Trees in Hazardous Locations: Appendix 12

Identifying Risks at Hazardous Locations Involving Trees

Once the decision is made to engage in a program to treat locations with tree-crash histories or potential, it is necessary to identify the locations and prioritize them for investigation on the basis of relative risk. This step varies from jurisdiction to jurisdiction, due partly to the fact that different data elements are available in the crash database, road inventory, and other data sources available for setting priorities. Some key considerations are discussed below.

Responsive Applications

Responsive applications require identifying locations on the roadway system where there are concentrations of tree crashes. Typically, errant vehicles strike different trees adjacent to the roadway, rather than a single tree. These areas may be where the clear zone is not adequate for the curvature of the road, the steepness of the sideslopes, or possibly where trees have grown large enough to become a roadside hazard.

One state DOT for example, investigates all fatal crashes on the state-maintained highways. Tree crashes comprise approximately 9 percent of the fatal crashes in this state. Therefore, when engineers investigate a fatal tree crash, they can consider developing projects that implement one or more of the strategies described in this guide.

It is also possible to identify trees in potentially hazardous locations through use of roadway logs, as well as analyzing public complaints, reported problems by police officers and the news media, and other systematic programs. For example, members of the public may call occasionally to complain about a tree being too close to the roadway, or blocking their view in a curve, or at an intersection.

EXHIBIT 12-1 A Tree in a Hazardous Location with the Scars to Prove It

Proactive Applications

Exhibit 12-1 shows a scarred mature tree very close to the traveled way of a rural highway. The shoulders are narrow and slightly rutted near the tree. The scars are probably from cars that had previously run into the tree. If the road is low volume, the frequency of crashes may not be sufficient to result in it being identified as a high-crash site in any systematic study. It is for cases such as these that agencies should consider developing proactive programs to address tree-crash problems.

While responsive application depends heavily upon crash data (i.e., identifying sites with clusters of tree crashes), the proactive approach depends upon detailed analyses of roadway data, the conduct of safety audits, or other methods to identify sites with a high potential for tree crashes. An information approach is relatively low cost. However, few agencies collect and maintain the data necessary to conduct a program of this nature. Data that may be used include roadway curvature along specific segments, horizontal clearance to roadside objects, sideslope information, and the presence of guardrail. Agencies may need to conduct special inventories.

Highway safety audits provide a method to identify locations with a high likelihood of tree crashes. Systemwide safety audits may not be economically feasible for agencies with large networks, so agencies may need to prioritize their roadways for using this procedure. It is also necessary to define what measures determine if a site has a high likelihood of tree crashes. Previous tree crash studies conducted by agencies allowed the agency to define the physical and operational characteristics that correspond to higher frequencies of tree crashes. This would allow the agency to extrapolate the results to other sites, where similar conditions exist. Planting and/or mowing guidelines may also be used to identify locations not in compliance, and hence potentially hazardous.

Investigate

Site investigations to determine potentially effective strategies contain two parts: risk assessment and effectiveness assessment. Risk assessment helps determine the relative risk of future tree crashes. Effectiveness assessment determines the local effects of removing trees, or the application of the other strategies. It may help for an agency to develop a structured approach to assess the risk and impacts.

Assess Risk

Assessing the risk of running off the road and striking roadside trees is a critical step in both the proactive and responsive approaches. In a responsive approach, the risk is assessed by reviewing the frequency of crashes involving trees, and determining the likelihood of the pattern continuing or worsening. It is necessary to review the frequency or rate of ROR crashes as well as hit-fixed-object crashes. While a responsive approach involves the use of crash histories, the proactive approach requires identifying locations that have a high likelihood of a vehicle striking a tree when it runs off the road. This approach usually relies upon identifying locations with similar physical and operational characteristics as the locations identified in the responsive approach. When assessing the risk of a site, there are several critical factors to consider. The following paragraphs discuss some of these factors.

Crash History
When assessing the risk of future events, a detailed review of the crash history at the site is important. Usually the location is identified as a candidate high-hazard location in a macroscopic review of crashes, by location. A more in-depth review of the crash history needs to consider the factors, other than location, associated with past tree crashes. The type and severity of each crash should be reviewed, along with contributing factors. For example, assume that a roadway segment had a high frequency of tree crashes due to vehicles leaving the roadway while exceeding the speed limit on a rural road. In that case, tree-oriented countermeasures may not mitigate the risks, since driver error and poor judgment are likely to remain as problems in the foreseeable future; no practical improvement to the clear zone may be sufficient to reduce the severity of these crashes. In such a case, countermeasures aimed at aggressive driving may be appropriate. On the other hand, an investigation should consider driver error and poor judgment in the risk assessment, but should not consider them as reasons to overlook a location.

