Green Line Reconfiguration

[737900er]

Except there are plans to rebuild the whole area

I thought the current project was to pretty much rebuild the existing mess as is.

 

[Riverside]

Dipping back into this convo for a hot second:

My objective with “Norumbega Junction” is two-fold:

1) Provide a transfer between Green and Indigo/Regional Rail that does not vulture frequencies on the Indigo/Regional Rail service

2) Provide an endpoint for short-turn services on the main line, without disrupting through-traffic and without relegating the short-turns to a low frequency branch terminus

Those objectives are intended to maximize network connectivity and efficiency, and are to some extent agnostic to the location.

Providing PnR relief to Riverside and other 128 PnRs is a nice-to-have, as is increased service to Weston. My focus is on what can/should be done here in order to improve the entire network, not just serve this local area.

As I’ve said previously, there are other solutions here. Setting Framingham instead of 128 as the short-turn “Urban Rail” terminus (which I’d argue can be justified) is one option. Extending the Green Line to Auburndale (kludgey but potentially feasible) could be another piece of the puzzle. An easily-navigated system of 128 business shuttles that link Weston/Waltham, Auburndale, Riverside, and the 128 business parks could be another — ensuring that MetroWest commuters have a non-automobile option for their journeys.

 

[Equilibria]

At this point, nothing but vague notions that bridges are nearing the end of life. When Allston first came on the radar screen, there was talk about this intersection being next.

That's not accurate. The plans to replace the bridges are specific and at (I believe) 75% design. Nothing about the interchange will meaningfully change, and the new bridges lock in the current layout (unfortunately) for decades to come.

 

[Highwayguy]

I know nothing about this area, do the plans have provisions to solve the ramp/traffic issues F-Line identified, or is that something that's not in the plans (regardless if we wished it might be)?

The 128/pike interchange is programmed and is between the 25% and 75% stages. Major reconstruction/ moderate reconfiguration of the interstate to interstate ramps and the Pike bridge over the Charles is called for, but access to the surface roads is pretty much going to remain as is. Overall interchange operations should be improved by the elimination of a few key bottlenecks, slightly improving surface street operations by reducing the spillback queues. There is also a separate project to reconstruct the Rt30/ River Rd signal, but l don’t know if its even at 10%, let alone programmed.

The Recreation Rd 128 onramp already provides excellent pike WB access (needs to be slightly modified if the superstation was built though), but an EB offramp is much tougher unless it was constructed to meet Orchard Ave (good luck with that).

Side note: Amongst those in the know, much better odds of this interchange job being completed prior to grounding the Allston viaduct rather than the other way around.

 

[Riverside]

The emerging common wisdom in this thread is that the best way to bring LRT to the Seaport is a subway that parallels the NEC between Washington St and South Station, originating in the general vicinity of Tufts Medical Center. This subway would connect to a subway from Back Bay and to a resurrected Tremont St Subway. An Aldgate junction built at “Bay Village” would enable services via all three combinations: Boylston-Back Bay, Back Bay-Seaport, Seaport-Boylston.

I’ve been thinking more about that last service, originating north of Park Street, coming down to Bay Village, hooking around, and coming back out to South Station and then the Seaport. My opinion on this has evolved over the years, from initial incredulity to what I’d call a healthily skeptical acceptance. But now my opinion has swung back in the other direction; I no longer think it’s a good service pattern or adequate solution to provide access to the Seaport, though I think that in practical terms it doesn’t change much as far as construction and planning.

The key argument was that a one-seat ride from the Green Line downtown stations would be convenient enough (and speedy enough) to draw riders away from the double-transfer journey required if accessing the Seaport via a Red Line trip to South Station. A Malden commuter, for example, would be able to make a single transfer for the Green Line at North Station, as opposed to transferring at Downtown Crossing and traveling one stop on the Red Line. I’m no longer convinced that the convenience would be high enough to draw riders away from the double-transfer-to-South-Station option.

First, I looked again at the length of the journey: 1.4 miles on Park-Bay Village-South Station. That’s nearly the distance from Park to Hynes, and with three times as many curves (albeit with fewer stations). That’s in contrast to the .5 miles on the Red Line; the Green Line is timetabled at 9 minutes between Park and Hynes today, meaning that it would almost be faster to walk from Park Street to South Station.

But the part that really moved me was considering the frequencies of this wraparound service. I’ve rambled at length about the relative frequencies of LRT services through Park Street, and I don’t intend to rehash that now. Rather, I’ll simply point out that Red Line headways pre-pandemic were 4.5 minutes, and my recollection is that the goal of the Transformation project is to reduce those even lower.

It’s extremely unlikely that a Seaport Green Line service through Park St would be afforded sufficient slots to beat that Red Line frequency. (Recall that Park Street needs to maintain strong frequencies to Nubian, Back Bay/Huntington, and Kenmore [plus Kenmore’s branches], in addition to whatever it devotes to the Seaport.) It is very likely that a transfer via the Red Line – even with the wait time – will be faster enough of the time to wash out most potentially redirected journeys.

