announcement header

Shree Khare
Risk Management Solutions Inc.
Monday, May 11, 2009
Mesa Laboratory, Damon Room
Lecture 11:00am

Data Assimilation Challenges in Real Time Catastrophe Risk Modeling

Statistical simulation models are used by financial companies to quantify and manage risk associated with catastrophic events like major landfalling hurricanes. When real events occur, reliable information about the distribution of insured losses is a critical element to the internal decision making processes of these companies. In this talk we begin by reviewing the fundamental concepts behind building such simulation models using US hurricane risk as an example. We then identify the real time data assimilation problem and discuss the observational information that is at our disposal. We then discuss the variety of ensemble assimilation methods that can feasibly be applied to this problem, show some preliminary results, and discuss some interesting research questions that arise in this context.