Redesign the Urban Ring

[belmont square]

I like the extension of the Mattapan line to Roslindale

Just out of curiosity, why? The MBTA's Route 30 bus, which currently links the two neighborhoods, carries only 2,000 passengers per day on a route serving relatively congestion free streets. Would a 4 mile trolley ride to Ashmont and the Red Line become the preferred path of travel to downtown for Roslindale residents over the current 1 mile bus ride to Forest Hills and Orange Line? Or do you think there is some untapped demand in travel between the two neighborhoods that would be willing to ride a trolley but not existing bus service?

 

[AmericanFolkLegend]

We could tear down all the tripple deckers, two and single family homes and start building some high density housing. (It's a fanasty world after all)

Triple deckers are quite dense.

 

[Teban54]

As suggested by @TheRatmeister, we're reviving this thread for transit pitches related to Urban Ring, instead of dumping them in the general Crazy Transit Pitches (or occasionally the Green Line Reconfiguration thread).

To kickstart some discussions, I'll start by posting this map showing some southwest quadrant alignments, originally written as a reply to @BlazingZack :

Copying over some notes I said over there:

Almost any combination of lines here can give you an Urban Ring, and/or an alternative D-E connector that serves LMA more directly than the Huntington Ave stations.

Your choice depends on too many variables, which I detailed in the original comment.

If anything, it's Exhibit A of how much of the "divisiveness" really depends on your level of optimism. The "bore baby bore" Reddit folks will think very differently from someone like F-Line who questions just about any C&C subway with utility relocation, not to mention TBMs.

 

[TheRatmeister]

So just to restart this thread of urban ring thoughts, I figure a good place to start would be listing some generally agreed upon conclusions.

Essentially mandatory nodes:

• LMA
• Nubian

Highly desirable nodes:

• Somewhere in Cambridge, either Kendall, Central, or Harvard
• Ruggles or Roxbury Crossing
• Sullivan
• Logan Airport, preferably the terminals
• West Station
• Kenmore/Fenway Park

Lower priority, primarily residential areas:

• Southie
• Assembly
• Seaport
• Allston/Lower Allston
• Brighton
• Union Sq
• Charlestown
• Chelsea
• Revere
• Everett
• UMass/Columbia Point
• Wonderland
• Widett Circle
• Suffolk Downs
• Cambridgeport

Mode choice:
Either GL compatible light rail or a separate subway line, not BRT.

Method of construction:
Surface/elevated whenever possible, C&C tunnel when it's not, TBM for the toughest sections.

The rest of this will be some of my random personal opinions:

• The rebuilding of the Tobin Bridge in the near-but-not-too-near future is a good opportunity, and would provide an excellent link to/between areas currently underserved by transit (Charlestown/Chelsea)
• Focusing on areas with strong potential for redevelopment provides an opportunity to fund the scheme. This includes areas such as Andrew, Widett Circle, Lower Allston, Inner Belt, that polluted industrial hellscape part of Everett, and Wonderland.
• Good corridors for C&C subway construction include Dorchester St/E Broadway (The Southie One), Broadway (The Chelsea/Revere one), Broadway (The Everett one, c'mon lets be more creative), Melina Cass Blvd (Although Elevated might be even better here)

And now my spicy takes:

• Automated Light Metro (à la Copenhagen or Vancouver) is the best mode choice. Smaller trains with very low headways are well suited for a route where many people will take short journeys involving transfers, and can decrease construction cost with smaller stations.
• An APM connecting to the BL would be the better solution for Logan Airport, allowing UR service to serve large, dense areas in Everett, Chelsea and Revere while also facilitating redevelopment of Wonderland and the Everett Hellscape™.
• While Kendall is a better node than Harvard, serving Kendall would likely mean sacrificing service to some combination of Lower Allston/Harvard/Central/Union Sq, which, IMO, tips the balance away from Kendall and towards Harvard or Central.
• The GJ would be better suited to a more local service, such as a light rail line/busway with frequent stops connecting to West Station/Sullivan. Urban ring service should find another ROW.
• The urban ring doesn't have to be a ring, a U-shape could be better.

So, with that said, here's my preferred Urban Ring:

 

[Riverside]

To bring over a few pieces that I contributed in the other thread:

Here is a broader overview of my take on [the question of getting from BU Bridge to Longwood], with the obligatory Pretty Map:

...

One question I think that would impact this: what kind of service are you running south of Fenway station? Is it surface-running in a dedicated ROW? In transit lanes? Is it grade separated? How far does it run -- LMA, Ruggles, Nubian? I think those would inform what kinds of investments and tradeoffs are reasonable between BU and Fenway.

To the points that @Teban54 raised in the previous thread:

Almost any combination of lines here can give you an Urban Ring, and/or an alternative D-E connector that serves LMA more directly than the Huntington Ave stations.

Your choice depends on too many variables:

• a) How fast do you want trains to traverse through this quadrant?
  • At-grade LRT with mixed street running (most likely on Park Dr and Mountfort St)?
  • At-grade LRT with all 13 grade crossings, but with dedicated lanes only?
  • Grade separate as much as possible, with maybe 1-2 grade crossings?
  • Fully grade separated?

On this point, I've favored going with two answers instead of one. I argue that a service coming from Kendall + Sullivan (if not farther) needs to prioritize being fast (which is why I send it down the Fenway Branch to Brookline Village), while there is a niche for a second service that runs more slowly over a shorter distance. My Transportation Dreams map depicted this as a Harvard <> Ruggles service, but I think you could also do a Kenmore <> Ruggles service.

