Green Line Reconfiguration

[TheRatmeister]

I'll need to go back and look at Fairmount more closely, but for now a couple quick points:

A single branch is unlikely to be able to absorb more demand originating south of Nubian.

The D Branch, operating 8-10 minute headways, carried just over 200,000 passengers per week in 2019. Assuming double Type 10s can carry twice as many pax as double Type 7s/8s and consistent 5 minute headways, which is definitely doable for a median running light rail line, that would put the very rough theoretical maximum number of weekly passengers for the Washington St light rail at around 600,000-800,000. In 2022, SL4/5 had ~75,000 weekly riders, meaning ridership could increase at least 8 fold with just one GL branch. I'm not convinced capacity is a major concern right out of the gate.

I find it hard to believe a new extension can run faster

Given that the existing branches have no signal priority and a bajilllion intersections, I don't find it hard to believe at all. To stress the point, here are the average speeds of some other tram systems:

City	Speed
Amsterdam	~10mph
Berlin	~11.5mph
Helsinki	~9.5mph in the City Center
Krakow	~11mph
Zurich	~10.5 mph
Prague	~12 mph (Within city)
Minneapolis (Green Line)	~14mph
Toronto (Queensway Streetcar)	~10mph

 

[Riverside]

Here in Berlin or here in Salt Lake City. Both are about 2 mile segments with ~10 minute travel times. Both feature 7-8 stops, close to the 8 or 9 that would make sense on the Washington St Line.

Given that the existing branches have no signal priority and a bajilllion intersections, I don't find it hard to believe at all. To stress the point, here are the average speeds of some other tram systems:

City	Speed
Amsterdam	~10mph
Berlin	~11.5mph
Helsinki	~9.5mph in the City Center
Krakow	~11mph
Zurich	~10.5 mph
Prague	~12 mph (Within city)
Minneapolis (Green Line)	~14mph
Toronto (Queensway Streetcar)	~10mph

These are interesting examples, definitely need to look at these more closely.

The only piece I'd say without first reviewing further is that we know from the examples of the D Line, GLX, and the Orange Line that full grade separation with 1-mile stop spacing can produce speeds upward of 16 mph (if not close to 20mph) -- all of which blow most of those examples out of the water. I appreciate that your point is that a surface line may be sufficiently fast, but I do think it's worth not losing sight of the size of the difference.

Depends on how brutal you want to be. The general principle is that alleys never have left turns, nor do quiet residential streets if it doesn't severely limit accessibility. Main streets like Melnea Cass and Mass Ave obviously keep their intersections. Here's the changes I would make from a surface level glance:

• Tremont St @ Oak St West: Change the intersection to only allow for right turns. This would mean EB Tremont->SB Shawmut and SB Tremont->EB Oak St only, no straight or left turning traffic.
• Washington St between Oak St and Marginal Rd: 4 lanes, transit runs on the 2 western lanes and vehicle traffic runs on the two eastern lanes. Crossover into the Median happens at the Marginal Rd intersection. This makes the transition between the Tremont St subway and the Washington St surface route seamless with no intersections or street running.
• Herald St + Marginal Rd: Traffic lights at both to act as one mega intersection despite being split by I-90
• William E Mullins Way and Traveller St: No left turns on Washington
• Perry, Savoy, Rollins, Waltham, and Union Park: No left turns on Washington
• Mystic St, Brookline St: No left turns on Washington
• Concord St, Worcester St, Springfield St: No left turns on Washington
• Northampton, Lenox, Newcombe, Thorndike, Ball Streets: No left turns on Washington
• Eustis/Williams and Dade Streets: No left turns on Washington
• Ruggles St: Extend through the parking lot to meet Harrison Ave, add traffic light
• Make Washington transit only between Ruggles/Eustis, widen to 4 lanes between Ruggles/Warren.

So with all these changes made, this would bring the total number of intersections the line would need to cross between Nubian and Eliot Norton Park down from ~20 (Depending on how you count it) to 7. Even if you're less aggressive than this it's still possible to reduce the number of car crossings substantially.

Some interesting thoughts here! I do think keeping Herald, E Berkeley, W Dedham/Reynolds, Newton and Mass Ave make sense. At least one of Brookline and Concord likely needs to be kept though, as they're one-way streets paired with Newton as well as offering access to parts of BMC. I also imagine most of these intersections will still see heavy pedestrian traffic, meaning you'll still need pedestrian crossings, but they're at least easier to manage than cars.

Definitely interesting proposals here!

My main concern is actually the point that Teban54 makes at the end: pedestrians. If I understand what you are suggesting, we would see distances of over 1500 feet between crosswalks. That's pretty far, and is a big change from the status quo. That seems like a significant barrier dividing the neighborhood down the middle. It may be workable along Prenzlauer Allee in Berlin, but that street is much wider and therefore already creates a significant division.

