Bus Rapid Transit and Traffic Safety Luis Antonio Lindau, Director, EMBARQ Brasil Nicolae Duduta, Associate Transport Planner, EMBARQ July 12, 2013
The impact of BRT systems on traffic safety
Factors that influence safety on BRT corridors
Relationship between safety and operational performance
Summary
Mexico City Guadalajara
Bogota Curitiba
Porto Alegre
Istanbul
Delhi
Ahmedabad
Vancouver
Brisbane
Rio, SP, BH
Pereira Cali
In Brazil it comprised:
road safety audits and inspections in 5 cities
190 km of corridors
more than 2 million pax per day
A global study
The impact of BRT systems on traffic safety
Factors that influence safety on BRT corridors
Relationship between safety and operational performance
Summary
Overall safety impact of a BRT Case study: Macrobús, Guadalajara
Calz. Independencia, 2007 (before BRT implementation)
Overall safety impact of a BRT Case study: Macrobús, Guadalajara
Reduction in the number of lanes
Shorter pedestrian crossings
Central median
Existing buses and minibuses replaced with a single operating agency
Monthly crashes before and after the implementation of the BRT
0
500
1000
1500
2000
2500
3000
3500
4000
4500
5000
0
50
100
150
200
250
Jan
-07
Mar
-07
May
-07
Jul-
07
Sep
-07
No
v-0
7
Jan
-08
Mar
-08
May
-08
Jul-
08
Sep
-08
No
v-0
8
Jan
-09
Mar
-09
May
-09
Jul-
09
Sep
-09
No
v-0
9
Jan
-10
Mar
-10
May
-10
Jul-
10
Sep
-10
No
v-1
0
Jan
-11
Mar
-11
May
-11
Mo
nth
ly c
ityw
ide
cras
hes
(e
xclu
din
g th
e B
RT
corr
ido
r)
Mo
nth
ly c
rash
es o
n t
he
BR
T co
rrid
or
Citywide
Crashes on the BRT
Before During BRT
construction During BRT
Impact on crashes by severity, per year
Impact on traffic fatalities
Annual fatalities in absence of BRT Annual fatalities with BRT Change in
fatalities
Baseline data
Modeled
baseline data Project data
Modeled
project data
Macrobus BRT 1 3.5 0.3 1 - 2.5
Impact on traffic injuries
Annual traffic injuries in absence of BRT Annual injuries with BRT Change in
injuries Baseline data Project data
Macrobus BRT 96.8 30 - 66.8
Impact on all crashes
Annual crashes in absence of BRT Annual crashes with BRT Change in
crashes Baseline data Project data
Macrobus BRT 2,341 1,010 - 1,331
5000
3194
1 BRT lane
2 general traffic lanes
Passenger per hour per direction (peak)
6
726
1 BRT lane
2 general traffic lanes
Crashes per year
Comparison between the bus lanes and the mixed traffic lanes
Overall safety impact of a BRT Case study: TransOeste, Rio de Janeiro
Av. das Americas 2010 Av. das Americas 2012
Overall safety impact of a BRT Case study: TransOeste, Rio de Janeiro
The layout of the BRT is very similar to Macrobus – center lanes, median stations, overtaking lanes
But the street is considerably wider (up to 100 m) and speed limits are as high as 80kmh
Impact on safety?
There is not yet enough data available to evaluate
TransOeste
However, the data so far show between 1.5 and 4.5
fatalities per month on Av. das Americas in 2012, some of
them involving BRT vehicles, which could mean an increase
in fatalities after BRT implementation
BRTs have the potential to significantly improve safety on
the streets where they are implemented
BRT systems where reductions in injuries and fatalities have
been observed include Macrobus (Guadalajara), Metrobus
(Mexico City), TransMilenio (Bogota), Janmarg (Ahmedabad)
Reductions range from 30% (Metrobus) to 70% (Macrobus)
But some systems have seen increases in fatalities (Delhi)
or have a concerning number of fatalities (TransOeste)
Overview of safety impacts
The impact of BRT systems on traffic safety
Factors that influence safety on BRT corridors
Relationship between safety and operational performance
Summary
Methodology
Crash frequency models
Road safety inspections
Discussions with BRT agency safety and operations staff
Factors that influence safety on BRT corridors
Statistical models that aim to explain the differences in
crash frequencies (or crash rates) at different locations (e.g.
intersections) using variables related to geometry, traffic
volumes, land use, etc.
