Portland: Facing the East-West Chokepoint

That red line (and the adjacent blue line that’s hard to see) is the east-west spine of Portland’s transit system. On the west, it is one of just two direct paths (street or transit) across the hills linking downtown and the the “Silicon Forest” suburbs to the west. But the slow operations across downtown makes this line much less useful than it looks. Credit: Travel Portland.

In 1986, Portland opened one of the first modern light rail lines in the US, the beginning of a light rail renaissance that built networks in mid-sized cities across the country. It was nice to be a leader — we’re used to that in Portland — but it also means that everyone has learned from our mistakes, while Portland still has to live with them.

Perhaps “mistake” is too strong a word, but the priorities of the early light rail designers certainly aren’t the priorities today.  Planners of the 1970s (when I was an enthusiastic teenage transit geek) confronted a city whose downtown consisted mostly of surface parking while prosperity fled to the suburbs.  Their top priority wasn’t even getting people to downtown.  It was making downtown a place worth going.

So they built a line that rushes into the city from the eastern suburbs, but then creeps across the inner city, making lots of stops for a net speed under 7 miles/hour.  For a while that was fine.  All those stations meant lots of development sites right next to the line, and downtown grew and prospered.

Today, the success at revitalizing downtown has created an opposite set of problems.  Downtown and the surrounding neighborhoods are so successful that working people can’t afford to live there.  Low income people are being pushed out to places where they face longer commutes.  Most important, the line now continues west out of downtown to serve the “Silicon Forest” suburbs to the west, so that it runs across downtown, not just into it.  Meanwhile, the development that the slow downtown segment was supposed to stimulate is largely done.

So the downtown segment of the line is becoming more of a barrier than a resource.

Transit in Portland benefits enormously from chokepoint geography.  Between the inner city and the western suburbs, there is a wall of hills pierced by one gridlocked freeway, one parallel arterial, and a light rail line.  This rail segment has prospered because the driving alternatives are terrible.

As always, chokepoint geography means: It’s worth spending a lot of money improving transit here, because so many trips, between so many places, go through this point.  A regional inbalance of jobs and housing (more jobs in the west, more housing in the east) has create a huge east-west travel demand right through this ch0kepoint.  The hills are still a barrier for drivers, but for transit the barrier is the slow downtown streets, and the 1970s assumption that the train needs to stop near every building.

As if the slow operations weren’t bad enough, there’s also a problem of capacity.  Portland’s adorable little 200-foot blocks, rightly credited with giving downtown such a human scale, limit the train lengths to 2 cars as long as they run on the surface.  The city is too big now, and its transit needs too dire, for such tiny trains to do the job.

The problem is being attacked at several scales.  The transit agency is gingerly suggesting that a few stations should be closed. Stations on the downtown segment are as close as 350 feet — far too close for bus stops, let alone rail stations.  (The newer north-south light rail line, whose designers learned from the mistake, has station spacing closer to 1000 feet.  Unfortunately, it is the east-west line that extends furthest into the suburbs and therefore serves the most people.)

But the problem is so big, and obstructs so much access to opportunity across the region, that it won’t be solved just with half measures. A long overdue study is looking at the complex of capacity problems, and while it’s looked at some half-measures, the only thing that solves all the problems is a new segment of subway under the core.  The long frequent east-west lines serving suburbs (and the airport) would go into a tunnel, rushing under downtown in perhaps 1/3 the time, so that transit would finally be viable for travel across downtown and not just to it.  The existing surface line would still be used to provide a more local service across downtown.

An early concept for the downtown rail tunnel (black) with existing light rail segments in red. The tunnel has six stations counting the endpoints. Too many?

I have been skeptical of many rail projects in my time, but the most defensible of all are these “core capacity” projects.  Like the excellent Los Angeles Regional Connector, this is a project that is in downtown but not for downtown. Its purpose is to unlock the potential for all kinds of access across the region.  To bypass the inevitable edge-core debate, it will have to be presented in those terms.

That’s why I’m a little skeptical of the earliest concepts.  As sketched the tunnel has six stations downtown.  They should at least study a line that just has three: the two edges of downtown — Lloyd Center and Goose Hollow — where the fast line would connect with the slow surface line, and just one station at the very center of the city, Pioneer Courthouse Square, where almost all of they city’s radial transit services meet.  This would make the new line barely half as long and much less expensive.

