Crazy Transit Pitches

[TheRatmeister]

Isn't that the exact same thing you said last time?

Anyway, in my unlimited-budget (or in other words, purely merit-based) crayon world, I actually prefer sending radial lines to both Everett and Chelsea, and there are two routes downtown that nicely pair with them: one via the Red X (aka the Financial District line) and the other via the route above (aka the Kendall-Copley line, which IMO is a way better alternative than the commonly suggested Mass Ave subway). In such a world, it makes more sense to pair Red X with Chelsea and Kendall-Copley with Everett instead, to reduce the detour from Chelsea to Kendall.

Quick and unpolished drawing of the proposal:

View attachment 50515

Remarks:

• A lot of decisions can be made about the Chelsea line's routing further north (surprisingly, Route 1 may have slightly higher density than Revere Broadway), which direction it turns to the north (west to Northgate/Linden or east to Wonderland), etc.
• The Lilac Line's stop spacing may be too close, especially near Albany St. I included that station both to offer a transfer to my "Nubian via I-93" El and to touch the corner of possible future developments to the north of Widett Circle, but I admit it doesn't seem to do the job well given the drawback of stop spacing.
• The Teal Line can skip Haymarket as both GL and OL stop there (and I'm not sure if the area has enough room for a third pair of platforms).
• I've also considered the idea of having the Red X (the Teal Line) take a detour to Seaport, but it's not explicitly shown here.
• I forgot to include Assembly station on the Lilac Line, even though I intended to.

If we're just crayoning a new subway route, what about some kind of blend between the two to make a U-shape, something like this?

S
Stops at:

• Medford Sq
• George St
• Harvard St
• Winter Hill
• Gilman Sq
• Union Sq
• Inman Sq
• The Port
• Mass Ave/GJ
• Hynes
• Back Bay
• Holy Cross
• Ink Block
• South Station
• Post Office Sq
• Aquarium
• North End
• Constitution
• Bunker Hill
• Sullivan Sq
• Encore
• Mystic Landing
• Carrington
• Glendale
• Woodlawn
• Linden

I'd really consider all of these more of god-mode pitches though, there's not really any shortcuts you can take to make any of these routes easier/cheaper besides a short elevated between Sullivan and Sweetsner Circle and a shallower bit under Broadway in Southie. Everything else needs to be bored anywhere from somewhat deep to very deep with large underground stations able to accommodate the ~400-450ft long trains you'd really want on a highly radial route. You'd be looking at a similar amount of bored tunnel for this line as for the BL to Brandeis extension and UR between Wonderland and City Point combined.

 

[Teban54]

If we're just crayoning a new subway route, what about some kind of blend between the two to make a U-shape, something like this?View attachment 50517S

I'm not too hot about a line to Medford Square. The 101's route closely parallels GLX, and density drops off a bit north of your Winter Hill station. Medford as a whole is pretty auto-centric and less dense, and the parts that aren't (along the Salem St corridor) would require a bus connection anyway. These regions may still warrant better transit in the form of improvements on the 101 corridor; or, if Medford Square in particular is the target, a surface or elevated route along Route 16 (or even I-93) may be possible, which feels appropriate for the population served. At the very least, the 101 corridor falls way behind any corridors involving Everett and Chelsea as candidates for another radial route, and I can't imagine it being chosen unless we have 3 radial routes to spare, if not 4.

If the intention is to give Gilman Square and Union Square Somerville a true HRT line to downtown, I actually think pursuing GLX improvements will be a much, much better use of resources. This can even include improving terminal capacity such that 2 bullets can terminate at Medford/Tufts instead of one (E).

(Also, the intention of my Kendall-Copley line is a combination of better connections to Kendall and better connections from Cambridge to Back Bay. It appears that your proposal will need an infill on the Red Line to do either of these things.)

 

[TheRatmeister]

I'm not too hot about a line to Medford Square. The 101's route closely parallels GLX, and density drops off a bit north of your Winter Hill station. Medford as a whole is pretty auto-centric and less dense, and the parts that aren't (along the Salem St corridor) would require a bus connection anyway. These regions may still warrant better transit in the form of improvements on the 101 corridor; or, if Medford Square in particular is the target, a surface or elevated route along Route 16 (or even I-93) may be possible, which feels appropriate for the population served. At the very least, the 101 corridor falls way behind any corridors involving Everett and Chelsea as candidates for another radial route, and I can't imagine it being chosen unless we have 3 radial routes to spare, if not 4.

