Seattle, WA: Seattle Link light rail at Rainier Beach station
Link Light Rail at Rainier Beach Station (photo by nabobswims)

Sound Transit recently revealed that as of September 19, 2020, Link Light Rail will run every 15 minutes during the day on weekdays and weekends, and every 8 minutes at peak. Link will still drop down to 30 minute headways in the evenings. This will be the first time since early April that Link will be running frequent service, as well as the first time since early January that Link will be more frequent at peak than Connect 2020 frequencies. While not as frequent as “normal” service, the restoration of frequent service is a welcome development. In recent months, restoration of Link service has been well behind that of Sound Transit and King County metro bus service, with many major corridors getting frequent bus service while Link still lacks frequent service at all (even at peak). This has been particularly bad for Kirkland riders, as back in March, Metro restructured route 255 to end at UW, with the expectation of frequent Link service to pick up riders headed to downtown.

With the drop in peak-hour ridership likely to continue long term as commuters adjust to the new normal of COVID-19 and offices embrace working from home, it seems unlikely that Link will return to 6-minute headways before East Link opens in 2023. When that happens, both lines will run every 8 minutes individually (barring unforeseen circumstances, as COVID-19 has shown), doubling the frequency north from downtown Seattle to every 4 minutes.

Sound Transit will also bring Sounder South to nine daily round trips, up from seven, while Sounder North will remain at two daily round trips. Sound Transit Express bus service will also be adjusted, with route 512 being restored to full service along with more minor changes to other routes. Routes 541, 544, 555, 556, and 567, which are currently suspended, will remain suspended as of September. Sound Transit stresses that service is subject to change even before the following planned service change in March 2021, similar to how it’s been for the past few months. Because of this, Sound Transit will not be providing printed route schedules for September 2020 (but will likely keep updating PDF schedules of its routes on the reduced service page).

52 Replies to “Link to run every 15 minutes starting September 19”

  1. I’m not seeing a definition anywhere for when daytime frequencies end and evening frequencies begin? Is the cutoff 7 PM? 8 PM? 9 PM? And is it earlier on weekends vs. weekdays?

    For people that need to travel around 7 PM’ish, details like this make a big difference.

    1. All will be PDFed. Pay no heed to those timetables at the stations that will never go into effect.

    2. Maybe we can put up some guerrilla signs. We should have thought of it sooner. Seven months delay is unacceptable. There are sixteen stations, so maybe forty schedule signs total.

      Waiting until the September service change to increase frequency seems misguided. Does ST think there will suddenly be more demand on September 21 than August 21? If not, then it still doesn’t fully understand the hardships people are facing now and the deterrent effect low frequency has on ridership. I appreciate ST may need a month’s lead time to rehire drivers and arrange logistics. But it didn’t wait for a biannual service change to reduce service, and the current level isn’t a regular service period, it’s a one-off reduction below it. So there could be a one-off increase to restore it, without the planets changing their orbits in indignation over a nonseasonal adjustment.

      1. It’s 15 northbound station schedules signs and 15 southbound station schedule signs. Maybe someone could volunteer their services to typeset each of them?

    3. The links have an asterisk defining “peak” — “Peak service runs approximately between 6-8:30 a.m and 3-6:20 p.m.”

      I’m not sure if that’s when the train starts from the first station or if it’s more generalized. I would agree that it should run a bit later but the time-of-day loads may be being monitored.

      I can understand how simple it is to add peak trains onto a base schedule (base-peak-base-peak-etc). Still, I’m surprised that it isn’t a 12-minute base.

    4. Given that the current schedule is 30 minutes weekday after 9pm, and this service change is generally an improvement across the board, I think it’s safe to assume that 30 minutes night service won’t start until 9pm or later. My money would be on 10 or 11pm.

      1. Considering that nightlife is pretty much closed right now and that very few people are flying, that’s probably ok for the short term. If anything, I would expect to see fewer people out and about at 8 PM, once the fall season kicks in and 8 PM isn’t daylight anymore.

        I do hope, thought, that the pre-COVID schedule gets completely restored once the vaccine comes out. Back when I was flying, I often chose return flights that would land at SeaTac around 10 PM on a Sunday night so that I could spend as much of the weekend as possible away, while still getting a full day’s work in on Monday. Coming home from these flights, 15-minute Link service at 11 PM is much appreciated. Getting picked up in any kind of car or shuttle at this time is a horrible mess, and with a 30-minute wait at the airport, the travel-time advantage of Uber/Lyft over public transit all but vanishes for much of the city (assuming, of course, that the public transit is running with decent frequency).

    5. I asked this on Twitter, and they said that Link will run every 15 minutes until 10 PM, which is when “evening” service starts.

