Aquilaat 3:32 local time.

Scientific arguments used in the trial

According to the Verdict Motivations (2013), thedefendants have performed an incomplete andapproximate risk analysis because the event was intheir ‘sphere of predictability’ based on scientificdata and current knowledge.

Many scientific arguments were discussed inthe trial. Here we briefly review those that, accord-ing to the Verdict Motivations, could have beenrelated by the defendants to the occurrence of animpending strong shock. These are: (a) ‘the seis-mic history of L’Aquila, in particular the recur-rence of large earthquakes preceded by swarms’;(b) the ‘peculiar time evolution of the sequence,with increasing number and magnitude of the earth-quakes focused at the same hypocentre and depth’;(c) the high seismic hazard of the L’Aquila area;(d) ‘the probabilistic study by Enzo Boschi andGiulo Selvaggi, indicating the region of L’Aquila

THE L’AQUILA TRIAL

by guest on February 15, 2015http://sp.lyellcollection.org/Downloaded from

as one with the highest probability of occurrence ofan earthquake of magnitude greater than 5.5 or 5.9,as with values among the highest in the national ter-ritory; even with a certainty judgement, probability(P) equal to 1, in the period 1995 to 2015’ (theauthors of the paper are indeed Boschi, Gasperiniand Mulargia (Boschi et al. 1995), not Selvaggi aserroneously indicated in the verdict motivations).Moreover, (e) the verdict motivations state that,based on the elements above, the main shock on 6April was in ‘the sphere of control and knowledgeof the defendants’, or in other words, ‘it was partof their sphere of predictability’ (Verdict Motiv-ations 2013).

According to the judge, this would suggest thatthe five arguments described above are indicatorsof an impending strong earthquake and that theseelements are closely related in a sort of conceptualmodel. Are these elements really linked in a clearand unique way, so that any expert could havedrawn the same conclusion from their analysis, atthe time of the meeting on 31 March? The answeris no: the scientific community worldwide hasbeen working on these issues for many decades,without providing a unique answer to be used forcivil protection purposes.

The first element concerns the (few) stronghistorical earthquakes in the region and their rela-tionship with the (frequent) seismic sequencesaffecting the same area. The historical earthquakes(Mw 6.6 in 1349, Mw 6.4 in 1461, Mw 6.7 in 1703and Mw 6.0 in 1762; Fig. 1) had been used to deter-mine the high seismic risk of the L’Aquila area, asdiscussed during the HRC meeting. However, theevidence of seismic sequences preceding historicallarge earthquakes is rather weak and cannot be con-sidered as a prognostic indicator of an impendingshock. Indeed, during the twentieth century, therewere at least 23 seismic sequences in the Abruzzoregion with only one followed by a strong earth-quake: the main shock on 6 April 2009 (Amato &Ciaccio 2012). A statistical analysis of the catalogueof Italian earthquakes between 1950 and 2010revealed that about 0.8% of the seismic sequences,having at least an ML 4 shock, was followed by astrong earthquake within a radius of 10 km andwithin three days (Marzocchi & Zhuang 2011).We can therefore conclude that a clear causal linkbetween swarms (or seismic sequences) and strongearthquakes cannot be demonstrated.

The second element is the ‘peculiar time evol-ution of the sequence, with increasing number andmagnitude of the earthquakes focused at the samehypocentre and depth’ (Verdict Motivations 2013).First of all, any ongoing seismic sequence is associ-ated with an increase in the number of earthquakeswith magnitudes larger than those characterizingthe background seismicity (Fig. 5). Therefore, the

analysis of the temporal evolution of seismicity doesnot permit the recognition of unambiguous peculiarfeatures to be used to forecast a large earthquake.Recent studies tackle this problem by computingthe probability of occurrence of a large earthquakefrom the spatio-temporal changes in the rate ofearthquake production. Although it is well docu-mented that the increase in probability caused bya large magnitude shock induces aftershocks, theopposite case (small earthquakes inducing strongevents) is still a matter of debate in the scientificcommunity (Marzocchi & Zhuang 2011). Indeed,there are models claiming that important seismicevents are preceded by seismic quiescence (Haber-mann 1988; Murru et al. 1999; Katsumata 2011).

