Green Line Reconfiguration

[Riverside]

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:

• 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.

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.

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

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 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:

1. 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.
2. 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.

 

[The EGE]

I'm agnostic about Harrison versus Washington for the northern half - much like the number of intermediate stops, it will probably come down to technical feasibility, costs, and what the community wants. (There's even an argument to be made for Shawmut to have a straight shot across the Pike; the 1940s plans for the El teardown called for a Shawmut subway.)

Certainly you need to keep quality surface transit - basically going back to how the 49 bus ran under the El. The route from the bus network redesign (a single Nubian-South Station route) would be fine, with extension on Summer or Congress busway if desired. That would work fine in the more pedestrianized environment, since it would mostly be handling local passengers and not be a load-bearing trunkline.

 

[Riverside]

the 1940s plans for the El teardown called for a Shawmut subway.)

A very specific question here, do you know where those plans are off-hand? I've heard of them but never seen them (or only saw them in generic detail).

 

[Charlie_mta]

A very specific question here, do you know where those plans are off-hand? I've heard of them but never seen them (or only saw them in generic detail).

All I have is an MTA System Map showing the proposed tunnel route with stations:

 

[The EGE]

I don't think it ever got particularly far in planning. This is the 1948 state law that directed the MTA to build the subway:

Acts and resolves passed by the General Court : Massachusetts : Free Download, Borrow, and Streaming : Internet Archive

Imprint varies

archive.org

 

[Teban54]

Thanks to you both, @The EGE and @Riverside ! Wow, I think a Nubian subway via either Washington or Harrison - if materialized - is a wonderful idea, and may have cascading effects on the entire Green Line Reconfiguration proposals.

Here are my comments:

Political concerns: How to get it done?

To me, this is the most concerning aspect of the whole proposal. Regardless of construction method, one simple question from stakeholders may be enough to kill the project altogether: Why bother doing the much more expensive tunneling, when BRT exists and can be converted to surface LRT at much more reasonable costs?

While I do conjecture that the overall benefits to the entire network will be worth the cost, we need to sell this in the right way. Factors that help build momentum in favor of a Nubian subway:

• Everything @The EGE said.
• Much better travel times to Nubian (compared to BRT and surface LRT), tied with the "equal or better" promise. A subway is practically the only way to achieve "equal or better" compared to the El, as @Riverside discussed here.
• Directly serving BMC and BUMC, if possible. Getting them onboard will be extremely helpful.
• Impacts on buses, which help communities south of Nubian all the way to Ashmont & Mattapan.
• Cut and cover to reduce cost.
• Pitching this as additional service to the SL4/5, not replacement. See section immediately below.

Washington vs Harrison

First of all, I agree that SL4/5 need to be kept, albeit with reduced frequency and capacity. Regardless of the route of the subway, having surface buses on Washington with dedicated bus lanes (basically what SL4/5 currently are) will do better in serving local stops in South End, much like the 42 bus today. I think this will be needed to ensure the project is politically feasible.

With that assumption, I think having the subway go down Harrison - at least near BMC/BUMC - will be the better choice. Think of it as an additional subway LRT branch that aims to serve Nubian and BUMC with rapid transit, rather than "just" replacing the SL4/5. The proximity to BUMC will be a huge boon in cost-benefit analysis, assuming feasibility and passable cost.

Bay Village station design

Most Bay Village designs recently, especially those in 2022, assume Nubian will be a surface LRT branch. Let's first do a quick recap of proposed designs:

• The EGE's designs (Nov 2022)
• Riverside's design 1 and design 2 (Mar 2022), and an updated design in Nov 2022
• And a bunch of old proposals I quoted here.

Some of these designs, such as those that place the Nubian branch station on surface level (on Marginal or over the Pike), no longer work with a subway to Nubian. Add in the choice of which road to dig under just south of the Pike - Washington, Harrison or even Shawmut - and that opens a whole bunch of questions:

• Does something need to go under the Orange Line ("Basement 3" level)? Most likely yes, since the Nubian subway needs to be at least Basement 2 or below to go under the Pike, and unless you only start descending E of Marginal/Shawmut, it needs to go under Orange Line as well.
• If going under Shawmut, can we just have a simple "+"-shaped station with Park-Nubian trunk on Shawmut and Back Bay-Seaport trunk on Marginal? Probably the simplest design possible (with a Back Bay-Park connection somewhere), but the Nubian branch platform needs to be at Basement 3.
• If going under Harrison immediately after the Pike, can we actually do the "start descending E of Marginal/Shawmut"? Between Shawmut and Harrison, descend to Basement 2 under the Pike, then have a flying junction. Problem is, for complete separation of the two trunks, this section needs to be quad-tracked, and Marginal isn't wide enough for that.
• Is a Park-Seaport connection needed? That will complicate the design even more.

I don't have answers to these questions yet, but hopefully they're inspiring enough.

Subway extension past Nubian?

Every now and then, someone suggests LRT from Nubian south following the 28's route, possibly all the way to Mattapan. While Blue Hill Ave does need better transit, there are several problems:

1. Too much street running, especially with Nubian-Bay Village as surface LRT
2. Warren St south of MLK Jr Blvd is too narrow

A common suggestion to resolve (2) is to build a short tunnel across that section. But now, we already have a subway to Nubian... So why not just extend it further to achieve that? (1) is also mitigated due to much better reliability north of Nubian.

Not necessarily suggesting subway to Mattapan, and even surface LRT to Mattapan seems overkill. A somewhat reasonable proposal (albeit still Crazy Transit Pitch) would be subway to Franklin Park. Dig under Warrent St to Grove Hall, then either rise to street level or continue as subway. If you really want to, you can even go further south on street level, although Franklin Park to Mattapan is as long as the C's street-running section.

In addition to serving the park and Roxbury/Dorchester, the route also potentially allows the surrounding bus network and ridership to be gravitated around Franklin Park, instead of (or in addition to) Nubian.

And ngl... "F train to Franklin Park" sounds fun, isn't it?

 

[Brattle Loop]

To me, this is the most concerning aspect of the whole proposal. Regardless of construction method, one simple question from stakeholders may be enough to kill the project altogether: Why bother doing the much more expensive tunneling, when BRT exists and can be converted to surface LRT at much more reasonable costs?

