Could King County Metro and Sound Transit be on a collision course over tunnel usage in 2016?
Could King County Metro and Sound Transit be on a collision course over tunnel usage in 2016?

As Zach wrote about earlier this week, King County Metro and Sound Transit will be running tests this morning to see how many buses Metro can squash into the tunnel during PM peak after Link trains move to 6-minute peak headway in 2016 (See page 106.).

Given the tunnel closure, the Highway 99 closure, and the SR 520 closure today, be sure to check to see if and where your bus is being re-routed. For those who have to transfer at SODO Station, catch route 21, 97, 101, 106, 131, 132, 150, or 594. Routes 21, 131, and 132 are over on 4th Ave S, so be sure to get off at the Lander stop.

Metro is right to try to get as much usage out of the tunnel as possible, without turning it into a worse crawl than now. Sound Transit is right to try to optimize wait+travel time for the state’s largest-ridership transit line (by orders of magnitude after 2016). Unless the two agencies work out a deal, neither may get what they want.

The problem is predominately during PM peak, when the vast majority of downtown riders are paying as they board, and roughly a third of them are still fumbling change. Both downtown street-level traffic and the transit tunnel slow to a crawl, and riders in the tunnel are treated to 2-8 minutes of being trapped on buses and trains waiting in between stations for the next platform to clear.

Allow me to offer a modest suggestion. Let’s call it the “6/10 Compromise“:

Two-car Link trains would run every 6 minutes during morning peak, as planned. ST would replace the 2-car trains with 4-car trains during PM peak, and run them at the all-day rate of 10-minute headway during PM peak.

During the morning rush to work, time is of the essence for commuters. Having the trains run every 6 minutes will make a difference in attracting choice riders. At the same time, the tunnel flows freely enough, with the vast majority of riders exiting at the nearest door, that the current number of buses being packed into a shorter window of platform use shouldn’t be a problem (but would need to be tested).

During PM peak, getting the buses and trains to flow better is a much larger problem. But if the current number of buses suddenly had an extra 2.5 more minutes of window in which to flow through, the delays would drop significantly.

Consider the wait+travel time for train riders: Riders waiting for 6-minute headway trains would be waiting an average of 3 minutes. Riders waiting for a 10-minute headway train would be waiting an average of 5 minutes. (It would actually be less than that, as riders would have an easier time learning the schedule.) If trains are delayed 2 minutes getting through the tunnel, then that wait-time savings is wiped out. Two minutes is the minimum delay getting through the tunnel during PM peak these days. If Metro pushes harder after the tests this Saturday, those delays could go up.

So, it would be great for riders if the heads of Metro and Sound Transit could sit down together and iron out a compromise that keeps the tunnel bus traffic at a level that avoids travel-time delays at least 90% of PM peak, but also allows the current number of buses to continue using the tunnel, until they all get kicked out in 2019.

As an added bonus to running 4-car trains every 10 minutes during peak, that is more seating than 2-car trains running every 6 minutes.

The 6/10 compromise could reduce commute time for train and bus riders alike, and improve the riding experience for everyone.

Update: The Northwest Urbanist got to participate in the tests, and has some video, along with a report here.

94 Replies to “A 6/10 Compromise?”

  1. Why not just keep link frequency the way it is now, and just run 4 car trains all the time? I don’t think that there is a crisis as far as overcrowding on trains as it is, and doubling the car count doubles the capacity immediately, and, at least in the short term, halves the crowding.

    1. I don’t think the number of cars per train effects bus capacity at all. The biggest question is does Sound Transit have enough cars to run 4 car trains with 6 minute headways.

      1. No it doesn’t, but that is of no matter. Until North Link reaches Northgate it does not need to (or at least, that’s what the models say).

        I think this is an great compromise.

    2. You are underestimating how badly the current arrangements work in the morning peak. It’s far from uncommon for Eastside buses to have to wait five plus minutes to access the tunnel in the morning peak.

      While change fumbling and other contributors to slow bus boarding are the biggest current problem in the tunnel, it’s not the only one. The trains have absurdly long dwell times, and the ridiculous spacing required between buses and trains means that a single train through the tunnel consumes a near four minute slot through the tunnel.

      I can’t speak for anyone but Eastside buses, but in my experience, the only real problem in the morning is getting to fourth. Would turning left onto Airport rather than winding through the ID work any better? I’d be sorry to give up the only decent connection to Link and the FHSC, but at least the latter’s interesting destinations might be just as well served by surface buses further into downtown. In the afternoon, there’s also the slog through 2nd Ave traffic and up to 5th is a nightmare, 5th is still no picnic, and there are also issues with stop capacity. The dwell times on those CT Double Talls is something to behold.

      Any solutions that involve speeding up bus boarding need to be robust in the face of failure. It only takes one wheelchair bound passenger having trouble getting safely secured to turn what ought to be a 45 second stop into a 3+ minute ordeal. Add that to the four minute hole that each train puts in the schedule, and your six minute headways are shot. How quickly can you recover? We have both a legal and a moral obligation to serve these passengers, so we need to be certain that whatever we assume accounts for them.

      It’s essential that we maximize the capacity of existing slots, by running vehicles of maximal length and by educating passengers to pack in better before we start reallocating slots.

