Crazy Transit Pitches

[ritchiew]

The industry-standard 82' radius

You've mentioned this number a few times lately, but where does it come from? Sorry if you cited it somewhere and I've missed it.
Also, what exactly is the benefit of following that standard? That's to let some specific train design through? Or allows certain speeds?

 

[JeffDowntown]

You've mentioned this number a few times lately, but where does it come from? Sorry if you cited it somewhere and I've missed it.
Also, what exactly is the benefit of following that standard? That's to let some specific train design through? Or allows certain speeds?

Track Design Handbook for Light Rail Transit, Second Edition | The National Academies Press

Read online, download a free PDF, or order a copy in print.

nap.nationalacademies.org

Some recommended reading.

The 25 Meter/ 82 foot standard allows modern systems to select among a range of standard LRV vehicles without extensive and expensive truck customization.

Older legacy systems (with tighter curves) have to do extensive (and expensive) engineering qualification for truck designs, that are then produced in low volumes further inflating the price per vehicle. Tight curves also put a speed and throughput penalty on the entire network. (Less bang for your tunneling bucks.)

 

[Riverside]

You've mentioned this number a few times lately, but where does it come from? Sorry if you cited it somewhere and I've missed it.
Also, what exactly is the benefit of following that standard? That's to let some specific train design through? Or allows certain speeds?

Track Design Handbook for Light Rail Transit, Second Edition | The National Academies Press

Read online, download a free PDF, or order a copy in print.

nap.nationalacademies.org

Some recommended reading.

See for example page 1-4 and in particular page 3-3 and 3-4, and 3-11.

I also highly recommend this master's thesis from MIT: Strategies for meeting future capacity needs on the light rail MBTA Green Line In particular, Appendix G (pg 209 of the PDF) provides a thorough overview of all of the constraining curves on the Green Line, as well as prospects for remediating them. The whole document is amazing though.

And yeah, @JeffDowntown is right on the money: modern standard curve radii make it possible to buy modern standard vehicles without customization, and makes it possible to run them faster.

In terms of "ruling radii" in a "Crazy Transit Pitches" future:

• Brattle Loop, 50 feet: the curve probably can't be widened, but potentially something clever could be done to convert the loop into a stub-end terminal, and avoid the sharp curve altogether
• Park St Inner Loop, 49 feet: difficult to widen, but could potentially be removed or reworked into a set of parallel tracks that could still enable turnbacks
• Kenmore Loop, 55 feet: I have yet to see any convincing proposal for widening this one
• Lake Street Yard, 50 feet, 49 feet and 45 feet: if revenue terminating ops can be relocated to the median of Comm Ave, then I think flexibility opens up to rework this plot of land as needed; for example, I wonder if you could "trim" the northern part of the yard to cut out the sharpest turns (see sketch below -- red is deleted track, light green is new track + platform); it looks like there is also some room to spare on the northern end of the plot, so potentially the curves could simply be expanded and otherwise left untouched
• Reservoir Yard, 50 feet: the curves going in to Prendergast Ave are the limiting factor here; they could potentially be expanded a bit, but not much
• Heath Street Loop, 50 feet: a number of solutions, including conversion to stub-end

The thesis linked above suggests that a reasonable target could be 60-feet radii for revenue tracks and 50-feet radii for non-revenue, which I agree with. It does mean that my example of the Kenmore Loop as the kind of status quo we'll need to live with isn't actually the most precise -- in that framework, it would either be the Brattle Loop or Reservoir Yard, or a stretch of southbound track just south of Haymarket which would be the ruling radius (depending on whether we're using the revenue or non-revenue number). Also, yeah, my recollection that the Kenmore Loop is 60 feet is just plain wrong, so I dunno what's up with that.

But regardless, either way, a Comm Ave <> Brookline Ave curve being slightly smaller than industry standard would not make a huge difference.

 

[Riverside]

Some meandering ramblings incoming...

(I’ve had a tough week IRL, so this is definitely some escapist coping at work here.)

Recently someone asked about possibilities that get opened up if the LRT fleet is split -- instead of finding one model that fits all needs, use multiple models for different lines. In general, my initial thought is that this does not enough benefit to outweigh the costs; separate fleets reduces operational flexibility and probably increases maintenance costs.

But, as a thought exercise, I want to game out the ruling radii if the existing network were split into "modern" and "legacy" systems. Building off the approach in the Green Line Reconfiguration thread, let's say that the "legacy" system is:

• B Line
• C Line
• "Fenway Branch" (Kenmore <> Brookline Village)
• Boylston Street Subway
• Park <> Boylston inner tracks

And the "modern" system is everything else, with additional modifications including:

• Extended Huntington Subway
• Elimination of Heath St Loop
• D <> E Connector
• Reroute of Huntington Subway to Pleasant Street Portal

For the "legacy" system, the ruling radius would probably remain 50-foot non-rev/60-foot rev -- maybe could be raised to 55-foot non-rev if Lake St and Reservoir can be cleaned up. On the other hand, perhaps you could keep the Park St loop as-is and live with the "legacy" fleet just using a 45- or 49-foot radius -- it would still require custom vehicles, but perhaps with a more modest set of customizations.

