and a goofy meme [they] posted suggesting that a commuter rail connection might be more valuable than rapid transit via the Seaport.
I don't know if I disagree with Alon about this principle in general, but I think I disagree with regard to Boston. The thing about airports is that your mode of transit -- car, rideshare, bus, train, peoplemover -- needs to bring you to your
terminal, not just the airport campus. The front doors of Terminals A and E are something like 2000 feet apart, comparable to Copley <> Arlington. And Terminal C is about 1000 feet (Boston City Hall <> Haymarket) away from a direct line connecting A <> E.
Most
North American airports with rail connections also have
peoplemovers to handle this particular flavor of the "last mile" problem, including SFO, JFK, EWR (Newark), ATL, DFW, DEN, and MIA. Those that don't are either smaller with single terminals (Providence, Midway, Portland), or have rail stations directly at the terminals, the prime (and only?) example being
Philadelphia.
The gold standard is probably Chicago O'Hare, where the subway station is something like 800 feet to Terminals 1 and 3, and less than 400 feet to Terminal 2. (But, even at O'Hare, Terminal 5 requires transferring to the peoplemover.) But O'Hare's core terminal area is more compact than Logan:
(Apologies for the atypical orientation on the Logan image -- the only way I could fit it all into the frame.)
My point is that, even in systems with a putative direct rail connection to the airport, it actually usually is a 2SR.
In this respect, Boston's current state is actually, arguably, better than average, since the [nominally] rapid transit line runs directly from the downtown core transit station to the terminals. The Silver Line actually most closely resembles the JFK Airtrain: you don't bring rail to the peoplemover, you bring the peoplemover to the rail.
That all being said... it is also clear to me that a peoplemover or otherwise broadly "light rail" (yeah I'm gonna call peoplemovers a flavor of light rail,
come at me bro) is often the right choice for serving the front door of individual terminals, especially compared to out-and-out mainline rail. Boston again offers an extreme example of why, but SFO is more visually obvious:
The SFO peoplemover stations sit on a circle with a radius of about 450 feet.
The North American mainline rail network's minimum curve radius is 410 feet, meaning the SFO peoplemover's ROW would be very tight (and slow) for mainline trains to navigate.
SFO's terminals are, at least, nicely arranged. Logan's are not. Drawing lines where a) the curve radius remains within that 410' max, b) platforms are long enough for mainline trains (see below), and c) all four-and-a-half terminals [including both sides of Terminal B's U] have station access is a fun little geometry challenge. Below is my best effort, and remember that those curves are substantially speed reducing:
Now, one note here is that our only example (that I can think of) of mainline rail serving individual terminals is Philadelphia. But, recall that SEPTA lives that EMU life, and often runs shorter trains. On the Airport Line, they must run them quite short, as the Terminals E & F platform is less than
200 feet long. (And I mean, damn, even our GLX platforms are longer than that, jeez.) If we shorten our platforms to 450' (a bit longer than our Red Line platforms), the problem becomes a bit more tractable, but it's still a tough fit:
My point is that airports are tight space and terminals are often not laid out linearly, which makes for a bad fit for mainline rail. If you are going to need to build a last mile peoplemover anyway, then keep the mainline rail where it goes best and build a bigger better last mile system.
(All of which is to say, I suggest we build a third Ted Williams bore and run something lighter than mainline rail directly from South Station to the individual terminals.)
