Sounders fans at ID/Chinatown Station, by Oran

Earlier this week I had a chance to sit down with Ron Tober, deputy chief executive of Sound Transit, to talk about Link reliability. Tober has extensive bus and rail experience – he oversaw the startup of Charlotte’s LYNX system, once headed up King County Metro, and has experience in several other cities.

On time performance has been much better in the last month than it was in the first quarter. Tober showed me more recent data – without weekend and night maintenance, trains have been on schedule about 90% of the time, and headway reliability has been well over 90%.

For now, though, that’s about all we’re going to get, and Sound Transit can’t do too much about it, largely because of the downtown tunnel.  More after the jump.

The tunnel agreement between Metro and Sound Transit was hashed out at a time when Sound Transit had very little political clout – just after it became clear that Link couldn’t be built in the original timeframe for the money available. The agency was essentially in a position to take what it could get – and what it got doesn’t provide many guarantees for service reliability. As a result, Metro can run up to 60 coaches per hour in the tunnel, even if they impact Link performance. Right now, Metro peaks at 56-57 per hour during rush hour, leaving Link waiting if a wheelchair needs to board or something goes wrong with a bus.

Tober said this agreement comes up for renegotiation in 2012, to be signed again in 2014, well before University Link opens. This is when Sound Transit will be able to work for changes – to improve Link’s performance, allowing trains in the tunnel earlier and later, and reducing the number of buses in preparation for increased service frequency. If we want all-night Friday and Saturday service, for instance, this is when to push for it. This is also when it’ll be determined how long buses remain in the tunnel – Tober says it’s possible joint operations will continue even after University Link opens, simply because there are so many more bus than train passengers, and they’d be put on congested surface streets.

In the meantime, Sound Transit is taking small steps to improve things. They’re looking at stationing wreckers (bus tow trucks) at both ends of the tunnel to clear broken down buses immediately. This summer, they’re also hiring a special incident response employee to work in the control center and ensure both riders and operators know what’s going on in case of a problem. This person would put up information on the electronic readerboards and website and trigger better audio announcements in case of a delay or closure. Supervisors at tunnel stations are also being trained soon to deal with problems on train equipment like they currently can with buses – so minor malfunctions don’t cause a delay while waiting for a technician.

The other major point of concern is overall train operation. Metro’s operators don’t have a strong incentive to be on time, and while many are fantastic, it only takes a couple of slow starts in the Rainier Valley to miss a light cycle and delay a train. While I didn’t get any specifics, it sounds like Sound Transit is looking at ways to improve this as well, and hopefully we’ll hear more in the coming months.

I also brought up concerns about the lack of realtime arrival information (actually, I wouldn’t shut up about it). Sound Transit’s train monitoring system didn’t go online as smoothly as originally hoped, and continues to be updated and worked on. As this provides data about where each train is, it could eventually lead to realtime countdowns at stations, but, of course, no promises. I stressed that these are very necessary during maintenance, when trains are running an unknown, low frequency schedule – but there’s no concrete plan.

In short, Link reliability is something Sound Transit is working on and it will improve over time, but it won’t run like clockwork until there are enough riders to justify removing buses from the downtown tunnel – potentially not until Link reaches Northgate, and high ridership buses such as the 71, 72, 73, and 41 don’t need to go downtown anymore.

I’ll be meeting with Tober again next week, so if you have questions, I’m happy to ask him then!

115 Replies to “How and When Link Reliability Will Improve”

  1. Great piece Ben, and glad to have you back. It seems the battle lines are being drawn between ST and Metro to remove the buses from the tunnel, and the solution will be a “political clout” job more than a “who’s ox gets gored” one. I can’t see where drawing more lines in the sand between Metro and ST is productive in the long run. Bus integration is essential to efficiently operate a rail based spine network. Maybe this is another good reason to merge the agencies, but I can think of many more to not merge.
    One question to ask Mr. Tober next week. Airport Link is in need of more riders to make projections and bring cost per rider down to budget estimates. Sometimes agencies make mistakes, and smart ones correct them and move on. The missing link of Link is a stop between Henderson and Hwy99, near South Center Mall, and available as a Park and Ride to intercept SOV’s off I-5 and I-405 (nearly as important as building Northgate for the northend). Have any plans been forwarded to the board to infill a P&R station at I-5/S.144th (with associated on/off ramps), or Hwy 599/S.133? If so, how soon, how much and how many riders?

    1. A few infill stations are something this blog has been proposing for some time. I believe the Hwy 599/133rd St station in Tukwila came in second after a Graham/MLK station (Boeing Access Rd and Convention Place coming in 3rd and 4th).

      I don’t think there is funding in ST1 or 2 for these stations, but I’m hopeful that there will be in ST3. If you want to give your opinions on this, email Marcus Clark at Sound Transit (

      1. There really isn’t funding.

        I don’t think there’s any problem with Airport Link ridership, though.

      2. Airport Link ridership, at least for commuters, is effectively maxed out until they do something about the parking mess at Tukwila. The P&R fills up by 9, and there aren’t anywhere near enough feeder routes to make bus + link feasible at that end.

      3. It’s not completely a lack of feeder routes. It’s also partially a resistance by the Metro route planners to modernize routes to reach Link and transition into being those feeder routes.

        I’ve been getting nowhere on trying to re-route the 132 to serve Tukwila International Boulevard Station (TIBS) or to get the 122/132 to serve TIBS or Airport Station. Executive Constantine hinted at providing a connection to TIBS at the South Park Bridge closure meeting, but the route planners said no such plan is in the works. There is a lot of vocal support for this re-route from my neighborhood, as those at the meeting could tell when the request was offered repeatedly.

