Chokepoints as Traffic Meters and Transit Opportunities

My post on the strategic value of chokepoints, using the example of chokepoint-rich Seattle, led to an interesting comment thread at the Seattle Transit Blog.  As often happens, discussion quickly turned to my references to rail and Bus Rapid Transit, as readers argued over whether my real agenda was to advance one of those modes.

As regular readers will know, it’s rarely that simple.  But chokepoints do point to an advantage for Bus Rapid Transit if you’re trying to do things cheaply.  That advantage is that a chokepoint that affects private vehicle traffic is effectively a kind of traffic meter.  In our* work for Seattle Dept. of Transportation in the mid 00s, for example, we noticed that congestion was actually worse at the chokepoints around the edges of downtown than right in the center of downtown.  The chokepoints were restricting the rate of flow of vehicles so that they couldn’t congest the core, exactly the way a system of freeway ramp meters can limit congestion on a freeway.

What that means is that in certain circumstances, if you give transit a clear advantage through a chokepoint, you may be able to get away with giving it less of an advantage elsewhere.  You need to be careful about this.  I’m certainly not saying that transit doesn’t need exclusive lanes through downtown Seattle; clearly it does, and the low congestion we observed is part of why that could be proposed without too severe a traffic impact.  But if you’re trying to create a transit advantage over a wide area of the city, it may well make sense to have a segment of exclusive transit right-of-way that gets you through a chokepoint, beyond which service can branch onto several outer corridors so that all those corridors can have direct access to the transit advantage.

 

Seattle qah chokepoints g.png For example, look again at Seattle’s Queen Anne Hill.  In the previous post, I showed an image of the hill and explained that it creates chokepoints on both its east and west sides, where it presses against bodies of water.

Here’s the same image just zoomed out a little.  The hill is in the lower center of the image.  Just north of it is the Ship Canal.  Downtown is at the bottom of the image, so all corridors linking downtown to the northwest quadrant of the city have to go around this hill.  On the west side of the hill, there’s just one path, which is to go through the narrow space between the hill and Puget Sound (the water body on the left of this image) then north and across the Ship Canal on a single bridge.

That’s two sequential chokepoints (marked in orange on the map) where you can give transit an advantage using exclusive lanes.  But once you’re north of the Ship Canal, you’re into a wider gridded area of medium density, so travel demand spreads out.  A Bus Rapid Transit solution can start by building just the segments that transit needs to get through the chokepoints.  This infrastructure can then be used by many buses between downtown and the Ship Canal, but north of the Ship Canal those same buses can spread out over several branches to completely cover the area.  That way, everyone from the whole area north of the Canal gets the travel time advantage of the exclusive segment into downtown, using a bus that serves their neighborhood directly.  (This is exactly how the Brisbane Busways work.)

This, friends, is one thing that rail transit doesn’t do well.  A rail solution to the same problem has to build rail not just from downtown to the Ship Canal but also on all the branches you want to serve north of there.  Or, alternatively, the rail line has to end just north of the Ship Canal, or follow just one path beyond it, which means everyone else has to transfer to buses at that point.

I am not saying that you should never build rail in this corridor.  You may have capacity or local impact requirements that only rail can meet.  But meanwhile, a good bus rapid transit solution would insist on really good exclusive paths through the chokepoints from downtown to north of the Ship Canal, but would not need as much infrastructure north of there.

So what am I saying?  Imagine, for the sake of argument, that Seattle had one single goal in its transit expenditures, which was to maximize transit’s travel time advantage over driving.  You might embrace this view, for example, if you believed that travel time is, in fact, a dominant factor in determining whether people use transit.

If, hypothetically, you were trying to do that with a fixed budget, you would focus like a laser on Seattle’s chokepoints.  You’d spend your scarce dollars to build Bus Rapid Transit segments through each chokepoint, but not necessarily entire exclusive lanes the whole length of each corridor you want to serve.  And you’d use rail only where you need rail’s higher capacity.  You’d also be likely to use rail in corridors where the demand doesn’t spread out over a large area, the way it does in this example north of the Ship Canal.

