1 Introduction

The unpredictability and severity of earthquakes classify them as the most dangerous natural hazards. Approximately 1.5–2.0 million people have lost their lives in 1000 deadly earthquakes recorded in the last century (Pomonis et al. 1993) whreas, over 200 fatal earthquakes have been experienced in the 21st century with a total loss of life of more than 0.7 million people [United State Geological Survey (USGS)]. The destruction due to such an event (in the future) can be minimized by reducing their effects through precise seismic hazard assessment and incorporating it in seismic designing. With the increase of population, the danger gets higher in many earthquake-susceptible regions, especially in growing nations, like Nepal, India, Haiti, etc. The Ahmedabad city falls in the central-eastern part of Gujarat state of India. It also falls under seismic zone III of the Indian seismic zone map of India (BIS: 1893–2016), The Gujarat state has experienced two large earthquakes; the 1819 Allah Bund Earthquake of magnitude 7.8 and the 2001 Bhuj Earthquake of magnitude 7.6 (2001 Bhuj Eq.) and several other earthquakes of magnitude ≤ 6 in the last 100 years (Rastogi et al. 2012). It is still seismically active.

Ahmedabad is the major city and the former capital of Gujarat. It is situated 32 km SW of the present capital city, Gandhinagar. The city is categorized in zone III in the Indian seismic zone map. The city is located in the rift basin named Cambay rift formed by the West Cambay and East Cambay boundary faults which are dipping towards the city (Wani and Kundu 1995). Ahmedabad is a rapidly developing city in the state of Gujarat; the population is in a rising trend, which has led to continuous horizontal and vertical development. The city suffered heavily in the 2001 Bhuj Eq., even though the epicenter was 250 km away and the depth of the earthquake was ~ 25 km. Many high-rise structures had damaged (Mishra 2004). The city has also experienced an earthquake of Mw 5.7 in 1864, with an epicenter ~ 80 km south (ISR 2013). The past seismicity suggests the vulnerability of structures from large earthquakes in the future (both from near- and far-sources) and emphasizes the need for micro-seismic hazard assessment first of Ahmedabad city. The microzonation of the urban area (both residential and industrial) was practiced, in 1954, at the Yokohama city of Japan. Worldwide, seismic microzonation is carried out in several countries, like Japan (Yokohama, Urmia, and Tokyo), United States (Central US), Nepal (Kathmandu), Australia (Perth), China (Hong Kong) and India. In India, micro-seismic hazard assessment was carried out in metropolitan cities such as Delhi (Sharma and Rao 2004), Guwahati and Sikkim (Nath et al. 2008), Bangalore (Anbazhagan and Sitharam 2008), Chennai (Ganpathy 2011) and Jabalpur (Rao 2014). The macro-SHA work was carried out by many researchers in parts (Peterson et al. 2004; Iyengar and Raghukanth 2006; Chopra et al. 2012b, a; Mohan 2014) in Gujarat. The predominant frequency was estimated by Mandal et al. (2005) through the H/V method in the central-eastern Kachchh. On a national scale, Parvez et al. (2003) have prepared the deterministic seismic hazard assessment map of India. National Disaster Management Authority (NDMA 2010) has also proposed a probabilistic seismic hazard map of India on NEHRP-A- site class (with Vs30 ≥ 1.5 km/s). Sitharam and Kolthyar (2013) have proposed the seismic hazard analysis of India using the areal sources. On a regional level, Chopra et al. (2012b, a) have carried out the Deterministic seismic hazard map of Gujarat in which surface PGA in Ahmedabad city is estimated. Bhandari et al. (2013) have also conducted a regional probabilistic seismic hazard assessment study in Ahmedabad. At the local level, the Gujarat State Disaster Management Authority, along with the Institute of Seismological Research and OYO International Corporation, Japan has conducted seismic microzonation in the Gandhidham-Kandla area, considering the tectonics, shear-wave velocity, and geotechnical data. Considering the geophysical, geological and geotechnical aspects, a micro-seismic hazard assessment of Ahmedabad is yet to be done.

In this work, we carried out a micro-level seismic hazard assessment in Ahmedabad through subsurface characterization using geotechnical and geophysical data. The strong motion data of large earthquakes is scarce in the study area to aid earthquake-resistant designing. In view of this, the strong motion due to large earthquakes is simulated by considering past seismicity, tectonic setting and various geotechnical, geophysical, and geological parameters. To recompense for the uncertainty, region-specific parameters are used and diverse scenarios are generated for near-field as well as far-field earthquakes. The local site effects are also incorporated by 1D-ground response analysis at each borehole location.