While I do conjecture that the overall benefits to the entire network will be worth the cost, we need to sell this in the right way. Factors that help build momentum in favor of a Nubian subway:

Absolutely a critical element. Recall that this is a corridor where the promised "equal or better" elevated replacement was a silver bus that, in addition to being worse in its mode, was shotgun-married to a completely different transit project, with the end result of being permanently broken in terms of downtown connectivity. (Phase III would have at least provided less-kludgy Green, Orange, and Red transfers, albeit still subpar.) The policymakers have been depressingly consistent in their willingness to do the least they possibly can get away with for this corridor, and considering the scale even having the city administration onboard (while necessary) would not be sufficient to avoid the need to get the approval of politicians with zero stake in the matter. That's difficult enough on any project (and requires spending political capital that then cannot be used on additional competing priorities, and no matter who, be it Wu, Healey, any of them, they all have fifteen top priorities at once). It's complicated in this instance by the fact that, while subpar to subway, surface LRT fixes one of the most glaring flaws of the Silver Line, its shitty connections to the rest of the system. A Green Line branch going only as far as North Station gets direct, inside-fare-control connections to all of the HRT lines, direct-via-tunnel connection to the northern Commuter Rail hub, cross-platform transfers to every other GL branch, and spreads out the transfer loads downtown, all while doing it at a fraction of the cost of tunneling from Bay Village to Nubian. Making the case that the perma-fucked Silver Line is not an adequate elevated replacement isn't that big an ask, politically; making the case that the replacement should be closer to the "spare no expense" than the "bare bones" end of the spectrum, when the city and state have a dozen priorities competing for transportation dollars and the General Court has fifty other things they want to do is way, way harder.

In that context, I'm not sure the selling points you've identified will be enough. The "better travel times" argument, to the extent that it's an "equal or better" (i.e., equity) argument, is not terribly strong. Nubian's continuing need for something is indicative of the fact that this corridor has not (at least to this point) had anything like enough political clout to get its equity arguments turned into on-the-ground realities. That has to change for those arguments to not be particularly vulnerable to cost criticisms (because excessive costs are themselves useable as a cudgel by other "deserving" claimants who want the money for their 'deserving' projects.) Direct service to the medical campuses is probably another "necessary but not sufficient" element. It's unquestionably a network add, it's also not that far of a walk from where surface LRVs would stop on Washington, since Harrison's not wide enough there to run surface LRT you'd be talking the cost of the whole-length tunnel to bring service right to the BMC/BUMC doorstep. (Could be somewhat mitigated if they were willing to pitch in financially, but not by much.)

"Cut and cover", if it's even a selling point (because its cheaper than TBMs), comes with a huge caveat in terms of disruptiveness. Harrison in particular has very little spare width to play with (it never gets to Essex Street Silver Line levels of nightmarish, but either way (Washington or Harrison) you're still going to be talking about a lot of businesses (and homes) not likely to be all-that-happy about the streets in front of them getting torn up (and potentially made temporarily impassible) for a while, so the NIMBY problem would be way, way worse than if your project was more "we're laying some rails and stringing some wires". (On that note, I thought I recalled F-Line saying that there was still a power interconnection from the elevated days under Washington; would a Harrison subway be able to feed off that, or would rebuilding the electrical system need to get added in too?)

Leaves us in the position where the main non-equity cost justification would be the network effects. I think it's inarguable that a subway would be a much, much more useful (faster, higher capacity) pipeline for running service in from Nubian, though it wouldn't be able to have as many stops as on the surface given that subway stations blow the cost out more than just tunneling feet. A lot would seem to turn on whether the impact of more capacity to feed into from the bus network (including the slack it would have to pick up from the walkshed potentially not served by wider station spacing) justifies the cost premium of the subway (because after all LRT, especially with Type 10s, would represent a capacity increase over the Silver Line anyway). That at least seems like something that can be modeled with at least some degree of empirical accuracy, which would at least give something to discuss rationally that doesn't involve mercurial politicians.

I think it's absolutely a debate worth having, just one that in the moment I find myself unconvinced that the overall benefits are worth the cost, at least, unconvinced that they are to the extent necessary to convince the politicians (and, unfortunately, that convincing is a necessary condition for building this thing). [For the sake of completeness, I will note that a thorough analysis of the costs and benefits would also consider the costs of a 'finite' amount of transportation money in terms of what else could be achieved with the money this would take. If doing this as subway means putting off BLX to Lynn for a decade, for example, is it still justified? Or should we settle for the surface option, acknowledging that it's inferior but that there's a reality in which we can't have both, say, because there's only so much money to go around. That's a quintessentially political calculus that every project has to deal with.]

Every now and then, someone suggests LRT from Nubian south following the 28's route, possibly all the way to Mattapan. While Blue Hill Ave does need better transit, there are several problems:

1. Too much street running, especially with Nubian-Bay Village as surface LRT
2. Warren St south of MLK Jr Blvd is too narrow

A common suggestion to resolve (2) is to build a short tunnel across that section. But now, we already have a subway to Nubian... So why not just extend it further to achieve that? (1) is also mitigated due to much better reliability north of Nubian.

Not necessarily suggesting subway to Mattapan, and even surface LRT to Mattapan seems overkill. A somewhat reasonable proposal (albeit still Crazy Transit Pitch) would be subway to Franklin Park. Dig under Warrent St to Grove Hall, then either rise to street level or continue as subway. If you really want to, you can even go further south on street level, although Franklin Park to Mattapan is as long as the C's street-running section.

I mean, you could, but it'd cost a boatload of money. There's not really anywhere immediately identifiable down there where you can stick a bus terminal, so while you could probably make some rationalizations with routes that cross the Warren/Blue Hill corridor not needing to distend as much to Fairmount, Orange, or Nubian, you'd have a hard time feeding that extended-subway pipe off of Blue Hill unless that was already LRT (because then you don't need a bus terminal), and, as discussed, if you put the cost and operational issues of LRT to Mattapan into your calculator, it doesn't make a happy face. Real-world, you'd run into a difficult-to-answer question about why this expense, if for a fraction of it we could get a ton more buses in bus lanes up from Four Corners and points south into the high-capacity subway pipe at Nubian. (Works out somewhat better if you can find a place to feed that pipe at scale somewhere near the park, but absent blowing up the Stop & Shop or - maybe less problematically - the furniture store on Columbia and doing an around-the-block loop with a turnaround loop, end station, and bus terminal on (well, the LRT stuff maybe under) what's now the Burger King's parking lot, I'm not seeing a place to fit it.)

