Green Line Reconfiguration

[Arlington]

General Green Line opps question: Even before we get into a full reconfiguration, is there a reason why each GL line (B, C, D, E) needs to turn around at a set station on the N side of the network? As I understand post-GLX it'll be D at Tufts/Medford, E at Union

What's the alternative to fixed termini?
I actually think E to Tufts and D to Union (based on route-length) is more likely
But certain points are better for turn-backs (Park St Loop, Gov Cen Loop, North Sta tail tracks)
You could also see a "G" Tufts to Governement Center inner loop if "local" demand was strong enough

 

[Riverside]

No transfer here [near Tufts Medical] really only affects riders from the South End to SWB which I've always argued is a very small number (although does SL4 ridership justify this? Asking for a friend.)

In short, quite possibly. According to the Better Bus Profiles for SL4 and SL5, some 35% of SL4 riders (1,050 riders) alight at South Station; some 1,710 alight at Temple Place on SL5, which is higher, but is also largely in line with the ratio of SL4 buses vs SL5 -- SL4 provides 56% as much service as SL5, and South Station's 1,050 is 61% of Temple Place's SL4.

Now, the question is whether those riders are continuing on to the Seaport or otherwise destined properly for South Station -- or whether they are destined for a Red Line transfer either way, and are agnostic to doing that transfer at South Station or Temple Place. Given the fact that the numbers line up pretty closely to the service ratios, it may well simply be the latter.

Of course, insert here vague statement about LRT connection to Seaport creating new demand, blah blah blah.

Also the loop at South Station is shown as in service. This is just me being too lazy to change it. Could do loop service if you want.

I would hope that the loop is maintained -- turnbacks are never a bad thing.

JumboBuc said:
General Green Line opps question: Even before we get into a full reconfiguration, is there a reason why each GL line (B, C, D, E) needs to turn around at a set station on the N side of the network? As I understand post-GLX it'll be D at Tufts/Medford, E at Union

What's the alternative to fixed termini?
I actually think E to Tufts and D to Union (based on route-length) is more likely
But certain points are better for turn-backs (Park St Loop, Gov Cen Loop, North Sta tail tracks)
You could also see a "G" Tufts to Governement Center inner loop if "local" demand was strong enough

There are a few overlapping reasons for the current terminus arrangement.

First, there is a bottleneck north of Park Street -- you cannot run (right now) as many trains between Park and GC as you can between Park and Boylston. That means you need to terminate one of the branches at Park Street.

Second -- each Green Line branch is loooong but some are reallllllly long. The Stonebrown Design Travel Time map is super illustrative here. The D Line runs all the way out to Riverside, 40 min from downtown. Given such a long route, and particularly since there isn't such strong demand north of downtown, it makes more sense to turn the D Line around at Government Center and get those trains back out where they're needed.

Likewise, the B Line traverses such a curvy route and makes so many stops that it also ends up taking 40 min to arrive at Park St. Unsurprisingly, the B Line is the one that currently terminates at Park St, the idea being to turn those trains around as quickly as possible and try to get them back out to Comm Ave to plug up all the delays.

If you look at the map, you can probably guess who terminates at North Station, and who (previously) ran all the way to Lechmere. (Answer: the C Line -- 30 min from downtown -- ran to North Station, while the E Line -- 22 min from downtown -- ran all the way to Lechmere.) Longer travel time to downtown means more opportunities for delays, which means more opportunities for their to be long gaps between trains in the Central Subway.

In theory, by running E trains to Lechmere, you had the best bet of reliable service north of North Station. (Of course, your mileage may vary, and it was hardly a perfect solution, since the E features the only street-running segment left in Boston.)

Now, this raises an obvious question -- if the D Line is already one of the longest lines, why in Charlie-on-the-MTA's name would that be the branch you run through to GLX? Well, there are two reasons.

First, while the E is the obvious choice, you need two branches to go through, so you have to pick one of the other three. One by one: the B Line is a disaster even in its current arrangement -- through-running would be even worse. The C Line is a decent enough candidate, but like the B Line, it's also more vulnerable to delays because it interacts with traffic so much. The D Line, though long, tends to be pretty reliable, in part because it has a ROW that is almost completely "sealed". Of the three, I think the D is the reasonable candidate.

Second, there's another strike against the C Line, and it's illustrated by a combination of its ridership numbers and frequencies. Check out page 29 of the 2014 Blue Book. The C Line has the lowest ridership of any Green Line branch, even though it has about as many stations as the D, and is noticeably longer than the E. The C Line's frequencies are also slightly lower than the rest of the network. I can't say that the frequencies cause the lower ridership or vice-versa, but it doesn't really matter either way. The point is that the C Line either underperforms or is underserved compared to the other branches; that makes it a bit less ideal to be paired with one of the new northside branches, where it looks like the demand will be more comparable to the existing D and E Lines; you want to match branches where the demand looks similar, so there we go.

As a fun addendum here, you can see from this diagram of D Line termini over the years from Wikipedia that these terminal assignments have hardly stayed constant over the years. So it's definitely an evolving phenomenon.

 

[fattony]

View attachment 9571

Maybe I’m missing something, but I don’t understand this E line to Seaport at all. It connects Huntington subway to Seaport and has a transfer with the D line, but that’s it. So from Seaport it’s 2 seats to the central subway and 3 seats to Kenmore. Same as today. Why bother tying it into the GL network at all?

