Tectonophysics 765 (2019) 146–160

Contents lists available at ScienceDirect

Tectonophysics
journal homepage: www.elsevier.com/locate/tecto

A comprehensive assessment of ground motions from two 2016 intra-slab T

earthquakes in Myanmar
⁎
Lin Thu Aunga, Stacey Servito Martina, Yu Wanga,b, , Shengji Weia, Myo Thantc,
Khaing Nyein Htayd, Hla Myo Aunge, Tay Zar Kyawc, Soe Minf, Kaung Sithug, Tun Naingh,
Saw Ngwe Khaingi, Kyaw Moe Ooj, G. Sureshk, Weiwen Chena, Phyo Maung Maunga,
Vineet Gahalautk
a
Earth Observatory of Singapore, Nanyang Technological University, Singapore
b
Department of Geosciences, National Taiwan University, Taiwan
c
Department of Geology, University of Yangon, Yangon, Myanmar
d
Gemological Institute of Myanmar, Yangon, Myanmar
e
Department of Civil Engineering, Sagaing Technological University, Myanmar
f
Department of Geology, Taungoo Univeristy, Taungoo, Myanmar
g
Department of Geology, Pathein University, Pathein, Myanmar
h
Department of Engineering Geology, Yangon Technological University, Myanmar
i
Department of Geology, Hinthada University, Hinthada, Myanmar
j
Department of Meteorology and Hydrology, Nay Pyi Taw, Myanmar
k
National Centre for Seismology, New Delhi, India

A R T I C LE I N FO A B S T R A C T

Keywords: We map the distribution of macroseismic intensities from the MW 6.9 Kani and the MW 6.8 Chauk intra-slab
Myanmar earthquakes 2016 earthquakes in 2016 in Myanmar using the 1998 European Macroseismic Scale (EMS-98) by interpreting data
Intra-slab earthquakes gathered from field surveys, community responses sent via social media to the Myanmar Earthquake Committee
Macroseismic intensity observations (MEC), and digital news reports. Our macroseismic maps for both events provide better spatial data coverage in
Strong motion analysis
Myanmar, India, and Bangladesh than community derived macroseismic maps (e.g., U.S. Geological Survey's
Local and regional ground motion behaviour
“Did You Feel It?”). In Myanmar, this was driven by improved telecommunication that has allowed social media
such as the Burmese language Facebook portal of the Myanmar Earthquake Committee (MEC) to reach into rural
areas from where reports of shaking effects from earthquakes have been previously unavailable. Our analysis of
both the macroseismic intensities and strong motion observations from India and Myanmar suggests the two
earthquakes had different source properties. The comparison of our intensity data with instrumental strong
motion records also suggests the peak ground motion-intensity relationship by Worden et al. (2012) generally
performs well for both earthquakes. In addition, ground motion behaviour within the Burma and Indian plates
can be related to different existing ground motion prediction equations (GMPEs) and intensity prediction
equations (IPEs) for subduction zones and for stable continental regions respectively. We therefore suggest these
effects will need to be considered in future regional seismic hazard models or Shake Maps for this region when
evaluating the impact of the future events.

1. Introduction et al., 2003; Benetatos and Kiratzi, 2004; Nagashima et al., 2012).
However, such instrumental data are sparse in Myanmar as efforts to
Over the past few decades, instrumental records of weak and strong improve and upgrade the country's instrumental capacities are still
ground motion from well-instrumented and seismically active regions underway (e.g. Hrin Nei Thiam et al., 2016). The very limited extent of
have yielded useful insights into the complexity of shaking during instrumental records, thus far, restricts the quantitative analysis of
earthquakes, its interdependence on rupture directivity, site conditions, earthquake ground motion behaviour in this part of Southeast Asia,
and building response (e.g. Erdik, 1987; Buchon and Barker, 1996; Wu unlike adjacent regions such as the Indian subcontinent (e.g. Hough

⁎
Corresponding author at: Department of Geosciences, National Taiwan University, No. 1, Sec. 4, Roosevelt Road, Taipei 106, Taiwan.
E-mail address: wangyu79@ntu.edu.tw (Y. Wang).

https://doi.org/10.1016/j.tecto.2019.04.016
Received 1 October 2018; Received in revised form 8 April 2019; Accepted 12 April 2019
Available online 26 April 2019
0040-1951/ © 2019 The Authors. Published by Elsevier B.V. This is an open access article under the CC BY-NC-ND license
(http://creativecommons.org/licenses/BY-NC-ND/4.0/).
Lin Thu Aung, et al. Tectonophysics 765 (2019) 146–160

et al., 2016; Sharma et al., 2016). It also limits the application of em- of generating great earthquakes (M > 8) every several hundred years
pirically determined Intensity Prediction Equations (IPEs) and Ground along its shallow north-east-dipping plate interface between ~18° N
Motion Prediction Equations (GMPEs) as these globally derived re- and ~24° N (e.g., Socquet et al., 2006; Wang et al., 2013; Wang et al.,
lationships have not been vetted for use in Myanmar with local ob- 2014; Steckler et al., 2016).
servations. Although background seismicity in the shallow part of the Arakan
In regions with similar constraints and limitations, macroseismic megathrust is sparse since the beginning of instrumental records, the
data from large modern earthquakes have been used as proxies to instrumental catalogue nevertheless shows abundant intermediate
analyze ground motion behaviour (e.g. Hough et al., 2002; Hough et al., depth (50 km - 150 km) seismicity within the subducted Indian slab
2016). However, the aging national telecommunication networks and (Storchak et al., 2013; Fig. 1). The focal mechanisms of these inter-
the internal political situation in Myanmar (The Asia Foundation, 3 mediate depth earthquakes display predominant N-S oriented P-axes,
April 2013; The World Bank, 2014; Open Democracy, 6 February 2014) subparallel to the strike of the down-going slab (Satyabala, 1998;
have been detrimental to the collection of macroseismic data for Hurukawa et al., 2012). North of 20°N, the depth of the intra-slab
earthquakes in the past several decades. In addition, macroseismic data seismicity is around 60 km - 90 km, with epicenters located along the
collected by well-calibrated online systems (e.g. Wald et al., 1999; eastern flanks of the Indo-Myanmar Range (IMR; also known as the
Bossu et al., 2016) have also yielded disproportionately low responses Indo-Burman Range), which represents the accretionary prism between
from this region, in part due to a lack of awareness of these systems the Burma plate and the Indian Plate. These earthquakes become even
coupled by poor internet coverage as has also been observed elsewhere deeper (> 90 km, and up to ~200 km) under the north-western portion
(Martin and Hough, 2015). Given these constraints, an alternative ap- of the Central Myanmar Basin (CMB). The CMB is a series of narrow,
proach to gather macroseismic data is essential. elongated, fore-arc and back-arc basins between the IMR accretionary
In the wake of major policy reforms implemented in Myanmar since prism and the Sagaing fault (e.g. Pivnik et al., 1998) (Fig. 1). During the
2011 (Mieno, 2013; Findlay et al., 2016), its mobile telecommunication past century, several significant earthquakes occurred within this sec-
network has rapidly expanded resulting in an increase in the number of tion of the subducted slab, causing varying grades of damage to set-
mobile subscribers, and in the exposure of a large proportion of the tlements in 1906, 1932, 1938, 1954, 1970, 1975, 1988, and 2016, as
population to social media via mobile broadband services. For example, plotted in Fig. 1 (Middlemiss, 1910; Gee, 1934; Gutenberg and Richter,
users of social networking portals in Myanmar grew from nearly 1- 1949; Tandon and Mukherjee, 1956; Tandon and Srivastava, 1974;
million active users in January 2014 to > 10-million active users in Rastogi and Singh, 1978; Mazumdar and Nageswaran, 1988; Gahalaut
2016, with > 80% of them accessing social media via their mobile et al., 2016). Intra-slab events were also possibly responsible for the
devices (Myanmar Times, 1 April 2016). Tapping this unique resource damage to the ancient city of Bagan in central Myanmar (formerly
for the first time in Myanmar, we elicited felt reports for both the MW Pagan; Fig. 1) in Common Era (CE) 324, 986, 1286, and 1290, accounts
6.9 Kani earthquake on 13 April 2016, and the MW 6.8 Chauk earth- of which are preserved in the written historical record (Shwe Gaing
quake on 24 August 2016 via the Burmese language Facebook portal of Thar, 1976; Nutalaya et al., 1985; Stadtner, 2011). The macroseismic
the Myanmar Earthquake Committee (MEC). The MEC was founded in effects of many of these aforementioned earthquakes are poorly cata-
1999 to facilitate research on seismic hazards and earthquake en- logued in Myanmar, possibly as a result of a lack of a political or an
gineering in Myanmar, and to increase the awareness of disaster miti- academic interest in rural areas, or because of the difficulty in col-
gation and preparedness within the country. We supplemented our field lecting macroseismic information from rural areas due to an under-
surveys conducted immediately after both earthquakes with these felt developed communication infrastructure in the mid to late 20th cen-
reports, as well as news reports from digital versions of conventional tury.
newspapers to map the macroseismic intensity distributions for both The Mw 6.9 Kani earthquake on 13 April 2016 and the Mw 6.8
earthquakes. We then compared our observations to available instru- Chauk earthquake on 24 August 2016 are the two most recent intra-slab
mental records of Peak Ground Acceleration (PGA) and Peak Ground events located beneath the western Central Myanmar Basin. Similar to
Velocity (PGV), and tested the validity of selected Ground Motion to other earlier intra-slab earthquakes, the Kani and the Chauk earth-
Intensity Conversion Equations (GMICEs) and GMPEs for Myanmar and quakes produced some damage in rural settlements near their epi-
the surrounding regions. centers. However, unlike previous events, these two recent earthquakes
offered a unique opportunity to collate macroseismic information in the
2. Seismotectonic background region to investigate ground motion and earthquake source properties.
These macroseismic observations were supplemented by instrumental
The seismotectonics of Myanmar and the surrounding region are records from the Myanmar National Seismic Network (MM), the net-
largely controlled by the on-going India-Eurasia collision (Tapponnier works of the National Center for Seismology (NCS) in India, and the
and Molnar, 1977, 1979; Hutchison, 1989; Le Dain et al., 1984). The Thailand Meteorological Department (TMD).
north-north-eastward motion of the Indian plate relative to the Eurasian
Plate creates two active plate boundaries along the eastern and the 3. Macroseismic observations
western margin of the Burma Plate, respectively (Fig. 1). To the east,
the plate boundary features the north-south striking, dextral Sagaing The 13 April 2016 and 24 August 2016 earthquakes in western
fault between the Sunda and the Burma plates; to the west, the Myanmar are referred to as the Kani and Chauk earthquakes, respec-
boundary between the Indian and the Burma Plate is characterized by tively in our study. To collate macroseismic information for both
oblique convergence along the Arakan megathrust, where the oceanic earthquakes in Myanmar, we combined the data gathered from post-
Indian Ocean lithosphere subducts obliquely beneath the Burma Plate earthquake field surveys in their epicentral regions with information
(Curray et al., 1979; Steckler et al., 2008; Zhang et al., 2013; Wang culled from digital media reports, and felt responses sent in via social
et al., 2014). Although previous seismological and tectonic studies media. The Facebook page of the Myanmar Earthquake Committee (see
suggest active subduction of the Indian Ocean lithosphere has ceased Data and resources) yielded a large number of felt and damage reports
along the western margin of the Burma Plate (Rao and Kumar, 1999; from Myanmar for both earthquakes within a few days of their occur-
Dasgupta et al., 2003; Satyabala, 2003; Nielsen et al., 2004), recent rence. Web analytics indicated as many as 10,000 visits to this platform
geodetic studies show an oblique motion of about 1 to 2 cm/yr across following the Chauk earthquake in terms of “Comments”, “Likes”,
the boundary of the Burma and Indian plates (Socquet et al., 2006; “Shares” or “Page Views”. A subset of these visitors provided their lo-
Maurin et al., 2010; Steckler et al., 2016). Most of this plate con- cations (i.e. the name of their village with the nearest town and district)
vergence is likely absorbed by the Arakan megathrust, which is capable and comments that described their experiences of the shaking. Some

