Typhoon Risk in Southeast Asia (14 Research Papers Included)

Background

Globally tropical cyclones are the most damaging weather events (Figure 1). Southeast Asia is a region that is prone to typhoons (tropical cyclones). The geographical location and warm waters of the Pacific Ocean make it a favourable environment for typhoon formation. Countries in Southeast Asia, such as the Philippines, Vietnam, Thailand, and Myanmar, are particularly vulnerable to these storms. The typhoon risk in Southeast Asia varies throughout the year. The official typhoon season in the region typically runs from May to November, with the peak activity occurring between July and October. During this period, the warm ocean waters and favourable atmospheric conditions contribute to the formation and intensification of typhoons. The Philippines, located in the western Pacific, is often directly affected by typhoons. The region experiences an average of 20 storms per year, with some causing significant damage and loss of life. Vietnam and parts of mainland Southeast Asia also face a substantial risk of typhoons, although they tend to be slightly less frequent and intense compared to the Philippines. Climate change is expected to increase the frequency of very powerful storms as the oceans warm. The role of El Nino is complex and regionally dependent.

Figure 1. Cumulative global economic losses by hazard since 2000. Tropical cyclones dominate. Aon Climate and Catastrophe Insight.

The financial costs associated with major typhoons in Southeast Asia can be substantial and very varied. It is challenging to provide exact figures due to variations in reporting and different sources of data. The estimated economic damage caused by, for example, Typhoon Haiyan in the Philippines was around USD12 billion. The storm affected multiple sectors, including infrastructure, agriculture, housing, and businesses. It is important to note that these figures represent estimates and the actual economic costs can vary. Additionally, the financial impact of a typhoon extends beyond immediate damages and includes long-term recovery and rebuilding efforts. The costs also encompass both direct damages and indirect losses, such as disruptions to businesses and impacts on the overall economy.

A catastrophe model is a mathematical tool used to estimate and assess the potential impact of hazards like typhoons on a given area. It combines various data sources, including historical storm data, meteorological data, topographical information, and exposure data (such as population density and property values), to simulate and predict the potential losses associated with a typhoon event. Catastrophe models use sophisticated algorithms and modeling techniques to simulate thousands or even millions of hypothetical typhoon scenarios. By considering factors such as storm intensity, track, and characteristics, these models take into account the vulnerability of different types of structures to estimate the potential damage. They can also incorporate other variables like the response capabilities of infrastructure systems and the effectiveness of preparedness measures. Ultimately, the models generate a range of potential outcomes, including estimates of property damage, casualties, and economic losses.

These models help insurance companies, reinsurers, government agencies, and other stakeholders assess and manage the financial risks associated with typhoons. By evaluating the potential losses under different scenarios, decision-makers can make informed decisions regarding risk mitigation, insurance pricing, emergency planning, and resource allocation. By using a typhoon catastrophe model, stakeholders can gain insights into the potential impact of typhoons on portfolios, regions, or specific locations. This information can help in developing risk management strategies, designing insurance policies, and assessing the financial resilience of an area in the face of typhoon-related risks.

IRIS: Imperial College Storm Model

Imperial College has built a typhoon catastrophe model for the Singapore Green Finance Centre (SGFC). The first step is to generate track and wind speeds of multiple events. Imperial College has published the basic version of the model as open source (Sparks and Toumi, 2024) as well as 13 other peer-reviewed papers (see below) that examine the physics of the problem and this has been used to build the new model. There are a number of innovations which make the model unique. Unlike other models used in the sector, we only model the decay from the lifetime maximum. This makes the model more accurate and also crucially easier to constrain for climate change and seasonal predictions. It has a lower bias than models currently used in the market.

Figure 2. shows a map of a sample of tracks with their wind speed. It shows the dominant role that the West Pacific has in the global role of tropical cyclones.

Figure 2. Map of the IRIS tropical cyclone tracks with wind speed (m/s).

The model can be used to assess the probability of tropical cyclone landfall in Asia and also at any point in the world. The probability is expressed as the return period which is the expected average time between events. IRIS data are already being used to inform Southeast Asia’s storm surge risk for a separate SGFC project. The model is configured for multiple applications including current base and future climate change risk.  Climate change risk users would need to specify their assumed warming as IRIS does not make climate projections on its own. This model is easily configured to bespoke climate scenarios.

Data Access

The basic model and sample data are open source (Sparks and Toumi, 2024). Converting these landfall wind speed probabilities into risk and financial loss requires the use of damage functions and exposure data. To achieve this, we have partnered with the OS-Climate initiative. This initiative is led by SGFC partners BNP Paribas and Goldman Sachs to enable open source tools and data for climate risk analytics. IRIS data are now freely available on their platform (under wind hazard, IRIS).  IRIS data will allow SGFC users to directly assess the current and future tropical cyclone risk to any specified exposure or critical infrastructure.

Worldwide Weather Attribution (WWA) and Global Media Coverage

The attribution science of extreme events to climate change is a rapidly emerging discipline which acts as a substantial policy driver. The WWA is the world leader in this field. Attributing tropical cyclone wind speeds has been challenging and not routinely done. However, IRIS has demonstrated a unique capability to achieve this, making it the model of choice for WWA. In 2024, IRIS was deployed in a number of high-profile cases across Asia and the USA, attracting extensive global media coverage. This included articles in The Straits Times on Hurricane Helene, TC Mayotte, Typhoon Shanshan, Typhoon Gaemi, and multiple landfalls in the Philippines, as well as coverage of Hurricane Milton in the Financial Times and The New York Times.

Testimonial

The IRIS Tropical Cyclone (TC) model is being used by the RISK ESG team in the Bank for the assessment of physical climate risk. Our analyses require a model that can provide synthetic TC events for current and future climate, for different years in the future. The IRIS team’s approach to perturbing tracks is well-suited to our use case, with an approach to climate-conditioning, through Potential Intensity, that avoids the effects of biases in CMIP6 data sets; and of course the model is demonstrably successful in simulating synthetic tracks. The IRIS data set is used both in event-based analyses and in the form of multi-return-period hazards maps; the validation that the return periods of IRIS have been through is therefore important for us. In both cases, the data is combined with models of wind vulnerability to assess how economic losses, and the associated financial risks, may increase under future climate scenarios. The model has been used as part of stress testing exercises, especially those focussed on the Western Pacific, and is also integrated into tooling intended to make physical climate tools available across the Bank. The model has proved to be extremely useful and we want to thank for authors for developing the model and making the output readily available.
Thanks again: we certainly very much appreciate your work and are trying to make good use of it!”

Applications

• Climate change scenarios stress testing
• ILS and catastrophe bonds
• Loss and Damage Fund
• Climate change litigation
• Climate change attribution