Over the past month the Seattle Department of Transportation (SDOT) has been quietly testing their new Intelligent Transportation Systems program on the Mercer Street corridor. Usually shorthanded to “signal retiming”, the $13m Move Seattle project creates a dynamic corridor that adapts to traffic conditions in real time. 32 signals have been updated with the new software on Mercer, Roy, and Republican Streets from 3rd Ave W to Fairview Ave N.

The primary management imperative seems to be to shorten signal times in periods of heavy traffic, theoretically reducing the cascading gridlock that occurs as queues form during lengthy cycles and spill over into neighborhood streets. In times of heavily asymmetric traffic, such as the end of major Key Arena concert late at night, the system could keep eastbound lights green far longer, and make the lesser amount of north-south traffic wait. Anyone who has ever driven Mercer Street on a sleepy Sunday morning will attest that the current system has been rigid and static, slow to drive even when no one is on it if you miss the green-light wave.

SDOT’s early data shows significant improvement in travel time reliability for most drivers, significantly better eastbound flow, and slightly worsened westbound flow (see SDOT images below). Eastbound in the PM peak, when volumes are highest and delays worst, travel time has decreased by 2.7 minutes and average speed has increased by 1.4 mph. Doing a little math, for a 1-mile corridor this implies travel time improvement from 12 minutes to 9.3 minutes, and average speed improvement from 5mph to 6.4mph.

These are significant relative improvements, but put another way $13m just went to making peak car traffic go from barely faster than brisk walking to about half the pace of casual cycling. To the average observer, or to the driver whose sense of time is dilated by rage, the corridor will still feel very slow. SDOT estimates that the average peak driver will save about $4 per year in reduced fuel costs.

So signal timing can work to improve east-west vehicle flow. But when it does so, it must also impede north-south traffic, and all transit in the area is north-south. A total of about 13,500 vehicles use Mercer during the PM peak, and their prioritization can’t help but downgrade the priority of the thousands of peak commuters on Routes 1, 2, 3, 4, 5, 8, 40, 62, 63, 63, 70, 309, C, and the SLU Streetcar. If aggregate flow improves sufficiently that total systemic collapse is avoided, then maybe everyone actually will be better off. But even at its best, it tinkers at the margins at great expense. The fundamental space constraints that generate traffic will remain, and those who think tech will save us from traffic will continue to be disappointed.

53 Replies to “Mercer Corridor Project Shows There Are No Silver Bullets”

  1. Ah. So that’s why I had to wait for a very long cycle behind 3 cars turning left onto Mercer, then the light turned green only long enough for one of us to go.

    Of course there is a magic bullet, and it rhymes with medicated pansit wane.

  2. Just wait until the SR99 tunnel opens. Mercer will become a de facto transition ramp between I-5 and the tunnel. The traffic problems on Mercer will grow worse, as will demands to retime the signals in ways that are unfriendly to people who walk, bike, or ride transit. The Bay Freeway lives on…

    1. What are those people doing now? Taking I-5 down to Spokane Street to the West Seattle freeway? 99 and 509 only lead to a few destinations, so while they’re useful to a minority of travelers the majority will stay on I-5. And those that leave will actually free up space on I-5 downtown compared to if they remained on it in the old system.

      1. Mike – I take it you have never entered or exited the Viaduct at Western. Many cars enter and exit 99 at the south entrance of the Battery Street Tunnel. Think of people in West Seattle/Burien/White Center/Ballard/Magnolia/QA. If the North Portal entrance/exit was to serve so little traffic, why have it? All the traffic directed to Mercer will be additional traffic on top of the current traffic on Mercer.

      2. “99 and 509 only lead to a few destinations,” you know, like West Seattle, Burien, the Dwamish Industrial Area, and our major airport .

        If I5 is backed up and 99 is not, plenty of drivers will cut across Mercer simply to get through downtown. Especially during peak, I’m bullish on much of the pass-through traffic to take the tunnel given the toll and lack of downtown exits should keep 99 following better than I5 … which is, of course, the entire reason the tunnel was built, to handle pass-through traffic.

