Metro is getting serious about restructuring their system for Northgate Link, so I figured I would as well. Here is a map of my proposed changes (click on the map for a full screen view):
I’ve tried to design a system that enables fast, frequent service to the Link stations and the UW. A balance is made between coverage versus speed and frequency. I’ve focused my efforts on the clusters of apartments that exist in the area, while still retaining a reasonable walking distance for everyone. As part of this restructure, I’ve tried to remove turns, which slow down buses. Effort has been made to consolidate routes and provide more of a grid, but given the geography of the area, it remains a challenge. Overlapping bus routes tend to occur where there is larger demand for transit.
Unlike previous proposals, all of the routes use existing layover space. This map also includes peak-only service.
I eliminate the 41, 64, 73, 76, 78, 309, 316 and 373. The following are new or modified routes:
Peak Only Changes:
63 — This provides some of the coverage currently provided by the 62 and 63.
77 — The 77 is truncated at Green Lake Park and Ride. It helps replace coverage on 15th Avenue NE in Maple Leaf (caused by the loss of the 73/373) while providing a faster, shorter express companion for the 67 and 348. It is also possible that it would be truncated at 145th (if ridership is not very high north of there). I could even see two versions (a 77 and 377). While peak-only, I would expect this to be bidirectional (like the current 373).
277 and Peak ST Buses — These are truncated, and sent to Memorial Parkway.
301/303/304 — I don’t get into the details here, but simply propose that they be truncated at Northgate. I’m ambivalent about the area — I’m not sure how much truncation should occur, given the effect this could have on the E.
308 — This is truncated at Northgate, which saves service hours while also providing a one seat connection from Northgate to 145th (between 15th NE and Lake City Way).
312 — Truncated at the Green Lake Park and Ride.
355 — Modified to no longer serve the U-District. This would speed it up, and allow it to be an express overlay for the southern part of the E. It would avoid the Linden detour of the E, and make no stops between 46th and 76th.
All Day Changes:
26 — The northern part of the 26 now goes to the UW, but is otherwise the same. If possible, it would through-route with the 74.
31/32 — Now sent through the U-District on The Ave up to the Green Lake Park and Ride. This provides Fremont users with a faster connection to Link (especially from the north), while doubling up service between 45th and 65th (one of the more densely populated parts of the city). It also means that a rider from Fremont would have two buses connecting them to Roosevelt: the 62 and the 31/32.
46 — This is a new route that has two parts: a direct connection between Greenwood and Northgate, along with an interesting tail. The tail provides east-west coverage on 65th, along with additional coverage for 32nd and 8th Avenue NW. If this western tail proves too expensive, then the route would be shortened in Greenwood, and lay over at 3rd Ave. NW, where the 5 Express lays over.
61 — This is the fastest connection between Lake City and Northgate. Some of the same bus stops in Lake City would serve the 312/522. Its short length and relatively fast speed gives Metro a lot of flexibility during rush hour. Increasing frequency on this route would be the cheapest way to increase capacity between Lake City and Link.
62 — Modified slightly to be faster east of Green Lake.
67 — The 67 no longer loops to Northgate. This avoids the costly button-hook of the existing 67, while balancing service better in the area. I go into more detail at the end of this post (since the issue is complicated).
71 — This replaces both the old 71 and 76 route. It is a major upgrade in Wedgwood, as it serves the apartments before the low density, single family neighborhood. This means that more riders get to their destination faster. For riders along 35th, it provides a much faster connection to Link than the 372. It is a short route, so it would make a live loop through View Ridge.
74 — The 74 goes to Children’s Hospital instead of Sand Point. To provide for some of the lost coverage on the 78, it loops around Laurelhurst before laying over. Riders on the rest of the line wouldn’t notice, while those in the nearby neighborhoods get home sooner. If there is low ridership on this small loop, it could always be cut.
75 — This has been modified to cover the old part of the 41 (thus eliminating a turn on both this bus and the new 61). It has also been extended to serve the U-District. This provides a faster connection to the U-District for people in Sand Point, Children’s Hospital and the U-Village (all of whom still have the option of catching a bus to the campus).
361 — Same route as the old 330, but since it always through routes with a 61, it is renamed. That way, someone will take a 61 or 361 in the same manner that folks take a 73 or 373.
345 — Modified to use 115th and Aurora. This saves a considerable amount of time, while still providing front door service to Northwest Hospital. It also provides additional service from Aurora to Northgate.
372 — Modified slightly to use the same loop through the UW as the 65, thus consolidating service between the UW and the U-Village.
522 — Truncated at Green Lake Park and Ride. There is more about the 522 at the end of this post.
I don’t have the tools or patience to come up with a detailed frequency plan. I will mention a few thoughts that I have though.
- Adding service gives you diminishing returns. For example, going from 30 minute frequency to 20 minutes is just as expensive as going from 20 to 15.