For the "modern" system, I think the Gov't Center <> Haymarket stretch still poses the most challenges. If we abandon the Brattle Loop (or otherwise reconfigure some sort of stub-end terminal there to short-turn northside trains), we are stuck with the 60-foot radius going southbound into GC. If that can be widened, and a few small curves in the Riverside and Reservoir yards can be modified, then you could have a "modern" system-wide 66-foot minimum radius.

Hmm.

The "legacy" system would have a total length of 10 miles (including the Fenway Branch and excluding an A Line, which happen to be broadly equal length). (An additional 2.5 miles for the Mattapan Line, but that should be covered by the Type 9s for a generation, so can be excluded from this discussion.) The "modern" system would start at about 17 miles with the existing components, with potential additions of:

1. Needham, 3.75 miles
2. Porter, 1.5 miles
3. Nubian, 2 miles
4. Grand Junction (Cambridge), 2.25 miles
5. Sullivan, 1 mile
6. Sullivan <> Chelsea, 3 miles
7. Watertown, 4.5 miles
8. Waltham/128, 8 miles

(System length isn't the perfect proxy for rolling stock needs, but I'm just using it for a rough estimate here.)

The existing system of 27 miles uses about 200 vehicles, which I believe is supposed to be handled by the order of the Type 10s, which will necessarily need to be usable on both the legacy and modern systems. Assuming at least some of the above extensions are built within the lifetime of the Type 10s, at some point there will be need to be another order -- so, the question becomes, at what point could/should subsequent orders be "modern-only"?

Let's assume 10/27, or 37%, of the Type 10 fleet gets devoted full-time to the legacy system. The modern system, with each progressing expansion, would increase in length from 17 miles to:

1. Needham, 21.75 miles, 128% of current
2. Porter, 23.25 miles, 136%
3. Nubian, 25.25 miles, 149%
4. Grand Junction (Cambridge), 27.5 miles, 162%
5. Sullivan, 28.5 miles, 168%
6. Sullivan <> Chelsea, 31.5 miles, 185%
7. Watertown, 36 miles, 211%
8. Waltham/128, 44 miles, 259%

Making totally unqualified wild guesses, the existing fleet could probably be "stretched" to handle extensions to Porter and Needham, especially with some initial frequency adjustments on the Needham branch. But once you add on the Nubian Branch, or any of the "Urban Ring" services, I think you're looking at another order of vehicles.

So... could the "Type 11" fleet be modern only? I think so. In terms of the modern system, really the only "problem spots" that would remain in use from the current system are the situations at Government Center. Everything else would either be new builds, or utilizing existing tracks with good radii. So, assuming isolated ops, the modern system could be moved off of custom-builds relatively early on.

Could an initial small Type 11 order be useful even without interoperability on the "legacy network"? Again, I think so. Assuming the T11s go into service at the same time as a Nubian branch, the modern system will be about 25 miles long, while the legacy system will only be 10 miles. By my math, a small order of T11s sufficient to "top off" the LRT fleet to levels roughly equal to today would have the T11s comprising about 27% of the overall fleet, leaving 73% of the fleet available and flexible to serve the 10 out of 35 miles (28%) that would require "special" cars. The modern system would be so much larger than the legacy system that the T11s would probably always go to the modern system anyway. (And there wouldn't actually be that many opportunities for vehicles to swap from one network to the other, since they would be largely physically separate by design.)

(And I think that the T currently operates under a restriction broadly similar to this: 13% of vehicles cannot serve ~19% of the system, due to Type 9s not being usable on the E Line.)

I think overall this boils down to, “The Type 10s should support both the modern and legacy systems, but the next order of vehicles — the hypothetical ‘Type 11’ — could/should potentially be built to handle only the modern system, with a 66-foot ruling radius, assuming that the order is being made due to system expansion, and not due to replacing the Type 10s.”

The only additional wrinkles I see are the potential “new” street-running builds; branches to Nubian, Allston/Harvard, and cross-Longwood would be done largely outside of existing ROWs and in more direct interface with the open environment. I haven’t looked through those alignments carefully enough to be sure that a 66-foot radius is possible on all of them. That said, alignments in those locations would need to avoid excessively sharp turns anyway in order to prevent wheel squeal. So, hopefully, the one problem would solve the other.

Finally, for completeness’s sake, these days I mostly am thinking about non-interlined Urban Ring services. So, if those get built first, then arguably you could justify a non-legacy Type 11 build that much sooner.

 

[Brattle Loop]

• Brattle Loop, 50 feet: the curve probably can't be widened, but potentially something clever could be done to convert the loop into a stub-end terminal, and avoid the sharp curve altogether

My ears are burning

I've never been convinced any kind of stub-end terminal is possible at all at Brattle.

There's an enormous structural element along the track (covered by the cinderblock wall on the narrow end of the westbound track), where Scollay's north entrance used to be. Unlike the cinderblock wall, that's probably immovable, and would basically prevent using the main wedge as a terminating platform on that end.