You could do pedestrian-only crossings, to provide intermediate crossings, which I suppose would be better than traffic intersections in that you can significantly shorten the light cycles. But it still means that the top vehicle speed needs to be low enough to brake in time to avoid hitting a pedestrian.

~~~

The good news, in the midst of all of this, is that a surface branch of the Green Line is more likely to be built first anyway. So there will be opportunity to iterate and improve on that infrastructure, testing its limits in real life, before a decision would need to be made about an I-93 alignment.

 

[HenryAlan]

You could do pedestrian-only crossings, to provide intermediate crossings, which I suppose would be better than traffic intersections in that you can significantly shorten the light cycles. But it still means that the top vehicle speed needs to be low enough to brake in time to avoid hitting a pedestrian.

The E-Line in Los Angeles has some of these, and they work pretty well from what I've seen. Pictured is the crossing between USC and Exposition Park, which has a ton of pedestrian activity, particularly during Saturdays in the Fall. I believe the bollards are retractable, to accommodate things like emergency or maintenance vehicles. I think the key element is that there is lots of space for pedestrians to gather while they wait for a train to pass.

 

[Teban54]

The D Branch, operating 8-10 minute headways, carried just over 200,000 passengers per week in 2019. Assuming double Type 10s can carry twice as many pax as double Type 7s/8s and consistent 5 minute headways, which is definitely doable for a median running light rail line, that would put the very rough theoretical maximum number of weekly passengers for the Washington St light rail at around 600,000-800,000. In 2022, SL4/5 had ~75,000 weekly riders, meaning ridership could increase at least 8 fold with just one GL branch. I'm not convinced capacity is a major concern right out of the gate.

While I do think you have a good point here, I'll first make a (somewhat minor) correction: In 2019, the D branch ran 9 tph (6-7 mins) during rush hours, and in 2022 it still ran 7 tph (8-9 mins) despite the slow zones. (I also checked the current schedules, and while the D particularly seems to average about 9 min most of the day, the B, C, E branches all seem to average 8-min headways. I'm not sure if the D is really less frequent than the others today.) Additionally, given that any Nubian branch will need to fit into the Green Line's schedule, I doubt you can have frequencies that differ substantially from the "baseline frequencies" of each GL branch -- which are defined as 8-9 mins today. So your headways probably need to be integer multiples of that, either 8-9 mins ("1 branch") or 4-4.5 mins ("2 branches"). The latter -- and the 5-min in your comment -- is where I become skeptical, particularly in terms of execution along a rather sizable street-running segment, thanks to the many pedestrian crossings; that is also where I proposed having both "express" and "local" services, instead of the "2 branches" all running down Washington surface.

Double checking the ridership numbers: Assume that 200,000 is the "practical capacity ceiling" of a 2-car Type 8 LRT line with 9 tph peak (the D actually ran triple Type 7/8s at that time), that each 2-car Type 7/8 has 3x as much capacity as a 60' bus (see below), and that SL4/5 ran combined 12.5 tph peak in Fall 2019 (Source: Bus Route Profiles). Then, the capacity ceiling of SL4/5 would be 92,593. In Fall 2019, SL4/5 had 95,016.7 boardings per week, meaning SL4/5 was roughly as crowded as the D.

Using the same calculation, assuming that each 2-car Type 10 carries twice as many passengers as a 2-car Type 7/8, a "baseline branch" would have a capacity ceiling of 400,000/week, or 4x as much as what SL4/5 already carried in 2019. This is still quite a substantial improvement, for sure, but not as much as the 8x that you mentioned. In fact, the difference is probably because you were implicitly assuming 2 branches' worth of service.

Finally, I'll remark that a surface light rail branch alone will already bring substantial induced demand compared to SL4/5. The 2003 PMT estimated Nubian light rail to have 34,300 daily riders, or 2.25x as much as SL4/5 in 2019 (15249.1). This means the present-day SL4/5 is severely underperforming. Given that the reports states "the majority of riders would be diverted from Washington Street Silver Line Bus Rapid Transit service", such induced demand appears to be mostly from South End, not Nubian. 2.25x as much ridership is still below the 4x as much capacity of a single branch, but I'm not sure how much the 2003 PMT figure accommodated future growth and TOD in South End, which we know has already been rapidly gentrifying today. Overall, it means attracting induced demand from Nubian -- particularly the ~50% bus riders going all the way to Ruggles -- will likely still be difficult considering to both capacity and speed.

Spoiler: Vehicle capacity calculations

Per 2014 Blue Book, each Green Line car has policy load of 101, and crush loads of 269 (Type 7) and 199 (Type 8). Each 60' bus has policy load of 79 and crush load of 104. Comparing a 2-car Type 7/8 to a 60' bus, the former has 2.55x policy load and 4.5x crush load. I therefore chose 3x as a blend of the two.

Aside: Do you have a citation for the D's 200,000 pax/week? The 2014 Blue Book included 2010-11 ridership numbers, which had 24,632 on the D's surface stations per weekday (boardings in both directions). (This extrapolates to 147,792 for the whole week, if we assume Sat and Sun each generates half the ridership of a weekday.) While this seems on the same scale as 200k, the real weirdness came when I tried the 2017-19 ridership data: Summing up "average ons" for all D branch stations in Fall 2019 got me a measly 5,590 boardings per day. Fall 2017 and 2018 had 14,520 and 13,737 respectively, but both are still a magnitude lower than the Blue Book data and your figure.

Given that the existing branches have no signal priority and a bajilllion intersections, I don't find it hard to believe at all.

My only concern is: If the T becomes willing and able to implement signal priority in the future, aren't they more likely to appear on the B/C branches first? Both branches also seem to have fewer intersections today than Washington St, and will probably remain true under crossing elimination proposals, yet they do little in speeding up the branches today. These things can be done, but the current state in Boston doesn't inspire confidence that they realistically will.