If the intention is to give Gilman Square and Union Square Somerville a true HRT line to downtown, I actually think pursuing GLX improvements will be a much, much better use of resources. This can even include improving terminal capacity such that 2 bullets can terminate at Medford/Tufts instead of one (E).

(Also, the intention of my Kendall-Copley line is a combination of better connections to Kendall and better connections from Cambridge to Back Bay. It appears that your proposal will need an infill on the Red Line to do either of these things.)

Good points all around. I think the bigger aim of the western half is to make a north-south cross-Camberville route which is something that's lacking right now (Inman Sq to Killian Court right now involves either two buses or 1 bus and 15 minutes of walking) and even with most UR proposals. Adding a stop at Technology Square (Or Friedman, whatever you call it) would fix the missing RL connection, that's just an accidental omission on my part.) An alternative could be instead of going to Medford Sq go from Gilman to Assembly and then close the loop (Well, London Circle Line style so not really closed but you get the point) at Sullivan.

 

[Teban54]

Good points all around. I think the bigger aim of the western half is to make a north-south cross-Camberville route which is something that's lacking right now (Inman Sq to Killian Court right now involves either two buses or 1 bus and 15 minutes of walking) and even with most UR proposals. Adding a stop at Technology Square (Or Friedman, whatever you call it) would fix the missing RL connection, that's just an accidental omission on my part.) An alternative could be instead of going to Medford Sq go from Gilman to Assembly and then close the loop (Well, London Circle Line style so not really closed but you get the point) at Sullivan.

A north-south connection through Somerville (and connecting Cambridge to Somerville) is indeed intriguing and necessary, but I have my reservations that a rail line, especially a grade-separated one, is the best option. I imagine such a connection will serve a lot of local demands, so a bus route seems much more realistic.

Idk, something like this?

Or this:

 

[Riverside]

I don't recall how recently this has been proposed/crayon-ed, but just based on the high-density sections of South Boston and the South End, there's a crazy corridor for a light rail line along Broadway from City Point/Broadway @ P St, to Back Bay, making stops at L/Broadway, Dorchester St/Broadway, D/Broadway, Broadway Station, Ink Block/ E Berkeley + Washington, and Back Bay.

This is just a fantastic line for a light rail subway but I think the engineering is just so far-fetched it couldn't happen. It would more or less be an urban ring half measure to replace the 9 bus. I'd have to look at what the 9 ridership is, and see if it's even a high demand area for transit, if replaced.

If Boston were in Europe, I suspect this route would indeed be a streetcar already (or still, uninterrupted from the early 20th century).

Anyway, in my unlimited-budget (or in other words, purely merit-based) crayon world, I actually prefer sending radial lines to both Everett and Chelsea, and there are two routes downtown that nicely pair with them: one via the Red X (aka the Financial District line) and the other via the route above (aka the Kendall-Copley line, which IMO is a way better alternative than the commonly suggested Mass Ave subway). In such a world, it makes more sense to pair Red X with Chelsea and Kendall-Copley with Everett instead, to reduce the detour from Chelsea to Kendall.

Quick and unpolished drawing of the proposal:

Remarks:

• A lot of decisions can be made about the Chelsea line's routing further north (surprisingly, Route 1 may have slightly higher density than Revere Broadway), which direction it turns to the north (west to Northgate/Linden or east to Wonderland), etc.
• The Lilac Line's stop spacing may be too close, especially near Albany St. I included that station both to offer a transfer to my "Nubian via I-93" El and to touch the corner of possible future developments to the north of Widett Circle, but I admit it doesn't seem to do the job well given the drawback of stop spacing.
• The Teal Line can skip Haymarket as both GL and OL stop there (and I'm not sure if the area has enough room for a third pair of platforms).
• I've also considered the idea of having the Red X (the Teal Line) take a detour to Seaport, but it's not explicitly shown here.
• I forgot to include Assembly station on the Lilac Line, even though I intended to.

Cool map! Yeah coincidentally I've been thinking more about the value of/need for a Back Bay <> Kendall spine. Of the 5 "downtowns" (Downtown, Back Bay, Longwood, Seaport, Kendall), Back Bay <> Kendall is the only one without a straightforward OSR crayon design.

