More on Cost Comparisons

Some of my past posts on cost comparisons are getting play on mainstream publications including Slate, Salon, The Economist, and The Atlantic Cities, and one of the consistent points I see is that the difference between the US and most of the rest of the first world is so glaring that projects that are locally considered boondoggles suddenly look good.

A list containing multiple projects at over a billion dollars per kilometer can legitimize anything below it. Thus projects approaching half a billion per kilometer look downright reasonable. In reality, Tokyo Metro said that there will not be further subway construction, and I have read elsewhere that it repeated this promise in advance of its impending IPO. And in Amsterdam, an inquiry into the North-South Line’s factor-of-2 cost overrun concluded the project should have never been built.

It seems that there’s an Overton window analog, in which higher costs legitimize the previous decade’s work, making it look good when at the time it was criticized for poor cost control. In fact, this could explain the decades-long trend toward increasing real costs – an explanation that is usually given in terms of rising wages and worker safety rules, but in reality poor countries build subways for not much less money than non-Anglophone rich countries.

I contend that the best practice should still be to compare with the average, rather than with either the worst (London Crossrail, Amsterdam North-South Line, Munich Stammstrecke 2) or the best (most projects in Spain). Being more expensive than one city could be a fluke. Being five or more times more expensive than upward of 90% of subway projects is less excusable.

Most interesting to me in this discussion is the explanations for US/Europe cost differences. Although most people regrettably keep comparing the US to China, never mind that European and Chinese costs are similar, some stay on target and avoid explanations that assume the entire first world is like the US. One comment on The Economist follows:

Observations on the public construction process, having seen it in action relatively up close:

1. Failure to embrace technology except in the most expensive cases. We are behind in construction techniques overall. We will bring in European methods when the case is made they are necessary. These methods therefore tend to be used when the expense is higher. This means we don’t upgrade technology overall, just at the costly fringes. Examples come from the methods used to construct the new tunnels in Boston; one used a method developed mostly by the Dutch because our domestic methods weren’t up to it.

2. Our project management is not equivalent. European large scale construction projects run more just-in-time. Even really big ones require very large things to be built and then to arrive on a schedule. Our system can’t handle that so we build in lots of slack expecting stuff will come late and will need to be adjusted – sometimes substantially – to fit the need. That is very costly.

3. Our system is very bad at prioritizing. My experience with this is mostly at the state and local level. I have seen very competent people working at both levels. They exist in a morass of work that needs to be done. They don’t have the resources to do things properly. They have to put repair, snow, etc. way, way, way ahead of planning.

4. My overall comment is this: Europeans understand they exist in a high cost environment so they squeeze out the inefficiency to be competitive. They focus on value-added design and on efficiency in planning and scheduling. We don’t.

Bad Defunding

The furor in the transit blogosphere about the House Republicans’ transportation bill proposal, defunding the Mass Transit Account and diverting the money to roads in order to shore up the Highway Trust Fund, deserves more scrutiny from the point of view of government effectiveness. Although the proposed defunding is clearly political and cultural (to many Tea Party Republicans, trains and buses are for hippies), the way in which it is done is a good reminder about what’s going on in US politics. The principle is that when government does not work, the people in government who propose to get rid of it are part of the same ineffective governance structure.

First, consider some recent projects or proposals for projects to expand transit. In Houston, as noted by rail critic Tory Grattis, of the four proposed light rail lines, the first two to be built are the less cost-effective two, while the more compelling Universities Line is saved for later. In Los Angeles, the Foothills Extension is being built before the Westside Subway. And the first California HSR contract to be tendered is for the northernmost initial construction segment, the segment that should be first on the chopping block if necessary to divert money to the more important Los Angeles-Bakersfield mountain crossing. In order to prevent smart scope changes from leaving the cost-ineffective parts out, the planners take the cost-effective lines hostage in order to make sure that they are built.

The same is true in the opposite direction. It was not so clear up until now because the big-ticket rail cancellations all involved just one major project per governor, but now that the Republicans are bundling all transit together, the pattern looks clearer. There is no accounting for good and bad projects here. Even cost-effective projects such as Second Avenue Subway would lose funds they need for completion. The goal is not to cut government waste; it’s to cut spending on people who the people who proposed the bill find distasteful, and the effect of quality of government services is about what you’d expect of any politicized government program.

