Saturday, July 10, 2010

Rapid streetcar, Portland, and the rail divide

There's an article on the light-rail advocacy site, which introduces a concept known as Rapid Streetcar.  Rapid Streetcar, as they define it, is a rail transit system which is a hybrid between "traditional" mixed traffic streetcars, and light rail systems.  With a rapid streetcar design, streetcar-class rail vehicles (which weigh considerably less than light rail vehicles) are operated in exclusive rights-of-way in urban areas, often in lanes which were previously dedicated to automobile service, and often in multi-car trains.  The page goes into detail on the advantages (lower construction cost, ability to run in mixed traffic if necessary, the possibility of tram-train like operation) and limitations (speed and capacity) of such systems, and discusses Portland's rail systems in great detail.

Of course, there's one catch:  Portland doesn't HAVE any Rapid Streetcar.  The article primarily uses Portland's streetcar system as an example of how such a system might be designed (focusing on details such as construction costs and techniques, station design, etc), while glossing over the fact that Portland Streetcar is, at present, a mainly mixed-traffic circulator system, not something marketed as rapid transit (or rapid anything).  Portland does have rail lines with service parameters that satisfy the conditions for Rapid Streetcar--but those lines (such as the Yellow Line) are running light-rail rolling stock, not streetcars.  There has been some discussion of Rapid Streetcar in Portland, but currently none exists.

As the article notes, "Rapid Streetcar" is not really a novel idea; many European tram systems have been operating in this fashion for decades.  It is more novel in the American context--the legacy streetcar systems which operated in the early 20th century before being replaced with busses and the automobile, were almost always mixed-traffic types; and many of the "revival" streetcars, such as the Portland Streetcar, are this way as well.

The Portland rail divide

The Portland metropolitan area has three transit rail systems in operation; this post compares the three systems (along with two "full metro" systems in other West Coast cities).  The remainder of the present article will ignore everything but MAX and Portland Streetcar.
In many ways, the two systems are dramatically different.  MAX is a traditional light-rail system, operated by TriMet--a mobility-focused transit agency which also provides bus service to the bulk of the metro area.  MAX uses light-rail class vehicles (such as the Siemens S70), usually running in two-car trains.  It runs almost exclusively in its own right-of-way, and in many places it is built to "light metro" standards, with grade separation or guarded crossings.  Other than the downtown area, average stop spacing on MAX is generally 3000-3500 feet (1km); and average speed on the system is about 20MPH (32km/h).  MAX operates on the proof-of-payment system, with tickets vended at platforms.  The fare structure for MAX is the same as for TriMet's bus system (with the added benefit of the Free Rail Zone downtown); service is branded similarly to the busses (both use the same colorschemes); and MAX is not marketed in any way as a premium or upscale service.  (Many of the MAX lines pass through parts of town regarded as blighted, and the "MAX causes crime" meme is unfortunately popular ).   

Portland Streetcar, on the other hand, is a different beast.  While TriMet drivers drive the trains around, the Streetcar is owned and operated by Portland Streetcar Inc (PSI), a subsidiary of the City of  Portland.   Unlike TriMet, PSI is focused on urbanist and community-development outcomes; on "placemaking".   The Streetcar uses streetcar-class vehicles (the Skoda 10T) operating in single-car units, and provides local circulator service through downtown Portland.  It operates in mixed traffic, mostly on streets of secondary importance.  Average stop spacing is about 750' (230m), or about once every three blocks.  Average speed is less than 7MPH (11.1 km/h); slower than many busses.  The Streetcar also uses the same fare system as MAX and the busses (and passes and transfers are interchangeable between the systems)--however, the Streetcar is marketed differently, with a brightly-colored colorscheme and styling aimed at more upscale riders.  Onboard amenities are similar.  The Streetcar also operates on a proof-of-payment basis, albeit with ticket machines onboard rather than at platforms.

