Can Rapid Transit Work Along Freeways?

Lately there’s been a groundswell of talk among transit advocates about the need to stop building rapid transit along freeways.  To honor the imminent opening of Portland’s Green Line, which runs mostly along the east side of that city’s I-205 freeway, I thought I’d weigh in on this a little, with some relevant pictures fresh from Berlin.

First the pictures.  Here’s a not-too-unusual example of rapid transit combined with a freeway, from the upscale Charlottenberg area of western Berlin.  Let’s start with a Google Earth shot (click to enlarge).

Charlottenberg freewayThe north-south freeway (Highway 100) is part of the innermost freeway ring around Berlin, very comparable to the Périphérique freeway that encircles Paris.  The freeway is six lanes, three in each direction.  An important S-Bahn rail rapid transit line runs down the middle of the freeway.  The line is also a ring all the way around the city, linking many of its major suburban nodes.

Charlottenberg freeway zoomin Zoom in.  Not counting the frontage road on the left, the right-of-way is about 70 metres (235 feet) wide.  (An entire Portland block, for example, would fit in this space.)  It’s so wide in part because in addition to six freeway lanes and an S-Bahn station, this segment has a separate set of tracks that serve intercity travel.  But the point is, there’s going to be a bit of a walk from a median rail station to any kind of development.

But if I stand at the station and look north, it looks like this.

It looks as if this wide strip has always been a transport right-of-way, as historic building faces clearly orient to it just as they would to a boulevard or river.  You may not want to live facing a freeway, but some people don’t mind, or find it worthwhile in return for living on a rail line.  And as soon as you step away from the open space in either direction, you are in a dense, well-maintained, relatively upscale part of the city.

Even European cities have freeways.  They usually ring the core city and extend outward from it, rather than penetrating the center as they do in North America and Australasia.  But they’re there, and important destinations develop along them.  In Europe, rail lines were often there first and freeways added later, while many American and Australian cities are now doing the opposite.  But in either case, the result is a very wide right of way with both a freeway and a rail line, and the land use opportunities and limitations that result.  The limitations include a longer-than-ideal walk between the rapid transit station and the surrounding development.  But there are also opportunities; in the New World these include the opportunity to get new transit close to development that has already occurred around the freeway, serving that development as it is while trying to catalyze it to become denser and less exclusively car-dependent.

When a New World city builds a rail line alongside a freeway, or even in its median, nobody expects the result to look like Berlin any time soon.  But it’s not obvious that these can’t be dense transit-oriented districts in their own style a few decades from now, if that’s what local government wants.  Yes, it’s always better to have the transit stop right inside the dense commercial core of a neighborhood.  But if we make that an absolute requirement in our thinking, we end up having to choose between either

  1. underground lines in places where we can’t afford undergrounding, or
  2. elevated lines in places where community impacts just won’t allow it, or
  3. lines at grade on arterial streets, even if in exclusive lanes, that can’t deliver a reasonable travel time given the long regional trips the line is intended to serve.

On most of the American rapid transit projects that I’ve been close to over the last two decades, any of those outcomes would have killed the project.

In the real landscapes of New World cities, a “no rapid transit along freeways rule” would have three really bad consequences:

  • It would mean dramatically slower rapid transit, because the freeway is where you are most likely to find continuous grade-separated space, which is where fastest operation is possible.  This is important when our main short-term purpose is to link key regional centers to each other and thus support further growth in those centers.
  • It would eliminate the cheapest alignment option for serving the existing demand in a corridor, effectively killing many lines that could do a great deal of good even if they didn’t fit the urban-design ideal.
  • It would essentially abandon the prospect of quality transit to freeway corridors, even where those corridors have already attracted significant commercial development and have even catalyzed large regional centers.

This last point is especially important.  It’s really not helpful to imagine a future divided between happy sustainable urban fabric and benighted freeway-oriented slums.  The media will always find it easiest to explain new stories, such as the urbanist agenda, by reducing them to old stories, such as class warfare.  So we will always be hearing that the sustainability agenda is elitist, especially so long as it can’t propose answers for redeeming the freeway landscape.   So much has been invested in freeway-oriented fabric that most cities can’t abandon those areas — either politically or economically.

I’m not saying that transit on freeways is always wonderful or that it doesn’t have disadvantages.  In corridors where there are freeway and non-freeway options, a lively debate is in order.  But I am saying that a blanket slogan of “no more rapid transit along freeways” would mean a lot less rapid transit, period.
On a positive note: you can find freeway+transit development models that look pretty dense and prosperous, even if you wouldn’t want to live there yourself.  Sooner or later we’ll figure out how to build this in the New World.  We’ll have to, because we’ll always have freeways and they’re often in places where rapid transit makes sense.

57 Responses to Can Rapid Transit Work Along Freeways?

  1. Cap'n Transit September 12, 2009 at 3:56 pm #

    A couple examples from Westchester County, NY:
    Fleetwood, Mount Vernon
    In both cases, the railroads pre-dated the highways.

  2. Cap'n Transit September 12, 2009 at 7:13 pm #

    Also: the CTA Congress Line UIC/Halsted station.

  3. Matthew Pennington September 12, 2009 at 8:11 pm #

    Not just the UIC/Halsted Station, but most of the Blue line outside of Downtown Chicago, as well as the Red line. Both lines were extended using the freeway right of ways.

  4. EngineerScotty September 12, 2009 at 8:50 pm #

    In Portland, there was originally a freight railroad through a shallow ravine called Sullivan’s Gulch.
    Said ravine was later widened to make way for the Banfield Freeway (US30, later I-84).
    And in between the UPRR tracks and the freeway, is MAX.
    If anything the Green Line along I-205 has an advantage over the Banfield segment, as it does border on the liveable areas. In the case of the Banfield, stations are connected via stairs and ramps to overpasses, generally–but there’s no direct access to the surrounding neighborhood except up.

