poster on board train of late-night bus service serving BART stations
Link in, Bus out?

Today is the last day to fill in Metro’s Late Night Transit Service Survey if you haven’t already. It takes only a few minutes. Notably absent from the list of routes in the survey, which includes the streetcars and other ST buses, is Link light rail, despite a public petition calling for expanded service hours.

Sound Transit says it needs the few hours when Link is closed for system maintenance. Reducing maintenance hours increases costs not just for operating later service but also for the maintenance itself. The challenge of dedicating enough time for essential maintenance is why many cities around the world shut down their subways overnight. It is not unique to Seattle.

What other cities do instead is run night buses that emulate rail service. Philadelphia, the Bay Area, and Toronto take this approach and it works well from my limited experience. After a night out on South Street in Philly I had to get back to my hotel which was on the other side of town off the Market Frankford Line, which closes between midnight and 1 am. As soon as the last train left the station, buses start running, serving all stations along the line every 15 minutes until the first train of the next day. The buses were well used enough that SEPTA began running all-night train service on weekends (once again). For people who are concerned for their safety of walking at night, the bus also stops in between stations.

Map of Toronto's Blue Night Network
Map of Toronto’s Blue Night Network

In Toronto, I had to catch an early morning flight at Pearson and was staying near Ossington Station on the Bloor-Danforth subway. Not a problem. Thanks to Toronto’s comprehensive Blue Night network I was able to complete the early morning trip to the airport on a one-seat ride. The 300 Bloor-Danforth night bus extends to the airport, eliminating a transfer that is required during the day.

Seattle already has the basis to do the same. It is called Route 97, the Link shuttle. Whenever Link service is expected to be disrupted for an extended period, a bus bridge is created to fill the gap or completely replace service. Route 97 operated for an entire day when Link was closed for system upgrades. An emulator bus like the 97 is easier to understand for most people than a combination of existing routes like the 7, 36, and 49 because of consistency with regular service. It serves the same stations and can accept the same fare media. A Link night bus could extend from Husky Stadium to the U District, filling the gap handled by other buses during the day. Imagine how liberating having Link, our busiest transit line, available 24/7 would be.

map of Route 97 Link Shuttle (pre U-Link)
Rt 97 map (2011 Metro route book)

The Link night bus would be slower than Link light rail because it can’t use the fast running ways that the train has. I estimate an all-stations trip from UW to the airport to take about 74 minutes, including 30-60 seconds dwell time at each station. That’s 30 minutes slower than Link but with better access than current night bus service. By my really ballpark estimates, if the service ran every 15 minutes between Angle Lake and U District stations from last train to first train, at least thirteen buses would be needed to operate the service and it would require at least 19,000 service hours annually. At $142 per vehicle hour, it would cost at least $2.7 million or about seven percent of what ST has budgeted to pay Metro for Link operations in 2016. Service every 15 minutes all night may be overkill but it would be needed when everyone is going home after the bars close; less frequent service could be run at other times.

Keeping our transit system in a state of good repair is necessary. Meanwhile, people want to be able to use transit around the clock for work and play, a sign of a increasingly vibrant city. Running a Link night bus is a promising solution that could balance both needs. Sound Transit is studying the feasibility of extending Link’s service hours. Whether it is a bus or a train or a combination of both, the goal should be to expand late-night transit service.

94 Replies to “Run a Link Bus Overnight”

  1. I 100% agree with you on this. For all those people out there clamoring for 24/7 link service just look at DC metro. They had to shut down for an entire day a few weeks ago and the decades of deferred maintenance will cost the city several billion dollars. I for one don’t want ST4 to be a multibillion dollar emergency maintenance package. Having a link shadow bus service (maybe every 30 minutes would better fit the demand) would fill the void without compromising the longevity of our system.

  2. I’d like to see late night Sounder alternative service as well. Last train at 6:20 is ridiculous.

    1. If you’re in the South Sound, the 578 handles this – it runs from downtown Seattle to Auburn, Sumner, and Puyallup stations. And you can catch it at 2nd and Jackson, about 50 feet from King Street station.

      I don’t know why its first stop is Federal Way rather than Tukwila and/or Kent, though.

      1. That’s the 150. It has its last run some time around 1:30, making it one of the latest-running suburban buses.

      2. At night the 150 takes 45 minutes, which is almost respectable compared to the hour it takes in the daytime. Northwest Kent has 24-hour employers which are on the 150. East Kent has the highest density in south King County, and although it’s off the 150 and their buses end earlier, it’s a short drive from Kent Station where the 150 is.