Finally – and this is less of a problem, but still a drawback – due to the frequency issue, it’s unlikely that more than one northside Green Line branch would through-run to the Seaport. So most Green Line riders would need to transfer as well. And coming from the south, we run into an additional problem: in its current configuration, Boylston can’t support the free northbound-southbound transfer that would be needed to accommodate Kenmore-Seaport journeys – riders will need to travel to Park St and doubleback, or the T will need to adopt a POP system that allows riders to exit to street level, cross and re-enter (which again will put this service at a disadvantage against a double-transfer via Red).

However, there is an option that would provide extremely high-frequency service, would travel direct from transfer points to the Seaport, would serve new and desired destinations along the way, could provide a one-seat ride to both the Seaport and the Airport, and could provide feeder service to multiple corridors outside of the core: the T7/Center City Link/Navy Line. The Downtown-Seaport corridor deserves its own trunkline, and a full-build modern BRT network can achieve that purpose.

There is – I grant – one non-trivial downside to using a Congress St route for transfers to the Seaport: Green Line short-turns at Park Street would still be stuck with a double-transfer. As it happens, I actually have a solution for that (more on that later), but the reality is that a Navy Line and a Park-Seaport branch of the Green Line could coexist – the point is that the Seaport Green Line branch needs to be relieved of the burden of being the primary connection, but it still can be a useful service (and there almost certainly could be capacity for it – but only at branch-level frequencies).

Moreover, enabling a Park-Seaport branch from a build perspective requires actually very little dedicated construction. The only additional requirement it adds is a leg of the Aldgate junction at Bay Village, which could either be included in the initial build for operational flexibility, or could be easily provisioned for and built at a later date.

So, my conclusion is that a Park-Seaport branch may still be valuable, but is insufficient. I think it's worth provisioning for it to be used in the future (particularly since it actually wouldn't require that much in terms of construction), but I've stopped prioritizing it in my crayon maps, and believe that there are better solutions more worthy of our time, attention, and ultimately investment.

 

[JeffDowntown]

I would point out that your hypothetical Malden commuter would stay on Orange all the way to Tufts, then transfer to any Green Seaport Direction service at the Tufts/Bay Village super station.

So the headway in that case is better. But it also raise the question of why include the Park/Seaport branch (and its schedule challenges in the Central Subway) if a simple transfer to the much higher frequency Orange at Tufts/Bay Village mirrors the service.

 

[Riverside]

I would point out that your hypothetical Malden commuter would stay on Orange all the way to Tufts, then transfer to any Green Seaport Direction service at the Tufts/Bay Village super station.

So the headway in that case is better. But it also raise the question of why include the Park/Seaport branch (and its schedule challenges in the Central Subway) if a simple transfer to the much higher frequency Orange at Tufts/Bay Village mirrors the service.

In theory, North Station will offer an easier Green-Orange transfer than Tufts/Bay Village, because it's a cross-platform transfer at North Station, vs what'll probably be at least one flight of stairs/escalator and a lengthy corridor at Tufts/Bay Village. But in general I agree with you: frequencies to the Seaport will always be highest at Tufts/Bay Village, so it will naturally form a strong transfer hub.

The reality is that Park-Seaport service helps serve the Seaport, but isn't uniquely well-suited, and arguably has a smaller impact than it looks on paper. Going one-by-one ("OSR" = one seat ride, "OOS" = out of system transfer):

​

Origin​	Wraparound​	Congress BRT & Back Bay-Seaport LRT​	Winner?​
Orange Line north, northside commuter rail, all but one Green Line north branch, Haymarket buses​	(Cross-platform) transfer at North Station (or Haymarket), every ~6 min​	transfer at street-level, every ~3 min​	Congress BRT/BBY-Seaport LRT, though both are viable​
One Green Line north branch​	OSR direct​	transfer at street-level, or at Bay Village, every ~3 min​	wraparound, though both are viable​
Red Line north​	transfer at South Station, every ~3 min​	same​	no difference​
Blue Line west​	transfer at Government Center, every ~6 min​	transfer at State street-level, every ~3 min​	Congress BRT/BBY-Seaport LRT​
Green Line branches from Kenmore​	transfer at Park St (or OOS at Boylston), every ~6 min​	if turning at North Station, transfer at street-level, every ~3 min; if turning at Park, double transfer via Red​	wraparound​
Orange Line south, Green Line branches from Huntington via Back Bay, Green Line from Nubian​	OSR direct and/or transfer at Bay Village (or Back Bay), every ~3 min​	same​	no difference​
Red Line south, Indigo Line, southside commuter rail​	transfer at South Station, every ~3 min​	same​	no difference​
Blue Line north​	transfer at Government Center, every ~6 min, OR transfer at Airport to SL3, every ~10 min​	transfer at State street-level, every ~3 min, OR transfer at Airport to SL3, every ~10 min​	Congress BRT/BBY-Seaport LRT​
northeast quadrant (Chelsea & Everett)​	OSR direct via SL3, every ~10 min or (hypothetically) OSR direct via Green Line, ~every 6 min​	OSR direct via SL3, every ~10 min or (potentially) OSR via BRT 111, every ~10 min​	draw​

So, that's one northern branch of the Green Line that gets favored via the Wraparound, plus the Kenmore trains coming in to Park Street.

