Crazy Transit Pitches

[Teban54]

With the exception of some of SoWa, the Broadway-Nubian route serves basically the same areas. The only people who would see a difference would be those who are starting/ending a journey at South Station already, but for those people a light rail line along Washington St would serve them. I don't think a line along Albany/93 is a substitute for a line along the Washington St corridor, nor is a line to Broadway and Seaport. Given this, it seems silly to build two essentially parallel routes, except one has a much worse walkshed due to its location along 93, that would have to share traffic between them, as opposed to a pair of lines serving somewhat different groups of people.

I'll start with this one by explaining why I feel a Nubian subway (whether via I-93, Harrison or Washington) is a necessity in the long term. Most of these points were raised in the discussion starting from here.

• Surface light rail on Washington is good for intermediate neighborhoods, but terrible for Nubian, where majority of the ridership comes from. The many intersections (and LRT's worse speed for handling them) mean you won't see much time savings from SL5, and it's really not "equal or better" than the El. People might still end up taking the bus to Ruggles.
• The low reliability of a streetcar branch also means you can't increase frequency meaningfully before bunching becomes commonplace, so it hurts capacity as well. Current SL4/5 have 4-5 min average headways during PM rush, better than any GL branch.
• For GL Reconfiguration, having a high-reliability, high-frequency faucet into Tremont is important due to reverse branching from Huntington. If you want to boost service on the important Huntington-SS-Seaport corridor, additional frequency for Bay Village-Park-GC-NS has to come from Nubian.

To be clear, I'm envisioning a Washington St streetcar as an initial build, and a more direct Nubian-downtown subway as a later, high-speed solution for Nubian that coexists with the Washington St local service.

You did make a good point about walkshed, but:

• The Nubian subway has a specific focus of offering a speedy ride to Nubian (and leaving local demand for Washington streetcar).
• It also serves BMC, a valuable node in itself that's just a little bit out of the way from Washington.
• Widett Circle has worse walkshed than a station at Albany/E Berkeley.
• And I mean, you can build a subway under I-93 and still hit Broadway, if you really prefer Broadway over South Station. Compared to a standalone Widett-Broadway grade-separated route and a Washington streetcar, this alternative essentially gives you three service patterns for little additional cost (and much better RL transfer).

I'm referring to the debate of Broadway vs South Station here. I understood your point arguing for South Station over Broadway to be that the line would connect with all regional rail routes (Assuming NSRL). My counterpoint is that with a RR infill station at Widett Circle, a line from Ruggles to Broadway would also connect with almost all, or with an extension to Kenmore/Lansdowne all of, the lines as well, essentially negating the advantages of a link to South Station for this purpose.

FWIW, I'm not concerned about Regional Rail lines at Ruggles and further north. The Huntington-South Station-Seaport subway essentially serves as a "semi-ring" in this case, and if you're building a Ruggles-Nubian subway (whether it continues underneath LMA or not), extending some form of service east to BMC seems like a piece of cake. So the specific demand patterns relevant here are: (Ashmont, Braintree, Fairmount, Old Colony) - (Nubian, BMC, maybe Ruggles and LMA).

To get to Nubian and BMC, any alignment south of South Station has to choose between missing Regional Rail (don't forget pseudo-rapid-transit Fairmount) or missing Red Line (or having a highly inconvenient transfer to either). South Station gets you both. Also, this distance is also short enough that I feel buses may be a better solution.

To get to LMA, there's always the option of transferring to the Huntington subway at South Station. Yes it's circuitous, but Broadway-Nubian-Ruggles-LMA is also quite a zigzag (and again, choice between RR and RL). You need to really get down to Andrew, Newmarket and JFK/UMass to offer significant distance reductions, at which point it feels more like a branch. I know @Riverside has long been skeptical that any circumferential solution for this will be competitive with the SS-Huntington transfer, and while I plan to look into this further, right now I'm not sure if the cost of specifically building this connection can be justified (especially with no ROWs). That's why my original proposal basically treats this as a byproduct of the Nubian subway (or a 50-50 purpose), not a dedicated circumferential connection.

(Thinking about it now, I wonder if there will be value in shifting Newmarket further north to Southampton St, such that it can be better aligned with a route to Andrew. Old Colony remains an issue, though. I guess the main point is: If we're building a dedicated line between southside lines and LMA, it's probably more worthwhile to look at connecting at Newmarket and JFK/UMass.)

 

[Teban54]

Another issue with a Widett infill is that it sits right at the start of the NSRL portal, as shown below. Even if there's room for a surface station through the spaghetti of tracks, you basically need to have separate platforms for Fairmount and Old Colony lines.

 

[TheRatmeister]

o get to LMA, there's always the option of transferring to the Huntington subway at South Station. Yes it's circuitous, but Broadway-Nubian-Ruggles-LMA is also quite a zigzag (and again, choice between RR and RL). You need to really get down to Andrew, Newmarket and JFK/UMass to offer significant distance reductions, at which point it feels more like a branch. I know @Riverside has long been skeptical that any circumferential solution for this will be competitive with the SS-Huntington transfer, and while I plan to look into this further, right now I'm not sure if the cost of specifically building this connection can be justified (especially with no ROWs). That's why my original proposal basically treats this as a byproduct of the Nubian subway (or a 50-50 purpose), not a dedicated circumferential connection.

