Bulletin of the Seismological Society of America, Vol. 99, No. 4, pp. 2155–2168, August 2009, doi: 10.

1785/0120080113

Tectonic Geomorphology of the Southernmost Sagaing Fault

and Surface Rupture Associated with the May 1930
Pegu (Bago) Earthquake, Myanmar
by Hiroyuki Tsutsumi and Takayuki Sato*

Abstract The Sagaing fault is a continental transform fault between the India and
Sunda plates that connects spreading centers in the Andaman Sea and the continental
convergence zone along the Himalayan front. Several M > 7 earthquakes occurred
along the fault in the last century, and Global Positioning System campaigns revealed
a right-lateral slip rate of 18 mm=yr, about half of the total India–Sunda displacement
rate of 35 mm=yr. However, there are few fundamental geologic data on the Sagaing
fault, such as detailed fault trace locations, information on late Quaternary slip rate, or
spatial extent of surface ruptures during historical earthquakes. We conducted geo-
logic field investigations along the southernmost 120 km long stretch of the fault zone
that ruptured during the M 7.3 1930 Pegu (Bago) earthquake. We found well-defined
tectonic geomorphic features across a deltaic lowland, including fault scarps, tectonic
depressions, stream offsets, and pressure ridges. The sense of displacement is predom-
inantly right-lateral strike-slip with vertical motion less than 1=5 of horizontal motion.
Based on right-lateral offsets of stream channels, terrace risers, and property bound-
aries, we estimate coseismic displacement during the 1930 earthquake as ≥ 3:0 m.
This in turn gives us a preliminary recurrence interval of surface-rupturing earth-
quakes on the Sagaing fault east of Yangon as ≥ 160 yr. Coulomb stress change in-
duced by the May 1930 Pegu (Bago) earthquake may have triggered the December
1930 Pyu earthquake immediately to the north. This study demonstrates that detailed
geomorphic investigations are key for a better assessment of seismic hazard for the
Sagaing fault.

Online Material: Field photographs of the southernmost Sagaing fault.

Introduction
The Sagaing fault is a continental transform fault more Pronounced tectonic geomorphic expression along the
than 1000 km long that accommodates right-lateral motion Sagaing fault had been noted on aerial photographs (Dey,
between the India and Sunda plates (Fig. 1). It connects 1968), satellite images (Le Dain et al., 1984), and portions
spreading centers in the Andaman Sea and a continental con- of the Sagaing fault were mapped in detail by previous stud-
vergence zone along the Himalayan front (Yeats et al., 1997). ies (Win Swe, 1981; Win Swe and Soe Thura Tun, 2007).
Portions of the Sagaing fault ruptured during large historical However, the precise surface trace locations, geologic slip
earthquakes in the past two centuries (Chhibber, 1934; Win rates, rupture extent of historical earthquakes, and paleoseis-
Swe, 1981; Le Dain et al., 1984); the fault is considered to be mic history of the Sagaing fault are poorly known. These
one of the most active in the world with a Global Positioning data are essential to better understand the active tectonics
System (GPS)-derived slip rate of 18 mm=yr (Socquet et al., of the eastern margin of the India-Eurasia collision zone.
2006). This is about half of the rate of motion between the In addition, the seismic hazard evaluation of the fault has
India and Sunda plates, which includes displacements at the important implications in Myanmar because the Sagaing
fault is located near population centers such as Yangon and
Arakan Trench and Arakan Yoma (Myanmar word meaning
Mandalay (Fig. 1).
mountain) in addition to the Sagaing fault (Fig. 1).
To better understand the seismic hazard posed by this
fault, we mapped the Sagaing fault based on stereo pairs of
*
Now at: Meiji Gakuin High School, Kita-Kyusyu 804-8558, Japan. CORONA satellite images available from the U.S. Geological

2155

Downloaded from http://pubs.geoscienceworld.org/ssa/bssa/article-pdf/99/4/2155/3673300/2155.pdf

by guest
 2156 H. Tsutsumi and T. Sato

project between the Myanmar Earthquake Committee, Kyoto

University, and Hiroshima University since 2007, we were
able to interpret 1:24000 scale aerial photographs and
1:63360 scale (1 inch
1 mile) topographic maps published
by the Myanmar government.
In this article, we describe tectonic geomorphic features
along the southernmost 120 km long portion of the Sagaing
fault and surface rupture associated with the 1930 Pegu
(Bago) earthquake. Based on these field data, we then dis-
cuss the slip distribution and source processes of the 1930
earthquake, the recurrence interval of large earthquakes on
the southernmost Sagaing fault, and coseismic Coulomb
stress changes that might have triggered a subsequent large
earthquake to the north.

