Urban Ring

[F-Line to Dudley]

Is there anything stopping us from, alongside or in place of the BU Bridge crossing, starting a branch or two of the Green Line out of North Station? I agree with Henry that there's a lot more value in building short spurs that could later be connected to each other.

Something like a phase one GLX to Chelsea, phase one Urban Ring over the bridge, then extend Union Square Green Line to Porter, connect Porter to the BU Bridge Spur and reroute those trains over to Chelsea. BLX to Lynn, Urban Ring to Lynn, and presto - you're done.

Absent or independent of that, the Orange Line doesn't need to be spurred. If we're going to do anything with the Orange Line, we should grab another five trainsets and double our headways on that line.

The Urban Ring doesn't have a clear integrity of concept. The city needs better radial distribution...great. The city's ideal radial distribution patterns and demand overlap on a map in a distinct ring area around downtown roughly conforming to the UR route...great. But since when did a one-seat radial ride become a necessity? How many people would actually ride this thing nearly end-to-end? That's where this falls part...where it gets force-fitted into the T's everything-must-be-a-one-seat-ride philosophy. People are going to ride this thing in segments, but there won't be much overlap of folks riding the SE quadrant in Southie and the NW quadrant between MIT and Lechmere. A 1-transfer ride on the subway will beat the one-seat roundabount every time. The T is severely overestimating the value of the one-seat and not taking into account that a line that doesn't reach any of the big half-dozen central transfer stations serves a fundamentally different purpose than the 4-1/2 lines that do. That error in logic is driving all the design decisions here.

Same as with the Silver Line. Why do people need rapid transit to Dudley so badly? Because it's the system's largest bus hub. The largest % of users are going to be using it as a radial line, even though it is not ring-shaped and Washington St. does have characteristics of a rapid transit surface branch in its own right. Hardly anyone needs a one-seat direct to Southie on it, but that feature became the unifying concept of the whole SL Phase III boondoggle and all of the project's design features were force-fitted around that.

Keep it simple. If it covers a radial transfer need, then it essentially is fulfilling the goals of the UR. Spurs fulfill those goals.

-- Green to Porter from Union. Red to Green/Orange and commuter rail, links 3 bus routes at Porter (incl. the 77) to 5 more at Union and 3 more at Lechmere.
-- Green to Sullivan stub. 12 bus routes + Orange at Sullivan to 4 bus routes + 2 other GL branches at Lechmere. The Green Line historically used to have 2 branches to Sullivan via Charlestown (albeit on different routings) serving different needs than the Orange Line, so there's not a redundancy here. The tracks to the GLX maintenance yard will also get halfway to Sullivan to begin with, so it's easy to continue. And this is your jumping-off point for Chelsea and Logan.
-- Green on Grand Junction from Lechmere. Pretty obvious why this is needed. MIT/CT2 corridor.
-- Green on Grand Junction via Kenmore. If B is buried underneath the reservation out to BU Bridge to form a connection. CT2 + 47. Reconfigure Kenmore so the B/D sides can loop there and it costs about $1B+ less than the cross-Brookline fantasy tunnel while linking the 66/65/60 at Brookline Village. Add E-to-D connecting track to BV and it links JP. Fistfuls and fistfuls of transfers.
-- Green to Dudley. Of course.
-- Green to South Station via some sanely-dug tunnel (probably under NEC from South End instead of through Chinatown).

And so on. Prioritize as you see fit, but it doesn't need to be a monolith to fulfill its radial transfer integrity-of-concept. And it damn sure doesn't have to be a totally different mode to fulfill its integrity-of-monolith inside a wholly invented one-seat integrity-of-concept.

 

[dshoost88]

OR in lieu of a publicly financed Urban Ring project, instead an organization pools financing from all the institutions that would benefit from ridership along an Urban Ring heavy-rail corridor and charge a premium for the service. Harvard, MIT, BU, all Longwood Hospitals, NU, WIT, Boston Medical, BCEC, Bio-Tech @ Seaport, Logan, Assembly Row, Tufts, Lesley, and back to Harvard.

Just my stream of consciousness, sorry. One can dream, right?

