NISTIR 7425

Accounting for Emergency Response in

Building Evacuation: Modeling
Differential Egress Capacity Solutions

Jason D. Averill
Weiguo Song
 NISTIR 7425

Accounting for Emergency Response

in Building Evacuation: Modeling
Differential Egress Capacity Solutions
Jason D. Averill
Building and Fire Research Laboratory

Weiguo Song, Ph.D.

Guest Researcher
Building and Fire Research Laboratory
and
State Key Laboratory of Fire Science, USTC, China

April 2007

U.S. Department of Commerce

Carlos M. Gutierrez, Secretary

Technology Administration
Robert Cresanti, Under Secretary of Commerce for Technology

National Institute of Standards and Technology

 Accounting for Emergency Response in Building Evacuation:
Modeling Differential Egress Capacity Solutions

Abstract: The impact of firefighter response on the progress of the building evacuation is
not typically considered. Responders use of the stairs while occupants are evacuating can
significantly increase total building evacuation time. To account for emergency
response, this analysis considered whether adding capacity through extra stairwell width
was equivalent to providing the same total egress capacity through an additional
stairwell. An egress simulation with a counterflow submodel was calibrated against
recent fire-drill experimental results to demonstrate the capability of the model to
produce meaningful evacuation results. The model was then applied to a hypothetical
50 story office building with 350 occupants per floor. When comparing equivalent total
width, additional stairwells outperform wider stairwells from the perspective of
evacuation performance, as well as firefighter ascent times. A third stairwell can
completely mitigate the effect of firefighter response or even improve the building
evacuation time compared to two stairwells with no firefighter response.

Background: The ICC Terrorism Resistant Buildings (TRB) Committee has proposed a
change to Section 403 of the International Building Code (IBC), which would require one
additional stairwell (one greater than otherwise required) for all high-rise buildings (other
than R-2) taller than 420 ft (128 m). The proposal would provide greater egress capacity
than currently required, recognizing that one stairwell may become unusable during
evacuation due to the introduction of smoke and heat and blockage by fire hoses once
suppression operations begin.

Objective: Perform computer egress modeling to provide quantitative comparisons of

different stairwell configurations with and without emergency response interaction.

Model Description
The computer model is a modified form of a biased random walk model without back
step.1 The model is defined in a two-dimensional grid, of which each site can be occupied
by a pedestrian or be empty. Each pedestrian can move to a neighboring site with certain
probabilities. Usually, every pedestrian has a drift to move to the preferential site. All the
possible configurations of downward walkers are demonstrated in Figure 1. The
movement probabilities can be calculated with Equation 1.
 (a) (b) (c)

(d) (e) (f)

(g) (h)

Figure 1: Basic Movement Rules

Pf is the probability of moving forward; Pl is the probability moving leftward; Pr is the

probability of moving rightward; and Ps is the probability of stopping. The values of
probabilities corresponding to Figure 1 are as follows:

(a) Pf=0.8 Pl=0.05 Pr=0.15 Ps=0.0

(b) Pf=0.8 Pl=0.05 Pr=0.0 Ps=0.1
(c) Pf=0.8 Pl=0.0 Pr=0.15 Ps=0.05
(d) Pf=0.8 Pl=0.0 Pr=0.0 Ps=0.2
(e) Pf=0.0 Pl=0.05 Pr=0.15 Ps=0.8
(f) Pf=0.0 Pl=0.0 Pr=0.15 Ps=0.85
(g) Pf=0.0 Pl=0.05 Pr=0.0 Ps=0.95
(h) Pf=0.0 Pl=0.0 Pr=0.0 Py=1.0

Note that all of the probabilities sum to 1.0, as shown in

(Eq. 1.)