Just doing some quick estimates, assuming an average speed of ~20mph the trip from Broadway to LMA would take around 9 minutes, or about 7.5 if a speed of 25mph is assumed. South Station to LMA via Huntington would likely take twice as long. As for a Washington St route, even with a pretty anemic average speed of 10mph the line from Park St to Nubian would take around 14 minutes. For the people who need to get specifically from downtown to Nubian, I think this is acceptable.

Surface light rail on Washington is good for intermediate neighborhoods, but terrible for Nubian, where majority of the ridership comes from. The many intersections (and LRT's worse speed for handling them) mean you won't see much time savings from SL5, and it's really not "equal or better" than the El. People might still end up taking the bus to Ruggles.

This is only a problem if were to just plop a streetcar line on Washington St today and just, hope for the best. Speed is very much a solvable problem, many minor cross streets can be eliminated, left turns can be restricted, and for when intersections/lights cannot be removed there is signal priority.

• The Nubian subway has a specific focus of offering a speedy ride to Nubian (and leaving local demand for Washington streetcar).
• It also serves BMC, a valuable node in itself that's just a little bit out of the way from Washington.

Yes, but going under Melnea Cass to Broadway also does both of these things. The big advantages of Nubian-South Station are the connections with the FiDi and Regional Rail. But if the former can be taken care of adequately by a surface light rail line and the latter can be accomplished at Widett Circle (Which again is not a given, but since it avoids the challenge of navigating to South Station while also leaving future options for extension open I'd imagine a station at Widett Circle would have to be basically impossible for this not to be the cheaper option), then why bother with South Station?

 

[Riverside]

Awesome discussion so far. I don't have time to weigh in on everything, but did want to briefly respond to this point:

As for a Washington St route, even with a pretty anemic average speed of 10mph the line from Park St to Nubian would take around 14 minutes. For the people who need to get specifically from downtown to Nubian, I think this is acceptable.

As I argue here, for riders transferring at Nubian, transferring to a surface line on Washington St probably won't have significant time-savings over riding all the way to Ruggles. For equal-or-better replacement of the El, we need to aim for service that expresses through the South End, making only one or two stops at most. The I-93 alignment potentially enables that without the need to tunnel all the way.

 

[Teban54]

Just doing some quick estimates, assuming an average speed of ~20mph the trip from Broadway to LMA would take around 9 minutes, or about 7.5 if a speed of 25mph is assumed. South Station to LMA via Huntington would likely take twice as long.

Thanks for doing part of what I wanted to do for the whole system

However, I'm not sure where you got the "twice as long" comment. My rough distance measurements on Google Maps give a distance of 3.21 miles for the Broadway-Widett-Nubian-Ruggles-LMA alignment to the E's LMA station, while Broadway-SS-Bay Village-LMA is 3.53 miles. Both rides will have exactly one transfer, at either Broadway or South Station, so the transfer time breaks even.

If anything, a connection at Andrew offers the most significant distance savings at 2.72 miles (whereas an extra 0.9 is added to all figures above due to Andrew-Broadway), which will probably become half of SS. This actually convinced me more about value of an Andrew/JFK alignment or branch - the biggest issue will probably be cost (since you're building a standalone connection at that point for this specific purpose), and for Andrew, loss of Regional Rail connections.

As for a Washington St route, even with a pretty anemic average speed of 10mph the line from Park St to Nubian would take around 14 minutes. For the people who need to get specifically from downtown to Nubian, I think this is acceptable.

This is only a problem if were to just plop a streetcar line on Washington St today and just, hope for the best. Speed is very much a solvable problem, many minor cross streets can be eliminated, left turns can be restricted, and for when intersections/lights cannot be removed there is signal priority.

Most of what I want to say have been mentioned in Riverside's comment above and @The EGE's initial comment that triggered the whole discussion (which I highly encourage you to read). TL;DR: Eliminating street crossings and signal priority are hard to do when you have busy commercial corridors with many pedestrians, which are as desirable as transit for urban neighborhoods like South End.

As another note, I can't imagine a surface light rail line getting from Washington St to South Station directly with any form of reliability.

Yes, but going under Melnea Cass to Broadway also does both of these things. The big advantages of Nubian-South Station are the connections with the FiDi and Regional Rail. But if the former can be taken care of adequately by a surface light rail line and the latter can be accomplished at Widett Circle (Which again is not a given, but since it avoids the challenge of navigating to South Station while also leaving future options for extension open I'd imagine a station at Widett Circle would have to be basically impossible for this not to be the cheaper option), then why bother with South Station?

Nubian-Broadway doesn't offer Nubian/Roxbury/Dorchester riders a speedy ride to downtown, which is by far the biggest destination node and offers more direct transfers to more places than can be done anywhere else (e.g. Blue, Orange north, Green, possibly Back Bay). You're basically forced to transfer at Broadway if that's your only "inbound" subway connection from Nubian, and that doesn't even match the established travel patterns. Likewise, it's also less ideal for BMC workers and visitors coming from these lines. Broadway gets you southside RL, Old Colony and likely better Southie transfers, but while they're valuable, they pale in comparison to all the other connections above combined.

There's also the elephant in the room of how you can even put a Regional Rail station at Widett Circle in the first place - realistically two stations, one for Fairmount and one for Old Colony. That area is right next to Amtrak's Southampton Yard, Red Line's Cabot Yard leads, and a whole lot of other things.

