Archive | August, 2012

The Past, Present, and Future of the Portland Transit Grid

Disclaimer: Zef Wagner is currently a Service Planning and Scheduling Intern at TriMet. The views expressed on this website are his own and do not necessarily reflect the views, plans, or policies of TriMet.

As Jarrett Walker notes in a recent post, this weekend marks the 30th birthday of the Portland transit grid system that enables crosstown travel without always having to go through downtown. This occasion seems like a great opportunity to discuss the past, present, and future of this incredibly important aspect of our transit network.

The Past

I don’t really have a whole lot to add about the formation and impact of the grid beyond what you will find over at Human Transit, so go read it now if you haven’t already. While MAX seems to get all the attention when people think of Portland’s transit innovations, the grid is arguably a much bigger deal. After all, most North American transit agencies are still stuck with radial networks that don’t make a lot of sense in today’s multi-centered or completely decentralized cities. The Portland eastside grid, established in 1982, has since inspired many other transit agencies (especially those in cities blessed with a grid street layout) to follow suit, among them Translink in Vancouver, BC, and Metro in Los Angeles.


1970 TriMet bus network

I really enjoy some of Walker’s examples of how difficult the TriMet system used to be for crosstown travel. It’s incredible that there was only one real north-south line in the entire SE quadrant (along SE 39th Ave), and even that was split into two lines at Hollywood. There was no continuous bus line at all along SE 82nd Ave, now home to one of the most consistently high-ridership lines (the 72) in the whole system. In any case, we should all be thankful that the grid system was established in the face of much opposition, and I echo Walker’s call for people to take the time to thank the planners who made it happen. (Full disclosure: one of them is my supervisor at TriMet, Ken Zatarain).

The Present

It is fitting that this Labor Day weekend marks the anniversary of the grid, because it is also when we will see a few new additions and enhancements to the grid. Even though the current restructures were implemented in the context of overall service cuts, they also show that even after 30 years there are still opportunities to make the grid work better.

Starting this Sunday, we will have a couple new crosstown connections that continue the work of building the grid:

The new Line 70 (click for map) will combine the current 70 and 73 into a new north-south crosstown running from Milwaukie all the way to the NE 33rd Ave/NE Columbia Blvd area. There has also always been a very wide gap in SE-to-NE service stretching from SE Grand Ave to SE 39th Ave. This won’t fill the gap completely, but it will be a big improvement and will create a nice connection between the inner SE and the Alberta neighborhoods. Frequency on NE 33rd will see a boost from current levels as well.

The new Line 87 (click for map) will also combine two previously short and disconnected lines, the 82 and 87. The new line will provide a complete north-south crosstown on 182nd Ave for the first time, with east-west segments to Gresham and along Airport Way, and another north-south segment down 102nd Ave to Gateway. The result is a rather strange-looking zig-zag route, but if you think of it as a few crosstowns stitched together, it makes sense as part of the grid.

The challenge of creating pure north-south crosstowns in East Portland and the eastside suburbs is that there just isn’t much space between Sandy and Powell for a viable stand-alone bus line. The 87 manages to be a long and useful route while also filling in a gap in continuous north-south service. The Airport Way employment district will have midday service for the first time and will be much better connected to the larger system. The bad news is it will run infrequently and only on weekdays, limiting its usefulness as part of the overall grid network, but over time this will hopefully improve if it performs well.

Even in the western suburbs and SW Portland, where a grid is challenging due to geography and land use patterns, we have seen elements of a grid slowly taking shape. This weekend’s service changes include some improvements to this grid as well. One very successful example is the 76 and 78, which have a combined segment from Tigard to Beaverton that acts as one of the few north-south crosstowns in the area. The 76/78 together are about as frequent as any of the “Frequent Service” branded lines, and weekend trips are being added this weekend to boost frequency even more.

Another change is combining the 47/48 with the 89 to create a pair of east-west lines from Hillsboro to Sunset Transit Center via Tanasbourne. The segments of the 47 and 48 that currently run to Willow Creek Transit Center will be deleted, but the 52 will remain to provide that north-south service. This is exactly the kind of change called for in designing a grid system: eliminating redundant service, simplifying routes, and encouraging connections.