This concludes the early-morning bucket-of-cold-water dump. In all seriousness, though, I appreciate your take on the corridor's needs and the situation, and I thank you for giving us an interesting proposal to chew on (even if my instinct is to immediately poke holes, I try to do so with the aim of being constructive, and I feel like sometimes I don't always make clear that even when I don't necessarily agree with a proposal, I appreciate and value the time, thought, and effort that goes into making them.)

 

[JeffDowntown]

Given the Washington Street corridor to Nubian is the original Boston Neck, you are looking at 4 centuries of poorly documented infrastructure under that route. I have to believe cut and cover would be a nightmare, and would result in huge cost overruns. It would probably be cheaper to go with a deep bore TBM, and deal with deeper stations. The fake math of the cut and cover crowd would claim that is more expensive, but they never get that right in Boston (key point, local contractors cannot execute TBM boring, so the contractors have to come from outside the local trades).

 

[Teban54]

I mean, you could, but it'd cost a boatload of money. There's not really anywhere immediately identifiable down there where you can stick a bus terminal, so while you could probably make some rationalizations with routes that cross the Warren/Blue Hill corridor not needing to distend as much to Fairmount, Orange, or Nubian, you'd have a hard time feeding that extended-subway pipe off of Blue Hill unless that was already LRT (because then you don't need a bus terminal), and, as discussed, if you put the cost and operational issues of LRT to Mattapan into your calculator, it doesn't make a happy face. Real-world, you'd run into a difficult-to-answer question about why this expense, if for a fraction of it we could get a ton more buses in bus lanes up from Four Corners and points south into the high-capacity subway pipe at Nubian. (Works out somewhat better if you can find a place to feed that pipe at scale somewhere near the park, but absent blowing up the Stop & Shop or - maybe less problematically - the furniture store on Columbia and doing an around-the-block loop with a turnaround loop, end station, and bus terminal on (well, the LRT stuff maybe under) what's now the Burger King's parking lot, I'm not seeing a place to fit it.)

This concludes the early-morning bucket-of-cold-water dump. In all seriousness, though, I appreciate your take on the corridor's needs and the situation, and I thank you for giving us an interesting proposal to chew on (even if my instinct is to immediately poke holes, I try to do so with the aim of being constructive, and I feel like sometimes I don't always make clear that even when I don't necessarily agree with a proposal, I appreciate and value the time, thought, and effort that goes into making them.)

Oh the subway-to-Franklin-Park proposal wasn't meant to be a super serious one, more of a shower thought following a subway to Nubian (if that even materializes). In reality, Warren St and Blue Hill Ave will likely continue to be served by buses and BRT, which will still benefit hugely from Nubian having true rapid transit and thus being a true anchor of bus routes further south.

And by no means was I suggesting Nubian subway needs to take priority over things like BLX to Lynn. A somewhat realistic, phased plan could be to have surface LRT via Washington first, then tunneling under either Washington or Harrison as a second phase.

 

[Brattle Loop]

And by no means was I suggesting Nubian subway needs to take priority over things like BLX to Lynn. A somewhat realistic, phased plan could be to have surface LRT via Washington first, then tunneling under either Washington or Harrison as a second phase.

Yeah, looking at my post I wasn't entirely clear that the BLX stuff was just a hypothetical example to illustrate the point that the cost side of the cost-benefit analysis would have to take into consideration what other projects might suffer if you spent your finite transportation dollars on this (or any given) project. A phased plan would be better, though would run the risk of the second phase not getting built if there weren't sufficient safeguards/triggers for getting it done (though you'd still get transit better than currently exists).

 

[themissinglink]

A somewhat realistic, phased plan could be to have surface LRT via Washington first, then tunneling under either Washington or Harrison as a second phase.

I feel like a phased tunneling plan is the most realistic chance of a grade-separated GL to Nubian becoming a reality. I'm concerned about how expensive it might be to cut & cover under Washington Street, though. I also wonder about the feasibility of using a TBM, although it would depend on the geological conditions of the area.

 

[Riverside]

I'm definitely going to sit down and give this a lot more of a think. @The EGE I am 100% in agreement with you and like I said have felt that way for a long time but despaired of the physical challenges and ensuing costs.

One question I'm spitballing on: is there any potential in the idea of two surface LRT lines, e.g. one on Washington and one on Albany? That could give you 4-min headways from Nubian while keeping the on-street segments at 8-min.

[insert here something something street widths something something increased journey time from roundabout path something something decreased journey time from reduced stops something something bonus points if you can shiv LRT underneath the 93 viaduct instead of those parking lots something something]

 

[Riverside]

It's 11pm on a Friday night, which means it's time to draw some lines on maps.

None of these ideas have been examined critically at all.

Setting aside a second surface route, let's instead think about building a subway to Nubian. A couple points for our consideration:

1) No matter how you do it, it will be expensive. Whether it's dealing with landfill or undocumented utilities or deep bores or unhappy residents, the cost will be high.

2) The number one goal is to run trains quickly and frequently from Nubian to downtown. A nice number two goal would be serving BU Medical Center.

3) Local service on Washington St is explicitly out of scope and can be handled via a shorter surface route.

Earlier in the week we were chatting about tunneling under Harrison. But what if we looked further east, and instead tunneled under Albany and 93? (You could maybe also save on tunneling costs by running in an open cut under 93.)

Believe it or not, this is only a quarter mile longer than a Washington St alignment. And I would argue that you could get away with having only two intermediate stations (e.g. at the southern and northern ends of BUMC), or maybe even only one. 2 miles of running track is equivalent to Beaconsfield to Longwood, which is timetabled at 7 min with two intermediate stops. So an alignment like this could potentially compete with the travel times of the El.