 

[F-Line to Dudley]

OR in lieu of a publicly financed Urban Ring project, instead an organization pools financing from all the institutions that would benefit from ridership along an Urban Ring heavy-rail corridor and charge a premium for the service. Harvard, MIT, BU, all Longwood Hospitals, NU, WIT, Boston Medical, BCEC, Bio-Tech @ Seaport, Logan, Assembly Row, Tufts, Lesley, and back to Harvard.

Just my stream of consciousness, sorry. One can dream, right?

Orange isn't well-suited to spurs. Or at least it hasn't been since the Atlantic Ave. El went away and all the original plans for spur Els to Southie, etc. were abandoned. Red can handle branches because they split well outside of downtown and the ridership craters bigtime after South Station, meaning most of the schedule mileage works just fine on branchline headways. Most of the stops Broadway-south have half the boardings of everything north of there, even the Alewife extension stops. I don't know how you'd be able to mix and match it on Orange when stops as far-flung as Malden Ctr. and Forest Hills hang within < 1000 boardings of a State St. or Haymarket. Except for some of the least important SW Corridor stops the boardings stay at a pretty even keel throughout, so branching would unduly punish headways at stops like MC or FH that need the full schedule. Which even with a full fleet and more capable signal system is going to put a crimp in things relative to headways on the rest of the line.

Orange is well-suited to going long-distance in each direction with fairly tight station spacing because of the express tracks in Somerville and Medford. If the Reading extension were to happen Track 3 would take over the commuter rail track and extend to between Oak Grove and Wyoming Hill, and theoretically you could go full quad-track between Sullivan and Malden Ctr. (but not CC to Sullivan because of the flyover over the CR tracks is only a max 3-wide). Forest Hills and Wellington would set up short-turn spots for rush hour extras coming from the terminals, and you could easily do a full-blown Reading-Needham run without overly taxing the line downtown.

 

[bigeman312]

-- Green to Dudley. Of course.

What could we call it.....

 

[omaja]

The Urban Ring doesn't have a clear integrity of concept. The city needs better radial distribution...great. The city's ideal radial distribution patterns and demand overlap on a map in a distinct ring area around downtown roughly conforming to the UR route...great. But since when did a one-seat radial ride become a necessity? How many people would actually ride this thing nearly end-to-end? That's where this falls part...where it gets force-fitted into the T's everything-must-be-a-one-seat-ride philosophy. People are going to ride this thing in segments, but there won't be much overlap of folks riding the SE quadrant in Southie and the NW quadrant between MIT and Lechmere. A 1-transfer ride on the subway will beat the one-seat roundabount every time. The T is severely overestimating the value of the one-seat and not taking into account that a line that doesn't reach any of the big half-dozen central transfer stations serves a fundamentally different purpose than the 4-1/2 lines that do. That error in logic is driving all the design decisions here.

I don't think having a disconnected spur routes would necessarily alleviate congestion at the over-saturated Downtown interchanges like a continuous urban loop ring line would. Not to mention branches and varieties of service along the same shared track is always much more complex/prone to operational issues and more confusing for passengers.

Look at the ring lines in places like Beijing, Berlin, Madrid and Moscow - some of the most efficient and reliable rail lines in operation.

 

[Matthew]

Well the point of the article that Henry linked to is that Moscow's Circle is too small in diameter, and therefore non-competitive with the radial lines for significant distances. And I think the Urban Ring plans fall in the same boat. Unless your destination or origin happens to lie near the Ring.

 

[czsz]

Orange isn't well-suited to spurs. Or at least it hasn't been since the Atlantic Ave. El went away and all the original plans for spur Els to Southie, etc. were abandoned. Red can handle branches because they split well outside of downtown and the ridership craters bigtime after South Station, meaning most of the schedule mileage works just fine on branchline headways. Most of the stops Broadway-south have half the boardings of everything north of there, even the Alewife extension stops. I don't know how you'd be able to mix and match it on Orange when stops as far-flung as Malden Ctr. and Forest Hills hang within < 1000 boardings of a State St. or Haymarket. Except for some of the least important SW Corridor stops the boardings stay at a pretty even keel throughout, so branching would unduly punish headways at stops like MC or FH that need the full schedule. Which even with a full fleet and more capable signal system is going to put a crimp in things relative to headways on the rest of the line

Isn't the solution to this simply increasing the number of trains?

