NIST Technical Note

NIST TN 2262

WUI Fire Evacuation and

Sheltering Considerations
Assessment, Planning, and Execution
(ESCAPE)

Alexander Maranghides
Eric D. Link

This publication is available free of charge from:

https://doi.org/10.6028/NIST.TN.2262
 NIST Technical Note
NIST TN 2262

WUI Fire Evacuation and

Sheltering Considerations
Assessment, Planning, and Execution
(ESCAPE)

Alexander Maranghides
Eric D. Link
Fire Research Division
Engineering Laboratory

This publication is available free of charge from:

https://doi.org/10.6028/NIST.TN.2262

August 2023

U.S. Department of Commerce

Gina M. Raimondo, Secretary

National Institute of Standards and Technology

Laurie E. Locascio, NIST Director and Under Secretary of Commerce for Standards and Technology
Certain commercial entities, equipment, or materials may be identified in this document in order to describe an
experimental procedure or concept adequately. Such identification is not intended to imply recommendation or
endorsement by the National Institute of Standards and Technology, nor is it intended to imply that the entities,
materials, or equipment are necessarily the best available for the purpose.

NIST Technical Series Policies

Copyright, Fair Use, and Licensing Statements
NIST Technical Series Publication Identifier Syntax

Publication History
Approved by the NIST Editorial Review Board on 2023-08-17

How to Cite this NIST Technical Series Publication

Maranghides A and Link E (2023) WUI Fire Evacuation and Sheltering Considerations: Assessment, Planning, and
Execution (ESCAPE). (National Institute of Standards and Technology, Gaithersburg, MD), NIST Technical Note
(TN) 2262. https://doi.org/10.6028/NIST.TN.2262

NIST Author ORCID iDs

Alexander Maranghides: 0000-0002-3545-2475
Eric D. Link: 0000-0002-7784-5023

Cover Page Photos

Left set: Photographs from the Camp Fire, 8 November 2018. Clockwise from top left:
Evacuating civilian vehicles exposed to fire on Pearson Rd, 09:41 (CAL FIRE);
Evacuation traffic on Skyway, 10:57 (Paradise Police Department);
Burned vehicles abandoned on Pearson Rd, 15:28 (U.S. Forest Service);
Civilians at the Optimo TRA, 12:07 (CAL FIRE).

Right set: Diagrams depicting four scenarios of fire impact to a WUI community and egress artery. See Fig. 11.
 NIST TN 2262
August 2023

Abstract
Impacts of wildland-urban interface (WUI) fires continue to rise in the U.S., as evidenced by the
string of devastating and record-breaking events occurring since 2017. As seen in several events
in recent years, WUI fires can impact communities quickly, leaving little to no time for civilians
to evacuate. Numerous events have also occurred internationally, including Australia, Canada,
Chile, Greece, and Portugal.
One example is the Camp Fire that occurred on November 8, 2018, in Butte County, CA. The
fire resulted in 85 fatalities and the destruction and damage of over 18 000 buildings, destroying
over 90 % of the buildings in the town of Paradise. Following the fire, NIST initiated a case
study to document and analyze fire spread and behavior, notifications, evacuations, and
defensive actions to support preparedness for future WUI fires. The NIST Camp Fire case study
has highlighted a number of potential challenges that intermix communities may face during
WUI fire events. The purpose of this report is to use the lessons learned from the NIST Camp
Fire case study to present a methodology and other considerations about WUI fire incidents that
can be used by small and intermediate-sized WUI communities to help develop notification and
evacuation plans.
The proposed methodology considers the spatial and temporal components of fire spread and the
resulting impacts of fire on evacuation to develop an evacuation triangle that can be used as the
foundation for notification and evacuation decisions by emergency managers. This report
provides communities a path forward for assessing, planning, and implementing a
notification/evacuation plan that leverages pre-fire conditions, local knowledge, and during-
event information to enhance the life safety of civilians and first responders.
While additional research will provide further refinements, specifically in the areas of weather
forecasting, fire spread modeling, and evacuation modeling, the proposed system outlines a path
for community leaders to effectively work with first responders before a fire to assess and
prepare the community for WUI fire events that can strike with little or no notice. The
methodology provides community leaders with a temporal context of WUI fire events that will
enable them to better evaluate different hazard reduction and risk management strategies to
enhance the life safety of residents and first responders.

Keywords
Camp Fire; community hazard reduction; disaster resilience; emergency notification; evacuation;
intermix; interface; notification; pre-fire planning; public safety; wildland-urban interface; WUI

iii
NIST TN 2262
August 2023

This page intentionally left blank.

iv
 NIST TN 2262
August 2023

Table of Contents
Abstract ...................................................................................................................................... iii
Table of Contents ........................................................................................................................ v
List of Tables ............................................................................................................................. vii
List of Figures ............................................................................................................................ ix
List of Camp Fire Examples ...................................................................................................... xi
List of Symbols, Abbreviations, and Acronyms ................................................................... xiii
Acknowledgments .................................................................................................................... xv
Executive Summary ................................................................................................................ xvii
1. Introduction ............................................................................................................................. 1
2. Community Evacuation and Alternatives ............................................................................. 5
2.1. Evacuation .......................................................................................................................... 5
2.2. Evacuation Alternatives ...................................................................................................... 6
2.2.1. Stay and Defend .......................................................................................................... 8
2.2.2. Shelter-in-Place ......................................................................................................... 10
2.2.3. Areas of Last Resort and Temporary Refuge Areas .................................................. 11
2.3. Defensible Space ............................................................................................................. 11
2.4. WUI Fire and Evacuation Modeling .................................................................................. 14
2.5. Evacuation Trigger Models ............................................................................................... 14
2.6. Evacuation Triangle .......................................................................................................... 15
3. Technical Challenges of Community Evacuations from WUI Fire ................................... 17
3.1. Distinguishing Characteristics of WUI Fire Disasters ....................................................... 17
3.1.1. Defensive Actions Affect WUI Fire Outcomes ........................................................... 18
3.1.2. The Asset is the Fuel ................................................................................................. 18
3.1.3. Fires Can Start New Fires ......................................................................................... 18
3.1.4. Notification Times Range from Minutes to Days ........................................................ 18
3.1.5. WUI Fires Have Limited Advanced Warning and Locally Variable Intensity .............. 19
3.1.6. No Standardized WUI Fire Shelters ........................................................................... 19
3.1.7. Wildfires Are Relatively Frequent .............................................................................. 20
3.2. WUI Fire Evacuations ....................................................................................................... 20
3.3. Compounded Uncertainties in Fire/Evacuation Predictions ............................................. 26
3.3.1. Progression of a WUI Fire Event ............................................................................... 26
3.3.2. Pre-fire vs. During Fire Modeling ............................................................................... 32
3.4. List of WUI Community Evacuation Challenges ............................................................... 35
4. Fire-Evacuation Temporal Relationships and Evacuation Failures ................................. 37
4.1. Primary Modes of Evacuation Failures ............................................................................. 37
4.1.1. Defining Failure .......................................................................................................... 37
4.1.2. Addressing Type 1 Evacuation Events ...................................................................... 40

v
NIST TN 2262
August 2023

4.1.3. Addressing Type 2 Evacuation Events ...................................................................... 40

4.2. Temporal Relationships Among Fire Progression, Notification, Evacuation, and Sheltering
....................................................................................................................................... 43
4.2.1. Minimum Time for Community Evacuation ................................................................ 43
4.2.2. Fire/Notification Timing Scenarios ............................................................................. 45
4.2.3. Evacuation Scenarios ................................................................................................ 50
4.3. Relationships Among Fire Ignition, Fire Growth, and Impact to Community .................... 58
4.4. Temporal Illustration of Full Community Evacuation Scenarios ....................................... 59
4.4.1. Scenarios 1a and 1b .................................................................................................. 60
4.4.2. Scenario 2.................................................................................................................. 61
4.4.3. Scenario 3.................................................................................................................. 61
4.4.4. Scenario 4.................................................................................................................. 61
5. Proposed Approach .............................................................................................................. 63
5.1. Mitigating Civilian Fire Exposures During Evacuation ...................................................... 63
5.2. Safety Areas – Wildfire Safety Zones and Community Sheltering Areas (Shelter in
Community) ............................................................................................................................. 64
5.2.1. Temporary Refuge Areas .......................................................................................... 64
5.2.2. Wildfire Safety Zones ................................................................................................. 65
5.2.3. Community Fire Shelters ........................................................................................... 68
5.3. Developing a Coupled Fire-Evacuation System ............................................................... 69
5.3.1. Trigger Zone Definitions ............................................................................................ 69
5.3.2. Determining Ignition Zone Widths .............................................................................. 74
5.4. Community Evacuation Options and Decisions ............................................................... 76
5.4.1. Shelter in Community ................................................................................................ 76
5.4.2. Partial vs. Complete Community Evacuation ............................................................. 77
6. Implementation ..................................................................................................................... 81
6.1. Assessment ...................................................................................................................... 81
6.2. Planning ........................................................................................................................... 84
6.2.1. Developing the Community Notification and Evacuation Plan ................................... 84
6.2.2. Accounting for Uncertainties and Including Safety Factors ....................................... 87
6.3. Execution .......................................................................................................................... 88
6.3.1. Pre-Planning and Normal Operations ........................................................................ 89
6.3.2. High Hazard Conditions ............................................................................................. 89
6.3.3. During a Fire .............................................................................................................. 89
7. Recommendations ................................................................................................................ 91
8. Summary ................................................................................................................................ 93
References ................................................................................................................................. 95
Appendix A. California Large-Loss Fire Statistics............................................................... 103

vi
 NIST TN 2262
August 2023

List of Tables
Table 1. Differences between private civilians and firefighters. .................................................. 10
Table 2. Characteristics of WUI fires compared to other selected natural disasters. ................. 17
Table 3. Comparison of evacuations for individual buildings vs. a campus vs. a larger
community. .................................................................................................................................. 21
Table 4. Evacuation capabilities of civilians at different locations. ............................................. 24
Table 5. Characteristics of evacuation and sheltering options. .................................................. 52
Table 6. Green Zone inner boundary distance, in miles, from edge of intermix community.
(1 mi = 1.6 km) ............................................................................................................................ 75
Table A-1. Recent WUI fire structure loss statistics in California (CAL FIRE State Responsibility
Area). ........................................................................................................................................ 103

vii
NIST TN 2262
August 2023

This page intentionally left blank.

viii
 NIST TN 2262
August 2023

List of Figures
Fig. 1. Evacuation triangle illustrating connectivity among evacuation trigger zones, available
time before fire reaches community, and evacuation decisions. ................................................ 16
Fig. 2. Primary wildland/WUI fire event components leading to community evacuation. ............ 27
Fig. 3. Linked modules associated with evacuation predictions. Uncertainties are compounded
and propagate from left to right and illustrated in red (not to scale). ........................................... 33
Fig. 4. Fire/notification/evacuation timeline scenarios as a function of evacuation status and
distance between the community and fire origin. ........................................................................ 45
Fig. 5. Flow chart depicting generalized evacuation scenarios. Red text indicates hazard. ....... 51
Fig. 6. Idealized relationship between ignition location, a) near or b) far, from a WUI community.
The fire front and ember exposures reaching the community are illustrated. The wind is directed
from left to right. (Figure from Ref. [7]). ....................................................................................... 58
Fig. 7. Temporal representation of ignition, fire exposure, evacuation warning, evacuation order,
and evacuation. The potential fire hazard to evacuees is indicated by color: green = low,
orange = moderate, and red = high. ........................................................................................... 60
Fig. 8. Conceptual illustration of three ignition zones around a WUI intermix community. Zones
may be asymmetrical because of fuels, fire history, topography, and prevailing winds. Fire
spread directionality and intensity may not be uniform from all directions towards the
community. .................................................................................................................................. 70
Fig. 9. Effect of fire spread deviation on community impact for ignitions near and far from a
community. a) fire spread deviation of 15° will affect whether the community is impacted, b)
similar deviation will not result in a no-impact scenario. Impacts of fuels and topography not
shown.......................................................................................................................................... 73
Fig. 10. First order assessment of initial impact of fire from nearby ignition resulting in partial
evacuation................................................................................................................................... 77
Fig. 11. Four second order WUI community assessments accounting for wildfire ignition and
egress artery locations and topography. ..................................................................................... 79

ix
NIST TN 2262
August 2023

This page intentionally left blank.

x
 NIST TN 2262
August 2023

List of Camp Fire Examples

Camp Fire Example 1. Introduction to the 2018 Camp Fire. ....................................................... 3
Camp Fire Example 2. Defensible space and exposures from neighboring parcels. ................ 13
Camp Fire Example 3. Inadequacy of existing infrastructure buildings as fire shelters. ........... 22
Camp Fire Example 4. Fuel treatments alone are likely not sufficient to protect existing
structures for use as shelters. ..................................................................................................... 23
Camp Fire Example 5. Paradise Police Department 911 dispatch evacuated. ......................... 26
Camp Fire Example 6. Ignition location and rapid fire spread. .................................................. 28
Camp Fire Example 7. Spot fires in Paradise. .......................................................................... 29
Camp Fire Example 8. Escalation of traffic gridlock. ................................................................. 30
Camp Fire Example 9. Burnover events that impacted evacuating civilians and responding
emergency personnel. ................................................................................................................ 39
Camp Fire Example 10. Evacuation of assisted living facilities. ................................................ 41
Camp Fire Example 11. Time of fire arrival and first official evacuation notification, by
evacuation zone. ......................................................................................................................... 46
Camp Fire Example 12. Simultaneous arrival of fire and evacuation notification, leading to
entrapment during evacuation in Concow. .................................................................................. 48
Camp Fire Example 13. Evacuation impacted by fire along egress artery. ............................... 49
Camp Fire Example 14. Range of exposure levels experienced at TRAs. ............................... 53
Camp Fire Example 15. Entrapment en route to the safety zone at Camelot Meadow in
Concow. ...................................................................................................................................... 55
Camp Fire Example 16. Entrapment during evacuation from the fire area. .............................. 56
Camp Fire Example 17. Safe evacuation from Paradise after shelter in TRA........................... 57
Camp Fire Example 18. TRA use during the Camp Fire. .......................................................... 65
Camp Fire Example 19. Natural areas used as wildfire safety zones. ...................................... 67
Camp Fire Example 20. Defensive actions at TRAs. ................................................................ 69
Camp Fire Example 21. Spot fire ignitions on Skyway and Andover Drive in Magalia.............. 72
Camp Fire Example 22. Humboldt Fire (2008). ......................................................................... 74
Camp Fire Example 23. Paradise fire history. ........................................................................... 81
Camp Fire Example 24. Impact of traffic gridlock beyond the immediate community. .............. 83

xi
NIST TN 2262
August 2023

This page intentionally left blank.

xii
 NIST TN 2262
August 2023

List of Symbols, Abbreviations, and Acronyms

AHJ
authority having jurisdiction

BO
burnover

FFL
fire front length

IC
incident commander

ICS
Incident Command System

IL
interface length

LE
law enforcement

PPE
personal protective equipment

SIP
shelter in place

TRA
temporary refuge area

xiii
NIST TN 2262
August 2023

This page intentionally left blank.

xiv
 NIST TN 2262
August 2023

Acknowledgments
The authors would like to acknowledge and thank Steve Hawks (California Department of
Forestry and Fire Protection, retired) for his thoughtful discussions, experience, and review of
this document. He also compiled the fire loss statistics for California, presented in Appendix A.
Thanks go to Karen Jackson (Texas A&M Forest Service, retired) for assistance with the
literature search and review of the document.
The authors appreciate the review and feedback from NIST WUI Fire Group colleagues Drs.
Glenn Forney and Kathryn Butler, WUI Fire Group Leader Tom Cleary, and Fire Research
Division Chief Matthew Hoehler.

xv
NIST TN 2262
August 2023

This page intentionally left blank.

xvi
 NIST TN 2262
August 2023

Executive Summary
Impacts of wildland-urban interface (WUI) fires continue to rise in the U.S., as evidenced by the
string of devastating and record-breaking events occurring since 2017. As seen in several events
in recent years, WUI fires can impact communities quickly, leaving little to no time for civilians
to evacuate. Numerous events have also occurred internationally, including Australia, Canada,
Chile, Greece, and Portugal.
Minutes matter during WUI fires. Fire can rapidly impact a community, impede evacuation
operations, and result in burnovers of evacuees and first responders. It is essential for
communities in fire-prone WUI areas to have pre-existing evacuation plans to efficiently use the
potentially limited available time during incidents. Pre-fire development of evacuation plans has
the added benefit of enhancing communication with surrounding/participating jurisdictions and
enabling effective communication pathways during rapidly developing WUI fire events.
Frequently, pre-planning is focused on scenarios where sufficient time for evacuation exists, and
solutions for scenarios where the fire will impact the community before there is sufficient time to
safely evacuate are underdeveloped or not considered. Community evacuations in response to
WUI fires are complex; they are influenced by dynamic conditions and numerous variables
ranging from fire behavior to human behavior. Findings from previous WUI fire case studies
indicate that even individuals who were well prepared for their primary response action to a WUI
fire, either to evacuate or to remain on their property, were overcome by fire when their primary
plan was not successful and they had not considered contingency plans. Similarly, at the
community level, citizens will benefit from a well-developed, pre-planned, community
evacuation response that includes contingencies to accommodate a wide range of scenarios and
random complications.
WUI intermix communities contain a significant amount of vegetative fuel scattered throughout
the community, and many also contain areas of moderate to high density structures and other
WUI fuels. Fire spread under these conditions can overcome the limited benefits of the small
areas of defensible space available on small parcels, and exposures can exceed structure
hardening and ignition mitigation attempts. The wide range of potential fire and ember exposure
intensities renders stay-and-defend or shelter-in-place responses at residences hazardous at best,
and deadly at worst.
WUI fires are different from other disasters primarily because the fuel that drives the event is
also the asset that is being protected. This presents a unique opportunity to influence both the
intensity and the impact of the disaster, with direct benefit to life safety. Targeted fuel reduction
and management to reduce potential exposures is an important aspect of community evacuation
preparedness and capability. Such approaches are particularly beneficial along egress arteries and
around pre-planned refuge areas for contingency use as a last resort.
In this report, technical challenges related to community evacuations are discussed with respect
to WUI fire event progression and life safety issues due to fire exposures potentially encountered
during evacuation. Lessons learned from a detailed case study of the evacuations during the 2018
Camp Fire in Butte County, CA highlight critical concepts of WUI fire evacuations, including:
• the relationship between the time and location of ignition and the time the fire will impact
the community,

xvii
NIST TN 2262
August 2023

• the number of civilians that can become trapped in intense fire exposures (i.e., burnovers)
during evacuation,
• the widespread use of temporary refuge areas (TRAs) in urgent attempts to shelter from
dangerous fire exposures and also as an evacuation and traffic management tool, and
• the benefits of a pre-disaster evacuation drill that provides a training opportunity for first
responders to implement traffic management tools like counterflow and to coordinate with
relevant partner agencies.
Twenty-four examples from the Camp Fire included throughout the report provide context for
the developed methodology and highlight real incident outcomes, complications, and
considerations.
While research has been conducted to further the understanding of different parts of the
evacuation process, standardized methodologies do not yet exist for defining trigger zones to
support evacuation response decision-making, specifically in the context of what communities
can accomplish during WUI fire conditions. This report outlines a technical approach to help
communities develop a comprehensive evacuation strategy that addresses spatiotemporal
considerations and constraints associated with WUI community evacuations. The methodology
and framework to develop these evacuation trigger zones are based on potential fire spread rates
and the required evacuation time specific to the local community. Finally, concepts for planning
and decision-making support tools that can be used in development of community notification
and evacuation plans are presented.
The trigger zone concept is based on a comparison between the required safe egress time of a
WUI community, WUI RSET or WRSET, and the available safe egress time, WASET. Both
measures are determined by fire spread parameters and expected impacts to the community and
egress pathways. Determining the elapsed time from fire ignition to the activation of emergency
notifications and evacuation orders (ITA) and the subsequent community evacuation time (ET)
can establish the minimum amount of time from ignition to safety (ITS), or WRSET. Successful
evacuation planning will require data collection of both ITA and ET values through training and
drills. Modeling may also be used to augment drills, but at this time will likely not be a substitute
for full community evacuation drills. Comparing RSET to the anticipated timeline of fire
progression (ASET) can determine the requisite evacuation (or alternative) response. In dire
scenarios, when WASET < WRSET, evacuation will not be possible without exposing civilians
to hazardous conditions. For these scenarios, implementation of pre-planned alternatives will be
necessary. Linking these critical evacuation timescales to the distance fire may spread in that
time can create designated geographic fire-evacuation trigger zones that correspond to different
responses.
This report is intended for existing small and intermediate-sized intermix communities and
isolated interface communities, although many concepts and incident considerations apply to
larger communities and urban interface areas as well. Specific emphasis is given to providing a
methodology to address fire events that impact a major portion of the community, resulting in
limited or no options available for safe evacuation. This is particularly important for
communities that may not have the resources or expertise to conduct or evaluate a more complex
evacuation analysis. While additional research is needed to optimize the concepts discussed here,
the presented information can inform communities and help develop/improve community
sheltering and evacuation planning.

xviii
 NIST TN 2262
August 2023

1. Introduction
Fires at the wildland-urban interface/intermix (WUI) pose a serious threat to the life safety of
residents, evacuees, and first responders, as evidenced from numerous events within the past
decade requiring rapid large-scale evacuations and resulting in destroyed communities and loss
of life.
Wildland fires can impact communities within minutes or hours after ignition. If a fire ignites
within a community, whether as an initial ignition or as a spot fire from a larger fire, it can
rapidly grow and significantly impact civilian evacuations within and from the community. Fires
in intermix communities are driven by both the built environment and the vegetative fuels
located throughout. While there are often efforts to manage fuels and fire exposures through
structure and parcel hardening, fires occurring in environments with high fuel densities can result
in dangerous fire exposures to evacuating civilians. Fire can block—or worse, entrap—
evacuating traffic and hinder first responder operations. There is a need to understand the
spatiotemporal relationships among fire ignition location, fire rate of spread, and community
evacuation.
Community evacuations in response to WUI fires are complex; they are influenced by dynamic
conditions and numerous variables. Evacuation has often been considered as a binary decision
made by civilians—either evacuate or stay, with those who choose to stay taking shelter or
protecting property. In the U.S., significant emphasis has been placed on the “Ready, Set, Go!”
[1, 2] education campaign that encourages awareness and evacuation, while a limited number of
communities have adopted variations of shelter-in-place responses [3]. In Australia, a “stay and
defend” approach is more widely considered as an option within the “Prepare Act Survive”
campaign [4, 5], which encourages advanced planning and deliberate action whether one decides
to stay or evacuate.
While research has been conducted to further the understanding of different parts of the
evacuation process, standardized methodologies do not yet exist for defining trigger zones to
support evacuation response decision-making, specifically in the context of what communities
can accomplish during WUI fire conditions. Advances in computing power are improving the
capability to run complex evacuation and fire spread models. However, limitations and
uncertainties persist in current state-of-the-art models. Uncertainties are compounded with each
step when models are combined and linked to capture the entire evolution of the event more
completely. Weather, fire spread, impact of fire on roads, and evacuation are all linked, making
attempts to precisely account for and predict the individual components difficult. Not every
specific outcome can be evaluated; changes in evacuation conditions, including road blockages
from varying fire behavior, are very difficult to predict before an ignition occurs or to model in
real time. Additionally, human behavior and response is influenced by many factors, making
predictions of the outcome from complex events difficult. The overall complexities call for a
simplified general approach with a heavy emphasis on flexible and adaptive pre-planning.
Lessons learned from a detailed case study of the evacuations during the 2018 Camp Fire in
Butte County, CA [6] highlight critical concepts of WUI fire evacuations, including:
• the relationship between the time and location of ignition and the time the fire will impact
the community,

1
NIST TN 2262
August 2023

• the number of civilians that can become trapped in intense fire exposures (i.e., burnovers)
during evacuation,
• the widespread use of temporary refuge areas (TRAs) in urgent attempts to shelter from
dangerous fire exposures and also as an evacuation and traffic management tool, and
• the benefits of a pre-disaster evacuation drill that provides a training opportunity for first
responders to implement traffic management tools like contraflow and to coordinate with
relevant partner agencies.
This report expands on these ideas in two parts. First, technical challenges related to community
evacuations are discussed with respect to WUI fire event progression and life safety issues due to
fire exposures potentially encountered during evacuation. Then, a technical approach is outlined
that can be used to help develop a comprehensive evacuation strategy that addresses
spatiotemporal considerations and constraints associated with WUI community evacuations.
Planning and decision-making support concepts for the development of community notification
and evacuation plans are then presented.
This report is intended for existing small and intermediate-sized1 intermix communities and
isolated interface communities, although many concepts and incident considerations apply to
larger communities and urban interface areas as well. Specific emphasis is given to providing a
methodology to address fire events that impact a major portion of the community resulting in
limited or no options available for safe evacuation. This is particularly important for
communities that may not have the resources or expertise to conduct or evaluate a more complex
evacuation analysis. While additional research is needed to optimize the concepts discussed here,
the presented information can inform communities and help develop/improve community
sheltering and evacuation planning.
Throughout this report, specific examples from the Camp Fire are presented to provide context
of real events from a WUI fire [6, 7]. They are presented in the following format, with this first
example containing background about the incident.

1
defined by the authors as communities with a total population less than approximately 30 000

2
 NIST TN 2262
August 2023

Camp Fire Example 1. Introduction to the 2018 Camp Fire.

The 2018 Camp Fire in Butte County, California rapidly impacted the communities of Concow,
Paradise, and Magalia, triggering widespread evacuation of 40 000 people. The maps above show
the location of Butte County in California, the final fire perimeter, and the local area around
Paradise.
The fire was the most deadly and destructive fire in California history, resulting in 85 fatalities and
more than 18 000 destroyed structures. The Camp Fire ignited at approximately 06:20 off Camp
Creek Road near the small community of Pulga in the Feather River Canyon, northeast of Concow.
After immediately impacting Pulga, the fire spread southwest over a ridge, spotting and burning into
Concow by 07:30, 6.4 km (4 mi) away. By 08:00 spot fires were igniting in Paradise, an additional
6 km (3.75 mi) west of Concow. The fire front impacted eastern Paradise forty minutes later.
A post-fire case study was conducted, resulting in two primary reports to date: the first on the fire
progression timeline, fire behavior, and identified civilian burnover events [7], and the second on
life safety aspects including notification, evacuation, traffic, temporary refuge areas, rescues, and
fatalities (collectively, NETTRA) [6]. Various examples from the Camp Fire are introduced in this
report to provide recent real-world examples that illustrate some of the considerations and
challenges that are presented here.

3
NIST TN 2262
August 2023

This page intentionally left blank.