Chokepoint-based strategies are especially important if you’re trying to find urgent short-term solutions for your city.  Best-practice transit agencies have short-term improvement programs running concurrently with their thinking about long-term major projects.  Sometimes, these short-term investments add up to a really clear advantage for transit that, if properly marketed, can make a big difference.  As Seattle readers know, the corridor I’ve used here as an example does have a “rapid bus” program in the works.  My point here is not to comment on the details of that program but just to point out the way a city’s geography creates opportunities.  Look for the same opportunities wherever you live.

* “our” refers to Nelson\Nygaard Consulting Associates, with whom I was associated at the time.

31 Responses to Chokepoints as Traffic Meters and Transit Opportunities

  1. Danny January 6, 2010 at 5:07 pm #

    Great post!
    I think the chokepoint example also explains why the XBL on the lincoln tunnel has provided such high capacity from bus transit, considering that it is one of the fastest ways into Manhatten given the low capacity of all the river crossings.

  2. Jason January 6, 2010 at 6:07 pm #

    “On the west side of the hill, there’s just one path, which is to go through the narrow space between the hill and Puget Sound (the water body on the left of this image) then north and across the Ship Canal on a single bridge.”
    The “narrow space” is actually between the hill and Magnolia, not the Sound. The corridor is known as Interbay and the arterial is 15th Avenue. Your point about the chokepoint still holds, of course.

  3. Cascadian January 6, 2010 at 6:35 pm #

    Westlake Ave (on the east side of Queen Anne Hill and the west side of Lake Union) and then the Fremont Bridge provide another route than the one you described (Interbay and the Ballard Bridge). Both get you around the hill and across the Ship Canal. There’s also the Aurora Bridge (Highway 99) that takes a much higher route on the east side of the hill.
    I guess your point might be that these two routes don’t go to the northwest quadrant of the city, defined by streets with the NW directional. But in practice if you’re trying to get across the Ship Canal west of I-5 you might consider any of these routes. 99 tends to be used for any trips north of the canal, sometimes including parts of Ballard and Fremont. The lower bridges really only serve the local areas. And even some of the NW streets are closer to the Fremont Bridge than the Ballard Bridge.
    Still, that’s only three options for the NW side of the city, along with 3 others further east (I-5, University, and Montlake bridges.)

  4. Jarrett at HumanTransit.org January 6, 2010 at 6:40 pm #


    You're right that corridors east of 15th NW can be fed through Fremont, which still leaves at least three (15th, 24th, and 32nd NW) that logically feed over the 15th NW bridge.    In referring to the whole NW quadrant of the city, I think I was saying that QA Hill defines chokepoints for the entire area, including the one you mention on the east side of the hill. 

  5. Pantheon January 6, 2010 at 6:59 pm #

    There has been a lot of debate over what to call the decade you refer to as the “00s”. I understand your reticence to use one of the proposed monikers, as the vociferous debate and entrenched views on this make the bus/rail debate look like a game of cricket.
    Nonetheless, I prefer the term “noughties”. It is the most linguistically accurate, but more importantly the most fun.
    If you want to really go out on a limb, you could use Slate’s new suggestion: the “Uh-Oh’s”!
    http://slate.msn.com/id/2239014/

  6. EngineerScotty January 6, 2010 at 7:26 pm #

    I like the “oh-sh*ts” myself, but then I’m just decade-nt.

  7. rhywun January 6, 2010 at 7:45 pm #

    I prefer “aughts” myself. Carry on.

