A presentation given at the 2016 Traffic Safety Conference during Breakout Session 12: Engineering Research Findings. By Karen Dixon, Senior Research Engineer, Roadway Safety, Texas A&M Transportation Institute

1. Reducing Lane and Shoulder
Width to Permit an Additional
Lane on a Freeway
June 7,2016
Karen Dixon, Ph.D., P.E.
Kay Fitzpatrick, Ph.D., P.E.
Raul Avelar, Ph.D., P.E.

2. Presentation Overview
• Background Information
• Operational Analysis
• Safety Analysis
• Questions
2

3. Typical Lane and Shoulder Widths
Agency Lane
Widths
Shoulder Widths
AASHTO 12 feet Right: 10 feet
Left: 4 feet (2 through lanes),
10 feet (≥ 3 through lanes)
TxDOT 12 feet Same as AASHTO
3

4. Operational Effects – Lane Width
Average Lane Width
(feet)
Reduction in FFS (mph)
≥ 12 0.0
≥ 11 – 12 1.9
≥ 10 – 11 6.6
FFS Adjustment based on Freeway Lane Width
Source: Highway Capacity Manual, Exhibit 11-8, p. 11-11 (TRB, 2010)
4

5. Operational Effects – Shoulder
Width / Lateral Offset
Right-Side
Lateral Clearance
(feet)
Lanes in One Direction
2 3 4 ≥ 5
≥ 6 0.0 0.0 0.0 0.0
5 0.6 0.4 0.2 0.1
4 1.2 0.8 0.4 0.2
3 1.8 1.2 0.6 0.3
2 2.4 1.6 0.8 0.4
1 3.0 2.0 1.0 0.5
0 3.6 2.4 1.2 0.5
FFS Reduction (mph) for Freeway Right-Side Lateral Clearance
Source: Highway Capacity Manual, Exhibit 11-9, p. 11-12 (TRB, 2010)
5

6. Operational Effects – Influential Factors
or Performance Measures to Note
• Physical dimensions (including minimum
values present) and bounding condition
(Barrier or Guardrail? Sideslope?)
• Heavy vehicle volume
• Balance of lane width (and distribution) with
companion shoulder width
• Free flow speed
• Lane capacity and lane location
6

7. Safety Effects – Lane & Shoulder Width
(3% to 11% increase in crashes)
Treatment
Traffic Volume
AADT (vehicles /
day)
Crash Type
(Severity)
CMF
Std.
Error
4 to 5 lane
conversion
79,000 to 128,000
one direction
All types (All severities) 1.11 0.05
All types (Injury and Non-injury
tow-away)
1.10 0.07
All types (Injury) 1.11 0.08
5 to 6 lane
conversion
77,000 to 126,000
one direction
All types (All severities) 1.03 0.08
All types (Injury and Non-injury
tow-away)
1.04 0.10
All types (Injury) 1.07 0.10
Base Condition: 4 or 5 lanes (12 feet wide) depending on initial roadway geometry
Crash Effects of Adding Lanes by Narrowing Existing Lanes and Shoulders
Source: Adapted from the Highway Safety Manual, Table 13-5, p. 13-10 (AASHTO, 2010)
7

8. Safety Effects – Lane Width for Texas
Urban Freeways (Increase of 7%)
Lane Width
(feet)
Number of Through Lanes (in one direction)
2 3 4 5
10 1.07 1.06 1.06 1.07
10.5 1.06 1.05 1.05 1.05
11 1.04 1.03 1.03 1.03
11.5 1.02 1.02 1.02 1.02
12 1.00 1.00 1.00 1.00
Base Condition: Lane width of 12 feet
Lane Width CMF Values for Texas Urban Freeways (Injury + Fatal Crashes)
Source: Roadway Safety Design Workbook
8

9. Safety Effects – Right Shoulder for Texas
Urban Freeways (Increase of 3% to 6%)
Shoulder
Width (feet)
Number of Through Lanes (in one direction)
2 3 4 5
≤ 6 1.06 1.04 1.03 1.03
7 1.05 1.03 1.02 1.02
8 1.03 1.02 1.01 1.01
9 1.02 1.01 1.01 1.01
10 1.00 1.00 1.00 1.00
11 0.99 0.99 0.99 0.99
12 0.97 0.98 0.99 0.99
Base Condition: Outside shoulder width of 10 feet
Outside (Right) Shoulder Width CMF Values for Texas Urban Freeways (Injury + Fatal Crashes)
Source: Roadway Safety Design Workbook
9

10. Safety Effects – Left Shoulder for Texas Urban
Freeways (Increase of 7% up to 16%)
Shoulder Width
(feet)
Number of Through Lanes (in one direction)
2 3 4 5
0 1.07 1.16 1.14 1.15
1 1.05 1.14 1.12 1.13
2 1.04 1.12 1.11 1.12
3 1.02 1.11 1.09 1.10
4 1.00 1.09 1.08 1.08
5 0.98 1.07 1.06 1.07
6 0.97 1.06 1.05 1.05
7 0.95 1.04 1.04 1.04
8 0.93 1.03 1.02 1.03
9 0.92 1.01 1.01 1.01
10 0.90 1.00 1.00 1.00
Base Condition: Inside shoulder width of 4 feet for 2 lanes in one direction, and 10 feet for ≥
3 lanes in one direction.
Inside (Left) Shoulder Width CMF Values for Texas Urban Freeways (Injury + Fatal Crashes)
Source: Roadway Safety Design Workbook
10

