Shallow Under 5th (Sound Transit)

[UPDATE:  To be clear, ST is not dismissing Sounder ridership. The author is.] One unexpected contention point in the ST3 plan is the precise location of Chinatown Station. The reference alignment places the station under 5th Avenue, steps from the existing station. However, Chinatown business that have already suffered through streetcar construction have no interest in disrupting business access yet again, although in the long run the neighborhood would presumably benefit from a closer station. SDOT thinks a 4th Avenue alignment may fund some work they already do there. Sound Transit, reasonably, is most focused on the easiest and cheapest way to complete the project. As usual, no stakeholder is primarily interested in the convenience of future riders.

This convenience matters: seamless transfers encourage ridership, and thousands of people will be switching between Link lines at this station every day. Indeed, this will happen as early as 2023 when East Link opens, but Sound Transit has already added an unnecessary escalator and flight of stairs (or two slow elevators) to that transfer experience to avoid a minor capital project.

Another consideration is Sounder, though about an order of magnitude smaller in ridership than Link by the time ST3 opens. Amtrak is probably another order of magnitude below that. But what is the right answer for riders? Sound Transit has four concepts for how these transfers might work: under 4th or 5th Avenue, either cut-and-cover or mined.

Deep under 5th

The 5th Avenue cut-and-cover station (above) creates more disruption for the neighborhood than the mined option, but it drops riders much closer to the surface. Northbound and southbound riders at the new station would be the same distance from the old station, though the lower one would have a longer ride to the surface. Sounder transfers are not a priority in this alignment. ST says the upper platform would be the northbound line, because it “best facilitates northbound-to-northbound transfers between the West Seattle-Everett line and the Ballard-Tacoma line, which is generally expected to be the highest volume transfer during the highly-concentrated AM peak period.”

ST fears that a side-by-side alignment, given the constraints of foundations under 5th, might not have the needed platform capacity.

The mined station (at right) plunges riders about 200 feet below the road, but construction would be less visible on the surface. This is deeper than Beacon Hill and much deeper than UW, implying elevator-only service. This may not be suitable for a high-ridership station. Transferring riders wouldn’t have to go all the way to the surface, although the complexities of emptying half of a crowded elevator are probably worse than simply taking it to the surface.

Under 4th, a cut-and-cover station would provide easy access to the current southbound track, but the northbound track requires going to the surface. Although the schematic isn’t clear, it appears to be possible in theory for the mezzanine to allow access to Sounder without having to cross 4th Avenue, though a surface path is inevitable. ST confirmed they are also looking at a stacked configuration here, as the BNSF retaining wall under 4th limits the available space.

Shallow under 4th
Deep under 4th

Finally, a mined 4th Avenue station requires elevators, so the mezzanine provides little value while worsening crowd flow. Once they’re on the elevator, you might as well have them walk through the beauty of Union Station instead of a rat maze.

Of the four options, it’s clear that a shallow station is better for future riders. A crowded train at a downtown station and major transfer point is likely to overload the elevators and cause long waits. Moreover, a mezzanine stop that facilitates transfers is likely to cause further delays if large enough to accommodate large crowds.

Meanwhile, the two shallow options are hard to reason about, partly because the stacked and center-platform configurations may be applicable under either avenue. It’s even hard to figure out which transfers matter. Clearly, Chinatown will be the focus for transfers between Bellevue and both SeaTac and West Seattle; but other north-south and south-north transfers are best made elsewhere.

As for north-north and south-south transfers, we don’t have enough information. To get from SeaTac to Northgate, one might transfer at the first opportunity, Sodo; but then there is half as much frequency as there is further north. Chinatown is an option, but depending on the relative speeds in the two downtown tunnels, and the transfer experience at Westlake, that station may work out better. There is similar uncertainty for South-South transfers.

In all, it’s hard to objectively measure the tradeoff between being really close to Link and really far from Sounder (5th), or being a medium distance from both, even though the Link transfers are more important. The difference to riders between these concepts is not all that large. If the contest turns out to be damage to downtown business vs. ease and cost of construction, this seems like a difference that can be patched with money. ST passes on some of the savings from the 5th avenue project to local businesses, they survive the construction period, and ST doesn’t get bad headlines for delays and overruns.

83 Replies to “Chinatown Station Locations”

  1. Thanks for focusing on this! It’s very important!

    The rider transfer hassles should push the deep station options into the “not suitable for further study” category. It’s a major inconvenience to build a platform much deeper than UW Station as a major transfer station.

    1. The fact that a 200′ deep station is still on the table, two years after ST3 passed, tells you more than you want to know about the Seattle Process. I made pretty much the same comment as Al S. in a speech last month to the Washington State Ridesharing Organization. You may as well just skip directly to Midtown, than force thousands of transfers from a platform 50′ deeper than Beacon Hill.

    2. It is not just a major transfer station, but a major station, period. Right now it carries close to 6,000 a day (roughly tied with SeaTac for fourth). It will continue to go up, as this is extended to the north, as well to the east. It is the first downtown station for many, and thus will always be popular, even if folks just walk to their destination (or take surface transit). It is nuts to consider an elevator for that kind of station.

    3. I’ve been harping on it at the handful of public meetings I’ve been to and every online feedback form I could find… no deep station at Chinatown/ID. (Really, no more deep stations, period.)

  2. The north-north and south-south transfers should be a primary design objective of SODO Station. Because it’s not tunneled, it’s much easier and cheaper to build for those transfers at that station.

    That would then leave this station to be designed more for Eastlink transfers and Sounder transfers (and bus and streetcar transfers above).

    What are the transfers that are more important? That depends on the volumes.