Finally, while it is not a new part of the grid, we should celebrate the extension of frequent service on the 9-Powell out to East Portland and Gresham. Previously, every other trip ended at 92nd & Powell, but now East County residents will have access to more frequent service, which means much easier connections to other lines in the grid.

So it’s all good news, right? Not really. While this weekend’s service change does a great deal to strengthen the grid in terms of network design and even adds trips in a few places, we still lack the overall network frequency needed to make the grid function well. The whole foundation of putting in the 1982 grid system was to replace a lot infrequent radial service with frequent crosstown service, encouraging people to connect from one frequent transit line to another rather than always expecting a one-seat ride. This is only possible with frequent service (at least every 15 minutes, preferably every 10 minutes), because otherwise people end up waiting so long to connect that it is no longer worth it. Right now only a handful of bus lines get better than 15-minute headways, and usually even that is only during peak rush-hour times.

We need a return to real frequent service (at least every 15 minutes all day, 7 days a week) for our core grid system to really work for everyone and for multiple purposes. Transit needs to be useful not only for M-F 9-5 workers, but also people working in the service industry on weekends or in the evenings. It needs to be useful for work and shopping and leisure activities. All of this requires high frequency during more than just rush hour for our high-demand transit lines.

Another problem is the lack of coordination between the more-frequent service in the more dense parts of the region and the less-frequent service out in sparsely-settled suburbia. Some elements of the grid, like the 87 or the new 21 on outer Sandy, might never warrant truly frequent service, so ideally they would be timed to connect riders easily to the frequent grid.

TriMet used to have a timed-transfer “pulse” system at suburban MAX stations, but the system was phased out because it was expensive to operate and because frequencies were increased on many of those lines. Pulsing is very expensive because it depends on very high reliability, which in turn requires a great deal of layover and recovery time in the schedule. An alternative to the pulse would be to try to match up frequencies so that, for example, a 30-minute-headway line connecting to a 15-minute-headway line would be scheduled to arrive a few minutes before every other connecting trip.

The Future

So where do we go from here? Besides the need to restore and expand the frequent service network, where do we still have gaps in the grid that need to be filled? Here are some ideas (since I don’t have Adobe Illustrator handy, please refer to this handy interactive TriMet System Map to follow along):

One clear candidate is that pesky gap in inner SE Portland. Even after this weekend we will still have well over a mile (from 12th to 39th) between north-south lines in one of the most dense and transit-friendly parts of the city. Part of the reason this gap exists is that there is also a gap in the arterial street grid. There is an odd discontinuity there, with 33rd, 28th, 30th, and 26th all acting as the main arterial at various points. For this reason, it is probably impossible to design a good bus line through that area.

SE 20th Ave, however, could work as a north-south line. What might such a line look like? One idea would be to have the 10 run on SE 20th Ave after going through Clinton, rather than going to downtown via Ladd’s Addition. This would undoubtedly anger Ladd’s residents, but they all live within 5 minutes of at least one frequent line to downtown anyway (the 4 and 14). I fail to see why they need extra buses cutting through the neighborhood, and cyclists might appreciate getting buses off such a popular bike route. The line would then run up 20th all the way to Sullivan’s Gulch. 20th might have to lose some parking for bus stops, but it wouldn’t really require major changes to accommodate bus service.

After Sullivan’s Gulch, where would it go? From a network design perspective, it should connect with the 8 going up NE 15th Ave, but the problem is that this new line would have no MAX connection. One solution would be to cut over to Lloyd Center MAX, then back to 15th, but that is quite a deviation. It is also unlikely that current riders of the 8 would be willing to lose their downtown service.

Another idea is to connect this new Line 10 with the northern part of the 44. It would cut over on Multnomah to connect with MAX, then go up Williams to serve the current path of the 44. That part of the 44 currently gets pretty infrequent service anyway, with decent peak service dropping down to once-an-hour service in the mid-day and evening hours. It is also very redundant with nearby Lines 4 and 6. Everyone on that stretch of Vancouver/Williams from Rosa Parks to Fremont lives within a 5-minute walk of the 4 or 6, which both provide frequent service to downtown. Therefore the 44 would be an excellent candidate for being turned into a crosstown.

