Seismic hazard analysis for public infrastructure in Metro Manila

Patrick Adrian Y. Selda1, R.A.C. Luna1, R.D. Quebral1, J.M.B. Gargullo1, K.J.R. Leobrera1, E.D.
Marasigan1, M.J.P. Armario1, and M.P. Baniquett1
1 AMH Philippines, Inc., University of the Philippines, Diliman, Quezon City 1101, Philippines.

ABSTRACT

Metro Manila is the Philippines’ National Capital Region and the nation’s political, economic and cultural center.
Several transport infrastructure, including the expansion of its current light rail transit (LRT), new elevated highways,
and the first subway system, are either under construction or in the detailed engineering design stage. Being situated
in tectonically active zone, the seismic design of this major infrastructure is a primary consideration. This paper
presents a seismic hazard analysis aimed at contributing to the robust and cost-effective design of public
infrastructure. The results of the study are also compared with the provisions of the National Structural Code of the
Philippines (2015), and the Philippine Earthquake Model (2017).

Keywords: Metro Manila; Philippines; infrastructure; Seismic Hazard Analysis; Valley Fault System

1 INTRODUCTION In order to provide a safe and cost-effective design of a

proposed major transport infrastructure in Metro Manila,
1.1 Study Area the geological and geotechnical conditions, as well as the
Metro Manila, the Philippines’ National Capital seismicity of the area, must be well understood.
Region (NCR), is the political, economic, and cultural
center of the Philippines. With a total area of
2 GEOLOGY AND SEISMICITY
approximately 620 km2, NCR houses several business
districts, schools, hospitals, and government offices. A 2.1 Tectonic Setting
census conducted in 2015 showed that the population in The Philippine Mobile Belt corresponds to the
Metro Manila is approximately 12 million which is complex boundary between the Eurasian Plate and the
steadily increasing every year. Metro Manila is thus Philippine Sea Plate. The Philippine Mobile Belt refers to
considered as one of the most densely populated areas in the portion of the Philippine archipelago bounded to the
Southeast Asia. With the continuous increase in the west by the Manila-Negros-Cotabato-Sulu Trenches and
population comes the need to alleviate the traffic to the east by the East Luzon Trough-Philippine Trench.
congestion and travel time in the city. Several mass The active 1200km long Philippine Fault, as well as many
transportation systems such as MRT, skyways, and other active seismic sources found within the Philippines,
subways are given the “go” signal by the Philippine is a physical manifestation of the surrounding tectonic
government for design and construction. plates’ opposing movements.
1.2 Significance of the Study
Geographically, Metro Manila is located in Luzon
Island. It is influenced by several earthquake generators, Project Area
among which is the Valley Fault System. Paleoseismic
and geologic data have shown that the Valley Fault
System is active. Trenching across the northern portion of
the fault, coupled with Carbon-14 dating, show that three
or four faulting events occurred during the last 1400 years.
An estimate of approximately 200 to 400 years has been
suggested as a possible recurrence interval for large
magnitude earthquakes (Nelson et al, 1995). Many
research studies estimated a possible magnitude 7 or more
for this fault and considering that no seismic event is
known after 17th century, it means that the active phase of Fig. 1. Tectonic Boundaries in the Philippines (Dimalanta and