The statistical models used in the operationalearthquake forecast (Jordan et al. 2011), a type ofseismicity-based probabilistic forecast in the shortterm (days, week), show that the short-term prob-abilities of a damaging event generally remainbelow 1% per day (Jordan et al. 2011) and are gen-erally too low to be used for civil protection actions.It is therefore not demonstrated that there is a‘peculiar’ type of seismic sequence that can beused as a predictive tool. Finally, it is important toemphasize that the discussion concerning the oper-ational earthquake forecasting is quite recent andstarted after the 2009 L’Aquila earthquake.

The third element considered as a risk indicatoris the long-term hazard of the region. The ItalianSeismic Hazard Map indicates that the area hasthe highest hazard in Italy. The map was shownand discussed at the meeting on 31 March and thisfact alone rules out any reassuring message fromthe scientific community. We stress again that theoccurrence of a seismic sequence in a high hazardregion worldwide is a frequent circumstance anddoes not provide any unambiguous indication forthe short-term forecast of a strong earthquake.

The fourth element refers to a paper publishedby Boschi et al. (1995) in the Bulletin of the Seis-mological Society of America which reported aprobability equal to 1 that a M . 5.5 earthquakewould have occurred near L’Aquila in the 5, 20and 100 years after the year of publication (1995).As for all scientific research, this article was writtento contribute to the scientific issues on earthquakeoccurrence and seismic hazard, and certainly notfor discussion and interpretation during a legalprosecution, and the discussion on the validity ofthose results pertains to the scientific debate. More-over, the article’s conclusions do not contradictthe INGV official statements and Enzo Boschi’sdeclarations before, during and after the HRC meet-ing, since the target area had a high probability ofoccurrence and high seismic hazard. In their paper,Boschi and colleagues adopted a probabilisticmodel using aGaussian distribution in a small region

M. COCCO ET AL.

by guest on February 15, 2015http://sp.lyellcollection.org/Downloaded from

surrounding L’Aquila where three earthquakes hadoccurred in 1646 (Mw 4.5), 1703 (Mw 6.7) and1762 (Mw 6.0), with an interval of about 60 yearsbetween earthquakes. The resulting probabilityvalues computed in the paper for three time intervals(5, 20 and 100 years) were always 1; that means cer-tainty in the period range 1995–2000 as well as1995–2095. However, the available data were rela-tively scarce and the statistical validity of theauthors’ conclusions was rather poor. Indeed, theywarned the reader about this anomalous result,admitting that the statistical sample was not signifi-cant (only three cases in many centuries) andsuggesting it could be due to earthquake clustering.The very different magnitudes of the three earth-quakes (Rovida et al. 2011), ranging from 4.5 to6.7 (i.e. a factor 2000 in energy), make the statisticalanalysis not significant.

The ‘elements of risk evaluation’, which accord-ing to the Verdict Motivations would have been pro-gnostic to recognize the imminent risk of a strongearthquake in L’Aquila, are common to many seis-mically active regions in Italy and worldwide. Thestatement that the effects of the devastating earth-quake were in ‘the sphere of predictability’ of theexperts means that they should have emphasizedthe need for prevention actions, which they actuallydid. Thus, it is difficult to understand why thisresponsibility is not extended to those policymakerswho did not promote any prevention initiative to

mitigate earthquake risk in the decades precedingthe L’Aquila shock.

The relationship among the so-called ‘risk indi-cators’ is difficult to assess from a scientific pointof view. It is appropriate for scientific discussions,papers and conferences, but is very critical whenbrought into a courtroom to individuate a posterioripossible legal responsibilities.

