Crazy Transit Pitches

I have more to say on the ongoing discussion about Urban Ring routings and objectives (those comments will take anywhere between minutes and weeks), but I wanted to post this one last piece about my "Binney & Gilmore" proposal:

Turns out, I may have overthought the part north of Lechmere. There seems to be a surprisingly easy way to build a flying junction that ties the Kendall-Binney-Land viaduct directly into Lechmere GLX:

Land & Binney to Lechmere.png


Basically, shift the current eastbound GLX track outward, with a split into southbound Kendall track. Northbound Kendall track flies over the EB track, and descends onto the original EB track before joining the westbound GLX track. You may need to take some space on McGrath to expand the viaduct, but it should be doable (e.g. by moving the bike lane and sidewalk underneath the viaduct).

This is basically a "harmonized" proposal that runs directly to Lechmere. From there, the easiest way to reach Sullivan is likely through the unbuilt yard lead at Brickbottom to the northern half of Grand Junction, essentially creating a Sullivan-Lechmere connection. There are some challenges with that, but it's nowhere near as complex if you don't need connections to Union Square and Kendall-bound Grand Junction. It should be quite doable without any tunnels: I may sketch it out another day or find some old posts about it.

In addition to having a same-platform transfer for all GLX riders, this alignment offers great flexibility and allows many potential service patterns (though whether you want to run them is another question):
  • Sullivan - Lechmere - Kendall, the standard Urban Ring service
  • (Everett/Chelsea) - Sullivan - Lechmere - Downtown, a "wrap-around" branch if Everett and Chelsea residents prefer the OSR over an Orange Line transfer
  • Medford/Tufts - Lechmere - Kendall, possibly a rush hour service if Somerville riders want an OSR to Kendall, and/or if GLX needs a capacity boost
The main problem is that it creates a bottleneck at Lechmere where Urban Ring interlines with GLX, potentially limiting the former's frequency. But I'm not sure if Urban Ring needs that much service.

This flying junction doesn't preclude Urban Ring from going further north via Gilmore Bridge if needed. For example, you can still do a deinterlined build to Sullivan via Gilmore (possibly as a later phase to increase capacity), but leave this flying junction as a non-revenue connection. Or you can run a streetcar branch into Charlestown.

---------------------------------------------

Some quick replies:

(Actually, I just realized: are you assessing the Fitchburg St alt, or the "harmonized" alt?)
I was mostly thinking of the Fitchburg St alt, though I don't think the difference matters too much for the summary table. The harmonized alt should score better on operational flexibility and a little bit better on cost, but possibly worse on speed.

As a side-note, I didn't mention the other major problem with using the Grand Junction, which is that now there isn't a (somewhat) good rail link between North Station and South Station for things like rolling stock transfers, which means that unless NSRL happens as far as I can tell any reuse of the alignment is dead in the water. If NSRL does happen, I think reusing the route as a busway/bike path/linear park actually makes more sense than rail service, as this is an area where feeding many local buses into enhanced local stops for the MIT campus and bajillion lab buildings nearby would make more sense then a couple consolidated stops.
View attachment 43948
This is a well-acknowledged problem, but I believe it was pointed out that future developments at Readville Maintenance Facility (?) may remove the need to use Grand Junction for equipment moves, freeing up the ROW for rapid transit. I think F-Line or someone else mentioned it before, but can't find it now.

---------------------------------------------

One piece that I do want to bring up now regarding Urban Ring.

For the moment, forget about ROWs and suppose you can build TBM tunnels anywhere without cost concerns, for either Urban Ring or other pieces of the rapid transit network. In other words, you're simply drawing a crayon map by connecting the dots.

Then... What should the goal of Urban Ring be?
  • Connecting employment centers?
  • Connecting residential areas?
    • Is it to serve transit deserts and let them transfer to another radial line, or to give these residents OSR to somewhere else?
  • Connecting existing transit hubs?
  • Just simply giving a transfer to all (or most) radial lines?
  • Trying to be competitive with a transfer in the downtown core?
  • Or a mix of the above?
    • If this is the case, how? How should we decide that workplace A deserves a direct connection to neighborhood B, but not workplace C to neighborhood D?
Personally, what I've always thought was that employment centers play a bigger role. For the most part, the majority of transit ridership still comes from people going to work (though whether we want to prioritize that is an assumption that may be challenged). And there happens to be a major employment center in most quadrants of the Urban Ring area: Kendall/Harvard, BU, Longwood, BMC (& Nubian?), Seaport, and Logan.

I think one important function of the Urban Ring should be to reduce transfers downtown and thus free up capacity there. For this to function well, 2-seat-rides vs 3-seat-rides will be significant. A Malden resident working at Kendall is much more likely to choose a Malden-Sullivan-Kendall ride via Urban Ring than Malden-DTX-Kendall. But if that Urban Ring ride becomes Malden-Sullivan-Central-Kendall or even Malden-Sullivan-Harvard-Kendall... Now distance itself matters much less, and it may not even be faster than DTX.
  • Obviously, the same can be said to someone working at Harvard. But this exercise shows that you can't really do Kendall and Harvard on the same route.
In the TBM-friendly world that I assumed, I think the ideal way to deal with residential transit deserts (looking at you, Allston) is to give them radial service, not circumferential service so that they can go all the way to Harvard in order to head downtown. They can always transfer to circumferential routes to get to their jobs.

