Bull Earthquake Eng

DOI 10.1007/s10518-016-9906-7

ORIGINAL RESEARCH PAPER

Seismic hazard assessment in continental Ecuador

H. Parra1,2 • M. B. Benito1 • J. M. Gaspar-Escribano1

Received: 14 April 2015 / Accepted: 3 April 2016

 Springer Science+Business Media Dordrecht 2016

Abstract A seismic hazard assessment study of continental Ecuador is presented in this

paper. The study begins with a revision of the available information on seismic events and
the elaboration of a seismic catalog homogenized to magnitude Mw. Different seismic
source definitions are revised and a new area-source model, based on geological and
seismic data, is proposed. The available ground motion prediction equations for crustal and
subduction sources are analyzed and selected for the tectonic environments observed in
Ecuador. A probabilistic seismic hazard assessment approach is carried out to evaluate the
exceedance probability of several levels of peak ground acceleration PGA and spectral
accelerations SA (T) for periods (T) of 0.1, 0.2, 0.5, 1 and 2s. The resulting hazard maps
for continental Ecuador are presented, together with the uniform hazard spectra of four
province capital cities. Hazard disaggregation is carried out for target motions defined by
the PGA values and SA (1s) expected for return periods of 475 and 2475 years, providing
estimates for short-period and long-period controlling earthquakes.

Keywords Seismic hazard  Ground-motion  Hazard disaggregation  Ecuador

& M. B. Benito
mariabelen.benito@upm.es
H. Parra
haparra@espe.edu.ec
J. M. Gaspar-Escribano
jorge.gaspar@upm.es
1
Universidad Politécnica de Madrid, Campus Sur UPM, Ctra. de Valencia, km 7.5, 28031 Madrid,
Spain
2
Universidad de las Fuerzas Armadas ESPE, Av. Gral. Rumiñahui s/n, Sangolquı́, Ecuador

123
 Bull Earthquake Eng

1 Introduction

Ecuador is located in the Pacific Ring of Fire, in the northwest of South America. It is
considered a high seismic hazard region, due to intense seismic activity, mainly caused by
the subduction of the oceanic Nazca plate and the presence of a complex system of active
faults that generate crustal earthquakes. The greatest magnitude event, recorded during the
instrumental period, occurred along the northern coasts in 1906, with a Mw magnitude of
8.8. Earthquakes of lower magnitudes and shallow depths occurred in the Andes Mountain
Range. They involved significant human and material losses, as they hit large cities (INEC
2014; IGM 2013). As reported, on the western side of the Central and South American
continents, inland, moderate-magnitude, shallow events are more destructive than large
subduction events, although these have the potential to trigger tsunamis that cause major
disasters.
Previous work in seismic hazard assessment of the study area includes several regional
studies involving different countries. The Regional Seismological Centre of South America
(CERESIS) published a seismic hazard map of South America in 1981 (Shedlock and
Tanner 1999), which was updated in 1995 in collaboration with the Pan American Institute
of Geography and History (Tanner and Shepherd 1997). Five Andean countries and four
European countries cooperated in the PILOT program over 3 years (1995–1997) whose
goals included compiling a complete earthquake catalogue and producing a unified seismic
hazard map for the Andean Region (Dimaté et al. 1999). In 1999, the Global Seismic
Hazard Assessment Program (GSHAP) published a global seismic hazard map, as a
compendium of works in different regions (Grünthal et al. 1999). More recently, the SARA
project was undertaken an initiative for developing seismic hazard and risk assessments of
the continent, led by the GEM foundation.
More specifically for Ecuador, the Ecuadorian Construction Code (CEC 2001) included
the first hazard zonation map for the country, differentiating four seismic zones as a
function of the PGA values estimated for a return period of 475 years (ranging between
0.15 and 0.40 g). This code was updated in 2011 (and officially approved in 2015) with the
publication of the Ecuadorian Construction Norm (NEC-11 2015), which includes a new
seismic hazard map with six different zones. The lowest and highest hazard zones are
located in the eastern and northwestern parts of the country respectively, with PGA values
ranging from 0.15 g to more than 0.50 g. As the process followed to obtain these maps is
not published, it is not possible to understand its scientific basis.
A new seismic hazard map for the country, and specific results for four province
capitals, are presented in this work. This paper includes a summary of the study, along with
the main results achieved in continental Ecuador. The preparation of the necessary cal-
culation inputs required the development of the following items: the generation of a
database with the seismotectonic information available; an update of the seismic catalog
that includes the compilation of information from historical earthquakes presented in
different studies; a new area-source model accounting for crustal, subduction interface and
subduction in-slab sources and a selection of ground-motion prediction models, proposed
for tectonic regions similar to Ecuador. These inputs represent the largest contribution of
labor towards a greater knowledge of seismic hazard in Ecuador. The calculation follows a
probabilistic area-source method, configuring a logic tree with a node that includes several
combinations of ground-motion models and using specific software for data processing.

123
Bull Earthquake Eng

2 Seismotectonic framework

The tectonic setting of the area is controlled by the subduction of the oceanic Nazca plate
underneath the South American continental plate. The location and geometry of the sub-
ducting slab exert primary control in: the active volcanism of the North Central Andes
Cordillera; the transpressional deformation in the upper plate; and the seismicity at dif-
ferent depth levels (Hayes et al. 2012; Guillier et al. 2001).
The Nazca Oceanic Plate penetrates under Ecuador with a N80–83E direction and
relative plate motions of 50 and 70 mm/year (Nocquet et al. 2009; Gutscher et al. 1999).
The Carnegie submarine ridge, which extends from the Galapagos Islands to the sub-
duction trench and has an approximate elevation of 2000 m above the ocean bottom, is
located in this plate. Several studies suggest that this ridge plays a relevant role in the
subduction mechanism, the degree of interplate coupling and the continental deformation
(Michaud et al. 2009; Goyes 2009; Trenkamp et al. 2002). The Grijalva fracture zone is
located to the south of this ridge, with an approximate orientation of N60E, which starts
subducting in the Gulf of Guayaquil. It constitutes the limit between the Neogene Nazca
crust, with depths below 3000 m to the North and the oldest oceanic crust (Oligocene) and
with depths above 3000 m to the South (Pedoja et al. 2009; Collot et al. 2009; Flüh et al.
2001; Gutscher et al. 1999).
Another important tectonic unit located in the continental territory is the North Andean
block (NAB), which may be the result of the conjunction of several factors, such as the
oblique subduction of Nazca plate, the coupling generated in subduction, and the sub-
duction of the Carnegie ridge, among others (Witt and Bourgois 2009). Its motion would
occur along an active fault system than runs from the Guayaquil Gulf (South) and extends
to the North through the Andes Cordillera (Nocquet et al. 2014; Bethoux et al. 2011; Witt
et al. 2006; Trenkamp et al. 2002).
The trench that marks the limit between both plates has an average depth of 3000 m,
reaching shallower values in its intersection with the Carnegie ridge (Chunga et al. 2009;
Trenkamp et al. 2002) (Fig. 1a).
Three seismogenic regimes may be differentiated in Ecuador: (1) a subduction interface
source, which is relatively shallow and related to the contact between oceanic and conti-
nental plates; (2) a crustal source, located on the continental plate, of shallow nature, which
suggest a depth under the Andes of up to *75 km (Guillier et al. 2001); and (3) a
subduction in-slab source, related to the deformation of the subducted oceanic plate at deep
levels (Fig. 1b).
The seismic information available for Ecuador goes back to 1541, from specific studies
of effects caused by historical events. The strongest event, located between the coastline
and the trench, took place on 31 January 1906 with Mw 8.8, followed by other events,
whose dates and magnitudes are: 14 May 1942, Mw 7.8; 19 January 1958, Mw 7.8; and 12
December 1979, Mw 8.1 (Chunga et al. 2009; Gutscher et al. 1999). However, events with
minor magnitudes that have caused most damage occurred in the Ecuadorian highlands,
due to its shallow depth and proximity to towns. They are associated mainly with the fault
systems that accommodate the motion between the NAB and the South American plate.
The most significant ones occurred on 05 August 1949 with Mw 6.8 and on 06 March 1987
with Mw 7.1. The first one caused innumerable human and material losses in the cities of
Pelileo and Ambato, while the second one mainly affected the Napo and Sucumbı́os
provinces, as well as producing some damage in the city of Quito (Segovia and Alvarado
2009; Rivadeneira et al. 2007; Egüez et al. 2003).