147
Lin Thu Aung, et al. Tectonophysics 765 (2019) 146–160

Fig. 1. Seismotectonic map of Myanmar showing

(MW > 5) instrumentally-recorded earthquakes
(color dots) between January 1904 and December
2016 (Storchak et al., 2013); significant intra-slab
earthquakes (black stars; M > 6.5 with focal
depths > 60 km); distribution of seismic stations
(black triangles) and USGS's focal mechanisms of the
2016 earthquakes. Dashed line represents the ap-
proximate boundary between the Indo-Myanmar
Ranges and the Central Myanmar Basin or Central
Myanmar Belt (CMB). Arrow head indicates the di-
rection of convergence of the Indian and Burma
plates and red stars represent the instrumental epi-
centers of the Kani (MW 6.9) and Chauk (MW 6.8)
earthquakes. (For interpretation of the references to
color in this figure legend, the reader is referred to
the web version of this article.)

Facebook users also uploaded photographs of damage. By analyzing experienced such severe earthquakes in their life-time. Therefore,
these reports collected via the MEC's Facebook portal, we were able to during our field surveys, we supplemented these with our own ob-
assign intensity and associate locations within Myanmar for 728 user servations and descriptions of damage or effects (e.g. small objects fell
reports from 105 locations for the Kani earthquake and 575 user reports down, heavy materials shifted, etc.), and the types of buildings affected.
from 65 locations for the Chauk earthquake (Table S1). Outside All of these observations were interpreted using outlined diagnostics
Myanmar, many of our sources, especially those from the Indian sub- (see Grünthal, 1998) to convert to the corresponding macroseismic
continent, are derived from newspaper reports, a resource used fre- intensities using the 1998 European Macroseismic Scale (EMS-98)
quently to study historical earthquakes in this region for which in- (Grünthal, 1998) and following procedures utilized in previous studies
strumental data is sparse or lacking (e.g. Singh et al., 2013; Martin and (e.g. Martin and Kakar, 2012; Martin et al., 2015; Martin and Hough,
Hough, 2015). In total, we have 505 and 239 reports of intensity for the 2016). For those locations where available observations were in-
Kani and Chauk earthquakes respectively, from Myanmar, Bangladesh, sufficient to assign an EMS-98 intensity based on predetermined diag-
Bhutan, China, India, Nepal, and Thailand (Table S2 and S3). nostics (see Grünthal, 1998), we instead indicate that the earthquake
Our post-earthquake field surveys in the epicentral regions for both was felt locally (F) or caused damage (D) (see Table S2 and S3). Each of
events were conducted within a week of each of the earthquakes. our intensity assignments in Tables S2 and S3 were assigned a quality
During the field surveys, we interviewed local people to ascertain their factor (see Musson, 1998). We do this to account for the uncertainty of
experiences during the earthquake. We also recorded building and the intensity assignment, the resolution of geo-spatial co-ordinates, the
ground damage, and other co-seismic phenomena. In general, we found truthfulness of the available accounts or a combination of all or some of
most people exaggerated the level of shaking because they had never the above. The adaption of the EMS-98 scale allows us to maintain

148
Lin Thu Aung, et al. Tectonophysics 765 (2019) 146–160

Table 1
Building typology in the epicentral region in Myanmar including their Romanized transliterations (A Shin Thu Mana, 2014) and their descriptions in English.
Building types (in Burmese) Romanized Transliteration English Description

Pha-Yar Pagodas (stupa or Buddha's image or temples) with massive masonry brick structures with traditional organic mud
mortar
Oak-Taik Massive masonry brick structures traditional organic mud mortar
Concrete-Taik Reinforced/Unreinforced concrete (RC) structures
Oak-Nyut/ Taik-Khan-Eain Brick-nogging. Timber frame filled with clay brick wall
Thit-Thar-Eain Traditional timber house
War-Eain Traditional bamboo house

uniformity with previous macroseismic compilations in and around the fatalities in eastern India and many people were injured in India and
Indian subcontinent (e.g. Martin and Szeliga, 2010). The EMS-98 has Bangladesh (World Health Organization, 2016). Our intensity data for
been shown to be consistent, with very minor differences, to the the Kani earthquake is forthcoming from 505 locations (Table S2). Al-
Modified Mercalli Intensity (MMI) scale (Musson et al., 2010). As though its epicenter was in between the Mingin and Taze Townships in
pointed out by Debbarma et al. (2017), despite the misleading nature of the Kalay District in the Sagaing Region, the felt intensity only reached
the word “European” in its name, the EMS-98 scale provides greater 3 EMS in Mingin Township while an intensity of 5–6 EMS was observed
flexibility than the MMI scale since the EMS-98 scale can be adapted to in Kani, Monywa, and Kalaywa Townships (Fig. 3a, Table S2). Most of
local societal and building conditions. Such adaptations (see Martin and the townships that experienced strong shaking were to the south and
Hough, 2016) do not affect the robustness of the interpreted intensities west of the instrumental epicenter (Fig. 3a). Among the areas that ex-
when compared to EMS-98 intensities determined by field surveys (e.g. perienced intensities in the range of 5–6 EMS, the most significant
Adhikari et al., 2017), or the intensities from well-calibrated internet damage occurred in Kani Township (Table S2). Brick nogging (Type A-
based algorithms (Bossu et al., 2016). B) buildings at a school and a monastery were gently warped, and
Our post-earthquake field observations in Myanmar found that masonry panels sustained Grade 1 damage. According to our field
structures in the epicentral regions were a mix of masonry, concrete, survey, ~20% of the walls of residential and public buildings (e.g.
and timber frame structures. In general, Oak-Taik (Massive earth brick schools) in this village developed cracks. Interviews with local eye-
structures; Type A), Oak-Nyut/Taik-Khan-Eain (Brick-nogging or witnesses in Kani Township led us to conclude that the buildings and
timber frame filled with earth brick walls; Type A-B), some Concrete- pagodas that were damaged by a Mw 5.4 earthquake on 27 November
Taik (Unreinforced concrete (RC) structures; Type B-C), and Pha-Yar 2015 sustained further damage during the Kani earthquake. The Kani
(pagodas and temples with massive earth brick structures) were da- earthquake was widely felt over a large region in Myanmar, especially
maged in both earthquakes (Grünthal, 1998; also see Table 1 and Fig. within the CMB (Fig. 2a). Shaking intensities diminished rapidly to the
S1). In the epicentral regions, the most significant damage was sus- east based on social media reports. Reports of isolated parapet or wall
tained by Oak-Nyut/Taik-Khan-Eain (Brick-nogging structures) owing collapses were available from Chittagong and Feni in Bangladesh
to the inhomogeneity between their timber frames and earth brick infill (Dhaka Tribune, 14 April 2016), and Jowai, Imphal, Shillong, and Sil-
walls, i.e., the brick wall being of lower ductility than the surrounding char in India (Prerna Bharati, 14 April 2016; Sangai Express, 16 April
timber frame during the earthquake. In contrast, the Thit-Thar-Eain 2016; Shillong Times, 13 April 2016). Minor Grade 1 damage, such as
(traditional timber houses; Type C-D) and War-Eain (bamboo houses; cracks in a few buildings or the displacement of small, unstable objects,
Type C-D) which are the main structural types in the epicentral regions were reported from several locations in Bhutan, Bangladesh, and north-
resisted the strong shaking in both earthquakes. We compared the eastern India. Tremors from this earthquake were also felt in southern
building types, vulnerability classes, and the damage sustained at each Tibet, and in the Kathmandu valley in Nepal. Seismic seiches were
location during the field survey to assign EMS-98 intensity (Grünthal, observed in several lakes and water bodies in Myanmar, such as at
1998). Gyobinauk, Hakha, Kalaywa, Mandalay, Myaungmya, Rathedaung,
The towns of Kani and Chauk, proximal to the instrumented epi- Sagaing, and Ya Thi Thaung (Table S2). Cracks also allegedly formed in
centers of the April (also known as the Mawlaik earthquake) and August a levee near Gayan on the Brahmaputra River in upper Assam, India
earthquakes, respectively, reported the highest intensities. The epi- (Dainik Purvoday, 13 April 2016). The Kani earthquake was also felt by
center of the Kani earthquake lies approximately between Kani and a few occupants of multi-storied buildings in northern, eastern and
Mawlaik towns. However, higher intensities and damage were reported southern India (Table S2).
in Kani and its environs rather than the Mawlaik area. For both
earthquakes, our data combining the results of the field survey, the 5. MW 6.8 Chauk earthquake
experiences of eyewitnesses, and reports from newspapers provide
better and denser spatial coverage (Fig. 2; Fig. S2) than the community The Mw 6.8 Chauk earthquake was located ~240 km south of the
based macroseismic maps (e.g., USGS's DYFI map, see Data and re- Kani earthquake at a depth of 82 km (see Data and resources), in the
sources) reiterating the observation made by Martin and Hough (2015). CMB. This earthquake caused three casualties due to the collapse of a
Although such transnational maps are routinely prepared in near-time river bank at Yenanchaung and a roof collapse at Pakkoku in central
by, for example, the USGS's DYFI, as noted by Van Noten et al. (2017), Myanmar. In addition, a tourist was injured in Bagan, Myanmar, and
national agencies are still best placed to do this. The denser coverage of many people were injured in eastern India and Bangladesh while run-
our dataset addresses this by incorporating responses received by the ning outdoors in panic (Table S3). We were able to assign macroseismic
Burmese language Facebook page of the MEC that is more widely intensities at 239 locations for this earthquake (Fig. 2b, Table S3). The
known in Myanmar than, for example, the USGS's DYFI system. highest intensity of 6 EMS was experienced in multiple areas, including
Salin and Chauk Townships in Magway District, in the CMB where re-
4. MW 6.9 Kani earthquake inforced concrete (RC) buildings were slightly damaged (Table S3). This
earthquake also triggered the eruption of mud volcanoes at Minbu and
The MW 6.9 Kani earthquake was located at a depth of 136 km and Kyaukpyu in the CMB and on the western coast of Myanmar, respec-
produced different grades of damage in Myanmar, India, and tively. In the epicentral region, different degrees of building damage
Bangladesh (see Data and resources). This earthquake resulted in 9 were observed to old masonry buildings and pagodas during our post-

