Infrastructure Light Rail Seattle

Light Rail Along Road Rights-of-Way: a Cheap Solution to an Expensive Proposition

» Seattle Mayor Mike McGinn proposes to build a new transit line to West Seattle and Ballard along the street.

The dividing line between what Americans reference as a streetcar and what they call light rail is not nearly as defined as one might assume considering the frequent use of the two terminologies in opposition. According to popular understanding, streetcars share their rights-of-way with automobiles and light rail has its own, reserved right-of-way.

But the truth is that the two modes use very similar vehicles and their corridors frequently fall somewhere between the respective stereotypes of each technology. Even the prototypical U.S. light rail project — the Portland MAX — includes significant track segments downtown in which its corridor is hardly separated from that of the automobiles nearby. And that city’s similarly pioneering streetcar includes several segments completely separated from the street.

In that context, it’s worth considering Seattle Mayor Mike McGinn’s recent argument for a scaled-down transit project that would extend from his city’s downtown to West Seattle and Ballard, a proposal he hopes to get before voters within two years. Unlike Seattle’s Central Link rail line, which opened in 2009, this new rail program would operate in street rights-of-way adjacent to moving automobiles; it would not include the expensive tunnels and viaducts that make Central Link a “mini-metro.”

As a result, some have labeled this plan little more than a streetcar, whose slow pace and minimal capacity make it more useful as a development tool than a transportation one. Others are convinced that the project will morph into a multi-billion dollar mini-metro like Link, a high-cost concept into whose face city budget experts are afraid to look.

But Mayor McGinn’s proposal is neither of those things — it’s an effort to build a cost-effective rail transit line on the model used by cities across Europe, known typically as tramways.

What makes Mr. McGinn’s plan — which, by the way, remains in the very early development stages — so different from those proposed by most cities is that it attempts directly to reduce significant road capacity for automobiles and replace it with space reserved for transit. Most light rail programs avoid that prospect by using existing rail rights-of-way for new lines, or by sending trains underground or above it. That’s because it’s considered treacherous to threaten to remove space now used by automobiles.

Indeed, if he goes forward with the proposal, the Mayor would be doing something that flies in the face of political expediency, since transit-friendly or not, most Seattleites continue to commute by private car. Yet Mr. McGinn claims he’s unworried about the implications of doing so; considering he won last year’s election partially by stridently opposing the construction of a $4 billion road tunnel under downtown, he probably should be taken at face value.

Mr. McGinn’s proposal is a reasonable one: by simply removing vehicle lanes and reserving space on the road for trains, you can build relatively fast light rail systems at the cost of streetcar lines. Other than over major physical barriers (the roughly 15-mile route suggested would require crossing two waterways), there’s little need to move earth or build new structures, saving tremendous amounts of money.

Unlike Central Link, which achieves very high average speeds compared to most urban rail systems, this project would feature only moderate speed improvements over existing bus services. But it would see a very large ramp-up in capacity and time savings over automobiles if intersections are properly designed. And it would encourage more people to ride transit because of clear station stops, frequent services, and comfortable trains.

All this at a much more reasonable price than would be possible if you wanted the type of full-scale, independent right-of-way featured by Link. Unlike equally cheap streetcars, these tram lines wouldn’t held up by surrounding traffic or required to have short trainsets because of limited street dimensions.

As I’ve mentioned before, it’s been relatively easy to implement such street-running rail in a number of European cities, and where it’s been done, it has often improved the quality of the surrounding streetscape, producing exactly the type of livable environment planners love to see around major new transit investments.

None of this is to suggest that there aren’t places where transit corridors requiring full separation of rights-of-way are advisable; building New York’s Second Avenue Subway as a street-running light rail line would be a disaster, simply because it wouldn’t be able to handle anywhere near the capacity required. But in Seattle, land of moderate densities and medium-height commercial corridors, what Mayor McGinn is suggesting is exactly the right investment to make — at the right price.

