Jeff Waters (pictured) is senior managerand meteorologist at catastrophe modeler Risk ManagementSolutions.

Superstorm Sandy is the second costliest landfalling U.S. stormon record, and ranks as one of the most destructive events in U.S.history. The storm, which was almost the size of Alaska, broughthurricane-force winds, rain, and catastrophic storm surge to theMid-Atlantic and Northeast, in addition to significant snowfall toparts of Appalachia and the Midwest. Thousands of flights werecancelled, and power outages that were reported in 15 statesimpacted almost 8.5 million homeowners and businesses, and causedsevere business interruption and contingent businessinterruption.

Sandy, like Hurricane Katrina and Hurricane Ike before it, haveemphasized the potential risk that coastal flooding poses to theinsurance industry. It drives the majority of U.S. flood tail risk,which impacts key risk management and rate filing decisions, fromcapital management to risk transfer. These storms have alsohighlighted that surge-driven damage can drive a large proportionof the total insured losses – sometimes exceeding the wind damagelosses. Sandy's greatest damage impact was from coastal flooding,which accounted for 65 percent of Sandy's total insured losses.

Using its U.S. North Atlantic Hurricane Model, RMS estimatesthat each year there is an 18 percent chance that a landfallingU.S. hurricane will cause coastal flood losses that exceed the windlosses – this probability nearly doubles along the Northeast coast.The RMS model also indicates that there is a 32 percent chance thatannual U.S. insured coastal flood losses will exceed $1 billion,and a 16 percent chance that they will exceed $5 billion.

One of the biggest sources of loss uncertainty for Sandycontinues to be how business interruption claims are ultimatelysettled. Although 93 percent of claims filed by businessesand homeowners were closed within six months of the storm, theremaining claims are expected to take more than a year to fullyresolve. Unlike windstorm claims, business interruptioncosts can escalate independent of structural damage, because thedamage caused by the floodwaters prevent the occupants fromreturning to their buildings.

RMS conducted research during Sandy that revealed that floodlosses from commercial and industrial structures might only bepartially covered. This is due either to incomplete take-up ofinsurance, i.e. less than 100 percent of locations having floodcoverage, or because of the common use of flood sub-limits andseparate attachment points. RMS research also revealed that claimsseverity is different in areas where wind and coastal floodinteract.

Modeling Storm Surge

Traditional approaches that attempt to predict surge magnitudebased only on the storms' landfalling intensity have been known tounderestimate the impacts of coastal flooding from surge. Sandyhighlighted the limitations associated with lower-resolution dataand aggregate modeling at ZIP-code level. Using coarser resolutioninput data to model a high-resolution hazard, like coastalflooding, yields less accurate results, with increased uncertaintysurrounding loss estimates.

This is because localized surge impacts are not captured asaccurately. As a result, coastal flood hazard and loss estimatesoften become more simplified and volatile as they are aggregated torepresent larger areas. The street-by-street pattern of flooding inManhattan and along the New Jersey coast – two areas where floodingclosely follows small-scale changes in topography – highlighted thevalue of data intensive modeling to enable accurate underwritingand assess flood loss.

The scale of coastal flood damage caused by Sandy continues toemphasize the need for catastrophe models that have physicalcoastal flood modeling capabilities that are linked to fulllife-cycle storm simulations. This type of advanced storm surgemodeling that uses high-resolution geospatial data to simulate theimpacts of wind-driven water in a hurricane system on coastalregions is necessary to be to enable (re)insurers develop a moreaccurate view of the storm surge risk and make sounder underwritingdecisions.

Regardless of when or where the next major coastal flood eventoccurs the risks associated with coastal flood have been heightenedby Sandy, Katrina, and Ike. As such, the industry needs to have theright tools available in order to model the magnitude and severityof catastrophic storm surge accurately. But these models need to betransparent and openly reflect where material uncertainties remain,allowing customizable views of surge risk that enable the industryto explore the risk. Only through models like this, combined withhigh-resolution data, will the industry's insight into coastalflood risk improve and ensure that the next Sandy is not a surprisewhen it hits.