Crazy Transit Pitches

[Alon]

Continuing from the conversation in the Red/Blue thread I sketched out some possible Blue Line West routes...

First, as BussesAin'tTrains says, the map doesn't work; you need to make it public.

Second, there's a Swiss slogan, Elektronik vor Beton. Or, as some German activists who are pissed at Munich's S-Bahn second tunnel's cost overruns suggest, Organisation vor Elektronik vor Beton. First, fix the organizational problems of the commuter trains: the long turnaround times, the loose scheduling, the ticket punchers, the low frequency (caused by high operating costs coming from having multiple conductors). This requires strategic concrete pouring, like full-length high platforms at all stations to reduce schedule risks, but judging by Fairmount Line construction costs it's very cheap, a couple million per station. Then, fix the electronics - in this case, lack of electrification, and in some cases long signal blocks. That also allows multiple infill stops, taking a lot of pressure off of the B branch, and maybe even diverting outer D branch riders. Then - and only then - should the region discuss new subway tunnels.

With the Red-Blue connection it's different, because there's no organizational fix. Normal first-world cities would just spend $200-odd million on the connection and call it a day. But with anything involving capacity to the west, there's ample track capacity, the MBTA just wastes it by running commuter rail like it's 1957.

The Blue Line should terminate at Charles/MGH, and that's it. This isn't New York. The street network looks like a spaghetti bowl, the baseball team goes almost a century without winning, I can be in a room with 30 people and be the only immigrant, and the subway system isn't at capacity. There are only two serious capacity problems on the subway: the Green Line, and the Red Line on the Cambridge side. The Green Line can be relieved by commuter rail given operating practices that postdate the Mad Men era, and the Red Line can be relieved by upgrading the signaling so that it can run more than 15 trains per hour at the peak. New subways are useful for new connections: Red-Blue, Blue to Lynn, Urban Ring, GLXes, Tremont Street Subway, Orange to Medford, Red to Arlington, whatever. They are not required for new capacity, and as a result, there's no need for a new subway trunk line paralleling the Green Line.

 

[George_Apley]

^ Does the Worcester Line have the track capacity to handle a few infills in Allston/Brighton/Newton without adding passing tracks? I've figured the best thing to do with the Worcester is to have an inner xMU short turning at Riverside and hitting all the Newton stops, plus Newton Corner, Boston (Brighton) Landing and Allston (West). If it can be squeezed in maybe one around Parsons Street in Brighton (Faneuil). Run them at 18-20 minute headways. But I thought that there would need to be several passing sidings installed to allow Framingham/Worcester trains to fly through.

I worry about adding too many infills to suburban commuter rail lines at the expense of transit time. A Union Square addition to the Fitchburg, for example, doesn't make sense to me. The more expensive option for that stretch (but more beneficial to the whole system) would be continuing the GLX to Porter and connecting with Fitchburg trains there (and whatever DMU inner-Fitchburg line that should be short-turning at 128 on the Waltham/Weston line).

 

[vanshnookenraggen]

Made the map public so let me know if you can see it.

Alon: While I agree with you about commuter rail improvements that's not the issue I'm talking about. It's a thought experiment looking at what could be done with the Blue Line. This is "Crazy" transit pitches, not "Reasonable" transit pitches.

 

[citylover94]

I agree that the route that swings down the grand junction looks like a great idea. It also provides a subway stop in the southern part of central Square/ Kendall Square border area which is quickly adding offices so I actually could see that being a really effective route.

In response to Alon if NYC took the stance that no one lives in Queens why do they need a subway when they were building it then the NYC subway would be much less extensive today. The transportation provided in an area has the greatest effect on how it will develop over anything other than zoning therefore if we want to encourage development adding subway service while up-zoning near stations is a great way to encourage growth where it can be best handled.

 

[davem]

I think the cost of crossing the Charles would outweigh the benefit.

I propose an alternate solution: rerouting the B over the Grand Junction, and having the BL run to Harvard and Watertown/Waltham.

https://www.google.com/maps/d/edit?mid=z5nSIU1nZ3tI.kUpx_Y7wbDzA

The BL would run along the Charlesbank routing, but with only two stops, express the rest of the way. At Kenmore it would use the old B platforms and run under Comm Ave to the BU Bridge.

