Last week Sound Transit released the Final Environmental Impact Statement (FEIS) for East Link. Press release here. After a group effort to look over the thing, there’s not a whole lot that’s particularly new, although over the next couple of weeks we might post on a few interesting details.

ST's Preferred Alignments

ST’s preferred alignment consists of alternatives A1, B2M, C9T or C11A, and D2A. The final decision on Segment C should come later this summer. The total cost is $2.3-2.7 billion (2007 dollars) for a surface alignment through downtown (C11A), or $2.6-3.1 billion for the tunnel option (C9T). C9T buys you an extra 1,000 daily riders (for a total of 50,000) in 2030.

The Executive Summary has one-pagers on each of the endless segment alternatives that ST considered.

60 Replies to “East Link FEIS Released”

  1. I was hoping someone could clarify why Segment B is surface. According to the EIS, it could be elevated for less money. Grade-separation and saving money simultaneously is a complete no brainer to me. I can only assume they were caving to pressure from Surrey Downs NIMBYs about noise or something…?

    Segment D could also be elevated for a bit more money. Obviously money is tight, but I think we’ll be kicking ourselves very hard in 20 years for not completely grade-separating this thing after pouring billions into it. We’re so close to have a completely grade-separated line here, let’s not throw away this opportunity.

  2. Page 28 of the executive summary:
    The East Link Project has the capacity to comfortably carry 600 persons per 4-car train and 800 persons with crowded conditions. Therefore, with the project, the center roadway would have a peak-hour capacity of up to 18,000 to 24,000 people per hour (equivalent to between 7 to 10 freeway lanes of traffic), which would more than double the person-carrying capacity of I-90 today.

    Apparently it’s being engineered to allow a train every two minutes.

    1. A highway lane can carry 2,000 to 2,200 vehicles per hour. The average car/SUV has a capacity of about 6 passengers. Therefore, with everyone seated, and only cars/SUV’s using the highway, one highway lane has the capacity of about 12,000 to 13,200 passengers per hour. That is more than the capacity of E. Link per track, even using the bogus capacity per train that ST uses.

      Vans have up to 15 seats per van. One highway lane can carry 2,000 vans per hour, giving the capacity of one highway lane with only vans as 30,000 per hour, far greater than East Link.

      One highway lane can carry up to 700 buses per hour. Buses have a capacity of up to 90 passengers per bus. That gives the capacity of one bus-only highway lane of 63,000 passengers per hour, many times higher than East Link.

      Of course, the reality is that ST plans to operate 3-car trains over the I-90 bridge every 7 minutes during peak hours in 2030. That means that the actual planned capacity of the center span will be 3,394 passengers per hour per track on East Link in 2030 — far lower than it is as a highway.

      East Link can NOT ever operate trains every 2 minutes in each direction as long as Central Link is also operating, and ST knows that, so the numbers they give are physically impossible unless they are talking about both tracks combined on the center span. And, even then, of course, they are using bogus capacity per train numbers.

      1. Norman, you know very well that your numbers will never be achieved either, yet you still post them. When combining traffic and the fact that most cars/ suv’s have one single person inside, East Link will always carry more people over i-90. But you can go ahead and sit in traffic while I wave to you as I cruise by on the Link.

      2. The average car/SUV carries 1.3 people. How wasteful and inefficient that 83% of the road capacity is completely wasted when SOVs are permitted to use the road. In order to make the roads efficient, lets require minimum occupancy of 4 during rush hours.

      3. Norman, I will believe you when the City of Bellevue, the County of King, and the State of Washington all mandate full-vehicle carpooling as a pre-condition for issuing vehicle tabs and accessing public rights of way. Any takers?

      4. And how many people does the average Link car actually carry at any point in time?

        Those ST figures are “capacity” — not what it will ever actually carry. You do understand that, don’t you?

      5. One highway lane can carry roughly the same amount of tractor trailers as buses. They are slightly longer than articulated buses (53′ trailer plus the cab; buses top out at 61′) and carry at most two people. Stopping distance is the same as or greater than buses. So I’ll make up a number and say that one highway lane can carry 500 semis per hour, or 1,000 people MAX.

        Further, semis and buses require a six-second spacing. During periods of high demand, SOVs see this six second following distance as “a great place to merge in to”. In order to maintain their six second following distance, the articulated vehicle is required to slow down, which decreases the capacity of the lane because the vehicles cannot operate at the maximum allowed speed.

        Link, of course, doesn’t have this setback and is free to operate at its maximum speed at all times. Link does not tow 53′ trailers that decrease the capacity to haul people.