Which brings me back to my original point: if you're an Allston/Brighton/Brookline commuter heading to the Seaport, and you've been deposited at Park Street, which are you going to do?:

1) Wait 6+ minutes for a OSR on the Green Line, which itself will take 7-10 minutes just to reach South Station
2) Wait 3 minutes for the Red Line, travel for 3 minutes to South Station, and wait 2-3 minutes for a Seaport LRT
3) Take advantage of your monthly pass's de facto free OOS transfer and walk the 9 minutes to South Station on foot

(The numbers for Options 1 and 3 change a little bit if this commuter can transfer at Boylston, but it's not a huge difference.)

Will some riders use Option 1? For sure. But I think enough will continue to use Option 2 that the impact simply won't be transformative. And the scenario I outlined above (how to go Kenmore - Seaport) can also be avoided if you can bring BBY-Seaport LRT up to meet the Green Line branches somewhere out in Brookline/Allston/Brighton. (Several possible ways to accomplish this, discussed elsewhere in thread, in the realm of possibility though still a tall order.)

Plus, minimum viable Congress BRT could literally be built in the next two weeks if they chose to. (Thank you, Orange Line Shutdown.)

So, yes. If there's spare capacity, I think a Park-Seaport service is fine. But the combination of Congress BRT and BBY-Seaport LRT is stronger, both for the Seaport and for the network overall.

I will say this though: the Wraparound concept is brilliant and innovative, and opened the doors to a wide range of possibilities that simply weren't in discussion 10 years ago. The problem of accessing the Seaport which the Wraparound tried to solve is very much a real one and will only get more urgent in the coming years. I ultimately feel the Wraparound is not the right answer, but it definitely asks the right questions.

 

[Brattle Loop]

The key argument was that a one-seat ride from the Green Line downtown stations would be convenient enough (and speedy enough) to draw riders away from the double-transfer journey required if accessing the Seaport via a Red Line trip to South Station. A Malden commuter, for example, would be able to make a single transfer for the Green Line at North Station, as opposed to transferring at Downtown Crossing and traveling one stop on the Red Line. I’m no longer convinced that the convenience would be high enough to draw riders away from the double-transfer-to-South-Station option.

First, I looked again at the length of the journey: 1.4 miles on Park-Bay Village-South Station. That’s nearly the distance from Park to Hynes, and with three times as many curves (albeit with fewer stations). That’s in contrast to the .5 miles on the Red Line; the Green Line is timetabled at 9 minutes between Park and Hynes today, meaning that it would almost be faster to walk from Park Street to South Station.

But the part that really moved me was considering the frequencies of this wraparound service. I’ve rambled at length about the relative frequencies of LRT services through Park Street, and I don’t intend to rehash that now. Rather, I’ll simply point out that Red Line headways pre-pandemic were 4.5 minutes, and my recollection is that the goal of the Transformation project is to reduce those even lower.

I may well be wrong, but my gut instinct is that people are generally lazy and will prefer to minimize transfers (especially longer and/or more crowded transfers) even at the expense of travel time. You're likely right that a double-transfer via a frequencies-enhanced Red might well result in a lower-aggregate-travel time than a GL-Seaport branch. As it stands now, though, an Orange Line rider (from anywhere, if at least one service ran Seaport-North Station) would go from two transfers to one, a north-side Commuter Rail rider on anything but Fitchburg would go from 3 to 1, GL western branch riders would go from 2 to 1 (possibly a bit distended at Park) or 0 (assuming some kind of BBY-Seaport service, not to mention simplifying the transfers from the Blue Line without relying on the SL3. Some of where the passengers go probably comes down to whether they care more about getting their the fastest/most efficient way, or if they care less that lower frequencies takes longer because it means they have to change less often. (Not to mention that some of the changes would be significantly easier than current access to Red.)

Relatedly, and while I think you have (per usual) a good, thorough, and interesting take on service to the Seaport, but the Seaport isn't the only (or necessarily the biggest) thing a GL-Seaport branch gets you; it adds a currently-nonexistent RT link to South Station. Right now the Seaport Silver Line passengers (excepting anyone who takes the counter-intuitive SL3 route to transfer to the BL at Airport) heading anywhere on the rest of the system get dumped in with all the CR passengers and everyone else relying on the Red Line. The forced transfer means extra passengers transferring at and clogging up the stairs and platforms at Park and DTX who don't actually need to transfer at those specific places. The Red Line Transformation project might well aim at improving the track and signal infrastructure to lower headway with the new cars, but that won't mean much in terms of service improvements if the trains are still getting clobbered with nasty dwell time penalties at Park and DTX, and the RLT project didn't include provisions for rebuilding those stations (or for building Red-Blue to pull those pax out to a different transfer). I think that element is at least as significant as the Seaport element, and more in favor of building it than just the Seaport alone.

