! " " # $ % & ' ' ( % ) * '
+ , - $ * & % &
. / 0 1 $ 2 3 4 5 $ 2 , 1 $ 6 7 4 8 3 4 9 7 : $ , 1 $ 6 , 4 ; $ 3 8 $ ; / 7 $ < , = 0 8 3 4 > 0 , $ ? @ . A $ ? 4 3 B 4 , 9 $ 3 8 $ ; / 7 $
C > 0 D 7 4 1 0 ; - $ 3 8 $ < , = 0 8 3 4 > 0 , E $ 0 > $ F 3 3 6 7 4 , ; 0 3 > $ 2 0 ; / $ ; / 7 $ " ; , ; 7 $ 3 8 $ < , = 0 8 3 4 > 0 , $ G H 1 0 > 7 1 1 E $
. 4 , > 1 6 3 4 ; , ; 0 3 > E $ , > : $ A 3 H 1 0 > B $ @ B 7 > F - E $ I 7 6 , 4 ; 9 7 > ; $ 3 8 $ . 4 , > 1 6 3 4 ; , ; 0 3 > E $ , > : $ ; / 7 $
C > 0 ; 7 : $ " ; , ; 7 1 $ I 7 6 , 4 ; 9 7 > ; $ 3 8 $ . 4 , > 1 6 3 4 ; , ; 0 3 > E $ J 7 : 7 4 , = $ A 0 B / 2 , - $ @ : 9 0 > 0 1 ; 4 , ; 0 3 > K
. / 7 $ F 3 > ; 7 > ; 1 $ 3 8 $ ; / 0 1 $ 4 7 6 3 4 ; $ 4 7 L 7 F ; $ ; / 7 $ D 0 7 2 1 $ 3 8 $ ; / 7 $ , H ; / 3 4 1 $ 2 / 3 $ , 4 7 $ 4 7 1 6 3 > 1 0 M = 7 $
8 3 4 $ ; / 7 $ 8 , F ; 1 $ , > : $ ; / 7 $ , F F H 4 , F - $ 3 8 $ ; / 7 $ : , ; , $ 6 4 7 1 7 > ; 7 : $ / 7 4 7 0 > K $ . / 7 $ F 3 > ; 7 > ; 1 $ : 3 $ > 3 ; $
> 7 F 7 1 1 , 4 0 = - $ 4 7 L 7 F ; $ ; / 7 $ 3 8 N F 0 , = $ D 0 7 2 1 $ 3 4 $ 6 3 = 0 F 0 7 1 $ 3 8 $ ; / 7 $ " ; , ; 7 $ 3 8 $ < , = 0 8 3 4 > 0 , K $ . / 0 1 $
4 7 6 3 4 ; $ : 3 7 1 $ > 3 ; $ F 3 > 1 ; 0 ; H ; 7 $ , $ 1 ; , > : , 4 : E $ 1 6 7 F 0 N F , ; 0 3 > E $ 3 4 $ 4 7 B H = , ; 0 3 > K
J 0 > , = $ O 7 6 3 4 ; $ 8 3 4 $ . , 1 5 $ P 4 : 7 4 $ Q * * *
CALIFORNIA PATH PROGRAM
INSTITUTE OF TRANSPORTATION STUDIES
UNIVERSITY OF CALIFORNIA, BERKELEY
! " # $ % & ' ( & $ ) * + , - % . / 0 * % 1 , % * 2 # % # / 3 & $ # * 4 & - ( - 5 # * , $ " *
6 # " # 0 % / & , $ 7 ! $ 8 . 5 8 # " * 9 . 5 5 & 0 & . $ * : , % # 0 * % 1 , % * ; ' ' # - % *
4 & - ( - 5 # * , $ " * 6 # " # 0 % / & , $ * 2 # 3 , $ " * , % * < = 5 % & 7 > , $ # *
: . = $ " , ? . = % 0
@ 9 4 7 ! A B 7 6 : : 7 C D E D 7 F G
9 , 5 & ' . / $ & , * 6 ; A H * : # 0 # , / - 1 * : # I . / %
> & $ " 0 , ( * B J * ; / $ . 5 " K * # % * , 5 J
CALIFORNIA PARTNERS FOR ADVANCED TRANSIT AND HIGHWAYS
!
UC Berkeley Safe Transportation Research & Education Center
University of California, Berkeley
Task Order 6222
Identifying Factors that Determine Bicyclist and
Pedestrian- Involved Collision Rates and Bicyclist and
Pedestrian Demand at Multi- Lane Roundabouts
FINAL REPORT
_________________________
May 2010
Prepared by
UC Berkeley Safe Transportation Research & Education Center
Lindsay S. Arnold, MPH1
Aimee Flannery, PhD, PE2
Lauren Ledbetter3
Tierra Bills4
Michael G. Jones5
David R. Ragland, PhD, MPH6
Laura Spautz7
Prepared for:
CALIFORNIA DEPARTMENT OF TRANSPORTATION
Division of Research & Innovation
1 UC Berkeley SafeTREC
2 George Mason University
3 Alta Planning & Design
4 UC Berkeley SafeTREC
5 Alta Planning & Design
6 UC Berkeley SafeTREC
7 Ibid.
FINAL REPORT: TASK ORDER 6222
STATE OF CALIFORNIA DEPARTMENT OF TRANSPORTATION
TECHNICAL REPORT DOCUMENTATION PAGE
TR0003 ( REV. 10/ 98)
1. REPORT NUMBER
CA10- 1093
2. GOVERNMENT ASSOCIATION NUMBER
3. RECIPIENT’S CATALOG NUMBER
5. REPORT DATE
May 2010
4. TITLE AND SUBTITLE
Identifying Factors that Determine Bicycle and Pedestrian- Involved
Collision Rates that Affect Bicycle and Pedestrian Demand at Multi-
Lane Roundabouts 6. PERFORMING ORGANIZATION CODE
7. AUTHOR( S)
Lindsay S. Arnold, Aimee Flannery, Lauren Ledbetter, Tierra
Bills, Michael G. Jones, David R. Ragland, Laura Spautz
8. PERFORMING ORGANIZATION REPORT NO.
UCB- ITS- PRR- 2010- 34
10. WORK UNIT NUMBER
193
9. PERFORMING ORGANIZATION NAME AND ADDRESS
University of California, Berkeley
Safe Transportation Research & Education Center ( SafeTREC)
11. CONTRACT OR GRANT NUMBER
Contract 65A0208 Task Order 6222
13. TYPE OF REPORT AND PERIOD COVERED
12. SPONSORING AGENCY AND ADDRESS
California Department of Transportation
Division of Research and Innovation, MS- 83
1227 O Street
Sacramento CA 95814
14. SPONSORING AGENCY CODE
15. SUPPLEMENTAL NOTES
16. ABSTRACT
This project examined the safety and demand issues for pedestrians and bicyclists at multi- lane roundabouts
through a literature review, case studies, in- field counts and surveys, focus groups, and video analysis. This
document presents research findings, synthesizes current information on best practices, and makes
recommendations to assist local agencies planning and designing safer multi- lane roundabouts. These findings
should help local agencies and Caltrans create roundabouts that better and more safely address the needs of
bicyclists and pedestrians. The current literature is referred to throughout the document to augment the
research team’s findings, especially for issues that were beyond the scope of this project. Key findings in the
areas of pedestrian and bicyclist avoidance of, behavior around, and collisions at multi- lane roundabouts are
presented along with recommendations for geometric design, design speed, sight distance, width of lanes,
signage and pavement markings, and operational recommendations.
17. KEY WORDS
roundabouts, multi- lane, pedestrian, bicycle, bicyclist,
collisions, demand, behavior
18. DISTRIBUTION STATEMENT
No restrictions. This document is available to the
public through the National Technical Information
Service, Springfield, VA 22161
19. SECURITY CLASSIFICATION ( of this report)
Unclassified
20. NUMBER OF PAGES
75
21. PRICE
N/ A
Reproduction of completed page authorized
2
FINAL REPORT: TASK ORDER 6222
DISCLAIMER STATEMENT
This document is disseminated in the interest of information exchange. The contents of this report reflect the
views of the authors who are responsible for the facts and accuracy of the data presented herein. The
contents do not necessarily reflect the official views or policies of the State of California or the Federal
Highway Administration. This publication does not constitute a standard, specification or regulation. This
report does not constitute an endorsement by the Department of any product described herein.
For individuals with sensory disabilities, this document is available in Braille, large print, audiocassette, or
compact disk. To obtain a copy of this document in one of these alternate formats, please contact: the
Division of Research and Innovation, MS- 83, California Department of Transportation, P. O. Box 942873,
Sacramento, CA 94273- 0001.
3
FINAL REPORT: TASK ORDER 6222
Acknowledgements
This work was supported by a grant from the California Department of Transportation ( Caltrans). The
authors would like to thank Maggie O’Mara, J. D. Bamfield, Janet Barlow, Roya Hassas, and Fred Yazdan for
their feedback and input.
4
! " # $ % & ' $ (
!
! " # $ % & ' ( ) * + , % * - . / # 0 * 1 % , " * % $ . 2 * , 3 % . 0 1 % 1 * / . 0 1 % # $ $ 4 * $ % 2 ( ' % & * 1 * $ , ' # . 0 $ % . 0 1 % 5 # + 3 + 6 # $ , $ % . , % / 4 6 , # 7
6 . 0 * % ' ( 4 0 1 . 5 ( 4 , $ % , " ' ( 4 8 " % . % 6 # , * ' . , 4 ' * % ' * 9 # * : ; % + . $ * % $ , 4 1 # * $ ; % # 0 7 2 # * 6 1 % + ( 4 0 , $ % . 0 1 % $ 4 ' 9 * 3 $ ; %
2 ( + 4 $ % 8 ' ( 4 & $ ; % . 0 1 % 9 # 1 * ( % . 0 . 6 3 $ # $ < % % ! " # $ % 1 ( + 4 / * 0 , % & ' * $ * 0 , $ % ' * $ * . ' + " % 2 # 0 1 # 0 8 $ ; % $ 3 0 , " * $ # = * $ %
+ 4 ' ' * 0 , % # 0 2 ( ' / . , # ( 0 % ( 0 % 5 * $ , % & ' . + , # + * $ ; % . 0 1 % / . > * $ % ' * + ( / / * 0 1 . , # ( 0 $ % , ( % . $ $ # $ , % 6 ( + . 6 % . 8 * 0 + # * $ %
& 6 . 0 0 # 0 8 % . 0 1 % 1 * $ # 8 0 # 0 8 % $ . 2 * ' % / 4 6 , # 7 6 . 0 * % ' ( 4 0 1 . 5 ( 4 , $ < % % %
%
! " * $ * % 2 # 0 1 # 0 8 $ % $ " ( 4 6 1 % " * 6 & % 6 ( + . 6 % . 8 * 0 + # * $ % . 0 1 % ? . 6 , ' . 0 $ % + ' * . , * % ' ( 4 0 1 . 5 ( 4 , $ % , " . , % 5 * , , * ' % . 0 1 %
/ ( ' * % $ . 2 * 6 3 % . 1 1 ' * $ $ % , " * % 0 * * 1 $ % ( 2 % 5 # + 3 + 6 # $ , $ % . 0 1 % & * 1 * $ , ' # . 0 $ < % % ! " * % + 4 ' ' * 0 , % 6 # , * ' . , 4 ' * % # $ %
' * 2 * ' ' * 1 % , ( % , " ' ( 4 8 " ( 4 , % , " * % 1 ( + 4 / * 0 , % , ( % . 4 8 / * 0 , % , " * % ' * $ * . ' + " % , * . / @ $ % 2 # 0 1 # 0 8 $ ; % * $ & * + # . 6 6 3 % 2 ( ' %
# $ $ 4 * $ % , " . , % : * ' * % 5 * 3 ( 0 1 % , " * % $ + ( & * % ( 2 % , " # $ % & ' ( ) * + , < % % A * 3 % 2 # 0 1 # 0 8 $ % # 0 % , " * % . ' * . $ % ( 2 % & * 1 * $ , ' # . 0 %
. 0 1 % 5 # + 3 + 6 # $ , % . 9 ( # 1 . 0 + * % ( 2 ; % 5 * " . 9 # ( ' % . ' ( 4 0 1 ; % . 0 1 % + ( 6 6 # $ # ( 0 $ % . , % / 4 6 , # 7 6 . 0 * % ' ( 4 0 1 . 5 ( 4 , $ % . ' * % %
& ' * $ * 0 , * 1 % . 6 ( 0 8 % : # , " % ' * + ( / / * 0 1 . , # ( 0 $ % 2 ( ' % 8 * ( / * , ' # + % 1 * $ # 8 0 ; % 1 * $ # 8 0 % $ & * * 1 ; % $ # 8 " , % 1 # $ , . 0 + * ; %
: # 1 , " % ( 2 % 6 . 0 * $ ; % $ # 8 0 . 8 * % . 0 1 % & . 9 * / * 0 , % / . ' > # 0 8 $ ; % . 0 1 % ( & * ' . , # ( 0 . 6 % ' * + ( / / * 0 1 . , # ( 0 $ < %
!
!
!
!
) * + , - % . # / ! ' ( 4 0 1 . 5 ( 4 , $ ; % / 4 6 , # 7 6 . 0 * ; % & * 1 * $ , ' # . 0 ; % 5 # + 3 + 6 * ; % 5 # + 3 + 6 # $ , ; % + ( 6 6 # $ # ( 0 $ ; % 1 * / . 0 1 ; %
5 * " . 9 # ( ' %
%
!
!
FINAL REPORT: TASK ORDER 6222
Executive Summary
Project Purpose
Caltrans and local agencies are installing roundabouts on roadways throughout the state of California.
