We use a population of 48669 double couple focal mechanisms of earthquakes, from 67717 events recorded in and around Iceland by the Icelandic Meteorological Office during the years 1995 to 1997. Magnitudes range between -1.8 and 4.8. The purpose of our study is to determine whether such a large mass of data has the potential to indicate the general tectonic field. Many reasons suggested that this may be not the case: uncertainties of determinations, perturbations in tectonic regimes, and so on. We considered two zones surrounding the major transform-rift zones north and south of Iceland: the Tjörnes quadrangle (66°-67°N, 16.5°-19.75°W) and the South Iceland Seismic Zone (SISZ) quadrangle (63.7°-64.25°N, 19.8°-21.1°W). Inversion was carried out based on a new direct method established by one of us (J.A.), using 10547 and 4413 double couple focal mechanisms in these two quadrangles respectively. In both cases, the reconstructed s2 axes plunge 78° or steeper, indicating dominating strike-slip mode. The average direction of extension (s3 trend) is N66°E for the Tjörnes quadrangle and N143°E for the SISZ one. Selecting only the data that fit significant quality requirements results in no or little change, with N66°E and N146°E trends respectively: 9831 and 1916 mechanisms are thus retained (respectively). This stability of the inversion indicates that the results are significant. The ratios F between principal stress differences average 0.6-0.7, indicating that in terms of magnitudes, s2 is closer to s1 than to s3. This is consistent with the association of normal and strike-slip faulting modes. Considering the general trend of plate separation and related extension in Iceland, that is, N104°E, these results in the main zones where transform faulting occurs north and south of the Icelandic rift (right-lateral and left-lateral respectively) are of particular interest. To the north, the trend of average extension is deviated counterclockwise of 38°. To the south, it is deviated clockwise of 42°. These deviations are in perfect agreement with the pattern of transform motions between the segments of the North-Atlantic oceanic spreading axis.
The South Iceland Seismic Zone is situated between two sections of the mid-atlantic ridge, i.e. the Reykjanes Ridge and the Eastern Volcanic Zone. It is no ideal transform fault, as it is not connecting both rifts at right angles and as the earthquakes do not occur on EW-trending left-lateral shear faults but on the conjugate, NS-oriented right-lateral, rupture planes. This is indicated by surface fault traces and aftershock distributions. The stress field permanently generated by rifting with slightly more than 2 cm per year around this transform fault-like zone is computed and superimposed by the stress field changes induced by a series of 11 earthquakes (M >= 6) between 1706 and 1912. Finally, the post-seismic stress field of 1912 is extrapolated to the present, to see where highest stresses might have accumulated. The stress is released by the series of events in the whole area, even though the ruptures planes are located on parallel NS-striking zones. The pre-seismic stress level for most events is high and pretty stable with the exception of situations when several strong shocks occur over a time span of several days, i.e. display typical main shock-aftershock patterns.
A GPS Network has been installed in North Iceland in order to measure surface displacements in the vicinity of the Husavik-Flatey Fault Zone (HFFZ). This fracture zone is a dextral transform fault which connects the rift of the Kobleinsey ridge with that of North Iceland. Geological studies indicate that the HFFZ is active since 7-9 Ma. This lineament is one of the most active in Northern Iceland. Permanent seismic activity has been recorded along the western part of the fault zone (Flatey segment). On the other hand the eastern part of the fault and its junction with the rift zone has no significant seismic activity.
The comparison between 1995 and 1997 shows horizontal and vertical displacements larger than the error estimated at a level of 95%. Far from the HFFZ the horizontal component of the displacement strikes parallel to the HFFZ and indicates a dextral movement of 1.7 cm/y. Near the transform fault displacements are deviated from the regional tendency and become progressively perpendicular to the HFFZ. We triangulated the GPS network and computed tensors of deformation. E-W extension has occurred in the rift zone while N-S compression has been observed on its western shoulder. Near the transform zone itself, tensors seem to have been deviated of 30° to the east in average.
Our measurements has pointed out that the Husavik fault is locked as suggested by the lack of seismicity. Mechanical modelling enable us to estimate the thickness of the elastic crust and a depth for the locked fault surface.
The Tjörnes Fracture Zone (TFZ) is a diamond shaped transform zone between the offshore Koblensey ridge and the onland North Iceland Rift Zone. Its southern border is a 100 km long N120°E dextral strike-slip zone, a few kilometres offshore North of the island for 2/3rds of its length. Historically, it is a seismically very active segment with dextral strike-slip and normal mechanisms.
Pull-apparts and pressure-ridges are characteristic of its onland section. The eastern tip is abruptly cut off by one of the fault zones of the North Iceland Rift Zone. Following the 1975 to 1984 Krafla rifting event, the seismic activity of the Husavik fault decreased considerably. It remained negligible for more than eight years. This change can be explained by a sudden stress increase, following the rifting episode, which would have blocked the fault. However, aseismic movement is not to be excluded.
The DIAPASON software, developed at the CNES, is used to form interferograms of SAR scenes from the ERS1 and ERS2 satellites covering the transform fault and its junction with the rift zone. We used 13 scenes, acquired between 1992 and 1998 on two parallel satellite tracks, in order to calculate 36 interferograms with temporal values of up to six years, i.e. potential displacements reaching 12 cm across the transform fault. The large number of interferograms allowed us to discriminate between crustal deformation and atmospheric or topographical artefacts, the first being always present with the same intensity regardless of temporal spacing of scenes and the latter being closely linked to the topography which has been modelled in detail in the area of study as no DEM sufficiently precise existed previously.
These observations will allow us to determine the behaviour of the transform fault after a major rifting episode, particularly with respect to the relative importance of aseismic movement and locked state for the TFZ.
Seismogenic faulting is commonly associated with zones of fluid overpressure. The origin and magnitude of the fluid overpressure are, however, not well known. In studies of seismogenic fault zones it is commonly assumed that only in zones of thrust faulting can the fluid pressure attain, and possibly exceed, that of the lithostatic (overburden) pressure. Using the appropriate equations for hydrofractures and the elastic properties of the host rock, the static fluid overpressure in a fault zone can be related to the aspect (length/maximum thickness) ratios of hydrothermal, mineral-filled veins. If the veins grow as self-similar structures their measured aspect ratios are a crude estimate of the fluid overpressure in the fault zone during vein formation. Using the average aspect ratios of 379 mineral-filled extension (mode I) veins from the Tjornes Fracture Zone, an active strike-slip fault zone (a transform fault) in North Iceland, the fluid overpressure, during their development, with reference to the minimum compressive principal stress is estimated at 20 MPa. Emplacement of such veins increases the minimum principal stress and can generate a temporary stress barrier to the propagation of subsequent hydrofractures. On meeting a subhorizontal stress barrier, vertically propagating hydrofractures may change into water sills where the fluid pressure is at or above lithostatic. In this model, stress barriers, and thus water sills, can form at any depth in, and in any type of, fault zones. For such a high fluid pressure, the product of the coefficient of sliding friction and the normal stress in the Modified Griffith Criterion becomes essentially zero and the driving stress associated with faulting equal to twice the in situ tensile strength of the host rock. For typical in situ tensile strengths of 2-3 MPa, the driving stresses for slip on overpressured fault planes is 4-6 MPa. These results are in good agreement with the commonly measured average static stress drops of 3-6 MPa during earthquakes.
In recent years the assessment of seismic hazard has increasingly relied on the identification and characterization of individual potential sources of future large earthquakes. The incorporation of individual physical sources into seismic hazard calculations complements the established practice of outlining homogeneous "seismogenic zones", allowing the highest hazard levels to be more precisely assigned without overprotecting nearby areas characterized by more moderate earthquakes.