Divide is perhaps too strong a word--as the two systems coexist fairly well.  The two agencies are on friendly terms, despite differing missions, and the systems are presented to the public as parts of a unified transit system.  Presently, they cross each other in two places (at Portland State University, and where the Yamhill/Morrison and 10th/11th couplets intersect), with a third crossing near the Portland Convention Center coming in 2011 when the Streetcar Loop project opens.  However, at the present time, the two systems are completely separate--separate routes, seperate rolling stock, separate yards and maintenance facilities, separate organizations, and separate missions.

Change coming

The opening of the Streetcar Loop project in 2011 won't change any of this.  However, when the Milwaukie MAX line opens in 2015, a shock to the system will come.   The new Caruthers bridge, which will ferry MAX across the river south of downtown, will also carry Streetcars as the Streetcar Loop project is completed to become a real (bidirectional) loop winding through both sides of the Willamette River.  For the first time, MAX trains and Streetcars will share the same tracks while in revenue service.  They won't share any stops--Streetcars will pass through the MAX stop on the west side of the bridge without stopping (and the lines diverge immediately after that), but they will share rails and catenary.

The next planned Streetcar project after the Loop is the Lake Oswego transit project, which everyone expects will select Streetcar as the LPA (locally preferred alternative).   This project, planned as a southward extension of the current Streetcar service down the Jefferson Branch line, is being billed by some  (including local planners) as Rapid Streetcar.  However, the predicted speed of the line--approximately 19MPH (30km/h), is on the low edge for a rapid transit line.  South of South Waterfront, the line is projected to be fairly fast; however the present Streetcar route takes ten minutes between OHSU and PSU.  (Planned enhancements to the line should shave a few minutes off that; however the numerous stops along this stretch of the route limit the impact of track improvements).   Further, the proposed project may suffer from capacity issues--as much of the route is single-tracked, effective headways are limited to 12 minute (5 trains per hour), or 700 pphpd.  Finally, one other significant issue is that whereas the #35 bus (which the line intends to replace) travels the transit mall, the Streetcar extension would instead stay on the existing alignment along the 10th/11th couplet--a five block difference.  Many of the commuters from West Linn who use the #35 either work downtown or plan to transfer to another line on the Mall; and are not well-served by a line which instead takes them to the Pearl District).

Streetcars on the Mall?

Given all that--what if the Lake Oswego Streetcar, instead of being an extension to the existing line, were instead routed on the Transit Mall?  Doing so would provide the following advantages (albeit with the indicated conditions).

  • Better replace the #35 service than a 10th/11th alignment.  Operation on the Mall would deliver riders to the core of downtown Portland's commercial district, and to all the transfer opportunities found there.
  • Would make running trains--i.e. more than one car--a possibility.  As the current Streetcar alignment north of PSU runs through existing, mature urban fabric, and in mixed traffic, the service is limited to single-car "trains" in order to minimize impacts.  The transit mall, OTOH, already serves 2-car MAX trains (equivalent in length to three Streetcars) with no difficulty.  Much of the line south of PSU could more easily be expanded to accommodate larger vehicles; permitting the LO line to have capacities closer to what MAX can deliver.
  • To further improve performance, the proposed Harbor Drive transitway (connecting the new Caruthers bridge to Lincoln Street) could be modified to accomodate Lake Oswego bound streetcars, thereby skipping many of the intermediate stops
In short, running the LO Streetcar on the Transit Mall would separate the local service (the existing Streetcar line and the loop) from the rapid transit service (the Lake Oswego line).

Now how much would you pay?

[Map courtesy of City of Portland/ ]

 But wait...there's more!

Last year, the  City of Portland and Portland Streetcar published a draft Streetcar System Plan, which was passed by the city council last September.  The plan is a document outlining where Streetcar service might go in the city of Portland in the future.  The document isn't intended to articulate future corridors for high capacity transit (see here for that); and limits its coverage to the city of Portland, not to the suburbs.  Most of the routes proposed are local circulators such as the current Streetcar; however the document does address Rapid Streetcar, further dividing it up into "enhanced local service" and "priority service", proclaiming the LO Streetcar extension to be the latter.  The Plan has this to say on the subject:

The rapid streetcar concept aims to combine the best features of streetcars and light rail transit (LRT) to achieve faster commute/travel times than streetcars and lower system costs than light rail. Streetcars are typically designed to go shorter distances in central cities, densely populated mixed-use centers and neighborhoods. Streetcars are also typically designed to operate in mixed traffic, preserving street traffic patterns.

