1

MICROZONING OF THE EARTHQUAKE HAZARD IN ISRAEL

PROJECT 9

SITE SPECIFIC EARTHQUAKE HAZARDS ASSESSMENT USING AMBIENT NOISE MEASUREMENTS IN HADERA, PARDES HANNA, BINYAMINA AND NEIGHBORING SETTLEMENTS

November, 2009 Job No 526/473/09

Principal Investigator: Dr. Yuli Zaslavsky

Collaborators

G. Ataev, M. Gorstein, M. Kalmanovich, D. Giller, I. Dan, N. Perelman, T. Aksinenko, V. Giller, and A. Shvartsburg

Submitted to: Earth Sciences Research Administration National Ministry of Infrastructures

Contract Number: 28-17-054 2

CONTENT LIST OF FIGURES ...... 3 LIST OF TABLES ...... 4 ABSTRACT ...... 5 INTRODUCTION ...... 7 Empirical approaches implemented in the analysis of site effect ...... 8 GEOLOGICAL OUTLINE ...... 10 Quaternary sediments ...... 12 Tertiary rocks ...... 13 Cretaceous rocks ...... 13 Structural position of the Top Judea Group ...... 13 DATA ACQUISITION AND PROCESSING ...... 14 IMPROVEMENT OF NAKAMURA TECHNIQUE ...... 18 1. Embedding of time series ...... 19 2. Decomposition of times series ...... 20 3. Grouping ...... 20 4. De-embedding (diagonal averaging) ...... 20 RESULTS ...... 23 Types of H/V ratios and developing of S-velocity model ...... 23 Profile A-A ...... 36 Profile B-B ...... 41 ...... 44 Profile C-C ...... 44 Distribution of H/V frequency and its associated amplitude for the fundamental and second resonance peaks ...... 47 SEISMIC MICROZONATION IN TERMS OF UNIFORM HAZARD ACCELERATION SPECTRA ...... 51 CONCLUSIONS ...... 61 ACKNOWLEDGEMENT ...... 63 APPENDIX A. WELL DATA IN THE STUDY AREA ...... 68 3

LIST OF FIGURES

Figure 1. The geological map of the study area and main geological structures. Location of boreholes used as examples and profiles for reconstructing cross sections...... 11 Figure 2. Topographical map of the study area and locations of the observation sites...... 17 Figure 3. H/V spectral ratios ± one standard deviation (shaded area) of automatic (a) and manual (b) selected time window...... 18 Figure 4. Times series of ambient noise recorded at site 357- (a); after applied SSA method with m=10 (embedding dimension) and q=1 (singularly value) – (b) and m=10, q=2 – (c)...... 21 Figure 5. Individual H/V spectral ratio obtained from ambient noise – (a); from time series after applying SSA method with: m=10 (embedding dimension) and q=1 (singularly value) – (b); and m=10, q=2 – (c). The analytical transfer function of soil column model is ...... 22 Figure 6. Individual H/V spectral ratio obtained from another time window of ambient noise (a); from time series after applying SSA method (for parameters see Figure 5)- (b) and (c)...... 23 Figure 7. Examples of V and H average Fourier spectra and corresponding H/V spectral ratios at sites whose surfaces (a) of calcareous sandstone of Kurkar Gr. and (b) of marl-chalk of Eocene- Senonian age create of the second frequency. frequency is related to Judea Gr. The shaded areas on the Fourier spectra show frequency ranges of resonance motion...... 24 Figure 8. Examples of V and H average Fourier spectra and corresponding H/V spectral ratios at sites whose surfaces (a) of calcareous sandstone of Kurkar Gr. and (b, c, d) of the Cenomanian chalk complex create of the second frequency. The first resonance frequency is related to the dolomite of Cenomanian-Albian age...... 26 Figure 9. Examples of V and H average Fourier spectra and corresponding H/V spectral ratios at sites whose surface of calcareous sandstone of Kurkar Gr. creates the resonance frequency...... 27 Figure 10. Examples of V and H average Fourier spectra and corresponding H/V spectral ratios at sites whose surfaces (a) of Judea Gr. and (b, c) of dolomite of Cenoman-Albian age create the resonance frequency...... 27 Figure 11. a) Lithostratigraphic section of the Binyamina well; b) velocity-depth section along refraction profile RL-1 ; c) comparison between H/V spectrum obtained at point 83 (red line) and the analytical transfer function computed using well data and velocities from refraction line RL-1 (black line)...... 29 Figure 12. H/V spectral ratio (red line) and the analytical transfer function ( black line) for point 85 in the Carmel coast...... 30 Figure 13. H/V spectral ratio (red line) and the analytical transfer function (black line) for point 311 (Atlit 2 well) in the Carmel coast...... 31 Figure 14. a) Lithostratigraphic section of the Wadi Ara 1 well; b) comparison between H/V spectrum obtained at point 287 (red line) and the analytical transfer functions (black and blue lines)...... 33 Figure 15. a) Lithostratigraphic section of the Karkur 3 well; b) comparison between H/V spectrum obtained at point 460 (red line) and the analytical transfer function (black line) computed using data from Table 4...... 33 Figure 16. a) Lithostratigraphic section of the Menascheh 1 well; b) comparison between H/V spectrum obtained at point 462 (red line) and the analytical transfer function (black line) computed using data from Table 4...... 34 4

Figure 17. Comparison between H/V spectral ratios (red lines) obtained (a) at point 132 (Caesarea 3 well) and (b) at point 131(Caesarea 2 well) and the analytical transfer functions (black lines) computed using data from Table 5...... 35 Figure 18. Schematic geological cross section along profile AA reconstructed on the base of H/V analysis ...... 39 Figure 19. H/V spectral ratios (black lines) and analytical transfer function (black dashed lines) for sites along profile A...... 40 Figure 20. Schematic geological cross section along profile B-B reconstructed on the base of H/V analysis ...... 43 Figure 21. H/V spectral ratios (solid line) compared with analytical transfer function (dashed line) for representative sites of profile B-B...... 44 Figure 22. Schematic geological cross section along profile C-C reconstructed on the base of H/V analysis ...... 46 Figure 23. H/V spectral ratios (black lines) and analytical transfer function (black dashed lines) for sites along profile C...... 47 Figure 24. Distribution of the H/V fundamental frequency and its associated amplitude over the study area. The red line indicates position of the geological cross section...... 49 Figure 25. Distribution of the H/V second resonance frequency and its associated amplitude. The red line indicates position of the geological cross section...... 50 Figure 26. Seismic microzonation map of the study area ...... 52

LIST OF TABLES

Table 1. Geotechnical data and soil column model for point 83 (Binyamina well) ...... 28 Table 2. S-wave velocity model based on the refraction line RL-2 and suggested soil column model for point 85 in the Carmel coast...... 30 Table 3. Geotechnical data and soil column model for point 311 (Atlit 2 well) in the Carmel coast...... 31 Table 4. Geotechnical data and soil column models for points 462 (Menascheh1 well), 460 (Karkur 3 well) and 287 (Wadi Ara 1 well)...... 32 Table 5. Geotechnical data and soil column models for points 131 (Caesarea 2 well) and 132 (Caesarea 3 well)...... 35 Table 6. S-wave velocity model for the study are...... 36 Table 7. Generalized soil column models for calculation of acceleration response spectra of the zones ...... 53

5

ABSTRACT

Ground motion amplifications due to soft soils, common in urban areas, are a major contributor to increasing damage and number of casualties. In this study, we present the next stage of the overall project “Microzoning of the seismic hazard in Israel” launched in 2001. To map site response characteristics across the study area of 280 km2 including the towns of Hadera, Pardes Hanna, Or Aqiva, Binyamina, Atlit, Tirat Carmel and Haifa, 375 ambient noise measurements have been carried out on different grid scales. Majority of measuring sites were spatially distributed each 500 meters. High variations in the observations led us to increase the density to a grid spacing of 250 m and in some sites even 150 m. Analysis of measurement results over the study area shows that horizontal-to-vertical (H/V) spectral ratios of ambient noise, which yield two resonance peaks at majority of sites, are categorized by shape considering both peaks and correlated with the geological features. Frequencies of two peaks are related to resonances of deep and shallow structures. Spatial variations of the frequency (0.27-8 Hz for the fundamental peak and 0.7-10 Hz for the second resonance peak) and H/V amplitude level (2-10 for both the fundamental and second resonance peaks) reflecting the geological complexity, are shown in four distribution maps. In the first approximation the fundamental frequency has general trend to increase toward the east in agreement with depth of the Judea Gr., which is fundamental reflector in the greater part of the study area from Hadera to Binyamina excluding the southwestern part of the study area where the Judea Gr. is dipping to a depth of more 800-900 meters and the shallow reflector (calcareous sandstone of the Kurkar Gr.) produces the fundamental resonance peak. To the north of Binyamina, at the Carmel Coast, while the Turonian-Cenomanian complexes are mostly eroded, the fundamental reflector is represented basically by dolomites of the Yagur Fm. Different geological structure is clearly reflected in both fundamental frequency and amplification maps. The shallow reflector varies over the study area. It is calcareous sandstone of the Kurkar Gr., clay-marl of the Yafo Fm., Eocene-Senonian marl-chalk or Cenomanian chalk in case of the Carmel coast. Data from representative boreholes and two refraction profiles integrated with H/V observations at corresponding locations are used to develop models of the subsurface at the measurements sites and obtain S-velocity model. We note that for the first time, shear-wave 6 velocities for Cenomanian chalk and dolomite are obtained from the refraction survey in the Carmel Coast. After testing at many wells, we concluded that S-wave velocity ranges used in the previous studies in Hashefela region are appropriate to the present study area. A fair agreement between depths of the top Judea Gr. according to the structural map and fundamental reflector is observed at borehole locations. Away from borehole in the central part of the study area and especially in the Carmel Coast difference in the depth estimations is very considerable. Results of subsurface modeling are illustrated by geological three cross sections. By comparison of the Uniform Hazard Acceleration Spectra calculated for 200 selected sites and considering the subsurface models constructed across the investigated area, we divided the area into 23 zones. Each zone is characterized by a generalized seismic hazard function representative the sites within that zone. For many zones the Israel Standard (IS-413) underestimates the acceleration in the period range 0.2-1 sec. 7

INTRODUCTION Examples of destructive earthquakes have clearly shown that local site conditions are an important factor in determining the seismic hazard specific to a given site. The fundamental phenomenon responsible for amplification of motion over soft sediments is the trapping of seismic waves due to the impedance contrast between sediments and the underlying bedrock. When the structure is horizontal layered (usually is referred to in the following as 1-D structures), this trapping effects only body waves travelling up and down to in the surface layers. When the surface sediments form a 2-D or 3-D structure, i.e., when lateral heterogeneities such as thickness variations are present, this trapping also affects the surface waves which develop on these heterogeneities, and thus reverberate back and forth. The interference between these trapped waves leads to resonance patterns, the shape and the frequency of which related with the geometrical and mechanical characteristics of the structure. Since 2001 the seismic hazard assessment (or seismic microzoning) studies have been carried out in the towns of Lod, Ramle, Qiryat Shemona, Kefar Sava, Petah Tiqwa, Hod Hasharon, Rosh Haayin, Dimona, Arad, Bet Shean, Haifa, Tiberias and other (see Zaslavsky et al., 2001-2009). In these studies we successfully applied the procedure developed by Shapira and van Eck (1993) to assess the site specific uniform hazard acceleration response. That procedure which we term SEEH (Stochastic Estimation of the Earthquake Hazard) is based on the stochastic method developed and used by Boore (1983), Boore and Atkinson (1987), Boore and Joyner (1991) among others. The process of hazard assessment involves: detailed mapping of the fundamental and other natural frequencies and amplitudes of H/V spectral ratios; compiling geological, geophysical and borehole data and integrating it with H/V observations to develop models for the subsurface of many sites across the study area. The subsurface model serves as input for computing the expected Uniform Hazard Site-Specific Acceleration Response Spectra (Shapira and van Eck, 1993) at the investigated sites. Lacking sufficient number of locally recorded strong ground motions, site specific hazard estimations are based on stochastic simulations. We applied the SEEH procedure (Stochastic Estimation of the Earthquake Hazard) developed by Shapira and van Eck (1993). This procedure produces a number of synthetic earthquake catalogues that represent the possible future seismic activity within 150 km of the investigated site. For each of the earthquake in a catalogue, SEEH implements the stochastic simulation method to generate 8 synthetic ground motions (accelerograms) expected to occur on the free surface of the studies site. The SEEH also incorporates the uncertainties associated with almost every parameter needed in the computations. At the final stage of the simulations, the synthetic horizontal accelerations propagate to the surface of the site through the soil layers constituting the site‟s sub-surface. The SEEH procedure uses a soil column model of the subsurface to compute the convolved effect of the specific site. The subsurface models for those sites are derived by integrating available geological, geophysical and borehole information relevant to the site and by conducting seismological surveys where we applied empirical techniques for site response estimating. The assembly of these synthetic accelerograms is used to predict spectral accelerations and the unified probability response spectra -10% exceedance in 50 years – for structures with 5% damping.

