Site effects of the Roio basin, L��Aquila

During the microzonation studies of the April 6th, 2009 L??Aquila earthquake, we observed local seismic amplifications in the Roio area??a plane separated from L??Aquila city center by mount Luco. Six portable, digital instruments were deployed across the plain from 15 April to mid-May 2009. This array recorded 152 aftershocks. We analyzed the ground motion from these events to determine relative site amplification within the plain and on surrounding ridges. Horizontal over vertical spectral ratio on noise data (HVSRN), aftershock recordings (HVEQ) and standard spectral ratio (SSR) showed amplifications at 1.3 and 4.0?Hz on quaternary deposits. Seismic amplifications in the frequency range of 4 and 6?Hz were also observed on a carbonate ridge of Colle di Roio, on the northwestern border of the plateau. A small amplification was noticed near the top of mount Luco, another rocky site. Large discrepancies in the amplification levels between methods have been observed for these sites, but the HVSRN, HVEQ and SSR gave similar results at the stations located in the Roio plain. On the rocky sites, the SSR was more reliable than the HVSRN at estimating the transfer function of the site, even if the resonance frequency seemed to be well detected by the latter method.     

Topographic site effects of Xishan Park ridge in Zigong city,Sichuan considering epicentral distance

Ground motion amplification induced by topography plays a vital role in engineering seismology. A topographic array of 8 accelerographs has been operating along the ridge in Xishan Park since 2007. The topographic site effects in Zigong city are studied based on the strong motion data of 2008 M_s 8.0 Wenchuan earthquake (the epicentral distance?=?225 km) and 2019 M_s 5.2 Zizhong earthquake (the epicentral distance?=?29 km). We compare the peak ground acceleration (PGA) of the two earthquakes and find that the PGA of Station 7#, which locates on a relatively steep slope, is amplified by 4.41 times comparing with the reference station in Zizhong earthquake, while this value is only 1.62 in Wenchuan earthquake. Fourier amplitude spectrum shows that the high frequency content of Zizhong earthquake is more abundant because of its smaller epicentral distance. By using the standard spectral ratio (SSR) method, we conclude that the amplification occurs because high-frequency ground motion is likely to resonate at small-scale features. Finally, the 3D numerical simulations are used to verify these conclusions. Our work indicates that more sophisticated numerical models need to be established for more accurate topographic site effects quantification. In addition, the influence of nearby topographic features should be considered when selecting reference stations.

     

Evaluation of site effects in the Aterno river valley (Central Italy) from aftershocks of the 2009 L��Aquila earthquake

A temporary network of 33 seismic stations was deployed in the area struck by the 6th April 2009, Mw 6.1 (Scognamiglio et?al. in Seism Res Lett 6/81, 2010), L??Aquila earthquake (central Italy), with the aim to investigate the site amplification within the Aterno river Valley. The seismograms of 18 earthquakes recorded by 14 of the 33 stations were used to evaluate the average horizontal to vertical spectral ratio (HVSR) for each site and the standard horizontal spectral ratio (SSR) between a site and a reference station. The obtained results have been compared to the geological and geophysical information in order to explain the resonance frequencies and the amplification levels with respect to surface geology of the valley. The results indicate that there is no uniform pattern of amplification, because of the complex geologic setting, as the thickness and degree of cementation of the deposits is highly variable.     

Experimental study of seismic behavior of two hilly sites in Tehran and comparison with 2D and 3D numerical modeling

Two hilly sites were selected to study seismic site response due to topography effects. The sites were selected in a manner to be as much as possible homogenous and free of the soft soil layers effects. The hills were instrumented by nine velocimetric stations to record microtremors and the obtained data were analyzed using horizontal to vertical spectral ratios. Some standard spectral ratio tests were performed on noise as well. Then the instrumented hills were modeled (both 2D and 3D) assuming a linear elastic constitutive behavior subjected to vertically propagating SV and P Ricker wavelets. All calculations were performed in time domain using direct boundary element method. Different transfer function components, amplification patterns and spectral ratios were calculated in frequency domain. The frequency of vibration, obtained by experimental studies, is between 4 and 5 Hz for both of the hills. The spectral ratios derived by numerical simulations were compared with the observed spectral ratios. They show relatively good similarities between the results of these two methods. The frequencies of vibration derived from different methods seem to be nearly identical. The agreement in term of resonance frequency between microtremors and numerical modeling suggests that noise measurements could represent a simple, even if preliminary, tool in order to identify possible topographic amplification.     