At BU a superstation would be constructed, allowing transfers between the Blue, the (B), and the commuter rail. The extended Blue line would take over "West" station and Boston Landing, speeding up CR trips.

This would also reposition Allston as an ideal housing solution for Kendall, while discouraging easy BU access. Downtown access is still preserved via North Station and a transfer to the Blue. It also alleviates congestion through the Central Subway, and positions heavy rail where needed (Allston and Watertown). The other benefit is that the GL trolleys will have an easier time navigating the at-grade intersections in Cambridge than a heavy rail train.

 

[FK4]

love it. totally expensive but totally worth it. would unlock watertown and waltham like nobody's business. both (waltham more so) are very much in need of better downtown access, and waltham in particular is poised to become one the next major waves of redevelopment.... but it's just a little too far, and hemmed in by walls of traffic in all directions, as it is right now. this would fix that, and everything else... now what would this whole thing cost (blue to moody and hvd sq, green along grand jnct)?

 

[Alon]

^ Does the Worcester Line have the track capacity to handle a few infills in Allston/Brighton/Newton without adding passing tracks?

Yes. Well, it depends on what frequency you want, but it's easy to accommodate with a 15-minute peak frequency. Given curves, the Worcester Line is optimistically 130 km/h territory, which means that the stop penalty for a FLIRT with level boarding is on the order of 1 minute. You need 2 minutes on each side to maintain adequate headways, which means that in a crunch, 11 local stations are possible before overtakes are required; for best comfort, you'd want to keep it to 10.

By the way: Zurich transit planners say that the best way to kill transit ridership is to run buses or trains every 11, 13, or 17 minutes. 18 is not much better. Acceptable headways are 10, 12, 15, and 20, repeating predictably all day, with infrequent (annual), well-publicized changes to the schedule.

I worry about adding too many infills to suburban commuter rail lines at the expense of transit time. A Union Square addition to the Fitchburg, for example, doesn't make sense to me.

Eh. First, the gains from better rolling stock are so large that losing a few minutes to extra stations will still be Pareto-better.

But second, sometimes, it's right to add a few minutes to long-range commute trips, as I discussed here. Most transit ridership is local; Union Square will always be a much more important node than South Acton.

Made the map public so let me know if you can see it.

Alon: While I agree with you about commuter rail improvements that's not the issue I'm talking about. It's a thought experiment looking at what could be done with the Blue Line. This is "Crazy" transit pitches, not "Reasonable" transit pitches.

Yes, I can!

As for crazy vs. reasonable: on the reasonable thread, I got sent to this thread for proposing railstituting the Green Line.

Anyway: if you're proposing something this outlandish, then think in terms of what should be built ahead of it. For example, I'll argue that the Grand Junction (as a ring with the NSRL, not as a reverse-branch of the Worcester Line) should come first, and this means the Blue Line extension should intersect it with a transfer point. Two points are reasonable: Kendall/MIT, with the Blue Line separated from the Red by a few blocks, and an East Cambridge stop located farther north, which would look awkward on a map.

Of course, this assumes it's worth the expense of setting up the Red-Blue connector for a Blue Line extension rather than for easy Red-Blue transfers, which it isn't.

In response to Alon if NYC took the stance that no one lives in Queens why do they need a subway when they were building it then the NYC subway would be much less extensive today. The transportation provided in an area has the greatest effect on how it will develop over anything other than zoning therefore if we want to encourage development adding subway service while up-zoning near stations is a great way to encourage growth where it can be best handled.

New York first planned a subway to Queens in the early 1910s. Between 1900 and 1910, Queens' population grew 85.6%, with tenements replacing farms in the first neighborhoods to urbanize. There was every reason to expect more urban growth: annual immigration was 1% of population, the US was not yet 50% urban but was rapidly urbanizing, there was a wide North-South wage gap so population growth was concentrated in the North, and the total fertility rate was still about 4. There was no zoning, at all, and zoning restrictions banning apartment buildings would be generations in the future.

Today, immigration is less than 0.5% (net of emigration it's 0.2%, but there was a lot of emigration in the 1900s too), the US is close to 100% urban, the South largely caught up with Northern incomes, and the total fertility rate for native-born Americans is 1.7. On the neighborhood level, Allston, Cambridge, and so on are already developed; redevelopment and infill are possible, but the only way you're doubling the population per decade is if you're plopping a few residential buildings in a commercial district, and that shouldn't really count. To say nothing of restrictive zoning in the most desirable areas.

