I thought that the mantra has been 15 min clockface during rush and 30 otherwise, at least to the major cities, ie Providence, Worcester, Lowell, Lawrence, and Salem, and out to somewhere between 128 and 495 for Fitchburg and Franklin/Foxboro, and that it would all be electric.
That's not the "mantra" at all.
:15 clockface
all hours on Urban Rail schedules [inside-128]
:30 clockface
all hours on Regional Rail schedules [128-to-495]
The leveling of the frequencies from the pronounced peak/off-peak skew of today to an even churn *is* what makes it real RER. And because the sets will be in constant rotation less equipment will be wasted sitting idle in yards between runs and consists will be put more evenly together (i.e. shorter but running more frequent) making them more portable for route reassignment (e.g. ends the dilemma of a train just in from Providence not being able to be re-badged as a Needham outbound because it's 8 cars instead of 5).
Surge slots can and will be used as gap-filler when there's a crowd-swallower slot needing a monster number of cars, but unlike today where that same 8-car set runs two-thirds empty for half its 3:00-7:00pm peak assignment because of one sardine-packed 5:00 on the schedule...a one-off surge slot only has to waste its seats once in the reverse-commute direction before either getting reassigned or broken up in the yard into a smaller RER consist.
And it does
not have to be all-electric. EMU's lower costs considerably, but since electrification will take a long time--especially North--it is implementable with the current diesel push-pull fleet. GO Transit in Toronto is implementing real RER right now with their diesels; it's not waiting for their own electrification effort to kick off. Merely, GO sees its wires as the follow-through for making it financially sustainable rather than any sort of project prerequisite before increasing service levels at all.
That is 16 EMUs an hour in the north, and 16 EMUs in the south. Fairmont taking all of Foxboro and Franklin gives it 8 tph, more than enough to turn it into a defacto HRT line. Lowell and Haverhill already have parallel subway lines, so I don't see a world in which more frequency would be needed, just bigger trains. So, how many additional EMUs will we need?
Again, that is
not what RER is. Frequency and headway consistency over length at all times except when special slot-specific needs arise.
You could make the argument that Riverside, Fitchburg (to 128) and Salem could use more frequent service, although BLX and GL to Chelsea might meet some of that need. How many Amtrak Regionals are going through the tunnel DURING RUSH, the only time access might be restrained? 2? 3? Over 2-3 hrs? If you add two tph on both Salem and Fitchburg, we are up to 21 tph if we can't fit 21 tph in TWO tunnels, then we shouldn't be spending the money.
So while there may be some ability to mix and match on the north end, the Fairmont feeds into one tunnel and Worcester and Providence into the other
Again, "DURING RUSH" is not what RER is about at all. And Amtrak has never been a rush phenomenon because NY-BOS travel times don't neatly correspond to rush to begin with. They run all day.
Please, please look at the Rail Vision FCMB slides or TransitMatters website first for an explanation of what RER is. All of this is explained nice and graphically.
From every design I have seen of the North portals, there are two, one fed by Fitchburg, and one by N/R/haverhill and Lowell. So if they do feed both tunnels, then they will have to split somewhere under. So if everything coming down from Anderson is going to the NEC /Worcester portal( I assume an OL conversion of the reading line-8 tph) and 6 tph coming from Fitchburg going to Fairmont and N/R splitting between them, depending on Amtrak needs. No more than two lines merging. Not rocket science.
They split at fairly shallow level into the portals. Northside's interlockings aren't as constrained as southside's, but since it's a conjoined beast the tunnel capacity limiter ends up the southside interlockings.
No, it's not rocket science. But you're looking 1 mile away from where the real issue is. It's not relevant what TPH you think you can cram into the north portals, because the TPH crammed through the south portals are where total NSRL capacity ends up a wash or slightly worse than the surface terminals and requires you to retain both surface terminals.
So what is left for diesels? The DE (6x day), Fitchburg outer (1-2 tph) outer N/R ( 2 tph) and whatever NH throws at us.
What am I missing? Where is all this demand coming from? Where are all the diesels?
It's not about the diesels. The only lines that are going to remain diesel are the Downeaster, some too-far runs like crossing the Cape Cod Canal every couple hours or taking some RIDOT subsidy to run past Fall River to Newport, and the lines whose electrification drags up the rear (Haverhill probably because of freight clearances, Old Colony because it's not worth it until the Dorchester-Quincy pinch is solved, Littleton-Wachusett because mileage = extra substations). But the ones dragging up the rear will eventually get done.