        The 578 doesn’t stop in Kent because Federal Way is on I-5 so that makes it more important. Or at least that seems to be the thinking.

      3. Mike,

        Isn’t it possible that ST is skipping Kent because Metro already goes that way? Going via Federal Way makes possible airport and West Kent travel to the 578’s destinations possible via the A-Line.

  3. I think it will be a good idea in the future if development starts to be more organized around link stations…

    One problem I foresee is that link light rail will be fairly dependent in some places on feeder buses, and those buses won’t run at night anyway. Even now, ST buses like the 512 will run to transit centers late at night when the feeder buses people need to get home aren’t running.

    Another point is that service every 15 minutes at 3am is overkill for the amount of traffic at that time. Every 30 minutes after midnight, and every hour after 2am would seem more reasonable to me.

    1. If there’s enough demand, then the feeder buses would also run overnight. SF and Toronto do this with timed transfers at key points.

      15 minutes is just an example for costing. Less frequent service just means it’ll cost less depending on how it’s scheduled.

      1. Even if the feeder buses don’t run all night, there’s still the TNC’s. If late transit service reduces the taxi bill from a $40 trip to a $5 trip, that’s still a big savings.

      2. Actually, the TNC coverage has improved significantly the past couple of years. When in doubt, I would suggest opening up the Uber app at 2 AM from the comfort of your home, and scroll the map down to Federal Way to see for yourself (without actually requesting a ride).

    2. 15 minute headways would be most valuable on Friday/Saturday nights. The rest of the week, I think 30 minutes would be sufficient. That would make the cost even less – I know 15 is just an example to baseline from.

      I’d suggest:

      Sunday-Thursday: 15 minute headways for ~1hr after last train (4 trips), then 30 minutes
      Friday-Saturday: 15 minute headways until ~3am, then 30 minutes

    3. If you’re only going to run the bus once an hour, you might as well not run it at all. Can you imagine arriving at the bus stop only to see that bus pulling away and knowing you have to wait 59 minutes for the next bus.

      That is terrible service.

      1. Happened to me once, at Montlake, when the 545 swung by just a bit ahead of my delayed 48. I guess I could’ve called a cab, but I was glad that yes, there would be another bus coming that evening.

      2. It’s different at night. If the choice is waiting an hour vs waiting five hours or walking for an hour or two, then an hourly bus is better than nothing. Especially if it’s dark and you’re worried about safety, and there are few venues open to wait in.

      3. Hourly service is lifeline service for people without other choices. When service frequency is that bad, people will plan their trips more carefully. Or course, if things happen and they miss their bus, it really sucks.

      4. In the early 1990s a couple of TriMet routes used to be every 120 minutes on Sundays. The demand for better service was there but TriMet didn’t have the resources for anything more than basic survival frequency. Today one of those is a 15 minute route.

        Terrible service is still service, and over time it can lead to something better.

        No service at all doesn’t have any ridership and thus can never prove its popularity.

  4. I totally forgot about the 97! I’ve used night buses in Vienna several years ago, but it was much more confusing because there are several lines, and since it’s mostly underground, it wasn’t intuitive for me where to go to catch them. We would not have those issues in Seattle.

    1. They do track and switch inspections. That takes time. I’m sure sometimes those inspections sometimes uncover problems that need to be fixed, so there is contingency time as well.

      As the system ages there will be more intensive maintenance needs. If we added 24/7 service now, there will be political pressure to maintain it 10-20-30 years in the future when proper maintenance standards would require more downtime.

      NYC has “24/7” service but it is hardly 24/7 on many lines with all of the major construction work going on. Nights and weekends are a constantly changing hodgepodge of closures, reroutes, and service exceptions.

      1. London seems to have quite a robust system for closures and re-routes to cover maintenance. It seemed like every night I rode they had a different re-route in place, but the explanatory signage for the detours was quite good explaining what alternate route to take or what station you’d have to exit closest to the station you wanted to be at. I imagine riders there are used to the system.

        However, those kinds of re-routes are a lot easier when you have an actual grid of subway lines like London, Paris, Seoul, or New York. With one spine there isn’t anywhere to re-route riders to other than buses.