 

[CSTH]

Continuing from the conversation in the Red/Blue thread I sketched out some possible Blue Line West routes.

Van, I like it. But for 'crazy' purposes it seems to me that if you're going to go across the Charles and that far west on the blue you might as well do it via north station.

As follow:

- From Bowdoin down Staniford to a new 'North Station' platform under Lomasney (a bit of a walk to GR and OR here is ok because of additional connections to them downstream - this is all about connectivity with the CR/NSRL and the Garden crowds)
- Then to the Charles either through CRP & Blossom St. or under Martha and Embankment Rds....accessing the Left Bank in a submerged tube under the Broad Canal ( i know: toxic yuck, co-gen plant, etc....i plead 'crazy')
- Then put a BL Kendall platform under Broadway and draw a connection to the RL platform under the Broadway / Main intersection (and dig a mini-yard under the Volpe lawn for good measure, before it gets redeveloped).
- Given the BL dimensions, there is plenty of room at Broadway and Gallileo for a sharp left turn onto the GJ alignment and thence Westward.

Why this would be a good:

- Eliminates most of the need for a BL-servicing Central Station in the NSRL
- Can still service MGH well, esp. from a potential CRP / Blossom St. alignment
- One-seat from NS to Kendall & RL takes pressure off Park St etc.
- Short path between BL and RL platforms @ kendall
- Less disruptive to dig under New Chardon and Martha than under Cambridge St, - both are wider and less lively, even if the path is longer...

I think it would be awesome

 

[George_Apley]

Van, I like it. But for 'crazy' purposes it seems to me that if you're going to go across the Charles and that far west on the blue you might as well do it via north station.

As follow:

- From Bowdoin down Staniford to a new 'North Station' platform under Lomasney (a bit of a walk to GR and OR here is ok because of additional connections to them downstream - this is all about connectivity with the CR/NSRL and the Garden crowds)
- Then to the Charles either through CRP & Blossom St. or under Martha and Embankment Rds....accessing the Left Bank in a submerged tube under the Broad Canal ( i know: toxic yuck, co-gen plant, etc....i plead 'crazy')
- Then put a BL Kendall platform under Broadway and draw a connection to the RL platform under the Broadway / Main intersection (and dig a mini-yard under the Volpe lawn for good measure, before it gets redeveloped).
- Given the BL dimensions, there is plenty of room at Broadway and Gallileo for a sharp left turn onto the GJ alignment and thence Westward.

Why this would be a good:

- Eliminates most of the need for a BL-servicing Central Station in the NSRL
- Can still service MGH well, esp. from a potential CRP / Blossom St. alignment
- One-seat from NS to Kendall & RL takes pressure off Park St etc.
- Short path between BL and RL platforms @ kendall
- Less disruptive to dig under New Chardon and Martha than under Cambridge St, - both are wider and less lively, even if the path is longer...

I think it would be awesome

That's such a roundabout way to do Red-Blue (which should be our highest priority when it comes to expansion).

- I'm still not convinced that the Blue Line needs a direct connection to the NSRL at all.
- The North Station-Red Line flank can be taken care of by continuing the GLX from Union to Porter and/or Green on Grand Junction running out of North Station with a free transfer (though not behind fare-gates) at Kendall.
- Three rapid transit lines at North Station is unnecessary. Particularly for the Blue Line. The Blue-Orange and Blue-Green transfer stations are only two stops away from North Station.

 

[winstonoboogie]

, and the Red Line on the Cambridge side. The Green Line can be relieved by commuter rail given operating practices that postdate the Mad Men era, and the Red Line can be relieved by upgrading the signaling so that it can run more than 15 trains per hour at the peak. .

For a half-hour period the Red Line runs 20 tph. In the PM there are 10 departures from Alewife between 4:43 PM and 5:13 PM, there are two extra Braintree trains that break up the regular alternating pattern of Ashmont/Braintree service. There is a also a brief burst of extra Braintree service in the morning as well at the "peak of the peak"

A constraint that limits to run more than this is the layout at Harvard that goes back to when the Northwest Extension first opened in 1983/84. There is a sharp curve to the immediate south of the station were the 1980s alignment cuts into the original 1912 tunnel. The combined station dwell time and travel time through the curve (which is followed by an interlocking where the two levels come back together) is about 3 minutes. Given the stopping distance for a six-car train of rapid transit equipment, the station and the curve are probably still going to be in the same signal block even in a CBTC set up.

