Wednesday, March 27, 2013

Review: Walkable City

I recently picked up Jeff Speck's new book, Walkable City: How downtown can save America, one step at a time. I highly recommend it. Jeff is a Massachusetts native who has worked as an architect, city planner and designer all over the country. The purpose of the book is best explained by this quote from the prologue:
This discussion is necessary because, since midcentury, whether intentionally or by accident, most American cities have effectively become no-walking zones. In the absence of any larger vision or mandate, city engineers--worshiping the twin gods of Smooth Traffic and Ample Parking--have turned our downtowns into places that are easy to get to but not worth arriving at.
I would characterize the book as a collection of well-footnoted facts assembled into a coherent narrative with compelling, clear prose. It was enjoyable and quick to read. You will be better informed afterwards. He lays out the case for walkability, and then follows it with advice for achieving walkability in your city. Being familiar with the topics, I knew about most of the facts presented, but did learn a bunch as well, including (but not limited to):

  • Portland, Oregon residents drive 20% less and save an estimated $2.6 billion because of that, which finds its way into the local economy instead of being pumped abroad.
  • Massachusetts residents with the lowest body mass index averages were located in Boston and the inner ring suburbs, while the highest averages were found outside the I-495 belt.
  • Toronto has the most linear feet of successful retail-fronted sidewalks; Sweden has the highest share of urban trips going to walking instead of driving; and sidewalk cafés stay open all year round in Denmark.
  • "Landscape urbanism" is actually a return of the dreaded "tower-in-the-park" style city planning that caused so much misery in the mid-20th century.
The book functions as a very accessible source of quick facts and pithy quotes for cross-referencing -- you will want to highlight many interesting passages for later. I think my favorite quote may be this one:
Since it is the only real constraint to driving, congestion is the one place where people are made to feel the pinch in their automotive lives. Were it not for congestion, we would drive enough additional miles to make congestion. So the traffic study has become the default act of planning, and more than a few large companies can thank traffic studies for the lion's share of their income. They don't want you to read the next few paragraphs. 
Traffic studies are bullshit.
Indeed. I've written about the problems with traffic studies several times. Jeff goes on to enumerate three main reasons. The first one is what we in the computer science business would call "GIGO": Garbage In, Garbage Out. The models are only as good as their inputs, and their design, for that matter. They can easily be tweaked to suit whatever agenda the modeler has in mind. For example, most cities assume 1% to 2% "background traffic growth" (here in Boston it's usually 0.5%) even when traffic levels are falling. The second objection is the old "conflict of interest" story: the firms which produce traffic studies are often the firms which get hired to implement the "solution." Hence they are motivated to call for more spending which will be routed into their own coffers. And finally, almost never do traffic studies consider the phenomenon of "induced demand", even after it has been observed and studied for over fifty years.

I would also add another problem with traffic studies: they focus on "vehicles throughput per hour" or "vehicle delay" instead of considering the flow of actual people. So the bias against buses or carpools continues. Although, nowadays traffic studies around here do tend to include some token mention of "pedestrian delay", "bicycle throughput" or "transit access" it's usually a short bit and tucked away.

I agree with much of what he writes, and he also relies heavily on the great work of others, such as Donald Shoup and Hans Mondermann. But I have to take issue with a small selection of ideas in the book; nothing that affects the larger point, but a few things that should be pointed out.

  • Jeff: Urbanity [in public transport] means locating all significant stops in the heart of the action, not a block away. Me: Yes and no. You definitely want to keep it convenient, but there's plenty of examples of successful transit which requires walking a block or two "from the action." Practically all subway stations (except for the shallowest cut and cover) require some amount of walking around to access. People will walk further to higher quality transit.
  • Jeff: Frequency is the thing that most transit services get wrong. [...] so ten-minute headways are the standard for any line that hopes to attract a crowd. If you can't fill a bus at that rate, then get a van. Me: He's absolutely right about the importance of frequency. But "getting a van" is not a solution to the high cost of frequency. First of all, the primary operating cost is paying a driver; and there's no discount for "van driver" over "bus driver." Second, if you are a public transit agency, then you have a fleet of hundreds of buses. You are probably not interested in adding yet another model of vehicle that needs its own maintenance manuals and technician training. You are definitely not interested in extra deadhead trips back to the garage because you need to switch out a van with a bus. In fact, it probably will end up costing you more money to equip a route with vans than with the buses you were already using.
  • Jeff: Few sidewalks without parking entice walking, yet cities routinely eliminate it in the name of traffic flow, beautification, and, more recently, security. Me: I was scratching my head over this one for a while, since there's plenty of examples of sidewalks without parking that are great. But I would say that they are mostly on small streets where people freely share the roadway (such as in the North End), or the sidewalks are plenty wide, or there is some other circumstance protecting walkers from traffic. I agree that if the street is too wide and there is fast-moving traffic, then a line of parked cars is better than a travel lane. But that's a workaround to deal with a bad situation. Unfortunately, the idea of putting street parking everywhere leads designers to create totally new streets that are way too wide in order to accommodate parking on one or both sides. Even when those streets could have had the opportunity to be smaller, calmer "shared streets," an idea which Jeff also promotes.
Of course, he gets right back on my good side: "Another reliable bellwether [of walkability] is the visible absence of push-button traffic signals. [...] Far from empowering walkers, the push button turns them into second-class citizens; pedestrian should never have to ask for a light." Although, Boston is the exception; we have walkability despite the omnipresence of push-button signals -- probably because everyone ignores them anyway.

