You are on page 1of 54

April-May 2015 (Mw 7.8, 7.

4)
Nepal earthquakes:
A perspective

Kusala Rrajendran
CEaS

What has led to our understanding (whatever little) has come out of collaborative effort

Studies for the past 15 years (major collaborator, C.P, Rajendran). Other
collaborators: Jaishri Sanwal, B. Kotlia, Mike Sandiford and Kristin Morel, MU,
Australia)
Students participating in the research work and field work to Nepal
Ms. Revathy Parameswaran (Ph. D student, works on source mechanisms of plate
boundary earthquakes) : Field surveys, source models
Mr. Thusasiraman (Ph.D student, works on site response of earthquakes): Field
surveys, Site response studies, tectonic Geomprphology.
Mr. Rishav Mallick (M.Tech student, works on source mechanisms): Field surveys
Others who participated in the field work
C.P. Rajendran, Faculty, JNCASR
Mathew Wood (Ph.D student with Mike Sandiford; works on landslides,
professional photographer)
We acknowledge the financial support from MOES and IISc

Three great earthquakes (1905, 1934, 1950) have occurred along the
Himalaya in the 1900s. The segment between the ruptures of 1905 and 1934
is considered as a gap, due for a great earthquake. (Rajendran et al 2015)

Results from other studies

Avouac, 2015

Estimated rupture area of major earthquakes along the Himalayas. The 1100 and
1413 AD events were both documented from paleoseismic studies. Bilham and
some coworkers believe that the 1505 historical earthquake must have had a
magnitude close to Mw 9 and with a renewal time of ~ 500 years (based on slip
estimates) they believe that a great earthquake is imminent.

Our studies have used various proxies to


infer the chronology of past earthquakes
1. Heritage structures (damage reports and
observations on temples)
2. Paleoliquefaction features

Major points of dissention with


Bilhams group
On the estimated size and location of
the 1505 earthquake, which finds no
mention in historic records
Believe that every large earthquake
may not originate on the detachment

3. Slip on faults observed in trench sections


4. Cave features (tilt on stalagmite deposits)

Missing slip may be overestimated as


partitioning on the splay faults are
ignored in the geodetic slip models.

Rajendran et al., 2015 concluded that the frontal thrust in central Himalaya may
have remained seismically inactive during the last ~700 years. Considering this long
elapsed time, a great earthquake may be due in the region.
Related publications:

Rajendran and Rajendran (2005, 2011);


Rajendran et al., (2013, SRL; 2015, JGR)

With these basic notions, let us approach


the recent Nepal earthquakes.

Organization of this talk


A brief overview of the tectonics of the Himalaya
Introduction to some terms that we need to
understand the source properties of earthquakes
The Nepal earthquake: field observations, damage,
deformation features
Source models and relation to seismogenic structures
Remarks on why the threat of an impending
earthquake continue to remain

Sketch from Avouac, 2015

Himalaya is the result of collision of India-Asia plates. Prior to collision, the Tethys
Sea, separated the two. The southern margin of Asia was an active margin with a
subduction zone (like Sumatra subduction zone). Age of the onset of collision is
estimated to be between 65 and 45 Ma. Paleolatitudes (variety of data including
paleimagnetic data) from sites just north and south of the suture zone suggest that
the collision occurred 44 and 55Ma.

At present, northern India is moving ~35 mm


year-1 along a N15 E azimuth (31 mm year -1
along a N10 E azimuth at the western
Himalayan syntaxis and 38 mm year-1 along a

N20E azimuth at the eastern Himalayan


syntaxis (Bettinelli et al., 2006). Crustal
shortening across the Himalayas, estimated

at 19 2.5 mm year-1 absorbs nearly half of


the current convergence rate between India
and Eurasia.

Convergence rates over the last 80 Ma;


kinematics obtained from the synthesis
of the magnetic anomalies of the
Indian Ocean (Royer and Patriat, 2002)

And the resulting features: ~ 2500 km long plate boundary, the uplifted
mountain ranges an the Tibetan Plateau..

India-Eurasia collision zone showing location of major topographic features. The red
arrow represents convergence between Bangalore (IISC) in southern India with stable
Eurasia. This convergence is characterized by shortening across the Himalaya, Tibetan
Plateau, and Tien Shan [Zhang et al., 2004].