        Metro is still in the mode of providing long express routes and long zig-zaggy neighborhood routes to downtown. Unless some of those routes are scavenged or truncated, they won’t be able to find the service hours to provide short local Link-connecting routes.

        And if Metro is saying its because every neighborhood is wanting to just keep their one-seat ride downtown, they are giving you a line of baloney. South Parkers want to fix the 132. Metro is ignoring us.

      4. Of the 4 stations mentioned, only 1 is ever likely, the Boeing Access Rd. station. It is the only segment in the final plans, and has the correct track circuits for a station. Sadly, as Ben stated, it will not happen.

    2. I went to Southcenter yesterday (5:15pm from downtown). I rode Link to TIB and transfered to the 128. It was fifteen minutes late and not quite as slow as molasses. It went out of its way to serve a high school & library, which I guess is important. The 140 left later and arrived at the same time, because I saw it en route. But both buses go every half hour till 10:30pm and are not synchronized, so that’s the biggest wait. It took 1 hr from downtown, +/- 5 minutes.

      I’d forgotten to write down the return schedule online, and there was no 128 or 140 schedule on the bus, and the 140/150 return stop had only a 150 schedule. I couldn’t even find the 128 return stop, although the route map seemed to say it was the same stop. I didn’t have half an hour to visit all the stops on the adjacent superblocks looking for it. So I just took the 150 back, at 9:35pm.

      The 150 took a speedy half hour. And I was glad to see it now runs every half hour till 10:30pm.

      So in summary, access to Southcenter via Link is shitty. The Burien-Renton RapidRide is severely needed. But even with it, given that Link to TIB is 35-something minutes, and the two-story walk to the bus bays, it’s hard to see how any Link-bus combination via TIB can ever match the 150.

      Which makes a 133rd station sound so much better. The 150 is the established route in the area, and if it could be truncated at 133rd/Interurban (I’m not sure where the proposed station is supposed to be), it would match or almost match the current 150 travel time. It would lose some time in Rainier Valley but gain it back in SODO. (It seems to me that trains go twice as fast through SODO as buses.)

      1. A couple clarifications. “It was 15 minutes late” was the 128.

        And the 150 is every half hour till 10:30pm (9:30 Sat, 8:30 Sun), then every hour until 12:33am. It used to be only once an hour in the evenings.

      2. If they opened a station at Boeing Access Road the 150 could start there instead of in the tunnel. Unfortunately this would truncate the portion of the 150 that is fast, i.e. from the end of the busway at Spokane St until it gets off the freeway in Tukwilla.

        The poor link/southcenter connectivity is a huge disappointment, although I more or less understand the historical politics and routing difficulties this represented. It is just unfortunate.

      3. Wait, isn’t there a direct bus between Tukwila and Southcenter?

        Also, really, we should be encouraging these businesses to relocate to downtown, instead of being out in the middle of nowhere. :)

      4. The direct bus is the 140, going from Burien to Renton. The other is the 128, which is not as direct but I remembered it because my friend lives near it. Both buses run every half hour, but one leaves five minutes after the other so you’re still in for a 20 minute wait sometimes. Given that Southcenter is just the other side of the freeway, it means you’re waiting longer than you’re riding.

        I thought about walking from Southcenter to TIB to see how long it takes, but didn’t this time.

      5. You would think the mall might provide a shuttle. But apparently they aren’t interested in attracting lowly transit riders.

      6. If memory serves me correctly, there is also a direct bus from the Tukwila Sounder station to the Tukwila Link station. I believe the bus only stops at the Sounder station when Sounder is running.

    3. I keep wondering… If Link will eventually stop at the following places:

      Westlake/Pacific Place, Northgate, Bellevue Square, Alderwood, and The Commons (Federal Way TC), will anyone care about Southcenter in 10 years? Probably not. True it’d be nice to have some easy way to get there until that time but I don’t think Link skipping over Southcenter is that big of a problem. How many shopping malls does Link need to stop at?

      1. Yeah, frankly, I don’t need to leave downtown Seattle for my shopping.

      2. Most of the bus slowness around Southcenter is due to how the streets are laid out. There’s a square of streets around the mall. Tukwila Bv is the north side, then a cross street goes over 405 to Southcenter Bv, which is usually where people are coming from. So both the 150 and 128 have to turn-turn-turn to get to the mall. Then the 150 turn-turns again to get to Kent. All these turns frustrate both bus riders and drivers. So, I can see Southcenter getting less and less popular for just that reason. On the other hand, it has “everything”, either in the mall or around it, moreso than other malls.

  2. I found this comment in the piece interesting:

    “Metro’s operators don’t have a strong incentive to be on time”

    Is that something that can be addressed? I know there are many variables outside the control of the operators, but it seems crazy that one of the biggest factors that determines John Q. Transit Using Public’s attitude about the service they are getting (“is my bus/train on time”?) doesn’t form the basis for some sort of incentive to those who are providing the service.

    1. I’m trying to find out more. They’re just throwing around ideas so far, but I’ll try to get something solid.

      1. Yeah it’s hard because you want drivers to drive as if time is important but you also want them to drive as safely as possible. I don’t know what systems are used in let’s say Germany or Japan but it would be interesting to know.

    2. I’m not a driver… but in their defense, the Link operators don’t have folks walking up to them to ask questions/for directions, they don’t have to handle wheelchair/carts, and riders are less likely to wave, run for, and pound on the side of the train to wait for them.

      It’s not exactly just laid back drivers.

      1. Amazingly, I’ve now seen two nit-wits pounding on the sides of trains as they are pulling away from University Street station southbound.

    3. Do LINK operators have displays which make it easy to tell how they are doing compared to their schedule? Something color-coded so they can tell if they need to be a bit more aggressive or if they should dwell at stations?

