67

BULLETIN

of the International Association of ENGINEERING GEOLOGY de I'Associat,on Internationale de GEOLOGIE DE L'INGENIEUR

N~

21,

61 - - 80

KREFELD 1980 I

SYMPOSIUM ON NATURAL AND INDUCED SEISMICITY ~$5o Paulo, Brazil. 24-25 September 19791 - SUMMARY OF PAPERS SYMPOSIUM SUR SISMICITE NATURELLE ET PROVOQUEE (SF.o Paulo, BrOsil, 24-25 Septembre 1979) - SOMMAIRE DES ARTICLES

ABGE - ASSOCIA~TAO BRASILEIRA DE GEOLOGIA DE ENGENHARIA (Brazilian Association of Engineering Geology)* Introduction When this symposium was proposed, ABGE's objective was to survey the present state of knowledge about natural and induced seismicity in Brazil through reports from specialists in the fields of (;cophysics, Seismology. Geological Engineering and Structural Engineering.

A series of factors have obstructed a more objective technical study of the problem, including: the absence of adequate registers of the estimated 250 earthquakes, both natural and induced, that have already occurred, -

The proposed themes attempted to cover aspects ranging from the theoretical conceptualization of the origin of natural and artificial events (analysing the instrumentation, the recordings and their seismic interpretations) to the application of seismological knowledge to large construction projects, paying special attention to Brazil's specific situation. At the ,same time we appealed to Unesco's experience in this field, through Prof. Robin D. Adam's participation. In reality, what was the importance of the Symposium on Natural and Induced Seismicity to the Brazilian technical environment'? Initially, it should be remembered that little is known about seismicity in this country. In fact only in the past few years have Brazilian technicians dedicated themselves to a more objective study of seismicity, specifically induced (connected directly to the filling of reservoirs created) by large dams.

the absence of a detailed tectonic map. At the moment, the regional scale map available dates from 1977, which was made by the DNPM, -

These reports of natural seismic events, until recently incomplete, were radically modified in their methodology, due to the construction of numerous dams and the consequent occurrence of earthquakes in some of them after the filling of the reservoir. These earthquakes indicated a greater necessity of observation, interpretation and correlation of these phenomena. We should remember that in the past few years around a dozen small earthquakes resulting from the filling of reservoirs have been catalogued and are being studied in Brazil. As a consequence of the more accurate studies of seismic phenomena, the tendency to attribute these earthquakes to purely atectonic causes is decreasing. Finally it is important that seismic studies in Brazil be executed efficiently since it has always been considered a country without profound seismic characteristics. This consideration led many hydroelectric power plant constructions to neglect the influence of seismic activity on civil structures. This image has its origin in the fact that this country is situated in a region of old crystalline shields which are generally stable, non-seismic blocks. However, it is known that certain seismic features of larger faults and rifts located in these shields, can be points of occasional activity of a small or medium intensity. For this reason, it is necessary that seismicity studies in Brazil be extended and carried out in more detail. More information must be gathered about geotectonic conditions. The number and quahty of both seismological stations and instrument networks must be increased, in both areas of structural instability as well as at the sites of large civil constructions.

the absence of a national committee or organ to select and concentrate all lithological, structural and hydrogeologicaI studies, in order to initiate and complete a geotectonic map of Brazil, a lack of integration of the methodology and terminology used by different specialists in the area, from geologists and seismologists to seismic engineer~

At the end of the symposium the positive balance was the following: -

The records of natural earthquakes in Brazil date from the 16th century, with references in various works and newspapers. However, due to inadequate instrumentation there were not any satisfactory recordings, and estimates were made from verbal reports or from in situ inspections after the event. Many earthquakes were attributed to particular geological phenomena, such as the settling of layers, and thus were assigned a non-tectonic origin.

the absence of a sufficient number of seismological stations. In Brazil there are stations in S~o Paulo, Rio de Janeiro, Natal and one in Brasilia opened in t971,

a large number of participants (220 technicians), the presentation of reports of a high technical level coveting various aspects of the problem,

-

the elaboration of the first Brazilian publication on seismicity, to be distributed free to ABGE members,

-

-

the seminar by Woodward-Clyde Consultants, San Francisco, California on reservoir induced seismicity,

-

the participation of Unesco, through Prof. Robin D. Adams, who analyzed natural and induced seismic risks,

-

the most important proposition, which was the implementation of the Brazilian Seismological Committee, with the objective of bringing together domestic experience and data on this subject.

This committee should be multidisciplinary and will be composed of technicians from universities, state and private companies (there are already more than 20 bodies interested). A working group coordinated by Professor Dinis da Gama, representing ABGE, was constituted to formalize the setting up of the Brazihan SeismoIogical Committee. The primary aim of the committee will be the compilation of a Brazilian seismic risk map, from which, through the interpretation of geological structures (faults, fractures, geosutures), the types of rock (lithologies), of the seismic characteristics, it will be possible to distinguish the areas of greatest and least seismic risk in Brazil. Also the ABGE wishes to thank CESP, specifically Dr. Jos6 Walter Merlo and Dr. AbrahAo Fainzilber and Dr. Tuyosi Itoo of AECESP for their organizational and financial assistance. Also we wish to acknowledge their openness in discussing induced seismicity which occurred in their dams at Paraibuna and Capivari, an attitude different from that of many companies and national departments which besides not supporting initiatives such as these refuse to discuss their work and problems.

* Postal address: I.P.T. - D.M.G.A., Cidade Universitaria, Caixa Postal 7141, 01000 S~o Paulo S. P. (Brazil)

68

UNESCO's Consultant believes that those responsible for dams, together with other specialists, should involve themselves with these risks of induced seismicity resulting f r o m the construction of dams.

N a t u r a l and i n d u c e d seismic risks Dr. R o b i n D. A d a m s U N E S C O , and The (England)

International

Seismological

Centre T h e e v o l u t i o n o f s e i s m o l o g y in Brazil

The risks of natural seismicity are real in any place and therefore, are not unusual in more stable areas such as Brazil.

Prof. Dr. H e r m a n n H a b e r l e h n e r

The risks of induced seisms, however, have to be studied carefully for it would be foolish not to incorporate antiseismic precautions into structures associated with large reservoirs. Unesco recommends the installation of seismic controls in all reservoirs with a depth of over 100 metres and a water volume of over 109 m 3 .

The first record of a tremor in Brazil dates from 1560 in S~o Paulo. Father Sim$o de Vasconcellos reported a heavy storm with thunder, lightning, strong winds and a "terrible" earthquake which shook all the settlements of the port ofS~o Vicente.

Engevix S. A.

Robin Adams, who is also a member of the International Seismological Centre of England, pointed out that his experience on the Latin American continent has been in countries neighbouring Brazil. These countries tend to suffer strong earthquakes so that everyone is aware of the existence of seismic risks. As a consequence it is not necessary to request the authorities to study earthquakes in order to minimize their consequences.

The second seismic occurrence in Brazil was in 1690, reported by Father Samuel Fritz. This happened in June affecting the left bank of the Amazon river, from the Rio Negro 300 leagues upriver.

tte praised the promotion of the Symposium in this country "where the seismic risk is not so evident".

In the XIX century, 53 earthquakes were registered by historians and the local press. Of these, 20 had an estimated magnitude of between IV and VI corresponding to accelerations of 0.07 g and " causing great damage.

The seismological specialist stated that there are many examples of destruction associated with natural seismic events. He has studied three large earthquakes. One of them at Haicheng, China in 1975 was the first major earthquake to be predicted, and caused much material damage. The others were recorded in New Zealand in 1968 with a magnitude of 7 and San Fernando, California (U.S.) in 197l. How are the effects of a natural earthquake related to its position and occurrence'? Adams explained that such phenomena occur principally at the edges of the large continental plates of the Earth's surface. Many natural earthquakes are felt at considerable distances. One in Alaska in 1964 was felt more than 1,000 km away, while an earthquake in Timor, Indonesia affected buildings in Adelaide, Australia, more than 3,000 km away. Also the earth tremor in Colombia in 1970 was felt here in Sgo Paulo, that is, at a distance of 3,000 km. tte reached the conclusion that large earthquakes can propagate their effects for long distances. For this reason, he emphasized the importance of the determination of an earthquake's effects, considering geological conditions, that is, the types of rocks and their structures (faults, fractures, ete). In regions like Brazil, where seismicity is low, the evaluation of seismic risk is difficult. The greatest danger, however, according to Adams, is related to isolated or sporadic tremors and to the effects of extremely violent earthquakes registered at long distances, which can be particularly damaging to structures. What can be done for these regions? Robin Adams answered: Any natural feature which demonstrates recent geological movement must be avoided. For regions where there is already evidence of seismic occurrences, look for tectonic features which may be found in other areas, suggesting that these too are seismic. Special attention should be paid to the foundations and a minimum degree of anti-seismic resistance should be incorporated in all structures. Induced seismicity, according to Adams, is mainly triggered by the filling of reservoirs and to a much lesser extent by nuclear explosions and underground mining. More than fifty earthquakes due to the filling of reservoirs have been registered. The most important of these are: China 1962, Kafiba, Central Africa in 1963, Greece in 1966 and Koyna, India in 1967, all of which had a magnitude greater than 6. After referring to Uneseo's suggestion, that is, that all reservoirs should be controlled and measured, Adams recommended the following studies in areas which could be affected by induced seismicity: a) a study of the historical seismicity of the region in which the reservoir is to be located, b) a geological and geomorphological survey of the reservoir and surrounding area to identify potentially active structures (faults, etc.); c) microseismicity sampling. Besides these studies, during the first two years of filling, scientists should undertake a geotectonic survey of the area's features. d) installation of a permanent seismograph network;e) undertaking of a geodetic survey and the installation of accelerographs to record the more violent tremors.

In the XVII century various historians registered 9 seismic occurrences. The tremor which shook Cuiabilon 28/I0/1746 with the probable intensity of IV - V coincided with the earthquake that destroyed Lima, the capital of Peru.