Commonly use a Poisson or negative binomial (Poisson-
Gamma) distribution
Crash frequency models
Crash frequency models
Variables (Xi) Coefficients (α, βi) P
Annual average daily traffic (AADT, thousands of vehicles) 0.016 0.074
Total length of all approaches to the intersection (L, meters) 0.003 0.010
Average number of lanes per approach 0.334 0.000
Cross street is through street (=1 if yes, =0 otherwise) 1.142 0.029
Major T junction (=1 if yes, =0 otherwise) 0.719 0.019
Constant -3.914 0.000
N = 133, LR χ2 (prob.) = 64.62 (0.000), Log likelihood = -141.580
Variables (Xi) Coefficients (α, βi) P
Presence of a center median (=1 if yes, =0 otherwise) -0.349 0.004
Total number of approaches to the intersection (m) 0.424 0.000
Average length of approaches to the intersection (Lavg, meters) -0.008 0.036
Average number of lanes per approach 0.492 0.000
Cross street is through street (=1 if yes, =0 otherwise) 0.820 0.000
Major T junction (=1 if yes, =0 otherwise) 0.748 0.008
Constant -1.197 0.002
N = 132, LR χ 2 (prob.) = 135.76 (0.000), Log likelihood = -141.580, chibar2 (prob.) = 341.99 (0.000)
Table 1: Severe crash frequency model for Guadalajara (Poisson)
Table 2: All crash frequency model for Guadalajara (negative binomial)
Factors influencing crash frequencies
Counterflow
Counterflow lanes were strongly correlated with higher crash frequencies across all our models
Factors influencing crash frequencies
Street width and intersection size and complexity
Metrobus Line 1, Mexico City
Road width and complexity of intersections were the most important predictors of crash frequencies.
Factors influencing crash frequencies
Location of bus lanes
Central median
Shorter pedestrian crossings
Fewer mixed traffic lanes
Some 4-way intersections turned into T junctions
The safety impact of large blocks
For each additional 10 m (30’) between signalized intersections: • 2% decrease in all crashes • 3% increase in severe crashes
Wider streets are more problematic for safety
Wide, complex intersections can easily become “black
spots”
The length of crosswalks is key for pedestrian safety (for
each additional meter a pedestrian needs to cross without a
median, there is a 2 to 3% increase in pedestrian crashes)
Counterflow is the most dangerous bus lane configuration
Overview of findings from data analysis
A detailed inspection of a street with the objective of
identifying safety issues
Involves walking along the entire length of a study site (in
our case, a BRT corridor) documenting problems related to
infrastructure and road user behavior
Road safety inspections
Fatalities on BRT corridors by road user type
Fatalities by Road User
Type 0%
Pedestrians 54% Car occupants
23%
Motorcyclists 10%
Bicyclists 5% Other
8%
The safest place to be on a bus corridor is inside the bus
The most dangerous: pedestrian crossing the avenue
Crossing in mid-block
Mid-block signalized crosswalk - TransOeste
Pedestrian signal timing
Pedestrian signal timing
Pedestrian signal timing
Pedestrian signal timing
Pedestrian signal timing
Pedestrian signal timing
Pedestrian signal timing
Pedestrian signal timing
Pedestrian signal timing
Pedestrian signal timing
Pedestrian signal timing
Pedestrian signal timing
Pedestrian signal timing
Pedestrian signal timing
BRT Safety recommendations
41 41
SPEED REDUCTIONS: PEDESTRIAN CROSSINGS: INCREASED SIGNAL TIME:
To increase safety on Av. das Americas, EMBARQ proposed…
Speed at Avenida das Américas
42
Speed
• It is very difficult to control speed through signage and enforcement • Streets should be designed for their desired speed
Pedestrian bridges
Pedestrian bridges rarely work, as pedestrians prefer to cross under them
In some cases, this is due to the fear of crime (e.g. research in Cape Town
showing pedestrians are concerned of being assaulted or robbed on overpasses)
Our research shows they are not effective for safety on typical urban arterials -
no significant impact on pedestrian crash frequencies
The exception is on high-speed, high-volume roads (e.g. Autopista Norte, Bogota,
where locations with pedestrian bridges had four times fewer pedestrian crashes
Improving the BRT routes Based on EMBARQ’s recommendations…
• Stations barriers were improved to avoid unregulated and unsafe pedestrian crossings
• Install lateral barriers in the platform doors
• The city is studying to reduce speed limits from 80km/h to 60 km/h in Avenida das Américas
2011: Signalized mid-block crossing on Eje 2 Oriente, with no traffic calming. Vehicles did not stop for pedestrians.