Obviously there are great arguments for every proposed downtown station: the university, the stadium, the train station.  But it’s going to be important to have clear conversations about the balance between downtown and regional benefit, and between the benefits to an already prosperous downtown and the need for reasonable travel times for low income people, who are increasingly pushed further out where they have to travel further.

I don’t know that a three-station solution is right, but I know it should be looked at.  It’s really easy to get around downtown on transit, from anywhere to one of the three stations that this minimal version would offer.  It’s really hard to get across the region, and every station we add to this project moves us back toward not solving that problem — not just by making the line slower but also by making it more expensive.  It’s a tough balance, and I hope we’ll have the debate.



25 Responses to Portland: Facing the East-West Chokepoint

  1. Roberta Robles July 9, 2019 at 6:45 am #

    I’m really going to be honest here; We need to be talking about developing new core urban centers outside of downtown. I get wha you are saying about network dynamics but you don’t address the critical issue: there are other urban centers across the region that we could be increasing densities. I’d rather not dig up downtown and instead spend it at new urban cores like Killingsworth, Vancouver, Gresham, Tigard and Hillsboro. Downtown is stuffed to the brim with success, so let’s evolve network planning, which means redistributing the main origins and destinations to other more equitable and lower cost real estate locations. That’s the core debate in NZ you seem to be loosing track of; the strength of their non-urban growth centers. Plus EQUITY… downtown had its chance let’s try these concepts in less expensive places.

    • anonymouse July 9, 2019 at 7:00 am #

      The thing with developing new core urban centers outside Downtown is that is just means even more travel through Downtown, making the subway that much more important. Employers want access to the whole regional pool of employees, employees want access to the whole regional job market. Particularly in a world of two-income households, people aren’t just going to be able to sort themselves by where their jobs are located. Which is really just the point that Jarrett was trying to make above.

    • Nilo July 9, 2019 at 10:31 am #

      Portland’s downtown is not filled. Chicago’s Loop is about a third bigger than Portland’s and has over 390k jobs, while Portland has 78k jobs downtown. This means plenty more job concentration still is possible in Portland.

      I’d also argue that it’s desirable based on these grounds. First the transit network as it stands is highly radial, and thus maximizing job access for all residents means putting jobs at or near the center of the network. Second, as Alon Levy has repeatedly observed a transit city is a highly centralized city. due to the dynamics of what makes transit efficient and what makes cars inefficient, meaning job dispersal leads to higher auto mode share, and poorer auto access for those jobs.

      Anyways Portland should build this tunnel and more skyscrapers in the center.


      • Jarrett July 11, 2019 at 5:30 pm #

        Portland’s street grid is not suited to Chicago skyscrapers. Blocks are too small.

    • Mike July 13, 2019 at 10:44 am #

      People need to get through the center to your beefed-up urban centers, so the tunnel actually benefits your goal. Seattle has a similar issue: the region explicitly targets future growth to a dozen edge cities like you’re suggesting (even though greater downtown is actually growing faster), and it decided a second downtown metro tunnel is necessary for both downtown circulation and to get people across the isthmus between the north end, south end, and Eastside. Both the north and south outlying counties want to attract more jobs and workers and think people will reverse-commute even from the other side of the isthmus. Portland could have a similar win-win-win effect, with increasing jobs and mobility in both downtown and your outlying centers and busy trains both directions all day.

      (There’s some doubt in Seattle about whether the outlying cities can really attract so many jobs, but they’re a lot further out than Porland’s. Everyone agrees the area within Lynnwood, Redmond, and Federal Way is solidly ready for larger urban centers and high-capacity transit, and Hillsboro to Gresham to somewhere south is in a similar situation.)

    • Mike July 13, 2019 at 11:32 am #

      Also, Seattle has “subarea equity”, which means that the north, south, east, and city subareas each have to pay for their own extensions out of their own taxes — there’s no cross-subsidy (in theory). But the second downtown tunnel was seen as benefiting all subareas because it does, and so they’re charged for their percentage of trains in both tunnels. Portland probably doesn’t have this funding situation — usually it’s just one regionwide tax base with central control — but it reinforces the idea that a downtown tunnel benefits all areas.

