Regional Rail (RUR) & North-South Rail Link (NSRL)

[TheRatmeister]

275 ft platforms are not long enough by today's construction standards. For light rail, GLX's specs called for 300 ft, so they're either built this way (Lechmere is about 350) or have provisions to extend them. For heavy rail, Malden Center's platform is about 410 ft, and Fields Corner 430 ft. Not a magnitude of difference, but it reduces the difference from NSRL to some extent.

The 275ft number is based on real systems in places like Vancouver and Copenhagen that get high capacity through high frequency automated service rather than fewer, larger trains. (And when there's a lot of people making a lot of smaller journeys that involve transfers, I think this is 100% the way to go.) You absolutely can have trains that size.

Disregarding the question of whether tunneling under downtown immediately implies $$$ when the ROW is very wide, if we assume that's true, then tunneling through Longwood will be even tougher. The hospital buildings are almost as tall and roads are much narrower, sometimes 35' between sidewalks.

Tunneling in downtown is expensive because your worksites are downtown. Simple as that. Once you get outside downtown there are more surface streets, and there are wider surface streets, making transporting equipment and materials easier and less disruptive.

but any Urban Ring proposal that completely avoids Kendall is likely a no-go.

Quite simply, I don't agree, and there's one table that summarizes why quite well:

A station that is dominated by commuter traffic to office districts is not a reliable source of ridership going forward. Given the choice of Kendall, an office district that is struggling to win back people who transitioned to WFH, and Harvard, a thriving student hub, I'd certainly pick the latter. Building a system tailor specifically to the needs of office working commuters is part of how we got into the mess we're in, it's time to change course.

I also think you're underselling NSRL's usefulness. It's almost never about allowing Providence-Lowell OSRs, but rather, improving efficiency (with through-running trains), access to multiple downtown destinations and subway connections, transfers for regional travels, and perhaps most important of all, allowing much higher frequencies on all trunks such as <=10 min Fairmount, 15-30 min Providence, and even allowing many urban segments (Fairmount Line, Newtons, Waltham, North Shore) to receive rapid-transit-level service.

How many of these things actually require the NSRL? There are no shortage of rail lines around the world that operate 15 minute or better frequencies out of city center terminus stations, providing rapid-transit like service to urban areas. London and its many termini is probably the best example of this, although there are plenty of other cities with similar situations. Yes, throughrunning is better, but is it so much better that it justifies the $18bn pricetag, and likely sacrificing non-radial transit? Again, we've now built a system further dependent on commuters, not for people making regular, everyday journies.

 

[Teban54]

Tunneling in downtown is expensive because your worksites are downtown. Simple as that. Once you get outside downtown there are more surface streets, and there are wider surface streets, making transporting equipment and materials easier and less disruptive.

TBM does not disrupt surface traffic except at station sites, and NSRL has a massive advantage in the number of stations.

Saying there are more and wider surface streets along the Urban Ring than in downtown is also a blanket statement. North, Central-Aquarium and South Stations all have Surface Rd which is relatively wide (not to mention Greenway space can be used), plus a rather robust road network with many cross streets and parallel streets. On the other hand, LMA almost certainly does NOT have many wide surface streets. The Allston-Brookline stations (Coolidge Corner, Harvard Ave (B), and Union Sq Allston) only have wide streets on the crossroads but not along the corridor itself (Harvard Ave), and the surrounding road network (especially parallel streets) is not nearly as robust as downtown for diverting traffic, especially given there had been a recent discussion on whether there needs to be more roads in this area.

Quite simply, I don't agree, and there's one table that summarizes why quite well:
View attachment 45667

I think there has to be an error in your spreadsheet. I compared 2019 Jan-Oct and 2023 Jan-Oct using the Gated Station Entries spreadsheet: (full list here)

There's virtually no difference between the ridership composition of all three stations, despite Kendall's ridership composition being so different from the other two.

A station that is dominated by commuter traffic to office districts is not a reliable source of ridership going forward. Given the choice of Kendall, an office district that is struggling to win back people who transitioned to WFH, and Harvard, a thriving student hub, I'd certainly pick the latter. Building a system tailor specifically to the needs of office working commuters is part of how we got into the mess we're in, it's time to change course.

While your first sentence may hold for downtown stations (though even there, GC and State do well), Kendall is not just an office district (or is not one at all). It also consists of many biomedical labs, whose needs for in-person operations and experiments remain; and MIT, whose students obviously still need to go to campus, and many of whose staff still work in-person in some capacity. In fact, both Kendall and LMA are exactly the kind of industries that should fare well in transit ridership post-Covid.

Specifically attributing Harvard's ridership to students also ignores MIT's student population. Also, students (undergrad and grad) generally have a tendency to stay as close to campus as possible (within their budget), especially at the level of Harvard and MIT, so they have lower needs for commute, especially on a subway. While an argument can certainly be made on increasing the mobility of commuting students so that they can live further away, this argument applies to MIT even more than Harvard due to its current lackluster transit - in fact, I know many MIT students living closer to Lechmere than Kendall.

If Harvard's recovery rate was actually 90%, a much more plausible explanation would have been from Cambridge/Watertown/Belmont/Allston residents via bus transfers and from recreational activities at Harvard Square. But regardless, its recovery rate is 51%, same as Kendall.

