Traffic count guidelines for widening roads ...

[stonefort]

 ..

[Riverside Frwy]

The number of lanes needed based on traffic demand can be calculated using a relatively simple formula(don't have the book right now) that also uses modifiers for Bus, trucks, RVs, and other heavy vehicles.

You also always round up, regardless of the decimal. If you get 3.2 lanes after doing the calculation, then you build 4 lanes, not 3.

[J N Winkler]

Matson's Traffic Engineering (1955) is a bit dated but is still a clear explanation of the basics; a well-equipped university library should have a copy.  An old edition of the Highway Capacity Manual should also have useful information (I wouldn't recommend going all the way back to 1950 because that edition talks about "practical capacity" whereas current design is based on the level-of-service concept, which was introduced in 1965).

In the US nowadays, the usual method is to begin with a facility type, a design hour volume (DHV) it has to accommodate, and a design-year level of service (LOS) that is to be used for that facility.  Facility type is a high-level project development decision ("we want to build a freeway in this corridor").  DHV comes from traffic forecasts.  Design LOS is an engineering policy decision ("we will build new facilities of this type to carry the DHV at a given LOS").  Typically the design year is the year the road is designed plus a design horizon (usually 20 to 25 years), and the intent is to provide enough capacity to accommodate normal traffic increase over the design horizon.

LOS is a letter-grade measure of how close the facility is to maximum capacity.  It correlates fairly closely to hourly volumes for multilane roads, and to time spent following other vehicles for two-lane roads.  For multilane roads, LOS A is an almost empty highway (and thus represents luxurious provision); LOS E corresponds to the highest volume at which stable flow is still possible; and LOS F is stop-and-go traffic.  Standard tables are available which relate LOS to hourly volume for various facility types.  Design policy in many states calls for design-year LOS values of B and D for new freeways in rural and urban areas respectively.

Design hour volume is intended to be a peak hour volume that recurs reasonably frequently.  A traditional rule of thumb is that DHV is the thirtieth highest peak hour volume in any given year.  There are ratios which can be used to convert annual average daily traffic (AADT) values into peak hourly volumes according to the type and location of a given highway facility, but these yield ballpark estimates only--every road is different and a given road may have a peak hourly volume significantly higher or smaller than that estimated from its AADT, as a result of particular local circumstances.  (State DOT traffic flow maps typically quote AADT, never peak hourly volumes, which are typically generated after additional engineering investigation and analysis.  In this forum we sometimes convert AADTs into peak hourly volumes to get an idea of whether a proposed road widening is justified, but this is no more than armchair engineering.)

Aside from this basic procedure for selecting lane count on a new-build facility, there are some general rules of thumb.  For example, engineers often speak of a trigger volume for widening from two-lane to four-lane divided.  This volume is essentially arbitrary because it is implicitly a choice of worst acceptable level of service on a crowded two-lane highway, and also reflects the highway agency's opinion as to whether intermediate solutions such as Super Two construction, addition of passing lanes, expansion to a 2+1 cross-section, etc. are viable.  In the US the trigger volume most commonly quoted nowadays is 10,000 VPD in level terrain (dropping to 7,000 VPD in rolling country and 5,000 VPD in mountains), but 4,000 VPD was extensively quoted in the 1930's and 1940's.

A single traffic lane on a freeway has a nominal maximum capacity of 2,000 VPLPH.  A freeway carrying these volumes in all lanes is operating at a bad level of service, however (usually LOS E).  This volume is also not consistent with the maintenance of two-second headways between vehicles.  It is useful mainly as a way of looking at a hourly volume or AADT value and assessing whether it is reasonable that a particular freeway could be carrying that much traffic (for example, you wouldn't expect to see 400,000 AADT on a six-lane freeway).

None of the foregoing relates to the issue of whether the road is worthwhile to build.  To address that question there has to be a separate layer of cost-benefit analysis, and often there is an iterative process where an ambitious proposal is ratcheted down in size until it becomes "affordable."  This can mean either that first cost fits the budget, or that the annualized rate of return (i.e., benefit net of cost discounted and paid out over time) is not significantly less than that of alternate highway investments that can be made with the same money.  It is always a political decision to allow the fate of a highway project to be determined by the outcome of a cost-benefit analysis, though many agencies try to take politics out of the equation by ruling that projects can be pursued only if their annualized rates of return are in excess of the long-term interest rate (which is used as a measure of the opportunity cost of the money spent to build them).

A useful text on cost-benefit analysis, available at most university libraries, is Roland McKean's Efficiency in Government.

[Alps]

Highway Capacity Manual, 2010. Each agency varies. In New Jersey, you don't consider widening a road until you have LOS E or F (Level of Service). In the West, you may consider widening in the middle of LOS C. It's like the grades you got back in school - A is good, F is failing. If you know daily traffic, a good rule of thumb is multiply by 8% in an urban area or 12% in a rural area to get the peak hour traffic. (It's lower in urban areas because traffic stays heavier throughout the day.) 10% is, of course, a good hedge bet for a small city or suburban area. Then you compare your peak hour traffic, divided by number of lanes, to the thresholds in the manual.

Yes, it would take a long time for the lay person to pick it up, just trying to read through the Manual. But that's why they pay me. (:

[J N Winkler]

This primer on capacity (1969) is dated, but still explains the basic concepts lucidly:

http://www.dot.ca.gov/newtech/researchreports/1969-1970/69-44.pdf

[kkt]

Matson's Traffic Engineering (1955) is a bit dated but is still a clear explanation of the basics; a well-equipped university library should have a copy.  An old edition of the Highway Capacity Manual should also have useful information (I wouldn't recommend going all the way back to 1950 because that edition talks about "practical capacity" whereas current design is based on the level-of-service concept, which was introduced in 1965).

(remainder deleted)

Thanks for the comprehensive post!