See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/226272046

Simulation of the Arabian Sea Tsunami propagation generated due to 1945 Earthquake and its effect on western parts of...

Article in Natural Hazards · January 2009 DOI: 10.1007/s11069-008-9261-3

CITATIONS READS 39 437

3 authors, including:

A. P. Singh B.K. Rastogi Institute of Seismological Research Institute of Seismological Research

55 PUBLICATIONS 311 CITATIONS 156 PUBLICATIONS 1,756 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

site response study View project

All content following this page was uploaded by A. P. Singh on 23 December 2013.

The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document and are linked to publications on ResearchGate, letting you access and read them immediately. Nat Hazards (2009) 48:245–258 DOI 10.1007/s11069-008-9261-3

ORIGINAL PAPER

Simulation of the Arabian Sea Tsunami propagation generated due to 1945 Makran Earthquake and its effect on western parts of Gujarat ()

R. K. Jaiswal Æ A. P. Singh Æ B. K. Rastogi

Received: 15 April 2008 / Accepted: 30 May 2008 / Published online: 1 July 2008 Ó Springer Science+Business B.V. 2008

Abstract The 1945 Tsunami generated due to Makran Earthquake in the Arabian Sea was the most devastating tsunami in the history of the Arabian Sea and caused severe damage to property and loss of life. It occurred on 28th November 1945, 21:56 UTC (03:26 IST) with a magnitude of 8.0 (Mw), originating off the Makran Coast of in the Arabian Sea. It has impacted as far as Mumbai in India and was noticed up to Karvar Coast, Karnataka. More than 4,000 people were killed as a result of the earthquake and the tsunami. In this paper an attempt is made for a numerical simulation of the tsunami generation from the source, its propagation into the Arabian Sea and its effect on the western coast of India through the use of a numerical model, referred to as Tunami-N2. The present simulation is carried out for a duration of 300 min. It is observed from the results that the simulated arrival time of tsunami waves at the western coast of India is in good agreement with the available data sources. The paper also presents run-up elevation maps prepared using Shuttle Radar Topographic Mission (SRTM) data, showing the possible area of inundation due to various wave heights along different parts of the Gujarat Coast. Thus, these results will be useful in planning the protection measures against inundation due to tsunami and in the implementation of a warning system.

Keywords Tsunami Á Arabian Sea Á Makran Coast Á Impact on Gujarat Coast Á Inundation limits

1 Introduction

Tsunami waves are the most underrated hazard affecting the population of the world living near the coastal belts. With the increasing rate of economic development of coastal , there is also an increase in socio-economic consequence resulting from the hazardous action of tsunami waves generated from submarine seismic activity and other causes. Therefore, the accurate early warning of tsunamis for a coastal community is of great

R. K. Jaiswal (&) Á A. P. Singh Á B. K. Rastogi Institute of Seismological Research, Sector 18, Gandhinagar 382 018, Gujarat, India e-mail: [email protected] 123 246 Nat Hazards (2009) 48:245–258

