The Denny Way Paving project will repave 12 blocks of Denny Way as well as add curb bulbs, no-turn-on-red signs, 8 new curb ramps, and bus stop changes. The project predominantly focuses on pedestrian safety.
Denny Way will be repaved from 5th Avenue (near Seattle Center) to Stewart Street (in SLU) for 11 blocks or around 4000 feet. Yale Avenue will also receive new asphalt paving from Denny Way to Howell Street for 1 block.
The sidewalk will be improved along with 8 new curb ramps and 21 rebuilt curb ramps allowing for easier ADA mobility. For safer intersection crossings, 28 “no turn on red” signs will be added and 25 stop bars will be painted at every signalized intersection. SDOT wll also add 8 curb bulbs to shorten crossing distances including 4 paint-and-post curb bulbs, 1 plastic curb-and-post bulb and 3 concrete curb bulbs.
Removed bus stops have red X, Existing bus stops circled in green; New Bus Stop with blue
For transit reliability, some minor bus stop consolidation will remove two bus stops in the westbound direction at Denny Way / 6th Ave and Denny Way / Dexter Ave N. A new bus stop in the westbound direction will be added at Denny Way / 7th Ave N. The stop spacing will increase from ~750 feet (0.14 mile) to ~1100 feet (0.20 mile), which is generally acceptable. The new bus stop is also slightly closer to the RapidRide E bus stop.
In the eastbound direction, the Denny Way / 6th Ave bus stop is removed. The stop spacing increases from ~800 feet (0.15 mile) to ~1200 feet (0.23 mile). While it isn’t great to have bus stops removed, hopefully the bus stop consolidation manages to help the bus reliability a bit.
The Late 8 campaign advocates proposed adding bus lanes in both directions on Denny Way to help alleviate Route 8 bus stuck in traffic. Typically the Route 8 in the afternoon is stuck eastbound because of many cars waiting to turn right on Yale Ave N. Those drivers are attempting to head onto the southbound onramp for I-5. For the repaving project, no additional transit lanes will be added/extended and it is unclear whether SDOT will be willing to reallocate more lanes for transit in the future.
Harrison Street and SLU transit future
The Harrison and Mercer Transit Access Project by SDOT will build a new east-west transit route through SLU. The new corridor will run from 5th Ave N to Fairview Ave N as well as connect to the Eastlake Layover Facility.
SDOT made plans in 2022 to reroute Route 8 to reach Lower Queen Anne via Harrison Street and then Mercer Street rather than Denny Way. In more recent documents, SDOT has clarified that it isn’t guaranteed to move the route to Harrison Street.
Specifically in the 2023 Q&A:
| Will Route 8 be moved to Harrison? |
| Metro and SDOT have made substantial investments to improve Route 8 reliability in the last five years and Metro remains committed to providing bus service on Denny Way. Metro’s long range plan also identifies bus service on both Denny Way and the Harrison/Mercer Street corridor. Shifting Route 8 to the Harrison and Mercer corridor is only one of several potential options to utilize an additional east-west pathway for buses through South Lake Union and Uptown. Before the corridor is ready for service in 2028, Metro will evaluate different service options to determine what routes will operate in the new corridor. This service plan will be informed by a community engagement process, which has not yet begun, as well as availability of service funding. |
Without shifting the high ridership bus Route 8 though, it’s unclear how SDOT will justify building the new transit street.
Route Denny Way/Boren Ave/MLK Way
Metro Connects (long range plan) suggested extending existing Route 106 from Renton via Rainier Beach to CID north on Boren Ave to SLU. The route would then run on Denny Way to Lower Queen Anne, taking over Route 8‘s portion on Denny Way in Belltown.
Perhaps one alternative for Harrison Street could have an extended Route 106 use the new transit street. This would allow Route 8 to continue on Denny Way without duplication.
Route 8 RapidRide and Split Future
 |  |
Further in the future, Metro Connects suggested splitting Route 8 into a north-south and east-west route, similar to RapidRide B‘s proposed split of the L. The new east-west route would run from Lower Queen Anne to SLU via Harrison Street, then to Capitol Hill via Denny Way and finally ending in Madison Park. The east-west route is a potential RapidRide candidate.
The remainder portion of Route 8 on MLK way would be taken over by the new corridor 3997 which would start at Kaiser Permanente on 15th Ave and end at Beacon Hill Station using College Street instead of terminating at Mount Baker Station.
Conclusion
The Denny Way repaving will add better pedestrian safety and access to the corridor as well as minor bus consolidations. Hopefully in the future bus lanes will be extended along the corridor helping Route 8. The future of east-west transit in SLU continues with the Harrison Street transit way but it remains unclear which bus route will be routed down that corridor.
Excellent article. The city really needs to add bus and BAT lanes on Denny. In the long run it makes sense to run a bus on Harrison/Mercer. But it should not be seen as a viable alternative to service on Denny (which should have more right-of-way).
A bus running on Harrison/Mercer should come from Boren (which should also have additional right-of-way). It would not have to be the 106 but for sake of argument I’ll call it the 106. It makes sense to have the 106 use Harrison/Mercer for a couple reasons. First, you avoids turns. The 8 and 106 would each go straight through Denny & Fairview (instead of both making turns there). This speeds up the buses and makes BAT/bus lanes easier. It also allows both buses to keep going the same general direction the whole time. From Jackson, Boren runs northwest. The bus would make a slight right on Fairview (as all vehicles do) and then make a series of turns to get up to Mercer. Thus it would be going northwest, then north, west, north and west. It is essentially going the same direction the entire time.