11. Operational Analysis
11

12. Source of Speed Data
• TxDOT sensors – per lane data
12
Sensor in Dallas

13. Number of Speed Sites
13
City
Lanes with
11-ft Width
Lanes with
12-ft Width
Total Number
of Lanes
Dallas 52 30 82
Houston 16 47 63
San
Antonio
15 43 58
Total 83 121 204

14. Locations of Speed Sites
14
San AntonioHoustonDallas

15. Speed Site Variables (slide 1 of 3)
• Number of general purpose lanes on freeway
(per direction)
• Posted speed limit (mph) – 60, 65, or 70 mph
• Lane width (ft) – 11 or 12 ft
• Left shoulder width (ft)
• Right shoulder width (ft)
• Median width (ft)
• Median type – grass or concrete barrier
15

16. Speed Site Variables (slide 2 of 3)
• To right and left of the lane:
– Lane = another lane
– Shoulder = either right or left shoulder
– Speed change lane
– Buffer + pylons (separates managed lanes from
general purpose lanes)
16

17. Speed Site Variables (slide 3 of 3)
• Ramps
– Distance to next ramp – downstream (ft)
– Distance to next ramp – upstream (ft)
– Number of ramps within 1.5 miles – downstream
– Number of ramps within 1.5 miles – upstream
• Curvature – engineering judgment (yes/no)
– Horizontal
– Vertical
17

18. Other Variables for Speed Study
• Speed – average operating speed per lane for
5-minute time period (mph)
• City – Dallas, Houston, or San Antonio
• Volume – 5 min/lane volume
• Natural light level – day or night
• Day of week – Wednesday or Saturday
18

19. Speed Data Filters
• Removed all HOV lanes
• Removed lanes with widths > 12 due to being
in a merging area
• Removed the following (served as surrogate to
identify free-flow speeds)
– Speeds < 50 mph
– 5-min volume > 250 veh
• Removed speed data that occurred within 30
minutes of sunrise or sunset
19

20. Number of Speed Measurements
11-ft Lane Width 12-ft Lane Width
Number of
Lanes
Frequency
Number
of Lanes
Frequency
2 0 2 3,005
3 108,715 3 110,575
4 123,684 4 87,119
5 108,296 5 125,903
All 340,695 All 326,602
20

21. Speed Data Analysis Technique
• Perform cross-sectional evaluation on speed,
volume, site characteristics
• Evaluate operational effects via statistical
analysis:
– Objective: model operational speed to isolate its
relationship with lane and shoulder width, after
accounting for other influential variables
21

22. Resulting Speed Model Variables
• Significant variables
– Volume
– Natural light level (day or night)
– Day of the week (Saturday or Wednesday)
– Use of the space neighboring the lane (To_Left
and To_Right)
– Median type (grass or barrier)
22

23. Final Model for a Freeway Lane
Speed = 67.80
– 0.00018V2
5-min/lane (between 0 and 250 veh)
+ 4.60(if grass median, concrete barrier otherwise)
+ 2.21(if 12 ft lane width, 11 ft otherwise)
+ 3.62(if buffer + pylons to left)
+ 2.03(if shoulder to left)
– 3.89(if speed change lane to right)
– 4.39(if shoulder to right)
– 2.00(if night, daytime otherwise)
– 1.47(if Wednesday, Saturday otherwise)
23
Identifies lane
placement –
only one of
these conditions
will apply

24. Examples of Predicted Speeds
24

25. Speed Study Key Findings
• Lane width – 12 ft lanes are about 2.2 mph
faster than 11 ft lanes (other variables held
constant)
• Shoulder width – for this dataset, shows
influence for left-most lane that are 11 ft wide
(1.1 mph / 1 ft change in shoulder width)
• Freeway speeds are lower at night (2 mph)
and higher on the weekend (1.5 mph,
Saturday compared to Wednesday)
25

26. Safety (Number of
Crashes)
26

27. Site Selection
27
• Site = one direction (similar to operations)
• Define freeway segments to coincide with
speed spot locations (when possible)
• Additional sites were also located
• No site included a ramp within its limits

28. Location of Safety Sites
28
San AntonioHoustonDallas

29. Crash Data
29
• Freeway crashes for years 2010 – 2013
• Data were disaggregated in 6-month periods
to correctly account for changes in cross
sections
• Unit of analysis is segment-years for each site:
1 segment-period:
(length of segment)x(analysis period length)