149
Lin Thu Aung, et al. Tectonophysics 765 (2019) 146–160

Fig. 2. Macroseismic maps (EMS-98) for the intra-slab earthquakes of 2016 in central Myanmar using data compiled in this study: (a) MW 6.9 Kani earthquake and (b)
MW 6.8 Chauk earthquake (see also Supplementary Materials). Red stars indicate the instrumental epicenters, and colored rectangles indicate the EMS-98 felt
intensities. Inset boxes represent location of epicentral regions in Fig. 3a and b. (For interpretation of the references to color in this figure legend, the reader is
referred to the web version of this article.)

earthquake survey (Fig. 3b). In the ancient city of Bagan, at least 397 visit. In the immediate vicinity, people lost their balance, loose objects
structures (Myanmar Times, 29 August 2016) including pagodas were fell from the shelves of a shop and diamond buds were dislodged from
damaged (Saw Htwe Zaw et al., 2017). At Taungdwingyi, ~103 km the tops of two pagodas, suggesting the macroseismic intensity ap-
southeast of the epicenter, the exterior brick wall of a century old, two- proached 6 EMS in this area. Unlike the Kani earthquake, damage from
storied monastery collapsed, exposing the internal wood frame floors the Chauk earthquake (5 EMS or higher) extended roughly 200 km from
and columns. This building was previously damaged and abandoned the instrumental epicenter, and as with the Kani earthquake, appears to
following the MW 6.6 Taungdwingyi earthquake on 21 September 2003 have extended further to the west compared to the east (Fig. 2b).
(Soe Thura Tun et al., 2003; Myanmar Earthquake Committee, 2003), Within this region there were reports of collapsed boundary walls such
and the ruins were in the process of being demolished at the time of our as at Sittwe, the largest city on the west coast of Myanmar. The shaking

150
Lin Thu Aung, et al. Tectonophysics 765 (2019) 146–160

was also strong enough to damage chinthe or leogryphs (protective time and place, their structures are mostly in the shape of an upturned
lions) at the entrance of the Shwe Mokhtaw Pagoda near Pakkoku funnel or bell, gilded in gold leaves, and composed of three main sec-
(Table S3). An isolated report of damage to the gateway of a monastery tions i.e. the spire, body, and base (Fig. 4a, c, and e). In the ancient city
was forthcoming from Mawlu in Sagaing division, ~643 km northeast of Bagan, the bell-shape is favored the most but hemispherical and
of the epicenter. Reports of shaking were received from a number of bulbous shapes also exist (Stadtner, 2005). Many of these pagodas in
locations to the south of 22° N latitude as far south as on the Ayeyar- Bagan were built around the 10th century, and have undergone nu-
wady delta. No accounts were received from Kayin, Mon, and Tha- merous structural modifications, suffered damage from natural phe-
nintharyi divisions in the south, or Kachin state in the north of the nomena and wars, and have been subject to haphazard repairs over the
country (Fig. 2b). However, in stark contrast to the Kani earthquake, course of their lifetimes (Stadtner, 2011). The extent and degree of
shaking from the Chauk earthquake was perceptible in high-rise modifications to these pagodas are often difficult to ascertain due to the
buildings in Bangkok, Thailand, ~1000 km away from the epicenter long and incomplete records of these structures.
(Bangkok Post, 24 August 2016) particularly in the Asoke, Sam Yan, We observed various grades of damage to pagodas during our field
Sukhumvit, and Ratchadapisek areas of the city. In the Indian sub- surveys after both earthquakes (Table S4). The most common damage
continent to the west and north of the epicenter, shaking was distinctly to pagodas was to the spire, i.e. the dislodging of the diamond bud and
felt within 1200 km of the epicenter despite disruption from monsoonal cracking at the base of the umbrella (Fig. 4f). As per the guidelines of
flooding. These far-field locations are predominantly in the foreland the EMS-98 (Grünthal, 1998), we do not assign intensities at locations
basin and the delta associated with the Ganga and Brahmaputra Rivers. where the only recorded damage was to pagodas, but we flag the da-
Within the Bengal basin in Bangladesh, shaking was strong enough to mage (D) instead (Table S2 and S3). However, separately, we classify
make people run outdoors in panic, resulting in numerous injuries. the earthquake damage to these structures based on the following in-
However, only minor building damage was recorded from one location dicators: (i) “collapse”: the spine collapsed, or heavy damage to the
in the Sunderbans in southern Bangladesh (Table S3): at Lord Hardinge upper part of temple, the body, or the base of stupa (Fig. 4b); (ii)
in Bhola district a school building developed cracks (Jagonews24, 25 “partial-collapse”: breaking and fall of the banana bud, the royal lotus,
August 2016). Seismic seiches were observed at Gopalganj and Pabna in and the protuberant coil (Fig. 4d); and (iii) “light damage”: the falling
Bangladesh (Table S3). The Chauk earthquake was also felt by a few of the diamond bud, the tilting of its umbrella and iron vane, or cracks
occupants of multi-storied buildings in northern and southeastern India. in the banana bud (Fig. 4f). We gathered 21 accounts of pagoda damage
for the Kani earthquake and 33 for the Chauk earthquake based on
social media responses and our post-earthquake field surveys (Table
6. Observations from damaged pagodas
S4). The locations of these pagodas, and the distribution of macro-
seismic intensities near these damaged structures are plotted in Fig. 3.
Pagodas that dot the landscape in Myanmar could potentially be a
Although these data are representative of a small population of pagodas
good proxy of the ground motions from historical and modern earth-
in the epicentral regions of both earthquakes, our observations show a
quakes. Although the style of the pagodas have changed with respect to

Fig. 3. Distributions of felt intensity and damaged pagodas in the epicentral regions observed following: (a) the MW 6.9 Kani and (b) the MW 6.8 Chauk earthquakes
(see Table S3). Colored squares represent the EMS-98 intensities at different localities; color triangles indicate the observed damaged pagodas; and red stars represent
(USGS) instrumental epicenters. Damage degree versus hypocentral distance for both events show a decrease in damage to pagodas with distance. Location of Fig. 3a
and b indicated by inset boxes in Fig. 2. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

151
Lin Thu Aung, et al. Tectonophysics 765 (2019) 146–160

Fig. 4. Structural typology of observed damaged pagodas following the 2016's earthquakes in Myanmar. The basic components of pagodas in general are base, body,
and spire and these are indicated through diagrammatic sketches and photographs (a-f), including examples of any damage. (a) Structural form of multi-story temple;
(b) collapsed spire of the Sulamani (multi-story) temple (Loc. 21° 09′ 54.05″ N and 94° 52′ 00.56″ E) due to the Chauk earthquake; (c) structural form of a common
solid stupa in Myanmar; (d) a partially-collapsed solid stupa at Kansiba village (Loc. 22° 32′ 36.00″ N and 94° 52′ 38.00″ E) due to the Kani earthquake; (e) structural
form of high-base stupa; and (f) light damage to the Shin Pin Paung Daw Oo (high-base) stupa (Loc. 21° 11′ 01.00″ N and 94° 55′ 13.00″ E) near Sin Phyu Kyun due to
the Chauk earthquake. Original photo in (b) is extracted from The Voice Daily Journal and other photos were taken during our field survey.