      6. If you want to use averages of what various transportation modes actually do carry, here is a good source, The Transportation Energy Data Book Edition 29:

        Table 2.12 on page 2-14

        Average passenger/car-SUV: 1.7
        Average passenger/transit rail car: 25

        So, if you want to use national averages for actual load factors you get:

        highway lane: 2,200 car-SUV/hour X 1.7 passengers/vehicle = 3,740 pphpd

        transit rail: 25 passengers/car X 60 train cars/hour = 1,500 pphpd (theoretical maximum Link cars/hour on East Link at 4-minute headways)

        East Link: 25 passengers/car X 26 train cars/hour = 650 pphpd (3-car trains every 7 minutes in 2030)

        So, if you want to use actual average load factors, then the I-90 bridge would average more than twice as many people per lane as East Link running 4-car trains every 4 minutes on each track. And almost SIX TIMES as many people per lane as East Link will actually be averaging in 2030, when it plans to run 3-car trains every 7 minutes.

        And, of course, in REALITY, there are a lot of buses and vans using the I-90 bridge — it’s not just cars and SUV’s, now is it?

      7. Norman, you missed my point entirely. The capacity of the lane decreases when a truck uses it. To quote you: “it’s not just cars and SUV’s, now is it?” No, it’s not. It’s cars, SUVs, vans, buses, and TRUCKS.

      8. You have no point. The accepted capacity of a highway lane is 2,000 to 2,200 vehicles/hour. This is actually observed on actual highway lanes.

      9. The accepted capacity of a highway lane is 2,000 to 2,200 vehicles/hour.

        One highway lane can carry up to 700 buses per hour.

        One of those numbers is wrong.

      10. Caveat is “up to”. You can have 300 buses and 1000 cars per lane per hour but you can’t have 700 buses + 2000 cars per lane per hour. It’s physically impossible.

      11. It’s fun to quote these stats, but they really don’t matter. With east link, you can get from Seattle to Bellevue in about 15 minutes, even when 520 and I90 are parking lots.

      12. Comparing the average car/SUV occupancy to the average rail car occupancy tells you nothing about the throughput at the peak bottleneck. The average car doesn’t change its number of riders on its trip – it will still be carrying 1.3 passengers on the I-90 bottleneck. By the way, the figure Norman cites already accounts for vans, carpools and the like. When you look at the non-HOV lanes, their average load is more like 1.0 passengers, so that makes the non-HOV lane capacity about 2000 people/hour.

        Link, on the other hand will likely have its peak loading across the I-90 bridge and at peak hours, so even if the average across an entire operating day and the entire length of line is 25/car, the load at the peak hour and peak point could well be 150-200/car, which at 800/train and 15 trains/hr does in fact accomodate 12,000/hr/direction, or significantly more than all the SOV lanes put together. So what if off-peak trains aren’t full – what is expensive is peak capacity, and Link lets us more than double the road’s peak capacity for the width of one lane. It’s complimentary.

    2. Tim,

      I think you’re off by a factor of 2. I assume the high end of the range assumes 200 people per car. That’s 15 4-car trains per hour in each direction.

      1. A 4 car train at 200 people per car carries 800 people. “24,000 people per hour” divided by “800 people” = 30 trains, or one every two minutes.

      2. East Link does not have the ability to operate one train every 2 minutes, because they have to share a track in the downtown tunnel with Central Link trains. So, if every other train in the tunnel goes to the East, and every other train goes to Central Link, the East and Central Link routes can have, at the minimum 4-minute headways, even if they operate at 2-minute headways in the downtown tunnel, which is not planned.

        I thought everyone understood this.

        However, there are no plans for headways under 7 minutes on East Link, nor for trains with more than 3 cars on East Link. The greatest capacity ST plans for East Link is 3-car trains every 4 minutes.

      3. Martin, I don’t see “in each direction” in the PDF, but four minute headways sounds more reasonable than two.

      4. Just noticed a typo in my last comment. Should read,

        “The greatest capacity ST plans for East Link is 3-car trains every SEVEN minutes.”

      5. Source, please.

        Not that it matters, because before Central Link went online every other train was proposed to terminate at Rainier Beach. Currently, only gap trains exhibit this behavior.

      6. I’m glad to see Norman pointing out that bottleneck capacity is not equal to system capacity. Since, after all, the number of vehicles an SOV lane can carry doesn’t account for the need for parking at the destination.

  3. 50,000 total boardings in 2030 on East Link. ST couldn’t even come close to projecting ridership on Central Link two years in advance — two years ago they projected about 32,000 boardings per day on Central Link in 2011. They will not come close to that.

    So, how accurate does anyone think that estimate for 2030 — nineteen years from now — actually is? Can anyone really have any idea of what ridership on any transit system will be nineteen years in the future?

