Monday, September 12, 2016

A single letter costs the T $2–3 million every year

Every night in Downtown Boston at about 12:45 a.m., a procedure, in theory, occurs to allow passengers to transfer between trains downtown and not miss the last train. (This dance is called "East-West"; the name probably goes back decades.) Here's how it should work (note that this is from an operations standpoint; passengers transfer as they normally would):
  1. The final Green Line trains from Lechmere, Boston College, Cleveland Circle, Riverside and Heath Street arrive at Park Street. 
  2. The last southbound Orange Line train waits at State Street for the last inbound Blue Line train.
  3. Once it arrives, the Blue Line train continues to Bowdoin, loops, and waits at Government Center. The Orange Line train proceeds south to Downtown Crossing.
  4. The last Alewife Red Line train leaves Downtown Crossing when this Orange Line train arrives and runs to Park.
  5. Passengers at Park transfer between Red and Green Line trains. Once this occurs, these trains are released, and a domino effect takes place.
  6. When the Ashmont-bound Red Line train gets to Downtown Crossing, the Orange Line trains waiting there are released. (There's no guaranteed last connection for Braintree passengers.)
  7. When the northbound Orange Line train gets to State, the Blue Line train there is released. There is a second meet (which is not necessary) between this train and the Lechmere Car at North Station.
  8. This is what the last train ballet should look like (thanks
    to Mark Ebuña for the screen grabs). These trains would
    remain stationary for more than 20 minutes. And that's
    on a good night.
  9. As these trains propagate out through the system, 56 "w" trip buses (the schedule notation of "w" means that a given bus will wait for the last train, although a few schedules use other letters) wait for transfers before making their last trips outbound, completing the domino effect.
The rail portion of this ballet, again in theory, should take about 8 minutes. The last trains out of Park Street are scheduled out between 12:45 and 12:53 (the later times because four Green Line trains have to all leave in succession on a single track). The system can then be shut for the night, leaving a bit more than three hours for track maintenance before the first trains the next morning.

Unfortunately, in practice, that's not how it works. As Marc Ebuña tweet-stormed recently, it takes a whole lot longer. And this costs the T a lot of money.

The last train connection is not guaranteed for passengers to Heath Street (who can take the 39 bus, which is held for connections at Back Bay), but it is guaranteed for Lechmere. Since there's no layover at Heath Street (since the Arborway terminus was abandoned), these trains have to turn back in to layover at Lechmere. The last train to Heath Street leaves Park at 12:30, arrives Heath at 12:47, and turns back to Park, with a scheduled arrival of 1:06. (In 2007, this train was scheduled 10 minutes earlier, and the "w" note was not present as recently as 2013, although I believe the T has guaranteed these schedules before then.) And this train is given the "w" notation, so that while every other train should be—if they're on schedule—ready to depart at 12:45, they wait for another 21 minutes before making this connection. And if the Lechmere train is late? The trains still wait. On September 4, for example, the Red Line waited 40 minutes.

This letter costs the T at least $3 million per year. The "w" notation reads:
"Last trips wait at some stations, primarily downtown, for connecting
service. Departure times are approximate."
This is entirely unnecessary. Earlier Lechmere trains are just as able to make the connections. There's a train scheduled to arrive Park at 12:41. If this train were the "w" train, it could drop passengers at Park and continue to Lechmere; any later train could pick up any passenger waiting, but connections for arriving passengers on such a late train would not be guaranteed, other than for buses meeting this train at Lechmere. Or these trains could be operated as non-revenue services, and the 39 bus, which connects inbound with trains at Copley, could provide this service. Other than a few late riders inbound on the E Line, no passengers would be adversely impacted, while every other passenger on the system waiting at least 20 extra minutes (the earliest the last Red Line has left Park Street in the past 30 days has been 1:09, the average has been 1:19—thanks for the coding from @MBTAinfo) would benefit. I'm usually not one to advocate for earlier service, but in this case, either publish a later, more truthful schedule, or run the service on time.

Then there are the costs, which cascade very quickly through the system, since the single Lechmere trip which operates late causes trains on every other line, and 56 bus trips, to all experience delays of at least 20 minutes. But the operators still get paid (overtime, in fact) and the power stays on and the inspectors keep the stations open and the operations staff stays on duty until the last trains pull in. It costs nearly as much money to keep a train stationary as it does to run it, and with overtime, it may cost more. In 2014, the T reported that a bus cost $178 per hour to operate, a heavy rail car $240 and a light rail car $264. Giving them the benefit of the doubt that they're operating single car green line trains, the cost per hour of four light rail cars (B, C, D and Mattapan), 30 heavy rail cars (five trains at six cars each) and 56 buses comes to $18,224 per hour, or $304 per minute. If the Lechmere car causes a 21 minute delay (as scheduled), this costs $6,384 per day, or $2.33 million per year. In fact, the average delay is more on the order of 34 minutes, which costs $3.77 million. This assumes that all delays are caused by the late Lechmere car; if we attribute 20% of the delays to other causes, there is still a direct operating cost of 1.9 to 3 million dollars per year. (These costs are likely even higher now.)

Two to three million dollars. All because of a "w" on the schedule.

But it gets worse. The T has precious little time between the end of service and the first trains the next morning; most lines aren't scheduled to be cleared for powering down or work until about 1:30, leaving only about three hours and thirty minutes until service starts in the morning. A 20 to 30 minute delay accounts for 10 to 15% of this time, meaning work crews have to wait for this unnecessary delay before performing maintenance.

Then there are the passengers. If you take the last train, the schedule—and any real time data—will show it coming at a certain time. But you'll either wind up standing on a platform for 20 to 30 extra minutes, or sitting on the train downtown for that amount of time. There is probably significant ridership loss from people who know how long the wait takes, and choose another mode. This fare revenue is probably minimal in relation to the operating costs, but certainly not zero, but the impact to passengers is more drastic. If we assume just 10 passengers per rail car and two additional passengers on each bus who don't transfer from a train), it amounts to 450 passengers each inconvenienced by 20 to 30 minutes. That adds up to 150 to 225 hours per night, or 50 to 75 thousand hours per year.

