Ten years ago yesterday, March 19, the University Link Extension opened to the public. The extension utilized a new 3.15 mile twin bore tunnel from the north end of the Downtown Seattle Transit Tunnel (DSTT) to the University of Washington, via Capitol Hill. U Link was arguably the first subway-like section on the 1 Line (then Central Link). Trips between the University of Washington (UW) station and Westlake station only took 8 minutes, compared to 20-30 minutes on a bus. The extension caused a significant Metro bus restructure that truncated numerous routes at UW.
Over the past 10 years, Link has been one of the fastest growing light rail systems in the country, by both ridership and miles of track. For a map of each Link extension, check out the Link Expansion Map created by Luke Billington.
- March 19, 2016: University Link Extension (from Westlake)
- September 24, 2016: Angle Lake Extension (from SeaTac/Airport)
- October 2, 2021: Northgate Extension (from UW)
- September 16, 2023: St Joseph T Link Extension (from Commerce Street)
- April 27, 2024: East Link Starter Line (between South Bellevue and Redmond Technology)
- August 30, 2024: Lynnwood Link Extension (from Northgate)
- May 10, 2025: Downtown Redmond Extension (from Redmond Technology)
- December 6, 2025: Federal Way Downtown Extension (from Angle Lake)
The party is not over yet! In just 8 days, the Crosslake Connection will begin open to the public. This extension will connect the 1 and 2 Lines and open new stations in Judkins Park and Mercer Island.
This is an open thread.
Excellent investment. Well spaced stops and high operating speed. Literally what they deny themselves elsewhere.
Capitol Hill is a real neighborhood, with character.
South Lake Union is just a lot of tall buildings. Who needs to serve that?
People work in those buildings, nominally at least.
People live in those buildings too.
But to maximize people going to a neighborhood, you need the widest variety of destinations there, meeting the needs of the largest cross-section of the population. And it needs to have an inviting character that pedestrians want to linger in, like a pre-WWII neighborhood with its small vertically-oriented buildings and art deco or traditional architecture with attention to detail, not soulless overscaled modernist monstrosities that feel designed for robots rather than human beings.
Capitol Hill, Ballard, the U-District, Old Bellevue, and even the West Seattle Junction (California Avenue) have this. SLU doesn’t: the buildings are all out of scale, and most of the businesses cater to the tech lunch crowds, not a full range of residents’ needs or shoppers’ interests or recreators interests. That’s why SLU feels so “We built it for you but you’re not happy with it”.
Either way, the current plan is for Ballard Link to tunnel under South Lake Union and stop at Aurora.
> the current plan is for Ballard Link to tunnel under South Lake Union and stop at Aurora.
fyi as of yesterday the new plan is to skip the SLU station. though it will still stop at denny (denny way and westlake)
Yes that’s my understanding too, WL.
The planned SLU station is so close to the 99 tunnel that it would have to be quite deep.
The loss though is that it would be the only place where E Line crosses the 1 Line. DSTT2 is at least 2 blocks from Third Ave Downtown.
If the station went away, I would suggest excavating at Denny Park and restoring it when finished. That may however have 4F problems (Federal rule about taking park lands).
Of course, any underground construction is fraught with potential risks from water and sewer lines to unstable soils to native sacred grounds or artifacts to drainage surprises like hidden creeks or rivers. .
That’s what I was intending to mean: they’d tunnel under SLU, but not have a station there until Aurora.
It seems like serving SLU was one entire motivation for Ballard Link, but now the plan serves neither Ballard nor SLU.
> If the station went away, I would suggest excavating at Denny Park and restoring it when finished. That may however have 4F problems (Federal rule about taking park lands).
they recently bought the south lake union discovery center so can use that land
“new plan is to skip the SLU station”
Those are scenarios. The board didn’t approve anything yesterday. The board asked staff to take the feedback and come back with more scenarios and information. To make it “the plan”, the board would have to introduce a motion in a regular monthly board meeting, accept public testimony on it, and vote on it.
No, Glenn. It will tunnel under Aurora and terminate at First West. You are confusing Denny Way Station with South Lake Union Station, which given the current plan is just barely within the 10 minute walkshed of Denny Way, without much development around it, and with a wide strip embargoed from development by the Bertha tunnel structure underground.
the skipping of SLU was a scenario for the Enterprise Initiative workshop; that is short of a plan.
The new Ballard EIS is overdue. I can see how each deep tunnel station is quite expensive, so I’m wondering what the EIS will show.
There’s a bigger topic here. What comes first — the Ballard EIS or the Enterprise Initiative resolution? The Everett EIS is also due to be published soon.
The Enterprise Initiative vote is May. I suspect the board will be too busy with that to finish the Ballard EIS by then.
The BLE EIS is in review by the FTA and the FTA is apparently waiting for guidance regarding recent executive orders.
I can think of plenty of other extensions that has “Well spaced stops and high operating speed”, but why would I bother when this is clearly a bad faith probe anyways, get it together Paul.
Correct. It’s the best segment of Link. Wish that was the standard for the rest of the system.
How can it be the standard if there is no other segment as good? That is like saying you wish LeBron James was the standard for basketball players. You can’t replicate that level of success because there are no other places in Seattle with that potential. It really can’t fail. You could just run on the freeway and skip every station in between (a terrible idea) and still end up with a bunch of riders. Serve South Lake Union instead of Capitol Hill? You would still get a bunch of riders. Or how about building a normal subway (like the Forward Thrust proposal). This means several stops in between. This would get the most riders (by far) and transform transit in that region. Another success! You really can’t lose.
In contrast, various corridors (like Ballard) are not as strong and can easily be screwed up. Other areas (like West Seattle) are much better suited for a Brisbane-style BRT system or just more mundane bus improvements. U-District to downtown was bound to be the best subway segment because of the fundamentals — something that can’t be duplicated in the rest of the region.
SODO, Chinatown, Westlake, Capitol Hill, and UW.
Why these? These offer the best value for a continued trip, and more riders.
SODO is mostly for busway transfers. Chinatown for sports and ferry/water taxi transfers. Westlake since its a major hub and best connection to the Seattle Center and SLU.
Capitol Hill since it’s pretty spaced already and its own place.
UW/U District since it’s a major university campus and has transfer potential to the East (Redmond, Kirkland, Woodinville, Bellevue, etc.)
The other stations? Useful stops if you are traveling within that area. Symphony is useful if you’re getting to Pike Place or the waterfront…but Chinatown isn’t that bad of a walk either.
But if we add a local train option (with express being low frequency and occasional), then we can add more stops without much of a hassle for riders who need to get in/out of downtown.
“Well spaced stops?” Are you trolling? If so, I’m taking the bait.
U-Link was a major improvement in transit mobility for the region. It contains most of what we should have started with (U-District to downtown). But other than the bus tunnel stations, the stops are not “well spaced”. They are way too far apart.
Sound Transit made various mistakes and ended up skipping First Hill. They didn’t even consider a station at 23rd & Madison, despite the obvious network advantages. Look at the Forward Thrust map in comparison. I’m not saying that is the ideal place to put the stations but there are a lot of them. They have three stations between downtown and the UW. We have one. Clearly the Forward Thrust proposal is better (for this section). There are a lot of good things with U-Link. Stop spacing isn’t one of them.
U Link takes 8 minutes. Bus takes 20-30 mins.
With more stops you’d bring up the U Link travel time to 15 mins. Then including getting in/out of the stations and the extra transfer/walking time compared to the bus… You end up losing time. Not everyone lives or work right next to a Link station. Link should move people quickly and further. A city bus is supposed to be responsible for short distance, high frequency, high stop density travel. Stations are also expensive and not needed everywhere, unless it’s a big stop for many people. Unfortunately many of these stations you propose are not destinations… So you have traffic heavy during peak hours or in one direction. Link should serve all day well spaced travel, and have effective FREQUENT and TIMED bus connections.
SKR,
One more stop would not add 7 minutes of travel time. Even two more stops would be unlikely to do that. Pinehurst is going to add like 1 or 2 minutes to travel times, so I would say it’s at most 4 extra minutes. That’d probably be worth it to have stations at First Hill and 23rd/Madison. I think that’d be a worthy trade off, and I would personally ride Link more often if it stopped at 23rd/Madison or thereabouts.
“U Link takes 8 minutes. Bus takes 20-30 mins. With more stops you’d bring up the U Link travel time to 15 mins. ”
A maximum of five more stops have been suggested:
– First Hill: Was in the ballot measure. Would significantly increase ridership: all those people going to the hospitals and clinics and people living in highrises. Both patients, staff, and visitors.
– Pine & Bellevue: Would serve southwest Capitol Hill, and area of solid 8-story density and much nightlife and shopping.
– 15th & Thomas: An urban village.
– 23rd & Aloha: Transfers from the 48, walk-ups from the residential area.
– 520: Transfers to express buses. Walk-ups from a potentially larger Montlake neighborhood.
If all this increases Westlake-UW travel time to 15 minutes, that’s still faster than a bus, and more frequent and reliable than a bus, so people would still ride it and be glad that it exists. And it would get additional riders going to all those new neigborhoods.
A full buildout like that could fully replace route 43, which Metro reduced and hasn’t fully filled the transit gap that created. In particular, people on 23rd have bad transfer options, and people at 15th & John going downtown don’t have adequate transit options either. (The 10 and 11 stop at different stops and each run every 20 minutes, so you have to choose bad service on one or bad service on the other.)
Another problem with U-Link spacing, beyond the missing 2-3 stations, is that the UW station is in the wrong spot. It should have been under the Husky Union Building, which is the literal and figurative, ahem, HUB of campus. It is much more centrally located to most dorms and classrooms, and is a place that the major riding demographic, students, is likely to find themselves anyway. It would also be just over half a mile (track distance) from the U-District station, a spacing associated with creating a walkable, high transit use urban neighborhood (the neighborhood in this case being campus+U-District). The UW station as built sacrifices a huge part of its walkshed to Lake Washington and land on the wrong side of the Montlake cut. Lack of similar spacing from Westlake through the already dense and transit friendly neighborhoods to the east (F. Hill, Cap. Hill, Central) is the real issue with the missing stations on U-Link, beyond simplistic “more people would have a short walk to the station.”
UW Station location is another example of Link being watered down for non-transit reasons. In this case it was because the university opposed the HUB location, and ST didn’t have eniment domain authority over a state university to override it. The university wanted Husky Stadium for its football alumni that donate to the university, UW Medical Center which is a significant transit draw on its own, and to keep subway stations out of the inner campus.
But this keeps happening repeatedly throughout the Link network, and the negative impact on Link’s ridership and usefulness is cumulative:
– No First Hill station.
