Bicyclist Safety Performance Functions for a U.S. Cityactresearchgroup.weebly.com/uploads/9/0/6/2/9062811/... · Bicyclist Safety Performance Functions for a U.S. City Krista Nordback,

Bicyclist Safety Performance Functions for a U.S. City

Krista Nordback, Ph.D., P.E. Wesley Marshall, Ph.D., P.E.

Bruce Janson, Ph.D.

Civil Engineering University of Colorado Denver

Active Communities Transportation (ACT) Research Group

Overview

• Safety performance functions (SPFs)

• Data

• Bicyclist SPF

• Prioritizing intersections

• Examples

• Conclusions

Safety Performance Functions

Safety Performance FunctionsSafety Performance Functions

The relationship of traffic volume to collisions.

AADT

Co

llisi

on

s

HSM SPF for 4HSM SPF for 4--leg signalized intersectionsleg signalized intersections

5,000

AADTmajor

Pre

dic

ted

Cra

sh F

req

uen

cy

20,000

AADTmin = 1,000 10

2

Roads in New Zealand and Australia

Turner et al. TRR 2011

Collisions = 0.0086 × Vehicle Volume0.25 × Cyclist Volume0.17 × Length of Segment0.37

0

0.05

0.1

0.15

0.2

0.25

0.3

0.35

0 500 1000 1500 2000 2500

Cra

she

s p

er

year

Bicyclists per day

HSM SPF for 4HSM SPF for 4--leg signalized intersectionsleg signalized intersections

5,000

AADTmajor

Pre

dic

ted

Cra

sh F

req

uen

cy

20,000

AADTmin = 1,000 10

2

Risk Performance FunctionRisk Performance Function

0.0000

0.0001

0.0002

0.0003

0.0004

0 5,000 10,000 15,000 20,000 25,000

Co

llis

ion

s/A

AD

T

Annual Average Daily Traffic (AADT)

Highway Safety ManualHighway Safety Manual

• No bicycle Safety Performance Function

(SPF)

Bicycle Collisions at intersections =

MV Crashes X f(intersection type)

• Bicyclist traffic

not included

Constant RiskConstant Risk C

ollis

ion

s

AADB

SPF

Co

llis

ion

s/A

AD

B

AADB

RPF

Safety Performance

Function

Risk Performance

Function

Random EventsRandom Events C

ollis

ion

s

AADB

SPF

Co

llis

ion

s/A

AD

B

AADB

RPF

Collisions/AADB ~ (AADB)-1

Data

AADB at AADB at 106 106

Signalized Signalized Intersections Intersections

in in Boulder, COBoulder, CO

Annual Average Daily Bicyclists (AADB)

• 3 hour manual counts adjusted by – Time of year

– Day of week

– Year

– University schedule

– Temperature

– Solar radiation

• Adjustment based on continuous bicycle count data at 25 stations from 2001 to 2011

AADB vs ADB

0

500

1000

1500

2000

2500

3000

3500

4000 A

vera

ge D

aily

Bic

yclis

ts

The 25 Intersections with Highest AADB

ADB 2003 AADB Upper Limit Lower Limit

Pyramid of InteractionsPyramid of Interactions

Safe Interactions

Potential for Collisions

Avoided Collisions

Collisions Occur Minor Injury Collisions

Fatal and Severe Injury

Collisions

Bicycle Collisions at Intersections in Boulder, COBicycle Collisions at Intersections in Boulder, CO

20082008--20112011 20012001--20052005

Bicyclist SPF

Safety Performance FunctionSafety Performance Function

7 8

4

0

5 6

11 10

AADT

Collisions

AADB

Moving AverageMoving Average

0

2

4

6

8

10

12

0 500 1000 1500

Bic

yclis

t-M

oto

rist

Inte

rse

ctio

n

Co

llisi

on

s in

5 Y

ears

Intersection AADB estimated for 2003

20 pt moving average

2001 to 2005 SPF2001 to 2005 SPF

Collisions = eε (AADT) β1 (AADB) β2

Parameter Estimate Standard Error

One Tailed P-Value

Wald 95% Confidence

Limits

Intercept Exponent

ε -9.07 1.85 <0.0001 -12.7 -5.4

AADT Exponent

β1 0.64 0.17 <0.0001 0.31 0.97

AADB Exponent

β2 0.53 0.14 0.0001 0.26 0.79

Dispersion 1/ν 0.54 0.17 0.28 1.01

Pseudo R2 = 0.26


0

30,000

60,000 0

2

4

6

8

10

0 400 800 1200 1600 2000

AADT

Co

llis

ion

s

AADB

8-10

6-8

4-6

2-4

0-2

0

2

4

6

8

10

12

0 500 1,000 1,500 2,000

Bic

yclis

t-M

oto

rist

Inte

rse

ctio

n

Co

llisi

on

s in

5 Y

ears


70,000 AADT

40,000 AADT

10,000 AADT


Risk Performance FunctionRisk Performance Function

0.00

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0 500 1,000 1,500 2,000

Colli

sion p

er

Bic

yclis

t


100,000

AADT

50,000

AADT

10,000

AADT

Intersection Prioritization

Intersection with 50,000 AADT Intersection with 50,000 AADT and 1000 AADBand 1000 AADB