If we don't feel like pissing off the hospitals by tunneling outside their front door, we could piss off everyone else and tunnel under a combination of Melnea Cass, the Mass Ave Connector, BioSquare Dr, and 93:

Even with potential cost savings from tunneling under the highway, this would still be a mammoth of a project, with a cost to match. I think it's possible that one could make a case for positive impact across the Green Line network (i.e. more than just positive impact for Nubian riders), but it's still probably a tall order.

But what if we can up the ante? What if, in addition to giving Nubian a speedy one-seat-ride downtown, we also provided a key piece of infrastructure to enable a southside LRT Urban Ring linking Nubian (and potentially beyond) with the Seaport?

Cut across to Broadway (somehow -- yes, big question mark here) and then cut across to South Boston Haul Road (yes, another big question mark here) and then run to the Seaport in the open cut of the current South Boston Haul Road. For less than half a mile of additional tunnel, you've built a fifth of an LRT Urban Ring. At that point, all you need to do (and it's definitely still a tall order, but all the same) is build a ~1 mile cut-and-cover subway between Huntington and Nubian, and then you have two full length circumferential LRT lines in dedicated ROWs originating at opposite ends of Longwood.

(Yes, this still doesn't solve the problem of getting across Longwood. But you can bet your CharlieCard, a Longwood West <> Cambridge <> Sullivan <> Airport line and a Longwood East <> Nubian <> Seaport line would be pretty fricking awesome on their own.)

Building a Nubian-Downtown subway that also can be used by the Urban Ring massively expands the scope of its impact. It would also massively expand the constituency that could lobby for it -- now this isn't just about expanding transit access for Dorchester, it's about linking together major employment centers (blah blah blah -- yes, I'm pointing out that this approach would bind together the fortunes of both wealthy and poor Bostonians, and it never hurts to get major institutions like Harvard, MGH/Brigham, and BU on your side).

In the name of crayoning perfectionism, I should acknowledge that, theoretically, you could utilize the Back Bay-South Station subway to connect the Urban Ring to the Seaport:

But to do that you'd need to tunnel under all of this, and also fit a flying junction of some kind into your alignment along Hudson St, which is already a finely threaded needle:

Of course, it's possible that cutting underneath the maintenance yards between 93 and Broadway will in fact be even hairier, so who knows.

tl;dr: Building a subway to Nubian will be Very Expensive no matter how you do it, so let's find a way to make it do more than "merely" connect Nubian with downtown -- if you're gonna Go Big, then you should properly Go Big.

 

[Teban54]

It's 11pm on a Friday night, which means it's time to draw some lines on maps.

None of these ideas have been examined critically at all.

Setting aside a second surface route, let's instead think about building a subway to Nubian. A couple points for our consideration:

1) No matter how you do it, it will be expensive. Whether it's dealing with landfill or undocumented utilities or deep bores or unhappy residents, the cost will be high.

2) The number one goal is to run trains quickly and frequently from Nubian to downtown. A nice number two goal would be serving BU Medical Center.

3) Local service on Washington St is explicitly out of scope and can be handled via a shorter surface route.

Earlier in the week we were chatting about tunneling under Harrison. But what if we looked further east, and instead tunneled under Albany and 93? (You could maybe also save on tunneling costs by running in an open cut under 93.)

View attachment 32432

Believe it or not, this is only a quarter mile longer than a Washington St alignment. And I would argue that you could get away with having only two intermediate stations (e.g. at the southern and northern ends of BUMC), or maybe even only one. 2 miles of running track is equivalent to Beaconsfield to Longwood, which is timetabled at 7 min with two intermediate stops. So an alignment like this could potentially compete with the travel times of the El.

If we don't feel like pissing off the hospitals by tunneling outside their front door, we could piss off everyone else and tunnel under a combination of Melnea Cass, the Mass Ave Connector, BioSquare Dr, and 93:

View attachment 32433

Even with potential cost savings from tunneling under the highway, this would still be a mammoth of a project, with a cost to match. I think it's possible that one could make a case for positive impact across the Green Line network (i.e. more than just positive impact for Nubian riders), but it's still probably a tall order.

But what if we can up the ante? What if, in addition to giving Nubian a speedy one-seat-ride downtown, we also provided a key piece of infrastructure to enable a southside LRT Urban Ring linking Nubian (and potentially beyond) with the Seaport?

View attachment 32434

Cut across to Broadway (somehow -- yes, big question mark here) and then cut across to South Boston Haul Road (yes, another big question mark here) and then run to the Seaport in the open cut of the current South Boston Haul Road. For less than half a mile of additional tunnel, you've built a fifth of an LRT Urban Ring. At that point, all you need to do (and it's definitely still a tall order, but all the same) is build a ~1 mile cut-and-cover subway between Huntington and Nubian, and then you have two full length circumferential LRT lines in dedicated ROWs originating at opposite ends of Longwood.

(Yes, this still doesn't solve the problem of getting across Longwood. But you can bet your CharlieCard, a Longwood West <> Cambridge <> Sullivan <> Airport line and a Longwood East <> Nubian <> Seaport line would be pretty fricking awesome on their own.)

Building a Nubian-Downtown subway that also can be used by the Urban Ring massively expands the scope of its impact. It would also massively expand the constituency that could lobby for it -- now this isn't just about expanding transit access for Dorchester, it's about linking together major employment centers (blah blah blah -- yes, I'm pointing out that this approach would bind together the fortunes of both wealthy and poor Bostonians, and it never hurts to get major institutions like Harvard, MGH/Brigham, and BU on your side).

In the name of crayoning perfectionism, I should acknowledge that, theoretically, you could utilize the Back Bay-South Station subway to connect the Urban Ring to the Seaport:

View attachment 32435

But to do that you'd need to tunnel under all of this, and also fit a flying junction of some kind into your alignment along Hudson St, which is already a finely threaded needle:

View attachment 32436

Of course, it's possible that cutting underneath the maintenance yards between 93 and Broadway will in fact be even hairier, so who knows.

tl;dr: Building a subway to Nubian will be Very Expensive no matter how you do it, so let's find a way to make it do more than "merely" connect Nubian with downtown -- if you're gonna Go Big, then you should properly Go Big.

I really like this proposal. Wow, this just keep getting more interesting.