Image above: Tram in Amsterdam, from Flickr user martin_vmorris (cc)

42 replies on “Light Rail Along Road Rights-of-Way: a Cheap Solution to an Expensive Proposition”

I’d argue the more common deployment of LRT in the states is as exactly this – running in the street in its own lane (Portland, Houston, downtown sections of Dallas, etc). Not sure how the mini-metro got declared more common; the only place that seems to have gone that way is Seattle, and don’t even they run in the street part of the way?

M1EK –

You’re exactly right, I didn’t mean to imply that mini-metro what’s actually done, but rather that most streetcar and light rail systems fall somewhere in between the two stereotypical definitions. Sorry about the confusion…

The broader point, though, is to suggest that street-running LRT can work very well, and that we shouldn’t be afraid of advocating for it!

All of this re-enforces the fact that technology does not make people switch to transit – it’s about making transit better the car in some way. What’s proposed here will allow transit to be quicker than the car (and possibly more comfortable) – that’s what will make it a winner.

I’m a strong critic of making automobile congestion worse so that rail transit appears better. That’s a lot of money to tread water. Doing so also tends to make the connecting bus service slower where applicable.

If you want to make people switch to transit, you should make driving and parking more expensive. This way, the government can collect money for transportation spending and limit pollution from congestion. This is much better than no money and more pollution.

Why do you conclude that it will increase congestion? There’s nothing in the proposal that suggests that it will. Obviously taking cars off the road faster than you take lanes away from cars will lead to less congestion.

As a bike commuter myself, I am a strong critic of what are called “bikelanes” due to the fact that drivers simply don’t respect them, even otherwise good drivers, this is primarily due to the minimal effort that goes into separating these lanes from car traffic. If lightrail where to replace a roadway however it would be a different story, since light-rail would run on a fixed guideway, bikers riding adjacent to the line would be safer. in this instance a bike lane would make sense, thou I still see great safety advantages in the lanes being separated somehow (either elevated as in the case of a sidewalk, or distanced a few feet from the LRT line, or maybe both), since a biker is also at risk due to his own mistakes.

BruceMcF–I’m not concluding that this Seattle project will. I have not studied this project at all. But street running LRT on arterial and heavily trafficked roads will likely increase congestion. Mathematically, rail will increase capacity on the road. But in practical terms the proportion of drivers transferred to transit compared to the proportion of roadway capacity reduced. Now, a road at 30% capacity won’t suffer. But one at 70% probably will if a lane is eliminated.

Nate, that’s not how it works. Take a road with two lanes each way, at 70% capacity. Devote one lane each way to rail.

If you’ve designed this at *all* competently, you get maybe 20% of the traffic switching to rail and 20% of the traffic *simply vanishing* (due to the flip-side of induced demand). Doing the computations you end up with a road at something like 84% capacity, which really isn’t a serious penalty. And those are pessimistic estimates for traffic change.

Now if you do something stupid like removing lanes from a *through route* and building a light rail line which *doesn’t go far enough to replace the trips*, you’ll get bad results. But here, Ballard is basically off in a corner and the light rail line will run far enough to be a sensible alternative for practically all the drivers. So then.

I’m talking about service on through routes, not redundant or lower-traffic parallel routes.

If traffic simply vanishes then there is a loss of productivity and economic output. Most public projects of this sort are built with the intent to do the opposite. Otherwise the traffic goes somewhere. And some in corridors, there is nowhere else for it to go.

Also, your capacity assumption is a bit simplified because one lane of through traffic is much different than 2 because 1 lane is subject to accidents shutting down the road. With only one lane, it also becomes the right turn lane whcih has lower capacity than through lanes. Parallel parking contributes to this as well. But again, I’m talking about arterials. Most street LRT tend to be on arterials when not in downtown, though. Of course, maybe Seattle is an exception.

Most of us Seattleites are experienced travelers who have seen and riden rapid underground metros in Europe, China and Japan. This is what Seattle needs, not a streetcar. Money is NOT an issue with me, it is HOW it is built.

