Four Protected Bike Signals Coming to Polk Street By May

The SFMTA has promised signals to separate southbound bike traffic from right-turning drivers at four intersections along Polk Street by May. Image: SFMTA

Today the SFMTA announced details about the first package of safety upgrades coming to Polk Street in the next few months. They include signals at four intersections that will give southbound bike traffic a separate phase from drivers turning right, making Polk the second street in SF to get the configuration.

By May, the SFMTA said it would install the bike signals at all four intersections in the Polk plan: Geary Boulevard, Ellis Street, Eddy Street, and Turk Street. The signals “will be implemented to address existing right-hook crash patterns,” the SFMTA said in an email announcing the upgrades.

The prevailing design of SF current bike lanes calls for people on bikes to merge with right-turning cars, putting them at risk of drivers who turn without looking. At the four Polk intersections, right-turning drivers will have a separate lane and signal phase. The configuration is widely used in cities like Amsterdam, and is planned for protected bike lanes on streets like Second.

The only street in SF that already has the configuration is Cargo Way in Bayview, where a two-way protected bikeway separated by a fence was installed in 2012. A similar configuration exists at Fell Street and Masonic Avenue, where a left-turn signal was installed to protect people in a crosswalk along the Panhandle’s mixed bike and pedestrian path.

As part of the first batch of improvements on Polk, the SFMTA said the conventional southbound bike lane will be extended from Union to Post Street by April. That space will apparently be created by narrowing traffic lanes.

When construction of the rest of the Polk project starts next spring, the southern segment of the bike lane will get green paint and a buffer zone. Many sections will run curbside, eliminating the risk of dooring.

The northbound Polk bike upgrades will also come next spring, with the construction of a raised bike lane from McAllister to Pine Street, which won’t include separate signal phases at intersections.

Pedestrian safety improvements are on the way this spring, include zebra crosswalks at 25 intersections and painted bulb-outs at five intersections. By summer, the SFMTA said it will install leading pedestrian intervals, which “allow pedestrians a few seconds of a ‘WALK’ signal before vehicles receive a green light at certain intersections.” By that time, daylighting will also be in place at “various intersections,” along with “new and relocated” loading zones to reduce double parking.

  • Mesozoic Polk

    Under CEQA, each change that adversely affects cars requires a minimum of 3.75 million pages of environmental review. How can SFMTA force this change on the neighborhood without publishing this necessary 15 million page document?

  • mx

    This design scares me. All over the rest of the city, we’re trying to teach right turning drivers to merge into the bike lane (in the dashed line portion) and make their turn from as far right as possible. Cyclists going straight can then pass right turning cars on the left. While people aren’t always very good at doing this (or signalling!), it’s at least how the rules of the road are written. That’s long been our strategy for prevent right hook collisions: cars should turn from as far right as possible and cyclists shouldn’t pass right turning cars on the right.

    Here, we’ll be asking people to do the exact opposite. Right turning drivers will get an arrow encouraging them to make their turns from the left, crossing over the bike lane without merging into it. I’m not convinced this design works so well on Polk, and I’m especially worried that it encourages the very same bad habits that we’re trying to get drivers to break at literally every other intersection in the city.

    I’m also concerned about the dashed yellow line through the intersection. Intersections with non-straight lanes have always felt unsafe to me, but I fully acknowledge that’s just a gut feeling. Is there any data on this anywhere?

  • hp2ena

    I agree. The weaving bike lane design does not feel safe at all. I think this kind of design would increase the chances of a right-hook, under two possible scenarios: when traffic and right turns get the green, some on bikes may confuse that as a green signal, whereas some drivers may turn right at a red signal (even if it is banned, I guarantee there will still be some drivers doing this) only to run onto a bicyclist’s path. Both may yield the same result: someone is going to get injured or killed as a result.

  • hp2ena

    Re. narrowing traffic lanes for a bike lane: Polk measures 68′ 9″. With 8′ parking and 12′ sidewalks on each side, that leaves 28′ 9″. Polk currently has traffic lanes that are 14′ 4.5″ wide, give or take. Muni buses require a minimum of 11′ to adequately operate. Assuming the SFMTA is striping a 5′ bike lane with a 1′ buffer, Polk is wide enough to have a southbound bike lane AND keep traffic lanes wide enough for buses to operate.

