Crazy Transit Pitches

Don't most railway ROWs have much gentler curves than existing bus routes? Such sharp curves that follow the street grid, such as the Harvard curve on the RL, or the Boylston Curve on the GL, require severe speed limits for such significant curves. Such limitations would make for longer travel times, or in some cases, frequency limitations (like that of the RL's Harvard curve, which IIRC seems to be capped at 3 min intervals in the trunk). Many ROWs tend to avoid such twisty streets and high grades, which allows quick travel to the outer terminals (Oak Grove (5.2/5.7 mi) in 17 minutes from Downtown, and West Roxbury HRT in 27 minutes (7.5/8.2 mi) on the OL), but today's built environment and rising seas would mean costly tunneling is the only way to reach today's hubs the same way the old historical ROWs did.

To tunnel under existing streets, would suggest that some bus routes/transit routes should first be straightened to run in more straighter paths to avoid the speed restrictions seen in the Harvard, Boylston, or Prudential curves in the subway system. (I do get the concern though, running street level transit from Ruggles or Lansdowne to LMA would leave LMA without direct service, so a subway following Huntington and Brookline Aves would be needed to directly serve it, but that Prudential curve severely limits the speed for the more "direct" ROW/tunnel routing to LMA, compared to OL being a lot faster to Ruggles)

The 57 routing requires a turn at Union Square in Allston, plus some more hard turns in Oak Square and Newton Corner (the turn at Oak Square is particularly sharp), following the route of the streetcar tracks. Only the turn by St. Elizabeths seems to be a more gentle one. Straightening such a sharp turn (like that of the Market St. crayon map), may end up tunneling under buildings instead of tunneling beneath the road itself.

A Mass Pike routing on the B & A mainline could allow for rapid transit metro to still have a top speed of 50 or 55 MPH (80 or 90 KMH) for much of the route, especially for the stretch just east of Newton Corner, meaning even with stops spaced ~3000 ft (1km) apart like GLX, it would still have an average speed akin that to GLX (20.5 MPH, 33 KMH w stops, but GLX has a 40 MPH/64 KMH speed limit), and get from Newton Corner to Back Bay in a little over 15 minutes, and Downtown in 20 - 25 minutes.

Many of the express buses 501/504/505, are timetabled to get from Newton Corner to Downtown in 34 minutes (the 501/503 and 502/504 were combined as such the express portion ends at Copley instead of South Station). If a rapid transit line results in curtailing of the express buses, a straightened route could speed commutes by 40% from Newton Corner. Such sharp curves and deviations along Market St. could penalize the travel times quite a bit.

It may also be possible to reconfigure and straighten the buses at Brighton Center significantly if the B & A routing is used for rapid transit (plus the same if Reservior - Brookline Village is converted to HRT). Brighton Center could get direct bus connections to 3 HRT lines via a frequent 86, plus a direct link to Central Sq. Cambridge, and a frequent 65 extension to Watertown or Boston Landing.

View attachment 49459
A few quick comments:
  • I think the curvature of tracks only plays a rather small part of travel speed, especially if the curve is also relatively gentle like those on the 57 (which can be done theoretically with TBMs). If anything, the number of stops matters a lot more for travel speed.
  • In many cases, the time savings from a more direct route gets eliminated by the extra time it takes to get to the station in the first place. (This alone does not speak against a B&A routing entirely, but it does play a part.)
 
A few quick comments:
  • I think the curvature of tracks only plays a rather small part of travel speed, especially if the curve is also relatively gentle like those on the 57 (which can be done theoretically with TBMs). If anything, the number of stops matters a lot more for travel speed.
  • In many cases, the time savings from a more direct route gets eliminated by the extra time it takes to get to the station in the first place. (This alone does not speak against a B&A routing entirely, but it does play a part.)

Comparing the travel times for the Orange Line from Roxbury Crossing to Back Bay, gives a travel time of 4.5 minutes to go 3 stops (1.5 miles, 2.5 km). The E branch of the Green Line, which runs on street level in dedicated lanes from Brigham Circle to Copley, takes 9.4 minutes from LMA to Copley, 1.6 miles (2.7 km). A similar distance, but twice as long on the Green Line E branch than the Orange Line, with 5 stops instead of 3. (Brigham Circle is 11.3 mins, but 1.9 mi, 3.1km). Even during off peak hours, it still takes 9.8 minutes to travel from Brigham Circle to Copley in late nights and early mornings with minimal traffic.