 

[bigeman312]

Orange isn't well-suited to spurs. Or at least it hasn't been since the Atlantic Ave. El went away and all the original plans for spur Els to Southie, etc. were abandoned. Red can handle branches because they split well outside of downtown and the ridership craters bigtime after South Station, meaning most of the schedule mileage works just fine on branchline headways. Most of the stops Broadway-south have half the boardings of everything north of there, even the Alewife extension stops. I don't know how you'd be able to mix and match it on Orange when stops as far-flung as Malden Ctr. and Forest Hills hang within < 1000 boardings of a State St. or Haymarket. Except for some of the least important SW Corridor stops the boardings stay at a pretty even keel throughout, so branching would unduly punish headways at stops like MC or FH that need the full schedule. Which even with a full fleet and more capable signal system is going to put a crimp in things relative to headways on the rest of the line.

I see your point. What about a spur at Ruggles and have the Franklin/Providence/Stoughton trains all stop at Forest Hills. This would at least partially offset the increase in Forest Hills headways. Additional fleet could offset the rest. This would cause a relative increase in headways for Green St, Stony Brook, Jackson Sq and Roxbury Crossing, which is okay. What do you think F-Line?

 

[bigeman312]

I see your point. What about a spur at Ruggles and have the Franklin/Providence/Stoughton trains all stop at Forest Hills. This would at least partially offset the increase in Forest Hills headways. Additional fleet could offset the rest. This would cause a relative increase in headways for Green St, Stony Brook, Jackson Sq and Roxbury Crossing, which is okay. What do you think F-Line?

By the way, I don't like the idea of a spur at Ruggles to Kenmore or wherever it is we are theoretically discussing. I do think it could be feasible, though. I much prefer the idea of a spur at Back Bay. If a spur at Ruggles were implemented, I would support it going to Dudley.

 

[F-Line to Dudley]

Isn't the solution to this simply increasing the number of trains?

Demand. Forest Hills currently has about 1000 more daily boardings than State. If you double service at State but leave it the same at FH because some branch forks off en route, the demand for FH trips is going to be more exaggeratedly under-served by the increased hordes flooding State. Or something to that effect. Obviously there are a lot of other variables in play that make it less simple, but FH and MC having near-downtown level boardings at stops near the ends of the line makes load-balancing branches a lot tricker. There's nothing comparable on the Red branches. Even Quincy Center's robust boardings are small potatoes compared to everything north of Broadway.


This is a big reason why the Green Line is so hard to dispatch. Pre-1959 when it was all streetcar-fed, it was a somewhat self-regulating beast where traffic entered the subway first-come/first-served, and the branches didn't have to be dispatched on an "on-time" start-to-finish schedule so much as get managed around the recommended headway. That was because there were no incredibly long-distance branches, and the boardings got much more diffuse after the branches forked. That changed when the Riverside Line was introduced. They thought they were adding a Mattapan-style interurban with light ridership on light enough headways that there'd only be a few cars occupying the branch at any given time, and that service increases at peak hours would loop at Kenmore for transfer Ashmont-style so the streetcar branches still got the lion's share of subway ridership and it could still self-regulate on very inapproximate schedules.

Well, didn't quite work out that way. The boardings on the D were so far above and beyond projections that it instantly became the single busiest and most car-hungry branch. And worse, the boardings were not diffuse interurban-type rides but were instead almost entirely headed downtown. Which means they couldn't regulate it with Kenmore short-turns lest every other branch get screwed up with the dwell times from cross-platform transfers. And with it being such a long-distance line it had to run on a clock...all the way downtown in mixed traffic. Hence, everything else had to get run on tighter schedules based on start/finish clock time instead of spacing behind the car ahead of them, and the signaling (still) isn't up-to-snuff for managing apples/oranges trips with precision.