• b) Related to (a): How much demand is there for through-LMA rides and to-LMA rides from the north? This determines how many people benefit from a faster ride.
  • Example O-D pairs: Cambridge-LMA, OL North-LMA, Riverside-Kendall/Harvard, OL South-Kendall/Harvard, Nubian-Kendall/Harvard, many non-commute trips and trips involving B branch

My general theory here is that farther destinations (e.g. Orange Line North) will be biased toward to-LMA journeys, while closer destinations such as Harvard and Cambridgeport will be more biased toward through-LMA journeys. I think the "squished" nature of the Urban Ring corridor also impacts this, as does the continued high-speed of the Red Line -- no Orange Line rider is going to transfer at Ruggles for Harvard, in my opinion.

• c) How much detour do you want to make to the southwest at LMA?
  • Which east-west road through LMA to use: Fenway-the-road, Longwood Ave, or Francis St?
  • Related to (b) in terms of how many riders will need to wind through the LMA detour
  • Also related to cost (financial and political), as Longwood Ave and Francis St will be more expensive
  • Note that street-running on Fenway-the-road may result in speed restrictions that offset advantages in travel time

In my opinion, an infill at Aspinwall ("Kent Square") on the Fenway Branch provides pretty good coverage to the southwestern LMA.

• d) How much detour do you want to make to the northeast at Kenmore?
  • Several alignments too: Park-Mountfort, D-Pike, and "Kenmore/Lansdowne" (D-Lansdowne-Brookline Ave-B)
  • Again, see earlier note on travel times for street-running sections
• e) Related to (d): How important do you think Kenmore is?
  • Compare through-Kenmore and to-Kenmore demands
  • Hitting every GL branch may be as good as hitting Kenmore? IMO, its function as a bus transfer hub is overrated: the only Kenmore bus that doesn't go to LMA is the 57
  • Kenmore/Lansdowne does offer transfer to commuter rail, but some designs allow you to do it at West Station
  • 2015-17 passenger survey data: 42% of Kenmore riders start their trip from there, 27% transfer from buses, 31% transfer from another GL branch
    • 2022 passenger survey data is available, but sadly it drops the "Previous Mode" field and the raw number of passengers (only shows percentages), so not nearly as useful for this purpose

I don't see Kenmore itself as a particular priority, though see (h) below.

• (f) How important do you think getting to the core of LMA is?
  • Current D and E stations require average walk of 10 minutes to most hospitals, while a more central station can do much better
    • @Riverside has a more detailed table that he may want to post?
  • But LMA has higher transit mode share than most other employment centers, suggesting the 10-min walk may not be deterring too many people from using transit?

Yup, I need to update that table and post it. Relative to expense, I'm not convinced a core LMA station is necessarily worth it.

• (g) How important is a convenient transfer to the B branch?

I would give this some weight, in part because we're also talking about a convenient transfer to the 57 bus; between those two, that's a lot of Allston Brighton that could have an easy 2SR to LMA. (There's also a related question about providing a link between West Station and LMA.)

• (h) Do you even want a continuous service at all?
  • One alternative is to simply terminate two legs of LRT services both at Brookline Village, avoiding the engineering challenges of cross-LMA altogether
  • Another is to have the northern half be LRT that ends at Brookline Village, and the southern half be BRT that goes through LMA surface roads, ending at Kenmore
  • Splitting the two services at Kenmore is also an option (if you can somehow short-turn Comm Ave trains)
  • Whether that's actually desirable depends on (b) and (e)

I think continuous service can be a good long term goal, but that there are plenty of discontinuous alternatives that are very viable. I'll also point out that the "squished" nature of the Urban Ring corridor means that journeys around the Ring through LMA are going to be less competitive against, e.g. the 1 bus, depending on your destination. So, there are some limits to the benefit of a continuous route.

In particular, I think that a "first gen build" that terminates multiple services at Kenmore could be a useful way to break this project into smaller chunks.

• (i) Of course: What's your budget?

Well that's the question of course. I think it might be more feasible and/or productive to frame this question as "where would these projects fall in your priority list?". Where do they fall in my priority list? Good question, stay tuned.

 

[Teban54]

This comment is inspired by the very first post in this thread by Van, and is a continuation of my comment in the Green Line Reconfiguration (GLR) thread. I decided to move here once I realized what I'm about to say have significantly more to do with Urban Ring (UR) and less to do with GLR. While my previous comment in the GLR thread was largely feedback on Van's and Riverside's points, this comment ended up incorporating much more of my personal opinions.

High-level philosophy of an Urban Ring

While Van's comment is technically more about GLR than Urban Ring, because it was branded as an Urban Ring proposal, a lot of the underlying motivations are more applicable to UR than GLR. In particular:

The purpose of this is two fold; the first being to take pressure off of the 4 major downtown transfer points (Park, State, Gov't Center, and DTX), and second to connect large job centers, primarily Kendall Sq and the Longwood Medical Area.

These are points I really want to stress, especially the first point about reducing transfers downtown. The same argument was conveyed by @JeffDowntown much more recently:

Simplistic reply on the goal of an Urban Ring, but it really is about total network capacity and efficiency. And a big barrier today is the transfer capacity at the four main downtown transfer stations.

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

As an example, a rider needs to see it as logical to go from Northeastern to Harvard by way of the Urban Ring rather than via Park or Downtown Crossing downtown transfers.

Therefore, I think the #1 goal of an Urban Ring is for circumferential trips to be faster than, or at least equally as fast as, trips that use a downtown transfer.

What about the other objectives?