    And, does it say anywhere in the FEIS how many of those 50,000 riders per day will be crossing the floating bridge? What is their projection for how many riders will be crossing the I-90 floating bridge on Link trains the first year East Link opens, and in 2030?

    1. Are you arguing that we should not do projections since we don’t know if they’ll be accurate?

      1. I am asking if you think these projections for 2030 are relevant in any way, since many regulars here say that ST’s 2009 projections of 32,000/weekday boardings on Central Link in 2011 are not “relevant.”

        So, if projections made two years in advance are not relevant, are projections made 19 years in advance relevant?

        Or just a load of nonsense?

        I am arguing that ST has proved that is has absolutely no idea how many people will board Link trains even two years in advance, let alone 19 years in advance. That projection of 50,000 boardings per day in 2030 is just a load of garbage. Nobody has any idea how many boardings East Link will average in 2030.

      2. They’re relevant to compare with other projections for other transportation projects that use the same underlying assumptions.

      3. ST’s underlying assumptions have been proven to be faulty by the gross inaccuracy of their ridership projections for Central Link. ST itself says it is going to come out with revised projections later this year.

      4. It’s an estimate, Norman. By its very definition it’s inaccurate. That’s kinda the whole point.

        Is it relevant? For the purposes of an EIS, yes. To politicians and transportation and land-use planners looking for a ballpark as they consider policy? Yes. For professional wonks working to improve modeling? Certainly.

        Other than that, not really. No one voted for ST2 because of a ridership estimate made two decades in advance (really, take a walk around Queen Anne and ask people who voted for ST2 what the East Link ridership estimate is for 2030—but prepare to be maced for being a loony). They voted for ST2 because connecting the region’s major employment, educational, and residential centers with big beautiful trains forever free from congestion struck them as a no-brainer.

      5. Norman,

        The actual ridership levels are irrelevant. So what if the figures aren’t reached until 2040? Everybody reading this will be dead by then, or moved to Arizona. You are hung up on made-up numbers. Get over it – surely you have something more important to do.

      6. [ot]

        Norm has a point about capacity vs reality but the problem I have with his argument is that we’ll never have 6 people in all of our cars going down the freeway EVER! Not only does an average car not hold 6 people (maybe in 1966) but in order for all of our cars to be full they’d all have to be group taxis driving around picking up single people. This is for all practical purposes an impossibility.

        However depending on outside factors (congestion, speed of Link, employment centers etc.) Link COULD reach it’s capacity. Will it? I don’t know and neither does anyone else but because it could perform at capacity it’s the only logical choice.

        Norms comment makes a bit more sense if he were talking strictly buses on I90. They COULD also reach capacity but the experience would still be as crappy as it is now.

      7. the problem I have with his argument is that we’ll never have 6 people in all of our cars going down the freeway EVER!

        He’s not talking usage, he’s talking capacity. He asserts that the freeways are capable of having cars with a capacity of 6 persons. His math is flawed, however, because he assumes that only passenger vehicles and buses use the freeway.

      8. The average car has about 5 seats. The average SUV has probably about 8 seats. I am estimating that the average capacity of cars and SUV’s on U.S. roads is about 6, since there are almost as many SUV’s on the road as cars.

      9. Actual ridership numbers are absolutely important, because they determine the cost-effectiveness of transit. East Link is going to cost around $3 billion. And you are of the opinion that it does not matter how many people will use it? How much are you willing to spend per boarding on light rail? Are you really such a “true believer” in little trains that cost is no object to you? The only thing that matters is that we actually build little trains, and how many people use them is irrelevant?

        I suspect that really is your opinion. I also suspect that most taxpayers do not share that opinion with you.

      10. I don’t know Norm, you may be wrong on that one because we seem to be building a lot of “little trains” that cost a lot of money. This isn’t a dictatorship, people vote for that stuff.

    1. The entire B2M-B7R comparison was also done in 2007 dollars; Arup said that’s what ST used throughout their own analysis.

    2. The numbers have to be consistent across all versions of the EIS so that they are comparable. The first EIS was began in 2007.

  4. I’m not sure I agree that a tunnel alignment in Bellevue will have more riders than the surface alignment. The surface alignment includes 2 stops in downtown Bellevue, and in general riders will be closer to most destinations at the BTC (roughly 108th-110th & NE 6th) plus they have the Main St stop, than the single tunnel station which is located at 110th & NE 4th. The transfer to BTC buses should be faster, and most riders have a shorter walk to all Bellevue destinations – and in fact it may capture more of the apartment/condo construction north of NE 8th St. I know they are factoring in a longer travel time for riders to/from the East, but the greater access to Bellevue provided by the surface route may be a net positive.

