Tuesday, March 12, 2019

A recent public earthquake prediction in Greece: geoethical issues

Gerassimos A. Papadopoulos*
(IAPG-Greece coordinator; Member of the Seismic Hazard Evaluation Committee, Greece)

* Gerassimos A. Papadopoulos served as the Committee Chairman (2006-2008) and still serves as Committee Member since 2017. He also served  (1994-2000) as Vice-Chairman of the European Advisory Evaluation Committee for Earthquake Prediction (Council of Europe), and as Member of the International Commission on Earthquake Forecasting established by the Italian Administration (2009-2011) in the aftermath of the 6th April 2019 L’Aquila earthquake.

Picture above:
The target area of radius of 225 km around Patras city according to the EQ prediction submitted to the Greek Permanent Scientific Committee for the Assessment of Earthquake Hazard and Risk. The star indicates the epicenter of the 25th October 2018 M6.8 earthquake. 

Gerassimos A. Papadopoulos
Earthquake prediction is a quite challenging topic. The socially sensitive nature of this topic calls for the establishment of Code of Ethics regarding the management of prediction results at the scientific, administrative and operational levels as well as the communication of predictions to the public. The seismological community put forward Codes of Ethics which are strongly recommended to be followed by scientists dealing with earthquake (EQ) prediction research, e.g. Allen et al. (1984), Council of Europe (1991). In the next lines extracts from Allen et al. (1984) are used.

Greece, a country of very high seismicity, since the beginning of 1980’s repeatedly experienced scientific, operational and geoethical issues due to public claims for successful predictions by the so-called VAN research team (University of Athens), based on supposed precursory seismic electric signals (SESs). To meet these challenges the Greek Government established (1992) a Permanent Scientific Committee for the Assessment of Earthquake Hazard and Risk (hereafter “the Committee”). In the last 20 years VAN does not submit predictions at the Committee for evaluation. Instead, it selects to upload relevant reports at the arXiv depository of Cornell University (USA), which accepts scientific reports which do not get prior review. Hence, VAN violates the rule of ethics saying that “scientists who wish to put forward predictions should submit them to the relevant organization”.  Since arXiv is openly accessible, VAN violates another basic rule of ethics: “scientists should seek to have predictions confidentially reviewed by their scientific colleagues and should ensure that such support is forthcoming before making them more widely known”. The arXiv editors violate also the Code of Ethics which says: “Scientific editors who are considering the publication of papers containing EQ predictions should take special precautions to ensure that adequate support from within seismological community has been obtained for these predictions”. The present short commentary is a personal view on the geothical issues raised regarding a recent case of a publicly announced EQ prediction in Greece.

On 24.01.2019, Prof. A. Tselentis, Director of the Institute of Geodynamics, National Observatory of Athens (IGNOA), who claims collaboration with VAN, sent to the Committee an e-mail containing an arXiv VAN report (Sarlis et al., 20.01.2019) saying that SESs were recorded a few days prior to the M6.8 EQ of 25.10.2018 in Zakynthos Island (Greece) (see the picture above). The report adds that new SESs recorded in Patras, NW Peloponnese, by the beginning of January 2019, combined with now casting EQ statistics, indicates an impending EQ of M≥6.0 in an area of radius 225 km around Patras (see the picture above) with the EQ potential score reaching at 80%. However, this is only an indirect approach of the probability for the EQ occurrence. The report adds: “The lead time of single SES is ≤11 days while for SES activities it varies from a few weeks up to 5.5 months … To estimate the occurrence time of the impending EQ, we currently analyze in natural time the subsequent seismic activity”. In an evaluation report that I submitted to the Committee the probability for the occurrence by chance of such an EQ in the target area and time frame was found ~67%.

The IGNOA Director sent also two diagrams (31.01.2019), without further explanation, on a seismicity acceleration model in the target area of 225 km indicating “time of failure” either 2019.053 or 2019.070, which I favorably determined as 23.02.2019 and 11.03.2019, respectively, counting from 01.01.2019. An alternative determination could be based on decimal time, i.e. 19.01.2019 and 25.01.2019, which implies that the message was sent after the “time of failure” had already overdue.