LRT typically functions as regional high-capacity transit (HCT), generally traveling in a separated right-of-way with relatively fast-moving, larger-capacity vehicles designed to rapidly transport large numbers of people between suburban and urban centers.

The rapid streetcar concept would apply some of the LRT features to streetcars to improve travel times while keeping capital costs lower. It would combine features of a semi-exclusive transitway and transit priority features within the street right-of-way to achieve faster travel times and maintain lower system
capital costs. This could introduce two new levels of service to Portland’s system.

Several corridors under consideration for the Streetcar System Concept Plan are prime candidates to introduce Enhanced Local Service. These corridors are major arterials with 4 to 5 lanes and on-street parking such as NE Sandy Boulevard and SE Foster Road.  In Portland there are potential corridors for introducing priority service. Currently, the region is undertaking a study to extend the existing streetcar system along a former railroad right-of-way from the South Waterfront District, through Johns Landing and south to Lake Oswego. SE Foster Road and 122nd Avenue are also candidates where there may be sufficient right-of-way width to introduce streetcar priority lanes.
I encourage all readers to download and read the draft plan--even if (or especially if) you think local service streetcars are a boondoggle or a weapon of gentrification.  (The document probably won't change your mind, but it may make you angry).

In the plan, numerous potential Streetcar corridors are considered.  Several screening functions are applied to exclude those which are technically unfeasible, unlikely to draw sufficient ridership, or unlikely to promote transformative land use outcomes; with the remaining corridors grouped into tiers.  More filtering was performed, with the following routes surviving all the cuts:
  • MLK from Alberta to the Loop.
  • Thurman/18th/Burnside/Sandy.  NW 18th/19th couplet, south to Burnside, east to Sandy to the Hollywood District.
  • The loop
  • The current route, with extension to Lake Oswego.
  • Sellwood Bridge, between LO route and Tacoma Street MAX.
  • Gateway Loop
  • Goose Hollow to 3rd/4th to Morrison Bridge to Belmont to 39th to Hawthorne to 50th. 
  • Broadway/Weidler from loop to Hollywood.
The plan assumes that the Powell Boulevard and Barbur Boulevard corridors will be constructed as light rail, and assumes that no sharing of track between MAX and Streetcars will occur, other than on the Caruthers Bridge.  If, however, Streetcars could use lower-speed MAX routes to get around, it may be possible to eliminate some redundant routes downtown, or use the existing Ruby Junction yards to service the Gateway Loop.  For example, if connectivity to St. Johns were desired, but extending the MLK line were thought to be difficult, it might be possible to lay tracks from Interstate only, and then have trains run down the Yellow Line--or even augment (or replace!) the Yellow Line itself with streetcar service, as a way of dealing with the Steel Bridge bottleneck.

