At first glance, the passenger rail systems in Seattle and Tokyo have nothing in common. Seattle has seven passenger rail lines (1 Line, 2 Line, S Line, N Line, First Hill Line, South Lake Union Line, Seattle Center Monorail). Meanwhile, Tokyo has 102 passenger rail lines. Seattle’s system has 130 thousand daily passengers while Tokyo’s has 41 million daily passengers. Despite these differences, there is one commonality between the transit systems: both Seattle and Tokyo each have two disconnected streetcar routes. (Note: “Tram” and “Streetcar” are used interchangeably in this article)

Tokyo’s passenger rail network. The two tram lines are highlighted: Arakawa Line in green and Setagaya Line in yellow.

Trams in Tokyo

Tokyo’s passenger rail network is one of the best in the world. Subways, regional trains, and Shinkansen high speed trains carry over a million people every hour. Before many of these services were built, Tokyo had a dense network of trams called the Tokyo Toden. At its peak, the Toden system had 41 routes. Over the past 75 years, almost all of these routes have been replaced by subways or buses, except the Arakawa Line. The Arakawa Line is a 12.2km (7.6mi) cross-town route that connects several neighborhoods and ten subway lines. The route primarily uses a dedicated right-of-way (ROW) but it does have a short segment in mixed traffic. Despite 5 minute frequencies during the day, the trams are consistently full. In 2018, the Arakawa Line carried over 48,000 passengers every day.

Minowabashi bound tram on the Arakawa Line at Oji-ekimae Station.
Shhimo-takaido bound cat themed tram on the Setagaya Line at Miyanosaka Station.

On the west side of Tokyo, trams also carry passengers on the Setagaya Line. This route was originally built as a branch of the Tamagawa Line (then a tram route). By the 1960s, the Tamagawa Line was plagued by traffic congestion as it ran in mixed traffic. A subway was built to replace the Tamagawa Line in 1969. As the Setagaya branch uses a dedicated ROW, it remained open. The Setagaya Line serves 10 stations along its 5km (3.1mi) route. Both terminus stations and one intermediate station have transfers to subway lines. In 2017, the route carried over 57,000 people each day. One of the trams on the route is cat themed to honor the nearby Gokoku-ji temple.

Streetcars in Seattle

Similar to Tokyo, Seattle also has two disconnected streetcar routes. The South Lake Union (SLU) Line opened in 2007 and travels between Lake Union and Westlake station, through the heart of SLU. The line was built to serve the quickly developing SLU neighborhood and connect it with the Link light rail system. The 1.3mi (2.1 km) route has sections in mixed traffic, bus only lanes, and a short section of dedicated ROW near Lake Union Park. Due to it’s short distance, poor reliability, and redundant bus service, the SLU Line only had about 500 passengers per day in 2024.

Two northbound streetcars on the South Lake Union Line stuck in traffic.

Opened in 2016, the First Hill Line connects Capitol Hill, First Hill, Yesler Terrace, the International District, and Pioneer Square. The route was built as a replacement for the cancelled First Hill Link station. The First Hill Line follows a direct path on Broadway between Capitol Hill station and Yesler Terrace. Due to the steep grade, it takes a several block detour between Yesler Terrace and the International District. Outside of a 1 block section, the 2.5mi (4.0km) First Hill line operates in mixed traffic. In 2023, the route carried about 3,200 people every day.

Observations

While the rolling stock on the tram routes in Tokyo and on the streetcar routes in Seattle are similar, just about everything else is different. Tokyo’s trams primarily run in dedicated ROW on longer routes with many destinations and transfers. Seattle’s streetcars run in mixed traffic on short routes that only have a few destinations. Seattle’s streetcars could be improved with the addition of transit only lane restrictions along their routes, but they are still short routes. A plan to connect the routes through downtown Seattle is all but cancelled. Unfortunately, Seattle’s engineers were really good at removing or paving over tracks in the mid-20th century, so there are not many (if any) dedicated track ROWs that could be used for a tram. This mistake continues to cost the City, as building new tracks is really expensive. Perhaps it is time for Seattle to re-evaluate the necessity of it’s streetcars.

This is an open thread.

