>> Research

>> Publication

>> People

>> Press Release

>> Services

>> Home

FIU Public Hurricane Loss Projection Model

What is the model? Why develop a CAT model? Components of the Wind Model
What can the model do? Participating Institutions Components of the Vulnerability Model
Potential Future Capabilities The Project Team Components of the Insured Loss Model
What can the model be used for? Model Design Selected Output

What is the model?

The model is essentially a very complex, state of the art, set of computer programs. The programs simulate and predict how, where and when hurricanes form, their wind speed and intensity and size etc, their track, how they are affected by the terrain along the track after landfall, how the winds interact with different types of structures, how much they can damage house roofs, windows, doors, interior, contents etc, how much it will cost to rebuild the damaged parts, and how much of the loss will be paid by insurers.

Design wise the model consists of three major components: wind hazard (meteorology), vulnerability (engineering), and insured loss cost (actuarial). It has over a dozen sub-components. The computer platform is designed to accommodate future hookups of additional sub-components or enhancements.

A presentation on the model in PDF format is available here.

What can the model do?

The model can assess hurricane risk, and predict annual expected insured residential losses in Florida for an insurance company, zip code, county, or for the entire state. Separate loss estimates are produced for structure, content, appurtenant structure, and additional living expenses. These losses can be reported for portfolios classified by construction type (e.g. masonry, frame, manufactured homes), by county or zip code, by policy form (e.g., HO-3, HO-4 etc.), by rating territory, and combinations thereof. Furthermore, the model can generate for a given portfolio of policies, the return time, the probability of exceedance, and the probable maximum loss.

The model can also immediately predict losses from a given event such as hurricane Katrina or Wilma. It can, for example, estimate losses from a Cat 5 hurricane landing in a major metropolitan area such as Miami, Tampa or Jacksonville. These estimates can be useful to the insurance companies, regulators and local and state government. Such estimates were provided to OIR for hurricane Katrina and Wilma last fall.
The model also has some limited capability to estimate the benefit (loss reduction) from certain mitigation efforts.

Potential future capabilities

Currently the model can estimate hurricane losses for only residential structures. The project team can develop additional modules if funding becomes available. It can design and produce components to estimate losses for commercial structures and high rises. Similarly, modules can be added to estimate benefit of mitigation for all types of construction. Model will then be able to estimate, for example, decline in losses in a two story frame house with gable roof if it had steel shutters or roof straps. Modules can be implemented to perform solvency and market analysis for insurance companies. The team can also develop similar models for other vulnerable coastal states.

The model can be used

  • by the state regulators to help evaluate rate filings

  • by insurance companies to assess hurricane risk and generate loss estimates that can be used as input in the rate making process

  • to assess hurricane risk and predict losses for counties and zip codes

  • to make Cat models affordable for smaller firms

  • to provide a state of the art wind hazard, vulnerability and insured loss models

  • to provide a check on the assumptions, analysis and results generated by the proprietary models

  • to help evaluate reinsurance risk for, e.g., the Florida CAT Fund

  • to assess the benefit of disaster mitigation strategies

Why develop a CAT model?

Traditional actuarial models and practices are inherently ineffective in dealing with low frequency, high severity catastrophic losses. Losses are predicted using recent past experience and limited data. This can lead to volatile premiums, and sharp periodic jump in premiums. This is bad for homeowners and the insurance firms. CAT models have a long term horizon, use more realistic models to estimates losses, better deal with low frequency events, and can potentially result in relatively stable premiums.

Participating Institutions

  • Florida International University/ IHRC (lead institution)
  • Florida State University
  • Florida Institute of Technology
  • Hurricane Research Division, NOAA
  • University of Florida
  • University of Miami

The project team

Dr. Shahid Hamid (PI and Project Director) Dept of Finance and IHRC, FIU
Dr. Shu-Ching Chen (Co-PI) School of Computer Science, FIU
Dr. Jean Paul Pinelli Dept of Civil Engineering, FIT
Dr. Mark Powell Hurricane Research Division, NOAA
Dr. Sneh Gulati Dept. of Statistics, FIU
Dr. Golam Kibria Dept. of Statistics, FIU
Dr. Kurtis Gurley Dept of Civil Eng, Univ of Florida
Dr. Steven Cocke Dept of Meteorology, FSU
Neil Dorst Hurricane Research Division, NOAA
Dr. Mei-Ling Shyu Dept. of Electrical & Computer Eng, Univ of Miami
Dr. George Soukup Applied physicist, AOML/NOAA
Bachir Annane CIMAS/UM/NOAA Hurricane Research Division
Dr. Mani Subramaniam Dept of Mechanical Engineering, FIT

 
 
 

 

 

 

 

 

 


 



• Plus over a dozen graduate students

Model design

The model consists of three major components: wind hazard (meteorology), vulnerability (engineering), and insured loss cost (actuarial). It has over a dozen sub-components. The major components are developed independently before being integrated. The computer platform is designed to accommodate future hookups of additional sub-components or enhancements.

Components of the Wind Model

  • Storm Track and Intensity Model: Generates the storm tracks and intensity up to close of land for simulated hurricanes based on historical initial conditions.
  • Inland Storm Decay Model: Estimates decay after landfall.
  • Wind Field Model: Generates open terrain wind speeds for each of the hurricane affected zip code.
  • Gust Factor Model: Generates peak gust wind speeds for each zip code.
  • Terrain Roughness Model: Corrects open terrain wind speed for terrain roughness.
  • Wind Probabilities Model: Generates wind probabilities for each zip code.
  • ArcIMS environment to visualize Florida GIS information and the associated data results over the Internet.

Components of the Vulnerability Model

  • Engineering simulation models: Simulates for each type of construction, all possible wind damages to the structure, interior, contents, appurtenant structure, as well as ALE.
  • Engineering damage model: Generates damage matrices for each construction type. Produces damage ratios for structure, contents, appurtenant structure, and ALE.
  • Engineering Mitigation Model: Generates vulnerability functions (damages matrices) for mitigated structures (e.g., with shutters, braced gable ends, hip roof, wall to roof straps etc.).

Components of the Insured Loss Model

  • Model for policy modifications: models wind deductibles, as well as policy limits etc.
  • Probabilistic insured loss actuarial model: Generates expected annual loss costs for each policy, or portfolio of policies, or by zip code, county, rating territory, construction type etc. Adjusts for deductibles and limits etc. Generates structure, content, AP and ALE loss.
  • Scenario based insured loss actuarial model: generates expected loss cost for a specific hurricane event affecting a given region.

Selected Output

 
Expected residential insured wind losses for simulated
hurricane landfalls ($million) based on 2003 exposure
Hurricane Category
Landfall Location
1
2
3
4
5
Jacksonville 30 88 503 4,267 9,557
Fort Pierce 59 207 1,394 6,043 10,847
Miami 179 523 3,124 12,786 25,849
Ft. Myers 90 505 2,017 8,808 14,582
Tampa 187 618 5,087 21,300 35,633
Panama City 10 29 211 1,102 2,532

 

 

 

 

 

 

Please note these are for residential policies (home owners and renters) and do not include commercial policies or condos.