Btw, @TheRatmeister I do very much appreciate you bringing up the topic for discussion, even though I might have sounded a bit harsh in earlier replies.
 
Last edited:
I have more to say on the ongoing discussion about Urban Ring routings and objectives (those comments will take anywhere between minutes and weeks), but I wanted to post this one last piece about my "Binney & Gilmore" proposal:

Turns out, I may have overthought the part north of Lechmere. There seems to be a surprisingly easy way to build a flying junction that ties the Kendall-Binney-Land viaduct directly into Lechmere GLX:

View attachment 43949

Basically, shift the current eastbound GLX track outward, with a split into southbound Kendall track. Northbound Kendall track flies over the EB track, and descends onto the original EB track before joining the westbound GLX track. You may need to take some space on McGrath to expand the viaduct, but it should be doable (e.g. by moving the bike lane and sidewalk underneath the viaduct).

This is basically a "harmonized" proposal that runs directly to Lechmere. From there, the easiest way to reach Sullivan is likely through the unbuilt yard lead at Brickbottom to the northern half of Grand Junction, essentially creating a Sullivan-Lechmere connection. There are some challenges with that, but it's nowhere near as complex if you don't need connections to Union Square and Kendall-bound Grand Junction. It should be quite doable without any tunnels: I may sketch it out another day or find some old posts about it.

In addition to having a same-platform transfer for all GLX riders, this alignment offers great flexibility and allows many potential service patterns (though whether you want to run them is another question):
  • Sullivan - Lechmere - Kendall, the standard Urban Ring service
  • (Everett/Chelsea) - Sullivan - Lechmere - Downtown, a "wrap-around" branch if Everett and Chelsea residents prefer the OSR over an Orange Line transfer
  • Medford/Tufts - Lechmere - Kendall, possibly a rush hour service if Somerville riders want an OSR to Kendall, and/or if GLX needs a capacity boost
The main problem is that it creates a bottleneck at Lechmere where Urban Ring interlines with GLX, potentially limiting the former's frequency. But I'm not sure if Urban Ring needs that much service.

This flying junction doesn't preclude Urban Ring from going further north via Gilmore Bridge if needed. For example, you can still do a deinterlined build to Sullivan via Gilmore (possibly as a later phase to increase capacity), but leave this flying junction as a non-revenue connection. Or you can run a streetcar branch into Charlestown.

---------------------------------------------

Some quick replies:


I was mostly thinking of the Fitchburg St alt, though I don't think the difference matters too much for the summary table. The harmonized alt should score better on operational flexibility and a little bit better on cost, but possibly worse on speed.


This is a well-acknowledged problem, but I believe it was pointed out that future developments at Readville Maintenance Facility (?) may remove the need to use Grand Junction for equipment moves, freeing up the ROW for rapid transit. I think F-Line or someone else mentioned it before, but can't find it now.

---------------------------------------------

One piece that I do want to bring up now regarding Urban Ring.

For the moment, forget about ROWs and suppose you can build TBM tunnels anywhere without cost concerns, for either Urban Ring or other pieces of the rapid transit network. In other words, you're simply drawing a crayon map by connecting the dots.

Then... What should the goal of Urban Ring be?
  • Connecting employment centers?
  • Connecting residential areas?
    • Is it to serve transit deserts and let them transfer to another radial line, or to give these residents OSR to somewhere else?
  • Connecting existing transit hubs?
  • Just simply giving a transfer to all (or most) radial lines?
  • Trying to be competitive with a transfer in the downtown core?
  • Or a mix of the above?
    • If this is the case, how? How should we decide that workplace A deserves a direct connection to neighborhood B, but not workplace C to neighborhood D?
Personally, what I've always thought was that employment centers play a bigger role. For the most part, the majority of transit ridership still comes from people going to work (though whether we want to prioritize that is an assumption that may be challenged). And there happens to be a major employment center in most quadrants of the Urban Ring area: Kendall/Harvard, BU, Longwood, BMC (& Nubian?), Seaport, and Logan.

I think one important function of the Urban Ring should be to reduce transfers downtown and thus free up capacity there. For this to function well, 2-seat-rides vs 3-seat-rides will be significant. A Malden resident working at Kendall is much more likely to choose a Malden-Sullivan-Kendall ride via Urban Ring than Malden-DTX-Kendall. But if that Urban Ring ride becomes Malden-Sullivan-Central-Kendall or even Malden-Sullivan-Harvard-Kendall... Now distance itself matters much less, and it may not even be faster than DTX.
  • Obviously, the same can be said to someone working at Harvard. But this exercise shows that you can't really do Kendall and Harvard on the same route.
In the TBM-friendly world that I assumed, I think the ideal way to deal with residential transit deserts (looking at you, Allston) is to give them radial service, not circumferential service so that they can go all the way to Harvard in order to head downtown. They can always transfer to circumferential routes to get to their jobs.

Btw, @TheRatmeister I do very much appreciate you bringing up the topic for discussion, even though I might have sounded a bit harsh in earlier replies.
Simplistic reply on the goal of an Urban Ring, but it really is about total network capacity and efficiency. And a big barrier today is the transfer capacity at the four main downtown transfer stations.