  2. Metro went to a high frequency model first for two very important reasons both related to social distancing: 1) Metro lacks the operational flexibility of Link, and 2) Metro lacks the protected operator environment of Link.

    The first reason is pretty obvious. Link can add space for social distancing at any given frequency simply by adding LRVs to each consist. Metro, with its fixed length buses can’t do this. To add social distancing space they can only add coaches, thus increasing frequency (and operators!) by default.

    The second reason is much the same. Because Metro doesn’t offer operator protections like Link, they are required to drop fares and go to middle-door boarding. While this “solves” the problem of operator protection (sort of), it increases ridership because of all the non-transit based ridership it generates. Fully half the ridership on the mid-day buses I have ridden lately have been riding for on-transit reasons.

    So, ya, frequency is great. But Not all frequency is created equal.

    1. I wonder how protected the air in the driver compartment of Link really is. If the ventilation system is circulating the same air between the driver compartment and the passenger compartment, any viruses on one side of the partition will easily spread to the other.

    2. Metro lacks the protected operator environment of Link.

      Right. Meanwhile, ST continues to run infrequently, which means more people ride on Metro buses instead of Link trains. This puts the Metro drivers at greater risk. Therefore, ST is endangering Metro drivers with its low frequency policy.

      1. Ha! How many layers of inefficiency are you willing to heap on the local transportation system just to cover for Metro’s shortcomings?

        No. ST should not be required to cover for Metro by running at a higher frequency than required. Instead the root cause of this problem should be addressed. Metro should add driver protection and start charging fares. Then they should address frequency to match true, transportation related demand.

        And it should be noted that operator protection should be added for more reasons than just CV-19. How long is Metro willing to keep our operator’s safety at risk?

      2. How many layers of inefficiency are you willing to heap on the local transportation system just to cover for Metro’s shortcomings?

        What are you talking about? Metro has barely cut service. There are buses running in the middle of the day that run more frequent than Link — even after this improvement! Holy cow, think about it. A subway train — the core of our transportation system — and they can’t bother to run it as often as the 70, let alone the 7, or the 3/4.

        But you want to blame Metro?

        Metro is not running the buses often because it makes things safer. They are running them more often because it makes things better. Good God, you’ve been commenting on this blog for how long? Yet you don’t understand the benefit of higher frequencies? Here is a primer: https://humantransit.org/basics/the-transit-ridership-recipe#frequency.

        It is obvious that it is ST that is coming up short, while Metro is at least attempting a decent level of service.

      3. ST should not be required to cover for Metro by running at a higher frequency than required. Instead the root cause of this problem should be addressed. Metro should add driver protection and start charging fares. Then they should address frequency to match true, transportation related demand.

        Metro already has the most important elements of protecting drivers: Social distancing and a mask requirement. Restoring fares will dramatically reduce social distancing from drivers. A shield will not do nearly as much to protect drivers.

        The best next thing Metro can do to protect drivers is to provide free masks and educate riders on how and when to wear them. That is in the works. Getting masks on “non-transit-related” riders is just as important in fighting the pandemic as getting them on “destinational” riders. Everyone needs to be wearing them when around other people in public. (I have no idea how you tell the difference between “non-transit-related” and “destinational” riders, but I recall ST was supposed to be starting up a program for free passes for various no-income riders, so fares would still not be abused as a way to get the “wrong” people off the *public* bus or train.)

        Is ST pitching in on that effort to help protect ST riders? Does making the schedule hard to find and having riders stand around with others who might not have their masks up for an average of 15 minutes before the train arrives a safe practice? How much is ST saving by not posting the real schedules? At least Metro posts their schedules at moderately-used stops, and provides printed schedules, so people can minimize the time they stand around waiting with strangers.

    3. It is also worth noting that ST dropped frequency on many of its buses, so it isn’t all about driver safety. It is about money.

      1. What ST did to Link’s frequency is disproportional to its bus frequency. Link is a core all-day service like the 512, 522, 550, 554, and 594. The suspended routes and runs are jettisoning some extra peak service, not core mobility. There may be a couple suspended evening trips here and there, but it’s not a 50% reduction in midday service, 66% in early evening and weekend service, and 50% in late evening service. You can still take the 512 and 550 every 15 minutes on weekdays and Saturdays, with only a few gaps.

        Looking at 8am to 10pm, weekdays on the 550 eastbound there are 20-29 minute gap at 6:33pm and 7:06pm. Westbound 20-minute gaps at 9:04am and 4:48pm, and a whopping 31-minute gap at 5:58pm. I think those gaps are because ST just crossed out runs without adjusting the ones around them. Weekend service has full frequency: 15 minutes Saturday, 30 minutes Sunday and evenings.

