Sound Transit’s ridership forecasts suggest that 15 trains per hour (every four minutes) between Northgate and the International District are sufficient for peak demand on that segment through 2030. This allows a train roughly every 8 minutes to Des Moines and every 8 minutes to Redmond.

Of course, forecasts are hard, and factors completely outside Sound Transit’s control might cause ridership to dramatically exceed or undershoot those forecasts. Things that Sound Transit does control, like Sound Transit 3 investments, would likely increase demand on this segment and could, optimistically, be open by 2030. Furthermore, the safest assumption is continued population and ridership growth after 2030. So what capacity is supported by the infrastructure under construction now?

In an effort to cut through some of the urban legends surrounding this question, I spoke with Marie Olson, Sound Transit’s Link Transportation Manager for Operations. As it turns out, Sound Transit’s signaling system is designed for a minimum 90 second headways. Ms. Olson believes that if the demand were there, 3 minute headways are achievable without significant deterioration in trip quality. SDOT thinks its signaling system on MLK’s surface segment can support 6 minute headways, so 3 minute headways neatly divides into 10 trains per hour to both the Rainier Valley and the Eastside. Planned fleet purchases are large enough to support 4-car trains at this rate.

Although going below 3 minutes is possible, due to the variability inherent with human factors and surface operations it “wouldn’t give our ridership as reliable a service.” The small windows to fit in delayed trains might cause them to bunch up, delaying riders. Furthermore, it would likely require additional investment in Traction Power Substations.

There’s also the question of where to put the trains. At 30 trains per hour, 10 of them can go neither to Bellevue nor the Rainier Valley. It’s possible that the pocket track planned for International District might be able to reverse trains, although it’s not really designed for that movement during peak operations. Alternatively, the extra trains could turn at Stadium, if Seattle were to accept a lower traffic level of service on Royal Brougham Way (or close it entirely). “It’s something we’d have to look at in simulation,” Olson said.

I specifically asked about the deleted Montlake vent shaft, often cited in STB comment threads as a constraint on tunnel throughput. Ms. Olson denied that it was one; in fact, when ST deleted the shaft they placed a signal at the midpoint between Capitol Hill and UW. This allows two trains in the three-minute segment between the stations, which means it has the same fundamental 90-second limit as the rest of the signaling system.

At Tokyo-style crowding levels, 20 trains per hour is 16,000 people per hour per direction. Of course, at that level of crowding headway limitations may be driven most by the dwell times needed to deboard. Here’s hoping that the system is popular enough to require that level of service.

184 Replies to “Capacity Limitations of Link”

  1. So can we now seriously discuss building a West Seattle – Ballard line without first spending billions on a new downtown tunnel? 3 lines at 6 minute frequency each is 2 minute headways, even at 5 minutes each it’s 1:40 minute headways (still above 90 seconds).

    I’m not saying that a new downtown tunnel won’t be needed eventually, but we should at least consider the option and weigh the pros and cons of doing the branches first. (I feel like ST has always implied that to build the new lines they need a new downtown tunnel).

    1. Exactly this. I fail to see how a Ballard line can’t fit in the current tunnel, as some of us have been saying all along. Why doesn’t this post address that question?

    2. +1
      A second downtown tunnel is great, but right now a lack of east/west mobility is needed more. I vote for Ballard Spur and SLU Gondola!

      1. Actually there will be a vote. Both directly when/if ST3 goes to the ballot and indirectly when the various ST board members are up for re-election.

      2. Ha, transit experts. You crack me up Lazarus. Hey, did you see the video of Brenda last week, it was boring! Get it, boring …

        Oh, wait, you are serious. Transit experts that forget to put in a station where Link intersects 520, or transit experts who can’t decide whether a station at NE 130th is a good thing or not. Transit experts my ass.

        This system is run by politicians, who have weird political fantasies (a spine will solve all our transit problems).

      3. @CS,

        Yes, there will be a vote, and that vote will be an up or down vote on whatever the transit experts and ST board decides to put in ST3. And right now both the technical experts and political pressures are aligned in near perfect unison against the Ballard Spur. It’s a near certainty that it won’t be built. Rapid Ride maybe, but LR no.

        I know this is a fan forum, and enthusiasm for transit is always good, but there really isn’t any good technical reason to build the spur.

      4. @Lazarus

        Where do you get your information on what the technical experts are recommending? [citation needed]

        As for the politics I know the city has been pushing for rail between Ballard and Downtown. Dow and some other political figures have been pushing for rail to West Seattle.

        Unfortunately politics in Olympia mean only the route between Westlake and Ballard via Interbay or one of the two streetcar options is affordable under the likely ST3 budget for North King. Option D (Queen Anne Tunnel) which seems to have a lot of support is out of reach.

        Dow is even more out of touch as there is no good reason West Seattle should get rail before Ballard. Furthermore given the cost the line will have to be so value engineered as to be an utterly worthless dinky shuttling between Delridge and SODO.

        Keep in mind that when I talk about Ballard/UW I mean as a separate line or eventual continuation of a Ballard/Downtown line as studied by Sound Transit last year (with a few more stations though), not as a spur off of Nothgate Link. I agree any such notion of a spur interlining in the U District is dead on arrival.

      5. @Lazurus — A subway line from the UW to Ballard would connect more people in a better way and be cheaper than any other route from Ballard to downtown. Is that technical enough for you?

    3. The Ballard Spur will not be built. It is an orphan line that would effectively block a second tunnel and expansion northward.

      The real goal should be Ballard-DT-WS in a second tunnel.

      1. It would only block a second tunnel if we were stupid enough not to design for the future. Oh wait… we pretty much always are.

      2. How the heck is Ballard/UW an “orphan line”? How exactly does it “effectively block a second tunnel and expansion northward”?

        Now I know some people like to call Ballard/UW “the Ballard spur” but I do not believe if it is built there will be any track connection between the line and the Northgate link tunnels.

        What this does tell us is passengers transferring from a Ballard/UW line or 520 buses won’t overload link between Westlake and UW or Brooklyn stations.

      3. As a further comment the funds in ST3 are limited. You can only go to Ballard from downtown if you build option B or one if the streetcar options.

        A half assed line to West Seattle might be possible, but it would eat all of the North King funds. In no sane world should West Seattle get rail before Ballard.

        Assuming Federal grants both a second DSTT and a Ballard/UW line are possible with likely ST3 funding levels. In addition using the monorail tax authority the City could fund either of these itself if ST3 fails to pass either with the legislature or the voters.

      4. Here’s another “orphan line”, the Sheppard Subway in Toronto. It has a similar length to Ballard-UW. It has 5 stations. It feeds into the main N-S subway in the north end of town and ends in a suburban shopping mall. It carries 50,000 a day.

        If Link does overflow as a result of a crosstown line, the logical solution would be to build a second downtown line but don’t use it as an excuse to hamper E-W connectivity. Would we shoot down a Denny subway for the same reasons? Hell no.

      5. A Ballard spur would carry roughly as many people as the north end of Link. It makes sense to split the line like so: https://www.flickr.com/photos/67869267@N07/9152772373/in/photostream/

        Every six minutes from Northgate (or Lynnwood or Everett) sounds good to me. We can get that to four with a little more effort (as the article mentions). That is great. That makes the transfer from the north really good, while carrying plenty of people. Pair the north end with SeaTac, and the Ballard line with the east side. This is a logical pairing, since one route goes mostly north/south and the other east/west. It also makes sense since folks from the far north and northeast will go the other way to the east side. It makes sense to go from Everett to SeaTac or Bothell to SeaTac via Link, but it doesn’t make sense to go from Everett to Redmond or Bothell to Bellevue via Link.

        But if you don’t interline the routes, then it isn’t the end of the world. Just minimize the distance that a rider has to travel to make that transfer and then time the trains accordingly.

        Either way the argument that we can’t build a Ballard to UW line because “we can’t handle the load” has just been blown out of the water. Great news indeed.

      6. Any Ballard – UW line will not be a spur off the spine. Quit calling it the Ballard ‘Spur’.

      7. @aw — Why not? Seriously, why not make a spur (https://www.flickr.com/photos/67869267@N07/9152772373/in/photostream/)?

        How can folks say, on the one hand, that we need a brand new, very expensive line connecting Ballard to downtown (or even West Seattle to downtown) but then turn around and say we can’t afford a spur? Either the Ballard area is big or it isn’t. If it is big, then it sure makes sense to connect it to the UW and downtown, via a spur. If it isn’t, then why are we even talking about building a light rail line from there to downtown? This just doesn’t make sense to me.

        Without a transfer, the UW to Ballard line (https://seattletransitblog.wpcomstaging.com/wp-content/uploads/2014/06/ST_Link-RGB-Lines.png) is still by far the fastest way to get to downtown from Ballard, let alone every spot along the way (as well as pretty much the entire northwest part of Seattle, indirectly via bus first). With a spur it is faster.

        What is the argument against the spur? That folks at Roosevelt and further north can’t handle six minute frequency? That’s crazy.

      8. It won’t be a spur because they won’t open up the tubes to put in a junction. They especially won’t do it under UW land as shown in your drawing. I also wonder what the vertical profile of an E-W line has to be in order to cross I-5. Does it go deep under the express lanes, or does it pass above the mainline and try to stay out of the way of the NE 45th and NE 50th ramps?

      9. I don’t see why they won’t “open up the lines”. Cities do that sort of thing all the time. Your point about the UW is well taken, so the connection point is moved east (there are a bunch of places where this could be done).

      10. By 45th, the express lanes have already returned to level with the general lanes.

        The highway cut isn’t deep at all at this point; dipping below there on an already-deep-bored line many blocks from the nearest station isn’t even in the same realm of complexity obstacles as, say, another Ship Canal undercrossing.

        As for the junction diagram, that is of course just one possible arrangement of many, but I would note that I took pains to ensure it would be feasible with curves as gentle as the main line’s, all work access from the Brooklyn station vicinity or from the public sidewalks of 15th, and a physical meeting point so deeply below (preserved, permanently empty) UW land that no construction rumbling could ever be felt from the surface lawn.

      11. Chris,

        The must be a connection between North Link and a Ballard-UW line built without Ballard-Downtown as people are suggesting. Even if you somehow find a place along it to store the trains along it so that they don’t have to travel to the Maintenance Facility to enter and leave service, they have to be able to get there for heavy overhauls. It would make no sense to have a heavy maintenance facility for a five mile line.

      12. Anandakos,

        Actually you don’t need a track connection, you can load the cars on a truck and drive them to the main heavy maintenance site. I believe this is how LA deals with heavy maintenance for at least one of its lines.

        Given how infrequently heavy maintenance needs to be done it really isn’t a big deal.

      13. Each LRV has to cycle through the Operations and Maintenance Facility at least once a week, based on my previous conversations with ST reps. If the trucking is an expensive or dangerous operation, that could be the Achilles Heel in any line that has no rail connection to the spine.