Pf +Pl +Pr +Ps = 1.0 (Eq. 1)

Stairwell evacuation also has its own characteristics, including right-side walking
tendency and firefighter avoidance. People in a stairwell have a right-side walking
tendency. There are three reasons for a right-side walking requirement in the model. First,
right-side walking is consistent with the traffic rules in the U.S. with which people are
familiar. Second, the staircase descends and rotates in clockwise direction, so right-side
walking is inner-side walking, which results in shorter walking distance relative to left-
side walking. Third, for firefighters, adopting a right-side walking strategy will avoid
collisions with occupants. So both pedestrians and fire fighters walk with a right-side
tendency. Fire fighters walk on the left side or middle only if the right side of the stair is
blocked by occupants. Firefighter avoidance means occupants make room for firefighters
when they encounter each other. That is a significant difference from the traditional
definition of counterflow where pedestrians with opposite walking direction compete
with each other for the walking opportunities.

Model Validation
The model was calibrated against data collected from an unannounced evacuation drill
observed in a six-story office building.2 During the evacuation, two stairwells (designated
“Stairwell A” and “Stairwell B”) were observed. The stairwells were in separate,
neighboring wings. The floor areas serving the stairs were nominally identical, with the
same number of elevators, stairwells, and exterior exit doors. The stairwells in each wing
were equally accessible from all rooms and floors. Both stairwells deposited occupants
into a lobby through a set of double doors, where they subsequently made their way
outside.

Figure 2 shows the geometry of the stairwell and tread. The stairwell had a width of
3.25 m and a length of 7.09 m. Occupants exited the stairwell through a 1.73 m wide
double exit door. In other floors, occupants entered the stairwell through a 0.91 m wide
exit door and merge with those from upper floors. In both Stairwells there were eight
steps per flight. The only exception was in Stairwell B between floor 2 and floor 1. From
floor 2 to floor 1.5 there were six steps, and from floor 1.5 and floor 1 there were 10
steps; this averaged out to eight steps for each flight. The steps in both stairwells had the
same tread and depth dimensions. The rise of each stair was 0.2 m (8 in), and the tread
was 0.28 m (11.1 in). The diagonal distance of each stair was 0.35 m (13.7 in).

There was no firefighter counterflow in Stairwell A. The walking speed of occupants

was determined by Figure 4, where the average speed is plotted against different
occupant density. The free walking speed, or maximum walking speed, is the speed of an
occupant when there is no obstacle nor occupants surrounding him. In fact, it is the speed
when occupant density is close to 0. In both Stairwell A and Stairwell B, the average
walking speed shows an approximately monotonic decrease with the increase of occupant
density. It is possible to obtain the free walking speed with the measured relation between
walking speed and occupant density.
 0.91m

1.73m
1.93m
2.44m

3.25m

(b)
3.13m

0.35m

0.20m
Down
1.52m

0.28m

3.25m
(c)
(a)
(a) stairwell over second floor (b) first floor (c)tread geometry
Figure 2: Floor and Tread Geometry

0.9
Wing A
Wing B
0.8

0.7
average speed (m/s)

0.6

0.5

0.4

0.3

0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8
2
density (p/m )

Figure 3: Occupant Speed Versus Density

The drill was conducted in 2005 on a sunny, warm, and clear day. DV-cameras were
used to record the egress process. The timeline of the egress drill was as follows.

• At t = 0 s, the fire alarm was activated.

• At t = 80 s, the first group of firefighters was sent into Stairwell B.
 • At t = 149 s, the second group of firefighters was sent into Stairwell B.
• The egress drill was completed after all people went out.

55
Egress Drill
50
Model (average of 10 runs)
45

40
Occupant number inside
35

30

25

20

15

10

-5
0 200 400
Time /s

(a) Stairwell A (no counterflow)

80 Egress drill
Model Run (average of 10 runs)
70

60
Occupants Number inside

50

40

30

20

10

-10
-50 0 50 100 150 200 250 300 350 400
Time /s

(b) Stairwell B (with counterflow)

Figure 4: Change of occupant number inside stairwell with time

Figure 4 shows the change in the number of occupants inside the stairwell as a function
of time for both stairwells. The entrance time for an individual occupant to the stairwell
was an input condition to the model. The occupant movement speeds, interactions with
other occupants and responders (if any) were calculated using the model. The differences
in occupant density between Stairwells A and B due to the firefighter counterflow were
well captured by the model.

