So, let's assume F-Line's Green LRT network is built, how weak are headways on the branches? Even with segregation in key parts of the system, certain areas will be crimped with trains coming through every two minutes. I was under the impression that 2 minute headways is the ceiling for smooth operations. Is this wrong?
For example... is the BU Bridge - Kenmore stretch going to cause headways on the feeders to suffer due to the number of lines? Say from Harvard, BC, Oak Sq, and at least two lines from the Grand Junction. That's getting heavy. What sort of max headways can the branches be looking at there? Would 5 branches only be able to manage 10 minute max headways, to keep the shared trunk at 2min headways?
Hard to say, because the whole system is going to look different with the D being load-spread down both Kenmore and Huntington, Huntington being fed into Boylston not Copley, the Seaport getting some thru-service action from Huntington/BBY, and so on.
What you can reliably assume is that the balance at Kenmore is going to tilt much heavier in the B direction with the load-spreading, and that the B direction itself will be vastly higher-capacity with the subway extension. If you're assuming BU Central gets built as a 4-track station to set up the line split, then the single-file trip time before the branching traffic clears itself out is the equivalent of Kenmore-Hynes. That's a large capacity ceiling to serve the diverging points there.
So consider, out of Kenmore:
-- The surface B outbound from St. Paul portal now needs to serve many fewer riders as a branch with the mainline subway now covering BU east campus. If headways stayed flat vs. today it would do its job equally well and suck up all growth.
-- If the Oak Sq. and Boston College lines split the frequencies 50/50 at Packards Corner (which is what they used to do), they both would cover out to Harvard Ave. equal frequencies as the B today. Harvard Ave. being the last stop before B demand drops off a cliff. In reality it's probably going to be 55/45 or 60/40 Boston College vs. Oak Sq. because one's a streetcar and one's a reservation, but whatever. There'd be a reduction in service up the hill on the B, which is the lightest-ridership portion. But that can be counterbalanced by extending C's up Chestnut Hill Ave. or Reservoir-diverging D's up Chestnut Hill Ave. to match BC frequencies to today. So assume the hill is the only place that takes a hit. Arguably, the hill
should take a hit today with beneficial installation of a short-turn track past Harvard Ave. So that may be old news by the time you build this.
-- Harvard's a shortie, and if it's going to live with a surface stub terminal for 20 years it'll have no more storage than Heath Loop does today. So E is the max frequency. Which is plenty good for, what, 3 or 4 stops at most after BU Central and 10 minutes at most to Harvard/Brattle Sq. from BU Bridge? You honestly could get by with a little bit less than that most hours of the day. It'll have more capacity to give when you build a river tunnel and go into the old Red Line tunnel with real tail-track storage...but we can't think about swallowing those kinds of expenses until much more important parts have been built downtown. Out-of-sight, out-of-mind...ceiling of E frequencies indefinitely.
-- The D gets
boosted headways from 2 trunks, and it needs it because it's now feeding Needham. Both trunks hit Longwood, so the only stop that may take a headway hit is Fenway. That specific location in the Fenway is not at a loss for surrounding stations, so not that big a deal. I think if Fenway Center gets built out and the walking path connects the Emerald Necklace with Kenmore/Brookline Ave. the center of gravity is probably going to tilt more towards Kenmore just because the unbroken wall of destinations is easier to plow through starting from that end.
-- If you find some way to thru-route on the 'boomerang', which of course we're not making assumptions about...you don't touch any Central Subway mainline capacity so your B's-to-D's are nearly unlimited.
-- Given all this the traffic skew at Kenmore can change dramatically. Say it's 40/35/25 D vs. B vs. C today (I have no idea what it really is). Under this configuration is can go 50/25/25 B vs. D vs. C with C's being flat, D's being sharply reduced out of Kenmore but gaining MORE frequencies at the Brookline Vill merge (and functionally Longwood because it's flanked), and B's taking the spoils. The gained subway throughput on the +1 extension to BU Central with its orderly track split and elimination of all E traffic from Boylston to Copley raises the ceiling even further over today. Because now you can effectively pump to Kenmore what you used to only be able to pump to Copley. Grand Junction + Harvard gets all these gains, BC/Oak's share stays flat as described above.
Now consider, out of Brookline Village:
-- Needham needs some of the lowest headways on the system. I'd say 7 minutes at peak, 10-12 off-peak. This is a modest increase BV-Newton Highlands over regular D headways, but all of it functionally comes from the Huntington trunk which can do it.
-- D's get some boost from new demand to BBY and (a little bit) to South Station.
-- Circuit service Downtown-->BV-->Kenmore-->Downtown gets some demand, especially at peak. Maybe this ends up taking most of the Kenmore D slots, and it just becomes cleaner to route nearly all Riverside trains down Huntington. Whatever works.
-- Forest Hills streetcar can run at sparse headways, loop at Park St. It did in the old days; it can forking off BV and back-tracking to South Huntington. 7 minutes peak, 10 off-peak.
