Tsunami Risk Reduction Measures Phase 2

November 2009

Cover pictures;

Initial water displacements (m) for the Seismicity of the study region for 1963- three northernmost Sunda Arc scenarios 2006, with symbols differentiating the of magnitude M 8.55, 8.53 and 8.60 magnitudes. respectively, as well as the M 8.86 Burma fault scenario.

Merged tsunami hazard Merged tsunami hazard Merged tsunami hazard map for Sri Lanka. map for the Philippines map for Eastern Indonesia

The designations employed and the presentation of material in this publication do not imply the expression of any opinion whatsoever on the part of the CCOP Technical Secretariat concerning the legal status of any country, territory, city or area or of its authorities, or concerning the delimitation of its frontiers or boundaries.

Note: The conclusions and recommendations of this publication have not been specifically endorsed by, or reflect the views of the organizations which have supported the production of this project, both financially and with content.

© Coordinating Committee for Geosciences Programmes in East and Southeast Asia, 2009

Project

Project: Tsunami Risk Reduction Measures phase 2 Document No.: 20061179-00-226-R Document title: Tsunami Risk evaluations for the Philippines Date: 6 November 2009

Client

Client: CCOP Technical Secretariat Client’s contact person: Niran Chaimanee Contract reference: Contract between CCOP and NGI of 17. April 2008

For NGI

Project manager: Kjell Karlsrud Prepared by: Bjørn Kalsnes Finn Løvholt, Sylfest Glimsdal, Daniela Kühn, Hilmar Bungum, Bjørn Vidar Vangelsten, Regula Frauenfelder, Helge Smebye, Gunilla Kaiser Reviewed by: Carl Bonnevie Harbitz

Summary

This report presents tsunami hazard analyses dedicated to parts of the Philippines. In agreement with the Philippine Institute of Volcanology and Seismology (PHIVOLCS), the tsunami hazard was studied within three regions, the Manila Trench, the South Western Mindanao, and for far field sources from the Northern Sulawesi. Available seismic and tsunami catalogues indicate that there is a high level of seismicity the last 100 years and a large number of historical tsunamis within the last 300-400 years. Using a scenario based approach; a merged tsunami hazard map is developed for earthquake generated tsunamis for the Manila Trench study area. The map gives more than 10 m maximum water elevation along large sections of western Luzon Island. In addition, a hindcast of the 1976 Moro Gulf

Summary (cont.) Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 4 tsunami is conducted, as well as one landslide generated tsunami scenario for Calapan - Mindoro Island, the latter for demonstration purposes. Sources located along Northern Sulawesi may also affect the Southern part of the Philippines.

Return periods for the individual earthquake scenarios are not quantified. Lower bound return periods for the earthquake magnitudes comparable to the scenarios are found within some of the study regions, obtaining 30-120 years from the regional seismicity. However, the scenario return periods are believed to be a few times longer than the lower bound. Moreover, the tectonic convergence rates support higher return periods of 100-1000 years for the scenario magnitudes in question.

Inundation modelling has been undertaken for the city of Batangas, which constituted the basis for a more detailed study for demonstration of local tsunami vulnerability and risk assessment. The study resulted in a critical facility map, as well as a mortality risk map identifying the most affected areas in the city. For a Mw 8.2 Manila Trench scenario with a 1 m sea level rise (high tide), the analyses indicates a loss of several hundreds of lives.

Due to the possibility for a relatively long warning time for the scenario investigated, a proper Early Warning System, including technical installations and a well organized dissemination system, as well as defined evacuation routes, may be effective for the most affected areas. A barrier of 2 m above sea level properly located will most likely reduce the wave impact significantly and hence reduce the consequences in the most exposed areas of Batangas city.

Contents Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 5

1 Introduction 6 2 Definitions 7 3 Historic background 9 4 Results from scenario simulations 12 4.1 South West Mindanao – Cotabato Trench 13 4.2 Northern Sulawesi 15 4.3 Manila Trench 20 4.4 Calapan landslide tsunami scenario – Mindoro Island, the Philippines 24 5 Batangas, the Philippines 27 5.1 Inundation model for the southern magnitude 8.2 Manila Trench scenario 27 5.2 Vulnerability and risk assessment 29 6 Mitigation measures 32 7 Acknowledgements 33 8 References 33