  5. A year and a half ago, I wrote an article with an idea for a late-night bus network. Many commenters, including the dearly-departed d.p., helped flesh it out and make it a real proposal. Just want to reiterate it because I still think it’s a good idea (including running 97 all night). It’s changed a bit after the Bus2Link restructure but the core is still valid, I think.

    Either way, we need a true all-night network because there are a lot of us who want to ride transit but can’t–at least not as often as we want–because service stops too early to and from too many places.

  6. A nighttime 97 is a great idea. However, I would suggest that it skip SODO & Beacon Hill. This would save the bus from having to zigzag through Beacon Hill to Mt Baker. Riders who use SODO and Beacon Hill (and most other stations in Seattle) are already familiar with bus service that is already in place (i.e. 7, 36, 124). After leaving downtown, the bus can head east on I-90 and access Mt Baker via the Rainier exit.

    Another alternative is to run an express 97 from UW, CH and downtown to SeaTac without stopping in the Rainier Valley. ST already has early Link trains running through the valley starting at 4:22a. Having an early 97 that runs through Rainier Valley would be redundant, especially if most riders are simply heading to the airport. Riders who need to access the RV from downtown, again, can do so with bus service already in place.

    1. The utility of Link is that it serves all these places on a single line. There are merits to having variants but my main argument against is simplicity. You can’t assume people will know alternatives if all they take is Link. Splitting it up will confuse people and make travel between the stations more difficult. It also splits resources between lines which increases costs.

      And people in RV need to go to the airport, too. The buses would not run when trains are running except for the very last trips which may overlap with the first trains.

      1. ‘You can’t assume people will know alternatives if all they take is Link”

        Having a modified version of the 97 is no different from introducing a new route elsewhere. Most people plan ahead and simply adapt to the new change, for better or for worse.

        “And people in RV need to go to the airport, too”

        I propose the 97 serve all stations except SODO & Beacon. The Rainier Valley will still be served. In the 4 o’clock hour, Link already serves the RV starting at 4:22. Having the 97 serve the RV is redundant and slow, especially when most riders will use it only to head to SeaTac.

        “You can’t assume people will know alternatives”

        But let’s be realistic. Most do. Good service is based on connecting riders to where they want to go and not where they could “possibly” go. Riders who currently travel from downtown to Beacon Hill & the RV after 1a already know they can take the 7 & 36 because. well. . they already do!

      2. Once the spine is extended to Lynwood, RTC and KDM stations, would you have a nighttime bus following the Red and Blue lines and stopping at every single station in order? How about when it extends from Everett to West Seattle and Ballard to Tacoman and Redmond to wherever?

      3. “introducing a new route”

        I’m fine with new night routes that shadow parts of Link as long as the system is easy to understand.

        “Link already serves the RV starting at 4:22”

        But not before then. I’m not proposing redundant bus service.

        “most riders will use it only to head to SeaTac”

        We do know the ridership on Link in the RV is not insignificant which is the whole point of having service that maintains those connections overnight and that doesn’t require backtracking downtown.

        “connecting riders to where they want to go”

        Link serves all those locations on one line. It’s pretty clear people want to go to those places. Now maybe a completely different grid of routes make more sense than a strict Link shadow, I’m fine with that (see first response)

        “Riders who currently travel…”

        I’m not talking only about current riders. I’m talking about the many more potential new riders who don’t travel because service is inadequate or confusing.

        “once the spine is extended”

        they should develop a network that makes sense. If you look at the BART example, they have timed transfers in downtown Oakland between the SF-Richmond line to the Fremont line. That’s one way of doing it.

      4. @aw

        I highly doubt a Link bus would go beyond Northgate. Maaaybe 145th. Unless demand for overnight service in North King/South Snoho counties jumps in ten years, I don’t see a need to serve that area with an overnight Link replacement. South county already has (and needs) overnight bus service, thus limiting the Link bus to SeaTac and no further. At least, one would think.. . .

      5. A full 97 owl would overlap the 49 and run between the 7 and 36. If it stops only at Link stations it will be a limited-stop express. Is that appropriate for a night owl bus? At night the 7 can get from Rainier Beach to 5th & Jackson in 30 minutes, and serving all the local stops since most of them are empty. Would people at Graham Street be happy that the 97 owl goes past their home without stopping?

