Thursday, August 19, 2010

The problem of wide corridors

One longstanding characteristic of public transit, is its linear nature.  Transit systems are generally (and ideally) organized in lines, on which some vehicle travels progressively from point A to B, and then in the reverse direction.  Not all such lines are geographically straight, of course, and many systems also feature loops with no discernible endpoints--but even in those cases, the vehicle makes progress towards some goal or direction, stopping along the way to let folks on or off.  Vehicles cannot be in more than one place at a time, nor do they (Amtrak's Empire Builder notwithstanding), split into two to serve different sets of destinations simultaneously.  Lines can branch, but you lose frequency on the branches--individual busses or trains running on the line can only serve one branch or the other, not both. (see Jarrett's excellent presentation for more on branching, and numerous other topics). 

But what happens when you have a wide corridor--a geographic area which is linear in shape, but where the key destinations don't lie in a convenient linear arrangement?


67 Map, courtesy of TriMet
Consider TriMet's 67 Jenkins/158th route, driven by our good buddy Al (who wants it known that he doesn't speak for TriMet, or his union, or anyone else but himself).   This route run between PCC Rock Creek and Beaverton Transit Center.  The primary routing is kind of zig-zag (the as-the-crow-flies direction is diagonal, but the streets for the most part are not), but that's close enough to linear for our goals.  However, the route contains two backtracking deviations, or as I prefer to call them, "cherries"--places where the bus leaves the main street, goes some distance to reach an important destination, turns around and backtracks, and then rejoins the primary service corridor.  One of these cherries, serving the Cornell Oaks business park, will be eliminated in September (a move praised by Jarrett).  The other, a shorter-stemmed cherry serving the Merlo/158th MAX station, remains in place. 

Or consider the proposed Barbur Boulevard corridor for a new MAX or BRT line.  Between downtown Portland and Tigard, there are several highly-important destinations (Tigard TC, Washington Square mall, PCC Sylvania, Barbur TC, Hillsdale, OHSU-Marquam Hill)--all of which, except for Barbur TC, aren't on Barbur.  A simple line parallel to the existing boulevard misses most of the action in the corridor--which is why various proposals for tunnels get floated about.  (I have no objection in principle to the proposal, but I seriously doubt that TriMet can afford it).

Or consider the Washington Square WES station.  Of course--there is no such thing; the nearest station to the mall is the Hall/Nimbus station (which isn't even located on Nimbus Avenue).  The mall is located about 2000' to the east, across a busy freeway and a major thoroughfare.  Connecting bus service is actually pretty good (between the 76 and 78, and the occasional 43, up to six busses per hour connect the Nimbus Avenue stop with the mall--that's more frequent than WES itself), but the transfer to the bus lines plying Hall Boulevard is a bit inconvenient, especially in bad weather.  WES was constrained to operate on the existing freight tracks, which lie on the opposite side of the freeway--but the mall is one of the most important destinations in the corridor. 

Wide corridors--sets of destinations which are clustered around some linear corridor but not sufficiently close enough to it (or across some barrier)--cause lots of problems for transit.  Serving them adequately requires diluting available service hours over greater numbers of route-miles--resulting in a net decrease in service quality.  In some cases, rapid transit lines may not be practical due to dispersed demand--and in the worst cases, the only sorts of services which are practical are inefficient "social service" routes.

What to do, what to do

Some possible solutions to wide corridors are discussed below.  These aren't solutions in the sense that they solve the fundamental problem; however they are different ways to provide service in a difficult situation.
Courtesy of

