Working With A Policy On Geometric Design Of Highways in the Field
Where to Get A Policy On Geometric Design Of Highways and What It Actually Covers
The main policy documents come from AASHTO in the US — specifically the Green Book, officially called A Policy on Geometric Design of Highways and Streets. It's updated every few years, currently on the 2018 edition with some addenda. You can grab it directly from the AASHTO store for around $120 to $150. If your agency has a library subscription, you can pull it for free through their portal. Canada uses its own version, the Geometric Design Guide, published by Transport Canada. Europe follows the Eurocode 7 and national annexes, which tends to produce tighter curvature standards than North American practice for the same design speed. The Green Book itself isn't free. It covers everything from design speed selection, stopping sight distance, superelevation, horizontal and vertical alignment, intersection geometry, and roadside clearances. That's the broad brush. Most engineers I know carry a printed copy or have the PDF open on a second monitor because the tables are where the actual work happens.
How to Use It, Not Just Read It
Pick a design speed first. This is the number that drives every other calculation. If you pick 65 mph, the minimum stopping sight distance is 1600 feet, the minimum radius for a given superelevation drops to specific values, and the lateral acceleration thresholds change. If you then realize your corridor can't support 65 mph without massive earthwork or right-of-way acquisition, you go back and either change the speed or accept a different horizontal alignment. Most projects I've worked on had the design speed negotiated between the client, the public, and the final alignment. The policy gives you the math, but the field reality is what matters. The typical workflow goes like this. You run a preliminary alignment inCivil 3D or a similar tool. You overlay it on LiDAR or topo survey data. You check each tangent and curve against the sight distance tables. You flag any sections where the stopping sight distance is less than required, then adjust the curve radius or elevation profile. You iterate until everything clears. In practice, this cycle takes 2 to 5 days on a mid-size road project, depending on terrain and how messy the existing conditions are.
A Real Problem I Faced With Sight Distance on a Curve
On a two-lane rural highway project in the Ozarks a few years back, we had a horizontal curve where the available stopping sight distance came up short by about 40 feet. The terrain made it impossible to flatten the vertical curve enough to meet the full 1600-foot requirement without taking out two more acres of ridge. The alternative was to lower the design speed to 55 mph, which would have required a wider right-of-way at several interchanges further down the line and cost significantly more. The workaround I used was to design a passing sight distance exception combined with a raised median at the curve, converting it effectively to a divided section at that point. The policy allows reduced sight distance if you can demonstrate through analysis that the crash risk doesn't increase materially and you provide other mitigation. We ran a conflict point analysis, showed that the curve was a free-flow segment with no major access points within half a mile, and submitted a variance to the state DOT. It passed on the third submittal after two rounds of comments. The lesson here isn't that the policy is flexible — it's that you need to know which sections actually allow exceptions before you commit to an alignment.
Get the Full Details
Common Mistakes Beginners Make
Using the wrong design speed for the context is the biggest one. People default to the posted speed limit, but the policy requires the design speed to be selected based on the physical characteristics of the road, not what drivers choose to drive. If the road geometry can safely accommodate 55 mph but the posted limit ends up being 45 due to traffic calming measures nearby, you still design to 55 unless there's a documented reason not to. Another frequent error is ignoring transition spirals. The Green Book recommends spirals for curves with radii below 3000 feet on high-speed roadways. I've seen projects skip them on curves around 1800 feet radius because the designer thought the superelevation runoff alone was sufficient. It is not. The transition zone needs a spiral to avoid jerk discomfort and to keep the drainage cross-slope from becoming inconsistent. A third mistake is treating superelevation tables as absolute minimums. They are, but only under ideal conditions. In areas with heavy truck volumes, ice-prone surfaces, or agricultural traffic, you typically want to increase the superelevation rate by 1 to 2 percent above the tabulated minimum. The policy acknowledges this in the supplementary guidance, but it's easy to miss if you're skimming the tables rather than reading the surrounding text.
What the Policy Gets Wrong
The Green Book assumes relatively uniform traffic composition. It doesn't account well for mixed freight and passenger vehicle fleets at extreme ratios, which is increasingly common on truck corridors. The friction values baked into the sight distance calculations assume dry pavement and standard tire tread. On routes where sanding and deicing chemicals create a persistent low-friction film at curve entries, the effective stopping distance can be 15 to 20 percent longer than the table value. No amount of widening fixes this — you need a combination of improved drainage, surface treatment, and potentially lower design speeds. The policy also lags behind current autonomous vehicle research. The assumption that human drivers will perceive and react within the standard time gaps is being questioned as sensor-equipped vehicles enter mixed traffic. The reaction time assumptions in the current edition haven't been revised to reflect this, though the next iteration is expected to address it. In the meantime, designing to the current standard is still the legal baseline, but it's worth noting where your project falls outside the assumed envelope.
Practical Tips That Actually Help
Keep a spreadsheet of your design speed versus minimum radius, stopped sight distance, and super elevation rates for quick reference. When you're doing a spot check during corridor screening, pulling these numbers from memory or flipping through the book costs more time than it saves. A lookup table cuts the screening phase from roughly 4 hours down to about 45 minutes on a typical 10-mile corridor study. Use a digital terrain model rather than relying on contour maps for sight distance checks. Contour-based analysis can miss subtle crown and drainage features that eat into sight triangles. A LiDAR-derived DEM with 1-meter or better resolution will catch those issues before you get to the design development phase. When you submit a design for review, include the sight distance diagrams at each critical point. Reviewers often ask for them anyway, so providing them upfront saves a round of revisions and typically cuts the review cycle by one week on average.
When the Policy Doesn't Apply
Arterial streets in dense urban cores, private roads, and parking lot circulation systems all fall outside the Green Book scope. For urban arterials, you generally follow the Unified Facilities Criteria or local street design manuals, which emphasize pedestrian crossing distances, curb radii, and intersection sight triangles over high-speed horizontal alignment. For local streets, the ITE Access Management Guidelines and municipal subdivision ordinances take precedence. The geometric design principles overlap, but the governing standards shift significantly, and applying the Green Book by default to an urban street project is a common source of overdesign and unnecessary cost. If you're working on a project that sits on the boundary between highway and local road — a collector that transitions into a rural two-lane, for example — the safest approach is to design each segment to the standard that applies to its intended operating speed and land use classification, then document the transition clearly in your design report. Ambiguity in that handoff is where most of the change orders I've seen originate.