Review and reference page

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 6

1 Introduction

The project “Tsunami Risk Assessment and mitigation in S&SE Asia – Phase 2” has been financed by The Norwegian Ministry of Foreign Affairs (MFA). The Coordinating Committee for Geoscience Programmes in East and Southeast Asia (CCOP), through their Technical Secretariat in Bangkok, acted as the project responsible institution towards MFA. The Norwegian Geotechnical Institute (NGI) had the role as Technical Executing Organisation (TEO). NGI subcontracted NORSAR for performing the seismological analyses required. The project was contracted in 2008 with four Asian countries: Indonesia, the Philippines, Vietnam, and Sri Lanka. The main goals of the project have been to reduce the tsunami risk in South and Southeast Asia by:

• Enhanced assessment of tsunami hazard and recommendations of risk mitigation measures in specified regions • Enhanced capacities of hazard assessment and risk reduction for regional, national, and local institutions

The detailed scope of work (SoW) for the invited countries the Philippines, Indonesia, Vietnam, and Sri Lanka varied according to the needs defined from previous tsunami hazard assessments and the capabilities of the individual countries. The SoW’s were agreed in project meetings with the countries in the early phase of the project. A map of the study area is shown in Figure 1.

This report presents selected tsunami analyses for the Philippines. These include tsunami hazard assessment due to earthquakes along the Manila Trench, a brief hindcast of the 1976 Moro Gulf event, demonstration of numerical methods for simulating landslide generated tsunamis, and demonstration of tools for tsunami risk evaluations and vulnerability for the city of Batangas.

In the complete project report (NGI, 2009) the findings for all the four countries and more elaborate details of the analysis relevant for the Philippines are given. For this purpose, NGI (2009) is extensively cited herein. It is emphasised that the hazard evaluations in this report considers only potential tsunamis of seismic origin. Modelling of tsunamis generated by landslides is included for demonstration purposes only. Moreover, it is stressed that a scenario based approach rather than a full probabilistic method is applied in this report.

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 7

Figure 1: Map of the study area including Indonesia, the Philippines, Vietnam, and Sri Lanka.

2 Definitions

Below, some definitions of technical key terms used in this text are given to help the reader. As far as possible, compatibility with the UNESCO-IOC tsunami glossary (UNESCO-IOC, 2006) is endeavoured. In addition, a brief definition sketch defining the parameters related to the tsunami inundation process is given in Figure 2.

• Fault - A fracture or a zone of fractures along which displacement has occurred parallel to the fracture. Earthquakes are caused by a sudden rupture along a fault or fault system; the ruptured area may be up to several thousand square kilometers. Relative movements across a fault may typically be tens of centimeters for magnitude 6.0-6.5 earthquakes, several meters for magnitude 7-9 earthquakes. • Flow depth – Water elevation above land during inundation (Figure 2). • Hazard - Probability that a particular danger (threat) occurs within a given period of time. Here, the tsunami hazard is the maximum water level associated with a scenario return period. • Inundation distance – Horizontal penetration of the tsunami from the shoreline. • Magnitude - A measure of earthquake size at its source. Magnitude was defined by C. Richter in 1935 as: “The logarithm to the trace amplitude in

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 8

0.001 mm on a standard Wood-Anderson seismometer located 100 km from the epicenter” The Wood-Anderson instrument measures the responses in the period range near 1 sec. Other magnitude scales have later been devised based on the responses measured in other period ranges, and on maximum amplitudes of specific wave forms. In this report, we mostly refer to the moment magnitude (with abbreviation Mw). The moment magnitude is based on the seismic moment computed directly from source parameters or from long period components in the earthquake record. Symbol M is also used for this magnitude. • Maximum water level – Here, defined as the largest water elevation above the still water level (see Figure 2). • Probability - A measure of the degree of certainty. This measure has a value between zero (impossibility) and 1.0 (certainty). It is an estimate of the likelihood of the magnitude of the uncertain quantity, or the likelihood of the occurrence of the uncertain future event. • Return period - Average time period between events of a given size in a particular region, cycle time. • Risk - Measure of the probability and severity of an adverse effect to life, health, property, or the environment. Quantitatively, Risk = Hazard × Potential Worth of Loss. This can be also expressed as “Probability of an adverse event times the consequences if the event occurs”. • Run-up height – Water level above the still water level at the inundation limit (see Figure 2). • Surface elevation – Here, defined as the water elevation relative to the mean sea (can be negative or positive). See Figure 2 for a definition sketch. • Threat - The natural phenomenon that could lead to damage, described in terms of its geometry, mechanical and other characteristics. The danger can be an existing one (such as a creeping slope) or a potential one (such as a tsunami). The characterization of a danger or threat does not include any forecasting. Here, the tsunami threat is mostly reported as the maximum water level. • Trench - Topographic depressions of the sea floor. • Vulnerability - (1) The degree of loss to a given element at risk, or set of such elements, resulting from an event of a given magnitude or intensity, usually expressed on a scale from 0 (no loss) to 1 (total loss). (2) Degree of damage caused by various levels of loading.

