EAP DRM KnowledgeNotes Working Paper Series No. 25 disaster risk management in east asia and the Pacific 63414 Tsunami risk managemenT in The conTexT of The Pacific islands By Dale Dominey-Howes & James Goff The general conTexT Tsunamis can be devastating. The 2004 Indian Ocean and 2011 Tōhoku disasters provide frightening exam- ples of the power of tsunamis. The Pacific has long been recognised as a place where tsunamis occur - the “Pa- cific Ring of Fire” (PRF) contains regions of volcanoes and large earthquakes associated with tectonic plate motions that are ideal breeding grounds for tsunamis. The Pacific Ocean covers an area of 30 million km2. Some 22 Pacific Island countries and territories (PICTs) are dotted throughout the Pacific and are vulnerable to varying degrees, to the effects of tsunamis generated locally, regionally and distantly1. We are all familiar with the large, trans-oceanic (‘distant’) tsunamis that occurred repeatedly in the 20th and 21st cen- turies in the Pacific (e.g., 1946 and 1957 Aleutian; 1952 Kuril-Kamchatka; 1960 Chile; 1964 Alaska and 2010 Chile events). These events are conspicuous in that they all originated in circum-Pacific locations, but less well-known local and regional sourced events are also significant. Tsunamis (local, regional and Pacific-wide) in the Pacific have claimed numerous lives, caused widespread damage to coastal infrastructure and heavily impacted natural ecosys- tems (Box 1). Large and destructive events can take years (perhaps decades) to recover from and can seriously affect long-term sustainable development – especially in rapidly developing economies. ■■ This is recognised by PICTs in their awareness of the need for tsunami risk mitigation measures. In particular, they recognise the need for improved community awareness and standard operating procedures (SOPs) for in- ternational and in-country communication of tsunami warning2. This working paper series is produced by the East Asia and Pacific Disaster Risk Management Team of the World Bank, with support from the Global Facility for Disaster Reduction and Recovery (GFDRR). The series is meant to provide just-in-time good practice examples and lessons learned from projects and programs related to aspects of disaster risk management. 2 disaster risk management in east asia and the Pacific Box 1. notable Pacific tsunamis and examples of their impacts and lessons learned3 2011: 11 March, Japan Earthquake: Pacific-wide 2010: 13 April, Cook Islands Submarine Landslide: Local* 2010: 27 February, Chile Earthquake: Pacific-wide 2007: 1 April, Solomon Islands Earthquake: Regional 2003: 25 September, Japan Earthquake: Local 2001: 23 June, Peru Earthquake: Local 1999: 26 November, Vanuatu Earthquake: Local 2009: 29 September, Samoa, American Samoa, Tonga impacts: 192 deaths; destruction of coastal infrastructure, etc. figure 1: The Pacific ocean with examples of local, lesson: good community regional, and distant tsunami sources. awareness and evacuation planning saved lives; heed Yellow oval = local source (earthquake & submarine landslide natural warning signs – Papua New Guinea 1998), red triangle = regional source (submarine caldera collapse – Kuwae (ku), 1452/1453AD; 1998: 17 July, PNG Healy (h), c.1280-1350AD); orange line = regional source 29/9/2009 South Pacific earthquakes; green lines = regional/ impacts: 2,205 deaths; distant sources, representative subduction zone segments from destruction of coastal various circum-Pacific Ocean country (CPOC) source areas. settlements Light blue filled black circles show locations of PICTs with dated lesson: Local tsunamis require prehistoric tsunami evidence. careful response planning (e.g. need for vertical evacuation options), heed natural warning signs Given the general context outlined, this Knowledge *Notable for the fact that this is currently not in NGDC database Note does the following. It: ■■ acknowledges the impacts and characteristics of re- Whilst this current state of awareness is encouraging, cent damaging tsunamis (with a particular focus on our understanding of the medium to longer-term (hun- Pacific and PICT events); dreds to thousands of years) recurrence of these events ■■ outlines what we understand the ‘tsunami risk man- is limited. We know almost nothing about the tsunami agement framework’ to involve; record for PICTs, nor the frequency and magnitude of ■■ outlines the challenges in identifying, assessing and locally and regionally sourced tsunamis in particular, as monitoring tsunami risk (using the risk manage- opposed to pacific-wide events. Given the general vul- ment framework identified as a guide); nerability of PICTs already noted, the lack of a detailed ■■ considers the variety of approaches and methods for and well-dated long-term record of all sources and prevention and mitigation of tsunami disasters; and events stands as a significant obstacle for the develop- ■■ identifies good practice cases and makes a series of ment of comprehensive tsunami risk mitigation mea- recommendations to move tsunami risk manage- sures4 (Figure 1). Successful ‘tsunami disaster risk re- ment forward. duction’ efforts require as a fundamental building block, a reasonable and reliable estimate of tsunami risk5. imPacTs of recenT Tsunamis & Their characTerisTics The majority of tsunamis are thought to be generated by earthquakes below the sea floor. Importantly how- ever, they may also be generated by volcanic eruptions, underwater landslides, asteroid/comet impacts in to Tsunami risk management in the context of the Pacific islands 3 the ocean and occasionally, meteorological conditions. pletely and huge damage was occurred although pre- However, things are not quite that simple, the Pacific paredness mitigated losses dramatically. also experiences unusually large tsunamis associated For PICTs such as Samoa, even a moderate event such with poorly understood processes operating at subduc- as the 2009 South Pacific tsunami caused by an magni- tion zones6. These include “tsunami earthquakes” where tude 8.1 earthquake was its worst natural disaster in at larger than expected tsunamis are generated by “slow” least