4
 NIST TN 2262
August 2023

2. Community Evacuation and Alternatives

This section highlights different aspects of community planning and response to WUI fires that
are important components of a comprehensive community response plan. Evacuation itself is a
key component, of course; however, alternatives are also an important consideration for
scenarios when evacuation is not possible. Such actions are commonly referred to as stay and
defend and shelter in place. The following sections present important life-safety considerations
and hazards of these approaches, including the concept of temporary refuge areas and areas of
last resort. Defensible space is a critical contributor to life safety in WUI fires, including for
evacuation and alternatives. However, defensible space alone is generally not a sufficient
solution.
Due to the dynamic nature of WUI fire evacuations, situational awareness is an important
component of the incident management. Traditional information updates come from sources
including radio reports from personnel in the field and 911 operators. Technology advancements
are increasing the amount of data potentially available, including live video feeds showing traffic
conditions, real-time observation of the current fire location via remote sensing, and social
media. There is a need to develop new systems to effectively gather, interpret, and use these
information streams. Planning for the dissemination of this information is an important
component of using these new technologies.
Computer models of fire spread and evacuations are becoming more sophisticated, with potential
to support advance planning and scenario development. This section includes a basic
introduction to models and references several recent research efforts to combine or couple fire
and evacuation models.

2.1. Evacuation
The most common life-safety response policy to WUI fires is evacuation of the threatened
population located in the anticipated path of the fire. This is the standard response implemented
by officials in the U.S. and Canada, and is the preferred action in Australia [8]. Successful
evacuation away from the hazard is the surest way to avoid exposure to flames, embers, heat, or
smoke.
There is limited standardization for both the planning and execution of evacuations for WUI
fires, although generalized conceptual guidance is available for all-hazards planning. The Federal
Emergency Management Agency (FEMA) has a document [9] that outlines planning
considerations. The document briefly touches on important aspects ranging from accessibility
and medical facilities to accommodating pets to evacuation shelters out of the hazard area. The
Federal Highway Administration (FHWA) also has a primer document describing general
planning and operations concepts for no-notice evacuations such as those due to wildfires [10].
While the document highlights the transportation component of evacuations, the report includes
additional background on no-notice incidents and general considerations for planning and
operations.
Evacuations are often initiated by emergency officials who issue notifications and instructions to
the affected populations using various tools such as reverse-911, the Integrated Public Alert &
Warning System (IPAWS), and the internet. Sirens and door-to-door notifications may also be
utilized. If time allows, evacuations may be conducted in phases, starting by notifying and

5
NIST TN 2262
August 2023

evacuating areas of the community that may be affected first. This is often facilitated by
establishing predetermined evacuation zones. In the U.S., most people evacuate using personal
vehicles. Communities may include provisions for mass evacuation transportation for individuals
without access or ability to utilize personal vehicles, whether via public transportation or more
specific arrangements for evacuation procedures.
A critical public safety aspect of evacuations is that they are done early, when conditions are
favorable before the fire impacts the community. Past fires have shown that attempting
evacuation when there is not enough time to do so safely can result in evacuees being overrun by
fire before reaching safety. Late evacuation can be generally defined as “an evacuation that puts
an individual, or group of individuals, at risk of encountering dangers associated with the
passage of a [wildfire]” [11], and may occur due to fast-moving fire, extended duration of
evacuation procedures, or delayed decision-making. Deaths due to fire exposures experienced
during evacuations have occurred in fires across the globe, with examples in the U.S. [12],
Australia [11, 13], and Europe [14-16]. The hazard associated with smoke exposure and
inhalation is complex and more widespread. Beyond acute exposures in the fire, smoke from
wildfires and WUI fires can travel for tens or even hundreds of kilometers downwind and can
impact civilians with preexisting respiratory conditions as well as the general population.
Adding to the challenges of timely evacuation is that, in many WUI fire events, evacuations are
initiated with little notice before the onset of hazardous conditions. While fire weather alerts,
such as Red Flag Warnings in the U.S., indicate conditions conducive to ignitions and rapid fire
spread [17, 18], these are only generalized pre-event advisories. Initial orders for evacuation are
typically reactionary, issued after a fire ignites. Alternatively, some individuals, or even entire
communities, may consider evacuation before the most hazardous fire weather conditions are
forecast to begin. This approach to “leave early” is recommended in Australia for days with
Extreme or Catastrophic Fire Danger Ratings [4, 19]. However, this approach may not be
feasible in all areas or for a large fraction of the population due to various socioeconomic
burdens imposed by evacuations.
The no-notice nature of many WUI fire evacuation events requires significant pre-planning [10].
During an incident, emergency officials will likely have to make decisions with incomplete
information and without time to develop a course of action coordinated with all the response
agencies involved. Development of a comprehensive evacuation/sheltering plan before an event
can increase the likelihood of an effective response, enhancing the safety of community
members. Even in the event of a catastrophic incident that exceeds the capacity of the written
plan, the existence of a plan can provide a foundation from which an appropriate response can be
enacted.

2.2. Evacuation Alternatives

In response to a threatening WUI fire, residents must decide to take protective action or not,
generally involving evacuation or sheltering. While early evacuation removes citizens from the
hazards presented by WUI fires, alternatives to evacuation may be beneficial in some
circumstances, either as contingency plans for individual citizens or for specific community/fire
scenarios [3, 20-22]. The variability of ignition locations, fire spread, and community specifics
dictate that evacuation cannot be implemented as a “one solution fits all” approach. Individuals
(and by extension, communities) must have multiple response plans for various scenarios or

6
 NIST TN 2262
August 2023

eventualities depending on the actual fire event progression. Fatalities have occurred in instances
where a person’s primary plan to evacuate was not possible or successful and they were not
prepared for an alternative action [23].
In the U.S., wildfire evacuations are predominantly achieved by use of personal vehicles, often at
the rate of two vehicles per household [24]. This can lead to significant traffic delays, further
increasing time required to evacuate and potentially exposing evacuees to the advancing fire.
Additionally, background traffic and various intermediate trips [25] that civilians may make in
response to a fire, including retrieving or meeting with family members [26, 27] or returning
home [28] to gather belongings, may further extend evacuation times. In some communities, the
evacuation of the residents to safety may not be a feasible response when evacuation capacity is
considered against the anticipated fire spread rate, and alternatives should be considered [3, 20,
29].
Two often-discussed alternatives to evacuation include “stay and defend” and “shelter-in-place”
(SIP) frameworks [22, 30]. There is some overlap between the approaches, however SIP is
generally regarded as more passive compared to the actions required by a stay and defend
approach [30, 31]. Sheltering choices may include staying at home (often including attempts to
protect the home) or heading to a refuge area or other safer space [20]. How people ultimately
decide what to do is an important component of ongoing research [32-34].
It is important to note that these approaches are not universal and may not work in all scenarios
[20, 22, 23, 30, 35]. Decisions about whether a home is defensible, whether a civilian is
physically able and prepared to defend a home, and whether it is safe to evacuate are nontrivial,
dynamic, and variable from individual to individual and location to location [20]. These
considerations cause difficulty for widespread implementation, including who determines (and
how) what conditions, locations, and personal preparedness are appropriate for SIP.
Consideration of alternatives to a complete evacuation are necessary as part of contingency
planning, primarily in response to situations where there is not enough time for safe evacuation
to occur. These are referred to as dire scenarios [36]. The idea of shelter-in-place is typically
employed under three circumstances: when it is the only option (i.e., entrapment), when
evacuating would lead to entrapment, or as a pre-determined course of action for refuge or to
protect the home [20]. Depending on the size of a community and the available evacuation
capacity, community response/evacuation plans could implement a hybrid of evacuation and SIP
[20]. For this to be a reasonable and safe path forward, additional research is required to identify
and design safe areas for community refuge, and to standardize an approach for assessing
whether a structure is defensible.
Comparisons of exposures experienced during evacuation should be considered with respect to
potential exposures in other places of refuge. In many WUI communities, the fuel density (both
vegetative and built environment) leads to high exposure hazards, making it unsafe to seek
refuge there, necessitating evacuation. Even if there is a risk of exposure during evacuation, this
may be less hazardous than taking refuge in the available locations. However, there may be
instances where combinations of roadside fuels, evacuation routes, or heavy traffic delays
increase the hazard to evacuees, and seeking refuge may be a better option.

7
NIST TN 2262
August 2023

2.2.1. Stay and Defend

“Stay and defend” refers to an alternative to evacuation in which residents prepare and intend to
remain at their property to defend against the fire and ember exposures, with the goal of saving
structures and lives. In Australia, stay and defend was gradually adopted into policy during the
1960s as a civil defense-like approach to public safety as development expanded into fire-prone
rural areas beyond the capacity of the existing fire service [37]. Subsequent disastrous bushfires
resulted in an increasing emphasis on individual responsibility and decision-making, with the
stay and defend approach evolving into “Stay or Go” after the 1983 Ash Wednesday bushfires,
and into “Prepare Act Survive” after the 2009 Black Saturday fires.
While not advocated by emergency officials in the U.S., some residents invariably decide to stay
during a wildfire and defend their property and surrounding neighboring properties.
The “stay and defend” approach can work under certain scenarios where there is an appropriate
balance among actual fire exposure levels, structure/parcel ignition resistance, and effective
defensive actions. Proponents of the stay and defend approach acknowledge the important
concept of suitability for staying and defending. Some structures or locations may not be
defensible [29]. There are many places in the interface and intermix where the density of
structures and other WUI fuels render a parcel indefensible, even by trained and equipped
firefighters. Building construction and local attributes of terrain and weather further influence
defensibility.
Significant preparations are necessary to implement this approach [38]; not all civilians will be
capable of defending their property, especially against high intensity exposures. However, this
crucial aspect of the “stay and defend” mantra was found to be largely misunderstood and
oversimplified by the public. Roughly one third of the 173 fatalities in the 2009 Black Saturday
fires occurred at homes that were not defensible [31, 39], including conditions where structures
were located close to significant wildland vegetation.
Furthermore, just because a structure may be deemed defensible, that does not address the need
for the resident to have appropriate training, fitness, and equipment to successfully defend the
structure [20]. The stay and defend approach essentially requires a civilian to undertake a skilled
and physically demanding professional job—firefighting. Miscommunication or
misunderstanding of the potential risks and exposure conditions that may be encountered can
result in inadequate preparation, which greatly increases the risk to life safety [31]. It should be
clear to residents that staying within the path of a WUI fire, whether planned or unplanned, can
have serious or fatal consequences. In past fires, even individuals who were well prepared to
defend their properties (including those with fire service training, firefighting equipment and
water supply, personal protective equipment [PPE], and defensible space) have died, indicating
the difficulty of this approach [23, 31]. Decisions to abandon plans to defend one’s structure (for
example, after determining the situation is beyond the capability of the resident to defend) and
then evacuating late can lead to some of the most dangerous scenarios in the middle of peak fire
activity.
While there are successful cases of civilians defending their properties, there has been no
systematic study to assess how many fatalities are associated specifically with stay and defend
practices or intentions. With some exceptions, e.g. Ref. [31], the intentions of those making this
decision are unknown. Additionally, information about harmful effects of smoke inhalation and
other delayed significant health issues after staying behind and defending residential properties is

8
 NIST TN 2262
August 2023

primarily anecdotal. No comprehensive analysis has yet been conducted to assess the
effectiveness of defensive actions by civilians, which would be especially useful within the
context of local conditions (fuels, topography, and local weather), quantification of fire and
ember exposures, and life safety.
All of this leads to important aspects of this approach—what is defensible, and what training and
equipment are required to do so effectively and safely?
Conceptually, defensibility includes exposure levels, structure construction and ignition
resistance, and the capacity to defend [38]. The level of potential exposures vary so dramatically
that it is difficult to summarize or know the conditions that residents will ultimately encounter.
Exposures can range from minimal fire and small ember fluxes to direct flame impingement
from a neighboring home that is fully involved. Even experienced firefighters can be surprised
by the intensity of the WUI fire front, especially in catastrophic events.
The capacity to defend is subject to personal capacity (physical and mental), equipment, and
local conditions of the property [38]. Physical fitness, water supply, and equipment availability
require long-term pre-planning and financial investment. In rural locations, exposures may be
more readily managed, as space may be available for fuel relocation, reduction, or removal.
Hazard reduction by individual property owners, even when implemented at a parcel level, does
not address overall exposures from adjacent parcels [40], which may be significant in intermix or
interface communities with increased fuel density.
Differences between a civilian, including all but the most well-prepared individuals, and
firefighters are listed in Table 1. From basics like appropriate firefighting training, experience,
and physical fitness, to details such as firefighting equipment (including tools, water supply, and
PPE) and situational awareness, there is a large gap between the two groups. The differences
point to the high potential for civilians to encounter situations that compromise their life safety.
Work is needed to develop explicit requirements that can be used by AHJs and civilians to
prepare residences, infrastructure, and communities for the safe implementation of stay and
defend policies and regulations [41].
When applied to today’s WUI, particularly in contrast to rural ranch settings, further challenges
arise, such as whether the property owner is present at home when the fire ignites, or whether
they will have to return toward or into the fire to get home and enact their defense plan.
If the expectation of stay and defend is to help put out spot fires, this is a very different exposure
level than an intense fire front with heavy ember shower and flame exposures from parcel and
neighboring parcel fuels, or even suppressing the structure after it potentially ignites. Residents
were surprised by the lengthy amount of time it took the fire front to pass in the 2009 Victoria
fires [39]. This can be the case in intermix and interface fires where the types of fuels present can
continue to burn for long durations, generate embers, and continually ignite additional fuels for a
long time, requiring continuous defensive and suppression efforts, likely in unhealthy conditions.

9
NIST TN 2262
August 2023

Table 1. Differences between private civilians and firefighters.

Preparedness/Response Attribute Typical Civilian Firefighter

Training and maintenance of proficiency Limited Mandatory training; experience
of WUI/wildland firefighting strategies gained through practice and
and tactics annual recertification
Physical fitness Variable Required, tested
Equipment Limited Available, maintained, tested,
and specialized
Standalone water supply (independent of Variable Available on apparatus and
community infrastructure) locally accessible sources
PPE and safety training, including Likely inadequate Standard and required
wildland fire shelter use
Situational awareness Limited to media, internet, and Fully integrated in ICS with an
radio scanners, and may be incident action plan (IAP)
dependent on electrical power
supply
Lookouts, Communication, Escape Unlikely Yes
Routes, and Safety Zones (LCES)
Operational support No Yes

2.2.2. Shelter-in-Place
Shelter-in-place is a protective action where individuals quickly shelter indoors in response to an
emergency. This action is accomplished faster than an evacuation because the person is already
located in or near the sheltering location and does not require travel. It is a common response for
other no-notice hazards, such as tornadoes or chemical releases. In special cases, shelter-in-place
may be a response during interior structure fires (e.g., high-rises, hospitals), but this action
depends on the specific design of engineered buildings, the fire scenario, and fire protection
systems (i.e., passive barriers and active suppression).
Currently, FEMA does not include wildfire in its guidance for SIP that includes other
emergencies including earthquakes, floods, hurricanes, and tornados [42, 43]; evacuation is the
recommended protective action [44]. Due to combustible construction materials, fuel
accumulation and agglomeration, high structure density, and existing ignition vulnerabilities
[40], many communities are not generally suitable for stay and defend, and nearly all are
unsuitable for a more passive SIP approach.
Shelter-in-place requires homes to stand alone through a WUI fire event without intervention. In
nearly all cases, current structures are not built to a standard where passive sheltering is a reliable
way to survive. Even in locations where fuels have been removed to avoid direct flame contact
and high radiative exposures, the threat of ignition from embers is very high. Using SIP as a last
resort, where a structure may provide temporary refuge from exterior exposures before it ignites,
has also been done in the past, sometimes as part of a stay and defend strategy. This is an
important option to keep in mind; however, significant risks are involved with sheltering inside a
burning building. Conditions can rapidly deteriorate once a home ignites, forcing occupants to

10
 NIST TN 2262
August 2023

escape outside where the initial exposures they were trying to avoid may still be hazardous.
Injuries and deaths of even experienced fire service personnel have occurred when determining
the transition point between sheltering inside and exiting into the outdoor fire [23].
Few communities in the U.S. have adopted a SIP approach for WUI fires. The community of
Rancho Santa Fe in San Diego County, California is often referenced as an example [3, 20, 30,
35]. However, even within the community where this approach is implemented there is variable
interpretation of the shelter-in-place terminology. Officials disagree whether shelter-in-place
means a) building construction and pre-fire preparations alone versus b) a default approach for
residents to remain at home during a wildfire. While the community was built to higher building
standards2 to resist ignitions, the current official guidance from the Rancho Santa Fe Fire
Protection District is to first evacuate if safely possible [45].

2.2.3. Areas of Last Resort and Temporary Refuge Areas

In some cases, such as in fast-moving fire scenarios, there may not be enough time to return to
one’s property to carry out a stay and defend or shelter-in-place plan, let alone evacuate. In other
scenarios, plans to stay and defend or shelter in place may fail, or evacuees may be overcome by
fire during evacuation. These outcomes have been observed in several past fires; the most
detailed analysis followed the 2009 Victoria bushfires in Australia when a royal commission
investigated the devastating fires [39].
One result of the Commission’s findings was that people must have contingency plans and
alternatives of last resort if their plan to leave early or stay and defend fails. Subsequently,
Australians formalized the Neighborhood Safer Place [46-48] and Community Fire Refuge [49]
ideas. Similar concepts have been implemented by local communities in the U.S., including the
community of Concow, CA involved in the 2018 Camp Fire, with pre-designated Wild Fire
Safety Zones and public assembly areas. While comparatively safer than surrounding areas, these
locations do not guarantee safety. It is not guaranteed that access to these locations will be
possible, or that they will be able to accommodate everybody seeking refuge there. Further, there
is no guarantee of the presence of first responders to facilitate or protect those taking refuge.
Similarly, open spaces (i.e., parks, parking lots) or select structures may be chosen in the
moment as temporary refuge areas (TRAs). Many of the same challenges regarding shelter-in-
place also apply to areas of last resort or TRAs.

2.3. Defensible Space

This section provides context on the implications of the concept of defensible space on structure
accessibility and defensibility as they relate to evacuation and rescue operations.
Defensible space is a structure-centric approach where fuels are managed/removed from the
surrounding parcel with differing specifications at increasing distances from the structure, often a
primary residential structure. The defensible space concept was conceived in large part to
provide a safer space for first responders to defend a structure and has been implemented for
decades. For example, it was first added to California state law in 1965 [50]. Several
2
While comparatively higher than other existing building codes, exact requirements are unspecified, as is the degree to which compliance of
these standards has been maintained years later. Many vulnerabilities remain even after WUI fire provisions in existing building codes are
implemented [40].

11
NIST TN 2262
August 2023

amendments have been made in the years since, with the most recent in 2021. The “home
ignition zone” and the Firewise program [51] were introduced in 1986 to educate homeowners
and communities about fuel reduction and vegetation maintenance within 200 ft (60 m) of
structures.
Two primary benefits of defensive space fuel reduction are:
1. exposure reduction to the structure and surrounding parcel area, and
2. a safer environment for trained firefighters to defend the residence.
While defensible space is an important component of structure survival and stay and defend or
shelter-in-place responses, defensible space alone is not sufficient to support these actions. A few
limitations of defensible space should be considered. Defensible space, although conceptually
presented as a single approach, is not defined by a one-size-fits-all distance or fuel treatment for
every structure; it requires adjustments based on the footprint of the structure and surrounding
fuels and topography [52]. One prominent limitation is that there are often discontinuities on
defensible space at property boundaries, and home ignition zones often overlap or extend beyond
property lines [29, 53]. Another is that, within a community, varying degrees of defensible space
may be implemented or maintained by property owners, and regulations and enforcement may
differ across jurisdictions.
Parcel-to-parcel exposures and fuel agglomeration can have significant impact on structure
survivability and access for evacuation and rescue operations [40]. Fire and ember exposures in
WUI settings can come from burning buildings, vegetation, and other parcel-level combustibles.
The parcel-to-parcel fire and ember exposures in high density construction (i.e., structures
separated by less than 25 ft [7.6 m]) can result in significant life safety hazards and should be
considered when planning for the evacuation of civilians; special attention should be placed on
the evacuation of civilians with reduced mobility. High exposures can also be encountered in
moderate density intermix settings with limited fire history, where fire exposures can impact a
residence from adjacent parcels with high parcel fuel loadings.
Wildland vegetation can also have a significant impact on residential properties, resident
evacuation, and first responder access. Intense fire exposures from wildland fuels can occur on
parcels where there is limited setback from property lines for the creation of effective defensible
space and no mechanisms are in place for a resident to manage the fuel loading beyond their
property line.

12
 NIST TN 2262
August 2023

Camp Fire Example 2. Defensible space and exposures from neighboring parcels.

Exposures from neighboring parcels must be accounted for when assessing the defensibility of a
property. The fully involved parcel (including structure, vehicle, and vegetation) seen above
illustrates the very high fire exposures that can be generated during WUI fires. Fully involved fuels
with flame lengths greater than 6 m (20 ft), as in the image above, would be difficult to contain even
with several firefighting apparatus and cannot be contained by defensive actions by residents.
In this scenario, the structure separation distance (SSD) was 13 m (43 ft) from the burning home
shown in the image to the neighboring structure. The structure to property line distance (SPLD) was
8 m (26 ft). Defensible space may be difficult to implement in moderate and high-density
communities where significant fire exposures can originate from neighboring parcels and structures
are spaced even closer than in this example.

Defensible space is an important component of a stay and defend or shelter-in-place approach,

but the life safety risks can be very high. In the context of evacuation and rescues, there is
significant value in managing the fuels around a residence. There is a fundamental difference
between designing, building, and maintaining a structure to meet existing minimum code
requirements and making a structure standalone with the necessary hardening to address both the
expected exposures from fire (flames) and embers [40]. Many existing building codes for
residential construction in the WUI are designed to work together with defensible space to allow
first responders safe access to defend the property. Defensible space can provide significant
value during assistance and rescue operations by creating a safer environment for
assisted/rescued civilians and first responders. Defensible space practices are therefore important
for residences, commercial facilities, and critical care facilities where additional time may be
necessary to evacuate mobility impaired residents.
Beyond the impacts on firefighter access and rescues from residences, fire exposures can pose a
hazard to evacuating civilians by causing the local closure of egress arteries. Closure of a main
roadway during community evacuation will impact civilian egress and first responder access
even if neither group becomes entrapped in a burnover event. While defensible space is often
focused on protecting structures, similar fuel reductions along egress arteries play a critical role
in limiting the potential exposures encountered during evacuations.

13
NIST TN 2262
August 2023

2.4. WUI Fire and Evacuation Modeling

WUI fires are community and incident specific. Extrapolation and generalization of results from
one incident to another community or scenario are limited due to sensitivity to local conditions
[54], including the fire scenario, community demographics, evacuation capacity and egress
routes, and many other factors. Previous case studies and generalized findings can identify
possible scenarios and challenges to look out for, but assessing how an individual community
can and will respond to those events must be focused on the local community in question. This
provides an opportunity for use of computer modeling as a tool to support decision-making and
advance planning by public safety officials by providing insight into potential scenarios and
outcomes.
Evacuation planning lies at the intersection between fire spread and evacuation modeling. To
provide a comprehensive approach, an integrated, coupled set of models is desired. A
computational tool that combines the relevant components of an evacuation event, namely the
fire, pedestrian movement, and vehicular traffic, would support evacuation plan development and
the decision-making of emergency officials [55]. Recently, Ronchi et al. have developed a
modular framework (called WUI-NITY) [56, 57] that integrates these three layers of modeling.
Through this framework, any of several models might be selected to accomplish prediction of
each fire or evacuation component. Ronchi et al. [55] present a broad comparison and evaluation
of many of the existing models in each of these three areas. Furthermore, they have developed a
list of important questions users/practitioners need to consider regarding the performance of
different model components [55, 58].
One major challenge is the amount of complexity in each layer of the combined modeling. For
the fire component these include aspects of fire spread uncertainty and factors like weather
forecasting and fire-weather coupling, spot fire ignitions, and treatment of structural/community
fuel types in current models. For the evacuation component these include aspects of evacuee
decision-making, departure timing, route choice, and destination choice [59]. Incorporation of
human behavior and decision-making into the models is also an ongoing challenge. Several
studies of human behavior during WUI fires have been done (see an overview [59] and detailed
reviews of the field [33, 60]). These complexities in predicting fire spread and the evacuation
process exist even before the fire impacts the evacuation routes or community; additional
complications can arise that are not able to be specifically predicted due to their stochastic
nature. Pre-planning and scenario gaming becomes key to understanding response options and
the potential consequences. Evacuation models that are uncoupled from the fire spread models,
specifically addressing scenarios where evacuation is not affected by fire, may provide valuable
insight to traffic management and the time required for evacuation in the best conditions.

2.5. Evacuation Trigger Models

The decision to order evacuations must be made by fire officials in the face of complexities and
uncertainties in evacuations and fire spread. Decision points are based on certain fire behavior or
weather events or extent of fire progression, often linked to physical landmarks to aid in
identification. When situation-specific criteria are met, response actions or evacuations are
triggered [61]. Standardized guidance for determining triggers is limited or non-existent; it is
often left to the experience and judgment of the incident commander or other emergency official
to consider required evacuation time against estimates of fire arrival time.

14
 NIST TN 2262
August 2023

Several models are in development to link components of fire and evacuation modeling to
systematically identify trigger locations and buffers by linking fire and evacuation models. An
early approach was introduced by Cova et al. [61] in 2005, inspired by a GIS-based decision
support tool for hurricanes (HURREVAC [62]). The model, named WUIVAC, calculates fire
spread predictions for an area of interest and determines buffers around the community
corresponding to the different time periods over which the fire is predicted to spread into the
community. These fire arrival times are compared to an estimated evacuation time, determined
independently, to identify the buffer location(s) where an evacuation should be initiated if the
fire reaches that specified location around the community. The model has been applied in a
research capacity [63, 64] to Julian, CA, a community in San Diego County that was affected by
the 2003 Cedar Fire. A traffic simulation model has recently been implemented in conjunction
with the WUIVAC model to further develop the evacuation time component of the trigger buffer
[65].
More recently, Mitchell et al. [66] explored defining evacuation trigger buffers using a method
they called PERIL, developed using the WUI-NITY model framework [56, 57, 67]. To
demonstrate the capabilities of this approach, the model was applied to two communities: one in
the United Kingdom and one in Colorado. Kalogeropoulos et al. [68] advanced PERIL into k-
PERIL. The new implementation generates a set of trigger buffers by stochastically changing the
input conditions to a range of user specified values. The resulting set of trigger buffers indicates
a confidence interval reflecting the variability and sensitivity of the wildfire/evacuation
simulations. Sections of the boundary that exhibit more variable fire behavior and higher
uncertainty result in a wider range of trigger buffers.
Both the WUIVAC and k-PERIL tools are promising research efforts that could provide a means
to apply many of the evacuation planning concepts discussed in this report. This approach has
not yet been implemented by communities in practice and is further discussed in this report in
support of expanding options for community evacuation plans. However, when using modeling
tools to plan and predict different aspects of community evacuation, model outputs need to be
assessed and interpreted by subject matter experts of the tools used to understand the limitations
and uncertainties associated with the predicted outcomes.

2.6. Evacuation Triangle

The evacuation triangle presented in Fig. 1 represents the connectivity among fire–evacuation
trigger zones (ETZs), available time before the fire reaches the community (or associated egress
routes depending on the geography and scenario), and evacuation decisions. The concepts of this
evacuation triangle will be useful before a fire, during the design stage of a community
evacuation plan, and during an actual fire event to link together these three key aspects of
evacuation. The concept of the triangle is intended to be continuously evaluated during a WUI
fire event, indicated by the arrows; all three aspects of the triangle are dynamic. Changing
conditions in fire progression and available versus required evacuation time (i.e., fire–evacuation
trigger zones) will influence evacuation decisions. It should be noted that, unless the fire is
contained or slows down, the available time typically only decreases as fire approaches a
community.