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 9

Figure 2: Definition sketch for tsunami parameters.

3 Historic background

Figure 3 shows source locations and source type for historical tsunamis recorded throughout the last 400 years of recorded history obtained by combining different tsunami and earthquake databases (more details are given in NGI, 2009). The database shows that a large number of tsunamis have occurred in the Philippines. Naturally, the results shown in Figure 3 should be used with caution, as there are large uncertainties related to these data. Most of the tsunamis are reported to be purely due to tectonic events.

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 10

Figure 3: Source location and year of occurrence for recorded earthquake generated historical tsunamis in the Philippines with source information. Yellow markers indicate non seismic or unknown sources, red markers indicate seismic sources. Large stars magnitudes M ≥ 8.5; small stars 8.5 > M ≥ 8.0; squares 8.0 > M ≥ 7.5; circles M < 7.5; asterisk, no magnitude reported. Upward-pointing triangles indicate volcanoes or combinations of volcanoes and other sources. Downward-pointing triangles indicate landslides or landslides and earthquakes. Diamonds indicate unknown sources.

In agreement with PHIVOLCS, two study regions (Manila Trench and southwestern Mindanao / Cotabato Trench) were defined as shown with red squares in Figure 4. The figure illustrates very clearly the high level of seismicity recorded since 1963, showing that earthquakes above magnitude 7.0 are not at all rare. At the same time Figure 4 also depicts some of the main transform faults and trenches.

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 11

Figure 4: Seismicity in the Philippines 1963-2006, with symbols differentiating the magnitudes. The largest red stars are for earthquakes above magnitude 8, the orange magnitudes 7.5-8, whereas the blue and cyan colored stars depict the smaller events. The black lines are transform faults and the green lines are trenches (subduction zones).

The tectonic situation in the Philippines can be briefly summarized as follows (Kremer et al., 2000): • In the Philippine region, there is an oblique convergence between Pacific Plate and Sunda block: • There is strain partitioning in the near-trench-normal subduction at the Philippine trench and left-lateral shear at the Philippine Fault (Barrier et al., 1991).

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 12

• In the northern Philippines most of the convergence is taken up at Manila trench (Rangin et al., 1999), where however the high strain rates are not matched by a corresponding seismic activity. The polarity here is opposite to the Philippine trench.

4 Results from scenario simulations

Results from tsunami simulations are briefly summarized below. The selection of the various earthquake scenarios are performed by combining information of historical tsunami records and seismicity. Below, the historical events from the regional tsunami database developed within this project as well as seismic focal mechanisms are shown. Table 1 shows the color legend for the figures. The analyses of the seismic data leading to the scenarios are elaborated by NGI (2009), which also includes the 3D visualization of the seismicity. For all the study regions in question, the lower bound return period for earthquakes with the same magnitude as the scenario earthquakes are given (here 30 – 120 years from the regional seismicity), however the individual scenario return periods are not quantified. These are in most cases expected to be a few times longer than the lower bound. Tectonically derived return periods for this region range from 100 – 1000 years. It is stressed that the range of the return periods is certainly small enough to warrant implementation of mitigation and precautionary measures.

Table 1: Color legends used for subsequent figures. The abbreviation m.w.l. indicate maximum water level. Tsunami surface elevation and Earthquake magnitudes run-up

The scenario earthquake magnitudes represent the upper range of expected, but still realistic future events, and represent so called ‘credible worst case scenarios’. The initial seabed displacement is computed by a standard dislocation model (Okada, 1985) and the tsunami propagation is modeled using the dispersive GloBouss model. Rough estimates of the maximum water level are computed using the method of amplification factors. Details on the computational methods are described by NGI (2009).