In Italy this new trend is largely the result of over 10 years of intense investigations spurred by the 23 November 1980, Irpinia earthquake. So far this effort has produced richer and more reliable earthquake catalogues, a better understanding of the modes and rates of present tectonic deformation at regional scale, and the first directly measured recurrence intervals for destructive earthquakes generated by the same source.
The new "Database of potential sources for earthquakes larger than magnitude 5.5. in Italy" lists over 300 of such sources derived from a combination of geological, instrumental and historical information. Each of the best defined sources is portrayed as a rectangle representing the surface projection of the fault plane at depth, and may be expressed at the surface by one or more active faults logically connected to the parent source by the database.
Geological Sources are constrained by geological and instrumental evidence regardless of the damage distribution of the associated earthquakes (in fact some of them have not been active historically). In contrast, Historical Sources are those that were defined exclusively through a quantitative analysis of the intensity reports of the latest earthquake they generated, without further geological or instrumental constraints: depending on the quality of the data these sources were divided in Type A, which are defined by an estimate of the location, size and orientation of the physical source, and Type B, for which only the source location and length could be estimated.
Inevitably, some or even many sources that were silent during historical times may have so far escaped this type of analysis, particularly in the case of moderate-sized earthquakes in slowly deforming areas. Others may have been mislocated due to the combination of uncertainties on their location and on the assessment of their size. Our next step will be to attempt filling in the gaps using direct geological evidence, seismicity patterns, observations of regional strain and geodynamic constraints.
A similar scheme is currently being implemented in Greece and Spain in the framework of the E.C. "Environment" project "Faust". We envision applications to other countries where the assessment of seismic hazard is necessarily based on the combination historical, instrumental and geological information.
The Gulfs of Patras and Corinth consist a system of WNW-ESE trending asymmetric graben structures and are located within one of the world's most seismically active zones. Marine geophysical studies in these gulfs have shown that the sedimentary cover of the Gulfs of Patras and Corinth are gas-charged while the Gulf of Corinth is considered as the home of submarine instability processes.
In the Gulf of Patras a large pockmark field was activated before and during the 5.4R earthquake which occurred on July 1993. During the 36 hour period prior to the earthquake the bottom water temperature abruptly increased on three occasions. It is considered that these changes were probably the result of upward migrating high temperature gas bubbles in the water column which were caused by the reduction in the pore volume in the sediments resulting from changes in the stress regime prior to the earthquake. Therefore it can be suggested that in seismic areas adjacent to pockmark fields, earthquake prediction may be achieved by automated monitoring the water temperature and/or the rate of gas venting in the pockmark field (Hasiotis et al., 1996). Pockmark fields were also detected in the Gulf of Corinth, near active submarine faults (Soter, 1995).
The existence of active gravitational mass movements in the Gulf of Corinth demonstrated by extensive damage caused to three submarine cables which were laid across the gulf between 1884 and 1957. Furthermore, historical observations give evidence of nearshore sediment failures at least four times during the past 2500 years in the same locations. These failures are associated with the destruction of the ancient cities of Eliki and Voura in 373BC (7.3R), as well as with submergences of long coastal strips during or after severe earthquakes in 1817 (6.5R), 1861 (6.7R) and 1965 (6.5R) (Papatheodorou and Ferentinos, 1997).
In the Gulf of Corinth the 6.1R earthquake on June 1995, caused coastal foundation damages and small scale subaerial to submarine sediment failures in at least four sites, in three fan delta deposits (Papatheodorou and Ferentinos, 1997). The sediment deformational types identified at the failure sites consist of ground cracking, rotational and elongated slides sediment gravity flows and extrusions of mixtures of water and sand (sand boils). SPT and CPT tests conducted in fourteen boreholes and subsequent liquefaction and slope stability analysis showed that the dominant instability mechanism in the fan delta deposits was liquefaction of shallow subsurface horizons. The liquefaction was caused by elevated pore water pressure enhanced perhaps by the presence of gas, resulting from the cyclic loading induced by the earthquake.
Additionally, red mud wastes discharged from a bauxite processing factory on the northern shoreline of the central Gulf of Corinth, have also been found in the base of slope and the basin, in the form of turbidity current deposits. This high water content red slurry seems to be very prone to failure, even during low to medium intensity seismic shocks. Short length sediment cores revealed that the number of turbiditic events in the surveyed area is from 2 to 5 over a period of 15 years.
Hasiotis T, Papatheodorou G, Kastanos N & Ferentinos G, Marine Geology, 130, 333-344, (1996).
Papatheodorou G & Ferentinos G, Marine Geology, 137, 287-304, (1997).
Soter S, In Ancient Helike and Aigialeia: Proc. Second Intern. Confer, in press, (1995).
Turkey lies within the Mediterranean sector of the Alpine-Himalayan orogenic system, which extends from Italy to Burma.Turkey. This system, identified with high mountain ranges and shallow, somewhat diffuse seismicity, constitute one of the most seismically active continental regions of the world with a long and well documented history of earthquakes. After the assessment of the tectonics and the seismicity of the region, encompassing the area between 30-37E longitudes and 33-37N latitudes, a probabilistic hazard analysis is conducted to evaluate the free-field earthquake ground motion parameters corresponding to different return periods. Portrayal of the seismicity and the tectonics of a region provides the essential information towards the assessment of seismic source zones. The epicentral maps of the historical damaging earthquakes in Turkey and vicinity clearly identifies the West, North and Southeast Anatolian regions as hazardous zones. The diffused character of the epicenters is partly due to the location inaccuracies associated with small magnitude events and hinders the detailed correlation of the epicentral locations with the neo-tectonic features. Almost all earthquakes in Turkey and its vicinity is associated with tectonic elements. The correlation of seismicity with the tectonic elements (seismo-tectonics) constitutes an important phase of the earthquake hazard assessment and, as such, several micro-plate tectonics models have been proposed. The bulk of Anatolia, Aegean Sea and Cyprus is located on the Anatolian (or Turkish) Plate. The northern boundary of the Anatolian Plate is the Anatolian Trough and the right-lateral, strike-slip North Anatolian Fault (NAF). The southern boundary of the Anatolian Plate is formed by the Hellenic Arc, south of Cyprus and the East Anatolian Fault (EAF), which joins the NAF at Karliova. The EAF, located between the Gulf of Iskenderun and Karliova, is a left lateral, strike-slip fault. The south-westward motion of the Anatolian Plate, relative to Africa, is taken up by the subduction along the Hellenic Trench. In Western Anatolia, the east-west trending grabens account for most of the seismic events of this region.
The seismicity compilations are based on the specially-compiled catalogs of historical, instrumental, USGS-NEIS, ISC and Kandilli Observatory data. Uniformity in magnitude is implemented by converting all magnitudes to moment magnitude. The data is biased with respect to the with respect to the reporting periods and magnitude ranges, and can only considered to be homogenous for magnitude 4 and above for the last several decades. For large and highly damaging earthquakes (Io>=VIII) which may be assumed to be completely reported, one can see periods (up to several) of no activity mixed with centuries of frequent activityAttention is paid to the cross-correlation of instrumental and historical earthquake data. Seismic sources are identified by using the macro-seismic locations of historic earthquakes and instrumental locations of the last 50 years' earthquakes. Delineation of the source boundaries is based on neo-tectonic elements and sudden variations in the homogeneity of the seismicity. Identified source zones are: Anatolian Trough Zone, North Anatolian Fault Zone, Caldiran Zone, Bitlis-Zagros Zone, Black Sea Escarpment, North-East Anatolia Zone (Pambak-Sevan Zone, Upper Caucasus Zone, Tabriz Fault Zone, Talish Zone, Soltanieh-South Parandak Fault Zone), East Anatolian Fault Zone, Bitlis Suture Zone, West Anatolian Graben Complex Zone, Cyclades and Fethiye Zone and Cyprus Zone (Kyrinea, Levant, Paphos and Southeast Cyprus). In addition a background seismic zone is defined to account for floating earthquakes not accounted by these sources. For the assessment of the recurrence relationships in these source zones the rates of occurrence in different magnitude groups are adjusted by determination of the period over which the data in a given magnitude group are completely reported.