Fig. 1 Historical great earthquakes of Makran Subduction Zone importance. One of the most deadly tsunamis ever recorded in the Arabian Sea occurred at 21:56 UTC (03:26 IST), on November 28, 1945, due to occurrence of a great earthquake of Mw 8.0 located at 25.204°N 63.420°E (Fig. 1), in the northern Arabian Sea, about 100 km south of and about 87 km SSW of Churi (Baluchistan), Pakistan (Pararas-Cara- yannis 2006). More than 4,000 people were killed by both the earthquake and the tsunami (Ambraseys and Melville 1982). The earthquake was also characterized by the eruption of a , a few kilometers off the Makran Coast, which are common features in Western Pakistan. It led to the formation of four small islands. A large volume of gas that erupted from one of the islands sent flames leaping hundreds of meters into the sky (Mathur 1998). The tsunami reached a height of 12 m in some Makran ports and caused great damage to the entire coastal (Pararas-Carayannis 2006). A great number of people were washed away due to the tsunami. The tsunami was also recorded at and Gawadar. The wave was 1.5 m high at Karachi, 2 m near Bombay, now called Mumbai (1,100 km away), 0.5 m in the Seychelles (3,400 km away), and caused notice- able effects at Karwar (1,500 km distant) and at Muscat. The transoceanic cable between India and England broke at eight places, indicating widespread slumping offshore. Part of moved, with one submarine slide shifting the coast 100 m landward. Coastal uplift at was 2 m. The wave had a run-up (i.e. flooding of dry coastal areas) height of 11.0– 11.5 m in Kutch, Gujarat (Pendse 1948). The first wave was recorded at 05:30 am, then subsequent waves arrived at 7:00 am, 7:15 am and, finally at 8:15 am. The last wave at 8:15 was the biggest. At 8:15 am, it was observed on Salsette Island, i.e. Mumbai (Newspaper archives, Mumbai). It was recorded in Bombay Harbour, Versova (Andheri), Haji Ali (Mahalaxmi), Juhu (Ville Parle), and Danda (Khar). At Versova (Andheri, Mumbai), 5 persons who were fishing were washed away. At Haji Ali (Mahalaxmi, Mumbai), 6 persons were swept into the Sea. At Danda and Juhu, several fishing boats were torn off their moorings. The tsunami did not do any damage to Bombay Harbour. Most persons who witnessed the tsunami said that it rose like the tide coming in, but much more rapidly. The height of the tsunami in Mumbai was 2 m. A total of 15 persons were washed away in Mumbai (Rastogi and Jaiswal 2006). Although large earthquakes along the Makran Subduction Zone are infrequent, the potential for the generation of destructive tsunamis in the Northern Arabian Sea cannot be ruled out. It is quite possible that historical tsunamis in this region have not been properly 123 Nat Hazards (2009) 48:245–258 247 reported or documented. Such past tsunamis must have affected Southern Pakistan, India, , , and possibly other areas as well. The sesimotectonics of the Makran Sub- duction Zone, historical earthquakes in the region, and the recent earthquake of October 8, 2005 are indicative of the active tectonic collision process that is taking place along the entire southern and southeastern boundary of the Eurasian plate as it collides with the Indian plate and adjacent microplates. Tectonic stress transference to other, stress loaded tectonic regions could trigger tsunamigenic earthquakes in the Northern Arabian Sea in the future. While earthquakes cannot be predicted in advance, once the signature of an earthquake is detected, it is possible to give about a few minutes to a few hours of notice of a potential tsunami to the coastal stations depending upon the location and how close or far it is from the earthquake epicenter. The timely and reasonably accurate early warning of the tsunami can save lives and also possibly mitigate the damage to properties. In view of this, an attempt is made here to simulate the tsunami waves generated due to earthquakes along the Makran Subduction Zone of the Arabian Sea that can affect the western coast of India.