As soon as the first e-mail message was received, the Committee Chair requested further information from the VAN team through confidential letters. However, VAN replied by just summarizing results in their arXiv report. In parallel, the Committee Chair sent a confidential letter to the Greek Civil Protection saying that for this EQ prediction there was no specific countermeasures to recommend beyond the usual ones. This confidential information reached regional authorities and fire brigade units in Western Greece. Unfortunately, the information did not remain confidential, on the contrary it got exaggeration through rumors spread in the area. Although no specific elements of the prediction were announced publicly, this is an important lesson learned: it is hard to keep confidential operational information which is circulated in written way among administrative services at various levels. On 05.02.2019 the convened Committee reviewed the situation and confirmed that because the prediction was not of operational use, countermeasures beyond the usual ones could not be recommended.

The next days, the IGNOA Director announced the prediction details through a series of TV and radio interviews saying that the EQ should be expected in a time interval of ~2 months (without specification on starting time), in a mapped circular area around Patras (from the map that appeared in the media I roughly estimated radius of ~110 km) and with magnitude of 6.0±0.5. Such a statement is significantly different from the one sent to the Committee. However, no scientific documentation was given on how those conclusions were achieved. Anyway, the public announcement of the prediction is another clear violation of the Codes of Ethics: “The news media are generally not the appropriate means by which to announce a prediction”. Moreover, the IGNOA Director failed to send his prediction to the IGNOA Scientific Council or at least to independent reviewers for evaluation before sending them to the Committee and spreading through mass media. This practice is another clear violation of the Code of Ethics: “Scientists should seek to have predictions confidentially reviewed…” (see above). At 9.30 am of 08.02.2019 the author of this commentary appeared at the Public Broadcasting System of the country and requested (a) for the Scientific Council of IGNOA to look after the predictions, and (b) for the Committee to reconvene for the evaluation of the new prediction version announced via mass media. The IGNOA Scientific Council never communicated with the Committee. The Committee reconvened on 12.02.2019 and reconfirmed that the prediction elements being available were not operationally usable beyond the usual countermeasures. The puzzle of the various prediction elements available to the Committee up to that date can be synthesized as follows: “SESs and EQ statistics indicate that an EQ of M≥5.5 is anticipated to occur in an area of radius 225 km around Patras in a time frame of 5.5 months counting from 03.01.2019. Now casting statistics indicates earthquake potential score of 80%”.

That the civil protection authorities failed to keep confidential that a prediction was submitted, combined with the public prediction announcement of the IGNOA Director, caused rumors and extensive social unrest in the entire Western Greece and the Ionian islands. Dr. G. Chouliaras, researcher at the NOA and collaborator of IGNOA as well as Committee member, is already under investigation by the Public Prosecutor since he uploaded in his personal Facebook account confidential letters of the Committee. Until writing these lines no EQ≥5.0 occurred in the target area and time frame. Regardless an EQ will occur or not, the chronicle of this EQ prediction is a clear case of multiple violation of the Code of Ethics adopted for EQ predictions by the international seismological community.


Allen, C. et al., 1984. Code of practice for earthquake prediction. IUGG Chronique, 165, Febr. 1984, 26-29.

Council of Europe, 1991. European code of ethics concerning earthquake prediction. In: Internat. Conf. on “Earthquake Prediction: State-of-the-Art, Strasbourg, 15-18 Oct. 1991”, Conclusions, 11-15.
Sarlis, N. et al., 2019. Geoelectric field and seismicity changes preceding the 2018 Mw6.8 earthquake and the subsequent activity in Greece. arXiv:1901.06658v1 [physics.geo-ph], 20 Jan. 2019, 8pp.

Other articles published in the IAPG Blog:

IAPG - International Association for Promoting Geoethics