Tsunami mapping in the Gulf of Guayaquil, Ecuador, due to local seismicity

Abstract

The North-Andean subduction zone generates
recurrent tsunamigenic earthquakes. The seismicity is
usually considered to be segmented because of different
specific morphological features of the Nazca Plate driving
the subduction motion. Most of the recent powerful
earthquakes in the margin were located in its northern part.
To the south, the region of the Gulf of Guayaquil, only
(undocumented) three events in 1901, 1933 and 1953 were
possibly powerful and tsunamigenic. Here we are interested
in the tsunami signature due to local seismicity. Two
realistic earthquake scenarios (Mw = 7 and Mw = 7.5)
taking into account the hypothesized segmentation of the
area are proposed. Their return period is supposed to be
intra-centenary. Then, a larger magnitude unsegmented
Mw = 8 scenario is computed (half-millennium return
period). The interior of the Gulf of Guayaquil as well as the
Santa Elena Peninsula are sheltered areas including
numerous coastal infrastructures and the city of Guayaquil.
It is predicted that potential flooding would occur at high
tide only for both segmented and unsegmented scenarios in
(1) south of Playas with however only a few centimeters of
wave height and (2) Chanduy (a few meters). Both are
important zones of coastal farms.

Full text

1 23
Marine Geophysical Research
An International Journal for the Study of
the Earth Beneath the Sea
ISSN 0025-3235
Mar Geophys Res
DOI 10.1007/s11001-014-9225-9
Tsunami mapping in the Gulf of
Guayaquil, Ecuador, due to local seismicity
M.Ioualalen, T.Monfret, N.Béthoux,
M.Chlieh, G.Ponce Adams, J.-Y.Collot,
C.Martillo Bustamante, K.Chunga,
E.Navarrete, et al.

1 23
Your article is protected by copyright and all
rights are held exclusively by Springer Science
+Business Media Dordrecht. This e-offprint
is for personal use only and shall not be self-
archived in electronic repositories. If you wish
to self-archive your article, please use the
accepted manuscript version for posting on
your own website. You may further deposit
the accepted manuscript version in any
repository, provided it is only made publicly
available 12 months after official publication
or later and provided acknowledgement is
given to the original source of publication
and a link is inserted to the published article
on Springer's website. The link must be
accompanied by the following text: "The final
publication is available at link.springer.com”.

ORIGINAL RESEARCH PAPER
Tsunami mapping in the Gulf of Guayaquil, Ecuador, due to local
seismicity
M. Ioualalen •T. Monfret •N. Be
´thoux •M. Chlieh •G. Ponce Adams •
J.-Y. Collot •C. Martillo Bustamante •K. Chunga •E. Navarrete •
G. Montenegro •G. Solis Gordillo
Received: 26 January 2014 / Accepted: 30 April 2014
ÓSpringer Science+Business Media Dordrecht 2014
Abstract The North-Andean subduction zone generates
recurrent tsunamigenic earthquakes. The seismicity is
usually considered to be segmented because of different
specific morphological features of the Nazca Plate driving
the subduction motion. Most of the recent powerful
earthquakes in the margin were located in its northern part.
To the south, the region of the Gulf of Guayaquil, only
(undocumented) three events in 1901, 1933 and 1953 were
possibly powerful and tsunamigenic. Here we are inter-
ested in the tsunami signature due to local seismicity. Two
realistic earthquake scenarios (M
w
=7 and M
w
=7.5)
taking into account the hypothesized segmentation of the
area are proposed. Their return period is supposed to be
intra-centenary. Then, a larger magnitude unsegmented
M
w
=8 scenario is computed (half-millennium return
period). The interior of the Gulf of Guayaquil as well as the
Santa Elena Peninsula are sheltered areas including
numerous coastal infrastructures and the city of Guayaquil.
It is predicted that potential flooding would occur at high
tide only for both segmented and unsegmented scenarios in
(1) south of Playas with however only a few centimeters of
wave height and (2) Chanduy (a few meters). Both are
important zones of coastal farms.
Keywords Tsunami Gulf of Guayaquil Local
seismicity Return period
Introduction
The Gulf of Guayaquil (hereafter GG) is located in the
Ecuador/Colombia margin, within the North Andean sub-
duction zone (Fig. 1) The margin generates recurrent
powerful earthquakes of magnitudes larger than M
w
=7
that are most of the time tsunamigenic (Kelleher 1972).
The Ecuador coastal area is relatively densely populated
(more than three million inhabitants in Guayaquil) and
hosts infrastructures such as oil refineries, natural gas
exploitation, as well as shrimp farms that were developed
along the GG within very flat and floodable land topogra-
phy. Distant (far-field) tsunamis, generated outside the GG
area, tend not to generate significant flooding there. To the
south, the destructive and extremely tsunamigenic 1960
M
w
=9.5 Chilean earthquake yielded only a 0.92 m runup
at the Santa Elena Peninsula (north of the GG) (NGDC
tsunami catalog). To the north, no tsunami is even reported
in the GG by NGDC for the destructive 1906 M
w
=8.8
Ecuador/Colombia earthquake. To the west, no trans-oce-
anic tsunami has impacted the GG coastal area (NGDC
data base). However, several small to moderate-sized tsu-
namis have occurred in the GG region in relation with M
w=
7–7.5 local earthquakes (1933 and 1953 events; Espinoza
M. Ioualalen (&)T. Monfret N. Be
´thoux M. Chlieh 
G. Ponce Adams J.-Y. Collot C. Martillo Bustamante
Ge
´oazur, Institut de Recherche pour le De
´veloppement, IRD,
UMR Ge
´oazur 7329, CNRS-IRD-UNS-OCA, 250 Rue A.
Einstein, 06560 Valbonne, France
e-mail: Mansour.ioualalen@geoazur.unice.fr
M. Ioualalen G. Ponce Adams C. Martillo Bustamante 
E. Navarrete G. Solis Gordillo
Escuela Superior Polite
´cnica del Litoral (ESPOL), Facultad de
Ingeniera en Ciencias de la Tierra, FICT, Guayaquil, Ecuador
K. Chunga
Escuela de Ciencias Geolo
´gicas y Ambientales, Facultad de
Ciencias Naturales, Universidad de Guayaquil, Guayaquil,
Ecuador
G. Montenegro
Gerencia de Exploracio
`n y Desarrollo de EP Petroproduccio
´n,
Guayaquil, Ecuador
123
Mar Geophys Res
DOI 10.1007/s11001-014-9225-9
Author's personal copy

1992; Lockridge 1984). Although these earthquakes and
the resulting tsunamis are poorly documented, they might
pose a threat to the region infrastructures. Consequently,
we aim to evaluate the tsunami threat for the GG that could
be triggered by local seismicity. Thus, we focus on tsu-
namis of relatively short wave arrival time (of at most
*1 h). Beyond this value and omitting the fact that the GG
is not severely exposed to distant tsunamis, the Pacific
Tsunami Warning Center (PTWC) system (relayed in
Ecuador by the Ecuadorian Navy, INOCAR) is effective.
For example, the information and evacuation procedures
have been executed efficiently during the 2011 Japan tsu-
nami. Consequently, neither trans-oceanic tsunamis nor
those triggered in Peru or in Central/North Ecuador will be
considered. To achieve this goal, we need to build reliable
(1) earthquake scenarios and (2) numerical simulations of
tsunami propagation and coastal impact. Earthquake rup-
turing scenarios will be derived based on the tectonics and
the seismicity of the area by using available earthquake
scaling laws as well as inversion of global positioning
system (GPS) data. A reliable tsunami propagation
numerical model will be used together with a good
description of bathymetry and adjacent topography. Once
scenarios are built, initial tsunami waves are computed
using Okada (1985)’s formulation. Then the propagating
wave and the runup distribution are computed using Fun-
wave, a fully nonlinear and dispersive tsunami propagation
model (Wei and Kirby 1995). A special effort is dedicated
to the construction of the computational grid with the use
of accurate bathymetric and topographic data sets. Runup
Fig. 1 Geodynamical sketch of
the Ecuadorian margin. Nazca
plate motion is from Trenkamp
et al. (2002) and Nocquet and
Mothes (2009). The stars
represent the great subduction
earthquake events which
occurred during the twentieth
century. The black dots are the
epicenters of events of
magnitude larger than M
w
=4
that were recorded by the
national permanent
seismological network of Peru
(IG-EPN) during the last
20 years. The dashed line
represents the southern limit of
the North Andean Block (NAB)
along the Dolores Guayaquil
Mega-thrust (DGMT)
Mar Geophys Res
123
Author's personal copy