– No 15th station. (When ST revived U-Link, it could have revisited the number of stations, and replaced the Broadway & Roy station the other alignment would have had. And it could have added Pine & Bellevue to replace Convention Place, and because it’s a dense urban neighborhood.
– No Lynnwood Link stops on Aurora.
– No Lake City stop. (The Lynnwood Link Lake City Way alternative would have had that.)
– Rainier Valley Link on MLK instead of Rainier Avenue.
– No Graham station in the initial segment.
– No Southcenter station.
– No stations on 99 at S 216th Street or between Kent-Des Moines and Federal Way stations.
– No stations on Bellevue Way.
– Overlake Village station not in the center of the Overlake Village retail area.
– Missing down escalators at several stations.
– A West Seattle plan that doesn’t serve West Seattle overall very well.
– Etc, etc, etc.
That’s a good summary of missed station opportunities, Mike!
Here is an article with the Forward Thrust diagram showing possible locations:
https://www.theurbanist.org/forward-thrust-part-1-no-city-island/
It had the Third Ave turn closer to Denny with stops near Pike and Broadway, 23rd and Union and MLK and Madison. Given the topography it may have actually been a tad unrealistic. It also skipped SE Seattle and SeaTac but ran to Renton and Lake City.
Overall, it’s an interesting layout that can be compared to ST2/3. Of course, development and commuting patterns were different in the mid 1970’s.
Even if the UW station was at the hub, UW Med Center and transfers from 520 buses would lose out without the construction of an additional station, costing additional money.
You’ve also got this tension from the outer subareas, who don’t really care about mobility within Seattle, they just want their trip to downtown to be as quick as possible because they view transit as solely for the purpose of getting downtown, and private cars for getting anywhere else (including congested/expensive to park areas like UW and Capitol Hill). To these people, every additional stop Link makes before reaching Lynnwood is a negative, even ignoring construction costs – an extra minute to get from Lynnwood to downtown, in exchange for train to service to destinations that they don’t see anyone in Lynnwood riding a train to, because they all have their cars. This is why Capitol Hill got only one stop. And I highly doubt the U district would have gotten a 3rd station, even if UW allowed it. The fact that they didn’t was just a convenient excuse.
If anyone wants a detailed breakdown of the route profiles, here is a document from the Municipal Archives about the 1970 proposal. It’s a little different than the 1968 proposal, notably the section in the arboretum is 100% underground due to the RH Tompson freeway being deleted from the long range plan. Other interesting things of note are that it would go over i5, not under, to get to first/capitol hill, as well as run parallel to Elliot above ground before crossing 15th ave presumably at the Armory.
https://archives.seattle.gov/digital-collections/index.php/Detail/objects/235842
A city bus is supposed to be responsible for short distance, high frequency, high stop density travel.
Yes, and the same is true with a subway. That is the whole point of building the subway system. There is an interesting discussion on Reddit about stop spacing:
The usual goal of choosing station location is to maximize utilization in its catchment area. You’d basically have to first assume how far people are willing to walk to get to a station (this differs from city to city), and then estimate which location will give you the most ridership given the catchment area. Practical concerns (e.g., land acquisition, route choice, etc) and political considerations then come into play, which would shift the location of the station around the general area, making compromises between all the factors.
And usually that’s it. If two stations close together means a higher utilization than further apart, they’ll be close together, and vice versa. In some cases, there might be a large gap between stations, and a station will be stuff in-between for “transit oriented development”.
In other words your subway line is going from A to B. Now maximize ridership between A to B. That’s it.
This is why it is common in dense, urban areas to have stations around 400 meters apart. That is the case for the old bus tunnel. It is worth noting that originally they planned to add a station at Madison but ran out of money. Since First and Capitol Hill are essentially still “downtown” the stop spacing should be similar (if possible). That should be the case for the UW as well. A station at 23rd & Madison is still plenty urban and is also critical from a bus intercept standpoint. Note that I don’t propose a station at Montlake which means you have a gap of about 2.4 km! That is a huge, very unusual gap. If anything, my proposal is a bit short on stations and we should add at least one in Montlake. The problem is, Montlake is a bit hemmed in with the parks. But there is a good case for even more than I proposed.
Again, this is just the way you normally do things in the parts of the world that have very good transit systems. I’m not sure what type of metro/subway line you are suggesting we build. What is your model city?
Stations are also expensive and not needed everywhere
Yeah. You can easily make the argument that our subway system is much bigger than it needs to be. But that is beside the point. For this section — the most urban in the entire state — there should be more stations.
Even if the UW station was at the hub, UW Med Center and transfers from 520 buses would lose out without the construction of an additional station, costing additional money.
Not running to the hub was unfortunate. But having only two stations in the UW was the bigger mistake. The “UW Station” should actually be in the triangle. There is already an underground connection from the hospital. The overpass from Rainier Vista to the triangle is fine. It makes for a pleasant walk. So all you would need is a way to get down to the platform. An alternative would be to build a tunnel from the north to the underground platform. That leaves the connection to the stadium. You could add another underground tunnel or have people come out on the top and then walk across the street. Very few people actually go that way except for game day. On game day the police control traffic so it really isn’t a big deal if a lot of people have to cross the street to get to the station.
The U-District Station is excellent. That basically means you need a station in between the two. Again, the HUB would be ideal. But since the train goes around the main campus, a station at around 40th & The Ave would work. The stations would not be as close together as the ones downtown but you manage to cover all the areas in between there. It also improves the overall transit network. Buses from the U-Village (and places to the north) run through campus and come out on 40th. They would serve the station easily — no more of this crap. This would make it easier for the 31/32 as well. They wouldn’t have to go all the way up to 45th. They would just keep going east-west, through campus. Likewise the buses that cross the University Bridge would connect to a station sooner.
yes, per RossB, there should have been three stations in the U District: UWMC, HUB, and NE 45th Street. Given the congestion on southbound Montlake Boulevard that the several agencies might have fixed but did not, a HUB station would have fixed the transfer issue; the network would have been stronger. ST was fiscally constrained and did not want another station; the Board of Regents did not want another station. The UWMC station could have been much better; it would have provided a below ground tunnel beneath Montlake Boulevard to reduce the transfer times. the UWMC did not want their tunnel beneath NE Pacific Street to be used by transit riders. the UW did not want buses running on the Rainier Vista. On the other hand, eventually, the UW did a great job improving the Burke Gilman Trail.
“This is why it is common in dense, urban areas to have stations around 400 meters apart. ”
Not at all. That is far too close for subway station spacing. The standard in the US for good walkability is 1/2 mile apart (800 meters), in Europe they tend to use 1,000 meter spacing (0.625 miles). In both cases you are assuming people will walk about 1/4 mile or 400-500 meters to the station (although walkshed is a gradient not a cutoff; look at the maps at this link to see where BART riders at each station come from, with circles on the map to clearly show the influence of 1/4 and 1/2 mi walk radiuses: https://www.bart.gov/sites/default/files/docs/StationMaps_NonHome_12thSt-Embarcadero.pdf).
From Stadium to Westlake is ~1800 meters, which is stations 600m apart, not 400m, and that is very close by most standards.
“Again, the HUB would be ideal. But since the train goes around the main campus, a station at around 40th & The Ave would work.”
The train goes not go around campus, it cuts straight across it, it even passes directly under the HUB! Google maps has a geographically accurate route if you turn on “Transit” under layers, which can be verified by checking King County’s GIS map (https://gis-kingcounty.opendata.arcgis.com/datasets/d75a66e2ec12429583bfc8686f382c66_0/explore?filters=eyJOQU1FX0wiOlsiU1QgTGlnaHQiXX0%3D&location=47.653180%2C-122.303059%2C15).
If you had stations at the hospital (south of the Triangle/NE Pacific), the HUB/HUB Yard, and 45th (slightly north of current U-District station) then they would be spaced about 0.5 miles apart – but I am not sure the demand at the hospital/stadium is worth a station if the HUB exists. Walkshed would be even worse with 3/4 cut off by water or parking lot, and for the 1/4 you do get all of your potential riders are going to be closer to the HUB or probably walk there rather than trek across NE Pacific. Don’t get me wrong, it would be phenomenal to serve the hospital directly, but a mid-campus station is so much more important. The HUB should have been the starting point and Medical Center the nice-to-have, not the other way around.
“The problem is, Montlake is a bit hemmed in with the parks. ”
If Seattle had done it right, the first section of Central Link would have been Judkins Park to Montlake,* with both those two train stations directly under freeway median bus stations. That would have allowed Eastside express busses on 520 and I-90 to drop their passengers and turn around on the freeway, saving huge service hour cost by not driving on streets, while riders got a quicker trip downtown after the connection. Stadium Station intercepts I-5 and the Sodo Busway, Montlake gets riders from U-District, Ravenna, etc. as soon as they cross Montlake Bridge, and Judkins takes riders from Rainer Valley off of the 7. You naturally build a rider base from bus connections for when the system extends out all those corridors. Urban stop spacing (average 1/2 mi) from Stadium to Aloha/23rd (depending on your route east of Westlake) means three stations across F. Hill/Cap. Hill/Central; then the Montlake/Judkins freeway stops are spaced farther based on terrain and land use. Once people got a taste of a real subway there would never have been an at grade section.
*Bonus points if the first section crossed the canal to UW or U-Dist, since it was always going to be the strongest corridor in the system, but saving the underwater tunneling for later would have been fine as long as the first route captured the key transfers at either end and covered the dense core properly.
Onux is exactly correct, though if the main campus station were “under the HUB”, it should actually be a couple of hundred yards toward the Med Center. That’s 24 hour demand that the underclass part of the University just doesn’t have.
But, for sure, UW Station is in a lousy place for everything except football games. Oh, and the Crew.
asdf2, exactly which is why I think the best policy for Seattle is to give in to the ‘Burbs, axe WSLE and DSTT2 as planned, keeping a Westlake to somewhere northwest automated “stub” with a small MF in the truck parking area west of Ballmer Yard north of the Magnolia Bridge. Use the WSLE money for roadway and access improvements to develop whatever makes the most sense along the lines of Ross’s and Martin’s BRT fan. Spend the rest of North King’s portion of the 2045 revenue stream to get as far toward Ballard as is possible. That might be Interbay or it might be Smith Cove.
Then, with greater taxing authority and the enthusiasm for transit that Seattle voters consistently exhibit “Go It Alone” without the other Sub-Areas. They will have their services, and they will be exactly what they asked for, but they will no longer be able to slag the decisions within the City.