      1. No they do not. They have a paper schedule that they use. Most use it to start their trips and don’t look at it after that. The running times are determined by the speeds they run at and by the controller spacing them out from Rainier Beach Northbound and Mt Baker Southbound.

      2. Hi Brett

        I am A Link Light Rail Operator and are dwell time at the Station’s is All governed by the city light system. From the time we open our doors of the train we push our call button and we then have 38 sec to load before our light starts to flash Lunar and we need to go.If we wait to long the system will time out and will loose our cascade of lights and need to stop at every light.If that happens then trains bunch up.The trains run very Often so if you don’t make that train take the next one…Our headway is 7 1/2 min at peak so we need to keep the trains going…

    4. “Metro’s operators don’t have a strong incentive to be on time”

      Yes, I found that sentence striking as well. Ben (or someone else), could you please elaborate? This isn’t a hostile or sympathetic question. I just don’t have any context. thx.

      1. Operators don’t get anything for being on time – or lose anything for being consistently late.

      2. They gain and lose their breaks. There is only a 5 minute break at the North end and a 15-20 minute break at the South end.

    5. The idea that somehow Metro drivers have less incentive to do a good job because they don’t work for Sound Transit is ludicrous. Please, someone, provide some evidence. That’s just nutty, in my opinion. Are you thinking that ST would somehow inherently have better relations with their workers, or that Metro drivers are biased in favor of Metro work? They come to work every day to do their job.

      1. As a Link operater, I take exception to the comment that Metro’s operators don’t have a strong incentive to be on time. I didn’t have to come over to Link from the bus side, I wanted to, just like all the others operating Link trains. We as a group are proud to be light rail operators. We have a nasty schedule, which includes a 5 minute turn-around time at the Pine Street stub on almost ALL trips. That’s all! If we are delayed northbound, or heaven help us we need to use the bathroom (we can only stop at Pine Street or the airport), we will loose time leaving Pine Street SB. The lights on MLK have been working a lot better over the past month, and I think some of the reason why in the comments here at STB. And no, we have only our train schedule with us, printed on paper, and there is no cutting corners to get everyone where they are going safely. I’m not going to be aggessive, and the dwell times on MLK are set by Seattle DOT, not the operators. We now have a 38 second dwell time to set the lights ahead of us.

      2. It’s not ludicrous at all.

        The contract Metro drivers operate under for Link is the same one they operate under for buses. It doesn’t prioritize or reward for good on-time performance.

        A contract specifically for rail operations, with ST, would have different goals, and have provisions for operator performance. Talk to the Tacoma Link operators!

        kwakiudl – you are almost definitely not part of the problem. But operator performance is absolutely an issue on Link.

    6. I take route #5 from 145th and Greenwood to work. I take the run that is supposed to leave Shoreline Community College at 11:29am. Why is it that this stop is so close to the beginning of the route yet the bus is always 10 minutes late? I’m sure almost everyone on the bus at that point is coming from the college so why is it always so late?

  3. Is the Bay system (Bay A, Bay B, Bay C, Bay D) leading to delays in the tunnel? If one eastside bus pulls up short at the first bay and is delayed for a minute or two, everything lines up behind that bus and gets delayed. Would the tunnel flow more smoothly if all buses went to the far end of the station? There might be a larger crowd at the bus stop, but the buses would likely get through the tunnel quicker.

    1. A friend and I had this idea to help solve that. Since the order of the buses in the Tunnel never changes, 8 LCD screens should be installed one bus length apart above the platforms. So before the buses even arrive, the screen will flash the bus number and the exact position where it’ll be stopping. It would eliminate the bay and “where is my bus, I can’t see the front board” issues as you’d get a minute warning to walk down to your spot. Additionally, RFID readers unique to each bus, can be scanned on enrty and exit to each station, ensuring the order is absolutly correct.

      The system could be made smarter as well. If bus #1 got ahead of buses #2 and #2, then the system would removed #1 from that pack and #2 would suddenly have the lead at the next platform, making it #1 while the original bus #1 would join the pack in front of it. (If that makes any sense.)

      1. This makes great sense! If there are as many as four buses to line up before the next train enters that station, let them.

      2. Yeah I was thinking about something like that, wouldn’t that be awesome? Of course, there won’t be GPS on buses for another few years, and at that point it’ll only be a few years until buses are out of the tunnel…

      3. GPS wouldn’t work in the tunnel, heck, even Cell Phones don’t. But the RFID system is in place. I would think that it would take 3 years to make such a system work though. (I remember the “Next train” signs at Tukwila station. Do you? they were controlled by a person with a clicker.)

      1. You wouldn’t operate buses on-street with bays unless you could do skip-stop operation. The tunnel is just another street. If the platoons are three buses or less, they can all stop in one place. If there’s a fourth it can stop again – but that would happen more rarely. Yes, please definitely ask about this – and while I like using real time signs to improve the concept, the bays could be removed without them and still have a significant benefit to operation.

      2. It could also be a detriment to operation. Those are 120m platforms – if someone’s expecting a 255 at one end, they’ll miss it if they’re at the wrong end when it stops.

    2. At least the inbound buses could go to the farthest bay. There’s no reason for a 550 going north through the tunnel to stop at the “eastside” bay when it’s dropping people off.

      1. you can’t have outbound buses stopping all over the station–blind or handicapped people take the bus and it would put them at a huge disadvantage if they couldn’t predict where to wait. Especially the blind–who I’ve watch getting trained on where to wait for the bus. No way could they make it from one end of the station to the other and catch their bus.

        Now, inbound buses like the northbound 550 could stop anywhere. The blind could find their way out of the station no matter where they were let off (they have textured strips on the floor…).

