Green Line Reconfiguration

[TheRatmeister]

View attachment Green Line Reconfiguration Services Diagram Criss Cross v1.1.png

I'd like to revisit this map. What if as a modification to this, the crossover point is at Boylston instead? And one station or two, I'm not sure how much it necessarily matters but we'll say one. That still gets you the connections to the Huntington Ave Subway, which would connect to the Tremont St Subway using the same Marginal Road or Stuart St connections as the tripod model, and to the Orange Line. This wouldn't use every platform at Boylston to its fullest extent, but it's hardly unimaginable that some Boylston St Subway trains would be better using the Park St or Gov't Center loops rather than continuing to Design Center. The one this this is really missing is the OSR from Huntington to Seaport however, and I'm not sure there's an easy way to add that back in. A flying junction at Boylston seems easiest but space there is very constrained so I'm not sure that's possible.

(And again, this assumes that the Essex St route works, which with no portals I think it probably does.)

 

[BosMaineiac]

I'm glad we're revisiting the Essex St subway route, IMO the superior routing for a Seaport light rail connection, albeit with clear engineering challenges. Here's how I think we can achieve a crossover point at Boylston, with inspiration from Riverside's post and SLPhase3 plans on the last page, broken down into phases of construction. As a prereq, let's assume Washington St has a surface LRT to Nubian, rising at Elliot Norton Park.

Phase 1: Marginal Rd subway from Huntington Ave @ Ring Rd to the existing Pleasant St portal, with two underground stops at Back Bay and Bay Village

Phase 2: Essex St subway roughly mirroring a combination of the Post Office Sq extenion and SLPhase3 plans. As F-Line mentioned, I agree a stacked platform station between Boylston and Chinatown will most likely be required for this.

Phase 3: Flying junction at the Marginal Rd subway to a Columbus Ave subway. I'm not sure if a station at Stuart/Arlington is necessary but you could have one there. Next, underpin the existing Boylston subway and link up with the Essex St subway inclines for a straight shot to South Station.

I don't have time right now but later I'll mark up a PDF to create some visuals.

 

[Riverside]

(Now, I must admit that I don't really believe that the T is actually hitting 30 tph through Park St right now [although I guess this is something I could verify empirically], so it may not be possible to actually reach 40 tph anymore. But I think that's all the more reason in favor of Criss Cross over Tripod.)

Well, serves me right for commenting without checking data: I grabbed a random day from the Transit Matters headways/departures data (March 5) and, yup, during the morning peak there were 30 trains per hour leaving Park Street going to Government Center. So, yes, the T is actually hitting 30 tph on the Outer Tracks, and I shouldn't have speculated otherwise.

An implicit assumption of your "4,000 vs. 7,800" is that you get the Marginal Rd subway for free (Back Bay - the Highway Interchange District), or that it's already built.

Right, to be clear, that is my point: I'm arguing that tunneling under Marginal Road is significantly less disruptive to the community than Stuart/Kneeland would be. It's "4,000 feet of disruption vs 7,800 feet of disruption". I agree that Stuart/Kneeland is worth evaluating, but I think Marginal Road will come out ahead in most analyses, in particular because:

• Less disruption to neighborhoods
• Less "uncertain" tunneling due to proximity to Pike
  • @F-Line to Dudley if you have a citation for Marginal Road being "clean", that'd be awesome
• Due to the above, seems very likely to be the only possible alignment that might be seen as low-risk-enough to do cut-and-cover (major expense savings)

The route feels more "central" than following the Pike. Obviously it depends on station locations, but the entire Stuart-Kneeland alignment is much closer to jobs, shops, restaurants and hospitals than Marginal Rd.

I know what you mean here, but I'm always pretty skeptical when this line of reasoning is brought up. The tunnel's proximity to the "Main St" of a neighborhood doesn't matter: it's the station locations that matter. In this case, the only possible intermediate station would be some counterpart to the current Arlington station -- such as at Arlington St itself, or at Bay Village. Realistically, there are two potential station locations: something near Park Plaza, and something near the Pleasant St Portal. Park Plaza is obviously close to stuff, but it's also quite close to Arlington. Bay Village at least fills a gap in coverage (including providing better access to the South End on the other side of the Pike).