        The further Link is extended, the less it will make sense for a single owl route to shadow it the whole way. But I wouldn’t make a hard assumption where it should end. People in Lake City and Northgate have been without night owls forever but that doesn’t mean they shouldn’t have one. Maybe it could go to Northgate, maybe 145th, maybe Shoreline. It could even go to Lynnwood because somebody will want to go there, especially when the live bands in Seattle let out on Fridays and Saturdays.

      6. There is no requirement that the buses run non-stop between stations. As I said, Philly’s rail owl buses stop between stations.

        The idea to run a Link shadow bus is mainly my response to the people demanding ST run trains later and cut into Link’s maintenance hours. I’m all for improving night service on bus routes across the city.

    2. The 97 serves all stations because it’s an emergency replacement and there’s no time to educate riders about a complex alternative, especially when ST/Metro’s poor communication means that riders waiting at the stations or bus stops are always the last to know about last-minute changes. The most complicated thing riders can handle at that point is the 3rd Avenue stop locations, since every station has multiple entrances but a single alternative bus stop.

      However, a planned night shuttle could have a slightly different route. Serving both SODO and Beacon Hill means you have to backtrack somewhere, either via Columbian Way or the steep, narrow, sharply-turning Holgate/Beacon road. UW Station is also out of the way in the middle of nowhere without any nightlife. So I could see a night shuttle that bypasses SODO and UW Stations, and also maybe Stadium Stations. It could run on Jackson/12th and Pine/Broadway/10th E. Another issue arises at that point though, that it’s overlapping the 36 and 49, so shouldn’t it make local stops at night and why not just use those routes in the first place? Alternatively, it could go on Eastlake and bypass Capitol Hill since the 49 serves both sides of Capitol Hill Station anyway, but then it’s not a seamless one-seat ride at one of the highest-use owl stations.

      Other cities I’ve seen — San Francisco, Chicago, London — have half-hourly night owls spaced a mile apart, sometimes serving the rail stations in a line or sometimes just the street grid. Either way is OK, and neither one is 15-minute frequency. So I don’t think we need 15 minutes here. Even just getting to a half-hourly or hourly frequency would be a major benefit.

      1. During service interruptions, the current 97 can remain the same. At night, however, a modified, faster route can be the new norm.

        ‘Another issue arises at that point though, that it’s overlapping the 36 and 49 . . ” This is one of the reasons why I suggest having a nighttime Link bus skip SODO & Beacon Hill (and Stadium too). The 7, 36 & 124 already cover these areas almost 24/7.

      2. I think the 7, 36, 49, and 124 are adequate replacements for an owl shadow, as long as the worst holes are filled: TIB to SeaTac, and Rainier Valley to SeaTac. The latter could be filled by taking the 7 or 36 the other way to the 124 at Jackson Street, but that’s where frequency gets more important because of the transfer.

      3. I was thinking along the same lines. The thru-routes of the 36/70 and the 7/49 actually fill the Link shadow role quite well, at least from Ranier Beach on northward. For the south end, just extend the late-night trips of the 124 to serve the airport, and I think that’s good enough.

        As to 15 minute service…right now, after 2 AM, we don’t even have a single route doing hourly service – even the highest ridership corridors have service only every 75-90 minutes or so. Let’s first get all-night service up to hourly on the major lines, and midnight-2 AM service up to half-hourly on the major lines, then talk about going beyond that.

      4. That still leaves the U-District to Nofthgate, but maybe Roosevelt BRT can be the savior there.

    3. just continue past sodo station, south on 6th to columbian.

      the long buses do not do well on mcclellan, however.

    4. I was wondering about maybe splitting this into a night 43 and 97 to deal with Beacon Hill and UW having rather different night demand than the areas further south.

      1. There’s no reason at all to have late-night service to Husky Stadium, with nothing to connect to. Just run the 49 from Capitol Hill to the U-district and call it good.

      2. You run the bus to each of the stations, in order, no exceptions. Anything else is not worth the extra complexity. The huge advantage for Central Link is its simplicity, and if its night bus skips certain stations, you’re gonna leave people stranded and pissed off.

      3. ^ This. My point exactly. It is practically Link 24/7 but with buses instead of trains operating overnight. It doesn’t get any simpler than that.

      4. Just run the 49 from Capitol Hill to the U-district

        43 goes to the U District too.

        The problem with the 49 is that it isn’t as close an approximation to where Link goes, while the 43 does go Capitol Hill to essentially Westlake (it’s one block off).

