Version classiqueVersion mobile

Risque, nature et société

Lucien Faugères
Christiane Villain-Gandossi

II. Étude de cas. La prévision des séismes (Grèce, Albanie)

Earthquake prognostics and prediction

Siasi Kociaj


La protection à long terme contre les dommages liés aux séismes est une nécessité, et dans cette perspective, la connaissance des mouvements du sol attendus est de première importance. Il s'agit de la compréhension des processus sismiques et de celle des modifications de la déformation pour le diagnostic sismique. Cette méthode comporte 4 axes principaux: la détermination du modèle complexe du foyer et l'évaluation du risque (sur le long terme, le moyen terme et le court terme), l'établissement d'un microzonage tenant compte des effets des mouvements et des effets de site, l'analyse des effets des séismes réels (l'expérimentation est en projet), enfin l'adoption des mesures permettant de limiter les dommages à venir (règlementation, planification de l'occupation du sol, mesures de préparation des populations).

Texte intégral

1We need long-term protection against future possible earthquake damage, and the knowledge of expected strong ground motions is of primary importance. Long-, medium- and short-term seismic hazard assessment requires: understanding the earthquake processes, modification of strong ground motions due to travel paths and local soil conditions, soil structure interaction investigations based on realistic ground motions parameters. The last ones may be used for risk analysis and long-term measures for earthquake protection. More information on such parameters of earthquake processes as the amplitude, frequency content and duration is necessary to predict strong ground motions in near and far fields conditions.

2Recently (Vogel 1986) a new approach for long-, medium- and short-term hazard assessment, risk evaluation and damage protection against it, called "earthquake prognostics" is developed as interdisciplinary research. Up to now four international seminars on these topics are held, and the next one will take place in Lagos, Nigeria, this year.

3This approach includes the following fields of research:

  1. Investigation of earthquake source processes,
  2. Earthquake hazard assessment by predictive modelling of earthquake ground motions,
  3. Risk analysis and evaluation,
  4. Measures and precautions to minimize damage by future expected earthquakes.

4This term "earthquake prognostics" has been introduced as the concept of earthquake hazard research with respect to long term protection against the earthquakes, closely integrating the seismology, engineering seismology and earthquake engineering together. From this point of view the organizational structure of our Seismological Centre is closed to this conception including three main research sections: Seismology, Engineering Seismology, and Earthquake Engineering. Based on our approach (Kociaj, 1989) this interdisciplinary concept imposes:

  1. The complex model of the focus and its long-, medium-, and short-term seismic hazard assessment,
  2. The evaluation of strong ground motions and of site effects by predictive modelling of strong ground motions at different levels,
  3. The seismic risk evaluation,
  4. Measures for minimizing damage by future earthquakes.

5The "earthquake prediction" aspects of research by different techniques are included in the first item of this approach, namely on the complex model of the focus and its medium, and short-term seismic hazard assessment.


6The determination of the complex model of the focus is of great importance for seismic hazard assessment at different levels and different recurrence periods. For this purpose it is necessary to distinguish three kinds of models for seismic hazard assessment:

Long-term model

7This model, called "seismotectonic" or geotectonic, considers the seismic hazard assessment, which is based both on geological and seismological data for the expected earthquake foci, their seismic potential and their time recurrence.

8Using geological mapping, neotectonics from geomorphological studies for mountainous regions, aerial photographs, and satellite images, the identifications of active faults and recent tectonic activity along them can be made. The studies on seismicity patterns and earthquake mechanism, through instrumental recordings, strong motions data and their comparison with geological ones leads to the determination of seismic active faults, seismic "gaps" or "quiescence" zones, which are important for predicting the foci of strong future earthquakes.

9It should be mentioned that such kind of investigations are already performed in our country including: fault plane solutions (Sulstarova 1986), seismic activity distribution (Kociaj 1986), seismotectonics and neotectonics (Aliaj 1979), evaluation of maximum expected earthquakes (Kociaj 1986) and up to the seismic regionalization of our country (Sulstarova et al. 1980), which represents the long-term seismic hazard assessment (on the national level). On the other hand, in the framework of micro-zoning studies (seismic hazard assessment on local level) in particular sites, these kind of investigations are being performed always for a surrounding zone of a site (generally with a radius of 60 km).