There were 2 earthquakes in Cear$, with a magnitude above VI, one in S~o Luiz and one in Cuiab~. The damage included destruction of wails and the opening of fissures in the ground. In 1910 J. C. Branner published the first study on earthquakes in Brazil and, in 1920, published a r e p o r t on 56 earthquakes which occurred between 1560 and 1912. tlilgard O'ReiUy Sternberg presented a report, in 1953, relating the occurrence of 39 seismic events in Amazonia between 1690 and 1953. The first two seismic shocks registered by a seismological station in Brazil occurred in 1915. One of these shook the region of the confluence of the I ~ and Solitudes rivers. The other occurrence was an earthquake of magnitude 7.25 in the Rio Purus region, corresponding to an acceleration of * 0.6 gThe 1915 seisms received very little publicity, the same happening with the other 87 tremors which shook the sparsely populated regions of Acre and Amazonia. Presented in Tab. 1 are the registers o f earthquakes between 1560 and 1977, distributed on the basis o f the units of the Federation and geographical regional division. Earthquakes of higher intensity, which could cause material damage are distinguished. These tremors, with few exceptions, were registered b y seismographic stations. It is to be noted that seismic activity in the north (with a lower population density) far exceeds that i n other regions, corresponding to 41 per cent of all seisms registered in Brazil 71 per cent of the greater shocks registered in Brazil occurred in the north, principally in the state of Acre. In 1962, in a r e p o r t on the viability of a hydroelectric project near Rio Branco (Acre) to the Water Division of the Mining and Energy Ministry, the following observations were made: "Earthquakes, hardly known in o t h e r parts of Brazil, affect the urban zone of Rio Branco quite frequently. At the end of 1961 an earthquake produced a severe fissure in a house on MaL Deodoro Rd. This fissure went through the yard of the house in a NW-SE direction, following the same general direction of the Acre River, near where the town's airport is located. The intensity of the tremor was greater along the alignment o f t h e fissure than in other parts of the city. From the damages observed in this brick house with a wellconserved pavement, the earthquake could be classified as VI - VII on the Mercalli-Sieberg scale. The seismic phenomenon in the state of Acre can be explained by the geographical position of the state. This region is in the Lower Andes Zone and is affected by the tectonic instability of the range. It is very possible that this instability also provokes a continuous, slow ascension of the land, thus explaining the forementioned coupling of the rivers." In the same report we spoke about the extensive landslides on the banks of the Acre river. The banks axe formed of montmorillonitic claystones and the slides are facilitated by a system of diaclases. We also think that seismic tremors contribute to the instability of the land in Rio Branco and that there is liquefactiofi of softs in the area.

6g

REGION

NORTH

STATE

Acre Amapa Amazonas Para Rondonia Roraima Subtotal

NORTHEAST

EAST

CENTER WEST

SOUTH

N ~ OF EARTHQUAKES OBSERVED 58 5 31 5 I 3 103

Alagoas Ceara Maranh'~o Paraiba Pernanbuco Piaui R. G. Norte

0 17 4 3 11 0 10

Subtotal

45

Bahia Esp. Santo Minas Gerais R. de Janeiro Sergipe

12 1 33 8 0

Subtotal

54

Goias Mato Grosso

2 11

Subtotal

13

Paran~ S. Catarina S. Paulo R. G. Sul

I1 2 15 8

Subtotal TOTAL

EARTHQUAKES RECORDI-D WITIt INTENSITY OR MAGNITUDE > 5,2 24 0 6 0 0 0 Subtotal

30 0 2 1 0 1 (Atlantic OceanJ 0 0

Subtotal

4 1 (Atlantic Ocean) 0 0 1 (Atlantic Ocean) 0

Subtotal

2 0 3

Subtotal

3 0 1 1 1

36

Subtotal

3

251

TOTAL

42

Tab. 1 The fact that until today the bulk of the public are aware only of Ixemors of low intensity combined with the lack of registration by domestic seismographic stations are the principal causes of the low level of interest in seismology in Brazil until recently. This lack of interest is also a reflection of the reduced number of seismological and seismotectonic studies that have been published. We know of 15 works which have appeared in 66 years. The majority of these works were published in this decade. Also there are a half dozen reports on seismicity at the sites of large construction projects, The growing concern with seismicity is a consequence of the shocks resulting from the filling of reservoirs created by dams, and also by the construction of nuclear power stations, the safe operation of which requires special precautions. At present there are 4 seismographic stations, located in Brasilia, Natal, Rio de Janeiro and S~o Paulo. Another station is being installed on the island of Trindade and a station is projected for Manaus. The Brasiha Seismological station is 20 km to the northeast of Brasilia and is operated by the department of Geosciences of the University of Brasilia. This station (ESB) uses 2 seismographic systems, the South American Array System (SAAS) and the World Wide Stan "dard Seismograph System (WWSSS). With the prevailing geological conditions of the Central Plateau and its distance from the Atlantic ocean, the SAAS (which began regular operation in 1972) is the best device for monitoring seismic activity in Brazil. S t r u c t u r a l and g e o t e c t o n i c p i c t u r e o f n a t u r a l seismicity Prof. Y o c i t e r u Hasui IPT S/A The drifting of continents has been admitted by many for a long time. To consolidate this idea, geological, paleontological and paleoclimatic evidence on the continental areas has been collected from the beginning of this century. L Wegener presented his great synthesis in 1915, proposing the existence of Pangea, which frag-

mented in the last 200 million years with the various parts moving to their present positions. From 1950 on, with the study of magnetism and the ocean depths, the hypothesis of the expansion of the ocean depths has gained credence. This hypothesis was proposed by I-[. H. Hess and formalized by R. S. Dietz in 1961. According to the theory, the continents move passively, maintaining continuity with plates which grow in mid-oceanic ridges and which are consumed in the undersea trenches. In 1967168 the akeady classic articles form Mackenzie and Parker, Morgan, Le Pichon, Isacks, Oliver and Sykes appeared. These articles defined the plates, their limits and kinematics and established what came to be called the plate theory or the new global tectonics. This theory unlqed multidisciplinary knowledge and opened investigative fronts with stimulating revelations, revolutionizing the geosciences. The Earth has a rigid skin, the lithosphere, which is from 50 - 100 km thick, and which covers the asthenosphere which goes down to around 700 km in depth. The lithosphere includes the crust and part of the mantle. Its conception is geological as opposed to compositional The crust is the external part with an average thickness of 35 km on the continent (up to 60 km in mountain ranges) and 6 km in the oceans (up to 25 km in the mid-ocean ridges). Its base is deFreed by the change of seismic velocities from 6.8 - 7.2 to 8.0 - 8.2 km]s (Mohorovicic's discontinuity). The lower limit of the lithosphere corresponds to the transition to the low seismic velocity zone of the mantle. In compositional terms, it can be said that the upper crust in the continental areas is granodioritic, and the lower part as well as in the oceans, is gabbroic / basaltic, amph~olitic or granulitic, while the mantle is peridotitic / serpentinitic or eclogitic. The lithosphere is segmented into plates which are generated in rift zones associated with the axis of the mid-oeean ridges and are consumed in the subduction zones, to which are related the submarine trenches and island arcs. Another type of limit is that represented by zones in which the plates move against one another, without

70

renewal or destruction and which correspond to tile transforr:.~ faults. The rift system of the mid-oceanic ridges stretches across the oceans ~itb ramifications which reach from the Arctic, Indian and Pacific Oceans respectively. "rt~e transform faults dislocate segments of the ril't system, producing the curvilinear geometry of the mid-oceanic ridges. These faults have directional dislocations (not transcurrentl following the rotation axis and with an angular velocity which depends on the situation of the segment in relation to the corresponding Eulerian pole. For the transform faults of the South Atlantic the pole is located midway between Labrador and Iceland. The subduction zones extend from the ocean trenches with variable inclination to depths of up to 700 km. They are distributed in South-l.iast Asia, a.round the Pacific, Caribbean and, it a o u l d appear, around the S~uth of the Atlantic. The continents axe nothing more than masses inte,m-ating these plates and are not subductcd, therefore being able to approach one another and collide. The dislocation speeds of the plates are calculated in various ways lmagnetic anomalies, surface seismic waves, direct measurementsl arid in some places reach up to 6 cm / year. The dislocation direction is determined by seismological methods, taking transform faults and rifts into consideration. If the geometry of the larger platcs (Africa, Eurasia, North America South America, The Pacific, India - Australia, Antarctica, Nazca Cocos and the Caribbean~ as weIl as their movements, has been reasonably well established, the same cannot be said for the causes and mechanisms. As for the interiors of plates, they are not aseismic as generally described. The seismicity of the region between the Himalayas and the Pamir - Baikal and the region of the ttawa'ian islands is impressive. A low frequency and generally less intense activity occurs throughout. One example is the eastern half of the United States where high intensity seisms were registered (10 mm in the extreme northeast in 1663, 12 mm on the borders of Kentucky, Missouri Illinois in 1812, 9 -- 10 mm in South Carolina in t886), while these areas figure as Zone 3 on seismic risk maps. Interplate ~ismicity has little or nothing to do with that on the edges and this is significant having been relegated to a secondar} level by specialists, ttowever, the great part of continental areas fall into this context, making seismological and neotectonic studies necessary for the understanding of natural seismicity. Brazil is situated on the western part of the South American plate, on the western part of which there is the subduction of the Nazca plate and where the Andes chain rises. Various records of seisms have been made in Brazil since 1912, the most complete and up-to-date being that of H. Haberlehner. All the records delineate various areas with seismic activity whose correlation with structural units and geotectonic entities should be perfected, at least to incorporate more recent data, with the purpose of elaborating a seismic risk map. In the western part of the Amazon region there are seismic loci in the area of the sedimentary basin of Acre. These are between 545 and 700 km deep and have magnitudes between 3.7 and 7.25 and are linked to the Benioff zone under the Andes chain. In the north of the country, the seismic areas coincide with the basins of the Mid-Amazon, the Takutu, the Maraj6 .and S~o Luiz/Barreixinhas. This seismicity is linked to the still active subsidence process. In the Brazilian North-East and the East of Bahia a relationship between the seisms and the pre-Cambfian basement faults has been established. The same is true for the centre-south of Minas Gerais and the north of Goi,qs. Such areas coincide with the moving belts developed around the edges of the Craton do Pammirim, during the Proterozoic (2.5 to 0.5 billion years). The most imposing features in these stretches are the geosutures reflected in a fault system. These geosutures had transcurrent movements at the end of the Brazilian Cycle (Upper Proterozoic) and axe partially redeveloped in a normal character in the Wealdenian reactivation of the SouthAmerican Platform. An analogous situation is seen in the territory of Mate Grosso, specifically around the Pantanal where the Brazilian belt is curved m such a way as to be parallel to the NW-SF geosuture of the oldest basement. The Pantanal depression developed here but its structural canals have not been established. The seismic areas of the east of S~o Paulo / south of Minas Gerais /