2 Tectonic and tsunamigenic characteristics of the Makran Subduction Zone

In order to fully understand the nature of the earthquakes that may generate tsunamis, the plate boundaries and their movement must also be understood. Tectonic activity due to plate movement is the principal cause of earthquakes, 80% of which occur along the plate boundaries in the oceanic crust. To the northwest of , for millions of years, the Indian plate has been drifting and moving in the northeast direction with the Arabian plate and subducting beneath the Iranian microplates of the Eurasian block and has created an active subduction zone along the Makran Coast of Pakistan. The subducting plate has a northward dip of greater than 20°, till of 27°N, then bending down to an angle of *300. Makran Subduction Zone is one of the largest sedimentary accretionary wedges on the earth, covered with up to 7 km of thick sediments. Due to sudden slumping along Makran accretionary wedges with large amount of sediments may generate large tsunami, somewhat similar to the 1992 Nicaragua earthquake of Ms 7.0–7.3 (Piatanesi et al. 1996) which generated a large tsunami due to fall of accretionary wedges. The 1,200 km long Makran Subduction Zone of Iran and Pakistan (boundary between Iran and Pakistan runs roughly N–S at about 62°E in the coastal region) is seismically not as active as the Himalaya or Sunda Arc, but has produced great earthquakes and tsunamis in the past. The Arabian plate is converging northward into Makran Subduction Zone with an average speed of 4 cm/year (DeMets et al. 1990). Oman oceanic lithosphere slips below the Iranian micro-plate. Thrust faults are oriented nearly perpendicular to the direction of convergence. A major fault in this region has produced several tsunamigenic earthquakes. Further south, the western side of the Indian tectonic plate is bounded by the Central Indian and Carlsberg mid-ocean ridges. This is a region of shallow seismicity. Five of the great earthquakes in Makran may have ruptured the plate boundary in four different rupture segments of lengths of about 200 km each in 1483 (58–60°E), 1851 and also 1864 (61–63°E), 1945 (63–65°E), and 1765 (65–67°E) (Byrne et al. 1992) (Fig. 1). The 1765 earthquake was felt strongly at Karachi in easternmost Makran. Two events occurred in 1851 and 1864 in the same area affecting the town of (Quittmeyer 1979) in eastern Makran. Out of all these earthquakes only the 1945 earthquake is known to have caused a large tsunami. It struck the coast of eastern Makran near Pasni, followed 123 248 Nat Hazards (2009) 48:245–258 by a large aftershock in 1947 immediately to the south. The list of historical tsunamis that affected west coast of India and vicinity is given in detail in Table 1. The oldest record of tsunami is available from November 326 BC earthquake near the Indus delta/Kutch region that set off massive sea waves in the Arabian Sea (Rastogi and Jaiswal 2006). The western Makran zone has no clear record of historic great earthquakes. An earth- quake in 1483 affected the Strait of Hormuz and northeast Oman and may therefore have occurred somewhere in western Makran, but exact location is not known. Modern instruments have also not detected shallow thrust events. Most earthquakes in this region occur within the down-going plate at intermediate depth. Absence of frequent earthquakes indicates either that aseismic subduction occurs or that the plate boundary is currently locked and experiences great earthquakes with long repeat periods. Evidence is presently inconclusive without Global Positioning System (GPS) measurements and knowledge of velocity structure. However, presence of well-defined late Holocene terraces along portions of the coasts of eastern and western Makran could be interpreted as evidence that both sections of the arc are capable of generating large plate boundary earthquakes (Byrne et al. 1992).

3 Data and methodology

Tsunamis are classified as long shallow-water gravity wave or long waves. As such, their propagation is strongly affected by ocean depth changes. For the modeling of tsunamis, open source bathymetry and topography data viz. General Bathymetric Chart of the Oceans (GEBCO), SRTM with C-Map, NHO chart, and RTKGPS (Real Time Kinematic Global Positioning System) data for coastal and near shore regions are used. In this study four nested domains, namely A, B, C, D are used (Fig. 2). The increased resolution is essential in order to simulate as best as possible the travel time and tsunami amplitude of the waves. The intermediate grid (B) allows for a better resolution all around the Arabian Sea. For grids A and B the model is run in the linear mode which, although not good enough for run-up estimates, is good enough for travel time estimates. Another reason for increasing the resolution as we go into shallower water is the fact that (Shuto et al. 1985, 1986) each tsunami wavelength should be covered by at least 20 grid points in order to diminish numerical dispersion (dissipation). Ramming and Kowalik (1980) found that 10 grid points per wavelength is sufficient if we are willing to accept a 2% error in the phase velocity. Still another reason is that numerical stability considerations require that the finite dif- 1/2 ferences time step be such that Dt B Dx/(2ghmax) , where Dx is the space discretization size, g the gravitational acceleration, and hmax is the maximum depth in the given grid. As the wave propagates into shallower waters hmax decreases and by decreasing Dx we can maintain a constant Dt (Goto and Ogawa 1992). In this study, the fault parameters of the earthquake used for the generation of tsunami are fault area length 200 km and width 100 km, angle of strike, dip, and slip 270°,15°, and 90°, respectively, focal depth 10 km, and magnitude of the earthquake M 8.0 (Mohan and Krishnamurthy 2007). Further for tsunami propagation we considered bathymetry, earth curvature, Coriolis force, ocean parameters such as tides, currents (speed and direction), and gravity waves (height, period, and direction). In this study for tsunami propagation we considered near shore bathymetry, land topography, coastal geomorphology, estuaries/ creeks/Inlets, bays, sand dunes, etc. The generation and propagation of tsunami waves are modeled using Tunami-N2 code. Tunami-N2 is authored by Fumihiko Imamura in Tahoku University, Japan, and developed 123 a aad 20)4:4–5 249 48:245–258 (2009) Hazards Nat Table 1 List of historical tsunamis that affected west coast of India and vicinity Sl. no. Date Location Long. °E Lat. °N Eq. mag Comment References

