above: a representation of why I-90 is a better choice

I’d imagine a fair portion of the people who read this blog already know some or all of these reasons that Link is going over I-90 before it goes over SR-520, but I thought I’d enumerate them for easy linking and just to fill in any holes.

I-90 offers a direct connection to the downtown Seattle transit tunnel. If you looked at my earlier tour of Central Link construction, I had a google maps link to the south transit tunnel entrance – you can see there the two tracks we’ve built, plus the space to either side where feeder tracks join with the I-90 center roadway. This kind of a connection offers us the opportunity to interline service – both trains going to the airport (or farther) and trains going to the eastside will come into downtown from the south and run on the same tracks in the tunnel.

It so happens that demand for the northern line (Northgate) is very close to the combined demand for an eastside line and a south line, so having East Link enter the tunnel from the south means that our commute patterns will much more efficiently use our infrastructure. This is also the big reason we didn’t pick buses for building from Seattle to Bellevue – they couldn’t efficiently interline with North Link to increase capacity there.

If we were to cross 520, we’d have two choices, both of them bad: One, we could build a surface level station to transfer at Husky Stadium, and force a transfer for commuters to already full trains coming in from Northgate. We’d create crush loaded trains. The other option would be to build a direct connection into the tunnel toward downtown – which would cost hundreds of millions on its own, potentially have large construction impacts on a residential area, and could be risky due to the depth. Such work would probably also delay University Link.

Even ignoring the capacity and technical issues in Seattle, the eastside would have a problem of its own. 520 is significantly north of downtown Bellevue, so trains would have to turn south first to serve the Bellevue downtown core, then north again to get to Redmond. When using I-90, we don’t have to go out of our way to serve south Bellevue, and the time between downtowns is lower.

Issaquah poses another problem with a 520 crossing.. We’re already planning to build to Redmond, but if we chose 520, later construction to Issaquah (part of the Sound Transit long range plan) would really necessitate an I-90 crossing anyway. With an initial I-90 crossing, it’s much simpler to continue east in or near the interstate right of way.

A 520 crossing would also impose any delays attached to construction of the new SR-520 bridge on Sound Transit’s schedule. The risk added by working with WSDOT on the project would likely also make Sound Transit less competitive for Federal Transit Administration grants.

All this, and I-90’s center roadway was built with conversion to high capacity transit in mind. I think it’s always been the clear choice, but hopefully this convinces more people who were worried about the decision!

50 Replies to “Why Link Will Cross I-90 First”

    1. Actually, it is a huge deal… the light-rail could go three-times as fast if it were under the lake. Why can’t light-rail have its own bridge across the lake?

      1. I shudder to think what the technical challenges alone of tunneling under Lake Washington might be. The cost would be astronomical as well.

        Even a new bridge would be very expensive on the order of a billion or two.

      2. It would at least be good to get an estimate. The length would be comparable to the proposed AWV replacement. The cost for that (although I wouldn’t hang my hang on pre-engineering estimates) are about the same as the floating bridge. No idea how the soil conditions compare but at least you don’t have utility lines, settling of buildings and such to worry about. Although tunnels filled with water (Brightwater for instance) are a big problem for TBMs. I think tunnels have a greater life expectancy and less maintenance. There was one outfit (Swedish I think) that proposed a submerged tube. Unproven and sounds like it would have all of the maintenance issues of a floating bridge and then some. Too bad a cable stay or suspension bridge was never seriously considered. WSDOT Decided an architecturally pleasing structure wasn’t in keeping with the neighborhood. I guess that means the new bridge needs to be at least as ugly as I-90 ;-)

  1. Can’t be that bad: compare the time you’ll save even with slowing down slightly on the bridge with traveling north to UW, getting out of a train, climbing a few stairs, getting on a train again, and so on.

    At first I was skeptical that I-90 light rail would work, as I have personally experienced the “hardship” of ST route 545. But, I gotta say, thanks to Ben, I now realize it’s probably better to put light rail on I-90 first than 520. Heck, maybe it might save me some time in the future instead of taking the 545!

