Archive | Transportation Operations

California’s New Performance Metrics & Getting What You Measure

The way we quantify how well our roads are (or aren’t) working isn’t something that tends to get a whole lot of play outside of the transportation wonkery, but it has a drastic effect on policies and livability. The most commonly used metrics to describe our system, including the infamous level-of-service metric, are drawn from something called the Highway Capacity Manual. See how the name of the manual doesn’t really imply that it’d be terribly useful for designing safe, welcoming local streets? Most jurisdictions don’t, and that is one of the main reasons why too many urban streets have become de facto highways.

Over the last several years, traffic engineers have increasingly been letting go of their long-held fondness for LOS and other traditional performance metrics, and in many cases are even leading the charge against them. The inadequacy of LOS as a primary measure of performance is perhaps most clear in California, where the state’s all-powerful Environmental Quality Act buttressed the importance of the metric by (ironically) requiring that environmental analyses consider LOS when evaluating the impacts of a project. So perhaps it’s not surprising that California has struck the most major blow to date to LOS, with new guidelines that evaluate projects based not on how much they will increase auto delay but instead on how much vehicular traffic they induce.

It’s hard to overstate how radically that this departs from the status quo. For many jurisdictions, an over-reliance on capacity-based metrics have produced policies that favor anything capable of moving one more car. California’s new standards appear to turn this idea on its head, favoring policies and land uses that create one fewer trip (or one fewer vehicular mile traveled). Thus, it would appear that analyzing a particular idea with California’s VMT-based methodology—whether to widen an intersection approach to include a turning lane, for example—might lead to the opposite conclusion as analyzing it with traditional metrics. Though the turning lane would certainly reduce delays and thus improve LOS, there’s plenty of evidence to suggest that it would also induce new traffic and thus be undesirable (or impermissible, even?) based on the VMT-based metric. That’s huge!

The VMT-based metric is neither perfect nor wholly complete. Success of the methodology relies heavily upon our ability to estimate the number of trips a project may generate which, as I described here, is something of an inexact science. The need to include trip length in these projections serves to widen the gulf between the data that’s needed and the data that’s available. And the state’s guidance [pdf] for utilizing the new methodologies do not appear to significantly improve upon the half-hearted methodologies engineers currently employ to evaluate safety ramifications of a given project. Finally, a scathing white paper [pdf] from UCLA’s School of Public Affairs suggests that the VMT-based methodology may not even be all that great at its purported goal of teasing out the environmental impacts of a project, although I’d hasten to challenge some of the assumptions their analysis is based upon.

Despite its shortcomings, the new VMT-based methodologies represent a big step forward and I’ll be curious to see how they’re applied by colleagues in California. Those of us who favor a multi-modal and safety-oriented approach are regularly stymied by traditional metrics that concern themselves with only capacity and delay, only as they pertain to autos, and only during the busiest 1% of the day. Though it leaves important considerations unaddressed, California’s new methodologies offer a way to overcome these hurdles. Time will tell what effects this will have, but there’s plenty of reason for optimism.

So what does this mean for Portland? That’s a good question. It’s now been two years since the city launched a project to update our performance standards, but sadly this effort seems to have disappeared into the same memory hole as bike share. Keeping with a storyline that’s becoming too familiar, others innovate while Portland waits.

The Story Told by the Blinking Hand

Joe Rose penned an interesting Q&A yesterday that takes up one of my favorite topics: the Flashing Don’t Walk (FDW) phase. While he caricatures the consequences of the widespread non-compliance with the FDW indication just a wee bit—it may cause motorists (gasp!) delays, and in extreme circumstances may even cause (gasp! recoil!) cycle failures, where all the cars don’t get through on a single green—he is correct that non-compliance is indeed widespread in downtown Portland.

That’s precisely what makes it a favorite topic: The FDW indication, at least when it comes with a countdown clock, is in clear violation of Brian’s First Rule of Traffic Engineering.

When people are disobeying your traffic control device en masse, there’s probably something wrong with your traffic control device (See: Circle, Ladd).

So what, then, could the issue be with something as universally accepted as a FDW indication? Perhaps some insight can be gleaned from comparing how signals handle pedestrians to how they treat vehicles.