Tectonic Setting of the Sagaing Fault

Northeastward movement of the India plate relative to
the Sunda plate predominantly controls the active tectonics
of Myanmar (Fig. 1). This oblique convergence is partitioned
into subduction along the Arakan trench and Arakan Yoma
and right-lateral movement on north-trending faults, such
as the Sagaing fault (Yeats et al., 1997). The wedge shaped
Burma plate is possibly a fore arc sliver separated from
the Sunda plate by the Sagaing fault (Curray et al., 1979;
Curray, 2005).
The Sagaing fault is generally interpreted as a transform
plate boundary between the Burma and Sunda plates (Curray
et al., 1979; Le Dain et al., 1984; Yeats et al., 1997; Curray,
2005). The fault transects Myanmar and extends southward
into the Andaman Sea where it connects to northeast-trending
spreading centers. Farther south of the spreading centers, the
Sumatra fault is a right-lateral strike-slip fault traversing the
island of Sumatra parallel to the subduction zone to the west
(Sieh and Natawidjaja, 2000).
The Sagaing fault has long been known as a major geo-
logic structure in Myanmar (e.g., Chhibber, 1934); however,
its geologic expression of recent movement has been recog-
nized only since the 1960s. Dey (1968) identified geomor-
phic evidence for recent movement of the Sagaing fault in
central Myanmar based on aerial photograph interpretation.
Figure 1. Tectonic setting of the Sagaing fault and surrounding Win Swe (1981) mapped the Sagaing fault zone in the Saga-
region. Locations of major structures are from Nielsen et al. (2004) ing Hills west of Mandalay and identified the fault as primar-
and Soe Thura Tun (2005). Rupture areas for the 1762 Arakan and ily a right-lateral strike-slip fault. Le Dain et al. (1984) first
2004 Sumatra-Andaman earthquakes are from Cummins (2007).
Locations and magnitudes of large historical earthquakes are from mapped the entire fault trace based on interpretation of
Chhibber (1934), Le Dain et al. (1984), and Yeats et al. (1997). Landsat images and seismicity. Sato (2003) prepared prelim-
Shaded relief map was drawn with the GMT software (Wessel inary active fault maps of the Sagaing fault with base maps at
and Smith, 1998). scales of 1∶50; 000 ∼ 1∶63; 360 based on interpretation of
CORONA satellite images. These maps were updated based
on interpretation of 1:24,000 aerial photographs in 2007.
Survey. During February and March 2002, we conducted field The fault extends from north of Lake Indawgyi southward
observations along the southernmost 120 km long onland along the Ayeyarwaddy River north of Mandalay and along
stretch of the Sagaing fault that is accessible from Yangon the eastern margin of the Bago Yoma to the Andaman
and Bago. We were able to identify recent tectonic geomorphic Sea (Fig. 1).
features as well as surface rupture associated with the 1930 A regional GPS survey suggests that the Sagaing fault
Pegu (Bago) earthquake. Through a collaborative research accommodates 18 mm=yr of right-lateral strike slip, half

Downloaded from http://pubs.geoscienceworld.org/ssa/bssa/article-pdf/99/4/2155/3673300/2155.pdf

by guest
 Tectonic Geomorphology of the Sagaing Fault and Surface Rupture with the May 1930 Pegu Earthquake 2157

of the 35 mm=yr relative motion between the India and tions from the Gulf of Martaban and the Sittang estuary
Sunda plates (Socquet et al., 2006). However, there is no in- (Chhibber, 1934). Most of the study area is less than 15 m
dependent late Quaternary geologic slip rate estimation for in elevation except for isolated low hills that stand a few tens
the fault. The Sagaing fault has been seismically active in of meters above the surrounding flat land. These hills are
the past 200 yr with several destructive earthquakes along composed of silt, sand, and gravel, probably correlated to
the fault, such as the 1839 Ava (Innwa), May 1930 Pegu the Irrawaddy Group of middle Miocene to early Pleistocene
(Bago), December 1930 Pyu, and 1956 Sagaing earthquakes age (Bender, 1983).
(Chhibber, 1934; Le Dain et al., 1984; Yeats et al., 1997; The study area was severely hit by the 5 May 1930 Pegu
Win Swe and Soe Thura Tun, 2007). (Bago) earthquake that occurred at 8:18 p.m. (Brown et al.,
1931). The town of Pegu (Bago) was ruined with a loss of at
least 500 lives. In Rangoon (Yangon), more than 50 people
Geologic Overview of the Study Area and the 1930 were killed. Isoseismals on the Rossi-Forel scale show a
Pegu (Bago) Earthquake
north-trending elongated pattern and the area of isoseismal
We conducted geologic field observations along the IX follows the trace of the Sagaing fault between Bago
southernmost section of the Sagaing fault from north of and Thongwa (Fig. 3). The magnitude was estimated at
Payagyi southward to the coastline (Fig. 2). Here, the Saga- M 7.3 (Gutenberg and Richter, 1949). Brown et al. (1931)
ing fault traverses the deltaic lowland built up by tidal accre- described the damage and surface deformation related to the

96°E 97°E

0 20km Fig . 4a
Fig. 4a

17°30’ N
Payagyi Nyaungkhashe

Bago

Fig. 4b Thanatpin SITTANG

S ITTANG
RIVER

Onhne
Pale
ay
ilw
ra

Fig. 4c
Sagaing
S

17°N
ag aingfault
ER
V
RI

Kayan
GO

fault
BA

Yangon

Syriam Thongwa
Fig. 4d

GULF OF
OF MARTABAN
MARTABAN 16°30’ N

Figure 2. Index map of the study area. See Figure 1 for location of the map. Locations of Figure 4a–d are shown.

