Seismic Hazard Assessment of the San Ramón

Fault in Santiago, Chile

Horacio Dominguez

Universidad Técnica Federico Santa María: Universidad Tecnica Federico Santa Maria
https://orcid.org/0009-0001-2043-5154
Pablo Heresi
Universidad de Chile
Gonzalo Montalva
Universidad de Concepcion

Research Article

Keywords: Deterministic Seismic Hazard Analysis, Probabilistic Seismic Hazard Analysis, Chilean
Subduction Zone, San Ramón Fault

Posted Date: November 27th, 2024

DOI: https://doi.org/10.21203/rs.3.rs-5389631/v1

License:   This work is licensed under a Creative Commons Attribution 4.0 International License.
Read Full License
 SEISMIC HAZARD ASSESSMENT OF THE SAN
RAMÓN FAULT IN SANTIAGO, CHILE
Horacio Domínguez F.1, Pablo Heresi V.2 Gonzalo Montalva A.3
1
Universidad Técnica Federico Santa María, Santiago, Región Metropolitana, Chile, 0009-0001-2043-5154,
[email protected].
2
Universidad de Chile, Departamento de Ingeniería Civil, Santiago, Región Metropolitana, Chile, 0000-0003-
4594-0621.
3
Universidad de Concepción, Departamento de Ingeniería Civil, Concepción, Región del Biobío, Chile, 0000-
0001-8598-7120.

Acknowledgments
The authors would like to acknowledge the financial support of the FONDECYT (Fondo Nacional de Desarrollo
Científico y Tecnológico) Iniciación en Investigación Grant No. 11230463 and the EASER (Evolution
Assessment of Seismic Risk) Project under Grant ACT240044 from ANID (Agencia Nacional de Investigación
y Desarrollo). Also, the corresponding author would like to acknowledge the financial support of Becas para
Estudios de Magíster Cientifico-Tecnológico from Universidad Técnica Federico Santa María. The records used
to evaluate active shallow crust models were obtained from the PEER NGA-West2 ground motion database.

Abstract
The Chilean seismic activity consists mainly of interface and inslab earthquakes caused by the subduction of
the Nazca Plate beneath the South American Plate. Nonetheless, several shallow crust faults also contribute to
the seismic hazard in the country. An example of such faults is the San Ramón Fault (SRF), an active shallow
crust system located at the eastern border of Santiago, whose seismic activity has recently been discovered. The
SRF has an average slip rate of 0.4 mm/year, and it has the potential to generate earthquakes with magnitudes
up to 𝑀𝑤 7.5. Although previous studies have examined the seismic hazard posed by the fault, a probabilistic
approach has not been undertaken. This research compares the seismic hazard posed by the subduction zone
with that generated by the SRF in Santiago, using state-of-the-art models that represent their geometry and
seismicity. We compute seismic hazard maps employing deterministic and probabilistic approaches to assess
the influence of the SRF on the study zone. Resulting hazard maps from a deterministic analysis present areas
where the intensities caused by an earthquake on the SRF are over three times greater than those from interface
and inslab events. However, under a probabilistic approach, the hazard of the SRF is less than 3% that of the
subduction zone, attributed to the fault's large return period. Finally, our results indicate that the influence of
the SRF on Santiago's seismic hazard strongly depends on the intensity measure considered.
Keywords Deterministic Seismic Hazard Analysis, Probabilistic Seismic Hazard Analysis, Chilean Subduction
Zone, San Ramón Fault.

1. Introduction
Chile is renowned for its high seismic activity. Since the advent of seismic record instrumentation,
approximately 20% of the tectonic energy released in the world has occurred within Chilean territory
(Maldonado et al. 2021). The seismic activity in Chile primarily results from its proximity to the subduction
zone of the Nazca Plate beneath the South American Plate. This subduction zone has generated several
megathrust earthquakes, including the 1960 Valdivia Earthquake, 𝑀𝑤 9.6 (Cifuentes and Silver 1989) and the
2010 Maule Earthquake, 𝑀𝑤 8.8 (Vigny et al. 2011), both being interface events. However, Chile has also been
struck by inslab events, such as the 1939 Chillán Earthquake, 𝑀𝑤 7.8 (Campos and Kausel 1990) and the 1945
Santiago Earthquake 𝑀𝑤 7.1 (Barrientos et al. 1997), and shallow crust events, such as the 1958 Las Melosas
Earthquake, 𝑀𝑤 6.3 (Campos et al. 2005), causing significant losses.

1
Santiago, Chile’s capital city, is home to approximately seven million people and represents the country's
economic center. Several earthquakes have struck the city, including the 1906 Valparaíso Earthquake, 𝑀𝑤 8.2
(Astroza 2007), the 1965 La Ligua Earthquake, 𝑀𝑤 7.4 (Norambuena 2006), the 1985 Algarrobo Earthquake,
𝑀𝑤 8.0 (Comte et al. 1986), and the events mentioned above. Each of these events has caused significant
structural damage and losses attributable to the amplification effects of the Santiago basin (Pilz et al. 2011).
Given Santiago’s seismic activity, the characterization of its seismic hazard is critical for Chile. However, the
recent discovery of the San Ramón Fault (SRF) introduces a new seismic scenario for the city. The SRF is an
active shallow crust reverse fault system located at the eastern border of Santiago. The main fault has a width
of up to 20 km and a trace length of up to 50 km (Ammirati et al. 2019), with surface expression (Estay et al.
2016). However, several authors have proposed different values for the fault’s width, length, and dip (Armijo
et al. 2010; Pérez et al. 2014; Estay et al. 2016; Ammirati et al. 2019; Santibáñez et al. 2019). Through
paleoseismologic studies, Vargas et al. (2014) identified two earthquakes of magnitude 𝑀𝑤 ~7.5 on the fault in
the past 17,000 − 19,000 years. Since 2017, at least 55% of the confirmed 30 km trace has been urbanized
(Easton et al. 2018), Santiago faces a significant seismic risk.
Shallow earthquakes occurring near populated regions, such as those that the SRF could generate, have caused
significant losses worldwide. Examples include the 1995 Kobe, Japan Earthquake, 𝑀𝑤 7.2 (Kanaori and
Kawakami 1996), the 1999 Chi-Chi, Taiwan Earthquake, 𝑀𝑤 7.6 (Shin and Teng 2001), and the recent 2023
Kahramanmaraş, Turkey Earthquake, 𝑀𝑤 7.8 (Dilsiz et al. 2023), all resulting in major structural damage,
several collapses, and significant economic losses. Considering the location and seismic potential of the SRF,
characterizing its seismic hazard is crucial for the safety of Santiago’s population.
Due to the significance of the SFR, its seismic hazard has been analyzed within a deterministic approach (Estay
et al. 2016; Ammirati et al. 2019), yet it has not been studied under a probabilistic approach. Furthermore, the
seismic hazard of the SRF has not been compared to the one posed by interface and inslab events. This
comparison is particularly relevant, as hazard maps of Chilean seismic design codes only consider interface
events for latitudes north of 50°S (about 90% of the Chilean continental territory), thus neglecting the SRF and
potentially underestimating the seismic hazard in Santiago.
In this article, we compare the seismic hazard posed by the SRF with the one generated by interface and inslab
events. We compute response spectra, hazard curves, and hazard maps using state-of-the-art geometry and
seismic potential models of each source. Hazard maps enable us to identify regions of Santiago where the
influence of the SRF is greater than that of interface and inslab events. These zones are referred to as influence
zones of the SRF hereafter.

2. The San Ramón Fault

The San Ramón Fault (SRF) was initially proposed by Brüggen (1950) as a potentially active normal fault of
approximately 40 km long at the piedmont of the Andes mountain chain. However, it was not until 2002 that it
was confirmed active and identified as a west-verging reverse fault (Rauld 2002). Tectonic activity in the SRF
has been confirmed by various authors (Pérez et al. 2014; Vargas et al. 2014; Estay et al. 2016).
Since the discovery of seismic activity in the SRF, several studies have been conducted to characterize its
geometric and seismic potential parameters. These include different techniques, such as analysis of
geomorphological and geophysical evidence (Armijo et al. 2010), paleoseismology (Vargas et al. 2014), and
hypocenter detection of minor events (Pérez et al. 2014; Estay et al. 2016; Ammirati et al. 2019).
Both geometric and seismic potential parameters of the SRF still have significant uncertainties. The length of
the fault trace was estimated at 30 km by Pérez et al. (Pérez et al. 2014); however, it is also thought to be
composed of four independent segments of 10 km (Estay et al. 2016) or even have a total length of 50 km (i.e.,
20 km further south added to the one proposed by Pérez et al. (2014)) (Ammirati et al. 2019). Its width has
been proposed to range between 30 − 40 km (Armijo et al. 2010), 10 − 15 km (Estay et al. 2016) and 20 km
(Ammirati et al. 2019). Likewise, the estimated dip angle of the fault ranges between 30 − 65° (Armijo et al.

