Putting Rail Lines in Highway Medians

North Americans are in love with trains that go in highway medians. A large fraction of urban rail construction since World War Two, both light rail and full metro, has used highway medians as cheap at-grade rights-of-way to extend train service, often deep into the suburbs. Some proposed longer-range lines are supposed to go in medians as well: Florida had reserved space in the I-4 median for Orlando-Tampa high-speed rail, and Xpress West planned to go from Las Vegas to the outskirts of the Los Angeles area in the I-15 median. The Texas Central Railway, a private group backed by JR Central planning high-speed rail between Dallas and Houston, is considering several alignments, but markets the route as following I-45 (no mention of median) in some public discussions. In nearly all cases, both urban and intercity, it borders on incompetent to design rail lines in highway medians; intercity lines frequently follow highways on one side, but even that tends to be overrated in American discussions in my experience.

Urban Rail

For urban rail, the reason to use highways is that, in most of North America, they’re everywhere, and they’re usually equipped with generous medians and shoulders, allowing relatively cheap placement of rail tracks. Of note, this is generally not the cheapest option: construction on extant (often disused) rail rights-of-way tends to be cheaper. However, in many cases, a rail right-of-way is unavailable, hosts heavy freight traffic, has been permanently turned into a trail, or has commuter trains without integration into the rest of the urban transit network. Examples include the Dan Ryan half of the Red Line and both halves of the Blue Line in Chicago, the Orange and Silver Lines in Washington, the outer ends of BART, the Spadina line in Toronto, and several light rail lines. Often they run on one side of the road, but more frequently they’re in the median, which was often reserved for it when the road was built (as in Chicago and Calgary).

The problem is that nobody wants to live, work, or hang out next to a busy grade-separated road. Living or working a kilometer or two away, with easy access by car, is great for the driver, but within close walking distance, there is just too much noise, pollution, and blight, and the pedestrian environment is unwelcoming. The transit-oriented development in Metrotown and Arlington could not have happened next to a freeway. Christof Spieler frames this as a decision of spending more money on routing trains near where people live versus staying on the easy rights-of-way. But this isn’t quite right: the Expo Line in Vancouver was assembled out of an interurban right-of-way and a city center tunnel, both out of service; the line’s high ridership comes from subsequent development next to Metrotown and other stations.

Other times, the routing comes from a deliberate decision to integrate the trains with cars, with large park-and-rides at the ends. This is common on newer light rail systems in the US (though not Canada, as Calgary prefers integration with connecting buses) and in the Washington and San Francisco suburbs. This makes things even worse, by extending the radius within which the environment is built for cars rather than for people, and by encouraging the same park-and-ride construction elsewhere, along abandoned railroads and greenfield routes, where the preexisting environment is not car-oriented.

I do not want to categorically say that cities should never build urban rail alongside highways. But I cannot think of a single example in which this was done right. Calgary is a marginal case: it did build light rail along highways, and had some success with transit-oriented development, but those highways are arterials rather than freeways, and this makes the pedestrian environment somewhat better.

The situation is somewhat different for suburban rail, but usually the scale of suburban rail is such that there’s not much new construction, only reappropriation of old lines. These lines are long and the environments low-density, making it hard justify the costs of new lines in most cases. Where new suburban rail is built, for examples the Grand Paris Express, and various airport connectors, it is typically in environments with such expected traffic density that the rules for urban rail apply, and we tend to see more underground construction or usage of extant rights-of-way.

Intercity Rail

The reasons favoring highway alignments intercity rail in the US are somewhat different. Tellingly, HSR in Europe is frequently twinned with motorways. It is not about integration with cars, since those alignments are rarely if ever meant to have major stops in their middle. Instead, it’s about picking a pre-impacted alignment, where there are fewer property takings and fewer NIMBYs. This logic is sound, but I often see Americans take it to extremes when discussing HSR.

The first problem is that roads are almost never as straight as HSR needs to be. The design standards I have seen after briefly Googling give the radius of a motorway capable of about 120 km/h as, at a minimum, 500-700 meters. With these curves, trains, too, are capable of achieving about 120 km/h – less at 500 meters without tilting, more at 700 meters with tilting. The most recent high-speed lines are built with a minimum curve radius of 7 km; about the absolute minimum that can be done, with design compromises and tilting trains, is 4 km. This implies that the trains have to deviate from the motorway alignment whenever it curves. In flat regions the road curves are much gentler than the minimum, but still too sharp for full-speed running. Both Florida HSR and Xpress West noted that the trains would have to slow down whenever the Interstate curved, because the need to run in the median would prevent them from curving gently enough to maintain full speed.

Of note, the European examples of HSR running in motorway alignments have it running alongside the roads, not in the medians. I invite the reader to spend a few minutes following French LGVs on Google Maps and seeing this. This is because there invariably have to be small deviations from the road, which in a rural area are trivial when one runs next to the road but require viaducts when one runs between the road’s two carriages.

There may also be an issue regarding reusing the Interstates. To transit supporters who view HSR as a replacement for freeways, this has an element of poetic justice, or just plain practical reuse of infrastructure they think is obsolete. I chanced upon this while looking up Interstate design standards, but I’ve seen similar proposals elsewhere, as well as dissimilar proposals making use of interstate terminology, as a reminder of past national greatness. It comes from the same place as proposals to reuse auto factories to produce rolling stock: there’s a romantic aspect in addition to or instead of an economic one.

