|
|
|
MICROZONING OF THE EARTHQUAKE HAZARD
IN ISRAEL
PROJECT 4
EMPIRICAL DETERMINATION OF SITE EFFECTS
FOR THE ASSESSMENT OF EARTHQUAKE HAZARD AND RISK
TO DIMONA & ARAD
November
2004 Job No 569/076/04
Principal Investigator:
Dr. Yuli Zaslavsky
Collaborators:
Dr. A. Hofstetter, M. Gorstein, M. Kalmanovich, G. Ataev, T. Aksinenko, D. Giller, I. Dan, N. Perelman, V. Giller, I. Livshits and A. Shvartsburg
Submitted to:
Earth Sciences Research Administration
Nation Ministry of Infrastuctures
and
The Ministry of Absorption.
Contract Number: 23-17-023
|
 |
| Part 1: |
| ABSTRACT |
The towns of Dimona and Arad are located in earthquake prone area, in the vicinity to the Dead Sea Transform, where several destructive earthquakes have been occurred in the past. They are recently developed towns and might be the place for future, heavily damaging events, due to the combination of site effects and urban development.
A project to map site effects across Israel undertaken by the Seismology Division of the Geophysical Institute of Israel in 2001 included microzonation in the towns of Lod-Ramla, Kiryat Shemona and Kefar Sava, as well as non-dense grid measurements in the Coastal Plain, and Ha-Shefela region.
In the Arad area we used borehole and seismic refraction data to construct the subsurface models, while in the Dimona area, where borehole information is not available, we utilized only results of the seismic surveys.
The analytical models were implemented into Stochastic Estimation of the Earthquake Hazard (SEEH) procedure to predict acceleration response spectra using ground motion simulations. In order to facilitate estimations of earthquake loss scenarios, we divided the Dimona and Arad areas into zones; each one is characterized by fundamental frequency and amplification. For each zone we adjusted the overall soil-column model for calculation of the response spectrum throughout the whole zone. In some zones, uniform hazard site-specific acceleration spectra obtained are significantly different from the ones prescribed by IS-413.
Six buildings of three and four stories in Dimona were temporarily instrumented with two horizontal component stations for various periods of time. The fundamental frequencies of the translational motions in the NS and EW directions and the frequency of first torsional mode were determined using ambient excitation. The vibration tests show that for three-storey-buildings the fundamental translational modal frequencies for either orthogonal axis is 6.5-7.5 Hz while for four-storey-buildings the first mode motion is 4.3-5.3 Hz. Empirical formula was proposed for calculating the fundamental frequency of three and four stories buildings. The estimated first-mode damping values vary from 2.2% to 3.4% for different buildings. We observed that the response of the soil is strongly influenced by the proximity of structures for distances roughly proportional to their heights. Owing to financial restrictions, we were unable to instrument buildings in Arad.
|
| 1.1 Iintroduction |
Earthquake hazard zonation for urban areas, mostly referred to as seismic microzonation, is the first and most important step towards a seismic risk analysis and mitigation strategy in densely populated regions with a high concentration of developed infrastructure. The objective of seismic microzonation is to quantify the spatial variation of the subsurface response that can be expected in the area for a typical earthquake. The expected ground motions have to be determined taking into account the influence of local soil conditions, i.e., so-called “site effects”. Nearly all recent destructive earthquakes (Spitak, Armenia 1988, Northridge 1994, Kobe 1995, Columbia 1999, Izmit, Turkey 1999 and many more) have provided additional evidence of the dramatic significance of site effects. During the last 20 years, a large number of observational studies have striven to evaluate the importance of the different factors involved in site response of soft soils, concentrated in the determination of the amplitude of the transfer function relating input motion to ground motion on top of the soft soils.
A project to map site effects across Israel undertaken by the Seismology Division of the Geophysical Institute of Israel in 2001 included microzonation in the towns of Lod-Ramla, Kiryat Shemona and Kefar Sava, as well as non-dense grid measurements in the Coastal Plain, and Ha-Shefela region.
We used a three-step approach to evaluate the local site conditions. In the first step, ambient vibration measurements were carried out in order to predict fundamental frequency and amplification factors of the unconsolidated sediments. In the second step, geological and geotechnical data available in the investigated area were collected and interpreted. Finally, analytical estimation of site effects were made and 1D numerical models were constructed
Nakamura (1989) hypothesized that site response could be estimated by simply evaluating the spectral ratio of horizontal versus vertical components of noise observed at the same site (non-reference-site technique). Most studies show that the H/V ratio obtained from microtremors coincides with response functions of near surface structures to incident shear waves (Ohmachi et al., 1991; Lermo and Chavez-Garcia, 1994; Zaslavsky et al., 1995; Gitterman et al., 1996; Konno and Ohmachi 1998; Mucciarelli and Monachesi, 1998; Chavez-Garcia and Cuenca, 1998; Toshinava et. al., 1997; Shapira et al., 2001). There is also another conclusion regarding microtremor horizontal-to-vertical spectral ratios. Recently, Field and Jacob, 1995, Bonilla et al., 1997, Horike et al., 2001 and Satoh et al., 2001 contended that estimates of the frequency of the predominant peak are similar to those obtained from standard sediment-to-bedrock spectral ratio of earthquake records, however the absolute level of site amplification does not correlate with the amplification obtained from this method. Based on measurements of explosions, earthquakes and ambient noise by reference and non-reference techniques, Malagnini et al. (1996) showed that the Nakamura technique failed to identify either the resonance frequency or its amplification. Nevertheless, Seekins et al. (1996) showed that H/V spectral ratio obtained from microtremors agree better with sediment-to-bedrock spectral ratio from S-waves than the microtremor ratio with respect to the reference site. Also Meneroud et al. (2000) show that the H/V ratio from microtremor measurements is very successful and gives the same result as more expensive and time-consuming methods.
The experience gained and results obtained from these investigations were used in the next stage of the general project aimed at an evaluation of site-specific ground motions in the towns of Dimona and Arad. These areas were chosen for the following reasons:
- Several destructive earthquakes have occurred in the Dead Sea area in the past. Due to the proximity to the Dead Sea Transform and taking into account the historical seismicity, the town of Dimona and its surroundings are considered a high seismic risk zone;) were used to calculate a map of 5% damped spectral accelerations (SA) levels 0.2 and 1-sec periods with a 10% probability of excedence in 50 years.
- Dimona is one of the Israel’s developing urban areas;
- In Dimona area we have a geological structure in which soft sediments directly overlay hard carbonates. Such conditions may cause significant amplifications owing to the high impedance contrast between soft soils and the firm basement.