Though we claim to prioritize pedestrians above other modes, our assignment of green time suggests otherwise. Pedestrians see a Walk phase for a given movement downtown for about 9 out of every 56 seconds during daytime hours, or about 16% of the time. By contrast cars will be assigned a green for 25 seconds, good for about 45% of the overall time. On the transit mall these percentages are often more skewed, since the pedestrian signals sometimes ‘rest’ in the Don’t Walk phase to quickly serve approaching transit vehicles. So if you were to assume random arrivals at an intersection, a car (or bike) is much more likely to see a ‘go’ indication than a pedestrian.

Of course, arrivals at intersections aren’t random; we specifically time our signals so a mixture of cars and bikes can progress smoothly through downtown at about 13 miles per hour. Since this is a little faster than most people not named ‘Bolt’ tend to travel on foot, the relative chances of arriving at a green indication are skewed further. And since turning movements on the downtown grid are what we’re concerning ourselves with here, let us not forget that these are commonly allowed on reds.

To be fair, our signals do some good things for pedestrians. The quick speed at which our downtown signals rotate through the indications (cycle length, in the jargon) serves to reduce pedestrian delay at the expense of auto capacity. And a happy consequence of the 13 mph progression speed for vehicles is that, if you were to walk opposite vehicular traffic at about 4.2 mph—a nice, brisk pace—you’ll catch all the Walk signals (and incidentally, you’d make it across an intersection in just more than half the time the countdown clock thinks it takes you). Still, though, relative to their wheeled counterparts, pedestrians are asked to endure a lot of delay at signals if they’re to follow the letter of the law.

To blame, of course, is the length of the FDW phase. Since pedestrians move more slowly than other modes, the time needed to clear them of the intersection—the clearance interval— is necessarily longer and thus the Walk time is naturally limited. Moreover, as the population ages average walking speeds are getting slower, so FDW times are getting longer. Not fair, but unavoidable, right?

A fun fact about Oregon is that we’re only one of a handful of states that treat the FDW and yellow lights with consistent logic, requiring the user to stop at both. But vehicles can proceed if it is unsafe to stop at a yellow, so as long as you know how to say, “Sorry, officer, I thought it’d be safer to keep going!” then the law really just means that you’re okay as long as you’re not still in the intersection at the onset of red. Still, though, this is better than the majority of states, which give vehicles far more leeway to enter an intersection on yellow.

But a when a countdown clock is present, it provides a lot of information, and so perhaps it is a useful ally here. Former Chicago DOT head John N. LaPlante examined the effects of countdown clocks [pdf] in a 2008 presentation to the Institute of Transportation Engineers, finding that more pedestrians began to cross during the FDW phase when a countdown clock was present, but fewer began walking late in the FDW interval, and fewer remained in the crosswalk at the onset of Don’t Walk. And guess what? Crashes decreased.

So what if we applied yellow-light logic to the FDW indication, allowing pedestrians to enter during this phase so long as they were clear by the end of the countdown? Revising our crossing laws to better reflect how people naturally interact with the indications makes sense from an engineering standpoint, and would be a powerful way to reinforce our stated claims regarding modal priority. And lest our fair city’s new-found fear of solving a problem with some new thinking cripple us, we can take comfort in the fact that those wily urbanists down in Salt Lake City provide precedent:

A COUNTDOWN CLOCK (displaying time in seconds remaining in the pedestrian crossing phase) in conjunction with the flashing orange UPRAISED HAND means that a pedestrian facing the signal indication may start to cross the roadway in the direction of the signal indication, but only if such pedestrian is able to safely walk completely across the street or to a safety island before the COUNTDOWN CLOCK shows no remaining time.

Traffic laws, right-of-way allocation, and signal timing combine to tell you just about everything there is to know about the transportation priorities of a particular place. That story is on full display when the countdown clock starts rolling.

Revealing a Certain Modal Bias

This announcement from the City has drawn some ire in active transportation circles:

Green street construction will close bicycle lanes and sidewalks on NE 122nd Avenue between NE Fremont and NE Shaver. Construction starts tomorrow (Tuesday, August 23) and will last about ten weeks. The work will also close on-street parking spaces in the construction zone.

At least one observer has suggested that any reasonable alternative bike/ped route takes you at least a mile out of your way.

Surely we can do better?