It’s for the same reason that I think very little of Chris Christie’s ideas about transportation management, despite my criticism of ARC. Although some of the rumors floated suggest that the depth of the cavern was one of the things that made Christie think the project was ridiculous, he made no effort to try fixing it. Instead, he canceled the project, trumpeted it as a cost saving measure, and proceeded to spend more money on freeways. Aaron Renn, hardly an orthodox leftist (he is a fan of Mitch Daniels), compared him to Chainsaw Al and called his style of governance kindergarten-level.

In analogy with technical and political transit supporters, there are technical and political road supporters. The technicals exist, in various driver magazines that support spending more money on maintenance and less on expansion, and do not mind raising tolls or gas taxes to pay for infrastructure. But the politicians do not care about cost-effectiveness, and have no problems supporting Big Dig-style projects. At the risk of overinterpreting blog comments, let me say that every road supporter I have seen express an opinion on the Big Dig thinks that it was a necessary project and should have been built regardless of the cost (for the record, $15 billion for what to the best of my understanding is about 200,000 cars per day). The thinktanks that support them care more about finding ways to convince people to want to pay more money for expensive freeways rather than about cutting the cost of construction or reducing environmental impacts.

Thus the House transportation bill is bad not only because it’s bad for transit, but also because it’s bad government. It’s not even selective worrying about cost-effectiveness, a charge often thrown by political transit supporters. It makes no attempt to decouple any funding from gas taxes, a decoupling that it necessary for the purpose of making it possible to tax pollution without demands from both APTA and the AASHTO that the revenues raised be plugged back into transportation. It makes no attempt to let go of projects that cost too much while maintaining those whose cost is adequate. It’s purely an exercise in muscle-flexing, a continuation of the US practice of not having a transportation policy that’s separate from the usual political and lobby bickering.

Why Long Island Should Get An HSR Spur

Having looked into why high-speed rail from New York to Boston should go through Providence, I want to explain why it should go through New Haven, rather than through any of the fanciful Long Island routings proposed most prominently by the Penn design group. Like Hartford, Long Island should have high-speed trains use the LIRR Main Line, but at medium speed rather than high speed, and with careful consideration to the much more important needs of commuter rail.

Although the LIRR Main Line shares one characteristic with the New Haven-Springfield line, namely that it is very good for 160-200 km/h but bad for 300, the reasons are subtler and less geometric. The most visible is NIMBYism. Even increasing the traffic of existing LIRR trains raised the ire of some suburbs along the Main Line, which opposed the three-tracking project (since canceled due to budget shortfall) on the grounds that extra train traffic would reduce quality of life and that eminent domain would be required. This is not Caltrain, whose local residents do not know what electric trains sound like; this is Long Island, which has lived with these trains for generations. Introducing HSR is asking for trouble.

Of course, the same could be said about any suburb that HSR needs to pass through. Connecticut is full of NIMBYs, just like Long Island. The reasons usually given for avoiding the existing Shore Line are that it’s too developed and has too much local opposition. But those are present on Long Island, and are worse because of the higher population density. For examples, compare Westport and Cos Cob with Brentwood and Farmingdale. The LIRR offers multiple straight rights-of-way, but all are going to have the same speed limits as heavily upgraded and modified tracks on the Shore Line – 250 km/h in the better parts, and 200 in the worse parts.

The Penn design proposal is not even the best Long Island proposal, for three reasons:

1. It insists on proceeding from Penn Station to Jamaica on the Lower Montauk Line. If a connection from the line to Penn Station opens, it’ll be far more useful for local rail, while intercity rail can use the Main Line. The difference between appropriating a Manhattan-accessible Lower Montauk Line for HSR and replacing the Lexington Avenue Line with a truck tunnel is one of degree, not kind; in both cases, local passenger rail is the most valuable use of the infrastructure.

2. It departs from the Main Line to use the Hempstead Branch (necessarily eviscerating commuter service) as well as abandoned tracks through endless residential suburbs, full of urban grade crossings. The Main Line has grade crossings and would need to be four-tracked, but the local NIMBYs actually supported grade separation, and multi-tracking at least could be sold as the local transit improvement project that it is.

3. Last and worst, it sharply veers north after stopping at Ronkonkoma, along a curve whose radius judging by the alignment map is around 900 meters (=150 km/h if superelevation and cant deficiency are set at normal HSR levels, or 170 km/h at cutting-edge levels). Then it crosses the Long Island Sound at its widest, so that it adds more than 20 kilometers to the New York-New Haven route length over the Shore Line, all at medium speed.