To better characterize the local site response of investigated areas, two main tasks were performed under this project. The first one consisted of the execution of ambient vibration surveys, giving the fundamental frequency and amplification of ground motion for the towns of Dimona and Arad. The second task consisted of the theoretical modeling in order to estimate seismic response during the occurrence of a strong earthquake.
|
| 1.2.1 Dimona |
The town of Dimona is situated in the northeastern part of the Dimona-Yeroham syncline, on a sedimentary fill that unconformably overlies Judea group hard carbonates of Turonian-Cenomanian age. The upper sedimentary section consists of clastic deposits of the Hazeva Formation of Miocene age. The geological units exposed in the investigated area are shown in the geological map (Fig. 1), which is a part of the geological map of Dimona (1:50000; Roded, 1996).
The Judea Gr. Sequence, 250m to 400m thick, has been divided in this study into three parts:
A lower part (unit c1,.) consisting of massive dolomites and limestones of the Albian Hevyon Fm., exposed in outcrops only in the Hatira anticline.
A middle part (unit c2, En Yorqeam, Zafit and Avnon Fms.; and unit c3, Tamar Fm.) consisting of well bedded dolomites and limestones with interbedded marls. This sequence outcrops in the northwestern and western part of the Dimona area.
An upper part (unit t, Derorim, Shivta and Nezer Fms.) consisting of limestones and dolomites of Turonian age. This unit is found in outcrops in the eastern part of the town of Dimona.
The Hazeva Formation is composed of terrigenous and river-lacustrine sediments. Calvo and Bartov (2001) subdivide it into three parts, based on the definition of two regional unconformities. This division has tectonic implications, The lower part (Ef’e and Gidron Fms.) was deposited during a relatively tectonicaly quiet period. The middle part (Zefa Fm.) shows some evidence of tectonic subsidence along the Arava graben and along the Negev-Sinai Shear Zone faults. The upper part (Rotem and Karkom Fms.) was partly syntectonically deposited.
In order to estimate the lithology of the sediments of the Hazeva Group in greater detail and its thickness in the Dimona area, we used information from the “Wadi Fai SH-01” borehole, located 2 km southeast of the area investigated. This is a unique borehole, which penetrates the clastic deposits. A description of the lithological-stratigraphic columnar section near the town of Yeroham made by Harash (1967) was also integrated into the investigation. According to Harash, this section consists of 15m of sandstone and conglomerate in the lower part of the Shualim member of the Zefa Fm., according to the classification of Calvo and Bartov (2001). The upper part (Mingar, Yeroham, Aroer and Ashalon members) of 15m thick is characterized mainly by numerous conglomerate-sand-sandstone-silt-clay sequences of fluvial environments with an oyster-bearing horizon. It corresponds to the Rotem Fm. in accordance with the classification of Calvo and Bartov (2001).
According to our interpretation, the lower part of the columnar section of Wadi Fai SH-01 well is correlated with the Mashaq member (carbonate section of Ef’e Fm.) and represented by alternations of marl and clay, and massive lacustrine limestone 60m thick. The middle part (the Zefa Fm.) consists of a 58m thick layer of coarse sandstone and conglomerate with a carbonate layers or carbonate cemented clastic beds in its uppermost part. The upper part, Rotem Fm., having a thickness of 30m is mainly characterized by sand-sandstone and conglomerate-silt-clay sequences.
The paleogeographical reconstruction by Garfunkel and Horowitz (1966) shows that several tectonically conditioned inland basins were connected by a river system. It originated in Trans-Jordan, flowed through the Arava, passed over the Mahmal anticlinal range and reached the sea through the northern Negev. The river running across the Avedat plateau followed a synclinal trend which enabled it to cross this elevated area. Tributaries came from the Yeroham-Dimona basin and Arad region.
The quaternary section is composed of Holocene alluvium deposits, mapped only in the western and northern part of the Dimoina area in the upper basin of Nahal Dimona and Nahal Aroer. They consist of soil, loess, clay, loam and gravel and are a few meters thick.
Figure 1. Geological map of the Dimona area
|
| 1.2.2 Arad |
The town of Arad is situated in the eastern parts of Aro’er basin which forms a synclinorium between the Kidod, Zohar and Dimona anticlines in the southeast and Ira-Kohal anticline in the west (Zohar, 1987). The geological map shown in Fig. 2 is a part of the 1:200 000 geological map (Sneh, 1998; Sheet 3). Data from three structural boreholes is available within the borders of the town of Arad (for location see Fig. 2). Limestone of Cenomanian age is found at a depth of 150m in the Kidod-1 borehole, while limestone of Turonian age is found at a depth of 65m in the Kidod-3 borehole . The Zohar West-7 borehole, located in the south-western part of the area, penetrates the limestone of the Top Judea group at a depth of 88 meters. From a depth of 43 to 88m yellowish marl with fragments of chalk and chert of the Menuha Fm. is found. The interval 27-43 meters consists of massive chert and is overlain by marl and chert of the Mishash Fm. The upper layer, from 0 to 22 m, consists of alternating marl and chalk of Ghareb Fm.
In general, the geological pattern in the Arad area is characterized by three lithological sections: limestone and dolomites of the Judea group constituting the basement, chalk and chert of the Mount Scopus group and clastic sediments of the Hazeva group. A detailed description of the Judea and Hazeva groups is given above (see Geology of the Dimona Area). Here we shall concentrate on the Mount Scopus group, which unconformably overlies the Judea group almost everywhere within the Arad area. The total thickness of this group varies from 30 to 100 meters towards the northeast.
Three sections may be distinguished in the Mount Scopus Gr. sequence:
- a lower section (unit sc, Menuha Fm. of Santonian age) consisting mainly of chalk and marl .Some phosphorite and chert layers are identified towards the top of the formation.
- a middle section (unit ca, Mishash Fm. of Campanian age) is subdivided into two members. A lower member of massive, brecciated and laminar chert and an upper member (Phosphorite series) with phosphoritic chalk, marl, chert and porcellanite.
- an upper section (unit ma, Ghareb Fm. of Maastrichtian age.) is composed of marly oil shale in the lower part and chalk in the upper part. The Hatrurim Fm. is composed of high-temperature minerals that are the metamorphic product of the Mishash and Ghareb Fms.. The Mishash Fm. (unit ca) appears everywhere in the town of Arad and is overlain by the Hazeva Fm.
The Hazeva Fm. near the town of Arad(Shahar,1973)is divided into three members (from bottom to top): Shahaq Conglomerate, up to 2m thick; clayey marl, up to 4 m thick; and sandstone with conglomerate horizons containing Maastrichtian or Middle Eocene chalk, up to 15m thick. Some 0.3m of the sand, from the surface down, has limey cement and locally, a finely laminated hard, white gray limestone occurs.
|
Figure2. Geological map of the Arad area
|
|
|
1.3 Methods used to determine site amplification |
The site response functions are best determined from recorded ground motion during an actual strong event by comparison with recordings at a nearby reference site located on rock (Jarpe et al., 1989). In most cases, mainly in regions where the seismic activity is relatively low as in Israel, this type of analysis is usually impractical. Many investigators evaluated site response functions from moderate to weak earthquakes motion (for example, Field and Jacob, 1992; Carver and Hartzell, 1996, Zaslavsky et al., 2000). Nakamura (1989) hypothesized that site response could be estimated by dividing horizontal component noise spectra by vertical component noise spectra. Results obtained by implementing the Nakamura technique (Field and Jacob, 1995; Mucciarelli, 1998; Zaslavsky and Shapira, 2000) support such use of microtremor measurements to estimate the site response for surface deposits.
