A Frank noted last week, tonight from 5:00 to 7:00 PM, at Ballard High School, Sound Transit and SDOT are hosting an open house to study the possibility of high-capacity transit from Downtown Seattle to Ballard. I’ll be there, and I’m looking forward to seeing lots of you guys there.

It’s worth reflecting on the problems with today’s transit when trying to design tomorrow’s. Zach has a great post coming up soon, looking at the average end-to-end scheduled speed of Seattle’s transit routes; I won’t steal his thunder, but suffice to say, most of the non-freeway routes, except Link, are pathetically slow. I was inspired by that to look a little closer at a route that’s among the alignments to be studied in this open house, namely Metro route 40. What parts of this route are slow, and what can we learn from this?

The table above is mostly composed of data about one sample Route 40 trip, northbound in the PM peak; it’s time-point data from OneBusAway, travel lengths computed from Google Maps, and speed computed by dividing the two. One caveat upfront: anyone who’s ridden the 40 knows the real schedule can be mess, particularly if it’s a bad traffic day downtown; nonetheless, these are decent ball-park numbers. There are two other data points, from two of Portland’s rail lines as they cross the Portland city center at around the same time: the Portland Streetcar North-South route, and TriMet’s MAX Red Green Line.

Here are some of my thoughts:

• Downtown is by far the slowest part of Route 40. Won’t be a surprise to anyone whose ridden a bus here, but downtown, and the downtown-like dense area that extends up to about Mercer Street, is just not a quick place to get around on a bus. To give a sense of how much of that is due to running on the street, Metro planners generally estimate it takes a bus five minutes less to travel the length downtown in the transit tunnel than an equivalent trip on the surface, so that’s maybe a 35-50% speedup, based on the first row of data above.
• Surface rail through downtown isn’t likely to be much faster. Even with well-optimized signal priority and a dedicated lane*, MAX is barely scheduled faster than Seattle’s surface buses. On a bustling, intimate, pedestrian-oriented street, there’s a limit to how fast a 30 ton vehicle can safely travel, and even excellent signal priority can’t guarantee you’ll make a dozen lights in a row. It seems unlikely that downtown Seattle will do much better, although Seattle’s blocks are longer, which may help.
• Where right-of-way is readily available, it barely matters. There are two places where surface right-of-way could easily be taken for (semi-)exclusive transit use: Westlake, north of Valley Street; and Leary, east of Ballard and west of Fremont. These streets are wide, have few signalized intersections and very little pedestrian activity, and traffic, including transit, moves pretty freely on them already. It doesn’t seem like ROW would make transit much faster, although signal priority could help by allowing transit vehicles to keep moving.

This is an open thread for anything related to the open house. See you there!

* The MAX Transit Mall is a pair of one-way streets, each with a painted transit lane and a general-purpose lane; this section of the streetcar runs in a curb lane in mixed traffic. All Portland rail services have off-board payment.

## 64 Replies to “Where and How Can We Make At-Grade Transit Faster?”

1. David L says:

One lesson here is how much difference could be made by TSP much more aggressive than SDOT has been willing to impose to date.

The slow speeds from 3rd/Virginia to Westlake/Highland are due to poor light timing, pure and simple. There are not that many stops in that area, although passenger volumes are reasonably high that time of day. But there is no TSP along that stretch and the bus is spending a lot of time waiting for lights.

Lights are also a cause of the relatively slow speed from Aurora to NSCC, but that is harder to fix because the lights are also holding up large volumes of cars.

The slow speeds on 24th in Ballard could be helped a little bit with a stop diet at the north end; take out the stops at 83rd and 77th. The south part of that stretch already has pretty good stop spacing.

1. Eric says:

Bus bulbs could also work wonders on that stretch on 24th. Seems like the bus often spends more time at a stop waiting to get back into traffic than it spends actually picking up and dropping off passengers.

1. Mickymse says:

How about we just start ticketing drivers who don’t let buses back in? And how about ticketing drivers who block intersections? Or turn corners illegally?

It’s amazing how much better traffic would flow in Downtown and its edges if we just enforced the laws that everyone is SUPPOSED to be following the first place.

2. djw says:

How about we just start ticketing drivers who don’t let buses back in

An easy, cost free (indeed, revenue generating!) way to improve transit service that should have been done aggressively, yesterday. Of course, like TSP, Metro can’t make it happen without the cooperation of other government agencies, which are not interested in transit. And so it goes…

3. Who exactly is going to do the ticketing? It’s probably cheaper to build a bus bulb than to hire more traffic cops.

4. djw says:

“Who exactly is going to do the ticketing?”