half a century with nearly 150 dead and 2.5% of earthquakes7 and by earthquakes that simultaneously the population left homeless. The final physical dam- generate submarine landslides8. age repair bill is expected to be around US$ 85 million In September 2009, yet another unexpectedly large (14% of GDP). However, when the additional costs of tsunami resulting from an unusual earthquake event maintaining basic social services and safety nets for the occurred in the South Pacific9. In essence, we are con- affected population and the costs for investing in disas- tinuing to experience larger tsunamis than anticipated ter risk reduction during the reconstruction process are by current numerical modelling scenarios. This is of considered, total economic cost is equal to about 21% of enormous concern for the Pacific (and PICTs) where GDP over the next three to four years13. attention has largely been focussed on subduction zone It is also important to remember that tsunamis do not events with little or no consideration given to regional respect geographic boundaries and may cause losses tectonic and submarine landslide sources that can be across entire (small) nations or the coastal zones of sev- equally important for individual PICTs10. This is sig- eral countries simultaneously. Impacts and effects of- nificant because, local and regionally generated events ten ‘ripple’ out across connected socio-economic and pose the greatest challenge for effecting warning alerts human-environment systems and transcend scales from and ensuring adequate community response (e.g. evacu- the local to the global. For example, the 2009 South Pa- ation)11. cific tsunami affected Samoa, American Samoa, Tonga It is not the purpose of this Knowledge Note to provide and even caused damage as far away as the Wallis and a comprehensive analysis of the impacts and effects of Futuna archipelago9,14. tsunamis of various magnitudes. However, in common with other types of natural hazards, tsunamis can cause extensive loss of life and injuries to survivors. Tsuna- WhaT does ‘Tsunami risk mis destroy and damage public and private coastal in- managemenT’ inVolVe? frastructure, lifelines, critical assets and infrastructure, It is important to outline what is understood as the ‘risk agricultural systems and produce, transport and com- management process’. By this, we mean, what steps are munication networks, natural ecosystems and the goods followed to understand and quantify the risk so that ap- and services those systems provide. Major tsunamis can propriate, context-specific, actions can be undertaken to lead to significant economic losses with recovery often reduce risk. Individual countries and organisations will counted in years to decades. Even then, a return to eco- use their own variations of a risk management standard, nomic growth does not compensate for the direct hu- but all such standards comprise the same basic ele- man and economic loss caused by the event. ments. Essentially, risk is defined by a simple equation: In the case of the 2011 Great East Japan Earthquake and Tsunami, the final toll will be immense. The num- risk = hazard x vulnerability ber of dead and missing exceeds 23,000 and economists have estimated the monetary loss at 16-25 trillion yen First, identify the tsunami hazard. The hazard must (US$198-310 billion)12, about 3-5% of Japanese GDP then be quantified. Once the hazard has been quantified and the effects of the Fukushima Daiichi nuclear plant and the characteristics determined, a range of possible remain unclear. The take home message from this event scenarios can be selected for analysis. Such scenarios is that even for the most well prepared Pacific country will likely be selected from a probabilistic assessment for tsunamis, it is impossible to prevent disaster com- of the hazard. For any given scenario of interest, the 4 disaster risk management in east asia and the Pacific likely effects of the tsunami may be assessed. To do this, constraints of the subduction system and its capacity first the exposure of people, infrastructure and assets to generate such large events were not fully recognised. (e.g. agricultural systems, communication and transport In many ways this problem hinges on a key issue re- networks, etc) in the forecast inundation zone must lated to assessing the tsunami hazard for any region – a be identified and then, critically, their vulnerability to lack of context. Until we have a better understanding harm evaluated. Once this is done, loss may be estimat- of how subduction zone geology/geophysics constrains ed by risk managers. Once likely loss is known for an maximum earthquake size, or we can be sure that our event of a given magnitude and probability, decisions knowledge of the tsunami record is complete, we must can be made about the appropriate risk management/ assume that any subduction zone can generate a large mitigation options to follow. tsunamigenic earthquake. We should therefore strive to improve understanding of past (historic and prehis- toric) events in order to better understand our future. Clearly, experts from different discipline fields will be involved in generating data associated with each step One recently developed method to undertake quantita- of this process. For example, generally speaking, earth tive tsunami hazard assessment is the Probabilistic Tsu- scientists are involved in the identification and charac- nami Hazard Assessment (PTHA) that is based on nu- terisation of the tsunami hazard (from not only earth- merical simulations of what are believed to be ‘plausible’ quakes but all possible tsunamigenic sources). Numeri- tsunami scenarios. Such an assessment has been conduct- cal modellers, oceanographers and coastal engineers will ed for SOPAC (Secretariat of the Pacific Community, model the tsunami from source to inundation. Social Applied Geoscience and Technology Division) member and human scientists and engineers will work to evalu- countries15. This approach accounts for large earthquake ate exposure and vulnerability of relevant people, in- occurrence on all subduction zones, even those that are frastructure