        You run the bus to each of the stations, in order, no exceptions. Anything else is not worth the extra complexity.

        What do you do in cases where there is no easy road alternative, such as Sodo to Beacon Hill to Mt. Baker?

      5. What do you do in cases where there is no easy road alternative, such as Sodo to Beacon Hill to Mt. Baker?

        Route 97, see my post and map at the top of the page. It has existed since Link opened and ran all day on Link’s opening weekend, to name a few cases it ran for extended durations.

      6. Creating special routes that operate only late at night is definitely extra complexity that is not worth it. That is a large reason why ridership is so poor on the existing night-owl routes 81-85. However, adding late-night trips to routes like the 7, 36, 49, and 70 is totally different. These are workhorse bus routes that already run frequent service all day, so late-night service is not a new service pattern people need to learn. Given the choice, I would rather see the late-night 49 boosted to run every 30 minutes than have a 49 running every 75 minutes, overlayed with a Link shuttle running every 60 minutes, which doesn’t even bother to stop in between the stations. South Seattle is kind of similar. You could operate a shuttle that stops at the Link stations along MLK, but adding extra trips to the 7 and 36 to achieve some kind of halfway-decent frequency is better. MLK and Ranier are close enough together that people who live on MLK can walk from the nearest route 7 stop. The only thing that’s missing is downtown to the airport, but a short extension of route 124 would do the trip. The former route 174 used to take about 45 minutes from the airport to downtown, so an extended 124 should be similar. If people don’t know to wait for the 124 after the last Link train has left for the night, that’s a problem that can easily be solved with decent signage. By contrast, a shuttle trying to stop at all of the Link stations would take considerably longer than this.

      7. The only thing missing is the Rainier Beach-Tukwila/Airport connection. Making people in Rainier Valley go downtown to get to anywhere in south King County along the A and 124 is ridiculously circuitous.

      8. Seattle’s geography and land use requires a shadow bus to do a lot of zigzagging and going through single-family areas it doesn’t stop at. An underground train bypasses all of that but a shadow bus can’t. That may be a bigger problem for Seattle than many other cities. Do we really want a bus going from Capitol Hill Station on John Street and 23rd to UW Station, then Pacific Street and Brooklyn Ave to U-District Station? Should it make local stops or just drive by uselessly on a detour? If it makes local stops we might as well resurrect the 43, but then people in the area will wonder why the 43 came back night owl but not midday when they really want it.

      9. Yes but that’s hardly bus service: it’s not two way, and it starts and ends at arbitrary hours that don’t match when people want to travel. It’s barely more than hitchhiking in cars that happen to be passing by their own reasons, not for you.

      10. Is there any reason not to use McClellan instead of jogging out of direction up 17th to College? I’m assuming it’s the hill down to Rainier but is that just for articulated buses or can the smaller ones also not handle it? It would be much more direct if they could manage that. Also a shame SoDo is out of line with the most direct bus route, requiring a loop to serve it before heading back up to Beacon Avenue. Other than that, a shadow bus all the way from Husky Stadium or the U District to the airport looks to be reasonably straightforward (I’m assuming Montlake–23rd–John–Melrose–Pike/Pine to Westlake from Husky Stadium, then following the 97’s route).

      11. Route 97, see my post and map at the top of the page. It has existed since Link opened and ran all day on Link’s opening weekend, to name a few cases it ran for extended

        What does the 97 do north of Westlake? I’m picturing something like a 43 as I don’t see what else you would do there.

      12. That map is old. I don’t have a U Link version of it but I presume it’ll either use the freeway or follow the 43 route.

    5. Mathwise, it makes zero service sense to run BOTH Line 97 and the existing KC Metro owl services. There isn’t enough demand. It’s either run existing service patterns, or cancel lines like the 7 and 36. The compromise would be to call Line 97 but have it operate between segments close to the existing 7 route which would make no one happy. In the old days owl routes in all cities were notorious for trying to serve all people all of the time, but that has been changing. The Twin Cities, incidentally, cancelled owl service on the 16 line between Minneapolis and St. Paul and runs the train hourly between the two cities, with no underlying local service (stop spacing is generally a half mile).

  7. Difficult to justify night service in a city where so much stuff closes early. I remember moving here from the Midwest when I was in my early 20s thinking, why on Earth does everything here close so stinking early. In college in Ohio, I’d go out for pizza, burritos, coffee, and hash browns at 3:30 am as a break from studying. Need fabric softener to throw in that load of laundry at 4 am? No problem-a grocery store will be open. Here… no, you can’t do that here. Keep the buses running for who? The barflies? Been in the burbs for a few years now, perhaps things in Seattle proper have changed a bit? Have they?

