Last time, I talked about pedestrians and their impact on traffic analyses. Continuing in that vein, this week I’m focused on u-turns. When examining intersection volumes, often one or more u-turns are a part of the turning movements. So how do you handle that typically small number of drivers?
For the software programs I use, u-turns can easily be integrated into a microscopic or macroscopic analysis. I suspect most people are also aware of the ease of adjusting your network to accommodate u-turns at one or more intersections. Therefore, the question is not how to analyze those turns, but rather should the u-turn movement be analyzed?
For my studies, I tend to consider the need to analyze u-turns based on a few factors:
- Knowledge of the situation
- Integral to the analysis
Are u-turns legal at the area in question? If only one or two are recorded and the movement is illegal, then I would typically just add the u-turn to the left turn movement and call it day. If the movement is legal, or you have a handful or more u-turns, then I tend to look deeper and try to get more knowledge of the study area.
Hopefully, as part of your study, observations of the intersection or intersections in question have been completed. Based on that or your other knowledge of the area (do you or a co-worker routinely drive through the study area), are u-turns common?
A call to the city engineer or local traffic engineer can also increase your knowledge of the area. He or she will likely be a lot more familiar with operations in the area than you are. If I determine that a u-turn is a common movement, I will then account for it in my analysis. A u-turn that is common will also be an indication to me of a potential operational issue to examine, i.e. why is the u-turn movement so common?
Finally, I consider whether the u-turn is or will be integral to my analysis. For instance, future access management on one of my projects would prevent a left turn movement from a side street. The easiest way to duplicate the movement in the future was a u-turn at an adjacent intersection. In another study, poor signing at an interchange led people to think the access to an entrance ramp was a left turn (typical diamond) and then discovered the ramp was a loop accessed from the right side – the u-turn at the next intersection was necessary to get back to the ramp. In these cases, analyzing the u-turn was definitely integral to the study and mitigation changed the number of u-turns, and the analysis results.
Mike’s Take: U-turns are typically analyzed if there are more than 20 of them in the peak hour, access management will increase the number of u-turns in the future, or if you’re analyzing a roundabout. Otherwise, typical practice is just to add u-turns to the left turning movement volume and skip analyzing them separately.