Empirical approaches implemented in the analysis of site effect

Site response functions are, by definition, equivalent to the spectral ratio with respect to a reference site, located on rock, when the data are strong ground motions. In regions where the seismic activity is relatively low, as in Israel, this type of analysis is usually impractical. The critical point of the reference site technique is that the choice of reference motion represents the true input motion to the soil site. According to Steidl (1993) when possible, the reference ground motion should be calculated by averaging several rock sites. Steidl et al., (1966) and Zaslavsky et al., (2002a) show that the assumption of flat response rock sites is often false, mainly due to weathering. Use these surface-rock sites with flat response as reference sites often lead to underestimation of the amplification by factor of 2 to 4 in frequency range 2 to 7 Hz. Furthermore, attempts to derive site response estimations from simultaneous recordings on sediments and on hard rock in an urban area may not be possible. Lermo and Ghavez-Garcia (1993) drew significant results from a non-reference technique, which is the receiver function technique, i.e., using the horizontal-to-vertical spectral ratios of shear-waves. Many studies report that the frequency dependence of site response can, thus, be obtained from measurements made at only one station at the analyzed site (Lermo and Chavez-Garcia 1994; Theodulidis et al., 1996; Seekins, et al., 1996; Malagnini et al., 1996, Zaslavsky et al., 1995 and many others). The implementation of this approach, however, still requires the rather frequent occurrence of earthquakes. 9

Spectral analysis of ambient vibrations is an alternative tool to quantify site effects. The idea of evaluating site characteristics from ambient noise records originated from the pioneer work of Kanai and Tanaka (1961). They pointed out that the predominant frequency of horizontal spectra of ambient noise is related to shallow, local geological conditions. Since then it has been reported that this technique has proved to be effective in estimating fundamental frequencies. However, in most cases, due to the influence of artificial sources from dense population, heavy traffic and industrial activities, resonance frequency cannot be directly identified in the ambient noise spectra (Zaslavsky et al., 2001). Kagami et al., (1982) proposed that the ratio of the horizontal components of the velocity spectra at the sediment site to those at the rock site could be used as a measure of microseism ground motion amplification. This technique is widely used for site response estimations (Rovelli et al., 1991; Field et al., 1990, 1992; Hough et al., 1990; Malagnini et al., 1996; Gutierrez and Singh, 1992; Dravinski et al 1995; Gaul et al., 1995; Zaslavsky et al., 1995, 2000; Shapira et al., 2001). Our experiments (Zaslavsky at al., 2002a) show that bedrock ground motion can be considered a good reference site with distances as small as 0.5-1.0 km from the soil site. Nakamura (1989) proposed the hypothesis that site response function under low strain can be determined as the spectral ratio of the horizontal versus the vertical component (H/V) of motion observed at the same site. He hypothesized that the vertical component of ambient noise is relatively unaffected by the unconsolidated near-surface layers. Hence, the site response is the spectral ratio between the horizontal component of microseisms and vertical component of microseisms recorded at the same location. In analogy to the other approaches, the vertical component of the surface motions retains the characteristics of horizontal components of ambient noise at depth on the bedrock (reference site). Many authors, among them Lermo and Chávez- García (1994), Seekins et al. (1996), Toshinawa et al. (1997), Chávez-García and Cuenca (1998), Enomoto et al. (2000), Shapira et al. (2001), Mucciarelli and Gallipoli (2004), Murphy and Eaton (2005), Maresca, (2006), show that the H/V spectral ratio technique can be a useful tool for the assessment of ground motion characteristics on soft sediments. The Nakamura approach has gained great interest, primarily due to the simplicity in its implementation. In this study, in order to evaluate empirically potential enhanced ground motion and assess seismic hazard in Hadera, Pardes Hana, Binyamina and neighboring settlements we apply spectral ratio from ambient noise technique. 10

In particular, the study focuses on the following objectives:  Improvement of the Nakamura‟s technique

 Empirical evaluation of the fundamental and other resonance frequencies and their associated amplitude levels obtained from ambient noise measurements by dividing of horizontal components by the vertical components of the Fourier spectra.  Producing maps of the distribution of frequencies and amplitudes.  Evaluation of the geotechnical characteristics (shear-wave velocity and thickness of sediments) for one-dimensional analysis of site effects by detailed comparison between the analytical and experimental site response functions (stochastic optimization algorithm).  Constructing geological cross sections and locating faults.  Extrapolating the derived theoretical models over the study area and integrating them into computing the uniform hazard site specific acceleration response spectra for 10% probability during an exposure time of 50 years and damping ratio of 5%;  Dividing the study area into zones based on the comparative analysis of the acceleration spectra.

GEOLOGICAL OUTLINE

The investigated area is situated along the Coastal Plain extending from Hadera to Haifa, about 47 km long and almost 10-12 km wide in the south and 3-4 km along the Carmel Mount (see Fig.1). The geological data of the region are collected from Gvirtzman (1965, 1969, 1970 and 1984), Fleisher and Gafsou (1993, 2000 and 2003) and the geological map of Israel to a scale of 1:200,000 (Sneh et al., 1998). Geothechnical information is gathered from 66 structural and oil wells from the database of the Geophysical Institute of Israel (see App. A) and about 200 boreholes from the Atlas of geological cross-sections (Ecker, 1999). The subject of the geological investigation are the hard carbonates of the Judea group (Cretaceous age), constituting the fundamental reflector in the study area, and the sediment cover rock overlying the carbonates (Tertiary and Quaternary ages). 11

Figure 1. The geological map of the study area and main geological structures. Location of boreholes used as examples, refraction lines (RL-1, 2) and profiles for reconstructing cross sections.

12

Quaternary sediments

The Quaternary sediments outcropping in the investigated area are represented by alluvium, dune sands, hamra and kurkar:  Sand dunes of Holocene age are exposed along the coastline as a strip up to 5km wide in Pardes Hanna, covering 25-40% of this area. Sand dunes usually overlay some sandstone and sometimes alluvium or loam. The thickness of these sediments varies from 2 to 15m.  Alluvium sediments of Holocene age are composed of sand, soil, gravel, clay and loess and developed along river beds ( Nahal Hadera, Barqan, Taninim and sea shore along Carmel) with thickness varies from 5m to30m, sometimes reaching 100m (Binyamina-Pardes Hanna).  Kurkar Group of Pleistocene age consists of marine and eolian calcareous sandstones named “kurkar”, some reddish silty-clayey “hamra”, silts, clays, loose sands, loam and conglomerates. The thickness of the Kurkar Gr decreases from about 180-200m near the shoreline to 0m in the eastern part of area. These rocks crop out along the coastline as a narrow strip up to 5 km wide, covering the Coastal Plain and the Israel Continental Shelf representing 5-7% of the area investigated. Two different representative provinces the west and east can be recognized in the Kurkar Gr. The western province is represented by calcareous sandstones (“kurkar”) with intercalation of clayey-silty sandstones (mostly “hamra”) and finer grained sediments. This unit covers the Coastal Plain and Continental Shelf. The eastern province is subdivided into three sub-units including the Rehovot Fm. (hamras, sands eolianes and some shales); Ahuzam Fm. (conglomerates); and Pleshet Fm. (marine calcareous sandstones). The Kurkar group unconformably overlays the clays of the Yafo Fm. of Pliocene age (in the west) or the Judea, Mt. Scopus, Avedat groups (in the east). The geological structure in the area between Binyamina and Haifa (the Carmel Coast) is considerably different. Here, the Kurkar Gr. directly overlies the Turonian-Cenomanian-Albian carbonates of the Judea Gr. Here it is only 50m thick near the coastline, wedging out eastward at the foot of the Carmel, where the Judea crops out. 13

Tertiary rocks

The Saqiye Gr. overlies unconformably the Avedat and Mt.Scopus groups and consists of deep marine marls of the Bet Guvrin Fm. and limestone of the Lakhish Fm. (Oligocene age). The sedimentary rocks of Pliocene age are represented by homogenous clay and marly clay of the transgressional Yafo formation. The sedimentation of the Yafo Fm. was dominated by a westward tilting of about two degree, which produced an increase in thickness from a few meters in the eastern area to more than 586m in the Hadera –Ramot Menashe syncline (well SH-CS-2).  The Avedat Gr. of Eocene age consists of massive soft silicified chalky limestone and marl from 0m up to 387m thick (Hadera-1well) related to the Adulam Fm. In the Coastal Plain area, the Avedat Gr. occasionally overlies the Albian Talme Yafe Fm., the Judea Group and Senonian Mt. Scopus rocks.  Mount Scopus Gr. is represented by marl-chalky facies of the „En Zetim and Ghareb Fms. and the limonitic shale of the Taqiye Fm. The thickness of the Mt. Scopus Gr. varies from 0 to 250m. It outcrops in the northeastern part of the study area, at the upper Hadera river.

Cretaceous rocks

The Judea group sequence has been divided into three parts consisting of massive dolomites and limestones of the Yagur Fm. (Albian-Cenomanian age); dolomites with interbedded marls of the Negba Fm. (Cenomanian age) and limestone, dolomites and marls of the Bina and Dalya Fms. (Turonian age). At the Carmel coast the Carmel chalk complex was deposited in an outher shelf environment. A hiatus, probably spanning the Upper Cenomanian to Lower Turonian, terminates the Cenomanian Carmel chalk complex. The upper part of Judea Gr. represented generally by the Bina Fm., consists of hard white to gray limestone and dolomite; containing rudist and coral fragments with a thickness varying from 20 to 160 m. It crops out along the foothills. The total thickness of Judea Gr. reaches 1092 m in the well Caesarea-3. The upper contact of the Bina Fm. is unconformably overlain almost everywhere by the „En Zetim Fm.

Structural position of the Top Judea Group

The study area, which occupies the Coastal Plain, Sharon and Carmel Mnt. regions, includes two major anticlinal trends: the Caesarea-Mnt. Carmel and Umm El Fahm ranges. The 14

Carmel anticline is the most westerly located structure in the northern Coastal Plain. The subordinate structures of the Carmel anticlines trend in a NE-SW direction with an asymmetrical, steep flexure on the southeastern flank. These are separated by the deep asymmetrical syncline of the Hadera –Ramot Menashe area. It descends to the area of the Hadera-1 well to an estimated depth of 1300 m below sea level. A pronounced west-oriented, faulted monocline is observed in the southern area. The Or Aqiva graben is situated between the Caesarea and Carmel structures. The graben of Neogene age is actively faulted. The vertical displacement of the Benymina fault is 500m. The top Judea structural map of the area depicts a complex structural configuration generated by a compressional regime of the Syrian Arc System, later superimposed by Neogene- Recent tectonic movements (Fleischer and Gafsou, 2003).

DATA ACQUISITION AND PROCESSING

To obtain instrumental observations of the areal distribution of ground motion characteristics within the study area, 375 ambient noise measurements were carried out during the period from January to September 2009. The work area of about 280 km2 includes the towns of Hadera (77000 inhabitants), Pardes Hana (about 30000), Binyamina (more than 10000 inhabitants) and adjoining settlements, and also the Carmel Coast (from Byniamina to Haifa). Normally spatial distribution of measuring sites about 500m enables enough resolution in the spectral ratios to identify correlation with geological features in a basin to depth of several hundred meters. In case of significant lateral variations of the results, density of the grid point spacing is increased to 250 m. Many measurements were made either at or close to the borehole sites. Site effect measurements, which were carried out during the Coastal plain and Hashefela projects (Zaslavsky et al., 2002 and 2003), are taken into account in both stages planning measuring sites and analysis of results (for locations see Fig. 9) . Ambient noise measurements are conducted using portable instruments (Shapira and Avirav, 1995) consisting of a multi channel amplifier, Global Positioning System (GPS) for timing and a laptop computer with 16-bit analogue-to-digital conversion card to digitize and store the data. In our experimental set-up, each seismograph station consists of three (one vertical and two horizontal) L4C velocity transducers (Mark Products) with a natural frequency of 1.0 Hz and damping ratio 70% of critical. The recorded signals are sampled at 100 samples per second and band-pass filtered between 0.2 Hz and 25 Hz. All the equipment: sensors, power supply, 15 amplifiers, personal computer and connectors are carried in a vehicle, which also serves as a recording centre. The seismometers are fixed on levelled metal plate placed directly on the ground. Prior to performing measurements, the individual seismometer constants (natural frequency, damping and motor constant) are determined using sine and step calibration signals, and then the frequency response functions of all channels are computed. This procedure allows evaluating change of natural frequency and motor constant (voltage sensitivity) during long time of measurements in harsh conditions in the free field. In the final instrumental test, all seismometers are placed at the same location and in the same orientation to record the same waves. It so happens that the differences between the seismic channels are marginal even without “correcting” for the instrumentation response. In fact, at frequencies below the natural frequency of the seismometers, correcting for the instrument response may increase variability and scattering. It is necessary to remind that all channels (horizontal and vertical) have practically the same frequency response function and amplifiers are set to the same gain level. Hence, spectral ratio may be calculated on the recorded signals. Moreover, it is possible to assess the predominant frequencies in the H/V spectra also at frequencies much lower than the natural frequency of the seismometer. To provide a good coverage of the study area, we design the measurement sites with a grid spacing of 500 m. Different surface sedimentary deposits, thickness of sediments and shear wave velocity contrast between sediments and bedrock are also considered in the design stage. High variation in the observations can lead us to increase the spatial density to a grid spacing of 250 m and in some locations 150-100 m. This dense network of monitored sites reflects (a) the high variability of the subsurface properties in the urban areas we study and (b) the relatively high sensitivity of H/V spectra to variations at depth. To study the characteristics of spectra of the ambient noise signals, we compute Fourier spectra and spectral ratios. The record length (time window) used for spectral calculations depends on the fundamental frequency. The basic criterion is to choose the minimal time window which yields spectra that practically do change when increasing the record length. We have concluded that at sites with fundamental frequencies of 1 Hz (or more) we should use a record of at least 30 sec. At sites with lower frequencies, the time window should be increased to 60 sec. The selected time windows are Fourier transformed, using cosine-tapering (1 sec at each end) before transformation and then smoothed with a triangular moving Hanning window. More 16 precisely, we apply “window closing” procedure (see Jenkins and Watts, 1968) for smart smoothing of spectral estimates so that any significant spectral peaks are not distorted. The H/V spectral ratios are 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)]:

S NS  f  S EW  f  ANS  f   AEW  f  (1) SV  f  SV  f 

The average spectral ratio for each of two horizontal components is computed, if the curves of average spectral ratios of the two components are similar then the average of the two horizontal-to-vertical ratios is defined as:

1  n S  f  n S  f   A f     NS i   EW i  (2) 2n  S  f  S  f   i  1 V i i  1 V i 

From comparison between the average H/V curves obtained from ambient noise recordings of different durations we conclude that recording for about 1-2 hours provides reliable calculation of the H/V function. A set of 25 - 50 selected time windows, for each site provide an H/V spectral function. Computations are made using program SEISPECT (Perelman and Zaslavsky, 2001), which was specially developed for routine analysis and site investigations. The SEISPECT is a MATLAB based application for spectral analysis and processing of ground motion recorded by a variety of seismic instruments. The main modules realized in the program are visualizing and editing the input data, selecting time window and computing Fourier spectra and H/V spectral ratios, saving and displaying results.