  • Longer routes.  One way to serve the major destinations in a wide corridor is to lengthen the route.  The "cherry" mentioned above is one case; serpentine routes such as a proposal for Maryland's "corridor cities" are another example.  Either way, lengthening the route to cover the greater distances imposed by dispersed destinations effectively slows down the line, making it less attractive to riders.
  • Splitting and branching.  If most of the trips are between destinations along the corridor and destinations elsewhere, branching may be an option.  Branching reduces frequency in the branches, of course, so doing so may also make the service less attractive; it makes travel between branches more difficult.   If the line requires fixed infrastructure, such infrastructure needs to be built on both branches.  One useful option here, though, is "open BRT"--the trunk can often be built to high busway standards, with the branches running in mixed traffic--if the branches are short, this arrangement often results in service of reasonable quality.
  • Transferring.  One reason "grid" or "fishbone" arrangements are useful is that intersecting lines can be used to provide service to destinations which are slightly off-track, while keeping the trunk line straight and fast.  This depends a lot on the quality of the transfers, obviously--if getting the last mile requires a 20-minute wait on the side of a road in the rain, this is not an attractive option.  If, on the other hand, transfers can be timed and waiting occurs in a protected space, this is an excellent option.

Mid-levels escalator (HK), image courtesy Wikipedia
  • Dedicated connecting service.  This is a subset of transferring, but the previous paragraph assumes a generic line that happens to pass close to a point of interest, but serves much else beyond.  A dedicated connecting service, on the other hand, provides point-to-point service between a station on the trunk line, and a given destination.  Such services may take the form of shuttle busses (such as the shuttles running between PCC-Sylvania and Barbur TC), a dedicated people mover (the Portland Aerial Tram functions in this fashion; even though its primary purpose is not connecting the Marquam Hill campus to the Portland Streetcar), or even a special-purpose transit line (such as the "Disneyland Subway" in Hong Kong).  With a dedicated service, there's ample room for creativity--the mode used might be outside the norm of services typically provided by the transit authority--and might not be funded by the transit authority anyways.  The dedicated service might even be passive--a skybridge or tunnel where people walk, for instance, but one which lets users avoid weather, obstacles, traffic, or difficult terrain. Generally, a destination has to be fairly major to justify this sort of infrastructure, but things like gondolas, shuttles, elevated walkways, public elevators and escalators, moving sidewalks, and such are all pressed into service to connect a transit line to the places people ultimately want to go.  (Many of these things additionally benefit pedestrians who aren't using the transit line as part of their trip).

Of course, as numerous writers cited above have pointed out, the best way to deal with this problem is to avoid it.  Good land-use planning can avoid widely-dispersed destinations--and result in land-use outcomes which are easier to provide service to.  The benefits accrue to well beyond transit--many other public goods, from police and fire protection to utilities, are cheaper and easier to provide if confined within a smaller space (It ought to be noted that automobiles, on the other hand, frequently benefit from dispersion--mainly due to their need for parking when not in use).  But there are many existing wide corridors which need service; and many popular corridors can be victimized by their own success--attracting development which widens the corridor as the closest lots become scarce and/or expensive.


  1. That cherry for the Merlo MAX stop is truly annoying for through passengers who aren't transferring at the Merlo stop, but it used to be even worse. Previously, the buses would go past the circle they currently use, and go into the Merlo yard to turn around. It was awful. So "annoying" is an improvement.

    It's about 900 feet from the MAX stop to the Jenkins-158th intersection. It doesn't seem like such a hardship to ask people to walk that distance to transfer between the 67 and MAX.

  2. dont forget multnomah village from the portland-tigard corridor or even ohsu.

    southwest portland bus service needs some real study for bus improvements, its not an easy area to serve. the street grid or lack there of is obviously one of the main culprits now for the poor service. sidewalks are needed big time. i've been thinking that the two branch legs of the 51 could be extended, one to multnomah village and the other to hillsdale which arent much further than their current terminus'.

    you have pretty extensive branching in SW, look at all the buses that stop in hillsdale then they begin to branch out from there.

    the 1-vermont is pretty bad with that large one way loop that runs one way in the morning and the otherway in the evening.

    somewhere i saw an old trimet map from the 70s that had even more little branches in SW portland particularly in the area around PCC, dont know anything about the service levels of these buses.

    dont get me started on erik "mr.-trimet-loves-rail-at-the-expense-of-buses" halstead's subway proposal. he has a lot of gall to propose that and take it to the level of the oregonian.

  3. @belmont--

    How many people generally transfer between the MAX and the 67 at the Merlo stop? The 67 also stops at Beaverton TC--it takes the bus 10-15 minutes to get between Beaverton and Merlo; it takes the train about seven.