Research indicates that while single- lane roundabouts may benefit bicyclists and pedestrians by slowing
traffic, multi- lane roundabouts may significantly increase safety problems for these users, especially those who
are disabled. This project examines the safety and demand issues for pedestrians and bicyclists at multi- lane
roundabouts through a literature review, case studies, in- field counts and surveys, focus groups, and video
analysis.
The key goals of the project are to:
1. Identify factors at multi- lane roundabouts that influence bicyclist and pedestrian- involved collision
rates;
2. Identify factors at multi- lane roundabouts that affect bicyclist and pedestrian demand; and
3. If effect found, recommend design treatments to mitigate these impacts on bicyclists and pedestrians.
Report Overview
This document presents research findings, synthesizes current information on best practices, and makes
recommendations to assist local agencies planning and designing safer multi- lane roundabouts. These
findings should help local agencies and Caltrans create roundabouts that better and more safely address the
needs of bicyclists and pedestrians. The current literature is referred to throughout the document to augment
the research team’s findings, especially for issues that were beyond the scope of this project.
Findings
Key analysis findings include:
Avoidance
! While 25% of bicyclists and 14% of pedestrians intercepted in the field stated that they would change
their route to avoid multi- lane roundabouts, in- field comparison counts did not show a significant
difference in pedestrian or bicyclist activity at roundabouts compared to traditional intersections
! Video analysis at three roundabouts showed an inverse relationship between motor vehicle volumes
and pedestrian volumes, as at most intersections.
! Self- reported comfort with multi- lane roundabouts differs by user mode. Bicyclists were more likely
than pedestrians to report feeling uncomfortable traveling through the roundabout, with 32 percent
of bicyclists feeling uncomfortable traveling through the roundabout, compared to 18 percent of
pedestrians. Most respondents felt comfortable traveling through the roundabout ( 60 percent of
bicyclists and 53 percent of pedestrians.)
! People’s comfort level at a multi- lane roundabout appears to be affected by the age of the
respondent, the motor vehicle, bicycle and pedestrian volumes at the roundabout, and also the
geometric configuration of the roundabout. Of the three roundabout locations surveyed,
5
FINAL REPORT: TASK ORDER 6222
respondents at the East Lansing roundabout reported being most comfortable walking and biking
through the roundabout ( 62 percent). These respondents were generally young ( 69 percent between
ages 18- 25) and the roundabout has a shared- use path around the perimeter. In Rehoboth Beach,
Delaware, 49% of people surveyed were comfortable using the roundabout. This roundabout has
significant bicycle and pedestrian activity ( 88 bicyclists and 89 pedestrians per hour during the
observation period).
! When given the choice of stop controlled, signalized and roundabout intersections, pedestrians
equally prefer signalized intersections and roundabouts, but bicyclists prefer signalized intersections
and not roundabouts. Neither bicyclists nor pedestrians prefer four- way stop- controlled
intersections. The preference for more typical intersection types is probably not related to familiarity
with these types and unfamiliarity with multi- lane roundabouts; all focus group participants were
familiar with multi- lane roundabouts, and in the case of Maryland, had to travel through one or more
multi- lane roundabouts to access the focus group location.
Collisions
! While data is limited, some studies suggest that multi- lane roundabouts have little effect on
pedestrian crash numbers— either positively or negatively, however pedestrian volume data is rarely
available to compare rates per pedestrian crossing.
! While there are no U. S. studies on the subject, non- U. S. studies have shown that circulating bicyclist-entering
vehicle collisions are the most common bicyclist collision type in multi- lane roundabouts.
! Bicyclist and pedestrian crash rates, measured by crashes per million bicyclists and pedestrians, vary
at different roundabouts in different locations ( e. g., 1.09 per million at East Lansing and 0.49 per
million at Santa Barbara.)
! European studies have shown that pedestrian and bicyclist crashes account for only 1 percent of the
total crashes at roundabouts. By contrast, bicyclist and pedestrian crashes in the case study
roundabouts accounted for a much larger percentage of total crashes ( 12 percent at Santa Barbara, 55
percent at East Lansing). This suggests that European roundabout design, bicycle and pedestrian
facility design, or driving, walking and biking behavior may have a role in reducing the number of
bike and pedestrian collisions.
! European studies have shown that the four factors with the strongest effect on total crash rates in
roundabouts are total traffic volume, proportion of vehicles entering from the minor road, operating
speed, and number of legs.
! Very few conflicts were observed in video analysis of case study multi- lane roundabouts.
Behavior
! Based on video observations at case study multi- lane roundabouts, pedestrians overwhelmingly chose
to cross at a crosswalk, between 41- 100% did not have to wait for a gap in traffic, depending on the
roundabout and leg.
! Between 33 and 100 percent of pedestrians observed in the video analysis had to wait to cross a
roundabout leg. The wait times averaged 3.6 seconds for crossing entering lanes and 5.6 seconds for
crossing exiting lanes.
6
FINAL REPORT: TASK ORDER 6222
! Multi- lane roundabouts with higher pedestrian volumes see less wait time for pedestrians and higher
yielding rates by motorists compared to multi- lane roundabouts with lower bicyclist and pedestrian
volumes.
! The majority of bicyclists observed riding in the circulating lane of a roundabout rode on the outside
edge of the lane, as opposed to the center of the lane, indicating discomfort, caution, or lack of
understanding of the appropriate way to ride through a multi- lane roundabout.
! When a shared- use path is provided around a roundabout, between 27 and 62% of bicyclists choose
to use the path, rather than travel through the roundabout on the roadway.
! Bicyclists would prefer multi- lane roundabouts with vehicle speeds that are close to average bicycling
speed ( 12 to 15 mph)
Recommendations
The comprehensive set of recommendations in this document is designed to aid engineers and planners in
determining where multi- lane roundabouts are appropriate, and how to best accommodate pedestrians and
bicyclists at these intersections through design features. The recommendations may also be used to evaluate
existing multi- lane roundabouts and determine if changes are necessary.
Recommendations are drawn from current FHWA, Caltrans and AASHTO guidance, emerging best practices
in the literature, and the results from our data collection and analysis. Chapter 5 contains specific
recommendations regarding the following design standards:
! Geometric Design
! Design Speed
! Sight Distance
! Width of Lanes
! Signage and Pavement Markings
! Operational Recommendations
! Example diagrams for addressing the needs of bicyclists and pedestrians at multi- lane roundabouts
We conclude with descriptions of innovative designs for roundabouts and recommendations for future
research.
7
FINAL REPORT: TASK ORDER 6222
TABLE OF CONTENTS
EXECUTIVE SUMMARY .............................................................................................................................. 5
Project Purpose........................................................................................................................ ............................... 5
Report Overview ............................................................................................................................... ..................... 5
Findings....................................................................................................................... ..................................................... 5
Recommendations ............................................................................................................................... ........................... 7
1. PROJECT BACKGROUND AND PURPOSE........................................................................................ 12
1.1. Introduction................................................................................................................... ................................ 12
1.2. Relevant Related Research....................................................................................................................... .... 12
1.2.1. Definition of a Roundabout .............................................................................................................................. 12
1.2.2. California Traffic Laws Regarding Roundabouts ........................................................................................... 13
1.2.3. Quality of Existing Vehicle, Pedestrian, and Bicyclist Statistics Regarding Roundabouts ........................ 14
1.2.4. Available Information on Existing Pedestrian and Bicyclist Behavior ........................................................ 14
1.2.5. Correlations between Pedestrian and Bicyclist Injuries, Fatalities, and Activity Levels and Multi- lane
Roundabouts ............................................................................................................................... .................................. 15
1.2.6. Accommodating Pedestrians and Bicyclists at Multi- lane Roundabouts..................................................... 15
2. METHODOLOGY.................................................................................................................... ............... 17
2.1. Selection of Case Study Roundabouts ........................................................................................................ 18
2.2. Bicycle and Pedestrian Counts..................................................................................................................... 18
2.2.1. Intersection Comparison Counts...................................................................................................................... 18
2.2.2. Corridor Count Comparisons.................................................................................................................... ....... 19
2.3. Collision Analysis ............................................................................................................................... ........... 21
2.4. Video Analysis ............................................................................................................................... ................ 21
2.5. Intercept Surveys........................................................................................................................ ................... 22
2.6. Focus Groups ............................................................................................................................... ................. 23
3. SUMMARY OF FINDINGS AND DISCUSSION..................................................................................... 25
3.1. Bicycle and Pedestrian Activity Levels at Multi- Lane Roundabouts ..................................................... 25
3.1.1. Milpas Corridor Analysis....................................................................................................................... ............ 27
3.1.2. Intercept Surveys........................................................................................................................ ........................ 27
3.1.3. Summary of Findings....................................................................................................................... .................. 28
3.2. Bicycle and Pedestrian Safety at Multi- Lane Roundabouts..................................................................... 29
3.2.1. Literature Findings....................................................................................................................... ...................... 29
3.2.2. Comparison of Collision Points between Signalized Intersections and Multi- Lane Roundabouts ......... 30
3.2.3. Crash Frequency and Severity ........................................................................................................................... 31
3.3. Bicyclist, Pedestrian and Driver Behavior.................................................................................................. 32
8
FINAL REPORT: TASK ORDER 6222
3.3.1. Video Analysis ............................................................................................................................... ..................... 32
3.3.2. Intercept Surveys........................................................................................................................ ........................ 33
3.3.3. Focus Group Analysis ............................................................................................................................... ........ 33
3.4. Summary of Findings ............................................................................................................................... .... 33
3.5. Limitations of Research ............................................................................................................................... 35
4. RECOMMENDATIONS................................................................................................................ .......... 37
4.1. Introduction................................................................................................................... ................................ 37
4.1.1. Documents Governing the Design of Multi- Lane Roundabouts in California .......................................... 37
4.2. Comparison of Bicyclist and Pedestrian Experience at Multi- Lane Roundabouts and Signalized
Intersections ............................................................................................................................... ........................... 38
4.3. Operational Issues for Pedestrians and Bicyclists at Multi- Lane Roundabouts .................................. 39
4.3.1. Issues for Pedestrians ............................................................................................................................... ......... 39
4.3.2. Issues for Bicyclists ............................................................................................................................... ............. 40
4.4. Considerations When Planning Multi- Lane Roundabouts...................................................................... 41
4.5. Designing Multi- Lane Roundabouts to Meet the Needs of Bicyclists and Pedestrians ..................... 42
4.5.1. General Design Goals ............................................................................................................................... ........ 42
4.5.2. Geometric Design of Multi- Lane Roundabouts ............................................................................................. 43
4.5.3. Design Speed ............................................................................................................................... ....................... 43
4.5.4. Sight Distance....................................................................................................................... .............................. 45
4.5.5. Width of Lanes ............................................................................................................................... .................... 51
4.5.6. Single- lane vs. Multi- lane........................................................................................................................... ........ 51
4.5.7. Signage and Pavement Markings for Directional Guidance.......................................................................... 52
4.5.8. Bicycle Facilities at Roundabouts.................................................................................................................... . 56
4.5.9. Pedestrian Facilities at Roundabouts................................................................................................................ 59
4.5.10. Splitter Island/ Pedestrian Refuge.................................................................................................................. 61
4.5.11. Grade- Separated Crossings...................................................................................................................... ....... 62
4.5.12. Operational Recommendations................................................................................................................ ...... 63
4.5.13. Accommodating Visually Impaired Pedestrians ........................................................................................... 63
4.5.14. Innovative Designs ............................................................................................................................... ........... 64
4.6. Illustrations of Recommended Design Treatments.................................................................................. 69
5. CONCLUSIONS.................................................................................................................... ................. 71
5.1. Summary of Findings ............................................................................................................................... .... 71
5.2. Additional Avenues of Research ................................................................................................................. 71
6. REFERENCES..................................................................................................................... .................. 73
9
FINAL REPORT: TASK ORDER 6222
LIST OF TABLES
Table 1: Data Collection by Roundabout Location .................................................................................................... 18
Table 2 Case Study Roundabouts Selected for Intersection Comparison Counts ................................................. 19
Table 3: Video Camera Locations and Recording Times........................................................................................... 21
Table 4: Summary of Roundabout Features, Conditions, and Operations ............................................................. 24
Table 5: Average Hourly Weekday Counts................................................................................................................... 25
Table 6: Average Hourly Weekend Counts .................................................................................................................. 25
Table 7: Weekday Ratios of Bicyclist and Pedestrian Counts at Roundabouts vs. Comparison Sites ................ 26
Table 8: Weekend Ratios of Bicyclist and Pedestrian Counts at Roundabouts vs. Comparison Sites................ 26
Table 9: Overall Ratios of Bicyclist and Pedestrian Counts at Roundabouts vs. Comparison Sites ................... 26
Table 10. Roundabout Avoidance...................................................................................................................... ........... 28
Table 11: Summary Collision Data 2002 through 2008.............................................................................................. 31
Table 12: Matrix Comparing Multi- Lane Roundabouts to Traditional Intersections............................................ 39
Table 13: Recommended Design Goals for Accommodating Bicyclists and Pedestrians at Multi- Lane
Roundabouts in Urban, Rural and Freeway Interchange Environments ................................................................ 70
LIST OF FIGURES
Figure 1: Illustration of How Research Answers Three Key Questions ................................................................. 18
Figure 2: Map of Count Locations in Santa Barbara................................................................................................... 20
Figure 3: Schematic of How Data was Analyzed......................................................................................................... 20
Figure 4: Bicyclist and Pedestrian Conflict Points at Signalized Intersection ......................................................... 31
Figure 5: Bicyclist and Pedestrian Conflict Points at Multi- Lane Roundabout ...................................................... 31
Figure 6: Five Key Radii that Affect Design Speed..................................................................................................... 44
Figure 7: Roundabout with unlimited sight distance................................................................................................... 46
Figure 8: Landscaped Splitter Island on Roundabout Approach Henrico County, VA ....................................... 47
Figure 9: Circulating Stream ( left) and Entering Stream ( right) from DIB 80- 01.................................................. 48
Figure 10: Approach Sight Distance ( Source: FHWA Roundabouts: An Informational Guide)....................... 49
Figure 11: Circulating Roadway Sight Distance ( Source: FHWA Roundabouts: An Informational Guide).. 49
Figure 12: Sight Distance to Crosswalk on Immediate Downstream Exit ( Source: FHWA Roundabouts: An
Informational Guide) ............................................................................................................................... ....................... 50
Figure 13: Bicyclist Controlling the Lane in Rehoboth Beach, DE.......................................................................... 51
Figure 14: Aerial View of Rehoboth Beach Roundabout........................................................................................... 52
Figure 15: Driver, Pedestrian, and Bicyclist Scanning on Approach to Roundabout............................................ 53
Figure 16: MUTCD Recommended Guide Signs for Roundabouts ( FHWA, 2008) ............................................ 54
Figure 17: MUTCD Recommended Exit Sign for Roundabouts ( FHWA, 2008).................................................. 54
10
FINAL REPORT: TASK ORDER 6222
11
Figure 18: Recommended Optional Fish- Hook Arrows on Lane Use Control Signs at Multi- Lane
Roundabouts ( FHWA, 2008).......................................................................................................................... ............... 54
Figure 19: Optional Fish- Hook Lane- Use Arrows ( FHWA, 2008).......................................................................... 55
Figure 20: Bike Facility Design from Caltrans DIB 80- 01......................................................................................... 57
Figure 21: Example of a Right- Turn Bypass Lane with a Yield at Exit Leg ( FHWA Guide) .............................. 59
Figure 22: Bicycle and Pedestrian Roundabout Undercrossing at Western Michigan University ....................... 62
Figure 23: Undercrossing Detail ( Western Michigan University) ............................................................................. 63
Figure 24: Netherlands Roundabout ............................................................................................................................. 64
Figure 25: Tractor trailer circulating through C- Roundabout ................................................................................... 65
Figure 26: Typical Dutch Design for a Turbo Roundabout ...................................................................................... 66
Figure 27: Jog in bicycle path on splitter island ........................................................................................................... 67
Figure 28: Toucan Crossing at Roundabout in England ............................................................................................ 68
Figure 29: Signalized Roundabout York, England ...................................................................................................... 68
FINAL REPORT: TASK ORDER 6222
1. Project Background and Purpose
1.1. Introduction
Caltrans and local agencies are considering installing roundabouts on roadways throughout the state of
California. While appropriately- designed roundabouts can reduce traffic collisions and therefore increase
safety effects, it is unclear whether bicyclists and pedestrians garner the same level of benefit as vehicle
occupants. Research indicates that single- lane roundabouts may benefit bicyclists and pedestrians by slowing
traffic. In general, multi- lane roundabouts reduce the number of pedestrian- and bicyclist- involved collisions
in before- and- after studies, but most analyses do not account for changes in pedestrian and bicyclist volume
and do not use rates. Globally, there are few studies of bicyclist and pedestrian behavior at roundabouts, and
even fewer studies conducted in the U. S. In response to current bicyclist and pedestrian data needs, this
research initiative sought to complete the following key tasks:
1. Identify factors at multi- lane roundabouts that influence bicyclist and pedestrian- involved collision
rates.