  5. Wad September 12, 2009 at 8:56 pm #

    Here’s the sordid history of the Los Angeles Green Line, the ugliest duckling of all light rail lines.
    First, the Green Line was conceived out of a consent decree brought on by the communities where the Century (105) Freeway was slated to be built.
    The Green Line was a transit mitigation measure in order to get the freeway built. It was never intended to solve any transportation problem.
    Second, the Green Line is a light rail line that was built by highway engineers. It was built for speed, but bypasses key transit demand areas so it can have a relatively high cruising speed.
    (I won’t even get into the politics of why the Green Line missed LAX, as that requires a long-form entry on its own. The short version is that LAX management fought to keep the train away.)
    What else besides the airport does the Green Line miss? Western Avenue, where there is a community college next to the freeway. This is 5,000-10,000 boardings lost. It misses Atlantic Boulevard, a key north-south arterial with the busiest north-south bus line east of downtown L.A. It also provides no access for Paramount residents, who must ride buses or park & ride at Lynwood and Downey to catch the train. These three stations would have added another 15,000 boardings to the Green Line.
    That’s not all. There’s a “last mile” problem at both ends of the line. The eastern terminal in Norwalk ends at the junction of another freeway. Otherwise, the transit center/park & ride is desolate. You need a bus to get out of there. Meanwhile, one mile east of the station is the Norwalk city and L.A. County Civic Center. This is where a courthouse and the L.A. County records are kept. There is also a major shopping center across the street. This is a major transportation destination. And a half-mile east of the Civic Center is the Norwalk/Santa Fe Springs Metrolink station, which would have been a transfer point for passengers coming from or going to Orange County.
    At the other end is a desolate industrial armpit of northern Redondo Beach. It still has some aerospace and light industry, but is a ghost town outside of rush hour. Had there been money to extend the line one mile south, the train could have reached the South Bay Galleria, which besides being a shopping destination also has a productive transit center.
    Can it get worse? Yes. It did. This consent decree had been fashioned in the late 1970s and construction on the freeway began in the mid-1980s. Then Southern California went through its worst economic crisis since the Great Depression. The government stopped supporting the aerospace sector, a grim capstone on Southern California’s deindustrialization. The train was on track to go where jobs had all but vanished.
    There should have been no reason for anyone to use the train.
    Yet today, the Green Line carries about 40,000 boardings, despite everything working against it.
    It’s L.A.’s catastrophic success.
    The Green Line is a teaching tool in not making the same mistakes in its bad design. Yes, 40,000 boardings are impressive, but they came at a huge toll and also have the problem of not being able to channel development. Regardless of ridership, the freeway is the determinant of the most optimal land use around the stations — not the Green Line. People are not going to want to live near a freeway, and the most optimal traffic would be geared around vehicles.

  6. Alexander Craghead September 12, 2009 at 9:46 pm #

    Nice commentary, Jarrett.
    Another issue I think we are dealing with here is that when one mentions “rapid transit” one gets many different visions of what that means. Pro TOD, pro urbanist supporters tend to see rapid transit as a land use tool, envisioning big pretty transit boulevards lines with medium and high rise residential and mixed use that faces those streets. Social supporters are more likely to see systems with the least footprint of all in their existing communities, a sort of hidden rapid transit that hopefully makes no noise and has little or no visibility. Mobility proponents are likely to envision grade separated systems operating at highway speeds.
    Historically, light rail has (in my view) been the most broadly appealing mode to transit modernists, and I believe it is precisely because it carves a middle path between these various points of view. It mixes street running and urbanist forms with quiet, pretty, and sometimes even artistic designs that are more welcome to communities and speeds that are at least comparable with existing bus services and often better.
    As systems (including Portland’s) mature, however, I think it is becoming clear that Using light rail systemwide for all rapid transit is not the best solution. Instead a variety of services can provide better levels of service. Thus we now have three kinds of rail transit, whereas just a decade ago we had only one. As the Portland Streetcar expands, light rail will likely continue to evolve into a more rapid, heavy rail character, and streetscape altering TOD roles will likely devolve to the streetcar.

  7. anonymouse September 12, 2009 at 10:40 pm #

    I’ve actually been to the Norwalk end of the Green Line a few times by bike. You really do need to take a bus there, as it’s very difficult figuring out how to get into the place! I do think there are some plans for infill stations, with Western, Atlantic, and Paramount being the logical choices of location. They’re definitely looking at extensions to LAX (and eventually Marina del Rey and Santa Monica), the South Bay Galleria, and they did a study of extending to Norwalk Metrolink. I think the problem at this point is that there are so many better transit projects closer to the urban core.

  8. Jarrett at September 12, 2009 at 11:55 pm #

    Actually, Wad, it sounds like your story supports my point, which is that the Green Line could have been an effective project, despite its freeway location, if it had been built with the intention of succeeding as transit.

  9. Alon Levy September 13, 2009 at 12:00 am #

    In Tel Aviv, the rail line and the highway were built at the same time. Both are somewhat to the east of downtown, though new commercial development has bypassed downtown in favor of the highway/rail corridor. In principle, the amount of development near the corridor is impressive, including the largest shopping mall in Israel, the military headquarters, and Tel Aviv University. The landscape from the line doesn’t look too different from the picture you show for Berlin.
    On this line, ridership is abysmal. The total ridership of Israel Railways, including both commuter and intercity traffic, is 38 million per year. The problem, I think, is that the main job centers are not oriented toward the rail line/freeway – and when they are, they’re not located conveniently. The military headquarters is located across the highway plus another arterial; there are pedestrian bridges, but they require walking in a labyrinthine mall (see link here).
    Conversely, running elevated may not be such a bad idea, when the road is wide enough. In Singapore, the pre-2003 rapid transit lines are mostly elevated over arterials. The arterials are still hard to cross and development is oriented inward, but it’s much easier to access the stations than the freeway-median stations of Tel Aviv; the roads are wide enough and the structures noise-resistant enough that this doesn’t cause blight.

  10. Ted King September 13, 2009 at 2:58 am #

    If you want a single system that does it all (gold-plating, undergrounding, freeway median, elevated, right-of-way salvage) look no further than BART ( ). This layout serves as an example of how a transit system can be both a kludge and serviceable. The Colma station (the one with the moth-balled platform) is the latest example of infill development (lots of housing) in the last couple of years. Just please let’s not go overboard on further expansions – the adjacent systems (CalTrain [ ] and VTA [ ]) close the loop pretty well.