Crayoning rapid transit down Everett's Broadway runs into a challenge in terms of where to go south of Sullivan. We've seen lots of suggestions to run it to Kendall and then potentially Longwood, which has its upsides, but I think missing Downtown and Back Bay stings pretty bad. But providing a OSR to Back Bay feels like a reasonable compromise.

If we're just crayoning a new subway route, what about some kind of blend between the two to make a U-shape, something like this?

From a philosophical perspective, U-shapes always seem at least a little inefficient to me. With an idealized straight line across a region, some percentage of riders will ride straight through as an OSR, without transfers. (For example, Broadway <> Harvard.) The more you curve the line, the greater the percentage of riders who will instead take a more direct bus route, and the lower the percentage of riders who will ride the train through the core... and thereby the greater number of transfers that need to happen on the core platforms.

With a U-shaped route, no one will ride it end to end, which costs you the effiency bump of that X percentage of riders who might otherwise ride straight through, and you increase transfers that need to occur in the core. I'm not saying it's never worth doing! Montreal and WMATA both have U-shaped lines that do alright, though both of them have "legs" that are further distant -- 3 miles or more (ish), while the route you're suggesting here is less than 2 miles apart for much of it.

 

[TheRatmeister]

A north-south connection through Somerville (and connecting Cambridge to Somerville) is indeed intriguing and necessary, but I have my reservations that a rail line, especially a grade-separated one, is the best option. I imagine such a connection will serve a lot of local demands, so a bus route seems much more realistic.

Idk, something like this?
View attachment 50531

Or this:
View attachment 50532

I think you're probably right that the need for such a route doesn't extend past Union Sq, but I think until there is demand between MIT, Inman Sq, and Union Sq. I've been thinking about how to rope in Chelsea, so what about a ribbon-crossover-ish style line like this?

With a cross-platform transfer at Sullivan this gives both branches an easy ride to Kendall, Back Bay, and Downtown. Or if you want to rope in Seaport instead of going Post Office Sq - South Station - Ink Block - Union Park - Back Bay, go Post Office Sq - Courthouse - Berkeley/SOWA - Back Bay.

 

[Teban54]

Crayoning rapid transit down Everett's Broadway runs into a challenge in terms of where to go south of Sullivan. We've seen lots of suggestions to run it to Kendall and then potentially Longwood, which has its upsides, but I think missing Downtown and Back Bay stings pretty bad. But providing a OSR to Back Bay feels like a reasonable compromise.

Yes, and that's an issue I faced when creating this very map too, and partly why so many crayon proposals just end up creating a branch of the Orange Line at Sullivan instead. I think the "best" radial route for Everett alone would be via the Red X to North Station, South Station and beyond, in a world where Everett is the only concern. But in my map above, I already used that routing for Chelsea.

Thus, I settled down on sending the Everett line to Kendall and Back Bay -- partly for the reason you mentioned (that Back Bay is still considered part of downtown), and partly because, in the real-world SLX Alternative Analysis, residents showed greater support for SL6 to Kendall than Haymarket. I had previously opined that I don't think this implies Everett residents see a circumferential route as a better alternative than a radial one in general. But this at least shows some support for Everett-Kendall, or that they would still be happy with such a route even if a radial one is better. I'm much less confident about Chelsea residents believing this, partly due to their existing radial route being more direct via the Tobin.

A nuance that's worth highlighting: Chelsea's traditional ROWs for radial vs. circumferential transit are fundamentally different, the former via Tobin and the latter via Grand Junction. In contrast, Everett uses largely the same ROWs (Lower Broadway) to get anywhere, radial or circumferential. This also means Everett residents are more likely to accept a "forced" transfer at Sullivan than Chelsea.

Another idea is to have one line serve both Chelsea and Everett:

This is probably the cheapest alternative, and the most effective in a world where you don't want to build the Kendall-Copley subway (or anything paralleling Mass Ave). The main drawback is that Everett residents go through a detour (4.1 miles, compared to 3.3 miles via Sullivan), although in my map above, they still have the same number of stops.

From a philosophical perspective, U-shapes always seem at least a little inefficient to me. With an idealized straight line across a region, some percentage of riders will ride straight through as an OSR, without transfers. (For example, Broadway <> Harvard.) The more you curve the line, the greater the percentage of riders who will instead take a more direct bus route, and the lower the percentage of riders who will ride the train through the core... and thereby the greater number of transfers that need to happen on the core platforms.