In this study, we focus on two previously cited approaches.
Horizontal-to-Vertical Noise Spectral Ratio
Nakamura (1989) proposed the hypothesis that the site response function under low strain can be determined as the spectral ratio of the horizontal versus the vertical component of motion observed at the same site. He hypothesized that the vertical component of ambient vibrations is relatively unaffected by the softer near-surface layers. Hence, the site response is the spectral ratio between the horizontal component of microseisms (Hh) and vertical component of microseism (Hv) recorded at the same location:
Rn(ω)=|Hh(ω)| / |Hv(ω)| (1)
In other words, the vertical component of the ambient vibration on the surface retains the characteristics of horizontal component at the bedrock (reference site).
Horizontal-to-Vertical S-Wave Spectral Ratio (Receiver Function)
This technique is based on Nakamura’s hypothesis for S-wave (Lermo et al., 1993):
Rn(ω)=|Ssh(ω)| / |Svs(ω)| (2)
where Ssh and Ssv , respectively, denote horizontal and vertical amplitude spectra computed at the same investigated site, from S-waves.
Receiver function was introduced by Langston (1979) to determine the velocity structure of the crust and upper mantle from teleseismically recorded P-waves. Langston made the assumption that the vertical component of motion is not influenced by local structure, whereas the horizontal components, owing to the geological layering, contain the P to S conversion. In the spectral domain this corresponds to a simple division of the horizontal spectrum by the vertical (equation 4). Most studies show that the H/V ratio obtained from ambient vibrations coincides with response functions of near surface structures to incident shear waves (Ohmachi et al., 1991; Lermo and Chavez-Garcia, 1994; Zaslavsky et al., 1995; Seekins et al., 1996; Gitterman et al., 1996; Konno and Ohmachi 1998; Mucciarelli and Monachesi, 1998; Chavez-Garcia and Cuenca, 1998; Toshinava et. al., 1997; Shapira et al., 2001, Zaslavsky et al., 2002b, 2003, 2004).
|
|
1.4 Data acquisition and procedure |
Ambient vibration measurements were carried out during the period from March to August 2004 in Dimona (W-199250; E-204725; S-551250 and N-555250) and from July to September, 2004, in Arad (W-218000; E-223000; S-572000 and N-576000). The work areas are approximately 19 km2 and 11 km2 for Dimona and Arad, respectively. The distributions of measurement points over the two towns are shown in Figures 3 and 4. The measurements points were selected to provide good coverage of the various sediments and sediment thicknesses. We planned the grid of measurement points of 250m*250m for both towns, increasing a density in the areas with high-frequency response. Ground motion (velocity time history) was recorded using the multi-channel digital seismic data acquisition system designed for site response field investigations (see Shapira and Avirav, 1995). The system includes: a multi-channel amplifier with band pass filters 0.2-25 Hz, GPS (for timing) and a laptop computer with analog-to-digital (A/D) conversion card. The seismometers (L4C) used were sensitive velocity transducers with a natural frequency of 1.0 Hz and damping at 70% of critical. The ambient vibrations motions were digitized at the ratio of 100 samples per second by a 16-bit A/D converter.
|
Figure 3. Map showing location of the measurement points and refraction lines in the Dimona area.
|
Prior to and during the measurements we checked and determined the transfer function of the round motion data, i.e., particle velocity. One vertical and two horizontal seismometers (oriented north-south and east-west) were installed at each site. These seismometers can also work at a reasonable range below the fundamental frequency, as demonstrated in Zaslavsky et al. (2003b). In that report the analysis of the horizontal-to-vertical spectral ratio shows that, for the instrumentation used, it is possible to obtain successful measurements up to 0.4 Hz. The seismometers were installed on leveled metal ground plates and connected to the data acquisition system by cables. All the equipment – sensors, power supply, amplifier, personal computer and connectors – were installed on a vehicle, which also served as a recording center.
|
Figure 4. Map showing location of the measurement points and refraction lines in the Arad area
|
At each site, the ambient vibration was recorded continuously for 90 minutes, creating data files of 3 minutes each of ambient vibration data. In Figures 5 and 6 we present examples of the locations of the seismic stations during the site investigation in the towns of Dimona and Arad. Prior to performing measurements we checked and determined the transfer function of the instrumentation in order to facilitate transformation of the record signals into true particle velocity system. These measurements also provide relative calibrations between the different channels of the entire monitoring system. Figure 7a presents seismograms and corresponding spectra of horizontal and vertical components for two sets of three-component seismometers recorded at Point 111. All horizontal seismometers were placed at the same location and in the same orientation. The amplitude spectra (Figure 7b) show that the measured motions are practically identical.
|
Figure 5. Locations of the seismometers during various sets of the site investigations in the Dimona area.
|
|
**************************
|
|
Figure 6. Locations of the seismometers during various sets of the site investigations in the Arad area.
|
|
**************************
|
|
Figure 7. Velocity seismograms and spectra of simultaneously recorded ambient noise by (a) – horizontal and (b) – vertical seismometers.
|
The signal treatment is described in the following. We selected two different time windows, consisting of 30 sec records for sites with resonance frequencies above 1 Hz and with resonance frequencies of 60 sec records for sites with resonance frequencies less than 1 Hz. The selected time windows were Fourier transformed (Perelman and Zaslavsky, 2003), using cosine-tapering (1 sec at each end) before transformation and then smoothed with a triangular moving Hanning window. The H/V spectral ratio was obtained by dividing the individual spectrum of each of the horizontal components [SNS(f) and SEW(f)] by the spectrum of the vertical component [(SV(f)]. To obtain systematic and reliable results from the spectra of microtremors, we used several time windows (60-70) that yielded a number of spectral ratios that, in turn, were averaged. We also experimented with computing the average of the spectral ratios and found the differences to be negligible.