  2. Can someone tell me with regard to East Link, will there be a transfer to the central link trains? Or will the East Lake converge directly with north-south bound trains? I am just having some difficulty wrapping my brain around how exactly the east link will connect to the central link system.

  3. anonymous:

    You won’t have to transfer at all! :)

    “Interlining” (that word I use) means that, in this case, your east link trains will come into Seattle and then just keep going on Central Link tracks northbound. Their first stop on Central Link tracks will be International District station, and they’ll just keep going north from there.

    The other direction, we’ll probably have two colors of train or something – red line to the airport, blue line to Bellevue/Overlake. You’ll see both red and blue trains at Northgate, and they’ll split south of ID station.

  4. Yeah, ballardcommuter, the slowdown won’t be a big deal at all.

    And the thing about the 545 – the 520 replacement project will speed that trip up an awful lot. I’ve been paying attention – I take it every day!

  5. Thanks for laying this out. I’ve heard or figured out some of these points, but not all of them, and certainly not all in one place. I-90 is clearly the better option.

    I am curious how eventual expansion of the system will work. For example, if there’s an extension to Issaquah, will that be a spur requiring a transfer in South Bellevue, or will it interline somehow with your “blue” line? What about extensions along the 405 corridor in both directions? I know that any Ballard/West Seattle extensions will require another tunnel.

  6. I assume the Issaquah extension would just overlap the Bel-Red line on the I-90 crossing increasing the headways between S.Bellevue and Seattle. The beginnings of a network would appear if these interlinings are merely different lines crossing paths on their way to different destinations:
    Something like that.

  7. if you are referring to the two empty lanes on either side of the tracks … those are the road ways for the busses …

  8. you will have

    UW – Airport
    UW – East Side (wherever it goes)
    East Side – Airport

    it makes the junction more complicated since trains will need to go 3 different directions … but it is simple engineering stuff.

    1. Gordon,

      Actually, it’s not “simple engineering stuff”. Because of the curved rise onto the transit structure above Royal Brougham that East Link will use, there is no possibility to connect westbound East Link to southbound Central Link and vice-versa via a “wye” between Stadium and International District.

      The may be the possibility of putting a “pocket track” at ID Station into which the occasional East Link trains could pull, reverse direction and travel south to the airport. And vice-versa of course. But it would probably play hell with the schedule.

      An expensive way to accomplish the same thing would be to use the space devoted to the HOV slip ramps just west of Rainier Station to provide a flying junction and elevated connection to just south of Mt. Baker station. It’s my understanding that East Link will take the entire center roadway for transit (buses will continue to operate among the light rail trains, but no private vehicles), so the slip ramps won’t be needed to let private HOV access to the main lanes for interchange with I-5.

      I’m not saying this is COST EFFECTIVE now, or even when East Link fires up, but I think the necessary space is available should eventual demand for East Side-Airport trips arise. There is no doubt that that train would be WAY up in the air. And the merger with Central Link on MLK would be pretty dicey.

      It wouldn’t be simple engineering. But it is at least possible.

  9. Gordon, those two roadways to either side would be converted to rail. Currently, the buses drive directly on the rails you see in the map I linked to.

    There will not be a wye that I’m aware of – no direct connection from the eastside to the airport until we build more of the long range plan. There’s very little demand for Bellevue-Airport to justify running trains in that route.

  10. Hey cas!

    If you haven’t, check out the old long range plan:

    (Also, I know they don’t show West Seattle, which is part of why I think a city/ST partnership would happen there.)

    Issaquah would interline with Bellevue on the bridge, adding service for Mercer Island and Rainier/I-90, and then turn northbound to go through downtown and serve Capitol Hill, UW, etc.

    The 405 corridor will probably get its own service eventually. My hope would be building, in pieces, a giant loop line from Burien, West Seattle, 2nd Avenue, Ballard (or Fremont), up to Lynnwood, and then back down through Bothell, Kirkland, Bellevue, Renton, Tukwila back to Burien. Hit both downtowns.