The wedge-side platform on the eastbound side isn't quite long enough for a current two-car train, and I'm not at all convinced you could regularly use it as a stub-terminal. It's also difficult to extend that tangent because of the proximity of the elevators.

Best bet would probably be using the abandoned second platform across the tracks behind the cinderblock wall, though I don't know how sharp the curve into there is. I also don't know if it would be feasible to use a platform like that where you'd basically have to cross the tracks to access it (or else reconfigure all that space where the cinderblock wall currently ends, though there is room back there to do it).

The other problem is that the tracks just aren't configured for two-way operation on the Brattle Loop. The two ends of the loop track aren't really on the same alignment, and you'd have to take out a bunch of 100+ year old support columns to connect them before you hit the Haymarket platform, and the end result would probably be so low-capacity that it wouldn't be worth doing.

 

[Riverside]

My ears are burning

I've never been convinced any kind of stub-end terminal is possible at all at Brattle.

There's an enormous structural element along the track (covered by the cinderblock wall on the narrow end of the westbound track), where Scollay's north entrance used to be. Unlike the cinderblock wall, that's probably immovable, and would basically prevent using the main wedge as a terminating platform on that end.

The wedge-side platform on the eastbound side isn't quite long enough for a current two-car train, and I'm not at all convinced you could regularly use it as a stub-terminal. It's also difficult to extend that tangent because of the proximity of the elevators.

Best bet would probably be using the abandoned second platform across the tracks behind the cinderblock wall, though I don't know how sharp the curve into there is. I also don't know if it would be feasible to use a platform like that where you'd basically have to cross the tracks to access it (or else reconfigure all that space where the cinderblock wall currently ends, though there is room back there to do it).

The other problem is that the tracks just aren't configured for two-way operation on the Brattle Loop. The two ends of the loop track aren't really on the same alignment, and you'd have to take out a bunch of 100+ year old support columns to connect them before you hit the Haymarket platform, and the end result would probably be so low-capacity that it wouldn't be worth doing.

I have no clue if this would work, and as outlined in the linked thesis there would need to be additional modifications to the curves to reach a 66-foot radius, but my general thinking was:

Current northbound Brattle Loop track is abandoned and/or converted to storage -- abandoned bits in red, potential storage/ancillary track in grey.

Southbound Brattle Loop track (on the western side of the loop, shown in "forest green") is flipped to run northbound, and a new curve of track (light green) at the north end connects to the northbound track just outside of Haymarket. At the opposite end, it runs into the reopened second platform and ends in a stub.

The often-photographed in-station curve is abandoned and filled in with a platform. This provides access to the second platform without needing to cross tracks.

A new crossover between the outer southbound track and the west side Brattle Loop track provides access to the stub platform for southbound trains.