    For opposite direction transfers on Link (Eastlink to/from the south), an option that could be added is to put the northbound tracks under 4th and the southbound tracks under 5th. That split sounds confusing at first, but it could enable level cross-platform transfers for going in the opposite direction (albeit a rider would have to walk fully under the King Street Station building for one of them).

    1. north-north transfers can be at IDS, and south-south transfers can be at the new SODO station. If the tracks are stacked at SODO as well, then southbound will be on the bottom, making it possible to do a same-platform south-south transfer at SODO just like the north-north transfer at IDS.

    2. For opposite direction transfers on Link (Eastlink to/from the south), an option that could be added is to put the northbound tracks under 4th and the southbound tracks under 5th.

      I love that the commenters on STB get into the weeds of transfer design and options, but I have a very strong suspicion from the public ST3 meetings I’ve been to that ST will not undertake any design that interrupts existing Link service more than some stretch of weekends. It seems to underlie some of the offhand alignment/construction comments I’ve heard. So I expect they will build a constrained or even crippled ST3 system (see: 200′ deep stations & some SODO or Chinatown/ID alignment options still on the table) before they disrupt existing Link service.

      1. JSG: but ST is disrupting current I-90 riders between 2018 and 2023 to achieve lower capital cost of East Link; are they still planning on disrupting current riders to install a crossover track at IDS?

      2. The planned track is a turnback track, not a crossover track. But yes, it is not as good as a flyover track that allows a train that has to be taken out of service to head straight back to the base. That turnback track will become a very expensive cheap-out.

      3. C’mon. Yes, they will do the minimum to connect the East Link tracks, but it seems they will do everything they can to avoid the ST3 alignments disrupting then-existing Link service. That’s just the sense I get, anyway. I would love if they did some creative/smart/future-proofing transfers or alignments.

  3. What’s the best option in my opinion?

    5th ave cut and cover (the first image) bar none. It’s not even close.

    First, there’s no point in building it next to Sounder if we won’t even connect it to Sounder, and ST even says that Sounder transfers is not a priority. Worse, there is the 4th ave viaduct rebuilding, making it the worst case for project costs. Finally, there is a retaining wall to deal with, and those are a pain, and the I-5 retaining wall was likely a determining factor in cutting Convention Place and First Hill from U-link. I think ultimately a 4th ave alignment will have no integration with the current system, and should just be seen as being included for completeness, and to show Chinatown why 5th ave is important.

    5th ave mined is scary because it sounds like a construction compromise that ST could find tempting. We don’t need a downtown station deeper than Beacon Hill, it just doesn’t make sense. And making the station not cut-and-cover makes construction of the station itself more complex.

    I think if you look at 5th ave cut and cover, you can only conclude that it has the #1 transfer environment, including a high-volume same-platform transfer! And a short mezzanine walk for the other transfers which, guys, we can deal with that. Much better than going to the surface. We usually think of mezzanines as silly and a waste of time, but that’s because we have mezzanines on a single-line system, and they are a waste of time. But rail-to-rail transfers drastically chance the mezzanine game.

    And lastly, the tracks are stacked, which is what we want! (see

    If they commit to stacked tracks on the new IDS, then they will be under pressure to build stacked tracks for the rest of the tunnel, and in SODO, and for an agency we fear will cheap out on rider transfer scenarios in the future, then stacked tracks is worth fighting for if it has a chance.

    1. I like the 5th stacked station idea. It does allow for level cross-platform transfers in one direction.

      I’m don’t think that a full mezzanine between the tracks is needed in that option. As long as the mezzanine goes under the current tracks, it can then terminate at a level platform for the bottom set of tracks. That would mean that only one escalator ride would be needed instead of two — and be cheaper and easier to build.

    2. IMHO, Sounder has potential and should be actually be emphasized over Link for longer trips. Tacoma to SEA is a good light rail trip. Tacoma to Seattle with the crawl through the Rainier Valley, less so! Therefore, it’s a shame that transfers are not a priority. It’s possible for long distance busses to be incorporated in to the scheme of things, kind of like NYC Port Authority station, too. Driverless vehicle dropoff bays too, if you’re bullish on that technology happening. Fact is the station complex will remain the busiest transit hub in the Pacific NW for a long time, and mode transfers shouldn’t be done ad hoc.

  4. As a future rider, I’m also interested in having an option that is politically palatable to the anti-tax austerity crew & ensuring the option taken minimizes their outrage machine’s basis in facts.

    It’s also worth noting that deep bore & mining actions in this part of the city can be … technically problematic… as Bertha found out when it ran into debris.

    1. The only option that’s politically palatable to the anti-tax crew is the repeal/voiding of ST3. They don’t care what option is chosen, they’ll scream about it being overpriced no matter what. They’re best ignored.

    2. Paul, I’ve read that out of ignorance or poor training ran herself into debris. But I think somebody at her controls knew there was a piece of metal jammed against her cutter and kept tunneling anyway. Good thing this kind of misplaced blame hardly ever happens in this industry.

      Mark Dublin

  5. The fact that a 200′ deep station is still on the table, two years after ST3 passed, tells you more than you want to know about the Seattle Process. I made pretty much the same comment as Al S. in a speech last month to the Washington State Ridesharing Organization. You may as well just skip directly to Midtown, than force thousands of transfers from a platform 50′ deeper than Beacon Hill.

    1. The “process” for this study is particularly flawed. The structural problem is that there is only a “stakeholders” committee and not a “users” or “riders” committee.

      Any other public investment would always have consideration of the users come first. Imagine if…

      … if the bicycle network had no design priority given to bicyclists.