But the most fundamental problem is that the contentious experiences of the freeway revolts and modern-day NIMBYism have soured Americans on any process that involves brazen takings. What I mean by brazen is that carving a new right-of-way, especially through a populated area, looks obvious on a map. In contrast, sticking to a preexisting right-of-way and incrementally widening it or straightening curves is less controversial, even when it involves eminent domain as well, and opposition remains much more local, based on the specific properties being taken, rather than stated in general principles. I am not completely sure why this is so; my suspicion is that widening and straightening are more easily justified as things that must be done, whereas a new right-of-way looks gratuitous.

In either way, Americans have convinced themselves that NIMBYs are a major obstacle to infrastructure construction. While zoning is a notoriously NIMBY-prone process, infrastructure often isn’t. In the English common law world, expropriations are if anything easier than in France, where farmers are especially powerful, or Japan, where rioters threatened to block the construction of Narita Airport. NIMBYs are good at getting their names out in the media, but when it comes to blocking construction, they are relatively powerless; California HSR is facing NIMBYs in the Central Valley, many of whom are conservative and politically opposed to the project regardless of local impact, but so far they have not managed to delay construction.

However, NIMBYs are a convenient bogeyman for public projects, as their motives are openly selfish. They give charismatic, authoritarian leaders the opportunity to portray their infrastructure projects as battles between the common good and backward-looking parochial interests. As I’ve noted multiple times before, New York’s livable streets community (which is similar politically to the set of HSR supporters in the US) tends to overblow the importance of NIMBYs to the point of seeing NIMBYs even when the concerns have nothing to do with NIMBYism: see, for example, the reaction to the opposition of two Harlem politicians to a plan to speed up only the whitest bus route through the neighborhood.


  1. Jonathan English

    Great post. I’ve always found the median fixation a bit strange and I agree that there’s a romantic aspect to it.

    I also agree that the NIMBY problem is very often overstated, especially when agencies decide on massively expensive solutions ostensibly to preclude NIMBY problems that haven’t even materialized yet. So-called NIMBY groups when it comes to infrastructure projects are often even fairly constructive. The Weston activists in Toronto, for example, were pushing for electrification and a service with reasonable fares and a stop in their neighbourhood in addition to the premium airport shuttle that was planned. Those things are actually good ideas and they probably helped speed the electrification discussion.

    When an infrastructure project genuinely will blight some people’s property, I have no problem with fair compensation in addition to any land acquisition. It’s far easier and cheaper to pay $50,000 or whatever to a dozen homeowners who will suddenly find themselves next to a high-frequency rail line than to plan a billion dollar tunnel from the outset to preclude their objections.

    • betamagellan

      In Chicago, at least, the desire to put trains in medians came from the same place as single-use zoning—looking at mid-century plans one will find that there wasn’t much to it besides “rail and highways are both transportation, ergo they belong together.” I’m fairly sure the ability to bundle rail and highway investments at once played into it—the major expressways in Chicago all have rail that was built either at the same time or within about twenty years of the roadway, or dedicated space meant for bus or rail expansion. It’s also worth noting that the median lines in Chicago tend to be fed by bus, not park-and-rides, and I’m sure this contributes to Chicago’s high distances and costs-per-trip relative to the other big northeastern American cities’ transit systems (though I’ve found the northwest side stations not bad for walking, in large part because the Kennedy’s a fairly narrow and primitive highway).

      Although I know little about its efficacy as a system, there’s a lot of admiration in the civil engineering community for the project management of the RTD light rail extensions and part of that came from effectively bundling of different transportation projects in one corridor, often alongside highway projects. I’m pretty sure it’s also a side effect of a planning process that sees rail mainly as a means of absorbing growth in automobile traffic while preserving suburban land uses.

  2. letsgola

    Following existing rail or freeway ROWs is the politically safe option. Even with CAHSR, they stuck to existing ROWs in the Central Valley (the 5, the 99, UP, BNSF) where they could and only went greenfield where topography forces it (Tehachapi, Soledad, Grapevine). Existing ROWs are also the legally safe option, because there’s no chance your eminent domain case will end up delayed in court for years and years. The CAHSR alignment through Santa Clarita, for example, may end up in court. Unless there is overwhelming political support for a project, politicians are just not going to take a new ROW through a built-up area.

    The Xpress West thing is less excusable because it’s out in the middle of the desert. I can’t recall if it was supposed to run in the median or on the side. Either way they probably would have been better off following UP (which wanted nothing to do with them) or at least taking a greenfield ROW that didn’t follow the 15 over Mountain Pass. Still, there’s little doubt they would have faced litigation from environmental groups if they tried that.

    Regarding urban transit, agreed. If US tunneling costs were lower, that might make it easier to follow routes other than freeways or old rail ROWs. LA’s Green Line suffers from an unpleasant environment in the median of the 105 (though it still does ok on ridership thanks to relatively dense surroundings).

    • Richard Mlynarik

      Even with CAHSR, they stuck to existing ROWs in the Central Valley (the 5, the 99, UP, BNSF) where they could and only went greenfield where topography forces it (Tehachapi, Soledad, Grapevine).

      The “existing ROWs in the Central Valley” business is in fact a public outcome disaster.

      The force behind it, as with everything connected to “public” civil engineering projects, is cost maximization. Designing and building a rail line in existing corridors is hugely more complicated, involving substantial utility relocations, entire freeway relocations, freight railroad relocations, complicated parallel grade separations and overcrossings of freight railroads, far more and far more complex road overcrossings and undercrossings, multi-level aerial passenger-hostile passenger stations with endless access bridges placed above freight yards, etc.