    While I was surprised to see Erik propose a subway, given his longstanding line of argument, the proposal needs to be considered on merits. Given the terrain, a subway might be a cost-effective way of doing it--and would have the advantage of being able to serve the most important destinations along the way, and have the performance necessary to make mass transit to distant suburbs like Sherwood or Tualatin an attractive option. The terrain is generally elevated, and thus water table issues are less likely to come into play. A grade-separated rail line would probably radically alter the bus topology; I imagine that many of the existing lines that now run down Barbur for a stretch would be feeders.

    That said--if MLR, a surface line, is too expensive...

  4. I couldn't agree more about the need for a tunnel in that corridor. Barbur really is the wrong place for LRT or any enhanced transit like BRT. All that Barbur serves is a forest (btwn Burlingame and Lair Hill) and endless strip mall autopia between (Burlingame and Tigard and beyond). I think the 44 corridor is a better corridor for a SW corridor high capacity transit line... it hits Hillsdale, Multnomah Village, Barbur TC (very close) and PCC. Just about all the riders on the 44-Capitol Highway get on or off at PCC, Barbur TC area, Multnomah Village, Hillsdale or Downtown... there are few getting on/off at intermediate stops that aren't one of these 5 town/regional/destination centers.

    I'd very much like to see a tunnel from Downtown Portland to Barbur TC (where it could surface) that has Washington Park-like subway stations at OHSU, Hillsdale and Multnomah Village. I think this is the most logical way to go. If I'm not mistaken the Robertson Tunnel wasn't all that expensive, I have it at about $133 million for 3 miles of track including 1 underground station... if thats the case I would bet it would be by far the cheapest option (looks to be about 4.5 miles), cheaper than surface along Barbur (and also without all the ranting and mitigation from Barbur autopia about the loss of driveway access). I think it is subway tunneling under city streets that is expensive, due to the nature of the construction digging under buildings and streets and dealing with all the buried utilities and adjacent property mitigation. Boring a tunnel under a hill seems like it would be rather cheap and with few conflicts.

    As far as where the line would go after it surfaces at Barbur TC (under my proposal), it could either go down Barbur to Tigard TC and beyond by surface, the road is quite a bit wider in this stretch, or go up Capitol Hill on Capitol Highway towards PCC also on the surface with a station for PCC then cut over and rejoin Barbur Blvd to go to Tigard TC. If you look at an aerial map, the backside of PCC is very close to Barbur Blvd so maybe there is a way to connect in.

  5. All depends on how much it would cost. And on whether Barbur is interested in re-developing at all. (I think that Barbur north of the Tigard interchange ought to be handed over to the city or county, and no longer maintained as an ODOT facility; there are other roadways of regional importance, such as Roy Rogers or River Roads, that might benefit instead from state maintenance should the authorities desire a trade).

    Will be interesting to see what they come up with.

  6. Given Portland's small grid street system and inability to support trains over 2 cars; a subway is the next logical step through the downtown area.

    The Blue Line is downright packed during peak hours. I just cringe at the cost of doing this, but I'd much rather have a subway line through DT than a dedicated transit bridge.

    Why spread yourself thin providing HCT along Barbur and 99W in Tigard? ROI is going to be minimal given its auto based landscape. Provide solid express buses during peak hours and timely buses thereafter and call it good.

    Seeing the curvature of the Earth is probably a sign that HCT like LR is not going to work well -- and Barbur Bvld. and Tigard definitely has that covered!

  7. A downtown subway implies that a high priority is placed on expediting crosstown trips. While in the long run, that's an excellent idea, in the short run that's probably not where you want to put money.

    For a subway to work, quite a few stops will need elimination (the subway idea discussed in the Metro HCT plan only had one stop between portals at Rose Quarter and Goose Hollow--a Pioneer Square stop). You could probably improve the performance of the existing surface alignment by getting rid of stops. Kings Hill station has long been recognized as a political concession. The stretch along first doesn't have significant cross traffic to deal with (passing under the Morrison and Burnside bridges) and is thus mainly stop-constrained. The E/W couplet between I-205 and 1st, OTOH, spends a fair bit of time waiting at lights.