Two of the most commonly considered objectives are:

• #2: Enhance connectivity to key nodes outside of downtown core (employment centers, recreational centers, etc)
• #3: Fill (mostly residential) transit deserts between existing lines

My opinion: Both of them should take lower priority than #1 when in direct conflict; However, they feed into #1 by virtue of being part of what makes a trip faster, especially #2. This is primarily due to the difference between two-seat rides (2SRs) and three-seat rides (3SRs).

Spoiler: Example of how #2 can make trips faster (plus other nuances)

Suppose I live in Malden and work at Harvard, but the Urban Ring route goes from Sullivan to Central. My choices are:

• (a) Malden - DTX - Harvard
• (b) Malden - Sullivan - Central - Harvard

Even though (b) covers a shorter geographical distance, it requires an additional transfer. This can already discourage some riders prioritizing comfort from using it, but more tangibly, it incurs a time penalty due to transfers (going through transfer walkways and waiting for a train).

A possible solution is: Make the Urban Ring go directly to Harvard! Now (b) becomes just a 2SR (Malden - Sullivan - Harvard), while still having a shorter distance. As long as the Urban Ring travels fast enough and is not too roundabout -- which itself is not a guarantee -- I will almost always choose (b). This frees up space at Downtown Crossing for other riders working downtown.

The problem is, Harvard is not the only job destination. Suppose my two roommates Mary and Tom both work in Kendall. (Commuters to Kendall are similar in number to (Central+Harvard) combined, and twice as much as each of them individually.) Now their choices are:

• (c) Malden - DTX - Kendall
• (d) Malden - Sullivan - Harvard - Kendall

(d) obviously loses: it covers a longer distance, meaning it would still have been slower even if transfers were instant. This is a problem that I call "backtracking". A similar problem exists on the residential end: even if the Urban Ring went to Malden, now my friend Bob who drives to Wellington needs to backtrack, and for my friend Christina who lives in Melrose, in a future with OLX to Reading, Urban Ring will not turn her trip from a 3SR into a 2SR.

I propose several high-level principles, many of which are illustrated by this thought exercise above:

• Connecting an Urban Ring directly to key nodes outside downtown (#2) is a great way to cut down travel time (#1), by turning 3SRs into 2SRs. Not everyone will get these "lucky" 2SRs, so it's best to maximize the number of people that can get them.
  • This also applies to bus hubs. Allowing people to hop from a bus to the ring route directly, avoiding the radial route at all, is very useful.
• There's value for a ring that's not too far from downtown. Anyone inside the ring will experience backtracking, which significantly degrades the experience, and their effects are much greater than anyone outside the ring. Thus, the "outer branches" need to carry enough ridership before a ring can be placed. (This applies to both origins/neighborhoods and destinations/jobs.)
• The ring needs to travel fast enough. This can mostly be achieved in two ways: Taking a more direct route between points; and achieving a higher speed, by having fewer stations, grade separation, etc.
• The ring needs to have convenient transfers to radial routes. Inconvenient ones will increase travel time and passenger experience; or worse, intersecting a radial route without a transfer station completely undermines the ring's functionality.
• In some cases, the ring itself can serve as a demand generator for future TOD. (Here's an example from Madrid.) However, how much this applies to Boston (and US cities in general) is another question.

In fact, several successful ring routes serve CBDs, and can often be a pseudo-radial line for these key nodes: Seoul's Line 2 (another explanation), Beijing's Line 10, Singapore's Circle Line, and Tokyo's legendary Yamanote Line.

However, these principles then introduce several tradeoffs:

• Ring radius: Do you place your ring closer to downtown to hit key nodes, gather higher ridership on outer branches, and reduce cost? Or do you place your ring further away from downtown, for greater distance savings (and possibly higher TOD potential)?
  • When two radial lines form an acute angle, the bigger your circle is, the shorter the distance. However, this also makes more people need to backtrack.
• Route choice: Do you take a more roundabout route to hit a key node X? Or do you take a straighter route that bypasses them, to help other passengers traveling on the same quadrant but not to X?
  • I often call this "to-X" vs. "through-X". The answer depends on the comparative demand for these two patterns, which in turns is depends on prominence of X itself.
  • This also manifests in where you place the station in X: a central location or a peripheral one?
• Stop spacing: Do you make more stops in X, to reduce travel times for the to-X crowd to reach the stations? Or do you fly through X with fewer stops, so that other passengers get a faster trip?
  • Again, this often depends on to-X vs. through-X.
• Target for competition, and "close calls": Sometimes, two radial routes form a ~120° angle, such that a circle can't achieve much distance savings. Does your ring try to cater to this crowd and achieve at least equal travel time, such that at least some riders will take it, freeing up some capacity downtown? Or do you just give them up and let them transfer downtown, in order to serve other riders (between two other radial routes) better?
  • Practical scenarios in Boston: OL North - LMA; OL South and Huntington - Kendall; Nubian - Seaport - Chelsea
• Perhaps the most important -- Budget: Do you go for an expensive build with a more ideal route and grade separation? Or do you settle for a cheaper route with drawbacks, so that you at least have something to ride in your lifetime?
  • This affects all tradeoffs above.

(I largely avoided the question of how "key nodes" should be defined: Employment centers? Recreational centers? Neighborhoods? Right now, I lean towards the first option, but not extremely strongly. I welcome attempts at changing my mind, which include future analyses I plan to do myself to answer this question.)

 

[Teban54]

Extended readings on circumferential lines
Here are some articles I found with a quick Google query (lol). I've read some of them, but not all, so I'll read the rest together with you guys.