  • The real problem with this design is that the traffic signals probably aren’t triggered but they’ll likely be timed. This means that cyclists will end up sitting at a red light when there’s no right turning traffic, and cars will sit at a red light when there’s no cyclists passing them. Also the bike lane on this block is frequently blocked by trucks unloading in front of the Culinary Academy. When the bike lane gets blocked, cyclists will have to merge into the vehicle lane. I doubt very many cyclists will merge back into the bike lane only to sit a red bicycle light at the corner of Turk. If this traffic signal is timed, it should be timed with a green wave so cyclists going down Polk Street won’t have to slam on their brakes at the bottom of the hill to yield to vehicles that may or may not be turning right.

    This would be the perfect intersection to have a dutch-style protected intersection design.

  • 99% of drivers don’t ever merge into the bike lane, and attempting to teach them something that’s non instinctive is a hopeless cause. A Dutch-style protected intersection design is a better way of avoiding bicycle and vehicle conflicts, because their design makes it obvious what drivers and cyclists are supposed to do. There’s no teaching involved, no dangerous mixing zones, and conflicts are easily managed because everyone can see everyone else.

  • Jeff Gonzales

    Well, cyclists aren’t exactly stopped by red lights. As for drivers being stuck at red lights, who cares? And the delivery trucks could be dealt with if only there were enough PCOs.

  • mx

    Well said. I see the same basic problem, on a lesser scale, with the “dashed green carpet” bike lanes near intersections (there are several on Polk, such as Polk and Hayes). Drivers have no earthly idea what dashed green and gray pavement means. However, many have started to get the message that they can’t drive on red and green pavement. So when right turning drivers approach a corner with a green bike lane, they don’t merge into the bike lane as the law requires, but instead turn across it, risking right hooks.

  • Seriously? You’re arguing that the SFMTA should add more red tape to their safety improvements? Other major cities are adding protected bike lanes in a matter of months instead of decades like the SFMTA. Fortunately California ditched the car-centric level of service requirement last August. So you are incorrect. Safety improvements to our streets that reduce the flow of traffic thankfully no longer require multiyear delays.

  • hp2ena

    Disclaimer: Mesozoic Polk is Bob Gunderson’s long-lost cousin (aka a parody account).

  • Because when drivers are stuck at red lights, some will try to make a right turn on red even theres a sign saying not to. How many cyclists were killed on Market & Octavia by vehicles making illegal right turns? How exactly will Polycystic ovary syndrome help with delivery trucks parked in bike lanes? That’s the first thing that came up when i Googled PCOs!

  • Bob Gunderson is a parody account?!?

  • M.

    Bobby G. cruises and runs over people around the Upper Haight. Meso (@MesozoicPolk) haunts Polk Gulch, trampling on still extant, more evolved species.

  • M.

    How? MorOnline outreach 😉

  • M.

    Indeed. How about requesting underground bike sensors (photo)? Plenty of cyclists run red bike lights and *many* drivers on Polk swing into right turns from the left lane and am not convinced the signals will stop that – at least not for a while. Creating a green wave is more critical for cyclists on the uphill (northbound) but timing one is tricky – but certainly not impossible – due to the uneven lengths of the blocks and the need to prioritize transit on several cross streets.
    As to Dutch intersections, the Polk PM’s response was, ‘The culture isn’t ready yet.’ Which translates to, ‘That’s begging for litigation.’ But no harm in asking for them again.

  • M.

    Folks for Polk asked for Intersection Crossing Markings (Through-intersection Striping) along with many other painted indicators as are implemented in *many* other towns. ICMs do help all road-users know where they belong in intersections and are an added visual buffer for people in crosswalks.

    We were told that there aren’t enough US studies of them and that they haven’t been used in SF but there are, including at the foot of Polk St. However, the markings here are generally easily overlooked single dashed lines; we prefer more vivid and robust double dashed lines as are done elsewhere, especially until our ‘culture’ changes.

  • M.

    Folks for Polk asked for Intersection Crossing Markings (Through-intersection Striping) for the cycle paths, along with many other painted indicators as are implemented in *many* other towns. ICMs do help all road-users know where they belong in intersections and cycleway ICMs are an added visual buffer to protect people in crosswalks.

    FFP was told that there aren’t enough US studies of them and that there are none in SF but there are (photo), including a rather sketchy one at the foot of Polk St. Most of the markings in SF are like that one, single dashed lines which can be easily overlooked; we prefer more vivid, robust double dashed lines as are done elsewhere, especially until our ‘culture’ changes.

  • M.