And FYI, the New York Subway (for ie: New York’s Second Avenue Subway) was dug underground even when there were parts of undeveloped NY blocks. NY planned ahead. Seattle needs to too.

I know it sounds unbelievable. But what if Seattle undergoes a major boom where millions are moving in every decade. PLAN AHEAD!


I would totally support that in theory, and money shouldn’t a be problem, considering how much the state pays for roads. But the state has regulation that is tying us up. We should start here, and improve transit bit by bit. Thats what even London and New York did.

Street running rail certainly has an appropriate place in circumstances with shorter distance alignments as a potential improvement to heavily patronized bus routes.

But what is the objective? to build a regional rail service or more local? If it’s the former, street rail will be a major shortcoming due to slow speed and capacity constraints.

Here, you lapse into exactly the kind of thinking in stereotypes rather than actual transit designs that Yonah talks about in the post.

It says in the article that the objective is a 15 mile tramway: for a 15 mile corridor, a tram makes fine speed, and if there is need for more capacity, then just increase frequency.

If the objective was to extend it further, a modern tram-train can run up to 62mph and still provide direct service to the main destinations along the tramway corridor.

> for a 15 mile corridor, a tram makes fine speed

15 miles seems a long way for a streetcar to be fighting traffic on shared right-of-way, esp if there are frequent stops.

A streetcar providing fill-in transportation in West or downtown Seattle seems like a fine idea. Not sure it’s so good for connecting them?

Bruce, I’m not sure what I’m stereotyping. A 15 mile long operating with fully exclusive lanes in the middle of an arterial is not likely to make good speed. Indeed, it will be faster than shared-lane streetcar type service. But if there is a continuous urban fabric with cross streets almost every block and a 30 mph speed limit, the service is not likely to average more than 15 or 16 mph with typical stop spacing. Fifteen mph is too slow to be competitive with automobile travel for the length of the line.

Capacity would be limited because headways would likely be limited to 5 minutes–and perhaps less frequently due to cross street traffic constraints. On street rail tends to also have short length platforms because of spatial limitations as well.

FWIW, the all-surface proposal for a project in my hometown was estimated to be about $70 million/mile with an optimistic (by my analysis) operating speed of 15 mph. So surface LRT ain’t necessarily a good value.

“A 15 mile long operating with fully exclusive lanes in the middle of an arterial is not likely to make good speed. ”

With proper signal priority (which Seattle has proved able to get right, unlike Toronto), yes, it is likely to make pretty good speed.

And on a route this crowded with auto traffic, where the *cars* are stopped every couple of blocks for signals, 15 mph actually is competitive with cars. Think about it.

The term “15 Mile” corridor is misleading. The proposed line line would run from West Seattle in the southwest quadrant of the city to Downtown and then continue north to Ballard in the northwest quadrant of the city. Thus the proposed alignment really serves two corridors: West Seattle to Downtown and Ballard to Downtown.

According to Google Maps West Seattle-Downtown segment is roughly 5.2 miles and the Ballard-Downtown segment is about 5-6 miles depending on the route. Note these numbers are super screwy because the final alignment hasn’t been selected. I picked endpoints at major mixed use hubs, but the city may extend the line further (hence the 15 mile figure cited above).

Regardless, of the final distance of the line, the overall point remains the same. Two 5-7 mile corridors serving opposite corners of the City to downtown would make totally reasonable commutes under the proposed technology. The express bus from both of these areas currently takes about 20 minutes. Assuming rail could shave 5-7 minutes off the commute, the time savings alone would make it would be a huge improvement over the bus. Of course you have all of the other benefits such as increased headways, more comfortable ride, system legibility that make fixed rail transit so appealing.

Moreover, the route speed improves even more when one looks at the geography and land-use patterns of the proposed alignments. The West Seattle segment has a major water crossing and then travels through an industrial district (long blocks and little cross traffic). The Ballard segment would most likely run along 15th Ave NW (industrial district, geographically constrained, little cross traffic) OR Westlake (waterfront, no cross streets). These factors would reduce choke points increasing the overall speed and reliability of the system.