    1. Well, college towns tend to cater to a later night crowd. Colleges are also really dense and have a captive audience.

      “why on Earth does everything here close so stinking early”
      “Been in the burbs for a few years now, perhaps things in Seattle proper have changed a bit?”

      I think you answered your own question :) – too many people move away from the city. In college I did a lot of stuff late at night; now I have lots more money but I want to sleep instead.

      I can vouch for Boston and DC being pretty dead at night. Much of NYC is deserted at night too. It is not nearly as 24/7 as it may appear, even if the subway runs that late I’ve found that night trains are fairly empty except Friday/Saturday.

      Unless a city has lots of non-standard shift workers who also have money to spend, late night dining/retail is not going to get much traction.

    2. Most QFCs and Safeways are open 24 hours now. The exceptions are in the most inner-city ones like the 50th & Brooklyn Safeway because of, er, violent people nearby. The flip side of that is that the ones that are always open and always have open bathrooms are located in less-walkable, less-transit areas. Seattle’s zoning has a lot of those places unfortunately.

      As for other late-night businesses that aren’t bars, I once had a list of “Seattle Late-Night Eats” because they were so few and random. But there’s starting to be more of them in; e.g., Capitol Hill and I think Ballard. (I was living in Ballard when I made the list, but I haven’t been in Ballard enough recently to say.) And of course, the increasing 24-hour supermarkets have premade deli sandwiches, so that’s something to fall back on.

      1. The 24 hour market thing is market dependent. Growing up in Kansas many supermarkets were 24 hours, even out in the suburbs of Wichita. But come to California and there are few 24 hour stores – mainly Wally and drugstores near hospitals (but many of them have now shifted to closing at midnight). Even some Wallys which were 24/7 are now closing at midnight or 11 pm as they have to get their stock price up.

      2. Mike Orr and Alex’s responses contradict each other:

        “Well, college towns tend to cater to a later night crowd. Colleges are also really dense and have a captive audience.”
        “Most QFCs and Safeways are open 24 hours now. The exceptions are in the most inner-city ones like the 50th & Brooklyn Safeway because of, er, violent people nearby.”

        Uh, yeah. So, 50th & Brooklyn, that’s U District, with a later night crowd in a dense neighborhood. Basically a college town. Back when I moved here, I spent a ton of time on the Ave. I was 22 years old, that was the place to be.

        Anyhow, back to my question, why do we need a 24 hour transit system when we clearly do not live in a 24 hour city? I can see running LIMITED transit until 2:30 am to get people home from the bars, but beyond that? Nothing’s open. There would be better ridership of select suburban routes during daytime hours if we improved those. I don’t see this as a good investment of our transit dollars. Let’s create something that helps more people get to jobs or school car-free.

      3. Nothing I suggested in my post requires a night bus to run 24 hours (or every 15 minutes) but if I’m going to spend nights writing a post I’m going to use the “best case/highest cost scenario”. People can then adjust costs and service down from there.

      4. So why don’t we close the freeways and arterials at night because hardly anybody wants to go anywhere. And most of those who do are coming from bars, so let’s make them show their bar hand stamp to get through the freeway entrance; no stamp, no travel. Never mind that other people are going to or from night jobs or spent a late evening at somebody’s house or got a call from a relative that has an emergency; they’re not deserving to travel because we didn’t anticipate they would.

        I never said night service had to be as frequent or comprehensive as day service; it just needs to be coverage service that should at least run hourly. When I was in Jersey City and had an early morning flight, I left my motel around 2am and took a feeder bus (half hourly?) to the Journal Square PATH station, PATH to Newark (15 minute), and a bus to the airport (hourly). That’s reasonable owl service. Not these 1-3 hour gaps before and after the magic 2:15 and 3:30am runs, and no practical way to transfer or go crosstown.

    3. From what I’ve seen, the need for late night service has primarily been generated by social event goers, i.e. bar/club patrons. The most vibrant nightlife right now is arguably Capitol Hill in the Pike St Corridor. Second to that, Ballard. Belltown is pretty strong too, along with Pioneer Square. However, I’ve noticed that Belltown & P.S. patronage aren’t the type to use the bus. However, people from all over the city flock to these neighborhoods at night and that’s where transit is weak.