The new 10/44 hybrid would be another odd-shaped line, but would provide a brand new connection from SE to N Portland and would be much more useful than what are now two redundant and infrequent radials to downtown. This is exactly the situation planners successfully challenged 30 years ago, but these lines show there is still work to be done.

Another clearly needed crosstown would be a line up and down 148th Ave in East Portland. This is needed to fill an extremely large gap in north-south coverage that leaves residents with a series of disconnected east-west lines. As more and more low-income people move out to East Portland, the need for car-free mobility is more urgent than ever, and north-south service will be key to meeting that need. The 71 along 122nd Ave has been pretty successful, showing there is demand for such a line.

The northern end of 148th Ave, up near Sandy Blvd, does have hourly bus service on the 23, but that line quickly leaves the arterial to meander about neighborhoods on its way to Gateway Transit Center. The 23 gets pretty poor ridership and is very unnecessary given that those neighborhoods are pretty close to the 77-Halsey. Whether or not the 23 was eliminated (preferable in order to get some service hours to invest in a new line), there could be a new line running all the way down 148th to Powell. That line by itself would be too short, so it could go southwest a bit and connect with another line like the 17 or 19.

My final example of how we could continue to build the grid is made possible by the long-overdue construction of the new Sellwood Bridge. Bus service can finally be restored! Rather than running yet another bus from Sellwood to downtown (which would duplicate the 35, 19, and Orange Line), why not use it for an east-west crosstown? The 43 is a very low-performing line that runs through SW Portland on Taylor’s Ferry Rd before going downtown via an extremely slow and frustrating route on Corbett Ave through Lair Hill. This line runs very slowly on narrow roads through neighborhoods surrounded on all sides by superior transit service on Barbur and Macadam. The numbers show that very few people ride the 43 on the Corbett, and it also means that few riders get on the Taylor’s Ferry segment knowing they are in for a long slog to get downtown.

If the Corbett segment were removed, the 43 could instead make a sharp turn from Taylor’s Ferry to Macadam, then cross the Sellwood Bridge. It could run on Tacoma St through Sellwood, connect with the Orange Line at the Tacoma station, then continue east on Johnson Creek Blvd, which currently has no transit service. The line could cut south to Clackamas Town Center, making a great new east-west crosstown from Washington Square to Clackamas Town Center. We often talk about the need for good, high-demand anchors for transit lines. This line would have two very good anchors, with lots of neighborhoods in between.

Once again, I would like to make clear that these are my own ideas, and do not necessarily reflect any actual plans, policies, or opinions of TriMet.

In conclusion, I think we need to take this time to celebrate the implementation of the grid 30 years ago, but we also need to start having a public conversation about how we can restore the frequency needed for the grid to work and how to fill in the gaps in the grid that still remain. These are not just questions for TriMet, but also for our candidates and elected officials at all levels of local government. Please chime in with your thoughts on these ideas and submit your own in the comments!

Street by Street, Out of the Mud

Wearing my Planning and Sustainability Commission hat, I’ve been following for some time an effort by the Portland Bureau of Transportation know as Street-by-street. The intent is to address the many unimproved streets in the City (many in East Portland, and a few areas in NE and SE Portland).

The source of the problem is that under County zoning rules, developers were not required to improve streets before building homes. So when areas were annexed to the City, they included streets that were not up to City standards. It’s incumbent on the property owner to bring the street up to standards before the City will take over maintenance.

It’s phenomenally expensive to meet full City standards (including sidewalks, curbs and storm water facilities), so the result has been that nothing happens, and these streets stay gravel or dirt.

Street-by-street is an effort to define some alternate standards that are better than mud, but not as expensive as the full treatment, on the theory that we’re all better off with something useful, even if not perfect.