With a U-shaped route, no one will ride it end to end, which costs you the effiency bump of that X percentage of riders who might otherwise ride straight through, and you increase transfers that need to occur in the core. I'm not saying it's never worth doing! Montreal and WMATA both have U-shaped lines that do alright, though both of them have "legs" that are further distant -- 3 miles or more (ish), while the route you're suggesting here is less than 2 miles apart for much of it.

Fully agreed. In addition, U-shaped lines are also generally more confusing: for example, the two terminals will need to be determined as "west" or "east", when in fact both "branches" are actually heading north.

Spoiler: One possible exception

The one exception I can see is if a chain of destinations forms a U-shape naturally. The closest example in Boston is the Green Line: now with GLX, it's already a U-shape, but doing so allows Somerville residents a one-seat ride to all of downtown, Copley and LMA. (It would also have been even more useful if southside GL riders have demand for jobs at Lechmere.)

However, I don't think that applies well to @TheRatmeister's initial proposal, due to the large number of criss-crossing lines inbetween the U (or so we hope). Everett residents may want MIT, but even without the Urban Ring or even the 101 bus, transferring to the Red Line at South Station (itself already a winding route) is still a better choice than continuing on this train that goes to South End, for god's sake. Likewise, Camberville residents going to Financial District, North End etc have the Red and Green lines. Compared to these, today's Green Line "works" largely due to lack of good north-south alternatives to Copley and LMA.

I think you're probably right that the need for such a route doesn't extend past Union Sq, but I think until there is demand between MIT, Inman Sq, and Union Sq. I've been thinking about how to rope in Chelsea, so what about a ribbon-crossover-ish style line like this?
View attachment 50535
With a cross-platform transfer at Sullivan this gives both branches an easy ride to Kendall, Back Bay, and Downtown. Or if you want to rope in Seaport instead of going Post Office Sq - South Station - Ink Block - Union Park - Back Bay, go Post Office Sq - Courthouse - Berkeley/SOWA - Back Bay.

That's an even more innovative idea for sure, but I see two issues.

Firstly, these "knot-shaped lines" (for lack of a better word) are even more confusing than U-shaped lines. Not that they don't exist - Singapore's "Blue Line" does exactly this - but it's natural to question its ease of use in practice. Just imagine seas of confused passengers at Sullivan.

Secondly, making Chelsea-OL transfer at Sullivan is a no-brainer. Let's not forget that they have the 111 bus today, which offers a relatively direct route all the way to Haymarket. Moving the transfer point to Community College already adds about 0.75 miles of detour per trip on rail, but that may still be acceptable. But transferring at Sullivan for OL downtown makes the rail trip 2.5 miles longer than the 111, and eliminates any advantage in speed. (And as I mentioned above, if the goal is to bring circumferential transit to Chelsea, other options are way cheaper.)

I think the gist of this and earlier discussion is: If one neighborhood needs to have a forced transfer at Sullivan, it should definitely be Everett, not Chelsea.

 

[Teban54]

I'd really consider all of these more of god-mode pitches though, there's not really any shortcuts you can take to make any of these routes easier/cheaper besides a short elevated between Sullivan and Sweetsner Circle and a shallower bit under Broadway in Southie. Everything else needs to be bored anywhere from somewhat deep to very deep with large underground stations able to accommodate the ~400-450ft long trains you'd really want on a highly radial route. You'd be looking at a similar amount of bored tunnel for this line as for the BL to Brandeis extension and UR between Wonderland and City Point combined.

I came back to this to run some numbers. The numbers alone are actually slightly in favor of the "two radial routes to Everett and Chelsea" combo over "BLX-Waltham + your UR" combo, with 7.4 TBM miles for the former and 9.15 TBM-equivalent miles for the latter.

Spoiler: Comparison of TBM lengths

This is comparing the costs of the following two combinations:

1. A Chelsea-NS-SS-Red X line, and an Everett-Kendall-Copley-South End-City Point line
2. A Kenmore-Longwood (D)-Coolidge Corner-Brighton-Watertown-Brandeis BLX, and a Wonderland-Chelsea-Community College-Harvard-BU-Fenway-Nubian-City Point line

Segments that are in both proposals are excluded from calculations: Tobin - Chelsea - Revere, and anything in South Boston.