      … if a sports arena design had no design priority given to the teams and the fans.

      … if a school plan had no design priority given to students.

      … if a playground plan had no design priority given to children.

      That’s how absurd this entire study process is.

      Some say that everyone is expected to be a rider, so that riders are theoretically represented on the committee. However, these stakeholders are primarily appointed to represent their organizations’ interests first — and riders second. I think that argument is thus very lame.

      I’ll keep asking for a Rider’s Committee review in this process in addition to a Stakeholder’s Committee review — even though the wall known as ST will continue to ignore me.

    2. It’s depressing. How can we get ST to recognize this basic fact. In my mind the most basic factors of a subway are good train-to-train transfers, frequency, speed, real-time signs, and reliability (both in the trains and escalators). The reason cities with comprehensive subways have less than 50% car ownership is it’s really easy to take a subway anytime, anywhere, and train-to-train transfers are faster and more convenient than train-to-bus or bus-to-bus transfers. So shouldn’t we do the same thing?

      Yes, we need a Riders’ Review Committee. Even if they’re nominally represented by the Stakeholders’ Committee, riders have such a direct experience with it and are critical to its success and it’s critical to their well-being, that there should be a separate riders’ review panel. There’s a panel reviewing the finances, and a panel of cities and larger employers and organizations, but there needs to be a panel of riders too. Especially when the board does not show great expertise in this area.

    3. Could it be, Mike, that given all the variables to date, the deep tunnel is best any engineer can also think of?


  6. I wish we could redo the current station to have the center platform we’ve long wished for, and then build the new southbound track further east, with a platform between the current northbound and new southbound track. Having the new northbound track a stairwell down from the current northbound and new southbound track would mean that all north-south transfers would then be no more than one flight of stairs.

    I wouldn’t worry about south-north as being a significant number of riders between the Chinatown stations, assuming decent south-north transfers at the Westlake stations — the most important transfer experience in the sytem. South-north transfers at Chinatown would be just to get between University Street and Pioneer Square Stations and Midtown Station, which can be done faster by walking than by taking two subterranean train rides.

    This two-center-platforms set-up would also make north-north transfers same platform or one flight of stairs.

    South-south transfers should mostly be happening at Westlake. But the ones that don’t happen there, at Stadium, or at SODO would be up one set of stairs and down another, and only impact riders who started at University Street, Pioneer Square, or Mid-town Station.

    Is there any hope of renovating to get center platforms?

    Oh, and deep elevator stations: -200

  7. I don’t have a dog in the fight as someone who rarely uses International District Station, but those deep platform seem like a big mistake for a busy station. It’s inconvenient enough at Mt. Baker with only elevators. For the price of all those elevators, can they just pay the merchants impact money while the 5th Avenue construction is going on?

  8. I don’t like this downplaying of Amtrak and Sounder. It seems to me if these services had convenient access to Link then more people might be interested in using them. Also will Link to Link transfers at ID be that much greater? Given Links transfer options at Westlake and Sodo, it would be interesting to hear what ST’s projections are for ID.

    1. We should give them some consideration because a multimodal transit network is a good thing, and one of the problems with the 20th century is that they often didn’t consider transit access to the airport or bus station, whereas in Europe the center of town is the train station and a multumodal hub. But there’s the rub: European train stations have several trains an hour to all over the region and country, so its usage is akin to the subway and hundreds of people are transferring to/from it every hour. Here, hundreds of people are transferring between Link lines and Metro buses, but only a few are transferring to Amtrak, Sounder, and the ferries because 80% of the time when Link comes there is no train or ferry to transfer to. That’s the problem with the F being detoured to the Tukwila Sounder station: most of the buses have no train, and none at all on weekends. The C detouring to Fauntleroy is the same issue (although I use it for the stop before it, Lincoln Park). The proposals for a second Tukwila Sounder station at BAR are equally flawed: most Link runs won’t have a train, and who would transfer there anyway, going from where to where? You aren’t going to transfer to Link going downtown or further because it’s faster to remain on Sounder until Intl Dist. Not many people are transferring from Sounder to Rainier Valley. And transferring from Sounder to the airport also sounds unlikely, especially when the Auburn and Kent Sounder stations have the 180 to SeaTac. So we need to keep these things in proportion, and not go next to Sounder and Amtrak simply for abstract multimodalism and a few people.

      1. I don’t know about that. ID is currently at 6000 and Sounder/Amtrak at 20,000+. With new station at ID there will be some cannibalization and some new passengers. Maybe 12,000 at most between the two ID stations. With these numbers there will be a hell of a lot of traffic (already significant numbers) crossing 4th everyday and I can’t help but think there will a significant part of it transferring between systems. Link offers easy access to the other DT stations along with Capital Hill and UW.

        It would be nice to know what the current numbers are for transfers though, and not something pulled out of thin air.

      2. Just because at BAR there may not be many transfers now doesn’t mean that in the future there won’t be.

        Too, we can’t really know how the future rail network wlll play out, and not putting a connection in at BAR will be a lost opportunity.

        Even from a redundancy standpoint let’s say, and I know this never happens, that there is an accident on MLK that blocks LINK from running south during the evening rush hour. If there were a hub at BAR the affected LINK riders could take Sounder south from downtown, and transfer to LINK at BAR. Again.. this never happens so I don’t know why I even bring it up.

    2. Link isn’t getting much further than Tacoma Dome. Maybe to Tacoma Mall if you’re lucky. Only Sounder could serve locations further out, as far as Olympia if you’re doing it big. (On the north side, it’s the only way to get to fast growing Marysville too–tracks already exist!). And we all know how silly it is to continue to depend on express busses for regional transport, given the increasing traffic congestion that will only get worse. Sounder remains an under-utilized commuter rail system, and should really get more attention.