15
NIST TN 2262
August 2023

Fig. 1. Evacuation triangle illustrating connectivity among evacuation trigger zones, available time before
fire reaches community, and evacuation decisions.

16
 NIST TN 2262
August 2023

3. Technical Challenges of Community Evacuations from WUI Fire

WUI fires have important differences distinguishing them from other disasters. These differences
have implications in managing the hazard of WUI fires, some of which can directly affect
evacuations. This section presents challenges regarding community evacuations to be considered
when planning community response to WUI fires.
Modeling of fire spread and evacuation may be able to assist with planning and decision-making
by emergency management officials. However, a thorough understanding of the limitations and
uncertainties of the modeling tool is required to adequately interpret the results.
A list of technical challenges is summarized at the end of this section.

3.1. Distinguishing Characteristics of WUI Fire Disasters

Several key aspects distinguish WUI fires from other natural disasters. Some characteristics that
distinguish WUI fires from other natural events that can lead to evacuation are presented in
Table 2. Each natural disaster included in this table can result in damage and destruction to a
community and infrastructure (including egress arteries). However, WUI fires are unique in that
the community itself is the fuel that propagates and, in many cases, intensifies the disaster. The
extent of disaster propagation in WUI fires is driven by the complex interactions of the fuels,
weather (wind, humidity), terrain, and defensive actions.

Table 2. Characteristics of WUI fires compared to other selected natural disasters.

Characteristic Hurricane Flood Tornado Earthquake WUI Fire

Built environment adds energy to No Yes a No No Yes
fuel event
Defensive actions during event No No No No Yes
change outcome b
Event energy can be managed No No c No No Yes
beforehand
Event starts other similar events No No No Yes Yes

Notification period Days Variable d Minutes None Variable d

Extent of evacuation Region Community/ Shelter-in- Shelter-in- Community/

Region place place City/Region
Building construction or sheltering Yes Yes Yes Yes Limited
standards
a
infrastructure failure; dams, levee systems (not individual buildings)
b
including residential and commercial structures or other infrastructure
c
amount of precipitation cannot be controlled; flood water can potentially be managed
d
minutes to hours to days

17
NIST TN 2262
August 2023

3.1.1. Defensive Actions Affect WUI Fire Outcomes

In WUI fires, the assets that are being protected (wildland vegetation and structures) are also the
fuel and energy sources that drive and often intensify the disaster. This has significant
implications on how WUI fire hazards can be mitigated and how WUI incidents can be
approached in the context of evacuation and disaster response. First responders (specifically
firefighters) can directly affect the outcome of WUI fires—defensive actions during the event
may alter the event progression and ultimately stop the fire spread through the community and
contain the fire in most cases. Therefore, it is critical that the specific defensive actions
performed in the community and their impacts are accounted for in the context of exposure
reduction as well as traffic management and rescue operations. This is different from the
approach to other disasters in which emphasis is placed on pre-disaster evacuation support or
rescues during and after the event.

3.1.2. The Asset is the Fuel

The fuel and asset being one and the same means that efforts to protect the community
beforehand through fuel reduction and structure hardening can affect the energy available to the
disaster, unlike the other events listed in Table 2. Wildland fires can be beneficial to the
ecosystem and are a recurring natural event. Attempts to manage wildland fire intensity and
frequency have been undertaken by federal, state, and local landowners for decades. The
transition from wildland fires to WUI fires can be mitigated by fuel treatments around and within
communities and by hardening structures and parcels throughout the entire community. These
actions can decrease ignition potential and reduce the intensity of a fire in the community,
potentially reducing losses from WUI disasters. Pre-fire hazard mitigation and its effects on
exposure intensity expand the range of potential outcomes as a wildfire impacts a community,
spanning from a reduction in intensity and interruption of fire spread to complete destruction of
the community and significant loss of life.

3.1.3. Fires Can Start New Fires

An additional distinction among natural disasters is the potential for WUI fires to initiate
additional incidents themselves. Spot fires ignited from firebrands, which can be transported
kilometers ahead of the main fire front, effectively accelerate the anticipated or previously
observed fire spread rate. These spot fires can generate new hazards in unexpected locations
deeper within the community, dynamically expanding evacuation requirements and potentially
impacting previously uncompromised evacuation routes.

3.1.4. Notification Times Range from Minutes to Days

On the continuum of notification periods, WUI fires are generally positioned between
earthquakes and tornados, which give shorter notice, and flooding and hurricanes, which
generally provide more lead time. In some instances, WUI fires may be characterized by a no-
notice timeline, where an ignition occurs close enough to a community that immediate action is
required, more akin to tornado timelines. In other cases, a fire burning in the wildlands may not
pose an immediate risk and provide additional lead time for evacuations.

18
 NIST TN 2262
August 2023

3.1.5. WUI Fires Have Limited Advanced Warning and Locally Variable Intensity
In the U.S., large-scale community and regional evacuations are commonly implemented when
hurricanes (more generally, tropical cyclones) impact populated coastal areas. Many coastal
states and communities have well-established evacuation plans and zones to guide preparation
and response for varying levels of storm intensities and impacts. Some notable components
associated with community evacuations ahead of a hurricane provide a contrast with WUI fires.
To inform the public about an approaching hurricane, a system of watches and warnings3 has
been established and implemented by the National Weather Service (NWS) [69, 70]. A watch is
issued for a specific area when hurricane conditions are a possibility within 48 hours. A warning
is issued when a hurricane is expected in an area within 36 hours. This is distinctly different from
WUI fire disasters in several ways:
1. While Fire Weather Watches and Red Flag Warnings are issued ahead of qualifying
weather days, these general warnings do not indicate that there is a fire. The actual
ignition location, direction of fire spread, and fire intensity are not determined until after
an ignition.
2. Hurricane watches and warnings are issued based on a standardized system for predicting
the track of the hurricane, including confidence and uncertainty analyses.
3. The lead time of hurricanes is typically measured in days, not hours or minutes.
4. There is a national decision support tool with clear temporal thresholds in place for
triggering evacuations for hurricanes (HURREVAC [62]).
Once a hurricane moves over land the storm decreases in intensity as the energy source (warm
ocean water) stops feeding the hurricane. In contrast, WUI fire intensity and fire spread is solely
dependent on local fuels, including the community itself, and therefore can decrease or increase
in intensity and/or spread rate. Fire spread through a community is not characterized by
widespread steady-state propagation; community fuels can generate very high local exposures.
Furthermore, fire intensity can vary dramatically based on local fuels, wind, and topographic
characteristics. Spot fire ignitions ahead of the main fire front can start new events or accelerate
the timeline of fire spread. In contrast, a hurricane does not start new hurricanes, and the
construction characteristics of the community do not affect the intensity of the hurricane. This is
also true of tornadoes, earthquakes, and floods. These two fundamental differences have
repercussions on the design and execution of evacuation plans for WUI fires as compared to
other disasters.

3.1.6. No Standardized WUI Fire Shelters

The lack of reliable and effective shelter also makes WUI fires different from other natural
disasters. There is no federal or state guidance available for the placement, construction, and
maintenance of shelters than can be used to protect civilians during WUI fires. The lack of
reliable community fire shelter design information, in addition to the concerns discussed in
Sec. 2.2 regarding stay and defend and shelter-in-place, make the use of that approach difficult at
this time. Without means to safely design and properly place and maintain WUI fire shelters in

3
The NWS watch and warning system also encompasses weather impacts beyond tropical cyclones.

19
NIST TN 2262
August 2023

communities, authorities having jurisdiction (AHJs) have fewer options for addressing the life
safety of civilians during WUI events compared to some of the other natural disasters like
tornadoes and hurricanes.
Information is available to the public for how to generally prepare for disasters. Publications and
websites like https://www.ready.gov/evacuation can provide useful information to the public for
general evacuation information and are designed to work together with specific community and
regional guidance; however, the differences highlighted above indicate a need to treat WUI fire
disaster evacuations differently.

3.1.7. Wildfires Are Relatively Frequent

Finally, AHJs must contend with the high frequency of wildland and WUI fires that threaten
communities relative to the frequency of some of the other natural disasters. The National
Interagency Fire Center (NIFC) reported a total of 58 733 wildland fires in 2021, the year with
the fifth fewest ignitions on record. This number includes all wildland fires, not just those that
reached a WUI community, and it does not necessarily include all fires that started within
communities. A closer look at wildfire statistics can provide insight into how many of this large
number of ignitions may develop into significant events that can impact communities.
Overall numbers and statistics of evacuations for wildland fires are not generally tracked or
available. One study by Beverly and Bothwell [71] identified 547 wildfire evacuation events in
Canada between 1980 and 2007, averaging 20 per year with a high count of 53 events in one
year. Updated data show approximately the same number of wildfire evacuation events (566)
occurring in Canada between 2008 and 2018 [72], averaging 51 per year. Beverly and Bothwell
point out that there are fundamental differences between Canada and locations such as
California, where the overlap between populated areas and fire occurrence is more widespread.
Based on data from the California Department of Forestry and Fire Protection (CAL FIRE)
compiled in Appendix A, an average of 34 fires per year led to reported structural losses in the
six-year period 2017–2022 in California, which may conservatively serve as a proxy for the
minimum number of evacuations in California. At least 11 large-scale fires in California between
2017 and 2019 required evacuation of more than 10 000 people [24].
The number of incidents that require evacuation for wildfires is thus greater than for hurricanes.
Historically, the upper bound for the number of Atlantic hurricanes in a season is 30 named
storms (2020), with a 30-year average of 15 [73]. Not all hurricanes will impact land and require
evacuation.

3.2. WUI Fire Evacuations

Evacuations encounter inherent constraints in the time available to evacuate and in the potential
for compromised evacuation infrastructure. Planning for a community-wide evacuation is
fundamentally different than that required for an evacuation from an indoor fire in a commercial
building, or for evacuation of a corporate or university campus in response to non-WUI fire
events such as a chemical spill or active shooter incident. Primary differences include:
1. The authority for mandatory evacuation;
2. Standards for exits, evacuation pathways, and building capacity;

20
 NIST TN 2262
August 2023

3. How far the building or campus occupants must travel and whether they need a vehicle to
evacuate;
4. The impact of evacuation on the surrounding local area and community; and
5. The extent of coordination required with the community.

Table 3 summarizes the differences listed above. The lack of standardization for community-
scale evacuations affects the design of plans for these events. This is further reinforced by the
limited accreditation infrastructure for the technical skills and tools necessary for the design of
such plans. An ongoing international effort, led by the Society of Fire Protection Engineers
(SFPE) Foundation, is working to address this issue for the WUI [74].

Table 3. Comparison of evacuations for individual buildings vs. a campus vs. a larger community.

Community or
Aspect Commercial building a Campus part of community
Codes used in design and Yes Yes No
construction addressing
fire
Regional coordination No No Yes
needed

Safety zone location Outside assembly area, Variable/undetermined Can be miles away and
typically in parking lot will likely require travel
in vehicle(s)
Evacuation impact on Low Low to Moderate High
overall community

Community road capacity Low Variable High

impact on evacuation

Potential impact of Parcel only or Local Local/Community Community to regional

evacuees
a
selected for comparison due to specific code requirements for evacuation and construction compared to single-
family residential type buildings

Evacuations in response to wildland or WUI fires need to address the life safety needs of both
civilians and first responders. This also applies to other disasters; however, the building
hardening standards currently available are typically not sufficient for buildings to withstand
potential WUI fire and ember exposures [40]. Therefore, buildings cannot be treated as de facto
shelters. Consequently, evacuation of the affected population is typically the key life safety
strategy and goal rather than sheltering in buildings.

21
NIST TN 2262
August 2023

Camp Fire Example 3. Inadequacy of existing infrastructure buildings as fire shelters.

The damage to Ponderosa Elementary School (a) and Feather River Hospital (b and c) illustrate just
two examples of existing infrastructure that were not adequate for use as WUI fire shelters during
the Camp Fire. Despite having more robust construction than typical residential structures, they are
not currently designed to withstand WUI fire exposures. Both the school and hospital buildings
ignited and were actively defended by firefighters, largely saving the structures. The damage to the
buildings was extensive, even with significant efforts by firefighters, and one defended hospital
building was destroyed. School children, hospital patients, and other susceptible populations cannot
shelter in place in existing infrastructure that is not designed specifically to withstand WUI fire and
ember exposures.

22
 NIST TN 2262
August 2023

Camp Fire Example 4. Fuel treatments alone are likely not sufficient to protect existing structures
for use as shelters.

The parking area of the Pine Ridge School in Magalia was a pre-designated public assembly point
(PAP) in the Paradise-Upper Ridge evacuation plan. The school campus, shown in the drone
imagery above, experienced an intense period of fire spread on the morning of November 9, 2018.
Before the fire, a fuel reduction and mastication program had recently been completed south of the
school buildings, indicated by the red dashed border in the map above. The photograph shows the
condition of the forest after the fire. The primary school buildings survived due to a combination of
reduced fire exposures from the adjacent fuel treatment and defensive actions by firefighters who
used the relative safety of the PAP to escape from a nearby burnover and were thus able to defend
the structures. Auxiliary temporary classroom buildings were ignited but suffered limited damage
owing to the actions of the firefighters.

23
NIST TN 2262
August 2023

A challenge arises from the combination of insufficiently hardened facilities and the potential for
a no-notice event. Evacuations can occur during the day or at night, during business hours or on
holidays. A key consideration is the ability of individuals to evacuate by their own means rather
than shelter in place. The range of evacuation capabilities of civilians from different locations is
presented in Table 4. Note that all types of locations might require specialized evacuation
considerations; there are individuals who will potentially require assistance or not have the
means to evacuate in all locations and population groups.

Table 4. Evacuation capabilities of civilians at different locations.

Location have means to do not have means require assistance

evacuate to evacuate to evacuate
Residence ✓ ✓ ✓
Work ✓ ✓ ✓
Schools variable ✓ ✓
Care Facilities employees ✓ ✓

WUI fire events can develop quickly. Where people will be and what the traffic conditions will
be at the beginning of an evacuation is related to the time of day the incident occurs. As is the
case with all disasters, Table 4 highlights that two key issues with evacuation of civilians are
access to transportation and the ability to self-evacuate. The potential for rapidly developing
events with short lead times presents several challenges for each of the civilian populations
identified in Table 4:
• Residences pose a particular challenge for evacuation of civilians with mobility
impairments because of the potential for many calls requesting evacuation assistance.
This is especially the case in retirement communities or other areas where the
demographics point to a less mobile population. The main evacuation challenge is that
large numbers of first responders must be available to respond to all the homes where
assistance is needed. Access to many residences may be compromised because some may
not be readily reachable, or access may be prevented due to fire. Civilians with limited
access to transportation but who are otherwise mobile also face challenges; however, in
these cases their mobility can enable evacuation with neighbors or public transportation,
or they could potentially walk to a safety zone or to a centralized evacuation location.
• Civilians at work can evacuate directly with their vehicles or with coworkers. However, a
challenge arises from subsequent intermediate trips after leaving work; for example,
when an individual decides to first return home to collect belongings or otherwise prepare
their home, or must pick up dependents (e.g., mobility impaired relatives or children)
from elsewhere. Traffic associated with these activities will impact road capacity and can
slow down overall community evacuation. Social tools, such as remote work during high
fire hazard weather events, may reduce road congestion and enhance evacuation.
• Evacuations of schools present several challenges, including that they can require parents
or guardians to pick up the children, and they may require a staggered evacuation if the
number of available buses is not enough to evacuate all students simultaneously. As

24
 NIST TN 2262
August 2023

suggested above, social tools such as remote work (distance learning) during high fire
hazard weather events may reduce road congestion and enhance evacuation.
• The evacuation of critical care facilities is a complex evacuation challenge for several
reasons, including:
o the need for specialized evacuation vehicles to transport patients with mobility
impairments and medical conditions; the vehicles must have sufficient capacity to
address all the facility residents.
o the need for partner facilities to accommodate the evacuees; this can be
particularly challenging when multiple facilities or an entire community is being
evacuated.
o the potential increase in time required for evacuation of these facilities owing to
mobility impairments and medical conditions of the evacuees.

Special consideration must also be provided for the life safety and potential evacuation of
emergency staff operating 911 dispatch and other communication and infrastructure facilities.
The hardening of such facilities is beyond the scope of this report; however, the safety
considerations for the life safety of evacuating civilians can also be used to address the
evacuation of emergency officials. The evacuation of first responders can have a significant
impact on response operations, for example, if 911 dispatch or the emergency operations center
needs to be evacuated for the safety of the first responders. Continuity of operations may be
impacted during evacuation, since these facilities have specialized equipment and infrastructure
that cannot be readily transported or replicated.
Potential solutions to some of these challenges are discussed in the evacuation planning section
of this report (Sec. 6.2). The authority to issue mandatory evacuations, enforcement of
evacuation orders, and the rights of civilians to not evacuate [75] are beyond the scope of this
report.

25
NIST TN 2262
August 2023

Camp Fire Example 5. Paradise Police Department 911 dispatch evacuated.

The Paradise Police Department (PPD) 911 operators/dispatchers started receiving calls related to
the Camp Fire by 07:07. Dispatch operations from the PPD building were suspended at 10:30 due
to the approaching fire, and the dispatch staff evacuated to Chico Police Department. It took them
approximately two hours to travel the 27 km (17 mi) route (the shortest route of 20 km [13 mi] was
blocked). Suspension of dispatch operations at PPD impacted response operations and public
notification. Radio communications from Chico to Paradise were patchy, often requiring officers to
relay messages between each other and with dispatch.
Fire reached the area surrounding the PPD building at 13:00. The parking lots around PPD, together
with the park to the east, protected the building against direct fire exposures. To protect against
firebrands, PPD was defended by several officers and firefighters. Defensive actions included
removal of debris from the roof, active suppression with a garden hose, and patrol by a fire engine.

3.3. Compounded Uncertainties in Fire/Evacuation Predictions

To convey the complexities associated with predicting an evacuation event, this section is
divided into two parts. The first part describes the general sequence of a WUI fire incident and
gives an overview of variables and consequences. The second discusses the use of models as pre-
event support tools and aids during WUI fire operations, in the context of the variables and
consequences presented in the first section.

3.3.1. Progression of a WUI Fire Event

WUI or wildfire events that impact communities can occur in a variety of scenarios with
different timelines and impacts to a community. A fire can start from a single point ignition and,
under the right conditions, grow rapidly, breach initial containment attempts, and impact

26
 NIST TN 2262
August 2023

communities (recent example: Camp Fire [CA, 2018]). Alternatively, a fire may burn for several
days with limited direct impacts to a community, before it quickly intensifies or shifts direction
due to wind or weather changes and spreads into the community (Waldo Canyon Fire [CO,
2012]). Long-duration and large area fires may impact multiple communities in sequence as the
fire continues to spread (Dixie Fire [CA, 2021]). Some fires may exhibit many of these timelines
for surrounding communities (Caldor Fire [CA, 2021]).
Figure 2 illustrates the relationships among fire, evacuation orders, and fire effects on
evacuation. The events are listed in a general chronological order from top to bottom, although
event specifics may result in temporal overlaps or loops. The figure shows two key
events/outcomes, one where fire affects evacuation the other where fire does not affect
evacuation.

Fig. 2. Primary wildland/WUI fire event components leading to community evacuation.

Not captured in Fig. 2 are several challenges associated with the links connecting the different
boxes, namely:
a) reliably determining which ignition scenarios will result in uncontrolled fires,
b) reliably determining whether a fire will pose a threat to one or more communities, and
c) reliably quantifying fire spread and its potential impacts on public evacuation.

27
NIST TN 2262
August 2023

Conditions that impact the capability to quickly contain a fire include the location of the ignition,
the fuels present (type, quantity, moisture content), topography, weather conditions (wind and
humidity), the accessibility and time needed to reach the area at or near the ignition location, and
the extent of available suppression resources. The above list alludes to the number of possible
scenarios that need to be considered.

Camp Fire Example 6. Ignition location and rapid fire spread.

The ignition of the Camp Fire (at approximately 06:15) was located 1.6 km (1 mi) northeast of the
community of Pulga in the Feather River Canyon. Visible from CA Highway 70 (photo above), the
actual ignition location along the high voltage electrical transmission lines was only accessible via
a narrow, winding, one-lane roadway difficult for emergency equipment to travel under the best
conditions. Furthermore, while the nearest fire station was a 12 km (7.5 mi) drive away and the first
engine arrived 13 minutes after being dispatched, additional resources had to travel more than 43 km
(27 mi) to access the ignition location.
Despite the rapid detection (06:25) and dispatch response (06:31), the challenging access and
location prevented the quick arrival of resources. Owing to the severe fire conditions at the time of
ignition (drought and high winds), the initial fire spread rapidly. The fire spread rapidly to the west,
cresting the ridge east of Concow, leaving little time to notify and evacuate the community. Attempts
to contain the fire east of Concow were not achievable given the intensity and size of the main fire
and long-distance spotting. Spot fires ignited within Concow by 07:20, 35 minutes after the first
engine arrived near the origin 6 km (3.7 mi) away.

These three challenges can be viewed as the capability to predict detailed fire behavior, typically
over a range of several kilometers. The smaller the community and the larger the distance from
the fire origin, the harder the prediction. Overall fire spread direction can often be inferred by
general wind and topography. The primary difficulty is the temporal component; quantifying the
rate of spread is difficult since the main/original fire front may be augmented by far-field

28
 NIST TN 2262
August 2023

spotting on the order of kilometers ahead of the front, making the fire advance much faster.
These spot fires can have significant impacts on community evacuation as discussed below
related to challenge c). The spread rate will also impact the width of the fire front that may
impact nearby communities.

Camp Fire Example 7. Spot fires in Paradise.

Thirty (30) confirmed spot fires ignited in Paradise between 07:49 and 08:30 (indicated by the red
points in map above) from embers ahead of the fire front, which landed as far as 4 km (2.5 mi) into
the community. Since the main fire front didn’t arrive at the east side of Paradise until 08:30, these
spot fires must have been ignited from fuels burning outside and upwind of the community. At least
35 additional spot fires ignited between 08:30 and 10:00 (black points), after the fire front arrival.
The spot fires were uncontained for four primary reasons:
1. high ignition potential of the fuels (drought),
2. amount of fuels present (intermix vegetation, limited/no fire history),
3. number and spatial distribution of spot fires compared to available firefighting resources, and
4. enhanced spread of spot fires due to weather conditions (low humidity and high wind).
Notably, only one of the four reasons, fuels management, is under the control of the community.
The widespread distribution of spot fires throughout the community forced the decision for the
simultaneous evacuation of the entire community at 08:03, rather than the pre-planned phased
evacuation of distinct zones. The widespread spot fires impacted the full-community evacuation
and, in several cases, caused burnovers that entrapped first responders and evacuating civilians (see
Camp Fire Example 9).

29
NIST TN 2262
August 2023

Challenges b) and c) differ in that a community may be impacted by fire after the community has
been evacuated. If fire behavior and fire spread can be reliably quantified to determine whether
one or more communities will be impacted by a fire, then the remaining task is to determine what
the impacts of that fire will be on evacuation. Here, the challenges are threefold. First, from b),
an estimate needs to be developed as to when the fire will reach the community. Then this
information needs to be processed in the context of the evacuation status. Finally, an assessment
needs to be made on the impact of that fire on an ongoing evacuation. How fire impacts a
community is very specific to the local conditions. The ignition potential of fuels, type of
exposure (fire vs. embers), locations of ignitions, availability of resources for defensive actions,
and impact of all of the above on egress arteries is scenario specific. Fire can impact evacuation
directly (smoke, flames [radiation, convection]) or indirectly (downed utility lines and poles,
other burned or burning obstructions). These impacts can result in the restriction or closure of
egress arteries, or worse, lead to civilian entrapments and burnovers. Evacuation flow restrictions
can then propagate along roadways and create traffic problems that require active management
during rapidly changing and deteriorating conditions.

Camp Fire Example 8. Escalation of traffic gridlock.

The Camp Fire impacted evacuation of Paradise, specifically the traffic, in two distinct ways. Firstly,
the ember showers that resulted in 30 spot fires throughout Paradise before the arrival of the main
fire front resulted in the need to evacuate the entire community simultaneously. Secondly, the fire
front, which extended along almost the entire eastern edge of the community, resulted in rapidly
deteriorating conditions in the eastern part of town, followed by the central and western parts.
The combination of traffic gridlock and fire impacts resulted in multiple burnovers that entrapped
evacuees and restricted evacuation of civilians and ingress of first responders. The rapidly
deteriorating egress conditions illustrate how difficult it is for civilians to travel though highly
congested areas after a fire has impacted a community.
The map figures below illustrate how quickly the traffic conditions deteriorated in the hour after the
first spot fire reached the community. At 08:00, traffic was flowing throughout Paradise (indicated
by green segments) as the first spot fires ignited in eastern Paradise. By 08:30, traffic throughout
Paradise had significantly deteriorated (indicated by yellow and orange segments), and fire overtook
traffic both on Pentz Road and upper Skyway (indicated by the red segments). These two burnovers
significantly affected traffic flow out of and through Paradise.
Each of the five egress arteries out of Paradise were closed due to fire at least once during the
evacuations on November 8. At 09:00, the first evacuation route to close was Pentz Road (the
easternmost artery). Conditions continued to deteriorate and between 11:30 and 13:00, three of the
four southbound egress routes were closed simultaneously.

(continues on next page)

30
 NIST TN 2262
August 2023

31
NIST TN 2262
August 2023

Considering the above three challenges, a) may be easiest to address. While there is significant
uncertainty in reliably determining which ignitions will result in uncontrolled fires, data are
available on both containment and non-containment events in many regions. Incorporation of
local expert knowledge may be able to increase the reliability of anticipating either type of event;
however, over-reliance on fire history may underpredict fire spread rates and intensity in future
incidents.

3.3.2. Pre-fire vs. During Fire Modeling

The purpose of this section is to highlight high-level issues associated with fire modeling, both
before and during the event, in the context of the challenges posed in the previous section. A
variety of models are in development and use for fire spread prediction, evacuee behavior, and
evacuation and traffic modeling. This section will not review all of the various types of models
available for these tasks, neither will it review model output uncertainties. The main focus of this
section is on the use of various models to address the challenges identified in Sec. 3.3.1, and how
the use of these models may impact different aspects of evacuation. This highlights the
challenges associated with linking multiple models (i.e., using the output from one as the input of
the next).
Several global constraints apply to using models, with some variation in their importance among
the different types of models. These constraints include:
1. The large number and broad range of scenarios to be considered
2. Trackable reliability (certification) of tools and accreditation of users
3. Individual component uncertainty and compounded total uncertainty
4. Interpretation and use of model outputs.

3.3.2.1. The large number and broad range of scenarios to be considered

Regardless of any predictive model use, there are many potential fire event and evacuation
scenarios. A very large number of possible scenarios may need to be considered as large
geographic areas may be involved with many possible ignition locations, different fire spread
directions, decisions on evacuation, and impacts of fire on evacuation. Planning decisions cannot
hinge on individual scenarios and must be flexible to accommodate the broad range of outcomes.
While models present a path forward to facilitate advance planning, the large number of
scenarios required may lead to exorbitant computational costs, especially for smaller
communities. The other constraints discussed below also affect the benefits provided by
fire/evacuation modeling.