      14. You need an O&M facility capable of routine maintenance accessible with a track connection. Heavy maintenance is a much less frequent occurrence.

        Essentially you need a Bellevue level facility directly accessible, not a SODO level one.

        For such a short line the facility need not be huge, even if trains run at fairly high frequency.

      15. I agree Sound Transit won’t open the Northgate link tubes to build a junction.

        The notion of a track junction in the U-District is a distraction from the essential utility of an E/W line between Ballard and the UW.

        I suspect some people tune out when they hear ‘Ballard spur’ because the notion of a junction is such a nonstarter. The spur notion also may be what Lazarus is on about when he claims an E/W line would ‘block’ a second N/S line.

        The real issue here is Ballard may end up with nothing more than a crappy and slow streetcar if the ST board can’t be convinced to take a serious look at Ballard/UW as a separate transit line.

      16. Good point Chris. They really are two different issues. First build a line from Ballard to the UW, then worry about whether the lines actually interline. Similar lines throughout the world interline that way, but somehow we can’t. Whatever. It isn’t the end of the world, as long as the transfer is fast. Mainly, it is crazy to me to build a subway line from Ballard to downtown that bypasses the UW. That seems nuts to me, but apparently, that is what a lot of people (like Lazurus) think makes sense because, uh, capacity!

        I can’t help but think that Lazarus is simply looking out for the poor Lynnwood rider, that might be inconvenienced by a reasonably successful subway system. Six minute frequency? Outrageous, I thought we were going to get four! Build the Ballard to downtown line via the UW, the cheapest and most effective way? Why, that would be horrible — I might have to sit next to someone on the way back in the evening.

        But what Lazurus and his like forget is that the Lynnwood rider might actually be headed somewhere besides downtown. The only guys I know from Snohomish County are like this. They work in places like Fremont, or Lake City. So of course they want a quick route to Ballard, and a stop at NE 130th. For them it doesn’t matter at all if these lines interline or not — all that matters is a quick transfer, because they will be going the other way. By the way, is the Husky Stadium Station or the Brooklyn Station center platform?

      17. @Chris Stefan,

        ST reps told me LRVs would have to go to the SODO O&MF once a week. The Bellevue facility will be an operations & storage facility. (This was one of the excuses for putting a turn-back track rather than a platform in the middle of ID Station.)

      18. Wait, I can answer my own question. From what I can tell, both stations are center platform. This is great news. The simplest, cheapest thing to do is just pop another hole in the wall and interline the routes north of Brooklyn Station.(like so: https://www.google.com/maps/d/edit?mid=z-ZcpzpzqRA0.k3CL2lbERZ8Y). Those heading to or from downtown just keep going. But those making the transfer just do so at the U-District station, via the center platform. Sounds good (really good) to me.

      19. Brent,
        Oops, I was under the impression the Bellevue yard was doing more than just being an operations base and storage yard.

        Also note how small the O&M facilities for most modern streetcar systems are. It is possible to build a fairly compact and inexpensive O&M facility as long as the number of cars is relatively small.

      20. Ross,
        Creating a junction is much more complex than just “putting a hole in the wall” especially considering the historic Neptune theater is just to the north of the station box. The site is very constrained not just by the Neptune but a historic apartment building to the South, UW tower to the West, and buildings along University Way to the East.

        The u-district junction idea is a non-starter and is a good way to get the professional staff at Sound Transit along with local political types to tune you out.

      21. Ross,

        To expand further I see Ballard/UW and Ballard/Downtown as separate lines with separate markets. I think we need to build both and both will have decent ridership if built.

        In addition the Ballard/UW line is something we can afford with a city-only measure or as part of ST3. As you are fond of pointing out it would improve downtown access for a wide swath of North Seattle including Ballard.

        Ballard/Downtown would serve Belltown, Lower Queen Anne/Seattle Center in addition to offering a faster ride to downtown for Ballard and additional capacity across the ship canal. There is also the possibility of serving upper Queen Anne, Fremont, Crown Hill, and/or Interbay.

        There is some redundancy between the two lines, especially if something like option D is selected for Ballard/Downtown. However I believe it would be a good investment to build both nonetheless.

      22. I agree completely with your last post. They are two different lines. I said as much down below, in defense of WSTT. I think those two projects, separately, make the most sense for Seattle (as I said in a different comment).

        As to the various options, I plan on writing a little post about it. But basically, we have three choices:

        1) Interline somewhere between the Brooklyn (U-District) station and the Roosevelt station. This is probably the cheapest solution. There are lots of places where the work could be done (there is roughly a mile between stations). Wherever we dig will likely be a lot cheaper and a lot easier then where we dug to build the Brooklyn Station (next to the Neptune and next to the biggest building in the U-District).

        2) Just go under and build a platform. This seems to be what lots of people want. This would be fine, but for Heaven’s Sake, make the transfer painless. If the transfer is like the one mentioned in Toronto (a thirty second transfer) then that would be great.

        3) A spur junction that is a combination of the two. This is trickier and more expensive, but would probably pay for itself over time. This would allow Sound Transit complete flexibility when it comes to routing. For example, during rush hour, the trains go to Lynnwood every three minutes, and folks from Ballard have to transfer. During the middle of the day, the trains go to Lynnwood and Ballard every eight minutes, and folks from Ballard don’t transfer. Trains would travel through the core (UW to downtown) every four minutes. It also means that trains go every eight minutes to the airport or Bellevue. Basically it means we don’t have to keep sending half empty trains to Lynnwood, just so we can have good frequency in our core or decent frequency on the east and south end.

      23. Ross, why don’t you just give up and admit that the only option that ST will consider now that Northgate Link has been mostly designed and is under construction is your #2.

      24. Assuming Ballard/UW is ever built #2 is the way to bet. I’d guess the station would be under 45th or 43rd. Hopefully connecting the two mezzanines won’t be too hard.

        Building the platform in such a way as to enable platform to platform transfers would be nice but I don’t think critical.

      25. Yeah, silly me. After spending billions on a light rail line that so far is hardly used, I guess it is crazy to think that they would spend a few million on a common, quite simple system to save thousands of riders thousands of minutes each day. I forgot who I was dealing with. Anyway, I look forward to the new awkward, inconvenient, but quite grand and spacious new set of stations.

    4. To get decent headways on any additional branches you would need to go below 3 minutes in the central section which comes with its own set of issues.

      At 20 trains per hour you only have 5 trains per hour over what is planned when ST2 is fully built out. This means 12 minute headways to Ballard and West Seattle. This also means you have no capacity for shorter headways without impacting reliability or adding new traction power substations.

      1. I’m not sure if I follow you Chris. I’m thinking:

        8:00 Ballard to Redmond
        8:03 Northgate to SeaTac
        8:06 Ballard to West Seattle
        8:03 Northgate to Redmond
        8:09 Ballard to SeaTac
        8:12 Northgate to West Seattle
        8:15 Ballard to Redmond


        Not that I think this is a great idea, but this means six minute frequency on the Ballard and Northgate ends, three minute frequency in the core (UW to downtown) and nine minute frequency for the south, east and West Seattle. I don’t see a twelve there anywhere.

      2. Ross,

        I’m assuming the north branch is at Convention Place. I sincerely doubt Sound Transit will build a junction in the North link tunnels. The only “easy” spot for a junction is where the pine street stub splits from the DSTT.

        I also assume 6 minute headways between Westlake and Lynnwood won’t be enough. ST is planning for 15 trains per hour at full ST2 build out. That only gives you 5 trains an hour to send to Ballard and 0 additional capacity for growth unless you accept the issues involved with headways below 3 minutes.

        5 trains per hour is 12 minute headways.

      3. OK, I think we are talking about two different things. I’m assuming “Ballard to downtown” means Ballard to the UW (https://seattletransitblog.wpcomstaging.com/wp-content/uploads/2014/06/ST_Link-RGB-Lines.png). So, here is the junction: https://seattletransitblog.wpcomstaging.com/wp-content/uploads/2014/06/ST_Link-RGB-Lines.png

        >> I also assume 6 minute headways between Westlake and Lynnwood won’t be enough.

        Right, if that is the only way to get from the UW to downtown. But with the spur, you will have three minute frequency between the UW and downtown, and six minute frequency from Lynnwood (or even Everett) and downtown. That is plenty.

        The UW to downtown is our core, and everything else is really a spur. If those trains are near capacity, it won’t because of folks from Lynnwood. It will be because of people from the U-District and Capitol Hill.

        Put it another way, imagine a train that is essentially North Link, but it has “exit only” stations in the U-District and Capitol Hill. Now, you are telling me that this train, with four cars, needs more than six minute frequency? I don’t think so.

      4. Ross,

        Others mentioned the possibility of branching at convention place rather than building a second transit tunnel downtown.

        Due to loads between Brooklyn and Westlake stations such branching wouldn’t work without increasing frequency to less than 3 minutes.

      5. Right — I agree with you. Branching at convention place probably wouldn’t work.

        But there is no need for increasing frequency to less than three minutes, as my schedule states. Either way you have three minute frequency along the core of our system (UW to downtown). That will carry the bulk of our riders. North of there it will trail off very quickly (especially for riders staying on the train).

        In other words, either way you carry the same amount of people from Brooklyn to Westlake. Identical! It is just that half the trains would be coming from the west.

      6. Ross,

        If branching happens North of Westlake and South of Northgate Convention Place is the only place it is likely to happen.

        I know you are obsessed with the idea of a branch in the University District, but that ain’t gonna happen. In fact if you ever want to see a Ballard/UW line talking of a ‘spur’ just hurts the cause.

        Indeed the association is so strong in some minds they assume one is advocating for the spur even when it is for a separate line with a transfer.

        If we want to avoid crappy and value engineered rail to Ballard or even worse no rail at all then advocates for Ballard/UW need to be taken seriously and not dismissed as a bunch of transit enthusiast nutters.

      7. A fair point, except…

        FWIW, I stopped loudly advocating for a branching structure a while ago, though Martin’s capacity research tempts me to start again. With everyone now admitting the peak-load-defining need will be between downtown and the U-District, to which throughput demand everywhere else will never hold a candle, it becomes almost insane to argue against branching! Every train sent northward, by definition, will be well under capacity. Forever. As often happens with long rail spines whose primary concentrations of demand sit concentrated in the middle.

        Branching, as a concept, is only a problem when you risk running individual branches too infrequently to be useful. When Jarrett Walker warns about this risk, he is speaking to the errors of Sacramento and Dallas and Baltimore, not to the Bostons or New Yorks or Londons that make extensive use of high-frequency branching for very-high-frequency trunks.

        When an east-west “spur” — which would happen to match service levels to need spectacularly well, in addition to being more convenient for just about everyone — has been debated, the stated “objection” has rarely involved service splitting or northward capacity anyway. The consternation always stems from a fear of junctions causing “service unreliability” at the junction point”. That fear has always been unfounded.