40 Replies to “Friday Roundtable: Tokyo’s Trams vs Seattle’s Streetcars”

  1. The South Lake Union Streetcar was built because Paul Allen wanted it.

    The First Hill Streetcar could do much better if it were frequent enough to be faster than walking. But to be more frequent, it would need more vehicles, and more storage space.

    The straight line from Capitol Hill Station to Yesler Terrace is certainly a good path that merits better than 12-minute headway. Perhaps running a bus route at 4-minute peak headway and 5-minute off-peak headway along Broadway could prove that point.

    There is no good reason for it to have lower ridership than the Swift Orange Line, except the Orange is more frequent and has enough space between its major destinations that it is much faster than walking.

    If I were designing a 1st Ave Streetcar from scratch, the idea of having it make a u-turn and go up Broadway would not be on my Bingo card for possible extensions. But serving most of the length of Jackson St would.

    That said, there is a reason why no bus routes dare to go west of 2nd Ave extension on Jackson St. They would get stuck in crowds before and after major events. The streetcar would get really good ridership at these times, if only it could turn back at Japantown Station. Adding a turnback just west of Japantown Station.

    While investing in any streetcar improvements, safety for cyclists should be front-of-mind. Has the City looked at options for making the tracks less deadly?

      1. Interesting. If buses were sent to First, hopefully they could turn on Yesler. That is better from a connectivity standpoint as well. For example assume the 24/33/124 is sent to First Avenue. Now think of someone going from Georgetown to Aurora. They get off the 124 at Prefontaine & Yesler, then catch the RapidRide E at the same stop. The 124 then turns on Yesler and heads to First Avenue.

      2. Part of the reason why buses don’t go further west than 3rd is that there are weight restrictions through much of Pioneer Square.

        How were they able to get the Benson line all the way to the waterfront in that case? Or is that one of the excepted streets?

      3. I think modern streetcars are just heavier with onboard batteries, safety features, etc.

        iirc, that was one of the major reasons for cost increases with the 1st ave streetcar: the original estimate was considering older, lighter models which wouldn’t necessitate a street rebuild and then deeper analysis revised that.

    2. The streetcars were a stupid idea. Modern trams a niche service. They make sense when you don’t have the money to invest in a metro (or you’ve already built one) and your buses are overwhelmed. For example it would make sense on Broadway in Vancouver (if they weren’t building a new SkyTrain line). It makes sense in Paris because they already have an excellent metro and many of the buses were overloaded with riders. Seattle never had that. We have “spines” instead. The handful of independent corridors (Eastlake, Aurora, Rainier) still don’t have enough riders to justify a streetcar. The whole idea with the streetcar is that you run it less often. That is very counter-intuitive. After all, when was the last time you wanted your bus (or train) to come less often? But in some circumstances (like Broadway in Vancouver or various corridors in Paris) it is a problem. Not for the riders on the corridor but the agency in general. They have to spend way too much money running buses too often. Riders don’t benefit when you go from running a bus every 3 minutes to every 2. Yet it costs the agency a fortune — money that could go into increase service somewhere else.

      Streetcars also have a tourist value. This is especially true of legacy streetcars. Streetcars in New Orleans and San Fransisco add charm. This was true of the Benson streetcars. It is a shame that they are gone right before the waterfront had its big makeover. The Benson streetcars made a lot of sense. Our current streetcars don’t.

      Part of the problem is the inflexibility with the routing. The South Lake Union routing had some value when it opened, now it is outdated. Yet combining it with other routes is impossible. It basically just duplicates what some of the buses do but in an awkward way. The First Hill Streetcar is better but it has flawed routing as well. The button hook is really bad. It would be trivial to fix it if it was a bus, but it isn’t. More to the point, if it was a bus you could easily reroute things in a much better way. As I suggested below, imagine this: Have the 49 turn west at Jackson, not Pine. Then straighten out the 60 so that it runs along Broadway the entire time. Now run the 49 more often and time them so riders get 6 minute headways along Broadway. This is obviously much better routing but hard to pull off with streetcars.