Rio tie Janeiro. of the south of S~.o Paulo / cast of Paran5. / ca~t ~f Santa Catarina and of Rio Grande do Sul coincide with stretches of the Mobile Coastal Belt, which suffered uplifts from the Upper Jurassic. These permitted the formation of dykes, the intrusion of alkaline bodies in the km~er and Upper Cretaceous / l'ertlary. the formation of taphrogenic basins in the Tertiary / Pleistocene and coastal morphogenesis. The relationship of earzhquakes with faults has not been verified, so that the definition of active faults in areas segmented by systems of discontinuities is very difficult. Attempts at correlation ~ i t h mapped faults were made in the east of Silo Paulo / Rio de Janeiro, and the coincidence of epicentrcs with the fault-line of Cubat~to (wich includes the kancinha, ltapeuna, Cubat~.o, Taxaquara I partially~ and ParMba) were suggestive. This brief outline here refiects the paucity of structural and tectonic information available in Brazil as well as the novelty of seismological studies. It is to be desired that, given the significance of seismic knowledge, progress in this sector of the science will be accelerated in the near future. Induced seismicity Geol. M. Sc. S o h r a b S h a y a n i Hidroservice Engenharia de P r o j e t o s Ltda. Propagation of seismic waves in rocks and structures (a) Mechanisms of propagation A seismic event is an oscillation or vibration of the Earth's surface caused by a transitory perturbation of the elastic equilibrium in the rocks on tlle surface or below it. The natural event can be of tectonic, volcanic or collapse origin. In considering rocks as elastic, isotropic and homogeneous material,. the mechanism of the seisms is explained by the elastic repercussion theory of Reid. According to this theory, the rock fractures when submitted to elastic deformations greater than it can withstand. These deformations result from slow dislocations of the lithospheric plates causing an accumulation of forces on the edges of the plate. These forces increase until they reach the rupture point of the rock. When the fracture occurs, energy is liberated in the form of elastic waves. Two principal types of elastic waves are identified and registered by seismographers: "'body waves" and "surface waves". Among the body waves, Primary waves P (compressional) and secondary waves S (shear) are distinguished, The P waves are faster, with the material particle moving parallel to the direction of the propagation of the wave. The S waves are slower with the material particle moving m a plane perpendicular to the direction of the propagation of the wave. The velocity of the P waves (vp) as well as the velocity of the S waves (vs) is expressed as a function of the elastic constants of the ma t e ri a l (b) Characteristics of soil vibration caused by seismic events The initial fracture point during a seism is called the focus and the surface point above the tbcus is the epicentre. The source of enemy liberation follows the fracturing of the fault as long as the rupture is propagated in a determined dkection, which depends on the magmtude of the energy liberated. The body seismic waves P and S, and Rayleigh and Love surface waves are generated by complex physical processes which occur during and after the fracture. The high frequency body waves (1 10 Hz) are the most important energy carriers up to 120 km from the focus. Generally, a normal or reverse fault will generate proportionally more energy in the form of P and Sv waves and a transcurrent fault will generate proport i ona l l y more energy in the form of SM waves. At greater distances from the fault (> 100 kin) the Rayleigh and Love waves are the more important energy carriers. -

(c) Interaction in the soil - structure system After the occurrence of an earthquake, the seismic waves propagate themselves through the layers of soil, presenting a signal related directly to the natural periods of the geological structure. This interaction affects the amplitude and spectral response of the soil vibration transmitted to the structure. The structure moves as a

71

function of its natural vibration patterns and other characteristics, stimulated by the soil vibration. As the structure moves it produces forces of reaction in the soil to which axe added those produced by the seismic waves. In this way the result of the incident seismic waves and the result of the force due to the structure are registered. Detailed knowledge of soil-structure interaction is more important for rigid structures than for flexible ones and for loose soil than for firm or consolidated terrain.

Induction of seismicity in civil construction The civil engineering works that could provoke small tremors or moderate to strong magnitude seismic activity are explosions for rock blasting, rock slides, water injection under pressure, the opening of artesian wells and the construction of large water reservoirs. Reservoir induced seismicity Seismicity is the result of a very slow geological process, and a long observation period is necessary to deduce the laws which govern such processes. Practically all the instrumental data available are from the XXth century which isn't enough to establish firmly the seismological pattern for each region of the world. Lane (1971) was possibly one of the first authors objectively to approach the subject of seismicity induced by the formation of artificial lakes. It is known that the filling of many large reservoirs has not caused significant seismic activity. However, in the cases of Kariba, Koyna, Kremasta and Hsinfengkiang magnitudes of 6 or more have been registered.

t~on. With rare exceptions, the occurrence of an earthquake received more journalistic speculation, more often than not the phenomenon being put down to the settling of layers. In these circumstances, the real nature of seismic phenomena in Brazil was unknown m popular media and its probable effects were ignored in civil construction. Later the construction of large scale dams, a consequence of the rapid economic development of Brazil from the 1960's, and more recently the building of nuclear power plants, obliged the technicaladministrative areas to make proiect and safety decisions involving and dealing quantitatively with this factor. The initial search for foreign consultants to help with seismic problems brought only partially satisfactory results, because of their ignorance of local geological and tectonic conditions and corresponding bibliographies. Similarly there were logistic problems with t,~strumentatien and field work under local conditions. At the same time, companies such as CEMIG, CESP, FURNAS, ELETRO-NORTE, NUCLEBRAS and more recently ITAIPUBinacional, signed agreements with the University of Brasilia to develop studies on natural and induced seismicity. This faciLitated a real participation of the UnB in the technological development of the country, allowing the application of its research, developed in a purely academic sphere, to local problems of national importance. The success of this association between the F. U. B. and Energy Companies is an auspicious example of the important role that the Brazilian University can play in the country's development when it is given administrative mechanism s which facilitate this participation.

These were the preliminary observations noted by Lane: (a) seismic activity seems to result from the filling of the reservoir and not from the actual construction of the dam: (b) in a locality where there is seismic activity, the number and magnitude of tremors increases with the rise in the water level; (c) the activity is not directly related to the weight of the stored water: (d) the level of seismic activity is not directly related to the maximum depth of the reservoir; (e) the seismic activity can start at the same time as the filling of the reservoir.

At the 2nd Colloquium on Inspection and Observation Systems for Dams at Salto Santiago on November of 1978, with the support of ElectrosuL, the initiative of Electrobras and Cemig to install a seismological centre in the Brasilia Seismological station was approved. The objective of this centre would be to create a data bank, standardize norms for data collection, share seismological instrumentation experience, orient research and thus coordinate efforts, in short, taking the maximum possible advantage of individual investments in this area of dam instrumentation.

Conclusions

Electrobras has not yet officially approved this project but the UnB, for all practical purpose, already fulfills a large part of the objectives approved in Salto Santiago.

At present, the filling of a reservoir is not necessarily the cause of localized seismic activity. Only a certain number of reservoirs demonstrate seismic activity associated with the formation of artificial lakes. Studying the relationships between induced seismicity and the construction of hydroelectric plants it can be established that there is no established seismicity pattern. The correlations between the filling levels do not check. The tremors can occur underneath the reservoir, up or down river from the dam or in the area of faults already known to exist around the reservoir. The tremors can increase in magnitude until the occurrence of a larger seism or there may be a 'large number of smaller tremors with no significantly large tremor being registered.

Summary of cases in Brazil In Brazil, some seismic phenomena have occurred, probably induced by the filling of dams such as Capivari-Cachoeira, Carmo do Cajur6, Porto Columbia and Volta Grande, Capivara, Maribondo and Paraibuna-Paraitinga.

Probably the first cases in Brazil where seismic activities connected with the filling of reservoirs were suspected was in Or6s (Cear~i) and Carrno do Cajuru (Minas Gerais in 1970). Taking into account tire fact that the activity in A~ude de Or6s was not studied instrumentally, Carmo do Cajuru is the first Brazilian reservoir where seismicity studies were developed, and will be discussed in detail in this paper. The most important, recent cases of induced seismicity are Volta Grande, Porto Colombo, 1974, Capivara, 1976 and ParaibunaParaitinga 1977. On the otherhand, among the most important examples of the filIing of reservoirs unaccompanied by seismicity we can mention ilha Solteira, Jupiff, Agua Vermelha, S~o SimS'o, Promis~o, Salto Osori6 and Sobradinho. S~o Sim~o (Cemig) was the t-trst BraziLian reservoir to be instrumented before being filled.

Instrumentation of dams. Collection and treatment of data

In comparative terms it can be stated that seismicity in Brazil is much less than the high seismicity in Japan, California or northern Italy; also it is less than the moderate seismicity of the eastern United States and the Reno Valley region.

The seismological instrumentation of a dam should begin before it is t'flled. Ideally the seismicity of the locality should be known with a high degree of certainty. Thus the seismic risk could be quantitatively evaluated and included as a parameter of calculation in the dam structure. This would be possible if one had historical data and local registers covering a century or more. As this is impossible in most places in Brazil, the installation of one to three seismographs in the region of the future reservoir should be done from 2 to 3 years before the projected construction. The objective of this instrumentation would be to verify any possible variation in the seismicity level coincident with the t-filing of the reservoir, thus indicating induced seismicity.

There is no record of human losses produced directly by seisms in Brazil. The maximum intensity of an earthquake in Brazil has not exceeded VI degrees on the Modified Mercalli Scale. Given this low seismicity, until recently seismological studies received little atten-

During the filling of the reservoir a network of seismographs covering the whole region of the reservoir should be installed. This network should have at least 3 seismographs, to be able to locate epicentres accurately, ftowever, 5 or 6 seismographs would be an

Studies on induced seismicity developed at the University of Brasilia Dr. J. A. Mendiguren U n i v e r s i t y o f Brasilia

Introduction

72

ideal minimum to assure precision in the localization of the seismic focus and adequately cover the whole region of the reservoir. The following are the successive stages of the installation and operation of seismographs in the reservoir area. 1) Choice of possible locations for the installation of the seismographs. 2) Definition of the configuration of the seismographic network. 3) Choice of the type of instrument. 4) Installation and maintenance of the instruments. 5) Daily operation of the seismographs. The locations most suitable for the installation of seismographs are those where the natural vibration of the Earth's surface is low, allowing the operation of the instruments with high sensitivity. In general, the most suitable locations are outcrops of compacted rock a long way from highways, villages, rivers, waterfalls, large trees, t h e sea, intensive agricultural activity or any other source of vibration. The people who operate the seismographs and make a preliminary interpretation of the seismograms are generally given a week's training at the UnB. The installation, calibration and maintenance of the seismological instruments is generally handled by UnB technicians; however, in some cases, such as Furnas, the actual company takes care of this work. In the case of reservoirs with induced seismicity, it is convenient to distribute questionnaires to the local population to get information about the intensity of any shocks fell In Carmo do Cajuru, Cemig maintains a group of around 20 observers. The information from these sources is analysed at UnB. An important aspect in data collection is the control of explosions around the reservoir. Forms containing the exact location of the explosion, the time and the charge are a great help, especially at the beginning of observations, in order to distinguish the registration of natural and induced events. The registers from different reservoirs are sent to UnB where they are analysed. After the data are processed by computer the UnB sends monthly or bi-monthly reports to the companies with data on epicentres, magnitudes, acceleration evolution of seismicity and its relationship to the geology and tectonics of the region, etc. Other special reports are given, at any time, when solicited by the companies. Apart from this, the seismological station maintains telephone contact with the operators of some reservoirs. Installation of the accelerograph Although not directly linked to the study of induced seismicity, the installation of accelerographs can also be considered in the seismological instrumentation of a dam. The accelerograph will only be used in the event of relatively strong movements, which would produce accelerations greater than a certain predetermined value, for example 10 cm/seg 2 . The registers of an accelerograph allow the measurement of forces to which a dam is subjected during a seismic movement. This information is important for future structural design. Generally two accelerographs are installed, one on the top and the other on one of the abutments or the base of the dam. In areas subject to high seismicity more accelerographs are installed. This goes also for concrete structures. In these cases the accelerographs are installed at differents heights in the dam. The present awareness of Brazilian hydroelectric companies indicates that at least one accelerograph will be installed on any future dam over 50 metres high. Studies in Carmo do Cajuru The Carmo do Cajuru reservoir, belonging to Cemig, is located approximately 100 km to the SW of Belo Horizonte, on the Pars River near the city of Carmo do Cajuru. The dam is 20 metres high and the reservoir's capacity is 193 x 106 m 3. This reservoir was filled in 1954; however, seismic activity only began in April 1970. The largest seism, registered in January 1972, had a magnitude of 4.6 on the Richter scale. In 1975 the UnB and Cemig installed a network of five seismographs. Most of the seisms occur at depths of less than 1500 metres. Their epicentres are, in the majority of cases, located outside the area covered by water. A