 

[Riverside]

I may well be wrong, but my gut instinct is that people are generally lazy and will prefer to minimize transfers (especially longer and/or more crowded transfers) even at the expense of travel time. You're likely right that a double-transfer via a frequencies-enhanced Red might well result in a lower-aggregate-travel time than a GL-Seaport branch. As it stands now, though, an Orange Line rider (from anywhere, if at least one service ran Seaport-North Station) would go from two transfers to one, a north-side Commuter Rail rider on anything but Fitchburg would go from 3 to 1, GL western branch riders would go from 2 to 1 (possibly a bit distended at Park) or 0 (assuming some kind of BBY-Seaport service, not to mention simplifying the transfers from the Blue Line without relying on the SL3. Some of where the passengers go probably comes down to whether they care more about getting their the fastest/most efficient way, or if they care less that lower frequencies takes longer because it means they have to change less often. (Not to mention that some of the changes would be significantly easier than current access to Red.)

Relatedly, and while I think you have (per usual) a good, thorough, and interesting take on service to the Seaport, but the Seaport isn't the only (or necessarily the biggest) thing a GL-Seaport branch gets you; it adds a currently-nonexistent RT link to South Station. Right now the Seaport Silver Line passengers (excepting anyone who takes the counter-intuitive SL3 route to transfer to the BL at Airport) heading anywhere on the rest of the system get dumped in with all the CR passengers and everyone else relying on the Red Line. The forced transfer means extra passengers transferring at and clogging up the stairs and platforms at Park and DTX who don't actually need to transfer at those specific places. The Red Line Transformation project might well aim at improving the track and signal infrastructure to lower headway with the new cars, but that won't mean much in terms of service improvements if the trains are still getting clobbered with nasty dwell time penalties at Park and DTX, and the RLT project didn't include provisions for rebuilding those stations (or for building Red-Blue to pull those pax out to a different transfer). I think that element is at least as significant as the Seaport element, and more in favor of building it than just the Seaport alone.

Potentially more to say later, but to be clear, I’m still very much in favor of an LRT subway going Seaport - South Station - Bay Village - Back Bay - Points West. That builds rail rapid transit along a strong straightshot axis that strings together multiple major destinations, and provides good access to the Seaport from the south.

From the north, full-build BRT (physically separated lanes, prepayment, raised platforms, shelters) along Congress St gets you the same-if-not-better access as the Wraparound does, at higher frequencies and (depending on which routes feed into Congress) OSRs from a larger number of northside corridors than a single northside GL branch. See the chart in my second post: the OSR convenience factor only impacts a fraction of potential riders; everyone else has at least one transfer, and Congress can offer significantly more frequent and probably faster service on that front.

(FWIW… if the concern is BRT vs LRT, it occurs to me that Congress BRT would have several potential connection points where it could connect to the LRT network. I personally think it’s overkill, and that an advantage of Congress BRT is its ability to accept routes from places rail won’t go anytime soon, eg the 111. But, the possibility is there, I suppose!)

 

[JeffDowntown]

Potentially more to say later, but to be clear, I’m still very much in favor of an LRT subway going Seaport - South Station - Bay Village - Back Bay - Points West. That builds rail rapid transit along a strong straightshot axis that strings together multiple major destinations, and provides good access to the Seaport from the south.

From the north, full-build BRT (physically separated lanes, prepayment, raised platforms, shelters) along Congress St gets you the same-if-not-better access as the Wraparound does, at higher frequencies and (depending on which routes feed into Congress) OSRs from a larger number of northside corridors than a single northside GL branch. See the chart in my second post: the OSR convenience factor only impacts a fraction of potential riders; everyone else has at least one transfer, and Congress can offer significantly more frequent and probably faster service on that front.

(FWIW… if the concern is BRT vs LRT, it occurs to me that Congress BRT would have several potential connection points where it could connect to the LRT network. I personally think it’s overkill, and that an advantage of Congress BRT is its ability to accept routes from places rail won’t go anytime soon, eg the 111. But, the possibility is there, I suppose!)

The problem with BRT in Boston is we all know what happens to a BRT pitch. You get value engineered into a bus stuck in traffic. No elevated boarding, no enforceable bus lanes, just a bus.

 

[Riverside]

The problem with BRT in Boston is we all know what happens to a BRT pitch. You get value engineered into a bus stuck in traffic. No elevated boarding, no enforceable bus lanes, just a bus.

I'm not indifferent to this concern, but I like to think that it is solvable. (There's a tragic backwards logic to it: "Let's propose the more expensive solution so that the less expensive solution doesn't get watered down into uselessness!") Like I mentioned above: minimum viable Congress BRT is achievable in a matter of days if they chose to. Start with enforced bus lanes, reroute the 4, and maybe extend the 7 to North Station (both adjustments aiming to be rolling stock neutral), and start advertising. From there, you have an established service that you can continue to enhance step by step.

 

[Brattle Loop]

I'm not indifferent to this concern, but I like to think that it is solvable. (There's a tragic backwards logic to it: "Let's propose the more expensive solution so that the less expensive solution doesn't get watered down into uselessness!") Like I mentioned above: minimum viable Congress BRT is achievable in a matter of days if they chose to. Start with enforced bus lanes, reroute the 4, and maybe extend the 7 to North Station (both adjustments aiming to be rolling stock neutral), and start advertising. From there, you have an established service that you can continue to enhance step by step.