Another thing that's perhaps even more detrimental is the T's alleged 25 mph restriction on speed-running segments, and 10 mph in mixed traffic (perhaps someone like @nbcoram can confirm?). While that's an organizational issue that can theoretically be changed, again, there don't seem to be signs for changes anytime soon. The nature of Bostonians to cross anywhere they like may also have contributed to the restrictions, and that will likely be worse on Washington St than both Comm Ave (so wide that it's unfriendly for pedestrians) and Beacon St (trees surrounding most of the C branch reservation).

 

[TheRatmeister]

My only concern is: If the T becomes willing and able to implement signal priority in the future, aren't they more likely to appear on the B/C branches first? Both branches also seem to have fewer intersections today than Washington St, and will probably remain true under crossing elimination proposals, yet they do little in speeding up the branches today. These things can be done, but the current state in Boston doesn't inspire confidence that they realistically will.

Quite possibly not. (My undersanding is that) The big problem with the B/C/E branches is that the roads in the area are managed by a bunch of different organizations, and traffic lights use incompatible protocols. The City of Boston, Town of Brookline, DCR, and MassDOT all manage some roads and lights in the area. As far as I know the City of Boston manages Washington St, its intersecting streets, and all the traffic lights along it, so getting TSP done here might actually be easier than the GL branches.

In fact, Washington St has TSP right now. I don't think it's at all intersections and used all the time, but the infrastructure might largely already be there right now.

 

[Teban54]

In fact, Washington St has TSP right now. I don't think it's at all intersections and used all the time, but the infrastructure might largely already be there right now.

I was surprised! Indeed, the Boston MPO's TSP Guidebook in December 2018 mentions this:

In 2008, the City of Boston worked with the MBTA to give priority to Silver Line buses at four signalized intersections along Washington Street. In 2012, the system was expanded to include an additional four intersections. TSP was also implemented at four intersections along MBTA bus route 57: at Washington and Brock Street/Lake Street; at Washington Street and Foster Street; at Cambridge Street and Gordon Street; and at Commonwealth Avenue and Babcock Street.

More recently, the MBTA has implemented pilot TSP projects on the Green Line, launching TSP at one intersection on each of the “B” and “E” branches in May 2017, and at one intersection on the “C” branch in June 2017. An additional pilot project in Cambridge, at the intersection of Massachusetts Avenue and Brookline Street, serves MBTA bus route 1. After four months in operation on the “B” and “E” branches, the MBTA reported an average green-light-time extension of 14 seconds, an average red-light-time reduction of eight seconds, and no demonstrable negative effects on non-prioritized traffic. On the “C” line, green-light time was extended by an average of 10 seconds, red-light time was reduced by an average of six seconds, and there was again minimal disruption to non-prioritized traffic.

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

As an unintended finding from Googling this, I also found a Route 57 Transit Priority Corridor Project that the the City of Boston launched in August 2023. Signal priority was mentioned as part of the transit priority toolkit (hence showed up in search results), but the project, still in its public inputs stage, focuses primarily on bus lanes. It also showed a few corridors where the existing bike lanes are proposed to be converted to bus/bike lanes as one of the options, but the streets are too narrow for both bus lanes and bike lanes separately.

 

[lainpimicaja]

I was surprised! Indeed, the Boston MPO's TSP Guidebook in December 2018 mentions this:

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

As an unintended finding from Googling this, I also found a Route 57 Transit Priority Corridor Project that the the City of Boston launched in August 2023. Signal priority was mentioned as part of the transit priority toolkit (hence showed up in search results), but the project, still in its public inputs stage, focuses primarily on bus lanes. It also showed a few corridors where the existing bike lanes are proposed to be converted to bus/bike lanes as one of the options, but the streets are too narrow for both bus lanes and bike lanes separately.

FWIW the T has all of their preferred TSP treatments and standards outlined in their TP Toolkit (see chapter 3, starting on page 30).

 

[Riverside]

Huntington Extended Subway Stations

So, if we extend the subway under Huntington, do we keep the existing stop locations? In theory, one might think that moving from a surface light rail line to a subway would have fewer stations, but, on closer inspection, that's not obvious to me.

It's about 6,300 feet from Francis St to Mass Ave (each alternating color shows a 500' stretch):

The existing stop spacing is roughly every 1500 feet. There is some variance to that, in part to accommodate surface conditions. For example, the Francis St stop (Brigham Circle) is actually located some 200' away from Francis St.

Simplifying things slightly, 1500' spacing would put subway stops at

• Francis St
• Longwood Ave
• Ruggles St
• Forsyth St

Things get a little stretched at the eastern end. If you measure 1500' from Mass Ave, the station locations don't change that much -- the LMA-equivalent station moves from Longwood Ave to St Alphonsus St -- so I'm willing to overlook the slight deviation from the idealized model.

1500' is comparable to the distance between most of the core downtown stations.

---

At 2000 feet, we have stations at

• Francis St
• Evans Way
• Parker St

One notable downside here is that you lose a station at Ruggles St, the potential location for an Urban Ring busway. This station location also means that LMA, MFA, and Northeastern riders all see their stop move significantly (and hospitals along Longwood Ave see a longer walk).

