Deterministic Seismic Microzonation of the NCT-Delhi (India) and Earthquake Engineering Implications

8 9 A deterministic seismic microzonation of the NCT Delhi (The capital of INDIA) and its earthquake 10 engineering implications is presented in this paper. The NCT Delhi with population density around 11 21,000/sq. Km has experienced several severe earthquake shakings in the past due to 12 earthquake occurrences in its vicinity and in the Great Himalaya. The exposed central quartzite 13 ridge, Badarpur-Okhala hillocks and River-Yamuna are responsible for the very large spatial 14 variation of sediment thickness (10 m to more than 300 m) in the NCT Delhi. The dynamic 15 properties of sediment layers over the quartzite basement at 158 sites, well distributed in the NCT 16 Delhi, are considered for seismic microzonation. First, we have finalised the maximum credible 17 earthquake (MCE) for each considered site based on the deterministic seismic hazard analysis. 18 Thereafter, acceleration time history at basement level is computed at each site using stochastic 19 finite-fault method with dynamic corner frequency and the geometry as well as rupture-dimension 20 of the respective MCE. The basement ground motion is numerically transferred to the free surface 21 using the rheological parameters and thickness of sediment layers overlying the quartzite 22 basement. Different maps of earthquake engineering interest like peak ground acceleration 23 (PGA), peak ground velocity (PGV) and peak ground displacement (PGV) at basement level and 24 the free surface level are developed and analysed for earthquake implications. The obtained 25 range of PGA (0.08-0.30g), PGV (3.34-26.58cm/s) and PGD (0.55-7.2cm) at the free surface and 26 fundamental frequency of the sediment deposit (0.4-7.0Hz) reveals that the NCT Delhi needs 27 special attention by the planners, engineers and decision makers for earthquake disaster