        The 512 weekdays southbound has a 20-minute gap at 9:20am, and 30-minute gaps at 10:15am, 10:45am, 3:00pm, 6:00pm, 40 minutes at 7:20pm, and 60 minutes at 9:30pm. Northbound has 30-minute gaps at 8:53am, 12:53pm, (route 510) 28 minutes at 3:05pm and 3:28pm, 23 minutes at 7:20pm, 40 minutes at 8:28pm. Northbound has 30-minute gaps at 8:23am, 12:53pm, (route 510) 23-27 minutes at 3:05pm and 3:55pm, 23 minutes at 7:20pm, and 20 minutes at 8:28pm. The 511 may fill in some of these. Again the oddness of some of the gaps seems to be because ST didn’t adjust the schedules around the suspended runs.

        If Link had reductions like that people wouldn’t be complaining as much. But Link has gone to a barely-tolerable 20-minute weekdays and an intolerable 30-minute weekends/evenings. And it’s rapid transit, which is supposed to be frequent as a circulator. That’s why people are screaming.

        I never thought I’d be so happy to see 15-minute service on Link. Compared to current service it’s a godsend. You don’t have to worry about a schedule or fit your life around its pulses, you just arrive when you’re ready and expect a 10-15 minute wait at worst.

        It gets worse with 2-seat rides, which might double your wait time, but even that’s better than a 15-minute bus transferring to a 30-minute Link. When I came back from Seward Park and Pritchard Park, I took the 50 northbound but did not transfer to Link because I would have had a 20-25 minute wait. Instead I transferred to the 7 and hoped it wouldn’t be overcrowded. I might have transferred to the 106 but I couldn’t remember if it was 15 or 30 minute Sundays.

      2. I didn’t get all the times accurate because it’s hard to transcribe them from the schedules and not jump to the northbound section without realizing it, but they still show the quantity of service and which hours have the most common gaps.

      3. amen. ST is being unnecessarily cheap. we have spent billions on Link; we should make it sing with frequency.

  3. With the drop in peak-hour ridership likely to continue…

    … makes this decision all the more odd. This is mainly an improvement in peak service. 8 minutes is decent — some might even call it frequent. But 15 minutes still sucks. Why not just run every 10 minutes all day? Going from 10 minutes to 8 minutes is an improvement, but not nearly as big an improvement as going from 15 to 10. This seems geared towards avoiding rush hour crowding that simply doesn’t exist. During the middle of the day, there will continue to be lots of buses that run more frequently than Link.

    This means that Kirkland riders heading downtown are still out of luck. Correct me if I’m wrong, but the old bus ran every half hour. Now you have a pair of vehicles running every 15 minutes. That is worse.

    A year ago, if I was downtown trying to get to the U-District, I would just take Link. Even if had to then take another bus, it would pay off, since the train is so frequent and so fast. Now I would check my phone, to see whether the train is supposed to be here soon. I would be scrambling, wondering if the trip down to the platform (and back up again) followed by the transfer would be worth it. If I have to wait longer, then I would just take the 70 or49 both of which are more frequent (and get me to more places in the U-District).

    This is a step in the right direction by ST, but a very small step, and one that is still geared towards a world that doesn’t exist now (a bunch of people crowded into the train at rush hour). At a minimum they should have 10 minutes service all day long.

    1. I agree that 8 minutes peak is probably overkill right now, but I suspect Sound Transit is thinking that ridership is slowly returning on a continuous basis, so you have to think about what the peak crowds might be in February, as opposed to today.

      As to the 255, 15-minute frequency on the train now matches the bus, which should make it theoretically possible to adjust the schedules for at least the eastbound trips to make the wait time for the train->bus connection negligible. This is something that, with a 10-minute train and a 15-minute bus, you simply cannot do.

      Of course, different people exit the station at different speeds, but this seems manageable. The escalators all move at a fixed speed, so you can conservatively assume the passenger to be standing on the escalators all the way up, and if somebody decides to walk up – or take the elevator – they’ll just have to wait a couple of minutes. A radio link between the bus driver and security staff can ensure that the bus driver knows exactly when the incoming train arrives, so they can wait in the unlikely event that the train is late.

      Of course, the flip side to all this is that some of the other bus connections (e.g. 372, 65) will be running every 20 minutes during some of the weekend periods when Link is running every 15 minutes. Connecting between a 15-minute service and a 20-minute service creates a phase problem where, no matter how you juggle the schedules, only one trip per hour can offer a connection with a wait time under 5 minutes.

      I suppose if Metro is going to be making bus service cuts this September anyway, they could run such routes every 30 minutes during the periods where they currently run every 20 and, timed properly, may actually improve Link connections. But, getting a good connection for both northbound and southbound riders might prove difficult, and such a move also comes at the cost of making the bus service harder to use for people who are just getting around north Seattle on the bus and aren’t riding Link.