EXHIBIT 12-2 Curve Direction and Crash Frequency8

Roadway Alignment
The horizontal alignment of a roadway section may have an influence on the level of risk for future ROR and tree crashes. Zegeer et al. developed a crash-prediction model that used length of curve, degree of curve, and traffic volume as key elements to estimate the number of expected crashes along a curve.¹ The AASHTO Roadside Design Guide² considers the impacts that the curvature of the roadway has on ROR crashes. It provides adjustments, to increase the clear-zone width, based upon design speed and the radius of the curve. Table 3.2, in the Roadside Design Guide, provides curve correction factors to increase the clear zone width by a factor of 1.1 to 1.5 depending on the curve radius and the design speed. O’Day shows that some parts of the roadway curve are more prone to ROR crashes than others (See Exhibit 12-2). Specifically, the outside shoulder and roadside had a higher involvement in total crashes and fatal crashes³. Zeigler et al. found that certain portions of a curve are more prone to tree crashes. In their report, Guide to Management of Roadside Trees, they reported that in Michigan, 77 percent of tree-related crashes on curves occurred on the outside of the curve and that most tree crashes involve right departures in left curves⁴.

EXHIBIT 12-3 Measuring Tree Diameter at (A) DBH height and (B) the recommended height (Divide the measured circumference by 3.14 to get diameter) *Source: Stott, C.B. "Techniques for Using Unpainted Steel Diameter Tapes." USDA Forestry Service. 1953.

Tree Size
The forestry industry uses tree trunk diameter as the most important measurement of a standing tree. ⁵ This attribute would also seem to be most important for representing potential crash severity. The forestry-industry standard-diameter is measured on the main stem at 4.5 feet above the ground on the uphill side of the tree (i.e., the diameter at breast height, or DBH). A measurement at this height is too high for traffic safety purposes, because it is rare for a vehicle to strike a tree that high above the ground. In addition, if a tree forks below the breast height, the standard treats each trunk as a separate tree.⁶

The diameter of the tree at bumper height is a better measurement for traffic safety purposes. However, the height of bumpers varies with the many variations of vehicle styles on the road, but we do know that it is less than the forestry standard height. With the wide variation of vehicle design, it is necessary to standardize at a reasonable height. A height of 2 feet above the ground is suggested as the representative "bumper" height. This provides a reasonable height above the ground to easily measure, and is in the bumper strike zone. Exhibit 12-3 shows a technique to measure the tree diameter at the recommended height and the DBH height.

Traditionally, engineers consider a tree with a diameter greater than the wood post used for sign supports as a hazard to vehicles striking them.⁷ This typical post size is 4 inches by 4 inches with a nominal width of 3 ^5/₈ inches. To simplify the measurements, some engineers have standardized the measurement to 4 inches. Therefore, trees with diameters of 4 inches measured at 2 feet above the ground are frequently considered hazardous to errant vehicles. It is easier to measure the circumference of a tree with a tape measure, rather than directly measuring the diameter. Exhibit 12-4 provides the recommended measurements to determine if a tree is hazardous to errant vehicles.

Sideslope
Sideslopes play an important role in allowing an errant vehicle to recover. Sideslopes steeper than 4:1 are considered non-recoverable and are not included in clear-zone calculations. On sideslopes of 4:1 or steeper, errant vehicles generally travel to the bottom of the slope or beyond before coming to a stop. If large trees, other fixed objects, or precipitous conditions are located on the slope, or too close to the toe of the slope, then the non-recoverable slope plays a role in bringing the errant vehicles to the area where the hazard is located. At times, the recommended clear zone extends into private property, thus reducing the DOT control over the safety of the situation. This condition compounds tree-removal issues, as well as legal responsibility. In some jurisdictions, the engineer’s responsibility may not end at the right-of-way line.⁸ While tree removal is an option, engineers should consider alternative strategies discussed in this guidebook.

The length of the sideslope is also a consideration when evaluating the alternatives. Flattening long steep sideslopes, common in mountains and foothills, is usually cost prohibitive. Flattening short steep slopes may have higher benefit-cost ratios than placing guardrail, considering the higher maintenance and crash cost associated with guardrail.

Traffic Volume
Traffic volume is a key component for determining the risk of tree crashes. Simply put, fewer vehicles provide fewer opportunities for tree crashes. While lower traffic volumes typically mean lower crash frequencies, lower volumes also tend to have higher crash rates. Zeeger found that roads with lower vehicle volume had higher tree-crash rates than roads with volumes higher than 4,000 vehicles per day.⁹ He suggested that it might be due to increased driver attentiveness, or physical conditions that limit driving speeds. Other possible explanations include better design and maintenance standards, for both the roadway and the roadside, often associated with roads having relatively high traffic volumes. The calculation of a "crash rate" (in terms of vehicle miles traveled) may not be the best measure of safety for tree crashes. Other measures such as tree crashes per mile, proportion of tree crashes, or a simple frequency of tree crashes may be better.