 

[Alon]

A constraint that limits to run more than this is the layout at Harvard that goes back to when the Northwest Extension first opened in 1983/84. There is a sharp curve to the immediate south of the station were the 1980s alignment cuts into the original 1912 tunnel. The combined station dwell time and travel time through the curve (which is followed by an interlocking where the two levels come back together) is about 3 minutes. Given the stopping distance for a six-car train of rapid transit equipment, the station and the curve are probably still going to be in the same signal block even in a CBTC set up.

First, CBTC, a.k.a. moving-block signaling, means not having any signal blocks.

Second, don't be helpless! Just add more signal blocks - when the train is slow, the stopping distances are reduced, allowing shorter blocks. New York has some truly nasty curves on the numbered lines, including some right next to stations, e.g. Union Square. The combined 4/5 trains run 24 tph peak, but the limit there is not the curves, but dwell times at Grand Central, which is several times as busy as Harvard or South Station.

Van, I like it. But for 'crazy' purposes it seems to me that if you're going to go across the Charles and that far west on the blue you might as well do it via north station.

Why this would be a good:

- Eliminates most of the need for a BL-servicing Central Station in the NSRL
- Can still service MGH well, esp. from a potential CRP / Blossom St. alignment
- One-seat from NS to Kendall & RL takes pressure off Park St etc.
- Short path between BL and RL platforms @ kendall
- Less disruptive to dig under New Chardon and Martha than under Cambridge St, - both are wider and less lively, even if the path is longer...

I think it would be awesome

1. What's the cost of this detour, and how does it compare with building an Aquarium station on the NSRL?

2. Given NSRL, who needs a one-seat ride from Cambridge to North Station? It's the same criticism I made of Van's plan - when you make a crazy proposal, consider the reasonable or semi-reasonable things that will happen first. Otherwise it's like, I don't, proposing crosstown subways in New York without considering transfers to Second Avenue Subway.

3. Kendall Square is underground, so unless you're blowing the budget by having Blue Line platforms go under the Red Line platforms, it will never have an easy a transfer as Charles/MGH.

 

[winstonoboogie]

First, CBTC, a.k.a. moving-block signaling, means not having any signal blocks.

Second, don't be helpless! Just add more signal blocks - when the train is slow, the stopping distances are reduced, allowing shorter blocks. New York has some truly nasty curves on the numbered lines, including some right next to stations, e.g. Union Square. The combined 4/5 trains run 24 tph peak, but the limit there is not the curves, but dwell times at Grand Central, which is several times as busy as Harvard or South Station.

.

There is no fixed block, but there is still the distance between the trains dictated by the braking rate of the equipment. The set up at Harvard with average peak dwell times and a 10 m.p.h curve literally starting at the southern end of the platform followed by an interlocking, is unlikely to allow more than 20 t.p.h, the present amount of peak of peak service.

 

[Alon]

There is no fixed block, but there is still the distance between the trains dictated by the braking rate of the equipment. The set up at Harvard with average peak dwell times and a 10 m.p.h curve literally starting at the southern end of the platform followed by an interlocking, is unlikely to allow more than 20 t.p.h, the present amount of peak of peak service.

You keep saying that, but no, actually, the equipment brakes faster in slow areas than in fast ones. Think about it.

 

[winstonoboogie]

You keep saying that, but no, actually, the equipment brakes faster in slow areas than in fast ones. Think about it.

There is also an interlocking that controls the cross-over south of Harvard. If a train is occupying the section, the next train can't enter until the one ahead has cleared the interlocking, since it could be routed for an opposing move to cross a train back at Harvard. It is 3 minutes from platform, through 10 m.p.h curve, and through interlocking to clear the route, CBTC won't change that..

 

[Matthew]

That only means that it needs 3 minutes to clear the interlocking when the paths are set to cross-over.

Under normal operation, that's not true.