17

Figure 2. Topographical map of the study area and locations of the observation sites.

18

IMPROVEMENT OF NAKAMURA TECHNIQUE

As already observed by many researchers, there is high scatter in the H/V spectra. The source of the scatter is debated between the researchers. Mucciarelli (1998), for example, claims that traffic is not a major reason for the scatter and Horice et al. (2001) used noise originated by passing traffic in their analysis. In a recent study, Parolai and Galiana-Merino (2006) showed that the influence of transients on the H/V spectral ratio is insignificant. Our observations indicate that the effect of transients is almost unnoticeable. In order to reduce the scatter and increase stability, our processing scheme involves a careful manual selection of the time windows from which we obtain the H/V functions. In selecting the time windows, the analysts follow the concept that sites with no site effect should exhibit spectra of the H and V components that are of the same level throughout the spectrum. At sites with significant site effect, the spectra of the two components should differ only within a certain limited frequency band, probably at the neighbourhood of the resonance frequency. Time windows with spectra that exhibit such or similar conditions are selected. Evidently, this practice has yielded an appreciated reduction in the H/V scatter. An example of the spectral ratio functions, obtained at sites in the town of Kefar Sava (Zaslavsky et al., 2009), from automatic and careful manual selection is shown in Figure 3.

Figure 3. H/V spectral ratios ± one standard deviation (shaded area) of automatic (a) and manual (b) selected time window. 19

In this work we also investigated the application of the Singular Spectrum Analysis to improve Nakamura's method. The SSA has a wide and multidisciplinary range of applications (Golyandina,et al., 2001; Cornel et al., 2006; Rodrigues, 2007). It allows a time series to be decomposed into different components, e.g. the signal itself, as well as various noise components, which can be subsequently removed from the data. Removal of the minor component of the data can lead to significant improvements in the identification of the system. In this report the use of SSA Technique is used to optimize the signal to noise ratio before computing the classical Nakamura spectral ratios. The SSA method or spectral decomposition of matrix has only recently been applied to time series analysis. In essence; the data are embedded in high-dimensional reconstruction embedding. Chosen a sufficiently high embedding space dimension m, the SSA allows decompose a time series s(t) to sum of decreasing importance time series pci(t), the so called

Principal Components pci(t):

m s(t)   pci (t)  SSA(s(t);m,m) i1 By this technique, one can perform de-noising at various levels simply considering the first q most important Principal Components:

q sq (t)   pci (t)  SSA(s(t);m,q) i1 The parameter q is called singularly value. Typically, the best results of de-noising are obtained with a high embedding dimension m and low number of principal components.

The SSA performs four steps: 1) the embedding step; 2) the decomposition step; 3) the re- composition step and 4) the de-embedding step.

1. Embedding of time series

At this step, the one-dimensional series is represented as a multidimensional series whose dimension is called the window length. The multidimensional time series (which is a sequence of vectors) forms the trajectory matrix (1). The sole parameter of this step is the window length 20

If we consider F = (f1, f2, …, fn) the time series of length n, and m, 1 < m < n , the window length, we obtain K = n –m + 1 lagged vectors of length m,

Xj = (fj, fj + 1, …fj +m - 1)‟, j = 1, 2, …, K and the trajectory matrix

X = [X1: … XK]‟ (1) 2. Decomposition of times series

This is singular value decomposition (SVD) step. This step is the singular value decomposition of the trajectory matrix X into a sum of rank-one bi-dimensional matrices. The first two steps together are considered as the decomposition stage of the method SSA.

The next two steps form the reconstruction stage. The point of this stage is the reconstruction of the original time series as the sum of principal components using the diagonal averaging procedure.

3. Grouping

The aim of this stage is to separate the additive components of the times series. It can be seen as separating times series into two groups: “original signal” and the “noisy” components. The idea is project the trajectory matrix over a q-dimensional space. The grouping step corresponds to splitting the matrices, computed at the SVD step, into several groups (m- the intended number of principal components) and summing the matrices within each group. The result of the step is a representation of the trajectory matrix as a sum of several resultant matrices.

4. De-embedding (diagonal averaging)

At this step only the q most important principal components are taken into account, while the least important ones are discarded. This allows achieving a de-noising of the signal. For this action must be performing transfers each resultant matrix into a time series, which is an additive component of the initial series. The corresponding operation is called diagonal averaging. It is a linear operation and maps the trajectory matrix of the initial series itself. In this way we obtain a decomposition of the initial series into several addictive components. 21

If we apply this procedure to the matrix XIk, where Ik = {ik1 , … ikp} is set of indexes, we obtain, series of component and therefore the initial series F = (f1, …fn) is decomposed into sum of m series:

m F   F (k ) , k = 1, …, K. (3) k 1 where K = n –m + 1

This m series represent the first m principal components and the equality in (3) only happens if m = n, i.e. this is sum of all series/principal components.

Figure 2a show a time series of ambient noise recorded at site 357 and two time series (Figures 4b,c) after applying the SSA method. We can see that amplitude and frequency content of the time series significantly differ from the original ones.

Figure 4. Times series of ambient noise recorded at site 357- (a); after applied SSA method with m=10 (embedding dimension) and q=1 (singularly value) – (b) and m=10, q=2 – (c). 22

Examples of individual spectra ratios for two different time windows are displayed in Figures 5 and 6. The main feature of spectral ratios is the two clear peaks appearing at frequencies 0.6 Hz and 2.0 Hz. Our investigations show that such an observation is associated with two impedance contrast; one of deeper and other of shallow strata. Figure 3 and 4 show that minor changes in parameters of the SSA method lead to significant changes in shape of the H/V ratios. Thus, as a result of changing parameter q from 1 to 2 (see Figure 5 and 6); the fundamental peak at frequency 0.6 Hz just disappeared. After numerous attempts to apply SSA method to ambient noise processing we concluded that results obtained after manual selecting time windows are considerably reliable. We think key problem of the applicability the SSA method in the Nakamura technique is impossibility to decompose original signal and noise.

Figure 5. Individual H/V spectral ratio obtained from ambient noise – (a); from time series after applying SSA method with: m=10 (embedding dimension) and q=1 (singularly value) – (b); and m=10, q=2 – (c). The analytical transfer function of soil column model is 23

Figure 6. Individual H/V spectral ratio obtained from another time window of ambient noise (a); from time series after applying SSA method (for parameters see Figure 5)- (b) and (c).

RESULTS

Types of H/V ratios and developing of S-velocity model

Our previous investigations (Zaslavsky at al., 2002-2006) in complete agreement to geological data show: the surface of hard base rock, that may produce the site response and here termed as main deep reflector, is the top of limestone and dolomite of Late Cretaceous age (Judea Gr.). There is correct only for the southern part of the studied area - Hadera-Binyamina. In the north of the study area – the Carmel coast zone – where limestone of Turonian age is eroded 24 at the most part of territory and deposits of Senomanian age are represented partly by chalky sequence (the Cenomanian chalk complex) and partly by dolomite, main deep reflector presents basically as dolomite of Albian age (Yagur Fm.) and partly as dolomite of Upper Cenomanian age and partly as limestone of Turonian age, that have been supposed by complicated blocks structure (Kafri and Ecker, 1964). Fourier spectra parameters and spectral ratios analysis over the study areas show good correlation with variations in the subsurface geology. The variation of average Fourier spectra and corresponding H/V spectral ratios that were obtained in different geological conditions are shown in Fig.7-10. We observe spectral ratios both with single frequency and with two frequencies and even with three peaks of frequency. Spectral ratios with two frequencies are distributed in the greater part of the study area. The Fourier spectra are characterized by two frequency ranges of resonance motion (Fig.7-8). a) b)

Figure 7. Examples of V and H average Fourier spectra and corresponding H/V spectral ratios at sites whose surfaces (a) of calcareous sandstone of Kurkar Gr. and (b) of marl-chalk of Eocene- Senonian age create of the second frequency. The first resonance frequency is related to Judea Gr. The shaded areas on the Fourier spectra show frequency ranges of resonance motion. 25

The first resonance frequency is related to the deeper hard rock reflector, and as noted above, it is different in Hadera-Binyamina and in the Carmel coast. The second peak is a product of the seismic impedance between alluvium and intermediate hard layer, which is also various in the different parts of study area. Thus, in the west of Hadera-Binyamina (the Coastal Plain), the surface of hard intermediate layer correlates with surface of calcareous sandstone of the Kurkar Gr. (Fig 7a). To the east (Hashefela), surfaces of clay - marl of the Saqiye Gr., then marl-chalk of Eocene- Senonian age create the second peak in the H/V spectra in accordance with geological structure (Fig 7 b). The first peak shows both low frequency and amplitude. The position of the second peak depends on thickness of intermediate and soft layers. The amplitudes of these peaks depend mainly on the S-wave velocity in the soft soil. In the Carmel coast, alluvium sediments cover the calcareous sandstone of the Kurkar Gr. in the western part (Fig 8a) and the Cenomanian chalk complex in the centre and in the eastern part (Fig 8 b, c, d). Seismic impedance between alluvium and these high-velocity layers creates second peak in the H/V spectra. Frequency range of the first and second peaks is 1-10 Hz, and H/V amplitudes are 4-9. Fourier spectra and H/V ratios with single peak that were observed in the study area also vary in accordance to geology conditions (Fig 9- 10). The same types of H/V spectral ratios are determined in the southern part (Coastal Plain) of Hadera-Binyamina. Here the depth to the deeper reflector is more than 850m. We obtained only one resonance frequency that was created by surface of Quaternary calcareous sandstone overlaying by alluvium (Fig 9 a, b). In the eastern part of study area at some sites located on exposure of Eocene –Senonian sediments the same results were obtained. Here a single peak of H/V functions is associated with the deep Judea Gr. and is characterized by low H/V amplitudes (Fig 10 a). In the Carmel coast, H/V functions show a single peak that related only to limestone- dolomite of Turon-Albian age. Such spectra and ratios are distributed on an outcrop of calcareous sandstone that overlies the hard carbonate of Turon-Albian age and also at some sites on the alluvium sediments, those overlaying the hard carbonate, and located very close to an exposure of limestone- dolomite (Fig. 10 b, c). 26

a) b)

c) d)

Figure 8. Examples of V and H average Fourier spectra and corresponding H/V spectral ratios at sites whose surfaces (a) of calcareous sandstone of Kurkar Gr. and (b, c, d) of the Cenomanian chalk complex create of the second frequency. The first resonance frequency is related to the dolomite of Cenomanian-Albian age. a) 27 b)

Figure 9. Examples of V and H average Fourier spectra and corresponding H/V spectral ratios at sites whose surface of calcareous sandstone of Kurkar Gr. creates the resonance frequency. a) b) c)

Figure 10. Examples of V and H average Fourier spectra and corresponding H/V spectral ratios at sites whose surfaces (a) of Judea Gr. and (b, c) of dolomite of Cenoman-Albian age create the resonance frequency. 28

To provide assessment of the Vs model for study area we have used dense ambient vibration measurements and all available geological information. Measurements were performed near existing well, where type of soil and their depths are known and in places where velocity data from refraction lines are available. Locations of wells and refraction lines, used in this study are shown in Fig.1. A detailed comparison of 1D analytical response functions and experimental curves of H/V ratios allows estimating shear wave velocities (Vs) of different lithological units and thickness of the sediments. Point 83 was observed in the Carmel Coast near Binyamina well that penetrates the dolomite of Cenomanian age at depth 36m. Information about the Vs of dolomite and alluvium could be derived by using borehole data and data from refraction survey (Ezersky, 2009), which was carried out near the well (Fig.1). Table 1 presents borehole data and velocity model from refraction line and from interpretation of H/V spectra of ambient vibration measurements at point 83. Calculated analytical transfer function is compared with the empirical H/V spectral ratio at point 83 (Fig.11). Both functions have the same resonance frequency and similar shape. We must note there is the first time when Vs of dolomite (2200m/sec) was obtained from the refraction survey and also was confirmed by ambient vibration measurements and was used for modeling as Vs of deeper reflector in the Carmel coast.

Table 1. Geotechnical data and soil column model for point 83 (Binyamina well)

Refraction survey Suggested soil Binyamina well data data, Rl-1 column model Depth Thickness, Vs, Thickness, Vs, Stratigraphy Lithology interval m m/sec m m/sec ,m Sand, silt 20 0 to10-17 370 20 310 Quaternary 10-17 Loam, gravel 16 480 16 500 to36-47 36-47 Cenomanian Dolomite 20 and 2210 2200 below

29

a) Binymina b) 0

Sand, silt -10

-20

m

, Loam,

h t gravel p -30

e D

-40 Dolomite

-50

10 c)

5

o

i t

a 3

R

l

a r

t 2 c

e p S 1

0.5 0.2 0.3 0.5 1 2 3 5 10 Frequency , Hz

Figure 11. a) Lithostratigraphic section of the Binyamina well; b) velocity-depth section along refraction profile RL-1 ; c) comparison between H/V spectrum obtained at point 83 (red line) and the analytical transfer function computed using well data and velocities from refraction line RL-1 (black line).

To estimate Vs of Carmel or Cenomanian chalk complex we used data (Table 2) from seismic refraction line RL-2, which was carried out near Kabara 3 well in the Carmel coast (see Fig. 1). Kabara 3 well reaches the Cenomanian chalk complex at depth of 38m. Soil column model for point 85, located on the refraction line RL-2, is also presented in Table2. Figure 12 shows average H/V spectral ratio at point 85 in comparison to the analytical transfer function calculated on the basis of the velocity model that is presented in Table2. 30

Table 2. S-wave velocity model based on the refraction line RL-2 and suggested soil column model for point 85 in the Carmel coast.