      A streetcar on First Avenue would improve ridership for the South Lake Union section. But it would require spending a lot more money on what is essentially duplicative service. In contrast if you shifted a few buses over a couple blocks it wouldn’t cost anything. While the routing would be an improvement for South Lake Union, it would do very little for First Hill. It would be a flawed route. If you ran a bus like the 24/33/124 along First Avenue then every trip combination is plausible. It remains a fairly straight line through the heart of the city. But with the streetcar it doubles back on itself. For example, no one would take it from Broadway & Pine to First & Pine (no matter how often it ran). There are just too many bad combination. The problem is that it is too hard to rearrange the streetcar routes.

      This inflexibility also extends to right-of-way. Jackson could really use BAT lanes. But the streetcar runs in the middle of the street and the buses don’t. That means either spending a bunch of money on buses with doors on both sides (for routes that involve trolleys and non-trolleys) or moving the tracks. Even minor things are difficult. It isn’t hard to imagine various BAT/bus sections for Broadway (similar to the tiny bus-lane section on Denny for the 8 that includes an island platform). But little things like that are made much more expensive because of the tracks.

      Clearly we made the wrong choice in adding the streetcars. As I wrote below, the best thing to do from a transit perspective is to get rid of all of them. But I wouldn’t go that far. Streetcars have a tourist value. I would get rid of the South Lake Union Streetcar and run the RapidRide H up there. This would greatly improve transit for those who currently ride the streetcar or RapidRide C. It would also improve the RapidRide H. Cyclists would also be much better off.

      The first thing I would to on Broadway is straighten out the 60 and time it with the streetcar. That way riders would have more frequent service on Broadway (6 minute midday is possible without extra funding). Eventually I would juggle around the 36, 49 and 60 (maybe like this) but that could come later (as part of a bigger restructure).

    3. The First Hill streetcar was because some activists claiming to speak for First Hill wanted a streetcar as mitigation for not getting the First Hill Link station, and wouldn’t accept anything else. So there was no consideration of whether a trolleybus might be just as effective and more cost-effective. The folly of mixed-traffic streetcars is it’s the worst of both worlds: the cost of rail but the travel time of a bus — or actually worse than a bus. The reason to spend money on rail is because it can be faster than buses, more frequent, and have more capacity. If it doesn’t do those, what’s the point? So the First Hill activists were blind to issues other than mode.

      A trolleybus could have used the existing wire, avoided the 14th Avenue detour, and shared the same bus stops as the other routes (7, 14, 26 on Jackson; 9, 60, and some 43s on Broadway) allowing people to wait less.

      1. “The First Hill streetcar was because some activists claiming to speak for First Hill wanted a streetcar as mitigation for not getting the First Hill Link station, and wouldn’t accept anything else.”

        Ah yes planning by a committee of people who never have lived riding a streetcar everyday. That’s how we like to do things in Seattle: Ask a bunch of people with a voice what they “think” should happen but don’t tell them what the travel time will be. Then once the specific alignment is chosen, add lots of cost to the project and lots of minutes to the operating round trip time by saddling it with design changes made for other reasons.

        Then once it’s open, operate it at a huge deficit for many years while everyone whines about its slow speed — as if it’s a surprising outcome.

        And speculate that extending it further will solve the problem, which bit won’t because the problem was always terrible travel speed on the original segment.

        Cynical, yes! But no extension is going to make them operate any faster — and that’s the fundamental problem with them.

    4. Buses already get killed by traffic around major events. I was on a 36 that sat at 3rd/Main for ten minutes after the Seahawks game until I bailed. I walked to the Link stop instead, saw the previous 36 that left 10 minutes before ours stuck at 2nd Extension and Jackson, and had a “laugh” (I was very mad). Maybe it was a one off, but it was ridiculously bad. Through car traffic from 2nd and Jackson was being continually prioritized over full buses that were backing all the way up to 3rd/Yesler.

      Surface transit needs to be actually planned for and prioritized if it’s going to be a good option for people at major events. I’ve seen really good stuff after UW games, but it was awful last weekend. It doesn’t matter if it’s a tram, a bus, a streetcar, a horse and buggy, or a taxi – it just has to be prioritized over cars.

    5. There is no good reason for it to have lower ridership than the Swift Orange Line, except the Orange is more frequent and has enough space between its major destinations that it is much faster than walking.

      Except the Swift orange line is 11 miles long. You can’t extend the first hill streetcar in any direction that makes sense an additional 9 miles.

