Irma: Using Cat in a Box Tools to Assess Scenarios
Sam GibsonSeptember 06, 2017
22:30 UTC Wednesday, September 6
Sam Gibson, director – Consulting, RMS
With Irma moving swiftly through the Caribbean as a Category 5 hurricane, currently producing sustained winds around 185 miles per hour (297 kilometres per hour), concern is building as to the potential impacts of another major hurricane landfalling in the U.S. Although Irma’s actual path as it approaches Florida is very uncertain, there are analytical techniques available to help gain insight into the range of potential damage.
The RMS Cat in a Box (CIAB) application is designed for the assessment of parametric contracts, and can evaluate the probability of storms intersecting a specific geographic region whilst having certain severity characteristics. The application will output the list of events which intersect these areas and produce the associated Exceedance Probability (EP) curve.
CIAB analytics can help to identify potential event tracks using hypothetical gates. These could, for example, be based on the projections from the National Hurricane Center (NHC) where their forecast provides a track uncertainty cone. Better yet, the RMS HWind forecast prototype analysis of Hurricane Irma can be used to visualize the forecast uncertainty of an ongoing tropical cyclone.
This forecast product is a clustering analysis of all the available ensemble global forecast model outputs (e.g. ECMWF and GFS) and it is used to determine the probability of the storm track passing through a location. Shown in Figure 1, the shaded field shows the probability of the center of Hurricane Irma being within 50 nautical miles of a certain position based on a large array of available forecast models. To complement the gridded cloud of probabilities, selected probabilities at selected population centers are extracted in a table at the top of the plot. Finally, overlaid on top of the shaded probabilities, five specific track scenarios for Irma are presented that are representative of the underlying models. These five scenarios have their own probabilities and each scenario provides important timing information across the probability cloud so clients can have an estimate of when hazards may occur.
By using these ensemble models, we are able to better understand the uncertainty of where the storm might track based on current atmospheric conditions up to five days in advance.
The track probability field can then be used to inform where boxes should be drawn in the CIAB application to capture the range of forecast outcomes. The common events intersecting all boxes are chosen, and the resulting set of events can represent a view of potential Irma scenarios, dependent on the confidence in the forecasts.
Using shapes based on this forecast (shown in Figure 2, above), we calculated the conditional Exceedance Probability curve below in Figure 3.
The conditional Exceedance Probability curve in Figure 3 above, shows that there is approximately a 17 percent chance of exceeding US$100 billion given:
a storm passes through all of the depicted boxes, and
the storm has the intensities that Irma is expected to maintain (I.e. Cat 5 through the first box, dropping to Cat 4 or above through the second box and Cat 3 or above through the third box as it moves eastwards)
Note that this possible US$100 billion loss is NOT an industry loss estimate for Irma, nor is it an appropriately calibrated RMS track selection for Irma — it is simply the conditional exceedance probability from a series of hypothetical modeled storms that pass through the boxes we have chosen. Since the only constraints are the track location and the category, the selection process will return all modeled events that qualify, and does not constrain these for similar characteristics to Irma (such as RMax, central pressure etc.).
As the situation unfolds, the RMS Event Response team will be following their established process for pre-U.S. landfalls, with the first stochastic track selections likely made on Friday if current timings of landfall forecasts persist.
You May Also Like
Sam has worked at RMS for 10 years, focusing on risk analytics and structuring for cat bonds. During his time at RMS, he has worked on dozens of risk financing projects for private and public entities alike. Such projects include the award-winning parametric cat bond issued by the New York State following Superstorm Sandy to protect the NY Metropolitan Transit Authority from storm surge in the future. Prior to joining RMS, Sam completed a master's degree in operational research and a bachelor's degree in mathematics, both from the University of Warwick.