The mass media and the presumed

act of tranquillization

The prosecution’s main accusation against the sevendefendants concerns their responsibility in deliver-ing ‘incomplete, imprecise and contradictory’ infor-mation in communicating risk to the population aswell as in the reassurance message that could haveled people to remain in their houses during thenight between 5 and 6 April 2009. During the seis-mic emergency, the population received informa-tion from local authorities and the regional DPChead officer as well as from scientists and decisionmakers, which were filtered and in some casesbiased by the media. Without any doubt, unclear,ambiguous and inconsistent messages disseminatedby the mass media in the two months preceding themain shock certainly did not help the citizens toget a clear idea about what was going on (Fig. 6).In one single day, on 3 April 2009, one could

The Earth is still shaking – Earthquake, requested state of emergency –Today schoolsare open. The police are looking for predictors of false allerts (Il Centro, April 2)

Evacuation exercises in schools – flyers with advice aredistributed in some schools (Il Centro, April 4)

Cialente confirms an emergency plan absolutely reassuring – Earthquake, back tonormality –The 15th of April scholars of De Amicisshould go back to their schools(Il Messaggero, April 3)

Fig. 6. Titles of articles that appeared in newspapers in the first days of April 2009 in the period between the HighRisks Committee meeting on 31March and the occurrence of the destructive earthquake on 6 April. Unclear, ambiguousand inconsistent messages disseminated by the mass media did not help the population to form a clear idea about whatwas going on.

THE L’AQUILA TRIAL

by guest on February 15, 2015http://sp.lyellcollection.org/Downloaded from

find on the news that ‘the earthquake doesn’t scareany more’ (Il Messaggero newspaper) and ‘it doesnot tranquillize the souls’ (Il Tempo newspaper).Without going into further details, a complete over-view of newspaper articles can be found at http://processoaquila.wordpress.com/media/rassegna-stampa-2009-2009-press-review and in the paper byAmato & Galadini (2014).

The mass media played a crucial role in commu-nicating risk to society and in particular in deliver-ing alarming or reassuring messages to the generalpublic during the days before and after the HRCmeeting. One of the journalists’ important sourcesof information on the ongoing seismic activity wasthe DPC. Because the communication of risk isone of the institutional duties of the DPC, theirexpertise is beyond question. It is therefore difficultto understand why the head of DPC and his deputywere convinced of the correctness of the formerlymentioned wrong concept of the discharge ofenergy as an inhibitor of strong earthquakes, sincethis statement was never included in the officialINGV reports, which were delivered directly toDPC without any filter by the media. This conceptwas directly mentioned by the vice-head of DPC(Bernardo De Bernardinis) in the aforementionedTV interview on 31 March 2009 before the HRCmeeting, when the journalist suggested to him thathe should ‘clink glasses full of local wine’ and staycalm. Reassuring messages were also delivered bythe regional DPC head officer, Daniela Stati, whoclaimed on 21 March that ‘she would have alreadymoved her family into a tent in the garden, if therewas imminent danger’ (TV interview, https://www.youtube.com/watch?v=PkOO8xU332o) andon 30 March stated in a press release issued byAgenzia Nazionale Stampa Associata (ANSA) ‘nofurther earthquakes of any magnitude are expected[. . .] Anything else is false’ (Verdict Motivations2013). After the conclusion of the HRC meeting on31 March, contradictory decisions were taken bythe mayor of L’Aquila and the dean of the univer-sity of L’Aquila. The former decided to keep theschools closed for safety verifications on the fol-lowing day, while the latter ordered the universityto remain open.

The INGV official reports (INGV official reports2009) did not minimize the seismic risk in theAbruzzo region, but underlined the presence ofactive faults and the presence of strong historicalearthquakes. They also stated that:

‘According to the present knowledge we can say thatthe seismic sequence of the past months has notchanged, that is, increased or decreased, the probabilityof occurrence of strong earthquakes in the area’.