Downloaded from http://pubs.geoscienceworld.org/ssa/bssa/article-pdf/99/4/2155/3673300/2155.pdf

by guest
 2158 H. Tsutsumi and T. Sato

Figure 3. (a) Rossi-Forel scale isoseismals, ground motion direction, and direction of collapse of buildings and pagodas associated with
the 1930 Pegu (Bago) earthquake compiled from Brown et al. (1931). (b) Ground-motion direction associated with northward-propagating
rupture along a north-trending right-lateral strike-slip fault (Koketsu, 1996).

earthquake in detail; we introduce some of their important fault from latitude N17°32′ southward to the coastline during
observations in the following sections. Although Brown et al. February and March 2002. We used a GPS receiver to record
(1931) and Chhibber (1934) suggested that the 1930 earth- the locations of the observation points. We took topographic
quake was caused by the movement of the Sagaing fault, profiles across fault scarps using a hand level, and we mea-
little geologic evidence on the surface rupture has been pre- sured the horizontal offsets of natural and artificial features
sented until now. across the fault zone using measuring tapes. We also inter-
viewed local people for information on ground deformation
and damage during the 1930 earthquake. In 2007, we were
Tectonic Geomorphology of the Southernmost
able to interpret 1:24,000 scale aerial photographs of the
Sagaing Fault and Surface Deformation Associated
with the 1930 Earthquake study area, which enabled us to complete a detailed active
fault map. In the following sections, we describe tectonic
We first mapped the surface trace of the Sagaing fault geomorphology of the fault zone and surface deformation
based on interpretation of CORONA satellite images. We then associated with the 1930 earthquake. To facilitate descrip-
conducted a 1 month long field survey along the Sagaing tion, we divide the study area into four sections (Fig. 4)

Downloaded from http://pubs.geoscienceworld.org/ssa/bssa/article-pdf/99/4/2155/3673300/2155.pdf

by guest
 Tectonic Geomorphology of the Sagaing Fault and Surface Rupture with the May 1930 Pegu Earthquake 2159

Figure 4. Detailed fault map of the study area. The base map is 1 inch
1 mile topographic maps published by the Myanmar govern-
ment. Contour interval is 50 ft; contour lines above 150 ft are not shown. (Continued)

Downloaded from http://pubs.geoscienceworld.org/ssa/bssa/article-pdf/99/4/2155/3673300/2155.pdf

by guest
 2160 H. Tsutsumi and T. Sato

Figure 4. Continued.

Downloaded from http://pubs.geoscienceworld.org/ssa/bssa/article-pdf/99/4/2155/3673300/2155.pdf

by guest
 Tectonic Geomorphology of the Sagaing Fault and Surface Rupture with the May 1930 Pegu Earthquake 2161

and several field photographs are shown in Figure 5. ⒺIn (1934) reported that large cracks appeared, from which water
addition, ∼30 field photographs are available in the elec- and sand were extruded, in parts of the town. Chhibber
tronic edition of BSSA. (1934) also noted that the most striking surface crack was
the one along the eastern foot of the Bago hill, along which
alluvium was piled up to the height of approximately 1 ft.
Wunbe-In to Bago
The crack was remarkably straight with a strike of N5°E.
In the northernmost section of the study area, the Sagaing This crack may have appeared as mole tracks along the sur-
fault is identified as continuous west-facing fault scarps. For face trace of the Sagaing fault we mapped in this study.
a 2.5 km long stretch between the Kanbe Chaung and the
Wagadok Chaung (Myanmar word meaning stream), fresh
Bago to Pale
west-facing scarps exist (Fig. 4a). The height of the scarps
is as large as 1 m but is about 20 cm on a younger alluvial plain A clear west-facing fault scarp about 0.3 m high is ob-
near the Wagadok Chaung (location 1, Fig. 5a). The higher served immediately north of Saingdi (location 8). About
scarps appear to have been produced by multiple faulting 2 km south of Saingdi, there is an anticlinal uplift west of
events. North of the Kanbe Chaung, the fault extends across the fault trace (location 9). This anticline is about 1 km long,
hilly terrain of the Bago Yoma. 50 m wide, and 1 m high, and the eastern edge of the anti-
West of Payagyi, the Sagaing fault cuts across an ancient cline is bounded by the Sagaing fault. North of Mokainggyi,
walled city called Payagyi Myo Haung (location 2). The land there is a sharp restraining bend of the fault trace across
enclosed by the wall is traversed by a less than 1 m high, which the direction of vertical displacement changes. At
west-facing fault scarp that is probably related to the 1930 location 10, the height of the pronounced fault scarp on a
earthquake. The wall appears to be offset right laterally fluvial terrace is 2.2 m, whereas right-lateral offset of a ter-
across the scarp on aerial photographs taken in 1953, but race riser between the terrace and modern stream channel is
we could not confirm or measure the offset due to recent ar- 9.5–11.8 m. This indicates that the horizontal component of
tificial modification for farming. The Shwedan In (Myanmar displacement is at least 5 times the vertical component be-
word meaning lake) is a tectonic depression or sag pong on cause the terrace riser is younger than the terrace surface
the down-thrown side of the fault. The height of the fault above. Between Mokainggyi and Pale, the surface trace of
scarp near the northern end of the Shwedan In is about 2.8 m the Sagaing fault is not clear due probably to erosion and/
(location 3). A clear west-facing fault scarp extends south of or sedimentation of the Paingkyun Chaung and its tributaries.
Shwedan Village (location 4, Fig. 5b). At location 5, two Between Zenyaungbin and Pale, we identified a fault
fault strands run parallel to one another. The west-facing west of and almost parallel to the Sagaing fault (Fig. 4b).
eastern scarp is 0.9 m high, whereas the east-facing western We describe this fault, named the Pale fault here, from south
scarp is 1.6 m high (Fig. 6a). Between location 5 and Tetkale, to north. In Pale, a clear west-facing scarp about 1 m high
fault scarps are obvious both on aerial photographs and in the extends to the north (location 14, Fig. 5d). There is no major
field. They are mostly continuous but alternate between east- north-flowing stream around the site, and this scarp is inter-
and west-facing scarps. The fault crosses two railroads that preted as tectonic in origin. Daw Own Tin, 48 years old when
were in place in 1930. Brown et al. (1931) reported damage interviewed in 2002, heard about the 1930 earthquake from
to the buildings of major railway stations such as the Tawa her grandmother who passed away in 1998 at the age of 80.
and Payagyi stations. However, there was no description of The grandmother noticed ground rupture across rice fields
damage to the railroads, and we do not know whether the along the present scarp. The Makaw Pagoda in the village
railroads crossed by the Sagaing fault were displaced during was severely damaged by the earthquake but was later recon-
the earthquake. structed. At location 13, a terrace riser of a west-flowing
From Tetkale to Bago, numerous creeks obscure the lo- stream channel is right laterally offset 8:6  0:3 m. North
cation of the fault trace. The city district of Bago has been and south of the Paingkyun Chaung, an elongated linear de-
developed on a bedrock hill. The eastern margin of the hill is pression about 10 m wide and 1 m deep extends along the
sharply defined by the Sagaing fault, which marks a bound- fault. North of the Paingkyun Chaung, there are numerous
ary against alluvial lowland to the east (between the road to isolated low hills. The strata exposed at a hill northeast of
the north and the road to Thanatpin). The hill is located at a Kyadet (location 12) is mainly composed of reddish yellow,
restraining bend of the right-lateral strike-slip Sagaing fault unconsolidated, fine sand to silt with round metamorphic
(Fig. 4a). South of the road to Thanatpin, clear faulted geo- gravels 1 ∼ 2 cm in diameter. At location 11, a small west-
morphic features are observed. At location 6, a terrace riser flowing creek about 1.5 m wide and 1 m deep is offset right
associated with an east-flowing stream channel is displaced laterally 13:8  0:3 m across the fault trace. A west-facing
right laterally 8:3  0:3 m. At location 7, an east-flowing fault scarp of the Pale fault extends northward to Tawa where
channel is deflected to the right 18:6  0:5 m (Fig. 5c), two buildings were completely destroyed, and seven persons
which is probably due to multiple earthquakes. were killed during the 1930 earthquake (Brown et al., 1931).
There are numerous accounts of surface deformation Discontinuous but clear fault scarps are identified between
during the 1930 earthquake in and around Bago. Chhibber Tawa and Zenyaungbin. The total length of the Pale fault