study on the possible correlation between the water level and seismic activity showed some affirmative results in 1975 and 1976. However, recent data do not confirm these early results. The focal mechanism of the earthquakes indicates that the seismic activity corresponds to transcurrent faults with nearly vertical fault planes and azimuth at about N 15 ~ E. This direction coincides with a relative maximum of fracturing frequency observed around the reservoir. The seismicity is produced by the liberation of tectonic compression forces which have an approximate direction N 60 ~ W. This direction coincides with the direction of the tectonic forces generated in the Atlantic Ocean range which dislocates the submarine plate in the direction of the Pacific. One special characteristic of the seismicity in Carmo do Cajuru is that it began 16 years at'let the reservoir was filled. Another is the fact that the majority of the epicentres are beyond the watercovered area. The seismicity of Carmo do Cajuru can be explained in the following way: the filling of the reservoir stabilized the transcurrent faults located immediately under the area covered by the water. The infiltration of the water to a great depth was relatively slow due to the low altitude of the reservoir. The neutral pressure took 16 :,'ears to reach values high enough to unbalance fractures situated one or more kilometres distant from the reservoir's axis. The fault occurred on nearly vertical, pre-existent fractures, releasing tectonic forces which probably originated from compression generated by the Meso-Atlantic range. It is clear that, according to this model, the means of filling the reservoir and the later control of the water level would not have altered the subsequent development of seismic activity. Similarly the dewatering of the reservoir would not immediately decrease seismic activity. Studies in Paraibuna-Paraitinga The Paraibuna Paraitinga reservoir which belongs to CESP is located in the NE of the state of S~o Paulo, 20 km from S~o Jos~ dos Campos. The maximum depth of the reservoir is 80 metres. The filling began in the middle of 1976 with the first tremors being felt in December of the same year. The seismic activity maintained a low level until November of 1977 when there was an earthquake of 3.4 Richter. After this, seismic activfty decreased with only some sporadic reactivations at a low level Unlike those observed in Carmo do Cajuru, these earthquakes had epicentres less than 1000 metres deep. As some of these epicentres axe located in the region of the Cuba~o-Paraibuna fault, initially these earthquakes were assumed to be related to movements along this fault. However, the focal mechanism of the seisms indicates that the fracturing is along NS or EW planes which do not coincide with the NE-SW direction of the Cubat~o-Paraibuna fault. As the predominant fracturing in the area has a N-S direction, we conclude that the seisms correspond to movements along this fracture as opposed to the Cubat~o-Para~una faults. The CubaNo-Paraibuna fault would not be associated with the occurrence of induced seisms. Consequently an accumulation of forces along this fault, due to earthquakes, is not to be expected. Recommendations Lxperience obtained since 1975 indicates that to perfect understanding of induced seismicity and its relation to the several parameters of a reservoir (depth, capacity, regional geology and tectonics) we should emphasize some aspects o f the studies. As to geological studies in the area" of reservoirs, we stress the importance of a detailed survey of the fractures and faults existing in the area, especially that which will be covered by water. The experience obtained at Carmo do Cajuru and Paraibuna-Paraitinga indicates that induced seisms correspond to movements along preexistent fractures observable at the surface. Another aspect of seismological studies which is important is the tectonics and geology of the region. These studies are essential for the interpretation of the state of forces in the area revealed by the focal mechanism of the shocks in terms of local tectonics. The determination of the direction o f tectonic forces by hydraulic fracturing would yield valuable information in the study of the nature of induced seismicity. A very important parameter in any

73

model of induced seismicity is the permeability of the medium. Although difficult to measure, permeability up to the greatest depth possible should be ascertained.

Richter determined that the magnitude would be the decimal logarithm of the maximum amplitude, registered in microns on a standardized seismograph, 100 km from its epicentre.

Laboratory studies on the variation of rock permeability as a function of pressure should be investigated.

This means that when the magnitude rises by one unit, the maximum amplitude increases tenfold.

Finally, another type of investigation which would complete seismological observations, would be on crust deformation due to the f'flling of reservoirs.

Energy l~erated by seisms

M a t h e m a t i c a l t r e a t m e n t of seismological data Eng. Dr. Carlos A. J. V. Dinis da G a m a S~o Paulo

IPT

Engineering criteria and seismological criteria For many years, engineering calculations for the design of structures located in seismic4egions were undertaken without consideration of the mechanisms involved in the origin of events. These calculations had considered only the application of vibrations of a certain amplitude and frequency which were correlated with the predicted seismic characteristics of the site. It can be observed that while seismology is more cor~cerned with the causes of this phenomenon, engineering is more interested in the effects. The present tendency is to integrate these two viewpoints. More recent models are supported by data which reflect variability and uncertainty about the mechanisms of earthquakes, of the mechanical properties of rocks and the surfaces of the faults that constitute these models. Obviously, stability is calculated in terms of probability, so that the probability of rupture of each design solution comes to be an incorporated risk in the project. As decision-making criteria are based on the search for solutions which make compatible the conflicting influences of cost and safety the use of reliability concepts permit the establishment of a compromise for each case. It can be concluded that probabilistic methods attempt to quantify the variability and uncertainty inherent in natural phenomena, incorporating them in projects in an economically acceptable way without sacrificing adequate security conditions. Deterministic methods are based on the choice of typical situations involving average parameters and properties of structures and terrains, aiming for a design which would resist a seism of a given intensity. The balance between cost and safety is generally achieved through various attempts, in each of which an earthquake of a given intensity would be considered, and the solution of the corresponding project is determined, later being quantified in terms of cost. The economic criterion is generally preponderant, both because of a desire for returns on investment or because of limitations on funds available for a project. Thus the security factor and the maximum seismic intensity which a structure can withstand are established within these limitations. Sometimes a certain seismic coefficient is selected "a priori" (either because it is common or recommended in the local construction codes) limiting the decision-making process to the choice of the best alternative between safety and economy. Deterministic methods for the treatment of seismological data Seismological data The principal source of seismological data is the seismogram. This consists of the record, as a function of time, of the movements of a determined spot where a sensor or geophone has previously been installed. Such a record is an accelerogram. It can be amplified on any time scale, in order to clarify the amplitude and frequency of movemerits, as well as the arrival times of different kinds of waves which pass through the earth in consequence of a seism. The interpretation of seismograms is aimed at determining the intensity of the seism at the local where it was registered (which also determines the magnitude of its focus) as well as the distance and direction of the epicentre in relation to the sensor. This determination is arrived at from the difference in the arrival times of P and S waves. Determination of the magnitude of seisms The magnitude of a seism is a measure of the energy it liberates.

As a result of the sudden liberation of the deformation energy stored in rocks, vibrations are transmitted in all directions. When these are recorded on appropriate equipment, the total quantity of energy dissipated can be calculated. There are various methods of calculating this energy. One very simple method re0uires the i~reliminary calculation of the local magnitude of the earthquake to determine the seismic energy. The calculation of the seismic energy can also be done by the direct method. This is done from the shape of the waves registered and utilizes the expressions of the elasticity theory. Intensity of a seism It is obvious that the destructive effect of a seism, the consequence of the seismic energy liberated, decreases at a greater distance from its focus. At each point of a terrain, the effects of a seism are different, according to the locality and the magnitude of the seism. Normally the intensity of a seisan is defined in terms of the damage caused in civil works, the volume of disturbances to the Earth's surface and the quantity and quality of reactions in living beings. Pseudo-static design methods As a result of the deterministic treatment of seismological data there are methods of structural calculation based on the consideration of the effects of seisms by way of the simple application of horizontal or inclined forces on structures. Some countries have adopted a seismic coefficient as a project criterion in their building codes. In many other countries the coefficient 0.1 is used. However, the choice of this parameter should depend on several factors, such as, for example, the region in question, the type of construction, its function and the importance of the losses its collapse would provoke. Probability methods for the treatment of seismological data Every year more than 300,000 seisms occur in the world, enough to enable a statistical trealynent of seismological data. As far as engineering application is concerned, it is important to determine the frequency of seisms of a given intensity in an area, and the number of seismic events of each intensity. Time distribution of seisms Using the methodology of series of events, the distribution of frequencies which represent the time interval between two seisms is defined using a time sequence of the occurrence of seisms (independent of the respective magnitude). A specific trend is sought. Various frequency distributions can be tested with the applicability of each one being researched statistically. This process should include four phases: (a) A trend between the time variable and the number of events is sought. If this is found, regression techniques are used to arrive at the desired forecast. (b) If there is no trend in the data, one looks for auto-correlation among them, thus making the forecast from immediately foregoing events. (c) If this auto-correlation is not demonstrated, the application of frequency distribution (Poisson) is attempted, using the Kolmogoroy - Smirnov test. (d) If the Poisson distribution cannot be applied, an empirical forecast model should be developed, generally using graphical representation of the variables involved. Probability methods of dynamic analysis The probability approach is justiffmble in modern engineering in virtue of the lack of knowledge about the real mechanical properties of materials used in construction. There is a ~nilar ignorance of the magnitude of the force (static and dynamic) to which these struotures will be subjected.