Even with a Huntington Ave Subway extension from Symphony to Brigham Circle, plus elimination of the Northeastern and Prudential stops, it still would be slower to ride the Huntington Ave. Subway from Brigham Circle to Copley, than the equivalent trip from Roxbury Crossing to Back Bay on the Orange. Removing Prudential and Northeastern stops might save 2 minutes of travel time, plus 1-2 minute off peak/3-4 minute peak periods of street running shifted underground and removal of 10 mph/16kmh at switches. The end result being 7 - 8 minutes from Brigham Circle (6 minutes from LMA) to Copley on the Huntington Line, compared with 4.6 minutes from Roxbury Crossing to Back Bay on the Orange.

I suppose the Huntington Ave. Subway would need conversion from LRT to proper HRT (assuming the curves are slightly smoothed to meet the Harvard Curve's turning radii), but even then, following the exact Green Line routing to Copley would still likely be slower from the equivalent GL stops, versus the OL stops at Roxbury Crossing/Ruggles. Roxbury Crossing to Chinatown on the OL is 8 minutes, I not sure how it would be possible, even with a HRT Green Line from Boylston to LMA, can also take 8 minutes, despite having the same # of stops after removal of the NEU/Pru stops.

----

Having more circuitous routes would still hurt legibility of the transit system and it's attractiveness (and hurt the viability of bringing HRT west of Newton Corner). In addition, bus routes can easily be reconfigured to bring riders to the train stations (hence why stop spacing on the B & A would be needed for connections with existing bus routes; to push riders from the 57 and onto rapid transit). It's also worth noting that Boston Landing station and its TOD has very little transit connectivity with other local bus routes today.
  • Increase frequency on the 86/65 buses to every 8 - 9 min peak/12 - 14 min off peak, gain frequent crosstown connections to 3 RTL lines.
  • Extend the 65 from Brighton Center to Brighton Depot/Boston Landing, more than doubling bus service from Brighton Center to Brighton Depot + bringing bus connections to Boston Landing's TOD.
  • Eliminate the Union Sq. detour on the 66 bus and stay on Harvard Ave. to bring riders directly to the Allston Depot station on the B & A.,
  • Reroute the 64 to Brighton Center, adding additional RTL connections to Central Sq. in Cambridge.
Finally, if bus service on Brighton Ave., is removed, and the B & A subsituting the 57 in Allston, it could make room for dedicated bike paths, as the 57 corridor lacks good cycling routes that take the shortest path downtown (which is today's 57 bus, which is the shortest route by cycling). The Charles River cycle path is closer to much of the B & A route, especially near Oak Sq. and Brighton, and east of Packards Corner, but leaves most of Allston Village and Union Sq. without a cycling connection.

Today, a trip from Brighton Center to Downtown at 8AM rush hour takes 47 minutes on the 501. A 2 seat ride on the 57 and the Green Line, often unreliable to due traffic congestion, bunching, transfers, and delays, takes a minimum of 40-45 minutes if not more. A RTL line following the B & A routing (with all stops) would take approx ~25 minutes from Newton Corner to downtown, and ~21 minutes from Brighton Depot. Assume 6 minute headways on the B & A, and the 86 and 65 running the aformentioned frequencies, for a combined headway of every 4 - 6 minutes, for the 4 minute bus trip from Brighton Center to Brighton Depot. This means that a full trip from Brighton Center to downtown, with a transfer at Brighton Depot to the B & A mainline, would take ~38 minutes at most, 10 minutes faster than the 501 today (20% faster), and still faster than the 57 bus + GL. If the transfer times are short, it is still possible to complete this trip in as little as a halff hour from Brighton Center (Washington St & Chestnut Hill Ave).

Any rail station that gets built in Newton Corner would likely cause the removal and the end of today's express buses. If Watertown doesn't get a branch off Newton Corner, or if a transfer is required to reach Back Bay, it would result in today's one seat ride on the 504 into one with two transfers.
 
Last edited:
I feel like we're stepping into the alignment too early here. I'm going to basically restart, assuming BL is extended to Kenmore under the Esplanade and go from there. Here's my fairly exhaustive list of possible destinations we'd possibly want to connect that are west(ish) of Kenmore:
  • BU
  • Fenway
  • West Station
  • LMA
  • Coolidge Corner
  • Boston Landing
  • Union Sq (Allston), I unilaterally renamed it to Hester Sq after Rita Hester to avoid confusion with the Cambridge Union Sq and I'll be using this name going forward.
  • Brighton Center
  • Oak Sq
  • BC
  • Newton Corner
  • Newton Centre
  • Newton Highlands
  • Brookline Village
  • Newtonville
  • West Newton
  • Auburndale
  • Watertown Sq
  • Watertown Arsenal
  • Waverly
  • Belmont Center
  • Waltham Center
  • Brandeis University
  • Riverside
Plotted on a map they look like this:
Screenshot 2024-04-11 at 22.56.08.png