and assets, and a range of experts such as not known to have generated large tsunamis. However, risk managers, engineers, social scientists, economists, these deal almost exclusively with reasonably simple sub- geographers, etc. might be involved in the quantifica- duction zone source scenarios using historical data as tion of the probable maximum loss (PML) associated their primary contextual source. The nature and extent of with the scenario. Lastly, risk managers have the task of larger events are in general extrapolated from these his- integrating all these datasets and decision makers de- torical data coupled with a rudimentary understanding of termine how best to mitigate the risk identified. All this the geophysical properties of the fault zone in question. should happen and the risk management actions should Scant consideration is given to Traditional Environmen- be in place before the next tsunami occurs. tal Knowledge (TEK) about past events and geological data concerning prehistoric tsunamis16. challenges in idenTifYing, While this approach offers a ‘general picture’ of the tsu- assessing & moniToring Tsunami nami hazard, it falls short of offering a comprehensive risks representation of the problems faced by PICTs. There is currently almost no grasp of local and regional volca- There are a number of important challenges that exist nic-related tsunamigenic sources and processes in the in relation to identifying, assessing and monitoring the Pacific (e.g., eruptions, caldera collapse, flank collapse risk from tsunamis. We discuss these in relation to each etc). For example, an eruption at Ritter Island, PNG in element of the risk management process identified in 1888 produced a significant local volcanic tsunami and the previous section. the 1452/1453AD eruption at Kuwae, Vanuatu pro- The first point is our limited understanding of tsunami duced a catastrophic region-wide tsunami only recently sources – the hazard assessment. Prior to the 2004 In- recognised in the geological record4,16. Many PICTs are dian Ocean disaster, few (if any) had imagined the pos- still volcanically active and form part of the PRF, while sibility of extremely large (Magnitude 9+) earthquakes those not directly associated with it are generally either along the Sumatra subduction zone. The geophysical linked to hot spot volcanism, mid-ocean ridges, or past Tsunami risk management in the context of the Pacific islands 5 tectonic activity. Perhaps most importantly, because water depth, speed, runup). Furthermore, modelling most PICTs are volcanic in origin they rise up steeply tsunami generation, propagation and inundation faces 1000’s of metres from the seafloor and as such are sus- the challenge of a lack of detailed offshore bathymetry ceptible to tsunamigenic landslides10 (see Box 2). and onshore topography including LIDAR datasets which both act as impediments to producing realistic Box 2. cook islands: an example of the inundation and runup forecasts. Equally, current inun- disjunct between existing assessments & dation models are based on bare earth topography and geological evidence fail to consider land surface roughness and vegetation, people and infrastructure. Finally, aspects of later arriv- While recognising that other sources exist, current PTHA data indicate that the Tonga-Kermadec Trench (TKT) is the most ing and receding waves, and the interaction of debris or significant source of tsunamigenic earthquakes for the Cook projectiles have not been included in many modelling Islands with 2000 year maximum amplitudes of around 1.7 m studies. for the northern islands and up to 2.8 m in parts of the southern group. A limited regional threat from the South Solomon and It is in the reporting of the results and dissemination New Hebrides trenches to the west and large earthquakes on the Kuril and Peru-Chile trenches may also represent a distant of the data through peer-reviewed journals and consul- tsunami threat16,17. tancy reports that often a necessary over-simplification takes place with many of the provisos, assumptions and limitations of source event characeteristics and model- ling approaches, not being fully considered, clearly ex- plained and justified and/or even overlooked entirely. Further, where such limitations are correctly described and discussed, these often fail to filter through to the emergency management professionals and wider pub- lic community for whom these studies are designed to benefit. The task of risk communication is fraught with difficulties and we acknowledge that experts do the best they can in the current circumstances18. Information transfer must be effective to ensure that the relevant It is therefore surprising to discover that a locally-generated tsunami (see Box 1) caused by a submarine landslide gener- messages to the different stakeholders are passed on and ated runup heights of up to 12 m on Mangaia10. understood. Assessing AND monitoring of tsunami risk for PICTs is still in The geographic remoteness, high cost of exploration, and its infancy. a relative lack of scientific interest in Pacific Island tsu- nami research have tended to act as barriers to detailed On balance, it seems reasonable to suggest that simple in-country PICT studies. Low population numbers and subduction zone events may represent as little as 50% a perceived limited infrastructure exposure have resulted of the potential tsunamigenic sources for some PICTs in a general lack of interest in understanding the complex (e.g. Cook Islands, Kiribati, French Polynesia)10. tsunami hazard and risk posed to PICTs. This is unfor- tunate since much PICT infrastructure, such as wharves While these challenges are not insurmoutable, and in- and airstrips, is in coastal or low-lying areas, and as such deed science advances by addressing such issues, for the is particularly vulnerable to tsunamis that can destroy time being all we can safely say is that we only have a the sole means of obtaining essential supplies4. Further, rudimentary knowledge of the risks posed by tsunamis to our knowledge, tsunami risk assessments have mostly to PICTs. failed to take in to consideration future sea level rise asso- Even the modelling of moderately simplistic tsunami ciated with enhanced anthropogenic climate change. As hazard scenarios is complex - all the way from source such, future tsunamis may well be even worse than our (e.g., earthquake, landslide, volcano) to inundation (e.g. current best estimates as inundation may change mark- 6 disaster risk management in east asia and the Pacific edly as sea levels rise. This assumption however, remains ence that has only been properly explored in the last 10 to be carefully considered. years or so. The next element of the risk management process that Significant efforts are underway to develop, test and has difficulties and challenges relates to the assessment validate (using post-tsunami building damage assess- of exposure and vulnerability. Quantifying PML for ments) semi-quantitative and quantitative engineering any given scenario relies on two fundamental datasets: models for tsunami damage assessment. Such models (1) data about who and what are ‘exposed’ to potential (in keeping with seismic, flood and wind hazards) are harm; and (2) an effective method for estimating the attempting to integrate damage functions and fragil- vulnerability (and resilience) of those exposed people ity curves established from empirical field evidence of and assets (so called ‘vulnerability assessment’). damage sustained by differently engineered structures in real disasters. Such models may then be applied in a Assessment of exposure is reasonably simple but is of- forecast sense to estimate damage and loss. ten hampered by a lack of detail. Population census sur- veys are often only undertaken once every five or ten Notwithstanding the difficulties faced by engineers in years and in rapidly developing countries, population establishing and validating such models, their applica- numbers can increase quickly – especially in urban and tion in forecast assessments is limited by the need for peri-urban centres. As such, having reliable estimates detailed datasets of the built environment at a high-res- of human exposure can be challenging including know- olution (building-by-building) scale. That is, what built ing the absolute numbers of adults and children, elderly environment characteristics are relevant for estimating and disabled, etc. For example, after the 1998 PNG tsu- vulnerability (e.g., building materials, number of floors, nami, authorities were unclear as to how many people design standards, etc.). had actually died (and their demographic composition) Studies of the September 2009 South Pacific tsunami, because they did not know how many people lived in as well as work by other agencies have made signifi- some of the affected communities at the time. cant advances in constructing and validating building Similarly, knowing the exposure of the built environ- fragility curves to aid in built environment vulnerability ment can be limited, especially where government re- assessments by undertaking post-tsunami damage as- cords of development planning and approvals are in- sessments of ‘typical’ Pacific structures20. This work is complete or non-existent. However, limited knowledge critical since it has demonstrated the value of the Pa- of the physical exposure of the built environment can pathoma Tsunami Vulnerability Assessment (PTVA) be overcome using modern satellite and air survey/pho- Model that is currently, to the best of our knowledge, tographic techniques. For example, freely available im- the only detailed model available to estimate building agery and data through open access platforms such as vulnerability to tsunamis21,22. OpenStreet Map can provide a reasonable mechanism The PTVA Model has been tested and applied in sev- for understanding building exposure. Another example eral locations including Australia, the Pacific coast of is the Pacific Catastrophe Risk Assessment and Financ- the USA, Greece, Italy, Malaysia, India and Sri Lanka ing Initiative that has established a substantive risk ex- and appears to be the most promising approach to such posure data base comprising of population, buildings, infrastructure vulnerability assessment. The challenge infrastructure and crops for risk modelling. The initia- in its use however, lies in its ‘data hungry’ inputs (e.g. tive is implemented jointly by SOPAC/SPC, the World building type, construction materials, number of floors, Bank and the Asia Development Bank and data is being ground floor plan layout, etc)23,24. launched in August 201119. Following on from the above, there are significant chal- Assessing the vulnerability of the built environment is lenges associated with calculating replacement costs for another issue altogether. The translation of data about physical infrastructure across PICTs. Without standard what is physically ‘exposed’ to tsunami inundation into methods/approaches and with widely variable costs in what is ‘vulnerable’ to damage and loss is an area of sci- each country, it is not a simple process to estimate fi- Tsunami risk management in the context of the Pacific islands 7 nal PML or replacement costs. Also, in supply-limited of the population are addressed, including young post-disaster situations, building material costs can es- and elderly, women, physically and mentally-chal- calate rapidly diminishing the relevance of pre-event lenged, religions, languages, culture; estimates. The strength of this methodology though is ■■ Disaster preparedness information that is a required that loss metrics have an associated probability of occur- part of school curricula, e.g. a scaffolded school cur- rence. A risk manager can therefore determine what re- riculum that reinforces and creates the appropriate turn period level of loss they would like to consider such educational messages throughout a student’s time as the 1:100 or 1:500 year event. However, one of the in the education system and one that serves to un- weaknesses of this approach is that it assumes a com- derpin other community awareness initiatives; prehensive knowledge of the country’s tsunami hazard ■■ Infrastructure and planning that is tsunami risk that as previously