123
 Bull Earthquake Eng

Fig. 1 a Tectonic framework and main physical characteristics of the study zone. Notation: NAB North
Andean block, NABL North Andean block limit (after Alvarado 2012); A–A0 -1 latitude cross-section,
Grijalva F. Z. Grijalva Fracture Zone, PE Esmeraldas Peninsula, PM Manabı́ Peninsula, PSE Santa Elena
Peninsula, GG Guayaquil Gulf, BB Borbón Basin, MB Manabı́ Basin, PB Progreso Basin, ChCC Chongón
Colonche Cordillera, WC Western Cordillera, IR Interandean Region, RC Real Cordillera, RV Reventador
Volcano, CC Cutucú Cordillera, CnC Condor Cordillera and b scheme of interface, in-slab and crustal
seismic sources, represented in the cross-section A–A0 and main geological units according to Vera (2013);
Notation: CP coast plain, AC Andes Cordillera, ER eastern region, ST subduction trench, CL coast line, WC
Western Cordillera, IR Interandean region, RC Real Cordillera

3 Seismic hazard assessment

The seismic hazard assessment developed in this study follows a probabilistic approach
that considers seismic area-sources. The inputs of the calculation include: (1) a seismic
catalog that covers 427 years, homogenized to moment magnitude Mw; (2) an area-source
model containing 21 seismogenic zones, which are characterized with corresponding
seismic parameters; and (3) ground-motion prediction equations (GMPEs), which have
been identified from amongst existing foreign models better able to reproduce the atten-
uation of the country. In addition, a simple logic tree is formulated in order to consider the
epistemic uncertainty related with GMPEs. These aspects will be discussed in detail in the
following sections.

123
Bull Earthquake Eng

3.1 Seismic catalog

The seismic catalog of this work contains records from 1587 to June 2014. It includes
instrumental and historic information published in specific studies and by different agen-
cies. The basic resource is the seismic catalog of the Instituto Geofı́sico-Politécnica
Nacional (IGEPN) of Ecuador. It contains information on events that occur until December
2009. This catalog is completed with other international catalogs, namely the International
Seismological Centre (ISC), and the National Earthquake Information Center-Preliminary
Determination of Epicenters (NEIC-PDE) catalogs. Additional information of historic
earthquakes is obtained from the specific works of Rivadeneira et al. (2007), Beauval et al.
(2010, 2013) and Alvarado (2012).
Several procedures are implemented with this catalog, so as to obtain a uniform size
parameter (homogenization), to remove foreshocks and aftershocks (declustering) and to
establish the periods within which the catalog can be considered complete (analysis of
completeness), as shown below.

3.1.1 Homogenization to Mw

The size parameter is converted from the original magnitude scale into moment magnitude,
Mw. The criteria published in Benito et al. (2012) to obtain Mw values from local mag-
nitudes and the correlations proposed in Beauval et al. (2013) to obtain Mw from the ISC
body wave magnitude (mbISC) and from the IGEPN duration magnitude (MD) are con-
sidered. Additionally, the relations Mw–mb (up to 6.0) and Mw–Ms (up to 8.0) proposed
by Utsu (2002) are used.
The resulting catalog contains 2879 events with magnitudes ranging from Mw 4.0 to
8.8. Most of them (93 %) correspond to magnitudes Mw B 5.5 and a few (0.9 %) to
magnitudes Mw C 7.0. The vast majority (97 %) of the records refer to events that
occurred since the 1960s (Fig. 2a, b), coinciding with a significant expansion of interna-
tional seismological networks.
The largest event recorded in instrumental period occurred in January 1906, with a
magnitude Mw 8.8, located along the subduction trench. There are 14 events prior to 1900
in the magnitude range Mw 6.1–7.6. Since 1900 the most notable earthquakes associated
with subduction sources have been located along the northwestern shorelines of the
country.
The distribution of seismicity shows two main sources: one associated with the sub-
duction of the Nazca plate and another related to crustal faulting that has been the source of
significant historical earthquakes occurring in the Andean Cordillera, where major cities
(e.g., Quito, Riobamba, Cuenca) are located.
The catalog is divided into three sets, corresponding to: crustal, subduction interface and
subduction in-slab events. Each division is organized according to geographical location,
focal depth and information of focal mechanisms. Crustal sources are limited by the Moho,
whose depth ranges from 75 km under the Andes (Guillier et al. 2001) to 40 km for the rest
of the country (Chunga et al. 2009). Additionally, three-dimensional representations of the
subduction surface (roughly the top of the subduction slab, as shown in Sect. 3.2.2) and
depth profiles crossing the trench trace perpendicularly, are plotted together with the
hypocenters to enable various events to be assigned to interface and in-slab subduction
sources.

123
 Bull Earthquake Eng

Fig. 2 Number of events per decade of the seismic catalog. a Number of events and b cumulative number
of events per magnitude range

3.1.2 Declustering

The catalog is declustered from foreshocks and aftershocks to verify the hypothesis that
seismicity fits to a Poisson model within each seismogenic area-source, as considered in
the probabilistic area-source method. The seismic catalog is analyzed to identify seismic
series, in Table 1 shows some of the series identified. Separated analyses are carried out for
crustal, in-slab and interface events attending to their temporal, spatial and magnitude
distributions.
The approach of Gardner and Knopoff (1974), using magnitude-dependent distance and
time windows to group events of the same series, is adopted and adjusted using data from
well-known earthquake series that occurred in the area of study. For each series, the main
shock magnitude and the largest time lag and distance of the related aftershocks are
obtained. These data are used to derive the time and space windows, obtaining the fol-
lowing equations:
Log T ¼ 0:68 M  1:57 ð1Þ

Log L ¼ 0:17 M þ 0:71 ð2Þ

Note that Eq. (1) is modified for magnitudes above 6.4 to avoid extremely wide time
windows, which is the same approach taken by Gardner and Knopoff (1974). The
declustering of the catalog shows that 28.4 % of catalog events are aftershocks. This

123
Bull Earthquake Eng

Table 1 Aftershock sequences

Date Mw T (days) L (km)

05 August 1996 6.3 512 58

21 May 2005 6.3 228 55
12 August 2010 6.3 136 49
18 August 1980 5.9 203 52
30 December 1988 5.3 112 46
09 October 2009 5.2 88 40
Mw moment magnitude, 11 April 1982 5.1 33 24
T sequence duration, and 27 October 2007 4.8 55 36
L distance of the aftershock 30 April 2009 4.1 14 24
furthest from the main shock

percentage is similar to the one obtained by Beauval et al. (2013) using the Reasenberg
(1985) declustering algorithm.

3.1.3 Analysis of completeness

Subsequently, the catalog is analyzed for lack of completeness, to account for the loss of
small-magnitude earthquakes as one goes back in time. The Stepp method (1972) is applied
for this analysis in half-degree magnitude intervals, using the declustered catalogs for
crustal and subduction sources. Table 2 shows the magnitude-dependent reference years
from which the catalog can be regarded complete for that magnitude interval. The refer-
ence years are drawn upon to compute seismicity rates, but do not imply any change in the
catalog composition.
The reference year for magnitudes Mw B 4.5, relates to the time that the IGEPN
seismic network provides solutions for low magnitude earthquakes and presents reasonable
coverage. The reference year of 1963 coincides with the beginning of distribution data
collected in worldwide seismic networks by international agencies. For magnitudes
between 6.0 and 7.0, the analysis marks a clear completeness beginning in 1910. For
magnitudes above 7.0, the analysis shows the completeness year in 1905 for the subduction
regime, while for the crustal regime, it is marked 1793, based on available historical
records. Those reference years will be used in a later stage to estimate seismic rates within
the period of completeness.

Table 2 Reference years that

Magnitude intervals (Mw) Reference years
establish completeness periods
for different magnitude intervals Crustal Subduction

4.0–4.5 1994 1994

4.5–5.0 1963 1963
5.0–5.5 1963 1963
5.5–6.0 1963 1963
6.0–6.5 1910 1910
6.5–7.0 1910 1910
[7.0 1793 1905

123
 Bull Earthquake Eng

3.1.4 Final catalogue

After the review, homogenization and declustering of the initial catalog, the final catalog
obtained for the region contains 2062 records with Mw C 4.0 (Fig. 3a). The highest
magnitude ones (M [ 6) are located along the coast (interface sources) and along the
Andes (crustal sources) (Fig. 3b).

3.2 Seismogenic zones and seismic parameters

A new seismogenic area-source model is developed for continental Ecuador. It is con-

strained by the following information: hypocenter distribution (taken from the final seismic
catalog); updated geological information at different scales (INIGEMM 2014); active
faulting data (Baize et al. 2015; Alvarado et al. 2014; Chunga 2010), which complement
the faults database and Quaternary folds of Egüez et al. (2003); bathymetry map (INOCAR
2014); slab geometry and relevant tectonic features (Hayes et al. 2012; Chunga et al. 2009;
Pontoise and Monfret 2004; Guillier et al. 2001; Penington 1981) and previously proposed
area-source models (Ortiz 2013; Alvarado 2012; Chunga 2010; Aguiar et al. 2009). The
seismogenic model contains separated crustal, subduction interface and subduction in-slab
area-sources.

3.2.1 Geometry and description of the seismogenic zones

3.2.1.1 Crustal zones The limits of the crustal area-sources give continuity to geological
structures and to the geological provinces proposed by Vera (2013), and are adjusted
considering the seismic information, the location and pattern of faulting and Quaternary
folds, and the main physical characteristics. Figure 4a, b show the crustal zones proposed
in this work superimposed on the geological map, and combined with the seismicity and
fault traces. In this set, coastal zones (C) are differenced from mountain ranges (CO and
CR), from the Interandean region (RI) and the eastern region (RO and A), which are named
with their respective symbols and followed by a sequential number (as shown in Table 3).
Figure 5 shows the number of earthquakes per area-source in the crust.

