Aurora’s Destiny Is Bus, Not Rail

Albuquerque Rapid Transit, a center-running BRT - the gold standard to which Aurora should aspire (Credit: Albuquerque Rapid Transit,
Albuquerque Rapid Transit, a center-running BRT--the gold standard to which Aurora should aspire to. (Credit: Albuquerque Rapid Transit)

Author’s Note: This article is in response to Doug Trumm’s excellent piece on Aurora rail, which was in turn a response to Seattle Subway’s vision map. I recommend reading both, perhaps even first.

Is rail the right solution for Aurora Avenue?

When identifying future rail corridors, a logical place to start is high preforming bus lines, and King County Metro’s RapidRide E Line is a great place to start. Upgrading the Aurora corridor to rail would have many benefits, but I’m skeptical rail is what is best for this corridor. I see five major drawbacks in a bus-to-rail conversion.

  1. Cost. Even gold standard, fully grade-separated bus rapid transit (BRT) would be significantly cheaper than rail because Aurora is already limited access and partially grade-separated between Mercer Street and N 85th St, and there is adequate right-of-way for bus lanes. Most of the value of light rail construction comes from the creation of new right-of-way, and that simply isn’t necessary on this corridor.
  2. Local opposition. One of the reasons local politicians have consistently killed light rail along SR-99 in every subarea is residents and small business owners express serious concern about multiple years of construction on their doorstep. BRT can be implemented in a shorter timeframe and with less costly disruption.
  3. Stop diet. The E Line currently has 17 stops between Denny Way and N 145th St, while even the most optimistic rail line would have much less. Under Trumm’s rail proposal, the corridor would offer 10 stops. While studies show that people are willing to walk further for a subway station than for a bus station, I’d wager there would be no difference between bus and rail if they offered identical frequency and speed. The stop diet required for rail conversion means that we will need to continue to run local buses along Aurora to maintain the same quality of service for the community.
  4. Capacity overkill. Almost every long-term vision includes a Ballard-Lake City extension of light rail, which should adequately address any E Line peak capacity issues by providing an “interceptor.” Off-peak, the lack of high ridership would likely result in trains running less frequently than buses would in a BRT line.
  5. Missed interline opportunity. Rail misses the opportunity to interline E Line and Swift into a single BRT line, instead forcing a transfer somewhere between Seattle and Everett, wherever the rail happens to stop.  Metro & CT’s route currently overlap in Edmonds, which not a logical transfer point.  As the 3 county area continues to grow into a single urban amalgamation, we need a transportation network that reflects this unity. While E+SWIFT could be prohibitively long for a single driver, self-driving bus technology should eliminate this issue in the future.

Rail certainly has many benefits, and if cost was no concern, I would happily support another north-south subway. And a great advantage of a rail option is the ability to improve station location by deviating from the existing SR-99 alignment. Unless, of course, Ms. Frizzle is your bus driver.

A BRT solution for the region

Our regional long-range plan already has the Aurora/SR-99 corridor marked as a single bus rapid transit corridor. (Credit: Sound Transit)
Our regional long-range plan already has the Aurora/SR-99 corridor marked as a single bus rapid transit corridor. (Credit: Sound Transit)

My vision for Aurora BRT would be a line operated by Sound Transit that breaks down the arbitrary borders between Metro, Community Transit, and Everett Transit. The line would provide continuous service from Seattle to Everett, better connecting contiguous neighborhoods on SR-99. The service features for this line would be:

  • Center-running but generally at-grade, leveraging the grade separation already built into Aurora Avenue between N 85th St and Mercer Street. Existing business access and transit (BAT) lanes would go away, so there is no impact on existing general purpose lanes. This design can make Aurora safer for all.
  • Elevated stations at major intersections, allowing the buses to bypass traffic while building light rail-quality stations for pedestrian and bicycle access. For example, I’d imagine elevated stations at N 85th St, N 105th St, N 130th St, and N 145th St within Seattle. 
  • Existing RapidRide and Swift stop spacing would be used. Link light rail would remain the fastest option for “long haul” trips between Snohomish County and Seattle.
  • A one-seat ride would be available for all neighborhoods between Seattle and Everett, absorbing the existing E Line and Swift Blue Line. Presumably, this would be funded and operated by Sound Transit, but it could be a joint operation between Metro and Community Transit as well.