The highest mountain range + the largest high-altitude plateau on earth;


seismically most active continental collision boundary

The youngest fault system that


accommodates the slip

Tapponier et al.,2001

And how are these thrust systems evolving?

Simplified geologic map of Himalayan arc (Avouac, 2015).


Section A A (next slide) shows structures in section

Critical Taper
In

geodynamics

the

concept is used to explain


tectonic observations in
accretionary wedges.

Applied to the Himalaya

Himalaya is seen as an analogue to the

wedge of sand that forms in front of a


moving bulldozer (Davis et al., 1983).
This wedge is critical, meaning that any

point within the wedge is at the verge


of failure.
Schematic geologic section across the Himalayas of central Nepal reflecting early
interpretation of the major thrust faults as crustal-scale parallel faults. Analogy with
the geometry of a sand wedge at front of a bulldozer

Landscape carved by such propagating wedges

Locked, breaks in great eqs


section

EQ
Aseismic slip
MCT: 2421 Ma; MFT: < 2 Ma
ITSZ Indus-Tsangpo suture zone; STD South Tibetan detachment; MCT Main Central thrust;
MBT Main Boundary thrust; MFT Main Frontal thrust

The faults get progressively younger towards south, but there are out-ofsequence thrusts (change in order of age) that are mapped.
For example, a physiographic break south of MCT has been identified as an out of
sequence thrust (younger than MBT).
Seismogenic potential of such thrusts have not been understood hitherto, as
there was no known association with large/great earthquakes.
The 2015 sequence may provide the first example.

Section of detachment

The South Tibetan detachment (north of


Mount Everest). Tethyan Sedimentary
Series in the hanging wall with marble,
calc-silicates, and more in the footwall
(Jessup et al., 2006).

View of a thrust sheet in the Nepal


sub-Himalaya

On which structure did the earthquake occur?


Geologic cross-section of
central Nepal Himalaya,
Source of 2015 eq?

showing

location

of

tectonic units and major


thrust faults. Figure from
Pandey et al. [1995].

The earthquake was shallow (<15 km); with a thrust faulting a 10 degree dipping plane.
The rupture (of the first eq) propagated from west to east.

So, what is the causative structure? Is there a surface rupture?


Did it propagate all the way to the Indogangetic plains as was expected?
Second earthquake occurred where the first rupture terminated.
It had a short rupture duration and aftershocks clustered in a small area.

Summary of plate motions and background seismicity

Motion of India relative to


Eurasia
models

by

two

different

Bettinelli

(2006).

Frontal fault defined by thrust


earthquakes
dipping

on

planes

shallow
(10-15);

shallow, within 30 km.

Great earthquakes originate on the detachment and propagate to the surface (eg.
1903, 1934). Surface ruptures have not been observed as the fault-propagation
folds absorb them before reaching the surface. Lack of surface ruptures is also
linked to the difficult field conditions and rapid weathering of outcrops. A limitation
for paleoseismologic studies in the Himalaya

A brief on focal mechanism

Normal fault: 1 vertical, e.g


gravitational load.

P-axis (~ Shmax)

T-axis (~ Shmin)

Thrust fault: 1 horizontal; 3


vertical. (e.g. Himalaya).

P- P axis (direction of compression)


T- T axis (direction of tension axis)
Beach Ball solution: A way of representing the style of
faulting. Point source assumption, match on the P-waves.
Does not work well with large ruptures. Hence, moment
tensor inversions and teleseismic waveform modeling.
Strike slip fault: 1 and 3
are also horizontal. (e.g.,
San Andreas Fault USA)

Teleseismic body-wave inversion allows modeling a longer part of


the waveform and is more representative of the rupture

Basic platform: Seismic Analysis Code (SAC)


Technique: construction of synthetic wave forms giving best match
with the observed seismogram
Stations are selected from II (IRIS/IDA), IU (IRIS/USGS) and GT (Global
Telemetric seismograph network) or Global Seismic Network (GSN)
Selection criteria: azimuth, epicentral distance (35-100) and signal
quality
Teleseismic Body wave inversion program by Prof. Kikuchi and Prof.
Kanamori, is used to calculate moment rate, maximum dislocation,
focal mechanism, rupture extent and moment magnitude.