In the meantime, you've got huge new sources of demand to tap which will require fileting sleek EMU's to the surface, too. Nevermind linear expansion of the system. First amongst all sources of traffic growth is
further implementation of RER scheduling well beyond the initial rollout being proposed. The most we can do now with near-term systemwide capacity is leveling the schedules through interlining. That's great and all for giving inside-128 :15 frequencies on most lines, and inside-495 :30 frequencies on most lines. But it's not perfect because there's a tradeoff of
frequencies + stop density vs.
travel time.
So, take the Franklin Line for example. Forge Park/Foxboro have to be interlined via the Fairmount Line because the NEC can't handle that many branch schedules. So :30 Foxboro frequencies + :30 Forge Park frequencies add up to :15 frequencies to Readville. Great! Except Forge Park, which is longish ~62-65 min. schedule, would have to start making all Fairmount stops for the sake of maintaining the Urban Rail frequencies. This adds 12-13 minutes, putting Forge Park in the 75+ minute range. They get lots more frequencies so it's eminently justifiable, but that's a stinky-long trip. And it calls into question the viability of ever extending to Milford or Woonsocket when that's the starting schedule.
There isn't enough extra slack capacity to pound that schedule back out using supplementals with any singular (surface OR tunnel) southside terminal, so Forge Park ends up being one of the quasi-losers of RER by having its schedule tanked in the name of frequencies. It's the better of the alternatives, but still a stark limitation.
How to fix that? If you had the terminal district capacity, you'd divorce Forge Park from feeding the Fairmount :15 schedules and backfill by other means. Foxboro's OK as-is...it's a pretty brisk schedule making all local stops, but it's only :30 frequencies unto itself. But now you go and change Forge Park to a judicious skip-stop schedule that skips all but the highest-demand Fairmount intermediates so end-to-end is doable (still at :30 headway) in 50 minutes. Fantastic! And you go throw more Readville or 128 short-turn trains at Fairmount Urban Rail to square back up to :15 after vacating the local stops from Forge Park trains.
That's a 33% traffic increase right there because you backfilled with locals the slots that turned express, and absolutely none of that traffic increase gets applied to the base inside-128/outside-128 frequencies which are still :30 or :15 same as they ever were. The increase all goes to improving travel times on the longest-haul schedule. Which is now short enough that Franklin ridership sees a new growth spurt from the schedule change, and you can probably hold the Woonsocket and/or Milford builds to an hour.
As
^above^ example illustrates, frequency and time no longer have to direct-compete with each other under a tight capacity cap like this initial RER rollout. New capacity can get assigned to the benefit of any pressing need, instead of having to be an exercise in canceling out sacrifices. That's why this RER push is 'initial' and why advocates like TransitMatters preach how getting the ops practices in place practically DEMANDS we build the NSRL build as a second act. It's the ability to double-barrel service that allows the system in the future to tackle the frequency vs. timeliness argument without both needing to compete with each other. Thus, the biggest source of traffic increases is not going to be the popular misconception of: blowing out frequencies bigger than Eastern MA's population actually is. It's going to be applied to things like this which live inside the default frequencies but fortify a given line's layer cake of service for the needs of specific audiences...rather than pitting audiences against each other. It's mostly a flexibility thing, but between the lines are a fuckload more trains you have to fit somewhere.
If you saw some of the less-appetizing skip-stop schemes in the FCMB Rail Vision slides, it's palpable just how much those factors are head-to-head with each other on the current system.
There are insane number of examples across the system where service increases can be applied to achieve best of both worlds on frequency/density vs. travel time. North, let's take the Eastern Route. Rockburyport's not bad for schedules. But if Urban Rail goes to Salem we're really going to need that Lynnport public reboot of River Works. We're really going to need that Salem State U. infill stop. You can even make a solid argument for an East Lynn revival spanning Lynn and Swampscott. But can you do all of those beneficial Urban Rail infills without stringing out the Regional Rail schedules on the branches if they're the ones contributing half-and-half to the :15 frequencies. How are we going to run this line if Portsmouth ever comes back? That actually wasn't a bad schedule on the clock when it got re-studied a dozen years ago, and a little stop-skipping on the main might put the Seacoast in 1:10 of Boston. If the terminal district cap were lifted, you could assign more trains strictly with the purpose of bringing Boston closer in
time without worrying about upsetting the apple cart on frequencies.
This is why I want data! Knowing how much of NS traffic goes down to NS under is key. If 10%, then we need a strong NS. If 90%, then nobody on the 2 diesels ph needs to go past Sullivan (yes, F-Line I know they can't turn around there, but BET is 2 min away) and nobody on Fitchburg needs to go past Porter. That leaves NH and DE (if none of the routes electrify, or we don't reduce the frequency of Lowell and Haverhill to make up for the NH runs.) 2-3tph? 4? 6? How many berths will we need? How many jobs can we fit into the CBD? Say we triple ridership. Still should fit. Quadruple? Still should fit, just bigger trains.