        That said, having the “eastside” inbound buses stopping anywhere might create a new set of problems. They’ll block the bay where all the outbound buses need to stop in order to pick people up. For example, the northbound 550 would block the bay needed by the 7x’s.. who then have to wait further back (and thus get the wrath of angry mobs trying to get the driver to open the door).

        Bottom line is the tunnel is full–changes to how and where the buses pick up and drop off passengers might just make things worse.

  4. Rather than having two separate bus bays, could the bays be merged at the far end, with buses pulling up one behind the other regardless of route number?

    Also, could we have a tow truck assigned to the tunnel, so buses could be towed out of the way if they stall more than a few minutes?

  5. Fantastic post! It addressed a number of things concerning tunnel reliability I have been thinking about recently.

  6. I’m still working on getting a connection between my neighborhood (South Park) and TIBS. The Executive hinted at that happening at the recent South Park Bridge closure meeting. Then, in conversation with Jack Latteman (Senior Metro Planner) afterward, Jack told me no such plan exists.

    Unfortunately, there is little the City of Seattle could do to help, since the route would run through northeast Burien and north Seatac.

    I think there is still pent-up demand of riders wanting access to Link from their neighborhoods, and plenty of people who don’t insist on that access consisting of a park&ride lot ruining the opportunity for transit-oriented development next to a station.

  7. You brought up real time arrival? I hope you weren’t too hostile considering ST has so much else to worry about. Thanks for the info on the rest of the operations though, interesting little read. I didn’t realize that the trains in the Valley can actually miss the light cascade. With all the fancy stuff they installed, one would think that it’d know where the train is, calculate its velocity, and arrange light timing accordingly.

    Happy to report that SD’s Trolley (actually light rail) doesn’t have real time arrival info either. But the 40-year-old BART and MUNI do. BART is also insanly cool.

    1. I’ve written to ST about the importance of real time arrival. I think it would be really interesting to create a new open source effort around this that ST and other transit agencies could use as the basis for real time arrival systems. There may also be existing projects that could be leveraged.

      It can be broken down into a set of smaller problems, each managed by an engineer from Soundtransit who will provide the necessary details, guidelines, data sources and limitations presented by the LINK system and express the preferences of the organization where appropriate. Soundtransit could exercise more control over the component selection and the platform on which the software would run since these involve direct costs to the organization.

      The practicality of this idea is *highly* debatable, but it could be an interesting way to get a high quality design and may even be a source of publicity for the organization. I’m not at all sure it would actually save money versus contracting the system out, but it could have intangible benefits. Okay, now you can tear my idea apart ;-)

    2. If the train is less than 2 minutes from arrival the readerboards could have a message that says “train will arrive in less than 2 minutes” instead of scrolling safety and security messages. When the train is about to enter the station the voice system could then repeat the “stand behind the yellow line” message.

  8. “Link waiting if a wheelchair needs to board or something goes wrong with a bus.”

    I just want to smugly say, told ya so. Long ago I warned of this problem, and the transit nerds on this blog dismissed my prescience prediction, saying there would be tunnel crews on standby to quickly take care of any such problem.

    Just want to say one more time, told ya so.

  9. Ben, have the April Link ridership numbers come out yet? Sorry if this has been asked and answered already . . . ..

      1. I remember a comment from Joni Earl on her CEO corner email last week talking about “preliminary” LINK numbers for April. She stated that there were many days in April that exceeded 20,000 riders and that LINK ridership gains were continuing.

      2. Yes, I’ve been told that there was a weekend day that went over 23,000, in fact.

    1. I just read the update – the numbers are down slightly for April, but of course one month does not a trend make!

  10. I don’t mean to re-hash old debates, but has there been a second look at whether it would be advantageous to have tunnel bus routes coupled?

    E.g. Have the 41 become the 101, completing the one-seat ride between Northgate and the sports palaces.

    With use of ORCA data, the best couplings could be figured out, and some riders would be able to go all the way through the tunnel without having to take the time to disembark and re-embark.

    Also, could the tunnel be removed from the Ride Free Area? That would make walking onto Link the cheapest way to get between tunnel stations.

    And one more thought. How about running the RapidRide Line C through the tunnel? Or are the RapidRide buses the wrong exhaust system for going through a tunnel?

    1. I think you’ll see the Metro Task Force look at ideas like that. They aren’t going to take on any other big questions like those until after that report, anyway.

    2. Ooh that would be awesome if they had buses go straight through. That would reduce delays for Link by a lot. I also think that the tunnel at least should be removed from the RFA, and they should have an option on the TVMs to get a bus ticket. Or, standardize the fares between ST and Metro!

      1. Or allow Link tickets to be used on tunnel bus routes. But then, it’s still not as good as people just getting an ORCA card.

    3. People already use Link to ride between stations inside the downtown tunnel. Fares are NEVER checked on Link within the tunnel, and people use Link just as they use buses inside the tunnel. I have found that between 10% and 15% of all Link boardings are people taking trips just between stations inside the downtown tunnel.

      If Central Link is now averaging around 20,000 boardings per day, that would mean that between 2,000 and 3,000 Link trips per day are just between stations inside the downtown tunnel — trips of about one mile or less. Just today, I saw two people get on my northbound Link car at University, which means they had to get out at Westlake, for a trip of about 1/3 of a mile. I don’t think anyone is letting a Link train pass by, and waiting for a bus, if they are taking a trip between downtown tunnel stations.

    4. It would be hard to find routes that have well matched demands. The thing is that typically the demand is mainly peak direction, so it makes sense to have most 41s return to Northgate directly after going through the tunnel in the morning rather than having such high off-peak frequency. So which buses have high demand in the off-peak direction? Maybe the University District buses, maybe the 550. I’d kind of like to see the 550 paired with the 545, not because that creates any sensible one seat rides, but because that would allow the 545 to be moved to the tunnel without increasing the number of buses in the tunnel during rush hour.