        I am expecting traffic on Mercer west of the portal to get much worse when the viaduct goes down, which will cause issues for the D line unless it gets bus lanes on Mercer. Would also help if the D could turn left from 15th onto Mercer from the far right (bus) lane, which should be doable with a dedicated signal. Both good uses of “early investment” ST3 money.

  3. It also severely screwed up pedestrian walk signals. Going kitty corner at Mercer and Terry yesterday took ~7 minutes. That is insane. My other regular walking routes have all roughly doubled in length (if I follow the crosswalk signs). Going to QFC went from 10 minutes to 20 minutes. Walking home from my friend’s apartment went from 20 minutes to 35.

    As far as I can tell, this change is actively hostile to everyone except drivers.

    1. Additionally, if you don’t push the beg button, the walk light won’t come on for the crosswalk parallel to the lanes that get a green light. This is incredibly hostile.

      1. Worse than this, inexplicably, “walk phases” only last a short time, then the red hand comes up, even though car and bike signals stay green for much longer.

        Basically, they expect walkers to never cross on a stale green. You must wait an entire light cycle til the next time around.

        My new policy is to cross on green, and ignore the decorative “ped signals” since they don’t function properly.

      2. I noticed this about two weeks ago, didn’t realize it was part of a retiming of the corridor. So much for Vision Zero.

    2. Does anyone know the rules? When doing traffic modeling, does SDOT have to include estimated impact on pedestrians (+ bikes/transit/…)? Do they have to include it in some of the reports afterwards?

      I feel that even if it is done, it is hardly ever reflected in media.

  4. IF it actually does reduce box blocking, it could conceivably be a wash or even a slight improvement for peak-hour transit (at least the routes most affected by box blocking). But that’s a very big if.

    But, of course, there’s a better way to reduce box blocking. Enforcement, anyone?

    1. Tolling I5 between 520 and I90 would do nicely. SDOT should be able to toll all the relevant exits

      (Actually tolling I5 to hit the through traffic would require action from WSDOT; I’m optimistic we’ll get a toll when the time comes to rebuild I5; looking at 99 and 520, the legislature is happy to toll a road if that means they don’t have to spend money up front for a megaproject)

      1. You can’t toll I-90 without adding capacity to it. It’s one of the nine primary north-south Interstates, and if any Interstate is embargoed from congestion tolling, it would be the primary north-south and primary east-west ones.

        You CAN however toll the exits at the point ownership passes from WSDOT to City of Seattle. Or put tolling equipment every few blocks on every street and let software levy timed occupancy tolls rather a dailypass.

      2. @Richard Bullington: I-90 is north-south? And I-90 *is* tolled in IL, IN, OH, NY, and MA.

      3. FYI, I-90 in Washington wasn’t built from scratch either. It was previously US Highway 10. The bridge was built in the 1940s and the highway terminated at either Rainier Ave or Dearborn Street. The part east of Lake Washington may have been an even earlier highway; it follows an Indian trail. The highway was renamed in the 1950s but the Interstate rebuild didn’t occur until the 1980s. I don’t know whether the eastern Washington alignment has changed or not.

      4. Ted Williams Tunnel is tolled, I-90, and not grandfathered. There is nothing special about the I-x0 or I-x5 interstates with regard to whether they are allowed to be tolled.

    2. I’d say let’s toll all the onramps and/or offramps in Seattle (heavier tolls in the more congested areas) and use the money to build more light rail/streetcars/I-5 lids/etc. We don’t need wsdot’a help to manage our City street congestion.

      1. WSDOT controls the on/off ramps too, as well as a certain amount of area around them (or at least they must agree to any changes there, not sure exactly how the ownership is worded). So SDOT could not do this without WSDOT agreeing.

  5. How about a trade? Let the monorail back into the public transit network, and let the Seattle Center administer tolls on Mercer?