The FMCB's response to a budget gap has been to push privatization, which is not guaranteed to fill any such gap, but will draw the ire of the unions and potentially degrade service. Yet various measures which this page has noted have fallen upon deaf ears. There's a lot of money to be saved at the T. There's a lot of very low-hanging fruit. (Like publishing a set of schedules without a "w" for the 12:47 departure from Heath Street. And who reads paper schedules, anyway?) This problem would be very easy to fix: the next schedules would be amended with a different note for the E Line, perhaps "x: last trip making connection downtown departs Heath Street at 12:25."

Making that change would go a long way towards paying for real, actual overnight service.

[Thanks to James Jay for noticing this, Marc Ebuña for burning the midnight oil, @MBTAinfo for the code and Stefan! for the maps.)

Wednesday, September 7, 2016

How MassDOT stacks the deck: Red-Blue edition

The Red-Blue connector is probably the biggest bang-for-your-buck piece of rail infrastructure in the Commonwealth of Massachusetts. With 1300 feet (¼ of a mile) of new subway, it would both provide a much better connection between East Boston, the airport and and the Red Line and serve as a major core capacity project. From the south, the Red Line, at rush hour, operates at peak capacity through downtown to Charles; as it drops passengers at South Station, Downtown Crossing and Park, it takes on transferring passengers and the load stays high. At Charles, however, there are many more destinations than origins, and demand drops. Right now, all Blue Line passengers destined for Kendall or beyond are forced on to the Green or Orange line and the Red Line at this high-utilization point. The Red-Blue connector would allow them to bypass this downtown congestion, reducing the strain on the near- or at-capacity central portion of the subway network. (Oh, and it would also allow a rethinking of Cambridge Street, which is incredibly dangerous for anyone not driving a car. But it has a pretty median.)

This page, in the past, has suggested that it may be cheaper to build an elevated Red-Blue connector, and also cast doubt (twice!) on the MassDOT's cost estimates. Their claim is that it would cost $750 million to make the extension; which is a cost per mile of $3 billion. This per-mile cost is double the cost of recent tunnel projects in Seattle and San Francisco (where, as you may be aware, they have earthquakes) and even more than the Second Avenue Subway in New York. It's a completely outlandish number.

And this is entirely by design.

The state is required to plan the Red-Blue connector, but they're not actually required to build it. Because MassDOT is, at some levels, a morass of incompetence (see Extension, Green Line), they operate under the assumption that nothing new should ever be built, even if there are dramatic improvements to the overall transportation network. Remember, these are the same people who look at ridership growth and declare it "basically flat." But not only do they want to do as little as possible, they actively stack the deck against their designs to come in so costly that they don't make sense to build. This is the idea: inflate the cost so much that it would not make any sense to build. It's deceitful. It's duplicitous. And at MassDOT, it's standard operating procedure.


(On the other hand, when MassDOT—or MassHighway—wants to build something, like the outlandish mutil-tunneling of rail lines in Dorchester to add highway capacity to the Southeast Expressway, they don't bother to put forth a cost estimate. Or remember when we didn't add a lane to 128 because it was going to cost too much? Yeah, me neither.)

Here's how it's done. As we discussed, there is an existing tunnel to Joy Street which was used until the 1950s to move East Boston cars to the old Bennett Street Yards in Harvard Square for heavy maintenance. The obvious solution is to use as much of this tunnel as possible—both to minimize digging in the street and to minimize disruption to current service—yet the state's two alternatives don't use it at all. Instead, with minimal justification, they propose a half-mile-long deep bore tunnel 50 feet below grade, tying in with the existing tunnel just west of Government Center. Using tunnel boring machines (TBM) makes sense for tunnels of any length, as the impact to the surface is significantly less. It is also, for longer tunnels, significantly cheaper than cut-and-cover methods. It's fine to have that as one alternative—there are certainly advantages to using a TBM—but the fact that the alternative analysis only mentions TBMs makes it, well, not really an alternatives analysis at all.

The benefits of a TBM, however, only accrue for longer tunnels. The marginal cost of an extra foot of TBM tunnel is relatively low, but the initial cost is very high. A cut-and-cover tunnel here would require 1300 feet surface impact. Using a TBM would require less, but only slightly. Why? Because you still have to dig launch and recovery boxes for the TBM, and where the tunnel needs to be wider for stations or crossovers, it has to be dug out. Considering the substrate in Boston (mud and clay) a TBM would have to build concrete rings as it digs, and any stations or crossovers between the tunnels would have to be dug out separately. And while the total disruption would be somewhat less than a cut-and-cover tunnel, the disruption would be more spread out and extend much further, from Charles Circle to or beyond Government Center, rather than from Charles to Joy Street. Utilities would be affected in either scheme, and it's possible that fewer utilities would be affected by a more-contained scheme between Charles Circle and Joy Street.

So in addition to carving up nearly as much street space, and over a longer distance, you'd also incur the cost of using a tunnel boring machine (they're not cheap). You'd be building nearly a mile of new tunnel, while only incurring the benefit of about a quarter of that. And the costs are therefore much higher. This only makes sense if you don't actually want to ever build anything.

MassDOT's plan would also build tail tracks under Charles Circle beyond the station. Tail tracks are important: it allows a terminal to continue to operate at full capacity even if a train is incapacitated: it can be shoved in to the tail track and out of the way until the end of service (or until service levels are decreased). However, they take up a good deal of space. This is less of an issue if there is extra space (like there is at Forest Hills, Wonderland or Oak Grove) or at Alewife, where the line was originally built to extend to Arlington (the tail tracks actually do cross the border). But in downtown Boston, the tail tracks require significant extra tunneling under Charles Circle, which is expensive and disruptive.

A different, more outside-the-box option is to create a "pocket track" before the final station. This serves the same purpose as the tail track—train storage near the end of the line—but rather than two long tails, it is in the middle of the two tracks short of the station. All this requires is that the tunnel be built wider here (the same width as the platform to the west requires) for this staging track. Additionally, by utilizing the existing layout at Bowdoin (with, perhaps, some modification for longer trains inbound, or even converting the station to outbound service only), the line would retain the utility of the loop there, which is eliminated in both of the state's alternatives. While this does result in slower operation in and out of the station, it allows redundancy for turning trains: if there is any congestion or another issue at Bowdoin, some or all trains can be turned temporarily at Government Center to maintain service on the rest of the line. Eliminating this loop eliminates any such redundancy. A pocket track and the retention of the loop are nearly impossible with a bored tunnel but with cut-and-cover simply requires a somewhat wider dig.

