Florida Commission on Hurricane Loss Projection Methodology Reviews, Approves Models Submitted Under 2011 Standards
Jun 24, 2013
The Florida Commission on Hurricane Loss Projection Methodology (“Commission”) met last week on June 18, 19 and 20, 2013 to hear modeler presentations on new loss projection software platforms. Computer models submitted under the Commission Standards and Acceptability Process for 2011 (“Standards”) were also evaluated.
To access complete meeting materials, click here.
A summary of each day’s meeting follows.
Day 1–June 18
Dr. Jack Nicholson, Chairman of the Commission’s Acceptability Process Committee and Chief Operating Officer of the Florida Hurricane Catastrophe Fund, noted a need for the Commission to address issues outside its traditional scope. In keeping with that objective, the first portion of meeting included a workshop on modeling storm surge and related new software platforms.
Commission Chair Lori Medders opened the meeting. Modeler software platform presentations followed introductions of Commission members and others in attendance.
Representatives from AIR Worldwide Corporation advised that, in January 2013, their company released its new software platform “Touchstone” to its clients and is anticipating Commission review and approval of the product in 2014.
While the model presented for review by AIR was CLASIC/2 V.14.0.1, AIR representatives requested that the Commission include Touchstone during its review of CLASIC/2, since the revisions, they explained, are not significant pursuant to Section II.B of the Standards. These may be small, but insignificant changes to loss costs due to changes in software, as well as possible issues of time required for model users to transition from CLASIC/2 to Touchstone in 2013 and 2014, it was explained. The Commission members questioned how significant these changes are.
EQECAT, Inc. has released its new platform, called “Risk Quantification and Engineering (“RQE”),” which models loss from a variety of catastrophic risks, including hurricanes, floods, earthquakes and others. EQECAT representatives on hand at the meeting explained that RQE is expected to improve model performance scalability and transparency, with minimal changes from the most recent version.
Risk Management Solutions (“RMS”) representatives discussed the company’s new platform, RMS (ONE), which will be released in April 2014. RMS (ONE) was described as a high-definition integrated modeling solution containing all aspects of modeling in a single platform.
RMS representatives explained that a transition period from Risk Link to RMS (ONE) is planned, during which clients will need to access both models enabled from a single platform. They added that the new model is designed to adapt to industry need for multiple views of risks.
The Commission’s Professional Team members then provided comments on the new software platforms, explaining that the changes are designed to improve the models by improving their implementation.
Dr. Nicholson expressed concern that these new changes could make it much more difficult to audit the models, however. Professional Team members responded that model auditing remains paramount, but will not be impossible, or necessarily more difficult with the new versions. Dr. Nicholson also noted the new models may need to have a more deliberative review.
A consumer representative participating via teleconference from Massachusetts asked the Commission for strict scrutiny of the review and expressed concern with the potential for model results to be manipulated.
The Commission also heard from AIR and RMS representatives on modeling storm surge. Currently, both companies have a storm surge model. Since estimation of storm surge is very location-specific, it was explained, considerations for Commission review of modeled storm surge should include the type of building features, a review of vulnerability framework similar to wind and actuarial considerations such as granularity.
Further, coordination with the National Flood Insurance Program will be critical, it was said. RMS representatives discussed the history of surge hazard technology, challenges in surge modeling and recommendations for Commission Standards, which they said should include hazard, vulnerability and financial considerations. ZIP code data does not provide a high-enough resolution, they added. Rather, storm surge review should be done at street level, which would likely double the size of the review process.
A representative from Florida International University provided an update on recent revisions to HURDAT, the North Atlantic hurricane database.
Dr. Nicholson then recapped the model review process, including the public meeting and the closed meeting for discussion of trade secrets. All notes from the closed meeting will be shredded, he said. The Commission will vote on the six Standards, either one-by-one, or in groups. A model will be considered acceptable if it meets all six Standards.
Next, the Commission held a closed meeting for discussion on the Applied Research Associates, Inc. (“ARA”) HurLoss Version 6.0.a (“HurLoss”) model.
Following that, ARA representatives presented a summary of HurLoss for acceptability by the Commission. After review and discussion of the general, meteorological, vulnerability, actuarial, statistical and computer standards, the Commission approved the model.
Day 2–June 19
The Commission continued its meeting on June 19, beginning with a closed meeting for discussion of trade secrets of AIR’s Atlantic Tropical Cyclone Model V14.0.1.
Dr. Medders opened the public portion of the meeting. Following introductions of Commission members and others in attendance, AIR representatives made a presentation regarding issues on the company’s 2009 model submission, noting that these have been corrected and have no effect on the output ranges. No changes are necessary to the currently released model, they said.
Next, AIR Worldwide representatives presented a summary of Atlantic Tropical Cyclone Model V14.0.1 for acceptability by the Commission. The Commission approved the model after discussion on the applicability of the various Standards.