74

If we add to these facts the doubt resulting from choice of a formula which does not fit exactly the physical phenomenon being analysed. it can be understood that probability methodology is more realistic than its determinist counterpart. The stability of structures comes to be characterized by a probability of rupture instead of a security factor. This probability can be computed in 2 different ways: (a) The hybrid method: various safety factors are calculated by the deterministic process, and the probability of rupture is considered to be the relative frequency of eases where the safety factor is less than unity. (b) The actual probabilistic method. With this method the distribution of characteristic probabilities of the resistance of structural materials and the tensions induced in these elements are determined, so that the probability of rupture corresponds to the area of intersection of the two distributions. Conclusions The treatment of seismological data, with a view to their use in engineering projects, should be approched bearing in mind the following aspects: (a) To determine the expected intensity of earthquakes by either deterministic or probabilistic methods. (b)To evaluate the possibility of collapse of the structures for different levels of importance of seimas, (c) To analyse the costs and damage to be expected as a consequence of the earthquakes, (d)To estimate the cost of palliative actions to prevent possible future Financial outlays. S e i s m i c c o n s i d e r a t i o n s in p r o j e c t s for e a r t h a n d r o c k f i l l dams L u i z A. V a l e n z u e l a P a l o m o Themag Engenharia Ltda Introduction In the present decade a rising interest in seismic risk can be noticed among specialists in dams. This tendency has spread to some countries traditionally considered as being non-seismic. This has been due to the occurrence of seisms induced by the filling of reservoirs. Although these cases have been of relatively low intensity, the cases of Koyna, India (Gupta and Rastogi 1976), Kremasta in Greece (Gupta and Rastogi 1976) and Hsinfengkiang (Hsu Tsung-Ho et at 1976) had magnitudes respectively of 6.5, 6.3 and 6.2. This preoccupation has also spread to Brazil due to the large number of dams being projected or already under construction, and also because of the occurrence of induced seisrns in at least 4 cases: Capirata, Porto Colombia, Volta Grande and Para~una-Paraitinga. Considerir,g that methods of seismic analysis of dams are still under research, it is understandable that this aspect of the project is subject to misunderstanding and some confusion. Frequently this gives rise to very different attitudes in those responsible for the projects, varying from greatly exaggerated precautions to the total disregard of seismic considerations. The behaviour of earth/mekfill dams in relation to seisms Survey of the real seimaic behaviour of earth dams The works of Ambraseys (1960) and more recently Seed (Seed et aL 1978) have complete surveys of dams all over the world. Rezendiz et al. (1972) compiled various studies including that of Ambraseys, arriving at the establishment of a relative frequency relationship of the occurrence of various kinds of rupture and damage in dams, caused by seismic action. It was possible to conclude that, although the cases of sliding or distortion through mass shearing of the dam walland/or of the foundations are mote frequent (60 per cent), there are other types of rupture which could be induced by seisms, and these should be considered in the analysis and seismic project of these works. Seed et al. (1978) presented a survey and analysis of the seismic behaviour of various earth dams. From the analysis of this survey, it

could be concluded that the occurrence of ruptures or great damage in earth/rockfill darns is related to vulnerable foundation conditions (for example: soft, sandy soils subject to liquefaction) or to large structures constructed by methods that today could be considered inadequate. Wail-built claystone dams with claystone or rock foundations have resisted strong seisms with accelerations varying from 0.35 to 0.80 g without apparent damage. Although some hydraulic earth dams have ruptured during strong seisms, there are numerous dams of this type that have satisfactorily resisted seisms of moderate intensity, that is, with accelerations of around 0.20 g resulting from seisms o f 6.5 to 7 magnitude. These dams were constructed by m e t h o d s that today would be considered inadequate. Partial conclusions

(a) Apart from the seismic stability o f slopes and the foundation of the wall of the dam, the seismic security analysis should also involve settlement, fault dislocation, rupture o f reservoir slopes; (b) practically any well-constructed dam on good foundations can resist moderate intensity seisms (with accelerations up to 0.20 g or more); (c) some darns with claystone or rock foundations and walls have resisted seisms with accelerations varying from 0.35 to 0.80 g for magnitudes of up to 8.25 without apparent damage; (d) hydraulic earth dams with adequate construction and controls have resisted moderate intensity seisrns ( m a x i m u m accelerations up to 0.20 g); (e) although there is evidence on the apparent seismic resistance of earth/rockfill dams, it should be noted that the majority of dams, principally those of great height and with large reservoirs, constructed in the last few years, have not been submitted to strong seisms, Thus the complete information on the real behaviour of these structures in relation to seismic activity is not available. Effects of foundations The phenomenon of vibrations in a structure due to seismic activity is a problem of interaction, and the seismic response of the structure will depend on the relative dynamic characteristics o f each of the elements in play. The application of the finite elements method has been the most powerful weapon in the analysis of the interaction foundation - structure - seism. This m e t h o d allows the consideration of more realistic cases, for example those with various layers of material in the foundation, and with different properties, such as more complex geometries. Another important aspect in seismic areas, and which influences the seismic response of structures and their foundations is related to the presence of Free sands and loose, sandy silts, materials which are subject to liquefaction. Partial conclusions A rapid analysis of the aspects m e n t i o n e d concerning the results of analytical studies of the seismic behaviour o f dams leads to the following conclusions: (a) analytical works clearly demonstrate that the seismic response of dams (or other structures) is a dynamic problem involving the interaction: structure - foundation - seism; (b) generally speaking, the accelerations acting on the base of an earth or rockf~ dam will be amplified, with the m a x i m u m accelerations being on the top. The accelerations at different levels will vary in magnitude and direction several times during the seism. The m a x i m u m values of the accelerations will be maintained only for a fraction of a second; (c) although they axe not completely satisfactory, there are approximate analytical methods which permit the estimation of the seismic response of a dam.

Seismic analysis of dam slopes Pseudo-static methods Although theoretical studies on the seismic behaviour of dams began in 1936, up to 1965 the only m e t h o d used to evaluate the seismic

75

stability of earth or rockfill slopes was that called "pseudo-static". In this method a force of inertia, generally horizontal, is included in a conventional stability calculation following one of the limit equilibrium methods. The safety factor is calculated conventionally, with a minimum value of 1.1 to 1.3 being accepted. The choice of the appropriate value of tile seismic coefficient k, which defines this force of inertia is the first problem in this method. In reality there is no value which can adequately represent the interaction between the structure (dam) and the activity (seism). In general, arbitrary values axe utilized, varying typically between 0.05 and 0.25 depending on the inferred seismicity of the region. The most common value used for many years was 0.I0. This was one of these "magical" engineering numbers (such as 1.5 as the .safety factor for slopes in the static case). Terzaghi (1950) commenting on the seismic coefficient mentioned values of k = 0.10 for severe seisms (IX on the Rossi - Forelli scale) k = 0.25 for destructive seisms (X on the same scale) and k = 0.50 for catastrophes. Attempts have been made to justify this method, supposing a rigid response of the structure. In this case the acceleration which would act on the structure would be equal to the acceleration of the terrain and consequently a value ofk which represents the maximum acceleration of the terrain should be adopted. This criterion was not accepted, because, when used in the conventional pseudo-static method, it gave very high inertia values.

Dynamic method This method and its variants were developed by the Berkley ~ o u p (University of California) especially Seed, Lee and ldriss. The "dynamic" method (Seed 1966, 1967, 1973 and 1979) was developed as alternative to evaluate deformation induced in dams by seisms. It is currently one of the most widely used methods. It is based on the utilization of the technique of finite elements and simulation in the laboratory of the history of tensions in elements and representative samples. The basic principles of this method axe: determination of tensions before the seisms; determination of the tensions induced by the selected seismic excitation, considering the non-linear dynamic properties of the mils; - tests of various representative samples of the dam which are then submitted to the combined effects of the initial static tensions and induced dynamic tensions. Their effects are determined in terms of the development of neutral pressures and unit deformations ; -

evaluation of the safety factor during and after the seism;

-

if the dam is safe, the global deformations of the dam are estimated utilizing the unit deformations obtained in representative elements.

Dynamic behaviour of soils

Methods based on the estimation of deformation

It is possible to differentiate between cases related to dams of claystone, dense sands or moderately loose sands which are not saturated and fine sands and silty, loose, saturated sands. In fact, the difference in behaviour of the last types of soil, is limited to the ~eater facility some .sandy soils have in developing large neutral pressures when involved in shearing especially in the case of seismic excitation. The Seed method manages to include this phenomenon in the seismic analysis "reproducing" in the laboratory the behaviour of representative elements of the dam and/or foundation. However, the laboratory procedures have been criticized recently (Castro 1969; Castro 1975; Casagrande 1975). It should be understood that these phenomena (liquefaction and cyclical mobility) are not completely understood at present, and the analytic methods available axe far from being flawless. They should be used cautiously.

Considering the aforementioned facts, Newmark (1965) suggested the adoption of criteria of deformation and proposed the utilization of a rigid block model (potential sliding mass) over a plane (rupture surface) with a given movement (seism), in order to estimate the deformation induced by a setsm. The movements and the beginning of the rupture of the slope would start when the inertia forces (accelerations) developed in a potential fault surface, were greater than a critical resistance (or flow).

These factors are particularly important for hydraulic earth dams and waste dams composed of materials which are more susceptible to liquefaction. The current practice seems to indicate the convenience of compacting the granular berms of the dams and compacting or removing very loose materials from the foundations in areas where they are significant. In the U.S.S.R., however, the general practice is to construct hydraulic earth dams without compacting the berms, even in seismic areas (Hydroprojekt 1973)

The usual project criterion based on obtaining a safety factor greater than 1.0 in a conventional stability calculation is not realistic. This is because dynamic behaviour studies on dams demonstrate that this factor can often fall below the value of 1.0 during a seism, leading to permanent deformations which, however, could be perfectly acceptable. Terzaghi (1950) had already called attention to the unreliability of this kind of analysis. Stable slopes could be obtained with a seismic security value of less than 1.0, and ruptures with values greater than 1.0 depending on the characteristics of the materials composing tire slope.

Later, the progress reached in methods of estimating induced seismic accelerations at different levels of a dam (Seed and Martin 1966; Ambraseys and Sarma 1967) comlSleted the ideas begun by Newmark, leading to the development of similar methods, but with wider possibilities of application. The Imperial College (London) group developed the basic ideas of Newmark, using a rigid block model sliding on an inclined plane, in order to estimate deformations. This model incorporated the predominant period of the seism. For the calculation of critical acceleration kc, methods of limit equilibrium in terms of effective tensions, are utilized. The characteristics of dynamic resistance of the materials are used through parameters of neutral pressure A and B, obtained from tests with cyclical loads. Although the Berkley, California group had a different approach, Makdisi and Seed (1978) published a study proposing a simplified method for the estimation of dam deformations. This method is similar to the previously mentioned ones. However, the deformations are calculated by double integration without using the block model. These authors also call attention to the fact that this method has only been applied to dams up to 60 metres high. Seed (1979) commenting on these simplified methods, concluded that the use of pseudo-static methods with a minimum SF of 1.15 and seismic coefficients k = 0.10 (for seisms of magnitude 6.5) and k = 0.10 (for seisms of magnitude 8.25) would guarantee the tolerable deformations (of up to 1 metre) for high, induced accelerations in the dam, as long as it is made of materials which have little or no loss of resistance due to deformations induced by the seism.

Partial conclusions (a) The pseudo-static method is arbitrary and does not take into consideration the principal characteristics of the real seismic behaviour of earth dams. The application of this method to some dams which burst proved to be inadequate; (b)the methods based on evaluation of deformation (Newmark 1965; Sarma 1975; Makdisi and Seed 1978) are relatively simple and have given satisfactory results especially for dams made from dense sands or claystone; (c) the dynamic method developed by Seed and his collaborators, although relatively complex, seems to have been successful in the analysis of numerous dams, principally those made of materials susceptible to liquefaction. However, some doubts persist as to wether the laboratory tests, on which the method is based, are representative.