      Finally, I take issue with your “orders [emphasis mine] of magnitude … after 2016″ claim. The 550 is on course for 2.5 million boardings this year, and (unless it’s deliberately made less attractive to accommodate Link) there’s little reason to expect a future decline. So orders magnitude more would be well over 200M annual boardings for Link no later than 2017. Can we hold you to that prediction?

      1. William, I don’t doubt that the 550’s ridership will continue to grow quickly, but even this year, Link ridership will go well past 10 million boardings. That’s a couple binary orders of magnitude lead, before U-Link opens. But I’m not trying to debate terminology.

        As a side note, ST needs the county’s permission to add more peak-direction 550 runs in the tunnel. Given than more 550s are / will be needed, Metro and the county have leverage on this point.

        I totally agree that ST should justify wasting half the platform space during joint operations. The wall in the stub tunnel will soon be gone as an excuse. I haven’t asked ST yet for its justification, given that this is mostly a post to bring out ideas and shoot some down.

        I hear you about the morning peak not working with 6-minute Link headway. So I’m glad you chimed in.

      2. It seems like just letting buses trail right behind the light rail cars would go a long way. That seems pretty harmless but maybe we’re missing something. Is there some federal regulation about spacing or something?

    3. The fleet is 62 light rail vehicles. It takes twelve trains to providie 10-minute headway on the 114-minute loop. Fifteen trains are needed for 7.5 minute headway, with no congestion, so realistically sixteen or seventeen trains. Nineteen trains are needed for 6-minute headway assuming no congestion, so realistically, up to twenty-two trains.

      There is no capacity reason to run 4-car trains at better than 10-minute headway any time except maybe after mega-events, as this is already more capacity than 2-car trains every 6 minutes.

      The main reason not to run 4-car trains all day is the cost of electricity, which, IIRC, is more than half the cost of an operator per platform hour. My estimate is that the extra elelctricity would cost on the order of $10,000 a day, or $2.5 million a year. It is, admittedly, a very rough guess.

      That said, I’m guessing 3-car trains would be enough, even at peak 10-minute headway, for at least the first year. ST plans for 2-car trains to be enough, at 6-minute peak headway, all the way to 2019. 3-car trains at 10-minute headway provide 18 LRVs per hour per direction capacity, which is almost as much as the planned 20 LRVs per direction per hour at peak.

      A million dollars or so of annual operational cost, to reduce impacts on dozens of bus routes, and make a good impression on voters, with much-smoother tunnel operations, and perhaps smoother street-level operations, in the run-up to a possible ST3 vote, is a wise investment, IMHO.

    4. From the 2014 Service Implementation Plan, it looks like Sound Transit won’t have enough cars to run 4 car trains, at least for every train during peak hours through 2019. Table 5 show that, at peak they’ll have 42 cars running in 2 cars sets, with 20 cars available as spares.

      1. That’s a pretty high spare ratio. But the SIP is just a projection, things will be adjusted when U-Link actually starts service.

      2. @aw

        If I understand the SIP correctly, I don’t see how it’s possible for ST to run every peak train as 3 cars- let alone 4 cars- while maintaining 6 minute headways. They’d have to buy and take delivery of more cars.

        Does anyone know when the cars for the Northgate Link Extension are supposed to be delivered?

      3. They haven’t even begun to think about the next order of cars yet.

        I think the RFP goes out next year?

        I would imagine that we’ll definitely see interest from:
        Kinkisharyo (for obvious reasons)
        Siemens (their S70 is proven all across the US)
        — the S70 comes in 80′ to 96′ (3 segment) and 120′ (5-segment (used in France))

        I would imagine that Kinkisharyo would use their new LA factory to construct new cars for us and Siemens already builds them in the US in Sacramento at their factory

        Alstom Citadis (Spirit) (they want to make a big move into the US market)
        Bombardier Flexity Link

        Outliers (unlikely):
        CAF (have had a lot of issues with delivery)
        Vossloh Citylink NET (making Tram Trains for Sheffield)

        NO way:
        Ansaldo Breda

      4. Its possible we’ll see proposals from all the vendors you’ve listed unless something in the RFP makes it unlikely a particular vendor will win.

        If Ansaldo Breda submits a proposal hopefully Sound Transit will find a way to disqualify them. I can’t think of one US city that has had a good experience with their equipment. Supposedly they lowball their bids which is why they keep winning contracts.

        I can’t say as I have any preference between Kinkisharyo, Siemens, Bombardier, or Alstom. They all build decent stuff. One concern might be that Alstom doesn’t have much experience in the North American market.

        I don’t really know enough about either CAF or Vossloh to have an opinion. A lack of experience on the part of the vendors plays a role here too.

        All things considered no matter who submits a proposal it will likely come down to a contest between Kinkisharyo and Siemens.

      5. I hope they will do something to make the ventilation system quieter in the next batch of cars. Or maybe Siemens knows how to make them quieter.

      6. Philip,

        Brent didn’t propose running four car trains — or even three car trains — on six minute headways. He proposed four car trains on ten minute headways.

        Twelve times four is forty-eight, which leaves three full length spare trains and one half length spare train, minus any cars which are in heavy overhaul of course.

      7. @Anandakos,

        I was responding to Alex’s comment, which I misread as advocating running 4 car trains with ST’s planned 6 minute headways. Re-reading it, I see Alex is proposing running 4 car trains at 7.5 minute headway. Looking at the SIP, running 4 car trains at 7.5 minute headways with the U-Link fleet could be possible though it would be tight- according to Table 5, there are currently 30 out of 62 cars in use at peak, so doubling each train would leave just 2 spare cars.