With a decent transfer directly over Elliott the D could become an all-day BRT version of the 15, and with two other less-frequent but useful lines from Magnolia, Red Paint could come to Elliott and Denny. Folks headed to Uptown or SLU would change to the train.
Once the City has command of its future — and the funds to shape it — the next step would be to extend the stub into First Hill, at least three stations away. Yes, you CAN have an “Uptown” station at Sixth and Seneca and with SkyTrain-type short bodied trains still make the turn up First Hill. If anyone has noticed , Sixth and Sixth and Seneca is the highest point west of the freeway within downtown Seattle. Yeah, you have to get under I-5, so it can’t be as shallow as would be preferable, but probably the platforms could be at -40 street level and it could pass under the highway. The ramps to and from the reversible lanes are at the same elevation as the southbound main lanes. They drop north of Seneca.
First station would be at Marion and Minor, and yes, it would be fairly deep, but probably not any deeper than Beacon Hill; you’re starting “up” from a high point. Turn south to a station between Harborview and Yesler Terrace and stop for a while.
Next, finish the line to Ballard, with a western approach as Ross and several of us have suggested. I think that the big “holding” tunnel makes a train tunnel difficult. There is also the siphon, but it’s 120 feet down so could be overpassed. Probably.
I’d say, “What the heck, make it lower and let it open!” It’s not like the Interstate Bridge where high stuff passes regularly. Give fifty foot clearance and you’ll pass 95% of craft without opening. To be aesthetic an opening bridge at that height would have to be a bascule with wing counter-weights. Like the Morrison Bridge in Portland. The weights are hidden in the structure.
Yeah, that means that the train enters Ballard way up in the air, but a western approach gives it some time to descend to a decent station height. Put the main station between 22nd and 24th above 56th and then curve at 22nd and run the guideway down Leary; it’s wide enough for supports.
An Automated Light Metro can have a lighter structure than full on Link, and there’s no overhead to blight the sky with wires. Put another station at 14th and Leary and let that be the terminal for a while.
Something to consider would be double-decking the Ballard Station and trackway, and include a diversion stub for a line toward 15th turning north to complete the original design.
Once that’s done and some more money has accumulated, turn back to First Hill and punch the tunnel out of the hill, run elevated down Rainier with stations at Dearborn, Judkins Park, College and Mt. Baker. The “missing Mezzanine” could be added to Mt. Baker for access at platform level from the ALM.
Finally, go back north and start working east from the stub at 14th and Leary. Put a station by Fred Meyer, and one at 42nd and 3rd NW. You’d probably have to take a row of houses on one side or the other of 42nd, and certainly a house on the east side of 3rd, because you’d be back underground from here on.
Underground stations are massively more expensive than elevated ones, even for SkyTrain-sized systems. So the next station would be for Fremont downtown, and it really couldn’t serve downtown very well. Put it between 38th and Fremont and the Troll Plaza, with access from new stops for the E on Aurora and also to Fremont and Troll Way. Leaving there it would have to curve pretty sharply to stay underground, but it would follow the Bridge Way right of way to 40th and Stone Way for another station there. The next would be about 45th and Wallingford.
Heading east from there the line would ideally have a station just east of the freeway for all the buildings around there, but perhaps just make the box initially. “New U-District” would lie east of the Spine tracks in the 45th Street right of way. An eastern entrance in the corner of the campus would be provided.
The line doesn’t end there, but instead continues on east under 45th to a station between the 45th Bridge and just east of McCarty Hall. Some sort of grade-separated Pedestrian access to University Village would duck under 45th while it’s still up in the air and then cross 25th.
Maybe it goes farther someday. To Children’s? Across the Lake?
This is doable physically and serves the City superbly. Yes, I get that dogleg lines are usually not a good idea, but Lake Union and Queen Anne Hill itself make any sort of grid between the western part of Seattle and North Capitol Hill/the University impossible. So wrapping around Queen Anne Hill will serve a lot of quirky rides, as well as providing some genuine relief capacity to the system.
Exactly Mike. None of that curler & flip flop crowd wanted on campus. The UW didn’t have much of a say on the Montlake location because it’s a State highway. They were so enthusiast about the connection to UW Medical they wouldn’t allow a station connection to the triangle garage. Obsensibly for “security reasons”. As for the Football boosters it doesn’t help tailgating so what’s the point. In the mid 70’s I was able to take a Metro bus direct from the DT Redmond P&R (is that lot still there?) and be dropped off right in front of the HUB. A few years later UW put the ixnay on that and adopted the no transit, no way , no how policy it holds today.
“This is why it is common in dense, urban areas to have stations around 400 meters apart. ”
Not at all.
What? You are saying it isn’t common to have stations 400 meters apart? Yes it is. In New York it is quite common. Paris too (of course). Even in Seattle it happens. Westlake and Symphony stations are about 400 meters apart. If we had built the station in Madison (and I wish we had) then you would have four stations in a row, all about 400 meters apart. That doesn’t mean 400 meters is ideal. Quite often this means a bigger gap following it. But depending on demand and the geography (including the other lines and their stations) it is often the best option.
The main point is that people have it backwards. As that Reddit post put it, in most of the world, the focus on speed *after* focusing on coverage. Basically you cover things as best you can, and then you think about making it faster. In the case of East Link, for example, you would want fast trains since it is bound to have a huge gap over the lake. But you still focus on covering the areas on either side as best you can. You never think “Oh wait, that will make long trips a bit slower” because increasing coverage saves way more time for riders.
This is the way they make most subways around the world. There are exceptions, but fairly rare outside of this country. This is telling, as the USA has the worst transit in the developed world. We have tried different things and they have often failed. You mentioned a great example: BART. BART has terrible stop spacing. They completely ignored the fundamentals. They prioritized long distance trips over short ones. To be fair, no one had every built a system like this. It was reasonable to assume it work well and transform the region. They were wrong. The experiment failed. They should have built a more traditional subway (in San Fransisco and East Bay) even if it never got out to the more distant suburbs. It would have been a much better system. To be fair, the one thing they got right was the fast trains. But that should have happened either way (for the crossing). But especially in the East Bay there simply aren’t enough stations. This is the case with BART and it is the case with Link.
The train goes not go around campus, it cuts straight across it, it even passes directly under the HUB!
I thought the station at the HUB was rejected because the trains would screw up the physics experiment. I just assumed it was diverted to the east (and didn’t notice how Google draws the lines). So yeah, the HUB would have been so much better.
I am not sure the demand at the hospital/stadium is worth a station if the HUB exists
It is because of bus integration. The other option would be to have a station at Montlake, on the other side of the bridge. But I don’t think that gets you that much. Most buses would probably keep going to the UW anyway. Given that, you might as well serve the hospital at the same time.
@Ross
“What? You are saying it isn’t common to have stations 400 meters apart? Yes it is. In New York it is quite common. Paris too (of course).”
That is exactly what I am saying, and the data is quite clear that 400m station spacing is very uncommon, not quite common, not even in Paris and New York. These two posts show graphs of subway/metro station spacing versus speed:
https://www.reddit.com/r/transit/comments/113n0ee/average_speed_of_various_metro_lines_around_the/
https://www.reddit.com/r/trains/comments/1l9lmy0/the_further_apart_the_stations_in_general_the/
As you can see, nowhere in the world are there subways lines with spacing less than 500m except for two lines in Paris, both very short (indicated by the error bars; shorter lines have a greater change in speed if trip time is a minute or two different.) Paris is by far the most closely spaced system, and its average is about 600-700m. NYC the average is 0.53 mi/850m. The average worldwide, as I stated, tends to 1,000m.
There are always exceptions, and NYC and Paris certainly have a few places where stations are 300-400m apart. However 1) These are enormous metro areas with some of the highest population/job density in their core areas; many(most?) walkable, urban metro areas with millions of people do not have places as dense as downtown Manhattan or Chatelet 2) These systems are very old in their core, and built under unique circumstances that restricted line length and depreciated average speed at the time: the NYC subway was approved when NY was just Manhattan and Brooklyn still an independent city, and commuting from outer boroughs was much rarer because of the need to take a ferry; in Paris meanwhile, the subway started when the city was still surrounded by a wall (!) and thus there were effectively no suburbs to reach because the boundary of the city was fixed and everything beyond was open for fields of fire or shanty towns. If you never expected your subway line to run very far then close spacing in the core and slower speeds are not a negative. Basically every city that started their system later or didn’t face these constraints has gone for longer station spacing, and even in NYC and Paris lines built more recently have gone for longer averages 3) Exceptions don’t prove the rule, and a particularly short spacing right near Wall Street or at a station originally only open on horseracing days doesn’t mean much. The graphs I link to show the real story.
“Even in Seattle it happens. Westlake and Symphony stations are about 400 meters apart. If we had built the station in Madison (and I wish we had) then you would have four stations in a row, all about 400 meters apart.”
Symphony to Westlake is 554m (as best I can measure without access to surveyed drawings) not 400m. A station on Madison in between Pioneer Sq and Symphony would be way too close spacing. Look again at those BART diagrams. The area within 1/4 mi of the station is completely covered in dots for people who started their trip walking to the station. The area from 1/4 to 1/2 mi is largely covered with dots for walking, biking, or bus, except for maybe at the outer edge. If people are *commonly* walking 1/4-3/8 of a mile to the station then you do not need to put stations 1/4 mi apart, because that would have them walking only 1/8 of a mi, and they *will* walk farther to good transit. You are better off spending the money on that station somewhere else where spacing would otherwise be too far.
Part of the issue here might be that most of us are referencing the interstation distance, usually taken as the distance along the tracks between platform centers. (I say might because I do not want to put words in your mouth, just have a good discussion and introduce a possible area of misaligned definitions). In dense systems like NYC or Paris, or unique conditions like the curve between Symphony/Westlake, you can end up with stations are within 400m of each other at a straight line distance due to overlapping or closely spaced lines, even though actual station spacing on a line is greater.
In the big picture we agree: coverage is more important that pure speed in building a high ridership system that supports (and is supported by) dense walkable urban neighborhoods. BART did get it wrong with far flung lines to the suburbs and stations every few miles instead of a solid network in SF/Oakland/Berkeley. Link is repeating the mistake, and as I said from the very start U-link was/is missing 2-4 stations. However, getting coverage does not mean you have to overdo it. Referencing once again those BART station access dot maps, station spacing of 800-1200m gets you the coverage you need. At that point the spacing/speed graphs I link to here come into play. Increasing spacing from 500-1000m takes average speed from 25kph to 35kph. Over a distance from Northgate to Westlake (4.3mi/6.9km) that is the difference between a 12min trip and a 16.5min trip – saving 4.5min/27% is a lot. Coverage should come first, but trip time influences mode share as well, so once minimum coverage is met you need to look at speed, and 400m station spacing is just too close and too slow.
