Slide1: CATASTROPHE INSURANCE IN THE PACIFIC Managing Disaster Risks in the Pacific Islands Region Summary Report SOPAC Miscellaneous Report 550 A SOPAC study in partnership with the World Bank and AusAID SUGGESTED READING Biukoto L., Swamy M., Shorten G.G., Schmall S., and Teakle G. 2001. Pacific Cities CD, Port Vila. GIS Hazards Dataset, Version 1.1. SOPAC Data Release Report 4. Blong R. 1992. Natural Perils in the Port Vila Area, Vanuatu. Macquarie Park Research Limited Report 4. Chung M., and Hill D. 2002. Urban Informal Settlements in Vanuatu: Challenge for equitable development. Report prepared for the Pacific Islands Forum Secretariat and UN ESCAP Pacific Operation Centre. Commonwealth Secretariat 2000. Small States: Meeting Challenges in the Global Economy. Report of the Commonwealth Secretariat/World Bank Joint Task Force on Small States. Garaebiti E., Shorten G.G., Regnier M., Naidu P., and Swamy M. 2002. Assessment and study of the Port Vila Earthquake, 2nd January 2002. SOPAC Joint Contribution Report 142. Hofstetter A., Shapira A., Bulehite K., Jones T., Mafi K., Malitzky A., Papabatu A., Prasad G., Regnier M., Shorten G., Singh A., Stephen M., and Vuetibau L. 2000. Frequency-magnitude relationships for seismic areas around the capital cities of Solomon, Vanuatu, Tonga and Fiji Islands. Journal of Seismology 4. 285-296. Howorth R. 1985. Baseline coastal studies, Port Vila, Vanuatu. Holocene uplift record and evidence for recurrence of large earthquakes. CCOP/SOPAC Technical Report 51. Kumar P. 1987. Tropical Cyclone Report – Tropical Cyclone Uma. Fiji Meteorological Report 87/8. Linnerooth-Bayer, J., Mace, M.J., and Verheyen, R. 2003. Insurance-related actions and risk assessment in the context of the UNFCCC. Background Paper for UNFCCC workshops. Unpublished. Pollner J.D. 2001. Catastrophe Risk Management: Using alternative risk financing and insurance pooling mechanisms. Report of the Finance, Private Sector & Infrastructure Unit, Caribbean Country Department, Latin America and the Caribbean Region, World Bank. Prevot R., and Chatelain J. L. 1984. Seismicity and earthquake risk in Vanuatu (Text), and 1983. Seismicity and seismic hazard in Vanuatu (Figures). ORSTOM Report 5-83. Regnier M., Morris S., Shapira A., Malitsky A., and Shorten G. 2000. Microzonation of the expected seismic site effects across Port Vila, Vanuatu. Journal of Earthquake Engineering 4 (2). 215-231. Schmall S. 2003. Participatory vulnerability assessment and community mitigation planning: Development of a participatory approach for risk mitigation in peri-urban settlements of Port Vila, Vanuatu. Unpublished report to SOPAC. Shorten G.G. 2001. Seismic risk in Pacific cities: Implications for planning, building code legislation, and urban search and rescue services. Proceedings, Australian Earthquake Engineering Society Conference: Earthquakes in the Real World, Canberra, ACT, 7-9th November, 2001. AEES, Melbourne. 2.1-2.9. Shorten G.G. 2002. Pacific Cities urban planning and risk management: Port Vila and peri-urban areas case study. Proceedings, Sustainable Urban Services: PECC Sustainable Cities Noumea Seminar, Noumea, New Caledonia, 4-5th November, 2002. PCEE, pp.107-116. Shorten G.G. 2003. Challenges for the application of earthquake engineering in the Pacific Islands. Proceedings, PCEE 2003: 7th Pacific Conference on Earthquake Engineering, University of Canterbury, Christchurch, NZ, 13-15th February, 2003. PCEE Paper 150. Shorten G.G., Goosby S., Granger K., Lindsay K., Naidu P., Oliver S., Stewart K., Titov V., and Walker G. 2003. Catastrophe insurance pilot study, Port Vila, Vanuatu: Developing risk-management options for disasters in the Pacific region. SOPAC Joint Contribution Report 147. Shorten G.G., and Schmall S. 2003. Disaster risk management in marginal communities of Port Vila, Vanuatu: Project Summary. SOPAC Miscellaneous Report 507. Shorten G.G., Schmall S., Granger K., and Naidu P. 2003. Blending custom knowledge and science to reduce risk in three settlements near Port Vila, Vanuatu. Proceedings, 2003 Australian Disaster Conference, Canberra, 10-12th September, 2003. Shorten G., Shapira A., Regnier M., Teakle G., Biukoto L., Swamy M., and Vuetibau L. (Compilers) 2001. Site-specific earthquake hazard determinations in capital cities in the South Pacific. Second Edition. SOPAC Technical Report 300. SOPAC 2000. SOPAC Sustainable Development Strategy. SOPAC Secretariat, Suva. Vermeiren J. 2000. Risk transfer and finance experience in the Caribbean. World Bank's Disaster Management Series No.2: Managing Disaster Risk in Emerging Economies, A. Kreimer and M. Arnold, editors. Walker G.R. 2003. Engineering disaster insurance schemes for the 21st Century. Proceedings, PCEE 2003: 7th Pacific Conference on Earthquake Engineering, University of Canterbury, Christchurch, NZ, 13-15th February, 2003. PCEE Paper 32. World Bank 2000. Pacific Regional Strategy. Papua New Guinea and Pacific Islands Country Unit, East Asia and Pacific Regional Office. Report No. 20370-EAP. 16