maps are provided along with a tentative identification of
the processes that are responsible for the wave amplifica-
tion and damping, which is a crucial aspect for delivering
maps that are as generic as possible.
Geodynamical and geological context of the Ecuadorian
margin
Along the Ecuadorian margin, the Nazca Plate is con-
verging in a N83°E direction towards the South American
Plate at 7.3 ±2.7 cm year
-1
(Nocquet and Mothes 2009),
which is in agreement with the previous estimation of
5.5–5.8 cm of Trenkamp et al. (2002). North of the equator
the margin turns from a NS into a NE–SE direction.
Consequently the convergence becomes oblique with
respect to the trench direction. This obliquity is accom-
modated by the escape in a 35°N direction of the North
Andean Block (the area between the Cordillera and the
trench) at about 7 mm year
-1
with respect to the stable
South America Plate (Nocquet and Mothes 2009; Fig. 1).
The overall motion of the North Andean Block (hereafter
NAB) occurs along a right-lateral strike-slip fault system
along the Dolores Guayaquil Mega-thrust (DGMT in
Fig. 1; Winter et al. 1993; Ego et al. 1996; Pararas-Cara-
yannis 2012) and consequently favors the opening of the
GG by NNE extension (Ego et al. 1996; Witt et al. 2006).
The GG therefore represents the southern boundary of the
NAB. Within the GG, the strike-slip movement may cause
deformation and underwater landslides (Pararas-Carayan-
nis 2012). However, a subsequent co-seismic displacement
cannot trigger a significant tsunami because (1) pure strike-
slip kinematics cannot generate a threatening tsunami in
the absence of vertical movement of the seafloor (2) the
expected magnitudes are assumed not to reach those of the
subduction cases that we consider here, and (3) an eventual
tsunami is assumed not to amplify significantly through the
slope effect at the coast since it is generated at very shallow
water within the GG. With regard to a potential landslide-
derived tsunami, the weakness of the bathymetric gradient
within the GG should not allow a sufficient acceleration of
a sliding volume to trigger a significant tsunami. Also,
since point (4) also applies for a slide, we do not believe a
threatening landslide-triggered tsunami is likely.
South of the GG (south of 3.5°S, Fig. 1) the Grijalva
Fracture Zone represents the main bathymetric feature on
the Nazca Plate, whereas further north on the Nazca Plate
(north of 2°S, Fig. 1) the Carnegie Ridge, a 200 km-wide
buoyant ridge carried by the down-going oceanic Nazca
Plate, subducts under the Ecuadorian central margin. This
wide and thick ridge segments the margin into three con-
trasting zones (Gutscher et al. 1999; Pedoja et al. 2006;
Collot et al. 2009). The ridge reveals an over-thickened
oceanic crust without evidence of frontal accretion
(Graindorge et al. 2004). North of the Carnegie Ridge, the
margin exhibits normal oceanic crust and the structure of
this segment suggests it was affected by transient phases of
erosion and accretion (Collot et al. 2009). South of the
Carnegie Ridge, the margin shows evidence for subsidence,
which becomes obvious in the GG (Calahorrano et al.
2008). A sedimentary prism faces the margin off the GG
and its width increases southward. These sediments come
mainly from the Andes, transported by the Guayas River,
but also from the Andean Western Cordillera (Calahorrano
et al. 2008). These structural variations along the margin
probably explain the segmentation in the earthquake dis-
tribution along the Ecuadorian margin, at least at relatively
short seismic return periods.
Along the northern segment of the Ecuadorian margin
(north of 0°–0.5°N), four megathrust earthquakes occurred
during the twentieth century (Fig. 1). They are listed here
for information only because their arrival time to the GG
are [1 h: The M
w
=8.8 1906 January 3 earthquake
(1.0°N, 81.5°W) ruptured the entire 400–500 km area
extending from the city of Esmeraldas to South Colombia
(Kelleher 1972; Kanamori and McNally 1982). The strong
earthquake triggered an important tsunami with local runup
ranging from 2 to 6 m. Wave heights of magnitude
0.8–1 m were observed at Bahia de Caraquez and on the
order of 0.4 m in Japan (Espinoza 1992). The M
w
=7.8
1942 May 14 event (0.3°S, 80.0°W, south of Atacames/
Esmeraldas) probably re-activated the southern 80 km
segment of the 1906 event (Beck and Ruff 1989; Kanamori
and McNally 1982; Mendoza and Dewey 1984). No tsu-
nami was reported from that earthquake. The M
w
=7.7
1958 January 19 earthquake (0.1°N, 79.3°W) ruptured
offshore Esmeraldas over a 110 km area (Kanamori and
McNally 1982). Tsunami waves ranging from 2 to 6 m
have been reported (Espinoza 1992). At this stage, it is
interesting to mention that the city of Esmeraldas hosts a
very specific site: The city is located immediately north of
the Atacames submerged promontory (Fig. 1). Ioualalen
et al. (2011) showed how the promontory contributes in
sheltering the area in case of tsunamis arriving from the
south (for example triggered by the 1942 earthquake) by
focusing the tsunami wave above the promontory (and not
at the coast). According to Ioualalen et al. (2011), the
focusing process does not apply for tsunamis coming from
the north (e.g., the tsunami triggered by the 1958 event).
This may be one reason that could explain the difference in
the tsunami signature between the 1942 and 1958 events.
The M
w
=8.2 1979 December 12 event (0.1°N, 79.3°W)
ruptured north of Esmeraldas, offshore Tumaco, Colombia
(Kanamori and McNally 1982). The rupture was estimated
to extend from 110 (Kanamori and McNally 1982)to
240 km (Mendoza and Dewey 1984). The tsunami runup
Mar Geophys Res
123
Author's personal copy

was estimated to be from 2 to 6 m mainly in the area of
Tumaco, southern Colombia (Espinoza 1992). Further
south, the M
w
=7.2 1998 August 4 event (0.6°S, 80.4°W)
happened offshore Bahia de Caraquez (Segovia 2001;
Fig. 1). The earthquake did not trigger a significant tsu-
nami; a tide gauge record at Manta (80 km south of the
epicenter) indicates a 30 cm rise of the water level
(Segovia 2001).
Contrasting with the seismicity studies of the northern
margin, the events that occurred in front of the GG are not
well documented. Four events are reported in the South
American SISRA catalog (http://www.ceresis.org/portal/
catal_hipo.php; Fig. 1). The locations of their epicenters
are approximate indeed: The M
w
*7.2 1901 January 7
earthquake that probably occurred offshore the Santa
Helena Peninsula at *(2°S, 82°W), the M
s
*6.9 1933
October 2 event at *(3.5°S, 80°W), the M
w
*7.5 1953
December 12 earthquake at *(3.4°S, 80.6°W) and the M
s
*7.5 1959 February 7 earthquake at *(4°S, 81.5°W). The
1933 earthquake generated a tsunami with runup on the
order of 2–2.5 m near Salinas/La Libertad (Fig. 1, Sta
Elena Peninsula) (Espinoza 1992). According to Lockridge
(1984), the 1953 event triggered a tsunami that was very
weak near Salinas (*0.2 m). The 1901 event is not doc-
umented regarding tsunamis.
Along the Ecuadorian forearc, the tectonics of the GG
area result from a combination of the subduction effects
and the NAB northward drift (Witt et al. 2006; Fig. 2). The
GG is structured by two main basins, the Esperanza and
Jambeli Basins, and six main active fault systems, the
Puna-Santa Clara Fault System, the Posorja Detachment
System, the Tenguel Fault, the domito fault system (DFS),
the Esperanza Graben and the Jambeli Detachment System
(Witt et al. 2006). The DFS delineates an area strongly
controlled by the subduction process from another area
strongly controlled by the escape of the NAB, inducing an
extensional component and subsidence of the Gulf. At
present day, the tectonics of the GG are dominantly
extensional (Witt et al. 2006). However, at the eastern end
of the GG, the Zampala Fault Zone is evident in the Puna
Fig. 2 Structural sketch of the
Gulf of Guayaquil. The
bathymetry compilation is
obtained from Collot et al.
(2006). Offshore structures are
from Witt et al. (2006). PSCFS
Puna-Santa Clara fault system,
IPBF Banco Peru fault, TF
tenguel fault, PDS posorja
detachment system, JDS
jambeli detachment system,
TDS tumbes detachment system.
Earthquake focal solutions that
are obtained from the global
centroid moment tensor
(GCMT) database available
since 1979, are superimposed.
The black dots represent the
epicenter locations obtained
from the USGS National
earthquake information center
(NEIC) located for the period
1973–2010. The two white stars
are the fault centroids of the
modeled source (scenarios of
Table 1)
Mar Geophys Res
123
Author's personal copy