        Successful high-volume merger precedents are legion, from the Brussels Premetro (an unending stream of merging light rail vehicles, with level junctions!) to Paris’s RER (massive commuter trains, headways as low as 90 seconds). Seattle objections were bullshit when the minimum headway was expected to be 2 minutes. They are super-bullshit when the headway is 3 minutes. Branching will better match supply to demand where needed, and that truth cannot and should not be “dismissed”.

      8. Furthermore, you should probably heed your own advice when you go on about closed systems with vehicle extraction pits for serious maintenance. The whole point of this line is how busy and multi-faceted and uninterrupted this mere 3-mile corridor is. And now you’re talking about installing deep-level extraction pits along it?

        Unlike branching, that approach is rare in the world for good reason. A track connection — even one never to be used in revenue service — is always preferable, and here would be obligatory.

        I agree with you that there’s no reason not to store trains overnight, and to perform inspections and minor work on them in-tunnel. But for serious maintenance, there’s going to be a track connection of some sort. What ST needs to get over is its irrational fear of those. But pushing for extraction pits and total maintenance isolation is going to get you pegged as the nutter.

      9. I’m not obsessed with the idea of interlining, I just think it makes sense. If I had to pick my priority of projects for Seattle, I would go with:

        1) UW to Ballard light rail
        2) The WSTT
        3) NE 130th Station
        4) Making the two lines (North Link and UW to Ballard) interline
        5) The Northgate Pedestrian Bridge over I-5

        The arguments against interlining are weak. It costs a bit more (maybe) and it cuts frequency just a bit for folks north of there. But as we all know (based on the studies as well as similar cities throughout the world) that will be a small proportion of the folks who come into downtown from the north. Meanwhile, d. p.’s suggestion is simply more expensive. It does nothing to the folks from the north (the interlining would only occur when the trains coming from the north drop their frequency anyway). Using his approach, you simply allow for interlining if you want it. But again, this is by no means a requirement, any more than the Northgate pedestrian bridge is a requirement for Northgate Link.

        Personally I think it strengthens the case, as it makes the idea of a bypass (as I’m calling it) pretty silly. Going from Ballard to downtown and skipping the UW is crazy, but one argument is that it might be just a little bit faster, because of the transfer. Take away the transfer and it just seems silly to avoid the UW.

        But I think you exaggerate the effect, or even the meaning of the word “spur” to the regular population. Conceptually, they are the same, whether there is a transfer or not. For example, I would call the Sheppard line (the purple line in Toronto) a spur, even though (as explained below) the two lines don’t interline. Again, that is really no big deal. If you look at what I wrote (https://seattletransitblog.wpcomstaging.com/2014/06/30/ballard-uw-should-be-the-next-light-rail-line-in-seattle/) there is only one paragraph that talks about interlining, and it makes clear that a simple transfer is just fine (although I don’t like the way that paragraph is written — I want to change that, especially in light of the information presented here).

        Meanwhile, the “Ballard Spur” pages imply no interlining at all (http://stb-wp.s3.amazonaws.com//srv/htdocs/wp-content/uploads/2014/06/Ballard-Spur.png, https://www.facebook.com/BallardSpur/photos/a.235444256516652.57688.235429863184758/253404268053984/?type=1&theater). That second picture implies the opposite. The second picture implies heading off to Children’s. But I wouldn’t say that idea (an idea I think is unnecessary) kills the basic idea, it just becomes something “nice to have”. There are variations, and some of them (like going west of Ballard or east of the UW) would be nice, but not necessary at all. The same is true with interlining.

      10. d.p.

        Where did I ever say a Ballard/UW line would not need an O&M base? I was simply saying a track connection was not necessarily needed to send cars to SODO for heavy maintenance (overhauls etc.). You can either provide for heavy maintenance at the on-line O&M base or you can stick them on a truck as this is not something that happens all that regularly. I’ll point out heavy maintenance for LA’s Green Line is done at the Blue Line O&M base. Cars are moved by truck as there is no track connection to the Green Line from the Blue Line.

        I don’t know where you get ‘extraction pits’ from. I certainly wasn’t advocating we do something like the Waterloo & City Line in London.it is far too early to know what form an O&M facility for Ballard/UW would look like, but even for 100% underground lines these are often built above ground for good reason and need decent truck access. Since the line is so short an O&M base for Ballard/UW would not need to be the sprawling complex the current SODO base is anyway.

        Again as these decisions haven’t been made yet it is too early to tell, but we many want to take advantage of the isolated nature of the Ballard/UW line and go with different rolling stock, third rail power, and automated operation. Say something like the Rotem cars on Vancouver’s Canada Line. If this is the case a track connection is pointless as the line would be using cars incompatible with the rest of the system.

      11. The point simply is I believe the notion of building a junction in the U-District is dead on arrival with Sound Transit staff and therefore is extremely unlikely to ever happen. There are no provisions made for adding a junction to the Northgate Link tunnels. Building such a junction is a somewhat complex and expensive project as well. Especially if you can’t access the site from above. Furthermore building the connection is likely to require at least a few weekend shutdowns on the spine from the junction point northward.

        Another issue is if you build a junction you pretty much can forget ever having a second line crossing the ship canal in any of our lifetimes or rail service to Belltown and Lower Queen Anne.

        Advantages to a separate line include:
        – the ability to have different service frequencies
        – the ability to run different length trains. This allows shorter and cheaper stations ala Canada Line for Ballard/UW.
        – the ability to use third rail power
        – the ability to use high-platform equipment
        – the ability to automate Ballard/UW operations
        – the ability to use true subway style cars for faster loading/unloading and greater capacity per foot of train length.

      12. I’ll point out heavy maintenance for LA’s Green Line is done at the Blue Line O&M base.

        That is patently, massively, and hilariously false.

        I’m from there, remember? There are two massive Green Line maintenance facilities, both on the D Line. The Blue Line has no access points on the city side of the harbor, and hasn’t since around 1950. What you claim is standard procedure would have required “trucking” railcars through the Sumner Tunnel or over the Tobin Bridge for decades, and I’m not even sure those two structures would have had the clearance.

        This is the what happens when you roam internet railfan boards and accept some stranger’s brainfarting as gospel. Not a shred of truth. Total b.s.

        The rest of your above logic meltdown is bringing me horrible acid flashbacks of trying to reason with Ben about this stuff. “You can’t build the Spur first, d.p.! Where will you put the maintenance base?”, he slobbered, insisting that any new expansion requires its own (regardless of track connections), and that the only place for one is in Interbay, requiring a Ship Canal tunnel just to tighten some screws.

        I’m sad to say that you’re more likely to see that reaction from ST than any other, and they’ll be correct that the mere 3-mile span of Ballard, Fremont, Wallingford, and the U-District presents no obvious gap into which to drop a fresh base. Which is precisely why it is such a high-value corridor. And which is why serious maintenance should be elsewhere.

        Whether revenue or non-revenue, it would be insane to build a short line with no physical connection to the main line. Toronto’s Sheppard “orphan” sure as hell has one. So does any other analogous example in the universe. Link is standard gauge and the platforms are low, so this wouldn’t even prevent you from using somewhat different stock.

        Anyone at ST who rejects this idea because of such a maintenance connection has already biased themselves for other reasons, and has about as much credibility as Lazarus’s mythical “experts”.

      13. Also, the notion that building something correctly now will somehow prevent us from doing something else later, if and when the need arises, is utterly ahistorical and is Ben-level “strategy” that is frankly beneath you.

      14. And lastly…

        Furthermore building the connection is likely to require at least a few weekend shutdowns on the spine from the junction point northward.



        Heaven forbid the half dozen weekend riders from Lynnwood should suffer extremely mild and temporary inconvenience so that we might have a better transit system for the rest of time.

        Have I mentioned that a non-revenue connection makes a new line far cheaper than from-scratch maintenance facilities on stub tunnels to who-knows-where? Seriously, I cannot understand why you’re flogging worst practices so hard here!

      15. Heh… Oh, look. You were referring to L.A.’s Green and Blue Lines.

        I was all prepared to mea culpa. But what’s that? It’s a non-revenue track connection!

      16. Again, sorry for the misreading and the Boston tangent. But my point stands — you saw an image of L.A. Metro taking delivery of new Green Line cars on a truck from the factory, and you (or someone else on the internet) jumped to the conclusion that no track connection is required.

        And yet there it is.

      17. d.p.

        I’m talking about Los Angeles not Boston. LA’s green line base doesn’t have the facilities to perform heavy maintenance like overhauls.

        In any case there is room for a maintenance facility somewhere in the industrial part of Ballard. For Ballard/UW it wouldn’t need to be large to store all of the cars as a 3 mile line doesn’t need many for even short headways.

        You could even provide heavy maintenance facilities as the modern streetcar barns manage to do so in a small footprint. It might even be preferable if the rolling stock is different.

        Unfortunately if Ballard/UW was a true branch off the ST2 spine I do think it makes the political case for a second ship canal crossing much harder..

        FWIW while I think Sound Transit is willing to consider Ballard/UW I don’t think they are interested in building any kind of junction with Northgate Link.

        Just to be perfectly clear I support building Ballard/UW in the ST3 time frame. Either directly as part of ST3 or as a city supplement to ST3.

        In any case we are all wasting way too many pixels debating a minor side issue.

        Other than the legislature not granting ST3 at all the biggest threat as I see it is Dow’s vision for rail to West Seattle claiming all of the north King funds. Given the high cost of building rail to West Seattle this is likely to be value engineered to the point of being even more useless than rail to West Seattle would be otherwise. At best Ballard might get a streetcar that has been value engineered to the point of boondoggle under this scenario.

        Ballard/UW is a much better idea.

      18. My point is that suggesting worst-practice scenarios — we should build a three-mile line with its own completely redundant all-contingency maintenance facility and no connection to the outside world, just like L.A. doesn’t — comes across as kookier than simply demanding that ST learn to future-proof and do the right thing.

        Suggesting that the “political case” for any future needs rest on us pursuing intentionally weak design in the near-term is equally kooky. That the kind of thinking that screwed over First Hill forever, and it reminds me of the people who don’t care what station-access mistakes are made in ST2 because “we’ll get it right in ST7”.

        System design should be built on the basis of real need, not by consulting the political equivalent of your Magic 8 Ball. Real need will either be demonstrated, or it won’t.

      19. Other than that, yes, we mostly agree. Which is why it is really important neither to give in to kooky insistences, nor to accept silly Sound Transit-isms as immutable.

    5. These ideas for an East-West route have the same problems as the current Bus Route 44. An east-west LRT route needs to continue east to UVillage shopping center and Chidren’s Hospital, not stop at UW. If you’re not going to do it right, don’t do it at all.

      And there would be no reason for it to interline with any other line. Just build the platforms right on top of each other like they do in Toronto. At St. George, Yonge-Bloor, and Yonge-Sheppard it takes about 30 seconds to transfer from the platform from one line to the other because they platforms are right on top of each other. There’s no walking down long hallways nonsense.