Much more speculative is slotting a second track along the second platform, in order to avoid creating bottlenecks -- marked with a dashed line because it's so speculative. I suppose that's a possible use for the otherwise abandoned east side Brattle Loop track -- rework the switches to provide access from the southbound track at Haymarket, and take turning trains out of service at Haymarket... not ideal but.

~~~

After sketching up all that, I found this: https://web.archive.org/web/2016120...ects_List/Gov Center Station Presentation.pdf

Page 19 has this:

Which I then doctored into this:

The crossover in red is mandatory, although it doesn't have to be in that specific location. The second orange track, as well as the short extension of the extant track, are optional -- the second track both requires expanding the footprint of the station and potentially would require too sharp of a curve, thus negating the benefit of the entire exercise.

The tangent stretch of the west side of the Brattle Loop should be enough to berth one T10 car (at 114'); you could maybe fit a pair of T10s by lengthening the track, but that would require more significant modification to the station in order to avoid the elevators etc.

But yeah, I really don't know. IIRC, F-Line once had an interesting idea for quad-tracking GC station, which would certainly be ambitious.

Actually, I suppose if you can figure out a dual crossover at Haymarket, you could turn both legs of the Loop into tangent tracks and point them inward toward the core of the station:

Aight... I need to stop escaping reality now and deal with IRL. Alas.

 

[Brattle Loop]

Southbound Brattle Loop track (on the western side of the loop, shown in "forest green") is flipped to run northbound, and a new curve of track (light green) at the north end connects to the northbound track just outside of Haymarket. At the opposite end, it runs into the reopened second platform and ends in a stub.

That area with the light green added near the Haymarket platform is full of ancient support columns. I don't know how feasible reconstructing that area (which would be essential to making the turnback work) would be.

But yeah, I really don't know. IIRC, F-Line once had an interesting idea for quad-tracking GC station, which would certainly be ambitious.

I'm pretty sure that idea would have required obliterating and relocating all the elevators and two of the staircases to the Blue Line, which I think would put it in God Mode territory.

I think your idea's probably feasible (dunno about the optional second stub track) if the connection can be made at Haymarket. I also think it would probably cost too much for the T to bother. (And as long as they insist on keeping the Park Street loop, the point is moot).

 

[Riverside]

Prompted in part by the headache I described in the Green Line Reconfiguration thread, I'm spitballing a bit here.

When we’ve talked about putting LRT on the Grand Junction (the “Gold Line” in this post), we usually have figured on connecting it to Sullivan via the Green Line Maintenance Facility (GLMF). This makes sense in a lot of ways, because the yard itself stretches close to halfway from Brickbottom Junction to Sullivan, so the thing’s practically already half built.

However, as I outlined in the Green Line Reconfiguration thread, that alignment creates headaches when trying to plot a northbound course from Cambridge through Brickbottom Junction to the GLMF. It’s somewhat feasible going southbound, but much harder going northbound, requiring either a massive viaduct crossing over the commuter rail maintenance facility, or a tunnel cutting through the heart of the junction.

What does that headache get you? To overstate the point somewhat, the GLMF alignment gets you a one-seat ride from Sullivan to Cambridgeport that manages to miss direct transfers with all three transit lines it crosses (Medford Branch, Union Branch, Red Line). The missed transfer to Red isn’t the end of the world, in part because Kendall is probably more of a destination for this service rather than a transfer node.

...

All of which is to say, a GLMF alignment for the Gold Line isn’t as simple as it first appears, and also has significant drawbacks.

...

So then we need to find a path that goes Sullivan <> East Somerville <> McGrath <> Kendall. And that… that might be doable.

Sullivan <> East Somerville is sorta straightforward. Washington St itself would require a subway or el, but New Washington St just to the south already has a freight rail track running along it, with potentially as few as two grade crossings and enough space between buildings to create a dedicated two-track surface ROW.

From the other direction, Kendall to McGrath Station should be relatively straightforward: tracks rise up from the Grand Junction ROW on to a viaduct that crosses over the Union Branch, putting an elevated platform perpendicular on the northern side of the Green Line above Somerville Ave Extension.

Getting from McGrath Station to East Somerville is hairier, but… well, there already is an elevated structure traversing most of that distance: McGrath Highway itself. If McGrath can be put on a road diet and dropped to surface level, and the elevated decking turned over to LRT, then you can have an el that isn’t any closer to houses/business than the current state.

Then you need to get between the McGrath Highway viaduct and the surface ROW on New Washington. This would require about 1000 feet of a greenfield elevated, either over Washington St proper, or over the adjacent lots (presumably purchased by the Commonwealth), with an elevated station above/above-and-north-of the current East Somerville station.

And… there you go. A Kendall <> Sullivan LRT line that provides transfers to both branches of the Green Line, and which serves actual neighborhoods instead of a no-man’s-land of maintenance facilities.

This alignment is about 1.3 miles, as opposed to the GLMF alignment which is about 1.1 miles. However, the East Somerville alignment would serve more riders, provide better transfers, avoid costly modifications to Brickbottom Junction, avoid operational conflicts with the Green Line, and would do so primarily by using rights-of-way that are already devoted to transportation.

...

Visual:

View attachment 34375

Coming back to this silliness, and partially inspired by an alternative @Teban54 put forward here, I've been noodling around with using a short tunnel under Fitchburg St to connect the Grand Junction line with the Medford Branch's yard leads, from which a flying junction could then grab the New Washington St alignment described above.