  8. They absolutely should do this, Sound Transit should run it in ST livery (not Metro), and it should be called something like Late Night Link Bus (not route 97)

    1. …and it should be called something like Late Night Link Bus (not route 97)

      Two points: 1) All of the signs for “Link is not working take this bus instead” say “board route 97 LINK SHUTTLE.” 2) The buses running on the 97 route also say “97 LINK SHUTTLE” on them.

      At least in this instance, Metro and Sound Transit are consistent.

      1. Route numbers (and service ) is nothing permanent. For advertising sake, the night time Link bus can easily be branded as something different.

    2. Oh yeah, one more thing, the late night bus should appear on the same schedule as the train schedule (with a highlight or other marking indicating it’s a bus instead)

  9. Late night transit is so important to people who rely on transit as their primary transportation. When missing the last bus means you have to take a very expensive taxi or simply walk several miles it makes it harder to rely on transit.

    Rapid Ride does a good job running about every hour overnight. It would be nice to see Link and future rail lines have the same idea.

    1. Those runs replace a couple of 33s that formerly served Union Station to Oregon City at those times. It wasn’t worthwhile to replace them with a MAX Orange Line train running part of the way and then a 33 to finish the route. You’ll notice that both runs turn into 33s south of Park Avenue MAX station.

      It would be a bit like keeping a couple of 71/72/73 late at night from Westlake to NE Seattle because a former late night slot in the timetable wasn’t worth operating with Link + 71/72/73 because one requires just a bus and bus driver while the other requires one bus and one train plus drivers for both.

      Adding completely new late night service is a different story than filling a slot in the timetable that former service had previously filled.

      1. It’s also based on where the yard is. Trimet feels there’s no sense in running an empty train as a deadhead.

      2. …which is why you see stuff like the 3:26am Blue Line train from Ruby Junction to Rose Quarter, which turns into a 4:09am train on the Orange Line to Milwaukie. If a train must be run you might as well run it with the lights turned on because otherwise the dispatching center has to be fully functional while moving 0 passengers.

        If they operated bus route 291 as a MAX train, it would mean keeping the entire system up and running for another hour or so, including dispatch and all that. As it is they only have two and a half hours between the 1 am weekday last train going to bed and the 3:26 am first train.

      3. The Link Control Center is staffed 24/7 because maintenance crews are out on the track when the trains aren’t.

      4. Hmm…, there is no direct replacement for the 71/72/73 trips that ran after the tunnel closed (on 2nd Ave), and hence after link stops operating, is there? Is that use case just replaced with a late-night route 70?

  10. I’m guessing that Metro doesn’t do this because unlike other cities, our light rail line was chosen on an arbitrary, or one could say, political basis, rather than a practical one. Look at a census map, or a map of trips or destinations and you sure wouldn’t start with a run out to Tukwila. With or without a tunnel, it simply isn’t that effective a line. It is a bit better now, as it finally connects the two most popular parts of our city, but just barely.

    So, while there is genuine value in having a system that mimics the day time service, I don’t think it should be a focus, especially when the daytime service spends a lot of its time wandering off to far off locations, or in a tunnel (some of which can be used by a bus, but some of which can’t). But, if we really want to serve people that missed that last train, I suggest the following:

    1) Run the 36 late at night. This covers Beacon Hill, one of the better performing stations (the 36 itself is one of the better performing lines).
    2) Same with the 7. This covers Rainier Valley. In some cases this means a pretty big walk, but not the end of the world (about a half mile). The 7 carries a lot of riders, so simply extending this might pick up a few.
    3) Same with the 49. The 7 pairs with the 49, so this works out nicely.
    4) Run an airport shuttle to downtown. I question the value of this. My guess is people would just pay for a private shuttle (since they are common and are way more likely to involve a minimum of waiting). This might be an interesting experiment, though. Maybe there are enough late night workers (or travelers) to justify a bus. Such a bus could easily swing by the Tukwila station, but I don’t see why it would.

    That leaves SoDo and the Stadiums, but those stations really don’t perform that well. SoDo has very few riders; Stadiums has a bit more riders, but a walk to Jackson isn’t the end of the world (around a half mile).

    1. Also, Metro and Sound Transit are different agencies so it’s easy for each to brush it off and say it’s the other agency’s responsibility, not theirs.