Blogposts

1. Moscow: Questioning the Circle Line (from Human Transit blog)
2. Mixing Circumferential and Radial Transit (from Alon Levy's Pedestrian Observations blog)
3. Mixing Circumferential and Radial Transit in the Other Direction (also from Alon Levy)

Even though #2 and #3 do not directly address purely circumferential lines, all three posts (and the comments!!!) provide very valuable insights. This particular comment under #1 elegantly explains why connecting to key nodes is important.

Academic papers
On Boston's Urban Ring in particular:

• "Analyzing the role of circumferential transit in relieving core congestion" (2002): I can't access this paper, but from the abstract, it specifically "explore the role and ability of circumferential transit to relieve core congestion in the central subway system of Boston", but also consider other systems.
• "Coming Around" (2003): This one does have open access, and is a very detailed paper that discusses MBTA's Urban Ring proposal from that era. But I haven't read the paper in detail to see if they're explaining the T's plans or discussing/proposing their own.

On non-Boston systems in general:

• "Long-term planning for ring-radial urban rail transit networks" (2016): Seems to be more of a mathematical model with simulations, which is more formal than needed for this audience, but I guess the high-level intuitions (which I have not read yet) may be helpful

 

[TheRatmeister]

I wanted to come back to this thread with a discussion of the Copenhagen Metro and why I think it's extremely relevant to an urban ring service.

Overview
The Copenhagen Metro is a relatively small system. It has 4 lines, with 2 radial lines, 1 loop line, and 1 radial line that runs along part of the loop. (There is also the S-Train network which is cool but I'm going to skip over that for right now.)

The network features a combination of underground, elevated, and at-grade stations, although the majority of the system is underground.

The Trains
This is where it starts to get interesting. The Metro uses the Hitachi Rail Italy Driverless Metro system, which means the trains are small, only 128ft long. (For reference, Red Line trains are around 420ft long.) And yet, the M3 of the Copenhagen metro can carry more passengers daily than the Red Line did in 2019. It does this through very high frequencies. M3 and M4 trains come every 2-3 minutes, and so headways on the combined section can be as low as 80-90 seconds during peak times.

The Stations
When you use the metro one of the first things you notice is that all the (underground) stations are very similar. They are (almost) all built in a 20m by 60m box, which makes the construction process fairly straightforward. Anywhere you can dig a 20x60m hole you can build a station, which means that you can easily site a station under just about any moderately wide street. They have no fare gates, operating on a proof-of-payment system, and are about as simple as you could build them.

The Money
The Copenhagen Metro is fundamentally tied to real-estate development. M1 and M2 stimulated huge new developments on the island of Amager, while the dirt excavated from the M3 tunnels was used to make new land in the harbor to be built upon and developed. M4 is designed to connect these new areas in the north as well as more areas currently under redevelopment in the south. Proceeds from land sales and development is how these projects have been funded.

Why is this relevant?
So why is this related to the Urban Ring? The urban ring is a service designed to make connections with other lines, and be used for many short journeys rather than a few end-to-end journeys. This high turnover rate and high number of journeys involve a connection mean that a system with smaller trains is not a problem, since there is not one busy section where the trains fill up and slowly empty out, and the high frequencies make connections shorter by reducing the amount of time spent waiting. The small stations in a 20x60m hole design makes stations easier as they can be sited under small, relatively narrow strips of land like on the SW corridor next to Ruggles, at Bellingham Sq in Chelsea, or under the Oscar Tugo Circle at LMA. and the funding model is also extremely relevant, as a design like the one I sketched up previously links a huge number of areas with massive sites awaiting development. Wonderland, Bell Circle, the Everett Hellscape™, Sullivan, Community College, the Inner Belt, Lower Allston/West Station, Widett Circle, and Andrew are all stations adjacent to huge sites that are either actively being redeveloped or are open to redevelopment that could be used to (At least partially) fund the project.

So how much complexity could the "Small station box" eliminate? Of the stations on my fantasy diagram, only the following could likely not be built using the station box method:

• MassArt/Wentworth if the extended Huntington Ave subway is built first
• Harvard
• Union Sq
• At-grade would likely be preferred for Community College, Sullivan, and Assembly
• Charlestown, Admirals Hill (Elevated as part of a Tobin Bridge replacement)
• Cheslea Creek, a stacked design would likely be required here

All the other stations are located on streets or adjacent to open spaces where a 20x60m space could be excavated.

 

[JeffDowntown]

I wanted to come back to this thread with a discussion of the Copenhagen Metro and why I think it's extremely relevant to an urban ring service.

Overview
The Copenhagen Metro is a relatively small system. It has 4 lines, with 2 radial lines, 1 loop line, and 1 radial line that runs along part of the loop. (There is also the S-Train network which is cool but I'm going to skip over that for right now.)
View attachment 49609
The network features a combination of underground, elevated, and at-grade stations, although the majority of the system is underground.

The Trains
This is where it starts to get interesting. The Metro uses the Hitachi Rail Italy Driverless Metro system, which means the trains are small, only 128ft long. (For reference, Red Line trains are around 420ft long.) And yet, the M3 of the Copenhagen metro can carry more passengers daily than the Red Line did in 2019. It does this through very high frequencies. M3 and M4 trains come every 2-3 minutes, and so headways on the combined section can be as low as 80-90 seconds during peak times.
View attachment 49610
The Stations
When you use the metro one of the first things you notice is that all the (underground) stations are very similar. They are (almost) all built in a 20m by 60m box, which makes the construction process fairly straightforward. Anywhere you can dig a 20x60m hole you can build a station, which means that you can easily site a station under just about any moderately wide street. They have no fare gates, operating on a proof-of-payment system, and are about as simple as you could build them.
View attachment 49611
The Money
The Copenhagen Metro is fundamentally tied to real-estate development. M1 and M2 stimulated huge new developments on the island of Amager, while the dirt excavated from the M3 tunnels was used to make new land in the harbor to be built upon and developed. M4 is designed to connect these new areas in the north as well as more areas currently under redevelopment in the south. Proceeds from land sales and development is how these projects have been funded.