2
2010; Pérez et al. 2014; Estay et al. 2016; Ammirati et al. 2019; Santibáñez et al. 2019). As for the potential
seismicity of the SRF, Armijo et al. (2010) proposed a return period of 2,500 − 10,000 years based on its slip
rate. Nonetheless, using paleoseismology techniques, Vargas et al. (2014) detected two significant events in the
past 8,000 and 17,000 − 19,000 years, suggesting a mean return period of 9,000 ± 500 years. These
earthquakes had coseismic slips of 4.9 ± 0.7 m and 4,7 m, consistent with moment magnitudes of
approximately 𝑀𝑤 7.5, the same value proposed by Ammirati et al. (2019), and similar to the 𝑀𝑤 6.9 − 7.4
proposed by Armijo et al. (2010). However, based on their proposed geometry, Estay et al. (2016) estimated a
maximum magnitude of 𝑀𝑤 6.2 − 6.7.
To the best of our knowledge, the seismic hazard posed by the SRF has been analyzed in two studies, both using
a deterministic approach. Estay et al. (Estay et al. 2016) considered four independent ruptures with events of
magnitude 𝑀𝑤 6.7. The resulting hazard maps show peak ground accelerations (𝑃𝐺𝐴) over 0.8 g, exceeding
the PGA values recorded in Santiago during the 2010 Maule Earthquake (𝑀𝑤 8.8). On the other hand, Ammirati
et al. (2019) considered a rupture of the entire SRF with a magnitude 𝑀𝑤 7.5 event, estimating expected 𝑃𝐺𝐴
values of ~0.8 g near the fault’s scarp.
Despite the seismic hazard that the SRF could pose, no regulation prohibits construction near the fault trace. In
fact, by the year 2017, 55% of the 30 km trace proposed by Pérez et al. (2014) had been urbanized (Easton et
al. 2018). Moreover, critical infrastructure such as three hospitals, several educational centers, a nuclear study
center, and crowded places such as commercial centers are located within 2 km of the fault. Since the SRF can
produce surface ruptures, nearby structures are at high risk of suffering considerable damage or collapse.
As mentioned before, seismic design maps of Chilean codes only consider interface events. Thus, comparing
the seismic hazard posed by the SRF with that generated by interface and inslab events enable us to examine if
there is an underestimation of the seismic hazard in Santiago. Moreover, the study of the SRF is even more
critical considering that, north of latitude 50°S (i.e., 90% of the Chilean territory), it represents the most
significant active shallow crust fault and is located farthest from the subduction trench. Consequently, we expect
the SRF’s influence zone to be an upper bound for the influence zones of other shallow crust faults north of
latitude 50°S.

3. Methodology
The study zone considered is the one investigated by Bustos et al. (2023). Their study estimated the time-
averaged shear wave velocity in the top 30 m (𝑉𝑆30 ) and the predominant period of the soil (𝑇 ∗ ) of 44 sites
between UTM coordinates 6,311,600 − 6,266,400 𝑁 and 318,100 − 360,900 𝐸. Using a moving average,
they interpolated the soil parameters to 185,661 sites in a grid with a separation distance between sites of 100
m. Figure 1 shows maps of both 𝑇 ∗ and 𝑉𝑆30 , as well as two study sites selected for the computation of response
spectra and hazard curves in subsequent sections. Study Site 1 is located in the municipality of Ñuñoa, close to
downtown Santiago, whereas Site 2 is located over the SRF’s trace. Table 1 provides the coordinates and the
site parameters of these two sites. Additionally, Figure 1 displays the trace of the SRF with its extension of 30
km (Armijo et al. 2010) and potential 20 km further to the south (Ammirati et al. 2019).

3
Fig. 1 Site dynamic parameters within the study zone (Bustos et al. 2023). (a) Predominant period of the site,
𝑇 ∗ ; (b) Time-averaged shear wave velocity in the top 30 m, 𝑉𝑆30
Table 1 Coordinates and site parameters of the two study sites.

𝒎
Study Site Longitude Latitude 𝑻∗ [𝒔] 𝑽𝑺𝟑𝟎 [ ] 𝒁𝟏.𝟎 [𝒎] 𝒁𝟐.𝟓 [𝒎]
𝒔
1 33.47°S 70.61°W 0.77 579 46.52 269.33
2 33.47°S 70.52°W 0.39 415 66.09 66.09

We computed the seismic hazard on the 46,631 sites within the study zone using both deterministic seismic
hazard analysis (DSHA) (Krinitzsky 2002) and probabilistic seismic hazard analysis (PSHA) (Esteva 1967;
Cornell 1968). Results on every site enable us to obtain hazard maps within the study zone. In addition, we
generated maps presenting the ratio between the seismic hazard of the SRF and that of subduction events. In
these maps, the influence zones of the SRF correspond to the regions where this ratio is greater than 1 (i.e., the
hazard imposed by the SRF is larger than the one resulting from subduction events).
To compute seismic hazard, we selected ground motion models (GMM) following the criteria proposed by
Bommer et al. (2010) for each seismic source type (i.e., active shallow crust, interface, and inslab). The log-
likelihood methodology proposed by Scherbaum et al. (2009) was used to compute the weights assigned to each
model within a logic tree, depending on the intensity measure considered (i.e., PGA and spectral accelerations).
To compute the log-likelihood of interface and inslab ground motion models, records from Chilean subduction
events with 𝑀𝑤 > 6.0 since 2010 were considered. Chilean models, such as IRRP17 (Idini et al. 2017) and
MBR17 (Montalva et al. 2017), and NGA-Sub models, such as AG20 (Abrahamson and Gülerce 2022),
KBCG20 (Kuehn et al. 2023) and PSHAB20 (Parker et al. 2021), were evaluated. However, IRRP17 (Idini et
al. 2017) model was discarded for inslab events, as it presents a poor constraint for high-magnitude events.
Regarding active shallow crust models, NGA-West2 (Ancheta et al. 2012) records with similar soil parameters

4
as those observed in the study zone (see Figure 1), Joyner-Boore distances (Joyner and Boore 1988) lower than
40 km and 𝑀𝑤 > 6.5 were considered. NGA-West2 models were evaluated (i.e., ASK14 (Abrahamson et al.
2014), BSSA14 (Boore et al. 2014), CB14 (Campbell and Bozorgnia 2014) and CY14 (Chiou and Youngs
2014)) as well as ASB14 (Akkar et al. 2014), Bindi14 (Bindi et al. 2014) and Zhao16 (Zhao et al. 2016). Figure
2 presents the cumulative weights obtained for each GMM, depending on the source type.

Fig. 2 Cumulative weights of the ground motion models computed from the log-likelihood methodology proposed by
Scherbaum et al. (2009) for (a) interface events; (b) inslab events; (c) shallow crust events.

3.1. Considerations for Deterministic Seismic Hazard Analysis

The deterministic event for the SRF is selected based on the worst-case scenario. Thus, a magnitude of 𝑀𝑤 7.5
is considered, given a rupture area of 50𝑥20 km2 dipping with an angle of 40° to the east and a coseismic slip
of 4.8 m (Ammirati et al. 2019). For interface and inslab events, two different scenarios are considered: (1) a
historical Chilean earthquake scenario; and (2) the worst-case scenario. The selected historical interface event
is the 2010 Maule Earthquake, 𝑀𝑤 8.8 (Vigny et al. 2011), yet it is placed in front of the study zone. The worst-
case scenario for interface events is a 𝑀𝑤 9.3 earthquake, also placed in front of the study zone; this corresponds
to the maximum magnitude of the zone according to the recurrence model proposed by Poulos et al. (2019)
(Zone 2 in Table 2). Both event geometries are computed from the Slab2 model of the Nazca subduction zone
(Hayes et al. 2018). On the other hand, the historical inslab event chosen is the 1939 Chillán Earthquake, 𝑀𝑤 7.8
(Campos and Kausel 1990), whereas the worst-case scenario of inslab events is a 𝑀𝑤 8.5 earthquake according
to the inslab zone proposed by Poulos et al. (2019) (Zone 7 in Table 2). Both inslab events are placed under the
study zone at coordinates 31.51°𝑆, 70.67°𝑊, and a depth of 71 km, based on the Slab2 model (Hayes et al.
2018). Figure 3a presents the epicenter of inslab events, while Figure 3b illustrates the interface ruptures placed
in front of the study zone. Within the deterministic approach, all computed intensities correspond to the 84th
percentile.