  11. Ben … I’d say that would be bad planning … a three-way junction shouldn’t be that cost prohibitive … and would be a boon in the long run. it is much more difficult to add a third connection AFTER the fact than at the same time.

    anyway … do east siders not fly? or are they too good for Light Rail?



  12. Gordon, I think the importance of a one-seat ride may be overstated when a transfer means going to the opposite platform of a light rail station. Certainly if the mode and location of the transfer are the same it removes quite a bit of headache. It’s not like a bus system, where everything is unintuitive and where there dozens stops on the average route.

    Ben S, would the addition of an Issaquah route effectively double the smallest possible headway for Overlake because of bus tunnel issues?

  13. Gordon, the issue is cost-benefit.

    First, there’s a capital cost problem. We already have ramps going directly into the tunnel in the direction we want them. For the opposite direction, we’d have to build new ramps entirely, in an area with a lot of other elevated infrastructure.

    The real issue, though, is operating trains that go from bellevue to the airport. Very few people want to do that – we’re talking about an order of magnitude fewer trips here. We’d have to tie up trains and service hours for a service that would get very little ridership, and drag down the overall efficiency of the system. The issue of transferring by crossing a platform is a pain, but it’s a very small pain compared to the pain we would feel when we tried to pass another ballot measure with the opposition talking about empty trains and bad ridership statistics.

  14. rizzuhjj, the headways I read about in the long range plan (2005) are 2.4 minutes for the tunnel (ID to Everett), 5 minutes for Tacoma, 7.5 minutes for Redmond, and 15 minutes for Issaquah. That’s twice our current Issaquah service (from Sound Transit).

    If we went down to 2 minutes in the tunnel, we could do 5 minute service to Tacoma, 5 minute service to Bellevue/Redmond, and 10 minute service to Issaquah. I don’t know if that’s feasible, but it would be adequate service.

  15. I find it humorous that “capital costs” are a problem for rail on 520. My goodness, this is the most expensive light rail project (in terms of dollars per mile) in the USA. You’d think they would have quit before starting if they were concerned about cost. Priority in design: 1) Engineering marvels, 2) costs, and 3) ridership. Who would plan a rail line underneath the UW campus without a station at the Hub? Who would decide not to put in a station at First Hill?
    Yes, I understand there was opposition to these proposals, but would you please stop pretending that Sound Transit’s decision-making process is perfect and its engineering genius flawless?

  16. Ben,

    If the light rail plan requires every line to go through downtown, doesn’t that in the end reduce the ability to provide more service on the ends? (Metro buses provide much more than 15-minute headways during the peak) I believe BART has this problem. Wouldn’t it make sense for Issaquah to enjoy a line that goes to Bellevue, since they are on the Eastside and likely many people live in Issaquah to be close to major employment centers in Bellevue, not Seattle. So how is Link to downtown going to help them unless there is an additional line to Bellevue that effectively splits the headway in Issaquah and thereby provide more service (7.5 minutes in the peak) from Issaquah to Eastgate?
    They might as well stick to the bus if they want to go to Bellevue, since using Link would mean transferring in Mercer Island then getting to the other side of the platform to reach an outbound train to Bellevue.

    The Link plan, as seemingly “regional” as it is, tries to force everyone to downtown Seattle. You build these beautiful rail lines that are constrained to mediocre headways on the tails because of a singular bottleneck- the hourglass we call Seattle. Under the ST plan, only several miles of the system would enjoy 2-minute headways. And the designed-in constraints would ever constrain it to such. Why are we trying to remake the problems of geography in our transit service?

  17. what is the projected travel time from Redmond to Seattle Westlake? Given the expected short headways, I don’t see nearly as much problem with a two-seat ride. I get most of your points though

  18. curmudgeon, we didn’t build First Hill because it had projected ridership of 5,000 a day and would have cost us $1.1 billion – $350 million for the station, and $750 million because we would have raised our risk so much that we’d have lost the FTA grant.

    The newest project in the US is always the most expensive, when you don’t adjust for inflation. According to the FTA, University Link (the most expensive per mile) is the most cost effective per rider the US has seen in 25 years.