 

[Riverside]

Just doing some quick estimates, assuming an average speed of ~20mph the trip from Broadway to LMA would take around 9 minutes, or about 7.5 if a speed of 25mph is assumed. South Station to LMA via Huntington would likely take twice as long. As for a Washington St route, even with a pretty anemic average speed of 10mph the line from Park St to Nubian would take around 14 minutes.

I know I'm cherry-picking things to respond to here, my apologies. But I wanted to offer some context on the numbers you are estimating. While actual speed of trains in motion can climb relatively high, the average speed of trains over the overall course of their run -- inclusive of accelerating, decelerating, and dwell times -- is significantly lower than the numbers you are using here. (I was really surprised the first time I looked at the data.) Courtesy of the Transit Matters dashboards:

Red Line: historical max of 21.2 mph, but on average about 18 mph

Orange Line: historical max of 18 mph, but on average about 16 mph

Blue Line: historical max of 20.5 mph, but on average about 19 mph

Green Line: historical max of 12.6 mph, average of 10.8 mph

Looking specifically at the D Line (to discount the surface line portions of the B/C/E): on April 18 2018, the average journey time from Riverside to Government Center was just shy of 45 minutes (and rises to just above 45 minutes if you discount the unusually fast night services):

45 minutes to travel 11.9 miles is about 15.8 mph.

And like I describe in my linked post, neither the surface lines nor SL4/SL5 come close to breaching the 10 mph barrier.

All of which is to say, the range we probably want to be using for back-of-the-napkin estimates is more like 15-20 mph rather than 20-25 mph. (At least, based on the data that I've seen.) With those figures :

LMA <> Ruggles <> Melnea Cass <> BUMC <> Broadway --> 2.9 miles = between 8.7 and 11.6 min

LMA <> Back Bay <> South Station --> 2.8 miles = between 8.4 and 11.2 min

Note that my distance measurements are not precise, so I'd say we can assume these two corridors are basically equal length. One significant difference between the routes would be number of stops; I usually sketch about 8 stops to South Station, and something like 5 stops to Broadway.

Will fewer stops result in huge time-savings? I'm not sure. Two examples for us to consider are DTX <> Jackson Square (3.1 miles, 7 stops, 11.75 min, 15.8 mph) and DTX <> JFK/UMass (2.6 miles, 4 stops, 7.75 min, 20.1 mph). Those speed both fall within our 15-20 mph range, and suggest that "Orange Line stop spacing" puts you at the lower end of that range and "Red Line stop spacing" puts you at the higher end. So, potentially we could infer that the Broadway alignment would be about 8.7 minutes, and the South Station alignment would be about 11.2 minutes.

Anyway, this has become a bit of a digression, and I'm not sure it materially changes the discussion you and @Teban54 are having, but I thought it was worth sprinkling in.

 

[Riverside]

Obviously, a double-deck, quad-"track" Ruggles-Nubian tunnel for both buses and LRT/HRT is the best, but in the event that it's too expensive and we have to pick one, I feel* that a tunnel aiming for regional connections (that still allows a quick transfer for Nubian bus riders) may have better value than a "modernized local streetcar network with a Central Subway-equivalent". Also, while a tunnel to Ruggles/LMA is obviously the best for buses, it can be argued that dedicated bus lanes on Malcolm X and/or Melnea Cass may be good enough, especially when you can't continue the tunnel into LMA proper without huge expenses. Politically, giving wealthy Cambridge/Somerville rapid transit for Urban Ring while leaving historically underserved Dorchester with a bus network also sounds problematic.

This last point (emphasis mine) has me a little confused. Even in a world where a subway is built from LMA across the Orange Line to Washington St (and points east), Dorchester is still "only" getting a bus network; riders still will have to transfer to get to Longwood, which arguably will be a step backwards assuming the T28 and other BNRD extensions are implemented. And a bus tunnel, with that many services being fed into it, would serve pretty adequately as a circumferential service to connect Nubian with the Orange Line and Longwood.

I agree that the tunnel would be expensive and hard to justify, but if anything it seems to me that killing two birds with one stone will improve its efficiency rating.

However, Nubian has demand for points northwest of LMA, such as BU/Kenmore, Kendall and Harvard. This part depends a lot more on the exact configuration of these services, such as whether a cross-LMA tunnel is built and whether the Charles River crossing uses BU Bridge or possibly a new tunnel. But [in the event that a through-running circumferential service from Cambridge somehow ends up getting to Ruggles], it seems like a waste to not let it go further to Nubian -- and an exclusive Ruggles-Nubian subway disrupts that.

This could be addressed by extending my H to Nubian, which I would feel better about if it had a tunnel for part of that (even if it has to share said tunnel with buses). But yes, I agree: Longwood creates a physical blocker that makes these corridors much more challenging:

So I started looking into a Nubian-South Station-Seaport Transitway connection... TL;DR: Not very encouraging in terms of engineering and political feasibility, though I do have a proposal that may work.

Yeah these are interesting ideas, especially repurposing Albany St itself. In general, I'm not super convinced of a rail OSR between Nubian and Seaport -- particularly if the Bypass Road can be used, I think express bus services could work nicely --, and the only reason I include the W on the map at all is because I think it would be relatively low-impact to add on to an I-93 build. Filling in the Albany Wye seems like an awful lot of headache for a problem that isn't super urgent.

