Crazy Transit Pitches

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:
View attachment 44811
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.
I agree that a Washington St streetcar is not the solution for Nubian. But it doesn't exist in a vacuum. We've been discussing 3 projects give or take:
  • Nubian-Downtown Subway (And then to Longwood)
  • Washington St Light Rail
  • Nubian-Southie/Seaport Subway (Again, then to Longwood)
I don't think any one of these projects by itself is enough. Therefore, we move to picking the 2 optimal projects. Let's evaluate each pair:

Downtown Subway + Southie Subway
Pros:
  • It's all subway, so more capacity, full grade separation, etc.
  • Southie gets service out of this as well, good for them.
  • Good connection to the commuter rail and FiDi at South Station
Cons:
  • Cost, obviously
  • Downtown being served by a branch is... not great, to say the least. If frequency is very high (Like I think it should be) on the main section of the line west of Nubian this can be mitigated somewhat, but it's definitely not ideal
  • If the downtown subway only goes as far as South Station, the only interchange is with the RL and either the SL or part of the reconfigured GL. For places like Nubian where Roxbury Crossing is only 15 minutes away this isn't the worst thing in the world, but for the intermediate neighborhoods it would be pretty annoying.
  • Leaving out Washington St makes the walkshed of potential stations a fair bit worse.
  • Local service on Washington St still sucks
Downtown Subway + Washington St LRT
Pros:
  • Lower cost than the previous option
  • Maintains a rapid connection to commuter rail and the FiDi at South Station
  • Adds connection to the OL at Park St/DTX, potentially the BL as well if some trains are extended to Gov Center, or if a pedestrian passage is built to connect Gov Center and Park St.
Cons:
  • This proposal features a fair amount of parallel service where it's arguably not necessary. Stations would be less than 500m apart south of Union Park St, hurting ridership at both.
  • No service to Southie
Southie Subway + Washington St LRT

Pros:
  • Provides improved service to the widest area,
  • Maintains connections to rapid transit at Park St/DTX, and potentially to the BL with extended service/pedestrian tunnel to Gov Center.
  • Likely the lowest cost option
Cons:
  • Loses connection to the Old Colony/Fairmount Lines, potentially the Worcester Line as well. Could potentially be mitigated with connection of the subway to Kenmore and a new station at Widett Circle, but like you have said this might be difficult. Given that the area is about to be leveled and used for a new yard, however, I'm still not convinced this is really a non-starter though.
  • Loses direct connection to the FiDi. It's like, a 10 minute walk max, I'm not particularly concerned about this point.
Based on this, I think option 3 is the way to go. It provides the most extensive improved service area with cons I would personally consider to be the easiest to work around, in addition to likely being the lowest cost option. The missing connection to the Old Colony lines is a bummer, but at a minimum it's something that is fixible with money. The same isn't true about the missing connection to Southie (And ultimately, hopefully, under the harbor to Logan, Chelsea, etc), which as stated previously would probably require branching after Nubian, cutting service to South Station and making the connection generally less useful in the first place.
 
I agree that a Washington St streetcar is not the solution for Nubian. But it doesn't exist in a vacuum. We've been discussing 3 projects give or take:
  • Nubian-Downtown Subway (And then to Longwood)
  • Washington St Light Rail
  • Nubian-Southie/Seaport Subway (Again, then to Longwood)
I don't think any one of these projects by itself is enough. Therefore, we move to picking the 2 optimal projects. Let's evaluate each pair:

Downtown Subway + Southie Subway
Pros:
  • It's all subway, so more capacity, full grade separation, etc.
  • Southie gets service out of this as well, good for them.
  • Good connection to the commuter rail and FiDi at South Station
Cons:
  • Cost, obviously
  • Downtown being served by a branch is... not great, to say the least. If frequency is very high (Like I think it should be) on the main section of the line west of Nubian this can be mitigated somewhat, but it's definitely not ideal
  • If the downtown subway only goes as far as South Station, the only interchange is with the RL and either the SL or part of the reconfigured GL. For places like Nubian where Roxbury Crossing is only 15 minutes away this isn't the worst thing in the world, but for the intermediate neighborhoods it would be pretty annoying.
  • Leaving out Washington St makes the walkshed of potential stations a fair bit worse.
  • Local service on Washington St still sucks
Downtown Subway + Washington St LRT
Pros:
  • Lower cost than the previous option
  • Maintains a rapid connection to commuter rail and the FiDi at South Station
  • Adds connection to the OL at Park St/DTX, potentially the BL as well if some trains are extended to Gov Center, or if a pedestrian passage is built to connect Gov Center and Park St.
Cons:
  • This proposal features a fair amount of parallel service where it's arguably not necessary. Stations would be less than 500m apart south of Union Park St, hurting ridership at both.
  • No service to Southie
Southie Subway + Washington St LRT

Pros:
  • Provides improved service to the widest area,
  • Maintains connections to rapid transit at Park St/DTX, and potentially to the BL with extended service/pedestrian tunnel to Gov Center.
  • Likely the lowest cost option
Cons:
  • Loses connection to the Old Colony/Fairmount Lines, potentially the Worcester Line as well. Could potentially be mitigated with connection of the subway to Kenmore and a new station at Widett Circle, but like you have said this might be difficult. Given that the area is about to be leveled and used for a new yard, however, I'm still not convinced this is really a non-starter though.
  • Loses direct connection to the FiDi. It's like, a 10 minute walk max, I'm not particularly concerned about this point.
Based on this, I think option 3 is the way to go. It provides the most extensive improved service area with cons I would personally consider to be the easiest to work around, in addition to likely being the lowest cost option. The missing connection to the Old Colony lines is a bummer, but at a minimum it's something that is fixible with money. The same isn't true about the missing connection to Southie (And ultimately, hopefully, under the harbor to Logan, Chelsea, etc), which as stated previously would probably require branching after Nubian, cutting service to South Station and making the connection generally less useful in the first place.
I'll make more comments later, but two major disconnects between your analysis and what I'm envisioning:
  • I typically assume the Nubian-downtown subway feeds into the Pleasant St incline and thus Tremont St subway, to at least Government Center. I do NOT assume it feeds into South Station.
  • I assume a fully Green-Line-Reconfigured world (due to those proposals being more practical and having greated impacts than southside ring service). In this world, a Huntington-Seaport subway via South Station solves any issues of Regional Rail to LMA.
 
  • I typically assume the Nubian-downtown subway feeds into the Pleasant St incline and thus Tremont St subway, to at least Government Center. I do NOT assume it feeds into South Station.
I'm referring to your proposal of South Station-Nubian here, "Subway" meaning subway line, not subway as in a tunnel. (And definitely not the sandwich shop) Apologies if that was unclear.
I assume a fully Green-Line-Reconfigured world (due to those proposals being more practical and having greated impacts than southside ring service). In this world, a Huntington-Seaport subway via South Station solves any issues of Regional Rail to LMA.
Fair point. It's still a bummer for other areas along the line, but it definitely makes skipping south station (And any infill station at Widett Circle) much less of a problem. Depending on what kind of development ends up happening there given that it sounds like the air rights are going to be used, it might still be worth it, but that's a separate question at this point.
 