“It is because of bus integration. The other option would be to have a station at Montlake, on the other side of the bridge. But I don’t think that gets you that much.”
As I noted above, the key for bus integration at a hypothetical Montlake station would be express busses on 520. Keep the busses moving on the freeway with a free connection from a dedicated freeway median platform direct to the train platform. I don’t see anything about the UW station as built/the triangle as key for busses. 45,67,75,372 all use E Stevens and basically stop in front of the HUB. 44 stops at U-District first so why stay on the slower bus to transfer later? 43 and 48 both pass Montlake first from the south, so why stay on the bus to cross the bridge to transfer if a Montlake station existed? Same for all of the Eastside expresses using 520.
I remember when I sat across a table at lunch across from a lifelong transit fan and rail engineer in Boston in 1989. He suggested to me that it’s not so much the station spacing as it is the number of times a rider must stop to reach their destination. He felt that the number was 10 intermediate stops or maybe 12. Once a line got too long, it no longer felt as useful to a rail rider.
From Westlake, 10 stops is Lynnwood (counting Pinehurst). From CID, 10 stops is Redmond Technology and TIBS (counting unbuilt infill stops) or Angle Lake (current). So that concept to me makes reasonable sense. Even 12 would be Ash Way, Downtown Redmond and Angle Lake using the methods above.
Of course, electric rail stops go much faster than diesel stops do so maybe it could be pushed to 14 or 15.
But 19 proposed stops between CID and Tacoma Dome or 16 or 17 from Westlake to Downtown Everett certainly begins to feel overly onerous. The continuation of express buses or hope that Sounder can offer alternative rather fit this line of reasoning too.
Places like New York and Chicago compensate with limited stop or A/B stop trains.
Maybe he just planted the idea in my head. But ever since, I’ve found station/ stop counting to be a valid assessment on urban rail utility. I’m sure our station planning would be different if ST had set a ceiling on the number of stations from Downtown Seattle rather than put them in because that’s what the best spacing is perceived to be.
This is an odd take. Will you not drive somewhere because you have to stop at too many stop signs? Yes, number of stops can impact perceived travel time, but the idea that someone won’t make the trip above a fixed number of stops is silly.
“This is an odd take. Will you not drive somewhere because you have to stop at too many stop signs”
If you’re forced to, then yes. However a normal person will look to alternatives. Is there a shortcut, or a freeway option available? They may also consider cost and taking the lower mile option, even if it is slower. That could include transit as an option too.
But for people who worry about time more than cost (which is a lot of people), public transportation needs to at least compete with a car. The more stops that a bus or train has, the slower it gets. A car has to stop too, but not nearly as long as a bus or train. A car also makes a direct route to the destination, while a bus may not… Possibly requiring transfers, and it may not drop you next to your destination either. This is true in high density and low density areas.
If we want to win ridership, transit has to be competitive with cars. Freeway travel needs a fast replacement, ideally connecting every major transit center. An express bus is good if it doesn’t get stuck in traffic. Link is good too, but if it runs super slow like it currently does – it’s not viewed as a valuable long distance travel option. For short distance, it’s tolerable but it still adds unnecessary travel time for trips that can easily be done by bus.
As for buses, buses have to stop for the same reason as cars – in addition to picking people up.. if changes were implemented to keep buses moving fast (except when picking up riders) and reduce dwelling time, it might as well even be faster than taking your car.
If someone wants to go to the UW, it could already be a tedious 30 min bus ride to the Link station from an area a few miles away from downtown. Do you think the average rider wants to put up with even more time on Link, when they previously got there super quickly from Westlake?
If adding more stops there should at least be a concept of express trains that bypass lower ridership stops, with timed transfers. That’d enable a fast trip between any two stops with just one transfer
All vehicles have to stop for traffic signals, congestion, and pedestrians.
A bus can skip congestion if given the right infrastructure.
Additionally a bus should be actively picking passengers up while it’s not stopped for other reasons. It should rarely be stopped just because it needs to pick passengers up. Good traffic signal algorithms will allow for red lights to line up with bus stop times, and to be green in advance to clear the intersection so the bus can proceed through at full speed when it reaches.
Sometimes it’s unavoidable to stop extra, but that’s where the congestion relief measures help: bus lanes, queue jumps, etc.
A train naturally doesn’t have these problems. It has the potential to move people really quickly through unusual pathways especially if it’s a subway. And it would always beat a car and bus. But that changes if you put a ridiculous number of stops. The current routing after Westlake at most could fit a First Hill stop (which isn’t even in the most useful center of First Hill anyways). And after that I don’t see any valuable stop to add.
Link is already slow in downtown, unfortunately, but that’s the nature of a downtown. People need stops to all the key parts of downtown.
For driving, how is this bypassed? Express lanes. Similarly Link needs an express version. Focus on the key stops, SODO, Chinatown, Westlake, Capitol Hill, and UW. Skip the rest. It moves people through quicker for those headed further south to the airport, or east to Bellevue. They can also transfer reliably to the next “local” train after deboarding, in a matter of minutes since downtown frequency is already 4 minutes.
“ This is an odd take. Will you not drive somewhere because you have to stop at too many stop signs? ”
When I drive, I do choose paths that minimize the number of major signals that I must pass through. I don’t think everyone does that — but some people probably do too.
“ Yes, number of stops can impact perceived travel time, but the idea that someone won’t make the trip above a fixed number of stops is silly.”
Have you ever ridden Chicago’s rail transit system from an end station? Several lines have at least 15 stops. I find it excruciating to ride. It may not ultimately drive riders away but it can get irritating.
@Onux — The graph shows *average* distance. This means some are smaller and some are bigger. To be clear, I never said 400 meters was average (although it appears to be for at least one line in Paris — one of the finest metros in the world). I merely said it was common to have stations 400 meters apart. Look at Toronto for example. It has an average of about a kilometer for some of its lines. Yet if you look downtown you can see that the stations are quite close together as you approach the water. Same goes for Copenhagen. It isn’t that had to find stations that are about 400 meters apart — sometimes closer. I’m not going to look at every system but so far I’ve always found stations like that in every system I’ve looked at. New York, Chicago, Toronto, London, Paris (of course), Madrid, Copenhagen and Seattle. They all have stations that are that close together. Look at the Forward Thrust map again. It sure looks like some of those stations are that close. That is just normal.
But that doesn’t mean it is the average or some sort of standard. That is because there is no standard. That is what the person on Reddit explained so well. The focus is on getting as much coverage as possible (for as little money as possible). That is the normal thing to do. Then you start thinking about speed.
Some agencies in America focus instead on speed first (for long range trips) but I’m saying that is a stupid approach and it has bit them in the ass over and over again. It is one of the many reasons America has such poor transit. We keep trying new things that don’t work. Or we see a mode overseas and ignore why it works there (and why they used that particular tool). Then we apply it here as if it is fundamentally better. We’ve done that with streetcars, light rail and BRT. I suppose you could call this an S-Bahn imitation but again, there is no attempt to figure out why it makes sense in Europe and is a huge waste of money (and poorly implemented) here.
A station on Madison in between Pioneer Sq and Symphony would be way too close spacing.
Of course. That is my point. We could have easily built a station there. Our system would have been better because of it. The walk from Madison would have been shorter. The integration with the buses would have been better. Is it ideal? Of course not. No system is ideal. Things got a little weird because of the bend in the line. It didn’t make sense to put a stop at 3rd & Pine so the Symphony Station (known previously as University Station) is a bit to the south. It is still pretty close to the Westlake Station. According to Google maps I measure the distance between the two stations at 420 meters (based on the symbols). By the way, this is the same distance between the old Convention Center Station and Westlake as well. With the station at University, the station at Madison would have been very close. But it would have filled in the gap between University Station and Pioneer Square which is they they proposed it in the first place. Again, this is common.
This is really a side point. I merely mentioned (in passing) that it is common to have stations really close together because coverage is more important than speed. That certainly doesn’t mean it is average and it certainly doesn’t mean it is a model. Because again, there is no model. Look at the Forward Thrust map. Some of the stations are really close together. Some of them are far apart. What is the average? What is the standard? Who cares! That isn’t the issue. The reason the stops are so close together in some areas is because it is important to cover them as best as possible. Sometimes, because of various geographic issues, you end up with stations really close together. The reason there is a big gap in Montlake is because it is so hard to cover that area. And I’m not saying that the Forward Thrust map is ideal but I’m saying it is a hell of a lot better than what we have because we ignored that fundamental idea. They have three stations between the UW while we have only one. It may seem easy to blame “The Spine” for this but ask yourself. If there were never any plans for the train to leave the city would they have only one station between the UW and Westlake? Of course not.
I’m not sure why you keep referencing BART given it is famous for flawed station spacing. But it does serve as an example of why the street grid mattes. Those little circles showing distance are meaningless. As Jarrett Walker has explained, you want to look at walking distance*. Typically this results in diamonds surround each stop. In the case of BART in San Fransisco though, you don’t have that, because the grid is turned at an angle. It gets complicated. In any event, I think it is obvious that BART would be much better off if the stations were closer together in San Fransisco (even though it isn’t the systems biggest weakness). The experiment failed. Only a handful of trips in BART dominate ridership and they are all within the urban core. For whatever reason the more distant suburban places account for very few rides** — despite having a system that is clearly designed for their benefit in terms of speed. Maybe it is because speed doesn’t really matter — even for them. Maybe they too would be better off with a system that is slower but involves a lot less walking.
Of course stop spacing is a balancing act. Ideally you have a system with multiple lines and stations spaced so that they complement each other. Lots of coverage and minimal overlap. There are a lot of considerations such as cost and local geography. It doesn’t make sense to put stations where there are very few people. But there is no balancing act when it comes to speed. If that is your focus, you are doing it wrong. Double track if you think there will be a ton of riders making really long trips. But if you try and skip stations because you are afraid riders from farther away will be delayed too much you will end up with a crappy system (like so many in this country).
*The book(s) cover this well and he has made several posts about it. Unfortunately, the website that holds his images no longer has them. So it isn’t as easy to understand.
**Just nine stations between San Francisco and Berkeley account for half of all rider on/offs. I borrowed that quote from here.
Have you ever ridden Chicago’s rail transit system from an end station? Several lines have at least 15 stops. I find it excruciating to ride. It may not ultimately drive riders away but it can get irritating.