  9. “…although in the long run the neighborhood would presumably benefit from a closer station”

    Not presumably, will. Sometimes kids don’t want to go to bed at bedtime but parents know they’ll have a better day tomorrow if they get their sleep. We are spending billions and billions of dollars to design a region-wide transit system that will hopefully be in use for the next century. These short-term local impact considerations are completely inconsequential and immaterial in comparison to the big picture and should have ZERO bearing on our initial design planning.

    That’s not to say that we shouldn’t do anything to mitigate the locate impact but it should be a separate, subsequent process. 1) Design the best, cost effective, future-proof station for users of the system. 2) Determine the impact the chosen plan will have to the local area businesses and mitigate appropriately.

    These ridiculous 200′ deep plans need to be thrown out and forgotten.

  10. I like the imagery in this post, in particular the scale of the 200 ft deep station options. It’s brilliantly subtle, yet the point comes across loud and clear, that super deep stations will be too far away from the existing station and that it’ll take way too long to get to the new station from either the ground or the existing station, especially if all but one elevator is out of service like at UW station.

  11. Given ST’s record of bending over backwards for ‘stakeholders’ (landowners) and screwing over riders, I have no doubt the 200ft deep option will be selected.

    1. If ST was in charge of building elementary schools, the playgrounds would be in caverns 200 feet deep and only reachable by elevator — dug out because the nearby landowners didn’t want the playground noise or the construction disruption of landscaping the playground. Of course, the whole district would be on the hook to pay for it.

      That’s how idiotic the deep station options are.

    2. Yes, except nimbys like schools and playgrounds: these and parks are the only non-residential things they think belong in single-family neighborhoods.

      I don’t think 200-foot stations are ST’s first choice. Beacon Hill was more about serving Stadium and SODO (rather than going around Beacon Hill), and initially there was no Beacon Hill Station planned. UW Station was going to be at the Burke Museum and maybe a second one near the University Bridge that might have been shallower.A deep station at Chinatown was not ST’s first choice, and ST’s official preference for West Seattle is still elevated. And it rejected the Queen Anne alternative for Ballard Link, and it has shown no enthusiasm for the 45th subway or Metro 8 subway. So in general, hardly any of ST’s original plans have deep stations. But community pressure for various alignments and the UW’s restrictions on campus routing led to more deep tunnels and potential deep tunnels. That indicates ST is deferring too much to “stakeholders”, rather than that ST wants to do them in the first place. Deep stations are also expensive, and ST isn’t that enthused about that.

      1. Don’t forget about the ridiculous zig-zagging on Pacific Hwy S to appease the concerns of fast food restaurants and strip malls. How can anyone look at this without laughing?

        The entirety of East Link is a whole other story of insanity.

      2. The diagram exaggerates the zigzagging. 99 and I-5 are five blocks apart at 208th, three blocks apart at KDM Road (and Link only goes two of them), and three more at 240th. So it’s not much more than the normal variation of swishing around obstacles.

      3. It could be one block, it would still be ridiculous. Link should run straight down Pacific Hwy S, period.

  12. I think it is wrong that ST is dismissing the Sounder ridership. To be clear, ridership on Link will dwarf Sounder; but this is a major transfer point for Sounder (the only connection between Link and Sounder) and a minor connection for Link. (Westlake is the major transfer point for link). There are several reasons for this:

    1) The downtown stations between Westlake and I. D. are pretty much identical. I doubt someone will bother with a transfer, no matter how easy it is to make it (things would be different if they built a stop at First Hill, but they won’t).

    2) Same and reverse direction transfers can happen at SoDo or Westlake.

    3) South Link to East Link will involve a small subset of the trips. For trips involving Rainier Valley, a transfer to Judkins Park makes more sense. For Tukwila and SeaTac, an express bus on I-405 makes sense. So much so that even for stations to the south (Angle Lake, etc.) it probably makes more sense to get off the train and transfer to a bus to downtown Bellevue (and might even make sense if you have to transfer again). So that leaves Beacon Hill to the East Side as the only significant trip on this segment.

    4) West Seattle to East Link is significant, but not huge. There will be some people from West Seattle taking this trip to the East side, just not a huge number. There will be far fewer going the other way.

    In contrast, this is the northern terminus for Sounder South, and the only station in Seattle. While a lot of riders walk to work from there, it makes sense that these folks are headed to other parts of downtown (Westlake, South Lake Union, Lower Queen Anne) as well as the UW.

    None of this means that we should build stations hundreds of feet under the ground (that would be stupid for a station like this) or ignore the significant number of Link to Link transfers that will occur here. But it would be a mistake to exaggerate the importance of Link to Link transfers, while ignoring the Sounder to Link transfers.

    1. I’m not going to quibble about about your skepticism east-south transferring, but I will say that anyone who isn’t on the 405 BRT but wants to go to SeaTac will almost certainly use Link instead of the 405 BRT. That includes riders going from Mercer Island, and in the Spring District, Overlake, and Redmond. (It probably extends to people from Eastgate and Issaquah too.) To use 405 BRT would mean a double transfer at BTC and at TIBS to get to SeaTac (405 BRT only goes to TIBS), compared to a single transfer at ID. That’s on top of likely less frequencies and reliability. Let’s also not forget that getting to South Renton Transit Center will mean that every 405 BRT bus is going to be spending an extra 5 to 10 minutes dealing with getting to and from that stop. Frankly, I don’t have high performance hopes for 405 BRT south of BTC given the current opeations proposal.