1. 326 BC Indus delta/Kutch region Alexander’s navy destroyed. Massive sea Lisitzin (1974) waves in the Arabian Sea due to large earthquake 2. 1008 Iranian Coast 60 25 Tsunami has been observed in the North Murty et al. (1999) Indian Ocean on the Iranian Coast from a local earthquake 3. 1524 Dabhol, Maharashtra, In. 73.2 17 Tsunami due to a large eq. caused Bendick and Bilham (1999) considerable alarm to the Portugese fleet assembled in the area 4. May 1668 Samaji—Delta of Indus 68 24 The town of Samawani (or Samaji) sunk Oldham (1883) into the ground with 30,000 houses during an earthquake 5. 16/06/1819 Kutch 71.9 26.6 7.8 The town of Sindri (26.6N 71.9E) and Macmurdo (1821) adjoining country were inundated by a tremendous rush from the ocean, and all submerged, the ground sinking apparently by about 5 m 6. 19/06/1845 Kutch 68.37 23.6 7.0 The sea rolled up the Koree mouth of the Nelson (1846) Indus overflowing the country as far westward as the Goongra river, northward to the vicinity of Veyre, and eastward to the Sindree Lake 7. 28/11/1945 Makran Coast 63.5 25.2 8.0 More than 4,000 people were killed on Murty et al. (1999) the Makran Coast by both the earthquake and the tsunami. Max. run up 17 m. The height of the tsunami in Mumbai was 2 m. A total of 123 15 persons were washed away in Mumbai 250 Nat Hazards (2009) 48:245–258

Fig. 2 The model domain with bathymetry and topography data

in Technical University (METU) and in the University of Southern California. It is an outcome of UNESCO TIME project. A leap-frog, semi-implicit time stepping integration scheme is used for the tsunami simulations. This allows the use of larger time steps while maintaining stability and accuracy (Morey et al. 2003; Rueda and Schladow 2002). However, if too large a time step is used and the Courant, Friedrichs, and Lewy (CFL) condition is violated, gravity waves may be slowed down (Bartello and Thomas 1996; Dupont 2001). The CFL condition states that the time step must be smaller than the time it takes for a wave to propagate from one grid point to the next. The model generates the water level displacement in model domain at given time intervals for all nested grids and maximum water level displacement at each grid cell independently of the time when it occurred. This array is the one used to examine the extent of inundation in the grids where the model is used in its non-linear mode.

4 The tsunami generation model

Typically, significant vertical deformation of the sea floor (i.e. a dip/slip earthquake) is required for tsunami generation. This deformation can be due to either isostatic rebound of an accretionary prism near a subduction zone or a change in crustal elevation (McCann 2006; Okal et al. 2003). The direction of movement, depth of deformation, length and width of the deforming fault or plate boundary, deformation dip and slip angles, and focal depth will determine the size of the tsunami (McCann 2006; Polet and Kanamori 2000; Zahibo et al. 2003). For example, a shallow subduction zone earthquake or an earthquake with a more vertical angle of deformation will usually displace a larger volume of water and consequently generate a larger tsunami (Bilek and Lay 2002; Polet and Kanamori 2000). The overlying geology also determines whether a tsunami will result from an earthquake (Bilek and Lay 2002; Kanamori 1972). There may be stronger motion at the sea

123 Nat Hazards (2009) 48:245–258 251

floor than the measured seismic moment would indicate. According to available evidence, the 1945 Makran Coast event was due to a submarine earthquake fault located on the Makran Subduction Zone. It is called a near-field tsunami, because it was generated close to the affected area. The initial condition consists of a sea surface deformation which itself, in this case, is due to a vertical displacement of the sea bottom. In this study the vertical displacement of the sea bottom is calculated with the Mansinha and Smylie method (1971), and is assumed to be equal to the tsunami initial profile with no modification. The initial displacement is generated in the exterior domain (A), and it is interpolated into the higher resolution grids B, C and D. The end result is an initial sea surface profile that extends smoothly from the exterior, lower resolution, domain into the higher resolu- tion domains. This is the sea surface condition at time t = 0 s. That is, the hypothesis is that the sea bottom displacement is immediately reflected in a sea surface displacement. The first step in the gravity wave formalism is to determine the static vertical displacement of the seafloor. It is in the first step that earthquake source parameters relate directly to tsunami generation. Because when an earthquake, or fault motion, occurred the elastic dislocation theory shows such deformation that seafloor just above the fault is uplifted while above the deeper end of the fault is subsided. Therefore the initial wave at t = 0 (Fig. 3a) for simulation purpose has been computed by the fault parameters.