Last edited:
I'm referring to your proposal of South Station-Nubian here, "Subway" meaning subway line, not subway as in a tunnel. (And definitely not the sandwich shop) Apologies if that was unclear.
I think the confusion arose mostly from the goal of the Nubian subway (whose nomenclature is solely to distinguish from the Washington St streetcar). To be very clear, I think of the Nubian subway as primarily serving Park St, Government Center, and possibly points further north (which of course get you OL and BL). This entire discussion about Nubian-SS-Seaport and the Albany Wye is solely to examine a low-cost addition to also let the subway serve South Station and Seaport as an add-on - I was not aiming for South Station.

So in case it wasn't clear, I think of the Nubian subway (or El) as inevitable in the long term. That's regardless of any circumferential service - its #1 role is as a radial route. And as I mentioned, if we were to look at any circumferential route to the south, I would aim for Andrew/JFK and have the line just terminate there (or perhaps go further east to Southie, but not to Logan from there if that means skipping Seaport).

In general, I'm skeptical of filling transit deserts with circumferential routes only. Forcing Nubian riders to take a subway to any of Ruggles, Broadway or Andrew for a transfer, as the only way to downtown with reasonable speed, sounds like a bad idea. Boston is nowhere close to those highway cities with hollow downtowns, and that's also where you have the most convenient transfers to northside lines.

The same applies to Southie, if not more, given that their current demand is heavily oriented towards downtown and the east-west spread of the neighborhood. Not only does putting a grade-separated station east of Broadway/Andrew require a tunnel, but a single station is unable to cover the entire neighborhood, so many riders will continue to take the 9 bus to Broadway for a more direct Red Line connection (or the 7 for OSR to work) unless they're going southwest. I plan to do more analysis on the bus ridership patterns in South Boston, but I'd guess that an east-west surface LRT from City Point to Broadway (or Andrew, possibly further north/west) may serve the local demand better, much like the Washington St streetcar for South End.
  • I'll also stress that an Albany St/I-93 route is not at odds with connections to Southie. If most people from Southie will take buses/BRT/streetcar via the 9 bus's route, there will be a intersection between that and I-93 for transfers. It's also only a block from Broadway RL, so it should be pretty easy to extend the BRT/streetcar by one stop.
I'd personally update your pros/cons analysis to the following: (any "Andrew subway" has the option for one-stop extension into Southie, though I personally don't expect it to be the main Airport-bound service; it may also go to JFK/UMass instead)
Lines BuiltProsCons
Nubian Subway + Andrew Subway
  • Full grade separation for both Nubian-downtown and South Shore-LMA
  • Faster connection for South Shore-LMA (also Fairmount and Ashmont, I'm just using "South Shore" for simplicity)
  • Better connection to Southie local routes (best of the 3 proposals)
  • Highest cost
  • Lack of South End local service (can be mitigated with "BRT" like current SL4/5)
  • Hard to get political will to directly build two subways in practice, without an "initial build" of Washington streetcar
Washington Streetcar + Nubian Subway
  • Lower cost
  • Grade separation to Nubian, with maximum radial capacity
  • Local service to South End
  • Minor effort to connect to Southie route via E Berkeley/I-93 station
  • Reasonable effort to also get Nubian-SS-Seaport "bonus service" via Albany Wye
  • Parallel service, but they serve fundamentally different purposes
  • No improvements in South Shore-LMA
  • Connection to Southie is less ideal than the other two proposals
  • Political challenges in arguing for a Nubian subway with a streetcar in place
Washington Streetcar + Andrew Subway
  • Lower cost, potentially lowest
  • Faster connection for South Shore-LMA
  • Better connection to Southie local routes (only via Andrew)
  • Local service to South End
  • Politically most likely, if there's will for Urban Ring
  • No grade-separated radial service to Nubian, which severely constrains speed and capacity to a major transit hub
    • Grade-separated capacity to Nubian (eastward) is in a direction that's not very useful for Nubian riders
  • Southie connection only at Andrew (without additional tunneling), when the greatest density and demand are to Broadway
All of the above
  • Full grade separation for both Nubian and South Shore-LMA, with maximum radial capacity to Nubian
  • Faster connection for South Shore-LMA
  • Better connection to Southie local routes
  • Local service to South End
  • Reasonable effort to also get Nubian-SS-Seaport "bonus service" via Albany Wye
  • Minor effort to have an additional Southie connection via E Berkeley/I-93 station
  • $$$$$$$$$$
  • Politically least likely

All three options have benefits and drawbacks, so it really depends on which one you value more: Nubian, or South Shore-LMA plus a bit of Southie? Personally, I'd prioritize Nubian (especially in a world where many Southie residents still need to take buses regardless).

Additional minor comments on why I dropped some of your points:
  • Branching east of Nubian shouldn't be a problem. It should be fairly easy to build Nubian station as a terminal (middle track or bilevel station), to allow both through-running Ruggles-Nubian-Andrew and terminating Nubian-GC routes without interference.
  • Parallel service shouldn't be a concern if they have different goals and different stop spacing (not to mention adding capacity). Nobody in NYC complains that the 4/5 Lex Express and 6 Lex Local (or any other express/local lines) are parallel, or that having local stations on the 6 hurts 4/5's ridership. Likewise, nobody thinks the presence of either Orange Line south of Sullivan or the 93 bus negates the needs for the other.
  • For the South Shore connections, I specifically focused on LMA because that's by far the most important one. Connections to Nubian and BMC likely have less demand, and are close enough that buses may be fine. Connections west of LMA (like Kenmore and BU) are far enough that other alternative routings may work better.
Ultimately, I think that a purely circumferential connection east of Nubian should focus on South Shore connections as its #1 goal, instead of also expecting it to serve as a Nubian radial route. Southie would obviously benefit from a service going southwest (mostly to LMA), but that's more of a bonus, and still reasonably achievable with the Albany alignment. For this reason, I also excluded Broadway from consideration: it doesn't do well at either serving Nubian or shortening connections for South Shore, and it doesn't even go deeper into Southie than you'd expect.
 
Last edited:
^ says it better than I probably can. Is that going to stop me from fussing over a few details? Not at all.
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.
Yes, I want to dig a subway (or otherwise grade-separate). I want a streetcar down Washington St too, but I have no expectation that it could be load-bearing for transfers coming in from south of Nubian.
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
The travel time improvements exist almost entirely regardless of mode -- if we can get travel times on an LRT surface line down to 13 minutes, we can do the same with a BRT surface line. But either way, it's still not a replacement for the El, as @Teban54 lays out.
I agree that a Washington St streetcar is not the solution for Nubian. But it doesn't exist in a vacuum. We've been discussing 3 projects give or take:
  • Nubian-Downtown Subway (And then to Longwood)
  • Washington St Light Rail
  • Nubian-Southie/Seaport Subway (Again, then to Longwood)
I don't think any one of these projects by itself is enough. Therefore, we move to picking the 2 optimal projects. Let's evaluate each pair:
Yeah I might also make more comments later, but a couple of points:

You mentioned "overlapping" walksheds between a Washington St surface line and an I-93 subway. I think this isn't really true, in large part because an I-93 subway would have very few stops; I've been putting one around Mass Ave to serve BUMC, and @Teban54 favors a stop at E Berkeley (though I do not), and that's it. The stations determine walksheds, and my design explicitly excludes additional stations to avoid overlapping walksheds.