For depicting earthquake ground motion severity, predictive empirical relationships for MSK intensity (attenuation relationships developed for Eastern Turkey and tested for Northern Iran), peak accelerations and spectral amplitudes (regression analysis based on Californian data, Boore et.al. 1994) are considered. For depicting earthquake ground motion severity, predictive empirical relationships for peak accelerations and the spectral amplitudes are considered. For the horizontal peak ground accelerations corresponding to different soil conditions the attenuation relationships provided by Boore at.al (1994) and Campbell and Bozorognia (1993) are used. For spectral amplitudes corresponding to different soil conditions at selected frequencies the regression analysis of Boore at.al (1994) is used. Probabilistic hazard is computed by SEISRISK III routine. The stochastic model in this routine assumes a homogenous Poisson process for earthquake generation. The pros and cons of this assumption are discussed. For the probabilistic hazard analysis SEISRISK III (Bender and Perkins, 1987) routine, in-house improved with graphical pre- and post-processors, is used. The stochastic model used in this routine assumes that the generation of earthquakes in the time domain follows a homogenous Poisson process. The pro's and con's of this assumption is discussed. The routine allows for variability of the source boundaries. A sensitivity analysis conducted indicated that this effect is not important considering the return periods and the geometry of the sources. The results are provided as contours of various ground motion parameters corresponding to different return periods. The contour maps provided for 5% damped 0.3 and 1.0 s pseudo acceleration spectral amplitude with a 10% of probability of exceedance in 50 years can be used for the rational construction of the site-specific design basis response spectrum.
Seismotectonic zonation studies in the Tell Atlas of Algeria, a branch of the Africa-Eurasia plate boundary, provide a valuable input for deterministic seismic hazard calculations. We delineate a number of seismogenic zones from causal relationships established between geological structures and earthquakes and compile a working seismic catalogue mainly from readily available sources. To this catalogue, for a most rational and best-justified hazard analysis, we add estimates of earthquake size translated from active faulting characteristics. We assess the regional seismic hazard using a deterministic procedure based on the computation of complete synthetic seismograms (up to 1 Hz) by the modal summation technique. As a result, we generate seismic hazard maps of maximum velocity, maximum displacement, and design ground acceleration that blend information from geology, historical seismicity and observational seismology, leading to better estimates of the earthquake hazard throughout northern Algeria. Our analysis and the resulting maps illustrate how different is the estimate of seismic hazard based primarily on combined geologic and seismological data with respect to the one for which only information from earthquake catalogues has been used.
Chaotic time-series arise in many fields, but one field in which their prediction is of critical importance is finance. A number of numeric tools was developed that enable technical analysts to forecast future trends in financial markets - the so-called "financial oscillators".
Time-series of earthquake parameters are also of a chaotic nature, often with comparable fractal dimensions, and testing the oscillators was a natural step. The first applications of these tools for seismic prediction, though encouraging, had two shortcomings: first, their outputs are qualitative, since the oscillators only indicate if a trend is rising, declining or stable ("buy", "sell" or "hold"); then, when we apply several oscillators to the same sequence the results are not always consistent.
Both quantitative output and consistency were achieved by integrating the oscillators in an artificial neural network (ANN). ANN's are software emulators of the nervous system and are adequate because of their mathematical universality, fault-tolerance, and ability to deal with semi-quantitative data such as the modified Mercalli intensity (MMI).
The ANN that was built was a partially connected, feedforward, backpropagation net with 30 input nodes (10 oscillators times 3 parameters: time, MMI and location), 8 hidden nodes and 3 output nodes, learning rate 0.35 and momentum 0.50. The ANN was trained with historical and instrumental seismic data from the Azores, between 1912 and 1993, for earthquakes with MMI greater than IV. It forecasted an earthquake to be felt in the Azores Central Group with MMI between VII and VII, between September 1997 and July 1998.
Data of a different nature were consistent with the ANN forecast.In late 1997, the fractal dimension of the geometrical distribution of epicentres in the Azores was 1.18 which indicated a crustal state of acceleration creep (Alves et al, 1998). In addition, it had been verified that the temporal evolution of this dimension, in the area, is parallel to the time-series for earthquake magnitudes and, at that time, the trend for dimension was rising (Alves, 1998). These data prompted the authors to state that "the fractal dimension for epicentre distribution seems to show that the Azores are currently in a phase of stress-buildup that precedes a major earthquake".
In July 9, 1998, an earthquake struck the Azores, being mostly destructive in the Central Group (MMI VII).
When this last earthquake is included in the ANN training set, the net forecasts an earthquake to be felt in the Azores Central Group with MMI between VI and VII, between August 2003 and August 2004. It is therefore advisable to continuously monitor the evolution of the fractal dimension of epicentral distribution.
This work was funded by project PRAXIS XXI 3/3.1/CEG/2619/95
Alves EI, Azevedo JMM & Dias JLF, Proc V Port Geol Congr (In Press), (1998).
Alves, EI, Proc V Port Geol Congr (In Press), (1998).
Mt. Merapi in Central Java is a strato volcano characterized by dome building. There were different dome growing episodes related to different sites on the southern and western upper part occasionally accompanied with eruptions as in 1992, November 1994, in January 1997, and in June 1998. Different dome growing episodes are also characterized with direction changes of avalanches and ardente nuee, from the southwestern slope in 1992 to the southerly slope in 1994 and since middle of 1998 abruptly to the western slope.
Seismic and tilt data were analyzed, to find out similar patterns before and during eruptive processes and for indications of expecting new dome growing episodes. For many events multi phases quakes (MP) and rock falls increase slowly some time before a bigger eruption and/or dome collapse happens. The main activity often is accompanied by other relevant seismic activity such as volcanic tremor, local tectonic quakes and seismic low frequency events.
Measurements with tilt-meters installed in a cave at Plawangan on low altitude and on the summit of Mt. Merapi in 1992 unraveled increasing surface deformation during landslide activity and before some eruptions. All of these instruments were destroyed from different volcanic activity in 1994 and 1998. Measurements using clusters of bore-hole tilt-meters installed on the flanks at higher altitudes through GFZ-Potsdam and BPPTK (VSI/Indonesia) in 1995 and 1996 provided some indications to deformation related to volcanic activity; due to topographic effects and ground water changes that clues might be still hidden in the data.
In some cases the seismic and the deformation pattern are similar during different dome growing episodes, but for some reasons, dome growth sometimes proceeds without strong relations to the seismic activity.
After the devastating Kobe earthquake (January 17, 1995, M7.2), Science and Technology Agency of Japan initiated the Earthquake Frontier Project. Within the framework of the project, two five-year sub-programs are related to short term earthquake prediction research based on electromagnetic methods. One is the International Frontier Program on Earthquake Research of The Institute of Physical and Chemical Research (RIKEN) and the other is the Earthquake Remote-Sensing Frontier Program of the National Space Development Agency of Japan (NASDA). The main scientific objectives of two programs are ; 1) to clarify whether electromagnetic precursors are useful for short-term earthquake prediction or not, 2) to understand what are the physical mechanisms of the electromagnetic precursors. In Japan, like in some other countries, many reports have been made by several research groups during the last two decades on the electromagnetic precursors of earthquakes. The efforts of these groups, however, have not been systematically planned or funded well. Actually, some precursory phenomena were accidentally discovered during observations for different purposes. The new Earthquake Frontier Project has enabled us to begin to unify these unrelated efforts into an integrated and coherent one, and to initiate international cooperative research activity. Current study contains the following methods; DC range telluric current observations (VAN method), ULF (1 mHz - 10Hz) three-component magnetic observations, ELF (223Hz) narrow band three-component magnetic observations, VLF(1 kHz - 30 kHz) two-component magnetic observations and LF (150 kHz -170 kHz) electric pulse observations. These methods are to detect "active preseismic signals" possibly emitted from the earthquake sources by various mechanisms. In contrast, we also monitor the often reported "preseismic anomalous transmission" of electromagnetic waves, such as temporal reception of FM radio waves (VHF) beyond the line of sight and the phase and amplitude anomalies of VLF and LF radio (and Omega) waves. They are caused by some disturbances in the ionosphere, which in turn, may possibly be caused by some electric charge production in epicentral areas. Furthermore, we try to incorporate satellite magnetic observation results. Already we are obtaining highly significant results, some of which will be presented, together with our compilation (covering the frequency from DC to VHF) on what actually happened electromagnetically in the case of the Kobe earthquake.