The same police officers who issue tickets for speeding, illegal left turns, and such.

5. d.p. says:

“Who exactly is going to do the ticketing?”

The same police officers who issue tickets for speeding, illegal left turns, and such.

Oh, so you mean “no one”.

6. djw says:

I don’t drive, myself, so I can’t speak to this from experience, but plenty of friends of mine who drive manage to get pulled over and ticketed from time to time.

7. JohnS says:

SPD does not prioritize the kind of enforcement Mickymse (wisely) calls for. In the past I’ve thought it would make more sense to move this kind of enforcement to SDOT; SPD tends to treat its ‘dedicated’ traffic force like the cavalry, making use of them wherever they need extra bodies.

8. Ben Schiendelman says:

In general, if the solution to something appears to be “issue tickets”, there’s an incentives problem. Redesign the street or just build bulbs. :)

9. FWIW, 24th Ave NW would need Dexter-ish bus bulbs to accommodate the bike lane as opposed to N 45th-ish ones, which are smaller and probably cheaper.

In any case, I think Dexter and N 45th have both turned out pretty well. I hope this isn’t just my perception, and that similar projects can move transit along similar streets.

10. Skylar Thompson says:

I’ve often thought that installing “bus runner” cameras would help. These cameras would come on when a bus is at a bus stop with its left turn blinker on. Any cars that pass would be photographed, license plates recorded, and drivers ticketed, in the same way that red light cameras work now.

There’s probably some complication that I’m missing, but it would be so satisfying when sitting on a bus that can’t merge because of single-occupant vehicles zooming by.

2. I’m actually sort of surprised that the “Crown Hill” section on this trip came out so much slower than the “Ballard” section, since the “Ballard” section includes the turn onto and off of Market. That’s the part of all the Ballard routes that always seems to crawl when I’m on them. But this chart is essentially based on one run, so it should be taken with a grain of salt. Also I’m in Ballard more often on weekends than during the evening commute, and I bet weekend traffic is concentrated on Market more than 24th.

1. David L says:

The issue with 24th, as noted above, is the combination of excessively frequent stops (between 70th and 85th) and trouble getting back out into traffic. It’s much worse at peak times, such as when this run was timed.

2. Jason says:

This chart makes you think that the problem is downtown, but I’d argue that these numbers need to be mixed in with ridership numbers (like this chart, but for 40: https://seattletransitblog.com/2011/10/29/ridership-patterns-on-route-14/) for each segment. If the bus is 25% full at 3rd and Marion, but 75% full on 36th and 1st, then even though the latter segment is “fastest” perhaps attention should be given to making it even faster so as to benefit the most number of people.

1. GuyOnBeaconHill says:

Do the 40’s runtimes between timepoint differ from the times used by the old 17/18/75 on the same segments? I’d like to see a study done on the actual run times, not the scheduled run times. Anybody who has ever taken those old routes knows that the 4 minutes scheduled to cross the Fremont Bridge can easily stretch to 14 minutes on a bad day, but 10 minutes to run up 24th Ave NW isn’t too hard to accomplish. Metro’s schedules seem to be built with an optimistic view of how long it will take to run from downtown to the closest neighborhood (Fremont in this case) and then there is some recovery time built into the segments further from downtown. I think there’s also some schedule padding often built into the last segment on many routes.

3. GuyOnBeaconHill says:

The Portland streetcar is burdened with some crucial errors made in the planning and design process, which I hope haven’t been duplicated in Seattle on the First Hill line. There are way too many stops in the Pearl District and NW, plus many of the stops are placed so that the streetcar has to be right at the intersection before it can open its doors. Too often, as the Portland streetcar is approaching a station, it has to wait for a car ahead to clear the platform area (which can take a full light cycle), then the streetcar pulls up to the platform, opens and closes its doors while the light is green and then as the doors close, the light turns red and the streetcar waits for another green. Tick, tick, tick….

The station placement is better on the CL line–usually at least one car length back from the intersection. It would also be helpful if the platforms were longer than the streetcars. That would allow a little wiggle room if there is a car stopped in the lane ahead of the streetcar at a red light.

1. d.p. says:

Check out the track map for the First Hill Street Car: http://zoom.it/Ge29

In particular, zoom in on the northbound stop right before crossing Pine and the southbound stop right before crossing Pike.

Also, recall that through traffic on Broadway is being reduced to a single lane, so cars are 100% guaranteed to be in front of the streetcar at any given red light.

Also, notice that the stops are shared for buses and streetcars.