So a primary goal of the Urban Ring needs to be to out-compete the downtown transfer stations for some significant portion of total system transfer load, hence freeing up capacity in the downtown transfer stations.

As an example, a rider needs to see it as logical to go from Northeastern to Harvard by way of the Urban Ring rather than via Park or Downtown Crossing downtown transfers.
 
I've also thought an Urban Ring roughly along the #66 bus route makes sense, and didn't even know it was considered. If anyone has the report where that was considered and rejected, I'd love to see it.

Additionally, if anyone goes by the main BPL, there's a temporary exhibit in the map room on the history of Boston transportation. The whole exhibit is great, but I saw this map "Existing and Proposed Circumferential Thoroughfares" from 1909. This is city planners a century ago trying to figure out better ring transportation on top of very radial roads. Just funny to see how long planners have been trying to cope with the regions weird layout. Much better image scan here:

1698677263144.png
 
I was able to find parts of the study here, obviously the whole thing would be nice though.
Thank you! And hey hey hey, look at what we've got here:

1698681213316.png


There truly is nothing new under the sun.

Anyways, I think building out stronger subnetworks is a good idea. Kendall, Longwood, Nubian, etc are all good places to do that. But the urban ring proposal doesn't do that very well either. Instead of building local transportation hubs they're used as nodes in a long, strung out, and (relatively) low capacity circumferential route. A much better place to start would be to use infrastructure like Roxbury's wide streets or Cambridge with the Grand Junction (Assuming NSRL gets built) as corridors for local service like buses and/or streetcar/tram routes, not as weird piecemeal bits of a franken-line.
I'd argue the Urban Ring via Grand Junction does a reasonably good job of building a Kendall network, an okay job building a Longwood network, but not really anything else. But I'm really not sure I follow the rest of the points you are making. IIRC, a "supertrain" (of 2 Type 10s coupled together) will have capacity similar to a Blue Line consist -- I don't really see that as low capacity?

Setting aside the question of circumferential transit in general, I put it to you (with full sincerity, and apologies if this comes across as snark): what would you build a Kendall subnetwork like?
Turns out, I may have overthought the part north of Lechmere. There seems to be a surprisingly easy way to build a flying junction that ties the Kendall-Binney-Land viaduct directly into Lechmere GLX:

Land & Binney to Lechmere.png
I had a similar thought about hooking in to Lechmere here -- "great minds, etc etc etc". I agree that -- if you are going to do a Binney alignment -- this is definitely the most parsimonious way to do it. The only downside that I don't think you mentioned is that this would not enable North Station <> Kendall service, which is always the one persistent objection in the back of my head when I think about deinterlining Gold and Green. (I mean, I guess you could Big Even Higher at this flying junction to add on North Station <> Kendall service, but...) On the other hand, Regional Rail infills at Sullivan and East Somerville could remove the need for North Station <> Kendall service altogether.
 
Last edited:
The main problem is that it creates a bottleneck at Lechmere where Urban Ring interlines with GLX, potentially limiting the former's frequency. But I'm not sure if Urban Ring needs that much service.
Hmm, let me "think out loud" here.
Northern branches from this transfer station at Lechmere would be:
  • Gold to Sullivan/beyond
  • Green to Medford
  • Green to Porter -> Watertown and/or Waltham and/or Arlington
Assuming 6 min headways to Sullivan, Medford, and Porter, then you're looking at 30 tph, which definitely is a lot of trains, but along a segment bracketed by straightaways and flying junctions, so it wouldn't be particularly unprecedented. And while it would require a total rebuild of Lechmere to do so, the road underneath is wide enough that Lechmere station itself could be quad-tracked if capacity becomes a concern in the future.

And that capacity pinch of 30 tph can get divided up differently as well. Slightly reducing the headways to 7.5 min each would then create capacity for a fourth service (or a doubling of some service), e.g.
  • Southbound:
    • 15 tph to North Station + Park St (Green)
    • 15 tph to Kendall (Gold)
  • Northbound:
    • 7.5 tph to Medford (Green)
    • 7.5 tph to Porter + one branch (Green)
    • 15 tph to Sullivan (Gold)
Finally, it's worth noting that if we stop thinking of this as a circumferential service and start thinking of it as a Kendall subnetwork, there's an obvious option to consider: reverse-branching the Medford Branch and running some trains directly to Kendall. I'm not saying this is must-have or anything, but particularly if those Kendall trains also provided a speedy OSR to Longwood, that could potentially be compelling.
All of which is to say, I think there are options here, but there's no reason I can see that this design should not preclude 6-min-or-better headways on the Sullivan <> Kendall service.

A summary table comparing the three alignments:
Excellent summary. The only points I'd disagree on is the Feasibility & Cost of the GLMF alignment and the E Somerville alignment: I would probably put GLMF at Bad to Medium because you need to tunnel under the densest part of the junction with a longer tunnel under active tracks that the East Somerville alignment's tunnel under Squires Bridge, with another tunnel west of the bridge to go from Lechmere -> Grand Junction.

The E Somerville alignment I would probably stick at Medium because you would still need to tunnel under active rails (just fewer of them) and under the bridge, and then again under the Medford Branch viaduct.