The horizontal-to-vertical spectral ratio AH/V(f) is obtained by dividing the individual spectrum of each of the horizontal components SNS(f) and SEW(f) by the spectrum of the vertical component SV(f). The average of the two horizontal-to-vertical ratios is defined as the site amplification function:
A(f)=1/2n[Σn-1(Sns(f)i / Sv(f)i)+Σn-1(Sew(f)i / Sv(f)i)] (3)
We have consistently observed that averaging the spectral ratio arithmetically or geometrically does not significantly change the results. It is worth noting the importance of this averaging procedure, the main problem related to the selection of time windows.
|
Empirical estimations of site effects were carried out by implementing H/V spectral ratio from ambient vibrations technique (Nakamura technique). The estimated response was compared with the response inferred from the receiver function technique using explosions recorded at three measurement sites.
|
| 2.1 Examples of the H/V spectral ratios from ambient vibrations in Dimona |
Figure 8 shows individual and average H/V spectral ratios from ambient vibrations obtained at sites located on the hard rock within the limits of outcrop according to the geological map. The scatter of individual curves is high but the variations in the average functions are small. H/V method does not produce amplification peaks in the 0.2-10 Hz frequency range.
|
Figure 8. Individual and average H/V spectral ratios obtained at sites located on the hard rock outcrop.
|
Experimental spectral ratios at points 120, 121, 126 and 127, distributed within adjacent areas of the exposed hard rock but remote from the outcrop line at distances of up to 500 m, contain no evidence of amplification at frequencies of engineering interest (see Fig. 9).The lack of a well-defined resonance may be explained by smaller depth to bedrock (roughly a few meters) and their location at the transition from sediment-filled valley to bedrock highland.
|
Figure 9. Individual and average spectral ratios showing no indications of site effects for sites distant from the outcrop line.
|
Our observations revealed clear and consistent peaks of spectral ratios traced from site to site. Figure 10 demonstrates the character of average Fourier spectra at Point 253, which produce amplification peaks in the frequency range 5-10 Hz. An increase in the spectral level of the horizontal components, shown on Fig. 10a, is clear at a frequency of 10 Hz, while the spectrum of the vertical component is approximately flat. Therefore, spectral ratios (Fig. 10b) show a prominent peak at this frequency with an amplification factor of 4. We observe a similar picture at point 33 (see Fig. 10a), where the increase in horizontal components produces a spectral ratio peak at a frequency of 5.5 Hz (Fig. 10b). Alternatively, in the case of Fourier spectra for Point 71, we get a deep trough at a frequency of about 8 Hz that is responsible for an amplification peak of 2.5.
A third type of amplitude spectra is presented in Fig. 10a (point 77). The main feature of this type of spectra is the divergence of horizontal and vertical components in the frequency band from 3 up to 10 Hz. In this frequency band horizontal components considerably exceed the vertical one. Spectral ratios calculated from amplitude spectra exhibit an amplification of 4.5 at fundamental frequency 7.5 Hz.
|
Figure 10. (a) Average spectra of NS (green line), EW (blue line) horizontal and vertical (red line) components; and (b) individual and average H/V spectral ratios showing fundamental frequency in the range from 5 to 8 Hz.
|
In order to complete the list of Fourier spectra typical for the investigated area we should add another, which actually is a combination of the previously described types. The H/V ratio from such spectra is localized by an increase in horizontal and trough in the vertical spectra simultaneously as shown in Fig. 11a (point 272).
|
Figure 11. (a) Example of Fourier spectra produced by high in the horizontal and trough in the vertical components; (b) individual and average spectral ratios at Point 272
|
A distinct tendency of the fundamental frequency to decrease is observed from the edge of the investigated area toward the center. Measurement points characterized by the lowest fundamental frequencies (from 1.4 up to 2 Hz) and amplifications (less that 3) are well grouped into structures with a linear strike oriented northwest–southeast. Examples of observed horizontal-to-vertical spectral ratios show well-pronounced peaks with low amplification (see Fig. 12).
|
Figure 12. Examples of individual and average H/V spectral ratios yielding the lower fundamental frequencies and amplifications in the study area.
|
Intermediate fundamental frequency from 2 up to 5 Hz accompanied with amplifications factors of 3-6 are widely represented over the investigated area. Some examples are given in Fig. 13.
|
Figure 13. Examples of individual and average H/V spectral ratio from ambient vibrations characterized by fundamental frequency from 2 to 5 Hz.
|
|
| 2.2 Comparison between ambient vibrations and explosion data
|
One problem in the methodology that makes use of explosion data for site response estimates is related to the shear-wave generation at the source, theoretically absent for isotropic source of structure. However, at the same time among the articles published on this topic efficient generation of SH waves during explosion (Kisslinger et al., 1961; Jones et al., 1962; and many others) are described. In conclusion, during an explosion, a significant amount of energy is carried by shear waves generated at the source or at the free surface directly above it. This means that, although the characteristics of earthquake and explosion signals are different, the use of artificial source can be a useful tool in making inferences on the seismic behavior of sites during earthquake (Malagnini et al., 1996; Zaslavsy, 2003; Zaslavsky et al., 2003).
During the measurements period we recorded three quarry explosions. In Figure 14 we present an example of the seismograms, spectra and the corresponding H/V spectral ratios (receiver function) from the explosion in Har-Tuv quarry (2004-04-08; 08:34, ML=2.3, epicentral distance R=76 km) recorded at Site 87 located on outcrop of hard rock,. We can see that amplitudes of horizontal and vertical ground motions (Fig. 14a) are approximately identical. In this case the spectrum calculated from vertical component of ground motion covers the spectra of horizontal ground motion in the frequency range from 0.5 Hz to 10 Hz (Fig. 14b). Spectral ratios of the horizontal components with respect to the vertical components are plotted in the Figure 14c. As clearly seen, in the frequency range 0.5-10 Hz there is no indication of site effect.
|
Figure 14. (a) Seismograms from the explosion in Har-Tuv quarry (2004-04-08; 08:34, ML=2.3, epicentral distance R=76 km) recorded at Sites 87; (b) amplitude spectra; (c) and the corresponding H/V spectral ratios.
|
Figure 15a shows a typical set of scaled seismograms from the explosion in the Arad quarry (2004-09-25; 09:14; ML =2.7; epicentral distance R=14 km) recorded at Site 9. The seismograms plotted with the same scale show that amplitudes of horizontal components are amplified with respect to the vertical component. Moreover, whereas horizontal components of records are saturated, vertical component of record are not saturated. Peak velocity amplitudes at the horizontal components are about 3 to 4 times larger than the vertical component. As Figure 15b demonstrates a "bump" in the spectra of the horizontal components is clearly seen in the frequency range 3.0-4.0 Hz. This effect is more pronounced in the spectral ratios. Figure 15c shows spectral ratio of the two horizontal components of the explosion motions computed with respect to vertical components. Examining the shape of the H/V spectral ratios at Site 2 we can see that there is amplification factor up to 6.0 in the frequency about 4 Hz.
|
Figure 15. (a) Seismograms from the explosion in Arad quarry (2004-09-25; 09:14; ML=2.7; epicentral distance R=14 km) recorded at Sites 9; (b) amplitude spectra; (c) and the corresponding H/V spectral ratios.