30. Summary Statistics
30
Variable Mean Min Max Total
Total Crashes 4.1 0 37 2202
KAB Crashes 0.8 0 6 407
Period Length (yr) 0.48 0.08 0.50 -
AADT (vpd) 152,163 200 281,450 -
Segment Length (ft) 1897.2 618 4510 -
Number of Lanes 3.8 2 6 -
All Lanes Width (ft) 45.2 24 73 -
Average Lane Width (ft) 11.8 10.8 12.5 -
Left Shoulder Width (ft) 9.1 1.3 22.9 -
Right Shoulder Width (ft) 10.3 2.0 14.8 -
Closest Downstream Ramp (ft) 1861.2 17 6938 -
Closest Upstream Ramp (ft) 1738.7 320 7170 -
N = 536

31. Resulting Crash Model Variables
 AADT
 Number of Lanes
 Right Shoulder
 Left Shoulder
 Distance to upstream ramps
 Distance to downstream ramps
31

32. Final Model for Total Crashes
32
NTotal
= 1.0027 × L × AADT0.539
× e −1.0243 Ramp_Up_D −1.0877 Ramp_Dn_D
× e −0.0241 NLane×Lane_WAvg
× e −0.0735 R_Shld_W −0.0646 L_Shld_W
A negative sign indicates a reduction in crashes
when these variables increase

33. Final Model for KAB Crashes
33
NKAB
= 0.0514 × L × AADT0.662
× e −1.5787 Ramp_Up_D −0.8659 Ramp_Dn_D
× e −0.0253 NLane×Lane_WAvg
× e −0.0956 R_Shld_W −0.0547 L_Shld_W

34. Safety Study Key Findings
Crashes vs. Lane Width
 5% fewer KAB crashes at 2-Lane (2-L)
freeways with 12 ft lane widths, compared to
2-L freeways with 11 ft lane widths
 12% fewer KAB crashes at 5-Lane (5-L)
freeways with 12 ft lane widths, compared to
5-L freeways with 11 ft lane widths
 Similar trends occur for Total Crashes
34

35. Safety Study Key Findings
Crashes vs. Number of Lanes
 24% fewer KAB crashes per additional lane
when the average lane is 11 ft
 26% fewer KAB crashes per additional lane
when the average lane is 12 ft
 Similar trends for Total Crashes
35

36. Safety Study Key Findings
Crashes vs. Shoulder Widths
 5% fewer KAB crashes per additional foot of
left shoulder width
 9% fewer KAB crashes per additional foot of
right shoulder width
 Similar trends for Total Crashes
36

37. Example Application of
Models
37

38. Example Problem #1
What is the expected daytime operating speed
on a weekday for a freeway lane with the
following characteristics:
• Hourly volume = 1200 vph (100 veh/5-min),
• Number of lanes = 3,
• Lane width = 12 ft,
• Median type = concrete barrier, and
• Lane position of interest = center lane.
38

39. Solution – Example #1
 Hourly volume rate of 1200 veh/hr corresponds to 100 veh/5-
min = (1200 veh/hr)/(20 5-min/hr periods)].
 Center lane in a 3-lane freeway section has a lane on both the
right and left.
 Wednesday can be considered a representative weekday.
Speed = 67.80 – 0.00018(100)2
5-min/lane + 4.60(0) + 2.21(1) +
3.62(0) + 2.03(0) – 3.89(0) – 4.39(0) – 2.00(0) – 1.47(1) =
66.7 mph
Conclusion: An operating speed of approximately 67 mph can be
expected for the freeway lane.
39

40. Example Problem #2
What is the expected total number of crashes and
number of injury crashes for a freeway section with the
following characteristics:
• Length= 0.5 miles
• Number of Lanes = 4 (average width of 12 feet each)
• AADT=150,000 vpd
• Right Shoulder Width = 10 ft
• Left Shoulder Width = 6 ft
• Distance to closest upstream ramp = 0.5 mi,
• Distance to closest downstream ramp = 1.0 mi
40

41. Solution – Example #2
41
𝑁 𝑇𝑜𝑡𝑎𝑙 = 1.0027 × 0.5 × 150,0000.539
×
𝑒 −1.0243 0.5 −1.0877 1.0 −0.0241 4×12 −0.0735 10 −0.0646 6 = 6.39
𝑁 𝐾𝐴𝐵 = 0.0514 × 0.5 × 150,0000.662 ×
𝑒 −1.5787 0.5 −0.8659 1.0 −0.0253 4×12 −0.0956 10 −0.0547 6 =1.08
Conclusion: A total number of approximately 6 to 7 crashes,
including 1 injury crash, can be expected for the freeway location.

42. Example Problem #3
42
𝐶ℎ𝑎𝑛𝑔𝑒 𝑖𝑛 𝐶𝑟𝑎𝑠ℎ𝑒𝑠
= e[ −2.53 ×10−2 × 55−48 + −9.56 ×10−2 × 8−10 + −5.47 ×10−2 × 1−6 ] = 1.33

43. Summary
• Speed predicted based on specific lane
• Adding lanes will result in:
– Increased capacity
• Narrowing lanes and shoulders will result in:
– Increased number of crashes and reduced speeds
• Methods can be used to minimize safety
impacts by optimizing lane and shoulder
width effects
43

44. Questions?
44