5 Tsunami propagation

Tsunami-N2 model is used for the propagation of tsunami waves for the event of November 28, 1945, that happened in the Makran region of the Arabian Sea. The tsunami propagation states at every 10-min interval are simulated. In this study, model outer domain has a horizontal resolution of 2502 m over the Arabian Sea including the Indian sub- (7– 26°N and 62–80°E). The simulation is carried out for a duration of 300 min and the Sea states at 0, 30, 60, 90, 120, 180, 240 min in the Arabian Sea are presented in Fig. 3a, b. Because of the variability in the bathymetry of the Arabian Sea and the earthquake that triggered the tsunami waves, the wave amplitude varies with the propagation of waves. At t = 0 min, the wave amplitude that is shown in the bar with different colors, next to the simulation figure, shows red at the point of epicenter. This indicates the wave height is in the range of 5–6 m on the land-ocean boundary. At t = 30 min, the wave starts propagating toward the Makran and the western coast of India. The wave amplitude varies with the forward motion of the tsunami waves. Boundary conditions play a significant factor in the separation of the land and ocean boundary. Further, the red color indicates in Fig. 3a, b that the water surface is higher than normal, while the blue means lower. Because of the fault geometry, the waves propagating to the Makran Coast begin with a receding wave, which explain why the Sea started to recede minutes before flooding the coast. From Fig. 3a, b, it could be observed that the tsunami wave propagated initially very fast in the Arabian Sea, and it became slower as it reached the shallow region of Gujarat Coast, Gulf of Kachchh, and Dwarka Coast. The tsunami strikes Jakhau and Porbandar coast with more than 2.5 and 1.5 m amplitude and 2.0 m at the Dwarka region. At t = 0, the source (Fig. 3a), it generates 6–7 m tsunami at the moment of the earthquake, and then the water is receded. At Dwarka, positive tsunami waves arrived within approximately 2 h 10 min and for the Gulf of Kachchh it takes 3 h 10 min. Based on these results, it is suggested that if the tsunami strikes in future during high tide, we should expect more serious hazards which impacts local coastal communities. But, the study of tsunami propagation along Gulf region shows that there is no impact during low 123 252 Nat Hazards (2009) 48:245–258

Fig. 3 (a) Tsunami wave travel times at 30, 60, 90, 120, 150 min. Tsunami amplitudes are in meters. (b) Tsunami wave travel times at 180, 210, 228 min. Tsunami amplitudes are in meters

123 Nat Hazards (2009) 48:245–258 253

Fig. 3 continued water in the inner gulf; however, a combined effect of high tide and tsunami wave has a great impact and it can inundate Mandvi area up to a greater extent. From Fig. 3a, b it is noticed that the tsunami generated due to an earthquake along the Makran Subduction Zone could reflect and refract the direct tsunami wave propagation through Murray ridge, Owen Fracture Zone, Carlsberg ridge, Laxmi ridge, Lakchhadeep, Maldives, and Diago Garcia Archipelago in the Arabian Sea. The tsunami run up at the coast is directly related to the beach slope, land elevation, settlement pattern, infrastructure facilities, etc. Figure 4 shows the maximum amplitude of the tsunami waves along Gulf of Kachchh at simulation time of 300 min. The bar on the right indicates the wave amplitude in meters. As the deep ocean tsunami approaches a distant shore, amplification and shortening of the wave occurs. This shows a trend that suggests the direction of the dissipation of wave energy. The dissipation is perpendicular to the major axis of the ellipsoidal fault plane. Another noticeable feature is the symmetry of the amplified waves. This can be attributed to the bathymetry of the ocean. A sea level gauge for a Tsunami Warning System (TWS) should be positioned to maximize warning time. Several factors such as population centers, locations where a 123 254 Nat Hazards (2009) 48:245–258