The same would be true if the 8 is split and becomes an east-west route (from Madison Park to Uptown). It would have a small bit at the end that goes north but it would be a consistent east-west bus. The straighter the bus the more effective it is and the better it serves the overall network.
I agree with Ross’s suggestion to extend whatever the “local shadow” on Martin Luther King Blvd is called all the way up Boren to SLU and then west to Lower Queen Anne. That gives a “back way” to Seattle Center events for southeast Seattle as well as direct access to the center of First Hill and SLU.
Obviously, this might be less useful after (and if DSTT2 comes about, but the bus line could certainly build a link between the neighborhoods in preparation for the re-routing of Line 1, should it happen. [We pray fervently that it does not, of course]. If the train cannabilizes the bus ridership too badly twenty years from now, shift its source, maybe to a Rainier express overlay.
I don’t think Ballard Link would have much effect on that route. Sure, that route would get riders from South Lake Union to Uptown and some of those riders would be replaced by Ballard Link but the bus would make plenty of connections that the train wouldn’t. Even if the train served First Hill (after serving Westlake) I wouldn’t call the bus route a “shadow”. It would share many of the same connections but be just different enough to warrant good frequency (let along existence).
Ross, sure, the bus wouldn’t be an exact shadow of an extended Ballard-Downtown-First Hill Light Metro, but they would certainly “compete” for trips between LQA or western SLU and First Hill.
That’s not an argument for or against establishing the bus line now in any case. It’s an unreservedly good idea.
I’ve been saying for years on this blog that for SLU and First Hill to be so built up and only a mile apart from one other, yet have no direct bus service between them, that’s a huge hole in the transit system. It’s an example of just how much of our bus network is inertia-driven. After all, back in the 1990’s, when SLU was a dump, going out of the way to serve it with buses going anywhere other than downtown would have been preposterous. Heck, if we had the money, a subway version of corridor 1074 would probably get much higher ridership per mile than the planned ST3 Link extensions. And, yet this corridor is so neglected, it doesn’t even get a basic bus!
Well, isn’t “a subway version of corridor 1074” exactly what we eventually want the Ballard-Downtown Automated Light Metro to do? No, it wouldn’t follow Boren and therefore would not directly serve the big cluster around Boren and Howell; a bus would still have to run. But the extension “into First Hill” serves exactly the same catchment areas at both ends, just not the middle.
About that big cluster of buildings…..
Is it possible to squeeze the tracks in the cut-and-cover tunnel section together enough that platforms could be added across Ninth Avenue for a Spine station there? That’s a LOT of riders three and a half blocks away…..
Yes, there are some similarities. But there are some differences (as you pointed out). A Boren bus would cover more of South Lake Union (by going up Fairview and turning on Harrison). As you mentioned the bus would serve more of Boren. The big difference is that the train would serve Westlake (to make a key train-to-train connection) whereas the bus wouldn’t.
In any event we shouldn’t have to wait 20 years for a subway line when we could have a bus route like this tomorrow if we wanted to. Vancouver didn’t wait for SkyTrain to run on Broadway — they built a highly successful BRT line. When SkyTrain is built they will still have buses there (just a lot fewer). Basing train lines on successful bus lines is nothing new. There are times when a train line ignores the grid (which can give it a significant advantage over surface transit) but quite often it just replaces a good bus route (that is typically slow but popular).
I’ve been hoping the 43 is made full time again on its new route (down Broadway thru First Hill to ID, instead of going downtown), that would accomplish this.
Not really. The 43 is a hodgepodge of a route. It overlaps the 8, 48 and streetcar, often in ways that are not appropriate given the demand and congestion (e. g. 23rd/24th through Montlake). It doesn’t enhance the network, it dilutes it. Buses that would otherwise run more often are dragged down because of it and similar buses. The 48 should run more often. The 8 should run more often. Broadway should have transit running every six minutes (or better). Overlaying a route that has bits and pieces of these core routes (and not necessarily the best parts) is not a good approach.
In contrast a Boren bus would definitely enhance the network because it would be a straight shot. No other bus runs that way. The problem right now is not that you have to transfer, it is that you have to go the wrong direction. Look at what it is like to get from Boren & Madison to Fairview & Harrison (https://maps.app.goo.gl/CfYLrt8LZci7qZJEA) via transit. Most options involve detouring downtown. Other options involve going the other way (east, then north). The only option that actually involves heading towards the destination is walking — the fastest option! It is a 22 minute walk and even if you catch the bus immediately it is faster to walk. Running the 43 more often wouldn’t fix that.
But it is more than just the direct connection. Consider a trip from Madison & 17th to Westlake & Harrison (https://maps.app.goo.gl/AkH85ayrnhBAb6kQ9). Once again most options involve going all the way downtown and then coming back. It isn’t terrible, but walking (36 minutes) is actually competitive. One of the fastest options is to walk up and catch the 8 and then walk again after the 8 reaches Westlake (which means 15 minutes of walking). In contrast if there was a Boren bus riders would take the G and simply transfer to the other bus. This would save a significant amount of time because you would actually have a real network in the area.