Modeling Scenario: The objective of this study was to understand the differences, if any,
between two commonly discussed strategies for increasing egress capacity: adding
additional width to existing stairs, or adding an additional stair. Therefore, a challenging
evacuation scenario was developed in order to clearly demonstrate any differences in
evacuation times. The modeling scenario selected was a high-rise office building with a
relatively high occupant load per floor. The model building was a 50 story high-rise
office building with 350 persons per floor. The evacuation mode was full-building
evacuation and input parameters were chosen to emphasize the effect of the primary
variable (stair width) on evacuation time. It was assumed that: (a) there were no
occupant evacuation initiation delays, (b) there were no mobility impaired occupants, and
(c) that occupants left by the nearest available exit.

A baseline scenario, where firefighters do not use the stairwells until the building is
evacuated, was calculated to quantify the impact of the emergency response on the
evacuation time. Subsequently, evacuation times were calculated varying three basic
parameters: stair width, response scenario, and location of the fire in the building.
Additionally, the ascent time for the firefighters is calculated. Note that ascent time does
not include the time that it would take to stage and mount an attack on the fire floor, nor
the physical impact that the ascent has on the overall ability of the firefighter to
subsequently fight the fire.

Evacuation times were calculated for two stair width configurations: (a) a two person
wide stair (roughly 1100 mm (44 in) clear width); and (b) a three person wide stair
(roughly 1700 mm (66 in) clear width). The results were also calculated for two fire
department response conditions. In the first scenario, referred to as the ‘counterflow’
scenario, firefighters ascend the stairs against the flow of the descending occupants until
they reach the fire floor, when the stairwell is taken out of service for use by occupants
above the fire floor. Occupants above the fire floor in the counterflow stairwell are
assumed to immediately exit the stairwell and transfer to the nearest available stairwell.
The counterflow scenario assumes that a new group of four firefighters enters the
stairwell every 5 min and begins to climb the stairs. Every 20 floors, the firefighters stop
and rest for 5 min (not in the stairwell) before resuming their climb to the fire floor.† In
the second scenario, referred to as the “closed stairwell’ scenario, firefighters close one
stairwell to occupant use after 10 min in order to maintain a dedicated attack stairwell.
Again, occupants are assumed to immediately exit the attack stairwell and enter the
nearest available stairwell. The firefighter ascent times are determined by the time when
the first firefighter reaches the fire floor. Finally, three different fire locations were
selected within the building. The fire high in the building was located on floor 50. The
fire in the middle of the building was located on floor 25. The fire low in the building
was located on floor 5.

†
This estimate was based on reports from the NIST WTC Investigation. See NCSTAR 1-8 “The
Emergency Response Operations” pp. 89 – 91, available at http://wtc.nist.gov. Individual firefighter climb
rates may vary and resting may or may not be necessary.
Results:
Table 1 shows the results of the baseline evacuation modeling scenarios. National model
codes would currently require two 1100 mm (44 in) stairwells in a 50 story office
building with an occupant load of 350 persons per floor (unless travel distance
requirements required an additional stair). Under ideal evacuation circumstances
(described above) and assuming that the fire service did not utilize the stairways during
the occupant evacuation time period, the building would evacuate in approximately
179 min (three hours). In the event that one of the two stairwells was out of service, the
evacuation time would approximately double (although this scenario was not modeled).
Next, the benefits of requiring additional egress capacity were evaluated. Two stairs
providing three flow lanes each (approximately 1700 mm (66 in) clear width) would
decrease the building evacuation time by approximately 40 min, or 22 %, compared to
two stairs each providing two flow lanes. Alternatively, three stairs, each providing two
lanes of flow (approximately 1100 mm (44 in) clear width), would decrease the building
evacuation time by 61 min, or 34 %.