INGV official reports and comments released duringinterviews simply report the formerly explained

concepts: (a) seismic sequences are often observedin Italy without generating damaging earthquakes;(b) this does not significantly change the probabilityof occurrence of a large magnitude earthquake; and(c) the region of L’Aquila has experienced in thepast many strong earthquakes and active faults arewell known in the area indicating that preventionactions are certainly needed. Since such declara-tions have a neutral character regarding the futureoccurrence and impact of a main shock, they canbe misused in a binary black–white scheme in thesense of ‘if there is no alert, we can feel either reas-sured or scared’.

Discussion

The evidence discussed in previous sections demon-strates that science was part of the trial. The use ofscientific results, characterized by use of a techni-cal language and uncertainties, can be extremelymisleading when brought in a courtroom. In thatcontext, the search for an absolute truth (beyondany reasonable doubt) can force the interpretationof uncertain scientific knowledge in one directionor in another. For instance, the interpretation ofthe premonitory role of a seismic sequence in thetrial and in the Verdict Motivations can be affectedby the human tendency to interpret facts a posterioriby reconsidering them with the help of updatedknowledge after their occurrence (one is alwaysmore clever afterwards). The basic principle ofscientific research, guided by doubts, experimen-tation, mistakes, continuous revisions and updates,and new discoveries, is incompatible with the prac-tice of a posteriori assessment of an intrinsicallyunlikely event, such as an earthquake, that occursirregularly and with a multisecular recurrence.

We have discussed in detail how the verdictmotivated the accusation for imprecise and ambi-guous communication of the imminent risk. Nowa-days scientists are forced to learn and adopt newapproaches to communicate risk and the associ-ated uncertainties. This requires a multidisciplinaryapproach involving complementary expertise andlong-term plans to create awareness and prepared-ness in populations living in earthquake-proneregions. Different actors with different roles areneeded to apply these plans, in which scientists areonly one of the key players. The lack of thisapproach amplified the difficulties in managingthe emergency before and after the L’Aquila mainshock. If it is hard for people to understand andaccept the uncertainty typical of science duringquiet periods, it becomes impossible during acrisis, when irrational behaviour tends to prevail.In these moments, the intermediation role of massmedia is crucial, easily bringing about opposite

M. COCCO ET AL.

by guest on February 15, 2015http://sp.lyellcollection.org/Downloaded from

extremes. This means that, in case of an evolvingseismic crisis affecting a poorly resilient society,either alarming or reassuring messages are con-veyed to the population, often fluctuating from oneextreme to the other.

During the weeks preceding the earthquake on 6April 2009, journalists often badly interpreted andtranslated what was really affirmed by seismolo-gists. This probably contributed to the tranquilliza-tion of the general public and certainly generatedconfusion and insecurity. This was analysed in detailby Mario Morcellini (Professor of Sociology ofCultural and Communicative Processes, and Direc-tor of the Department of Communication and SocialResearch, ‘La Sapienza’ University of Rome) forthis particular case (personal communication). Hereviewed over 50 years of empirical literature oncommunication models and media organizations toclaim that journalistic interpretations can be con-sidered as a deforming lens, capable of altering thecorrect perception by the citizens, especially in situ-ation of major risks. Even in the context of risk com-munication, the question of ‘journalistic bias’ ableto construct their own portrayals of reality as wellas influence perception of citizens exposed to aperceived risk, has been the subject of many theoret-ical studies and much empirical research, the impli-cations of which are discussed in Slovic (1986),Lichtenberg & MacLean (1991), Leiss (1996),Kitzinger (1999), Wahlberg & Sjoberg (2000) andMurdock et al. (2003).