I've tried to color code them based on which ones make sense to serve together, (Black can serve multiple), but it's very much not perfect. I think there are 3 clear segments of any route:

The Western Segment
This segment basically involves stringing together whatever destinations around Allston/Brookline you think are most important. You could follow the D branch to hit Fenway, LMA, and Brookline Village, or you could swap BV out for Coolidge Corner. Maybe you're feeling a more BU/West Station routing, that works too. I'd personally pick Fenway-LMA-CC or BU-West Station, but there is plenty of room for other options. For the most part these can be combined with any of the other

The Central Segment
This is where things diverge pretty wildly. I think these are the main possible routes:
  • The Watertown Route: From West Station or Coolidge Corner, hit either Boston Landing or Hester Sq before going under the Charles to serve the Arsenal and either Watertown Sq or Newton Corner
  • The 57: From West Station or Coolidge Corner, hit either Hester sq or Boston Landing, Brighton Center, and Newton Corner
  • The D Branch: From Coolidge Corner or Brookline Village (roughly) follow the D branch to BC, Newton Centre, and Newton Highlands
  • The Hybrid: From West Station hit either Hester Sq/Boston Landing, Brighton Center, BC, Newton Centre, and Newton Highlands
The Terminus
And here are some possibilities to cap the line off
  • The I-90: From Newton Corner run along I-90 to hit Newtonville, West Newton, Auburndale, and then terminate at Riverside.
  • The D Branch (Continued): From Newton Highlands run to Riverside
  • The Fitchburg Line: From Watertown Sq run to Waltham and Brandeis/Roberts
  • The Northern Hook: From Watertown Sq run to Waverley and/or Belmont Center

So what could be some final routes?
  • The full D: Pretty much just the D Branch replacement, maybe with a couple modifications like a relocated Fenway and BC diversion, ends at Riverside
  • The partial D: West Station-Hester Sq-Brighton Center-BC-D Branch ending at Riverside
  • The Northern Option: West Station-Arsenal-Watertown Sq-Belmont
  • The Hook: Follow the 57 to Watertown Sq then swing north to Belmont
  • The "Blue" in Blue Line: Follow either the 57 or the Northern route to Watertown Sq, then run to Waltham and Brandeis
  • The Newton: Follow either the 57 to Newton Corner or go Boston Landing-Arsenal-Newton Corner, then run in the median and/or elevated along I-90, diverging after Auburndale to end at Riverside.
With potentially the exception of the "Partial D" due to the weird interaction with the D branch I think any of these could be viable in a TBM world and it would really come down to ridership. I'd suspect the "Blue" would come out on top but I'm honestly not sure.
 
^^ This is part of a larger point I'm coming to appreciate: most proposals for new transit in Greater Boston are constrained to the ROWs that were first "claimed" by railroads in the 19th century. (As early as 1850 in some cases.) Beyond the obvious fact that the region has changed since then, it's also worth noting that those railroads were, in many cases, built specifically to avoid existing settlements. Most if not all of the existing settlements in 1850 are still major neighborhoods today, so are going to be chronically underserved if transit remains constrained to its historical ROWs.

Regarding Riverside's concern about transit being constrained to historical ROWs from the 1850s leaving several hubs underserved into the present day and beyond.

I went through the painstaking task of getting 10 minute walksheds for every single intersection of the street grid with all historical ROWs, regardless of if a station exists or not, historically or currently. This way I can plot the entire areas served by all historical ROWs and all areas that are underserved without existing rail corridors/ROWs that can be used for straight-ish line, high speed, direct rapid transit service.

Areas that are notably underserved by all historical railway ROWs (I've bolded some major bus routes)

In Allston-Brighton/Brookline: Oak Sq., Brighton Center, Washington Sq., Coolidge Corner. Covering parts of the B, C, 65, 66, 86, and 70.
In Cambridge/Belmont/Watertown/Arlington: Much of the area near Mt. Auburn St., eastern Belmont, parts of Cambridge St. (rts: 73/71/69/74/75/78)
In Charlestown/Somerville/Medford: Portions of the 87, 95, 134, and 101 bus routes. All of the 92 and the 93. Covers Clarendon Hill, Ten Hills, Winter Hill, almost all of Charlestown, parts of Medford.
In Malden: Surprisingly most of Malden is covered by the historical ROWs, only areas near the border and more remote areas are underserved, such as portions of the 99, 106, and 411 routes.
In Everett/Chelsea/Revere/East Boston: Much of the 109, 104, 97, 110, 111, 116, and 117 routes are underserved by existing ROWs, covering large areas of these towns/cities/neighborhoods. Very little of Everett and Revere are near existing ROWs. Chelsea and Eastie have mostly good coverage except for small portions.
In Boston (points south). Longwood Medical Area, parts of JP and West Roxbury, a large portion of South End and the bulk of Roxbury, plus portions of Rozzie, Hyde Park and Dorchester. Covers: the 39, SL4/SL5, 34, 42, 28, 21, 201/202, 15, 31, 47, 8, 9, 11, 7, and 66 buses.