      Ultimately, the only good solution to all this is, as you say, to run Link more frequently. But, 15 minutes is better than 30, and is at least a place to start.

      1. I think it will be very hard to time the bus to train combination, just because the trip up from the platform varies person to person, as you mentioned. If you time if for an average person, then half your riders have a long wait (up to 14 minutes). If you time it for the slowest walker, then the total trip is really long — much longer than the old bus. I like to walk up the escalators, and I walk fast — it would seem like a bad random connection to me. Then there is the possibility of the bridge going up and making that trip even longer. My initial trip on the train is challenging as well. I would try and time it, just like I would time the old half hour bus ride. So the trip would be slower, but I have more options for timing it — it just sounds worse to me.

        Nevertheless, I wouldn’t reverse things. It would be a bad idea to go back and forth. The increased frequency helps people who are traveling within Kirkland, and it helps those going to the UW. But it was designed for a frequent Link, not one that runs every 15 minutes.

      2. My thought is that even the slowest person will still move upward at least the speed of the escalator. So, you can time how long it takes to each the upper level standing on the escalator, without walking, then assume 2-3 mph for the remaining portions of the walk. The few people that need more time than this walking the flats can make it up by taking the elevator up, rather than the escalator.

        I think the travel-time difference between walking up the escalator and standing on the escalator all the way is about 2-3 minutes, which not a terrible amount of time for a faster walker to spend waiting on the bus.

        Not perfect, and definitely not as good as simply running Link more frequently. And, I don’t think there’s any way to time the inbound connection, with traffic being too unpredictable. But, it’s better than just picking schedules out of a hat and ending in a world where every 15 minutes, the bus leaves right as the train is approaching, leaving everyone with a 10-13 minute wait for the next bus.

    2. The 255 was running direct to downtown every 15 minutes or better between about 5:30am and 7:30pm on weekdays and every 30 minutes evenings and weekend daytimes. Peak of peak it was about every 7 minutes.

      That was way better than two 15 minute frequency services with a cumbersome transfer.

      And a week ago when SDOT was doing repairs the Montlake bridge, the management that determined that the transfer to Link was the way for Kirkland riders to access downtown ran the 255 up over the University bridge and terminated it at 15th Ave NE, about 0.6 miles away from the Link station. They could have just as easily terminated it at 5th & Pine at the Link station. The only reason no one cares is that no one is riding.

      Let’s see what they do next weekend when the entire 520 bridge is closed for construction. My guess is that anyone who wants to go from Kirkland to Seattle will be expected to ride across I-90 and I-5 past downtown all the way to the U-District. Maybe they will again terminate it 15 minutes from any Link station. It may be faster to ride ST-535 up to Bothell (does it even run on weekends any more?) and then on ST-522 which still goes all the way downtown.

      It seems pretty pointless to run the 255 at the frequency they are running when it is pretty useless. It is also pretty customer insensitive that if they want to reduce coaches on the weekend, instead of running every 20 minutes, they eliminate random runs so that 3 coaches come with 15 minute gaps and then there is a 30 minutes gap.

      It’s crazy that they do not have the ability to make more customer friendly changes for frequency reductions and service disruptions. We know that there will be periodic Montlake disruptions for the next several years due to construction. A hugely more sensible service pattern would be to serve a downtown Link station whenever there is a construction disruption on 520 or the Montlake bridge or Husky stadium.

      1. I agree, in the short term, the 255 reroute is rediculous. Fortunately, in another year or so, they’ll be able to send the 255 to the U-district station at Brooklyn/43rd anytime Husky Stadium station is inaccessible.

    3. Normal peak is 8 minutes. 6 minutes was a temporary boost that started when the Ride Free Area ended or thereabouts and was intended to end with U-Link or Northgate Link, and it would revert to 8 minutes until Lynnwood Link. The fact that it has sometimes continue running at 6 minutes is extra; ST never promised to keep it. So ST has a reasonable argument to reduce peak service to 8 minutes peak hours if trains aren’t full; it just doesn’t have have an excuse to run less than every 10-15 minutes off-peak because that damages people’s mobility and the purpose of having Link in the first place.

      1. So ST has a reasonable argument to reduce peak service to 8 minutes peak hours

        But the argument should be rooted in science, with feedback from the public, not an arbitrary number. We know that running less often saves money. How much money? We know that running more often adds ridership. How much ridership? (I can make a guess if ST is to lazy to look at the literature). I also know how much riders pay (on average), which is around two bucks per rider. Then we can compare the net cost, and decide, as a representative democracy, whether it is worth it or not. My guess is it is. My guess is that the net cost is minimal, and one of the more cost effective improvements that can be made.