1700605103309.png


Downtown Subway + Southie Subway
...
Cons:
  • ...
  • Downtown being served by a branch is... not great, to say the least. If frequency is very high (Like I think it should be) on the main section of the line west of Nubian this can be mitigated somewhat, but it's definitely not ideal
  • If the downtown subway only goes as far as South Station, the only interchange is with the RL and either the SL or part of the reconfigured GL. For places like Nubian where Roxbury Crossing is only 15 minutes away this isn't the worst thing in the world, but for the intermediate neighborhoods it would be pretty annoying
I think Teban54 cleared this up, but any Downtown Subway we are discussing here would hook into the Green Line at the Pleasant St Portal. That means there wouldn't be a train leaving Nubian for Park St via I-93 every two minutes -- its frequencies will be more modest in order to integrate into the larger Green Line system. So there will be extra capacity along an I-93 subway, so if you want to hook in a branch to Andrew, Broadway, or South Station, that will be just fine.

(One reason in favor of an I-93 <> South Station connection would be to utilize the excess capacity north of South Station in the Seaport to send a second route down to Nubian -- that would allow for much higher combined frequencies between Nubian and downtown.)
Southie Subway + Washington St LRT

...
Cons:
  • Loses connection to the Old Colony/Fairmount Lines, potentially the Worcester Line as well. Could potentially be mitigated with connection of the subway to Kenmore and a new station at Widett Circle, but like you have said this might be difficult. Given that the area is about to be leveled and used for a new yard, however, I'm still not convinced this is really a non-starter though.
I'm really confused what you're talking about here. What alignment of a "Southie" Subway are you imagining here? The loss of connection to OC/Fairmount and Worcester -- is that because this alignment misses South Station? And how does Kenmore mitigate this?
 
I'm really confused what you're talking about here. What alignment of a "Southie" Subway are you imagining here? The loss of connection to OC/Fairmount and Worcester -- is that because this alignment misses South Station? And how does Kenmore mitigate this?
I'll also come back to all of this later, hopefully with a full plan since there's a lot of ideas being thrown around, a lot of confusion, and it's all just kind of a mess. But for now I'll just touch on this point: I'm imagining two options here (Whether or not these are mutually exclusive has already been mentioned, we'll leave that for another time.) One alignment follows either Albany or Melnea Cass and 93 to South Station, the other follows Melnea Cass to Widett Circle before either meeting the RL at Andrew or traveling north under the South Boston Bypass to meet the RL at Broadway and then terminate at Seaport. Both have pros and cons that have been discussed at length, again, for another time.

My basic point here is that if option 2 is selected, this new line doesn't connect to South Station, therefore missing out on the Old Colony Lines, the Fairmount Line, and the Framingham/Worcester Line. (The Franklin and Providence/Stoughton Lines, along with SCR Phase 2 if it happens all travel through Ruggles and the new line could meet them there.) Therefore for those travelling in from the South Shore looking to get to jobs at Longwood or to Nubian, this routing is not great. As @Teban54 has pointed out, the GL Reconfig means that you could get the GL from South Station to Longwood, it would just be slower by way of being the GL. If we want to make all these connections possible, an extension to Kenmore/Lansdowne would connect with the Worcester Line, and an infill station at Widett Circle would connect with the OC/Fairmount lines.
 
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
The idea of an El along this alignment is a really interesting one.

Linking into a Herald St Subway: probably unnecessary, assuming you can repurpose the spot where I've previously suggested putting a portal. Then you run in a viaduct over the open cut or over Herald St. But yes, the elevation changes are hairy. Thinking about this more, I think this would be challenging. You can either hook in to the Pleasant Street portal, or you can hook into the portal location linked in the strikethrough text. The problem with Pleasant Street is that I think you'd have to have at the very least a grade crossing at Oak St, if not also Marginal (and I don't know where you'd put the platform other than at surface level on Shawmut). And the problem with the other portal is that you'd need to rise from -3 under the Orange Line to... at least +1 (if over Albany) if not +2 (over 93). That's something like 72' feet of rise or more, needing to weave in between a number of overpasses.

Building a wye in the 90/93/Hudson triangle: this is super tempting, in part because it would create a pair of "perfect" Aldgate junctions:

1700606717303.png


Giving both Park St and South Station one-seat-rides to both Back Bay and Nubian.

That being said, I think the plot is too small to handle the two portals that you'd need, and I don't think there's enough space to also do a flying junction underneath Hudson. (Just using a very rough estimate of 240' horizontally to accommodate 1 "level" of elevation change, I don't think it quite fits.)

But I think the plot could hold a single portal, which brings us back to your interesting idea here:
The only way I could make this remotely work is by running the rail ROW alongside Albany St at the same level, while either putting Albany St (south of Herald) on a road diet, or modifying the southbound on-ramp to start later and/or be steeper. Terrible drawing here, not to scale:
Albany Wye 2.png

The point where Albany St starts rising is shown here. The idea is to let the rail ROW "slide out" from under I-93 to Albany north of that point, and move the on-ramp to start to its south. So it may be engineeringly feasible... But whether it's politically feasible to reduce road space for Albany and the on-ramp at this block is another question.
which potentially becomes more interesting if the I-93 Subway is actually an elevated over either 93 or Albany or both.
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.)
I'd be keen to see an analysis of Andrew vs JFK/UMass for this terminal. You could "simply" run an El all the way down 93 to JFK/UMass directly. That saves you from needing to add a Regional Rail infill (though I guess it almost certainly would be cheaper to build the infill rather than build nearly a mile of elevated).
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
Well, not that I ever particularly care to prioritize parking lots, but if they really want to keep them, it should be possible: the highway viaduct is well over 200 feet wide, and we'd only need 25-30' of it for tracks.

As for the grade crossings, yeah, that's where this idea potentially becomes much more expensive. Part of the appeal of "I-93 Under" is the possibility that some of it could genuinely require no tunneling at all, for about 2000 feet (.37 miles) between Herald St and Randolph St. Grade separating Traveler St and E Berkeley St would undercut that benefit. (Which doesn't make the idea DOA, of course, just a little less enticing.)

But now you've got me thinking about a "I-93 Over" alignment instead. It's definitely intriguing -- 93 and Melnea Cass are both wide enough corridors that a modern elevated could maybe work. But like I said above, I have trouble hooking in both at the northern end and at the southern end.

@TheRatmeister, just saw your reply -- I hadn't realized we were actually discussing a out-and-out subway to Andrew. That makes a bit more sense now, thank you!
 