Fig. 3 a Epicenters of the seismic catalog compiled in the present study, with Mw C 4.0 and b Epicenters
of the Mw C 6.0 earthquakes contained in the catalog. The color represents the tectonic regime related to
each event

123
Bull Earthquake Eng

Fig. 4 Crustal zoning proposed for the seismic hazard calculation in this study. The defined area-sources
are superimposed to a geological structures (geological map of the Republic of the Ecuador, provided for
this study by the INIGEMM), and geological provinces (segmented lines) identified by Vera (2013); and to
b digital elevation model (30 m), quaternary faults and fold traces (Egüez et al. 2003) and shallow
seismicity according to the catalog prepared for the study

The coastal area-source zones are located in the coastal plain up to a height close to
1200 m. Zone C2 includes hills of maximum altitudes of about 850 m along the coastline
(Chongón-Colonche cordillera), whose uplift is related to the subduction process. The area
presents high seismicity, which is uniformly distributed, with a maximum recorded Mw
magnitude of 6.6. Zone C4 includes the Gulf of Guayaquil, which constitutes a defor-
mation zone of the Ecuadorian forearc, characterized by low-angle normal faulting. The
largest event recorded in Zone C4 corresponds to a magnitude Mw 5.6. Zone C3 is formed
accretion of sedimentary rocks and old oceanic basement. Its seismicity is in the Mw
magnitude range 4.0–4.8.
Area-sources covering the Andes Cordillera include the mountain ranges (zones CO1,
CR1 and CR2) and the Interandean valleys (zones RI1, RI2, RI3 and RI4). Zone CO1,
which is located along the Western Cordillera, is composed mainly of oceanic terraces, the

123
 Bull Earthquake Eng

Table 3 Distribution of Mmax for the seismogenic sources

Area-source Code Mmax obs MG Mmax ext MGR Mmax D

Crustal seismicity
Coast 1 C1 4.6 7.1 5.9 1.2
Coast 2Aa C2A 4.8 4.8 4.8 0.0
Coast 2Ba C2B 6.6 7.2 6.8 0.4
Coast 3 C3 4.8 7.2 7.0 0.2
Coast 4Aa C4A 4.6 4.6 4.6 0.0
Coast 4Ba C4B 5.6 7.2 6.9 0.3
Western Cordillera CO1 6.4 7.2 6.7 0.5
Inter-Andean Region 1 RI1 7.3 7.1 6.9 0.4
Inter-Andean Region 2 RI2 5.0 6.6 5.8 0.8
Inter-Andean Region 3 RI3 6.4 6.9 6.6 0.3
Inter-Andean Region 4 RI4 7.6 7.1 6.7 0.9
Real Cordillera 1 CR1 7.4 7.2 6.8 0.6
Real Cordillera 2 CR2 5.1 7.3 6.9 0.4
Eastern Region 1 RO1 5.4 7.1 6.8 0.3
Eastern Region 2 RO2 7.0 7.5 7.1 0.4
Amazon Region A1 4.8 5.2 5.0 0.2
Interface seismicity
Interface North Aa IFN_A 5.9 5.9 5.9 0.0
Interface North Ba IFN_B 8.8 (8.1) 8.8 8.8 0.7
Interface Center Aa IFC_A 5.3 5.3 5.3 0.0
Interface Center Ba IFC_B 7.8 7.9 7.9 0.1
Interface South IFS 7.5 (7.1) 7.5 7.5 0.4
In-slab seismicity
In-slab North ISN 7.2 6.8 7.2 0.4
In-slab North Central ISNC 6.5 7.5 7.0 0.5
In-slab South Central ISSC 7.5 7.4 7.5 0.1
In-slab South ISS 7.5 7.9 7.7 0.2
Code, id code for each zone; Mmax obs, maximum magnitude observed in the zone (in brackets, the second
highest magnitude value); MG, estimated Mw with geological criteria, using the correlation of Stirling et al.
(2002); Mmax ext, extreme Mw of the distribution estimated with geological criteria, using the correlation of
Stirling et al. (2002). MGR, Mw observed by tendency of the curve of Gutenberg–Richter, Mmax, expected
maximum Mw magnitude; D, estimated values by the criteria described above
a
Zones with double fit: the minimum magnitude in the A sections is 4.0. However, these values are higher
for sections B, marking the straight line threshold of lesser slopes observed in the fits

Macuchi Island Arc and Pliocene-Quaternary volcanoes. Its seismicity is homogeneously

distributed. The largest recorded event located in this zone corresponds to a (historic)
magnitude Mw 6.4. The CR1 and CR2 zones cover the Real Cordillera, which is older and
higher than the Western Cordillera. It is formed mainly by rocks accreted in an East–West
direction. Fault systems that could accommodate the movement of the NAB are located in
zone CR1. This zone presents the largest number of events among the zones of the Andes
area, in a range of magnitude Mw 4.0–7.4. Zone CR2 corresponds to the lowest part of the
Real Cordillera, where Quaternary volcanic structures are not observed (Vera 2013;
Guillier et al. 2001). It shows a high seismicity not completely homogeneously distributed.

123
Bull Earthquake Eng

Fig. 5 Number of events in the

cortical seismogenic sources with
Mw [ 4.0

The RI1, RI2, RI3 and RI4 zones, located in the Interandean Valley, are formed by
geomorphological depressions, filled in with detritus and volcanic sediments (IGM 2013;
Egüez et al. 2003). Maximum altitudes range approximately between 2200 and 3000 m.
The largest number of historical earthquakes are located in an estimated range of mag-
nitude Mw 5.8–7.6.
The eastern region, which represents the back-arc zone, is modeled by zones RO1 and
RO2 in the Eastern foothills of the Real Cordillera. Zone RO1 is formed by andesitic
volcanoes, being the Reventador volcano the oldest and most active one in this sector.
Zone RO2 contains the Cutucú Cordillera and the Cóndor Cordillera, and shows the
highest seismicity between crustal zones, with events of magnitudes up to Mw 7.0. Finally,
Zone A1 runs across the eastern part of Ecuador, which forms the Amazon River head-
waters and shows lower seismicity, in a range of magnitude Mw B 4.8.

3.2.1.2 Subduction zones Subduction interface area-sources begin on the edge of the
subduction trench, while the in-slab area-sources contain deep events (Fig. 6a, b). Several
regional and local works are considered for the determination of its dip, such as Vallée
et al. (2013), Hayes et al. (2012), Manchuel et al. (2009), Trenkamp et al. (2002) and

Fig. 6 Area-source model for the subduction regime proposed in this study. a Interface area-sources with
its corresponding seismicity and b in-slab area-sources with its corresponding seismicity

123
 Bull Earthquake Eng

123
 Bull Earthquake Eng

b Fig. 7 Variation in the dip angle a depth contours showing the dip variability of the interface and in-slab
area-sources (in km); b cross-section A–A0 , at -1450 , and c cross-section B–B0 , at -3550 , where the
approximate dip angle of the roof of the slab is represented, and the corresponding seismicity. Adopted
notation: CP coastal plain, AC Andes Cordillera, ER Eastern Region, ST Subduction Trench, CL coast line,
WC Western Cordillera, IAV Interandean valley, EC Eastern Cordillera, RC Real Cordillera, CnC Condor
Cordillera. Vertical coordinate of topographic profile is exaggerated

Guillier et al. (2001), among others. This analysis suggests a variation in the dip angle,
which can be due to the presence of the Carnegie ridge or the differentiation between an
older and deeper oceanic crust in the South of the Grijalva fracture zone and a shallower
oceanic crust North of this, enabling seismicity differentiation at different depth levels
(Fig. 7a–c).
In the subduction zones, a heterogeneous distribution of seismicity is observed (Fig. 8).
The zones Interface North and in-slab South show greater seismicity, with 281 and 295
earthquakes (Mw C 4.0) respectively. The largest earthquakes recorded, of magnitude Mw
8.8 and 8.1, are located in the Interface North zone, while seven events in the magnitude
range Mw 6.8–7.5 are located in the in-slab South zone. In the southwestern part of the in-
slab North Central zone, there is a concentration of deep earthquakes, with depths larger
than 130 km. In the Interface South zone, there are 240 events, four of them with Mw
magnitude ranging between 6.8 and 7.5, at the southwest of the Gulf of Guayaquil. The
zone with less seismic activity is the in-slab North zone, although the largest events are in
the Mw interval of 6.5–7.2.