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 13

4.1 South West Mindanao – Cotabato Trench

Historical tsunamis and seismicity for the South West Mindanao is shown in Figure 5. The Philippine earthquake on the 16th of August 1976 was, although very large, not associated with the dominant tectonic feature of the Philippine trench, but with a less prominent trench system in the Moro Gulf (North Celebes Sea), the Cotabato Trench. Bathymetric data reveal the presence of a trench striking north-south in the region of the earthquake and curving west—north-west to east—south-east paralleling the southern Mindanao coastline. Since the 1976 event caused a locally destructive tsunami, this location is chosen for a scenario.

Figure 5: Upper panel: source location of historical tsunamis with maximum run-up / water levels colour-coded. Lower panel: map view of fault planes (red rectangles: shallow fault planes; green lines: trenches, black lines: transform faults) and CMT solutions colour-coded according to magnitude. See Table 1 for legend. The initial surface elevations for a scenario mimicking the 1976 Cotabato earthquake and tsunami are shown in Figure 6. The earthquake parameters, i.e. the shear strength, dip angle, length, width, and slip were first based on the paper by Stewart and Cohn (1979), and next updated after preliminary numerical tsunami simulations and discussions (Okal 2009, pers. comm.). Moreover, Okal (2009, pers. comm.) pointed out that the localisation and strength of the 1976 source hypocentre was highly uncertain due to the limited quality of the seismic traces.

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 14

Figure 6: Mw8.1 Cotabato trench scenario. Initial water displacement in meters.

Figure 7 shows the simulated maximum surface elevations and the estimated maximum water elevation using the method of amplification factors (NGI, 2009) for the Cotabato Trench scenario. Maximum water levels of 3-8 m are found in the Moro Gulf and along the South West Mindanao, which are roughly consistent with reported run-up heights / maximum water levels from field surveys indicating a range of about 3.5-9 m with the highest run-up in Lebak close to the source (Badillo and Astilla, 1978).

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 15

Figure 7: Upper panel, maximum simulated surface elevation and directivity for the Mw 8.1 Cotabato Trench scenario. Lower panel, estimated maximum water level (m) using the method of amplification factors for the Mw 8.1 Cotabato Trench scenario.

4.2 Northern Sulawesi

The seismicity (Figure 8) in the Northern Sulawesi region is too complex to be easily summarized. A number of crustal blocks take part in a rotational plate motion. Scenarios are defined as dip-slip earthquakes located along the Minahassa Trench subduction zone north of Sulawesi. Two scenario locations north of the Minahassa trench are shown in Figure 8 and Figure 9. Since the seismicity is low

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 16 at shallow depths northwest of the Gorontalo region, the faultplane is buried here, whereas the fault plane north of Minahassa peninsula is situated at the surface. Proposed scenarios are shown in Figure 9 (initial water displacements). Both scenarios are pure dip slip events along the North Sulawesi Trench.

It was found that magnitude 7.9 earthquakes represent ‘credible worst case scenarios’ for this area. The lower bound return period for a magnitude 7.9 scenario within the North Sulawesi study region in Figure 4 is approximately 30 years, but again we stress that the return period of the individual scenarios are believed to be a few times longer. For a more complete discussion, see NGI (2009).

Figure 8: Upper panel: historical tsunami run-up heights / water levels colour- coded. Lower panel: map view of fault planes (red rectangles: shallow fault planes; green lines: trenches, black lines: transform faults) and CMT solutions colour-coded according to magnitude. See Table 1 for legend.

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 17

Figure 9: Mw 7.9 scenarios in the North Sulawesi Trench, the colour bar shows initial water displacements in meters.

Figure 10 shows the simulated maximum surface elevations for the two magnitude 7.9 scenarios. Estimated maximum water level in the Celebes Sea using the method of amplification factors are shown in Figure 11. For both scenarios, large maximum water level heights of more than 10 m are found along the northern coastline of Sulawesi, with a maximum of 18 m in the northern Gorontalo region. Moreover, significant maximum water level of 2-7 m is found in the far-field even as far north as in Malaysia and the Philippines.

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 18

Figure 10: Maximum simulated surface elevation and directivity for the eastern Mw 7.9 Sulawesi scenario (upper panel) and the western Mw 7.9 Sulawesi scenario (lower panel).

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 19

Figure 11: Estimated maximum water level (m) using the method of amplification factors for the eastern Mw 7.9 Sulawesi scenario (upper panel) and the western Mw 7.9 Sulawesi scenario (lower panel).