Alright, here's a quick and dirty diagram to (roughly) show my thoughts. It's super late and I need to go to bed, but basically here's some bullet points:

  • This map makes a couple assumptions:
  1. The GL reconfig happens, with the Huntington Ave subway following I90 to Atlantic Ave, where it connects with the existing Transitway
  2. The Fairmount Line is rapid-transitified
  • The Washington St Tram is not intended to take passengers from all of south Boston. There are other lines and hubs better suited to that. The OL, Fairmount Line, and RL all have stations better suited to collecting local bus service from Roxbury and Dorchester and funneling it into downtown.
  • With this plan, my opinion is that an added connection between BUMC and South Station does not justify the costs.
  • The Washington St Streetcar uses the quad-track section of tunnel between Boylston and Park and the extra platforms at Boylston to remain almost entirely separate from the rest of the green line, avoiding delay contagion.
  • A pedestrian tunnel between Park St and Government Center is constructed to allow for transfers to the Blue Line.
  • The Gold Line/Urban ring here operates much more like a "pure" orbital route, being entirely grade-separated to enable frequent, automated service. The section from Nubian-Southie can almost certainly be done entirely elevated or C&C, the section from Nubian to Ruggles can maybe be done C&C, and the section from Ruggles-Cambridge would likely need to be bored.
  • Between the new orbital route and a reconfigured GL, a weaker, highly radial network with a few single points of failure has been turned into a stronger, more resilient mesh network.
  • Yes, there would be more 2 seat rides than might be considered ideal, but this structure also results in the mesh network mentioned above. High-quality transfers at Boylston and Widett Circle in particular would make the ride into downtown (relatively) seamless, despite the transfer.
  • I'm sure there's something I missed or something wrong with this map. Like I said, tired. Don't take it too seriously.
Alright, that's enough for tonight.
South Boston Metro.png
 
The idea of an El along this alignment is a really interesting one.

Linking into a Herald St Subway: probably unnecessary, assuming you can repurpose the spot where I've previously suggested putting a portal. Then you run in a viaduct over the open cut or over Herald St. But yes, the elevation changes are hairy. Thinking about this more, I think this would be challenging. You can either hook in to the Pleasant Street portal, or you can hook into the portal location linked in the strikethrough text. The problem with Pleasant Street is that I think you'd have to have at the very least a grade crossing at Oak St, if not also Marginal (and I don't know where you'd put the platform other than at surface level on Shawmut). And the problem with the other portal is that you'd need to rise from -3 under the Orange Line to... at least +1 (if over Albany) if not +2 (over 93). That's something like 72' feet of rise or more, needing to weave in between a number of overpasses.

Building a wye in the 90/93/Hudson triangle: this is super tempting, in part because it would create a pair of "perfect" Aldgate junctions:

View attachment 44835

Giving both Park St and South Station one-seat-rides to both Back Bay and Nubian.

That being said, I think the plot is too small to handle the two portals that you'd need, and I don't think there's enough space to also do a flying junction underneath Hudson. (Just using a very rough estimate of 240' horizontally to accommodate 1 "level" of elevation change, I don't think it quite fits.)

But I think the plot could hold a single portal, which brings us back to your interesting idea here:

which potentially becomes more interesting if the I-93 Subway is actually an elevated over either 93 or Albany or both.
Here's what I'm thinking of right now: Run the rail line on the western edge of I-93 (and/or over Albany), and delete the Herald St southbound on-ramp.

I'll start by saying that I don't really see the point of the Herald St on-ramp (someone else can correct me). Cars going southbound onto I-93 can do so at the next on-ramp at Frontage Rd (the interchange with Melnea Cass), and it's also a smooth ride down Albany St with just two intersections. Also, after further considerations since the time I made that proposal, I think trying to keep the on-ramp while also accommodating for an above-ground rail crossing over the Pike will be challenging (the ramp can only start south of E Berkeley if you want to give enough clearance to Traveler and E Berkeley).

With that said, suppose we have an El over Melnea Cass. Start by turning the El to the median between I-93 and the SB Melnea Cass off-ramp:
I-93 El 1.png

Then cross the off-ramp to land on the western edge of I-93, which is also where the highway rises from at-grade to a viaduct:
I-93 El 2.png

Getting space on the same level as I-93 can be a bit tricky. Looks like a double-tracked rail ROW requires 30-35' (Longfellow Bridge has 30', GLX has 35'). This section of I-93 is very wide with 5 lanes, but also has the off-ramp coming up to the south (and is not too long after the I-90 on-ramp from the east, by highway standards), so I'm not sure if a road diet is politically feasible.

Taking one lane plus the shoulder would give you enough space, but if you can't, then you're looking at taking space over Albany. Luckily, the place that's the most intrusive to Albany is also the most desolete section, only abutting the MBTA bus garage and an athletic field. So I feel cautiously optimistic about "widening" the I-93 viaduct here but for transit use.
I-93 El 3.png

Luckily, the part where you get to the apartments is also the part where you have more space available from the (former) Herald St on-ramp, so you'll only need minimal space over Albany, if any:
I-93 El 4.png


The problem comes... once you get to Herald St.

The "wye" part towards South Station is easy. You can take over the on-ramp's space (with some widening) to land on the same level as Albany St, continuing my proposal above. But turning towards Hudson/Marginal is very hard because of a Pike on-ramp, meaning you'll have to go down 2 levels in about 280' of straight-line distance (from Albany level to below Pike and the on-ramp). That sounds... impossible, even if you can somehow figure out how to cross Albany and the southbound Albany on-ramp to their west.

(The figure below is not drawn to scale, and the shape of the curve is not precise due to limitations with Paint 3D. Measurements are for 115', about the minimum outer radius. You'll get about 300-350' of length along this alignment, better than 280', but still not enough.)

I-93 El 5.png


So you'll basically have to split it to an above-ground branch next to Albany St, and a tunnel that turns to Herald St. There's 720' between Traveler St and Herald St, which is just enough for the ramp to rise from Albany level to leaving enough vertical clearance for Traveler St. If light rail can handle steeper grades than cars, this may be enough to let the rail ROW descend to surface level (Albany) first, then slide underneath I-93 and split into two, with one going underground and the other remaining at street level to cross the Pike:

(Each marker is 240'; haven't checked turning radius yet, tracks not to scale)
I-93 El 6.png

Seems plausible. However, this crucially relies on the grades actually being feasible, and even then, it will be a very steep climb.
 
Alright, here's a quick and dirty diagram to (roughly) show my thoughts. It's super late and I need to go to bed, but basically here's some bullet points:
I appreciate you drawing out the map, though I myself had inferred most of it in the prior exchanges.

I remain unconvinced that OL, RL and Fairmount are adequate replacements for quality transit to Nubian. The first two already exist, and don't seem to solve the issue. The last may reduce demand on several routes, but won't knock down Nubian enough to the point where it won't have 14 different bus routes through it, 7 of which have sub-15 min headways (no other bus hub has this many). As an analogy, I don't think GLX and future rapid transit to Everett/Chelsea along the Eastern Route reduce the need for Harvard and Sullivan to have a rapid connection downtown, nor do they justify turning Red and Orange lines into streetcars. Not to mention a notable amount of people do start their journey at Nubian.
  • If anything, this is reminiscent of the recent (resurfacing) debate about whether Regional Rail to Lynn negates the need for BLX.
There are still some misunderstandings about what GL Reconfiguration is exactly (there's usually a Huntington-Park-GC branch - refer to Riverside's recent map and this Bay Village interchange design), but that's minor, and I certainly don't fault you for that given how long it has taken this board to collectively come up with these proposals.

  • With this plan, my opinion is that an added connection between BUMC and South Station does not justify the costs.
At the risk of beating a dead horse... No, the added connection we're talking about isn't for South Station.

  • The Gold Line/Urban ring here operates much more like a "pure" orbital route, being entirely grade-separated to enable frequent, automated service. The section from Nubian-Southie can almost certainly be done entirely elevated or C&C, the section from Nubian to Ruggles can maybe be done C&C, and the section from Ruggles-Cambridge would likely need to be bored.
I have to say, the cost of full grade separation from Ruggles all the way to probably Sullivan (and definitely at least north of the Red Line) will likely far outweigh the cost of a Nubian-downtown grade-separated route.
 