EDIT: Replaced the images to also include the Harvard Branch (thanks for the tip @ Riverside's blog post) (While the Harvard Branch is mostly overwritten unlike the MedfordSq/DedhamMall-WestRox Branches, the former junction is still barely visible to this day).

1712968039538.png
1712968048574.png


EDIT: Diagram of all the railway ROWs
 
Last edited:
^ This is an excellent analysis. I've been drafting a blog post on this, so I'm gonna see if I can finish that up first and then I'll add more comments after. But, one of the tl;dr's here IMO: the underserved areas highlighted on your map overlap remarkably with the shorelines and existing settlements of pre-Industrial Revolution (and pre-aggressive landfill) Boston.
 
I like that idea in principle (and indeed, my analysis was virtually leaning towards that idea), but my concerns are:
  • Cost, obviously. I think there may be an intermediate between your two possibilities: The ridership is enough to justify a surface or elevated rapid transit extension along B&A and the Pike, which sees 5-min frequencies instead of 10-15, but not a deep bored subway.
  • The 57 alignment (even if starting after West Station) omits Boston Landing, which I feel has potential to become a major node or even destination on its own thanks to the TOD (with even greater future potential as currently discussed here).
If we're using the TBM card anyway, I suppose you can do something like this:
View attachment 49443
Personally, I'd prioritize Arsenal over Brighton on your way to Watertown Square (and then possibly to Waltham or Newton Corner). If we're doing crayon maps, let Brighton be served by one of the light rail lines.
 
I feel like we're stepping into the alignment too early here. I'm going to basically restart, assuming BL is extended to Kenmore under the Esplanade and go from there. Here's my fairly exhaustive list of possible destinations we'd possibly want to connect that are west(ish) of Kenmore:
  • BU
  • Fenway
  • West Station
  • LMA
  • Coolidge Corner
  • Boston Landing
  • Union Sq (Allston), I unilaterally renamed it to Hester Sq after Rita Hester to avoid confusion with the Cambridge Union Sq and I'll be using this name going forward.
  • Brighton Center
  • Oak Sq
  • BC
  • Newton Corner
  • Newton Centre
  • Newton Highlands
  • Brookline Village
  • Newtonville
  • West Newton
  • Auburndale
  • Watertown Sq
  • Watertown Arsenal
  • Waverly
  • Belmont Center
  • Waltham Center
  • Brandeis University
  • Riverside
Plotted on a map they look like this:
View attachment 49475
I've tried to color code them based on which ones make sense to serve together, (Black can serve multiple), but it's very much not perfect. I think there are 3 clear segments of any route:

The Western Segment
This segment basically involves stringing together whatever destinations around Allston/Brookline you think are most important. You could follow the D branch to hit Fenway, LMA, and Brookline Village, or you could swap BV out for Coolidge Corner. Maybe you're feeling a more BU/West Station routing, that works too. I'd personally pick Fenway-LMA-CC or BU-West Station, but there is plenty of room for other options. For the most part these can be combined with any of the other

The Central Segment
This is where things diverge pretty wildly. I think these are the main possible routes:
  • The Watertown Route: From West Station or Coolidge Corner, hit either Boston Landing or Hester Sq before going under the Charles to serve the Arsenal and either Watertown Sq or Newton Corner
  • The 57: From West Station or Coolidge Corner, hit either Hester sq or Boston Landing, Brighton Center, and Newton Corner
  • The D Branch: From Coolidge Corner or Brookline Village (roughly) follow the D branch to BC, Newton Centre, and Newton Highlands
  • The Hybrid: From West Station hit either Hester Sq/Boston Landing, Brighton Center, BC, Newton Centre, and Newton Highlands
The Terminus
And here are some possibilities to cap the line off
  • The I-90: From Newton Corner run along I-90 to hit Newtonville, West Newton, Auburndale, and then terminate at Riverside.
  • The D Branch (Continued): From Newton Highlands run to Riverside
  • The Fitchburg Line: From Watertown Sq run to Waltham and Brandeis/Roberts
  • The Northern Hook: From Watertown Sq run to Waverley and/or Belmont Center

