Dexter Williams, P. E., DRW Consultants, LLC
Access Management Regulations
The Access Management Regulations for VDOT maintained roads are a marked departure from past practice, and I believe the AMR in general represent an improvement over past practice. However, a number of theoretical principles have been established as minimum criteria. While many of these theoretical principles have merit, I believe there is limited evidence to justify some of the minimum criteria specified in AMR.
There are three geometric criteria of the Road Design Manual Appendices F (Principal Arterials) and G (Minor Arterials, Collectors & Local Streets) that I believe have no foundation in practice. Following are summary comments about these three elements (more complete technical explanations follow thereafter):
Driveway Flares On Curb And Gutter Streets. AMR requires flares on the curb returns of driveways on curb and gutter streets. Flares are 48 by 12 foot wedges of pavement on both sides of the driveway. I doubt that even 1% of existing driveways on curb and gutter arterials and collectors have flares. I have found no demonstrative evidence of benefits of curb and gutter streets with driveway flares vs. without. In 34 years of looking at VDOT road plans, I do not recall ever seeing a VDOT curb and gutter road construction project with driveway flares. Given no demonstrated advantages of flares on curb and gutter streets, I find it remarkable flares are required on curb and gutter streets given that the inherent disadvantages as follows:
Flares add 1152 to 1776 square feet of impervious area for every driveway, add right of way for VDOT maintenance, and create a zigzag pattern for sidewalks and street curb lines.
Because driveway flares are inherently intrusive, the addition of flares on driveways reconstructed by VDOT reconstruction projects will increase some individual property damages from $100,000’s to over a $1,000,000.
Some property owners looking to upgrade their properties cannot provide flares without destroying parts of their properties.
Rigid Signal Spacing: The Half-Mile Standard. There is a theory that half mile signal spacing on four lane arterials can provide the capacity of six lanes. The most often and primarily cited source for this claim is Colorado Access Control Demonstration Project, 1985. The primary subject of this document is Arapahoe Road in Arapahoe County which was completed as a four lane highway in 1984 with planned half mile signal spacing (the 1985 report shows mostly vacant land around the new road). Since then, the road has been widened to six lanes, carries 56,500 to 67,600 vehicles per day, and now experiences LOS F at many intersections. This isn’t much different from six lane Midlothian Turnpike (52,000 to 63,000 vehicles per day) which doesn’t have nearly half mile signal spacing. When Arapahoe Road was built, modern signal coordination equipment such as that operating on Midlothian Turnpike didn’t exist. Arapahoe Road is now planned to have many of the traffic signals replaced with grade-separated interchange designs. If the Colorado Access Control Demonstration Project proved anything, it proved that rigid signal spacing criteria (e.g., half-mile) are not a solution to the limited traffic capacity of at-grade, traffic-signal controlled arterials. Grade-separated interchanges have obvious capacity advantages with prices in the tens of millions range per location.
Half Mile Signal Spacing From Interchanges. AMR requires half-mile signal spacing on cross streets from the ramp junctions with limited access interchanges. The VDOT web site references the “Virginia Tech Access Spacing Study: 2007”, which is a study of crash rates on limited access interchange crossroads based on the spacing of the first access point from the interchange. The study doesn’t show any decrease in crash rates beyond 990 foot access spacing, either signalized or unsignalized. In other words, the Virginia Tech study doesn’t show any safety improvements for half mile spacing versus 1,000 foot spacing from limited access interchange ramps.
Flares are specified to be 12 feet wide and 48 feet long, with the consequent relocation of the driveway radius 12 feet from the roadway pavement.
The primary source for the imposition of flares on curb and gutter streets is page 398, Geometric Design of Highways and Streets, 2004, published by AASHTO (American Association of State Highway and Transportation Officials): “Flared driveways are preferred because they are distinct from intersections, can properly handle turning movements, and can minimize problems for persons with disabilities”. This is a subjective claim without documentation. There is no reference or evidence provided to document any advantage because a flared driveway design is distinct from intersection design. Driveways without flares can and routinely do properly handle turning movements: adequate driveway radii and turn lanes are a fundamental part of driveway design. The AASHTO reference again provides no supporting evidence for this claim of design advantage with flares because they are distinct.