Medium-term model

10It is not the same to expect an earthquake in, let say, 100 years to come or in the next 10 years. So to reveal the foci of future earthquakes it is necessary to use many techniques of "earthquake prediction" research for the determination of visco-elastic properties and temporal variations of physical rock properties, associated with stress accumulations in focal zones as the key factor for earthquake generation. But in-situ stress measurements of regional stress distribution that may indicate earthquake-prone areas are still insufficient to monitor temporal changes of stress field, temporal variations of physical rock properties. Using geophysical methods, in many cases we tried to "predict the foci of next earthquakes". As the stresses cause the strain accumulation in viscoelastic media of focal zones, long-term changes of strain should be investigated by ground-geodetic surveys (triangulation, trilateration and levelling, sometimes combined with precise gravity surveys).

Short-term model

11This model is very useful especially for "pure earthquake prediction".

12Critical stresses prior to earthquakes should be easily observed using space-geodetic techniques. Short-term changes of strain can be investigated by continuous recording and monitoring of strain events in earthquake source regions, requiring high precision recording instruments and sufficient long term stability of recordings (tiltmeters, strain meters, gravity meters, sea-level and underground water level recorders). Using indirect methods (such as seismic wave velocity changes, resistivity changes, electro and magnetomechanics effects, anomalous earthtides etc.). Not all cases we tried "prediction" have been successful. In our country such kind of investigations were carried out only after strong earthquakes occurrences (Kociaj 1981), especially after that of April 15, 1979 (Ms=6.9). It was observed that at the eve of this earthquake temporal changes of b values (fig. 1) and Vp/Vs ratios (fig. 1) were significant for an area of a radius of about 200 km from the hypocentre.

13By radon gas emission measurements carried out in Austria far away from this focus, it was observed that the anomalous content of radon gas emission can be correlated with this earthquake.

14That means that the area to be taken into consideration for such kind of investigation, for large earthquakes should include the territory of several Balkan countries. From this point of view, only cooperative studies can contribute to solve this problem.

Complex model

15To assess quantitatively the seismic hazard of the earthquake foci a simulation model should be performed comparing the observed data with laboratory tests and theoretical aspects of earthquake generation.

16In some cases data can be used as a natural experiment on induced seismicity.

17From these data a complex model of earthquake foci can be derived. Based on such a model, the complete earthquake process or strong ground motions generated by earthquake foci can be determined. This process can be recorded by strong ground instruments in different field conditions. But the number of strong motion records on near-field conditions is very limited. As our strong motion network has been set up recently (from 1986) we have got only one good record, the one of the January 9, 1988 earthquake (M=5.4, Io=VII) which hit Tirana city. This record is a typical near field one (see fig. 2). Based on strong motion data of the April 15, 1979 earthquake (Petrovski 1980), and an instrumental one (Kociaj 1983) this earthquake represents a typical multishock. The highest acceleration (0.2g) was recorded on sand stones in Ulcinj. Comparing these two earthquakes it can be seen that their durations were different: 6 sec (for January 9, 1988) and 28 sec (for April 15, 1979). Their intensities were different, 7 and 9+ degrees, respectively.

18Another way to determine the complex model and to compare it with observed data is under implementation in our country, starting from strong motion records, using different stochastic models and synthetic accelerograms. Such an attempt was made recently for the two above mentioned earthquakes (Pitarka 1990).

19One of the main outputs of the complex model are their strong motions which can be used as input motions for further purposes.

Fig. 1. Vp/Vs ratios - april 15 1979.

Fig. 1. Vp/Vs ratios - april 15 1979.


20The evaluation of strong ground motion and site effects is linked closely with the influence of the transmission path and local site conditions on earthquake strong motion generated by the focus. Such a technique which is linked as well as with earthquake hazard assessment by predictive modelling of earthquake ground motion is one of the main tasks of engineering seismology for microzoning purposes. The experience gained recently by our strong earthquakes and of the world's (as that of Spitak in Armenia and in San Francisco) showed, once more, the importance of such investigations in active areas and especially for microzoning purposes.

21Such kind of investigations have been going on in our country from 1982. Up to now we have performed seismic microzoning of Vlora, Durres, Shkodra, Korea towns and Tirana city.


