Several months ago I obtained travel time data for the Downtown Seattle Transit Tunnel (DSTT) which Metro used for performance analysis of the tunnel. With the elimination of the ride free area (RFA), a key component of the bi-partisan compromise to pass the $20 dollar CRC, this information is now much more relevant.

The data includes DSTT travel times during the AM and PM peak for both northbound and southbound travel which was generated using archived data from Metro’s vehicle tracking system. Three data points, fall 2008, fall 2009 and spring 2010 were included. The data, which was provided to me in PDF format, was post processed to create the graphics in the slideshare above (download ppt here).

The first four slides show the percent of peak trips that took equal to or less than that amount of time to travel through the tunnel. So for example in fall 2008 71% of AM southbound trips took less than or equal to 8 minutes, dropping to 56.1% in the fall 2009 and rebounding very slightly to 56.5% in the spring of 2010.

The next four slides compare the data points (fall 2008 to fall 2009 and fall 2009 to spring 2010), which gives you the percent change in trips which took equal to or less than that amount of time to travel through the tunnel. From the pervious example this means 14.9% few trips were completed in less than or equal to 8 minutes between fall 2008 and fall 2009, and .4% more trips were completed in less than or equal to 8 minutes between fall 2009 and spring 2010.

Below are some observations after looking at the data:

  • AM service in both directions degraded with the introduction of Link but has not seen additional degradation since then.
  • PM service in both directions degraded with the introduction of Link, with this trend continuing in a significant way between the fall 2009 and spring 2010 data points. The spring 2010 service change eliminated the route 194 (4 buses per hour southbound and northbound) and added the 76, 77, 216, 218 and 316 (roughly 10 buses per hour northbound and  7 buses per hour southbound in the PM peak) to the DSTT, increasing the number of buses in the DSTT by 6 buses per hour northbound and 3 buses per hour southbound.
  • PM southbound travel is the slowest with only 57% of trips being completed in equal to or less than 9 minutes. Holding all else equal it appears that the addition of Link had roughly the same impact on travel times as the spring 2010 service changes.
  • PM northbound travel has seen the most significant degradation of  service, partly because it initially had very good performance and partly because of the addition of the out of service security sweep of Link at Westlake. Travel times are now more in line with those of PM southbound travel.

All this data jives with my personal experiences and the internal report Bruce reported on last week. The introduction of Link certainly increased travel times for buses in the DSTT, but the addition of buses in spring 2010 also had a significant impact on PM travel times.

49 Replies to “DSTT Travel Time Data”

  1. Here are a couple of additional thoughts on this.

    First, it is significant that the PM slowdown is more significant than the AM slowdown. It seems to show that it takes people longer to board buses than to leave. Intuitively that makes sense – you have to find where on the platform your bus stops, some buses stop twice, some people ask the driver questions or aren’t organized, drivers wait for a runner – boarding simply takes longer than leaving the bus. In the morning, most people in the tunnel are leaving the bus. The tunnel had adequate capacity to absorb extra bsues. But in the afternoon most buses are loading in the tunnel, so the capacity is lower than in the morning, and adding a few more buses exacerbated the problem. In fact this clearly demonstrates the clearest limitation of BRT vs. light rail. While a busway can theoretically accomodate 600 buses/hour, only if there are no stops. With stops, the effective capacity decreases significantly and DSTT may be there.

    Eliminating the RFA is likely to make this problem much, much worse if they aren’t able to implement off-board fare payment or dramatically increase the ORCA adoption – though I’m afraid that even increasing ORCA adoption won’t be enough unless they can add a reader at the rear door and permit rear door boarding. Either fare payment needs to be POP, or the tunnel routes should continue to be pay as you leave, and maybe POP within the tunnel – otherwise requiriment payment on boarding in the tunnel during rush hour is likely to plus the DSTT. It’s already bad when a 70’s bus loads at Westlake after 7pm.

    Final thought – let’s eliminate the security sweep at Westlake, at least during the PM peak. Virtually no subway system in the world does this, and anyone dumb enough to stay in the car will be back at Westlake in under 10 minutes.

    1. Very good points Carl. Thanks. The data certainly does show that boarding is slower even when people aren’t paying. As of now I’m very skeptical that pay as you enter at one door will be acceptable in the tunnel. Operations are already too slow in my opinion.

      1. It won’t be acceptable, which is why no other transit system in the world (that I’m aware of) does this.

        Use all doors, pay off board. We need to keep repeating this until Metro gets it.

    2. I’ve said it before and I’ll say it again. Buses are the only form of mass transportation where people pay the driver. Streetcars, light rail, subways, trains, and airplanes all require people to buy a ticket beforehand. They either use proof-of-payment onboard, or you show a ticket just to get into the terminal. We need to move buses into the same category if we want them to be reliable and easy to use. Drivers should concentrate on driving, not on fare enforcement, end of story.