I believe BRT fits the scale of SR-99. Look at the shape of growth corridors in Shoreline Town CenterEdmonds, and Seattle’s urban villages Aurora-Licton Springs and Bitter Lake. These are long, narrow stretches of density, better served by multiple bus stations than a single rail station. Unlike a freeway, development faces the street so a BRT station will not serve an ocean of asphalt. Nonetheless, there is ample existing right-of-way to repurpose for bus lanes, which makes a true BRT upgrade much easier, both technically and politically, than most of the other RapidRide+ corridors. Put another way, I am arguing the current SR-99 “stroad” is well suited for BRT.

A BRT solution for Seattle

Possible elevated alignment. (Doug Trumm via Google Maps)
Possible elevated alignment. (Doug Trumm via Google Maps)

In Seattle, any enhancements to bus infrastructure will also benefit Metro routes that use Aurora to get into downtown, such as Routes 5, 26, and 28. There are three options for serving downtown:

  • At grade. This is the cheapest option but the most uninspiring. It likely doesn’t make sense to build excellent BRT north of the Ship Canal and then have buses get stuck in downtown traffic. There may be a way to do at-grade well, likely with a grade-separated crossing of Denny Way and then strong enforcement of a bus mall along 3rd Avenue.
  • Elevated. Trumm sketches out what an elevated alignment would look like for a new line, and I think BRT would actually be superior to rail for this alignment because rubber tired buses would much more capable of handling the sharp turns and steep grades that this alignment would require. Additionally, First Hill can be seamlessly served as an elevated busway could transition directly into the Madison BRT bus lanes in First Hill.
Westside Transit Tunnel concept. (Ross Brinkley / Seattle Transit Blog)
Westside Transit Tunnel concept. (Seattle Subway / Seattle Transit Blog)
  • Tunnel. The last option is a new bus tunnel. A north-south tunnel quickly turns into Seattle Subway’s Westside Transit Tunnel concept–dating back to before the Sound Transit 3 package scope was widened–which means it could also serve others routes such as the Delridge RapidRide. Most other tunnel alignment options quickly run into issues with tunneling under the Ship Canal or I-5, and if we are going through the expense of tunneling under either of those, then we should probably just go ahead and build rail.
  • Flying Bus Actually, probably a bad idea.

We hope you loved this article. If so, please consider subscribing or donating. The Urbanist is a 501(c)(4) nonprofit that depends on donations from readers like you.

AJ McGauley was a temporary transplant in the Seattle area, living in East King with his lovely wife for five years before returning to the great Midwest. Having lived in nine very different cities in the six years prior to moving to Washington, he discovered the wonky side of urbanism after reading The Urbanist and is interested in why cities grow (or don’t grow) in different ways. He worked for Sound Transit and can still be found riding transit for fun.

Inline Feedbacks
View all comments
Howard Metzenberg

It’s ultimately about capacity, which is why I am with Doug Trumm on this one. A train adds real capacity to the Seattle urban area, running in its own right-of-way, and can foster genuine urban villages on the giant parking lots of Aurora Avenue. Buses are for shorter links to trains.

Aurora to Everett is the region’s longest strip mall, much of it occupied by dying retailers. A train does much more to raise the value of land along the corridor. Urbanist governments could be acquiring land along this corridor through voluntary acquisitions, not eminent domain, with both short term and long term uses in mind for that land.

transit rider

Shoreline got away with an overload of stops, in part because Metro wasn’t willing to run local service underneath, despite having some existing routes that could be altered to do just that.

E Line and Swift overlap in Shoreline, not Edmonds. Chalk that up to being a newcomer.

BRT should have been the choice to do the Everett loop (128th – Boeing – downtown Everett), and the north loop would have opened next year, at a fraction of the cost of light rail. Instead, because Snohomish County Executive Dave Somers, who reportedly has never ridden BRT, refused anything other than rail, the region is ponying up an extra $1 billion and northerners are stuck waiting for at least several more years before an indirect route to and from Everett is projected to open. In short, politics got in the way of what made sense for commuters, density, and what should be most importantly, taxpayers.


Excellent article. I especially like the Ms. Frizzle reference. I agree with all of your points.

It is easy to overestimate the importance of the Aurora corridor. But the E owes its success to three things:

1) It is very long.
2) It is very fast.
3) There are lots of stops.