While we’re getting nitpicky, the West Seattle segment will also likely have to cross major freight rail track serving the Port of Seattle so that part would have to be grade separated too resulting in further speed efficiencies.

Are you referring to MLK Way on the Central Link LRT? LRT is more competitive there than in my example because I’ve heard the speed limit is 35 mph and the station spacing is more like one mile or so.

Certainly the system is more competitive with cars after the line is built (at least at rush hour), but I’m not sure it would competitive compared to MLK prior to construction–which was the treading water point I made earlier.

True… Europe has invested in short segments of tranways and streetcar-styled lightrail. But ALL European cities ALSO have vast, underground subway/metro lines.

Ask yourself: What is more important to the city? Money… or Convenience/Quality? If you want a convenient city-wide rapid rail system, then you’ve got to build most of it underground, or at least keep the trains as far away from the surface-streets as possible.

Uselly I rate a light rail project as a street car if there were streetcar tracks embeded in the street from a eary 1920’s streetcar line that got torn up. Such as if they put a light rail line along a street that I knew had streetcar line in it becomes a simply a restored streetcar line.

They had a interview with a man who was 90 years old who was one of the last Trolley drivers in Norfolk they let him vist the new Norfolk light rail line. Well he said that it looked like a modern computerized version of the streetcar that he used to drive in the 1940’s. The humor to about it is that Norfolk’s light rail line runs along several sections of Norfolk’s old 1930’s streetcar system.

If light rail where to run in west seattle it would likely run along California Ave SE, this being the best street in terms of population density and width, yet even this street only opens up to four lanes around the intersection of SW Alaska Street. because of this I believe the line would be best put underground, though this option is much more costly it is likely the best one considering street width as well as West Seattle’s hilly topography. South of Fontleroy Way SW I believe the underground line should make a SE turn ending near White Center, a section of seattle badly in need of development.

If there’s no appropriate four-lane ROW (and no set of fully closeable ‘alleys’) then indeed it becomes time to consider underground or elevated. (People really hate losing local access entirely.)

So, “what Mayor McGinn is suggesting is exactly the right investment to make — at the right price”. Well, don’t stop there, Yonah, tell us what hs is proposing and what the price is- because McGinn certainly hasn’t.

Or maybe you’re endorsing his rather vague “we will build it right and the price will be cheap”. Why sure, I’ll go for that!

That is, once I’ve learned the details and seen them vetted by a few outside experts.

Its really weird how hard it is for people to understand there is no plan, especially when we consider that McGinn himself has said he has no plan, and his spokesman Matassa has said they’re not even discussing it.

And Yonah has just shown us how the legend of McGinn grows, while McGinn himself maintains perfect deniability. Whatever happens, it won’t be his fault, and he never promised you anything different.

Thanks to Serial Cat Owner for reminding us that there is, as of yet, no plan from this brand new mayor. At this point it is all pie in the sky. there are 2 water crossings to contend with (both navigable requiring high bridges or opening bridges), strange topography, uncertain routing through downtown, whether to serve the Seattle Center, and so much else. It is doubtful that we are much advanced beyond the magic marker on a map stage here as of yet. Vote “YES” in 2011? Not a chance.

No, as I noted in the post, the project remains in the conceptual stage. We should see more information about where it might go over the next few months.

But you can imagine generally where it might go — from downtown, north through Queen Anne, and then over into Ballard, probably along 15 Ave NW; and from downtown, south and then west over Harbor Island, and into West Seattle along California Ave.

The problem with running the line along 15 Ave NW is that the line would go directly north into Ballard without ever stopping in Fremont. As for the north part of this proposal, I believe a line would be at its best if it started either at the future UW LRT station or at the Brooklyn LRT Station, then headed west along the north side of Ship Canal all the way into Ballard. Unfortunately both proposals would most likely mean skipping Wallingford, another important neighborhood in terms of population. So while the 15 Ave NW corridor might be the easiest to build I believe it would give less bang for the buck than alternative proposals.