Technical difficulties

How technically (and politically) feasible is any of this?  Both systems use the same rail gauge (standard) and power system (overhead catenary supplying 750VDC), so two key requirements for mixing the systems are satisfied.   The following technical issues and restrictions do exist:
  • The signalling systems are presently different; unfortunately, I'm not aware of the technical details, or how much it would cost to retrofit them to be the same (I imagine the Streetcar, with far fewer miles of track than MAX, would be converted). 
  • The Skoda streetcars are about 8" narrower than MAX trains--introducing a 4" or so platform gap were a Streetcar to use a platform designed for MAX.  That gap, too wide to simply instruct users to mind, could be bridged with platform extenders.  (For obvious reasons, MAX trains cannot use Streetcar platforms--but there are many other reasons MAX can't run on streetcar routes).
  • MAX trains are too heavy to run on the current Streetcar route, which has a smaller railbed.  (The use of the lighter-grade construction is one thing that produces the cost savings attributed to Rapid Streetcar).  Nor could MAX take many of  the corners successfully.  
  • The Skoda 10T streetcars have a top speed of 40-45MPH (64-70km/h); making their use on the high-speed parts of the MAX network (anything west of Goose Hollow, the I-84, I-205, and Airport segments, and the Milwaukie Line south of Holgate) inappropriate.  Were a faster vehicle available, this would not be an issue.
  • According to Curt Ailes at Huston Street Racing, who interviewed Oregon Iron Works president Chandra Brown on the subject, the 10T streetcars cannot be coupled for in-service operation.  They have couplers, but Ms. Brown suggested that they aren't suitable for operational use.  (Perhaps they are only for towing the streetcars; but the model in question lacks the electronics for tandem operation of multiple units, and/or the horsepower for one to pull another).  Skoda makes other streetcars (such as the 14T and its successors) which can be formed in trains; but current practice in the North American streetcar market seems to be to use streetcars are larger busses, and to stick with light rail or metro for high-volume applications. 
In addition, there may be political objections as well.  As noted above, the two agencies have different missions:  TriMet has an operations focus (claims by its critics to the contrary nonwithstanding), and PSI is focused more on land use.   PSI board member (and Portland planning commissioner) Chris Smith indicated, in a conversation on, that the former agency isn't interested in the land use goals of the latter, noting that:

I think it would be fair to characterize streetcar supporters (including the City of Portland) as believing that as soon as the streetcar system is handed over to TriMet it will stop expanding. So unless TriMet changes its culture and mission I don't see the organizations being merged for a long time. 

While increased joint operation of the Streetcar and MAX certainly doesn't require a merger between the two organizations, Smith has a valid point:  Many in Portland's transit community take a dim view of Portland Streetcar, in particular it's land-use focus--and may extrapolate negative views towards mixed-traffic rail and transit-oriented development into opposition to streetcars (the technology), even in applications where they make sense from a transit-focused perspective.  And, there are probably many on the TriMet operations side who would look unfavorably to an additional vehicle type to deal with.

Final thoughts

Does running Streetcar technology and light-rail technology on the same tracks, beyond the planned Caruthers bridge concurrence, ultimately make sense?  It's hard to say--for specific lines, such as Lake Oswego, a good argument can be made.  Elsewhere in the system, arguments get sketchier.  Long range, it may be the case that a downtown subway is built to handle light rail, and that the surface routes become exclusively served by streetcar.  It may be the case that rapid streetcar is selected for future transit corridors such as Powell, which already runs parallel to light rail.

It's also possible that the Lake Oswego project won't survive the DEIS process in its current form; and Streetcar technology is for the foreseeable future only used for local circulators--making the whole subject moot.

But in planning for the future, it's good to keep an open mind, especially in an era where budgets are tight.  And just because a vehicle looks like a streetcar, doesn't mean it has to act like one.


  1. Great post!!! I absolutely agree with you on LO streetcar via Transit Mall. I've been harping about this on PortlandTransport and have written on comment cards about the TriMet bridge asking for a connection/switch between the SoWa streetcar tracks and the LRT alignment into downtown, at the west end of the new bridge. At least allow for the possibility to route the LO line via the Transit Mall. The LO streetcar needs to run express where there are duplicate routes, plus OHSU tram riders need a fast trip into downtown. I personally feel the Skoda/Inekon/OIW streetcars are inappropriate for the LO line (they are great for a standing-based downtown circulator), something like a modern PCC would be better suited for LO in my opinion.

    Why are they planning to add streetcar tracks later on the new Sellwood Bridge? They should be installed during the construction of the bridge. You dont want to have to close that key link to install the tracks a few years later. I do think that a Tacoma St Streetcar is an important route (plus no existing bus service on route) and of course once youve got tracks on the bridge you'd only need, what, another 1/2 mile of track to get it to the Tacoma St MAX station? Tacoma St Streetcar seem to get a lot of route for not much new track. On this topic of the Sellwood Bridge, I'd like to see TriMet add several crosstown bus lines between SW Portland and SE Portland/Milwaukie when the new bridge reopens. For an example, the recent article in the Oregonian about Neil McFarlane indirectly mentioned that he takes 3 buses to get from SW Portland to the TriMet HQ, 43->9->70 (it mentioned him taking the 43, a bus over the RI bridge and the 70)... thats crazy.