How many of these things actually require the NSRL? There are no shortage of rail lines around the world that operate 15 minute or better frequencies out of city center terminus stations, providing rapid-transit like service to urban areas. London and its many termini is probably the best example of this, although there are plenty of other cities with similar situations. Yes, throughrunning is better, but is it so much better that it justifies the $18bn pricetag, and likely sacrificing non-radial transit? Again, we've now built a system further dependent on commuters, not for people making regular, everyday journies.

Except that commuter rail is much more successful than subways and buses at bringing back its riders: (source)

Commuter rail, meanwhile, stands in contrast to both subways and buses. Its average ridership is the lowest of the three primary modes, but it has displayed undoubted success at attracting passengers since the lowest of the lows.

As recently as March 2022, the commuter rail system had a lower rate of ridership recovery than both MBTA subway and buses. And since the turn of the year [2023], it's been effectively tied with buses for the best rebound, attracting about 68% as many passengers in February 2023 as it did in February 2020.

Improving commuter/regional rail, including NSRL, is far from having a singular focus on commuter ridership. The shift to a "regional rail"-like schedule post-Covid has certainly helped improve ridership, and they show that you can - and you should - have a quality regional rail system even if you want to attract non-commuters.

While I'm not well-versed in regional rail operations to comment on whether 15-min frequencies for all lines out of South Station (plus <=10 for Fairmount) is feasible, I'm not sure if London is the best example, as they just built their own version of NSRL in Crossrail.

Also, why are commutes not counted as "regular, everyday journies"? If anything, that's still the biggest source of regular, everyday journeys. I certainly go to work more often than I go to supermarkets or dine out at restaurants in other neighborhoods, and I'm sure many people still do, especially those whose companies are calling them back in person or pushing a hybrid schedule.

 

[bigeman312]

I agree with all this, 100%.

Ironically though, in the long term, I don’t believe there should be a continuous Cambridge <> Sullivan <> Chelsea <> Airport service. Sullivan <> Airport via Chelsea is lengthy and roundabout enough that Orange + Blue is gonna be too competitive.

Chelsea <> Kendall is worthwhile for the OSR to Kendall, and Chelsea <> Airport for the same reason. If you want to have a Sullivan <> Airport service pinging back and forth, that seems reasonable to pick up local riders transferring to Orange/Blue, but I don’t see that as a load-bearing service, nor do I see it redirecting any Orange + Blue transfers away from State.

I agree that bus routing that covers too much of the ring in one route shouldn’t be the goal, but I am envisioning longer routing than you are. I don’t believe that these routes require end-to-end riders to be successful. I’m envisioning the target to aim for to be roughly six routes that each cover roughly 1/3 of the ring. What are your thoughts on that? What I mean in practice is that in a world with a bus-based Urban Ring, the target routes would be something like:

• Orange (e.g. Sullivan) <-> Red (e.g. South Station)
• Red (e.g. Kendall) <-> Blue (e.g. Airport)
• Green (e.g. Kenmore) <-> Orange (e.g. Sullivan)
• Orange (e.g. Ruggles) <-> Red (e.g. Kendall)
• Red (e.g. JFK/UMass) <-> Green (e.g. Kenmore)
• Blue (Airport) <-> Orange (e.g. Ruggles)

(Mods feel free to move this discussion to Urban Ring)

 

[Riverside]

I don’t believe that these routes require end-to-end riders to be successful.

In general, I agree with you. But (very very theoretically) once a bus travels further than its longest-distance rider from the start of the route, then it is doing the work that another bus could be doing, at the expense of its own reliability. The longer the route, the more opportunity for delays.

There are, of course, many many many reasons why it's not as simple as that. You obviously can't create separate services for every pair of endpoints; the location of feasible turnback and layover points has an impact, and you also need to balance efficiency of equipment use in consideration of demand.

So, yes, you don't need end-to-end riders, but I think the lack thereof has to trigger additional scrutiny for a route.

I’m envisioning the target to aim for to be roughly six routes that each cover roughly 1/3 of the ring. What are your thoughts on that?

In general, this is indeed my rule of thumb when thinking about circumferential designs. Once you go more than about 1/3rd of the way around a (perfect/idealized) circle, you are traveling a further distance than it would be to go directly to the center of the circle and then back out. So, in general the farthest I see a rider taking a circumferential service is about 1/3rd of the way around the circle. (More elaboration here.)

But... the Urban Ring corridor is not a perfect circle.

Here is a perfect circle, centered on Park St, that roughly hits key Urban Ring nodes, including Airport, Sullivan, and Nubian.

But that circle misses certain key nodes: it runs wide of Kendall, runs short of reaching Longwood, misses almost all of Southie, to say nothing of the Seaport, and literally does not even make it past the Mystic, leaving Chelsea untouched.

To stretch that circle to reach Chelsea, Sullivan, and Longwood, we get an oval (though, annoyingly, this particular oval I've drawn hits most of the nodes pretty well, but still only grazes LMA):

And, as you know, the actual Urban Ring corridor distorts this even further, with bulges out to the Logan Terminals, Everett, and probably Andrew, and tighter hewing inward to Kendall:

One of the consequences of this distortion is that not all segments of the Urban Ring Corridor are created equal. As I've mentioned, the Sullivan <> Airport segment has to diverge so much from the "perfect circle" that it (likely) would be just as fast to continue the Orange + Blue transfer at State.