Table 1: No fire in the building, full building evacuation (baseline)

Scenario Occupant Evacuation Time (min)
Two 44 in stairs 179
Two 66 in stairs 141
Three 44 in stairs 119

The results in Table 1 do not account for the impact of the emergency response on the
building evacuation time. Assuming that the firefighters do not use the elevators, they
are likely to employ one of two strategies in order to initiate suppression activities:
walking up the stairs against the flow of descending occupants (counterflow) and closing
the attack stair when reaching the fire floor; alternatively, the fire service may close one
of the stairwells to occupant use upon arrival (assumed to be 10 min). The impact of the
two fire service strategies, along with the fire location (high, medium, and low in the
building) and the two stairway configurations, on the overall evacuation time is
summarized in Table 2. Figure 5Figure 7 show the same information graphically.
 Table 2: Summary of Modeled Evacuation and Ascent Times
Firefighter Ascent Time Occupant Descent Time
Scenario (min) (min)
Closed Closed
Counterflow Counterflow
Stairwell Stairwell
Floor 50

Two 66 in stairs 129 41 161 274

Fire on

Three 44 in stairs 108 41 136 174

Floor 25

Two 66 in stairs 67 25 221 274

Fire on

Three 44 in stairs 61 25 156 174

Two 66 in stairs 12 11 N/A 274

Floor 5
Fire on

Three 44 in stairs 12 11 N/A 174

 Figure 5: Evacuation and Response Times for Fire at Floor 50
No Emergency Response Occupant Evacuation Time with Emergency Response Firefighter Ascent Time
300

Two 66 inch stairs

250
Two 44 inch stairs
baseline occupant Three 44 inch stairs
evacuation
200
Time (minutes)

150

100

50

0
Occupant - Occupant - Occupant - Close Firefighter - Firefighter - Close
Baseline Evac Counterflow Evac Stairwell Evac Counterflow Stairwell Ascent
Time Time Time Ascent Time Time

Figure 6: Evacuation and Response Times for Fire at Floor 25

No Emergency Response Occupant Evacuation Time with Emergency Response Firefighter Ascent Time
300

Two 66 inch stairs

250
Two 44 inch stairs
baseline occupant
Three 44 inch stairs
evacuation
200
Time (minutes)

150

100

50

0
Occupant - Occupant - Occupant - Close Firefighter - Firefighter - Close
Baseline Evac Counterflow Evac Stairwell Evac Counterflow Stairwell Ascent
Time Time Time Ascent Time Time

Figure 7: Evacuation and Response Times for Fire at Floor 5

No Emergency Response Occupant Evacuation Time with Emergency Response Firefighter Ascent Time
300

Two 66 inch stairs

250 Two 44 inch stairs
baseline occupant
evacuation Three 44 inch stairs
200
Time (minutes)

150

100

50
N/A
0
Occupant - Occupant - Occupant - Close Firefighter - Firefighter - Close
Baseline Evac Counterflow Evac Stairwell Evac Counterflow Stairwell Ascent
Time Time Time Ascent Time Time
The impact of counterflow on total evacuation time is dependent upon the location of the
fire and the configuration of the stairwell. The reason that the fire low in the building has
a more significant impact on the building evacuation time than a fire higher in the
building is due to the fact that the firefighters close the stairwell for use above the fire
floor once they arrive at the fire floor.

Table 3: Impact of Emergency Response on Evacuation Time*

Difference in Evacuation Time
Difference in Evacuation Time
Comparing Alternative Stair
Comparing Same Stair
Configuration With Emergency
Configuration With and Without
Scenario Emergency Response
Response to Two 44 in Stairs
Without Emergency Response
Closed Closed
Counterflow Counterflow
Stairwell Stairwell
Two 66 in
14 % 94 % -11 % 52%
Floor 50
Fire on

stairs
Three 44 in
14 % 46 % -24 % -3 %
stairs
Two 66 in
57 % 94 % 23% 52 %
Floor 25
Fire on

stairs
Three 44 in
31 % 46 % -13% -3 %
stairs
Two 66 in
N/A 94 % N/A 52 %
Fire on
Floor 5

stairs
Three 44 in
N/A 46 % N/A -3 %
stairs
* Positive indicates longer evacuation times and negative indicates shorter evacuation times.

Accounting for emergency response in high-rise buildings may significantly increase the
expected building evacuation time. Table 3 shows percent increase in total building
evacuation time when the impact of emergency response is calculated compared to
building evacuation time when the emergency response is neglected. Closing a stairwell
after 10 min has the effect of proportionally increasing the evacuation time: removing
one of the two stairwells nearly doubles the evacuation time (94% increase). For a
building with three stairwells, removing one-third of the stairwells increases the
evacuation time by 46%.