  6. Politically, it makes no sense whatsoever for SDOT to make it a lot harder to cross Mercer, while making Mercer slightly faster for out-of-town commuters to travel.

  7. Cheap, easy, and cost-effective-cubed: Reserved streetcar lane from Fred Hutch to Fairview, signal pre-empts onto and off of the lake-shore.

    While everybody’s working on the savage rest of the Mercer problem: Just take a couple of thirty second phone calls and do it.

    Mark Dublin

  8. 13 million to upgrade software for 32 signals. That seems excessive unless it included the software and infrastructure to upgrade a larger area. Minimal advantage to one group countered by disadvantage to all others does not seem like a project worth expanding. One hopes that opening up more through streets after the tunnel opens will help more than this. Unless that attracts even more cars…

    1. If it shortens the signal cycle it might be helping everyone out, especially the north-south travellers

    2. A six-lane boulevard necessarily has long signals, and Mercer is the largest street in the area alongside Denny Way and Aurora. It’s impossible to coordinate the signals perfectly for all north-south and east-west trips, because if you give priority to one direction it intrinsically degrades the cross direction. Since Mercer is the largest street and a freeway entrance and has a ton of cars that will gridlock the area if the signals aren’t robust for them, it’s going to get priority, end of story. The issue of transit not getting priority is because there’s no transit on Mercer; otherwise it could enjoy the priority. But Metro wants to put transit on Thomas, and urbanists think that’s a more wakable and pleasant option than Mercer. So there you go. It may be possible to improve the timings a bit but don’t expect a perfect situation. I have walked across Mercer several times the past three months from Eastlake to Dexter (north across Mercer at Fairview, west across 9th and Westlake). The timings are long but not unusual for a six-lane boulevard.

  9. The main problem with how Mercer has always worked, and this timing has now exacerbated is that the people walking and bicycling, particularly N/S but also E/W due to the beg buttons, get stuck waiting minutes. During heavy traffic (e.g. PM peak when EB car traffic is mostly sitting there not moving), people walking and bicycling regularly cross Mercer against the light. Most of the time they are fine but with eleventy billions lanes to cross this is just asking for someone to get killed.

    SDOT knows this. They know that when they horribly prioritize cars at the expense of people on foot, people ignore the signals. This is especially true when the cars are not moving, yet sitting there with a green for 4 minutes, but is also true when there is very little traffic for 4 minutes and people run across between large gaps.

    It is negligent for SDOT to know this will happen and is happening and do nothing about it. In fact they are making it worse/more likely it will happen. They are purposely trading danger of people walking for a few seconds of convenience for people driving (and spending millions of our tax money to do it). Their justification is that when people walking do what they fully expect they will do and someone is injured/killed, SDOT will just say “they were crossing against the light, it’s their fault they got killed”.

    That is not Vision Zero. Vision Zero designs for safety of all users and doesn’t let Kubly put his hands over his ears/eyes and chant “I can’t hear you, I can’t see you. La la la la la la … We put up signals, if people ignore them they deserve to die!”

    1. For many walkers – including me – a pedestrian overpass would be a blessing. There doesn’t have to be one at every intersection over Mercer at SLU but one would be great. It would make walking safer and driving more efficient..

      1. I tend to agree with this, at 1 or 2 strategic locations-perhaps midblock? I don’t know that a ped overpass brings down the neighborhood any more than the whole Mercer mess itself already does. (You’d have to remove the expressway access to *fix* it, which is probably not gonna happen). There is (kind of/sort of) one here that doesn’t look too bad: https://www.google.com/maps/@47.6245473,-122.3499122,3a,75y,285.53h,96.77t/data=!3m7!1e1!3m5!1sR1IDTZXHa6c2CCqu-mSgpQ!2e0!6s%2F%2Fgeo2.ggpht.com%2Fcbk%3Fpanoid%3DR1IDTZXHa6c2CCqu-mSgpQ%26output%3Dthumbnail%26cb_client%3Dmaps_sv.tactile.gps%26thumb%3D2%26w%3D203%26h%3D100%26yaw%3D112.53503%26pitch%3D0%26thumbfov%3D100!7i13312!8i6656!6m1!1e1