    Generally good news in here, working to create more.
    Here’s a postscript to SFSB’s post on 26 Feb. ‘SFMTA Cuts Block of Polk Bike Lane Fought By Visionless Mayor’s Optometrist:’
    A recent visit to Hiura & Hiura Optometry elicited their defense of the ‘just one letter’ written 2 years ago and a false denial that they continued to spearhead opposition to any change to their block of Polk Street. The younger Hiura was reminded that public money and safety are being jeopardized by ignoring reams of data documenting how cycling infrastructure benefits everyone. He was asked why the City should be coerced into changing their plan based on nothing more than his hunch that it *might* temporarily affect business. His reply to the query was, ‘The City can do whatever it wants.’ We’ll take that as a request to go back to ‘the City’s’ final plan for a raised cycleway northbound from McAllister to California, not ending at Pine. Polk was never just about a street, and this sure isn’t just about one block.

  • mx

    You raise an interesting point about Muni buses. It seems unclear to me why we’re compromising the design of Polk St. to accommodate 40-foot buses when we’re already building a major transit corridor one block over. Is it perhaps worth considering relocating the 19-Polk and shrinking lanes on Polk to scale down the street, lowering speeds and bringing things to a more human scale? You’d also eliminate all the problems that occur trying to fit bus stops into the design along with protected bike lanes.

    And if people complain about reduced bus service, the answer is that we’re providing much better transit on Van Ness soon enough. Besides, after the Polk St. redesign is completed and lanes are removed from Van Ness, traffic will inevitably be so horrendous that nobody will tolerate sitting on the 19-Polk anyway.

  • Prinzrob

    Video or microwave detection would work better as it can distinguish between bikes and cars, and it could be oriented further down the block to detect bikes and give them a green light by the time they reach the intersection.

  • Prinzrob

    “FFP was told that there aren’t enough US studies of them”

    That’s BS. Bikeway intersection markings have already been approved both federally and by the state:

    And they are referenced in the NACTO guide the SF endorsed:

    They’ve already been used in many locations all around the Bay Area, including in SF.

  • hp2ena

    How else would you call someone Bob? 🙂

  • I’m dubious about this, myself, especially since I used to ride Cargo Way as part of a morning exercise loop both before and after the bikeway was put in. But maybe I’m not clear on the details for Polk.

    So if I’m a cyclist, I approach the intersection in the bikelane. What I fear will happen next: 90% of the time the special bikelane light is going to be red. It’s red when crossing traffic has right of way, but it’s also red whenever the north/southbound auto traffic has right of way. The only time I’m going to have a right of way to continue without stopping is when there was a cyclist ahead of me to set off the separate signal. Or, since I am comfortable with riding in traffic, I’ll move into the main traffic lane and ride through on the ‘regular’ green.

    End result: we install “beg buttons” in the street for cyclists which will have the effect of slowing down cyclists that follow the traffic signals and create another “we gave cyclists their own signal but they still ride through it” People Behaving Badly segment opportunity.

    Granted, the “beg button” detector on Cargo Way works surprisingly well (especially since you can see where it’s embedded right now) to trigger the special light cycle to start when the next opportunity comes along.

  • Cargo Way has sensors embedded in the ground that work pretty well–basically looks like one of the in-ground parking detectors that were installed around town: a 3-inch diameter epoxy/rubber filled hole.

    “Work well” here means the bike signal shows up at the next opportunity and I am not hanging around waiting for 3 cycles of the light as I get my bike positioned correctly. “Works well” also means I KNOW I will have a red bike-light on every approach to the intersection (and the shitty chain-link fence divider blocks auto-bike visibility in both directions (unless someone’s stolen that section of chainlink again)).

  • TL;DR version of previous comment: I hope this doesn’t mean “Four Intersections on Polk turned to Mandatory Stop via nearly always Red Bike Signal.”

  • Dark Soul

    Balanced Safety….and like the traffic signals will be obeyed.


Detailed Polk Street Designs: Plans for Safe Bicycling Still “Lackluster”

The SFMTA and the Planning Department presented detailed plans for Polk Street at the project’s final open house meeting yesterday. The new aspects include specific locations of bulb-outs, dedicated bicycle signals, left turn prohibitions, loading zones, and new trees and landscaping. Plans to improve bicycle infrastructure are still composed of a mix of protected, buffered, […]

Parking-First “Save Polk Street” Crowd Attacks Van Ness BRT

“Save Polk Street” has aimed its parking-first agenda at Van Ness Bus Rapid Transit. A couple dozen speakers protested the project an SFMTA hearing last week, distributing fearmongering flyers [PDF] claiming that removing some parking and banning left turns would “kill small businesses,” back up car traffic, and make the street more dangerous. The long-delayed Van Ness […]