Organization of this talk


A brief overview of the tectonics of the Himalaya
Introduction to some terms that we need to
understand the source properties of earthquakes
The Nepal earthquake: field observations, damage,
deformation features
Source models and relation to seismogenic structures
Remarks on why the threat of an impending
earthquake continue to remain

Source mechanisms for the 7.8 and 7.3 earthquakes

April 25th, 2015


Mw 7.8

May 12th, 2015


Mw 7.3

Pink stars: Post May 12; Blue stars: Pre May 12


Scaled to size 3.3 to 7.8

Kathmandu
city
Arniko Highway

Gorkha

East ward directed


rupture

Lumbini

Bhaktapur
Route map followed by the team during the course of the field work.

Bhaktapur

Outskirts of Kathmandu
city

Kathmandu city

Swayambhunath Temple,
Kathmandu

Outskirts of Kathmandu city

Damage to heritage structures


Pashupathinath Temple (15th century)
caught the attention mostly because

the main shrine survived both the


1934 and 2015 events. The only
notable damage was in the main
entrance where the wooden pillar has
detached from the brick-frame.

Poonsva Mahadev (believed to be more


than 200 years old) stands adjacent to
Pashupathinath. Sustained severe damage
in 1934.

Bhaktapur

Sikhara- style Temple on Darbar Square; the largest


Sikhara tower ever built in Nepal, using bricks. Damaged
in 1934 and dome was replaced later.

Note: The tower was not affected by


the 1833 earthquake, north of
Kathmandu

What remains

On April 25th, there were 67-70 tourists within the Dharara and many perished.
Words of the shop-owners who work ~10 m away: We were inside, at the farther
end of the shop, when the shaking began. We could not move. However we managed
to somehow cross the length of the shop (~2m) and head to the door. We saw the

tower sway twice. And then it collapsed completely in the next 3-5 seconds It killed
almost everyone inside

Kathmandu Valley

Liquefaction
Sites studied in some detail

Subsidence

Slumping on the Araniko Highway

The central part of the road slumped by almost 90 cm. Ground cracks could be traced
these cracks all the way into the alleys Kinematic GPS surveys on both sides of the road

so as to create a Digital Elevation Model (DEM) of the surface. We learned that the
slumped part of the road used to be a lake-bed many years ago, which was filled up to
make the highway. Numbers are GPS location numbers for reference.

Extensive damage

Subsidence of
Highway

Extensive damage

Liquefaction
Sites studied in some detail
Ambient noise measurements for site response studies made at
10 locations including the sites shown here

Kathmandu valley (600 sqkm): filled by lacustrine


deposits derived from the surrounding mountains.

Liquefactions at Chagunarayan

Some far field effects

Everest shaken but no change in height

The 25 April rupture did not get


close to Mt. Everest but the
elastic

strain

from

distant

rupture did. The figure shows


that it lies close to the zero line
of height change. If anything its
height sank 1 or 2 mm in the

earthquake.

http://cires1.colorado.edu/~bilham/2015%20
Nepal/Nepal_2015_earthquake.html

~ 600 km away
North: ~ 2mm

East: ~ 3mm

24 hour positions using final orbits


(Blue and using rapid orbits,
Majenta).
Source: Nevada Geodetic Lab.

Vertical: ~ 8 mm

Source parameters

(H= source depth)

(var.=residual
between original
and synthetic
waveforms)

Maximum slip is
east of the source
location

(Strike, dip, rake)

Maximum slip : 4.5 m

(var.=residual
between original
and synthetic
waveforms)
(H= source depth)

Maximum slip is at
the source location

(Strike, dip, rake)

Maximum slip : 3.41m

USGS finite fault model

A potential surface Rupture? And where to look for?


Surface Ruptures of great Himalayan earthquakes have not been mapped so far.
Only suggestion for a surface rupture comes from the recent palseoseimological
studies (Sapkota

et al., 2013, Nature Geoscience) that suggest that the 1934

earthquake must have indeed produced a surface rupture. It is believed to have


ruptured along the Main Himalayan Thrust.
If the 2015 earthquake also ruptured in a similar way (as proposed by current
geodetic models), we should find evidence for rupture in the plains, but the initial
aerial surveys and satellite images could not suggest any.
Our search for surface rupture was based on our understanding of the geometry of
faults and the INSAR images that marked out uplift and subsidence.