 

[Tallguy]

So, the big question on Nubian GL service is the eventual ridership of maxed out Fairmount and how much of the bus ridership from BHA, Morton etc is diverted to Jackson/Ruggles by BRTish service. Much of that ridership might go east to ride Fairmount as well. Nubians function as a mode junction might be greatly diminished in the longer term

 

[Teban54]

This last point (emphasis mine) has me a little confused. Even in a world where a subway is built from LMA across the Orange Line to Washington St (and points east), Dorchester is still "only" getting a bus network; riders still will have to transfer to get to Longwood, which arguably will be a step backwards assuming the T28 and other BNRD extensions are implemented. And a bus tunnel, with that many services being fed into it, would serve pretty adequately as a circumferential service to connect Nubian with the Orange Line and Longwood.

You brought up a very good point. In a world where you literally can't get a circumferential route across LMA at all, any southside circumferential route has to be limited to LMA-Ruggles-Nubian (maybe to BMC and/or Andrew/JFK), so it makes a lot more sense to allow buses to use this tunnel after merging from Nubian.

The calculus changes entirely if you're willing to build a cross-LMA bored tunnel, though. Regardless of whether you're using it for Huntington-D service or for northside Urban Ring (I prefer a mix of both), it seems too precious to not run some kind of rail on it, at which point mixing them with buses seems infeasible. Imagine if you have a half-ring from Chelsea all the way to LMA tunnel (or even Ruggles) - and then stop there, while buses run in a different tunnel from Nubian to Ruggles and LMA. That's where the political issues arise.

A harmonious solution, though, is to have an initial build for buses and later repurpose it for a continuous ring with a cross-LMA tunnel, similar to what Seattle did. And of course, whether we can actually afford to build a cross-LMA subway is the elephant in the room. It depends on whether Boston - as a region that's known for both its healthcare industry and public transit - is satisfied with letting its healthcare workers forever endure a 10-min walk from the nearest rapid transit stations.

Yeah these are interesting ideas, especially repurposing Albany St itself. In general, I'm not super convinced of a rail OSR between Nubian and Seaport -- particularly if the Bypass Road can be used, I think express bus services could work nicely --, and the only reason I include the W on the map at all is because I think it would be relatively low-impact to add on to an I-93 build. Filling in the Albany Wye seems like an awful lot of headache for a problem that isn't super urgent.

To be fair, it's not like the Albany-Broadway alignment doesn't have its own list of issues either:

• Turning from I-93 to Traveler St, while rising from underground to a viaduct over Cabot Yard, all within a short span of 300-700'; and maintaining grade separation, which requires you to cut through both surface roads and underground road tunnels near Traveler St, while keeping this pedestrian trail accessible
• Descending from the viaduct back to the streetcar portal at Foundry St, which is at the same level as Cabot Yard and just 110' from the nearest track, while still giving enough vertical clearance for Red Line trains; or destroy the Foundry St portal and build another one, while somehow achieving the same elevation change before hitting Broadway's station box with a starting point above the Foundry St portal, not to mention structural modifications and grade separation
• Turning from Broadway streetcar level to Track 61, where any alignment needs to make a turn under some buildings, and likely with a tight curve that limits speed and reduces time savings
• Creating space on Track 61 for two tracks, while keeping the road functional and dealing with MassDOT, Massport and CSX to coordinate removal of freight service, especially with planned freight traffic to Marine Terminal (as F-Line explained here and here)
• And last but not least, actually putting the station at Seaport - yes your proposals are awesome, but they still take work, and still miss Courthouse

Some are more doable than others, but a quick glance suggests that things may be even more problematic than the Albany Wye proposals, especially if my first proposal can actually work. If the main criticism of Albany Wye is that a short section on the map is engineeringly difficult in practice, the same applies to the jump between Albany/I-93 and Track 61, not to mention issues with Track 61 itself.

On the other hand, an emerging theme from my exchange with @TheRatmeister above is: The Nubian-Seaport lag is likely the weakest quadrant in the whole Urban Ring and offers the least potential time savings, largely due to how the rapid transit network to the south is already going diagonally NE. I'm even inclined to argue that it doesn't need to be built at all - any southside circumferential route aiming to connect RL branches, Fairmount and Old Colony may be better off going to Andrew or JFK, while any bored tunnel from Logan (if it happens) is probably better off hitting South Station and Back Bay via the Huntington-Seaport subway.

 

[Riverside]

I'll start with this one by explaining why I feel a Nubian subway (whether via I-93, Harrison or Washington) is a necessity in the long term. Most of these points were raised in the discussion starting from here.

• Surface light rail on Washington is good for intermediate neighborhoods, but terrible for Nubian, where majority of the ridership comes from. The many intersections (and LRT's worse speed for handling them) mean you won't see much time savings from SL5, and it's really not "equal or better" than the El. People might still end up taking the bus to Ruggles.
• The low reliability of a streetcar branch also means you can't increase frequency meaningfully before bunching becomes commonplace, so it hurts capacity as well. Current SL4/5 have 4-5 min average headways during PM rush, better than any GL branch.
• For GL Reconfiguration, having a high-reliability, high-frequency faucet into Tremont is important due to reverse branching from Huntington. If you want to boost service on the important Huntington-SS-Seaport corridor, additional frequency for Bay Village-Park-GC-NS has to come from Nubian.

To be clear, I'm envisioning a Washington St streetcar as an initial build, and a more direct Nubian-downtown subway as a later, high-speed solution for Nubian that coexists with the Washington St local service.