3.2.2 Seismic parameters

After the definition of area-sources, the parameters which characterize their seismic
activity are estimated using a Gutenberg–Richter model (Gutenberg and Richter 1944)
truncated to a minimum magnitude of 4.0 (Fig. 9a, b). Annual rates and b-values are
calculated using a maximum likelihood procedure following the method of Weichert
(1980). A clear slope change is noted in some zones (IFN, C2) where it was necessary to
perform a double fit to account for the lower slope at larger magnitudes (Fig. 9c, d).
The maximum expected magnitude (Mmax) in each area-source is modeled using a
uniform probability distribution, between a lower threshold ML and an upper limit Mu. In
this way, the aleatory uncertainty inherent to the Mmax is captured, considering all the
values that this parameter can take between the two limit values (ML, Mu). The half-width
interval D = (Mu - ML)/2 is a dispersion measure of this parameter, so that:
ML ¼ Mmax  D and ð3Þ

Mu ¼ Mmax þ D ð4Þ
In most zones, the determination of ML and Mu is not trivial, so it is necessary to make a
number of considerations and adjustments, based on the observed seismicity and the
available data from existing faults that allow estimating maximum magnitude values with
geological criteria. A different problem is located in crustal zones and in subduction zones,
as described below.
In crustal zones, the correlations proposed by Stirling et al. (2002) are applied in order
to estimate Mw from the length of active fault segments (as contained in the Quaternary
faults database of Egüez et al. 2003 and the complementary information of Chunga 2010).
This magnitude, estimated with geological criteria, is called MG. Depending on the number
of faults in each zone, different approaches are employed:

123
 Bull Earthquake Eng

Fig. 8 Number of events of magnitude Mw C 4.0 in the interface and in-slab seismogenic sources

Fig. 9 Examples of the Gutenberg–Richter curves obtained for some area-sources, showing the cumulative
annual seismicity rate. a, b Simple-fit examples; c, d double-fit examples. a In-slab South (ISS), b Inter-
Andean Region (RI3), c Interface North (IFN) and d Coast 2 (C2)

1. In most crustal zones, there is a sufficient number of known faults to compose a

frequency distribution of MG values (Fig. 10a). Mmax is then identified as the average
magnitude of the distribution of MG values (MG ) and D as the difference between
Mmax and the extreme value (Mmax ext) of this distribution. In this case:
Mmax ¼ MG ð5Þ

D ¼ Mmax ext  MG ð6Þ

2. In some crustal zones (C1 and RI2), there is only one fault to estimate MG, which is
associated with Mu (Fig. 10b). The maximum observed magnitude (Mmax obs) is
identified with ML. Then, Mmax and D are:

123
Bull Earthquake Eng

Fig. 10 Examples of the procedure followed for the determination of maximum magnitudes for a crustal
zones, with an adequate number of active faults, to adjust the magnitudes derived by geological criteria
(depending on the length) to a frequency distribution. b Crustal zones with a single fault in its interior and
c subduction zones with MGR [ Mmax obs. [Ln(N): natural logarithm of the theoretical cumulative number of
earthquakes N within the period covered by the study]

Mmax ¼ ðMmax obs þ MG Þ=2 ð7Þ

D ¼ ðMG  Mmax obs Þ =2 ð8Þ

3. In three zones (RI1, RI4 and CR1), the maximum observed magnitudes, as given by
historical earthquakes, are Mw 7.3 (1868), 7.6 (1797) and 7.4 (1834) respectively.
These values are higher than the MG values derived for these zones. This can be
indicative of an overestimation of the Mw value assigned to those historic events. In
these cases, more credibility was assigned to the Mmax estimated by geology, whilst
considering a deviation D that covers the total range up to the Mmax observed. Hence,
Mmax is the average of the distribution of MG values, MG, and D is the difference
between Mmax and Mmax obs (which is the maximum magnitude derived from historical
records).
Mmax ¼ MG ð9Þ

D ¼ Mmax obs  MG ð10Þ

The assessment of Mmax for the interface and in-slab subduction area-sources is dif-
ficult due to the lack of data constraints for establishing possible ruptures and slip rates and
hence, the estimation of maximum magnitudes. In this regard, GPS measurements are
required in order to determine the maximum credible earthquakes along the trench. As a

123
 Bull Earthquake Eng

‘‘less bad’’ solution, Mu is estimated from extrapolation of the Gutenberg–Richter model to

the axis of Magnitude when considering the plot of theoretical number of earthquakes in
the period of completeness (estimated from the rate, as in Fig. 10c) versus magnitude. The
cut-off point may give an estimation of the upper limit for M following the trend of the
seismicity.
The magnitude that is obtained in the cut-off point is called MGR. The following cases
are identified:
1: If MGR [ Mmax obs ; then Mmax ¼ ðMGR þ Mmax obs Þ=2and ð11Þ

D ¼ ðMGR  Mmax obs Þ =2 ð12Þ

2: If MGR  Mmax obs ; then Mmax ¼ Mmax obs ; and D ¼ Mmax  MGR : ð13Þ
3. If MGR = Mmax obs, then Mmax = Mmax obs, and the D value is identified as the dif-
ference between the two highest magnitude values observed in the zone.
The final Mmax and D values estimated with the different approximations, together with
the seismic parameters obtained for each zone are shown in Table 4.

3.3 Ground-motion prediction equations selection

Unfortunately, there are no strong motion records available in Ecuador to develop a ground
motion model constrained with local data, or to calibrate foreign models with local data.
For this reason, different GMPEs formulated for subduction and crust in other parts of the
world, are analyzed to select those that fit better with the tectonic environment of Ecuador.
Few studies have been published regarding the suitability of GMPEs for South America.
The publication of Arango et al. (2012) discusses the GMPEs applicability for Central and
South America (Chile-Peru) subduction zones, using databases of both regions. Among the
conclusions, the Youngs et al. (1997) model is presented as still suitable for modelling the
attenuation of subduction interface and in-slab events. For crustal zones, the McVerry et al.
(2006) model developed with data from earthquakes in New Zealand, seems to adjust in an
acceptable manner to the data available in South America, especially for intermediate and
long spectral periods. Finally, the Zhao et al. (2006) model developed for both crustal and
subduction zones, from records of Japan, is capable of predicting ground motions in the
South America region.
Another criterion is the one published in the NEC-11, which mentions the use of
GMPEs valid for subduction sources, among others Youngs et al. (1997), and Zhao et al.
(2006). The work of Beauval et al. (2014) developed for the seismic hazard calculation in
Quito, considers the use of Zhao et al. (2006), for three tectonic regimes, and of Youngs
et al. (1997), for tectonic subduction regimes, showing the robustness and stability of the
first for all frequency ranges in the crustal regime.
Aguiar et al. (2014) uses the Zhao et al. (2006) model for the accelerations calculation,
after the shallow earthquake occurred in 12 August 2014, in the vicinity of Quito, showing
acceptable results compared to the accelerations recorded in an instrumental way
(Table 5).
In the absence of additional information for the selection of other models, the results of
Arango et al. (2012) are adopted in this work. The characteristics of these models, included
in Table 5. Figure 11 present a comparison of the PGA values obtained as a function of
rupture distance (the shortest distance from the source rupture to the calculation site) for
magnitudes Mw 5.0 and 7.0 and different tectonic regimes. It was observed that the model

123
Bull Earthquake Eng

Table 4 Seismic parameters

Area-source Code m0 Mmax D a b rb N_ (m0) N_ (m0)/area

Crustal seismicity
Coast 1 C1 4.0 5.9 1.2 10.744 1.988 1.645 0.320 2.47E-05
Coast 2Aa C2A 4.0 4.8 0.0 17.416 2.960 0.291 2.539 6.10E-05
Coast 2Ba C2B 4.9 6.8 0.4 10.958 1.590 0.704 0.268 6.44E-06
Coast 3 C3 4.0 7.0 0.2 15.648 3.027 0.949 0.677 3.49E-05
Coast 4Aa C4A 4.0 4.6 0.0 17.853 3.330 0.428 1.823 7.44E-05
Coast 4Ba C4B 4.7 6.9 0.3 13.295 2.300 1.138 0.235 9.60E-06
Western Cordillera CO1 4.0 6.7 0.5 15.720 2.881 0.520 1.308 5.35E-05
Inter-Andean Region 1 RI1 4.0 6.9 0.4 7.959 0.978 0.306 0.258 1.79E-05
Inter-Andean Region 2 RI2 4.0 5.8 0.8 18.516 3.766 0.948 0.618 2.04E-04
Inter-Andean Region 3 RI3 4.0 6.6 0.3 9.414 1.467 0.466 0.334 1.20E-04
Inter-Andean Region 4 RI4 4.0 6.7 0.9 11.293 1.601 0.269 0.602 9.71E-05
Real Cordillera 1 CR1 4.0 6.8 0.6 13.250 1.859 0.206 1.514 8.36E-05
Real Cordillera 2 CR2 4.0 6.9 0.4 18.156 3.446 0.544 1.553 2.07E-05
Eastern Region 1 RO1 4.0 6.8 0.3 15.588 2.830 0.463 1.398 5.07E-05
Eastern Region 2 RO2 4.0 7.1 0.4 15.086 2.152 0.185 2.946 6.25E-05
Amazon Region A1 4.0 5.0 0.2 12.678 2.430 1.209 0.378 6.23E-06
Interface seismicity
Interface North Aa IFN_A 4.0 5.9 0.0 14.059 1.870 0.099 6.609 1.10E-04
Interface North Ba IFN_B 6.0 8.8 0.7 8.739 1.000 0.370 0.142 2.37E-06
Interface Center Aa IFC_A 4.0 5.3 0.0 14.551 2.090 4.479 4.479 1.14E-04
Interface Center Ba IFC_B 5.4 7.9 0.1 10.290 1.300 0.241 0.241 6.12E-06
Interface South IFS 4.0 7.5 0.4 13.994 1.869 0.107 6.205 1.05E-04
In-slab seismicity
In-slab North ISN 4.0 7.2 0.4 10.487 1.541 0.286 0.690 8.43E-06
In-slab North Central ISNC 4.0 7.0 0.5 12.179 1.640 0.171 2.653 5.33E-05
In-slab South Central ISSC 4.0 7.5 0.1 13.465 1.821 0.123 4.441 6.70E-05
In-slab South ISS 4.0 7.7 0.2 13.209 1.665 0.094 6.414 3.17E-05