There are actually two street grids in the area and Metro largely ignores both of them. The main one is based on cardinal directions. The second one is mostly downtown but includes Madison. It is based on the waterfront (which at that point is at a roughly 45 degree angle). Having a frequent bus on Boren (along with the frequent bus on Madison and the spine on 3rd) would create a transit grid to match the second street grid. This would be combined with a grid for the streets that (mostly) are north-south-east-west. We actually have a lot of those pieces — they just need to be altered slightly. We need to consolidate on certain corridors (e. g. Broadway) and stop sending every bus downtown like it is 1980 and that is the only transit destination. We also need to complement Link and not try and compete with it. We need more of a grid (actually two grids).
Yes, there is duplication among bus routes; duplication is reciprocal; it goes both ways. Montlake is a congestion point; both routes 43 and 48 go through it. With U Link, Metro chose to nearly delete Route 43 and split Route 48; new Route 45 has does well; Route 48 has regressed in productivity. Route 43 duplicated routes 8 and 48; routes 8 and 48 duplicated Route 43.
RossB discusses the transit grid. In north Seattle it was pretty good, though some mistakes were made. Transit grids have to use Link stations. Route 48 may not use them as well as Route 43 could have. North Capitol Hill riders can take Route 48 to UW Stadium station pretty well; but they can not take Route 48 to downtown very easily; the transfers to/from radial routes are awkward; at John Street, one can transfer to reach the Capitol Hill station; the G Line transfer takes a long walk.
If bus lanes are too “extreme”, what about HOV lanes? Make the right lanes HOV/bus/right turn only, so it still prioritizes buses and carpools (and receives higher utilization than bus-only), while lessening the traffic impact on the general purpose lane.
The same % of individuals who respect “fare required” on transit will equal those who respect “hov only”
No enforcement and both are a joke.
what do you think the rate of HOV violation is?
I think politically it’s still be easier to implement hov lanes in some corridors.
This probably isn’t the best one though as you’d have it change from hov to bus lane which would lead to a mess
I would be OK with HOV, but my guess is there are only a handful of people who carpool. I think a more realistic compromise is “bus and freight” lanes. There are plenty of people delivering goods that are stuck in traffic. They could make for a powerful ally.
The thing about HOV: it’s hard to enforce (you have to actually peek in windows to see if a car is allowed to be there), which encourages cheats. And, even among valid users, most them aren’t real carpools. Rather, you get lots of cases like a parent taking a kid to school, a family going out to dinner together, a solo passenger taking an Uber somewhere, or somebody picking up a family member from the airport.
I’ve personally benefitted from this. For example, if I choose to take Uber to the airport, rather than transit, I get the 405 HOV lane all the way, in a way I wouldn’t, taking my own car. Yet, the fact that many others get to do the same is part of the reason the 405 HOV lane often gets clogged.
Ultimately, bus lanes on city streets should be just for buses, and if you have to add some traffic to it, commercial trucks is a good place to start, because there aren’t that many of them and it’s much harder to cheat than with HOV.
I wonder how Seattle can rethink SLU to work like downtown — with Third Ave for buses and other streets for moving cars. Putting every function in one street makes no one happy. Now that John, Thomas and Harrison cross over the 99 tunnel it seems possible.
I would even be cool with one way couplets for both buses and cars. One way streets are often panned in the new urbanism world but I think they can have useful merit in some situations.
The real challenge is probably I-5 though. That river of traffic combined with the very steep hill makes things challenging.
It would be cool to look to a transit addition like a funciculat or gondola. But that requires years of planning and a funding source.
I’ve also toyed with a decked street with transit running above traffic. It’s sounds fantastical and expensive — but it sure would help Denny buses get through SLU quickly.
Jarrett Walker did a writeup on couplets, and his thesis is that it’s generally better for bus routes to be on regular two way streets. The reason is that real trips are bidirectional, so running the bus on a couplet guarantees having to walk an extra block in either one direction or the other to catch the bus.
Of course, sometimes, the street grid is set up such that running the bus on a one way couplet is the only way to serve an area efficiently, in which case, you do what you have to do. But, even then, the reason for making the street one way is usually about accommodating more left turning cars, so it’s essentially a small sacrifice on the quality of the bus route for increased car throughput.
Yes, both SDOT and NACTO assert correctly that transit is better off on two-way streets. In the One City Center exercise, SDOT said it wanted frequent two-way all-day service on both 3rd and 1st avenues; then they were dreaming of the CCC Streetcar; now, the CC Streetcar is dead; Seattle should ask Metro to shift some already funded bus routes to 1st Avenue from 3rd Avenue. Unfortunately, SDOT has not followed its own suggestion and is placing the J Line on the Roosevelt couplet. All the transfers would have been better with the J Line using the current Route 70 pathway. SDOT is placing both transit and bike priority on the couplet; it will be slower for transit. There will be friction between cyclists and bus patrons going between the buses and sidewalks on the bus-bike humps.
The Roosevelt couplet makes sense if the intent is to continue north past the UW. Which the J line originally did, but then got shortened due to funding cuts. But, by then, it was too late to adjust the routing to be closer to campus, since being “on the way” for continuing northward wasn’t going to matter anymore.