    6. Should have kept Jefferson Base, at least the land for future use, not necessarily the building. Although modern streetcars are bigger than what was used there, still would have been able to accommodate more than the current fleet. Location, just a few blocks from the First Hill line, but nobody saw the First Hill Line coming when it was sold.

      1. “Should have kept Jefferson Base, at least the land for future use”

        What they should have done is not use incompatible car designs for the streetcars and for Link. If they used compatible car designs, they could be stored and maintained at Link facilities with little added real estate needs, and have a common spare parts inventory making them cheaper to operate and maintain.

        The Siemens cars used as light rail cars on Link also work just fine as streetcars when used that way on streetcar operations.

      2. The Siemens cars used as light rail cars on Link also work just fine as streetcars when used that way on streetcar operations.

        Exactly. The author forgot that Seattle also has tramway service on Martin Luther King Jr Way. With reserved right-of-way no less.

        Now, grant, that reserved right-of-way is not well enough protected; private vehicles intrude on it in manifold ways and far too often for anyone’s good.

        But they’re still trams in reserved right-of-way. And a stubbed and swollen toe in the Regional Metro that is Link.

      3. What they should have done is not use incompatible car designs for the streetcars and for Link.

        What they should have done is not built the streetcars. Does anyone really think they were a good idea? But yes, given they went ahead and built them they should have at least used the same type of cars.

      4. Do the Portland streetcars use the same vehicle type as MAX? Allen modeled the SLU streetcar after the Portland streetcars and the Pearl District.

      5. “Do the Portland streetcars use the same vehicle type as MAX? ”

        Mostly no with a bit of yes. Reading up on the specs. I see that they do have identical track gauges and power systems. That enables both technologies to run on the Tilikum Crossing bridge using the same tracks.

        However, the streetcar track ballast (what’s beneath the rails) is apparently not designed to handle the weight of MAX vehicles, and the vehicles are not the same width — and the trains are not the same length.

        The downtown streetcars in San Francisco, San Jose, Salt Lake City and San Diego also have some deliberate compatibility with their light rail systems.

      6. Yes, Portland was conned into making the same mistake as Seattle.

        The streetcars in both Seattle and Portland are 8 inches narrower than the light rail cars. This is enough to make the platforms incompatible but not enough to make a difference in street space consumed.

        Car length is ≈91 feet for Link and MAX light rail cars and 66 1/2,feet for the streetcars. Shorter versions of the Siemens light rail cars have been built for such operations as the Atlanta streetcar and they perform fine in that role, if for some reason you must have a streetcar that has less capacity than a Metro articulated bus.

        As to the track construction differences cited, that’s something that was put out by the group promoting a streetcar revival and repeated in a number of documents since, but I have yet to see anyone show what those actual differences are.

        When they put in the Portland streetcar, they used some sort of polymer bond in the concrete that supposedly made it cheaper and faster to install and had not been used in the USA before. I have not seen any evidence that this couldn’t be used for light rail lines in the street as well, nor have I seen any evidence that the cost difference between installing a streetcar line track and a light rail track in the street is meaningfully different. You all lived through the construction effort for the first hill line, and it didn’t seem especially different than when they put MAX on the transit mall. In both cases the city will want to move utilities, get a good subrisdbed down, etc.

        The weight per axle of both cars winds up being in the same range. Skoda 10T3 used in Portland is 63,400 lbs. The Siemens SD660 as used in Portland is 108,000 lbs. That’s 16,850 lbs per axle for the Portland streetcar and 18,000 lbs per axle for the MAX light rail car. This weight difference is meaningless when it comes to rail vehicles, and if your track is unable to tolerate that level of weight difference you’ve got serious problems with your track. Among other things it means the weight of your passengers would have to be carefully controlled, and there’s no way you’d be able to change car designs without very carefully managing the weight.

        I would point out that the Portland Skoda streetcars were built before the new burn-through limit requiring heavier floors in streetcars in the USA, so right there you’ve probably exceeded that 2,000 lb per axle difference.

        To me, it’s pretty obvious from these numbers there really is no meaningful difference in the track construction requirements. We’re talking margin of error type difference in weight once you look at the weight per axle.

