Systematic Evaluation of Run-Off-Road Crash Locations in Wisconsin

 by Alex Drakopoulos and  Ertan Örnek

Marquette University

Department of Civil and Environmental Engineering


 MU ITE Chapter  | MU ITE Exec.  | ITE WI  | ITE.org  | Alumni page   |ASCE MU chapter  |MU Civil Events-Pictures  |Older MU ITE Pictures  | MU ITE Open House Nov 2002


Systematic Evaluation of Run-Off-Road Crash Locations in Wisconsin

EXECUTIVE SUMMARY

The project focused on developing a methodology to identify Run-off-Road (ROR), nonintersection

crashes that occurred on the two-lane undivided portions of State Trunk

Highways (STH) in the State of Wisconsin. The objectives of this project were to

calculate crash statistics and identify locations requiring safety improvements using a

state-wide systematic methodology.

The project addressed two of the seven action plans listed in the 2000 Wisconsin

Department of Transportation (WisDOT) Strategic Highway Safety Plan, namely:

Improve data and decision support systems; and

Keep vehicles on the roadway/minimize the consequences of leaving the

Project objectives were met by developing state-wide and highway-specific statistics

(crash rates, crash densities and other statistics) for a variety of Run-off-Road crash

characteristics, and proposing strategies to evaluate the safety performance of all STH

on these statistics.

Two strategies for the systematic identification, stratification and analysis of crash

locations were proposed. The first strategy required two sequential steps:

1. Multiple rankings of all State Trunk Highways based on selected crash

characteristics are used to select a limited number of STH for examination.

2. Particular segments of the selected highways that require further examination for

potential treatment are determined with the help of the "floating highway

segment" algorithm PRÈCIS. (Details in Appendix I.)

The second strategy required processing all undivided STH through the PRÈCIS

algorithm and selecting segments with high crash rates for treatment.

In meeting project objectives major emphasis was placed on producing a user-friendly

methodology, using existing databases in an automated manner. The proposed strategies

were based on a simple tabular format, using a locational reference in wide use within

WisDOT. GIS-based maps presenting PRÈCIS results provided simple locational

references, and crash rates at any point along a STH were presented by color-coded

continuous lines parallel to the centerline with additional line charts providing more

detailed crash rates at any mile point. The three databases used in the project (crash,

Metamanager, and State Trunk Highway Log) preexisted and were consistently updated

each year. The developed methodology to produce tabular state-wide statistics is

automated to a great extent, requiring minimal labor for annual table updates. With some

additional work on automating the state-wide application of PRÈCIS tables and the GIS

maps displaying the information they contain, a set of maps could be created annually

with minimal labor expenditure.

A wide array of crash statistics were produced for Run-off-Road crashes on the 9,471

miles of undivided STH in the state of Wisconsin. A quick reference to appropriate

tables is provided in Table 4.

Between 1998 and 2000 a total of 60,345 crashes occurred on the 9,471 miles of

undivided STH. Most (34,604) took place on rural highways (8,901 miles), where most

crashes (21,947) occurred at non-intersection locations. Among those, 11,803 were Runoff-

Road crashes including 207 fatal and 4,972 injury crashes.

Relationships between two-lane rural highway Run-off-Road crash characteristics and

severity were examined in detail in Appendix H. A high percentage of fatal crashes

(62.2%) occurred during nighttime—only 44% of all ROR crashes occurred during

nighttime. Most fatal crashes (74.8%) occurred on dry pavements with a relatively small

percentage (9.4%) occurring on pavements covered with snow, slush or ice. Statistics for

all ROR crashes were 53.8% and 33.3% respectively. A large percentage of fatal crashes

involved overturning vehicles (40.5%)-the overall percentage was 26.4% for ROR

crashes. A disproportionate number of crashes involving motorcyclists were fatal (6.5%

of all fatal crashes) given that motorcyclists were involved in 1.9% of all crashes.

The state-wide crash rate for undivided STH was 144 crashes/ 100MVM; the rate was 99

crashes/100MVM for rural highways and 368 crashes/100MVM for urban highways.

Two-lane rural highways had a rate of 96 crashes/100MVM; the non-intersection rate

was 62 crashes/100MVM and the ROR rate was 34 crashes/100MVM.

The state-wide crash density for undivided STH was 2 crashes/mile/year (crashes/mi/yr);

urban density was 15 crashes/mi/yr; rural was 1.3 crashes/mi/yr. Two-lane rural was 1.2

crashes/mi/yr. On the same highways, the non-intersection crash density was 0.8

crashes/mi/yr and ROR density was 0.4 crashes/mi/yr.

It is recommended to include additional years of crash experience in ROR crash statistics

calculations. A systematic review of results produced in the course of the present effort

would be desirable; highway segments selected for safety upgrades should be

documented and available through a WisDOT safety clearinghouse. It is recommended

to generate a set of maps, one for each STH, to display crash rates generated through the

PRÈCIS algorithm, and make them available through the safety clearinghouse.