noted is rarely, if ever, achieved. aware – e.g., coastal setbacks, coastal landscape parks, open ground floor plans, building construc- Lastly, we recognise the value of using indigenous or tions styles; sea walls, breakwaters, barriers and traditional building practices as a possibly sustainable natural berms; coastal forest/mangroves and buffer and affordable best practice. Further research should be vegetation; where appropriate, no build zones are done to consider the viability of incorporating lessons designated or coastal communities are transitioned learned in Indonesia, Southeast Asia, and Japan into re- inland or to higher ground. covery planning and construction. ■■ Whole-of-government (and cross-sectoral) risk and In terms of monitoring risks, we would suggest that multi-hazard warning and disaster management after each major tsunami has occurred, relevant au- plans and policies that are mutually supportive; thorities should carefully review lessons learned and the ■■ An informed and proactive tourist industry that knowledge generated by post-event analyses in order to ensures not only the safety and well-being of their determine if our understanding of the risk has changed. clients, but also safeguards the sustainability of a PICTs GDP (tourism is the largest and fastest growing sector in the Pacific. It is estimated that aPProaches for PreVenTion & tourism income in the Pacific is around US$2 bil- miTigaTion lion per year and represents a significant part of the GDP of PICTs [e.g. 2/3 of Palau’s economy], port Prevention and mitigation of natural hazards is a stan- and harbour facilities for tourism and trade need to dard element of the risk management process. How- acknowledge marine risks)25; and ever, in the context of tsunamis it is simply not possible ■■ Annual memorial/hazard awareness days. to ‘prevent’ them from occurring. As such, risk manage- ment has to focus entirely on mitigating the effects of Emerging lessons from the recent Japan disaster suggest events when they occur. the following: Tsunami mitigation: - what does this mean? In practice, ■■ In risk management, there is always a significant mitigation for PICTs will have several elements. These trade-off between risk tolerance and acceptable include having: mitigation in that countries could never afford or ■■ Detection, monitoring and early warning systems may want to choose not to pay the 100% mitiga- (regularly maintained and tested to ensure they are tion costs for very infrequent hazardous events, but operational); it should be acknowledged that there are potentially ■■ Science-based tsunami hazard maps showing flood- catastrophic consequences associated with underes- ing impact (where, when, how big), locations of timating the hazard; important facilities (infrastructure, lifelines, evacu- ■■ Complete reliance on physical mitigation measures ation sites or safe areas, police/fire/hospitals, etc); (such as tsunami sea walls, tsunami forests and tsu- ■■ Educated, prepared and responsible coastal com- nami defence gates) is unwise as they may easily be munities (and relevant stakeholders). All segments overwhelmed and overtopped26. Again, we do rec- 8 disaster risk management in east asia and the Pacific ognise that such measures do lessen the impacts of ■■ At the international and regional level, the UNES- events in places; CO-IOC organised two Pacific-wide tsunami exer- ■■ For many political and practical reasons, ‘set back’ cises in 2006 and 2008 to encourage countries to pre- is frequently unrealistic. As such, exposure cannot pare for the next tsunami. The next Exercise Pacific be reduced thus increasing the importance of other Wave (PacWave) 2011 will take place 9-10 Novem- mitigative strategies such as community educa- ber as a multi-scenario exercise to allow countries tion, evacuation planning and practices and vertical to practice responding to local and regional sourced evacuation structures. tsunamis. For PICTs, earthquake sources along the ■■ There is a need for a paradigm shift in disaster risk New Hebrides and/or the Philippines Trench can management: one that accepts worst case scenarios be used. PacWave06, in which 44 countries partici- and includes them in planning and mitigation sce- pated, has been credited with saving lives in the 2009 narios. This recognizes the importance of under- South Pacific and 2010 Chilean tsunamis. standing that simple hazard assessments invariably ■■ At a national scale, prior to the 2009 South Pacific do not contemplate the complexity of the real world. tsunami, the Government of Samoa had developed an effective tsunami early warning system, working With specific regard to PICTs, a recent assessment of the with communities to raise public awareness and to tsunami capacity of SOPAC member countries by the practice evacuation drills and exercises. These ac- Australian Government Bureau of Meteorology (BoM) tions helped to save lives during the 2009 South has been undertaken2. This report highlights through a Pacific tsunami since the population in many cases process of consultation, that mitigative options like those knew how to respond14. In American Samoa, Sep- we identified above are entirely appropriate. However, of tember has been declared as ‘Disaster Awareness the 14 SOPAC member countries involved in this assess- Month’. During this month in 2010, the disaster ment, 13 or more recommended the need for robust and management office conducted numerous outreach and training activities for government and non- effective international and in-country SOPs for tsunami government agencies and schools, so that by the 29 warnings and dissemination to make communities aware. September, many people knew about tsunamis. Significantly, none of these 14 countries have complete ■■ At a local scale, the Island Council of Mangaia, Cook coverage for the effective dissemination of warnings to Islands, has passed a law stating that all new hous- their national communities and none have comprehen- es should be built inland and uphill away from the sive training programs for their officials involved in tsu- coast, in part taking on-board the results