    The UW told Sound Transit they wouldn’t allow stations on campus. That’s their prerogative.

    Take your anti-Sound Transit rant elsewhere.

    multimodal man, I’m sorry you have a problem with the region’s layout, but Seattle’s an order of magnitude larger job center than Bellevue, and that’s going to continue to be the case. Bellevue largely developed as a node because congestion across the bridges created a high marginal cost for the last mile of a previously cheaper commute, as well as because of the limitations of the 1989 CAP initiative. With the removal of CAP and now East Link, a lot of that bottleneck will be gone. Seattle has the ports and the financial center (and that’s not likely to change), so it’s likely that Bellevue’s growth will taper off.

    If you look at commute patterns today, more people go from Issaquah to Seattle downtown than go from Issaquah to Bellevue downtown. Note that I specify downtown – that’s where the light rail line will go, so that’s all that really matters to the comparison.

    In the future, as well, we’ll see a 405 line (as you can see if you go to the Sound Transit long range plan). That’ll mean one transfer (and rail to rail transfers aren’t really an issue like bus to bus transfers are – I’m going to write about that at some point here. It’s likely that a 405 line and an Issaquah line will happen in a similar timeframe.

  19. multimodal man – one more thing. There’s nothing keeping us from running trains Issaquah-Bellevue-Redmond if it looks like there’s demand when we’re building that line. So don’t worry about it too much, you’re talking about ST3 or ST4.

    anonymous, which Redmond? From Overlake to Westlake would likely be 35 minutes, and from Redmond to Westlake would be 40. The two seat ride isn’t as much of a problem as the crush load issue – but it’s still silly to solve the much harder problem of connecting 520 when we have I-90 ready made.

    What we don’t want to do, in general, is replace current one-seat ride bus service (like the 550) with two seat rail rides. That doesn’t end up giving us much benefit – regardless of whether a two seat ride is “that bad”, it does significantly impact ridership.

  20. As an Eastsider who goes downtown AND to the airport, the I-90 routing offers me nothing.

    Of all the trial balloons being floated for the next ST referendum, none of them offer me anything.

    Why would I vote for it again?

    The I90 routing is essentially being built for carless Microsofties who live in North Seattle. It’s useless for most Eastsiders.

  21. brad, if you live on the eastside and go downtown and to the airport, considering that’s exactly where we’re building… what are you talking about?

  22. I think he’s saying that because he lives more north (rural Redmond), it’s quicker for him to take the 520 in a SOV.

    Brad, you could always take the 545 to downtown and then get on the link to the airport or take a feeder bus to the closest light rail station and get to downtown or the airport and save loads on gas or parking. Maybe in 12 years from now you’ll end up moving closer to a light rail stop, who knows?

    Those in Ballard, Queen Anne, Belltown and other Seattle neighborhoods are going to have to take a bus to light rail, too. Of course, bus access in the city is much easier/closer.

  23. Brad, do you commute into downtown, or just go in on weekends? Once 520 is replaced, it will be cheaper for you to park in Redmond and take the train into downtown than it will be for you to pay the 520 toll.

  24. Rizz- Thanks for saying it out loud.

    I AM NOT MOVING. Period.

    Most of you guys sit at your computers, dreamily pondering the day when all of the people in the suburbs will “get it” and tear down our homes and move to TOD near a PnR.

    And you are entirely wrong.

    No wonder there is such a huge disconnect on transit issues in this region. It’s your way or the highway, right? (pun intended)

  25. Rizzz- Also, the 545 is a joke if you don’t work at Microsoft. It’s routing is insane for an express bus.

    Ditto the 554.

    But you knew that already.

  26. Brad… what are you complaining about? You seem to be lashing out, not suggesting anything or being constructive.

  27. Can anyone answer this question?
    What is the expected total number of people per day to cross the I90 inner lanes if they are used for rail compared to the total average number of people per day that cross now by car and bus?
    I am curious about how much more efficient using those lanes for rail is expected to be and why. Thanks.