You did make a good point about walkshed, but:

• The Nubian subway has a specific focus of offering a speedy ride to Nubian (and leaving local demand for Washington streetcar).
• It also serves BMC, a valuable node in itself that's just a little bit out of the way from Washington.
• Widett Circle has worse walkshed than a station at Albany/E Berkeley.
• And I mean, you can build a subway under I-93 and still hit Broadway, if you really prefer Broadway over South Station. Compared to a standalone Widett-Broadway grade-separated route and a Washington streetcar, this alternative essentially gives you three service patterns for little additional cost (and much better RL transfer).

FWIW, I'm not concerned about Regional Rail lines at Ruggles and further north. The Huntington-South Station-Seaport subway essentially serves as a "semi-ring" in this case, and if you're building a Ruggles-Nubian subway (whether it continues underneath LMA or not), extending some form of service east to BMC seems like a piece of cake. So the specific demand patterns relevant here are: (Ashmont, Braintree, Fairmount, Old Colony) - (Nubian, BMC, maybe Ruggles and LMA).

To get to Nubian and BMC, any alignment south of South Station has to choose between missing Regional Rail (don't forget pseudo-rapid-transit Fairmount) or missing Red Line (or having a highly inconvenient transfer to either). South Station gets you both. Also, this distance is also short enough that I feel buses may be a better solution.

To get to LMA, there's always the option of transferring to the Huntington subway at South Station. Yes it's circuitous, but Broadway-Nubian-Ruggles-LMA is also quite a zigzag (and again, choice between RR and RL). You need to really get down to Andrew, Newmarket and JFK/UMass to offer significant distance reductions, at which point it feels more like a branch. I know @Riverside has long been skeptical that any circumferential solution for this will be competitive with the SS-Huntington transfer, and while I plan to look into this further, right now I'm not sure if the cost of specifically building this connection can be justified (especially with no ROWs). That's why my original proposal basically treats this as a byproduct of the Nubian subway (or a 50-50 purpose), not a dedicated circumferential connection.

(Thinking about it now, I wonder if there will be value in shifting Newmarket further north to Southampton St, such that it can be better aligned with a route to Andrew. Old Colony remains an issue, though. I guess the main point is: If we're building a dedicated line between southside lines and LMA, it's probably more worthwhile to look at connecting at Newmarket and JFK/UMass.)

Quoting all of this because I agree with all this and think it is well-put. F Line to Nubian replaces the 49 and local streetcars of days of yore, and G Line to Nubian replaces the El. The I-93 alignment is the only one that I've come across that can come close to replacing the El without requiring a subway the full length of the South End.

And yes -- to your last point -- this discussion has definitely made me consider more carefully the value of a connection in the Andrew/Newmarket/JFK/UMass vicinity. JFK/UMass seems pretty difficult, but Andrew at least is a relatively short straight shot down Southampton St. And yes, it would be nice if Newmarket station were closer to Southampton St in such a scenario.

Alternatively, perhaps you could use the airspace above the various highways to string together a dedicated ROW for a connection to Andrew:

On the other hand, it's not like Southampton St is a picturesque residential neighborhood, so if you're gonna do all the above, you could just build an elevated over Southampton and be done with it.

 

[Riverside]

Our replies fired at nearly the same time. Let me respond to your latest now.

You brought up a very good point. In a world where you literally can't get a circumferential route across LMA at all, any southside circumferential route has to be limited to LMA-Ruggles-Nubian (maybe to BMC and/or Andrew/JFK), so it makes a lot more sense to allow buses to use this tunnel after merging from Nubian.

The calculus changes entirely if you're willing to build a cross-LMA bored tunnel, though. Regardless of whether you're using it for Huntington-D service or for northside Urban Ring (I prefer a mix of both), it seems too precious to not run some kind of rail on it, at which point mixing them with buses seems infeasible. Imagine if you have a half-ring from Chelsea all the way to LMA tunnel (or even Ruggles) - and then stop there, while buses run in a different tunnel from Nubian to Ruggles and LMA. That's where the political issues arise.

A harmonious solution, though, is to have an initial build for buses and later repurpose it for a continuous ring with a cross-LMA tunnel, similar to what Seattle did. And of course, whether we can actually afford to build a cross-LMA subway is the elephant in the room. It depends on whether Boston - as a region that's known for both its healthcare industry and public transit - is satisfied with letting its healthcare workers forever endure a 10-min walk from the nearest rapid transit stations.

I like your thinking on this. Yes, I think a phased approach is the best way to think about this -- which means that the objective becomes "design a tunnel that is useful for a radial Dorchester network in the short-turn, and circumferential service in the long run", which may be a challenge in its own right, but at least it's more clearly defined.