The tangent part of the tunnel would only need to be 650' (assuming you do an open cut southwest of McGrath Hwy to drop to -1), followed by a pair of ~250' tunnels to connect to the yard leads to the north (with open cut descents), and about ~250' of tunneling to the south to connect to the yard. I eyeball that at 650' of double track, plus cumulatively about 1000' of single tracks.

For a SWAG at costs, I'll use my metric of "$1B per [double track] mile", and arbitrarily assume that a single track tunnel is... I dunno, 75% the cost of a double-track tunnel? 650 feet = .12 miles, and 1000 feet = .19 miles, so

$1.00B x 0.12 miles = $120.0M
+ $0.75B x 0.19 miles = $142.5M
= $262.5M

(Comparing this to one of the elevated alts, eg via McGrath & Poplar, that's about 1760 feet, about 30% of a mile, and assuming costs similar to subways, that would be $300M.)

The two things that terrify me about this version of the silly idea:

1. The silliness that would be needed to maintain the yard connections
2. Tunneling under Squires Bridge, particularly trying to thread a needle between the support pillars -- I think there's a path but...

 

[Teban54]

Coming back to this silliness, and partially inspired by an alternative @Teban54 put forward here, I've been noodling around with using a short tunnel under Fitchburg St to connect the Grand Junction line with the Medford Branch's yard leads, from which a flying junction could then grab the New Washington St alignment described above.

View attachment 43523

The tangent part of the tunnel would only need to be 650' (assuming you do an open cut southwest of McGrath Hwy to drop to -1), followed by a pair of ~250' tunnels to connect to the yard leads to the north (with open cut descents), and about ~250' of tunneling to the south to connect to the yard. I eyeball that at 650' of double track, plus cumulatively about 1000' of single tracks.

For a SWAG at costs, I'll use my metric of "$1B per [double track] mile", and arbitrarily assume that a single track tunnel is... I dunno, 75% the cost of a double-track tunnel? 650 feet = .12 miles, and 1000 feet = .19 miles, so

$1.00B x 0.12 miles = $120.0M
+ $0.75B x 0.19 miles = $142.5M
= $262.5M

(Comparing this to one of the elevated alts, eg via McGrath & Poplar, that's about 1760 feet, about 30% of a mile, and assuming costs similar to subways, that would be $300M.)

The two things that terrify me about this version of the silly idea:

1. The silliness that would be needed to maintain the yard connections
2. Tunneling under Squires Bridge, particularly trying to thread a needle between the support pillars -- I think there's a path but...

Great idea! While the thought of extending the tunnel south just for the yard leads may seem silly at first glance, I think there's great value in that it allows Urban Ring trains to access the Green Line maintenance facility, thus adds a lot more operational flexibility.

This also allows the line to be built in phases, with the first phase being Sullivan-BU for example, without an immediate need for its own yards nor a connection from BU to the Green Line system.

Both of these would be trivial under the GLMF alignment, but a lot less so here, especially if a connection to Commonwealth Ave isn't built (like you proposed a few times). Basically, even if Urban Ring is operationally independent from the Green Line - which has its own merits, save for a possible GL Grand Junction branch - the non-revenue connection between the two light rail systems can be important.

Aside: I'm surprised at how "cost-effective" the tunneling alternative seems to be. I guess the culture on this forum is to avoid tunnels as much as possible, but this could be a case where a short subway is totally worth it. (The effectiveness of a tunnel underneath GLMF for the more mainstream alignment is another question.)

 

[Riverside]

Great idea! While the thought of extending the tunnel south just for the yard leads may seem silly at first glance, I think there's great value in that it allows Urban Ring trains to access the Green Line maintenance facility, thus adds a lot more operational flexibility.

Yeah, to clarify, the “silliness” is more that it’ll be a complicated tunnel + flat junction to build — I have a feeling it would have some slightly ridiculous build requirements.

This also allows the line to be built in phases, with the first phase being Sullivan-BU for example, without an immediate need for its own yards nor a connection from BU to the Green Line system.

Yeah, exactly. This approach would also make it trivial to enable Lechmere <> Sullivan/beyond service; while I don’t personally think that’s the best service pattern to run, this would at least enable it.

Basically, even if Urban Ring is operationally independent from the Green Line - which has its own merits, save for a possible GL Grand Junction branch - the non-revenue connection between the two light rail systems can be important.

Yeah, definitely agree about the utility of the non-revenue connections. That said, a Sullivan <> GLMF connection could still be enabled via access from the east (along the historical alignment of the Grand Junction). But you would still need some sort of connection for Medford Branch trains, which could be annoying.

(Oh hmm actually, I have an idea for that. This won’t make any sense written out, I’ll need to sketch it out, but potentially you could provide the needed access by running trains counterclockwise out from the eastern end of the GLMF.)

Aside: I'm surprised at how "cost-effective" the tunneling alternative seems to be. I guess the culture on this forum is to avoid tunnels as much as possible, but this could be a case where a short subway is totally worth it.

Well, yes and no, I think. The name of the game is of course gonna be the mitigations. (Am I thinking of the right word there? Been a long week.) This alignment wins some points for hooking into existing ROWs and partially leveraging the existing flying junction between Medford and GLMF. But I think tunneling directly under Squires Bridge — perpendicular to the bridge and at an angle to the street and rail ROW — could get dicey fast.

(And also having to tunnel under the Medford Branch viaduct — that could get complicated, depending on how much you are able to use the existing cross-under.)

But, in terms of absolute elevation shift, the tunnel wins out because — AFAIK — it only will be necessary to go down to -1, whereas the el will either need to up to +2 (assume that ground level = 0), or will need to go a more roundabout route to bypass the bridge.