5
Fig. 3 (a) Study zone to compute hazard maps and study sites to compute response spectra and hazard curves. The pink
rectangle represents the study zone, the red continuous line corresponds to the 30 km long trace of the SRF (Armijo et al.
2010), the red dashed line represents the 50 km trace of the SRF (Ammirati et al. 2019), and the target symbol the epicenter
of the inslab events for deterministic assessments. (b) Ruptures for 𝑀𝑤 9.3 (yellow dashed area) and 𝑀𝑤 8.8 (white and
dashed blue area) placed in front of the study zone for deterministic assessments

To compute hazard maps, acceleration spectral ordinates at vibration periods of 0.3 s and 3.0 s, denoted Sa(0.3
s) and Sa(3.0 s), respectively, are considered. These values are selected to study the impact of the vibration
period on the influence zones of the SRF. Hazard ratio maps will be computed as the ratio of the 84 th percentile
intensity of the SRF and the one of the subduction events for each intensity measure.

3.2. Considerations for Probabilistic Seismic Hazard Analysis

We employed logic trees to represent the SRF’s geometric and recurrence models to consider epistemic
uncertainties. In the first place, the length of the fault trace is split into two branches of 30 km and 50 km, with
weights of 0.3 and 0.7. We assign more weight to the 50 km branch as there is a consensus on this value. The
30 km branch is further divided into two branches with equal weight to consider the maximum magnitude of
the fault. The possible maximum magnitudes are 𝑀𝑤 6.9, based on the fault’s length (Thingbaijam et al. 2017),
and 𝑀𝑤 7.3, based on the seismic moment of a 30𝑥20 km2 rupture area with a coseismic slip of 4.8 m. Similarly,
the 50 km branch splits into maximum magnitudes of 𝑀𝑤 7.3 and 𝑀𝑤 7.4.
The fault is assumed to follow a characteristic event recurrence model (Schwartz and Coppersmith 1984) with
a constant recurrence rate between 𝑀𝑤 6.6 and the maximum magnitude. Therefore, each model branch is
divided into three new branches corresponding to return periods of 8,500, 9,000, and 9,500, according to
Vargas et al. (2014), with weights of 0.25, 0.50, and 0.25, respectively. As discussed by Estay et al. (2016),
although a characteristic event recurrence model is not demonstrated for the SRF, it provides a valid first-order
approximation of its potential seismic hazard. For each resulting model, we considered unique values of 40°
east dipping and 20 km width. The logic tree for the SRF model is presented in Figure 4. Considering the seven

6
ground motion models for active shallow crust events (see Figure 2), the SRF model is compounded by 84 logic
tree branches for each intensity measure.

Fig. 4 Logic tree of the SRF geometry and seismicity for PSHA calculations. Each model considers a 40° dip angle, a 20
km rupture width, and 𝑀𝑤 6.6 minimum magnitude for a characteristic event recurrence model

For the subduction zone, we considered a unique model geographically divided into seven zones as proposed
by Poulos et al. (2019). However, only zones 2, 3, 6 and 7 were used as a site-to-source distance limit of 600
km is considered. Recurrence models for each zone are also obtained from Poulos et al. (2019), where bounded
Gutenberg-Richter models are fitted with a minimum magnitude of 𝑀𝑤 5.0. The parameters for the recurrence
law of each zone are shown in Table 2. Finally, the subduction geometry is obtained from the Slab2 model
(Hayes et al. 2018). Given that the subduction model is made up from interface and inslab sources, its logic tree
is compounded by 30 logic tree branches (i.e., 5 GMMs for inslab events and 6 GMMs for interface events).
Hazard maps obtained from the PSHA methodology will display the probability of exceedance of an intensity
level in a time window of 50 years. Similar to the DSHA methodology, spectral accelerations with vibration
periods of 0.3 s and 3.0 s, Sa(0.3 s) and Sa(3.0 s), are selected as intensity measures, and 1.0 g and 0.1 g as
their respective intensity levels. Hazard ratio maps of each intensity measure will be computed from the
exceedance rate of the corresponding intensity level of the SRF and the subduction events.
Table 2 Bounded Gutenberg-Richter model parameters for subduction seismic zones (Poulos et al. 2019). The minimum
magnitude for all the zones is 𝑀𝑤 5.0.

Seismic Zone Event Type a b Maximum 𝑴𝒘

1 Interface 5.15 0.87 9.2
2 Interface 5.36 0.88 9.3

7
 3 Interface 4.26 0.77 9.6
4 Inslab 5.11 0.90 8.6
5 Inslab 6.37 1.04 8.4
6 Inslab 6.12 1.11 8.4
7 Inslab 5.32 0.98 8.5

4. Results
4.1. Deterministic Seismic Hazard Analysis
We computed response spectra at Sites 1 and 2 and hazard maps within the study zone through DSHA. For the
SRF, we only considered a worst-case scenario event, whereas, for interface and inslab events, we included
both historical and worst-case earthquakes in our analysis.
Figure 5 illustrates the response spectra at both study sites for the worst-case scenario for the SRF and the
historical earthquake magnitude scenario for interface and inslab events. For vibration periods shorter than 0.9
s, the seismic hazard at Site 1 is mainly controlled by the inslab event, while for longer periods, it is controlled
by the SRF. At Site 2, where the hazard posed by the SRF significantly increased compared to Site 1, the inslab
event poses a greater hazard than that generated by the SRF at vibration periods between 0 − 0.1 s, whereas for
longer vibration periods, the hazard is clearly controlled by SRF. Thus, influence zones of the SRF will appear
for vibration periods longer than 0.1 s and reach Site 1 for vibration periods over 0.9 s. Note that at both sites,
the interface event does not control the hazard at any period.

Fig. 5 Deterministic response spectra for both study sites considering the worst-case scenario for the SRF, 𝑀𝑤 7.5, and
historical earthquake magnitude scenarios for interface event, 𝑀𝑤 8.8, and inslab event, 𝑀𝑤 7.8. Each map presents the
location of the corresponding study site and the inslab epicenter (see Figure 3 for the location of interface rupture)

Figure 6 presents response spectra for both study sites considering the worst-case scenario for the SRF, interface
event, and inslab event. Compared to Figure 5, response spectra of interface and inslab events are greater, as
expected. In this case, the hazard at Site 1 is controlled by the inslab event for vibration periods shorter than
1.8 s. At Site 2, the SRF hazard is greater than that of interface and inslab events for vibration periods longer
than 0.13 s. Consequently, influence zones of the SRF will manifest for periods longer than 0.13 s and will not
reach Site 1 for periods shorter than 1.8 s when considering worst-case scenarios for interface and inslab events.

8
As in Figure 5, the interface event has no influence on the seismic hazard at both study sites under a
deterministic approach.

Fig. 6 Deterministic response spectra for both study sites considering the worst-case scenario for the SRF, 𝑀𝑤 7.5, interface
event, 𝑀𝑤 9.3, and inslab event, 𝑀𝑤 8.5. Each map presents the location of the corresponding study site and the inslab
epicenter (see Figure 3 for the location of interface rupture)

Figure 7 shows hazard maps derived from DSHA across the entire study zone, considering historical earthquake
magnitudes for interface and inslab events. In particular, the 84th percentile intensity for Sa(0.3 s) is displayed
for (a) the SRF and (b) the envelope of the interface and inslab events. Notably, a 𝑀𝑤 7.5 event on the SRF
could cause Sa(0.3 s) intensities greater than 2.4 g in a zone that covers up to 5 km west from the 30 km fault
trace, reaching values of over 3.4 g at the north of the trace. At Site 2 the 𝑀𝑤 7.5 event on the SRF could trigger
Sa(0.3 s) intensities similar to 3.0 g. It is noted that the SRF hazard decreases rapidly with source-to-site
distance, reaching a value of approximately 1.8 g at Site 1. In contrast, the subduction hazard map presents a
more uniform Sa(0.3 s) intensity, as almost the entire study zone is in the 1.8 − 2.6 g range. The circular-alike-
shaped intensity contours suggest that the inslab event hazard is greater than that posed by the interface event.
The 84th percentile intensity at Site 1 is greater than 2.2 g, thus greater than that posed by the SRF. However, at
Site 2, the hazard posed by the subduction events is lower than that produced by the SRF. This means that an
influence zone of the SRF is expected to appear but not reach Site 1. This is consistent with what was seen in
Figure 5, where influence zones of the SRF would appear for vibration periods longer than 0.1 s and reach Site
1 for vibration periods longer than 0.9 s.