152
Lin Thu Aung, et al. Tectonophysics 765 (2019) 146–160

Table 2
PGA and PGV values for the 13 April 2016 Kani earthquake.
# Station name Code Net Lat. Long. R PGA (cm s−2) PGV (cm s−1)

N E km Z N-S E-W Z N-S E-W

a
1 Tamu TMU MM 24.230 94.300 140 14.53 40.27 25.57 1.24 3.87 3.50
2 Hakha HKA MMa 22.642 93.600 140 21.09 56.69 42.95 2.53 4.41 5.52
3 Mandalay MDY MMa 22.016 96.112 175 1.41 1.32 1.03 0.18 0.44 0.24
4 Nyaung-U NAU MMa 21.206 94.916 210 16.48 9.81 11.32 1.01 1.05 0.70
5 Silchar SIL NCSb 24.781 92.803 280 26.9 79.2 75.7 2.05 5.52 5.24
6 Kohima KOHI NCSb 25.720 94.108 300 33.1 67.6 56.3 4.47 9.46 8.86
7 Belonia BELO NCSb 23.248 91.447 350 13.2 55.3 58.2 0.55 3.27 3.29
8 Mokokchung MOKO NCSb 26.321 94.516 360 35.3 75.1 86.2 2.58 6.30 8.85
9 Agartala AGT NCSb 23.889 91.246 380 11.8 28.7 43.6 0.73 2.31 3.56
10 Nay Pyi Taw NPT MMa 19.779 96.138 390 1.57 1.16 1.24 0.21 0.26 0.27
11 Jorhat JORH NCSb 26.743 94.251 410 40.6 104 129 2.72 5.86 4.98
12 Shillong SHL NCSb 25.567 91.856 410 19.1 76.1 41.7 0.87 1.69 1.55
13 Tezpur TEZ NCSb 26.617 92.800 445 52.1 152 123 2.05 4.74 4.91
14 Itanagar ITA NCSb 27.144 93.722 465 72.4 128 142 6.69 6.96 13.8
15 Guwahati GUWA NCSb 26.193 91.691 470 18.4 44.8 39.6 0.77 1.57 1.11
16 Lekhapani LKP NCSb 27.333 95.846 480 4.89 15.3 18.6 0.39 1.41 0.94
17 Dibrugarh DIBR NCSb 27.468 94.911 485 18.8 39.2 41.1 0.88 2.63 2.80
18 Kyaing Tong KTN MMa 21.286 99.590 525 0.36 0.69 0.64 0.12 0.22 0.26
19 Tura TURA NCSb 25.517 90.224 540 80.9 74.5 112 2.23 3.13 4.31
20 Dhubri DHUB NCSb 26.020 89.995 590 47.2 50.3 14.3 1.88 1.59 0.09
21 Yangon YGN MMa 16.865 96.153 705 4.34 2.98 2.79 0.32 0.16 0.14
22 Shiliguri SLGI NCSb 26.700 88.417 763 5.31 24.3 26.1 0.29 1.12 1.12
23 Gangtok GTK NCSb 27.319 88.601 785 2.83 4.63 6.76 0.29 0.37 0.41
24 Valmikinagar VAL NCSb 27.317 83.867 1200 0.529 1.44 1.97 0.05 0.14 0.17

a
Myanmar National Seismic Network (MM), Myanmar.
b
National Centre for Seismology (NCS), India.

Table 3
PGA and PGV values for the 26 August 2016 Chauk earthquake.
# Station name Code Net Lat. Long. R PGA (cm s-2) PGV (cm s-1)

N E km Z N-S E-W Z N-S E-W

a
1 Nyaung-U NAU MM 21.206 94.916 48 116.8 82.31 86.21 3.41 5.61 7.27
2 Mandalay MDY MMa 22.016 96.112 200 2.27 2.81 3.25 0.33 0.22 0.25
3 Nya Pyi Taw NPT MMa 19.779 96.138 210 21.08 32.41 39.11 1.02 2.39 4.18
4 Hakha HKA MMa 22.642 93.600 215 2.05 3.63 3.18 0.34 0.57 0.44
5 Tamu TMU MMa 24.230 94.300 370 0.54 0.92 1.15 0.06 0.09 0.11
6 Belonia BELO NCSb 23.248 91.447 415 8.05 18.9 22.2 0.33 0.97 0.82
7 Imphal IMPH NCSb 24.831 93.946 440 0.33 0.50 0.41 0.04 0.04 0.06
8 Agartala AGAR NCSb 23.889 91.246 475 1.36 2.37 3.65 0.17 0.31 0.40
9 Yangon YGN MMa 16.865 96.153 480 1.74 2.70 3.29 0.20 0.43 0.50
10 Chiang Mai CHTO TMDc 18.810 98.940 515 4.9 – – – – –
11 Kyaing Tong KTN MMa 21.286 99.590 520 1.04 1.71 1.47 0.32 0.45 0.43
12 Kohima KOHI NCSb 25.720 94.108 535 0.78 0.91 1.52 0.04 0.09 0.10
13 Shillong SHL NCSb 25.567 91.856 585 3.28 23.3 14.1 0.07 0.27 0.19
14 Mokochong MOKO NCSb 26.321 94.516 600 0.68 1.06 0.10 0.03 0.06 0.05
15 Jorhat JORH NCSb 26.743 94.251 650 1.22 2.65 2.88 0.05 0.12 0.14
16 Guwahati GUWA NCSb 26.193 91.691 655 2.68 8.43 7.38 0.06 0.07 0.09
17 Tezpur TEZP NCSb 26.617 92.800 655 3.96 7.10 7.43 0.09 0.02 0.02
18 Tura TURA NCSb 25.517 90.224 675 7.01 8.53 8.47 0.17 0.22 0.22
19 Itanagar ITA NCSb 27.144 93.722 695 3.18 6.87 4.38 0.10 0.20 0.16
20 Lekhapani LEKH NCSb 27.333 95.846 725 0.30 0.53 0.58 0.03 0.03 0.03
21 Dibrugarh DIBR NCSb 27.468 94.911 730 0.42 1.23 1.07 0.04 0.08 0.09
22 Dhubri DUBH NCSb 26.020 89.995 735 1.22 4.91 4.68 0.07 0.17 0.15
23 Ziro ZIRO NCSb 27.594 93.850 730 2.13 5.09 4.72 0.08 0.33 0.25
24 Tawang TAWA NCSb 27.594 91.867 790 1.87 3.70 3.02 0.08 0.23 0.15
25 Siliguri SLGI NCSb 26.700 88.417 895 2.10 6.12 6.99 0.06 0.30 0.25
26 Nongkai NONG TMDc 18.060 103.150 955 1.96 – – – – –
27 Bangkok BKK TMDc 13.670 100.060 995 1.47 – – – – –
28 Nakonayok NAYO TMDc 14.320 101.320 1025 5.88 – – – – –
29 Prachuab PRAC TMDc 12.470 99.790 1090 0.78 – – – – –
30 Valmikinagar VAL NCSb 27.317 83.867 1300 0.362 1.06 0.94 0.03 0.06 0.07

a
Myanmar National Seismic Network (MM), Myanmar.
b
National Centre for Seismology (NCS), India.
c
Thailand Meteorological Department (TMD), Thailand.

153
Lin Thu Aung, et al. Tectonophysics 765 (2019) 146–160

Fig. 5. Highest peak ground motions for the 2016 intra-slab earthquakes recorded by the Myanmar (MM), Thai (DMD), and Indian (NCS) seismic networks. PGA
observations for the (a) MW 6.9 Kani and (b) MW 6.8 Chauk earthquakes; PGV observations for the (c) MW 6.9 Kani and (d) MW 6.8 Chauk earthquakes. Focal
mechanisms and epicenters (red stars) for both earthquakes are from the USGS. Distribution of peak ground motions shows the strong NNW directivity effects of the
Kani earthquake and weak SSW directivity effects of the Chauk earthquake. (For interpretation of the references to color in this figure legend, the reader is referred to
the web version of this article.)

distinct decrease in the grade of damage with an increase in epicentral network began operation in 2017 (Wang et al., 2018). For the Kani
distance for both earthquakes (Fig. 3). earthquake, the Hakha (HKA) and Tamu (TMU) stations in Myanmar
are the closest stations (~140 km) to the west and north-west of the
epicenter, respectively (Fig. 1). These two stations show similar hor-
7. Recorded ground motions and earthquake source properties izontal peak ground motions, with the higher PGA value of 0.06 g
(56.69 cm s2) recorded at HKA (Table 2). In stark contrast, recorded
We present peak ground velocity (PGV) and peak ground accelera- ground motions were much larger to the northwest of the epicenter in
tion (PGA) values at 24 sites for the Kani earthquake (Table 2), and at India with the largest PGA of 0.16 g (152 cm s2) at Tezpur (TEZ) station
30 sites for the Chauk earthquake (Table 3) recorded by strong motion even though it is > 400 km away from the epicenter (Table 2). Inter-
and broadband stations in Myanmar, India, and Thailand (Fig. 5). Al- estingly, the farthest (~1200 km) Indian station at Valmikinagar (VAL)
though sparse at the time of both the Kani and Chauk earthquakes, the recorded a PGA value of 0.002 g (1.97 cm s2), which is higher than
recently upgraded National Myanmar Seismic (MM) network (Hrin Nei those (< 0.002 g) recorded at Mandalay (MDY), Nay Pyi Taw (NPT),
Thiam et al., 2016) provided good instrumental observations, particu- and Kyaing Tong (KTN), at distances of ~175 km, ~390 km, and
larly strong motion records. We do not have data from the seismic ~525 km, respectively in Myanmar. For the Chauk earthquake, the
network maintained by the Earth Observatory of Singapore as this

154
Lin Thu Aung, et al. Tectonophysics 765 (2019) 146–160

Fig. 6. Focal mechanisms and wave form analysis for selected strong motion stations for the 2016 intra-slab earthquakes. P-wave acceleration spectra amplitude at
(a) the TEZ station and (b) the YGN station for the Kani (black) and Chauk (red) events. The azimuthal projection of the two stations on the two focal mechanisms is
indicated by the arrows and dots on the beach balls (Kani in black and Chauk in red). For the Kani earthquake, three component waveforms (c) from the YGN and (d)
the TEZ stations are shown with the peak velocity appearing in the upper panel and peak accelerations appearing in the lower panels. The peak recorded PGV or PGA
values are printed above each individual waveform. Locations of TEZ and YGN stations are shown in Fig. 5. (For interpretation of the references to color in this figure
legend, the reader is referred to the web version of this article.)