Seismic characterization and monitoring of Fucino Basin (Central Italy)

The Fucino basin (Central Italy) is one of the largest intramountain alluvial plain in the Apennines range. It has a tectonic origin related to the presence of important systems of faults located in its northern and eastern edges. Some of these faults are still active and capable of generating strong seismic events. Site effects related to the soft soils filling the basin can be very important. In this paper we show the preliminary results of a seismic network installed in the Fucino area in order to collect information about site amplification effects and geometry of the basin. We analyze ambient seismic vibrations and recordings of about 150 local earthquakes mainly related to the seismic sequence of the April 6th 2009 Mw 6.3 L’Aquila event. Moreover the strongest events of L’Aquila sequence were analyzed at the three permanent strong-motion stations operating in the area. Using standard spectral techniques we investigate the variation of resonance frequencies within the basin. The ground motion recorded in the Fucino plain is mainly characterized by strong energy at low-frequencies (f < 1 Hz) affecting both horizontal and vertical components. This is particularly evident for stations deployed in correspondence of very thick deposits of sedimentary filling, where a significant increase of ground-motion amplitude and duration is likely caused by locally generated surface waves. The amplification at low-frequencies (<1 Hz) on the horizontal components can reach up a factor of 10 in comparison to nearby stiff sites. However, we found evidences of seismic amplification phenomena also for stiff sites surrounding the basin, including stations of the Italian strong motion network. The independent geological information and the shallow shear-velocity profiles available for the basin can be combined with resonance frequencies for deriving representative geological sections to be used as base for future numerical 2D–3D modeling of the basin.     

3D Response analysis of an instrumented hill at Matsuzaki, Japan, by a spectral method

Strong ground motion observed at an instrumented hill site is first analysed through the standard (SSR) and the horizontal-to-vertical (HVSR) spectral ratio techniques. A reasonable agreement is found between these approaches. The observations are then compared with 3D numerical simulations, performed with a highly efficient numerical code based on a spectral method, that allowed for reasonable computer times also on a PC. The observed amplification is significantly higher than that computed with a 3D homogeneous model of the mountain, suggesting that local response is governed by large-scale and small-scale soil heterogeneities rather than by topographic site effects. The introduction of a local near-surface inclusion of nonhomogeneous soil material under one of the recording stations has not significantly improved the numerical results. The observed data are also compared with the results of simplified simulations, either using 2D homogeneous models or coupling the 3D response with a 1D local soil profile. The results of such simplified approaches are discussed and their usefulness is emphasised.     

Seismic response of deep Quaternary sediments in historical center of L’Aquila City (central Italy)

On 6 April 2009 a M_w=6.1 earthquake produced severe destruction and damage over the historic center of L’Aquila City (central Italy), in which the accelerometer stations AQK and AQU recorded a large amount of near-fault ground motion data. This paper analyzes the recorded ground motions and compares the observed peak accelerations and the horizontal to vertical response spectral ratios with those revealed from numerical simulations. The finite element method is considered herein to perform dynamic modeling on the soil profile underlying the seismic station AQU. The subsurface model, which is based on the reviewed surveys that were carried out in previous studies, consists of 200–400 m of Quaternary sediments overlying a Meso-Cenozoic carbonate bedrock. The Martin-Finn-Seed's pore-water pressure model is used in the simulations. The horizontal to vertical response spectral ratio that is observed during the weak seismic events shows three predominant frequencies at about 14 Hz, 3 Hz and 0.6 Hz, which may be related to the computed seismic motion amplification occurring at the shallow colluvium, at the top and base of the fluvial-lacustrine sequence, respectively. During the 2009 L’Aquila main shock the predominant frequency of 14 Hz shifts to lower values probably due to a peculiar wave-field incidence angle. The predominant frequency of 3 Hz shifts to lower values when the earthquake magnitude increases, which may be associated to the progressive softening of soil due to the excess pore-water pressure generation that reaches a maximum value of about 350 kPa in the top of fluvial-lacustrine sequence. The computed vertical peak acceleration underestimates the experimental value and the horizontal to vertical peak acceleration ratio that is observed at station AQU decreases when the earthquake magnitude increases, which reveals amplification of the vertical component of ground motion probably due to near-source effects.     