9
Fig. 7 Deterministic seismic hazard maps for Sa(0.3 s) considering the worst-case scenario of the SRF, 𝑀𝑤 7.5, and the
historical earthquake magnitude scenario for interface event, 𝑀𝑤 8.8, and inslab event, 𝑀𝑤 7.8

A map of the hazard ratio between the SRF and subduction events for Sa(0.3 s) is presented in Figure 8,
considering the worst-case scenario of the SRF and historical earthquake magnitudes for subduction events.
Yellow lines indicate a ratio value of 1, meaning equal hazard posed by the SRF and the envelope of subduction
events. Therefore, zones between the fault and the yellow lines represent the influence zone of the SRF for
Sa(0.3 s) under a deterministic approach. This influence zone has a width of approximately 4 km west of the
fault trace, opposite to the fault dip direction. The influence zone covers Site 2, where the hazard posed by the
SRF is 140% of that generated by the subduction events. However, at Site 1, the SRF hazard is lower than 80%
of that of subduction events. Another influence zone can be seen at the northern area of the study region. This

10
influence zone is produced by site effects, as this area corresponds to that with small 𝑉𝑆30 values, of
approximately 280 m/s, and a long predominant period of the soil, between 2.2 and 4.2 s (see Figure 1).

Fig. 8 Map of deterministic hazard ratio for Sa(0.3 s) considering the worst-case scenario of the SRF, 𝑀𝑤 7.5, and the
historical earthquake magnitude scenario for interface event, 𝑀𝑤 8.8, and inslab event, 𝑀𝑤 7.8. Yellow lines represent the
limits of the influence zone of the SRF

The 84th percentile intensity for Sa(0.3 s) is shown in Figure 9, considering the worst-case scenario of the SRF
and the envelope of the worst-case scenario of subduction events. As expected, a considerable increase in the
seismic hazard is observed on the subduction map. Due to subduction events, Sa(0.3 s) intensities within almost
the entire study zone are over 2.5 g. The intensity at Site 1 is over 3.2 g, higher than the 1.8 g posed by the SRF
event. However, at Site 2, the intensity is approximately 2.8 g, lower than the Sa(0.3 s) generated by the SRF,
3.0 g. Therefore, same as when considering historical-case scenarios for subduction events, an influence zone
is expected when considering worst-case scenarios, yet it is not expected to reach the proximities of Site 1.

11
Fig. 9 Deterministic seismic hazard maps for Sa(0.3 s) considering the worst-case scenario of the SRF, 𝑀𝑤 7.5, interface
event, 𝑀𝑤 9.3, and inslab event, 𝑀𝑤 8.5

Figure 10 presents the hazard ratio maps considering the worst-case scenario of the SRF and subduction events.
In this case, three influence zones of the SRF are observed. The two greater zones are primarily produced by
the short site-to-source distance from the SRF, as these are located at the proximities of its trace. Within the
three influence zones, the hazard ratio reaches values over 120%. This width of the greatest zone is
approximately 4.0 km to the west of the trace. The smallest of the three influence zones is produced by site
effects.

12
These influence zones indicate that the intensities due to the SRF event could reach similar values to those
generated by a 𝑀𝑤 8.5 inslab event under Santiago at certain areas of the city. However, the hazard ratio
decreases rapidly with the distance from the SRF trace, as the seismic hazard posed by the fault also decreases
with source-to-site distance. In contrast, the hazard from subduction events is slightly uniform within the study
zone.

Fig. 10 Map of deterministic hazard ratio for Sa(0.3 s) considering the worst-case scenario of the SRF, 𝑀𝑤 7.5, interface
event, 𝑀𝑤 9.3, and inslab event, 𝑀𝑤 8.5. No influence zone of the SRF is observed

Similar to Figure 7, Figure 11 shows hazard maps obtained through DSHA considering historical earthquake
magnitudes for interface and inslab events. However, the 84th percentile intensity of Sa(3.0 s) is displayed in
this case. It can be observed that the SRF event can generate Sa(3.0 s) intensities over 0.30 g on the proximities
of its trace. Notably, in the northern region of Santiago, Sa(3.0 s) intensities reach a value of 0.26 g due to site
effects. On the other hand, the hazard map of subduction events presents intensities ranging between 0.08 −
0.12 g, relatively low values compared to the hazard posed by the SRF. This finding indicates the presence of
an influence zone of the SRF.
Site 1 is affected by a Sa(3.0 s) intensity of approximately 0.18 g from the SRF event, whereas the subduction
events generate intensities similar to 0.10 g. Consequently, the influence zone of the SRF will reach Site 1, and
it is expected to extend beyond it, consistent with the findings in Figure 5. Additionally, at Site 2, the SRF
generates an intensity of over 0.28 g, while subduction events produce intensities of approximately 0.10 g. This
indicates that, from a deterministic approach, the seismic hazard posed by the SRF in terms of Sa(3.0 s) is up
to 2.8 times greater than that generated by subduction events in the proximities of the fault trace.

13
Fig. 11 Deterministic seismic hazard maps for Sa(3.0 s) considering the worst-case scenario of the SRF, 𝑀𝑤 7.5, and the
historical earthquake magnitude scenario for interface event, 𝑀𝑤 8.8, and inslab event, 𝑀𝑤 7.8

Figure 12 maps the hazard ratio between the SRF and subduction events for Sa(3.0 s), considering the worst-
case scenario of the SRF and historical earthquake magnitudes for interface and inslab events. As discussed
previously, a large influence zone of the SRF is observed, covering more than half of the study zone and
extending almost 30 km west of the fault trace. The hazard ratio reaches values over 3.2 near the SRF and
greater than 1.6 for more than 10 km west of the fault trace. These results notably differ from those obtained
for a vibration period of 0.3 s, highlighting the effect of the considered intensity measure when evaluating the
influence zone of the SRF.

14
Fig. 12 Map of deterministic hazard ratio for Sa(3.0 s) considering the worst-case scenario of the SRF, 𝑀𝑤 7.5, and the
historical earthquake magnitude scenario for interface event, 𝑀𝑤 8.8, and inslab event, 𝑀𝑤 7.8. Yellow lines represent the
limits of the influence zone of the SRF

Figure 13 illustrates the hazard map of Sa(3.0 s) considering the worst-case scenario of the SRF and subduction
events. Although the hazard posed by the subduction events increased with respect to the historical earthquake
magnitude events, the maximum intensity is 0.18 g, still lower than the maximum intensity generated by the
𝑀𝑤 7.5 event on SRF, which is over 0.32 g.

15
Fig. 13 Deterministic seismic hazard maps for Sa(3.0 s) considering the worst-case scenario of the SRF, 𝑀𝑤 7.5, interface
event, 𝑀𝑤 9.3, and inslab event, 𝑀𝑤 8.5

Figure 14 maps the hazard ratio between the SRF and subduction events for Sa(3.0 s), considering the worst-
case scenario of the SRF and the subduction events. The influence zone of the SRF covers over 12 km west of
the fault trace and even over 20 km in the northern region. From Figure 13, the Sa(3.0 s) caused by the
subduction events at Site 1 is approximately 0.15 g, lower than the 0.18 g generated by the SRF. Consequently,
the hazard ratio at Site 1 is similar to 1.2. As for Site 2, the intensity generated by the worst-case scenario of
the subduction events is again approximately 0.15 g; thus, the hazard ratio is 1.8.

16
Fig. 14 Map of deterministic hazard ratio for Sa(3.0 s) considering the worst-case scenario of the SRF, 𝑀𝑤 7.5, interface
event, 𝑀𝑤 9.3, and inslab event, 𝑀𝑤 8.5. Yellow lines represent the limits of the influence zone of the SRF

Comparing the maps of hazard ratios of Figures 8, 10, 12, and 14, we observe that although the influence zone
of the SRF from a deterministic perspective depends, as expected, on the selected events for the subduction
zone, the vibration period has a much greater impact. For Sa(0.3 s), the influence zone of the SRF reaches sites
within 4 km west of the fault trace. On the other hand, for Sa(3.0 s), the influence zone of the SRF extends up
to 10 to 30 km west of the fault trace, depending on the subduction events considered. These results indicate
that the influence zone of the SRF is greater for long-period spectral ordinates than for short-period ordinates.
Moreover, the generation of influence zones is affected by site effects, as the region that presents small 𝑉𝑆30
values and a long predominant period of the soil (northern area of Santiago, see Figure 1) is prone to have
hazard ratios over 1.0.