2 Geology and tectonic setting

The Ahmedabad city is located in a Cambay rift, formed in the early Cretaceous (Biswas 1982, 1987), with east and West Cambay margin faults as the rift shoulders (Fig. 1). The development of the Cambay rift, especially tectonic, is closely associated with the development in the Indian Plate (Biswas et al. 1994). The rifting propagated from east to west. The first rift stage started with extensional faulting from the east along the NNW–SSE older-Aravalli trend developing an easterly tilted half-graben in which the deltaic clastics had deposited. This was followed by widespread volcanism which was a syn-rift event and the effects of rifting were subdued by outpouring basaltic lava, resulting in the cover of a very thick lava pile over the Cretaceous sediments, which formed the basement of the Tertiary deposits in the second rift cycle. The basin evolved fully during the later cycle when the graben subsided by the extensional faulting to accommodate the huge thickness of marine/lagoonal fine clastic sediments from Late Paleocene to Mid-Miocene till the rifting process ceased (Biswas 1999). The Cambay basin extends broadly in an NNW–SSE direction in the onshore and the offshore parts of Gujarat (Fig. 1). The basin width narrows down towards the northern part, indicating a possible decline of the rift propagation. Northward, it swings to NE merging into the Rajasthan basin (Mehr and Chamyal 1997). oil and natural gas corporation (ONGC) through acquired and analysed data from the geophysical survey (mainly on the Tertiary sediments), revealed the basement configuration and delineated the cross faults that were responsible for the division of basin into sub-basins/tectonic blocks. These faults behaved differently during the deposition due to differential uplifts along with various bounding and transverse faults (Wani and Kundu 1995).

Fig. 1
figure 1

Geology and structural map of Cambay rift basin (after Wani and Kundu 1995) overlapped with the seismicity up to 50 miles from the year 1668 to 2012

Full size image

The faulted basin is filled with approximate 4000 m thick sediments of Tertiary to Quaternary in age. Based on the subsurface drilling and detail seismic surveys, ONGC (Biswas 1982; Wani and Kundu 1995) inferred different faults and depths to different geological layers. Due to the alluvial cover, these faults could be potentially detected by geophysical surveys as these are not always easy to decipher (Dixit et al. 2010; Mohan et al. 2017a, b; Danda et al. 2017).

The city of Ahmedabad is situated on the banks of Sabarmati River, within 50 km wide Cambay basin (Fig. 1), enclosed by two NNE–SSW trending faults: West Cambay and East Cambay Faults.

3 Methodology

The present study is targeted to conduct a micro-seismic hazard assessment and to predict surface ground motion parameters for the Ahmedabad city through characterizing the near-surface. The entire scheme is divided into five main parts: the methodology proposed in the Mohan et al. (2017a, b, 2018) is considered in the current study. The estimation of surface strong ground motion is divided into five major parts:

  1. (1)

    Assessment of earthquake potential of the study area based on past seismicity and major tectonics.

  2. (2)

    Demarcation of the engineering bed layer (EBL) (in case if bedrock is deeper). The layer is marked based on a geotechnical property (SPT N value > 80), geophysical property (400 m/s ≤ Vs ≤ 750 m/s) and uniform distribution of soil type in the study area.

  3. (3)

    Preparation of soil models (soil modeling) using the geotechnical and geophysical parameters.

  4. (4)

    Assessment of the strong motion at EBL through simulation considering far-field and near-field earthquake scenarios using region-specific ground motion parameters.

  5. (5)

    Equivalent linear 1D-ground response analysis for the computation of strong motion at the surface. The flowchart showing methodology has been given in Fig. 2.

Fig. 2
figure 2

The flow-chart showing methodology of estimation of SGM at the surface level

Full size image

3.1 Estimation of the earthquake potential of the study area

The BIS classified the Ahmedabad city under seismic zone III (BIS 2016). The historical seismicity map prepared from the catalog of the Institute of Seismological Research, Gandhinagar suggests very few earthquakes in the vicinity of Ahmedabad city. A total of 26 earthquakes were observed in Ahmedabad city from 1821 to 2016 with the maximum magnitude of M5.7 that occurred in 1864. The Frequency–Magnitude range plot of the earthquakes has been shown in Fig. 3. Besides local earthquakes, Ahmedabad has also experienced large earthquakes from the Kachchh region. During the 2001 Bhuj earthquake, several multi-story reinforced concrete-frame buildings collapsed in Ahmedabad [80 buildings with 1021 apartments and 82 other houses collapsed (Mishra 2004)]. In total, 1103 units or families were affected by the collapsed structures in the city (Mishra 2004).

Fig. 3
figure 3

Frequency–Magnitude plot of the Ahmedabad seismicity from 1821 to 2016

Full size image

For designing earthquake-resistant structures, it is important to predict the expected peak accelerations during strong earthquakes. The buildings can be of different types; low story (1–4 story), mid-story (4–7 story) and high-rise (> 7 stories) buildings. The near-field earthquakes possess most of the energy in the high frequency/low period range. Therefore, near-field earthquakes are critical for low story buildings. Similarly, due to the low frequency/high period, the far-field earthquakes were found more critical for mid-rise and high-rise buildings. The East Cambay Boundary Fault has the potential of M ≤ 6 earthquake (Rastogi 2014). Therefore, based on recent past experience of the 2001 Bhuj Earthquake, the seismic potential of the Eastern Cambay Boundary fault and the presence of low, mid and high-rise buildings in Ahmedabad, the near-field and far-field earthquake scenarios are considered for this study.

