Upload
chester-gordon
View
219
Download
0
Tags:
Embed Size (px)
Citation preview
Investigation of the Safety Effects of Edge and Centerline Markings on Narrow, Low-Volume Roads
Lance Dougald
Ben Cottrell
Young-Jun Kweon
In-Kyu Lim
Outline
• Background• Purpose• What we did• What we found• Next steps
2
Pavement Markings
3
Introduction and Background
Evolving Problem – Increase in housing/retail development – Higher traffic volumes on nearby narrow roads
without markings– Thought: Improve safety with low cost
striping until higher cost design improvements could be implemented
4
Introduction and Background
• VDOT Guidelines:– Centerlines
• ≥ 500 vpd, ≥18 ft width
– Edgelines• Primary and Secondary routes w/no curb & gutter• Minimum 20 ft width• Centerlines present
* where an engineering study indicates a need
5
Introduction and Background
• MUTCD requirements:– Centerlines
• Urban arterials and collectors ≥ 20 ft width,
ADT ≥ 6,000 vpd (recommended ≥ 4,000)• All two-way streets with 3+ traffic lanes
– Edgelines• All freeways and expressways• Rural arterials ≥ 20 ft width, ADT ≥ 6,000 vpd
(recommended ≥ 3,000)
6
Purpose and Scope
• Initially: to develop guidelines for marking edge and centerlines on low volume roads (≤ 3,000 vpd and 16-20 ft wide pavement)
• Two phase process– Phase I: investigate safety effectiveness of
markings using cross-sectional crash data– Phase II: before/after pilot study of edge and
centerline applications at candidate sites and B/C analysis.
7
Methods
1. Conduct literature review (focusing on rural/suburban low volume roads)
2. Obtain information from other state DOTs
3. Develop inventory of current edge and centerline markings and database of crash history on narrow roads
4. Perform crash analysis (5 years of data – width, AADT, and presence of pavement markings) [cross-sectional analysis]
8
Lit Review and Survey Results
• Variability found in past research– Crashes– Speed– Lateral positioning
• Variability in state DOT practices/policies– Majority follow MUTCD – Lower width/ADT thresholds from states that
maintain secondary road system
9
Safety Impact of Edgelines on Rural Two-Lane Highways in Texas
• A before/after comparison study found edgelines reduced accident frequency up to 26%
• highest safety impacts on curved roadways with lane widths of 9-10 ft
10
Benefit-Cost Analysis of Lane Marking
• paint costs $0.04 /linear-ft in rural areas and $0.07/linear-ft in urban areas
• existing longitudinal pavement markings reduce crashes by 21%
• edgelines on rural two-lane highways reduce crashes by 8 %
11
Lure of Low Cost Markings
Using $0.07 per linear-ft:• $740 per mile centerlines • $1,480 per mile center and edgelines
VDOT HSIP average crash values• PDO: $9,000 B/C=6/1• Injury :$55,000-$275,000 B/C=37/1-
185/1 • Fatal :$5,000,000 B/C=33,784/1
12
Kentucky DOT Guidelines
13
Use of Edge Line Markings on Rural Two Lane Highways. 2008.
Pavement Width (ft)
Lane Width (ft)
Centerline EdgelinePaved
Shoulder Width (ft)
28 12 Yes Yes 2
27 12 Yes Yes 1.5
26 11 Yes Yes 2
25 11 Yes Yes 1.5
24 11 Yes Yes 123 10 Yes Yes 1.5
22 10 Yes Yes 1
21 9 Yes Yes 1.5
20 9 Yes Yes 1
19 8 Yes No 1.5
18 8 Yes No 1
17 7.5 No Yes 1
16 7 No Yes 1
15 6.5 No Yes 1
14 6 No Yes 1
State DOTs’ actions to increasing volumes on narrow roads
• Wyoming adds centerlines• Delaware was under pressure in two
counties to establish a low AADT criterion for centerlines
• Oregon adds edgelines if there are documented safety problems
15
Inventory/Crash History
• HTRIS – 3 subsystems– Roadway Inventory (200,000 segments)– Accident/Crash (FR-300)– Traffic Monitoring System (CCS)
• 2004-2008 Crash Data– Eligible narrow and low volume segments– Presence of pavement markings (Google)
16
17
HTRIS
RDISubsystem
Select segmentsmaintained in2004 - 2008
2004 RDI2005 RDI2006 RDI2007 RDI2008 RDI
Select segmentsTwo-lane, Two way
& Undivided
Select segmentspavement width
16ft - 20ft
TMSSubsystem
Select segments with AADT
1 - 3,000 veh.
Select segments onPrimary & Secondary
2004 TMS2005 TMS2006 TMS2007 TMS2008 TMS
ACCSubsystem
Compile RDI, TMS and ACC
data
Select segments with crashes in
2004 - 2008
AddMarking Inventory
2004 ACC2005 ACC2006 ACC2007 ACC2008 ACC
Step 1
Step 2
Step 3
Step 4
Step 5
Step 6
Step 7
Step 8
Step 9
Join TMS to RDI
Join ACCto RDI
Step 10
Number of Lanes & Facility
Type
Route System
Pavement Width (ft)
Length (mi)
Total Crash
Frequency2004-2008
Two-lane
Undivided
Primary 16 20 6617 11 818 162 53619 68 21820 1,456 5,182
Secondary 16 122 25217 9 1218 351 32719 32 4020 285 166
Total 2,516 6,807
18
Inventory/Crash History
• Matrix:– Pavement width (16’,18’,20’)– AADT band (<500, 501-3000)– Presence edge/centerlines– Number/length of segments– Crashes
• Total• Road Departure• Density
19
Road Sections by Two AADT Bands and Presence of Pavement Markings
AADT Bands Centerline Edgeline Number of
SegmentsLength (miles)
≤ 500No
No 1,910 496Yes 12 6
YesNo 27 25Yes 232 194
501-3,000No
No 171 40Yes 6 3
YesNo 212 88Yes 2,227 1,180
Total 4,797 2,033
20
No Markings and Edgelines Only
21
Centerlines Only and Both Markings
22
23
Crash Analysis
Question:
Are narrow roads with pavement markings safer than those without pavement markings?