INTRODUCTION 50 51
The seismic microzonation of an area taking into account the source, path and site effects is 52 essential for the prediction of relevant seismic parameters for the earthquake engineering 53 designs, land use planning, retrofitting, seismic disaster reduction, building insurance and risk 54 assessment (Oprsal et al., 2005;Wang, 2008; Anbazhagan and Sitharam, 2008; Shiuly and 55 Narayan, 2012). Earthquake engineers estimate design forces considering PGA and the response 56 spectra as per building to be designed (IS-1893:2002 (Part 1)). The response spectra to some 57 extent takes into account the effects of fundamental and higher modes of vibration of structure. 58 However, under double resonance condition, the dynamic forces during earthquake may be much 59 larger than that of predicted using current practice and building may not survive. The dynamic 60 force may increase by a factor of 3-5 times under double resonance condition (Romo and Seed 61 1986; Kumar and Narayan, 2018). For example, the unexpected selective damage to the high-62 rise buildings in the Ahmedabad city at epicentral distance more 350 km took place due to the 63 occurrence of double resonance during the 2001 Bhuj earthquake (Narayan et al., 2002). The 64 seismic microzonation of NCT Delhi (The capital of India) seems essential considering highly 65 lateral variation of local geology (thickness and rheological parameters of the sediment deposits 66 above the quartzite rock) due to the presence of exposed central quartzite ridge, Badarpur-Okhala 67 hillocks and the Yamuna river, high population density (11,320 per sq km), the accelerated 68 6 due to anthropogenic activity and many lines and natural ponds have been altered or obliterated. 169 The oldest exposed geological section in the region is middle to upper Proterozoic Delhi Super-170 group. The Delhi Super-group is overlain by older Alluvium (unconsolidated Quaternary 171 sediments) of Late Pleistocene and recent Alluvium Holocene epoch. The Delhi Super-group 172 composed of gritty quartzite, quartzite, arkosic grit with lean intercalations of micaceous schist. 173 Delhi Super-group rock intruded through quartz and pegmatite veins. The older Alluvium mainly 174 composed of occasionally white micaceous, yellowish-brown, medium to fine sand, silty-clay, silt, 175 clay and kankar. The Recent Alluvium is limited to the flood plain of Yamuna stream and primarily 176 comprises grey micaceous medium to fine-grained sand, intercalations of clay and sediment 177 along fine nodular kankar. NCT Delhi has mainly three extensive Geomorphological units called 178 exposed rock Quartzite, older Alluvial smoothly undulating surface along rolling topography and 179 low lying surface of Yamuna River flood plain (Kazim et al., 2005). studies, it has been found that a few major geomorphological features known as Lahore-Delhi 196 edge, Delhi Haridwar ridge, Himalayan Frontal Fold region and Delhi axis of folding are following 197 the regional trends (Srivastava and Roy, 1982). Criss-cross lineaments near Delhi (Hukku, 1966;198 Mehta et al., 1970 andSharda, 1996) indicate the complexity of the zone probably due Mathura fault is trending in NE-SW direction. Figure 1 depicts that the study region is surrounded 202 both the deterministic and stochastic aspects of the ground motion (Brune, 1970;Hanks and 241 McGuire, 198;Boore, 2003). The stochastic aspects of ground motion are modelled as Gaussian 242 white noise with the specified underlying spectrum (Boore, 1983;2003). The deterministic aspects 243 are defined by the mean Fourier spectrum as the multiplication of the omega square source 244 model, path effect and site effects (Brune, 1970). The extension of stochastic method to finite-245 (1) 257 Where = 4 3 is scaling factor, is the radiation coefficients averaged over the range of 258 azimuth and take off of angle, is free surface effect, ρ is the density (g/cc) of crust at the focal 259 depth and β is the shear wave velocity (km/s) in the source zone. , is a normalization factor 260 that aims to conserve high-frequency spectral level of ij th sub-fault. 0 , ( ), 0 , and , 261 represent ij th sub-fault corner frequency, seismic moment and distance of site from the sub-fault, 262 respectively. The terms ( , ) and Q(f) represent the geometrical spreading and quality factor, 263 respectively. ( ) and (− ) represent the spectral amplification factors and high frequency 264 spectral decay, respectively and k is the Kappa value (Anderson and Hough, 1984). The corner 265 frequency for a particular sub-fault is computed using the following formula 266 9 0 ( ) = 4.9 * 10 6 ( ( )) −1/3 1/3 ( The estimated S-wave velocity for the sediment layer of thickness 100 m above the basement at 328 site-114 is 689 m/s (Table 3). 329 The density (ρ) in gm/cc of each sediment layer is computed in terms of Vs (m/s) using an 330 empirical relationship (Eqn. 4) developed by Kumar and Narayan (2020). 331 The quality factors (Q) for sediment layers with S-wave velocity in range of 175 m/s to 610 m/s 333 are obtained using the empirical relation proposed by Iyasen (1996) and for Vs more than 610 334 m/s, Q is taken as simply 10% of Vs (Rao et al., 2006). 335 = 0.08 + 6.99 (5) 336 In order to incorporate frequency dependent damping in the time domain simulations using GMB-337 EK rheological model, it is assumed that the obtained S-wave velocity and quality factor in each 338 layer are measured in the field using the signal with 1.0 Hz frequency (Emmerich and Korn, 1987). (0.87 cm) at the basement level; which is in accordance with the Brune's model (Brune, 1970). 410 Further, the obtained different sediment amplification factors for a particular parameter (say PGA) 411 in the case of ground motion due to MCEs on MDF, MF and SF at site114 may be due to the 412 change of spectra with magnitude, fault parameters, focal mechanism and epicentral distance. 413