of recent nami warning and response. This suggests that at pres- geological studies of past tsunamis on the island27. ent, these PICTs are not as well prepared as they would ■■ In Guam, American Samoa and the USA, there like, or need to be and that their current tsunami risk is a community-based programme called Tsuna- mitigation processes are inadequate and limited. How- miReady which requires communities to have re- ever, in April 2010, as a positive sign of developing an dundant warning and alerting communications, integrated approach to tsunami risk management, the tsunami response plans, tsunami hazard and evacu- Solomon Islands obtained funding to undertake a coor- ation maps and signage, and active community and dinated, whole-of-government, private sector and NGO school education programmes. Similar efforts are approach to tsunami management by completion of a in place in the Philippines and being undertaken holistic National Tsunami Response Plan. in Commonwealth Caribbean countries. These are the essential components of an effective and suc- good PracTice cases cessful end-to-end tsunami warning. There is no single example of comprehensive good In reality, good practice will have to be ‘context specif- practice. However, elements of good practice have been ic’. No single model will fit all situations and given the adopted at international, national and local scales. diversity of PICTs, a single approach to good practice Tsunami risk management in the context of the Pacific islands 9 would not be appropriate. However, it is still possible to al tsunami warning centres. All such organisations ‘map out’ those concepts we see as contributing to good contain a wealth of existing experience, knowledge practice and these include: and skill in relation to disaster risk reduction across a range of hazard types and wherever possible, this ■■ Whole-of-government willingness to act and work col- existing regional experience and knowledge should laboratively to reduce risk; be utilised; ■■ A commitment to long-term funding of warning and ■■ Building of PICT skills and capacities of agencies and disaster risk management offices (within an ‘all-haz- especially its staff through training and outreach, es- ards framework’). Critical core stakeholders sup- porting disaster risk reduction are; the scientists pecially dedicated, multi-year, sustainable in-country providing hazard and risk information, tsunami training and outreach on tools, equipment and knowl- warning centres providing continuous threat assess- edge. Engaged and motivated personnel will lead to ment in real-time, and disaster management offices improved tsunami warning and response efficien- and first responders that are responsible for public cies across the region. safety. These stakeholders must work together and seamlessly in an emergency. An important mecha- Specific activities and programmes that constitute part nism for building a strong mitigation program is of the ‘map’ for reducing tsunami risk include: (1) build- the establishment of Tsunami Coordination Com- ing sustainable operations for tsunami early warning mittees that will oversee tsunami mitigation in a and alerting and tsunami emergency response through country or jurisdiction; well-known and exercised plans, procedures, and proto- ■■ Inter-PICT government and regional cooperation – cols; and (2) education with a focus on adaptive strate- shared technical and communications infrastructure gies that work for different languages and cultures. (e.g. seismic stations, coastal sea level gauges, deep- ocean DART© buoys, and in the future real-time GPS recommendaTions & WaYs networks, WMO (United Nations World Meteoro- logical Organization) Global Telecommunications forWard System, RANET communications project, etc.), and During the UNESCO-IOC ICG/PTWS-XXIII held information sharing during real events of observations in Apia, Samoa, Pacific countries endorsed its 2009- and impacts, will lead to improved tsunami detection 2013 Medium Term Strategy (MTS) to document the and threat assessment across the region; essential components and strategies for achieving tsu- ■■ Intergovernmental and regional organisations continue nami preparedness. The PTWS is envisioned as an “An to provide significant contributions to assist countries interoperable tsunami warning and mitigation system based to build their capabilities for hazard and risk assess- on coordinated Member State contributions that uses best ment, early warning and dissemination, and aware- practices and operational technologies to provide timely and ness and education. At the United Nations intergov- effective advice to National Tsunami Warning Centres. As ernmental level, the UNESCO-IOC coordinates a result, PTWS communities at risk are aware of the tsu- the global tsunami system that includes, since 1965, nami threat, reduce risk, and are prepared to act to save lives.” the Pacific Tsunami Warning and Mitigation Sys- The MTS builds from the ITSU (PTWS) Master Plan tem (PTWS). The UNESCO-IOC’s ITIC (In- (1999, revised 2004) that summarises the mitigation of ternational Tsunami Information Centre) serves as tsunami hazards in the Pacific. its regional focal point for training, technical assis- tance, and awareness. Regional organisations such The PTWS MTS is comprised of three Pillars sup- as SPC/SOPAC, support PICTs disaster risk re- ported by four foundational elements. duction and national disaster management offices, The Pillars are: and SPREP (South Pacific Regional Environment Programme) supports PICT national meteorologi- ■■ Risk Assessment and Reduction: hazard and risk cal services under the WMO in their role as nation- identification and risk reduction 10 disaster risk management in east asia and the Pacific ■■ Detection, Warning and Dissemination: rapid de- First, we need better science (e.g. geological, geophysical, tection and warning dissemination down to the last oceanographic, engineering, social, TEK, historical, envi- kilometre ronmental, etc.), better integration of the data and com- ■■ Awareness and Response: public education, emer- munications of these scientific results