3.3.2.2. Certification of tools and accreditation of users

The issues of model performance, limitations, and validation apply independent of the model
type or complexity. The same applies to the accreditation or training of the user. It should be
noted that the availability of a model does not necessarily make it the appropriate choice for
addressing a specific problem. This is particularly important to consider for models initially
designed for one application that are later utilized for another. Examples include structural fire

32
 NIST TN 2262
August 2023

spread models expanded to the WUI as well as wildland fire models applied to the WUI. Models
must be verified and validated for the new application before they can be trusted to provide
correct and useful results.

3.3.2.3. Individual component uncertainty and compounded total uncertainty

The effects of a wildland or WUI fire on the evacuation of a community are the outcome of a
complex sequence of events. Small changes in one event or input may significantly affect the
outcome. Quantifying the uncertainties of evacuation predictions requires an understanding of
the uncertainties of all the key components that impact the evacuation. As fire spread impacts
evacuation decisions and the evacuation impacts traffic, this entire system can be viewed as a
linked system resulting in compounded uncertainties. Each step must account for the
uncertainties associated with the input from the previous step.
Figure 3 illustrates the complex relationships among the various modules and how they
ultimately affect community evacuations. The coupled weather and fire models impact AHJ
notification and evacuation decisions, which impact civilian evacuation decisions. These
individual decisions, together with the actualized impacts of fire on egress arteries, impact traffic
management decisions and, finally, civilian evacuation.

Fig. 3. Linked modules associated with evacuation predictions. Uncertainties are compounded and
propagate from left to right and illustrated in red (not to scale).

Figure 3 illustrates the interconnectivity of the components that ultimately drive evacuation
conditions. Note the two-way arrow between the weather module and the fire behavior module;
large fires can affect local weather, and some models incorporate a feedback loop to address this
coupling [76]. Each component listed in the figure (white boxes) represent complex systems. The

33
NIST TN 2262
August 2023

polygon in red illustrates the ever-increasing prediction uncertainties (not drawn to scale in the
figure). Note that there is an initial uncertainty associated with weather forecasting and the red
polygon does not start at zero.
While wildfire modeling has been actively researched for close to 50 years, the number of
variables, stochastic fire behavior, and ever-changing local fuel and weather conditions make it
difficult to reliably predict landscape-scale fire spread at the resolution of interest to the
evacuation problem. Furthermore, there is frequently a difference in the temporal scale of
interest between wildland fires and WUI fires (days vs. hours or minutes, respectively), implying
that the application of wildland fire models to specific WUI applications may not be as direct as
it seems.
The limitations of spatially resolving wildland fire behavior can also have significant
consequences on evacuations. In a large wildfire, far-field spotting ahead of the flaming front
may ultimately have little impact on the overall wildfire; however, spot fires can have significant
impact on a community. Spotting can result in multiple ignitions within the community that can
directly or indirectly impact evacuations, challenge firefighting resources, and complicate
evacuation decision-making and notifications.

3.3.2.4. Interpretation and use of model outputs

The interpretation and application of evacuation modeling results depends on the use of tools
well-suited to address the specific module needs. It must account for the large number of
scenarios and the compounded uncertainties as discussed above.
If a model is validated and tested for the conditions of interest and the uncertainties are known,
modeling may provide insights for scoping different scenarios; however, the linking of multiple
models may still result in large uncertainties, jeopardizing the utility of the results. The
variability of fire behavior at the scale of interest to evacuation needs to be acknowledged, as
small disturbances/changes can result in significant impacts on fire spread pathways. While past
fires may provide validation opportunities if sufficient data is available, changing conditions
(weather, drought, and fuels buildup) can result in unprecedented fire behavior, particularly for
communities that have not experienced fire in a long time and have no directly applicable fire
experience. Community evacuation drills may provide opportunities to collect non-emergency
data to further support model development and validation, such as recent work by Gwynne et al.
[54].
The inherent limitations of models with very large uncertainties may result in a more
conservative evacuation approach as an understanding is gained of how variable WUI fire spread
and evacuation events can be. The challenges and complexities associated with developing
reliable evacuation predictions, which are highlighted in this section, point to a need for a
simplified evacuation approach that leverages the known uncertainties to create an
implementable evacuation system focused on life safety.
Interpretation of model outputs requires expert knowledge of the model inputs and architecture.
For example, one desired output of a model may be an optimized evacuation time estimate.
However, in cases where evacuation of the community is simultaneous with the fire, alternative
traffic control strategies may need to be implemented to prioritize life safety and reduce the
number of vehicles exposed to or stuck in hazardous conditions. This strategy was used by first

34
 NIST TN 2262
August 2023

responders during the Camp Fire when vehicles were ushered onto secondary roadways within
the town of Paradise in order to move the end of the line of traffic away from the advancing fire
front [6]. A recent study of evacuation traffic modeling found that evacuation strategies like
phased (or zoned) evacuations and implementation of contraflow where feasible can effectively
reduce the number of vehicles exposed, although they may not necessarily reduce overall
evacuation times [77]. If the focus of the analysis is on reduced evacuation time, certain
strategies or tangential goals may be overlooked.

3.4. List of WUI Community Evacuation Challenges

The following list summarizes the issues presented in this section that make planning for and
executing WUI evacuations challenging. There is a need for a simplified adaptive approach to
address the formulation of evacuation plans, particularly for small and medium-sized intermix
communities. Additional components of a comprehensive evacuation plan include notification of
civilians and first responders, and situational awareness of emergency officials and the public.

Before the Fire

1. Large number of possible fire scenarios (ignition location, fuel presence, fuel moisture
content, weather).
2. Chaotic behavior, in which small perturbations of variables can result in large changes in
predicted event outcomes.
3. Difficulty in characterizing, quantifying, and analyzing the large number of different fire
scenarios.
4. Complexities of modeling and predicting human behavior in evacuations and response to
emergency situations.
5. Difficulties in how to account for the uncertainties in the methods used to generate the
different scenarios/predictions.
6. Difficulties in how to use/implement the findings from the above-mentioned
scenarios/predictions.
7. Need to characterize and quantify the possibility of non-containment of the fire (to
address the large number of ignitions that do not result in catastrophic events).
8. Need to develop contingencies for events like loss of communication and power.
9. Need to develop contingencies for potential closures or obstructions of egress arteries.
10. Need to evaluate evacuation through high-hazard wildland areas (which may result in
burnovers), an issue that is particularly important for remote intermix communities.
11. Need to evaluate evacuation pathways that lead through urban areas for intermix
communities adjacent to or near a large urban area.
12. Need to develop evacuation plans that address the above issues.
13. Need to disseminate the evacuation plans to first responders and the public.

35
NIST TN 2262
August 2023

During the Fire

1. Limits in situational awareness, including dynamic outages in data sources and
communications.
2. Integration of rapidly changing conditions into ongoing evacuation activities.
3. Large uncertainty in fire spread during incidents.
4. Communication to first responders and civilians of any changes to the evacuation plan.

36
 NIST TN 2262
August 2023

4. Fire-Evacuation Temporal Relationships and Evacuation Failures

Successful evacuation, meaning evacuees are not exposed to hazardous fire conditions during
evacuation, is a function of the temporal relationship between fire spread and the evacuation
process. To better understand the impact of this relationship, potential failure modes are
presented, followed by sample timeline scenarios that may lead to these evacuation failures.

4.1. Primary Modes of Evacuation Failures

Recognizing different ways in which evacuations can fall short of their objectives can be useful
in identifying potential contingencies to maintain life safety. The use of the term “evacuation
failure” in this report is not intended to convey or assign blame but rather to highlight scenarios
that result in undesirable outcomes. Such outcomes may be non-life threatening or can include
injuries or fatalities.

4.1.1. Defining Failure

Failures of evacuation events can be divided into two types based on whether the shortcomings
impact life safety:
Type 1: Undesirable Evacuation Consequences – No impacts to life safety
This classification reflects situations where there were no direct threats to life safety of the
general population4 during the evacuation, but the evacuation was characterized by other
undesirable results. Two examples of this include:
1. Prolonged evacuation, extending beyond the expected duration; this may highlight needs
for adjustment to the existing evacuation plan.
2. Evacuation conducted when fire does not end up impacting the community, and the
associated:
a) economic cost of evacuation (personal and commercial), and
b) evacuation fatigue, potentially resulting in resistance to evacuate in future events.
Type 2: Evacuation Failures – Impacts to life safety
This type of failure can be described when residents experience high fire exposures at their
residences or during evacuation. Causes associated with these scenarios include:
1. Inability to effectively communicate evacuation orders to residents in a timely fashion,
delaying the start of evacuation.
2. Fire ignition near the community resulting in only a short time to safely evacuate when
the resulting fire behavior and rate of spread outpace the evacuation process.
3. Underestimation of the fire rate of spread, or changing conditions, resulting in fire arrival
at the community sooner than anticipated.
4. Underestimation of the time required to evacuate the community or part of community.
5. Underestimation of the impact of fire on egress arteries.

4
Note that evacuations may induce a larger health burden on susceptible subpopulations, such as hospital patients and individuals with
disabilities, compared to the general population as a whole.

37
NIST TN 2262
August 2023

These types of failures will also impact first responder operations, primarily through
prioritization of rescues ahead of fire suppression or control, including access for rescues that
may be restricted by fire.
Residents can experience high fire exposure conditions in several situations. The three primary
scenarios that can result in injuries or fatalities are:
1. An inability to evacuate owing to reduced mobility (e.g., physical or medical factors) or
lack of access to a vehicle or other transportation.
2. High exposures at one’s residence experienced after a decision to stay (whether to shelter
in place or stay and defend).
3. High exposures experienced during egress (i.e., burnover)
a. during a late or delayed evacuation after an initial decision to stay or after
accomplishing specific tasks like getting kids from school.
b. being overrun by fire due to rapid fire spread or due to traffic or other evacuation
delay.
Injuries and fatalities of individuals can occur at or near residences or during evacuation.
Exposure conditions can vary dramatically over short distances (on the order of 10 m [30 ft]) and
within short time frames (on the order of a minute). Some fire exposures can be short in duration,
such as from a burning bush, while others can last an hour or longer, such as from a burning
structure. Smoke exposures can last for hours or days.

High fire exposures at residences

Injuries and fatalities at home can be associated with structure ignition or deteriorating local
conditions. These high exposure conditions can occur when the structure cannot withstand the
incoming exposures (fire and embers) or extensive burning occurs in the vicinity of the home
(fire/smoke exposures). Affected residents may be inside or outside of the structure, whether
taking shelter, conducting defensive actions, or attempting to evacuate.

High fire exposures during egress

Injuries and fatalities during travel can result from the burning of high fuel loads present along
key egress roadways. Potential causes of high fire exposures for evacuees include:
• Traffic delays leading to extended time in high hazard areas.
• Road closures far ahead of the fire front (due to spot fires), causing delays and burnovers
when traffic extends back into the fire area.
• Burnover conditions can occur even without heavy traffic, resulting in civilians getting
trapped.

38
 NIST TN 2262
August 2023

Camp Fire Example 9. Burnover events that impacted evacuating civilians and responding
emergency personnel.

Entrapment/burnover events are defined as life-threatening situations where planned escape routes
or safety zones are inadequate or compromised and individuals are overtaken or trapped by fire,
often resulting in equipment damage and personal injury or death. The post-fire case study identified
23 such events that occurred in the first 26 hours of the Camp Fire [6, 7, 78], 17 of which involved
an estimated combined total of up to 500 civilians. A total of seven civilians were killed in three of
the 23 events. The locations of the 23 burnover incidents are shown in the map above, occurring
throughout the fire area.
Out of the 23 identified burnovers, 11 impacted primary egress arteries during the peak of the
evacuation, roughly between 08:00 and 12:00. These events are indicated by the circled burnover
ID points in the map above. Fire overtook evacuees who were stuck in gridlocked traffic in five
instances, and intense fire impeded or trapped moving traffic in six instances. The closure of
Concow Road, the sole egress route in Concow, significantly affected the evacuation of that
community. In Paradise, some egress arteries were closed for several hours, impacting both civilian
egress and first responder access and operations. At 09:45, two hours after the first spot fire ignited
in Paradise and a little over three hours after the fire was reported, two of the four southbound egress
arteries were closed due to fire. By 11:45, during the peak of the Paradise evacuation, three of the
four were closed due to fire, significantly impacting evacuation.
The concurrence of evacuation and fire impact on the community significantly affected the life
safety of evacuating civilians. The formation and use of TRAs significantly enhanced the life safety
of entrapped individuals.

39
NIST TN 2262
August 2023

4.1.2. Addressing Type 1 Evacuation Events

Type 1 failure events when fire does not end up impacting the community would ideally be
avoided but should not be viewed only as “unnecessary evacuations.” Instances like this can
benefit community preparedness and experience and identify improvements to the evacuation
plan and execution. Given uncertainties in fire spread predictions, a small change may have
resulted in a direct hit to the community. Improved fire spread predictions, coupled with
improved evacuation modeling, may help to reduce the uncertainty of predictions in the overall
evacuation system. However, this is difficult to achieve because of the chaotic nature of WUI
events, the stochastic ignition and nonlinear impacts of spot fires, and the complexities
associated with traffic redirection of evacuees occurring in real time. While advancements will
be achieved in each component, an overall reduction in uncertainties to reliably predict what will
happen is a long-term goal.
The quantification of the economic impacts of an evacuation when the community does not
experience a fire are beyond the scope of this report. There is a need for economic modeling to
provide guidance for the benefit cost of repeated evacuations compared to the probability of a
community getting impacted by the fire. This information would help inform the public and help
AHJs further refine community evacuation thresholds and education campaigns.

4.1.3. Addressing Type 2 Evacuation Events

Type 2 scenarios, when fire directly affects evacuations, can result in injuries or fatalities. This
section addresses the three scenarios identified in Sec. 4.1.1.

Inability to evacuate and exposed to fire

Scenarios in which civilians are unable to evacuate owing to mobility impairments or a lack of
access to transportation can be further divided into two separate categories:
a) Events that occur when there is sufficient time to safely evacuate the community before
the fire affects evacuation.
b) Events that occur when there is little or no time between when an evacuation order is
issued and when the fire impacts the community.
Events in category a) might be addressed by implementation of evacuation programs for people
who need assistance. A primary issue to address in this case is the potential need for
simultaneous evacuation of many civilians and multiple care facilities. The
transportation/evacuation resources in the community should be assessed together with the time
necessary to get mutual aid resources on location.
Events in category b) are difficult to manage for three distinct reasons:
• Fire may rapidly restrict access to the area for first responders
• Fire and smoke may slow down evacuations of local civilians
• Traffic may be directed out of the area using contraflow to increase egress capacity,
making ingress of first responders difficult and dangerous.

40
 NIST TN 2262
August 2023

Camp Fire Example 10. Evacuation of assisted living facilities.

Two individual facilities illustrate a wide range of scenarios. One assisted living facility (c, above)
on the eastern edge of Paradise was located in one of the first areas of town impacted by spot fires
and the fire front (select nearby spot fires are indicated in the figure above). Urgent evacuation of
140 residents and staff was accomplished using various vehicles, including many first responder
vehicles, concurrent with evacuation of the neighboring hospital and the community. Fire was
observed spreading onto the property by 08:00. Law enforcement requested additional evacuation
support by 08:34, and all residents were evacuated by 09:00. Firefighter actions at the main building
extinguished several spot fires after the residents had been evacuated, and several detached
residences were destroyed by the fire.
A second, smaller, skilled nursing facility located at the center of town (b, above) was not directly
threatened by the initial fire impacts to Paradise. Evacuation assistance was first requested at 10:30.
After several hours, presumably related to first responder prioritization, threat levels, and
availability of transportation (vehicles and access), the facility was evacuated between 14:00 and
15:00 as fire was approaching. A mutual aid task force of a dozen ambulances arrived from two
hours away and evacuated the patients. The building was reported to have ignited as the evacuation
was being completed. The facility was destroyed by the fire.

41
NIST TN 2262
August 2023

An approach to address the Category b) events where evacuation is difficult is to harden the
structure or facility. This is not to create a fire shelter, but rather to briefly extend the time
available for evacuation by reducing the ignition potential as much as possible. The WUI
Structure/Parcel/Community Fire Hazard Mitigation Methodology (HMM) [40] is an example of
a comprehensive approach to address structure and parcel hardening for both fire and ember
exposures. While HMM was not developed explicitly for commercial facilities, the approach can
be applied and adapted to assess the exposures and address the vulnerabilities. Hardening the
facility against fire does not imply that the facility will necessarily be suitable as a fire shelter.
Ventilation, power, and other tenability and access issues necessary to create a fire shelter are
beyond the scope of this report. Examples of items that would need to be addressed beyond what
is explicitly listed in the HMM include hardening of ventilation systems, hardening of the
roofing assembly, and instituting a requirement for the parking of commercial transport vehicles
(e.g., minibuses, ambulances) at a safe distance from the structure.
Along with the development of evacuation plans for individuals with disabilities, the hardening
of individual residences is also necessary. Individual households and multi-patient care facilities
can challenge first responders in different ways. Evacuating multiple individual households is
intensive in time and resources, requiring many distinct stops. In many cases, evacuations from
these residences or facilities may be accomplished by vehicles or fire engines; however,
responding to many individual calls around the community is problematic because of the number
of vehicles needed, the challenges in traveling to the locations, and the potential difficulties of
getting the civilians out of the fire. For multi-patient facilities, large capacity vehicles are
typically necessary for transport. If such vehicles are not present at the facility, they need to be
located and driven there. Medically vulnerable patients require specialized transport and staff to
support them. Evacuation of patients during high exposure conditions, potentially resulting in
burnovers, is hazardous for the patients and first responders.

High fire exposures after decision to stay

High exposures may be experienced at a residence after a decision to stay, whether to shelter in
place or to actively defend against fire. Cases that result in injuries or fatality can hopefully be
avoided by appropriate education and individual planning. Education and information campaigns
highlighting the dangers of wildfires, together with the limitations and priorities of first
responders during these events, can be used to inform the public. It is likely that some people
will chose not to evacuate; however, those people should be aware that these alternative
approaches can be extremely dangerous even for well-prepared individuals, as discussed in Sec.
2.2.

High fire exposures during egress

Avoiding situations where evacuees experience high exposures during egress, i.e., burnovers,
that result in injuries or fatality is a primary goal of an evacuation plan. One approach to mitigate
these events is to reduce the potential for high fire exposure along the key egress routes and
arteries. This maintains tenability of the routes and allows evacuees to remain in their vehicles to
reach safety. Making this an effective approach requires fuel thinning and vegetation removal
along evacuation corridors and continued maintenance of these fuel treatments over time.
Hardening or burying utility infrastructure along roadways will also reduce potential for
obstructions. Two challenges associated with mitigating exposures to egress routes are the

42
 NIST TN 2262
August 2023

access/rights to conduct the fuel treatments (involving rights-of-way and private property) and
the expense, both short-term to perform the initial clean-up and long-term to maintain the
treatments. Access for fuel treatments is particularly important, since high radiative and flame
exposures can occur tens of feet from burning fuels and therefore will necessitate access to
properties well beyond the typical rights-of-way. Some fuels, such as structures close to the road
that may impact egress arteries when they burn, can be very difficult or effectively impossible to
remove for multiple reasons (e.g., critical infrastructure, historical status, or ownership).
Hardening may improve the fire behavior of these structures.
A second approach to mitigating the risk of high exposures for evacuees is to assemble residents
at a wildfire safety zone or other preplanned and identified safer place. A distributed network of
wildfire safety zones can reduce travel time for residents seeking shelter when there is no safe
route for full evacuation.

4.2. Temporal Relationships Among Fire Progression, Notification, Evacuation,

and Sheltering
The relationship between the time fire arrives the completion of evacuation influences how the
fire may affect evacuees. Ideally, evacuations occur before being affected by the fire, both in the
community and along the egress routes. This section presents the baseline minimum time
required for evacuation, beginning from detection and extending through decision-making by
emergency officials, notification of the public, and transportation out of the hazard area.
Five fire/notification scenarios are presented with varying times of fire arrival during the
evacuation. Scenarios range from very dire situations where fire ignites or arrives very near a
community without any notification, causing immediate threat to life safety, to scenarios where
fire spreads into the community after evacuation.
Then, five evacuation scenarios are described and related to the fire/notification timelines. The
evacuation scenarios range from immediate shelter in place to a safe evacuation from the
community without evacuees being exposed to fire.

4.2.1. Minimum Time for Community Evacuation

There is a minimum amount of time needed to execute an evacuation. In fire protection
engineering this is often referred to as the required safe egress time (RSET). RSET includes time
for detection, alarm, pre-movement, and evacuation. In the WUI, the WUI RSET (WRSET) [55,
58, 66] includes additional steps not typically encountered or that are typically much shorter in
building evacuation timelines, including the time required to assess the ignition/fire situation,
communicate this information to the incident commander and emergency operations center,
decide on the required evacuations, begin the notification and evacuation processes, and conduct
the evacuation. A minimum WRSET is defined here as the time needed to evacuate a community
utilizing all available tools (like contraflow) in the absence of any direct or indirect fire impacts.
This is the best-case evacuation and would occur only in conditions where the fire reached the
community after the entire evacuation was completed.

43
NIST TN 2262
August 2023

Ronchi et al. [58] proposed a WRSET time as

WRSET = 𝑡𝑑 + 𝑡𝐹𝐷𝑆 + 𝑡𝐹𝐷𝐼 + 𝑡𝑁 + 𝑡𝑝𝑟𝑒𝑝 + 𝑡𝑓𝑜𝑜𝑡 + 𝑡𝑣𝑒ℎ + 𝑡𝑟𝑒𝑓 (1)

where 𝑡𝑑 is the time elapsed from ignition to detection, 𝑡𝐹𝐷𝑆 and 𝑡𝐹𝐷𝐼 represent time for fire
department situational assessment and intervention, and 𝑡𝑁 is the time required for notification.
The remaining terms relate to the evacuee timeline, including preparation time 𝑡𝑝𝑟𝑒𝑝 , movement
time on foot and in vehicles, 𝑡𝑓𝑜𝑜𝑡 and 𝑡𝑣𝑒ℎ , and the time to be on-boarded at a place of refuge
𝑡𝑟𝑒𝑓 .
To make the timeline more salient to the incident commander (IC), Eq. (1) can be reformulated
to expand some terms and condense others in order highlight different components of the
evacuation timeline as viewed from the IC perspective:
WRSET2 = 𝑡𝑜𝑟𝑑𝑒𝑟 + 𝑡𝐼𝑇 + 𝑡𝑑𝑖𝑠𝑠 + 𝑡𝑝𝑟𝑒𝑝 + 𝑡𝑡𝑟𝑎𝑛𝑠 (2)

The first term of WRSET2 in Eq. (2), 𝑡𝑜𝑟𝑑𝑒𝑟 , is the time it takes the IC to request an evacuation
order. This term includes the time from ignition to detection plus the time for situational
assessment and decision making. In relation to Eq. (1), 𝑡𝑜𝑟𝑑𝑒𝑟 = 𝑡𝑑 + 𝑡𝐹𝐷𝑆 + 𝑡𝐹𝐷𝐼 .
The time for information transfer 𝑡𝐼𝑇 and dissemination 𝑡𝑑𝑖𝑠𝑠 are terms representing how long it
takes for the evacuation request to get to the agency responsible for the evacuation and how long
it takes for that agency to implement the orders and disseminate the information to the public.
For this time estimation, the 𝑡𝑑𝑖𝑠𝑠 term represents the time to the start of the dissemination (e.g.,
the first notification or reverse-911 call), not the complete notification. This is intentional to
provide an absolute minimum total time for evacuation for planning purposes. This information
by itself does not provide an estimate for the time needed to inform the majority of a certain
population. These terms are an expansion of Eq. (1); 𝑡𝑁 = 𝑡𝐼𝑇 + 𝑡𝑑𝑖𝑠𝑠 .
The last two terms 𝑡𝑝𝑟𝑒𝑝 and 𝑡𝑡𝑟𝑎𝑛𝑠 represent the minimum time for civilians to get out of the fire,
including preparation and transport, respectively. The transport term incorporates both
movement on foot and in vehicles from Eq. (1); 𝑡𝑡𝑟𝑎𝑛𝑠 = 𝑡𝑓𝑜𝑜𝑡 + 𝑡𝑣𝑒ℎ . To first order, a minimum
𝑡𝑡𝑟𝑎𝑛𝑠 can be estimated or modeled based on scenarios in which evacuation is not impacted by
fire (i.e., no burnovers or other road closures due to fire or fire effects), or it can be established
through evacuation drills. The no-fire scenarios should account for traffic and the utilizations of
traffic management systems, such as contraflow, where applicable. A full community evacuation
drill may provide valuable information and insight towards quantifying an evacuation time
𝑡𝑡𝑟𝑎𝑛𝑠 .
WRSET2 also differs from WRSET in that WRSET includes the time to be onboarded at the
place of refuge. While part of the overall evacuation process, this time is not directly related to
the IC decision to order evacuations and has been omitted from WRSET2.
While WRSET2 represents a minimum time required for evacuation, there is no minimum time
between a fire ignition and fire reaching a community. The worst case conditions are scenarios
where there is not sufficient time to safely evacuate communities, referred to as dire scenarios
[36]. These scenarios need to be characterized, understood, and incorporated into community
evacuation plans. This report will highlight these scenarios and outline implementable solutions
to manage the risk to evacuating civilians.

44
 NIST TN 2262
August 2023

An additional consideration is the time of day when the evacuation takes place, which influences
where people will be and what they may be doing. Equation (3) calculates 𝑇𝑒𝑣𝑎𝑐 , the time of day
at which the evacuation is completed. The clock time is set by defining 𝑇𝑖𝑔 as the time of day of
ignition (24-hour hh:mm).
𝑇𝑒𝑣𝑎𝑐 = 𝑇𝑖𝑔 + 𝑊𝑅𝑆𝐸𝑇2 (3)

Considering the actual time of day as well as the elapsed time may highlight evacuation/traffic
issues like rush hour or dropping off/picking up children from school.

4.2.2. Fire/Notification Timing Scenarios

As discussed in the previous sections, exposures (fire and smoke) can impact ongoing
evacuations in certain fire ignition/spread and evacuation scenarios. In some cases, rescues may
be necessary. The timing between ignition and community impact has a direct effect on the time
available to evacuate and the time for emergency personnel to respond to requests for evacuation
assistance. In this section, five scenarios are defined in space and time by the presence of fire,
notification status, and ability to safely egress. Fire/notification scenarios FN1, FN2, and FN3
have the potential to expose evacuees to high hazard conditions. To protect civilian and first
responder life safety, the goal is to operate within scenarios FN4 or FN5.
The schematic in Fig. 4 denotes the general timelines for each scenario.

Fig. 4. Fire/notification/evacuation timeline scenarios as a function of evacuation status and distance

between the community and fire origin.