It’s a tricky topic. The answer is usually context sensitive. One big factor is how big blocks are, how signals are tuned and how stops are spaced.
Seattle is either cursed or blessed by small blocks. The east-west streets in SLU are only spaced about 400 feet from each other. Compare that to the Avenues in Manhattan spaced about 800 feet apart. Roosevelt and 11th Ave north of 45th is in the 250 foot range. Small blocks mean less walking between one-way streets and it means more signals closer together in a place like SLU.
My interest here is transit speed and reliability. You did not explain whether or not buses on one-way couplets move faster or slower. It’s really a function of signal timing in a place like SLU. It’s the speed of Second and Fourth Avenues that helps make buses on those streets faster because a bus in a bus lane can drive through every signal between stops, for example.
The biggest problem with one way streets are probably at the ends. That’s where it may add travel time.
SDOT just extended the Pike-Pine one-way couplet with five Metro routes on it and that didn’t appear anti-transit, did it? In the other extreme, think about how awful the Pike-Pine Metro bus speeds would be if they were both two-way streets.
The Pike-Pine one-way system was extended to gain more space for sidewalks, a bike lane, and transit-priority lane. You can’t really do that with two-way streets when they’re so narrow. You’d have to move all the cars or all the buses to one street, and then there would be complaints that the other street is shortchanged. You’d end up with one GP lane in each direction and no room for transit-priority lanes or half that other stuff.
The Roosevelt couplet makes sense if the intent is to continue north past the UW.
Yes, which is what the J should eventually do. The 67 should be moved to Ravenna/University (The Ave) and share the same pathway as the 45 starting just south of 65th. This means the 45 should turn south on Roosevelt. It means you would have at least two buses going from 65th & Roosevelt to Ravenna/University and the campus (https://maps.app.goo.gl/YxFeBbgmnCbeECrNA). Eventually a new bus (the 77) should run from Lake City to the U-District and follow the same pathway as well (https://maps.app.goo.gl/CocSrbJpNGxaZ2Vw8).
Meanwhile the RapidRide J should start just north of the Roosevelt Station and not leave the Roosevelt/Eastlake/Fairview corridor until Boren (https://maps.app.goo.gl/UDHkKCrUEAkTzVVQA). Thus the J would still make an excellent connection to Link but it would also stay on the main north-south corridor that extends alongside Lake Union. It would mean a spine on one corridor and RapidRide service on another. It would solve the awkward situation we have now where someone west of the U-District Station heading south (to Eastlake or South Lake Union) has to walk east only to have the bus go back west before rejoining the main corridor that the rider crossed on foot a while ago (https://maps.app.goo.gl/4N3Ro4ejKcw5iz237). The two surface transit corridors (one RapidRide the other a spine) complement each other and work well with Link. If you are close to the Ave and headed south it is quite likely you will just walk to the Link station. If you are farther way (on or past Roosevelt) then you are more likely to want a bus that just goes straight south.
It’s really a function of signal timing in a place like SLU.
I don’t think there is a street in the county that is timed for the buses. To do so is difficult. Buses travel at irregular speeds because they make stops at an irregular basis. You can do things that benefit the buses (https://nacto.org/publication/transit-street-design-guide/intersections/signals-operations/transit-signal-progression/) but other things have a bigger influence. Traffic Signal Priority (TSP) is one, of course. The light stays green a bit longer then it would normally or it turns green a bit early. Another is just the signal phases favoring the corridor the bus happens to run on. Even if the traffic signals aren’t timed very well for a bus on Aurora you have to figure that more time is spent (on a typical cycle) with the light favoring Aurora traffic then a particular cross street. Then there how many phases there are. A lot of streets have a phase for left turns which means the cycle is really long.
I think in the case of Third it really has nothing to do with timing. First the north-south streets are favored over the east-west streets. Second I don’t think there are any left turn signals on any of the intersections. If there are some there there aren’t many. This means the phases are relatively short and odds are the bus will be be able to get a green light. But that is different than the lights being “timed” (so that if you drive a certain speed you’ll make all of the lights).
I don’t think Denny is quite as clean but it is similar. I’m pretty sure the signals favor it in most cases. There are a lot of places where you can’t turn left (but plenty where there is a left turn arrow). Fairview has a lot of left-turn signals and thus long cycles. I’m not sure there is a lot that can be done about that.
“I don’t think there is a street in the county that is timed for the buses. To do so is difficult.”
Here I think it’s more that bus stops are laid out to match signals. An example is when 2-3 signals are close to each other and coordinated. Bus stops are often placed before the first signal and after the last rather than stop in the middle so the bus can get through those signals without having to stop.
It’s not the same as active signal priority. That’s when the timing changes in realtime because a transit vehicle is approaching. Priority really only works between stops — which is why Link on MLK does do pretty well. It’s not practical to anticipate transit vehicles beyond just the next stop and Link stops are pretty far apart.
But being designed for active transit signal priority is different than having a street timed for buses. When stops are sited to get through the signals between those stops, they are operating as de facto “signals timed for buses” in a conceptual way — even if the Metro and SDOT chose the bus stop sites because of pre-existing signal timing.
Meh. Jarret Walker talks a lot, and some of it is good. Some of it isn’t.