 

[George_Apley]

Eh, that would require a long back track that no one would make.

They would if it was their only 2-seat ride to the Seaport.

 

[vanshnookenraggen]

Maybe I’m missing something, but I don’t understand this E line to Seaport at all. It connects Huntington subway to Seaport and has a transfer with the D line, but that’s it. So from Seaport it’s 2 seats to the central subway and 3 seats to Kenmore. Same as today. Why bother tying it into the GL network at all?

I agree, it isn't ideal. But without the Essex St alignment you can't have ideal. My first alternative, the B/C going to the new subway, would fix this. But it is more complex and expensive. But maybe it would serve riders better. Only a full O/D study will tell us the branches which should serve SBW.

 

[Charlie_mta]

I agree, it isn't ideal. But without the Essex St alignment you can't have ideal. My first alternative, the B/C going to the new subway, would fix this. But it is more complex and expensive. But maybe it would serve riders better. Only a full O/D study will tell us the branches which should serve SBW.

I'm not giving up on the Essex Street alignment. The new tunnel would begin on Boylston Street, halfway between Arlington and Charles Streets, tying into the existing Green Line tunnel at that point. Then it heads east along Boylston Street dropping at a -5% grade, its tracks arriving well beneath Boylston station at 66 feet below street level. Continuing east along Essex St at -5% grade, the tunnel tracks pass under Washington Street at 80 feet below street level. At that point it's on a sag vertical curve and begins rising at +4% grade to pass its tracks under the SB X-way tunnel at 40 feet below street level, about enough depth for clearance under the X-way tunnel. The new tunnel continues east along Essex St at +5% to 6% grade to reach and match the busway tunnel elevation at SS.

I'm thinking the deep-bored tunnel might be deep enough to mitigate or avoid impacts on existing subways and building foundations. To me its the best alignment for ridership catchment, T station access, and direct routing. I've looked at other alignments to the south and they all involve at least two 90 degree tight turns, whereas this Essex one only requires one sharp curve at the junction with the busway tunnel.

 

[vanshnookenraggen]

I'm thinking the deep-bored tunnel might be deep enough to mitigate or avoid impacts on existing subways and building foundations. To me its the best alignment for ridership catchment, T station access, and direct routing. I've looked at other alignments to the south and they all involve at least two 90 degree tight turns, whereas this Essex one only requires one sharp curve at the junction with the busway tunnel.

But deep boring, at the depth where you can be safe, requires it do be like 100 feet down. You don't have the clearance to get there and back up.

Even if it was possible to build what you describe it would require expensive shoring up of every building and the tunnels in the area. None of this is impossible, it's just really expensive. As F Line pointed out it was these constraints which made the SL tunnel too much of an ask.

In order to get a modern station platform under Essex St you are going to need a bi-level station. You can get a single level down there at the grades you describe but not a bi-level. Even if you went back down Boylston St far enough to get 100 feet below Essex St you still need to get back up to Atlantic Ave. It doesn't work.

As you point out any alignment south of here requires sharp 90 degree turns. That's a non starter. So to me the only thing that works is to tie this into a new Green Line tunnel from Huntington Ave or even Hynes. More construction, but way easier.

 

[Charlie_mta]

Sharpening up my calcs a bit here, a tunnel could be (at track level) 110 feet below grade at Boylston and Tremont if a -6% grade is used (= 1280' x .06 plus 20' depth where it joins the existing Green Line tunnel, plus 13 ft ground elevation rise). That would allow 110' track elevation below Washington Street as well. Then the tunnel would rise up at about 4.5% to end up 40' below grade (at track elevation) at the X-Way tunnel. So where the tunnel crosses under Washington and Tremont Streets. the tunnel roof is about 90 feet below street level. I'm no expert on tunnel engineering, but just looking at the geometry.

 

[F-Line to Dudley]

Sharpening up my calcs a bit here, a tunnel could be (at track level) 110 feet below grade at Boylston and Tremont if a -6% grade is used (= 1280' x .06 plus 20' depth where it joins the existing Green Line tunnel, plus 13 ft ground elevation rise). That would allow 110' track elevation below Washington Street as well. Then the tunnel would rise up at about 4.5% to end up 40' below grade (at track elevation) at the X-Way tunnel. So where the tunnel crosses under Washington and Tremont Streets. the tunnel roof is about 90 feet below street level. I'm no expert on tunnel engineering, but just looking at the geometry.

Yeah...the geometry is doable with LRT. At a speed penalty. But you don't get rid of the cost chew from underpinning mitigation.

AGAIN, FOR EMPHASIS: structural underpins were the Design fatality that caused the no-rec rating. You can engineering-maths a "Look! I proved physical feasibility!" all you want. SL III did that. It didn't matter because MassDOT could not fund 100% Design...nevermind the Build...by its lonesome without the committed FTA match funding. And FTA match funding was removed because the underpins were a cost blowout. You can't square that project fatality with a personal preference for Essex.