highest PGA of ~0.1 g (86.21 cm s2) was recorded at the Nyaung-U same station differ by about 200 km, the associated geometry spreading
(NAU) station in Myanmar, ~48 km from the epicenter. Unlike the high difference (< 30%) for the two earthquakes can be ignored compared
PGA/PGV recorded in India for the Kani earthquake, the peak ground to their actual amplitude differences. As displayed in Fig. 6a and d, the
motions for the Chauk earthquake (Table 3) in north-eastern India were P-wave spectra of the two events at TEZ station show a dominant fre-
much smaller. For instance, the Tezpur (TEZ) station only recorded a quency at 5-10 Hz (corresponding to resonance frequency of low-rise
PGA of 0.007 g (7.43 cm s2) for the Chauk event, which is > 20 times buildings), but the amplitude of the Kani earthquake is about 20 times
smaller than that recorded for the Kani event (0.16 g; 152 cm s2). larger than the Chauk earthquake. This can be partly explained as a
In general, the recorded PGA/PGV values for the Kani earthquake consequence of the difference in sampling the radiation pattern from
are much larger than the Chauk earthquake (Fig. 5), despite the mag- the source. The dots on the beach balls in Fig. 6a represent the pro-
nitudes and depths of the two events being quite similar. To better jection of the TEZ station, which show that the TEZ station is close to
understand the PGA/PGV patterns and earthquake source properties, the P-wave nodal plane on the focal mechanism of the Chauk earth-
we selected two representative strong motion and broadband stations to quake but not for the Kani earthquake. The sampling of the P-wave
analyze their waveforms. Fig. 6 presents the records for TEZ and YGN radiation pattern is less different for the YGN station (Fig. 6b), but yet
stations, which are located about 400 km -700 km away from the two the amplitude of the Kani earthquake is systematically larger than the
events. Although the epicentral distances of the two earthquakes at the Chauk earthquake, despite the Chauk earthquake being located 200 km

155
Lin Thu Aung, et al. Tectonophysics 765 (2019) 146–160

closer to the station. The observations at the YGN station indicate that first converted the instrumental PGV and PGA values to macroseismic
the Kani earthquake radiated more high frequency (5–10 Hz) energy intensity (EMSPGV and EMSPGA) using both aforementioned GMICEs.
than the Chauk earthquake, suggesting a higher stress drop for the Kani We then compared these intensities to the observed macroseismic in-
earthquake. tensity data (EMSOBS) in the immediate vicinity (≤ 10 km) of these
The P-wave radiation samplings at the TEZ and YGN stations in stations. Fig. 7 shows the comparison between the EMSPGA and EMSPGV
India and Myanmar respectively, for the Kani earthquake have similar respectively with their paired EMSOBS for both earthquakes using the
strengths (red dots on black beach balls in Fig. 6a;b). However, the W12 and the AK07 relationships. In both figures, EMSPGA and EMSPGV
recorded waveform amplitude at TEZ station is much larger than the derived from the W12 model show a better fit than the AK07 model in
YGN station (Fig. 6c;d, e.g. 52.1 cm s2 vs 4.34 cm s2). This azimuthally the region of Myanmar and north-eastern India. The AK07 model, in
dependent pattern of PGA/PGV (Fig. 5) strongly suggests that the Kani general, predicts higher EMSPGA and EMSPGV than our EMSOBS. On the
earthquake had a northward rupture directivity. Site effects should not other hand, with the W12 model, the EMSPGA and EMSPGV values are in
have played a significant role in shaping the PGA/PGV patterns for the general agreement with our EMSOBS, especially for points within
Kani earthquake, as we do not observe such a pattern (i.e. larger PGA/ Myanmar. When we compare EMSPGV and EMSPGA using the W12
PGV in India) for the Chauk earthquake. In short, the intensity and model, our data suggest the mean difference between EMSOBS and
waveform data can be roughly explained by a larger stress drop and EMSPGA is −0.91 (Fig. 7b).
strong rupture directivity of the Kani earthquake. Further geophysical On closer examination of these data from the Kani and Chauk
investigation is needed to refine the details of the source processes of earthquakes individually, we find this −0.91 residual appears to be
the two earthquakes. largely influenced by instrumental observations in northeastern India
during the Kani earthquake. In one extreme case (TEZ, India; Table 2),
8. Discussion the difference between the EMSPGA and the nearby EMSOBS is about 4
(Fig. 7b). This station is > 445 km away from the hypocentre. Using the
8.1. Relationship between felt intensity and pagoda damage W12 relationship, the EMSPGV at the same location is ~4.7 while the
EMSOBS value is lower (~3 EMS). The same effect is also observed at the
Although the damage and reconstruction history of pagodas has Itanagar station (ITA; Table 2), where the recorded PGA is 142 cm s2
been used as the earthquake chronology in many cities of Southeast (EMSPGA = 7) and PGV is 13.8 cm s−1 (EMSPGV = 6). The observed
Asia (e.g. Thawbita, 1976; Win Swe, 2006; Saw Htwe Saw Htwe Zaw, intensity at this location, however, only approached 3 EMS (Table S2).
2006; Maung Thein et al., 2009; Wang et al., 2011; Yuan and Li, 2013; In contrast, the PGA recorded during the Chauk earthquake shows
Pananont et al., 2017), using pagoda damage as a proxy for earthquake lower residuals than those from the Kani event (Fig. 7b). In a previous
ground motions has not been tested thoroughly. Part of the reason is section, we have also shown that the Kani earthquake radiated more
that the varying architectural and structural characteristics of these high frequency energy than the Chauk earthquake. This coupled with a
pagodas (Fig. 4a, c, and e), including their heights and states of health, suggested northward rupture directivity produced stronger ground
makes it complicated to use observed damage to these structures as a motions in north-east India. Ground motions can vary over short dis-
uniform diagnostic. The varying frequency content and duration of tances, for example in connection with local geology. Given that we do
earthquake ground motions further complicates the process. Despite not know the exact locations of the felt report used to determine
these complex conditions mentioned above, our preliminary compar- macroseismic intensity with respect to the location of the instrument,
ison of the spatial distribution of different grades of damage to pagodas although we are confident of its location more generally, we expect
with the observed macroseismic intensity in their vicinity there will be differences in the observed (EMSOBS) and converted
(i.e. ≤ 10 km) show a rough relationship between pagoda damage and (EMSPGV, EMSPGA) intensities as we note above.
the observed intensity (Fig. 3, Table S4). Most of the pagodas classified Overall, our analysis suggests the W12 model is more suitable than
by us as “collapsed” were concentrated in those regions where in- AK07 to reproduce the observed felt intensities from instrumental re-
tensities reached and exceeded ~5 EMS. In contrast, the spatial dis- cords during these two intra-slab events, especially in Myanmar.
tribution of pagodas that sustained “partial-collapse” and “light damage” Previous analysis of the 2015 MW 7.8 Gorkha earthquake in Nepal also
appears to correspond to ~4 to ~5 EMS. The number of pagodas that showed a similar result wherein EMSPGA converted from instrumental
sustained “light damage” decreased once the observed intensity was strong motion observations in northern India was in good agreement
close to 6 EMS. However, there were also remained numerous instances with EMSOBS (Hough et al., 2016). Although the source characteristics
wherein pagodas with a “partial-collapse” damage grade appeared when of these earthquakes are very different i.e. interface rupturing versus
the observed intensity reached 6 EMS (Table S4). Our data on pagodas intra-slab ruptures, both examples suggest the Worden et al. (2012)
damaged in these two earthquakes are far from complete since we do relationship is a suitable empirical model to convert between the
not conduct a comprehensive survey of all pagodas in the epicenter macroseismic intensity and peak ground motion, especially from PGV in
areas. Our preliminary comparison, however, does suggest a relation- the case of these two events.
ship between the grade of pagoda damage and the observed macro-
seismic intensity, which is clearly evident in the damage decay plots 8.3. Ground motion attenuation behaviour
(Fig. 3). The same damage-intensity relationship for pagodas may not
apply to other types of earthquakes such as shallow crustal events along We test different GMPEs and IPEs to look for an appropriate em-
the strike-slip Sagaing fault (Fig. 1). Crustal events may generate pirical ground motion equation that can be applied to these intra-slab
ground motions with different frequency contents compared to intra- earthquakes. We tested five IPE's: two for active shallow crustal regions
slab events. A more detailed analysis of pagoda damage patterns from (Atkinson and Wald, 2007 [AW07Cal]; Atkinson et al., 2014
other recent earthquakes is needed before applying this relationship to [AWW14WUS]), two for stable continental regions (Atkinson and Wald,
evaluate the ground motion and intensity of historical earthquakes in 2007 [AW07EUS]; Atkinson et al., 2014 [AWW14EUS]), and the India
Myanmar. model (Szeliga et al., 2010 [India]). We also test an additional three
GMPEs for subduction zone intra-slab earthquakes, including Youngs
8.2. Relationship between felt intensity and peak ground motion et al., 1997 (Y97), Atkinson and Boore, 2003 (AB03), and Zhao et al.,
2006 (Z06). The PGA computed from the AB03, Z06 and Y97 models is
We use our macroseismic intensity observations and the recorded then converted to intensity using the previously tested W12 model
peak ground motion values to test two widely used GMICE's models i.e., (Worden et al., 2012). We applied the global Vs30 model (Allen and
Atkinson and Kaka, 2007 (AK07) and Worden et al., 2012 (W12). We Wald, 2009) to estimate the site effects from shallow soil layers, if the

156
Lin Thu Aung, et al. Tectonophysics 765 (2019) 146–160

Fig. 7. Comparison between observed intensity (this study) and predicted intensity for the Kani and Chauk earthquakes using globally used IPEs: (a) converted from
PGV and (b) from PGA. We use red diamonds and green squares to indicate intensity converted using the W12 model (Worden et al., 2012). For the AK07 model
(Atkinson and Kaka, 2007) we use an un-inverted and an inverted grey triangle to indicate ground motion values predicted using our intensity dataset. (For
interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

Fig. 8. Intensity versus distance decay diagram with GMPEs

and IPEs of different tectonic settings (a) for the Kani earth-
quake and (b) for the Chauk earthquake. Red dots represent
macroseismic intensity data west of the Arakan megathrust,
i.e. on the Indian plate, and blue dots represent intensity data
from Myanmar. Solid stars represent strong motion data from
the Myanmar National Seismic (MM) network (see Tables 2
and 3) converted to intensity using Worden et al. (2012).
Dashed line on the inset map represents the Arakan Mega-
thrust and colored lines represent IPEs and GMPEs discussed
in the text along with issues related to directivity. The lower
bound of the three colored ribbons show the intensity pre-
dicted by GMPEs after removing the −0.91 mean residual
between the EMSPGV and EMSOBS estimated in Fig. 7. See
main text for details. (For interpretation of the references to
color in this figure legend, the reader is referred to the web
version of this article.)

