Martin’s post about signal priority on the First Hill Streetcar brought to mind an exchange I had with SDOT staff about signal priority on the RapidRide D line, which will replace Metro’s current Route 15. In particular, I raised an issue that an STB commenter had mentioned to me, namely that the inbound left turn from Elliott Ave W onto W Mercer Pl (where the bus turns from Interbay to make the deviation through Lower Queen Anne) is, under its current configuration, likely to cause unacceptable delays:

Regarding the D Line (Downtown-Uptown-Ballard), there is a signal at Elliott & W Mercer Pl that, from 2 PM to 8 PM, presents a very long (3-4 minutes, I believe) red signal to traffic turning from Elliott to Mercer. This will have a dramatic effect on the speed and reliability of the inbound D Line unless it is fixed. What kind of signal priority will the D Line receive at this light? If it’s the usual ~15 second hold-or-advance-green it will be completely inadequate at this light.

Here is the response I received from Enrique Garcia, Signal Timing Engineer at SDOT:

The reason for the wait time on the southbound left turn movement is that, in the PM peak commute hours, there is a high influx of northbound traffic. Our goal is to provide progression on the main arterial for this large traffic volume. This may result in side street and left turn movements having to wait a bit more. The idea is that once you are on a main arterial with good progression, the stopping should be minimal. I monitored this location and although the wait time can be a few minutes for the southbound left turn, the signal clears out the queue once the movement is serviced.

We are currently working on the re-timing of signals for the installation of the D Line Metro project. The specifics have not yet been finalized, so at this time we cannot say what kind of signal priority will be implemented at this location. We will, however, be taking an in-depth look at the timing and adjusting it to accommodate all modes and directions of transportation as efficiently as possible.

This answer, while totally reasonable as a discussion of engineering trade-offs, does not address* the policy question of whether good signal progression for arterial car traffic is more important than ensuring speed and reliability on a bus route that is being held forth by Metro and the City as a flagship BRT route. Ultimately, it doesn’t matter whether your transit runs on steel or rubber tires, what matters is whether it’s frequent, reliable, and a speedy way to get in to, out of, and (most importantly) between dense urban areas. SDOT aren’t to blame for RapidRide’s 15-minute off-peak headways, but they have have a lot of say in how fast and reliable the C, D and E lines will be, and Elliott & Mercer is where we shall find out exactly how much the City of Seattle is going to Walk, Bike and Ride the talk.

* And I’m not faulting Garcia — I wouldn’t expect an SDOT staff member to address a policy question such as I’m doing in this post.

55 Replies to “Signal Priority at Elliott & Mercer”

  1. There is only one optimal solution for that intersection … build a flyover for SB traffic. All other solutions will cause delays in one direction or another.

    1. If your definition of “optimal” doesn’t include cost, then yes. NB car traffic isn’t so massive, nor RR so frequent, that aggressive signal priority there would gridlock Elliot. I’m told that the signal in question used to have a shorter cycle in the peaks — the prolonged red is part of why SDOT had to lengthen the SB turn pocket recently.

    2. Bingo, Gordon,

      If there’s ANYwhere a flyover makes sense it’s that intersection; you don’t even have to “land” the exit. Just anchor it to the hillside and drive on.

      Of course, the good burghers who live in the five blocks west of Queen Anne on 15th West Mercer will have a cow because it will guarantee that some traffic previously turning at Harrison or Thomas will come up the hill via Mercer.

      The other admittedly more spendiferous option is to move the bus line one block south and build the flyover for buses only to Republican. It would be a pretty steep grade but would avoid the Mother of All Traffic Jams at Queen Anne and Mercer.

      1. A flyover there would be very spendiferous and dramatically increase traffic speeds and volumes on a narrow arterial with no sidewalks. It as an optimal solution for moving buses and cars, terrible for residents, bikes and pedestrians.

        When both Norman and d.p. agree on something (as in the thread below, discussing how gratuitous that long red for the turn is), you should probably take note. This intersection does not need a flyover. It needs the signal timing SDOT had before they decided to treat Elliott as a freeway for northbound peak traffic.

      2. It needs the signal timing SDOT had before they decided to treat Elliott as a freeway for northbound peak traffic.


        The “short burst” greens also did a much better job of distributing traffic on Mercer at the crest of the hill. You rarely got the 3-block backups that you now get after releasing many minutes of pent-up traffic at once. Traffic that, of course, prevents the queue-jump-less bus from pulling out of its stop at 3rd West.

        Frankly, SDOT and Metro need to tell Lower Queen Anne that they can have parking on West Mercer OR they can have RapidRide through the heart of their neighborhood, but they can’t have both. Detour + no bus lanes is why RapidRide will win no hearts and no converts.