  11. Cap'n Transit September 13, 2009 at 5:51 am #

    I just want to clarify that the UIC/Halsted station is one where the surrounding development is definitely dense and walkable. I’m not convinced of the others; certainly some of the Dan Ryan stops feel uncomfortable to me as a pedestrian.
    The Robert Taylor Homes might have qualified as dense development, but their tower-in-a-park configuration was pretty anti-pedestrian. How are the replacement developments that way?

  12. Eric September 13, 2009 at 3:16 pm #

    Well said Jarrett. There’s no way we’ll build robust transit networks in North American cities without light rail in freeway medians. Sorry to break anyone’s hopes here, but there’s just no way. At least, not until cities are willing to tax themselves substantially for transit and use eminent domain while they are at it. (Good luck!) Some transit advocates just don’t have a clue about the costs and caveats involved, or of the all important need for existing right-of-way.
    I general, existing underutilized rail right-of-way is your first choice (which is why we built the South Corridor first here in Charlotte) but there’s only one or two of those available in most U.S. cities. After that, you only have the freeways to work with. In the former you have aging warehousing and brownfields to create TOD’s, in the latter you have grayfields (and often greenfields) alongside the existing freeway you can use for TOD (as is the case for the planned Independence Blvd. Silver Line in Charlotte). And, yes, our local government wants these TOD’s to be DENSE to make a serious impact off the bat…which is not that easy pull off in other places where you will have to contend with small parcels along the slow thoroughfare corridors sandwiched between a lot of NIMBY neighborhoods. Those aren’t for light rail in Charlotte. In Charlotte, those are for the streetcar.

  13. Pantheon September 13, 2009 at 6:26 pm #

    I usually agree with what you write in your blog. For example, you really opened my eyes to the limitations of streetcars. But I have serious issues with this.
    First, you are setting up a bit of a “straw man” argument. If someone seriously made the argument that there should be “no more transit along freeways” then even I would disagree with them. There might be some instances where it could be the best option, or at least a reasonable one. But how about the argument that we shouldn’t build ALL rapid transit along freeways? Because it certainly seems like in North America, that is what is happening. The Portland MAX mostly goes along the freeway, as do the lines in San Diego, and the Seattle Eastlink is currently proposed to run in the middle of the freeway.
    Let’s be honest about this. These are not the best transit solutions for the people who ride the trains. They are put in place because they are cheap, plain and simple. The most expensive aspect of building light rail is acquiring rights-of-way, and the freeway eliminates the need for that as it is an existing right-of-way. Now I fully understand that a transit engineer has to work with the money he has – he cannot unilaterally raise taxes to build a transit dream. And to that extent, perhaps rail along freeways is the best we can do with the money we have. But I do not believe that the political will is a constancy. If oil goes to $200, $500, or $1000 a barrel then we will have all the money and political support we want to build the best possible systems.
    In the case of MAX, it is about the cheapest light rail system they could have built, and it doesn’t serve the city very well, if at all. It combines the two worst features of light rail: at-grade in mixed traffic downtown (slow), and following the freeway everywhere else (inconvenient). The net effect of this is that it is useless (or nearly so) as a means of getting around downtown and useless as a means of getting from downtown to any part of the city I would want to go. Its only purpose is as a suburban commuter rail system for downtown workers. I’m not saying suburbanites shouldn’t have transit, but a line from downtown to Clackamas that does not effectively serve the points in-between is a missed opportunity.
    I am also skeptical of the argument that if we build light rail along the freeway, eventually development will occur there to justify it. Certainly some will, but there is no way of predicting how much. Instead of speculating, why don’t we look at an existing case study? The original MAX line to Gresham was built along the freeway in 1986. It has now been running for 23 years, long enough to have the desired positive effects. I don’t know what the area was like two decades ago, but today it is a shady area full of Taco Bell’s and KFC’s, ghetto apartments and down-on-their-luck people. If this is the future of freeway-oriened transit development then count me out.
    I am not an expert on Europe AT ALL, but my gut tells me this is an apples-to-oranges comparison. Perhaps the development in Berlin occurred after the rail line but before the freeway? Also, the Berlin freeway looks like it has fewer lanes of traffic than American freeways.
    Also, remember that the likelihood of new development along the freeway is determined by the attractiveness of living near a station. And that, in turn, is determined by how much of the city is accessible through the network. In this way, transit networks are cumulative: each new line enhances the value of the existing lines, because people on the existing lines can now access the areas along the new line. So in the case of a comprehensive transit network like Berlin, if I live near a station by the freeway, I have access through the transit network to all parts of Berlin – not just downtown but Neighborhood A, Neighborhood B, Neighborhood C, etc. – precisely because SOME parts of the rail network go straight through the desirable neighborhoods, instead of along the freeway. But if I live next to a station by the freeway in a suburban commuter network like Portland, what do I have access to through that network? Downtown and the Clackamas mall. That’s it. I can’t use the network to visit my friends in Neighborhood A or go to a restaurant in Neighborhood B, because the light rail doesn’t go there, it is ALL along the freeway.
    This suggests to me that living at a light rail station along the freeway in the New World will never be as attractive as in Berlin unless and until we build the expensive, politically difficult rail lines that involve acquiring rights-of-way through neighborhoods, knocking down some houses, and raising taxes. There is no way around it, and believing otherwise is, dare I say, a bit naive. And if this option isn’t attractive, then the development will not come, or it will not be the kind of development we want.
    So to summarize, I am not against building light rail along freeways in all cases. But in most cases I think it is not very effective, and that is where all the light rail is being built now. I would rather have a city with one super-premium light rail line that serves real neighborhoods at an elevated grade, rather than a bunch of cheap, mediocre lines along freeways. And I am skeptical that we will have the kind of development you are hoping for along these freeways just because they are served by one of these mediocre lines.

  14. Tabitha September 13, 2009 at 10:04 pm #

    While I appreciate your point that we cannot abandon the spaces around freeways, I agree with Pantheon that the development patterns of Berlin, especially considering that the buildings you are showing look older than the freeway, won’t be the same patterns for Portland.
    I think that this development pattern might be an interesting way to keep housing near transit affordable, so in that sense it does sound like a good thing. But there can be very real health consequences associated with living near dense auto traffic and so even if we do get some serious TOD going down, those people might be more likely to suffer from cancer and so forth. Maybe that’s just something we have to accept, because it’s hard to convince anyone that the long-term health costs are a serious consideration, but gosh-darn-it, we can afford to be a little idealistic in the blogosphere and think about some of the human costs. Also, those empty spaces near transit stops but slightly away from the main street are often not conducive to feeling safe, which may have disproportionate influence on female ridership.
    You make some good points, but I’m not on board.