It wouldn't be exactly the same on a heavy rail line, but you can see that branching gets more difficult the less alike the branches are in ridership/service characteristics and the more high-ridership outliers there are on the outer branch stops. We don't have the alt routes or express tracks downtown to manage that smoothly. Signals can increase capacity on Red/Orange/Blue, but they don't change the very nature of the service patterns you could cram through downtown on those lines. They'll always be end-to-end lines. Green, on the other hand, has been functionally nuked since 1959 and simply festering in its own brokenness for a half-century since. An improvement back to 1959-level relative sanity would FEEL revolutionary just because it's been that long since dispatching worked correctly. Remember, until the early 50's it handled 8 full-time branches and multiple flavors of short-turn service off those branches just fine. It "lost" that capacity because the schedule give-and-take between branches now has to march to Riverside's beat, but on the same first-come/first-serve signal system as before. The only way to kinda sorta make it work was to slash service levels and tie the dispatchers' hands so everything is ruled by per-run on-time performance clock instead of headways. And that's a futile, futile fight when the branches have no signal priority and the subway is all short manual stop-and-go blocks that can't auto-tune the train spacing.

 

[Max Power]

The Urban Ring doesn't have a clear integrity of concept. The city needs better radial distribution...great. The city's ideal radial distribution patterns and demand overlap on a map in a distinct ring area around downtown roughly conforming to the UR route...great. But since when did a one-seat radial ride become a necessity? How many people would actually ride this thing nearly end-to-end? That's where this falls part...where it gets force-fitted into the T's everything-must-be-a-one-seat-ride philosophy. People are going to ride this thing in segments, but there won't be much overlap of folks riding the SE quadrant in Southie and the NW quadrant between MIT and Lechmere. A 1-transfer ride on the subway will beat the one-seat roundabount every time. The T is severely overestimating the value of the one-seat and not taking into account that a line that doesn't reach any of the big half-dozen central transfer stations serves a fundamentally different purpose than the 4-1/2 lines that do. That error in logic is driving all the design decisions here.

Same as with the Silver Line. Why do people need rapid transit to Dudley so badly? Because it's the system's largest bus hub. The largest % of users are going to be using it as a radial line, even though it is not ring-shaped and Washington St. does have characteristics of a rapid transit surface branch in its own right. Hardly anyone needs a one-seat direct to Southie on it, but that feature became the unifying concept of the whole SL Phase III boondoggle and all of the project's design features were force-fitted around that.

Keep it simple. If it covers a radial transfer need, then it essentially is fulfilling the goals of the UR. Spurs fulfill those goals.

-- Green to Porter from Union. Red to Green/Orange and commuter rail, links 3 bus routes at Porter (incl. the 77) to 5 more at Union and 3 more at Lechmere.
-- Green to Sullivan stub. 12 bus routes + Orange at Sullivan to 4 bus routes + 2 other GL branches at Lechmere. The Green Line historically used to have 2 branches to Sullivan via Charlestown (albeit on different routings) serving different needs than the Orange Line, so there's not a redundancy here. The tracks to the GLX maintenance yard will also get halfway to Sullivan to begin with, so it's easy to continue. And this is your jumping-off point for Chelsea and Logan.
-- Green on Grand Junction from Lechmere. Pretty obvious why this is needed. MIT/CT2 corridor.
-- Green on Grand Junction via Kenmore. If B is buried underneath the reservation out to BU Bridge to form a connection. CT2 + 47. Reconfigure Kenmore so the B/D sides can loop there and it costs about $1B+ less than the cross-Brookline fantasy tunnel while linking the 66/65/60 at Brookline Village. Add E-to-D connecting track to BV and it links JP. Fistfuls and fistfuls of transfers.
-- Green to Dudley. Of course.
-- Green to South Station via some sanely-dug tunnel (probably under NEC from South End instead of through Chinatown).

And so on. Prioritize as you see fit, but it doesn't need to be a monolith to fulfill its radial transfer integrity-of-concept. And it damn sure doesn't have to be a totally different mode to fulfill its integrity-of-monolith inside a wholly invented one-seat integrity-of-concept.

I agree that these spurs will handle the vast majority of the needs of the urban ring. Especially north of the Charles. I like the idea of Sullivan becoming a key transfer point. My only concern is linking Dudley/Ruggles/Kenmore. That's a huge transit gap. I don't think a green line spur to Dudley (I'm assuming it's coming from downtown via washington?) will be enough to handle that in the long run. I think that will cripple the green line going through park/government center. Some of this traffic should be diverted from the downtown crush. Are there any solutions to this problem w/o tunneling deep bore?