I don’t think Yonah understands the topography of Seattle. Queen Anne, for example, is a hill ascended by a counterbalance in the days of traction, and still avoided by some drivers today because of the steepness. It’s one heck of a bus ride, though!

A line from West Seattle to Ballard (for which no need has ever been revealed) has to pass two navigable waterways, i.e., requires two high-level crossings. The McGinn alternative, obviously, is to use the Ballard Bridge (or the Fremont Bridge) for one of these crossings, putting a bottleneck in the system.

It’s possible that McGinn, being a new-comer to Seattle, is as confused as Yonah about some of what happens in Seattle. If you haven’t spent time trying to get around in town, it’s easy to underrate the volume or nature of traffic on some streets, or the number of geographic problems and bottlenecks to overcome.

No offense serial catowner, but I’m not “confused” by the topography of Seattle, and neither is Mike McGinn. There are several alignments that travel through (or near) Queen Anne and which avoid the steep hills you’re discussing, such as along 15 Ave NW or along Westlake Ave.

I’ve made clear here that there is not yet a specific alignment, but that doesn’t mean there aren’t potential options.

Second, the point of the line isn’t to connect West Seattle to Ballard. It’s to connect West Seattle to downtown and Ballard to downtown; there is certainly a need for both of those routes. It just so happens that you can do so by through-running trains from Ballard to West Seattle.

There are certainly geographic boundaries and potential bottlenecks, but there’s nothing about them that are insurmountable.

Does anyone have any daily traffic counts for California Ave and 15 St NW? Knowing that might give a better clue off the bat as to whether a fully-exclusive lane LRT service could be implemented without effects I described above.

Looking at California Ave from Google Maps, it appears the road has 2 lanes of through traffic, 2 permanent parking lanes and an alternating turn lane. It looks to be about 50 feet wide. Unless one is willing to lose all parking on that street, I don’t see how a fully exclusive lane service could operate. Additionally, it appears that California is one of the few north-south streets that isn’t somehow cutoff. So traffic might be forced on to it at some point regardless of attempts to find an alternative route.

Given the Seattle area’s large size, moderately-high density and geographic bottlenecks, I think they’re beyond using street running LRT for regional/rapid transit. It’s not like Phoenix.

My points were-

a) going around the base of Queen Anne hill is not going “through” Queen Anne. Going out Elliot and up 15th misses a huge segment of Queen Anne and Westlake misses even more. That’s fine as long as you’re under no illusions that you’re going “through” Queen Anne.

b) these are actually two different lines we’re talking about (West Seattle-City Center) and Ballard. But for starters, where should the Ballard Line go? Maybe to the University. It makes sense to talk about buying the same cars and overhead for both lines, but each will be better if it is considered independently of the other.

A majority of the density on and around Queen Anne Hill is right at the Southern base of the Hill in the Uptown area. You certainly can swing through the Uptown/Seattle Center area before heading out to Elliot and 15th.

True the top of the Hill does have an urban village but it in no way has the density to justify the expense of an underground light rail line to serve it.

Serving Ballard via the Elliot/15th corridor is a plan that has been talked about off and on for at least 45 years. This corridor was part of the Forward Thrust plan, the monorail plan, the City ICT studies, etc.

The travel demand between Ballard and downtown is much greater than an E/W corridor to the University. So connecting to downtown first makes sense. Past that an E/W line to the UW through Wallingford makes a certain amount of sense as does continuing up 15th to Crown Hill then swinging East to connect to Greenwood, Aurora, Northgate, and Lake City (see the city ICT study). The route to Northgate has the advantage of wider street ROW for most of the route and a more favorable topography.