      1. Drivers don’t have much, if any, power to enforce fares. I could walk on to a bus and not pay a fare. Or easier, walk right off a pay-as-you-leave bus. What are they gonna do, call the cops? Except RapidRide or Link of course where the chance of being caught is much, much higher.

        And your statement is not true. On Muni Metro if you pay cash outside the subway, then you pay into the farebox and the driver gives you a transfer. Similarly to MBTA’s Green Line, where it’s fare gates in the subway and pay as you board elsewhere.

      2. A more accurate statement would be to say that Metro is (nearly) unique in requiring on-board payment inside a station. Lots of cities have on-board payment for streetcars and other surface vehicles, but they universally shift to off-board payment (or proof-of-purchase) once the vehicle enters the elevated/underground section.

        It’s bad enough that we have on-board payment in the DSTT after 7pm, but doing so all day would simply be ridiculous. Off-board payment needs to be a requirement, and the DSTT platforms need to become a fare-paid zone. That’s the only way that eliminating the RFA can become a net financial win, rather than loss.

  2. It will be interesting to see what the travel times are going to be when all bus trips have to return to the street in 2016 when light rail heads to Capitol Hill and Montlake.

    1. That’s not been decided yet. The tunnel will still have capacity for both buses and trains, even after U-link starts. There just extending the line a couple of stops and maybe upping the frequency a little.

    2. The real point at which buses will probably have to leave the tunnel is around 2020 when East Link starts.

      1. Northgate in 2020… East Link is a little further off. But yes, U Link shouldn’t cause the buses to leave the downtown tunnel. It may be possible to run Link at current service levels, just with 4 car trains, and handle Capitol Hill/Husky Loads.

  3. I think the security sweep is more of a carry over of the ‘lost and found sweep’ bus drivers are required to do at the end of a run.
    It’s amazing the junk people leave on a bus.
    Seems like the tunnel is pretty damn reliable from the data. How’s that compare to surface trips between IDS and CPS uring the same times/directions?

    1. What a Hoot.
      I tried ST Trip Planner fro IDS to Westlake tunnel stations.
      It gave me a choice of 13 minutes on the 132 or 15 minutes on the 545 going down 4th Ave.
      Are they boycotting their own trains?

  4. Good post, Adam. Now I wonder if you or anyone else can give us this statistic:
    What is the cost of one minute of service in the Downtown Seattle Transit Tunnel?

    I think that if the transit agencies had to account for the dollar cost of operating delays in the DSTT, service would rapidly start to get faster.

    Mark Dublin

      1. I believe scheduled travel time is usually 8-9 minutes, with NB pm scheduled for 10 minutes. That is what I have been told. Haven’t ever actually timed it myself.

  5. What’s wrong with pay as you leave for all buses after they leave the Downtown Transit Tunnel? No cost, no bother. It would help too to order buses with an extra down and phase these in on all tunnel routes.

    1. Two reasons why not.

      First, pay as you leave sucks. Everyone hates it. It’s confusing for old and new riders alike. It’s easy for people to leave without paying. You can only use one door for exiting. And for buses that have a big destination, now you have a 5 minute pause in the middle of the route while everyone pays, and that’s just annoying.

      Second, there’s a much better and equally cheap alternative, which is off-board payment in the tunnel (probably excluding CPS). You enter the DSTT, you buy a ticket using a TVM or ORCA, and you board your bus or train. The platform level of the DSTT becomes a fare-paid zone, so inspectors can walk around the platforms and ask to see people’s proof of payment. This is how grade-separated transit systems work the world over; I see no need for us to do anything different.

      1. I agree that payment before entering the platform is the ideal. Add turnstiles and you don’t even need much fare enforcement. However, that’s expensive and Metro doesn’t have any spare change right now. We already have pay as you leave signs on all the buses. I’ve been on a packed 550 and it’s still less than 2 min at Mercer Island and S. Bellevue. I’m pretty sure you could unload an entire artic in 5 min. with pay as you leave. That’s an average of about 4 seconds per person and an Orca tap takes less than a second. I’m not saying it’s the ultimate answer just that it’s the path of least resistance and solves the immediate issue of time delay in the tunnel.

      2. Given that the $20 car tab’s approval was largely a result of agreeing to scrap the RFA, I wouldn’t call that the “path of least resistance”… it has the potential to seriously damage both Metro’s and the council’s credibility.

        I agree that turnstiles are expensive, but proof of payment shouldn’t be so bad. All you need to do is hire a few more rent-a-cops. (And honestly, some of the tunnel stations could use them anyway.)