Yeah, sure. But those riders typically make up a tiny subset of the overall ridership no matter how you try and accommodate them. It is fine if you want to run express buses or commuter rail, but even double tracking your standard metro (so that some trains can skip stops) is usually not worth it. What is a really bad idea is to leave out some stops just to favor those riders. I can’t think of any system in the world where they think “We have too many stations — it takes too long to get a long distance”. The Paris Metro isn’t going to have a stop diet. Metros are just so fundamentally fast for the vast majority of trips that people take that you want as many stops as you can afford.
Instead the various agencies focus on infill service. They add a bunch of stations, often relatively quite close to existing stations, just to get a little more coverage. A good example is the Second Avenue Subway. They also try to build different lines going different directions so that riders can avoid having to go out of their way for a common trip. A good example is the Interborough Express. But no one in New York or Paris regrets all those existing stations. I’m not saying there aren’t stations that are now obsolete. But having the stations so close together — having such thorough coverage even though some trips take a long time — is a very good thing.
On the other hand it is common for American agencies to regret the overly large stop spacing. BART, DART, Denver RTD and yes Link are like that. In various discussion boards you will have people regretting the lack of stations. Or they will be unfamiliar with the system and ask why there is such big gaps. This happens a lot with people from overseas — it is like asking us why we still don’t use the metric system. The answers are usually complicated but the conclusion is always the same. The system would be better if there were more stations (and it was more like a European metro).
Consider East Link for example. There will be one new station in Seattle (at Judkins Park). A trip from Downtown Bellevue to Downtown Seattle will be quite fast. But imagine we had a station at Rainier & Jackson. The line would be a lot more useful. Yes, riders from the East Side would be delayed — ever so slightly — on their trip into the city. But some of them would take advantage of that stop. I’m not saying they should have done that — I’m just saying it would be a much better system if we had a station there. But it was cost, not speed that was the reason there isn’t a station there.
@ Ross:
East Link does not have an excessive number of stops. I have even said that they should have added one underneath the 12th Ave bridge with elevators or escalators both up (to 12th Ave S) and down (to Dearborn).
RE Symphony vs Madison: Rather than add a Madison stop, Symphony Station would have been better if it was exactly halfway between Pioneer Square and Westlake entrances. But it is where it is probably for other reasons. And I’d gladly support punching through a walkway further southeast from the mezzanine from Symphony to get an exit closer to the Central Library (in lieu of DSTT2).
There really aren’t any definitive rules about station spacing as you say. To me, it’s about potential ridership demand. If a station can generate 4000 new boardings I don’t much care how far away the next station is. But if a station spacing is 2 miles and there’s still anemic ridership, sure I’d suggest even further stop spacing. But I would encourage ST to make sure that track grades and curves could accommodate more possible infill stations in the distant future. It seems like a modest cost to do that.
Spoken like a selfish man who lives within the walkshed of one of those rare stations.
His point is a train can’t serve everyone, and trying to do so is ABSURD.
Buses can serve every single surface street and connect you to a station in just 1-2 trips! If running very frequently throughout the city, everyone can get to the nearest station in a matter of minutes. But that also means people have to walk. Transit is all about walking to your stop. If you pick up everyone right at their doorstep, it might as well be able to cycle to your destination.
Often times it doesn’t even make sense to user Link for inner city mobility if we kicked cars out and made buses really good. But for a trip like UW to Symphony, for example, that’s where Link gets useful. Seattle isn’t like NYC. We made the roads and buses can get around if we kicked out cars, and timed bus stops to match with pedestrian crossing times / red lights.
Link should have 1km spacing at most (600-800m if the destination is critical), like most civilized systems. And that would be within downtown. Once you’re out of downtown and lose linear / radial density, it should be even further spaced and have express buses / P&R connections. It’s meant to take you somewhere far enough so you can hop onto the bus system in that area and get to whichever place you need to go.
If you just look at a map, the areas after Westlake are not nearly as dense as downtown. There’s zero reason to build more stations when it doesn’t even serve most of the people in a walking distance.
Instead of building more billion dollar stations, we could instead for much cheaper:
Make sure every street in Seattle has these three type of buses:
* N-S bus route along the entire street
* E-W bus route along the entire street
* One seat connectors to the most convenient Link station (or any other major destination)
High frequency, smaller buses instead of articulated. Perhaps automated trolley buses if that technology is available and trusted.
Kick SOVs off downtown streets except for people that live there / service vehicles, and trim down street parking as much as possible. Add bus only lanes / roads, TSP
For suburbs, you can’t serve every street. Add DART service to demanding neighborhoods, add express routes that focus on connecting two major transit centers (hubs), and add local routes that focus on commercial centers, key destinations, and apartment complexes. Essentially a hub and spoke design, with lower frequency and more articulated buses.
If you just look at a map, the areas after Westlake are not nearly as dense as downtown.
So you are saying we shouldn’t have built a metro in the first place?
His point is a train can’t serve everyone, and trying to do so is ABSURD.
Agreed. No one is saying otherwise. It is pretty easy to argue the line is too long — it goes too far from the urban core. But that isn’t the point. A metro should serve as much of the area along the route as possible. Of course there are times when there is basically nothing between the gaps. East Link is like that. There is no point in having a station in the middle of the lake. But otherwise you should do your best to cover it as well as possible. Otherwise you have to ask yourself if it is worth building.
Link should have 1km spacing at most (600-800m if the destination is critical), like most civilized systems.
Most civilized systems cover things in the way I describe. Even North American systems have stops closer than 1 km. New York, Chicago, Boston, Toronto, Montreal — the stops are all closer than that. Go across the pond and look at European systems. Paris, Rome, London, Madrid. The stations are close together in the urban core.
That’s because once you commit to the expense of building a line you want to increase coverage as much as possible. Imagine you are building a five mile metro line in the middle of the city. It is a normal subway — it is significantly faster than a bus or driving. All other things being equal, how much of the area along the line should you try and cover? All of it! Every last inch. You don’t want someone thinking “Damn, it is too far to walk to a station — I guess I’ll take a bus — or drive”.
If there is reasonably fast, parallel service, then you might end up stretching things out a bit. You just assume that for some trips they will take the bus (or tram) on a parallel street (they won’t take the metro at all). Of course if there are multiple metro lines then you can also get by with wider stop spacing (just as you would with a bus). Cost is always a big factor and you might end up shortchanging some areas (this happened with our downtown tunnel — we were supposed to have a station at Madison but it became too costly). Likewise, some areas have very little by them — for example they are next to a body of water. Ironically, this often leads to stations being fairly close together (less then 400 meters) because there is a big gap after that (and you might as well put the station closer to the other one). But the goal is to serve as much of the pathway as possible.
Of course you have to consider the impact of the network (and potential crossing buses and trains). You can see this with the Canada Line (a fairly typical metro in terms of stop spacing). In some places there is a really big gap. Usually this is because there is no cross street and there is a big park in the middle. It is worth noting that they plan on adding infill stations at 33rd and 57th despite this. Sometimes the stations are really close together (400 meters). That’s because it is essential to cover both areas (and both streets have significant crossing bus service). There is probably only one station they should have added — at West 16th. The gap between Broadway and King Edward is too large. I have no idea why they didn’t (no system is perfect). The point is, it would be much better with that station.
Once you’re out of downtown and lose linear / radial density, it should be even further spaced and have express buses / P&R connections. It’s meant to take you somewhere far enough so you can hop onto the bus system in that area and get to whichever place you need to go.
Right, but to do that you need to have the stations in the city. Otherwise you are just forcing another bus trip — and not necessarily a good one. Just consider that Canada Line example again. Assume you are trying to get from 16th & Cambie (where there should be a station) and Richmond Hospital. It takes about 45 minutes, and that is with 27 minutes of walking. You might get lucky and shave some time off that trip (by taking two buses) but it isn’t going to be great. If you move to the side it gets worse. This must be the “almost” in the almost perfect grid. This is a trip from a fairly dense inner suburb to the heart of the city. This is bound to be a fairly common trip and there are bound to be a lot of similar trips. It is the type of trip that gets people into their cars (note how fast it is to drive despite all of the intersections and traffic lights). This is clearly a flaw with an otherwise very good transit system.
Now would it better if they skipped half the stations so that riders from Richmond could get downtown faster? NO! Absolutely not. You would cripple ridership. Not only the Canada Line (which carries over 100,000 riders despite being much smaller than Link) but the overall network as well. Even in Vancouver — a system that is much better than ours in every respect — they could use more stations on their metro.
But this particular flaw with the Canada Line is nothing compared to what exists in our system. This is what it is like to get from a typical suburban location to First Hill. Two transfers, over 45 minutes. About a third of the time is spent just getting from Capitol Hill to First Hill. That means riders would save around ten to fifteen minutes if there was a stop there. Or consider something that is bound to require a transfer: Roosevelt to Garfield High School. This is a typical urban trip from one high school to another. The trip takes about a half hour, at noon. By far the longest section is trying to get from a station to the high school (on a bus). What is true of Garfield is true of everywhere on 23rd. If there was a station at 23rd & Madison it would not only serve an urban neighborhood but it would dramatically improve the transfer for all of the people that live to the south. These are the types of trips that allow an urban line to carry massive number of riders. It is the main reason our system lacks high ridership. Not the fact that it is light rail or runs on the surface — it is the lack of stations.
We were so focused on trying to get to distant suburbs that we failed to properly cover the urban core. There are fewer potential riders in the suburbs. We are basically chasing specs of gold while ignoring the nuggets under our nose. Not only that, but the suburban riders come out behind as well. Sure, there trip to downtown is a little bit faster. But their trip to First Hill is much, much slower. There just aren’t enough stations.
Instead of building more billion dollar stations, we could instead for much cheaper:
Yeah, there is a very strong argument that we shouldn’t build any rail outside the city. There is even a strong argument that we went too far. But if we do build something, it should have lots of stations — otherwise it probably isn’t worth building.
Just look at West Seattle Link. If West Seattle Link consisted of a couple dozen station (and several branches) then it would be a very different conversation. No one in their right mind would say we would be better off with buses. But there are so few stations (three) that the vast majority of potential riders in West Seattle would be better off with better buses. What is true of West Seattle is true of Issaquah-Kirkland Link (for much the same reason). Tacoma Dome and Everett Link are slightly different. There aren’t enough stations to justify the cost but even if they added a bunch of stations they wouldn’t get that many riders. Everett and Tacoma are just too small to justify a subway system.
“Most civilized systems cover things in the way I describe. Even North American systems have stops closer than 1 km. New York, Chicago, Boston, Toronto, Montreal — the stops are all closer than that. Go across the pond and look at European systems. Paris, Rome, London, Madrid. The stations are close together in the urban core.”