Downloaded from http://pubs.geoscienceworld.org/ssa/bssa/article-pdf/99/4/2155/3673300/2155.pdf

by guest
 2162 H. Tsutsumi and T. Sato

Figure 5. Photographs of typical tectonic geomorphic features of the Sagaing and Pale faults in the study area. The red arrows indicate
the location of the fault. The yellow arrows show piercing points for lateral offset. See Figure 4 for locations of the photographs.

Downloaded from http://pubs.geoscienceworld.org/ssa/bssa/article-pdf/99/4/2155/3673300/2155.pdf

by guest
 Tectonic Geomorphology of the Sagaing Fault and Surface Rupture with the May 1930 Pegu Earthquake 2163

(a) (m) W Loc. 5 Vertical exaggeration: 20 times E

2

1 1.6 m

0.9 m
0
0 100 (m)
(b) W Loc. 17 E
(m)
0.5
0.5 m
0
0 100 (m)

(c) (m) “earthquake

W Loc. 22 crack creek” E
0.5

0
0 10 0 (m)

Figure 6. Topographic cross sections across the Sagaing fault in the study area. See Figure 4 for locations of the cross sections.

is about 17 km. The Pale fault lies 1:5 ∼ 2 km west of and fault trace is largely obscured by erosion and sedimentation
parallel to the Sagaing fault north of location 12. To the by the Kayan Chaung and the Kamapadaw Chaung (Fig. 4c).
south, the fault trends N10°W, more northwesterly than the South of Tamangyi, tectonic geomorphic features of the
Sagaing fault (Fig. 4b). Sagaing fault are well preserved. At location 18, an irrigation
canal that seems to have originated from a natural stream
channel dissecting the alluvial plain is right laterally offset
Pale to Thongwa 3:2  0:3 m. East of Shanzu, there is a north-northeast-
trending, west-facing, 2.5 km long fault west of the main
The main trace of the Sagaing fault becomes clear south trace. Southward from Shanzu, the fault trace is mostly rec-
of Pale along the eastern margin of a north-trending low hill ognized as left stepping en echelon faults rather than contin-
on which Moganein is located (Fig. 4b). At location 15, an uous faults, probably related to the greater thickness of
east-trending terrace riser is offset right laterally 5:8 unconsolidated sediments to the south. At location 19, a
0:2 m. Around location 16 east of Moganein, the fault trace broad stream channel is offset by the fault. The northern edge
is recognized as left-stepping, east-facing scarps less than of the channel is offset right laterally 3:3  0:2 m. South of
50 cm high (Fig. 5e). At location 16, there is a 2 ∼ 3 m high this area, there is no geomorphic evidence for cumulative dis-
and 20 m long pressure ridge. Monk U Chit, 80 years old, placement of the fault, suggesting that the land surface there
identified north-trending cracks at the eastern base of the may have experienced only the 1930 earthquake. At loca-
pressure ridge after an earthquake that occurred when he was tion 20 east of Yebawgan, a 10 m wide channel cutting into
a child. These cracks were subsequently filled by sediments the alluvial plain is offset right laterally 1:6  0:3 m. The
and disappeared. vertical offsets at locations 18–20 are less than 20 cm.
Farther south, the Sagaing fault bounds the eastern mar-
gin of the Zwegaik In (Fig. 4b). The height of the west-facing
scarp is about 0.5 m at location 17 (Fig. 6b). Two males, both Thongwa to the Coast
age 80 when interviewed in 2002, live near the lake and In Thongwa, the fault trace is unclear due mostly to ar-
experienced the 1930 earthquake. They witnessed water tificial modification (Fig. 4d) although Brown et al. (1931)
shooting out of the ground during the earthquake, and a lake reported numerous accounts of fissures associated with the
appeared on the down-thrown side of the fault after the earth- 1930 earthquake near the village. Immediately north of the
quake. We interpret that the water vented to the surface Aungzeyone Pagoda, the northern and southern property
due to liquefaction filling a portion of the preexisting depres- boundaries of a house are sharply offset right laterally (loca-
sion (Zwegaik In). They also noted that the lake subsequently tion 21). The amounts of offsets were measured at 1:6  0:1
started being filled by sediments and gradually became shal- and 1:5  0:1 m, which we believe represent offset during
lower with time. The Kabauk In is a pull-apart basin located at the 1930 earthquake because straight features make a sharp
a right step of the Sagaing fault (Fig. 4c). A clear west-facing turn exactly at the fault line. U Thein Han, 80 years old when
scarp continues southward from Kanbe. West of Kayan, the interviewed, was 8 years old when he experienced the 1930