  2. Yes thanks for adding this, I was going to make a comment on Martin’s post about that.

    Just because transit has TSP, the befits of it can vary significantly, especially in situations in which that priority directly conflicts or crosses large vehicle volumes flows, which this is a perfect example of.

    For the SLU Streetcar the intersections of Valley and Western and Valley and Fairview is another perfect example. The Streetcar has it’s own phase in the signal cycle, but because of this it often has to wait for a significant portion of the signal cycle till it’s phase comes up.

  3. Elliott might be the most misspelled name in Seattle.

    I’d love that route to have signal priority of course.

    1. Heh heh–I was just thinking that as well! (In addition to Sand Point and Green Lake both being two words, not one–and Wedgwood, not Wedgewood. Elliott still takes the prize, though.)

      1. Montlake/Mountlake Terrace is one I have trouble with.

        Apparently, all that was need for immortality in the 18th and 19th centuries was a midshipman’s position on an voyage of exploration.

      1. iPhones automatically add the apostrophe, whether or not it should be there. It is annoying.

  4. On policy: Yes. We need a strongly worded, clear policy that transit travel always trump car travel. I’d even add metrics to it: each transit passenger minute should be worth X SOV passenger minutes. This number should be high, as we care more about future use than how long someone waits right now.

    On this signal: Crazy. 15th has a clear bottleneck at the Ballard Bridge, and the entire stretch of 15th south of that point is simply a car storage lot preparing for that bottleneck. Someone speeding NB past this intersection will just find themselves in the bridge queue.

    1. (and by 15th I include Elliott as well – speeding up on Elliott just makes you hurry up and wait on 15th)

      1. True efficiency requires that one person minute = one person minute, regardless of whether that person is a transit passenger, an SOV passenger, a bicyclist, or a pedestrian.

      2. Not true in the slightest. Keep in mind how many people you can move on the same road if they all rode buses. We need to build in an incentive to move more people to mass transit, or travel for everyone is slower.

    2. I’ve wondered if the fastest way to serve Ballard (short of a new subway, etc.) is actually via SR 99, straight up to Fremont, with a transfer station there, then out 39th / Leary into Ballard. Wouldn’t that be faster than 15th and the Ballard Bridge when it’s backed up? The Aurora Bridge never goes up. The traffic on Leary never seems that bad. Such a line could swing up to serve the area of 15th/Market and west from there, or whatever makes the most sense in Ballard.

      Clearly that isn’t the plan with RapidRide, etc., but I wonder if it wouldn’t actually be faster and more reliable. Is the Uptown connection super important?

  5. I drive through this intersection all the time, heading south on Elliott and turning left to go up to Mercer. In the afternoons, when the red for my left turn is very long, there are virtually ALWAYS significant gaps in norththbound traffic on Elliott, during which the left turn light remains red, so no traffic at all goes through that intersection for long periods, which is a total waste, obviously. You would think SDOT would be able to have sensors tell when northbound traffic on Elliott has a long gap in it, and give at least a short green light for turning left from southbound Elliott during those gaps. That intersection is very inefficient.

    At the same time, southbound traffic on Elliott in the pm is the off-peak direction, so there should be many fewer bus riders heading south on Elliott there than heading north. So, you are not impacting nearly as many commuters as you would be if that situation existed in the morning commute.

    Also, within just the past couple of years, at that intersection there is now a green light at almost every cycle for traffic heading east through the intersection, because of that new development west of Elliott. So, the traffic southbound on Elliott turning left to Mercer now has to wait for the few cars coming from the west to cross that intersection, which adds about another 30 seconds wait in each cycle. In effect, that was always a 3-way intersection until a few years ago, when that new development went in, and now it has another street entering the intersection from the west, which was never given a green light in the past.


    1. RapidRide passengers here are only impacted in the southbound direction (as northbound buses are coming from Mercer and don’t have to wait for any signals to merge). This situation is therefore just as unacceptable in the (crowded) counter-peak direction as it would be in the (slightly more crowded) peak direction.

    2. In regards to the signal issue, the problem is that when a gap is detected the signal goes through a process of steps. First, the pedestrian phase is cut short and goes to the flashing hand. Second, after a set time (potentially 10 to 15 seconds or more depending on the width of the roadway), the yellow for NB traffic starts ( 3 to 4 more seconds). Then, maybe after an all-red time of 2 seconds, can the SB left turn begin.

      This is why permissive phasing (you take a left turn when a gap appears rather than wait for a signal) is more efficient. Unfortunately, due to other policy constraints, dual left turn lanes always get a protected/prohibited phase.