Please re-read the previous posts about how every "Yeah, but..." retort clinging vainly to Essex alignment fails to address this point that it doesn't get built without restored funding. It's getting tiresome to have to keep repeating this. It is doubleplus impossible with the insertion angles from the Transitway to dive deep enough fast enough to avoid a major cost-blowout structural underpin along the Essex corridor, and that was the objectionable thing. This is an unfunded mandate, not a Crazy Transit Pitch. Take "But I want Essex!!!" wishful thinking dreams to the Crazy Transit Pitches thread. Absolutely NONE of those will get funding commitments restored on this unfunded mandate of a project, and that is the be-all/end-all fix job here.

You have multiple not-Essex paths to debate to the cows come home in this Reimagine thread. But not Essex...never Essex. Yet another "But I want Essex!" repeat is a waste of time and energy, because it can't be done within without Fed help and the Feds have drawn their line in the sand that they won't help unless we avoid structural underpins. There isn't an alternate universe that'll make it so; those paths have long been sealed.

I agree, it isn't ideal. But without the Essex St alignment you can't have ideal.

^Put this quote on a T-shirt^ if absolutely nothing else brings the point home. You can't have your perfect-as-crow-flied 2D straight line on a map. Can't. So figure out a viable compromise. The fact that the viable compromise 'jogs' all AVOID the speed-clobbering grades of the Chinatown underpin means they move faster for their greater tunneling length. 2D crow-flies Google Maps perfectionism doesn't matter when the Essex alignment was performance-compromised out-of-box. So most of the Alternatives, while a compromise on straightness, are indeed for-reals "equal or better" on the clock Boylston-SS. Multiples of alignments would accomplish equal-or-better, so in the real world we're actually sitting pretty here!

While I don't personally agree with Van that one right-turn angle down Tremont to 'a' (pick-'em/no-preference) more-feasible side street is enough of a performance demerit to flat-out mandate a bypass subway from the west for variety of reasons (mainly because it's hardly a system-worst curve and the Tremont Tunnel was inherently zippy)...at least "bypass must be so" is a fully valid Project Alternative to debate away in this thread. Or even something you could string together have-cake/eat-it-too on the installment plan so one particular 'jog' build eventually turns into a 'bypass'. I have no objection to poking around that one for all it's worth, since it is buildable without committing any of the same structural underpin sins. It's the "But I want Essex so reality's going to bend to make it so if I wish hard enough!" that advances the discourse nowhere. You can't do it. We don't even have the internal funding mechanisms to ill-advisedly go rogue from the Feds and do it if it truly meant as much to the planners as it does to some individuals on this board. Move on and hash out what we can build, because this thing that's 15 years tardy will never advance if Essex stanning drags the planning to a standstill. Stanning for Essex over anything buildable ensures nothing will ever get built.

 

[F-Line to Dudley]

That was my question, too.

A second question: Why terminate Blue at Kenmore? Why not continue down Brookline Ave and terminate at Brookline Village to pump D/Needham through Huntington? If heading into the heart of Longwood, D/Needham riders would transfer at Brookline Village to Blue, sure, but travel times would be a wash because of a much more central station location at Brookline/Longwood compared to the current D Longwood. Provides a faster connection for northside RL and OL riders heading to Longwood, and probably reduces a lot of demand for shuttles/future BRT connecting to Lansdowne.

One practical reason: Charles-Kenmore is the project area for Storrow Dr. midsection trade-in. Therefore anything beyond is by funding fiat going to be a detached effort because the set Storrow project area is where funding will be calculated on offsets for displaced cars-to-transit, while anything beyond is its own separate animal.

Second practical reason: there isn't a ready consensus for where you'd take it past Kenmore. There are pros/cons for Longwood, pros/cons for West/Allston and beyond that haven't been studied to enough degree to settle out. So for that reason alone you're building the consensus part (i.e. if/when we agree as a City + State that Storrow midsection is expendable) before the not-yet-consensus part.


For Crazy Transit Pitching purposes you can crayon-away, because the tail tracks up Brookline Ave. will be total choose-your-adventure on where you opt to go next. It's more for bureaucratic realities that Charles-Kenmore is a one-shot project and past-Kenmore is a disconnected second-shot project. You can tandem-plan them in close succession if that's worth it, but it won't be a "monolith" build into the suburbs. There will be multiple installments...therefore when we talk BLX-Kenmore in terms of gravitational influence on Green Line we're limiting scope to just the Charles-Kenmore leg and not speculating yet about what goes past because installment-after-next prospects are a tad beyond prediction right now.

That's all. There's no functional limitation on dreaming away. It's just a slotting/phasing practicality.

 

[Charlie_mta]

Yeah...the geometry is doable with LRT. At a speed penalty. But you don't get rid of the cost chew from underpinning mitigation.
AGAIN, FOR EMPHASIS: structural underpins were the Design fatality that caused the no-rec rating. You can engineering-maths a "Look! I proved physical feasibility!" all you want. SL III did that. It didn't matter because MassDOT could not fund 100% Design...nevermind the Build...by its lonesome without the committed FTA match funding. And FTA match funding was removed because the underpins were a cost blowout. You can't square that project fatality with a personal preference for Essex.