the Valley Fault System is approaching. Yumul, 2016).
2.2 Regional Geology The Valley Fault System is a system of active faults
The underlying formation in Metro Manila is the that cuts through Metro Manila. It consists of the West
Pliocene-Pleistocene Guadalupe Formation, known as the Valley Fault (WVF) and East Valley Fault (EVF), which
Guadalupe Tuff Formation (GTF). This consists of the can generally be classified as strike-slip faults. West
lower member Alat Conglomerate and upper member Valley Fault is the nearest active fault in Metro Manila
Diliman Tuff. The Alat Conglomerate is a group of which extends from the southern Sierra Madre to
massive poorly sorted round pebbles and small boulders Tagaytay over a distance of 110 km. East Valley Fault, on
conglomerate and sandstone with medium to thin bedded the other hand, extends over a distance of about 30 km.
mudstone or shale, while the Diliman Tuff is a volcanic Lubang Fault is an active left-lateral strike-slip fault
ejecta with some amount of tuffaceous sandstone, that passes through the Verde Passage between Mindoro
tuffaceous siltstone, and shale. This formation, locally and Batangas and continues west towards the Manila
referred to as “adobe”, stretches from Quezon City and Trench. The Lubang Fault is a major branch of the
extends to the Province of Cavite in the south. Philippine Fault.
The Philippine Fault (PF) an active left-lateral strike-
2.3 Seismicity slip fault that cuts across the entire Philippine archipelago
The Philippines accounts for 3.2% of the world’s over a distance about 1200 km. It is the most active
seismicity. It is situated in the Circum-Pacific Belt a.k.a. earthquake generator in the country and has been the
“Ring of Fire”, where 80% of the world’s earthquakes source of several devastating earthquakes.
occur. Philippine seismicity is mainly related to plate
subduction and in part to strike-slip motions along trans-
3 GEOHAZARDS
current faults.
A search of the United States Geological Survey 3.1 Liquefaction
(USGS) earthquake database was conducted for Liquefaction is a phenomenon wherein loose,
earthquakes with magnitudes of at least 4.0 within 100 km saturated, cohesionless soil is subjected to cyclic shear
of Metro Manila. There was a total of 442 events since stress that results in an increase in pore water pressure and
1900. The list included only 5 strong events between reduction of the effective stress to zero. This results in the
magnitudes 6.0 and 8.0—the strongest of which was a fluid behavior and near zero shear resistance of the soil.
magnitude 7.5 earthquake that occurred in August 20, The map shown in Figure 3 indicates the varying degrees
1937. This was a shallow 15km deep earthquake; the of liquefaction susceptibility for different areas of Metro
epicenter of which is approximately 69.4 km away from Manila.
the center of Metro Manila. According to the Philippine
Institute of Volcanology and Seismology (PHIVOLCS), 3.2 Ground Rupture
this was caused by the 19km wide, 56km long strike-slip In the event of an earthquake, the rupture along the
Laguna-Banahaw Fault. surface of an active fault may cause severe damage to
structures or other developments around the vicinity of the
trace. The ground rupture hazard map of Metro Manila
shows minor risk of ground rupture for most areas in
Metro Manila except for those near the West Valley Fault.
Philippine One study estimated the casualty, both dead and injured,
Fault to be around 147,100. Several offices, residential, mid-
rise and high-rise structures are also expected to be
heavily damaged.