      1. U-Village and Children’s Hospital are decent destinations, but not essential. The walk catchment is pretty limited (by lots of green space). The improvements to the bus system are good, but not nearly as big as those to the west. Frequent buses will soon go along 65th, which is a pretty fast way to hook up to Link for a fair number of people. I don’t see a huge number of people who will have a huge improvement in bus service because of stations in the U-Village or Children’s. This is in part because North Link lies east of I-5. On the other hand, the stops on the west side are huge from a bus feeder perspective, because heading west (past I-5 is terrible) while heading north/south is great (see “Complements the Bus Network” on https://seattletransitblog.wpcomstaging.com/2014/06/30/ballard-uw-should-be-the-next-light-rail-line-in-seattle/).

        By all means we should allow for future expansion, both east and west (most people consider the western end of this at 15th NW, but extending to 24th makes a lot of sense). But neither is essential for this first round.

        As to your other point, I think that is a very good point. It isn’t essential that we have a spur (although I think it would be nice). What is essential is that the transfer is easy. I could easily see the trains being timed. The train coming from the north will be grade separated, so it will be consistent. The train coming from the south won’t, but the train heading to Ballard can simply wait for it (since more people will make that sort of connection). Thus you will minimize the transfer time with a Ballard to downtown ride.

        I would prefer a spur, but I would take a 30 second transfer any day.

      2. If you’re going to UVil and Childrens then you’ll need some space for a turnaround and a yard. Magnuson and the north end of UW athletic fields are the largest open areas, and I don’t think they are large enough.

      3. Fortunately, Link trains can reverse on stub tracks, like they already do at Pine Street and at the airport, so there won’t need to be a turnaround. And yes, we’ll need a yard somewhere, but we already knew that.

      4. Jack,
        Actually the yard for Ballard/UW wouldn’t need to be huge. The line is fairly short, 3 miles or so. Even at very high frequencies you wouldn’t need very many trains. Say maybe 6 2 car trains plus another 2 cars for spares.

      5. Ross,

        Why would you bother with timed transfers on Ballard/UW if trains on both lines are reasonably frequent? There simply is no need if headways are short.

        If both lines have a 4 minute headway then the worst case is a 4 minute wait.

      6. @Chris — Right, which is four minutes longer than necessary. But if, for some reason, this city can’t figure out how to interline their routes then we should make the transfer as painless as possible. Make sure it is a thirty second walk, and try and time it for a minute, maybe two (to get the slow pokes). I just don’t want to see folks having to run to the other platform, and try and keep the doors open. Get Toronto style (two minute) frequency and maybe it doesn’t matter.

        But anyway, the four minutes is only at peak. What is it off peak, eight minutes? All these things add up. We already have about 500 cuts in our system, no reason to add 500 more.

  2. What you heard from Marie is exactly what my understanding of the situation is: they could control down to 90 sec headways, but practically they won’t go below 3 mins. This is mainly driven by the situation on MLK (SDOT) and variability in service.

    The only thing I would add is that it is my understanding that ST won’t hold loaded LRV’s in the tunnel between Cap Hill and Husky Station. Combined with the lack of storage tracks at Husky Station this means that Joint Ops at 6 min headways better be darn good.

    thanks for a good post. It’s about time that the myth of the vent shaft limiting capacity goes away. I have no idea how that got started.

    1. With all due respect, lazarus, erring on the side of safety is not a sin.

      Now, we have an answer to what has been a perfectly reasonable concern.

      The missing shaft is no wives’ tale. But ST seems to have a good work-around, that might even have positive spin-off effects.

  3. Is that midpoint signal the reason why CHS-UW travel times slipped from 6 to 8 minutes?

      1. The point is that this overly long express stretch is now divided into two defined “signal blocks”. That is new information and is really good to know. I do not think the intention is to halt trains at that midpoint signal, except under the most extraordinary circumstances.

      2. d.p.

        Eastbound MAX stops midpoint in the west portion of the tunnel if the station isn’t cleared. Link will be stopping there, too.

      3. I have trouble imagining how MAX would ever have an uncleared station, given the appalling crappiness of their frequencies.

        I have a new theory about Portland: they think their trains make them look cooler and “more urbane” when stopped than when moving.

    1. Originally it was 3 minutes Westlake to Capitol Hill, and 5 minutes Capitol Hill to U. For some reason they added a minute south of CHS and subtracted a minute north of it.

      1. Originally it was 6 minutes for the whole thing. That number was on every diagram and in every sales pitch.

        With any luck, “8 minutes” represents mere reflexive schedule padding, limited only to the period in which UW functions as a terminus and only for the sake of operational accounting, rather than an arbitrary and permanent slowdown or a case of false advertising.

      2. Seriously, I’m getting a little scared here, because I actually care about the process by which your synapses fire.

        You’ve long been overly differential to the “authority” spin of the week, but “not remembering 6 minutes” is some “Oceania has always been at war with Eastasia”-level shit!

      3. There’s an inconsistency in the documents then because I was going by the North Link and Lynnwood Link tables, which said 3 minutes Westlake – Capitol Hill, 8 minutes Westlake – UW, and 12 minutes Westlake – U-District. I don’t read these project updates or 2-page flyers much. Perhaps ST has a problem propagating new estimates to the whole organization. Do you want to sue to get your “approximate 2 minutes” back? I never expected the estimates to be certain until testing started. I just knew the travel time would be similar to the I-5 buses, but more reliable and frequent.

      4. It would be more productive to vent your wrath at the travel time between stations. Rainier Valley is 2-3 minutes between stations, or 4 minutes in the Columbia City – Othello gap that has a deferred station. But Capitol Hill is underground so it should be faster. Here’s your best argument for more stations in Capitol Hill. How can you go four minutes without a station in Capitol Hill when that’s the outside maximim gap in Rainier Valley? The only answer can be the cost of the stations, since underground is 2-3 times as expensive, but if so that should have been explicit in the planning.

      5. Well, since your prior comment revealed for the first time that ST intends to take 4 minutes just to go one short stop across the UW campus — seriously? — maybe it’s time to start keeping a running, referenced list of examples of this agency having no fucking clue what it is doing!

      6. p.s. ST has been heavily advertising a 13-minute running time from Westlake all the way to Northgate. That has been an enormous piece of the “look how awesome we are” sales pitch, a spin that aims to obfuscate for the public how anti-urban and often useless the lines might be.

        This isn’t about “internal estimate propagation troubles”. 13 minutes is the big fucking pitch!

        And now, suddenly, 45th Street will take nearly as long as Northgate was supposed to take? How far away is Northgate now? Slower than the 41? Slower than the 41 with traffic?

        Cracker-riding Christ! This isn’t a rounding error!

      7. I challenge you to drive between Westlake and 45th & Brooklyn in 12 minutes even at 3 AM.

      8. So now — according to you — it’s going to take a full 12 minutes from Westlake to Brooklyn. But it will only take 3 minutes more to reach Northgate (almost twice the distance)?

        So the initial segment has slowed by a couple of minutes from its bold marketing claims, and the next stop has slowed even more, but the Northgate stretch will miraculously shrink in time between now and the opening date?

        Just how much b.s. will you swallow before you start demanding answers?

      9. And Chris, the 41 does its highway stretch — tunnel all the way to Northgate exit ramp — in as little as 10 minutes in times of low traffic.

        This suddenly-growing Link runtime (and conspicuous ST promise-breaking) is a real problem. The line hardly stops anywhere. Its urban-trip coverage is pathetic. Speed is all it has going for it!

        Also, I car2go and I’m cheap. You bet your tochas I’ve driven downtown to 45th faster than that.

      10. A 10 minute ride on the 41 between Westlake and Northgate Transit center is a purple cow. It may be possible in theory, but I’ve yet to ride a 41 that has done it.

        Driving between Westlake and 45th and Brooklyn in under 12 minutes is impossible, even at 3 AM on Sunday morning unless you are willing to both run red lights and excessively speed. Sure you can get between a downtown freeway entrance and the 45th street exit quickly in light traffic, but you lose a huge amount of time waiting for traffic lights to get to the freeway entrance or to get from the exit to your destination.

        Considering most people are traveling at peak, midday, early evening, or on a weekend afternoon when traffic between downtown and Northgate on I-5 pretty much sucks the fact that amazing speeds are possible in the wee hours of the morning really isn’t relevant.

      11. Firstly, I have been on that 10-minute 41 bus.

        It’s express, remember? It comes directly out of a tunnel, remember? Google Maps, which is not based on excessive speeding, gives you 7 minutes for that stretch of highway right at this very moment.

        Second, I can easily drive downtown to the U-District in 7 or 8 minutes, and have many times.

        Car2go, remember? Cheap, remember? I check my drive times. I drive like a Masshole, for sure, but neither I speed on the highway portion nor “run red lights”. Again, perfectly conservative Google is offering my 8 minutes from 5th/Olive to 45th/Brooklyn in the middle of the lovely drizzly afternoon.

        Thirdly, we’re talking about a subway, and we’re comparing it to A) existing buses that use the tunnel, and that enjoy unimpeded express-lane access in the “primary” direction; and B) specific on-the-record travel-time promises that were a big part of the “limited stop express line” sales pitch. So comparing it to my particular style of driving is irrelevant.

        I’m a little aghast that “can’t believe” very fast trips ever happen on fast highways in the present. That was the whole original point of highways. In both these examples, 98% of the journey can take place on the highway, and the distances really aren’t that far. Of course 8-minute point-to-point trips can exist around them!

        The whole point of the subway is to be better, faster, more convenient, go more places, and be far more reliable at all times. Which is why running times are so important to the pitch. And now it will suddenly take 12 minutes to reach 45th Street instead of then prior 8 minutes?! Northgate was 13, now 15, maybe soon to be 20!?

        This is not a matter to be casually shrugged off!

      12. I’m guessing the confusion is simply because of the difference between the UW (Husky Stadium) and the U-District (Brooklyn) stations. From downtown I think 6 minutes to Husky Stadium, 8 minutes to Brooklyn and 13 minutes to Northgate all make sense. These are all consistent with our spacing and our long dwell times. Everything else is probably just rounding differences. There are twists and turns in our line, but most of those are close to the stations (making their effect on speed negligible) There are long straightaways on the big gaps (Capitol Hill to Husky Stadium, Roosevelt to Northgate) where speed matters and some twists and turns for the shorter sections (coming out of Westlake as well as between Husky Stadium and Brooklyn to avoid stuff in the UW). I think in general this will be pretty close to the original estimates (13 minutes from Northgate to Westlake, etc.).

        As mentioned, we could shrink the dwell times to get better speed, but of all things to worry about in our system, speed is probably last on my list.

      13. Mike was explicit about the revised travel times he has seen referenced. The next segment to open contains only two stops. 6 minutes versus 8 minutes is hard error to make.

        And he claimed the new U-District statistic is 12 minutes, which would genuinely qualify as “not that fast”.