A route similar to the Penn design route but using the more feasible Main Line alignment would be 9 minutes slower than the optimal Shore Line route – 41 versus 32 minutes – with stops at Jamaica and Hicksville, enforced by unfixable track curvature near the stations. But in addition to the extra travel time, fixing the alignment through New Rochelle, Darien, and Bridgeport is far cheaper than a long undersea tunnel. A better Long Island route would follow the Main Line to the end and tunnel near Greenport, trading deeper waters for shorter tunneling and a route length comparable to that of the optimal Shore/I-95 alignment, so it could achieve a comparable trip time. But even that’s unneeded: it’s 15 km of deep tunneling, whereas if one is willing to slightly compromise on trip times, the only Connecticut tunneling required for a Shore Line fix is 3 km in Bridgeport.

The other problem is what to do about commuter service. The Providence Line’s traffic level is low enough and its average interstation is long, allowing a blended plan. Shared tracks between New Rochelle and Penn Station would see more commuter traffic, but intercity trains would go slower anyway, and there is more room for four-tracking. The Ronkonkoma Branch’s 10-minute peak service requires at least one overtake between Hicksville and Ronkonkoma and probably two, in addition to four-tracking the Main Line; this is feasible, but less than optimal, and the overtakes would have to be constructed in more constrained locations than those available on the Providence Line. East of Ronkonkoma commuter service may need to be cut, but this is less of a problem on account of its low traffic. On the other hand, the Main Line west of Hicksville is not a problem with four tracks, and neither is the New Haven Line – express commuter trains could weave in and out.

On the benefits side, offering Long Island service to Boston that doesn’t go through New York is better than not doing so. However, the difference in benefits with New Haven, while positive, is smaller than it seems. The New Haven Line has almost as much ridership as the LIRR Main Line, and Stamford is a bigger edge city than Mineola and Garden City. On top of that, since the optimal LIRR option connects to the Shore Line in the far east of Connecticut, there is no hope for service to Hartford except on legacy track. On balance, the advantage of the LIRR option is just service to Jamaica, a larger draw than those smaller cities and suburbs, but there the time saving is the smallest.

On top of that, does such a small benefit really justify the cost? Having some high-speed trains run through to Jamaica, Mineola, Hicksville, and Ronkonkoma at lower speed requires re-electrifying the LIRR with catenary, which is a fraction of the cost of all those urban grade separations and 1-2 order of magnitude cheaper than an undersea tunnel and land connections. On a similar note, since half an undersea tunnel is of no use, it’s harder to break construction into small chunks if it is necessary, putting it at a disadvantage against a route consisting of cutoffs and modifications of the existing line. The route of 1834 may work now that we can build tunnels, but the cost structure favors that of 1846 and 1852.

Formal Announcement

Those of you following my Twitter feed already know this: I’m moving to Vancouver sometime in this summer for a two-year position at UBC. Expect a dramatic reduction in the number of posts expressing exasperation with local transit planning beginning sometime in July or so. I will try to keep writing about the Northeast and not just about Vancouver.

High-Speed Rail Should Serve Providence

The most straightforward part of constructing greenfield tracks for high-speed rail on the Northeast Corridor is east of New Haven. There are good legacy lines to hook into, and good Interstate corridors to follow when the legacy lines are too curvy. It’s also the segment with the biggest variation in alignment options, which boil down to going through Hartford and going through Providence. Both the Penn design proposal and the Amtrak proposal go through Hartford and avoid Providence, and this is a bad idea, for both costs and benefits reasons.

See here for a very early and rough draft of my HSR map, which goes through Providence; there are significant issues with this map west of New Haven, but it’s fairly accurate east of Haven. It uses I-95 between New Haven and the state line, and transitions to the legacy line around Kingston; Hartford would be served on the legacy line, which would be electrified. I have not seen detailed drawings of Amtrak’s proposal, but here is the detailed Penn design map, going through Hartford: the idea is to use a combination of I-91 and a heavily upgraded legacy line, and then transition to I-84 in Hartford and then I-90, while retaining the Shore Line for slower service to Providence. The latter option turns out to be inferior, essentially because full HSR is easier to build through Providence whereas a medium-speed branch is easier to build to Hartford.

First, the cost side. Because the portions of the Shore Line used by the Providence option are straight and already built to high standards, minimal upgrade work is required there. The bulk of the cost would be constructing high-speed track along a mostly flat, not very developed right-of-way, with two and a half painful segments (New London, the cutoff east of New Haven, and the Connecticut River crossing as the half). East of New London the median is available, cutting costs further. All in all, this is 125 km of largely at-grade track, and about 60 km of cheap electrification to Hartford.