Yeah, some of my commentary was based along the lines of "don't give them ideas", so that they don't just send silver buses down Congress and call that all the service needed. Congress BRT, I should note, is a good idea for additional service, but there are far too many examples of times when the powers that be have decided that 'additional'-caliber services should just serve the trunk roles. I think Seaport LRT + Congress BRT would provide good, complementary service options.

 

[737900er]

South Station-Seaport LRT is probably worthy of more than one branch frequency, especially considering it wouldn't share platforms with the BRT. If not from Park, those trains would have to come from somewhere else, which leads to the question of what that would be and if any of them would be better than Park.

Obviously one option would be just South Station-Seaport short turns.

Should half/most Huntington Subway service run to Seaport instead of Park? I don't think so given the historical connections Highland and Huntington have to Park Street

What kind of contortioning would be required for another/enhanced Aldgate to enable Nubian-Seaport. Would two branches to Nubian be overkill? Running Nubian to Park and South Station seems a little more logical to me.

 

[Riverside]

South Station-Seaport LRT is probably worthy of more than one branch frequency, especially considering it wouldn't share platforms with the BRT. If not from Park, those trains would have to come from somewhere else, which leads to the question of what that would be and if any of them would be better than Park.

Obviously one option would be just South Station-Seaport short turns.

Should half/most Huntington Subway service run to Seaport instead of Park? I don't think so given the historical connections Highland and Huntington have to Park Street

What kind of contortioning would be required for another/enhanced Aldgate to enable Nubian-Seaport. Would two branches to Nubian be overkill? Running Nubian to Park and South Station seems a little more logical to me.

I have grander designs in progress (will share soon enough), but a few points:

• Huntington can take higher frequencies than it has now, all the more so if the subway is constructed
• The demand profile between Riverside/Needham and downtown (via Huntington) probably merits a layer-on short-turn service, on top of the stuff coming in from the Highland Branch
  • For example, a layer-on service originating in the Seaport, traveling down Huntington and beyond to Reservoir, and cutting across Chestnut Hill Ave to reach Boston College, giving that end of Brighton a OSR to the Seaport
• Think of the Seaport transitway as a trunk of its own, and pick a target frequency (e.g. 15 trains per hour, for four-minute headways), and from there consider where to send services
• Nubian-Seaport might be worthwhile, but the contortion comes with the station design at Bay Village -- enabling a transfer between Nubian-Seaport and Orange would either require a double-back (head east to the transfer statoin before doubling back west to go to South Station), or would require an additional set of platforms to the east of the Orange Line (and most/every design/sketch I've seen of a Green-Orange transfer at Bay Village puts the Green platforms to the west)
  • If your Nubian-Park frequencies are high enough on their own, you certainly could run excess Seaport trains down, bypassing Bay Village altogether, though it's not clear if that's the best use of those trains

For my money, send Seaport trains to Huntington (Longwood) and then see where to go from there. Devote slots in the Transitway to creating a high-freq high-capacity spine between Seaport, South Station, Back Bay, Pru, and LMA, linking together all four of Boston's "downtowns", as well as two major railroad stations, plus a same platform transfer from Logan.

Combine that with 8-15 tph coming in from the Highland Branch (and maybe Heath), and now Huntington + Back Bay has doubled east-west LRT capacity from today, with a train under Huntington as often as we have trains under Boylston now.

 

[The EGE]

I've been playing with the two-trunks model (Boylston and Bay Village) following Riverside's exploration. I'm going to make a couple self-indulgent posts about it. I've been trying to keep to some general rules:

• No more than 30 TPH on any single track, or through any single switch. This would allow fully automatic train operations.
• Grade separate junctions whenever possible
• No less than 7.5 TPH (8-minute headways) on any given service
• Complete separation of Boylston and Bay Village trains - an issue on one trunk line should not cascade into the other
• Trains headed the same place should share a platform
• 300-foot platforms
• Reuse the existing flying junctions as much as possible

Unsurprisingly, Bay Village once again becomes the critical infrastructure to figure out. If you want both Nubian and Seaport branches, then you have four lines feeding in, with at minimum four pairs of routings (the Back Bay - downtown - Seaport triangle, plus downtown to/from Nubian). You're also severely limited by the Orange Line and the Pike cut, both of which block routings at certain levels. It's helpful to think of there being four possible levels:

• Ground level (blue on my maps)
• Shallow (green on my maps) - rails about 20 feet below the surface. This is the level of the Pike cut, and able to pass over the Orange Line.
• Mid (yellow/orange on my maps) - rails about 40 feet below the surface. This is the level of the Orange Line, and able to pass under the Pike.
• Deep (red on my maps) - rails about 60 feet below the surface. This is able to pass under the Orange Line.

Immediately we see two things: Seaport-Back Bay tracks have to be either be shallow or deep to avoid the Orange Line, and downtown-Nubian tracks have to be deep to pass under the Orange Line and Pike - or be on the surface.