22The natural hazard becomes a risk when man, man-made structures, and other properties become involved. Risk analysis and its evaluation are basic requirements for developing the measures to prevent the damages that can be caused by future earthquakes. For this purpose different aspects can be considered:

3.1. The experience gained from past earthquakes

23The experience gained from past earthquakes by analysing aseismic measures and building materials used are very important. It should be mentioned that such studies are just at the beginning in our country. Very interesting are studies carried out for some strong earthquakes as those of June 1, 1905 (Kociaj et al. 1980), November 30, 1967 (Sulstarova et al. 1980) and April 15, 1979. Very interesting are the studies comparing outputs of microzoning studies for Shkodra town with the damage caused in this town by two earthquakes (June 1, 1905 and April 15, 1979). Very interesting is the experience gained by the January 9, 1988 earthquake which hit Tirana city just when the microzoning studies for this city were finished (Kociaj et al. 1988).

24But there are some aspects of risk analysis that we are not able yet to study as tests on dynamic loading of structures by experimental earthquake simulation, and studies on dynamic soil behaviour in laboratory conditions.

25As for the studies on site effects that can be caused by liquefaction phenomena, there is an experience gained from past observations in many sites and from microzoning studies of some coastal areas (as Vlora and Durres).


26Based on the above mentioned analysis, the most important measures to prevent the damage by future earthquakes may be considered as follows:

4.1. Aseismic regulations

27Recently in our country a new aseismic code is approved taking into consideration many of the outputs of seismic microzoning studies carried out in our country.

28These regulations are going to be applied from July 1990.

4.2. Proper landuse planning

29Based on dynamic behaviour of soils and structures, the proper land-use planning in seismic areas is very important. For this purpose, microzoning maps are very useful for engineering, and so the microzoning outputs for Tirana city and Durres town were used in new urban planning maps. After some destructive earthquakes it was observed that very strong aseismic measures to minimize the damage by future earthquakes were undertaken in epicentral areas, meanwhile from a seismological point of view (long-termmodel) these areas, for a certain period of time, are safer because the largest accumulated seismic energy was released during the previous strong earthquake.

4.3. Earthquake preparedness

30As is known, "earthquake prediction" is still very difficult and not yet reliable. So all the information on "earthquake prediction", if false, may create a lot of troubles in economic life and in human psychology. We should minimize the damage that can be caused by strong earthquakes (even when earthquake preparedness plans have to be worked out in active areas).


31As the density of the population increases,, seismic risk becomes eminent, and, as a matter of fact, seismic hazard assessment, seismic risk studies, and the necessary precautions to be taken must be extended and more accurated.

32It is of importance to treat the questions not yet attacked in the scientific research in our country. For this purpose it is necessary:

331. To determine the complex geotechnical models by new equipments and methodologies in order to determine the focal parameters, and by analytical solutions making use of the recordings of the strong earthquakes as well. The analytical and physical modelling of the processes taking place in the earthquake focus includes the synthetic accelerograms, subject of continuous comparison with the results of real observations or laboratory tests.

342. To start the complex studies, collaborating with other geological-geophysical and geodetic institutions of the country in order to determine the physical and visco-elastic models, intending to help seismic hazard assessment in time making the first steps in earthquake prediction. The classification of earthquake foci according to the stresses created so far is very important for the determination of the seismic hazard in the course of time. For this purpose it is necessary to widely use the regional geophysical methods and those of prospecting, in order to determine the physical properties of the focal zone and consequently the strength of the expected event. On the other hand, in order to determine the stressed state in the focus in the course of time, indirect methods used for prediction may be used.

35To determine the visco-elastic properties in the focus, the deformations of the Earth's crust in the focal zone area must be determined and measured continuously.

363. To assess the seismic hazard of a site, the studies on the transmission path must begin making use of the results of the geological-geophysical studies of particular regions and of the entire territory.

37To improve the methods and techniques for the determination of a geotechnical model, rapid methods such as electromagnetic and physical-mechanical properties of rocks in field conditions (including cross-hole) should be used.

38One may begin the study of the dynamic properties of soils and rocks in field and laboratory conditions especially for their behaviour in nonelastic stage.

39To extend and deepen the mathematical modelling methods one, two and three dimensional for dynamic response analysis of ground profiles during earthquakes and the influence of the geomorphological and topographic factors on and subterranean on the strong motions.

404. To undertake the studies for soil structure interaction as an important element to determine the seismic risk combining the laboratory modelling of the response of the structures in shaking tables with the mathematical model, selecting the most favourable models of these interactions.

