Microtremor measurements in Yanbu city of Western Saudi Arabia: A tool for seismic microzonation

1 Introduction

Yanbu city represents one of marine ports of western Saudi Arabia where it is located along the Red Sea eastern coast. It belongs to Al-Madinah province about 350 km north of Jeddah city (Fig. 1).

Close

Figure 1

Location map for Yanbu city.

Export to PPT

The central part of the city refers to the old town, while modern Yanbu is a short distance away. In 1975, the Saudi government selected both Yanbu and Al-Jubail, both small towns on the Persian Gulf coast, to be developed as modern industrial cities and established the Royal Commission for Jubail and Yanbu. The town is expanding as an industrial and commercial center with three large oil refineries, a petrochemical complex, and a large desalination plant that plays an important role in diversifying Saudi Arabia's economic base.

The behavior of the ground motion during an earthquake is generally well explained by the geological surface structure in the place where the phenomenon is studied. Past and recent observations have shown that the damage caused by strong earthquakes is more important in sedimentary basins than on hard rock structures. Unfortunately, Yanbu city lies within the Red Sea's active tectonic environment and is affected by the present-day geodynamic processes acting in the Red Sea region. The resulting structures either normal or transform faults run parallel to and/or across the Red Sea. Some of these faults extend inland over tens or hundreds of kilometers (Al-Shanti, 1966; Pallister, 1984). The relative movements along such faults can cause large and damaging earthquakes. Historical information (Poirier and Taher, 1980; Ambraseys and Milville, 1983; Ambraseys et al., 1994) in association with the recent studies (Merghelani, 1981; El-Isa and Al-Shanti, 1989; Al-Amri, 1995) around Yanbu have revealed its significant level of earthquake activity which should be taken into account for the strategic plans in the future. Recently, the city was affected by the occurrence of the moderate earthquake at Harat Lunayyir (M_w 5.7) earthquake swarm on May 19, 2009 (Al-Amri and Fnais, 2009). Earthquake ground-shaking intensity has been participated at Yanbu city (MMI = 5) in spite of 130 km distance from the earthquake location.

It is well known that, local site effects play an important role in the damage occurring during a destructive earthquake (Singh et al., 1988; Graves, 1993; Bakir et al., 2002; Sørensen et al., 2006). The lateral variation in the site effects are mainly due to the local site conditions such as type and nature of sediments, water saturation, basin geometry, and thickness of sediments. In addition, it is realized that the presence of sharp lithological boundaries between the bedrocks and the overlying sediments cause strong impedance contrasts that affect the local site responses. Different approaches have been carried out to evaluate the local site response (Nakamura, 1989; Lermo and Chavez-Garcia, 1993; Field and Jacob, 1995). In this study, the evaluation of the local site effects for Yanbu metropolitan city was conducted according to the most popular and world-wide approach through the calculation of the horizontal to vertical spectral ratio (HVSR) for the recorded ambient vibrations. These spectral ratios show the fundamental frequency and the associated amplification factor. As proved through the intensive investigations, thick sedimentary cover usually generates amplitude spectra with peaks at low frequencies (<3 Hz) while thin sediments generate peaks at high frequencies (>10 Hz). The broad spectrum with constant amplitude is usually associated with the crystalline outcrops (Nakamura, 1989, 2000; Bard, 1994; Bard et al., 2004). Then, the estimated fundamental frequencies and the corresponding amplifications are compared with the surface geology.

2 Geological and tectonic setting

The available geological information for Yanbu city and the surrounding area comes from the geologic map of Yanbu (Fig. 2) with scale 1:250,000 (Pellaton, 1979). Most of the area is occupied by outcrops of tertiary–quaternary marine and continental coastal plain sediments. Marine quaternary deposits are represented mainly by reef terraces which lie several meters above sea level. Sand and mud in the lower zones are intermixed with contemporaneous alluvial material. While continental quaternary deposits are represented by: (1) sandy mantle covering a wide area which has a composite origin incorporating fluviatile and Aeolian transport; (2) gravely or sandy spreads dissected by very close drainage; and (3) gravely spreads related to the degradation of the older terraces.

Close

Figure 2

Geologic map for Yanbu area.