Great analysis @Teban54. I need to check the handbook — if memory serves, you can do 7%, but only if the length of the incline doesn’t exceed… 500 feet, I think? In which case, you need a gentler grade.
 
I appreciate you drawing out the map, though I myself had inferred most of it in the prior exchanges.

I remain unconvinced that OL, RL and Fairmount are adequate replacements for quality transit to Nubian. The first two already exist, and don't seem to solve the issue. The last may reduce demand on several routes, but won't knock down Nubian enough to the point where it won't have 14 different bus routes through it, 7 of which have sub-15 min headways (no other bus hub has this many). Not to mention a notable amount of people do start their journey at Nubian.
Is Nubian a bus hub because it's the best place for a bus hub, and it will always be one, or is it a bus hub because... it just is? Is it really more suited to the role than Ruggles, Jackson Sq, or Roxbury Crossing, for example? Food for thought, if not super relevant. Anyways, a look into Nubian's most frequent bus routes:

  • 1: Some of its ridership would likely switch to using the gold line to BUMC or Cambridge, if it existed, but in general it would be less affected by the new lines.
  • 8: Provides additional service along Washington St, as well as connections to JFK, Ruggles, and Longwood. Would most likely lose significant ridership to the Gold Line
  • 15: This route already connects to the Fairmount Line at Upham's corner and the OL at Ruggles. Most passengers going downtown today ride to Ruggles, but in an ideal world Uphams corner would have frequent service as well.
  • 19: Same as the 23, but it actually meets the Fairmount Line at Four Corners/Geneva. Again, Ruggles/Fairmount Line for downtown travel
  • 23: Its southern section parallels the Fairmount Line quite a bit. With enhanced service it would likely take over a good chunk of the 23's ridership in this area. It also continues through Nubian to serve Ruggles, which is already the best way to connect to downtown along this route.
  • 41: Provides an important crosstown service, connecting to most lines in Dorchester and Roxbury. Passengers heading downtown already transfer at Jackson Sq or JFK, and with better Fairmount Line service Uphams corner as well.
  • 42: Provides a local service along Washington St. Most downtown trips probably start with walking a bit further to the OL already. Not likely to be replaced by a new rapid transit line.
  • 44: Primarily a local service along Walnut St and Humboldt Ave. Downtown passengers connect at Jackson Sq or Ruggles.
  • 45: Provides a local service paralleling the Fairmount line. With enhanced Fairmount Line service this route would likely bleed a significant amount of ridership. Again, downtown passengers generally connect at Ruggles.
  • 66: This route would almost certainly see huge ridership losses along the southern section in particular if the gold line existed
So, now let's look at the data. I've included boardings/alightings at Nubian for all these bus routes, as well as data for Ruggles where applicable since many of these routes also connect there.
Screenshot 2023-11-22 at 12.30.11.png

Of the routes that (frequently) stop at Nubian, only the 1, 15, 23, 42, and 66 break 1000 weekly boardings/alightings for 2022. Of these, the 42 is mainly a local service, and the 1 and 66 are likely to lose some and a lot of ridership to an orbital rail line respectively. This leaves the 15 and 23, which both also stop at Ruggles. As the data shows, there seems to be two large groups of riders using these routes: 1 group travels to Nubian to connect with with other buses, likely the 1 and 66. The other group stays on until Ruggles, likely connecting to the OL to travel downtown. It's unclear how many transfers are made from local buses to the Silver Line, but unfortunately this data is not collected so there isn't really a good way to check without spending a few days interviewing and counting people at Nubian.

But back to whether or not a Washington St LRT is suitable for the traffic volume. Let's assume weekly boardings at Nubian quadruple from 2022, just to give ourselves plenty of margin. That gives us around 22,500 boardings per week, or an average of around 3-4k per day. This is about in line with some of the busiest GL stops such as Harvard Ave, Coolidge Corner, Fenway, and Brookline Village, which recieve around 8TPH during peak hours. I don't think 12TPH during peak hours along a Washington St route is unreasonable, providing steps are taken to sufficiently isolate this route from the other GL routes. Again I think this is doable, given the quad-track between Boylston and Park and 4 platforms and 2 turnaround loops at Park St.
I have to say, the cost of full grade separation from Ruggles all the way to probably Sullivan (and definitely at least north of the Red Line) will likely far outweigh the cost of a Nubian-downtown grade-separated route.
Last note for now: This is a separate problem/discussion mainly about the best way to continue an urban ring into Cambridge. I don't really think it has much bearing on the Washington St LRT or subway to Downtown/Southie.
 
@TheRatmeister I feel bad but I think I'm mostly going to bow out of the discussion of the Washington St LRT line for the moment. I really appreciate the depth of your analysis and I think you are giving good food for thought. I have some other projects that I want to devote my (currently somewhat limited) "ArchBoston time" to, so I am mostly going to focus on those. I have a couple of small points I'd love to raise and get your thoughts on:
The GL reconfig happens, with the Huntington Ave subway following I90 to Atlantic Ave, where it connects with the existing Transitway
As @Teban54 alluded to, this is part, but not all, of the Green Line Reconfiguration. Sending all Huntington trains to South Station reduces connectivity to the Kenmore branches, the Blue Line, and the northern end of downtown, so my/our GLR design reverse-branches to provide direct connection from Huntington to both Park St and South Station. I don't think this radically impacts the main arguments you are making though.
A pedestrian tunnel between Park St and Government Center is constructed to allow for transfers to the Blue Line.
I'd be very interested in any ideas/sketches for this. Expanding the underground footprint between Park St and Government Center has looked appealing for decades, but threading a path under this oldest part of Boston has always been tricky. If you have ideas, would love to see them!
Is it really more suited to the role than Ruggles, Jackson Sq, or Roxbury Crossing, for example? Food for thought, if not super relevant.
Well, for one thing, Nubian is closer than Ruggles or Roxbury Crossing is, resulting in shorter routes. Of course, taken to its logical conclusion, that also points to bus hubs along the Fairmount Line -- which I agree are valuable and often underestimated, though I also think are insufficient to wholly negate Nubian's role.
 
I'd be very interested in any ideas/sketches for this. Expanding the underground footprint between Park St and Government Center has looked appealing for decades, but threading a path under this oldest part of Boston has always been tricky. If you have ideas, would love to see them!
I'm by no means an expert in this, I'm generally coming at these projects with a spatial planning perspective rather than a civil engineering perspective, so take basically everything I say that's more engineering focused with a grain of salt. That all being said, an underground connection probably needs to follow Tremont St. If the space between the GL tracks is wide enough perhaps some kind of passageway could fit there, though I'm not sure if this is the case, it doesn't seem super likely. The next-best option would probably be to go further down, likely below the RL/BL. Obviously from a cost perspective this is not ideal, but for a pedestrian tunnel it's not completely unworkable, the distance is quite short and it's not like this passageway needs to be massive. In terms of plausibility, this project would directly serve a bunch of politicians so I think it gets an immediate boost there.
 