There's also several decisions to make:

• Is Nubian essential? It's the fly in the ointment from a technical perspective, but it's highly desirable to improve transit on the corridor.
• Are we willing to tunnel under the Orange Line? This would be the most technically demanding aspect, but without it you probably can't have everything grade-separated.
• Does everything have to be grade-separated? Highly desirable given the frequencies, but makes everything else more difficult.
• What lines stop at Bay Village? You don't have to stop at all - the Orange Line is right there, and you'll have a stop for the Nubian line nearby - but both Boylston and Park suck for opposite-direction transfers.. If you do stop, you can either have platforms only for the Nubian-downtown and Back Bay-Seaport routings (which gives you transfers between Nubian and most points), or platforms for every routing.

I've chosen to only include designs that have a Nubian branch, which drives some of the other decisions. I don't believe that a full-transfer configuration is feasible; even if you forgo grade separation, the platforms have to be after lines join, which means your transfers are pretty terrible. Having the Nubian branch run under the Pike is less than ideal - that means you're three levels from the street and two from the Back Bay-Seaport platform. My best effort with that so far involves the Nubian platform having left-hand running, with an additional flyover required:

If you're willing to have a level junction, you can have all Seaport trains serve Bay Village:

There are variations on that theme, like dropping the Seaport-bound tracks to the deep level (with a triple-stacked platform) for grade separation. It's also possible to shift the surface and shallow levels north (assuming a modernization of the Josiah Quincy School) to ease the curves.

If you just have transfers between downtown-Nubian and Back Bay-Seaport, then you get to my preferred design. Nubian platform over the Pike, with the Back Bay-Seaport platform under Marginal. You do have to close off Marginal (except for pedestrians) and Oak at Shawmut, but I don't think that's the end of the world. The only tricky part is the Seaport-bound track from Boylston, which has to drop under the Orange Line.

I tried playing with no-station configurations, but they don't have any real advantage - there still has to be something that dips below the Orange Line.

 

[The EGE]

So now that we've gone to this insane civil engineering effort, what does it get us? This:

At full build-out, you have two east-west trunklines with 30 TPH, fed by 11 branches. The Boylston Street Subway is fed by the B and C, plus a new branch through Beacon Park to Harvard, all looping on the inner tracks at Park Street. Needham, Hyde Square, Nubian, and Seaport all feed into the outer tracks, through-running to the GLX branches plus the Grand Junction. Because they use entire separate tracks through Park Street and Boylston, each trunk can run at full capacity.

The E Branch is put into a subway from Brigham Circle to Northeastern. A new subway under Longwood Avenue carries the D to Huntington Avenue.

The Marginal Street tunnel, Bay Village junction, Nubian Branch, and connector to South Station have all been discussed here.

The Seaport end has options - Design Center, Southie, even a tunnel to Logan. Here I just show Design Center - Southie will probably use the proposed T7, while a harbor tunnel would be a whole separate project.

There are options for the Harvard branch; this uses Malvern, the planned Stadium Way, and then a tunnel under the Charles to reach the old tunnel box at Harvard. The Grand Junction, as discussed previously, could branch off the Union Square Branch. I've also included an extension to Porter; further extension is of course possible.

Along with Bay Village, this is the other Crazy Engineering Pitch. The Grand Junction dives under Comm Ave, with a 4-track transfer station with the buried B. Follow whatever alignment is plausible to reach Fenway (I show it under Park Drive). A wye near Longwood Avenue joins with the D subway, allowing Grand Junction trains to terminate at LMA. Fenway station could be retained for use by baseball specials from Riverside.

The one thing that's discussed a lot that I didn't include is a branch from Lechmere to Chelsea via Sullivan. There's no good alignment for the inbound track at Red Bridge to run it into the subway, and running it only to the Grand Junction wouldn't have any good transfer to the rest of the line. I think Chelsea would be better served by proper BRT to Sullivan/Kendall/downtown (and eventually true rapid transit), rather than having it awkwardly attached here.

Frequencies could look something like this.
Green Line: every 6 min on each service
A Harvard - Park Street
B Boston College - Park Street
C Cleveland Circle - Park Street

Magenta Line: every 7.5 min on each service
D Riverside - Seaport
J Needham Junction - Porter
E Hyde Square - Mystic Valley Parkway
K Hyde Square - Seaport
F Nubian - LMA via Grand Junction
L Seaport - LMA via Grand Junction

Combined headways are 3 min (20 TPH) between Packards and Kenmore, 2 min (30 TPH) Kenmore to Park. 4 min (15 TPH) Newton Highlands to MFA, 4 min Hyde Square to MFA, 2 min MFA to Back Bay, 2 min Boylston to Lechmere. South Station to Seaport would have 22.5 TPH; you could probably add an additional South Station-Seaport shuttle if 30 TPH on even headways was desired.

 

[Riverside]

Love it @The EGE, and I'm tickled that my work was a useful jumping-off point. A lot of great ideas in here, and a lot of convergence with ideas I've been refining -- in fact, I have a draft right now called "Simplifying the Green Line Reconfiguration" that looks remarkably similar to what you've put together here. I need to draw out a few visualizations for that post, but will try to share this weekend.