  4. Reason I keep asking for some comments by people with first-hand knowledge about Link, both “Operations” and “Office.”

    Considering the advantage offered by Link’s reserved and signal-preempted right of way, in addition to its additional on-board space and built-in driver protection, I can’t see what ST would have to gain by keeping service below capacity.

    From the now daily-failures in everything from wi-fi to paper-trash pick-up at my complex, I have no problem believing that most of what ST isn’t doing is because somebody else is still waiting for some third party to get their COMCAST back.

    It’s also from daily personal experience that I’m stressing things that, individually and together, if modestly led and encouraged, operating personnel can figure out and fix for themselves. But just for starters…..

    What’s the truth about train speed and signal priority on MLK between Mt. Baker Station and Rainier Beach? Did I hear they were running single-track for awhile? STB Portland Division: what’s MAX top speed in comparable conditions?

    Promise I’m not teasing Dow about this, but undercuts provide employment too! For crews whom Link will always offer a car-free ride to work.

    Mark Dublin

  5. Pour one out for 6-minute headways. Little did we know that January 3, 2020 would be the last ever day that stations from Stadium southward would ever see them. I think Alex is right about the 8-minute headway on the ‘1 Line’ continuing indefinitely, combining with the 2 Line in 2023 to form a 4-minute trunk from Int’l District to Lynnwood. On the plus side, I think we’re probably also done forever with the 2-car trains of yore.

    1. I think there is an outside chance that going to 6 minutes is possible between Northgate Link opening and East Link (Line 2) opening — but only if demand warranted it. Still, all of the ST2-related environmental studies assumed 4/8 minutes as you describe, and that’s what was used to plan all the system basics like vehicle purchases and facility design.

      1. I think there is an outside chance that going to 6 minutes is possible between Northgate Link opening and East Link (Line 2) opening

        Yeah, I could see that. Maybe by summer of 2022, things are back to normal, more or less. There has been widespread vaccination, and Covid is under control in the U. S., sitting somewhere between Polio and Measles. Running the trains every six minutes from Northgate seems like it would be barely adequate, especially with the various truncations. Metro (and Community Transit) will be loathe to repeat the 255 debacle — they won’t truncate unless ST commits to high frequency.

    2. It is really hard to say. ST has always been chintzy with frequency. Like so many things, they don’t seem to understand the importance of it. So it is quite possible that you are right — ST won’t invest in more frequency, even when there is decent demand, and obvious ridership gains to be had. By the way, this is why I also think it is crazy to assume that trains will run frequently all day to Fife; if ST is too cheap to run trains through the core frequently, it is hard to see them running them in areas where the ridership per service time is much, much worse. But I digress.

      I think it is possible that they will run trains during peak every six minutes, just because they did that before. Of course they did that in large part to deal with capacity issues, and four car trains can carry a lot of people. It is unlikely there will be much crowding on the south line, but the east line could involve a little. Not to the point of leaving people on the platform (that never seemed likely) but just a lot of crowding. I could see some complaints by wealthy people, which could lead to more trains during rush hour.

      1. Let’s be real. ST is not going to Fife in the lifetime of anyone now old enough to be interested in this blog. Its economic plan is destroyed and any possible remediation will die outside Seattle.

        Federal Way will be the southern terminus for fifty years, if indeed it even gets there.

      2. To be clear, Seattle may pay ST to build and run something through SLU, but that will be the only extension beyond Lynnwood-Federal Way-Redmond that will be built; “ST3” will be stillborn.

        And of course, as many people have argued, that is enough.

      3. “ST has always been chintzy with frequency.”

        In my (admittedly limited) experience, Sound Transit’s commitment to Link frequency all the time has been superb (until 2020). It’s always easy to say that it should be more frequent. But I think about recent trips to San Francisco and Washington DC.

        I went to take the blue line in DC once on a weekend. The next one was in 17 minutes. It turns out that it just runs every 20 minutes on weekends, pretty much all the time. It made me really happy that Link (normally) ran every 10 minutes or better nearly all the time (even on a Sunday at 9:30pm), and only after 10pm did it drop down to 15 minutes (and ran every 20 minutes in the earliest hour of the morning, that’s it).

        I was flying back from Oakland on a Sunday only to realize that the start of service made the connection to the airport closer than I would have liked. I had taken for granted the fact that Link ran on Sundays starting at 5 or 6am depending on where you start, and while running at 12 minute headways for much of the morning, that is double the max frequency of BART all Sunday. BART doesn’t start running until 7:40ish to 8 on Sunday, and only runs every 24 minutes. Quite the disappointment, and a wasteful way to operate a built-out subway system.