Fluids are suspected to play a major role in earthquake mechanics, especially in the case of the weak San Andreas Fault. Numerous physical and chemical parameters have been measured on groundwaters because they are easy to sample and monitor with a good time resolution as a function of seismic activity. However, space resolution is very poor and in most cases, groundwater is not directly involved in faulting. We present here the first isotope study comparing deformation zones (gouges, breccias, slickensides ...), and vein fillings with their hosts and the fluids associated with these materials, as sampled by fluid inclusions. We are investigating ca. 300 samples of various lithologies from over 25 localities along the San Andreas and adjacent faults, for analyses of C, O, H, Noble Gases, Sr, and Pb isotopes, and Trace Elements concentrations.
Calcite fills most of the veins and repetitively occurs as an accessory mineral in deformation zones. The C- and O-isotope systematics of carbonates from veins, deformation zones and their hosts form a large and consistent trend suggesting percolation by external fluids of similar compositions and origin. In each sample site, from most to least isotopically depleted, thus from most to least infiltrated, are the deformation zones, the vein fillings and the host rocks, respectively. Veins are not injected more than once in contrast to the deformation zones. The isotopic trend shown by veins and deformation zones follow those shown by limestones, marbles, gneisses, granites and basalts, matching their increasing content of metamorphic/ deep crustal carbonates. Cracked fluid inclusions exhibit helium isotope ratios 0.1 to 2.5 times the ratio in air, indicating that past fluids percolating through the SAF system contained mantle helium contributions of minimum 1 to 32%, similar to that measured in present-day groundwaters associated with the fault (Kennedy et al., 1997).
The infiltrated deformation zones, veins and host rocks show that fault zones in the San Andreas system maintain a higher permeability than adjacent regions. Our stable isotope, noble gas, and trace element systematics both show consistent evidence for the involvement of mantle-derived fluids in faulting, supporting the assumption of a deep source of fluids at lithostatic pressure percolating through and weakening the fault zone (Rice, 1992), in addition to the infiltration of deep metamorphic water and CO2. Different fluid supplies contributed by different lithologies and reservoir must be taken into account for assessing the recurrence time needed to rebuilt pressure during the seismic cycle.
Kennedy BM, Kharaka YK, Evans W, Ellwood A, DePaolo DJ, Thordsen J, Ambats G, Mariner RH, Science, 278, 1278, (1997).
Rice JR, Academic Press, 475, (1992).
The quest for geochemical precursors of earthquakes has largely been concentrated on gases monitored in the vicinity of active faults. Although some clear anomalies have been found, gases often yield equivocal results because they are prone to perturbations by environmental factors. Furthermore, the exact nature of the interactions between these gases, rocks and groundwaters during earthquake preparation is not easy to establish.
The use of dissolved elements in thermal springwaters appears to give simpler geochemical signals. For instance, we carried out major element analysis in a Pyrenean springwater sampled across a time series of 598 days encompassing an earthquake of magnitude 5.2. A chlorine anomaly was clearly apparent 5 days before the earthquake, and lasted for about two weeks. Nevertheless, the amplitude of this anomaly was limited as Cl concentration rose at only 35% above background values. Other works published in the literature also show that major anions display seismic-related anomalies with limited amplitudes, commonly less than a factor of two of background concentrations.
However, as a general rule in geochemistry, trace elements should follow major element variations in a much more magnified way. Ultratrace analyses of the Alet springwater time-series have therefore been carried out using a double-focusing ICP-MS attached to an ultrasonic nebuliser with its desolvatation unit. Lead concentrations show a tenfold increase and isotopic compositions shift toward anthropogenic values four days prior to the quake. These anomalies lasted for one week only, but according to the expectations, the lead elemental anomaly was 30 times more intense than the Cl one. The time-shift between the Cl and Pb geochemical anomalies and the earthquake, combined with hydrogeological constraints are then used to infer where and when strain changes went beyond the threshold required to temporarily mix normally independent aquifers. It appears that the Cl anomaly can be explained by the limited mixing of two different deep springwaters, whereas the Pb one probably results from an increase of superficial water contamination of the Alet hydrogeologic system. It thus appears that groundwater geochemical tracers can be an interesting alternative to geophysical tools to decipher the crustal deformation occurring during the earthquake preparation processes.
Widespread radon anomalies, kilometres size, have been detected at the surface (50 cm depth) and the subsurface (in galleries, tens to hundreds of meters depth) of the volcanic edifice of Tenerife, Canary Islands (Martin et al., 1997). The overall association with geological and structural features of the island, with thermal anomalies and geochemical indicators of latent volcanic activity indicate that radon anomalies are also reflecting a similar connection to the volcanic system. A reasonable assumption is therefore that temporal variation of radon at the subsurface in Tenerife might reflect changes, among other things, in the volcanic gas upflow system within the volcanic edifice.
Initial SSNTD monitoring, using one month integration time during 1 to 3 years, showed two temporal patterns, seasonal and random, both at surface and at the subsurface (Martin et al., 1994). A radon anomalous area was selected for detailed monitoring, applying SSNTD for 2-3 day integration. Results, from several surface and subsurface stations, showed highly correlated short term fluctuations several days long.
Recently an electronic continuous monitoring system, based on a NaI(Tl) sensor, was installed at the subsurface of the same radon anomaly, at 400 m from the gallery entrance and 100 m below surface. Using an integration time of 10 minutes several patterns are observed: a) quiet and uniform levels interpreted as background, with 30-50 pCi/l, spanning intervals days to several weeks long, b) intense fluctuations exhibiting several peak forms - symmetrical, asymmetrical, single, multiple, etc.
An additional alpha-meter sensor (15 min. integration time) was installed at the same site during one month. A very high correlation was found between the record of both sensors. The high quality of the obtained data using such electronic systems will enable a new insight into the applicability of radon as a monitor of subtle volcanic manifestations.
Martin MC, Steinitz G & Soler V, Rare Gases Geochemistry IV Int. Conf. (Rome, Italy). Abstracts, O10, (1997).
Martin MC, de la Nuez J, Quesada ML, Ahijado A, Carracedo JC & Coello J, International Volcanological Congress, IAVCEI. Ankara (Turkey). Abstracts, (1994).
A current debate rages on whether earthquake prediction is just difficult or forever infeasible. But among earthquakes, aftershocks are doubtless the most predictable events, and thus studies of aftershocks are essential to understand earthquake occurrence. An earthquake releases part of the stress that accumulates with plate motion. But the earthquake also raises the stress at some sites near the fault rupture. Recent studies of a suite of California and Japanese earthquakes (the 1983 M=6.7 Coalinga, 1987 M=6.0 Whittier Narrows, 1989 M=7.0 Loma Prieta, 1992 M=7.3 Landers, and 1994 M=6.7 Northridge events in California; and the 1995 M=6.9 Kobe and 1997 M=6.3 Kagoshima shocks in Japan) have shown that regions where the Coulomb stress is calculated to have risen sustain increased rates of seismicity-including aftershocks and subsequent mainshocks-and regions where the stress drops exhibit reductions in the rate of earthquakes. The stress changes can thus be regarded as advancing or retarding the time to the next rupture, and so such calculations have potential for quantifying earthquake hazard. Earthquake-nucleation constitutive relations derived from laboratory experiments further permit translation of stress changes into earthquake probability changes. The observed nonlinear dependence of seismicity rate change on stress change is compatible with such a state-dependent formulation for earthquake occurrence. This approach augments the 'time-predictable' model of earthquake hazard assessment, in which the probability of an earthquake of specified size and location drops after the rupture, but does not change elsewhere. Earthquake stress triggering offers an alternative approach to probability estimation that is more strongly time-dependent and location-specific, enabling both near-real-time and long-term hazard assessment.
