Green Line Reconfiguration

[bigeman312]

I dont disagree, and nothing in your post is inconsistent with what I said.... Hyde Square isn't Arborway and isn't really Centre St either, save for a few feet. How about eminent domain for the 711 and have it go to Canary Square? Maybe not so crazy and if there is the width to do it (which there might be, if you had S Huntington + 711 parking lot + Centre)....

I was using Hyde Square to refer to the general area (which includes Canary Square), not a specific intersection. Terminating anywhere in that area (be it S Huntington & Perkins, S Huntington & Centre, Perkins & Centre, Barbara St, Day St, etc) would qualify as an extension to Hyde Square. I’m not advocating for a specific iteration of that extension, just for that type of extension in general, which would be very popular among JP residents if built.

What I wrote wasn’t intended to be inconsistent with what you wrote. Just further context for the discussion.

 

[Riverside]

I've been trying to write a version of this post for, like, a month now, and it just keeps on getting away from me. So I'm going to try to just get something of it out now -- I can always come back to the thread and add more later!

A little less than a year ago, I took my suggested model for a Green Line Reconfiguration network and flipped it: instead of sending Kenmore trains to Park Street Inner Loop and splitting Huntington trains between Downtown and the Seaport (the "Tripod" model herein), the "Criss Cross" model sends all Kenmore trains to the Seaport and all Huntington (+ Nubian) trains through Downtown (using all four tracks at Park Street). At the time, I still somewhat preferred "Tripod", though there were things I liked about Criss Cross; I also felt that Criss Cross had some major drawbacks.

Since then, my thinking has somewhat shifted, both in that I've become more skeptical of certain aspects of Tripod, and more optimistic about Criss Cross. For example, Tripod's reliance on isolating the B/C on Park St Inner raises questions in my mind about "graceful degradation": I think the Inner Loop can support two full-time branches, but what if I'm wrong? In that scenario, you would need to start intermixing the streetcar-subway routes with the dedicated-ROW routes, and then we're sorta just back where we started.

For Tripod to work as intended, the B/C need Very High Reliability to be able to squeeze through the single loop. But, if the B/C can achieve Very High Reliability, then they could be viable for providing service to the Seaport; one of my biggest objections to Criss Cross was "if we're going to build this modern subway to the Seaport, why kneecap it with unreliable streetcar-subway routes?", but that sets up a false dichotomy. In either model, the B/C need Very High Reliability and the model needs to be accommodating when their reliability falls short. In that respect, Criss Cross outstrips Tripod because it truly does isolate the Kenmore services.

That isolation pays dividends elsewhere, too. Criss Cross allows all four tracks at Park St to be devoted to Huntington + Nubian frequencies, setting the ceiling higher than anything Tripod could do. Criss Cross enables additional branches from Kenmore (e.g. through-running to Harvard, which @Teban54 has noted seems to be very promising), and ultimately means that you get truly double the capacity of today's subway. (And, indeed, this is me coming full circle: the exact same argument -- splitting Kenmore and Park services into entire separate trunks to create two separate lines with maximal capacity -- was the topic of one of my first deep dives into GLR details, four years ago; Criss Cross recapitulates that same concept.)

There are some other upsides that I see in Criss Cross, and I might post about those later. But the key piece here is that "Tripod" vs "Criss Cross" is a false choice: the vast majority of GLR projects are agnostic to the Tripod vs Criss Cross model. And in fact, nothing in either model needs to preclude the option for subsequent conversion to the other. Which is to say, a First Generation GLR might look like Tripod, and eventually be expanded into a Second Generation version that looks like Criss Cross.

Alright, time for some pretty diagrams. (Well, some diagrams. Beauty is in the eye of the beholder.) Lower-res screenshots here for ease of quick reading, full images attached.

View attachment Green Line Reconfiguration Services Diagram Current State v1.1.png

View attachment Green Line Reconfiguration Services Diagram Tripod v1.1.png

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

And a system diagram (with some minor differences from the diagrams above; this diagram, for example, assumes a subway through the South End):

View attachment Project Criss Cross Minima v1.3.3.png

And a sketch of how the "Copley Criss-Cross" would work; unlike last year, where I suggested using the short segment of subway under Exeter Street, here I suggest simply swinging out into Copley Square and continuing down Dartmouth. The existing Copley station would close but would be able to remain in situ. And, in a key difference from last year, here I propose having separate stations at Copley and Back Bay. Yes, they'll be very close together. I think that's okay.

 

[ritchiew]

I've been trying to write a version of this post for, like, a month now, and it just keeps on getting away from me. So I'm going to try to just get something of it out now -- I can always come back to the thread and add more later!

A little less than a year ago, I took my suggested model for a Green Line Reconfiguration network and flipped it: instead of sending Kenmore trains to Park Street Inner Loop and splitting Huntington trains between Downtown and the Seaport (the "Tripod" model herein), the "Criss Cross" model sends all Kenmore trains to the Seaport and all Huntington (+ Nubian) trains through Downtown (using all four tracks at Park Street). At the time, I still somewhat preferred "Tripod", though there were things I liked about Criss Cross; I also felt that Criss Cross had some major drawbacks.

Since then, my thinking has somewhat shifted, both in that I've become more skeptical of certain aspects of Tripod, and more optimistic about Criss Cross. For example, Tripod's reliance on isolating the B/C on Park St Inner raises questions in my mind about "graceful degradation": I think the Inner Loop can support two full-time branches, but what if I'm wrong? In that scenario, you would need to start intermixing the streetcar-subway routes with the dedicated-ROW routes, and then we're sorta just back where we started.

For Tripod to work as intended, the B/C need Very High Reliability to be able to squeeze through the single loop. But, if the B/C can achieve Very High Reliability, then they could be viable for providing service to the Seaport; one of my biggest objections to Criss Cross was "if we're going to build this modern subway to the Seaport, why kneecap it with unreliable streetcar-subway routes?", but that sets up a false dichotomy. In either model, the B/C need Very High Reliability and the model needs to be accommodating when their reliability falls short. In that respect, Criss Cross outstrips Tripod because it truly does isolate the Kenmore services.

That isolation pays dividends elsewhere, too. Criss Cross allows all four tracks at Park St to be devoted to Huntington + Nubian frequencies, setting the ceiling higher than anything Tripod could do. Criss Cross enables additional branches from Kenmore (e.g. through-running to Harvard, which @Teban54 has noted seems to be very promising), and ultimately means that you get truly double the capacity of today's subway. (And, indeed, this is me coming full circle: the exact same argument -- splitting Kenmore and Park services into entire separate trunks to create two separate lines with maximal capacity -- was the topic of one of my first deep dives into GLR details, four years ago; Criss Cross recapitulates that same concept.)

There are some other upsides that I see in Criss Cross, and I might post about those later. But the key piece here is that "Tripod" vs "Criss Cross" is a false choice: the vast majority of GLR projects are agnostic to the Tripod vs Criss Cross model. And in fact, nothing in either model needs to preclude the option for subsequent conversion to the other. Which is to say, a First Generation GLR might look like Tripod, and eventually be expanded into a Second Generation version that looks like Criss Cross.

Alright, time for some pretty diagrams. (Well, some diagrams. Beauty is in the eye of the beholder.) Lower-res screenshots here for ease of quick reading, full images attached.

View attachment 60762

View attachment 60757

View attachment 60763

View attachment 60758

View attachment 60764

View attachment 60759

And a system diagram (with some minor differences from the diagrams above; this diagram, for example, assumes a subway through the South End):

View attachment 60765

View attachment 60760

And a sketch of how the "Copley Criss-Cross" would work; unlike last year, where I suggested using the short segment of subway under Exeter Street, here I suggest simply swinging out into Copley Square and continuing down Dartmouth. The existing Copley station would close but would be able to remain in situ. And, in a key difference from last year, here I propose having separate stations at Copley and Back Bay. Yes, they'll be very close together. I think that's okay.

View attachment 60761

This looks really interesting! The "Tripod" plans always seemed to have some odd routing and reverse branching that made it hard to see where the clear benefits would be. This "Criss Cross" plan definitively fixes a lot of that. (But I'm still kind of mulling it over)

There are a couple of sections a little unclear, though. Could you explain the Yellow Line routing from Back Bay to South Station? Is that just going under Marginal/Mass Pike/Surface Road/Essex? And what does the Brookline Village intersection look like? Sorry if I'm just missing something.

 

[Riverside]

This looks really interesting! The "Tripod" plans always seemed to have some odd routing and reverse branching that made it hard to see where the clear benefits would be. This "Criss Cross" plan definitively fixes a lot of that. (But I'm still kind of mulling it over)

Yeah. Tripod would probably work better if it tried to do one fewer thing. For example, if it were only the B or C running on the Inner Loop, then I might be less skeptical. Or if there weren't a branch to Nubian, or if it didn't need to handle a branch to Heath St/Hyde Square. Remove any one of those, and I think you could swing it.

I think the concept maybe comes through a little more clearly in the diagram I made for the Transportation Dreams contest:

Standalone, it's a concept for a subway that centers on Longwood + Back Bay, with separate branches to Downtown and the Seaport, which I think is reasonable enough and is a simple enough trunk-and-branch system. But the reality is that the overall network would look like what I diagrammed in my last post, with distinct-but-not-fully-separate systems that are forced to intertwine.

There are a couple of sections a little unclear, though. Could you explain the Yellow Line routing from Back Bay to South Station? Is that just going under Marginal/Mass Pike/Surface Road/Essex? And what does the Brookline Village intersection look like? Sorry if I'm just missing something.

You aren't missing anything -- those parts are actually somewhat imprecise by design

In general, yes, the Gold Line hugs the northern side of the Mass Pike, running in a shallow tunnel under Marginal Road until Harrison Ave. From here, the exact alignment will need to be determined through more detailed study. One potential option might be to thread under the highway ramp spaghetti (potentially even in an open cut) before tunneling under Lincoln to Essex with a sharp hook around to meet the existing Transitway tunnel under Atlantic.

A similar alignment could go under Hudson, or could do some zig-zagging under Washington and Kneeland. Increasingly, I think that this segment may be best addressed by "biting the bullet" and using a TBM. This area is a real headache beacuse you have to be very three-dimensional to thread around the Big Dig tunnels. But, ultimately we're talking about somewhere between a half mile and full mile of difficult tunneling that "unlocks" a system that is mostly composed of extant and easier tunneling.

(This is the part that stresses me out the most, in particular the intersection of Lincoln, Essex, and Surface Road [with the highway tunnel underneath]. I think there are enough options in this general area I'm pretty sure something will work, but I wish I had something other than a TBM that I felt fully confident in. The only saving grace is that basically any possible way to hook into the Transitway will face the same difficulties; at least the Green Line Reconfiguration provides the benefit of tying it into a systemwide rework like the Green Line Reconfiguration allows you to articulate the benefits across a much larger region.)

Spoiler: the underlying difficulties of hooking into the Transitway

For those who are curious, the Transitway creates two pinchpoints for connecting it to the larger network. First, the existing Transitway's loop goes about as far south under Atlantic Ave as you can go, because the highway tunnel underneath is rising as it heads south, and there's not enough headroom to extend the Transitway further south above the highway. So you need to turn; you could turn east under South Station, but obviously that's pointing in the wrong direction, so mostly that means you turn west down Essex (which was provisioned for in the original tunnel construction). The problem with Essex is that just two blocks down, you're faced with the other highway tunnel, running directly underneath Surface Road. So, either you need to dive down pretty sharply as you traverse Essex St, or you need to do a very tight turn on to Hudson, Lincoln, or South St... or you need to mess with the highway tunnel itself, somehow. But, in any case, if you want to hook into the Transitway, it needs to be via Essex St [or somehow under South Station, which would be its own can of worms], and if you need to reach Essex St, you need to contend with the highway tunnel under Surface Road.

(My good Lord, would this be easier if the Big Dig had been built with consideration for Seaport transit to pass through the Big Dig's build area, and not just terminate at South Station. [Silver Line Phase III ran into the same problems I describe above. There just wasn't adequate consideration given, IMO.])

As for Brookline Village, there's a similar story in that the imprecision is intentional so as to not get bogged down in making premature decisions about alignments. There, however, I'm fully confident that there are options that wouldn't break the bank (many of which are discussed up thread). I do see something confusing in my diagrams, though, which is the name "Brookline Village": there are a range of ways to connect D + E, and they vary in cost and complexity (and in terms of how they interact with potential designs for subway{s} through Longwood more directly).

Depending on the selected alignment, it's quite possible that Criss Cross's Huntington-Highland Subway would intersect the Heath-Kenmore branch somewhere on the above map but not actually that close to the current Brookline Village station. So I can see how that would be a mildly confusing simplification on my part.

 

[JeffDowntown]

Yeah. Tripod would probably work better if it tried to do one fewer thing. For example, if it were only the B or C running on the Inner Loop, then I might be less skeptical. Or if there weren't a branch to Nubian, or if it didn't need to handle a branch to Heath St/Hyde Square. Remove any one of those, and I think you could swing it.

I think the concept maybe comes through a little more clearly in the diagram I made for the Transportation Dreams contest:

View attachment 60795

Standalone, it's a concept for a subway that centers on Longwood + Back Bay, with separate branches to Downtown and the Seaport, which I think is reasonable enough and is a simple enough trunk-and-branch system. But the reality is that the overall network would look like what I diagrammed in my last post, with distinct-but-not-fully-separate systems that are forced to intertwine.

You aren't missing anything -- those parts are actually somewhat imprecise by design

In general, yes, the Gold Line hugs the northern side of the Mass Pike, running in a shallow tunnel under Marginal Road until Harrison Ave. From here, the exact alignment will need to be determined through more detailed study. One potential option might be to thread under the highway ramp spaghetti (potentially even in an open cut) before tunneling under Lincoln to Essex with a sharp hook around to meet the existing Transitway tunnel under Atlantic.

View attachment 60797

A similar alignment could go under Hudson, or could do some zig-zagging under Washington and Kneeland. Increasingly, I think that this segment may be best addressed by "biting the bullet" and using a TBM. This area is a real headache beacuse you have to be very three-dimensional to thread around the Big Dig tunnels. But, ultimately we're talking about somewhere between a half mile and full mile of difficult tunneling that "unlocks" a system that is mostly composed of extant and easier tunneling.

(This is the part that stresses me out the most, in particular the intersection of Lincoln, Essex, and Surface Road [with the highway tunnel underneath]. I think there are enough options in this general area I'm pretty sure something will work, but I wish I had something other than a TBM that I felt fully confident in. The only saving grace is that basically any possible way to hook into the Transitway will face the same difficulties; at least the Green Line Reconfiguration provides the benefit of tying it into a systemwide rework like the Green Line Reconfiguration allows you to articulate the benefits across a much larger region.)

Spoiler: the underlying difficulties of hooking into the Transitway

For those who are curious, the Transitway creates two pinchpoints for connecting it to the larger network. First, the existing Transitway's loop goes about as far south under Atlantic Ave as you can go, because the highway tunnel underneath is rising as it heads south, and there's not enough headroom to extend the Transitway further south above the highway. So you need to turn; you could turn east under South Station, but obviously that's pointing in the wrong direction, so mostly that means you turn west down Essex (which was provisioned for in the original tunnel construction). The problem with Essex is that just two blocks down, you're faced with the other highway tunnel, running directly underneath Surface Road. So, either you need to dive down pretty sharply as you traverse Essex St, or you need to do a very tight turn on to Hudson, Lincoln, or South St... or you need to mess with the highway tunnel itself, somehow. But, in any case, if you want to hook into the Transitway, it needs to be via Essex St [or somehow under South Station, which would be its own can of worms], and if you need to reach Essex St, you need to contend with the highway tunnel under Surface Road.

(My good Lord, would this be easier if the Big Dig had been built with consideration for Seaport transit to pass through the Big Dig's build area, and not just terminate at South Station. [Silver Line Phase III ran into the same problems I describe above. There just wasn't adequate consideration given, IMO.])

As for Brookline Village, there's a similar story in that the imprecision is intentional so as to not get bogged down in making premature decisions about alignments. There, however, I'm fully confident that there are options that wouldn't break the bank (many of which are discussed up thread). I do see something confusing in my diagrams, though, which is the name "Brookline Village": there are a range of ways to connect D + E, and they vary in cost and complexity (and in terms of how they interact with potential designs for subway{s} through Longwood more directly).

View attachment 60799

Depending on the selected alignment, it's quite possible that Criss Cross's Huntington-Highland Subway would intersect the Heath-Kenmore branch somewhere on the above map but not actually that close to the current Brookline Village station. So I can see how that would be a mildly confusing simplification on my part.

Quick note on the Chinatown Gold Line connection to the transitway. This has been reviewed up thread many times (sorry cannot find the reference). You have to access the transitway via Essex street -- you can not come at it from any location further south. There is no way to get across the tunnels of the Big Dig to get to the transitway from the south that will get you to the level of the transitway.

So you have to thread your way up to Essex Street west of the Big Dig area, via Harrison, Hudson, etc. That is the transitway access point planned for the cancelled Silver Line Phase III tunnel.

 

[Riverside]

Quick note on the Chinatown Gold Line connection to the transitway. This has been reviewed up thread many times (sorry cannot find the reference). You have to access the transitway via Essex street -- you can not come at it from any location further south. There is no way to get across the tunnels of the Big Dig to get to the transitway from the south that will get you to the level of the transitway.

So you have to thread your way up to Essex Street west of the Big Dig area, via Harrison, Hudson, etc. That is the transitway access point planned for the cancelled Silver Line Phase III tunnel.

Yes, to clarify, I was suggesting using a TBM to reach Essex St, to use the same insertion point as the other options.

 

[ritchiew]

Standalone, it's a concept for a subway that centers on Longwood + Back Bay, with separate branches to Downtown and the Seaport

Centering transportation on Longwood/Back Bay makes a certain amount of sense. But putting a trunk there and Downtown on a branch always seemed fundamentally flawed. Downtown is by far the biggest destination, and so shouldn't be on a branch. In that Tripod model, too many people waiting for a train along the Huntington trunk will have to wait for the correct train. That effectively halves the useful frequencies for lots of people. That also adds wait time on any transfer on a branch line. I'd also expect the Downtown-bound trains to be crowded and the Seaport-bound trains underutilized. (I think this is textbook "reverse branching," but it's a little harder to say when you've effectively got a network with multiple trunks.)

Your Criss Cross plan looks obviously better to me because it avoids a lot of that. For most people going in most directions, riders can just hop on whichever train is next, and they'll be headed in the right direction. People along the B or C branches might miss their OSR to Park Street, but if frequencies are good and the transfer is easy, that doesn't matter much.

I think that this segment may be best addressed by "biting the bullet" and using a TBM. This area is a real headache

I'll add a couple more potential problems with TBMs here (a little bit overlapping with what you already mentioned). First, TBMs are good on straight lines but tend to have wide minimum turning radii, especially for the larger bores you'd need for trains. Other people here will know more than I do, but I've seen a turning radius of 10x the bore diameter flagged as a real achievement, and that was in really optimal tunneling conditions. The turns tends to be wider than that, and will be in our suboptimal soft landfill. With a probable turn radius of several hundred feet, you wouldn't be able to do the turns you drew. It's unlikely you could do any turn onto Essex.

Second, the connection to the Transitway at South Station is extremely shallow. That would mean the TBM couldn't go too deep. Running a TBM closer to the surface level increases risk for anything nearby. It also wouldn't go deep enough for hard rock and would be going through a lot of soft landfill. All of that adds a lot of risk, plus expense in risk mitigation. TBMs have gotten a lot better at this in recent years, with all kinds of tricks. But this is a challenge.

To be clear, I'm not doubting we could build a tunnel from Back Bay to South Station. Of course we can. I would love to hear people's ideas on how to do this cheapest.

Depending on the selected alignment, it's quite possible that Criss Cross's Huntington-Highland Subway would intersect the Heath-Kenmore branch somewhere on the above map but not actually that close to the current Brookline Village station. So I can see how that would be a mildly confusing simplification on my part.

I see. Yeah, that's where I got confused. And that's a more minor part of the plan, so hand waving there is fine by me.

 

[TheRatmeister]

More thoughts later, but what about revisiting some of the early 1900s planning done about an extension to Post Office Square from Arlington? The interchange at Boylston would be a bit challenging but maybe there's something there.

As for the South Station connection, somewhere probably in this thread I laid out a plan that uses surface streets for such a connection. Maybe it's worth revisiting that idea in some form as well.

 

[JeffDowntown]

Yes, to clarify, I was suggesting using a TBM to reach Essex St, to use the same insertion point as the other options.

Understood -- but you need to be aligned from the west down Essex street (for a few blocks) to get there. The Transitway is quite shallow -- and shallow depths are fully compromised by the Big Dig tunnels south of Essex Street -- basically a classic "you can't get there for here" situation. Otherwise you are using a TBM to destroy a Big Dig tunnel.

 

[Riverside]

Centering transportation on Longwood/Back Bay makes a certain amount of sense. But putting a trunk there and Downtown on a branch always seemed fundamentally flawed. Downtown is by far the biggest destination, and so shouldn't be on a branch. In that Tripod model, too many people waiting for a train along the Huntington trunk will have to wait for the correct train. That effectively halves the useful frequencies for lots of people. That also adds wait time on any transfer on a branch line. I'd also expect the Downtown-bound trains to be crowded and the Seaport-bound trains underutilized. (I think this is textbook "reverse branching," but it's a little harder to say when you've effectively got a network with multiple trunks.)

This is a point that would be worth closer studying and modeling. In principle, I agree with you, but I think it's worth highlighting that South Station (on the Seaport branch) is arguably the better access point to Downtown -- the central employment axis of downtown is along Congress St, not Tremont St:

[analysis is from a post I did a couple of years ago, where, coincidentally, we also talked about a surface alternative to South Station, as @TheRatmeister was mentioning]

So I think it's possible that at least some Huntington -> Downtown riders (under Tripod) would be "opportunistic" and just take the first train that comes; for a large swath of commuters, it's a 5-10 minute walk from Park Street anyway, so a 5-10 minute walk from South Station wouldn't be any worse. Also worth noting that even "half frequencies" along Huntington would still be something like 4 min headways. And the dedicated ROWs elsewhere on Tripod's Gold Line network would better support longer trainsets than today, so crowded platforms would clear out that much faster.

But yes, to your point: Criss Cross is just simpler overall.

I'll add a couple more potential problems with TBMs here (a little bit overlapping with what you already mentioned). First, TBMs are good on straight lines but tend to have wide minimum turning radii, especially for the larger bores you'd need for trains. Other people here will know more than I do, but I've seen a turning radius of 10x the bore diameter flagged as a real achievement, and that was in really optimal tunneling conditions. The turns tends to be wider than that, and will be in our suboptimal soft landfill. With a probable turn radius of several hundred feet, you wouldn't be able to do the turns you drew. It's unlikely you could do any turn onto Essex.

Second, the connection to the Transitway at South Station is extremely shallow. That would mean the TBM couldn't go too deep. Running a TBM closer to the surface level increases risk for anything nearby. It also wouldn't go deep enough for hard rock and would be going through a lot of soft landfill. All of that adds a lot of risk, plus expense in risk mitigation. TBMs have gotten a lot better at this in recent years, with all kinds of tricks. But this is a challenge.

To be clear, I'm not doubting we could build a tunnel from Back Bay to South Station. Of course we can. I would love to hear people's ideas on how to do this cheapest.

Re turning radii: to clarify (and to @JeffDowntown's point), my suggestion of TBM would assume a very different alignment, something more like this:

In this example, I've suggested cut-and-cover (blue) to handle a sharper turn under Harrison, then switching to TBM (red) for a much longer curve up to Essex. To your point, I'm not sure what the best way to tunnel under Essex would be (TBM vs C&C), though my vague recollection is that it wasn't specifically the tunneling under the highway that posed so many difficulties. (The good news is that Essex St isn't on landfill, so that simplifies things a little bit.)

But yeah, there's no denying that this part will be challenging. That's why I think it's important to articulate a value proposition that stretches system-wide -- i.e. articulate that building this one particularly difficult tunnel would allow expansion of light rail to Needham, Harvard, Nubian, and enable high frequencies to Longwood. Plus the benefits to the Seaport itself, and relief on the Red and Orange Lines to provide additional access to South Station.

(I think at one point I was looking at stealing a couple of tracks from South Station, using them to dig out a pit, and then access the transitway from the east. You'd need to do some work to thread the line from Marginal Road through the highway spaghetti into the South Station approach, but I think that part would be doable. I think I concluded that it wasn't viable, though I don't remember why.)

More thoughts later, but what about revisiting some of the early 1900s planning done about an extension to Post Office Square from Arlington? The interchange at Boylston would be a bit challenging but maybe there's something there.

I think it would still face a lot of the challenges faced by the Silver Line Phase III alignment, wouldn't it?

As for the South Station connection, somewhere probably in this thread I laid out a plan that uses surface streets for such a connection. Maybe it's worth revisiting that idea in some form as well.

I remain skeptical for all the usual reasons, but agree that the cost-benefit analysis changes if we're talking about services that already run at street-level elsewhere. If some streets could be reclaimed into transitways, and maybe if you could reclaim most of Essex St to add a portal...

I sketched this up to capture the idea of separate single-track routes. The aim is to isolate tracks from crossing high-traffic segments, which it mostly succeeds at, except going southbound at Kneeland St. This design steals a lane/parking from Atlantic Ave and descends on the "roof" of the underlying highway tunnel to join up with the extant transitway subway's loop. Coming the other direction, much of South Street plus the eastmost block of Essex St, are pedestrianized with a sealed single track median running down the middle. This would have crossing gates at pedestrian crossings. The surrounding blocks would either be pedestrianized or converted to dead-ends. I'd want to put crossing gates at Beach St too but I'm not sure if traffic levels would permit that, and in turn I'm not sure whether Beach St could be pedestrianized along these two blocks. The reworked intersection at Kneeland would align the east-west pedestrian crossing cycles at Atlantic Ave with the train crossing cycles. Track continue south within "railroad land" south through the South Station approach area.

Alternatively, maybe you could just steal a lot of space along Atlantic (probably including the sidewalk):

Like I said, this is the part that stresses me out the most.

 

[Riverside]

I'm gonna stop now (for my own sanity as much as anything else) but this perhaps seems like the least painful:

1. Portal + median on a pedestrianized Essex St (I guess it doesn't even have to be pedestrianized though)
2. Traffic light
3. Run along the western side of Surface Road where there's wider sidewalk and parking, with semi-sealed ROW
4. Add some sort of pedestrian crossing option at Beach St
5. Traffic light
6. Steal space on Albany Street and the ramp to I-90 until you're able to cut over to Hudson St; relocate the ramp slightly to the south which I think would be doable
7. Portal + median on pedestrianized Hudson St

 

[Teban54]

Hey guys, I'm back! I've sadly been away from all the excellent discussions (both in this thread and elsewhere on archBoston) the last few months, as my IRL schedule simply didn't allow catching up on everything on this forum. I'll gradually dig through several months of archives, so you can wish me good luck

Back to the subject matter, here's what I like about Criss-Cross (in decreasing order of importance):

• Kenmore, the B, and to a lesser extent the C, are no longer reduced to a "shuttle". Especially given how insanely dense the B's walkshed is!
• Full frequencies on the Seaport Transitway
• Much better transfer between the Central Subway routes and the Huntington routes
  • This can be helpful for residents on the westside branches -- Arlington currently has a fair amount of branch-to-branch transfers
• Much less operational reliance on Park St loop, whose capacity is unknown; and allows more operational flexibility
• Potentially open up more branches on the Seaport end in the long term (e.g. Logan Airport, Chelsea, City Point)
• Allows Harvard-Seaport OSR
• Resolves the debate of "Jamaica Plain LRT vs. Huntington reliability"

Here are the main drawbacks and challenges:

• Huntington's frequency is limited to "2 letter bullets" / 3-4 min, as a branch into Tremont
• Loss of South Station - Huntington OSR
  • Mainly affect Old Colony riders, and those coming from Logan Airport (!) and Seaport
• Greater unreliability and uncertainty for Seaport
  • May not be a big issue: the uncertainty exists in the Central Subway regardless, and Criss-Cross makes Seaport trade reliability for frequency, which is well worth it
• The Red-Central Subway transfer point is moved south from Park St to South Station

Frequency on the Huntington Ave subway is really what concerns me the most. By now, I firmly believe that this is the most important tunnel that's yet to be built. LMA alone gets you nearly as many workers as Back Bay, and that's not counting the multiple schools along Huntington Ave itself and MFA, all of which are big destinations. Assuming each "letter bullet" has 6-8 min headways / 8-10 tph, this limits Huntington to 3-4 min headways -- which looks okay until you factor in capacity constraints of LRT. Remember, this trunk will probably carry the equivalent to Cambridge's Red Line in the reverse-peak direction.

You can alternatively send 3 bullets down Huntington, but at the cost of a single bullet to Nubian, which I'm not fond of either. And that's the key issue with Criss-Cross: It forces a binary choice between Huntington and Nubian.

How that compares to Tripod's choice between Tremont and Seaport is anyone's guess, but I'd argue that the latter is not a binary one: Tremont can get supplemental service from Nubian under Tripod, but that can't be said for Huntington under Criss-Cross.

 

[Teban54]

With that said, there is an alternative that lets you add supplemental service to Huntington under Criss-Cross.

Presenting to you, a very quick and rough drawing of the 2x2 / "Full Criss-Cross" model:

Disclaimer: The number of routes ("bullets") per branch matches my intentions. However, the exact routes and terminals are just an approximation, subject to debate, and not the focus of this proposal.

Frequency Assumption: Each letter bullet runs 8-10 tph with 6-8 min headways. (The 6 min / 10 tph scenario is very aggressive and may exceed current capacity limits; the 8 min / 8 tph scenario is much more realistic.) Each track carries no more than 4 bullets, for a maximum tph of 40 (aggressive) or 32 (realistic).


The key idea is a giant interchange in the vicinity of Copley - Back Pay that allows all destination pairings:

• Kenmore - Park St (terminal): Lime, (C)
• Kenmore - Seaport: Silver, (A) (B) (J)
• Huntington - Park St (and north): Green, (D) (E)
• Huntington - Seaport: Magenta, (N)

The primary pairings are exactly like the Criss-Cross plan: Silver and Green. The other two are supplemental pairings, Lime, (C) and Magenta, (N), with only 1 bullet each. They serve the following purposes:

• Add capacity to Huntington -- by utilizing Park St loop for revenue service. Every Lime train terminating at Park St means room for another train on Huntington in the form of Magenta.
• Recover some useful OSRs that are lost under Criss-Cross, for passengers willing to wait for a single bullet (up to 6-8 mins). This applies primarily to Magenta (Huntington - South Station and points east), but also to a lesser extent, Lime (Park St for northside Red Line transfers).

The obvious drawback is (arguably) too much interlining. The system becomes similar to DeKalb and CPW in New York, which have their fair share of operational issues; and erases most benefits of creating two isolated systems, which both Tripod and Criss-Cross achieved.


Note: Theoretically, by moving another Kenmore bullet from Silver to Lime, you can fully restore maximum capacity on Huntington by having two Magenta bullets. I didn't do that here, because it exacerbates the interlining issue even more.

 

[Teban54]

A couple comments on other topics being brought up:

Engineering between Bay Village (Marginal St) and South Station (SL Transitway)

I had previously raised some doubts about this segment in December 2023. At that time, the prevailing proposal -- before most people paid heavy attention to it -- was F-Line's idea from 2014 to 2016, which I illustrated below:

Without the Hudson St issue, on paper it meets most restrictions that were mentioned above. Assuming you can C&C across the Big Dig by trenching under the existing tunnel (which IIRC is what the F-Line proposal implied), this should theoretically stay clear of the tunnel otherwise, though I'm not sure about the Big Dig's exact dimensions near Chinatown. The turn radius from Essex St to the linear park also looks reasonable, especially if C&C can be done.

The main issue is tunneling under Hudson St, which I had elaborated in my Dec 2023 comment. One alternative is to stack the tunnels on top of each other, trading depth for width, and pinching it between the two tall buildings. Whether that's feasible... I don't know.

But my most preferred approach is actually pretty similar to what Riverside said here -- turn onto Lincoln St:

The turn from "Marginal Rd extension" to "Lincoln St extension", under the block of spaghetti highway ramps, should be easily doable. At the very least, there's currently an access road that cuts under the ramps, so a tunnel can be buried directly beneath it. Lincoln St itself is wide enough for a subway, with 60' building-to-building.

My only concern is with the turn from Lincoln to Essex, which is quite tight and may need to interfere with the Big Dig.

Lastly... I myself really don't think that a surface alignment anywhere around here is worth it. That would be like making Red-Blue run in the median of Cambridge St just to save costs.

 

[Teban54]

Centering transportation on Longwood/Back Bay makes a certain amount of sense. But putting a trunk there and Downtown on a branch always seemed fundamentally flawed. Downtown is by far the biggest destination, and so shouldn't be on a branch. In that Tripod model, too many people waiting for a train along the Huntington trunk will have to wait for the correct train. That effectively halves the useful frequencies for lots of people. That also adds wait time on any transfer on a branch line. I'd also expect the Downtown-bound trains to be crowded and the Seaport-bound trains underutilized. (I think this is textbook "reverse branching," but it's a little harder to say when you've effectively got a network with multiple trunks.)

For one thing, even under Tripod, the Tremont St subway (say, North Station to Bay Village) still sees much higher frequencies than the Seaport "branch". This is because Tremont will see supplemental trains from Nubian: in fact, most Tripod GLR proposals, including Riverside's and mine, have just as many trains running BV-GC as there are trains on Huntington.

But even if we focus on the Tremont-BBY-Huntington route itself:

• The Orange Line already duplicates the Tremont St subway heavily, and their north-south segments are virtually interchangeable. I think capacity of Tremont St subway is second-order to capacity of Huntington Ave subway for that reason, if we're only considering how many trains run through that stretch. (Huntington stands out more from OL due to LMA/MFA and Prudential.)
• Any service arrangement, Tripod or Criss-Cross, will already be a net improvement for Huntington-Tremont (including Brookline Village and west). Today, any such station pair only sees one Green Line branch. But any GLR proposal can send two bullets doing Huntington-Tremont alone, doubling the number of trains. Having additional Huntington-Seaport trains doesn't take that away.

Lastly, if Seaport trains being underutilized is your concern, that will only be exacerbated by Criss-Cross -- where all trains from Kenmore feed into Seaport. I personally don't think Seaport trains would be underutilized in either scenario, anyway, especially when considering South Station's utility (as Riverside said) and Old Colony/RL South transfers. South Station has the highest ridership of all rapid transit stations, after all.

 

[Riverside]

Great to have you back! At least some of these points were ones I anticipated

Here are the main drawbacks and challenges:

• Huntington's frequency is limited to "2 letter bullets" / 3-4 min, as a branch into Tremont

....

Frequency on the Huntington Ave subway is really what concerns me the most. By now, I firmly believe that this is the most important tunnel that's yet to be built. LMA alone gets you nearly as many workers as Back Bay, and that's not counting the multiple schools along Huntington Ave itself and MFA, all of which are big destinations. Assuming each "letter bullet" has 6-8 min headways / 8-10 tph, this limits Huntington to 3-4 min headways -- which looks okay until you factor in capacity constraints of LRT. Remember, this trunk will probably carry the equivalent to Cambridge's Red Line in the reverse-peak direction.

You can alternatively send 3 bullets down Huntington, but at the cost of a single bullet to Nubian, which I'm not fond of either. And that's the key issue with Criss-Cross: It forces a binary choice between Huntington and Nubian.

How that compares to Tripod's choice between Tremont and Seaport is anyone's guess, but I'd argue that the latter is not a binary one: Tremont can get supplemental service from Nubian under Tripod, but that can't be said for Huntington under Criss-Cross.

So I actually think this isn't quite as drastic as it seems.

One of the big things to highlight at the outset is the capacity boost from full usage of Park St quad tracks. One of the downsides of Tripod is that you're limited in your ability to use all four tracks flexibly. As a result of this, we've assumed that Huntington + Nubian trains would be capped at 30 tph, operating primarily on the outer tracks, while we try to fit at least 15 tph from Kenmore on the Inner Track -- totaling 45 tph across all four tracks. Now, I've been looking at some older schedules recently, and it does seem like the Green Line did sometimes advertise the equivalent of ~42 tph across all four tracks (90s and early 00s). But I think 45 tph would be a stretch beyond that. If 40 tph is the ceiling, and we need 15 tph to provide 8-min headways to B/C, then that's only 25 tph available for Huntington + Nubian, which caps us out at just under 6 min headways on each. (I also think 40 tph under the Tripod model is far from certain, so 6 min headways is a best case.)

So, that's one "bullet" for Huntington <> Park under Tripod, with room for one or mayyyybe two "bullets" for Huntington <> Seaport. (I'm a little skeptical that a 2-min cadence of Park-Seaport-Seaport-Park-Seaport-Seaport would be doable.)

Under Criss Cross, you're guaranteed 2 full "bullets" for Huntington <> Park and mayyyybe a little extra (see below).

So, to me, the outcomes are about the same: you can guarantee 2 "bullets" (3-4 min headways) for Huntington, and maybe squeeze a little bit more. In terms of squeezing extra capacity, I think Criss Cross's simpler design gives you better chances.

To put it another way:

View attachment Green Line Reconfiguration Capacity Equations.png

What's going on with the [8,15] ranges, you ask? This brings us to your point about Huntington vs Nubian. Under Tripod, the network is complicated enough that simple interval ratios seem necessary to make it viable. For example, giving Nubian and Huntington an equal 15 tph each, or running a 10 tph Huntington - Park service alongside a 20 tph Huntington - Seaport service.

Under Criss Cross, though, I'd argue there's a lot more flexibility because of the full availability of the Park St quad tracks. For example, 25 tph to Huntington and 10 tph to Nubian, if we are concerned about Huntington frequencies, though personally I'd go for the inverse of 20 tph (3 min) to Huntington and 15 tph (4 min) to Nubian. In any case, this flexibility means there's a range of potential frequencies, which the brackets are meant to encompass.

Back to the equations: if we assume equal splits and simple resonances, Tripod and Criss Cross provide the same number of trains to Huntington at the low end of M, with Tripod edging out Criss Cross at high end of M (though see below about 40 tph and 20-25 tph). If we get funkier with the scheduling, both designs can provide 2-min headways on Huntington if M is 40 and Nubian gets minimum frequencies of 8 tph. But if M is 30 (worst case scenario but possible), then Huntington gets stuck with 8-min headways to downtown, which seems like a non-starter. (And would be harder to achieve 3-min combined frequencies with interleaved Seaport trains -- you would either need to live with 4 min, or be okay squeezing one train in 1 minute after the previous. Again, precision scheduling that seems risky under Tripod.)

Overall, my point is that while I agree that Tripod provides some additional options for increasing frequencies by layering in Seaport trains, on the balance I think that Criss Cross's guaranteed 15+ tph branching out of a simple 30 tph trunk is a better bet than Tripod's interline-with-Nubian-and-interline-with-Seaport.

A further note on "cheating" capacity at Park Street to get to 40 tph: today, the Outer Tracks theoretically carry 30 tph. If we imagine that a South End subway is built, and the Nubian Branch receives 20 tph, a Huntington Subway could send 10 tph to the Outer Tracks (matching today's capacity of 30 tph) and send 10 tph to the Inner Tracks to loop. In practice, even I don't advocate for out-and-out 3-min headways to Nubian, so if Nubian instead gets 15 tph (4-min), Huntington could send its own 15 tph on to the Outer Tracks, plus another 10 on to the Inner Tracks, getting you to 2.4-min headways. And, until a subway to Nubian is built, realistically, I think the most it would get is 12 tph (5 min), maybe even 10 tph ( ), which in turn means that Huntington could run 18-20 tph on the Outer Tracks, plus 10 tph to the Inner Loop, which gets you to (just about) 2-min headways along Huntington.

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

20-25 tph: Truth be told, 20 trains per hour (see note below about rolling stock) along Huntington seems pretty reasonable to me. That being said, the clearest example I could find of Tripod increasing frequencies on Huntington was an increase from 20 to 25 tph; this would be a frequency increase from 3 min to 2.4 min. That's not trivial, but I wouldn't say it's transformative, and it's moving toward the level of precision scheduling that I worry about Tripod's ability to deliver.

Maximizing frequencies on Huntington: If we are concerned that 3-min headways would be insufficient, then I think we would need to be talking about a straightforward Highland Branch <> Huntington <> Seaport line, full stop. There are things about such a line that appeal to me, but the lack of Blue or Kenmore transfer really kneecaps it.

Rolling stock: with Type 10 supertrains, each Green Line consist would be able to hold ~800 passengers. I think that 6-car Red Line trains can hold ~1,000 passengers? Which means that Green Line trains at 20 tph/3 min would have a 16K/hr capacity, and Red Line trains at 15 tph/4 min would have a 15K/hr capacity. Again, that seems reasonable to me; if we believe we need the 30 tph/2 min/24K per hour capacity, then I think we're back to the idea of a standalone Huntington <> Seaport line (but that seems like overkill to me).

Supplemental frequencies along Huntington: well, soooo we could potentially solve one problem with another here. Huntington is wide enough that it could probably support a quad-track subway. And/or, Huntington will continue to have transit ROW at surface level (if only to have bus lanes for the 39). Which means the Green Line doesn't need to be our only Huntington service. We could basically just keep a version of today's E Line mostly in place. (You would need to do a little fussing around Back Bay/Copley Junction and at the Mass Ave/Huntington intersection, but I think it would be doable.) Or we could quad-track the new Huntington Subway and run Gold Line Seaport trains on one set of tracks and Green Line Downtown trains on the other. This would also let you address...

 

[Riverside]

Loss of South Station - Huntington OSR

• Mainly affect Old Colony riders, and those coming from Logan Airport (!) and Seaport

The South Station <> Longwood 1SR: So, setting aside the quad-track/upstairs-downstairs idea I just mentioned, I do want to talk about the South Station <> LMA 1SR question in terms of my original map above. This was also one of my biggest hesitations. But a few things changed my mind:

• The eastern half of Huntington (from MFA to Back Bay) would be mostly accessible via a high-freq Regional Rail South Station <> Back Bay <> Ruggles service
• The western end of the LMA (including the behemoth that is Brigham) can be served reasonably well with the existing Longwood ("Chapel") station plus an infill at Francis

That doesn't solve all of the problems, but it does knock out a chunk of them. (And yeah, keeping a surface E branch of the Gold Line probably solves the rest.) But there is one other option that gets opened up under Criss Cross:

Using the streetcar-subway superpower -- brief street-running: this would not be my first choice. But if a system is designed to be flexible enough to accommodate disruption from the streetscape, it means you can (on a limited basis) do brief street-running if needed. So, if push came to shove, a Gold Line branch from Kenmore could turn east at "Chapel" to head directly into LMA proper. This would obviously not be my first choice, but it's a capability that Criss Cross has that Tripod doesn't.

Logan <> Huntingon journeys: This could be addressed a couple of ways:

• Use my suggestion of a Congress St-busway-using SL1 being extended to Boylston -- relocating the transfer from South Station to Boylston, but keeping a 2SR overall
• Enhance the Logan Express service to Back Bay + Copley, relocating the transfer from South Station to Copley, but also keeping a 2SR overall

Greater unreliability and uncertainty for Seaport

• May not be a big issue: the uncertainty exists in the Central Subway regardless, and Criss-Cross makes Seaport trade reliability for frequency, which is well worth it

Yeah, and I think the point I made last year stands: the branches to Harvard and Fenway Branch provide some reliability buttressing, since they'll be running primarily in dedicated ROWs.

The Red-Central Subway transfer point is moved south from Park St to South Station

I think this actually comes out to a wash, or maybe even is better than Tripod?

• Tripod
  • Park St: B, C, half-of-Huntington, Nubian
  • South Station: half-of-Huntington
• Criss Cross
  • Park St: Huntington, Nubian
  • South Station: B, C (plus new branches)

If anything, I think Criss Cross distributes the load better?

Presenting to you, a very quick and rough drawing of the 2x2 / "Full Criss-Cross" model:

While I do think this system just runs into too much of the "DeKalb Problem", it is definitely an elegant solution. (I think part of the problem is that each of your four endpoints itself have, like, 3 branches coming out of them, which means this network would be really complicated.) And yes, one thing that is nice about Criss-Cross as I've presented it is that it would be relatively straightforward to complete the "Full Criss Cross".

But my most preferred approach is actually pretty similar to what Riverside said here -- turn onto Lincoln St:

The turn from "Marginal Rd extension" to "Lincoln St extension", under the block of spaghetti highway ramps, should be easily doable. At the very least, there's currently an access road that cuts under the ramps, so a tunnel can be buried directly beneath it. Lincoln St itself is wide enough for a subway, with 60' building-to-building.

My only concern is with the turn from Lincoln to Essex, which is quite tight and may need to interfere with the Big Dig.

I should've given you credit here, since yes the Lincoln St alignment arose out of discussions I was having with you. But yes, the piece that worries me here is the curve from Lincoln to Essex.

 

[TheRatmeister]

I think it would still face a lot of the challenges faced by the Silver Line Phase III alignment, wouldn't it?

maybe? It wouldn't need to dive under Boylston station though so maybe not, I'm not sure. I'm also not sure if the plan was designed with the Washington St Subway in mind or not.

More to come later whenever I have time to read a novel lol.