Island, off the Guayas Estuary. Dumont et al. (2005)
demonstrated from a morphological study that this fault
system is an active strike-slip fault, accommodating the
NAB drift with a mean slip rate of 5.5–8 mm year
-1
.
The active deformation of the GG is very complex, due
to coeval effects of subduction, extension and strike-slip
movements. This complexity is evident from the variety of
focal mechanism solutions available in the area (Fig. 2).
The inverse solutions could correspond to deeper events
located on the interplate zone while the normal or strike-
slip mechanisms are related to the NAB. Nevertheless, we
consider that the most energetic events occurred at the
seismogenic zone between the Nazca and South American
Plates. In the present study, we take the option to consider
only interplate earthquakes. The reasons follow: (1) Sub-
duction tsunamis are likely to exhibit larger waves because
the Green’s effect of the seafloor slope (a 1/4-powered
slope law of wave amplification) is more important when
triggered at deep water (vs. a tsunami triggered by an
earthquake located within the shallow water continental
shelf). Besides, interplate earthquakes are expected to be
larger than intraplate ones; (2) contrary to the interplate
area, the geometry of the upper-plate faults is not yet
documented. Consequently, in this area, there is no infor-
mation to constrain any rupture scenario while interplate
earthquakes are very much studied and numerous scaling
laws are available (see later). Besides, several research
cruises in the area of the GG (Collot et al. 2002) allow a
good description of the interplate geometry. Useful
parameters can then be derived.
The Grijalva Fracture Zone and the Carnegie Ridge are
two major structural features of the Nazca Plate that enter
the Ecuador trench. It has been shown that the thick Car-
negie Ridge appears to act as a buttress against the
southward propagation of the largest earthquakes ruptures,
which instead propagated northward (Collot et al. 2002).
The Grijalva Fracture Zone operates a NE-trending, 700 m
high-scarp that separates the morphology of the ancient
Farallon lithosphere to the south from the younger Nazca
lithosphere to the north (Witt et al. 2006). We therefore
postulate that this main structure may act as a barrier for
seismological rupture propagation. Consequently, a first set
of tsunami source scenarios will be based on this seg-
mentation. We will place the area of margin seafloor
deformation within the area bounded by the Carnegie
Ridge to the north and the Grijalva Fracture Zone to the
south.
Then, we will construct a further larger magnitude
scenario that breaks the segmentation hypothesis. We
believe that the parameter that differentiates the two sets of
scenarios (segmented vs. unsegmented hypothesis) is the
earthquake return period. The segmented scenarios rely on
the recent observed earthquakes (during approximately the
last century), whereas the unsegmented scenario (of larger
magnitude) is based on unregistered potential events
(approximately during the last half-millennium).
Building possibly-tsunamigenic local earthquakes
scenarios within the Gulf of Guayaquil
Segmented earthquake scenarios (short return period)
We propose here characteristic earthquake scenarios rep-
resentative of the local and registered seismicity. We
consider two interplate earthquake scenarios of magnitude
M
w
=7.0 and M
w
=7.5 that embrace the magnitudes of
the three historical events of 1901, 1933 and 1953. We
speculate that the two scenarios are sufficiently represen-
tative in terms of magnitude with a minimum M
w
=7.0
that could possibly trigger an observable and significant
tsunami, and a realistic M
w
=7.5 earthquake. We specu-
late that the study of these two segmented scenarios is
sufficient in terms of tsunamigenesis in the case we are able
to well constrain them; we will dedicate a special effort to
include subsurface data (the interplate geometry) and their
analysis (described later). As a result, we do not build
further ensembles of scenarios (and computing- and
memory-costly simulations) that would necessarily require
constructing a statistical model to assess tsunami mapping.
By proposing to model the tsunami effects that could likely
be produced by two realistic earthquake scenarios based on
local historical earthquakes of M
w
=7.0–7.5, we place
ourselves in a deterministic point of view.
Most of the megathrust subduction events along the Nazca
Plate generate a tsunami, which can be very destructive as
related with the 1877 and 1867 great earthquakes of southern
Peru-northern Chile (both of magnitude M
w
=8.8), the 1960
Valdivia mega-thrust event (M
w
=9.5, the greatest earth-
quake ever recorded) and more recently, the 2010 Maule large
event (M
w
=8.8) (Contreras-Reyes and Carrizo 2011)or
closer to Ecuador, the above-cited 1906 Ecuador-Colombia
megathrust event of magnitude M
w
=8.8 (Kelleher 1972;
Kanamori and McNally 1982). In the Grijalva-Mendan
˜a
segment in southern Ecuador-northern Peru (2°–10°S), a few
reported historical events of magnitude between 7.5 and 8 for
the last 500 years have been reported (Carena 2011). How-
ever, the absence of great events may be interpreted either by
creeping where low seismic energy is released regularly or by
a locked zone ready to break into a sudden high seismic energy
release (Kelleher 1972).
In order to estimate potential tsunamigenic earthquakes
in the region of the GG on the northern edge of the Gri-
jalva-Mendan
˜a segment, we should estimate the width and
length of the rupture of such events. The space–time
behavior of seismicity patterns was one of the tools
Mar Geophys Res
123
Author's personal copy

commonly used by seismologists to identify zones of future
large earthquakes, i.e., their rupture sizes (length and
width). These zones corresponded to places where the
background seismicity varies with time either as seismic
gap, seismic quiescence, foreshock activity, precursory
seismic swarm or in a doughnut pattern (Kanamori 1981).
However, from nowadays, these patterns are not widely
used for estimating future rupture dimensions because they
vary strongly from event to event (Kanamori 1981).
The greatest-magnitude tsunamigenic earthquake recor-
ded in the study zone occurred in 1953. Recent tsunamigenic
megathrust earthquakes (2004 Sumatra earthquake; 2010
Maule, Chilean earthquake; 2011 Tohoku-Oki, Japanese
earthquake) generated trans-oceanic tsunami waves. Their
coseismic ruptures exceeded 500 km length and the updip
limit of their ruptures is at shallow depths (Ammon et al.
2005; Chlieh et al. 2007; Simons et al. 2011; Lay et al. 2011).
However, Ioualalen et al. (2011) observed fresh fractures on
the sea bottom related to the major Tohoku 2011 event,
attesting clearly that the coseismic rupture reached the sea-
floor. Generally, the updip limit of a rupture plane is a dif-
ficult parameter to estimate, in particular before the rupture
slip, and varies from one place to another. In central Chile,
Moscoso et al. (2011) estimated the uppermost (shallower)
dip limit of the Maule 2010 event to be 15 km depth.
Although wide-angle seismic data cannot define directly the
updip limit of a rupture zone, they can provide valuable
information on the rocks of the backstop in a region that is
shown by other methods to be the updip limit. In the GG,
Calahorrano et al. (2008) estimated the uppermost limit of
the backstop to 5–6 km depth (Fig. 3).
We estimate the slip displacement D, the length Land
the width Wof the rupture plane for interface subduction-
zone earthquakes, following scaling laws of Blaser et al.
(2010) and Strasser et al. (2010) (their length unit trans-
posed from km to m), with Land Wbeing functions of
magnitude M
w
and defined as:
log10ðLÞ¼0:57Mwþ0:63;ð1Þ
and
log10ðWÞ¼0:46Mwþ1:14:ð2Þ
According to Hanks and Kanamori (1979), M
w
is related
to the seismic moment M0¼lLWD(Dbeing the slip
displacement and M
0
transposed from Dyne cm into N m
units) as:
Mw¼2
3log10ðM0Þ6:03:ð3Þ
We obtain for D:
log10ðDÞ¼0:47Mwþ7:28 log10 ðlÞ:ð4Þ
Considering a typical medium shear modulus of l¼
3:31010 Pa, the slip displacement (in meters) becomes:
log10ðDÞ¼0:47Mw3:24:ð5Þ
Thus, for a given magnitude M
w
, we are able to constrain
the rupture parameters L,Wand Dalong the interfacing
subduction zone(Table 1).
Fig. 3 Seismic profile S-18 sampled during research cruise SIS-
TEUR operated by Ge
´oazur in 2000 and processed in Calahorrano
(2005) and Calahorrano et al. (2008). The vertical axis is depth (in
km), yellow ellipses correspond to the hypocenter locations for
scenarios M
w
=7.0 and 7.5 (Table 1). The subduction channel is
located with the red curve. The profile track is located in red at the
lower left corner
Mar Geophys Res
123
Author's personal copy