The book is pitched at smaller cities, places trying to pick up population and compete with Boston, New York, San Francisco, etc. But I would say there's plenty Boston could still learn from it, and I'm sure folks would agree about some of the other "star" cities as well. If more city planners read this book, great. But it's also written for a wide audience, so just about anyone with any interest in cities, or walkable towns, will find this book enjoyable and enlightening. And that really could be anyone; we are all affected by the fate of the cities in our economy. And if city planning (and traffic engineering) of the last century has shown us anything, it's that the people with the most credentials can sometimes have the least sense. We need ordinary people to become informed and involved.

Sunday, March 24, 2013

Late night service on the T

Results are in from a survey regarding late-night service on the MBTA, which received an overwhelming response rate. From the Globe:
More than 85 percent of those who answered the survey said they would be willing to wait at least 10 to 19 minutes for a late-night bus or train. More than half said they would be willing to pay at least double the normal fare for night-owl subway service. [...]
The MBTA Rider Oversight Committee, volunteers who ­advise the T on customer service issues, will use the survey results to recommend possible late-night service strategies in the next few months. Reid Sprite, cochairman of the committee, said he and other members will identify which routes would be most popular with the after-hours crowd.

“We’re figuring out whether the current daytime routes meet nighttime needs,” Sprite said.
These are good signs from a survey, and also from the ROC. There is no reason why night time service demand would necessarily match day time service demand. In many comparable cities, night time service is provided in a significantly different manner from normal service.

  • For example, during the day, the MBTA's radial network is used heavily to bring people to and from the downtown Boston CBD. But as anyone who has walked there after dark knows, much of that same area is a ghost town at night.
  • Many of the subway routes are not paralleled by roads and therefore are very difficult to efficiently replicate with a bus. Instead of going station-to-station, a similar, but more efficient bus route should be identified.
  • The most important, and yet most expensive, aspect of service provision is frequency. That becomes completely unaffordable when the volumes are much lower -- such as at night.
  • In order to save money, the agency may choose to operate routes at lower frequency, but this could make transfers terribly difficult. Nobody wants to get stuck watching the bus you need leave too soon, facing a 30 minute wait at night.
  • The answer is to take a page from suburban transit services and use timed transfer points: all the intersecting buses come together and exchange passengers, and then all leave simultaneously to their next destination.
  • Some additional operating cost can be made up with a fare surcharge. But only to a point, as you don't want to discourage ridership.
  • When the MBTA ran night owl service in the past, it was discontinued due to cost. But it's not clear whether they could have saved money by adopting some common sense changes, and attracted more riders through better advertisement (not to mention the real-time tracking which is now available). And even for the cost that remains, I believe that even one drunken driving death avoided each year would be enough to make it all worthwhile.
As for running the subway trains later, let's get that out of the way. It is possible to run trains 24/7 even on a two track system like we have. Just look at the Chicago "L" which operates the Red and Blue lines at all hours even though they run through two track tunnels inside the "Loop". Heck, even in San Juan I witnessed seamless single-tracking of a portion of their elevated Tren Urbano for day-time maintenance purposes. And the excuse that the T gives is downright lame: "it's the oldest subway in America." No, it isn't. Only the portion of the Green Line between Scollay Square and Boylston can lay claim to being built in 1897 (plus the remainder of the Tremont Street tunnel). The rest was built out in stages over the course of the 20th century, and is "technically younger" than a large part of the NYC subway. Plus, everything gets renovated and refurbished every so often, or would if not for deferred maintenance.

But all of that is irrelevant. The real question should be this -- which is more efficient and effective: operating the subway later, or replacement Night Owl bus service? And I believe that the answer is the latter, for now. The capacity and frequency needs seem suited for bus service. It's easier to implement timed transfer pulses when all the vehicles can come together at the same place (the T does do a timed transfer between subways and buses currently, but it requires more personnel). Later subway service would probably require changes to maintenance schedules and some physical alterations, which may be a good idea anyway, but will have to wait until there's funding. And I think the most compelling argument for Night Owl buses is simply this: that's how other cities do it, including some big names like Paris and London, as well as more comparably-sized cities like San Francisco and Montreal.