Experimental study of the ground motion on a large scale topographic hill at Kitherion (Greece)

We present the results of an experiment about the effects on the seismic ground motion of a hill of large dimensions (6km long, 3km large and 700m high) near Corinth (Greece). We installed 7 seismometers across the hill, and analyzed the ground motion with different methods: the classical spectral ratios (CSR) and the horizontal to vertical spectral ratios calculated both on noise (HVNR) and earthquake data (RF). The amplification at the resonant frequency (around 0.7Hz) is low, as expected, but one of the two stations installed at the top exhibits large amplification (up to a factor of 10) at 3Hz. The comparison between the H/V results and the CSR ones suggests that the former method is able to point out the fundamental frequencies of a hill. Amplifications and spatial localization predicted by numerical 3D modeling are consistent with the experimental data only at frequencies lower than 1Hz.     

Separation of source and site effects by generalized inversion technique using the aftershock recordings of the 2009 L��Aquila earthquake

We exploit S-wave spectral amplitudes from 112 aftershocks (3.0 ≤ M_L ≤ 5.3) of the L’Aquila 2009 seismic sequence recorded at 23 temporary stations in the epicentral area to estimate the source parameters of these events, the seismic attenuation characteristics and the site amplification effects at the recording sites. The spectral attenuation curves exhibit a very fast decay in the first few kilometers that could be attributed to the large attenuation of waves traveling trough the highly heterogeneous and fractured crust in the fault zone of the L’Aquila mainshock. The S-waves total attenuation in the first 30 km can be parameterized by a quality factor Q_S(f) = 23f ^0.58 obtained by fixing the geometrical spreading to 1/R. The source spectra can be satisfactorily modeled using the omega-square model that provides stress drops between 0.3 and 60 MPa with a mean value of 3.3±2.8 MPa. The site responses show a large variability over the study area and significant amplification peaks are visible in the frequency range from 1 to more than 10 Hz. Finally, the vertical component of the motion is amplified at a number of sites where, as a consequence, the horizontal-to-vertical spectral ratios (HVSR) method fails in detecting the amplitude levels and in few cases the resonance frequencies.     

Two and Half Dimensional Simulation of Ridge Effects on the Ground Motion Characteristics

— Seismic responses of weathered and non-weathered ridge models were simulated to study the ridge effects on the ground motion characteristics. The range of ridge slope from 19.98° to 45° was considered to produce a possible set of generalized results. 2.5-D modeling based on parsimonious staggered grid approximation of elastodynamic wave equations was adopted in simulations. Computed results reveal an increase of amplitude of incoming waves with both elevation and the slope of the ridge. Further, the characteristics of surface waves are highly ridge slope dependent. The analysis of responses of weathered and non-weathered ridge models reveals that ridge has caused a strong generation of surface waves near its top. The surface waves are not dominating on the top of the ridge but at some lower elevation. The increase of weathering of ridge further intensified the ridge effect. Analysis in frequency domain, based on spectral ratio method, does not indicate any pattern in the spectral amplification factor and is very much sensitive to slope, source focal mechanism and location. However, on an average there is a continuous decrease of amplification with slope in the vertical component and increase in the transverse component, and it is increasing in the radial component up to slope =38.0° and thereafter decreasing.     

Seismic response of L��Aquila downtown from comparison between 2D synthetics spectral ratios of SH, P-SV and Rayleigh waves and observations of the 2009 earthquake sequence

This is the first part of a study on the seismic response of the L’Aquila city using 2D simulation and experimental data. We have studied two velocity-depth models with the aim of outlining the behavior of a velocity reversal in the top layer, which is associated with the stiff Brecce de L’Aquila unit (BrA). In this setting, the SMTH model is topped by a layer with about 2:1 impedance contrast with the underlying layer while the NORV model has no velocity reversal. We have simulated the propagation of SH and P-SV wavefields in the range 0–10 Hz for incidence 0°–90°. Earthquake spectral ratios of the horizontal and vertical components at six sites in L’Aquila downtown are compared to corresponding synthetics spectral ratios. The vertical component of P-SV synthetics enables us to investigate a remarkable amplification effect seen in the vertical component of the recorded strong motion. Sites AQ04 and AQ05 are best matched by synthetics from the NORV model while FAQ5 and AQ06 have a better match with synthetics spectral ratios from the SMTH model. All simulations show this behavior systematically, with horizontal and near-horizontal incident waves predicting the overall pattern of matches more clearly than vertical and near-vertical incidence. The model inferences are in agreement with new geological data reporting lateral passages in the top layer from the stiff BrA to softer sediments. Matches are good in terms of frequency of the first amplification peak and of spectral amplitude: the horizontal components have spectral ratio peaks predominantly at 0.5 Hz in the simulations and at 0.7 Hz in the data, both with amplitudes of 4, while the vertical component spectral ratios reach values of 6 at frequencies of about 1 Hz in both data and simulations. The vertical component spectral ratios are very well matched using Rayleigh waves with incidence at 90°. The NORV model without the velocity reversal predicts spectral ratio peaks for the horizontal components at frequencies up to 6 Hz. The reversal of velocity acts as a low-pass frequency filter on the horizontal components reducing the amplification effect of the sediment filled valley.     

