The original version of this post by Roger Rudick first appeared in Streetsblog SF and is shared with permission. Additional reporting by Max Harrison-Caldwell.
A speeding car and subsequent crash killed Andrew Zieman as he stood on the sidewalk near Sherman Elementary School on Franklin and Union streets last November. In the wake of Zieman’s death, the San Francisco Municipal Transportation Agency promised a lane reduction and other “quick-build” measures to help reduce speeds and narrow the one-way, five-lane arterial (three running lanes, with two parking lanes). But apparently the lane reduction was quietly removed from the plan over the summer.
The culprit, it seems, is an engineering and planning culture that’s still using Level of Service, an arcane and thoroughly debunked metric that analyzes projects based solely on expected traffic congestion at intersections.
First, some background: Elias Batshon, longtime owner of the Sherman Market at the corner of Franklin and Union, says he has seen “hundreds of accidents” at the intersection and that he witnessed the deadly collision that pinned the 30-year-old Zieman against a building, across from the school where he worked. “People go crazy going down that street,” he added, referring to drivers speeding on Franklin to make the green lights.
Shortly after Zieman’s death, Sup. Catherine Stefani made it clear that San Francisco streets need to be safer, especially near schools.
SFMTA responded with a quick-build plan that included sharper corners, new crosswalks, daylighting, and most significantly, a lane reduction.
The agency had recent experience to lean on. It recently slimmed down California Street in the Richmond District with a lane reduction, or road diet. The initial results showed more safety without more congestion, and the agency has touted success on other city streets as well.
As seen in the images below, SFMTA presented the plans to the Pacific Heights Residents Association back in July. Shanan Delp told Streetsblog that SFMTA planners told him a lane reduction was fundamental to the plan. But then it was pulled. “I got the strong impression a road diet was a logical solution to this problem, so it’s unsettling to see this so casually tossed aside,” he said.
Delp, who lives on Franklin and regularly rides his bike or walks with his two small children, demanded to know why. He received this response Tuesday from Shannon Hake, an SFMTA planner:
Thanks for reaching out, and I understand your frustration and disappointment. This is the same process we go through with every Quick-Build project — determining which of our tools works to address the specific needs on a corridor. In the case of Franklin, our original scope included the entire suite of tools in the Quick-Build toolkit (including a road diet), and we anticipated developing a design with any/all of those treatments. Once we began the design process to respond to the community’s needs, we modeled the impacts of all the potential treatments. In the case of a lane reduction, our modeling showed that the impacts to that change were significant and would lead to major queues along Franklin Street and spillover impacts to adjacent streets. [Emphasis added.]
“To say that I’m disappointed in this change is a massive understatement,” wrote Delp in his reply to Hake. “When we did our walkthrough a few months ago, the door was wide open for a road diet (perhaps even more than wide open, it was emphasized as a no-brainer), and it really breaks my heart that this was pulled from the proposed design. And pulled because of a generalized desire to ‘move traffic’ that always seems to trump all other outcomes.”
So what was it about SFMTA modeling that led to the determination that a lane reduction would cause major queues and traffic spillover — and that these “significant impacts” matter more than conditions that led to the death of Zieman? (There were also school children on the sidewalk that morning.)
In an internal email obtained by Delp, SFMTA traffic engineer Alan Uy wrote this to planning staff:
I assumed vehicles would turn right on Vallejo and onto Van Ness as soon as they realized the road diet. Of course, this degrades the eastbound left turn Levels of Service at Vallejo/Van Ness from C to F. There’s no EBLT [eastbound left turn] at Broadway/Van Ness. If these diverted vehicles turned before Broadway, the northbound Levels of Service at Broadway/Van Ness would be a worse F. [Emphasis added.]
UPDATE, Sept. 14: SMFTA communications staff emailed Streetsblog the following: “It is important to note that no one is killing a project in this email, it’s important to note that safety is the top priority for all of our projects. The public expects us to keep an eye on traffic flow on the city’s few arterial roads even if pure [Level of Service] is not the top Key Performance Indicator.”
For readers not familiar with Level of Service (LOS), it was a legally mandated car-centric metric for large project engineering that is behind much of today’s traffic carnage. Essentially, focusing on LOS pushes every intersection to be built so traffic speeds during peak periods are nearly the same as in the middle of the night. This practice has resulted in a vicious circle of road widening that increased crossing distances, traffic speeds, and ironically, traffic congestion throughout American cities. In many places, designs to maximize Level of Service have made it almost impossible to get around without a car.
SFMTA head Jeffrey Tumlin helped lead the charge to end mandates for using LOS to measure projects back when he was a consultant. “So a street [ends up] three times wider than it ever needs to be,” he told an audience during a SPUR presentation in 2016. “But now walking on the street is a miserable experience and transit will not work, so the street drags down adjacent property values.”
Engineers have used Level of Service in California environmental analyses for years, but in 2013 the state passed a bill that was supposed to end the practice. However, it took years to weed it out, yet it clearly hasn’t been eradicated yet.
Including, apparently, among engineers at SFMTA.
“Franklin Street is a garbage design and this quick-build project is a once-in-a-generation opportunity to make strides toward making it a livable street,” said Delp. “If we fail to seize this opportunity we’ll wait for another generation before we make it a street people want to walk on.”
Roger Rudick is the editor of Streetsblog San Francisco.