By requiring long-term construction closure of the line west of Government Center, the state's plan would also require new construction of a terminal station there. Government Center does have a crossover to its east, but it is a single crossover, which would not be able to handle the rush hour Blue Line schedule. This would require a new double crossover to be installed in what is a narrow section of tunnel. Utilizing the existing tunnel past Bowdoin would preclude this extra cost, as trains could continue to loop there during construction. There would be no disruption when the extension opened—not even a weekend shutdown to tie in new tracks. The same can't be said for the State's scheme.




Here's a quick rundown on the major elements required to build the state's version of the Red-Blue Connector:

  • A launch box for the TBM
  • Cut-and-cover tail tracks west of Charles Station
  • Main access to the Charles Station (planned 50' below grade)
  • Charles Station, proposed as a sequentially excavated cavern but with no explanation of how that will be done in the fill-and-clay substrate in the area.
  • Emergency egress from Charles Station
  • Crossovers east of Charles Station
  • New Bowdoin Station (alternative 2 only)
  • Modification of existing Bowdoin Station and trackage to serve as ventillation (alternative 1 only)
  • Receiving shaft for Bowdoin station
  • Cut-and-cover track for connection to existing track at Government Center
  • New crossover west of Government Center to allow it to serve as a terminal station
  • Total cut-and-cover of approximately 800 feet, assuming the Charles Station can be built below grade (I'm skeptical).
And here's a rundown of the elements required for a cut-and-cover tunnel from Joy Street to Charles Circle:
  • Main access to Charles Station (20' below grade)
  • Charles Station cut-and-cover
  • Secondary access to Charles Station (possible because it would require only 20' of vertical circulation, rather than 50')
  • Cut-and-cover crossovers and pocket track
  • Cut-and-cover connection to existing tail tracks at Joy Street
Here's my total-guess cost estimate for the cut-and-cover costs (and I think many of these are quite high):

1. Utility Relocation: 0.25 miles at $100m/mi = $25m
2. Cut-and-cover tunnel (mostly 40' wide): 0.25 mi at $600m/mi = $150m
3. Rail systems: 0.25 mi at $100m/mi = $25m
4. Egress, NFP130, etc: 0.25 mi at $200m/mi = $50m
5. Station, 1 at $50m = $50m
6. Street rebuilding, 0.33 mi at $75m/mi = $25m

This totals to $375 million, or half of what the state's plan would cost. There's no way to know much the state thinks it would cost, because they didn't bother to analyze this alternative as part of their alternatives analysis.

Now, the state did address the difference in cost between a bored tunnel and a cut-and-cover tunnel. Well, sort of. The draft environmental impact report has the clause:
The resulting total cost (direct plus offsets) to construct a cut-and -cover tunnel shell is about 1.2 times the cost of the mined tunnel method. This differential may slightly decrease when the balance of construction scope (e.g., station components common to both Build Alternatives) is considered. Based on this relative cost differential and the associated environmental and social impacts, schemes utilizing mining methods were selected for further development and evaluation. 
Uh, this isn't really how an alternatives analysis works. In a complex construction project, 20% is basically a margin of error. As we've seen in some other deep boring construction, it's not uncommon to have an unforeseen obstruction which can dramatically increase the cost of a project. This can also be an issue with a cut-and-cover tunnel, although Cambridge Street was widened in the 1920s so the utilities there are a bit less complex than the centuries-old sewers under many Boston streets. The point of an alternatives analysis is too look at different alternatives and see which is the most appropriate. In this case, there are not enough alternatives, and very little actual analysis. I wouldn't be surprised if a full analysis showed that 1300 feet of a cut-and-cover tunnel was a good deal less expensive than twice as much deep bore tunnel and the additional track connections involved.

And there is no information in this report about how they are going to build the Charles Blue Line Station 50 feet underground in the substrate of Boston with sequential mining. The Second Avenue Subway is being built in hard Manhattan schist which can be blasted apart while still maintaining structural integrity above. I'm not an engineer with this sort of experience, but given that there is no explanation of how it would take place, I'd have to doubt its veracity. The current idea is explained that the tunnels would be bored and then the station areas would be mined out in between the bored tunnels, using their structure to support the road above. I guess that could work. But it seems to add several steps (and thus increase the cost) by building the tunnels only to hack them apart to build space for stations and crossovers in between. When asked in 2011 if it was inflating the costs of the project, MassDOT was very defensive in claiming that they weren't, and that Very Important People said the same thing, yet they didn't actually explain why they chose the scheme that they did (beyond "we hired someone") and as I outline here, it seems they put their thumb on the scale.

Even giving them the benefit of the doubt that a bored tunnel is the best option, the cost estimates seem out of hand. The cost of the 72nd Street Station cavern—which is 1300 feet long, the length of the Red-Blue connector from Charles to Joy Street—plus the track connections to 63rd Street is $431 million, significantly less than the Red-Blue connector. This, for a project taking place 100 feet below the street in New York City, which may be the most expensive construction market in the world.

To put it another way, I find it very hard to fathom that a ¼ mile cut-and-cover tunnel with a single station (for which the headhouse is already built) connecting in to an existing tunnel would cost three quarters of a billion dollars. Or $3 billion per mile. Some more comparisons? The cost of the Longfellow Bridge—twice as long, and rebuilding a century-old bridge while maintaining transit service—is one third the projected cost of Red-Blue. That can't be right. The Big Dig cost less than $3 billion per mile, to build highway tunnels three times as wide, over and under several active railroad tunnels, with more ventilation and dozens of ramps. And the Red-Blue connector would cost as much? Please.