Metro Manila

Lubang Fault
Valley Fault
System

Fig. 2. Distribution of Historical Earthquakes around Metro

Manila (United States Geological Survey).

2.4 Potential Earthquake Generators

From the maps produced by PHIVOLCS and Mines
Fig. 3. (Left) Liquefaction Susceptibility Map and (Right)
and Geosciences Bureau (MGB), the following potential
Ground Rupture Hazard Map of Metro Manila (PHIVOLCS).
earthquake generators were considered in the seismic
hazard analysis:
4 SUBSURFACE CONDITIONS 5 SEISMIC HAZARD ANALYSIS
An area’s site class provides strength-deformation Seismic Hazard Analysis (SHA) is the process of
characteristics and, in effect, is an indicator of how the quantifying the overall seismic hazard of an area in
subsurface will behave when subjected to dynamic terms of acceleration. The probabilistic approach
loads such as earthquakes. (PSHA) in performing SHA quantifies seismic hazard
Among the different dynamic properties, it is at different levels of risk depending on the recurrence
generally suitable to use the shear wave velocity of the interval or return period of the design ground motion.
upper 30 m (VS30) to determine which category the site PSHA also considers multiple seismic sources
class falls under. The range of VS30 values that simultaneously and accounts for uncertainties related to
correspond to each site class are tabulated in the distance, time, recurrence, and size (magnitude).
National Structural Code of the Philippines (NSCP) In performing SHA, empirically-formulated
2015 Table 208-2. attenuation models are utilized to determine the
expected surface acceleration by estimating how
Table 1. Site Class Categories (NSCP 2015 Table 208-2). seismic waves propagate and travel from source to site.
Average Soil Properties for Top 30 m of Soil Profile Attenuation models are commonly referred to as
Soil Profile Soil Profile Name /
Type Generic Description Shear Wave Velocity,
𝑽𝑺𝟑𝟎 (m/s)
SPT, N-value
Undrained Shear Ground Motion Prediction Equations (GMPE), and
Strength, 𝑺𝒖 (kPa)
𝑆𝐴 Hard Rock > 1500
these equations were formulated using globally-
𝑆𝐵 Rock 760 to 1500 acquired earthquake information (e.g. epicenter
Very Dense Soil and
𝑆𝐶
Soft Rock
360 to 760 > 50 > 100 location, depth, and magnitude). In this study, the New
𝑆𝐷 Stiff Soil Profile 180 to 360 15 to 50 50 to 100
Generation Attenuation West 2 (NGA-West2) GMPE’s
𝑆𝐸 Soft Soil Profile < 180 < 15 < 50
𝑆𝐹 Soil Requiring Site-specific Evaluation developed by the Pacific Earthquake Engineering
Research (PEER) Center were used for fault systems,
In this study, geophysical tests (e.g. Downhole and the BC Hydro GMPE (Abrahamson et al., 2016)
Seismic Test, PS Suspension Logging) were conducted was used for subduction zone sources.
across Metro Manila in order to measure in-situ seismic At present, Filipino engineers typically extract
wave velocities of the GTF. Other elastic properties seismic design accelerations from the NSCP, which
(which can further be used in numerical modeling) can utilizes a generalized national-scale response spectrum
then be determined using empirically-derived equations developed in 1994. Design accelerations for seismic
correlated to the measured velocity of the in-situ analysis can be taken from the response spectra
seismic waves. generated from PSHA. In anticipation of the varying
From the results of the geophysical tests, the code requirements when designing different types of
average measured VS30 of northern Metro Manila is structures, target response spectra shall represent the
determined to be 856.50 m/s, which just barely falls uniform hazard of ground motions with 2%, 5%, and
under site class SB. On the other hand, the average 10% Probabilities of Exceedance (PoE) in 50 years.
measured VS30 of southern Metro Manila is determined These ground motions have return periods of 2,475,
to be 675.00 m/s, which is on the higher bound of site 975, and 475 years, respectively. In light of this,
class SC. PHIVOLCS conducted a national-scale PSHA in 2017
These findings suggest that Metro Manila is and published the Philippine Earthquake Model
generally underlain by soft rocks, which is essentially (PEM)—an atlas of maps containing isolines of equal
what was expected of the GTF. Not only are these spectral acceleration.
findings consistent with the information gathered from
sampling, but they are also consistent with the 6 RESULTS
generalized Metro Manila VS30 site model map (left According to NSCP 2015 Chapter 208.5.3.2, the
image) published by PHIVOLCS. ground motion to be considered in seismic analysis for
typical vertical structures should, as a minimum,
correspond to one having a return period of 475 years.
This return period is one of the hazard levels provided
in both the NSCP 2015 and the PEM 2017. As such, the
site-specific 475-year seismic hazard generated from
this study shall be compared to both the NSCP 2015
and PEM 2017 response spectra.
The following figures show the generated 5%-
damped 475-year response spectra of the northern and
southern areas of Metro Manila from all three studies.

Fig. 4. Metro Manila VS30 Map (Left image from PHIVOLCS).