3.2 Establishment of EBL from geotechnical and geophysical data

3.2.1 Geotechnical investigations

Twenty boreholes were drilled in the Ahmedabad city up to the depths of 80 m (10 boreholes; BH-01–BH-10), 50 m (one borehole; BH-20) and 40 m (nine boreholes; BH-11–BH-19). The lithology of the boreholes drilled up to the depth of 80 m is given in Fig. 4.

Fig. 4
figure 4

Lithologs of the 11 boreholes drilled up to the depth of ~ 80 m in the Ahmedabad city

Full size image

The clay is present at the surface in the western and southeastern part of Ahmedabad city whereas the rest of the city is covered by the sand at the surface (Fig. 5). The Northeast–Southwest and East–West soil cross-sectional profiles (locations marked in Fig. 5a) are shown in Fig. 5b and c, respectively.

Fig. 5
figure 5

a Location map of the soil profiles prepared in Ahmedabad City, b East–West soil Profile and c NE–SW Profile

Full size image

The alternate layers of sandy and clayey units with more stiffness were found up to 20 m depth, denser 20 m onwards. At shallower depths of around 10 m, the corrected N-values reach > 50 (Very stiff/ Refusal), while at deeper depths the values reach > 100 (as same as a rock) (Sairam et al. 2018).

  1. (6)

    These layers are classified as per the engineering geological classification based on average values of soil properties obtained from laboratory tests according to Indian Standards.

3.2.2 Geophysical investigations

In Ahmedabad, the shear-wave velocity (Vs) was estimated at 55 sites (having length of 60 m) using Multichannel Analysis of Surface Wave and at 11 sites with the PS-logging (Sairam et al. 2018) (Fig. 6). A good match of shear wave velocity (1D models) has been found at sites where both MASW and PS logging has been conducted. The Vs30 was computed at each site (Sairam et al. 2018). The eastern portion adjacent to the Sabarmati River has shown higher velocity (300–360 m/s) than the western side (265–300 m/s). Only one site has shown a Vs30 of about ~ 2005 m/s and two sites have shown the Vs30 of ~ 380 m/s. On average, the Vs30 has been found varying from 265 to 360 m/s (Fig. 6) in the Ahmedabad city and can be categorized in D-type soil (stiff soil) in the national earthquake hazards reduction program (NEHRP 2003) scheme. The two sites that showed higher Vs30 can be classified as C-type soil (dense soil).

Fig. 6
figure 6

The Vs30 distribution map of Ahmedabad City prepared based on Vs30 data from Sairam et al. (2018). The Vs30 values vary from 265 to 360 m/s fall under the category of NEHRP D-site class (stiff soil)

Full size image

Based on the seismic survey, Kaila et al. (1990) have suggested the deep basement/bedrock (of up to ~ 6 km) in Ahmedabad city, which is located in the Cambay rift. Therefore, the site effect on the strong motion can be estimated considering the engineering bed layer (EBL) in place of basement/bedrock. The shear wave velocity estimated through MASW survey and PS logging (Sairam et al. 2018) has shown a layer with Vs ~ 500 m/s in the Holocene sediments in the entire city. At similar depths, the geotechnical investigation from the boreholes has shown a layer with an N-value of more than 80 blow counts. Therefore, this layer (with a Vs. of 500 m/s and N-value of ~ 80 blow counts) and present in the entire city has been considered as the engineering bed layer. The EBL is demarcated at about 24–54 m depth in Ahmedabad city. The depth of the EBL in Ahmedabad city is variable, being shallower (24–35 m) (in the west-central part and deeper (44–54 m) in the central and the south–eastern parts (Fig. 7 and Table 1).

Fig. 7
figure 7

EBL distribution in the Ahmedabad City

Full size image
Table 1 The depth of EBL demarcated in different boreholes based on geophysical survey and geotechnical testing
Full size table

3.3 Soil modeling

A total of 20 boreholes were drilled in the Ahmedabad city: 10 boreholes are drilled up to 80 m depth (BH-01–BH-10), one borehole (BH-20) up to 50 m depth and 9 boreholes up to 40 m depth (BH-11–BH-19). The lithology is classified based on the grain size analysis (maximum content of clay, silt or sand) for the soil modeling. If any of the soil types (Gravel, Clay, Sand, or Silt) occupies > 50% of the soil, then the soil is designated as that type of soil. In case, all soil types show < 50% presence, then the soil (Sand, Silt, Clay or Gravel with the highest presence is designated as the soil type.

Based on the soil investigation from geotechnical data and geophysical investigations, soil models of 20 boreholes were prepared for the ground response analysis. The properties like wet density, N-value, plasticity index estimated from the borehole data and Vs (estimated through MASW survey and PS logging) have been considered to model the soil. The depth of the groundwater table (estimated from borehole drilling in the city) is found to vary from 8 m to more than 30 m in the city and is considered for preparing soil models. The prepared soil models for the boreholes BH-01 and 08 are given in Tables 2 and 3, respectively. A 3–18 m thick clay has been found (from the surface) in the Ahmedabad city, west to Sabarmati River and in the SE part, whereas sand covers (2–30 m) the rest of the city, east to the river. The remaining soil models (18 nos.) have been given as supplementary information.