Why is this relevant?
So why is this related to the Urban Ring? The urban ring is a service designed to make connections with other lines, and be used for many short journeys rather than a few end-to-end journeys. This high turnover rate and high number of journeys involve a connection mean that a system with smaller trains is not a problem, since there is not one busy section where the trains fill up and slowly empty out, and the high frequencies make connections shorter by reducing the amount of time spent waiting. The small stations in a 20x60m hole design makes stations easier as they can be sited under small, relatively narrow strips of land like on the SW corridor next to Ruggles, at Bellingham Sq in Chelsea, or under the Oscar Tugo Circle at LMA. and the funding model is also extremely relevant, as a design like the one I sketched up previously links a huge number of areas with massive sites awaiting development. Wonderland, Bell Circle, the Everett Hellscape™, Sullivan, Community College, the Inner Belt, Lower Allston/West Station, Widett Circle, and Andrew are all stations adjacent to huge sites that are either actively being redeveloped or are open to redevelopment that could be used to (At least partially) fund the project.

View attachment 49614
So how much complexity could the "Small station box" eliminate? Of the stations on my fantasy diagram, only the following could likely not be built using the station box method:

• MassArt/Wentworth if the extended Huntington Ave subway is built first
• Harvard
• Union Sq
• At-grade would likely be preferred for Community College, Sullivan, and Assembly
• Charlestown, Admirals Hill (Elevated as part of a Tobin Bridge replacement)
• Cheslea Creek, a stacked design would likely be required here

All the other stations are located on streets or adjacent to open spaces where a 20x60m space could be excavated.

One big drawback of the Copenhagen station design is only one elevator per station -- and even there they are frequently out of service.

 

[TheRatmeister]

One big drawback of the Copenhagen station design is only one elevator per station -- and even there they are frequently out of service.

I believe this is only true for the older M1/M2 stations, the M3 Cityringen stations generally have 2 parallel elevators at the end of the platform to provide redundancy. Regardless of how common it is, it's definitely possible to have redundant elevators with the Copenhagen station design.

 

[TheRatmeister]

I wanted to come back to this thread with a discussion of the Copenhagen Metro and why I think it's extremely relevant to an urban ring service.

Overview
The Copenhagen Metro is a relatively small system. It has 4 lines, with 2 radial lines, 1 loop line, and 1 radial line that runs along part of the loop. (There is also the S-Train network which is cool but I'm going to skip over that for right now.)
View attachment 49609
The network features a combination of underground, elevated, and at-grade stations, although the majority of the system is underground.

The Trains
This is where it starts to get interesting. The Metro uses the Hitachi Rail Italy Driverless Metro system, which means the trains are small, only 128ft long. (For reference, Red Line trains are around 420ft long.) And yet, the M3 of the Copenhagen metro can carry more passengers daily than the Red Line did in 2019. It does this through very high frequencies. M3 and M4 trains come every 2-3 minutes, and so headways on the combined section can be as low as 80-90 seconds during peak times.
View attachment 49610
The Stations
When you use the metro one of the first things you notice is that all the (underground) stations are very similar. They are (almost) all built in a 20m by 60m box, which makes the construction process fairly straightforward. Anywhere you can dig a 20x60m hole you can build a station, which means that you can easily site a station under just about any moderately wide street. They have no fare gates, operating on a proof-of-payment system, and are about as simple as you could build them.
View attachment 49611
The Money
The Copenhagen Metro is fundamentally tied to real-estate development. M1 and M2 stimulated huge new developments on the island of Amager, while the dirt excavated from the M3 tunnels was used to make new land in the harbor to be built upon and developed. M4 is designed to connect these new areas in the north as well as more areas currently under redevelopment in the south. Proceeds from land sales and development is how these projects have been funded.

Why is this relevant?
So why is this related to the Urban Ring? The urban ring is a service designed to make connections with other lines, and be used for many short journeys rather than a few end-to-end journeys. This high turnover rate and high number of journeys involve a connection mean that a system with smaller trains is not a problem, since there is not one busy section where the trains fill up and slowly empty out, and the high frequencies make connections shorter by reducing the amount of time spent waiting. The small stations in a 20x60m hole design makes stations easier as they can be sited under small, relatively narrow strips of land like on the SW corridor next to Ruggles, at Bellingham Sq in Chelsea, or under the Oscar Tugo Circle at LMA. and the funding model is also extremely relevant, as a design like the one I sketched up previously links a huge number of areas with massive sites awaiting development. Wonderland, Bell Circle, the Everett Hellscape™, Sullivan, Community College, the Inner Belt, Lower Allston/West Station, Widett Circle, and Andrew are all stations adjacent to huge sites that are either actively being redeveloped or are open to redevelopment that could be used to (At least partially) fund the project.

View attachment 49614
So how much complexity could the "Small station box" eliminate? Of the stations on my fantasy diagram, only the following could likely not be built using the station box method:

• MassArt/Wentworth if the extended Huntington Ave subway is built first
• Harvard
• Union Sq
• At-grade would likely be preferred for Community College, Sullivan, and Assembly
• Charlestown, Admirals Hill (Elevated as part of a Tobin Bridge replacement)
• Cheslea Creek, a stacked design would likely be required here

All the other stations are located on streets or adjacent to open spaces where a 20x60m space could be excavated.