Antonello Ciccozzi, a consultant whose anthro-pological analysis was used by the judge to assessthe effect of the reassuring messages as the real(contributory) cause of the deaths, affirms insteadthat ‘it is demonstrable that the media did notdistort anything concerning the obtained infor-mation: the tranquillization came from the scientificinstitutions’ (Ciccozzi 2013). We disagree with hisconclusions and believe that the evidence discussedabove confirms that the media play an importantrole in determining an ambiguous risk perception.Evidence from over 30 years of research on riskcommunication demonstrates that, in impendingrisk situations, the public’s trust in public institu-tions, decision makers and scientists should neverbe taken for granted. Moreover, the risk communi-cation guidelines issued by major risk agenciessuch as the US Environmental Protection Agency(US EPA 1988), the US Nuclear Regulatory Com-mission (NRC, 2004) and UK Resilience (UKResilience 2006) clearly indicate that building,maintaining or restoring trust and credibilitybetween the general public, scientists and decisionmakers must be assumed as one of the main goalsof any effective risk communication strategy.

Unfortunately, there are no clear answers to thequestions that seismologists are frequently asked:

‘Is a stronger earthquake expected?’, ‘How willthe seismic sequence evolve?’, ‘Will the seismicactivity cease soon?’, ‘Should we keep calm?’ and‘Can you reassure the citizens?’. A risk-preparedsociety should raise different questions such as:‘Do I live in a resistant building, capable of protect-ing me from a damaging earthquake?’ and ‘Arethere evacuation plans in place that will allow meto leave my house safely after a major shock?’. Ingeneral, from the media point of view, the lack ofspectacular news (‘The dog that doesn’t bite is notnews’, Hough 2009) represents a negative factor.A prudent, sound and realistic information basedon (uncertain) scientific data is not suited to them,especially in the case of a seismic crisis whenpeople are scared and want absolute certainty andreassurance. The tendency of unconsciously dis-torting technical and scientific knowledge provokesthe propagation of mostly incomprehensible, butsometimes even tranquillizing information. The lin-guistic complexity of science and the cultural dis-tance from the civil society constitute elements ofweakness, favouring the blaming of the scientificcommunity.

The L’Aquila case represents a clear example inwhich the results of decades of scientific researchwere applied to characterize the seismic hazardand the vulnerability of the territory. The scientificresults were also transferred to decision makers,since the new seismic hazard map of Italian territorybecame a law of the Italian Republic. Furthermore,these results were transferred to local authoritiesin the years preceding the 2009 earthquake. Thisdemonstrates that all the conditions were in placeto define and apply prevention actions and prepared-ness initiatives involving the population. Thisstrategy is still recognized as the most effectiveapproach to mitigate seismic risk in any place of theworld. For this reason, the statement of the VerdictMotivations (2013) that ‘the only defence fromearthquakes that consist in reinforcing buildingsand increase in that way their capacity to resistshaking seems as obvious as useless’ is wrong anddangerous.

We agree with all those who believe that theL’Aquila trial and its verdict, by invoking lastminute risk communication and convicting the seis-mologists and engineers, diverts attention from theresponsibilities of those who, for decades, refused tocontrol and reduce building vulnerability. Vittorini(2011), a medical scientist and local councillorof L’Aquila, noted that the information on the seis-mic hazard in the Abruzzo region was available butit was left in the desk drawers of the politicians andadministrators. He seems to assign the responsibil-ity of the lack of awareness of L’Aquila citizens tothe past administrations for having ignored and noteven partially used the available information for

THE L’AQUILA TRIAL

by guest on February 15, 2015http://sp.lyellcollection.org/Downloaded from

implementing the necessary measures for land-useplanning.

The L’Aquila earthquake caused much contro-versy and certainly did not help to strengthen thecentral role of prevention and of reducing the vul-nerability of buildings. A modern defence strategyfrom seismic disasters requires citizens to have arealistic perception of the potential damages of aseismic event in their region and of the securitylevel of their homes and working places. Unfor-tunately, a reduced perception of seismic risk iscommon to many regions in Italy even today, par-ticularly in high hazard areas (Crescimbene et al.2013).