2. Identify factors at multi- lane roundabouts that affect bicyclist and pedestrian travel demand.
3. If effect found, recommend design treatments for multi- lane roundabouts to mitigate impacts on
bicyclists and pedestrians.
This project examines the safety and demand issues for pedestrians and bicyclists at multi- lane roundabouts
through a literature review, case studies, in- field counts and surveys, focus groups, and video analysis. The
report concludes with recommendations for design guidelines for multi- lane roundabouts as well as for
circumstances under which multi- lane roundabouts should or should not be installed.
1.2. Relevant Related Research
Roundabouts have long been used in many parts of the world and continue to gain popularity in places with
little previous experience with roundabouts. Though roundabouts have generally been proven to decrease
the number and severity of automobile collisions, there is little, if any, consistent data on the safety of non-motorized
users in roundabouts. A literature review was conducted to consolidate research on modern
roundabouts, pedestrian and bicyclist behavior, and the interaction that occurs when these users attempt to
navigate multi- lane roundabouts. Some conclusions can be drawn about the safety impacts on pedestrians
and bicyclists at multi- lane roundabouts, mostly from European and Australian experience. Additionally,
there are a number of innovative treatments and recommendations aimed at making multi- lane roundabouts
safer for more vulnerable users. A full annotated bibliography is included as an Appendix.
1.2.1. Definition of a Roundabout
Though modern roundabouts are circular intersections, they are different from traditional traffic circles. The
modern roundabout has several unique characteristics, the most prominent being the rule that drivers ( both
motorists and bicyclists) yield on entry. Other key characteristics include a central island with deflection,
which forces motorists and bicyclists to slow down, and a splitter island which separates traffic on the entry
and exit legs. The articles reviewed in this section not only define these characteristics, but also provide
design guidelines that can be followed when designing a modern roundabout.
12
FINAL REPORT: TASK ORDER 6222
1.2.2. California Traffic Laws Regarding Roundabouts
California Vehicle Code ( CVC) does not have specific provisions that govern how motorists ( referred to as
" drivers of vehicles" in the CVC), pedestrians and bicyclists must use a roundabout. However, it does have
general provisions governing all users that are applicable to roundabouts.
Generally, motorists and bicyclists must yield to pedestrians within crosswalks, but pedestrians are required to
exercise due caution when crossing a roadway at a marked or unmarked crosswalk. ( CVC Section 21950)
When bicyclists walk their bikes across a crosswalk, they become a pedestrian with the same rights and
responsibilities as other pedestrians. Bicyclists riding in crosswalks are not discussed in the CVC, and it is
therefore unclear whether a motorist's duty to yield to pedestrians in a crosswalk also applies to bicyclists
riding in a crosswalk.
CVC Section 21950.
Right- of- Way at Crosswalks
21950. ( a) The driver of a vehicle shall yield the right- of- way to a pedestrian crossing the roadway within
any marked crosswalk or within any unmarked crosswalk at an intersection, except as otherwise provided
in this chapter.
( b) This section does not relieve a pedestrian from the duty of using due care for his or her safety. No
pedestrian may suddenly leave a curb or other place of safety and walk or run into the path of a vehicle
that is so close as to constitute an immediate hazard. No pedestrian may unnecessarily stop or delay traffic
while in a marked or unmarked crosswalk.
( c) The driver of a vehicle approaching a pedestrian within any marked or unmarked crosswalk shall
exercise all due care and shall reduce the speed of the vehicle or take any other action relating to the
operation of the vehicle as necessary to safeguard the safety of the pedestrian.
( d) Subdivision ( b) does not relieve a driver of a vehicle from the duty of exercising due care for the safety
of any pedestrian within any marked crosswalk or within any unmarked crosswalk at an intersection.
When riding on the road “ at a speed less than the normal speed of traffic moving in the same direction at that
time,” bicyclists have the same rights and responsibilities as drivers, with the exception that they must " ride as
close as practicable to the right- hand curb or edge of the roadway," except under circumstances provided in
CVC Section 21950.. ( CVC Section 21202) These circumstances, which may apply to bicyclists traveling
through roundabouts at some point, are:
1. When overtaking and passing another bicycle or vehicle proceeding in the same direction.
2. When preparing for a left turn at an intersection or into a private road or driveway.
3. When reasonably necessary to avoid conditions ( including, but not limited to, fixed or moving
objects, vehicles, bicycles, pedestrians, animals, surface hazards, or substandard width lanes) that
make it less safe to continue along the right- hand curb or edge, subject to the provisions of Section
13
FINAL REPORT: TASK ORDER 6222
21656. For purposes of this section, a " substandard width lane" is a lane that is too narrow for a
bicycle and a vehicle to travel safely side by side within the lane.
4. When approaching a place where a right turn is authorized.
These situations may apply at some point to bicyclists riding through roundabouts. The CVC requires drivers
approaching a yield- controlled intersection to yield the right of way to any motor vehicles that are in the
intersection or close enough to create a hazard. ( CVC Section 21803) As stated in CVC Section 21200, the
same provisions for drivers of vehicles apply to bicyclists, except where they are specifically not applicable.
1.2.3. Quality of Existing Vehicle, Pedestrian, and Bicyclist Statistics
Regarding Roundabouts
A group of articles addresses the availability of data used to conduct statistical analyses of collisions at
roundabouts, and identifies gaps in current practices. Auto collision data is generally available, but not always
for specific roundabouts. The available information regarding pedestrian and bicyclist collisions is often
ambiguous or incomplete, and pedestrian and bicyclist- involved collisions, in general, are not always reported.
In the United States, lack of data may be mainly due to the fact that there are few roundabouts for
observation, particularly multi- lane roundabouts, and to the limited number of pedestrians and bicyclists
traveling through these roundabouts. In addition, the distinction between single- lane and multi- lane types is
usually not made in existing data collected on roundabouts. While pedestrian and bicyclist volumes and
collisions can be modeled and simulated, there is a lack of substantiated data on these specific users within
roundabouts.
1.2.4. Available Information on Existing Pedestrian and Bicyclist
Behavior
The literature reviewed concerning existing walking and bicycling behavior focuses mainly on two topics:
route choice and perceived risk. The route choice articles attempt to define variables that influence route
selection for pedestrians or bicyclists when these users are faced with alternative routes, although the articles
do not specifically discuss roundabouts. There is a general consensus that route directness is a primary
consideration in determining route choice. However, there is evidence that non- motorized travelers, and
bicyclists in particular, are willing to travel additional distance in exchange for other benefits that they find
significant ( Harvey et al. 2008, Howard & Burns 2001, Aultman- Hall et al. 1997). Some significant variables
include travel time, safety, and pleasantness. Westerdijk et al. ( 1990) used a multiattribute utility model to
quantify these tradeoffs and found, for example, that bicyclists were willing to travel an additional 250 meters
( 820 feet) to gain one extra point for traffic safety on a theoretical 7- point scale. Although these studies do
not explicitly address roundabouts, they still provide insight into how far out of their way pedestrians and
bicyclists will go in order to travel a more comfortable route.
This issue of self- reported comfort is also addressed in the articles that discuss cyclists’ perceptions of risk.
Parkin et al. ( 2006) studied perceived risk over an entire bike journey and concluded that roundabouts, two-way
auto traffic, and the number of parked vehicles on the street are all factors that increase perceived risk for
cyclists. Moller and Hels ( 2008) developed a model for variation in cyclists’ perceived risk, specifically at
roundabouts. They found that the most significant variables were gender, having experienced a near- collision
in the past year, auto volume through the roundabout, and the existence of a cycle facility. While the
existence of a cycle facility in a roundabout decreased perceived risk, this study did not attempt to compare
perceived risk with measures of actual risk or investigate how this perception affects behavior. However, the
route choice studies imply that these perceptions of risk might be a significant variable affecting route choice.
Two reports specifically address how non- motorized users react to roundabouts. One study conducted in
New Zealand included a comprehensive survey of bicyclists regarding their attitudes toward multi- lane
14
FINAL REPORT: TASK ORDER 6222
roundabouts ( Campbell et al., 2006). The authors found that while 85% of the survey respondents identified
themselves as experienced cyclists, 93% felt that multi- lane roundabouts were a hazard and a deterrent to bike
riding. Over 60% said they were willing to make some attempt to avoid multi- lane roundabouts. Novice
cyclists demonstrated even higher levels of aversion to multi- lane roundabouts.
1.2.5. Correlations between Pedestrian and Bicyclist Injuries,
Fatalities, and Activity Levels and Multi- lane Roundabouts
Much of the existing literature on roundabout safety supports the well- documented conclusion that
roundabouts have the potential to increase both motor vehicle capacity and motor vehicle safety. But much
less attention has been paid to the impact of multi- lane roundabouts on pedestrians and bicyclists. Most
studies, especially in the U. S., have found that there is too little data to conduct meaningful analyses of bike
and pedestrian collisions in roundabouts. However, there have been some studies on this relationship outside
the U. S.
Generally, the effect of multi- lane roundabouts on pedestrian safety perceptions and behavior is unclear,
although some studies have found that roundabouts result in no significant change in levels of pedestrian
safety. Observational studies have found that pedestrians are more likely to hesitate at multi- lane
roundabouts than at other types of intersections ( Harkey & Carter, 2006) and that visually impaired
pedestrians experience longer waiting times and more risky crossings at multi- lane roundabouts ( Ashmead et
al., 2005).
Roundabout design is a critical factor in safety for all users; if the roundabouts are designed to allow for
speeds in excess of 25 mph, more collisions can occur. In addition, multi- lane roundabouts have more cyclist
collisions when compared to comparable single- lane roundabouts, as a result of a greater difference in speeds
between modes ( Furtado, 2004). Several studies ( including Furtado, Brüde & Larsson ( 2000), Harkey &
Carter ( 2006), Shen ( 2000), and USDOT FHWA ( 2000)) have found that multi- lane roundabouts are
perceived as more dangerous, and often result in more collisions for all users when compared to single- lane
roundabouts. Exposure rates were not available for these studies. This leads to a conclusion that multi- lane
roundabouts can significantly increase bicyclist safety risk. Chapter 5 of the U. S. DOT FHWA publication,
“ Roundabouts: An Information Guide,” ( 2000) states that adding an additional lane to a one- lane roundabout
is likely to increase overall injury crashes by 25%. Brude and Larsson ( 2000) found that in Sweden, bicycle
collisions were six times more frequent on multi- lane roundabouts compared to single- lane roundabouts.