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 20

4.3 Manila Trench

Historical tsunamis and seismicity for the Northern Philippines are shown in Figure 12. Bautista et al. (2001) describe the Manila trench as a straight line from 13-18°N which swerves abruptly to ESE at latitudes lower than 13°N because of a collision of micro-continental fragments with Mindoro and Panay islands. Hayes and Lewis (1984) state that the rate of subduction along the northern Manila Trench is probably not extended so far south.

Figure 12: Left panel: tsunami wave heights colour-coded; Right panel: map view of fault planes (red rectangles: shallow fault planes; green lines: trenches, black lines: transform faults) and CMT solutions colour-coded according to magnitude. See Table 1 for legend.

Two magnitude 8.2 scenarios are defined in Figure 12 (each one composed of three single faults) west of Batangas and Subic Bay, overlapping in their middle parts. For the northernmost scenarios, two additional magnitude sizes are proposed, resulting in north-south shrinkage of the fault plane, giving in total four scenarios for the Manila Bay.

The lower bound return period for a magnitude 8.2 earthquake along the Manila Trench is about 120 years, obtained from a Gutenberg-Richter relation on the full domain in the right panel in Figure 12. However, the return period for the individual scenario is expected to be a few times longer, as elaborated by NGI (2009). Lower bound return periods for the magnitude 8.0 and 7.6 earthquakes are about 75 and 30 years, respectively, and again the scenario return periods should be clearly longer than the lower bound. The initial water elevations for the four scenarios are shown in Figure 13.

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 21

Figure 13: Manila Trench Scenarios, initial water displacements (m). Upper panels Mw 8.2 scenarios; lower left panel Mw 8.0; lower right panel Mw 7.6.

Figure 14 shows the simulated maximum surface elevations for the Manila Trench scenarios. Estimated maximum water levels using the method of amplification factors are shown in Figure 15. As expected, the magnitude 8.2 scenarios give more widespread effects than the scenarios with smaller magnitudes. Still, the smaller scenarios also display more than 5 m maximum water level along the shoreline facing the source area.

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 22

Figure 14: Maximum simulated surface elevation and directivity for the Manila Trench scenarios. Upper left, southern Mw 8.2 scenario. Upper right, northern Mw 8.2 scenario. Lower left, Mw 8.0 scenario. Lower right, Mw 7.6 scenario.

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 23

Figure 15: Estimated maximum water levels using the method of amplification factors for the Manila Trench scenarios. Upper left, southern Mw 8.2 scenario. Upper right, northern Mw 8.2 scenario. Lower left, Mw 8.0 scenario. Lower right, Mw 7.6 scenario.

Figure 16 shows a merged plot of the estimated maximum water levels obtained by assembling largest maximum water level for all the four Manila Trench scenarios. Coastlines on the western part of Luzon Island facing the earthquake sources are subject to large maximum water levels of more than 10 m. Further south, also the Mindoro Island is strongly affected.

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 24

Figure 16: Estimated maximum water levels using the method of amplification factors merged for the Manila Trench scenarios.

4.4 Calapan landslide tsunami scenario – Mindoro Island, the Philippines

Two tsunami scenarios generated by landslides located northwest of the city of Calapan, Mindoro Island, the Philippines are analysed here, Figure 17.

The simulated landslides have a length of 1 km, a width of 1 km, and thicknesses of 20 m and 40 m were considered. The landslides are rounded to avoid artificial grid noise, giving volumes of about 26 and 52 Mm3. The maximum velocity of the slide is 20 m/s with a total run-out of 6 km in the South North direction.

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 25

The scenario computations serves both the purpose of a demonstration of modelling techniques for landslide generated tsunamis, as well as illustrating the required size and dynamics for a landslide to generate waves comparable to the 1994 Mindoro Island tsunami (Imamura et al. 1995). However, these computations should not be interpreted as a hindcast of the 1994 event for several reasons. First, the geological evidence for past landslides is not available, and it is not known if this tsunami was generated by the strike slip earthquake or a landslide. Second, no major landslide has been detected to these authors knowledge. Conducting the necessary seabed investigations for detecting landslides would require large resources. Finally, inundation simulations and explicit comparison with run-up from field investigations are not conducted.