 

[Matthew]

Well, didn't quite work out that way. The boardings on the D were so far above and beyond projections that it instantly became the single busiest and most car-hungry branch. And worse, the boardings were not diffuse interurban-type rides but were instead almost entirely headed downtown. Which means they couldn't regulate it with Kenmore short-turns lest every other branch get screwed up with the dwell times from cross-platform transfers. And with it being such a long-distance line it had to run on a clock...all the way downtown in mixed traffic. Hence, everything else had to get run on tighter schedules based on start/finish clock time instead of spacing behind the car ahead of them, and the signaling (still) isn't up-to-snuff for managing apples/oranges trips with precision.

Couldn't this be alleviated a little bit by using somewhat more dynamic scheduling; for example, by changing destination and sending trains back out on whichever route needs it next, when they reach their downtown terminus and turn around?

 

[Commuting Boston Student]

Given infinite feddybux, I'd be screaming up and down the Central Subway at anybody who would listen to me that we need express tracks running downtown. As it is, I don't think such a proposal could make it out of cost/benefit analysis, and rightly shouldn't, but maybe 50 years from now?

As for the Urban Ring being a full line instead of several branches - I think that the single-seat ride is a fringe bonus on the real benefit of having the Urban Ring be a complete line. That benefit is, rolling stock distribution. If we have 12 Urban Ring trains and seven or eight short spurs to distribute them between, things are going to get a lot messier than if we can have six trains going one way and six trains going the other way. Very few people will stay on for more than one 'spur' worth of line, but that's okay, because emptying a train only to fill it with new and different people works out great.

 

[F-Line to Dudley]

Couldn't this be alleviated a little bit by using somewhat more dynamic scheduling; for example, by changing destination and sending trains back out on whichever route needs it next, when they reach their downtown terminus and turn around?

Making everything a run-as-directed only works if all branches have roughly equal fleet needs with roughly equal spread of cars in the outer yards. They don't...the D uses far more cars than any others. More than C+E combined. And the truncated E doesn't have its own carhouse anymore so they have to ration the assignments in tiny Lechmere yard very carefully. This wasn't a problem in the old days when A, B, C, E were all more or less equal-weight and each had carhouses at the ends (BC and Reservoir of course semi-shared), the other former branches (2 South End branches to Egleston and City Point, 2 North Station branches to Sullivan flanking different parts of Charlestown) split up the rest of the pie, and the longest and most schedule-uncertain street-running runs to Arborway and Watertown were backed up at rush hour by short-turns at Heath and one of the loops at Braves/BU Field, Union Sq., or Oak Sq. Just wave 'em in and out at whatever spacing makes them hit that "8 minutes to Park St." sign at the portal at a rate that will get them to Park in 8 minutes, and that's all there was to it.

A super-exaggerated metaphor for what it's like dispatching the GL pre-1959 vs. post-1959: Harvard bus tunnel. Everything for half a century turns there first-come/first-served, and it's all orderly and civilized for the 10 or so routes that use it. But say when the Alewife extension of the Red Line was built it went right through the bus tunnel and every mode had to share the platform. Still first-come/first-served. The effects on RL (a.k.a. the D) schedule-keeping are obvious. But what about the buses? What if, the only way it could ever plausibly work is if they whacked two-thirds of the routes and crippled the headways on the 77 so every single bus was bursting at the seams and choking on its own dwell times? Because that's the only way everything could behave on first-come/first-served basis.

Green's got a LOT of room to grow if it only had signals/dispatching that were actually designed-to-task for mixing the kind of loads and schedule precision the D requires with the headways and schedule fudge factor the streetcar branches (but especially the B) require. But we're now in Decade #6 where it does not.

GLX actually won't complicate things as much as feared because of the conversion of the inner inbound track at Park St. to thru service and the new Innerbelt carhouse. But for damn sure there's going to have to be liberal amounts of short-turning at GC/Brattle Loop at crush load to pad everything that could go wrong getting from Riverside to Route 16 on-time.

 

[bigeman312]

I have a semi-related question:

What is the most heavily patronized section of the 66?

(i.e. Dudley to Roxbury Crossing, Roxbury Crossing to Brigham Circle, Brigham Circle to Brookline Village, etc.)