And then read your description of the Portland streetcar’s “crucial errors” again.

Yup.

4. Doug says:

add to one of the major reason for being slow isn’t just road capacity or lights or infrastructure but rather the density of people riding from downtown to fremont. every stop takes SO LONG because the buses are PACKED. They mostly seem to be running short (40 ft?) buses on the route…and getting on or off can take on the order of minutes as people rearrange and shove their way on or off…

1. aw says:

So, a “rapid streetcar” with off-board payment, multiple wide doors and level boarding could help that considerably.

1. Doug says:

completely agree, i guess my overall point is that it’s not just traffic and infrastructure causing the slow down and the resolution you suggest is one way to accomodate. However, in the meantime, certainly the size of bus or the frequency of service should be looked at (with all the appropriate caveats applied…lack of funding etc.)

2. Ben Schiendelman says:

Haha. Thanks aw – exactly. That’s why it’s in the top three corridors in the TMP.

3. d.p. says:

AW: No, really, it couldn’t .

4. d.p. says:

…Which is precisely the mathematically-supported point of Bruce’s entire post.

But don’t take our words for it. You too can experience the total inability of streetcars to provide speed improvements when confronted with frequent lights and in the absence of any sort of dedicated right-of-way.

Simply take any bus to Westlake, board a SLUT car, and note how long it takes you to get anywhere on it!

2. d.p. says:

As a near-daily user of the 40 (at many different times of day), it has been my experience that boarding and exiting delays are far less problematic on the route than an many others.

One advantage of a frequent route through mostly transit-savvy areas is that most of the ridership uses ORCAs and knows to exit from the back when possible, while the slow bike/wheelchair/clueless-cashpayer phenomenon loading is less likely to happen on your bus every single time (the way it seems to with infrequent routes or certain trolley lines).

The route is already much faster than the old 17, partly for the reasons above, partly because the straight line it travels saves 1/2 mile over Nickerson, and partly because Nickerson was a particularly awful street for buses pulling back into traffic (both before and after the road diet).

The 40 is actually so shockingly fast and efficient that if a bus gets lucky with the lights in SLU and Belltown or with the bottlenecks of Fremont Place and the Mercer crossing is likely to make the trip in 20-22 minutes, beating the heavily-padded schedule by as much as much as 8 minutes (even when carrying a moderate-heavy passenger load).

But man, can those lights and those bottlenecks be killers. The Mercer cycle is arguably worse for crossing traffic now that it ever was before the two-way project. The lights of the Denny Regrade are awful, but you can’t really do much about them because the cross-streets carry transit too.

A streetcar following the 40 corridor might save, at most, a minute or two in boarding time. But it can’t queue-jump the entirety of Fremont, and as Portland has amply shown, it will crawl through downtown even if you attempt to give it priority treatments.

It will solve nothing.

1. Nathanael says:

Give it gates like a proper railroad line and it will go quite quickly through downtown…. how would that be for “priority treatment”?

2. d.p. says:

It would be “not fucking going to happen”.

5. reality based commute says:

Metro has been asking the city to look at transit contra flow lanes on 2nd and 4th, but so far the city has balked, probably because they would have to take parking. Things will only get worse in downtown Seattle with the viaduct coming down and seawall construction starting.

1. Ben Schiendelman says:

2nd, 4th and 5th should probably all be two way anyway. That would help.

1. David L says:

I strongly disagree. That could be beneficial for pedestrians and local delivery traffic, but it would kill the only working N/S corridor available, and severely delay suburban buses. Having 2nd and 4th to get car and some bus traffic into and out of town will only get more important once the viaduct closes.

2. I’m fairly unconvinced by the idea that two-way streets are inherently better than one-ways for pedestrians and cyclists. One-way streets have simple intersections with relatively few movements and conflicts so they can have short, simple light cycles. If a street is converted to two-way and needs to have extra turning phases that can cut severely into pedestrian walk times.

Many US one-way streets are lousy places to walk because they were made one-way as one part of projects to maximize vehicle throughput. But the efficiency of one-way streets could actually work for pedestrians, by allowing narrower streets and shorter and simpler light cycles. On a one-way street you can do this without totally destroying vehicle throughput, which means it’s politically possible to do.

6. RossB says:

So surface based transportation is slow. Got it. I wonder how fast the streetcar is. I wonder if it can match the super fast speed of 15 miles per hour (which is the 40 at its peak). Anyone who has ever ridden a bike along a bus path knows that buses are slow. You do a little leap frogging (annoying both the bus driver and the biker) and then pretty soon the biker gets way ahead. Fifteen miles an hour is not that fast on a bike but apparently it is smokin’ fast for a bus.