        I hope Mike is wrong, but there is very little suck-creep that I wouldn’t put past Sound Transit.

  4. Martin said “There’s also the question of where to put the trains. At 30 trains per hour, 10 of them can go neither to Bellevue nor the Rainier Valley.”.
    I suppose sending 10 trains per hour to from Ballard, entering the tunnel from the expanded Convention Center Station, then continuing down the busway to Spokane St wouldn’t be the worst thing in the world to have happen. 6 minute headways from W.Seattle to Ballard would be 8,000 PPHPD.

    1. If you want to do anything with Convention Place, you need to get organized right now. The Convention Center expansion is planned to close off the station in the next several years.

      (And a Convention Place – SLU – Belltown – Ballard line sounds pretty good.)

    2. At 3 minute headways there are 20 trains per hour. While the math works for a 3 way split to West Seattle, Des Moines, and Redmond the math to the North doesn’t really work without decreasing headways below 3 minutes in the DSTT which may cause reliability problems and additional traction power substations. In addition there is the issue of the IDS to SODO segment which due to the road crossings is limited to 6 minute headways.

      1. I missed that memo on 6 min. headway limits going down the busway. I guess all the buses going down there with 1-2 minute cycle times on the signals will all have to change when they too get ‘kicked’ off the busway for capacity issues.
        just sayin’
        or is this another urban transit myth being born?

      2. Wait Chris, you’re saying that the 6 minute headway limit is due to the crossings in SODO? I had thought that it was due to the MLK at-grade segment. It was my understanding that the signal cascades needed to keep the trains moving in both directions without hopelssly snarling traffic on MLK was the cause of the limitation.

      3. Two artic buses following each other down the busway, 6 seconds apart are identical in length to a 4-car Link trainset. Why on earth would the rules change just because it’s a train instead of two buses?

  5. The additional trains should travel down airport way and rejoin Central Link south of Rainier Beach. A Duwamish Bypass Will be necessary, and it will help Central Link not just with travel times but also capacity.

    1. There’s even a flying junction at the Maintenance Facility provided to merge/divide the routes.

      I have a hard time believing that the ST Board “permanently” removed the Duwamish Bypass from the regional plan a couple of months ago. IF climate refugees fill up South King County, the Bypass will become necessary. And they very well might.

      1. “Climate refugees.”

        Apocalypse miraculously fails to disturb local government stability, the revenue stream, even the subarea equity! Indeed, global meltdown has no effect locally except to get us so very many trains!

        This truly is the dumbest pet prediction in the history of STB. And Ben used to write here.

        Anyway… TG, there is no capacity problem to the south. So all a bypass would do is to halve service where it is actually needed, at great cost, and in pursuit of geometrically-impossible ridership from the distant sprawl. It is dead.

      2. “I have a hard time believing that the ST Board “permanently” removed the Duwamish Bypass from the regional plan a couple of months ago”

        May I ask where you heard of this? If true, that is a tragic blow to the potential of Central Link.

      3. Let’s put it this way I’d prioritize a gold plated light rail subway to West Seattle ove a Duwamosh bypass. A Graham street infill station will do much more to fulfill the potential of Central Link than the bypass will.

      4. Chris Stephan

        A bypass will save everyone south of TIBS 12 mins in travel time. (Including everyone going to the Airport, Federal way and even Tacoma) It literally doesn’t take 12 mins to WALK (not bus, walk. I think bus #8 on MLK is completely redundant for this reason) from Othello Station to Graham St. Also a bypass will create the potential for a Georgetown station.

      5. 1. It takes 17 minutes to walk to Graham from any other stop. Funny how the people getting this wrong are invariably the same ones flogging 35-mile subways to nowhere. You people haven’t a clue how usable rapid transit works.

        2. Georgetown is tiny, is surrounded by industry and emptiness, and will remain that way. And somehow in spite of the tininess, this former company town manages to sprawl in a way that makes buses and bikes and cars permanently better suited than the rail it won’t need in a million years.

        3. Federal Way ridership predictions are crap, because the land use is crap. It doesn’t matter if you cut minutes off if the journey. Which still wouldn’t be competitive with express buses, btw.

        4. Tacoma is 20 minutes further regardless. No one will ever ride this inter-city, at least not without whopping regret. Sounder. Express buses. Period.

        5. Where are you getting “12 minutes faster” anyway? You’d only bypass a couple of miles of track and a handful of stops. You still need to move those remaining miles. Subways aren’t teleporters.

        6. For the airport, like anywhere else, frequency is as vital as speed. And your plan would inevitably reduce that frequency, leaving riders with a trip that takes functionally just as long.

        7. You’d also wind up cutting frequency in the Rainier Valley, where (as the suburban geometry-deniers have failed to notice) the majority of ridership and ridership growth is. That urban ridership will only get better when urban access improves.

        Sorry, South County Reality Deniers, but you are wrong on every point.

      6. dp

        -Someone working for ST told me the bypass would cut 12 mins at an ST3 meeting.
        -no, it takes 11 mins to walk that distance according to google maps. It would probably take me less than 10 mins.
        -What’s the point of a train that is so slow it can’t even replace a bus?

        The whole point of building a train really is to provide speedy, reliable trips. Link, in its current from, fails to do so. (Thanks to MLK), seriously, try travelling outside of North America. In the rest of the world, trains are faster than the buses they replace.
        (I’ve lived and studied in Japan for 10+ years, and even interned at a Train manufacturer there, and trust me; in the rest of the world,a train that is slower than the buses it replace would be considered pretty pathetic)

  6. Thanks for doing this research, Martin!

    So, now we know what the signaling was designed for. That still doesn’t tell us the shortest headway the signaling system can tolerate through the tunnel, or on MLK. It can handle 6 minutes on MLK. She didn’t say it couldn’t handle 5 minutes. Or 4 minutes.

    Also, the claim has arisen a few times about a limit on the number of trains on the bridge between Seattle and Mercer Island. Is there anything to that?

    An ancillary question would be what the maximum dwell time at each platform in the DSTT would be, before trains behind have to be held. (I’m thinking more about after the buses are kicked out, since this is a post about long-term minimum headway.) Expected dwell time would also be of import. If dwell time at any one of these stations can become the headway limiter, it would be nice to know now, while there is still time to renovate stations before East Link opens.

    1. In general if a rail transit line has headways shorter than 5 minutes you start having problems with car/train interactions at grade crossings. Either not enough time is allowed for cars to cross the line or trains tend to bunch due to waiting for cars.

      Due to grade crossings in Bellevue, SODO, and on MLK they limit the headways on these segments.

  7. The answer about the allowed throughput in the U-Link tunnel has me wondering about whether running 6-minute peak headway with joint ops in the DSTT is really a safe idea. ST is saving wear and tear on LRVs by not running 3-car trains all day, and saving money from not pulling lots of trains in and out of service. Maybe it is safe, but I’m trying to wrap my mind around how. Joint ops has been, and will continue to be, unreliable. The trains certainly do not run at perfect intervals as it is.

  8. And now, a point of irony:

    Any time a line from UW to the eastside was suggested back in the day (roughly 2011 or before), the response on this blog was that it would put too much extra passenger traffic on Link, at peak, in the peak direction.

    Now, there is a big push for the Ballard Spur, which would have the same problematic effect, or even more. We have also downplayed ridership from the eastside to UW (depending on whether it is the alignment we like) and played up ridership between Ballard and UW, which only exacerbates the problem if our ridership guesses and the original fears have any merit.

    Ben, care to chime in?

    1. Well the recent corridor studies by ST show a line over 520 would be a loser. The only route with any ridership to speak of is to Redmond. Even then the cost is higher and the ridership lower than East Link.

      Using 520 for East Link would have also meant either bypassing Bellevue or doing a time wasting U turn in order to serve it.

      1. 520 would give 19-21,000 daily riders according to ST’s most recent study. run a line from sandpoint to kirkland and include another 4-5 stations and ridership jumps to 25,000+ which would be a good number.

      2. 20,000 riders per day is less than East link at greater cost. Going via Kirkland and Sand Point would be even more expensive. The addition to ridership is questionable as central Kirkland really isn’t all that big.

        At least at this time another cross lake rail line doesn’t make much sense. It certainly didn’t make sense as an alternative to East Link.

        We may see the 520 crossing in ST3 simply because there aren’t many better projects other than completing East Link to downtown Redmond.

      3. This study was done in july where the numbers are in addition to east link which serves downtown and not UW/Childrens/Pill Hill, etc. It will not cannibalize east because east link is too long a route to get to the U. Redmond/Kirkland/N.Bellevue will opt for bus before Link to get to UW. 2-3 stops in Kirkland and a couple 405 park-n-rides for folk north would provide more than enough to bring an additional 5,000 riders. If ST is willing to spend billions for WS or Ballard with comparable ridership potential then this is a good bet for a line. 520 will never fly because to difficult to configure for Kirkland and points north. A Kirkland line would be much easier to configure and connect to Redmonad and N. Bellevue.

      4. LES,

        Bill was asking why East Link didn’t use 520 instead of I-90 particularly as the capacity arguments Ben was making against the idea have proved to be invalid. Way back when many East Link opponents were suggesting Sound Transit should have used 520 instead. Mostly it was yet another “keep light rail the hell out of our neighborhood” argument from Southwest Bellevue NIMBYs.

        To your point I think a Sand Point crossing is an absolute fantasy. If Sound Transit decides to advance a second lake crossing, it almost certainly be via 520 and go to Redmond.

        The costs of a Sand Point crossing are unknown, expensive tunneling through downtown Kirkland would be required, and on the Seattle side there would be fierce opposition to anything but a tunnel between University Village and Sand Point (also expensive).

        The crossing itself would be problematic as it would need to be a floating tunnel to avoid fierce opposition. Floating tunnels while in theory possible are a largely untried technology.

        For all the cost and NIMBY firestorm you’d ignite you might get a handful more riders than a 520 crossing.

        Note that at this time I don’t believe there are any rail projects on the Eastside worth spending money on beyond completing East Link to downtown Redmond.

      5. Just like the myth of the Montlake shaft, I would like ST to address the “myth” of a feasible/unfeasible bridge to Sandpoint. Every blogger is an expert, at least they like to think they are; a simple study would go along way in addressing issue. And you’re probably right, a line at this point is unrealistic and probably wouldn’t even make ST4. But a 520 line will never happen; sure it’s already got a bridge but the remainder of project would not be cheap, unobtrusive and easy to configure. Spending billions doesn’t bother me, what does bother me is not utilizing the billions that are spent to maximize the removal of the most cars/buses from the roads and serve the areas with the most potential riders. I don’t see ST3 heading in that direction.