2012: Speed hump installed before the pedestrian crossing, slowing traffic down and allowing pedestrian to cross safely.
More closely spaced signalized intersections can help avoid unsafe speeds
But drivers may disregard signals where the only conflict is with pedestrians
Speed reductions
There is often a considerable difference between how
streets are meant to be used and how people actually use
them:
Jaywalking, including crossing under pedestrian bridges,
jumping over guardrails, destroying guardrails
Crossing on red (pedestrians) or driving on red (drivers)
The key to designing safer streets is understanding road
user behavior and how the design of the infrastructure can
be better adapted to that
Overview of findings from inspections
The impact of BRT systems on traffic safety
Factors that influence safety on BRT corridors
Relationship between safety and operational performance
Summary
Most safety recommendations involved speed reductions,
traffic calming, or additional traffic signals
This raises the question of how safety countermeasures
would impact the operational performance of a BRT
We illustrate the relationship between safety and
operational performance with a case study from TransOeste
Safety, operations, and capacity
The relationship between safety and operational performance
TransOeste BRT, Rio de Janeiro, Brazil
Methodology
Identified safety issues
Proposed countermeasures
Applied a microsimulation model to test the impact of safety
countermeasures on:
Operating speeds
Travel times
Speed variability
The relationship between safety and operational performance
TransOeste BRT, Rio de Janeiro, Brazil
Main safety issues:
High speed road: 80 km/h speed limit
Few crossing opportunities for pedestrians
Jaywalking is a common problem
The relationship between safety and operational performance
TransOeste BRT, Rio de Janeiro, Brazil
Main safety recommendations:
Lowering the speed limit to 60 km/h
Adding mid-block signalized crossings
Improving signal timing to reduce pedestrian delay
The relationship between safety and operational performance
TransOeste BRT, Rio de Janeiro, Brazil
Indicator Service Baseline 60 km/h 60/30 km/h Complete
Speed (km/h) Express 32 31.5 29.6 29.6
Local 25.6 25.6 25.4 25.4
Travel time (min) Express 71 72 77 77
Local 89 89 89 89
Speed variability* Express 0.19 0.18 0.16 0.16
Local 0.16 0.15 0.15 0.16
* Speed variability is defined here as the ratio of the standard deviation to the mean commercial speed, for all
vehicles generated in the simulation. A lower speed variability coefficient indicates more reliable service.
Key finding:
Slight negative impact on operating speed and travel time
The relationship between safety and operational performance
TransOeste BRT, Rio de Janeiro, Brazil
* Speed variability is defined here as the ratio of the standard deviation to the mean commercial speed, for all
vehicles generated in the simulation. A lower speed variability coefficient indicates more reliable service.
Key findings:
Potential for significant safety benefits
No impact on passenger capacity
Magnitude of operational impacts negligible compared to safety benefits
Performance indicator
Baseline Project Change Source
All crashes n/a n/a -15% to -67% Elvik and Vaa 2004
Operating speed 32 km/h 29.6 km/h -2.4 km/h Microsimulation
Travel time 71 min 77 min + 6 min Microsimulation
Speed variability 0.19 0.16 - 0.03 Microsimulation
Capacity (pphpd) 18,800 18,800 No impact Based on Hidalgo et al. (2011)
More resources on this topic WCTR Presentation:
C4-6 Public Transport Safety (Slot 6g/Room g/13:40-15:20 Tue.)
More detailed presentation on the link between safety and operational
performance using the TransOeste case study
Publications available online:
Understanding the Road Safety Impact of High Performance BRT and Busway Design Features
Technical paper including the methodology and main findings from the
crash frequency models shown here
Available online at: http://www.brt.cl/understanding-road-safety-impact-of-
high-performance-bus-rapid-transit-and-busway-design-features-2/
Traffic Safety on Bus Corridors
Safe design concepts for BRT and Busways
Available online at: http://www.embarq.org/en/traffic-safety-bus-corridors-
pilot-version-road-test (English, Spanish, Portuguese)