To solve the Okada (1985) formulation we need further
constraints, i.e., the level of coupling of the area, the
earthquake scenarios’ centroids, their focal depths and the
interplate dipping angles, the ruptures being assumed as
pure thrusts (rake angle is set to k=90°) and the fault
strike direction being aligned with the trench (/=10°).
We assume here pure subduction earthquake scenarios for
which their hypocenters and directions of propagation are
located within the subduction channel (along and across
strike). Fortunately, we have at our disposal a seismic
profile (track SIS-18), very characteristic of the area, that
has been sampled north of the Grijalva Fracture Zone
during the SISTEUR survey operated in the year 2000 by
Ge
´oazur (Fig. 3; Collot et al. 2002). The N110°E profile is
derived from a 109.3 km deep multichannel seismic (MCS)
reflection and refraction track data that have been pro-
cessed through prestack depth migration (PSDM) by Cal-
ahorrano (2005) and Calahorrano et al. (2008) in order to
obtain a depth-image (Fig. 3). We have placed the hypo-
centers along the subduction channel with respect to the
width of the rupture plane (Fig. 3). The ultimate question
concerns the positioning of the hypocenters in these sce-
narios along the interplate in the SIS-18 seismic profile that
should provide us with their respective focal depths and
mean dip angles (Table 1).
For that purpose we have used a GPS inversion proce-
dure in order to locate horizontally possible coupling areas.
We have used a combination of continuous and campaign
GPS data installed since 1994 and expressed in a stable
North Andean Block (NAB) reference frame (from Valle
´e
et al. 2013) to determine the interplate coupling along the
southern Ecuador subduction zone. We used a back-slip
approach (Savage 1983) and the analytical formulation of
Okada (1992) to perform non-linear inversions based on a
stochastic simulated annealing algorithm (Chlieh et al.
2011). The megathrust geometry follows the trench axis
and a precise 3D geometry (Font et al. 2013). The slab
interface is meshed into source elements of 20 920 km
embedded in a elastic half-space. The maximum slip deficit
rate is bounded by the relative Nazca/NAB plate conver-
gence rate of 46 mm year
-1
and the rake is fixed to the
average slip vector of megathrust CMT solutions found in
the Harvard catalog. Geodetic inversions are performed
driven by the minimization of a cost function that is a
weighted summation of the RMS fit to the GPS data and a
smoothing factor that homogenises the slip distribution:
Cost ¼rmsðÞ
2þkDcðÞ
2ð6Þ
where Dc is the average difference of coupling/decoupling
between adjacent cells. The coefficient kcontrols the
smoothness of the solution. The interseismic coupling is
defined as the ratio of the slip deficit rate and the plate
convergence rate. Consequently, an interplate coupling of 1
corresponds to full locking while an interplate coupling of
0 corresponds to creeping at the plate convergence rate.
global positioning system (GPS)-derived inversions
indicate a high heterogeneity along the South Ecuadorian
subduction zone (Fig. 4). Offshore Salinas, the interplate
coupling is confined in the shallow portion of the slab
interface at \12 km depth. That shallow coupling extends
southward to 3°S. The shallow portion of the slab interface
could be then taken as a potential region for a M
w
=7.0–7.5
earthquake scenario. The absence of deep coupling excludes
that this source could be deeper than 12 km. As a direct and
crucial consequence, the lack of deep coupling down the
slab necessarily excludes a large rupture plan width and
therefore a possibly large rupture plan length if we take into
account the above scaling laws. Such scaling consideration
excludes a large slip and finally a M
w
[8earthquake.
However a M
w
=8 scenario is possible if we ignore the
scaling laws. This is possible because, precisely, for this
magnitude we place ourselves outside the domain of validity
of the laws (scenario in the next section).
Besides, and importantly, the weak coupling (0.2–0.3 in
Fig. 4) within the 0–12 km coupled area also indicates that
most of the subduction rate is controlled by creeping. The
Table 1 (Okada (1985)’s input parameters for the three earthquake
scenarios of magnitude M
w
=7.04, 7.53 and 8.00 (S7.0 and S7.5 and
S8.0). S7.0M, S7.5M and S8.0M are identical but computed for high
tide (?90 cm relative to mean sea level): Longitude and latitude of
segment centroid (x
0
,y
0
); the centroid depth d(related to the ocean
bottom while Fig. 4indicates depths relative to the earth level ref-
erence, i.e., including ocean depth), the fault strike angle /(clock-
wise from North); the fault rake angle k(counterclockwise from
strike); the fault dip angle d(dip counted clockwise from the hori-
zontal plane); the maximum fault slip D; The segment lengths along
and across strike (L,W); the medium shear modulus lis taken equal
to 3.3 910
10
Pa
Parameters S7.0, S7.0M S7.5, S7.5M S8.0, S8.0M
x
0
81.1579°W 81.0868°W 81.2578°W
y
0
2.7764°S 2.7868°S 2.7605°S
d(km) 5 6 4
/10°10°10°
k90°90°90°
d6.5°6.5°6.5°
D(m) 1.12 1.93 5
L(km) 42 80 150
W(km) 23 39 39
M
0
(J) 3.60 910
19
2.00 910
20
9.67 910
20
k
0
(km) 23 39 39
s
0
(s) 225 479 308
g
0
(m) -0.24, ?0.37 -0.43, ?0.64 -1.11, ?1.63
M
0
is the seismic moment. The three last lines are model outputs: the
characteristic wave length and period (k
0
and s
0
) and the initial trough
and crest amplitudes (g
0
)
Mar Geophys Res
123
Author's personal copy

residual (coupling) is coherent with.the M
w
=7 and 7.5
scenarios for a centennial return period. A further M
w
=8
scenario would apply for earthquakes of larger return
period (half a millennium, next section). These conclusions
are in agreement with the fact that no large earthquakes are
mentioned within the GG area in the recent period.
From Fig. 4, we have located the area to be ruptured for
our two segmented scenarios as well as the position of the
centroids which, in turn, prescribe the focal depths and
dipping angles obtained from the profile SIS-18 (Table 1).
Unsegmented earthquake scenario (large return period)
Here, we propose a larger magnitude scenario for the GG
area. Similarly to the previous segmented scenarios, we
have placed our ruptured area within the coupled area
(Fig. 4). We have however extended the ruptured area over
the southern borders of the Carnegie Ridge and the Grijalva
Fracture Zone (Fig. 4; Table 1). Since the coupling does
not exceed 0.2–0.3, the 5 m slip we used (Table 1) repre-
sents an approximate half-millennial return period. We
believe it is not reasonable to extend such amount of slip
(and the return period) in the absence of any paleo-tsunami
record data to validate it. Similarly to the 2011 Tohoku-Oki
earthquake/tsunami event, we have placed the updip of our
ruptured area very close to the trench so that we could
expect a maximum of seafloor deformation (and thus of the
initial tsunami wave amplitude), but also a larger tsunami
signature at the coast by enhancing the potential 1/4–
powered Green’s law.
The other fundamental issue concerns the necessity or
not to extend laterally (southward and northward) our
M
w
=8.0 scenario, considering that North/Central Ecua-
dor and North Peru host strong (but limited) asperities.
Historical North/Central Ecuador earthquakes never rup-
tured across the Carnegie Ridge because it is a 50 km
decoupling area acting as a barrier to the southward
propagation of large earthquakes like the 1906 one (Noc-
quet et al. 2014; Collot et al. 2002). Recent GPS mea-
surements indicate the same characteristics for North Peru,
which hosts similar barriers that strongly reduce a potential
rupture in a cascade of multiple asperities (Nocquet et al.
2014). Ultimately, the sizes of the asperities of North Peru
and Central/North Ecuador are much smaller than those of
Japan, Sumatra and Chile, thus excluding a M
w
=9 mag-
nitude there (Nocquet et al. 2014).
Considering the points raised above, we consider that
our M
w
=8.0 scenario is an extreme scenario indeed,
especially bearing in mind that there is no record of such
magnitude in the GG.
The numerical procedure and the computational
domain
Once we have built a representative range of realistic
earthquake scenarios we need to construct robust numerical
simulations for deriving tsunami runup and flooding maps.
Fig. 4 Interseismic coupling along the Southern Ecuadorian subduc-
tion zone derived from GPS inversion. Iso-contours of the slab
interface (dashed lines) are reported at 5-km depth intervals. The
interseismic coupling model indicates that from Salinas and along the
Gulf of Guayaquil, the interplate coupling appears to be relatively low
(about 0.2–0.30) and confined on the shallow portion (\12-km depth)
of the slab interface. This supports the potential for a shallow large
seismic event. The computed vertical seafloor displacements pro-
duced by the aM
w
=7.0, bM
w
=7.5 and cM
w
=8.0 earthquakes
are reported every 10 cm (see Figs. 6,7,11 for details)
Mar Geophys Res
123
Author's personal copy