On the other hand, the physical alignment of Sullivan <> Kendall <> LMA runs in nearly a straight line:

Meaning that a circumferential service from Sullivan/East Somerville to Longwood would potentially be competitive with the OSR radial journey via the Green Line.

The Chelsea segment underperforms for circumferential journeys, while the Kendall segment overperforms.

So, to the specific routes you propose:

• Orange (e.g. Sullivan) <-> Red (e.g. South Station)
  • I would split into a Sullivan <> Logan Terminals route and the current SL3
  • See note below
• Red (e.g. Kendall) <-> Blue (e.g. Airport)
  • I would truncate this in Chelsea, for a Kendall <> Chelsea route
  • See note below
• Green (e.g. Kenmore) <-> Orange (e.g. Sullivan)
  • As mentioned above, this route makes sense and probably could even be extended south to Longwood proper
• Orange (e.g. Ruggles) <-> Red (e.g. Kendall)
  • Yup, this is solid
• Red (e.g. JFK/UMass) <-> Green (e.g. Kenmore)
  • This quadrant is always such a headache. Kenmore <> JFK/UMass seems a little long, so I might split into Kenmore <> Nubian (see note) and LMA <> JFK/UMass, but I think there are a lot of moving pieces here
• Blue (Airport) <-> Orange (e.g. Ruggles)
  • I think this one is tricky. I've been sorely tempted by a limited stop Nubian <> Logan service designed for a maximally speedy connection to Logan and to the Seaport
  • But just in terms of distance, Washington St LRT + Blue is actually a little shorter between Nubian and Airport station than a route via Logan and the TWT would be; the question will then be whether the Washington LRT would be fast enough to compete
  • But... if we measure to Logan Airport itself as the destination, the calculus changes a lot -- the distance becomes basically the same between the two, but the route via TWT would be a 1SR as opposed to the 3SR (Wash + Blue + shuttle) via downtown
  • So, I would split this into SL3/SL1, plus a Ruggles/Nubian <> Seaport route, potentially with an extension through the TWT but with a question mark next to it

NOTE: if we stop thinking of them as "Urban Ring" routes, several of the routes I've laid out have some obvious potential extensions:

• SL3 could be extended to Everett or even into Malden, or have a separate Malden <> Logan Terminals with significant interlining between Chelsea and Airport
• A Kendall <> Chelsea route could extended along the 116 corridor into Revere. This route would be a bit long, and probably we could come up with a faster Revere <> Kendall alignment via the Tobin Bridge
• Alternatively, a Kendall <> Chelsea route could be rerouted into downtown Chelsea proper for its terminus, e.g. via Everett Ave
• A Kenmore <> Nubian route is already being proposed: the T28. Just as the Malden <> Logan or Kendall <> Revere routes I've suggested would actually be radial routes based out of Logan and Kendall, the T28 will become a radial route based out of Logan, but like the others will also fill a circumferential gap for part of its route

(A lot of this is very similar to the initial proposal for Urban Ring Phase 1, though I've suggested some changes.)

To summarize my suggestions (with a couple extra tossed in at the end):

Terminus	via	Terminus	optional extension to
South Station	Seaport	Chelsea	Everett
Logan Terminals	Chelsea	Sullivan	Malden (branch)
Kendall	Sullivan	Chelsea	Downtown Chelsea or Revere
Sullivan	Kendall	Kenmore	LMA
Kendall	LMA	Ruggles	Nubian
Kenmore	LMA	Nubian	Dorchester
LMA	Nubian	JFK/UMass
Nubian/Ruggles	[something]	Seaport	Logan Terminals
Revere	Tobin Bridge	Kendall
Harvard	Allston	LMA

 

[Teban54]

MBTA has released a Request for Information (RFI) for the Regional Rail Future Innovative Operating Contract, with an explicit mention of electrification:

COMMBUYS - Bid Solicitation

The main potential roles which could be procured separately or together as part of a Regional Rail contract or contracts would be:

• Passenger service operations
• Legacy rolling stock maintenance
• New rolling stock provision, maintenance and finance/leasing
• Infrastructure management (including O&M and dispatch)
• Delivery, finance, and long-term maintenance of major capital works including electrification

Questions for interested parties also include:

4. Has your organization previously been involved in a major project to decarbonize or electrify a new or existing railroad?

Sample interview questions (for different roles):

• Are you willing to assume risk delivering or integrating electrification work?
• What are you views on a P3 approach that includes rolling stock and the necessary electrification infrastructure and support facilities (like RSSOM)? In that case how would you ensure the input of the train operator is cost effectively incorporated in the design/performance requirements?
• How practical/ cost effective would a move to predictive maintenance be with the legacy Diesel fleet?

 

[ritchiew]

MBTA has released a Request for Information (RFI) for the Regional Rail Future Innovative Operating Contract, with an explicit mention of electrification:

COMMBUYS - Bid Solicitation

Questions for interested parties also include:

Sample interview questions (for different roles):

This RFI looks like the start of finding who will take over operating commuter rail when Keolis's contract ends in 2026. Is that what this is, or something slightly different?