|
Figure 16a shows three components of seismograms obtained from the explosion in Samaria region (2004-04-29; 6:58; ML=1.9; epicental distance 115 km) recorded at Site129. Again, amplitudes of horizontal ground motion are larger than the amplitude of vertical ground motion. Comparing the horizontal and vertical spectra obtained on this site (Figure 16b) we notice that the frequency range where the two spectra deviate are 1.5-2.5 Hz. At these frequencies, the spectral amplitudes for horizontal components are higher than the spectral amplitude for the vertical component. This feature, clearly visible at the spectral ratio (Figure 16c), relates to amplification of ground motion.
|
Figure 16. (a) Seismograms from an explosion in the Samaria region (2004-04-29; 06:58, ML=1.9, epicentral distances R=115 km) recorded at Sites 129; (b) amplitude spectra; (c) and the corresponding H/V spectral ratios.
|
The average site response functions observed at Sites 87, 9 and 129 for different data set (ambient vibrations and explosions) are shown in Figure 17. Both response functions from explosion and ambient vibrations for Site 87 (Figure 17a) in the frequency range 0.5-10 Hz show no site effects. It is clear that these two curves are almost flat and do not exhibit any significant amplification. Figure 17b shows average H/V spectral ratios obtained from explosion and ambient vibrations recorded at Site 129. In this case the fundamental frequency determined from explosion and ambient vibration are the same. The explosion and ambient vibration spectral ratios for Site 9 look similar in the frequency range from 0.5 to 10 Hz.
|
Figure 17. Comparison of different estimates of site amplification based on H/V spectral ratio techniques applied to explosions and ambient vibrations measurements. Green line is the ambient vibration spectral ratios. Red line represents the spectral ratio computed over explosion.
|
|
| 2.3 Examples of H/V spectral ratios from ambient vibration in Arad |
Fundamental resonance frequency in the town of Arad varies in the range 2-8 Hz, while amplifications are very monotonous, basically fluctuating within the range a factor of 2-3 and in some place reaching factor 4. Several examples of the spectral ratios obtained in the town of Arad are shown in Fig.18.
|
Figure 18. Example of individual and average spectral ratios obtained in the town of Arad
|
|
|
2.4.1 The town of Dimona |
In seismic microzonation we want to display the variations in seismic response of the subsurface and, subsequently, determine the soil is being amplified to a level that may damage existing buildings or other structures at that location. Results of processing 275 ambient vibration measurements carried out in the Dimona area were integrated into maps showing distributions of the fundamental frequency and the largest amplification of the soil that will occur at this frequency.
The fundamental resonant frequencies in the Dimona urban area are distributed in the range between 1.5 and 9 Hz (see map in Fig. 18). At sites located at the outcropping bedrock consisting of limestones and dolomites of the Turonian-Cenomanian age we observed flat site response with no resonance frequency. We should note, however, that an area with no site response is not limited by outcropped bedrock proper. The width of the belt with no resonant frequency outside the bedrock can reach 600 meters. A strip of higher fundamental frequency (from 5 up to 10 Hz) adjoins the area without the site effects practically along the whole boundary. Moving away from outcropped bedrock toward the axis of the syncline, which strikes from the southwest to northeast, the fundamental frequency decreases gradually down to 3 Hz. This is in accordance with the geological map of the area. The asymmetrical shape of the syncline is distinctly visible in the contours of the resonant frequency. Another deep linear structure was revealed in the analysis of ambient vibration measurements. It is indicated by the belt of lower (1-2 Hz) fundamental frequency values directed northwest–southeast. This structure may be interpreted as a channel erosion of paleorelief as is confirmed by geological information.
Figure 19 depicts maximum amplification contours within the study area. We observed significant variations in the amplification level from factor 2 up to 10. These variations could be related with the variation of the impedance between the bedrock and the overlying sediments. The highest amplifications (over factor 5) are attained at sites located in the northeastern part of the study area; probably because of relatively low S-wave velocity of the heterogeneous upper soft layer. Apparently, the amplification values exhibit a general decrease to the factor 3 to the southwestern part of the area. The smaller values (of factor less than 3) are obtained within the low-frequency belt crossing the investigated area from the northwest to southeast. It should be noted that the belt of low amplification response can be divided into two areas. The first shows amplification less than factor 2, while the second one exhibit peaks with amplitude from factor 2 to 3.
|
Figure 19. Distribution of the fundamental resonance frequency over the town of Dimona.
|
|
**************************
|
|
Figure 20. Distribution of the maximum amplification level over the town of Dimona.
|
|
| 2.4.2 The town of Arad |
The contour map presented in Fig. 20 is created on the basis of 110 ambient vibration measurements and shows variation of fundamental frequency throughout the town of Arad. The site response functions exhibit peaks at resonance frequencies between 2.5-8 Hz, which, in turn, are correlated with the sediment thickness. The higher fundamental frequencies occur in the eastern part of the study area, bordering the outcrop of limestones and dolomites of Turonian-Cenomanian age constituting the bedrock. Like as in the case of Dimona an area with flat response function is wider than rock outcrop. Toward the southwest and northwest resonance frequency of site response gradually decrease correlating with dip of the bedrock. The lower resonance frequencies (range from 2.5-3 Hz) are attained in the southeastern part of Arad. Sharp change of resonance frequency near the northeastern edge of the study area from 6 down to 3 Hz indicates the presence of fault. This finding is in agreement with geological data.
The map of maximum amplification is displayed in Fig. 21. Except the areas adjoining the exposed bedrock where site response functions do not exhibit amplification, we observed the variation of amplification level from factor of 2-3 at majority of sites. Only in the eastern part of the area are distinguished a few spots with amplification a factor of 3-4.
|
Figure 21. Distribution of the fundamental resonance frequency over the town of Arad.
|
|
**************************
|
|
Figure 22. Distribution of the maximum amplification level over the town of Arad
|
|
| 3. CONSTRUCTION OF THE SUBSURFACE STRUCTURE |
The empirical method that we implemented to evaluate the site response, i.e., H/V spectral ratios from ambient noise and explosions, reveals the fundamental mode of vibration of the soil column. In order to obtain a more general representation of the site response function, it is thus desirable to evaluate transfer function of the soil column analytically by modeling the subsurface. Modeling the observed spectral ratios with one-dimensional transfer functions requires constraint of the geological subsurface structure with a variety of geotechnical data. The layers can be characterized by their shear-wave velocity Vs, thickness H, density and damping, according to the SHAKE program (Schnabel et al., 1972) that we used in our study. Response function obtained analytically by choosing a subsurface model with parameters constrained by the available geological and geophysical information and similar in shape to the empirical evaluation of site response of a site is the site response function.