To be fair, it's not like the Albany-Broadway alignment doesn't have its own list of issues either:

• Turning from I-93 to Traveler St, while rising from underground to a viaduct over Cabot Yard, all within a short span of 300-700'; and maintaining grade separation, which requires you to cut through both surface roads and underground road tunnels near Traveler St, while keeping this pedestrian trail accessible
• Descending from the viaduct back to the streetcar portal at Foundry St, which is at the same level as Cabot Yard and just 110' from the nearest track, while still giving enough vertical clearance for Red Line trains; or destroy the Foundry St portal and build another one, while somehow achieving the same elevation change before hitting Broadway's station box with a starting point above the Foundry St portal, not to mention structural modifications and grade separation
• Turning from Broadway streetcar level to Track 61, where any alignment needs to make a turn under some buildings, and likely with a tight curve that limits speed and reduces time savings
• Creating space on Track 61 for two tracks, while keeping the road functional and dealing with MassDOT, Massport and CSX to coordinate removal of freight service, especially with planned freight traffic to Marine Terminal (as F-Line explained here and here)
• And last but not least, actually putting the station at Seaport - yes your proposals are awesome, but they still take work, and still miss Courthouse

Some are more doable than others, but a quick glance suggests that things may be even more problematic than the Albany Wye proposals, especially if my first proposal can actually work. If the main criticism of Albany Wye is that a short section on the map is engineeringly difficult in practice, the same applies to the jump between Albany/I-93 and Track 61, not to mention issues with Track 61 itself.

Yep, this is all fair, and in full disclosure looks at the details more closely than I have. The Track 61 notion hinges on it being a relatively easy add-on to an I-93 subway*, and if it's not, then that's that.

* I've been saying "I-93 subway" but I'm actually not sure it would even need to be a full-out subway. Need to flesh that out more.

On the other hand, an emerging theme from my exchange with @TheRatmeister above is: The Nubian-Seaport lag is likely the weakest quadrant in the whole Urban Ring and offers the least potential time savings, largely due to how the rapid transit network to the south is already going diagonally NE. I'm even inclined to argue that it doesn't need to be built at all - any southside circumferential route aiming to connect RL branches, Fairmount and Old Colony may be better off going to Andrew or JFK, while any bored tunnel from Logan (if it happens) is probably better off hitting South Station and Back Bay via the Huntington-Seaport subway.

Yes, I think this is really interesting. Perhaps a Nubian <> Seaport (<> Airport) leg is trying to do too many things at once. Because I think you're right: when you tease out the different objectives, Andrew and JFK make for stronger circumferential transfer points, and a cross-harbor tunnel would almost certainly be most effective if it can serve both South Station and Back Bay. (Although, to be fair, you could probably find a way to mix radial and circumferential services in that tunnel -- it doesn't have to be "either/or". But regardless, I think your larger point stands.)

 

[Teban54]

So, the big question on Nubian GL service is the eventual ridership of maxed out Fairmount and how much of the bus ridership from BHA, Morton etc is diverted to Jackson/Ruggles by BRTish service. Much of that ridership might go east to ride Fairmount as well. Nubians function as a mode junction might be greatly diminished in the longer term

So I looked at the 2015-17 passenger survey data for all passengers boarding SL4/5 at Nubian. Here's where they come from:

• From another bus: 2406 (72.6%)
• Walking or cycling: 740 (22.3%)
• Driving alone: 111 (3.4%)
• Others: 57 (1.7%)

Yes, Nubian's main role is as a bus transfer hub, but more than a quarter of passengers still start their transit ride at Nubian - which is good enough with me, and almost certainly suppressed due to poor service on SL4/5.

While Fairmount may indeed siphon off many folks from the 15 (Upham's Corner) and 23 (Four Corners/Geneva), I don't think it's a true replacement for rapid transit to Nubian even for bus riders. For one, Fairmount only gets you to South Station until NSRL, and even then it may not offer the best transfers to non-Red rapid transit lines that a rapid transit line via Tremont St subway can give you. Not to mention the walk to a Fairmount station, especially for those who live near a bus stop.

For another, Warren St up to the Grove Hall/Franklin Park vicinity isn't even close to Fairmount Line, and accounts for a significant chunk of bus ridership. Here's the 28:

The 23 has similar boardings north of their merge, not to mention 14 and 19. Not everyone who gets off the 28 at Nubian rode it all the way from Mattapan - far from it. As seen here, passenger turnover on the entire corridor is extremely high.

As of this report, 1640 inbound passengers per weekday get off at Nubian, while 1500 get off at Ruggles and 610 get off inbetween. (For reference, 1420 board along Warren St.) This indicates significant demand for true rapid transit service despite SL4/5 being available, and I don't think all of them will move over to Fairmount even with quality service there.

 

[Teban54]

Alternatively, perhaps you could use the airspace above the various highways to string together a dedicated ROW for a connection to Andrew:

View attachment 44779

On the other hand, it's not like Southampton St is a picturesque residential neighborhood, so if you're gonna do all the above, you could just build an elevated over Southampton and be done with it.

This actually got me thinking if fully elevated rail in the median of I-93 and Melnea Cass is feasible for a Nubian-downtown service. You'll certainly face a lot less NIMBYs than Newton. Tying back to Harold may be difficult, but you may even be able to build a true wye in the triangular region of I-90/I-93/Hudson St to tie directly into the Huntington subway, eliminating the need for a parallel subway under Harold (though the elevation change will be rough).

But yeah, this did convince me that an El for Andrew-BMC-(portal before Nubian?) may be quite feasible, via either your alignment or Southampton St, with accompanied Newmarket shift and an infill for Old Colony near Andrew.

• (The Old Colony infill is too close to JFK/UMass, but may not be a big issue with EMUs. It may also mess up with F-Line's Red X proposal, but that may be salvageable with some property takings west of the ROW to widen it.)

Re: whether "I-93 under" actually needs be a subway, I see two (minor) issues for surface running at street level:

• Grade crossings at Traveler St and E Berkeley St, requiring at least open cuts
• Elimination of a whole bunch of parking lots that span the whole I93-under today, which are all well-utilized

 

[TheRatmeister]

TL;DR: Eliminating street crossings and signal priority are hard to do when you have busy commercial corridors with many pedestrians, which are as desirable as transit for urban neighborhoods like South End.