Also, I might be reading this wrong, but some of this looks kind of bad. This looks like the continued outsourcing of everything, while giving up on building inhouse expertise and ability in the MBTA. That lack of institutional capacity is routinely cited as a main cause for lots of transit problem in America, like outrageous infrastructure costs. (This is a major theme in the Transit Costs Project). I think it was @F-Line to Dudley that was pointing out that electrification is the perfect project to start building up that institutional capacity, for a bunch of reasons: there is some inhouse expertise to build on; it's a large but repetitive project, allowing a team to build, iterate, and improve; it is a (relatively) straightforward problem with limited options and known solutions. But everything in this RFI looks like MBTA really wants to just contract the problem out.

 

[F-Line to Dudley]

This RFI looks like the start of finding who will take over operating commuter rail when Keolis's contract ends in 2026. Is that what this is, or something slightly different?

It's putting feelers out ahead of the operator renewal bid for a substantial change in responsibilities in the operator if Regional Rail/Rail Vision is implemented. And an RFI instead of RFP because they need to know if the expansion of responsibilities is going to kill off competitive interest in the eventual bid or jack the prices up more than they can bear.

Also, I might be reading this wrong, but some of this looks kind of bad. This looks like the continued outsourcing of everything, while giving up on building inhouse expertise and ability in the MBTA. That lack of institutional capacity is routinely cited as a main cause for lots of transit problem in America, like outrageous infrastructure costs. (This is a major theme in the Transit Costs Project). I think it was @F-Line to Dudley that was pointing out that electrification is the perfect project to start building up that institutional capacity, for a bunch of reasons: there is some inhouse expertise to build on; it's a large but repetitive project, allowing a team to build, iterate, and improve; it is a (relatively) straightforward problem with limited options and known solutions. But everything in this RFI looks like MBTA really wants to just contract the problem out.

Now isn't the time to bring in-house arguably the only competently-operating division of the agency. Yes, they're the largest commuter rail operator on the continent who still outsources ops, and that's sub-ideal. No, we don't want to change that until the rapid transit and bus divisions have been rehabilitated out of crisis. For at least one more 10-year term it's going to be prudent to keep CR ops outsourced. Keolis has shown they're at least competent at what they do; the T in-house, not so much.

Construction I agree needs more in-house oversight. Not necessarily full-on "MBTA Construction", just lots more professional in-house project managers so individual projects are getting watched before they accrue cost blowouts. We're already getting killed on the costs of station renovations which are going for double, triple, quadruple what they could be, and level boarding is going to trigger dozens more of such instances. The consultants are bleeding us dry there. And the outsourcing of electrification to consultants-managing-consultants-managing-contractors already has very bad omens in this hemisphere by how mismanaged Caltrain's retrofit was and how much it's overcomplicating GO Transit's more sprawling electrification plans. They have to strive for more efficiencies before the iterative process of Regional Rail renovations/retrofittings begins.

 

[ritchiew]

Now isn't the time to bring in-house arguably the only competently-operating division of the agency. Yes, they're the largest commuter rail operator on the continent who still outsources ops, and that's sub-ideal. No, we don't want to change that until the rapid transit and bus divisions have been rehabilitated out of crisis. For at least one more 10-year term it's going to be prudent to keep CR ops outsourced. Keolis has shown they're at least competent at what they do; the T in-house, not so much.

Construction I agree needs more in-house oversight. Not necessarily full-on "MBTA Construction", just lots more professional in-house project managers so individual projects are getting watched before they accrue cost blowouts. We're already getting killed on the costs of station renovations which are going for double, triple, quadruple what they could be, and level boarding is going to trigger dozens more of such instances. The consultants are bleeding us dry there. And the outsourcing of electrification to consultants-managing-consultants-managing-contractors already has very bad omens in this hemisphere by how mismanaged Caltrain's retrofit was and how much it's overcomplicating GO Transit's more sprawling electrification plans. They have to strive for more efficiencies before the iterative process of Regional Rail renovations/retrofittings begins.

Yeah, I agree. I should have been a little clearer. I wasn't saying we should totally ditch outsourcing commuter rail ops. (I mean, we should. Like you say, this is sub-optimal, but not best to change now).

I was specifically concerned about the construction parts of the RFI. MBTA is looking for companies interested in "Delivery, finance, and long-term maintenance of major capital works including electrification." And asking companies "Are you willing to assume risk delivering or integrating electrification work?" Those look like significant expansions of what private companies would be doing with CR, and looks like a step in the wrong direction.

It is especially worrying that the MBTA is looking for companies to assume financial risk of construction projects. This is one of the practices the Transit Costs Project points to specifically as a big factor in rising construction costs. Those researchers found that trying to push the cost of risks onto private companies leads to higher initial bids, but does basically nothing to actually control eventual costs.

 

[TheRatmeister]

Hold on tight we're jumping threads!

I also don't know that I would expect, for example, SCR trains to make stops at Hyde Park or a Cummins Hwy/Mount Hope infill; if the express tracks are there, probably they'd use them, I would think?

My assumption would be peak hours express, off peak local for a consistent service all day. A (Fairly pessimistic, TM Regional Rail calls for way more off peak trains) service pattern for Cummins/Hyde Park/Readville might look like:

Line	Peak Trains	Off-peak trains
Franklin	2	1
Foxboro	2	1
Providence	2 (+2 Express)	2
New Bedford	0 (2 Express)	1
Fall River	0 (2 Express)	1
Total	6 TPH	6 TPH

Even if you ditch the Providence peak trains and SCR off-peak trains, that's still 15 minute headways all day, plus the Fairmount Line with however many TPH you want to shove down there. I don't think you could really aim for much better with bus lanes or an LRT/Streetcar, and OLX would probably cap out around 10TPH so not a ton of room for improvement there either.