(Actually, I just realized: are you assessing the Fitchburg St alt, or the "harmonized" alt?)
I was mostly thinking of the Fitchburg St alt, though I don't think the difference matters too much for the summary table. The harmonized alt should score better on operational flexibility and a little bit better on cost, but possibly worse on speed.
Cool, yeah in that case I stand by my assessment that the East Somerville Feasibility & Cost should be Medium at best. The Fitchburg St alignment requires a tunnel under a dual active ROW + a highway bridge, and then requires a second tunnel underneath the Medford Branch viaduct (which would in turn impact GLMF access to/from the Medford Branch) -- to me, that makes it hard to describe as "no major difficulties," unfortunately.

Annoyingly, there are of course more variations here that would impact feasibility and cost. A GLMF alignment becomes significantly more feasible (albeit lower capacity) if you can fit a flying junction in at the western end (e.g. by shifting the tracks northward so that there's room for an "on-ramp" on to the eastbound track, slotted in between the LRT and mainline tracks), and then use the yard leads (including the as-of-yet unbuilt one) to run east:

1698681370809.png


[updated table below due to char limits]
 
AlignmentGLMF with full separationGLMF with interlining under bridgeEast SomervilleBinney & Gilmore
Length and SpeedGood

  • Shortest route (1.92 mi)
  • 4 grade crossings
  • 1 stop
  • Fastest (gentle curves)
Good

  • Shortest route (1.92 mi)
  • 4 grade crossings
  • 1 stop
  • Fastest (gentle curves)
Medium

  • Slightly longer (2.00 mi)
  • 4 grade crossings
  • 2 stops
  • Medium speed (several curves near East Somerville and McGrath)
Medium for a viaduct, Bad forsurface

  • Longest route (2.41 mi)
  • 8 grade crossings
  • 2 stops (corrected)
  • Slowest speed (primarily due to route length and possible grade crossings)
Ridership and TransfersBad

  • Low catchment (1 stop at Cambridge St and/or Twin City Plaza)
  • No GLX transfer
  • No Lowell transfer
Bad

  • Low catchment (1 stop at Cambridge St and/or Twin City Plaza)
  • No GLX transfer
  • No Lowell transfer
Medium to Good

  • Medium catchment (Twin City Plaza (McGrath) and East Somerville)
  • Easy GLX transfers
  • Possible Lowell transfer (via East Somerville infill)
Medium to Good

  • High catchment (Lechmere and Kendall North)
  • GLX transfer, but long walk
  • No Lowell transfer
Feasibility and CostMedium to Good

  • Tricky infrastructure (Brickbottom Junction)
  • Reuse most surface railroad ROWs
  • Short tunnel likely required
  • Reconfigurations for GLMF and CRMF needed
  • Politically feasible
Good

  • Tunnel under active ROW with relocation
  • Reuse most surface railroad ROWs
  • Use planned designs for GLMF access
  • Short tunnel required
  • No reconfigurations for GLMF and CRMF needed
  • Politically feasible
Medium to Good (without Brickbottom connection)

  • No major difficulties Tunnel under active ROW, tunnel under Squires Bridge, tunnel under Medford Viaduct
  • Reuse most surface railroad ROWs, but may need viaduct over Yard 10 lead
  • Short tunnel required
  • No or less impacts to GLMF and CRMF
  • Politically feasible
Medium for a viaduct, Good for surface

  • No major difficulties
  • Greenfield ROW
  • No tunnels, but may need long viaducts
  • No impacts to GLMF and CRMF
  • May face political challenges (Land Blvd El, Gilmore Bridge)
Operational FlexibilityGood

  • Easy access to Green Line and GLMF
  • Possible Phase 1 for GJ service (as GL branch)
Medium to Good

  • Easy access to Green Line and GLMF
  • Possible Phase 1 for GJ service (as GL branch)
  • Interlining creates potential capacity impact
Medium

  • Can access GLMF, but tedious
  • GJ Phase 1 less likely
Bad

  • No access to Green Line and GLMF
  • GJ Phase 1 unlikely

Parenthetically -- for a Phase 1 Lechmere <> GJ service compatible with the Fitchburg St alignment, I need to look more closely, because I think in theory there should be enough space.
 
What should the goal of Urban Ring be?
Right. And this is why I've tried to start talking about "Gold Line" and other monikers; I think the term "Urban Ring" potentially leads us to try to do too many things at once. To me, the priority for a "Gold Line" line along the Grand Junction corridor is to build out a Kendall subnetwork; Longwood I believe needs to be tackled with a separate set of solutions, and the other employment centers are important, but definitely not as important. I think the stronger need is for a pluricentric network, rather than a circumferential service.
I think one important function of the Urban Ring should be to reduce transfers downtown and thus free up capacity there. For this to function well, 2-seat-rides vs 3-seat-rides will be significant. A Malden resident working at Kendall is much more likely to choose a Malden-Sullivan-Kendall ride via Urban Ring than Malden-DTX-Kendall. But if that Urban Ring ride becomes Malden-Sullivan-Central-Kendall or even Malden-Sullivan-Harvard-Kendall... Now distance itself matters much less, and it may not even be faster than DTX.
Yes, I think the 2-seat vs 3-seat difference is something I didn't appreciate for a long time.

There's a little bit of geometry that is relevant. Imagine you are standing on a circle that orbits downtown at a distance of 2 miles. You are commuting to another location that is also on this circle. If your destination is directly opposite you (e.g. 12 o'clock <> 6 o'clock), then the most direct route is straight across the circle -- i.e. through the core on a single radial line.