mo, minimum magnitude to compute the rate; Mmax, expected maximum Mw magnitude; D, Delta; a, b,
parameters of the Gutenberg–Richter relationship; r b, Beta sigma; N_ (mo), events rate with Mw C mo; N_
(mo)/area, events rate with Mw C mo, divided by the area
a
Zones with double adjustments

of Youngs et al. (1997) gives larger accelerations than the model of Zhao et al. (2006),
whereas the model of McVerry et al. (2006) predicts larger accelerations than the model of
Zhao et al. (2006) for low magnitudes and lower accelerations for high magnitudes.

3.4 Seismic hazard estimation

Seismic hazard is estimated according to the PSHA methodology using the software
CRISIS 2012 v5.5 (Ordaz et al. 2013). This program allows the implementation of the
probabilistic method with an area-source model assuming that earthquake occurrence
follows a Poissonian process. Area-sources are subdivided in triangular subsources up to a
threshold size specified by the user. Source-to-site distances are calculated respecting

123
 Bull Earthquake Eng

Table 5 Characteristics of the ground motion prediction equations used in this study
Model Code Area Regime N record Mmin Mmax Rmin Rmax Type
of R

Zhao et al. ZH06 Japan CR, IT, IB 2763 5.0 8.3 0 300 Rrup
(2006)
McVerry et al. McV06 New Zealand CR, IT, IB 435 5.0 7.0 6 400 Rrup
(2006)
Youngs et al. Young97 World IT, IB 476 5.0 8.2 8.5 550 Rrup
(1997)

Regime tectonic regime, CR crustal, IT interface, IB in-slab, Mmin minimum magnitude, Mmax maximum
magnitude, R distance, Rmin minimum distance in km, Rmax maximum distance in km, Rrup shortest
distance to the fault rupture

compatibility with the magnitude definition of GMPEs. CRISIS allows the use of rupture
distance by modelling rupture areas as circles of magnitude-dependent radius R, using the
general form:
A ¼ K1 expðK2  MÞ ð14Þ
2
where A is the source area (in km ), M is the magnitude, and K1 and K2, are constants that
for the present study correspond to the model developed by Wells and Coppermith (1994).
Additionally, a GMPE viewer is included that facilitates checking the performance of
different GMPE choices. Results may be obtained for a grid of calculation points, in terms
of exceedance probabilities in user-defined exposure periods and ground motion levels
(Ordaz et al. 2013).
The calculations are made for points of a grid that cover all the territory, with a
separation of 0.1 both in longitude and latitude. Results are obtained for PGA and SA
(T) for T = 0.1, 0.2, 0.5, 1.0 and 2.0s and for return periods of 475, 975 and 2475 years.
Expected accelerations are obtained for rock conditions, thereby excluding local amplifi-
cation effects that may be significant. This approach is usual for hazard mapping of large
areas. A more local scale, site-specific approach would require a further detailed assess-
ment of site effects that is beyond the scope of this work.
A simple logic tree with a node to quantify the inherent epistemic uncertainty in the
GPMEs is configured for the calculation. The attenuation node is composed of four
branches, corresponding to different GMPE combinations. Each branch combines models
for crustal, interface and in-slab zones. All branches are equally weighted because there are
no criteria to favor a certain branch over the others (Fig. 12). It is noteworthy that each
logic tree branch does not correspond to a single GMPE, but to a combination of the three
of them, one for each tectonic regime.
The aleatory uncertainties associated with the seismic parameters (b and Mmax) of each
zone are also considered.

4 Results

4.1 Hazard maps

The first results obtained from the hazard calculation are maps expressed in terms of PGA
and SA (1s) for three return periods: 475, 975 and 2475 years, which correspond to ground

123
Bull Earthquake Eng

Fig. 11 Representation of the selected GMPEs for different magnitudes. a Interface, b in-slab and c Crustal
sources

Crustal (McV06) + Interface (Young97) + In-slab

Crustal (McV06) + Interface (ZH06) + In-slab (ZH06)

GMP
Crustal (ZH06) + Interface (Young97) + In-slab (Young97)

Crustal (ZH06) + Interface (ZH06) + In-slab

Fig. 12 Logic tree elaborated to capture the inherent uncertainty in the GMPEs and assigned weights

123
 Bull Earthquake Eng

motions with an exceedance probability of 10, 5 and 2 % in 50 years, respectively. These

maps represent the mean values obtained with the four combinations of GMPEs and the
distributions taken to quantify the random uncertainty of the seismicity parameters. Fig-
ure 13 shows the PGA maps for the three return periods, and Fig. 14 shows similar results
in terms of SA (1s). The analysis of the maps shows that:
• The highest PGA values are predicted, for the three return periods, in areas close to the
Ecuadorian coastline, especially in the Santa Elena Peninsula, Manabı́ Peninsula and
Esmeraldas Peninsula. High values are also obtained in the Interandean region and
South-East of the Ecuadorian continental territory, specifically in the region of the
Condor Cordillera, where, in addition to the local faulting, the subduction effect is
appreciated.
• For a return period of 475 years, the PGA reaches maximum values of 0.76 g in the
zone of the Esmeraldas Peninsula and 0.4 g in the Interandean and South-Eastern
regions. PGA values decrease eastwards, reaching values lower than 0.15 g.
• The forms of PGA isolines for return periods of 975 and 2475 years are quite similar,
although the values are higher for larger return periods. Maximum values of 1.0 and
1.37 g, respectively, are reached in zones with greater hazard.
• In the hazard maps in terms of SA (1s), the highest values are observed on the coast,
making evident the great influence of the subduction activity for this structural period.
Maximum values of 0.61, 0.87 and 1.28 g are reached for return periods of 475, 975
and 2475 years, respectively.

4.2 Specific results in four capitals

Four cities are selected to present specific results due to a combination of factors, such as
their geographical distribution, their location within the physical, geological and tectonic
characteristics of the country, their population density, the damage caused by historical
earthquakes at these cities and their location in the hazard maps for different return periods.
These cities are: Esmeraldas (Northwest Coast), Quito (Northern Andes Cordillera),
Guayaquil (Guayaquil Gulf) and Loja (South Andes Cordillera). Figure 15 shows the
hazard curves obtained in these cities for six representative ground motion parameters:
PGA, SA (0.1s), SA (0.2s), SA (0.5s), SA (1s) and SA (2).
In addition, Fig. 16 shows the Uniform Hazard Spectrum (UHS) obtained in these cities
for the three return periods (475, 975 and 2475 years).
The study of each city is completed with a hazard disaggregation analysis for the target
motions given by the expected PGA and SA (1s) values. Seismic hazard is disaggregated in
magnitude using distance bins of width 0.5 magnitude units and 30 km in distance,
respectively. As a result, the controlling earthquakes that contribute most to the hazard in
every city, for short and long structural periods, and exceedance probabilities of 10 %, and
2 % in 50 years, are determined. Each controlling earthquake is characterized by a mag-
nitude and distance bin (M–R). Disaggregation results are shown in Fig. 17, which displays
the hazard contribution of different M-R bins, for the target motion given by the PGA, for
the return period of 475 years. Similar results for SA (1s) are shown in Fig. 18.
In addition, although the graphics are not shown here, the analysis for a return period of
2475 years was also carried out.
A summary of the controlling earthquakes is shown in Table 6. It is interesting to note
the changes that are observed depending on the return and/or structural period. It is also
noteworthy that in the case of Guayaquil, two modes are identified in Figs. 17 and 18,

123
Bull Earthquake Eng

Fig. 13 Seismic hazard maps of Ecuador in terms of PGA obtained in this study for return periods of
a 475 years, b 975 years and c 2475 years

123
 Bull Earthquake Eng

Fig. 14 Seismic hazard maps of Ecuador in terms of SA (1s) obtained in this study for return periods of
a 475 years, b 975 years and c 2475 years