 

[Teban54]

Hold on tight we're jumping threads!

My assumption would be peak hours express, off peak local for a consistent service all day. A service pattern for Cummins/Hyde Park/Readville might look like:

Line	Peak Trains	Off-peak trains
Franklin	2	1
Foxboro	2	1
Providence	2 (+2 Express)	2
New Bedford	0 (2 Express)	1
Fall River	0 (2 Express)	1
Total	6 TPH	6 TPH

A problem is that in a Regional Rail world, rerouting Franklin (and thus Foxboro) trains to Fairmount Line is a near certainty, as the line's flat junctions affect operations too much. The question is how many trains will go there. I think people like @F-Line to Dudley have been pretty insistent that almost all Franklin trains will need to be rerouted.

Ideally, this should have little to do with service at Hyde Park (if Franklin trains are gone, it means their slots will be replaced by a Providence/SCR train, meaning Hyde Park will still see the same or even higher numbers of trains). But the real question is what's the maximum capacity that SW Corridor can handle (don't forget Amtrak) -- which I don't know, but it's definitely low enough for Franklin-via-Fairmount to be on the table. Perhaps someone else can chime in?

 

[TheRatmeister]

A problem is that in a Regional Rail world, rerouting Franklin (and thus Foxboro) trains to Fairmount Line is a near certainty, as the line's flat junctions affect operations too much.

Readville already needs to be rebuilt pretty badly, is a flying junction really too much to ask for? The flyover is already there, you'd just need to move the station a bit to the north.

But the real question is what's the maximum capacity that SW Corridor can handle (don't forget Amtrak)

Assuming quad-track between Readville and Forest Hills, since the alignment seems to be wide enough until you get very close to Forest Hills Station itself, and the existing triple track north of Forest Hills, Amtrak can have the middle track of the 3-track section all to itself, with the other two being used for CR. For the quad-track area a standard middle track express, outer track local setup can be used. That essentially means the question is: "How many TPH can you put through a double tracked railway?" To which the answer is: At least 12TPH, which is what the RER B in Paris runs through Port Royal station. The Elizabeth Line in London runs even more trains through Stratford with 16TPH at peak times, the North River tunnels between NYC and NJ sees up to 24TPH, and the Munich S-Bahn runs 28 TPH through Donnersbergerbrücke station at peak times. The last one is absolutely absurd even by European standards and probably not within reach of the NEC, but even the 12TPH the RER-B runs would be enough for the hypothetical service pattern I threw together.

 

[Riverside]

tl;dr: if Hyde Park can be quad-tracked, then yes I think that cumulative "show-up-and-go" ("SUAG") Purple Line frequencies should be achievable there one way or another. But that's a medium-sized "if".

Readville already needs to be rebuilt pretty badly, is a flying junction really too much to ask for? The flyover is already there, you'd just need to move the station a bit to the north.

I think this could be doable, but keep in mind it will take something like 2000 feet to descend from the flyover to ground level (based on the existing NEC <> Franklin leg). In fact, it'll probably be longer, because you need to fit an 800' platform in there, which will need to be level -- I'm not 100% sure about this, but I think that inclined platforms such as the Franklin Line platform at Readville are considered inaccessible and really not built anymore unless absolutely necessary. So, it seems doable but likely very costly.

Assuming quad-track between Readville and Forest Hills, since the alignment seems to be wide enough until you get very close to Forest Hills Station itself,

This is where the question mark for Hyde Park is. The ROW at Hyde Park looks about 70 feet wide, into which 3 tracks (12' each, so 36') and two platforms (one at about 9', the other at 16') and one ramp (9') have been fit. Four tracks will require 48', plus platforms at the now-apparently-standard 12'... which puts you at 72 feet already. The ROW widens at the River St bridge, so you can probably replace the ramp with some elevators, and fit stairs in... somewhere. But that's still pretty tight. (My measurements are imprecise and inconclusive, but I think are sufficient to highlight the potential concern here.)

And the slot for the fourth track is located on the western side of the ROW (look at the location of the catenary posts), pretty much plowing directly through the southbound platform. Keeping all other tracks in place, it looks like there might be just enough room to fit in 12' of fourth track and 12' of platform, but it is very constrained.

Also keep in mind that Amtrak uses this stretch (and is aiming for quad-tracking to enhance their ability) to overtake local trains at near-maximum speeds. That means that the inner tracks (for Amtrak expresses) need to stay in place, without any speed reducing curves around new platforms. Unfortunately, that means that we're limited in our ability to use the eastern side of the ROW, which is much less constrained due to the parking lot.

It looks like a similar story at Cummins Hwy, though Metropolitan Ave looks a bit more promising.

Anyway, that's where the question mark is, IMO. If you can fit 4 tracks + 2 side platforms, great! I agree that getting high frequencies will be relatively trivial. If you can't fit them in, then it's a more difficult conversation.

 

[TheRatmeister]

This is where the question mark for Hyde Park is. The ROW at Hyde Park looks about 70 feet wide, into which 3 tracks (12' each, so 36') and two platforms (one at about 9', the other at 16') and one ramp (9') have been fit. Four tracks will require 48', plus platforms at the now-apparently-standard 12'... which puts you at 72 feet already. The ROW widens at the River St bridge, so you can probably replace the ramp with some elevators, and fit stairs in... somewhere. But that's still pretty tight. (My measurements are imprecise and inconclusive, but I think are sufficient to highlight the potential concern here.)