      To make up for it (not!), the property acquisition costs are as high or higher. The difference is that the private freight railroads and the city father types are much more cagey about driving rewarding bargains than the yokel tea party “this means then end of the family farm” astroturfers.

      We also, in a very small way, have our dim-witted friends at the Sierra Club to thank for agreeing to endorse the language crafted by the engineering contractors that was California’s 1999 Proposition 1A: the “existing transportation corridors” business there sure sounded to the non-analytical cadres like something that would “end sprawl” and “not impact parklands” and all that yeah-yeah good stuff jive, so sure, good boy, roll over for a superficial promise like you always do?

      It really is as simple as “follow the money” for this hyper-mega-project. HSR in the US doesn’t belong in privately owned railroad corridors, and rail above 200kmh doesn’t belong in any urban area. These two insane public policy choices (and the juicy compounding interactions between them!) are, however, just the $100 billion bonanza that one would like to think was beyond anybody’s wildest dreams of avarice, but instead is the fiscal and environmental disaster that California and US federal taxpayers are being reamed for.

        • Nathanael Nerode

          There’s really no route through the Central Valley which doesn’t impact someone. The whole thing is built up. Richard M. delusionally believes that there were cheaper routes through the CV. There weren’t. He’s nuts.

          • Joey

            Every route impacts someone, but building a full-speed alignment through urban areas has a disproportionately large number of impacts, and by extension, mitigation costs. Also trying to switch from one side to the other of UP’s tracks or SR-99 generally involves large complex structures with lots of straddle bents.

  3. Matthew

    Boston has the opposite case where rail lines were used to provide highway right of ways. The Red line through Dorchester, the Framingham line within rt 128, and Orange line north. And in the case of the Orange line south, the original plan was highway with median rail that got converted into rail with park.

    Honestly, all the rail with highway stations are fairly sucky to be around. I guess back bay is an exceptional case because it’s covered. Savin hill is an alright area to be around but the moment you crest the hill overlooking the interstate it gets miserable.

    Rockridge in Oakland might be the only successful “median rail” station I can think of…

    • Roland S

      I’ve heard Rockridge cited a couple of times. What is successful about it that is not successful at, say, Irving Park or Rosemont in Chicago which have a similar elevated setup?

      • enf1234567890

        It’s been too long since I’ve been to Irving Park or Rosemont for me to remember what they’re like, but what Rockridge has is the sense, as you walk down the stairs from the station, that you are emerging into a real, living place.

        Rockridge opens directly onto a thriving neighborhood commercial street (College Avenue), which is fairly narrow and is slow and difficult to drive on. There are no freeway ramps that lead directly onto College or that are running parallel to the freeway at that location to widen its right of way. And the freeway is high off the ground, high enough that when you look back at the station, what you see is its mezzanine level below the freeway, not the track/roadway level above, so it looks like you are looking at a station, not a freeway.

        It all stands in direct contrast to MacArthur, where the freeway is wide and low, the street is wide and lifeless, and the station opens onto a parking lot plaza that no one lingers in.

        • betamagellan

          I’ve never been to Oakland, but here’s my opinion of the Kennedy Expressway stations:

          Irving Park is still a real, dense-ish neighborhood despite the expressway, but most commercial activity is a few-blocks’-wall to short bus ride away from the expressway, and it’s still under a wide concrete viaduct.

          The street geometry around Montrose actually isn’t terrible (i. e. no traffic going onto the expressway, very little exiting the expressway and then from only one direction), but the immediate area is very light industrial/gas station filled.

          The best on the northwest side is probably Jefferson Park—it’s not near any on-off ramps and the area could be a nice little kernel of density were the locals willing to upzone, but even so the local streets are all fairly accessible.

          Cumberland is okay in a suburban skyway sense (a major intersection, but there are bridges across it), and if you’re a reverse commuter it’s easy to walk to the nearby office park, provided it isn’t too far from the station.

          I think the issue of platform environment is pretty important, too—Irving Park and Montrose are terrible and windswept, whereas Jefferson Park and Cumberland are more shielded and not uncomfortable at all.

          • Roland S

            Harlem would be great if only the sidewalk on the overpass were wider. The station is pretty spacious and sheltered from traffic, and bus passengers have their own depot so they’re not hanging out on the Harlem sidewalk. If they modified the loop ramp and built a retail pavilion over the top, it would be perfect.

        • Richard Mlynarik

          That’s spot on about Rockridge. It’s the lack of freeway access and (by US standards) lack of station parking — along with the lack of tender Redevelopment Agency attention in the 1970s, and also some much less savory racial politics and segregation — that work their puny little ways against the massive inherent freeway blight. Also that College Avenue is perpendicular to the monster.

        • calwatch

          The other thing not mentioned is that SR 24 has limited truck use due to restrictions in the Caldecott Tunnel and general lack of warehouse facilities in Central and Eastern Contra Costa County. The I-580 median stations serve interregional truck traffic coming from the Port of Oakland.

  4. Eric

    Any examples where freight traffic was/should be diverted to a newly built freeway ROW so that the old rail ROW could be used exclusively for urban transit?