Refraction survey data, Soil column model for RL-2 point 85 Depth Thickness, Vs, m/sec Vs, m/sec interval, m m 14 170 0 to 28 100 20 190

28 and below 980 140 960

2200

10

5 o

i t

a 3 R

l

a r

t 2

c

e p

S

1

0.5 0.2 0.3 0.5 1 2 3 5 10 Frequency , Hz

Figure 12. H/V spectral ratio (red line) and the analytical transfer function (black line) for point 85 in the Carmel coast.

Calculated depth of the main reflector (dolomite of Yagur Fm.) for point 85 is 174m, calculated thickness of the layer is 140m which we correlate as the the Cenomanian chalk complex. According to data from wells Atlit 1 and Atlit 2 (seeTable3) that located in the north- west of the Carmel coast (see Fig.1), thickness of the Cenomanian chalk complex are 175m and 138m. We construct a velocity model for point 311, measured near Atlit 2 well, using Vs for chalk of the Cenomanian chalk complex that is presented in Table2. Atlit 2 well is located on outcrop of calcareous sandstone of Kurkar Gr. H/V function shows a single peak at 1.4 Hz with relatively low amplitude, which is typical for case of high-velocity uniform lithological section. 31

Figure 13 shows good agreement between H/V spectral ratio at point 311 and calculated analytical transfer function.

Table 3. Geotechnical data and soil column model for point 311 (Atlit 2 well) in the Carmel coast.

Suggested soil column model for Atlit 2 well data point 311 Stratigraphy Lithology Thickness, m Thickness, m Vs, m/sec Quaternary Calcareous 32 33 740 (Kurkar Gr.) sandstone Cenomanian Chalk 138 138 950 Albian Dolomite 680 and below 2200 (Jagur Rm.)

10

5

o

i t

a 3

R

l a

r

t 2

c e

p S 1

0.5 0.2 0.3 0.5 1 2 3 5 10 Frequency , Hz

Figure 13. H/V spectral ratio (red line) and the analytical transfer function (black line) for point 311 (Atlit 2 well) in the Carmel coast.

As we noted above, main deep reflector in southern part of study area is top of limestone and dolomite of Late Cretaceous age (Judea Gr.) and its Vs=1970m/sec was already detected by refraction survey in the Parsa area, located in the Dead Rift system (Zaslavsky et al.,2000), and in the Dimona area (Zaslavsky et al., 2004). In our previous works in Hashefela region (Kefar Sava, Petah Tikva, Lod Valley) we also had demonstrated that chalky limestone of Turonian age with Vs=1000-1400m/sec is not confirmed as the deeper reflector, but as an intermediate layer or part of this layer (Zaslavsky et al., 2005-2007). In this study we present examples of three wells that located in the eastern part of the study area (Hadera-Binyamina) (see Fig.1), and reach the fundamental reflector and whose 32 lithostratigraphic sections contain chalky limestone of Turonian age. Suggested soil column models for wells are presented in Table 4. Fig. 14 shows the lithostratigraphic section of Wadi-Ara 1 well and H/V spectral ratio at point 287, situated near the well, in comparison with two analytical transfer functions (black and blue lines). We observed H/V spectra with two peaks at 4.5Hz and at 8.5Hz, that consistent with analytical transfer function (black line) calculated on the basis of the subsurface model that contains the alluvium sediments, chalk of Late Eocene- Santonian age and chalky limestone of Turonian age (seeTable4). Analytical function shows only one resonance frequency 8.5Hz (blue line) was computed on the basis of the model where chalky limestone of Turonian age is not included. So, the first resonance frequency- 4.5 Hz relates to the main reflector at depth of 82m. Fig-s 15 and 16 present the lithostratigraphic sections of Karkur3 well , of Menascheh 1 well , the H/V spectrums at points 460 and 462 that located near those wells, and the corresponding analytical transfer functions calculated on basis of the suggested velocity models. Table 4. Geotechnical data and soil column models for points 462 (Menascheh1 well), 460 (Karkur 3 well) and 287 (Wadi Ara 1 well).

Suggested soil column models Menascheh-1 Karkur 3 Wadi Ara 1 well

well well

Stratigraphy Lithology

m m m

Vs, m/sec Vs, m/sec Vs, m/sec Vs,

Thickness

Thickness Thickness Thickness Sand, clay, loam, Quaternary sandy loam, silt 67 380 33 210 7 250 sandstone, gravel Oligocene- Middle Marl 188 860 Eocene Lower Eocene- Chalk 360 980 93 930 15 950 Santonian

Turonian (Judea Gr.) Chalky limestone 25 1340 65 1260 60 1230

Cenomanian Dolomite 2100 2100 2100 (JudeaGr.)

33

a) Wadi Ara 1 b) 0 Alluvium 10 Chalk

5

-40 Chalky

o

i

m t , limestone

a 3

h

t

R

p l

e

a

r D

t 2

c

-80 e p

S Dolomite 1

-120 0.5 0.5 1 2 3 5 10 20 Frequency , Hz

Figure 14. a) Lithostratigraphic section of the Wadi Ara 1 well; b) comparison between H/V spectrum obtained at point 287 (red line) and the analytical transfer functions (black and blue lines).

Karkur 3 0 a) Alluvium b)

10 -50

Chalk 5

o i

-100 t

a 3 m

R

,

l

h

t

a r

p

t 2

e

c

e D

p -150 Chalky S limestone 1

-200 0.5 0.2 0.3 0.5 1 2 3 5 10 Dolomite Frequency , Hz

-250

Figure 15. a) Lithostratigraphic section of the Karkur 3 well; b) comparison between H/V spectrum obtained at point 460 (red line) and the analytical transfer function (black line) computed using data from Table 4.

34 Menascheh 1 0 a) Alluvium

b)

Marl 10 -200

m

, h

t 5

p

e D

o

i t

a 3

R

-400 l

Chalk a r

t 2

c e

p S 1 -600 Chalky limestone 0.5 0.1 0.2 0.3 0.5 1 2 3 5 10 Frequency , Hz Dolomite -800

Figure 16. a) Lithostratigraphic section of the Menascheh 1 well; b) comparison between H/V spectrum obtained at point 462 (red line) and the analytical transfer function (black line) computed using data from Table 4.

We achieved good agreement between the analytical transfer functions and the empirical H/V spectrums for points 460 and 462. Our modeling shows S-velocity for the chalky limestone that varies from 1230 m/sec up to 1340 m/sec. The range of Vs for marl-chalk layer is 860-980 m/sec, that well agrees with results of velocities for Oligocene- Senonian rock sections in our previously works (Zaslavsky et al., 2005-2007). We must note that layer of chalky limestone of Turonian age with Vs=1340m/sec in the Menascheh 1 well does not influence on the model, because of its thickness 25m that less than 10% of the depth to the main reflector (see Table 4). Next examples demonstrate the measurement at sites near wells in the western part of the study area, where clay and marl of Saqiye Gr. (Yafo Fm.) are widely presented (see Fig.1). The Table 5 contains data of wells Caesarea 2 and Caesarea 3, which penetrate the Judea Gr. and below, with suggested velocities models for points 131 and 132 that were measured near the wells. Obtained S-velocities for marl-chalk layers of Senonian rock sections (see Table 4) was used. We also used Vs for alluvium sediments and Vs for calcareous sandstone from the seismic refraction and downhole surveys, which were carried out in the power station in the town of Hadera (Ezersky, 2008). We note that the wide range of velocities 600m/sec-750m/sec for 35 sediments of Saqiye Gr. (Yafo Fm.) obtained in this study may be explained by the densification of marine clay with the depth and by increase of a part of marl in the sediments. Calculated analytical transfer functions are compared with the empirical H/V spectral ratios at points 131 and 132 and good agreements are obtained (see Fig.17).

a) b)

10 10

5 5 o

o

i i

t t a

a 3 3 R

R

l l

a a

r r t

t 2 2

c c

e e

p p S S

1 1

0.5 0.5 0.1 0.2 0.3 0.5 1 2 3 5 10 0.1 0.2 0.3 0.5 1 2 3 5 10 Frequency , Hz Frequency , Hz

Figure 17. Comparison between H/V spectral ratios (red lines) obtained (a) at point 132 (Caesarea 3 well) and (b) at point 131(Caesarea 2 well) and the analytical transfer functions (black lines) computed using data from Table 5.

Table 5. Geotechnical data and soil column models for points 131 (Caesarea 2 well) and 132 (Caesarea 3 well).

Suggested soil column models Caesarea 2 well Caesarea 3 well Stratigraphy Lithology Thickness, Thickness, Vs, m/sec Vs, m/sec m m Sand 6 180 5 140 Quaternary (Kurkar Gr.) Calcareous 99 740 119 700 sandstone Pliocene Clay, marl 343 750 75 620 (Saqiye Gr.) Senonian Marl, chalk, 42 950 120 820 (Mt.Scopus Gr.) chert, Turonian 112 and 120 and Limestone 1900 (Judea Gr.) below below

Based on results of recent and the previous investigations (Zaslavsky at al., 2004-2008), we constrained the Vs range for main lithological units, which represented in the study area (Table6). 36

Table 6. S-wave velocity model for the study are.

Lithological unit Vs Silt, terrestrial clay (Holocene) 160-250

Sand and sandy loam (Kurkar Gr., Pleistocene) 200-350

Sandstone, gravel, conglomerate (Kurkar Gr., Pleistocene)) 350-550

Calcareous sandstone (Kurkar Gr.) 700-750

Clay, marl (Saqiye Gr., Yafo Fm.) 600-750

Marl, chalk (Oligocene, Eocene, Senonian) 800-1000

Chalky limestone (Turonian, Judea Gr.) 1000-1400

Limestone (Turonian, Cenomanian, Judea Gr.) 1900

Chalk (Cenomanian chalk complex) 900-1000

Dolomite (Cenomanian, Judea Gr.; Albian, Jagur Fm.) 2000-2200

We developed Vs models over the study area and used H/V spectral information in estimating layers thicknesses where well data or other data are not available. The analytical transfer functions are consistent with the observed H/V functions of each site. The dense and uniform grid helps us to obtain models that are consistent with all other geological information.

Profile A-A Profile A-A crosses the center part of study area (Hadera-Byniamina) from the west to the east and is presented in Fig.18. The direction of profile was chosen while taking into account the distribution of the resonance frequency together with the goal to demonstrate new geological structure elements derived from H/V measurements. Comparison between H/V spectral ratios and analytical transfer function at selected sites are shown in Fig.19. In the western part of profile (Caesarea range) we observed spectral ratios with two frequencies (points 131-13). The first resonance frequency (0.4-0.6 Hz) is associated with the deep Judea Gr. and the second frequency (6Hz - 10Hz) is correlated with the thin alluvial sediments (about 5-10m) that overlay of the Quaternary calcareous sandstone (Fig 18). Points 37

131 and 78 are characteristic examples (Fig. 19). Here presence of two reflectors is confirmed by borehole data from wells Caesarea1, Caesarea3 and well K-34 (Kafri and Ecker, 1964). Deeper and shallower reflectors are marked on the profile by blue and green lines (Fig.18). Clay and marl of Yafo Fm. form the most part of the lithostratigraphic section. Next part of profile, from point 28 –to 298, shows example that sometimes we observed identical H/V ratios that correspond to different geological conditions. Modeling and analysis of geology allow us to come to the right conclusion. Here we suggest three blocks, boarded by faults, with different depth to the main reflector. The first block is presented by the point 28 (Fig 18, 19). H/V ratio has a single peak 0.7Hz with relatively low amplitude. We have interpreted this frequency as related to the deeper reflector- limestone of Judea Gr. Our interpretation was confirmed by the modeling on the sites 12, 37, not included in the profile, but also located on this block. Point 12 has peak of resonance frequency at 0.7Hz, is placed on the outcrop of calcareous sandstone, near the well 55/A (Kafri and Ecker, 1964). Well reached the clay of Yafo Fm. at the depth of 110m. Point 37 (0.6Hz) is located near the well Hadera 4 (Kafri and Ecker, 1964), that reached the clay of Yafo Fm. at the depth of 80m. Having known thicknesses of calcareous sandstone at these points, we obtained good agreement between observed H/V functions and calculated transfer functions only assuming the frequencies 0.6Hz-0.7Hz are responsible for the deeper reflector-Judea Gr. Calculated depths to the reflector are 330-430m, but according to the geological data (Fleischer and Gafsou, 2000) the depth to the top Judea Gr. exceeds of 1200m. The part of profile (points79, 2) also is characterized by H/V ratios with one peak (1.3-1.5 Hz), that is created by seismic impedance between alluvium and marl-chalk of Oligocene- Senonian age (Fig.18, 19). Absence of the frequency responsible for the deeper reflector indicates that the depth to the main deeper reflector is more than 850m. This zone is marked on the profile and on the maps of distribution of resonance frequency and its amplitude (Fig.-s 18, 24). We suggested here also absence of the clay of Yafo Fm. It may be indirect confirmed by absence these sediments in the well Menashet 3, located in two km to north. The low first resonance frequencies (0.32Hz-0.4Hz) appear in the H/V functions obtained at points 219-221, for which models for two reflectors were constructed. Point 222 is shown as example (Fig. 19). We interpreted the block, presented by points 79-221, as tilted deep structure (Fig. 18). 38