Other aspects of the seismic project for earth and rockfill dams Eventual fault dislocation in foundations Although this type of incident is relatively uncommon, lately it has been receiving a lot of attention from specialists, due to some "near" ruptures caused by fault dislocation. One example of this is the Hebgen darn in the U.S.A. The 1959 West Yellowstone seism (M = 7.1)was accompanied by a vertical dislocation of 6 metres in a normal fault transverse to the dam axis. A series of other fortunate events (for example an average

76

subsidence of 3 metres in the reservoir, accompanied by the same subsidence in the dam) prevented the rupture of the dam which needed only minor repairs (Sherard et al. 1974).

fill dams, especially those built with compacted clay materials and/ or rockfill and having good foundations, have a significant, inherent resistance.

Another case was the gravity dam at Lower Crystal Springs in California. There was a dislocation of the San Andreas fault 200 metres above the dam, associated with the 1906 San Francisco seism (Sherard et al. 1974).

(c) Many hydraulic earth dams have resisted seisms with maximum accelerations of up to 0.20 g. Project details in accordance with modern construction techniques (use of predominantly sandy materials, berm compaction, etc) could significantly increase this resistance.

Sherard et aL (1974), in their study on potentially active faults in dam foundation~ defined an active (or potentially active) fault as "a fault for which there is sufficient evidence of dislocation in the recent geological past to anticipate future dislocations during the working life of a dam (approximately 100 years)". By the definition adopted by the U. S. Atomic Energy Comission (1973), in studies for the location of nuclear reactors, an active fault should have one of the following characteristics: (a) dislocations on or near the surface at least once in 35,000 years or recurring movements in the last 500,000 years; (b) macro-seismicity measured instrumentally with enough precision to demonstrate a direct relation with the fault; (c) a structural relation with an active fault having the characteristics of (a) or (b) so that there would be a reasonable expectation that a dislocation in one would be accompanied by a dislocation in the other. Loading berm An ample loading berm is fundamental in the seismic design of a dam. It should be sufficient to compensate eventual settlements, slides, dislocations of the foundation and seism-induced waves in the reservoir. Transition zones and f'dter

(d) Besides the stability of the slopes and the foundation of the wall of the dam, other factors should be considered in the seismic analysis of these structures (such as settlement, eventual dislocation of faults in the foundation, rupture of slopes in the reservoir, seismicinduced waves, etc). (e) The seismic response of a dam is a dynamic problem of interaction of structure/foundation/seism, and is greatly influenced by the geometric and dynamic characteristics of these factors. (f) The conventional pseudo-static m e t h o d s of analysis of slope stability are arbitrary and do not consider the fundamental characteristics of the seismic behaviour of dams. There are, at present, methods based on the evaluation of deformations induced by the seism, which seem to give a more satisfactory solution to these problems. Simplified methods such as those of Sarma, Makdisi and Seed are particularly suitable for soils which do not develop large pore pressures as a result of seismic activity. For soils susceptible to liquefaction, Seed's "dynamic" m e t h o d seems to be the solution, although doubts remain about the laboratory tests on which the method is based. (g) The selection of the seismic m o v e m e n t s to be applied to a project should be done through the application of probabilistic methods and project optimization criteria even though there are difficulties resulting from the absence of sufficient, adequate data.

Current practice considers this point as to be the most important line of defence. Efforts should be made to provide sufficiently wide transition zones, made of adequate materials (essentially non-cohesive) and less emphasis should be given to special characteristics of the core-material.

Seismological evaluation for engineering projects

Diverse aspects of the seismic project

Introduction

Details of the top (project with conservation zoning against the passage of water); details of the upper part of the dam (using materials more resistant to the formation of cracks which can occur in this area of low confinement tensions); details of the work program, avoiding, when poss~le, interaction with auxiliary concrete works, such as conduits etc.

A multidisciplinaty team is necessary for the evaluation of potential seismic damage and the establishment of a land utilization plan in order to reduce risks in seismically active areas. This team is composed of specialists in geology, engineering geology, soil mechanics, seismology, structural engineering and linduse planners. One example of a detailed seismological study was that done for the Auburn dam in California, involving 50 technicians.

Selection

of the characteristics

of seismic

agitation

The classic studies of Newmark (1965) and Seed (1966) incorporated deterministic criteria in as m u c h as that they suggested the adoption o f a seism or a family o f seisms (described by certain parameters such as acceleration, velocity, dislocation, duration, ere) to be applied in the locality of the dam. These would be chosen according to subjective criteria, for example considering the largest probable seism in the area. A similar approach was suggested by Valenzueli (1978) in commenting on probable Brazilian criteria in seismic projects.

Geol. H a m i l t o n A. C o s t a H i d r o s e r v i c e E n g e n h a r i a de P r o j e t o s L t d a .

The seimaological evaluation should include studies to answer 6 basic questions: 1) Is there a seismic risk? 2) Where will it occur? 3) How big will it be? 4) What is the frequency of the event? 5) What will be the consequence? 6) How are the effects to be minimized? Stages of a seismological study and seismic risks in engineering projects 1) Geological reconnaissance and general tectonics -

However, the convenience of adopting statistical and probabilistic criteria is increasingly emphasized in the selection of seismic movements. This is similar to the selection of floods in Hydro-electric projects. Lomnitz (1974), Rezendiz (1975) and Esteva (1976) among others developed these ideas in great detail. The objective of these probabalistic techniques is to determine the probability of the occurrence of a seism of a determined magnitude in a determined number of years. This magnitude, in turn, should be related to probable values of engineering parameters (generally m a x i m u m accelerations and velocities and predominant periods). Final conclusions (a) The majority of earth and rockfill dams, principally the high ones built in recent years still have not been submRted to strong seisms. Therefore, we have little experience of the real seismic behaviour of these structures (b) From available evidence it can be concluded that earth and rock-

-

geological photo-interpretation (lineations and faults) aerial and ground reconnaissance elaboration o f a general seismological map (historical, instxumental)

2) Quaternary geology and geochronology -

mapping and sampling o f Quaternary formations opening of trenches Criteria of active faults

3) Microseismic observations -

localization of seismological stations network operation analysis and interpretation of results localization of epicentres with magnitude values

4) Selection of project earthqu',tkes -

elaboration of the tectonic m o d e l localization, magnitude and repetition frequency of project earthquakes

77

5) Selection of the seismological parameters of the project -

maximum acceleration, velocity, dislocation, duration of movement and response spectra

6) Preparation of final report Geological reconnaissance and general tectonics Geological data allied with Quaternary geology today seem to be one of the best methods of evaluating the seismicity of an area. Considering that earthquake data for periods of 2,000 to 3,000 years are available in very few places in the world (Japan, China) any extrapolation of seismological data to. other parts of the world should be made very cautiously. With this, a thorough geological-tectonic study, in which all the lineations and faults are localized and classified, is of great importance in any projected seismological study. Simultaneously all seismological registers should be analysed, thus instrumental registers should be combined with historical data, making for a statistical study of the frequency and magnitude of earthquakes, the depth of hypocentres and the geographical location of the epicentres. Quaternary geology and geochronology The historical geology of the fault in the late Quaternary period is one of the most promising sources tbr statistical studies of the frequency and location of earthquakes.

location in the last 35,000 years or more than one in the last 500,000 years. Similar definitions are found in the International Atomic Energy Agency (1972) and Krinitzsky (1974). An inactive or dead fault is a fault which was active during the orogenic period but is no longer active in the present seismotectoni~ regime and consequently neither affects the last Holocene deposits, nor demonstrates associated seismic activity. Microseismic observations The study of small earthquakes is today an important element in the investigation of seismic risks in engineering projects. The observation of the specific locality as well as the region in general, using sensitive instruments, provides very useful information. These small tremors, called micro-earthquakes are defined as those with a local magnitude of M less than 3 on the Richter scale. The instruments now being used for studying micro-earthquakes are capable of registering earth movements amplified more than one million times, in a frequency band from 5 to 25 Hz. Recent laboratory studies have demonstrated that the physical mechanism of micro-earthquakes is essentially the same as that of larger ones in the same way that a larger fracture (strong earthquake) is accompanied by smaller earthquakes (micro-earthquakes). At present the instruments for microseismic observations are portable and precise (high gain and high frequency). They can be operated by radiotelemetry or telephone directly to a central locahty where the registers are recorded. Data analysis is rapid and routine, using high speed digital computers. The programs provide characteristics such as location, frequency, spectrum, speed of seismic wave, magnitude.

Seismic-tectonic relationships in California, where the relationship of active faults to earthquakes is one to one, are apphcable to other parts of the world. For example, in Turkey all the principal seismic areas can be easily recognised, even without seismological registers, by field studies of the Quaternary faults (Allen 1975). In China the association of big earthquakes with active faults is probably one of the best examples of how Quaternary history can identify seismicity without relying on seismological instrumentation.

Selection of project earthquakes

On the otherhand, in Japan the largest earthquakes do not demonstrate such a direct relationship with faults considered to be active. However, some of the larger faults such as the Median Tectonic Line and the Itoigawa-Shizuoka demonstrate considerable evidence of Holocene dislocations and therefore, should be considered sources of large, future events.

The tectonic model determines the patterns followed by the seismic waves from the hypocentre to the seismographic station and is the basis for the calculation of the arrival times of the seismic waves in each station.

The detailed geology of Quaternary formations, their geomorphological aspects and the opening of trenches is an important detail in the realization of a seismological study. In this stage of the studies, criteria for the classification of faults referring to their activity should be established. At present there is not a universally accepted definition of an active or non-active fault. 9Definitions vary from those based on a low degree of activity or recurrence, to definitions which are limited to faults with historical dislocations. The original definition of Willis and Wood (1923) includes four essential elements: 1) Active faults have been dislocated in the present seismo-tectonic regime. 2) Active faults have the potential or probability of dislocating in the future.

To be able to borated. This fication of the magnitudcs to

do this, first a regional tectonic model must be elatectonic model will serve as a basis for the identimore probable sources of earthquakes, the values of be expected and the respective period of recurrence.

The models are better when based on the results of large-scale seismic refraction and reflection investigations. However, they can be based on the regional geology as the velocity characteristics of different rocks can be estimated with reasonable accuracy. Selection of seismological parameters of the project Using the previously defined projected earthquakes as a basis, it is necessary to estimate the characteristics of the seismic movements produced in the work site by these earthquakes. These characteristics include the values of maximum accelerations, velocities and dislocations, duration of the movement and spectral response. These parameters will naturally be used in all the stability calculations of the works incorporated in a project. Also it will be useful as orientation for dynamic laboratory tests, especially for trlaxial dynamic tests on sand and/or silts. These tests are aimed at discovering the liquefaction potential under a cyclical load.