in an understandable gency planning and response and actionable way to enable meaningful public policy and to convince an often-busy citizen to pay attention to natu- The supporting foundational elements are: ral hazards and take action to prepare themselves and their community. This is not just about numerical modelling or ■■ Interoperability: free, open and functional exchange pure research, but using science to improve decision-mak- of tsunami information ing and to ultimately implement a sustainable and practi- ■■ Research: enhanced understanding and improved cal end-to-end warning and mitigation package. technologies and techniques ■■ Capacity Building: training and technology transfer Second, BUT at the same time, national and local gov- ■■ Funding and Sustainability: resources to sustain an ernments and communities need to get better at do- effective PTWS ing risk management. This can be achieved through improved government policy systems, better warnings, Within each Pillar, prioritised activities, guided by the community-wide consultation on awareness and educa- PTWS’s foundational elements, are to be undertaken tion and an integrated approach to tsunami risk within with the aim of making at-risk populations safer. an all-hazards framework. After the PTWS MTS was approved, the PTWS These two points have been made in some detail by Implementation Plan was developed setting forth Pri- the recent Australian Government BoM report dealing orities of Action to fulfil the strategy. Thus, PICTs are with PICT tsunami preparedness2. recommended to adopt the MTS and use the Imple- To fill in the existing knowledge gaps, our key recom- mentation Plan Priorities as a guide to improving their mendations are: tsunami preparedness. 1. Improve basic understanding of the hazard. This can During the ICG/PTWS-XXIV in May 2011, a tech- be achieved by: nical workshop was convened to share experiences and lessons learned from the recent regionally-destructive ■■ improving our long-term understanding of the tsunamis (2009 South Pacific, 2010 Chile, 2010 Ja- hazard by undertaking detailed country-by-country pan) and to discuss and elaborate on how effective the geological, archaeological and TEK investigations PTWS, both as a system and individually as countries, of past historic and prehistoric tsunamis in PICTs; has been in providing early, timely warnings to com- ■■ investigating all tsunamigenic sources (e.g. incor- munities at risk. Outcomes from the Workshop are in- porate volcanic- and landslide-related data into re- tended to serve as a catalyst for improving the system. gional and national tsunami databases for PICTs); The Workshop discussions and outcomes were used to ■■ using integrated source to inundation numerical formulate the PTWS Working Group recommenda- modelling that is underpinned by comprehensive tions to the ICG/PTWS-XXIV. geological, historical and TEK datasets to ensure that PTHA are robust; For PICTs, the relevant PTWS Working Groups are a ■■ undertaking studies of socially-oriented perceptions Regional Working Group for the South West Pacific, of tsunami hazard and risk to improve community Seismic Data Sharing for the South West Pacific Task understanding across the region; Team, Warning Communications Task Team, Exercise ■■ testing the assumption that PICTs are automatical- Pacific Wave 2011 Task Team, Enhanced Products Task ly ‘vulnerable’ to tsunamis. This is an over-simplis- Team, Risk Assessment and Risk Reduction Working tic generalisation. It appears that many PICTs have Group, and Awareness and Response Working Group. a built in cultural and system resilience (e.g. Samoa: There are two interrelated, but clear ways forward. developed in response to the 2009 South Pacific Tsunami risk management in the context of the Pacific islands 11 tsunami). We need to understand this so that this ■■ Providing telephonic communications services, existing resilience can be protected and enhanced, whether by fixed line or mobile (voice and/or text), and current vulnerabilities reduced or eliminated. that is affordable to everyone. It should not happen that many run out of mobile phone credits after a 2. Improve governance structures, processes, policies and disaster, such as happened in Samoa 2009 imme- protocols. This can be achieved by: diately after the tsunami. Government can work ■■ Ensuring that there is a genuine integration of an with the telecommunications industry to encourage all-hazards approach to risk management and bet- fair pricing or consider subsidies to allow for greater ter integration of Disaster Risk Reduction (DRR) use. with Climate Change Adaptation to ensure better ■■ Providing robust communications for warning and outcomes for communities and governments with emergency services, and for transmitting impor- limited financial and human resource capacity28; tant monitoring (earthquake and tsunami) data to ■■ committing to ongoing, permanent educational and warning centres. Communication before for receiv- outreach programs to improve community under- ing alerts, during for monitoring impact, and after standing; for declaring ‘all-clear’ safety and search-and-rescue ■■ incorporating other important stakeholder groups response to disasters is a critical necessity. in DRR for PICTs (e.g. tourists: tourism is a sig- nificant component of the GDP of many PICTs. If Last, the Australian Government BoM report high- tourists are not considered, significant market losses lighted that no PICT has a comprehensive grasp of the can occur when tourists go elsewhere following a tsunami risk from end-to-end. Furthermore, no PICT hazard/disaster, thus compounding the problems has a comprehensive tsunami risk management frame- faced during recovery efforts). work in place. As such, much work remains to be done and we advocate that many international and regional Tsunami risk management needs to adopt a ‘coupled organisations can play a significant role in addressing human-environment systems framework’ in conjunc- the gaps that exist at present. tion with the risk management process. Climate change and extreme climate events researchers are already us- endnoTes ing this approach. The tsunami science community has not yet caught up with other hazard scientists in using 1. Méheux, K. et al. 2007. Natural hazard impacts in small island developing states: a review of current knowledge and future re- this approach effectively. In essence, we are not good at search needs. Natural Hazards, 40, 429-446. understanding the human and environmental context in 2. Australian Government Bureau of Meteorology 2010. Re- which tsunami hazards become disasters nor the resil- gional consolidated report, SOPAC Member Countries Na- ience within coastal communities and ecosystems and tional Capacity Assessment: tsunami warning and mitigation how to preserve and enhance that resilience29. systems. BOM, Melbourne. 