Site effects in the Aterno River Valley (L’Aquila,Italy): comparison between empirical and 2D numerical modelling starting from April 6th 2009 Mw 6.3 earthquake

In this paper, we focused our attention on a cross-section of the Aterno River Valley where a good quality geological and geophysical dataset allowed to reconstruct accurately the geometry and the Vs profiles along all the plane of the section. Its trace is deliberately aligned close to the strong motion stations that recorded the Mw 6.3 (April 6th 2009) L’Aquila earthquake. We analysed strong and weak motion data available at these latter stations as well as at one of the temporary stations installed during the Microzonation activities and located on outcropping bedrock, in proximity of the cross-section. We used the H/V technique to select a reliable reference site and once we found it, we applied the SSR technique to compute amplification functions in correspondence of two strong motion stations. In turn, for both sites we performed a site response numerical modelling with two different 2D codes and we compared simulated versus experimental transfer functions. We found that the cross-section is well constrained based on the very reasonable agreement between results of numerical modelling and earthquake data analysis. We pointed out also a strong amplification of the deposit at the centre of the valley due to the constructive interference of S and surface waves, not predictable by means of 1D numerical modelling. We also compared the H/V as well as the SSR obtained from strong motion data with the ones computed from weak motion finding evidences of non-linearity in soil behaviour.     

Site Response in Las Vegas Valley, Nevada from NTS Explosions and Earthquake Data

We report site response in Las Vegas Valley (LVV) from historical recordings of Nevada Test Site (NTS) nuclear explosions and earthquake recordings from permanent and temporary seismic stations. Our data set significantly improves the spatial coverage of LVV over previous studies, especially in the northern, deeper parts of the basin. Site response at stations in LVV was measured for frequencies in the range 0.2–5.0 Hz using Standard Spectral Ratios (SSR) and Horizontal-Vertical Spectral Ratios (HVR). For the SSR measurements we used a reference site (approximately NEHRP B ``rock' classification) located on Frenchman Mountain outside the basin. Site response at sedimentary sites is variable in LVV with average amplifications approaching a factor of 10 at some frequencies. We observed peaks in the site response curves at frequencies clustered near 0.6, 1.2 and 2.0 Hz, with some sites showing additional lower amplitude peaks at higher frequencies. The spatial pattern of site response is strongly correlated with the reported depth to basement for frequencies between 0.2 and 3.0 Hz, although the frequency of peak amplification does not show a similar correlation. For a few sites where we have geotechnical shear velocities, the amplification shows a correlation with the average upper 30-meter shear velocities, V₃₀. We performed two-dimensional finite difference simulations and reproduced the observed peak site amplifications at 0.6 and 1.2 Hz with a low velocity near-surface layer with shear velocities 600–750 m/s and a thickness of 100–200 m. These modeling results indicate that the amplitude and frequencies of site response peaks in LVV are strongly controlled by shallow velocity structure.     

Urban seismic stations: soil–structure interaction assessment by spectral ratio analyses