      1. One problem with route pairing freeway routes is that they are often estimated timepoints. Years ago, I drove a 150 that was scheduled for 20 minutes longer than it took me to go from Tukwila P&R to SODO/Spokane. If it was through routed, I would have had to wait somewhere for 20 minutes to pick up the schedule for the trip out of downtown Northward. This type of scheduling occurs with all of the freeway trips into downtown to allow for freeway variations.

      2. True, but this doesn’t matter as much with frequent buses like 550/545, which have 10 minute headways during peak hours. If the buses come regularly enough, who cares if they’re off schedule? During off peak hours they could be de-coupled.

  11. As a corollary to tunnel route coupling, could the ORCA data be used to identify which routes have people merely transferring downtown, or even passing through downtown, and make the downtown routes with the highest percentage of riders actually heading elsewhere become the tunnel routes, kicking the routes for which downtown is the destination for most riders upstairs?

    1. I know that’s one of the reasons ORCA was created, but pass riders aren’t even entirely on it yet – give it time!

  12. Terrific post Ben! Regarding Real Time Arrival Info, is it possible to get a more specific answer on why it hasn’t gone online as hoped? I believe they use GE Transportation Systems (I could be wrong)I’m curious if it is an issue with whoever sold them the system not implementing it properly or not providing enough support.

    My feeling when you are having trouble with an existing technology (meaning there are plenty of North American cities with Real Time functioning flawlessly)that if you are having trouble getting something to work, finding people who know how to make it work shouldn’t be that difficult.

    1. There aren’t any North American cities with bus/train operations. Realtime arrival systems “guess” quite a bit, and they’d be off a lot of the time because of delays.

      1. Since they already have data for certain points before stations, they could compile all of that and have it guess for the time between stations, correcting itself if the train got off schedule. Now that we know that it’s about 90% on time these days, it wouldn’t be that often that the guess would be wrong anyways.

      2. Buses are not responsible for the lack of real time train information.

      3. Actually, as I understand it, they’re a big part of the problem. The performance issues that joint operations cause means realtime arrival information can’t be easily implemented the way it is in other systems.

      4. This might be a silly suggestion, but I’m wondering if, since we apparently can’t accurately estimate the amount of time it will take for a bus or train to get between its current location and the station the user is in, it might be useful just to have displays that show the current location of the next train / bus on a given route, and leave it to the user to estimate how long it will take that train / bus to get there. Something sort of like the old interface, displayed on screens in the stations…

      5. The tunnel is not the same as the rest of the system. In the tunnel buses can be detected as they enter, and real time signs can be based on the real-time detection rather than on polling the buses every couple minutes as is done everywhere else. There ought to be a way to develop real-time bus information in the tunnel that’s better than what’s available elsewhere.

      6. “can be” and “ought to be a way” are things that cost money on top of a standard realtime arrival system.

      7. Can they say, “The next train is two stations away”, rather than giving the exact minutes. That would be almost as good.

  13. Thank you for this information. Now I have a better understanding of why the delays and about this agreement mentioned about the buses in the tunnel–a reality that is unfortunate. (What about making 3rd Ave. a bus-only mall?) and taking more buses out of the tunnel sooner?

    Question: I take the train northbound from Othello Station at 0545. That time of day the trains are every 15 minutes until shortly after 0600. Having an extra train or two beginning earlier in the morning to increase frequency northbound would be advantageous. The trains are more frequent going southbound, and I understand why, so could southbound frequency somehow ramp up to increase northbound frequency?

    Also the reader board on the Othello southbound platform has had the incorrect time for months. Could it be addressed and corrected?

    You guys are awesome. The blod is great. Thanks for all the coverage and information and commitment to improve public transit here in the Puget Sound region.

    1. The clock at the top of the escalator at the station entry on between 6th and 5th is still on standard time. I’ve let the folks at the info booth know, to no avail…

      1. Speaking of clocks I’d love to see the station clocks in Westlake and Pioneer Square put back into service.

      2. I see that clock every day. It says to me “that clock isn’t on their master clock” and is instead some crazy one-off thing.

  14. “Tober says it’s possible joint operations will continue even after University Link opens, simply because there are so many more bus than train passengers, and they’d be put on congested surface streets.”

    My immediate reaction to reading this: “Oh, HELL no.” I am so looking forward to getting those buses out of there.

  15. Ben, what I want to know is if ST has been talking with SDOT about potentially making Third Avenue a full “transit mall” with 24/7 bus and bike only access. This could alleviate bus congestion that is moved from the tunnel, allowing priority for trains below.

    1. I just read Othello’s comment and he mentioned the same thing. I forgot to add this. Last year, I contacted Bill Bryant who leads the transit program at SDOT, and he told me that extending peak hour access to buses and bikes will be addressed in the update of the Transit Master Plan.

    2. I don’t know, but that discussion wouldn’t start until 2012 anyway.

  16. “In short, Link reliability is something Sound Transit is working on and it will improve over time, but it won’t run like clockwork until there are enough riders to justify removing buses from the downtown tunnel – potentially not until Link reaches Northgate, and high ridership buses such as the 71, 72, 73, and 41 don’t need to go downtown anymore.”