I agree with your concerns. In addition to underpinning Boylston Station and the Orange Line tunnel and station at Essex Street, an Essex GL alignment would require underpinning the Central subway between Charles and Tremont Streets, as the new GL tunnel would be underneath it along that stretch. The one thing that keeps me from ruling out an Essex alignment is that the previous study for the bus tunnel was a different facility on a shallower vertical alignment (as shown in the EIS). If I were approaching this as the MBTA, I would conduct a separate design study for an LRT deep tunnel along this route and see what results.

 

[Riverside]

About a month ago, I wrote at some length about the downsides of certain branches of the Green Line only being able to run into Park Street. In particular, trains running on/under Commonwealth (whether from Boston College, Oak Square, Allston, Harvard, or Cambridge) and trains coming from Nubian via Washington could only be fed from a "faucet" at Park Street. By contrast, all of the (potential) northside branches, along with service on Beacon, the Highland Branch, and along Huntington, could be supplemented with short-turns (e.g. Kenmore) and diversions out of the Central Subway (e.g. to the Seaport). As I explained there, I think that there is a real need to find some way to feed extra trains onto these branches from sources other than Park Street; otherwise, you risk needing to give them undue priority in the Central Subway, crowding out other services. That post kicked off some interesting discussion, including some discussion of diverging the Central Subway somewhere near Copley or Hynes to feed into a second subway under Stuart or along the Mass Pike (corridors which previously we had only really considered feeding from Huntington).

On further consideration, I think I've identified a few very modest projects that could be added into the mix as "relief valves", which would provide the flexibility for Nubian and Commonwealth to bring them on-par with the rest of the network.

To be very clear: I don't think any of these are must-haves or must-builds. Rather, I think they should be added to the universe of projects under consideration, and deployed if (and only if) the need becomes clear. (Which is the approach that underpins the entire Reconfiguration Concept.)

1. Turn Most "Kenmore Division" Trains at Park Street

I don't think this needs to be an ironclad proposal, but I think it should be the default. By dedicating the inner tracks to "Kenmore Division" trains, and reserving the outer tracks to "Tremont Division" trains, the "local streetcar" services of the B/C can be isolated from the "sealed ROW LRT" services of a combined D+E (assuming the Huntington Subway is extended), and capacity north of Park Street can be reserved for Tremont Division trains to through-run to the north, where the corridors are generally better matches for the Tremont Division "sealed ROW" flavor of LRT. On its own, I don't think this move is a sufficient "relief valve," but I think it's a useful part of the conversation.

2. Add a Nubian-South Station Leg East Of Bay Village

As laid out previously, one possibility would be to construct a "flexi-junction" at Bay Village (near the current Tufts Medical Center station), that would allow north-originating trains to a) loop, b) head to Back Bay, c) head to Nubian or d) head to South Station. This is an elegant solution that solves a number of problems. However, it still leaves Nubian being fed exclusively by Park trains (and possibly some de facto short-turns at Bay Village). Nubian merits high headways, and by splitting the Green Line trunk so close to downtown, this will inevitably eat up a sizable fraction of the capacity at Park Street.

This can be ameliorated by adding a south-to-east connector between the Washington Street corridor and the Bay Village-South Station corridor, enabling Nubian to be served both by Park-originating trains and by Seaport-originating trains. As I explained earlier, current SL4 ridership demonstrates that there is some market for Nubian-South Station service, although it's worth noting that this LRT version would miss the transfer opportunity at Tufts, so it's not perfect. Still, by giving Washington Street two sources, it would no longer be necessary to devote such a sizable share of Park trains to Nubian, freeing up capacity and flexibility for the rest of the network.

3. Build a Small Yard at the Public Garden Portal

This one I am the least confident in. As can be seen on vanshnookenraggen's track map, there is a short stretch between Arlington and Boylston where the tunnel widens, and there currently is a solitary lead track. This dates back roughly 100 years, from before the Huntington Subway was built, and when streetcars from Heath and Arborway (and actually Brookline Village) ran along Boylston at street level before entering through a portal at this location.

Building a small proper yard here would allow the T to keep one or two trainsets "on-call" during peak hours; in the event of severe inbound delays on the Commonwealth branch(es), it would be possible to dispatch one of these on-call sets from Public Garden outbound in an effort to fill in the gaps caused by the cascading delays. These would need to be "run-as-directeds," as I can't imagine regularly short-turning services at Arlington station, but it would still give the Commonwealth branch(es) somewhere to receive supplemental short-turns from, as opposed to relying solely on Park Street.

There are two major question marks here. First, it's unclear what kind of structural work would be needed underground to actually build the yard. On a map, obviously it looks straightforward enough, but presumably that area is "a very three-dimensional space", given that it has supported multiple portals over the years; it may be a lot of work to re-lay tracks. Second, I'm not entirely sure the space is long enough to hold a three-car trainset, with space for leading and trailing crossover switches. Now, if used solely for the operating model I've described above, technically you wouldn't need access from the inbound tracks, so that could simplify things, but it's still a bit unclear.

continued below

 

[Riverside]

4. Add Kenmore Short-Turns on Commonwealth

"But this is impossible!" - me, several weeks ago. As the Byrds' song goes, "I was so much older then, I'm younger than that now."

There are several ways to add Kenmore short-turns on Commonwealth, ranging from low-build to high-build. None of these are mutually exclusive -- definite potential for mix-and-matching and/or sequential builds here.