I'm mulling how much to respond to right now during my lunch break versus later on today...

IIRC you and I have gone back-and-forth previously about the "Seaport wraparound"; I remain skeptical (as described upthread here and here) but we probably don't need to rehash it here. I do have some sketches for (I think) a simpler design for Bay Village that becomes a bit easier (and a bit more transfer-friendly) if we remove Boylston-Seaport from scope -- I'll find those and post them later.

Regarding Longwood: I like the way you've approached this, and appreciate that you are giving Longwood radial service rather than solely circumferential service, as I waxed eloquent about over the summer. In that same post, I outlined a potential surface route using separated lanes and ROWs to cross Longwood -- I've noodled on that more and more in recent weeks, and actually see it as an increasingly reasonable and high-impact solution -- minimal Crazy Engineering required. It would lead to a slightly different network topology than what you've outlined here, but the ideas are similar for sure.

More later!

 

[Aprehensive_Words]

Nice stuff!

FWIW, one argument in favor of the third Tufts-adjacent stop configuration is that it sets you up to deck over that entire block of the Pike -- the concrete and disruptions needed to build that bridge aren't too far removed from what you'd need to put in the footings a la South Station for a full deck -- creating an opportunity to sell off air rights to help finance the project or generate more affordable housing.

 

[Riverside]

The Bay Village junction offers probably the most potential and likewise the most complexity of the whole shebang. Putting the Nubian platform under the Pike is an idea I hadn't thought of before. Yes on the one hand I agree that that platform is problematically deep, in terms of street access and transfers; on the other hand, it's the first design I've seen that would provide a street-level entrance south of the Mass Pike, which is quite tempting.

One variable that I think is worth giving more consideration than we usually do is staging: it's more likely that Bay Village will get built in stages, and so it's worth coming up with designs that can be built in stages and still be viable/sensible (i.e. low-impact future-proofing). Particularly given that Green-to-Nubian is the most "realistic" aspect of this proposal (at least in that it is on the mainstream radar), I think it is hard to swing making its share of Bay Village the most complex.

Back in the spring, I put together a design that was meant to achieve all the objectives you put out, and also be friendly to phased construction over the course of a couple of decades. The final build is detailed here and some options for interim builds is discussed here. In all cases, I've placed the ("north-south") subway station under Tremont St, which has a few benefits:

• modestly wide street with somewhat set-back buildings (at least one set of which is not historical, which does provide some additional flexibility during construction)
• well-aligned for continuing to Back Bay, and better positioned to partially fill the gap between Tufts and Back Bay by expanding the walkshed
• well-aligned for a gentle curve around to align with an extra-wide segment of the NEC to use for a portal without disrupting neighborhoods on Washington
  • using the westernmost tracks at the station creates sufficient running space to descend under the Pike without requiring the platform to be excessively deep

In general, the benefit of the "Tremont" station is that it can serve both Nubian and Back Bay, on their own and in tandem, regardless of which gets built first.

(The 1-over-3 design that I'm using means that Park trains from both branches board on the same platform, albeit different tracks.)

The downside of the design I came up with is that Back Bay-Seaport LRT trains can't stop at Bay Village. I explained in my previous posts why I felt that was an acceptable compromise. However, in the intervening months, I've concluded that I don't believe Park-Seaport wraparound service is worth prioritizing (upthread here and here).

That's led me to a relatively minor redesign that adds a "Marginal" station for Back Bay-Seaport service; this becomes feasible by eliminating the grade changes needed to accommodate the Park-Seaport leg of the flying junction -- creating more "running room" to enable a level platform at "Marginal".

This design ends up being relatively simple:

• Nubian service ducks deep under the Pike and uses the lower level of "Tremont" station
• A simple flying junction allows Park and Seaport services to merge going to Back Bay
• Back Bay-Park service uses stacked platforms on the upper and lower levels of "Tremont" station, consolidating all Park-bound trains on a shared platform
• Seaport service uses a shallow platform at "Marginal" station

Aside from Nubian service, all tracks are built shallow just below street-level, except for a divedown on the westbound track for the flying junction and the lower level of "Tremont" station.

For transfers: the "Green-Aqua" transfer would be something like 150' or 200' between platforms, which is vaguely on-par with the Red-Orange transfer at DTX; the "Aqua-Orange" transfer would be something like 450', which is on the long side but not as long as Southbound-Orange-Blue at State (~600'). The "Green-Orange" transfer is long, at 600', and is comparable to the Winter Street Concourse, but is mitigated in part by strong transfer opportunities at North Station and Back Bay.

And just for good measure: as you can see, this design still enables Park-Seaport service if you really want it, via surface tracks accessible through a Y on the Nubian branch. This is probably too slow if Park-Seaport is a "load-bearing" service, but as I've argued previously, I don't think Park-Seaport will be suitable for that purpose under any design, so it's not worth trying to fit the square peg into the round hole.

In general, I think this design enables reasonable transfers, modest construction impacts and costs, staged construction that doesn't require full buy-in, no grade crossing, and minimal speed-reducing-geometries. The caveat to all this is that I'm always second-guessing myself on the grades here -- I've gone through the numbers at least three times and so I'm pretty sure that everything fits, but I still wish I were more confident.

