Friday, February 17, 2017

Good intentions, bad plans, and $7 million wasted

The Auburndale Station is a mess. It is planned for a rebuild, which it sorely needs. But thanks to the MBTA's planning process (which ignored little things like rail operations), that's a bloated mess. There have been two public meetings four years apart, and during that time a plan has been put forward which bakes in bad design and pays no mind to any larger-scale issues on the Worcester Line. It's the T's planning process at its worst, which is saying something.

The villages of Newton developed in the mid-1800s along railroad lines; the concept of the commuter—the term, indeed!—began with the Boston and Albany's "commuted" season fares in the early 1840s. If anyone can lay claim to commuters, it's Newton, even if they now enjoy some of the worst Commuter Rail service in the region. In the 1960s, when the Turnpike was planned, Newton fought a losing battle against it (way too much background here). The old stations were replaced with rickety stairs and narrow platforms, and by the early 1970s, there was minimal train service on the line. Since then, however, the number of passengers on the Worcester Line has grown many-fold (from 600 in 1972 to 16,000 today), yet the line infrastructure generally still dates to the 1960s. With dozens of steps to the platform, these rail stations were inaccessible for anyone with mobility needs, and inhospitable to others. And the single platforms were only accessible by these stairs from a bridge, cut off from the portions of the village centers not cut off by the Turnpike.

Several years ago, the local state representative, Kay Khan, worked with then-Congressman Barney Frank to earmark federal money to build an accessible station at Auburndale. While this goal is laudable and the need is clear,  due to a combination of and overall lack of vision for the line and possibly some incompetence, the team retained by the MBTA specified a project which provides few benefits with a high cost. The Worcester Line should not be fixed piecemeal, but needs an overarching vision, which is currently lacking. Still, this should not be an excuse for the lack of understanding which has led to the current state of this project.

The Newton stations are the only ones on the Worcester Line—and on pretty much any Commuter Rail line in Boston—with single platform on one side of double tracks (on the south side, which is Track 2). As such, the stations have no reverse-peak service: there's enough traffic on the line that trains can't run in and out on the same track at rush hour, so, for instance, there's no inbound train leaving Auburndale between 1:12 and 7:31 in the afternoon. The obvious—and best—solution would be to build a new facility with platforms on both sides of the tracks, although such projects—like the recent station in South Acton—cost about $10 million each.

Building a single platform in situ on the south side would not be much cheaper, since it would still require ADA accessibility which, in the case of Auburndale, would require an overpass and redundant vertical circulation, and elevators cost about $1 million each. The actual platform only costs about $2 million, but getting there costs significantly more. It may be possible to build elevators from the current bridges, but the current sidewalks leading to the bridges are too steep to meet ADA requirements, so additional bridge work would be required. In any case, it makes sense to build a platform on both sides. (There's also the question of building a gauntlet track to allow infrequent wide freight trains to bypass the platform.)

A somewhat cheaper option in Auburndale is to build a single platform on the north side of the railroad (adjacent to Track 1); this is what is proposed. This would be significantly less expensive because it is adjacent to the local street and requires minimal vertical circulation: just a couple of small ramps instead of elevators since the platform would lie only about three feet vertically below the sidewalk. It doesn't solve the reverse-peak issue and still only provides one platform for service, but it at least puts that platform in a much more accessible location. If you only have money for one platform, this makes a lot of sense, with one major caveat: you have to rebuild West Newton and Newtonville on that side as well. If you don't, it's nearly impossible to serve platforms on Track 2 at Auburndale and Track 1 at the other Newtons, and even if you can, it requires an expensive interlocking and signal changes to do so. Without an interlocking, Auburndale would lose all peak commuter service, which is used by 325 passengers per day (the busiest of the Newton stations). With an interlocking, the cost of building the station triples.