One potential issue I can see is that in a pessimistic world where value engineering is a concern, the I93-to-Nubian subway may end up only serving Urban Ring and not a GL Nubian branch. One can argue that it does bring Nubian faster transit, just far from the best alignment (either at Broadway or South Station). Then, with the GL branch concern removed, they may even continue to lobby for the whole thing (now just Urban Ring) to be downgraded to surface LRT or BRT.

Obviously, if you're already tunneling up to E Berkeley/W 4th, connecting it to Bay Village is only another half a mile, but worse value engineering has happened before...

This also works less well with a phased approach. If Phase 1 already brings surface LRT down Washington St, with OSR to downtown subway stations, it's a harder sell politically to turn the South End riders back onto buses again, with no subway stations within walking distance, even if Nubian gets fast subway LRT.

 

[JeffDowntown]

The big issue with moving over to the Albany alignment is you lose about 1/2 the transit catchment. Rather than running through the middle of the neighborhood, you run along the edge with one side basically no mans land.

 

[Riverside]

The big issue with moving over to the Albany alignment is you lose about 1/2 the transit catchment. Rather than running through the middle of the neighborhood, you run along the edge with one side basically no mans land.

Well, yes, but I guess I'm saying, the point of building a subway to Nubian isn't about serving the South End -- it's about getting from Nubian to downtown on timeframes that are comparable to the El.

And on the southern end of the corridor (i.e. south of W Dedham St), you're still within the 10-min walkshed of either the Orange Line, the Albany corridor, or both:

At the northern end, my late-night sketch excluded any stations on the Albany corridor, but if you wanted to fill in the gap, a station at Harrison & Herald works nicely:

And in fact if you put at headhouse at Bay Village that's near the southern end of the complex (i.e. about a block south of the Orange Line entrance today), that gap gets nearly filled with a 10 min walkshed, and completely if you do a 15 min walkshed:

And either way, Washington St itself would maintain some level of surface transit anyway: a resurrected 49 (or continued SL4/SL5) would join the 7 and 92 & 93 as unusual group of radial bus services that run directly into downtown (and hopefully across downtown as well, but that's neither here nor there).

To be clear, I think a subway under Washington or Harrison (with some separate solution in an ideal world for Urban Ring LRT) would be better. But I also don't think it makes as big of a difference as it might first appear on paper -- it's the stations that determine the catchment, not the path of the route itself. If you look at the combined coverage of the Orange Line and the Albany corridor, most of the South End gets coverage anyway.

This also works less well with a phased approach. If Phase 1 already brings surface LRT down Washington St, with OSR to downtown subway stations, it's a harder sell politically to turn the South End riders back onto buses again, with no subway stations within walking distance, even if Nubian gets fast subway LRT.

Couple points:

• As discussed above, depending on what you do on the northern end of the corridor, most of the South End will still have a subway station within a 10 minute walk
  
  
• If a Phase 1 Surface LRT actually gets built, then this whole shebang goes back to Square 1, for many reasons. Honestly I am skeptical about that happening though, and I'm suggesting that a more ambitious plan might be more likely to get built first
  
  
• Buuuuuuut.... even if you did build a surface LRT first........ you might still have options:
  • 1: have two branches: "F to Nubian via Washington" and "G to Nubian via BUMC".
    • You'd be looking at 7.5-min headways on each branch, which isn't great, but isn't the end of the world.
    • You could potentially offset that with additional UR trains that run to the Seaport, especially if they go via South Station -- the current SL4/SL5 ridership split suggests that a fair number of riders are opportunistic and will be happy with either South Station or DTX/Park as destinations/transfer points
    • You could also potentially offset it by doing something like a 2:1 split, eg 6 min headways via BUMC [10 tph] and 12 min headways via Washington [5 tph] supplemented by SL5 v2.0 buses
  • 2: this is the rare scenario where you could potentially run the LRT line somewhere not in the Central Subway; you may recall that I suggested a Greenway streetcar called "the Rose Line", and you could potentially connect a Washington streetcar to that
    • I'm hesitant to mix "normal service" from a Washington line with the "tourist/entertainment/leisure service" of a Greenway trolley, but we could play with it
    • Or we could instead route Washington streetcars into the Seaport subway via a portal along Hudson St, which actually the more I look at it, the more I'd say that's the way to handle that objection, should it arise. It's far from perfect, but I think it's sellable

But like I said, I think it's unlikely that we'll encounter this scenario -- if there are already LRT tracks on Washington, then we just live with bus lanes and BRT for the southern Urban Ring.

One potential issue I can see is that in a pessimistic world where value engineering is a concern, the I93-to-Nubian subway may end up only serving Urban Ring and not a GL Nubian branch. One can argue that it does bring Nubian faster transit, just far from the best alignment (either at Broadway or South Station). Then, with the GL branch concern removed, they may even continue to lobby for the whole thing (now just Urban Ring) to be downgraded to surface LRT or BRT.

Obviously, if you're already tunneling up to E Berkeley/W 4th, connecting it to Bay Village is only another half a mile, but worse value engineering has happened before...

I agree with you: it's a high-risk-high-reward strategy. It's a proposal where potentially the whole is greater than the sum of its parts, and where -- exactly as you outline -- if you start to shave off pieces, the whole thing collapses. It will require leadership that actively wants to invest in transit (which Baker never has).

The problem is that this requires a study that, AFAIK, has never been done before, and what's more would be very ambitious, taking into account:

• The cost of LRT with dedicated lanes on Washington
• The benefits to riders of a surface OSR from Washington to Park St/Gov't Center/North Station
• The cost of a subway under Washington
• The benefits to riders of a subway OSR from Washington to Park/GC/NS
• The cost of a subway in the Albany corridor (via Albany and via the Mass Ave Connector)
• The benefits to riders of a subway OSR from Nubian to Park/GC/NS and from Nubian to Seaport
• The cost of either a surface or subway LRT connection from Nubian to Huntington (LMA)
• The added benefit of an LRT subway line from Huntington (LMA) to Seaport vs surface BRT options, and whether that has domino effects on the viability of an Albany corridor

The biggest question of course being what it would cost to tunnel under Washington. If we can somehow figure out that it would be an easy job, then all this other stuff fades away; if OTOH it's a hard job that might not pass cost/benefit, that's when this Go Big Or Go Home strategy steps up to the plate.