        I agree that Brent’s proposal for 4 cars trains running at 10 minute headways during the PM peak should be possible.

      8. The new Panamá Metro cars are Alstom, level board (high platforms but with pantograph and overhead power rail). They seemed pleasant enough; smooth acceleration and deceleration.

        I hope that whatever is selected in the next round they get something better looking than our Kinkisharyos. Those things are so clunky-looking they look like overgrown streetcars rather than a “rapid” transit line. There are certainly LRT vehicles out there that have an appearance more befitting our core rail lines. Perhaps it’s my design background, but even the appearance of speed can help to pull in choice users.

  2. I have a better solution: Eliminate change fumbling (and *all* fare issues) in the Tunnel by installing turnstiles. Let’s remember that the reason Tunnel stations were built with open access is because of the Free Ride Area and pay-leave going outbound, which no longer exists. The Tunnel was never intended to have on-board fare collection be a factor, except perhaps at night when the FRA wasn’t in effect.

    1. The issue is Link, ST express, and Metro buses all have different fares.

      Perhaps install readers at all doors for buses using the tunnel. Use turnstiles or require off board payment in the tunnel. Charge the highest fare for the first tap and refund the balance when the rider taps off.

      Some other reforms:
      Make ORCA cards cost $1
      Offer 1, 3, 5, 7, 14, and 30 day passes (30 day passes replace monthly passes)
      Increase the number of locations one can purchase an ORCA card
      Eliminate paper transfers
      Add a cash surcharge

    2. There are two unaddressed issues with this idea: two-zone Metro fares and Orca revenue allocation.

      Since the tunnel buses cross a fare zone boundary, you need to be able to charge a two-zone fare for those leaving the city, but only a one-zone fare for those staying inside (including those catching the 7X and 41). Turnstiles can’t enforce that.

      Also, if fare paid with Orca is collected at one point for all modes, then the agencies don’t know how to distribute that revenue put of the revenue pool.

      I have solutions for both:

      1. Reform the Metro fare structure to be level-of-service-based (express/limited/local) instead of distance-based. Every bus that enters the tunnel is considered an “express” and costs a premium.

      2. Use estimates from the onboard counters to apportion revenue.

      1. Another option would be to require tap on / tap off like Link. This would require readers at all doors on tunnel buses but would be cheaper than turnstiles. It would also avoid charging express fares for in-city trunk routes like the 41, 71, 72, and 73.

      2. Reform the Metro fare structure to be level-of-service-based (express/limited/local) instead of distance-based. Every bus that enters the tunnel is considered an “express” and costs a premium.

        I’ve proposed a similar idea many times in the past. However, it’s important to make sure that you’re not creating perverse incentives.

        For example, consider the 41, and the 71/72/73/74. From Metro’s perspective, it’s actually cheaper to have passengers ride the 41 than it would be to have the same passengers ride the 40 all the way from downtown to Northgate. The total cycle time for the 41 is much less than the 40, so you can provide a given level of service with fewer buses (and thus at a lower cost). Similarly, it’s better for Metro if passengers ride the U-District expresses, rather than riding the 43/49/70 between downtown and the U-District.

        I think an express premium makes the most sense for the services that are actually more expensive to provide, namely the peak-only expresses. I think it’s entirely fare to charge double for a ride on a one-way bus: you’re paying both for the trip to your destination, and the deadhead back to your origin. Even if the bus has an in-service deadhead, peak buses are still more expensive than all-day ones. The number of vehicles that Metro needs to own is bounded by the number of distinct buses that are operating at peak-of-peak, and peak buses often require paying some number of drivers to do nothing for a portion of their shift.

        Of course, this doesn’t directly address the problem at hand. But it does help it indirectly: it reduces demand for peak expresses (which lets Metro reduce service), and it raises more money that Metro can use for its other services.

      3. Aleks: it’s also important to figure out how many of the 41 and 7X riders are paying with their own unsubsidized money. I’m not sure about the 41, but I guarantee that percentage is in the low double digits, or perhaps even single digits, for the 7X.

        Keep in mind that there’s also the 66 (which in my scheme would be branded as 66L—Eastlake Limited), which would retain its current fare structure.

      4. Also, I understand the attraction of tying the express fare to the marginal cost of providing that specific service, but this is a complex system, and just like all complex systems the total is not necessarily equal to the sum of its parts.

        Incremental gains in Econ 101-style efficiency in one area of the system might actually have a deleterious effect on the system as a whole. Not saying you’re wrong, but it’s important to keep that attitude in check.

      5. Kyle,

        I don’t mean to suggest at all that the price of service should be tied to the marginal cost of providing it! Rather, I’m talking about incentives. When there is more than one way to get between two points, I think it’s wise to use pricing as a way of encouraging people to choose the option that is cheaper for Metro to provide. This way, if people follow the pricing signals, then Metro will be able to use its budget to serve a greater number of riders than it could otherwise serve.

        You’re making a separate point, which is that the price of service should be tied to the ability and willingness of customers to pay for that level of service. To put it another way, you’re arguing for price discrimination, i.e. charging more for certain services largely because the riders of those services are willing and able to pay more.