Points 7-298 show the spectral ratios with two peaks of resonance frequency (Fig.19). Sharp changes of the first resonance frequency from 0.32 Hz to 0.6 Hz (points221-7), from 0.8Hz to 0.36Hz (points298-138) and changes in the shape of the H/V spectra suggest the existence of faults, which divide the blocks with different depths of the main reflector (Fig.18). The second frequency changes from 1.2 Hz to 1.4 Hz. The uplifted geological structure that derived from ambient vibration measurements is shown in Fig.18. Calculated depth to main reflector is about 300 m. According geological data (Fleischer and Gafsou, 2000), this part of the profile corresponds to the syncline with depth to top Judea Gr. of 900m. The difference with supposing models is significant. Shift in the fundamental frequency from 0.35 Hz up to 0.45 Hz between neighboring points 292 and 146 corresponds to the vertical shift in the reflector depth of 90 meters and is accompanied by fault, which is also mapped in the structural map of the Judea Gr. (Fleischer and Gafsou, 2000). Next structural block occupying segment of profile AA from point 138 to point 160 is characterized by frequency 0.43-0.46 Hz. Point 290 shows frequency 0.6 Hz and located on the eastern side of the NS oriented fault running at the foothills of the Samaria mountains. Further to the east, a gradual increase in depth of the top Judea Gr. is observed that completely agrees with geological data. We note that according to the information from wells situated along the eastern segment of profile AA (karkur-38, 3, 101 and Ein Shemer wells) the Turonian chalky limestone layer is clearly detected within the Judea Gr. Furthermore, with decrease in the total sediment thickness above the dolomites, the contribution of the limestone layer becomes more significant and should be taken into account in the model calculations as intermediate hard layer. Its S-velocity as we know is in the range 1100-1400 m/sec. H/V spectral ratios in the segment of profile AA from point 111 to 269 yield fairly high amplitude of the second resonance peak related to the low velocity quaternary sediments over Eocene-Senonian marl-chalk layer. In addition, we note that the last segment of profile AA oriented strictly east-west, runs within the same structural block and no faults are identified up to the eastern edge. 39

Figure 18. Schematic geological cross section along profile AA reconstructed on the base of H/V analysis

40

10 10 10 131 78 28

5 5 5

o

o o

i

i i

t

t t

a a

3 a 3 3

R

R R

l

l l

a

a a

r r

r

t t

2 t 2 2

c

c c

e

e e

p

p p

S S S 1 1 1

0.5 0.5 0.5 0.1 0.2 0.3 0.5 1 2 3 5 10 0.1 0.2 0.3 0.5 1 2 3 5 10 0.1 0.2 0.3 0.5 1 2 3 5 10 Frequency , Hz Frequency , Hz Frequency , Hz 10 10 10 79 222 7

5 5 5

o

o

o

i

i

i

t t

t

a a

3 a 3 3

R

R

R

l

l

l

a

a

a

r

r

r

t t

2 t 2 2

c

c

c

e e

e

p

p

p

S

S S 1 1 1

0.5 0.5 0.5 0.1 0.2 0.3 0.5 1 2 3 5 10 0.1 0.2 0.3 0.5 1 2 3 5 10 0.2 0.3 0.5 1 2 3 5 10 Frequency , Hz Frequency , Hz Frequency , Hz 10 10 10 298 138 146

5 5 5

o

o o

i

i i

t

t

t

a a

3 3 a 3

R

R R

l

l l

a

a a

r

r

r

t t

2 2 t 2

c

c c

e

e e

p

p

p

S S S

1 1 1

0.5 0.5 0.5 0.1 0.2 0.3 0.5 1 2 3 5 10 0.1 0.2 0.3 0.5 1 2 3 5 10 0.1 0.2 0.3 0.5 1 2 3 5 10 Frequency , Hz Frequency , Hz Frequency , Hz 10 462 290 111

5 5 5 o

A o

i

i

o

t

t

i

t

a a

a 3

R

R

R

l

l

l

a

a

a

r

t r

2 r t

2 t 2

c

c c

e

p

e

e

S

p

p S S 1 1 1

0.5 0.5 0.5 0.1 0.2 0.3 0.5 1 2 3 5 10 Frequency , Hz 0.1 0.2 0.5 1 2 5 10 0.1 0.2 0.5 1 2 5 10 Frequency (Hz) Frequency (Hz)

460 289 461

5 5 5

o

o

o

i

i

i

t

t

t

a a

a

R

R

R

l

l

l

a

a

a

r r

r

t t

289 2 t 2 2

c

c

c

e

e

e

p p

p

S

S S 1 1 1

0.5 0.5 0.5 0.1 0.2 0.5 1 2 5 10 0.5 1 2 5 10 20 0.5 1 2 5 10 20 Frequency (Hz) Frequency (Hz) Frequency (Hz) Figure 19. H/V spectral ratios (black lines) and analytical transfer function (black dashed lines) for sites along profile A. 41

Profile B-B As seen from Fig. 1, profile B-B has W-E direction and crosses several faults. To demonstrate in the best way a series of faults detected in the eastern part of the study area, direction of the profile was slightly changed to the southeast. Figure 20 shows cross section along profile B-B reconstructed on the base of H/V ratio analysis. Figure 21 displays examples of the H/V spectral ratios for representative sites along the profile, for instance, at sites located on different sides of suggested faults. As seen from this figure, the majority of H/V spectral ratios show two resonance peaks. The fundamental one is associated with the limestone and dolomite of the Judea Gr. (Cenomanian-Turonian age). The shallow reflector is represented by calcareous sandstone together with the Yafo marl at the Coastal Plain; and by marl-chalk of Oligocene- Eocene-Senonian age in Hashefela region. We note that calcareous sandstone of the Kurkar Gr. is the only reflector at the western Coatal Plain, where the Judea Gr. deepens at a depth of more than 850-900m and produces the single peak whose frequency and amplitude strongly depend on the lithological composition and thickness of the alluvial deposits. The fundamental peaks at sites 184 and 498 in Fig. 21 yield frequencies 4.5 Hz and 1.8 Hz respectively. Sites 159 and 157 located at the Kurkar ridge show no site effects. A prominent peak at frequency 0.28 Hz and amplitude of 2.5 at site 271 indicates a change in the fundamental reflector on the significantly deeper one, which is obviously the limestone and dolomite of the Judea Gr. According to the analytical models calculated for site 271, the reflector depth is 870 meters. We suggest the existence of a fault between sites 498 and 271. The thickness of the soft sediments over the shallow reflector equal to 40 meters provides the second resonance peak at frequency 3-4- Hz. According to the geological information the Yafo clay and Paleogene marl-chalk complex are intermediate layers. The upper part of the section including the Quaternary sediments and depth of the Yafo Fm. could be estimated from water wells and the geological cross section No 1 (Ecker, 1999). Thickness of the marl-chalk complex is calculated. Starting from this place and to the east we are speaking about the Judea Gr. as the fundamental reflector. H/V ratio obtained at Site 155 shows the fundamental peak at frequency 0.33 Hz corresponding to a depth of 770m and no second resonance peak due to location of this site on the Kurkar outcrop. Up to site 106

(f0=0.45 Hz), the gradual increase in the fundamental frequency and, accordingly, decrease in the thickness of sediments above the fundamental reflector is observed. Sharp increase in the fundamental frequency up to 0.55 Hz and 0.6 Hz at sites 108 and 357 (we show site 357 in Fig. 42

21) is accompanied by a fault with vertical shift of about 100m. This fault is mapped in the structural map of the Judea Gr. (Fleischer and Gafsou, 2003); however it is shifted approximately 1 km to the east. Furthermore, in agreement with the geological cross sections by Kafri, the Yafo Fm. is wedging out and the subsurface model is represented by the soft Quaternary sediments and the Middle Eocene chalk over the Judea Gr. Despite the fact that along this segment of the profile B-B there are no deep wells reaching the Judea Gr., we have included into the subsurface model Turonian chalky limestone of 30-50m thick as an intermediate layer. It does not influence significantly on the analytical transfer function for sites located at this segment of the profile, however it is found in all the wells reaching the Judea Gr. located in the eastern part of the study area but to the north of profile B-B (Karkur-38, 145, 101, 146, En-Shemer and etc.). In Maanit 5/A well located on the profile thickness of the Turonian limestone is 60 meters. Further eastward along the profile the reflector depth steadily decreases, however, between sites 133 and 197 we detect fault. This fault has EW direction and it is one of the series of faults dissecting the eastern part of the area (see map in Figure 27) into blocks of different depth. However, while crossing profile B-B within the same structural block it does not show considerable vertical displacement. Site 193 and 283 are located at the different structural blocks and we can see both vertical displacement of 60 meters and change in the H/V ratio parameters indicating changes in the subsurface structure. Site H282 located at (well Maanit 4A) is located on outcrop of the Mt. Scopus Fm. and produces low amplitude peak and frequency 2 Hz. The last segment of the profile B-B is out of the study area with the regular measurement grid, but a few ambient noise measurements carried out within the framework of the Hashefela project (Zaslavsky et al., 2002) allowed to continue profile B-B for a few more kilometers to the east.

43

Figure 20. Schematic geological cross section along profile B-B reconstructed on the base of H/V analysis 44

5 184 5 498 5 271 5 155

o

o

o

o

i

i

i

i

t

t

t

t

a

a

a

a

R

R

R

R

2 2 2

l

l l

l 2

a

a

a

a

r

r

r

r

t

t

t

t

c

c c

1 1 c 1

e

e e

e 1

p

p

p

p

S

S

S S 0.5 0.5 0.5 0.5

0.5 1 2 5 10 0.5 1 2 5 10 0.1 0.2 0.5 1 2 5 10 0.1 0.2 0.5 1 2 5 10 Frequency (Hz) Frequency (Hz) Frequency (Hz) Frequency (Hz)

5 292 5 106 5 357 5 113

o

o

o

o

i

i

i

i

t

t

t

t

a

a

a

a

R

R

R

R

l

l

l

l

a

a a

2 2 a 2 2

r

r

r

r

t

t

t

t

c

c

c

c

e

e

e

e

p

p p

1 1 p 1 1

S

S

S S

0.5 0.5 0.5 0.5 0.1 0.2 0.5 1 2 5 10 0.1 0.2 0.5 1 2 5 0.1 0.2 0.5 1 2 5 0.1 0.2 0.5 1 2 5 Frequency (Hz) Frequency (Hz) Frequency (Hz) Frequency (Hz)

5 197 5 193 5 283 5 282

o

o o

o

i

i i

i

t

t t

t

a

a a

a

R

R R

R

l

l l

l

a

a a

2 a 2 2 2

r

r r

r

t

t t

t

c

c c

c

e

e e

e

p

p p

1 p 1 1 1

S

S S S

0.5 0.5 0.5 0.5 0.1 0.2 0.5 1 2 5 0.1 0.2 0.5 1 2 5 0.10.2 0.5 1 2 5 10 200.1 0.2 0.5 1 2 5 10 Frequency (Hz) Frequency (Hz) Frequency (Hz) Frequency (Hz)

Figure 21. H/V spectral ratios (solid line) compared with analytical transfer function (dashed line) for representative sites of profile B-B.

Profile C-C

Profile C-C (location see in Figs. 1, 24) crosses the northern part of the study area- the Carmel coast zone- from west to east. H/V spectral ratios and corresponding analytical transfer functions at sites along profile are shown in Fig. 23. 45

As it was note above, in the Carmel coast main deep reflector presents basically as dolomite of Albian age (Yagur Fm.) and partly as dolomite of Upper Cenomanian age and partly as limestone of Turonian age. The important feature of this profile derived from H/V measurements is the sharp change of the main reflector from dolomite of Albian age (Jagur Gr.) to dolomite of Cenomanian age, what connected with lithologic replacement of chalk (Carmel chalk complex) to dolomite in the Cenomanian age (Fig. 22). Western part of profile (points 87-242) is characterized by H/V spectral ratios with two peaks of frequencies (Fig. 23). The first (1.3-2 Hz) is associated with the dolomite of Yagur Fm and the second peak of spectra (2.5-4 Hz) is produced by seismic impedance between the alluvial sediments and calcareous sandstone of Kurkar Gr. and Cenomanian chalk complex, which together form here an intermediate hard layer. The corresponding H/V amplitudes of both peaks reach a level of 5-6. The exception is the point 231 that was observed on outcrop of calcareous sandstone (Fig. 22). The H/V function shows a single peak at 1.3 Hz with low amplitudes (less than 2.5). Taking into account the geology data (Kafri, U, Ecker,A, 1964) and results of H/V measurements, part of the profile between points 242 and 243 we interpreted as change reflector zone. H/V ratios obtained at points 243, 244 show single resonance frequencies with high amplitudes. The structure of the eastern part of the profile presents a simple two-layer model, when the soft layer of alluvium overlays dolomite of Cenomanian age. Point 245 does not show the site effect, because is located on the outcrop of dolomite.

46

Figure 22. Schematic geological cross section along profile C-C reconstructed on the base of H/V analysis

47

10 10 10 87 231 240

5 5 5

o o o

i i i

t t t

a a

a 3 3 3

R R R

l l l

a a a

r r r

t t

t 2 2 2

c c c

e e

e

p p p

S S S 1 1 1

0.5 0.5 0.5 0.1 0.2 0.3 0.5 1 2 3 5 10 0.1 0.2 0.3 0.5 1 2 3 5 10 0.1 0.2 0.3 0.5 1 2 3 5 10 Frequency , Hz Frequency , Hz Frequency , Hz

10 10 10 242 243 244 5

5 5

o

i

t o

o

a i

3 i

t

t

R

a a

3 3

l

R

R

a

l

r l

t 2

a

a

c

r

r t

t 2 2

e

c

c

p

e

e

S p

p

S S 1 1 1

0.5 0.1 0.2 0.3 0.5 1 2 3 5 10 0.5 0.5 Frequency , Hz 0.1 0.2 0.3 0.5 1 2 3 5 10 1 2 3 5 10 20 Frequency , Hz Frequency , Hz

Figure 23. H/V spectral ratios (black lines) and analytical transfer function (black dashed lines) for sites along profile C.