So, anyway, I can think of several possible routes from Ballard. All of these are grade separated, so that they could be faster than a jogger (or maybe even a biker):

1) Ballard to the UW in a tunnel. I think this is the shortest distance, which is why it may be the best value. You could add a stop at Fremont, which would really help that neighborhood.

2) Ballard to the UW on an elevated line above the Burke Gilman. I have no idea if there are big issues with this route, but I wouldn’t be surprised if there are plenty. I think this route might be the cheapest if there the issues aren’t deal breakers.

3) Ballard to downtown via Interbay. It is hard to say whether underground or elevated makes more sense. That is why they do studies (especially cost studies). It makes sense for the part that goes right by the railroad tracks to be elevated (above the tracks). The neighbors are used to trains. There are only a few obstacles along there. On the other hand, elevated across the canal would require a very tall bridge (as tall as Aurora) or a drawbridge (which would be rather unfortunate). Furthermore, tunneling in Ballard and downtown makes sense (given the existing density). I personally like elevated rail (since it is a more enjoyable experience — almost as much fun as a ferry) but sometimes it just doesn’t pencil out (neighbors complain about the noise, pillars are expensive, etc.).

4) Ballard to downtown via Fremont and Westlake. I don’t this route makes sense. It further isolates the west side of town, which needs high speed transit more than anyone. Aurora should be (and hopefully will be) used at some point to serve more areas (like Fremont). I think it wouldn’t take much (a little tweaking of the bridge) to get reasonably fast service for those folks. Right now, the 5 is probably the fastest way for someone to get from Fremont to downtown (even if you include the extra walk time).

1. Mickymse says:

#4b) Use Dextrer, not Westlake. Look at the density and development along that corridor over the last few years and plans for more!

Then, as you approach Fremont, transition up to Aurora where a new bridge would be high enough and SHORT enough to be relatively cheap.

1. Pete Lorimer says:

Interesting. That map explains why we’ve had the problem extending the Burke-Gilman trail past Fremont. The original rail right of way was cut in two by the Fremont arm of the ship canal, so where the trail’s natural extension would then be on the south side of the canal. I never before understood why there wasn’t right-of-way past Fremont.

2. Ben Schiendelman says:

I found a really old map of rail lines in Seattle that made it clear to me – cool to realize how much of our infrastructure is based on decisions from 150 years ago, isn’t it?

3. Andreas says:

@Pete: You’re right that the SLS&E originally went straight across on Northlake/Ewing—and continued to even after the canal was built—but by 1918 it also ran on into Ballard, with the diagonal 8th Ave bridge (shown on that map) replacing the Ewing bridge.

The problem has never been that there wasn’t any ROW into Ballard. The ROW was and still is there. The problem is that unlike the rest of the old SLS&E line, that segment of ROW is still active, and the operator has been fighting the Burke-Gilman extension tooth & nail.

2. Mike Orr says:

An elevator to a flyer stop on Aurora would be cheaper than a new bridge, and probably cheaper than an all-day express route to Fremont.

1. JohnS says:

And have a heck of a view to boot :)

2. Eric says:

I would love to see something like this happen. I work in Fremont and the 358 passes almost literally over my head on the Aurora Bridge. If I want to actually ride that bus I have to walk a mile uphill to the nearest stop at 46th Street.

3. Ben Schiendelman says:

That’s a really cool idea!

4. RossB says:

I completely agree. We should leverage Aurora to serve Fremont. It would be cheaper and faster than most alternatives.

3. PhillipG says:

2) Ballard to the UW on an elevated line above the Burke Gilman. I have no idea if there are big issues with this route, but I wouldn’t be surprised if there are plenty. I think this route might be the cheapest if there the issues aren’t deal breakers.

The other routes are at least plausible, but I don’t see the logic behind this one. You lose almost half the possible walkshed by running along the canal, and you don’t serve 45th St or Market St.

1. RossB says:

The main advantage is cost. It is basically built already. But you are right about losing 15th and Market. That is a logical spot to put a station. I’m less concerned about losing 45th. Remember, light rail is about a limited number of stops. You don’t really serve “45th” (meaning the length of it) you serve one stop along there (OK, maybe two). I can’t think of a stop along 45th that would compare to Fremont, even though the street has way more people along it than does Pacific (although, to be fair, Pacific is growing quite a bit).