123
Bull Earthquake Eng

Fig. 15 Hazard curves for the cities of a Esmeraldas, b Quito, c Guayaquil and d Loja in terms of PGA and
SA (T) in (g) for T = 0.1, 0.2, 0.5, 1 and 2s

respectively, which can be associated with two controlling earthquakes, one of moderate
magnitude and near source and the other corresponding to a more distant source and higher
magnitude. Both controlling earthquakes are included in Table 6.
Once the results for each of the cities were revised, the following aspects could be
highlighted:
• In Esmeraldas, for a return period of TR = 475, a clear controlling earthquake is
identified by the pair (Mw 7.50, R = 30–60 km) for short and long structural periods.
A second mode is observed, identified by the pair (Mw 6.50, R = 0–30 km), although
with lower probability density. As both crustal and interface sources are located at short
distances in Esmeraldas, it is not evident to ascertain the source responsible for these
controlling earthquakes. For TR = 2475 years, the controlling earthquake is identified
with (Mw 6.50, R = 0–30 km), which could be associated with the influence of the
local faulting activity, possibly due to the Esmeraldas (Northern section) fault, located
in the crustal zone C2. This zone presents a lower rate and thus its contribution is
noticeable at long TR values, which is not a typical feature.
• In Quito, the pair (Mw 6.25, R = 0–30 km) represent a clear controlling earthquake
both for PGA and SA (1s) and for the two return periods. These events show that
hazard in Quito is dominated by faults located in the source zone RI 3. Alvarado et al.
(2014) undertook a study of the Quito fault system, locating fault segments whose
rupture lengths could be related to magnitudes in the range of Mw 6.0–6.4.

123
 Bull Earthquake Eng

Fig. 16 Uniform Hazard Spectrum (UHS) obtained in four cities of Ecuador, for return periods of
475 years, 975 years and 2475 years. a Esmeraldas City, b Quito City, c Guayaquil City and d Loja City

Fig. 17 Hazard disaggregation results in the cities of Esmeraldas, Quito, Guayaquil and Loja.
Contributions of the (Mw, R) pairs are represented for the target motion given by PGA for the return
period of 475 years. a Esmeraldas City, b Quito City, c Guayaquil City and d Loja City

123
Bull Earthquake Eng

Fig. 18 Hazard disaggregation results in the cities of Esmeraldas, Quito, Guayaquil and Loja.
Contributions of the (Mw, R) pairs are represented for the target motion given by SA (1s) for the return
period of 475 years. a Esmeraldas City, b Quito City, c Guayaquil City and d Loja City

Table 6 Disaggregation Results: Controlling Earthquakes

Units (g) Return period 475 Return period 2475

PGA SA (1s) PGA SA (1s)

Capital city Controlling Mw R (km) Mw R (km) Mw R (km) Mw R (km)

earthquake

Esmeraldas CE1 7.50 30–60 7.50 30–60 6.50 0–30 6.50 0–30
Quito CE1 6.25 0–30 6.25 0–30 6.25 0–30 6.25 0–30
Guayaquil CE1 7.25 60–90 7.25 60–90 6.50 0–30 7.25 60–90
CE2 – – 6.75 60–90 – – 6.50 0–30
Loja CE1 7.00 60–90 7.00 60–90 7.25 60–90 7.25 60–90

• In Guayaquil, a clear controlling earthquake emerges for the two return periods given
by the pair (Mw 7.25, R = 60–90 km), which could be identified with a subduction in-
slab earthquake in the Intra South Central zone (ISSC). An interesting case is observed
for TR 2475 and SA (1s) given by the pair (Mw 6.50, R = 0–30 km), which could be
associated with the influence of the activity of the local faulting, possibly to the Carrisal
or Colonche faults, located in the crustal zone C2. This may be due to the rate of
seismicity of the crustal zone (C2), which is lower than the rate of the subduction zone

123
 Bull Earthquake Eng

(ISSC), therefore its contribution to the hazard will be more noticeable in this period.
For SA (1s) and TRs 475 and 2475, another pair are observed (Mw 6.75,
R = 60–90 km and Mw 6.50, R 0–30 respectively), although with lower probability
density, which could be associated with the subduction interface and crustal activity.
• In Loja for TR of 475 and 2475 years, a controlling earthquake emerged, which is
associated with the pair (Mw 7.0–7.25, R = 60–90 km), for both structural periods,
which might be related to the in-slab activity of the Intra South source zone. A second
mode is observed, for a return period of TR = 475 years, and SA (1s), identified by the
pair (Mw 6.50, R = 60–90 km), although with lower probability density.

5 Discussion

The most recent hazard maps of Ecuador have been published by the Ecuadorian Con-
struction Norm (NEC-11), which was based on a probabilistic approach, assuming a 10 %
exceedance probability in 50 years (return period of 475 years). This map determines six
seismic zones, in which the acceleration in rock covers a range from PGA = 0.15 g in the
North-East, to PGA C 0.50 g at the VI zone, corresponding to the Ecuadorian coast.
However, NEC 11 sets a saturation value of 0.50 g for seismic design of buildings located
along the northwestern coast although the acceleration value originally obtained for this
area was larger. The process followed to obtain these maps is not published, so it is not
possible to determine potential causes leading to different results from those obtained in
the present work.
The comparison between the PGA results on rock conditions for a return period of
475 years obtained in the present study and those published in the NEC-11 shows a similar
range of values, both in eastern Ecuador and in the central highlands. However, in the
southern part of the country, the PGA values obtained in this study are higher than those
presented in NEC-11 and are closer to the acceleration values contained in the Peruvian
Design Seismic Resistant Norm (Norma E.30, Diseño Sismoresistente). Meanwhile, the
PGA values obtained for the North of Ecuador are lower than those of the NEC 11, but
more consistent with the Colombian Seismic Resistant Norm (NSR-10, Reglamento
Colombiano de Construcción Sismoresistente).
Table 7 presents the PGA values obtained in this study for four provincial capitals for a
return period of 475 years, compared to the ones given by NEC-11. Minor differences are
observed in the cities of Guayaquil and Quito, where our study shows slightly higher
values, with differences of 0.02–0.03 g. This difference is accentuated in Loja, where our
values exceed the PGA value provided in NEC-11 by approximately 0.13 g. The value of
0.38 g, given in this study, is more congruent with the values presented in the North border
of the Norma E.30. Finally, it is not possible to develop a precise comparison for
Esmeraldas due to the saturation value of 0.50 g adopted by the NEC-11.
The determination of the maximum magnitude expected in each zone directly affects
the variability of acceleration levels, which are subsequently obtained in the hazard

Table 7 PGA (g) values for cities of this study

Esmeraldas Quito Guayaquil Loja

This study 0.65 0.43 0.42 0.38

NEC-11 C0.50 0.40 0.40 0.25

123
Bull Earthquake Eng

calculations. This paper provides a PSHA assessment, taking into account all available
geological and seismological data and associated uncertainties.
The foreign models selection for the ground motion prediction, with applicability limits
both in low and high magnitude generated for some regions, still constitutes one of the
main uncertainty sources in seismic hazard analysis. The generation of greater amounts of
local strong motion data and models would facilitate reducing this source of uncertainty in
future studies.

6 Summary and conclusions

A seismic hazard calculation following a zoned probabilistic model (PSHA) has been
presented in this paper. For the development of this study, a new seismic catalog and a new
seismogenic area-source model are elaborated. The GMPEs selected are among the ones
existing in specialized literature, which consider the attenuation characteristics of Ecuador.
The seismic catalog covers the period 1584–2014 and it is prepared by collecting data
from different agencies. The size parameter is homogenized to a moment magnitude, Mw.
The catalog is declustered for foreshocks and aftershocks, to accomplish the basic
hypothesis of earthquake independence required in the area-source Poissonian model. The
resulting final catalog contains 2062 records with Mw C 4.0.
A seismogenic area-source model is developed combining different geological and
seismic information types, which allows identifying the boundaries of grouped zones into
three tectonic regimes: crustal, subduction interface and subduction in-slab. Particular
attention was placed on the analysis of the subduction dip in order to generate an area-
source model consistent with the physical, geological and tectonic knowledge of the study
region.
The seismicity in each zone is modeled applying the Gutenberg–Richter law to the
estimation recurrence model. In some coastal zone cases, the seismicity pattern required a
double fit, reflecting a lower slope in high magnitude ranges, somehow approaching a
characteristic earthquake model.
Due to the lack of availability of local accelerograms for the development of GMPEs,
foreign models formulated for subduction and crustal sources in the world are selected for
their tectonic affinity with Ecuador. The related epistemic uncertainty is considered
through a logic tree whose branches are combinations of models for the three tectonic
regimes set forth above.
Seismic hazard is estimated in rock conditions for five spectral accelerations 0.1s, 0.2s,
0.5s, 1.0s and 2.0s, plus PGA, allowing maps with these parameters to be generated for the
whole country, in addition to UHS and hazard curves for four provincial capitals: Quito,
Esmeraldas, Guayaquil and Loja. UHS and hazard maps are obtained for return periods of
475, 975 and 2475 years. New contributions include: results of disaggregation for the four
capital cities; providing (Mw, R) pairs that contribute most to the hazard and representing
the controlling earthquakes.
Analyzing the hazard maps, the following can be concluded: for all the return periods,
the maximum PGA values are located on the Ecuadorian coast, followed by high values in
the central highlands and in southeastern Ecuador. The isolines are parallel to the coastline,
with acceleration values decreasing towards the East. The morphology of the maps for SA
(1s) changes, presenting lower gradients along the coastal areas, and evidencing the
influence of the subduction sources.