414
In the past, some of the scientists have used average spectral amplification (ASA) caused by 415 sediment deposit to transfer the predicted PGA at the basement level to compute the same at the 416 free surface. For example, the computed PGA at free surface at site2, site114 and site96 using 417 ASA are 2.0, 1.6 and 1.5 times larger than that obtained at free surface based on the wave 418 propagation, respectively (Table 1). In the case of PGA prediction using ASA, the over prediction 419 of the PGA is increasing with the increase of sediment thickness, which is obvious one. So, it may 420 be concluded that basement ground motion should be transferred to the free surface using 421 seismic wave propagation taking into account the rheological parameters and thicknesses of the 422 sediment layers above the basement. The pattern of spatial variation of PGA* at the free surface in the NCT Delhi is shown in figure 12. 497 The range of PGA* variation in the NCT Delhi is 0.12g to 0.53g, which is much larger than the 498 range of PGA at free surface (0.08g-0.30g). Further, the obtained PGA* is larger than PGA at all 499 the sites. This may be due to the obtained range of ASA variation for all the sites of the NCT Delhi 500 is 2.25-4.89. So, it may be concluded that the basement/bedrock ground motion should be 501 We obtained the lower PGV (≤8 cm/s) at localities from Bhaktawarpur to Wazirabad on the central 516 ridge and surrounding area (sites152, 153, 96, 156, 154, 115, 128, 129). In the eastern region, 517 PGV≥15 cm/s was obtained at site near Jaitpur Police station (site143) and at rest of the sites 8 518 cm/s <PGV<15 cm/s (sites101, 102, 103, 155, 157, 158, 124, 131, 132, 135, 136). We got large 519 PGV (≥15 cm/s) at localities of the western region like Nazafgarh (site40), Dwarka (sites56, 57), The spatial variation of PGD at the free surface in the NCT Delhi is shown in figure 14. The lowest 526 PGD value of the order of 0.55 cm is obtained in Pusta-4, Usmanpur (site129) and highest one 527 as 7.2 cm in Narela locality (site76). The computed effect of sediment thickness based on the 528 seismic wave propagation on the PGD is very interesting. The thick sediment deposit is amplifying 529 the PGD and reverse is the finding in the case of shallow sediment deposit. For example, 530 amplification of PGD is obtained at localities like Qutubgarh (sites12, 13) and Jharoda Kalan 531 (site19) where sediment thickness is more than 300 m; and deamplification is obtained at localities 532 lying on central ridge from Bhakhtawarpur (site152) to Pusta-4, Usmanpur (site129) where 533 sediment thickness is less than 30 m (Table 1). Almost no amplification of the low frequency 534 seismic waves due to shallow sediment deposit may be the reason behind this observation. We 535 obtained very less PGD (<1.0 cm) at localities from Bhaktawarpur to Wazirabad on the central 536 ridge and surrounding area (sites152, 153, 96, 156, 154, 115, 117, 99), which are underlain by 537 either out-cropping or shallow quartzite rock. In the eastern region, 1.0 cm<PGD<3.0 cm is 538 obtained at sites which are near or east of the Yamuna River (sites131, 132, 135, 136, 140, 141,  539   150, 138, 142, 143) and at rest of the sites PGD is <1.0 cm. In the western region, large PGD (>3 540 cm) was also obtained in localities like Nazafgarh (site40) , Jharodha (sites18, 19), Karol Bagh 541 (sites75, 76), Jalkhor, Puth-Khurd, Dariyapur Kalan (sites12, 17, 26, 27, 28, 45, 49) and Narela 542 Mandi (sites43, 44, 144, 60, 61, 63, 66, 79). At rest of the localities of the western region, the 543 range for the PGD variation is 1 cm to 3 cm. 544 545

EARTHQUAKE ENGINEERING IMPLICATIONS 546 547
Most of the buildings of the NCT Delhi can be grouped in to two categories namely "B" type and 548 "C" type, as per MSK intensity scale. "B" type buildings are ordinary brick buildings and stories ≤4 549 and "C" type buildings are mostly well build RC buildings. In order to achieve a specified level of 550 performance of the building when exposed to seismic hazard, the performance-based design 551 reflects a more general design criterion. Design based on displacement can be regarded as a 552 subset of performance-based design. The pseudo spectral acceleration (PSA) corresponding to 553 the resonance frequency of building can increase by a factor more than 4 under double resonance 554 condition (Kumar and Narayan, 2018). The same may be the amplification scenario for the 555 velocity and displacement response spectra. So, an increase of level of damage to a structure 556 under double resonance condition may be equivalent to an increase of intensity value by a factor 557 of 1-2 units, as was observed in Ahmedabad city during the 2001 Bhuj earthquake (Narayan et 558 al., 2002). Therefore, we have also considered the PGV and PGD in order to infer the expected 559 level of damage to medium and high-rise buildings, respectively under double resonance 560 condition. The expected grade of damage (G1-G5) which may occur to the buildings in the NCT 561 Delhi as per predicted ground motion parameters is described taking in to consideration the MSK 562 intensity scale and a relation between acceleration and the intensity (IS: 1893 (Part 1), 2002). 563 a.