The Roadside Design Guide considers volume when determining recommended clear zones. Figure 3.2 in the Guide defines four volume categories with categories of average daily traffic: less than 750, 750-1,500, 1,500-6,000, and greater than 6,000 vehicles per day.¹⁰ These adjustments are based primarily upon economics, because it is not considered practical to apply the same standards in all conditions. While economics play an important role when defining systemwide policies and standards, specific locations with higher tree-crash frequencies, and continued risk of future events, require larger clear zones. For instance, if a very low-volume section continues to experience a high frequency of tree crashes, the documented hazard should supersede guidelines that would suggest that no further improvement in the clear zone was required.

Distance of Tree from Road
The answer to the question "How far away from the road is enough?" is elusive and varies depending on the situation. Zeigler found that 85 percent of the trees involved in tree crashes were within 30 feet of the traveled way. The number of cases in Michigan peaked at around 10 feet. However, the distances were not stratified by speed or variables, and the cause for peaking at a distance of 10 feet¹¹ may be a function of available cross-section rather than other factors. The width of the clear zone is often at the core of the disagreements between engineers and conservation organizations. In assessing the risk of future crashes, the distance of the tree from the road is a parameter to consider, rather than a goal to achieve.

EXHIBIT 12-5 Combination of Roadway and Roadside Characteristics Can Increase the Recommended Clear Zone This is a site of a fatal tree crash where the combination of the curve and side slope reduced the clear zone below recommended distance.

Other Considerations
These critical factors are just a few of the physical and operational characteristics that an investigation should consider when assessing the risk of a tree crash. Other items such as lane width and delineation, shoulder condition, truck volume, level of enforcement, driver expectancy, and actual vehicle speed (not speed limit) could influence the risk of tree crashes. The investigating engineer needs to determine if these, or other factors, as well as their interaction with each other, affect the risk of future events at the location, or along the segment.

The interaction between many variables often makes it necessary to apply good engineering judgment. Data may suggest that a policy that establishes a clear zone at 30 feet will suffice for 85 percent of potential tree crashes. However, this may not be a fiscally feasible approach. It may also not be environmentally responsive; and it may conflict with the values of many communities. Exhibit 12-5 shows a location of a fatal tree crash where the combination of the curvature of the road, narrow shoulders, and steep sideslopes (steeper than 3:1), may have contributed to a the vehicle rolling over and striking two trees more than 20 feet from the edge line. Considering the speed, the curvature and the sideslope, the 20 feet of available space is less than the recommended clear zone.

Legal considerations of roadside hazards are an important factor for agencies. Some agencies have lost liability cases concerning roadside tree crashes, although the tree is located on private property. Even in states with contributory negligence laws, if trees are intentionally planted near the highway, they can be described as absolute nuisances and the contributory negligence defense may not apply.¹²

¹ Zegeer, C., et al. "Cost-Effective Geometric Improvements for Safety Upgrading of Horizontal Curves." Federal Highway Administration, Washington D.C. October 1991.

² American Association of State Highway and Transportation Officials. Roadside Design Guide. Washington, D.C. January 1996.

³ O'Day, J. Identification of Sites with a High Risk of Run-Off-Road Accidents UM-HSRI-79-39. University of Michigan, Highway Safety Research Institute. Ann Arbor, Michigan. 1979.

⁴ Zeigler, A. J., Guide to Management of Roadside Trees, Federal Highway Administration. Washington, D.C. December 1986.

⁵ "Estimating Tree Diameter." University of Minnesota Extension Service Website. February 21, 2001.

⁶ "Measuring Standing Trees Determining Diameter, Merchantable Height, and Volume." Ohio State University Extension Website. February 21, 2001.

⁷ Federal Highway Administration. Vegetation Control for Safety - A Guide for Street and Highway Maintenance Personnel. FHWA 90-003 Washington, D.C. 1990.

⁸ Fitzpatrick, J., Sohn, M., Silfen, T., and Wood, R. The Law and Roadside Hazards. Michie Company. For the Insurance Institute for Highway Safety. Charlottesville, Virginia. 1974.

⁹ Zegeer, C., Stewart, R., Reinfurt, D., Council, F., Neuman, T., Hamilton, E., Miller, T., and Hunter, W. "Cost-Effective Geometric Improvements for Safety Upgrading of Horizontal Curves." Federal Highway Administration, Washington D.C. October 1991.

¹⁰ "Special Study — Motor Vehicle Collision with Trees Along Highway, Roads and Streets: An Assessment." National Transportation Safety Board. Washington D.C. May 1981.

¹¹ Ziegler, op cit

¹² Fitzpatrick, op cit