Daniels et al. have conducted some comprehensive studies of the effects of roundabouts on bicyclists in
Flanders- Belgium. In their 2008 study, they investigated whether a safety effect could be quantified and if
this effect was influenced by particular characteristics of the roundabout location. They include both single-and
multi- lane roundabouts in their sample. Their study of before- and- after roundabout installation reveals
that roundabouts increased injury collisions involving bicyclists by 27% and severe injury collisions by up to
46%. Roundabouts constructed inside built- up areas had a negative effect on bike safety, as did roundabouts
that replaced previously signalized intersections. Daniels, et al. followed up with a 2009 study to determine if
bicycle facilities within roundabouts have any effect on bicyclist safety. They arrived at the unexpected
conclusion that roundabouts with cycle lanes increased bicycle injury collisions significantly ( as opposed to
roundabouts with separate cycle paths, grade separated paths, or no bicycle facility) and suggested that a clear
distinction should be made between roundabouts with cycle lanes and those with other types of facilities.
1.2.6. Accommodating Pedestrians and Bicyclists at Multi- lane
Roundabouts
Several articles were reviewed that specifically address accommodations for pedestrians and bicyclists within
roundabouts. Some of the literature consists of general design manuals, which note that special
considerations need to be made for non- motorized users when designing any roundabout. Several of these
15
FINAL REPORT: TASK ORDER 6222
discuss the relatively new practice of roundabout signalization for pedestrian access. Inman and Davis ( 2007)
discovered that roundabout signalization has mainly focused on improving traffic operations and not
necessarily on benefits to other users. However, it has been shown that certain pedestrians, particularly
pedestrians with disabilities, may require special treatments to safely and efficiently travel through
roundabouts. Rouphail et al. ( 2005) and Schroeder et al. ( 2008) both used simulation models to study the
effects of pedestrian signalization treatments on roundabout operations. Schroeder, et al. found that delays
for all users could be mitigated using a two- stage pedestrian signal or a HAWK ( High- Intensity Activated
Crosswalk) system ( more recently renamed a Pedestrian Hybrid Beacon in the 2009 MUTCD), both of which
minimize the red time for auto traffic. Rouphail et al. suggest that a mid- block crossing downstream of
exiting traffic minimizes the possibility of disruptive queues forming; but the tradeoffs between traffic
operations and increased pedestrian travel distance have not been examined. There is also still some dispute
regarding when and where pedestrian signalization should be implemented. The U. S. Access Board has
proposed guidelines, referred to as “ PROWAG,” that would require pedestrian- actuated signals at all multi-lane
roundabout crossings. ( Access Board, 2005) However, there are some who believe that further research
must be conducted before any generalized guidelines should be enforced ( Baranowski 2005). Additional
research is underway as part of NCHRP 3- 78A, “ Crossing Solutions at Roundabouts and Channelized Turn
Lanes for Pedestrians with Vision Disabilities” and as part of a project of the National Institutes of Health,
National Eye Institute, “ Blind Pedestrians’ Access to Complex Intersections,” which may provide additional
information to engineers and designers of roundabouts.
Three studies propose new ideas for road treatments and roundabout design intended to benefit pedestrians
and bicyclists. The first study evaluates a new road treatment to audibly alert visually- impaired pedestrians to
the presence of yielding vehicles at multi- lane roundabouts ( Inman et al., 2005). Two other papers present
new versions of a roundabout design that are more amenable to non- motorized users. Campbell, et al. ( 2006)
introduce the concept of a cyclist- roundabout ( or C- roundabout), which has very specific geometric
guidelines intended to reduce the 85th percentile auto circulating speed to 30 kilometers/ hour ( 19
miles/ hour), which reduces the differential between cars and the typical bicyclist to 10 kilometers per hour.
The most distinctive feature of the multi- lane C- roundabout is the narrow entry lanes, which encourages
operators of heavy vehicles to travel in a single file. Campbell, et al. also suggest using economical vertical
deflection devices ( such as speed humps) at entry legs, but concede that these may be opposed by drivers of
buses, emergency vehicles, and other heavy vehicles. Another novel multi- lane roundabout design, the turbo-roundabout,
was described by Fortuijn in 2003. The turbo- roundabout design prohibits lane changing among
the circulatory traffic and has been implemented in the Netherlands. This design benefits non- motorized
users by lowering the circulatory speed and reducing the number of potential conflict points.
Much more research is needed on the best way to accommodate pedestrians and bicyclists at roundabouts.
But some valuable work has been done to show that a combination of innovative solutions and efficient use
of current treatments may ultimately benefit all users.
16
FINAL REPORT: TASK ORDER 6222
2. Methodology
Between January 2007 and February 2009, the researchers collected both qualitative and quantitative data
regarding pedestrian and bicyclist usage of multi- lane roundabouts. Methods used to collect data were:
! Bicyclist and pedestrian counts at four case study multi- lane roundabouts and at nearby comparison
sites.
! Comparison of bicyclist and pedestrian volumes along three corridors, one of which contained a
multi- lane roundabout.
! Summary of police- reported collision data at two case study multi- lane roundabouts.
! In- field intercept surveys of bicyclists and pedestrians at three case study multi- lane roundabouts.
! Video documentation of bicyclists and pedestrians at three case study multi- lane roundabouts.
! Four focus groups held in two communities with multi- lane roundabouts.
Figure 1 illustrates how our different data collection efforts work together to answer the three main questions
in the report:
1. Do pedestrians and bicyclists avoid multi- lane roundabouts? If so, why?
2. Are pedestrians and bicyclists more likely to be involved in crashes or more severe crashes at multi-lane
roundabouts than at other types of intersections? Why?
3. Given the answers to the above questions, where should multi- lane roundabouts be installed? Where
should they not be installed? And what type of specific design treatments are recommended for
improving safety, mobility and comfort of bicyclists and pedestrians at multi- lane roundabouts?
17
FINAL REPORT: TASK ORDER 6222
Figure 1: Illustration of How Research Answers Three Key Questions
2.1. Selection of Case Study Roundabouts
Data was collected at five case study multi- lane roundabouts. Bicyclist and pedestrian counts were conducted
at all five locations, while more in- depth analysis ( surveys, video documentation, focus groups) were
conducted at a subset of locations.
Table 1: Data Collection by Roundabout Location
Location
Bike & Ped
Counts and
Comparison
to Nearby
Intersections
Bike & Ped
Counts and
Comparison
to Nearby
Corridors
Bike & Ped
Collision
Summary
Intercept
Surveys
Video
Documentation
Focus
Groups
Santa Barbara, CA
Milpas Road & Hwy 101 X X X
not
conducted not conducted X
Annapolis, MD
Spa Road, Taylor
Avenue & MD 450
X not
conducted not available X X not
conducted
Kentlands
( Gaithersburg), MD
Kentlands Boulevard &
Market Street
X not
conducted not available not
conducted not conducted X
Rehoboth Beach, DE
Rehoboth Avenue &
Grove Street
X not
conducted not available X X not
conducted
East Lansing, MI
Shaw Lane & Bogue
Street
X not
conducted
X
X X not
conducted
2.2. Bicycle and Pedestrian Counts
2.2.1. Intersection Comparison Counts
Researchers conducted bicyclist and pedestrian counts at four case study multi- lane roundabouts and at
nearby comparison sites to determine if bicyclists and/ or pedestrians were avoiding roundabouts. Corridor
counts were used at the Santa Barbara, CA site because of roadway configuration and presence of the U. S.
Route 101. The researchers hypothesized that if bicyclist and/ or pedestrian counts were higher at
comparison sites than at nearby roundabouts, that may indicate that bicyclists and/ or pedestrians were
avoiding the multi- lane roundabouts.
Methodology
The count methodology consists of comparing bicycle and pedestrian volumes at a roundabout to
comparable signalized or stop- controlled sites within a half a mile. The ideal case study site had standard
multi- lane geometry, was surrounded by a grid network of streets, and had similar land uses at the
roundabout and at potential comparison sites within half a mile of the roundabout. Very few multi- lane
roundabouts met all of these conditions.
18
Table 2 lists the final case study sites that were selected. More information on the selection and
characteristics of the comparison sites is available in Identifying Factors that Determine Bicyclist and Pedestrian-
Involved Collision Rates and Bicyclist and Pedestrian Demand at Multi- Lane Roundabouts: Year One Report.
FINAL REPORT: TASK ORDER 6222
Table 2 Case Study Roundabouts Selected for Intersection Comparison Counts
Location Roundabout
Intersection
Standard Multi- Lane
Geometry
Surrounded by
Grid Network
Similar Land Uses at
Roundabout and
Comparison sites
Annapolis,
Maryland
Spa Road/ Taylor
Avenue/ MD 450 yes yes yes
Kentlands
Development,
Gaithersburg,
Maryland
Kentlands Boulevard
and Market Street yes yes yes
Rehoboth Beach,
Delaware
Rehoboth Avenue and
Grove Street
No – only major east-west
movement has
multiple lanes
no
no comparison sites
selected, due to lack of
grid network
East Lansing,
Michigan
Shaw Lane and
Bogue Street
No – right slip turn is
provided yes yes
Two- hour bicyclist and pedestrian counts were conducted during peak weekday and weekend periods at each
of the case study locations. At each case study location, counts were conducted at a roundabout location and
at one or two comparison sites. Field observations included identification of users by mode, general age, and
other attributes, and quality and extent of the existing bicycle and pedestrian facilities ( such as quality of
sidewalks, bicycle facilities, driveways, etc.) within three blocks or 1,000 feet of the roundabout, or within
proximity of the actual count locations.
2.2.2. Corridor Count Comparisons
In addition to the counts conducted at the four east coast locations, the project team conducted counts along
parallel corridors in Santa Barbara, California. The Santa Barbara counts were conducted to compare bicyclist
and pedestrian counts along three parallel corridors, one of which contains a multi- lane roundabout.
Methodology
To determine whether bicyclists and pedestrians avoid multi- lane roundabouts, researchers conducted counts
along three parallel corridors in Santa Barbara. The three corridors represent three different types of
interfaces with Highway 101: an undercrossing, signal- controlled on- off ramps, and a multi- lane roundabout.
Counts were conducted north of 101, at the interface of 101, and south of 101 during weekday and weekend
peak hours. Count data was collected at a total of nine locations. Figure 2 shows the corridors and count
locations that were sampled.
19
FINAL REPORT: TASK ORDER 6222
1. Milpas/ Yanonali
2. Milpas Roundabout
8. State/ 101
7. State/ E. Cota
6. Garden/ Cabrillo
5. Garden/ 101
3. Milpas/ Cabrillo
4. Garden/ E. Cota
9. State / Cabrillo
Figure 2: Map of Count Locations in Santa Barbara
These corridors were selected with the assumption that a large volume of pedestrian and bicycle movement
occurs between the commercial and residential areas northwest of the highway and the beach. By examining
bicyclist and pedestrian flows from north to south and south to north along the three corridors, we hoped to
establish whether interfaces with Highway 101 affect bicyclist and pedestrian movements.
At all count locations except the Milpas roundabout, trained counters recorded the number of bicyclists and
pedestrians that passed a screenline. At the Milpas roundabout, counters recorded turning movements.
Figure 3: Schematic of How Data was Analyzed
To analyze the counts, we counted the number of bicyclists and pedestrians that traveled toward the interface
of Highway 101 and compared it to the total number of bicyclists and pedestrians that were counted at that
20
FINAL REPORT: TASK ORDER 6222
interface. Using Figure 3 as an illustration, we compared the sum of Count A and Count B to Count C. If
Count C was higher than the sum of Count A and Count B, then it was assumed that bicyclists and
pedestrians were being funneled toward the interface to cross Highway 101. If Count C was lower than the
sum of Count A and Count B then it is assumed that bicyclists and pedestrians had a destination off the
corridor or were avoiding the interface with Highway 101.
2.3. Collision Analysis
Researchers collected historical collision data for two of the five case study roundabouts. This collision data
was compared with estimated levels of bicycle and pedestrian activity levels to calculate a collision rate.
2.4. Video Analysis
The video documentation methodology consisted of video- recording pedestrian, bicyclist, and vehicles at the
roundabouts for a period of time and reviewing the videos to look for particular behaviors. Videos were
recorded at the following three sites:
Maryland— Annapolis, Spa Road/ Taylor Avenue/ MD 450 Roundabout
Delaware— Rehoboth Beach, Rehoboth Avenue and Grove Street Roundabout
Michigan— East Lansing, Shaw Lane and Bogue Street Roundabout
At each roundabout, a video camera was set up at the center median of the roundabout, facing out, at each of
two or three approaches to each roundabout, and allowed to record for the periods of time indicated in
Table 3. The locations of the cameras are summarized in
Table 3 below:
Table 3: Video Camera Locations and Recording Times
Camera Location Number Approach Total Recorded Time ( min)
Rehoboth Beach, DE 1 Southeast 186.15
2 Southwest 209.77
3 Northwest 125.13
East Lansing, MI 1 South 156.75
2 North 170.60
21
FINAL REPORT: TASK ORDER 6222
22
Annapolis, MD8 1 East 30.56
2 North 11.87
3 Northwest 147.69
Total recording time for all locations, minutes: 1,038.52
Total recording time for all locations, hours: 17.31
The video review was conducted in two phases using a video playback tool developed by the California
Partners for Advanced Transit and Highways ( PATH). In the first phase, analysts played the video and
recorded the times of each type of event using numbers ( pedestrian = 1, bicycle = 2). The recorded times
and events were then exported to an Excel spreadsheet. In the second phase, analysts re- played the video and
recorded behaviors associated with each event in the spreadsheet, based on a pre- established protocol. The
behaviors reviewed included:
Pedestrians: whether pedestrians were in a group or not, crosswalk crossing times and directions, delay at
crosswalk, position with respect to crosswalk, and pace ( normal or running).