King’s Cross/St. Pancras to Angel on the Northern Line is 1.2km. To Farringdon on the District/Circle/Hammersmith lines (oldest stretch of subway in the world!) is 1.7km. Regent’s Park to Oxford Circus on the Bakerloo is 980m. Holborn to Bond St on the Central Line, right underneath Oxford St, the **busiest shopping district in Europe**, is 2.13km for 3 stations, average 710m. Shadwell to Tower Gateway on the DLR, 1.2km. St. Paul’s to Chancery Lane, 1.05km.
These examples all fit within a radius no larger than Westlake to Madison & MLK, at the center of a city with 50% more people than all of Washington State, covering a density of population, jobs, retail and entertainment unheard of anywhere in the densest part of the urban core of Seattle, let alone anything between downtown and UW (or inside of downtown or U-District for that matter). Can you find examples of stations closer than this in central London – yes, of course. But you can also find station spacing that is greater not too much farther out. 1km, or greater, is in fact the average station spacing for almost every system in the world. For London it is 1.3 km across the system, including “suburbs” that are probably denser than any “urban village” in Seattle. The median distance is 1.05km. In the central part of London (inside of the Circle line) there are the crazy Charing Cross-Embankment examples of a few hundred meters, but it averages to 800m. This is all right in the range of 1/2mi to 1 km I have been saying.
First, let me state where we are in agreement: Subway systems (aka mass transit, rapid transit, heavy/light rail, etc.) work best when then serve reasonably dense and walkable urban areas. They also work best when they provide good, continuous bands of coverage to that dense, walkable urban fabric (this can also be stated as the walkshed of each station should meet, or slightly overlap).
Based on this agreement, we both also agree that the U-link extension was built without enough stations. The First Hill/Capitol Hill/Central District areas are a dense, walkable urban neighborhoods, yet the 1.8km between Westlake and Capitol Hill stations and the 3.6km between Capitol Hill and UW stations result in walksheds that do not meet. As a result some of the dense, walkable urban fabric is unserved because the stations are too far apart/needed stations are absent, and this results in a much less useful system and a large amount of lost ridership potential.
Where you are making a mistake is concluding that if stations can be too far apart then it is always better if they are closer together. It is not, for two reasons:
1) Coverage should not be sacrificed for speed, but speed is still important. Trip time is a factor people evaluate when making a mode choice (along with others: price, convenience, waiting time, etc.) If your speed gets to be too slow, then people will chose other means of travel, even if there is a station close by.
2) The objective is for walksheds to meet or gently overlap, but it is not necessary to overlap completely. Walkshed is a (relatively) fixed geometric factor, a result of human biology (mechanics of walking) and psychology (perceptions of time and distance). Pretty much everyone will walk at least 1/4mi or 400m to a rapid transit station, and that most people are willing to walk up to 3/8-1/2mi or 600-800m, with ridership dropping off at the edge of that distance and certainly beyond it.
Neither #1 or #2 are absolute. Two separate subway systems with similar stop spacing can have different average speeds based on technology, terrain, etc. Walkshed is a gradient that not only varies based on fixed conditions (hill or no hill, street connectivity, sidewalk and station entrance locations) but over time (how far you will walk to a concert as a young person will be different than how far the same person will walk to work every day when old).
Still, there are broad trends that hold. From the diagrams on the two Reddit posts I shared, there is a steep drop in average speed as station spacing goes below ~1000m. The stations get to be so close together that trains don’t have time to get up to speed before they have to start braking for the next station. This gives up one of the main advantages of having dedicated right of way without traffic. What is the point of not needing to stop at several intersections for cross traffic if you are going to stop at them anyway for over-close stations?
Another broad trend is that once your stations are within 1/2-3/4 mi (800-1200m) of each other then you will not gain any additional appreciable riders by making them closer. Look at Page 8 of the pdf here:
https://www.bart.gov/sites/default/files/docs/StationMaps_NonHome_Fremont-WalnutCreek.pdf
You have very valid criticism of BART and its overall topography (neglect of core urban areas for long lines to widely spaced subway stations). The examples I am going to point to are all in the dense core, however. Notice how there are almost no people walking to the station from SW of 6th street (~1/4mi from the station) but the area NW of the station to almost 1/2mi out is packed solid. Why? Because 6th St is halfway between the Powell and Civic Center stations (1/2 mi spacing, look at that….) and so people south 6th walk to Civic Center instead. You can see the same on Page 7, with continuous zones of people walking to Montgomery station from almost a mile away on 2nd and Columbus (two streets that point basically right at Montgomery BART) but empty zones with no riders ¼ mi away just south of Embarcadero/Powell stations (even though we can see on Page 8 that there are plenty of BART riders from 4th St/5th St going to Powell.)
On Page 9 you can see this even more dramatically. Rockridge is in a much less dense area, full of single family homes not downtown SF office towers. But along College Ave (a vibrant mixed use neighborhood business street) you have a solid mass of purple dots showing people walking to the station from a quarter mi north to a quarter mi south, and decent coverage among all modes almost exactly to a half mile down the street in each direction.
Quick note, you are correct as Jarret Walker notes that walksheds cannot be perfect circles due to actual walking distances. But the example above are in areas with highly connected streets (no cul-de-sacs) and thus pretty representative of what you can get.
You are correct that there is no single answer where every station must be laid out the exact same distance as the one before. But putting #1 and #2 above together we can see that there is an optimum: station spacing of ½-3/4 miles allows for the walk up station usage that supports good urbanism without affecting trip time too much. More importantly, those two points show that in addition to making stations too far apart, you can also make them too close together. Below ~1/2mi or 800m, costs begin to outweigh benefits. Your average speed gets lower and lower, but you do not gain any riders, because everyone in the walkshed of a station is already in the walkshed of the adjoining stations.
Thus placing a station at Madison would not have made the system better, it would have made it worse. With a station just ~325m from two other stations, you get:
-Spending the capital cost of building an underground station, which is expensive. Capital dollars are finite, and an extra station downtown means another underground station (or multiple elevated stations) are cut elsewhere.
-Spending the operating costs of an underground station, forever. They require maintenance, the cost of electricity for lights and fans, repairs, etc.
-A ride that is slower by 1-1.5 minutes for everyone crossing downtown, tens of thousands of people a day, every day, forever.
-The extra operating costs from a slightly longer journey.
In return for all of these negatives you get:
-Zero extra riders.
Is it literally zero, probably not. There might be a super lazy person working at 3rd and Madison who will ride link as long as they don’t have to cross an intersection to get to the station. But as a practical matter, a station between Pioneer Sq and Symphony would not add any riders to the system that you could detect from ridership statistics. The evidence from the BART origin maps shows clearly that everyone who would walk to a Madison station is already walking to Pioneer Sq or Symphony.
You might have a good argument why Madison is a better location than the other two for bus connections (Rapid Ride G). But that is an argument for moving a station, not for adding one.
Finally, it is simply not common to have stations 400m apart. I have already explained why NYC and Paris have some for historical reasons relating to city size in the late 1800s. Those cities do not build stations that closely anymore, and Seattle does not stop at fortifications along the Lake Washington Ship Canal. As for London, on or inside the circle line, there are exactly 5 stations spaced less than 400m, 6 more spaced less than 500m, and only 10 between 500-600m. That’s only 21 stations as close or closer than Link is spaced downtown, out of 76 stations, in one of the areas most densely packed with subway stations, with more density than downtown Seattle has ever seen or maybe ever will. That same area has more stations with spacing of 900m or greater. The most common spacing is 700-800m (800m, half a mile, there it is again…) And that is only the stations in the very core, there are about 200 other stations outside it, but still well within the city, virtually all of which have longer spacing. Even in a city with an old packed network 400m spacing is not common.
Can you find examples of stations closer than this in central London – yes, of course.
That is my point! I never said that 400 meters was average or ideal. I simply said it is quite common to see stations that close together — or even closer — especially in a downtown area. That’s because sometimes it just works out that way. The fact that a lot of metros *routinely* have stops that close shows it really isn’t a bad thing. Again, that is the case in Seattle.
Coverage should not be sacrificed for speed, but speed is still important.
Yes. In other words, speed is a secondary consideration. That is what I’ve been saying the whole time (and what the writer on Reddit said). First focus on coverage. Then speed.
Trip time is a factor people evaluate when making a mode choice (along with others: price, convenience, waiting time, etc.) If your speed gets to be too slow, then people will chose other means of travel, even if there is a station close by.
Right. And the “other means of travel” is a regional rail line or express bus. That is, assuming there are enough riders to justify the express bus or regional rail line. If there aren’t, then we shouldn’t worry about such a small subset of riders.
The objective is for walksheds to meet or gently overlap, but it is not necessary to overlap completely.
Yes, and I wrote that. I also cited Walker who wrote the same thing. You want to have coverage without overlap. It is a balancing act and it there are a number of different factors. We can delve into it but we would basically just be rehashing Walker’s blog post (or Chapter Five of his book). Oh, and no walkshed is a circle unless there is no street grid. It is typically a diamond but with BART it is actually a diamond to the south and a half-hexagon to the north.
From the diagrams on the two Reddit posts I shared, there is a steep drop in average speed as station spacing goes below ~1000m.
Yes, and I didn’t mean to suggest otherwise. My point is, it doesn’t matter. Basically, speed is a “nice to have”. There probably is no correlation between speed and the quality of a metro. If anything, there is an inverse relationship. The systems that perform poorly on various metrics are those tend to be fast. That might not be fair — it would probably be the U. S. systems (like BART) dragging down the average. Whatever. The point being, having too many stations is a tiny problem compared to having too few. Which brings us back to Madison.
Yes, it would be wasteful to have a station at Madison. The stop spacing is poor. But you dismiss the value of the station (and why they wanted to add it in the first place). There is a major bus route crossing it. Our biggest buildings are up the hill. It would get a lot of riders which means that a lot of riders are spending extra time enduring the longer walk or they have given up on the system. At some point, someone up the hill just says “screw it” and decides to drive. There are a lot of people up the hill. It is not worth the money but it does add value. You overstate your case when you write:
A ride that is slower by 1-1.5 minutes for everyone crossing downtown
A minute to a minute and a half??! Dwell time should be about 20 or 30 seconds. In terms of speeding or slowing down, the effect is minimal *because* it is so close to the other stations. In terms of speed you actually want a bunch of stations close together, then a big gap, then another bunch (instead of them all being spread out equally). A station like BAR will mean the train will have to come to a stop after hitting top speed. Then it will reach top speed again. The impact on travel times will be much bigger than that at Madison (except the dwell time might actually be short since so few people will use it and a lot of people would use the Madison station).