For that purpose, we need accurate ocean bathymetry and
coastal topography data sets that will be used for the
construction of the computational domain and a well-val-
idated tsunami propagation and runup model (Ioualalen
et al. 2007). The issue is particularly crucial within the GG
where numerous rivers and estuaries connect to each other.
An misrepresented connection, e.g., an artificial river ter-
mination, may eventually provoke a tsunami buildup and
force a local seiche resonance. An accurate digitized
topographic data set of 30 m spatial resolution has been
used (Souris 2002) (Fig. 5). For the bathymetry, accurate
and digitized marine charts have been provided by INO-
CAR (Fig. 5). Elsewhere ETOPO-2 20-grid bathymetry has
been considered (ETOPO-2 2001). The computational
domain, extending from 81.83°W to 79.61°W in longitude
and from 3.5°S to 1.8°S in latitude, covers the entire GG
and most of the significant vertical co-seismic displace-
ments (Figs. 6,7). A 100 m uniform grid spacing has been
taken for both horizontal directions (2,376 91,980 grid
points) along with a 0.125 s time step to avoid any
numerical instability. A total simulation of 6 h has been
performed (172,800 time steps) in order to allow the first
wave crests to reach any single point of the computational
domain, in particular the farthest river points to the
northeast of the domain. Finally, Funwave, a fully non-
linear Boussinesq (dispersive) tsunami propagation/runup
model has been used (Wei and Kirby 1995; Wei et al.
1995). The model has been throughoutly validated for
different case studies (Ioualalen et al. 2007,2012).
Numerical results
With our set of local segmented earthquake scenarios
M
w
=7.0 and 7.5, we believe we have a comprehensive set
of scenarios that are likely to be representative of the tsu-
nami threat for the GG at a centenary return period. For
both scenarios, our simulations do not indicate any inun-
dation or significant coastal wave heights over the entire
coastal domain for a zero mean sea level. For M
w
=7.5 the
tsunami wave height never exceeds 1–2 m along the entire
computed coast (Fig. 6) (even weaker for M
w
=7.0, not
shown). Considering our 100 m grid resolution and that the
land topographic slope is always larger than 2/100, no
flooding operates. Nearly all along the coast, there is no
process, e.g., focusing, that would be likely to operate a
significant wave amplification of the initial 64 cm wave
(Table 1). The Santa Elena Peninsula is an exception
because there is convex-shaped bathymetry there (see the
100 m isobath in Fig. 6) that is likely to trigger a focusing,
but it is not sufficient: the wave does not overpass the local
5/100 topographic slope. Rather, the peninsula acts as a
diffractor that hosts a shadow zone to the north of the
peninsula (Fig. 6). Offshore Chanduy and Playas, the
bathymetry exhibits wave attenuation processes (concave-
type 100 m isobath in Fig. 6) that could eventually limit
the wave height at the coast. The Green’s slope effect
operates everywhere (with the exception of north of Santa
Elena) but the amplification is also not sufficient to over-
pass the beach slope. Then, considering our interplate
constraints, in particular the positioning and azimuth of the
ruptured areas, no significant wave energy enters the GG
through the eastern side of Puna Island (the most opened
side) (Fig. 6). The wave impact at the GG interior could
possibly be more significant if both rupture zone scenarios
were located more southward. Then, in such a case, we
could expect that the very complex coastal geometry and
the presence of a wide range of estuary extensions could
potentially host a wide frequency spectrum of bay reso-
nances. However, such a scenario would not be coherent
with our seismic constraints. In our scenarios, the only path
allowing the wave entrance to the GG is located to the west
of Puna Island and it is very much confined. Consequently,
only a small part of the wave energy could be redistributed
within the GG interior; the inner part of the GG appears
sheltered. The latter issue is crucial because the interior of
the GG hosts aquaculture farms and fisheries and of course
the densely populated city of Guayaquil.
Although the scenarios are based on relatively strong
earthquakes, there is no (even very) local tsunami impact.
Fig. 5 Data sets used for the construction of the computational
domain: the land topography is obtained from the 30 m resolution
grid of Souris (2002)(brown) and from the SRTM 90 m resolution for
Northern Peru (grey). The bathymetry is derived from digitized
marine charts obtained from INOCAR (blue and green for high
resolution charts). In red is the INOCAR coastline and ETOPO-2
completes the bathymetric field offshore (black dots)
Mar Geophys Res
123
Author's personal copy

The impact of a tsunami depends on various parameters:
(1) the earthquake magnitude and focal mechanisms
(parameters of Table 1), and (2) the local (or large-scale)
wave amplification factors (focusing, shoaling, and bay
resonance mainly). Here, although our scenarios are based
on relatively strong earthquakes, there is no (even very)
Fig. 6 Simulation S7.5
(Table 1): Runup map for the
Gulf of Guayaquil along with
the initial wave (red lines
represent uplift and blue lines
represent subsidence, both at
0.2 m contour intervals). The
runup scale is voluntarily
limited to 2 m for the regional
map (top) for a better
identification of sheltered and
exposed areas and for an easier
comparison with the other
simulation (it can eventually
exceed this value greatly). The
background bathymetry is
plotted in grey at 100 m contour
intervals. The white star
represents the fault centroid
Fig. 7 Same as Fig. 6for
simulation S7.5 M (Table 1).
Areas in rectangles (a) and
(b) are enlarged in Figs. 8and 9
Mar Geophys Res
123
Author's personal copy

local tsunami impact. The bathymetric and topographic
shapes do not encourage specific wave amplification.
Regarding point (2), these are not parameters of control.
The only option we have is to evaluate the tsunami sig-
nature at different tides. For our scenarios, the main limi-
tation of the wave impact is the inability of the wave to
overpass the local coastal topographic slope and inundate
further inland. The local beach topography along the con-
sidered zone is often reduced to a positive beach slope
followed in land by a negative slope (area of Santa Elena,
South of Playas). Consequently, there is a possiblity that at
high tide the wave overpasses the dunes and inundates the
inland regions of lower altitude. This ‘dunes’ barrier appears
critical. In our computations we treat dunes as rigid bodies
while they are not always rigid in reality. A simulated wave
could be eventually damped through absorption by sediment.
On the other hand, the long tsunami wave (of nearly hori-
zontal uniform velocity) provokes sediment extraction,
possibly strengthening the water energy contributing to an
easier overpassing of the dunes. Besides, the dunes are
Fig. 8 Enlargment of area
(a) of Fig. 7for S7.5M (region
of Chanduy). The 2 and 4 m
land topographic contours are
highlighted in white. The
associated satellite image is also
reported (Canduy is Chanduy)
(Courtesy of Google Earth,
Inc.). The computed inundated
area is reported in red in the
image where the green parcels
are shrimp farms. The red arrow
corresponds to the wave
entrance path. Here the 2 and
4 m contours are highlighted in
white
Mar Geophys Res
123
Author's personal copy