In this work we present and discuss the results of ambient seismic noise analyses computed at four sites where seismic stations, managed by the INGV (Italian Institute for Geophysics and Vulcanology) and the DPC (Italian Department of Civil Protection), are installed inside buildings. The experiments were performed considering different types of installation: sensor located at the bottom of a school, directly installed on rock (case 1); sensor located at the bottom of a medieval fortress, built on an isolate hill, directly installed on rock (case 2); sensor installed on the foundations of a medieval fortress, built on an isolate hill (case 3); sensor installed on the foundations of a school, built on alluvial deposits (case 4). Since recent works proposed the use of spectral ratio techniques to study the dynamic characterization of buildings, ambient seismic-noise measurements were performed for each site close to the stations (at the base of the structures), at the top of the structures and outside the buildings. In order to check the source of vibrations both horizontal to vertical spectral ratio (HVNR) and standard spectral ratio (SSR) techniques were applied. For all stations the results from ambient seismic noise were compared to those obtained from earthquakes (HVSR). In order to detect preferential directions of amplification, for each site average HVNRs and HVSRs were computed considering one azimuth for each set of 5°. We obtain different results for different types of installation: in cases 1 and 2, where the sensors are directly installed on rock, the vibrations of the structure do not affect the noise measures performed close to the stations, which show flat HVNR in the whole frequency range: in both cases the eigenfrequency of the building is given by the HVNR calculated from the measures performed at the top of the structure. In cases 3 and 4, where the sensors are installed on the foundations of the considered structures, both the amplification peaks between 5 and 9 Hz (case 3) and between 5.5 and 7 Hz (case 4) include the contribution of the free oscillations of the buildings. In particular, in case 4, HVNRs performed outside building highlight possible soil–structure resonance effects in case of an earthquake.     

Site Effect and its Relationship to the Intensity and Damage Observed in the June 27, 1998 Adana-Ceyhan Earthquake

The site response at 15 stations in the Adana-Ceyhan region (Southern Turkey) is calculated from the recordings of aftershocks of June 27, 1998 Adana-Ceyhan earthquake (M_S=6.2) by using the Standard Spectral Ratio (SSR) and the Horizontal-to-Vertical Spectral Ratio (HVSR) methods. While the two methods are in good harmony at a few stations in determining the site effects, they show differences on the estimated amplifications or on the site resonance frequencies at most stations. It was not clear which one of the two methods underestimates or overestimates the amplification values. We observe that at some stations, where the local site conditions are rather complex, the vertical component records are strongly influenced from the local soil conditions. Thus, the HVSR method fails at these stations. The SSR method underestimates the amplifications at some stations since the rock site, selected as reference site, has its own site response and/or the path correction we applied, considering the geometrical spreading factor only, is insufficient. At the sites where high intensity values were observed, we found high amplifications. The fundamental soil frequencies characterize the damage properties observed in the Adana-Ceyhan earthquake. The fundamental soil frequency is nearly at 1.1 Hz at the Ceyhan site, where severe damage was observed in the 5–6 story buildings, while the fundamental soil frequency is between 3–6 Hz at the Adana site, where damage was in the low-story buildings. Therefore, in addition to inefficient construction practices, it is clear that the resonance effects have also contributed to the observed damage.     

Simulating the strong ground motion of the 2022 MS6.8 Luding,Sichuan, China Earthquake

Stochastic finite-fault simulations are effective for simulating ground motions and are widely used in engineering to determine the impacts of ground motion and develop relevant predictive equations. In this study, the source, path, and site amplification coefficient of western Sichuan Province, China, and stochastic finite-fault simulations were used to simulate the acceleration time series, Fourier amplitude spectra, and 5% damped response spectra of 28 strong-motion stations with rupture dist...     

Evaluation of site effects using numerical and experimental analyses in Città di Castello (Italy)

In the paper the results of the numerical and experimental analyses, at a site in the Umbria Region (Città di Castello—PG), are finalized and the evaluations of site effects are shown. The aim of the work was to compare the two types of analyses and give some methodologies that may be used at the level of urban planning. Therefore a series of geological, geomorphological (1:5.000 scale), geotechnical and seismic analyses have been carried out, to characterize the lithological units and to identify the areas affected by site effects. The expected seismic inputs were provided and two-dimensional numerical analyses have been done. An experimental analysis, using the registrations of small events, has been done. The results for the two approaches were performed in terms of elastic pseudo-acceleration spectra and amplification factors. The results are given as a ratio between spectral intensity, calculated using the pseudo-velocity spectra, in the periods of 0.1–0.5 s and 0.1–2.5 s of output and input. The results have been analyzed and compared, to provide a methodology that may be comprehensive and precise.     