There's a couple of these; they're identically titled and dated, but are in fact different.

Thanks! I'll have to take a look.

I'd like to revisit this map. What if as a modification to this, the crossover point is at Boylston instead? And one station or two, I'm not sure how much it necessarily matters but we'll say one. That still gets you the connections to the Huntington Ave Subway, which would connect to the Tremont St Subway using the same Marginal Road or Stuart St connections as the tripod model, and to the Orange Line. This wouldn't use every platform at Boylston to its fullest extent, but it's hardly unimaginable that some Boylston St Subway trains would be better using the Park St or Gov't Center loops rather than continuing to Design Center. The one this this is really missing is the OSR from Huntington to Seaport however, and I'm not sure there's an easy way to add that back in. A flying junction at Boylston seems easiest but space there is very constrained so I'm not sure that's possible.

(And again, this assumes that the Essex St route works, which with no portals I think it probably does.)

I'm glad we're revisiting the Essex St subway route, IMO the superior routing for a Seaport light rail connection, albeit with clear engineering challenges. Here's how I think we can achieve a crossover point at Boylston, with inspiration from Riverside's post and SLPhase3 plans on the last page, broken down into phases of construction. As a prereq, let's assume Washington St has a surface LRT to Nubian, rising at Elliot Norton Park.

Phase 1: Marginal Rd subway from Huntington Ave @ Ring Rd to the existing Pleasant St portal, with two underground stops at Back Bay and Bay Village

Phase 2: Essex St subway roughly mirroring a combination of the Post Office Sq extenion and SLPhase3 plans. As F-Line mentioned, I agree a stacked platform station between Boylston and Chinatown will most likely be required for this.

Phase 3: Flying junction at the Marginal Rd subway to a Columbus Ave subway. I'm not sure if a station at Stuart/Arlington is necessary but you could have one there. Next, underpin the existing Boylston subway and link up with the Essex St subway inclines for a straight shot to South Station.

I don't have time right now but later I'll mark up a PDF to create some visuals.

Yeah, I mean, I myself am ultimately not in favor of the Full Criss Cross model; if it were really just four services, it might be doable, but each "corner" of the Criss Cross (Kenmore, Prudential, Boylston, Bay Village) would have multiple services of their own, plus the Nubian <> Downtown services creating yet another leg. That seems like an incredibly tall order. (If the whole system were automated, that could work, but obviously that won't happen here.) But, I do see its appeal, so I definitely see the value in gaming out ways to do it.

I was an early "Essex Stan", but at this point I do genuinely believe that the Marginal Road alignment is better (and not just for being cheaper/easier); I think I've pretty much laid out the reasons why, so I won't belabor that point.

But the only thing I do want to highlight, as folks consider how to achieve the Full Criss Cross: improbably, downtown Boston has not one but two fully intact flying junctions that sit unused. No matter how you cut it, flying junctions are harder and more expensive to build than simple flat tunnels. Designing around those, ensuring that both get used, seems like a good pay to supercharge any cost-benefit analysis.

(I think BosMaineaic's suggestion does end up using both? But I'm not clear what happens on Boylston between Charles on Tremont?)

 

[TheRatmeister]

Just to go way back to the original premise of this discussion, the capacity of the Park Street loop, I'm not sure that's actually unknown. It seems like, at least for a 2 week period, 3 branches, so about 20-22 TPH, used the Park Street loop continuously. (I believe the D was turned at Kenmore during that time.) That's definitely enough for the B/C branches, probably plus a restored A branch as well since that would be okay cannibalizing the B to some extent. That is to say, I think the concerns about tripod being unsustainable are mostly unfounded, and the Park Street loop can actually sustain these operations.