Here's a set of map fragments to show some features from those cities' networks:


San Francisco MUNI has a good clear map which is largely based on their day-time routes, but simplified and consolidated. Note in particular the "L" which has to go over the Twin Peaks mountain in order to replicate the trolley service which runs in a tunnel. And also the abundance of timed transfer "circles".

Montreal recently expanded their night bus vehicle hours by 73%. The black dots are subway stations; notice how little prominence they receive, except for those used as transfer points. The STM actually ends normal service earlier than the MBTA, because it can count on night buses. The "747" route goes to the Airport.

Vancouver (Canada) TransLink sends out pulses of night buses from downtown until 3 a.m.

Austin (Texas) Capital Metro operates a set of night buses until 3 a.m., focused around 6th Street, of course.
The night bus networks are simpler than the day-time networks, but they exploit the riders' existing "transit geography" knowledge to put together sensible routes. They rely on timed transfers or pulses. They are focused on serving areas of high night time demand, and the agencies publish clear maps and generally try to do a good job of disseminating information and posting signs to help guide users.

If the MBTA decides to take another shot at providing night time service, and I think they should, then they ought to take some cues from other agencies who already operate such networks successfully.

Friday, March 15, 2013

"There is a very strong element of psychology behind traffic patterns."

I checked back in on the so-called "Carmageddon" of late 2012 -- the closure of one lane of the Alexander Hamilton Bridge from Manhattan to the Bronx -- to see how it panned out. From the Bergen Record, Warnings staved off traffic gridlock from George Washington Bridge. Quote (italics added):
It was supposed to be a traffic nightmare. It turned out to be a dream. 
Predictions of five-mile backups and unprecedented delays on the New Jersey side of the George Washington Bridge appear to have scared thousands of daily commuters away from the span. Except for the recent congestion caused by superstorm Sandy’s assault on the region’s transit system, motorists have seen faster-than-normal rush-hour commutes ever since the July start of a major construction project in New York City, according to transportation officials, experts and traffic data. 
“We see this time and time again,” said Jim Bak, director of community relations at INRIX, a company that collects data on traffic along the Interstate 95 corridor and other highways throughout the country. “There is a very strong element of psychology behind traffic patterns.” 
In the last few years, he said, similar worst-case warnings in Los Angeles, Seattle and London have had the same effect — actually reducing traffic to below-normal levels on roads that were expected to be clogged. It’s a public relations strategy that state transportation officials in New York, who aired the predictions in media outlets throughout the region before the July 15 start of the project, acknowledge opens them to “cry wolf” criticism. But they say there was no effort to mislead the public and that the backups might have matched predictions if people hadn’t changed their routines.
 They called it "crying wolf," I called it "Henny Penny." Same difference. More:
At 8 a.m. on a typical weekday, it takes 11 to 13 minutes on average for a vehicle to travel eastbound from the intersection of Routes 80 and 95 in Hackensack to the deck of the George Washington Bridge, according to INRIX. But in the months from the start of the Alexander Hamilton Bridge project until Sandy struck, that drive took an average of between four and six minutes.

There were nearly 6,400 fewer daily Manhattan-bound trips on average in the 2.5 months after the July 15 start of construction, compared with the period in 2011, according to officials at the Port Authority of New York and New Jersey, the agency that operates the GWB. They also said the warnings appeared to be the cause.

The Lincoln and Holland tunnels, the Port Authority’s other Hudson River crossings, meanwhile, saw no appreciable increase in traffic over that time, said Steve Coleman, a spokes¬man at agency.
I'll note that while the original over-hyped, alarmist coverage of the closure was widespread, the follow-up report was buried and barely noticed.

Obviously this fits well with their public relations strategy if they are intentionally raising the alarm in an effort to manipulate traffic patterns. It wouldn't do for the public to grow wise to their scheme, would it?
The Casey Overpass (source)

But as a result, the public still believes in doomsday predictions when it comes time to discuss the closure of collapsing highways and outdated overpasses. Cries of "gridlock!" and "nightmare!" fly at public meetings. Local residents form a sort of "Stockholm Syndrome" attachment to the 1950s era infrastructure which mutilated their community. There's a great fear of bringing the roadways back to human scale because the supposedly trustworthy officials have spent so many years scaring the public with outlandish claims of impending disaster.