Bicycles: whether bicyclists were in a group ( 2 or more bicyclists) or not, riding or walking the bicycle,
location within the roundabout ( on the road, on the crosswalk, on the sidewalk or on the multi- use path),
position in the lane ( center or edge), direction ( with traffic, against traffic), and whether the bicyclist changed
behavior ( e. g. from walking the bicycle on the sidewalk to riding the bicycle on the road).
Vehicles ( with respect to pedestrians): whether motorists yielded to pedestrians, or forced pedestrians to wait
for a gap in traffic in order to cross.
The data was then analyzed to look for patterns in behavior. The video recordings were also used to estimate
vehicular volumes based on 15- minute counts, and vehicles entering, exiting, or circulating in the roundabout
were counted separately.
2.5. Intercept Surveys
During the summer of 2008, bicyclists and pedestrians were surveyed at three of the case study multi- lane
roundabouts. The purpose of the surveys was to find out which characteristics attract or deter bicyclists and
pedestrians from multi- lane roundabouts, and to provide guidance for the placement and design of multi- lane
roundabouts to accommodate all transportation modes.
Surveys were conducted on Friday, August 22, 2008 in the early afternoon. At each location, bicyclists and
pedestrians were asked to participate in a ten- question survey. The survey asked questions related to the
following topics:
! Method of traveling through the roundabout
! Comfort of traveling through the roundabout9
8 Please note, the camera batteries failed on two of the approaches in Annapolis which resulted in low
recording time for the East and North approaches.
9 Survey question: “ What was your comfort level as you traveled through the roundabout? A. Comfortable B.
Neutral ( neither comfortable nor uncomfortable) C. Uncomfortable”
FINAL REPORT: TASK ORDER 6222
23
! Preference for different types of intersections, and reason for preference
! Demographic information
2.6. Focus Groups
The primary purpose of the focus groups was to solicit information from a variety of roadway users regarding
bicycling and walking through multi- lane roundabouts. It is important to note that all other data collection
efforts for this project sample only people who were already using roundabouts. The focus groups were
intended to sample all types of roadway users and not only those who were already using roundabouts.
Methodology
A total of four focus groups were held: two in Kentlands, Maryland, and two in Santa Barbara, California.
Focus groups were held in January 2009 during the weekday morning and weekday evening, each consisting
of nine participants. Participants ranged in age from 18 to over 65. Slightly more women than men
participated in the groups ( 20 women versus 16 men). Participants were recruited by posting flyers at local
establishments in each community, as well as through ads posted on craigslist. org.
The focus groups were conducted in five sections, each concentrating on a particular topic area:
! Section 1 – Pedestrian and Bicycle Behavior
! Section 2 – Understanding of the Operations and Self- reported Comfort with Roundabouts
! Section 3 – Bicycle Design Options
! Section 4 – Pedestrian Design Options
! Section 5 – Final Comments/ Suggestions
The groups were conducted by a lead facilitator with the help of an assistant. Participants were paid a $ 50
cash honorarium for their participation at the end of the focus group. Comments were recorded using a
digital audio recorder, and the assistant also took notes. Comments were later transcribed from the digital
recordings. In most cases, comments were transcribed word for word as provided by the participants. But in
some cases, comments were paraphrased to capture their essence. As with most focus groups, short
discussions between participants often arose when a particular topic was addressed, and in these cases those
comments were recorded which reflected the nature of the discussion related to the designated topic.
FINAL REPORT: TASK ORDER 6222
Table 4: Summary of Roundabout Features, Conditions, and Operations
Location Lane Configuration Bicycle
Facilities Pedestrian Facilities
Speeds
( Approach
Circulating
Exit)
Land Uses
Notes
Santa Barbara, CA
Milpas Road & Hwy 101
2 circulating, exiting
and some circulating
lanes are delineated
5 legs
Bike lanes
on Milpas.
Sidewalks, high visibility
ladder crosswalks.
Pedestrians prohibited
from crossing southeast
and northwest legs at
roundabout.
25- 30 mph
n/ a
25- 30 mph
( posted)
Single family,
industrial,
grocery store
Non- standard design.
Circulating roadway is oval,
lanes are delineated, and
deflection at southeast and
northwest legs is smaller
than recommended.
Intersection includes on and
off- ramps for Highway 101.
Annapolis, MD
Spa Road, Taylor Avenue
& MD 450
2 circulating, 4 legs none
Sidewalks, high visibility
ladder crosswalks, brick
pavers, pedestrian
warning signage
22 mph
17 mph
17 mph
( radar)
National
cemetery, single-family
and
duplex, some
retail and office
Drivers entering from Taylor
Ave outside lane are not
permitted to circulate the
roundabout.
Kentlands ( Gaithersburg),
MD
Kentlands Boulevard &
Market Street
2 circulating, 4 legs none
Sidewalks, colored
pavers, sidewalks on
Kentlands are 2 car-lengths
back
n/ a
Large retail
stores, surface
parking lots.
Kentlands Blvd narrows
before it enters roundabout.
Market Street serves as
driveway to adjacent retail
stores.
Rehoboth Beach, DE
Rehoboth Avenue &
Grove Street
1.5 circulating ( only
inside lane travels
around entire
roundabout), 4 legs ( 2
multi- lane, 2 single-lane),
exiting and
circulating lanes are
delineated
none
Sidewalks, brick pavers,
pedestrian warning
signage
25 mph
18 mph
14 mph
( radar)
Retail,
residential
Rehoboth Avenue is a
heavily traveled
thoroughfare with access to
the beach from state
highway Route 1A to the
west. Bicycle, pedestrians,
fixed route transit, and auto
traffic is high and as such
drivers are more aware of
bicycle and pedestrian
presence.
East Lansing, MI
Shaw Lane & Bogue
Street
1.5 circulating ( only
inside lane travels
around entire
roundabout), 4 legs
Side path
Sidewalks, high- visibility
ladder crosswalks, “ Yield
to Peds in X Walk” sign
29 mph
19 mph
19 mph
( radar)
University
24
FINAL REPORT: TASK ORDER 6222
3. Summary of Findings and Discussion
3.1. Bicycle and Pedestrian Activity Levels at Multi- Lane Roundabouts
Do multi- lane roundabouts deter bicyclists and pedestrians?
The collected data from roundabouts did not show a significant difference in pedestrian or bicyclist activity at
roundabouts compared to surrounding traditional intersections. As can be seen in Table 5 and Table 6, even
when considering motor vehicle volumes, average hourly bicyclist and pedestrian counts at roundabouts were
not universally higher or lower than at comparison sites.
Table 5: Average Hourly Weekday Counts
Bicycle Volumes Pedestrian Volumes Motor Vehicle Volumes
Roundabout
Comparison 1
Comparison 2
Roundabout
Comparison 1
Comparison 2
Roundabout
Comparison 1
Comparison 2
East Lansing 13 5 11 89 35 81 972 413 552
Rehoboth
Beach 59.5 n/ a n/ a 43 n/ a n/ a 2054.5 n/ a n/ a
Gaithersburg 2.5 1.5 1 28 30.5 99.5 1049 254.5 257.5
Annapolis 7 7 9 14 20.5 1.5 2221.5 3368 3520.5
Source: Field counts by authors.
Table 6: Average Hourly Weekend Counts
Bicycle Volumes Pedestrian Volumes Motor Vehicle Volumes
Roundabout
Comparison 1
Comparison 2
Roundabout
Comparison 1
Comparison 2
Roundabout
Comparison 1
Comparison 2
East Lansing 15.5 3.5 n/ a 60 20 n/ a 602.5 198 n/ a
Rehoboth
Beach 88 n/ a n/ a 89 n/ a n/ a 2348 n/ a n/ a
Gaithersburg 3.5 7 7.5 28 89.5 286.5 1079.5 300 319
Annapolis 6 3.5 2.5 27.5 58 2 2377 2901 3017
Source: Field counts by authors.
However, it is difficult to determine from this data if bicyclists’ or pedestrians’ route choices are influenced by
the location of the roundabouts or a variety of other possible factors. The case study locations vary in terms
of land use, attractors, street networks, demographics and other factors that affect bicyclist and pedestrian
25
FINAL REPORT: TASK ORDER 6222
26
activity levels. It was also difficult to select comparison sites that had traffic volumes, land uses and activity
centers similar to those of the roundabout sites. Not surprisingly, the ratio of pedestrian and bicyclist activity
levels at roundabouts and traditional intersections varied across study locations. ( See Table 7 and Table 8 on
page 26.)
Among overall ratios ( Table 9), the highest roundabout/ comparison site ratio was in East Lansing, where
approximately two bicyclists were counted at the roundabout for every one bicyclist counted at the
comparison sites. At this site, 1.6 pedestrians were counted at the roundabout for every one pedestrian
counted at comparison sites. Bike lanes and sidewalks are provided on the approach to this roundabout, and
a bike ramp at each leg permits bicyclists to ride onto the sidewalk and use it while traveling around the
circular part of the roundabout. Though this roundabout is multi- lane with two entry lanes, its configuration
is not typical. The circular portion of the roadway has two lanes, with an outer lane striped as right- slip lanes
and an inner lane for vehicles 10 traveling past the first exit leg. The site is located near a large university.
The lowest ratio of bicyclist and pedestrian activity levels was found at the Gaithersburg site. The
roundabout at this site is close to housing developments, and the immediate vicinity consists of auto- oriented
retail stores and large parking lots. Motor vehicle volumes are significantly higher at the roundabout than at
comparison sites, suggesting that bicyclists and pedestrians may be avoiding motor vehicle traffic rather than
the roundabout itself. However, the East Lansing site also had high volumes of motor vehicle traffic at the
roundabout compared to the comparison sites.
Table 7: Weekday Ratios of Bicyclist and Pedestrian Counts at Roundabouts vs. Comparison Sites
Bicyclists Pedestrians
East Lansing 1.6 to 1 1.53 to 1
Rehoboth Beach n/ a n/ a
Gaithersburg 2.0 to 1 0.43 to 1
Annapolis 0.88 to 1 1.27 to 1
Source: Field counts by authors.
Table 8: Weekend Ratios of Bicyclist and Pedestrian Counts at Roundabouts vs. Comparison Sites
Bicyclists Pedestrians
East Lansing 4.43 to 1 3 to 1
Rehoboth Beach n/ a n/ a
Gaithersburg 0.48 to 1 0.15 to 1
Annapolis 2 to 1 0.92 to 1
Source: Field counts by authors.
Table 9: Overall Ratios of Bicyclist and Pedestrian Counts at Roundabouts vs. Comparison Sites
Bicyclists Pedestrians
East Lansing 2.19 to 1 1.64 to 1
Rehoboth Beach n/ a n/ a
Gaithersburg 0.71 to 1 0.22 to 1
Annapolis 1.18 to 1 1.01 to 1
Source: Field counts by authors.
10 Including bicyclists who choose to ride on the roadway, rather than on the separated path.
FINAL REPORT: TASK ORDER 6222
3.1.1. Milpas Corridor Analysis
Bicyclists and pedestrians did not appear to avoid the Milpas roundabout. On average, there were 122%
more bicyclists and 29% more pedestrians counted at the roundabout than would be expected from counts
on both ends. Bicyclists and pedestrians traveling on Milpas Street must travel 0.38 miles out of their way if
they wish to avoid the Milpas roundabout.
This suggests that bicycle and pedestrian activity may be more highly related to factors such as attractors, land
use, directness of bicycle routes, and other factors, rather than the presence of a multi- lane roundabout.
Bicyclists may be using the roundabout as a connection point to destinations such as the northern end of the
corridor where counts were high, or a grocery store. Additionally, the bicyclist volumes at the Milpas
roundabout were much higher than at the other Highway 101 interfaces, suggesting that the roundabout may
be a more desirable interface than either undercrossings on State or Garden Street. It may also be that Milpas
is a more popular destination.
These results should be considered in light of the limitations of the experiment. Variations in land use along
the corridors, destinations at the interfaces with the highway, and with the types of alternative routes available
to bicyclists and pedestrians may be primary reasons for the differences seen between the corridors. In
particular, the grocery store located at the Milpas Roundabout may be a key destination that attracts both
bicyclists and pedestrians. There are no nearby key destinations at the State Street or Garden Street
undercrossings. This finding suggests that land use/ attractors/ destinations may in fact override any effects
of intersection type on pedestrian and bicyclist route choice. All three of the interfaces with Highway 101
( undercrossing, signalized ramps, roundabout) are somewhat challenging for bicyclists, therefore it is possible
that the bicyclists using the three corridors are more experienced as cyclists.
3.1.2. Intercept Surveys
Intercept surveys found that self- reported comfort with multi- lane roundabouts differed by user mode.
Bicyclists were more likely than pedestrians to report feeling uncomfortable traveling through the
roundabout, with 18 percent of walkers saying they felt uncomfortable traveling through the roundabout
compared to 32 percent of bicyclists. However, a large percentage of respondents said they felt comfortable
traveling through the roundabout. Fifty- three percent of walking respondents reported being comfortable
traveling through the roundabout, compared to 60 percent of bicyclists. This study surveyed only those who
were using the roundabout and does not include people those who avoid the facility and take other routes
instead. Further research is necessary to find out more definitively if bicyclists and pedestrians avoid these
roundabouts
The final question related to respondents’ comfort of using roundabouts was: “ In general, do you change
your route to avoid traveling through a roundabout?” This question was asked separately about walking and
biking, of both pedestrians and bicyclists. The results are illustrated in Table 10. The majority of
respondents would not change their route to avoid the roundabout if walking.