But this gets back to the original point: Coverage first, speed second. There are several ways to increase speed:
1) Increase the acceleration of the vehicle.
2) Increase the top speed of the vehicle.
3) Reduce dwell time.
4) Skip stops by double tracking.
The first doesn’t typically matter with metros. It does matter with regional rail though (look at Caltrain). Likewise, top speed doesn’t matter for a lot of metros. They don’t have any big gaps. In our case it does matter. East Link is going to have a big gap (because of the lake). Never mind our stop spacing — we should have fast trains just to cross the lake.
Dwell time don’t matter that much if you have very few stops. This is why commuter/regional rail often has long dwell times and fast trains. It is why metros often have the opposite. Our dwell times are way too long. They should be shorter. We sometimes forget what we are building. It is a metro. It is expensive like a metro. It should have shorter dwell times like a good metro.
Double tracking is an option if you have enough riders going long distances. It is probably a good idea in L. A. (I haven’t looked at their long range plans). It is silly in our case. Almost all of the density is within the city. We have freeways that connect to the suburbs. Run express buses and commuter trains for those riders.
I want to explore this idea a bit more. It is easy to assume that someone in Lynnwood benefits by having so few stations. Many do. It is quite fast if you are headed to downtown. But imagine we go overboard the other direction. Imagine we have so many stations that it takes a while to get from Lynnwood to Seattle. So we run express buses (like they do in New York). Problem solved.
Except not everyone who rides the train from Lynnwood works downtown. What if they work at someplace a little bit obscure, like, say the Central Area. Now the express bus isn’t much good. The train is too slow. Except wait. One of the stops they added is at 23rd & Madison. This is actually quite a bit faster for that trip. They get to their destination several minutes before they would now *despite* the fact that they spend more time on the train. So even many of the suburban riders benefit with a system that has a lot more stops.
Of course you can go overboard. But again, it is so rare that I don’t think I’ve ever seen it. The Paris Metro has lots of stops very close together. Yet it is widely considered one of the best subways in the world. It is far more likely that you will screw up and not have enough stations, rather than too many.
Of course the most common mistake (especially in the U. S.) is overexpansion. It even has a section in the critique of light rail systems (https://media4.manhattan-institute.org/sites/default/files/economics-of-urban-light-rail-CH.pdf). This really isn’t about stop spacing although they often go together. BART would be fine going to Pleasanton as long as it covered East Bay a lot better (and there was a better metro in San Fransisco). But part of the reason it doesn’t cover East Bay very well is because they were focused on going to places like Pleasanton. It is worse for some cities though as they struggle to provide service even after building the line there (Denver comes to mind). This was before the pandemic — they just didn’t get many riders. Now Seattle is trying very had to join this club (with light rail no less).
“And the “other means of travel” is a regional rail line or express bus”
You forgot about cars. And that’s what most people are opting to ride instead because transit is too slow.
Just rode the 271.
Just 14 minutes from the Husky Stadium to Bellevue Transit Center.
But apparently I’m supposed to ride Link for 30+ minutes instead.
It would make more sense to through-route the eastside with SLU-Ballard, and Tacoma with Everett, and recognize that northenders will ride the express buses between UW and the eastside, given the choice.
But ST does not want to make operators swap out mid-route. Instead, they would rather make thousands of riders make lengthy transfers mid-journey.
For that specific trip during the day, Route 271 usually makes sense. But station walk time and effort and wait times are important parts of travel time too.
When Route 271 drops to every 30 minutes at 7:30 pm Link is going to be a better bet if you have to wait long for a bus.
If you boarded Link at Northgate and was headed to South Bellevue any time of day it would not. The extra time to keep changing and waiting for vehicles (including the time inside UW Station) takes lots of minutes.
Having a robust set of transit routing options gives a rider the opportunity to choose what makes the most sense given their specific travel times and destinations.
My unrealistic dream is that the 271 (270) gets bus lanes along it’s entire route minus the 520 Bridge where the HOV lanes are fine, and a new flyover (or tunnel) is built connecting the 520 HOV lanes and the 405 express lanes, and the 270 is routed onto it. I believe the 270 has much potential to become a faster than driving option during rush hour, as if Seattle also adds bus lanes and signal lines priority for it along mountlake and pacific it would be entirely insulated from traffic.
@Exponent
> My unrealistic dream is that the 271 (270) gets bus lanes along it’s entire route minus the 520 Bridge where the HOV lanes are fine, and a new flyover (or tunnel) is built connecting the 520 HOV lanes and the 405 express lanes, and the 270 is routed onto it.
it’s not actually that unrealistic. or at least the 520 west leg (seattle) to 405 south leg (bellevue/renton) hov direct connector is the one wsdot wants to build the most for that intersection. though of course they have other projects to build right now.
All of Bellevue Way needs bus lanes. Ideally Stride S1 uses Bellevue Way to reach the transit center. Yeah.. it makes that HOV transit ramp near the TC useless but the connection possibilities are way better at S Bellevue Station. It would also speed up the many Metro buses following Bellevue Way such as 240, 270, 556…
Or if Stride is meant to be a faster connection, leave it the way it is… But add interlined low-ish frequency Metro express service that makes these additional pathways through SeaTac airport, Renton core, Southcenter, and Bellevue Way (revised 560).
And S King County needs some sort of better connection to Eastgate, Factoria, and Issaquah. It’s quite a difficult trip at the moment. Almost every other major city connection is solved except that one.
Stride stopping at S Bellevue essentially solves this issue. The 111 does the job here, but it requires an additional transfer for anyone south of Renton Highlands.
I’m sure those who live in Issaquah or Eastgate would appreciate the easier ride to the airport as well.
The 271 is not going away (it is being renamed though). Eastside – UW trips will still mostly be faster by bus. East Link will just be another option, and one that has a better frequency and span of service (especially weekends/off peak).
Yes, this. East Link will work for some trips from the East Side but not all of them. The key is for the two transit agencies to create a system that takes advantage of Link while also providing trips that are significantly faster. This means running buses like Redmond to the UW instead of Redmond to downtown.
But new Route 270 will have longer headway and waits in the weekday peak periods; ST Route 556 is being deleted. So, in combination, there will be significantly less service between the U District and BTC in September. IMO, that should not be.
I took a round trip on Route 271 on Wednesday at midday; the loads were good.
per RossB,
In September 2026, ST is planning retaining routes 542 and 545 as is. That is wasteful. Instead, Route 545 should be consolidated into a very frequent Route 542 between Bear Creek and the U District. They are not even revising routes 542 and 545 to serve the bus stops on Cleveland Street next to the downtown Redmond Link station. they will leave Route 545 in the congestion of the I-5 general purpose lanes and on 4th and 5th avenues in downtown Seattle. Note that September 2026 is long after the FIFA visitors will have left.
Jack,
There will be one less trip on the 270 from Bellevue to U District during the 8 AM hour, and the 270 will also take a better route through northeast Bellevue instead of Medina. Yes, the 556 loss means that there will be significantly less service, but this is only in peak hours. I think this mostly fine, though I do wish some peak hour 556 trips would be retained to U District. Some riders (from south of Bellevue TC) will have a faster trip via Eastgate buses to Mercer Island Link, so overall I’d take the “wait and see how it shakes out, and add more trips if needed in the peak hours as needed”.
Try riding the 271 during rush hour. It won’t be a happy 14 minutes.
For many people traveling between Husky Stadium and Bellevue TC, those aren’t their endpoints. If their endpoint is Roosevelt, Spring District, Overlake Village, East Main, Shoreline, or a bus transfer from those stations, Link is dramatically better than using the 271 for part of the trip.
“The 271 is not going away (it is being renamed though).”
It’s also moving to north Bellevue Way instead of Medina.
Some people like me think that will be faster, as well as serving the continuous row of apartments along Bellevue Way; i.e., higher ridership than the low-density houses in Medina. Others think the Bellevue Way routing will be slower — and that would make the future 270 less competitive with Link.
“But apparently I’m supposed to ride Link for 30+ minutes instead.”
You’re not. The 271 is being renamed to 270, and should be slightly faster in Bellevue, in fact.
The problem is that, outside of weekday, daytime hours, Metro plans to run the 271 infrequently enough that all the time advantage over Link gets squandered by the time you add in the wait time. This is bad, as the bus has to run at a minimum frequency, otherwise you do just end up with an empty people and people riding Link all the way around.
In my opinion, the 270 has to run at least every half hour to be worth running at all, and that, if push comes to shove, it is better to make the last trip earlier in the evening than run hourly service after 6 PM on weekends. The counter argument is that the 270 is not just about Bellevue to U district travel, it also provides coverage service along Bellevue Way, which would otherwise have no service at all. And, the coverage role favors service with more span, less frequency.
The problem is that ST is deferring the bus restructure for the 2 Line opening to the fall. Once that happens, the 556 becomes an all-week route with 15-minute headways. Both the 271 and 556 take ~20 minutes to get from UW Station to BTC, not sure about the 270 since it’ll be on a new pathway but it will be running much less often than the 556. Things will get better soon at least.
The 566’s UW-BTC segment will be deleted, so the 270 will be the only route between Bellevue TC and UW.
Ah, I missed that the UW segment was being dropped while the suburban portion is retained. That is a huge transit hole that ST should address, since they have way more resources than Metro.
It is addressing it with Link.
It remains to be seen how many people take Link vs the 270 between Bellevue TC and UW/U-District. It may be half and half, with different people taking one or the other because they prefer different features.
Isn’t the proposed 270 running much more often than the 556? The 556 is single direction peak only
I suspect most UW-BTC riders will prefer the 270 unless they are already on Link since the bus will be much faster. It will save around 10 minutes, plus the time to get in and out of the station (20-25 minutes by bus, 30-35 via Link)
The 270 proposal is here:
https://cdn.kingcounty.gov/-/media/king-county/depts/metro/documents/projects/east-link-connections/routes/270.pdf?rev=0bdcf9daf51849ddb210ee5e782fd12c&hash=39859D7C2B27981807E4F044C784C5FE
The 271 has really poor weekend headways anyways so it’s possible the 270 will be no worse, but it’s really disappointing to see “30-60 minutes” all weekend, except evenings when it’s guaranteed 60 minutes.
It really gets to a larger question of why ST sees fit to serve some urban centers but not others. Why are U-District – Redmond and Bellevue – Issaquah blessed with STEX service, but not Bellevue – U-District?