I'll start by saying that I don't really see the point of the Herald St on-ramp (someone else can correct me)
I think, though am not at all sure, that it's actually there in part to collect surface traffic from north of the Mass Pike:

1700659781115.png


Looking more carefully at your proposal, one difference from what I was imagining is that you are suggesting an elevated that is at the same elevation as I-93, I think? I was imagining just running above it altogether. But again, that creates problems with the grade changes at the northern end.

I was thinking about whether the southern half of the Aldgate junction at the Albany Wye could go under Herald St. The problem with that -- and it's less of an engineering problem, more of a design elegance problem -- is that it makes the (rather expensive) under-Pike tunnel that I propose near Shawmut Ave somewhat redundant; if you're gonna go over the Pike for your northern portal, it's gonna seem silly/wasteful to build a second subway and second underpass for the western portal.

The whole thing looks very alluring but I just am not sure it can work.
 
Is Nubian a bus hub because it's the best place for a bus hub, and it will always be one, or is it a bus hub because... it just is? Is it really more suited to the role than Ruggles, Jackson Sq, or Roxbury Crossing, for example? Food for thought, if not super relevant. Anyways, a look into Nubian's most frequent bus routes:

  • 1: Some of its ridership would likely switch to using the gold line to BUMC or Cambridge, if it existed, but in general it would be less affected by the new lines.
  • 8: Provides additional service along Washington St, as well as connections to JFK, Ruggles, and Longwood. Would most likely lose significant ridership to the Gold Line
  • 15: This route already connects to the Fairmount Line at Upham's corner and the OL at Ruggles. Most passengers going downtown today ride to Ruggles, but in an ideal world Uphams corner would have frequent service as well.
  • 19: Same as the 23, but it actually meets the Fairmount Line at Four Corners/Geneva. Again, Ruggles/Fairmount Line for downtown travel
  • 23: Its southern section parallels the Fairmount Line quite a bit. With enhanced service it would likely take over a good chunk of the 23's ridership in this area. It also continues through Nubian to serve Ruggles, which is already the best way to connect to downtown along this route.
  • 41: Provides an important crosstown service, connecting to most lines in Dorchester and Roxbury. Passengers heading downtown already transfer at Jackson Sq or JFK, and with better Fairmount Line service Uphams corner as well.
  • 42: Provides a local service along Washington St. Most downtown trips probably start with walking a bit further to the OL already. Not likely to be replaced by a new rapid transit line.
  • 44: Primarily a local service along Walnut St and Humboldt Ave. Downtown passengers connect at Jackson Sq or Ruggles.
  • 45: Provides a local service paralleling the Fairmount line. With enhanced Fairmount Line service this route would likely bleed a significant amount of ridership. Again, downtown passengers generally connect at Ruggles.
  • 66: This route would almost certainly see huge ridership losses along the southern section in particular if the gold line existed
So, now let's look at the data. I've included boardings/alightings at Nubian for all these bus routes, as well as data for Ruggles where applicable since many of these routes also connect there.
View attachment 44846
Of the routes that (frequently) stop at Nubian, only the 1, 15, 23, 42, and 66 break 1000 weekly boardings/alightings for 2022. Of these, the 42 is mainly a local service, and the 1 and 66 are likely to lose some and a lot of ridership to an orbital rail line respectively. This leaves the 15 and 23, which both also stop at Ruggles. As the data shows, there seems to be two large groups of riders using these routes: 1 group travels to Nubian to connect with with other buses, likely the 1 and 66. The other group stays on until Ruggles, likely connecting to the OL to travel downtown. It's unclear how many transfers are made from local buses to the Silver Line, but unfortunately this data is not collected so there isn't really a good way to check without spending a few days interviewing and counting people at Nubian.

But back to whether or not a Washington St LRT is suitable for the traffic volume. Let's assume weekly boardings at Nubian quadruple from 2022, just to give ourselves plenty of margin. That gives us around 22,500 boardings per week, or an average of around 3-4k per day. This is about in line with some of the busiest GL stops such as Harvard Ave, Coolidge Corner, Fenway, and Brookline Village, which recieve around 8TPH during peak hours. I don't think 12TPH during peak hours along a Washington St route is unreasonable, providing steps are taken to sufficiently isolate this route from the other GL routes. Again I think this is doable, given the quad-track between Boylston and Park and 4 platforms and 2 turnaround loops at Park St.

Last note for now: This is a separate problem/discussion mainly about the best way to continue an urban ring into Cambridge. I don't really think it has much bearing on the Washington St LRT or subway to Downtown/Southie.
This is a truly excellent analysis. I'm genuinely sorry if it sounds like I'm shutting down everything you write, but here are some comments:

The most glaring issue (other than the omission of the 28) is - I think there has to be something wrong in your methodology, because I just did the analysis myself and came up with completely different numbers that's much higher:
Nubian and Ruggles bus riderships Fall22.png

Color code: Red = major radial routes (above 3000/week), Blue = minor radial routes, Green = crosstown routes. Ruggles routes that don't go to Nubian are not shown.

I used the bus ridership data in the MBTA dataset "MBTA Bus Ridership by Time Period, Season, Route/Line, and Stop", which groups ridership by time periods like early morning, AM peak, etc. Note there's also a similar dataset named "MBTA Bus Ridership by Trip, Season, Route/Line, and Stop" which seems to break it down to each individual trip on each route, which I didn't use.

According to the guide, the second dataset's "average_ons", "average_offs" and "average_load" refer to average per vehicle, not average per day; so I assume the first dataset does something similar, and get "total_ons", "total_offs" by multiplying the average with the number of trips recorded during each period. As a practical example, the "average_ons" for route 28 at Ruggles on Saturday (where there's only one time period, OFF_PEAK) is 7.885, but it's hard to believe that the entire route only loads 7.885 passengers from Ruggles all day. Multiplying this by the 100 trips per Saturday gives 788.5 boardings at Ruggles, which is much more realistic.

I sum over all weekday time periods to obtain total ons and offs per weekday, then multiply it by 5 and add the figures on Saturday and Sunday. Direction 0 is outbound, and direction 1 is inbound. As a sanity check, my figures are in the same ballpark as the MBTA Bus Route Profiles (MBTA's numbers for the 23 have 1400 inbound alightings at Nubian per weekday pre-Covid, while mine has 1270.8 and yours seems to be far below that.)

Let's do the same for Harvard: (includes the Harvard busway, the 1's stop at Mass Ave @ Holyoke St, the 66/68/69's stop at Johnston Gate and Dawes Island, and Eliot St @ Bennett St)
Harvard bus riderships Fall22.png


And Sullivan:
Sullivan bus riderships Fall22.png


Both Harvard's and Sullivan's passenger volumes are only half of Nubian's (and that's before we consider people getting off at Ruggles). I think the numbers speak for themselves. I'd also say they have roughly similar compositions of local and crosstown routes, in both numbers of routes and ridership, and both Harvard and Sullivan will be impacted by a circumferential subway route.

Now let's compare Nubian's updated numbers (Fall22) with Coolidge Corner et al (Fall19, the most recent rail ridership data, where "average_ons" and "average_offs" are per day):
Nubian vs subway stations.png


The passenger volume at Nubian alone is 10 times that of each major GL surface station, and approaches the scale of subway stations with major bus hubs like Kenmore, Ruggles, Sullivan and Maverick. Even if you exclude crosstown Nubian buses, it's still greater than all 4 GL surface stations combined - especially if you also consider that many bus riders take them to Ruggles today but would probably transfer to a Nubian subway if it exists.