For the tsunami propagation the GloBouss model described by NGI (2009) is used. The slide is modelled through sink-source distribution where volume displacement of the water is taken into account. For the computations, a merged bathymetry with a grid resolution of approximately 100 m (0.001o) containing two fine grid data sets outside Calapan and Puerto Galera, respectively (kindly provided by PHIVOLCS), merged with the GEBCO 1’ bathymetric data is used. The maximum surface elevation of the two slide scenarios (Figure 17), and some mariograms (time history of surface elevation) are shown (Figure 18). The figures reveal that the order of magnitude of the tsunami elevation is probably comparable to the 1994 event close to Calapan, whereas the simulated waves in the far-field seem to be clearly smaller. The latter may indicate other or a separate generation mechanism. The distribution of the surface elevations seems not to support a landslide source, at least not for the landslide location chosen here. Other landslide locations have not been tested.

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 26

Figure 17: Upper panel, situation map, yellow marker indicates the landslide position. Mid and lower panels, simulated maximum surface elevations for the two landslide scenarios of 26 and 52 Mm3, respectively.

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 27

Figure 18: Time series evolution of the simulated landslide generated waves. Upper panel 26 Mm3, lower panel 52 Mm3. The locations of the time series are given in Figure 17.

5 Batangas, the Philippines

5.1 Inundation model for the southern magnitude 8.2 Manila Trench scenario

Here, tsunami inundation simulations for the city of Batangas are presented, using input from the wave propagation model for the southern magnitude 8.2 Manila Trench scenario (see above). The inundation model ComMIT (Titov and Synlokis, 1995; 1998; ComMIT 2008) is used for this purpose, using a sequence of nested grids with the finest resolution approximately 30 m. The bathymetry is refined from the publicly available GEBCO 1’ grid, however, information from ASTER data is used for generating the local topography for Batangas. It is noted that both ASTER and SRTM are hampered with artificial effects due to elements such as forest and buildings, which causes additional land elevation to be smeared out over each pixel (of approximately 30 m length). Some of these effects, due to large industrial structures, were removed using information from the QUICKBIRD photographs. However, although it is expected from the QUICKBIRD image that additional artefacts in the topography remained, it was

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 28 not feasible to perform further corrections to the applied topography as this would require field investigations.

Two cases were investigated, a first simulation with no correction to the water level, and a second simulation including a 1 m sea level rise. The tsunami arrival time in the city of Batangas is about 1 hour 45 minutes for these cases.

The first scenario serves as a hazard map for a tsunami generated by southern Mw 8.2 Manila Trench scenario for a normal sea state (mean tide). Figure 19 shows the simulated water level for the overland flow as well as the flow depth, illustrating that the inundation is quite limited for this case.

The second scenario was constructed mainly for demonstration purposes, including the effect of a high tide (up to 0.7 m obtained from online tide tables, www.pgyc.org/tide%20tables.html). Thus the addition of 1 m water elevation due to tides alone may be considered overly conservative. The excess 0.3 m may therefore be interpreted due to other effects such as sea level rise due to warming or storm surges, but one should really keep in mind that the main purpose is for demonstration of the risk assessment method described below. Compared to the first study, it is shown that the addition of the 1 m extra water elevation (mainly due to high tide), has a distinct effect, giving clear increase in the inundated distance in various low lying areas, Figure 20.

Figure 19: Inundation simulation for the southern Mw8.2 Manila Trench scenario using the ComMIT model, maximum water level (m) in Batangas. No effects of due to tide, sea level rise, or storm surge are included in this figure.

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 29

Figure 20: Inundation simulation for the southern Mw8.2 Manila Trench scenario using the ComMIT model, maximum water level (m) in Batangas. For demonstration purposes, the mean sea is elevated 1m due tide, sea level rise, or storm surge for the simulation displayed in this figure.

5.2 Vulnerability and risk assessment

The example vulnerability and risk assessment of Batangas City combines a manifold of methods, including building vulnerability, population exposure, and mortality, as well as the framework to combine these elements.

5.2.1 Critical facility map

Building use (such as hospitals, emergency services, tourist facilities) was assessed for an example of a total of 74 buildings using information available from GoogleEarth (http://earth.google.com/) and GoogleMaps (http://maps. google.com/). This ‘building use’ was then categorized into 16 categories. The data was plotted as a ‘Map of critical facility distribution’ exemplified for the 74 surveyed buildings. The combination with the flow depth information yields the “Critical facility map” (Figure 21).

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 30

Figure 21: ‘Critical facility map’ for an example of 74 assessed buildings in Batangas City.