A layman's staging plan for Red-Blue. Simplified, a bit.
A layman's plan (as follows) would involve a shallow cut-and-cover tunnel, likely using slurry walls to support the excavation. (See Dig, Big). The tunnel would be 40 feet wide at the Charles Station and to the east for the crossovers and the pocket track, where it would taper to 20 feet for the connection to the existing trackage at Joy Street. The Cambridge Street roadway is at least 64' wide between Charles Circle and Joy Street; with parking it is 80 feet wide. Assuming the construction could be completed in two phases (two 20-foot-wide excavations) with a 5 foot buffer around each, this would leave 34 feet for road for traffic during any construction, enough for two lanes of travel in each direction (or two in one direction, one in the other and an emergency vehicle access lane). Cambridge Street is currently a horror show for cyclists and not much better for pedestrians, and the project would allow a complete street to be built in its stead.

Trackwork. Click for full size.
Would there be traffic headaches during construction? Sure, just as there have been with the Longfellow Bridge adjacent to the project area. Would it be apocalyptic? Much like the Longfellow, it would not. And the effects would stretch only from Charles Circle to Joy Street, rather than the state's plan, which would have impacts extending from west of Charles Circle to Government Center, at least. In any case, either scheme will have short term traffic issues, but a long-term benefit, both with fewer vehicles and the potential to build a "complete street" with separated bicycling facilities and better pedestrian facilities. And get rid of the damn median!

There is no logical reason that the Red-Blue connector should be, per mile, the most expensive subway construction in the country. Unless it's by design. And—yes, to point a finger at MassDOT—that's exactly what I think has happened.

tl;dr: this is why we can't have nice things.

Monday, August 29, 2016

What to do with Central Square

Traffic-wise, Central Square is a mess. Squeezed in to the streets are about 30,000 vehicles on Mass Ave and Prospect Streets, bus routes—most of which terminate in or near Central—serving more than 30,000 daily riders, thousands of cyclists and countless pedestrians going to and from work, home, businesses and transit. (This leaves out the tens of thousands of Red Line riders moving through under the street.) The street has been rebuilt many times, most recently between 2006 and 2009, to widen the sidewalks and realign Lafayette Square at the east end of the area. Sitting as it does adjacent to Kendall, Central has seen more traffic (of all types) in recent years, and often devolves in to gridlock at peak times.

That's a lot of space for cars, isn't it?
Still, the Square is remarkably car-oriented for a community where the majority of residents don't drive as their main means of transport. Bike lanes are an afterthought, and cyclists jockey for space as buses, taxicabs and parked cars pull in and out, crossing and frequently blocking the bike lane. It is one of the most dangerous locations in the city for cyclists, which is no surprise to anyone who bikes there. For pedestrians, crosswalks are frequent and Mass Ave and Prospect Street have five second leading pedestrian intervals, but sidewalks are still congested, especially near transit stops which often fill with riders if a bus is a few minutes off of its headway.

Back in the day, transit riders boarded streetcars in the center of the street
in Central Square (these were not exclusive lanes but rather "safety zones"
where passengers could board streetcars while automobiles passed on the
right; cars could pass on either side of the platform.
And transit riders? They have it worst. Long queues can form entering and exiting the too-narrow subway entrances at Pearl Street. Bus riders have a small shelter on Mass Ave, which is often inadequate for the number of riders waiting for the multiple routes which board there, and riders on Route 70 are forced to board buses a block away from the Square, on Green Street, with minimal shelter, narrow (just five feet wide!) sidewalks and on a grungy back street which is often so choked with traffic the bus can barely manage a crawl between the stops.

Unlike most other parts of the 1 Bus route, there are parallel streets in Central which could be used to alleviate traffic on Mass Ave and provide safer options for cyclists and pedestrians and better conditions for transit riders. It would require a major rethinking of how street space is used, changing the direction of Green Street and moving eastbound traffic one block to the south. That hurdle aside, Mass Ave could be reapportioned to allow for a safe, separated bicycle facility, bus stop consolidation at a single point adjacent to the Red Line (not, for many riders, a block away), and a transit-only facility stretching several blocks, free of the traffic snarls that routinely hold up buses. It would also (gasp) reduce some street parking, but the majority of businesses in Central cater to walk-in traffic, and there is ample parking at the too-numerous parking lots nearby and at the ugly-and-should-be-torn-down-for-housing Green Street Garage.

So, how do we create a Central Square where pedestrians, cyclists and transit riders are put first, and not an afterthought?

1. Green Street flips from west to east. This allows all of the traffic from Mass Ave to be shunted south on Pleasant Street by the Post Office and then left on to Green. (Franklin would probably also be flipped from east to west, which would have the added benefit of eliminating a the Franklin/River intersection, which has very poor sight lines.) Green would be two lanes wide, with one lane for through traffic and the other for deliveries and drop-offs and potentially parking between Magazine and Brookline. While this would increase traffic on Green Street, it would be mitigated by removing most if not all of the buses (more on that in a moment). Furthermore, there are very few residential buildings on Green Street, which is really mostly a service corridor for Central Square, so the impact of any additional traffic would be minimal. The street is 24 feet wide, which is wide enough for two 12-foot travel lanes.

Once past Pearl Street, traffic would be able to filter back to Mass Ave. Some traffic would take Brookline Street, mostly to zigzag across to Douglass Street and Bishop Allen. Traffic destined to Main Street could turn here, or signals could be changed to allow a straight-through move on Sidney Street. Traffic going towards Boston could continue on Green Street as far as Landsdowne, where the diagonal street allows for less severe turns.

What about westbound traffic via Bishop Allen and a transit-only corridor? There are several reasons this is suboptimal. First, it's probably good from a political and practical sense to have some vehicular access to Mass Ave. Otherwise you wind up with some dead-ended narrow streets abutting the square. Second, the right turn for through traffic from Mass Ave to Bishop Allen is very hard to figure out. The Sidney Extension-Main-Columbia turn would be implausible increased traffic. Douglas Street is only 20 feet wide and is probably too narrow for trucks. Norfolk is 24 feet, but then you're creating a busy turn right in the middle of the square. Finally, complete streets include cars. They just don't make them the priority.

2. Mass Ave eastbound is rerouted to Green Street. As described above, all traffic from Mass Ave eastbound would be diverted to Green Street at Pleasant. Traffic wishing to turn left on to Prospect would take a right on Pleasant, and a left at Western. Light timings would be changed at Western to allow for additional traffic. Mass Ave Westbound would remain as is.