But if you want any services connecting the Kenmore and Seaport, or more than those 3 branches, such as adding a Harvard service, you would need two more tracks that go... somewhere. I think the options are:

1. Essex St Subway connecting the Boylston St Subway to the Transitway
2. Charles St Subway connecting to a quad-tracked Stuart/Kneeland St Subway
3. Pike Subway extension connecting Hynes to Back Bay, then with quad-track all the way to Tremont St
4. Post Office Square Extension, kind of a dark horse but why not?

 

[BosMaineiac]

Just to go way back to the original premise of this discussion, the capacity of the Park Street loop, I'm not sure that's actually unknown. It seems like, at least for a 2 week period, 3 branches, so about 20-22 TPH, used the Park Street loop continuously. (I believe the D was turned at Kenmore during that time.) That's definitely enough for the B/C branches, probably plus a restored A branch as well since that would be okay cannibalizing the B to some extent. That is to say, I think the concerns about tripod being unsustainable are mostly unfounded, and the Park Street loop can actually sustain these operations.

But if you want any services connecting the Kenmore and Seaport, or more than those 3 branches you would two more tracks that go... somewhere. I think the options are:

1. Essex St Subway connecting the Boylston St Subway to the Transitway
2. Charles St Subway connecting to a quad-tracked Stuart/Kneeland St Subway
3. Pike Subway extension connecting Hynes to Back Bay, then with quad-track all the way to Tremont St
4. Post Office Square Extension, kind of a dark horse but why not?

Doesn’t the Park St loop have too tight a radius for Type 10 vehicles? If trains use a loop there to turn back I think it has to be a reactivated outer loop, which will interfere with GC/north bound trains and thus reduce capacity overall.

 

[HenryAlan]

Doesn’t the Park St loop have too tight a radius for Type 10 vehicles? If trains use a loop there to turn back I think it has to be a reactivated outer loop, which will interfere with GC/north bound trains and thus reduce capacity overall.

I'm not sure on that -- the Type-10s, with all their short sections seem like they should be able to take that turning radius so long as the tunnel itself is wide enough for clearance. That said, perhaps there is room for a second pocket track inside the loop, with cross over, which would enable turning at least some trains there, regardless of loop viability.

 

[TheRatmeister]

Doesn’t the Park St loop have too tight a radius for Type 10 vehicles? If trains use a loop there to turn back I think it has to be a reactivated outer loop, which will interfere with GC/north bound trains and thus reduce capacity overall.

No, the Type 10s are designed to fit around the Park St loop. The T has stated that it's needed for operation flexibility with the current service levels, so removing it from service is not an option and altering it is obviously out of the question given the tight space underground.

I'm not sure on that -- the Type-10s, with all their short sections seem like they should be able to take that turning radius so long as the tunnel itself is wide enough for clearance. That said, perhaps there is room for a second pocket track inside the loop, with cross over, which would enable turning at least some trains there, regardless of loop viability.

But using the inner tracks as bay platforms could also be an option, although I suspect that would have an even lower capacity compared to the loop since the driver would need to swap ends.

 

[F-Line to Dudley]

I'm not sure on that -- the Type-10s, with all their short sections seem like they should be able to take that turning radius so long as the tunnel itself is wide enough for clearance. That said, perhaps there is room for a second pocket track inside the loop, with cross over, which would enable turning at least some trains there, regardless of loop viability.

There's no room for a pocket turnback track. The current pocket w/inspection pit is only 1 car (and will probably go away at some point because it can't fit a Type 10), and the area is riddled with support poles.

Don't forget, Park used to have 2 loops...the current inner-to-inner track one, and also an abandoned wider outer-to-outer track loop. The outer loop has crossing conflicts so wouldn't be able to take as much traffic as the inner loop (reason why it was mandated that the Type 10's be compatible with the inner loop), but if you ever needed a future where a lot of branches were turning at Park reactivating the outer loop for a little more turning capacity is trivial. The only change from its previous existence is that you'd just need to have a diamond where it crosses the to-be-installed thru service crossover on the inner track, and have GLTPS in-place so the more complicated variety of movements within all that infrastructure are stop-enforced.