Export to PPT

The distribution of tertiary sediments was controlled by the syndepositional section across the coastal plain from the Precambrian basement in the east to the coast in the west would be: (1) beds of shale and conglomerate sandstone alternating with siltstone, (2) thin tertiary reefal limestone with the Miocene coral, rising above the coastal plain and alternating with sandstone and silts, (3) shale covered by sand and gravel, and (4) raised reef limestone terraces of quaternary age.

Depending on the lithological variations through 106 boring logs in the city of Yanbu (Al-Haddad et al., 2001), the subsurface soil column at Yanbu city can be classified as an extremely variable complex and is composed of the following profiles:

• Extremely loose to loose fine to medium sand deposits from the surface to 15 m depth with a very shallow groundwater table (less than 1 m).

• Thin top layer of compact shattered coral (2–5 m with SPT-N values from 20 to 50) underlain by thick layer of very loose to loose fine sand with some gravel and/or silt (6–8 m with SPT-N values between 5 and 25) below the ground level.

• A dense to very dense layer of silty sand with gravel and cobbles.

• Coral limestone interbedded with coralline sand and fines. It is of medium dense to very dense especially at deeper depths (35 km south of Yanbu).

3 Data acquisition and processing

Seismic noise measurements were acquired at 85 sites during the period from 3rd to 20th August 2009 at the intersected points of two mutually perpendicular sets of profiles covering the metropolitan zone of Yanbu city (Fig. 3 and Table 1).

Close

Figure 3

Distribution of microtremor measurements at Yanbu area.

Export to PPT

Noise measurements were collected (Fig. 4) using five digital stations of 24-bit Quanterra Q₃₃₀ digitizer equipped with high-performance portable Very Broad-Band (VBB) triaxial Streckeisen STS-2 velocimeter (with flat response from 8.33 mHz up to 50 Hz) with GPS timing. Generally known is the fact that man-made seismic noise represents a problem for seismic noise survey especially within densely urbanized areas in the form of traffic or industrial activities. To avoid recording of such transient signals into the measurements, special precautions should be taken whenever possible to measure as far as possible away from sites close to heavily traffic streets, boulevards and heavy-duty machines. In some cases, there are no possibilities in finding an appropriate site, then the measurements should be acquired in the early morning when there is relatively little human activity throughout the city. During the acquiring period of data the used seismometers have been installed in the early morning to collect the desired data after twenty-four hours of continuous recording. The measuring points are spread out within a range of 450 m vertically and 900 m horizontally forming a regular grid to achieve good quality of the contour maps for the fundamental frequency and amplification factor (Fig. 3).

Close

Figure 4

Microtremor record at Yanbu city.

Export to PPT

Throughout the current work, all the sensors used were calibrated, installed in good coupling with soil, isolated thermally well against temperature changes using thick foam box and covered to reduce the interference of wind. Then, these sensors were oriented horizontally and vertically leveled. Ambient seismic noise was recorded with 200 Hz sampling rate for 24 h of continuous recording. This long duration of recording guarantees the statistical stabilization of the signal (Picozzi et al., 2005, 2008). All the experimental conditions for the current work were controlled mainly by the precautions of European SESAME research project (Chatelain et al., 2008; Guillier et al., 2008). Keeping in mind, higher fundamental resonance frequencies would be observed on sites having thin sediments, whereas the lower values of fundamental frequencies would be observed over the thick sediments (Nakamura, 1989; Atakan 1995; Lacave et al., 1999; Bard et al., 2004).

The collected data have been processed using the J-SESAME software developed within the framework of the great European project SESAME. The H/V spectral ratios were computed according to the following methodology: (1) base-line correction; (2) band-pass filtering to retain the frequencies in the range from 0.2 to 20 Hz; (3) windows of 50 s length were automatically selected using an anti-STA/LTA trigger algorithm and tapered with a 5% cosine function before the computation of spectra; (4) Fourier spectra were calculated for each noise component and smoothed using Konno–Omachi window having a smoothing constant b-value of 30; (5) the resulting spectral amplitudes of horizontal components were geometrically averaged and divided by the vertical spectra to calculate the H/V function. Moreover, stability of the peak in the H/V curve have been checked through three tests conducted at every measuring point, where the same frequency peaks were picked for each site at different times. Sensitivity tests were also carried out on the selected time window lengths during the data processing; the results declared low dependence of the window length and therefore, a high stability was achieved.