If the answer to this question is known, that seems like a very good starting point for an Urban Ring-related spur.

 

[HenryAlan]

I don't have the stats on this, but I wonder if branching the OL would result in less bus ridership to MC and FH. Couldn't that mitigate against the problem you've described F-Line? I'm assuming a radical redesign of bus routes to go along with the branching, which ultimately results in fewer passengers accessing the system at the heavy use stations, but boarding instead along the new branches.

 

[F-Line to Dudley]

Given infinite feddybux, I'd be screaming up and down the Central Subway at anybody who would listen to me that we need express tracks running downtown. As it is, I don't think such a proposal could make it out of cost/benefit analysis, and rightly shouldn't, but maybe 50 years from now?

As for the Urban Ring being a full line instead of several branches - I think that the single-seat ride is a fringe bonus on the real benefit of having the Urban Ring be a complete line. That benefit is, rolling stock distribution. If we have 12 Urban Ring trains and seven or eight short spurs to distribute them between, things are going to get a lot messier than if we can have six trains going one way and six trains going the other way. Very few people will stay on for more than one 'spur' worth of line, but that's okay, because emptying a train only to fill it with new and different people works out great.

Well, they did have 4 thru tracks Boylston-Park and GC-Haymarket in the old days. Pre-1941 E's also split off at the old grade-separated portal on Boylston St. instead of needing Copley Jct., and of course pre-'59 loads from Kenmore to Copley were nothing like they are today. And almost every branch had short-turns. There's a number of things beyond signaling they can to do restore lost capacity.

Minor/modest $$$:

-- D to E surface connection at Brookline Village. Pad rush-hour service with this, load-spread away from Kenmore via the E during Sox games. Emergency bypass. Alt peak-hour routing if they ever do the Needham branch.
-- Short-turn on the B. I am hoping like hell whenever MassHighway funds the Comm Ave. reconstruction from Packards to Warren that relocates the reservation into the center that they build a Blandford-style turnback track between Harvard Ave. and Griggs. How much better would that make the school bus if they could augment service without caring about making it all the way up the hill on-time?
-- GC 4-track reconfiguration. It's a shame that on such a humongous platform it can't have a Park-style track layout where thru trains can get to/from the loop platform. Put down a track split on the Park St.-facing side that cuts across the platform straight to the loop platform and segregate branch boardings by track. It would change pedestrian access a bit, but the giant wedge has so much room they can easily do this if they wanted. And may have to think about it if the Park-GC tunnel itself can't be widened to 4 track.

Mega $$$:
-- Widen the Park-GC tunnel for continuous multi-track ops. I have serious doubts they could ever make it 4-track because of the burying ground. But if an engineering assessment says 3-track works, do it. Make it double-track for inbound runs so trains turning at GC or NS can stay on-schedule, leave the lone outbound single for trains beginning their runs.
-- Move the E off Copley Jct. This is a considerable dig...but that branch could split off at Boylston in the old tunnel, hang a left under Marginal Rd. (1965 Pike-cleared fill...they could pretty much dig sideways from the Pike retaining wall), go to Back Bay with a new stop, and then rejoin the Huntington tunnel at the Prudential curve. Run a subset of D service over the D-to-E connecting trackage. As for $$$...remember, 1965-cleared urban renewal land abutting interstate retaining wall, with no spaghetti utilities underground. This will be cheaper than either the SL Phase III tunnel through Chinatown or the billion-dollar UR tunnel through Brookline/Longwood. By a couple $B.
-- Bury the E to Brookline Village, route all D service on it. Eventually think about subwaying to Brookline Vill. and totally bypassing Kenmore with full grade separation and a parallel-load subway. It'll work. Northeastern-Brigham Circle is an easy dig empty of underground construction. Brigham-BV thornier, but worth it for closing that gap. Now you can do full "circuit" service in the subway. Turn at GC or NS, go to BV, back inbound via Kenmore, turn back at GC or NS. Whole second Central Subway for probably the same cost or less than doing the truly impossible Silver Line III dig. There's where all the Urban Ring extra capacity comes from, and if SL Phase III gets grafted on here with a trolley tunnel snaking to SS under the NEC, then you've joined the Transitway to the whole works too.
-- Make the grade-separated branches heavy-rail: D to Blue via "Riverbank Subway" under Storrow, GLX to Red via half of the N-S Link and Columbia Jct. (the "Red X" plan where 2 north and 2 south branches switch off each other every-other-train at JFK on the full grade separated 4-track junction). I think this is a real stretch since Blue would have issues handling that many stops end-to-end unless everything from Lynn and you'd probably have to sacrifice the Needham branch (and its grade crossings). Then, on the remaining GL....back to the future, pre-1958. All branches (including many new routes) as streetcar feeders. Enough other transit systems are building these that clearly there's some lasting modern-era wisdom to what BERy wrought in 1897.