3) Active faults demonstrate evidence of recent activity by physiographic features. 4) Active faults can be associated with earthquakes. Nowadays most professionals prefer to use geological evidence, rather than historical registers, to check long term activity on a fault. An active fault is a fault which has experienced a relative dislocation during the present seisrno-tectonic regime and which has a large possibility of doing so again in the future. For the Bureau of Reclamation, a fault which has moved in the last 100,000 years is active; for the Corps of Engineers this period is 35,000 years; for the State of California it is 10,000 years and for the U.S. Nuclear Regulatory Commission it is 500,000 years. The U.S. Nuclear Regulatory Commission (1975) coined the expression "capable fault" for the studies of localities for atomic reactors- These would be faults which have presented only one dis-

References ALLEN C.R. (1975) city. Geol. Aug. 1975.

: Geological Criteria for Evaluating SeismiSoc. of America, Bulletin, 86, 1041,

ALLEN C.R. (1976) : Geological Criteria for Evaluating Seismicity. Vide Lomnitz and Rosenblueth (1976). AMBRASEYS N.N. (1960a) : The Seismic Stabihty of Earth Dams. Proc. 2nd Conf. Earthquake Engng., Tokyo and Kyoto. AMBRASEYS N.N. (1960b) : A Note on the Effect of Surface Loading on the Shear Response of Overburdens. J1 Geophysical Res.

78

AMBRASEYS NaN.(1960c) : On the Shear ResponseofTwo-D> mensional Truncated Wedge Subjected to an Arbitrary' Disturhance. Bull. Seis. Soc. Am., 50. AMBRASEYS N,N.(1972) : Discussion on Lane's paper:Seismic activity at manmade reservoir. Lane, 1971. Proc. Inst. Civ. Engrs., 5 1 , 7 6 7 - 7 6 9 . AMBRASEYS N.N.(19731 : Dynamics and Response of t:oundation Materials in Epicentral Regions of Strong Earthquakes. Proc. 5th World Conf. Earthquake Engng., Rome.

COOK N.G3V. (1975) : Seismicity associated with mining. 1st Int. Syrup. Induced Seismicity, Alherta, Canada. DALY W . - J U D D W . - M E A D R . ( 1 9 7 7 ) : Evaluation of Seismicity at 1;.S. Reservoirs. USCOLD Committee on Earthquakes. Rep., 28 pp. DAVINO A. (1972) : Ausculta~;~o Sismica de Reservatorios barragens sismog~nicas. ' VIII SeminSrio Nacional de Grandes Barragens, S~o Paulo, Brazil.

AMBRASEYS N.N. - SARMA S.K. (1967) : The Response of Earth Dams to Strong Earthquakes. Geotechnique, 17.

DONOVAN N.C. - BORNSTEIN A.E.(1978) : Uncertainties in Seismic Risk Procedures. Jl Geot. Engng Div. ASCE 104, GT7.

ATHAVALER.N.(19751: Induced Seepage along a coastal parallel System as possible cause of the Koyna earthquakes. Bull. Seismol. Soc. Am., 65, 183 -191.

ESTEVA L.(1976) : Seismicity. In: Seismic Risk and Engineering Decisions, C. Lomnitz and E. Rosenblueth (eds). Developments in Geotechnical Engineering, 15.

ATTEWELL P.B. - FARMt'R I.W. (1976) : Principles of Engineering Geology. Chapman and flail, Ltd., London.

FRANC~A P.O. (1978) : Seismic Activity Induced by Reservoir Impounding. Ph. D. Thesis, Imperial College, London.

BATH M. (1970) : Artificial release of earthquakes. Scientia, Vol. CV NDCCIII--DCCIV. 706 -724.

GAMA C.D. et al. (1979) : Sismologia de struFao Pesada, Jan. 1979, Silo Paulo.

BE'CK J i . (1976) : Weight induced stresses and the recent seismicity at lake Oroville, California. Bull. Seismol. Soc. Am., 66, 1121 -1131.

GOODMAN RE. (1973) : Pore pressure in intact and jointed rocks induced by deformation. Int. Coll. on Seism. Eft. of Res. Imp. (UNESCO), London, 1 p.

BEItRMAN D. I1976) : A chcia previu a hora e o lugar de um terremoto e salvou milhares de vidas. UNESCO, O Correio, 4,7. BERROCAL J. (1973) : Seismic activity in some areas of Brazil. Int. Coll. on Seism. Eft. of Res. Imp. (UNESCO), London, 1 p. BOLT B.A. (1978) : Earthquakes. WM. Freeman and Company. San Francisco. BON1LLA bl.G. 1,1970) : Surface Faulting and Related Effects. Chapter 3 in: Earthquake Engineering (R.L. Wiegel, ed.). Prentice-HaU. BOORE D.M. (1977) : The Motion of the Ground in Earthquakes. Scientific American, Dec. 1977. BOZOVIC A. (1973) : Review and appraisal of case histories related to reservoir induced seismicity. Int. Coll. on Seism. Eft. of Res. Imp. (UNESCO), London, 3 pp.

Barragens.

Con-

GOUGH D J . - G O U G K W . I . ( 1 9 7 0 - 1 ) : Stress and deflection in the lithosphere near lake Kariba - I. Geophys. J.R. Astr. Soc., 2 1 , 6 5 - 7 8 . (iUPTA H.K. (1973) : Investigations of the behaviour of reservoir associated earthquakes undertaken at the National Geophysical Research Institute, India. Int. Coll. on Seism. Eft. of Res. Imp. (UNESCO), London, 2 pp. GUPTA H.K. - RASTOGI B.K. - NARAIN ft. (1972) : Some discriminatory characteristics of earthquakes near the Kariba, Kremasta and Koyna artificial lakes. Bull. Seismol. Soc. Am., 6 2 , 4 9 3 - 5 0 7 . GUPTA H.K. - RASTOGI B.K. - NARAIN H. (1972) : Some Features of the Reservoir-Associated Seismic Activities. Bull. Seismol. Soc. Am., 6 2 , 4 8 1 - 4 9 2 . GUPTA H.K. - RASTOGI B.K. (1976) : Dams and Earthquakes. Elsevier Scient. Publ. Co., New York.

BRITO S. (1974) : Seismic activity around the Cajuru reservoir. Proc. II Int. Cong. Int. Assoc. Engng. Geol., Sgo Paulo, Brazil, II, 5.1 -5.9.

HANDIN J. - RALEIGH C.B. (1972) : Man-made earthquakes and earthquake control. Proc. Syrup. Percolation throu~h Fissured Rocks, Deutsche Gesellschaft fiir Erd- und Grundban, Stuttgart, F.R. Germany, T. 2D, 1 - 1 0 .

CADMAN J.D. (1978) : Sismicidade induzida e a seguranqa de grandes, barragens. XlI Semin~io National de Grandes Barragens, S'ao Paulo, Brazil, 393-403.

HANDIN J. - NELSON R,~.. (1973) : Why is lake Powell aseismic ? Int. Coll. on Seism. Eft. of Res. Imp. (UNESCO), London, 2 pp.

CASAGRANDE A. (1975) : Liquefaction and Cyclic Deformation of Sands - A Critical Review. Harvard Soil Mechan~s Series No. 88, Harvard University.

HARPSTER R.E. (1978) : Selected Clays Used as Core for a Rock.tiff Dam Designed to Cross a Potentially Active Fault. Clay Fills, Inst. Civ. Engrs., London.

CASTLE R . O . (1976) : Seismicity and faulting attributable to fluid extraction. Engng Geol., p. 151.

HATANAKA M. (1955) : Fundamental Considerations on the Earthquake Properties of the Earth Dams. Disaster Prevention Res. Inst., Kyoto University Bull. No. 11.

CASTRO G. (1969) : Liquefaction of Sands. Harvard Soil Mechanics Series No. 81, Harvard University. CASTRO G. (1975) : Liquefaction and Cyclic Mobility of Saturated Sands. Jl Geotechn. Engng Div. ASCE 101, GT6. CHOPRA A.K. (1967) : Earthquake Response of Earth Dams. J1 Soil Mech. Fdns. Div. ASCE 93, SM6.

HEALY J.H. - RUBEY W.W. - GRIGGS D.T. - RALEIGH C.B. (1968) : The Denver earthquakes. Science 161, 3; 848, 1301-1310. HOUSNER G.W. (1959) : Behaviour of Structures During Earthquakes. Jl Engng Mech. Div. ASCE Oct. 1959.

CLOUGH R . - CHOPRA A.K. (1966) : Earthquake Stress Analysis in Earth Dams. Jl Engng Mech. Div. ASCE 92, EM2.

HOUSNER G . W . - SCOTT R.F. (1979) : Earthquake considerations in Darn Design. Water Power and Dam Construction, July 1979.

CLUFF L. - SLEMMONS D. - WAGGONEER E.(1970) : Active Fault Zone Hazards and Related Problems of Siting Works of Man. Proc., 4th Int. Syrup. Earthquake Engng, I, Bulletin, Nov. 1970, India.

HOWELLS D.A. (1972) : Man-made earthquakes: The reservoir designer's problem. Proc. Syrup. Percolation through Fissured Rocks, Deutsche Gesellschaft fiir Erd- und Grundbau, Stuttgart, F.R. Germany, T. 2E, 1 - 5 .

CLUFF L. et al. (1972) : Site Evaluation in Seimnically Active Regions - An Interdisciplinary Team Approach. Proc. Int. Conf. Microzonation for Safer Construction, Research and Application, Seattle, Wash., Oct. 30, 1972.

HSU Tsung-Ho et at. (1976) : Strong Motion Observation of Water-Induced Earthquakes at Hsinfengkiang Reservoir in China. Engng Geol. 10, No. 2--4, special issue (W.G. Milne, ed.).

CLUFF L.S. - BROGAN G.E. (1974) : Investigation and Evaluation of Fault Activity in the USA. 2rid Int. Cong. Int. Assoc. Engng Geol., S~'o Paulo, Brazil.

HUBBERT M . K . - RUBEY W.W. (1959) : Role of fluid presssure in mechanics of faulting. Bull. Geol. Soc. Am., 70, 115-166.

79

HYDROPROJEKT(1973J : Contruq~'odeBarragensdeTerrapelo M6todo de Aterro Hidr~ulico. Relat6rio do Organismo Sovietico de Trabalos tlidrSulicos. Hydroprokekt Seminar, Coyne & Bellier, Moscow. IAG-Instituto Astron6mico e Geofisico de S~o Paulo (1979) : Atividade Sismica induzida por Reservat6rios ttidrel6tricos. (Unpublished). IDRISS I.M. - SEED tt.B. (1967) : Response of Earth Banks During Earthquakes. Jl Soil Mech. Fdns. Div. ASCE 93, SM3. IDRISS I . M . - S E E D t I . B . ( 1 9 7 4 ) : Effects of Local Geologic and Soil Conditions on Damage Potential During Earthquakes. 2nd Int. Cong. Int. Assoc. Engng Geol., S~'o Paulo, Brazil. IPT (1977) : Relat6rio No. 10.678 para Companhia Energdtica de S~o Paulo. Nov. 1977. JAEGER J.C. - COOK N.G. (1969) Mechanics. Methuen, London.