  2. Social Engineer July 9, 2019 at 3:20 pm #

    Is it any surprise that Jarrett supports the same three-station concept from the 2009 HCT System Plan, the one that he studied while he was an employee at Nelson Nygaard?

    Isn’t one of the objectives of a tunnel to serve dense activity centers in Central City, not just create the fastest possible trip through downtown? Don’t we want to serve existing (+future) households and jobs while increasing transit reliability through grade separation?

    For example, the Pearl District has over 7,000 residents, 15,000 jobs, 2,000 affordable housing units (many clustered in the north end where existing transit service is not adequate), and massive development potential at the Post Office site. Why should it be bypassed by MAX regional rail for a third time?

    • John Charles Wilson July 10, 2019 at 6:24 pm #

      I think Jarrett’s idea is to use the existing surface trains to access activity centers, and use the tunnel as an “express” to bypass them.

    • Jarrett July 11, 2019 at 5:29 pm #

      Actually, I left NN in 2005, before that plan was done.

  3. maccoinnich July 9, 2019 at 5:26 pm #

    I completely agree with the analysis that the MAX is far too slow through the Central City, and that severely limits how useful it is for cross-regional trips. A tunnel is the only way that we can increase the frequency, speed and capacity of trains, all of which are limited by the Steel Bridge bottleneck and/or the short length of our downtown blocks.

    Any tunnel will need to have fewer stations than we have today, or even than we will have once TriMet closes some of the original stations. It would be a mistake to place too many stations on the route… but it would also be a mistake to place too many. The proposed Union Station stop would be approximately 3,300 ft north of Pioneer Courthouse Sq; a larger stop spacing than BART in the Market St tunnel or LINK in the Downtown Seattle Transit Tunnel, both of which are also at the center of regional rail networks. Similarly, the two Lloyd District stations are approximately 2,300 ft apart, which is close to the spacing of BART in the Market St tunnel.

    The Lloyd District and the Pearl have seem enormous amounts of housing and job growth in recent decades, and are planned to have more. I don’t think that serving them adequately has to compromise the project goal of speeding up the MAX through the core.

    • maccoinnich July 9, 2019 at 5:34 pm #

      *also be a mistake to place too few.

      Dang it.

      • RossB July 10, 2019 at 8:02 am #

        I knew what you meant 🙂

        I also agree with your points.

  4. R. W. Rynerson July 9, 2019 at 6:53 pm #

    Wait a minute! Wouldn’t San Diego’s 1981 opening be considered the first modern LRT in the U.S.?

    The block length issue has an interesting sidebar. Don McDonald, the Edmonton engineer who also led Tri-Met’s project, hoped that the reliable Siemens-DuWag Frankfurt car selected by Edmonton, Calgary and San Diego would also win the Portland contract. Don was too ethical, however, and the process allowed for a variety of cars to be considered. Bombardier came in with a winning design from Belgian carbuilder BN that was longer than the SD car. And as it turned out, a two-car BN train fit a downtown block. It would have taken 2½ of the SD cars to do the same thing.

    Don adjusted to his disappointment, noting that “the Bombardier bid was low enough that I could afford to send someone to their factory to keep an eye on things.” It might seem odd to younger readers, but this was after a decade of equipment fiascos in the U.S. A new system starting out with a new car design was scary.

    I won’t try to settle the subway issue from out of town, but I know the Steel Bridge well and an alternative needs to be underway before there is a serious problem.

    • Jarrett July 11, 2019 at 5:28 pm #

      You got me on San Diego. I’ve revised the text.

  5. Ben July 9, 2019 at 8:34 pm #

    Here in Osaka I never really thought about how far apart the metro stops are because walking from one stop to another never really seems like a big deal (e.g. if I want to enjoy the weather instead of taking the train).

    So I measured on Google Maps. The distance between most stations? 500m minimum, but more often 800-1000m — that’s in the center of the city. And in reality, if you’re taking the train to your destination the theoretical maximum walking distance is half the station spacing.