157
Lin Thu Aung, et al. Tectonophysics 765 (2019) 146–160

selected GMPE model also incorporated Vs30 to estimate the site ef- been observed in previous intermediate depth earthquakes in the Indian
fects. Our previous analysis has already shown that the EMSPGA, in subcontinent (e.g. Martin and Kakar, 2012; Gahalaut et al., 2016). Al-
certain situations, tends to be overestimated. We adopt the W12 PGA- though this result is not unusual, this is the first time that we observe
intensity model (Worden et al., 2012) as most global GMPEs estimate such a distinct pattern in the Myanmar-India-Bangladesh region using a
the peak accelerations (including PGA) instead of PGV. As a result, a combination of a dense distribution of macroseismic intensities and
comparison of the Worden et al. (2012) model to three subduction zone strong motion recordings.
intra-slab related GMPEs may also result in an overestimation of the If the hypocenter of future similar sized intra-slab events were closer
predicted intensity when compared to both the ground observations to Bangladesh and north-eastern India, or if the magnitude of future
and the intensity predictions from the other five IPEs. intra-slab events is much greater (regardless of focal depth), such ani-
Fig. 8 shows the observed macroseismic intensity versus hypocen- sotropic attenuation behaviour could significantly impact rapid earth-
tral distance for both earthquakes. The large scatter of our observations quake damage assessments i.e. over-estimate damage in Myanmar and
is likely caused by the combination of different tectonic settings with under-estimate the damage in adjacent parts of Bangladesh and India.
different ground motion attenuation characteristics in Myanmar and We observed such an effect in the 2016 Kani earthquake, where
India, the nature of the rupture processes of both earthquakes, and site building damage was reported in both India and Bangladesh at loca-
effects within the sediment basins (e.g. Bengal Basin and Central tions at large distances from the epicenter. The change in the intensity
Myanmar Basin). None of the IPEs nor GMPEs we selected could explain attenuation behaviour across the Arakan megathrust therefore may
our observations since the observed intensities displayed significant require the incorporation of additional correction factors in the ground
scatter with distance. While most of these models either over- or under- motion prediction models, to produce a more realistic and reliable
estimate the observed intensity, the IPE derived for the Indian sub- earthquake Shake Map for example, or to produce better seismic hazard
continent (Szeliga et al., 2010, green line in Fig. 8) appears to provide a maps across the Indian-Burma plate boundary.
median prediction for both earthquake events with an uncertainty of
intensity, approximately ± 2 levels of intensity on the EMS-98 scale. 9. Conclusions
To test whether the different tectonic settings between Myanmar
and India affect the macroseismic intensity distribution, we divide our Using data gathered from social media responses, post-earthquake
dataset of observations from both the Kani and Chauk earthquakes field investigations, and news reports available via the internet, we
along the Indian-Burma Plate Boundary (e.g. Arakan megathrust; generated macroseismic intensity maps across the Burma and Indian
Fig. 1) into two sub-datasets for the Indian Plate and the Burma Plate, Plate boundary for modern intermediate depth earthquakes in this re-
respectively. Unmistakable differences are evident in our plot of in- gion. Our macroseismic maps provide better spatial coverage than
tensity as a function of distance (Fig. 8) between observations on the community-based macroseismic maps (e.g. USGS DYFI map) due to the
Burma plate with those on the Indian plate. Notably, the observed in- better reach of social media in Myanmar, and are largely comparable
tensities on the Indian plate are distinctly higher at large hypocentral with previous work in the Indian subcontinent (Martin and Hough
distances than those at similar distance on the Burma plate, which is (2015) and (2016)).
indicative that they are controlled by different attenuation behaviour in Despite the multiple factors affecting a pagoda's structural response
the Indian and the Burma plates respectively. The IPE AW07EUS gen- to earthquake ground motions, our data from these two earthquakes
erally matches the trend of far-field observations from Bangladesh and suggests a weak causal relationship between pagoda damage grades and
India, suggesting the crustal attenuation behaviour of the eastern In- the intensity in the immediate area (Fig. 3). However, given our small
dian plate shares similar characteristics to central and eastern North sampling size and that shallower earthquakes elsewhere in the region
America, a view also shared by previous studies (e.g. Johnston, 1996). would produce different frequency ground motions that would cause
On the other hand, the Z06 and AB03 models for intra-slab earthquakes different grades of damage. Further work is required to be able to sa-
agree well with our observations within the Burma Plate, even though tisfactorily incorporate damage to pagodas as a macroseismic intensity
we only used EMSPGA, not the EMSPGV, to estimate intensity. The use of diagnostic.
EMSPGA from GMPE models is the likely reason for the overestimation Our macroseismic observations display strikingly different spatial
in the macroseismic intensity by 1 unit on the EMS-98 scale (Fig. 7), patterns to the east and west of their epicentral regions. This can be
and causes a small upward shift of our intensity predictions on the interpreted as evidence for different ground motion characteristics ex-
Burma Plate (Fig. 6). If we adjust our three GMPE prediction models by hibited by the Burma and Indian plates. Within the Burma Plate, based
removing the 0.91 overestimation from the W12 relationship (Fig. 5b), on our observations the ground motion prediction equations for sub-
both Z06 and AB03 models remain the best models that explain the duction zone intra-slab earthquakes, i.e. Zhao et al. (2006) is most
EMS-98 intensity distribution within the Burma Plate. Their predicted appropriate. On the other hand, our observation suggests the best IPE
intensity also shows good agreement to the intensity from the AW07Cal for the eastern Indian Plate region is the intra-plate model for the
IPE model that is commonly used for shallow earthquakes in the active Central and Eastern U.S. (Atkinson and Wald, 2007).
crustal regions. The good agreement between our observations on the The distribution of EMS-98 intensities from both events also sug-
Burma Plate, and the predicted intensity from both Z06 and AB03 gests significant influence from local site effects in large sedimentary
model suggests the Burma plate, at least in the CMB area, displays si- basins, such as the Ayeyarwady Embayment, the southern part of
milar intensity attenuation characteristics for intra-slab earthquakes as Central Myanmar Basin, and the Bengal basin in Bangladesh. Moreover,
compared to other subduction zones around the world (Atkinson and waveform records indicate a strong north-northwestward rupture for
Boore, 2003; Zhao et al., 2006). the Kani earthquake and a larger stress drop, which explain larger PGA/
It is not surprising that the Indian Plate and the Burma Plate share PGV and macroseismic intensities in northeastern India. Site effects and
different attenuation characteristics, since the majority of the Indian earthquake source properties (such as stress drop and rupture direc-
subcontinent is a stable continental shield, similar to the tectonic set- tivity) should be considered in the regional seismic hazard models in
ting under the central and eastern United States (e.g. Johnston, 1996; the future.
Pubellier et al., 2008). The Burma plate, especially within the Central
Myanmar Basin, is mainly underlain by a Mesozoic basement mantled 10. Data and resources
by deep Cenozoic sediments (e.g. Pivnik et al., 1998), where seismic
waves attenuate much faster than in the Pre-Cambrian continental The preliminary seismological information, including the focal
shield. The resultant anisotropy in the spatial distribution of macro- mechanisms, were obtained from the USGS web page for the Kani
seismic effects and instrumentally recorded ground motions have also earthquake (https://earthquake.usgs.gov/earthquakes/eventpage/