Out of Lechmere:
-- Union/Porter I would assume keeps flat headways for as long as Porter is the terminus since it's mainly a radial tie-in of other transit nodes like Red and 77, plus a Somerville Ave. de-clogger. If further branches to Waltham and Watertown are considered I would think Waltham would be no greater than Needham, and Watertown *moderately* greater by +1-2 minutes of frequency at most, not a whole lot different from the Oak Sq. line. Let's not get carried away: Arsenal St.'s tippy-top potential will never be Cambridge-level, not even if H2OTown scores the redev of its dreams on the scuzzy side of the street out to the Square. So figure the Union Branch's tippy-top ceiling is about equivalent to the B after it spits out of the St. Paul portal. Pleasantly brisk frequencies, but not brisk-brisk.
-- Medford probably grows a bit on raw demand and sees more growth.
-- Chelsea/Airport gets some big ridership demanding appropriate frequencies. BUT, you are routing a minority of those frequencies around the entire half-Ring. Maybe total frequencies go a bit higher than Medford because of that, but it would probably match Medford on frequencies out of Lechmere and the Central Subway.
South End: Obviously heavy heavy service, which makes that 4-track flying junction built 118 years ago a real lifesaver.
-- Washington St., being a streetcar, gets the lightest service. Better than Arborway, better than Oak Sq. But I would say not exceeding today's E just for the sake of dispatching sanity and fact that Dudley Sq. probably isn't going to have more than maybe Heath x2's amount of storage. They do, after all, need to store a shitload of buses at that terminal too so parking spots need rationing.
-- Assume something resembling today's D + E gets pumped out Downtown-Back Bay.
-- Assume something D-equivalent gets pumped out to the Seaport. And that the Seaport gets more overall augmented by BBY<-->South Station thru-routes.
^^Adjust ratios accordingly where you think the most growth is going to be. But you get the picture: Kenmore-level out to Tufts, and cherry on top of BBY<-->Seaport skipping downtown. Transitway is going to be nice and flush.
Central Subway:
-- Kenmore-Boylston: Significant capacity gains as noted by taking E's off Copley and throttling the D's out of Kenmore by slicing/dicing trunks. I like the 'circuit' service idea for rush hour mainly because it's so much easier to dispatch Boylston-->Kenmore-->BV-->TMC-->Boylston inside the CBD rather than all the way long-haul out to Riverside or Needham. So also consider the dynamics in play of short-turns and short-'circuits' for managing the loads at peak where demand is heaviest, then running things long-haul on the off-peak to keep up appearances on the fringes while keeping up frequencies on the trunks.
-- Park. Streetcars from Arborway and Oak: loop. Do not put them through the GC pinch. Remember: inner track is also a thru track now so you have traffic sorting at Boylston merge. I'd stick the Kenmores on the inner and South Enders on the outer. And just let tiny Arborway be the one that crosses traffic to loop.
-- Park-GC.
Ooof. OK, this is a problem because of the tunnel expandability issue. Which is why you really really want to explore all options. Even if you can only wring THREE tracks of width, and make it 2 eastbound, 1 westbound (the mismatch being a traffic management thing for the routes at ends of their runs turning at GC)...that's a monumental improvement. Otherwise, I think you can get by but be prepared to do more Seaport-->Huntington and short-turning of trains inbound of Lechmere at GC/Brattle Loop.
-- GC-Haymarket. Reconfig GC station to split back into 4 tracks so you've got a third straight station with traffic separation. This mitigates the pinch between Park-GC a bit, which is why you can survive if it's not expandable. And it's doable. But if you
can expand the Park-GC tunnel...you essentially have 2x the Central Subway capacity between Boylston merge and GC loop.
^^Lots of things should be looping at GC from the west. Washington St., and BC, for sure. Needham...sure. 'Circuit' service...sure.
-- Brattle Loop-Haymarket. I think Haymarket's easy enough to quad up for a 4th consecutive station with traffic sorting options, but you're largely home free at this point. Some stuff from the east like Grand Junction running 'circuit' GC-->Kenmore-->BU Central-->Lechmere should be looping at Brattle Loop and not continuing to Park.
-- North Station. Seaport-North Station rush-hour short-turns. The subway "North-South Link". Do it. Off-peak, maybe have Cleveland Circle revert to its traditional spot here and have South End stuff running longer-haul.
-- Lechmere. Some stuff will short-turn here on shift changes because of the carhouse, but otherwise whatever continues on from North Station is diverging somewhere past Lechmere.
It's rough, because so many moving parts are fitting together. And the traffic management is so very very very different than today because of the alternate trunks, alternate routings, re-shaped traffic, short-turns, reshaped by time-of-day traffic, etc. etc. It doesn't even remotely resemble today's Green Line because of the shape-shifting capability and ability to serve up destination pairs for any demand pattern. You don't necessarily even have to know today what those demand patterns are going to be, only where the blending can handle it.
But I think you can give everything headways no worse than today with no loss of mission-critical frequencies. And only one worrisome limiter: the Park-GC 2-track pinch. Which is augmentable by boosting on the couple routes (Huntington<-->Seaport, Brattle Loop<-->Lechmere diverging routes) that don't touch it...less-ideal as that sounds.