      6. @LES and Chris — I think when they finally crunch the numbers, what makes sense is to use what we are building now. Buses traveling very quickly in their own lane (or very close to it). The only problem we have now is the “last mile”. Fixing that, even with the most expensive, fancy solution (a bus tunnel underneath the ship canal and then connecting to Husky Stadium) would be much cheaper than another light rail line and just about as useful. Kirkland really isn’t that big. It just doesn’t look like most suburbia (more of a grid) so folks assume it is bigger than it is. The biggest cluster of density there has fewer people than Lake City, and last time I checked, no one is planning a tunnel connecting it to the UW, even though it would be cheaper.

      7. Ross,

        I don’t disagree, however East King has a huge wad of money to spend and not that much to spend it on. The numbers for UW to Redmond via 520 were about the best of a sorry lot of potential rail projects.

        For this reason I think there is a chance it might show up in ST3.

      8. Well, I don’t know what makes the most sense for East King, but I would be willing to bet it isn’t another crossing. First you have to push Link to Redmond. Then, I don’t know, a lot of bus service. Maybe some BRT. Sound Transit wants it on 405, but I like the idea of putting it on the old railway line. Either way, I think ST bus service is quite popular, especially on the east side. Let’s not forget the only reason we even got Sound Transit (and rail) in the first place is because, just to paraphrase, there was just enough bus service for the suburbs and just enough rail for Seattle (http://community.seattletimes.nwsource.com/archive/?date=19961106&slug=2358535).

      9. Ross,
        Again I don’t disagree. I think BRT between Bellevue and Totem Lake via the ERC is a huge win. But at this time I can’t say what the final list of projects for East King will look like other than completing East Link between Overlake and Downtown Redmond will be in there.

      10. To rehash the arguments about why turning the old rail line into a bus/train corridor is a bad idea:

        1) The path is not straight. It’s full of twists and turns. Any vehicle using would likely be limited to around 30 mph. It would be substantially slower than I-405, especially once the new HOT lanes open.

        2) The existing corridor is of great use as a walk/bike trail and should remain that way. While it is probably technically possible to squeeze in a bus/rail corridor and a trail side by side, it would be impossible to do so without a huge loss for trail users and nearby homeowners. Not only would have to bulldoze a bunch of trees surrounding the trail, but safety concerns would inevitably limit east/west crossing points, both for trail access and general east/west connectivity through the neighborhood. (As it is, there are several informal east/west pedestrian crossing points that don’t show up in the street map. With a parallel transit corridor, all of these would likely have to be fenced off). Furthermore, the noise of buses and trains roaring through would be a significant detriment to the surrounding environment, which currently gets very little car noise from nearby streets.

        3) The level of transit demand in the area for the forseeable future does not justify any level of service that can’t reasonably be met by buses down regular streets or I-405.

        Instead, of giving up the trail for transit, the trail needs to be expanded. It should connect south to Bellevue and north to Woodinville, with a connection to the Sammamish River Trail.

    2. I’m not following your logic on this one. The article states that 4 min headways can handle demand and that 3 minute headways are possible without additional captial investments or potential service issues.

      3 minute headways would easily add enough capacity to handle transfers from a line ST estimates at 24k riders/day (Ballard Spur) – even if they blow that number out of the water (they would.)

      1. As d.p. has said elsewhere we should be so lucky as to max out Link between Westlake and Brooklyn station.

        While packed rail transit doesn’t always mean getting the money for capacity relief is easy or a short process (see the Second Avenue subway), it certainly makes the justification easier.

        By the way the key to blowing ridership estimates for Ballard/UW out of the water is providing decent urban stop spacing (1/2 mile) at reasonable locations for transfers. At the very least there should be a station in the commercial heart of Wallingford near Meridian and a station between Fremont Ave and Aurora.

      2. The 2030 numbers are for a built-out ST2 (since no other segments are expected to be open by then), not for the extra passengers transferring from the Ballard Spur.

        Really, my question is to Ben, who seems to have changed his tune a full 180 degrees, from denouncing any east-west line through UW because of its impact on peak U-Link capacity, to becoming one of the biggest backers of a line that will do what he fought so hard against all these years.

        I’m fine with the Ballard Spur, FWIW, so long as another north-south line is in the works, for when (not if) North Link reaches its rush hour capacity limit. I just want to know why Ben changed his mind.

      3. Ben who? Seriously, you claim that (“the response on this blog”) didn’t like East Link because it would put stress the system. I never heard anyone say or write that. Seriously, I don’t know who said that. Perhaps Ben was wrong. Bring it up with him (if he even said what you claim he said). Who cares? Do you think I try and point out when Bailo is wrong about something five years after the fact (by the way, nice to see you commenting, John).

        Keith Kyle is the outreach director for Seattle Subway and he has answered your question quite well. This blog has endorsed the Seattle Spur for the same reason. The worries about putting stress on the system have pretty much been obliterated by this article, which came from Sound Transit. That is three organizations who officially have said there is nothing to worry about.

        I really don’t get this attitude. It is crazy to me. I’ve never seen a subway line built like this (or a freeway line or practically anything). When they built I-90 they knew it was eventually going to put pressure on I-5, but they didn’t build a second east/west freeway at the same time. When a company opens, they hope they will get swamped — they hope that they will have a lineup out the door — they don’t purposely discourage people from coming. No rock band in the world says “Well, let’s make sure this album isn’t too good, otherwise it will be hard for our fans to get hold of it or see us live”.

        Build the most important, most popular stuff first. Then, if it is crowded, supplement it with other service. There will be plenty of support for that. But I think it is nuts to build a crappy system (or an overly expensive system) because we are worried that four car trains going every three minutes (with no extra cost) or two minutes (with a little work) are not enough. That is basically what Toronto has. Toronto! A city that has over 2 and a half million people, and has the second most popular transit system in North America.

      4. Ok – just to be clear though. 4 minute trains from Northgate will meet demand in 2030 and they have the option to run 3 minute trains to handle additional demand from the Ballard Spur.

        There is no capacity issue in this regard.

      5. “When a company opens, they hope they will get swamped”

        They hope they will be at capacity, not overcrowded. But a business is not a subway line. A business just wants to put butts in seats to make money, and doesn’t care what people do who can’t get in. If Safeway is full you can go to another Safeway or QFC or Trader Joe’s.

        But the purpose of a subway is to meet the mobility needs of the area, so it does matter if people are being left behind. If there’s only one subway, their alternative is a slower and less reliable bus. It should be designed so that it never reaches capacity for thirty or fifty years, except unusual spikes like 4th of July and Seahawks victory parades. That gives a sufficient period of time where it’s adequate, and sufficient lead time to build a parallel line before it gets overcrowded.

      6. Oops – my last comment was for Brent. Lots of action this Saturday morning. :)

        Thanks for the support Ross – I’m currently the President of Seattle Subway. ;).

      7. Awesome Keith — congratulations and sorry for not properly recognizing your promotion (I did a sloppy search).

        @Mike — Sorry, but come on. Look, if the choice is:

        1) Build a really good subway that can handle tons of people.
        2) Build a really good subway that is crush loaded.

        Then obviously I vote for one. But that isn’t the choice. That isn’t analogous. I’m saying if the choice is:

        1) Build a really good subway that might be crush loaded if Lynnwood ever gets bigger than Seattle.
        2) Build a mediocre subway that never has that problem, because it isn’t that good.

        Then I vote for one. Seriously, it isn’t just that Ballard to the UW is cheaper, it is that it is better. This will be a common conversation thirty years from now, if we fail to build the right thing:

        Sally: “Hey Joe, you live in Lynnwood, how come you don’t take Link to work? I take it downtown and it is sweet. I can really stretch out, too — it is like a ferry (until I get to the UW, then I have to move my stuff)”

        Joe: “Well, I work in Ballard, so by the time I go all the way downtown and all the way back, I’m just better off driving”

        Fred: “Yeah, I hear you, I live in Ballard and work in Redmond. I tried Link, but I have to take a bus that doesn’t run that often on the other end, so I just drive 520”

        Sally: “Can’t you just take a bus on 520 from the U-District?”

        Fred: “I tried that, too. But I have to take the 44, which is way slower than driving, or I have to take Link downtown, then back up. It’s not too bad in the morning, because the connection is pretty easy, but at night it is hard since the downtown to Ballard train doesn’t go that often. Driving isn’t that bad — I just wait until the traffic dies down”

        That’s the thing. Just count the stops. Going from Ballard to downtown via the UW is a minor detour. There are only six stops between Ballard and Westlake via the spur. Going the other way is not. Roosevelt to Ballard has four stops in between (via the spur) but going downtown first means nine stops in between, with similar stop spacing (four to Westlake, followed by Belltown, Uptown, Queen Anne, Fremont, East Ballard). As d. p. said, the station is not a teleporter. You still have to get there. That is all assuming we build the best, most expensive line (Corridor D).

        As Keith said, we can have three minute frequency with four car trains. That is huge! With a little bit of work, we can up that to 2 minutes. Holy smoke that is a lot of people. Cities much bigger than us have less frequency.

        I really don’t get it Mike. As you said, the purpose of a subway is to meet the mobility needs of the area, so it does matter if people are being left behind. But that is happening *right now*. Our current subway is so weak that it leaves just about everyone behind. Two car trains traveling way less than allowed headways means that it simply isn’t good. I’m being left behind — most of the city is being left behind. It will get better (a lot better) once it gets to Northgate, but much of the city will still be left behind. The best way to allow them to ride this super expensive line is to stop worrying about too much success, but start trying to achieve it.

      8. Just to be clear I want both Ballard/UW and Ballard/Downtown. I believe the population and ridership is there to support both. I would build both before I’d build rail to West Seattle.

      9. @Chris — That is why I like the WSTT and Ballard/UW. It is great for West Seattle and great for Ballard and great for folks along Aurora (including South Lake Union, where pedestrians will soon be better connected to the grid). They compliment each other quite well. In the long run it could lead to another train to Ballard. I’m with you, though, in that I don’t think West Seattle rail will ever make sense. But I want to build something really good so that folks in West Seattle see how good bus service can be (like a lot of people, they assume that trains are always better than buses — they aren’t).

    3. @BW,

      You are correct, the concern with any sort of feeder or connecting line north of the cut has been the capacity issues that arise from adding more directional riders to a commute that is already highly directional in nature. This is why the I-90 crossing is preferable to a SR520 crossing, and why a direct Ballard-DT line is preferable to the Ballard Spur (amongst other reasons).

      Additionally, all this buz you hear about the Ballard Spur is coming from the fan base and not the technical experts. They have yet to weigh in publicly yet, but they will surely follow the data and not be swayed by emotion.

      1. Seriously, Lazarus, shut up! You’ve been proven wrong on the urban-mobility precedents, on the cost:benefit analysis, on the time savings per passenger, and now on the de facto total lack of capacity constraints proportional to the kind of transit demand you’re ever going to see in a small and decentralized city like this one. North Link would have to become one of the busiest subway lines in the Western Hemisphere to run out of space. So knock it off!

        All you have left to stand is your obsession with the radial structure, which in your mind seems to reign supreme no matter how badly any individual line’s access-shed. Which is a success-hobbling prospect from Mountlake Terrace to Interbay.