The scenarios represent sequential temporal interactions among the fire environment,
response/rescue attempts and community evacuation. After an ignition, time is always necessary
to assess the fire spread and local conditions, make decisions on evacuation and response, inform
the relevant agencies/AHJs and initiate the notification process. Even if all of these tasks are
executed effectively and efficiently, there is still a non-zero minimum time between the fire
ignition and the beginning of the public notification (𝑡𝑜𝑟𝑑𝑒𝑟 + 𝑡𝐼𝑇 + 𝑡𝑑𝑖𝑠𝑠 > 0). The area the fire

45
NIST TN 2262
August 2023

covers in that period is represented by FN1. The number of residences, commercial properties,
and civilians impacted will depend on what is in the area impacted by the fire during that initial
time interval between ignition and the start of notification. While an ignition far away from a
community may limit or eliminate the number of civilians involved in scenarios like FN1, the
overall impact of the fire on the community may be larger as the fire spreads and grows.

Camp Fire Example 11. Time of fire arrival and first official evacuation notification, by evacuation
zone.

The plot above, adapted from Ref. [6], illustrates the time gap between time of fire arrival and the
time of official notification for each evacuation zone for the Camp Fire as it spread through Concow
and Paradise. Early in the incident, before 08:45, evacuation notifications were sent after or
concurrent with the arrival of the fire in each evacuation zone (FN1 and FN2). This is represented
in the plot by data points below the x-axis. At 07:45, 80 minutes after the first report of the fire,
evacuation notifications were being sent at the same time as fire arrived within the notified
evacuation zones. The positive gap illustrated on the right side of the graph, after 08:45, shows how
notifications were issued ahead of the fire as the incident continued to evolve, giving the civilians
in these zones more time to evacuate before the arrival of fire (FN3).

46
 NIST TN 2262
August 2023

4.2.2.1. Fire/Notification Scenario FN1

Fire near or at residence, no official notification → exposure/entrapment during egress
In FN1, the situational assessment of the fire by the IC and emergency officials is concurrent
with fire impacts to a part of the community. In this scenario, civilians may see or be impacted
by the fire before they receive an official notification or evacuation orders. These conditions can
result in entrapments and burnovers during evacuation and limit the emergency response
resources available to perform rescue assistance. Life-threatening fire conditions may require the
formation of TRAs. If the fire origin is near the community, the area of community impacted can
be small; however, the affected area can increase as the fire spreads and ignitions occur farther
into the community (see Sec. 4.3).

4.2.2.2. Fire/Notification Scenario FN2

Simultaneous arrival of fire and official notification → exposure/entrapment during egress
This is similar to FN1 with the addition of an official notification of the fire from emergency
services who have been able to conduct at least a preliminary situational assessment and
formulate an alert message. Notification can occur through various means and may include door-
to-door messaging. The presence of first responders providing notifications can facilitate
evacuation. However, the presence of first responders cannot be interpreted as an ability to
reduce exposures through defensive actions. Again, high hazard conditions may require the
formation of TRAs.

47
NIST TN 2262
August 2023

Camp Fire Example 12. Simultaneous arrival of fire and evacuation notification, leading to
entrapment during evacuation in Concow.

The map above shows the area of Concow, a rural community located between the origin of the fire
in Pulga and the town of Paradise. Red data points indicate individual fire observations before 08:00.
The red shaded areas roughly indicate the area of main fire activity in 30-minute intervals after
ignition. Note the significant number of spot fires ahead of the main fire activity.
The IC requested evacuation of Concow at 07:37, seven minutes after the first 911 calls were
received reporting spot fires in the area. All responding law enforcement officers (LE) were directed
to Concow to begin evacuations. Due to the location and the scattered spot fires ahead of the main
fire front, the 911 calls were the first indication to the IC that the fire was within Concow.
Firefighters on the ridge between Concow and Pulga observed the fire front spreading west at 07:30,
indicated by the intermediate shaded polygon.
One of the first firefighters to access the Concow area conducted drive-by notifications of residents
using the vehicle siren and public address speaker between 07:40 and 08:00. At the same time, law
enforcement was directing civilians to seek shelter in the designated Wild Fire Safety Zone at the
Camelot Meadow. Multiple spot fires grew rapidly and within minutes created impassable
conditions, entrapping evacuating civilians and first responders at several locations and resulting in
multiple burnover events and the formation of multiple TRAs (see Camp Fire Examples 15 and 19).

4.2.2.3. Fire/Notification Scenario FN3

No fire near/at residence, official notification, egress → exposed to fire on route to safety
This scenario may occur when residents do not experience fire at their residence or other starting
point of evacuation but get caught in one or more high fire exposure events (i.e., burnovers)
during their evacuation to safety. As in FN1 and FN2, high hazard conditions may require the
formation of TRAs.

48
 NIST TN 2262
August 2023

Camp Fire Example 13. Evacuation impacted by fire along egress artery.

Civilians evacuating from western Paradise (area highlighted in blue) and from points north in
Magalia who left before fire reached their part of the community were potentially caught in several
burnovers that occurred on Skyway (BO #4, BO #12; indicated by the red shaded areas) [7]. The
burnovers were a result of spot fires that ignited well ahead of the main fire. Heavy traffic from all
areas of Paradise was utilizing Skyway as an evacuation route because other egress arteries to the
east were closed due to the advancing fire. Traffic delays in combination with the rapid expansion
of spot fires led to multiple burnovers and use of TRAs during evacuations. The map above shows
the two burnover areas that affected Skyway between 08:30 and 14:00, and the fire perimeter
observed via satellite at 10:45 (yellow shaded area) [7, 79].

4.2.2.4. Fire/Notification Scenario FN4

No fire near residence, official notification → early egress, or shelter in community wildfire
safety zones without experiencing high exposures
This is the desired evacuation scenario, where evacuation orders are issued and received with
enough time to safely evacuate the civilian population before fire impacts the community or
egress routes.

49
NIST TN 2262
August 2023

4.2.2.5. Fire/Notification Scenario FN5

No fire near residence, no official notification → early evacuation
This scenario also represents a safe evacuation. This occurs when civilians are aware of a fire
event, elect to evacuate before official orders are issued, and are able to get to a safe location
without fire impacting their evacuation. If possible, this may be the best approach for susceptible
subpopulations who need more time to evacuate. In some scenarios, there is potential for
congestion and delays due to increased evacuation traffic from shadow evacuees, who are
individuals who evacuate from locations outside of those specified in evacuation orders [80, 81].

4.2.3. Evacuation Scenarios

In certain fire incidents there may be no solution that avoids fire exposures to evacuees because
the time to exposure is shorter than the time required for the entire population to reach a safety
zone or evacuate (fire/notification scenarios FN1, FN2, and FN3). Understanding these scenarios
can support the development of evacuation plans designed specifically to reduce high fire
exposures to as many residents and first responders as possible.
Once a civilian decides to evacuate, additional decisions will be needed to select an egress route
and destination, whether the destination is a wildfire safety zone or some location outside of the
fire area. Depending on local egress routes and the starting point, egress from the community
may require a longer travel distance or time than reaching a safety zone. There are several
evacuation scenarios that an individual may encounter as a fire event develops:
E1. Shelter in place.
E2. Become entrapped during evacuation to safety zone.
E3. Evacuate to safety zone.
E4. Become entrapped during evacuation from the fire area.
E5. Safe egress from the community or to a safety zone.
The flowchart in Fig. 5 illustrates these five simplified evacuation outcomes. These evacuation
scenarios are related to the fire/notification scenarios described above and are expanded on in the
following subsections to provide context for the relationships among egress options, TRAs, and
safety zones. To first order, risk of exposure in scenarios E2 and E4 is proportional to travel
distance in the presence of fire, assuming potential exposure hazards are equal. This drives the
need for a distributed community wildfire safety zone system to reduce the travel distance to safe
areas.
Table 5 summarizes evacuation and sheltering options. Early evacuation (shaded green) is the
only low hazard option—one that avoids exposure to the fire. Two options are shown for
sheltering in buildings: residences and shelters. Both options are shaded gray to indicate the large
range of potential risk based on local conditions. The next two columns describe evacuations in
hazardous conditions. Evacuations that result in entrapments and burnovers are shaded red as the
highest risk outcome, while evacuations where TRAs are formed are shaded orange indicating
there is some safety benefit of TRAs. The last column describes the shelter-in-community
option. This scenario is also shaded gray, since there is a range of possible fire exposure
scenarios depending on the placement and access of wildfire safety zones.

50
 NIST TN 2262
August 2023

Fig. 5. Flow chart depicting generalized evacuation scenarios. Red text indicates hazard.

51
 August 2023
NIST TN 2262
Table 5. Characteristics of evacuation and sheltering options.
Evacuation/ Evacuate and shelter in Evacuate and shelter in
sheltering Shelter in place in Shelter in community Evacuate in hazardous TRA – not caught in designated safe
option Evacuate early residence (SIC) conditions/burnovers burnover building
Descriptions Partial or full evacuation Residents shelter in their Residents shelter in a Entrapped in a burnover Directed by first Residents shelter in
before fire reaches home designated wildfire during evacuation responders to take shelter designated wildfire
community safety zone in TRA shelter
Life safety No exposure to fire Limit travel in Limited or no fire A TRA may be formed Reduced fire exposure No fire or smoke
enhancements potentially hazardous exposure in designated only if local conditions exposure inside specially
conditions safety zone permit engineered building
Life safety Limited hazard Can result in entrapment, Hazard with accessing Very hazardous; can May experience fire and Hazard while accessing
hazards associated with potential injuries and/or death; local Safety Zone; result in injuries and/or smoke exposures, local shelter; hazard will
high-volume traffic; may may require rescue hazard will increase with fatalities although less severe than increase with distance
experience smoke distance traveled, burnover conditions away, proximity of fire,
exposures proximity of fire, and and fuels and topography
fuels and topography between residence and
between residence and shelter location
SIC location
Travel By vehicle or mass No travel required (if at By vehicle or on foot By vehicle or mass By vehicle or mass By vehicle or mass
required transit home during incident) transit transit transit
52

Notes Road network must be Can evacuate after the First responders may No standards or design
able to accommodate the fire intensity has relocate civilians guidance exist for the
partial or full evacuation subsided or not evacuate. between TRAs during design, construction, and
before hazardous the event to address maintenance of such
conditions result in If property is prepared, safety and road capacity facilities specifically for
burnovers either in the resident is able and issues. WUI applications.
community and/or in the equipped, and exposure
egress corridors. levels permit, defensive Such facilities will be
actions may save expensive to design and
Early evacuation plans residence (although maintain and may be
must be developed in likely hazardous to beyond the reach of most
parallel with trigger residents). small communities.
points for shelter in
community. Retrofit of existing
facilities will likely also
This may be the desired be very costly.
option for mobility
impaired residents and
critical care and medical
facilities unless a shelter
in community option
exists within reach and
can be accessed with in-
house mobility options.
 NIST TN 2262
August 2023

Camp Fire Example 14. Range of exposure levels experienced at TRAs.

The photographs above show images from two very different fire scenarios experienced in TRAs
during the Camp Fire. On the left, the Pearson Road TRA was formed with high urgency in extreme
fire exposure conditions burning over civilians evacuating in their vehicles. Exposures were so high
that several vehicles ignited, and fire shelters were deployed inside a fire engine to block the
radiation. Fortunately, an unbuilt, cleared residential lot was present to provide refuge for vehicles
with support from a fire engine and dozer, which reduced the exposure levels enough to enable
survival.
The Optimo TRA, pictured on the right, was formed with less urgency in response to traffic
congestion and roadways blocked by fire. The location was at a paved parking area. With fuels set
back at a greater distance, shelter inside commercial buildings, and support from fire engines, the
exposure was less extreme than at the Pearson Road TRA.

4.2.3.1. Scenario E1: Shelter in place

In this response scenario, the resident(s) seek shelter in their residence. Three distinct outcomes
can result from this approach are discussed in this section. E1 scenarios can result in residents
being exposed to very hazardous and life-threatening conditions. They should be carefully
addressed during pre-fire evacuation planning.

E1-A: Defend the structure/property

The resident(s) made the decision to stay and defend their property well before the fire event and
invested in extensive pre-fire planning. The structure has been hardened for fire and ember
exposures and the parcel has defensible space that has been prepared and well-maintained. This
approach requires a pre-fire assessment by a subject matter expert to determine whether the
structure and parcel can be hardened to the necessary level so as to provide a safe environment

53
NIST TN 2262
August 2023

for the resident. This may not be possible, depending on local conditions, and is the reason why a
pre-fire assessment is necessary.5 To enable their structure protection capabilities and enhance
life safety, residents will likely need their own water supply, personal protective equipment
(including firefighting garments and respiratory protection), and adequate physical and mental
fitness for the task. Some of the hazards associated with the stay and defend approach are
discussed in Section 2.2. This outcome assumes that the residents do not become entrapped. A
scenario where a defended structure results in entrapment of residents is defined as scenario
E1-C, described below.

E1-B: Inadequate preparation

A homeowner may elect to stay (FN5) or be forced to stay (FN1) in an unprepared
property/structure for a number of reasons, including mobility impairment, lack of transportation,
or concerns about property loss (including an uninsured property). A homeowner may also have
a preconceived notion that their home will offer more protection than trying to evacuate during
the fire. Staying in an unprepared residence can be very dangerous and may expose residents to
life-threatening conditions when the property is exposed to fire and embers. Partial structure
hardening may also provide a false sense of security. If fire reaches these types of properties,
residents in these scenarios will likely require evacuation assistance. There may be a short time
window for first responders to reach the residence before the fire arrives and limits access to the
residence.

E1-C: Entrapment
In this scenario, the resident(s) cannot evacuate because the structure, vehicle, parcel, or
immediate surroundings are on fire. Residents will require rescues. High fire exposures may
make timely access by law enforcement and firefighters difficult or impossible. Entrapment
situations are not limited to the early stages of the event. While firefighting equipment is
designed to tolerate higher exposures than unprotected vehicles, severe conditions will also
restrict access by firefighters, limiting rapid response. Law enforcement equipment is not
designed to tolerate the same thermal exposures as fire apparatus, and law enforcement personnel
are typically not equipped with fire resistant clothing or PPE. Therefore, access into or through
locations of high fire exposure may be more limited for other first responders than it is for
firefighters.

4.2.3.2. Scenario E2: Entrapped during evacuation to safety zone

In this scenario, the resident elected to seek shelter in a safety zone. However, local conditions
deteriorated rapidly, and they became entrapped on the way. This scenario points to the need for
multiple safety zones distributed throughout the community and the need to communicate the
evacuation information to residents as quickly as possible so that they may complete evacuation
before they are impacted by fire.

5
Cases where residents should not stay include, but are not limited to, high density construction and residences near untreated wildland
fuels. HMM can provide additional context for these scenarios; however, the assessment should still be performed by a subject matter expert.

54
 NIST TN 2262
August 2023

Camp Fire Example 15. Entrapment en route to the safety zone at Camelot Meadow in Concow.

Civilians evacuating from the area circled in blue, west of the egress artery, were up to 2 km (1.2 mi)
straight line distance and 4 km (2.5 mi) driving distance away from the pre-designated Wild Fire
Safety Zone at Camelot Meadow (TRA-A, indicated with a blue square and outline). These civilians
were caught in two burnovers (BO #1 and #2, indicated with red circles and outlines) and took
shelter in two TRAs (B and C) on their way to the meadow.
Two firefighters in a pickup truck were scouting out the fire and evacuating civilians in the west
portion of Concow. Returning toward the exit (1, in yellow text), they were blocked by fire and
debris on Hoffman Road with 10 to 15 civilian vehicles following them (BO #1) (2). The firefighters
deployed fire shelters to shield civilians as they moved them to a TRA in the creek (TRA-B) while
several vehicles were igniting. A dozer was able to access the TRA and clear the obstructed roadway
(3). However, the group was unable to reach the Camelot Meadow, and instead had to take refuge
in a second TRA (C) at the intersection of Hoffman Road and Concow Road (BO #2) (4). After 24
minutes, fire activity subsided enough that they could convoy (5) to the safety zone at the meadow
to join the group already taking refuge there (6).
The two burnovers that occurred before residents could reach the designated safety zone highlights
the need for a distributed wildfire safety zone system that would reduce the travel distance between
areas of relative safety.

55
NIST TN 2262
August 2023

4.2.3.3. Scenario E3: Safe evacuation to safety zone

In this scenario, the resident safely reaches a nearby safety zone. The safety zone may be close to
their residence, but not necessarily in their direct evacuation path out of the community.

4.2.3.4. Scenario E4: Entrapment during evacuation from fire area

E4-A: Evacuation from community
This scenario can occur when a resident tries to egress directly from their home, workplace, or
other location in the community, and is caught by fire during evacuation. Like the E2 scenario,
the resident does not reach a safe area but instead gets caught in a burnover while in transit. The
density and placement of safety zones, together with the accessibility of these zones from
different parts of the surrounding community, will influence the prevalence of this scenario.

Camp Fire Example 16. Entrapment during evacuation from the fire area.

The Camp Fire presents multiple examples of civilians becoming entrapped during their attempted
evacuation from the fire. The two photos above show areas where vehicles were abandoned in the
roadway when evacuees were overcome by fire during their escape from the initial fire impact in
eastern Paradise. TRAs were formed to enhance life safety in both cases pictured, a) on Bille Road,
and b) on Pearson Road.

E4-B: Evacuation from safety zone

This scenario differs from the previous situation in that the burnover exposures could readily be
avoided by staying in the safety zone longer until conditions are safe for further evacuation.

4.2.3.5. Scenario E5: Safe evacuation from fire area

E5-A: Evacuation from community

This scenario is frequently associated with early warning and early evacuation, which can limit
the exposure of residents to hazardous conditions. This preferred scenario is achievable in a
number of fire ignition/fire spread and community evacuation scenarios.

56
 NIST TN 2262
August 2023

E5-B: Evacuation from safety zone

This scenario is similar to the previous one. First responders at the safety zone may escort or
direct residents out of the community when it is safe to do so. The enhanced situational
awareness of first responders limits the potential of encountering dangerous conditions during
further evacuation.

Camp Fire Example 17. Safe evacuation from Paradise after shelter in TRA.

Several TRAs during the Camp Fire were maintained beyond the duration of the fire exposures in
the immediate area. This was done so evacuees could wait safely until the egress routes were
confirmed passable and additional transportation could be arranged for those without vehicles. The
map above highlights two examples. Due to the numerous roadways blocked by fire, abandoned
vehicles, or other obstructions (indicated by the red × marks), several hundred evacuees took refuge
in the TRAs at the parking lots of the Paradise Plaza shopping center (TRA-K), CMA Church
(TRA-V), and Optimo restaurant (TRA-S). After first responders were able to both coordinate a group
of public transit buses to facilitate evacuation and identify a passable egress route, evacuees were
escorted in convoys from the Optimo and CMA Church to consolidate at the Paradise Plaza. This
occurred at about 16:30, indicated by the blue arrows on the map above. From there, a convoy was
led out of the fire area to Chico. The first vehicles left at 17:00, seven hours after the TRA was first
initiated. Several transit buses remained to collect later evacuees until 23:00. The evacuation routes
taken from Paradise Plaza (K) are marked by the green pathways.
Later overnight, a similar convoy evacuation event occurred from the Rite Aid TRA (AA). Between
03:30 and 04:00 on November 9, after the fire front intensely burned through Magalia, dozens of
people were escorted from Magalia to Chico on a route prescribed by first responders. This
evacuation route is indicated by the red pathway in the map above.

57
NIST TN 2262
August 2023

4.3. Relationships Among Fire Ignition, Fire Growth, and Impact to Community
The extent of the fire front reaching a community will influence the initial area that needs to be
evacuated. The conceptual diagrams in Fig. 6 illustrate idealized fire spread scenarios with
ignition locations at two distances away from the edge of a WUI community. In both cases the
fire is not contained before reaching the community. In the case where the ignition occurs near
the community, Fig. 6a, the fire front length (FFL) represents only a small fraction of the
community interface length (IL). The resulting initial impact on the community is relatively
small, with FFL/IL<<1, and the extent of the high ember flux exposure zone downwind of the
initial fire front also covers a small area of the WUI community.

Fig. 6. Idealized relationship between ignition location, a) near or b) far, from a WUI community. The fire
front and ember exposures reaching the community are illustrated. The wind is directed from left to right.
(Figure from Ref. [7]).

58
 NIST TN 2262
August 2023

In the second case, illustrated in Fig. 6b, the wildland fire ignition occurs far from the
community. The fire has enough fuel and distance to develop a fire front that represents a large
fraction of the interface length of the community. In addition to the extended length of direct fire
front assault, the high ember exposure zone represents a large fraction of the community and is
illustrated as having a longer and deeper reach into the community. This deeper reach is related
to the higher overall intensity of the fire front (assuming identical fuels, wind, and topography).
The increased area of initial high ember flux exposure has the potential to overwhelm
firefighting resources and enables the fire to rapidly establish itself throughout the community.
Looking at the relationship between the distance of the fire ignition from the community and
community size, one can visualize that there is a “sweet spot” where the fire ignites far enough
away to grow and impact the community with a full-length fire front (bottom scenario in Fig. 6),
but close enough to reduce available evacuation time. This can be considered a worst-case
evacuation scenario.

4.4. Temporal Illustration of Full Community Evacuation Scenarios

Figure 7 provides an illustration of the temporal relationship of fire and evacuation for five
scenarios in which the entire community is evacuated. To illustrate the progression of time,
columns indicate sequential, evenly spaced time intervals (a–i). Conceptually, this allows the
comparison of relative evacuation times among scenarios.
In this illustrative set of community evacuation sequences, a baseline evacuation without any fire
impacts is assumed to take four time intervals (see Scenario 4). This best-case scenario assumes
that the community has an evacuation plan and that the plan has been rehearsed by first
responders and communicated to residents. Although these assumptions are not critical to the
scenarios illustrated in the figure, the baseline total evacuation time would likely be larger if
these systems are not in place. Under each sequence, the four rows indicate:
1. the level of fire activity within the community, specified as low intensity (F-L) and high
intensity (F-H). In this illustration, the period of low fire intensity is assigned one time
interval (e), and high fire intensity is assigned two time intervals (f, g). These durations
will differ in real WUI fire event, but are kept uniform in this example to enable
comparisons between the five scenarios;
2. the status of evacuation warnings (W);
3. the status of evacuation orders (O);
4. whether evacuation is ongoing (E) or extended from the expected baseline evacuation
duration (E-E) due to traffic and complications from the fire. Colors in the evacuation
row indicate the potential fire hazard to evacuees: green is low, orange is moderate, and
red is high.
Figure 7 illustrates the impact of fire on evacuation and the benefit of getting civilians out early.
The first two scenarios, 1a and 1b, have similar outcomes and are the most hazardous for
evacuees, since there is high potential to directly expose a significant fraction of the population
to fire, smoke, and possible burnover conditions. The relevance of Scenarios 1a and 1b for
communities in high WUI fire hazard areas is that there are fire spread/evacuation conditions
under which there is insufficient time to fully evacuate the community without placing large

59
NIST TN 2262
August 2023

fractions of civilians at risk. These dire scenarios highlight the need for a second tier of
evacuation planning. Scenarios 2 and 3 expose progressively fewer civilians to hazardous
conditions. The timeline represented in Scenario 4 enables all civilians to evacuate before the fire
arrives and is the ideal evacuation outcome.
More information about each sequence is given in the sections below.

Fig. 7. Temporal representation of ignition, fire exposure, evacuation warning, evacuation order, and
evacuation. The potential fire hazard to evacuees is indicated by color: green = low, orange = moderate,
and red = high.

4.4.1. Scenarios 1a and 1b

Scenario 1 is characterized by dire situations with rapid impacts of fire on the community. In
Scenario 1a, the fire starts very close to or within the community and rapidly grows to impact
part of the community or the community as a whole. Community-wide evacuation orders are
issued shortly after ignition. Once the fire arrives, there may be a period of low intensity (F-L)
fire growth within the community that affect ongoing evacuation. Conditions deteriorate, and
high-intensity fire conditions impact the evacuation (red E). Civilians can be trapped during
evacuation if burnovers occur. Fire/smoke and possible burnovers during high fire exposure

60
 NIST TN 2262
August 2023

conditions (F-H) slow down evacuation, extending the duration of evacuations (E-E) past the
peak fire activity. Evacuation after the fire peak is further slowed by fire-related obstructions
such as downed utility lines and poles and abandoned and burned vehicles. There is significant
potential for injuries and fatalities. This scenario can also occur when the fire reaches the
community before evacuation orders have been issued, possibly because of spot fire ignitions far
ahead of fire further away from the community resulting in new fires within the community.
Scenario 1b is similar to 1a, except that the fire starts at some distance from the community. The
community is issued an evacuation warning (W) but is not ordered to evacuate until fire reaches
the community. In this case, there may be time to adjust tactics, operations, or decision-making
to take advantage of the warning time. If no adjustments are made, the net fire impact on
evacuation is similar to Scenario 1a. Total evacuation time is longer than the baseline time
requirements (i.e., Scenario 5) because civilians are impacted by fire, smoke, and potential road
closures and burnovers.

4.4.2. Scenario 2
In Scenario 2, the fire starts far away from the community, as in Scenario 1b. In this case,
however, the decision to evacuate the community is made earlier. Part of the evacuation occurs
before impact from the fire. Contraflow and other traffic management tools can be used to
expedite evacuation. A smaller percentage of the population may be impacted by smoke/fire if a
significant fraction is able to evacuate before conditions deteriorate. However, civilians are still
caught within the fire during evacuations. Total evacuation time is extended by the combination
of fire and traffic, and evacuation continues during and after peak hazard conditions.

4.4.3. Scenario 3
In Scenario 3, the fire again starts far away from the community, but the evacuation orders are
issued sooner than in Scenarios 1b and 2. In this case, an even larger fraction of the population is
able to evacuate without being impacted by the fire. Only the last quarter of the normal
evacuation window occurs within fire. The impacts of the fire extend the evacuation times of the
final evacuees.

4.4.4. Scenario 4
As with Scenarios 1b, 2, and 3, the fire starts far from the community in Scenario 4. Early orders
to evacuate provide time for the community to be fully evacuated before the fire arrives. It is
likely that only a small fraction of all possible combinations of fire ignition location, fire spread
rate, and community evacuation circumstances will result in this scenario.

61
NIST TN 2262
August 2023

This page intentionally left blank.

62
 NIST TN 2262
August 2023

5. Proposed Approach
This section presents a framework methodology to assist communities and emergency officials in
developing a comprehensive WUI fire response plan for evacuations that includes alternative life
safety measures, such as shelter-in-community plans. The framework includes substantial pre-
planning actions to mitigate the potential fire exposures civilians might encounter during
evacuations. Mitigation includes fuel management along egress arteries and the designation and
maintenance of wildfire safety zones throughout the community to be used in dire scenarios
when there is insufficient time to fully evacuate. The approach presented in the following
sections defines a set of fire–evacuation trigger zones based on WRSET and WASET, which are
determined from anticipated fire spread rates and community evacuation times. Identifying these
trigger zones or decision points before a fire incident can facilitate preparedness and training of
the local community.