Table 2 The Soil model prepared for the borehole BH-01
Full size table
Table 3 The Soil model prepared for the borehole BH-08
Full size table

3.4 Estimation of strong motion at EBL

Aki (1966, 1967) proposed an expression for a seismic wave spectrum radiated from complex faulting and correlated the seismic moment to the rigidity, slip, and fault area. The concept was used for ground motion estimation for high frequencies. Hartzell (1978) proposed the concept of division of rupture along a finite fault into sub-faults/sub-sources. These sub-faults can be deliberated as a point source. The concept covered the distributed and inhomogeneous rupture and the effect of the geometry of the fault. In this concept, the resultant strong motion is assessed at a site by summation of the contribution of sub-sources considering the time delays of sub-sources due to propagation of the rupture and source to site distance. This concept was further followed by Irikura (1983, 1992); Midorikawa (1993); Irikura and Kamae (1994). Beresnev and Atkinson (1998a) implemented the concept of a stochastic approach in the strong motion simulation by dividing the fault into sub faults considered as sub-sources that follows the ω2 source spectrum. The systematically delayed sub-source effects were then included in the time-domain to estimate the strong motion at the interested site. The prepared source code was named FINSIM (Beresnev and Atkinson 1998b); Motazedian and Atkinson 2005) have further modified this source code by introducing the theory of dynamic corner frequency. In this concept, the corner frequency is assumed as a function of time and the total energy emitted from the fault is independent of sub source size unlike the static corner frequency (Motazedian and Atkinson 2005). The modified approach, therefore, is very less dependent on the sub-source size. This approach is named as stochastic finite fault modeling technique (SFFMT) for strong motion simulation. The revised source code is named as EXSIM. In the Stochastic finite fault modeling technique, the rupture history controls the frequency content of the simulated time series of each sub-fault. The advantage of the modified stochastic finite fault model is that it conserves moment with the single triggering of each sub-fault as compared to multiple triggering in previous approaches (e.g. Silva and Darragh 1995; Beresnev and Atkinson 1998a). In the current work, strong motion is simulated at the EBL using the stochastic finite fault modeling technique. This method was successfully applied to simulate strong motion in various parts of the world. The detail has been given in Table 4. The requirement of two types of scenario earthquakes has been deliberated in the current study (1) near-field earthquake scenarios and (2) far-field earthquake scenarios. Input strong motion is simulated at EBL using stochastic finite fault modeling technique of Motazedian and Atkinson (2005) at Vs of 500 m/s. The requirement of two types of scenario earthquakes has been deliberated in the current study (1) near-field earthquake scenarios and (2) far-field earthquake scenarios. Input strong motion is simulated at EBL using the stochastic finite fault modeling technique of Motazedian and Atkinson (2005) at EBL (Vs of 500 m/s).

Table 4 The List of simulated earthquakes worldwide using Stochastic Finite Fault modeling Technique (Motazedian and Atkinson 2005)
Full size table

3.4.1 Input parameters for simulation of strong motion at EBL

In the historic past, Ahmedabad city has experienced two earthquakes of magnitudes 3.7 (in 1843 and 1897), one of 4.6 (in 1840) and one of 5.7 (in 1864) (Oldham 1883). Several other small shocks were also experienced in and around Ahmedabad (ISR 2013). Besides, large earthquakes that occurred in Kachchh district (Mw 6, 1956 Anjar earthquake 2001 Bhuj Eq. and 1819- ABEQ), at a distance of ~ 250 km from the city, had affected Ahmedabad. Several multi-story reinforced concrete-frame buildings were collapsed in Ahmedabad (Mishra 2004). The famous minaret of Jama Maszid (mosque) located in Ahmedabad was damaged during the 1819-ABEQ that occurred in the NW part of Kachchh, Gujarat (Martin and Szeliga 2010).

For proper designing of earthquake-resistant structures, it is important to estimate the level of accelerations expected during future large earthquakes. Ahmedabad has experienced the effect of both near-field and far-field sources in the past and has low-story, mid-story, and high rise buildings. Therefore, consideration of both the sources (near-field as well as far-field) become essential for the estimation of possible seismic hazard of the city. The unavailability of strong motion recordings in low seismicity areas suggests the simulation with well-justified input parameters (seismological, geophysical, geological and geotechnical) as the alternate option.