      1. I certainly agree for anyone already living along a Link Station. But, it’s less clear-cut for people who aren’t. Many of them, their first leg of any transit trip to the airport requires getting to Bellevue Transit, whether by bus or by Uber. Once there (by some means other than Link), you may as well hop on the bus if it’s coming first, of if the bus and train and coming through around the same time.

        Eastgate, neither of the transit options to the airport are all that great, but, of the two, Link feels like the least bad. For this market, the I-405 bus is severely weakened by the fact that you can’t get to it from I-90 without a long detour north to Bellevue Transit Center. Catching Link at South Bellevue P&R feels more “on the way”, even if it means going west to Seattle.

      2. An initial concept of the 405 BRT is shown in this presentation from September:

        The proposed design at TIBS looks like no buses coming north on Pac Highway (Seatac) will be able to stop at the eastbound platform. The ramp weaving with both BRT stops underneath the Pac Highway overpass looks like a big problem and it may not be operationally viable to build, and it’s on a WSDOT roadway so the initial design may never get clearance. In fact it’s so messy that I could see ST ultimately giving up on making a TIBS stop and instead routing buses directly to Seatac, working with the airport to build a looping/ reserve direction segment and stop at Seatac Airport Link

        I’m also a bit dubious that 2024 will be the opening date. The designs of these things is not yet adopted and we are already approaching 2019. Eastlink opens in 2023.

      3. I still don’t understand the Renton routing. Diagram #2 on the “Transit Speed and Reliability” slide seems to show new lane construction rather than the bus route. My general understanding is that the bus will get off the freeway right at the Rainier/Grady intersection, go into the transit center and out (crossing Grady twice), and then get back on the freeway. This will be slower than in-lane freeway stations, but still much faster than the current slog through Renton from NE 10th Street to SW 15th Street.

        I see the relocated Renton Transit Center is called the “South Renton Transit Center”. Is that just ST’s working name or the official Renton/Metro/ST name? Will Renton no longer have a “Renton” transit hub but only a “South Renton” one? Then it should be called “Renton”.

        “The proposed design at TIBS looks like no buses coming north on Pac Highway (Seatac) will be able to stop at the eastbound platform.”

        I assume both-way bus stops will remain where they are under the Link station. The biggest transfer here is between 405 and Link. Even some of the people from Pacific Highway will be able to come via Link.

        Look at those big beautiful red roundabouts (three of them) at 85th. +1 for aesthetics; -1 for the giant car-coddling interchange.

        “Branding: Creation of the BRT name and graphics.”

        So we are getting a new brand. Although the Swift station picture suggests at least some potential to argue for merging with Swift. Whatever name is chosen, I think it had better not have “BRT” in it. That would just add confusion over why with three similar services (Swift, ST, RapidRide), one is BRT and the other two are not.

      4. I-405 BRT has its flaws, clearly. But the 560 (a similar bus) does not carry that many riders, despite taking roughly the same amount of time as Link would to get from downtown Bellevue to SeaTac. If you look at the numbers from the 560, you can see that SeaTac does not dominate as a destination. The 560 is mostly a suburbs to downtown Bellevue bus, not a suburbs to SeaTac bus. The two SeaTac stops handle about 300 people a day, while the Bellevue downtown station does about double that. This could explain why ST decided to skip SeaTac — it just wasn’t worth it. They figured they could increase the already relatively high ridership from Westlake Village/Burien (which exceeds SeaTac).

        But there is nothing preventing ST or Metro from running an express from SeaTac to downtown Bellevue, with additional stops after (not before) Bellevue. That would compliment the new I-405 BRT, and provide a huge time savings over either the BRT or Link. Not counting the transfer, a trip from SeaTac to downtown Bellevue would take 38 minutes. That is competitive with a bus right now — even with the detour to Renton.

        Those on Mercer Island and South Bellevue would definitely take the train, but those are suburban areas — not destinations in themselves. There simply aren’t that many people headed to the airport, and for suburban riders headed to downtown Bellevue, there will be better options.

      5. ST gave the community two options for 405’s western terminus: SeaTac or Burien. The community said Burien because SeaTac has Link and Burien has no regional service except the soon-to-be-deleted 560. Going to SeaTac and then Burien is apparently too much of a detour.

      6. During heavy airport traffic, SeaTac is less of a detour and more of a time-sink that ruins reliability for the whole line, unless the BRT buses get their own lanes and bays at the airport.

  13. I’m not a structural engineer, but how about a variant on Shallow under 5th.

    Once the new station box is dug, undermine the Eastern rail path on the existing station, and add a new Rail path at this lower level.
    Bring the ST3 lines to the lower platform.
    Move the existing lines one space to the east.

    This would give a full over / under station with a central platform on both the upper and lower lines.

    The Western platform of the existing station is now available for “Specials”, or “Out of Services” trains that have not yet been taken to one of the OMF’s

    1. This would involve shutting down existing service to realign the DSTT to the new platforms. Not to mention re-laying the rails. There’s no way you could do that fast, it would be a multi-month downtime. We can’t lose IDS—really break the entire system—for that long, especially since East Link will be up and running by that point.

      1. Actually, since we will have two lines through Downtown, we could possibly assign all the train service south/east of Downtown onto the new line temporarily if switching tracks can handle the loads. The line from Snohomish / UW could be turned around at Pioneer Square temporarily although turning trains around quickly would need to be carefully managed.

    2. “undermine the Eastern rail path”

      Undermining Link is what ST is doing at UW Station with the unreliable escalators and elevators and distant bus stops.