5.1. Mitigating Civilian Fire Exposures During Evacuation

Two strategies for mitigating fire exposure risk potential during evacuation are fuel management
and a community system of wildfire safety zones. Their implementation supports evacuation
planning by addressing scenarios where there is insufficient time to safely evacuate the entire
community. In many cases, both strategies will likely be necessary, and they can work together
to address specific community needs and leverage local community attributes (e.g., commercial
parking lots, parks). Importantly, these two strategies are not substitutes for programs like
“Ready, Set, Go!” and fire-evacuation scenarios (like Scenario E5) where there is sufficient time
to safely evacuate.
One strategy for reducing the risk of fire exposures during evacuation is to mitigate the potential
fire hazard presented to the evacuating public by managing fuels within the community and
along egress arteries. A long-term commitment will be required to maintain the fuels within the
community and to prevent buildup or accumulation of fuels along egress arteries. The goal is to
prevent high fire exposure conditions that could potentially result in burnovers during
evacuation. Collaboration with multiple landowners to carry out fuel treatments may be required.
Treatments along egress arteries may need to reach well beyond the boundaries of the
community to ensure a continuous corridor for evacuating civilians until they reach safe
locations outside of the fire.
A second strategy is to create a distributed safety zone system within the community. The goal is
to reduce risk by enabling civilians to get to lower-hazard locations with minimal/limited travel,
thus reducing potential exposure opportunities. Travel time to these locations will be a function
of road conditions, population density, and safety zone density. Note that this second approach
calls for safety zones and not WUI fire shelters.6 While the design and use of shelters that fully
protect inhabitants from WUI fires may be a viable solution in the future, there are significant
technical/science gaps to enable their design at this time. WUI fire shelters will also be
significantly more expensive to construct and maintain and therefore may not be a readily
implementable option for many existing WUI communities.

6
A WUI fire shelter is not synonymous with an evacuation shelter.

63
NIST TN 2262
August 2023

5.2. Safety Areas – Wildfire Safety Zones and Community Sheltering Areas
(Shelter in Community)
Community wildfire safety areas provide shelter from high fire exposures. These areas can be
organized by the relative degree of protection they provide: TRAs, wildfire safety zones, and
community fire shelters. Safety areas may be used for two reasons—to take immediate refuge
from high exposure entrapments, and to manage traffic and prevent civilians from encountering
high exposures. If a safety area is within reach, residents may divert there to take refuge until
high exposure conditions along the evacuation route have improved sufficiently to continue
toward a safe location.
The safety areas discussed in this section are intended to reduce thermal exposures (radiation and
convection) to evacuees in order to prevent civilian injuries and the ignition of vehicles.
Although a reduction of fire exposures is likely in these areas, evacuees can still expect to
encounter significant smoke and ember exposures. Respiratory protection in the form of N95 or
N99 masks can help by significantly reducing particulate exposures, but these devices will not
remove the harmful gases in smoke. In addition to communication and preparedness, a
comprehensive evacuation plan should also evaluate the smoke exposure that may be incurred in
wildfire safety zones. Early community evacuation (when possible) is most likely a better option
than the use of safety areas.
The development of community safety areas should be included in the overall evacuation plan.
Their use will likely require first responder resources, which will remove firefighters from
suppression tasks to focus on the shelter-in-community location(s). In the event of a fast-moving
fire in which people are unable to evacuate in time, life safety is the priority.

5.2.1. Temporary Refuge Areas

There are fundamental differences among TRAs, wildfire safety zones, and community fire
shelters. TRAs are locations that are used in crisis situations during the event as makeshift
locations for emergency relief. They are not pre-designated areas and they may provide only
limited protection.
TRAs can be divided into two subcategories based on their use in WUI fire events. First, a TRA
may be established in response to a situation in which evacuating civilians are already trapped in
a high exposure area and a readily accessible place with reduced fire exposures needs to be
rapidly identified to shelter evacuees. Second, a TRA can be used to manage traffic and civilians
in order to prevent civilians from encountering high exposures during evacuation. For example, a
TRA can be implemented to keep evacuating civilians in a safe location (such as a large
commercial parking lot) and block traffic from entering a hazardous road section. This is a
critical traffic management tool that can be used by first responders during evacuation.

64
 NIST TN 2262
August 2023

Camp Fire Example 18. TRA use during the Camp Fire.

The NIST post-fire case study identified 31 separate TRAs that collectively provided refuge to more
than 1200 civilians during the first 24 hours of the Camp Fire [6]. The TRA locations were binned
into five categories: 14 parking lots (e.g., above right), 7 roadways, 6 structures, 3 natural areas
(e.g., roadside creek, seen above left), and 1 maintained natural area (e.g., ballfield, maintained
meadow).
The Paradise Plaza parking lot TRA in Paradise (a) and the Hoffman Road TRA at the creek crossing
in Concow (b) are seen in the pre-fire aerial imagery and post-fire photos above.
TRAs were implemented by first responders for two reasons—to take immediate refuge from high
exposure entrapments, and to manage traffic and prevent civilians from encountering high
exposures. Within the first two hours of the fire’s arrival in Paradise, multiple TRAs were formed
in roadway intersections or similar areas of last resort. See Camp Fire Example 14.
After the initial fire front, hundreds of civilians were still evacuating when egress routes were
blocked by fire and debris (abandoned and burned vehicles, downed trees and utility poles) and fire
was still burning through the town. First responders established several large TRAs in parking lots
of commercial shopping areas as places for people to wait until the roadways were safe to pass. See
Camp Fire Example 17.

5.2.2. Wildfire Safety Zones

Wildfire safety zones are pre-designated locations characterized by open space with limited or no
fuels present. They may provide additional protection relative to TRAs due to the reduced fuel
and their pre-designated status. With pre-designated safety zones, residents can be aware of
locations to seek reduced exposures within and around their community when hazardous fire and
smoke conditions may impact their safe evacuation.
The presence of wildfire safety zones does not imply that their use will result in the lowest
exposure hazards in all situations or that they should be treated as the default evacuation option.
Safety zone definition has long been an important topic for wildland firefighters, although there

65
NIST TN 2262
August 2023

are currently no U.S. standards for the design, sizing, density, or placement of community
wildfire safety zones. Based on calculations of radiative exposures exclusively from vegetative
fuels [82], a rule of thumb that can be implemented by firefighters in the field [83] is that a safe
distance is four times the expected flame height. However, additional factors influence the
exposure level, including convective heating, wind, slope, and protective clothing or shelters [84,
85]. Recent work by Campbell et al. [86] has incorporated adjustment factors for slope and wind
conditions, implemented in a GIS tool for calculation and visualization of potential safety zones
for wildland firefighters. More research is needed to further define adequate safety zones by
incorporating fuels from the exposures from the built environment (e.g., residences, commercial
buildings, and vehicles)
Wildfire safety zones might be reached by civilians on foot or in vehicles. Their design and
sizing should reflect the needs of the local population and consider the specific characteristics of
the community. The sizing of wildfire safety zones must consider the fraction of the community
being served (the number of civilians and their vehicles) and potential fire exposures from
surrounding fuels. Flashy vegetative fuels and thick forest can generate significant fire
exposures. While relatively short in duration compared to a building fire, vegetative fire
exposures must be factored into the creation of an exposure reduction buffer around the usable
core of a wildfire safety zone.
There are significant benefits in establishing wildfire safety zones well before a fire as part of a
comprehensive evacuation plan. Signage used to identify the wildfire safety zones and to direct
civilians to them can help the community become more familiar with their locations and
potential use. Civilians and mutual aid first responders will also benefit from reviewing the
evacuation plans and maps with clearly demarcated wildfire safety zones.
Communities can use a variety of existing locations as potential wildfire safety zones. Wildfire
safety zones should contain limited or no fuel and can be natural or manmade geographic
features. Examples of areas that may be evaluated for potential use as wildfire safety zones
include clearings, gravel areas and parking lots, bare earth lots, and well-maintained parks or
other irrigated green areas. Areas to avoid using as wildfire safety zones include heavily wooded
areas with understory fuels, areas near combustible structures (e.g., outdoor auxiliary features
like gazebos), areas of flashy fuels (e.g., unburned tall dry grass), and high-density residential
areas, which can ignite and result in structure-to-structure fire spread and very high exposures.
Commercial buildings can act as buffers for radiation, but ignition of the buildings and their
contents is possible even if firefighters are present. The presence of firefighters to reduce or
protect safety zones from surrounding exposures cannot be guaranteed in rapidly developing or
large-scale incidents.

66
 NIST TN 2262
August 2023

Camp Fire Example 19. Natural areas used as wildfire safety zones.

The photos above show two examples of natural area safety zones on Concow; a) Camelot Meadow
and b) Crain Memorial Park. Both locations were indicated in the existing pre-fire evacuation plans
for the Concow area and had signage indicating their intended use as public assembly points during
fire incidents.
The Camelot Meadow was minimally maintained as a 3.2 ha (8 ac) natural grass meadow; during
the Camp Fire, the safety zone was temporarily unusable while the fire burned through it.
Afterwards, an estimated 70 to 85 civilians took refuge in the burned meadow in addition to several
first responders. The photo above shows the condition of the meadow one year after the fire.
Crain Memorial Park was another natural safety zone in Concow, characterized by a maintained
field. It’s use during the Camp Fire was undetermined.

The placement of wildfire safety zones must be readily accessible to civilians in the area, and
they must avoid high-hazard locations such as topographic features like chimneys and narrow
canyons. If possible, high fuel load areas and dangerous topographic features should be avoided
between the residences and the wildfire safety zone. This may be difficult to accomplish in
intermix communities with high fuel loads and limited fire history. Special consideration should
be given to the sheltering of civilians living in high fuel load areas in the perimeter of

67
NIST TN 2262
August 2023

communities where limited time to egress may be available, and particularly when there is only a
single egress route.
The proposed intent of wildfire safety zones is to reduce fire exposures to evacuating civilians in
limited evacuation time scenarios and to get civilians to safety with the least amount of high fire
exposures. This approach calls for a very distributed system of safety zones. In that context,
having four safety zones of 2 ha (5 ac) each, distributed in high hazard intermix area, will likely
provide greater accessibility than one single zone of 8 ha (20 ac). A distributed system provides
more options and contingencies to both civilians and first responders and may reduce overall
congestion by reducing travel distances and simplifying routes to the nearest safety zone.
While new communities will have options for the placement and sizing of wildfire safety zones,
existing communities will need to leverage available opportunities that may allow rezoning or
creation of suitable parks or other open spaces. Continued maintenance of wildfire safety zones,
particularly those that utilize natural areas such as meadows, needs to be considered since fuel
buildup can impact the usability of the zone.

5.2.3. Community Fire Shelters

There are no standards currently available for the construction and maintenance of commercial
or residential buildings for use as fire shelters. Community fire shelters should be designed to
withstand ember storms and direct fire impingement. Shelters will need to consider tenability,
including conditioned and filtered ventilation, electricity, water, and meet accessibility
requirements.
Commercial and residential buildings have both been used in past WUI events to shelter
evacuating civilians. In one case, people were already located at the place of refuge (a casino)
during the fire [12]. In two others, universities enacted their shelter-in-place response and
students and faculty from across campus had to get to the designated building (gymnasium) to
seek shelter [87, 88]. In these three cases, only people already on the general premises used these
refuges, not the public at large. The Camp Fire provides a fourth example in which several
buildings, including commercial and residential structures, were used as TRAs for the general
public while being actively defended by firefighters [6]. Existing structures provide limited
protection and should not be considered as standalone fire shelters. Depending on exposure
levels, they may require active defensive measures to maintain their viability. Defensive actions
can also be used to enhance the protection provided by TRAs and wildfire safety zones.

68
 NIST TN 2262
August 2023

Camp Fire Example 20. Defensive actions at TRAs.

Just over half of the TRAs used during the Camp Fire (17/31) had the benefit of defensive actions
to support the tenability of the location and the safety of the occupants and first responders.
Defensive actions in four of the 17 cases were efforts directly related to life safety. One dramatic
example is the use of a fire engine monitor nozzle to spray over an estimated 70 to 100 civilian
vehicles assembled at the Bille Road TRA as fire burned around and over the evacuees.
In eight cases, the primary defensive actions were aimed at protecting infrastructure and commercial
buildings that were used intermittently to shelter civilians. Five cases were characterized by
exposure reduction in the area within or surrounding the TRA to reduce losses and enhance access
to the TRA. In these situations, the civilians sheltering in the TRA were not in immediate danger.
At the Optimo TRA, pictured above, there is record of at least three instances of firefighters
suppressing the ignition of the commercial building adjacent to the TRA while the area was occupied
by evacuees. The building was ultimately destroyed in the fire by another ignition after the TRA
had been evacuated.
The remaining 14 TRAs without defensive actions highlight that, in many cases, there were not
enough or the right type of resources to do so. This includes TRAs initiated by law enforcement or
fire personnel without a fire engine, limited or unavailable water for suppression, and intensity of
exposures that prevented safe action.

5.3. Developing a Coupled Fire-Evacuation System

To enable the development of a simplified and implementable evacuation plan for small and
medium size intermix and isolated interface communities, an evacuation trigger zone concept
defined by buffer zones around the community [61, 63, 64, 66, 68] is outlined in this section.
These proposed zones can be used as a basis for notification and evacuation decision-making
when coupled with fire spread information/estimates.

5.3.1. Trigger Zone Definitions

A three-zone system is presented here to support specific notification and evacuation thresholds.
Zone widths are driven by the temporal relationships between evacuation requirements and
anticipated/potential fire spread rates. Two minimum evacuation times, 𝑡𝑒𝑣𝑎𝑐 , are used to
develop the zone widths:

69
NIST TN 2262
August 2023

• Minimum time needed for Partial Community Evacuation, 𝑡𝑒𝑣𝑎𝑐,𝑃

• Minimum time needed for Complete Community Evacuation, 𝑡𝑒𝑣𝑎𝑐,𝐶
Both 𝑡𝑒𝑣𝑎𝑐,𝑃 and 𝑡𝑒𝑣𝑎𝑐,𝐶 specifically assume no direct (e.g., flames, smoke) or indirect impacts
(e.g., burned and downed trees or utility poles) of fire on evacuation but do include elapsed time
from ignition to detection, detection to assessment, decision to evacuate and notification. The
three proposed zones are named in a color-coded set—Purple, Red, and Green—arranged in
expanding areas around the community as diagrammed in Fig. 8. Ignitions or fire spread into the
different zones correspond to different courses of action and evacuation approaches.
While conceptually treated and shown as concentric areas in this discussion and figure, the true
shape will depend on fire spread rates and are expected to have irregular shapes [63, 64, 66]
influenced by wind, fuel types and loadings, and topography. Operational trigger buffers may
differ from the calculated locations to facilitate identification using specific landmarks or
geographic features [89].

Fig. 8. Conceptual illustration of three ignition zones around a WUI intermix community. Zones may be
asymmetrical because of fuels, fire history, topography, and prevailing winds. Fire spread directionality
and intensity may not be uniform from all directions towards the community.

70
 NIST TN 2262
August 2023

Purple Ignition Zone: The Purple Zone is the innermost zone. Ignitions within the Purple Zone
are close to the community and can quickly generate hazardous conditions for localized portions
of the community. Because of the proximity of the ignition to the community, there will be little
time to safely evacuate before conditions in the impacted area become unsafe. Partial evacuation
and/or shelter-in-community responses will likely be needed to reduce overall fire exposure
hazards to the civilians immediately impacted. Based on community size, layout, fuels, fire
history, topography, and prevailing winds, the Purple Zone may be small and localized or may
even be non-existent.
Red Ignition Zone: The Red Zone represents the area in which a fire ignition spreading towards
the community may not leave sufficient time to evacuate all parts of the community before the
fire arrives. Shelter-in-community was discussed in Sec. 5.2 as an approach to address the life
safety of civilians who cannot safely evacuate out of the community or the immediate hazard
area. The Red Zone outer boundary, bordering the Green Zone, is defined by 𝑡𝑒𝑣𝑎𝑐,𝐶 . The Red
Zone is distinguished from the Purple Zone in that the ignitions occur farther from the
community but still within the 𝑡𝑒𝑣𝑎𝑐,𝐶 temporal threshold. Ignitions in the Red Zone can generate
a large fire front that exposes a large part of the community to significant fire effects.
Green Ignition Zone: The Green Zone represents the region in which a fire is determined to
pose a potential threat to the entire community and there is sufficient time to conduct a full
community evacuation before the fire arrives. The inner border of the Green Zone, bordering the
Red Zone, is defined by the minimum time required for a full community evacuation 𝑡𝑒𝑣𝑎𝑐,𝐶 . An
example corresponding to a similar scenario is shown in Fig. 6b. Topography, accessibility of
ignition locations, and the anticipated chance of containment can also be used to establish the
width of the zone to identify which fires should initiate an early evacuation. The outer edge to
the Green Zone is based on fire behavior and fire spread rates. Local characteristics including
fire history and the presence of watershed/fuel breaks and other topographic features could be
used to establish the Green Zone outer boundary.
Surrounding Region: The surrounding region, illustrated by the brown area in Fig. 8, is defined
as everything beyond the Green Zone. The area beyond the immediate community trigger zones
can be viewed as monitoring of regional-scale fire activities. The proximity of a fire to the outer
edge of the Green Zone, direction of fire spread, topography, and weather influences will be used
to determine the issuance of evacuation warnings.

One approach to simplify the large number of possible ignition scenarios is to divide the zones
into quadrants or sectors. Local knowledge, historical wind directions, topography, and fire
history can be studied to understand expected directionality and create wind direction thresholds
for the different quadrants. A review of all quadrants/sectors should be performed for three
reasons:
1. Fire may occur during an unusual weather event,
2. Spot fire ignitions may result in fire “jumping over” the community and burning back
from the other direction,
3. Locally unprecedented fire behavior exceeds historical fire spread and intensity.

71
NIST TN 2262
August 2023

Camp Fire Example 21. Spot fire ignitions on Skyway and Andover Drive in Magalia.

In the early stages of the Camp Fire, several spot fires ignited west of Paradise, several miles ahead
and downwind of the main fire front. The ignition locations in canyons provided upslope fire spread
pathways against the prevailing wind, directing fire back toward the community. Local winds may
also have been affected by the canyon topography. These spot fires accelerated the timeline of fire
impacts to the community, exposing evacuees on Skyway to fire in two places. The map above
shows the fire perimeter as observed by satellite at 10:45 with the relevant spot fires and spread
directions indicated.

Data of evacuation clearance times could be collected through evacuation exercises and
supported by traffic modeling. While such data collection is non-trivial and may not represent a
realistic worst-case scenario (such as evacuating at night in smoke without streetlights), or
account for all human behavior, it may provide a realistic way to bound an absolute minimum
evacuation time. With this evacuation time in hand, the remaining part of the ignition trigger
zone development is determining the fire spread rate and direction coupled with the relationship
between ignition location and size of the community. A fire resulting from an ignition far away
may deviate slightly from its projected path and miss a small community several miles
downwind. In contrast, a fire igniting closer to and spreading towards the community will require
a significant redirection to miss the community. This concept is illustrated in Fig. 9, where fire
spread direction is altered by 15 degrees to compare the two scenarios.

72
 NIST TN 2262
August 2023

Fig. 9. Effect of fire spread deviation on community impact for ignitions near and far from a community. a)
fire spread deviation of 15° will affect whether the community is impacted, b) similar deviation will not
result in a no-impact scenario. Impacts of fuels and topography not shown.

Figure 9 illustrates how an ignition far from a community may result in scenarios where a
community may be missed, and a “false evacuation” may take place. An additional consideration
is that distant ignitions can generate larger fire fronts and more aggressive fire spread under the
right conditions.
The two illustrated scenarios are idealized. Spot fire ignitions ahead of the fire may result in
significant impacts to a community evacuation even if the spotting is several kilometers away
from the community.

73
NIST TN 2262
August 2023

Camp Fire Example 22. Humboldt Fire (2008).

The map above shows the fire history, in yellow, from 1911 to 2018, before the Camp Fire. The
Humboldt Fire is individually highlighted in red. On Wednesday, June 11, 2008, the Humboldt Fire
ignited near the northwest area of the perimeter and spread rapidly to the south and east under strong
north winds. A change in the wind direction on Friday threatened to push fire up the canyons deeper
into the town. The fire destroyed 254 structures, including 74 homes in the southern reaches of
Paradise. Nearly one-third of Paradise was evacuated during this incident, complicated by the fire
impacts to three of the four egress routes.

5.3.2. Determining Ignition Zone Widths

Widths of the trigger zones, 𝑑𝑧𝑜𝑛𝑒 , are determined by estimates of fire spread distance covered
during the time required for evacuations:
𝑑𝑧𝑜𝑛𝑒 = 𝐹𝑆𝑚𝑎𝑥 × 𝑡𝑒𝑣𝑎𝑐 (4)
where 𝐹𝑆𝑚𝑎𝑥 is the fire rate of spread and 𝑡𝑒𝑣𝑎𝑐 is the time required for community evacuation
(WRSET2). First, an assessment is made to determine the inner most part of the zone based on
ignition location, expected effective fire spread rate, and community evacuation particulars.
Then, an assessment of the outermost zone is developed. Because the potential for rapid fire
spread exists during many severe wildfire and WUI events, attention should be placed on
quantifying the expected 𝐹𝑆𝑚𝑎𝑥 .
Table 6 lists the distance from the inner boundary of the green zone and the boundary of an
intermix community based on a range of fire spread rates 𝐹𝑆𝑚𝑎𝑥 and different 𝑡𝑒𝑣𝑎𝑐 durations.
The calculations are the simple multiplication of Eq. (4); however, the values emphasize the
potentially extensive distance/area of concern to accommodate evacuations. Table 6 shows that if
a community requires two hours to evacuate (partially or fully, depending on the scenario) and

74
 NIST TN 2262
August 2023

the expected fire spread rate 𝐹𝑆𝑚𝑎𝑥 is 4 mi/h (6.4 km/h), then the inner most green zone
boundary should be set at 8 mi (12.8 km) from the boundary of the community. This scenario
outlines a case where sufficient time is provided for a community to evacuate before the fire
arrives. Considerations such as fuels along egress arteries, long range spotting, and other
conditions that may impact different parts of the community evacuation still need to be
addressed. The range of fire spread rates and evacuation times listed in the table are not
unprecedented. Recent WUI fire events have been within these bounds. For example, the
evacuation of Paradise during the Camp Fire took at least four hours, and the effective fire
spread rate from ignition to the first spot fires in Paradise of 7 mi (12 km) in 1.5 hours, or
4.6 mi/h [7].

Table 6. Green Zone inner boundary distance, in miles, from edge of intermix community. (1 mi = 1.6 km)

𝑭𝑺𝒎𝒂𝒙 (mi/h)
𝒕𝒆𝒗𝒂𝒄,𝑪 or 𝒕𝒆𝒗𝒂𝒄,𝑷
(hours) 1 2 4 6
0.5 0.5 1 2 3
1 1 2 4 6
2 2 4 8 12
4 4 8 16 24

Increasing the width of the trigger zones can be used to address uncertainty in fire spread rates
and can be viewed as engineering safety factors. Additionally, the width of zones can be used to
create temporal fire containment thresholds. Rapid fire spread under high winds or dry
conditions can result in a fire covering more than a mile in 15 minutes. Quick containment will
be dictated by time of day (impacting availability of aerial suppression), detection time,
accessibility, staffing, and environmental conditions. If an ignition in the Green Zone cannot be
contained (including spot fires), a full community evacuation should be considered when the fire
reaches the boundary of the Green/Red Zone. There is a possibility that the fire may be contained
in the Red Zone; however, the likelihood of that outcome should be weighed against the potential
for long-range spotting and other conditions impacting containment. Spotting of the fire from the
Green Zone into the Red Zone with limited chance of containment should also be considered.
The outer edge of the Red Zone intersects with the inner boundary of the Green Zone and marks
the location where complete or partial evacuation may not be accomplished before the fire
reaches the community. An ignition in the Red Zone that has potential to be contained presents a
difficult situation with respect to response actions. If the fire cannot be contained, there will be
insufficient time to safely evacuate the community. Consideration should be given to activation
of evacuation procedures in all but select cases where the ease and speed of access may make
containment highly probable. These cases should be explicitly defined, as even small reductions
in the number of available first responder resources could result in fire impacting the community
during evacuations.
The inner boundary of the Red Zone is dictated by the outer Purple Zone boundary. The Purple
Zone can be used in locations where fires will have limited impact on the community as a whole
and only partial evacuation or partial shelter-in-community may be necessary. These include

75
NIST TN 2262
August 2023

scenarios where only a small fraction of the community is impacted owing to local conditions
and layout of the community with respect to the fire spread. If such scenarios do not exist or
cannot be reliably developed for a particular community, then a Purple Zone does not have to be
used and the inner boundary of the Red Zone will abut the community boundary.
The evacuation timeline previously described in Sec. 4.2.1 can be used to facilitate the
development of the evacuation plan using trigger points. The time estimates are cumulative.

1. Ignition/first observation, 𝑡𝑑
2. Situational assessment
𝑡𝑜𝑟𝑑𝑒𝑟
3. Decision making 𝑡𝐹𝐷𝑆
4. Order to evacuate ITA
5. Communication of evacuation order, 𝑡𝐼𝑇 ITS
6. Activation of notification systems, 𝑡𝑑𝑖𝑠𝑠
7. Evacuation duration (including necessary time for evacuation
of critical care facilities/schools/hospitals), (𝑡𝑝𝑟𝑒𝑝 + 𝑡𝑡𝑟𝑎𝑛𝑠 ) ET

There is a minimum time from the observation/notification of an ignition to the activation (ITA)
of public notification systems. This time can be viewed as an operational baseline; any
evacuation time (ET) necessary for the public to reach safety must be added to the ITA time. The
total time from ignition to the time a civilian reaches safety (ignition to safety, ITS, or WRSET)
can be rewritten as the sum of ignition to activation (ITA) and the evacuation time (ET).
ITS = WRSET = ITA + ET (5)
Equation (5) can be used to assess/characterize a scenario where the evacuation time involves
civilians leaving the community or a scenario where civilians are directed to shelter in
community. The goal is to have ITS less than the time from fire ignition until fire impacts egress
arteries. While some time savings may be had with potential improvements of situational
awareness and decision making (reducing ITA), the time required to evacuate the community
will likely have the largest impact on developing the ignition zone thresholds.

5.4. Community Evacuation Options and Decisions

The risks of sheltering within the community should not be compared to the low exposure risks
associated with an early evacuation, but rather to the realistic outcomes of a specific wildfire or
WUI ignition and fire spread scenario where there is not sufficient time to safely evacuate part of
or the entire community. Not all options will not be available for all fire scenarios. Direct
comparisons among all options may not provide the necessary context for risk management.

5.4.1. Shelter in Community

Large communities with 𝑡𝑒𝑣𝑎𝑐,𝐶 times of several hours (e.g., 4 h) require significant lead time to
accomplish a safe early evacuation. Combined with fast fire spread rates (e.g., 4 mi/h), this
results in a significant Green Zone ignition radius (e.g., 16 mi) to fully evacuate. An alternative
framing is that, for a community that needs four hours to fully evacuate, any ignition within

76
 NIST TN 2262
August 2023

16 mi of the perimeter of the community should trigger a shelter-in-community response rather

than a full evacuation if:
• egress arteries will not provide the desired life safety conditions to evacuating civilians,
• designated safety zones are in place, and
• the public is informed of the shelter-in-community response.
The shelter-in-community approach provides a risk management tool for scenarios where
evacuation may expose civilians to high hazard conditions (i.e., burnovers). When conditions
permit (time is available, egress corridors are accessible and remain open) evacuation may be the
less hazardous approach. In other conditions, shelter-in-community or a combination of
approaches may enhance life safety overall. Partial evacuation may also be considered,
particularly for medical care facilities and other susceptible civilian populations.