    • Alon Levy

      I’m not sure how solid this example is, but the inner San Bernardino Line. Right now, there are two parallel ROWs: one along I-10, single-tracked without the possibility of expansion since all space has been deeded for car or bus lanes; one in its own ROW, which has enough space for everything. LA’s plans are actually the other way around: keep all the freight in the dedicated ROW, where UP is investing, and keep the passenger rail on I-10. This is wrong – for UP, unless it’s planning on double track the whole way, the two options are about equally good; for passenger service, only one of these routes offers the possibility of high frequency, closely spaced urban stops, and future TOD.

      • letsgola

        The plan is for double track the whole way along the UP ROW. http://www.theaceproject.org/project_area.php

        Unfortunately it probably is the better ROW for rapid transit, as it goes closer to downtown Alhambra & San Gabriel. It’s probably better from UP’s point of view too because it provides better access to their yard at the west end.

        They could always just acquire more ROW along the 10; it’s not like they didn’t just do it to widen the 5 in Norwalk and Santa Fe Springs.

        • calwatch

          It’s still single track through San Gabriel, even after the trench construction. UP doesn’t have the money to add the second track. The San Bernardino Line study contemplates running Metrolink express service along the UP track – http://www.scribd.com/doc/238440316/San-Bernardino-Line-Study-presentation – which would probably work if UP management ever becomes more friendlier to passenger rail, and there were two tracks along the Alhambra/San Gabriel Trench corridor.

          • Anandakos

            The UP certainly has money to add a second track. They are fabulously wealthy and a money spinner since they got SP completely integrated into their systems. it’s just a question of “is this the best place to spend $300 million? How long will it take to pay for itself?”

            Remember, the UP has its old route through Riverside as a relief valve.

      • Anandakos


        You’re forgetting Cal State LA. I rode a mid-morning trip out of San Bernardino the LA and saw perhaps 25% of the train empty at Cal State. While running the AM expresses via Alhambra might make a lot of sense if a track were added, the mid-day service depends strongly on the Cal State ridership.

        Bypassing the ElMonte Busway section also doesn’t solve the problem of the LOOOOOONNNNNGG stretch of single track from the south end of Baldwin Park at Amar Road to Irwindale Avenue. The Baldwin Park station is even single-tracked. Since that trackage is almost on peoples’ front steps there’s no way to improve that bottleneck.

        There’s a similar long stretch of single-track through Covina and La Verne between the Covina Station and just west of the Pomona Station. Again, there’s insufficient room to double-track at all in a good part of it and only by moving the existing track over in the rest.

        East of Claremont there are several stretches of single-track, but there’s plenty of room to double track as needed.

        Finally, it looks to me like the “trench” through Alhambra is ready for a second track although there’s a stretch where the existing track is in the middle which might be problematic. Perhaps UP would go in 50-50 with Metrolink with the understanding that it would be used only for passenger trains during the peak hours. So far as serving Alhambra and San Gabriel, they’re pretty close in and the areas around the UP tracks are very industrial. Would there be ridership anywhere near what CSLA attracts?

        There’s certainly no room for platforms down in the trench. What a horrid environment for a station anyway.

  5. dr2chase

    I have some sympathy with the SF Peninsula NIMBYs for CA HSR. We lived a short walk from the Menlo Park rail station for a few years, and the rail is immediately adjacent to people’s back yards through parts of Palo Alto, Menlo Park, and Atherton (and maybe other places). I’m not sure what speed is contemplated for that stretch, but I can certainly understand that it could be plenty annoying to people living adjacent to the tracks if it was much higher than current trains.

    On the other hand, train noise falls off rapidly with distance (we lived on Laurel St then, we live a little further from the local commuter rail now and I’ve not heard anyone in our neighborhood complain about it in twenty years — and there was a near unanimous complaint against a kennel with noisy barking dogs across the tracks a few years ago) so it might make sense to think hard about making actual abutters whole, and to tell the rest to take a hike.

    • Alon Levy

      To be honest, in many cases noise emissions are actually going to decrease, because of electrification. The diesel engines are really noisy, and their noise tends to be independent of speed, so in relative terms it’s especially bad at low speed. For what it’s worth, suburbanites along the Northeast Corridor complaint about freight train noise more than about passenger train noise; the passenger trains are a lot faster, but they’re also a lot quieter at any given speed. The story I remember, I believe due to Adirondacker, is in 110 mph SEPTA territory – someone complained that the noise is unbearable, but after probing it turned out that the noise they were complaining about was at times no passenger trains were running, so it was exclusively about freight.

  6. Amanda Clark

    Actually, VTA light rail I think has done a decent job with rail in the median of Highways 87 and 85. That particular stretch actually gets you from south San Jose to downtown (Santa Teresa to downtown) at a pretty good clip, and traverses an awful lot of suburbia. VTA light rail is a shitty, horribly designed system, but if you’re going to have light rail service south San Jose, then travelling down the medians isn’t a bad way to go.

    The flip side is that obviously walking is a pain-I used to live next to the Snell station, but it still took me almost 30 minutes to travel through the maze of suburban streets to get there on foot.

  7. AlexB

    In the situations where lines run in highways, I think the key difference is whether the stations are also located in the highway median. The stations in Chicago on the Blue Line in the Kennedy Expressway are horrendous. On the other hand, I imagine the stops on Burnside Avenue at its eastern end are made more attractive to riders because the line operates more as an “express” on I-84 between there and downtown. Highway alignments used in this way can theoretically be used to support high ridership and walkable neighborhoods.

    In general, running transit lines (with stations) in a highway median is probably only rarely effective. However, I can’t think of a single place where the city has made the necessary investments to even attempt to create high ridership at a highway station, such as encouraging much denser development and mitigating the effects of being near the freeway. The new elevated walkways at Tyson’s corner might work, but the Silver Line is above an arterial, not an expressway, so it’s a little different.