  14. The best configuration is the one that isn’t on the table – all four tracks on the same level with three shared center plateforms between them.

    Put the outer tracks against the east and west retaining walls and make use of that center breakdown lane in the middle of the existing IDS station. Because a breakdown lane makes sense when you are talking about buses, but you don’t need it when you are talking about light rail.

    Keep the underground mezzanine for those transfers that aren’t cross platform, and extend it to KSS to facilitate those transfers too.

    And if you want to get real fancy, weave the lines so the NB and SB lines are next to each other, because North-North and South-South are probably going to be the dominant transfers.

    But it’s not on the table…..

    1. I don’t think it’s simple to widen the existing ID station for two sets of tracks, at least without serious disruption to current operations. If it is possible, I’d heartily support your three-platform proposal.

      The whole appeal of putting the station under 4th is that it would create a single station with a unified mezzanine beneath all three sets of tracks. That isn’t what we see in the 4th Ave cut-and-cover proposal, with the tracks placed where the unified mezzanine should go.

      If the original concept isn’t possible, then I might have a slight preference for 5th Ave, especially a three-platform proposal, with a mezzanine to King Street at a later date.

      1. I would gladly accept a 48 hour interruption with a bus bridge to join existing tracks. Especially in light of how much money would be saved by not having to dig.

        Go to you tube and look up ‘Chicago’s Five-Track Deval Diamond Replaced in Under Two Minutes’ They took 48 hours to get it done. I know ST could do something similar just adding a switch to an existing track.

      2. Ah, it seems I misremembered the diagram where I first saw the concept:

        Turns out the 4th Ave cut-and-cover proposal is pretty close, it just needs a tunnel from the mezzanine to the northbound tracks of the current line (or at least a direct, sheltered connection on the surface), and a connection from the mezzanine to King Street Station. Given those things, 4th Ave cut-and-cover is a pretty good proposal.

    2. This is exactly what I was considering when I was reading this post. Re-configuring the entire station a bit and taking over the middle lane\of the existing station to get more width. The biggest issues I could think of would be new elevators, platforms, and realigning a track (maybe both) which seems like a reasonable compromise for the scheme of things.

      I know there would be service impacts, but roughly in my head it seems like if they staged the construction with the new tracks first then fixing the old, it could be less disruptive.

      This solution seems clear enough, that there is likely something I don’t understand.

  15. I know it’s just a bit of romantic fancy, but I’d love to see the reactivation of the Union Station building as a part of any plan.

    1. It would make a nice regional bus hub. Incorporate Bolt, Greyhound, Dungness Line, charter busses, ect. Doesn’t have to be reactivated as a rail station — probably will never be, unless by some miracle high speed rail to downtown Seattle happens.

      1. Sound Transit is not going to kick themselves out of their own offices, nor would they want to put up with crowds. I agree though, Union Station would make a nice transit facility, and it is ironic that it is being used for offices for our transit agency (!) instead of for the public.

    2. Could have been a southern bus terminus and the new convention center basement could have been a northern bus terminus.

      1. I wish they would have done this at the convention center. It’s a shame the express lane connection will disappear, it would have made a great express bus terminal.

    3. Paul, your idea is-seriously- far and away best one tonight. A working railroad station with years of use behind it. Directly across the street from another one. Would hate to be the spokesman arguing that Sound Transit needs an office building more.

      But also think that the public-meeting group we need now has to put the public in the same room with actual tunnel engineers. Because Jackson’s history as a former beach for an actual lagoon will have far and away the most authority.

      We need to know what the ground around the pictures looks like, to see what we can do to drown fewer people. Shoreline Management Act, you know. Does “200 feet” mean anything specific, like where the real bottom of the lagoon (still) is?

      Remember under Seattle CBD, history and geology has handed us some underground things that could be useful, or useful to avoid or fix. Find 3D implications of the BN Tunnel from base of Belltown down to Jackson. Where the wheels of freight locomotives go over Link pantographs by about five feet.

      Really would like to see it become standard that the tech details come before the public first thing, and stay there throughout the project. Because it’ll make all the work a lot easier. With same amount to argue about, but finally on same as to the space-time continuum.

      Especially where we’ll have to keep the rescue diving bells and miniature submarines for when either climate change (forget Federal disaster funds for that one!) or Mt. Rainier’s present of a tsunami gives us our lagoon back, except this time with grass skirts, flower necklaces, ukeleles, alligators, crocodiles, sharks and Hawaii Five-O.


  16. Why can’t the new line share the tracks in the existing China town station? The buses in the tunnel has proven there’s capacity for more trains. The roads into and out of the tunnel can be re-purposed without any new digging.

    1. While it is possible that the tracks could be shared with improved electronics, Sound Transit has already determined that future capacity will require a second downtown tunnel. It’d be very difficult to go back on that decision.

    2. That’s a valid question.

      First, trains need more “spacing” than buses. That’s because trains are much longer and because trains take more time and distance to safely start and stop.

      As I understand how ST3 developed, it has to do with a few things:

      – With Ballard/SLU Link, ST did not design a way for those tracks to connect to the existing light rail tracks — not at Convention Place, not at UW or anywhere else. To do that now is pretty dang expensive and messy now since it would all be in a tunnel.

      – With the addition of the West Seattle Link, that would put three lines coming in south of ID. If ST was to meet its goal of 6-minute spacing on each line, the three lines would have a spacing of 2 minutes each. Given switching requirements and slow speeds through the switches ass well as occasional loading delays, having trains less frequent (say every 3 minutes) creates doubt that it could meet the forecasted demand. I’ll mention that the EIS for Lynnwood Link and East Link have both lines at 8 minutes at peak hours (4 when combined between ID and Lynnwood). It’s not clear when the 4/8 minute frequency won’t be enough capacity.