  8. Anandakos January 6, 2010 at 8:17 pm #

    Jason,
    Unfortunately, the Rapid Ride bus is proposed to go up Mercer Way RIGHT into the soup of Lower Queen Anne. Gag!
    There is insufficient street capacity on either Mercer Place (two lanes and narrow ones at that) or the Queen Anne, First Avenue North couplet to give a bus lane. It might be OK to turn off Elliott at Harrison to give some access to LQA and Seattle Center to the Rapid Ride; the bus would miss the worst of the mess right at Queen Anne and Mercer. But not Mercer Place/West Mercer.

  9. Adam Parast January 7, 2010 at 1:37 am #

    Jarrett I agree with you completely from a technical perspective but the simple fact is that Metro BRT solution, “RapidRide”, is likely to be just about the lowest form of BRT possible. For example mid-day frequency will be 15 minutes, no new bridges will ever be built just for the bus, business access lanes are only used in the peak direction, and less than half of the stops will actually have station level amenities.
    As I have written before on STB this is the problem with BRT. (http://seattletransitblog.com/2009/12/09/the-lowly-bus/)
    You’re talking about a level of bus based transit that will never exist in the corridor because Metro simply doesn’t have the money or the vision to accomplish it.
    Metro is pursuing Rapidride, I believe, mostly because it is getting free TSP from the city of seattle (via bridging the gap) and free buses from the FTA as well as money for stations and other communication expenses. If there wasn’t any of this money Metro won’t do it.
    So while I said I agree theoretically this is another example of how BRT over-promises and under-delivers… at least in the US.

  10. Jarrett at HumanTransit.org January 7, 2010 at 1:39 am #


    Thanks, Adam.  As I said, I was not commenting on Rapid Ride in particular, just exploiting its geography to talk about the principle. 

  11. Alon Levy January 7, 2010 at 3:54 am #

    The non-US rail solution would be to terminate the trains north of the bridge and build a bus terminal right next to the station. In the German-speaking world the transfers would be timed.
    BRT’s main advantage in Seattle is that it can use existing streets in downtown. But even then it would need at a minimum signal priority to be able to run at decent speed. Otherwise it would get stuck behind too many signals; without signal priority, buses can be slow even when the streets are not congested.

  12. Joshuadf January 7, 2010 at 7:35 am #

    It’s probably worth mentioning that the Seattle city council president wants to extend the Seattle Streetcar from South Lake Union to Fremont, probably on the east side of Queen Anne Hill (Westlake Ave N). It remains to be seen whether the rest of the council and/or the mayor will play along, though.

  13. Casey Hildreth January 7, 2010 at 1:34 pm #

    Thank you Jared for writing so eloguently on the value of congestion and the importance of transit/car differentials as a factor in mode split.
    I would like to point out that the geography and physical barriers in Seattle are indeed a major provider of chokepoints, but equally so are the abundance of shifting grids and multi-leg, angled intersections. As you know, these are what contribute to the chokepoints in downtown along Denny, Westlake, and south near King Street Station – not hills or waterbodies.
    Capturing public space at these intersection locations – as is being done in San Francisco, New York City, and other places (Sandy Blvd in Portland, Mass and Main in Boston, H Street, et al in DC) – has already entered the U.S. urban design lexicon in a short timeframe. Expanding these open space concepts to include higher transit priority through vehicle chokepoints – think mini transit-only diagonals – should be a priority for the transportation profession as well.

  14. anonymouse January 7, 2010 at 4:38 pm #

    A rail based solution would involve a transfer north of the chokepoint, trading the higher speed of rail (55 or 70 mph in a grade-separated corridor) for the delay incurred by a transfer. The high speed and reliability of the rail link would make the area around its terminus desirable for residences, and all those people making the transfer would make it a good retail location, since once they get off the train on the way home, they might as well stop by a store before getting on the bus to go home. The transfer point thus becomes a secondary center, something you don’t necessarily get with just plain BRT.