You don't get it. It COSTS more to turn outside an ops base. Let's say, since we're trading in way-spurious examples, that you're jealously guarding that penny jar you want to raid from drawbridge SGR to pay for Urban Ring bridges because all money is one big slush fund and blah blah blah (bear with me, everyone else). Well...you're emptying that jar a penny at a time with every :15 and :30 frequency train you cut out at Porter instead of going to the integrated terminal with the staff + cash rooms in the NS building and all of Keolis World Headquarters right there at BET. BTW...that crew room in North Station: those folks can and will run downstairs just as easily when the tunnel is open to stage a crew turnover. All-day, every day...take-a-penny with each Porter-turning train. How efficient an operation are you running doing it this way here...and Sullivan, and elsewhere. Drip, drip, drip.
And this is a universal railroading problem, not T specific. Amtrak's long-distance trains are such cromulent money-losers in part because the routes that don't have big Amtrak facilities at the end of the line cost so damn much more in between-run chores, staff expenses, and ground transportation when a train (e.g. an LD that only runs a few times per week) has to get remotely re-crewed. Freight carriers have to religiously monitor crew shifts for federal hours-of-service limits because if they 'can' the train in the middle of nowhere far from finishing their runs they're bleeding money on taxi fare. Pan Am on the northside is an industry-wide joke for canning its trains so much they lose up to a couple million dollars in unnecessary cab/Uber/Lyft fare per year because they're too lazy to staff up their trains properly enough to finish on time.
It doesn't have to be because of slop-ops that they're losing money. All of the time a crew is out-of-service because you cut them at Porter instead of cycling them along is money wasted. All of the time you're carrying extra cars instead of being able to swap sets in the yard within the turnaround time of a :15 Waltham set is money wasted.
Also..."nobody on Fitchburg needs to go past Porter."
WHAAAAA??? That may be news to everyone coming in who has to use this contraption called the Orange Line. Assloads of people take Orange every day, and it does not run through Porter. You can direct-transfer to it at North Station. It's an exercise in torture to reach on Red. I personally lived in North Cambridge and used Porter for commuting for 9 years. Any morning I had to find my way to Orange I'd try to time it for the Fitchburg freebie instead of fighting my way to DTX. There was always a crowd on the platform.
Green isn't going to substitute it. For one, the Union Branch (E) sees exactly 1/3 the headways of North Station (C+D+E), so the notion you can just shove everybody on there like they'd get fooled by the Foldgers Crystals switch is patently false. As in last post, Harvard + North Cambridge destinations and the 77 are big drivers for GLX reaching Porter. The Fitchburg Line is not and never was a driver for Porter because NS has and always will have minimum 3x the trolley frequencies + Orange. Sullivan CR and Sullivan GL are driven by the humongous bus terminal. It's the same Orange Line there, and NS will always have multiple times the Green frequencies.
While everything electric from the north should fit in the tunnel, the same in not the case for the south. Worcester and Providence will be pumping 9 car EMU crowdswallowers and Acela 2.0 will be filling SS (and OC as well)for a long, long time.
No. Nothing about a "crowdswallower"
vehicle makes it unsuitable for the tunnel. The tunnel as-designed will be able to take a max length push-pull set of Kawasaki bi-levels if hauled by a locomotive that's in electric mode (straight electric or dual-mode).
As before, RER practices mean even-keeled frequencies and keeping as much equipment in rotation as possible. So a lot of rush-hour will be satiated by the higher frequencies. However, in the absence of a traditional peak/off-peak divide and the limitation of running all-local on a lot of routes...there will be need for surge slots. A surge slot
can be a super-size crowd-swallower. But it can also be an express extra since staging expresses is a little harder with the headway requirements. It can be a headway-corrector when there's a service disruption, as it'll be more important to maintain the default headway than follow the paper schedule. And it can be a throttle-up/throttle-down headway corrector on one of the mercifully few lines (Haverhill in particular) that has to swim upstream against freight interference. Any of those surges are easier to stage from the surface terminal, because attempting to pair-match on-the-fly with an instant tunnel slot causes more problems than it solves. And they're easier being able to grab-and-go a trainset on-demand from BET...to NS, not Porter.
I'm sorry for assuming the group reading this would presume that it would go to BET
As am I for assuming it was self-evident a much traffic-heavier system needed to pay basic mind to ops and cost efficiency and not go out of its way to make things harder for itself.
What is driving this crusade for service cuts at Porter, anyway...since money, logistics, transfer destinations, and rapid transit frequencies apparently have zero to do with it?