There is a place in Seattle’s transportation system for streetcars, and a place for Link-style light metros. This is a very high-ridership corridor with Ballard and West Seattle rapidly densfying and places between Downtown Seattle and these places showing promise of densifying if light rail is built through there. Also, there is not doubt that in the future this route will be extended beyond Ballard and West Seattle farther out into the corners of the city and into the suburbs. We will need the capacity that grade-separated rail can give us (light rail on the street except in isolated instances is limited to 2-car trains and lower frequencies) and the people of Ballard and West Seattle deserve a fast link to Downtown and other neighborhoods. The most common complaint I have heard about Central Link is the slow speed of the segment along MLK; while on the elevated and underground segments it can go 55mph, Link can only go about 30 on MLK. In Seattle we don’t need a Portland MAX-style light rail system; we need a NYC Subway-style system.

If the entire route is to be street-median running, with no high-speed segments, why not use streetcar-type vehicles? Saves a lot of money to build (in many cases, no need to rebuild roadbeds or relocate utilities; which you generally have to do to support the weight of heavier LRT trains).

Taking auto lanes out of service for LRT is not that unusual: As noted, much of MAX was constructed that way–the downtown segments, the Burnside Street segment, the Intertate Avenue segment, are all examples where auto lanes were removed for trains.

What part of “grade separated” and “the cheapskate always pays the most” don’t people understand?
The one advantage Seattle has in being 25-40 years behind many other cities is that we can learn from their experiences: so let’s PLEASE tax ourselves enough to spend the requisite amount of money to build it right, from the start – grade separated. But let’s also be clear that the other in-city routing needs must be considered and a prioritised list created before we jump in and go to WS and Ballard just because we think it’d be cool. The routing through Downtown must be decided first and foremost.

In LA I like to drive next to the Blue Line and the East LA Gold Line for fun and see if I get hit and die. I’m not dead yet!

Unless they build another bridge over the Duwamish River for the light rail they are going to have a terrible bottleneck on the West Seattle bridge if they take out any traffic lanes to devote to rail

“And that city’s similarly pioneering streetcar includes several segments completely separated from the street.”

Portland’s MAX is not at all pioneering. It is referred to as the ‘Karlsruhe Model’ – using trams on both heavy-rail ROW’s and on inter-city, at grade streets. This sort of transportation integration is not new to America, either. I’m kind of disappointed that you forgot to mention this little tidbit. Otherwise, I really enjoy your blogs.

The entirely grade seperated Canada Line (adjusted for miles and USD) is cheaper per mile than the proposed at-grade Portland-Milwaukie MAX line which is almost all at grade. I think we have come to the point where after utility relocation, building small inefficient couple block long segments at a time, full street width recontructions, working around business concerns and holidays, traffic remediation, etc that the costs for at-grade rail lines start to become comparible to grade seperated lines. I believe bored tunnel through a hillside is the cheapest rail line per mile to build.

Yonah, another example of Seattle proposing to take away auto lanes is on Broadway on First & Capitol Hills…

@poncho–The Canada Line is a poor example. You neglect to mention that the insistence on PPP for the line resulted in a consortium using heavy-rail subway cars from Rotem, incompatible with Bombardier-built SkyTrain cars (with linear induction propulsion). Also, the stations on the Canada Line were built with short platforms to save money. The line is way too close to reaching design capacity, and reconstruction of stations to allow 4-car trains has been estimated as costing at least 40% of original construction cost. Also, the difficulty of reclaiming that lost capacity will likely force construction of either streetcar or LRT in the Arbutus corridor, and sooner than anticipated, so the total infrastructure cost for moving the same design capacity will be significantly more than the MAX line in question.

This Seattle proposal–is this intended to be streetcar, permanently distinct from LINK? Or is it intended to be an introductory service, intended to be upgraded incrementally to light rail? Large parts of Seattle are built on clay, and a line built for a single articulated streetcar won’t have the load-bearing capacity to carry a 4-car light rail train. Factors like curve radius, grade, and track spacing would have to be considered. An at-grade line with light rail geometry can be designed to accommodate grade separation in the future, but the provision doesn’t mean it can’t start as a streetcar.

Leave a Reply