I haven’t been able to find much costing estimates on per-mile cost for elevateds; it seems like it should be cheaper than tunnels, but the limited data I’ve found suggests that it’s around the same order of magnitude.

But, your underlying point remains: the actual raw distance that needs to be covered here is quite short. Tunneling under Brickbottom Junction proper (as I’ve proposed previously) is daunting due to the complexity of what’s on the surface, but the distance there would also be rather short.

I’m gonna take a look to see how difficult it would be to tunnel parallel to the ROW under the bridge rather than at an angle; at least then you wouldn’t need to move any support columns. (I hope.)

 

[Teban54]

Another concern that was raised the last time we discussed this: I think F-Line mentioned the track between East Somerville area and Sullivan is currently used by freight trains and can't be removed easily. That's why his own proposal involved an open cut beneath the freight ROW while heading down Grand Junction. There's a chance this can kill the entire proposal outright.

 

[Riverside]

Another concern that was raised the last time we discussed this: I think F-Line mentioned the track between East Somerville area and Sullivan is currently used by freight trains and can't be removed easily. That's why his own proposal involved an open cut beneath the freight ROW while heading down Grand Junction. There's a chance this can kill the entire proposal outright.

Yeah, my take on this is that there is ample room for three tracks along the New Washington corridor, to say nothing of an elevated. Again, this option would very likely be more expensive than the GLMF alignment, but potentially you get more bang for your buck, with transfers to both Green branches. Plus I have my eye on a Lowell Line infill at East Somerville, to provide a connection to Kendall, mirroring the role Porter currently plays.

Let me try sketching out some of the alts I tossed out last night...

 

[Riverside]

A bit more tunneling in this one, but Squires Bridge is traversed instead via the Somerville Ave Extension, which hopefully should reduce the potential impact.

Tunnel is longer in part because of the need to traverse the parking lot of 61 Medford Street; in my previous sketch, the Gold Line stayed within the railroad ROW south of McGrath, so the open cut was more feasible, but I assume 61 Medford will want their parking lot back.

Also worth noting that the curves are still a bit dicey here. I want the segment under Squires Bridge to be tangent, and for the curve only to start once you are clear of any of the support structures. I think there is just barely space to do so with an 82' curve while still leaving a tangent stretch to the south to build a station. But it's a little tight.

(I did look at putting the platform underneath Somerville Ave Extension, but I think it is too narrow to do so without impacting the Bridge. Doesn't mean it's not doable, just more expensive.)

In terms of yard access, what I was thinking about last night is what you potentially get by handling yard access from the east. Upon sketching it out, I'm less convinced there's anything worthwhile in it, but:

From GLMF to:

• Sullivan: K -> B
• Medford: H -> E *
• Union: current route (forgot to put in a marker, oops)
• Grand Junction: K -> C -> D -> G
• Lechmere: K -> C -> D -> E -> Lechmere, or
• Lechmere: J -> use existing crossover

To GLMF from:

• Sullivan: B -> K, or
• Sullivan: A -> D -> E -> H
• Medford: E -> H *
• Union: currently unbuilt provisioned route (forgot this marker too)
• Grand Junction: G -> D -> C -> K
• Lechmere: J

The * next to E and H is that I've just sorta scribbled in the idea of adding splits to the yard from the viaduct itself, i.e. going south from Medford, you'd first hit E where there'd be a full crossover with the outer tracks which would split off and down into a tunnel, and then a bit further on there would be a second split with outer tracks dropping down to ground level to head into the yard. And I really have no idea if that would work.

But yeah -- my point is that GLMF <> Gold access could be handled by some sort of connection near Sullivan -- either B or C or both. But that still leaves unsolved the issue of running trains in and out of the Medford branch.

 

[Charlie_mta]

Your straight shot route under the Squires Bridge (see your map below) is preferable, and probably doable.
Looking at Google street view from the Somerville Ave extension, it looks like a straight shot tunnel could be threaded between the bridge piers, left of the one way sign in the Street View shot.

 

[Riverside]

I went back to my design for a Kendall <> GLMF <> Sullivan alignment, and realized it may be possible to harmonize it with the concept for a Kendall <> East Somerville <> Sullivan alignment:

The western half of the junction remains as I proposed in February. The eastern half uses a similar topology, but a different alignment. For the westbound track, instead of using the GLMF -> Union yard lead, a new (tightly squeezed but I think still 82') curve connects to the Medford yard leads. (See detail below.) The eastbound track still uses a tunnel, but instead curves around to also connect to the general vicinity of the Medford yard leads.

The key benefit of this harmonized design is that it enables a future via-East Somerville build, while still enabling initial partial builds to enable, e.g. Lechmere <> Grand Junction service as a Phase 1.

Is this tunnel any better than my previous design's? Maybe. Certainly it's less busy aboveground for much of it, so that could make it easier to build.

The hook-in with the yard leads would require non-trivial modification:

By my reckoning, an 82' curve that clears the property line would, unfortunately, align with the "wrong" track in the yard leads. We could tighten the curve somewhat, or try to purchase a small corner of the parcel in order to shift the entire curve to the west.

Here, I've proposed a slightly inelegant design where the revenue track switches from the "southbound" lead over to the northbound lead very briefly before doing a flat crossing again over the southbound lead as trains continue through the curve.

This unfortunately shifts the tunnel further northeast, and may require cutting under the viaduct, as I've tried to vaguely suggest here.

That all being said, this is a relatively large parcel, all owned by the T, so I think there's room to play around here to find something that works.