The bike lanes for the Roosevelt Couplet are different than those on Pike/Pine. For Pike/Pine the bike lanes are on the left. This means that the bus stops (on the right) don’t interfere with the bike lanes at all. In contrast they put the bike lanes on the right on Roosevelt (and 11th/12th). As a result it requires island bus stops (https://maps.app.goo.gl/GuPrRv2hriZCdBVQA). That means there is less space to add BAT lanes.
To be fair there are a number of reasons for that. Going south the couplet becomes a two-way street and then goes over the bridge. At some point cyclists need to be on the right so they can go over the bridge. If the bike lane was on the left (north of the bridge) it would be awkward or dangerous to get over to the right.
One alternative is to make both Roosevelt and 11th/12th two-way streets. But instead of general-purpose traffic going both ways it would be bus-only going north on Roosevelt and bus-only south on 11th/12th. (This is known as contraflow.) It would mean that you would not need island bus stops. The bike lanes would stay where they are (but with more protection) while the bus stops move to the curb (on the other side of the street). The only drawback is getting to and from the couplet. The buses would basically be doing a crossover (Allen Iverson would be proud). You could make the cars do a crossover (while the buses go straight) but I think that would be more hazardous (for people walking). I think it would be simpler for the buses to do it. A northbound bus would cross the bridge in the left lane (hopefully in its own lane) and then take a left on 41st, followed by a right on Roosevelt (in its own lane). Thus it wouldn’t wait for traffic — only pedestrians. A southbound bus would be traveling in its own lane and make a simple right and then a left (onto Roosevelt). This is the only place (other than 41st itself) where a bus would have to mix with traffic but it wouldn’t be too bad.
Another alternative is to make both streets two-way and add bike lanes on one street and the buses on another. I don’t think this would work that well though. Roosevelt would likely be the bike street to retain the downhill pathway. That means those going uphill would have to turn left on 41st or 42nd to continue on the protected bike lanes. Or we could redundantly add them on the other street, but that means that we are back to island bus stops.
I think the best approach is contraflow and the second best is just to keep the current setup and add as much in the way of right-of-way as possible. It is two lane each way and it doesn’t need to be. Just as Westlake will soon be one lane each way, so can the Roosevelt couplet.
Here I think it’s more that bus stops are laid out to match signals.
Yes, but that is for each individual signal. It has more to do with backups as opposed to making the lights. Imagine a bus stop on a crowded street right before the intersection. Traffic is stopped at the light and the bus is stuck behind the traffic. The light turns green and the bus approaches the intersection. It finally access the bus stop before the intersection. It sits there a while as people get on and off. Just as it is about to pull out the light turns red. Thus it missed the light twice. This is why they tend to put the bus stops after the intersection (although there are plenty of exceptions).
Priority really only works between stops
Yes, of course. It is worth noting that it isn’t like a fire truck. Sometimes the train (or bus) is just out of luck. It is basically a sliding scale in terms of priority. For example imagine a simple two-phase signal. The light turn east-west for 30 seconds and then north-south for 30-seconds. A bus is traveling north-south and arrives 20 seconds into the east-west phase. The priority is set for 5 seconds. This means that the bus has to wait five seconds but then the signal switches 5 second early. The same thing would happen if it arrived at the tail end of the north-south phase (the north-south phase is extended to 35 seconds).
I think the train does well for several reasons. First there is a general preference for travel along that corridor (it isn’t split 50-50). Second there is likely a bigger priority for the train than there is for a bus (e. g. ten second instead of five). Third the large gaps help the train. But I don’t think it is signal priority in this case as much as being able to take advantage of signal timing. Basically the train operates as an express through Rainier Valley. Thus it can run though several traffic lights at the same time. If the lights were random then signal priority could “fail” randomly and the big gap wouldn’t help. But that isn’t reality. Worse case scenario the train has to wait at a traffic light and then soon makes every other light (since they are timed). That is why signal timing is really only a big deal with express service (like Link in Rainier Valley) and not typical urban surface transit. It is only in those cases where a bus (or train) can take full advantage of it.
“I think the best approach is contraflow…”
Contraflow cones with some basic problems. Crossing pedestrians often don’t look both ways. Any blockages will stop buses completely as they couldn’t swerve into the adjacent travel lane. There are often parking issues too — including loading zones and handicapped spots. Then there is the issue of designing the end point operations — likely requiring special signal phases getting added so buses can use the lanes with buses having to wait until the phase arrives.
Finally, contraflow has the same design issue about bus directions on different street that others think is already inferior.
So it’s a big expense in redoing the entire street pavement, signs, parking rules and new signal heads to add the reverse direction — yet doesn’t appear to me to save signifucant travel time for a rider once completed. I think a much better investment is to improve sidewalk and bus stop lighting along with paid fare areas at bus stops with TVMs (fast boarding) , and to make geometric changes to help bus speed and flow.
Contraflow comes with some basic problems. Crossing pedestrians often don’t look both ways.
That has nothing to do with it being contraflow. It has more to do with changing a street. Make Second Avenue two-way and some people won’t look both ways. There is a learning curve for sure. But it is worth noting that the various places where we do have contraflow are not exactly high pedestrian-accident places (two-way Aurora is). People tend to look both ways even if there is only one lane going one direction.
Any blockages will stop buses completely as they couldn’t swerve into the adjacent travel lane.