So what did the MBTA do? They, of course, proposed to rebuild Auburndale on the north side, and to install an interlocking east of the station—just a mile east of the current CP 11 (see Weston Switch at Dave's glossary)—to allow trains to move from one track to another. Setting aside the operational difficulties of having two interlockings a mile apart and switching trains frequently back and forth, the interlocking—and associated signal changes—costs a lot of money. Here's the cost breakdown they presented:

Site Work 436,138
High Platform 1,733,094
Station Canopies $810,000
Parking Lot Modifications $1,685,750
Track and Interlocking $6,685,750
Station Electrical $179,156

Now, let's break this down in to three parts. The station itself (site work, platform and canopies) costs $3.16 million. The parking lot modifications to create ADA accessible spaces costs another $1.69 million. This accounts for 42% of the total cost of the project. The rest, 58%, is for the interlocking and track and signal work associated with it. This work is entirely unnecessary. First of all, there is already a perfectly good (or at least good enough) interlocking one mile west, so this won't have any operational efficiencies for the rest of the line (and will likely cause operational issues; the project team admitted that they have not modeled the schedule impact of this). Second, the line is likely to need new signals within the next decade, so this would likely be good money thrown after bad: the signals would have to be coordinated with that project, or replaced, and the interlocking is in a sub-optimal location so close to the current switch at CP 11.

Now remember: a north side station works if the other two Newton stations also had north side platforms. And the actual cost to build a platform here costs only $3 million (this is about the right ballpark: at stations like South Acton, for instance, each platform costs $3 million and the vertical circulation costs another $4 million). If you build all three stations, you save $6.7 million by not rebuilding the interlocking, and using CP 6 in Brighton and CP 11 in Weston to move trains back and forth as needed. You also have trains on a long-enough section of track that others can pass without encountering suboptimal signal aspects. (In other words: think of passing a tractor on a country road. If there's a long straightaway with good sightlines, you can easily keep up your speed, change lanes, and make the pass. If there's just a short section, you have to slow down, make sure there is enough room, the tractor may pull to the side of the road, and you pass at a lower speed. This is what a mile between interlockings would entail.)

$6.7 million should be enough to build a north-side platform at West Newton and Newtonville. Newtonville is easy: there's actually an old, low-level platform on the north side which would provide a suitable base for a high level platform, which could be connected to the sidewalk by stairs and short ramps. There is 35 feet between the sidewalk and the edge of the track, plenty for a platform and vertical circulation. West Newton is a bit more difficult: it's only about 700 feet between bridge abutments, and the T prefers to build high level platforms 800 feet long (although an eight-car train is only 680 feet long, and Yawkey Station is that length, with tapered platform ends to accommodate the site). In addition, some excavation would be required to remove the granite blocks on the north side (these were the original supports for Washington Street which, before the Turnpike was built, crossed diagonally) although these might provide a suitable base for a high level platform. But the parking lot already has accessible parking, and there is ample room to build ramps and a platform.

The issue is not that we don't have the money, it's that we're going to spend it in about the most wasteful way possible. The question is how to—and whether we can—reallocate this money. The Auburndale Station has about $3 million of federal dollars earmarked for it, so that likely could not be reallocated. Much of the rest of the money is included in the state's five-year Capital Investment Plan (CIP), a document released by the state. That money could, theoretically, be reallocated, although it would be a political process, and there is, apparently, no guarantee that the money would be reallocated to the other Newton stations (which are not in the current CIP). But here's the rub. There are three ways you can spend $11 million on the Auburndale Station:
  1. Spend ~$4 million on the Auburndale Station, and $7 million on an interlocking which has not yet been modeled and may overall degrade service on the Worcester Line and no guarantee you could provide even the current level of service.
  2. Spend ~$4 million on the Auburndale Station, and the remaining $7 million on similar improvements to West Newton and Newtonville. This would actually improve service on the line (local trains serving high-level platforms would have shorter dwell times, improving accessibility, service speed and reliability) and you could certainly provide the current level of service.
  3. Spend $11 million on the Auburndale Station, but instead of building an interlocking, build platforms on both sides with ADA accessibility. 
The first is wasteful. Either of the other two is a good start towards better service and accessibility in Newton.