At other meetings, folks stand up and protest against new development -- housing that is so desperately needed, commercial space to attract more jobs -- because they were told it would "worsen congestion." The consultants are stifled by their very own ridiculously oversimplified "level of service" metric which labels intersections with "A" to "F" letter grades. The residents want to know why the wonks aren't doing their homework well enough: surely they could get a better grade if only they put in the effort?

In both cases, we need more honesty from traffic engineers and city officials in charge of transportation. The first step is to make it clear that "Traffic is NOT like water." There is no "strong element of psychology" dictating how water flows, so that analogy simply breaks down. The honesty starts when traffic engineers stand up and admit that motorists adapt to real world conditions, that there are complicated feedback processes which are extremely difficult to model, and that it is not an exact science.
The use of "level of service." Credit: Andy Singer (source)

Traffic engineers need to admit that the letter grades given out by "level of service" are extremely misleading. What the grades really measure is the volume of cars using the roads and intersections, not the amount of "work" that an engineer does. They can tinker around the margins but the core cause of an "F" grade is that there are too many cars trying to get through that space. There is only one way to accommodate all those cars: increase the number of lanes. This is how "level of service" was often used in the past: to justify the widening of roads, the creation of grade separations, and the destruction of the communities around them. In many cases, the road widening induced even more demand than before, leaving the area just as congested and more miserable than ever.

Since road widening is no longer acceptable in urban communities, there is only one other choice available: to manage demand. Instead of holding communities hostage to dictated traffic projections, often based on false premises, the communities need to turn that around and stake out what street design is acceptable and what is not. Then the traffic engineers come in as public servants, not masters, and make the safest, best use of the allotted space. Streets are supposed to connect communities, not divide them.

Demand management goes further than that. Unintentionally, the officials in New York have demonstrated a certain principle: If you increase the perceived cost of a particular route, then traffic will decrease on that route. In the case of the Alexander Hamilton Bridge closure, the officials went out of their way to scare people into thinking that there would be an increased "time cost" for going over the George Washington Bridge. It worked. But we don't need to create crises to manage demand. There's a natural way to do it, the same way we manage demand in nearly all other contexts: putting a price to valuable road space. It's been demonstrated to work well in London, Stockholm, and Singapore, as well as other places. And it's the only real solution to congestion.

There's other tools in the demand management box: designing streets so that people feel comfortable walking or biking instead of getting into a car; improving public transportation; building communities where distances are at human-scale instead of at automobile-scale; and reducing the number of parking spaces. These strategies don't directly attack congestion the way road pricing does -- and may not appear to reduce congestion at all. But they do enable more people to live and get around without struggling constantly against congestion.

It is generally not possible to "solve congestion" non-destructively without demand management, in an urban environment. The reason is simple geometry. In a city, most any destination of significance will attract more people than roadways could handle if they all came in cars. There just isn't enough space. The cars that do attempt to come will quickly fill up the available space, and people will call it congestion. It will be congested if few people walk there, and it could still be congested even if most people walk there. Because as long as the demand exists, available slots will be filled with cars, no matter how well the roadways are designed.

The geometry problem, illustrated (source)
These days, most large projects filed with the BRA include a "Transportation Demand Management" section. They tend to use similar language (copy-n-paste is my bet) and say nice things about getting their employees to walk, bike or use transit. And it's all well and good to get people thinking about this. Developers claim the standard is to "mitigate their own impacts." But individual developers can't do much. They are usually asked to tinker with signal timing at nearby intersections, or redesign some street flows.

But that can only go so far. Real action has to be taken at a larger level; maybe community-wide, maybe city-wide, maybe even region-wide. This is because the key factors driving road demand are coming from those levels. For example, zoning rules which dictate "minimum parking requirements" cause traffic to increase in the vicinity of new developments. Zoning rules which force communities to sprawl cause people to use their cars more. Zoning rules which make for miserable streets discourage walking. The construction or widening of a nearby highway induces additional demand on the roads leading to that highway. The deterioration of public transportation forces people to resort to automobiles. And the implementation of road pricing will have to come about through action by elected officials and activism by local advocates.

The first step is honesty from the experts and the transportation officials serving us. The answers are not easy. Many won't want to hear them. But we should demand them.