The ‘Critical facility map’ was further refined by differing between the importance of the facilities in the acute/short-term, i.e. right after a hazardous event, and in the long-term, i.e. weeks up to several months after an event while re-establishing activities are ongoing.

The present example study shows the following results for the southern Mw8.2 Manila Trench scenario (with 1 m sea level rise) for Batangas City: • Batangas City was historically confined to the area of the ‘Poblacion’, approximately 1 km from the present shoreline. The region occupied by large industrial plants (First Gas Power Corporation, Caltex Refinery, Cocochem, etc.) today, was originally marshland, including large fish pond areas. Parts of this area, mainly to the South of the present harbour and on the estuary island of the Calumpang River, seem to have become more densely populated later. • The above mentioned facilities (refineries, power plants, coconut oil industry, etc.) will pose an imminent environmental threat when impacted by a tsunami (damaging of pipes, leakage, seepage from waste water basins, etc.). • This would, expectedly, have a negative impact on tourism, as it would presumably lead to polluted beaches and near-shore life. • The historical settlement development prevented, to a large extent, the establishment of critical infrastructure close to the shoreline.

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 31

• The main arterial roads are situated well beyond the tsunami impacted areas. • The existing critical infrastructure, such as the harbour and its associated buildings (Philippine Ports Authority (PPA) Building, Bureau of Customs, etc.) are protected by the main harbour infrastructure with a harbour basin confined by high walls. Also the modern piers are of solid nature and reach several meters above the water level. • There are some affected critical facilities, such as the Lyceum International Maritime Academy, or the PagAsa Centre (Hope Centre).

5.2.2 Mortality risk map

Empirical data shows a correlation between maximum flow depth and the percentage of people killed. Such data are used to derive a relation between flow depth and mortality with upper and lower bounds to include the uncertainty in the data and influences of other parameters, such as, e.g., physical environment and vulnerability, on the mortality (NGI, 2009f; Eidsvig and Medina-Cetina, 2008).

Hence, the flow depth is used to calculate a mortality percentage map, i.e. how many percent of the people present (in each raster cell) would perish due to a given flow depth. The percentage map is finally used to calculate the number of fatalities, i.e. the mortality risk map (Figure 22), by combining the percentage map information with the average population density map.

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 32

Figure 22: Mortality risk map for Batangas City for the “southern Mw8.2 Manila Trench scenario”. Spatial raster resolution = 33 m x 33 m, i.e. 1 cell = 1089 m2.

Based on the southern Mw8.2 Manila Trench scenario (with the mean sea elevated 1 m due to tide, sea level rise, or storm surge), the calculation procedure outlined above indicates that approximately 500-600 persons would be lost due to the tsunami. The most affected regions would be: • The area at the river mouth of the Calumpang River, mainly in the districts (‘Barangays’) of “Wawa” and “Malitam”, • The shore and near-shore areas just SSE of the southern pier area of Batangas City port (‘Barangay’ “Cuta”), • The shore settlements NE of the main port basin (‘Barangay’ “Sta. Clara”). The high number of fatalities in all these areas is mainly the result of a combination of three factors: (a) dense population, (b) high building vulnerability, (c) medium high flow depth. The study was performed under the assumption of no tsunami warnings and did not consider any evacuation plans.

6 Mitigation measures

The evaluation of appropriate tsunami mitigation measures presented in this report is based on a literature study performed to establish state-of-the-art. The study has looked for documents outlining general tsunami mitigation strategies as well as documentation of mitigation measures implemented in practice. To be effective, tsunami mitigation needs to be part of and integrated into the general development planning for the society. Both “soft” mitigation measures such as awareness building, education, planning (development/evacuation/emergency) as well as “hard” techniques such as warning systems, barriers, design principles for buildings, land use, and infrastructure must be considered.

Further evaluations are related to what mitigation measures a society chooses to implement based on a number of factors beyond the risk a tsunami poses. In a wider context other risks, as well as the general cultural, societal, institutional, economical, and environmental situation, need to be taken into account.

Possible mitigation measures for the city of Batangas are identified. Based on the tsunami character (arriving time from source, wave height), the following mitigation measures are therefore found appropriate for Batangas City (and adjoining areas): 1. A detailed Early Warning System (EWS). The system must include appropriate ways of warning dissemination, proper evacuation routes to safe areas, and plans for evacuation of specific vulnerable parts of the population. In addition a good education system established for the population in question is required. Regular exercises of evacuation should be included in the EWS. A study made by Sugimoto et al. (2003) concludes that the mortality decreases significantly with increasing alarm

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 33

time. The death toll for a non-evacuation situation was found to be 10 times higher than for a situation where evacuation was considered, even though the evacuation time was as little as 10 minutes.