Now imagine that your destination is instead a mere half mile clockwise; in that case, a circumferential route along the circle is the most direct route (compared to going into the core and coming back out). So, the question becomes -- at what point does it become faster to go in-and-out rather than around? (Let's assume all possible services travel at equal speeds and equal frequencies, and let's assume that there is no transfer penalty -- if you need to transfer in downtown, you know there will always be a train ready for you.)

And I think under idealized circumstances, the answer is simple: if your destination is more than approx. one-third of the way around the circle, you're better off going in and back out. Why one third?

C=2πr

To swap in some approx. figures: the circumference of a circle is a little more than 2 x 3.1, or ~6 times the radius -- 6r. Going in and out of downtown in this idealized system will always be a distance of 2r. 2r is a third of 6r. So, if you are going more than a third of the way around the circle, it makes more sense to do a radial + transfer + radial journey, rather than a circumferential one.

Now, Boston distorts this ideal network in a few different ways. For one, all of our crosstown subway lines "bend", meaning, ironically, that most of them already provide a one-seat ride from end-points one-third of the circle apart:

1698681433651.png


On the other hand, the Urban "Ring" corridor is actually kinda squished: Sullivan and Longwood are both over 2 miles from downtown, while Kendall is closer to 1.5 miles, forming an oval shape. In that respect, a Grand Junction corridor really runs parallel to the Orange Line between Sullivan and Ruggles, rather than forming the arc of a circle, which probably causes it to overperform compared to what we might expect for a circumferential service.

But let's look at a journey where the Urban Ring corridor might underperform: Sullivan <> Airport. Via Orange + Blue, this is about 4.2 miles; via Chelsea it's... 4.6 miles. Uh oh. Now, it is true that Chelsea provides a one-seat ride, while Orange + Blue requires a transfer. But, 1) if you are coming from, e.g. Malden, then this will be a 2SR either way -- either Orange + Blue or Orange + Gold -- so the OSR washes out, and 2) this journey is less than a quarter of the way around the circle -- a circumferential route should absolutely smoke the in-and-out option -- so this is a significant underperformance.

(The reason, though, that it still makes sense to run service through Chelsea is because it's not really a circumferential service, but rather a radial service from Kendall, which it definitely does a good job with.)

So, going back to your point about freeing up capacity downtown by moving transfers to the Urban Ring: I absolutely agree. But the key piece to keep in mind is that the Urban Ring corridor does not do this equally for all transfers. The northern half is most effective at shifting commutes to Kendall (and to a lesser extent, Sullivan) outside of the core:
  • northern Orange: Sullivan
  • northern Green: maybe around Brickbottom?
  • southern Green: somewhere near BU and/or Longwood
  • southern Orange: Ruggles
(Although Ruggles is a little dicey -- Orange + Red is longer but Red will be fast, and an Urban Ring route might need to trudge through Longwood)

The southern half, I argue, is most effective at shifting certain Longwood commutes out of the core:
  • northern Orange: Sullivan (maybe -- this is close to an even split, depending on where in the LMA you are going)
  • northern Red: Kendall
  • southern Orange: Ruggles
  • southern Red: JFK/UMass or Andrew or Broadway (depending on where/how you do it)
Part of the allure of the "Urban Ring" concept is that it links together so many employment centers, as you note. But I argue that it does this somewhat accidentally -- really, you could draw out the Urban Ring corridor by creating sketching separate radial subnetworks for each of these employment centers, and find where those subnetworks overlap.

To me, that means that our first priority really should be thinking about what those theoretical radial subnetworks would look like, and focus on finding good solutions for them. If we end up with a circumferential service along the way, that's great, but I think it should be a secondary priority.
 
Additionally, if anyone goes by the main BPL, there's a temporary exhibit in the map room on the history of Boston transportation. The whole exhibit is great, but I saw this map "Existing and Proposed Circumferential Thoroughfares" from 1909. This is city planners a century ago trying to figure out better ring transportation on top of very radial roads. Just funny to see how long planners have been trying to cope with the regions weird layout. Much better image scan here:

View attachment 43954
Great find! But looks like this is just for roads, right?

I'd argue the Urban Ring via Grand Junction does a reasonably good job of building a Kendall network, an okay job building a Longwood network, but not really anything else. But I'm really not sure I follow the rest of the points you are making. IIRC, a "supertrain" (of 2 Type 10s coupled together) will have capacity similar to a Blue Line consist -- I don't really see that as low capacity?
This was answered recently here:
Type 10 seating hasn't been finalized yet, but a 2-car supertrain is expected to seat at much as a 3-car Type 7/8/9 consist...so roughly 132 seats for 2 cars or 66 seats for 1 car. A 6-car Blue Line set seats 210, roughly 37% more than a Type 10 deuce. A 6-car Orange Line set seats 255, a 48% increase over a Type 10 deuce. It's a fairly sizeable capacity difference between LRT and HRT.

(More replies to other comments coming at some point.)
 
But looks like this is just for roads, right?
Yes and no. I mean, yes, they are just roads. But that was 1909, so roughly all of those proposed new or widened streets would have also carried streetcars. And that was also at a time when major streetcar routes were getting turned into subways. The current Red Line tunnel in Cambridge would have started construction that year, for example. And whatever did or didn't get built back then is still kind of dictating options today for subway alignments, unless you want to deep bore the whole thing.
 