The reliability of the actual H/V curve obtained with the selected recordings should be tested throughout the following consequences: (i) for a peak to be significant, it is recommended checking that the following condition is fulfilled: f₀ > 10/l_w (window length). This condition means that, at the frequency of interest, at least 10 significant cycles in each window should be present, (ii) large number of windows and of cycles is needed. The total number of significant cycles: n_c = l_w · n_w · f₀ be larger than 200, and (iii) an acceptably low level of scattering between all windows is needed. Large standard deviation values often mean that ambient vibrations are strongly non-stationary and undergo some kind of perturbations, which may significantly affect the physical meaning of the H/V peak frequency. Therefore it is recommended that σ_A(f) be lower than a factor of 2 (for f₀ > 0.5 Hz), or a factor of 3 (for f₀ < 0.5 Hz), over a frequency range at least equal to [0.5f₀ and 2f₀].

Generally, as noticed in urban environments that, the H/V curves exhibit local narrow peaks – or troughs of industrial origin, related to some kind of machinery (turbine and/or generators). Such perturbations are recognized by the following general characteristics: (i) they may exist over a significant area (in other terms, they can be seen up to distances of several kilometers from their source), (ii) as the source is more or less “permanent” (at least within working hours), the original (non smoothed) Fourier spectra should exhibit sharp narrow peaks at the same frequency for all the three components, (iii) reprocessing with less and less smoothing: in the case of industrial origin, the H/V peak should become sharper and sharper (while this is not the case for a site effect peak linked with the soil characteristics), (iv) if other measurements have been performed in the same area, determine whether a peak exists at the same frequencies with comparable sharpness (the amplitude of the associated peak, even for fixed smoothing parameters, may vary significantly from site to site, being transformed sometimes into a trough), and (v) another very effective check is to apply the random decrement technique (Dunand et al., 2002) to the ambient vibration recordings in order to derive the “impulse response” around the frequency of interest: if the corresponding damping (z) is very low (below 1%), an anthropic origin may be assumed almost certainly, and the frequency should not be considered in the interpretation.

4 Results and discussion

H/V spectral ratios for the selected windows have been computed at 85 of surveyed sites (Fig. 5a–c). The presence of clear peak of H/V curve (Fig. 5a) is considered as an indicative of the impedance contrast between the uppermost surface soil and the underlying hard rock, where large peak values are generally associated with sharp velocity contrasts (Bard et al., 2004) and is likely to amplify the ground motion. Fig. 5b illustrates two of peaks that reflect the presence of two large impedance contrasts while, Fig. 5c present some of broad peak or plateau-like curve and this could be related to the presence of an underground sloping of the interface between softer and harder layers. These results declared that the underground structure of the site exhibits significant lateral variations in thickness and dynamic properties, which lead to a significant 2D or 3D effects.

Close

Figure 5

(a) Examples of clear peak of spectral ratio (H/V) curves. (b) Examples of two peaks of spectral ratio (H/V) curves. (c) Examples of multiple and broad peaks of spectral ratio (H/V) curves.

Export to PPT

Close

Figure 5

(a) Examples of clear peak of spectral ratio (H/V) curves. (b) Examples of two peaks of spectral ratio (H/V) curves. (c) Examples of multiple and broad peaks of spectral ratio (H/V) curves.

Export to PPT

According to Fig. 6, the resonance frequency f₀ varies from 0.25 to 7.9 Hz within the study area. The central part has lower values of f₀ (less than 1 Hz) where 50% of f₀ values lie in this area (Fig. 7). These low values indicate large thickness of sediments. While, there are some localized areas that have higher values of f₀ (up to 6.7 Hz) distributed throughout the southeastern, southern, and southwestern parts of the area. The presence of high f₀ at the southern part is limited to Al-Majd Sporting Club, Al-Maktaba, and Al-Shati secondary school sites. Parolai et al. (2001) stated that, the resonance frequency becomes lower in areas where the basement depth is greater and higher where it is shallower. Accordingly, the presence of higher and lower values reflect variation in the thicknesses of sediments through the area suggesting 2D or 3D basins in-between (Fig. 8). In general, the values of f₀ increase due north.