  15. Stuart Donovan September 14, 2009 at 1:33 am #

    Interesting discussion and good comments.
    Designing transit to avoid highways reflects the latter’s negative impacts on surrounding environs and its tendency to support vehicle dependent development patterns.
    However, as several posts have noted, many key destinations are located adjacent to or in close proximity to highways, which also make for cost-effective transit corridors.
    They key issue would seem to be: do we deploy transit to support travel to these large destinations notwithstanding the otherwise poor transit environment?
    The answer, I suspect, is highly context sensitive. But if we can do so cost-effectively (relative to other potential investments) then I don’t see why we would not.
    Also, the co-location of transit and highway infrastructure might become increasingly important should time-of-use road-pricing become more widely deployed.
    Such corridors are exactly where more accurate road pricing signals might solicit changes in travel behaviour, precisely because a convenient transit option exists.
    I’m certainly onboard the JW bus on this one. Pragmatic outcomes should govern transit investment; I suspect co-locating transport rights-of-ways often makes good sense.

  16. Jarrett at September 14, 2009 at 2:24 am #

    In response to these two comments:
    1. The social-impact objection. I think there are ways to construct freeway-oriented density such that housing is buffered from some of the negative impacts of the freeway. Nobody’s proposing the ghastly freeway-side apartments that got built so much in the 50s and 60s. But the real answer to these social impact concerns is that we appear to live in a market-based economy in which people choose where to live based on how their options and desires interact with their means. I don’t see that changing in the next few decades and as I can assure you from my recent tour of Prague and Eastern Berlin, it’s not like communism did this any better than capitalism does. I’m sure there are people living on freeways who wish they could afford not to. I’m sure there are others who don’t mind, or even like it. And although I can be as idealistic a leftist as anyone, I think we’ll continue to be a society where people get to make those decisions based on both their desires and their financial means.
    2. The “straw man” objection. There has indeed been quite a bit of categorical argument that rail along freeways is always a bad idea. Recent examples include the Transport Politic post I linked in the first paragraph. See also this response to this post at The Overhead Wire:

  17. Jennifer September 14, 2009 at 7:08 am #

    Even when transit lines along freeways result in stations being a little further for walkability, that doesn’t rule out all transit users – what about bikers? I wonder if the future will be divided less by those who drive vs. walk/transit, and more along the lines of those willing to bike to live a bit further away from the transit hubs, and those who only walk.

  18. Damien Goodmon September 14, 2009 at 10:44 am #

    The green line is more of a BART-type line (hybrid commuter and urban rail) than a “tradition” urban rail line. And in the car-centric far-flung suburbs like Lakewood (areas where for the most part we shouldn’t be trying to drastically increase density), freeway ROW lines can be attractive, because in large part, speed and travel time savings are arguably more important and attractive to suburban commuters and most of the riders are arriving at the station by bike or car. When the lines enter the central business district the lines just need to break off the freeway and continue down grade separated tracks either elevated or underground (again, like BART).
    Some American regions have been using light rail in freeway medians when exiting the CBD while building at-grade tracks in the CBD, which depending on the region and alignment causes the line to lose a significant amount in travel time savings. A point that Jarrett is good at pointing out.
    Back to the Green Line (which avoids the CBD completely, but does serve one of LA’s many dispersed commercial areas – LAX/El Segundo), many of the cities on the eastern end of the line are very suburban and filled with mostly single-family homes. Accordingly, Paramount is actually a station the line could and does live without. Atlantic is necessary for the reasons specified, along with the Norwalk Civic Center and Metrolink station. You’re also right about Western, but that station is more towards the west. And I do think it is reasonable to believe if those four stations were built the line would see an additional 10-20K riders a day. Getting closer to LAX (Aviation & Century) might be worth another 5-10K (greatly depends on whether there’s a people mover).
    On the west end of the line in El Segundo/Redondo Beach, blame it mostly on 15 years of poor city planning. Once the aerospace industry left town they should have moved to add some residential uses and encourage basic pedestrian-oriented planning in new developments around the stations. The first phase of Campus at El Segundo (or whatever it’s called) is a textbook example of poor planning on El Segundo’s part. It actually places the offices on the parcel of the large land area that is the farthest away from the Mariposa station. The section also needs some improvements in station placement (i.e. Douglas vs. Rosecrans).
    I think I’ve indirectly made my point about how difficult it is to get changes in land use in suburban areas. The area has many attractive qualities, a major transit line investment, and economic conditions ripe for a change in planning strategy and still they can’t get it right.

  19. Jeffrey Bridgman September 14, 2009 at 10:41 pm #

    Random comment:
    Foreclosure crisis = opportunity to obtain ROW?

  20. Jeffrey Bridgman September 14, 2009 at 10:49 pm #

    I agree with Jennifer. But I’d add, what about bus riders?
    The advantage of the MAX running down the interstate is faster service. Since Portland already has frequent bus service, why not improve and rely on frequent east-west service to connect more walkable neighbors and the MAX? I’m not very familiar with Portland, but I do imagine there are people wanting to travel east-west across the green line who’d also benefit from a transfer to a quick ride downtown.

  21. EthanJ September 14, 2009 at 11:57 pm #

    In the US, it’s probably best to build along the SIDE of the freeway. Instead of putting the rail down the median, engineers should convert the median to freeway lanes and run both tracks in existing lanes on one side of the freeway. That way at least the “railside” of the freeway is truly walkable and transit served. Stations could be constructed away from (between) major interchanges and arterial roads, allowing TOD infill away from car-centric stripmall sprawl.
    Also, I live in Berlin. If you look at the history, 80 years ago those fat ROWs were ALL rail. Older unused tracks were taken out to make room for the autobahns. They built the freeways along the outside of existing rail ROWs, not in-filling a freeway median.

  22. Jarrett at September 15, 2009 at 12:22 am #

    I think many modes of station access are important. Portland’s east-side MAX lines do already operate in the context of a high-frequency grid of local lines. Most stations in this area have a frequent service more or less perpendicular to the light rail alignment.
    Note, however, that light rail frequency has been improving but there has been little investment in new frequency on these connecting lines. Most are still at policy headways of 15 minutes all day.