Last but not least, this design is a little less friendly to a "McGrath" station, but it still is potentially doable -- put the southbound Gold Line platform in the subway, and the northbound Gold Line platform at ground level. The parcel is enough that the track could be straightened for a 250' tangent and still hook in to the necessary curves. (Sketch, not to scale, below.)

As discussed previously, ideally a second Gold Line track could be added north of the westbound Union track to avoid the intermingling. But, again, that could be built in phases potentially. Likewise, a future build could expand the northbound tunnel to add a second track and abandon the cross-under at McGrath altogether to enable fully isolated running. But, like I said, that's left open to the future.

 

[TheRatmeister]

Also on the Storrow topic: Assuming Storrow goes (Which it should) and some form of transit is built along the route (Which it should be), What if the C Branch was extended along Beacon St to Charlesgate Park, where it would turn north, then follow what is now Storrow Drive, making stops at Landsdowne, Kenmore, Charlesgate/Harvard Bridge, Fairfield St, Clarendon St, (One of these two should probably be named Esplanade), Hatch Shell, Charles MGH, and somewhere between Science Park GL and Lederman Park. From there, it would travel over Leverett Sq on a new viaduct before diving into the tunnel where it would serve North Station and Haymarket before terminating using the Brattle Loop at Gov. Center.

I have not given a ton of thought to this, so I'm sure there's flaws I haven't seen yet though.

 

[Teban54]

Also on the Storrow topic: Assuming Storrow goes (Which it should) and some form of transit is built along the route (Which it should be), What if the C Branch was extended along Beacon St to Charlesgate Park, where it would turn north, then follow what is now Storrow Drive, making stops at Landsdowne, Kenmore, Charlesgate/Harvard Bridge, Fairfield St, Clarendon St, (One of these two should probably be named Esplanade), Hatch Shell, Charles MGH, and somewhere between Science Park GL and Lederman Park. From there, it would travel over Leverett Sq on a new viaduct before diving into the tunnel where it would serve North Station and Haymarket before terminating using the Brattle Loop at Gov. Center.

I have not given a ton of thought to this, so I'm sure there's flaws I haven't seen yet though.

Very innovative proposal!

The first thought I have is: What does it do that a more conventional Blue Line Extension doesn't? In this case, I think it boils down to:

• Having an additional faucet for a Green Line branch to free up capacity
  • Very much needed, but there are many other ways to do this
• Connecting the Science Park area to Charles/MGH, Kenmore and further south
  • Could be a useful link for GLX riders, but it doesn't look like your C extension can stop at Science Park anyway, and if the connection is at North Station, it won't be much different from transferring to Blue Line at Government Center
• Serve local stations in West End
  • Most of the area seem to be within the walkshed of Science Park already, and it's not clear where you can put additional stations
• Less need for grade separation, allowing trolleys to run at grade with level crossings
  • But that also limits speed and service quality, while still not being ideal for pedestrians

In comparison, benefits of a Blue Line Extension (via a capped cut) include:

• Minimal disruptions to travel patterns of existing C branch riders, who likely prefer getting to Copley, Park St etc via a more direct route
• Potential for a future HRT extension beyond Kenmore

I still think BLX is a better choice, but it's certainly something to think about, and I imagine both will be examined should a Storrow road diet take place.

BTW, adding a Station at Lansdowne is likely very difficult as discussed here.

 

[Riverside]

Also on the Storrow topic: Assuming Storrow goes (Which it should) and some form of transit is built along the route (Which it should be), What if the C Branch was extended along Beacon St to Charlesgate Park, where it would turn north, then follow what is now Storrow Drive, making stops at Landsdowne, Kenmore, Charlesgate/Harvard Bridge, Fairfield St, Clarendon St, (One of these two should probably be named Esplanade), Hatch Shell, Charles MGH, and somewhere between Science Park GL and Lederman Park. From there, it would travel over Leverett Sq on a new viaduct before diving into the tunnel where it would serve North Station and Haymarket before terminating using the Brattle Loop at Gov. Center.

I have not given a ton of thought to this, so I'm sure there's flaws I haven't seen yet though.

This is exactly what I've been noodling on for the past week or so.

In my opinion, getting to Charles/MGH is easy -- or at least doesn't represent any new problems beyond those of a Blue-to-Kenmore via Riverbank alignment. (And indeed, some options open up since an extension of the Beacon Street Line could accommodate less grade separation.)

Where you go after that gets less clear.

A) just terminate at Charles/MGH -- this would sorta be a "boutique" streetcar service, so I'm not sure it needs to bend over backwards to do more than its core purpose.

B) turn east down Cambridge St and use the subway of the Red-Blue Connector (assuming the Blue Line has been rerouted out of it somehow). The big question mark here is what to do when you reach Gov't Center, which will depend heavily on where/how the Blue Line is rerouted.

C) head north to hook into the Green Line portal at Nashua Street. I definitely see the appeal here, and it's nice how tidily it makes use of the Brattle Loop. Two major downsides I see are that it's somewhat roundabout (2.8 miles vs 2.2 miles), and that it doesn't segregate the street-running super-extended C Line from the rest of the Green Line, which makes me concerned about reliability. The biggest benefits I see to riders here are the direct transfer to the Orange Line, and access to downtown.

D) continue to hug the coast of the Charles River and do some sort of surface terminal on/near Nashua St/North Station

E) as a variant on the Cambridge St alignment, you could build a median on Cambridge St and then do something on New Chardon and/or Sudbury St and terminate at Haymarket on the surface

For my part, I'd probably lean toward either options A or E -- I don't think this would be a "load bearing" service, but rather a way to increase transit access to the Esplanade and give C Line riders a bit of consolation for losing their OSR into downtown.