        I think that’s a reasonable approach, too, but you can also take it too far. It all depends on the elasticity of demand for the premium services. For example, let’s say that Metro charged $100/ride for the 71/72/73, under the assumption that everyone is using a U-Pass, anyway. If 100% of riders stay on the express buses and pay premium fares, then that’s great! But if 100% of riders divert to the 49/70, that’s not so great. All of a sudden, the 49/70 are more overcrowded than it’s ever been, and it’s virtually impossible to access Eastlake or 10th Ave E by bus.

        Putting this all together, I think premium fares are most appropriate for services that meet the following criteria:

        – The premium service has desirable qualities for riders (i.e. shorter ride, fewer transfers).
        – There is an expectation that some proportion of riders will be willing to pay the higher fare.
        – If most or all existing riders shifted away from the premium service and towards the non-premium alternative, Metro would be better off.

        Using these three criteria, I think there’s a strong argument to be made for charging a premium fare for many one-way peak expresses, and especially for the services that run along I-5 south of Seattle (i.e. they compete with Link). But I don’t think it’s appropriate to charge a higher fare for buses like the 41, or the 71/72/73, or the 15/17/18, simply because Metro actively wants riders to choose these alternatives over the equivalent local routes.

    3. I think you’re all over-thinking this.

      Simply reprogram the existing TVM’s in the Tunnel to sell Metro 1-zone and 2-zone tickets.

      When drivers enter the tunnel, they place a bag over the fareboxes… forcing all customers to flash a transfer or tap their ORCA.

      It’s a good stop-gap solution for the few more years the buses remain in the tunnel.

      That being said I support Chris Stefan’s other reforms including lowering the ORCA cards cost $1 (LA does this), eliminating paper transfers and adding a cash surcharge.

      1. Simply reprogram the existing TVM’s in the Tunnel to sell Metro 1-zone and 2-zone tickets.

        We really need completely off-board payment.

      2. The TVMs we had here in Portland happily sold 1, 2 and 3 zone tickets when TriMet still had 3 zones, so it isn’t as zones are an issue with printed tickets from TVMs.

        If they move the tunnel to entirely off-board fare payment, then they will probably need more ORCA readers anyway. So, why not have separate readers for Link and Metro buses?

        Of course, this is one example of how it would be rather nice if there were a single fare structure across the multiple agencies (so someone could buy a single printed ticket valid for buses or Link).

      3. It would make so much sense on many levels to have a single fare structure inside King County, such that the fare is the same from point A to point B, and you can take whichever vehicle comes first, or is fastest, or is less crowded, regardless of which brand is painted on it, regardless of whether it is rubber tired or steel wheeled. Multiple fare structures and fares create complexity, and increase costs, and frustrate riders – those costs far exceed any incremental revenue you generate, and even worse when it causes agencies to operate redundant service. If you want a distance element to the fares, then create zones. But why make Link distance based and not the buses? Just because you can do it doesn’t make it a good idea or worthwhile. If we had a single agency determining the fare policy, by now we would have figured out how to make ORCA universal and gotten rid of impediments like the $5 media charge which increase operating costs by discourage ORCA use

      4. I’ve been doing a lot of research on train fares, and noticed an ironic trend: Instead of making distance-based fares feasible on all-day frequent train lines, smart cards have enabled higher-distance fares to be easily defeated (except for station-specific surcharges). All one has to do to defeat Link’s $2.75 top fare is to get off the train at Rainier Beach Station, tap off, tap on, and get back on. For those paying per ride, that saves the rider 50 cents. Distance-based fares serve a purpose to utopia-minded theoretical engineers, but don’t seem to achieve much in reality, besides longer lines at the ORCA vending machines.

        The overwhelming majority of all-day passenger train systems in the US have a flat fare that is equal to the local bus fare.

      5. Most systems in Europe have a zone-based system and the bus fare and rail fare are identical for travel within or between zones. Generally the city limits are all in one zone so it is effectively a flat fare system for the city, with additional charge for going into the suburbs/region. They do not do strange things like charge you more to ride the train than a bus, or have the bus be flat fare while the train charges by distance.

      6. That assumes 1. You get transfer credit for the first leg, which is not the case in other cities with distance fares, 2. You are willing to spend 7.5-15 minutes (or up to 30 minutes if you rely on a bus connection) to save 50¢. This loophole can easily be fixed. In Asia, the case is opposite the US and distance fares are the norm to recover a greater share of operating costs. It’s hardly theoretical. It’s just different goals they want to achieve with the fare structure.

    4. Turnstiles are not happening for just 3-4 more years of joint operation, and ST has lower fare evasion rates than its peer agencies using the proof-of-payment system. Turnstiles are not feasible for the at-grade stations. They are not happening.

      1. +1 – Proof of Payment (PoP) is sufficient to drive down fare evasion to reasonable levels, especially with a few reforms.

        Metro’s RapidRide fare enforcement is effective, but there are limits. Paper transfers don’t have any indication that a reduced fare or youth fare was paid so there is no way for fare enforcement to check eligibility. They also don’t appear to have the authority to ask for proof of age for youth fare (A “youth” with substantial facial hair and other features that indicated he was older than 18 tried to pay his fare on my E Line in front of Fare Enforcement. The FEO indicated that they were not able to check ID – Sound Transit fare enforcement can and does check ID – I spoke with at least one person who tried to cheat by paying a Youth fare on Link, was caught, and paid a fine). And then there’s the whole issue of paper transfer fraud which makes any estimate of fare evasion highly suspect. They simply can’t know how much transfer fraud is being perpetrated. The system *could* be improved, but long term I believe management wants to go cashless/ORCA only, which would be fine by me.