4.2. Probabilistic Seismic Hazard Analysis

We computed uniform hazard spectra (UHS), seismic hazard curves, and seismic hazard maps through PSHA.
The UHS is computed for both study sites and four hazard levels corresponding to probabilities of exceedance
of 0.25%, 0.5%, 2.0%, and 10.0% in 50 years (i.e., return periods of 19,975, 9,975, 2,475, and 475 years).
To evaluate the effect of the SRF on the probabilistic seismic hazard, our calculations considered two cases: (1)
both sources, the SRF and subduction events; and (2) only subduction events.
Figure 15 shows the mean UHS obtained at each study site. Results at Site 1 show a negligible effect of the
SRF on the seismic hazard, as both curves are almost equal regardless of the hazard level. Something similar is
observed at Site 2, however, a slight difference between the curves appears for very long return periods of
19,975 and 9,975 years (i.e., probabilities of exceedance of 0.25% and 0.5% in 50 years). At both study sites,
the difference between the UHS considering both seismic sources and the UHS considering only subduction
events varies along the vibration period. This suggests that the influence of the SRF, in a probabilistic approach,
depends on the vibration period, aligning with the conclusions drawn from a deterministic analysis.
In Figure 15, we compared mean UHS from the logic trees and observed a negligible influence of the SRF.
However, it is also important to explore whether a UHS of individual branches of the SRF logic tree could
exceed the UHS of certain branches for subduction events. This would indicate a non-negligible contribution

17
of the SRF, at least for some combinations of source models. To address this issue, the UHS of each branch of
the SRF and subduction logic trees are shown in Figures 16 and 17 for probabilities of exceedance of 0.25%
and 0.5% in 50 years, respectively. In these figures, light orange curves represent individual realizations of
UHS from the SRF logic tree, while light blue curves represent those from the subduction logic tree. The
continuous black line represents the mean UHS of the subduction events, and the dashed black line is the mean
UHS of the SRF.

Fig. 15 Uniform hazard spectra computed at both study sites, considering probabilities of exceedance of 0.25%, 0.5%,
2.0%, and 10.0% in 50 years. Red curves represent the UHS considering both SRF and subduction events, while blue
curves represent the UHS considering only subduction events. Each inset map presents the location of the corresponding
study site

As can be observed in Figures 16 and 17, the contribution of the SRF to the seismic hazard at both study sites
is negligible, regardless of the variations in the length of its trace, maximum magnitude, recurrence period, and
ground motion model considered. Notably, although the dispersion of the subduction UHS is attributed only to
the different ground motion models, it is considerably higher than that observed in the UHS of the SRF models,
especially on Site 1. Looking only at the SRF uniform hazard spectra, a set of realizations present considerably
higher levels of intensity than the mean value for a probability of exceedance of 0.25% in 50 years. These
realizations correspond to those computed by the ASK14 (Abrahamson et al. 2014) ground motion model. The
same is seen for a probability of 0.5% in 50 years, however the difference is considerably lower.

18
Fig. 16 Uniform hazard spectra considering a probability of exceedance of 0.25% in 50 years (i.e., 19,975 years return
period) of each logic tree branch, computed at both study sites. Light orange curves represent the different realizations of
the SRF, and light blue curves the realizations of the subduction models. The black continuous line represents the subduction
mean UHS, and the dashed curve is the SRF mean UHS. Each inset map presents the location of the corresponding study
site

Fig. 17 Uniform hazard spectra considering a probability of exceedance of 0.5% in 50 years (i.e., 9,975 years return period)
of each logic tree branch, computed at both study sites. Light orange curves represent the different realizations of the SRF,
and light blue curves the realizations of the subduction models. The black continuous line represents the subduction mean
UHS, and the dashed curve is the SRF mean UHS. Each inset map presents the location of the corresponding study site

We also computed seismic hazard curves at both study sites, considering two different vibration periods, 0.3
and 3.0 s. Three cases were taken into account: (1) both sources, the SRF and subduction events; (2) only
subduction events; and (3) only SRF events.
Figures 18 and 19 present the results of the mean seismic hazard curves for vibration periods of 0.3 and 3.0 s,
respectively. As in Figure 15, considering both seismic sources and only considering subduction events yield
very similar hazard curves for both intensities, highlighting a negligible influence of the SRF. Moreover, the

19
seismic hazard curve of the SRF presents considerably lower values of mean recurrence rate, 𝜆, than that of
subduction events. Note that the recurrence rate values on the y-axis are presented on a logarithmic scale.
At Site 1, the ratio between the mean seismic hazard curve of the SRF and the mean seismic hazard curve of
subduction events for Sa(0.3 s) reaches its maximum value for Sa(0.3 s) = 3.81 g, with a ratio of 0.0080. This
indicates that, at Site 1, the mean hazard posed by the subduction zone is at least 125 times greater than that
generated by the SRF when considering Sa(0.3 s) as the intensity measure. At Site 2, the intensity at which the
hazard ratio presents its higher value is Sa(0.3 s) = 8 g, resulting in a ratio of 0.1407. Thus, the hazard posed
by the subduction zone at Site 2 is over 7 times greater than that generated by the SRF when considering Sa(0.3
s) as the intensity measure. Note that 8 g is the upper bound of the considered intensity, thus this ratio could
increase for higher intensity levels of Sa(0.3 s).

Fig. 18 Seismic hazard curves for Sa(0.3 s) at both study sites. Red curves represent the case where both SRF and subduction
events are considered, blue curves represent the case considering only subduction events, and yellow curves the case
considering only the SRF. The gray dashed line represents the mean recurrence rate of the SRF source model, 1⁄9,000
years-1. Each inset map presents the location of the corresponding study site

Similarly, when considering Sa(3.0 s) (Figure 19), the hazard caused by subduction events at Site 1 is at least
83 times greater than that posed by the SRF. Compared to the obtained for Sa(0.3 s), this result highlights the
importance of the vibration period in the influence of the SRF. At Site 2, the hazard caused by the subduction
is over 8.5 times greater. However, same as for Sa(0.3 s), the maximum hazard ratio occurs at the upper bound
of the considered intensity, 1 g for Sa(3.0 s). Thus, in these cases, the maximum ratio is biased by the upper
bound of the intensities selected by the authors and should not be used to compare the influence of the vibration
period.
Figures 18 and 19 also display a gray dashed line representing the mean recurrence rate of the SRF source
model (see Figure 4), 1⁄9,000 years-1. As can be seen, the seismic hazard curve of the SRF is limited to this
value as an upper bound. Thus, the recurrence rate selected for the SRF, as well as the characteristic event
recurrence model, have a significant impact on the seismic hazard results under a probabilistic approach.

20
Fig. 19 Seismic hazard curves for Sa(3.0 s) at both study sites. Red curves represent the case where both SRF and subduction
events are considered, blue curves represent the case considering only subduction events, and yellow curves the case
considering only the SRF. The gray dashed line represents the mean recurrence rate of the SRF source model, 1⁄9,000
years-1. Each inset map presents the location of the corresponding study site

As with the UHS, seismic hazard curves were also computed for all realizations of both SRF and subduction
logic tree models. Figures 20 and 21 present the seismic hazard curves for vibration periods of 0.3 and 3.0 s,
respectively, considering the SRF individual branches (light orange), the subduction individual branches (light
blue), and their respective means in continuous (subduction) and dashed (SRF) black lines. In Figure 20, some
realizations of the SRF overcome the seismic hazard posed by realizations of the subduction model on Site 2.
These again correspond to the logic tree branch computed using ASK14 (Abrahamson et al. 2014) ground
motion model. Even though the same branches present higher hazard on Site 1, these do not overcome any
seismic hazard curve of the subduction model. When looking at the subduction realizations on Site 1, some
seismic hazard curves present a considerably low hazard in comparison to the other realizations. Although they
seem like a unique seismic hazard curve, they are indeed two different logic tree branches: IRRP17 (Idini et al.
2017) interface model combined with AG20 (Abrahamson and Gülerce 2022) and PSHAB20 (Parker et al.
2021) models for inslab events.
From Figure 20, it is important to note that seismic hazard curves computed for the SRF present a much higher
dispersion on both study sites than the subduction model. On the other hand, in Figure 21, the dispersion of the
seismic hazard curves is higher for the subduction model. In contrast with Figure 20, in Figure 21, a considerable
number of realizations of the subduction model are overcome by realizations of the SRF on Site 2. Even more,
the mean seismic hazard curve of the SRF overcomes some realizations of the subduction model.
Although the intersection of seismic hazard curves of the SRF and the subduction model suggests the possibility
of the SRF generating a greater seismic hazard than subduction events under a probabilistic approach, this
occurs for recurrence rates lower than 3𝑥10−5 year-1 (i.e., a return period of thirty thousand years). At Site 1,
the intersection between seismic hazard curves from the SRF and the subduction models is also seen for a few
realizations. Regarding only the subduction model, some seismic hazard curves present a considerably higher
hazard level. These curves correspond to the five seismic hazard curves computed considering the KBCG20
(Kuehn et al. 2023) model for interface events.