Fig. 4 Maximum wave amplitude (in meters) along Gulf of Kachchh tsunami may occur, travel time or propagation speed, and wave dissipation are considered when calculating warning time. Knowing where a tsunami will originate is essential to determining where a gauge should be installed. Because the tsunami wave propagates in both directions, those toward Makran Coast would first observe the receding wave, whereas those toward western parts of India would observe sudden rise in water. Such features of tsunami propagation were reproduced by computer simulation (Fig. 4). It shows that the water depression, or receding wave, propagate toward Makran Coast, whereas to the west, say toward western part of India, high water is traveling. The Makran Subduction Zone has infrequent large earthquakes. The frequency of occurrence of moderate to large thrust mechanism earthquakes is more in the eastern part of the Makran Coast. Less number of thrust events are known to have occurred along the plate boundary of the western Makran, probably suggesting the region is locked. If it is locked will the earthquake be catastrophic? If it is so the region should be closely moni- tored as the western region may be the site for the next large earthquake. Impact of tsunami along the west coast of India could be devastating as there is rapid growth, development, and larger concentration of human population. Also, the western coast is the most vul- nerable to tsunamis from the Makran region.

6 Inundation mapping

The western coastal region of India could be struck by tsunami waves due to occurrences of tsunamigenic earthquakes (magnitude [ 7.0) from the Makran Coast (Arabian Sea), Persian Gulf (Western most side of Makran Coast), Gulf of Aden, Diego Garcia Region (Indian Ocean), and Socotra Island Region (near North African Coast), though tsunami generated only from Makran Subduction Zone could severely affect the west coast of India. Historical seismic records indicate occurrences of earthquakes up to magnitude 8.0 in these regions but most of the earthquake locations were seismically inactive for the last 30 years or so. Since November 2005, an earthquake of magnitude 6.1 occurred at the Persian Gulf on 27th November 2005; there were frequent earthquakes of magnitude 6.0–6.5 in these regions. It has been observed that most of these regions have become 123 Nat Hazards (2009) 48:245–258 255 seismically active (Bapat 2007). Some of the events were monitored from tsunami point of view. It was observed from the records based on the tide gauges data in some ports of Gujarat Coast that the heights of wave coinciding with these events were in the range 55–65 cm. In this view inundation mapping is crucial for Gujarat, situated on the western coast of India, as this state is having the longest coastline (*1,600 km) in India and is also near to Makran Subduction Zone that could produce tsunamigenic earthquake in near future. The elevation data sets are the most important input for inundation mapping in tsunami prone areas. Preparation of a vulnerability map could inform coastal community and others about susceptibility to inundation corresponding to various wave-heights. The Gujarat state has important installations like ports, jetties, industries along the coast, and also other socio-economical perspective, which can be affected by tsunami trigged due to such events, and hence the determination of possible inundation areas is important. Inundation model is prepared for coastal parts of Gujarat belt on the basis of existing topographic and bathymetric (water depth) data sets. For preparation of the inundation map, high resolution of 3 arcs second or 90 m SRTM data set has been used. The Geo- graphical Information System (GIS) software is used to plot elevation data and different heights of tsunami waves depend upon the same. The different colors in Fig. 5 present various elevations as 0–2, 2–5, and 5–10 m are shown by green, pink, and dark blue colors, respectively, over the parts of Gujarat state. From this figure we can infer that the Gulf of Kachchh region has very low elevations, and therefore inundation would be more due to tsunamigenic conditions than other costal parts of Gujarat. In economic point of view, Gujarat coastal regions are very important where a number of ports/jetties and various industrial installations have been made. Gulf of Kachchh region is ecologically also very important as the various ports, viz. Kandla, Mandvi, Mundra, Dwarka, Okha, and Marine

Fig. 5 Possible inundations due to various wave heights (in meters) 123 256 Nat Hazards (2009) 48:245–258

National Park, are located in it which inhabits coral reefs, , etc. Smaller rise or fall in the water level can affect this valuable sensitive species and ports. Further, Banni (23.7°N, 69.56°E) region would inundate due to high tsunami wave but Rann of Kachchh could inundate due to low tsunami wave. From Fig. 5, it is also evident that the south east of Porbandar and north of Bhavnagar areas in Saurashtra could face inundation due to high tsunami wave. In south Gujarat, costal area from Khambhat to Navsari would inundate due to high tsunami wave amplitude, where also major ports and jetties are located. On the one hand, the Gulf of Kachchh and Gulf of Khambhat could be severely hit up to maximum extent of inundation and casualties would be too high, and on the other hand Rann of Kachchh could also inundating up to full extent but here casualties would be low as these regions are salty regions and there is not much population.