    You are definitely correct Jarrett — 1000 FEET (300m) spacing is way TOO CLOSE. Especially with “downtown” being 2,000m across at most? So if you have stops on the very edges of downtown, you really only need a single stop in the middle. That puts everything within a 15 min walk at most (at a very casual pace) if you look at the walk radius between stations and consider that most of the time you won’t be walking to the far reaches of the radius anyway.

    Add to that the high percentage of bicycle ridership and the excellent bus system in Portland and there’s really no legit reason to put train stations so close together. Use trains for what they’re good for — distances that are too uncomfortable for most people to walk, and which need more capacity than bus — and let the other modes fill in the gaps.

    I hope they listen to you.

    • Andrew July 10, 2019 at 5:05 am #

      I’m guessing that was between Matsuyamachi and Tani-yon? I’ve spent a lot of time around that area, and those two stations do seem oddly close. The local neighborhood association probably lobbied hard for the former when the Nagahori-Ryokuchi Line was built.

      As for Portland, I agree that six stations is too many. But in addition to the stations Jarrett suggests, I would also keep Union Station if for nothing other than connectivity to Amtrak and commuter rail.

    • John July 12, 2019 at 5:32 pm #

      I found Japan a bit of an eye opener in terms of how the rail system has developed as a patchwork of different companies. Clean – yes, punctual – yes, integrated – hit and miss with stations on adjacent lines constructed some distance apart where they could have formed an integrated hub with centralised planning. Apparently the “Act on Enhancement of Convenience of Urban Railways” aims to mitigate this by building connecting routes where feasible in Tokyo, Osaka and Nagoya areas.

      As regards Portland or anywhere else, the walk from commuter rail to light rail needs to be short, well signposted and preferably under cover otherwise you can’t reasonably expect people to use it with other options.

  6. Jacob Manaker July 9, 2019 at 10:03 pm #

    A three-station plan seems too sparse. In the plan you cite, Park Hill and Pioneer Square are definitely too close together. Pioneer Square enables transfers to the surface lines, so Park Hill has got to go. But I’m much more wary of eliminating the stop at Union Station.

    At the very least the tunnel should be built with provisions to retrofit a station there. The mainline rail station isn’t going to move, and any future transit lines that leverage Portland’s freight rail lines *will* want transfers to the subway. Adding the stop later will cost orders of magnitude more than building it now, so it seems a relatively low-cost hedge against any future S-bahn-type service.

  7. RossB July 11, 2019 at 8:38 am #

    This article seems a bit vague. It isn’t clear how much time it would save to skip the two stops, or for that matter, how much time the tunnel will save. Let me take a stab at it:

    According to Google, it takes 17 minutes to get from one end of downtown to the other. If it takes 1/3 the time, that is roughly 6 minutes. That seems about right.

    Skipping three stops is mostly about avoiding dwell time. Assuming 30 seconds per stop (which is likely a bit generous) that is 90 seconds. Acceleration/deceleration adds some time, but not much, since the distances are short anyway (even with three stops a train won’t reach maximum speed). So add 30 seconds (again, likely a big generous).

    So basically you are talking about saving a couple minutes. You go from 17 minutes, to 6 minutes, to 4 minutes. That seems like a heavy price for those last 2 minutes. An 11 minute savings is huge; a two minute savings — not so much.

    Meanwhile, you give up a lot. Again, it isn’t clear what the plan is in terms if mixing the trains. You have the “local” (that runs on the surface) and the “express” (that runs underground). Is the idea to have half local, half express on each line? Or is it to favor the express, since that is where the city is investing? If you skip the stops, then the latter is difficult. Demand would be higher for the train “that makes all the stops”. An express is cheaper to run, since it completes its run faster. If you are running lots of locals, then this costs the agency money.

    Meanwhile, one of the stated purposes of this project is to deal with crowding. The local trains can only serve so many people. Because the two trains share the same track, this will actually reduce the number of local trains that can be run. If half the trains are local, and have express, the assumption is that lots of people will switch to the express. But if people prefer the local train then it will be more crowded than when you started. It makes sense to move as many people to the express as possible, since it will have a lot more capacity. The best way to do that is to have all the stops.