    A streetcar line that I think makes a lot of sense for the streetcar system plan is a branch off the MLK line down Alberta to about 33rd, so that Alberta Street can have direct service into downtown. I suppose you could even extend it a little further to Kennedy School for a terminus. The beauty of an Alberta Streetcar is that you aren't cannibalizing any bus routes since the 72 is a very different animal.

    It seems the Gateway Loop is a PDC wish/equity (throwing a bone) to East Portland since they have been left out on the urban projects of inner Portland. As much as I see the equity reason for a Gateway streetcar, Gateway and vicinity simply doesnt have any of the urban built form required of a streetcar.

  2. Average speed is less than 7MPH (11.1 km/h); slower than many busses.

    I understand what you're getting at (calculating average speed as how long it takes for the vehicle to cover a set distance) but I personally never liked calculating average speeds of various forms of public transportation in that way - obviously it'll be slower than a private car that doesn't make multiple stops, but the average speed that the streetcar is going to travel *when in motion* is probably just under 15mph, and its average speed downtown is probably comparable to buses when you don't factor in their out-of-downtown speeds.

    And point of minor interest - the assistant manager of operations at PSC had previously been an assistant manager of rail transportation at TriMet until he retired from that role. I have no idea if that has any impact on the integration of the two systems or not.

  3. @Jon;

    It's my understanding that the new Sellwood Bridge will include streetcar tracks (though not catenary, which can be added later at far less expense and inconvenience); Mayor Adams has been given to much posturing about how the bridge will be "streetcar ready". Of course, I don't particularly LIKE the idea of replacing the Sellwood in its present location; I'd rather it be built a few blocks south, parallel with Golf Junction--so that regional traffic doesn't need to pass through the Sellwood District; and the existing bridge can be kept as a pedestrian/bike facility.

    @Cameloparalis: Private car average speed is generally far lower for urban trips than most motorists think. I live in Beaverton, and its 11 miles between my house and Pioneer Square. Google Maps estimate it takes 20 minutes to drive; my experience is that unless its late at night, it takes at least 25 minutes. Unless you and your destination are next to a highway, driving isn't as fast as people thing.

  4. Regarding the LO extension: time gained from changing the alignment from 10th/11th to 5th/6th would be significant but minor. Most riders will find themselves with a new extra transfer at the PSU stop, but mall transit vehicles would be coming frequently enough so that the time penalty should be held to around four minutes.

    According to TriMet's Spring 2010 passenger census, a number of northbound route 35 riders continue past the mall. About three dozen use the Interstate/Rose Quarter stop which generally offers more efficient connections because buses do better than LRT in getting through downtown.

    The main time penalties with the proposed LO extension are because of the inherent slowness of streetcars relative to buses, forced transfers, and longer walking distances at the south end of trips.

    While the vehicle design speed limit for Skodas might be 75 kph (a little less than 47 mph), the maximum operating speed for the local system is much lower - I believe 35 mph. If PSI wanted faster speeds, we'd have to spring for heavier duty roadbeds and use catenary instead of suspended overhead wire.

    Another factor which is not normally considered by prospective riders is that streetcars cost more to run than buses. We can expect the annual operating cost of Highway 43 transit to jump by about $2.5 to $3 million at least when the extension comes on line without getting a compensating increase in ridership for many years. TriMet will have no choice but to offer less service elsewhere in the system or to raise fares --- again.

  5. I was getting it from the Oregonian article a day or so ago about the Sellwood Bridge which said "The deck must accommodate a future streetcar." The way I interpreted it was building the bridge to streetcar weight requirements and lane width requirements and not necessarily track installation. I sure hope I interpreted 'streetcar ready' incorrectly.

