Fully agree on all of this, and probably have more to say later (because of course I do). Everything works way better with a subway to Nubian, and you've articulated quite well the impact that would be felt well beyond Washington Street.The more I crayon things out, the more convinced I am that GL to Nubian needs to be built as a subway.
My experience as a planner in San Francisco is that busy commercial corridors do not work well for speedy transit, even with dedicated lanes and TSP. Speedy transit is often at cross purposes with other desirable aspects of the corridor, particularly pedestrian activity and safety:
Realistically, the 23-minute scheduled time between Nubian and Temple Place is not likely to substantially decrease, even with relatively aggressive transit priority measures on the surface. You'll save a little time using the subway north of the Pike, but trains are less nimble than buses and tend to be slower when dealing with intersections. For many of the same reasons, you're limited as to the effective frequency you can run. One branch at 8-minute headways is probably fine; two would probably end up with a lot of bunching.
- Closely spaced intersections are great for pedestrian convenience and slowing traffic, but they make effective TSP near-impossible. Herald to Nubian has 15-20 likely crossings, depending on how you do the accounting, in 1.3 miles. Compare that to 17 / 2.3 miles on the C and ~25 / 4 miles on the B, both of which are torturously slow even with a dedicated median.
- Similarly, you want to be slowing vehicle speeds for ped/bike safety and comfort.
- Because of high pedestrian volumes, you want your walk phases to be as long as possible - preferably 3.0 feet per second or even 2.8 fps, rather than the old standard of 3.5 fps. That means longer wait times for transit vehicles.
- Transit takes up space, especially if you're doing center-running with boarding islands. That eats into the space available for protected bike lanes, wider sidewalks, parklets, and necessary commercial loading.
- For transition to/from a subway, a portal takes up a lot of space. It's about 500 feet where nothing can cross the transit ROW, and the walls mean it's somewhat wider than just the tracks.
Thinking about service patterns, it's essentially a balancing game: Huntington + Nubian = Tremont + Seaport. (That's true whether or not you have direct Nubian-Seaport service.) Huntington is going to be full up - you're likely to have services from Hyde Square, Needham, Riverside, and Harvard or West Station. Obviously you need to use every available slot for Tremont as well. So, if you want more than 8-minute headways to the Seaport, you need to have more than one service going to Nubian.
A Nubian subway lets you run a balanced pattern: Huntington trains are evenly split between Tremont and Seaport, and Tremont trains evenly split between Huntington and Nubian. (Assuming you don't run Seaport-Nubian, which isn't a bad idea but would require a flying junction under the Pike.) 4-minute headways in a subway lets Nubian actually be the Roxbury bus hub in the way it was until 1987. Trips to downtown, Kendall, Seaport, etc change at Nubians, rather than every single bus route having to be extended to Ruggles. (Obviously some routes would run to Ruggles/LMA, but some routes could terminate at Nubian while still having a rapid transit connection.
If you do commit to build that subway, it's not that ridiculously difficult. Less than 2 miles of tunneling from the existing subway, and cut-and-cover is probably doable except the brief section under the Pike. You can have anywhere from 2 to 4 intermediate stations between Bay Village and Nubian. If you're wiling to do some more difficult tunneling (perhaps in conjunction with a redevelopment of the Cathedral housing), you can run part of the subway on Harrison to directly serve BMC/BUMC.
View attachment 32309
The thing that has always terrified me is the complexity of tunneling. On the one hand, Washington runs down the spine of the original Shawmut Peninsula, so it's not landfill, but on the other hand, that also means there are that many more years for things to have been buried under there.
That being said, I don't think I've considered what the cost-benefit ratio looks like when you take into account the systemic improvements this subway would offer -- I'd have to think about that.
Two additional ideas for you to consider crayoning:
- Why not tunnel down Harrison exclusively? Then you can do fully cut-and-cover without needing to cut across one of the blocks at a diagonal. In your crayon above, it would only entail relocating your E Dedham station 600 feet west, so very little impact to service area. (I myself would potentially include an additional stop at E Berkeley, but again, the difference would only be 500 feet.) I agree that BUMC is a worthwhile target.
- While it wouldn't need to be LRT, you'd still want some sort of surface transit on Washington or one of the parallel streets; as far as I know, the El always had a local route running underneath. To your point, I think you need speedy journeys on sub-5-min headways from Nubian to downtown, so I wouldn't suggest having 1 surface LRT branch and 1 subway branch, but you could maintain a BRT service with local stop spacing; if relieved of the need to connect Nubian with downtown, a bus service could probably function pretty well there with bus lanes etc.