2. David L says:

You have potential stops with huge walksheds at 46th/Phinney and 45th/Wallingford. All those people on the 44 are going somewhere! Meanwhile, the 26 that serves the area around the trail has negligible ridership. The 31/32 do better, but most of those passengers are headed to Fremont.

3. asdf says:

I don’t see how you could possibly build an elevated line over the Burke Gilman without closing the trail for up to 5 years while the line is under construction, so I consider this proposal a non-starter. And even then, it wouldn’t be wide enough, unless you think single-track is the way to go.

4. I can think of lots of problems with “Ballard to UW along the BGT”. First, the BGT is mostly right along the water and has lousy walkshed (lousy walkshed is OK for a bike path, especially if you get a nice grade in return, but not OK for transit stops). Second, the BGT has some tight turns and narrow sections (it doesn’t precisely follow the former rail line). Third, the BGT (or “missing link”) crosses under the following bridges: Ballard, Fremont, Aurora, University, I-5 (at the point of this crossing there are already four levels in play: the two decks of I-5, the BGT, and the side street it’s crossing over). If you’re building elevated build it over streets, Chicago-style. This gets you out from under the low bridges and into useful walksheds.

7. PhillipG says:

2) Ballard to the UW on an elevated line above the Burke Gilman. I have no idea if there are big issues with this route, but I wouldn’t be surprised if there are plenty. I think this route might be the cheapest if there the issues aren’t deal breakers.

The other routes are at least plausible, but I don’t see the logic behind this one. You lose almost half the possible walkshed by running along the canal, and you don’t serve 45th St or Market St.

1. PhillipG says:

Should have been a reply to RossB…

1. Ben Schiendelman says:

If you’d like, re-reply it there and I can delete this.

2. PhillipG says:

Ben-

Reposted it above, so this subthread can be safely deleted.

8. Surface transport in a crowded urban area is no faster than walking? What a surprise. Actually, Boston, New York, London and Paris had this same problem around 100 years ago, and they developed an ingenious solution to it. It has since been implemented in most cities around the world, but you probably don’t want to hear about it. It could be expensive.

1. Lloyd says:

“… an ingenious solution to it. It has since been implemented in most cities around the world, but you probably don’t want to hear about it. It could be expensive.”
And that’s our problem – we simply cannot see beyond the ends of our noses. We are NOT solving a problem for this budget cycle or for the next decade, but for 3/4 of a century or more.

1. Ben Schiendelman says:

Indeed!

2. Nathanael says:

Ah, you mean elevated railways!

Yes, that was a snarky comment. The fact is that New York built an extensive network of elevated railways, and *several decades later* built subways to replace *some of* them.

9. Joseph Singer says:

The chart of speed mostly agree with what I’ve found between downtown and Capitol Hill checking with a speedo app. Link light rail isn’t all that speedy either. I’ve never seen anything higher than 38 m.p.h. traveling from Sea-Tac to Westlake. The highest transit speed I’ve encountered is 40 m.p.h. on a 66 express. It’s not so much the traffic light problem as it is that the bus (on non express) often stops every two blocks in some areas.

1. David Seater says:

Link should have been well over 50mph on the section along I-5. I’m surprised you only measured 38mph.

1. Ben Schiendelman says:

Yeah, I’ve personally seen 58mph. :)

2. Oran says:

How does your speedo app know the speed underground?

3. J. Reddoch says:

And I’ve been on some buses where the driver hit 65 to 70 mph.

1. asdf says:

The 594 often cruises at 65-70 mph, once it gets on the freeway.

10. Well, now I know why I can walk home almost as fast as the green/yellow MAX lines or buses in some cases. (I live just over a foot bridge at the Union Station stop and commute from the “City Hall” stop – about 8 blocks from the PSU South stop). Portland’s big clog on the transit mall seems to be unfamiliar drivers that get into transit only lanes, causing buses to miss signals.

11. The MAX green line thru downtown Portland isn’t very fast, because it stops every 4 or 5 blocks and the lights are timed to 12 mph.

In fact, the parallel bus lines on the Portland transit mall are faster. For example, the 8 will get you 1.0 mile in 7 minutes (average speed 10 mph: http://goo.gl/maps/1ufdM), while the Green line takes 10 minute to go the same distance on the same street: http://goo.gl/maps/T52fB

Don’t get me started on the slow-as-molasses streetcar

So Seattle should be able to speed up buses to 10 mph even thru Downtown, with exclusive bus lanes on a pair of streets and 12 mph signal timing similar to Portland. And even higher average speeds could be achieved by spacing the bus or train stops farther apart while raising the signal timing speed to 15 or 20 mph.