2000' is an unusual spacing without many comparable ones across the system. Assuming an infill at Kent Square/Netherlands, the Fenway Branch from Fenway to Brookline Hills would be the most similar.

---

3000' spacing mirrors the Orange Line stations to the east, providing stations at

• Francis St
• Ruggles St

This option regains the Ruggles St connection, but loses more direct service to Northeastern and to Longwood Ave.

Alternatively, you could move the pair of new stations further east -- reducing the spacing between the Northeastern stop and Symphony, but centering on the current Northeastern and LMA stops, which I think are the highest among this quartet of stations. (I am struggling with pivot tables tonight, so someone should check me.)

This option (likely) eliminates a Ruggles St connection, but, as mentioned, regains access to more of Northeastern and LMA.

3000' is comparable to Orange Line Southwest Corridor spacing, as well as spacing on the Ashmont Branch, and (to a varying extent) the spacing on GLX.

---

So which one is best?

In general, the 3000' options seem too far apart to me. That's a lot of walks which will be made significantly longer (almost certainly longer than the potential time savings on the train).

The 1500' option honestly does not strike me as unreasonably close. High ridership stretch, elongated linear destination zone, and comparable to Boston's original downtown. And, even with the benefits of a "heavy metro" design, this subway would still be served by LRT vehicles, running shorter consists than HRT: having four stations means you can having four trains loading and unloading passengers simultaneously; fewer stations mean fewer vehicle doors opening simultaneously. So, 1500' still seems like a reasonable contender.

In general, 2000' strikes a good happy medium. Existing walks are lengthened somewhat, but not severely. New stations focus on points of higher ridership. The only fly in the ointment is the potential loss of an Urban Ring connection at Ruggles St. There are, of course, many ways to address that; while my (current) Urban Ring proposal calls for a crossing via Ruggles St, other proposals do not.

So, one thing I think this examination does reveal is that the value of a Ruggles St station really is the question to answer regarding stop spacing. If it's not needed, then you have a lot more flexibility, and probably the 2000' option would be fine. But if the Ruggles St station is required, then (IMO) the only option is to continue with 1500'.

For the moment, I plan to continue crayoning using the 1500' model. One simple reason: I don't think the potential public discussion about (what is essentially) stop consolidation is particularly useful at this point. Insofar as I have a platform for advocacy, I'd like to keep attention focused on the idea of an extended Huntington Subway itself, and not get too caught up in station placement. (That's not to forestall such discussion here; I'm just saying that I see it as secondary.)

 

[The EGE]

I think spacing depends a lot on what the plan is for serving the LMA - i.e, whether the Huntington mainline is the primary access from the rapid transit network. If it is, then you need station entrances as close to Francis and Longwood as possible, which lends itself to a 1500' spacing. The current configuration with the staggered platforms at Brigham Circle is pretty bad for this.

If there's an actual LMA station via new construction - D-E connector via Longwood Avenue (my preferred configuration), Urban Ring, Blue Line via Brookline Avenue - then you have more leeway on the Huntington stop locations. My crayon map is 2000' spacing, with stations centered between Forsyth and Hemenway, at Evans Way, and at Francis.

 

[TheRatmeister]

Huntington Extended Subway Stations

So, if we extend the subway under Huntington, do we keep the existing stop locations? In theory, one might think that moving from a surface light rail line to a subway would have fewer stations, but, on closer inspection, that's not obvious to me.

It's about 6,300 feet from Francis St to Mass Ave (each alternating color shows a 500' stretch):

View attachment 48970

The existing stop spacing is roughly every 1500 feet. There is some variance to that, in part to accommodate surface conditions. For example, the Francis St stop (Brigham Circle) is actually located some 200' away from Francis St.

Simplifying things slightly, 1500' spacing would put subway stops at

• Francis St
• Longwood Ave
• Ruggles St
• Forsyth St

View attachment 48971

Things get a little stretched at the eastern end. If you measure 1500' from Mass Ave, the station locations don't change that much -- the LMA-equivalent station moves from Longwood Ave to St Alphonsus St -- so I'm willing to overlook the slight deviation from the idealized model.

1500' is comparable to the distance between most of the core downtown stations.

---

At 2000 feet, we have stations at

• Francis St
• Evans Way
• Parker St

View attachment 48972

One notable downside here is that you lose a station at Ruggles St, the potential location for an Urban Ring busway. This station location also means that LMA, MFA, and Northeastern riders all see their stop move significantly (and hospitals along Longwood Ave see a longer walk).

2000' is an unusual spacing without many comparable ones across the system. Assuming an infill at Kent Square/Netherlands, the Fenway Branch from Fenway to Brookline Hills would be the most similar.

---

3000' spacing mirrors the Orange Line stations to the east, providing stations at

• Francis St
• Ruggles St

View attachment 48973

This option regains the Ruggles St connection, but loses more direct service to Northeastern and to Longwood Ave.

Alternatively, you could move the pair of new stations further east -- reducing the spacing between the Northeastern stop and Symphony, but centering on the current Northeastern and LMA stops, which I think are the highest among this quartet of stations. (I am struggling with pivot tables tonight, so someone should check me.)

View attachment 48983

This option (likely) eliminates a Ruggles St connection, but, as mentioned, regains access to more of Northeastern and LMA.

3000' is comparable to Orange Line Southwest Corridor spacing, as well as spacing on the Ashmont Branch, and (to a varying extent) the spacing on GLX.

---

So which one is best?