Slide5: The uplifted coral terraces around Port Vila are clear indicators of a long history of large prehistoric earthquakes. Taken together with instrumental recordings of seismic events, local microzoning studies, and the lessons of the January 2002 Port Vila earthquake, the information can be used to estimate probable future damage from earthquakes. The worst-case earthquake might cause losses of USD$215 M. Modelling of the resulting worst-case tsunami hazard suggests a scenario 6-7 m wave impacting more than 200 buildings in the central business district of Port Vila to cause USD$26 M damage, as well as totally destroying some 1,500 buildings in peri-urban coastal settlements on Mele Bay. Earthquakes Larger magnitude earthquakes are likely over longer return periods. The maximum anticipated earthquake for the Vanuatu archipelago is magnitude Mw 8.1 on the Richter Scale. Shocks of this size occur in the wider Vanuatu region - usually along the New Hebrides Trench to the west of the Group - with a return period of less than 100 years. Closer to home, the uplifted coral terraces on the western side of Mele Bay indicate that there has been a Richter magnitude Mw 7.5 to 8 event in the near vicinity of Port Vila every 1,000 years on average, uplifting the area by 1-2 m each time. The 1,000-year event has about a 5% probability of occurring within a 50-year time-frame. Intensity, a measure of the actual physical effects and damage of an earthquake at a particular place, is measured on the Modified Mercalli (MM) Intensity Scale. The Richter magnitudes of historical earthquakes near Efate can be compared with the past record of intensities actually felt in Port Vila. From this, it is possible to predict that the city can expect to experience average earthquake intensities up to MM 9-10 in the long term. At these intensities there would be general destruction of masonry structures and even some well-constructed wooden buildings. Lessons from the 2002 earthquake show that buildings on thick alluvial deposits, reclaimed land, or high ridges and plateaus will experience even higher intensities than on sites with rock foundations. The Richter magnitude Mw 7.1 earthquake that hit Port Vila on 2nd January 2002 was the largest here since 1927. It caused local intensities averaging MM 7-8, and as high as MM 8-9 in places, and resulted in around USD$10 M in losses to the city. Damage included widespread minor structural damage and the destruction of a 3-storey secondary school, a multi-storey Government building, and a vital highway bridge, as well as blocking the port access road by major landslides for several weeks. The loss curve predicts that the 100-year earthquake would cause some USD$76 M damage to buildings in Port Vila and its peri-urban settlements, while infrastructure losses would add another USD$21 M, for a total of USD$98 M losses. The 500-year event (10% probability in 50 years) might result in damage of USD$215 M. On this scale, it’s obvious that the 2002 Port Vila earthquake - causing losses of only 1.3% of all assets - was a relatively small magnitude event that can be expected to recur frequently. Tsunamis Fifteen minutes after the 2002 earthquake, Port Vila was struck by a tsunami of significant size. The observed effect at the shoreline varied across the harbour, reaching a maximum peak-trough height of 3 m. The tsunami fortunately coincided with a spring low tide, narrowly averting the flooding of the central business district with over 1 m of water. A tsunami scenario for Port Vila was generated from a hypothetical Richter magnitude Mw 8.1 earthquake located 50 km west of the capital, in the same location as the January 2002 earthquake, and equivalent in magnitude to the largest earthquake recorded in the region to date. The return period of such an event is assumed to be an order of magnitude longer than that of the Mw 8.1 earthquake because not all large earthquakes necessarily generate tsunamis. The point of origin of the tsunami, the actual causative mechanism, and the travel path of the wave are all critical variables that introduce uncertainty. In the worst-case scenario, it’s anticipated that 212 buildings would be inundated around Port Vila Harbour as well as several entire peri-urban villages and settlements at the head of Mele Bay. The level of flooding has been defined for each building within the inundation zone. The maximum tsunami run-up in Port Vila Harbour would extend up to 200 m beyond the shoreline. The wave height is predicted to rise up to 6 m above mean sea level. Altogether, some 28,000 m2 of floor space in the central business district would be inundated, causing anticipated losses of more than USD$24 M, including losses to infrastructure. More than 1,500 buildings in the peri-urban area fringing Mele Bay are also at risk from the worst-case tsunami. The predicted maximum tsunami run-up extends up to 2,000 m beyond the Mele Bay shoreline. The wave height here would rise up to 7 m above mean sea level. Replacement values are not necessarily in line with the averages for the city area and losses are difficult to calculate but could reach USD$50-65 M in real terms. 9 The January 2002 earthquake caused only about 1.3% loss (USD$10 M) to Port Vila assets, although much higher intensities, and therefore larger losses should be expected in the longer term.


