251
Tectonophysics, 224 (1993) 251-255 Elsevier Science Publishers B.V., Amsterdam
Extended Abstract
Social impacts of earthquake
prediction in Greece *
Hiroaki Yoshii Bunkyo University, 1100 Namegawa, Chigasaki-shi, Kanagawa 253, Japan
(Received and accepted September 7,1992)
The VAN group (Professors Varotsos, Alexopoulos, and Nomicos) located four seismic electric signals @ES) on August 31st, 1988, and gave their prediction in telegrams to the Greek Government; to Dr. Tazief, a French scientist; and to a Japanese co-researcher on September lst, 1988. The text described two possibilities. One suggested that an earthquake was imminent 300 km northwest of Athens and that there was also an indication of an M5.8 event 240 km west of Athens (Fig. 1). Two days later, Dr. Tazief, who was convinced of the validity of these forecasts, made the prediction public, suggesting that contingency plans should be made for against the possibility of an earthquake at these centres. The announcement was made in the conviction that the Greek Government would fail to respond to such a warning. The mass media in Greece reacted strongly to this warning and published detailed articles described the reaction of seismologists to these views and quoting the VAN group’s comments. Lead stories in national newspapers on September 4th referred to VAN’S prediction of a large earthquake in the area of the Patras or Corfu. A news report of September 5th quoted Dr. Tazief as saying that the danger was very high. Further mass-media reports suggested that it was the moment of truth, as far as VAN'S predictions were concerned, the subject being treated as a serious social issue. The implication was that the VAN
* Based on Yoshii (1990). 0040-1951/93/$06.00
forecast had become a kind of pubhc warning, despite the fact that it was not an official announcement. On September 22nd, 1988, the predicted earthquake finally occurred at a point 250 km west of Athens, measuring a shock of M5.2-5.5. However, the reaction of the mass media was muted as damage was slight. Indeed, the effect was insufficient to prove publicly the success of VAN’Sprediction. However, on September 30th and October 3rd. the VAN group recorded two successive SES events. They again despatched two telegrams saying that seismic activity was
\‘
Fig. 1. The areas (dots) in Greece where the VAN Group predicted an earthquake on September l$t, 1988.
0 1993 - Elsevier Science Publishers B.V. All rights reserved
Knew a little
Fig, 2. A pie-chart suggesting that about two-thirds of a sample of people at Valtholomio and Zakinthos received prior warning of VAN'S prediction.
likely to continue at the same focal points. On this occasion the mass media displayed little interest. What was the local resident’s reaction during this period? For example, how many pea@@ were aware of VAN’Sprediction prior to the destructive earthquake on October 16th? Figure 2 suggests that about two-thirds of those questioned at Valtholomio and Zakinthos had received _priur warning i. Of those who knew of the prediction, 80% had read of it in the newspapers, whilst 25% had been informed by friends or acquaintances. Other sources of information were television (22%), radio @%), and family members (7%). The residents, on average, received the prediction via 1.5 different channels of communication. The most common source was the mass media (especially newspapers1 supplemented by word of mouth (Fig. 3).
<
“‘No
anxiety
[Italian example1 :ireek example1 (n=103) !n=103) Fig. 4. The degree of anxiety caused, on the left, by a warning in Italy and, on the right, by the Greek warning.
It might be asked what the reaction was of those who learned of the possibility of an earthquake. Figure 4 attempts to chart the degree of anxiety in the population. Of those questioned 26% expressed feelings of great anxiety. An identical percentage said they were “anxious”. The question is, was such a level of anxiety very high or not? The left-hand segment of Fiie 4 shows the levels of anxiety recorded for Italian citizens when they were warned of the possibility of an earthquake on television in 1985. Of those questioned 60% became “very anxious” and 30% expressed their feelings as “anxious”. It can therefore be seen that the degree of anxiety in Greece was not high when compared with the Italian case 2.