The Cambay rift is enclosed by the West Cambay Boundary fault and East Cambay Boundary fault (Fig. 1). In 1864, an earthquake of magnitude M5.7 occurred along the East Cambay Boundary fault. This fault with a dip of 60° has the potential to generate an earthquake of M ≤ 6 (Rastogi 2014). Thus, near-field scenario earthquakes of magnitudes Mw 5.5, 6.0 and 6.5 (at the maximum) have been considered along this fault. Wells and Coppersmith (1994) have proposed relations to estimate the rupture dimensions based on worldwide dataset. The same relations have been used to estimate rupture dimensions in the present study. The West Cambay Boundary fault is an NNW–SSE trending fault that has a strike direction of 330° and dips of around 60° (Wani and Kundu 1995). So, for near-field, two scenarios are considered, one from ECF and another from WCF.

The Mw 7.6, 2001 Bhuj earthquake was located ~ 250 km NW of Ahmedabad. Also, Mw 7.8, 1819 Allah Bund earthquake occurred ~ 270 km from the city. Effectively, earthquakes with Mw 7.0, 7.8 and 7.6 have been considered for the far-field earthquake scenarios. The Wells and Coppersmith (1994) relationship has been used to estimate the rupture dimension for Mw 7.0 and 7.8 scenario earthquakes and the rupture dimensions proposed by Yagi and Kikuchi (2001) for 2001 BEQ have been considered for simulation of Mw 7.6. The quality factor proposed by Chopra et al. (2010) and Gupta et al. (2012) from local earthquakes has been used for compensating the path effect for near-field earthquake scenarios. The quality factor estimated by Singh et al. (1999, 2004) and Bodin et al. (2004) are used for the far-field earthquake scenarios. The Kappa values (ĸ), which takes into account the near-surface wave attenuation/Fall-off of the high frequency (> 1 Hz) Fourier amplitude spectrum (Anderson and Hough 1984) was considered around 0.03 for areas covered by Quaternary alluvium (Rapolu and Mandal 2014; Kumar 2015; Chandler et al. 2006). We took Kappa values of 0.02, 0.025 and 0.03 for the present work.

The stress values for shallow earthquakes of the stable continental region (Mw 6.7 Koyna earthquake in 1967, Mw 5.7 Bhadrachalam earthquake in 1969, and Mw 5.4 Broach earthquake in 1970) is around 60 bars (Rastogi 2016). Therefore, stress drop values of 60, 80 and 100 bars have been used in the near-field scenario earthquakes. Bodin and Horton (2004) computed the stress drop of 160 bars for the 2001 Bhuj Eq. thus, a stress drops of 140, 160, and 180 bars have been used for the far-field earthquake scenarios.

The strong motion due to near-field earthquake scenarios (Mw 5.5, 6.0 and 6.5 at East Cambay Boundary fault and far-field Eqs. (Mw 7, 7.6 and 7.8 at the eastern part of KMF) are generated using model parameters as shown in Tables 5 and 6, respectively and considering site amplifications suggested by Boore and Joyner (1997) for generic soils at Vs30 (average Vs in top 30 m subsurface layer) of 520 m/s (Table 7). Although it is not possible to predict earthquakes but the seismic scenarios based on past seismicity and estimated source parameters of the past earthquakes in the area can provide a reasonable seismic hazard for the region. Therefore, a total of 108 and 81 input parametric combinations have been considered for estimating the strong motions at EBL for near-field and far-field earthquake scenarios, respectively.

Table 5 The selected model parameters for simulation of SGM due to near-field earthquake scenarios
Full size table
Table 6 The selected model parameters for simulation of SGM due to far-field ground earthquake scenarios
Full size table
Table 7 Frequency dependent amplification generated for the Vs30 of 520 m/s (Boore and Joyner 1997)
Full size table

3.5 Ground response analysis

The equivalent-linear ground response analysis is a popular technique of estimation of ground response due to an earthquake. Therefore, it has been applied at each borehole site to estimate the surface strong motion. In the equivalent-linear analysis, the transfer functions are computed for a horizontally polarized vertically propagating shear wave passing through a soil column with layers of different stiffness, frequency, density and damping. Initially, the small values of strain, damping, and stiffness are considered for obtaining the response of the soil column. From this initial estimate, shear strain for every layer is computed. Then, for each layer, the effective shear strain (which is magnitude dependent) is calculated as some part of the maximum shear strain. In the present study, it is (taken as 0.65 due to the considered scenario earthquake of magnitudes 5.5–7.8 earthquakes. Afterward, the values of stiffness and damping at the effective shear strain are determined from damping and shear modulus reduction curves. The process is iterated until the difference of damping and stiffness values in two consecutive iterations reaches less than a set of acceptable values (Kramer 1996). The following four inputs are required for ground response analysis : (1) Soil classification (on the basis of soil type, density, thickness and N value/shear-wave velocity), (2) depth to the base rock/EBL, (3) strong motion at base rock/EBL, and (4) damping and shear modulus with shear strain curves. The ground response analysis is done on SHAKE 2000 (Ordonez 2012). The SHAKE 2000 can convert the supplied outcrop strong motion to base rock motion before ground response analysis with its inbuilt capability by supplying the option at the beginning itself. The ground response analysis has been conducted for both far-field earthquake scenarios and near-field earthquake scenarios.