I try not to pull this out much because it's generally not helpful, but I'm going to play the "But Europe" card here. Whether you go to Amsterdam or Milan, Munich or Budapest, you will find no shortage of wide streets and boulevards along important commercial corridors with no shortage of pedestrians cut in half by tramways, with almost all the minor streets not crossing the tracks. This does not mean pedestrians cannot cross the tracks, quite the opposite in fact. There are plenty of pedestrian crossings both at and in between stops, and even without them people are free to cross so long as there isn't any traffic. (Tram, car, or otherwise.)

If we limit the cross-street intersections along Washington St to Melnea Cass, Mass Ave, Newton St, Monsignor Renolds, E Berkeley St, and Herald St, for example, now TSP is looking much more feasible. To me this is another case of something that is in fact not difficult, so long as we choose not to put cars first in our street design.

 

[Riverside]

I try not to pull this out much because it's generally not helpful, but I'm going to play the "But Europe" card here. Whether you go to Amsterdam or Milan, Munich or Budapest, you will find no shortage of wide streets and boulevards along important commercial corridors with no shortage of pedestrians cut in half by tramways, with almost all the minor streets not crossing the tracks. This does not mean pedestrians cannot cross the tracks, quite the opposite in fact. There are plenty of pedestrian crossings both at and in between stops, and even without them people are free to cross so long as there isn't any traffic. (Tram, car, or otherwise.)

If we limit the cross-street intersections along Washington St to Melnea Cass, Mass Ave, Newton St, Monsignor Renolds, E Berkeley St, and Herald St, for example, now TSP is looking much more feasible. To me this is another case of something that is in fact not difficult, so long as we choose not to put cars first in our street design.

I am intrigued but skeptical. If you can show me a European system where trams like this can average 15-20 mph over the length of their journey, I will be more convinced.

My skepticism is that, fundamentally, moving a couple hundred people in a single 1.5 miles in less than 8 minutes seems at odds with an environment that mixes with pedestrians. Regardless of mode, if your ROW is unsealed, you need to be able to stop reasonably quickly, which caps your speed. It’s not (just) the cars — it’s the pedestrians.

Plus, even if we can create a European team on Washington St, you still have the competing needs of the 49 bus vs the El in terms of stop spacing. I’d need to do the math, but at first glance I am skeptical that you can achieve the El’s travel times while still providing adequate local service.

(Also, based on my limited understanding, those European trams exist in cities where whole swaths of the city have been designed to make cars unnecessary, not just a single corridor. I don’t think we’re anywhere close to that here, so even if the Washington St corridor became devoid of cars, there’d still be cars from other neighborhoods passing through.)

 

[737900er]

I am intrigued but skeptical. If you can show me a European system where trams like this can average 15-20 mph over the length of their journey, I will be more convinced.

My skepticism is that, fundamentally, moving a couple hundred people in a single 1.5 miles in less than 8 minutes seems at odds with an environment that mixes with pedestrians. Regardless of mode, if your ROW is unsealed, you need to be able to stop reasonably quickly, which caps your speed. It’s not (just) the cars — it’s the pedestrians.

Plus, even if we can create a European team on Washington St, you still have the competing needs of the 49 bus vs the El in terms of stop spacing. I’d need to do the math, but at first glance I am skeptical that you can achieve the El’s travel times while still providing adequate local service.

(Also, based on my limited understanding, those European trams exist in cities where whole swaths of the city have been designed to make cars unnecessary, not just a single corridor. I don’t think we’re anywhere close to that here, so even if the Washington St corridor became devoid of cars, there’d still be cars from other neighborhoods passing through.)

I was actually in Phoenix of all places a few weeks ago and this reminds me of their light rail design, which is mostly center-running. I think the whole system averages like 18mph, but they get it by also limiting pedestrians to only crossing at the major intersections (crossings every 1,200ft whereas the current B/C/E is 500-800ft) and having stations every mile. I think they're running at 35mph between stations too. Like you said, that's not really appropriate for this corridor.

 

[HenryAlan]

I was actually in Phoenix of all places a few weeks ago and this reminds me of their light rail design, which is mostly center-running. I think the whole system averages like 18mph, but they get it by also limiting pedestrians to only crossing at the major intersections (crossings every 1,200ft whereas the current B/C/E is 500-800ft) and having stations every mile. I think they're running at 35mph between stations too. Like you said, that's not really appropriate for this corridor.

I was also recently in Phoenix, and was fairly impressed by their bus system of all things (at least in the places I went). The thing to remember with Phoenix, though, is that almost none of it is instructive to a discussion about Boston. On the example regarding pedestrian crossings, they can get away with that because there are almost no pedestrians in the vast majority of locations. And the blocks are so large, that 1200 feet might be a typical distance between cross streets. We won't see that in many places here, though.

 

[TheRatmeister]

I am intrigued but skeptical. If you can show me a European system where trams like this can average 15-20 mph over the length of their journey, I will be more convinced.

My skepticism is that, fundamentally, moving a couple hundred people in a single 1.5 miles in less than 8 minutes seems at odds with an environment that mixes with pedestrians. Regardless of mode, if your ROW is unsealed, you need to be able to stop reasonably quickly, which caps your speed. It’s not (just) the cars — it’s the pedestrians.