For Combination 1, segments that likely require TBM are:

• Everett Upper Broadway: 1.6 miles (from Woodlawn NW corner to Sweetser Circle)
• CRMF to Broadway RL: 3.8 miles
• Congress St subway (Fort Point Channel to the west of Tobin Bridge): 2.0 miles

With a total of 7.4 TBM miles.

For Combination 2, segments that likely require TBM are:

• Cambridge St to Harvard: 1.7 miles
  • Starting from the Grand Junction intersection, because I can't tell your preferred routing just by looking at your map. If you want to not use Grand Junction and dig your own tunnels, this goes up to 2.2 miles.
• Harvard area to river crossing: 0.5 miles
  • F-Line's initial suggestion of a Harvard branch assumes reusing the old RL Harvard tunnel. That's inapplicable if you want the line to continue north beyond Harvard: you'll probably need a new tunnel, because you'll need to go deeper than the new, bi-level RL Harvard station.
• BLX from Longwood (D) to Arsenal St, via Brighton: 3.3 miles

With a total of 5.5 TBM miles. Additionally, the following tunnel segments can plausibly be done with C&C with minimal neighborhood impacts, though TBM may be needed due to other political or practical concerns:

• Comm Ave subway and curving onto Fenway: 0.6-1.2 miles depending on alignment (your map probably uses the 0.6 version)
• Subway under Fenway (the road), from Fenway (the station) to Ruggles: 1.0 mile
  • I can't tell if you intended to use Fenway the road or Longwood Ave, especially if the interchange with Huntington is at MFA. If using Longwood Ave, cost is likely higher.
  • (Though in fairness, I had considered building an El over this corridor before, so...)
• Ruggles to Nubian: 0.6 miles
  • Using my earlier proposal that may be C&C-able
• Nubian to Melnea Cass: 0.5 miles
  • This assumes an El over Melnea Cass. If not, cost is higher.
• BLX Riverbank from Charles/MGH to the D's Fenway portal: 2.0 miles
• Arsenal St where Watertown Branch is encroached, until just east of Watertown Sq: 1.1 miles
• Pleasant St from Howard St (where encroachment begins again) to Seyon St (where Watertown Branch's traces begin again): 1.1 miles
• Elm St and Charles River crossing near Waltham, where the rail ROW is encroached again: 0.4 miles

With a total of 7.3 miles. If we assume the cost ratio of C&C to TBM is 1:2, this means the entire Combination 2 is worth "9.15 TBM-equivalent miles".

This analysis also does not consider any segments on the surface, elevated or in trenches. But Combination 2 has more of those too, most of which due to the Harvard-West Station trench, Watertown Branch reactivation (if you can even get back all the property rights and remove the surface-level parking lots, that is), and Nubian-Andrew El. The only notable such component in Combination 1 is from Sweetser Circle to CRMF.

The key insight is that C&C is cheaper, but it's not free.

Of course, specifics of individual segments may result in smaller adjustments up and down. But this illustrates that the cost of your "BLX+UR" combo may be bigger than it first seems.

 

[Riverside]

Another idea is to have one line serve both Chelsea and Everett:

This is probably the cheapest alternative, and the most effective in a world where you don't want to build the Kendall-Copley subway (or anything paralleling Mass Ave). The main drawback is that Everett residents go through a detour (4.1 miles, compared to 3.3 miles via Sullivan), although in my map above, they still have the same number of stops.

This is the route that I've considered the most often. North of Everett Sq is, I think, closer to God Mode, but I do actually think everything south of it is slightly more realistic/achievable.

Yes, the detour for Everett is less ideal, but I think it's actually not any worse than Porter <> Downtown via Union Sq vs Porter <> Downtown via Harvard.

Also, a slight realignment would (IMO) bring more people into the walkshed (and in the case of Everett, potentially provide more riders with a OSR to Downtown in walking distance): travel east-west via Revere Beach Parkway (probably an el) and then north-south via Route 1, with an additional Chelsea station closer to Bellingham Square or Chelsea Square (i.e. where Chelsea residents live in much greater density).

A Red-X service to Chelsea + Everett Sq seems vaguely reasonable, but I admit that extensions beyond toward, for example, Glendale or Revere seem like a taller order.

 

[Teban54]

Yes, the detour for Everett is less ideal, but I think it's actually not any worse than Porter <> Downtown via Union Sq vs Porter <> Downtown via Harvard.

Actually, no.