Earthquakes occur due to faulting and fracturing of rocks; as a consequence, an extensive documentation of the distribution and geometry of natural faults has important implications for the evaluation of seismic hazard in tectonically active settings. Most models for fault growth and scaling are based on analysis of faults which display dip-slip (i.e. reverse, normal) and/or strike-slip kinematics. By contrast, relatively little information is derived from faults displaying oblique-slip kinematics. Yet, transtension and transpression play a fundamental role during deformation of the Earth's lithosphere, and any displacement zone margin with an irregular geometry is bound to exhibit oblique convergence and/or divergence. Therefore, studies from oblique-slip faults can contribute to a more complete understanding of the fault propagation processes.
Detailed investigation on mesoscopic transpressional faults from the Salinian Block of California and transtensional faults from the Southern Apennines of Italy reveal a complex kinematic history of fault propagation. Faults initially nucleate as isolate segments, which are later kinematically and mechanically linked via development of diffuse deformation zones and localised oblique connecting splays. The geometry of observed mesoscopic faults is similar to that of the host, larger structures, thus suggesting that the produced fault patterns are scale independent. The overprinting relationships among minor fault-related fabrics permit to define a relative chronology within fault arrays, thus enabling for a general sequence of structural stages to be correctly established. Based on minor fabrics and their overprinting relationships, a 3D kinematic deformation model of oblique-slip fault growth by segment linkage is presented. The proposed model indicates that, as the fault segments overlap and interact, the intervening volume of rock is subject to experience distributed and localised propagation of connecting fault splays, thus enabling to predict which areas are more prone than others to experience faulting.
The modes of fault interaction are crucial for relating the fault propagation processes to seismic hazard assessment. The seismicity related to activity of the large structures which host the mesoscopic faults described in this study is not random, but rather appears concentrated in the areas where fault segments are connected by oblique fault splays, outlining the high vulnerability of fault overlap zones. The recognition of capable and active overlapping oblique-slip fault segments is therefore a critical step towards an improved evaluation of seismic hazard in tectonically active regions.
0461
Large earthquakes with surface rupture commonly occur along faults with non-optimum failure orientations, such as the San Andreas Fault System in California. Such earthquakes, nucleating along faults at depth, can only propagate to the Earth's surface if the stresses during faulting exceed the strength of the uppermost crust. The strength of upper crustal faults is approximated by the ambient stress, which increases linearly with depth, as known from borehole measurements. This stress gradient may not be extrapolated beyond the base of the hydrostatic fluid pressure regime, as suprahydrostatic fluid pressures can reduce the stresses required for faulting. The base of the hydrostatic fluid pressure regime is expected to have a minimum depth of 3 km in areas with a geothermal gradient of ~30°C/km. The average differential stress of 45 MPa at this depth must be exceeded during faulting for surface breakage to occur. However, the maximum stress difference permissible for the reactivation of unfavourably oriented faults is limited by the strength of their wall rock.
Based on failure criteria for the wall and fault rock, a comprehensive set of algorithms has been developed to constrain maximum stress differences for compressional shear failure along unfavourably oriented faults. These algorithms can be applied to cohesive, as well as incohesive faults, whether they are locked or creeping. Maximum stresses for fault reactivation are calculated for typical granite and sandstone wall rock strengths and for different fault-rock types. The simultaneous evaluation of these maximum stresses against the minimum stresses required for earthquake surface rupture constrains the maximum fault angles for such hazardous events. These maximum fault angles, enclosed by the fault plane and the maximum principal stress direction, are estimated to be less than 65°. Hypocentral shear stresses required for earthquake surface rupture are about 30 MPa if a shear stress drop of 10 MPa is assumed. For faults with long interseismic periods, and thus inferred high cohesive strength, the predicted reactivation angle is ~55°. Applying these results to the San Andreas Fault System in Southern California suggests that the San Bernardino area, parts of the Elsinore Fault Zone, and the San Jacinto Fault provide the most probable hypocentral sites for future large earthquakes that disrupt the surface.
The volcanic hazards in Greece are generally very few apart from the areas along the volcanic arc and especially the areas of active volcanic centers, such as Methana, Milos, Santorini, and Nisyros. The impending hazards of these areas could be characterised as low relative to other volcanic zones of the planet, however they exists, which automatically implies to put on some measures for reducing the risks. Despite the fact that volcanic action along the volcanic arc seems limited, however there is increased risk for seismic-volcanic activity, that is a "volcanic activity-excitation" with or without any apparent indication on the surface, but with simultaneous occurrence of seismic activity, due to ongoing activity in the volcanic center. The occurrence of such seismic activity even if it does not include high magnitude earthquakes- usually not over than 5.5R- however ought to the possible small depth of the focus- 1 to 5 km- it can be proved especially hazardous for the neighboring towns which are also characterised by high seismic vulnerability. The seismic-volcanic hazards are controlled mainly by two significant factors and especially by the big fault zones and some particular geological formations with problematic seismic response. Especially in Greece, along the volcanic arc the volcanic activity was controlled or accompanied, in almost all the cases, by large fault zones which were active prior, during or after the volcanic activity such as the outstanding examples in the volcanic region of Milos and Nisyros. These fault zones consist the vulnerable, or the high risk areas. Similar case is the 1992 earthquakes in Milos island, which caused reactivation of large faults with a surface occurrence. Similar phenomena were also observed in the island of Nisyros during the period 1996-1997. Another factor which increase the seismic-volcanic hazards is the presence of volcanic formations which show problematic response. Especially the occurrence of volcanic products, such as tuffs, ash and other pyroclastics which consist the non-cohesive rocks and also with small thickness overlying solid rocky lavas, represent a negative soil-dynamic frame, unsuitable for founding any constructions, except if special construction measures are taken. Additionally, the morphological dips compose the favorable conditions for landsliding phenomena particularly in cases of earthquakes. The historical documents which referred to the occurrence of earthquakes during the periods of volcanic activity or in the intervals, the recent examples of catastrophical earthquakes in volcanic centers in the domain of Greece (i.e., Milos 1992, Nisyros 1996), as well as the above mentioned documents, imply the necessity of input and evaluate the data regarding the seismic-volcanic hazards which together with the rest of the data will help in the better, long term land-use planning of the volcanic areas.