• Compared roads with:– No markings– Centerline markings only– Edge-line markings only – Edge and centerline markings
24
Crash Analysis
ANOVA:• Developed individual models for widths:
– 16’, 18’, 20’• Crash frequency (# of crashes)• Crash density (crashes/mile)• Crash rate (crashes/mile/vehicles)
25
Number of segments and marking presence
Number of Segments EdgelinesAbsent Present
Centerlines Absent 2,081 18Present 239 2,459
26
Crash frequency and marking presence
Crash Frequency(5-year crashes per
segment)
EdgelinesAbsent Present
Centerlines Absent 0.13 0.28Present 1.43 2.03
27
Crash density and marking presence
Crash Density(5-year crashes per 0.5
mile)
EdgelinesAbsent Present
Centerlines Absent 0.15 0.19Present 1.16 1.48
28
Crash Rate by marking presence
Crash Rate(5-year crashes per 0.5 mile per 1,000 vehicles)
EdgelinesAbsent Present
Centerlines Absent 0.77 0.96Present 1.11 1.11
29
Number of segments by marking presence and pavement widthNumber of Segments Edgelines
Absent Present
16 feet Centerlines Absent 118 4Present 21 56
18 feet Centerlines Absent 737 9Present 61 284
20 feet Centerlines Absent 1226 5Present 157 2119
30
Crash frequency by marking presence and pavement width
Crash Frequency(5-year crashes per
segment)
EdgelinesAbsent Present
16 feet Centerlines Absent 0.42 0.25Present 2.62 2.66
18 feet Centerlines Absent 0.20 0.33Present 1.70 1.77
20 feet Centerlines Absent 0.05 0.20Present 1.16 2.04
31
Crash rate by marking presence and pavement width
Crash Rate(5-year crashes per 0.5 mile
per 1,000 vehicles)
EdgelinesAbsent Present
16 feet Centerlines Absent 1.69 0.41Present 2.66 2.01
18 feet Centerlines Absent 1.11 1.65Present 1.58 1.56
20 feet Centerlines Absent 0.47 0.17Present 0.72 1.03
32
Crash Analysis Results
ANOVA:• No statistical difference found for crash
frequency, density, and rate for each of the 4 marking scenarios
• Shortcoming: accounting for AADT, segment length
33
Safety Performance Function (SPF)
A safety performance function (SPF) is an equation used to predict the average number of crashes per year at a location as a function of exposure and, in some cases, roadway or intersection characteristics (e.g., number of lanes, traffic control, or median type)
SPF estimated by the negative binomial model
34
Crash Analysis
• Separate SPFs developed for 4 cases:• No lines present • Centerlines only• Edgelines only• Both lines present
35
Crash Analysis• SPF results
• 3 segment groups appear different
36
0
1
2
3
4
5
6
7
8
0 500 1,000 1,500 2,000 2,500 3,000
5-Y
ear
Cra
sh F
requ
ency
per
0.5
Mile
AADT
Crash Analysis• SPF results for 95% upper limit
• However, 3 curves are statistically identical
37
0
2
4
6
8
10
12
14
16
0 500 1,000 1,500 2,000 2,500 3,000
5-Y
ear
Cra
sh F
requ
ency
per
0.5
Mile
AADT
Study sites vs all VDOT roads
no. road milesno. crashes 2004-2008
study sites 2,033 4,797VDOT roads 60,066 251,609% 3.4% 1.9%
38
Pavement Marking Inventory
• how districts maintain a pavement marking inventory in terms of software, format, data quality and frequency of updating the inventory vary
• 2 districts did not have an inventory for these roads
• a uniform, up to date pavement marking inventory may have value
39
Discussion: Perspectives
• HJR 243:– “the Virginia Department of Transportation be
requested to revise its standards for the provision of centerline pavement markings to include all appropriate secondary roads having a pavement width of 18 feet or more where official traffic counts indicate a minimum of 500 vpd.”
40
Discussion: Perspectives
• The term “appropriate” may allow for guidance to be developed on what roads to mark and how to mark them
• Flexibility in HJR 243 to mark more roads? Mark fewer roads?
• Interpretation on the word “appropriate” is needed
41
Limitations
• Before/after Empirical Bayes study ideal– Select and mark roads then wait for after data– Low number of crashes typical– Large number of road sections = impractical
• Driver’s behavior not addressed– Exploratory study utilizing data from VTTI’s
naturalistic driving study
42
Conclusions
• High variability exists in state DOT practices for installing pavement markings on narrow roads
• Based on cross-sectional analysis of crash frequency, density, rate, and SPF prediction, there appears to be no statistical difference between segments with and without centerlines and/or edgelines
43
Recommendations
1. VDOT’s Traffic Engineering Div. should consider developing a statewide process for a pavement marking inventory.
2. VDOT’s TED should consider asking the Office of the Attorney General for an interpretation/opinion of the term “appropriate” in HJR 243.
44
Recommendations
3. VCTIR staff should consider an exploratory study to determine if data from VTTI’s naturalistic driving study may be used to evaluate driver behavior on narrow roads with and without centerlines and/or edgelines.
45
Investigation of the Safety Effects of Edge and Centerline Markings on Narrow, Low-Volume Roads
http://www.virginiadot.org/vtrc/main/online_reports/pdf/14-r3.pdf