123
 Bull Earthquake Eng

Looking at specific outcomes for the capitals, hazard curves, UHS and controlling
earthquakes, it is apparent that the greatest hazard would be expected in the city of
Esmeraldas, due to the subduction interface process; followed by Quito, where the hazard
would be dominated by crustal activity; Guayaquil, which would be influenced mostly by
the source in-slab subduction; and finally, the city of Loja, where the hazard would be
dominated by in-slab activity.
The study results may have different applications, such as: the direct definition of the
seismic action (and not through spectral forms) for the earthquake-resistant design for both
normal- and special-importance structures. In addition, the controlling earthquakes
obtained by disaggregation allow the calculation specific response spectra for singular,
critical structures such as dams, bridges, etc. Finally, these results can serve as input for
future seismic risk studies, for application in the elaboration of emergency plans, post-
event actions, and mitigation measures.

7 Data and resources

The geological information for this study was provided by the Instituto Nacional de
Investigación Geológico, Minero, Metalúrgico of Ecuador; the mapping database was
made available by the Instituto Geográfico Militar of Ecuador; the seismic catalog updated
to December 2009 was given by the Instituto Geofı́sico of the Escuela Politécnica Nacional
of Ecuador; for the depuration and the calculations of seismicity parameters, software
developed by the Grupo de Investigación de Ingenierı́a Sı́smica (GIIS) of the Universidad
Politécnica de Madrid (UPM) was used; the calculation of seismic hazard was carried out
with the CRISIS 2012 v5.5 software (Ordaz et al. 2013) and ArcGIS 10.1 was accessed for
the management of graphic and alphanumeric databases.

Acknowledgments The authors express their gratitude to the following institutions that provided support
on specific themes for the realization of this study, in working meetings held in Guayaquil (March 2014),
Quito (March, April, and July, August 2014), and Madrid (September 2014): Centro de Investigaciones
Cientı́ficas of the Universidad de las Fuerzas Armadas ESPE of Ecuador, Instituto Oceanográfico de la
Armada of Ecuador, Instituto Geográfico Militar of Ecuador, Instituto Espacial Ecuatoriano, and Instituto
Nacional de Investigación Geológico, Minero, Metalúrgico of Ecuador; they also express their gratitude to
Dr. Roberto Aguiar, Dr. Chunga Kervin, and to the members of the Grupo de Investigación de Ingenierı́a
Sı́smica of the Universidad Politécnica de Madrid: Yolanda Torres and Dr. Alicia Rivas. H. Parra works in
the laboratories of the Grupo de Investigación de Ingenierı́a Sı́smica of the Universidad Politécnica de
Madrid, with a scholarship awarded by the Secretarı́a Nacional de Educación Superior, Ciencia, Tecnologı́a
e Innovación of Ecuador; the reviews provided by the anonymous referees and especially the thorough and
constructive review of Dr. D. Slejko, which the authors sincerely appreciate.

Compliance with ethical standards

Conflict of interest The authors declare that they have no conflicts of interest.

References
Aguiar R, Castro C, Garzón Ch, Yanchatuña W, Cumbal L, La Fave J (2009) Magnitud máxima en zonas
fuentes para estudios de peligrosidad sı́smica del Ecuador. Rev Cienc ESPE Univ Fuerzas Armadas
12(2):109–121
Aguiar R, Rivas A, Benito M, Gaspar J, Trujillo S, Arciniegas S, Villalva P, Parra H (2014) Aceleraciones
Registradas y Calculadas del Sismo del 12 de agosto de 2014 en Quito. Rev Cienc ESPE Univ Fuerzas
Armadas 16(2):139–153

123
Bull Earthquake Eng

Alvarado A (2012) Néotectonique et cinématique de la déformation continentale en Equateur. Thèse de

doctorat (Ph.D.), Université de Grenoble, France
Alvarado A, Audin L, Nocquet J, Lagreulet S, Segovia M, Font Y et al (2014) Active tectonics in Quito,
Ecuador, assessed by geomorphological studies, GPS data, and crustal seismicity. Tectonics
33(2):67–83. doi:10.1002/2012TC003224
Arango M, Strasser F, Bommer J, Cepeda J, Boroschek R, Hernandez D, Tavera H (2012) An evaluation of
the applicability of current ground-motion models to the South and Central American subduction
zones. Bull Seismmol Soc Am 102(1):143–168. doi:10.1785/0120110078
Baize S, Audin L, Winter T, Alvarado A, Moreno L, Taipe M et al (2015) Paleoseismology and tectonic
geomorphology of the Pallatanga fault (Central Ecuador), a major structure of the South-American
crust. Geomorphology 237:14–28
Beauval C, Yepes H, Bakun W, Egred J, Alvarado A, Singaucho J (2010) Locations and magnitudes of
historical earthquakes in the Sierra of Ecuador (1587–1996). Geophys J Int 181:1613–1633. doi:10.
1111/j.1365-246X.2010.04569.x
Beauval C, Yepes H, Palacios P, Segovia M, Alvarado A, Font Y, Aguilar J, Troncoso L, Vaca S (2013) An
earthquake catalog for seismic hazard assessment in Ecuador. Bull Seismol Soc Am 103:773–786.
doi:10.1785/0120120270
Beauval C, Yepes H, Audin L, Alvarado A, Nocquet J, Monelli D, Danciu L (2014) Probabilistic seismic-
hazard assessment in Quito, estimates and uncertainties. Seismol Res Lett 85(6):1316–1327
Benito M, Lindholm C, Camacho E et al (2012) A new evaluation of seismic hazard for the Central America
region. Bull Seismol Soc Am 102(2):504–523. doi:10.1785/0120110015
Bethoux N, Segovia M, Alvarez V, Collot J, Charvis P, Gailler A, Monfret T (2011) Seismological study of
the central Ecuadorian margin: evidence of upper plate deformation. J South Am Earth Sci 31:139–152
CEC (2001) Código Ecuatoriano de la Construcción. Requisitos generales de diseño: peligro sı́smico,
espectros de diseño y requisitos mı́nimos de cálculos para diseño sismo resistente. Registro Oficial No.
382 de 2 de agosto de 2001. http://www.normaconstruccion.ec/. Accessed 21 de enero del 2015
Chunga K (2010) Terremoti crostali e zonaoine sı́smica delĺEcuador attraverso lı́ntegrazione dei dati geo-
logici, sismologici e morfostrutturali. Tesis Doctoral, Università degli Studi dell’Insubria, Sede di
Como, Italia
Chunga K, Michetti A, Pazmiño N, Martillo C, Romero A, Quiñonez M, di Geologia Ambientale G (2009)
Estimación de máximos niveles de sismicidad para el litoral ecuatoriano a través de la integración de
datos geológicos, sismológicos y sismotectónicos. Int J Oro y Pet 19:46–57
Collot J, Michaud F, Alvarado A, Marcaillou B, Sosson M, Ratzov G, Pazmino A (2009) Visión general de
la morfologı́a submarina del margen convergente de Ecuador-Sur de Colombia: Implicaciones sobre la
transferencia de masa y la edad de la subducción de la Cordillera de Carnegie. In: Collot J, Sallares V,
Pazmiño N (eds) Geologı́a y Geofı́sica Marina y Terrestre del Ecuador, 1st edn. Argudo & Asociados,
Guayaquil-Ecuador, pp 47–74
Dimaté C, Drake L, Yepez H, Ocola L, Rendon H, Grünthal G, Giardini D (1999) Seismic hazard
assessment in the Northern Andes (PILOTO Project). Ann Geofis 42(6):1039–1055
Egüez A, Alvarado A, Yepes H, Machette M, Costa C, Dart R (2003) Database and Map of Quaternary
faults and folds of Ecuador and its offshore regions. Open-File report. USGS-International Lithosphere
program. Accesed 11 Apr 2015
Flüh R, Bialas J et al (2001) South American lithospheric transects across volcanic ridges. In: Flüh E, Bialas
J, Charvis P (eds) GEOMAR report 101, Guayaquil-Ecuador, pp 4–74. ISSN: 0936-5788
Gardner JK, Knopoff L (1974) Is the sequence of earthquakes in southern California, with aftershocks
removed, Poissonian? Bull Seismol Soc Am 64(5):1363–1367
Goyes P (2009) Fondos Marinos de Soberanı́a y Jurisdicción del Ecuador. In: Collot J, Sallares V, Pazmiño
N (eds) Geologı́a y Geofı́sica Marina y Terrestre del Ecuador, 1st edn. Argudo & Asociados,
Guayaquil-Ecuador, pp 131–149
Grünthal G, Bosse Ch, Sellami S, Mayer-Rosa D, Giardini D (1999) Compilation of the GSHAP regional
seismic hazard for Europe, Africa and the Middle East. Ann Geofis 42(6):1215–1223
Guillier B, Chatelain J, Jaillard E, Yepes H, Poupinet G, Fels J (2001) Seismological evidence on the
geometry of the Orogenic System in central-northern Ecuador (South America). Geophys Res Lett
28:3749–3752
Gutenberg B, Richter C (1944) Frecuency of earthquakes in California. Bull Seismol Soc Am 34:185–188
Gutscher MA, Malavielli J, Lallemand S, Collot JY (1999) Tectonic segmentation of the North Andean
margin: impact of the Carnegie Ridge collision. Earth Planet Sci Lett 168:255–270
Hayes G, Wald D, Johnson R (2012) Slab1. 0: a three-dimensional model of global subduction zone
geometries. J Geophys Res. doi:10.1029/2011JB008524
IGM (2013) Instituto Geográfico Militar-Atlas del Ecuador. IGM, Quito-Ecuador