 7 CONCLUSION
The probabilistic seismic hazard analysis of Metro
Manila carried out in this study yielded higher values of
peak ground acceleration for various return periods,
compared to the provisions of NSCP 2015 and the
Philippine Earthquake Model (PEM) 2017.
It is therefore essential that site-specific seismic
hazard analysis is undertaken for essential facilities and
critical structures, in order to come up with robust and
earthquake-resistant design.

ACKNOWLEDGEMENTS
Fig. 5. 5%-Damped 475-Year Response Spectra (Northern Area).
The authors would like express their utmost
gratitude to the different agencies/institutions for the
tremendous effort they have put into research which
provided essential information to this study (e.g. MGB,
PHIVOLCS, and USGS). For without any of which,
this study may not have been completed; to GEM and
PEER, for their extensive research in seismology and
for providing the necessary tools (e.g. OpenQuake,
GMPE’s) in performing SHA; to ASCE and ASEP, for
publishing standards that outline the design process of
structures where SHA can be used as input. The authors
would also like to thank AMH Philippines, Inc. and the
organizers of the 16th ARC for providing the
Fig. 6. 5%-Damped 475-Year Response Spectra (Southern Area). opportunity to present this study.

Evidently, the PEM 2017 prescribes much lower REFERENCES

values than both the code-prescribed and site-specific Association of Structural Engineers of the Philippines. 2015.
response spectra. This is true throughout the whole National Structural Code of the Philippines Volume 1:
period domain for both site classes. Comparing this Buildings, Towers, and Other Vertical Structures. Association
study’s response spectra to what is prescribed in the of Structural Engineers of the Philippines, Inc., Quezon City,
code, the Peak Ground Accelerations (the y-intercept or Philippines.
Baker, J.W. (2013). Introduction to Probabilistic Seismic Hazard
the spectral acceleration at 0th period) do not differ Analysis. White Paper Version 2.0.1.
much. Focusing on the shorter period range (up to 0.2s), Daligdig, J.A., Punongbayan, R.S., Besana, G.M., and Tuñgol,
spectral accelerations are also very close. It should be N.M. (1997). The Marikina Valley Fault System: Active
noted, however, that rocks or very hard/dense soils (site Faulting in Eastern Metro Manila. PHIVOLCS Professional
classes SB and SC) tend to resonate with high-frequency Paper 01.
seismic waves, and not so much with low-frequency Galgana, G., Hamburger, M., McCaffrey, R., Corpuz, E., and
Chen, Q. (2007). Analysis of Crustal Deformation in Luzon,
seismic waves. This is the reason the response spectra Philippines using Geodetic Observations and Earthquake
on site classes SB and SC are expected to peak Focal Mechanisms. Tectonophysics 432: 63-87.
somewhere within 0.1s to 0.3s, and is also the reason Japan International Cooperation Agency, Metropolitan Manila
why the response spectrum for the southern area of the Development Authority, and Philippine Institute of
Metro Manila peaks at a longer period compared to that Volcanology and Seismology. (2004). Earthquake Impact
of its northern counterpart. Reduction Study for Metropolitan Manila, Republic of the
Philippines Final Report.
As the period increases past the shorter period range, Philippine Institute of Volcanology and Seismology. (2017). The
the spectral accelerations of the code-prescribed Philippine Earthquake Model.
response spectra plateaus and becomes much higher Thenhaus, P.C., Hanson, S.L., Algermissen, S.T., Bautista, B.C.,
than the other response spectra thereafter (near-source Bautista, L.P., Punongbayan, B.J., Rasdas, A.R., Nillos,
effects play a significant role in this finding). As a J.T.E., and Punongbayan, R.S. (1994). Estimates of the
result, the generalization of the code-prescribed Regional Ground Motion Hazard in the Philippines.
Proceedings of the Conference on Natural Disaster
spectrum is very much highlighted in this scenario. Mitigation in the Philippines.
When designing for specific infrastructures within United States Geological Survey. ANSS Comprehensive
the Metro, this study’s site-specific response spectra Earthquake Catalog Database.
can be used as long as near-source effects are accounted
for prior to analysis.