Last but not least, another consequence regardsthe exposure of the scientist involved in technicalcommittee, as well summarized by the journalistAnna Meldolesi in an interview to the newspaperIl Centro on 23 October 2012:

‘The next time a scientist is asked to evaluate the risk offlooding or earthquakes, he will first evaluate theproper risk of his statements before communicating.We run the risk that a phenomenon will arise, like indefensive medicine, where doctors choose for thepatient the therapy which implies the lower risk forthemselves from a legal point of view, instead of thebest one for the patient’.

In the same line, Marzocchi (2012) wrote:

‘When the next earthquake crisis occurs, seismologistswill be afraid to say or to do anything. I expect thisverdict may affect also many other scientific fieldswhere important decisions have to be made underlarge scientific uncertainty’.

Opting for the worst possibility does not provide auseful service to citizens, resulting in continuousand useless one night evacuations, as it happenedafter the Mw 4.8 earthquake in the Garfagnanaarea (Tuscany, Italy) in January 2013. In the pastfew years in Italy, there have been several one dayschool evacuations after very small (even ML 2)earthquakes due to scared teachers and local auth-orities. Most of the time, these evacuations werenot followed by any subsequent action for effectiverisk reduction (building performance evaluation andretrofitting), but only by a removal of the imminentrisk: no more swarm, no more risk – a dangerouspractice.

We believe that progress in the mitigation ofnatural risk should be based on knowledge of thehazard of the territory, on the reduction of vulner-ability and awareness of risk exposure. This canonly be achieved through the joint action of scien-tists, institutions, local authorities, the media andsociety. In any seismic country in the world, thebest defence against earthquakes is safe buildingsand citizens’ preparedness to face the occurrenceof an earthquake.

On 10 November 2014, while this paper wasabout to be published, the Appeal Court acquittedthe experts Giulio Selvaggi, Enzo Boschi, FrancoBarberi, Michele Calvi, Mauro Dolce and ClaudioEva on all counts because no crime had been com-mitted; the court upheld a guilty verdict againstBernardo De Bernardinis and issued a suspendedtwo-year sentence. The verdict’s reasoning is expec-ted in 90 days. This verdict closely reflects ourposition about the innocence of scientists and wethink that the arguments analysed in this paper arestill worth discussing in the light of public andmedia reaction to the Appeal Court’s decision.

The authors would like to thank their colleagueG. Selvaggi for long discussions about the trial, an eventthat had and still has a major impact on the interactionbetween the scientific community and civil society bothin Italy and abroad. The authors participate in the INGVworking group for the information management on theL’Aquila trial (http://processoaquila.wordpress.com/),where many documents on the issues raised in this paperhave been collected. Our thoughts go once again to theearthquake victims and their relatives: we know that nosentence could ever compensate for the loved ones lostin this disaster.

References

Amato, A. & Ciaccio, M. G. 2012. Earthquake sequencesof the last Millennium in L’Aquila and surroundingregions (central Italy). Terra Nova, 24, 52–61,http://dx.doi.org/10.1111/j.1365-3121.2011.01037.x

Amato, A. & Galadini, F. 2014. Gli argomenti dellascienza nel processo dell’Aquila alla ‘commissionegrandi rischi. ANALYSIS, Rivista di cultura e politicascientifica 3/2013 (in Italian with English abstract).

Arcoraci, L., Battelli, P., Castellano, C., March-

etti, A.,Mele, F.,Nardi, A. &Rossi, A. 2011. Bollet-tino Sismico Italiano 2008–2010. Poster presented at30th National Congress, Gruppo Nazionale di Geofi-sica della Terra Solida, November 2011, Trieste, Italy.

Baratta, M. 1901. I Terremoti d’Italia. Arnaldo ForniEditore, Turin.

Boncio, P., Lavecchia, G. & Pace, B. 2004. Defining amodel of 3D seismogenic sources for Seismic HazardAssessment applications: the case of central Apennines(Italy). Journal of Seismology, 8, 407–425.

Boschi, E., Gasperini, P. & Mulargia, F. 1995. Fore-casting where larger crustal earthquakes are likely tooccur in Italy in the near future. Bulletin of the Seismo-logical Society of America, 85, 1475–1482.