  15. Jarrett at HumanTransit.org January 7, 2010 at 9:25 pm #


    All true except that there would be on difference in maximum speed for rail vs bus in a grade separated corridor of this length with this station spacing.  Both would top out around 55 mph, and differences in travel time due to acceleration would be slight. 
    Transfer locations make excellent retail nodes, as you point out.  This is not necessarily a sufficient reason to force a transfer if a less expensive solution provides the same mobility without that requirement.

  16. Pantheon January 7, 2010 at 10:39 pm #

    To be a little more blunt, transit agencies are not developers. It gets a little tiring to hear people constantly talk about the development consequences of transit, as if the actual act of moving people around were merely a secondary consideration. It is a pollution of the mind, a pernicious and often unconscious way of thinking about transit that needs to be stamped out.

  17. EngineerScotty January 7, 2010 at 11:10 pm #

    Back in the old days, transit operators built theme parks to attract riders to streetcar lines.
    Now its condos. 🙂
    TOD is useful when it’s in ADDITION to (and as a result of) the act of moving existing riders and serving . Westside MAX has lots of TOD, but connecting Beaverton and Hillsboro to Portland is a worthy transit goal even if Orenco didn’t turn into a pile of tacky townhouses. 🙂 If a transit line passes through an existing brownfield or greenfield, turning it into high-density housing and getting folks out of car-dependent living isn’t a bad idea.
    OTOH, transit lines whose primary purpose is to serve speculative development (the SoWa streetcar extension comes to mind) are a bit more dubious.

  18. Placemaking Institute January 8, 2010 at 1:03 pm #

    Higher transit use is not an input but rather an output in economic development and urban redevelopment; the benefits of great mixed-use places will outweigh all the negative externalities associated with cars and, while the effects of such developments may not be immediately felt, in a couple of generations it will be – Just like it took the pro-road/autocentric lobby a couplathree generations for their myth to be firmly inculcated. The arguments used by those who argue against our broadening our modality? Could be applied to roads at the Interstate inception in the mid-50s.

  19. Pantheon January 8, 2010 at 5:13 pm #

    I don’t understand what you are trying to say. What do you mean by input and output? Negative externalities? Broadening our modality? I regret that while I am relatively fluent in English, I am unfamiliar with the rather peculiar tongue of faceless committees and bureaucratic organizations.

  20. Jarrett January 8, 2010 at 8:47 pm #

    Placemaking Institute isn’t all that bureaucratic. He’s saying that it will may take a couple of generations before sustainable urban development and transport become standard, just as it took a couple of generations for the car+sprawl model to become standard.

  21. Placemaking Institute January 9, 2010 at 9:48 am #

    Exactly, and thanks, Jarrett, for untying my purportedly rather peculiar tongue. Unless a bureaucracy can consist of one person, the Placemaking Institute is most definitely not at all in any way bureaucratic; in fact one of its intents is to satirize such faceless bureaucracies, at times by co-opting their language.

  22. Pantheon January 9, 2010 at 3:31 pm #

    Rest assured that I never meant to imply you were bureaucratic, or faceless. Only that you speak in the same tongue as those who are. And I do very much appreciate satire, even if it must come at the expense of clarity.

  23. Placemaking Institute January 10, 2010 at 10:13 am #

    But I took being labeled a faceless bureaucracy as a compliment!

  24. Nathanael January 14, 2010 at 4:36 pm #

    “This, friends, is one thing that rail transit doesn’t do well.”
    Well, given
    (1) that you have a fairly uncongested pre-fabricated mess of asphalt roads to spread out onto;
    (2) that you don’t have a pre-fabricated mess of rails to spread out onto;
    (3) that a humungous pile of buses is actually sufficient capacity for the corridor and doesn’t overstrain the busway.
    If (3) is false, you do better to run a rail line across the bottleneck until you hit a wide-open area where you can build a major transfer center.
    In suburban London, arguably (1) and (2) are both false and it may be easier to run spread-out rail services than spread-out bus services!