- As with our current network, there are many instances where a rider can take more than one bus to a particular destination. Someone in Lake City, for example, can take the 65 or 75 to the UW (from the same bus stop). In general, these areas with multiple routes are high ridership locations, while areas with only one bus route are not. Thus there is value in spreading the improved service. Right now, the 75 runs every 15 minutes, and the 65 has 10 minute headways. It wouldn’t cost any more to run them both every 12 minutes. Rides from Lake City would the same, if not better (if they could time the two routes). Riders along Sand Point Way would see a bigger gain than the loss felt by those along 35th.
- At the same time, there are some corridors with very high ridership throughout, and only one bus. The 44 comes to mind.
Then there is the issue of coverage. The whole point of a coverage route is to reduce the walking distance to a bus. But if there is a nearby bus that is a lot more frequent, people will walk to it. Again, this is where diminishing returns matter. If a frequent bus goes from 12 minutes to 10, it isn’t likely to attract more riders from the coverage bus. On the other hand, a coverage bus that goes from 30 minutes to 20 will likely see a considerable increase in ridership. That is why the coverage routes that are relatively close to other routes are peak-only. The 63 and 77 save about five minutes of walking. Very few people will use them if they run every half hour. Since it would be tough to justify more frequent service in the middle of the day, they are rush-hour only. The 71 and 74 are different. The walking distances are larger, and they go to different locations than the more frequent buses. They could probably do OK with half hour service, although I would try to do better. With all of that in mind, here are some hopes for the system:
26 and 74 — The buses are tied together, which complicates things. Both routes are shorter, but not a lot shorter. I think we could get 20 minute frequency, but it would be hard to justify 15.
31/32 — Change from 30 minute to 20 minute frequency, with 10 minute combined frequency from Fremont to the U-District.
44 — As mentioned above, the 44 is the only bus serving a very popular corridor that is likely to get a lot more popular. This is one case where added service is justified, even with big diminishing returns. Running the bus every 7 and a half minutes, or even every 6 minutes would change the nature of the route, and likely lead to a big increase in ridership. Of course, at that frequency, speed improvements are even more important, both to save money and reduce bus bunching.
45 — Like the 44, I think we will see a significant increase in ridership, justifying 12 minute, if not 10 minute frequency. However, the 46 would provide an additional way to get to Link for many of those riders, so it isn’t clear whether we can afford both.
46 — As mentioned above, there are two parts to this bus: the section from Greenwood to Northgate, and the tail (to Sunset Hill). The frequency of the Greenwood to Northgate section should match the frequency of the 45 if possible. The tail is complicated. From the standpoint of serving Sunset Hill, any service is welcome. It is about a ten minute walk from 32nd to 24th, which is about twice as long as the distance saved by buses like the 63 and 77. But the core of the route — crosstown service on 65th — only makes sense if it is frequent. One of the key values of a grid is that it makes for straightforward two-seat rides. But that is only popular if both buses are frequent. I think at a minimum this needs 20 minute service at the tail. That would mean half the buses truncated in Greenwood, thus providing 10 minutes service between Northgate and Greenwood. If we can’t afford service that frequent to Sunset Hill, I would just truncate it in Greenwood.
61 — The 61 provides the fastest, cheapest connection between Lake City and Northgate. You can make a good case for running it every 10 minutes, which would match the current 41. But changes to this bus depend on both the 75 and 522. The 522 runs every half hour, but it does provide another way for riders to get to Link. Likewise, if the 75 runs every 10 minutes then it might be tough to justify 10 minute service here as well.
62 — When I started this, I assumed that the 62 would see a huge increase in frequency. Now I’m not so sure. Other changes reduce the need for higher frequency here, and it is one of the longer, more expensive routes. The 31/32 provides another way to get from Fremont to Roosevelt. The new 71 would serves much of 65th. The new 75 would provide Sand Point riders with another (albeit slower) way to get to Link. I would still try and squeeze in another run, so that the bus could run every 12 minutes, but there are probably better values elsewhere.
71 — I think the new 71 would be very popular. Right now it runs every half hour, and it is considerably longer (serving the U-District instead of Roosevelt). I think the bus should run every 20 minutes, if not 15.
75 — This is a bit complicated, as much of the corridor is currently served by the 41, which runs every 10 minutes. Running the 75 every 10 minutes would be great, but I’m not sure we can afford it. If we can’t, then I would prefer better balance, with both the 65 and 75 running every 12 minutes, and the 61 running at least as often (since it is the cheapest route).
77 — A lot depends on how often the 348 runs. If the 348 runs every 20 minutes (as suggested above) then this could get by with 20 minute frequency. That would mean 10 minute frequency (or better) between 125th and 175th. But 20 minute service would be barely adequate from a coverage standpoint for those on 15th, between 65th and 125th. As mentioned above, I could see two versions. One would truncate at 145th, while the other went to 175th. The longer version (the 377) would run every twenty minutes, opposite the 348. The shorter version (the 77) would run opposite the 377 (enabling 10 minute frequency). That would occur at the peak of peak, while the 377 would be the only bus during shoulder periods (e. g. 2:00 PM to 3:30 PM).