As a side-note, I didn't mention the other major problem with using the Grand Junction, which is that now there isn't a (somewhat) good rail link between North Station and South Station for things like rolling stock transfers, which means that unless NSRL happens as far as I can tell any reuse of the alignment is dead in the water.
Not true at all. The T has trackage rights on the CSX Worcester-Ayer line, and has indeed used that for equipment transfers when the Grand Junction has been out-of-service. The only prerequisites for moving the swaps permanently out to Worcester County are:
  • Build a large-enough southside maintenance facility at Readville to handle repairs and inspections that today can only be done up north (already in design)
  • Upgrade the Worcester-Ayer track to Class 3/60 MPH to cut a 5-hour trip in half (actively being done by CSX as part of their post- Pan Am acquisition renovations)
  • Have enough padding in the rolling stock roster to inoculate against north vs. south shortages, so there doesn't need to be so many swaps to balance the numbers.
  • Reduce swaps from 1-2x a day on the Grand Junction to 1-2x a week through Worcester County, so the crew hours stay relatively par. The Downeaster swaps sets with Southampton Yard 1-2x per week, so the goal would be to have "mega" Amtrak+T lash-ups doing the weekly swaps.
  • A southside base for Maintenance-of-Way equipment (like the Alewife shed). The Readville maint facility may end up accommodating that.
That's it...a very manageable bucket list that's already well underway. You can take it off the RR network immediately to convert to rapid transit with just the ^above^. Zero NSRL prereqs.

If NSRL does happen, I think reusing the route as a busway/bike path/linear park actually makes more sense than rail service, as this is an area where feeding many local buses into enhanced local stops for the MIT campus and bajillion lab buildings nearby would make more sense then a couple consolidated stops.
A frickin' rail trail on an underutilized ROW asset that links some of the biggest radial transfer destinations on the system?!? Look...I know you're well-invested in your further-out orbital, but let's not be insane. There are big things to exploit on the GJ. Do something different and Green Line-attached with it if the real Ring needs its utterly perfect integrity-of-concept. The routing isn't garbage.
 
A frickin' rail trail on an underutilized ROW asset that links some of the biggest radial transfer destinations on the system?!? Look...I know you're well-invested in your further-out orbital, but let's not be insane. There are big things to exploit on the GJ. Do something different and Green Line-attached with it if the real Ring needs its utterly perfect integrity-of-concept. The routing isn't garbage.
At no point did I suggest converting it to a rail trail, I said that for transit use a busway (As originally proposed) might make more sense than surface rail transit, and since you've got a linear path, you can probably fit some bike lanes and greenery in there as well to make it a proper urban space. Obviously if a surface light rail network north of the Charles ever gets going then conversion to light rail is on the table, but frankly I don't consider that incredibly likely, at least in comparison to South Boston.
 
At no point did I suggest converting it to a rail trail, I said that for transit use a busway (As originally proposed) might make more sense than surface rail transit, and since you've got a linear path, you can probably fit some bike lanes and greenery in there as well to make it a proper urban space. Obviously if a surface light rail network north of the Charles ever gets going then conversion to light rail is on the table, but frankly I don't consider that incredibly likely, at least in comparison to South Boston.
With all due respect, I think you're making the underlying assumption that any light rail running on the surface means a glorified streetcar with minimum grade separation (that's not too different from a busway). That doesn't have to be the case, especially for a service via Grand Junction. Grade separation at junctions is easy other than Main St.

Also FWIW, Cambridge already has initiatives for a linear park along Grand Junction with bike lanes. Some parts have already been built (at Galileo Galilei Way) while others are in active planning. This is all being done without even touching the Grand Junction ROW (and I believe without compromising the ability to double track it).
 
Great find! But looks like this is just for roads, right?


This was answered recently here:
Type 10 seating hasn't been finalized yet, but a 2-car supertrain is expected to seat at much as a 3-car Type 7/8/9 consist...so roughly 132 seats for 2 cars or 66 seats for 1 car. A 6-car Blue Line set seats 210, roughly 37% more than a Type 10 deuce. A 6-car Orange Line set seats 255, a 48% increase over a Type 10 deuce. It's a fairly sizeable capacity difference between LRT and HRT.
Thank you for laying hands on that! Okay yeah, fair enough about the capacity difference.
 
With all due respect, I think you're making the underlying assumption that any light rail running on the surface means a glorified streetcar with minimum grade separation (that's not too different from a busway). That doesn't have to be the case, especially for a service via Grand Junction. Grade separation at junctions is easy other than Main St.
You're right that it doesn't have to be the case, but for Mass Ave and Main St especially I'd be very surprised if the cost/benefit analysis works out in favor of full grade separation vs signal priority.

And again we wrap around to one of my original points about light rail being somewhat weird here, if we're building a fully grade separated alignment that connects to the OL at Sullvian, why are we not just making this an OL branch that can then run up to Malden using the existing stations, or even downtown, depending on how complicated of a spaghetti junction we end up with at the CR yard? (Which was actually the recommendation made back in the MIS study for rail along this section, for the same reasons.) We'd be paying 85-90% of the cost for half the capacity at that point. The whole advantage of light rail is that it can transition to street/median running with low floor boarding, but if we just ignore all of that and make the whole route grade separated, why are we even bothering?