I think the options here would either be reconfigure the area to add the new track on the eastern side, and remove 24 Factory St, or add the new track on the western side without reconfiguring and destroy 30 and 34 Business St and eat into the reservation on the side a bit as well. In terms of destroying buildings the first is better, in terms of existing infrastucture the second is better. If Amtrak wants 4 tracks they will need to pick one at some point, but honestly even the two residential buildings option isn't that bad in terms of required eminent domain.

I think this could be doable, but keep in mind it will take something like 2000 feet to descend from the flyover to ground level (based on the existing NEC <> Franklin leg). In fact, it'll probably be longer, because you need to fit an 800' platform in there, which will need to be level -- I'm not 100% sure about this, but I think that inclined platforms such as the Franklin Line platform at Readville are considered inaccessible and really not built anymore unless absolutely necessary. So, it seems doable but likely very costly.

The minimum height of a bridge over a railway is 22 feet, so we'll assume that's the height difference we need to cover. 4% grades are not uncommon on even high speed railways, so we'll use that as the assumed grade. (AFAIK no freight trains run on this section of the NEC so we don't need to consider them here.) This gives us a descent distance of 550ft. which wouldn't even push the southern end of the side platforms north of Milton St. Even with a very conservative 2% grade that's still only 1100 ft.

 

[F-Line to Dudley]

The minimum height of a bridge over a railway is 22 feet, so we'll assume that's the height difference we need to cover. 4% grades are not uncommon on even high speed railways, so we'll use that as the assumed grade. (AFAIK no freight trains run on this section of the NEC so we don't need to consider them here.) This gives us a descent distance of 550ft. which wouldn't even push the southern end of the side platforms north of Milton St. Even with a very conservative 2% grade that's still only 1100 ft.

FRA max grade is 2% and they only grant exemptions in cases of extreme hardship (like a fixed bridge replacement for a movable bridge). There's no way you'd get a waiver here, so picking and choosing grades is not an option in this country.

 

[Riverside]

I think the options here would either be reconfigure the area to add the new track on the eastern side, and remove 24 Factory St,

Right, but that's what I'm saying won't work: everywhere else except the station, the space to add the fourth track is on the western side, where there's even an empty slot under the catenary towers/arches/I-can't-remember-the-term; if you add the track on the eastern side, then all of the tracks need to do one "jog" over to reach it, which kills the speed benefit.

or add the new track on the western side without reconfiguring and destroy 30 and 34 Business St and eat into the reservation on the side a bit as well. In terms of destroying buildings the first is better, in terms of existing infrastucture the second is better. If Amtrak wants 4 tracks they will need to pick one at some point, but honestly even the two residential buildings option isn't that bad in terms of required eminent domain.

We're in agreement that there's a real possibility that 4 tracks + station would require eminent domain. But in terms of "if Amtrak wants 4 tracks, they will need to pick one at some point," they won't need to pick one, since the third option would be to close the station. I'm pretty sure that would also be the least expensive option as well, so I'm sure it'll look favorable to certain parties.

The minimum height of a bridge over a railway is 22 feet, so we'll assume that's the height difference we need to cover. 4% grades are not uncommon on even high speed railways, so we'll use that as the assumed grade. (AFAIK no freight trains run on this section of the NEC so we don't need to consider them here.) This gives us a descent distance of 550ft. which wouldn't even push the southern end of the side platforms north of Milton St. Even with a very conservative 2% grade that's still only 1100 ft.

I'm not sure how you can fit 1100' of running distance between the existing flyover and Milton St.

I'm not personally super invested in building out a design here -- I think it's potentially doable, I think it'll be complicated if it is doable, and it's not super clear to me that it's needed.

 

[TheRatmeister]

since the third option would be to close the station

I can tell you right now that won't fly. Case in point: Heath St. If the MBTA can't rid themselves of a street-running light rail service that's slower than the 39 bus on the same route, I have a hard time seeing how they will get rid of a regional rail service that is more than twice as fast as the competing bus service.

FRA max grade is 2% and they only grant exemptions in cases of extreme hardship (like a fixed bridge replacement for a movable bridge). There's no way you'd get a waiver here, so picking and choosing grades is not an option in this country.

Fair enough, I should have known better than to expect reasonable rules from the FRA.

I'm not personally super invested in building out a design here -- I think it's potentially doable, I think it'll be complicated if it is doable, and it's not super clear to me that it's needed.

Ultimately Hyde Park does need some form of high speed radial transit service, and Readville will need to be rebuilt. The 32 is slow and overcrowded, and Readville has only low level platforms, one of which is not even flat, some waiting shelters of a quality best described as "Questionable", and some station signage about 40 years out of date. I don't think either of those things are in dispute. So in my mind the question of "New transit line or frequent CR service" is a simple cost comparison between a flyover at Readville, which as you pointed out would likely mean moving the Milton St bridge up towards Westinghouse Plaza, and the construction of a new transit route, either an OLX parallel to the NEC or a light rail line using some mixture of Hyde Park ave where it's wide enough and a dedicated alignment where it's not. I have a hard time believing either of those options are cheaper than a single road bridge and railway incline.