The AASHTO Geometric Design of Highways and Streets, 2004 does reference another AASHTO publication: Guide for the Planning, Design, and Operation of Pedestrian Facilities. That publication includes recommendations on driveway design for pedestrian facilities, but it does not include any reference to driveway flares in the section on driveway design. In addition, Urban Street Geometric Design Handbook, 2008, includes no reference to driveway flares in the recommendations on driveway design.
If VDOT institutes flare design on condemnation projects, the relative impact on damages can be significant. Recent experience on Rt. 17 includes two properties that were settled in the $300,000 to $500,000 range with barely acceptable functional access from the driveway to site parking. With flare designs, damages would be increased to the point of total take with costs in the millions of dollars.
The inapplicability of flare driveway design to urban, curb and gutter streets should not be construed to diminish the advantage of flare design on rural, ditch and shoulder roads. This is particularly true for rural roads with insignificant shoulders, where the flare design provides a comfortable degree of separation between driveway turning vehicles and substandard shoulders and ditches and thus prevents deterioration of minimal condition rural roads.
RIGID SIGNAL SPACING – THE HALF MILE EXPERIENCE
The Colorado Access Control Demonstration Project, 1985, is the primary citation to support the claim that a four lane road with half mile signal spacing can produce the same capacity as a six lane road. The report includes extensive photography that shows a largely vacant environment adjacent to Arapahoe Road just after the road was built.
At the time Arapahoe Road was built and the analysis was done in 1985, Arapahoe Road carried 20,000 to 45,000 vehicles per day (vpd). This is relatively moderate compared to the 56,500 to 67,600 vpd in the 2006 study that recommended grade separation of signalized intersections on Arapahoe Road. In other words, the 1985 study was done before development and traffic had accrued on Arapahoe Road to the 60,000 vpd range associated with six lane highways. The 1985 conclusion that a four lane road with half-mile signal spacing could provide the capacity of a six lane road has proven premature: the road with half mile signal spacing has been widened to six lanes with resulting LOS F at many intersections.
One of the most interesting things about the Colorado Access Control Demonstration Project is its widespread citation without any current publication of the document. Various requests to VDOT and the Highway Research Council indicated that VDOT doesn’t have a copy. A copy was secured by local government inter-library transfer from the Colorado State Publications Library.
VDOT’s direct source for this claim of six lane capacity on a four lane road with half mile signal spacing is Access Management Manual, 2003, published by the National Academy of Sciences, Transportation Research Board. The only citations in this document of actual practice to support half mile signal spacing are Colorado Access Control Demonstration Project and Access Management Study for Wisconsin State Trunk Highway 50, Suburban Kenosha, Wisconsin, Aug. 1998. The diagram above is a simplification of diagrams originally included in Colorado Access Control Demonstration Project.
A check of the status of State Trunk Highway 50 showed that Wisconsin DOT has an ongoing program of capacity enhancement on Highway 50 that includes access management design and construction. However, Highway 50 doesn’t have fixed signal spacing. Signal spacing is as follows (in feet): 650, 2700, 1550, 3400, 1980, 1200, and 1950. This is generally greater spacing between signals than found on Midlothian Turnpike, but it not 2640 feet between signals.
In summary, half-mile signal spacing as a cure for over-capacity roads is a theory that doesn’t have any documented evidence to support it. 25 years of experience on Arapahoe Road has proven that half-mile signal spacing doesn’t provide six lane capacity on a four lane road.
This is not to say that practical highway design to reduce turning movement conflicts and enhance capacity is without merit. In general, maximizing distance between signals enhances capacity. However, a blind adherence to a simplistic spacing theory disproven by experience is an exercise is dogma, not engineering achievement.
HALF MILE SIGNAL SPACING FROM INTERCHANGES
The calculations for signal spacing from the interchange ramp were originally done in 1982. While the design principles for these calculations may lead to the conclusion that half-mile signal spacing is necessary for ramp junction to signal spacing, the “Virginia Tech Access Spacing Study: 2007” does not.
This is not to say that signalized access can’t be spaced one half mile from an interchange ramp, but there is no documented safety advantage in the “Virginia Tech Access Spacing Study: 2007”. The study clearly shows that there is no decrease in accident rates beyond 990 to 1320 feet.