Since the wave height is rather limited, raising a barrier at the shore in the district of the exposed population may be a proper protection of the exposed population. A barrier of 2 m above sea level properly located will most likely reduce the wave impact significantly and hence reduce the consequences in the most exposed areas of Batangas city correspondingly.

7 Acknowledgements

The NGI and NORSAR staffs in charge of producing this report greatly acknowledge the contributions made from the coordinating committee CCOP under the guidance of Mr. Niran Chaimanee, and from our partner in the Philippines, the Philippine Institute of Volcanology and Seismology (PHIVOLCS), under the leadership of Dr Renato Solidum.

8 References

Badillo, V.L. and Astilla, Z.C. (1978), Moro Gulf tsunami of 17. August 1976. Report to the special committee on tsunami warning system, National Committee on Marine Sciences, National Science Board, Philippines Barrier, E., Huchon, P., and Aurelio, M. (1991). Philippine Fault: A key for Philippine kinematics. Geology, 19, 32-35. Bautista, B.C., Bautista, M.L.P., Oike, K., Wu, F.T., and Punongbayen, R. S. (2001). A new insight on the geometry of subducting slabs in northern Luzon, Philippines. Tectonophysics, 339 (3-4), 279-310. ComMIT (2008). Software and documentation, URL: http://nctr.pmel.noaa.gov/ComMIT/index.html Eidsvig, U. and Medina-Cetina, Z. (2008). Quantitative Risk Assessment for Tsunamigenic Rockslides: The Åknes Case Study. ICG report 2006-2-3 for the EU Project IRASMOS. Hayes, D.E. and Lewis, S.D. (1984). A geophysical study of the Manila Trench, Luzon, Philippines: 1. Crustal structure, gravity and regional tectonic evolution. J. Geophys. Res., 89 (B11), 9171-9195. Imamura, F., Synolakis, C.E., Gica, E., Titov, V.V., Listanco, E., and Lee, H.J. (1995). Field Survey of the 1994 Mindoro Island, Philippines Tsunami. Pure and App. Geoph., 144 (3-4), 875-890. Kreemer, C., Holt, W.E., Goes, S., and Govers, R. (2000). Active deformation in eastern Indonesia and the Philippines from GPS and seismicity data. J. Geophys. Res., 105 (B1), 663-680.

Document No.: 20061179-00-226-R Date: 2009-11-06 Page: 34

NGI (2009). Tsunami Risk Reduction Measures phase 2 – Main Report. NGI report no. 20061179-00-3-R. Okada, Y. (1985). Surface deformation due to shear and tensile faults in a half- space. Bull. Seismic Soc. of Am., 74 (4), 1135-1154. Okal (2009). Epicenter location of the 1976 Moro Gulf tsunami, personal communication. Rangin, C., Le Pichon, X., Mazzotti, S., Pubellier, M., Chamot-Rooke, N., Aurelio, M., Walpersdorf, A., and Quebral, R. (1999). Plate convergence measured by GPS across the Sundaland/Philippine Sea Plate deformed boundary: the Philippines and eastern Indonesia. Geophys. J. Int., 139, 296– 316. Stewart, G.S., and Cohn, S.N. (1979). The 1976 August 16, Mindanao, Philippine earthquake (MS = 7.8) – evidence for a subduction zone south of Mindanao. Geophys. J. R. Astr. Soc., 57, 51-65. Sugimoto, T., Murakami, H., Kozuki, Y., Nishikawa, K., and Shimada, T. (2003). A human damage prediction method for tsunami disasters incorporating evacuation activities. Nat. Haz., 29, 585-600. Titov, V. V. and Synolakis, C. E. (1995). Modeling of breaking and nonbreaking long-wave evolution and runup using VTCS-2. J. Waterw., Port,Coast., Ocean Engrg., 121(6):308–316. Titov, V. V. and Synolakis, C. E. (1998). Numerical modeling of tidal wave runup. J. Waterway, Port, Coastal, and Ocean Engineering, 124(4):157– 171. UNESCO-IOC (2006). Tsunami Glossary