Slide6: In order to develop estimates of future damage, computer-modelling of cyclonic winds, based on historic records of past cyclones back to 1950, were calibrated against the known effects of Cyclone Uma which caused around USD$65 M damage to Port Vila in 1987 – probably the largest single recorded disaster in Vanuatu to date. Storm surge and storm wave modelling, based on Cyclone Beni in 2003, indicates that the potential damage from these secondary hazards is more limited. The predicted worst-case cyclone, together with associated surge and wave events, might cause losses to the city approaching USD$415 M. Cyclones A statistical analysis of tropical cyclones in the vicinity of Efate was used as the basis to generate random events in a cyclone modelling system. The model was run to produce the equivalent of 5,000 years of storms. A topography-wind zone grid and model were set up over the city area and a damage-loss analysis was performed, using wind speed-damage vulnerability models developed specifically for building classes identified in Port Vila. The loss analysis was calibrated using known Cyclone Uma losses. By applying the wind and vulnerability models over a large number of cyclone events, the results could be integrated to estimate the levels of damage expected for Port Vila in the future and the relative frequency of occurrence of those damage levels. The model was re-run to measure the potential benefit of retrofitting and upgrading of Class B and Class C buildings. The total losses expected to buildings and infrastructure combined might be more than USD$170 M in the 100-year event cyclone. The 100-year event has about a 50% chance of occurring in a 70-year lifetime. Damage levels were computed for each of the 42 topographic wind zones imposed over Port Vila. Damage levels were calculated for each building class for recurrence intervals of 50, 100, 450 and 1,000 years. The relative differences between the cells largely reflect the terrain differences between locations. The model was also used to estimate the frequency (recurrence interval) of levels of damage, expressed as percentage of buildings damaged to a particular level. For example, 10% of Class A buildings would be impacted to a level of 50% damage once every 455 years but Class C buildings would experience the same level of damage more frequently - every 88 years - underlining the importance of raising building standards. Catastrophe Insurance is one of a number of options that can be drawn upon to manage risk in the Pacific. Disasters are generally measured by both their social and environmental impact and their economic effects. The real tragedy of disasters is the subsequent individual hardship endured by those affected, and this hardship generally has a strong financial component. Mitigation of the financial aspect of hardship is usually the primary objective of national disaster insurance schemes. The lessons of Cyclone Uma in 1987, and more recently the January 2002 earthquake, provide a glimpse of a possibly devastating outcome for Vanuatu in the future. Uma had destructive effects on agricultural crops, while in Port Vila the Government infrastructure and tourism sectors suffered major physical and business interruption loss. While it can be assumed that insurance will make a major contribution there are uninsured losses to be taken into account. For some 12 months after Uma, tourist arrivals were down, while activity in other sectors such as construction and some professional areas increased significantly. Uma had a severe impact on the Government’s finances in the form of lower revenue and higher expenditure. Real economic growth in the 1987/88 period stagnated. However, in terms of GDP, the economy in Vanuatu did not show a decline but, on the contrary, grew moderately. Government relief, donations of aid, charities, and insurance are the main mechanisms by which hardship can be reduced. Each of these spreads the financial losses beyond those directly affected, so reducing the individual impact. Insurance is a proactive form of mitigation whereby an individual or organisation is able to transfer for a fee all, or part, of the financial risk of loss in advance of a disaster. If a disaster strikes and loss is suffered then the relevant financial compensation is made. Because of the international transfer of risk that occurs internally within the insurance industry through the reinsurance process, the losses can be widely spread around the world. As a result, the national economy will be subjected to a lesser financial shock. Before launching into a regional scheme, a conceptual framework should be developed. The existing level of household and business private insurance must be ascertained for each country involved. If infrastructure and utilities have been privatised and are insured to a level commensurate with the needs of the nation, then these would not need to be covered further. Important considerations include setting limits to premiums (and thus coverage), the necessity for developing a pool or fund, and the form of management and administration. Because premiums are likely to be determined by issues of affordability, the design objective for a scheme will essentially be to determine a set of policy conditions that satisfies the limitations on premiums. SOPAC is working in many areas to demonstrate a comprehensive approach to risk management within the Pacific Island Countries. This approach includes practical steps to reduce vulnerability through the Pacific Cities project, upgrading of building standards for urban preparedness and response initiatives, and through advocacy to carry the issues to the highest levels of government. Operating in conjunction with civil society organisations in participatory community projects in peri-urban Port Vila, SOPAC has demonstrated that concrete risk-treatment solutions are achievable. Complete destruction of a house by cyclonic winds on a ridge overlooking Port Vila Harbour.