For the bicycling question, three- quarters of respondents overall would not change their routes. Respondents
at Rehoboth Beach were the most likely to take another route when bicycling, in order to avoid the
roundabout. Also, people more familiar with bicycling ( i. e., those who were surveyed while bicycling) may be
generally more comfortable using a roundabout than other users. For example, 16 percent of people who
were bicycling when given the survey reported that they would change their route if walking, whereas 28
percent of people who were pedestrians at the study roundabout said they would change their route if they
were on bicycles.
27
FINAL REPORT: TASK ORDER 6222
Table 10. Roundabout Avoidance
All Data East Lansing, MI Rehoboth Beach, DE Annapolis, MD
if I were walking…
No. of
Participants %
No. of
Participants %
No. of
Participants %
No. of
Participants %
I would change my
route 12 14% 3 7% 8 22% 1 25%
I would not change
my route 75 86% 43 93% 29 78% 3 75%
Total 87 46 37 4
if I were biking…
No. of
Participants %
No. of
Participants %
No. of
Participants %
No. of
Participants %
I would change my
route 22 25% 7 15% 14 38% 1 25%
I would not change
my route 65 75% 39 85% 23 62% 3 75%
Total 87 46 37 4
3.1.3. Summary of Findings
Multi- lane roundabouts most likely do pose a deterrent to both bicyclists and pedestrians. Bicyclists are more
likely than pedestrians to change their route to avoid a multi- lane roundabout. Intercept surveys showed that
14% of respondents would change their route when walking to avoid a multi- lane roundabout, and 25% of
respondents would change their route when biking.
Intercept survey results show that pedestrians equally prefer four- way stop lights at intersections to
roundabouts, while more bicyclists reported preferring four- way stop lights at intersections. This is
supported by literature indicating that multi- lane roundabouts are considered a risk factor for bicyclists
( Furtado, 2004; Bruce and Larsson 2000).
Comparison counts of bicyclists and pedestrians at multi- lane roundabouts and nearby signalized
intersections did not show a consistent pattern of usage. However, this is may be due to the variations in
land use between the comparison sites, including the existence of important attractors such as grocery stores
Comparison of three freeway interfaces along parallel corridors in Santa Barbara showed that more bicyclists
were counted at the freeway interfaces than at the endpoints, suggesting that bicyclists were being funneled
toward these interfaces. Interestingly, the highest percentage increase was seen at the Milpas Roundabout,
( 122% at the roundabout for all count times combined) suggesting that issues of connectivity, directness, and
land use are more important than traffic considerations.
Similarly, a higher number of pedestrians were counted at the Milpas Roundabout than would be expected
from counts on each end, suggesting that the roundabout does not serve as a hindrance to walkers. On
average, 29% more pedestrians were counted at the roundabout than would be expected from the counts on
either end. In comparison, pedestrians along State Street and Garden Street were actually seen in lower
numbers at the highway interface than would be expected by the counts on either end. It is likely that for
most pedestrians— especially those shopping or making other discretionary trips— noise and land uses factors
near US 101 discourage walking.
28
FINAL REPORT: TASK ORDER 6222
3.2. Bicycle and Pedestrian Safety at Multi- Lane Roundabouts
Are pedestrians and bicyclists more likely to be involved in crashes or more severe crashes at multi-lane
roundabouts than at other types of intersections?
3.2.1. Literature Findings
As stated in Section 1.2. Relevant Related Research, several studies conducted in the U. S. and abroad have
examined roundabout safety but few focus on bicycle/ pedestrian/ auto safety. In a broad sense, researchers
have demonstrated that roundabouts typically perform better in terms of crashes and crash rates than
traditional forms of traffic control. Some specific insights into collision numbers, rates and severity can be
gleaned from a review of this literature.
Collision data at 39 U. S. roundabouts over 3.8 years showed that bicyclist and pedestrian collisions each
accounted for approximately 1% of collisions both single- lane and multi- lane roundabouts, with a total of five
reported pedestrian crashes and eight reported bicycle crashes ( NCHRP 572: Roundabouts in the United States).
While this study does not examine pedestrian and bicyclist volumes, it is assumed that bicyclist volumes are
much lower than pedestrian volumes, and that therefore the crash rate for bicyclists ( crashes per bicyclist
traveling through the roundabout) is even higher than indicated by the total crash numbers.
A study for the Australian Road Research Board analyzed bicycle crash data at multi- lane roundabouts over a
ten- year period ( 1995 to 2004). The study found that multi- lane roundabouts are safer for bicyclists than
traditional intersections and made some key points about bicycle collisions:
! Cyclists are over- represented by a significant factor in injury crashes at multi- lane roundabouts
! The predominant crash type is entering vehicle- circulating cyclists at multi- lane roundabouts ( 68% of
total bicycle crashes)
! Nighttime crashes with cyclists accounted for 25% of all cyclists’ crashes
! Thirty- nine of the 58 reported bicyclist crashes were injury crashes ( 67%)
! At locations with higher cyclist traffic, cyclist crash rates are lower ( drivers are more aware of their
presence)
! There are indications that reducing the speed differential between vehicles and cyclists should reduce
cyclists’ injury rates
! In the United Kingdom and the United States, studies have found that higher approach and entry
speeds have been found at roundabout locations with more approach visibility.
Research on pedestrian safety at multi- lane roundabouts is unfortunately limited and to some extent dated.
Brilon conducted a study of 32 newly- constructed single- lane roundabouts in Germany in the 1990s. While
he noted a 40% reduction in crash frequency and an even more impressive reduction in injury crashes, there
was only a small reduction in pedestrian crashes at the study locations. ( Brilon, 2005)
A review of the safety study by Lalani investigated the performance of 38 roundabouts in London, England
in the 1970s ( 20 mini- roundabouts; 9 small roundabouts; 5 large roundabouts and 4 double mini-roundabouts).
The authors compared the before and after safety performance of roundabouts with an
average study period of 19 months and found that pedestrian crashes were reduced by 46%, compared to the
traditional intersections that were replaced by the roundabouts. ( Lalani, 1975)
29
FINAL REPORT: TASK ORDER 6222
The most comprehensive study conducted in the U. S. was recently completed as part of NCHRP 3- 65
Roundabouts in the United States, published as NCHRP Report 572. As noted above, the occurrence of
pedestrian crashes was minimal at the study locations with only five reported pedestrian crashes out of 726
crashes collected from 55 study locations. This limited amount of data greatly hindered the researchers’
ability to draw any conclusions regarding pedestrian before/ after crash occurrences at roundabouts, except to
say that they accounted for approximately 1% of total crashes at the study locations.
3.2.2. Comparison of Collision Points between Signalized Intersections
and Multi- Lane Roundabouts
According to a synthesis of twenty- eight non- U. S. studies, roundabouts reduce injury accidents from between
30 and 50 percent, and fatal crashes from between 50 and 70 percent. ( Elvik) Many researchers attribute this
reduction in crashes to the reduction in potential collision points at roundabouts as compared to
conventional traffic control strategies.
When looking at bicyclist and pedestrian collision points at a traditional intersection in comparison to a multi-lane
roundabout, a different picture emerges. As can be seen in Figure 4 and Figure 5, bicyclists have more
collision points at multi- lane roundabouts than they do at the intersection of two four- lane roads. Pedestrians
have the same number of collision points at both intersection types.
At a signalized intersection and at a multi- lane roundabout, pedestrians have 12 potential collision points
where motorists must yield.
Bicyclists see an increase in collision points in signalized intersections compared to multi- lane roundabouts.
Bicyclists who travel the roundabout like a motor vehicle are exposed to 16 potential collision points where
motorists must yield to bicyclists. Bicyclists who travel the roundabout on a sidewalk or path, crossing like a
pedestrian, experience 12 potential collision points where motorists must yield. By comparison, bicyclists
traveling a four- way intersection like a motor vehicle see 12 collision points.
30
FINAL REPORT: TASK ORDER 6222
31
Figure 4: Bicyclist and Pedestrian Conflict Points at Signalized Intersection
Figure 5: Bicyclist and Pedestrian Conflict Points at Multi- Lane Roundabout
3.2.3. Crash Frequency and Severity
Table 11: Summary Collision Data 2002 through 2008 summarizes the collision data collected for the East
Lansing, Michigan and Santa Barbara, California roundabouts. 11 Collision data were collected for 2002
through 2008 for both locations. The two study locations are multi- lane roundabouts, each with two
circulating lanes. An estimated annual number of pedestrians and bicyclists was extrapolated from the bicycle
and pedestrian counts collected at the roundabouts during this study. As can be seen from the table, the
collision rate at the East Lansing location is more than double the collision rate at the Santa Barbara location.
It should be noted that the roundabout in East Lansing Michigan is designed with a multi- use cycle path that
allows bicyclists to travel around the roundabout separated from vehicle traffic, except when crossing the legs
or approaches to the roundabout. The Santa Barbara, California roundabout is an oval shaped roundabout
that serves the exit ramps from Highway 101. These geometric differences may account for the difference in
collision rates.
Table 11: Summary Collision Data 2002 through 2008
Roundabout location
# ped/ bike
collisions
# vehicle
collisions
Percent
ped/ bike
Estimated
annual # peds &
bikes**
Estimated collisions
per million
bicyclists/ pedestrians
East Lansing, MI* 5 42 11.9% 895,000 1.09
Santa Barbara, CA* 5 11 45.5% 1,825,000 0.47
* East Lansing collision data is from January 2002 to January 2008; Santa Barbara collision data is from 2002 to 2008
** Preliminary estimates, subject to change.
11 Collision data was not available to researchers for the Delaware and Maryland roundabouts.
FINAL REPORT: TASK ORDER 6222
3.3. Bicyclist, Pedestrian and Driver Behavior
How do bicyclists, pedestrians and drivers interact at multi- lane roundabouts? How does this relate
to safety, comfort of bicyclists and pedestrians?
3.3.1. Video Analysis
Results of the video analysis indicate an inverse relationship between vehicle traffic volume and pedestrian
volumes at roundabouts. Typically, the roundabouts with the least vehicle traffic volume had higher
pedestrian and bicycle volumes. Although we cannot conclude causality, it is possible that there would be an
adverse effect on pedestrian demand at locations where roundabouts are placed with the purpose of
optimizing road capacity.
Pedestrian Delay
At the Maryland roundabout, of 63 pedestrians observed, 31% waited to cross the entering lane and 29%
waited to cross the exiting lane. For those who waited, the average wait time before crossing entering lanes
was 3.9 seconds, while the average wait time before crossing exiting lanes was 4.1 seconds. This difference
was statistically significant for some of the roundabout legs only. At the Delaware roundabout, of 53
pedestrians, 43% waited to cross entering lanes of the roundabout and 57% waited to cross exiting lanes of
the roundabout. For those who waited, the average wait time to cross entering lanes of the roundabout was
4.7 seconds, while the average wait time to cross exiting lanes of the roundabout was 6.5 seconds. At the
Michigan roundabout, of 288 pedestrians observed, 98% waited when crossing entering lanes and all
pedestrians waited when crossing exiting lanes. These wait times were in the range of 0.1 to 17.7 seconds.
For those who waited, the average wait time for crossing entering lanes of the roundabout was 3.6 seconds,
while the average wait time for crossing exiting lanes of the roundabout was 5.6 seconds.
While the average wait times for both entering and exiting lanes at the Maryland roundabout were relatively
consistent, pedestrians at the Delaware and Michigan roundabouts experienced wait times 2 seconds longer
when crossing exiting lanes than when crossing entering lanes.
Overall, pedestrians experienced longer wait times when crossing exiting lanes than when crossing entering
lanes.
Yielding Behavior
Overall, most drivers who could yield to pedestrians at crosswalks did so, however some drivers did not yield.
The percentage of drivers who did not yield to pedestrians varied depending on location and lane and ranged
from 0- 100%. Lower driver yielding percentages tended to occur at roundabouts with lower pedestrian and
bicycle traffic volumes. This indicates a proportional relationship between vehicle yielding and pedestrian
demand at roundabouts. It also suggests that drivers pay more attention to pedestrians and bicyclists when
there is higher pedestrian and bicyclist volume.
Risk- Taking Behavior
Using the pedestrian assertiveness parameters as defined during the video analysis, it is difficult to pinpoint
factors that impact pedestrian risk- taking behaviors. However, not surprisingly, our observations of
pedestrian level of assertion and crossing pace suggest that pedestrians’ level of comfort may be related to
traffic volumes. For bicyclists, their level of assertiveness was measured by identifying the cyclist’s chosen
position in the lane, i. e., whether they rode in the center or edge of the lane. Results of the video analysis
suggest that bicyclists prefer to negotiate a roundabout on a separated bicycle path when such a path is
available. The analysis also shows that 76% of bicyclists did not ride in the center of the lane when travelling
on the road, choosing to ride at the outer edge of the lane instead.
32
FINAL REPORT: TASK ORDER 6222
3.3.2. Intercept Surveys
The results from the intercept surveys seem to be in agreement with those of the video documentation and
focus group portions of this project. Since these were intercept surveys, the pedestrians and bicyclists
interviewed were ones who were willing to use a multi- lane roundabout. Overall, 53 percent of pedestrians
reported being comfortable traveling through roundabouts and 60 percent of bicyclists reported being
comfortable traveling through roundabouts. On average, 30 percent of bicyclists biked on the roadway with
vehicles through the roundabout ( 8% in East Lansing; 54% in Rehoboth Beach; 0% in Annapolis).
3.3.3. Focus Group Analysis
As mentioned in the literature review, the presence of roundabouts in general is among the factors that
increase perceptions of risk for non- motorized travelers. ( Moller and Hels 2007), and this study’s focus
group results generally confirm this idea. Members of the focus group walked and biked through
roundabouts, but they said they find them risky and “ scary.” Bicyclists are concerned about conflicts with
motorists in the roundabouts while pedestrians are concerned about crossing the entry and exit lanes.