The damping ratio (β) and modulus reduction (G/G max) versus shear strain curves of Schnabel (1973), Sun et al. (1988) and Seed and Idriss (1970) for rock, clay, and sand, respectively, which were used in the seismic microzonaton study of nearby Gandhinagar region (Mohan et al. 2018) have been used in the present study to conduct ground response analysis. These curves are presented in Figs. 8 and 9. These curves are provided in the SHAKE database and used during calculations by selecting the options. The ground response analysis is conducted for all parametric combinations of both earthquake scenarios and the surface strong motion is generated.

Fig. 8
figure 8

The Modulus reduction versus shear strain curves used in the analysis for various soil types

Full size image
Fig. 9
figure 9

The Damping ratio versus shear strain curves used in the analysis for various soil types

Full size image

The PGA values at EBL, as well as the surface level of all the 20 boreholes, have been computed from the surface strong motion. The estimated mean PGA values due to all near-field earthquake scenarios at each borehole at a surface level have been given in Table 8 and due to far-field earthquake scenarios in Table 9. The PGA distribution map for surface level due to near-field earthquake scenarios has been given in Fig. 10. The Bureau of Indian Standard [IS 1893 (Part 1)] has proposed the conversion factor from 5% damping to other dampings based on material type, therefore the response spectra are calculated on the surface of every borehole for damping of 5%. In the case of a near-field earthquake, a total of 108 response spectra are generated at each borehole considering different parametric combinations. The mean, mean+standard deviation (SD) and mean-SD response spectra are also estimated at each borehole. All response spectra computed due to near-field earthquake scenarios at each borehole have been shown in Fig. 11. The spectral acceleration maps at different periods (0.2 s, 0.55 s, 1.0 s and 1.25 s) have also been prepared given in Fig. 12. The amplification of ground motion between EBL and surface levels due to near-field earthquake scenarios has been given in Fig 13. In the case of far-field earthquake scenarios. The PGA distribution map for surface level due to far-field earthquake scenarios has been given in Fig. 14. In the case of a far-field earthquake, a total of 81 response spectra are generated at each borehole considering different parametric combinations. The mean, mean+standard deviation (SD) and mean-SD response spectra are also estimated at each borehole. All response spectra computed due to far-field earthquake scenarios at each borehole have been shown in Fig. 15. The spectral acceleration values at each borehole at the surface level have also been extracted from the response spectra and the spectral acceleration distribution maps have been prepared for periods of 0.2 s, 0.55 s, 1.0 s and 1.25 s. The spectral acceleration (SA) distribution maps for far-field earthquake scenarios have been shown in Fig. 16. The amplification of ground motion between EBL and surface levels due to far-field earthquake scenarios has been given in Fig. 17.

Table 8 The borehole-wise mean PGA values (at surface) and amplification due to near-field scenario earthquake
Full size table
Table 9 The borehole-wise mean PGA values (at surface) and amplification due to far-field earthquake
Full size table
Fig. 10
figure 10

The surface peak ground acceleration distribution map of Ahmedabad city area for near-field earthquake scenario prepared using SHAKE program

Full size image
Fig. 11
figure 11figure 11figure 11

Borehole-wise normalized spectral acceleration plots with Mean and Standard Deviation for Near-Field earthquake consideration

Full size image
Fig. 12
figure 12

The spectral acceleration distribution maps of the Ahmedabad City for near-field earthquake scenario

Full size image
Fig. 13
figure 13

Mean Amplification due to near-field earthquake scenario

Full size image
Fig. 14
figure 14

The surface peak ground acceleration (in cm/s2) distribution map of Ahmedabad City for a far-field earthquake with the location of damaged buildings

Full size image
Fig. 15
figure 15figure 15figure 15

Borehole-wise normalized response spectra plots with Mean and Standard Deviation for Far-Field earthquake consideration

Full size image
Fig. 16
figure 16

The spectral acceleration distribution maps of the Ahmedabad city area for far-field earthquake scenario at: a 0.2 s, b 0.55 s, c 1.0 s, and d 1.25 s

Full size image
Fig 17
figure 17

Mean Amplification due to far-field earthquake scenario

Full size image

4 Results and discussion

The ground response analysis is conducted using soil models prepared from 20 boreholes of the Ahmedabad city (Fig. 4). The EBL has been established between 24 and 54 m (Table 1 and Fig. 7). The input strong motion (at EBL) when passing within the soil layers (present above EBL) during ground response analysis, either gets amplified or attenuated on the boundaries of every soil layer. Finally, the resultant strong motion is obtained at the surface.

