Insuring against natural catastrophes is no simple feat. Aside from the massive scale of damage brought by each single event, multiple risks are present, all of which must be taken into account by re/insurers’ models.
James Cosgrove (pictured), meteorologist and senior analyst for event response at catastrophe risk modelling firm RMS, explained how weather disturbances, such as Hurricane Florence in the US, bring multiple risks and how advances in data gathering and modelling helps re/insurers improve coverage.
RMS estimated the insured losses for Hurricane Florence, which made landfall on September 14 in North Carolina, at between US$2.8 billion and US$5 billion, factoring in post-event loss amplification. The figure includes losses to the National Flood Insurance Program (NFIP), which have been estimated at between US$800 million and US$1.2 billion. Meanwhile, overall economic losses were estimated to fall between US$6 billion and US$11 billion.
“Like [Hurricane] Harvey, Florence demonstrated how wind, storm surge, and inland flooding can all be generated by a single event, with the largest contribution to total loss in both instances coming from inland flooding (both coastal storm surge and inland flooding),” Cosgrove said.
To simulate the precipitation, run-off, and pluvial and fluvial flows through the southeast US and Virginia, RMS used its US Inland Flood HD Model.
According to Cosgrove, this “allows for robust modeling of inland flood risk from all sources, including tropical cyclones. By combining results from the North American Hurricane Model and the US Inland Flood HD Model, our clients can model the risk from tropical cyclones comprehensively, including the effects of wind, coastal surge, and inland flooding.”
These models are a result of monitoring weather conditions both on the ground and in RMS’s facilities in London and Florida.
“After a landfalling hurricane or other natural disaster, RMS routinely sends scientists and engineers into the affected region to survey the damage and collect ground truth,” Cosgrove said. “This field reconnaissance in the immediate aftermath of an event serves several purposes: refining our understanding of the event, validating hazard observations, evaluating damage, and providing insight into the prolonged impacts associated with affected communities. This increased understanding helps improve and enhance our modeling solutions, which subsequently supports future loss estimates.”
Interestingly, Hurricane Florence made landfall in the US at roughly the same time Typhoon Mangkhut hit the Philippines. However, the storms were vastly different. Florence was slow and plodding, with Category 1 winds of 90mph, but accompanied by torrential rain, dumping a record-high of over 30 inches of rain on North Carolina.
Meanwhile, Mangkhut moved fast, whipping its path with Category 5 (the highest level) winds of 165mph.
“Our event response teams in London and Florida are constantly monitoring global conditions and are able to respond to multiple/simultaneous events,” Cosgrove said. “Our automated data collection systems are continuously ingesting hazard data in real-time, including observations which can be used to refine and validate hazard estimates (e.g. wind speeds) from our modeling process as an event unfolds.
“Our teams have been through highly active hurricane seasons, such as in 2004, 2005, and more recently last year in 2017, when multiple landfalling hurricanes impacted different parts of the US within a matter of months or weeks.”