There are two broad options for connecting the Transitway to the Green Line system as a whole: Either you connect from Park St or you connect from Kenmore (let's ignore a new E line trunk for now). the south side, in my opinion, has far more capacity to give than anything routed through Park St. Given the narrow street geometry near Boylston, I don't see any sort of track curve happening to bring anything from Kenmore down south to reconnect to the Tremont subway. Far better, in my mind, to simply follow Van's suggestion of using the widened tunnel near Arlington to bury a new tunnel under Boylston and Chinatown. The costs are high, yes, but if it means actually being able to feed the Seaport rather than sacrificing overall system capacity I think it's worth it.If they pick up Transitway-Downtown for re-study as LRT, they'll work up the Essex St. Alternative again simply because that was the best-studied one when it was Silver Line Phase III. But its chances are not good because of the expense incurred by structurally underpinning Boylston Station and Chinatown Station (with a double-decker Silver platform, making it 3 stories deep!!!), and the mitigation costs incurred by the digging under very narrow Essex.
The killer flaws that need to be corrected to net a cost estimate that's actually buildable are:
That probably means you're doing a South End jog, hitting Orange-Tufts Med. Ctr. with a connecting concourse, and looking for wide and/or urban renewal streets (Kneeland, Marginal...more than one way to try it). Then hitting the pre-provisioned trajectory into the Transitway at the north tip of Chinatown Park. And keep in mind: as BRT the Essex tunnel was to be so godawful dog slow that nearly all service would be forced to loop from either side at Boylston for dispatching sanity. So a trolley on fixed track is going to make this trip faster than SL Phase III ever would've even if the path is no longer straight as the crow flies and looks underwhelmingly indirect on a 2D map. SL Phase III would've been way, way worse than the Transitway crawl.
- Fewest structural underpins possible. Offset, rather than stacked, stations.
- Fewer touches of old infrastructure. 20th c. widened streets, 1960's urban renewal property >>> "Old Boston" 19th c. buildings + poorly documented utilities.
- Less duplication of infrastructure. Re-use outer Boylston platforms + some/all of Tremont tunnel instead of building twice, resorting to ham-fisted geometry to make transfers.
Now let's bring back the new E-Trunk. This new tunnel gives us another way of reaching the Seaport from the south: merely continue the tunnel forward with a wye where you would otherwise turn to hit Boylston. This lets you route traffic from the north and south into the Seaport. I like it as a solution, and will probably integrate it into future thoughts. Under this system you'd essentially be trading one E-Trunk branch for a Seaport branch at Park St.It's less about having a "favorite" routing than maximizing the build odds...because we should've already had this but those bulleted pitfalls above torpedoed the cost. It's a billions-dollar project, but we want a "buildable" digit stuck onto the front.
As above, think efficiency + minimal pitfalls to lift up the Seaport connection to the realm of buildability. Tremont tunnel is 4 tracks...right there you've got grade separation for 2 branches. Can you utilize 2 tracks for the Seaport branch and 2 tracks for the Dudley branch? Or can you plow that 4-track tunnel past Tufts station down Tremont, set up 2 of the tracks to go to the Seaport + Dudley, and leave a 2-track stub to continue west as the "new" E to Back Bay + Prudential. Such that the Huntington tunnel eventually gets extended to become an "alt. spine".
For the record, I was considering one of the two tracks to continue west under Stuart or Marginal, and dedicate the other to Seaport/Dudley traffic.
Look again at the map. Only three lines are actually going into Lechmere from the north. There's a bit of dispatching trouble possible at the new maintenance yards, but at no point would more than three lines ever be sharing tracks:As in my last reply, that's way way too many northern branches to dispatch...and the fact that the non-UR additions are subject to the chaos of hard-to-control grade crossings offset too far from nearest traffic lights means schedule dominoes are going to start falling regularly.
- Urban Ring, Malden to Downtown, Malden to Kendall across the bridge into Sullivan
- Medford Branch, Porter Branch, Malden to Downtown through Lechmere into Downtown.
- Porter Branch, Malden to Kendall and Urban Ring along a short stretch between the maintenance yards and where the Grand Junction splits off.
Now, that's a lot of line interweaving but there's no point at which it becomes a crisis, especially given that two of the lines are completely traffic separated.
How is traffic getting routed at Kenmore then? If it's continuing straight into Back Bay then I don't see the point of any of it. At what point does someone get on the train along the Urban ring to go all the way through Kendall and BU just to get to Park St? If the traffic is bounding back out along the C or D, then reconfiguring the tracks is going to be a monster to allow that. And if we're bounding back out along the B then Kenmore is a terminal station on the Urban Ring? That seems reasonable, but I'd be worried about trains sitting on active B platforms.The Cambridge Ring is absolutely positively essential. Kenmore is the tie-in destination, because the south-half Ring that's going to have to be BRT originates there. Filet service between the Kenmore circuit and Harvard Branch as suitable, but anyone waiting on the westbound platform at Kendall surface station must be able to hit Kenmore on at least every other train. Bio-metropolis over in Cambridge needs to be able to pick up a quick transfer to Longwood; that shouldn't even be a question.
Hence why I didn't include it. There's a dozen ways out from there, all of them reasonably priced and of reasonable ridership to alleviate Red Line congestion at the end, but none of them that are of high importance.Watertown's pretty straightforward. It'd be about 3/4 mile of street-running (though perhaps with reservation'ed platforms) on Arsenal St. at the very tail end...not enough length to kill an inbound schedule to Lechmere. But other than that the routing is very direct and structurally simple (2 shallow duck-unders of Sherman St. and Fresh Pond Pkwy. being the most concrete poured). On cost and buildability it's a good one. It's just not as mission-critical as the biggies. We REALLY REALLY need the Seaport connection, the NW+NE Ring quadrants, Dudley streetcar, the Porter transfer, Needham (for all the well-documented RER-driven reasons), and the E Back Bay relocation. Watertown just doesn't have any oxygen when that's the five-alarm needs list.
Also, I start to get nervous extending any line out from there too far. You quickly start looping back on yourself over absurdly long dispatch distances if you send D or E out there, which you have to if you want to maintain branches as traffic separated.
I absolutely agree that keeping completely traffic-separated branches apart from non-separated branches is critical, but I don't see any way to really do that unless we're limiting ourselves to two northside branches for all time (or biting the bullet and picking out new downtown tunnel locations. An alternative is to send all of the Urban Ring/Malden branches down the Grand Junction, but I don't think that riders there are going to love needing to make odd transfers to get downtown.As I said, that's just too many northern branches. And that's a bad thing if the extras exert any destabilizing effect on a load-bearing branch like the Urban Ring segments. Or Medford, for that matter...that one's got insane potential to completely blow its projections out of the water and require more service. And you can't do much more than is already going up. Lechmere is a pretty snug fit across the street, so there isn't a turnback until Brattle Loop. The problem with having crossing-heavy branches added to the 2 grade separated GLX's and the 2 very nearly grade separated Ring routes is that if one of these extra builds stubs its toes in crossing traffic in Malden it's already started dragging the rest down for several stops before it ever gets the chance to dump out at Brattle. And that's doubly ungood for whatever other branches are running at much denser frequencies when they get tripped up (like two 3-min. headway 3-car trains becoming late because of a 6-min. headway 2-car train, and relative numbers of riders impacted).