Originally posted as a comment on The Transport Politic. I'm sure I'll write about the Silver Line again.
The Silver Line has so many problems.
First off, there's this bizarre notion that people from Roxbury and Mattapan need to get to the airport. All the time, forever. It's probably not the case. I'm not one to make brash generalizations, but here's one: the people who generally use the airport are folks from affluent and/or student-infested parts of Boston. At higher rates, anyway, than the Roxbury-Mattapan. For this community, access to downtown Boston and better-than-bus service is probably paramount. It would make much more sense to take this $114m and build a spur of the blue line in to the actual airport. Build a loop to the terminals. Heck, build it in to Central Parking, where there are elevated walkways to all the terminals. Eliminate the shuttle bus (which I once tried to take before a long weekend and it was packed to the gills with college students with dozens more waiting to board).
View Logan Blue Line Spur in a larger map
Second, the Silver Line from South Station to the airport is very slow. The tunnel is fast enough—the speed is slow but it's grade separated, so it works. The problems arise once the buses reach the surface. The then cross D Street and proceed to drop the trolley poles and switch to petroleum. The route then takes a convoluted backtrack loop back towards South Station, across D Street, through several lights, before the bus can finally turn down in to the tunnel. $15b and they couldn't build a ramp straight to the airport, which would have been rapid.
Then, the airport. Since Boston is the furthest northeast city in the country, Logan has never developed in to a hub airport. Thus, no one has ever built one big terminal. So the airport is a hodgepodge of terminals, each with an access road which gets choked with traffic. Sure, the buses can sometimes bypass these queues, but they still have to go through the loops in to terminals A, B, C, and E (with two stops in B—Terminal D doesn't really exist). Ten of fifteen minutes later, they loop back in to the tunnel and a mess of roads before looping back to South Station.
I'm very glad that "Phase III" has been all but nixed by the Feds. A $1.4b tunnel would not fix the main issue that trips are scheduled to take 38 minutes to go from South Station to the airport and back. As the crow flies, this is just over a mile. A Blue Line spur to the airport, with stops at Maverick and Aquarium, would tie in to the rest of the system with trip times of maybe eight minutes, tops, with faster loading and more capacity, to boot.
But that's really actually not the worst part. Again, I'll start by explaining that I am happy that the $1.4b tunnel from pretty much nowhere to pretty much nowhere with a couple sharp curves thrown in was not funded. From Boylston Station on the Green Line five blocks south is a disused tunnel for streetcars, and the plan was to basically decimate that tunnel to build it to bus loading gauge. Here's the thing—the tunnel ties in to the Green Line—to a four track alignment to Park Street Station—and is grade separated, underground!, at the junction. Basically, if you turn back one line of the green line at Park Street, you could add in another without increasing capacity on the congested central subway.
And this tunnel would tie in splendidly with light rail down towards Mattipan. You build a new portal at Tremont and Oak and cut diagonally across the Turnpike and NEC from Shawmut to Washington. Washington Street is wide enough for trolley cars to not interfere with parked cars by occupying the center lanes. (Washington Street once had the elevated above it.) Stations in the center of the tracks, proof of payment ticketing perhaps, and you don't impede traffic significantly, which could pass stopped trains.
Getting through Roxbury might be fun—but you could use the old elevated right-of-way for one or both tracks of a light rail line. Or tunnel underneath if you had the dollars. From there, Warren Street has two lanes each way plus a wide median, so congestion wouldn't be a major issue) to Quincy Street, where you'd then have to build on a two-lanes-plus parking street to Blue Hill Avenue for less than half a mile. Cut parking to one side of the street and build wide lanes and you'd be fine.
Then you hit Blue Hill Avenue. The Avenue is three lanes each way plus a wide median all the way down to Mattapan, where you could connect with the High Speed Line. Trolley tracks could be in a separate median (like Comm Av or Beacon Street in Brighton and Brookline). In fact, this was the case in the past.
View Boylston-Roxbury-Mattapan in a larger map
For a heck of a lot less than 1.4b. (Portland can build a streetcar for $25m/mile. This is 7.5 miles. Make that $30, throw in $50 for a new portal at Tremont and $25m for a bridge across the Pike, and the cost is $300m.