Downloaded from http://pubs.geoscienceworld.org/ssa/bssa/article-pdf/99/4/2155/3673300/2155.pdf

by guest
 2164 H. Tsutsumi and T. Sato

earthquake. He identified cracks immediately east of the steps and bends that are related to local uplift or subsidence.
Aungzeyone Pagoda that collapsed during the earthquake. The fault trace is commonly composed of series of left-
Although he noticed numerous sand blows, there was no stepping en echelon faults that correspond to Riedel shears
severe damage to houses in Thongwa. Clear west-facing, for right-lateral strike-slip faults.
10 ∼ 20 cm high scarps continue south of the Aungzeyone We suggest that the dip of the Sagaing fault in the study
Pagoda. area is close to vertical based on the observations that the
Between Thayagon and Tagundaing-anauk, the Hmaw- fault trace is remarkably straight, that the direction of vertical
wun Chaung meanders along the fault trace, and most of the offset changes from place to place, and that isoseismals of
tectonic geomorphic features are lost. U Chet, 83 year old of the 1930 earthquake are elongated along the fault trace.
Tagundaing-anauk, experienced the 1930 earthquake when Transtensional deformation such as linear troughs and sag
he was 11 years old. He heard a loud sound at the time ponds is commonly observed along the fault zone. This is
of the earthquake. He noticed abundant sand blows but probably because the right-lateral strike-slip Sagaing fault
did not notice surface break. Nobody was killed, and damage in the study area trends more northerly (N1°W) than the fault
was minimal in the village. to the north (N10°W), which results in a divergent compo-
South of Tagundaing-anauk, a west-facing scarp with nent of slip across the fault in the study area (Fig. 1).
maximum height of ∼50 cm extends to the south. Southward In addition to the main trace of the Sagaing fault, there
from the Hmawwun Chaung, the fault trace is marked by are faults branching from the main trace. The Pale fault is a
transtensional deformation; linear depressions develop along 17 km long, predominantly right-lateral strike-slip fault
the fault zone. Occasionally en echelon, west-facing scarps branching northward from the Sagaing fault. To the south,
less than a few tens of centimeters high are observed. there is a 2.5 km long southward-branching fault near Shanzu.
South of Tada, a creek flows along the linear depression.
The villagers call this creek earthquake crack creek because
Extent of Surface Rupture, Slip Distribution, and
it appeared after the 1930 earthquake. At location 22, a linear
Source Processes Associated with the 1930 Pegu
depression about 10 m wide is observed (Figs. 5f, 6c). U Kya
(Bago) Earthquake
Nyunt, 83 years old, told us that furniture fell down, but there
was no damage to his house, and nobody was killed or in- The Sagaing fault in the study area is marked by fresh
jured in Myagaing during the 1930 earthquake. At loca- and well-defined tectonic geomorphic features. We were able
tion 23, a clear west-facing scarp cuts across rice fields. to confirm that the Sagaing fault south of Bago ruptured dur-
Rice paddy dikes parallel to the coastline are systematically ing the 1930 Pegu (Bago) earthquake by interviewing local
offset right laterally, but the amount of offset is less than 1 m. people who experienced the earthquake. North of Bago, we
The coastline on the CORONA images taken in May 1965 do not have an eyewitness account of the surface rupture dur-
was at location 24, and since then the coastline has migrated ing the earthquake, and the fresh geomorphic features there
∼3 km southward. The fault trace that was visible across the could be related to the December 1930 Pyu earthquake. The
tidal flat on the CORONA images can be traced now in the December earthquake caused severe damage to Pyu located
field down to location 25. Because the faulted geomorphic about 130 km north of Bago (Fig. 7). Most of the masonry
features gradually become less distinct, and the amount of buildings in Pyu were wrecked, and about 30 people were
offset is smaller to the south, we believe that the southern killed (Brown and Leicester, 1933). The isoseismals are
termination of the 1930 surface rupture lies close to this centered on Pyu, and the area of isoseismal VIII extends
point. south to the latitude of Pyuntaza. The December earthquake
caused no damage to Bago or Payagale located 4 km north of
Payagyi (Brown and Leicester, 1933) in contrast to the May
Discussion earthquake for which the area of isoseismal VIII extends
north to Payagyi. We thus suggest that the fresh tectonic geo-
Fault Trace Geometry and Sense of Movement
morphic features in the study area between Wunbe-in to the
of the Sagaing Fault
coast are associated with the May 1930 earthquake. Lack of
The surface trace of the Sagaing fault in the study area is damage descriptions north of Bago in Brown et al. (1931) is
remarkably straight with a general strike of N1°W. The trace probably due to the absence of large population centers;
is continuous except for places where sedimentation or ero- Payagyi is the only village in the study area north of Bago
sion of modern stream channels obscured tectonic landforms. described in Brown et al. (1931).
Systematic right-lateral offsets of streams and terrace risers Horizontal displacements associated with past surface-
suggest that the Sagaing fault is predominantly a right-lateral rupturing earthquakes along the Sagaing fault in the study
strike-slip fault. The vertical component of displacement is area are recorded as offsets of streams, terrace risers, and
much smaller than the horizontal component, and the direc- property boundaries. We plot the right-lateral offsets mea-
tion of vertical displacement (east side up versus west side sured in the field in Figure 8; the values range from less
up) changes from place to place, a commonly observed fea- than 1 m (location 23) to more than 15 m (location 7). South
ture for strike-slip faults. The fault trace contains small-scale of latitude 17° N, the fault scarps are mostly less than 20 cm