        The fact that Link has been a mostly dependable 10-minute system is something ST should be praised for, especially given Seattle’s relative population and how recently its even had a subway. Sound Transit was beating out many larger cities that have had larger subway systems for decades. That’s part of what’s so disappointing about the slow return of Link frequency. It gave me the uncomfortable DC and BART vibes that I never associated with Sound Transit before. But with the beginnings of returning frequency, I am more hopeful that as things recover, dependable 10 minute frequency will return in time.

      4. “Sound Transit’s commitment to Link frequency all the time has been superb (until 2020).”

        Yes, it has been! Link faithfully kept to 10 minutes until 10pm like real subways do, and that made it easier to get around than other American light rails and BART that run every 15 minutes on their lowest-frequency branches and sometimes even drop to 30 minutes.That’s what’s been so alarming in the current service: ST said it was cutting because of vandalism when there were no fares, but then fares came back and the frequency didn’t, and there was no explanation as to why or what the criteria would be to restore it, as if the board had lost its commitment to frequency.

      5. @Alex (and Mike) — I meant chintzy from a standard (international) practices standpoint. I didn’t mean that they would be worse than other cities in the United States.

        I read an interesting article a while back. It was a comparison of health care systems in the industrialized world. In there, they talked about the conventions that are held where experts in the field compare different models, and different ideas. Apparently, when in the middle of an argument, if you point out the U. S. does things that way, it is strong argument for doing things differently. We are that bad when it comes to national health insurance. As it turns out, the same is true for public transit.

        Running the trains every ten minutes is not “frequent”, it is barely adequate. Most cities (around the world) know that. The experts know that — there are numerous studies to support better frequency. Spending billions on trains to Tukwila or Angle Lake is dubious. Running the trains more often is sensible.

      6. It’s a reasonable incremental step. Yes, the NYC subway, London tube, and Moscow and St Petersburg metros would never have more than 5 minute headways until after 8:30pm, but we have to start somewhere. Link has achieved 10 minutes, which is better than most American light rails. Metro’s 10-minute bus network is where its 15-minute network was twenty years ago. First you have to get all core routes up to 15 minutes, then 10 minutes, then 5 minutes. Moscow and St Petersburg have 5-minute headways on all forms of transit: metro, trams, trolleybuses, and autobuses — if it’s in service, it comes every 5 minutes, at least until the last one or two metro stations where it might drop to 20 minutes. But King County and the US are not nearly ready for that. First we need to work on a 15-minute baseline, then a 10-minute baseline, then we can work on a 5-minute baseline. Link will reach 4 minutes in the north side when East Link opens, so that will be a major step.

      7. I wouldn’t expect frequency to match that of Toronto, let alone London or Paris. But I would expect it to be similar to Vancouver. There are a lot of similarities between the two systems. Seattle and Vancouver are roughly the same size. Both cities have most of the density in the city proper (we aren’t like L. A. or Phoenix). Both rail systems are largely grade separated and thus fast. Both are North American cities (nations without the history of public transit success that is common in Europe and Asia). One of the big differences is that the Vancouver branches occur well outside the core city — where density is low. In contrast, both of our branches occurs right from downtown. One covers the south end of Seattle and the southern suburbs, while the other will cover our most important suburb (and the only place outside Seattle with significant density).

        But the biggest difference is frequency. Inside Vancouver proper, the trains run frequently. The Expo Line runs 2–3 minutes during peak, and 3–5 minutes the rest of the time. This high frequency extends all the way to Burnaby and New Westminster — suburbs well outside the urban core. The Expo Line is their “core” line — their “Central Link”, if you will; it has ridership of about 300,000 per weekday, or most of the rail ridership.

        The Millennial Line — which doesn’t split, and doesn’t go downtown — runs 3–4 minutes during peak. During the middle of the day, it runs every 6 minutes, in the evening 8 minutes, and late at night 10. While not as impressive as the Expo Line, it is still a lot better than Central Link, or East Link.

        Only the Canada Line comes close to the weak frequencies of Link. In Vancouver itself it is still better, with 3 minutes at peak, 4 minutes midday, 6 minutes in the evening and 10 minutes late at night. But for the branches outside the city, these numbers are poor. There are technical issues with the Canada Line, and this may have something to do with it.

        When you look at the overall system, it is clear that Vancouver has much higher frequency all day long. This will continue, indefinitely. Even when Link is split, and the trains run at peak every 4 minutes to the north (and 8 minutes to the east and south) the Expo Line will be running every 2-3 minutes. More importantly, in the middle of the day every line in Vancouver is frequent — 4 minutes at worst. Late at night frequency is 4 minutes (Expo) or 10 minutes (Canada and Millennium) . In contrast, ST plans to run trains to Rainier Valley every 10 minutes in the middle of the day, and 15 after 10 PM. That is a huge difference.