What you are describing makes it sound like only a single train can enter the interlocking at a time. But that's not necessary under any signal system even slightly more advanced than the most primitive fixed block setup.

 

[winstonoboogie]

That only means that it needs 3 minutes to clear the interlocking when the paths are set to cross-over.

Under normal operation, that's not true.

What you are describing makes it sound like only a single train can enter the interlocking at a time. But that's not necessary under any signal system even slightly more advanced than the most primitive fixed block setup.

It's an automated interlocking that is operated remotely by the dispatcher by selecting routings. If a northbound train is brought onto the platform at Harvard and crossed back, the track must be kept clear back to the cross-over, if the train is crossed back first to the southbound side, the track must be clear from the crossover to the southbound platform. Because of the two elevations at Harvard station, the distance from the cross-over to the platform is greater than what you might normally want, and it involves moving through a 10 m.p.h. curve. There is a single manual cross-over at Kendall and you have to go all the way to Park St. to find the next automated cross-over. They don't schedule to cross trains back at Harvard, but they want to maintain the ability to cross a disabled train or cross a delayed train remotely. They could remove the automated interlocking and go to a manual switch that is not protected by signals, but that is a large loss of flexibility

 

[Matthew]

Again, it's nothing new to have an interlocking that is controlled remotely, but that locks itself in place whenever trains are too close, preventing unsafe routings. You can have multiple, non-intersecting, paths through such an interlocking that allow multiple trains to pass at high-frequency without conflict.

I know you know this very well, so I'm puzzled why you think that this is an objection to frequent service on the Red Line.

 

[winstonoboogie]

Again, it's nothing new to have an interlocking that is controlled remotely, but that locks itself in place whenever trains are too close, preventing unsafe routings. You can have multiple, non-intersecting, paths through such an interlocking that allow multiple trains to pass at high-frequency without conflict.

I know you know this very well, so I'm puzzled why you think that this is an objection to frequent service on the Red Line.

An automated system is going to maintain the ability for a dispatcher to cross-back a train at Harvard at any time, which requires keeping the track clear from the interlocking to the platform until the train has cleared the platform (northbound) or cleared the interlocking (southbound). The distance and speed from the platform to the interlocking at Harvard is much longer than standard for an automated interlocking (like Alewife, Park St., Ashmont, North Quincy Quincy Center, Quincy Adams, or Braintree on the Red)

 

[Alon]

An automated system is going to maintain the ability for a dispatcher to cross-back a train at Harvard at any time, which requires keeping the track clear from the interlocking to the platform until the train has cleared the platform (northbound) or cleared the interlocking (southbound). The distance and speed from the platform to the interlocking at Harvard is much longer than standard for an automated interlocking (like Alewife, Park St., Ashmont, North Quincy Quincy Center, Quincy Adams, or Braintree on the Red)

Why is it so important to keep the platform clear until the preceding train has left the interlocking (or vice versa)? A train can enter the platform and dwell at it while the interlocking is switched. If for some reason it's super-important to have off-schedule short-turns (which is one of the reasons I bring up organization here as more important than electronics), it's perfectly possible to switch a northbound train to the southbound platform for turnback even while the preceding train is dwelling at the northbound platform.

 

[winstonoboogie]

Why is it so important to keep the platform clear until the preceding train has left the interlocking (or vice versa)? A train can enter the platform and dwell at it while the interlocking is switched. If for some reason it's super-important to have off-schedule short-turns (which is one of the reasons I bring up organization here as more important than electronics), it's perfectly possible to switch a northbound train to the southbound platform for turnback even while the preceding train is dwelling at the northbound platform.

If a northbound train is on the platform, it allows the dispatcher to route it back southbound from that track, which it couldn't do if the next Alewife train had already passed the interlocking and was committed to the route. Conversely, a southbound train won't have clearance to their platform until the previous southbound train has cleared the interlocking, as that gives the dispatcher the option to set up a route to cross a northbound onto the southbound, which they could not do if a southbound train was already committed to entering the southbound platform. Think about how the interlocking works at a terminal which has a much more standard short distance from the switches to the platform. A southbound train at Braintree would not pass through the interlocking until the previous southbound train had completed its move through the plant and was at the platform on one of the two tracks. That takes under a minute at Braintree, so it is not going to reduce capacity, but its the same idea.