I see @Teban54 has replied, let's take a look...

Yeah, I agree with everything said there. And, like I said, I don't think this would be a load-bearing service, meaning I think it really only makes sense if the Blue Line is extended west via some other alignment and if the C Line somehow gets booted from the Central Subway.

 

[TheRatmeister]

Very innovative proposal!

The first thought I have is: What does it do that a more conventional Blue Line Extension doesn't? In this case, I think it boils down to:

• Having an additional faucet for a Green Line branch to free up capacity
  • Very much needed, but there are many other ways to do this
• Connecting the Science Park area to Charles/MGH, Kenmore and further south
  • Could be a useful link for GLX riders, but it doesn't look like your C extension can stop at Science Park anyway, and if the connection is at North Station, it won't be much different from transferring to Blue Line at Government Center
• Serve local stations in West End
  • Most of the area seem to be within the walkshed of Science Park already, and it's not clear where you can put additional stations
• Less need for grade separation, allowing trolleys to run at grade with level crossings
  • But that also limits speed and service quality, while still not being ideal for pedestrians

In comparison, benefits of a Blue Line Extension (via a capped cut) include:

• Minimal disruptions to travel patterns of existing C branch riders, who likely prefer getting to Copley, Park St etc via a more direct route
• Potential for a future HRT extension beyond Kenmore

I still think BLX is a better choice, but it's certainly something to think about, and I imagine both will be examined should a Storrow road diet take place.

BTW, adding a Station at Lansdowne is likely very difficult as discussed here.

In general I'd sum the advantages of this proposal over a Blue Line extension as:

• Lower cost (Obviously) Tunneling that close to the river would likely result in various engineering difficulties that would significantly inflate costs. Compared to heavy rail, light rail would not require full grade separation and regular pedestrian crossings could allow for easy access to the park, unlike heavy rail which would need to be in a cutting, with limited bridges accessing the park. You're right, it's not perfect, but compare something like the E Branch along Huntington Ave with its fairly regular crossings to the GLX/Lowell Line through Somerville. The first is definitely preferable for a public space in my opinion. Speed may be limited in sections without a dedicated alignment, but the relevant distance here is under 2 miles, and if we're being honest it wouldn't be any slower than the central subway already is.
• Kills two birds with one stone: It provides transit over the Esplanade and additional coverage in Back Bay while relieving some pressure on the Central Subway.
• Better integrates with an urban park, see the Wertheimpark in Amsterdam for an example of how this can look.

I will say, from reading your post I still get the idea that you're at least partially in the mindset of a heavy rail line, with large, complex stations. This would not be that. This would be an at-grade tram route, like the rest of the C Branch already is. A new station at Landsdowne would not be a complex structure, just a basic median street stop, same at Science Park.

 

[TheRatmeister]

This is exactly what I've been noodling on for the past week or so.

In my opinion, getting to Charles/MGH is easy -- or at least doesn't represent any new problems beyond those of a Blue-to-Kenmore via Riverbank alignment. (And indeed, some options open up since an extension of the Beacon Street Line could accommodate less grade separation.)

Where you go after that gets less clear.

A) just terminate at Charles/MGH -- this would sorta be a "boutique" streetcar service, so I'm not sure it needs to bend over backwards to do more than its core purpose.

B) turn east down Cambridge St and use the subway of the Red-Blue Connector (assuming the Blue Line has been rerouted out of it somehow). The big question mark here is what to do when you reach Gov't Center, which will depend heavily on where/how the Blue Line is rerouted.

C) head north to hook into the Green Line portal at Nashua Street. I definitely see the appeal here, and it's nice how tidily it makes use of the Brattle Loop. Two major downsides I see are that it's somewhat roundabout (2.8 miles vs 2.2 miles), and that it doesn't segregate the street-running super-extended C Line from the rest of the Green Line, which makes me concerned about reliability. The biggest benefits I see to riders here are the direct transfer to the Orange Line, and access to downtown.

D) continue to hug the coast of the Charles River and do some sort of surface terminal on/near Nashua St/North Station

E) as a variant on the Cambridge St alignment, you could build a median on Cambridge St and then do something on New Chardon and/or Sudbury St and terminate at Haymarket on the surface

For my part, I'd probably lean toward either options A or E -- I don't think this would be a "load bearing" service, but rather a way to increase transit access to the Esplanade and give C Line riders a bit of consolation for losing their OSR into downtown.

I see @Teban54 has replied, let's take a look...

Yeah, I agree with everything said there. And, like I said, I don't think this would be a load-bearing service, meaning I think it really only makes sense if the Blue Line is extended west via some other alignment and if the C Line somehow gets booted from the Central Subway.

The reason I think option C is what should be done is not necessarily because it's the best, a terminus at Charles/MGH would likely make more sense in that regard and neatly cap the service, but because quite simply good luck telling all those rich Brooklineites that no, if you want to get to Malden now you've got to change trains twice. Not happening.

But you're right that it's not a "load-bearing" service. It's not designed to be, just like the current C branch isn't. It's more of a hybrid between a local bus service and a cross-city rapid transit line. I'm also less concerned about the isolation problem seeing as the route would not use the central subway, which is by far where the worst delays occur and where contagion is incredibly bad. The line also still functions well without the turn into downtown should it suddenly become necessary because of service interruptions, but this should obviously be avoided whenever possible to minimize crowding at Charles/MGH and State/DTX.