4a. Bus Lanes

Commonwealth Ave is ridiculously wide. There would be political battles to fight, for sure, but in terms of the physical space, there's no reason why dedicated bus lanes could not be provided, upon which short-turning bus service could be run, terminating at the street-level bus station in Kenmore Square. Some consideration would need to be given to informing riders whether they should wait on the bus platform or the train platform for the next arrival (ideally, such platforms could be shared), but that's easy enough to solve, and would end up providing service that is roughly comparable to Beacon short-turns. Now, to be clear, it's still a bit imperfect -- the transfer at Kenmore wouldn't be cross-platform; however, Kenmore will need to be a seamless upstairs-downstairs bus-LRT transfer station by this point anyway -- and presumably some of these riders would be transferring to another bus anyway. So it doesn't need to be fatal.

(Note that if a subway is extended under Commonwealth Ave, then the bus lanes could just eat the existing track reservation.)

If there ever is resurrected service to Oak Square, presumably the trolley tracks would get dedicated lanes, upon which the 57/57A could continue to run as normal, providing short-turn service along there.

4b. Transitways

Instead of trying to eat parking lanes or existing auto lanes, the Commonwealth reservation could be rebuilt as a transitway with tracks embedded in pavement. This would solve the "which platform to wait on" problem, and would avoid the political battles of bus lanes. But it would come at pretty substantial cost, as it would be a rebuild of the reservation: beyond having to pour a whole lot of pavement, my understanding is that the ROW is currently just a little bit too narrow for buses... which would mean that all of the stations would need to be rebuilt as well. This solution probably is one to strategically deploy at certain locations, as opposed to the whole corridor.

4c. Add A Street-Level Terminus At Kenmore

I know this sounds like Crazy Transit Pitches, but hear me out...

The distance between Kenmore Square and the Blandford portal is super short -- roughly 700 feet. Comm Ave west of the intersection is 4 lanes + parking + reservation, and to the east it's 7 lanes + parking. There is width aplenty here, and only a very very short length to traverse.

Let's assume that the "short-turn tracks" diverge just west of Blandford Street station, jutting out into Comm Ave. At Blandford, inner tracks would serve through trains, and outer tracks would serve short-turns (meaning that Blandford itself could offer a cross-platform transfer). Including the stretch along the Blandford platforms, it would only be 1300 feet of street-running that would need a dedicated lane (and the last 100 feet or so would be within the Kenmore terminal area anyway). Put another way, that's, like, worst case scenario, two blocks of on-street parking that would need to be eliminated, which really isn't that much.

There a few challenges here.

First, there it's very questionable whether there would be enough space for a proper loop track at Kenmore Upper -- best case scenario would be a loop from the outside of the current bus terminal to the middle lane of westbound Comm Ave north of the station; that looks like it is a 105 ft semicircle, which is just a smidge tighter than Lechmere Loop's 110 ft.

So it would probably need to be a stub-end track or two. That's a bit easier from a rail operational perspective, but the bus operations might need to be tweaked, to avoid wrongway conflicts between arriving buses and departing trolleys (since obviously the buses are always going to be using loops rather than stub-ends).

Second, there's room at Kenmore, but not, you know, a lot of room. This would all but certainly require expanding the footprint of the station, and it might be difficult to accommodate three-car trains. (Although hopefully short-turn services could be run using doubles instead.) Assuming a two car trainset of ~150 feet, and then 90 feet of running track for a crossover, plus some space for merging bus lanes... I think it's doable, but it would require planning (and almost certainly a rebuild of the Kenmore surface station, which might be in order anyway.)

A very rough render of what this station might look like is below. Solid green lines are represent trolley tracks, solid yellow lines indicate bus paths, the dashed segments indicating overlap between the two; the grey boxes are platforms, and the thin blue lines indicate some level of semi-permeable isolation of the transit lanes. Not illustrated would be transit priority signals, in particular at the northeastern entrance to the Commonwealth/Beacon/Brookline intersection.

Obviously there's lots of room for variability here -- if you were willing to lose that pretty median just west of the station, you could elongate the platform, or separate the trolley lanes from the bus lanes. If you were willing to eat the northernmost eastbound lane just south of the island, you could probably convert the lane into a bus platform, and split the existing berth into a narrower 2nd bus platform and a 2nd bus lane.

The final challenge is that this option doesn't help you if you build a subway under Commonwealth; there's no way you're gonna build a "reverse portal" near BU just to access a dinky surface terminus. So, if you build this, then it likely means a) Harvard and Grand Junction trains will never use it, and b) you would need to maintain the surface ROW even if you built a subway, and your short-turn Boston College and Oak Square services would continue to run above ground to it.

I don't think either of those are the end of the world. Harvard and Grand Junction would both be supplemented by additional services. Having short-turn services run above ground while through-services run underground is a bit inelegant, but every station would still have a platform with "full" frequencies: a new "BU Under" station would have through-running A & B, plus Harvard and Grand Junction, and stations west of the Commonwealth portal would have short-turns and throughs on the same platform anyway.