Thursday, March 14, 2013

Boston MPO memo fails to understand purpose of HOV lanes


The Boston MPO recently released a memo titled Screening Regional Express Highways for Possible Preferential Lane Implementation. Quote:
The preferential lanes envisioned in this study would usually constitute specialized, separate lanes, generally constructed in the median of an expressway. The adjacent main travel lanes could continue to accommodate their full lane capacities, up to around 2,200 vehicles per hour, depending on the design standards of the roadway. With preferential lane eligibility set to allow fewer vehicles, perhaps 1,500 per hour, buses and other users of the preferential lane would be able to travel at posted speeds. Depending on traffic conditions, users of the general-purpose lanes would continue to experience delays either as a result of heavy traffic or from queues forming at bottlenecks.
It is sometimes suggested that an HOV lane could be implemented by converting a general-purpose lane for the exclusive use of HOVs. Since it is assumed that any preferential lane eligibility rules would result in fewer vehicles in the preferential lane than in the general-purpose lanes, the result would be a reduction in total expressway capacity. Reducing the capacity of a congested expressway would seriously worsen congestion and queuing within and leading to the capacity-reduced corridor, as well as on nearby surface roadways.
Here's the author's reasoning:

  • General purpose lanes have a capacity of 2,200 vehicles per hour
  • HOV/preferential lanes should be limited to 1,500 vehicles per hour
  • 1,500 is less than 2,200
  • Therefore, general-purpose lanes should never be converted to HOV lanes.
Have you spotted the flaw yet? It's the measurement unit: vehicles per hour. The author considers that all vehicles are equal with regards to congestion. Therefore a single-occupancy car is equal to a four-person carpool, and both are considered equal to a bus carrying 50 people. This memo blithely ignores the entire purpose of HOV lanes -- which is to expedite the movement of people, not pieces of metal.

Monday, March 11, 2013

Guest Street and the 64 bus

Guest Street Planning Study height and density map. Purple lines are bus routes, yellow stipples show walking distance from Union Square up to a quarter-mile. The density is far from the square.
The recently proposed car-free apartment building on North Beacon Street happens to fall within the Guest Street planning area which was re-zoned last year by the BRA and a local group of residents in the production of the Guest Street Planning Study. The main feature of this area is going to be the $500 million investment by New Balance to rebuild their headquarters and create a mixed-use district. The planning study features many forward looking ideas in an attempt to tackle the difficult problem of building a brand-new neighborhood on formerly industrial land. For example, it will permit heights up to 150' and FARs of up to 4.0 in an effort to bring life and vitality to an area devoid of anything currently. Unfortunately there are some problems, which the study readily admits:
Factors that are detrimental to its attractiveness include the area’s relative lack of public transportation access to downtown and the western suburbs, lack of identity and visibility from major streets in the area, access that is less than clear, and a public realm that is lacking.
There are two bus lines which can be said to pass through the study area, both of them infrequent and unreliable: the 64 and the 86. However, two key bus routes are nearby, both of them top-5 in ridership: the 57 and the 66. The study says:
Preserve and expand bus service. In order for the Study Area to develop at all, it is imperative that MBTA bus service be improved to provide access to Downtown, the Back Bay, Longwood/Fenway, and Cambridge. 
But it does not have any specific recommendations, beyond the new commuter rail station which is now slated to be built. There are also New Balance employee shuttles to Harvard and Kenmore, which will be upgraded according to the PNF, but they do not seem likely to be usable by the general public.

The chart at the top of this post shows the permitted heights and FARs for new development in the planning area, overlaid with the public transit infrastructure currently available. The location of the future commuter rail station at Everett Street is also shown. The yellow stipples reaching out from Union Square trace a walking path of approximately one quarter-mile (~400 meters) from the center of the square. The key problem can be easily observed: almost none of the higher density parcels can be reached within a short walk from Union Square. This seems to be the result of an unfortunate compromise which means that the existing transit resources will be furthest from the developments which need them most.

A quarter-mile is a bit of an arbitrary choice but it's a commonly cited rule of thumb when trying to figure out how far people are willing to walk to transit. What's also as important is frequency: the kind of high frequency that the key bus routes 57 and 66 provide will attract people to walk to them for many purposes. But the infrequent (and often unreliable) 64 and 86 buses will probably only be tolerable for rush-hour commuting purposes, if even that. The commuter rail station, while a net benefit, will also only provide infrequent commute-oriented service.

At this point, the question is: will the Guest Street area evolve into a car-dependent insular subdivision, or will it be able to integrate into the city's transit network in a way that reduces demand for automobile travel? In the former vision, the density will come paired with enormous parking garages (in addition to what's already being planned) and will be stifled by the already jammed capacity of North Beacon Street and Union Square. But if good alternative transportation options are supplied, then we can avoid that fate while adding much needed housing and commercial space.

I believe that to achieve good public transportation in the Guest Street area, the 64 bus route will have to be boosted and remade from a meandering coverage route into an efficient ridership-oriented line. The 86 could also stand to see some improvements. And New Balance should consider coordinating their shuttle operations with the MBTA.