      In theory, ST could have looked at other capacity-enhancing options in developing ST3, but te public never got to see those options analyzed or discussed. That includes things like longer trains and platforms to allow for the 4/8 frequency, the operation of both West Seattle and Ballard as stub lines, a third line operation from West Seattle that would create a combined 2 minute 40 second frequency (maybe adding additional outer platforms jutting out at stations from the DSTT tracks so two trains could load at the same time) or other operational strategies that overcrowded systems around the world have had to build as overcrowding occurs. ST could have also considered above-ground options (like Chicago’s “el”) or at-grade options (like Portland’s Downtown Transit Mall one-way pairs replicated on Second and Third Avenue) — but they didn’t.

      That’s not to say that ST won’t have to come up with more money to get the full tunnel built. I don’t think the general public and even the elected officials fully understand that ST3 projects were designed and budgeted with only 10 percent contingencies even though the recommended industry standard is 30 percent contingencies at the early stages.

      New downtown subways next to several blocks of tall buildings aren’t cheap or quick to build. Light rail tunnels have to be bigger than traditional third rail tunnels (catenaries on light rail trains) which add to the subway cost. Right now, there is a light rail downtown subway under construction in Los Angeles and in San Francisco — both projects which have evolved over decades and have construction periods of about 8-10 years. Both projects have also seen major cost increases as well.

      With this in mind, I predict that ST will have to ask the voters for more money, and my guess is that the vote will be between 2024 and 2028. Of course, the opening of Northgate Link in 2021 and East Link in 2023 will probably so wildly embrraced that the voters may go along (although the subarea equity issue will be difficult to negotiate since there won’t be a general sentiment for expansion elsewhere). If that vote passes, we’ll get the subway. If that vote fails, ST will be forced to drastically cut costs somewhere — like forgoing a Ship Canal crossing, forgoing the last station in West Seattle or putting more of the track at or above ground through Downtown or SLU — or like revisiting if and how the trains could all be put into one tunnel.

    3. When East Link starts, both lines will run at 6 minutes peak, or 3 minutes combined in the tunnel. That’s the limit of the tunnel’s capacity without capital improvements. ST considered making those improvements in ST3 to bring the capacity up to 1.5 minutes, but rejected it when it decided to build the second tunnel. That was a good decision I think. There has been concerns about Link possibly being overcrowded between downtown and the U-District, and it’s better to have spare capacity than to get overcrowded. The second tunnel will offload the considerable Rainier/SeaTac/Tacoma/SLU/Ballard traffic. Also, it prebuilds capacity for a future fourth line. We were incredibly lucky that the DSTT was built in the 1980s when costs were lower. That meant it didn’t have to be included in the ST1 budget, and that may have been a factor in getting ST1 passed. Now we can do the same again.

      1. Mike, here is the Address of Lynnwood Link’s latest project folio that says 4 minutes:

        Here is where ST says that a train will go to Downtown Redmond at 8 minutes:

        Certainly, ST3 says 3/6 minutes. Still, it’s not clear what year the higher frequencies are supposed to begin.

      2. Ah yes, I forgot that six-minute peaks was a temporary measure between U-Link and Northgate Link to accommodate capacity needs until Northgate Link opened. This was decided back at the beginning of U-Link (or was it earlier when the Ride Free Area ended or something?), before they know they would also need three-car trains and the cars that haven’t arrived yet.

  17. Thank you for the insight into this!! I also don’t think forcing thousands of transfers via elevators is the best way… but if necessary, it can be done right with spacious elevator lobbies and fast over-sized elevators. Covent Garden in the London Underground sees 40,000+ riders a day and the only access is 4 elevators or 193 stair steps.

    For elevators, I don’t think you would have a first stop at mezzanine and a second stop at ground level. I think everyone would have to get off at the mezzanine and either make their way to the transfer line or make their way to an exit via escalator or another elevator.

    Also, how about high speed rail? Are they future proofing for that?? :P

  18. Martin, this is a great post, and of course I think a correct one. The 5th Ave option is better across the board because of the volume of LRT transfer activity that will take place here. While Ross may be right that Westlake will have more, this is the transfer for eastsiders going to/from the airport, so a very effn big deal.

    The problem ST needs to solve therefore is construction mitigation. They have a couple good stories to tell about building underground facilities in places that manage and mitigate disruptions. Perhaps they haven’t done a good enough job describing how this would work for the neighborhood.

    The best example is the Pine St stub tunnel, which was built as part of the initial segment to allow trains to reverse direction after serving Westlake. This was a cut & cover tunnel directly under Pine St in front of the Paramount Theater, in one of the busiest parts of town. They did some limited closures of the street so they could scrape the surface and build a temporary bridge deck to carry the street. After that, all the construction took place beneath the surface out of view. The job was staged thru the tunnel entrances, sparing the folks on the surface from noise, dirt, traffic, etc.

    The Paramount theater was as close to the job site as the new buildings along 5th. It stayed open throughout the schedule and ran a full slate of shows.

    The same could be done at IDS. The primary impacts could be isolated to the inital few weeks while the street surface is lowered and decked over. And the job could be staged from the southern tunnel portal. I don’t know if this is how ST plans to do it, but it’s what the city and the neighbiorhood should be asking for.