I figured I'd elaborate on what alternative sites could or likely couldn't also work well with the "small box" design:

• Kenmore is probably a no due to the GL tunnels
• A combined Fenway/St Marys could work. The walk to the latter would be a bit on the longer side but not too bad.
• BU could work quite well, Comm Ave is wide enough that you could potentially not even need to touch the B branch tracks
• The Grand Junction could work fairly well. A Waverly St stop could be at grade, Pacific St appears to have just enough room, Mass Ave has a big parking lot, Gallileo Gallilei has the GJ park and is also just really wide by itself, and there's a wide open space behind Memorial Way if a station is desired there. (This then gets into the Harvard vs Kendall debate but that can be saved for another time.)
• Seaport could get difficult with all the tunnels already there, but you might be able to find a big enough space.
• The Chelsea Greenway/SL3 busway is generally wide enough, although the location isn't optimal.
• The only place you could make this method work at Logan appears to be where the Terminal E parking is currently, but that's not very central. Better than the current BL stop though.
• There is no shortage of space at Newmarket
• Maybe you could get something under 93 at JFK/UMass? Failing that you could dig up part of the busway

 

[JeffDowntown]

I believe this is only true for the older M1/M2 stations, the M3 Cityringen stations generally have 2 parallel elevators at the end of the platform to provide redundancy. Regardless of how common it is, it's definitely possible to have redundant elevators with the Copenhagen station design.

Agreed. More just pointing out that you can get too simple in the station design. There is a limit to how cheap you can go and still maintain functionality. But even a redundant elevator version is still a very efficient design (which we would find some way to muck up, I am sure).

The problem with a standardized design is it take out a lot of the opportunity for Project Consultants managing the Design Consultants, managing the Construction Consultants to pad their bills.

 

[samsongam]

Hey everyone,
I’ve been following your conversations for the past few months and got super excited about a bunch of the ideas shared on the forum. I went ahead and tried to map out a somewhat realistic route for a continuous urban ring with the lowest barrier to entry.

The route is inspired by bus priority expansions in the city such as the Columbus Ave bus lanes, dedicated busway for the SL3 through Chelsea, upcoming BHA bus lanes, and upcoming SL6 for Everett and onto Sullivan Square. It seems like there is more of an appetite for the more affordable bus infrastructure upgrades rather than undertaking large metro/subway expansions or major tunneling in the near future.

My thinking is that by creating these (mostly) center running bus corridors that we could expand incrementally in the future with dedicated transitways (https://www.google.com/maps/d/u/0/edit?mid=1_yyYlVQgwanvpk79VS0kCkcaYyuVcgo&usp=sharing) that would run trolleys/LRT (maybe type 10’s? Not sure) as well as buses. Priority would be for LRT but would also ensure a separated route for several high ridership bus lines that may remain necessary. There are similar corridors in Melbourne, San Francisco, Hong Kong, Portland, Utrecht.

A couple of informing opinions:

1. Southie - This would provide the whole of southie with access to faster transit. I’m not all that familiar with South Boston, but this route pretty much follows the 10 bus.

1. Southie to Ruggles - Bus lanes on Melnea Cass have been part of the discussions on its upcoming reconstruction, but it’s hard for me to find info on what’s going on with that since previous plans were canceled.It seems like a great opportunity to connect Southie to Roxbury and Ruggles. I have a couple different options here for reaching Nubian which would be a high priority.

1. Ruggles to Grand Junction - As discussed previously, space is hardest to come by between Ruggles and the Grand Junction. I support the idea of a dedicated transitway from Louis Prang and Fenway all the way to Fenway Station as proposed by @Riverside and @Teban54 above. Park Drive to Mountfort Street/St.Mary’s is a big challenge without mixed traffic. I like the idea of extending Mountfort over the commuter rail before St. Mary’s St. and then over the area in brown. This whole thing is an investment so it’s a bit yikes but seems like it would be more affordable than going to Kenmore or having to build more tunnel
   
   
   
   
   
   
   
   

1. Crossing the Charles - The most recent plans from the will cross over the turnpike and then under the bu bridge. A significant amount of landwork will be down here, doesn’t seem like a commuter would be helpful here being so close to west station but would this make sense as a stop to connect to GL B Branch? It’s on a curve but there’s some space and work being planned.The bridge crossing would have to be expanded. A connection to West Station (potentially through to Harvard?) would be great but it’s really difficult to see with how squished it will all get after grounding the pike.

1. Grand Junction - This ROW would do the majority of heavy lifting through Cambridge as in the Is there space on the Grand Junction ROW for both a commuter rail line and 2 bus lanes throughout? Assuming West Station revives rail use here like the plans say, how much space would we have to use this corridor for anything other than heavy rail? It seems especially tight in East Cambridge

1. Crossing the Inner Belt - The easier way to go here is via the Inner Belt, someone posted a map on the tracks that would make that connection possible similar to the lighter yellow option I have ( sorry I wasn’t able to find the exact post). However since we’re thinking BRT first, this seems pretty impossible to run through the inner belt. My alternative would be to go under or over the Fitchburg and GL tracks and connect into Mcgrath Highway which is going to be grounded. While a busway is unlikely at this point in time due to low transit demand, this project would certainly change that assumption and the design of the project is ongoing.This route also gives us an easier connection to the green line in this area at East Somerville Station on the Green Line.