    In a related note, transit activists typically overlap with urbanists who want the whole package: fast, reliable and electric rail mass transit combined with parking maximums, high density zoning, and pedestrian amenities. There is an older, somewhat discredited concept that aims to build electric rail transit as a redundant backup system in the event of a gas price spike or to provide an alternative to highway congestion. The purpose of these systems is to provide alternatives to suburbanites, not change the development pattern. This is where the park n rides of San Francisco, Atlanta, DC, Chicago come from. That model is not completely wrong, but cities like Dallas and Denver that have built sprawling light rail networks, often in or next to highways and maximum headways of 15 or more minutes, should not claim to be building transit that supports dense, walkable communities. Houston’s example makes the most sense to me: focus rail transit on the busiest corridors close to downtown, and use HOV lanes and park n rides to provide alternatives to suburbanites.

    Rail transit can solve many problems, but rarely all at the same time. Those who try to sell these lines to the public should make their goals and expectations clear.

  8. D. Malcolm Carson

    As with most things, “it depends”. It’s probably safe to say that rail transit in a freeway median is very unlikely to spur transit-oriented development, but that’s not its only purpose. The Green Line in Los Angeles was built at the same time as the freeway, so the choice wasn’t in the median versus another alternative, it was in the median or nowhere. It has decent ridership today because it’s now the fastest way to traverse that corridor on public transit. People don’t generally walk to it, and there isn’t any transit-oriented development around it, but they do connect to it from local bus lines, several of which are among the most heavily used in the County. As LA’s rail system gets further built out, it will get further use as a connecting line. A parallel rail line along an old rail right of way (Expo) is being built now, and it will likely do much better in terms of spurring development and creating walk-able communities, but it’s also going to come in at about double the cost. I guess the point is that the more you can route a rail line through existing walkable dense communities, the better, but I wouldn’t “look a gift horse in the mouth” and reject an opportunity to build one in a freeway median to the extent that it presented itself.

    • calwatch

      Well in order to build the Century Freeway, the courts ordered a transitway be built within it, which is the Green Line. It’s another example of transit by consent decree, but at least it wasn’t a busway.

  9. Pingback: The Problem With “Infrastructurism” | Streetsblog.net
  10. Tom

    It appears that no one has ever designed/planned rail transit systems = It is the MONEY and Schedule. No one plays the best way – only the least cost ways –
    Basic transportation planning is the that transit should avoid paralleling or being coincident with other major rider modes…BUT try to get a ROW = eminent domaine/time/cost or go underground the try to get the station sites = ED/$/Time or underground…
    Oh Yeah then there are the existing mainline/Class1 Rail corridors = dealing with the 100+year old entrenched federally protected companies = Time/Money – one asks what to do and provides a signed blank check and a 20 year schedule for the RR company whenever they choose…then Wait…Freeways/Interstates belong the the USDoT and therefore land is available….schedule can be made even though it won’t work the way the “planners” know…change the model to allow service of the whole service area will go to the freeway route, even knowing they won’t.

  11. Pingback: Today’s Headlines | Streetsblog Los Angeles
  12. MichaelF

    Metra’s STAR line proposal explicitly calls for building a new commuter rail line in the median of the Jane Adams Tollway which does not have a median. I believe the Center for Neighborhood Technology’s Chicago Transit Future proposal calls for Blue Line extension in medians that also don’t exist. I understand, without agreeing with, putting a rail line in a median that already exists or will be built anyway but who would rebuild an expressway just to create a median for a rail project?

    • Roland S

      The Addams Tollway is currently being rebuilt. After this is completed in 2016, the inside shoulders will be wide enough to accommodate buses comfortably. At a future date, the tollway can add a few feet of pavement to the outside of the roadway and shift the lanes outward (the overpasses are already designed for this) and then turn the two inside shoulders into a rail or bus transitway like the Kennedy currently has. Things get more complicated at stations though…

      • MichaelF

        Thank you. That does make more sense than the impression I was getting from Metra’s website.

  13. Stephen

    Of course you’re usually right. However, I can think of a few places where the best route for a future rapid transit line would be along/within a highway. Both are relatively dense places that grew up around a highway.

    The first is in Houston, where the geography of the city is literally defined by its highways – “urban” means within the 610 Loop. If you were going to serve the Energy Corridor and Katy with rapid transit – and seeing how they’re growing, I imagine some day Houston might want to – there’s really no other option but I-10. It’s a huge fucking highway (it’s that massive thing that you always see in urbanist horror tweets about highways), but it’s the only thing that goes through the heart of the Energy Corridor.

    The second is in/outside of Bucharest. The northern suburbs grew up along DN-1, and they’re becoming quite dense. It’s also not a super oppressive highway – it’s at grade and there’s a fair amount of development directly adjacent to it. I believe the plan is to run one of the metro lines along/under/above it (can’t remember which), and this seems like a good plan to me.

    • Anandakos

      I-10 is a “huge fucking highway” precisely because TXDOT consumed the old MKT (“Katy”) trackage just to the north of it. It would have been a sockdolager line for commuter rail, passing right between all the huge oil company offices along 10 west of the West Belt, at Kirkwood, Dairy-Ashford,and Eldridge. But it’s just the Katy Freeway service road today.

      There is an HOV facility in the middle of the freeway, but it’s completely unsuited to rail because of the grades at the interchanges.