  23. Jarrett at September 15, 2009 at 12:33 am #

    Yes, I agree that rail on one side of a freeway is better than in the median. Portland, for example, now has a lot of rail along freeways but always on one side.
    While the Berlin example obviously has a different history, I stand by my use of it simply to open some room for North American imaginations about what freeway+transit has to look like, and what the livability outcomes might be.

  24. Wad September 15, 2009 at 4:52 am #

    Yes, Jarrett, it’s a good supportive example.
    I can give you two other freeway-transit mess-ups from the L.A. area. These involve buses, though.
    The El Monte Busway and the Harbor Transitway both have major design flaws, but the former is successful while the latter has been an abysmal failure.
    The problem with the El Monte Busway is no stops between Cal State Los Angeles and El Monte. The intent has been to provide a very fast connection between the San Gabriel Valley and downtown L.A. It does, and the busway is well-used.
    One analyst has said the shared busway/carpool lanes move as many people during rush hour as all the mixed-flow lanes! Half of the ridership is on buses.
    The problem, though, is that by not having stops between CSULA and El Monte, it overburdens parallel local bus lines, which take more than an hour to go between El Monte and downtown L.A. There should have been stops at least at Atlantic Boulevard and Rosemead Boulevard.
    The Harbor Transitway has more stops, but its biggest problem is that it has been an enormous failure. There are about 8 lines on it carrying a combined 4,000 riders along the I-110 Freeway.
    It does have connections at Slauson Avenue, Manchester Avenue and a transfer to the Green Line, but it misses important roads like Vernon, Florence and Century.
    The other problem has been that Metro has always charged extra express step-ups to use the freeway. The poorest transit riders in L.A. County live around the freeway and most won’t pay any extra fare for the time savings.
    Another factor is that the Harbor Transitway stops are in very high-crime communities. The Blue Line serves the same communities, but the line is well-policed and unlike the Transitway, the stations aren’t entombed.
    What would help the Transitway ridership is to treat it like an Orange Line-style busway, where there’s only a local base fare and a police presence would check for fares and provide security.
    One thing that needs to be fixed for any type of freeway-transit alignment: noise. High-speed traffic is painful to the ears. There needs to be some sound barrier in ways that don’t produce an echo chamber and also don’t make the station unsafe.

  25. M1EK September 15, 2009 at 6:15 am #

    The problem with this theory is it simply doesn’t fit the observed facts – such as the fact that the light rail lines built in the last couple of decades with the highest ridership are not actually those built along freeways. You run the risk of ignoring the absolutely essential nature of the walk from the station (rather than a transfer to a shuttle-bus for the last leg of the trip), giving up the ability to serve urban development well, and all in service of the questionable goal of serving freeway development very poorly as well.

  26. Jarrett at September 15, 2009 at 7:26 am #

    Please cite some of the specific “observed facts” you have in mind.
    Obviously our impressions may vary based on which cities we know well and on
    the outcomes each project intended. I’m talking here specifically about
    light rail as opposed to streetcar projects, where the intent is rapid
    transit and reasonably fast operation is therefore important on at last part
    of the corridor.

  27. M1EK September 15, 2009 at 7:27 am #

    Portland’s more urban lines do better than their freeway-routed ones. DART and Houston do very very well with lines routed in the street (where necessary).
    I’ll throw it back to you – show me a freeway-routed light-rail line that can compete with those startys.

  28. Jarrett at September 15, 2009 at 10:25 am #

    Not sure what you mean by “Portland’s more urban lines.” Portland’s main
    eastward and westward rapid transit corridors are both long lines that
    depend crucially on freeway segments (plus a tunnel in the westside’s case)
    to achieve reasonable travel times for the longer regional trips that they
    are designed to serve. To compare this to Houston’s current short line —
    really more of a streetcar at the scale it operates on now — is to mix
    apples and oranges. Sure, you can get good patronage statistics if you
    serve only very short trips within a dense city, but this post is really
    about the *rapid transit* function, connecting large dense centers to each
    other over significant distances within a large urban region.

  29. M1EK September 15, 2009 at 10:30 am #

    I don’t regard Houston’s line as a streetcar – lately, streetcar has become synonymous with non-reserved-guideway. Portland leaves the freeway as soon as reasonably possible on their older lines. Nowhere did I suggest that NO freeway running should be allowed; I’m with the guys who say “use it when you have to, but get off it as soon as there’s an urban corridor to travel through”.
    To make this more clear where I’m coming from, I’m from Austin, and we’re about to finally open a disastrous ‘urban’ commuter rail line which, although it doesn’t technically follow freeway ROW, is going to fail for the same reasons as if it did (failure to run along urban corridors which could support walk-up traffic and walk-from-train-to-office trips). Our 2000 LRT proposal would probably be viewed by you as unfavorably as Houston’s, which seems ridiculous given the ridership numbers ours would have put up and theirs is putting up now.
    I’m with you on the streetcar-in-shared-traffic nonsense, by the way. It seems to me like you’re making a similar mistake here as their partisans do, though – if the train is ‘fast’ due to running along the freeway but then requires a shuttlebus transfer for most office workers because it didn’t run through enough density to drop them off within a short walk of their office, it seems like the rail solution offers no real mobility advantages.

  30. Jarrett at September 15, 2009 at 11:46 am #

    Sounds like we’re actually in agreement, just focusing on different things.

  31. Alon Levy September 15, 2009 at 3:25 pm #

    I suspect that for bike commuters, being able to fearlessly cross the streets leading to the train station is as important as for pedestrians.

  32. Alon Levy September 15, 2009 at 3:27 pm #

    Jarrett, I think the main comparison is the Maryland side Metro lines and the Virginia side lines. The Maryland lines use freeway ROWs more, and have promoted far less TOD (I’m not sure about ridership) than in Virginia. Maryland has nothing like Arlington or Alexandria.