Unless they swerve into oncoming traffic (just like with any two-lane road). In any event this would not be a big issue on this street as for most of the corridor there would be only one bus. The buses would just wait for the other bus (the way buses do in most cases). Once in a blue moon a bus is stalled and traffic cops would be called out to wave the other bus around.
There are often parking issues too — including loading zones and handicapped spots.
You have it backwards. Contraflow avoids the typical issues involved with a transit mall. Cars and trucks are allowed on the street 24/7. They are only allowed one direction, that’s all.
Then there is the issue of designing the end point operations — likely requiring special signal phases getting added so buses can use the lanes with buses having to wait until the phase arrives.
I already covered that. No special signals would be needed.
Finally, contraflow has the same design issue about bus directions on different street that others think is already inferior.
Yes, that is a trade-off. So you are suggesting what? We turn Roosevelt into a transit-mall (north of Eastlake)? Good luck with that. There are dozens of parking lots people are used to accessing. You still have trouble with the “endpoints” (at least northbound). You still have trouble with people not looking both ways. You have many of the disadvantages of contraflow and only one advantage — the bus stops would be on the same street. But even that isn’t that much of an advantage. Think about where the buses would be running if they ran contraflow — northbound on Roosevelt and southbound on 11/12. That means to get from a northbound bus stop to a southbound bus stop you don’t need to cross the street. This would be a huge improvement over today, where the bus is on the far side of both streets. Yes, running on the same street would mean the stops would be closer, but not a lot closer.
Sure, a full-time 24-hour, 365-days-a-year Roosevelt transit mall would be fantastic (and better than contraflow). I just don’t think it is realistic. Contraflow is.
@Mike h,
“Meh. Jarret Walker……”
Jarret Walker is the Carl Sagan of transit.
And anyone who actually met or had to deal with Carl Sagan knows exactly why I mean by that comment.
@ Ross:
After all that disruption and cost and negotiation with the storefront businesses, does it actually save riders significant travel time? I don’t think it will.
We shouldn’t build contraflow lanes unless there’s a travel time benefit. We know that Seattle has had a recent parade of transit projects put out there that don’t save much travel time.
After all that disruption and cost and negotiation with the storefront businesses, does it actually save riders significant travel time?
Of course it would. That is like asking if the Third Avenue Transit Mall saves a significant amount of time. Of course it does.
We know that Seattle has had a recent parade of transit projects put out there that don’t save much travel time.
First of all that isn’t true. Of course they save time. Secondly the only major drawback is that they don’t go far enough! Since the subject of this post is supposed to be about the 8 I assume you are talking about the added bus lane and island bus stop here: https://maps.app.goo.gl/cjiLuQBzc1WzeD9Z7. Does this save any time? Obviously it does. You can see the long line of cars on the curbside lane. If you’ve ever used that bus stop (and I have) it is clear that it makes a huge difference. The bus passes all of those cars and gets to the bus stop a lot sooner than it would with BAT lanes (as everyone in the right lane is trying to turn right to get to the freeway). Does that mean the 8 is fast now? No! That is because the “fix” wasn’t enough. You need more right-of-way. Adding it won’t be easy. Consider the various options (starting with the best ones):
1) Transit mall. I think this is unrealistic. I don’t think we can kick the cars off of Denny.
2) Contraflow. Same problem. It is less disruptive but only by a little bit. I don’t think we can kick the cars off one direction. It would also mean that buses would be stuck in traffic going the other direction (as there is no obvious pair for Denny).
3) BAT lanes. This is the only realistic solution but it means that the eastbound bus has to do a bit of a weave. It has to run curbside then switch over to using the middle lane (to serve that stop and avoid the backup). It also means that unlike a transit mall/contraflow the bus still has to deal with traffic.
Again, I want to be clear. I think BAT lanes would help a lot. Just because implementing them wouldn’t be easy and it wouldn’t be as fast as a transit mall doesn’t mean they aren’t worth adding.
But it is also worth noting that one of the problems we have is that we think BAT lanes are ideal. We add them and then get surprised that the bus isn’t as fast as we think it should be. That is the case with RapidRide G. The part where the bus runs in the middle of the street is fine. The part where the bus runs in BAT lanes (downtown) is not. This is where the backup is and this is where there are a transit mall or contraflow would make a huge difference. So consider the trade-offs (between the two) downtown:
Both a transit mall and contraflow are just as effective when it comes to congestion (and a lot more effective than BAT lanes). Both allow buses to run very fast. The only advantage of a transit mall is that a bus can pass a stalled bus (which should happen rarely). The drawback of a transit mall is that it is extremely disruptive. You aren’t just taking a lane — you are taking all the lanes. In contrast contraflow on an existing one-way street involves taking the same amount of space as BAT lanes.
In the case of Roosevelt it would actually take less. Thus it would be less disruptive than BAT lanes but a lot more effective. That is because of the bike lanes. Because of the bike lanes we need island bus stops which add extra space. With contraflow the bus stops would be on the curb, allowing more space for a wider sidewalk and/or parking.
We should use ever tool in the toolbox to make the buses faster. In some cases (like Denny) that means BAT lanes. In other cases (like Third) that means a transit mall. (It is worth noting that the transit mall is going to be extended and in my opinion should be extended all the way to Denny). In other cases — especially when we have existing one-way streets — contraflow is the way to go. But simply throwing up our hands and saying “it won’t make a difference” is guaranteeing failure.