Eastbound traffic patterns for traffic to and from Mass Ave. The dashed line shows where traffic would be allowed but not encouraged; signs would direct through westbound traffic to Mass Ave to proceed to Landsdowne Street, but right turns from Brookline to Mass Ave would be permitted. For simplicity, not all traffic movements are shown.
Eastbound traffic patterns for traffic to and from Mass Ave. The main change would be the split for traffic destined to Sidney Street and Pearl Street, where turning traffic would share a separate lane with buses before turning left.
3. A two-way busway would be built on the south side of Central Square from Pleasant to Sidney. Eastbound buses would be exempt from the turn to Green Street and instead proceed directly down Mass Ave. East of Pearl Street, this busway would allow for some general traffic: right turns from Brookline Street to Mass Ave and left turns from Mass Ave to Pearl. The busway would also allow emergency vehicles to bypass gridlock in Central Square, creating BRT elements in one of the most congested areas of the 1 bus route (as opposed to, say, the Silver Line, which has bus lanes in the least congested part of the route). Bus stops would be consolidated between Pearl and Essex for betteraccess to transit. (This 160-foot long section could accommodate four 40-foot buses.) Buses would be able to loop as follows:

  • 1 is a through route. CT1 should be eliminated. Short turns could be made via Pleasant-Green-Western.
  • 47 would go left from Brookline to the busway. Loop would be made via Pleasant-Green-Western. A single-bus layover would be retained at the end of Magazine Street. This would eliminate the need for passengers to walk a block to transfer.
  • 64, when not operating through to Kendall, would loop via University Park, but instead of serving stops on Green Street, it would loop back to the busway. Left turns would be allowed for buses from Mass Ave to Western.
  • 70 would loop via University Park as above, making inbound and outbound stops on Mass Ave, eliminating the walk to Green Street and the inadequate boarding facilities there.
  • 83 and 91 would use a left-turn lane for buses only on Prospect Street (currently a painted median) to allow access to the busway. An actuated signal there would allow a left turn phase when necessary (approximately once every ten minutes, which would have a negligible effect on other traffic). Buses would then loop and layover in University Park like the 70. This would allow these routes to serve the growing University Park area, which has seen significant development in recent years. 
A busway, a cycletrack, a travel lane and even some parking! Emergency vehicles would be able to use the busway, too.
4. Eastbound bus stops would remain largely where they are on the south side of the street, but any pull-ins and bulb-outs would be removed to allow vehicles to maneuver more freely. (The additinal crossing distance would be mitigated by the bus platform mid-street.) Westbound bus stops would be placed in the center of the roadway; one between Pearl and Essex (approximately 160 feet long) and another east of Sidney Street (60' long, for the 1 Bus only), where those buses (and Pearl Street turns) would be shunted to the left. These stops would be ten feet wide, significantly wider than the current stops on Green Street. Pedestrians transferring between the Red Line and westbound buses would have to cross just the westbound traffic lanes of Mass Ave, no longer making the trek to and from Green Street. West of Essex Street, the bus lanes would jog to the right to allow clearance between the headhouse and elevator for the main entrance to the Central Square station. The bus platform—which could be raised to allow level boarding akin to the Loop Link in Chicago—would span the distance between the Pearl and Essex crosswalks, allowing access from both ends of the platform. (The bus platform for eastbound buses could also be raised.)
MBTA bus routes shown, including loops for routes terminating in Central. At non-rush hours, the 64 bus would follow the route of the 70. The dashed blue line shows the ability for the 1 bus to short-turn (today known as the CT1). Other buses, such as the MASCO shuttle, could also use the busway.
East of Pearl, the busway would allow some general traffic (left-turning cars to Pearl Street). Through traffic would remain on the north side of the street, and the cycle track would have no vehicular crossings between Brookline and Prospect.
5. In between the westbound bus stop and the westbound traffic and loading zone lane would be a 10- to 12-foot-wide cycletrack, running from Sidney to Inman. Except where adjacent to the firehouse, it would be raised above grade and separated from traffic. At either end, a separate bicycle signal phase would allow cyclists to move from existing bicycle facilities to the center of the roadway. This would eliminate the constant conflicts between cyclists, motorists and buses. Bicycle traffic calming measures would be required in the vicinity of the bus stop at Pearl Street with high pedestrian traffic, but cyclists would otherwise have an unobstructed trip from Sidney to Inman (with traffic lights at Brookline and Prospect, where a bicycle phase might be necessary for right turns). For turns to Prospect and Western, bike boxes would be provided to allow two-stage turns. For turns to minor streets, cyclists could use areas adjacent to crosswalks. Since bike lanes in Central Square are frequently blocked by vehicles, this would wholly eliminate these issues.

A bus-only facility would dramatically improve facilities for transit passengers, a cycletrack would eliminate car-bike conflicts make biking through Central much safer, and a bus platform would decrease crossing distances for pedestrians. And there would still be ample room for taxis and loading zones on the westbound side of the street.
Why not side-of-street cycletracks? A few reasons. First, putting the cycletrack in the middle of the street means that you don't have right-hook issues (although right turns from the cycletrack are trickier). Second, bus stops. Central is one of the busiest bus transfers in the MBTA system without an off-street facility (think Alewife, Harvard, Forest Hills, Ashmont, Sullivan, Kenmore, etc). You'd need large floating bus stops and really need to pull the cyclists back from the street. Third: pedestrian traffic. There are a lot of pedestrians in Central Square. A successful cycletrack would need significant separation from the sidewalks to avoid becoming choked with pedestrians. This is a lot easier to do in the middle of the road than it would be alongside the sidewalks. Finally, the busway creates the need for a buffer between the westbound travel lane and the buses, which is a perfect place for the cyclists. You do have two points of conflict on either end of the cycletrack to transition from the existing lanes (which can be signalized) but otherwise have relatively clear sailing for cyclists devoid of the current maze of turn lanes, parking spaces and taxi stands.