Let's take a closer look at the 64 bus as it currently exists:


View in a larger map

The dark blue line is the route of the 64 bus between Oak Square in Brighton and University Park/Central Square/Red Line connection in Cambridge. The bus travels another 2 minutes to Kendall Square at rush hour. In Cambridge there is a one-way pair formed by the Western Avenue and River Street bridges. The translucent blue strips are connecting bus routes. The yellow areas on the map are sites that have been identified for development in the near or long-term future. In Allston/Brighton we have the Guest Street area, the Harvard IMP area, the Beacon Park area, and in Cambridge the K2C2 study area. The nice thing about the 64 is that it can help connect these new developments with access to the existing jobs and activities in Kendall, Central, Union, Oak Square and Allston Village, and an easy link to the Red Line and all that implies.

Nowadays, there is no consistent frequency of the 64 bus, its schedule is largely dictated by the time it takes for the assigned vehicles to complete their roundtrips. At the weekday peak, there are four buses providing approximately 23 minute headways, which reduces to two buses on 30 minute headways off-peak. In practice the headways vary from as little as 15 minutes to as much as 30 minutes during the peak, and can reach 45 minutes off-peak. The weekend service was recently reduced and features hour headways and nearly non-existent service on Sundays. It is the portrait of a bus route in decline, although the ridership is surprisingly resilient in my experience.

If the Guest Street area is going to develop as a real neighborhood instead of an island in a sea of parking lots, then the 64 bus is going to have to change. There is a mutually reinforcing positive relationship that needs to be jump-started: transit oriented development requires good transit, but the cost of providing that good transit needs good development to justify it. The service improvements will probably have to be phased and given a chance to build up ridership. Here's a few suggestions

  • The Hobart Park jog needs to go. The route forms an S-shape just north of Oak Square, where the bus suddenly turns away from Faneuil Street and sweeps around Hobart Park and under the Mass Pike. This could be accomplished in two ways, either by turning directly onto Brooks Street from Faneuil Street, or by just staying on Faneuil Street all the way to Market Street, which has the benefit of shortening the trip. This change saves resources, makes the route more sensible, and avoids tiny Hobart Street which is really not well suited to carrying buses.
  • A decision needs to be made about the current Guest Street segment, which now serves the Stop-n-Shop supermarket. Either it needs to be eliminated in favor of a pure North Beacon Street trip, or the Guest Street-running portion should be lengthened all the way to Market Street in order to serve the new developments more closely.
  • Cambridge may need to reconsider the design of the Central Square bus station, which accommodates the 64 bus. The geometry of that intersection is the reason why the 64 bus must currently divert onto Magazine Street before arriving at Central Square, instead of the more intuitive River Street approach. Increased frequency may require shuffling things around.
  • It should be possible to provide consistent 20 minute off-peak headways by adding a single bus to the route. This starts to bring it within the reasonable range for all-day usage, at least for the initial phase.
  • If the other suggestions are adopted and the route is shortened, then using the T's current assumptions, it should be possible to provide 15 minute peak headways by adding only two buses to the peak hour fleet. Again, not ideal, but getting better.
  • Other ideas could include: a 64A which short-turns at Market Street and heads back to Kendall Square; or, split the route into a coverage-oriented route 64 and a ridership-oriented route 63 (to pick an available number),
I've drawn a possible route on this map:

View in a larger map


In conclusion, the Guest Street planning study recommendations have a lot of potential to revitalize a lifeless piece of Allston/Brighton. It is unfortunate that the greatest intensity of planned development is far away from the existing transit resources, and perhaps that should be addressed directly in a revision to the study. But regardless of that, there will be a significant amount of new development taking place outside the quarter-mile "walk shed" of existing frequent transit lines. The commuter rail station will help a little bit but it is no substitute for a frequent route. The 64 bus route must see significant upgrades if this area is going to be able to reach its full potential.

Saturday, March 9, 2013

Larger vehicles should have fewer stations-per-mile

On the topic of station consolidation along the Green Line, one thing I've always noticed is that the 57 bus tends to outpace the "B" branch on the shared corridor except in the case where the bus has to make every stop. The train makes every stop on almost every trip except late in the evening when demand is lower. The difference can be quite stark: I've made it from Park Street to Harvard Avenue in under 20 minutes after midnight, but I've been stuck for over 50 minutes on the same leg during the afternoon rush, on what should nominally be a 25-30 minute trip.

The reason why I (and others) advocate for station consolidation is to reduce the fixed overhead associated with making a stop. But I think there's a more general case to be made. Let's look at the capacities of the vehicles. The planning capacity of a typical MBTA bus is 54 people (75 at crush load) and the capacity of a typical 2-car Green Line train is 202 people (468 at crush load). The train could easily be carrying between 4 and 6 times the number of people as the bus.

Now consider two stations that are too close together (for example, Saint Paul Street and Pleasant Street). There are people who will prefer one or the other station for small reasons and will be inclined to ring the stop request bell at that station. On the train, the probability is much higher that a person will ring the bell for the first station and also another person will ring the bell for the second station. On the bus, with fewer people onboard, it is likelier that one or the other can be passed up.