Anyways, good to know that there's some progress on greenifying the Grand Junction, it's a kind of a lousy part of an already lousy area, especially towards the southern portion.
 
Last edited:
The whole advantage of light rail is that it can transition to street/median running with low floor boarding, but if we just ignore all of that and make the whole route grade separated, why are we even bothering?
Light rail allows you to run service with 95% grade separation but one single grade crossing, such as Main St. Heavy rail requires 100% grade separation.

The difference of 95% vs 100% doesn't make things much worse in service and reliability, but has serious impacts on the mode choice, especially in a world where the last 5% is really hard.

if we're building a fully grade separated alignment that connects to the OL at Sullvian, why are we not just making this an OL branch that can then run up to Malden using the existing stations, or even downtown, depending on how complicated of a spaghetti junction we end up with at the CR yard?
To be clear, I'm not opposed to converting a Grand Junction service to heavy rail in the future (if you can figure out grade separating Main St and has the $$ to do it). I wish it can happen, in fact.

But I don't think an Orange Line branch is the right choice, because it introduces reverse branching from northside OL. You will permanently cut capacity south of Community College in half, when ridership is at its highest in downtown Boston. There's no guarantee that the Malden-GJ trains will be nearly as well-utilized as Malden-downtown trains; in fact, it appears very unlikely. Despite Kendall (plus Harvard, plus Longwood) being major job destinations, they pale in comparison to downtown and Back Bay, not to mention connections to other lines (especially Blue) and mainline rail.

Running a downtown-GJ wraparound has similar issues, this time cutting capacity north of Sullivan in half.

You're right that it doesn't have to be the case, but for Mass Ave and Main St especially I'd be very surprised if the cost/benefit analysis works out in favor of full grade separation vs signal priority.
I pass by the Mass Ave/Grand Junction crossing on a near daily basis, and I can say that grade separation of whatever service is there will pass the cost-benefit analysis (assuming MIT doesn't put buildings above GJ all the way to Mass Ave).

There's a lot of traffic on Mass Ave. A lot, to the point where it seriously impacts the 1 bus's performance. And Grand Junction sits between two traffic signals that are already very close (Vassar St and Albany St). Adding a third signal will be bad for everyone - GJ rail, cars on Mass Ave, the 1 bus, and pedestrians.

I'd even say it has the highest priority for grade separation along the entire Grand Junction.
 
But I don't think an Orange Line branch is the right choice, because it introduces reverse branching from northside OL. You will permanently cut capacity south of Community College in half, when ridership is at its highest in downtown Boston. There's no guarantee that the Malden-GJ trains will be nearly as well-utilized as Malden-downtown trains; in fact, it appears very unlikely. Despite Kendall (plus Harvard, plus Longwood) being major job destinations, they pale in comparison to downtown and Back Bay, not to mention connections to other lines (especially Blue) and mainline rail.
You're right that branching takes capacity, but we're not exactly running the OL at full capacity, nor do we really plan to. If GJ-Wellington gets trains every 2.5 minutes instead of an optimistic 5 we're still well within what the line can handle, especially since this section is already triple tracked and could probably handle an extra service terminating at Wellington without significant modification, today. (Except Assembly, but that's not an incredibly difficult fix. Or you could just not fix it and some trains run express or something.)
I pass by the Mass Ave/Grand Junction crossing on a near daily basis, and I can say that grade separation of whatever service is there will pass the cost-benefit analysis (assuming MIT doesn't put buildings above GJ all the way to Mass Ave).

There's a lot of traffic on Mass Ave. A lot, to the point where it seriously impacts the 1 bus's performance. And Grand Junction sits between two traffic signals that are already very close (Vassar St and Albany St). Adding a third signal will be bad for everyone - GJ rail, cars on Mass Ave, the 1 bus, and pedestrians.

I'd even say it has the highest priority for grade separation along the entire Grand Junction.
I have no doubt that the benefit side is there, but if the cost is so disproportionate that it outweighs even a large benefit it's not happening. Anyways, I'd assume any GJ crossing on Mass Ave would be connected with the Vassar St light and/or the Albany St light, thereby minimizing the number of new waits. With some lane magic you could even do left turns simultaneously from both Vassar/Albany and GJ, although the street might need a bit of widening for that, it's not totally clear. Either way, there's plenty of ways you can make the crossing not a huge problem, making any costs for full grade-separation harder to justify.

Light rail allows you to run service with 95% grade separation but one single grade crossing, such as Main St. Heavy rail requires 100% grade separation.

The difference of 95% vs 100% doesn't make things much worse in service and reliability, but has serious impacts on the mode choice, especially in a world where the last 5% is really hard.
Yes, but mostly-but-not-100% grade-separated light rail also means you lose the flexibility of buses that are able to more freely enter and exit the alignment, which if we're building a local network around a college campus and office district, somewhere where there are going to be lots of less popular journeys rather than one or two trunk routes, this flexibility is highly desirable. It also means that private entities can operate buses of their own along the route, something that is not really possible with light rail. (At least, not without a significant investment) It's the difference between being able to start a journey from one of the dorms or campus buildings along the river, then join the busway for a quick journey to the main research/education facilities, or the difference between service that brings you right to Kendall/MIT or leaves you a 5 minute walk away.