Spoiler: Porter's distances

I checked the two routes for the sake of completeness. Porter to Park St via the Red Line is about 4.33 miles, whereas Porter to Park St via GLX is 4.26 miles. Given that I'm just casually mapping on Google Maps, the difference is well within the margin of error due to curves and methodology. Another factor to consider is that the Red Line is extremely fast, owing to its wide stop spacing, especially when considering closeness of the GL stops within downtown.

Part of the reason is that RL vs. GL is not exactly "U shape vs. straight line", but actually opposite sides of a parallelogram. The Green Line makes a turn within downtown at North Station (not easily noticeable), whereas the Red Line does so at Harvard (much more easily noticeable):

Also, a slight realignment would (IMO) bring more people into the walkshed (and in the case of Everett, potentially provide more riders with a OSR to Downtown in walking distance): travel east-west via Revere Beach Parkway (probably an el) and then north-south via Route 1, with an additional Chelsea station closer to Bellingham Square or Chelsea Square (i.e. where Chelsea residents live in much greater density).

I had considered that, but a few reasons why I didn't:

• It appears that turning from Route 1 to Revere Beach Pkwy requires going through a high school and its stadium. I have no idea if it's advised to build an El above a stadium, but if not, this requires a short capped-cut tunnel if not outright eliminates the feasibility of this routing:

• If you want to place a Bellingham Square station to the south of the big turn near Chelsea City Hall, the station's site is not affected by alignment further north, via either the Eastern Route or Revere Beach Pkwy. Both alignments make this turn regardless.
  • Also, this "Bellingham Square" station would be only 1300-1900 ft from a plausible "Chelsea Sq" station, which is too close IMO, especially if Everett residents need to go through this section. You can move the "Bellingham Sq" station north, but it can't be on the curve, and being north of the curve would no longer be convenient for the Chelsea City Hall area.
• Someone mentioned this last time: It may not look that hot now, but Chelsea commuter rail station has huge TOD potential. That huge block that currently hosts Market Basket and its giant parking lot can easily be replaced with denser buildings in the future, not to mention other blocks in the area. Since Chelsea seems like a pretty YIMBY city, it's reasonable to suspect that by the time any of these crayon lines get built, such TOD would be ready to take place if not already.
• Last but not least, Chelsea commuter rail station is already built. While we have considered relocating the CR station back to the Bellingham Sq area before, I'm still not sold that it's worth the hassle and/or likely in practice. After all, the Market Basket is already a popular destination, and Chelsea station has higher ridership on SL3 (~800 per weekday as of Fall 22) than Bellingham Square (~520/wkdy) or Box District (~430/wkdy).
  • Yes, BS + BD combined has higher ridership than Chelsea (Market Basket), but just barely. More importantly though, I think this highlights Chelsea station is doing surprisingly well given the (lack of) density in its surrounding area.

A Red-X service to Chelsea + Everett Sq seems vaguely reasonable, but I admit that extensions beyond toward, for example, Glendale or Revere seem like a taller order.

I'm actually not sure if terminating the line at the "true" Everett Sq would be much easier than at Glendale or north. Placing a station directly at Everett Sq will require a subway (unless you can convince the residents who rejected an El last century to accept it this time), so whatever economic and social impacts result from a Glendale extension would still manifest from an Everett Sq station, just maybe smaller in magnitude.

This is made worse if you use Revere Beach Pkwy, as the turn to Everett Sq gets sharp enough that you can't avoid going under houses, making TBM a requirement. In contrast, using the Eastern Route at least allows the theoretical possibility of a fully cut-and-covered route with a turn like this:

(Not drawn to scale)

Another possibility is to continue onto the Saugus Branch to Malden and beyond. While the thoughts of reactivating that ROW for transit is intriguing (and has been discussed several times before), that kind of defeats the purpose of having an Everett Sq station in the first place; besides, the route seems too roundabout coming from Chelsea to Malden.

 

[TheRatmeister]

Firstly, these "knot-shaped lines" (for lack of a better word) are even more confusing than U-shaped lines. Not that they don't exist - Singapore's "Blue Line" does exactly this - but it's natural to question its ease of use in practice. Just imagine seas of confused passengers at Sullivan.

I'm not convinced this is an issue. Proper signage and announcements go a long way, and regardless frequent users of the line will get used to it. I would argue that how the GL branches have worked in the past terminating all over downtown is just as convoluted but I wouldn't consider that a major problem.