  25. Nathanael January 14, 2010 at 4:44 pm #

    “I think the chokepoint example also explains why the XBL on the lincoln tunnel has provided such high capacity from bus transit, considering that it is one of the fastest ways into Manhatten given the low capacity of all the river crossings. ”
    It certainly does!
    The funny thing, however, is that New York City is full of examples of doing the same thing with rail transit, even *more* efficiently. Consider the tunnels and bridges between Brooklyn and Manhattan: services spread out, merge for the tunnels, and spread out again. For a larger-scale example, consider the massive Penn Station project, connecting tunnels to New Jersey to tunnels to Long Island. Of course, failure to use through-running has lost some of the potential of that one.
    These were all built in an era when there was a massive network of rails on either side of the bottleneck, in contrast to today’s “massive network of overbuilt roads” present in many cities. The very same chokepoint argument demanded rail construction, not road construction.
    From this we see that making use of existing infrastructure saves money. Well, duh.

  26. Placemaking Institute January 16, 2010 at 11:06 am #

    The sad thing is that many, many “natural” transportation chokepoints within this country once had railway throughputs – But they were ripped out.

  27. Ari March 9, 2010 at 9:24 am #

    @EngineerScotty
    You said:
    “[T]ransit lines whose primary purpose is to serve speculative development … are a bit more dubious.”
    Most transit lines in this country are based on lines which were built as streetcars in the early 1900s, and most streetcar companies made their money by buying land along the line and then, once there was transit, selling it at higher prices. That’s speculative development. Of course, once government takes on this role, it’s a bit more muddled, but there is a long (if not all-that-recent) history of transit and development.

  28. EngineerScotty March 9, 2010 at 10:44 am #

    Ari,
    Certainly; OTOH with the early 20th century streetcar lines, the same entity which engaged in the real estate speculation also financed the line construction.
    What many object to with TOD is the appearance of public money being used to benefit private developers. Of course, in most cases, the evidence of outright corruption is not there–there is a defensible mobility reason for putting the line where it is, and the involvement of the adjacent property owners, who can be taxed to finance construction (how much of the Portland Streetcar was funded), is often a win-win.
    With some projects, though… net mobility benefits are hard to discern, or depend on the new TOD becoming an occupied and vibrant community. When this fails to occur, the line becomes–in retrospect–a boondoggle, and too many of these and people start to question the competence and honesty of the public officials responsible.

  29. Alon Levy March 10, 2010 at 2:37 pm #

    Ari: yes, and those developers neglected the streetcars once they’d finished selling the land. Many lines were dying even before the National City Lines conspiracy.
    On mainline railroads it was even worse. Many railroads, especially in the western US, got land for cheap by refusing to haul farmers’ product, forcing them into bankruptcy. Southern Pacific engaged in so much extortion that popular writers of the late 19th century portrayed it as an octopus. This corruption then became one of the major selling points of the good roads movement: rural populists saw good roads as an alternative to the railroads, or at least a way of reaching multiple railheads, creating more competition.

  30. EngineerScotty April 21, 2010 at 7:43 am #

    I note another interesting use of chokepoints (in Hong Kong) here–essentially, since all transit services have to go through the tollbooths at bridges in tunnels in HK, the tollbooths are turned into transit centers.

  31. Frosty February 10, 2012 at 1:35 pm #

    One of the biggest Seattle chokepoints not mentioned in the essay or comments is I-5 under the Convention Center and on either side of it. There are seven lanes of traffic and entrances and exits that require weaving at up to 60 mph. This area extending from the Ship Canal bridge south to the exit to I-90 is used by local traffic to get around the core of the city. Traffic backs up on the freeway and on city streets due to gridlock. It isn’t clear to me that it’s possible to widen the freeway under the Convetion center.
    Another obstacle is southbound I-5 north of the Ship Canal. Cars unintentially slow as the grade increases over the bridge leading to backups reaching 145th St. at times.