  5. As long as Microsoft is still in business, travel time to Redmond is going to continue to be very important. Already, there are enough riders going to Redmond from Capitol Hill to fill an articulated 545 bus every 10 minutes. When East Link get built, we will need to make a decision on whether to continue to provide the 545 for those riders, or ask them to take Link instead (there is also the connector, but it doesn’t provide nearly enough capacity to meet the demand).

    The travel time of Link getting through Bellevue will have a big impact on this decision. A slow crawl means ST will be almost forced to maintain its existing service hours on the 545. A quick pass through Bellevue in a tunnel means ST may be able to get away with running the 545 only during peak hours and using 542 and Link for off-peak trips.

      1. +1, duplicative service is bad, right? Even if it is much faster than the bright shiny new thing that is replacing it, and costing more to operate.

      2. Car runs light, hits train. Train hosed, bus works fine.

        Duplication is a good thing in a large system–it prevents people from getting stranded when one line goes down.

      3. Mike: The difference is that utilization on the 545 is *much* higher during commute periods. At 8:30 AM at the Capitol Hill stop, it’s standing room only. A few hours later, no pair of seats has two passengers.

        In my experience, demand to the airport is much more consistent throughout the day, since flights don’t follow a 9-to-5 schedule.

        Running a peak-only bus from Seattle to Microsoft via 520 would satisfy most of the 545’s current ridership for a fraction of the cost, and it could save a significant amount of time — Capitol Hill to OTC is 17 minutes now, and would probably be more like 33 minutes via Link.

        Running the 194 peak-only would *not* satisfy most of the ridership, and it’s unclear that it would even save much time.

        FWIW, I would be happy to see service consolidation *now*. Delete the 545 off-peak, and use those service hours to increase service on the 550 and extend it from Bellevue to OTC. The ride would be longer, but if that could get us double the off-peak frequency on both routes, I think it would be a very worthwhile tradeoff.

      4. Tim,

        Often, the best kind of redundancy is the kind that is kept in reserve until it’s actually needed.

        In Boston, the MBTA has a fleet of shuttle buses which are only used if one of the rail lines break down. When this happens, these buses (and a number of T personnel) are immediately deployed to all relevant locations. They often have headways of under 2 minutes, which makes sense when you compare the capacity of a bus (even an artic) to a 6-car Red Line train.

        In contrast, if your redundancy story is “use this other line”, then what happens if the other line is already full? Now you’ve got two problems.

        Now, redundancy of infrastructure does make sense. If 520 collapses, we still have I-90. If the AWV collapses, we still have I-5. If a tree blocks one track between SODO and Stadium, we can still run limited service on the other track. But infrastructure takes years to build, whereas emergency service changes can (and do) go into effect within minutes.

    1. The time difference through Bellevue is 2 minutes between surface and tunnel. The train still has to stop in the tunnel station and the tunnel has two very sharp curves limiting speed.

      Link will not and connot be time-competitive with bus service on SR-520 because the routing via I-90 and Mercer Island is double the distance of a direct route on 520. You can make the argument that Link provides the necessary accessibility, and those riders who desire a duplicate express service should pay the full costs of that express bus, and charge a premium for that service. I would make that precise argument for the 59x expresses which parallel Sounder for trips between Tacoma and Seattle.

    2. 2040? I wouldn’t actually bet on Microsoft still being in business. It probably will, but it may well be a shrinking operation.

      There’s probably other stuff in Redmond though.

    3. The important thing to remember is that most of the service hours (and thus most of the cost) happens off-peak, but most of the riders are riding at peak. Thus, by eliminating off-peak bus service between DT Seattle and Redmond, you get most of the financial benefit of deleting the 545, but you get most of the ridership of keeping it.

      In general, I think this is a good middle ground. During periods of high demand, have lots of redundant service to get people where they want as fast as possible. During periods of low demand, it’s more important to reach every destination than to preserve the fastest point-to-point routing.

  6. A1 has four traction power substations, two on either side of the I-90 bridge. I wonder if LRV’s with battery capacity could cut the need for one or two of those?

  7. The FEIS makes clear that a tunnel (C9T) funding agreement must be reached between the City of Bellevue and ST or else the surface version (C11A) will be built instead. Anyone who thinks a tunnel through downtown Bellevue is better should be writing to the Bellevue city council to get an agreement in place.

  8. Did anyone notice that the FEIS essentially rules out B7 and B7R?

    The B7R route was not one of the alternatives that the FTA determined results in “Least Harm” and thus, will not be considered as an option, since it has more problems than a number of other B-segment alternatives.

      1. I believe so, but now it’s got the official backing of WSDOT and the FTA, it’s no longer just ST’s opinion.

Comments are closed.