21
Fig. 20 Seismic hazard curves for Sa(0.3 s) considering each logic tree branch of SRF and subduction models, computed at
both study sites. Light blue curves represent the realizations of the subduction models, light orange curves represent the
different realizations of the SRF, and darker curves represent their mean hazard curves. Each inset map presents the location
of the corresponding study site

Fig. 21 Seismic hazard curves for Sa(3.0 s) considering each logic tree branch of SRF and subduction models, computed at
both study sites. Light blue curves represent the realizations of the subduction models, light orange curves represent the
different realizations of the SRF, and darker curves represent their mean hazard curves. Each inset map presents the location
of the corresponding study site

22
Figure 22 shows seismic hazard maps for the probability of Sa(0.3 s) exceeding an intensity level of 1.0 g in
50 years. The hazard posed by the SRF is significantly lower than that induced by subduction events. In
particular, the probability that the SRF causes a Sa(0.3 s) greater than 1.0 g within the study zone is
approximately 0.35%, whereas this probability reaches 28% for subduction events. Consistent with the
deterministic approach, the hazard generated by the SRF decreases rapidly with increasing source-to-site
distance. On the other hand, the hazard due to subduction events is relatively uniform within the study zone,
and it is more conditioned by site effects.

Fig. 22 Seismic hazard maps for the probability of exceedance of 1.0 𝑔 in 50 years, considering a vibration period of 0.3 𝑠,
of the SRF events and subduction zone events

23
The significantly different probability values of the hazard maps in Figure 22 suggest a lack of influence zones
of the SRF for Sa(0.3 s) within a probabilistic approach. Indeed, Figure 23 confirms that assertion, as the ratio
between the hazard posed by the fault and that of subduction events is smaller than 1 throughout the entire study
zone. Moreover, the maximum ratio is approximately 1.5%. Thus, the probability of Sa(0.3 s) exceeding 1.0 g
due to subduction events is 67 times greater than that due to SRF events on the proximities of the faults trace.
In the western region of the study zone, this probability is over 200 times greater for subduction events
compared to SRF events.

Fig. 23 Seismic hazard ratio map for the probability of exceedance of 1.0 𝑔 in 50 years, considering a vibration period of
0.3 𝑠

Figure 24 displays seismic hazard maps of the probability of Sa(3.0 s) exceeding 0.1 g in 50 years due to the
SRF and subduction events. The probability that an SRF event causes an intensity greater than 0.1 g is 0.25%
near its trace and under 0.1% for more than half of the study zone. Due to subduction events, Sa(3.0 s) has a
probability of exceeding 0.1 g in 50 years of at least 4% in the entire study zone, reaching values of over 20%
in the northern region due to the site effects of this area (see Figure 1).

24
Fig. 24 Seismic hazard maps for the probability of exceedance of 0.1 𝑔 in 50 years, considering a vibration period of 3.0 𝑠,
of the SRF events and subduction zone events

Figure 25 maps the seismic hazard ratio between the SRF and the subduction zone for Sa(3.0 s) exceeding 0.1
g. As in Figure 23, the ratio is considerably lower than 1 in the entire study zone. Therefore, no influence zone
of the SRF is displayed. Furthermore, the seismic hazard ratio only reaches 2.5% near the fault trace and
becomes lower than 0.5% for half of the study zone. This means that subduction events are 40 times more likely
to produce an Sa(3.0 s) exceeding 0.1 g in 50 years for half of the study zone. Although the ratio is still
considerably low, these results are significantly higher than those obtained for Sa(0.3 s) as the intensity measure,
once again highlighting the importance of the vibration period in the influence of the SRF.

25
Fig. 25 Seismic hazard ratio map for the probability of exceedance of 0.1 𝑔 in 50 years, considering a vibration period of
3.0 𝑠

5. Discussion
The case study of the SRF is highly relevant to Chile. Given the fault location, to the east of the country’s most
populated and economically relevant city, its trace urbanization, and possible surface ruptures, an activation of
the fault supposes a significant risk of several structural collapses and major losses. Moreover, given its location
relative to the subduction trench and its seismic potential, studying the influence zone of the SRF is expected
to provide an upper bound for the influence zones of other shallow crust faults in the country, particularly for
latitudes north of 50°S (about 90% of the Chilean continental territory).
The study zone of this investigation mainly lies to the west of the SRF trace. Given the fault’s eastward dip, the
seismic hazard decreases considerably with distance to the trace. It is important to recognize that the influence
zones of the SRF would be higher if the fault were to dip to the west. Thus, the results of this study apply to
sites located on the footwall block of reverse shallow crust faults in Chile.
The model for the SRF encompasses a significant degree of epistemic uncertainties. As described in Section 2,
various researchers have proposed different values for the fault geometric parameters, including trace length
(e.g., 30 km, 4𝑥10 km, and 50 km), width (ranging between 10 to 40 km), and dipping angle (ranging between
30° to 65°). The same occurs with SRF seismicity. In the literature, the fault’s maximum magnitude ranges
between 𝑀𝑤 6.7 − 7.5, and its recurrence period varies between 2,000 − 10,000 or 9,000 ± 500 (the latter is
based on only two previous events). Although some of these uncertainties were incorporated into the logic tree
model of the fault for probabilistic analyses, resulting response spectra, hazard curves, and hazard maps should
be interpreted with caution. It should be noted that the deterministic approach considered the fault’s maximum
length and magnitude proposed in the literature. Given the uncertainties surrounding the SRF, it is plausible
that its actual length is shorter, and its maximum magnitude is lower than those used in this investigation. In
the future, further studies of the SRF and its seismic activity are warranted to reduce its associated epistemic
uncertainty and produce more accurate estimations of the seismic hazard in Santiago.
Results obtained in this investigation exhibit a great dependency on the selected approach. Under a deterministic
analysis, influence zones of the SRF extend 30 km west of the fault trace. Moreover, the seismic hazard posed

26
by the SRF reached over 3.2 times that induced by subduction events. However, from a probabilistic
perspective, the influence of the SRF is negligible, as the seismic hazard ratio is systematically lower than 3%.
These discrepancies between deterministic and probabilistic results prompt the simple question of which
approach to use. It should be noted that while the goal of this article was to compute and compare the seismic
hazard of the SRF and subduction zone within both deterministic and probabilistic approaches, the utilization
of this information will ultimately depend on the specific decision-making context, as pointed out by McGuire
(2001) and Bommer (2002), among others.
Regardless of the approach, the influence of the SRF depends on the considered vibration period. Under a
deterministic approach, both response spectra and hazard maps present a significantly greater impact of the SRF
at longer vibration periods. For example, when considering historical earthquake magnitude scenarios, hazard
ratio maps show influence zones of the SRF extending 4 km to the west for Sa(0.3 s) and 30 km to the west for
Sa(3.0 s).
The same trend is observed when utilizing a PSHA. Although the fault has no major influence on the seismic
hazard in the study zone, UHS, seismic hazard curves, and hazard maps show an increased contribution of the
SRF when the vibration period increases. This is observed in the differences between seismic hazard curves
when changing the vibration period.
The seismic hazard curves shown in Figures 18 and 19 demonstrate a clear upper limit for the seismic hazard
of the SRF. This limit, shown in the figures as a return period of 9,000 years, corresponds to the mean recurrence
rate assigned to the seismic model of the fault (see Figure 4). Considering that the fault has a non-activation
time of approximately 8,000 ± 500 years, a non-Poissonean (i.e., time-dependent) recurrence model could
improve the results and increase the seismic hazard of the SRF.
The impact of the mean recurrence period on the PSHA results leads us to conclude that other active shallow
crust faults, with lower recurrence periods, may potentially generate a considerably higher seismic hazard than
that obtained for the SRF in this investigation. In that case, the SRF would not serve as an upper bound for the
influence zones of active shallow crusts in Chile, as previously suggested. However, when looking at shallow
crust faults in Chile proved to be active in the database compiled by Maldonado et al. (2021), none presents a
lower recurrence period than the SRF (although many faults do not have this information). Thus, the SRF study
case is still an upper bound given the state-of-the-art information.
The results of a PSHA for each realization of the logic trees for the SRF and subduction events lead to
observations similar to those discussed in the previous paragraphs. The SRF generates a lower seismic hazard
at both study sites than the one induced by subduction events. However, some realizations of the subduction
model yielded seismic hazard curves on Site 2 that reached values lower than those computed for all SRF
models. Although this indicates that, for specific models, an influence zone of the SRF would appear, the
intersection of the seismic hazard curves occurs at recurrence rates lower than 3𝑥10−5 years-1, meaning return
periods over thirty thousand years. At such a time scale, it is highly uncertain if our models can predict the
potential intensities throughout the city.