Boschi, E., Guidoboni, E., Ferrari, G., Mariotti, D.,Valensise, G. & Gasperini, P. (eds) 2000. Catalogueof strong Italian earthquakes from 461 B.C. to 1997.Annals of Geophysics, 43, 609–868, with databaseon CD-ROM.

Cavasino, A. 1935. I Terremoti d’Italia nel Trentacin-quennio 1899–1933. Appendix to Volume IV, Ser.III Delle Memorie del R. Ufficio Centrale di Meteoro-logia e Geofisica. Instituto Poligrafico dello Stato,Rome.

M. COCCO ET AL.

by guest on February 15, 2015http://sp.lyellcollection.org/Downloaded from

Ciccozzi, A. 2013. Parola di scienza. Il terremotodell’Aquila e la Commissione Grandi Rischi. Un’ana-lisi antropologica. DeriveApprodi, Rome.

Convocational Letter 2009. http://processoaquila.files.wordpress.com/2012/10/2-cgr-convoc-310309.pdf (oin Italian) (accessed 27 November 2014).

Crescimbene, M., La Longa, F., Camassi, R. & Pino,N. A. 2013. D2.6 Report on risk perception. DPC-INGV-S2 project. Istituto Nazionale di Geofisica eVulcanologia, https://sites.google.com/site/ingvdpc2012progettos2/deliverables/d2_6 (accessed 27November 2014).

Dipartimento della Protezione Civile 1999. Censi-mento di vulnerabilita degli edifici pubblici, strategicie speciali nelle regioni Abruzzo, Basilicata, Calabria,Campania, Molise, Puglia e Sicilia, 3 volumes. Dipar-timento della Protezione Civile, Rome.

DISS Working Group 2010. Database of IndividualSeismogenic Sources (DISS), Version 3.1.1: A compi-lation of potential sources for earthquakes larger thanM5.5 in Italy and surrounding areas. Istituto Nazionaledi Geofisica e Vulcanologia, http://diss.rm.ingv.it/diss/ (accessed 27 November 2014).

Galadini, F. 2013. I terremoti in Abruzzo e la cultura sis-mologica tra la fine dell’Ottocento e i primi del Nove-cento. In: ‘. . . pareva quel giorno dell’UniversalGiuditio. Il terremoto aquilano del 1703 tra indaginestorica e sviluppo della sismologia moderna’. Kirke,Cerchio, XVII–CIV.

Galadini, F. & Galli, P. 2000. Active tectonics in thecentral Apennines (Italy) – input data for seismichazard assessment. Natural Hazards, 22, 202–223.

Habermann, R. E. 1988. Precursory seismic sequence:past, present and future. Pageoph, 126, 2–4.

Hough, S. 2009. Predicting the Unpredictable: TheTumultuous Science of Earthquake Prediction. Prince-ton University Press, Princeton, New Jersey.

INGV official reports 2009. INGV reports to DPC(February 17 and March 12, 2009), http://processoaquila.wordpress.com/prima-del-processo/ (accessed27 November 2014).

Jordan, T. H., Chen, Y.-T. et al. 2011. Operationalearthquake forecasting: state of knowledge and guide-lines for utilization. Annals of Geophysics, 54,315–391, http://dx.doi.org/10.4401/ag-5350

Katsumata, K. 2011. A long-term seismic quiescencestarted 23 years before the 2011 off the Pacific coastof TohokuEarthquake (M ¼ 9.0).EarthPlanets Space,63, 709–712.

Kitzinger, J. 1999. Researching risk and the media.Health, Risk & Society, 1, 55–69.

La Longa, F., Camassi, R. & Crescimbene, M. 2012.Educational strategies to reduce risk: a choice ofsocial responsibility. Annals of Geophysics, 55,445–451, http://dx.doi.org/10.4401/ag-6110

Leiss, W. 1996. Three phases in the evolution of riskcommunication practice. The Annals of the AmericanAcademy of Political and Social Science, 545, 85–94.