The corridor has a lot of ridership, but does not have the highest ridership per mile. This is important. Experts who say we should consider the current ridership of a corridor always emphasize the ridership per mile.

But speed is important too. Ridership between Ballard and Bellevue is not very high right now. It has gone down considerably the last few years. This would make the argument for East Link very weak. Yet you can’t ignore that fact that it now takes a very long time to get from downtown Bellevue to downtown Seattle. Ridership on East Link will surely exceed that of the 550, many times over. Yet an Aurora subway line would not have a significant speed advantage over the E. The E, by its very nature, is an express, making only two stops between 46th and downtown. It travels in its own lane much of the way. It is only from outside the city (past 145th) that a train would be significantly faster. But the biggest gain in speed would be from skipping stops. If we really think this is a good idea, then we could run additional express buses.

The other big problem with the corridor is that it doesn’t have major destinations along the way. The train serves UW and Capitol Hill on the same line. It will soon serve Northgate as well as the U-District. The E has nothing like that. To be clear, it does have people going from say, 145th to 80th — just nothing like the numbers that go from Northgate to UW, or Capitol Hill to Roosevelt. The big destinations are on the main Link line.

Many of the riders on the E make a transfer to those destinations. If you are at 130th heading to the UW, for example, then you take the E down to 46th, and slog your way over on the 44. But once Link gets to Lynnwood, it makes way more sense to catch a bus to the 130th station. As you get further north, the train becomes a better and better option for downtown as well as those other destinations. In comparison, speeding up the E (or building a train) only improves the trip to downtown. What is true for 130th is true for other corridors. As you get farther north, cutting over (to the main Link line) becomes more popular.

Which brings me to the Everett to downtown Seattle idea. I agree that ST should do a better job bridging the service gap between agencies. But Community Transit will soon do the heavy lifting. Swift (which runs from Everett to the King County line) will go into King County, and over to the 185th station. This will enable fast trips from the Snohomish County SR 99 corridor to downtown Seattle, as well as those other major destinations (UW, etc.). It also eliminates part of the overlap. Folks from 185th and 192nd headed up to places like Everett Community College will have a one seat ride. The overlap is not huge, but you probably don’t have huge numbers of people going across (e. g. 160th to Edmonds CC) — not enough to justify a huge expenditure.

Which gets me to the last point. There is no question that high quality BRT would be much cheaper (and arguably better) than light rail. But it wouldn’t be cheap. If you want to run buses in the center, then you need to add a lot of bus stops. If you also want to keep the same number of general purpose lanes (two each direction) that presents a challenge. In some cases (46th) there isn’t room, which means that you have to widen the overpass. You also have to have a way for riders to get from the middle of the street to the side. I suppose you could have a beg button and a crosswalk (which would slow down drivers there). On other parts of the route, you could add center stops by taking the bidirectional turn lane. But some of the more popular stops (e. g. 130th) are cross streets, which means that you have a turn pocket, for cars turning left. Hard to see getting rid of those, which means you would need to spend more money to widen the street. Again, it would still be a lot cheaper than building rail, but it wouldn’t be cheap. I’m not sure if would be a high priority, compared to improving other corridors (like the 7, 8, 44, etc.).


Without a doubt there are people taking trips along SR 99. I don’t want to suggest that the amount of overlap that CT will create is completely adequate. I would probably extend the E up to Edmonds CC (with money from ST). But the inconvenience of making a transfer is rather minor (or at least it will be) and it isn’t the end of the world to make one. There are plenty of similar transfers in our system. For example, if you are at 65th and Roosevelt, and want to get to Eastlake, you have to take two buses, even though it is the same corridor. My guess is more people make that trip than make a cross county SR 99 trip. Or how about Lake City to Eastlake — again, the same basic corridor — with lots of people, but it takes two buses and at some point involves going way out your way. I could go on — you get the idea.

There are issues with really long buses. They are much more prone to bus bunching. Even trains have issues that ripple through the system (because of the fixed headways). You also end up creating three seat rides, or watering down the system. Swift will go to 185th, which will mean people in Snohomish County will have a fast, frequent connection to Link. If you keep similar buses, then you are either spending too much on this one corridor, or frequency will suffer. If you don’t, and only have buses on cross streets (like 185th) then you are asking riders to make a lot of three seat rides, simply because there aren’t that many cross streets. There is no simple solution, but I think running a bus from Everett to Seattle on that corridor would be overkill, and messy.