6. Conclusions
Evaluating the seismic hazard posed by the San Ramón Fault (SRF) is crucial to Santiago, Chile’s capital city.
Previous shallow earthquakes occurring near populated regions worldwide have caused significant damage,
losses, and several fatalities. Moreover, the urbanization along the SRF trace, its proximity to critical
infrastructure, and its potential for surface ruptures increase the seismic risk of the city. In this context, this
article presents a comparative analysis of the hazard generated by subduction events (i.e., interface and inslab
events) against that induced by the SRF, using both deterministic and probabilistic seismic hazard assessments.
We study the “influence zone of the SRF”, which is defined as the region within Santiago where the seismic
hazard of the SRF is greater than that of the subduction zone. Given the particularities of the SRF (i.e., its
location relative to the subduction zone and its potential for nucleating large-magnitude events), it serves as an
upper bound for the influence zone of similar faults in Chile, located north of latitude 50°S. We explore the

27
impact of the considered analysis (i.e., deterministic vs probabilistic) and the selected intensity measure (i.e.,
short-period vs long-period spectral acceleration ordinates) on estimating the influence zone of the SRF.
The influence zones of the SRF obtained in this study vary significantly depending on the approach considered
for the seismic hazard analysis. When considering a spectral acceleration ordinate at a vibration period of 3.0 s
and a deterministic analysis, the influence zone of the SRF extends 30 km to the west of the fault trace, with a
seismic hazard ratio between the SRF and the subduction zone peaking at over 320%. However, there is no
influence zone under a probabilistic approach, and the seismic hazard ratio remains under 3%. This is mainly
attributed to the large mean recurrence period assigned to the SRF (i.e., 9,000 years) and the characteristic event
recurrence model chosen for a time-independent probabilistic approach. This model provides a first-order
approximation of the seismic hazard induced by the SRF. Therefore, further studies of the potential seismicity
of the SRF, especially its magnitude-recurrence relationship, are warranted.
The influence of the SRF on the seismic hazard within the study region obtained in this research shows a strong
dependency on the vibration period. For example, when performing a deterministic analysis, in both response
spectra (i.e., considering historical and worst-case scenarios for subduction events), the SRF controlled the
seismic hazard for long vibration periods at both study sites. Furthermore, the hazard ratio obtained in the study
zone reached 3.2 for a long vibration period and only 1.4 for a short vibration period. When performing a
probabilistic analysis, the effect of the vibration period is not easily observed on the UHS, as the influence of
the SRF is considerably lower than that of the subduction zone. However, seismic hazard curves of the SRF
have a notable difference when changing the vibration period, and the maximum hazard ratio slightly increases
from 0.8% to 1.2% for Sa(0.3 s) and Sa(3.0 s) at Site 1, respectively.
The findings of this study will serve as guidelines for future revisions of seismic design codes and urban growth
management policies in Chile. Although the current understanding suggests a low likelihood of an earthquake
occurring on the SRF in the next 50 years, as its return period has been estimated as 9,000 years (Vargas et al.
2014), its potential consequences could be devastating, especially in the fault's proximity. Ground motion
intensities may far exceed those of large-magnitude inslab or interface events, highlighting the need for
preparedness and risk mitigation measures.
Finally, this investigation focused on the ground-shaking intensities that may affect the city of Santiago during
earthquakes in the SRF or the subduction zone. However, an earthquake occurring in the SRF may also trigger
surface ruptures and landslides in the proximities of the fault. Although these additional earthquake-induced
hazards were out of the scope of this article, they may pose a higher seismic risk to people, structures, and
infrastructure near the SRF and warrant future investigations.

References
Abrahamson N, Gülerce Z (2022) Summary of the Abrahamson and Gulerce NGA-SUB ground-motion model
for subduction earthquakes. Earthquake Spectra 38:2638–2681.
https://doi.org/10.1177/87552930221114374
Abrahamson N, Silva WJ, Kamai R (2014) Summary of the ASK14 ground motion relation for active crustal
regions. Earthquake Spectra 30:1025–1055
Akkar S, Sandıkkaya MA, Bommer JJ (2014) Empirical ground-motion models for point- and extended-source
crustal earthquake scenarios in Europe and the Middle East. Bulletin of Earthquake Engineering 12:359–
387. https://doi.org/10.1007/s10518-013-9461-4
Ammirati J-B, Vargas G, Rebolledo S, Abrahami R, Potin B, Leyton F, Ruiz S (2019) The crustal seismicity of
the western Andean thrust (Central Chile, 33°-34° s): Implications for regional tectonics and seismic
hazard in the Santiago area. Bulletin of the Seismological Society of America 109:1985–1999.
https://doi.org/10.1785/0120190082
Ancheta T, Bozorgnia Y, Darragh R, Silva WJ, Chiou B, Stewart JP, Boore DM, Graves R, Abrahamson NA,
Campbell KW (2012) PEER NGA-West2 database: A database of ground motions recorded in shallow
crustal earthquakes in active tectonic regions. In: Proceedings, 15th world conference on earthquake
engineering. pp 24–28

28
Armijo R, Rauld RA, Thiele R, Vargas G, Campos J, Lacassin R, Kausel E (2010) The West Andean Thrust,
the San Ramón Fault, and the seismic hazard for Santiago, Chile. Tectonics 29.
https://doi.org/10.1029/2008tc002427
Astroza M (2007) Reinterpretación de las intensidades del terremoto de 1906. VI Congreso Chileno de
Geotecnia, Valparaíso, Chile
Barrientos S, Kausel E, Campos J (1997) Sismicidad de profundidad intermedia y peligro sísmico en Santiago.
In: VIII Congreso Geológico Chileno. pp 645–649
Bindi D, Massa M, Luzi L, Ameri G, Pacor F, Puglia R, Augliera P (2014) Pan-European ground-motion
prediction equations for the average horizontal component of PGA, PGV, and 5 %-damped PSA at
spectral periods up to 3.0 s using the RESORCE dataset. Bulletin of Earthquake Engineering 12:391–
430. https://doi.org/10.1007/s10518-013-9525-5
Bommer JJ (2002) Deterministic vs. probabilistic seismic hazard assessment: An exaggerated and obstructive
dichotomy. Journal of Earthquake Engineering 6:43–73. https://doi.org/10.1080/13632460209350432
Bommer JJ, Douglas J, Scherbaum F, Cotton F, Bungum H, Fäh D (2010) On the selection of ground-motion
prediction equations for seismic hazard analysis. Seismological Research Letters 81:783–793.
https://doi.org/10.1785/gssrl.81.5.783
Boore DM, Stewart JP, Seyhan E, Atkinson GM (2014) NGA-West2 equations for predicting PGA, PGV, and
5% damped PSA for shallow crustal earthquakes. Earthquake Spectra 30:1057–1085.
https://doi.org/10.1193/070113EQS184M
Brüggen J (1950) Fundamentos de la Geología en Chile. Instituto Geográfico Militar, Santiago
Bustos J, Pastén C, Pavez D, Acevedo M, Ruiz S, Astroza R (2023) Two-dimensional simulation of the seismic
response of the Santiago Basin, Chile. Soil Dynamics and Earthquake Engineering 164:107569.
https://doi.org/10.1016/j.soildyn.2022.107569
Campbell KW, Bozorgnia Y (2014) NGA-West2 ground motion model for the average horizontal components
of PGA, PGV, and 5% damped linear acceleration response spectra. Earthquake Spectra 30:1087–1115
Campos J, Kausel E (1990) The large 1939 intraplate earthquake of southern Chile. Seismological Research
Letters 61:43
Campos J, Ruiz S, Pérez A, Ruiz J, Kausel E, Thiele R, Saragoni GR, Sepúlveda S (2005) Terremotos Corticales
de las Melosas 1958, Chusmiza 2001 y Curicó 2004: Un Análisis Comparativo de los Terremotos de
Northridge 1994 y Kobe 1995. Nuevos Antecedentes para el Peligro Sísmico en Chile. Jornadas Chilenas
de Sismología e Ingeniería Antisísmica
Chiou BS-J, Youngs RR (2014) Update of the Chiou and Youngs NGA model for the average horizontal
component of peak ground motion and response spectra. Earthquake Spectra 30:1117–1153
Cifuentes IL, Silver PG (1989) Low-frequency source characteristics of the great 1960 Chilean earthquake.
Journal of Geophysical Research 94:643–663. https://doi.org/10.1029/JB094iB01p00643
Comte D, Eisenberg A, Lorca E, Pardo M, Ponce L, Saragoni GR, Singh SK, Suárez G (1986) The 1985 Central
Chile earthquake: A repeat of previous great earthquakes in the region? Science 233:449–453.
https://doi.org/10.1126/science.233.4762.449
Cornell CA (1968) Engineering Seismic Risk Analysis. Bulletin of the Seismological Society of America
58:1583–1606
Dilsiz A, Gunay S, Mosalam K, Miranda E, Arteta C, Sezen H, Fischer E, Hakhamaneshi M, Hassan W,
Alhawamdeh B, Andrus S, Archbold J, Arslantürkoglu S, Bektaş N, Ceferino L, Cohen J, Duran B, Erazo
K, Faraone G, Feinstein T, Gautam R, Gupta A, Haj Ismail S, Jana A, Javadinasab Hormozabad S,
Kasalanati A, Kenawy M, Khalil Z, Liou I, Marinković M, Martin A, Merino Y, Mivehchi M, Moya L,
Pájaro Miranda C, Quintero N, Rivera J, Romão X, Lopez Ruiz M, Sorosh S, Vargas L, Velani P, Wibowo
H, Xu S, Yilmaz T, Alam M, Holtzer G, Kijewski-Correa T, Robertson I, Roueche D, Safiey A (2023)
StEER: 2023 Mw 7.8 Kahramanmaras, Türkiye Earthquake Sequence Preliminary Virtual
Reconnaissance Report (PVRR). StEER- February 6, 2023, Kahramanmaras, Türkiye, Mw 78