Downloaded from http://pubs.geoscienceworld.org/ssa/bssa/article-pdf/99/4/2155/3673300/2155.pdf

by guest
 Tectonic Geomorphology of the Sagaing Fault and Surface Rupture with the May 1930 Pegu Earthquake 2165

high, and there is no geomorphic evidence for cumulative

displacement on the Sagaing fault probably because the del-
taic flood plain there is very young. We suggest that the mea-
sured offsets represent displacements during the 1930 Pegu
(Bago) earthquake; the 1930 coseismic displacement is about
3 m north of Thongwa and gradually decreases to the south.
North of latitude 17° N, the measured offsets are larger than
5 m, but the features measured there tend to be older than the
floodplain south of 17° N. We do not know whether these
offsets are associated only with the 1930 earthquake or the
1930 and preceding earthquakes. We suggest that the maxi-
mum coseismic displacement of the entire 1930 Pegu (Bago)
earthquake rupture was at least 3 m.
Although we were not able to measure horizontal offsets
north of Bago, fresh topographic scarps are present in the
northernmost part of the study area (location 1). We thus sug-
gest that a section of the southernmost Sagaing fault at least
120 km long from north of Payagyi to the coastline ruptured
during the 1930 Pegu (Bago) earthquake. We calculate the
moment magnitude for the earthquake using the relations
Mo
μAs (Brune, 1968) and log Mo
1:5Mw  16:1
(Kanamori, 1977), where Mo is the seismic moment,
μ
3:2 × 1011 dyne=cm2 is the rigidity, A is the area of
the fault rupture, and s is the average fault displacement.
Assuming the length (120 km), width (15 km), and average
slip (3 m) for the 1930 Pegu (Bago) earthquake, we can cal-
culate the magnitude of the earthquake as Mw 7:4. For com-
parison, Gutenberg and Richter (1949) estimated M 7.3 for the
earthquake.
The collapsed direction of buildings and pagodas and
direction of ground movement compiled from Brown et al.
(1931) are shown in Figure 3. Near the fault trace, the col-
lapsed direction of buildings and pagodas was southeast
(Bago), east and east-southeast (Tawa), and west-northwest
(Tongyi). This suggests that the direction of ground shaking
was east–west to northwest–southeast in these areas. In con-
trast, the ground shaking direction was north–south away
from the fault at Insein, Yangon, Syriam, and Kyauktan. This
ground shaking pattern is consistent with that associated with
strike-slip (left- or right-lateral slip) on a north-trending ver-
tical fault (Fig. 3; Koketsu, 1996). In comparison, the ground
shaking pattern for dip-slip (normal or reverse slip) on a
north-trending fault would be east–west both near and away
from the fault (T. Iwata, personal commun., 2008). This ob-
servation supports our field observation that the southern-
most Sagaing fault moved right laterally during the 1930
Pegu (Bago) earthquake.
Figure 7. Coulomb stress changes for the Sagaing fault caused Based on observations of historical surface ruptures,
by the May 1930 Pegu (Bago) earthquake calculated by Coulomb Nakata et al. (1998) proposed that earthquake rupture prop-
2.6 (Toda et al., 1998). Three meters of uniform right-lateral slip
was given for the 120 km long and 15 km wide vertical rectangular
agates in the same direction as fault trace branches. For the
fault plane for the May 1930 earthquake. Calculations are made for Sagaing fault in the study area, the Pale fault branches north-
receiver fault planes with strike (N10°W), dip (80° W), and width ward from the main trace of the Sagaing fault, and there is a
(15 km), assuming a shear modulus of 3:2 × 1011 dyne cm2 , Pois- fault east of Shanzu that branches southward. If the branch-
son’s ratio of 0.25, and apparent coefficient of friction μ0
0:4. See ing pattern is a reliable indicator of rupture direction, it sug-
Figure 1 for location of the map.
gests that the 1930 earthquake rupture initiated somewhere
between the branching points and propagated bilaterally to

Downloaded from http://pubs.geoscienceworld.org/ssa/bssa/article-pdf/99/4/2155/3673300/2155.pdf

by guest
 2166 H. Tsutsumi and T. Sato

Right-lateral offset (m)

0 5 10 15 20

17°
Payagyi 30’ N

Bago
Bago
Loc. 6 Loc. 7
Pale fault

Loc. 10
Loc. 11
Pale Loc. 13

Loc. 15
Sagaing fault
Sagaing fault

proposed 17°N
rupture
initiation
point

Loc. 18
Loc. 19
Loc. 20

Thongwa Loc. 21

Loc. 23

proposed
rupture 16°
propagation 0 20km 30’ N
direction

Figure 8. Right-lateral offsets measured along the Sagaing fault in the study area. Bars with open circle indicate preferred values with
error bars, whereas bars indicate ranges of right-lateral offset.