Separation of motor vehicle traffic and bicyclist and pedestrian traffic using a cycle track or shared use
pathway was seen as the best solution to bicyclist- vehicle conflicts, as expressed by both bicyclists and
pedestrians. This finding is echoed in the video documentation, which shows the majority of bicyclists using
a side path when one is available, and most bicyclists using the roadway when a side path is not available.
Enhanced crossing treatments were seen by focus group participants as potentially helpful to pedestrians.
Pedestrians in the focus group reported that they mostly navigate roundabouts assuming that drivers will not
yield to them, like at other intersections. Most wait until all vehicle traffic has cleared from the roundabout
before traveling across the entrance or exit of a roundabout leg. Many use the splitter island refuge to
perform a two- stage crossing. A few participants noted that they had changed their route to deliberately
avoid a roundabout.
Data gathered during this study show that non- motorized users-- particularly bicyclists-- are uncomfortable
using multi- lane roundabouts. Yet people are still using roundabouts, despite this discomfort. Whether this
discomfort relates to actual increased risk or just perception of increased risk is unclear from the focus group
comments. Bicyclists’ assertiveness as measured in the video analysis portion of this project shows that many
bicyclists are cautious when using multi- lane roundabouts. The majority of observed bicyclists rode on the
edge of the lane as opposed to the center of the lane, indicating a higher level of caution, discomfort, or lack
of understanding of the proper way to navigate a roundabout.
3.4. Summary of Findings
Findings from this study raise several important issues that should be considered when designing multi- lane
roundabouts:
Avoidance
! While 14% of pedestrians and 25% of bicyclists intercepted in the field stated that they would change
their route to avoid multi- lane roundabouts, in- field comparison counts did not show a significant
difference in pedestrian or bicyclist activity at roundabouts compared to traditional intersections
! Level of comfort with multi- lane roundabouts differs by user mode. Bicyclists were more likely than
pedestrians to report feeling uncomfortable traveling through the roundabout, with 32 percent of
bicyclists feeling uncomfortable traveling through the roundabout, compared to 18 percent of
pedestrians. Most respondents felt comfortable traveling through the roundabout ( 60 percent of
bicyclists and 53 percent of pedestrians.)
33
FINAL REPORT: TASK ORDER 6222
! People’s comfort level at a multi- lane roundabout appears to be affected by the age of the
respondent, the motor vehicle, bicycle and pedestrian volumes at the roundabout, and also the
geometric configuration of the roundabout. Of the three roundabout locations surveyed,
respondents at the East Lansing roundabout were most comfortable walking and biking through the
roundabout ( 62 percent). These respondents were generally young ( 69 percent between ages 18- 25)
and the roundabout has a shared- use path around the perimeter. In Rehoboth Beach, Delaware,
49% of people surveyed were comfortable using the roundabout. This roundabout has significant
bicycle and pedestrian activity ( 88 bicyclists and 89 pedestrians per hour). The relationship between
comfort at roundabouts and age should be explored further.
! When given the choice of stop controlled, signalized and roundabout intersections, pedestrians
equally prefer signalized intersections and roundabouts, but bicyclists prefer signalized intersections
and not roundabouts. Neither bicyclists nor pedestrians prefer four- way stop- controlled
intersections. The preference for more typical intersection types is probably not related to familiarity
with these types and unfamiliarity with multi- lane roundabouts; all focus group participants were
familiar with multi- lane roundabouts, and in the case of Maryland, had to travel through one or more
multi- lane roundabouts to access the focus group location.
Collisions
! While data is limited, some studies suggest that multi- lane roundabouts have little effect on
pedestrian crash numbers— either positively or negatively.
! While there are no U. S. studies on the subject, non- U. S. studies have shown that circulating bicyclist-entering
vehicle collisions are the most common bicyclist collision type in multi- lane roundabouts.
! Bicyclist and pedestrian crash rates, measured by crashes per million bicyclists/ pedestrians, vary at
different roundabouts in different locations ( e. g., 1.09 per million at East Lansing and 0.49 per
million at Santa Barbara.)
! European studies have found that pedestrian and bicyclist crashes account for only 1 percent each of
the total crashes at roundabouts. By contrast, bicyclist and pedestrian crashes in the case study
roundabouts accounted for a much larger percentage of total crashes ( 12 percent at Santa Barbara, 55
percent at East Lansing). This suggests that European roundabout design, bicycle and pedestrian
facility design, or driving, walking and biking behavior ( cultural acceptance, training, laws, and
familiarity) may have a role in reducing the number of bike and pedestrian collisions.
! European studies have shown that the four factors with the strongest effect on total crash rates in
roundabouts are total traffic volume, proportion of vehicles entering from the minor road, speed
limit, and number of legs.
! Very few conflicts were observed in video analysis of case study multi- lane roundabouts.
Behavior
! Based on video observations at case study multi- lane roundabouts, pedestrians overwhelmingly chose
to cross at a crosswalk, and many did not have to wait for a gap in traffic.
! Between 33 and 100 percent of pedestrians observed in the video analysis had to wait to cross a
roundabout leg, depending on the location. The wait times averaged 3.6 seconds for crossing
entering lanes and 5.6 seconds for crossing exiting lanes.
34
FINAL REPORT: TASK ORDER 6222
! Multi- lane roundabouts with higher bicyclist and pedestrian volumes see less wait time for
pedestrians and higher yielding rates by motorists.
! The majority of bicyclists observed riding in the circulating lane of a roundabout rode on the outer
edge of the lane, as opposed to controlling the lane, indicating discomfort, caution, or lack of
understanding of the proper way to navigate.
! When a shared- use path is provided around a roundabout, the majority of bicyclists choose to use
the path, rather than travel through the roundabout on the roadway.
! Bicyclists would prefer multi- lane roundabouts with vehicle speeds that are close to bicycling speed
( 12 to 15 mph).
3.5. Limitations of Research
There are limitations to the conclusions that can be drawn from this study, due to both the scope of the
research effort and the nature of the subject.
Lack of relevant data is the greatest obstacle to better understanding pedestrian and bicyclist safety at multi-lane
roundabouts in the U. S. Multi- lane roundabouts remain uncommon in this country and many are
relatively new, with little crash history. Where crash history is available, pedestrian and bicyclist volumes are
generally not collected, making it difficult to calculate a crash rate that accounts for pedestrian or bicyclist
exposure to motor- vehicle traffic. Because of limitations in police collision report forms, official crash
reports from pedestrian and/ or bicyclist- involved collisions at roundabouts may not be sufficient to conduct
a detailed analysis. For example, the CHP 555 form used in California does not contain fields that specifically
apply to roundabouts and it may be difficult for an officer to accurately describe a roundabout collision. The
form is also limited in terms of the number of fields that apply to pedestrians and/ or bicyclists, and efforts
have been underway for years to improve reporting of pedestrian and bicyclist- involved collisions. In
addition, because crashes, and especially crashes involving pedestrians, are rare, a long observation period is
required to observe enough crashes to conduct a meaningful analysis. As more multi- lane roundabouts are
constructed in the U. S., there will likely be more pedestrian and/ or bicyclist- involved crashes at roundabouts.
While unfortunate, these crashes will provide more data for future studies.
For this study, we referenced the literature on the European multi- lane roundabout experience, although it
must be noted that roadway design features will not necessarily translate well from one country to another.
The public’s lack of familiarity with roundabout operation, as noted in the focus groups, suggests that
Americans are still generally confused by roundabouts, especially multi- lane roundabouts. Attitudes toward
walking and bicycling also vary considerably between the U. S. and various European countries-- factors which
may affect the safety performance of a roundabout.
This analysis relies on both observational and self- reported data. As with any research that involves self-reporting,
including surveys and focus groups, self- reporting bias may affect the validity of findings. The
topic of this study is not sensitive or controversial enough to suggest that respondents might avoid truthful
responses. The survey questions were designed and written to minimize bias, and focus groups were
conducted in an informal atmosphere, with participants assured of the purpose of the meeting.
Through the comparison counts, the Milpas corridor analysis, surveys, and focus groups, we sought to
ascertain whether pedestrians and/ or bicyclists actively avoid traveling through multi- lane roundabouts. The
focus groups confirmed that these users sometimes avoid roundabouts, but their route choice is also greatly
influenced by factors of directness, land use, the existence of attractors, and of alternative route choices. In
the extreme case pedestrians and bicyclists will avoid a roundabout with the consequence of not walking or
35
FINAL REPORT: TASK ORDER 6222
biking. This would result in zero pedestrian or bicyclist- involved crashes but it is difficult to quantify the true
impact on these users. Above all, land use factors seem to be the most influential variable in pedestrian’s and
bicyclists’ choice of route, relative to multi- lane roundabouts.
36
FINAL REPORT: TASK ORDER 6222
4. Recommendations
4.1. Introduction
This chapter presents recommendations for installing multi- lane roundabouts and recommends design
features to accommodate bicyclists and pedestrians at multi- lane roundabouts. The design recommendations
are based on several existing design resources:
! FHWA’s Roundabouts: An Informational Guide FHWA- RD- 00- 067
! Caltrans Design Information Bulletin ( DIB) 80- 01
! AASHTO Guide for the Planning, Design, and Operation of Pedestrian Facilities
! AASHTO Guide for the Development of Bicycle Facilities
! California Manual on Uniform Traffic Control Devices
! Proposed Amendments to Federal MUTCD
! Draft Public Rights- of- Way Accessibility Guidelines ( 2005)
! FHWA memo on Public- Rights- of- Way ( 2006)
We have used insights gathered from our research to clarify and supplement recommendations presented in
the documents above and in emerging international and domestic research. Engineers should use
professional judgment when applying these recommendations to multi- lane roundabouts. Most treatments
are already accepted for use at other locations, but have not yet been applied or studied for multi- lane
roundabouts specifically.
This chapter is divided into the following sections:
Section 4.2. A comparison of conflict points, capacity and other elements of multi- lane roundabouts
to signalized intersections
Section 4.3. A summary of operational problems faced by bicyclists and pedestrians when
navigating multi- lane roundabouts
Section 5.4. Considerations for meeting the safety, comfort and mobility needs of bicyclists and
pedestrians when planning and designing multi- lane roundabouts, including locations where
roundabouts may not be appropriate.
Section 4.5. Recommended designs for accommodating bicyclists and pedestrians at multi- lane
roundabouts
Section 4.6. A summary of recommended designs and illustrations of example roundabout designs
4.1.1. Documents Governing the Design of Multi- Lane Roundabouts in
California
Roundabout design in California shall follow Federal Highway Administration’s technical publication,
Roundabouts: An Informal Guide ( Guide), published in June 2001, and the Caltrans Design Information Bulletin
( DIB) 80- 01 and its Attachment A, both of which specify guidelines and considerations for roundabout
37
FINAL REPORT: TASK ORDER 6222
design. The text provided in DIB 80- 01 Attachment A shall govern in every instance where conflicts arise or
ambiguities exist between the Guide and the Highway Design Manual ( HDM) or the Traffic Manual.
As stated in Caltrans DIB 80- 01, “ roundabouts need to be evaluated and designed on a case- by- case basis,
taking into consideration the physical characteristics of the location, the orientation of the approaches to the
circular intersection, the existing and proposed intersection operating conditions, plus the safety and mobility
needs of all motorists, bicyclists, and pedestrians that will be using the facility.
Deputy Directive 64, R- 1, issued October 2008, requires the safety and mobility needs of bicyclists and
pedestrians to be addressed in all projects on the State Highway system, regardless of funding. As stated in
DD 64- R1, " The intent of this directive is to ensure that travelers of all ages and abilities can move safety and
efficiently along and across a network of ‘ complete streets.’” DD- 64 R1 defines a complete street as " A
transportation facility that is planned, designed, operated, and maintained to provide safe mobility for all
users, including bicyclists, pedestrians, transit riders, and motorists appropriate to the function and the
context of the facility." In compliance with DD 64 R- 1, all roundabouts located on state highways must
address the needs of bicyclists and pedestrians in the planning, design, operation and maintenance stages.
In addition, several other design guides are applicable to the planning and design of roundabouts, including
the American Association of State Highway and Transportation Officials ( AASHTO) Guide for the Planning,
Design, and Operation of Pedestrian Facilities ( AASHTO Ped.), published in July 2004, the AASHTO Guide
for the Development of Bicycle Facilities ( AASHTO Bike), published in 1999, U. S. Access Board in Draft
PROWAG, and the California Manual on Uniform Traffic Control Devices ( CA MUTCD), published in
September 2006.
The FHWA has a Notice of Proposed Amendment regarding changes to the next edition of the MUTCD
which went out for comment in July 2008 and a Final Rule for the next edition of the MUTCD is anticipated
in 2009. Proposed amendments for the next edition of the MUTCD include recommendations for
roundabouts signing and striping.
4.2. Comparison of Bicyclist and Pedestrian Experience at Multi- Lane
Roundabouts and Signalized Intersections
The user experience for bicyclists and pedestrians greatly varies between multi- lane roundabouts and
traditional intersections. Three primary factors that contribute to bicyclists’ and pedestrians’ discomfort at
roundabouts are: ( 1) the fact that crossings are uncontrolled, ( 2) the higher number of conflict points for
bicyclists at multi- lane roundabouts and ( 3) the longer distance required to travel through these intersections.