        It should be pointed out that adding PoP to the entire system would give Metro Police a potent tool to address Rider conduct issues. By default, if you don’t have PoP, you can be removed from any bus, given a warning/written a ticket, and given a lecture for whatever disruptive behavior you are engaging in. (Metro Police, BTW, don’t have the ability to read fares but do carry Android smart phones – FareBot, anyone?) Sure, PoP and Fare enforcement cost money, but it isn’t just about ensuring revenue. Most drivers of RapidRide I’ve spoken with report improved passenger behavior after fare enforcement is added. But that improvement virtually disappears when fare enforcement goes away. The E Line, after 7pm, is largely a free ride zone to all but the dedicated who dutifully file on the front of the bus. (Telling drivers to ask people to board the front after 7pm is pointless. It’s a recipe for fare disputes which we are expressly forbidden from engaging in. This is a common issue with night RapidRide drivers)

        PoP is the solution. Fare enforcement can be applied randomly and sampling can determine where limited enforcement resources are best utilized. Just do it already.

      2. @Velo
        How does the E compare to the old 358 as far as rider behavior goes?

        I noticed the A is nothing like the old 174 even in the evenings.

        I agree going PoP system wide might solve many problems especially if both police and FEOs have the ability to check fare media.

  3. Not allowing on board payment in the tunnel seems like it would make things faster for everyone. Either turnstiles as mentioned above or a requirement to purchase tickets prior to boarding would work. Otherwise i’d kick Metro out of the tunnel for good.

    1. Yeah, turnstiles are probably overkill. There are already fare enforcement officers in the tunnel for Link, you could prob spread them out and add a few to cover the tunnel busses. Improving the operations in the tunnel will ultimately save money (drivers waste less time), so you’d use those savings to pay for the enforcement.

    2. I don’t see banning on-board payment as happening, as that would probably lead to the return of the dark days of PAYSTTE. However, a ban on cash payment in the tunnel is feasible. It would probably force everyone to get an ORCA, but I’m totally down with that. Can’t afford an ORCA from the machine? I hear that the low-income ORCA may end up being free, if you qualify. Down with cash fumbling in the tunnel! SInce the county council is unlikely to take the political risk for 3-4 more years of joint operation, make it a prelude to the creation of a cash–fumbling free zone throughout the Central Business District, permanently.

      Metro is already taking a huge hit for the low-income fare program. Reap the investment by doing all the things we couldn’t do before because of the $5 cost of an ORCA cards.

      1. Just lower the cost of the fucking cards. Other agencies using the same vendor only charge $1 for a card.

  4. What about putting some of those “Expect Delays” signs in every station? It’s a solution for delays on roads with construction.

  5. In Boston, the Harvard Square MBTA station serves both the Red Line (subway) and several bus lines. Because of how the tunnel is constructed, there are a small number of trolley buses that drive through a tunnel with their right side against a wall. The MBTA deals with this situation in a very interesting way. The buses in question have three doors: two on the right, and one on the left. You board the bus using the two back doors (left in the tunnel, right on the surface), without paying, and you exit the bus at the driver’s door, paying as you exit.

    I wonder if we couldn’t do something similar here. To be precise:

    – Kick out local buses (i.e. the 106).
    – Outbound buses (starting in the tunnel) go back to “pay as you exit”, 100% of the time (no more of this “before 7PM” BS). However, if you board one of these buses after the express segment, you’re allowed to pay as you enter, so long as you retap your ORCA card (or show your transfer) on your way out as well.
    – Inbound buses (ending in the tunnel) are unaffected.
    – Signs, audio reminders, and Metro personnel all communicate the message that you shouldn’t board buses in the tunnel unless you’re planning on exiting after the nonstop segment, and that Link is cheaper anyway.
    – At least for the first month or so, security guards in tunnel stations will stand outside outbound buses, and will ask any passenger who exits the bus to show their transfer or their ORCA card. Anyone who fails to provide proof of payment will be handled in the same way as Link/RR non-payers.

    The advantage of doing this, compared to other proposals, is that it requires no capital spending (unlike installing turnstiles, say), and preserves high frequencies on Link during all periods. The disadvantage, of course, is that it would be confusing. But that might not be the end of the world, especially if relatively few passengers board an outbound tunnel bus after the express segment.

    1. This brings up a good point. Didn’t Metro run tests of removing PAYE and decide its impact was small enough to remove it? If so this means that we should be skeptical of their tests, and possibly that removing PAYE was a mistake. I’ve always thought we should have waited for full POP before we removed PAYE, and that the evening commute would be awful without PAYE. It certainly had its problems, but it makes the best use of limited downtown curb capacity.

      1. In my experience, PAYSTTE took longer for most riders than the time-cost of getting through downtown with PAYE. We have many better ways to mitigate PAYE which have been discussed at length here and many times before.

  6. It’s also worth reiterating Zach’s excellent article on the subject of tunnel buses.

    I think the basic principles still hold true. Maintain common corridors, but don’t waste the tunnel (an all-day resource) with peak-only routes.

    Kicking out the 77, 101, 102, 106, 150, and 255 would reduce PM peak loads by 10 northbound and 16 southbound buses per hour. Adding the 554 would only increase PM peak loads by 2 southbound buses per hour, since the 554’s frequency during peak is lower than other periods.

