Trees in Hazardous Locations

The Type of Problem Being Addressed

Collisions between vehicles and trees are a major type of traffic fatality. According to Fatal Accident Reporting System (FARS) data for 1999 (Exhibit III-1), 10,967 fatal crashes involved a fixed object. Trees were the objects most often struck, involving 3,010 fatal crashes, or about 8 percent of all fatal crashes.

Exhibit III-2 shows the distribution of fatal crashes by roadway functional class. Fatal tree crashes were most prevalent on local rural roads, followed by major rural collectors. Of all fatal tree crashes, 90 percent occurred on two-lane roads and 5 percent on four-lane roads (see Exhibit III-3). While reducing tree-related fatal crashes will require addressing all classes of streets and highways in urban and rural areas, rural two-lane roads will receive much of the focus in the development of any program to reduce tree-related highway fatalities.

Specific Attributes of the Problem

Tree crashes are strongly correlated with traffic volume, roadway geometry, and overall roadside condition. Zegeer et al. (1990) included a detailed analysis of crashes involving specific types of roadside features. For average daily traffic (ADT) categories of 1,000 vehicles per day (vpd) and below, 22 to 24 percent of fixed-object crashes involve striking trees (Exhibit III-4). This compares to 16 percent involving tree crashes for roads with ADTs of 1,000 to 4,000 vpd, and 11 percent for ADT above 7,500 vpd.

Conversely, the percent of crashes involving utility poles, signs, and guardrail increases as ADT increases, which reflects increased numbers of such roadside features on higher-volume, generally higher-class roads. More insights are gained by examining the relationship between tree and other fixed-object crashes and traffic volume, as well as by looking at the frequency per-mile of such crashes. Given that total crashes increase as ADT increases, the frequency per-mile of crashes involving trees and other fixed objects increases as ADT increases. Exhibits III-5a through 5c demonstrate the relationship among tree crashes (per-mile per-year); ADT; distance of trees from the road; and "tree coverage" (i.e., percent of the roadside with one or more trees). The study was based on data for a 5,000-mi sample of mostly rural two-lane roads (Zegeer et al., 1987).

Exhibit III-5b corresponds to roadway segments having tree coverage of 15 to 30 percent and average tree distances of 0 to 30 ft from the roadway under various ADT categories. Here, sections having 15 to 30 percent tree coverage between 0 and 12 ft and having ADTs above 4,000 were found to average 0.25 tree crash per mile per year. Actual values for a given section will vary, depending upon roadway geometry (e.g., roadway width, roadway alignment); traffic factors (e.g., percent trucks); and driver factors (e.g., percent of drinking drivers, young drivers).

Perhaps the most important point illustrated by Exhibits III-5a through 5c is the relative infrequency of tree crashes on two-lane highways, even where traffic volumes are higher, tree coverage is significant, and the trees are close to the road. In such cases, one might expect an average of one tree-related crash per mile every 3 to 5 years. A "high-crash" segment may be one in which no more than two or three tree-related crashes occur over a 5-year period. Indeed, one should not expect to find many locations where a specific tree represented a repeated, significant hazard over a 3- to 5-year period.

Other characteristics of fatal tree crashes (from FARS 1998 and 1999 data) include:

• About 56 percent of fatal tree crashes occurred under nighttime conditions. This is particularly significant given that much more traffic occurs in daylight hours versus night hours.
• Nearly half of all fatal tree crashes occurred on curved roads. Most road mileage is tangent; hence, this finding is particularly significant.
• Of the 1,562 fatal tree crashes in 1998 where alcohol use was suspected, 45 percent of crashes were cited as alcohol involved.

The legal issues resulting from tree crashes are complex and will not be covered in this guide. However, they should be considered when implementing the strategies in this guide. See Appendix 4 for further discussion and associated references. Furthermore, care must be taken when doing detailed site analyses to identify the role of the tree in crashes, to determine if the tree is really in a hazardous location, or if the hazard lies with some other design or environmental feature (Appendix 15).