the north and south (Fig. 8). If this had been the case, east– quakes would be 3m=18mm=yr
160 yr or longer. The
west shaking along the fault trace to the north and south of fault has accumulated elastic strain of about 1.5 m during the
the epicenter would have been significantly amplified. Much past ∼80 yr.
less damage in Thongwa and villages farther south compared
to Bago can be explained by a smaller population and de-
creasing slip to the south. Coulomb Stress Change due to the May 1930 Pegu
Although there is no paleoseismic data for the south- (Bago) Earthquake and Its Relation to the December
ernmost section of the Sagaing fault, we can estimate a pre- 1930 Pyu Earthquake
liminary recurrence interval of large surface-rupturing The Pegu (Bago) earthquake on 5 May 1930 was fol-
earthquakes like the 1930 event. The horizontal slip rate of lowed by great shocks on 3 and 4 December 1930 (Chhibber,
the Sagaing fault is 18 mm=yr from GPS analysis (Socquet 1934). The severest shock occurred at about 1:20 a.m. on the
et al., 2006), while coseismic horizontal slip during the fourth and killed about 30 people in Pyu (Chhibber, 1934).
1930 earthquake was at least 3 m. Therefore, assuming char- The magnitude for the largest shock is estimated at M 7.3
acteristic slip, the recurrence interval of 1930-type earth- (Gutenberg and Richter, 1949). The May and December

Downloaded from http://pubs.geoscienceworld.org/ssa/bssa/article-pdf/99/4/2155/3673300/2155.pdf

by guest
 Tectonic Geomorphology of the Sagaing Fault and Surface Rupture with the May 1930 Pegu Earthquake 2167

events were only 7 months apart and probably occurred on Okumura, and Shinji Toda are greatly acknowledged. Shigeru Sueoka pre-
adjacent segments of the Sagaing fault. pared shaded relief maps in Figures 1 and 7. Comments on an early version
of the manuscript by Bob Yeats and Nobuhiko Sugito greatly improved the
To explore the relations between the events, we calcu- manuscript. We thank Kelvin Berryman, Aron Meltzner, and an anonymous
lated Coulomb stress changes for the Sagaing fault caused by reviewer for their through reviews of the manuscript. Funds for purchasing
the May earthquake (Fig. 7). North of the study area, the the CORONA satellite images were provided by the United Nations. This
Sagaing fault trends more westerly (∼N10°W) and marks work was partially supported by Kyoto University Active Geosphere Inves-
tigations for the 21st Century Centers of Excellence Program (KAGI21),
a topographic boundary between the Bago Yoma and alluvial
which was approved by the Ministry of Education, Culture, Sports, Science,
lowland of the Sittang River. Our calculation shows that and Technology (MEXT) of Japan.
stress increased by several bars immediately north of the
May earthquake rupture. Previous studies demonstrated that References
a stress increase on the order of a few bars is large enough to
trigger failure on nearby favorably oriented faults (e.g., Stein Bender, F. (1983). Geology of Burma, Gebrüder Borntraeger, Berlin-
Stuttgart, 293 pp.
et al., 1992; Toda et al., 2005). Therefore, it is possible that
Brown, C. J., and P. Leicester (1933). The Pyu earthquake of 3rd and 4th
the May earthquake triggered the December earthquake. December, 1930, and subsequent Burma earthquakes up to January,
1932, Mem. Geol. Surv. India 62, 1–140.
Conclusions Brown, C. J., P. Leicester, and H. L. Chhibber (1931). A preliminary note on
the Pegu earthquake of May 5th, 1930, Rec. Geol. Surv. India LXV,
The Sagaing fault is a right-lateral strike-slip fault that 221–284.
accommodates much of the shear component of the highly Brune, J. N. (1968). Seismic moment, seismicity, and rate of slip along major
fault zones, J. Geophys. Res. 73, 777–784.
oblique convergence between the India and Sunda plates. For Chhibber, H. L. (1934). The Geology of Burma Macmillan and Co., London,
a better assessment of seismic hazard from the Sagaing fault 538 pp.
near Yangon, we conducted geologic field investigations Cummins, P. R. (2007). The potential for giant tsunamigenic earthquakes in
along the southernmost 120 km long stretch of the fault zone the northern Bay of Bengal, Nature 449, 75–78.
that ruptured during the M 7.3 1930 Pegu (Bago) earthquake. Curray, J. R. (2005). Tectonics and history of the Andaman Sea region, J.
Asian Earth Sci. 25, 187–232.
We identified well-defined tectonic geomorphic features Curray, J. R., D. G. Moore, L. A. Lawver, F. J. Emmel, R. W. Raitt,
across the deltaic lowland, including fault scarps, tectonic M. Henry, and R. Kieckhefer (1979). Tectonics of the Andaman
depressions, stream offsets, and pressure ridges. The sense Sea and Burma, in Geological and Geophysical Investigations of Con-
of displacement is predominantly right-lateral strike-slip. tinental Margins, J. S. Watkins, L. Montadert, and P. W. Dickerson
Geomorphic observations and documentation by local peo- (Editors), American Association of Petroleum Geologists Memoir,
29, 189–198.
ple suggest that the Sagaing fault from north of Payagyi to Dey, B. P. (1968). Aerial photo interpretation of a major lineament in the
the coastline ruptured during the 1930 Pegu (Bago) earth- Yamethin-Pyawbwe quadrangle, Union of Burma J. Sci. Tech. 1,
quake. Based on right-lateral offsets of stream channels, ter- 431–443.
race risers, and property boundaries, we can estimate at least Gutenberg, B., and C. F. Richter (1949). Seismicity of the Earth, Princeton
3.0 m coseismic right-lateral displacement. This in turn gives University Press, Princeton, 273 pp.
Kanamori, H. (1977). The energy release in great earthquakes, J. Geophys.
us a preliminary recurrence interval of surface-rupturing Res. 82, 2981–2987.
earthquakes on the Sagaing fault east of Yangon as 160 yr Koketsu, K. (1996). Disastrous earthquakes in California and Hyogo-ken
or longer. Coulomb stress change by the May 1930 Pegu nanbu earthquake, Kagaku 66–2, 93–97 (in Japanese).
(Bago) earthquake may have triggered the December 1930 Le Dain, A. Y., P. Tapponnier, and P. Molnar (1984). Active faulting and tec-
tonics of Burma and surrounding regions, J. Geophys. Res. 89, 453–472.
Pyu earthquake immediately to the north. Similar geo-
Nakata, T., K. Shimazaki, Y. Suzuki, and E. Tsukuda (1998). Fault branch-
morphic investigations are key for a better understanding ing and directivity of rupture propagation, J. Geog. 107, 512–528 (in
of active tectonics and seismic hazard of the Sagaing fault. Japanese with English abstract).
Nielsen, C., N. Chamot-Rooke, and C. Ranginthe ANDAMAN Cruise Team
(2004). From partial to full strain partitioning along the Indo-Burmese
Data and Resources
hyper-oblique subduction, Mar. Geol. 209, 303–327.
Sato, T. (2003). Fault morphology of the Sagaing fault in Myanmar—Surface
All data used in this article were collected by the authors
rupture associated with the Bago earthquake of May 5, 1930, Master’s
or came from published sources listed in the references. Thesis, Hiroshima University, Higashi-Hiroshima, Japan, 140 pp.
Sieh, K., and D. Natawidjaja (2000). Neotectonics of the Sumatran fault,
Acknowledgments Indonesia , J. Geophys. Res. 105, 28295–28326.
Socquet, A., C. Vigny, N. Chamot-Rooke, W. Simons, C. Rangin, and
We are grateful to Yoshimitsu Yamada and Toshiya Akasaki at the Yan- B. Ambrosius (2006). India and Sunda plates motion and deformation
gon Office of Fukken Co., Ltd., for their help during our stay in Myanmar, along their boundary in Myanmar determined by GPS, J. Geophys.
without which our field investigation would not have been possible. We also Res. 111, B05406, doi 10.1029/2005JB003877.
would like to thank our guide U Thiha Soe and driver Tun Min Aung for Soe Thura Tun (2005). The Sagaing Fault—A Desk Study Report on Seis-
their kind cooperation in the field. We thank our colleagues at Myanmar motectonic Implications in Myanmar, Myanmar Earthquake Commit-
Earthquake Committee, especially Win Swe, Soe Thura Tun, Tint Lwin tee, Yangon, 112 pp.
Swe, and Thura Aung for making aerial photographs and topographic maps Stein, R. S., G. C. P. King, and J. Lin (1992). Change in failure stress on the
available and for discussions on geology and tectonics of the Sagaing fault. southern San Andreas fault system caused by the 1992 magnitude