    University District service won’t be ideal until Brooklyn station opens, but could someone check my reasoning? It seems to me that the 70-series is really two routes: Wedgwood/Lake City/Jackson Park to the U-District, and U-District to downtown. Wedgwood/Lake City/Jackson Park service to downtown is provided during rush hour primarily by the 76/77/79 (and other routes, like the 522 in the case of Lake City, or the 64 in the case of Wedgwood). I suspect the U-District-Downtown leg is responsible for most of the 70-series’ popularity, and it’ll be mostly obviated when UW station opens. Might the 70-series be rerouted and truncated to UW station, with the 70 (and later a possible streetcar) picking up the slack for Eastlake service evenings, nights, and Sundays? Brooklyn won’t be a good place to terminate the 70-series anyway, since part of their role is also local service along University Way, and the core of the Ave would be skipped if the 70-series ended at 45th. (How many routes will terminate at Brooklyn at all? The 271, the 70, the 133 and Sound Transit routes in place of a loop, a split 25 or 30, the routes currently using the 11th/12th loop? UW station will always be a more important transit center because of the importance of the UW and the proximity to 520.)

    1. The problem is the large amount of traffic between the U District and Husky Stadium. It would be hard to get that many more buses in there, although I suppose you could do it. When North Link opens they should do a complete redesign of the whole network, wiping the slate clean and starting from scratch. They should try to make it more of a grid system with Link as the spine, as opposed to the current mishmash of infrequent routes.

    2. “University District service won’t be ideal until Brooklyn station opens,”

      True that!
      North and South campuses are *completely* different worlds. Folks on the Ave or in Denny Hall or even the Quad probably aren’t going to walk down to the station; they’re gonna get on the 70-something bus on 15th.

      UW station will be for “south campus”; probably not for “north campus” or the U-District residents/workers/visitors.

      1. Yeah, I can’t imagine even the majority of students using U-Link to commute to school until Brooklin Station opens, which is why I hope that stop is somehow expedited to open before the rest of North Link (not counting on it though).

      2. It would not be possible to open that station earlier than the rest of North Link, no. There aren’t crossovers there.

    3. Then there is the silly-ness of rerouting the 76 so it also goes through the tunnel south to north when heading northbound. The 76 was the antimatter of the 70s series. If you worked at the south end of downtown, it was a wonderful alternative…

      The re-route added 20 minutes each way for south downtown commuters and took away a wonderful alternative for those times when events at either end of town impacted you.

      Now they fall in line with all the other 70s series… great for the anal retentive, but a real loss for those of us who worked around the city and county buildings and Columbia Tower area. I still have not heard a good reason for doing that switch…

      1. Major impact, on the 76 especially. Silly decision, I’d agree. I loved the 76 when I lived in Ravenna – always 10 minutes of mmore faster than the 71.

      2. Ben,

        FYI, the move happened a few months back and was the major conversation among the 76 ers for weeks… the best reason anyone could come up with (all of us being consumers only) was they wanted to keep the 70s series clumped together to add capacity N bound to the the common spots. Problem with that answer is that by the time the 76 joins the thicket of NB in the tunnel, it fills with those who can opt for those runs, and the South- end workers have just accepted their commute now takes 40 minutes more. I spent a year using the 76 on the old route on and off wit the 64 for work at an office midway between both on 5th Ave… I have since moved on… but see some of the old 76ers on my 64 commute run…

        Maybe you can find a better answer than they have.

  17. Sound Transit needs to decide whether it wants to work with Metro to tackle the tunnel problems, or whether they want to make things as bad as possible to support their case for kicking the buses out. If the buses come out, a whole lot more transit riders will lose than gain for at least a decade or two.

    Standing in the station and watching, there are tons of seemingly obvious operational problems.

    Buses enter the tunnel in platoons, but out of order. The first bus has to stop at the first bay, and the buses after need to wait for that bus to leave before they can get to the second bay. When the tunnel was built, the solution to this was the large staging areas at Convention Place and International District stations, where buses were supposed to be dispatched in platoons, but in the proper order to arrive at the right bays. A smart signal system could automate that. Another approach would be to eliminate the two bays – but that would require normal street operating rules (the fourth bus in a platoon would need to stop twice), or use of the dynamic signs to let people know what buses are loading in what order. (That would also help those who just want to circulate downtown and now need to run back and forth between bays to catch the first bus that arrives).

    Then there are what seem to me ridiculously conservative operating rules established when light rail began. For example, buses have passed each other in the tunnel for more than a decade without incident. But when light rail began, new rules forbid bus drivers from passing each other because they might get too close to a light rail train going in the opposite direction — which is *less* dangerous than for buses in the opposite direction because (1) rail cars don’t wander laterally as buses do, and (2) rail cars can’t enter the passing lane in the opposite direction. In theory, drivers are supposed to be able to call the dispatcher to pass, but what I’ve seen in practice is reluctance to pass unless a supervisor is on the scene to OK it. Solutions include lightening up the passing rule (since the rules are wrecking operation without enhancing safety in any provable way), encouraging people to call the dispatcher when needed, or locating operating staff in key tunnel locations to keep things moving. Having someone in the station or dispatching who is dedicated to keeping things moving can make a big difference.

    Metro’s had some of these problems for years because they never took optimizing the tunnel seriously, since there were never enough tunnel buses to strain capacity. But it’s a different story now with joint operation, and these are problems that could be solved with some attention and determination. (When rail takes over the tunnel at high frequencies, there will be a whole new set of operation problems, given no cross-over tracks that would allow problems at individual stations to be isolated and service to continue – but that’s a problem for another day).

    1. And MORE so to the point… On the streets and in the tunnels Metro drivers come within inches of each other bumper to bumper… but the 2 minute or 90 second or whatever silly bubble around each train makes NO sense to me. We trust our drivers to come within inches everywhere else, but G0dforbid we let a coach enter the station before the train has completely left the area. They land them closer together at SEATAC for gosh sakes, and thats WITH Jetwash.

      SERIOUSLY… So bizzare to be five coaches back in the tube until one can no longer see taillights of the train in front.