Secondly, making Chelsea-OL transfer at Sullivan is a no-brainer. Let's not forget that they have the 111 bus today, which offers a relatively direct route all the way to Haymarket. Moving the transfer point to Community College already adds about 0.75 miles of detour per trip on rail, but that may still be acceptable. But transferring at Sullivan for OL downtown makes the rail trip 2.5 miles longer than the 111, and eliminates any advantage in speed. (And as I mentioned above, if the goal is to bring circumferential transit to Chelsea, other options are way cheaper.)

A .75 mile detour is around 2.5 minutes, I'd consider that a non-issue. But you're right that moving the transfer point between the branches to CC makes more sense. That's what happens when you spend a grand total of 5 minutes smashing something together quickly lol.

 

[Riverside]

I'm actually not sure if terminating the line at the "true" Everett Sq would be much easier than at Glendale or north. Placing a station directly at Everett Sq will require a subway (unless you can convince the residents who rejected an El last century to accept it this time), so whatever economic and social impacts result from a Glendale extension would still manifest from an Everett Sq station, just maybe smaller in magnitude.

This is made worse if you use Revere Beach Pkwy, as the turn to Everett Sq gets sharp enough that you can't avoid going under houses, making TBM a requirement. In contrast, using the Eastern Route at least allows the theoretical possibility of a fully cut-and-covered route with a turn like this:

A small thing I can reply to quickly here, but to clarify, my point about terminating at Everett Square wasn't considering feasibility, but really just thinking about density/distance from the core etc etc. (This was also a "gut feel" comment that may not hold up under scrutiny.)

 

[Riverside]

(This post was going to have maps, but I figure everyone can use their imagination well enough.)

The year is 2099.

The General Manager of the MBTA is now a Cabinet-level position, reporting directly to the President of the United States, and you have just received an urgent page to report to the Oval Office. You enter and she immediately gets down to business. "The Green Line must be converted to Blue Line-style heavy rail immediately. This is a matter of national security."

"National security, Madam President?" you repeat in disbelief.

She nods. "National security." The gears in your head start turning.

"We should be able to convert the subways without issue, and most of the branches too... it will mean building a viaduct through downtown Needham, or just digging it up for a subway or trench."

"No", she interrupts. "The villages of Needham must remain substantively unaltered with no prolonged disruption. This is also a measure of national security."

"...how is that a matter of national security, Madam President?"

"That is classified on a need-to-know basis."

"...and I don't 'need to know'?" you say skeptically.

"You do not."

You puff out your cheeks and exhale. "Well, if that's a priority to leave Needham untouched, the light rail tracks can just remain in situ, we'll power down the wires and halt service --"

"No," the President interrupts again. "Needham must also remain served by rail transit, with minimized degradation of service."

"A matter of national security, is it?" you ask.

