Egress Modeling Workshop - 9th IAFSS Symposium, 2008

S.M.V.Gwynne
Hughes Associates, Inc.

The workshop concerns the application and development of egress models. It is designed for students, researchers, model developers, model users and those that sit in judgment of the results produced. It will primarily relate to the availability of suitable data and the validation process required of the data, the models, the users and the third party observers.

The workshop will be held on the Sunday 21st September before the main conference activities begin. The workshop is scheduled to run between 1-4pm, providing three hours of presentation and discussion time.

The panel will consist of model developers, model users, consultant engineers, and data collectors. The panel members will be drawn from the US and from Europe and will include

David Purser (Hartford Environmental Research, UK)
Rita Fahy (NFPA, US)
Barbara Lane (Arup, UK)
Hubert Kluepfel (TraffGo-HT, Germany)
Erica Kuligowski (NIST, US)
Daniel Nilsson (Lund University, Sweden)

along with Steven Gwynne (Hughes Associates, Inc., US) who will be the moderator and a panel member.

Each member will be asked to give a 10-15 minute informal presentation; i.e not requiring Powerpoint or overhead slides. This informal format is preferred as it will encourage greater participation from the audience. Approximately 90 minutes will be taken up with presentations, while 90 minutes will be dedicated solely to discussion involving the audience. These two discussions may be separate or intermingled depending on how the workshop progresses. A summary of the panel member responses and of the discussion will be produced and distributed to attendees after the event.

Below is the list of questions. These questions are purely to prompt responses from the panel and promote discussion during the workshop. They may be answered individually, in combination or can simply be used as a guide to the topics of interest: data availability and validation. They are a guide as to the expected subject-matter of the workshop.

Data-Related
Primary Questions:	Secondary Questions:
What are the key data requirements for the current use, development and validation of egress models? Is there currently sufficient data for egress models to be developed/applied with any degree of reliability? Who should be collecting this data? What format should this data be in? What measures should be taken to allow egress models to be applied, in the absence of this data; i.e. discuss the concept of engineering judgment?	Given the proliferation of CCTV cameras, should we exploit the data being captured on a daily basis? Does the non-emergency nature of this additional data pose issues that need to be addressed? How can we encourage those collecting data to (1) collect the necessary data required for model development; (2) provide the data in the correct format? What mechanisms or incentive can be provided to the data collectors in order for them to do this?

Model Development
Primary Questions:	Secondary Questions:
How do we ensure the validity of the egress models used in the field? Particular attention should be given to the data employed in these models. Should it be left to the model developer alone to validate the egress model or should third-parties be involved? What are the key developments required in egress modeling?	In what situations can the assumed default values associated with egress models be employed with any confidence? What are the basic output requirements of an egress model? Should this necessarily include sophisticated graphics? Can an egress model be as accurate as a fire model? Is sufficient guidance provided on the development and application of egress models? If not, what is missing? Who should provide this guidance, if required? Are there any circumstances in which it is acceptable for a simple model to be employed to gauge egress performance?

Users/Observers
Primary Questions:
How can we ensure that users are sufficiently qualified to produce reliable results (i.e. how can we ensure that the user is validated and sufficiently informed)? By increasing accessibility of the models (e.g. by simplifying the use of the model) are we encouraging under-qualified users? Is it possible for those responsible for assessing results to have sufficient familiarity with the range of egress models applied to reliably assess the results produced? Should expert panels/developers be organized to assist in this process? What information should accompany simulation results in order for them to be judged reliably?

General
Primary Questions:
Given the current status of egress models, can we reasonably apply them to investigate actual incidents more forensically? If not, what developments need to be made? Should a greater burden be placed on the predictive capabilities of the model, or should the burden of anticipating behavioral response be placed on the user? What have we learned from the application of egress models to the Rhodes Island, Gothenburg and WTC incidents? In particular, what additional insight was provided by the egress models and what data was required in order for this to take place?