East of Sidney Street, Mass Ave westbound would split, with left-turning traffic to Sidney and Pearl to the left of the bus stop island for the 1 bus (the floating stop on the south side of the street would serve the 1, as well as routes short-turning at University Park). The westbound bike lane could be cycletracked inside parking. Both bike lanes would have a signal phase at Sidney to allow a safe transition from the side of the street to the center-street cycletrack.
6. The westbound lanes of Mass Ave would be 22 to 24 feet wide, allowing the current travel lane as well as a wide area for a loading zone for area businesses, a taxi stand and other pick-ups and drop-offs near the transit station. These uses would no longer conflict with bicycle traffic. Some street parking could be provided, but it is probably best relegated to side streets nearby or parking lots (there are generally few on-street spots today anyway).

All of this might increase traffic congestion for some drivers. (Horrors!) But it would benefit the large majority of users of Central Square who arrive by transit, walking or biking, or a combination of all of them. Central once had transit stops in the center of Mass Ave (for streetcars), and it's time that those users were the priority for the heart of Cambridge, not an afterthought.

Thursday, August 25, 2016

The odd history of the 66 Bus Wiggle

One frequently-mentioned (and usually fallacious) argument is that we need to rework our bus routes because they follow the same routes as they did in the days of the streetcars. (This is a fallacy because, in most cases, the streetcars followed the path of least resistance: straight, wide roads with mixed uses and density. In Houston the uses had changed so drastically since the streetcars it made some sense, in older, denser cities, it doesn't.) Then there's the case of the 66 bus in Boston, Brookline, Boston again, and then Cambridge, particularly the two-sides-of-a-triangle "wiggle" to Union Square in Allston. (Here's a visual primer on the 66)

This isn't entirely true. There are a lot of reasons why the
66 bus sucks. But this shows the wiggle well.
Via here.
The wiggle stems from a route realignment in 1989, where several routes were rejiggered, which is the type of route realignment I think the T should do more of. Before 1989, there were several routes which terminated in Union Square in Allston, not because it is a major activity locus, or a major transfer point, but because back in the day, there was a car barn there and BERy decided to start and end routes there. Back then, the following routes served Union Square Allston:


  • 57 Kenmore to Watertown Yard
  • 63 Cleveland Circle to Central Square via Western
  • 64 Oak Square to Central Square
  • 66 Allston to Dudley (note that there was never a streetcar line across the Anderson Bridge; this was always a bus line)
  • 86 Union Square Allston to Union Square Somerville via Harvard


After the changes, except for the 57 and 64, the routes were split apart and recombined at Union to better serve the needs of the traveling public (what a thought!). The 63 was shifted  combined with the 86 to form the current 86 bus (which was extended to Sullivan in 1981 when the Orange Line was realigned). The 66 was then extended to Harvard Square to cover the section in Lower Allston the 86 bus missed. This better focused service towards Harvard (where subway connections were, since 1985, available in both directions, and which has more terminal capacity than Central) and provided a one-seat ride between Brookline and Harvard for the first time, leaving us with the current routes:
  • 57 Kenmore to Watertown Yard
  • 64 Oak Square to Central Square
  • 66 Harvard to Dudley 
  • 86 Cleveland Circle to Sullivan
Initially, the 66 was routed straight through on Harvard Ave. Apparently there were protests (not sure by whom) and that the route no longer served Union Square and it was realigned, and since then has cost through-riding passengers (the vast majority) several minutes of travel time. The gist of the protest, as far as I can tell, is that you couldn't get to Union Square on the 66. But this is not a valid argument. The 66 intersects the 57, another high-frequency bus route, which serves Union Square. If you are unable to make the (short) walk to Union, you can instead utilize the transfer feature of the bus network. Most users can walk the 1200 feet (a 4 minute walk) to Harvard Ave; others can use the every-ten-minutes 57 bus and transfer. Apparently, planners at the time kowtowed to these complaints. It probably costs the T tens of thousands of dollars in operating costs every year, and likely reduces ridership as potential riders choose other modes because of the length of the route.

A straightened route would consolidate several stops at Harvard and Commonwealth, which could be rebuilt as a high-amenity stop (larger shelter, higher curb, real-time arrival display, signal priority, etc). The few passengers who need to get to Union Square and can not make the walk could, instead, ride the 57 (or even the 64). No stop would lose service (the entirety of the wiggle duplicates other routes) and it would make the 66 faster and more reliable.

Tuesday, August 16, 2016

The problem with the CT1

The 1 bus is one of the busiest routes in Boston. It runs along Massachusetts Avenue, touches three subway lines (and the Silver Line), and is an important crosstown route, despite frequent bus bunching and traffic delays. The bus is chronically overcrowded; I've regularly counted 78 people on a 40 foot bus, even with frequent service. It is supplemented by the CT1 "Limited" service route, but the CT1 is poorly planned and integrated, and winds up being a waste of resources on the route. (Speaking of resources, we've argued in the past that the corridor should have bus lanes on the Harvard Bridge and in Boston, with a more equitable allocation of space for corridor users.)

The CT1 is barely a limited service route. The two routes overlap between Central Square in Cambridge and BU Medical Center in Boston. In theory, the CT1, by making fewer stops, should be able to make the trip significantly faster than the slower 1 bus. What follows is an exhaustive list of stops that the 1 bus makes that the CT1 bus does not:
  • Mass Ave at Albany Street
  • Mass Ave opposite Christian Science Center
  • Mass Ave at Columbus Ave
That's it. 

In general, a limited stop route should serve no more than half the stops that the local service does. (For instance, limited-stop routes in Chicago make only about one in four stops the local buses serve; the Twin Cities is similar.) But in this case, the local route makes 13 stops, and the limited route makes 10. A few years ago, several poorly-utilized stops on the 1 bus were cut. (This included the particularly inane stop in the median of Commonwealth Avenue which required crossing the same number of streets as stops within a few hundred feet at Beacon and Newbury. The stop at Columbus Ave is within 500 feet of the Mass Ave Station and could be similarly consolidated. For those of you keeping track at home, that's less than a two minute walk.) If most of the stops are served by both buses, there's really no point in having the two separate routes overlap and not make the same stops. Cut Columbus and consolidate Sidney and Albany in to one mid-block stop and you can have both buses make the same stops.