      1. At Elliott and Mercer, the through-car-traffic signal remains green for almost a minute after the walk-hand turns red.


        In this town “pedestrian safety” just means “severely truncated legal time for pedestrian movements.”

      2. I would say the obvious solution would be to have the green for left-turning off Elliott to Mercer twice as often, but shorter. It really is ridiculous to have lines of vehicles in both left-turn lanes waiting at a red light with no traffic heading northbound through the intersection for periods of up to 30 seconds at a time. And this is a regular occurrence. I would say the traffic engineer quoted in this article [ad-hominem] has not stood at that intersection and watched it for a few afternoon peak hours for several days. The red light for the left turn lanes should and could be much shorter than it currently is.

      3. Believe it or not, that’s precisely the way it was until a year or two ago.

        45 seconds red. 15 seconds green. Repeat.

        At all times.

        Then they did a repaving/rechannelization project between there and the Magnolia Bridge. Traffic engineers got involved, and now we have the current situation.

        The old way was arbitrary, but it worked quite well. Studies or no any time SDOT changes anything, cycles grow longer and slower, usually to their detriment. Buses usually suffer the most. We are moving in the wrong direction.

      4. I think the change was made when they made Mercer cross Elliott because of the development on the west side of Elliott. That traffic signal now gives eastbound traffic through that intersection its own green light. Before that development, that was just a driveway, and it never had its own signal, as it does now.

        And, of course, all that new development has meant a lot more traffic through that intersection.

        But, you are right. They really screwed it up. It is really poorly operated now, and it should be changed.

      5. That was definitely a part of it.

        The change also occurred concurrently with the extension of the BAT lanes in the couple of blocks leading up to the Ballard bridge, though again this benefits only peak express and Magnolia-bound buses. The core, much higher-volume 15/18 service got the short end of the stick).

  6. Does anyone know at what level does policy needs to be set?
    Who do we pin down and have them say, “yes, buses > cars, we should put in lanes & signal-priority wherever possible, even if it slows down cars, and even if it slows down a lot of cars”, or have them say, “no, we’re trying to keep everything moving as smoothly as possible, even if it means buses are stuck in traffic”.

    And, if they say the latter, who do we appeal to to change their minds?

    1. Someone more government saavy than me will come along with the full answer, but I assume it would be either the Mayor or the City Council. There is always an appeal with these two entities, generally ending in an election.

  7. There is a possibly very big issue coming pretty soon with this intersection that nobody has mentioned yet: the deep bore tunnel.

    A lot of southbound traffic on Elliott going straight through that intersection is heading to the Alaskan Way viaduct. There is a convenient onramp to the viaduct off of Elliott just south of the Battery Street tunnel, and a lot of traffic uses this (of course it is all screwed up now with the viaduct truncated and reduced to only 2 lanes by the stadiums, but a lot of vehicles still use Elliott to get onto the viaduct southbound).

    There will be no onramp to the deep bore tunnel off of Elliott. So, if you don’t want to use surface streets all the way past downtown (I think ths surface Alaskan Way will have about 24 traffic signals), and you want to use the deep bore tunnel to bypass downtown, you will have to make that left-hand turn off Elliott onto Mercer to access the tunnel from Mercer. This is likely going to greatly increase traffic on Mercer, at the same time as Mercer is going to be made into a 2-way street past Seattle Center, with only 2 lanes eastbound, instead of the current 4.

    A lot more traffic on Mercer, with half the lanes. This could potentially back up Mercer all the way back to Elliott at times, meaning traffic could not make that left turn from Elliott to Mercer even on a green light!

    A lot of us on Queen Anne think the deep-bore tunnel, combined with the Mercer screw-up is going to make it extremely difficult to even get off the hill any time near the peak traffic hours. Buses will get caught up in this (potential) mess, just like everyone else.

    1. Good one Norman. I suspect RR-D can hold it’s breath, gnash its teeth, and push the ‘gimme my green light’ button all they want, and it’s still going to be another Seattle Clusterfuck.
      Funny, the shortcut will end up following the 13 wire up and over the hill. Don’t tell the truckers.

    2. Not everyone else. There is always the option to hop on the bike and take the new Thomas St. overpass to the Elliot Bay Trail. There’s a signed route that goes all the way to Ballard.

      If could just give bikes a route across the Ballard Bridge that wasn’t overly hostile, this option would become even better.

    3. It’s actually worse than that. Until the Elliot-Western connector is built in 2018, which will connect the Elliot-Western couplet in Belltown to the surface street on the waterfront, there won’t even be a good way for cars to get to Alaskan Way from Interbay, so everybody will either have to go down Denny and through the downtown grid, or up Mercer and through the DBT. From 2016 to 2018, we’ll have two years where cars and transit are totally screwed-up through Lower Queen Anne.