The TL;DR
If we want an orbital line with more central, limited stops, rail is a great choice, but again, that's not the goal. The goal is a network that serves Kendall, MIT, and the surrounding area. The GJ is a great tool to help do this, but it's not the only part needed If you want to utilize the GJ as a part, but not the only element, of a local transportation plan you should (IMO) use the flexibility of buses to fill this role. If instead you're after an orbital line meant to better connect the city, use heavy rail in order to meet demand. To me light rail takes the worst of both elements, being limited to a single alignment and having to deal with grade crossings.
 
The TL;DR
If we want an orbital line with more central, limited stops, rail is a great choice, but again, that's not the goal. The goal is a network that serves Kendall, MIT, and the surrounding area. The GJ is a great tool to help do this, but it's not the only part needed If you want to utilize the GJ as a part, but not the only element, of a local transportation plan you should (IMO) use the flexibility of buses to fill this role. If instead you're after an orbital line meant to better connect the city, use heavy rail in order to meet demand. To me light rail takes the worst of both elements, being limited to a single alignment and having to deal with grade crossings.
Admittedly, the GJ is supposed to serve as a part of an orbital line and a line that serves West Station, Kendall, Sullivan, and Chelsea. It can be both, which sort of makes it a compromise but also a useful and underutilized asset since it just happens to already be there. While having drawbacks, the GJ could certainly be a shared-use ROW for light rail and buses. IMO, the best part of the grand junction is it allows you to scale up over time to increase capacity. This is what gives the alignment the best bang for the buck. It can be started out as a busway, progress up to light rail with a connection at the northern end, then the southern and soon you have a viable northwest quadrant of the Urban Ring while also serving intra-Cambridge mobility.

The problem with a heavy rail urban ring line is it’s too all or nothing. It wouldn’t be very useful if it were built in small pieces and it would be far less flexible than an LRT alignment. This sort of practicality and consideration for real-world practice is what keeps this idea in crazy transit pitches (and even bleeds into real-world planning docs), as opposed to God mode where an automated HRT deep-bore urban ring belongs. Not to knock the idea since it’s well thought out and probably a more ideal solution in a perfect world, but alas we do not live in one.
 
The problem with a heavy rail urban ring line is it’s too all or nothing. It wouldn’t be very useful if it were built in small pieces and it would be far less flexible than an LRT alignment. This sort of practicality and consideration for real-world practice is what keeps this idea in crazy transit pitches (and even bleeds into real-world planning docs), as opposed to God mode where an automated HRT deep-bore urban ring belongs. Not to knock the idea since it’s well thought out and probably a more ideal solution in a perfect world, but alas we do not live in one.
I'm not sure I totally agree. Yes, it's a lot more all or nothing than a slow conversion of a surface route, but it absolutely could be built in phases. A line from Southie to Roxbury Crossing for example would absolutely still be useful for a lot of people, and the deep bored section from ~Andrew-Roxbury Crossing is about the same length as the proposed tunnel under Longwood for the Urban Ring project, it's hardly unprecedented. No doubt the project would take a lot of time to reach full completion, but I don't think it's nearly as outrageous or unrealistic as you've said.
 
You're right that branching takes capacity, but we're not exactly running the OL at full capacity, nor do we really plan to. If GJ-Wellington gets trains every 2.5 minutes instead of an optimistic 5 we're still well within what the line can handle, especially since this section is already triple tracked and could probably handle an extra service terminating at Wellington without significant modification, today. (Except Assembly, but that's not an incredibly difficult fix. Or you could just not fix it and some trains run express or something.)
I can see the reverse branch working with capacity increases north of Sullivan Square. Or better yet, branch in both directions, so what you really have is two heavy rail lines that briefly share a ROW between Sullivan and Community College, with Malden/Downtown, Malden/Grand Junction, Chelsea/Downtown, Chelsea/Grand Junction as possible service patterns. That said, I don't see it as the best option. What do you do with it once it reaches BU? My impression is that you want to send it to Southie via Longwood, Nubian, South Bay. But one of the key ideas supporting LRT on the Grand Junction is that it is much more versatile for branching. When you reach BU, why limit the service options to a single destination, when there are multiple services that would be valuable. Grand Junction as a trunk line for three to four light rail services gives us a much better upgrade across the entire system, rather than just a single, fresh, HRT line.
 
I can see the reverse branch working with capacity increases north of Sullivan Square. Or better yet, branch in both directions, so what you really have is two heavy rail lines that briefly share a ROW between Sullivan and Community College, with Malden/Downtown, Malden/Grand Junction, Chelsea/Downtown, Chelsea/Grand Junction as possible service patterns. That said, I don't see it as the best option. What do you do with it once it reaches BU? My impression is that you want to send it to Southie via Longwood, Nubian, South Bay. But one of the key ideas supporting LRT on the Grand Junction is that it is much more versatile for branching. When you reach BU, why limit the service options to a single destination, when there are multiple services that would be valuable. Grand Junction as a trunk line for three to four light rail services gives us a much better upgrade across the entire system, rather than just a single, fresh, HRT line.
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

Screenshot 2023-10-31 at 15.35.00.png
 

Back
Top