Spoiler: my hot takes on the above set of questions, though am pretty rambly and repetitive so caveat lector

• The cost of LRT with dedicated lanes on Washington
• The benefits to riders of a surface OSR from Washington to Park St/Gov't Center/North Station
  • I actually worry about this combination the most -- I think it'll hinge on where Nubian trains terminate downtown; if it's North Station, we're good, but if they assume Nubian trains will be turning at Park, we'll be in trouble
• The cost of a subway under Washington
• The benefits to riders of a subway OSR from Washington to Park/GC/NS
  • I think the benefit will be high, given the significant reduction in travel time and improvement in reliability; the wildcard is the cost, which I worry will be too high or too unpredictable to justify
• The cost of a subway in the Albany corridor (via Albany and via the Mass Ave Connector)
• The benefits to riders of a subway OSR from Nubian to Park/GC/NS and from Nubian to Seaport
  • I suspect the benefit of the toward-Park component will be similar to a Washington subway -- you lose some catchment, but you gain better access to jobs, and so will be overall pretty high; the benefit of the toward-Seaport piece I think will be high but not sure if it'll hit the same level
  • Even so, the amount of tunneling whether via Washington or Albany/93 will be broadly the same; so, theoretically, Albany's connectivity to the Seaport could create a "greater benefit for same tunneling" scenario
  • So, the cost may be very high, but the benefit will hopefully also be maximally high
• The cost of either a surface or subway LRT connection from Nubian to Huntington (LMA)
• The added benefit of an LRT subway line from Huntington (LMA) to Seaport vs surface BRT options, and whether that has domino effects on the viability of an Albany corridor
  • Again, it's a similar story here; we know that LMA <> Nubian presents a particular set of headaches, and so the question becomes, "If you can lock in an entire subway route from Huntington/LMA to the Seaport, does that change the calculus on any individual pieces?"

 

[Teban54]

The big issue with moving over to the Albany alignment is you lose about 1/2 the transit catchment. Rather than running through the middle of the neighborhood, you run along the edge with one side basically no mans land.

Coming back to this: I think "Albany vs Washington vs Harrison" and "combining Bay Village-Nubian and Urban Ring SE quad" should be seen as two separate decisions.

• The former is more for reducing cost of the subway, by digging under part of I-93.
• The latter is for increasing benefits of the subway and making it more justifiable, compared to a surface GL branch to Nubian.

In theory, you can build a subway under Washington or Harrison for full catchment, and still have it serve double duty for a GL branch and part of the Urban Ring, like this:

Still looks less "ring"-y than the SE Urban Ring could have been otherwise (via Andrew or something), and may have a few engineering challenges - you either need a 3-way flying junction under the Pike where the Nubian subway branches off the Huntington-Seaport subway, or need to figure out the Broadway alignment and tie it back to Washington/Harrison. But of all Nubian subway proposals, I'm inclined to think this maximizes both utility and catchment.

 

[Riverside]

Coming back to this: I think "Albany vs Washington vs Harrison" and "combining Bay Village-Nubian and Urban Ring SE quad" should be seen as two separate decisions.

• The former is more for reducing cost of the subway, by digging under part of I-93.
• The latter is for increasing benefits of the subway and making it more justifiable, compared to a surface GL branch to Nubian.

In theory, you can build a subway under Washington or Harrison for full catchment, and still have it serve double duty for a GL branch and part of the Urban Ring, like this:
View attachment 32633
Still looks less "ring"-y than the SE Urban Ring could have been otherwise (via Andrew or something), and may have a few engineering challenges - you either need a 3-way flying junction under the Pike where the Nubian subway branches off the Huntington-Seaport subway, or need to figure out the Broadway alignment and tie it back to Washington/Harrison. But of all Nubian subway proposals, I'm inclined to think this maximizes both utility and catchment.

Yeah, I could get on board with this. (I just realized I never responded to your thoughtful piece about the political factors needed to make this happen, so I'll go back and take a look.)

I think the revelation to me out of all of this is realizing that Washington vs Harrison vs Albany all end up being pretty much the same running time between Nubian and a junction at Bay Village. Albany "feels" so far removed from Washington, but it really isn't. So that's actually a pretty interesting range of flexibility.

And yes -- to your point -- an "Urban Ring" could travel along Washington and still hook over to either South Station or Track 61. I agree that it's less "Ring-y", but I don't think it actually ends up making a huge difference: in terms of Red Line crossing, it's a choice between Broadway and South Station -- if we're really talking about some sort of grade-separated ROW, I don't see how you go to either Andrew or JFK/UMass. So the "Ring" is already going to hew closer to the core along here anyway.

One "fun" thing about an alternative via Broadway: this is the rare situation where you could maybe justify reusing the Broadway streetcar level (which generally is oriented the "wrong way" to be useful) -- the footprint (including the northern incline) looks something like this:

Build a viaduct over the Cabot Yard (a railfan photographer's dream), and you could look at insertion points either from Herald St or from further south:

But regardless, I think you're right: largely independent of which street it goes under, a Subway-to-Nubian could potentially be integrated into a larger Urban Ring network, expanding its potential benefit and constituency.

 

[Riverside]

2: this is the rare scenario where you could potentially run the LRT line somewhere not in the Central Subway; you may recall that I suggested a Greenway streetcar called "the Rose Line", and you could potentially connect a Washington streetcar to that

• I'm hesitant to mix "normal service" from a Washington line with the "tourist/entertainment/leisure service" of a Greenway trolley, but we could play with it
• Or we could instead route Washington streetcars into the Seaport subway via a portal along Hudson St, which actually the more I look at it, the more I'd say that's the way to handle that objection, should it arise. It's far from perfect, but I think it's sellable

This has gotten stuck in my mind. I'm less wild about Washington streetcars going straight into the Seaport subway; you lose the connection to the Orange Line, and this route is already transfer-poor.

You could maybe achieve a Washington <> Orange transfer by adding yet another pedestrian tunnel to the sprawling Bay Village/Tufts Medical Center complex, with a headhouse at Washington & Marginal, but that's not a great transfer experience, and would be something like 1,000 feet from the Green platforms under Tremont.