Distribution of H/V frequency and its associated amplitude for the fundamental and second resonance peaks

A data set of the ambient noise measurements was used to construct the distribution maps of the H/V fundamental and second resonance frequencies and maximum relative amplification in the study area. During campaign of systematic ambient noise measurements in the Hadera- Binyamina-Hof HaCarmel area, we took the opportunity to revise available results presented in the Coastal Plain and Ha-Shefela reports (Zaslavsky et al., 2002, 2003) to supplement the present database. Frequency and amplification maps constructed for two H/V resonance peaks shown in Figs. 24, 25 integrate all experimental data. The first, fundamental peak is associated with the regional reflector – Judea Gr. Clearly, the main geological structure is reflected in the measurement results: thus, the frequency distribution within the Carmel Coast zone, located on the western slope of the Carmel anticline, follows the morphology of the Turonian to Albian limestone and dolomite and is characterized by the prevalent frequencies 2-3 Hz. The higher frequencies up to 8 Hz are attached to the Judea outcrop in the east. Several faults directed NW- SE are clearly identified in the Carmel coast. In the area to the south of the Binyamina a deep- seated transverse fault, which is mapped by decrease in the fundamental frequency, the prevalent 48 frequencies are 0.3-0.5 Hz. Within this low frequency field there are some areas of the higher values up to 1 Hz bounded by SW-EN and W-E faults are interpreted as uplifted blocks of the Judea Gr. Areas with no fundamental frequency are revealed in both central and southwestern parts of the Hadera-Binyamina area. The northern one may be explained by too thin alluvial layer over reflector on the uplifted block. The reason for absence of the fundamental frequency in area located 2 km to the south of the mentioned one, as well as in the southern part of the Coastal plain, is dipping of the top Judea Gr. to the depth more than 850 m that is beyond the resolution of our method. To the east of the fault of N-S strike, which is subdivided into minor segments, we observe increase in the fundamental frequency from 0.5 Hz up to 8 Hz associated with rising of the Judea Gr. A series of sublatitudinal faults dividing the main fault into separate parts which delineate the structural blocks of different depth and also shifted are detected also in the structural map of the Judea Gr. The Carmel Coast is distinguished on the map of the H/V amplitude associated with fundamental frequency owing to its higher values (up to 8). Such values are revealed at those sites where thick silt layer together with calcareous sandstone overlays directly dense dolomites and chalks. The area south of the Binyamina faults shows the low to moderate amplitude values (up to 2.5) depending on lithology and thickness of the sediments overlying the Judea Gr. We observe a general increase of the first amplitude values toward the east, where the Judea Gr. is rising and sediments over reflector are thinning. The distribution of the second resonance frequency and its amplitude (Figs. 24-25) strongly depends on the geotechnical characteristics of the alluvium layer above the shallow reflector and underlying rock, which is represented by calcareous sandstone of the Kurkar Gr. together with clay-marl of the Yafo Fm. at the Coastal Plain and by Eocene-Senonian marl-chalky complex in the central and eastern parts. We note that despite of calcareous sandstone of the Kurkar Gr. is the only reflector at the Coastal plain, while the Judea Gr. keeps dipping to depths of 800-1500 meters; we consider it as a part of the shallow one in the map of spatial distribution of the second H/V frequency. The range of the second frequency is 0.7-10 Hz. The frequency contours has south-north strike along the coastline in agreement with the strike of the Kurkar ridge. Sites with no resonance frequency are found at exposed calcareous sandstone of the Kurkar Gr. in the western part and at outcrop of the marl-chalk in the eastern part.

49

204000

A

2.5

4

6

8

site effect site

No No

B

200000

196000

C

C

192000

A

B

188000

204000

A

Hz

0.3

0.4

0.5

0.6

0.8

1

1.5

2

3

4

5

7

site effect

No No

B

200000

196000

C

C

192000

A

B

188000

700000

704000

708000

712000

716000

720000

724000

728000

732000

736000

740000 744000 748000

Figure 24. Distribution of the H/V fundamental frequency and its associated amplitude over the study area. The red line indicates position of the geological cross section.

50

204000

A

B

site effect

No

2

2.5

3

4

6

8

9

200000

196000

C

C

192000

A

B

188000

204000

A

Hz

B

site effect

No

0.7

1

1.5

2

3

4

6

8

10

200000

196000

C

C

192000

A

B

188000

700000

704000

708000

712000

716000

720000

724000

728000

732000

736000

740000 744000 748000 Figure 25. Distribution of the H/V second resonance frequency and its associated amplitude. The red line indicates position of the geological cross section. 51

The amplitude associated with the second resonance H/V peak varies from 2.0 to 10. The highest amplitude values are attained at sites where soil column consists of the low velocity alluvium layer over the dolomites in the Carmel Coast area. We note that the second amplitude reaches maximum values of 6 in the Hadera-Binyamina part and areas of increased amplitude are visibly associated with channels of the Hadera River. Areas with no site effects are correspond to outcrops of the Kurkar ridges, which are parallel to the coastline, and Eocene-Senonian deposits in the northeast.

SEISMIC MICROZONATION IN TERMS OF UNIFORM HAZARD ACCELERATION SPECTRA

The Stochastic Estimation of Earthquake Hazard (SEEH) method, developed by Shapira and van Eck (1993), presents the seismic hazard in terms of the site specific uniform probability acceleration response spectrum for a single degree of freedom oscillators with 5% damping. Calculation of spectral acceleration values is based on seismic activity characteristics in the area around the observation site (usually within a radius of 100-150 km) and on the response function at the same site calculated using SHAKE (Schnabel, 1972) program. The SEEH method (Stochastic Estimation of the Earthquake Hazard) 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, et al., 1996; Shamir et al., 2001). Synthetic accelerations at the investigated site are used to compute the acceleration spectrum on a damped oscillating system (i.e., the acceleration response spectrum). The multitude of simulated ground motions provides the data for probabilistic estimations. Based on the Stochastic Estimations method of Earthquake Hazard (SEEH), we computed the seismic hazard for 260 sites in the Hadera-Binyamina-Carmel Coast area for 10% probability of exceedance during an exposure time of 50 years (corresponding to recurrence activity of 1 every 475 years) and damping ratio 5%. The spectrum was computed under the assumption that the sediments are not susceptible to non-linear effects. By comparison of the Uniform Hazard Acceleration Spectra and in consideration of the constructed subsurface models across the investigated area we subjectively divided the area into 23 zones as shown in Fig. 26. 52

204000

17

No site effect site No

f2=0.7-0.8 Hz; A2=3.0; Hz; f2=0.7-0.8

f1=0.35 Hz; A1=2.5; A1=2.5; Hz; f1=0.35

19

f2=1.3-2 Hz; A2=3-3.5; Hz; f2=1.3-2

f1=0.5-1 Hz; A1=2-2.5; A1=2-2.5; Hz; f1=0.5-1

f2=1.6-2.5 Hz; A2=3-6; Hz; f2=1.6-2.5

f1=1-2 Hz; A1=4-7; A1=4-7; Hz; f1=1-2

f2=8-12 Hz; A2=4-6; f2=8-12

f1=5-10 Hz; A1=3.5-4.5; A1=3.5-4.5; Hz; f1=5-10

f2=2-3 Hz; A2=2.5-3.5; Hz; f2=2-3

f1=0.7-2 Hz; A1=2-4; A1=2-4; Hz; f1=0.7-2

f2=2-4 Hz; A2=2-3; Hz; f2=2-4

no first peak; no first

f2=1-2 Hz; A2=2-3; f2=1-2 Hz;

no first peak; no first

f2=3-5 Hz; A2=4-8; f2=3-5

f1=2-3 Hz; A1=6-10; A1=6-10; Hz; f1=2-3

f2=5-8 Hz; A2=4-7; f2=5-8

f1=3-6 Hz; A1=5-7; A1=5-7; Hz; f1=3-6

f2=11.5-3 Hz; A2=4.5-6; Hz; f2=11.5-3

f1=0.5-1 Hz; A1=4-6; A1=4-6; Hz; f1=0.5-1

f2=2.5-5 Hz; A2=3.5-6.5; Hz; f2=2.5-5

f1=0.35-0.5 Hz; A1=2.5; A1=2.5; Hz; f1=0.35-0.5

f2=3-6 Hz; A2=5-9; Hz; f2=3-6

f1=1.5-2.0 Hz; A1=3-4; A1=3-4; Hz; f1=1.5-2.0

f2=1-6 Hz; A2=2-4; Hz; f2=1-6

f1=0.3-0.6 Hz; A1=2-2.5; A1=2-2.5; Hz; f1=0.3-0.6

f2=1.1-1.5 Hz; A2=4.5-6; Hz; f2=1.1-1.5

f1=0.3-0.5 Hz; A1=2-2.5; A1=2-2.5; Hz; f1=0.3-0.5

f2=5.5-10 Hz; A2=2.5-3.5; Hz; f2=5.5-10

f1=0.3-0.5 Hz; A1=2-2.5; A1=2-2.5; Hz; f1=0.3-0.5

f2=1.7-3 Hz; A2=3-5; f2=1.7-3

f1=1-1.8 Hz; A1=4-6; A1=4-6; Hz; f1=1-1.8

f2=2.5-3 Hz; A2= 5-6; A2= Hz; f2=2.5-3

f1=1.5-3 Hz; A1=3-4; A1=3-4; Hz; f1=1.5-3

f2=2-3 Hz; A2=3-3.5; Hz; f2=2-3

f1=1.1-2 Hz; A1=3-4; A1=3-4; Hz; f1=1.1-2

f2=4-6 Hz; A2=2-3; f2=4-6

no first peak; no first

no second resonance peak; resonance no second

f1=4-10 Hz; A1=3-9; Hz; A1=3-9; f1=4-10

no second resonance peak; resonance second no

f1=3 Hz; A1=2.5-3.5; Hz; A1=2.5-3.5; f1=3

no second resonance peak resonance no second

f1=0.9-1.5 Hz; A1=2.5; A1=2.5; Hz; f1=0.9-1.5

f2=3.5-7 Hz; A2=2.5-3; f2=3.5-7 f1=1.5-2.0 Hz; A1=2.5-3; A1=2.5-3; Hz; f1=1.5-2.0

16

16 15

20

5

4

18

12

10

9

2

1

22

19

13

8

20

11

7

6

200000

23

21

17

16

15 14

3 12

18

16

4 15

9

5

11 7

4

7

5 2

8

19

11

14 196000

11

4

8 13

10

5

4

16

3 1

7

9

1

8

9 6

6 7

2

192000

3 21

2

9 16

22

23

188000

700000

704000

708000

712000

716000

720000

724000

728000

732000

736000

740000 744000 748000

Figure 26. Seismic microzonation map of the study area 53

Table 7. Generalized soil column models for calculation of acceleration response spectra of the zones

Thickness, Density, Vs, Damping, Acceleration spectrum and the Zone m g/cm³ m/sec % Israel Building Code IS-413

1.0 40 1.9 750 2

g Zone 1

,

n 0.8

o

i

t

a r

e 0.6

l

e

c c

150 2.0 1000 1 a

1 0.4

l

a

r

t c

e 0.2

p S 0.0 - 2.3 2200 - 0 1 2 3 Period, sec

1.0 25 1.8 400 3

g Zone 2

,

n 0.8

o

i

t

a r

20 1.9 750 2 e 0.6

l

e

c

c a

2 0.4 l

120 2.0 1000 1 a

r

t c

e 0.2

p S 0.0 - 2.3 2200 - 0 1 2 3 Period, sec

1.0

g Zone 3

,

n 0.8

60 1.9 750 2 o

i

t

a r

e 0.6

l

e

c

c a

3 0.4

l

a

r

t c

e 0.2 p

- 2.3 2200 - S 0.0 0 1 2 3 Period, sec 54

Thickness, Density, Vs, Damping, Acceleration spectrum and the Zone m g/cm³ m/sec % Israel Building Code IS-413

1.0

g Zone 4

,

n 0.8 o

20 1.8 450 3 i

t

a r

e 0.6

l

e

c

c a

4 0.4

l

a r

t c

e 0.2 p - 2.3 2200 - S 0.0 0 1 2 3 Period, sec

1.0

g Zone 5

7 1.5 150 5 ,

n 0.8

o

i

t

a r

e 0.6 l

e c

130 1.9 950 1 c a

5 0.4

l

a r

t c

e 0.2

p S - 2.3 2200 - 0.0 0 1 2 3 Period, sec

35 1.6 225 4 1.0

g Zone 6

,

n 0.8

o

i

t

a r

15 1.9 750 2 e 0.6

l

e

c

c a

6 0.4

l

a r

120 2.0 990 1 t c

e 0.2

p S 0.0 - 2.3 2200 - 0 1 2 3 Period, sec

55

Thickness, Density, Vs, Damping, Acceleration spectrum and the Zone m g/cm³ m/sec % Israel Building Code IS-413

1.0

g Zone 7

6 1.5 130 5 ,

n 0.8

o

i

t

a r

e 0.6

l

e c

25 1.8 480 3 c a

7 0.4

l

a

r

t c

e 0.2

p S - 2.3 2200 - 0.0 0 1 2 3 Period, sec

1.0

g Zone 8

20 1.5 160 5 ,

n 0.8

o

i

t

a r

e 0.6

l

e

c c

90 2.0 960 1 a

8 0.4

l

a

r

t c

e 0.2

p S 0.0 - 2.3 2200 - 0 1 2 3 Period, sec

1.0 4 1.5 130 5

g Zone 9

,

n 0.8

o

i

t a 90 1.9 750 2 r

e 0.6

l

e

c

c a 9 0.4

280 1.9 770 2 l

a

r

t c

e 0.2

50 2.0 943 1 p S 0.0 - 2.3 2200 - 0 1 2 3 Period, sec

56

Thickness, Density, Vs, Damping, Acceleration spectrum and the Zone m g/cm³ m/sec % Israel Building Code IS-413

1.0

10 1.6 180 5 g Zone 10

,

n 0.8

o

i

t

a r

e 0.6

190 1.8 440 3 l

e

c

c a

10 0.4

l

a

r t

550 2.0 930 1 c

e 0.2

p S 0.0 - 2.2 1900 - 0 1 2 3 Period, sec

40 1.8 430 3 1.0

g Zone 11

,

n 0.8

o

i t

310 1.9 880 1 a r

e 0.6

l

e

c

c a

11 0.4 l

280 2.0 960 1 a

r

t c

e 0.2

p S 0.0 - 2.2 1900 - 0 1 2 3 Period, sec

65 1.8 400 3 1.0

g Zone 12

,

n 0.8

o

i t

170 1.9 770 2 a r

e 0.6

l

e

c c

130 2.0 960 1 a

12 0.4

l

a

r

t c

60 2.0 1400 1 e 0.2

p S 0.0 - 2.2 1900 - 0 1 2 3 Period, sec

57

Thickness, Density, Vs, Damping, Acceleration spectrum and the Zone m g/cm³ m/sec % Israel Building Code IS-413