7 Summary and conclusion

The model has successfully simulated the propagation of tsunami event of November 1945. The model results are qualitatively consistent with the reported damage. Our model gives maximum amplitude along the creeks at the coast of Gujarat. So the most vulnerable areas of the coast need to be provided greater protection when planning for preparedness. Tide gauges should be installed where the tsunami amplitude is greater. For a tsunami generated from Makran Subduction Zone, which propagates across the Arabian Sea or Indian Ocean, there is half an hour to 1 h arrival time for more accurate and reliable warning for Gujarat Coast and western coast of India. Any large earthquake in the world can be located in 7– 15 min using seismic body waves recorded on the global seismic network. For accurate estimation of earthquake size, a few tens of minutes may be needed until surface waves are recorded around the globe. Because there is only half an hour to 1 h time before tsunami arrival at the Gujarat coast and western coast of India, it is very important to actually confirm the tsunami generation. For this purpose, sea level monitoring systems, located on western coasts and offshore of India, are necessary. The seismic and sea level data need to be shared in real-time, using satellite communication. An effective tsunami early warning system is achieved when all the persons in vul- nerable western coastal communities are prepared and respond appropriately, and in a timely manner, upon recognizing that a potentially destructive tsunami is approaching. Timely tsunami warnings issued by a recognized tsunami warning center are essential. When these warning messages are received by the designated government agency, tsunami emergency response plans must already be in place in coastal communities so that standard and efficient actions are immediately taken for evacuation, if necessary.

Acknowledgments The authors would like to acknowledge the TARU (research and consultancy group) for providing bathymetry data set for this experiment. Dr. M. V. Ramana Murthy of ICMAM, Chennai, is gratefully acknowledged by first author for discussion on TUNAMI-N2 model. Authors are thankful to Dr. Tad S. Murty and anonymous reviewers for their valuable suggestions that have helped in improving the manuscript. The work was carried out under a project sponsored by Dept. Sc. Tech., New Delhi.

References

Ambraseys NN, Melville CP (1982) A history of Persian earthquakes. Cambridge University Press, New York Bapat A (2007) Tsunamigenic vulnerability of west coast of India. In: Proc. 4th Indian National Conference on Harbour and Ocean Engineering (INCHOE). Published by National Institute of Technology, Karnataka, Surathkal, I, pp 426–431 123 Nat Hazards (2009) 48:245–258 257