So what could be some final routes?
  • The full D: Pretty much just the D Branch replacement, maybe with a couple modifications like a relocated Fenway and BC diversion, ends at Riverside
  • The partial D: West Station-Hester Sq-Brighton Center-BC-D Branch ending at Riverside
  • The Northern Option: West Station-Arsenal-Watertown Sq-Belmont
  • The Hook: Follow the 57 to Watertown Sq then swing north to Belmont
  • The "Blue" in Blue Line: Follow either the 57 or the Northern route to Watertown Sq, then run to Waltham and Brandeis
  • The Newton: Follow either the 57 to Newton Corner or go Boston Landing-Arsenal-Newton Corner, then run in the median and/or elevated along I-90, diverging after Auburndale to end at Riverside.
With potentially the exception of the "Partial D" due to the weird interaction with the D branch I think any of these could be viable in a TBM world and it would really come down to ridership. I'd suspect the "Blue" would come out on top but I'm honestly not sure.

Good start to the analysis!

Point of revision: I would argue that for your list to be "fairly exhaustive," it should include at least the following three:
  • Audubon Circle
  • Washington Square
  • Cleveland Circle
Revising your map and analysis would illuminate another routing possibility.
 
Good start to the analysis!

Point of revision: I would argue that for your list to be "fairly exhaustive," it should include at least the following three:
  • Audubon Circle
  • Washington Square
  • Cleveland Circle
Revising your map and analysis would illuminate another routing possibility.
Fair point. I don't think Audubon Circle changes too much, as I said there's plenty of room for the start of the route to just go wherever. I would probably pin Washington Sq and Cleveland Circle (Reservoir) onto the 'D branch' central segment since I don't think you can really connect them with Brighton without some pretty crazy curves.
 
Fair point. I don't think Audubon Circle changes too much, as I said there's plenty of room for the start of the route to just go wherever. I would probably pin Washington Sq and Cleveland Circle (Reservoir) onto the 'D branch' central segment since I don't think you can really connect them with Brighton without some pretty crazy curves.

No lip service to tunnel under C-Branch as one of the alternatives?
 
I went through the painstaking task of getting 10 minute walksheds for every single intersection of the street grid with all historical ROWs, regardless of if a station exists or not, historically or currently. This way I can plot the entire areas served by all historical ROWs and all areas that are underserved without existing rail corridors/ROWs that can be used for straight-ish line, high speed, direct rapid transit service.

Areas that are notably underserved by all historical railway ROWs (I've bolded some major bus routes)

In Allston-Brighton/Brookline: Oak Sq., Brighton Center, Washington Sq., Coolidge Corner. Covering parts of the B, C, 65, 66, 86, and 70.
In Cambridge/Belmont/Watertown/Arlington: Much of the area near Mt. Auburn St., eastern Belmont, parts of Cambridge St. (rts: 73/71/69/74/75/78)
In Charlestown/Somerville/Medford: Portions of the 87, 95, 134, and 101 bus routes. All of the 92 and the 93. Covers Clarendon Hill, Ten Hills, Winter Hill, almost all of Charlestown, parts of Medford.
In Malden: Surprisingly most of Malden is covered by the historical ROWs, only areas near the border and more remote areas are underserved, such as portions of the 99, 106, and 411 routes.
In Everett/Chelsea/Revere/East Boston: Much of the 109, 104, 97, 110, 111, 116, and 117 routes are underserved by existing ROWs, covering large areas of these towns/cities/neighborhoods. Very little of Everett and Revere are near existing ROWs. Chelsea and Eastie have mostly good coverage except for small portions.
In Boston (points south). Longwood Medical Area, parts of JP and West Roxbury, a large portion of South End and the bulk of Roxbury, plus portions of Rozzie, Hyde Park and Dorchester. Covers: the 39, SL4/SL5, 34, 42, 28, 21, 201/202, 15, 31, 47, 8, 9, 11, 7, and 66 buses.
What are some (potential) (very) high ridership transit corridors that would have the most difficulty in getting a high speed, rapid, potentially grade separated dedicated transitway to support them?

Some potential valuable transit connections listed here are part of BNRD, others are routes that mostly avoid historical ROWs or run perpendicular to those.