The Commission then held a closed session for discussions of trade secrets relating to RMS’ North Atlantic Hurricane Model RiskLink 13.0 (Build 1509).
During the public portion of the Commission meeting, RMS representatives presented a summary of North Atlantic Hurricane Model RiskLink 13.0 (Build 1509) for acceptability. The Commission approved it after discussion on the Standards.
Day 3–June 20
On the third day of the meeting, the Commission reviewed EQECAT, Inc.’s Florida Hurricane Model 2013a, which features a new client server-based software and a global multi-peril platform. With empirical data as its basis, the model defines a hazard, propagates hazard sites, and estimates damage and loss. A wind field model that performs locational adjustments has been refined to increase resolution every five minutes. It was explained that this model is subjected to a validation process, which includes post landfall information.
Changes to EQECAT’s model since 2011 have included:
- An update of the probabilistic hurricane database;
- A doubling of simulation time period from 150,000 to 300,000;
- Increase in resolution of time-stepping in the Winfield calculation from 15 minutes to 5 minutes;
- An update in the model’s ZIP code database to March 2012;
- Updated mitigation measures;
- An update of financial models to use discrete calculations instead of numerical integration
All Commissioners waived the reading of the entirety of each model standard in regard to EQECAT’s model. However, the Commission expressed that there was difficulty verifying Standard G-1: “Scope of the Computer Model and its Implementation” during the on-site visit. Notwithstanding, the Standard was verified during EQECAT’s closed session in front of the Commission.
Commissioner Hugh Willoughby requested that an individual vote be held for Standard M-5: “Landfall and Over-Land Friction Weakening Methodologies,” inasmuch as he felt EQECAT’s response had been somewhat glossed-over and not given appropriate attention. Commissioner Willoughby was the lone “no” vote to the question as to whether EQECAT satisfied standard M-5, however, the Commission ultimately determined that EQECAT Florida Hurricane Model 2013a complied with all standards.
After approval of the EQECAT model, the Commission considered general business matters, nominating and electing Floyd Yeager as the Commission’s Vice-Chair. With Mr. Yeager’s appointment as Vice-Chair now leaving a vacancy on the Executive Committee, that position was filled by Kristin Pilzecker, who was appointed by Chair Medders.
The Commission then discussed its 2013 committee assignments. These are as follows:
General Standards Committee:
Kristin Piltzecker (Chair), Barry Gilway, Bob Lee, Jai Navlakha, Jack Nicholson
Meteorological Standard Committee:
Hugh Willoughby (Chair), Randy Dumm, Bryan Koon, Jack Nicholson, Floyd Yeager
Vulnerability Standards Committee:
Randy Dumm (Chair), Bob Lee, Jai Navlakha, Robin Westcott, Floyd Yeager
Actuarial Standards Committee:
Floyd Yeager (Chair), Randy Dumm, Bob Lee, Kristin Pilzecker, Barry Gilway
Statistical Standards Committee:
Lori Medders (Chair), Bryon Koon, Robin Westcott, Hugh Willoughby, Floyd Yeager
Computer Standards Committee:
Jai Navlakha (Chair), Bryan Koon, Kristin Pilzecker, Robin Westcott, Hugh Willoughby
Acceptability Process Committee:
Jack Nicholson (Chair), Randy Dumm, Barry Gilway, Kristin Pilzecker, Robin Westcott
An update on the 2013 Florida Legislative Session was provided by Leonard Schulte, Florida Hurricane Catastrophe Fund Director of Legal Analysis and Risk Evaluation. Mr Schulte noted the passage of SB 1770, which requires the appointment of a structural engineer to the Commission by Florida’s Governor. He also discussed legislation relating to an appropriation for the update of the Florida Public Hurricane Loss Projection Model.
Executive Committee Prioritization of Issues
Dr. Nicholson stated that he intends to propose a method for addressing information not included in initial submissions through the Commission’s Acceptability Process Committee. He also suggested that the Commission review models in an open session prior to conducting a closed, non-public session.
Chair Medders suggested the Commission should address storm surge modeling and asked how proactive the Commission should be in that effort. The Chair also raised the issue as to whether the Commission is the appropriate venue for reviewing mitigation discounts.
Commissioner Bob Lee, an Florida Office of Insurance Regulation actuary, asked whether the factors used to evaluate coastal and inland exposure should be examined. Other Commissioners felt this issue is already sufficiently addressed. However, there was concern as to whether modelers are provided with the correct questions to respond to this issue. The Commission agreed that simply asking for differences related to coastal and inland exposures is not sufficient.
The Executive Committee designated all of the above mentioned issues as priorities for further discussion.
Should you have any questions or comments, please contact Colodny Fass& Webb.
To unsubscribe from this newsletter, please send an email to Brooke Ellis at email@example.com.