  10. Why would someone put themselves through the torture of that daily struggle? It would turn me into a raging idiot. Is it as simplistic as people want to park their car within a 30 second walk of their office chair? Because, it’s only a 10-15 minute walk from Westlake to the heart of SLU, or a simple Monorail trip to Lower QA and vicinity, and I’m assuming that’s where most people are going (and getting to Westlake is way easier and insulated from the Mercer Mess).

  11. I think that one core problem here is unbridled upzoning in South Lake Union adding tens of thousands of jobs and residents. All of the traffic planning done for the area was done ten years ago, and the impact of all these new jobs and residents was not fully included. I even believe that the forecasts were “frozen” for fear of legal challenges about the AWV replacement tunnel. Now that the tunnel opening is assured, it’s time for Seattle to analyze and deal with this unanticipated increase in traffic and transit demand. The begins by first revising and publicly disclosing what forecast traffic volumes are projected to be in 2020 and 2030 and 2040.

    One upcoming solution: We are close to having the reopening of Harrison, Thomas and John Streets. This will create entirely new circulation patterns. Mercer and Denny will be less important for traffic once these open. Again, SDOT needs to present a new to show what happens to traffic before additional strategies can be analyzed like reallocating how these three east-west streets are used..

    A final solution is financial: a special assessment is needed for new South Lake Union buildings. That assessment should be partially paying for this new transit tunnel planned in the area as these owners will be getting direct benefit. As it stands now, the ST residents are gifting a new transit tunnel for South Lake Union in 2035 while developers make hundreds of millions without forking over any capital transit improvements to Seattle. With more localized capital funding for the area, a variety of additional actions can be created such as pedestrian grade-separated connections using the Houston method (under the street at the basement level), the Vegas method (overpasses to handle the crowds) or the Chicago method (raise the traffic lanes 10 feet and have an at-grade pedestrian undercrossing). Parking can be removed for transit lanes. More frequent streetcars can be bought and added. A new link station can begin designed that has a regional commuter transit hub including layover space for buses at midday.

    A final measure would be to put a cap on the number of parking spaces in these new buildings. If the spaces are there, the traffic will increase! If building owners want more parking spaces in the buildings than a cap permits, then they should be kicking in more money to pay for mitigations to the increased traffic that these spaces create.

    1. SLU property owners pay taxes like the rest of us … and they pay more, because the buildings are generally new, very large, and on expensive land. And doesn’t all new construction pay TIM fees that cover transportation impacts?

      I agree on capping the addition of new parking spaces. No reason to add parking if the street grid can’t handle the traffic. And the rebuilding of the street grid with Harrison, Thomas and John should make the grid more resilient. I think Metro is looking at moving the 8 off of Denny on to one of these streets to get it out of SOV traffic – might be a good opportunity to give one of these streets the “3rd Ave treatment” where private vehicles are prevent from driving more than 1 block on the street.

      Pedestrian grade-separated connections seems wildly unnecessary. We don’t need it downtown, and we don’t need it in SLU.

    2. ” the ST residents are gifting a new transit tunnel for South Lake Union in 2035 while developers make hundreds of millions without forking over any capital transit improvements to Seattle.”

      ST residents are paying for transit infrastructure improvements for themselves. They’re not “gifting” developers anything. Chicago did not built its El and Metra network for developers; it did it so its residents could get around, you know, like people do in Europe. How much the developers make is irrelevant. The jobs and housing they’re building benefit the region.

      1. Most of the Chicago “L” network existed before CTA was created. Almost all of the tracks for Metra were also in existence before it was taken over by RTA. It was built by private interests at a time when building and running transit was profitable.