More thoughts on systemwide design later today.

 

[Riverside]

Below is the post I'd been drafting for a little while, extending many of the core concepts that @The EGE laid out:

• No more than 30 TPH on any single track, or through any single switch to allow fully automatic train operations.
  • Achieved, with some stretches where potentially higher volumes could be accommodated, but not required
• Grade separate junctions whenever possible
  • Achieved in core; some junctions in the "Inner Belt" zone may be level
• No less than 7.5 TPH (8-minute headways) on any given service
  • Yes, and though I won't really go into it here, I think a takt around 7 or 8 minutes (7.5 or 8 tph) sits in a sweet spot and is useful to apply systemwide -- the knock-on effects of combining trains in increments of 15 tph are actually very beneficial; most importantly, I think 8-minute headways represent a tipping point where riders won't bother waiting for a dedicated service, causing a ridership death spiral
• Complete separation of Boylston and Bay Village trains - an issue on one trunk line should not cascade into the other
  • Yep, I've used Emerald and Green for this as opposed to Green and Magenta but the concept is identical
• Trains headed the same place should share a platform
  • Largely achieved; Bay Village has three different platforms heading to Back Bay (including Orange), and there is a short stretch of Huntington where you could catch an inbound Aqua Line train on the surface, or go downstairs to catch a Green Line train
• 300-foot platforms
  • I admit I haven't checked this carefully but I think is achievable in most places
• Reuse the existing flying junctions as much as possible
  • Yup, and I put the Kenmore Loop into full-time revenue use as well

Longwood:

One key point of convergence is our use of a cross-Longwood corridor for combined radial and circumferential service. I lay out here the details on how I would achieve this: basically reclaim half of Park Drive (north of Fenway station) and all of Fenway (the road, between Fenway station and Huntington) as dedicated transit ROW, with Forsyth St converted to a transit mall. I love @The EGE's idea for a tunnel, especially one that serves both radial Highland service and circumferential Grand Junction service, and indeed I think my system can map quite elegantly into that build. The more central location is very tempting. I think the surface routing along Fenway still provides significant value, but at a fraction of the cost. (And allows for the inclusion of BRT service -- i.e. to Logan Airport -- as well.)

Chelsea:

The one thing that's discussed a lot that I didn't include is a branch from Lechmere to Chelsea via Sullivan. There's no good alignment for the inbound track at Red Bridge to run it into the subway, and running it only to the Grand Junction wouldn't have any good transfer to the rest of the line. I think Chelsea would be better served by proper BRT to Sullivan/Kendall/downtown (and eventually true rapid transit), rather than having it awkwardly attached here.

I haven't given up the ghost on Grand Junction-Chelsea service; a stop at McGrath Hwy/Twin Cities Plaza can provide an interchange with a Union/Porter branch. What's more, I believe Grand Junction really covers two overlapping corridors: a north-south Sullivan-Kendall-Longwood corridor, and an east-west Lechmere-Kendall-Allston corridor. I run Gold as the former, and Green as the latter, meaning Chelsea commuters could transfer at McGrath or Cambridge St to double back.

That being said, I do agree that Chelsea (and Everett, to a lesser extent) are best served by radial BRT, circumferential LRT, and electrified express mainline rail. Hence my use of a "Navy Line" BRT service running through Congress St downtown.

Riverside:

Shout-out to @737900er (I think it's you?) for suggesting Riverside service could be short-turned at Kenmore, an idea I've adopted here (with allowance for small amounts of through-running).

Silver Line:

Though not my main focus, I've done a couple of "unusual" things with the Silver Line. I'm skeptical we'll ever get LRT through the Ted Williams Tunnel in my lifetime, but I'm somewhat more optimistic about dedicated bus lanes (or at least partially restricted); that means that we will need BRT for at least some of our circumferential service, so let's make the best of it.

The Silver Line's role in this system is serving Logan from non-Downtown locations: Kenmore, Longwood, Nubian, Seaport, Chelsea, Everett, Wellington; it is a sister service to the Gold Line (Gold and Silver ring, get it?), with enough overlaps at the end to minimize transfers for short/medium journeys. With the possible partial exception of the TWT, the Silver Line is designed to run on roads with enough space for dedicated (usually center-running) separated bus lanes.

Inspired indirectly by the revised Bus Network Redesign, this Silver Line design utilizes center-running bus lanes on Albany St to create a speedy bypass from the southern quadrant to the eastern quadrant; this allows us to "cheat" and add an extra radial line running from South Station to Nubian, even though the Silver Line is overall functioning as a circumferential service. Previously I'd looked to South Boston Haul Road for this purpose, but it has awkward connections to other services; Albany, by contrast, directly serves Boston Medical Center, and can connect to South Station and then leverage the surface BRT infrastructure on Summer St.

Running via Albany also allows the current SL3 service to be stitched into a longer more cohesive corridor (extending from South Station to BMC and Nubian), providing Chelsea commuters with OSR service to more locations.

And now, without further ado...