The strong motion for near-field earthquake scenarios (Mw 5.5, 6.0 and 6.5) along ECF and WCF are generated using the Stochastic finite fault modeling technique. To reduce uncertainty in parameters, a suite of parameter based on past studies in the region are used. A total of 108 parametric combinations ( 2 faults (ECF and WCF), three magnitudes, two quality factors, three stress drops (Table 5) are used for the estimation of strong motion at EBL due to near-field earthquake scenarios. The ground motion of each of the 108 input parametric combinations for the near-field earthquake scenarios with PGA varying between 52.4 and 111.4 cm/s2 is applied at EBL. The mean surface PGA is thus estimated from 100.7 to 278.9 cm/s2. The PGA is found high in western Ahmedabad in comparison to eastern Ahmedabad due to low shear wave velocity (Fig. 6) and the presence of clay with medium plasticity (Fig. 5b). The maximum PGA (of 278.9 cm/s2) has been estimated in the west-central part of Ahmedabad at BH-1. The PGA distribution at the surface is given in Fig. 10 and Table 8. The response spectra (on 5% damping) at the surface level of each borehole and the mean normalized response spectra at each borehole are given in Fig. 11. The natural frequencies of buildings depend on two main parameters (mass and stiffness) which are related as: f = 1/2π \(\surd (K/M)\) ; K is stiffness and M is the mass of the building. Therefore, taller buildings tend to be more flexible and have low natural frequencies in comparison to shorter buildings. In general, the building with 0.2 s corresponds to a natural period of 1–2 story building, 0.55 s for 4–5 story building, 1 s high rise building and 1.25 s for large story building. Therefore, the mean SA distribution maps for 0.2 s, 0.55 s, 1 s and 1.25 s period (large story building) are prepared at the surface level and are shown in Fig. 12. The mean PGA amplification factor is also calculated at all the boreholes. It is varying from 1.6 to 3.3 (Fig. 13 and Table 8). The maximum amplification is observed at boreholes BH-01 and BH-17 due to low shear wave velocity (Fig. 6) and the presence of clay with medium plasticity (Fig. 5 b). Moreover, the stiffness contrast between the layers also creates the amplification factor. The presence of a 6 m clay layer followed by sand at borehole BH-01 also suggested an abrupt increase in ground motion/amplification.

The response spectra of all the boreholes are computed at 5% damping from 0.1 to 4.0 s. The maximum mean SA of 680 cm/s2 (70% more than BIS suggested values National code) is observed at 0.2 s (Fig. 12a) (corresponds to 1–2 story buildings), in the western part of the Ahmedabad city area around borehole BH-01 (comprising of the clayey top layer with low shear wave velocity). The mean SA is found in the range of 600–650 cm/s2 near BH-7 and BH-11 (in the south-western part). At the period of 0.55 s (analogous to the natural period of 5–6 story buildings) higher SA is observed in the north-western, west-central and southern parts of the Ahmedabad city. The near-field earthquakes generate low energy at high periods/low frequencies, the maximum mean SA of 280 cm/s2 is calculated (28% less than National code) at BH-10 and BH-19 in the west-central part of the Ahmedabad city (Fig. 12b). The maximum mean SA of 130 cm/s2 (50% less than the National code) at BH-07 is computed in the south-western part of the Ahmedabad city at 1.0 s (corresponding to the natural period of 10 story buildings) (Fig. 12c). The maximum SA of 115 cm/s2 (less than 50% of the National code) at BH-11, is computed in the south-western part of the Ahmedabad city at 1.25 s (corresponding to the natural period of 12–15 story buildings) (Fig. 12d).

In the case of far-field earthquake scenarios, the input motion with a mean PGA (estimated from 81 input parametric combinations) of 107.8 cm/s2 is applied at EBL. The PGA at the surface level has been computed through ground response analysis. The computed mean surface PGA is found to vary from 118.4 to 161.5 cm/s2 (Table 9). The highest mean PGA of 161.5 cm/s2 is observed in the west-central portion and the lowest PGA of 118.4 cm/s2 is estimated in the north-eastern part of Ahmedabad. The maximum mean PGA (161.5 cm/s2) is observed at BH-01 (comprising of clay with medium plasticity and low Vs values) and the minimum at BH-03 (comprising of sand with high density and higher Vs in the top layer). The mean PGA amplification factor is also calculated at all boreholes and is found to vary from 2.2 to 3.0 (Table 9), with maximum amplification at BH-01.

The mean response spectra are determined at all the boreholes at 5% damping from 0.1 to 4.0 s (Fig. 15). The SA maps at different periods (0.2, 0.55, 1.0 and 1.25 s) are prepared (Fig. 16). At 0.2 s (corresponding to 2–3 story buildings) (Fig. 16a), the maximum mean SA of 500 cm/s2 is observed (~ 28% more than the National code) in the west-central part of Ahmedabad city at BH-01. At periods 0.55 s (corresponding to 5–6 story buildings), the maximum mean SA of ~ 270 cm/s2 (31% less than the national code) is computed in the southeastern part of the city at BH-06 (Fig. 16b). At periods 1.0 and 1.25 s, the highest mean SA of 110 cm/s2 and 100 cm/s2 (~ 60% less than the national code), respectively are observed in the southeastern part (at BH-06) and northwestern part (at BH-14) of the city (Fig. 16c, d) and lowest in the west-central part of the city (at BH-02).