 

[HenryAlan]

But in terms of "if Amtrak wants 4 tracks, they will need to pick one at some point," they won't need to pick one, since the third option would be to close the station. I'm pretty sure that would also be the least expensive option as well, so I'm sure it'll look favorable to certain parties.

That's an almost impossible sell, though I agree it will be one of the studied options. Hyde Park and Rozzie folks are still smarting from the eliminations of Mt Hope and Clarendon/Metropolitan stations. To get rid of yet another station on the stretch between Forest Hills and Readville is just not credible. If anything, the others need to come back, or perhaps a compromise station at Cummins Hwy. RUR should mean more stops, not fewer.

 

[F-Line to Dudley]

That's an almost impossible sell, though I agree it will be one of the studied options. Hyde Park and Rozzie folks are still smarting from the eliminations of Mt Hope and Clarendon/Metropolitan stations. To get rid of yet another station on the stretch between Forest Hills and Readville is just not credible. If anything, the others need to come back, or perhaps a compromise station at Cummins Hwy. RUR should mean more stops, not fewer.

Fairmount's well within the walkshed of the NEC stop and is getting sharply higher frequencies, though...so it's not true that Cleary Sq.'s train transit shares are on a one-way trip to oblivion. They're better than they've ever been, with :15 service yet to come. South Coast Rail Phase 2 would've skip-stopped HP all the time anyway so the future was never looking bright on the NEC side of the neighborhood.

Mt. Hope is not one they're "smarting" over. It was very lightly patronized and was sited with very poor access to the street grid.

 

[Tallguy]

tl;dr: if Hyde Park can be quad-tracked, then yes I think that cumulative "show-up-and-go" ("SUAG") Purple Line frequencies should be achievable there one way or another. But that's a medium-sized "if".

I think this could be doable, but keep in mind it will take something like 2000 feet to descend from the flyover to ground level (based on the existing NEC <> Franklin leg). In fact, it'll probably be longer, because you need to fit an 800' platform in there, which will need to be level -- I'm not 100% sure about this, but I think that inclined platforms such as the Franklin Line platform at Readville are considered inaccessible and really not built anymore unless absolutely necessary. So, it seems doable but likely very costly.

This is where the question mark for Hyde Park is. The ROW at Hyde Park looks about 70 feet wide, into which 3 tracks (12' each, so 36') and two platforms (one at about 9', the other at 16') and one ramp (9') have been fit. Four tracks will require 48', plus platforms at the now-apparently-standard 12'... which puts you at 72 feet already. The ROW widens at the River St bridge, so you can probably replace the ramp with some elevators, and fit stairs in... somewhere. But that's still pretty tight. (My measurements are imprecise and inconclusive, but I think are sufficient to highlight the potential concern here.)

And the slot for the fourth track is located on the western side of the ROW (look at the location of the catenary posts), pretty much plowing directly through the southbound platform. Keeping all other tracks in place, it looks like there might be just enough room to fit in 12' of fourth track and 12' of platform, but it is very constrained.

Also keep in mind that Amtrak uses this stretch (and is aiming for quad-tracking to enhance their ability) to overtake local trains at near-maximum speeds. That means that the inner tracks (for Amtrak expresses) need to stay in place, without any speed reducing curves around new platforms. Unfortunately, that means that we're limited in our ability to use the eastern side of the ROW, which is much less constrained due to the parking lot.

It looks like a similar story at Cummins Hwy, though Metropolitan Ave looks a bit more promising.

Anyway, that's where the question mark is, IMO. If you can fit 4 tracks + 2 side platforms, great! I agree that getting high frequencies will be relatively trivial. If you can't fit them in, then it's a more difficult conversation.

A platform can have a 1% grade, so an 800ft platform can be 8ft lower on one end.

 

[Teban54]

We're in agreement that there's a real possibility that 4 tracks + station would require eminent domain. But in terms of "if Amtrak wants 4 tracks, they will need to pick one at some point," they won't need to pick one, since the third option would be to close the station. I'm pretty sure that would also be the least expensive option as well, so I'm sure it'll look favorable to certain parties.

I can tell you right now that won't fly. Case in point: Heath St. If the MBTA can't rid themselves of a street-running light rail service that's slower than the 39 bus on the same route, I have a hard time seeing how they will get rid of a regional rail service that is more than twice as fast as the competing bus service.

The elephant in the room: On the surface, Hyde Park station is very close to Fairmount station. In a world where Hyde Park is closed, it can still be claimed that the neighborhood sees high frequency regional rail via Fairmount. This in turns presents an even stronger argument for closing Hyde Park for an uninformed decision maker, precisely due to the additional costs for quad-tracking it.

(Another case in point: Nubian. The MBTA absolutely has precedent in getting rid of a rail service that is more than twice as fast as the competing bus service, during the same era as Heath St. It's worth noting that ~40% of Hyde Park residents are black, which reminds us of Nubian. I think Heath St only illustrates a case where the incumbent (VA Hospital) has strong political power, which does not seem to describe Hyde Park, and it doesn't even describe Jamaica Plain.)

Of course, the big question is: How many people at Hyde Park specifically seek the Orange Line corridor, such as Roslindale, Forest Hills, LMA and Back Bay?

Spoiler: Analysis of demand from Hyde Park

I used two sources to analyze where Hyde Park residents go to: Census OnTheMap, and BNRD Origin-Destination Dashboard.