    To make this deal fair to Metro, Sound Transit would agree to take over the 101, 150, and 255. Sound Transit still wins by getting to run Link at their desired frequency, and by moving the 554 to the tunnel (and its dedicated I-90 ramps). Metro wins by jettisoning three frequent (i.e. expensive) bus routes, without actually reducing service.

    1. Instead of cramming every bus into the tunnel how about dropping people off on the busway at say Stadiurm Station and letting them take LINK in?

      1. John, Given the choice of having the 169 continue up to Rainier Beach Station via Sunset/MLK very frequently, or having it go all the way to SODO or Stadium Station via Spokane St less frequently, which would you prefer? There is also the option of having it go all the way up to the Seneca St. exit, but even less frequently. Whichever way, it is high time the 169 connect to Link somewhere, and for the 101 to leave the tunnel.

        I am a fan of decoupling my 132 and having it start at Stadium or SODO Station, FWIW. I wait way too long at Lander for a bus showing up at some random time to get me to South Park. I’d take the reliability and a transfer over that situation any day of the week. I try to ride the 60 instead now, when it is available, but the 60 route path is set to change when the 106 changes, with the 60 serving new purposes I can support, like the connection between White Center and the Othello business district. My thoughts on the subject might be colored by my dream of moving next to Othello Station.

      2. As far as the 169 and Rainer…my thinking is along the same lines as here and in regards to express buses and Kent Station. If you can get a bus to a rapid rail line, there is no reason for the bus to make the trip all the way into downtown!

        Ideally, by this time, we would have had a LINK station at the South Renton Park and Ride and that would have been a hub linking Southeast King County to Seattle and Bellevue.

        But until that day comes, yes, I think we can make due with Rainier Beach as the hub, at least to Seattle.

        So yes, do away with the 101/106 routes, and just have the 169 start at Rainier Beach and then head to South Renton P&R and then Benson to 108th as the 169 now does. Add a RapidRide from Covington to Kent East to Renton to Rainer making key stops like Lake Meriidan, 240th .. as well.

    2. Hmm, if Sound Transit took over the 101 and the 150, would it be more motivated than Metro is to turn them into shuttles?

      1. I don’t think that necessarily follows. Metro has done some major reorganization already to connect buses to trains outside of downtown, with more in the pipeline if the county council doesn’t get in the way. The A and F Lines were designed specifically with Link connections in mind.

        I’ve seen much less movement from ST to push out-of-downtown connections. For starters, the only bus route advertised at any of the Rainier Valley stations is the 97. ST has expressed zero interest in eliminating the 545 and moving the hours to the 542 (which I can understand, given its high ridership and the fear of the transfer penalty). The 592 is a much larger waste of tax dollars than any other route we beat up on here. ST has priced Sounder to encourage riders to choose express buses. For ST, connections to Link and Sounder is something left for Metro to handle.

      2. “If Sound Transit took over …” Psst, all they would do would be to slap ST stickers on the bus, but the bases, drivers, supervisors, mechanics, etc., etc., etc., would still be the same. When it comes to their bus system, ST is a transit agency in name only. They operate nothing.

    3. One large problem with ST taking over the 101 and 150 is that it would be perpetual. Since it would come out of ST’s south King subarea budget, a few more years might be added to the projected timeline for Link to reach downtown Federal Way. I don’t see ST going for such a bargain just to get these routes out of the tunnel for three years. Moreover, I don’t see that ST would necessarily turn these routes into Link connectors. It might just push them upstairs.

    4. I’ll mention again my idea for the SODO Transit Station where routes like the 101, 102, 106, 150 would terminate and riders would make connections to Link. I would also terminate the D and E RapidRide lines at SODO which would offer even more opportunities for easy connections at SODO. Then, add in a major bike share facility at SODO and riders could be onto their final destination in a matter of minutes.

      With the SODO transfer hub, a route like the 101 could lower its headways to 20 minutes without adding any extra service hours just by avoiding the trip through the Busway and DSTT.

    5. They really ought to optimize the tunnel for buses that use the I-90 dedicated ramps and south busway. The 255 makes little sense in the tunnel since it has a painfully slow route to/from I-5 all the time. The other I-5 buses at least can use the dedicated ramp for half their trips.

  7. I am very much against any solution that involves running less frequently. Which would not help the delays caused by cash-fumbling bus passengers anyway. Platforms need to be prepaid zones, period.

    Too often we let the perfect become the enemy of the good. “But someone might not pay for their two zone fare!” “But ORCA revenue might not be allocated perfectly!” You know what? The vast majority of people are honest and want to do the right thing. Making it a prepaid zone would immediately cause most people to pay correctly. Random fare enforcement occasionally would help. Allocation of revenues would be derived from the onboard counter totals.

    The main problem I see with this is that basic tapping of ORCA cards introduces the question: which fare are you paying? Since they are different amounts. But perhaps the way to do that is to just charge the bus fare to the card, then when Link users tap off, readjust their totals to the correct Link fare — which already happens anyway.

    This is not a perfect solution either, but something needs to be done. And that something absolutely should not be “returning to pay as you go” or increasing the time between trains which penalizes train riders because various transit agencies around here can’t get their act together.

    1. Just to be clear: Would you be willing to have the train trip take two minutes longer, if it means you’ll be waiting two minutes fewer for the train to come?