        Anyway, it doesn’t sound like the experts Martin consulted are nearly as enslaved to tunnel vision as the “experts” you keep pulling from your the base of your warped skull.

      2. “Info. “Vetted.” “Experts.”

        These words do not mean what you seem to think they mean.

      3. Lazarus, you cannot claim to argue from authority without evidence or citations, especially when you are continuously proven wrong!

      4. North Link would have to become one of the busiest subway lines in the Western Hemisphere to run out of space.

        At one point in time, the busiest subway in the world was in the western hemisphere, at least if measured in terms of riders per mile of track. They added a bit of track since then so I’m not sure how they rank these days.

        However, the Puget Sound region would have to absorb the entire population of the state of California before it got to the population density of São Paulo.

      5. Lazarus,

        Again [citation needed]. Who are these supposed ‘experts’? Sound Transit staff? Metro staff? City of Seattle staff? Washington Policy Center Staff? John Niles? Kemper Freeman? The Bavarian Illuminati? Skull & Bones? Who?

  9. I think first week’s operations on U-LINK will give us a world’s worth of new questions. Followed by preliminary answers after several years.

    For instance:

    1. Will the grade crossings on MLK have to be undercut or bridged over to keep service reliable on the entire remainder of the system?

    2. Will entire escalator and elevator system in DSTT and elsewhere have to be rebuilt and upgraded to handle increasing passenger loads?

    3. How will passenger information be upgraded to meet seriously increased demands?

    4. Will there finally have to be passenger restrooms either in or near stations?

    I also have a feeling that the mezzanines that have sat empty all these years will finally have some use.

    There’s nothing wrong with constantly looking at real-time data and projecting. But more important is constant attention to developments minute by minute, and solving, and learning from, events as they happen.

    Mark Dublin

    1. My answers:

      1) No. These are pretty reliable, and if not, they should simply allow for a little extra dwell time before going through the shared sections. As it is, they should probably shrink the dwell time considerably in most areas. But if a train arrives early at, say, Mount Baker (or anything before I. D.) have it wait a little bit before synching up with the other trains.

      2) No, but it would be really nice if they were.

      3) No, we will muddle along as we always have. We are making progress, but I don’t see Link being more complicated than the buses (quite the opposite).

      4) That would sure be nice.

      I get your point, though. As we create a system that is a lot more dependent on transfers, the value of the little things goes up considerably. If you have a wait a couple minutes for the next train or bus, it would be nice to be able to duck into a bathroom. Likewise, minimizing the time spent switching from seat to seat is very important. I sure hope they do pay more attention to these things.

    2. Great questions Mark!

      1) If it does, ST could look at longer trains at the current frequencies.
      2) DSTT stations need a comprehensive assessment before ST3 so that improvements can be put into the measure. It would be smart to do strategically, as riders in existing corridors would perceive a greater benefit to ST3.
      3) The lack of prominent recognition of platform directions is terribly needed. That won’t depend on U-Link and could be done tomorrow with the message signs in the stations.
      4) Systems around the US have closed restrooms for security reasons (reducing places where explosives can be placed).

  10. Thanks Martin. We should bookmark this page when someone says a Ballard to UW spur would overload the tunnel.

  11. This is an outstanding article. I really appreciate you getting answers to questions that many of us (including me) have wondered about for years. Thank you very much. This clears up everything.

    Well, just about everything. Is East Link limited to six minute headways as well? I’ve heard conflicting stories (more rumors) going either way. This would be the logical place to send trains if we wanted to bump up the frequency.

    1. Don’t forget that East Link will have an at-grade segment. It won’t be as long as MLK and it will have fewer station stops, so it may not affect miniimum headways as much.

      1. Oh, that is probably it. The headways for Central (AKA South) Link are based on the signal prioritization on Rainier Valley. The city doesn’t want the train screwing up cross traffic by giving it priority too much. I would assume the same is true for East Link. The rumors about the limits on the bridge are probably unfounded. The limits are probably based on the roads.

  12. The wildcard is Everett and Tacoma. They will certainly increase ridership, perhaps using up the 4-minute to 3-minute difference. Perhaps less, perhaps more, but the point is we don’t know, and we especially don’t know when UW, Lynnwood, and Overlake haven’t opened yet. So it’s worth being cautious. A subway that’s overcrowded soon after it opens is not an achievement, it’s a failure. Also, a second DSTT puts us in a very good position in any case. Instead of worrying about possibly reaching capacity, we’ll have plenty of capacity for all the expansion lines that have been suggested, or several bus routes which would dramatically improve citywide transit. So we should just put the second DSTT first, not try to avoid it.

    1. No, a subway line that’s crowded as soon as it opens is a great success in comparison to no subway line at all. It might be a failure in comparison to two subway lines with equal overall ridership and costs, but that’s not what we’re comparing it to. The Ballard-UW line is being argued for because it will get more ridership for less money. And what’s more, given the size of the Sound Transit package, it’s very possible the choice will be between Ballard-UW and nothing in that corridor. Given that, it’s a runaway success.

      1. I agree. Stop worrying about being too successful, and start trying to be successful. We have a very, very long way to go before we reach the limit of this system.

    2. Tacoma is limited by the six minute headways. So three or four minute limits are pretty much meaningless. If they are crush loaded (and I doubt they ever will be) then it is because the train route is not grade separated.

      Everett is a different story, but Everett is much smaller. It is about half the size of Tacoma. Likewise, I’m pretty sure the suburbs in between are much smaller (Kent and Federal Way have more people than Lynnwood and Shoreline, etc.). Everett is also a city, which means a substantial number of people there have no interest in commuting to Seattle. Link will be reasonably fast, but it will be no silver bullet. You get diminishing returns with light rail as you go further and further out, because each stop takes time. I think you’ve already reached that point in Federal Way (let alone Tacoma) which is why folks there want to keep their express buses. To be fair to Everett, it won’t suffer as much by that phenomenon. Even if there are express buses, there are already good stations (at Lynnwood and Mountlake Terrace) that hook in quite well to express buses. Getting from Everett to Seattle will still take a long time, though. The light rail isn’t magic. It only seems that way when operating in the city (“Holy smoke — Ballard to UW in six minutes — it must be going 90 miles an hour!”). No, it is still plodding along, but way faster than the alternatives. The distance is just surprisingly small, that’s all (less than three miles). But Everett to Seattle is still 30 miles. That puts it very close to “maximum comfortable commute” even if you are driving directly at 60 MPH right to your destination. Link won’t do that — most riders will have to get to the station, then take a train that will take a while before the train actually gets to Seattle. I’m guessing most trips will be close to an hour, if not more. I don’t think we need to worry about huge numbers of people flooding our system because of that.

      WSTT is a solid suggestion on its own right. It is by far the best solution for West Seattle, and compliments the Ballard to UW line. Someone from 15th and Market will still probably go via the UW if headed downtown, but if headed to lower Queen Anne or Madison Street (which includes a great connection to Madison BRT) then that is the way to go. Plus if you are in upper Ballard (say, Ballard High School) you can keep going on the bus downtown, or transfer to Link. But that’s just Ballard. It is huge for Queen Anne (both upper and lower) as buses spend most of their time slogging through downtown traffic. Likewise it is huge for Magnolia and West Queen Anne for the same reason (and because heading north to Link would take a while). Then there is the Aurora corridor, which would finally get a quick ride to downtown. This includes the edge of South Lake Union as well, which will eventually have a decent street grid. We can’t forget Belltown, which would finally be connected to the rest of the system — finally we connect the most densely populated area in the city. This is good. Very good.

      So, yes, the WSTT is a fantastic idea. It would improve transit mobility by quite a bit. But Ballard to UW is still a bit better by my reckoning. Either way, they compliment each other quite well. I don’t buy the argument that we need to worry about capacity, but yes, adding the WSTT after building the Ballard to UW line (or at the same time) means we can forget about that issue. It also means that we could build out two train lines to Ballard (one via the WSTT and one via the UW) which could easily make sense eventually. Extend the UW to Ballard line a bit west (to 24th NW) and extend the other line a bit north (to NW 65th or beyond). I think we will probably build other things first (light rail to the Central District and South Lake Union) but that is still something to work towards in the long run.

    3. OK, I estimated wrong. I forgot that the people from the Everett stations will already be on Link as soon as Lynnwood Station opens because the buses will be truncated there. The Everett extension would add some new riders beyond that, but not enough to add several more trains per hour. And in the south end, we have to assume the express buses will still be running until ST says otherwise.

      However, I’m still hesitant to put a Ballard spur in the Central Link tunnel because that would precommit 100% of the capacity and leave no room for flexibility.

      As to why Ballard is lower priority than Everett or Tacoma in the Central Link tunnel, it’s because of the promises ST has explicitly made (to Lynnwood and Des Moines) and implicitly made (to Everett and Tacoma), and the suburban boardmembers.

      1. If the WSTT and the Ballard Spur are built simultaneously, the capacity argument pretty much goes away, as Link from Ballard to downtown pretty much becomes a wash in travel time vs. the direct bus.

        Although, I still think the argument that so many people will be riding from Lynnwood that there will be no room for people from Ballard, even with 4-car trains running every 3 minutes is B.S. The limit of one daily Lynnwood rider per parking space, combined with the fact that Lynnwood TC only will only have a couple thousand parking spaces (with no room to feasibly increase this by a factor of 10 or more), practically guarantees it.

      2. You don’t think anyone will use the feeders? That the only people who will use Lynnwood Station will drive there? That there will be no more riders from Edmonds and Mukilteo and Mill Creek? That the existing Edmonds-Lynnwood buses will be empty even though CT has made 200th a frequent corridor?

      3. No, we think that none of those cumulative methods of arrival will accrue tens upon tens of thousands of passengers at rush hour (i.e. equivalent to every one of the best-performing bus routes in the entire metropolitan area combined). Which is what would be required to remotely “pack” the trains.

      4. Also, I decided to take a look at BART, just for giggles. It is always hard to pick an ideal comparison, but I picked Fremont, California. Fremont has 240,000 people, which is bigger than Tacoma, and more than twice as big as Everett. There are probably more people in Fremont than there are in the entire “North Link” area, north of Northgate. So, anyway, greater San Fransisco is much bigger than Seattle, and East Bay is huge. As luck would have it, this also connects to a couple airports and a major university (Cal) just like us (UW). Great, so I suppose a “Fremont, CA” would just swamp our little light rail line, right?

        Wrong. There are 8,300 people a day on the train, or basically a bit over one trip. Well, I’m sure it picks up along the way. After all, this goes through Union City (71,000) and Hayword (150,000) and San Leandro (85,000) which is way bigger than our northern suburbs:

        Fremont — 8,300
        Union City — 4,600
        South Hayword — 3,100
        Hayword — 5,100
        Bay Fair — 5,700
        San Leandro — 5,789

        It is really quite striking how poorly this system, built for speed — designed to quickly move people from the far off suburbs into the city — fails. By “fails” I mean fails to carry more people than a typical bus (say, Metro 41). These numbers are astonishingly low, and I see no reason why our suburbs, which are smaller, and will be connected to attractions that are smaller, and will take longer to travel, should come even close to those numbers.