Some I can think of:
  • Radial service to downtown Everett, Chelsea, Revere, and Charlestown
  • Crosstown service serving Malden, Medford, and Davis Sq. (also possible to reroute to Arlington Center)
  • Watertown & Lower Allston radial service
  • LMA transit hub - extending to Watertown Sq. via Brighton, Ashmont via. Forest Hills/Rt. 39 or via Heath St./Jackson Sq./Columbus Ave. transitway, South Boston via South End, JFK UMass
  • Chronically slow, bunched, and overcrowded 1 and 66 buses.
  • Washington St. El replacement and Blue Hill Ave. transitway via Warren St, Nubian, and South End.
  • 15 and 47 buses.
  • Parts of the B and C branches
  • Huntington Ave. subway extension.
  • Northern Somerville (89/93 combo)
Most of these were streetcar suburbs that were entirely dependent and grew off of the streetcar, and lacked direct regional rail service. One thing I also notice is how while the Grand Junction ROW is often used for proposals for crosstown service on the north side (especially Kendall, Everett, Chelsea, and Eastie), the south side has no crosstown ROW available. As such, many possible crosstown corridors are on the south side intersecting near LMA. On the northside, Malden, Medford, Arlington, and Revere lack an "outer ring" dedicated crosstown route.

Just list possible high ridership corridors with the most difficultly in getting dedicated transitways. (EDIT: Added Harvard Branch from @ Riverside's tip in blog post)

1712968965820.png
 
Last edited:
What's interesting (And very much not coincidental) is that this is basically a map of former streetcar lines that are now busy bus routes.
They're essentially all the corridors constantly being tinkered with in Reasonable and Crazy Transit Pitches, given they don't have a nearby rail ROW that can be used and fall back on.

The only bus routes that could theoretically be absorbed outright by a railway ROW are the 87, 88, 80, 77, 32, and 36 routes (106/108 is debatable). Them being that for those particular routes, those railway ROWs are abandoned or too narrow for widening, and/or not being used for local HRT (or LRT) service currently.

Viewing it all through a lens of segregating railway ROWs/tunnels from major local bus/streetcar routes makes it much more visible on perspective. (Plus it would help in seeing what corridor can be used to knock out most of the unserved areas with fewer stones).
 
Last edited:
Been a hectic day but here is my blogpost: https://railsroadsriverside.com/the-long-lasting-impact-of-the-transit-land-grab-in-boston/

Key points:
Which brings us to our first point: the large majority of the T’s (rapid transit) route miles run on the same paths that were carved out before 1890 (many before 1870, and quite a few as early as 1855).

What’s more: many common proposals to expand the T simply reactivate ROWs that were first carved out in the 19th century (in some cases, as much as 170 years ago).
...
The wetlands surrounding Boston Neck were easier to go through than the neighborhood itself, which doomed the neighborhood to miss out on the “transit land grab” of the 19th century, which continues to govern the location of rapid transit to this day.
...
The dual examples of Cambridge and Boston Neck demonstrate that the construction of railroad ROWs has frozen in time the idiosyncratic mid-19th century divisions between “old” and “new” settlements.
...

[A]reas already-settled by the mid-19th century were bypassed by the new railroad ROWs that now serve as our primary space for transit.
I argue that the Cambridge Subway, the Washington St El, and the East Boston Tunnel provide illustrative examples.
 
I intend to share my thoughts on this great ongoing discussion over the weekend, but I want to throw the questions out first for thoughts and opinions, as these are things I'm genuinely unsure about.

1. Suppose you want to build an underground subway line immediately beneath an active mainline rail ROW. What engineering options are there?
  • With the way this question is currently phrased, I can't see cut-and-cover being an option (though I'm happy to be corrected). So is TBM the only choice?
2. What are the benefits of building a subway line that follows such a ROW, as opposed to any other alignment (e.g. under streets or even under buildings)?
  • One possible advantage I could see is the relative lack of underground utilities, if continuously running mainline rail means no opportunity to put utilities under the ROW. But how relevant is that if we're not doing C&C in the first place?
Now, the intuition behind the questions.

Much of the discussions above assumes that following a railroad ROW makes it easier for rapid transit expansions. However, for ROWs with active mainline rail service, it appears that the assumption is only true if rapid transit tracks can be laid next to mainline rail (like GLX), which requires the ROW to have sufficient width.

That's not necessarily true with every ROW. Some, such as certain parts of the B&A, don't have room for 4 tracks, and even those that do may not have room for platforms at ideal locations. You need to go either above ground or underground, but while the former is much cheaper, it may not always be feasible due to neighborhood concerns. And when it comes to the latter... My uncertainties with the two questions above are what make me wonder if it will even be a preferred option, compared to tunneling under any other alignment.

In essence, this thought exercise is a reexamination of the idea that "follow rail ROW = easy, greenfield ROW = hard".
 
Last edited:
I intend to share my thoughts on this great ongoing discussion over the weekend, but I want to throw the questions out first for thoughts and opinions, as these are things I'm genuinely unsure about.

