P2.54 Probabilistic Forecasting of US Market Insured Loss

Wednesday, 18 April 2012
Heritage Ballroom (Sawgrass Marriott)
Mark A. Saunders, University College London, Dorking, Surrey, United Kingdom; and A. S. R. Lea

Tropical Storm Risk (TSR) Business data help (re)insurers to better manage their real-time risk from live tropical storms worldwide. TSR's new probabilistic forecast market loss out to 120hrs lead product was introduced in beta mode for the 2011 hurricane season. This innovative real-time product gives realistic probabilities for a hurricane's market wind loss exceedance (ie the chance the insured market wind loss will exceed different thresholds) out to 5 days before landfall. These probabilistic forecast losses are provided by coastal state. The product employs the TSR ‘ensemble forecast windfields' product and the TSR innovative loss methodology. It models and includes the uncertainty in forecast storm track, the uncertainty in forecast storm intensity and the uncertainty in the insured loss for each ensemble forecast windfield. The product provides user-friendly information with probabilistic forecast wind losses listed in excel chart and spreadsheet format and normalised to 2010 societal conditions. It is available within 30mins of a public tropical cyclone forecast advisory being released and is updated every 6 hours. The product performed well for hurricane Irene (2011). Its accuracy is comparable to the early post-event loss accuracy from Cat modelling companies for Florida-striking hurricanes. The key features, accuracy, timeliness, availability and business benefits of the ‘TSR probabilistic forecast market loss' product will be presented and described with examples.
- Indicates paper has been withdrawn from meeting
- Indicates an Award Winner