In general, the 3000' options seem too far apart to me. That's a lot of walks which will be made significantly longer (almost certainly longer than the potential time savings on the train).

The 1500' option honestly does not strike me as unreasonably close. High ridership stretch, elongated linear destination zone, and comparable to Boston's original downtown. And, even with the benefits of a "heavy metro" design, this subway would still be served by LRT vehicles, running shorter consists than HRT: having four stations means you can having four trains loading and unloading passengers simultaneously; fewer stations mean fewer vehicle doors opening simultaneously. So, 1500' still seems like a reasonable contender.

In general, 2000' strikes a good happy medium. Existing walks are lengthened somewhat, but not severely. New stations focus on points of higher ridership. The only fly in the ointment is the potential loss of an Urban Ring connection at Ruggles St. There are, of course, many ways to address that; while my (current) Urban Ring proposal calls for a crossing via Ruggles St, other proposals do not.

So, one thing I think this examination does reveal is that the value of a Ruggles St station really is the question to answer regarding stop spacing. If it's not needed, then you have a lot more flexibility, and probably the 2000' option would be fine. But if the Ruggles St station is required, then (IMO) the only option is to continue with 1500'.

For the moment, I plan to continue crayoning using the 1500' model. One simple reason: I don't think the potential public discussion about (what is essentially) stop consolidation is particularly useful at this point. Insofar as I have a platform for advocacy, I'd like to keep attention focused on the idea of an extended Huntington Subway itself, and not get too caught up in station placement. (That's not to forestall such discussion here; I'm just saying that I see it as secondary.)

In general I would favor shorter stop spacing. Huntington is an area with extremely high node density and ridership. Here's the combined daily boardings from the GL and 39 along Huntington:

Spoiler: Caveat

I've been unable to find reliable GL surface ridership data more recent than 2013. The 2019 Rail Ridership data from the MBTA Blue Book seems to be massively undercounting ridership, only predicting ~800 boardings at Brookline Village, whereas in 2013 there were over 3,000. If anyone has any insights or other more recent data please let me know.

With stops mostly at the existing locations, and assuming a revived Arborway Branch because the ridership and urban fabric is definitely there, LMA would become one of the busiest stops on the entire GL, even beating out some stops on the Central Subway like Boylston, Haymarket, and Government Center. This would likely be further compounded by increased ridership from the existing D branch, where the ~2500 riders from Longwood and some of the ~3000 from Brookline Village would probably switch over as well. This ridership data also suggests that Northeastern and Brigham Circle would be powerhouse stops as well, moving even higher up to near the top of the list for busiest surface stops. MFA is a bit of an exception, but it's enough ridership combined with its excellent location on Ruggles St and potential for extreme seasonal or event demand that it's still worth keeping in my opinion. I wouldn't make any significant changes to the existing stop locations or spacing.

 

[HenryAlan]

having four stations means you can having four trains loading and unloading passengers simultaneously; fewer stations mean fewer vehicle doors opening simultaneously.

This is probably an underrated concept, and indeed a strength of LRT subways. The question is probably more connected to how many people might be boarding each of those trains at 1500 feet vs the other distances, and whether the trains can handle the larger numbers associated with fewer doors without suffering a dwell time penalty. I suspect the passenger density along that corridor is more than enough to justify four geographically close stations.

 

[Riverside]

There are, of course, many ways to address [the loss of a radial/circumferential connection at Ruggles & Huntington]; while my (current) Urban Ring proposal calls for a crossing via Ruggles St, other proposals do not.

This got me thinking. The latest iteration of my crayon/GLR map has three services using Ruggles St: an H Harvard <> Ruggles service, an L Brigham Circle <> Mattapan service, and an M Brigham Circle <> BU Med Ctr service. The L and M run along Huntington before turning on to Ruggles, so aren't a concern here. So, really, the 2000' spacing would only impact something like my Harvard <> Ruggles service.

But, truth be told, I was always a little skeptical of my H service running all the way to Ruggles. Ruggles is a transfer hub, but isn't really a destination on its own.

Particularly since its use of The Fenway road means that it doesn't serve the hospitals of LMA very well, the potential strength of the H rests on its ability to serve something like the Allston <> Nubian-ish market (below). And truncating at Ruggles doesn't achieve that goal particularly well. I had previously been imagining this as a through-Longwood service, but H riders needing to transfer at Ruggles undercuts the strength of that service.

All of which is to say, maybe it wouldn't be so bad to reconfigure these services to not require a transfer at Huntington & Ruggles. Something like this:

H: Harvard <> Brigham Circle
L: Brigham Circle <> Nubian via Ruggles
M: Brigham Circle <> BU Med Ctr via Ruggles

H gets more direct service to Longwood, and through-Longwood journeys are still possible through a cross-platform transfer at LMA station. (Which, yes, should be renamed.) And given that the BNRD is relocating a number of bus transfers from Nubian/Ruggles directly into the LMA (more on that below), having the Urban Ring services do the same has benefits as well.

One downside here is that Regional Rail commuters transferring at Ruggles will have a less-direct route to BU and to Harvard's Allston campus. Transferring at Back Bay and doubling back to Lansdowne or West Station isn't the worst solution, but it's not great especially for Allston, where an additional transfer at West Station will probably be required. So, it'll definitely a question of comparing pros and cons.