Region of NCT Delhi underlain by shallow/out-cropping quartzite rock 564
The obtained PGA, PGV and PGD in the central ridge and surrounding regions is less than 0.12g, 565 8.0 cm/s and 1.0 cm, respectively (Table 1). The fundamental frequency of sediment deposit is 566 mostly larger than 2.0 Hz (Kumar and Narayan, 2020). For example, it is more than 5 Hz at 567 Bakhtawarpur, JNU and Shalimar Bagh (sites152, 153, 156), Karol Bagh (sites115, 154) as well 568 as more than 2 Hz at site128 and site129. Under non-double resonance condition, many B-type 569 buildings may suffer with G1 and few G2 grade gamage and few low-rise C-type buildings may 570 suffer with G1 grade damge. However, under double resonance condition, the low-rise B-type 571 buildings may suffer with G3 grade and low-rise C-type buildings may suffer with G2 grade damge. 572 However, the high-rise and medium-rise buildings are safe in this region due to less values of 573 PGV and PGD. However, relatively larger PGA (0.12-0.18g) and PGV (<10 cm/c) are obtained at 574 Delhi Univ. (site116; F0 around 2.5 Hz) and sites lying east of river Yamuna (sites131,133 with F0 575 1.3-1.5 Hz). In these localities under non-double resonance condition, many B-type buildings may 576 suffer with G2-G3 grade gamage and many low-rise C-type buildings may suffer with G1-G2 577 grade damge. However, under double resonance condition, B-type buildings may suffer with G3-578 G4 grade and low-rise as well as medium-rise C-type buildings may suffer with G2-G3 grade 579 damage. Arjun Nagar (site136) and near Ghazipur (site140). As mentioned above, in the localities with 589 F0>2.0 Hz and under non-double resonance condition, many B-type may suffer with G1-G2 grade 590 and few C-type may suffer with G1 grade damage. However, the B-type and low-rise C-type 591 buildings in these localities may sufer with G2-G3 grade and G2 grade damage, respectively 592 under double resonance condition. On the other hand, the medium-rise buildings in localities 593 falling east of river Yamuna like Gita colony, Gokulpur (site134), Mansarovar Park (site135), Arjun 594 Nagar (site136) and near Ghazipur (site140) may suffer with only G2 grade damage since PGV 595 is less than 8-10 cm/s. 596

597
The range of F0 of sediment deposit in the Chhatarpur basin and nearby semiclosed basin is 1.8 598 -3.2 Hz and the range of obtained PGA is 0.12g -0.17g. In the localities like Silampur (site130), 599 Harsh Vihar (site132) and Gazipur (site141), the range of PGA is same but range of F0 of 600 sediment deposit is 1.30 Hz to 1.50 Hz. So, many B-type buildings in these localities may sufer  Map depicting important localities, considered sites (stations) and basement depth variation in NCT Delhi (Modi ed after CGWB (2010-11)). Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.  Deterministically predicted acceleration, velocity and displacement time histories at free surface (left panels) and basement level (right panels) at site2 (Mandhela Khurd) with sediment thickness 320 m using corresponding MCE as MW7.1 on Mahendergarh Dehradun fault.

Figure 5
Deterministically predicted acceleration, velocity and displacement time histories at free surface (left panels) and basement level (right panels) at site96 (CISF Rd., Mahipalpur Extn.) with sediment thickness 9 m using corresponding MCE as MW7.1 on Mahendergarh Dehradun fault.   Map depicting the variation of peak ground acceleration (PGA) at basement level in the NCT Delhi region.
Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors. Map depicting the variation of PGV at basement level in the NCT Delhi region. Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors. Map depicting the variation of PGD at basement level surface in the NCT Delhi region. Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors. Map depicting the variation of PGA at the free level in the NCT Delhi region. Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors. Map depicting the PGA* obtained using ASA at the free surface and its spatial variation in the NCT Delhi region. Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors. Map depicting the variation of PGV at the free surface level in the NCT Delhi region. Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors. Map depicting the variation of PGD at the free surface level in the NCT Delhi region. Note: The designations employed and the presentation of the material on this map do not imply the expression of any opinion whatsoever on the part of Research Square concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries. This map has been provided by the authors.