Making this cahnge would require the cooperation of MassDOT and the politicians in Newton and elsewhere. There would have to be promises made—perhaps even legislation passed—reallocating the $6.7 million from the interlocking specifically to the West Newton and Newtonville stations. If you build Auburndale and build the interlocking, you waste $6.7 million on the interlocking to build a $4 million station. But if you build Auburndale without the interlocking, you waste $4 million on a barely-usable station. Unless you build two platforms, Auburndale, West Newton and Newtonville are joined at the hip. You can spend $11 million and get an attractive, accessible station at Auburndale, or spend the same $11 million and get three stations for the price of one. 

This process should have never gotten to this point, of course. The project management team is mostly to blame: they ran amok with a design which has become far too expensive and provides little, if any, benefit. In addition, the fact that the MBTA lacks any long-term vision for Commuter Rail or the Worcester Line leads to these piecemeal, wasteful approaches like this. The corridor needs a long-term vision, which is something which should be in the wheelhouse of the Worcester Line Working Group.

Coming back to Auburndale, however, there are two preferable solutions: a two-platform station, or improvements to West Newton and Newtonville. The costs are about the same, and the benefits are much higher than an interlocking you don't need. Mistakes were made. We can either double down on the mistakes—and waste $7 million taxpayer dollars—or we can make the best of the situation, spend the same amount of money, and come away with a lot more to show for it: either a two-platform Auburndale or accessible stations throughout Newton.

This has gone from being an engineering issue to a political one: and this is why we elect political officials.  As we say in Patriots Nation: Do Your Job.


  1. What would it take to build a center platform with a long ramp from the bridge for all these stations? Would you have to chop a lane off the Turnpike to relocate the southern track?

    1. As a resident of Aburndale, I can say that that idea would require either moving the entire turnpike a bit (as unfortunately, there are only 2 lanes to begin with) to make room for moving the southern track further south, or to remove a good amount of earthen wall to make room for the northern side of the track.

      Problem with the first idea is that would disrupt traffic for a year or more. The second idea would be an issue because it would also involve interference with the support of several bridges, which currently are built with their primary structures on the earthen walls. It's also worth noting, that's ignoring the movement of the streets and the houses built along them.

    2. I just realised a third alternative: Move the northern track north and the southern track south. That would be the best, but would require making room by completely removing existing track, halting movement on the commuter rail for a month.

    3. It's doable without something drastic like moving the Turnpike, but there are some issues.

      First of all, the distance from the Turnpike to the edge of the rail is 26 feet, so about 20 feet to the edge of the train. This is plenty wide enough for a platform (at its widest point, the platform at Yawkey is just 15 feet wide, and it has a higher passenger load than Auburndale). So putting a platform there is not an issue.

      What may be an issue is the fact that the current station at Auburndale is on the inside of a curve, and high level platforms have to be built on tangents (or on the outsides of curves) to allow boarding at the doors at the ends of the cars. The Auburndale Station would have to be shifted under the Auburn Street bridge to allow the station to be build on a tangent. That's doable.

      So how do you get to the station? The issue with bridge ramps is that the sidewalks on both the Woodland Road and Auburn Street bridges are too steep and have too much of a cross-slope to allow ADA accessibility. So even if you built ramps, it would not provide accessible access without modifying the approaches to the bridges. An up-and-over solution like Yawkey or South Action would work; both of those stations cost about the same as is budgeted for Auburndale for two platforms and elevators. You could still have stairs from Woodland and Auburn, and an overpass, probably in the vicinity of Ash Street.

      It may be possible to do this with elevators on only one side (ramps are possible, too, but actually aren't that much less expensive and aren't preferred by the T). The parking lot is about 10 feet above the railroad, so a ramp to clear over the railroad on the north side would be significantly shorter than on the south side (and there's more real estate to build a ramp on the north side). This would mean a mobility-impaired individual would use the ramp on the north side to cross a bridge above the tracks, then use an elevator to descend to the south side platform. This would probably save some, but not a lot, of money.