 

[F-Line to Dudley]

Right, to be clear, that is my point: I'm arguing that tunneling under Marginal Road is significantly less disruptive to the community than Stuart/Kneeland would be. It's "4,000 feet of disruption vs 7,800 feet of disruption". I agree that Stuart/Kneeland is worth evaluating, but I think Marginal Road will come out ahead in most analyses, in particular because:

• Less disruption to neighborhoods
• Less "uncertain" tunneling due to proximity to Pike
  • @F-Line to Dudley if you have a citation for Marginal Road being "clean", that'd be awesome
• Due to the above, seems very likely to be the only possible alignment that might be seen as low-risk-enough to do cut-and-cover (major expense savings)

Marginal was re-laid in its entirety when the Pike came through in 1964-65 and required widening of the rail ROW. Prior to that the South End street grid abutting the rail canyon was very different, owing to just how many blocks were nuked in the name of urban renewal. There was a street there serving the same general function as Marginal bookending the canyon, but it was shifted a few dozen feet and ended at an awkward-angled 'square'-like intersection with Harrison and old Broadway instead of wrapping around onto Hudson. See Historic Aerials' 1950's topos of the area. Total nuke zone, so the utilities there are absolutely no older than 60 years on-the-dot and much better mapped than any streets a few blocks north on the 'old' street grid.

 

[Riverside]

Just to go way back to the original premise of this discussion, the capacity of the Park Street loop, I'm not sure that's actually unknown. It seems like, at least for a 2 week period, 3 branches, so about 20-22 TPH, used the Park Street loop continuously.

Good call! I pulled the Transit Matters data for an example day during this period, and yup, looks like it was about 20-22 TPH: https://dashboard.transitmatters.org/green/trips/single/?from=70158&to=70200&date=2024-12-11

That being said:

No, the Type 10s are designed to fit around the Park St loop. The T has stated that it's needed for operation flexibility with the current service levels, so removing it from service is not an option and altering it is obviously out of the question given the tight space underground.

This is the part that gets me though. "Operational flexibility" seems like it could mean lots of things, and it's unclear to me whether they are just matters of best practice (i.e. things that could be solved in a Green Line Reconfiguration world without physical modifications), or if they are just the practical realities of interlining multiple surface-subway routes at close headways. To put it another way, does whatever X that exists at 30 tph that requires an extra track/a turnback track also exist at 20 tph? If it does, then it suggests that the 20 tph during the diversion was viable for the short-term but not the longer term.

(Of course, that also raises the question of whether it would be a problem to send 25-40 tph from Kenmore to the Seaport.)

In any case, I think the Criss Cross vs Tripod dynamic does elucidate underlying comparative strengths: Tripod may be viable, but ultimately it requires Park Street to absorb trains from three trunks: Kenmore, Huntington, and Nubian; Criss Cross requires it to absorb only two. Tripod limits Kenmore branches to whatever can fit in the Park St Inner Loop; Criss Cross provides a dedicated subway that's able to be fully isolated. Those dynamics stay true regardless of the Inner Loop's actual capacity.

 

[Teban54]

(Of course, that also raises the question of whether it would be a problem to send 25-40 tph from Kenmore to the Seaport.)

I'd guess that's much more feasible than turning 25-40 tph at Park St, simply because the Seaport Transitway has a lot more options on the eastern end to terminate trains.

Obviously I'm not a train operator myself, but AFAIK, a lot of the time-killers in terminal operations actually come from the "software" side: ensuring that everyone gets off, adjusting displays and announcements to change directions, etc. Seaport has at least two such terminals to do this: Silver Line Way and Design Center. And SLW makes it very easy to set up separate terminating vs. through-running eastbound platforms.

In any case, I think the Criss Cross vs Tripod dynamic does elucidate underlying comparative strengths: Tripod may be viable, but ultimately it requires Park Street to absorb trains from three trunks: Kenmore, Huntington, and Nubian; Criss Cross requires it to absorb only two. Tripod limits Kenmore branches to whatever can fit in the Park St Inner Loop; Criss Cross provides a dedicated subway that's able to be fully isolated. Those dynamics stay true regardless of the Inner Loop's actual capacity.