~~~

[I was gonna post something further about integrating a Dorchester BRT network into this whole thing but now I've gone deep down the mapmaking rabbit hole, so I'll go ahead and hit the "post" button now, and come back with more thoughts later.]

 

[Riverside]

I think spacing depends a lot on what the plan is for serving the LMA - i.e, whether the Huntington mainline is the primary access from the rapid transit network. If it is, then you need station entrances as close to Francis and Longwood as possible, which lends itself to a 1500' spacing.

Yes, definitely agree with this, and I agree that Francis and Longwood are the keys.

In general I would favor shorter stop spacing. Huntington is an area with extremely high node density and ridership. Here's the combined daily boardings from the GL and 39 along Huntington:

Spoiler: Caveat

I've been unable to find reliable GL surface ridership data more recent than 2013. The 2019 Rail Ridership data from the MBTA Blue Book seems to be massively undercounting ridership, only predicting ~800 boardings at Brookline Village, whereas in 2013 there were over 3,000. If anyone has any insights or other more recent data please let me know.

View attachment 48994

With stops mostly at the existing locations, and assuming a revived Arborway Branch because the ridership and urban fabric is definitely there, LMA would become one of the busiest stops on the entire GL, even beating out some stops on the Central Subway like Boylston, Haymarket, and Government Center. This would likely be further compounded by increased ridership from the existing D branch, where the ~2500 riders from Longwood and some of the ~3000 from Brookline Village would probably switch over as well. This ridership data also suggests that Northeastern and Brigham Circle would be powerhouse stops as well, moving even higher up to near the top of the list for busiest surface stops. MFA is a bit of an exception, but it's enough ridership combined with its excellent location on Ruggles St and potential for extreme seasonal or event demand that it's still worth keeping in my opinion. I wouldn't make any significant changes to the existing stop locations or spacing.

Thanks for pulling the numbers! Yeah, when I was staring at the pivot tables last night, it seemed like the official figures for Prudential were, like, an order of magnitude above the other stops? I thought I was just getting confused, but maybe there's something weird in the data there.

In any case, I think you make a compelling case here. MFA sits on the lower side, so if we had to drop one, then I'd probably choose that and use the 2000' spacing. On the other hand, dropping MFA would mean adding about 900 riders each to Northeastern and LMA, which would send LMA's numbers in particular up into the stratosphere. Which brings us back to The EGE's point about the ultimate location of an LMA subway station.