      3. They already have plenty of rent-a-cops standing around looking bored in the tunnel stations. At least this would give them something to do.

        The real problem with going to POP in the tunnel is all the different fare levels. If you’re riding Link, the fare is somewhere between $2.00 and $2.75 depending on distance. If you’re riding the 550, it’s $2.50, and if you’re riding a Metro bus it could theoretically be $2.25, $2.50, or $3.00. If they enforce fares on the platform and not on the vehicle, you have the opportunity for systematic underpayment by purchasing a fare for a service other than the one you’re going to use.

      4. To implement POP within the tunnel, you’d need to deal with the interagency fare policies. For example, if someone takes the 255 into the tunnel, then transfers to the 550, the transfer from the 255 isn’t supposed to be valid, since the 550 is an ST route. But, within the tunnel, the transfer would have to be a valid POP, or else everyone coming off the bus who paid cash would be subject to a ticket!

        Therefore, if you’re going to have POP inside the tunnel, it seems unavoidable that the 550 and Link will have to accept Metro transfers. I sincerely hope a stupid issue like this doesn’t become the deal breaker to prevent POP in the tunnel from happening.

      5. Turnstiles don’t solve the problem of riders waiting to the last second to decide which bus (or train) they will take. Nor will bus tickets from ORCA VMs deal with that problem.

        There’s nothing that will work better (or cheaper) than a collection of well-branded ORCA readers at each bay, starting to incentivize the use of e-purse, putting ORCA readers in the hands of the tunnel security force, having lots of signs and announcements that POP is required on board all buses in the tunnel, and that the $3 “tickets” Metro is giving out are not POP. (This is one more reason I think the ticket giveaway is a disastrous idea.)

      6. To implement POP within the tunnel, you’d need to deal with the interagency fare policies.

        My reaction to this is a big “whatever”. Back before ORCA, if you had a Metro transfer, you could use it to board ST (modulo the fare difference) and the agencies would just deal. When you get in the tunnel, you buy a tunnel ticket, and that’s that. ST and Metro can split the revenue based on their usual cost-sharing agreements.

        But, within the tunnel, the transfer would have to be a valid POP, or else everyone coming off the bus who paid cash would be subject to a ticket!

        That is an interesting point. You’re right: tunnel buses would probably have to hand out transfers to everyone who didn’t use ORCA, and that transfer would have to be valid POP. But I don’t think that’s a huge problem. And anyway, it’s not like fare inspectors are going to jump on people who immediately exit the bus; they’re going to wait till the platform is quiet and then audit the people who are waiting.

        There’s nothing that will work better (or cheaper) than a collection of well-branded ORCA readers at each bay

        The political will to end the RFA is currently much stronger than the political will to incentivize ORCA. As much as I want both, I don’t want to hold up the former on the latter’s account.

    2. Pay-as-you-leave costs around 22 service hours cumulative on weekday Metro routes, per the study several posts back. That’s almost half what Metro predicts converting from the RFA to pay-as-you-enter (47 daily service hours) would cost Metro bus routes downtown.

      1. If downtown the peak boarding is in a few stations why does pay as you leave (dispersed alighting) cost service hours?

      2. Pay-as-you-leave involves people slowly shoving to the front, kinda like on an airplane, except slower, since only some of the people are getting off at any given stop.

  6. Great post and yes, it has been frustrating leaving or entering Seattle on buses that use the DSTT. My personal view is that there are way too many holdups for no good reason. I don’t see why buses have to wait for a train in front to not just leave the platform ahead but arrive at the next station before the bus is cleared to advance.

    Also, the barriers that have to lower before buses can leave the tunnel southbound and enter the I-90 are pointless in my book and seem to really slow down buses leaving the DSTT. Why do they do this? There are no trains in the way on the I-90!

    I think that bus drivers should have more discretion to enter a DSTT station once a departing train has left the station ahead of them. I don’t think they should have to wait until a train has reached Pioneer Square for example before they clear a bus stopped before University Street to enter University Street.

    It used to be really nice to take the DSTT and it still is outside of commuter hours but it is not a pleasant experience during commuter hours.

    1. Yeah I remember riding through the tunnel back with the dual mode Bredas were still in use. The tunnel used to be fast and reliable. Now whenever I ride though the tunnel it’s an aggravating process.

    2. “I don’t see why buses have to wait for a train in front to not just leave the platform ahead but arrive at the next station before the bus is cleared to advance.”

      They don’t. They only have to wait for the train to enter the next signal block, which, when leaving a station, is the tunnel between stations.

      “Also, the barriers that have to lower before buses can leave the tunnel southbound and enter the I-90 are pointless in my book and seem to really slow down buses leaving the DSTT. Why do they do this?”