The national disaster management authority (NDMA) has proposed the probabilistic seismic hazard assessment map of India at A-type soil class (Vs30 of 1500 m/s) (NDMA 2010) and suggested a PGA value of ~ 140 cm/s2 in Ahmedabad at 2% probability of exceedance (2% PE) in 50 years. Parvez et al. (2003) have conducted the deterministic seismic hazard assessment of India based on the Modal Summation technique of Costa et al. (1993) and suggested the design ground acceleration of ~ 150–300 cm/s2 in Ahmedabad city at the surface level. Chopra et al. (2012b, a) have conducted the deterministic seismic hazard assessment of Gujarat considering far-field earthquake using Stochastic finite fault modeling technique and suggested the PGA of 41 cm/s2 at the surface level. Rao et al. (2012) have conducted the classical deterministic seismic hazard assessment of Ahmedabad city and suggested the PGA in the range of 140–440 cm/s2 at rock level. Trivedi (2011) has conducted the de-convolution analysis considering Far-field earthquake considering single accelerogram of 2001 Bhuj earthquake recorded at the passport office building of Ahmedabad and suggested the PGA of the order of 96–194 cm/s2 that matches with the PGA estimated through far-field earthquake scenarios in the present study (100–162 cm/s2). The results of the present study match well with the regional study of Parvez (2003) and the study carried out by NDMA (2010), after incorporation of soil effects (amplification from A to D type soil class).

Ahmedabad has already experienced damages due to two large earthquakes in the past (1819-ABEQ and 2001-BEQ). The shaking minarets of the Jama Masjid located in the central part of Ahmedabad (near Lal Darwaja area) (Fig. 12) were damaged due to the 1819-ABEQ. The current investigation has also suggested the high PGA in the range of ~ 135 cm/s2 at the Jama Masjid site due to the far-field earthquake scenarios, which is analogous to 1819 or 2001 earthquakes (Fig. 14). In Ahmedabad, 80 buildings of 1021 flats and 82 other houses had collapsed during the 2001 Bhuj Eq. (Mishra 2004) mainly in the west-central part [in the vicinity of BH-01, BH-17, and BH-19 (Fig. 14)] of Ahmedabad city. In the present investigation, the maximum PGA values of 135–150 cm/s2 are also estimated in the west-central part (Fig. 14).

Most of the collapsed buildings were close to BH-01 and BH-19 where clay is present at/near-surface, with a very low N-value, low shear wave velocity and medium plasticity. The clay with interbeds of silty sand is present in the subsurface. In the Loma Prieta earthquake (Ms 7.1) that occurred on October 17, 1989, the maximum (up to 80%) loss of lives occurred about 50 miles away (at Treasure Island) from the fault rupture zone. Ferritto et al. (1993) from Naval Civil Engineering Lab, California have conducted the ground response analysis at Treasure Island, which is a naval base. The soil of Treasure Island consists of bay mud deposits, a high plasticity silty clay (Ferritto et al. 1993). The study concluded that high plasticity clay deposits can be a possible source of ground motion amplification. The Clay with medium plasticity found at BH-01 and BH-19 (in the western central area of Ahmedabad) might be one of the causes of devastation.

5 Conclusion

The micro-level seismic hazard assessment is carried out at Ahmedabad city through characterization of near-surface using seismological, geotechnical and geophysical investigations. The ground response analysis is carried out at 20 boreholes varying in depths from 40 and 80 m. The study highlighted the following points:

  1. (1)

    In Ahmedabad city, 3–18 m thick clay (from the surface) is found west of the Sabarmati River and sand layer with a thickness of 2–31 m is found in the eastern part of the city, east of Sabarmati River (except SE part, where clay thickness ~ 15 m is present).

  2. (2)

    The mean surface PGA of the order of 101–279 cm/s2 (maximum PGA, 70% higher than the national code) is computed for the near-field earthquake scenarios and 118–162 cm/s2 due to far-field earthquake scenarios.

  3. (3)

    Comparatively higher PGA is computed in the western portion of the Ahmedabad city (due to near-field earthquake scenarios) than the eastern portion. The highest PGA of 279 cm/s2 is found in the west-central part of the Ahmedabad city due to near-field earthquake scenarios. The highest mean PGA of 162 cm/s2 is observed in the west-central portion of the city due to far-field earthquake scenarios. The damage during the 1819 Allah Bund and 2001 Bhuj earthquakes have occurred in the west-central part of the city.

  4. (4)

    In the case of near-field earthquake scenarios, the mean SA is found to be 70% more than the National code for 0.2–0.3 s (2–3 story buildings), ~ 30% less for 0.55 s (5–7 story buildings) and ~ 50% less for 1–1.25 s (10–15 story high rise buildings). For far-field earthquake scenario, the SA is found to be 28% more than the National code for 0.2–0.3 s (2–3 story buildings), ~ 31% less for 0.55 s (5–7 story buildings) and ~ 40% less for 1–1.25 s (10–15 story high rise buildings).

  5. (5)

    The amplification factor of 1.6–3.3 is found from both earthquake scenarios in the Ahmedabad region.