Not that anyone really waits for the CT1, anyway. Passengers, for good reasons, generally will get on whichever bus comes first unless the next is visible. If a 1 bus pulls up, get on the 1 bus; it's rare for it to lose so much time at two or three stops that it gets caught by another. The CT1 is really more of a short-turn of the 1 bus (the inimitable Miles on the MBTA agrees as to its lack of usefulness), serving the busier portion of the route between Boston Medical Center and Central Square. Yet the schedules aren't integrated, so, at times, two buses are scheduled to leave Central Square within a couple of minutes with a subsequent 10-plus minute gap. 

For visual learners, this chart shows the combined 1 and CT1 bus headways
at Central Square, and a moving average of five buses. By combining the
1 and CT1, the effective headway could be reduced significantly. In other
words, the orange line shows the average headway of the bus (what would
be possible if the routes were combined and better dispatched) while the dots 
(blue and orange) show the effective headway of service provided today.
What this creates is a situation where resources go underutilized. Often a bus will leave Central Square packed to the gills, and another will leave mostly empty two minutes later—and invariably catch up with the bus in front of it—and then no bus will run for 10 minutes. Yet if the two routes were combined, rush-hour service could be provided every six or seven minutes (down from wait times as long as eight minutes in the evening and ten in the morning) at rush hours and 10 to 12 during the midday (current wait times are as long as 15 minutes midday). This is an issue related to the poor interlining of the 70 bus which this page has discussed in the past. The effective headway of the bus—the longest headway during any given time—is longer than it would otherwise be. 

Last year, through Cambridge's participatory budgeting system, voters there overwhelmingly supported signal priority for the 1 bus, and, according to Twitter, it is currently being installed. This is important, as it will allow better schedule adherence for buses which otherwise get hung up at the many lights through the city (traffic, on the other hand, is another question this page will attempt to answer in coming days). Better dispatching is important as well to allow short turns when two or three 1 buses run back-to-back (which happens all the time).

The CT1 may have made some sense when the 1 bus made more stops, but today it just serves to gum up the works. After 22 years, it's time to axe the CT1 and improve the 1 bus. Relieving the route of a few extraneous stops was a good start. Cambridge has taken another step forward with signal priority. All-door boarding and pre-payment would be easy at major stops, since most are adjacent to stations with fare machines (and others, like MIT, could have machines installed). Loop Link-like platforms and stations would help as well (Loop Link is an example of where the city and its transit agency actually talk to each other). And dedicated lanes? Well, that's probably further off, but should be part of an iterative process. Otherwise, we'll waste most of the small, but important, improvements to the 1 bus so far.

Wednesday, August 3, 2016

Service, Not Storage

One of the many issues with the Allston project that I have been way too involved with is that the State maintains that they need a midday layover yard for Commuter Rail equipment. Why? Because they plan to add capacity to South Station and, since they can't stack out-of-service trains in the terminal, need somewhere to store them when they're not in use in the middle of the day. (Off-topic but relevant: the need for and cost of which could be obviated completely simply by building the North-South Rail Link; thru-running is so much more efficient that in Philadelphia SEPTA runs 44 trains per hour through its four track tunnel at rush hour while the MBTA peaks at 32 trains combined on 22 tracks at North and South stations.) The relevant issue, however, is how silly it is to build large rail yards on prime real estate in order to not run service!

The supposed "need" for this whole facility could be obviated simply by running more trains in service in the midday. (Some layover space could be built between Cambridge and Everett streets in Allston without impacting transit operations and development of the Beacon Park Yards.) If you have more trains in service, you don't need storage for them. Considering that 75% of the costs of running Commuter Rail in Massachusetts are fixed, much of the marginal cost of providing increased service would be made up for by the opportunity cost of not building such an unnecessary facility. Most every other major Commuter Rail line runs more frequent midday service than the MBTA, even on lines to major anchor cities like Worcester, Providence and Brockton. In English: you have the trains, and the track, and the stations. Just run more darned trains already!

With that said, you still need to figure out where to run these darned trains. Obviously, increasing service on current lines to large cities and "Gateway Cities" makes sense. But there's actually a way to increase service to Western Massachusetts without any major investment in track, stations or additional equipment. Right now, several train sets begin and end their day in Worcester. These trains could, instead, begin their trips further west, providing service to Springfield the Pioneer Valley.

The Commonwealth and Feds recently spent $83 million to upgrade the Connecticut River Line for passenger service, and the Boston and Albany main line already hosts Amtrak trains (albeit at a pitiful top speed of 59 mph). So you might as well get some use out of it! There is a vague plan to provide commuter service in the Pioneer Valley soon linking Greenfield, Northampton, Holyoke and Springfield, and connecting with upgraded Springfield-Hartford-New Haven "Knowledge Corridor" service. Service couldn't start overnight, steps would include working out a track use agreement with CSX and qualifying crews west of Worcester. But the track (with the exception of layover facilities in Greenfield; I assume trains could be stored overnight on tracks in Springfield's station), stations (with the exception of Palmer where you might want to build a new station) and trains are in place. It's not a big leap to running service.

Here's what a schedule would look like for the trips serving Boston, Springfield and Greenfield. Train numbers are shown for current Amtrak or MBTA Commuter Rail service. (Amtrak 448/449 is the Lake Shore Limited to and from Chicago via Albany, 55/56 is the Vermonter from Saint Albans to Washington D.C.)

Eastbound
       Train # →         MBTA 508    MBTA 552      NEW      AMTK 56    AMTK 449  
Dep Greenfield--5:45--13:36--
Dep Springfield5:456:4513:0014:3517:33
Arr Boston8:209:0715:20--20:01
Westbound
       Train # →           NEW      AMTK 448    AMTK 55    MBTA 521    MBTA 551  
Dep Boston9:3812:50--17:0519:35
Arr Springfield12:0015:1815:1519:3522:00
Arr Greenfield----16:1516:2220:35

Rightly or not, Western Mass often feels like it gets the short end of the bargain when it comes to transportation funding. There has been hundreds of millions of dollars spent on infrastructure (most notably the Knowledge Corridor and Springfield Union Station), yet very little service to show for it. This mistake would be compounded by overbuilding layover facilities in Boston and siloing operations in Eastern and Western Massachusetts. Any passenger movements would need to accommodate CSX freight traffic between Worcester and Springfield, and in the long term, much more increased service may require a larger investment to re-double track the B&A line to Springfield (and to increase line speed where possible as well).