Plus, even if we can create a European team on Washington St, you still have the competing needs of the 49 bus vs the El in terms of stop spacing. I’d need to do the math, but at first glance I am skeptical that you can achieve the El’s travel times while still providing adequate local service.

(Also, based on my limited understanding, those European trams exist in cities where whole swaths of the city have been designed to make cars unnecessary, not just a single corridor. I don’t think we’re anywhere close to that here, so even if the Washington St corridor became devoid of cars, there’d still be cars from other neighborhoods passing through.)

15-20mph is not realistic for a surface system, nor is matching the El's travel time of about 8 minutes, unfortunately. Unless we want to dig a subway, we're just not getting that. However, with an average speed of ~10MPH, which is definitely achievable, travel times from Nubian to Park St would be around 13 minutes, improved from around 20 minutes from Nubian-DTX on the Silver Line currently. And that's not even counting the fact that if you want to go somewhere other than Downtown Crossing you need to make an incredibly unweidly surface bus-subway transfer, as opposed to the quick GL-RL transfer at Park, or the slightly longer walk along the Winter St concourse to DTX. It's still 5 minutes slower than the El, but it also provides a more local service. The El only made 3 stops along this route, compared to 9 currently, or 7 with the cuts I propose in a couple paragraphs. I think it's a fair trade.

From some quick surface level research it seems like the systems of Amsterdam and Vienna both have an average speed of around 10MPH, including some mixed traffic segments. I would also challenge the notion that trams exist because most of the city is designed to make cars unnecessary. Having just gotten back from Milan, there's no shortage of cars on the main roads. You'll still see no shortage of cars besides median-running trams, the big difference is that you won't see them turning left onto or off of side streets.

The last thing I'll touch on is stop spacing and local service. To be honest, I think the stop spacing of the Silver Line isn't bad. (If it wasn't so damn slow) The table shows data from Fall 2022 of weekly boardings/alightings at all the SL4/5 stops. (Except Chinatown gate, for some reason) I've also included stop spacing data. Based on what's available here, I think Worcester Sq and Lenox St stops should be eliminated. This would bring the average stop spacing along the Washington St trunk from 320m to 410m, which I think is generally a good mix between the extremely tight stop spacing of bus services of ~200m and the much wider spacing of a subway line of around ~1000m. (For reference the stop spacing on the B and C branches is around 350m)

 

[737900er]

I was also recently in Phoenix, and was fairly impressed by their bus system of all things (at least in the places I went). The thing to remember with Phoenix, though, is that almost none of it is instructive to a discussion about Boston. On the example regarding pedestrian crossings, they can get away with that because there are almost no pedestrians in the vast majority of locations. And the blocks are so large, that 1200 feet might be a typical distance between cross streets. We won't see that in many places here, though.

Totally agree. I was just trying to make the point that to get the 15-20 mph needed it probably has to be built like that, but it's almost certainly not appropriate for the Washington Street corridor with stop spacing and pedestrian crossings considerations.

 

[Teban54]

15-20mph is not realistic for a surface system, nor is matching the El's travel time of about 8 minutes, unfortunately. Unless we want to dig a subway, we're just not getting that. However, with an average speed of ~10MPH, which is definitely achievable, travel times from Nubian to Park St would be around 13 minutes, improved from around 20 minutes from Nubian-DTX on the Silver Line currently. And that's not even counting the fact that if you want to go somewhere other than Downtown Crossing you need to make an incredibly unweidly surface bus-subway transfer, as opposed to the quick GL-RL transfer at Park, or the slightly longer walk along the Winter St concourse to DTX. It's still 5 minutes slower than the El, but it also provides a more local service. The El only made 3 stops along this route, compared to 9 currently, or 7 with the cuts I propose in a couple paragraphs. I think it's a fair trade.

I appreciate your analysis of current SL stops in the rest of the comment. That being said, I still think there needs to be a greater emphasis on demand from Nubian itself than the intermediate stops.

Nubian is one of the biggest bus hubs in the entire system, if not the biggest one. It's comparable to Harvard and Sullivan (and I'd argue it's more crucial than Ruggles). However, Harvard and Sullivan have 3-stop and 2-stop rides to the edge of the downtown core respectively, and Nubian used to have a 3-stop ride itself.

One can make a strong argument that Charlestown and Cambridge neighborhoods have local demands. But a hypothetical Red Line that makes 7 stops from Harvard to Charles/MGH will likely cause riots:

And that's before we consider that such a line needs to run in street medians, and slow down due to safety hazards or signals at pedestrian crossings (even with car crossings removed).

Your analysis does show nicely that SL Washington today has just as much demand at intermediate stations as Nubian, but I think that indicates demand for both "express" and "local" services, not a choice between them. Just like how Orange Line and the 93/T7 bus can co-exist, Red Line and the 1 bus can co-exist, and the C and D branches can co-exist to Reservoir. And not to forget that people boarding at Nubian had likely already endured a bus ride, so a seemingly small number like 13 minutes on this route can still mean a much longer ride overall. (All of this is before considering other beneficial factors, such as boosting capacity of Tremont St subway and serving BMC.)

Bottom line is, Nubian really shouldn't be treated any worse than Harvard and Sullivan. Especially given the social justice factors involved. A temporary solution to improve "the 49 bus" is very much needed, but it shouldn't be the ultimate goal.