      • Joey

        How steep are the grades at the interchanges? Modern EMUs can climb 3.5% grades easily and probably 4.5% without too much trouble. High power/weight trains like LRVs can climb even steeper grades.

      • Nathanael Nerode

        There are a lot of examples where freeways were built next to rail lines, or (worse) replacing rail lines. The freeways have uniformly killed the neighborhoods around them, while the rail lines tended to be supportive of them.

        Not sure how to reverse this, but these are often examples where the freeway route is a good railway route, it’s just a terrible freeway route and the freeway should be torn out.

  14. Alon Levy

    In Tel Aviv, some of the rail mainlines are in freeway medians. The connection between Tel Aviv Center and Tel Aviv South/HaHagana, enabling nationwide through-running, was built simultaneously with the Ayalon Freeway. A recent line constructed to serve Bat Yam is also in a freeway median. The results are not great. There is a lot of new development in Tel Aviv next to Ayalon, but it’s unwalkable, and walking across from the Tel Aviv side to the Ramat Gan side is so hard that they offer special tickets to cross the fare barriers just to walk across, without accessing the platforms. The last-kilometer connectivity is poor: for example, there is a station called Tel Aviv University, but it’s a long walk uphill from the station to campus.

    • Moshe S.

      On the other hand, HaShalom connectivity to jobs (Kirya and Azrieli accessible without even crossing the street) is excellent.

      • Alon Levy

        Well, yes. The problem is that a lot of what they’re building in that area is completely unwalkable (Namir Road, ugh). You can go between buildings on bridges above the roads, but you can’t really cross the street at grade level.

  15. Andre Lot

    There is a lot that can be done to abate noise and make the surroundings of highways places that are far more convenient and nice for offices or residences. It all depends on how the areas adjacent to highway ROWs are managed.

    Two European examples pertinent to this thread:

    1) U-Bahn running on central lanes of A40 in Essen, Germany (http://goo.gl/maps/Yecw5)

    2) Amsterdam Zuidas, major real estate development surrounding A10 in Amsterdam (http://goo.gl/maps/Pnd5n)

    You can use very tall noise barriers that greatly increase quality of life on adjacent property of highways or railways.

    As for CAHSR project, following I-5 would actually reduce costs a lot on Central Valley: there is virtually nothing built to the west of A5, the terrain is very flat, and the costs of adapting couple dozen overpasses and junctions would be minimal. Of course this would left out the valley disconnected as well.

    • Alon Levy

      Of course following I-5 would reduce the cost of CAHSR. I don’t think anyone doubts that. What is in doubt is how much cheaper the construction would be. The Central Valley segments that are being tendered right now are already pretty cheap, with mainly at-grade construction.

      As for Amsterdam-Zuid, what mode of transportation do people who work in the adjacent office buildings usually take? I’m asking because it looks fairly auto-oriented to me, with what look like large parking lots.

      • Andre Lot

        There is a major railway and overground metro station (Amsterdam Zuid), plus secondary stations nearby, 2 “fast tram” lines converging there, and also easy access from Amsterdam ring road A10. They also built 4 major segregated bike paths passing nearby.

      • betamagellan

        I actually work there and I think the majority of people are probably coming from regional rail, with bike, metro/sneltram, car and tram all contributing. While there’s some structured parking incorporated into the buildings, the main surface lot is for the adjacent recreational area, and the Google Earth photo’s out-of-date—much of the surface area north of De Boelelaan is currently under construction. Though that probably will increase the total amount of parking, due to new structures incorporated within new buildings, I think it will drive regional and metro growth more—long-term plans have Amsterdam-Zuid as the main hub, and the N-S metro link through the center of Amsterdam will be going there, too.

        It also helps that there are no exits directly near Amsterdam-Zuid, and that the A10 is fairly puny by American standards (and split in two, so you’re never going under more than two lanes). That said, the walking environment at the next station to the east, Amstelveenseweg, is terrible in a fairly typical American highway-oriented transit sense.

      • EJ

        The main reason CAHSR along I-5 would be cheaper is, besides being slightly shorter than the current route, you’ve got a MUCH smaller number of grade separations to deal with, as you’re not going through major urban areas. You also don’t have to deal with crossing any railroad spur lines. In addition, you’ve got less land to acquire (you would have to do some acquisition, for example in a few spots where I-5 is too bendy to follow exactly, but there’s a lot less than the current route).

        The downside of not going through major cities (and yes, by the standards of most of the US, Bakersfield and Fresno are major cities), is of course you’re not serving that population.

    • Joey

      I-5 would be cheaper than the current alignment, but probably not much cheaper than a West-of-99 alignment which skirted the developed areas rather than passing through them, It boils down to the fact that a lot of the costs come from grade separations with land acquisition only being a small portion of the budget.

      • Anandakos

        Going through the centers of Bakersfield, Fresno, and Merced are features, not bugs. CalHSR is NOT for people who want to go between San Francisco and LA; well, not primarily for them. Airplanes will serve that market better for anyone whose origin or destination is not within a walk or short cab ride of the terminals. Since they’ll have to get to and from the train stations they might as well get to and from one of the four airports in the LA basin or three in the Bay Area and make the hop in an hour.

        It’s for people who want to go between either of those end points and one of the intermediate cities or between the intermediate cities. Air travel for those trips sucks!

        Therefore, having HSR dive right into the hearts of those cities where many people will be similarly be able to take a short cab ride, the bus or walk is much better for more travelers that having it stop in West Boondocks and making people drive and park there.