    1. Yeah, the First Hill Streetcar has always been fairly productive. That is pretty much what you would expect with an urban line like that. Despite the weird button hook it still makes sense as a way to get from the south end of downtown/CID to First Hill. It also works if you just want to catch a bus along Broadway. The routing isn’t ideal but it is still pretty good. If you replaced the functionality with buses* it would likely increase ridership. But given the novelty factor of the streetcar it makes sense to keep it.

      The South Lake Union Streetcar is another matter. The ridership is extremely low for an urban route. It would be much better to just get rid of it and send the RapidRide H to South Lake Union. You could pave over the tracks, move the bike lanes and add BAT lanes on Valley (https://seattletransitblog.com/2024/10/04/replacing-the-south-lake-union-streetcar/). Transit Riders would be much better off. Bike riders would be much better off. The city could sell the streetcar depot. It would just be much better all around.

      *I would extend the 49 along Broadway to Jackson and then have it turn west to the same layover. I would straighten out the 60 as well. Then I would run the 49 more often and synch up the routes to get a combined 6-minute headway along Broadway. This would be better for transit riders but not as good for tourists (or other streetcar fans).

    1. That map gives me a headache, not even King County Metro has that many buses in their city, let alone Sound Transit. Dang in Tokyo there are more trains than buses in Seattle, also speaking of which I’m planning a field trip for the blog to go to Tokyo in the near future to explore the lifestyle and transit there, who would like to go?

      1. Tokyo is the biggest city in the world. It isn’t that surprising they have a first class transit system. But my guess is smaller cities are much better than comparable ones in the US. Even if they mostly get by with buses, my guess is the bus system is just a lot better.

  4. Who’s been to the new T&T at Lynnwood Crossroads (old Sprouts location), Yesterday, I rode the Swift Orange Line from Edmonds College to Hwy 99, and I noticed as it was a cold day, the ride took TEN minutes, like I feel like 68th needs more lanes, as the reason why it took 10+ minutes just to go to the new T&T was mainly due to the school buses dropping off students (and of course peak congestion), overall I did like the new T&T, but I wasted over 60 bucks on groceries there, too expensive. But my main point of this reply is just the congestion on 68th Ave at peak, and I want to know why a ride from Edmonds College to 99 took ten minutes.

    1. Most of the suburbs have little in the way of right-of-way despite fancy “BRT” lines. Shoreline is an exception. But Bellevue and Lynnwood could use a lot more red paint.

  6. At this point, just remove the streetcars and improve buses, they should be illegal in the US, as well as the useless T Line. They don’t cause anything but trouble, not to mention ST now regrets building the MLK segment at-grade, all to just save money, but guess what they now will have to spend more now, ST what have you done?

  7. I will be super interested to see Wilson’s views on both this topic and dstt2. She has played it pretty close to the vest so far aside from being “protransit”.

  8. “The downtown streetcars in San Francisco, San Jose, Salt Lake City and San Diego also have some deliberate compatibility with their light rail systems.”

    I rode the F line in San Francisco, but I didn’t know San Jose, San Diego, or Salt Lake City had streetcars.

    1. The San Diego light rail line originally had street running sections with fold out staircases, much like Toronto’s streetcar. I think the last of that was eliminated in the early 1990s, converting it to act a bit more like light rail in those areas.

      They also had to separate their operation a bit more from freight service once operated by San Diego & Arizona Eastern on one section on their line.

      1. Here’s a photo I found of their original “station” arrangement, though this is the south end of the line at the border. Note: no station platforms even here. It’s set up to board at street level, back before anyone tried to say there was any difference between “light rail” and “streetcar”.

        https://www.flickr.com/photos/megaanorak/52383376955/

    2. Salt Lake has a short, infrequent streetcar line in South Salt Lake that operates as a trunk connecting the Fairmont neighborhood with the Central Pointe station that serves all three of their light rail lines.

  9. Would high-speed rail actually be a reasonable commuting option for say Tacoma to Everett? Taking a look at Alon Levy’s analysis from a couple of years ago (https://pedestrianobservations.com/2024/05/17/fares-on-high-speed-trains/) we could expect it to cost between $30 and $50 per round trip (approximate European and Japanese prices) or more.

    Granted that’s less than the IRS mileage cost for that trip but I feel like folks will balk at paying ~$1000/month in HSR fares to get to work versus ~$20 in fuel costs.