  33. Doug Allen September 15, 2009 at 3:59 pm #

    When Portland’s first light rail was conceived in the early 1970’s, I think it is fair to say that more thought was given toward service design than was given toward development potential, at least more in proportion to how planners seem to be thinking today. As a result, the inner portion of the initial east-side line, which parallels I-84, was thought of as a fast trunk route, to be fed by intersecting bus routes. This concept, I believe, is one important reason why light rail was chosen rather than the exclusive busway or HOV project that was originally agreed-to by TriMet, the City of Portland, and the Oregon Department of Transportation. Light rail was felt to be much more suitable as a component in a system in which many patrons would be expected to transfer.
    The busway proponents emphasized the “single-ride” aspect in which local buses in the suburbs would then travel express to downtown. Some planners advocated that the initial line should follow what is now essentially the newly opened Green line (making a right-angle turn to follow I-205) that is being criticized in some quarters. The reason was that they wanted transit to be in place early in order to influence development in the Clackamas area to be more “transit oriented.” The deal had already been cut to have an empty busway right-of-way included in the construction of I-205, so many assumed that busway was pre-ordained.
    However, when TriMet gave the project more thought, and hired Wilbur, Smith Associates to study light rail in the corridor, the advantages of a light rail route became apparent. By continuing the initial route straight east to Gresham, through relatively low-density existing development, but on a street with a wide right-of-way due to its status as a former interurban rail route, the line was expected to serve the existing population of potential riders. Serving Gresham had a political advantage in that the line was expected to be paid for largely from money from the abandoned “Mt. Hood Freeway” which was seen by some as serving the same area. Furthermore, the County Commission had been the one local government that had consistently advocated light rail, so continuing east through Multnomah County made sense.
    Unfortunately, during the “Ronald Reagan Recession” of the early 1980’s, TriMet cut way back on the planned intersecting cross-town service, so the full potential of light rail as trunk service within a “multi-destinational” system has never been realized. Still, it is well-patronized, and the initial route is the most cost-effective transit in the TriMet system.
    I think the pendulum has swung way too far in the direction of tailoring transit to serve developers. It is time to think more about fundamental system designs that serve existing development in a cost-effective way. For the most part, that is what Portland’s light rail system does. As others have pointed out, freeways are often the only affordable locations to build higher-speed rail trunk routes. Portland’s second rail line was able to use an abandoned railroad for a considerable distance, which was great, but that was a one-time opportunity. Perhaps the greater criticism of Portland’s newer light rail lines should be their often weak integration with the bus network, rather than their reasonably fast stretches along freeways.

  34. Adam N. September 15, 2009 at 6:28 pm #

    I haven’t heard much from actual daily users of freeway transit stops here, but come on. They suck. I live in Chicago and work in the inaccessible-by-L UofC in Hyde Park (aka where Obama’s house was). The Dan Ryan L stop on the Red Line drops me off in the middle of the freeway where I have spent far too much time waiting for a bus and wishing it was just a little closer so I could walk.
    Let me tell you. After three years of waiting for that bus in the middle of a freezing and dirty overpass, I have had it. Most people can’t even of it for that long. This commute has even actually made my coworkers cry. How’s that for observable evidence?

  35. Jarrett at September 16, 2009 at 12:19 am #

    Of course there’s lots of observable evidence in that vein. I’ve been
    riding transit for almost 30 years, and have spent lots of time at
    freeway-side and freeway-median stops. Sure, waiting next to a freeway
    isn’t pleasant, and access to platforms near freeways can still be
    difficult. But when I think about current freeway-oriented lines and ask
    where else they could have been where they would have been (a) as fast and
    (b) anywhere close to the same construction cost, the answer is usually
    nowhere. There would have been no project.

  36. Adam N. September 16, 2009 at 4:50 pm #

    Thanks for responding. Alas I know something is better than nothing and I can’t argue with your point at all. It’s just so discouraging to use this type of transit sometimes. However, I suppose that not all new transit lines will kept being built in the middle of highways. Even the proposed Chicago Red Line extension to 130th is supposed to be built on ROW that isn’t in a highway (a big change from the original extension plan). Good post, I really enjoyed this one. Thanks.

  37. calwatch September 16, 2009 at 10:10 pm #

    Portland’s light rail is one of the slowest in the country. The roughly 3.2 miles between Lloyd Center and the entrance to the Robertson Tunnel takes a scheduled 30 minutes, and often longer during rush hour – a pathetic average of 7 miles an hour. Basically, this discourages folks riding from the eastside to work in businesses towards the Galleria, and westside residents from working in the Lloyd District. TriMet is right to increase speeds on MAX, but ultimately what needs to be done is to either elevate or sink it through downtown, with fewer stations (instead of, for instance, one on BOTH sides of the transit mall – a mistake the Portland Mall light rail didn’t make.)

  38. Wad September 16, 2009 at 10:25 pm #

    I notice something similar with the San Diego Trolley. Despite having its own street one block north of the commercial heart of downtown, the Trolley has one of the most painfully slow journeys crossing downtown.
    It takes about 20 minutes to go from 12th & Imperial to America Plaza.
    It’s not the stops as much as it is the lack of signal synchronization. The blocks are very small, and the train has to wait about one minute if it doesn’t catch a green.

  39. Jarrett at September 16, 2009 at 10:55 pm #

    I don’t think anyone in Portland would dispute this. Light rail was
    originally very much about stimulating downtown Portland, so the design
    focused on trips terminating there rather than flowing across. Eventually
    I’m sure some sort of downtown subway segment will be needed, much like San
    Francisco’s Muni Metro subway (which wasn’t built until the 1970s).

  40. Jarrett at September 16, 2009 at 11:00 pm #

    True enough, although small blocks have so many virtues for pedestrian life
    that I’d never identify them as a net disadvantage. Much of the success of
    Portland’s downtown, and the promise of San Diego’s, lies in the good
    fortune of those small blocks, which mandate a certain human scale to
    development even in the most brutalist eras of architecture. You’re right
    that small blocks mean more signals, not to mention a shorter maximum train
    length. But in a properly timed signaling system the number of signals
    shouldn’t matter to speed so much as the overall speed limit of the street,
    and this must be low because of the close mix with pedestrians.