1. Suppose you want to build an underground subway line immediately beneath an active mainline rail ROW. What engineering options are there?
  • With the way this question is currently phrased, I can't see cut-and-cover being an option (though I'm happy to be corrected). So is TBM the only choice?
You could dig under one track at a time and "lane-shift" around the temporarily single-tracked railroad. Stabilize the soil around the remaining one track while pouring a complete half of a tunnel, then shift the track onto the completed tunnel roof (properly stabilized) while doing the other track. There are definitely places around the world that have done this.
2. What are the benefits of building a subway line that follows such a ROW, as opposed to any other alignment (e.g. under streets or even under buildings)?

  • One possible advantage I could see is the relative lack of underground utilities, if continuously running mainline rail means no opportunity to put utilities under the ROW. But how relevant is that if we're not doing C&C in the first place?
It's very relevant, because it wouldn't be C&C. It would be a capped cut. C&C includes that 10-20 ft. 'sandwich' layer where all the under-road utilities go. You obviously wouldn't need that on a rail ROW that has no dug utilities (only cable conduits a few inches below ballast). Capped cuts are MUCH cheaper than C&C because they don't deal with the extra layer...you just have the bare tunnel roof supporting the new trackbed on top. The only halfway tricky spots would be where the railroad is already elevated on an overpass above a cross street (e.g. the B&A on the 2250 ft. between North Beacon and Brooks St.) where the tunnel would obviously have to dip deeper to undercut the streets below.
Now, the intuition behind the questions.


Much of the discussions above assumes that following a railroad ROW makes it easier for rapid transit expansions. However, for ROWs with active mainline rail service, it appears that the assumption is only true if rapid transit tracks can be laid next to mainline rail (like GLX), which requires the ROW to have sufficient width.

That's not necessarily true with every ROW. Some, such as certain parts of the B&A, don't have room for 4 tracks, and even those that do may not have room for platforms at ideal locations. You need to go either above ground or underground, but while the former is much cheaper, it may not always be feasible due to neighborhood concerns. And when it comes to the latter... My uncertainties with the two questions above are what make me wonder if it will even be a preferred option, compared to tunneling under any other alignment.

In essence, this thought exercise is a reexamination of the idea that "follow rail ROW = easy, greenfield ROW = hard".
The B&A also has the option of "lane-shifting" onto the right lane of the Pike during construction staging. East of Newton Corner the Pike is 8 lanes. A temp 1-lane reduction to 7 lanes is only going to marginally impact one direction of travel at one peak. The bigger issue might be the schedule constraints of single-tracking, but to treat that you just have to keep the temporary single-tracking to short distances as you move along. Hopefully by this point the B&A will have enough crossovers inside of 128 that you can stage it one interlocking at a time to minimize the OTP hit.
 
Here's a philosophical question.

Suppose you are a planner, and you're deciding between improving transit for two neighborhoods, A and B. They're similar in population, density and distance from downtown.

It's easier or cheaper to bring a significantly higher quality of transit to A. However, B is significantly more transit-dependent or disadvantaged.
Here's a rough list of different levels and qualities of transit, from lowest to highest:
  • Infrequent bus route in mixed traffic to a rapid transit transfer station
  • Frequent bus route in mixed traffic to a rapid transit transfer station
  • Frequent bus route in side-running bus lanes to a rapid transit transfer station
  • Frequent bus route in center-running bus lanes to a rapid transit transfer station
  • Streetcar in dedicated lanes to a rapid transit transfer station
  • Streetcar in dedicated lanes to the edge of "extended downtown" (e.g. Back Bay, North Station/Haymarket, Seaport, TMC)
  • Streetcar in dedicated lanes to "downtown core" (e.g. Park/DTX/State/GC, perhaps also South Station)
  • "Metro-lite" light rail to downtown core
    • "Metro-lite" is a term that I currently use to describe light rail that's largely or completely grade-separated and has long stop spacing, resembling HRT. Examples include today's Green Line trunk, most GLR proposals, and many parts of LA's and Seattle's light rail networks.
  • Heavy rail to downtown core
The list is imperfect, as a few other categories do not fall cleanly within the tier list. (e.g. frequent bus route to downtown or extended downtown; streetcar in mixed traffic; metro-lite or HRT to a transfer station or extended downtown; circumferential service that doesn't hit downtown but does hit more than one transfer points)
Common indicators of higher levels of transit dependency include:
  • Lower income, or greater low-income population
  • Lower number of cars pers household, or greater car-free population
  • Greater minority population
  • Lower levels of working from home
  • Higher share of trips by transit today
    • There's a caveat to this: transit share is often heavily correlated with existing transit service, regardless of demographics
  • Lower share of trips by car or by bike
    • Same caveat as above
  • Higher share of non-commute trips by transit today

You can only choose one neighborhood (e.g. spending significantly more on B to get similar levels of service, or getting a lower level of service for B with similar cost, compared to A).