3. http://www.ngdc.noaa.gov/hazard/tsu_db.shtml 3. Improve communications infrastructure for infor- 4. Goff, J. et al. 2011. Palaeotsunamis in the Pacific. Earth Science mation sharing and event monitoring. This can be Reviews, 107, 141-146. achieved by: 5. Dominey-Howes, D. 2002. Documentary and geological re- cords of tsunamis in the Aegean Sea region of Greece and their ■■ Providing basic electronic information access to all potential value to risk assessment and disaster management. communities, especially through the Internet and Natural Hazards, 25, 195-224. computers. This is essential for providing knowl- 6. ‘Subduction zone’ - where two crustal plates collide on the sea- edge-building opportunities for everyone. Better and floor and one plunges back into the earth’s interior. This can generate large earthquakes. For further information see: http:// faster access at greater bandwidth will allow many pubs.usgs.gov/gip/dynamic/understanding.html more to learn much more, whether it is about tsu- 7. Kanamori, H. & Kikuchi, M. 1993. The 1992 Nicaragua earth- namis, hurricanes or other hazards, or about health, quake: A slow tsunami earthquake associated with subducted agriculture, education, or other day-to-day topics. sediments. Nature, 361, 714–716. 8. Goto, K. et al. 2010. Historical and geological evidence of 20. Reese, S. et al. 2011. Empirical building fragilities from ob- boulders deposited by tsunamis, southern Ryukyu Islands, Ja- served damage in the 2009 South Pacific tsunami. Earth-Sci- pan. Earth-Science Reviews, 102, 77-99. ence Reviews, 107, 156-173. 9. Goff, J. & Dominey-Howes, D. 2011. The 2009 South Pacific 21. Dall’Osso, D. et al. 2009a. A revised (PTVA) model for assess- tsunami. Earth-Science Reviews, 107, v-vii. ing the vulnerability of buildings to tsunami. Natural Hazards 10. Goff, J. 2011. Evidence of a previously unrecorded local tsu- and Earth System Sciences, 9, 1557-1565. nami, 13 April 2010, Cook Islands: Implications for Pacific 22. Dall’Osso, D. et al. 2009b. Assessment of the vulnerability of Island Countries. Natural Hazards and Earth Systems Science, buildings to damage from tsunami (in Sydney). Natural Haz- 11, 1371-1379. ards and Earth System Sciences, 9, 2015-2026. 11. Bird, D. & Dominey-Howes, D. 2008. Testing the use of a 23. Dominey-Howes, D. et al. 2010. Estimating probable maxi- ‘questionnaire survey instrument’ to investigate public percep- mum loss from a Cascadia tsunami. Natural Hazards, 53, 43- tions of tsunami hazard and risk in Sydney, Australia. Natural 61. Hazards, 45, 99-122. 24. Dominey-Howes, D, & Papathoma, M. 2007. Validating a 12. http://www.kantei.go.jp/saigai/pdf/201106061700jisin.pdf tsunami vulnerability assessment model (the “PTVA Model) 13. World Bank East Asia and Pacific economic update 2010, vol. using field data from the 2004 Indian Ocean tsunami. Natural I. Hazards, 40, 113-136. 14. Dominey-Howes, D. & Thaman, R. 2009. UNESCO-IOC 25. http://www.unohrlls.org/UserFiles/File/Pacific_Regional_ International Tsunami Survey Team Samoa. UNESCO-IOC Synthesis-MSI5-Final.pdf & Australian Tsunami Research Centre Misc. Report No. 2: 26. Port and Airport Research Institute (PARI) 2011. Executive 172pp. Summary of Urgent Field Survey of Earthquake and Tsunami 15. Thomas, C. & Burbidge, D. 2009. A Probabilistic Tsunami Disasters by the 2011 off the Pacific coast of Tohoku Earth- Hazard Assessment of the Southwest Pacific Nations. Geosci- quake. Unpublished report. 7pp. ence Australia Professional Opinion No. 2009/02. 27. Goff, J. & Dominey-Howes, D. in press. Hazardous Processes: 16. Goff, J. et al. 2011. Palaeotsunami precursors to the 2009 South 13.31 Tsunami. In: Geomorphology of Human Disturbances and Pacific tsunami in the Wallis and Futuna archipelago. Earth- Climate Change. Treatise of Geomorphology, Elsevier. Science Reviews, 107, 91-106. 28. Gero et al. 2011. Integrating community based disaster risk re- 17. Thomas, C. et al. 2007. A Preliminary study into the Tsunami duction and climate change adaptation: examples from the Pa- Hazard faced by Southwest Pacific Nations. Risk and Impact cific. Natural Hazards and Earth System Sciences, 11, 101-113. Analysis Group, Geoscience Australia. 29. Turner II, B.L., et al. 2003. Illustrating the coupled human- 18. Slovic, P. 1987. Perceptions of risk. Science, 236, 280-285. environment system for vulnerability analysis: three case stud- ies. Proceedings of the National Academy of Sciences 100, 19. http://web.worldbank.org/WBSITE/EXTERNAL/COUN- 8080-8085. TRIES/EASTASIAPACIFICEXT/EXTEAPREGTOPRU RDEV/0,,contentMDK:22739959~pagePK:34004173~piPK: 34003707~theSitePK:573964,00.html east asia and the Pacific region The World Bank 1818 H St. NW, Washington, D.C., 20433 http://www.worldbank.org/eap GFDRR is able to help developing countries reduce their vulnerability to natural disasters and adapt to climate change, thanks to the continued support of its partners: ACP Secretariat, Arab Academy, Australia, Bangladesh, Belgium, Brazil, Canada, Colombia, China, Denmark, Egypt, European Union, Finland, France, Germany, Haiti, India, Ireland, Italy, Japan, Luxembourg, Malawi, Mexico, The Netherlands, New Zealand, Norway, Portugal, Saudi Arabia, Senegal, Spain, South Africa, South Korea, Sweden, Switzerland, Turkey, United Kingdom, United States, Vietnam, Yemen, IFRC, UNDP, UN/International Strategy for Disaster Reduction and The World Bank.