Green Line Extension to Medford & Union Sq

Agreed, and also in this article:

"Pesaturo noted that those two branches serve the Longwood Medical Area, and this will minimizing transfers for riders who take the commuter rail to North Station, as well as the Green Line Extension. When service begins on the Green Line Extension, the B and C Lines will terminate downtown at Government Center (currently, B trains terminate at Park Street and C trains terminate at North Station)."

N Station to LMA is definitely a big demand driver... I don't have numbers but GLX is serving mostly residential demand so it makes sense to connect that with employment demand which I would guess is mostly centered on the E branch, inner D and perhaps the B branch.
That makes sense - good that they did consider at least NS and LMA connections. But it still seems to me like the D's Fenway/Kenmore connections make it better for jobs access than the E, but I suppose that's up for debate.

Interesting that the new map you posted from Ruggles shows the B turning at GC and C turning at N Station, while the article says both will turn at GC.
 
Are the stations going to have faregates? Originally they weren't going to and the Green Line was going to proof-of-payment, but has that changed with the delays to AFC 2.0?
 
I think they modeled it based on current state of chaos-ops (since GLX is going to predate all of the GLT optimizations), and simply didn't like what they saw. Too many hopelessly late/bunched D's needing to bail early at GC...too many run-as-directeds needing to be summoned to pick up the slack. Swapping to Union isn't a cure-all, but staunches some bleeding for the years they'll still be running old-style ops with the new branches.
 
I actually ran the numbers using MBTA origin-destination-transfer data during my Masters research. D and E were unquestionably the two branches to run through to GLX based on minimizing transfers - something like 5k less same-direction transfers. They're about equal as far as reverse-direction ridership, so transfers-wise it doesn't make much difference which branch goes to which northside terminus. There could be any number of internal reasons for the switch - perhaps it worked out better for scheduling.

Last I heard, only Lechmere will have faregates. How they're going to do fare payment before AFC 2.0 is active is a big question - the island platforms means you can't even do front-door-only boarding.
 
And I'll just note that the T put those revised maps up with (apparently) no indication that the GLX stops aren't in service yet.
 
Agreed, and also in this article:

"Pesaturo noted that those two branches serve the Longwood Medical Area, and this will minimizing transfers for riders who take the commuter rail to North Station, as well as the Green Line Extension. When service begins on the Green Line Extension, the B and C Lines will terminate downtown at Government Center (currently, B trains terminate at Park Street and C trains terminate at North Station)."

N Station to LMA is definitely a big demand driver... I don't have numbers but GLX is serving mostly residential demand so it makes sense to connect that with employment demand which I would guess is mostly centered on the E branch, inner D and perhaps the B branch.

Yeah, it's reasonable to suppose that more people would be trying to get from Somerville to Longwood vs. Kenmore.
 
I’m assuming it’s up for discussion once the GLT work is complete.
 
The length thing may be a labor/union rule. IE, maybe running the Ds all the way out would have required an extra break or something
 
The length thing may be a labor/union rule. IE, maybe running the Ds all the way out would have required an extra break or something

Highly unlikely, because the routes are so variable as-is by time-of-day loading. And they don't cut 'em early because somebody is 45 seconds late for a scheduled break; they finish the trip to destination before changing crews.

It's most likely a schedule thing. Feed end-to-end through chaos effects at peak loading, and headways likely got too gapped out too often running the D to Medford. Which would've required extra run-as-directeds to be staged within Downtown when gaps got too long, which required too much extra staff on-standby. And created another set of problems on the outflow with excessive bunching after the gaps were plugged.

I’m assuming it’s up for discussion once the GLT work is complete.

Yes. When chaos effects are tamed with the full slate of dwell-taming and signal priority improvements, they can not only look at fresh data for what the highest-capture pairings are but also probably implement way more variety for short-turn supplementals. It's just that GLX is going to open a half-decade before the main thrust of that work gets underway, so chaos is going to reign per usual in the meantime. I still think a Union terminus will require more run-as-directeds to be dispatched for plugging blown headways than we were accustomed to seeing before when D's went no further than GC, but it was the do-least-harm of the two termini for how often that would have to be invoked.
 
I actually ran the numbers using MBTA origin-destination-transfer data during my Masters research. D and E were unquestionably the two branches to run through to GLX based on minimizing transfers - something like 5k less same-direction transfers. They're about equal as far as reverse-direction ridership, so transfers-wise it doesn't make much difference which branch goes to which northside terminus. There could be any number of internal reasons for the switch - perhaps it worked out better for scheduling.

