NEW YORK (BestWire) — Developers of catastrophe models will spend years digging through the data cache left behind by Hurricane Sandy that one modeller called a “treasure trove” of information, which will help address challenges in modelling Northeast storms by adding layers of validation.The storm data fill in what has historically been a sparse area of experience, said Ryan Ogaard, a senior vice president of product management at RMS, noting the area Sandy struck is where the industry has the least amount of information and “the most uncertainty in our models.”“The most information we have is in Florida. That stands alone,” Mr Ogaard said.“If you look at maps of the storm footprints we have with detailed information, Florida is fully coloured in with these storm footprints. Then you have a dense population of storms that we have data for in the Gulf.“As you go up the Atlantic Coast it gets more and more sparse, until you get to the north of New Jersey there are only two or three storms that we have reasonable data from. Nothing of this scale. A big blank spot suddenly has a storm footprint.”Sandy made landfall on the evening of October 29, just southwest of Atlantic City, NJ, as a post-tropical cyclone. Total insured damages from the storm are expected by RMS to be up to $25 billion, which is one of the largest industry estimates.The storm was unusual for a number of reasons, all of which will be dissected by modellers in the coming months and years, said Tim Doggett, principal scientist at AIR Worldwide.He said Sandy isn’t likely to cause major ripples in the model.“Sandy was a very large storm. It had very large wind radii and we see that actually pushed the envelope of what we’ve had in our historical experience,” Mr Doggett said.“So that will be a lesson learned for us, and we’ll probably look to see if we can make sure those larger storm radii are properly captured in our event sets.”Other points that will be analysed include Sandy’s low central pressure that resulted in onshore winds weaker than their potential; Sandy’s track from east to west; and the girth of the storm, which caused surge to extend itself over several tidal cycles. ““That’s an important thing to recognise and make sure the model is capturing correctly. All of the little intricate details of the event are things that we go back and validate again,” Mr Doggett said.Sandy also highlighted some storm features that models don’t consider, like indirect losses due to power failure, transportation disruption and forced evacuations, said Jeffrey Tennis, manager of catastrophe analytics for Lockton. He said models are helpful in quantifying direct losses, but care needs to be used when viewing cat model output.“At this point, the industry is so sceptical of catastrophe models, few are surprised the models failed to get it right,” Mr Tennis said.“At the same time, the industry has greater awareness that models do not consider every possible source of loss. [Hurricane] Katrina highlighted the importance of levee failure.“[Hurricane] Ike demonstrated that hurricanes can cause meaningful damage further inland than expected. In this case, Sandy exposed how events can cause damage to an insured, even if the damage didn’t occur at the insured property.”