The information on the subsurface distribution of S-wave velocities and thickness of the soil layers is obtained from geophysical surveys (refraction lines) performed and processed by Ezersky (2004, reference). Five seismic refraction profiles were carried out within the study area (for location see Figure 3). The resulting depth sections are presented in Figures 22-26. The horizontal axes on the sections represent the distance in meters along the refraction line, whereas the vertical axis shows the depth in meters.
It is desirable that the information inferred from the refraction profiles will be cross-referenced with nearby borehole data. However, we do not have borehole information within the study area. Thus, in estimating the site response function analytically we relied upon our measurements constrained by the refraction surveys data and the regional geology. In the other words we used the refraction profiles for calibration of our measurements at points located along these profiles in order to obtain optimal velocity model of subsurface. The main criterion of the appropriateness of the inferred model was a better fit of the analytical estimation with empirical one. The fitting procedure was carried out using an iteration algorithm developed by Mickenberg (Zaslavsky et al., 2004). In turn, fixing the S-wave velocities enabled adjusting the thickness of sedimentary layers for sites where information on the subsurface structure is not available.
|
| 3.1 Calibration of experimental spectral ratios by refraction lines in Dimona |
The spectral ratio technique applied at measurement sites 26, 15 and 163 located at RL-2 line suggests fundamental frequencies 5.3; 4.4 and 3.7 Hz with amplifications of factor 6, 4.5 and 5 correspondingly, as seen from Fig. 22a. The seismic section of S-waves along line RL-2 is shown in Figure 22b. Analytical transfer functions for Points 26 and 15 calculated with direct use of thicknesses and velocities given in the seismic section yield fairly agreement with corresponding experimental spectral ratios. The lower layer in the section having Vs=1970 (or 2000 m/sec in our model) was assumed as a reflector represented by hard limestone and dolomite of the Judea Group. At Point 163, then we should note that there is a disagreement between P- and S-wave sections in the interpretation of thickness of the second layer. According to the P-wave section the boundary between the second and third layers is found at a depth of about 60 m vs. 90 m according to the S-wave section. This depth provides a better agreement of analytical function with empirical spectral ratio. S-wave velocity of 1650 m/sec in the northern part of the profile probably indicates the appearance of an additional layer occurring at the bottom of the second one. The thickness of this layer, according to our calculations, is no more than several meters. The presence of this high-velocity layer in the model does not significantly alter the fundamental frequency and amplification factor of the analytical transfer function. However, we will show below that this layer appears also in the RL-4 refraction line and is needed for reconstruction of the subsurface structure of the investigated area. The parameters of both the geophysical and analytical models are shown in Table 1. Density and damping values were taken from literature sources. We should note that physically reasonable variations in density have little impact on the modeled frequency at which predicted resonance peak occurs, and their effect on the predicted amplitude of the fundamental resonance peak is within the uncertainties of the observation results.
In the RL-4 seismic section the Vs of the lower layer are different for the western and eastern part, as seen from Fig. 23b. Correlating this section with the others available in the study area we revealed that layers with similar S-wave velocities were identified in RL2 refraction line. Vs= 1290 m/sec may be correlated with the second layer in RL2 profile, while Vs= 1680 with the third layer. 1-D models constructed on the basis of RL-4 data, considering mentioned intermediate layers, yield analytical transfer function with fundamental frequency and amplification factor similar to those obtained empirically at Points 269 and 216. The models are presented in Table 2. For analytical and empirical transfer functions see Fig. 23a.
The seismic section along RL-3 line is available only for P-wave (Fig. 24b). Intervals of S-wave velocity were obtained using inter-correlations between Vp and Vs in the region. Measurement points 164 and 27 located on this line reveal very similar fundamental frequency and amplification (see Fig. 24a). Thus, the model inferred for Point 164 is also suitable for Point 27. We suggested that the two upper layers appearing in the section are underlain by a layer of about 30 m thick with Vs of 1700 m/sec. The optimal model is shown in Table 3.
As demonstrated in Fig. 25a, the spectral ratios obtained at Point 162 and 161 located on refraction line RL-1, show a fundamental frequency of 2.7 Hz and 2.1 Hz with amplifications of 4.5 and 3.0 correspondingly. From the seismic section of the S-wave along RL-1 line are identified three layers of different velocities as shown in Fig. 25b. The procedure of fitting the analytical transfer function to the empirical one allowed us to adjust the thickness of the third layer above the reflector. The resulting models are shown in Table 4
The resulting depth section along line RL-5 presented in Fig. 26b differentiates three layers with different velocities. Three ambient vibration measurements were carried out along this profile. We show two of them in Fig. 26a (points 271 and 270), since point 191 is identical with 271. In order to reach a reasonable match between analytical and empirical estimations and bearing in mind the others refraction profiles we merged second and third layers with velocities 490 and 780 m/sec into one with Vs equal to 700 m/sec and underlain it with the layer with velocity 1100 m/sec, which, in turn, overlies the reflector. Both the optimal and geophysical models are presented in Table 5.
Figure 23. (a) Experimental (solid line) and analytical (dashed line) transfer function for Points 26, 15 and 163;
(b) S-wave velocity depth section along refraction line RL-2.
|
Table 1. Geophysical and analytical model for calculating transfer functions at points located along RL-2 refraction line.
|
**************************
|
|
Figure 24. (a) Experimental (solid line) and analytical (dashed line) transfer function for Points 269 and 216; (b) S-wave velocity depth section along RL-4 refraction line.
|
Table 2. Geophysical and analytical model for calculating transfer functions at points located along RL-4 refraction line.
|
**************************
|
|
Figure 25. (a) Experimental (solid line) and analytical (dashed line) transfer function for Points 164 and 27; (b) S-wave velocity depth section along RL-3 refraction line.
|
Table 3. Geophysical and analytical model for calculating transfer functions at points located along RL-3 refraction line.
|
**************************
|
|
Figure 26. (a) Experimental (solid line) and analytical (dashed line) transfer function for Points 162 and 161; (b) S-wave velocity depth section along RL-1 refraction line.
|
Table 4. Geophysical and analytical model for calculating transfer functions at points located along RL-1 refraction line.
|
**************************
|
|
Figure 27. (a) Experimental (solid line) and analytical (dashed line) transfer function for Points 271 and 270; (b) S-wave velocity depth section along RL-5-6 refraction line.
|
Table 5. Geophysical and analytical models for calculating transfer functions at points located along RL-5-6 refraction lines.