1.0

g Zone 13

55 1.7 350 4 ,

n 0.8

o

i

t

a r

e 0.6

270 1.9 700 2 l

e

c

c a

13 0.4

l a

400 2.0 990 1 r

t c

e 0.2

p S 0.0 - 2.2 1900 - 0 1 2 3 Period, sec

1.0 55 1.7 370 4

g Zone 14

,

n 0.8

o

i

t a

10 1.9 740 2 r

e 0.6

l

e

c

c a

14 0.4 l

150 2.0 920 1 a

r

t c

e 0.2

p S 0.0 - 2.2 1900 - 0 1 2 3 Period, sec

1.0 35 1.7 350 4

g Zone 15

,

n 0.8

o

i

t

a r

e 0.6

l

e c

450 1.9 840 1 c a

15 0.4

l

a

r

t c

e 0.2

p S 0.0 - 2.2 1900 - 0 1 2 3 Period, sec

58

Thickness, Density, Vs, Damping, Acceleration spectrum and the Zone m g/cm³ m/sec % Israel Building Code IS-413

10 1.6 190 5 1.0

g Zone 16

,

n 0.8

o

i

t a

50 1.7 360 4 r

e 0.6

l

e

c

c a

16 100 2.0 990 1 0.4

l

a

r

t c

e 0.2

65 2.0 1100 1 p S 0.0 0 1 2 3 - 2.2 1900 - Period, sec

1.0 3 1.5 150 5

g Zone 17

,

n 0.8

o

i

t a

5 1.7 340 4 r

e 0.6

l

e

c

c a 17 0.4

15 1.8 620 2 l

a

r

t c

e 0.2 p

35 2.0 1400 1 S 0.0 0 1 2 3 - 2.2 1900 - Period, sec

1.0 45 1.7 300 4

g Zone 18

,

n 0.8

o

i

t a

270 1.9 860 1 r

e 0.6

l

e

c

c a 18 0.4

40 2.0 1150 1 l

a

r

t c

e 0.2

p S - 2.2 1900 - 0.0 0 1 2 3 Period, sec

59

Thickness, Density, Vs, Damping, Acceleration spectrum and the Zone m g/cm³ m/sec % Israel Building Code IS-413

20 1.6 280 4 1.0

g Zone 19

,

n 0.8

o

i t

8 1.7 350 4 a r

e 0.6

l

e

c

c a

19 35 2.0 900 1 0.4

l

a

r

t c

e 0.2 p

60 2.0 1250 1 S 0.0 0 1 2 3 - 2.2 1900 - Period, sec

1.0

g Zone 20

8 1.5 150 5 ,

n 0.8

o

i

t

a r

25 1.6 200 5 e 0.6

l

e

c

c a

20 0.4 l

90 2.0 950 1 a

r

t c

e 0.2 p 70 2.0 1300 1 S 0.0 0 1 2 3 - 2.2 1900 - Period, sec

1.0

g Zone 21

20 1.6 250 5 ,

n 0.8

o

i

t

a r

e 0.6

l

e c

30 1.7 350 4 c a

21 0.4

l

a

r

t c

e 0.2

p S 0.0 - 1.9 750 - 0 1 2 3 Period, sec

60

Thickness, Density, Vs, Damping, Acceleration spectrum and the Zone m g/cm³ m/sec % Israel Building Code IS-413

1.0 15 1.5 160 5

g Zone 22

,

n 0.8

o

i

t

a r

e 0.6

5 1.7 350 4 l

e

c

c a

22 0.4

l

a

r t

16 1.8 480 3 c

e 0.2

p S 0.0 - 1.9 700 - 0 1 2 3 Period, sec

1.0

g Zone 23

7 1.6 180 5 ,

n 0.8

o

i

t

a r

e 0.6

l

e c

4 1.7 350 4 c a

23 0.4

l

a

r

t c

e 0.2

p S 0.0 - 1.9 700 - 0 1 2 3 Period, sec

The selected zones are distributed as follows: from zone 1 to zone 8 are located on the Carmel Coast, the rest are in the Hadera-Binyamina area. We note that the last three zones (21- 23), located at the Coastal plain, are selected because of different subsurface structure, i.e. the shallow calcareous sandstone of the Kurkar Gr. is the fundamental reflector instead of Turonian- Albian carbonates. The calculated acceleration spectrum for each zone is depicted in Table 7 together with the generalized soil column model. For comparison, also are plotted the design spectra required in the same area by the current Israel Standard 413 (IS-413) and for ground conditions that meet the BSSC (1997) soil classification scheme. We note that the shape of the hazard functions exceed significantly those prescribed by IS-413 code in the period range from 0.2 to 1 sec. in zones from Zone 5 to Zone 8 (the Carmel Coast) and from Zone 15 to Zone 20 61

(the Hadera-Binyamina area). It should be noted that fundamental periods of many of the buildings in the study area also have the same diapason. Together with this, there are zones in both areas where the acceleration spectra are comparable or slightly exceed the design spectra in different period ranges (zones 3, 4, 9-13).

CONCLUSIONS

This report presents a next study of the overall project on microzoning of Israel including the towns of Hadera, Or Akiva, Pardes Hanna, Binyamina, Atlit and the Carmel Coast. The experiment discussed in the present study had the following goals:  to improve the Nakamura‟s technique, in particular to reduce high scatter in the H/V spectra and increase stability;  to estimate site effect in situ using the ambient noise measurements of ground motions and produce maps of the distribution of frequency and amplitude of the H/V resonance peaks;  to evaluate the geotechnical characteristics for one-dimensional analysis of site effects by detailed comparison between the analytical and experimental site response functions, reconstruct geological cross sections and locate faults.  to extrapolate the derived theoretical models over the study area and integrate them into computing the uniform hazard site specific acceleration response spectra for 10% probability during an exposure time of 50 years and damping ratio of 5%;  to divide the study area into zones based on the comparative analysis of the acceleration spectra.

The following main conclusions were reached:  We have to state that our attempts to apply the Singular Spectrum Analysis (SSA) to ambient noise processing have not been successful; manual selecting time windows shows a much more reliable results. We think a key problem of the applicability the SSA method in the Nakamura technique is impossibility to decompose original signal and noise.  A great part of the measurement sites produces H/V spectral ratios with two resonance peaks. In the southern part of the study area from Hadera to Binyamina, the first resonance or fundamental peak is associated with limestone and dolomite of Late Cretaceous age (Judea Gr.). At the Carmel coast, the fundamental reflector is mainly 62

represented by dolomite of Albian age (Yagur Fm.), partly by dolomite of Upper Cenomanian age and partly by limestone of Turonian age. The second resonance peak correlated to the shallower impedance contrast varies from the calcareous sandstone of the Kurkar Gr. at the Coastal plain including the western part of the Carmel coast, clay-marl of the Saqie Gr. and Eocene-Senonian marl-chalk in Hashefela. Alluvial sediments and the Cenomanian chalk forms the shallow impedance contrast in the central and eastern parts of the Carmel Coast.  Fourier spectra and spectral ratios analysis over the study areas show good correlation with variations in the subsurface geology.  Spatial distribution of resonance frequency of both the fundamental and second resonance peaks is entirely determined by morphostructure of the appropriate reflector. At the Coastal plain, the fundamental reflector associated with the Judea Gr. is too deep to be detected using Nakamura‟s technique, so the shallower one (Kurkar) is the only reflector here. Ranges of the fundamental and second resonance frequencies are 0.2-8 Hz and 0.7- 10 Hz respectively. Ranges of H/V amplitudes of both resonance peaks are practically the same, 2-9 units.  While there is a general correlation with geological data on the reflector depth, some sharp divergences were revealed. So in the central part of the study area, areas with the increased fundamental frequency values unambiguous indicate uplifted blocks, while the geological data are opposite.  S-wave velocities for lithological units represented in the study area were derived at borehole sites using refraction survey data and verified throughout the study area. Obtained for the first time by refraction survey S-velocity of Cenomanian dolomites and Carmel chalk complex confirmed our earlier estimations. While applying the S-velocity structure to ambient noise observations, we assumed deviations in the velocity values in order to reach the best fit between analytical models and H/V ratios. Such deviations as a rule were justified by the wells lithological description and geological data;  analysis of spatial distribution of the fundamental frequency and constructing a number of the cross sections enabled more accurate tracing the earlier mapped faults and also identifying several new faults and estimating the vertical displacement;.  ambient noise studies with horizontal-to-vertical spectral ratio can yield information relevant to the field of earthquake hazard assessment and microzonation. Seismic zones 63

map is proposed based on comparative analysis of Uniform Hazard Site-specific Acceleration Spectra that meets the criterion of accepted hazard in the Israel Standard 413. Generalized model, which encompass the gross features of the geology and incorporate borehole data, is proposed for each zone.  the characteristic acceleration response spectra computed using SEEH procedure on the base of the generalized subsurface models exceed those prescribed by Israeli Code in the period range 0.2-1 sec in ten of twenty three zones.

ACKNOWLEDGEMENT

The Earth Sciences Research Administration of the Ministry for National Infrastructures sponsored this report. The study was conducted under Contract No. 28-17-054. We are most grateful to Dr. A. Hofstetter for fruitful discussion. We thank Y. Menahem for assistance in preparing this report. 64

REFERENCES Boore, D. M., 1983. Use of seismscope records to determine ML and peak velocities, Bull. Seism. Soc. Am., 74, 315-324. Boore, D. and G. Atkinson, 1987. Stochastic prediction of ground motionand spectral response parameters at hard- rock site in eastern North America, Bull. Seism. Soc., 77, 440-467. Boore, D. M., Joyner, W.B., 1991. Estimation of ground motion at deep-soil sites in the eastern North America, , Bull. Seism. Soc., 81, 2167-2185. Chávez-García, F. J., and Cuenca, J., 1998. Site effects and microzonation in Acapulco, Earthquake Spectra, 14(1): 75-93. Dravinski, M., Ding, G., and Wen, K.-L., 1995. Ground motion amplification in the San Fernando Valley from dense array measurements of microtremors, 5th International Conference on Seismic Zonation, Nice, France, 723-730. Carniel R., Barazza F. and Pascolo P., 2006. Improvement of Nakamura technique by singular spectrum analysis, Soil Dynamics and Earthquake Engineering, Volume 26, Issue 1, January 2006, 55-63 Ecker, A., 1999. Selected geological cross-sections and subsurface maps in the coastal aquifer of Israel, Atlas, Jerusalem. Enomoto T., Kuriyama, T., Abeki, N., Iwatate, T., Navarro, M., and Nagumo, M., 2000. Study on microtremor characteristics based on simultaneous measurements between basement and surface using borehole, Proc. of 12th World Conference on Earthquake Engineering, Auckland, New Zealand, 2000, No 1522, pp 8. Ezersky, M, 2009. Seismic refraction survey at the Benyamina area, GII Report No 217/450/09. Field, E. H., Hough, S. E., and Jacob, K. H., 1990. Using microtremors to assess potential earthquake site response: a case study in flushing meadows, New York City, Bull. Seism. Soc. Am., 80: 1456-1480. Field, E. H., Jacob, K. H., and Hough, S. E. 1992. Earthquake site response estimation: a weak- motion case study, Bull. Seism. Soc. Am., 82: 2283-2306 Fleischer, L., Gelberman, E. and Wolff, O., 1993. A geological-geophysical reassessment of the Judea Group, IPRG Report No. 244/147/92. Fleischer, L., Gafsou, R., 2003. Top Judea Group digital structural map of Israel. GII Report No. 873/55/99. Gaull, B. A., Kagami, H., and Taniguchi H., 1995. The microzonation of Perth, Western Australia, using microtremor spectral ratio, Earthquake Spectra, 11: 173-191. Golyandina, N.;. Nekrutkin V., and A. Zhigljavsky A., 2001. Analysis of Time Series Structure: SSA and related techniques. Hapman&Hall/CRC, New York - London. Gutierrez, C. and Singh, S. K., 1992. A site effect study in Acapulco, Guerrero, Mexico: Comparison of results from strong-motion and microtremor data, Bull. Seism. Soc. Am., 82: 642-659. Gvirtzman, G. and Reiss, Z., 1965. Stratigraphic nomenclature in the Coastal Plain and Hashefela regions, GSI Report OD/1/65. Gvirtzman, G., 1969. Subsurface data on the Saqiye Group in the Coastal Plain and Hashelphela regions, Israel, Geological Survey of Israel, OD/1/69. Gvirtzman, G., 1970. The Saqiye Group (Late Eocene to Early Pleistocene) in the Coastal Plain and Hashephela regions, Israel, Ph D. Thesis, Hebrew University Jerusalem. Gvirtzman, G. , Shachnai, E., Bakler, N. and Ilani, S., 1984. Stratigrahy of the Kurkar group (Quaternary) of the Coastal Plain of Israel, GSI Current Research 1983-1984, 70-82. Hofstetter, A., T., van Eck, A. Shapira, 1996. Seismic activity along fault branches of the Dead Sea-Jordan transform system: The Carmel-Tirtza fault system, Tectonophysics, 267: 317-330. 65