I'm confused. Why should the total tph across all 4 tracks at Park St matter under Tripod, if, under regular operations, the two inner tracks and the two outer tracks can be largely independent of each other?

(Or, in other words: Assuming we can comfortably send 30 tph from Boylston Outer to Government Center, why should it affect those trains if we add another 10-20 tph from Boylston Inner to Park St Inner?)

One possible answer is signaling, but I remain unconvinced. The NYC subway, despite its notoriously old-school operational practices, has no trouble with two trains departing or entering a platform on express and local tracks respectively at the same time. Unless there's a need for a train to switch between the inner and outer tracks, I don't see why other trains at Park St can't do this.

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

From an earlier comment:

I know what you mean here, but I'm always pretty skeptical when this line of reasoning is brought up. The tunnel's proximity to the "Main St" of a neighborhood doesn't matter: it's the station locations that matter. In this case, the only possible intermediate station would be some counterpart to the current Arlington station -- such as at Arlington St itself, or at Bay Village. Realistically, there are two potential station locations: something near Park Plaza, and something near the Pleasant St Portal. Park Plaza is obviously close to stuff, but it's also quite close to Arlington. Bay Village at least fills a gap in coverage (including providing better access to the South End on the other side of the Pike).

For one, I imagine that if there's one crucial intermediate station that any "Essex Criss-Cross" or "Stuart-Kneeland Criss-Cross" will need, it's not Arlington -- but an interchange with the Orange Line, around either Chinatown or Tufts Medical Center. An Arlington replacement (or the ability to stop at present-day Arlington) pales in comparison.

Let's also not forget that, especially for Criss-Cross, the goal of any intermediate stations is not so much to "add something onto the T map for the first time", but "as a replacement for Kenmore/B/C riders heading downtown". Such riders treat the intermediate stations as destination nodes, whether they're for Public Garden and Boston Common, shops at the eastern end of Newbury St, jobs at Park Plaza, Asian restaurants in Chinatown (including for the Asian populations living in Allston), or TMC. Stopping at Bay Village and maybe Chinatown Gate is a huge inconvenience for the first three, and a minor inconvenience for the last two.

Yes, having a Bay Village station really helps nearby residents, especially those south of the Pike. But the existence of a Bay Village station is practically guaranteed thanks to the Nubian branch, regardless of where the second east-west LRT trunk will be and which routes it serves. I doubt the benefits of additional service at Bay Village outweigh the inconvenience of riders who now need a transfer for their old commute patterns.

(This discussion can still be had for Tripod and rerouting the D/E to Bay Village, but at least they would still stop at Boylston.)

But the only thing I do want to highlight, as folks consider how to achieve the Full Criss Cross: improbably, downtown Boston has not one but two fully intact flying junctions that sit unused. No matter how you cut it, flying junctions are harder and more expensive to build than simple flat tunnels. Designing around those, ensuring that both get used, seems like a good pay to supercharge any cost-benefit analysis.

(I think BosMaineaic's suggestion does end up using both? But I'm not clear what happens on Boylston between Charles on Tremont?)

Interestingly, Tripod actually makes much better use of both flying junctions, particularly the Pleasant St incline, than Criss-Cross:

• Tripod: The Pleasant St junction sorts out Huntington trains (taking Bay Village West) and Nubian trains (taking Bay Village South). Then, the Boylston junction separates both from Central Subway trains. Both junctions can operate at full capacity in theory.
• Criss-Cross: The Boylston junction can operate at full capacity, but the Pleasant St junction only sees Nubian trains. Firstly, it remains debatable whether Nubian needs two branches in the first place. Even if we assume it does (which is my preference), that still means Criss-Cross will inherently see much lower utilization of the Pleasant St junction, in terms of tph passing through it.

This is not an argument against Criss-Cross, not even close; but while I do think that utilizing at least one of the two flying junctions is hugely beneficial and should be factored in, it shouldn't be a constraint to aim to utilize both. Especially when a part of the analysis for Pleasant St comes from Nubian, which has been largely detached from discussions of Full Criss-Cross above.