Census OnTheMap (where residents work)
I drew the following polygon as the analysis area, which captures neighborhoods near Hyde Park and Fairmount stations. (FYI, the region has 7,699 primary job takers in 2.201 mi^2 for a density of 3,497 jobs/mi^2, way less dense than the urban neighborhoods I've examined recently, so the success of the 32 is quite anomalous -- even though the total number of residents is still 3/4 of Chelsea's.)

Now let's look at where they work. Switching from "primary jobs" to "all private jobs" (as locations of public jobs are often not accurately reflected):

The top destinations are:

1. LMA South (south of Longwood Ave): 342 jobs (4.8%)
2. Financial District (east of Washington St and south of Milk St, contains South Station): 242 jobs (3.4%)
3. Haymarket-State (north of Milk St): 138 jobs (1.9%)
4. LMA North (north of Longwood Ave, contains Beth Israel East but otherwise mostly schools): 134 jobs (1.9%)
5. BU Medical Campus: 119 jobs (1.7%)
6. MGH: 116 jobs (1.6%)
7. Islington to University Park (mostly suburban industrial parks): 94 jobs (1.3%)
8. Hyde Park West (between SW Corridor and Fairmount ROW): 88 jobs (1.2%)
9. Back Bay (south of Boylston St) and Prudential: 87 jobs (1.2%)
10. Kendall: 73 jobs (1.0%)
11. Harvard: 72 jobs (1.0%)

Grouping them into broader regions, and comparing their accessibility from Orange Line, NEC and Fairmount:

• LMA: 476 jobs, OL ≈ NEC >> Fairmount
• Downtown (Shawmut Peninsula): 496 jobs, accessibility is a bit fine-grained
  • 358 jobs at FiDi and MGH are similar for all three alignments (MGH via a Red Line transfer), with NEC weakly favored for its speed
  • 138 jobs at Haymarket-State: OL > NEC > Fairmount
• Cambridge: 145 jobs, similar for all three alignments (via a Red Line transfer), with NEC weakly favored for its speed
  • If an Urban Ring connects to Kendall and/or Harvard, OL and NEC gain an advantage due to transferring at Ruggles
• BUMC: 119 jobs, Fairmount >= OL >> NEC
  • Urban Ring can reduce difference between them
• Back Bay: 87 jobs, NEC ≈ OL > Fairmount

Overall, 701 jobs have a clear preference either OL or NEC over Fairmount. A majority of them are in LMA, which is almost as popular as the entire downtown combined (a smaller number come from Haymarket and Back Bay). In contrast, only 119 jobs arguably weakly prefer Fairmount over OL/NEC (BUMC), and even that depends heavily on quality of bus transfers.

(Comparing OL and NEC is a much harder question. The 138 jobs at Haymarket-State would obviously prefer OL. Most of the remaining jobs can be reached in slightly shorter time by NEC, but OL may still offer higher confidence in service levels and better fare/infra integration with rapid transit. The Red Line jobs are slightly closer to OL, but NEC's time savings will likely cancel that out.)

---

BNRD Origin-Destination Dashboard (where residents go for all purposes)
The BNRD dashboard covers trip data for all origins and destinations in Fall 2019. While the data itself is more powerful, it comes with some nuances that makes it often harder to use in practice:

• Less flexible customization: In this case, selecting "Hyde Park" as the origin automatically includes Readville
• Overwhelmed by short trips: Because it covers all trip purposes and all modes (including driving, walking and biking), for any origin, the highest destinations will always be the same neighborhood and adjacent ones (likely for grocery, errands, etc)
• Worse visualization

Nevertheless, I managed to create a visualization of all its data with origin set as Hyde Park (which includes Readville). You can view the map here:

Note: The points are centers of regions from the original BNRD dashboard, not the precise locations that people go to. For example, the dot near Nubian, named "Roxbury Lower Roxbury", contains the entire region east of OL Ruggles station. Also, the dots are colored by "Destination Geography", which is loosely based on neighborhoods, not by transit lines.

It appears that travel demand from Hyde Park bifurcates into two corridors -- which are precisely:

• The NEC corridor: Roslindale, Forest Hills, LMA, Ruggles-Nubian
• The Fairmount Line corridor: Mattapan, Dorchester, Grove Hall

Considering that the job destinations data did not show nearly as much demand in Mattapan and Dorchester, we can assume they're recreational destinations. (Indeed, I was able to verify that 52% of trips to Mattapan and Dorchester are "home-based non-work", whereas that number drops to 35% for Fenway/LMA, 37% for downtown, and 26% for Cambridge.)

By my approximate counts, 18224 trips would favor OL/NEC, and 14540 trips would favor Fairmount. (To make my life easier, I only counted "geographies" that seemingly have substantial demand, counted the entire Roslindale/West Rox as favoring OL, the entire South Boston as favoring Fairmount, and only did manual splits for downtown and Roxbury.)

Overall, while Hyde Park's demand for Fairmount Line is substantial (slightly leaning towards non-commutes), there's just as much or even higher demand to stay on the NE Corridor (slightly leaning towards commutes). This means having access to both Fairmount and NEC/OL is important.

The question then becomes: Will decision makers (whether it's Amtrak, the T, or someone else) be aware of this fact, and is it important enough to justify the higher costs of quad-tracking Hyde Park? Or will they be completely comfortable with using Fairmount station as an excuse to close Hyde Park station when situation warrants?