Site amplification of S-waves in Bursa City and its vicinity, Northwestern Turkey: comparison of different approaches

Microearthquake recordings of local events have been used to quantify the site effects in the vicinity of Bursa City, northwest Turkey. Since the city is located near the southwest branch of the western extension of the North Anatolian Fault (NAF) zone, the importance of the seismic hazard in the region becomes progressively more important. The accelerograms of 69 earthquakes that were recorded on different geologic units, massive limestone, slope deposit and Quaternary sediment were analyzed to estimate the response of the recording sites. Site amplification functions were obtained by using three different approaches (standard spectral ratio, SSR; horizontal to vertical, H/V ratio and generalized inversion method, GIM) and the differences between the methods were also evaluated. We found large discrepancies between the SSR and H/V ratio methods, specifically; the former yields almost three times higher amplitudes than those obtained in the latter approach. Station located within the Bursa Quaternary basin (SCKR) is characterized by the largest estimates of the amplification amplitudes (8.0, 4.5 and 4.0 for SSR, H/V ratio and GIM, respectively) in all the three methods. On the other hand, stations located on deep limestone geological unit (SIGD and SKAY) show the least amplification level, ranging between 1.0 and 1.6. Three methods are able to identify resonant frequencies of the sites, although the absolute amplitudes of the amplification function are obtained different from each method.     

Frequency variation in site response as observed from strong motion data of the L��Aquila (2009) seismic sequence

Previous works based mainly on strong-motion recordings of large Japanese earthquakes showed that site amplification and soil fundamental frequency could vary over long and short time scales. These phenomena were attributed to non-linear soil behaviour: the starting fundamental frequency and amplification were both instantaneously decreasing and then recovering for a time varying from few seconds to several months. The recent April 6, 2009 earthquake (M _W 6.3), occurred in the L’Aquila district (central Italy), gave us the possibility to test hypotheses on time variation of amplification function and soil fundamental frequency, thanks to the recordings provided by a pre-existing strong-motion array and by a large number of temporary stations. We investigated the intra- and inter-event soil frequency variations through different spectral analyses, including time-frequency spectral ratios and S-Transform (Stockwell et al. in IEEE Trans Signal Process 44:998–1001, 1996). Finally, analyses on noise recordings were performed, in order to study the soil behaviour in linear conditions. The results provided puzzling evidences. Concerning the long time scale, little variation was observed at the permanent stations of the Aterno Valley array. As for the short time-scale variation, the evidence was often contrasting, with some station showing a time-varying behavior, while others did not change their frequency with respect to the one evaluated from noise measurements. Even when a time-varying fundamental frequency was observed, it was difficult to attribute it to a classical, softening non-linear behaviour. Even for the strongest recorded shocks, with peak ground acceleration reaching 0.7 g, variations in frequency and amplitude seems not relevant from building design standpoint. The only exception seems to be the site named AQV, where the analyses evidence a fundamental frequency of the soil shifting from 3 Hz to about 1.5 Hz during the mainshock.     

Spectral Ratio Estimates for Site Effects on the Horst–Graben System in West Turkey

Recordings of micro- and moderate-size local earthquakes have been used to quantify site effects in the central-west Turkey which contains one of the world’s best examples of a rapid intra-continental extension with its high population and industrial potential. We analyzed 436 earthquakes with local magnitudes ranging between 2.0 and 5.6 using three component digital recordings from 32 stations. Site functions were obtained using two different spectral ratio approaches (horizontal to vertical spectral ratio, HVSR, and standard spectral ratio, SSR). HVSR estimates of transverse and radial S-waves were compared with one another. Epicentral distance, magnitude and back-azimuth dependencies of site functions were also evaluated. In general, HVSR values from transverse and radial S-waves are similar within a factor of 2. The back-azimuth dependencies of transverse S-wave HVSR results are more significant than distance and magnitude dependencies. On the other hand, averaging of transverse and radial S-wave HVSR results eliminates systematic back-azimuth dependencies caused by source radiation effects. Distributions of HVSR estimates along ~N–S linear array, which traversed main grabens in the region with a station spacing of 3–4 km, reflect subsurface geological complexities in the region. The sites located near the basin edges are characterized by broader HVSR curves. Broad HVSR peaks could be attributed to the complexity of wave propagation related to significant 2D/3D velocity variations at the sediment–bedrock interface near the basin edges. The results also show that, even if the site is located on a horst, the presence of weathered zones along the surface could cause moderate frequency dependent site effects. Comparison of HVSR and SSR estimates for the stations on the graben sites showed that SSR estimates give larger values at lower frequencies which could be attributed to lateral variations in regional velocity and attenuation values caused by basin geometry and edge effects.