Which one will you choose? Or, will you hold off until you have the means to do both?
  • Cost-effectiveness
  • From the perspective of the entire metro area, better service to A may allow and encourage more newcomers to move there, possibly doing more to help with the housing crisis and/or encourage car-free living styles
  • Capitalize on network effects of rapid transit more easily and quickly
  • Boost confidence for other regions that they might get similar treatments in the future
  • Equity; does not give the impression of favoring the rich
  • Reduce inequality across the whole region
  • Boost confidence for other regions that they'll not be shafted
(Note that quantity of bullet points for the arguments above is not associated with quality.)

Bonus question: How will your answer be affected by the following factors (each factor can lean in favor of either A of B)?
  • Historical transit service: If the quality of transit we're providing is merely a restoration of what used to be there, in either the recent past or the distant past
  • Access to alternative transit service
  • Commonality of such service in other cities: i.e. Whether the service proposed is "what every transit-friendly city should have"
  • Duration that the transit improvement has been considered: If the expansion has been discussed for decades, but kept getting pushed back
  • An emphasis of direct downtown access: Does the answer change substantially if one of them gets a one-seat ride to downtown and the other does not?
 
Last edited:
I'm going to be annoying and say that there isn't enough information to make a decision.
It's easier or cheaper to bring a significantly higher quality of transit to A. However, B is significantly more transit-dependent or disadvantaged.
That would suggest that a project in area B would have more ridership and thus a better return on investment. If so, how large of a difference are we talking? If that is the case then the cost difference needs to be examined more closely. Are we dealing with a minor difference like LRT on a shorter vs longer corridor, for example, or a difference of at least an order of magnitude, say surface LRT vs deep level subway? If it's the latter then the ridership difference better be large.
An emphasis of direct downtown access: Does the answer change substantially if one of them gets a one-seat ride to downtown and the other does not?
Depends on whether people in A or B want to go to downtown or not, or more realistically what proportion of trips are to downtown vs other major centers.
Commonality of such service in other cities: i.e. Whether the service proposed is "what every transit-friendly city should have"
Look to other cities to see what's feasible, what can work, and what problems a solution might bring. Don't blindly follow the bandwagon lest you end up with a monorail.
Duration that the transit improvement has been considered: If the expansion has been discussed for decades, but kept getting pushed back
Something that has been discussed for longer may have higher levels of public support and a relatively solid plan already and thus be easier to get built, whereas a newer project would have a longer planning phase and consultations with involved stakeholders.
Access to alternative transit service
Is the alternative transit service significantly lower quality than the proposed service, or does it fill a different role?
Historical transit service: If the quality of transit we're providing is merely a restoration of what used to be there, in either the recent past or the distant past
If there's existing infrastructure that can be reused or reactivated in one location, whereas all-new infrastructure would be needed in the other location, that likely pushes the cost difference more towards an order of magnitude and thus more ridership would be needed to recoup costs. The other way a restoration could tilt the scales is if a former service has impacted the urban environment, say a former streetcar suburb with high densities along a linear corridor, or a series of concentrated centers located around a former mainline railway.
 
All else equal, I’d choose A.

Transit expansion happens on a timescale much larger than human movement.

Let’s look at our most recent rapid transit expansion in the region: the Green Line Extension.

What percentage of residents who lived there when it broke ground (2012) were still there when it opened (2022)? What percentage of residents who lived there when planning began (2006) were still there when it opened? What percentage of residents who lived there when the state committed to it (1990) were still there when it opened?

I don’t know the exact number to these questions, but I’d bet low (maybe 20%), very low (maybe 10%), and extremely low (maybe 1%). A thing that has remained constant though is that Somerville is the most densely populated municipality in New England. Generations of people have come and gone and the demographics (especially socioeconomic) has changed, but the transit-oriented nature of the built environment has not.

Therefore, in my estimation, when planning for large scale (multi-year) transit expansion, it makes sense to serve the most theoretical people for your buck rather than focusing on the exact composition that exists at this moment in time. Most of the current residents won’t see the actual expansion completed while residing there anyways.

And, to further my perspective, that is only focusing on the opening of the branch! A large-scale transit expansion will hopefully be operational for generations. Nobody who used the East Boston Tunnel in 1904 is still alive today, yet generations of people continue to use it and East Boston continues to be (and will continue to be) a transit-oriented, urban neighborhood.

There are a lot of unknowns in your hypothetical, but controlling for everything not mentioned, I firmly believe you prioritize building transit where its cost-effective and there is a high density of humans (living, working, etc), first and foremost.
 

Back
Top