      This is not general public driving… these are well trained METRO, PIERCE and CT DRivers who are trusted to keep metal from mingling at all OTHER parts of the Journey, but not allowed to enter the rarified air of the Link within the station.

      AND the folks here worry about lag time paying, or dropping a lift for a wheelchair… yet “Ding DING” and NO BUS until the train is safely away down the tube.

      THERE is your schedule delays big time… Since the Train and its rule of the roost, not a 71 has made it to Westlake on time for my commutes. NOT ONE.

      I am amazed at this process. It screams to me, ” we trust you out there, up there, and among yourselves, but really do NOT trust you can keep from hitting the train.

      1. Hear hear. In Boston, even light rail operates with line of sight control in stations and achieves 50-second headways. Here there is only one sight limitation in the entire tunnel. I’d like the people who set standards to sometimes have to prove the case that ultra-conservative policies have measurable effects on safety.

    2. Greenwood Rider, just from the beginning of your piece, I feel like you didn’t even read what I wrote.

  18. It seems to me that one of the congestion solutions would be to add a bit more station capacity. It seems a bit silly to have that huge middle sections between northbound and southbound track where nothing is currently allowed to happen. There’s enough space there for another entire platform plus bus pull-outs.

    However, to my mind, the way to tackle this capacity problem would be:

    1. Make the tunnel part of the Ride Free Area. That way, to the passenger’s wallet, it doesn’t matter if they are taking a train or a bus.

    2. Convert the “bases” at the south end of the tunnel and north end of the tunnel into bus terminals – basically massive tranfer stations combined with bus layover facilities. Those then become the start and end of the ride free area. These will also need to have a few light rail tracks added to them for train layover and turning (it doesn’t take much time for a train operator to walk to the other end of a train, but it does take some time, and you need it to happen somewhere off the main line).

    3. The tunnel becomes train only.

    4. Due to the elimination of the bus routes in the tunnel, it is now possible to increase the frequency of trains in the tunnel. The tunnel is basically grade separated from city streets (except for the very south end) so fairly frequent operation should be possible. Once every two or three minutes shouldn’t be too hard to manage. Perhaps once every 5 minutes would be possible to convince SoundTransit to operate. This high frequency of train service in the tunnel provides vastly improved transit times due to the low wait period. It isn’t that much different than arriving in the tunnel and getting directly onto your bus. It also provides extremely good connectivity between a huge number of routes on the north and south side of downtown.

    5. Sell the current SLUT cars to Portland and buy a current streetcar design that is just as wide as a light rail car (new streetcar lines in Europe are never built as narrow as SLUT or Portland’s Streetcar – those narrow widths are only reserved for antique lines). Rebuild SLUT platforms to use the wider car design.

    6. Extend SLUT south and connect it to the track in the transit tunnel.

    7. SLUT now becomes a central city circulator, not just a branch line route, and is able to help with keeping the frequency up inside the transit tunnel. This, plus full LINK trains, plus LINK trains that circulate only in the transit tunnel and turn at the terminal ends, would provide very good frequency and connectivity that should pretty much eliminate the need for buses in the tunnel.

    As for the need for crossovers to bypass problems at individual stations, the problem there is that if you are running high frequency, you need to have both tracks operate in a single direction. There really isn’t enough capacity for a section of track to become bi-directional.

    TriMet in Portland runs a three mile tunnel with no crossovers, but they have built three track sections at each end of the tunnel so that a problem train can be quickly removed from the tunnel and placed in an out of the way location. This is another purpose for the train terminal and turning spots at each end of the tunnel.

    However, I doubt very much anyone is going to want to spend the money to accomplish all this.

  19. Regarding buses in the tunnel, you also have to look at it from the bus riders’ perspective. Yes, buses ran on the surface during renovation and 3rd Ave was transit-only at peak hours, but it still took 20 minutes to get through downtown — two or three times as long as using the tunnel. I’m still pleasantly surprised whenever I see the tunnel open after 7pm and on Sundays because it’s such a major benefit.

    So, kicking out the buses is really telling their riders they have to suffer stop-and-go traffic and streetlights because their house isn’t on a rail line where the cool people live.

    Not that I think buses should remain in the tunnel to hinder the trains. But it all suggests that maybe Portland is right, and we should get rid of the RFA for buses but add it for trains. The fare inspectors may be de facto staying out of downtown, but all they’d have to do is officially stay outside downtown. Although, it would suck for riders who intended to stay downtown but changed their mind after they got on the train, they’d have no way to tap in unless readers were installed on the train. But then the reader wouldn’t know where they are. Or would it? The train presumably knows where it is, and certainly the driver does, so he could press his “arrival” button and it would be synchronized with the reader. How does Portland handle train payment, is it all a flat rate?

    1. I think removing lower-ridership routes from the tunnel is a good step as Link builds more ridership, but yeah, I don’t want to kick a bunch of people out just to improve performance for a smaller group.

  20. Below is a response from ST on reliability. I requested the info some weeks ago. Overal it’s not good, but looking at the overall trip times of 37 minutes scheduled, they are late (over 40 minutes per trip) only about 7% of the time. So, indeed, it looks like when the bugs are worked out, and construction ends, things will get much better.

    In response to your email of May 9, 2010 regarding a monthly summary of Central Link on-time performance, I am happy to provide the following data we have on record.

    Sound Transit defines on time performance as a train arriving at SeaTac/Airport Station and Westlake Station +/- 3 minutes of the scheduled arrival time. The performance measures for February, March, and April are:

    February 2010 71.24%
    March 2010 73.17%
    April 2010 65.90%

    The average on-time performance for the referenced three months is approximately 71%. Construction and some maintenance activities impacted Link service to some degree on 35 service days during the February – April reporting period, which contributes to the 71% on-time performance number. It should be noted that most of these service impacts occurred on weekends and outside of peak service hours. Another challenge that affects our on-time performance is joint rail and bus service in the Downtown Seattle Transit Tunnel (DSTT). Maintenance and other issues that may delay bus service traveling in the tunnel will delay rail service.