        BART only gets decent numbers when it gets close to the city. Every station over 10,000 is in San Fransisco, Oakland or Berkeley. Some of those are several times bigger than Fremont (there are a couple in San Fransisco that are over 40,000). That’s right, one station has more ridership than the entire section I mentioned (a section shared by two lines). I don’t know how else to say it — we really don’t need to worry about suburban riders swamping out system — they won’t.

      5. Wait, you mean the model (Wizard of Oz) could be way off, and Lynnwood Stn won’t generate 25,000 per day?
        … but, but, we want to be just like BART, don’t we?
        All these headway calculations are starting to make my head hurt.

      6. Even the feeder buses are still limited by the amount of parking they serve. Add up all the parking spaces at Everett Station, Eastmonth P&R, and South Everett P&R, and the 20,000 daily riders at Lynnwood Station still doesn’t pencil out.

        The local feeder routes (that people can walk to) run once or twice an hour and carry around 10 people per trip – again, a negligible contribution to a goal of 20,000 riders per day.

      7. As of 2011 there are 19,000 weekday transit trips going past Northgate on I-5. Note that this does not include routes using other paths such as the 5, Rapid Ride E or the 522.

        To give some comparison there are as of 2011 33,000 weekday transit trips crossing the Ship Canal on I-5. Again this does not represent transit using other corridors including the 71/72/73/74 trips using Eastlake.

        Sound Transit expects to add 50% new riders with the completion of Lynnwood Link. Additional 2935 riders would come from population growth.


  13. Martin:

    One thing not mentioned here for sub 3 minute headways are the fire code issues in the DSTT.

    Its my understanding that the open mezzanines would be a problem for sub 3 minute service. They would need to close them for fire code reasons. Can you verify that with ST?

    What suprised me is that they were talking 90 second service at all. That would certainly be enough to handle Ballard/WS but would preclude future expansion. I don’t think we have enough information to fully form opinions on that front.

    Lots of plusses and minuses vs biting the bullet and building the WSTT. Would love to see ST do a full breakdown of that decision point.

    1. The WSTT still makes sense. I mention this above (https://seattletransitblog.wpcomstaging.com/2015/03/21/capacity-limitations-of-link/#comment-604108). They compliment each other quite well. I think Ballard to UW light rail is more important, but the WSTT is huge. If you tried to merge the WSTT with the existing tunnel, you don’t save that much money. You would only save yourself the section from Westlake to the SoDo busway, you still need to do all the digging on the north end as well as the ramp on the West Seattle freeway on the south end.

      You also skip Madison — that is only one stop (and pretty close to other stops) but that is still an excellent stop that adds value on its own.

      Then you run buses in the tunnel, even though the tunnel isn’t really designed very well to handle both (although it does now and maybe it could be improved). I’m not sure how much the tunnel can handle from a mixed perspective, but I’m guessing not that much.

      If you run rail instead, you encounter all the problems with rail on the west side. It is very expensive, and not very convenient (compared to one-seat buses) for West Seattle. Ballard is a bit better, but then what about Aurora? Do we add rail along there (that would cost a bunch). Eventually rail through there, maybe. Eventually. But we are a long ways from there. Meanwhile, you are trying to get all these lines (rail from Ballard, Aurora, UW, West Seattle and Bellevue) to all work together in a system that Sound Transit has said could have lots of problems going that frequently. In other words, I think you could get things mucked up in a hurry.

      The great thing about the WSTT is that it gets us something great in the short run (really fast bus service) while allowing for good things in the long run (another train line to Ballard or even West Seattle light rail if it ever became big enough) at a reasonable cost. If we were committed to West Seattle light rail, Aurora light rail and Ballard to downtown light rail, then we should certainly explore the idea of sharing the line, but that is way too much money to spend right now.

      There is nothing terrible about having two lines through downtown. Lots of subway systems have that, and it usually isn’t about capacity. The second line just covers a different part of downtown. This line will do that. Maybe it overlaps a bit too much, but not too bad, really. If we really want to maximize the value of the WSTT, then move the Madison stop up to First Hill. But that would cost money; a lot less than light rail to West Seattle, but still substantial money.

      1. Yeah, the more I think about it the more I agree. The WSTT stands on its own as a good idea regardless of the potential ability (with significant capital improvements) to jam a few more trains into an existing tunnel. Ballard/DT and DT/WS are only part of the justification for the WSTT.

    1. It is implied in the main body of this post:

      Alternatively, the extra trains could turn at Stadium, if Seattle were to accept a lower traffic level of service on Royal Brougham Way (or close it entirely). “It’s something we’d have to look at in simulation,” Olson said.

      As SODO has grade crossings there is car/train interaction. Past a certain point either you don’t allow cars enough time to cross the rail line or you end up holding trains to allow cars through. You either create a traffic jam on the road or bunching on the rail line.

      1. Thanks Chris. E.Link trains will have pealed off by Royal Brougham, so the peak of the peak may see two trains two minutes apart, then nothing for 4 minutes. W. Seattle, if built, would likely never need more than 6 min headways.
        Alternately, if not built, those Ballard trains could continue down the busway to the to the BNSF/UP yard, rejoining Airport Link at the south end of Boeing Field and serve as the express trains to Seatac airport. Maybe even a Georgetown and new terminal stop at the airport.
        I welcome a real study on headways down the busway, and what happens to the buses when train headway get reduced.

  14. I would point out that in heavily-used, high-frequency rail systems, failure occurs regularly and occurs when something delays only one train — from a jammed door, from an on-board incident, from a train control glitch, or whatever. Having technical capabilities that allow for more frequency that the schedules have is essential unless the riders want trains off-schedule for a few hours anytime a regularly-occuring delay arises.

    1. Yes, I’m sure that subway systems like Toronto and Vancouver (which have very high frequency) are prone to failure, and the fact that so much of the system (all of those buses and streetcars) are dependent on them have caused so much aggravation that folks have given up and just drive instead. Except they don’t. Those cities are number two and three per capita in North America in transit ridership. But I guess folks don’t care about schedules (which is probably true, since the trains come so often that looking at a schedule is pointless).

      Sorry about the snark, but given the choice between service so frequent that you can throw away the schedule, or consistent, infrequent service, I will take the former every time. We should absolutely spend money (especially if isn’t that much) to avoid disruptions, but we should also strive for very high frequency. For a system like ours (which will never be able to come close to the vast majority of locations) that is our best chance of success (and would be following the model so successfully applied by the two cities I mentioned).

      1. I generally agree with you RossB. My point is simply that it’s good to leave a little time in the schedule — say 30 seconds or a minute — as a cushion for train operations. The major systems that I’ve ridden around the US, like BART and Boston and Chicago, have a little space in train arrivals even at peak hour.

      2. Yeah, I agree with you there Al. I think Sound Transit does as well. That’s why they basically say “three minutes is fine, but below that things get messy”. I’m fine with that. They could, potentially get as low as 90 seconds, but I doubt that will happen. If they really wanted to, they could go two minutes with a thirty second cushion, but I think they will stick with three minutes. The difference between four minutes (which is all they think they need) and three minutes is huge (an increase of 33% in throughput).

  15. I really wish that ST would be more transparent about its peaking assumptions and the STB could discuss it. The percent of daily boarding that occurs by commuters and students at the PM peak hour doesn’t seem to be widely discussed (sources, base year) in any reports yet the capacity issue is brought up on the STB at least monthly. The bulk of the demand analysis by ST was done for environmental studies and funding applications (and this means that ST — as disclosed in the DEIS — assumed Metro and Community Transit route structures that prioritize riding Link to get the numbers higher) and it based on year 2000 commuting behavior, which as we all know had a higher percentage of people commuting and coming home from UW at a PM peak hour than people do today. The heavily used segments disclosed in the Lynwood Link DEIS ridership report are between Capitol Hill and IDS, — which frankly are expected to have high-frequency buses as an alternative if things get crowded — so the analysis is mathematically stacked to be high. Finally, these peak assumptions are for 2035 which is well beyond 10 years after opening ST2. We won’t know what the actual use will be — but in all fairness to us concerned riders and taxpayers, we deserve to explore how this peak is really going to happen based on field findings and not multipliers based on data that is now 15 years old.

    For reference, the report is here:

    1. If 95% of North Link ridership will be peak-hour commuters until the end of time, then we have literally no business building and running a costly full-scale rapid transit system in that direction.

      But if the ridership estimates are better spread throughout the day, the “over capacity” myth can equally take a hike, thank you very much.

      1. There are issues pulling in a number of directions, so the forecasting isn’t necessarily high. Downtown parking costs, major employer attendance policies, parallel and feeder bus routes, fare structures, the cost of fuel — all of those affect projections. I remember reading on here that the 2035 household forecast numbers have been exceeded in Capitol Hill already, so that means that the forecast is low. Honestly, forecasting is full of so many assumptions that agencies do themselves a disservice by not merely reporting ranges. The ST3 planning studies actually did present ranges, to ST’s credit! Still, each of these assumptions affect demand and it would be good to know what these are.

  16. Good point. To me the most important thing is the technical aspect of this report. People have said various things about our headways and potential headways. It is really nice to hear the true story, and to know that we could run four car trains very frequently.

    Estimating the number of riders is always going to be tough. So many things can change. If you ran a subway line through South Lake Union ten years ago it would pick up hardly anyone. Now it would pick up tons of people. But yes, it would make sense to try and update the information, even if will always be a rough estimate.

    When you say “the heavily used segments disclosed in the Lynwood Link DEIS ridership report are between Capitol Hill and IDS”, what percentage of riders will that be? I couldn’t find that in the report. But that is just another reason (along with the technical data, which shows we can dramatically increase frequency at little to no cost) that a line from the UW to Ballard is just fine (and that worries about crush loading are ridiculous).

    1. Careful, you might chafe Lazarus’s “experts”.

      Seriously, though, it increasingly sounds like I should return to arguing for through-routing the Spur trains. Demand on the “peak load” portion of the line — that would be south of UW, i.e. below the merge point — is simply so ridiculously much higher than it will ever be north of the merge point. That is self-evident.

      So in a sane world, we would design our service to take advantage of that!

    2. It should also be noted that as the U-district is a big destination in and of itself, a good number of people who got on the train further north will get off there, freeing up even more capacity for U-district to downtown trips.

      1. Exactly — or head west. That’s the part of this that seems nuts. Some people are really concerned that all of these folks from Lynnwood will be flooding into the city, but none of them want to go to Ballard, Fremont or the UW.

    1. That only works in a Socialist State, where they have been brainwashed that it is feasible – even with 10 car trains.
      Seattle is unique and we’re all coming around to that idea after 25 years of outreach.

Comments are closed.