“But it is also worth noting that one of the problems we have is that we think BAT lanes are ideal. We add them and then get surprised that the bus isn’t as fast as we think it should be”
That’s how the US is different from other countries. Other countries are more willing to give a BRT line or streetcar full transit lanes for its entire length, in order to maximize speed and ridership and minimize driving. Mixed traffic is limited to small segments that are too narrow to convert such as historical areas. The US finds any excuse not to convert GP lanes or parking lanes, so they end up with a bus/streetcar that’s slower than its potential, has less ridership than its potential, and is not as transformative on mode share as its potential.
I wonder how Seattle can rethink SLU to work like downtown — with Third Ave for buses and other streets for moving cars. Putting every function in one street makes no one happy. Now that John, Thomas and Harrison cross over the 99 tunnel it seems possible.
Ideally you would have one street for cars/trucks, one for buses and one for bikes. Of course it is never that neat. One nice thing to remember is that BAT lanes are bike lanes. They aren’t protected (since cars can use them for short stretches) but they provide more protection than your average street. Ideally their would be a nearby parallel street with bike lanes as well.
In any event in this case it would make sense to add BAT lanes on the Denny/John/Thomas corridor as well as the Harrison corridor. Mercer then becomes the street for cars. Instead of a future bus running on Mercer it could run on Roy (which is actually better since it is covers more of the area). It is a fairly simple set of improvements that would still allow those driving to get around.
That is what is so frustrating about this. Al is the first person to suggest Denny (or the Denny/John/Thomas corridor) to be a transit street. Most of us just want enough in the way of BAT lanes to have the bus run reasonably fast. It is a four-lane road. If they just added BAT lanes it would make a big difference while still allowing cars to do their car thing.
As far as Harrison goes, that is the plan. Turning it into a transit street has been in the works for a very long time. The problem is it doesn’t go through. It runs into the Seattle Center to the west and the freeway to the east. Having an east-west bus detour just to follow that tiny section just doesn’t make sense. On the other hand having a bus from go west on that section and then north is quite appropriate. The bus would not be making a detour but rather going the same general direction the entire way.
Whoa. This is underwhelming.
SDOT and Metro were to be collecting concepts on improving the Route 8 flow. Where are they? Perhaps the STB could ask about that. The stop consolidation may help a bit. The no right on red will slow Route 8; Route 8 buses use the outside lanes to serve stops; right turning vehicles must wait for pedestrians; both will further slow Route 8. Where is the juice? Repaving is the time to implement speed and reliability projects. (Note SDOT missed that opportunity on northern Greenwood Avenue North; there some pullouts remain that should have bus bulbs; Metro had a Route 5 project outreach, but I have seen no action; in fall 2020, the Saturday Route 5 headway was lengthened to 20 minutes, so we went backwards).
Yale Avenue will be repaved. Well its connection with the southbound ramp to I-5 is a leading cause of congestion on eastbound Denny Way. That ramp attracts a long queue that backs up onto Denny Way. Will Yale Avenue be blocked? Will the ramp be closed or tolled? Otherwise, eastbound Route 8 will remain slow and unreliable.
Even in the long run, it may not make sense to have two frequent crosstown routes three blocks apart (1,250 feet). Service hours are scarce and will always be scarce. See service guideline four on route spacing. Given the geometry of walking speed and waiting, transit ridership will be maximized with frequency on one corridor rather than split on two so close together.
Metro Connects is not funded. It has many concepts that do not make sense with constrained service budgets. SDOT relied on the MC concept of service on Harrison Street and applied for a PSRC study grant. That does mean it makes transit sense. The existing Route 8 pathway serves Belltown; there are towers under construction on its pathway. The existing Route 8 pathway includes the Queen Anne and 1st Avenue North couplet with electric trolleybus overhead. In the AWV Enhanced Trolley Scenarios of about 2008, it was envisioned with overhead. Later, in the U Link project P2, Route 8.11 was tested; it was not proposed for the ordinance.
The PSRC will fund a study; it will not fund service hours, electric trolleybus overhead, pavement, or other infrastructure. Why does SDOT think it can move a bus route better on the two-lane Harrison Street any better than on a four-lane Denny Way? SLU local streets should be teeming with pedestrians; all will be slow. Both streets feed ramps to SR-99. Traffic will equilibrate and congest all of them.
MC notions. They are unaffordable. Note that Route 3997 would not serve Capitol Hill Link and be only four blocks from frequent service on 23rd Avenue (in fall 2024, Transit took Route 10 from Capitol Hill station; why? I think riders want to reach Link with short transfer walks).
Regarding the Harrison Street concept, a retired service planner and hiking buddy of mine coined an adage: one cannot take a study across town.
Regarding the new bus stop and the no right turn on red, one thing I’ve wondered is why the westbound 7th Ave bus stop isn’t on the far side rather than the near side
Too many driveways outside the Shell station?
Route 3997 should be extended to Capitol Hill station. There’s no reason not to make it viable to go north to a Link station as well as south. Transferring at 15th to transfer again six blocks further doesn’t make sense, nor do people want to walk up the hill eastbound. It just creates an inefficiency in the transit network that benefits nobody. Some might be concerned about overservice on those six blocks, but they need to get over their fear of ultra-frequency, and recognize that it’s one of the densest areas in the city.