5.4.2. Partial vs. Complete Community Evacuation

Communities can be evacuated in their entirety or partially. Partial evacuation can occur when
only part of the community will be impacted by fire or when only part of the community can
safely evacuate before fire conditions prevent continued safe evacuation. Figure 10 illustrates a
first order assessment of fire impact to the community potentially permitting a partial evacuation.
There are benefits and limitations with both full and partial evacuation options. In both cases it is
essential that clear and timely evacuation information is conveyed to the public. An advantage of
a partial evacuation or phased evacuation targeting people who will experience hazardous
conditions first is reduced traffic on the egress arteries resulting in more rapid evacuations and
reduced gridlock. A full evacuation may be easier to communicate to all residents through mass
notification systems like sirens; however, a simultaneous full community evacuation may result
in significant traffic gridlock and ultimately slow down evacuation.

Fig. 10. First order assessment of initial impact of fire from nearby ignition resulting in partial evacuation.

77
NIST TN 2262
August 2023

If there is limited time to fully evacuate before the fire arrives and heavy fuel loadings are
present along the egress arteries, there is a higher risk of civilians possibly becoming trapped by
fire during evacuation. This is particularly important for intermix communities with limited fire
history and heavy vegetative fuel loadings where high ember exposures can result in multiple
ignitions within the community ahead of the main fire front, challenging suppression capacity
and partial evacuations of the impacted areas.
Figure 11 illustrates four second order assessment scenarios that can be developed/considered
based on ignition locations, wind direction, effects of topography, and location and condition of
evacuation arteries. The number of scenarios to be developed will grow significantly in number
and complexity as the community is subdivided into more zones. The diagrams in Fig. 11
represent idealized situations of a two-zone WUI community. In all four cases, wildfire ignition
occurs near the community (in the Red Zone), offering little time for evacuation in the areas
immediately impacted by fire and embers. The areas of direct fire and ember exposures
determine the zone where shelter in community may be implemented if safe evacuation cannot
be achieved.
Figure 11a and b show how sheltering in the community may also be necessary for areas of the
community that are not directly impacted by the fire if egress routes pass though the fire’s
projected path. Figure 11c shows a partial impact to the community, where areas that are not
immediately impacted have access to a safe egress route and may evacuate without direct
impacts from the fire. Figure 11d shows a similar configuration to that of Fig. 11c, with the
inclusion of sloped terrain. In Fig. 11d, the slope may result in upslope fire spread that impacts
egress routes on the upslope egress artery.
The complexities described above will also carry through to the implementation of the response.
The reliability and time lags associated with situational assessment and changes in wind will
further complicate real time response (see Sec. 6.3). Evacuation challenges can be partly
mitigated by public involvement in the development of the evacuation plan, creating or engaging
volunteer organizations to help with the evacuation (specifically, but not limited to, traffic
management), extensive communication campaigns, and evacuation drills. The development of
social engagement strategies for dissemination of evacuation and notification plans and public
engagement of the community are beyond the scope of this report.

78
 NIST TN 2262
August 2023

Fig. 11. Four second order WUI community assessments accounting for wildfire ignition and egress artery
locations and topography.

79
NIST TN 2262
August 2023

This page intentionally left blank.

80
 NIST TN 2262
August 2023

6. Implementation
Implementation of the proposed approach is presented as three phases in the following sections:
assessment, planning, and execution. During the assessment phase, communities should collect
any existing evacuation plans, as well as various fire- and evacuation-related community and
demographic data, to evaluate the needs and capabilities of the community to respond to a WUI
fire. These results will be incorporated into the evacuation plan in the planning phase. Based on
the capacity of the community to evacuate and the potential fire behavior, trigger buffers and
decision points can be pre-planned for a range of scenarios. Once a plan is developed, the
execution phase includes continual advance planning and maintenance, in addition to any
incident response.

6.1. Assessment
The primary purpose of the assessment phase is to collect the community attribute data necessary
for the planning, development, or revision of a community notification and evacuation plan. A
WUI Community Hazard Framework is specifically built for that purpose [90]. It is likely that
much of the information identified in the WUI Community Hazard Framework is already
included in various community and local government documents; however, collecting all the
necessary data in a centralized digital location will facilitate a more comprehensive and effective
development of notification and evacuation plans. Any existing notification and evacuation
plans, along with the supporting material used for their development, should also be identified.
This data collection will enable an assessment of conditions and attributes within the community
and its surroundings. Demographic data and information regarding senior citizen and medical
care facilities will enable the identification of susceptible populations. Data on vegetative and
built environment fuel densities will provide context of possible fire behavior, particularly when
viewed together with fire history, topography, and weather statistics on high wind events.

Camp Fire Example 23. Paradise fire history.

The map figure below shows the fire perimeters of the recorded wildfires in northern Butte County
from 1911 to 2018, shaded to indicate the number of times each area burned. The red outline
indicates the extent of the Camp Fire. Forty-two percent of the Camp Fire footprint had not burned
in the last 100 years, including the area in and immediately around Paradise.
Lack of historical fire does not, by default, translate to low fire hazard. The absence of fire activity
in the intermix community contributed to significant fuels build up. The fuels built up together with
the severe lack of precipitation and strong winds all contributed to the severe fire behavior that
caused the loss of life and structure destruction.
Historically, fires in Concow had been contained upwind of Paradise before ever crossing the West
Branch Feather River canyon (approximately 240 m deep, 800 m rim to rim). During the Camp Fire,
the combination of wind and drought caused the fire to be uncontainable in Concow. Paradise was
impacted by both an ember assault that caused 30 of spot fires (see Camp Fire Example 7) and a
very intense and extensive fire front.
The absence of historical fire should not be viewed as a benefit or absence of hazard, but rather
assessed in the overall context of fuels, topography, and local weather (wind and drought).
Communities that have not regularly experienced fire and have extensive vegetative fuels
accumulation may therefore be prone to severe WUI fire events. Past fire history alone cannot solely
be used to predict the severity of future events. (continues on next page)

81
NIST TN 2262
August 2023

Details of the road network, fuel, topography, and fire history can provide context for evacuation
hazards and can be used to identify potential burnover locations. Data collection on possible
wildfire safety zones and access to these zones will also be critical in the development of the
community evacuation plan. Destination locations for evacuees should be identified and road
capacity and potential fire hazards should be documented for the entire travel corridor from the
community. Gridlocks and road restrictions can occur many miles away from the evacuating
community and can result in traffic getting backed up all the way to the community.

82
 NIST TN 2262
August 2023

Camp Fire Example 24. Impact of traffic gridlock beyond the immediate community.

The schematic above diagrams the road network around Paradise and the number of lanes available
for evacuation traffic flow in and out of the network during the Camp Fire. The nearest sizable
communities, and the locations that residents were familiar with, included the cities of Chico and
Oroville located on CA Highways 99 and 70. The four southern evacuation routes from Paradise all
merge with CA-99 or CA-70. Due to existing traffic in the neighboring communities, the restrictive
flow through urban areas, and the extent of feasible implementation of contraflow, the net result
was the reduction of 17 incoming lanes into 7 available outgoing lanes. The widespread merging
and traffic restrictions experienced outside of the fire area resulted in backups that reached from
Chico all the way back into Paradise, impacting the ability of evacuees to get out of the fire area and
compromising the life safety of evacuees.
Large-scale evacuations are often taxing on roadway networks and infrastructure. Evacuation plans
need to account for evacuee travel and need to address potential bottlenecks and restrictions even if
these occur many miles from the community being evacuated.

Information on the status and hardening of critical infrastructure should also be collected. This
includes any structure hardening and fuel treatments around critical infrastructure like water
pumping stations, telecommunication towers, and electrical distribution equipment. The purpose
of collecting this information is to identify infrastructure hardening needs and to understand the

83
NIST TN 2262
August 2023

potential failure of these critical systems. Hardening of key infrastructure systems is part of
developing a reliable notification and evacuation plan.
Information on the time needed to evacuate the community should be collected from any
previous evacuations. If limited data exists on community evacuation, an evacuation drill may
provide critical information for the development of the community evacuation plan. Evacuation
of part of the community will provide useful information on minimum evacuation times;
however, it may not necessarily identify critical traffic flow/congestion issues that may only
manifest at higher traffic flows. The primary purpose of the evacuation drill is to determine
minimum ignition to safety (ITS) timelines without any direct or indirect impacts from fire. Such
a drill will also provide a training opportunity for first responders to implement traffic
management tools like contraflow and to coordinate with relevant partner agencies. Community
participation in the development of the evacuation plan will also provide practice and training for
residents. The determination of ignition to notification activation (ITA) time can be determined
in one or more separate first responder exercises.
Information on first responder staffing, availability and timing/arrival and capacity of mutual aid
resources should also be collected. This information can then be reviewed in the context of
egress and ingress arteries. Discussion with the community and regional fire agencies will help
identity which fires and under what conditions fires may get out of control. This assessment will
be enhanced by knowledge of regional fire history, vegetive fuel distributions, and historical
weather data.

6.2. Planning
Community- and regional-specific data collected during the Assessment phase will influence the
design and implementation of the plan. This section provides a high-level overview of the
workflow for development of community notification and evacuation plans. In the U.S.,
evacuation decisions during a fire incident are typically made by the IC and executed or enforced
by a number of different agencies. Therefore, it is critical to develop the plans with and integrate
and inform all AHJs impacted before an event. The presented methodology is developed for
small and medium size intermix, and isolated interface, communities (conceptually on the order
of 30 000 residents or fewer).

6.2.1. Developing the Community Notification and Evacuation Plan

The notification plan should work hand in hand with the decisions and expectations set forth by
the evacuation plan and vice versa. The means of notifying large fractions of the community
should be identified and consider population density, demographics, and infrastructure.
Notification plans should consider opt-out, rather than opt-in, notification systems to increase
participation rates. Specific consideration should be given to the notification of critical care
facilities and groups that may need additional assistance. Infrastructure hardening throughout and
surrounding the community may be necessary to ensure a reliable and resilient notification
system, particularly if power is cut off (intentionally or accidentally). Developing contingencies
accounting for loss of power, internet, and phone services and the potential evacuation of the
Emergency Operations Center or people in other emergency management roles will result in a
more resilient notification plan.

84
 NIST TN 2262
August 2023

Community evacuations can be summarized in the following two questions:

• Under what conditions (what fire and when) should an evacuation be initiated?
• Who should be evacuated and where should they evacuate to?
To address the above two questions the development of the evacuation plan is divided into three
primary steps.
1. Identify the Green/Red Zone threshold indicating scenarios with insufficient time
to evacuate.
2. Develop evacuation scenarios for fire ignitions within the Red Zone.
3. Identify the Purple Zone adjacent to the community where ignitions may have
localized effects on only a portion of the community.

6.2.1.1. Step 1 – Identify the Green/Red Zone threshold

Data collected in the assessment stage is used to identify the threshold for the Red/Green Zones.
The ITS data is used together with estimated fire spread rates (𝐹𝑆𝑚𝑎𝑥 ) for the expected worst-
case conditions to identify the boundary of the Green and Red Zones. By establishing this
boundary there is a spatial threshold around the community that defines the early evacuation and
full community evacuation scenarios. Any fire outside of this boundary that is heading towards
the community will trigger a community-wide evacuation. The radius of the Green Zone, as
discussed previously in Sec. 5.3.1, can be increased to address uncertainties in evacuation time
and fire spread rate and can be viewed as an engineering safety factor. For reference, a zone
depth of 1 mi, for a fire that travels at 4 mi/h provides only 15 minutes of “additional”
evacuation time before the fire impacts the edge of the community. It is for this reason that a
reliable community evacuation time (ITS) needs to be determined in the assessment stage.
The exact location of the ignition in the development of the Green Zone fire scenarios is less
important than the combination of maximum fire spread rate and direction of fire spread.
Topographic features and climatology of strong winds can help identify general scenarios. As
mentioned in Sec. 5.3.1, dividing the Green Zone into sectors may help with the design process.
The probability of ignition and the likelihood of containment do not factor in the development of
the Green Zone. The first step in the development of the Green Zone is to determine the inner
boundary—the distance from a community that will allow enough time for a full evacuation. To
make the zone useful, criteria need to be further developed to identify which ignition in the
Green Zone will warrant the full community evacuation.
Wind, topography, and fuels are the primary drivers that influence fire spread. Since topography
does not change and vegetative fuel buildup is a long-term process, the primary variables that
need to be characterized are fuel moisture, wind, and firefighting response. Note that fire spread
rate is not directly included since it was prescribed for the establishment of the Green Zone/Red
Zone boundary. Local fire history and weather records can be used to establish thresholds for
fuel moisture and wind speed parameters. With respect to fire department response, two
considerations are the availability of first responders and accessibility of the ignition location.
Availability of first responders should be considered in the context of an ongoing regional fire
storm which may reduce the response capacity below typical performance levels.

85
NIST TN 2262
August 2023

The established fuel moisture, wind, and firefighting response characteristics can be used to help
triage ignitions in the Green Zone. Slower moving fires occurring during lower winds or wetter
fuel conditions will take longer to reach the community and may be successfully controlled by
aerial resources; however, these fires should not be dismissed and need to be monitored carefully
(see Sec. 6.3 on execution).

Additional Fire Considerations

There are several specific cases where additional alternative or supplementary ignition zones
might be developed. The first is the development of scenarios for lower fire spread rates, again
using the above approach. Such scenarios may provide context for non-catastrophic events.
Fires, however, can generate their own wind and gain momentum so in many ways one of the
most important components to consider for non-extreme events is the effectiveness of
suppression.
Another is the extent of the fire front when the fire reaches the community. There are many
factors that drive the extent and intensity of the fire front as a function of time. The fire front that
impacted the town of Paradise in the 2018 Camp Fire was over 1 km (1.5 mi) in length, 11 km
(7 mi) from the origin.
Reliably predicting fire spread is challenging; however, there could be certain cases where
topographic features and other natural breaks may be used to refine or create “exclusion zones”
with the Green Zone. Fires in these exclusion zones should not pose a threat to the community,
although they should be carefully monitored. The development of potential exclusion zones
should carefully consider extreme fire behavior and long-range spotting that can take place over
several miles.
Refinements of fire “restarting” after a large fuel break are beyond the scope of the initial zone
development and introduce complexities and unknowns and that may increase risk by
inadvertently underpredicting detailed fire behavior that may negatively influence evacuation
decisions.

Additional Evacuation Considerations

Consideration should be given to the evacuation of critical care facilities to avoid the need for
simultaneous use of resources for evacuation of multiple critical care facilities. Evacuation plans
should include accommodation of patients on a full community evacuation. Additional
preplanning should address evacuation assistance with the mobility impaired population.
Communities should consider the use of trained community volunteers for traffic management
and assess the potential for leveraging existing infrastructure (e.g., buses/trains) for mass
evacuations. Coordination with neighboring jurisdictions can help to avoid gridlock in
surrounding communities from impacting the evacuation from the community in the path of the
fire.

6.2.1.2. Step 2 – Develop evacuation scenarios for ignitions in the Red Zone
While evacuations ideally will take place without impacts from fire, there are scenarios where
the coupling between fire ignition/spread and time to evacuate the community will generate
potentially hazardous conditions. A fire that ignites in the Red Zone will pose an evacuation

86
 NIST TN 2262
August 2023

challenge if it spreads towards the community and cannot be contained. For ignitions at the outer
limit of the Red Zone (at the interface with the Green Zone), evacuation may be completed
without fire impacts if the assumed maximum fire spread rate is estimated correctly. Ignitions
that occur closer to the community will pose an ever-increasing evacuation risk. The goal of Step
2 is to identify and characterize scenarios where there is insufficient time to safely evacuate, and
develop evacuation solutions, or alternatives, to manage these higher exposures and to reduce the
overall fire exposures to civilians.
To develop lower risk solutions, the data and characteristics of egress arteries and the availability
of possible wildfire safety zones will need to be assessed. Egress arteries will need to be
evaluated for capacity, accessibility, and potential for burnovers (i.e., fuels and fuel setbacks).
Similarly, existing locations for the establishment of wildfire safety zones should be assessed for
size, exposures, evacuee capacity, surrounding civilian population, and accessibility.
The option of clearing and maintaining fuel reductions along egress arteries should be reviewed
together with the option of implementing a system of distributed wildfire safety zones. The
analysis of egress arteries and the implementation of a distributed wildfire safety zone system
can be used to develop the shelter-in-community response and to assess the feasibility of partial
evacuation options. Fuel treatments to enhance access to safety zones and the implementation of
community and parcel hardening programs like HMM [40] should also be considered.

6.2.1.3. Step 3 – Identify the Purple Zone

This is an optional step that addresses a specific scenario in which fire impacts a small fraction
of the community. Evacuation planners can use sectors in the Red Zone to further develop likely
community exposure scenarios (including size of wildfire front and relationships to egress
arteries) to determine if any Purple Zones can be developed which will lead to the zoning of the
community in order to accomplish partial community evacuations (or partial sheltering in
community). Fire ignitions in Purple Zones are reserved for fire ignitions near or within the
community where local conditions will contain the fire to only part of the community. Specific
consideration should be given to spot fire ignitions within the community and the availability of
resources to control the initial fire and any spot fires. Fires that ignite within the Purple Zones
may still require large (complete community) evacuations if they develop into community
conflagrations or occur during a fire storm when multiple regional fires coincide with a high
wind event and first responder resources are extended on multiple incidents.

6.2.2. Accounting for Uncertainties and Including Safety Factors

The above outlined three step process allows a community to establish preliminary boundaries
for the ignition zone boundaries as a function of ITA, ITS, and maximum expected fire spread
rate. Developing trigger conditions that reflect realistic worst-case scenarios requires high quality
inputs for ITA, ITS and 𝐹𝑆𝑚𝑎𝑥 . The first two values can be supported through a combination of
exercises. The fire spread rate and direction are the most challenging to predict. Regional fire
history in similar fuels and topography can be used to bound limits of fire spread rates. In some
cases, fire spread direction may be relatively straightforward to predict based on topography and
prevailing winds. However, the conditions during the actual event will ultimately drive the fire
and determine if the community gets impacted. There is significant value in preparing

87
NIST TN 2262
August 2023

beforehand, understanding the spatiotemporal relationships between fire ignition/fire spread and
decision-making, and developing evacuation options that reduce fire exposure risks to civilians
and first responders.
The fast fire spread rates that can occur during high wind wildland/WUI fire events, together
with the necessary time to evacuate communities, will determine the trigger zones. These
ignition zones will likely span areas that are many miles away from the community boundaries.
Care should be taken so that 𝐹𝑆𝑚𝑎𝑥 is not underestimated and to provide realistic estimates for
the time required for community evacuation, accounting for scenarios such as a nighttime
evacuation or other adverse conditions that could extend evacuation times. Ongoing roadway
construction or other temporary closures of egress arteries should be addressed in the
development of the evacuation plans, either directly or in the form of a safety factor in the ITS
value. This is particularly true for small to medium size communities with limited egress routes.
For example, if a community only has six egress lanes and construction has closed two of the six,
the impact of that reduction will be very significant. The evacuation plan should contain
provisions for revisions and adjustments based on changing egress route conditions.
Community engagement and public education are critical components of a successful
notification and evacuation system. Such efforts:

• Communicate the impact and cost of evacuations and inform the public of the risks.
• Communicate the scenarios and options/limitations so the public understands what they
should do and how they will get the necessary information.
o This will inform the public on how little time may be available in certain scenarios and
the value of being prepared (programs like “Ready, Set, Go!”).
o Create mechanisms to inform seasonal or temporary residents (including visitors) of
the notification/evacuation plans.
• Help conduct evacuation exercises.
• Garner public acceptance and support for:
o the implementation of fuel treatments along egress arteries, including on private
property, if necessary,
o infrastructure hardening,
o installation/ improvements to mass notification systems,
o creating/establishing wildfire safety zones, and
o participation in volunteer programs to manage evacuation traffic.
• Encourage planning for early evacuation of critical care facilities and other residents
requiring assistance.

6.3. Execution
The implementation and execution of the evacuation plan can be divided into three temporal
categories; during pre-planning and normal operations, just before and during high hazard
conditions (e.g., during Red Flag Warnings or days with critical fire weather), and during a
wildfire/WUI fire.

88
 NIST TN 2262
August 2023

6.3.1. Pre-Planning and Normal Operations

Activities occurring well before the fire season or high hazard conditions include:
• Maintenance of egress arteries and wildfire safety zones
• Maintenance and upgrades to first responder communication, public notification, and
traffic management equipment
• Establishment of communications channels for the dissemination of fire/notification and
evacuation information7
• Training of first responders, including publics works, law enforcement, and volunteers on
principles of WUI fire safety
• PPE for law enforcement, public works and volunteers
• Updating the evacuation plan, specifically ITS and, by extension, the Red/Green Zone
boundary based on any evacuation route alterations (maintenance/closures)
• Monitoring fire activities in the surrounding region and keeping awareness of scenarios
of reduced first responder staffing that may impact early fire containment in non-high
hazard conditions.

6.3.2. High Hazard Conditions

High hazard conditions outlined in the evacuation plan will likely include Red Flag Warnings,
high wind events, regional fire storms, or other emergencies or disasters that may deplete or
reduce local first responder resources. If high hazard conditions are forecast, AHJs should inform
the community of pending conditions and use the opportunity to communicate evacuation
scenarios and restate where evacuation data will be available. Communication with surrounding
jurisdictions located within the Green and Red Zones will be critical for rapid and effective
situation assessment in the event a fire ignites within or spreads into the zones.
Active fires outside the Green Zone that have the potential to spread into the zone should be
closely monitored for direction and rate of spread.
Communication of changing conditions to the public is critical. AJHs should inform the public
using established communication channels and keep information current.

6.3.3. During a Fire

A fire burning within the identified ignition zones will activate the emergency management
response. The evacuation and notification plan developed by first responders and community
officials, pre-event training, hardening of egress arteries, and implementation of wildfire safety
zones will provide input to facilitate and support community evacuation decision-making. The
evacuation plan, even if it is not followed exactly, because of different actual fire spread rates or
other deviations from the assumed/planned conditions, will serve as a foundation for real-time

7
These channels should be clearly conveyed to the public before high hazard events and the proposed channels should be used and remain current
during an event. This will limit/avoid conflicting information during rapidly changing conditions. Multiple unused channels may cause confusion
and may not be effective if staffing limitation will prevent them from being kept current during a severe wildland/WUI event.

89
NIST TN 2262
August 2023

decision making. Experience from training exercises, pre-fire preparations, and communication
with participating agencies will enable effective dissemination of information, resulting in
reduced exposure hazards to civilians and first responders. Continual assessment of conditions to
change/adapt evacuation thresholds, and the use of the evacuation triangle (Fig. 1), will provide
spatiotemporal context for evacuation decisions.

The following factors will impact the evacuation decisions made by the incident commander:

Fire in the Green, Red, or Purple Zone

In the event of an ignition near the community, important parameters include the location,
accessibility, time of day, availability of resources, weather, and direction and rate of spread of
the fire (towards the community or not).

Fire Containment
Rapid fire spread under high wind and dry fuel conditions can result in a fire covering more than
one mile in 15 minutes. Early containment will be dictated by time of day (influencing the
availability of aerial suppression), accessibility, staffing, fuels, topography, and local weather. If
an ignition in the Green Zone cannot be contained (including spot fires) when it reaches the
Green /Red Zone boundary, then a full evacuation should be considered. There is a possibility
that the fire may be contained in the Red Zone; however, this should be weighed against long-
range spotting and the previously described conditions impacting containment. If available,
historical data on fire containment under similar conditions may provide supporting information,
although changes in fuels (i.e., fuel loading or moisture content) may result in more aggressive
fire behavior that what has been experienced historically.

Situational Assessment and Evacuation Decisions

As the event develops, field observations (from fire department, law enforcement, and dispatch
[emergency 911 calls]) and other data streams, should be used to enhance situational assessment
and determine the current fire spread rate. Information will be communicated to partner agencies
and the public using the established channels as evacuation decisions are adapted to the current
situation. Increasing fire spread rates may require a shift in the Green/Red Boundary and call for
an evacuation when the fire is further away from the community than initially planned for.

90
 NIST TN 2262
August 2023

7. Recommendations
The concepts presented in this report can be used by AHJs and WUI communities as needed.
There are three technical recommendations aimed at informing future research that will help
enhance the development of community notification and evacuation plans.
ESCAPE R1. Understand the relationship between fire spread and duration of wind events.
This may impact evacuation projections in the future.
ESCAPE R2. Understand the relationship between wind events and effectiveness of initial
containment.
ESCAPE R3. Develop methodology for assessing the performance of wildland fire spread
models using pre-fire predictions and post-fire fire spread data.

One further recommendation to facilitate evacuations, agnostic to any particular evacuation plan,
is explicitly listed here particularly because of its potential beneficial impact and relative ease of
implementation.
ESCAPE R4. Social tools like remote work during high fire hazard weather events may
reduce road congestion and enhance evacuation if residents are already located at a common
point with family (i.e., home) and preparation time to onset of evacuation is reduced.

91
NIST TN 2262
August 2023

This page intentionally left blank.

92
 NIST TN 2262
August 2023

8. Summary
Wildland fires can impact communities quickly, posing a serious threat to life safety of residents,
evacuees, and first responders, as evidenced from numerous events within the past decade. WUI
fire events have led to rapid large-scale evacuations and have resulted in destroyed communities
and loss of life. The need for WUI community evacuations can become apparent minutes to
hours before a community is impacted by fire. The rapid onset, range of scenarios, and
complexity of WUI fire incidents and evacuations calls for dedicated pre-planning of the
emergency response and contingencies.
This report offers an overview of existing practices and concepts related to community
evacuation and alternatives including stay and defend and shelter in place. Evacuation beyond
the fire area will always be the safest; staying within the fire area, by choice or by circumstance,
can be dangerous or deadly. However, recent events show that there may not always be time to
fully evacuate the community before fire impacts it. This possibility is particularly important to
consider during pre-planning. In addition to scenarios where there is sufficient time for
evacuation, solutions should be sought for dire scenarios in which the fire impacts the
community faster than the time it will take to safely evacuate.
Community evacuations present numerous challenges, from the large number of potential fire
scenarios and variability of fire behavior to the stochastic events as an incident unfolds,
including spot fire ignitions, egress obstructions, and human behavior. Advances in computing
power are improving the capability to run complex evacuation modeling simulations to support
evacuation planning. A discussion of the many components needed to reliably predict
evacuations highlights the challenges associated with managing uncertainties and the large
number of scenarios to be considered. Limitations on the state-of-the-art fire and evacuation
models means that their outputs must be carefully interpreted within the broad scope of possible
evacuation events. The overall complexities and associated uncertainties of these models call for
a simplified general approach with a heavy emphasis on flexible and adaptive pre-planning.
Various evacuation complications and considerations are presented in this report for multiple
spatial/temporal scenarios based on findings and examples from actual WUI events, particularly
the recent Camp Fire in Butte County, CA in 2018. The lessons learned from the Camp Fire have
been used to highlight potential challenges that should be considered in the context of each
individual community and to outline various intermix community evacuation scenarios. Several
scenarios are characterized by insufficient time to safely evacuate the community before the fire
impacts evacuation, compromising life safety of evacuating civilians and responding emergency
personnel. The potential for such dire situations means that communities must have several
options available to enhance life safety when evacuation is not possible.
Just as important is the ability of communities to identify these scenarios as they occur. To
address this, a methodology to link fire spread and community evacuation actions using a set of
evacuation trigger zones is presented. Adapting the community evacuation response to the
evolving situation based on trigger zones can mitigate civilian fire exposure risk. A key need in
pre-planning is to identify critical temporal/spatial thresholds in the evacuation continuum where
complete community evacuation will not be possible before the community egress arteries are
negatively impacted by fire. The proposed methodology was developed specifically to help small
and medium size WUI communities define these zones and pre-plan for different evacuation
scenarios. The report offers a path forward for the assessment of existing communities for

93
NIST TN 2262
August 2023

evacuation and notification planning, along with considerations for developing the evacuation
plan. Also included is a discussion on executing the evacuation plan, including monitoring of
actual fires and the use of real time data to adjust the planned evacuation and notification actions
as necessary.
The report provides authorities having jurisdiction (AHJs) of small and intermediate-sized
intermix communities with context of WUI fire events that will enable them to better evaluate
different hazard reduction and risk management strategies in order to enhance the life safety of
residents and first responders. This is particularly important for communities that may not have
the resources or expertise to conduct or evaluate a more complex evacuation analysis. While
additional research is needed to optimize the concepts discussed here, the presented information
can inform communities and help develop and improve community sheltering and evacuation
planning.