The building evacuation time for a fire very low in the building in the counterflow
scenario will asymptote to the ‘closed stairwell’ scenario since the ascent time
approaches zero. The percentage impact of counterflow on occupant evacuation time for
a fire high in the building (where no occupants are forced to change stairwells) is similar
for both stairwell configurations (two 1700 mm (66 in) and three 1100 mm (44 in) stairs)
at 14 %; however, as the baseline occupant evacuation time for the three stair
configuration is shorter than the baseline occupant evacuation time for the two stair
configuration (119 min versus 141 min), the total impact is less for the three stair
configuration.

Therefore, the rightmost two columns in Table 3 show the impact of emergency response
when the calculations are normalized on a common basis (in this case, 179 min for two
1100 mm (44 in) stairs with no emergency response). Three 1100 mm (44 in) stairwells
with emergency responders will result in faster occupant evacuation times than two 1100
mm (44 in) stairs without emergency responders (as shown by the negative values for
evacuation times for three 1100 mm (44 in) stair rows in Table 3). Two 1700 mm (66 in)
stairs with emergency responders will result in longer evacuation times than two 1100
mm (44 in) stairs without emergency responders, with the exception of the fire high in the
building and firefighter counterflow scenario because there are no occupants above the
fire floor when the firefighters arrive.

Emergency response strategy is a trade-off between minimizing the arrival time of the
firefighters to the fire floor in order to conduct rescue and suppression activities and
minimizing the occupant evacuation time. The strategy of closing a stairwell upon arrival
of the fire department will approximately double the evacuation time if only two
stairways are otherwise available (141 min with both 1700 mm (66 in) stairs available
compared to 274 min if one stairwell is removed from occupant service after 10 min). If
three stairways are available, the strategy of closing a stairwell will increase the occupant
evacuation time by 46 % (119 min with three 1100 mm (44 in) stairs available compared
to 174 min if one stairwell is removed from occupant service after 10 min). On the other
hand, closing one stairwell can dramatically reduce the response time for the emergency
responders for fires higher in the building; for a response to a fire on floor 50, it would
take firefighters 129 min to reach the fire floor for a 1700 mm (66 in) stairway with
counterflow versus 41 min for an open 66 in (1700 mm) stairway; 108 min for a 1100
mm (44 in) stairway with counterflow versus 41 min for an open 1100 mm (44 in)
stairway.

If the counterflow strategy is employed by the firefighters, the impact of the occupants on
firefighter response time is dependent upon the location of the fire in the building. For
fires high in the building (floor 50), two 1700 mm (66 in) stairs require 19 % longer
evacuation time (129 min versus 108 min) than three 1100 mm (44 in) stairs. For fires in
the middle of the building (floor 25), two 1700 mm (66 in) stairs require 10 % greater
evacuation time (67 versus 61 min) than three 1100 mm (44 in) stairs. For a fire on floor
5, no simulation was performed.

Conclusions:
Subject to the assumptions and limitations of this simplified analysis, it is apparent that if
the fire department uses a stairwell (either by removing it from occupant use or walking
upward against the flow of the occupants) for emergency response prior to completion of
the evacuation, an additional stairwell restores all or more of the capacity lost to fire
department suppression operations. Further, when comparing equivalent total width,
additional stairwells outperform wider stairwells from the perspective of evacuation
performance. No cost analysis was performed for this study.

Acknowledgments: The authors wish to acknowledge the contributions of Jessica

Kratchman, Erica Kuligowski, and Richard Peacock, NIST.

1
“Evacuation behaviors at exit in CA model with force essentials: A comparison with social force model.”
Song, W., Yu, Y., Wang, B., and Fan, W. Physica A, Vol. 371 (2006) pp. 658-666.
2
“Emergency Responder Counterflow in Stairwells.” Averill, J. and Kratchman, J. Partners for
Protection: Fire Protection Engineers and the Fire Service. Oct. 17 and 18, 2006, Ellicott City, Md.