And so on and so on.

 

[F-Line to Dudley]

I don't have the stats on this, but I wonder if branching the OL would result in less bus ridership to MC and FH. Couldn't that mitigate against the problem you've described F-Line? I'm assuming a radical redesign of bus routes to go along with the branching, which ultimately results in fewer passengers accessing the system at the heavy use stations, but boarding instead along the new branches.

Well, FH is an easier problem to solve because they could always put the trolley tracks back to help serve that need. But also consider that FH boardings are probably lower overall now vs. when it was a major rail transfer stop, so there's an inhibited demand effect from the 39 "temporary replacement". That's a case where I think demand is going to far outstrip supply for a long time, service increases or not. It's got far less service than it had in 1985, and it's grown a ton since 1985. MC I'm less certain about since I don't know what the ridership growth curve has been there from 1975 on. That one would probably have its demand fully-enough satisfied by modest service increases. But I don't know quite how explosively it's growing.

As for bus routes, I don't think a "radical" reconfiguration is possible. At least not with the core roster of downtown routes. Nearly every bus on the system is an old streetcar route, which in turn was an old horsecar route. They follow established square-to-square travel patterns that have been more or less in-place since post-Reconstruction and post-landfilling the city. Relocating an established route to a new path is still a big controversy 25 years later with the El/SW Corridor/Silver Line, and that's at rapid-transit spacing. I don't know how you could do a radical makeover of even more localized bus routes without creating way more problems than it solves. These routes are more than habit...they're DNA after 90-130 years in service. JP, Rozzie, West Roxbury farmland became an extension of the city when they became streetcar suburbs on those same exact routes on same exact streets. We found out the hard way in urban renewal that you can't just tell neighborhoods to be something else because it makes my map look nicer and expect them to adjust healthily. That always ends badly.

 

[Matthew]

Making everything a run-as-directed only works if all branches have roughly equal fleet needs with roughly equal spread of cars in the outer yards. They don't...the D uses far more cars than any others. More than C+E combined. And the truncated E doesn't have its own carhouse anymore so they have to ration the assignments in tiny Lechmere yard very carefully. This wasn't a problem in the old days when A, B, C, E were all more or less equal-weight and each had carhouses at the ends (BC and Reservoir of course semi-shared), the other former branches (2 South End branches to Egleston and City Point, 2 North Station branches to Sullivan flanking different parts of Charlestown) split up the rest of the pie, and the longest and most schedule-uncertain street-running runs to Arborway and Watertown were backed up at rush hour by short-turns at Heath and one of the loops at Braves/BU Field, Union Sq., or Oak Sq. Just wave 'em in and out at whatever spacing makes them hit that "8 minutes to Park St." sign at the portal at a rate that will get them to Park in 8 minutes, and that's all there was to it.

Well, the Blue Book vehicle count for AM Peak is:

"B" - 20 vehicles, "C" - 14 vehicles, "D" - 21 vehicles, "E" - 17 vehicles, "Run as directed" - 3 vehicles. So I'm not sure what you mean by "D" = "C" + "E" combined there. Plus, the "B" line is easily the most heavily ridden.

I don't know which portal is 8 minutes to Park Street, it's at least 15 minutes from Kenmore normally. That is, if your train isn't held up by the dispatcher for "headway adjustment" which, when they say it, honestly sounds like a load of B.S. to me.