      If I could think of a way to persuade Metro management to delay RapidRide D until 2018 when that problem will be somewhat ameliorated (and they’d probably have enough cash to run the D line at 10-minute headways), I’d have written a post advocating that a long time ago.

      I actually think the Mercer Corridor Project is awesome, but I have a different post lined up for that.

      1. If Mercer really becomes gridlocked and Denny gets worse too, that would convince more people that a subway is the only solution.

    4. I’ll reiterate the post near the top. If BRT is going to be the long-term solution for transit to and from Ballard — and where is the money going to come from for a “Green Line” LRT? — then investing in a flyover at Republican (two way, ideally, with a $2000 fine for use by private automobiles) would be a very good use of funds.

      Combined with bus-only lanes between Denny and Republican on First North and Queen Anne this would insulate Ballard Service from the messes on both Mercer AND Denny.

  8. This situation is like San Francisco’s official “transit first” policy – unless a car driver complains, in which case the driver gets their way.

    1. James – surely you jest: Geary St and VanNess Avenue are all disasters currently with the BRT “Improvements” unlikely to improve much of anything because, as in Seattle, SOVs will continue to be allowed to park all along these BRT routes. The Central Subway will likely be another unmitigated money sink. Then, there’s MUNI’s labor relations, poorly maintained equipment, lack of even basic customer service from a majority of bus and train drivers and on and on…
      Don’t get me wrong, I love me some MUNI: 3-8 minute service on most routes all day, a grid pattern of service, etc, but… the return on investment from these 3 projects is dubious at best, I’d say.

    2. There’s enough room actually to do that on Elliott/15th West, but the sad fact is that Geary has many one mile stretches with more people living along it those who live adjacent to the entire length of the D-Line.

    3. I rode one of the Van Ness routes in the AM peak last November, and the center BRT lanes make it less bad but not good. It’s not like a bus sailing down Eastlake when there’s no traffic. However, it’s better than the core Chinatown area where buses are packed full and take two minutes to load at each stop, then take another two minutes to get to the next stop because of traffic congestion.

  9. Folks, discussion of bypassing Uptown is off-topic for this thread. I didn’t want to talk about it, primarily because it’s a waste of time (the decision has already been made), but being as it’s apparently what everyone wants to talk about, I will write a post laying out the reasons for the Uptown deviation (complete with facts about travel times, reliability, and ridership patterns) in a couple of weeks.

  10. I like the point made by the mayor of Bogota in Urbanized where he said it was the job of the government to ensure equal rights to all citizens so that means a transit vehicle carrying 100 people deserves 100x more share of the road space/time.

    Maybe framing it this way still won’t justify public transit priority to us Americans because of the incredible vastness and capacity for selfishness that we have engrained in our culture.

  11. The reason for the wait time on the southbound left turn movement is that, in the PM peak commute hours, there is a high influx of northbound traffic. Our goal is to provide progression on the main arterial for this large traffic volume. This may result in side street and left turn movements having to wait a bit more.

    Indeed, E-W traffic in this city is nearly always lower on the list of priorities, regardless of the time of day, because our transport infrastructure is distorted in order to cope with commuter traffic.

    The NS-EW dilemma here is reflective of another, larger policy tangle: to what extent should the mobility needs of Seattle residents be subordinated to the needs of commuters who find their livelihood in Seattle yet choose not to live in the city? From a purely fiscal perspective it seems Seattle benefits from the daily ebb and flow of nonresidents but at the same time this benefit is traded for costs, as demonstrated in the case at hand. In terms of daily impact on Seattle’s livability by commuters the generally compromised ability to travel E-W is one of the most visible and frequently annoying aspects of accommodating ex-pat workers.

  12. This is a problematic light, but what about Dravis or Leary? Dravis, while not as much of a delay as Leary or Mercer Pl, is annoying, because the bus has to “exit” 15th, wait at a light (not too bad of a light), then “enter” 15th. Has Metro considered putting a stop under the bridge on either side? There’s more than enough room for one, with stairs and a ramp. Plus, they could extend the BAT lane north of Dravis.

    Leary, on the other hand, you have to exit 15th down to Leary, wait at the light, which can take a couple minutes, then reenter 15th, which requires a difficult merge on the bridge. Unfortunately, to avoid this, would require a rebuild of the bridge approach to add a transit pulloff over Leary with a stop. The biggest advantage, however, would be right after the bridge has come up, when being trying to merge onto 15th from Leary takes 10+ minutes.

Comments are closed.