29
 Earthquake. https://doi.org/https://doi.org/10.17603/ds2-7ry2-gv66
Easton G, Inzulza J, Pérez S, Ejsmentewicz D, Jiménez C (2018) ¿Urbanización fallada? La Falla San Ramón
como nuevo escenario de riesgo sísmico y la sostenibilidad de Santiago ,Chile. Revista de Urbanismo
38:1–20
Estay NP, Yáñez G, Carretier S, Lira E, Maringue J (2016) Seismic hazard in low slip rate crustal faults,
estimating the characteristic event and the most hazardous zone: Study case San Ramón Fault, in southern
Andes. Natural Hazards and Earth System Sciences 16:2511–2528. https://doi.org/10.5194/nhess-16-
2511-2016
Esteva L (1967) Criterios para la construcción de espectros para diseño sísmico. In: Jornadas Sudamericanas
de Ingeniería Estructural y 3er Simposio panamericano de estructuras. Caracas, pp 3–8
Hanks TC, Kanamori H (1979) A moment magnitude scale. Journal of Geophysical Research: Solid Earth
84:2348–2350. https://doi.org/https://doi.org/10.1029/JB084iB05p02348
Hayes GP, Moore GL, Portner DE, Hearne M, Flamme H, Furtney M, Smoczyk GM (2018) Slab2, a
comprehensive subduction zone geometry model. Science 362:58–61
Idini B, Rojas F, Ruiz S, Pastén C (2017) Ground motion prediction equations for the Chilean subduction zone.
Bulletin of Earthquake Engineering 15:1853–1880. https://doi.org/10.1007/s10518-016-0050-1
Joyner WB, Boore DM (1988) Measurement, characterization, and prediction of strong ground motion. In:
Earthquake Engineering and Soil Dynamics II, Proc. Am. Soc. Civil Eng. Geotech. Eng. Div. Specialty
Conf. pp 27–30
Kanaori Y, Kawakami SI (1996) The 1995 7.2 magnitude Kobe earthquake and the Arima-Takatsuki tectonic
line: implications of the seismic risk for central Japan. Engineering Geology 43:135–150.
https://doi.org/10.1016/S0165-1250(97)80006-5
Krinitzsky EL (2002) How to obtain earthquake ground motions for engineering design. Engineering Geology
65:1–16. https://doi.org/https://doi.org/10.1016/S0013-7952(01)00098-9
Kuehn NM, Bozorgnia Y, Campbell KW, Gregor N (2023) A regionalized partially nonergodic ground-motion
model for subduction earthquakes using the NGA-Sub database. Earthquake Spectra 39:1625–1657.
https://doi.org/10.1177/87552930231180906
Maldonado V, Contreras M, Melnick D (2021) A comprehensive database of active and potentially-active
continental faults in Chile at 1:25,000 scale. Scientific Data 8:20. https://doi.org/10.1038/s41597-021-
00802-4
McGuire RK (2001) Deterministic vs. probabilistic earthquake hazards and risks. Soil Dynamics and
Earthquake Engineering 21:377–384. https://doi.org/10.1016/S0267-7261(01)00019-7
Montalva GA, Bastías N, Rodriguez-Marek A (2017) Ground-motion prediction equation for the Chilean
subduction zone. Bulletin of the Seismological Society of America 107:901–911.
https://doi.org/10.1785/0120160221
Norambuena A (2006) Estudio de los efectos del terremoto de La Ligua del 28 de marzo de 1965. Memoria
para optar al título de Ingeniero Civil, Facultad de Ciencias Físicas y Matemáticas, Departamento de
Ingeniería Civil, Universidad de Chile, Santiago, Chile
Parker GA, Stewart JP, Boore DM, Atkinson GM, Hassani B (2021) NGA-subduction global ground motion
models with regional adjustment factors. Earthquake Spectra 38:456–493.
https://doi.org/10.1177/87552930211034889
Pérez A, Ruiz JA, Vargas G, Rauld RA, Rebolledo S, Campos J (2014) Improving seismotectonics and seismic
hazard assessment along the San Ramón Fault at the eastern border of Santiago city, Chile. Natural
Hazards 71:243–274. https://doi.org/10.1007/s11069-013-0908-3
Pilz M, Parolai S, Stupazzini M, Paolucci R, Zschau J (2011) Modelling basin effects on earthquake ground
motion in the Santiago de Chile basin by a spectral element code. Geophysical Journal International
187:929–945. https://doi.org/10.1111/j.1365-246X.2011.05183.x

30
Poulos A, Monsalve M, Zamora N, de la Llera JC (2019) An updated recurrence model for chilean subduction
seismicity and statistical validation of its Poisson nature. Bulletin of the Seismological Society of America
109:66–74. https://doi.org/10.1785/0120170160
Rauld RA (2002) Análisis morfoestructural del frente cordillerano de Santiago Oriente, entre el río Mapocho y
la quebrada de Macul. Memoria para optar al título de Geólogo, Departamento de Geología, Universidad
de Chile, Santiago, Chile
Santibáñez I, Cembrano J, García-Pérez T, Costa C, Yáñez G, Marquardt C, Arancibia G, González G (2019)
Crustal faults in the chilean Andes: Geological constraints and seismic potential. Andean Geology 46:32–
65. https://doi.org/10.5027/andgeov46n1-3067
Scherbaum F, Delavaud E, Riggelsen C (2009) Model selection in seismic hazard analysis: An information-
theoretic perspective. Bulletin of the Seismological Society of America 99:3234–3247
Schwartz DP, Coppersmith KJ (1984) Fault behavior and characteristic earthquakes: Examples from the
Wasatch and San Andreas Fault Zones. Journal of Geophysical Research: Solid Earth 89:5681–5698.
https://doi.org/https://doi.org/10.1029/JB089iB07p05681
Shin TC, Teng TL (2001) An overview of the 1999 Chi-Chi, Taiwan, earthquake. Bulletin of the Seismological
Society of America 91:895–913. https://doi.org/10.1785/0120000738
Thingbaijam KKS, Martin Mai P, Goda K (2017) New Empirical Earthquake Source‐Scaling Laws. Bulletin of
the Seismological Society of America 107:2225–2246. https://doi.org/10.1785/0120170017
Vargas G, Klinger Y, Rockwell TK, Forman SL, Rebolledo S, Baize S, Lacassin R, Armijo R (2014) Probing
large intraplate earthquakes at the west flank of the Andes. 42:1083–1086.
https://doi.org/10.1130/G35741.1
Vigny C, Socquet A, Peyrat S, Ruegg JC, Métois M, Madariaga R, Morvan S, Lancieri M, Lacassin R, Campos
J, Carrizo D, Bejar-Pizarro M, Barrientos S, Armijo R, Aranda C, Valderas-Bermejo MC, Ortega I,
Bondoux F, Baize S, Lyon-Caen H, Pavez A, Vilotte JP, Bevis M, Brooks B, Smalley R, Parra H, Baez
JC, Blanco M, Cimbaro S, Kendrick E (2011) The 2010 Mw 8.8 Maule Megathrust Earthquake of Central
Chile, Monitored by GPS. Science 332:1417–1421. https://doi.org/10.1126/science.1204132
Zhao JX, Zhou S, Zhou J, Zhao C, Zhang H, Zhang Y, Gao P, Lan X, Rhoades D, Fukushima Y (2016) Ground‐
motion prediction equations for shallow crustal and upper‐mantle earthquakes in Japan using site class
and simple geometric attenuation functions. Bulletin of the Seismological Society of America 106:1552–
1569

Statement and Declarations

Funding
This work was supported by ANID (Grant numbers 11230463 and ACT240044) and Universidad Técnica
Federico Santa María (Becas para Estudios de Magíster Científico-Tecnológico).

Competing Interests
The authors have no relevant financial or non-financial interests to disclose.

Author Contributions
Horacio Domínguez F.: Methodology, Software, Investigation, Writing – Original Draft, Visualization. Pablo
Heresi V.: Conceptualization, Validation, Writing – Review and Editing, Supervision. Gonzalo Montalva A.:
Writing – Review and Editing.

Data Availability
The datasets generated during and/or analyzed during the current study are available from the corresponding
author upon reasonable request.

31