My experience bears this out: the Green Line rarely skips stops under normal service; the 57 bus will almost always be able to skip some stops even at the height of rush hour.

That accounts for alighting, but what about boarding? The likelihood that someone will be waiting at the station and flagging down the vehicle is related to the "strength" of the route: how well it attracts riders. The top 3 characteristics that make it attractive to riders are: frequency, where it goes, and perceived reliability. The Green Line is scheduled to run at higher frequency than the bus during its entire span. In addition, the train seems to draw more riders than the bus for a number of reasons, including: it goes downtown, it has a larger presence in the minds of riders (unfair but true), and higher perceived reliability (at least for those folks without smartphones). Again, if you have two closely spaced stations, then for small reasons people will prefer one or the other and the train will likely have to stop at both. But in this case it has less to do with capacity (except, perhaps, if riders fear they may be passed up by a full bus).

So to wrap it up, I think there is a natural bias towards longer station spacing (up to a reasonable point) for larger capacity vehicles, where the stop request bell is going to be pressed virtually all of the time anyway. In addition, people tend to be willing to walk further to faster, higher quality service, and spacing out stations further will speed up the Green Line a bit. Building higher quality platforms with good access at both ends (something the Green Line sadly lacks, but ADA requires) will also help shorten those walking distances while not sacrificing service quality.

Monday, March 4, 2013

Against the proposed widening of Melnea Cass Boulevard

The Friends and Neighbors of Melnea Cass Boulevard have put out an excellent op-ed piece: Widening Melnea Cass a bad idea.
As concerned members of Friends and Neighbors of Melnea Cass Boulevard, we write to highlight an urgent situation that many Roxbury residents may not be aware of. On Wednesday March 6, 2013 [at the Boston Water & Sewer Headquarters] an important public meeting will be held by the Boston Transportation Department to discuss its plan to widen Melnea Cass Boulevard (MCB), an ill-conceived project that will be detrimental to Lower Roxbury. The project would widen MCB by about 40 feet to add bus lanes, providing little public benefit because there are currently no plans to augment existing MBTA bus routes. The proposed project is at best a knee-jerk response to the availability of federal monies designated for such planning, which might be withdrawn if not used in the near future.
A little history: the land through which MCB now runs was once a neighborhood, but it was bulldozed and cleared as part of "urban renewal" and in preparation for the never-built Inner Belt expressway.

Comparison of the urban fabric in 1955 and the desolation found in 1995 (thanks to HistoricAerials)
The Inner Belt was stopped, but not in time to save this portion of Roxbury. Since then, for whatever reason, the land has lain underused. You can see in the pictures that a crosstown boulevard was established where none was before. The land around this boulevard is largely vacant or some sort of truck or car parking lot. For an area which is just over 2 miles away from downtown Boston, this seems strange and unnatural. The nearby Dudley Square has also suffered since the Orange Line was relocated to the Southwest Corridor about half a mile west. The question weighing on everyone is: what can be done about this situation?

Here are BTD's primary talking points for this project:
Transportation should work well for all modes:
  • Safe, efficient, calmed traffic
  • Lower speeds
  • Good bicycle and pedestrian connections
  • Effective transit
And one of the proposed cross-sectional designs:
And here's a sample of a proposed intersection:
I'm having a great deal of trouble reconciling these pictures with the notion of "safe, efficient, calmed traffic", "lower speeds" and "good pedestrian connections."

You're looking at 7 lanes of traffic just to get across the street. That's on the order of 80 feet of pavement. If you want to know how that will feel, just go to the corner of Melnea Cass Boulevard and Tremont Street currently. MCB widens out to about 77 feet at this point, and to make matters worse, Tremont is also really wide. It's not pleasant to be a pedestrian there. This plan looks a lot like Commonwealth Avenue just east of the Mass Pike bridge. That was rebuilt not too long ago. It used to be even more dangerous. It's been improved, but it's still pretty nasty. And that's with a much wider transit reservation.

Also, just in case you didn't have enough pavement, there's actually 9 lanes with parking. Those sidewalks are looking awfully tiny. Unfortunately, tiny sidewalks are a widespread problem in Boston, but why perpetuate that on what is supposed to be a "Complete Streets" design?

Where did those tall buildings come from in the rendering? Always fun to see what the artists imagine. But let's think about it. Suppose that kind of investment was attracted and allowed to flourish along MCB? What are the possibilities?