It is well known that the island of Zakynthos is situated in the outer part of the Hellenic Arc and very close to the convergence limit of the two plates, the European and the African one. This geotectonic setting of the island and the intense neotectonic deformation which is represented mainly with faults which are also the cause for the occurrence of the seismic shocks. Despite the severe problems created by the earthquakes, so far there has not been attempted any systematic mapping and study of the faults. Therefore there is not a complete view of the seismotectonic frame. As a result it is not plausible to make any effort to organise an antiseismic plan and survey for the reduce of the seismic hazards. Within the framework of a neotectonic research it was distinguished and plotted in a small scale, all the faults and fault zones. Based on this elaborate tectonic analysis it was shown that the most active faults and fault zones are: 1); The Zakynthos fault with a general orientation NE-SW which intersects and displace Pleistocene formations, limits alluvial formations. This fault basically cut across the town of Zakynthos. 2); The faults in the area of Mpohalis-Gerakari with a general orientation NW-SE which intersect and displace the Pliocene and Pleistocene formations. 3); The fault zone in Bolimnes consists of 3-4 faults which cut even younger to recent formations of scree and fans. The fault zone of Bolimnes is directly associated with the seismic activity. 4); The fault zones on either side of the Keri bay with a general orientation E-W, which create the topographical lowering of the bay and intersects even the recent formations. The classification and the precise mapping of the active faults is one of the first and essential elements for the antiseismic planning and the survey of the island, considering that: 5); Areas with the highest risk are precisely surveyed. These areas correspond to the traces of the faults and the fault zones. 6); The correlation of the faults and the earthquake focus gives a clearer view of the recent deformation. 7); It is possible to form appropriate aid plans in case of an earthquake, so as to considerably reduce the effects of the shock or generally any geodynamic episode. 8); Large scale development plans or public utility works can be designed in such a way so as to be unaffected by fault reactivation and the effects of the consequent seismic activity, especially regarding that their normal utility is considered necessary.
The North Aegean area is characterized by high seismic activity. Epicenter distribution of earthquakes with magnitudes greater than 4.0, during the period 1963-1992 reveals that the main seismically active fault zones, correlate well with the dominant morphological features of the area. In order to gain an initial insight to the seismotectonics of the area the fault plane solutions of major earthquakes in the area were examined. The solutions proposed by previous authors were grouped per event and the best solution for each event was selected taking into consideration the epicenter distribution and the morphological and tectonic features of the area.
Next, the data recorded by a 10-station temporary seismic network installed in the North Aegean were analyzed. Epicenter distribution of small magnitude earthquakes, recorded during the period September 1993 - December 1995, delineates the major fault zones and tectonic structures in the area (North Aegean and Sarros troughs, Sporades basin). Focal mechanisms of selected microearthquakes in association with the epicenter distribution and depth cross-sections indicate NNW-SSE extension in the area. The main tectonic features of the North Aegean and Sarros troughs are delineated by NE-SW fault zones and connected or divided by secondary fault zones trending NW-SE. The calculated fault plane solutions indicate predominantly normal faulting but strike slip or reverse faulting is also locally observed.
The high seismic activity the corresponding seismic hazard implications in the area were emphasized by two major earthquakes which caused minor damage. The first one with magnitude 5.9 occurred in May 1994 in the southeastern part of area between Lesvos and Chios and the second with 5.4 magnitude in May 1995 near the town of Arnea in the Chalkidiki peninsula. The detailed study of these events and their aftershock sequences recorded by the local network provided additional valuable information about the tectonics of the area.
Lisbon is placed in a relatively high seismic risk area. This risk is associated to several seismogenetic origins both intraplate and interplate which gave rise to some important historical earthquakes such as the one on the November 1st 1755, and many registered earthquakes. The big earthquake of 1755 was the first one to have a very complete description of the effects produced. The effects caused by the 1755 earthquake show, in Lisbon, a narrow relationship between the damage distribution and local geology. The most damaged areas, located on the East side of Lisbon, were the ones that have alluvial subsoil followed by the areas that have Miocene subsoil. The Miocene formations are composed of over-consolidated hard soils and soft rocks, such as sands, clays and sandstones that are decompressed near the surface. The West side, where the damage was less significant, is mainly composed of Cretaceous limestones and Neo-Cretaceous basalts. The Downtown Lisbon was one of the most affected areas.
In order to minimise the risks it is very important to evaluate the role of the different parameters involved. The knowledge of the geotechnical behaviour of the materials that compose the subsoil is one of the important aspects of this multidisciplinary work, essential to the knowledge of the site effects.
The seismic modelling must be based on the detailed geology knowledge. In urban areas, such as Lisbon, where it is very difficult to find an outcrop, the geological engineering boreholes take on a great importance. They allow a detailed knowledge of the local geology as well as a geotechnical characterisation of the materials.
Following the work done in Almeida et al (1997) in the area of the Castelo de São Jorge hill, the detailed geological-geotechnical mapping for the Martim Moniz area and the 1-D seismic modelling was preformed. This area is part of one old branch of the Tagus River, the Arroios tributary. The valley deeply carved in Miocene formations has a thick cover of alluvial and coverage materials. The 1-D seismic modelling was calculated at different points of the valley allowing a better visualisation of the seismic behaviour of this area and an understanding of the effect of the different behaviours of the materials (lithology, thickness and compaction, among others).
Almeida I, Lopes I, Almeida F & Teves-Costa P, Caracterização Geotécnica da Colina do Castelo. Abordagem preliminar para a estimativa do risco sísmico. 3º Encon. sobre Sismologia e Eng. Sísmica, IST, Lisboa, 129-136, (1997).
If T is a random variable defined as the time of failure of a system(that is, the time of occurrence of a particular event), with probability distribution function F and probability density function f, the hazard function or failure rate is defined as r(x)=f(x)/(1-F(x)).r(x)dx might be thought of as the instantaneous probability of failure at x, given survival to x. (see, e.g. Prakasa Rao, 1983). The study of this function has a great interest, particularly when the particular event measures the occurrence of an earthquake in a certain area of study (Rice and Rosenblatt, 1976), because the hazard function performs the risk of occurrence of a new earthquake at time x. The estimation of the hazard function by means of nonparametric estimates has been quite important in the statistical literature in the recent years (see, e.g. Hassani et al., 1986), but there are not too many papers about hazard function estimation in dependence situations (that is the case when we observed earthquake or microearthquakes occurrences). In this work, we study a nonparametric estimation method of the hazard function in a dependence context (that is, when we work with data that pertains to a time series). We propose a particular estimator, we comment its asymptotic properties and we study the bandwidth selection problem. We apply our theoretical results with an application to a set of observations of the time interval between earthquakes in a certain area of Spain.
Hassani S., Sarda P. & Vieu P., Rev. Statist. Appl. (Approche non paramétrique en théorie de la fiabilité), 35, 27-41, (1986).
Prakasa Rao, Nonparametric Functional Estimation, (1983).
Rice J & Rosenblatt M, Sankhya (Estimation of the log survivor function and hazard function), 38 A, 60-78, (1976).
Historical data and recent instrumental measurements negate conventional opinion of Lithuania as an a-seismic territory. More than 40 strong (intensity of 5-7, MSK-64 scale) earthquakes were reported from the Baltic Region since 1602. Several hundred low-intensity (M up to 3) seismic events were registered by seismic stations since seventies. Majority of seismic events align along the first-order faults that points out to their modern activity. Neotectonic amplitudes of the faults vary in order of 10-20 m. Potential field transformations and available DSS data show soling of these faults into the lower crust. The seismic activity is not uniform over the Lithuanian territory and varies significantly. The territory of Lithuania was divided into the areas of different potential seismic activity evaluated in MSK-64 scale. The pattern of areas shows the best correlation to the heat flow and crystalline basement lithologies. A potential seismic activity in the eastern and northern Lithuania was assessed as of 6-7 (MSK-64 scale) and is a lower one in the remaining territory (less than 5). The higher activity in these parts might be accounted to more mafic composition of the earth's upper crust and the lower heat flow. 11 potential seismically hazardous zones were defined in Lithuania, which coincide with deep-seated faults. The definition of the zones was based on seismological observations, neotectonic and geophysical characteristics, He measurements in ground-water etc. Depths of seismic activity, magnitudes and influence distances were assessed for the zones. All that information has been compiled in the Map of Seismic Activity of Lithuania. Some ecologically dangerous industrial objects are located on the seismic zones or are situated within the territories of assigned higher seismic activity (Ignalina NPP, Kedainiai and Jonava Chemistry Factories, Mazeikiai Oil Refinery etc.).