123
 Bull Earthquake Eng

IGEPN, Instituto Geofı́sico de la Politécnica Nacional de Quito Ecuador. http://www.igepn.edu.ec

INEC, Instituto Nacional de Estadı́stica y Censos. http://www.ecuadorencifras.gob.ec. Accessed 30 Nov
2014
INIGEMM (2014) Instituto Nacional de Investigación Geológico, Minero, Metalúrgico of Ecuador. http://
www.inigemm.gob.ec. Maps provided for this study
INOCAR (2014) Instituto Oceanográfico de la Armada of Ecuador. http://www.inocar.mil.es
ISC, International Seismological Centre. http://www.isc.ac.uk. Accessed 01 July 2014
Manchuel K, Pontoise B, Béthoux N, Régnier M, Font Y et al (2009) Sismicidad e implicaciones estruc-
turales en el área de Esmeraldas (Norte de Ecuador): a partir de los experimentos Sublime y Esmer-
aldas. In: Collot J, Sallares V, Pazmiño N (eds) Geologı́a y Geofı́sica Marina y Terrestre del Ecuador,
1st edn. Argudo & Asociados, Guayaquil-Ecuador, pp 167–179
McVerry G, Zhao J, Abrahamson N, Somerville P (2006) New Zealand acceleration response spectrum
attenuation relations for crustal and subduction zone earthquakes. Bull N Z Soc Earthq Eng 39(1):1–58
Michaud F, Pazmiño N, Collot J (2009) El karst submarino de mega depresiones circulares de la Cordillera
de Carnegie (Ecuador): posible origen por disolución submarina. In: Collot J, Sallares V, Pazmiño N
(eds) Geologı́a y Geofı́sica Marina y Terrestre del Ecuador, 1st edn. Argudo & Asociados, Guayaquil-
Ecuador, pp 29–45
NEC-11 (2015) Norma Ecuatoriana de la Construcción. Registro Oficial No. 413 del 10 de enero de 2015.
http://www.normaconstruccion.ec/. Accessed 21 de enero del 2015
NEIC-PDE, National Earthquake Information Center Premliminary Determination of Epicenters. http://
earthquake.usgs.gov/. Accessed 02 July 2014
Nocquet J, Mothes P, Alvarado A (2009) Geodesia, geodinámica y ciclo sı́smico en Ecuador. In: Collot J,
Sallares V, Pazmiño N (eds) Geologı́a y Geofı́sica Marina y Terrestre del Ecuador, 1st edn. Argudo &
Asociados, Guayaquil-Ecuador, pp 83–94
Nocquet J, Villegas-Lanza J, Chlieh M, Mothes P, Rolandone F, Jarrin P et al (2014) Motion of continental
slivers and creeping subduction in the northern Andes. Nat Geosci 7:287–291. doi:10.1038/NGEO2099
Norma E.030, Diseño Sismoresistente, Servicio Nacional de Capacitación para la Industria de la Con-
strucción del Gobierno del Perú. http://www.sencico.gob.pe. Accessed 21 Jan 2015
NSR-10, Reglamento Colombiano de la Construcción Sismoresistente, Cámara Colombiana de la Con-
strucción. http://camacol.co. Accessed 21 Jan 2015
Ordaz M, Martinelli F, D’Amico V, Meletti C (2013) CRISIS2008: a flexible tool to perform probabilistic
seismic hazard assessment. Seismol Res Lett 84–3:495–504
Ortiz O (2013) Sismotectónica y Peligrosidad Sı́smica en Ecuador. Tesis de Master, Universidad Com-
plutense de Madrid, Facultad de Ciencias Geológicas, Madrid-España
Pedoja K, Dumont JF, Ortlieb L (2009) Levantamiento Cuaternario costero del Arco de Talara (Ecuador y
norte del Perú): cuantificaciones con las secuencias de terrazas marinas. In: Collot J, Sallares V,
Pazmiño N (eds) Geologı́a y Geofı́sica Marina y Terrestre del Ecuador, 1st edn. Argudo & Asociados,
Guayaquil-Ecuador, pp 107–129
Penington W (1981) Subduction of the Eastern-Panama basin and seismotectonics of northwestern South
America. J Geophys Res 86:10753–10770
Pontoise B, Monfret T (2004) Challow seismogenic zone detedted from an offshore-onshore temporary
seismic network in the Esmeraldas area (Northern Ecuador). Geochem Geophys Geosyst
5(Q0209):1–22. doi:10.1029/2003Gc000561
Reasenberg P (1985) Second order moment of central California seismicity, 1969–1982. J Geophys Res
90:5479–5495
Rivadeneira F, Segovia M, Alvarado A, Egred J, Troncoso L, Vaca S, Yepes H (2007) Breves fundamentos
sobre los terremotos en el Ecuador. Instituto Geofı́sico Escuela Politécnica Nacional, Quito-Ecuador
Segovia M, Alvarado A (2009) Breve análisis de la sismicidad y del campo de esfuerzos en el Ecuador. In:
Collot J, Sallares V, Pazmiño N (eds) Geologı́a y Geofı́sica Marina y Terrestre del Ecuador, 1st edn.
Argudo & Asociados, Guayaquil-Ecuador, pp 131–149
Shedlock K, Tanner J (1999) Seismic hazard map of the western hemisphere. Ann Geofis 42(6):1199–1214
Stepp J (1972) Analysis of completeness of the earthquake sample in the Puget Sound area and its effects on
statistical estimates of earthquake hazard. In: Proceeding of the first International confernce on
Microzonazion vol 2, pp 897–910
Stirling M, Rhoades D, Berryman K (2002) Comparison of earthquake scaling relations derived from data of
the instrumental and preinstrumental era. Bull Seismol Soc Am 92(2):812–830
Tanner J, Shepherd J (1997) Seismic Hazard in Latin America and the Caribbean, Pan American Institute of
Geography and History, Final report vol 1, pp 1–140
Trenkamp R, Kellogg J, Freymuller J, Mora H (2002) Wide plate margin deformation, South Central
America and Northwestern South America, CASA GPS observations. J Soth Am Earth Sci 15:157–171

123
Bull Earthquake Eng

Utsu T (2002) Relationships between magnitude scales. In: Lee W, Kanamori H, Jennings P, Kisslinger C
(eds) International handbook of earthquake & engineering seismology, part A. Elseiver, Amsterdam,
pp 733–746
Vallée M, Nocquet J, Battaglia J, Font Y, Segovia M et al (2013) Intense interface seismicity triggered by a
shallow slow slip event in the Central Ecuador subduction zone. J Geophys Res 118(6):2965–2981
Vera R (2013) Geology of Ecuador. Gráficas Iberia, Quito-Ecuador
Weichert H (1980) Estimation of the earthquake recurrence parameters for unequal observation periods for
different magnitudes. Bull Seismol Soc Am 70(4):1337–1346
Wells D, Coppermith K (1994) New empirical relationships among magnitude, rupture length, rupture
width, rupture area, and surface displacement. Bull Seismol Soc Am 84(4):974–1002
Witt C, Bourgois J (2009) Relaciones entre la evolución de la cuenca del Golfo de Guayaquil-Tumbes y el
escape del Bloque Nor-Andino. In: Collot J, Sallares V, Pazmiño N (eds) Geologı́a y Geofı́sica Marina
y Terrestre del Ecuador, 1st edn. Argudo & Asociados, Guayaquil-Ecuador, pp 95–106
Witt C, Bourgois J, Michaud F, Ordoñez M, Jiménez N, Sasson M (2006) Development of the Gulf of
Guayaquil (Ecuador) during the Quaternary as an effect of the North Andean block tectonic escape.
Tectonics 25(TC3017):1–22. doi:10.1029/2004TC001723
Youngs R, Chiou S, Silva W, Humphrey J (1997) Strong ground motion attenuation relationships for
subduction zone earthquakes. Seismol Res Lett 68(1):58–73
Zhao J, Zhang J, Asano A, Ohno Y, Oouchi T, Takahashi T, Fukushima Y (2006) Attenuation relations of
strong ground motion in Japan using site classification based on predominant period. Bull Seismol Soc
Am 96(3):898–913. doi:10.1785/0120050122

123