        • Alon Levy

          That’s not really true. 3-hour HSR service tends to get much more ridership than air service on the same city pairs, even for people who don’t live right next to the train stations.

          • Nathanael Nerode

            If you care at all about trying to reduce sprawl in the sprawlsvilles of Stockton, Fresno, and Bakersfield, however, you really really want the intercity/HSR train station to be DOWNTOWN.

          • calwatch

            Taiwan built greenfield HSR stations outside of the major cities between Taipei and Kaoshiung, leaving the traditional station locations to conventional speed rail. So there is precedent.

  16. threestationsquare / Anon256

    Actually the outer ends of BART are mostly not in highway medians; they tend to be at grade or elevated in their own ROWs. The main exception is the Pittsburg/Bay Point line (Macarthur thru Lafayette and also at Bay Point itself), but that’s at least partly for topographic reasons; the highway and BART line follow the same passes through the hills, and the BART line diverges from the highway in the flatter and more populated area around Walnut Creek and Concord. (There’s not much excuse at Macarthur though, which is indeed pretty unpleasant to access from its surroundings.) Of course there are plenty of other places in the world where topography leads adjacent rail and freeways being the optimal arrangement (the Watford Gap comes to mind). I suppose road+rail bridges and their approaches can be viewed as a special case of this.

  17. Anandakos

    “intercity lines frequently follow highways on one side”

    Actually, the highways followed the rail lines. Nowadays the interstates strike off on their own, but the early network essentially replaced the highways next to the railroads, so they “follow” the rail lines, at least in the Midwest, just a mile or so away so they pass around the small towns strung along the tracks.

      • Nathanael Nerode

        But new-build HSR lines should generally follow the old rail lines! The patterns of development mean that that’s the only way to reach the correct city centers.

        What they do between the cities is another matter, but you see what I mean?

        • Joey

          That’s true of larger areas, but in small cities it may not be desirable or even possible to construct a HSR alignment that reaches the city center. Example: Hanford, CA – the BNSF route goes through the city center but it’s pretty built up and has curves which make it unsuitable for high speed operations. The currently planned route correctly diverts it around the city with a peripheral station.

          There’s another solution which involves routing the mainline around the city with connections to legacy tracks so that stopping trains can reach the city center – used extensively in Italy and less so in France. It keeps the noisy express trains out of the built-up areas while keeping the benefits of a downtown station. This typically only works well when the legacy tracks are well-maintained and electrified though.

          • Max Wyss

            An example would be Avignon TGV, which is now about 10 minutes by bus from Avignon Centre. Eventually, there will also be a rail connector, shortening that time even more.

          • Nathanael

            Well, yeah, that’s OK for places as small as Hanford… but not for places as big as Bakersfield.

          • Joey

            How important is serving downtown Bakersfield? Bakersfield is large but not particularly dense. The downtown area does have some destinations but it requires routing trains through large stretches of developed area. So you’re left with the dilemma: go through the expense and disruption of routing all trains (including expresses) through Bakersfield or build a lot of extra track in order to have a bypass and a station loop. This is different than, say, Fresno where it’s relatively easy to build a station loop that isn’t too long.

      • Max Wyss

        Actually, another example of HSR following a highway is Frankfurt Airport – Köln.

  18. Alex Bailey

    The median running East Link slated to open in 2023 in the Seattle area is an example of freeway running rail done right. The alignment is used because it is easily the cheapest/most direct way to get between two big destinations (downtown Seattle and downtown Bellevue). The two planned freeway running stations are designed to be well insulated from the roadway and both include at least one station entrance that leads to a lidded area. In particular, since I-90 goes into a tunnel just east of the planned “Rainier Station,” the area around the eastern entrance will not be subject to the roar and footprint of an interstate freeway.

    I also think the Red Line on the south side of Chicago is a good example of a successful freeway alignment. Notably, the Red Line greatly out performs, ridership wise, the adjacent and similar Green Line, despite the Green Line having ostensibly better walksheds. And this was even the case back when the Green Line had more competitive frequencies. What I suspect makes the Red Line work is that it readily connects with high frequency east/west bus routes, it’s fast and frequent, a relatively high percentage of its potential patrons are “transit dependent”, and most importantly once you get to either side of the expressway there is both density and a complete street grid, which means that the walksheds for the station are still fairly big.

    These two examples, particularly the first one, show how freeway alignments can make a lot of sense in some cases. But in most other cases BRT using HOV lanes is a much cheaper and more flexible way to provide, fast, frequent and reliable transit out towards suburban areas. Without heavy capital expense, buses can serve different local corridors before combining to form a frequent fast corridor on the existing freeway as they head towards the CBD.

    • Alon Levy

      But the really successful parts of the Dan Ryan branch are the ones farthest south, where there’s no Green Line competition… south of Sox-35th, the busiest stations on the line are 95th, 79th, 69th, and 87th (link), the southernmost four. The inner parts of the South Side underwent a huge population decline, so the South Side Elevated had fewer people to serve: since 1950, Woodlawn’s population has gone from 80,000 to 23,000 in 2010, Washington Park’s from 55,000 to 11,000, Fuller Park’s from 17,000 to 3,000, Grand Boulevard’s from 115,000 to 22,000, and Douglas’s from 80,000 to 18,000. Farther south, the population declines were less steep (link): Roseland, hosting 95th, kept increasing in population until 1980, and has since only lost one third of its population. Armour Square, hosting Sox-35th and Cernak-Chinatown, went from 23,000 to 10,000 people but has regained some population and is now at 13,000, and this combination of immigration and gentrification is boosting transit usage: tellingly, these two stations are seeing large recent rises in ridership, whereas the southernmost four are all seeing declines, especially 95th.