38
FINAL REPORT: TASK ORDER 6222
Table 12: Matrix Comparing Multi- Lane Roundabouts to Traditional Intersections
Multi- Lane Roundabout Signalized Intersection of two
four- lane roads*
Signalized Intersection of
two four- lane roads with
right- slip turn*
Bicyclist
distance to
traverse
Generally longer than signalized
intersections Shorter than roundabouts Shorter than roundabouts
Pedestrian
distance to
traverse
Generally longer than signalized
intersections Generally shorter than roundabout Can be the same as
roundabout
Bicyclist delay
Function of motor vehicle volumes
and yielding behavior, but can be
shorter than signalized intersections
Function of signal timing Function of signal timing
Pedestrian delay
Function of motor vehicle volumes
and yielding behavior, but can be
shorter than signalized intersections
Function of signal timing Function of signal timing
Motor vehicle
conflict points Fewer, fewer T and head- on conflicts More, T and head on conflicts are
possible
More than a roundabout, T
and head on conflicts as well
as conflicts associated with
right slip turn
Bicycle conflict
points
More, many merging/ weaving
conflicts
Fewer merging/ weaving conflicts
than a roundabout
More than a roundabout,
some crossings are
uncontrolled
Pedestrian
conflict points
Same number, crossings are
uncontrolled
Same number as a roundabout,
crossings are controlled
More than a roundabout,
some crossings may be
uncontrolled
* These intersection types are lower capacity than multi- lane roundabouts.
As referenced earlier in this report, a reduction of conflict points in comparison to a traditional intersection is
a commonly- cited benefit of roundabouts. This is not true of multi- lane roundabouts. Bicyclists experience
more conflict points at multi- lane roundabouts than at traditional four- way intersections, and pedestrians face
an equal number or slightly fewer collision points See Section 3.2.2. for more details.
4.3. Operational Issues for Pedestrians and Bicyclists at Multi- Lane
Roundabouts
Several design challenges must be addressed when designing multi- lane roundabouts to address the safety and
mobility needs of bicyclists and pedestrians. The challenges for each mode of travel are discussed separately
to clarify the issues unique to each mode, as follows:
4.3.1. Issues for Pedestrians
Multi- lane roundabouts pose the following challenges for pedestrians:
! Pedestrians crossing the exit lane must be able to correctly judge whether a circulating motorist is
going to exit, to correctly judge the speed of that motorist and judge whether the driver is going to
yield.
! At multi- lane roundabouts, pedestrians often have to make judgments about more than one vehicle
when crossing.
39
FINAL REPORT: TASK ORDER 6222
40
! Every crossing represents a potential “ multiple threat” scenario, in which a motorist in the lane
nearest the pedestrian yields and blocks the view of the pedestrian, while the motorist in the second
lane does not yield. 12
! Pedestrians traveling counterclockwise around the roundabout must look behind them to check for
circulating vehicles that may exit.
! Motorists approaching the roundabout are looking to the left for a gap in traffic and are less likely to
notice pedestrians trying to cross from their right.
! A pedestrian’s path of travel is longer at multi- lane roundabouts than at a signalized intersection.
! Pedestrians who are blind or who have low vision may have greater difficulty discerning appropriate
crossing times, resulting in more risk and delay, because the sound of other vehicles traveling around
the roundabout, or a vehicle yielding in one lane, may mask the sound of a vehicle approaching the
crosswalk.
! Bicyclists using shared side paths or sidewalks to travel around a roundabout can cause conflicts with
pedestrians, in particular pedestrians with disabilities.
! Pedestrians crossing the entry lanes must be able to correctly judge the speed of approaching vehicles
and judge whether the drivers are going to yield.
4.3.2. Issues for Bicyclists
Bicyclists can travel through a multi- lane roundabout three ways: ( 1) riding on the sidewalk or separated
shared use path, ( 2) like a motorist, controlling the lane and circulating on the roundabout within a traffic
lane, or ( 3) dismounting and walking through the roundabout as a pedestrian.
Bicyclists who ride on a separated path around a roundabout have similar challenges to pedestrians, with
three key differences:
! Bicyclists travel several times faster than pedestrians. Motorists are not expecting to see a person
travel so quickly into a crosswalk, and may not be able to yield in time.
! Unlike pedestrians, bicyclists cannot easily stop or jump out of the way if they notice a motorist that
is not going to yield.
! The California Vehicle Code does not specifically protect bicyclists riding on sidewalks and in
crosswalks. Bicyclists riding crosswalks are not given the same legal rights and responsibilities as
pedestrians and, as such, may be found liable if they are involved a collision while riding in a
crosswalk.
Multi- lane roundabouts pose the following challenges to bicyclists who control the lane:
12 Passing a vehicle that is stopped for a pedestrian in a crosswalk is a violation of CVC Section 21951, which states, “ Whenever any
vehicle has stopped at a marked crosswalk or at any unmarked crosswalk at an intersection to permit a pedestrian to cross the roadway
the driver of any other vehicle approaching from the rear shall not overtake and pass the stopped vehicle.”
FINAL REPORT: TASK ORDER 6222
41
! Bicyclists must control the lane before they enter the roundabout to avoid becoming caught in a
“ right hook,” a situation in which a motorist turns right, across the path of a bicyclist traveling
straight. Entry leg speeds must be slow enough for bicyclists to be able to control the lane safely.
! Theoretically, once motor vehicle volumes reach a certain magnitude, there are no gaps in traffic
large enough to accommodate a bicyclist.
! Bicyclists must be able to correctly judge the speed of circulating motorists to find a gap that is large
enough for them to safely enter the roundabout. This task is particularly difficult if the circulating
motorists are traveling at a much higher speed than the bicyclists. In addition, if circulating speeds in
a roundabout are much higher than 20 mph, drivers behind a bicyclist may become impatient, and
may pass the bicyclist and turn in front of him, creating more risks for the bicyclist.
! As a circulating bicyclist approaches an entry lane, a driver waiting to enter must notice the bicyclist,
properly judge the bicyclist’s speed, and yield to him/ her if necessary. In a location where there are
few bicyclists, motorists may not even register that there is a bicyclist approaching. If a bicyclist is
hugging the curb, s/ he may be outside the motorist’s cone of vision.
4.4. Considerations When Planning Multi- Lane Roundabouts
Planners and designers should consider the safety, comfort and mobility of bicyclists and pedestrians when
determining if a multi- lane roundabout is appropriate, and should design multi- lane roundabouts that balance
the needs of drivers with the needs of bicyclists and pedestrians. In some cases, a multi- lane roundabout may
not be the best option for bicyclists and pedestrians, and alternative designs should be considered.
Designers and planners should consider alternatives to multi- lane roundabouts in the following
situations:
! As noted in the Guide, “ heavy pedestrian or bicycle movements in conflict with high traffic volumes”
may preclude a roundabout at a specific location. ( Section 3.3.2) The Guide also lists general ways in
which issues can be resolved, including design features, operational changes, and in some cases,
specific mitigation actions. The Guide does not go into further detail.
! Planners should consider alternatives to multi- lane roundabouts located within the walkshed or
bikeshed of schools serving younger students. If multi- lane roundabouts are placed within the
walkshed or bikeshed of school serving younger students, designers should consider signalizing the
roundabout. 13
! Planners should consider alternatives to multi- lane roundabouts located within walking vicinity of a
senior center, nursing home, or other facility that serves the elderly. If multi- lane roundabouts are
placed within walking vicinity of a facility that serves the elderly, designers should consider
signalizing the roundabout. 14
13 While there is little specific information on the safety of children navigating roundabouts, this demographic crash involvement rates
are highest among males aged 5 to 9 years old ( PedSafe), and elementary school- aged children have a narrower field of vision than
adults and often over- estimate their physical abilities ( uwhealth. org)..
14 While there is little specific information on the safety of seniors navigating roundabouts, collisions involving seniors are much more
likely to include severe injuries and fatalities than collisions involving younger pedestrians. Pedestrian collisions resulting in death are
greater than 20 percent for adults over 75 years of age as compared to less than 8 percent for pedestrians under age 14 ( Pedsafe).
FINAL REPORT: TASK ORDER 6222
42
Designers and planners should provide enhanced pedestrian and bicycle facilities at the following
locations:
! Multi- lane roundabouts within a quarter mile walking distance of pedestrian generators, ( e. g. transit
stops, shopping districts, universities, etc…) should be designed with best practice pedestrian and
bicycle treatments, including treatments identified in this document.
! If pedestrian generators are developed within a quarter mile walking distance of an existing multi- lane
roundabout, the existing roundabout should be evaluated for accommodation of pedestrians and
necessary treatments should be installed.
In general, all roundabouts should be designed to meet the following recommendations:
! Multi- lane roundabouts should be designed to accommodate bicyclists of all abilities, through the use
of lowest practical design speeds and best practice bicycle treatments, including treatments described
in this document.
! Multi- lane roundabouts should be designed to accommodate pedestrians of all abilities, through the
use of lower design speeds, signage, striping and traffic control devices to increase yielding at
crosswalks, and best practice pedestrian treatments including treatments described in this document.
! Particular attention should be given to accommodating pedestrians with disabilities at multi- lane
roundabouts. In particular, engineers and designers are urged to follow the progress of NCHRP 3-
78, which is conducting a very large- scale in- field test of various roundabout treatments to enhance
the experience of visually impaired pedestrians. Findings from the study are expected in December
2009.
4.5. Designing Multi- Lane Roundabouts to Meet the Needs of Bicyclists
and Pedestrians
This section of the report is intended to help engineers and planners better understand ways in which multi-lane
roundabouts can be designed to accommodate pedestrians and bicyclists. The recommendations
presented in this chapter are a compilation of emerging best practices from across the United States and from
other countries, as well as recommendations based on results of our research gathered through focus groups,
surveys and field observations.
4.5.1. General Design Goals
The design recommendations in this section are based on the following design goals.
1. Design roundabouts to accommodate on- street bicyclists by reducing the speed differential between
circulating motorists and bicyclists. We recommend a 25 mph maximum circulating design speed. 15
2. Design approaches and exits to the lowest speeds possible, in order to reduce the severity of
potential collisions with pedestrians.
15 A 2004 FHWA study that collected field data in 21 locations around the United States measured the 85th percentile speed of
bicyclists at 14 mph. http:// www. tfhrc. gov/ safety/ pubs/ 04103/ index. htm
FINAL REPORT: TASK ORDER 6222
43
3. Design roundabout approaches, circulating lanes and exits to encourage bicyclists navigating the
roundabout like motor vehicles to control the lane. This approach reduces the chances of a bicyclist
being cut off by a “ right hook.”
4. Utilize the most practicable and effective tools to maximize yielding rate of motorists to pedestrians
and bicyclists at crosswalks.
5. Provide separated facilities for bicyclists who prefer not to navigate the roundabout on the
roadway. 16
6. Use appropriate signing, roadway markings and geometric design to clearly indicate to drivers,
bicyclists and pedestrians, the right- of- way rules and correct navigation at a multi- lane roundabout.
4.5.2. Geometric Design of Multi- Lane Roundabouts
Through literature reviews, observations, and surveys, three key components of geometric design have been
identified that affect safety and perceived safety of pedestrians and bicyclists at multi- lane roundabouts:
! Vehicle design speed ( on entry, circulating, and on exit)
! Visibility of pedestrians and bicyclists ( both to see vehicles and to be seen by drivers)
! Width of roundabout lanes
These components are addressed in the following three sections.
4.5.3. Design Speed
Vehicle speeds on approach to, into, through, and on exit of a roundabout are influenced by the various radii
chosen by the roundabout designer. The FHWA guidance document specifies five specific path radii to be
selected by the designer to meet the overall objectives of roundabout performance. These are identified in
Figure 6.
R1 – entry path radius
R2 – circulating path radius
R3 – exit path radius
R4 – left turn path circulating radius
R5 – right turn path radius
16 Even though a separated path may be provided, some bicyclists may choose to travel through the roundabout with motor vehicles,
and are legally allowed to do so. Roundabouts should, whenever possible, be designed to accommodate bicyclists on the approaches
and circulating roadway.
FINAL REPORT: TASK ORDER 6222
Figure 6: Five Key Radii that Affect Design Speed
Source: FHWA Roundabouts: An Informational Guide
Information from NCHRP Report 572 has further refined the speed prediction models which were
previously based on AASHTO Policy on Geometric Design of Streets and Highways.
V = 15R( e + f ) Equation 1
where
V = speed ( mph),
R= radius ( ft),
e= superelevation ( ft/ ft), and
f= side friction factor.
NCHRP Report 572 further simplified the relationship between speed, radius, superelevation and side friction
factor by assuming the use of common superelevation factors of + 0.02 and - 0.02:
V = 3.4415R0.3861 , for e = + 0.02 Equation 2
V = 3.4614R0.3673 , for e = – 0.02 Equation 3
where V = predicted speed ( mph), and
R = radius of curve ( ft).
Speed data collected at several study locations helped researchers of NCHRP Report 572 to determine the
variations that exist between basic horizontal speed estimates and driver selected speeds through
roundabouts. Two additional suggested changes were made to the exit and entry speed prediction models, in
which the AASHTO relationship was found to over- estimate these speeds at roundabouts.
Selection of design speed is typically related to factors in the environment in which the roundabout is to be
placed, posted speed limits on the approaches, capacity requirements, and available sight distance for entering
drivers. Current guidance from the FHWA Roundabout Guide recommends higher entry speeds for locations
in rural areas and for urban double- lane roundabouts ( 30 and 25 mph, respectively) ( FHWA Guide).
44
FINAL REPORT: TASK ORDER 6222
Recommendations
1. When designing roundabout radii, use AASHTO relationship between radius,
superelevation, and friction to estimate speeds ( see Equation 1, above) rather than
relationships from NCHRP Report 572.
Rationale: The AASHTO equation provides higher predicted speeds than the equation in NCHRP
Report 572. Given the relationship between bicyclist and pedestrian injury severity and speed, we
recommend using the more conservative AASHTO equation.
2. Design roundabout entries that slow motor vehicles to spee