Bartello P, Thomas SJ (1996) The cost-effectiveness of semi-lagrangian advection. Mon Weather Rev 124:2883–2897. doi:10.1175/1520-0493(1996)124B2883:TCEOSLC2.0.CO;2 Bendick R, Bilham R (1999) A search for buckling of the SW Indian coast related to Himalayan collision. In: Macfarlane A, Sorkhabi RB, Quade J (eds) Himalaya and Tibet: mountain roots to mountain tops. Geol Soc Am, Special paper 328, pp 313–322 Bilek SL, Lay T (2002) Tsunami earthquakes possibly widespread manifestations of frictional conditional stability. Geophys Res Lett 29:1673. doi:10.1029/2002GL015215 Byrne DE, Sykes LR, Davis DM (1992) Great thrust earthquakes and aseismic slip along the plate boundary of the Makran subduction zone. J Geophys Res 97:449–478. doi:10.1029/91JB02165 DeMets C, Gordon RG, Argus DF, Stein S (1990) Current plate motions. Geophys J Int 101:425–478. doi: 10.1111/j.1365-246X.1990.tb06579.x Dupont F (2001) Comparison of numerical methods for modeling ocean circulation in basins with irregular coasts. Dissertation thesis, McGill University, Montreal Goto C, Ogawa Y (1992) Numerical method of tsunami simulation with the Leap-Frog Scheme. Translated for the Time Project by Shuto N. Disaster Control Research Center, Faculty of Engineering, Tohoku University Kanamori H (1972) Mechanism of tsunami earthquakes. Phys Earth Planet Inter 6:346–359. doi: 10.1016/0031-9201(72)90058-1 Lisitzin E (1974) Sea level changes. Elsevier Oceanographic Series, No 8. New York Macmurdo C (1821) Account of the earthquake which occurred in India in June 1819. Edinb Philos J 4:106–109 Mansinha L, Smylie DE (1971) The displacement fields of inclined faults. Bull Seismol Soc Am 61: 1433–1440 Mathur SM (1998) Physical geology of India. National Book Trust of India, New Delhi McCann WR (2006) Estimating the threat of tsunamigenic earthquakes and earthquake induced landslide tsunami in the . Paper presented at NSF Caribbean Tsunami Workshop, World Scientific, San Juan, Puerto Rico Mohan RV, Krishnamurthy RR (2007) Early warning system development for tsunami and storm surges in Indian ocean. http://siteresources.worldbank.org/CMUDLP/Resources/Day4-3-ram_mohan_kris.pdf Morey SL, Martin PJ, O’Brien JJ, Wallcraft AA, Zavala-Hidalgo J (2003) Export pathways for river discharged fresh water in the northern Gulf of Mexico. J Geophys Res 108:3303–3318. doi: 10.1029/2002JC001674 Murty TS, Bapat A, Prasad V (1999) Tsunamis on the coastlines of India. Sci Tsunami Hazards 17(3): 167–172 Nelson C (1846) Notice of an earthquake and a probable subsidence of the land in the district of Kutch, near the mouth of Koree, or the eastern branch of the Indus in June 1845. Geol Soc Lond Q J 2:103 Okal EA, Alasset PJ, Hyvernaud O, Schindele´ F (2003) The deficient T waves of tsunami earthquakes. Geophys J Int 152:416–432. doi:10.1046/j.1365-246X.2003.01853.x Oldham TA (1883) Catalogue of Indian Earthquakes. Mem Geo Surv India 19(Pt 3):163–215 Pararas-Carayannis G (2006) The potential of tsunami generation along the Makran subducation zone in the Northern Arabian Sea case study: the earthquake and tsunami of November 28, 1945. Presentation at 3rd tsunami symposium of the tsunami society May 23–25, 2006, East-West Center, University of Hawaii, Honolulu, Hawaii, vol 24(5), pp 358–383 Pendse CG (1948) The Makran earthquake of the 28th November, 1945. Indian Meteorol Dept Sci Notes 10:141–145 Piatanesi A, Tinti S, Gavagni I (1996) The slip distribution of the 1992 Nicaragua earthquake from tsunami run-up data. Geophys Res Lett 23(1):37–40. doi:10.1029/95GL03606 Polet J, Kanamori H (2000) Shallow subduction zone earthquakes and their tsunamigenic potential. Geophys J Int 142:684–702. doi:10.1046/j.1365-246x.2000.00205.x Quittmeyer CR (1979) Seismicity variations in the Makran Region of Pakistan and Iran relation to great earthquakes. Pure Appl Geophys 117:1212–1228. doi:10.1007/BF00876216 Ramming HG, Kowalik Z (1980) Numerical modeling of marine hydrodynamics. Elsevier, New York, 368 pp Rastogi BK, Jaiswal RK (2006) A catalog of tsunamis in the Indian Ocean. Sci Tsunami Hazards 25(3): 128–143 Rueda FJ, Schladow SG (2002) Quantitative comparison of models for isotropic response of homogeneous basins. J Hydraul Eng 128:201–203 Shuto N, Suzuki T, Hasegawa K, Inagaki K (1985) Summary of a study of numerical techniques on the tsunami propagation and run-up. In: Murty TS, Rapatz WJ (eds) Proc Int Symp Inst Ocean Sci, BC, Canada, pp 88–92

123 258 Nat Hazards (2009) 48:245–258

Shuto N, Suzuki T, Hasegawa K, Inagaki K (1986) A study of numerical techniques on the tsunami propagation and run-up. Sci Tsunami Hazards 4:111–124 Zahibo N, Pelinovsky E, Yalciner AC, Kurkin A, Koselkov A, Zaitsev A (2003) The 1867 Virgin Island Tsunami: observations and modeling. Oceanol Acta 26:609–621

123

View publication stats