Close

Figure 6

Contour map for f₀ for Yanbu city.

Export to PPT

Close

Figure 7

Relation between number of records and their values of f₀.

Export to PPT

Close

Figure 8

3D model for the subsurface basement at Yanbu city.

Export to PPT

Contour map for H/V amplitude ratio A_max (Fig. 9) illustrated that, the eastern part of the area has values less than 2.5 while the western part has values less than 3. Whereas these values increases towards the central zone of Yanbu metropolitan where it has a value of 5. By referring to the surface soil distribution it can be observed that the central part characterized by the thick section of soft sediments (sabkhat) amplifies the ground-shaking intensity five times more than that of basement rocks. However, it is to be noted that these values give the lower bound estimates of amplification of H/V ratio for a given site. Generally, it could be mentioned that, there is significant variation in this parameter within the study area.

Close

Figure 9

Contour map for H/V amplitude ratio A_max for Yanbu city.

Export to PPT

Several studies show the relationship between the velocity structure beneath the recording site and the fundamental frequency obtained from HVSR (horizontal-to-vertical spectral ratios of microtremors) analysis (Ibs-von Seht and Wohlenberg, 1999; Parolai et al., 2002; Motamed et al., 2007). We used the sedimentary layer parameters obtained by Al-Haddad et al. (submitted for publication) through shear-wave velocity profiles, where a sample of ten boreholes have been drilled through the study area. Three of these boreholes are presented within the area of interest (Table 2). The weighted shear-wave velocities were computed according to the following formula: $$V_{\text{s}(\text{av})}=\frac{{\sum }_{i=1}^n{d}_i}{{\sum }_{i=1}^n\frac{d_{\text{i}}}{V_{\text{si}}}}$$ where d_i and v_si denote the thickness (in m) and the shear-wave velocity (in m/s) of the ith layer, in a total of n layers, existing in the same type of stratum (d_i and v_si were determined by borehole measurements).

Then, the values of fundamental resonance frequency can be computed at these three localities through the application of the following equation (Bard, 2000): $$f=\frac{V_{\text{av}}}{4H}$$ where H is the total thickness (in m) of the sediments overlying the half-space and V_av is the average shear-wave velocity (in m/s) in the sediments. Using the above equation and the velocity model in Table 2, the resonance frequency f₀ at these sites is estimated to be equal to 6.7 Hz at Al-Shati secondary school; 4.8 Hz at Al-Maktaba, and 6.3 Hz at Al-Majd Sporting Club where V_av values are 467.83 m/s; 566.04 m/s and 752.74 m/s respectively. On the other hand, the resonance frequency f₀ estimated using the microtremor measurements, close to these boreholes, are 6.7 Hz, 5.2 Hz, and 6.5 Hz, respectively (Fig. 10) that reflects an excellent agreement between the estimated values for f₀ from two independent techniques.

Close

Figure 10

H/V results at three of geotechnical boreholes at Yanbu city.

Export to PPT

5 Conclusions

Although Yanbu city is located close to Red Sea floor spreading active zone; it is not affected by recent instrumentally strong earthquakes. Moreover, most of the city is built on thick soft sediments, which could reasonably amplify the earthquake ground motion in the case of an event. No site effects analyses have been carried out in the city, therefore, it is believed that a detailed study of the local response of Yanbu city should be of some concern in terms of seismic hazard. The current work represents the first attempt to derive useful information on local ground motion amplification in the urban area of Yanbu.

Analysis of microtremor measurements show that the resonance frequency varies considerably through Yanbu city. Values of resonance frequency increase as the basement depth decreases. The central part of the city has low values of f₀ as compared to other parts of the city suggesting basinal shape with great thickness of sediments. The results of microtremor data are in congruence with the shear-wave velocity profiles through Yanbu city.

Finally, it is declared that, the using of microtremor measurements represents a powerful tool for microzonation studies in spite of their tendency to underestimate the level of ground motion amplification when compared with earthquake records.