  41. CroMagnon September 17, 2009 at 8:19 am #

    Based on the peak hour ridership levels, I don’t see how Portland can expand the system beyond this new extension. Their at capacity and I think that can only be corrected with a downtown tunnel to eliminate the short block length constraints. Are there any significant constraints outside of downtown? The Yellow comes to mind as difficult, but if even the Blue or the Red could accomodate longer platforms outside of downtown, it’s worth it.
    I’ve often wondered why the Portland system has been such a poster child–its been used to sell LRT as a one size fits all approach throughout the US. From what I’ve read, I feel the integrated planning and construction of the system and the high ridership are what makes Portland transit immpressive, not the system per se. If I were designing a regional rail system for a metro area of well over 1 million people, I would avoid building a downtown surface alignment at all costs, and also not have build the Interstate Ave line; because as soon as the line becomes too popular, you have to rebuild to increase capacity. I suppose when Portland first started the system, doing downtown surface might seem reasonable without a strong ridership precedent–but not the Interstate Ave line. I think many transit agencies are starting to realize that attempting to operate rapid-like LRT in mixed-traffic (even with dedicated lanes) is ultimately short-sighted: they’re too slow, too dangerous, and too unreliable.

  42. EngineerScotty September 17, 2009 at 9:55 am #

    Prior to the recent construction of the Transit Mall line, running west from the Steel Bridge to Union Station, and then down the transit mall (5th and 6th) to Portland State University, the biggest bottleneck on the line was the surface street grid, which has trouble as headways approach 3.5 minutes or so. It was pretty saturated with the rush-hour complement of 8 blues, 4 reds, and 4 yellows per hour, so the addition of 4 green trains would put it at three-minute average headways, well above its capacity. That’s one reason why the Yellow trains, as well as the Green, moved to the new mall alignment–now the First Avenue and Yamhill/Morrison sections are back to five minute headways on average (8 blue and 4 red trains/hour during rush) and the mall alignment has a peak average headway of 7.5 trains per hour–4 yellow and 4 green. (According to an Oregonian report, there may be an additional two trains per hour which circulate on the Mall on weekdays; nothing on Tri-Met’s website mentions this however.)
    Right now, the bottleneck is probably the Steel Bridge, which is serving 20 trains per hour in each direction. Even there, however, there is an option–according to reports, two additional tracks might be laid on the bridge, which would double its train capacity.
    The slog through downtown is annoying for crosstown riders–but there isn’t a significant capacity problem downtown with MAX. Tunneling has been discussed in town, but there are no plans currently in the works, and no funding for such a thing at the present time.

  43. Jarrett at September 17, 2009 at 10:17 am #

    Engineer Scotty’s response below is right on, I think. My understanding of the Portland Mall redesign is that there’s now plenty of capacity to add a Milwaukie line and even a Barbur line, as these would be southward extensions of Yellow and Green using trips that already exist on the mall. I agree that Interstate doesn’t make much sense as it is; it was always Tri-Met’s down payment on a line to Vancouver WA, but I won’t even comment on the Columbia River Crossing to which that prospect is now attached.

  44. CroMagnon September 17, 2009 at 11:37 am #

    Thanks, I didn’t know about the new downtown section. Still, I’d think the block length constraint would still be putting a lid on capacity at peak hour since I’ve heard the cars are at or beyond comfortable capacity at the heaviest loading point. (I know they were a few years back, at least.) The future lines would likely increase total system usage, so the Red and Blue might have get more despite the seperate feeding routes.

  45. EngineerScotty September 17, 2009 at 12:13 pm #

    The block length constraint is still there, and will be so until and unless the downtown segment is augmented with a subway.
    Longer trains are of course a way to add capacity at peak hour, but I’d rather have shorter headways, assuming we can afford it. (The disclaimer is to head of any snarky suggestions of replacing MAX with BRT…)
    I misspoke a bit when I said that the biggest bottleneck in the system is the Steel Bridge–that’s only the biggest physical bottleneck. The biggest operational bottleneck, of course, is money. Right now, the limit on Tri-Met’s capacity is availability of rolling stock and ability to hire drivers; not physical infrastructure.

  46. Alon Levy September 17, 2009 at 3:29 pm #

    If I were designing a regional rail system for a sprawling metro area, I’d just lift design principles from Calgary, whose rail ridership-to-population ratio is among the highest in North America and by far the highest in the Sunbelt. In particular:
    – The downtown alignment would be on the surface, in a transit mall, closed to car traffic. Undergrounding can come later – Calgary is planning for it now that the system’s maturing.
    – All freeway construction would freeze, as it would compete with the light rail sytem.
    – Stations would be plain, with automatic fare collection (i.e. no station agents or conductors), except possibly proof-of-payment inspectors.

  47. EngineerScotty September 17, 2009 at 6:40 pm #

    Portland actually comes close:
    * Nowhere that MAX trains run, do cars run, except for the shared crossing on the upper deck of the Steel Bridge. MAX runs in dedicated lanes on Yamhill and Morrison and 1st, and in a transit mall shared with busses on 5th and 6th. These streets do have single car lanes in some places, but MAX gets its own right of way.
    * MAX stations are plain, with auotmatic fare collection, no station agents, no conductors, and fares are enforced with proof of payment inspectors.
    Where Portland falls short is the “no freeways” part, as the city has a fully connected freeway network. No new freeways have been built in town since the 1980s (there have been modernizations and widenings to existing freeways since then), but downtown can be approached by freeway from four directions, and a freeway loop circles the downtown area. While closing part (or all) of this is occasionally proposed, that obviously goes nowhere.

  48. calwatch September 17, 2009 at 8:11 pm #

    The other problem with the Steel Bridge is that it is subject to delay occasionally due to bridge lifts, which ripples through the system. If Fleet Week comes, it’s game over for MAX, since trains are going to be delayed for 20 minutes or more while the bridge goes up and down. Imagine the backup of trains when that happens. (It’s the same single point of failure as is the case with BART, when one slow train leads to 30 minute cascading delays system-wide.)
    The new bridge for Milwaukie trains will provide some redundancy, although ultimately a parallel eastside line, perhaps along Grand/MLK or 11th/12th, should be built. Not only will this help revitalize the eastside, it will give MAX a way out when the Steel Bridge is being worked on. This is probably a better option than adding more tracks on the Steel Bridge.

  49. EngineerScotty September 17, 2009 at 11:17 pm #

    Tri-Met actually handles Steel Bridge failures reasonably well, and planned lifts (such as the Rose Festival fleet) have not in the past posed any problem. Whether or not the Green Line will break the present contingency plans (shuttle busses across the numerous other bridges in town), I don’t know–but Tri-Met is adept at handling single-point failures on MAX.