  19. If we want to direct northbound transfers to one station and southbound transfers to another, we’ll have to think about how to communicate that on the map. I think I’ve seen this in one or two cities but I don’t remember clearly. Usually maps have a box or connecting lines marking the group of possible transfer stations. Then you’d want something additional to show the recommended station. Maybe it was BART where I saw it, because I vaguely remember seeing something MacArthur for northbound transfers and Lake Merritt for southbound transfers, but I don’t remember what it looked like.

    (And I’ve gone between El Cerrito and San Francisco on Sundays when there’s no through train, and found that MacArther has a timed transfer with the other train waiting across the platform or coming in a couple minutes. That’s the kind of transfer quality we should emulate.)

    1. Notably, MacArthur isn’t in Downtown Oakland and is above-ground. The trains from the two lines going in the “same direction” are stopping across the platform from each other. Train drivers can see if people are walking the roughly 20 feet across the platform before closing the doors — not very possible if two lines are on different levels.

      That’s exactly the setup I’d like to see at SODO. It’s the next best thing to having a UW/ Northgate/ Snohomish train go directly to Rainier Valley/ Seatac/ Tacoma. It’s much cheaper and easier to build there (no tunnel) and gives enough spacing in the schedule to accommodate the occasional train running two or three minutes late (not practical once East Link trains are also operating on the tracks).

      Sure East Link riders can’t take advantage of it but they’ll need more time to get from one train to the other anyway because the anticipated station layouts make timed transfers inefficient.

    2. I don’t think you have to do anything. Transfers *can* happen at every station, it is just that some transfers are better at some stations and worse at others. Experienced travelers figure it out, everyone else just muddles along. It is no different than any other public transportation system, or for that matter, any transportation system. I know the best way to drive to work, and I know which bus to take. It may not be what Metro or Google recommended, or the obvious one on the map, but having done it a few times, I can tell it is better.

      For what its worth, I have yet to find a good map of the exits to Westlake Station, despite the thing being one of the oldest (and best) transportation investments in the region. I remember my commute would occasionally go through Nordtrom’s (or maybe it was Frederick and Nelson’s) and I would know exactly how to walk through the clothing department. Now I feel like a tourist every time I go downtown, stumbling around until I get to the surface, where it then takes me a while to get my bearings, and walk to my destination. That is just life in the big city, as they say.

    3. Hong Kong MTR has good maps for this. Many transfers are twinned across two stations with paired cross-platform transfers (Hong Kong’s narrow geography helps). The map uses concave and convex lines to denote the best transfers.

      The system (and it’s map) became much more complicated a few years ago when the MTR merged with the KCR (a rapid regional rail system akin to Paris’ RER); however, the MTR’s original Red Tsuen Wan, Green Kwun Tong and Blue Island Lines–as well as the much newer Purple Tseung Kwan O Line–all use the classic twinned cross-platform transfers.

  20. Easy transfers should be the #1 priority. It really pisses me off when we make decisions to save a little money or temporary inconvenience in exchange for permanently handicapping a system, or at least an alignment, that will (should) last 100+ years, maybe a lot more. People like Bruce Englehardt have called me anti-transit or anti-rail because of this when in fact I more pro-rail than he apparently is. I am for well designed, effective rail, not half-assed crap rail.

    Also, I go to the international district quite frequently, and generally I try to avoid 5th. It isn’t an essential street for the neighborhood by any stretch of the imagination. There are easy detours to avoid it. We can put up with losing it for a bit for a much better system.

    1. I am all for well designed just as much as you are but I disingenuous to destroy peoples livelihoods because you don’t like a street. It’s not what’s businesses that are on the street, but how many people have to cross that construction zone to visit businesses on other streets beyond it as well. You essentially divide an economic district in half for the length and some breadth as far as the road is impassable.

  21. I don’t understand why the 4th Ave cut and cover is so deep. What about moving the new link platforms under 4th up to roughly the same level as the Sounder and existing ID station elevation? Then there could be a connector corridor underneath and connecting all platforms (Sounder, Amtrak, old link, and new link). Super simple to understand for transferring riders. To transfer you go down and circulate east or west follow signs until you get to the correct platform and then go back up, just like numerous european stations. The new train tubes would have to snug up next to the old BNSF tunnel to get over top of the DSTT (the DSTT ducks under the BNSF tunnel at about 4th and Wash.).

    I don’t undertand how they make either cut and cover station (under 4th or 5th) to work with the existing link rails and the East Link overpasses south of the DSTT entrance.

    1. This section directly beneath 4th is mostly open already. between the lightrail line and the sounder lines is a giant underground parking garage that surrounds the older of union station to the entrance beneath Seattle Blvd, a bridge built in 1901. That area beneath 4th Ave may be owned by BNSF.

  22. I haven’t read all the comments, so someone may have pointed this out. What does having the railhead 200 feet deep at IDS do to Midtown?

    It will already be deep even with the cut-and-cover design; the necessity to under-run the DSTT trackway wil put the railhead three stories down there, and Madison is at least 45 feet higher at Fifth or Sixth than is Pine.

    Sure, a mezzanine could consume 18 to 20 feet, but you’re looking at nearly 100 between the railhead and street level if the tracks stay level.

    Light Rail Trains can manage gradients of about 6% in dry tunnels, and it’s about eight blocks or 2400 horizontal feet between a platform whose south wall is at Marion and one whose north wall is at Jackson. That gives you about 130 to 140 feet of potential vertical change, just about the vertical difference of a platform three stories below Pine and one two-and-a-half stories below Jackson. This assumes Martin’s “in-the-middle” mezzanine, which is brilliant.

    So a deep-mined platform at IDS would require a Midtown platform at elevator-only depths as well.

    That is completely unacceptable. It would be insane to be blunt.

Comments are closed.