Yeah, looks like D and E are good.
 

Attachments

  • Screenshot 2021-04-09 172358.jpg
    Screenshot 2021-04-09 172358.jpg
    622.7 KB · Views: 214
FWIW, here's the results from my thesis:

The MBTA and its predecessors have changed the eastern terminals of surface branches on
numerous occasions. For decades, the B and C branches were run through to Lechmere as
the primary trunk routes. In recent years, the E Branch has been matched to Lechmere
so that the branch can be served partially from Lechmere Yard. With new data available,
passenger flows as well as operational considerations should be used to match branches to
terminals.

ODX data can be used to estimate the number of passengers forced to transfer between
branches — a key consideration. An origin-destination matrix (with the Green Line simplified
into segments that share the same services) was created using October 2016 data, and
a spreadsheet lookup table was used to count all OD pairs that would require transfers in
a given scenario. Because not all destinations are inferred, the aggregated inferred trips of
each segment were scaled up to match MBTA boarding counts. A second version was made
to estimate transfers required after the GLX is in operation, with Lechmere data scaled to
approximate additional origins and destinations on the two branches.

The current pairing (B — Park Street, C — North Station, D — Government Center,
and E — Lechmere) is fairly efficient, with 10,023 of 209,369 passengers estimated to have
to change Green Line trains. (Of those, 2,635 travel between points west of Kenmore, or
between the E Branch and points west of Copley, and would have to transfer under any ser-
vice pattern.) While keeping one branch per terminal, the least-transfer pairing (B — Park
Street, C — Government Center, D — North Station, and E — Lechmere) would require
9,840 transfers, just 183 fewer transfers than the existing configuration.
Extending an additional branch to Lechmere has been proposed as legally-required mitigation
for delays with the GLX. As well as providing more frequent service to Lechmere,
this would reduce the number of passengers that need to transfer. Extending the B Branch
to Lechmere would halve the number of transfers to 5,302.

Although the GLX will increase daily Green Line boardings by about one-sixth, it will
involve multiple through-running routes and thus could actually decrease the number of
transfers required (as well as substantially reducing bus-rail transfers at Lechmere). The
current planned service (B — Park Street, C — North Station, D — College Avenue, and E
— Union Square) would reduce transfers slightly to 9,813. Switching the B and C terminals
plus switching the D and E terminals would reduce this further to 9,066. It may be necessary
to add a second service to the Medford Branch during rush hours to handle crowding. If
the C Branch was extended to College Avenue at all hours, transfers would fall to 4,131.
 
The length thing may be a labor/union rule. IE, maybe running the Ds all the way out would have required an extra break or something

No, nothing like that exists. Remember in the old days the D ran all the way to Lechmere and operators used to bet on who could get the last train back to Riverside the fastest. Used to be a lot more fun back then :)
 
Last edited:
Looking more at the new spider map, I see they fixed one of my pet peeves: the bizarre wide, sweeping curve in the green line between Haymarket and North Station. It’s been replaced with two 45 degree turns so that the green and orange run parallel for that segment. Looks much better IMO.
 
Looking more at the new spider map, I see they fixed one of my pet peeves: the bizarre wide, sweeping curve in the green line between Haymarket and North Station. It’s been replaced with two 45 degree turns so that the green and orange run parallel for that segment. Looks much better IMO.
I think it was originally drawn that way to represent the elevated portion of the GL thru the Bullfinch triangle that ran separately from the OL. So this was 15 years overdue
 
Looking more at the new spider map, I see they fixed one of my pet peeves: the bizarre wide, sweeping curve in the green line between Haymarket and North Station. It’s been replaced with two 45 degree turns so that the green and orange run parallel for that segment. Looks much better IMO.

Now they need to fix the Blue Line so that RBX, famed as the shortest, straightest subway connection yet to build, doesn't map as an elbow.
 
Can they also fix the location of Savin hill while they're at it correcting the accuracy? It's always kind of been odd how they portray Savin Hill as a separated location.
 
Can they also fix the location of Savin hill while they're at it correcting the accuracy? It's always kind of been odd how they portray Savin Hill as a separated location.

The Braintree line doesn't stop there, so the way it's displayed here is an effective way to illustrate it. Otherwise you'd need some special notation just for that stop.
 
Last edited:

Back
Top