Even the Silver Lines alone, with their terrible conditions today, still have 2-4 times ridership from Nubian than each of the GL surface stations. That's not considering additional passengers they'll pick up in South End, further crowding the buses.

So if you assume only 50% of the Nubian bus riders will take the Nubian subway (due to transfers to crosstown routes, Fairmount, etc), it still gives you an important station in the entire network. That's not even considering induced demand - almost as many bus riders get off at Ruggles as they do at Nubian, likely for the rapid transit connection, and they will likely take the Nubian subway. The faster, more reliable ride will also encourage mode shift.
 
Last edited:
(continued from above)

One last main point about Nubian that I want to address:
Is Nubian a bus hub because it's the best place for a bus hub, and it will always be one, or is it a bus hub because... it just is? Is it really more suited to the role than Ruggles, Jackson Sq, or Roxbury Crossing, for example? Food for thought, if not super relevant.
If anything, I'd say Nubian is the opposite of what you're saying. It's the natural bifurcation point between two important bus corridors, the 15 and the 23/28, and also combines several other routes that radiate out to different directions (1, 66, 42, etc).

On the other hand, what you're describing sounds exactly like Ruggles and Wonderland. Half of the bus routes at Ruggles are really Nubian routes, but were extended to Ruggles in 1987 solely to provide the rapid transit connection (with additional reorganizations - can you believe the 28 didn't exist until 1987, and was only launched as a supplement to the 29???), the other half are really LMA routes, and then minor exceptions like 22 and 43. Ruggles is the posterboy of a "man-made" bus hub just because of the rapid transit connection, and offering high-quality transit to Nubian would greatly free up resources from buses running between Nubian and Ruggles. Similar conclusions apply to Wonderland, where half of the bus routes would have loved to terminate at Lynn.

--------------------------------

(not as related to above)

Last note for now: This is a separate problem/discussion mainly about the best way to continue an urban ring into Cambridge. I don't really think it has much bearing on the Washington St LRT or subway to Downtown/Southie.
While the cost of a deep bored subway from LMA to Cambridge is indeed a separate problem, if we assume C&C is cheaper than TBM (which may or may not be true), then consideration of such a long deep-bored tunnel would indicate a lower degree of emphasis on cost and practicality. This, in turns, would imply that a heated, one-or-the-other debate between Nubian-downtown subway and Nubian-Andrew subway is less necessary, as such a debate only arises due to budget concerns.

Well, for one thing, Nubian is closer than Ruggles or Roxbury Crossing is, resulting in shorter routes. Of course, taken to its logical conclusion, that also points to bus hubs along the Fairmount Line -- which I agree are valuable and often underestimated, though I also think are insufficient to wholly negate Nubian's role.
The issue with relocating bus hubs to the Fairmount Line is that riders want to go to Warren St. The initial BNRD proposal eliminated the 44 and 45 routes, and had the 14 and 19 replace them, skipping Warren St. However, community opposition made the T drop this plan and reintroduced the 44 and 45 in the revised proposal.

As I mentioned above, if riders only care about rapid transit connection, the 29 would have been the winner in the post-SW-Corridor era instead of the "supplemental" 28. I'm inclined to think that most (if not all) Dorchester routes have high turnovers due to local demands, and that's not something that can be easily solved with Fairmount-related integrations.

But I do agree that Fairmount Line can use better bus connections. I had considered it myself, but the conclusion seems to be that what we have today (and BNRD) is good enough - Newmarket has T8/10/41, Uphams Corner has T15, Four Corners has 19/T23 (and maybe T16), Talbot Ave has T22, Morton St has 21/26, and Blue Hill Ave has T28/29/30/T31. A few changes can be made, but given that a lot of the BNRD plans (26, 29) in the Mattapan area were reverted due to community feedback, I'm not sure about feasibility.

I think, though am not at all sure, that it's actually there in part to collect surface traffic from north of the Mass Pike:
Yes that's what I noticed too, but such traffic can easily continue on Albany St and Frontage Rd (with relatively few intersections) until they get to the Melnea Cass on-ramp. If anything, it may create fewer merging conflicts on I-93. (Edit: there's also an on-ramp at Lincoln St that eventually merges into I-93 south of Melnea Cass, so it can also handle traffic north of the Pike.)

Looking more carefully at your proposal, one difference from what I was imagining is that you are suggesting an elevated that is at the same elevation as I-93, I think? I was imagining just running above it altogether. But again, that creates problems with the grade changes at the northern end.
Yes, my proposal is at the same level as I-93, or more of an "Albany St El". This is primarily due to grade changes to the north. It should also be feasible to run an El above I-93 until you get to E Berkeley, but that would be a relatively short "above I-93" section and I'm not sure if it'd be worth it.

I was thinking about whether the southern half of the Aldgate junction at the Albany Wye could go under Herald St. The problem with that -- and it's less of an engineering problem, more of a design elegance problem -- is that it makes the (rather expensive) under-Pike tunnel that I propose near Shawmut Ave somewhat redundant; if you're gonna go over the Pike for your northern portal, it's gonna seem silly/wasteful to build a second subway and second underpass for the western portal.

The whole thing looks very alluring but I just am not sure it can work.
While the Herald St subway (with the tunnel under OL) seems redundant, I'm not sure if there's really a better option. Parcels over the Pike will probably be redeveloped (not sure if Albany St is affected, but I do remember one of the parcels further west was to be made a park over Pike), so I don't think running an El over the Pike for any section is feasible. Even if you could, there's no good space for a portal, as you have three close intersections (Shawmut, Washington, Harrison), one of which you'll almost certainly have to cross at grade. Starting your descent west of Shawmut is way too close to the Bay Village interchange.

An El over Herald St faces the same problem - grade crossing with at least one of the north-south streets.
 
Last edited:
This is a truly excellent analysis. I'm genuinely sorry if it sounds like I'm shutting down everything you write, but here are some comments:

The most glaring issue (other than the omission of the 28) is - I think there has to be something wrong in your methodology, because I just did the analysis myself and came up with completely different numbers that's much higher:
View attachment 44857
Color code: Red = major radial routes (above 3000/week), Blue = minor radial routes, Green = crosstown routes. Ruggles routes that don't go to Nubian are not shown.

I used the bus ridership data in the MBTA dataset "MBTA Bus Ridership by Time Period, Season, Route/Line, and Stop", which groups ridership by time periods like early morning, AM peak, etc. Note there's also a similar dataset named "MBTA Bus Ridership by Trip, Season, Route/Line, and Stop" which seems to break it down to each individual trip on each route, which I didn't use.

According to the guide, the second dataset's "average_ons", "average_offs" and "average_load" refer to average per vehicle, not average per day; so I assume the first dataset does something similar, and get "total_ons", "total_offs" by multiplying the average with the number of trips recorded during each period. As a practical example, the "average_ons" for route 28 at Ruggles on Saturday (where there's only one time period, OFF_PEAK) is 7.885, but it's hard to believe that the entire route only loads 7.885 passengers from Ruggles all day. Multiplying this by the 100 trips per Saturday gives 788.5 boardings at Ruggles, which is much more realistic.