And yeah, I think HenryAlan sums it up well here:

This is probably an underrated concept, and indeed a strength of LRT subways. The question is probably more connected to how many people might be boarding each of those trains at 1500 feet vs the other distances, and whether the trains can handle the larger numbers associated with fewer doors without suffering a dwell time penalty. I suspect the passenger density along that corridor is more than enough to justify four geographically close stations.

 

[Teban54]

Huntington Extended Subway Stations
[...]

I've had some thoughts on Huntington Ave's stop spacing as well, though I came to slightly varied conclusions. Due to busy IRL matters, I'll share my thoughts in pieces across various comments over the next several days.

Stop spacing

First, a soft disagree on this:

The 1500' option honestly does not strike me as unreasonably close. High ridership stretch, elongated linear destination zone, and comparable to Boston's original downtown. And, even with the benefits of a "heavy metro" design, this subway would still be served by LRT vehicles, running shorter consists than HRT: having four stations means you can having four trains loading and unloading passengers simultaneously; fewer stations mean fewer vehicle doors opening simultaneously. So, 1500' still seems like a reasonable contender.

In addition to density within this region, another factor to consider is speed for through-Huntington traffic. In this case, it contains significant demands from Brookline, Reservoir, Newton, Riverside and Needham. They see the following numbers of stops between Brookline Village and Copley (exclusive on both ends):

• Today, via Kenmore: 4 stops
  • At most 5 stops with an Aspinwall infill
• Huntington reroute: 7 stops, roughly with today's stop spacing (with Fenwood Rd and Riverway eliminated)

That a shorter, more direct route results in more steps is counterintuitive, and I'm afraid will hurt ridership. Also, Mission Park has been omitted in the discussion above, but whether it should be retained does matter.

I'll also note that stop spacing of 3000' upwards is not unique to HRT lines. GLX averages about 3400' per stop between Medford/Tufts and East Somerville. That is indeed excessive on Huntington given the region's job density (not just LMA, but also Northeastern, MFA and many other colleges), but it's not without precedent.


LMA stops
@Riverside's initial comment assumes a station at Francis St no matter what. I'd like to challenge that assumption.

If you want to only keep 1 stop in LMA around the hospital area, I'd suggest the midpoint between Longwood Ave and Francis St:

This is the current northern exit of Brigham Circle. The reason why this works is the abundance of pedestrian paths in the area, which is pretty self-explanatory from the diagram:

• To Longwood Ave: Similar walking distance
  • The top green path is similar to the red path
• To blocks between Longwood Ave and Francis St (e.g. Dana-Farber): Similar walking distance, if not slightly shorter
• To Francis St: 500' longer, but not too much
  • This is mostly Brigham and Women's, which already enjoys a shorter walk than the other hospitals anyway

If there's another stop at Ruggles St, the only ones whose walks become significantly longer are MassArt, which isn't as major of a ridership generator as the hospitals. I'm not saying there should only be one station within this block, but if there is, I think this a better location than either Longwood Ave or Francis St intersections.


I'll continue with more thoughts later, but a very minor point:
Arborway

assuming a revived Arborway Branch because the ridership and urban fabric is definitely there

I'll check this again, but I'm very sure this can't be done with dedicated transit lanes. Several parts of Centre St and South St are 40-45' wide, and those that are 45' wide already have dedicated bike lanes in one or both directions. Even with all parking removed, there's literally not enough room for 2 LRT lanes, 2 travel lanes and 2 bike lanes. To make things worse, the road network in the region is fundamentally broken and heavily depends on Center St and South St, so you can't make travel lanes one-way either. Shared bus/bike lanes are much more feasible, which would argue against LRT conversion.

 

[TheRatmeister]

I'll check this again, but I'm very sure this can't be done with dedicated transit lanes. Several parts of Centre St and South St are 40-45' wide, and those that are 45' wide already have dedicated bike lanes in one or both directions. Even with all parking removed, there's literally not enough room for 2 LRT lanes, 2 travel lanes and 2 bike lanes. To make things worse, the road network in the region is fundamentally broken and heavily depends on Center St and South St, so you can't make travel lanes one-way either. Shared bus/bike lanes are much more feasible, which would argue against LRT conversion.

These are not unsolvable problems though. Yes, they will make designing dedicated transit lanes more difficult and will have a high cost in terms of political capital. But if we can be confident in the results, which I think we can be, there's a reason Arborway service lasted until 1985 and why the 39 is still so popular, then at some point you just have to make the decision to use that political capital to force it through, just like the Hyde Square extension or the Columbus Ave bus lanes. Both have almost all the same problems, with the exception of South Huntington being just wide enough for bike lanes, but I don't think there's much hesitation at the government level about ultimately pushing either project through despite opposition from drivers. Arborway service can also count on the support of some local leaders and residents who will still remember the original streetcar service, and through quality public outreach and events I think local businesses would at least lessen in their opposition, if not turn the corner completely. In my mind, it's very much a "So long as we choose to think it's not possible then it will never be" project.