She nods. "Exactly."

~~~

In the past, I've argued against the conversion of the D Branch to heavy rail due to its impact on a potential branch to Needham. The D Branch is largely grade-separated (a prerequisite for modern HRT), but the Needham Line travels at grade through downtown Needham. As alluded to above, grade separating the downtown portion would require a costly/disruptive viaduct, tunnel, or trench.

But recent talk of a "Needham Trolley", pinging between a Green Line station at Needham Junction and an Orange Line station near Millennium Park (h/t to @TheRatmeister, though I think there are others that aren't coming to mind), made me look again, and notice something that should've been obvious to me years ago:

Needham has four train stations. Three of them can be served via (mostly) grade-separated ROWs. Screenshot from OpenRailway Map:

(There would be an additional crossing, not marked here, at Oak Street in Newton Upper Falls.)

These crossing fall into three categories:

• A pedestrian grade crossing just west of Hersey
  • Could be eliminated as part of a rebuild of Hersey station
• Street crossings at Oak, Gould, and Webster Streets
  • Significantly more disruptive, but this corridor will require non-trivial work anyway to convert it to a rapid transit ROW. With careful planning, potentially the impact could be reduced
  • Gould St might also get a station, which again offers options to fold a crossing into the station
• Grade crossings in downtown Needham
  • Much harder
  • Numerous residential abutters
  • Short distances between crossings reduces running length for grades, making longer stretches of grade separation -- potentially over a mile, from Heights to Needham's Oak Street

That third category is the rub. But... Hersey, Junction, and Heights can all be reached without touching the grade crossings in downtown: Heights via an extension from Newton Highlands, and the others via an Orange Line extension. It's only Needham Center that would require the downtown grade separation.

According to the 2018 passenger counts, the 4 stations in question have a combined ridership of 1,440 riders. Of these, less than 16% come from Center. To put another way, 84% of current Needham commuters could be served by grade separated rapid transit without build an el/tunnel/trench through downtown.

(And, in fact, the addition of a new station at Gould St may expand access -- or at least increase the number of residents in walking distance of a station.)

So... in this distant, hypothetical future, in which there is no time for mitigations, no time to build community support... and in which there is somehow drastic urgency to run heavy rail over the Highland Branch... and all the other challenges of the LRT -> HRT conversion have been otherwise accounted for... in this scenario, it might just be viable to run HRT to Hersey, Junction, and Heights, and run a (cute) "Needham Trolley" between Heights and Junction, to serve downtown.

The large majority of riders would maintain their service. And while Center riders would need to transfer, a trolley could justify the addition of infill stops at Rosemary St and Oak St, improving pedestrian access to transit for more members of the community, and provide improved within-town transit in a downtown that is well-configured for heavy pedestrian and transit orientation.

As is hopefully obvious, this is a contingency plan for a contingency plan for a contingency plan. I remain steadfast in my opinion that a Green Line/Gold Line extension to Needham Junction remains the best way to serve Needham in the context of the larger system.

But... if you really had to... you could probably swing HRT to 85% of riders without needing to disrupt downtown.

 

[Teban54]

Question: Would a trench through Needham really be that expensive and/or disruptive? After all, there aren't that many grade crossings, and similar tasks are also needed for OLX to Reading (see recent discussion) which are much more NIMBY than Needham.

 

[kdmc]

That was a fun read, but I want to pick at this assumption:

According to the 2018 passenger counts, the 4 stations in question have a combined ridership of 1,440 riders. Of these, less than 16% come from Center. To put another way, 84% of current Needham commuters could be served by grade separated rapid transit without build an el/tunnel/trench through downtown.

Are Needham CR ridership patterns really reflective of what Needham GL ridership patterns will look like? The former is 1hr peak headways, the latter is, what, <7min peak headways?

1hr headways work for... commuting. 7min headways work for commuting, but also for errands, appointments, dining, going for drinks with friends, etc, all things that I imagine would benefit ridership for "downtown village" stops like Needham Center. Under that context, I imagine a 1SR vs 2SR makes a bigger difference, no?

 

[TheRatmeister]

Question: Would a trench through Needham really be that expensive and/or disruptive?

Given that there isn't a historic station building to worry about, and the current Needham Center is surrounded by parking lots you can just dig up, I'd lean towards no.

 

[Riverside]

Are Needham CR ridership patterns really reflective of what Needham GL ridership patterns will look like? The former is 1hr peak headways, the latter is, what, <7min peak headways?

1hr headways work for... commuting. 7min headways work for commuting, but also for errands, appointments, dining, going for drinks with friends, etc, all things that I imagine would benefit ridership for "downtown village" stops like Needham Center. Under that context, I imagine a 1SR vs 2SR makes a bigger difference, no?

Great point, and it is true that I am making a slight conflation: a Needham Trolley + HRT 2SR would be replacing an LRT 1SR, and would not be replacing today's CR 1SR; compared to the CR 1SR, the Needham Trolley concept does fine, but you're right that it would be a more dramatic regression if it is following decades of an LRT 1SR.

If we were building the system from a blank slate, though, rapid transit terminals at opposite ends of downtown, with some sort of surface route (trolley or bus) pinging between them, would seem reasonable enough. In this scenario, imagine Needham Heights/Junction as being less like Newton Highlands and more like Quincy Center or Malden Center.

But yes, that is a drawback, for sure.

Re @Teban54's question about the actual infeasibility: yeah I dunno. @F-Line to Dudley has strong opinions on the topic. I'm sorta agnostic on the question ultimately: it should be converted to LRT in the near-future because that would be the least expensive way to get Needham trains out of the NEC. In the long-term, a lot of things will need to change before a full-on conversion of the Green/Gold Line in general, and the Highland Branch in particular, becomes a pressing concern. When/if that happens, the situation in Needham will be significantly different from today regardless.

All of this does make me reconsider whether an OLX all the way to Needham Junction makes more sense than I'd previously thought. Hersey and (to a lesser extent) Junction both have ridership figures on par with the Roslindale/West Roxbury figures.