Lichtenberg, J. & MacLean, D. 1991. The role of themedia in risk communication. In: Kasperson, R. E. &Stallen, P. J. M. (eds) Communicating Risks to the

Public: International Perspectives. Springer, Dor-drecht, 4, 157–173.

Marzocchi, W. 2012. Putting science on trial. PhysicsWorld, 25, 17–18.

Marzocchi, W. & Zhuang, J. 2011. Statistics betweenmainshocks and foreshocks in Italy and Southern Cali-fornia. Geophysical Research Letters, 38, L09310.

Meeting draft report 2009. Draft of the minutes reportof the 31 March 2009 HRC meeting, http://processoaquila.files.wordpress.com/2012/11/3b-bozza-cgr-310309eng.pdf (accessed 27 November 2014).

Minutes report 2009. Riunione Commissione GrandiRischi, L’Aquila, 31 marzo 2009, http://processoaquila.files.wordpress.com/2012/10/4-cgr-310309.pdf(accessed 27 November 2014).

Murdock, G., Petts, J. & Horlick-Jones, T. 2003. Afteramplification: rethinking the role of the media in riskcommunication. In: Pidgeon, N., Kasperson, R. E. &Slovic, P. (eds) The Social Amplification of Risk. Cam-bridge University Press, Cambridge, 156–178.

Murru, M., Console, R. &Montuori, C. 1999. Seismicquiescence precursor prior to the 1983 Nihonkai-Chubu (M7.7) earthquake, Japan. Annali di Geofisica,42, 871–882.

NRC 2004. Effective Risk Communication. The NuclearRegulatory Commission’s Guidelines for externalRisk Communication, NUREG/BR-0308. NuclearRegulatory Commission, Washington DC.

Reasenberg, P. 1985. Second-order moment of centralCalifornia seismicity, 1969–1982. Journal of Geophy-sical Research, 90, 5479–5495.

Rovida, A., Camassi, R., Gasperini, P. & Stucchi, M.(eds) 2011. Catalogo Parametrico deiTerremoti Italiani,version CPTI11, December 2011. Istituto Nazionale diGeofisica e Vulcanologia, Milano, Bologna, http://emidius.mi.ingv.it/CPTI, http://dx.doi.org/10.6092/INGV.IT-CPTI11 (accessed 27 November 2014)

Slovic, P. 1986. Informing and educating the public aboutrisk. Risk Analysis, 6, 403–415.

Stucchi, M.,Meletti, C.,Montaldo, V.,Crowley, H.,Calvi, G. M. & Boschi, E. 2011. Seismic HazardAssessment (2003–2009) for the ItalianBuildingCode.Bulletin of the Seismological Society of America, 101,1885–1911, http://dx.doi.org/10.1785/0120100130

UK Resilience 2006. Communicating Risk Guidance.HMSO, London.

Verdict Motivations 2013. Tribunale di L’Aquila,Sezione Penale, Dott. M. Billi. Sentenza n. 380/2012.http://processoaquila.files.wordpress.com/2013/01/sentenza-grandi-rischi-completa-1.pdf (accessed 27November 2014).

Vittorini, V. 2011. In: Protezione Civile e responsabilitanella societa del rischio: chi valuta, chi decide, chigiudica. Edited by theDepartment of Civil Protection ofthe Presidency of the Council of Ministers and CIMAResearch Foundation. Edizioni ETS, Pisa, 136–140.

US EPA 1988. Seven Cardinal Rules of Risk Communi-cation, OPA-87-020. US Environmental ProtectionAgency, Washington DC.

Wahlberg, A. A. & Sjoberg, L. 2000. Risk perceptionand the media. Journal of Risk Research, 3, 31–50.

THE L’AQUILA TRIAL

by guest on February 15, 2015http://sp.lyellcollection.org/Downloaded from