  1. Large parking lots are built under, around or behind those buildings; as current zoning dictates. The street environment stays car-oriented. Therefore everyone drives, all lanes of traffic are jammed, and the bus ridership is anemic. Nobody is happy, but drivers stare hungrily at those bus lanes and commence the "empty lanes attack" through political channels. Eventually, the bus lanes get turned over to general traffic, just like so many trolley lanes before them. The greenspaces become neglected and dangerous.
  2. A thriving transit-oriented development corridor is created. The bus lines function smoothly and efficiently so people use them and don't feel the need to use their car for every trip. Heck, suppose it even gets used for the Urban Ring! But still plenty of traffic is attracted to those lanes. Either they're uncongested and traffic flows at dangerously high speeds, or enough through-traffic is induced to jam them anyway. Either way, the majority of people making use of the corridor are on foot and yet they are squeezed into a small amount of the space, and put at risk.
BTD's plan is to take the existing 120' right-of-way and re-arrange it in such a way that it adds an extra 24' of pavement in the center for the bus lanes. This implies shifting the north side of the road further north onto the current bike path and line of trees. If they intend to keep MCB as a car-oriented corridor while adding a set of dedicated bus lanes, this is probably the best way to go about it, although it does sacrifice a lot of trees.

Is there a better way? I think so. If the neighborhood wants to become more walkable, more bike-friendly, and more transit-friendly, then I think the best solution is not to widen the pavement at all. Instead, if BTD really wants to show its dedication to "Complete Streets" then they will take two of the existing travel lanes, and convert them into dedicated bus lanes.

Some might argue the old trope that this will cause "gridlock" and traffic jams as drivers senselessly pile into one another without thought. That's nonsense. If the corridor is going to be reimagined and reinvented, then it will also change the way people travel along the corridor just as radically. People respond to their environment. If you want to have a highway, then by all means, keep the road wide. But if you want it to be a walkable place to build a neighborhood around, then you'll want to keep the lane count down.

My plan also saves the trees and costs a whole lot less money. It can be implemented with paint. Then, the city needs to allow the development of the vacant lots and the parking lots. Lower or outright remove any minimum parking requirements in the zoning code to help ease traffic. This area has been neglected by the city for far too long. It's time to clean up the mistakes of the 1960s, not repeat them.

Sunday, March 3, 2013

Light rail and traffic

Streetcars on Brighton Avenue 1940 
A note on Brad Plumer's article Do light-rail systems help cut down on traffic?. Anyone familiar with the Downs-Thomson paradox and related concepts would understand that light rail or any public transit system is not a "congestion reliever" in the simple sense that most people use. That's because in most congested contexts, any driver who decides to leave the car at home and switch over to transit will likely be replaced by another car on the road. This is not a hopeless outcome however--where before there was one person able to travel, there are now two. So the overall level of transportation activity has increased even though road congestion has remained the same, in this scenario.

Public transit in dedicated lanes can enable more people to travel despite congestion on the roads. If you want to actually tackle the congestion itself, then as Brad points out, road pricing seems to be the only option that has been proven to work. But I want to get to his last point which is this:
The other point is that mass-transit systems can lessen road traffic, but only if they’re part of a broader shift by a city to move to denser development. Researchers have found that people living in more compact cities with better transit options do drive less overall. But that shift doesn't happen overnight. And not all light-rail systems are even built with this sort of development in mind.
Boston happens to be one of those places with "light rail" and denser development built up along those corridors, thanks to the fact that most of it was built before zoning became a weapon used against cities. So if you take a look a Commonwealth Avenue you find something interesting. Despite the fact that the road has 6 lanes of travel (4 inner, 2 outer), MassDOT records very low traffic levels on the order of 12,000 AADT. By comparison, Washington Street in Brighton Center carries 14,500 AADT with only 2 lanes, and Harvard Avenue is all the way up at 18,100 AADT with only 2 lanes as well. Why isn't Commonwealth Avenue more like Brighton Avenue which has 4 lanes and sees volumes of 23,000 to 28,000 AADT?

First, for a variety of reasons, Commonwealth Avenue (in Allston/Brighton) is not a heavily used through-route, unlike Harvard Avenue or even Washington Street, so the traffic on it is largely coming from or going to nearby places. Second, although the development along the avenue is at city levels of density (~80-120 dwelling units/acre), because it grew up around the streetcar, there's not as much parking as such quantities of units would bring under current zoning codes. So trips are much less likely to be made by automobile. This seems borne out by the fact that the current "B" line is the busiest branch (~30,000 boardings/weekday) and that just over 50% of residents who live along this corridor get by without access to a car.

So in congruence with Brad's point, I think that you have here an example of how a century of streetcar service, combined with the long-ago shift to denser development, has lessened road traffic in the present day. Now the question for many developing light rail systems is: can you also achieve that, and perhaps on a shorter timescale?