        To be clear — there are physical limits on both branches (we can’t run the trains down Rainier Valley or through Bellevue every 4 minutes). But when East Link is built, both branches should run every six minutes, all day long, and transition to ten late at night. It is nuts to spend so much money on a high end system, and then sacrifice ridership — and rider convenience — by running infrequently. The extra cost for running more frequently is minimal. The benefit of running more frequently is huge. Running infrequently is just being chintzy.

      8. “I wouldn’t expect frequency to match that of Toronto, let alone London or Paris. But I would expect it to be similar to Vancouver.”

        Yes. If you’re in central New York or London you have a choice of several corridors (tracks) going several directions every 2-5 minutes midday. I would not expect this in Seattle, but I would expect a 10-minute or 5-minute minimum. 6 minutes would be close enough to 5 minutes to call it a day.

        The ultimate issue is how the agencies weigh frequency in their strategic plan. If they value it highly as they should, then you can go to a station anytime confident a train/bus will come within 5 minutes (or 10 minutes). You can travel when you’re ready rather than working around infrequent pulses. You can do more errands in a day, and transit is more competitive with driving. In contrast, if the agencies put frequency at the bottom of the list or leave it as an afterthought, then transit is inconvenient and not very useful, , and only riders with no other choice or transit mania will show up. The first vision maximizes people’s mobility, minimizes vehicle miles driven, and is best for the environment. “Mobility” means both actual rides and the availability of a potential ride. You may not ride a train/bus today at 10am, but the fact that it runs every day at 10am means it will be there on the day you need it, and somebody else will ride it today.

        ST has been impressive at keeping Link at 10 minutes until 10pm, higher than peer light rails. I want a commitment it will stick to it long term. 6 minutes would be even better but that’s a later step. For Metro, I want to see the current 10-15 minute service in Seattle, and proportionally in the suburbs. Metro’s long-range plans are a starting point to where/how much it should be. And its “Frequent” routes should be frequent evenings and weekends too.

        As I’ve visited peer cities in the 80s, 90s, and 00s, I’ve been struck by how ubiquidous frequent service is in San Francisco, Chicago, and Vancouver. Almost every route you look at runs at least every 10-15 minutes daytime, 20 minutes evenings, and there are half-hourly night owls a mile apart. The only 30-minute routes are a few coverage routes. in contrast, Metro was horrible before 2012. Only a handful of 15-minute daytime corridors, and very few after 7pm or on Sundays. (Eastlake, Rainier, Jefferson, Fremont/34th, and Ballard/Leary, Pine/Bellevue, and 3rd/Cedar or QA/Mercer are the ones I recall.) I used to try to live in one of those segments, and it was hard because the Capitol Hill, Fremont, and Ballard ones only reached the edge of the neighborhood with few housing choices within walking distance. I lived at 56th/University, which was probably the best, with full-time frequent service and night owl to downtown, Eastlake, and the Ave. (And I’d walk down from the Vogue to Fairview to catch 1:45am or 2:20am.) The only other places where you could really do that were Jefferson and Jackson/Rainier.

        Metro has gotten a lot better since 2012. But its frequent network is still not as comprehensive as San Francisco, Chicago, or Vancouver.

        Vancouver’s Skytrain is unusually frequent. That’s partly because it’s driverless, had funding from the federal government, and Bombardier offered a discount to showcase Canadian technology. We can’t replicate those here. Driverless must wait until MLK is sunk into a trench, which ST has no plans to do; federal funding is still a trickle; Boeing is unlikely to get light rail mania; the Oregon streetcar manufacturer is out of business I think; and Biden supports Buy America even though all the low-cost rail technology is outside the US and we’re not a big enough market to expand it here. Still, ST could commit to 6-minute frequency now with the trains it has.

        The branch concept doesn’t apply to Link very well. A traditional branch is a Y from the city center, like New York’s A/C/E, London’s District Line, or BART’s Daly City-East Bay fan. Link will have a reverse Y, from Lynnwood/128th to Intl Dist. That mainly benefits people in the U-District/Northgate/Lynnwood/Capitol Hill axis, not people in the central core. It makes sense because the U-District is the second downtown and Northgate is the third designated urban center, but it’s not a traditional branch benefit like inside the London circle or central Manhattan. And when the Everett-West Seattle pattern starts, they may not be branches as much as two lines overlapping in the middle. So I’m not really sure how to apply traditional branch concepts to Link.

        In Portland I used to look at where I could live with 7.5 or 5-minute service, and when there were only two or three lines you had to live between downtown and Gateway, which gives not many housing choices, the Banfield stations have limited walksheds, and it’s a mile north of the Hawthorne district or East Burnside/Stark. I guess you could live west until Beaverton; i didn’t know much about that area. These are the kind of situations I think are too limited, where sub-15 minute transit is available in only a small part of the city.