      There is also the issue of wide freight rights. CSX very infrequently runs wide freight on this corridor which would impact the platform edge. There are several options to mitigate this, including a gauntlet track which would allow freight cars to shift slightly off the platform (this is planned in Winchester) or building a mini-high platform with a retractable edge or negotiating a reduction in trackage rights with CSX (unlikely without a huge "ransom").

      As opposed to the anonymous poster, none of this would require moving bridges, streets or tracks; it could all be built within existing real estate. (I have no idea what the "movement of the streets and the houses built along them" even means.) This is totally doable, and probably for the same ballpark cost as is currently programmed for the project.

      The main hurdles are political.

    4. At the Worcester Line Working group, we discussed and explored the idea of a center island platform. Although Ari is generally correct with regards to the real estate directly at the station, the problem is the LEAD into the station on both sides. The tracks need to start 'separating' well before the station, and the bridge abutments are the constraining factor for that geometry. To the west, the tracks go under the Mass Pike, and then both the Woodland Road and Auburn Street underpasses are not wide enough. Theoretically those could all be fixed...but that raises the price exponentially.

    5. Excellent point, Dave. This is especially true given express trains at the station. The leads at Littleton, for example, are quite short, but all trains stop there, so it doesn't matter that there's a speed restriction. There is currently a 55 mph speed restriction on the Auburndale curve, so you certainly wouldn't want to make it any more severe than that. And there's certainly not room given the bridge abutments. Plus which, platforms are (relatively) cheap. Vertical circulation is expensive.

  2. Ari:

    Thanks for such an excellent description. Well presented. Clear. Understandable, even if the substance boggles the mind.

  3. Distance from the Pike isn't an absolute blocker at any of the 3 stations. You need a 50 ft. wide ROW to fit 2 tracks sandwiched around a recommended 24 ft. island. Minimum allowable platform width for an island is 12 ft., though that is discouraged by the Mass Architectural Board anywhere it's not absolutely pinned in with no give.

    Island platform would be technically feasible at Auburndale, but because the station is on a slight curve it would be costly to do in its current location. The platform slabs would need high-precision custom cutting, and the tracks would have to spread apart much earlier on the straightaways bookending the station requiring more landscaping. To keep an island build within acceptable cost you'd most likely need to shift the entire station east between the Auburn St. and Comm Ave. block where the tracks are tangent and the platform slabs can be generic prefab. But that pulls the station far away from the parking lot, spaces it too close to West Newton, and spaces it too far from Riverside (which would get a commuter rail platform adjacent to the Green Line platform if inside-128 Indigo Line service were initiated on the Worcester Line). Since the inbound track already has decent separation from Pike noise and road spray, side platforms in the current location end up working best.

    Newtonville's an easy one to do up as a center island. Plenty of room to track spread around a tangent island in the current location if the earthen embankment along Washington St. were shaved back a little. That ancient 19th century embankment is starting to have significant erosion issues, so it needs a taller rebuilt retaining wall to begin with as an eventual state-of-repair item. Center island would also provide much better protection from Pike road spray than the current side platform and its inadequate sound wall. Also, by pulling the platform further away from the Pike you can extend the full westbound breakdown lane another third of a mile to the Star Market air rights and have it rest right up on a guardrail against the bridge abutments...a little bit of operational improvement on the road side of the fence.

    West Newton's a TBD. Ideally you'd want an island here for the same Pike-proximity reasons of it being safer and more comfortable for passengers on the platform to have more separation from the road. Probably doable because that stretch also has an erosion-prone embankment on the Washington side that's easy to shave, but overall a much tighter fit than Newtonville because of the closely-spaced overhead bridges. Reconfiguration of the Pike Exit 17 "Circle of Death" rotary has been studied for eons but never advanced by MassHighway, so the final answer for West Newton's design could well involve giving that rotary reconfig long-overdue traction as a transpo two-fer. Pedestrian accessibility is always going to be awful at that stop with the as-is rotary, so might as well shoot for a multimodal package deal to fix all that ails that location.

    1. West Newton and Pike Exit 17 are different interchanges. West Newton is Exit 16.