.png)
By Enrico Ronchi, Steve Gwynne, Guillermo Rein, Rahul Wadhwani, Paolo Intini
The number of people who need to evacuate from wildfires keeps rising, year after year globally, posing a serious challenge to fire and emergency services and affecting thousands of longstanding and new communities around the world. Decisions made during community planning, property upkeep, emergency planning, public education, responder training and the evacuation itself all heavily rely on the information available; i.e., the evidence on which the wildland urban interface (WUI) response is based.
To address this issue, the U.S. National Institute of Standards and Technology (NIST) funded a multi-disciplinary project conducted by Lund University, the National Research Council of Canada, Imperial College and the Fire Protection Research Foundation, aimed at developing a novel integrated framework for modelling wildfire urban evacuations.1 Such a tool would allow increased situational awareness for both evacuation planning and decision support during an incident, depending on its use.
An integrated approach requires considering and integrating the key modelling layers in WUI evacuations, namely fire spread, pedestrian movement and traffic movement. This was a first attempt to address the three key modelling areas in an integrated simulation system. The outcome of the project was a design specification for a system incorporating a suite of tools that can be used in unison to forecast the outcome of an incident to different degrees of refinement, given the constraints and resources available.
Methodology
The first step was conducting a review of recent incidents to ascertain the core scenarios that should be examined and the underlying factors involved. Evaluating the case studies used a similar approach to the one adopted in fire safety engineering — engineering timelines of available and required safe egress time (the concept of WASET vs. WRSET). A literature review evaluated the state-of-the-art of modelling tools for the simulation of different modelling layers of WUI fires.
The models in the review were chosen according to the most authoritative reviews in the last 10 years2-6 and from current practice. Documentation associated with each model (e.g., produced by developers, users and third parties) provided an in-depth understanding of the model accessibility, assumptions, capabilities and limitations. When possible, model developers were also contacted directly to make sure that the information retrieved represented an accurate representation of model capabilities.
Starting from the analysis of recent incidents, the research took into consideration issues associated with WUI fires at varying scales and in scenarios of increasing complexity. This work identified variables affecting the system requirements (spatial and temporal scale, population involved, variables affecting the fire evolution and characteristics of the road network, etc.).
In essence, what model functionality is required to integrate a model into such a WUI evacuation framework? The dual use of the models (planning and decision support for real-time application) was also taken into consideration when providing such information. This was important because it influenced the constraints on the model use — time, computational expense, required output, etc.
Recommendations for each of the three modelling areas (fire, pedestrian and traffic) concerned the most-appropriate modelling approaches to be used at different levels of granularity and in different timeframes (and given different constraints). Reviews of each type of model have been performed with a systematic approach consisting of five steps (see Figure 1).
The first step involved creating a template for model evaluation common to all types of models. The second step was to modify this template to address the specific features of each model type (e.g., the specifics of fire models, etc.). That led to performing a review of the main variables, sub-models and key requirements for integration of each model component in parallel with the template definition (Step 3).
The scope of this review was to identify the benchmark characteristics that a model might need (for all three types of models) for it to be integrated with the other modelling layers for WUI fire evacuation applications (Step 4); i.e., to function and exchange information. This was performed in relation to the final use of the integrated toolkit (i.e., evacuation planning or real-time decision support).
The final step consisted of the review of existing models, considering the characteristics of a benchmark model (Step 5) by setting the criteria for the evaluation and identifying a set of questions for the assessment.
The analysis of existing modelling tools was eventually used to develop an agenda for the future research activities to develop a comprehensive, integrated, multi-physics modelling framework for WUI fire evacuation scenarios.
The Multi-Physics Modelling Framework
The modelling framework was built by identifying the key data exchange requirements between models. The integrated system would have to be highly coupled to involve the required exchange of information between the models (to represent the impact of fire on traffic, or of people on the traffic, etc.) and between the system and the user (meeting user needs). Ignoring the connection between the components would both reduce the integration of the system and artificially isolate the simulated conditions from each other, potentially producing results that diverge from expectation.
Table 1 shows the assessment of the required inputs/outputs exchange between different models is presented. The exchange of information between the models is not symmetrical, primarily due to the different theoretical and empirical maturity of the three subject domains, but also the sophistication of the models currently available and the computational limitations associated with each domain.
This formed the basis for the design of a set of system components that, through suitably configured data exchange enabled system functionality, might produce output of interest, given the need to simulate WUI scenarios.
Conclusion
The emergency response to WUI fires includes the ability of the affected community to prepare for the hazards, adapt to the evolving conditions of the incident, and recover from disruptions in the immediate aftermath of the incident and in the longer term. This is achieved through the efforts of the community itself and emergency responders. To ensure that this preparation and response is adequate, professionals must understand the effectiveness of the pre-incident decisions and decisions taken during the incident needs so they can assess these decisions before they are finalized — that is, before they are put into practice.
Both design and emergency response are key elements in addressing the occurrence, development and impact of WUI incidents. Efforts to inform and improve these elements will affect the frequency and severity of such WUI incidents. This work addressed this need by presenting a system specification for a toolkit able to provide numerical evidence to support the design and emergency response processes.
Further information about the modelling framework, the specifics of the required data exchange and their implications can be found in the final report associated with the project.1
Enrico Ronchi is with the Department of Fire Safety Engineering, Lund University, Sweden. Steve Gwynne is with the National Research Council of Canada. Guillermo Rein is with the Imperial College, Imperial College, London, UK. Rahul Wadhwani is with Victoria University, Australia. Paolo Intini is with the Polytechnic University of Bari, Italy.
References
1E. Ronchi, G. Rein, S.M.V. Gwynne, R. Wadhwani, P. Intini, and A. Bergstedt. 2017. “e-Sanctuary: Open Multi-Physics Framework for Modelling Wildfire Urban Evacuation.” Quincy, MA: Fire Protection Research Foundation.
2S. Gwynne, E.R. Galea, M. Owen, P.J. Lawrence, and L. Filippidis. 1999. “A review of the methodologies used in the computer simulation of evacuation from the built environment.” Build. Environ., 34(6)741–749.
3E.D. Kuligowski, R.D. Peacock, and B.L. Hoskins. 2010. “A Review of Building Evacuation Models, 2nd Edition, NIST Technical Note 1680.” Washington, DC: National Institute of Standards and Technology.
4A J. Pel, M.C.J. Bliemer, and S.P. Hoogendoorn. 2012. “A review on travel behaviour modelling in dynamic traffic simulation models for evacuations,” Transportation, 39(1), 97–123.
5A. L. Sullivan. 2009. “Wildland surface fire spread modelling, 1990–2007. 3: Simulation and mathematical analogue models.” Int. J. Wildland Fire, 18(4), 387–403.
6A. L. Sullivan. 2009. “Wildland surface fire spread modelling, 1990–2007. 2: Empirical and quasi-empirical models,” Int. J. Wildland Fire 18(4), 369–386.