|
**************************
|
|
|
3.2.1 A-A’ south-north cross section |
The models constructed for points located along the refraction lines together with distribution of fundamental frequency and amplification level were used to estimate subsurface geological structure of the area investigated. Figure 27 depicts a simplified sketch of the geological cross sections of the area investigated constructed using results of ambient vibration analysis. For location of profile A-A' and B-B', see Figures 19 and 20. Measurement points distributed along A-A’ profile oriented southwest-northeast yield range of fundamental frequencies from 1.5 up to 5.5 Hz and amplification factors of 2-5. Point 58 located at outcropped rock we obtained flat site response with no amplification. The main feature characterizing the A-A’ profile is decreasing of fundamental frequency from the edges to the central part. This general trend can be correlated with the reflector depth that is supposed to increase in that direction in accordance with geological data. Four layers above reflector are identified along the A-A' profile. The uppermost layer detected by all refraction lines has an S-wave velocity varying from 280 up to 440 m/sec and thickness from 5 up to 26 meters. According to the geological data this layer consists of alluvium, loam and sand of the Upper Hazeva group (Calvo and Bartov, 2001). The second layer with Vs=600-800 m/sec, associated with sandstone and clay of the Upper Hazeva Gr., reaches a thickness of 50 m in the central part of the AA profile. In the northern part of the profile (Points 26, 15,163 located along RL-2 refraction line) the second layer is wedging out and replaced by underlying third layer with S-wave velocity of 1100 m/sec. This layer represented by conglomerate and coarse sandstone of the Middle Hazeva Gr., is distinguished in RL2, RL3 and RL1 refraction lines and its occurrence is probably wide-spread occurrence within the study area. Its thickness varies significantly and is strongly tied to the reflector depth. In the part of the section between points 74 and 122 this layer is eroded.
Between points 26-163 the reflector with Vs=2000 m/sec dips from a depth of 25 m down to a depth of 90 m over a distance of 250 meters. This dipping is reflected in the fundamental frequency, which decreases from 5.7 Hz (point 26) down to 3.7 Hz (Point 163). We also observe a slight decrease in the amplification level from factor 6 to 5 between points 26 and 163 connected with the contribution of the soft upper layer to the total thickness of sediments above the reflector. It is important to note that according to refraction survey interpretation, S-wave velocity in the south part of the RL-2 (Point 163) the profile is 1650 m/sec in contrast to 1970 m/sec detected in the northern part and assumed to be a reflector. It may be correlated with the appearance in the soil profile of an additional high velocity layer directly above the reflector. The presence of this layer is revealed in the western part of line RL-4 (Point 269), as shown in Fig. 23b. We suppose this layer may be correlated with micritic mudstone of the Lower Hazeva Gr. (Calvo and Bartov, 2001).We should note that in spite of the fundamental frequency and amplification level at points 74, 269, 94, and 122 being very close to the those at point 163, the subsurface structure looks significantly different. The reason for that is redistribution of the velocities and thicknesses of layers in the soil profile. It is very important to bear in mind that the site response function must be systematic with the empirical evaluations and with available information about the subsurface.
The tendency of the fundamental frequency to decrease is maintained also for Points from 81 to 161 and is connected with the increasing total thickness of sediments above the reflector. Analytical models for measurement points from 81 to 161 were correlated with refraction survey along profile RL-1 (see Fig. 25a). Taking into consideration that the thickness of the second layer with Vs of 600-800 m/sec does not change significantly we concluded that total sediment thickness decreases at the expense of the third layer with Vs=1100 m/sec.
Points 52-97 are located in the area characterized by the lowest both fundamental frequency and amplification level. Analytical models for these points were correlated with RL-5 refraction line, which shown in Fig. 26b. Total sediment thickness in the area reaches 160-170 meters, and then it begins to decrease. In the southern part of the A-A’ profile, at Point 245 located near the transition to bedrock outcrop, ambient noise measurements yield an amplification of factor 4.5 at fundamental frequency 5.5 Hz.
|
|
3.2.2 B-B’ west-east cross section |
Figure 29 depicts a geological cross-section B-B' (for location see Figs. 19 and 20) over the Dimona area oriented northwest–southeast. It is accompanied by H/V average spectral ratios observed at points located along this profile. The spectral ratios from ambient vibrations yield a resonance frequency varying in the range of 3-7 Hz with associated amplifications of factor 3.5-5.5. The B-B' profile crosses the syncline of asymmetric shape, formed by limestones and dolomites of the Top Judea Group. According to our estimation the thickness of sediments overlaying Top Judea Gr. assumed to be a reflector could reach 90 meters. In order to obtain a more detailed picture of the morphology of the hard rock basement and the structure of sediment layers above the reflector, analytical models were constructed at every point along profile B-B. Again, like as in the case of the A-A' profile, ambient vibration observations compiled with the refraction survey data were used for this aim.
Spectral ratios from the two sites on the bedrock (Points 113 and 56) are flat at frequencies from 0.2 to 10 Hz. We should say that all measurement points along the profile, with the exception of 113 and 56, are located on the layer consisting of alluvium, loam and sand. Its S-wave velocity, as indicated above, varies from 280 to 440 m/sec according to seismic refraction data. Points 77 and 57 situated near the transition from outcropped bedrock to sediment-filled structure show well-defined high frequency resonance peaks at 7.5 and 6.5 Hz correspondingly. The uppermost soft layer in the soil profile at these points has a thickness of 13m and it is underlain by layer characterized by Vs=1100 m/sec and represented by conglomerate and coarse sandstone. Similar geological situation and distribution of the frequency and amplification allow us to draw a parallel with refraction line RL2 and RL-3, stretch this layer along B-B' profile, and adjust the analytical models. For the same reason we incorporate into cross section layer with a velocity of 1700 m/sec described as micritic mudstone of the Lower Hazeva Gr. The presence of this high velocity layer with a thickness up to 15 m influence insignificantly on both fundamental frequency and amplification, and we cannot detect its thickness reliably. Thus, we show the top of this layer in the cross section by a dashed line. Only near point 94 can we identify this layer evidently by correlation with the western part of the RL-4 refraction line.
From point 123 westward we constructed a model according to the RL-4 line. Here the layer with Vs=1100 m/sec is eroded and sandstone and the clay layer (Vs=600-700 m/sec) is appearing in the section. In the section of the profile between Points 123 and 203 amplification level values vary from factor of 6 down to factor 3 in the frequency range of 3-3.5 Hz. These fluctuations are expressed in changes of total thickness of sediments from 65 (Points 94, 79) up to 90m (Point 76). Between Points 203 and 77 the layer with Vs=600-700 m/sec is wedging out and we return to the models constructed in accordance with RL2 refraction lines. Sharp decrease in the depth to the bedrock to 25m matches the increase of the fundamental frequency up to 7.5 Hz at Point 77.
|
In the case of Arad we calibrated H/V spectral ratio obtained at points located close to the Zohar West-7 (ZW-7) borehole drilled down to a depth of 88m and along the seismic refraction lines. Results of seismic survey were provided by Ezersky. The location of the borehole and seismic lines one can see in Fig. 4. The seismic section of P and S-waves along the Line2 is presented in Fig. 30.
|
Figure 30. Velocity depth section along Line 2
|
At points 83 and 84 situated at the edges of the seismic profile Line 2 we obtained very similar fundamental frequency and amplification values and that is in agreement with velocity depth section, which does not yield significant changes in thicknesses of layers.
|
Figure 31. Average spectral ratios obtained for points 83 and 84 located at seismic profile.
|
Initial parameters for developing the subsurface model at point 2 located directly at Zohar West-7 borehole are shown in Table 6. S-wave velocities used were taken from seismic refraction carried out at a distance of about 800m from the drilling site.