Having the cross-lake 2 Line open will shift many Eastside transit trips. Metro and ST have developed the new route structures as best as they can, but surely there will be issues with some buses over utilized or underutilized. It seems normal to me to expect some frequency tweaking to happen once the new major demand flows can be observed rather than guessed/ estimated. If nothing else, traffic congestion may affect schedules — and if it’s less than expected it may be possible to just reduce excessive layover times to improve frequencies.
Did Metro and ST tweak frequencies and schedules with some of the other recent restructures?
Once the new routes are running for a few weeks sounds like a good time to start submitting requests for adjustments. It probably will have more impact once the new routes are running.
In September, ST proposes to truncate Route 556 at BTC; it will not extend to/from the U District.
“Isn’t the proposed 270 running much more often than the 556?”
People are getting confused over the 556 restructure. Read all of Skylar’s link. This is part of the ST Express September 2026 service change. The board will vote on that in the next month or two, but it’s expected to pass without further changes.
The 554 and 556 will be replaced with a new 556 from Issaquah Highlands P&R to City Hall, local stops on Gilman Blvd and Sunset Way (replacing the 208), Issaquah TC, South Bellevue station, the south Bellevue Way stops (replacing the 550), to Bellevue Transit Center. The rest of the route to the U-District will be deleted. It will run every 15 minutes daytime every day, and 15-30 minutes until 1am.
This and the Metro’s East Link Connections restructure were planned jointly. ST withdrew from the process in the middle but is honoring the Issaquah service it indicated. The article describes Metro’s Issaquah routes. This is already approved and is being implemented in phases. Route 203 was in the last service change last August. The rest of it will be in next September’s service change. (At the time the Issaquah-Bellevue route was number 554, but it has since been renumbered to 556.) To wit:
Route 203 (already running): South Bellevue station, Factoria, Newport Way, Issaquah TC.
Routes 215, 218, 269: Combined 15-minute express service from Mercer Island station to Issaquah Highlands P&R (no stops in central Issaquah). From there they’ll diverge:
Route 215: Every third trip will continue to North Bend (replacing the 208), thus 90-minute service there.
Route 218: Extra peak-hour service going no further.
Route 269: Will continue north to Sammamish, Marymoor Village station, and Bear Creek P&R.
Regarding the U-District, Medina, and north Bellevue Way:
Route 270 from Bellevue TC to north Bellevue Way and the U-District. 15 minutes weekdays, 15-30 minutes weekday evenings, 30-60 minutes weekends, 60 minutes weekend evenings.
Route 249 will take over Medina service. The full route is from South Bellevue station to Bellevue TC, NE 8th/12th, 84th Ave NE, 24th, 92nd, Northup Way, South Kirkland P&R, to Spring District station. Every 30 minutes weekdays, 45 minutes weekends, 60 minutes every evening.
The only ways from Bellevue TC to UW/U-District will be the 2 Line or route 270.
Route
Did Metro and ST tweak frequencies and schedules with some of the other recent restructures?
I think the schedule changes are applied more system wide. There are a number of different factors they consider (crowding, performance, equity). The methodology and route measurements are covered in the service evaluations. But I don’t think it matters whether a route has been around a long time or is brand new.
It reminds me of software development. It has been decades since I was a software-process nerd and I only have a vague ideas of the particulars. But some of the basics haven’t changed since they are intuitive (and they apply to any organization). You’ve got your process to make the software and then when it is all done, you can analyze the process as well as the results. What did you do right and wrong — how can you do it better next time. It is fairly common but people often just run through this step (as if it is another bureaucratic meeting they just have to endure). People generally just want to move on and tackle the next project. If they do a better job it is probably just the individuals becoming more skilled, not the organization as a whole.
As for the transit agencies, it is well known that ST is a dysfunctional organization — I wouldn’t expect them to do a great job in this department given all their other issues. Metro used to be very well run but I don’t think this is the case any more. The quality of the project depends highly on who gets placed in charge. Some managers are brilliant. Some are not. I don’t think they spend a lot of time analyzing their previous changes to see if they worked out. This means that they aren’t improving their process. You can expect future changes to be just as bad (or to be just as random, depending on who is in charge). Likewise, riders have to live with the poor routing just as they have lived with poor routing that goes back decades.
A lot of the changes over the last few years have been reactionary. This makes sense. There have been a lot of big changes. Each Link expansion comes with a set of new changes. RapidRide changes can also alter things (RapidRide G is an extreme example of this). While necessary, I don’t think this is the best way to do things. I think they should take an iterative approach. Maybe assign teams based on area. Of course there is overlap but groups can meet with other groups. So assign someone to the north end of Seattle for example. Everything north of the ship canal. We should alter the routes system wide every year. Then go back and look at what worked and what didn’t. I don’t think we are doing that. I think the Link expansions are masking the weaknesses with the routing and especially the routing changes. As much as I think we need a system-wide restructure we also need to have a process in place that allows for a cycle of positive changes.
Let’s not forget the escalator debacle of UW station that happened shortly after opening.
I had hoped that it would inspire ST to not cut escalators and instead provide more escalator redundancy. However, even with the giant PR hit that ST took over the escalator failures, ST continues to chop escalators as their first cost-cutting move to this day.
How’s that working for you Husky Football fans?
“Trips between the University of Washington (UW) station and Westlake station only took 8 minutes, compared to 20-30 minutes on a bus. ”
I realize this post is about going from UW stadium to Westlake, but mention of that does make me miss going from the Ave to Westlake/CPS in the mornings via the 7X express buses. Often it would only take about 10 minutes. Yes, traffic etc could lengthen that. But getting a view of everything in the distance as the bus zoomed across the I-5 bridge and then having the bus “sneak” into CPS a few minutes later always felt pretty great.
Yeah, I agree. People forget how fast the express buses from the U-District were. A lot of people suggested we make minor changes to them when U-Link was implemented (and make the big changes when the train got to the U-District). But Metro got really aggressive. As a result, it was slower for a lot of people heading downtown but a lot easier to get to the U-District (since the buses were running more frequently). It got a lot better when the train got to Northgate.
Not that the train was faster. The old bus would often beat it (when the express lanes were in its favor). I remember telling my son that they were going to finally run a train to the UW. He said something like “Why will they do that? That is the only trip that is actually fast in this city!”
I think it is a very good example of how the purpose of a subway is often misunderstood. By building a train to the U-District you save a lot of operational costs. But more importantly, it is about the trips along the way. An express train from the U-District to downtown wouldn’t add much value. In fact it would be worse than the old buses (that served some stops to the north and east of the U-District before sharing those trips). It is about the trips along the way. The 41 was faster from Northgate to downtown. But Link is much, much faster from Northgate to the UW, or U-District to Capitol Hill (or a bunch of other combinations). It isn’t about replacing express buses to downtown with something that can avoid traffic. It is the shorter, non-express trips that make a metro a worthy investment.
This.
This is why I think was Seattle link is still worth it (aside from the price tag,) even though it’ll make many trips slower the amount of bus service they be able to run in West Seattle once it’s done will be a huge upgrade from what it’s like currently
That’s only if you’re going in the same direction as the express lanes. Otherwise it was 30-45 minutes even with an express on Eastlake or the regular I-5 lanes.
Would love to see a Fremont/Stone Way light rail stop either going south or in a UW Ballard line.
I think the most realistic option would be serve Fremont as part of an east-west line (from Ballard to the UW). It is a bit tricky. On the one hand you want to connect to Aurora (for the Aurora buses). On the other hand, Fremont is well below it. But I think you could pull it off. Basically put the station right next to the troll. Depending on the entrances and depth of the station it would be a bit of an uphill walk, but it wouldn’t be that bad. Likewise, if they add a stop at Aurora there for the RapidRide E (which they should have done years ago) then it would be a fairly short transfer.
There would also be a stop in Wallingford. So if you were in between stations (at say, Stone & 40th) you could take the 62 to a station.
U link is undeniably pretty cool but it’s not very good for getting to the hospital. The 43 was a mess but it was super convenient to UWMC. Good downtown access to/from rhat facility is long gone…
It would have been so much better if they simply put the station in the triangle. It really is crazy that they didn’t. There is already a pedestrian connection to the hospital. I realize there were bound to be issues but this was a major failure that just shouldn’t have happened.
UW wouldn’t even allow a connection from Link to the Medical Center via the triangle garage. UW played every card it had, including the ruse about the physics building, to fight ST every step of the way.
Yeah, and ST rolled over like a scared puppy-dog. At some point you push back. Get the regents involved. Hell, get the governor involved. I get the physics thing but the rest of it is just BS. Just because they run the UW like a multinational corporation doesn’t mean it isn’t a public university. They have a mandate to support the public good and that includes transit. If you can’t do it behind the scenes then take it to the streets. Get the transit riders union involved — hold marches. But instead they just rolled over and took the path of least resistance (which is much worse for riders).
IIRC the UW wanted Link to cross the ship canal somewhere around the University Bridge and just keep going north. Move along, nothing to see here. So the two stations around the campus are a win. I’d say the big win for transit was to getting WSDOT to restore the flyer stop on the new Montlake Lid. The original plan was for buses to just bypass Montlake. The reconfigured lid design is a win win.
I never bought into the physics build story. Look how much construction the UW has done on campus. All of the heavy equipment and earth moving has to create more seismic noise than the short duration of very predicable vibration from a TBM.
IIRC the UW wanted Link to cross the ship canal somewhere around the University Bridge and just keep going north.
Yeah, yeah, it could have been worse. A lot worse. At one point they wanted to follow the freeway from downtown to the U-District. Less expensive to be sure, but so much worse.
For what it’s worth, the 1985 Forward Thrust proposal had only two stations in the U-District: https://www.theurbanist.org/content/images/www-theurbanist-org/wp-content/uploads/2018/05/17013207411_cce9745488_o.jpg. It skips the heart of campus as well. I would probably slide the southern station to the triangle. While I think the Forward Thrust proposal is light in the U-District, it is so much better in other ways (three stations in the Central Area — not one) that it is a huge improvement over what we built.
Yes indeed Ross. That would have been the right solution but I’ve grown a little weary of thinking st will make the right choice.. they’ve beaten me down.
I recall (but I might be wrong) the first RM-Transit video I watched was about U Link extension opening. The bike&ped bridge was impressive to me.
Hello,
I run the King County Metro Transit radio scanner (as an offshoot of rosecitytransit.org) and am planning on taking local transit up from Portland for the Link opening. I am wondering if anyone would be willing to host me next Friday and Saturday night. I could get a hotel but would love to stay with someone.
Also, is there plans for a transit meetup? YouTuber/blogger Miles in Transit has said he’s coming for the event too.