      1. I’m not always traveling the entire length of the tunnel or even entering the tunnel. So I won’t always have to wait those two minutes of tunnel delays — but reducing the number of trips would affect all of my trips. So, yeah. I’d rather have more trains per hour.

  8. I love the suggestions about moving bus payment off the buses to speed up service. These changes should be applied not only in the tunnel but also on all parts of the frequent transit network!

    I would like to add that the solution to wheelchair boarding delays has also been known for some time: passive wheelchair restraints. Simply put, the wheelchair rider turns their chair to face the rear of the vehicle, backs up against a bulkhead and locks the wheels. This system has been used on bus systems all over the world and there is no reason why it can’t be used on bus routes in our city. I would have some concerns about bus routes that travel up steep hills, but perhaps there could be optional straps on the buses for when buses travel on them. Alternately, the buses that utilize this system could be assigned only to relatively flat routes or explicitly branded for these routes, such as with Rapidride.

      1. It is cost neutral if we specify passive restraint on all new bus orders. Alas, that has not been the plan, and Metro isn’t sold on universal passive restraint.

  9. While I would love for POP to become the fare system in the tunnel (and elsewhere), there are still serious implementation issues, beyond just the morass of different fares. Most of the tunnel buses are crush-loaded during peak. About the only option for fare enforcement officers is to be waiting at various way-outside-downtown stops. But those who don’t want to get caught will stay on the bus to avoid getting checked when they see the FOE. They’ll also mass-text their network of fare evading friends to avoid getting off at that stop.

    Or the FOEs will have to ride the still-mostly-crushloaded bus further along, and have to wait for a less-frequent return bus when they are done.

    The buses themselves are not designed well for FOEs, unfortunately.

    1. The worst habitual fare evaders aren’t usually riding crush loaded peak buses.

      Depending on the route if you want to sweep a crush loaded peak bus there are opportunities to do so. For example FEOs can check people getting off of a 41 at Northgate and sweep the few remaining passengers on the bus.

  10. If we are discussing tunnel ops, then I have a question for the turnstile proponents. Where are you putting them? You either have to put them at the initial entrance to the station and move the TVM’s to the outside of that or you would need to put a turnstile in front of the elevators that lead to the platform level.

    These stations weren’t built for turnstiles. To put them in would require more construction than just simply adding them.

  11. The future of the DSTT is “bus only.” There is no reason to start reducing frequency of service on your only blue chip line in an attempt to delay the inevitable. Metro needs to start reducing the number of buses using the DSTT so that Link can run freely. Link is the future, Metro shouldn’t be working against it.

    Fewer buses in the DSTT during peak times will solve the problem. It is time for Metro to get out of ST’s way. And Metro certainly shouldn’t be wasting taxpayer dollars trying to deny a future that we all know is just a few more years away.

    1. Would you be okay with 3-car trains running 7.5-minute peak headway (which is more capacity than currently planned, but well within fleet size), and then ramping up the headway to 6 minutes in 2019?

      1. No. If Link is running 6 minute headways during the am peak, then they should be running 6 min headways during the pm peak.

        Buses In the DSTT have always been fickle and unreliable. Metro has never been able to operate the DSTT to its pre-opening promises. At this point Metro should not be forcing compromises and additional costs onto ST. They just need to align themselves with the future.

      2. It takes two, Lazarus. Sound Transit shouldn’t take the attitude that it will do what it is going to do because it can, and push costs onto Metro. As a taxpayer and bus rider, I’m strongly opposed to running up Metro’s operating costs when Metro is broke.

        Let me rephrase my question: Would you be okay with Link running 7.5-minute AM and PM peak headway (with sufficient capacity) until 2019? especially if it enables travel time through the DSTT to be held below nine minutes most of the time?

        Would an agreement that Metro will continue to reduce the number of buses it runs in the tunnel until that threshold is met matter to you?

      3. Nope. I would not be ok with ST spending limited tax dollars to develop two different operating schemes, one for am and one for pm. And I would not be ok with Metro forcing ST to reconfigure their trainsets twice per day, every day of the week, for years just to delay the future that everyone knows is coming in just a few short years.

        This truly is a Metro problem. Metro just can’t run buses in the DSTT to the reliability levels that everyone expects from Link. Metro needs to get out of the way.

        Metro just doesn’t seem to get what real high capacity, high frequency, reliable transportation is..

      4. Adding one car to each train set and leaving peak headways at 7.5 minutes is the smartest idea yet. It solves the capacity issue, it doesn’t require any more operators, and you don’t have to kick any buses out of the tunnel. Then we should reduce headways once the tunnel is exclusively for Link.

      5. If ST takes more slots in the tunnel, their share of tunnel costs changes. So I’m not sure this is a case of ST pushing costs onto Metro.

      6. Since the same taxpayers pay for the tunnel debt payments regardless of whether the check is cut by Metro or ST, I’m not terribly concerned about that. It is the increasing cost to taxpayers, when adding both agencies’ costs together, that concerns me, not to mention the cost to riders, in lost time.

        But my higher concern is that some are egging ST on to behave badly, and that cannot end well.

  12. Wish I hadn’t been away from my computer all day. For a lot of years of work on subject under discussion, I’ve got to say something.

    Not one single piece of recognition here that for its quarter century of existence, the Downtown Seattle Transit Tunnel has never for one day been run with the discipline, control, and training that were presumed in its design!