    Of course, getting rid of stops will prove politically difficult...

    Regarding the Barbur corridor: I think the shape of the corridor will depend greatly on what communities further out want. The City of Portland, I think, will happily upzone Barbur within the city limits. Tigard, on the other hand, may be a tougher issue--the traffic there sucks, but a lot of the locals want to see the highway widened, not mass transit installed.

  8. yeah but a downtown subway would be like 10-20 times the cost of the willamette river bridge.

    in addition to longer trains and more capacity, a subway allows you to go fast between stations too, even if there are more or less the same number of stations as now with a subway. and of course the biggest time savings would be between rose quarter and union station/old town by avoiding the steel bridge.

    it seems to me a subway would go under holladay with a single station for lloyd district, a single station for rose quarter/convention center, then under the river, then run north-south under 5th or 6th stopping at union station, burnside, pioneer courthouse square, city hall, portland state. a branch off this main north-south line around city hall running east-west down either jefferson or columbia to goose hollow with a station at art museum/sw10th/sw11th.

    i say avoid barbur altogether, i couldnt think of a street worse for improved transit (except maybe mcloughlin). there is nothing on barbur (except for the barbur TC) that is any benefit for transit or TOD. all the 20 minute neighborhoods and destinations are in a row along capitol highway... link them with a robertson-like tunnel. and dont forget pretty much the whole south side of barbur is cut off by I-5, so you'd have poor connectivity to the neighborhoods to the south of barbur.

    BTW the new TIP for 2011 mentions BRT (sounds like lite-BRT) in the Downtown to Gresham corridor (via Powell to 122 to Division)


  9. say avoid barbur altogether, i couldnt think of a street worse for improved transit (except maybe mcloughlin). there is nothing on barbur (except for the barbur TC) that is any benefit for transit or TOD. all the 20 minute neighborhoods and destinations are in a row along capitol highway... link them with a robertson-like tunnel. and dont forget pretty much the whole south side of barbur is cut off by I-5, so you'd have poor connectivity to the neighborhoods to the south of barbur.

    I know that some will argue that this is an excellent reason to send MAX down Barbur--it's an opportunity to transform the place into a rather ugly sprawlevard into something nicer. I'm not sure I really agree with that notion, but it will be on the minds of many. Plus, Barbur is amenable to surface transit if a tunnel is impractical (especially a tunnel that continues south of Hillsdale/Burlingame); whereas Capitol, not so much.

    It will be interesting to see what the priorities for the Barbur HCT corridor are: Are they

    * Faster/more efficient service between Tigard and Portland?
    * Better service to poorly-served major destinations such as OHSU or PCC?
    * High-capacity rapid transit which makes longer commutes from points south of the Tualatin River practical on transit?
    * Land-use improvements on Barbur?

    How these questions are answered will go a long way in determining what sort of transit (BRT, surface LRT, a subway) is appropriate for the corridor. Obviously, a subway wouuld be great; but also expensive.

  10. i agree if there was something to build on with barbur but when its endless sprawl of tire salons and lube shops, KFCs, 40s motels, used car lots, and every building low density set back with ample parking, i think its a complete lost cause and waste of time and money to think it can be made into TOD.

    i strongly think they need to focus TOD and transit in places that already have a DNA suited to density, pedestrians and transit.

    interstate ave had/has much more to work with for transit and TOD than barbur... more urban, street grid, sidewalks, some historic up-to-the-sidewalk storefronts, fewer setbacks, dead businesses (like the old motels), walkable old houses right behind interstate, an already transit heavy neighborhood (N, inner NE), etc. Barbur has none of that.

    if you want to bring transit in and completely make over a neighborhood into a denser urban area, i think the springwater corridor is the place for it, given the private ROW directly adjacent to properties to the north and south and not cut off by 'traffic sewers' like most other places. we need places close to Portland where TOD can surround the station on all sides and not just on one side of the station where it also has to contend with the noise, ugliness and obstruction of a freeway or major road.


Keep it clean, please