    I dont have a problem with the regional traffic going through Sellwood, I think it is good for the retailers. The problem I see is that traffic engineers give so much priority to through traffic and keeping it uninterrupted and fast that they ruin the street turning it into a 'traffic sewer' and remove the on-street parking. Slow, calm through traffic that treats Tacoma St as a Main street (with on-street parking) is the way to go IMO.

    As a frequent streetcar rider, yeah its slow, but speaking for myself, the slow speed doesnt deter me from using it.

  6. @RA:

    Streetcars are only inherently slower than busses in mixed traffic, where the inability of the train to maneuver around obstacles becomes an issue. Give the vehicles an exclusive ROW, and this ceases to be an issue--the faster acceleration of electric railcars may give rail an advantage. (Of course, electric trolleybusses have better acceleration than electric rail, but Portland doesn't have any of those).

    The main issues are: a) the transfer, and b) The optional nature of many bus stops. While #35 has lots of scheduled stops between LO and PSU, many of them aren't made due to nobody waiting or pulling the cord. If the bus actually had to stop twenty times en route, no way would it get between LO and downtown in twenty-odd minutes. The streetcar has lower stop density, but stops at all stations. (To put it another way, the #35 provides good service to commuters from OC and West Linn precisely because few people north of Lake Oswego use it; thus it acts like an express bus between LO and downtown--it's trip time is about the same as that of a private auto).

    Transfers heading inbound are probably not going to be an issue, as both at LO and at PSU one is transferring onto a more frequent service. I'd be more concerned about the reverse direction, where you're transferring onto LESS frequent service; the average wait time depends on the headway of what you are transferring onto, not what you just came off of.

    TriMet thinks, in the planning docs, that the Streetcar will save money on operating costs. Whether their analysis is reasonable or not, it's hard to say: an interesting question is whether or not TriMet should do things which attract new riders, but add expense. The organization considers it an important part of its mission to increase ridership--whether this is a worthwhile endeavor (in sort of environmental and sustainability goals) or a waste of money (and perhaps evidence of empire-building or corruption) is an interesting question. (And sounds like a good topic for a future post).


    An issue with the Sellwood is that its actual likely users, its design, and its funding are all in disagreement. It's designed like a local neighborhood collector (including its streetcar service, which is not regionally important given that Milwaukie MAX will be nearby); it's widely used by Clackamas County commuters, and Multnomah County taxpayers will be footing the lion's share of the bill.

  7. What I meant by "inherent slowness" of streetcars was primarily the 35 MPH maximum operating speed. In the LO case, buses can and do cruise 43 south of the bridge at speeds in the 40's. Buses also have routes where they operate at freeway speeds. Think of CTRAN and the 96.

    The LO project's "no-build" option includes TriMet's long standing plans to upgrade the 35 to Frequent Service. It's 3rd on the TIP list for new conversions. While the agency's current problems forced it to seriously downgrade FS's definition, it would still include at least study of stop consolidation. The official so-called "Enhanced Bus" option includes elimination of half the stops north of LO and - inscrutably - diversion of the bus to 10th and 11th.

    I believe that the degradation of the northbound trips is significant. We're facing about 10 minutes total just for new forced transfers if TriMet holds to 12 minute streetcar headway. We agree that southbound transfers will require more time, especially in the late evening or on weekends.

    Actually, streetcars only stop on request, making their operation significantly different from MAX. They tend to stop at most stops anyway, particularly in the FRZ. Interestingly, there isn't much schedule deviation between peak and quiet hours.

    The project team's conclusion that streetcar will have lower operating costs on a per rider basis is based on an extraordinarily rosy(??? apocrophyllic???) prediction of incredible ridership growth in the corridor far in excess of TriMet's service area as a whole. This is absolutely unbelievable given the demographic and geographic realities of West Linn and Lake Oswego. Those predictions are necessary if TriMet and Metro are to get all that free federal money. The basic fact is that streetcars cost more to operate than buses and will remain so until $300 per barrel crude.


Keep it clean, please