constantly moving and their migration may make potential
tsunami passages. Here we will not treat these problems
since we focus on rigid bodies only, but we will try to provide
estimates of tsunami inundation for a varying tide.
Simulations S7.0 and S7.5 were performed for a zero
mean sea level. We submit again the same simulations but
at high tide. On average, in the GG area, the two highest
tide are on the order of 90 cm above mean sea level. We
then take this value and apply a (uniform) static sea level
rise on our bathymetric file (Simulations S7.0M and S7.5M
in Table 1). We take into account neither the tidal variation
in space and time nor tsunami/tide interactions because
those are beyond the scope of the study (Kowalik et al.
2006; Kowalik and Proshutinsky 2010). Simulation S7.5M
indicates two identified threatened areas (Fig. 7): The area
of Chanduy (a fishing village) (Figs. 7,8) with an impor-
tant runup and inundation, and the area between Playas and
El Arenal with an important inundation and a weak runup
(Figs. 7,9). As for S7.0, simulation S7.0M does not exhibit
any inundation (not shown).
The first wave peak arrives at Chanduy area 45 min after
the earthquake occurrence (Fig. 10). Here, the wave is
Fig. 9 Same as Fig. 8for
Playas area (b) of Fig. 7
Mar Geophys Res
123
Author's personal copy

trapped during 1 h and radiates within the estuary. The
successive ongoing depressions (wave troughs) of the main
(offshore) tsunami do not enter the estuary because their
wave periods are different from the those of the inner
trapped wave radiating at very shallow water. The first
effective depression entering the estuary occurs 3 h after
the earthquake. Then a water leakage operates within the
estuary for 1 h. The runup reaches 4–4.5 m within the
estuary which may cause human and marine equipment
damage because Chanduy is a very active (pure) fishery
harbor. Besides, most of the shrimp farms are flooded
(Fig. 8) which would be a severe economic issue.
It takes 70 min for the first wave to reach the Playas area
(Fig. 10). The area is composed of dunes at the beach
followed by a land depression. Thus, most of the time, the
tsunami cannot overpass the dunes. However, there are two
locations of low land topographic gradient where a passage
of the wave is possible (channels A and B in Fig. 9). Then,
as for the Chanduy area, the wave radiates in the interior.
Similarly to Chanduy, the water wave is trapped because
only the very narrow passages A and B allow an offshore
water leakage. During the inundation, water spreads out
from A and B within the land interior, explaining why the
water elevation is very low (\20 cm, Fig. 9). Conse-
quently, there is no real threat to human damages consid-
ering the weakness of the wave height. However, the
simulation exhibits an extended inundation, mainly over
the shrimp farms (Fig. 9). It also indicates that it could be
easily avoided if the basins are surrounded by raised paths
of at most 10 s of centimeters high. This is actually the
case because the farms are already protected against other
(climatic) flooding. Our simulation shows that the existing
protection is sufficient indeed for tsunami issues.
These prospective inundation maps are difficult to val-
idate with observations of former earthquakes/tsunamis.
The segmented scenarios generate very weak runup at the
region of Salinas/La Libertad, i.e., the northern side of the
Sta Elena Peninsula (a few cm in Fig. 7). Observations
mention a runup of at most 2.5 m for the M
s
*6.9 1933
October 2 earthquake (Espinoza 1992) while a much
weaker runup (approximately 0.2 m) is reported for the
possibly much stronger M
s
*7.5 1959 February 07
(Lockridge 1984). Although the absolute positioning of the
two events are not clear, the 1959 event is estimated at
lower latitude (4°s vs. 3.5°S for 1933). We interpret this as
follows: if an earthquake occurs south of the Sta Elena
Peninsula (e.g., possibly the 1959 event), the peninsula acts
as a wave diffractor for the northern side of the peninsula
(see above and Fig. 7), generating a sheltered shadow zone.
To the contrary, if the earthquake occurs further north (e.g.,
the 1933 event), the peninsula (including its northern side)
is a wave focusing area (wave amplification). Consequently
it should not be surprising that the relatively weak 1959
earthquake triggered a larger tsunami at Salinas than did
the stronger 1933 earthquake. Our segmented scenarios are
positioned more like the 1959 event, thus the few simulated
cm at Salinas/La Libertad are more in agreement with the
20 cm mentioned for the 1933 event. The above results
apply for tsunamigenic earthquakes of return period on the
order of a century.
For larger return periods (approximately half a mille-
nium), the unsegmented M
w
=8.0 scenario (without tide)
does not exhibit a drastic change compared to segmented
ones except that the entire coast from Sta-Elena Peninsula
to the entrance of the GG is subject to an approximately
2 m runup without, however, any significant inundation
(Fig. 11). Similarly, Isla Puna is an efficient barrier against
the tsunami propagation within the GG interior, i.e., the
city of Guayaquil is sheltered. The area of Chanduy and its
estuary are inundated but at a lower level compared to the
M
w
=7.5 scenario at high tide (Fig. 11a) as well as for the
area south of Playas, where the tsunami wave does not
overpass the barrier of sand dunes (Fig. 11b). Only the
uninhabited and unexploited area south of Engabao
encounter significant runup (approximately 5 m) and land
inundation.
The passage from zero mean sea level to high tide (from
scenario S8.0 to scenario S8.0M in Table 1) exhibits
Fig. 10 Snapshots for simulation S7.5M (Table 1) at times t =0, 45 and 70 min after earthquake occurrence. Here we are not concerned with
the wave amplitudes. The light blue color represents wave crests while dark blue is wave troughs (depression)
Mar Geophys Res
123
Author's personal copy

similarities with the one from S7.5 to S7.5M (Fig. 12): The
area south of Playas becomes inundated. The inundation
zone is slightly more extended, however, and the water
level is on the order of tens of centimeters (compared to
centimeters for S7.5M). Moreover, the inundated area of
Chanduy is similar while comparing S7.5M and S8.0M.
Fig. 11 Simulation S8.0
(Table 1): runup map for the
Gulf of Guayaquil along with
the initial wave (red lines
represent uplift and blue lines
represent subsidence, both at
0.2 m contour intervals). Here,
the runup scale is limited to
4 m. The background
bathymetry is plotted in grey at
100 m contour intervals. The
white star represents the fault
centroid. Areas in rectangles
(a) and (b) are enlarged
Mar Geophys Res
123
Author's personal copy

Besides, the tide does not affect the inundated area south of
Engabao significantly, i.e., the inundation is comparable
with the one of the zero mean sea level case study (S8.0).
Consequently, with the exception of this latter area, the
tsunami signature of the S8.0M unsegmented scenario is
comparable to the one of the segmented S7.5M.
Fig. 12 Same as Fig. 11 for
simulation S8.0M (Table 1)
Mar Geophys Res
123
Author's personal copy