So I've been thinking more about routing a Washington streetcar route up the Greenway. I think it's not quite a workable proposal, but it fares better than I expect the closer I look at it, so I'm not sure...

First, though, let me go through the benefits of keeping a Washington streetcar out of the Central Subway (assuming the construction of a subway to Nubian via BUMC):

Near-homogenous operating environment for Green Line network: in a Reconfigured Green Line, the Washington Street Branch would be very much the odd one out, being the only stretch of lengthy street-running (save perhaps for an extension to Hyde Square, depending on final build details), and therefore the only stretch was surface stop spacing rather than rapid transit stop spacing; with a subway via BUMC, both southern branches of the Green Line (this and Huntington) would have similar operating environments and similar needs.

Frequency independence: keeping a Washington streetcar out of the subway frees it from competing for slots in the subway, and increases the flexibility of its scheduling as it does not need to coordinate with other routes.

Parsimonious and higher frequencies for all: assuming a 30 tph limit between Park and GC (somewhat arbitrary but fine for now), removing the Washington streetcar from the mix means that you can see higher express frequencies to Nubian and potentially simpler frequencies across the system overall -- two potential distributions being:

• 15 tph to Park <> Huntington (then splitting among Riverside, Needham, Heath as you wish) and 15 tph on Park <> Nubian, giving you 4-min headways (which is probably the closest we've seen yet to "equal or better")
• 10 tph on Park <> Nubian, and then 10 tph each of two of the three Huntington branches (Riverside, Needham, Heath, with the third being picked up as a Seaport service)

Potential division of rolling stock: again, as an effect of the homogeneity of operating environments, keeping the Washington streetcar out of the Central Subway means that there are no concerns about running full-length doubleset Type 10 supertrains. I haven't looked carefully -- it may indeed be possible to place surface stops on Washington such that these 240' sets wouldn't cause any problems (blocking intersections, etc), but the problem is side-stepped altogether if the streetcar is kept out of the Central Subway.

These benefits -- particularly that increase in frequencies to reach "equal or better" -- seem non-trivial to me. And again, this goes back to the previous discussion about there being two separate objectives for rapid transit on Washington, existing in slight tension between each other: local service in the South End, and connecting the Dorchester bus network to downtown. Applying separate solutions potentially creates the best outcomes for each objective -- we've seen this apply when looking south of the Pike, so why not north of it as well?

"Why not?", he says...

The first reason "why not" is that running a Washington streetcar into the Central Subway is by far the simplest solution. Running to Government Center (which should be very achievable) gives you transfers to all the heavy rail services, and puts much of downtown within a 10-minute walkshed. Whether that simplicity is worth losing the benefits articulated above is a separate question, but it's one very obvious reason not to do anything crazy here.

Beyond that, the next reason "why not" is that there simply aren't a lot of good alternatives. As mentioned above, running into the Seaport subway misses transfer opportunities, and also doesn't provide great access to downtown. Creating a new LRT subway branching north from South Station is a fun idea, but hard to justify for just one line (unless you wanted to convert the T7 and T111 to LRT as well, which becomes a Whole Nother Thing).

So that leaves us with surface alternatives.

Getting from Washington St to South Station is mostly doable: Washington <> Kneeland <> Greenway, all of which are wide enough that they could be reconfigured with center-running lanes (I think). Running along Washington up to Kneeland would give a good transfer to the Orange Line at Tufts Medical Center, and an okay transfer to the Green Line at its Bay Village Tremont platforms.

Once at South Station, there are basically two corridors available: the Greenway via Aquarium and Congress St via Post Office Sq.

The Greenway, as I've described previously, could probably see a lane stolen in each direction to devote to some form of transit. The disadvantage there is that you don't serve the Financial District as directly (but, as I'll discuss below, it may not be any less direct than the Tremont Street Subway is).

Congress St runs through the heart of the Financial District; north of Post Office Square, it's wide enough to devote transit lanes, while to the south it'd be necessary to steal the parking lanes on Congress & Pearl, where the pavement is about 35 feet wide: say 12 feet for a streetcar lane, 10 feet for an automobile lane, with leftover space for bike lane(s), expanded sidewalks, and platforms for at least one pair of stops perhaps at Franklin St or High St.

On the other hand, Congress St is already being eyed for a BRT corridor for the likes of the T7 (to Southie and Charlestown), and I argue it'd be worth extending the T111 at least to South Station, and potentially include a service to Logan via TWT. The BNRD calls for 8 bph on the T7 and 12 bph on the T111, which puts us as 20 bph, or a bus every 3 minutes in each direction. (And that's before we think about Logan services). I don't think we really want to try to put more vehicles through that corridor.

This could be addressed potentially by extending a Washington streetcar to replace one of those northern routes -- e.g. extend it to Charlestown (and thereby complete our off-set rebuild of the Els of old) -- or reroute the T111 via the Greenway instead of Congress St -- but that expands the scope of this project and requires that much more to disentangle.

To me, this means the question comes down to, "Could a Washington streetcar extended to Haymarket via the Greenway provide good and sufficient transit?"

And, it bears acknowledgement, that this option would be an improvement on today's Silver Line: larger vehicles with greater capacity, smoother ride, OSR deeper into downtown, direct transfers to the Blue Line, and probably no need to transfer for most commutes into downtown. (And there is some insane scenario where the T balks at integrating a new branch into the subway but where the City gets enough funding and gumption to just build their own damn trolley along this route and call it a win.)

The question then becomes, would a Washington-Greenway-Haymarket surface streetcar be better than a Washington-Park-Government Center subway streetcar? And this is where I falter a bit because... on the face of it, it seems like it should be obvious that the subway streetcar beats out the surface streetcar under any scenario. And yet, as I look closer, I don't actually think it's that obvious.