Horike, M., Zhao, B., and Kawase, H., 2001. Comparison of site response characteristics inferred from microtremors and earthquake shear wave, Bull. Seism. Soc. Am., 91; 1526-1536. Hough, S. E., Friberg, P.A., Busby, R., Field, E. F., Jacob K.H & Borcherdt R.D., 1990. Sediment-induced amplification and the collapse of the Nimitz Freeway, Nature, 344: 853- 855. Kafri, U., and Ecker A., 1964. Neogene and Quaternary subsurface geology and hydrogeology of the Zevulun Plain. Geology Survey of Israel, Bull. No 37. Kagami, H., Duke C.M., Liang, G.C., and Ohta, Y., 1982. Observation of 1- to 5-second microtremors and their application to earthquake engineering. Part II. Evaluation of site effect upon seismic wave amplification deep soil deposits, Bull. Seism. Soc. Am., 72: 987-998. Kanai, K., and Tanaka, T., 1961. On microtremors VIII., Bull. Earthquake Res. Inst. 39: 97-114. Lermo, J., and Chávez-García, F. J., 1993. Site effect evaluation using spectral ratios with only one station, Bull. Seism. Soc. Am., 83: 1574-1594. Lermo, J. and Chávez-García, F. J., 1994. Are microtremors useful in site response evaluation?, Bull. Seism. Soc. Am., 84: 1350-1364. Malagnini, L., Tricarico, P., Rovelli, A., Herrmann, R. B., Opice, S., Biella G., and de Franco, R., 1996. Explosion, earthquake, and ambient noise recording in a pliocene sediment-filled valley: inferences on seismic response properties by reference- and non-reference-site techniques, Bull. Seism. Soc. Am., 86: 670-682. Maresca, R.; Galluzzo, D.; Del Pezzo, E. H/V spectral ratio and array techniques applied to ambient noise recorded in the Colfiorito basin, Central Italy. Bull. Seism. Soc. Am. 2006, 96(2), 490-505. Mucciarelli M., Monachesi, G., 1998. A quick survey of local amplifications and their correlation with damage observed during the Umbro-Marchesan (Italy) earthquake of September 26, 1997. Journal of Earthquake Engineering, 2 (2): 325-337. Mucciarelli M., and Gallipoli, M. R., 2004. The HVSR technique from microtremor to strong motion: empirical and statistical considerations, in Proc. of 13th World Conference of Earthquake Engineering, Vancouver, B.C., Canada, Paper No. 45 Murphy, C., and Eaton, D., 2005. Empirical site response for POLARIS stations in Southern Ontario, Canada, Seismological Research Letters, V.76, No. 1, 99-109 Nakamura, Y., 1989. A method for dynamic characteristics estimation of subsurface using microtremor on the ground surface. Quarterly Report of Railway Technical Research 30(1): 25-33. Parolai, S., and Galiano-Merino, J.J., 2006. Effect of transient seismic noise on estimates on H/V spectral ratios, Bull. Bull. Seism. Soc. Am., 96, 228-236. Rodrigues P. C., 2007. Principal Component Analysis of Dependent Data , 15th European Young Statisticians Meeting (15thEYSM ) September 10-14, 2007, Castro Urdiales (Spain) Rovelli, A., Singh S. K., Malagnini, L., Amato, A., and Cocco, M., 1991. Feasibility of the use microtremors in estimating site response during earthquakes: some test in Italy, Earthquake Spectra, 7: 551-561. Schnabel, P.B., Lysmer, J., and Seed, H. B., 1972. SHAKE – a computer program for earthquake response analysis of horizontal layered sites. Rpt. No. EERC 71-12, Earthquake Engineering Research Center, Berkeley, CA. Seekins, L. C., Wennerberg, L., Margheriti, L. and Liu, Hsi-Ping, 1996. Site amplification at fife locations in San Francisco, California: a comparison of S waves, codas and microtremors, Bull. Seism. Soc. Am., 86: 627-635. Shamir, G., Bartov, Y., Sneh, A., Fleischer, L., Arad., A., Rosenshaft, M, 2001. Preliminary seismic Zonation in Israel, GSI Report No.GSI12/01, GII Report No. 550/95/01(1), pp 10. 66

Shapira, A. and van Eck, T., 1993. Synthetic uniform hazard site specific response spectrum, Natural Hazard, 8, 201-205. Shapira, A., and A. Hofstetter, 1993. Source parameters and scaling relationship of earthquakes in Israel, Teconophysics: 217-226. Shapira, A. and Avirav, V., 1995. PS-SDA Operation Manual, IPRG report Z1/567/79, 24pp. Shapira, A., Feldman L., Zaslavsky, Y., Malitzky, A., 2001. Application of a stochastic method for the development of earthquake damage scenarios: Eilat, Israel test case, The Problems of Lithosphere Dynamics and Seismicity, Computational Seismology, V.32, 58-73. Shapira, A. and Shamir, G.,1994. Seismicity parameters of seismogenic zones in and around Israel, IPRG Report Z1/567/79/(109), 22pp. Steidl J. H., Tumarkin, A.G. and Archuleta R. J. 1996. What is reference site? Bull. Seism. Soc. Am., 86: 1733-1748. Shtivelman, V., 1999. Using surface waves for estimating shear wave velocities in the shallow subsurface onshore and offshore Israel, European Journal of Environmental and Engineering Geophysics, 4:15-35. Sneh, A., Bartov, Y. and Rosensaft, M., 1998.Geological Map of Israel 1: 200,000. Sheet 2, Geological Survey of Israel. Stivelman, V., 1998. Geophysical site investigation in the Haifa port extension project area using shallow water seismic surveys, GII Report 838/49/48. Stivelman, V., 1998. P- and S-wave refraction surveys near the power stations in Haifa, Ashqelon and Hadera., GII Report No. 592/86/97. Theodulidis, N., Bard, P.Y., Archuleta, R. and Bouchon, M., 1996. Horizontal-to-vertical spectral ratio and geological conditions: the case of Garner valley downhole in Southern California, Bull. Seism. Soc . Am., 68: 767-779. Toshinawa, T., Taber J. J., and Berrill, J. B., 1997. Distribution of ground motion intensity inferred from questionnaire survey, earthquake recordings, and microtremor measurements- a case study in Christchurch, New Zealand, during 1994 Arthurs pass earthquake. Bull. Seism. Soc. Am., 87: 356-369. Y. Zaslavsky, A. Shapira, G. Ataev, M. Gorstein, T. Aksinenko, M. Kalmanovich, N. Perelman, and R. Hofstetter, 2009. Using Ambient Noise Measurements in the Process of Assessing Earthquake Hazards in Urban Areas: Examples from Israel, Authors / Editors: T. Miura and Y. Ikeda, Earthquake Engineering: New research, Nova Publisher Zaslavsky, Y., (Principal Investigator), Gorstein, M., Kalmanovich, M., Giller, V., Livshits, I., Giller, D., Dan I., Shapira, A., Fleischer, L., Leonov, J., and Peled, U., 2001. Microzoning of the earthquake hazard in Israel. Project 1: Seismic microzoning of Lod and Ramla, GII Report No 569/143/01. Zaslavsky, Y., (Principal Investigator), Shapira, A., Gorstein, M., Kalmanovich, M., Perelman, N., Giller, V., Livshits, I., Giller, D., Dan I., Aksinenko, T., Ataeva G., Shvartsburg, A., 2002. Microzoning of the earthquake hazard in Israel. Project 2: Site effects in and around Qiryat Shemona, GII Report No 569/252/02. Zaslavsky, Y. (Principal Investigator), Shapira, A., Gorstein, M., Kalmanovich, M., Aksinenko, T., Ataev, G., Giller, D, Dan I., Giller, V., Livshits, L., Perelman N., and Shvartsburg, A., 2003. Microzoning of the earthquake hazard in Israel. Project 3: Empirical determination of site effects for the assessment of earthquake hazard and risk to Kefar Sava. GII Report No 569/345/03. Zaslavsky, Y. (Principal Investigator), Hofstetter, A., Gorstein, M., Kalmanovich, M., Ataeva G., Aksinenko, T., Giller, D, Dan I., Giller, V., Livshits, I., Perelman N. and Shvartsburg, A., 2004. Microzoning of the earthquake hazard in Israel. Project 4: Empirical determination of 67

site-effect for the assessment of earthquake hazard and risk to Dimona and Arad, GII Report No 569/076/04. Zaslavsky, Y. (Principal Investigator), Ataeva G., Gorstein, M., Hofstetter, A., Kalmanovich, M., Giller, D, Dan I., Perelman N., Aksinenko, T., Giller, V., Livshits, I., and Shvartsburg, A., 2005. Microzoning of the earthquake hazard in Israel. Project 5: Empirical determination of site-effect for the assessment of earthquake hazard and risk to Beit Shean and Afula, GII Report No 569/175/05. Zaslavsky, Y. (Principal Investigator), Ataeva G., Gorstein, M., Hofstetter, A., Kalmanovich, M., Giller, D, Dan I., Perelman N., Aksinenko, T., Giller, V., Livshits, I., and Shvartsburg, A., 2006. Microzoning of the earthquake hazard in Israel. Project 6: Site-effect and seismic hazard assessment for Petah Tikva, Hod haSharon and Rosh haAyin towns. GII Report No 569/237/06. Zaslavsky, Y. (Principal Investigator), Gorstein, M., Ataeva G., Kalmanovich, M., Perelman N., Giller, D, Dan I., Aksinenko, T., Giller, V., Livshits, I., and Shvartsburg, A., 2007. Microzoning of the earthquake hazard in Israel. Project 7: Empirical determination of site- effects for the assessment of earthquake hazard and risk to the Southern Sharon and Lod valley areas. GII Report No 569/303/07. Hofstetter, A. (Principal Investigator), Ataeva G., Aksinenko, T., Dr. Y. Zaslavsky, Gorstein, M., Kalmanovich, M., Giller, D, Dan I., Perelman N., Giller, V., Livshits, I., and Shvartsburg, A., 2008. Exploration of Sedimentary Layers and Reconstruction of Subsurface Structure by Ambient Vibration measurements for Microzoning of the Afek Area. GII Report No 508/388/08. Zaslavsky, Y. (Principal Investigator), Aksinenko, T., Ataeva G., Gorstein, M., Kalmanovich, M., Giller, D, Dan I., Perelman N., Giller, V., Livshits, I., and Shvartsburg, A., 2008. Empirical Determination of Site Effects for Seismic Hazard Assessment in Kishon Graben area near Carmel Fault. GII Report No 510/389/08. Zaslavsky, Y. (Principal Investigator), Kalmanovich, M., Dr. M. Ezersky, Gorstein, M., Perelman N., Dan I., Giller, V., Ataeva G., Aksinenko, T., Giller, D, and Shvartsburg, A., 2008. Site effects from ambient noise measurements and seismic hazard assessment in Nothern Tel Aviv. GII Report No 519/401/08. Zaslavsky, Y. (Principal Investigator), Gorstein, M., Kalmanovich, M., Dan I., Perelman N., Giller, D, Ataeva G., Aksinenko, T., Giller, V., and Shvartsburg, A., 2009. Microzoning of the earthquake hazard in Israel. Project 8: Site effect and seismic hazard microzonation across the town of Tiberias. GII Report No 502/416/09. Zaslavsky, Y. (Principal Investigator), Aksinenko, T., Giller, V., Shvartsburg, A., Gorstein, M., Ataeva G., Kalmanovich, M., Perelman N., Giller, D and Dan I., 2009. Site Effects in the town of Akko: Implications for Seismic Hazard Assessment. GII Report No 552/471/09.

68

APPENDIX A. WELL DATA IN THE STUDY AREA

EW NS Number Well name TJUD 194250 734500 84 ATLIT-1 -49.(ER) 194915 733896 85 ATLIT-2 202840 708870 157 BARKAI-1 54.(T) 205040 710520 158 BARKAI-3 23 193660 714570 174 BINYAMINA -23.(T) 203570 702090 188 BQ GARBIYE -132 193000 715222 205 BT.HANANYA-1 -32.(T) 190556 713033 221 CESAR-2 479 190190 711518 222 CESAR-3 227 191535 712750 234 CS-1 -473 192275 712350 235 CS-2 -746 190940 712950 236 CS-3 -465 190320 712140 237 CS-4 -243 190340 712440 238 CS-5 -262 193620 714800 239 CS-6 -22.(T) 206620 720950 308 DALIA-1 -159 206100 721620 309 DALIA-2 -191 202060 707500 392 EN SHEMER.A 25 201280 707270 393 EN SHEMER.B -39 207250 718700 445 GAL-ED-2 -175 205430 718780 446 GAL-ED-3 -300 198900 712400 483 GV.ADA -563.(E) 189509 705868 536 HADERA-1 -1300.(E) 193830 719090 563 HF CARMEL-8 -16.(T) 193000 720000 564 HF.CARMEL-7 -34.(T) 192370 717420 662 KABARA-3 -28.(T) 192020 716720 663 KABARA-4A -32.(T) 193900 716400 664 KABARA-84 -4.(T) 200570 707090 671 KARKUR-101 -64 200660 707350 672 KARKUR-145 -82 200580 706530 673 KARKUR-146 88.(T) 199780 707370 674 KARKUR-38 -143 191750 715720 805 M.MIHAEL-4 -34.(T) 201250 703270 814 MA'ANIT-1 -173.(E) 200930 702140 815 MA'ANIT-2 -306 200580 701320 816 MA'ANIT-3 -338 201840 704590 817 MA'ANIT-4A -54 200940 704170 818 MA'ANIT-5A -170 202450 707380 819 MA'ANIT-A 46.(T) 202460 706150 820 MA'ANIT-B -23 198630 707970 837 MENASHE T-1 -578 202120 706800 838 MENASHE T-2 3 195800 711900 839 MENASHE T-3 -840 204780 705260 840 METZER -20 194600 719800 848 MN ZVI -8.(T) 194610 722140 849 MN ZVI-6 -27.(T) 200590 721930 881 N.TUT-1 -460 202266 721994 884 NAHAL TUT-4 -478.4 201300 708480 931 P.HANA KR. -61 195280 715460 1094 SHUNI-2 60 195700 716330 1095 SHUNI-3 28 69

EW NS Number Well name TJUD 200160 705430 1126 SR.MEN-1 -223 202030 705890 1127 SR.MEN-2 -82 191420 715780 1182 TANIN T-2 -38.(T) 192550 715850 1183 TANIN T-3 -23.(T) 191450 717350 1184 TANIN T-4 -37.(T) 193300 717500 1185 TANIN T-5 -18.(T) 199754 721574 1197 TIRELY-7 -512 202370 709030 1260 V.ARA-1 23 204450 709740 1261 V.ARA-2 63.(T) 204520 710090 1262 V.ARA-3 71 198220 701770 1309 ZEITA -455 194710 720790 1321 ZN.YAAKOV -15.(T) 194100 718450 1322 ZN.YAAKOV-3 -12.(T)