Geometry of Fault

The region near the star (mainshock) slipped up to 4 m displacing the


boundary between Nepal and Tibet decimeters southward.
A striking feature of the earthquake is the apparent absence of slip on the
main frontal thrust faults in the northern plains of Nepal.
http://cires1.colorado.edu/~bilham/2015%20Nepal/Nepal_2015_earthquake.html

INSAR image of the source of the 2015 earthquake


Mapped surface rupture (~ 100 m)
Subsidence : 0.5 m

Uplift: 1 m

Source: http://topex.ucsd.edu/nepal/
Expected rupture along the uplifted area; we search along accessible sites.

This geodetic method uses two or more synthetic aperture images to generate maps of
surface deformation.

Uplift-Erosion-extrusion driven model for the tectonic evolution of Himalaya

Chanel Flow

Beaumont et al., 2001

Coupled thermal- mechanical models show that channel flow and ductile extrusion
may be dynamically linked through the effects of surface denudation focused at the
edge of a plateau that is underlain by low-viscosity material.
Uplift- followed by denudation- followed by extrusion of underlying, partially molten
material shapes the tall mountain ranges.

PT2: Physiographic break controlled by extrusion

Chanel Flow
Earthquakes
(Hodges et al., 2004).

PT2 is defined as a sharp physiographic break characterized by expressed by a


variety of landform changes: (1)Narrow and steep-walled gorges in the north to
alluviated valleys in the south (2) abrupt decrease in hill-slope gradient (3)

abrupt transition from landslide covered hill-slopes to weathered hillslopes.

Wobus et al., (Nature, 2005) proposed this as an out-of sequence thrust ( Out of

sequence in age), but its seismogenic potential not known.

Mw 7.8
Mapped Rupture

Mw 7.3

Slope map showing Physiographic transition, PT2, is prominent break in


hillslope gradient between yellow arrows (Wobus et al., Geology 2003).

N
E

PT2

Projected station locations

Himalayan crust above


MSL
Two-dimensional north-south
structure of Vp and Vp /Vs .
White dots: microseismicity.
(base map from Monsalve
et al., 2008)

Vp/Vs image shows a region of relatively low ratio north of latitude 27.5 N at depths
from the surface to 40 km BSL. Note that both the earthquakes occurred on the
boundary of high-low velocity. The aftershocks of the second earthquake did not

propagate towards the low velocity region. Are we seeing evidence for channel flow?

PT2?

(~8000 m)

Implications.

Rupture propagated along the emerging out-of-sequence thrust. The first large
earthquake for which slip is modeled at source (for other earthquakes slip is
inferred)
If indeed there are such out-of-sequence thrusts, it would imply that the ~ 21 mm
year-1 slip estimated to be absorbed at the MFT, is underestimated and the total
shortening rate across the range would subsequently exceed the geodetic rate.
The lesson from the 2015 earthquake, to improve on the geodetic models.
Accomodate for potential previous 2015-type earthquakes.

Organization of this talk


A brief overview of the tectonics of the Himalaya
Introduction to some terms that we need to
understand the source properties of earthquakes
The Nepal earthquake: field observations, damage,
deformation features
Source models and relation to seismogenic structures
Remarks on why the threat of an impending
earthquake continue to remain

What next? Projections by the Bilhams group

http://cires1.colorado.edu/~bilham/2015%20Nepal/Nepal_2015_earthquake.html

Pink zones : potential for great earthquakes.


Indian Kashmir is sufficiently stressed to host one or more Mw8 or greater
earthquakes (Mmax=8.9).

Similarly, Himalayan segments in Sikkim, Bhutan and Arunachal have not


sustained rupture for many hundreds of years.

And what we think

The 1505 earthquake has not ruptured the central segment (Rajendran et al.,
2015); last earthquake here was ~ 700 years B.P. Slip potential exists.
Slip potential also for Kashmir Himalaya, Bhutan and Arunachal segments.
Not all large earthquakes propagate to the plains, they can emerge through out-of
sequence thrusts, as we believe it happened in 2015. Reassessment of missing slip?
Many more known unknowns could surprise us in future?

Nature continues with its acts..

But life goes on

You might also like