[continued below]

 

[Riverside]

[continued from above]

Reach of service: the Surface Alt could reach as far as Haymarket (and maybe North Station, but I think that gets hairier and for less gain); the Subway Alt might be able to go as far, but depending on what happens on the northside, it may be desirable to turn one of the southside branches at Government Center, and I think a Washington streetcar would be the leading candidate. (Recall that in this scenario we also have a subway to Nubian via BUMC, and that branch would probably get the priority to consume the through-running slots.) This is a toss-up but might lean in favor of the Surface Alt

Frequency of service: as noted above, the Surface Alt has more flexibility for frequencies; the conditions on Washington St itself probably keep the ceiling relatively modest, so I wouldn't expect much better than 5-min headways, but that's slightly better than the Subway Alt which, needing to compete with a Nubian subway plus the Huntington services, is probably capped at 7.5 minutes; again, probably a toss-up leaning in favor of the Surface Alt

Transfers: mostly equivalent between the two, with a slight edge to the Subway Alt because of the better transfers to the southern half of the Green Line (though this isn't a huge edge, as the T9, T8, and T1 would provide potentially faster two-seat rides between the South End and the Boylston and Huntington corridors)

Reliability: clear win for the Subway Alt on this one, as the Surface Alt will have twice as much street-running (even with dedicated lanes); but, the story is still a little more complicated because the Subway Alt will have to compete with a number of other routes in the subway -- if strategic turnbacks were included in the Surface Alt, there could be more flexibility to recover from delays; the Surface Alt would also be immune to any cascading delays on the rest of the LRT network

Journey time: So. This is where I think things get really interesting. Obviously the Subway Alt will traverse the 1 mile to Government Center faster than the Surface Alt will reach Rowes Wharf (same distance). But the employment heart of downtown is oriented along Congress St, not Tremont St (see below) -- meaning that commutes from Park St and to a lesser extent Government Center will require a "last-mile" walk of ~10 minutes. (This also means that South Station itself is a meaningfully competitive destination of its own right.)

The Orange Line today is timetabled at 5 minutes between Tufts Medical Center and State. Our origination point is a bit to the south, at the intersection of Washington & Marginal; the Subway Alt will probably take a little longer, given the curve it will navigate to access the Pleasant Street Portal, so probably comparable to today's Arlington-Government Center journey, which is 7 minutes. A commute from Arlington (as a stand-in for Bay Village) to the southern half of Post Office Sq today is 14 minutes: 5 min Green to Park St + 9 min walk.

The speed of the the Surface Alt will depend on how fast it can buzz through all the intersections it would hit. At the slow end of 5.5 mph (roughly equivalent, I think, to today's street-running E Line), a journey from the origination point to South Station (.75 miles) would take just over 8 minutes. Add a 5-min walk to reach the same point in Post Office Square, and it's a 13-min journey, which is basically equal to the Subway Alt. If we are more optimistic and assume 8 mph, that drops the journey time to 10.5 minutes, which gives the Surface Alt an edge in terms of overall journey time to reach downtown jobs, with the slower transit time being offset by the shorter walk.

If we assume for the moment that the above analysis suggests that a large number of downtown jobs are more or less equally accessible via the Subway Alt (shorter ride + longer walk) and via the Surface Alt (longer ride + shorter walk), then we turn to the jobs which are uniquely accessible by each alternative. And it turns out, the Surface Alt edges out here, slightly.

Using a 1000' radius around each stop (to approximate a 5 min walkshed), the Subway Alt serves 65,000 jobs:

The Surface Alt comes in slightly higher, at just shy of 75,000 jobs:

(And it should be noted that a Surface Alt on Congress St instead of the Greenway blows both other options out of the water, with nearly 88,000 jobs in a 5 min walkshed and probably the shortest average last-mile walk for commuters.)

So, this suggests that for a majority of Washington <> Downtown commuters, the Surface Alt and Subway Alts will be pretty similar in overall journey time, and that the Surface Alt probably scores mildly stronger in terms of access to jobs.

~~~

To summarize: on closer inspection, it looks like a Greenway Surface Alt would exceed a traditional Subway Alt in terms of:

• Frequencies
• Journey time to many Financial District jobs
• Number of jobs accessible within 5 min walkshed
• Number of OSR destinations

And a Greenway Surface Alt would enable improved operations on the Green Line proper by:

• Increasing frequencies to Nubian to match frequencies of the El
• Improved travel times between downtown and Nubian, getting closer to El travel times
• Improving overall network reliability with minimized street-running with surface stop spacing
• Homogenized rolling stock needs, with all branches running primarily in dedicated ROWs with space for full-length trains

So... yeah. To me, that is a thought-provoking set of costs and benefits. The lightbulb moment for me really was in thinking about the last-mile walk from the current Green Line stations -- it's not like they don't provide access to downtown, but it's not like they are ideal. And then I got to thinking about the character of the different lines -- a Washington streetcar really will be the odd one out in a Reconfigured Green Line, and it really was much more similar to what I had described for a Greenway-running Rose Line.

And that brings up one final interesting piece to consider: one of the benefits I had described of a "Rose Line" was its ability to "placemake" as a destination unto itself -- becoming part of your visit to Boston, not just a way to get around during your visit to Boston. In some respects, using the Greenway route for a conventional transit service takes away from that, potentially. (For those curious about what I'd do at the north end [*rimshot*] of the Rose Line, see the spoiler box.)

But in other respects, there's something worth pondering about the possibility of a placemaking Rose Line being extended through the South End to Nubian, knitting together the city and making the tacit statement that "the heart" of Boston doesn't stop at the edge of downtown.

Spoiler: Rose Line to the North End

Whither the Rose Line to the North End? I have a solution for that, and in some ways it's better than the original proposal. Instead of a North Station <> South Station bidirectional service, there is now the full-time bidirectional Haymarket <> Nubian service, and a unidirectional loop running counter-clockwise from North Station to Haymarket to Hanover St to Commercial St to North Station. The North End loop is short enough that it won't feel "out of the way" to go "the wrong way round," especially since you can get off "early" at North Station if you wish. Decreasing to a single track also reduces the impact on the neighborhood, hopefully reducing opposition. It also allows for taut headways on what I see as the key part of the journey -- eliminating the slightly-too-long walk from Haymarket/North Station.

Most of the time, the Haymarket <> Nubian and North End Loop would run as separate services. But during evenings and weekends -- without the rush hour pressure on Nubian <> Downtown service -- I'd consider extending the Nubian service through the Loop, to provide the one-seat ride from South Station for those visiting Boston for an evening out. (The dedicated North End Loop would continue to run to provide a timely North Station -> North End OSR.)