: Fundamentals of Rock

JANSEN RS. - DUCKLETH G.W. - BARRET K.G.(1976) : Problems of Hydraulic Fill Dams. XIIth Int. Cong. Laxge Dams, Mexico. JOINT

PANEL ON PROBLEMS CONCERNING SEISMOLOGY AND ROCK ~,IECtIANICS (1972) : Earthquakes related to reservoir filling. Div. of Earth Sciences, National Research Council, Washington, D.C., 24 pp.

KAMIL H. - BERTERO V.V. (1977) : A Nonlinear Seismic Design Procedure for Nuclear Facilities. Rep. No. UCB/ EERC-77/11, Univ. of California, Berkeley. KISSLINGER C. (1975) : Review of theories of mechanisms of induced seismicity. First Nat. Syrup. on Induced Seismicity, Alberta, Canada. KRINITZSKY E.L. (1974) : State of the Art for Assessing Earthquake ttazards in the United States; Fault assessment in earthquake engineering. Miscellaneous Paper, S - 7 3 - 1 , Rep. 2, U.S. Army Engng. Waterways Exp. Station, Miss. LANE R.G.T. (1971) : Seismic activity at man-made reservoir. Proc. Inst. Cir. Engrs., 50, 15-24. LANE R.G.T. (19731 : Investigations of Seismicity at Dam/Re~ervoix sites. Int. Coll. on Seism. Eft. on Res. Imp. (UNESco), London,1 p. LENSEN G.J. (1973) : A Model producing Seismic effects by reservoir impounding and a proposal for monitoring to predict man-made earthquakes. Int. Coll. on Seism. Eft. of Res. Imp. (UNESCO), London, 1 p. LEE K.L. - WALTERS tt.G. (1972) : Earthquake induced Cracking of Dry Canyon Dam. ASCE Annual and Nat. Envtl. Engng Meeting, Houston, USA. LOMNITZ C. (1974) : Global Tectonic and Earthquake Risk. Developments in Geotectonics 5. Elsevier Scientific Publishing Co. LOMNITZ C . - ROSENBLUETH E. (eds.) (1976) : Seismic Risk and Engineering Decisions. Elsevier Scientific Publishing Co. MAKDISI F . I . - SEED H.B. (1978) : Simplified Procedure for Estimating Dam and Embankment Earthquake-Induced Deformations. J1 Geol. Engng Div. ASCE 104, GT7. MONONOBE N. - TAKATA A. - MATSUMURA M. (1936) : Seismic Stability of the Earth Dams. lind Int. Cong. Large Dams, Washington, USA. NEWMARK N. (1965) : Effects of Earthquakes on Dams and Embankments. Geotechnique 15. (Previously presented at ASCE Conf., 1963). NEWMARK N.M. - ROSENBLUETH E. (1973) : Fundamentals of Earthquake Engineering. Prentice tlall Inc., Englewood Cliffs, N.J. NUR

A.

BOOKER J.R. (1972) : Aftershocks caused by pore fluid flow? Science, 175, 4024, 885-887. -

OBORN [..E.~1974) : Seismic Phenomena and Engineering Geology. 2nd Int. Cong. Int. Assoc. 1r Geol., S~o Paulo. Brazil. OKAMOTOS.(1973) : Introduction to Earthquake Engineering. John Wiley and Sons, New York. OWEISS I.S. - DOBRY R.(1978l : Establishment of Equivalent Linear Model and Site Period of a Soil Profile. 1st Caribbean Conf. on Earthquake Engng, Port of Spain, Trinidad. POLYAHOV S. (1974) : Design of Earthquake Resistant Structures. Mir Publishers, Moscow. RALEIGH C.B.(1972) : Mechanical behaviour of fissured rocks due to percolation: man-made earthquakes. Proe. Syrup. on Percolation through Fissured Rocks, Deutsche Gesellschaft ffir Erd- und Grundbau, Stuttgart, t:.R. Germany. RESENDIZ D. (1975) : Seismic Design of Earth and Rockfill Dams. General Report 5th Panam. Conf. Soil Mech. Fdn. Engng, IV, Buenos Aires. RESENDIZ D. - ROSENBLLETH E. - MENDOZAE.(.1972) : Diseffo Sismico de Presas de Tierra y Enrocamiento-Estado del Arte. Informe No. 300, Instituto de Ingenieria, M,~xico. ROTHE J.P. (1973) : Man-made lakes: Their problems and environmental effects - a geophysics report. In: Ackermann W.C., White G.F., Worthington t'.B.(eds.), Geophysical Monograph, 17, Amer. Geophys. Union, Washington. D.C., 441-454. SARMA S.K. (1973) : Stability Anlaysis of Embankments and Slopes. Geotechnique 23, 3. SARMA S.K. (1975) : Seismic Stability of Eaxth Darts and Embankments. Geotechnique 25, 4. SARMA S.K. (1978) : Response Characteristics and Stahility of Earth Dams During Strong Earthquakes. P h . D . Thesis, University of London. SAVAGE W. (1978) : La relacion entre Microsismos, Fallas Activas y la Occurencia de Temblores Mayores. Woodward-Clyde Consultants, San Francisco, Ca. SEED H.B. (1966) : A Method for Earthquake Resistance Design of Earth Dams. J1 Soil Mech. Fdns. Div. ASCE 92, SM1. SEED H.B. (1967) : Slope Stability During Earthquakes. Jl Soil Mech. Fdns. Div. ASCE 93, SM4. SEED H.B. (1973) : Stability of Earth and Roekfill Dams I)uring Earthquakes. In: Hirschfeld R.C., Poulos S.J. (eds.), Embankment Dam Engineering, Casagrande Volume: John Wiley and Sons. SEED lt.B. (1979) : Considerations in the Earthquake-Resistant Design of Earth and RockftU Dams. 19th Rankine Lecture, British Geotechn. Soc., Geotechnique, 29. SEED H.B. - MARTIN G.R. (1966) : The Seismic Coefficient on Earth Dam Design. Jl Soil Mech. Fdns. Div. ASCE 92, SM3. SEED

SEED

IDRISS I.M. (1969) : Influence of Soil Conditions on Ground Motions During Earthquakes. J1 Soil Mech. Fdns. Div. ASCE 95, SM1.

H.B.

LEE K.L. - IDRISS I.M. (1969) : Analysis of Sheffield Dam Failure. J1 Soft Mech. Fdns. Div. ASCE 95, SM6.

H.B.

SEED H.B. - WHITMAN R.V. - DEZFULIAN H. - DOBRY R. IDRISS I.M. (1972) : Relationships Between Soil Conditions and Building Damage in the 1967 Caracas Earthquake. J1 Soil Mech. Fdns. Div. ASCE 98, SM8. SEED H.B. - IDRISS I.M. - LEE K.L. - MAKDISI F.I. (1975) : Dynamic Analysis of the Slide in the Lower San Fernando Dam During the Earthquake of Februaxy 9, 1971. J1 Geot. Engng Div. ASCE 101, GT9. SEED H.B. - MAKDISI F j . - ALBA P. (1977) : The Performance of Earth Dams during Earthquakes. Rep. No. UCB/ EERC -77/20. University of California, Berkeley.

80

SERAFIN J.L. (1970) : Behaviour or rock masses: Structural foundations including deformation and seismic effects of reservoir loading. General Report Theme 8, 1st Cong. Int. Soc. Rock Mech., Belgrad. SHENG Chung-Kang et al. (1973) : Earthquakes induced by reservoir impounding and their effect on the Hsinfengkiang Dam. Scientia Sinica, 2, Peking, China, 44 pp. SHERARD J.L. - CLUFF L.S. - ALLEN C.R.(1974) : Potentially Active Faults in Dam Foundations. Geotechnique 24, 3. SIMON R.B.(1972) : Earthquake Interpretations (3rd ed.). Colorado School of Mines. SIMPSON D.W. (1976) : Seismicity changes associated with reservoir loading. In: Milne W.G. (ed.) Induced Seismicity, Engng Geol., 10 ( 2 - 4 ) , 123-150. SLEMMONS D . - McKINNEY R.(1977) : Definition of Active Fault. Miscellaneous Paper S - 7 7 - 8 , U.S. Army Engr Waterways Exp. Station, Miss., Soils and Pavements Laboratory, May 1977. SNOW D.T. (1972) : Geodynamics of Seismic Reservoirs. Proc. Int. Syrup. on Percolation through Fissured Rocks, Deutsche Gesellschaft f~ir Erd- und Grundbau, Stuttgart, F.R. Germany, T. 2T., 19 pp. SNOW D.T. (1973) : Reservoir-Induced Seismicity. Int. Coll. on Seism. Elf. of Res. Imp. (UNESCO), London, 2 pp.

u.S. ATOMIC f'NERGY COMMISSION (1973~ : Nuclear Pox~er Plants, Seismic and Geologic Siting Criteria. Federal Register, 38, 218. U.S. NUCLEAR RI'(;ULATORY COMMISSION (1975) : Reactor Site Criteria Title 10, Chapter l , C o d e of Federal Regulation - Energy. 1 1 0 - 1 - 1 0 0 - 6 , VALENZUEI, A L. (1976) : Um Novo M6todo para Anfilise de Estabilidade' de Taludes considerando Superficies de Ruptufa Quaisquer. XI Semin'5,rio de Grandes Barragens, Fortaleza. VALENZUELA L. (1978a) : Seismic problems in Brazil. 1st Caribbean Conf. on Earthquake Engng, Port of Spain, Trinidad, I 1 pp. VALENZUELA L. 11978b) : Seguran~a Sism~ca de barragens e a pr.~tica brasileira. XII Semin,~rio Nacional de Grandes Barragens, S~o Paulo, Brazil, III, 4 0 5 - 4 2 5 . WANG Miao-Yueh et al. (1975) : Mecanism of the reservoir impounding earthquakes at Hsinfengkiang and a preliminary endeavour to discuss their cause. Academia Sinica, Peking, China. WILLIS B. (1923) : A Fault Map of California. Seismot. Soc. America. Bulletin, 13, 1-12. WII_.LMORE P.L. (19737 : The scope and limitations of seismicity studies in the assessment of dam-sites hazards. Int. Coll. on Seism. Eft. of Res. Imp. (UNESCO), London.

SYKEES L.R. - FLETCHER J.P. - SBAR M.L. (1973) : Contemporary stresses, intraplate earthquakes and seismic risk associated with high pressure fluid injection wells in New York State. Int. Coll. on Seism. Eff. of Res. Imp. (UNESCO), London, I p.

WttlTMAN R.V. (1971) : Influence of Earthquakes on Stability Proc. 1st Int. Conf. on Stability in Open Pit Mining, Vancouver, Canada.

TERZAGHI K. (1950) : Mechanisms of Landslides. Geol. Soc. Amer. Engng Geol. Volume.

WOOD H.O. (1912) : The Earthquake Problem in the Western United States, Scismol. Soc. America Bulletin, 6, 181-217.