0

2

4

6

8

10

12

0 500 1,000 1,500 2,000

Bic

yclis

t-M

oto

rist

Inte

rsect

ion

Colli

sions

in 5

Year

s


50,000

AADT

Intersection with 50,000 AADT Intersection with 50,000 AADT and 1000 AADBand 1000 AADB

0

2

4

6

8

10

12

0 500 1,000 1,500 2,000

Bic

yclis

t-M

oto

rist

Inte

rsect

ion

Colli

sions

in 5

Year

s


50,000

AADT

3

2001 to 2005 Intersection Bicyclist Collision Rank

Examples

Policy ScenarioPolicy Scenario

0

2

4

6

8

10

12

0 500 1,000 1,500 2,000

Bic

yclis

t-M

oto

rist

Inte

rsect

ion

Colli

sions

in 5

Year

s


50,000

AADT

3/1000 =

0.003 Collisions/cyclist


0

2

4

6

8

10

12

0 500 1,000 1,500 2,000

Bic

yclis

t-M

oto

rist

Inte

rsect

ion

Colli

sions

in 5

Year

s


50,000

AADT

3/1000 =


1/140 =

0.007 collisions/cyclist


0

2

4

6

8

10

12

0 500 1,000 1,500 2,000

Bic

yclis

t-M

oto

rist

Inte

rsect

ion

Colli

sions

in 5

Year

s


50,000

AADT

3/1000 =



0

2

4

6

8

10

12

0 500 1,000 1,500 2,000

Bic

yclis

t-M

oto

rist

Inte

rsect

ion

Colli

sions

in 5

Year

s


50,000

AADT

3/1000 =


4/1700 =


Lower Motorist VolumesLower Motorist Volumes

0

2

4

6

8

10

12

0 500 1,000 1,500 2,000

Bic

yclis

t-M

oto

rist

Inte

rsect

ion

Colli

sions

in 5

Year

s


100,000

AADT

50,000

AADT

10,000

AADT

3/1000 =


1/1000 =


ConclusionsConclusions Bicycle volumes are needed to predict crashes.

• AADB & AADT significantly predict collisions. • Cyclists are safer in numbers. • Method for prioritizing intersections • Policy implications

Future Work

• Improve AADB estimates. • Method to measure infrastructure safety.

Questions?Questions?

Krista Nordback [email protected] 303-556-5247 Wesley Marshall [email protected]

mailto:[email protected]

mailto:[email protected]

Committee Chair:

Bruce Janson, Transportation Engineering

Committee Members:

Jake Kononov, Transportation Engineering

Kevin Krizek, Planning

Wes Marshall, Transportation Engineering

Debbi Main, Health and Behavioral Science

Keith Molenaar, Civil Engineering

Funding Provided by:

National Science Foundation (IGERT)

American Association of University Women

Centers for Disease Control and Prevention

Association of Schools of Public Health

Federal Highway Administration

Colorado Department of Transportation

Women’s Transportation Seminar

Intersection PrioritizationIntersection Prioritization

0 2 4 6 8 10 12

Arapahoe AV. & 30th

Walnut Ave & 30th Street

Arapahoe Ave & 33rd street

Foothills Pkwy. & Arapahoe

Baseline Road & 30th St

Folsom St & Colorado

University & Broadway

Valmont St. & Airport

Arapahoe & 55th St

1719 Canyon Blvd

28th Street & Arapahoe Avenue

Arapahoe & 9th St.

Marine & Arapahoe

Bicycle-Related Collisions 2001-2005

0

2

4

6

8

10

12

0 500 1,000 1,500 2,000

Bic

yclis

t-M

oto

rist

Inte

rsec

tio

n C

olli

sio

ns

in 5

Yea

rs


Intersections with 55,000-80,000 AADT

70,000 AADT

0

2

4

6

8

10

12

0 500 1,000 1,500 2,000

Bic

ycl

ist-

Moto

rist

Inte

rsec

tion

Coll

isio

ns

in

5 Y

ears


Intersections with

55,000-80,000 AADT

70,000 AADT

0

2

4

6

8

10

12

0 500 1,000 1,500 2,000

Bic

ycl

ist-

Moto

rist

Inte

rsec

tion

Coll

isio

ns

in 5

Yea

rs


Intersections with

25,000-55,000 AADT

40,000 AADT

0

2

4

6

8

10

12

0 500 1,000 1,500 2,000

Bic

ycl

ist-

Moto

rist

Inte

rsec

tion

Coll

isio

ns

in 5

Yea

rs


Intersections with 0-

25,000 AADT

10,000 AADT

Collisions per Intersection

Bicyclist Safety Performance Functions for a U.S. Cityactresearchgroup.weebly.com/uploads/9/0/6/2/9062811/... · Bicyclist Safety Performance Functions for a U.S. City Krista Nordback,

Documents