This graph is the simulated average arterial through vehicle speed for 4 lanes (top line) and 3 lanes (bottom line). The LOS ranges indicated are for the arterial LOS 1994 HCM. This graph is for an ADT of 23,000 and used a typical urban arterial daily flow profile volumes for the hourly breakdown (the profile comes from an ITE handbook for arterials).

I found it interesting because 1) it does show four lane always above 3 lane (which we knew), 2) it indicates that at the highest volumes we have the most similar but worst operation, and 3) it shows that four-lane operations drop more quickly for the peaks than the three-lane. So, you have average speed reduction simulated for most of the day (and it's pretty consistent), and similar operations during the peaks. If your measure of feasibility is speed reduction (and safety) - then the non-peak-hours (and the peak hours to a much much smaller extent) support that. In both cases (IA and WI) we have found that this difference gets bigger as volumes go down (peak and non-peak) - which isn't surprising. But if you base feasibility on a minimum reduction in travel speed - then the peak-hour is the most likely hour to support this, and the non-peak-hours say the opposite.

This graphic or something similar could possibly be used to show the the safety/operations tradeoff that needs to be considered. I have no doubt that it also tell us other things. I would expect that the lines would just become further apart (and maybe the 4LUD drops would be smaller) as volumes decreased.

Keith K. Knapp

Assistant Professor
University of Wisconsin - Madison
Engineering Professional Development
432 North Lake Street #713
Madison, WI 53706