    This seems to assume that cross town traffic exceeds demand within the city, but again, I don’t see any data for that. I would also be very surprised if that is the long term trend. If Portland grows, it is likely to grow right in the heart of the city (the way most coastal cities have grown recently).

    The only argument for a stop diet that makes a lot of sense to me is cost. As Jacob said above, if that is really an issue, then the thing to do is allow the stations to be retrofitted. Very rarely (if ever) does an agency regret putting in a downtown station. But they certainly regret leaving one out (e. g. First Hill in Seattle).

    • DL July 11, 2019 at 5:23 pm #

      From what I remember of the VTA analysis of the now discontinued express trips on the Santa Teresa Line, the time it took to stop (including doors opening, people get on and off, and acceleration/deceleration so the whole deal) was 40 seconds per stop. Not sure if that applies here neccesarily but it gives you a more concrete starting point

    • Mike July 13, 2019 at 11:14 am #

      It’s not crosstown trips vs downtown trips, but crosstown trips plus Pioneer Square trips and all the transfers there and Lloyd Center-Pioneer Square trips and Lloyd Center-Goose Hollow trips vs the other downtown stations. The remaining surface trains function like the streetcar lines, so the question is whether any of its unique stations are so important they should be added to the express line. The current service is *really* slow because of both the Steel Bridge and so many stations — this solves the two biggest problems and still gives access to a lot of downtown. If Ben is right and it’s a 15-minute walk between the stations, then the longest walk is 7.5 minutes, which is borderline reasonable. It should ideally be 5 minutes and that would serve more of the triangle area outside the walk circles, but it’s close, and there are the surface trains for other parts. I could even see the surface segment turned into a shuttle. That seems better in some ways, and it would allow the metro lines and the shuttle line to independently have their own optimal frequencies rather than a compromise.

  8. Trebor July 13, 2019 at 10:14 am #

    I’m inclined to agree that three stops is too few. It would be hard to eliminate the Rose Quarter station given its importance for bus transfers (though not having a stop there would justify doubling frequencies on the Yellow-Orange Line). Park Ave would presumably be comparatively cheap given that it would be build on open land, and would presumably generate a lot of trips due to its useful proximity to PSU.

    Finally, I am not sure that the real utility of the tunnel would be for travel from eastern homes to western places of employment. Land use in Washington County is pretty abysmal. OTOH, people who live to the west and who work in the Lloyd would be huge beneficiaries of this plan–whether it involved six stops or three.

  9. Bill Henry July 13, 2019 at 5:31 pm #

    This concept map looks like it has 2-car length stations. Metro mentioned 4-car stations in the materials they released. That raises a lot of questions about how the stations are configured — both new and existing.

  10. Richard L Bullington July 14, 2019 at 1:23 am #

    The tunnel should “belly out” to the north to about the grain elevator just north of Broadway. This allows a much shallower and more useful station at Rose Quarter and placement of the “Union Station” station at about 9th and Marshall. This puts it adjacent to Union Station using the underpass under the Broadway Bridge approaches and very close to the large new buildings in the North Pearl. Have another station at 8th and Couch for the southern Pearl, and the Pioneer Square, Park Avenue and Goose Hollow stations as shown. Yes, this is one MORE stop, but it means the tunnel serves all of downtown Portland.

    Park Avenue is very important because it means that riders from the Northeast lines can transfer to the Orange Line (or for that matter to the Green Line extension) AFTER the slow meander around Union Station and down the Transit Mall. Riders from the west wouldn’t have to “double back” to Pioneer Square, and those from the Orange Line could transfer to the high speed line at the third stop downtown.

    All of the stations on the main trunk — :PDX to Elmonica — will have to be lengthened to four cars to make this work. It’s probably not possible to do that for the Blue Line east of Gateway, so the Green Line will have to become the four-car train partner of the Red Line. That’s a shame because there are a LOT of apartments which have grown up along the Blue Line that will lose their service to the west end, but I really can’t see how you’d provide four-car stations down Burnside.

  11. Richard L Bullington July 14, 2019 at 1:25 am #

    I just realized that the Blue Line could become the “Local” line preserving service from East Portland to the West End, though without the express speed-up.

Leave a Reply