Uncovering Canada’s Catastrophic Risks

Compared to other countries, Canada has been fortunate to avoid large-scale economic destruction from natural disasters, such as events witnessed last year in Japan, New Zealand and Thailand. However, while the country is fairly well protected against such perils, depending on geographic location and other factors, there is still significant exposure risk. A thorough understanding of the risk landscape, integrated with catastrophe models, can help claims professionals better manage resources and response, even in advance of the event.

Of all disasters throughout Canadian history, natural perils have accounted for 70 percent, with flooding being by far the greatest – and most costly – source in the 20th century. To put this in perspective, in 2009 insurance payments for all events in Canada totaled $5.3 billion – of that amount flooding was responsible for $1.3 billion. The Red River, with its long history of flooding, caused extensive damage in 2009 in Manitoba due to ice jams and snowmelt, and additional precipitation on flat ground, while in the 2011, the Assiniboine River caused widespread damage in both Manitoba and Saskatchewan. To a lesser degree, but common nonetheless, are tornadoes. Surprisingly, Canada has the second largest amount of tornadoes (next to the U.S.), concentrated mostly in the southern regions. And, while rare, they can be severe, such as the F5 recorded in Elie, Manitoba in 2007.

But, perhaps the most regularly occurring natural perils in Canada are earthquakes, with approximately 4,000 recorded each year. Half of these occur in or off the shore of British Columbia. Over the past century, at least nine earthquakes in this area have registered a magnitude greater than 7 – notably, the (magnitude) M7.3 earthquake that hit British Columbia’s southern Vancouver Island in 1946, and the M8.1 Queen Charlotte earthquake – the largest in Canada’s history – which occurred off the coast of nearby Graham Island three years later. The presence of the active Cascadia subduction fault in an active boundary between the North American Plate and the subducting Juan de Fuca Plate in the Pacific Ocean makes western Canada, especially southwestern British Columbia, the most earthquake-prone region, especially susceptible to major, damaging events.

Since 2004, four earthquakes exceeding M8.5 have struck worldwide- two in Sumatra (M9.1 in 2004, M8.6 in 2005), one in Chile (M8.8) in 2010, and last year’s Great Tohoku (M9.0) in Japan. Seismologists have suggested that this pattern of megaquakes may actually occur in clusters. If this current cluster follows the one that occurred between 1950 and 1965, which saw six events ranging from M8.6 to M9.5, then it appears we may be only about halfway through, with the largest quake perhaps yet to happen. The Cascadia subduction zone is part of the “Circum-Pacific Ring of Fire” where these types of earthquakes occur most frequently on a global basis. While the Cascadia subduction zone converges only half as fast as tectonic plates in Chile or Japan, it does have a history of generating magnitude 8.5 or larger earthquakes, although not since 1700. And, the active Cascadia subduction zone is capable of producing the same size of earthquakes as in Indonesia and Japan.

While things tend to be quieter in southeastern Canada, they are not risk free. Because it is located within a stable continental region of the North American Plate, the rate of earthquake activity has been relatively low. Yet, the slow movement of this plate combined with other forces acting upon it creates sufficient stresses to have caused earthquakes of M6.0 over the decades in the St. Lawrence and Ottawa valleys. In fact, seismic sources zones along the St. Lawrence River consist of individual clusters of historical concentrations, such as the Western Quebec and Charlevoix seismic zones, which have consistently generated some of the largest earthquakes in southeastern Canada. However, earthquakes of this region seldom rupture faults all the way to the ground surface lending significant uncertainty to their causative structures.

The Canadian Risk and Hazards Network has determined that a significant earthquake is Canada’s greatest potential natural disaster. Fortunately, past earthquakes have not resulted in extensive property damage, primarily because a large event has yet to strike a heavily populated, and, therefore, highly insured area with significant exposures. But if an earthquake were to strike near one of Canada’s major urban areas, even a M6.0 would be devastating. Vancouver and Victoria would be most at risk: Estimated insured losses could approach $100 billion, with Vancouver sustaining about 20 percent of losses. Total economic damage could exceed $400 billion in that region alone, but Montreal, Ottawa, Toronto and Quebec are also vulnerable.

Because Canada experiences damaging earthquakes only every few generations, a much longer time scale than say, in California, there is little or no social memory of these devastating events. As a result, there is considerable resistance to spending the time and money to adopt earthquake-resistant designs, update code standards, retrofit building stock (especially in the older cities), or implement other preventive measures. Would high-rises, even if they are code compliant, be able to meet expectations for usability and repair/rebuilding costs? In the older cities such as Vancouver, would the numerous masonry buildings survive and the ongoing efforts to move hazardous overhead electric transmission systems underground be completed in time?

Practising Model Behaviour

Of course, it is not yet possible to know when and where a major earthquake will occur, but history tells us it is inevitable. Insurance professionals can turn awareness of this into action by embracing scenario planning and using catastrophe models. Re/insurers appreciate the value of a model’s simulation sets, built from past events, to help them establish rational expectations about risk, but claims professionals can and should also be taking advantage of the technology both before and after disaster strikes to reduce surprises and better allocate resources. So, if there’s a major earthquake in a low seismic region like Quebec, even in the absence of recent history, claims managers can look at a model’s hazard maps to determine such details as how hard the ground might shake and how the soil would behave, and how structures of a specific vintage and construction are likely to react to these elements.

This data is incorporated into models in a different manner than would be in seismic code requirements. Using models in conjunction with building code products can reduce the potential for bias judgment that comes from relying too heavily on building codes alone. For example, the methodology for incorporating soil amplification of earthquake motions into a base map of ground motions is known to have a bias to the high side, which meets code standards, but is not a true estimation. Catastrophe modelers change the reference conditions for soil to amplify or de-amplify the earthquake motions, thereby reducing the degree of bias.

Information derived from a model can also help claims managers pre-negotiate with suppliers, plan alternative transportation routes and provide backup to areas that could be disrupted. When an earthquake or other disaster does strike, they can compare the hazards sets against their portfolio and plan staffing resources accordingly, such as the number and experience level of adjusters to deploy to the site. With time a critical component to claims processing, a model can make getting on the ground faster and more efficient.

Claims managers have a choice. Canada’s current state of seismic inactivity can make them either vulnerable or prepared. As General Douglas MacArthur said, “There is no security on this earth; there is only opportunity.”

Paul Thenhaus is senior geologist of model development for catastrophe modeling firm EQECAT, Inc. (Tim Trainor, client relationship direc
tor, assisted with this article.)