1. Sullivan to Chelsea - This section pretty much just follows the SL6 preferred route with side running bidirectional bus lanes through the Encore and Soccer Stadium (fingers-crossed) sites then into Everett with bus lanes where space becomes tight again. I have a potential bridge over Sweester circle in case of light rail. Crossing the rail tracks from 2nd street are another challenge in the future for LRT. We’re onto the Chelsea busway from here and relatively smooth sailing.
   
2. Chelsea to Airport or Wonderland - Here we run into the Chelsea Creek issue like @TheRatmeister mentioned to support their routing up to Wonderland. While I like this connection for Revere stop and avoiding the costs of crossing, I feel like it doesn’t really make for a helpful connection compared to Airport. BLX to Lynn brings in new riders but unless riders want to go Lynn -> Everett or Chelsea it doesn’t beat a downtown transfer. My more pricey crazy fantasy solution to crossing the river is a tunnel under the creek into the ROW connecting to Martin A. Coughlin Bypass Rd which would be a dedicated transit way connecting to Airport station on the west side of the East Boston greenway.

What the line is good at -

1. Reducing downtown transfers for movements between Cambridgeport, Kendall, MIT <-> BU, LMA, Northeastern, Roxbury
2. Greatly expanding transit access and connections for South Boston, Roxbury, Everett, Chelsea, Revere
3. Building on current infrastructure projects in planning over the next few years

What this line is bad at -

1. Being Fast - Many Many Grade Crossings, could be greatly improved with signal priority. But mixing different bus routes on the same transitway also brings challenges.
2. Misses some big nodes like Harvard, Kenmore

Anyways, I wanted to share these ideas I have been picking up from browsing through the forum and wanted to get your informed opinions on why this is or isn’t a good idea.

 

[TheRatmeister]

Hey everyone,
I’ve been following your conversations for the past few months and got super excited about a bunch of the ideas shared on the forum. I went ahead and tried to map out a somewhat realistic route for a continuous urban ring with the lowest barrier to entry.

Essentially my thoughts boil down to: If we want to build a mostly surface LRT Urban Ring this is a very good way to do it, but is that what we should do? My main argument against is that a huge chunk of this route would be build on dedicated ROWs like the Grand Junction, SL3 Busway, and the Newburyport/Rockport line, as well as places where elevated/at-grade running could and should be considered, such as along Melnea Cass or through the Everett Hellscape™ (And future redevelopment on the site). There are also segments where it's generally agreed that a tunnel is probably necessary, such as through Longwood and likely between Nubian/Ruggles to avoid narrow streets and a roundabout route. So, given that, on the core section for this route you're essentially trading 2 miles of tunnel under Southie, .65 mi between Longwood and BU, and a shallow cutting under part of the GJ for reduced reliability, longer headways, and longer travel times. An advantage of this route at least is that it doesn't rely on interconnecting with the GL and has relatively little median-running, so high floor trains could potentially be used to at least address most of the capacity constraints of LRT, so there is that I suppose.

If you accept the higher upfront costs of something like a fully grade separated Hitachi Light Metro system, you'll see major benefits in speed and reliability from full grade separation, extremely low headways (Possibly as low as 90 seconds during peak periods) which is really useful when a huge chunk of the line's ridership is coming from transfers, and reduced operating costs from full automation.

One more thought on Airport before I leave: Would an Urban Ring actually meaningfully reduce travel times to the Airport (In a post Red Blue Connector and Logan APM world)? I don't think it would. Sullivan-Airport would take about 17-18 minutes on the UR assuming 20 mph average speeds, compared to around 25 minutes with an APM. GL is similar with Inner Belt/Airport travel times being about 20 minutes vs the hypothetical ~27 minutes with an APM, and RL would essentially be a wash from Kendall. So there is about 7 minutes of savings on trips from OL North and GLX to the airport, but that's really it. RL would see no improvement, BL would obviously see no improvement, and OL/GL would only be improved coming from the North, GL branches and the SW corridor would remain unchanged for travel times. Given the fairly substantial difficulty of this segment, is that really worth the cost?

 

[TheRatmeister]

Moving this because we're well off topic now.

Your Harvard/Inman route looks like it has stop spacing of about .66 miles, while your GJ route has spacings of .44 miles. The only scenario I've seen so far where the average (not maximum) speed hits 20+mph is when the stop spacing is at least 1 mile. For the conditions you are describing, something in the neighborhood of 17mph seems much more reasonable.

Fair point, ~17-18MPH is probably a more reasonable estimate. I don't think it really changes the conclusion though, it just puts UR (Harvard) at ~20 minutes compared to ~18 minutes for GLRC. With the terrible North Station transfer, extra minute or so on the CR train between CC and North Station, and more central location at Longwood it still ends up being a wash for the most part, and it would still come down to your specific final destination. People going to Boston Childrens would probably take the UR, people going to MassArt would probably take the GL. If anything this spreads the load going to LMA onto two services, which is probably not a bad thing to have.

 

[Riverside]

If anything this spreads the load going to LMA onto two services, which is probably not a bad thing to have.

In general, this is where I land -- for a number of reasons, we will probably see distributed load/journeys to Longwood, which isn't bad.

I still think the wash argument doesn't quite make sense. Assuming a starting point at Community College (though see below) and an ending point at one edge of LMA (e.g. either along Huntington or along Brookline Ave etc):

Option A: NSRL through running to Back Bay + transfer to GLR Gold Line (or transfer at South Station if your train doesn't continue to Back Bay) to LMA

Option B: Urban Ring from Community College to Harvard to LMA

Option A is about 4.8 miles, about 3.2 of which are at "mainline/RER speeds" and 1.6 are at "rapid transit speeds" (or 2.1 mainline + 2.7 rapid transit)

Option B is roughly 5.5 miles, all at "rapid transit" speeds. Whether that's 20 mph, 16 mph, or 12 mph, the rapid transit speeds are almost certainly going to be the slower of these two categories.