      Those yellow barriers are to prevent unauthorized vehicles from entering the tunnel, they have nothing to do with the trains. Thank Homeland Security.

      1. I am referring to the pointless barriers leaving the tunnel, not the ones entering it. They don’t need to be there, surely. I have been on buses that crawl through the narrow space of the downed barrier and proceed on the HOV lanes at around 30 miles an hour from a standing start uphill southbound.

        I just don’t think we need the barriers when leaving the tunnel, and yes, the trains are certainly not responsible for them being there!

      2. Without some kind of exit barrier, someone could drive the wrong way through the exit and bypass the entrance barrier. Therefore negating the point of installing barriers in the first place.

      3. @Zed: The logistical aspects of suicide bombing are dead easy. Our protection, except at targets with exceptionally high profiles, can only be that the psychological aspects of suicide bombing are a lot harder.

      4. I don’t get it. There’s nothing like this for the Great Northern tunnel. That tunnel must be weaker and a better target?

        The bus tunnel barriers have to be about stopping dummies, not stopping terrorists.

      5. The problem isn’t the presence of the exit barrier, it’s the fact that it goes up after every bus even if there’s another bus immediately behind it. It should go up only after the queue of buses has passed.

  7. I would like to see a data point just outside the DSTT included in this report to see what delays, if any, are occuring before busses arrive inbound at CPS or IDS.

    Once we see any delay northbound into IDS (one example), in the afternoon peak, a backup of both busses and trains rapidly snowballs into a mess that is near impossible to clear until after rush hour even when traffic northbound starts flowing again.

    Fixing that part of the ballet is a key, I think, to getting maximum throughput in the DSTT now and after ending the RFA.

    1. Yeah this is a issues I have heard about as well, which I have never seen because I don’t often don’t take the bus past Westlake on the north end of the DSTT. The conflict between buses and Link at the Pine St stub was fairly well documented in the internal report Bruce wrote about last week. Don’t think it addressed this issue.

      1. That’s a serious admission for a transit professional of your caliber, Adam!

        Might be a good idea to spend a Friday pm rush hour just riding back and forth between CPS and Westlake, and another one doing the same between IDS and Stadium.

        Also, since you’ve got street cred as a respected analyst and journalist, you might be able to stand alongside supervisors at both portals through rush hour.

        I guarantee a lot of good ideas will occur to you.

        Mark Dublin

      2. @Mark. I actually wanted to do this before I posted this but working in Kirkland makes it really hard for me to get downtown in time to watch the peak start to pick up. I was able to get catch the tail end of the rush hour a few weeks ago and I just hung out on the SB Westlake platform from about 30 minutes and watched operations there and up into the Pine St stub.

  8. Sometime in the early ’90’s, I remember standing on the northbound platform at University Street and seeing a beat-up white van come out of the tube from the south. As it passed, I got a good look at the two occupants, and wasn’t encouraged: they looked ragged and tough.

    I called Control on the emergency phone. I was told that a supervisor intercepted the van at Westlake, and simply directed the driver up the cut-and-cover and out at Convention Place.

    I also recall a tour group I was with during Tunnel construction a couple years earlier. A full-sized trailer for an eighteen-wheeler was parked at the platform.
    I pointed out to a Seattle city councilman and Metro Council member just what size load of explosives could arrive at Westlake with nothing to stop it.

    He told me I worried too much.

    Problem with not gating outbound lanes is that anything everybody can drive through, somebody will.

    However, it’s a legitimate question why buses have to stop and wait for barriers to drop, while the barrier is already down so trains can roll through at 30. Same question holds for the multiplicity of stops buses have to make that trains don’t, including stopping for running passengers after leaving platforms.

    The question in everybody’s mind should be: why is fast, smooth service reserved for trains and forbidden to buses?

    Mark Dublin

    1. I spend time at Westlake Station southbound around 1700-1730 2 or 3 days a week. Runners should never be accorded ANY courtesy as they are a menace to timekeeping.

  9. Adam: when routes 76, 77, 316, 216, 218 became DSTT routes, routes 194 AND 174 left; the increase in trips per hour was less than you state.

    1. From what I have been able to find the northern portion of the 174 (north of TIB station was replaced with the 124 in September 2009 which now runs on 3rd Ave. This change would be reflected in the Fall 2009 data point not the Spring 2010 data point.

      http://metro.kingcounty.gov/up/sc/rideralert/ra-092009.html#new

      The 76,77,316, 216 and 218 moved to the tunnel in the in the Spring 2010 which are reflected in the Spring 2010 data point.

      http://metro.kingcounty.gov/up/sc/rideralert/ra-022010.html#tunnel

      This means the change in buses per hour for the Spring 2010 service change in the post is accurate.

Comments are closed.