In any case, rail service in Massachusetts has long been focused on Boston, with a minimal statewide transportation plan (well, beyond taking donations from Peter Pan, buying them buses and having them get stuck in traffic on the Pike). The state has taken hundreds of millions of Federal dollars (and local match) to upgrade the line in the Pioneer Valley, but it barely runs any service. It would be a politically wise move to better serve Western Mass and, given traffic and tolls, would probably attract significant ridership, too.

Tuesday, June 28, 2016

Remember: changing road pricing can have unintended consequences

A few years ago, I wrote about how the 1996 changing of the toll structure of the Turnpike in Newton dramatically affected traffic. In that case, changing the toll by $1 created a new calculus where many commuters took an alternate route to avoid the toll, leading to traffic on side roads. The state will soon change the toll structure on the Tobin Bridge, going from $2.50 in one direction to $1.25 in each way. They claim it will be revenue neutral (it will likely be somewhat revenue positive, actually; see below), but there is not talk of the traffic impact, because the amount you're charging is the same, so it won't change the traffic, right?

Wrong. There may be a major traffic impact.

First, the revenue projections. It is very likely that revenue will actually go up. Today, it costs $2.50 to go south on the Tobin, and it is free going north. For a motorist coming south on 95 in Peabody, it is usually only 2 to 3 minutes longer to loop around on 128 and 93 versus the trip straight down Route 1. (At rush hour, it depends more on traffic, but the majority of travel on the Tobin Bridge is at non-rush times.) The 128-93 route is about five miles further, but even assuming 50¢ for gas (most motorists don't figure in the full marginal cost of a mile traveled, many probably discount the extra gas anyway) it is still a saving of $2 for three minutes of time, a rate of $40 per hour. That's generally worth it.

Which is much of the reason why, in 2015, there were 51,000 northbound vehicles daily on the Tobin, but just 34,000 going southbound. (Some of this may be explained by things like the location of on- and off-ramps and traffic patterns, but most of it is likely due to the toll.)

Equalizing the toll will change this calculus dramatically. Many motorists who had balked at the $2.50 toll may be more willing to part with $1.25 to save a couple of minutes. And while some will avoid the bridge northbound, it will be far fewer than if the toll were flipped and it was charged full rate northbound, and free coming south. Guessing wildly, I'd guess that 5,000 motorists will use the bridge coming south, and 5,000 will abandon it going north. This means a lot of new toll revenue for the state.

Currently, the state collects $2.50 from each of the vehicles using the bridge southbound (we'll assume that the higher rate for larger vehicles offsets the discounted toll charged to Charlestown and Chelsea residents). With 34,215 vehicles counted per weekday in 2015, this amounts to $85,537.50 in toll revenue. My wild-guess assumption is that there will be 5,000 more northbound travelers (39,215) and 5,000 fewer southbound travelers (46,108), each paying $1.25, for a total of $106,653.75, or an additional $20,000 in toll revenue daily. Even discounting lower traffic on weekends and holidays, this will probably add in the neighborhood of five to six million additional dollars of toll revenue for the state.

The tolls will be fairer and make more sense and raises more money for the state, which can always use more money for infrastructure. This is a win-win …

… unless it has an unforeseen impact traffic. The bridge itself is nowhere near congested, especially coming southbound, where peak-hour traffic counts average just 3000 per hour (1000 per lane per hour), well below the 1500 where congestion begins in earnest (in the chart above, you can see how the northbound traffic levels off at about 4000 per hour, even dipping slightly during the 4 to 5 p.m. peak, which may be due to heavy traffic on roadways accessing the bridge reducing throughput). The issue with more inbound traffic is at the end of the bridge.

Most of the traffic (about 85%) stays on the loop ramp to the Leverett Connector and O'Neill Tunnel. An additional stream of traffic is added from Rutherford Avenue, and there is considerable merging and sorting of this traffic. With only two lanes, this is much nearer capacity; adding more vehicles may create merge traffic which will cause significant backups. The traffic on the Leverett Connector and the tunnel should be a zero-sum game, shifting users from I-93 to Route 1, much like the Turnpike toll removal didn't necessarily increase traffic, but changed where it got on and off of the highway. Even minor changes can have consequences. This one may work fine. Or it may not.

Still, I'm for this change, as it is sensible policy (even if it might have some unintended consequences). In fact, the Commonwealth should explore avenues to toll all the highways leading in to Boston at a rate equal to the Turnpike. Tolls on the Turnpike are not a detriment to the local economy, which seems fine to be churning alone just fine. And other than federal policy (which may be changing), there is no logical reason why Turnpike commuters should have to pay $5 a day to get from 128 to the city while I-93 commuters get in for free. And while the dollars from the Tobin change are relatively small, charging a sort-of congestion charge for other highways leading in to Boston could bring in big dollars. I-93 could be tolled at $2.50 in both directions from 128 to the city, with a lower toll for Route 2 ($1.25), perhaps waived for commuters parking at Alewife. The harbor tunnel tolls, currently $3.50 one way, could be reduced to $1.25 each way to match the other tolls.

At this rate, and assuming that 10% of travelers would carpool, take transit or use side roads to avoid the tolls (although with electronic tolling, it's harder to simply avoid a toll booth), this would increase the equity of the transportation system, while at the same time raising more than $600 million annually for road maintenance. Considering the age and state of many of the roadways, bridges and tunnels in the Commonwealth, this money could be spent making sure that the roadways are maintained statewide. Unlike a vehicle mile tax, this is not a new concept; it's one which has been in place on roadways throughout the Commonwealth for nearly a century (far longer if you account for the 19th century incarnations). And unlike a statewide gas tax, this targets users of the state's most crowded and overtaxed infrastructure, and may be a factor leading drivers to consider other modes. The Turnpike users already pay tolls. It's high time others did a well.