In a world with "Kenmore Upper", I'd be more optimistic than ever about full restoration of the A Line at least to Newton Corner -- but rather than through-running, most such trains would short-turn at Kenmore Upper, making the A Line less of a "branch of the Green Line" and more of a connector service between major transfer nodes at Kenmore and Newton Corner (or Watertown); this service wouldn't impact capacity in the subway, and also would be operationally isolated, meaning that the cost of street-running wouldn't propagate throughout the network.

Is "Kenmore Upper" a surefire winner? Definitely not. Is it mandatory or high priority? Not at all. Is it possibly feasible as a matter of last recourse? I think it might be. The key is that it's a very short distance between the Green Line tracks on Commonwealth and the Kenmore surface station, and that there is physical space at Kenmore for an expanded station, which could include a small LRT terminus.

Certainly, to me, this seems much more feasible than a bypass of Copley via Hynes and Back Bay, and much more feasible than trying to hook the B Line into the existing Kenmore loop. There is no present need for it, but if there ever were, I think it could work.

Conclusion

Stepping back from the idea of "Kenmore Upper" -- which is obviously the extreme outlier here -- my point is that some combination of all three of these could be used to provide Commonwealth service originating from Kenmore (as opposed to Park), and without the need for absurd blowout construction costs. Combine that with a Nubian-South Station connector (possibly a bit more costly), and then you would truly have a versatile network where every branch and every trunk can be fed by multiple sources, ensuring that every stretch has as much flexibility as needed to support proper high-frequency service.

 

[davem]

Hi. This was always kind of my baby, so I might as well reply...

Riverside, when you're running your calculations I think it's important to consider a Oak Square via Common Ave and Boston College via Comm Ave as one line together, not two. Demand dramatically drops off beyond Harvard Ave, hence why the idea of either a stub extension to Union Square or a Griggs Street yard have always been such good ideas. The only reason every B train goes all the way to BC is because the T has no other good option to turn trains before there. (The packards corner turnback is useless for all but emergencies). The fatal flaw in your numbers is that Harvard Ave - Chestnut Hill Ave does NOT need full rapid-transit headways, even at rush.

This is especially true if you assume in the process of the reconfiguration the tracks on chestnut hill ave through cleveland circle are reconfigured. This could allow more load from BC to be shed via a dinky BC - Kenmore loop via Highland Branch short turn at rush hours. You could probably get away with 3:2, or even 2:1 oak square vs boston college at rush. Either way, adding oak square does not mean we have to add very many more trains through BU.

A harvard ave branch is similarly going to peel demand off inbound comm ave trains (with people instead going outbound to transfer to the red line), as are the allston and brighton CR stops. Speaking of a harvard branch - it wouldn't be too much more trouble to keep running under the pike after the BU Bridge, and have it tie into the C instead of the B. That way you could have the "H Line" just bounce back and fourth between Harvard and Kenmore. This way it isn't fed from Park at all. A lot of rush hour load is also going to be taking the CR from the Brighton and Allston stations, bypassing the light rail system altogether.

All this is to say, I think you're placing too much emphasis on keeping up the TPH on the comm ave line, and artificially inflating how many rush hour trains would actually need to be running. There are much easier ways to not overwhelm park st.


Now as for worries about the B and C blowing their headways because of surface running: I don't think that's as much of a concern either. If we're talking about
- building a connector between Brookline Village and Huntington
- building a new subway beneath huntington
- building a new subway beneath marginal road
- building a superstation at bay village
- building a connector from south station to bay village
- laying tracks and rehabbing the whole transitway
- building a layover yard somewhere in the waterfront

I don't think some "easier" ways to alleviate many of the concerns with street running are too crazy:
- burying the comm ave line out to Packards corner (or union square)
- building a bi-directional turnback + 1-2 storage leads at Packard's Corner or Union Square (if an Oak Square train is mashed up in traffic, dispatch a new train from here on schedule, then turn back the oak train upon arrival)
- eliminate 1/3 of the crossings on the C
- a trolley reservation between packards/union and St E's
- signal prioritization
- raising potential TPH through park (better signals, connecting the inner outbound platform so any train can access any birth on any platform, and not have to wait for it's specific slot)


Also going to do a quick info dump of some stuff I created a while back, before I disappear for another couple years...

Bay village (this would tie in with the adjacent OL station). Van's map a couple pages back missed the full potential of what you can do here. The only think I don't like about this is that you can get an "inbound" train to Park Street from EITHER platform, which could be confusing. An island platform would alleviate this issue, but as of right now the northern platform at the western end is a full level higher than the southern platform, which means it would be a split island, only connected at the eastern side. I guess you could have a flight of stairs on the west side? IDK, I'm sure someone can figure it out

My full-build map of how the reconfigured GL would be routed. After living in brighton for years I determined an oak square branch would be crazy to try to build, and I actually was sending the BL out to Harvard...

Full Size

The SL loop and essex st lead location. This does scale if you lay it over google maps:

My Plan for connecting into the transitway. I spent a bunch of time driving and walking down there and was convinced the ramps from 90 were low enough to do this on top of (based on this view, which is already "two levels down", and then the tunnel just keeps descending)

Obviously you can use the essex street leads as well, but if it's possible to eliminate that S-curve it would be better.

 

[bigpicture7]

Hi. This was always kind of my baby, so I might as well reply...

