Mega Prepared?

In any given year, Canada can be impacted by numerous natural disasters caused by severe thunderstorms, winter storms and tropical cyclones.

With the vast size of the country and its varied topography and geology, the risk of atmospheric perils differs by region and by season. Couple that with steady population growth, ongoing development and increasing concentrations of exposure value in at-risk regions, and the potential for insured loss from these perils is increasing.

Although sizable catastrophic events from atmospheric perils do occur – such as the Great North American Ice Storm of 1998, which caused insured losses in Canada of about $3 billion (adjusted to 2013 dollars) – many smaller events, of various sizes, intensities and duration can have a substantial impact on insured losses. Aggregating the risks from various perils can help risk managers prepare for the uncertainty of numerous events, at wide-ranging loss levels, occurring in different regions and in all 12 months of the year.

THE SCENARIO

Consider a mega-disaster year in Canada, which results in insured losses of $6.4 billion from severe thunderstorms, winter storms and tropical cyclones, a sum that falls at an exceedance probability (EP) near 1% (100-year return period). This modelled year encompasses 84 events in all:

• 74 severe thunderstorm events – hail, straight-line wind and/or tornadoes – impacting all provinces, except Prince Edward Island, with total insured losses in Canada of $5.9 billion;

• seven winter storms – wind, frozen precipitation (snow, ice pellets, freezing rain) and/or freezing temperature – affecting all 10 provinces, with insured losses of $217 million; and

• three tropical cyclones – high wind, with insured losses totalling $213 million, all in Quebec and Atlantic Canada.

The map below shows the loss severity throughout Canada of the three perils combined. The loss gradient is based on “forward sortation areas,” geographical regions determined by Canadian postal codes. [Click image below to enlarge]

Note that significant losses occurred in the west-central and central regions of the country, areas historically at high risk of severe thunderstorms, as well as in areas with high population density in Ontario and Atlantic Canada.

Insured losses from severe thunderstorms dominate the modelled year – reflecting historical natural disaster experience in Canada – although both winter storms and tropical cyclones contribute losses that could have noteworthy impact to regional portfolios.

AFFECTED EXPOSURE

The various sub-perils of severe thunderstorm, winter storm and tropical cyclone inflict damage by different mechanisms, adding a layer of complexity to the impact on exposures. Also, vulnerability to atmospheric perils in Canada differs region to region because of varying building codes and standards, varying building construction and varying climates and weather conditions (see map below). [Click image below to enlarge]

Wind is a risk common to all three atmospheric perils, although for the tropical cyclone peril, wind duration – much longer for a slower-moving tropical storm than for a severe thunderstorm or winter storm – can become critical. Construction type affects a building’s vulnerability to wind.

As well as physical damage from atmospheric perils, time element losses can be a factor. The 1998 ice storm, for example, crippled the electrical infrastructure in the region affected, resulting in power outages lasting from days to months, considerably impacting the ability of many businesses to operate.

In this simulated mega-disaster year, with $6.4 billion in estimated insured losses, losses occurred in every month of the year except March and December. May experienced the greatest losses by far – $4.5 billion, representing 71% of the total for the year – due to an off-the-chart windstorm that month, with $4.3 billion in losses (see graph below). [Click on image below to enlarge]

The modelling also reveals the season-specific characteristics of the various perils, with winter storms in the fall and winter months, severe thunderstorms (and related sub-perils) in warmer, perhaps humid, spring and summer months, and tropical cyclones in the late summer and fall months. (A winter storm with less than $1 million in insured losses, not visible on the chart, also occurred in October).

The 10 most costly events of the modelled year, shown in the table below, account for $5.9 billion, or more than 90%, of total losses in this simulated 1% EP year. [Click image below to enlarge]

The most devastating severe thunderstorm of the modelled year, the $4.3 billion event in May, occurred just 12 days after an $82 million storm struck Alberta and 18 days before $371 million in damage resulted from a storm in Alberta, Manitoba and Ontario. This underscores the real risk of atmospheric peril event clustering in Canada.

The costliest winter storm, with insured losses of $164 million, brought wind gusts in excess of 100 kilometres/hour in coastal regions and snowfall of almost a metre in major metro areas, spreading losses across six provinces from Ontario to Prince Edward Island. The costliest tropical cyclone in the simulated year left $179 million in losses in Nova Scotia, New Brunswick, PEI, Newfoundland and Labrador and Quebec.

IS THE INDUSTRY PREPARED?

The various severe thunderstorms, winter storm and tropical storm events in this mega-disaster loss scenario exemplify the extensive and widespread damage that can result from atmospheric perils in Canada, particularly the potential risk of loss from multiple smaller events. Importantly, with an annual EP of about 1%, the scenario year is not an extreme tail event; far greater losses are possible.

Responsible risk management includes acquiring a comprehensive and realistic view of risk, as well as preparing for a wide range of loss scenarios. Using model scenarios – such as a multi-peril 1% EP year – to probe a portfolio’s strengths and weaknesses will help a company respond effectively when disaster does strike. Being attentive to best practices can help ensure the most realistic loss estimates are achieved.

• Strive to obtain highly accurate exposure data. Atmospheric perils cause widely variable events that can have different impacts in different regions. Collecting detailed information for the properties in a portfolio – including location, primary building characteristics (such as construction type, occupancy, building age and height), and a true replacement value – will help refine the loss results and help get closer to the true vulnerability of individual buildings.

• Use visual intelligence to understand a company’s data. Geospatial analytics can provide rich and intuitive information on risk accumulations and how they relate to potential hazards, allowing for more-informed risk management.

• Anticipate business interruption losses. Use catastrophe models to calculate business interruption after natural disasters – a potential source of significant loss – as a function of downtime, damage, building size and architectural complexity.

• Be aware of non-modelled sources of potential loss. Damage attributable to inland flooding and coastal storm surge, extra clean-up costs and loss-adjustment expenses – which, for example, are not modelled in this scenario – can result in sizable insurance losses.

Although no model can predict what catastrophes will occur and when, it can be revealing and informative to simultaneously analyze multiple models across different perils, and across different regions. The careful analysis of model results can help risk managers prepare for myriad contingencies, ensuring scenarios like this simulated mega-disaster year will not be unexpected.