That seems like Option A will be notably shorter, and also will spend more time via a higher-speed service. Am I missing something?

(FWIW, "Option C," going via Grand Junction, would I think be most favorable using the above heuristics.)

As for Community College: I think basically all of the proposed designs call for portals north of Gilmore Bridge, and assume a direct gradient downward to pass under the Charles River and reach the needed depth for North Station Under. For example:

I'm not saying that an NSRL station at Community College is impossible, but it's definitely more complicated, and complicating.

That being said, building RER stations at East Somerville (potentially straightforward), Sullivan (more complicated), and somewhere on the Fitchburg Line (big question mark) would allow you to run an Urban Ring service either via Grand Junction or via your preferred Harvard alignment. So you could potentially get the same overall system design you are proposing, even if manifested somewhat differently.

 

[TheRatmeister]

As for Community College: I think basically all of the proposed designs call for portals north of Gilmore Bridge, and assume a direct gradient downward to pass under the Charles River and reach the needed depth for North Station Under. For example:

View attachment 51609

I'm not saying that an NSRL station at Community College is impossible, but it's definitely more complicated, and complicating.

That being said, building RER stations at East Somerville (potentially straightforward), Sullivan (more complicated), and somewhere on the Fitchburg Line (big question mark) would allow you to run an Urban Ring service either via Grand Junction or via your preferred Harvard alignment. So you could potentially get the same overall system design you are proposing, even if manifested somewhat differently.

You're definitely right about NSRL making the UR-RER transfer harder. If I were to propose an alternate option I think it would be one station at Sullivan for the Newburyport/Rockport Line and another station at Inner Belt, with the Lowell line re-reconfigured to run alongside GLX before joining the Fitchburg line.

Option A: NSRL through running to Back Bay + transfer to GLR Gold Line (or transfer at South Station if your train doesn't continue to Back Bay) to LMA

I think you're generally right about Back Bay being faster but there are a few complications:

• Not all RER trains will go to Back Bay and if you need to transfer that probably wipes out any time savings
• The downtown stops are going to have pretty high dwell times simply from the large passenger numbers which could slow the journey down enough to where it's competitive again
• The GL-RER/OL transfer at Back Bay probably won't be great and that will add a decent chunk of time as well. A transfer at Sullivan or Inner Belt would probably be quicker.

 

[Riverside]

another station at Inner Belt, with the Lowell line re-reconfigured to run alongside GLX before joining the Fitchburg line

Conceptually, I like this idea, as it's a very strong transfer node (and in general the lack of a station near here creates a gap). That said, I'm not sure how you would a) run the Lowell Line through the Brickbottom Junction flying interchange, and b) fit all the necessary platforms, without essentially rebuilding the whole junction. I haven't looked at it closely, but it definitely does not seem trivial.

Not all RER trains will go to Back Bay and if you need to transfer that probably wipes out any time savings

Right, that's why I mentioned the alternative of transferring to rapid transit at South Station. (Also worth noting that the transfer between a Fairmount-bound and Back Bay-bound train would be the easiest in the whole system -- cross-platform, very high frequencies [waiting 3 min max, 1.5 min on average], potentially even with the possibility of coordinated schedules enabling passengers to just walk across the platform into a waiting train. Obviously it's tough to increase a 2SR into a 3SR, but like I said I just think it's worth noting that this would be the least painful iteration of that possible.)

The downtown stops are going to have pretty high dwell times simply from the large passenger numbers which could slow the journey down enough to where it's competitive again

I think the dwell times would need to be pretty high for that to happen, though I haven't looked at the math closely.

The GL-RER/OL transfer at Back Bay probably won't be great and that will add a decent chunk of time as well. A transfer at Sullivan or Inner Belt would probably be quicker.

Respectfully, this is where you lose me. Gold <> Orange or Gold <> RER at Back Bay will be up-over-down, the same as Sullivan. The "over" distance at Back Bay will be a bit longer, it's true, due to the need to cross the Mass Pike -- something like 160-240 feet, depending which RER platform you're coming from. That's a shorter distance than the CR <> Red transfer at South Station today. At typical walking speeds, at most the 240 feet "over" section would take 55 seconds.

Broadly speaking, the Gold <> Orange or Gold <> RER transfer at Back Bay doesn't look that different from, say, the SB Orange -> SB Red transfer at DTX. It's not the cross-platform ideal described above, but I don't think it reaches the level of being "not great". And, in any case, either Option requires you to do a transfer somewhere; the difference between Back Bay's ~200 feet and Sullivan's ~130 feet (?) comes down to 16 seconds, which seems effectively equivalent for our purposes. (I do grant, though, that a South Station <> Gold transfer would have a more severe time penalty.)

(I definitely don't see how a transfer at Inner Belt could be better. Even if you rebuild the entire junction to put platforms as close to an idealized crosspoint here, I don't see a way to fit tangent platforms on the Lowell Line while still maintaining a reasonable curve radius for it to join the Fitchburg Line.)

In any case, there's a way to achieve your goal of an RER <> LMA journey that bypasses downtown altogether: Sullivan <> Grand Junction <> LMA (even with circuitous zig-zags at East Somerville and BU Bridge) beats out everyone at 4.2 miles. The Grand Junction ROW hews remarkably close to a straight line from Sullivan/East Somerville to LMA. If we're willing to dig an expensive grade-separated ROW from East Somerville to Harvard, one along the Grand Junction corridor wouldn't be much longer.