41On the other hand, during the determination of the seismic risk, apart from the losses caused by the behaviour of the foundations and structures, special attention must be drawn to the quantitative evaluation of the dynamic instability of the soil which may cause such phenomena as: liquefaction of sands, the tixotropy of clays, the activisation or the active slips, rock falls and the active faults crops out on Earth's surface.

425. For earthquake risk reduction of expected events in relation to the precautions it must continuously improve the aseismic design considering the occurred events and the improvement of the building materials.

43To take close consideration in urban planning the results of the engineering seismology and earthquake engineering, clearly distinguishing the appropriate, improper and dangerous sites.

44Collaboration with the governmental and administrative authorities must prepare the population for such an event according to common programs for the reduction of human losses and economic ones.

45Earthquake prediction can be seen as a part of earthquake prognostics.



1. Sh. Aliaj. - Sizmotektonika dhe kriteret gjeologjike te sizmicitetit te Shqiperise, Disertacion, T., 1979.

2. Final Report of Working Group B, Vulnerability and Seismic Hazard, UNDP Project RER/79/014, Unesco, Skopje, 1982.

3. K. Ishihara, M. Ansal. - Final Report of Working group D, Dynamic behaviour of soils, soil amplification and soil-structure interaction, UNDP Project RER/79/014, Unesco, December 1982.

4. Kociaj, S. - Epiqendrat e termeteve te Shqiperise dhe rreziku sizmik, Disertacion, T., 1976.

5. S. Kociaj, E. Sulstarova. - The earthquake of June 1, 1905, Shkodra, Albania, Tectonophysics 67, 1980.

6. S. Kociaj. - Karakteristikat themelore te termetit te 15.4.1979 sipas te dhenave instrumentale, Termed i 15 prillit 1979, "8 Nentori", T., 1983.

7. S. Kociaj. - Some features of anomalous seismic activity before great earthquakes in Albania. Proceedings of the seminar on earthquake prediction and instrumentation, Thessaloniki, Greece, UNDP/UNESCO Project RER/79/014, 1981.

8. S. Kociaj - Mbi mikrozonimin sizmik, Qendra Sizmologjike, 1985.

9. S. Kociaj, LI. Duni, A. Pitarka. - Mbi mikrozonimin sizmik te sheshit te ndertimit ku do te ngribet hoteli i ri Vlore, Raport teknik, Qendra Sizmologjike, T., 1985.

10. S. Kociaj. - Rreziku sizmik i kores se Tokes ne Shqiperi dhe saktesimi i tij per sheshet e ndertimit, Disertacion, T., 1986.

11. S. Kociaj, A. Pitarka. - Raport teknik per perdorimin e metodes se elementeve te fundem per reagimin dinamik te elementeve topografike, Qendra Sizmologjike, 1987.

12. S. Kociaj, A. Pitarka. - Check up on the seismic hazard assessment taking into account the influence of local site conditions, XXIV General Assembly of ESC, P2, Sofia, 1988.

13. Mikrozonimi sizmik i qytetit te Vlores, Qendra Sizmologjike, 1984.

14. Mikrozonimi sizmik i qytetit te Durresit, Qendra Sizmologjike, 1985.
Mikrozonimi sizmik i qytetit te Shkodres, Qendra Sizmologjike, 1986.
Mikrozonimi sizmik i qytetit te Tiranes, Qendra Sizmologjike, 1987.

15. Projektkushti teknik i projektimit per ndertimet antisizmike, Qendra Sizmologjike, 1987.

16. E. Sulstarova, S. Kociaj. - The earthquake of November 30, 1967, Dibra, Albania, Tectonophysics 67.

17. E. Sulstarova, S. Kociaj, Sh. Alia. - Rajonizimi sizmik i RPSSH ne shkallen 1:500 000, "8 Nentori", T„ 1980.

18. E. Sulstarova - Mekanizmi i vatrave dhe fusha e sforcimeve tektonike ne Shqiperi, Disertacion, T., 1986.

19. A. Vogel - Earthquake prognostics, Proceedings of 2nd seminar, Berlin, 1986.

© Éditions de la Sorbonne, 1996

Conditions d’utilisation :

Cette publication numérique est issue d’un traitement automatique par reconnaissance optique de caractères.
Rechercher dans OpenEdition Search

Vous allez être redirigé vers OpenEdition Search