  50. Chris H September 24, 2009 at 12:21 am #

    One thing that really struck me about Boulevard Périphérique in Paris is how much better it seemed to fit into an Urban environment than an American freeway. I think the lack of shoulders, relatively narrow lanes and frequent capping helped a lot. The first two seem to be true for the example Berlin autobahn.

  51. Beige September 24, 2009 at 8:40 am #

    I don’t know how small the blocks are there, and I appreciate that really giant blocks will be a problem, but when I’m walking in the center of my city, the intersections are just another thing to dread. Another place to wait, wait, wait for the traffic signals to cycle, and they are the dangerous place where you have to look out for the motorized traffic to avoid being smushed. By bicycle, the intersections are again the dangerous places, where you have to pay attention and try to figure out how far left to merge before you get there to avoid hanging up the right-turning traffic. And again, another place to stop and wait, surrounded by idling cars.
    Certainly you could have too few, as you’ll find outside of the city where no one at all is walking, but it never struck me that you’d want even more. Unless Milwaukee counts as a city with many small blocks.

  52. Alon Levy September 24, 2009 at 4:13 pm #

    The Peripherique also conveniently segregates Paris from the suburbs, which are populated by poor people and minorities.

  53. Scott Workman September 28, 2009 at 8:42 am #

    Eric, good point on this….I might have missed a bunch of relevant posts but what is also important when you have a rapid line making use of an existing right of way (freight, highway, etc), you must have connectivity to that rapid line. As you mention above regarding Charlotte, this would be a street car line (more desirable) or bus line (less desirable) serving the rapid line stop and points perpendicular to the rapid line. That is where the TOD can happen without being next to a major transit-way.

  54. Steve Schijns November 12, 2009 at 10:49 am #

    Agree that freeway corridors offer the path of least resistance (and hence least cost / disruption) to new rapid transit lines but the key practical issue has been, in my planning and design experience, the ability to retrofit effective, accessible, and functional stations. Freeway corridors are not only not planned to accommodate transit station needs, but are rife with features that inherently conflict with transit stops. Even in a brand new “integrated” multi-modal transportation corridor it is difficult to reconcile the different functional needs of the freeway and the transit line.
    From a transit user’s perspective, it’s all about the station; how you travel between stations doesn’t really matter. It could be on a bus in a bus lane, an LRT vehicle on a track, or even a bus in mixed freeway traffic if it moves and travel times are reliable. But take a typical freeway and try to retrofit a station that has a parking lot, walking / cycling accessibility, movement in and out by local buses in all directions, is compatible with surrounding land uses, and doesn’t disrupt the freeway traffic operations (and doesn’t cost a zillion dollars!) – it generally can’t be done.
    So then you push the transit stop to some location between interchanges and you end up with an isolated location that is inaccessible, doesn’t accommodate transfers, fails to serve transit users well in more than one direction, and conflicts with established land uses.
    Alternatively, avoid stations entirely and use the freeway corridor as an express conduit between the central city and some off-line suburban transit hub. Success at last!

  55. Andrew November 15, 2009 at 4:22 pm #

    Transit along freeways can be done well and it can be done badly. If the density surrounding the freeway is high and there are good pedestrian connections to nearby areas it can work fine. For instance, look at the Spadina subway in Toronto: near Eglinton West, Glencairn, and Lawrence West stations, there is medium to high density right up to the edge of the freeway, there are numerous pedestrian bridges crossing the freeway and there are footpaths parallel to the freeway on either side for much of the length. These stations are quite well designed. Yorkdale is somewhat worse because it is near a large shopping mall with big parking lots, and the only connection to the mall is via an elevated bridge; the connections to surface streets are unpleasant. Wilson is very badly designed; it is in an industrial area, next to a subway line, an airport and some big box stores and cut off from a nearby residential area by a parking lot. It is really only convenient for passengers transferring to buses via designated bus bays.

  56. A transport geek in middle school October 4, 2011 at 12:00 am #

    Of course rapid transit can work along freeways! They do it in Perth where I live and the each of the two freeway lines (Joondalup and Mandurah) have about double the ridership of the conventional railway lines (Fremantle, Midland,Thornlie/ Armadale), although in Perth we use heavy rail instead of light rail, and the Mandurah Line is more intercity. For freeway transit to be successful it needs to be frequent(ours is every 5 min during peak hour and every 15 min during off-peak and weekends), very fast (The Joondalup line has an average speed of about 60 km/ and Mandurah Line nearly 90 km/h), and have good feeder buses. I admit we aren’t very good at TODs. Cockburn Central tries its best with apartments near the station, but the freeway still separates them from the station. Wellard isn’t on the freeway but it isn’t very dense. Still, the lines are very busy, with peak trips full and some off-peak and even contra-peak services standing-room only. For more info click on my name

  57. Ben Smith May 16, 2012 at 10:09 pm #

    This comment is a little late to the game, but still I would like to share it. While rapid transit can work along highways, I think there are two things which need to be kept in mind:
    1. Stop spacing and speed
    2. Densities around stops and interchanges
    For the first point, while a grade separated transit route stopping every 800m or so may prove to be very competitive speed wise along an urban street or avenue, along a highway it is a different story. For example, a stretch of subway in Toronto runs along an unfinished highway with stops about every 1km. With the exception of traffic being backed up at the ends of the route (which I will admit, I get a shot of ecstasy watching traffic backed up as I pass them from the train), it cannot pick up enough speed to compete with traffic when flowing smoothly. Therefore stops need to be spaced out enough to allow the trains to at least keep up with flowing traffic.
    The other thing to keep in mind is if stops are located at interchanges with major arteries, which would make the most sense as this is where people connect to the service, that density is limited to allow highway traffic to absorb on to the artery smoothly. Whenever I’m in the US and travelling along the Interstates in the suburbs and in between cities, I have to give credit to how they have planned the areas around highway exits. Yes they are pedestrian unfriendly and an incredible waste of land, but they allow the influx of traffic from the highway to absorb on to the local roads relatively smoothly.
    To conclude, while rapid transit can work along highways, it needs to adjust its role to better suit its environment. If a route is to operate primarily along such a corridor, it is probably better to construct it as a commuter line or suburban metro, instead as a city subway line.