Table 6. Initial parameters for calculation of analytical transfer function at Zohar West-7 borehole
By means of the optimization algorithm developed by Mickenberg we found missing S-wave velocities for chalk layer that yield the best similarity between the analytical and the observed fundamental frequency and maximum amplification at point 2. Vs for reflector is the same as in the Dimona area. The optimal model derived considering measured fundamental frequency and amplification, thicknesses of layers taken from borehole information and velocities from refraction survey, is shown in Table 7.
Table 7. Optimal analytical model for Point 2 located at Zohar West-7 borehole
Optimal analytical and experimental transfer functions for point 2 are presented in Fig. 32.
|
Figure 32. Comparison between analytical transfer function (dashed line) and experimental spectral ratio (solid line) for point 2.
|
Suggested velocity structure combined with our observation and geological information was used to construct geological cross section along profile A-A', the simplified sketch of which is depicted in Fig. 33. Gradual increase of the observed fundamental frequency from 2.5 Hz up to 9.5 Hz in the direction of southwest-northeast is explained by decrease of the total sediment thickness above limestones and dolomites of the Judea Gr. Marl-chalk layer of the Ghareb Fm. identified in Zohar West-7 borehole is thinning out between the 52 and 53 points according to the geological data. The highest fundamental frequency (9.5 Hz) we obtained at point 108. Between points 108 and 107 situated at a distance of 100m each from other sharp decrease in fundamental frequency (from 9.5 Hz to 3 Hz) occurs. It indicates a strong vertical shift in the basement accompanying the fault. The presence of faults is confirmed by geological data. We should note here, that changes in subsurface structure below the A-A' profile practically do not count for amplification level.
|
| 4. PREDICTION OF SITE-SPECIFIC ACCELERATION SPECTRA AND SEISMIC ZONATION IN THE TOWNS OF DIMONA AND ARAD |
The procedure of subdividing a region into clearly distinguished sectors with the same level of hazard parameters within its borders, named seismic microzonation, is linked to engineering or land-use planning purpose. Besides these practical objectives, it is also essential in microzonation projects to perform basic scientific research in order to understand the significance of specific phenomena such as the influence of certain soil- and subsoil parameters on the potential shaking.
Very generally, the usual techniques applied in microzonation studies can be classified as direct, empirical or numerical methods. Recently numerical methods have been increasingly applied, because of their flexibility and efficiency. It should to be borne in mind, however, that such synthetic numerical methods have to be validated through real data. This validation can be obtained when a strong earthquake occurs in the region or by examination of historical events. In practical cases, empirical methods are the most commonly used. The analysis of ambient vibrations has become a very attractive method in microzonation because it can be used in areas with moderate and low seismicity but not negligible seismic risk.
Considering criteria such a proximity to the Dead Sea Transform, the historical seismicity, developing urban area and potentially unfavorable soil conditions, the towns of Dimona and Arad were identified for which microzonation studies are advisable.
Since the local acceleration data from strong earthquakes is insufficient to estimate directly the design acceleration spectrum needed to estimate the ability of buildings at a certain site to withstand seismic activity, we must therefore, resort to the use of synthetic data. For this purpose Shapira and van Eck (1993) developed the SEEH method (Stochastic Estimation of the Earthquake Hazard) which is based on the generation of synthetic seismic events using events assembled in simulated earthquake lists and local seismological characteristics such as mechanism and strength of the event, epicenter location, mechanical and dynamic characteristics of the propagation paths etc. (Shapira and Hofstetter, 1993; Shapira and Shamir, 1994; Hofstetter, 1996; Shamir et al 2001). Hence, the regional information is used to synthesizing many accelerograms for the surface of the underlying bedrock and then they are convolved with the response function of the site under investigation, to yield the expected accelerations on the free surface of that site. The final result of the application of the SEEH method is the uniform hazard, site specific acceleration spectrum for the investigated site in which the same probability of occurrence applies to each point. This design acceleration spectrum is calculated for a different probability of occurrence over different period of years with different damping ratio.
|
Preliminary seismic zonation in the town of Dimona was based on the distribution of fundamental frequency and corresponding amplification level of ground motions (for these maps see Figures 19 and 20). For many measurement sites, the analytical site response functions were calculated using SHAKE program, and using H/V spectral ratios, geological and geophysical information. Moreover, for 48 measurement sites distributed within the Dimona area and grouped into zones, the uniform hazard site-specific acceleration spectra for a probability of exceedence of 10% during an exposure time of 50 years and a damping ratio of 5% were computed. The results of our computations are shown in Figure 34.
|
Figure 34. Uniform hazard site-specific acceleration spectra for different sites within the selected zones of Dimona. Spectrum according to the Israel Building Code (PGA of 0.11g) included for reference is shown by dashed black line.
|
Comparison of the acceleration spectra curves shows that the maximum of spectral acceleration changes from 0.3g to 0.85g within the study area. Some zones characterized by different geotechnical parameters but which do not differ significantly in spectral acceleration functions were joined. Generalized transfer functions and acceleration spectra for each unit were calculated.
A map of preliminary seismic zonation for the town of Dimona is shown in Fig. 35. An overall response spectrum, which accounts for amplification effects, was adopted for each zone. The parameters of the generalized soil-column models together with transfer function and acceleration spectrum for six zones selected in the town of Dimona are presented in Table 8.
|
Figure 35. Map showing zone division in the town of Dimona.
|
|
Three zones were selected in the town of Arad on the basis of acceleration spectra calculated for 22 sites. Each zone is characterized by a fundamental resonance frequency of the soil columns without respect to the amplifications, since variations in this parameter are not significant over the whole study area. The uniform hazard site-specific acceleration spectra calculated for a probability of exceedence of 10% during an exposure time of 50 years and a damping ratio of 5% and grouped into three zones are shown in Fig. 36. The spectrum prescribed by IS-413 is shown in this figure.
|
Figure 36. Uniform Hazard Site-Specific Acceleration Spectra for different sites within the selected zones of Arad. Spectrum according to the Israel Building Code (PGA of 0.15g) included for reference is shown by black dashed line.
|
Maximum of spectral acceleration varies from 0.4g to 0.65g within the Arad area. The map of the seismic zonation in Arad is presented in Fig. 37.
|
Figure 37. Map showing zone division in the town of Arad
|
The parameters of the generalized soil-column models together with transfer function and acceleration spectrum for six zones selected in the town of Arad are shown in Table 9.
|
|
|