If the branches operated at different headways, then I could see there being an issue. But they don't. They all are supposed to hit 5 minute headways at peak (in theory) with homogeneous equipment. So in an ideal situation (assuming Copley was a flying junction and signalling wasn't shit), they would all be able to interleave in the Central Subway at approximately 1.25 minute headways regardless of the number of cars it takes to maintain that on each branch. Of course, in the real world, there's delays, and the various problems with signals and Copley. That's why they should all run as directed and the dispatcher should choose the next destination based on both the forward and the backward headways of all the branches at that moment in time.

 

[F-Line to Dudley]

Well, the Blue Book vehicle count for AM Peak is:

"B" - 20 vehicles, "C" - 14 vehicles, "D" - 21 vehicles, "E" - 17 vehicles, "Run as directed" - 3 vehicles. So I'm not sure what you mean by "D" = "C" + "E" combined there. Plus, the "B" line is easily the most heavily ridden.

I don't know which portal is 8 minutes to Park Street, it's at least 15 minutes from Kenmore normally. That is, if your train isn't held up by the dispatcher for "headway adjustment" which, when they say it, honestly sounds like a load of B.S. to me.

If the branches operated at different headways, then I could see there being an issue. But they don't. They all are supposed to hit 5 minute headways at peak (in theory) with homogeneous equipment. So in an ideal situation (assuming Copley was a flying junction and signalling wasn't shit), they would all be able to interleave in the Central Subway at approximately 1.25 minute headways regardless of the number of cars it takes to maintain that on each branch. Of course, in the real world, there's delays, and the various problems with signals and Copley. That's why they should all run as directed and the dispatcher should choose the next destination based on both the forward and the backward headways of all the branches at that moment in time.

That doesn't solve the problem of where the cars are going to come from. To do an all-RAD setup you'd have to have yards with equal supply at the end of each branch. They don't, and haven't had that since pre-'59. Riverside is larger than all other yards/layovers combined. That means all equipment staging moves are at root shaped along on the D's schedule. Which is why the Green Line isn't a free-for-all. They don't even use the Chestnut Hill Ave. tracks between the B and C outside midday shift-changes because of traffic impacts, so tiny Lake St. carhouse has to manage its allotment miserly without being able to phone in reinforcements from down the street. If there's cascading delays on the B, they can't just flush it clean by running as many RAD's as it takes to de-clog...they don't have the equipment for RAD's to begin with. They're dependent on things hitting their schedule slots. The E is the same way. Heath has no storage except for the inner loop only used for disablements, Lechmere is tight on space, and Lechmere is only fed by that line so it does take a schedule change to move stuff from elsewhere. North Station is where shift changes get staged on the inbound side, so while that's got some flex mid-shift about 4 times a day it gets stuffed solid and/or emptied dry by the change on/off-peak. 6 times when there's a Garden or Fenway event, because game-day RAD's get staged from there. Currently Lechmere is the base for the RAD's because it's the one place equally accessible inbound to all branches without truncating any runs, but that does make it more difficult to fire limitless ammo at GC-turning B's consequence-free. GC loop has no storage whatsoever except for disabled trains on the extra track to Haymarket. So there's no secure base where they could fire RAD's at-will at the outbound B. Unless anyone really thinks lengthening that run even further to Lechmere is going to make the school bus run any smoother.

At most, the C is the only line where most hours of the day there's ample space at each end and reassignments can be done without having to abort a revenue run midway through. Most hours. Not 10:00, 4:00, 7:00 or any night there's a game. Not surprisingly, the C is the line that seems to have fewer "standing by for schedule adjustment" pauses...because it already is the path of least resistance for RAD's. It won't work unless all lines have roughly equal flex like that. So do we wage war on land-takings at BC and JP to build more yard space on the B and E, buy an assload more cars, tell every D rider that they're looping at Park St. from now on, and embrace pre-1959 run-as-directed managed chaos? Or do we expend the energy doing some pretty cut-and-dried work to make the trains run sorta more on-time?


This is why expensive-ass GLX has that expensive-ass Innerbelt carhouse attached. They've needed this ever since Arborway and Watertown got cut from the system. But it's still going to be operating a D-centric schedule because that's the only line that'll link it to another carhouse, so in no way does it negate the need to get out of the 19th century with dispatching or put the subway on a more RAD-centric dispatching. It's just bringing back some of what was lost in the 80's.