1. Systematic Identification of High Crash Locations
TRB 2002
Reg Souleyrette, Ali Kamyab, Zach Hans, Aemal Khattak

2. Problems … no consideration of type, under using data we have
HES
Hazard
Elimination
Safety
Program
Rate rank
Top 200
Overall
rank
Engineering
study (top
dozen or so)
Frequency rate
$loss rank
1999
1998
1996
1997
Problems … no consideration of type, under using data we have

3. 1996
1997
1998
1999
Opportunity …

4. Methodology Identify Problem Types
Collect/develop Databases (e.g., curves)
Rank Locations
Share Results with Field Personnel*
Validation
Education
Adjustment
Use for low-cost mitigation
Follow up studies of causal factors/design implications, better ranking
“… low cost opportunities can be better identified by familiarity with a site than by automated network screening.” Hauer, 2002

5. Potential Study Topics
Safety impact of…
Elderly drivers
Horizontal curve characteristics (e.g., degree, radius)
Speed limits of 50 mph or more on expressways
Traffic volume and traffic mixture
Speed limit
Shoulder surface (e.g. paved, unpaved)
Number of access points per mile
Pavement markings
Signalized turning bays
Crossing the centerline/head-on
Turn lanes in creating traffic turbulence and weaving
High crash locations…
During Wet weather conditions
For Run-off-the-road crashes
For Fixed-object crashes
At Urban 4-lane undivided roadways
At Signalized intersections
At Stop-signed intersections
Rural expressway intersections
Choose those where
Promising treatments are possible
Safety community perceives a problem

6. Expressway Intersections
Study Topics
Head-on
Horizontal
Curves
Urban, Four-lane
Undivided Corridors
Four-lane, Rural
Expressway Intersections
Fixed
Objects

7. High Crash Curve Methodology
Identify
GIS segments
proximate to
“curve”
crashes
[ALAS]
Define Curves
(bearing/manual)
[DGPS CL]
[GIMS Cartography]
[Aerial Photos]
Join
Crashes/
Curves A
Field
Personnel
Change Design
Derive Curve
Radius and
Length E
Significance
Test/Causal Factors
Rank Locations
(Freq, Sev, Rate)
Implement
Mitigation

8. Curves: Data Integration
DGPS Driven Centerline
“On Curve” Crashes
Aerial Photography
Cartography

9. Curves: Developing the Database
Horizontal Curves
1. Manual Definition
2. Change in bearing
between road segments

10. Curves: Calculating Curve Parameters
Derived from new horizontal curves GIS

11. High Crash Curves: Validation
Present Lists and Maps
to Field Engineers
Sometimes Educate
Misconceptions
Omissions
Sometimes Adjust
Curve definition
Human error

12. High Crash Curves: Results
Statewide average = 1.4/mvm, f=1.9, $55K
Top 30 average = 11.5/mvm, f=9.5, $490K
Worst = 27.4/mvm, f=14, $960K
5% of crashes occur at top 30 locations (1% of curves)
11% of fatalities occur at top 30 locations
Curve length and degree of curvature are significant causal variables.

13. High Crash Curves: Mitigation
BEFORE
AFTER

14. Type
Curve
Rural
Intersections
Fixed
Object
Head-On
4 ln. Urban
Undivided N
3024
326
43,000
3321 40
Statewide Rate
1.4
0.30
4.8
0.5
3.4
Top 30 Rate
11.5
0.76
25.5
0.9
3.6
HCL Rate
27.4
1.32
35.5
1.2
5.5
Statewide Frequency
1.9
3.5
5.0
3.0
206
Top 30 Frequency
9.5
13.0
15.0
4.0
231
HCL Frequency
14.0
27.0
23.0
408
Statewide Loss
$55K
$140K
$130K
$810K
$2,100K
Top 30 Loss
$491K
$990K
$770K
$1,500K
$2,200K
HCL Loss
$960K
$4,100K
$1,300K
$4,000K
% of type
1.0
9.2
0.07 75
% of crashes 35
0.60
3.1 84
% of fatalities
11.0 81
1.60
6.4
58*

15. Fixed-Objects Struck: Important Factors
Interstate: terrain, pavement type, barriers
US Hwy: barriers and surface width
Other Primary: terrain, pavement
Farm: shoulder & pavement type
Local: terrain, pavement, #lanes, speed limit
… and, as expected, functional class matters

16. Fixed-Object Struck: Potential Problem Locations
US Highways, No Median Barrier, Narrower Surface Widths

17. Closure Curve database developed Systematic Approach that recognizes:
Problem type
Mitigation potential
Public/professional perception/input
Treat locations that never would have made it to the SICL

18. Thank you Sponsor: Iowa Highway Research Board
Report: