Crazy Transit Pitches

This image shows the version proposed by the MIS study, but you're right, I don't think it's the best option, the GJ would better serve as a trunk line for services that branch out, and I've already given a few reasons why I don't think light rail is actually the best plan for that. I've spread it out over a few comments, but I'll condense it all here: My general opinion is that the best transit mode for a converted GJ is a busway for these reasons:
  • Buses can handle steeper gradients than light rail, making grade separation via over/underpasses easier
  • Additional buses can be operated by private entities such as labs, universities, or other companies, which is really useful in a place like Kendall.
  • Buses give you the most flexibility to merge routes on and off the ROW
  • And of course, buses can utilize the existing road network

View attachment 43997
TL;DR:
  • A bus corridor for Kendall has better options than tearing up Grand Junction and paving a busway
  • A micro-level "Kendall network" really shouldn't be separated from a macro-level "circumferential route", especially for mid-range destinations
  • For such a service, a Grand Junction heavy-metro LRT line achieves the best balance in the real world
I'd argue that, if your overall goal is specifically to build a busway to serve Kendall, Grand Junction is a bad alignment, because:
  • Need to tear up tracks and pave roads - which cost $$
  • Limited catchment along the route (at the western edge of the office area, and not too many residents nearby)
  • Hard to do a busway through Brickbottom and GLMF/CRMF
    • IMO, this is why any service using this part of Grand Junction (Brickbottom) really should be rail, especially GL-compatible rail
  • Bad transfers to the Red Line
Your 4 reasons in favor of buses seem very generic, and may not apply as well in the context of Kendall. Some may even conflict with each other (e.g. grade separation vs. merges and private entities).

Instead, how about this route?
Kendall Transitway.png


The main feature is a dedicated busway along Galileo Galilei Way, Binney St and Land Blvd, either in dedicated lanes or as an elevated (preferred but $$). The blue section near Kendall can also be used as an alternative route, greatly enhancing Red Line transfer at the cost of grade separation.

Compared to a Grand Junction-only busway, this tackles almost all its difficulties one by one:
  • Uses existing road infrastructure - can start running immediately if you don't need grade separation
  • Covers a much greater part of Kendall (Binney St labs) and East Cambridge
  • Avoids the mess of Brickbottom Junction, while still providing feasible ways to Sullivan
  • Much better Red Line transfer possible
  • And, as a bonus, also connects to Lechmere (GLX and developments there)
But take a step back. Isn't this basically recreating SL6?
1699001022130.png


The effectiveness of such a feeder network with buses is also questionable. The 64 bus, which connects Kendall to Cambridgeport, Allston and Oak Square, doesn't have great ridership. The 85 to Union Square is even worse. The CT2 does better than both, but it has a lot more catchment and is decidedly NOT a local feeder route for Kendall. And even though the Bus Network Redesign initially had the 83 bus to Kendall, it was reverted back to Central in the final draft due to community feedback.

And who's to say that such a bus network alone, as the only enhancement for Kendall, is really the best way to satisfy someone going to Kendall from… Sullivan? Kenmore? Chelsea? South End?

Which leads to another question…

Is there really a need to completely separate a micro-level "Kendall feeder network" from a macro-level "circumferential route"?

Let's first try to sketch out some characteristics of both.

Kendall network:
  • Goal: Bring people to Kendall
  • Should ideally have wide coverage to a variety of places
    • On paper, this suggests many branches
  • Slower service may be OK??? That's a big question mark - more on this soon
  • Red Line transfer is not as important
Circumferential line:
  • Goal: Bring people away from the transfer stations downtown
  • Should ideally hit most transfer stations and transit hubs
    • On paper, this suggests one continuous line that crosses all other lines is sufficient. But…
  • Should also hope to hit as many destinations (employment centers) as possible
    • To maximize 2-seat-rides and minimize 3-seat-rides
  • Speed is crucial
  • Transfers to other lines are crucial
… Is it, really? Let's take a look at the "variety of places":
Kendall coverage.png


The ring boundaries are not supposed to be exact, and nodes at the corners can swing either way. Places along the Red Line are excluded.

But the high-level picture is, these "variety of places" can be roughly divided into tiers, each with different needs:
  • Red:Immediate neighborhoods
    • Parts of Cambridge, Charlestown, Lechmere/GLX
  • Blue:Transfer nodes (mostly)
    • Sullivan/North Station, Kenmore/GL branches, Ruggles/Back Bay, Nubian
    • Some of these are near the red/blue boundary, but still
  • Green:Faraway neighborhoods
    • Everett, Chelsea, Medford, Allston/Brighton/Brookline, Roxbury/Dorchester
  • Don't forget there are places even further away that are not shown:
    • Revere/Lynn, Malden, Watertown, Newton, Forest Hills, Mattapan, Quincy, etc
I'd argue that: By the time you get to the blue ring (or red/blue border) and especially its transfer hubs, the needs for transit to Kendall change fundamentally. At that point, speed becomes important.

Many transfer hubs happen to fall here. And by virtue of them being transfer nodes, these passengers often come from other transit modes with origins that are much further away. They're also more likely to have the option to continue riding into downtown and transfer there.

For someone arriving at Sullivan, Ruggles, etc, what they need is not a non-grade-separated service with frequent stops (like T101, even though that's better than the status quo). What they'll most likely need is:
an orbital line with more central, limited stops
Yes, that's what I'm saying: At that distance, a micro-level Kendall subnetwork coincides with a macro-level circumferential route.

So we really should be taking a layered approach even when looking at Kendall's needs. What we need are probably a fast connection with long stop spacing that happens to also serve as an "orbital line meant to better connect the city", plus short-to-medium-haul feeder buses for nearby neighborhoods (red ring, maybe some of blue ring, mayyybe Kenmore) and the handful connections that aren't served by the former.

Fortunately for Kendall, these two objectives (mid-haul micro network + macro ring route) align pretty well - a single spine serves most of the blue ring. That's a feature, not a bug. No need to go to two extremes of "buses vs. deep-bored HRT" just to achieve perfect separation.

Is "heavy-metro circumferential LRT down Grand Junction" the best way to do this? No, but it's the most realistic compromise between cost, flexibility and (sufficient) capacity.
  • First of all, does a macro ring route need to connect at Kendall in the first place? Not necessarily: Kendall, Central and Harvard each has its own merits. But I believe Kendall has strong arguments going for it - such as how it also double duties as a much-needed Kendall subnetwork.
  • I'll say this for the 128th time: A 95% grade-separated, 2+ Type 10 sets "heavy metro" LRT through Grand Junction is the most realistic option through Kendall. And it gives you much more flexibility that HRT doesn't have, at the cost of losing 1/3 of HRT's capacity:
    • Ability to get something running first, and then incrementally do grade separation later while not subjecting to low capacity forever
    • Possibility of branches - there are already strong arguments for branches on both ends, Sullivan/North Station and Kenmore/West Station. Yes, buses can do branches too, but if LRT can do the same thing, why not?
The Achilles' Heel of Grand Junction is: Its transfer with Kendall sucks. This is definitely a huge negative for Kendall on a circumferential route, but it also gives the only solution that doesn’t involve either a (most likely TBM) tunnel or an excessive amount of street running.

In an ideal sandbox world, the following alignment is probably ideal for a purely macro-level circumferential route. But good luck seeing this fully-TBM route happen in 2223.
TBM Ring.png
 
TL;DR:
  • A bus corridor for Kendall has better options than tearing up Grand Junction and paving a busway
  • A micro-level "Kendall network" really shouldn't be separated from a macro-level "circumferential route", especially for mid-range destinations
  • For such a service, a Grand Junction heavy-metro LRT line achieves the best balance in the real world
I'd argue that, if your overall goal is specifically to build a busway to serve Kendall, Grand Junction is a bad alignment, because:
  • Need to tear up tracks and pave roads - which cost $$
  • Limited catchment along the route (at the western edge of the office area, and not too many residents nearby)
  • Hard to do a busway through Brickbottom and GLMF/CRMF
    • IMO, this is why any service using this part of Grand Junction (Brickbottom) really should be rail, especially GL-compatible rail
  • Bad transfers to the Red Line
Your 4 reasons in favor of buses seem very generic, and may not apply as well in the context of Kendall. Some may even conflict with each other (e.g. grade separation vs. merges and private entities).

Instead, how about this route?
View attachment 44103

The main feature is a dedicated busway along Galileo Galilei Way, Binney St and Land Blvd, either in dedicated lanes or as an elevated (preferred but $$). The blue section near Kendall can also be used as an alternative route, greatly enhancing Red Line transfer at the cost of grade separation.

Compared to a Grand Junction-only busway, this tackles almost all its difficulties one by one:
  • Uses existing road infrastructure - can start running immediately if you don't need grade separation
  • Covers a much greater part of Kendall (Binney St labs) and East Cambridge
  • Avoids the mess of Brickbottom Junction, while still providing feasible ways to Sullivan
  • Much better Red Line transfer possible
  • And, as a bonus, also connects to Lechmere (GLX and developments there)
But take a step back. Isn't this basically recreating SL6?
View attachment 44104

The effectiveness of such a feeder network with buses is also questionable. The 64 bus, which connects Kendall to Cambridgeport, Allston and Oak Square, doesn't have great ridership. The 85 to Union Square is even worse. The CT2 does better than both, but it has a lot more catchment and is decidedly NOT a local feeder route for Kendall. And even though the Bus Network Redesign initially had the 83 bus to Kendall, it was reverted back to Central in the final draft due to community feedback.

And who's to say that such a bus network alone, as the only enhancement for Kendall, is really the best way to satisfy someone going to Kendall from… Sullivan? Kenmore? Chelsea? South End?

Which leads to another question…

Is there really a need to completely separate a micro-level "Kendall feeder network" from a macro-level "circumferential route"?

Let's first try to sketch out some characteristics of both.

Kendall network:
  • Goal: Bring people to Kendall
  • Should ideally have wide coverage to a variety of places
    • On paper, this suggests many branches
  • Slower service may be OK??? That's a big question mark - more on this soon
  • Red Line transfer is not as important
Circumferential line:
  • Goal: Bring people away from the transfer stations downtown
  • Should ideally hit most transfer stations and transit hubs
    • On paper, this suggests one continuous line that crosses all other lines is sufficient. But…
  • Should also hope to hit as many destinations (employment centers) as possible
    • To maximize 2-seat-rides and minimize 3-seat-rides
  • Speed is crucial
  • Transfers to other lines are crucial
… Is it, really? Let's take a look at the "variety of places":
View attachment 44105

The ring boundaries are not supposed to be exact, and nodes at the corners can swing either way. Places along the Red Line are excluded.

But the high-level picture is, these "variety of places" can be roughly divided into tiers, each with different needs:
  • Red:Immediate neighborhoods
    • Parts of Cambridge, Charlestown, Lechmere/GLX
  • Blue:Transfer nodes (mostly)
    • Sullivan/North Station, Kenmore/GL branches, Ruggles/Back Bay, Nubian
    • Some of these are near the red/blue boundary, but still
  • Green:Faraway neighborhoods
    • Everett, Chelsea, Medford, Allston/Brighton/Brookline, Roxbury/Dorchester
  • Don't forget there are places even further away that are not shown:
    • Revere/Lynn, Malden, Watertown, Newton, Forest Hills, Mattapan, Quincy, etc
I'd argue that: By the time you get to the blue ring (or red/blue border) and especially its transfer hubs, the needs for transit to Kendall change fundamentally. At that point, speed becomes important.

Many transfer hubs happen to fall here. And by virtue of them being transfer nodes, these passengers often come from other transit modes with origins that are much further away. They're also more likely to have the option to continue riding into downtown and transfer there.

For someone arriving at Sullivan, Ruggles, etc, what they need is not a non-grade-separated service with frequent stops (like T101, even though that's better than the status quo). What they'll most likely need is:

Yes, that's what I'm saying: At that distance, a micro-level Kendall subnetwork coincides with a macro-level circumferential route.

So we really should be taking a layered approach even when looking at Kendall's needs. What we need are probably a fast connection with long stop spacing that happens to also serve as an "orbital line meant to better connect the city", plus short-to-medium-haul feeder buses for nearby neighborhoods (red ring, maybe some of blue ring, mayyybe Kenmore) and the handful connections that aren't served by the former.

Fortunately for Kendall, these two objectives (mid-haul micro network + macro ring route) align pretty well - a single spine serves most of the blue ring. That's a feature, not a bug. No need to go to two extremes of "buses vs. deep-bored HRT" just to achieve perfect separation.

Is "heavy-metro circumferential LRT down Grand Junction" the best way to do this? No, but it's the most realistic compromise between cost, flexibility and (sufficient) capacity.
  • First of all, does a macro ring route need to connect at Kendall in the first place? Not necessarily: Kendall, Central and Harvard each has its own merits. But I believe Kendall has strong arguments going for it - such as how it also double duties as a much-needed Kendall subnetwork.
  • I'll say this for the 128th time: A 95% grade-separated, 2+ Type 10 sets "heavy metro" LRT through Grand Junction is the most realistic option through Kendall. And it gives you much more flexibility that HRT doesn't have, at the cost of losing 1/3 of HRT's capacity:
    • Ability to get something running first, and then incrementally do grade separation later while not subjecting to low capacity forever
    • Possibility of branches - there are already strong arguments for branches on both ends, Sullivan/North Station and Kenmore/West Station. Yes, buses can do branches too, but if LRT can do the same thing, why not?
The Achilles' Heel of Grand Junction is: Its transfer with Kendall sucks. This is definitely a huge negative for Kendall on a circumferential route, but it also gives the only solution that doesn’t involve either a (most likely TBM) tunnel or an excessive amount of street running.

In an ideal sandbox world, the following alignment is probably ideal for a purely macro-level circumferential route. But good luck seeing this fully-TBM route happen in 2223.
View attachment 44106
I advocated for this routing during the SL process. You needn't have a stop at Sullivan as you can do OL at Com College
 
TL;DR:
  • A bus corridor for Kendall has better options than tearing up Grand Junction and paving a busway
  • A micro-level "Kendall network" really shouldn't be separated from a macro-level "circumferential route", especially for mid-range destinations
  • For such a service, a Grand Junction heavy-metro LRT line achieves the best balance in the real world
I'd argue that, if your overall goal is specifically to build a busway to serve Kendall, Grand Junction is a bad alignment, because:
  • Need to tear up tracks and pave roads - which cost $$
  • Limited catchment along the route (at the western edge of the office area, and not too many residents nearby)
  • Hard to do a busway through Brickbottom and GLMF/CRMF
    • IMO, this is why any service using this part of Grand Junction (Brickbottom) really should be rail, especially GL-compatible rail
  • Bad transfers to the Red Line
Your 4 reasons in favor of buses seem very generic, and may not apply as well in the context of Kendall. Some may even conflict with each other (e.g. grade separation vs. merges and private entities).

Instead, how about this route?
View attachment 44103

The main feature is a dedicated busway along Galileo Galilei Way, Binney St and Land Blvd, either in dedicated lanes or as an elevated (preferred but $$). The blue section near Kendall can also be used as an alternative route, greatly enhancing Red Line transfer at the cost of grade separation.

Compared to a Grand Junction-only busway, this tackles almost all its difficulties one by one:
  • Uses existing road infrastructure - can start running immediately if you don't need grade separation
  • Covers a much greater part of Kendall (Binney St labs) and East Cambridge
  • Avoids the mess of Brickbottom Junction, while still providing feasible ways to Sullivan
  • Much better Red Line transfer possible
  • And, as a bonus, also connects to Lechmere (GLX and developments there)
But take a step back. Isn't this basically recreating SL6?
View attachment 44104

The effectiveness of such a feeder network with buses is also questionable. The 64 bus, which connects Kendall to Cambridgeport, Allston and Oak Square, doesn't have great ridership. The 85 to Union Square is even worse. The CT2 does better than both, but it has a lot more catchment and is decidedly NOT a local feeder route for Kendall. And even though the Bus Network Redesign initially had the 83 bus to Kendall, it was reverted back to Central in the final draft due to community feedback.

And who's to say that such a bus network alone, as the only enhancement for Kendall, is really the best way to satisfy someone going to Kendall from… Sullivan? Kenmore? Chelsea? South End?

Which leads to another question…

Is there really a need to completely separate a micro-level "Kendall feeder network" from a macro-level "circumferential route"?

Let's first try to sketch out some characteristics of both.

Kendall network:
  • Goal: Bring people to Kendall
  • Should ideally have wide coverage to a variety of places
    • On paper, this suggests many branches
  • Slower service may be OK??? That's a big question mark - more on this soon
  • Red Line transfer is not as important
Circumferential line:
  • Goal: Bring people away from the transfer stations downtown
  • Should ideally hit most transfer stations and transit hubs
    • On paper, this suggests one continuous line that crosses all other lines is sufficient. But…
  • Should also hope to hit as many destinations (employment centers) as possible
    • To maximize 2-seat-rides and minimize 3-seat-rides
  • Speed is crucial
  • Transfers to other lines are crucial
… Is it, really? Let's take a look at the "variety of places":
View attachment 44105

The ring boundaries are not supposed to be exact, and nodes at the corners can swing either way. Places along the Red Line are excluded.

But the high-level picture is, these "variety of places" can be roughly divided into tiers, each with different needs:
  • Red:Immediate neighborhoods
    • Parts of Cambridge, Charlestown, Lechmere/GLX
  • Blue:Transfer nodes (mostly)
    • Sullivan/North Station, Kenmore/GL branches, Ruggles/Back Bay, Nubian
    • Some of these are near the red/blue boundary, but still
  • Green:Faraway neighborhoods
    • Everett, Chelsea, Medford, Allston/Brighton/Brookline, Roxbury/Dorchester
  • Don't forget there are places even further away that are not shown:
    • Revere/Lynn, Malden, Watertown, Newton, Forest Hills, Mattapan, Quincy, etc
I'd argue that: By the time you get to the blue ring (or red/blue border) and especially its transfer hubs, the needs for transit to Kendall change fundamentally. At that point, speed becomes important.

Many transfer hubs happen to fall here. And by virtue of them being transfer nodes, these passengers often come from other transit modes with origins that are much further away. They're also more likely to have the option to continue riding into downtown and transfer there.

For someone arriving at Sullivan, Ruggles, etc, what they need is not a non-grade-separated service with frequent stops (like T101, even though that's better than the status quo). What they'll most likely need is:

Yes, that's what I'm saying: At that distance, a micro-level Kendall subnetwork coincides with a macro-level circumferential route.

So we really should be taking a layered approach even when looking at Kendall's needs. What we need are probably a fast connection with long stop spacing that happens to also serve as an "orbital line meant to better connect the city", plus short-to-medium-haul feeder buses for nearby neighborhoods (red ring, maybe some of blue ring, mayyybe Kenmore) and the handful connections that aren't served by the former.

Fortunately for Kendall, these two objectives (mid-haul micro network + macro ring route) align pretty well - a single spine serves most of the blue ring. That's a feature, not a bug. No need to go to two extremes of "buses vs. deep-bored HRT" just to achieve perfect separation.

Is "heavy-metro circumferential LRT down Grand Junction" the best way to do this? No, but it's the most realistic compromise between cost, flexibility and (sufficient) capacity.
  • First of all, does a macro ring route need to connect at Kendall in the first place? Not necessarily: Kendall, Central and Harvard each has its own merits. But I believe Kendall has strong arguments going for it - such as how it also double duties as a much-needed Kendall subnetwork.
  • I'll say this for the 128th time: A 95% grade-separated, 2+ Type 10 sets "heavy metro" LRT through Grand Junction is the most realistic option through Kendall. And it gives you much more flexibility that HRT doesn't have, at the cost of losing 1/3 of HRT's capacity:
    • Ability to get something running first, and then incrementally do grade separation later while not subjecting to low capacity forever
    • Possibility of branches - there are already strong arguments for branches on both ends, Sullivan/North Station and Kenmore/West Station. Yes, buses can do branches too, but if LRT can do the same thing, why not?
The Achilles' Heel of Grand Junction is: Its transfer with Kendall sucks. This is definitely a huge negative for Kendall on a circumferential route, but it also gives the only solution that doesn’t involve either a (most likely TBM) tunnel or an excessive amount of street running.

In an ideal sandbox world, the following alignment is probably ideal for a purely macro-level circumferential route. But good luck seeing this fully-TBM route happen in 2223.
View attachment 44106
In general, I agree, and you've definitely convinced me that, at least for the GJ section, light rail is at worst "fine". My only remaining issue is still the bigger urban ring project. Assuming the intention is to roughly follow the 2002 MIS, it's not like we're choosing between a bored tunnel or light rail. We would be choosing between a longer bored tunnel or a shorter bored tunnel and some surface running, since we all agree that Longwood badly needs more transit capacity and since there really isn't a great way to do surface running or cut and cover, the streets are too narrow, a deep-bore tunnel would essentially have to happen. So, given that, I think it makes more sense to do full heavy rail in a fully bored tunnel, rather than compromise with a half surface light rail route. We'd already be paying the fixed costs for a bored tunnel, so let's just go all the way to Brickbottom.
 
In general, I agree, and you've definitely convinced me that, at least for the GJ section, light rail is at worst "fine". My only remaining issue is still the bigger urban ring project. Assuming the intention is to roughly follow the 2002 MIS, it's not like we're choosing between a bored tunnel or light rail. We would be choosing between a longer bored tunnel or a shorter bored tunnel and some surface running, since we all agree that Longwood badly needs more transit capacity and since there really isn't a great way to do surface running or cut and cover, the streets are too narrow, a deep-bore tunnel would essentially have to happen. So, given that, I think it makes more sense to do full heavy rail in a fully bored tunnel, rather than compromise with a half surface light rail route. We'd already be paying the fixed costs for a bored tunnel, so let's just go all the way to Brickbottom.
But the deep station work for your long bored tunnel is going to be a huge extra cost. Light Rail gives you the option of surface, including stations, where you can, deep bore where you must.
 
But the deep station work for your long bored tunnel is going to be a huge extra cost. Light Rail gives you the option of surface, including stations, where you can, deep bore where you must.
Yes, and then you learn the lesson Seattle is learning right now about what that does to your future capacity, with trains that are the equivalent of 4 type 10 cars long, and there's still not enough room. Longer trains mean more expensive stations everywhere, even on the surface, and not spending the money on grade separation upfront will make it much harder to retrofit, meaning you'll miss out on potential for automation and super high frequency service, again very important on an orbital route where people will be making a lot of transfers.
 
since we all agree that Longwood badly needs more transit capacity and since there really isn't a great way to do surface running or cut and cover, the streets are too narrow, a deep-bore tunnel would essentially have to happen.
Is Francis St too narrow for Cut and Cover? Of course there will be difficulties with partially shutting down a street in front of a hospital, at least Fenwood rd could provide some relief being closely parallel with 3 connecting streets.
 
Yes, and then you learn the lesson Seattle is learning right now about what that does to your future capacity, with trains that are the equivalent of 4 type 10 cars long, and there's still not enough room. Longer trains mean more expensive stations everywhere, even on the surface, and not spending the money on grade separation upfront will make it much harder to retrofit, meaning you'll miss out on potential for automation and super high frequency service, again very important on an orbital route where people will be making a lot of transfers.
Every rapid transit line in metro Boston has seen extensive extension of the grade separated section over their lifetime.

It happens (and gets funded) when demand requires it.
 
Is Francis St too narrow for Cut and Cover? Of course there will be difficulties with partially shutting down a street in front of a hospital, at least Fenwood rd could provide some relief being closely parallel with 3 connecting streets.
Even assuming it's not too narrow, you immediately run into the next problem of: What happens to the north and south of the C&C tunnel? C&C north of Longwood is not an option, and south of Longwood you'd end up with very few routing options once you pass Nubian, not to mention you'd need to either pass under the SW corridor or find room on Tremont St... somewhere.

Every rapid transit line in metro Boston has seen extensive extension of the grade separated section over their lifetime.

It happens (and gets funded) when demand requires it.
So if we're doing a two-stage project, why would we do a light rail stage, then a light rail stage that costs as much as heavy rail, when we could just do two heavy rail stages? I'd really argue that the Cambridge section of an orbital route is not the most important, it doesn't really make sense to start there when you'd almost certainly see way more ridership in the rapid transit starved area around Nubian and the South End.
 
I'll start by saying this: The current climate for transit planning in the US doesn't seem conducive for deep-bored new heavy rail lines. Even projects with bored tunnels often opt for light rail to operate in them, whereas our Red-Blue is moving forward with C&C (which is a good choice here but still unable to keep the cost low). Whether it's solely because of cost, or because most cities who are actively working on major expansions don't have the demand for heavy rail, I don't know. But as much as I'd love to see more heavy rail lines and more TBM tunnels, it makes me feel pessimistic about their future.

We would be choosing between a longer bored tunnel or a shorter bored tunnel and some surface running, since we all agree that Longwood badly needs more transit capacity and since there really isn't a great way to do surface running or cut and cover, the streets are too narrow, a deep-bore tunnel would essentially have to happen.
That's a very good point - there will be fixed costs the moment a TBM is involved. That being said, from what I know (which is admittedly not too much), the variable cost is still significant. A short 1-mile bored tunnel to take care of tricky parts (like Longwood) has very different costs from a full 13.5-mile Urban Ring as shown below - and make no mistake, I really want that to be feasible.
1699033286315.png

Is Francis St too narrow for Cut and Cover? Of course there will be difficulties with partially shutting down a street in front of a hospital, at least Fenwood rd could provide some relief being closely parallel with 3 connecting streets.
F-Line (in 2015) was very insistent on a cross-Longwood tunnel via either Longwood Ave or Francis St simply being "impossible":
Longwood is even tougher. Longwood Ave.'s a 2-lane street (so is Francis), very tall NU and hospital buildings right up to curbside its entire length. Then crossing the Muddy River at its widest point. Then bang-bang interfacing with the D right next to the river and walking path. Going along Fenway via Louis Prang St. has similar issues of waterproofing around the Fens and a river crossing right where it interfaces with the D.

All that stupidity was part of the south half of the Phase III study's cross-Brookline tunnel recycling the old I-695 tunnel alignment. It's deader than dead for exactly the same reasons the 695 schematics weren't worth the paper they were printed on.


You don't have an easy interface, but BV has the street width for cut-and-cover tunneling and trajectory off Huntington. And the fewest by far wetlands to square. If you have to raze any buildings the Pearl block w/ that New England Institute of Art eyesore and the hospital office buildings are the most disposable/rebuildable of any. The apartment buildings flanking the air rights garage shouldn't be an issue, though. They're separated by 125 ft. on either side of the ROW with just that flimsy deck over the tracks between them, and set back ~30-40 ft. from curbside.

It's not without problems, but I'd much rather take my best shot here because cross-Longwood really really really isn't going to work. We found that out 45 years ago when the increasingly desperate tunnel concepts the state proposed as a last-ditch effort to save I-695 were on their final death spiral. The very inclusion of that routing in the UR Phase III grade separation concepts betrays considerable amount of un-seriousness on the state's part because they know this better than anyone.
Not sure what's the implied construction method here (C&C, TBM or anything). I do hope a TBM tunnel there can happen, but still.

Coming back to @TheRatmeister 's point:
since we all agree that Longwood badly needs more transit capacity and since there really isn't a great way to do surface running or cut and cover, the streets are too narrow, a deep-bore tunnel would essentially have to happen.
There are "cheap build" crayon proposals that serve Longwood (less well) without expensive tunneling. The consensus seem to be something like this:
1699034048213.png

The general idea is to rely on existing Longwood (D) and Longwood Medical Area (E) stations - which are still a 10-min walk away from the heart of LMA, but not terribly further than Financial District to nearby stations. The orange route assumes a short Commonwealth Ave subway extension to BU bridge. The brown route assumes a Huntington Ave subway and the D-E connector, which is generally seen as C&C-doable possibly except for the Muddy River crossing. Other parts like Ruggles St and Melnea Cass are also candidates for C&C tunnels if needed. The two routes can either be through-run or both terminate near Brookline Village (since they serve different travel patterns that mostly end at LMA).

Yes, this is very circuitous, far from ideal, and doesn't serve Longwood as well as we hope to. But this means a TBM-free network is at least feasible, and many C&C tunnels are also mostly free of utilities (most notably under the B & E reservations).

Yes, and then you learn the lesson Seattle is learning right now about what that does to your future capacity, with trains that are the equivalent of 4 type 10 cars long, and there's still not enough room. Longer trains mean more expensive stations everywhere, even on the surface, and not spending the money on grade separation upfront will make it much harder to retrofit, meaning you'll miss out on potential for automation and super high frequency service, again very important on an orbital route where people will be making a lot of transfers.
Their 4-car set can still hold 776 passengers, not too far off from a Red Line set (this is the best I can find). They're also running at 8-min frequencies during peak, so it sounds more like a frequency issue before a capacity one.

Also, Seattle is still building the main spines of their network. That would have been equivalent to our Red and Orange lines running light rail. The characters of our Urban Ring will be very different, and by the very nature of it being used for short-haul transfers, its ridership will be spread out among different segments and may not translate entirely into peak crowdedness.

Genuine question: Do we have any evidence that an Urban Ring (even the fully tunneled approach in the MIS) will have so much ridership that warrants high-frequency heavy rail? I honestly find that hard to believe, at least in the foreseeable future. Jobs in downtown Boston still far outweigh all other nodes (Kendall, Longwood etc). On the other hand, Singapore's Circle Line (which I'm familiar with, and which I think draws many parallels to Boston) uses 3-car heavy rail as compared to 6-car sets on their main radial routes, and while they're crowded, they're still not as crowded as the radial lines.
 
when you'd almost certainly see way more ridership in the rapid transit starved area around Nubian and the South End.
That just means Nubian needs better radial transit options, and I'd argue that's a much bigger priority for them than a circumferential one. The need for a Nubian-downtown connection isn't eliminated by a ring route, which generally doesn't do as well at serving transit deserts as radial ones do.

(This also applies to Everett and Chelsea, but they have a different set of complications.)
 
Last edited:
Genuine question: Do we have any evidence that an Urban Ring (even the fully tunneled approach in the MIS) will have so much ridership that warrants high-frequency heavy rail? I honestly find that hard to believe, at least in the foreseeable future. Jobs in downtown Boston still far outweigh all other nodes (Kendall, Longwood etc). On the other hand, Singapore's Circle Line (which I'm familiar with, and which I think draws many parallels to Boston) uses 3-car heavy rail as compared to 6-car sets on their main radial routes, and while they're crowded, they're still not as crowded as the radial lines.
The MIS study predicts around 200,000 daily weekday ridership on a heavy rail route from Wellington-Nubian via Kendall/Kenmore/Longwood, which would give it roughly as many riders as the rest of the OL. They predict BRT service would add another 75,000 or so, which assuming the whole route was built as heavy rail, would likely make it the most used line in the system. Not unsurprising given that the full line would hit the airport, Kendall, MIT, Everett, Chelsea, BU/Fenway, Longwood, Ruggles, Nubian, BU Medical, and UMass Boston. Lots of heavy hitters there. I think Boston is somewhat unique in that regard by having so many major destinations spread out in a neat orbital ring that's a decent distance out from downtown, and why it's really why I think heavy rail is super important, these aren't minor destinations and this isn't just an orbital route, it's a line that serves most major points in the city outside of downtown, as well as some of the largest suburbs like Everett, Revere, and Chelsea.

I will say take these numbers with a grain of salt, they also predicted around 225,000 daily CR boardings by 2025, even in 2019 it wasn't topping 100,000 daily.

I will also take slight issue with your distances, I believe the original plan calls for a tunnel around 2 mi long, and just with some rough estimates a full ring with the more conservative distance would likely only be around 9ish miles of deep bore tunnel, 5 for the section from ~Andrew to Sulivan, and another 4 ish from Chelsea back to City Point. Sullivan-Chelsea could likely be done all above ground, and C&C should be possible along E Broadway and Dorchester St in Southie. Going via Seaport would add another mile ish, but I really think a neighborhood of ~35k with no rapid transit should be prioritized for a new line over a wealthy business district of less than 5000 residents with transit access that's already pretty good, not to mention some of the major redvelopments happening in some of the industrial areas and potential for a pretty major increase in density over time.
 
Last edited:
The MIS study predicts around 200,000 daily weekday ridership on a heavy rail route from Wellington-Nubian via Kendall/Kenmore/Longwood, which would give it roughly as many riders as the rest of the OL. They predict BRT service would add another 75,000 or so, which assuming the whole route was built as heavy rail, would likely make it the most used line in the system. Not unsurprising given that the full line would hit the airport, Kendall, MIT, Everett, Chelsea, BU/Fenway, Longwood, Ruggles, Nubian, BU Medical, and UMass Boston. Lots of heavy hitters there. I think Boston is somewhat unique in that regard by having so many major destinations spread out in a neat orbital ring that's a decent distance out from downtown, and why it's really why I think heavy rail is super important, these aren't minor destinations and this isn't just an orbital route, it's a line that serves most major points in the city outside of downtown, as well as some of the largest suburbs like Everett, Revere, and Chelsea.

I will say take these numbers with a grain of salt, they also predicted around 225,000 daily CR boardings by 2025, even in 2019 it wasn't topping 100,000 daily.
Not sure if I trust the practicality of these numbers - I agree they're probably too high. 200k is about the daily ridership on the Red and Orange lines pre-Covid (Red slightly above, Orange slightly below).

But even if we take it at face value that Urban Ring has as many riders as Orange Line, there's another caveat: the load, or the number of passengers on each train, is different from ridership.
OL load Fall19.png

This chart shows how many passengers are on the train at each station (summed over all trains during AM peak, I believe), except for the last column which shows the total number of boardings. Ridership is the sum of blue and orange bars in the last column. The maximum load, which determines the capacity needed per train set, is 54-62% of total boardings.

Clearly, OL's ridership has a very standard "single peak" distribution. People board at either ends of the line, and generally get off in the downtown core between North Station and Back Bay (with some more alightings at Ruggles from the north).

Urban Ring, on the other hand, will have much, much higher turnover. A counter-clockwise train from Airport during AM peak will probably pick up passengers in Everett and Chelsea, dump at least half of them at Sullivan in exchange for new passengers, dump a lot of them at Kendall (or Central/Harvard) with perhaps a bit more boardings, pick up more when it crosses the GL branches, drop off a huge chunk at Longwood, and pick up again at Ruggles and Nubian heading to BUMC and Seaport. The same 200k passengers will be distributed a lot more evenly along the route with a much lower peak load in the car, unlike OL.

So Urban Ring will probably not need cars as big as OL, even if they have the same ridership. The billion dollar question is, by how much? My guess is that the load will be low enough that 2-car or 3-car Type 10s can handle them, if other circumstances call for LRT.

You did make me feel more optimistic about the prospects of a heavy rail Urban Ring materializing eventually, so kudos. However, I still don't think that's enough to move a heavily deep-bored UR out of the God-Mode thread yet.

I will also take slight issue with your distances, I believe the original plan calls for a tunnel around 2 mi long, and just with some rough estimates a full ring with the more conservative distance would likely only be around 9ish miles of deep bore tunnel, 5 for the section from ~Andrew to Sulivan, and another 4 ish from Chelsea back to City Point. Sullivan-Chelsea could likely be done all above ground, and C&C should be possible along E Broadway and Dorchester St in Southie.
Yeah, the 13.5 mi ring is mostly for illustrative purposes. The point was to show that it's still very different than C&Cs on specific short sections (especially if you're doing deep bore from Andrew all the way to Sullivan), and 9 vs 13.5 doesn't really change that.

Going via Seaport would add another mile ish, but I really think a neighborhood of ~35k with no rapid transit should be prioritized for a new line over a wealthy business district of less than 5000 residents with transit access that's already pretty good, not to mention some of the major redvelopments happening in some of the industrial areas and potential for a pretty major increase in density over time.
This goes back to the fundamental question of whether Urban Ring should primarily serve residential areas or employment centers. My opinion is still the latter. It sounds more natural to have a circumferential route such that riders from other lines (with a much bigger ridership base) can make a single transfer to work, compared to neighborhoods on the route (smaller ridership base) having a station but still needing a transfer to a radial line when heading downtown.

And Seaport is primarily a destination node, not a neighborhood. Using Fall 2022 data, during AM peak on an average weekday across all of SL1/2/3, Courthouse station sees 368.1 alightings outbound, but only 65.5 boardings inbound. (SL3 alone from Chelsea and Airport drops off 69.1 passengers.) The same applies to WTC, though the contrast is less extreme (outbound offs are 2x inbound ons).

South Boston obviously needs better transit options (the 7 bus was probably the most frequent route during AM peak pre-Covid), but my hypothesis is that it's better served with light rail that's largely along the proposed T7 route (and possibly T9 too).
 
Last edited:
So Urban Ring will probably not need cars as big as OL, even if they have the same ridership. The billion dollar question is, by how much? My guess is that the load will be low enough that 2-car or 3-car Type 10s can handle them, if other circumstances call for LRT.
One problem though is that with lower frequencies, like you see on the OL or RL, and smaller trains you'd likely get a few areas with much higher ridership, such as between Nubian and Longwood, where you have employees/students commuting in from Roxbury, Dorchester, and all along the OL transferring onto this new line in a very small area, which could lead to significant overcrowding at times. There's two potential solutions here: Bigger trains or more trains. I think the latter is the better option here, given that the large trains will be as you said much less useful on the rest of the network, and increase station costs significantly. I think smaller trains, (See Copenhagen or Vancouver as an example) with super high frequencies made possible by automation (That can only be accomplished with total grade separation) are a better solution as they allow you to build smaller stations, reducing construction costs, and run service very frequently, decreasing the amount of time people need to wait for transfers. Higher frequencies will also translate to better ridership across the rest of the line, somewhat mitigating the problem of trains that are quite empty for big chunks of the line, but in some ways for residential areas that can be a feature, not a bug.

Which leads me to the question of residential vs commercial areas. If this were 2019, I'd say mainly commercial areas, obviously stopping at residential areas along the way. But things have changed with WFH taking off, way fewer people are making "traditional" commutes downtown for 9-5 office jobs, and this is probably a trend that will continue going forward. With that in mind, I think new transit should primarily focus on bringing people to recreational, shopping, and other residential areas.

Not to mention business areas are disproportionately wealthy, and you very much run the risk of creating a transit line for rich people and leaving everyone else behind, which is... Not great.
 
Not sure if I trust the practicality of these numbers - I agree they're probably too high. 200k is about the daily ridership on the Red and Orange lines pre-Covid (Red slightly above, Orange slightly below).
FWIW, I think it's pretty common for ring routes to have among the highest riderships, looking at say, Moscow or Tokyo for example. And the rings have the highest riderships for a lot of the reasons you explain: lots of people doing short trips and transfers. So if MBTA was predicting Red Line level usage, I'd believe it. Though obviously "red line level usage" has chanced since that report.

Also FWIW, a brief look around the world, the ring routes around the world run full metro. Paris is building one now. One notable exception is the planned IBX in New York, which is planned to be light rail. That's not exactly a "ring" route, but seems to serve a similar purpose. None of this is me supporting one mode over another, just there are a lot of ring examples out there that'd be interesting to look at.
 
FWIW, I think it's pretty common for ring routes to have among the highest riderships, looking at say, Moscow or Tokyo for example. And the rings have the highest riderships for a lot of the reasons you explain: lots of people doing short trips and transfers. So if MBTA was predicting Red Line level usage, I'd believe it. Though obviously "red line level usage" has chanced since that report.

Also FWIW, a brief look around the world, the ring routes around the world run full metro. Paris is building one now. One notable exception is the planned IBX in New York, which is planned to be light rail. That's not exactly a "ring" route, but seems to serve a similar purpose. None of this is me supporting one mode over another, just there are a lot of ring examples out there that'd be interesting to look at.
True, but Paris, Moscow and Tokyo also have a much bigger network and greater political will to build transit than Boston or most places in the US.

I believe NYC's Interborough Express is using light rail primarily because of a short non-grade-separated section near All Faiths Cemetery. In other words, they indeed have 95% grade separation, but uses light rail to avoid doing the last 5% at disproportionately greater cost.
 
True, but Paris, Moscow and Tokyo also have a much bigger network and greater political will to build transit than Boston or most places in the US.

I believe NYC's Interborough Express is using light rail primarily because of a short non-grade-separated section near All Faiths Cemetery. In other words, they indeed have 95% grade separation, but uses light rail to avoid doing the last 5% at disproportionately greater cost.
That's not totally true. The existing line is 100% grade separated, but the proposal seems to suggest that the freight line should remain entirely intact, therefore resulting in some weird grade crossings, bi-level sections, and other oddities. I really don't see why you couldn't just run conventional rail service along the line and install passing loops to allow for simultaneous passenger/freight service, since to my knowledge, (Correct me if I'm wrong) there aren't actually that many freight trains that use the line.

It's like if Boston proposed building rapid transit along the Fairmount line that in order to not disturb the corridor, ran entirely parallel or above the existing tracks, rather than just, using the existing tracks.
 
That's not totally true. The existing line is 100% grade separated, but the proposal seems to suggest that the freight line should remain entirely intact, therefore resulting in some weird grade crossings, bi-level sections, and other oddities. I really don't see why you couldn't just run conventional rail service along the line and install passing loops to allow for simultaneous passenger/freight service, since to my knowledge, (Correct me if I'm wrong) there aren't actually that many freight trains that use the line.

It's like if Boston proposed building rapid transit along the Fairmount line that in order to not disturb the corridor, ran entirely parallel or above the existing tracks, rather than just, using the existing tracks.
I think it's because the freight tracks still needed to be used by freight. As for why they didn't just run mainline rail, no idea.
 
I think it's because the freight tracks still needed to be used by freight. As for why they didn't just run mainline rail, no idea.
That's exactly my point, and frankly I'm slightly worried something similar could happen with the Fairmount Line.
 
That's exactly my point, and frankly I'm slightly worried something similar could happen with the Fairmount Line.
For Fairmount Line in particular, I'm less worried. Commuter rail is already running, not to mention there's more than enough political will to push for more frequent service and electrification.
 
That's not totally true. The existing line is 100% grade separated, but the proposal seems to suggest that the freight line should remain entirely intact, therefore resulting in some weird grade crossings, bi-level sections, and other oddities. I really don't see why you couldn't just run conventional rail service along the line and install passing loops to allow for simultaneous passenger/freight service, since to my knowledge, (Correct me if I'm wrong) there aren't actually that many freight trains that use the line.

It's like if Boston proposed building rapid transit along the Fairmount line that in order to not disturb the corridor, ran entirely parallel or above the existing tracks, rather than just, using the existing tracks.
I think it's because the freight tracks still needed to be used by freight. As for why they didn't just run mainline rail, no idea.
There's not a lot of freight on the line. One daily New York & Atlantic RR local between Fresh Pond Yard (at the junction with the Lower Montauk Branch) to the car float at the end of the line, one daily CSX yard-stocker job from the NEC to Fresh Pond, and an overnight P&W yard-stocker job from the NEC to Fresh Pond. The entire line is double-track width with some sections that are tri-track width. The freight interference is virtually nil. They could do a frequent LIRR shuttle trivially. The Rube Goldberg-esque LRT crazy quilt is designed to tank the project because the MTA doesn't want to build it, and the Governor wants to make a show of trying to build it but won't actually be around when it's (not) built. Tankapalooza FTW.


For Fairmount Line in particular, I'm less worried. Commuter rail is already running, not to mention there's more than enough political will to push for more frequent service and electrification.
There'd be no daytime freight on the Fairmount Line, with exception of the twice-daily backup moves at Readville Jct. off the Franklin Line connector in/out of the CSX yard. They have to cross the would-be Southampton diamond to get to South Boston, and there's just too much commuter and Amtrak traffic during the day for any slots to be available. If Track 61 freight ever returns, it's understood that that would be a fully nocturnal job just like the fully nocturnal jobs to the cold storage warehouse @ Widett Circle that used to run up until a couple years ago when they stopped receiving by rail.
 
Last edited:

Back
Top