    There will always be a few folks who are willing to make that commute, but will there be enough to fill significant transit-oriented development around HSR stations?

    2. I believe very few people commute this far every day. It is expensive to super-commute by any mode. I don’t think HSR is supposed to serve much of daily commute. It serves travel demand like intercity business trip current served by airline.

      1. Yes, although it can also double for those types of commutes. Something like Everett to Tacoma is fairly rare. But Tacoma to Seattle is not. Baltimore to DC is similar (and a lot bigger). The commuter trains (MARC) and Amtrak trains run quickly and frequently between the cities. I could easily see the same thing for this region.

    3. I guess that depends on what you mean by “HSR”?

      People are calling Brightline Florida “high speed rail” but its average speed is about the same as Berlin’s RE1 commuter trains. At an average speed of 70 mph including station stops something like RE1 would be more than enough to attract people away from driving for that distance of service, and it’s probably what most of Europe would use for commuting of that type. The next level up in speed isn’t covered by the regional transit pass, and would probably stop in both of those cities, but would be just an intercity express.

      High Speed Rail of the European type does have its daily users, but those passengers would be of the type that currently commute from Portland to Seattle (there are a fair number, believe it or not) rather than Tacoma to Everett.

      1. The terms for fast trains are pretty vague, which doesn’t help the discussion (high speed rail, higher speed rail). The terms are focused on speed and tend to ignore frequency. What if you had one really fast train running from Seattle to Portland every day and it passed by Tacoma at noon. That isn’t really effective.

        Some have called for trains capable of running 250 mph between Vancouver and Eugene. I think that is completely unnecessary and a distraction. Making the trains more reliable, frequent and a bit faster would be a much better value. Implement this and it would be great. It isn’t like L. A. to San Fransisco. It isn’t that far from Seattle to Portland or Seattle to Vancouver. You don’t need to be that fast before the advantages of rail outweigh those of flying or driving. The one exception is Portland to Vancouver. People might keep flying between those two cities. My guess is relatively few people do. Driving between there would become even less attractive.

        Likewise, relatively few people go between Everett and Tacoma every day. Those cities are just too small. Not only that, but people who “super commute” are way more likely to spend most of their days working from home. I know someone who lives in Seattle and works for the state. His office is in Olympia. He goes down there maybe three times a month — often staying overnight. This was the deal when he was hired. He does spend a lot of time on the road but often that is going to places like Yakima or the Bellingham (since he works for the state). Janitors and baristas are not super-commuting; it is generally higher paid office workers. They negotiate their commute.

        In contrast there are still a lot of people commuting from Tacoma to Seattle every day. Any improvement to Sounder would help quite a bit. If this was combined with improvements to Portland and Vancouver it would be great. Improving trips like Everett to Tacoma would be a side benefit.

    4. Did just that tyoe of commute when I was doing Y2K work at the State for about 3 years (Bothell to Olympia).
      The makeup of IRS/AAA mileage cost calculations was driven home when I realized I would essentially “use up” my car in only a few years. “Sunk cost” gets a little bit ‘squishy’ when you’re doing it at that accelerated rate.
      This driving ustification game is always played when comparisons are made, and it’s a perfect example of cognitive dissonance.
      That’s a 90 minute commute driving, by the way.

      I experimented with the bus, Metro-Pierce-Intercity. Spent almost as time on the bus (or waiting), as I did at work.
      Found a vanpool from the Roosevelt district made up of State employees who did the commute daily.
      They peferred the social life of Seattle. Everyone was married, which was why they tolerated the commute, because the joke was “Olympia was where single people go to die alone”.

      The other thing about driving that distance is “Highway Hypnosis”.
      That’s a real hazard.

  10. “Unfortunately, Seattle’s engineers were really good at removing or paving over tracks in the mid-20th century, so there are not many (if any) dedicated track ROWs that could be used for a tram.”

    Early 1900s era streetcar served very different purpose than urban rail transit today.

    Their reason to exist was because back then combustible engine was not good enough to power large transit vehicle, so almost every transit route was built as streetcar. Ever since trackless trolley and later diesel engine became available, streetcars were obsolete.

    Even if those tracks were not removed, it would still take tremendous effort to modernizes them for today’s use. Most of those services were slower than any KCM bus today and has no right of way either.

Comments are closed.