345/346/347/348 — Change from 30 minute to 20 minute frequency. These routes are not that time consuming, and that is only one additional bus per hour.
The 522 is not part of the North Link Connections Mobility Project, but it is certainly in scope. Changes to the 522 effect Metro’s network. I think it is essential that the 522 serve Roosevelt, as the stop at 20th and 85th is more popular than any stop north of Lake City.
The dynamic between the 522 and 312 is less than ideal. On Lake City Way, the 522 acts as an express; it makes only one stop south of 125th. The 312 makes several. In general, it makes sense to run both the express and regular bus during rush hour, and then run the regular bus the rest of the day. Serving all the stops in the middle of the day increases ridership, thus enabling better frequency when it is needed the most. The 5 and 5 Express is a good example of this dynamic.
But in this case, it is the 522 that runs in the middle of the day, while the 312 does not. This would be easy to fix if it wasn’t for the fact that two different agencies (with two different budgets) operate the routes. Making matters even more complicated is that the 522 serves Woodinville, while the 312 does not.
One solution would be to create a different version of the 522 (e. g. the 523). This would only operate during off-peak hours. Ideally it would have all the stops that the 312 has, but at a minimum, it would serve 80th and 15th. This stop would help plug the biggest coverage hole caused by the loss of the 73.
If ST doesn’t do this, then Metro could pick up the slack. They could simply run the 312 all day, opposite the 522. This would provide 15 minute frequency for the bulk of users, while still serving the 312 stops all day (with half hour service). To save money, it could be truncated in Kenmore, or even Lake City. Of the five communities outside downtown (Lake City, Lake Forest Park, Kenmore, Bothell and Woodinville), Lake City has by far the most riders, followed by Kenmore. With a significant number of Lake City riders north of the layover spot (by Fred Meyer, on 133rd) and Kenmore itself with decent ridership, truncating at Kenmore is probably the best choice.
Straightening the 67
As with every change, there is a trade-off. There are winners and losers with a 67 that goes to 145th. Here is what we lose with this change:
- Fewer buses to Northgate. Maybe, but there will still be plenty. From Roosevelt Way and Northgate Way, you currently have the 75 (4 buses an hour), plus the 347/348 (2) and the 67 (6) for a total of 12 buses an hour on that corridor. But with the proposed changes, you would have the 347/348 (4 buses an hour) and the new 61 (6 buses an hour). If 10 per hour is not enough, then the cheapest fix would be to increase the frequency of the 61, which would mean regular service more frequent than what exists now. At 5th Avenue, you add another bus, which will run at least 4 times an hour. That is somewhere between 14 and 18 buses an hour on 5th Avenue NE, headed to Northgate. That is plenty.
- Two seat ride from Northgate to Maple Leaf or Northgate to Roosevelt. Fortunately, riders will have alternatives. As mentioned, a two seat bus ride would be very frequent. Many would just walk to their destination, or walk five minutes to catch the bus. For longer destinations, it makes more sense to use Link for parts or all of the trip, not only because Link is faster, but also because the combination is likely to be extremely frequent. There will be lots of buses on 5th, connecting to the Northgate Station, and lots of buses connecting to the Roosevelt or U-District Station.
- A two seat connection to the Northgate Station. This doesn’t matter, since those riders will just go to the Roosevelt Station. Since most riders will be heading south on Link (well, all riders will be heading south initially) this is usually faster overall.
In contrast, here is what is gained:
- The bus avoids the time consuming twists and turns to get to Northgate. Even if it gets just as many riders per hour, those riders get to their destination faster.
- Dramatic increase in frequency to Link along an important corridor. If the 347/348 runs every ten minutes (combined) that would be a substantial improvement, with 6 buses an hour headed to a Link Station. But that is nothing compared to the buses on the other corridor (which would be up to 18 per hour without the 67). Adding another 6 buses to the Roosevelt/Pinehurst/15th corridor would be a major improvement.
- One seat ride to Roosevelt and the U-District, from Pinehurst up to the city border. Trips like this would be much easier. This is a gain that matches the loss mentioned in the second item above. But the gain is bigger than the loss. As mentioned, riders along the old corridor can take an extremely frequent set of buses (or Link) to make it a two seat ride. Others will simply walk an extra five minutes. In the case of Pinehurst (and places north) the alternatives are much worse. Walking is unrealistic, and the transfer isn’t nearly as common.
- It makes it harder to get rid of the 73. Metro tried to get rid of the 73 before, when they added the looping 67. The objection wasn’t so much the extra five minute walk (from Roosevelt to 15th) it was the ugly transfer, in the middle of a car sewer. This is not a destination. This is not like making a same direction transfer in the U-District, or even Northgate. This is a terrible place to make a transfer, when you just want to keep going the same direction, on the same road, the way that a car or bike would.
Straightening out the 67 hurts some riders a little, while improving transit dramatically for a lot more.