        I don’t disagree that dense jobs and housing are good. I do think that unplanned, unanticipated dense jobs and housing is not so go, though. It’s going to cost us more dearly in the 2020’s and 2030’s to build this tunnel when the right kind of transit planning could have been preparing for this line before SLU took off. In fact, the City of Seattle forced the current ST3 alignment because no one thought of it when the Ballard-Downtown study was done a few years ago! Seattle just does not want to think ahead!

      2. Al is right – Chicago’s L history is similar to Seattle’ streetcars, with most of the networks privately built by people interested in making money in either real estate or electricity.

        And sure it would be nice to build transit before there is growth, and maybe we’ll get lucky and UK District will wait to boom until after Link opens. But complaining that an economic boom is inconveniently placed is a rich man’s problem

  12. This article didn’t address the fact that now pedestrian and bike lights go red up to 30 whole seconds before parallel car traffic. I wrote SDOT on this subject and their response was it was in testing phase. A hilarious example is where Taylor meets Mercer. This light goes red for pedestrians a good 20 seconds before the bike light–we all cross anyway while drivers honk and honk. Plus, now lights won’t change for pedestrians without hitting the beg button in one of the most walked neighborhoods.

    SDOT has shown that cars>people even in an urban village. of Lower Queen Anne/SLU.

    1. At Terry and Valley/Roy, I just missed the light to cross Valley. So I hit the pedestrian cross button and waited. The pedestrian signal along Valley started flashing red, then turned green for a bit, then begun flashing red a second time, before turning green again for a bit. Finally, the third countdown actually resulted in a red traffic light for Valley and allowed me to cross. I did not time it, but it was well over a minute. And this is for a two lane side street, which SDOT apparently wants to become a high speed bypass of Mercer, to and from Fairview.

      Dori Monson’s “War on Cars” continues, with non car drivers still losing immensely!

      1. I would love to strand Dori on that Mercer median, waiting eternally for the light to change but then just missing it every time. Like Boston’s fabled Charlie on the MTA, he can be “the man who never returned.” ;-)

        “He may wait forever in the midst of Mercer/ he’s the man who never returned.”

  13. I think Mercer has reached that third world point where just getting rid of the lights, signals, crosswalks, etc. and putting in uncontrolled roundabouts with random pedestrians scurrying through the traffic would flow better.

  14. Things should become increasingly interesting when 3,000+ Expedia employees start commuting to/from Interbay. I think we’ll need another tunnel…

    1. I’m not an expert on this, off the top of my head… 1.4 gallons of gas at ~6 lbs/gal is about 8.4 pounds of gasoline… gasoline mostly supplies carbon to CO2, with the oxygen supplied by the air. Adding in the mass of all the oxygen in a CO2 molecule gets me to 30.8 pounds. So that indeed doesn’t seem quite right.

  15. One simple fix would be to have the north-south signals on Dexter precede the left turn signals onto Mercer. As it is, eastbound Mercer is nearly always at gridlock when the signal changes, meaning all the left turning traffic either follows the law (right!) and doesn’t turn or blocks the box (and the bike lane and bus lane on Dexter). This has never made sense since they changed the signal to its current configuration, including the bike signal. Since right turning traffic has to await the bike signal going red, this would allow the eastbound traffic to clear the intersection, leaving room for both right and left turning traffic off Dexter before the next green for Mercer.

  16. Lets create some word associations: transit with arterials, density with crosswalks away from arterials. This is a real case to throw some buses at the pedestrian problem. Start as far south as Whole Foods (or further), stop on the 1/4 mile crossing Mercer out to the shore businesses and return. Force ties into the new BRT on Eastlake. How about an overhead pedestrian ramp with a moving sidewalk, sheltered from weather. The goal isn’t to move pedestrians out and away, but to give them a reason to not play dodgeball in the traffic.

    Time the changes with the new tunnel opening and consider retiring the SLUT. Its done its job focusing density along the corridor. We need buses with variable routes and getting rid of the rails helps the bikers.

Comments are closed.