3997? I assume you mean 1997. Yes, the 1997 is goofy. It would be silly to head towards Capitol Hill Station and then stop a few blocks short. I don’t think it is practical either (there is no layover space there). Maybe the folks who drew the line forgot where the station was. In any event there are a number of better alternatives like:
1) Just ending at Madison. That doesn’t connect to as many places but it still connects to two Link stations, covers what needs to be covered and makes a one-seat connection that doesn’t exist now (Beacon Hill to the CD). This means it can run more frequently (or other buses can).
2) Take a right on 19th and go up to Galer. This would be a way to cover 19th if we got rid of the 12 while the 10 ran twice as often (https://seattletransitblog.com/2023/08/30/high-frequency-network-surrounding-rapidride-g/). The drawback is that the 10 and 12 wouldn’t intersect. That leads to this sort of thing:
3) The 10 would follow its current pathway from downtown to 15th & Thomas. But instead of continuing north on 15th it would turn on Thomas and go over to 19th and then go north again on 19th. This other bus would go on 15th (thus the buses would essentially weave).
Each of these have their advantages and disadvantages and each make more sense then what they proposed.
“Take a right on 19th and go up to Galer. This would be a way to cover 19th if we got rid of the 12”
That means 19th would have a route with no significant destinations.
I can’t see how the MLK portion of the 8 can get split off into a useful route without requiring many more service hours.
I think the 11 should just get sent down Denny to combine with the 8. That could run at 20 minutes for now, for 10 minute frequency from LQA to Madison Valley. I think that would match our existing service hours more closely to ridership.
That means 19th would have a route with no significant destinations.
It would connect to Madison Park, Garfield High School and Beacon Hill. It is common for a route on a grid to lack major destinations. That is arguably one of its major features. The routes are not geared towards guessing where people want to go with infrequent one-seat rides. It is instead focused on providing frequent connections (https://humantransit.org/2009/04/why-transferring-is-good-for-you-and-good-for-your-city.html).
I’m not saying it is an ideal route. But it would be consistent in the sense that it is a coverage route. A bus on 19th is coverage if there is a bus on 15th and 24th. Same thing goes with a bus on MLK (since there is a bus on 23rd). So you are basically just extending a coverage route. That being said the connection to Beacon Hill would attract people that would otherwise ignore a route like this. A trip from Madison Park to various parts of Beacon Hill takes a while even with the speedy (and frequent) RapidRide G (https://maps.app.goo.gl/1PPqjp54ACxapWtz5). If you are a few block away it gets much worse (https://maps.app.goo.gl/pP52gda21cPJ2WUY6). Thus the route could attract riders who otherwise have to go all the way downtown (or make a three-seat ride) just to get to various parts of Beacon Hill.
I can’t see how the MLK portion of the 8 can get split off into a useful route without requiring many more service hours.
I can and it isn’t that different than the other idea you mentioned. Start with this: https://seattletransitblog.com/2023/08/30/high-frequency-network-surrounding-rapidride-g/. That means combining the 8 and 11 (while making other changes that result in buses running a lot more often).
Now backfill service on MLK with something less frequent. Connecting MLK with Beacon Hill looks nice but I could see a different branch than the one you suggested. The 27 has a service mismatch between the areas to the west (Yesler Terrace) and the areas to the east (Lescii). This would be a good place for a branch. Have a bus go down Yesler and then head north on MLK and go up to Madison Valley. That would connect much of the Central Area and provide another way to get downtown (for those in the hard to serve areas of MLK). Not only is it a good branch from a ridership standpoint but each branch takes about the same amount of time. This makes managing the routes a lot easier.
It is important to note that this approach — like your approach — would treat service on MLK as more coverage in nature. This is as it should be. Right now — in many cases — we are splitting service equally and the result is a system so poor it is losing riders. Riders have multiple 20-minute buses to choose from (which go to the same place) and they are essentially ignoring them and taking the few frequent options available (even if they require a long walk). Or they are just calling a cab. The 8/48 situation isn’t quite that bad but it is similar. MLK is quite close to 23rd and both buses run every fifteen minutes. They both go to the same place (Mount Baker Station). One way or another you need consolidation even if it means more frequent service on one pathway and less frequent service on another.
“It would connect to Madison Park, Garfield High School and Beacon Hill. It is common for a route on a grid to lack major destinations.”
Those are two seat rides. A route needs to serve at least some of people’s everyday needs. Transferring is substandard if you have to walk a block or more and wait 15-30 minutes for a continuing bus, like from the 48 to the 8 or 11.
Ah that is a good idea. The 27 branch looks pretty good and makes more sense than sending the MLK portion of the 8 down to Mt Baker. And the southern half of the 8 is already well-covered by the 48/14/4 anyway.
Like 10 years ago a number of the properties along the south side of Denny between maybe 5th and Westlake were still just surface parking lots. Always thought the city should have purchased them — or at least easements along the ROW — and widened the street to add a transit lane. Having sat for 20 minutes without moving in the past, it seems like bypassing congestion for even two or three blocks would do wonders for rush hour speed and reliability.
In Metro Connect Appendices, 1074 was listed as comparable to 8,106 in 2025 route but comparable to route 38 in 2040. Not sure why is that.