(n=103)
Fig. 3. Sources of information alerting the public to VAN’sprediction. This survey was conducted over August 2-8, 1989. Samples amounted to 69 and to 88 at Zakinthos. The survey was directed at residents in the centre of the towns and to passers-by on the basis of a questionnaire. The procedure was not baaed on statistkxl random sampling. collected at Valtholomio
’ In 1985, the Central Government in Italy suddenly decided to issue earthquake warnings just after a small earthquake occurred, based on historical data relating to the target area. The warnings said that a large earthquake might happen within 48 hr. In the event, no major earthquake occurred, so the forecast failed. See Yoshii (1989).
SOCIAL
IMPACTS
OF EARTHQUAKE
dt
[ 1talian
example1
253
PREDICTION
is
to happen
certain
-Al,ittle/no [Greek example]
Fig. 7. The degree of credence expressed in relation to the VAN prediction.
risk
Fig. 5. The perceived certainty as to whether or not an earthquake might occur. The italian case is on the left, the Greek example on the right.
This reaction may be related to how the Greek residents perceived the validity of the prediction 3. Figure 5 shows that, of those questioned, a large proportion were half in doubt that an earthquake would occur. Although 37% of the sample thought there was a strong possibility of such an occurrence, it was probable they were influenced by the fact that the earthquake had actually occurred. Moreover, only 6% of the population decided on evacuation, compared with about 60% in the case of the Italian warning. The actions taken by the Greek citizens are shown in Figure 6. Those who received the infor-
3 The survey was carried out after the earthquake and those questioned tended to answer as if they had greater faith in the truth of the VANprediction.
mation took between one and two courses of action. Half of those questioned attempted to obtain more information through newspapers and television, and, at the same time questioned their relatives and friends. Those who consulted their immediate family comprised 35% of the sample. Action to secure furniture and fixtures was widespread at Zakinthos, but low at Valtholomio. Evacuation was decided on by only 6% of the population. In short, almost the entire population decided to stay put, with a complete lack of panic or confusion. The response to the earthquake warning was almost entirely on the level of social discussion, a situation *which was quite different to that of the Italian warning. On October 16th, 1988, the predicted destructive earthquake occurred, registering 1M6.0. More than 2000 houses were destroyed or made uninhabitable. After the event, there was widespread acknowledgement in the mass media that the VAN
0 1,lriea
1 4. Secured
10 get mornation
an earrnquane
10
20
30
40 50 60 70%
warnil
furniture
‘7.Took measures
to strengthen
Fig. 6. Countermeasures
the house
taken by Greek resident after receiving the
(n=103) VAN
prediction.
254
Do no,t issue warnings
Le/t the children aged evacuate (n=154)
and
Fig. 8. Likely behaviour if residents again learn of a VAN earthquake prediction.
prediction had been valid. As a result, the population changes their attitude to warnings from this source. For example: (1) More residents were inclined to believe VAN’Spredictions (Fig. 7). Indeed, half of those questioned said they “believed” or “completely believed” the group’s warning. (2) A large proportion said they would react more positively to VAN’s prediction on a future occasion and said they would evacuate their families 4 (Fig. 8). (3) In relation to the level of concern, 80% of the residents expressed the opinion that the Greek Government ought to issue warnings following a prediction by the VANgroup (Fig. 9). The prediction by the VAN group highhghted two controversial issues; namely, the differences in formulating a strategy on earthquake countermeasures and different perceptions of VAN’Sprediction methods. Needless to say, the aim of earthquake prediction is to mitigate damage, but the current level of the science is not sufficient for its practical application on a general basis. the VAN group have attempted to achieve a break-
4
the author visited Iris’s prefectural office, he asked an officer in the technical division what he would do in the event of a VANwarning on a subsequent occasion. The reply was that he would do nothing in his present capacity unless the warning was issued by Central Government. However, on a personal level, he would evacuate his family. When
Fig. 9. Answers to a questionnaire as to whether or not the Greek government should issue warnings following a VAN pronouncement.