 

[Teban54]

there's a reason Arborway service lasted until 1985 and why the 39 is still so popular

Neither the pre-1985 Arborway branch nor today's 39 had dedicated transit lanes, nor did they have to face the political challenges of either removing bike lanes (a really, really bad look) or turning the street one-way (much bigger issue than removing parking).

just like the Hyde Square extension

South Huntington Ave is notably wider. Today, it already has 2 dedicated bike lanes and parking lanes on both sides. You can already extend the E to Hyde Square by just removing parking (which is already a daunting task, but at least more realistic). But to go further south requires even more than that.

or the Columbus Ave bus lanes

The existence of Columbus Ave bus lanes already means there's no hope for dedicated bike lanes on the corridor. Correct me if I'm wrong, but I don't think bike lanes existed on that corridor previously, anyway. That's a distinct problem from taking away existing bike lanes and putting bikes back to car traffic.

but I don't think there's much hesitation at the government level about ultimately pushing either project through despite opposition from drivers

Not just from drivers. You'll also get opposition from cyclists. That's the point.

 

[HenryAlan]

The existence of Columbus Ave bus lanes already means there's no hope for dedicated bike lanes on the corridor. Correct me if I'm wrong, but I don't think bike lanes existed on that corridor previously, anyway. That's a distinct problem from taking away existing bike lanes and putting bikes back to car traffic.

You are correct, there were never bike lanes there. However, I would definitely argue that the current arrangement is less bike friendly than before, because there really is no option but to take the lane if you bike that stretch, and many are not comfortable doing so. The plan is to establish alternative parallels routes for biking, so that there are still connections between Roxbury and JP. Not enough of this has happened yet, and if you look at the street grid, it's not hard to see why the city is having difficulty figuring out the routes. That said, as a bike rider and also a transit advocate, I will always favor bus priority lanes over bike infrastructure if there is no option for having both.

 

[JeffDowntown]

You are correct, there were never bike lanes there. However, I would definitely argue that the current arrangement is less bike friendly than before, because there really is no option but to take the lane if you bike that stretch, and many are not comfortable doing so. The plan is to establish alternative parallels routes for biking, so that there are still connections between Roxbury and JP. Not enough of this has happened yet, and if you look at the street grid, it's not hard to see why the city is having difficulty figuring out the routes. That said, as a bike rider and also a transit advocate, I will always favor bus priority lanes over bike infrastructure if there is no option for having both.

For the Columbus Avenue corridor with the center running bus lanes, why would we even consider bike lanes? There are bike paths in the Southwest Corridor Park running adjacent to Columbus. This corridor has had some of the best (and heavily used) bike infrastructure in the city for over 30 years!

 

[TheRatmeister]

Neither the pre-1985 Arborway branch nor today's 39 had dedicated transit lanes, nor did they have to face the political challenges of either removing bus lanes (a really, really bad look) or turning the street one-way (much bigger issue than removing parking).

My point is more that there is transit demand. Despite being slow they have still attracted massive ridership.

That's a distinct problem from taking away existing bike lanes and putting bikes back to car traffic.

I think you're really overstating the quality of the existing bike "Infrastucture" (With masssive air-quotes) on Centre St and South St. Of the ~6700ft route between Arborway and Hyde Square, there are no protected bike lanes, about 2000ft has narrow bike lanes in both directions, and another ~500ft has a bike lane in one direction only. It should also be stated that these bike lanes are, not coincidentally, often located on the wider areas of the route. At least 600-700ft of the existing two way bike lanes could be preserved even with transit lanes, and if it was desired much of the rest could be converted to one direction.

As a side note, it's completely unscientific but I think it's at least somewhat telling that while going through on Street View, I saw a grand total of 2 cyclists (One of whom was not using the bike lanes), one parked bike, and a Bluebikes dock more than 75% full. I'm not convinced this is a major cycling route with a huge volume of cyclists just waiting to protest the removal of their barely existent lanes.

For the Columbus Avenue corridor with the center running bus lanes, why would we even consider bike lanes? There are bike paths in the Southwest Corridor Park running adjacent to Columbus.

I would claim that this argument actually works even better with Centre/South Streets. Not every street needs to have bikes, especially if you want a vibrant pedestrian-focused area.

 

[JeffDowntown]

I would claim that this argument actually works even better with Centre/South Streets. Not every street needs to have bikes, especially if you want a vibrant pedestrian-focused area.

The Southwest Corridor infrastructure can service the Centre/South Streets corridor. But to make that argument, you need to improve the safety of the major connecting streets for bikes. They are barely safe to drive!