Georisc-NZ is a joint French-New Zealand programme studying active faults in New Zealand towns. It aims to accurately locate potential areas of surface rupture, as part of seismic hazard evaluation. Georisc-NZ is mainly based on Ground-Penetrating Radar (GPR) imagery. Locations of the March 1998 survey were selected to test the GPR imagery in a variety of tectonic and sedimentary environments; several radar frequencies were also tested.
Known and previously unknown strands of faults have been identified in the selected sites. In the streets of Wellington (capital city) and suburbs, GPR images reveal the branching geometry of the Wellington strike-slip fault and precise its accurate location within 1 m. Another leg of the survey was devoted to the reconnaissance of the Poukawa Fault Zone that cuts through several North Island towns. Strike-slip, reverse and normal faults have been imaged, corresponding with previous trench observations. Sites were also chosen to examine active strike-slip faults within South Island towns. Their complex sub-surface geometry of several strands and step-overs has been illustrated by GPR imagery. Technical limitations of GPR, imposed by sediment and fault characteristics, have been identified during the survey. Penetration depths and GPR resolution were strongly affected by sediments types (worst in clayey gravels, swamp/marine deposits, volcanoclastics), by the groundwater level and by the gentle dip of fault planes. Alluvial and colluvial sediments in conjunction with steeply dipping faults proved more suitable for imaging.
GPR thus provided accurate data on the location and characteristics of several faults within urban areas. This approach, accompanied by advanced GPR processing, helped in targeting areas for other seismic hazard studies.
Most seismically active faults in the Eastern Alps are reactivated structures from Miocene. From the distribution of fault length together with the depth of the brittle crustal layer the calculation of a maximum credible earthquake is possible. If we assume a rupure process of the whole faults, higher magnitude values are obtained than from the data of catalogues even considering historical data. As the time of human observation of earthquakes is extremly short compared to any geological process, the data may be incomplete and additional methods are needed for hazard estimation. A discrimination of hazardous faults is attempted by the comparison of fault traces with earthquake epicenters, the comparison with fault plane solutions, the offset of quaternary sediments (for example in the Vienna basin or the Lavanttal in Austria) A comparison of reconstructed Miocene convergence rates with the recent plate motion indicates a continuous convergence with the velocity similar to the Miocene. Frequency analysis of fault lengths show that most faults have lengths between 10 and 30 km. Large fault zones like the Inntal- Salzach-, Ennstal, Mur-Mür- Vienna Basin- and Lavanttal and Periadriatic fault in Austria display variable segmentation with about 100 km maximum length of individual segments. The drawback of this approach is that we have yet only few sound data on segmentation processes, so that further research is needed.
Shear-wave splitting (seismic bi-refringence) in the crust is controlled by crack densities and crack aspect ratios in fluid-saturated microcracked rock (Crampin, 1994). The hypotheses for 'stress-forecasting' earthquakes are that: (1) the build-up of stress before earthquakes causes progressive changes in crack geometry until a level of cracking known as fracture criticality is reached when the rock loses shear-strength, the rock fractures, and there is an earthquake; and (2) the approach to criticality can be recognized by analyzing shear-wave splitting (Crampin, 1998). The evolution of such fluid-saturated microcracked rock under changing conditions can be modelled by anisotropic poro-elasticity, or APE (Zatsepin & Crampin, 1997). APE-modelling supports both hypotheses (Crampin & Zatsepin, 1997). Note that changes in shear-wave splitting monitor the effects of stress on the rockmass, where such changes can stress-forecast the likelihood of future events within a region whose size depends on the magnitude of the future earthquake, and where the longer the increase continues the larger the potential event. Note that stress-forecasting does not predict the time, place, and magnitude of future earthquakes. However, it can be argued that a stress-forecast crescendo of increasing risk is the optimum scenario for mitigating hazard by allowing precautionary measures to be taken. The difficulty in testing these hypotheses has been that recognizing changes by analyzing shear-wave splitting requires a swarm of small earthquakes, to provide a source of shear waves within the shear-wave window of at least one digital three-component seismic station, comparatively close to the epicentre of a potential large earthquake. These requirements are severe. Until recently, changes had been recognized serendipitously before only five earthquakes worldwide (two in California, one in Arkansas, and two in China). The seismicity of the transform zone of the Mid-Atlantic Ridge in SW Iceland and the extensive seismic network of the SIL Project (Stefansson et al., 1993) now allow changes in shear-wave splitting before earthquakes to be recognized routinely. As a consequence, since 1996, changes similar to those observed elsewhere have been identified (in retrospect) before several earthquakes (and the Vatnajokull volcanic eruption) at several seismic stations in Iceland. This means that it can only be a question of time before a future large earthquake is stress-forecast in Iceland. It is anticipated that a large earthquake will be stress-forecast before EUG99.
Crampin, S, Geophys. J. Int, 107, 185-189, (1991).
Crampin, S, Geophys. J. Int, 118, 428-438, (1994).
Crampin, S, Trans. R. Soc. Edin., in press, (1998).
Crampin, S & Zatsepin, SV, Geophys. J. Int, 129, 495-506, (1997).
Stefansson, Ret al, Bull. Seism. Soc. Am, 83, 696-716, (1993).
Zatsepin, SV & Crampin, S, Geophys. J. Int, 129, 477-494, (1997).
Nisyros Island belongs to the eastern edge of the Aegean Volcanic Arc and it is located at about 10 miles south of the island of Kos, in Dodekanese, Greece. It is composed of Quaternary volcanic rocks, which include a first stratovolcano type sequence of lavas and pyroclastics (A,B,C,D), followed by the rhyolites of Nikia, which all together constitute the first period of the Nisyros volcano building process with final event the formation of the caldera. The second post-caldera period of Nisyros volcanism includes the massive extrusions of the dacitic lavas of Prof. Ilias. No major volcanic eruption is known during the historical period on Nisyros, with the exception of some phreatic explosions in 1422, 1830, 1871 and 1888.
The only major destructive earthquake of this century occured in 1933, with extended damage at the Emporio village, which was semi-abandoned. In 1996 an earthquake activity started with magnitudes between 4 and 5 R scale, which in July 1996 damaged about 30 houses in the western part of Mandraki, the capital of Nisyros. The damaged houses occur all along a narrow street along a valley developed parallel to the hill of Panaghia Spiliani Monastery, which is following a NNW-SSE fault.
The above fault is one of the active faults of Nisyros. Four major normal fault zones can been distinguished in Nisyros with an average throw of about 100 m. The neotectonic structure comprises five blocks, each one with a kinematic character of relative uplift -horst- or relative subsidence -graben. The NW-SE fault zone of Prof. Ilias-Mandraki implies an extension in the NE-SW direction, whereas the NE-SW fault zone of Ag.Irini implies an extension in the NW-SE.
G.P.S. stations were established on Nisyros Island over all five neotectonic blocks, aiming to a monitoring of the crustal deformation. Measurements started in June 1997 and by October 1998 they showed very important displacements, both in the horizontal and the vertical axes, of the order of 10 to 40 mm/per year. Additionally: i) the maximum uplift was observed in a NW-SE direction running parallel to the major Prof.Ilias-Mandraki fault and ii) the horizontal displacement of the block west of the Prof. Ilias-Mandraki fault zone is directed towards the SW whereas the block to the east is directed towards the east. Finally, the southwestern area of Argos is directed to the SE, diverging from the Ag. Irini fault zone. The overall neotectonic kinematics is in agreement with the above G.P.S. measurements and indicates an updoming of the central part of the island together with a lateral escape in three opposing directions from the intersection area of the major main fault zones. The general conclusion is that the seismic risk is related to the activation of the Prof. Ilias-Mandraki fault, which may produce extended damage to the capital of Nisyros whereas the volcanic risk is limited in the area of the craters within the caldera.
By studying the atmospheric processes forming the weather condition it was given till last time little attention to seismic processes or it was thought, that the development of the atmospheric processes is not connected with the conditions of the earth crust. There is wide spread the meaning, that there are possible situations in zones of seismic activity, when the quick basic systems change and called by it sharp atmosphere pressure change are provoking the earthquake ripened in the nature depends from the atmospheric processes.