    Given both of their attitudes to the importance and possibility of any better performance than to date- they’re both doubtless right given both of their present management of the facility.

    But given the impossibility of ridding the project of either or both of them- before throwing out three decades of work and hundreds of millions of dollars worth or capital, at least let’s somebody approach Tunnel operations as if serious improvement is possible.

    Another force is needed here. ATU Local 587 would be a several thousands of “somebodies”, many of whom vote in King County, including some who think not just that improvement should be achieve, but can. Anybody know the level, direction, and energy of any action from organized labor here?

    Mark Dublin

    1. Mark, I’m now curious to know what the original plan was to manage the DSTT vs. how Metro has deviated from that plan. Have you written anything in detail about that before?

  13. I think Brent’s analysis is solid given the inputs from the 2014 SIP, but as I see it the SIP is entirely unrealistic. I use Link in the afternoon peak every day, and it is … uncomfortable. I can’t imagine that only 25% more capacity by running 2-car trains more often will be anything like adequate once U-Link is online.

    Perhaps it’s just a placeholder. But I wouldn’t want to lock in low frequency due to an agreement with Metro if it turns out to be totally inadequate. But I think I could go with 3 cars every 7.5, as that will be the long term frequency in the RV anyway.

    1. Martin’s concerns on capacity are spot on. Look at load factors (passenger-to-seat ratio) on page 106 of the SIP. In 2018, predicted actual load factor exceeds design load factor. Uh oh.

      Actions that slow down tunnel movement and increase the time for each train to perform the loop increases the number of LRVs needed to maintain the same capacity. Shortening trains and increasing frequency works against the ability to increase capacity, given the limited fleet. Running four-car trains at peak-of-peak, at decreasing headway as the peak loads grow, is the real way to manage capacity. If it becomes unmanageable, even with carefully right-sizing the trains throughout peak, then kicking all buses out of the tunnel will become justified, as that would shorten the loop time by double-digit minutes.

      Martin is right that ST shouldn’t lock itself into a situation where Link can’t keep up with ridership. Whoever said “The opening of University Link in 2016 requires more frequent light rail service in the Downtown Seattle Transit Tunnel to meet expected demand. The current volume of buses and trains in the tunnel during peak periods is not optimal and service delays are frequent. Clearly, some reduction in tunnel bus volume will be needed to accommodate expanded Link service, requiring an operations plan for buses that are moved to surface streets.” on page 108 of the SIP demonstrates a rather poor grasp of capacity management using a limited fleet.

      1. “given the limited fleet”

        Right now Sound Transit has a lot of extra trains because they bought train cars with future expansion in mind.

        I don’t think the quote on page 108 demonstrates a poor grasp at all. In fact, they’re not even directly addressing “capacity”. If you read closely, what they’re saying is that increased frequency is needed to meet demand. Demand is not the same as capacity. They’re taking a stance regarding how to meet this demand (more frequency) but it’s an easily defensible stance given that increase frequency reduces wait times which in turn increases ridership.

      2. Okay, Alex, I overstated my conclusion. But saying the increased demand *requires* more frequency is inaccurate, given the option of lengthening trains, unless they think four car trains running more frequently than 7.5 minutes are needed to meet demand. ST needs to look at the options and defend their choice of option.

  14. There should be a way for one-time users – someone coming from the airport, for example – to get a point to point ticket from the vending machine that included both Link and Metro transfer. Present system very confusing for visitors and one reason for present Westlake clogup.

      1. The concept is great, but cost a bit steep, especially if there is not also one available for children, seniors, and students, like Portland has. Works well for coming from the airport, but not so convenient when not near an Orca vending machine. What is the outlook for the all-day fare program becoming permanent after the pilot period over in September?

  15. Why not just get bus riders to buy their fare on the platform instead of on the bus? Put machines in there next to the bus stops, and require customers to use them during the evening peak. That could definitely speed up boarding.

    1. ST and Metro aren’t going to install bus ticket machines in the tunnel for just three years of usage.

  16. Are there any logical through routes that could be combined? It seems kind of silly to both north and south going through separately if they could be combined, except that going one direction they empty out and become faster, and going the other way they fill up and become slower.

    1. I’m not totally sure of this, but I bet most of the tunnel routes are about at their length/reliability limits already. In a sense, that’s why we’re spending so much on Link, building traffic separation and often grade separation so the important destinations can be served with enough reliability for through-routes and downtown transfers to work. There’s also something of a problem finding suitable partners. The 101 might be short enough to through-route, but all the routes heading north are more frequent.

      To be sure, there are a few really long through routes on the surface, like the 5 and 21. I don’t think either side of that through route suffers from reliability problems the way the 7[123] does, though.

      1. The 5/21 seems like a good Rapid-Ride candidate. Rode the 21 to Seattle recently when I had just missed the C line at Westwood Village and didn t want to wait at this creepy place full of shady, drunk people even in the middle of the day. It took less time to get to Seattle than the red bus.

      2. What about eliminating the 7[123] and combining it with a bunch of the southern routes, but only for the UW to tunnel segment? The tails (Jackson Park for the 73, etc.) become separate routes. That is what is going to have to happen when Link to the UW opens anyway, isn’t it? That way the 7[123] become a more manageable piece from UW to the downtown tunnel, which could be combined with some of the other routes going south.

        This eliminates the convenience of a one seat ride on the north end, but the through routing from the UW to downtown might add convenience for other riders.

Comments are closed.