Conclusion
The North-Andean subduction block hosts a segmented
seismicity due to a very specific morphological structure.
North of 0°–0.5°N, the North Ecuador-South Colombia
segment generates powerful tsunamigenic earthquakes due
to the subduction of the Nazca plate beneath the North-
Andean subduction block. The central zone in front of the
Carnegie Ridge is characterized by moderate seismicity.
Southward, between the Carnegie Ridge and the Grijalva
Fracture Zone, significant earthquakes (M
w
[6.9) occur-
red during the last century, e.g., 1901 and 1933, which
were possibly tsunamigenic. Within the Gulf of Guayaquil,
the seismicity is instead dominated by relatively large focal
depths ([20 km; Fig. 4) for subduction earthquakes or
local strike-slip kinematics that are both assumed to be less
tsunamigenic than those occurring near the trench. Our
study concerns the tsunami mapping of the Gulf of
Guayaquil for relatively short wave arrival time, i.e.,
shorter than an hour. The reasons follow: (1) According to
the NGDC data base, there is no record of tsunami impact
on the GG coastal due to distant earthquakes, i.e., propa-
gating from North Peru or Chile to the south, from North
Ecuador- South Colombia to the north, nor for trans-oce-
anic tsunamis; (2) the PTWC warning system has proved
its efficiency in Ecuador for arrival times larger than 1–2 h.
Consequently, we are interested in tsunamis that could
possibly be triggered within the southern area located
between the southern border of the Carnegie Ridge and the
Grijalva Fracture Zone. Considering the relative confine-
ment of area, the largest earthquake magnitude that we
could obtain with the use of pertinent seismic scaling laws
is at most M
w
=7.5. Rupture parameters are deduced from
the subsurface morphology (interplate geometry) of the
area, the use of appropriate scaling laws and a GPS
inversion procedure. A further available seismic profile
allowed us to derive coherent focal depths and dipping
angles, thus considerably reducing the number of scenarios.
At the end, only two representative earthquake scenarios
have been retained, a M
w
=7.0 and a M
w
=7.5, yielding a
deterministic tsunami mapping. The two scenarios have
been computed at mean and high tide levels.
It is fair to say that a comprehensive tsunami mapping of
the area would require paleo-tsunami data for many rea-
sons. Among them: (1) The lack of observations concern-
ing the 1901, 1933, 1953 and 1959 events makes it difficult
to validate our results; (2) paleo-records would place our
mapping into the context of much larger time scales and
eventually could provide us with information concerning
return periods, and consequently further scenario magni-
tudes. The analysis of our GPS coupling maps indicate that
our M
w
=7.0 and M
w
=7.5 scenarios correspond to return
periods on the order of a century.
In order to increase our mapping validity we further
processed a M
w
=8.0 scenario by ignoring the above
segmentation assumption. We found that the M
w
=8.0
tsunami signature is quite similar to the one of the
M
w
=7.5. Considering the weak interplate coupling in the
area, such a scenario corresponds to a return period of
approximately half a millennium. We certainly could
process an even more extreme scenario of much larger
return period: however, we believe that without any paleo-
tsunami record verification, it is not reasonable to extend
indefinitely those time scales for a purely academic goal.
Besides, recent GPS measurements invalidate the unseg-
mented assumption that could eventually yield the prop-
agation of seismic rupture from North Peru or North
Ecuador—South Colombia into the Gulf of Guayaquil
(Nocquet et al. 2014). Our mapping concerns return period
of approximately 0–500 years. Beyond such time scales
we believe that it would be more reasonable to process
first a paleo-tsunami study (core sampling and analysis)
and then (and only then) simulate larger scenarios if
necessary. In particular the area of Chanduy might be
inundated severely. Past a certain earthquake magnitude,
the tsunami wave might overpass the 4 m topographic
threshold and then inundate the entire area to the north-
west of the Sta Elena Peninsula. A prospective simulation
has been processed (a M
w
=8.5) but it is not shown here
in the absence of any paleo-tsunami record or any evi-
dence of possible unsegmented events. Informal discus-
sions with Ecuadorian paleontologists suggest that they
did not observe any trace of paleo-tsunami fossils in that
particular area, some of their analyzed cores (extracted for
other purposes and thus unreferentiable) representing
several millennia of sedimentary stratification. Again, we
want to stress that the simulation of long return period
earthquakes/tsunamis in the GG (larger than a millen-
nium), could be processed only in conjunction with an
effective paleo-tsunami analysis.
Acknowledgments The authors wish to acknowledge with thanks
the Editor Amy Draut and Dr. George Pararas-Carayannis along with
two anonymous reviewers for their efforts and comments which
contributed to a substantial improvement of the first draft manuscript.
We acknowledge the support of the French Agence Natonale pour la
Recherche, ANR, under the project REMAKE.
References
Ammon CJ, Ji C, Thio H-K, Robinson D, Ni S, Hjorleifsdottir V,
Kanamori H, Lay T, Das S, Helmberger D, Ichinose G, Polet J,
Wald D (2005) Scaling relations of earthquake source parameter
estimates with special focus on subduction environment, rupture
process of the 2004 Sumatra-Andaman Earthquake. Science 308
(5725)1133–1139. ISSN 0036-8075
Beck S, Ruff L (1989) Great earthquake and subduction along the
Peru trench. Phys Earth Planet Inter 57(3–4):199–224
Mar Geophys Res
123
Author's personal copy