158
Lin Thu Aung, et al. Tectonophysics 765 (2019) 146–160

us20005hqz/executive; last accessed 15 September 2018) and the (2), 497–510 (doi:10:1785/0120060154).
Chauk earthquake (https://earthquake.usgs.gov/earthquakes/ Atkinson, Gail.M., Wald, D.J., 2007. “Did you feel it?” intensity data: a surprisingly good
measure of earthquake ground motion. Seismol. Res. Lett. 78, 362–368.
eventpage/us10006gbf/executive; last accessed 15 September 2018). Atkinson, Gail.M., Worden, C.B., Wald, D.J., 2014. Intensity prediction equations for
Our intensity assignments are based on responses received via the social North America. Bull. Seismol. Soc. Am. 104 (6), 3084–3093.
media account of the Myanmar Earthquake Committee (MEC) of Benetatos, C., Kiratzi, A., 2004. Stochastic strong ground motion simulation of inter-
mediate depth earthquakes: the cases of the 30 May 1990 Vrancea (Romania) and of
Myanmar Engineering Society (i.e., 13 April (https://www.facebook. the 22 January 2002 Karpathos Island (Greece) earthquakes. Soil Dyn. Earthq. Eng.
com/MyanmarEarthquakeCommittee/photos/pcb.595316197292099/ 24 (1), 1–9 (ISSN 0267-7261).
595311890625863/?type=3&theater; last accessed 15 September Bossu, R., Landes, M., Roussel, F., Steed, R., Mazet-Roux, G., Martin, S.S., Hough, S.E.,
2016. Thumbnail-based questionnaires for the rapid and efficient collection of
2018) and 24 August earthquakes (https://www.facebook.com/ macroseismic data from global earthquakes. Seismol. Res. Lett. https://doi.org/10.
MyanmarEarthquakeCommittee/photos/a.119139391576451/ 1785/0220160120.
656250071198711/?type=3&theater; last accessed 15 September Buchon, M., Barker, J.S., 1996. Seismic response of a hill: the example of Tarzana,
California. Bull. Seismol. Soc. Am. 86 (1), 66–72.
2018), and digital news media (Table S2, S3). Our post-earthquake
Curray, J.R., Moore, D.G., Lawver, L.A., Emmel, F.J., Raitt, R.W., Henry, M., Kieckhefer,
ground and interview surveys were carried out in the epicentral regions R., 1979. Tectonics of the Andaman sea and Burma. In: Watkins, T.S., Martadet, L.,
within a week of the mainshocks. Early instrumented seismicity was Dickerson, P.W. (Eds.), Geol. Geophy. Invest. Conti. Marg., AAPG Memoir, no. 29, pp.
extracted from relocations by the ISC-GEM project (Storchak et al., 189–198.
Dasgupta, S., Mukhopadhyay, B., Mukopadhyay, M., 2003. Earthquake forerunner as
2013) available online (http://www.isc.ac.uk/iscgem/request_ probable precursor - an example from North Burma Subduction zone. J. Geol. Soc.
catalogue.php, last accessed December 2016). Instrumental ground India 80 (3), 393–402. https://doi.org/10.1007/s12594-012-0157-1.
motion data are available (please contact corresponding author) from the Debbarma, J., Martin, Stacey S., Suresh, G., Ahsan, A., Gahalaut, Vineet K., 2017.
Preliminary observations from the 3 January 2017, MW 5.6 Manu, Tripura (India)
Myanmar National Seismic (MM) network operated and maintained by earthquake. J. Asian Earth Sci. 148, 173–180.
the Department of Meteorology and Hydrology (DMH), Myanmar, the Erdik, M., 1987. Developing a Comprehensive Earthquake disaster Masterplan for
National Centre for Seismology (NCS), India, and the Thailand Me- Istanbul. In: Issues in Urban Earthquake Risk. NATO ASI Series, vol. 271. pp.
125–166.
teorology Department (TMD), Thailand. The original photo in Fig. 4b is Findlay, R., Park, Cyn-Young, Verbiest, Jean-Pierre A., 2016. Myanmar: Building eco-
extracted from The Voice Daily Journal (15 August 2016). nomic foundations. Asian Pac. Econ. Lit. 30 (1), 42–64. https://doi.org/10.1111/
Supplementary data to this article can be found online at https:// apel.12133.
Gahalaut, V.K., Martin, S.S., Srinagesh, D., Kapil, S.L., Suresh, G., Saikia, S., Kumar, V.,
doi.org/10.1016/j.tecto.2019.04.016. Dadhich, H., Patel, A., Prajapati, S.K., Shukla, H.P., Gautam, J.L., Baidya, P.R.,
Mandal, S., Jain, A., 2016. Seismological, geodetic, macroseismic and historical
Acknowledgements context of the 2016 Mw 6.7 Tamenglong (Manipur) India earthquake. Tectonophysics
688, 36–48. https://doi.org/10.1016/j.tecto.2016.09.017.
Gee, E.R., 1934. The Dhubri earthquake of 3rd July, 1930. Mem. Geol. Sur. India 65.
We would like to sincerely acknowledge Mr. Shan Htay (Senator of Grünthal, G. (Ed.), 1998. The European Macroseismic Scale EMS-98. vol. 15. Conseil de
No. 1, Kani Township constituency, Sagaing Regional Parliament), Ms. l'Europe, Cahiers du Centre Europeén de Geodynamique et de Seismoligie,
Ei Hmone Nanthar Myo (University of Yangon), and Ms. Thin Hlaing Oo Luxembourg, pp. 101.
Gutenberg, B., Richter, C., 1949. Seismicity of the Earth and Associated Phenomenon.
(UNDP) for their co-operation during the post-earthquake field survey Hafner, New York.
of Kani area, and Ms. Hnin Ei Win (Yangon City Development Hough, S.E., Martin, S.S., Bilham, R., Atkinson, G.M., 2002. The 26 January 2001 M 7.6
Committee) for arranging the accommodation during field work in the Bhuj, India, earthquake: Observed and predicted ground motions. Bull. Seismol. Soc.
Am. 92 (6), 2061–2079.
Bagan area. We would also like to thank Ms. Su Su Myint (UN Habitat, Hough, S.E., Martin, S.S., Gahalaut, V., Joshi, A., Landes, M., Bossu, R., 2016. A com-
Yangon) and Mr. Pyae Sone Aung (Earth Observatory of Singapore) for parison of observed and predicted ground motions from the 2015 Mw7.8 Gorkha,
their help collecting and sorting data from the Facebook portal of the Nepal, earthquake. J. Inter. Soc. Prev. Mitig. Nat. l Haz. https://doi.org/10.1007/
s11069-016-2505-8.
Myanmar Earthquake Committee. We also wish to thank Mr. Soe Thura Hrin Nei Thiam, Htwe, Yin Myo Min, Kyaw, Tun Lin, Pa, Pa Tun, Min, Zaw, Hninn Htwe,
Tun (Chairman of Resource and Environmental Conservation Su, Aung, Tin Myo, Lin, Kyaw Kyaw, Aung, Myat Min, de Cristofaro, J., Franke, M.,
Committee at Pyithu Hluttaw (the Lower House), Myanmar Parliament) Radman, S., Lepiten, E., Wolin, E., Hough, S.E., 2016. A report on upgraded seismic
monitoring stations in Myanmar: Station performance and site response. Seismol.
and Mr. Saw Htwe Zaw (vice-president of Myanmar Earthquake
Res. Lett. 88, 926–934. https://doi.org/10.1785/0220160168.
Committee, Myanmar Engineering Society) for their valuable sugges- Hurukawa, N., Tun, Pa Pa, Shibazaki, B., 2012. Detailed geometry of the subduction
tions and discussions on damage caused by the earthquakes. Our thanks Indian plate beneath the Burma Plate and subcrustal in the Burma Plate derived from
joint hypocenter relocation. Earth Planets Space 64, 333–343.
are also extended to Ms. Kannikan Poolcharuansin (Thailand
Hutchison, C.S., 1989. Geological Evolution of South-East Asia. Clarendon Press, Oxford,
Meteorological Department) for sharing the ground motion data of pp. 368.
TMD's seismological stations from Thailand. We would also like to Johnston, A.C., 1996. Seismic moment assessment of earthquakes in stable continental
thank Koen van Noten and an anonymous reviewer for constructive regions-I. Instrumental seismicity. Geophys. J. Int. 124 (2), 381–414.
Le Dain, A.Y., Tapponnier, P., Molnar, P., 1984. Active faulting and tectonics of Burma
comments that improved this manuscript. This study was financially and surrounding regions. J. Geophys. Res. 89 (B1), 453–472. https://doi.org/10.
supported by the National Research Foundation Singapore and the 1029/JB089iB01p00453.
Singapore Ministry of Education under the Research Centers of Martin, S.S., Hough, S.E., 2015. The 21 May 2014 Mw 5.9 Bay of Bengal earthquake:
macroseismic data suggest a high-stress-drop event. Seismol. Res. Let. 86 (2A),
Excellence initiative. Part of this research was supported by the 369–377.
Ministry of Science and Technology, Taiwan (MOST) (108-2116-M-002- Martin, S.S., Hough, S.E., 2016. Reply to, “Comment on ‘Ground Motions from the 2015
001-MY3 to Y.W). This is Earth Observatory of Singapore contribution Mw 7.8 Gorkha, Nepal, Earthquake Constrained by a Detailed Assessment of
Macroseismic Data’ by Stacey Martin, Susan E. Hough, and C. Hung” by Andrea
number 241. Tertulliani, Laura Graziani, Corrado Castellano, Alessandra Maramai, Antonio Rossi.
Seism. Res. Lett. 87 (4), 957–962. https://doi.org/10.1785/0220160061.
References Martin, S.S., Kakar, D.M., 2012. The 19 January 2011 Mw 7.2 Dalbandin earthquake,
Balochistan. Bull. Seism. Soc. Am. 100 (4), 1810–1819. https://doi.org/10.1785/
0120110221.
A Shin Thu Mana, 2014. New Method English Grammar, 2nd edition. Nee Thit publishing, Martin, S.S., Szeliga, W., 2010. A catalog of felt intensity data for 589 earthquakes in
Yangon, pp. 1–672. India, 1636-2008. Bull. Seismol. Soc. Am. 100 (2), 536–569. https://doi.org/10.
Adhikari, S.R., Basyal, G., Dixit, A., Landès, M., Bossu, R., Martin, S., Hough, S.E., 2017. 1785/0120080329.
Towards a unified intensity map for the 2015 Gorkha, Nepal, earthquake. Earthquake Martin, S.S., Hough, S.E., Hung, C., 2015. Ground Motions from the 2015 Mw 7.8 Gorkha,
Spectra 33, S4. https://doi.org/10.1193/120716EQS226M. Nepal, earthquake constrained by a detailed assessment of macroseismic data.
Allen, T.I., Wald, D.J., 2009. On the use of high-resolution topographic data as a proxy for Seismol. Res. Lett. 86 (6), 1524–1532.
seismic site conditions (Vs30). Bull. Seismol. Soc. Am. 99 (2A), 935–943. Maurin, T., Masson, F., Rangin, C., Min, U.T., Collard, P., 2010. First global positioning
Atkinson, Gail M., Boore, David M., 2003. Earthquakes and their application to Cascadia system results in northern Myanmar: constant and localized slip rate along the
and other regions. Bull. Seismol. Soc. Am. 93 (4), 1703–1729. Sagaing fault. Geology 38 (7), 591–594. https://doi.org/10.1130/G30872.1.
Atkinson, Gail.M., Kaka, S.I., 2007. Relationship between felt intensity and instrumental Mazumdar, K., Nageswaran, B., 1988. An interm report on the 6th August 1988 earth-
ground motion in the central United States and California. Bull. Seismol. Soc. Am. 97 quake of Indo-Burma Region. Geol. Sur. India 1–13 (unpublished report).