 

[TheRatmeister]

This is the part that gets me though. "Operational flexibility" seems like it could mean lots of things, and it's unclear to me whether they are just matters of best practice (i.e. things that could be solved in a Green Line Reconfiguration world without physical modifications), or if they are just the practical realities of interlining multiple surface-subway routes at close headways. To put it another way, does whatever X that exists at 30 tph that requires an extra track/a turnback track also exist at 20 tph? If it does, then it suggests that the 20 tph during the diversion was viable for the short-term but not the longer term.

The "operational flexibility" in question that is needed (presumably) comes from both capacity limitations at the Government Center loop, or from any other circumstances which may require trains to be turned at Park St, such as the diversion in question (realities of the track layout and running this many branches) and the need to regulate headways by turning some trains early. If trains can't use the loop, there is no possibility for trains to be turned back at Park St, which is a problem for both of this circumstances.

In both of these situations, the complete decoupling of Tremont St service from Boylston St service is useful. The number of trains using the Park St loop is not affected in any way by the service to/from Nubian or Huntington. The reactivation of the outer loop plus the use of the Government Center loop could maintain the short-turn flexibility on the Huntington/Nubian routes while also maintaining separation. The Park St inner loop and Kenmore loop could adequately support the Tremont St branches. In other words, even operated independently both sets of branches can maintain short-turn flexibility. (You know, it's almost like the people who built the Tremont St subway knew what they were doing.)

I'm confused. Why should the total tph across all 4 tracks at Park St matter under Tripod, if, under regular operations, the two inner tracks and the two outer tracks can be largely independent of each other?

And +1 on this. Operations of both trunks are separate and independent.

improbably, downtown Boston has not one but two fully intact flying junctions that sit unused. No matter how you cut it, flying junctions are harder and more expensive to build than simple flat tunnels. Designing around those, ensuring that both get used, seems like a good pay to supercharge any cost-benefit analysis.

Beware of sunk-cost fallacy. The existence of those flying junctions gives us great options to build new transit going forward, but it should not dictate what that transit looks like.

 

[Riverside]

Don't have time tonight to respond to things (will try tomorrow), but one other point did occur to me, which is the length of trainsets. Current 2-car sets are about ~150 feet long. A single T10 supertrain will be 114 feet, and a supertrain doubleset ("hypertrain"?) will be just under 230 feet, the length of GLX platforms. It looks to me like the Park Street platforms are about 390 feet long. So it would be possible to double berth a pair of supertrains, but not a pair of hypertrains.

I'm not sure if the T is currently double-berthing trains at Park Street, but I know they have in the past, so that's another variable to consider when estimating future throughput capacity.

 

[TheRatmeister]

I'm not sure if the T is currently double-berthing trains at Park Street, but I know they have in the past, so that's another variable to consider when estimating future throughput capacity.

I don't think so? It's quite operationally complex and not super practical if you want the trains to, you know, stay still while people board. This was not a concern in the early 1900s.

What I'm more sure about is that it's not a problem. Government Center operates just fine as a terminus for 2 branches without double-berthing, and the rest of the central subway can handle 3-4 through branches without it either.

 

[Teban54]

I'm not sure if the T is currently double-berthing trains at Park Street, but I know they have in the past, so that's another variable to consider when estimating future throughput capacity.

I don't think so? It's quite operationally complex and not super practical if you want the trains to, you know, stay still while people board. This was not a concern in the early 1900s.

What I'm more sure about is that it's not a problem. Government Center operates just fine as a terminus for 2 branches without double-berthing, and the rest of the central subway can handle 3-4 through branches without it either.

I do agree with @TheRatmeister that it shouldn't be a problem. The T is already planning for concrete, real-world goals of running 2-car Type 10 hyper trains on today's Green Line, so they'll have to consider the fact that they can't berth two trains on the same platform (even if they do now -- which I think can happen, but am not sure at all and never paid attention to it).