When I signed up for Hubway in August 2015, I did not anticipate how I would start paying attention to intersection design. In hindsight, at least a minor uptick of interest was inevitable. Knowing how to identify safe intersections tends to come in handy when 'playing car' as I pedal around town.
My post about a traffic intersection in Cambridge back in April reflected this new interest. In thinking over the traffic light pattern at Beacon and Kirkland Streets, I stumbled upon a decent rule of thumb for anyone seeking to improve safety at a stoplight: extend the length of the red light. However, other approaches may work better for different styles of intersection. I recently came across such an example at a traffic signal on the border of Cambridge and Arlington.
The intersection in question is at the meeting of Alewife Brook Parkway and Mass Ave. Each street supports two lanes of traffic in both directions. There is additional lane space for cars preparing a left or right turn. The complexity is eased with specific left or right 'turn only' arrows.
At these traffic lights, the standard pattern is to allow cars facing each other on one street to simultaneously turn left until the arrow goes red. Then, the lights for the same street turn green for cars waiting to go straight or to turn right. Next, the pattern repeats for the other street. Again, the left turns are made first and the cars going straight or turning right go next.
At this particular intersection, however, the order was reversed. Here, cars went straight or turned right first before those waiting to turn left were allowed. When the other street got its turn, this order was repeated.
I'm sure this pattern was not invented in Cambridge (or Arlington) but it was my first-ever exposure to it. It made me wonder why anyone would bother to reverse a well-established traffic pattern. I never heard anyone cite the 'prioritization' of the left turn arrow across two lanes of traffic as a primary factor in a car crash. It never felt dangerous to me when I turned left using such an arrow. Surely, these types of intersections were not considered major traffic issues.
Why fix what isn't broken? The only logical explanation suggests the intersection became safer after the change. Otherwise, the new pattern would be reversed as soon as it became clear the changes were harmful.
I studied the intersection for a short time until I figured it out. The intersection's improved safety came into play if a car made a mistake by running the red light; the new pattern made it less likely a law-abiding car from the other street would accelerate directly in front of it. (1)
Consider a car turning left. In the traditional pattern, the turning car exits the intersection directly in front of the cars waiting just behind the stop line. These waiting cars are the next ones through so, as the light signal changes, all the possible action is condensed into two corners of the intersection. In other words, the cars exiting are leaving the intersection in the very same space the next cars are waiting to accelerate into.
If the car turning left runs the red, the signal will beckon a waiting car into the space the light-running car is turning left into. In this case, the law-abiding driver has the width of the crosswalk to stop in time without getting into a crash.
In the new pattern, the next cars through are on the street the car turning left is entering. These cars are also waiting to go straight (or turn right). However, the stop line where they wait is several car lengths behind where the left-turning car is driving. Even if the left-turning car runs the light and surprises the next drivers through, the chances of a collision are reduced because there is more space between the next cars through and the car running the light while turning left. The action is spaced into four corners of the intersection instead of two. If a law-abiding driver starts moving forward before noticing a car running through a red 'left turn only' arrow, there is half an intersection worth of space to stop without getting into a crash.
By exchanging the order of left turn arrows and standard straight/right green lights, the traffic engineers created a larger safety margin in the event of a driver running the red light. The best part of this method is how the safety margin is increased by shifting the burden of stopping to the cars more likely to stop (those already at rest) and affords them more space to hit the brakes than in the design of the original intersection pattern.
A few weeks later, I recognized the new pattern at another Cambridge intersection. This one is at Mass Ave and Sidney Street in Central Square. Studying the impact at this intersection revealed an additional dimension to the safety aspect I did not consider the first time. Here, it became evident the design is better not just for cars but for pedestrians as well.
I suspect this is because most pedestrians assume they can safely cross an intersection as long as they are walking with traffic. I'm not sure why this is. Perhaps pedestrians expect all intersections to mimic those found at major intersections in big cities (which automatically put a walk signal on to align with the direction of traffic). In any event, the end result at this intersection is, once traffic stops on Sidney Street, pedestrians tend to resume their strolls along Mass Ave and start to cross without looking.
The problem is if the left turn arrows come on first, pedestrians are often wandering directly into the path of a turning car. To make matters worse, since the car is using a turn arrow, there is an expectation of an unimpeded driving path. It is also harder for these cars to stop since they are usually picking up speed as they make the left turn.
The new pattern protects pedestrians from making this understandable yet careless mistake. When the flow of traffic on Mass Ave stops, pedestrians will wait at the curb assuming the cars on Sidney Street will go next. When cars turn left from Mass Ave instead, the chances of a pedestrian being in the crosswalk are significantly lower in comparison to the original pattern.
Of course, this intersection (like many others in Cambridge) already relies on the rule of thumb to 'hold the red' I referenced at the start of this post. The signal for pedestrians comes on a couple of seconds before the green light for cars. This is the key to making the new pattern run smoothly. By allowing pedestrians to start crossing first, the potential danger of cars turning right is almost entirely eliminated because a car will see the pedestrians before starting the turn. If the walk signal and the green light came on simultaneously, the result of adjusting the traffic pattern would be a shift rather than a reduction in risk.
So, although there are many safety improvements to be made through improved intersection design and constant reevaluation of accepted patterns, I remain sure the single biggest step forward is to simply hold the red light. Just an extra second or two would be enough at most intersections, I'm sure, to give everyone involved the chance to stop moving, take an extra look, and pass through the intersection once it is safe to do so.
Footnotes / imagined complaints
0. Let's look at the other case: a car about to drive straight...
This case is not as clear cut but the basic principle of having more space to stop applies. In the traditional pattern, the next cars through are waiting to turn left onto the same street the first car is driving on. If a car driving straight runs the red light, the next cars through will either appear directly in front or attempt to turn left onto the same street.
In the new pattern, the next cars through are those turning left onto the cross-street. These cars will cut into the straight-driving car's space. There is not a major change in terms of space here. But there is an advantage, at least for the law-abiding left-turning driver, because the law-breaker is approaching from directly in front. Once the law-breaker crosses the stop line, it becomes obvious that the red light is being run. The left-turning driver can see this and stop (and has half the intersection to do it).
In the traditional pattern, the left-turning driver would have to see the other driver coming in from the left, a relatively unnatural angle compared to looking straight ahead. The danger comes much quicker as well- the law-abiding car would get struck just as soon as the car crosses the stop line.
It occurred to me during this post that, for perhaps the first (and maybe last) time, it would be very helpful if I included an illustration or two in this post. Oh well...
The improvement to safety is less well-defined than the case in the main body of the post. But I think there is a good argument there. Regardless of whether the case in this footnote further improves the intersection, I think the benefits seen from the first example are more than enough to justify changing the traffic pattern.
1. The first draft of this paragraph was a little, er, less balanced...
"The only logical reason I could come up with- as usual- was to protect drivers from their own insatiable greed. In the more common traffic pattern, a car that runs the light while turning left is in the path of the cars on the opposite side waiting to drive forward. Until the full turn is complete, the idiot driver is in danger and temps the law-abiding drivers to demonstrate their understanding of traffic jurisprudence."
How to explain this? Let's see...I'll blame alcohol!
In any event, I think the final version is more, er, concise.