through and, in so doing, have reached a certain level of practical application. Traditional reseachers and administrators tend to insist on orthodox countermeasures to the earthquake threats, such as the design of buildings resistant to seismic shocks and training and education in appropriate behaviour in the event of an earthquake. Such authorities do not like to depend on such doubtful unproven “fuzzy” technology as earthquake prediction. Moreover, in many cases there are vested interests. On the other hand, the VAN group is only concerned with the discussion of the scientific feasibility of earthquake prediction as the result of SES detection. As a result, disputes are always conducted at cross-purposes. It wilI be seen that society might have no need of prediction technology if the population lived and worked in buildings designed to resist seismic shock, and if individuals could master the requirements of emergency response. In practice, the realisation of such ideals is very difficult. As a result, populations at risk, including that of Greece, have a paramount need for earthquake prediction, as well as for other countermeasures. However, the priority accorded earthquake prediction depends on national policy. The controversy between the VAN group and traditional seismologists, has, at times, been very emotional, making it difficult to obtain a consensus of expert opinion and to structure an allocation system for research funding.
255
SOCIALIMPACTS0~ EARTHQUAKEPREDICTION
The second controversial issue is that relating to the split in the perception of the VAN group’s prediction. The members of the team regard their forecasts as being on the practical level, a claim disputed by traditional seismologists and by the Government. If the standard of prediction technology is at an early stage of development, with a relatively low prospect of success, then society neglects such warnings. On the other hand, if the predictions are well-proven, then the Government will issue warnings without reservation. But where the predictions are within an intermediate band of probability and accuracy, as in the case of VAN'S work at present, then the population finds it very difficult to respond appropriately and unemotionally. Nevertheless, although it is inevitable that social dispute will arise in such circumstances, it must be regarded as a normal stage in the gradual acceptance of innovative technology. In conclusion, the author would advance two policy implications: First, earthquake prediction should be accepted as one of the major counte~easures against earthquake in Greece. Such an approach is seen to be especially effective in the case of rural areas. The reasons are as follows: (1) VAN'S prediction technology could be improved in the near future, based on its empirical rule characteristics. (2) Buildings in Greece, especially those in rural districts, are not well constructed to resist earthquakes. Furthermore, it would be very difficult to modify them to resist seismic shock. Earthquake prediction is therefore an effective countermeasure in the saving of lives in rural areas. (3) Even where a prediction subsequently proves to be false, residents in the target area
tend to be patient. They are more concerned that they might be unaware of the likelihood of a major earthquake than an alarm has proved to be false. Actual experience in the context of warnings in Italy and precautions in Japan 5 have indicated this. (4) Although those in opposition to the release of such warnings have claimed that the result will be panic, such a reaction seldom occurs. This fear falls within the category of the so-called “disaster myth”. Second, it is necessary to prepare a response plan to relatively imprecise “fuzzy” earthquake prediction. Such measures should be divided into three stages: Warning, precautions, and the issuing of basic information, according to the probability and accuracy of the prediction. It is natural to restrict such warnings to special cases where the prediction is likely to be accurate and highly probable, with every expectancy of widespread and damage. As a first step, such target areas should be confined to rural districts. References Yoshii, H., 1989. Social impact of earthquake warnings. Proc. 2nd Japan-US Workshop on Urban Earthquake Hazards Reduction. Yoshii, H., 1990. Social Impacts of Earthquake Prediction in Greece. Disaster Management, 3(l): 3-7.
’ The Izu-Oshima Kinkai earthquake, registering M7.0, occurred on January 14th, 1978, in Japan. Four days later, the Government of Shizuoka prefecture issued aftershock information which called on residents to take precautions against aftershock. As the warning was spread by word~f-mouth the text became distorted in its progression from person to person, becoming represented as a serious danger. AIthough many residents did evacuate, the population did not criticise the Governor’s decision to issue after-shock information.