Riverside, when you're running your calculations I think it's important to consider a Oak Square via Common Ave and Boston College via Comm Ave as one line together, not two. Demand dramatically drops off beyond Harvard Ave, hence why the idea of either a stub extension to Union Square or a Griggs Street yard have always been such good ideas. The only reason every B train goes all the way to BC is because the T has no other good option to turn trains before there. (The packards corner turnback is useless for all but emergencies). The fatal flaw in your numbers is that Harvard Ave - Chestnut Hill Ave does NOT need full rapid-transit headways, even at rush.
. . .

Davem, thank you for all of the thought you've put into this. In the spirit of future-proofing, though, I am wondering if you've considered the dramatic number of housing units being added in the vicinity of the Washington St. stop on Comm Ave? I agree, based on long time riding, that demand dramatically drops off after Harvard Ave., but during rush (pre-pandemic) it was still moderately busy up until Washington St....then, yes, a huge drop off the rest of the way to BC. Even though up-to Washington was busy, I still don't dispute, under existing conditions, what you claim. However, there are no fewer than 7 substantial residential developments (approved or u/c) within 3 blocks of Washington St. (along Comm. and along Washington), all of which cite the Washington St. stop in their transit impact studies. They're all posted here on aB. A rough guesstimate is 3,000 new units among them. Anyway, maybe you're already aware, but just thought I'd bring attention to it.

 

[Riverside]

Hey @davem, always great to hear from you. Still love those maps, all these years later. I think the tl;dr here is that I've basically concluded that the problem I saw is one that can be solved relatively modestly -- setting a Kenmore surface terminal aside, I think some judicious bus lanes to provide short-turn service would be enough to solve any real problems. And then separately a South Station-Nubian connector (which I do still feel needs to be in the mix for consideration).

In practical terms, I think we're largely in agreement here, at least in terms of the top priority builds. My whole point earlier this year was to widen the conversation to include some consideration for the additional problems I laid out, and I think at this point it's clear that there are several possible solutions, of varying Craziness™.

I have some responses to your points, but let me be clear that these are generally speaking academic concerns -- not intended to be voracious disagreement, just some hashing out of details.

Riverside, when you're running your calculations I think it's important to consider a Oak Square via Common Ave and Boston College via Comm Ave as one line together, not two. Demand dramatically drops off beyond Harvard Ave, hence why the idea of either a stub extension to Union Square or a Griggs Street yard have always been such good ideas. The only reason every B train goes all the way to BC is because the T has no other good option to turn trains before there. (The packards corner turnback is useless for all but emergencies). The fatal flaw in your numbers is that Harvard Ave - Chestnut Hill Ave does NOT need full rapid-transit headways, even at rush.

The point about treating Oak Sq and BC as one common (ha!) branch is well-taken. And for sure, the current stops at Harvard Ave, Allston St, and Warren St would all see a significant siphoning off to a resurrected A Line. But I'm not sure I agree that Washington St, Sutherland Road and Chiswick Road are as anemic as you suggest here. Looking at the Blue Book numbers, they all do better than most C Line stops (Washington St in particular). In fact, I'd argue that the current B Line is underserved -- if anything, the demand west of Harvard Ave seems "normal" and the demand to the east seems like overload.

And while I agree that a resurrected A/B pair wouldn't require the same frequencies as today, my point is that, absent any supplemental source like Kenmore, you rapidly (boy am I full of puns today) end up with extremely low frequencies -- standing on the platform waiting for 12-15 minutes (especially when you remember trains coming much more frequently Way Back When) creates a death spiral for transit usage. I understand that you don't believe this will be a problem, but I was concerned that there wasn't a good solution if you're wrong.

But that's where short-turns come in -- maybe you boost frequencies via a Kenmore surface terminus, or maybe you acknowledge that riders don't really care whether their ride runs on asphalt vs rails and offer better integration of short-turn buses and thru-running trains. So, even in that worst case scenario, the problem can be addressed.

I was thinking about this over the weekend (working on a related project), and it occurred to me also that the A and B lines are not interchangeable in this scenario: the A has two generations of forced transfers at Kenmore that it has become (somewhat) accustomed to, which means if they get a thru trolley every 12 minutes and a short-turn bus every other 12 minutes, it'll still be an improvement in service. So, that's another possible "relief valve" if degradation of service on the B becomes a concern -- prioritize frequency maintenance on the B.

(And I think that [degradation of service] is non-trivial -- the whole point of Reconfiguration is to improve the Green Line, and I think it'll be hard to convince people that "less frequent trains = improvement," even if rationally we knew that the corridor didn't need those frequencies.)

A harvard ave branch is similarly going to peel demand off inbound comm ave trains (with people instead going outbound to transfer to the red line), as are the allston and brighton CR stops. Speaking of a harvard branch - it wouldn't be too much more trouble to keep running under the pike after the BU Bridge, and have it tie into the C instead of the B. That way you could have the "H Line" just bounce back and fourth between Harvard and Kenmore. This way it isn't fed from Park at all. A lot of rush hour load is also going to be taking the CR from the Brighton and Allston stations, bypassing the light rail system altogether.