I first encountered this as part of the Cully Plan, where pilot treatments were proposed to address the large gaps in the street network there. More recently the Planning and Sustainability Commission got a complete briefing on the program, and today City Council had a work session on the topic.

Except that today, when it got to Council it was called “Out of the Mud (and Dust)”. I think I detect the Mayor’s messaging hand at work…

And I was happy to see that one of the recommendations from our Commission was included, lowering the target speed for “shared streets” (where a single 16-foot paved area is intended to serve people in cars, on bikes and on foot) from 20mph to 15mph.

You can check out the full briefing presentation here (PDF, 1.5M).

I know some pedestrian advocates who think settling for anything less than fully separated sidewalks is a bad idea, but I’m inclined to think that half a loaf (at less than half the cost), particularly on streets intended to serve 500 or fewer auto trips per day, is better than none.

What do you think?

Five reasons why BRT may have advantages over rail

Once more into the bus/rail breech, my friends.

In various comments and articles, I’ve enumerated various advantages that bus rapid transit has over equivalent-service rail in some circumstances; this post is simply a collection of these. It doesn’t constitute an endorsement of bus over rail for any specific project or system, hence the word “may” in the title–that analysis needs to be done on a case-by-case basis. And this is a one-sided post; the corresponding advantages that rail has over bus are not listed. Not because they don’t exist or are not important, but simply because I wanted to collect many of the good technical pro-bus arguments in one place. (I’m limiting myself to technical arguments for the most part; sociological or political arguments such as “trains cause gentrification” or “rail is just pork for developers” are not included).

A bit of terminology: This article refers to “Class A”, “Class B”, and “Class C” transitways, which refer to the isolation of the transitway from other traffic. Very roughly:

  • Class A is a grade-separated transitway (or one with absolute crossing priority), such as the various freeway-adjacent sections of MAX, and much of the Blue Line between Beaverton and Hillsboro. There are no examples of class A bus in the Pacific Northwest; North American examples can be found in Ottawa and Pittsburgh.
  • Class B is surface operation in an exclusive right of way where the transit vehicle may need to stop at crossings, such as MAX through downtown, along Interstate and Burnside, and in downtown Hillsboro. Much of the EmX line in Eugene is an example of Class B BRT.
  • Class C is ordinary mixed traffic operation–such as the bulk of TriMet’s bus operations as well as the Portland Streetcar. Generally, plain class C bus is not considered BRT, but a type of bus service that is is commonly referred to as class C+ bus (or by other names such as “rapid bus”)–this refers to mixed traffic bus that enjoys enough enhancements (off-board fare collection, all-door boarding, signal priority, limited stop spacing, prominent stops) that it is a materially better product than local bus. Mixed-traffic streetcar systems can also have signal priority (and be class C+); the Portland Streetcar does not do this however.

A claim was made in a thread at Human Transit that for class A and B operation, rail is almost always preferable; this is a partial rebuttal to that, but the content is important enough to emphasize that it deserves a post of its own.

After the jump…
The reasons
BRT enjoys these advantages over rail:

  • Topology advantages. One thing that BRT does easily but rail cannot do is operate in an “open” configuration–meaning vehicles travel in a transitway for part of their journey, and then filter out into the existing street network without need for any special off-transitway infrastructure. Trains can only run where there are tracks and switches, but busses can mix between a busway and local operation. Open BRT isn’t useful for every situation–it works best in cities with the main transit nodes concentrated in one place and transit demand dispersing outward from there–but many cities do have that transit topology.
    Related to the open busway issue is the capability to handle large number of discrete vehicles. I’ll ignore some of the more extravagant claims concerning high-traffic freeway bus lanes like the Lincoln Tunnel XBL (Exclusive Bus Lane) between Manhattan and New Jersey; the XBL is a special case (there are no stops in the tunnel, after all). But here in Portland, the transit mall is able to handle–easily–180 busses per hour and 10 trains per hour, in each direction. Even if both 5th and 6th streets were dual tracked, and MAX trains limited to a single car-length, I can assure you that the Mall could not handle 180 trains per hour. For a long corridor–say Gateway to downtown–you wouldn’t want to run 180 busses per hour; if you had that much demand over such a distance that’s an obvious candidate for rail. But for highly branching topologies, the high vehicle throughput is a major consideration.
  • Partial operation: The ability of busses to run on ordinary streets has a second set of advantages. It permits easier phasing–agencies building a busway or bus lane can open half of it when it’s done, and have busses run in the completed sections of the busway and on local streets the rest of the way, and then shift additional sections of the route into the busway when it completes. BRT also lets you get around hard parts with mixed traffic operation. Obviously, mixing in a mixed-traffic (class C) section adversely impacts the overall reliability of a route, but it’s a way to utilize existing (expensive) infrastructure such as bridges or tunnels. Rail, OTOH, can only go where there are tracks, and mixed-traffic operations compromises rail far more than it does bus.
  • Costs. For Class C/C+ operation; bus is way cheaper to install–it’s just ordinarily local bus service, possibly with changes to traffic signals and nicer stations. The equivalent rail technology is mixed-traffic streetcar. Streetcar may be better suited to placemaking and land-use transformations, but the performance characteristics of mixed-traffic streetcar are generally the same as ordinary bus service; but streetcar requires installation of tracks. BRT also lets you do class B cheaply–if planners are willing to taking a traffic lane. Class A infrastructure will require major capital construction regardless of mode; a good argument can be made that rail is more suitable for class A operation for this reason (though that isn’t an open-and-shut case; Class A open BRTs like Melbourne are highly effective). But if political will exists, class B bus can be installed as cheaply as class C+. And even if a transit agency does decide to pour concrete–replacing asphalt with a road surface more able to withstand the axle loads of bus–you don’t have to relocate the utilities under the pavement or rebuild the roadbed. Which brings us to…
  • Less prone to catastrophic failure. BRT doesn’t break down as easily or as spectacularly when the line gets blocked or closed. This benefit is most often discussed in the context of streetcars vs local bus (where obstacles along the route are plenty), but even class A and B transit lines are impacted by events such as accidents, breakdowns, power or control failures, and maintenance of the right of way. With a BRT, vehicles can simply navigate around, leaving the transitway if necessary. Rail often requires “bus bridges” to be set up when a line is taken out of commission. Generally, many sections of track–up to the next switch on either side of the incident–will need to be bridged. This happens all the time in Portland–and frequently results in bus runs being cancelled so a bus bridge can be assembled, angering just about everybody, both bus and train riders alike.
  • The ability to pass. BRT makes it far easier to mix express and local services and provide skip-stop service. Busses can simply pull out of the busway for stops; only a little more pavement and real estate is needed to enable passing. The transit mall is a fine example of this capability, using skip-stop operation to achieve a high vehicle throughput.. Adding express tracks or passing sidings greatly complicates the design of a rail line, as switches need to be installed and signals changed for any change in track topology.

President Obama expedites the CRC (corrected)

Note: The original version of this article identified the source as Oregonian reporter Joseph Rose; the Oregonian article was written by Charles Pope. Portland Transport regrets the error.

This evening, Joseph Rose Charles Pope reported that President Obama has ordered the $3.5B Columbia River Crossing to be given “expedited” status. While not providing any additional funding, “expedited” status permits red tape to be cleared out of the way–and in many cases, local objections to be overridden.

This would appear to be bad news for project opponents. When the Tappan Zee Bridge in New York received expedited status last year, the result was the transit components of the project being jettisoned. Whether a similar situation could happen here–with the CRC being built without the transitway, is unclear–in the Tappan Zee case, the transit part was greenwashing that none of the major project stakeholders (including the State of New York) really cared about. Here, of course, the city of Portland insists on light rail as a condition of the project, and the project cleared state regulatory hurdles by virtue of a light rail siting statute (resulting in the infamous declaration from the state Supreme Court that the CRC was a “light rail project”). OTOH, it may well be the case that Salem and Olympia could override any local objections should they choose.

On the other hand, it must be remembered that the Bush Administration expedited the CRC four years ago, and the project is still stuck in planning limbo.