The carried out investigations and long-years considerations made by the author upon the weather anomalies and seismic activity on the Tashkent seismic - geophysics range (The Middle Asia) and in another regions of the Earth (1980 - 1996 years) are pointing that there exists definite correlation connection between seismic - tectonic and atmospheric processes. This connection displays in the weather deviation from the usual process in one or another region and precedes (accompanies) the seismic activity in seismic zones.
As is known, the earthquake creation is connected with the sharp hit moving of the mountains breed in the earth crust. In the period of the preparation of the earthquake under the influence of tectonic strains and huge pressures there are changing the volume and the form of separated parts of the mountain breed and there is going decreasing of the density of the materials on account of the creation of micro splits, and as the consequence, the increase of the volume (dilatans'y), what moves to the earth surface deformation on the big square.
On this ground there is changing sharply the warm stream income intensity and the earth surface albedo, which are defining the dynamics of atmosphere processes and the weather condition before and after the earthquake. The meteorological consequences are as deeper and longer especially in the epicenter zone as more the magnitude of expected earthquake. These circumstances are giving the possibility to estimate and to separate the effective features of the meteorological harbingers of the earthquakes and to develop the method of the forecast of its place, time and magnitude and so it opens the possibility to consider the problems of weather cataclysms creation from the new scientific positions.
The complex using meteorological and radon earthquakes predictions according to our preliminary estimations reliability of forecast is 90 - 95% when the method would be used.
Tirana-Durres Region is most important inhabited region of Albanie, where more than one third of the population of the country is concentrated (about 1 million people). Two biggest cities situated at both ends of this region: Tirana (the capital) and Durres (the biggest port of the country), nowadays are going to be linked together as a great metropolitan area. The new development have to take into consideration the environmental impact due to expected strong ground motion of future earthquakes. Durres City with about 100,000 inhabitants, situated on Durres Bay on Adriatic Coast, on very poor thick Quaternary sediments represents on of the oldest cities. Durachium, the antique name of the Durres, was one of the most important cities of Rome and Byzantine empires. During its long history of civilization the city suffered from strong earthquakes. One of the strongest earthquakes was the earthquake of 1273, aster which inhabitants of Durres were forced to leave the city and settle at Berat. The latest disastrous earthquake was that of December 17, 1926 (M=6.2, Io=IX).
Taking into account its seismotectonic setting on the Adriatic seacoast, there is one of the interest the experiences got by another big earthquake, that of Albania-Montenegro earthquake of April 15, 1979 (Ms=6.8, Io=IZ+). From seismic hazard point of view, Durres City represents on of the most dangerous areas of coastal zone in Albania, with a lot of problems concerning soil instabilities during future earthquakes representing important expected environmental impacts as:
o Fault rupturing on free surfaces,
o New and induced landslides
o Induced liquefaction phenomena on poor sandy deposits of Durres lagoon and Durres Bay area.
Tirana City, the capital of Albania, situated on a Quaternary depression (Tirana - Ishmi depression) and partially on Tirana Hills (of Neogene age), at legs of dajti mountain (Cretaceous age) (1600 m high), represents an area where expected seismic hazard has to be taken into consideration for its long term and medium term further development.
Tirana City has to experience different strong ground shakings during future excepted earthquakes, which consequences and characteristics will depend on many factors as:
o The local soil conditions
o Great thickness of gravel deposits overlying baserocks
o Small thickness of surface layers of poorer sediments
o Poor layers of former swamps with shallow underground water level
o The characteristics of active faults surrounding the City of just beneath it
o The underground topography (steep slopes and shaepness) of baserocks.
The results of studies carried out and accomplished at the end of 1987 year, were checked by the consequences of a moderate earthquake of January 9, 1988 (M=5.5, Io=VII+) that hit Tirana City. All these phenomena have to be taken into account by urbanists and decision market for an appropriate land use and a sustainable development of these cities and the future metropolitan area, in the framework of physical and urban planning.
The Mw=7.7, July 16, 1990 earthquake is the strongest earthquake to have occurred in Luzon, Philippines this century and perhaps during historic times. The earthquake was generated by left-lateral strike-slip faulting along the Digdig Fault, one of the northern splays of the Philippine Fault. The continuous curved surface rupture extended over 110 km making it one of the largest strike-slip events recorded in the world. Since the establishment of global seismological networks, this is the only recorded seismic event in the region that has produced surface ruptures which clearly manifested the sense of movement along the fault thus allowing a seismological analysis substantiated by geological observations. An inversion of teleseismic body-wave data is conducted using the method of Nábelek in order to obtain a detailed model of the rupture process. 25 Broad-band teleseismic P and SH records from IRIS and GEOSCOPE global networks were used to have the best azimuthal coverage. Information obtained from the surface faulting and the results of a relocation of aftershocks reported by NEIC using the Joint Hypocenter Determination provide constraints on fault length and slip distribution at the surface. The source is complex and several subevents had to be introduced in order to obtain the best fit in the waveform modelling.
Protection and conservation of historical city of Potenza by evaluation of seismic hazardFabrizio Terenzio GizziThe presence in Italy of various and different cultural tradition has allowed the development of an unequalled Cultural Heritage the which higest expression is in historical cities. In Italy, among geological hazard that set serious constraints on the conservation of the Cultural Heritage there are the seismic events that certainly cause the heaviest losses. Nowadays, the protection against seismic events is only possible via the probabilistic definition of event occurrence, i.e. the seismic hazard. In fact, the investigation of seismic precursors (variation of the magnetic properties of the rocks, ground deformation, gas emission, etc.) has not yet provided a unambiguous view of the space-temporal evolution of stress processes involving the Earth's crust. From this viewpoint the study of the seismic hazard, based on the statistical analysis of national seismic catalogues, offers a valuable tool for the identification of areas subject to higher risk and, consequently, having a higher priority in the protection of the population and in the static restructuring of building structures.The city of Potenza has been affected by several earthquakes with maximun local felt of the IX degree MCS and it is characterizated by a frequent seismicity owing to seismogenetic structures near the site (seismic period of year 1990). The seismic hazard can be calculated by means of a detailed knowledge of the seismic history of a Potenza site. However, often the historical documentation is inexistent or insufficient. In such cases it is possible, by means of analitical laws of attenuation of the epicentral intensity with the distance, to reconstruct the intensity detected at a given location.In such cases the local seismic history is reconstructed using a probabilistic technique. This method introducing the concept of site felt probability of a epicentral event. Starting from this introduction and using the relation given by Magri et al. (1994), it is possible to calculate the probability that, for a given earthquake characterised by an epicentral intensity Io and a distance r from the examination site, the intensity Is, assessed at the site is greater than or equal to the chosen value I. After this evaluation, by means the direct method (Cornell,1968, Dargahi-Nourbary,1989) has been assessment the seismic hazard relatively to the interval of completeness determinated with CUVI method. This work contributes to the definition of strategies aimed to a careful evaluation and prevention of the seismic risk (in place of the consolidated a posteriori approach) and of a rational territorial planning.
Boschi et al, Catalogo dei forti Terremoti, 1, (1995).
CNR-PFG, Atlas of isoseismal map of italian earthquakes, 2A, (1985).
CNR-PFG, Catalogo dei terremoti italiani dall'anno 1000 al 1980, 2B, (1985).
Cornell CA, Bull. Seism. Soc. Am, (1968).
Gizzi FT, Le strutture castellane normanno-sveve della Basilicata, (1998).
Index of EUG 10 Volume
Further EUG 10 Information
Index of the Journal of Conference Abstracts
Cambridge Publications Home Page
Last Updated on Wednesday, March 17, 1999.
© 1997 Cambridge Publications