    1. Hi Nathan!

      In terms of capacity, at least, there’s another factor. Before we close the center roadway to traffic, we’re adding HOV lanes to each direction of the outer roadway – doing this project and closing the center lanes already results in a slight increase in total bridge capacity. The center roadway is bottlenecked by the ramps/lights at either end (which was painfully clear during the I-5 closures last year) – so we only see 1800-1900 vehicles per hour today, and that’s pretty much maxed out. The HOV lane additions will add more capacity than that in each direction (remember that the center roadway only goes one direction at a time).

      Light rail offers an eventual capacity of 12,000 people per hour per direction. While our initial service will be less than half that, ridership takes time to build and service can easily be added as that ridership increases. Also note that the highest ridership route across I-90, the 550, would be replaced by East Link.

      Here’s a neat graphic showing the comparison and mentioning the rather dramatic increase in capacity that East Link would add to the roadway:

      To answer your question more directly: I believe the I-90 center roadway gets 20-25,000 users daily. With light rail, assuming $2/gallon fuel, we projected 45,000 riders per day in 2030. We’ll have much, much higher ridership with $4.39/gallon (today). Those numbers also don’t include transit oriented development – I don’t even know if they include some of the massive residential construction already going on in downtown Bellevue.

      As the center roadway lanes are already essentially maxed out today, light rail offers actual ridership well over double the existing use, and eventual capacity an order of magnitude higher than the existing lanes.

  28. Does anyone know of any (post-Bogue Plan) mass transit research for a route in a straight line between Downtown Seattle and Downtown Bellevue?

    Sure, it would be costly, but by my rough, completely unscientific measurements, it would be a 6 mile trip, instead of an 11-12 mile trip. Let’s say a straight line from 5th and Pike downtown to Bellevue Way and NE 6th Street in Downtown Bellevue, instead of the two circuitous highway routing schemes now being debated.

    1. In terms of cost per rider it would likely be something in the range of an order of magnitude more expensive than East Link. You’d be talking about a $20-30B project rather than a $3-5B project.

      For the money you’d spend on a line that might get you 50,000 riders per day, you could build lines elsewhere getting you 400,000 riders per day. It’s just not worth it.

  29. Are you saying not now for 520 or never? I ask because if ever, doesn’t the bridge need to be built differently?

    A better trend would be whatever gets people who live in Seattle and work in Redmond to just move to Redmond.

    1. There is no room for rail on the plans for a six lane 520 replacement and there’s no way that the public is going to go back to four lanes for traffic in the future. Traffic volumes are only going to go up. We’re not going to lose cars because we run out of oil. We use the same amount of gas today that we did in the ’70s (and it’s going down) and there are way more cars on the road. Electric cars, CNG powered vehicles, ethanol all offer alternatives to gasoline and the other main uses of oil, home heating and electrical generation, are easily shifted to other sources.

      The only way I see that rail could cross the lake along the 520 corridor would be building a separate bridge. Maybe if it was just for rail it could get pushed through but the expense would be prohibitive, well over a billion in today’s dollars. Think what the impact would be on East Link if WSDOT told ST the fees for using the center roadway on I90 are going to have to cover half the cost of the new 520 bridge.

      Even if they did have funding to build a rail only bridge there would be a whole host of other problems. The alignment would have to be along the current 520 bridge; the new one is being built 100′ to the north. So on the west end it would have to go directly into a tunnel making it even more expensive.

      They’re replacing the bridge and redoing all of the interchanges on the west side. I think it is now or never for a rail ROW over 520. One idea at Montlake is a tunnel under The Cut. If that were designed to accommodate rail in the future, and by designed I mean added not traded for vehicle lanes then the option would still be on the table but there’s zero support for building the bridge this way. And I don’t think there’s any talk of making the Portage Bay portion even “rail ready”.

      1. Sigh. I wrote a letter to Obama asking that all stimulus moeny for permanent and difficult road infrastructure components like bridges and tunnels at least be built not precluding rail based mass transit options.

        I also get incredibly smug smirks from anyone living in or near Montlake that ‘we’ won’t have any light rail over the 520 bridge and that is a good thing in their view.

        Maybe connecting up toward the north will happen….and over between the eastside stations.