Blaser L, Kru
¨ger F, Ohrnberger M, Scherbaum M (2010) Scaling
relations of earthquake source parameter estimates with special
focus on subduction environment. Bull Seismol Soc Am
100:2914–2916
Calahorrano A (2005) Structure de la Marge du Golfe de Guayaquil
(Equateur) et proprie
´te
´s physiques du chenal de subduction, a
`
partir des donne
´s de sismique marine, re
´flexion et re
´fraction.
Ph.D document, Universite
´Pierre et Marie Curie-Paris 6
Calahorrano A, Sallares V, Collot J-Y, Sage F, Ranero CR (2008)
Nonlinear variations of the physical properties along the
southern Ecuador subduction channel: results from depth-
migrated seismic data. Earth Planet Sci Lett 267:453–467
Carena S (2011) Subducting-plate topography and nucleation of great
and giant earthquakes along the south american trench. Seismol
Res Lett 82(5). doi:10.1785/gssrl.82.5.629
Chlieh M, Avouac J-P, Hjorleifsdottir V, Song T-RA, Ji C, Sieh K,
Sladen A, Hebert H, Prawirodirdjo L, Bock Y, Galetzka J (2007)
Coseismic and afterslip of the great Mw 9.15 Sumatra-Andaman
earthquake of 2004. Bull Seismol Soc Am 97:S152–S173
Chlieh M, Perfettini H, Tavera H, Avouac J-P, Remy D, Nocquet J-M,
Rolandone F et al (2011) Interseismic coupling and seismic
potential along the Central Andes subduction zone. J Geophys
Res 116:B12405. doi:10.1029/2010JB008166
Collot J-Y, Charvis P, Gutscher M-A, Operto S (2002) Exploring the
Ecuador-Colombia active Margin and interplate seismogenic
zone. Eos Trans Am Geophys. Union 83(17), 185, 189–190
Collot J-Y, Michaud F, Legonidec Y, Calahorrano A, Sage F,
Alvarado A, y el personal cientifico y te
´cnico del INOCAR
(2006) Mapas del margen continental centro y sur de Ecuador:
Batimetria, relieve, reflectividad acu
´stica e interpretacio
´n geol-
o
´gica. Publicacio
´n IOA-CVM-04-POST
Collot J-Y, Michaud F, Alvarado A, Marcaillou B, Sosson M, Ratzov
G, Migeon S, Calahorrano A, Pazmino A (2009) Visio
´n general
de la morfologı
´a submarina del margen convergente de Ecuador-
Sur de Colombia: implicaciones sobre la transferencia de masa y
la edad de la subduccio
´n de la Cordillera de Carnegie. In: Collot
J-Y, V. Sallares, and A. Pazmin
˜o, Publicacion CNDM-INO-
CAR-IRD, PSE001-09 (eds) Geologia y Geofisica Marina y
Terestre del Ecuador desde la costa continental hasta las islas
Gapapagos, Guayaquil, Ecuador, pp 47–74
Contreras-Reyes E, Carrizo D (2011) Control of high oceanic features
and subduction channel on earthquake ruptures along the Chile-
Peru subduction zone. Phys Earth Planet Inter 186:49–58.
doi:10.1016/j.pepi.2011.03.002
Dumont J-F, Santana E, Vilema W, Pedoja K, Ordonez M, Cruz M,
Jimenez N, Zambrano I (2005) Morphological and microtectonic
analysis of quaternary deformation from Puna and Santa Clara
Islands, Gulf of Guyaquil, Ecuador (South America). Tectono-
physics 339:331–350
Ego F, Sebrier M, Lavenu A, Yepes H, Eguez A (1996) Quaternary
state of stress in the northern Andes and the restraining bend
model for the Ecuadorian Andes. Tectonophysics 259:101–116
Espinoza J (1992) Terremotos tsunamige
´nicos en el Ecuador, Acta
Oceanografica del Pacifico. INOCAR 25, 66–82. Geologica del
Peru, 32(2), 225–238. http://hdl.handle.net/1834/2181
ETOPO-2 (2001) (World Data Center for Marine Geology and
Geophysics. 2-minute gridded global relief data) (2001).
National Geophysical Data Center (NGDC), NOAA Satellite
and Information Service. http://www.ngdc.noaa.gov/mgg/fliers/
01mgg04.html
Font Y, Segovia M, Vaca S, Theunissen T (2013) Seismicity patterns
along the Ecuadorian subduction zone: new constraints from
earthquake location in a 3-D a priori velocity model. Geophys.
J Int doi: 10.1093/gji/ggs1083
Graindorge D, Collahorrano A, Charvis P, Collot J-Y, Be
´thoux N
(2004) Deep structure of the Ecuador convergent margin and the
carnegie ridge, possible consequence on great earthquakes
recurrence interval. Geophys Res Lett 31:L04603
Gutscher M-A, Malavieille J, Lallemand S, Collot J-Y (1999)
Tectonic segmentation of the North Andean margin: impact of
the carnegie ridge collision. Earth Planet Sci Lett 168:255–270
Hanks T, Kanamori H (1979) A moment magnitude scale. J Geophys
Res 84(B5):2348–2350
Ioualalen M, Asavanant J, Kaewbanjak N, Grilli ST, Kirby JT, Watts
P (2007) Modeling the 26 december 2004 Indian Ocean tsunami:
case study of impact in Thailand. J Geophys Res 112:C07024.
doi:10.1029/2006JC003850
Ioualalen M, Ratzov G, Collot J-Y, Sanclemente E (2011) The
tsunami signature on a submerged promontory: the case study of
the Atacames Promontory. Ecuador Geophys J Int
184(2):680–688
Ioualalen M, Perfettini H, Yauri Condo S, Jimenez C, Tavera H
(2012) Tsunami modeling to validate slip models of the 2007
Mw8.0 Pisco earthquake, Central Peru. Pure Appl Geophys.
doi:10.1007/s00024-012-0608-z
Kanamori H (1981) The nature of seismicity patterns before large
earthquakes. In: Ewing M (ed) Series 4: earthquake prediction—
an international review, AGU geophys. Washington D.C, Mono.,
pp 1–19
Kanamori H, McNally KC (1982) Variable rupture mode of the
subduction zone along the Ecuador-Colombia coast. Bull
Seismol Soc Am 72:1241–1253
Kelleher J (1972) Rupture zones of large South American earthquakes
and some predictions. J Geophys Res 77:2087–2103
Kowalik Z, Proshutinsky A (2010) Tsunami tide interactions: a cook
inlet case study. Cont Shelf Res 30:633–642
Kowalik Z, Proshutinsky T, Proshutinsky A (2006) Tide-Tsunami
interactions. Sci Tsunami Hazards 24(4):242
Lay T, Ammon CJ, Kanamori H, Xue L, Kim MJ (2011) Possible
large near-trench slip during the great 2011 Tohoku (Mw 9.0)
earthquake. Earth Planet Sci Lett 63. doi:10.5047/eps.xxxx.xx.
xxx
Lockridge PA (1984) The criteria for Identification of tsunami events
likely to produce far-field damage, A data application. In :
Brennan AM, Lander JF (Eds.) Second UJNR Tsunami Work-
shop, Honolulu, Hawaii, November 1990, pp 43–50
Mendoza C, Dewey JW (1984) Seismicity associated with the great
Colombia-Ecuador earthquakes of 1942, 1968 and 1979: impli-
cations for barrier models of earthquake rupture. Bull Seismol
Soc Am 74:577–593
Moscoso E, Grevemeyer I, Contreras-Reyes E, Flueh ER, Dzierma Y,
Rabbel W, Thorwart M (2011) Revealing the deep structure and
rupture plane of the 2010 Maule, Chile earthquake (Mw =8.8)
using wide angle seismic data. Earth Planet Sci Lett 307:147–155
NEIC: http://www.neic.usgs.gov
Nocquet J-M, Mothes P (2009) Ge
´ode
´sie, ge
´odynamique et cycle
sismique en Equateur. Pub INOCAR, Ecuador In geologia y
sismicitad en la zona costera del Ecuador, Spec
Nocquet J-M, Villegas-Lanza JC, Chlieh M, Mothes PA, Rolandone
F, Jarrin P, Cisneros D, Alvarado A, Audin L, Bondoux F,
Martin X, Font Y, Re
´gnier M, Valle
´e M, Tran T, Beauval C,
Maguin
˜a Mendoza JM, Martinez W, Tavera H, Yepes H (2014)
Motion of continental slivers and creeping subduction in the
northern Andes. Nat Geosci. Published online: 2 March, 2014.
doi:10.1038/NGEO2099
Okada S (1985) Surface displacement due to shear and tensile faults
in a half-space. Bull Seismol Soc Am 75:1135–1154
Okada Y (1992) Internal deformation due to shear and tensile faults in
a half-space. Bull Seismol Soc Am 82:1018–1040
Pararas-Carayannis G (2012) Potential of Tsunami generation along
the Colombia/Ecuador Subduction Margin and the Dolores-
Guayaquil mega-thrust. Sci Tsunami Hazards 31(3):209–230
Mar Geophys Res
123
Author's personal copy

Pedoja K, Dumont J-F, Lamothe M, Ortlieb L, Collot J-Y, Ghaleb B,
Auclair M, Alvarez V, Labrousse B (2006) Plio-Quaternary
uplift of the Manta Peninsula and La Plata Island and the
subduction of the Carnegie Ridge, central coast of Ecuador. J S
Am Earth Sci 22:1–21
Savage JC (1983) A dislocation model of strain accumulation and
release at a subduction zone. J Geophys Res 88:4984–4996.
doi:10.1029/JB088iB06p04984
Segovia M (2001) El sismo de Bahia del 4 de Augusto de 1998:
Caracterizacion del mecanismo de ruptura y analisis de la
sismicidad en la zona costera. (Titulo de Ingeniera Geologia
tesis): Escuela Politecnica Nacional, Quito, Ecuador
Simons M, Minson SE, Sladen A, Ortega F, Jiang J, Owen S., Meng
L, Ampuerto J-P, Wei S, Chu R, Heimberger DV, Kanamori H,
Hetland E, Moore AW, Webb FK (2011) The 2011 Magnitude
9.0 Tohoku-Oki earthquake: Mosaicking the megathrust from
seconds to centuries. Science 332. doi:10.1126/science.1206731
Souris M (2002) Repu
´blica del Ecuador: modelo nume
´rico del
relieve, escala 1:2500000, Mode
`le nume
´rique calcule
´a
`partir des
cartes topographiques de l’ IGM (1:50000, 1:100000). IGM
Editors, Quito, Ecuador
Strasser FO, Arango MC, Bommer JJ (2010) Scaling of the source
dimensions of interface and intraslab subduction-zone earth-
quakes with moment magnitude. Seism Res Lett 81:941–950
Trenkamp R, Kellogg JN, Freymueller JT, Hector Mora P (2002)
Wide plate margin deformation, southern Central America and
northwestern South America, CASA GPS observations. J S Am
Earth Sci 15:157–171
Valle
´e M, Nocquet J-M, Battaglia J, Font Y, Segovia M, Re
´gnier M,
Mothes P, Jarrin P, Cisneros D, Vaca S, Yepes H, Martin X,
Be
´thoux N, Chlieh M (2013) Intense interface seismicity
triggered by a shallow slow-slip event in the Central-Ecuador
subduction zone. J Geophys Res Solid Earth 118(6):2965–2981
Wei G, Kirby JT (1995) A time-dependent numerical code for
extended Boussinesq equations. J Wtrwy Port Coast Ocean Eng
121:251–261
Wei G, Kirby JT, Grilli ST, Subramanya R (1995) A fully nonlinear
Boussinesq model for free surface waves. Part 1: highly
nonlinear unsteady waves. J Fluid Mech 294:71–92
Winter T, Avouac J-P, Lavenu A (1993) Late Quaternary kinematics
of the Pallatanga strike-slip fault (Central Ecuador) from
topographic measurements of displaced morphologic features.
Geophys J Int 115:905–920
Witt C, Bourgois J, Michaud F, Ordonez M, Jimenez N, Sosson M
(2006) Development of the Gulf of Guayaquil (Ecuador) during
the quaternary as an effect of the North Andean block tectonic
escape. Tectonics 25(3):22
Mar Geophys Res
123
Author's personal copy