94
 NIST TN 2262
August 2023

References
[1] International Association of Fire Chiefs (2023) Ready, Set, Go! Available at
www.wildlandfirersg.org. Accessed: May 2023.

[2] CAL FIRE (n.d.) Prepare for Wildfire – Ready, Set, Go! Available at
https://www.readyforwildfire.org/prepare-for-wildfire/ready-set-go/.

[3] McCaffrey S., Rhodes A., Stidham M. (2015) Wildfire evacuation and its alternatives:
perspectives from four United States' communities. International Journal of Wildland
Fire 24:170–178. https://doi.org/10.1071/WF13050

[4] Australian Capital Territory Emergency Services Agency (2009) Bushfire Survival Plan:
Prepare. Act. Survive. Available at https://esa.act.gov.au/sites/default/files/wp-
content/uploads/bushfire-survival-plan.pdf.

[5] Fire and Emergency Services Authority of Western Australia (2011) Prepare. Act.
Survive.: Your guide to preparing for and surviving the bushfire season. Available at
https://www.parliament.wa.gov.au/publications/tabledpapers.nsf/displaypaper/3814042a1
95d0712a147f5514825791a00045feb/$file/4042.pdf.

[6] Maranghides A., Link E., Mell W., Hawks S., Brown C., Walton W. (2023) A Case
Study of the Camp Fire — Notification, Evacuation, Traffic, and Temporary Refuge
Areas. NIST Technical Note 2252. (National Institute of Standards and Technology,
Gaithersburg, MD). https://doi.org/10.6028/NIST.TN.2252

[7] Maranghides A., Link E., Mell W., Hawks S., Wilson M., Brewer W., Brown C.,
Vihnaneck B., Walton W.D. (2021) A Case Study of the Camp Fire — Fire Progression
Timeline. NIST Technical Note 2135. (National Institute of Standards and Technology,
Gaithersburg, MD). https://doi.org/10.6028/NIST.TN.2135

[8] McLennan J., Ryan B., Bearman C., Toh K. (2019) Should We Leave Now? Behavioral
Factors in Evacuation Under Wildfire Threat. Fire Technology 55:487–516.
https://doi.org/10.1007/s10694-018-0753-8

[9] U.S. Department of Homeland Security (2019) Planning Considerations: Evacuation and
Shelter-in-Place – Guidance for State, Local, Tribal, and Territorial Partners. Available at
https://www.fema.gov/sites/default/files/2020-07/planning-considerations-evacuation-
and-shelter-in-place.pdf.

[10] Zimmerman C., Brodesky R., Karp J. (2007) Using Highways for No-Notice
Evacuations: Routes to Effective Evacuation Planning Primer Series. FHWA-HOP-08-
003. (U.S. Department of Transportation, Washington, D.C.). Available at
https://ops.fhwa.dot.gov/publications/evac_primer_nn/primer.pdf.

[11] Haynes K., Handmer J., McAneney J., Tibbits A., Coates L. (2010) Australian bushfire
fatalities 1900–2008: exploring trends in relation to the ‘Prepare, stay and defend or leave

95
NIST TN 2262
August 2023

early’ policy. Environmental Science & Policy 13:185–194.

https://doi.org/10.1016/j.envsci.2010.03.002

[12] Mutch R.W. (2007) FACES: The Story of the Victims of Southern California's 2003 Fire
Siege. (Wildland Fire Lessons Learned Center).

[13] Handmer J., Tibbits A. (2005) Is staying at home the safest option during bushfires?
Historical evidence for an Australian approach. Environmental Hazards 6:81–91.
https://doi.org/10.1016/j.hazards.2005.10.006

[14] Viegas D.X., Almeida M.F., Ribeiro L.M., Raposo J., Viegas M.T., Oliveira R., Alves
D., Pinto C., Jorge H., Rodrigues A., Lucas D., Lopes S., Silva L.F. (2017) O Complexo
de Incêndios de Pedrógão Grande e concelhos limítrofes, iniciado a 17 de junho de 2017.
(Centro de Estudos sobre Incêndios Florestais (CEIF/ADAI/LAETA), Coimbra,
Portugal).

[15] Diakakis M., Xanthopoulos G., Gregos L. (2016) Analysis of forest fire fatalities in
Greece: 1977–2013. International Journal of Wildland Fire 25:797–809.
https://doi.org/10.1071/WF15198

[16] Molina-Terrén D.M., Xanthopoulos G., Diakakis M., Ribeiro L., Caballero D., Delogu
G.M., Viegas D.X., Silva C.A., Cardil A. (2019) Analysis of forest fire fatalities in
Southern Europe: Spain, Portugal, Greece and Sardinia (Italy). International Journal of
Wildland Fire 28:85–98. https://doi.org/10.1071/WF18004

[17] Wall T., Brown T. (2020) Red Flag Warnings [info sheet]. (Desert Research Institute).
Available at https://www.drought.gov/sites/default/files/2020-10/RedFlagFlyer508C.pdf.

[18] Treisman R. (2023) Seeing a red flag warning in your weather app? Here's what to do.
(National Public Radio). Available at https://www.npr.org/2023/04/14/1169979511/red-
flag-warning-fire-prevention-tips.

[19] Australasian Fire and Emergency Service Authorities Council (2022) Australian Fire
Danger Rating System. Available at https://afdrs.com.au/. Accessed: May 2023.

[20] Cova T.J., Drews F.A., Siebeneck L.K., Musters A. (2009) Protective Actions in
Wildfires: Evacuate or Shelter-in-Place? Natural Hazards Review 10(4):151–162.
https://doi.org/10.1061/(ASCE)1527-6988(2009)10:4(151)

[21] Paveglio T.B., Boyd A.D., Carroll M.S. (2012) Wildfire evacuation and its alternatives in
a post-Black Saturday landscape: Catchy slogans and cautionary tales. Environmental
Hazards 11:52–70. https://doi.org/10.1080/17477891.2011.635185

[22] Paveglio T., Carroll M.S., Jakes P.J. (2008) Alternatives to Evacuation—Protecting
Public Safety during Wildland Fire. Journal of Forestry 106(2):65–70.

[23] Hart D. (2014) Lessons Learnt From the Black Saturday Bushfires. (Bushfire
Cooperative Research Centre, East Melbourne, Victoria).

96
 NIST TN 2262
August 2023

[24] Wong S.D., Broader J.C., Shaheen S.A. (2020) Review of California Wildfire
Evacuations from 2017 to 2019. UC-ITS-2019-19-b. (UC Office of the President:
University of California Institute of Transportation Studies).
https://doi.org/10.7922/G29G5K2R

[25] Toledo T., Marom I., Grimberg E., Bekhor S. (2018) Analysis of evacuation behavior in
a wildfire event. International Journal of Disaster Risk Reduction 31:1366–1373.
https://doi.org/10.1016/j.ijdrr.2018.03.033

[26] Liu S., Murray-Tuite P., Schweitzer L. (2012) Analysis of child pick-up during daily
routines and for daytime no-notice evacuations. Transportation Research Part A 46:48–
67. https://doi.org/10.1016/j.tra.2011.09.003

[27] Liu S., Murray-Tuite P., Schweitzer L. (2013) Incorporating Household Gathering and
Mode Decisions in Large-Scale No-Notice Evacuation Modeling. Computer-Aided Civil
and Infrastructure Engineering 29:107–122. https://doi.org/10.1111/mice.12008

[28] Wilkinson C., Erikesen C., Penman T. (2013) Into the firing line: civilian ingress during
the 2013 "Red October" bushfires, Australia. Natural Hazards 80:521–538.
https://doi.org/10.1007/s11069-015-1982-5

[29] Cova T.J. (2005) Public Safety in the Urban–Wildland Interface: Should Fire-Prone
Communities Have a Maximum Occupancy? Natural Hazards Review 6(3):99–108.
https://doi.org/10.1061/(ASCE)1527-6988(2005)6:3(99)

[30] Paveglio T.B., Carroll M.S., Jakes P.J. (2010) Adoption and perceptions of shelter-in-
place in California's Rancho Santa Fe Fire Protection District. International Journal of
Wildland Fire 19:677–688. https://doi.org/10.1071/WF09034

[31] Handmer J., O’Neil S., Killalea D. (2010) Review of fatalities in the February 7, 2009,
bushfires. (Bushfire CRC). Available at
https://www.bushfirecrc.com/sites/default/files/managed/resource/review-fatalities-
february-7.pdf.

[32] McCaffrey S., Wilson R., Konar A. (2018) Should I Stay or Should I Go Now? Or
Should I Wait and See? Influences on Wildfire Evacuation Decisions. Risk Analysis
38(7):1390–1404. https://doi.org/10.1111/risa.12944

[33] Strahan K., Gilbert J. (2021) Protective Decision-Making in Bushfire Part 1: A Rapid
Systematic Review of the ‘Wait and See’ Literature. Fire 4(1):4.
https://doi.org/10.3390/fire4010004

[34] Strahan K., Gilbert J. (2021) Protective Decision-Making in Bushfire Part 2: A Rapid
Systematic Review of the ‘Leave Early’ Literature. Fire 4(3):42.
https://doi.org/10.3390/fire4030042

[35] Mutch R.W., Rogers M.J., Stephens S.L., Gill A.M. (2011) Protecting Lives and Property
in the Wildland-Urban Interface: Communities in Montana and Southern California

97
NIST TN 2262
August 2023

Adopt Australian Paradigm. Fire Technology 47:357–377.

https://doi.org/10.1007/s10694-010-0171-z

[36] Cova T.J., Li D., Siebeneck L.K., Drews F.A. (2021) Toward Simulating Dire Wildfire
Scenarios. Natural Hazards Review 22(3):06021003.
https://doi.org/10.1061/(ASCE)NH.1527-6996.0000474

[37] Reynolds B.T. (2017) A History of the Prepare, Stay and Defend or Leave Early Policy in
Victoria. Ph.D. Thesis. (RMIT University, Melbourne). Available at
https://core.ac.uk/download/pdf/98662407.pdf. Accessed: April 2023.

[38] Penman T.D., Eriksen C., Blanchi R., Chladil M., Gill A.M., Haynes K., Leonard J.,
McLennan J., Bradstock R.A. (2013) Defining adequate means of residents to prepare
property for protection from wildfire. International Journal of Disaster Risk Reduction
6:67–77. https://doi.org/10.1016/j.ijdrr.2013.09.001

[39] Teague B., McLeod R., Pascoe S. (2010) 2009 Victorian Bushfires Royal Commission
Final Report (Parliament of Victoria, Melbourne), Vol. I–II. Available at
http://royalcommission.vic.gov.au/Commission-Reports/Final-Report.html. Accessed:
May 2023.

[40] Maranghides A., Link E.D., Hawks S., McDougald J., Quarles S.L., Gorham D.J., Nazare
S. (2022) WUI Structure/Parcel/Community Fire Hazard Mitigation Methodology. NIST
Technical Note 2205. (National Institute of Standards and Technology, Gaithersburg,
MD). https://doi.org/10.6028/NIST.TN.2205

[41] Blanchi R., Leonard J., Haynes K., Opie K., James M., Dimer de Oliveira F. (2014)
Environmental circumstances surrounding bushfire fatalities in Australia 1901–2011.
Environmental Science & Policy 37:192–203.
https://doi.org/10.1016/j.envsci.2013.09.013

[42] Federal Emergency Management Agency (2021) Shelter-in-Place [pictogram]. Available

at https://www.fema.gov/sites/default/files/documents/fema_shelter-in-
place_guidance.pdf.

[43] Federal Emergency Management Agency (n.d.) Protective Actions Research—Wildfire.

Available at https://community.fema.gov/ProtectiveActions/s/article/Wildfire.

[44] Department of Homeland Security (2022) Wildfires. (Updated 31 May 2022). Available
at https://www.ready.gov/wildfires. Accessed: May 2023.

[45] Rancho Santa Fe Fire Protection District (n.d.) Shelter-in-place. Available at

https://www.rsf-fire.org/shelter-in-place/. Accessed: May 2023.

[46] CFA (Country Fire Authority) (2022) Neighbourhood Safer Places. (Updated 7
December 2022). Available at https://www.cfa.vic.gov.au/plan-prepare/your-local-area-
info-and-advice/neighbourhood-safer-places. Accessed: May 2023.

98
 NIST TN 2262
August 2023

[47] CFA (Country Fire Authority) (2020) CFA Neighbourhood Safer Place – Bushfire Place
of Last Resort Assessment Guide. Available at
https://www.cfa.vic.gov.au/ArticleDocuments/533/Guideline-CFA-NSP-BPLR-
Assessment-Guideline-July-2020.pdf.aspx?Embed=Y.

[48] NSW Rural Fire Service (2017) Neighbourhood Safer Places: Guidelines for the
Identification and Inspection of Neighbourhood Safer Places in NSW. Available at
https://www.rfs.nsw.gov.au/__data/assets/pdf_file/0017/26135/NSP-Guidelines.pdf.

[49] Emergency Management Victoria (2022) Community Fire Refuges. (Updated 8 July
2022). Available at https://www.emv.vic.gov.au/responsibilities/bushfire-shelter-
options/community-fire-refuges.

[50] California Public Resources Code § 4291 (2021)

[51] National Fire Protection Association Firewise USA. Available at http://www.firewise.org.

Accessed: May 2023.

[52] City of Austin/Travis County Joint Wildfire Task Force (2014) Wildfire Ready Austin
(Austin, TX). Available at
https://www.austintexas.gov/sites/default/files/files/Watershed/wildfire/Firewise-before-
and-after-the-fire.pdf. Accessed: May 2023.

[53] Maranghides A., McNamara D., Mell W., Trook J., Toman B. (2013) A Case Study of a
Community Affected by the Witch and Guejito Fires Report: #2 – Evaluating the Effects
of Hazard Mitigation Actions on Structure Ignitions. NIST Technical Note 1796.
(National Institute of Standards and Technology, Gaithersburg, MD).
https://doi.org/10.6028/NIST.TN.1796

[54] Gwynne S.M.V., Ronchi E., Cuesta A., Villa J.G., Kuligowski E.D., Kimball A., Rein G.,
Kinateder M., Benichou N., Xie H. (2023) Roxborough Park Community Wildfire
Evacuation Drill: Data Collection and Model Benchmarking. Fire Technology 59:579–
901. https://doi.org/10.1007/s10694-023-01371-1

[55] Ronchi E., Gwynne S., Rein G., Wadhqani R., Intini P., Bergstedt A. (2017) e-Sanctuary:
Open Multi-Physics Framework for Modelling Wildfire Urban Evacuation. FPRF-2017-
22. (Fire Protection Research Foundation, Quincy, MA). Available at
https://www.nfpa.org//-/media/Files/News-and-Research/Fire-statistics-and-
reports/WUI/RFWUIEvacuationModelingFramework.pdf.

[56] Ronchi E., Wahlqvist J., Gwynne S., Kinateder M., Benichou N., Ma C., Rein G.,
Mitchell H., Kimball A. (2020) WUI-NITY: a platform for the simulation of wildland-
urban interface fire evacuation. FPRF-2020-11. (Fire Protection Research Foundation,
Quincy, MA). Available at https://www.nfpa.org//-/media/Files/News-and-Research/Fire-
statistics-and-reports/WUI/RFWUINITY.pdf.

[57] Wahlqvist J., Ronchi E., Gwynne S.M.V., Kinateder M., Rein G., Mitchell H., Benichou
N., Ma C., Kimball A., Kuligowski E. (2021) The simulation of wildland-urban interface

99
NIST TN 2262
August 2023

fire evacuation: The WUI-NITY platform. Safety Science 136.

https://doi.org/10.1016/j.ssci.2020.105145

[58] Ronchi E., Gwynne S.M.V., Rein G., Intini P., Wadhwani R. (2019) An open multi-
physics framework for modelling wildland-urban interface fire evacuations. Safety
Science 118:868–880. https://doi.org/10.1016/j.ssci.2019.06.009

[59] Kuligowski E. (2021) Evacuation decision-making and behavior in wildfires: Past

research, current challenges and a future research agenda. Fire Safety Journal
120:103129. https://doi.org/10.1016/j.firesaf.2020.103129

[60] Elhami-Khorasani N., Kinateder M., Lemiale V., Manzello S.L., Marom I., Marquez L.,
Suzuki S., Theodori M., Wang Y., Wong S.D. (2023) Review of Research on Human
Behavior in Large Outdoor Fires. Fire Technology. https://doi.org/10.1007/s10694-023-
01388-6

[61] Cova T.J., Dennison P.E., Kim T.H., Moritz M.A. (2005) Setting Wildfire Evacuation
Trigger Points Using Fire Spread Modeling and GIS. Transactions in GIS 9(4):603–617.
https://doi.org/10.1111/j.1467-9671.2005.00237.x

[62] National Hurricane Program (2023) HURREVAC. Available at www.hurrevac.com.

Accessed: May 2023.

[63] Dennison P.E., Cova T.J., Mortiz M.A. (2007) WUIVAC: a wildland-urban interface
evacuation trigger model applied in strategic wildfire scenarios. Natural Hazards
41:181–199. https://doi.org/10.1007/s11069-006-9032-y

[64] Larsen J.C., Dennison P.E., Cova T.J., Jones C. (2011) Evaluating dynamic wildfire
evacuation trigger buffers using the 2003 Cedar Fire. Applied Geography 31:12–19.
https://doi.org/10.1016/j.apgeog.2010.05.003

[65] Li D., Cova T.J., Dennison P.E. (2018) Setting Wildfire Evacuation Triggers by Coupling
Fire and Traffic Simulation Models: A Spatiotemporal GIS Approach. Fire Technology
55:617–642. https://doi.org/10.1007/s10694-018-0771-6

[66] Mitchell H., Gwynne S., Ronchi E., Kalogeropoulos N., Rein G. (2023) Integrating
wildfire spread and evacuation times to design safe triggers: Application to two rural
communities using PERIL model. Safety Science 157.
https://doi.org/10.1016/j.ssci.2022.105914

[67] Ronchi E., Wahlqvist J., Rohaert A., Ardinge A., Gwynne S., Rein G., Mitchell H.,
Kalogeropoulos N., Kinateder M., Benichou N., Kuligowski E., Westbury A., Kimball A.
(2021) WUI-NITY2: the integration, verification, and validation of the wildfire
evacuation platform WUI-NITY. FPRF-2021-13. (Fire Protection Research Foundation,
Quincy, MA). Available at https://www.nfpa.org//-/media/Files/News-and-Research/Fire-
statistics-and-reports/WUI/RFWUI-NITY2.pdf.

100
 NIST TN 2262
August 2023

[68] Kalogeropoulos N., Mitchell H., Ronchi E., Gwynne S., Rein G. (2023) Design of
stochastic trigger boundaries for rural communities evacuating from a wildfire. Fire
Safety Journal 140:103854. https://doi.org/10.1016/j.firesaf.2023.103854

[69] National Weather Service (n.d.) Hurricane and Tropical Storm Watches, Warnings,
Advisories and Outlooks. Available at https://www.weather.gov/safety/hurricane-ww.
Accessed: April 2023.

[70] Demuth J.L., Morss R.E., Morrow B.H., Lazo J.K. (2012) Creation and Communication
of Hurricane Risk Information. Bulletin of the American Meteorological Society (BAMS)
93(8):1133–1145. https://doi.org/10.1175/BAMS-D-11-00150.1

[71] Beverly J.L., Bothwell P. (2011) Wildfire evacuations in Canada 1980–2007. Natural
Hazards 59:571–596. https://doi.org/10.1007/s11069-011-9777-9

[72] Natural Resources Canada (2020) Wildland fire evacuations. (Updated 15 July 2020).
Available at https://natural-resources.canada.ca/climate-change/impacts-
adaptations/climate-change-impacts-forests/forest-change-indicators/wildland-fire-
evacuations/17787. Accessed: May 2023.

[73] National Hurricane Center (n.d.) Tropical Cyclone Climatology. Available at

https://www.nhc.noaa.gov/climo/. Accessed: April 2023.

[74] SFPE Foundation (2023) SFPE Foundation Wildand-Urban Interface Working Group
Initiative. Available at https://www.sfpe.org/foundation/wui3/wui. Accessed: May 2023.

[75] Fairchild A.L., Colgrove J., Jones M.M. (2006) The Challenge of Mandatory Evacuation:
Providing For and Deciding For. Health Affairs 25(4):958–967.
https://doi.org/10.1377/hlthaff.25.4.958

[76] Bakhshaii A., Johnson E.A. (2019) A review of a new generation of wildfire-atmosphere
modeling. Canadian Journal of Forest Research 49(6):565–574.
https://doi.org/10.1139/cjfr-2018-0138

[77] Zhao B., Wong S.D. (2021) Developing Transportation Response Strategies for Wildfire
Evacuations via an Empirically Supported Traffic Simulation of Berkeley, California.
Transportation Research Record 2675(12):557–582.
https://doi.org/10.1177/03611981211030271

[78] Link E.D., Maranghides A. (2023) Burnover events identified during the 2018 Camp
Fire. International Journal of Wildland Fire 32(6):989–997.
https://doi.org/10.1071/WF22115

[79] NASA Earth Observatory (2018) Camp Fire Rages in California. "Landsat 8 image of
Camp Fire at 10:45am local time." Available at
https://earthobservatory.nasa.gov/images/144225/camp-fire-rages-in-california.

101
NIST TN 2262
August 2023

[80] Zhao X., Xu Y., Lovreglio R., Kuligowski E., Nilsson D., Cova T.J., Wu A., Yan X.
(2022) Estimating wildfire evacuation decision and departure timing using large-scale
GPS data. Transportation Research Part D: Transport and Environment 107:103277.
https://doi.org/10.1016/j.trd.2022.103277

[81] Zhang Z., Herrera N., Tuncer E., Parr S., Shapouri M., Wolshon B. (2020) Effects of
shadow evacuation on megaregion evacuations. Transportation Research Part D:
Transport and Environment 83:102295. https://doi.org/10.1016/j.trd.2020.102295

[82] Butler B.W., Cohen J.D. (1998) Firefighter Safety Zones: A Theoretical Model Based on
Radiative Heating. International Journal of Wildland Fire 8(2):73–77.
https://doi.org/10.1071/WF9980073

[83] National Wildfire Coordinating Group (2022) NWCG Incident Response Pocket Guide
(IRPG). PMS 461 (NFES 001077). Available at
https://www.nwcg.gov/sites/default/files/publications/pms461.pdf.

[84] Page W.G., Butler B.W. (2017) An empirically based approach to defining wildland
firefighter safety and survival zone separation distances. International Journal of
Wildland Fire 26:655–667. https://doi.org/10.1071/WF16213

[85] Butler B.W. (2014) Wildland firefighter safety zones: a review of past science and
summary of future needs. International Journal of Wildland Fire 23:295–308.
https://doi.org/10.1071/WF13021

[86] Campbell M.J., Dennison P.E., Thompson M.P., Butler B.W. (2022) Assessing Potential
Safety Zone Suitability Using a New Online Mapping Tool. Fire 5(1):5.
https://doi.org/10.3390/fire5010005

[87] Weiss K.R., Chawkins S. (2018) "A long night of fear in the college gym." Los Angeles
Times, 15 November 2008. Available at https://www.latimes.com/archives/la-xpm-2008-
nov-15-me-college15-story.html. Accessed: May 2023.

[88] Mejia B. (2008) "Pepperdine University defends ‘shelter in place’ decision during
Woolsey fire." Los Angeles Times, 13 November 2018. Available at
https://www.latimes.com/local/lanow/la-me-ln-pepperdine-shelter-20181113-story.html.
Accessed: May 2023.

[89] Li D., Cova T.J., Dennison P.E. (2017) Using reverse geocoding to identify prominent
wildfire evacuation trigger points. Applied Geography 87:14–27.
https://doi.org/10.1016/j.apgeog.2017.05.008

[90] Maranghides A., Link E., Mell W., Hawks S., Wilson M., Brewer W., Brown C.U.,
Vihnaneck R., Walton W. (2021) A Case Study of the Camp Fire – Fire Progression
Timeline; Appendix C. Community WUI Fire Hazard Evaluation Framework. NIST TN
2135 Supplement. (National Institute of Standards and Technology, Gaithersburg, MD).
https://doi.org/10.6028/NIST.TN.2135sup

102
 NIST TN 2262
August 2023

Appendix A. California Large-Loss Fire Statistics

Data compiled from the California Department of Forestry and Fire Protection (CAL FIRE) for
fires in the State Responsibility Area8 of California between 2017 and 2022, listed in Table A-1,
show that a small number of fires contributed the majority of structure losses. Only 1 % of all
fires resulted in structure loss. High-loss fires, defined as incidents with 100 or more damaged or
destroyed structures, accounted for 18 % of fires with losses, but 95 % of structure losses over
the six-year period.

Table A-1. Recent WUI fire structure loss statistics in California (CAL FIRE State Responsibility Area).

Number
of fires Number of Total lossesa Percentage of losses
Number with Total large-lossb from large- Percentage of from large-loss
Year of fires losses lossesa fires loss fires large-loss firesc fires/total losses
2017 3470 41 12061 12 11565 29% 96%
(12/41) (11565/12061)

2018 3504 36 24227 4 23731 11% 98%

(4/36) (23731/24227)

2019 3086 19 802 2 549 11% 68%

(2/19) (549/802)

2020 3501 45 10621 11 10005 24% 94%

(11/45) (10005/10621)

2021 3054 35 3535 5 3036 14% 86%

(5/35) (3036/3535)

2022 — 28 876 3 544 11% 62%

(3/28) (544/876)

Total 16615 204 52122 37 49430 18% 95%

(37/204) (49430/52122)
Note: Total number of fires was not yet available for 2022.
a
total losses = damaged + destroyed structures
b
large-loss fires defined as total losses ≥100
c
number of large-loss fires divided by number of fires with losses

8
Areas of the state of California where CAL FIRE is responsible for fire suppression and prevention, rather than local or federal agencies.

103