159
Lin Thu Aung, et al. Tectonophysics 765 (2019) 146–160

Middlemiss, C.S., 1910. The Kangra earthquake of 4th April 1905. Mem. Geol. Sur. India earthquake catalogue (1900–2009). Seism. Res. Lett. 84 (5), 810–815.
38 (Government of India (1981 reprint). Win Swe, 2006. Earthquake hazard potentials in Myanmar: a science to public welfare
Mieno, F., 2013. Toward Myanmar's new stage of development: transition from military outlook. In: The Symposium on Tectonics, Seismotectonics, and Earthquake Hazard
rule to the market. Asian Econ. Pol. Rev. 8 (1), 94–117. Mitigation and Management of Myanmar, (Yangon, Myanmar, abstract).
Musson, R., 1998. Intensity assignments from historical earthquake data: issues of cer- Szeliga, W., Hough, S.E., Martin, S.S., Bilham, R., 2010. Intensity, magnitude, location,
tainty and quality. Ann. Geof. 41 (1), 79–91. and attenuation in India for felt earthquakes since 1762. Bull. Seisomol. Soc. Am.
Musson, R.M.W., Grünthal, G., Stucchi, M., 2010. The comparison of macroseismic in- 100, 570–584. https://doi.org/10.1785/0120080329.
tensity scales. J. Seismol. 14, 413–428. Tandon, A.N., Mukherjee, S.M., 1956. Manipur-Burma border earthquake of 22nd March
Myanmar Earthquake Committee, 2003. Survey Report on Taung Twin Gyi Earthquake, 1954. Indian J. Meteoro. Geophys. 7 (1), 27–36.
22 September 2003. Myanmar Engineering Society (unpublished report). Tandon, A.N., Srivastava, H.N., 1974. Earthquake Occurrence in India, Earthquake
Nagashima, I., Maseki, R., Shinozaki, Y., Toyama, J., Kohiyama, M., 2012. Study on Engineering, Jai Krishna 60th Birth Anniversary Commemoration Volume. Indian
perfromance of semi-active base solution system using earthquake observation re- Society for Earthquake Technology (ISET).
cords. In: International Symposium on Reliability Engineering and Risk Management Tapponnier, P., Molnar, P., 1977. Active faulting and tectonics in China. J. Geophys. Res.
2012, pp. 437–442 Yokohama, Japan, Conference Paper. 82, 2905–2930.
Nielsen, C., Chamot-Roke, N., Gangin, C., 2004. From partial to full strain partitioning Tapponnier, P., Molnar, P., 1979. Active faulting and Cenozoic tectonics of the Tien Shan,
along the Indo-Burmese hyper-oblique subduction. Mar. Geol. 209, 303–327. Mongolia, and Baykal regions. J. Geophys. Res. 84, 3425–3459.
Nutalaya, S., Sodsri, S., Arnold, E.P., 1985. SEASEE series on seismology, Thailand, Shwe Gaing Thar, 1976. Historical Earthquake and the Bagan, Sanay Literature Group,
Southeast Asia Association of Seismology and Earthquake Engineering and U. S. Geol. Yanogn, Myanmar (in Burmese). pp. 267–293.
Survey 2, 1–403. Thawbita, 1976. Chronology-Earthquakes of Burma. J. Burm. Res. Soc 59 (1–2), 97–99.
Pananont, P., Herman, M., Pornsopin, P., Furlong, K.P., Waldhauser, F., Kosuwan, S., The World Bank, 2014. International Development Association Project Appraisal
2017. Implications of the Seismo-Tectonics damage patterns of the 2014 Chiang Rai, Document on a Proposed Credit in the Amount of SDR 20.60 Million (USD 31.5
Thailand, earthquake. In: 14th Annual Meeting of Asia Oceania Geosciences Society, Million Equivalent) to the Republic of the Union of Myanmar for a
(Singapore, abstract). Telecommunications Sector Reform Project. report no. 80558-MM. pp. 57.
Pivnik, D.A., Nahm, J., Tucker, R.S., Smith, G.O., Nyein, K., Nyunt, M., Maung, P.H., Maung Thein, Than Myint, Soe Thura Tun, Tint Lwin Swe, 2009. Earthquake and tsunami
1998. Polyphase deformation in a Fore-Arc/Back-Arc Basin, Salin Subbasin, hazard in Myanmar. J. Earthq. Tsun. 3 (2), 43–57.
Myanmar (Burma). AAPG Bull. 82, 1387–1856. Van Noten, K., Lecocq, T., Sira, C., Hinzen, K.G., Camelbeeck, T., 2017. Path and site
Pubellier, M., Chamot-Rooke, N., Ego, F., Guezou, J.C., Konstantinovskaya, E., Rabaute, effects deduced from merged transfrontier internet macroseismic data of two recent
A., Ringenbach, J.C., 2008. Structural map of Eastern Eurasia. Commission for the M4 earthquakes in northwest Europe using a grid cell approach. Solid Earth 8,
Geological Map of the World 1 (12.500.000 scale). 453–477.
Rao, N.P., Kumar, M.R., 1999. Evidences for cassation of Indian Plate subduction in the Wald, D.J., Quitoriano, V., Dengler, L., Dewey, J.W., 1999. Utilizations of the internet for
Burmese Arc region. Geophys. Res. Lett. 26, 3149–3152. rapid community intensity maps. Seismol. Res. Lett. 70, 680–697.
Rastogi, B.K., Singh, D.D., 1978. Source parameters of the Burma-India border earthquake Wang, Y., Sieh, K., Aung, T., Min, S., Khaing, S.N., Tun, S.T., 2011. Earthquakes and slip
of July 29, 1970, from body waves. Tectonophysics 51 (3–4), 77–84. rate of the southern Sagaing fault: insights from an offset ancient fort wall, lower
Satyabala, S.P., 1998. The historical earthquakes of India. In: Diggles, M.E. (Ed.), Burma (Myanmar). Geophys. J. Int. 185, 49–64. doi: 10.111/j.1365-246X.2010.
International Handbook of Earthquake and Engineering Seismology, pp. 1–5 Chap. 1. 04918.x.
Satyabala, S.P., 2003. Oblique plate convergence in the Indo-Burma (Myanmar) sub- Wang, Y., Shyu, J. Bruce H., Sieh, K., Chiang, Hong-Wei, Wang, Chung-Che, Aung, Thura,
duction region. Pure Appl. Geophys. 160, 1611–1650. Lin, Yu-nung Nina, Shen, Chuan-Chou, Min, Soe, Oo, Than, Lin, Kyaw Kyaw, Tun, Soe
Saw Htwe Zaw, 2006. Hazard assessment in multi-hazard design. In: The Symposium on Thura, 2013. Permanent upper plate deformation in western Myanmar during the
Tectonics, Seismotectonics, and Earthquake Hazard Mitigation and Management of great 1762 earthquake: Implications for neotectonic behavior of the northern Sunda
Myanmar, Yangon, Myanmar, (abstract). megathrust. J. Geophys. Res. Solid Earth 118, 1277–1303. https://doi.org/10.1002/
Saw Htwe Zaw, Ornthammarath, T., Poovarodom, N., 2017. Seismic reconnaissance and jgrb.50121.
observed damage after the Mw 6.8, 24 August 2016 Chauk (central Myanmar) Wang, Y., Sieh, K., Tun, S.T., Lai, K.-Y., Myint, T., 2014. Active tectonics and earthquake
earthquake. J. Earthq. Eng. https://doi.org/10.1080/3632469.2017.1323050. potential of the Myanmar region. J. Geophys. Res. Solid Earth 119 (4), 3767–3822.
Sharma, B., Chopra, S., Chingtham, P., Kumar, V., 2016. A study of characteristics of Wang, X., Wei, S., Wang, Y., Phyo, M.M., Hubbard, J., Banerjee, P., Huang, B.-S., Oo,
ground motion response spectra from earthquakes recorded in NE Himalayan region: K.M., Bodin, T., Foster, A., Almeida, R., 2018. A preliminary 3D shear-wave velocity
an active plate boundary. Nat. Hazards. https://doi.org/10.1007/s11069-016- model for Myanmar region. J. Geophys. Res. Solid Earth 123. https://doi.org/10.
2543-2. 1029/2018JB016622.
Singh, S.K., Suresh, G., Dattatrayam, R.S., Shukla, H.P., Martin, S.S., Havskov, J., Iglesias, Worden, C.B., Gerstenberger, M.C., Rhoades, A.D., Wald, D.J., 2012. Probabilistic re-
A., 2013. The Delhi 1960 earthquake: epicenter, depth and magnitude. Curr. Sci. 105 lationships between ground-motion parameters and modified Mercalli intensity in
(8), 1155–1165. California. Bull. Seism. Soc. Am. 102, 204–221. https://doi.org/10.1785/
Socquet, A., Vigny, C., Chamot-Rooke, N., Simons, W., Ranginand, C., Ambrosius, B., 0120110156.
2006. India and Sunda plates motion and deformation along their boundary in World Health Organization, 2016. Emergency situation report (ESR -1), 14 April 2016.
Myanmar determined by GPS. J. Geophys. Res. Solid Earth 111 (5), 1–11. https://doi. http://www.searo.who.int/entity/emergencies/mmr-earthquake-sitrep1.pdf?ua=1,
org/10.1029/2005JB003877. Accessed date: 28 September 2018.
Soe Thura Tun, Min Swe, Tin Maung Nyunt, Win Naing, 2003. September 22 Earthquake Wu, Y., Teng, Ta-liang, Shin, Tzay-Chyn, Hsiao, Nai-Chi, 2003. Relationship between
in the Taungdwingyi Area, Western Central Myanmar, School Family Day National peak ground acceleration, peak ground velocity, and intensity in Taiwan. Bull.
Seminar. Depart. Geol., University of Yangon, Yangon, Myanmar (abstract). Seismol. Soc. Am. 93 (1), 386–396.
Stadtner, D.M., 2005. Ancient Pagan. vol. 286 River Books, Bangkok. Youngs, R., Chiou, S., Silva, W., Humphrey, J., 1997. Strong ground motion attenuation
Stadtner, D.M., 2011. Sacred Sites of Burma: Myth and Folklore in an Evolving Spiritual relationships for subduction zone earthquakes. Seismol. Res. Lett. 68, 58–73.
Realm, Bangkok. vol. 2011978-974-9863-60-2. Yuan, J.L., Li, S.C., 2013. Study of the seismic damage regularity of ancient masonry
Steckler, M.S., Humayun Akhter, S., Seeber, L., 2008. Collision of the Ganges- pagodas in the 2009 Wenchuan earthquake. WIT Trans. Built Env. 132, 421–432
Brahmaputra delta with the Burma Arc: implication for earthquake hazard. Earth (ISSN 1743-3509).
Planet. Sci. Lett. 273, 367–378. Zhang, L.P., Gang, S.A., Hong Sheng, M.A., Wang, X.Z., Hai, L.Z., 2013. The plate contact
Steckler, M.S., Mondal, D.R., Akhter, S.H., Seeber, L., Feng, L., Gale, J., Hill, E., How, M., geometry investigation based on earthquake source parameters at Burma arc sub-
2016. Locked and loading megathrust linked to active subduction beneath the Indo- duction zone. Sci. China 12. https://doi.org/10.1007/s11430-012-4578-x.
Burman Ranges. Nat. Geosci. 5. https://doi.org/10.1038/NGEO2760. Zhao, J.X., Zhang, J., Asano, A., Ohno, Y., Oouchi, T., Takahashi, T., Ogawa, H., Irikura,
Storchak, D.A., Giacomo, D.D., Bondár, I., Engdahl, E.R., Harris, J., Lee, W.H.K., K., Thio, H.K., Somerville, P.G., Fukushima, Y., Fukushima, Y., 2006. Site classifi-
Villaseñor, A., Bormann, P., 2013. Public release of the ISC-GEM global instrumental cation based on predominant period. Bull. Seismol. Soc. Am. 96 (3), 898–913.

160