See, this is why I wanted to bring up the topic -- I think hooking a Harvard Branch into the Beacon subway is a really interesting idea, and wasn't something I had thought of before! I think I'd want to see a close analysis of the 3D geometry there... that's a fair amount of spaghetti to contend with, and especially if you wanted to avoid a flat junction. But you're not wrong -- that would not be a huge realignment, and would also allow you to short-turn any "Urban Ring" trains coming in from Grand Junction.

With respect to improvements to the B/C/E's street-running reliability -- no arguments in terms of the value of those improvements, but I still think that the B and C (and any resurrected A) are fundamentally different in character from the rest of the network, due to their close stop spacing and shorter travel distance. This has knock-on effects on everything from boarding methods, station design, walkshed tolerance, to overall travel time and journey planning. What's best for the B/C won't always be the best for the rest of the system, and the rest of the network is only going to become more like the D as time goes on. That's why I see value in treating them like a separate network as much as possible, in terms of the 100-year vision.

By the way -- that other project that I alluded to (still in progress) draws heavy inspiration from your J/K/L network. I've reworked it a fair bit, so it may not necessarily be recognizable, but I'm definitely drawing on your ideas there! So thanks again for that, and all these other great ideas over the years.

 

[F-Line to Dudley]

Also, a station at Porter would be incredibly expensive, since it would have to be under the commuter rail station.

Or tunnel under Somerville Ave, with an underground station under Somerville Ave next to the commuter rail station. I'll try to lay that out tonight.

No, it wouldn't be incredibly expensive...and no, you would never opt for a Somerville Ave. dig over the Fitchburg for any reason ever. There's no utility layer under the CR tracks like there is in a 15-25 ft. 'sandwich' layer under every street. Utility relocation and abatement is frequently the most expensive part of subway construction, and that is not operative here if you stay under the RR. Staying under the RR means portaling-down at the Beacon St. overpass when the quad-track Fitchburg ROW runs out of real estate, and shimmying under the tracks in a shallow box tunnel with the roof forming the new Fitchburg track bed. That's it. No utility layer, no utility relocations. The only station costs are outfitting the new GL level, refitting the replacement CR level, and ramping-up/down egresses to each of them at the current lobby entrance. Red mezzanine would handle all station access as before with one simple change of the faregates being rearranged to put the current-CR/future-GL access door behind prepayment.

You could possibly even value engineer a total rip-the-band-aid shutdown of the Fitchburg Line for a 6-12 month box tunnel construction phase to save even more while speeding the construction. Ayer-Wachusett are accessible to full schedules run off the Lowell Line and Pan Am Stony Brook Branch if they were expressed Anderson<==>Lowell before taking their speed penalty on the Stony Brook. Littleton-Porter can subsist as a dinky shuttle to the Red Line for the year via a temporary single-side platform erector-setted the other side of Mass Ave., and by utilizing the Alewife Maintenance of Way facility for storage and layover plug-in power while North Station/Boston Engine Terminal are severed.

Extending Union-Porter with 1-2 intermediates is not even as expensive as equivalent-distance chunking on the current GLX, because there's no deep ROW pit here needing mass-rearranging except for the final 1000 ft. and no Community Path anywhere here. It's not a megaproject; the "extremely expensive" scaremongering is lazy at best, disingenuous at worst.

 

[Charlie_mta]

The only station costs are outfitting the new GL level, refitting the replacement CR level, and ramping-up/down egresses to each of them at the current lobby entrance. Red mezzanine would handle all station access as before with one simple change of the faregates being rearranged to put the current-CR/future-GL access door behind prepayment.

Yes, as long as the new GL station center platform (under the Fitchburg Div) can be built around the supports of the existing station building (and more support added as needed). Its hard to tell without an actual design, but I'm guessing it can probably be done.

 

[F-Line to Dudley]

Yes, as long as the new GL station center platform (under the Fitchburg Div) can be built around the supports of the existing station building (and more support added as needed). Its hard to tell without an actual design, but I'm guessing it can probably be done.

The stilts of the building are 4 tracks wide, because it bowlegs over the ancestral ROW which was quad all the way through Belmont Center. The tunnel footprint would only be somewhat wider than the island platform here, with staging not needing to touch the building stilts. The only question is whether you do some manner of delicate construction staging so 1 temp track of the Fitchburg can stay in-service through the duration, or sever the whole works for a year with "rip the band-aid" approach and alternate shuttle routes for the outbound stops.

Some additional space on the extension gets saved by the Green Line's electrical being tied into the existing Red Line/TT substation in that overhang building (the vent windows seen on the photo), meaning you can anchor the power draw for the extension with distributed power upgrades to the Porter + Alewife RL sub power rooms and the next-nearest GLX sub rather than needing to construct an all-new substation. That's a significant cost saver for this particular extension being able to span existing 600V interconnects rather than create all-new substation sources like the current GLX has had to.

It's also potentially feasible to have the tunnel construction intentionally lower the level of the Fitchburg trackbed a couple feet before reconstructing it on the roof, making the Porter-Beacon St. canyon here able to be capped over at sidewalk level for an air rights park (and, crucially, a direct Somerville Ave. driveway for Lesley U./Porter Exchange so inadequate Roseland St. no longer has to be tasked with acting like a thoroughfare between Beacon & Mass).

 

[RandomWalk]

There’s a center pier on the Mass Ave bridge to contend with, unless you are expecting the tunnel box to stop short of it.