    Because of the operating factors mentioned above, Sound Transit uses headway performance to measure productivity. Headway performance is monitored against our advertised operating frequencies of every 7.5 minutes during peak hours and every 10 or 15 minutes at other hours. The average headway performance for the referenced three months shown below is 89%. In other words about 89% of the time our passengers did not have to wait more than three additional minutes beyond the scheduled headways for their train. The monthly headway performance is noted below:

    February 2010 90.23%
    March 2010 88.77%
    April 2010 87.12%

    We’ve interpreted the information you requested as end to end performance. End to end performance is measured by considering the actual time trip time when measured against the scheduled trip time of 37 minutes. Those figures are:

    February 2010 92.8%
    March 2010 92.3%
    April 2010 93.0%

    The average number of trips performed within the scheduled time for the referenced three months is 92.7%.

    We hope you find this information useful. Thank you for your interest in Sound Transit.

    1. From what Tober showed me, May is already significantly better than this.

  21. Many good comments above re: Tunnel improvements. However, chief reason for poor schedule performance for buses is that buses have to spend so much time in regular traffic- making them into huge, slow, clumsy automobiles.

    Like any other automobile, every bus is at the mercy of the worst non-transit driver on the road- as well as everything else from weather to sporting events to visiting politicians that stalls traffic.

    Any transit vehicle keeps schedule in direct proportion to how effectively things are kept out of its way. Want to improve bus schedule performance? Get onto your elected officials to start reserving lanes for transit and getting serious about signal pre-empt.

    Be prepared to exert more pressure than the multiple vocal interests who constantly oppose these measures.

    Couple of cost figures elected reps need to hear: I’m told a bus operating hour costs $100. Divide that by 60 to get price of one minute delay. And start sending your politicians bills for your own hourly time lost when your bus is late.

    Tell them you appreciate not being killed by dangerous buses, but don’t like being robbed by slow ones.

    Mark Dublin


  22. this piece attracted comments from some who know transit pretty well: Dublin, Skehan, and Greenwood rider. Tober is very good.

    the tone of the post and the comments could acknowledge that it is not ST v. Metro, but ST and Metro working together. ST and Metro are considering changes to the tunnel operating rules. Metro employees run ST services, Link and bus. of course, operators have incentive to run on time; as a commenter stated, they get a longer break. the lack of complete grade separation along MLK Jr. Way South means that Link itself cannot run perfectly.

    a major issue is the ST choice to build Link south first in 2001. that means that the higher ridership north Seattle routes would be bus longer and that the Link trains would actually reduce the effective transit capacity of downtown until it was serving markets on sides of the DSTT.

    we should keep the riders in mind, on both modes. to help riders, the tunnel throughput should be maximized. it is more reliable and 10 minutes faster than the surface. this advantage will grow over time as the surface streets congest.

    as Greenwood pointed out, real time information could be provided for both bus and rail, but at a cost. both agencies are making those decisions.

    it is not the extent of Link that will end joint bus-rail operation, but a much shorter Link headway. even when Link reach Northgate, expected in 2020, with four car trains, headways will probably be long enough to allow joint operation. ST Route 550 will be in the DSTT for all of that time and more. Routes 101 and 150 are good trunk lines. when routes 41, 71, 72, 73, 74 leave the tunnel, there may be other bus trips that make sense to put in the tunnel or maybe improved frequency on existing tunnel bus routes.

    interlining routes in the DSTT is worthy of consideration, but requires serious restructuring that is not under official consideration today. routes in the I-5 and I-90 corridors only have a reliable path in the peak direciton with the flow of their reversible lanes, at least until tolling or the completion of the third stage of R8A. the length of interlined routes should not be too long to preserve reliability, as the inbound trip of one route becomes the outbound trip of the second. it implies that transfers will be imposed at transit centers outside downtown. that has costs and opportunities.

    1. If the 101 is such a good trunk line, why is it being cut back? And if it is being cut back, why not truncate it at Rainier Beach Station to keep or improve headway on the route, while expanding the hours back to the hours one would expect on most all-day routes? Could this happen on weekends, at least as an experiment?

      Yes, average travel time to downtown would increase slightly, but people who need that route would have it available any time they need it, and would have much quicker access to other destinations along Link.

  23. Ben,
    This may have been implied in some of the comments above (I read almost all of them yesterday) but when you meet with Tober again I’d like to make sure this clarification on real-time info is covered: Posting countdowns to the next train in the stations has almost no influence on what I do – I’m already waiting at the station. A relatively simple data feed to the web saying where the trains are (I hope they know where the trains are!) would be useful and actually help me decide when to leave home, a restaurant, work, etc. Unlike BART and other systems all of the Link trains go to the same place (except if they’re going out of service I guess) so the whole next train countdown thing is kind of cool but not very useful. As I walk toward an underground station – before cell service cuts out – it would be useful to know if the train just passed and I have time before the next train to do something else like make a call above ground, grab a snack, etc.


  24. ScottA, I’d disagree on necessity of real-time information at the stop. The intangible value of real-time arrival info and the additional feeling of control it gives is great – it has made my use of transit on trips abroad much easier.

    A message saying “Train arriving in X minutes” would be a significant improvement and a reasonable request.

    I’ve had to skip using the restroom at the Tukwila light rail stop (before light rail went all the way to the airport). I was in a rush to get back into town from the airport, and waiting 15 minutes in case I missed the train would’ve made me late. That kind of experience puts people off using transit.

Comments are closed.