I sum over all weekday time periods to obtain total ons and offs per weekday, then multiply it by 5 and add the figures on Saturday and Sunday. Direction 0 is outbound, and direction 1 is inbound. As a sanity check, my figures are in the same ballpark as the MBTA Bus Route Profiles (MBTA's numbers for the 23 have 1400 inbound alightings at Nubian per weekday pre-Covid, while mine has 1270.8 and yours seems to be far below that.)

Let's do the same for Harvard: (includes the Harvard busway, the 1's stop at Mass Ave @ Holyoke St, the 66/68/69's stop at Johnston Gate and Dawes Island, and Eliot St @ Bennett St)
View attachment 44858

And Sullivan:
View attachment 44859

Both Harvard's and Sullivan's passenger volumes are only half of Nubian's (and that's before we consider people getting off at Ruggles). I think the numbers speak for themselves. I'd also say they have roughly similar compositions of local and crosstown routes, in both numbers of routes and ridership, and both Harvard and Sullivan will be impacted by a circumferential subway route.

Now let's compare Nubian's updated numbers (Fall22) with Coolidge Corner et al (Fall19, the most recent rail ridership data, where "average_ons" and "average_offs" are per day):
View attachment 44861

The passenger volume at Nubian alone is 10 times that of each major GL surface station, and approaches the scale of subway stations with major bus hubs like Kenmore, Ruggles, Sullivan and Maverick. Even if you exclude crosstown Nubian buses, it's still greater than all 4 GL surface stations combined - especially if you also consider that many bus riders take them to Ruggles today but would probably transfer to a Nubian subway if it exists.

Even the Silver Lines alone, with their terrible conditions today, still have 2-4 times ridership from Nubian than each of the GL surface stations. That's not considering additional passengers they'll pick up in South End, further crowding the buses.

So if you assume only 50% of the Nubian bus riders will take the Nubian subway (due to transfers to crosstown routes, Fairmount, etc), it still gives you an important station in the entire network. That's not even considering induced demand - almost as many bus riders get off at Ruggles as they do at Nubian, likely for the rapid transit connection, and they will likely take the Nubian subway. The faster, more reliable ride will also encourage mode shift.
You're right, I have indeed goofed. I've only added the weekday numbers once rather than 5 times. I'll redo the analysis and then come back with new thoughts.
 
Screenshot 2023-11-23 at 07.28.02.png
Screenshot 2023-11-23 at 07.40.29.png

Here's the fixed data for SL4/5 and Nubian/Ruggles to start off with. Anyways, a big focus for bus routes at Nubian is the 15, 23, and 28 which all mainly serve Dorchester and Mattapan. From the data it seems like many of these riders end up connecting onto the Silver Line, although that is speculation since unfortunately transfers aren't recorded in any way. An LRT replacing SL4/5 would probably struggle, especially during peak hours, to handle this demand.

But maybe it doesn't have to.
Screenshot 2023-11-23 at 08.06.50.png

It seems like of the 3 bus routes mentioned, around 50-55% of their boardings/alightings come from stops within a 10 minute walk of existing Fairmount Line stations. (When you exclude Nubian/Ruggles) It wouldn't be 100%, but with a proper service a very large chunk of these riders would likely take the Fairmount Line either downtown, or to connect with an orbital line that heads to Longwood/Cambridge at Widett Circle given that the 1 and 66 make up most of Nubian's remaining local bus ridership. With that load (mostly) taken off, I think the LRT would fare much better, and serve mainly demand along the Washington St corridor itself rather than connecting passengers from Dorchester at Nubian. This is a pretty soft conclusion though, without numbers to determine how many SL4/5 riders transfer onto the Dorchester buses it's pretty hard to give a definitive answer, and as far as I know this data doesn't currently exist.
 
Last edited:
View attachment 44887View attachment 44889
Here's the fixed data for SL4/5 and Nubian/Ruggles to start off with. Anyways, a big focus for bus routes at Nubian is the 15, 23, and 28 which all mainly serve Dorchester and Mattapan. From the data it seems like many of these riders end up connecting onto the Silver Line, although that is speculation since unfortunately transfers aren't recorded in any way. An LRT replacing SL4/5 would probably struggle, especially during peak hours, to handle this demand.

But maybe it doesn't have to.
View attachment 44891
It seems like of the 3 bus routes mentioned, around 50-55% of their boardings/alightings come from stops within a 10 minute walk of existing Fairmount Line stations. (When you exclude Nubian/Ruggles) It wouldn't be 100%, but with a proper service a very large chunk of these riders would likely take the Fairmount Line either downtown, or to connect with an orbital line that heads to Longwood/Cambridge at Widett Circle given that the 1 and 66 make up most of Nubian's remaining local bus ridership. With that load (mostly) taken off, I think the LRT would fare much better, and serve mainly demand along the Washington St corridor itself rather than connecting passengers from Dorchester at Nubian. This is a pretty soft conclusion though, without numbers to determine how many SL4/5 riders transfer onto the Dorchester buses it's pretty hard to give a definitive answer, and as far as I know this data doesn't currently exist.
Great analysis!

I intend to revisit this with more detailed followups in a few weeks once I get a chance to look at the data, but a few quick conjectures based on my current assumptions:
  • There actually is data on how many SL4/5 riders boarding at Nubian transfer from other buses, as I mentioned here. Those are less reliable than the ridership data as they're based on surveys with sample sizes in the 200-300 range, but looks like 73% of them come from other buses. Not only does that indicate demand from Nubian itself that's still notable (27%), but it shows that lack of a quick rapid transit connection is really holding the corridor back, as the numeber of riders on Nubian radial buses that ride to Ruggles is 2x the number of SL5 riders that board at Nubian after transferring from another bus.
  • I don't think it's reasonable to make the assumption that most or all riders currently boarding at stops near Fairmount stations alight at Nubian. Like I said in the linked comment, turnover on many Nubian routes is very high, much more so than your average radial route. It's very possible, if not likely, that a significant chunk or riders boarding near Uphams Corner, Four Corners and especially BHA stations get off at or before Warren St, while a new group of riders board from Grove Hall and transfer at Nubian. The latter will not be served by Fairmount.
  • Likewise, I think the impacts of a new rapid transit route (Fairmount) on nearby bus routes may be overestimated. Delvin did an analysis of ridership of Somerville/Medford buses after GLX opened: Only the 80 really plummeted, but it literally follows GLX at every stop (and it still has half the ridership). Even the 88 and 90, which run very closely to GLX, didn't suffer that much, while routes that cross a single GLX station (86, 87, 89, 91, 94, 96) and parallel routes (87, 89, 101) didn't see much noticeable dent at all. I'll remark that the 87's and 101's corridors are roughly as far from GLX Medford/Tufts branch as the 28 is from Fairmount, and the BNRD even makes both of them 15-min corridors (87 west of Union Square), despite running parallel to GLX. So I think there will still be significant demand to Nubian, even after improvements on Fairmount.
  • Lastly, there are social justice issues at hand that go beyond just ridership numbers. More than 3/4 of riders boarding SL4/5 at Nubian are persons of color. And this is a corridor that had rapid transit taken away and was slapped with a bus for the promised "equal or better" service. Given much higher ridership than the GL surface branches (Nubian is simply not comparable to Harvard Ave regardless of how many riders you siphon off via Fairmount), I don't think giving them equal treatment as the GL branches (which would have been bustituted in an alternative universe) is justified.
 
Last edited:

Back
Top