      9. “must wait until MLK is sunk into a trench”

        I think driverless cars technology has as advanced to the point where that’s no longer the case. Going driverless on a fixed route with only a few intersections in known locations is much easier than going driverless on a car that drives on arbitrary city streets

        The bigger problem is 1) no transit agency wants to be the first, for liability reasons. 2) The union would go on strike if ST ever tried to do it, and they’d be left with no bus drivers or mechanics.

    3. Arthur, what’s your “call” on the date when Sound Transit and the rest of us first realized that COVID was not just a typo that left the “R” out of the crow family?

      Rendering our every prediction our own individual and collective best guess. Which, if we know what’s good for us, we’ll make as flexibly-adjustable and reversible as possible from here on.

      From COVID’s onset, every day I’ve personally been seeing our strongest counter-force: Something, like my COMCAST, stops working? On the street, on the road, or on the phone, a person either figures out a remedy, or provides direction to my next step.

      So priority: Max equipment- and passenger-handling-training for everybody in Operations. Close second? To-the-minute passenger information. The “Ambassadors” I’ve heard mentioned- any possibility? And every rule, whatever’s easiest to obey.

      Personally, sad that years and change of residence have cost me all my inside sources. So I am wondering if STB and all the transit agencies, especially ST, can develop a solid and mutually-respected inside-info-privy “press corps” to keep readers and contributors posted on things exactly like the new railcars?

      Would save a lot of money on advertising space. Not to mention consultant’s fees. Let alone as a preventive remedy for Hell’s own most COVIDIAN evil misfortune in delays and mistakes.

      And is “meme” the right word for this calculation: “What is the cost of one minute where a bus or train that should be moving, isn’t?” Or is region-wide spray-paint graffiti the correct mode of delivery. Bright orange “$100!” should do it.

      Mark Dublin

    4. Four-car trains running every 7.5 minutes is the same capacity as three-car trains every 6 minutes. ST barely has enough LRVs to run that much capacity now, maybe. If 3-car trains are gone, so is 6-minute headway on The Segment Formerly Known as Central Link.

      1. I’m pretty sure it’ll be 3-car every 7.5 minutes, which is only 8% less capacity than the pre-pandemic peak of 2.6 cars every 6 mins.

      2. It appears that ST should have taken delivery of about 30 new Siemens light rail vehicles in the past year. Given the pandemic and the need for testing, my guess is that about 15-25 of them are ready for service. I’ve seen them both sitting in the yard in SODO as well as making test runs on MLK. Adding that to the 62 Kinkisharyo vehicles would mean that they now have 77-87 vehicles to use now. That should mean that ST has 25-40 percent more light rail vehicles available today than in early 2019.

      3. Yes, but as the Siemens cars come online, ST will be taking the Kinkisharyo’s out of commission and trucking them to Bellevue for retrofitting. The overall fleet won’t increase much until those retrofits are done. The older cars are running 2007-era tech and need a substantial upgrade prior to East Link.

  6. Forgive me if this has been discussed elsewhere, but I couldn’t find any information on the topic:

    What is the status of integrating the new Siemens railcars into service?

    First delivery was over a year ago, and announcements at that time projected their entry into service Q1 2020.

    It appears that it’s also taking a while for Valley Metro in Phx to get them up and running. Like Link, they opened with Kinkisharyos before selecting S70s for round 2.

    1. 15 minutes. See the first sentence. The previous article was unclear about weekends but now it’s been confirmed.

  7. I’m still trying to get a good “read” on this: Is Sound Transit’s problem that it doesn’t have enough railcars to run good service, of that they’ve got competence lapses that other systems have solved?

    If it’s the second one, I’d hate to see STB readership and transit ridership give up and resign ourselves to the world of sub-optimal. Again, my sense of the times including but not limited to Mike Lindblom’s employer, is that while events are taking their toll of our every system, people at every level have both the skills and the will to repair, regroup, and restart.

    So if in fact our problem is not enough train-cars….can we make any headway by running the ones we have better? Or at least start organizing to see to it that our elected officials start to do so? Starting out by finding out what they need us to do to help them get into action.

    Mark Dublin

    1. In 2019 Link ran 10 minutes with 3-car trains and 6-8 minutes with a combination of 2- and 3-car trains. It doesn’t have enough trains to run 6-8 minutes with all 3-car trains so the extra peak runs are 2-car.

      1. It appears that ST should have taken delivery of about 30 Siemens light rail vehicles in the past year. Given the pandemic and the need for testing, my guess is that about 15-25 of them are ready for service.

Comments are closed.