    2. Metro doesn’t do this because it has a limited budget and many other transit needs, especially in the suburbs where there’s no Prop 1. And you can’t blithely say “Metro doesn’t have a route from Rainier Valley to Tukwila because nobody wants to go there.” A grade-separated train can just zip between them and go on its way, while a bus has to navigate the streets and in the daytime gets bogged down by traffic.

      Skipping Stadium and UW Stations is a different issue; it gets into what’s around them and when it’s open. The sports teams and college have long closed before night owl starts, and there’s nothing else to walk to. That’s different from commercial and residential areas that have a scattering of people wanting to travel. (UWMC is working 24 hours, but my experience at Harborview suggests that shifts don’t start or end between 11:30pm and 6am.)

      1. Of course people want to get from Rainier Valley to Tukwila, just not enough to justify a light rail line, or, for that matter, late night service. Per service minute, even on an express grade separated rail line, their are parts of the line that are hard to justify. They become extremely hard to justify for late night bus service, when all of the speed advantages of the line disappear. Not only that, they get much worse. From Mount Baker, a train will beat the 7 to downtown every time during the day, but a bus hitting every Link station will arrive downtown well after the 7. The only reason to run a bus line like this is to make it a bit easier for folks who miss that lass train. But as any late night transit user knows, if you miss that last train, having to figure out where the buses go is the least of your worries.

        As you said, Metro has a limited budget — it can’t run buses everywhere. There are better alternatives than the 97. It is one thing to run it in the day time (when riders assume that a train will be there, or have become dependent on the train route) it is another to run it late at night, every night. By the way, how many people rode the 97 on the day when it replaced Link?

    1. The big issue is ADA paratransit shadow coverage for overnight buses. The transit agency still has to pay the ADA contractor. Also, MBTA was keeping overnight service was its existing staff, which meant a yard has to remain open, maintenance crews would not be able to work on certain buses, and backup coverage is still required for supervising and extra board. KC Metro already runs owl service so the marginal costs are just the driver and the bus, with some additional paratransit area possible when Link goes to Northgate.

  11. Why does the 97 deviate to S. College street from McClellan street? The bus can hit both Beacon Hill and Mount Baker stations without the deviation. That’s actually bigger than the SODO deviation.

    1. Mcclellan is too steep for 60 foot buses, and probably even for 40 foot buses. Back when the 38 ran from Mt Baker to SODO it used 30 foot buses as I recall.

  12. Toronto has the best overnight bus/streetcar service in North America, as is evident from the map in the article. Nearly 100% of the population lives within a 15-minute walk of at least one all-night transit route that maintains headways of 30 minutes or less all night. Even NYC doesn’t come close to that. In Staten Island for example, 0% of the population lives within a 15-minute walk of such a route.

    If Seattle is looking for a night bus service pattern to copy, it should copy Toronto’s.

    1. If Seattle is looking for a mass transit system to copy, it should copy Toronto’s. Or Vancouver’s (which was copied after Toronto’s).

  13. The DSTT is closed at night, too, so the night-owl 97 has to be surfaced. I don’t know if the SODO busway is closed overnight, but I’d rather wait for a bus in a more populated area, so move it out to 4th I guess…

  14. And there’s no reason to have a 30-60 second dwell time at each station. The stations should just be treated like any other bus stop. If no one wants to get on or off a particular station, there’s no reason for the bus to stop at all. That’s how Toronto’s overnight buses work – like any daytime bus route.

    That would speed things up and require fewer buses to maintain a given headway.

    1. Buses are just slower than trains at letting people on and off. The 30-60 second figure is an average. Some stops, the bus would blow by without stopping at all. Others would have dwell times of 2 minutes or more, depending on how many people are getting on and off.

    2. asdf2 is right. This is an average assigned to each stop for purposes of determining the running time. It doesn’t mean the bus would always stop at every station for 30 seconds. You can’t use the end-to-end run time straight out of Google Maps.

      I didn’t go into detail of determining the proportion of riders using ORCA cards, round-trip Link tickets, and cash.

  15. I’m arriving pretty late at SeaTac for a flight soon, so I looked up the Link schedule to make sure I could still make it. The schedule shows trains running to UW until 3:04 AM and to Beacon Hill until 3:49. Is this accurate??

    1. I realize now it might be using my local time zone to display the schedule. That’s pretty odd…

Comments are closed.