Run-Off-Road Collisions

Type of Problem Being Addressed

General Description of the Problem

The 1999 statistics from the Fatality Analysis Reporting System (FARS) show that nearly 39 percent of the 37,043 fatal crashes were single-vehicle ROR crashes on various road types (see Exhibit III-1).


EXHIBIT III-1
Single-Vehicle ROR Crashes as a Percentage of All Fatal Crashes
Single-Vehicle ROR Crashes as a Percentage of All Fatal Crashes

For two-lane, undivided, noninterchange, nonjunction roadways exclusively, there were 8,901 (24 percent) single-vehicle ROR crashes. Exhibit III-2 shows how single-vehicle ROR crashes on two-lane roads are distributed by roadway functional classification. There are more than twice as many ROR fatal crashes on rural roads than on urban roads, partly due to the higher speeds on rural roads and to the greater mileage.

EXHIBIT III-2
Distribution of Single-Vehicle ROR Fatalities on Two-Lane, Undivided, Noninterchange, Nonjunction Roads by Highway Type
(Source: 1999 FARS Data)
Distribution of Single-Vehicle ROR Fatalities on Two-Lane, Undivided, Noninterchange, Nonjunction Roads by Highway Type (Source: 1999 FARS Data)

Exhibits III-3 and III-4 show the distribution of ROR crashes by first harmful event and most harmful event for the same accident and roadway type, the latter being of higher severity (i.e., death) and the former being the first event or object hit, which may or may not result in injury or fatality. Attention should be focused on the first harmful event for strategies that deal with eliminating or protecting drivers from various roadside objects and to the most harmful event for strategies that minimize the severity of crashes when collisions with such objects occur. As noted, the objects that are hit most often are trees.

EXHIBIT III-3
Distribution of Single-Vehicle ROR Fatalities for Two-Lane, Undivided, Noninterchange, Nonjunction Roads by First Harmful Event
(Source: 1999 FARS Data)
Distribution of Single-Vehicle ROR Fatalities for Two-Lane, Undivided, Noninterchange, Nonjunction Roads by First Harmful Event (Source: 1999 FARS Data)

EXHIBIT III-4
Distribution of Single-Vehicle ROR Fatalities for Two-Lane, Undivided, Noninterchange, Nonjunction Roads by Most Harmful Event
(Source: 1999 FARS Data)
Distribution of Single-Vehicle ROR Fatalities for Two-Lane, Undivided, Noninterchange, Nonjunction Roads by Most Harmful Event (Source: 1999 FARS Data)

* Other includes events that each represent less than 0.5 percent of the total first harmful events: bridge parapet end, immersion, shrubbery, longitudinal barriers (concrete or other), pedal cycle, other noncollision, fire hydrant, snow bank, fell/jumped, other type of nonmotorist, vehicle occupant struck or run over by own vehicle, impact attenuator/crash cushion, railroad train, or gas inhalation.

Specific Attributes of the Problem

While vehicles are more likely to leave the roadway along curves, most ROR fatalities on all roads and on two-lane rural roads are on tangent sections, as shown in Exhibit III-5. For all roads, 42 percent of the 1999 ROR fatal crashes were on curves and 58 percent on tangents. For two-lane rural roads, the percentage of ROR fatal crashes on curves increased to 50 percent. The fact that more crashes occur on tangents for all roads most likely reflects the fact that most road sections are tangent. However, it is clear that both tangents and curves have significant problems and warrant treatment. As seen below, strategies are suggested for both curve and tangent sections.

As would be expected, roadside features cause the most damage in a ROR crash.FARS data for all roadway classes indicate that the most harmful event is most likely to be an overturn (42.1 percent of 1999 ROR single-vehicle fatalities), an impact with a tree (25.4 percent), an impact with a utility pole (7.2 percent), or an impact with a ditch or embankment (4.9 percent). Most other roadside objects (e.g., culverts, posts, or guardrails) are the most harmful event in 2 percent or less of the fatalities. For two-lane rural roads, the percentages for most harmful event are similar—an overturn (44.5 percent of 1999 ROR single-vehicle fatalities), an impact with a tree (28.7 percent), an impact with a utility pole (8.0 percent), or an impact with a ditch or embankment (5.0 percent). As all of these features are either necessary elements of the roadway, commonly found along the roadside, or both, strategies are needed to protect the vehicle and its occupants when it has failed to remain on the roadway.

EXHIBIT III-5
Distribution of Single-Vehicle ROR Crashes between Tangent and Curved Sections
Distribution of Single-Vehicle ROR Crashes between Tangent and Curved Sections