      The lesson here is that high transit ridership is best achieved in intact urban neighborhoods. Urban expressways are likely to contribute to neighborhood decline, via a combination of pollution and noise making the area undesirable, neighborhood splitting, and greater ease of middle-class flight. I’m at least somewhat comfortable saying that the Dan Ryan Expressway helped accelerate the South Side’s decline, which means that even though the Dan Ryan branch has decent ridership, the expressway’s effect on overall city transit usage has been negative.

      • Alex Bailey

        I wouldn’t disagree with the assertion that the Dan Ryan accelerated the decline of those neighborhoods. Freeways almost always have giant negative impacts on the surrounding areas and your point about it accelerating middle class flight (by enabling the suburbs) is a good one. Moreover, I think examples of situations where it is better deal to take the expressway and rail together over no expressway and no rail are rare.

        But despite the negative effects of the Dan Ryan expressway, the Dan Ryan Red Line itself has a lot of ridership, enough to justify very high frequency. In that sense it is a success, even if it can’t remotely negate the deleterious effects of the expressway.

        In this case, the counter factual of interest is not whether the Red Line justifies the Dan Ryan Expressway, but whether it would have been a better use of dollars to ignore building the Dan Ryan L line and instead to have extended the Green Line further south, say for example by continuing west from Ashland & 63rd to the rail corridor about a mile west and then running south from there until you hit 95th and Wood. Thinking about it now, I think that could have been a much better use of dollars.

        But inevitably there is a trade-off between mobility and cost. And the political economy is such that most freeways are not going anywhere anytime soon. Thus, in some cases it makes sense to build at least part of a transit line along a freeway because it is cheaper, more politically viable, and/or makes the line more direct, even if stations along freeway alignments tend to have limited value.

  19. Aussie_fred

    With respect to rail in medians, you might consider the experience of Perth, Western Australia. The biggest problems seems that they under-estimated the ridership potential and now need more train-sets and better parking at the “far” stations.
    As a side issue the traditional rail service (Bunbury) is declining because access to the newer (ie faster) suburban rail (with parking) is offering faster (and cheaper) journeys.

  20. Patrick R

    A city with successful intra highway urban rail lines is Perth in Western Australia. These work with integrated bus interchange stations and to a lesser extent park’n’ride. The stations serve a dispersed population and are widely spaced and the newer trains get up to 130kph, faster than the 100kph speed limit on the freeway. Which is great marketing to drivers, even when not gridlocked. http://en.wikipedia.org/wiki/Transperth_Trains

  21. Pingback: Excuses for High Construction Costs | Pedestrian Observations
  22. Ron Wm. Hurlbut

    Although I agree that there are speed constraints I’d like you to consider this idea as an initial step towards higher speed rail:

    Trains don’t have to be going mach II, they only have to be going 30%-60% faster than the highway traffic.

    Another factor is the North American practice of running big trains infrequently versus small trains frequently.

    Instead of operating loco-hauled heavy rail trains a couple of times per day, it would make more sense to run light rail (Possibly Narrow Gauge) EMU’s more frequently.

    Perhaps a couple of overnight Transporter type trains like the RhB has.

    Running through the Hwy 401 corridor connects all of the major Ontario Cities except *boo-hoo* Ottawa.

    The O&Q route could be utilized at some future date for cross country Mach II service but in the meantime the heavily congested Hwy 401 corridor would get some relief.

    The ROW is already there, electrification would be green and Narrow Gauge would be less expensive to build.

  23. marvin gruza

    In tel aviv israel there was a river,
    rail lines “stubbed” at either end but were not continuous (thus not allowing for an obvious haifa-tel aviv-airport-jerusalem line)

    over the time the river has become more and more channeled until they are now looking into burying it in a pipe
    a major 4+ track rail line now runs through
    a major highway straddle the rail line

    the city has boomed to both sides

    perhaps a key element is that the trains and highway are depressed as opposed ground level or elevated

    • Alon Levy

      This isn’t completely right.

      1. The stub-end station going north, Tel Aviv Center, is south of the Yarkon, and goes back to the 1950s.

      2. The air-rail link was built very recently – the legacy line to Jerusalem doesn’t pass by the airport.

      3. The Ayalon Railway, opened in the early 1990s, has three tracks and not four, and this is a constraint on train capacity. There are plans for reducing the river to a pipe in order to add tracks, but they’re not going anywhere.

      4. The Ayalon Freeway is a major barrier, both physical and psychological. There is CBD development on both sides, but these are separate CBDs in separate areas: the Diamond Exchange in Ramat Gan is just east of Tel Aviv Center, whereas the Tel Aviv CBD is just west of HaShalom, and is centered around an arterial road slightly west of the highway. Crossing the highway is difficult, to the point that at Tel Aviv Center you can buy a cheap ticket just to let you cross on foot, since the footbridge to the tracks is one of the few pedestrian crossings of the highway.

  24. Pingback: Refocusing the Urban Renewal Conversation | Itinerant Urbanist
  25. Pingback: Why do the best transit projects face the strongest opposition? - Mobility Lab
  26. Pingback: Suburban Metros and S-Bahns | Pedestrian Observations

Leave a Reply to letsgola Cancel reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.