Discussions and guidance by Tomotaka Iwata, Takashi Nakata, Koji 7:4 Landers earthquake, Science 258, 1328–1332.

Downloaded from http://pubs.geoscienceworld.org/ssa/bssa/article-pdf/99/4/2155/3673300/2155.pdf

by guest
 2168 H. Tsutsumi and T. Sato

Toda, S., R. S. Stein, P. A. Reasenberg, J. H. Dieterich, and A. Yoshida Yeats, R. S., K. Sieh, and C. R. Allen (1997). The Geology of Earthquakes,
(1998). Stress transferred by the 1995 Mw
6:9 Kobe, Japan, shock: Oxford University Press, New York, 568 pp.
Effect on aftershocks and future earthquake probabilities, J. Geophys.
Res. 103, 24543–24565.
Department of Geophysics
Toda, S., R. S. Stein, K. Richards-Dinger, and S. B. Bozkurt (2005). Fore-
Kyoto University
casting the evolution of seismicity in southern California: Animations
Kyoto 606-8502, Japan
built on earthquake stress transfer, J. Geophys. Res. 110, B05S16, doi [email protected]‑u.ac.jp
10.1029/2004JB003415. (H.T.)
Wessel, P., and W. H. F. Smith (1998). New, improved version of Generic
Mapping Tools released, EOS 79, 579.
Win Swe (1981). A major strike-slip fault in Burma, Contrib. Burmese Geol. Department of Geography
1, 63–72. Hiroshima University
Win Swe, and Soe Thura Tun (2007). A Field Trip Guide to the Sagaing Higashi-Hiroshima 739-8522, Japan
Fault Zone in the Sagaing Hills Area, Central Myanmar, on the Aus- [email protected]
(T.S.)
picious Occasion of International Workshop on Seismotectonics in
Myanmar and Earthquake Risk Management, 25–26 January 2007,
Myanmar Earthquake Committee, Yangon, 43 pp. Manuscript received 29 May 2008

Downloaded from http://pubs.geoscienceworld.org/ssa/bssa/article-pdf/99/4/2155/3673300/2155.pdf

by guest