        Fine, all should not devolve into a downtown seattle.

        But I really wish we would at least build for possibilities when we make these investments.

      2. Sigh. A highway with more than 2 lanes in each direction is a waste of money. Pretty much always. If you’ve filled up that many lanes, you can easily support a rail line, which will carry far more people.

        Pity they aren’t designing it with two lanes in each direction plus a rail line in each direction. Well, here’s to delay in the SR 520 project; maybe delay will lead to a saner result.

        Obviously any bottleneck is a natural for rail; it should run over both bridges in the long run.

  30. above: a representation of why I-90 is a better choice

    The diagram starts with the premise that the line starts where the I90 alignment ends. Take the I90 alignment and flip it. What you end up with is Issaquah/Factoria up 148th Ave NE through Crossroads to Overlake then Bell-Red to Bellevue and final 520 picking up Kirkland. It would be much faster to the U district, about the same to downtown and faster to the airport.

    Either alignment can work. The big risk with 520 is it ties into a mega roads project which is in disarray. The advantage could be that since this is all being built from scratch it would be the chance to do it right for a minimum of added cost. The proposal for a SOV tunnel under the Montlake Cut becomes a tie in to North Link. It’s unfortunate that’s not a political reality. Even more unfortunate is that what we’ll likely end up with is a highway project that’s the most expensive ever built per miles of vehicle lane and one that’s only marginally more effective than what we have today. I know the bridge needs to be replaced before it sinks but when all the concrete is pour the “improvement” is about four miles of HOV lanes which just pushes congestion out to I-5, 405 and the U district.

  31. Ben’s explanations of why I-90 and never SR 520 for Link light rail are entirely correct. Rail on I-90 has been in planning since the 1960s, and locked down as a local government decision since the 1990s. Even the Forward Thrust rail mass transit plan voted down in 1968 and 1970 had the cross-Lake train along what is now the I-90 alignment.

    One additional tidbit is that the original design for the I-90 bridge from the 1970s assumed BART style heavy rail would eventually be installed, since this was the kind of rail transit assumed to be in the region’s future at that time. And now it turns out that light rail cars and the catenary infrastructure for power are HEAVIER than the same equipment for heavy rail.

    There is tremendous momentum for government agencies at all levels to follow through on the ST2 vote and build passenger rail tracks on I-90. However, the environmental review process has at least a year left to run before a final Record of Decision is issued.

  32. Ben,

    Nice Job moderating this list! You have a very good sense of what needs to be said to keep people on task. i am a bus driver and would like to chime in. I live in Columbia City and the south end rail line is really working great for me. My commute is 1/3 the time it used to be.

    East Link needs to go over both bridges! The East Side, Bellevue, Redmond, and Issaquah are all interconnected to Seattle. In one light rail vote that failed research for a line from Ballard along the 44 route to UW and over 520 to Overlake was REALLY exciting to me. This line needs to exist as well as the Seallte, I-90, Bellevue, Eastgate, Issaquah, Issaquah Highlands line. Connecting it all would be the 560 bus route converting to rail and and continuing over the west seattle bridge on the seattle side and continuing North-South in the 405 corridor to Kirkland Redmond Bear Creek. BOTH Bridges!! When you are criticizing one bridge over the other, try to think not about how we use transit now, instead try to imagine how the cities will evolve with a transit system that has several inter-crossed well planned routes. Watch MLK over the next few years, drive down it tomorrow if you have not been down that road in a while. It is amazing to see the growth, but what is more exciting is what is to come now that the train is moving. That is what will be great imagine how the cities will change when the train is traveling over both brides to ballard to west seattle north south. that is the vision, Do not settle for one bride!

    Look at the Eastgate and the Issaquah Highlands Park and Rides! This is well planned out transit! The route 212 is one of the best buses in the system. It gets to downtown from the Issaquah Highlands Park and Ride faster than a 7 from Rainier Beach but it needs the capacity of a Light Rail train to fully come to fruition. Keep the debate going it is really exciting!

    Paul Margolis
    Transit Operator
    Local 587

Comments are closed.