Going through Hartford is about equally hard. The New Haven-Springfield line is built to low-speed standards, with grade crossings and curves that are good for 200 km/h rather than 300. It avoids the river crossings of I-95, but I-84 and I-90 are a bit curvier and follow more rugged and urbanized terrain, and the urban segment through Hartford looks harder than that through and immediately east of New Haven. Per kilometer it could cost about the same, but 200 km of new track are required.

The costs by themselves are not a huge deal. The New York-New Haven segment requires new grade-separated junctions, multiple bypasses, and some urban tunneling. In contrast, mostly at-grade track costs $20 million per kilometer or not much more, so despite the large difference in length, the difference in cost is about $2.5 billion vs. $4 billion.

However, there’s also a constructability argument for I-95, which is that it can be done in segments more easily, using portions of the Shore Line before the full line opens. This could be useful if money were made available in very small chunks. The Hartford route could be done partly on an electrified Springfield line, but Hartford-Boston has to be done in one go.

But a bigger issue is that going through Providence has two advantages over going through Hartford without regards to costs. First, Providence is a larger city than Hartford: its metro area is about 20% larger than Hartford’s, and the central city is 40% larger and denser. Although the Hartford option passes near Worcester, there is no way to bring a station into Worcester itself without excessive tunneling; the Penn design plan puts the station at the edge of the built-up area, 7 kilometers from downtown Worcester. The Providence option passes through much smaller New London, but it can at least be served by a station that’s within the city, one km from the present station.

The other advantage is how to serve the city that does not get to be on the HSR mainline. The Springfield line is easy to upgrade, since it is straight enough for medium speed, and grade crossing protection good for about 180 km/h is relatively cheap. This would give Hartford very good service to New York – about half an hour to New Haven, and a little more than another half an hour to New York. The Shore Line in contrast is curvy and slow and already has a fair amount of superelevation and cant deficiency, making future upgrades much harder. Providence would still get better trip times than today coming from better rolling stock and higher speeds west of New Haven, but better trip times than about 1:45 to New York are only possible with trains with high degree of tilt, which tend to be a maintenance nightmare.

For the record, my original proposal above is from 2009, and I only accepted my current job at Brown in 2011. However, in the interest of full disclosure, by 2009 I already knew that Brown had one of the best departments in my field, whereas Hartford doesn’t have a research university of comparable quality. I don’t think it biased my choice – the idea of following the present alignments and serving present lines as much as possible appears elsewhere in the plan as well as in my regional rail proposals for New York and Boston – but then again nobody thinks their own choices are biased.

MBTA Mode Shares

As a followup to my claim in my first post about improving the MBTA about the low mode share of commuter rail for trips into Boston, here are some figures about commuter rail use, by sector. All numbers exclude commuters from inner suburbs and from Boston itself, since those would use the subway. Only Boston-bound commuters are included. I’m providing wider and narrower numbers, wider numbers corresponding to the entire sector defined by a commuter line and narrower numbers including only towns within reasonable range of a commuter station.

All commuter rail numbers come from the Bluebook and are averages as of February 2009, from page 74; be careful not to use numbers from the map on page 70, as they inflate the ridership on the Providence Line. All commute market numbers come from the 2000 census; I do not believe commuting patterns have changed so radically as to significantly alter the picture.

Update: we obtain the following table, explained below (see also computation error fix for Haverhill):

Line Wide market Narrow market Riders Share of wide Share of narrow
Old Colony 43,587 34,934 20,907 48% 60%
Providence, Stoughton 23,297 17,490 11,719 50% 67%
Franklin 14,899 12,747 7,043 47% 55%
Worcester 19,997 15,581 7,479 37% 48%
Fitchburg 16,544 13,358 5,883 38% 44%
Lowell 15,551 11,912 5,586 36% 47%
Haverhill 19,196 16,902 8,922 46% 53%
Newburyport, Rockport 26,926 25,534 13,230 49% 52%

Old Colony Lines (including Greenbush)

Commute market (wider): all of Plymouth County; Cohasset, Weymouth, Randolph, Holbrook, and Avon towns in Norfolk County

Commute market (narrower): including only the above Norfolk County towns and Plymouth County’s towns of Abington, Bridgewater (including East and West), Brockton, Hanson, Hingham, Hull, Kingston, Lakeville, Middleborough, Plymouth, Rockland, Scituate, and Whitman

Commuter volume (wider): 43,587

Commuter volume (narrower): 34,934

Inbound commuter rail ridership: 12,065 (excluding JFK-UMass, Quincy, and Braintree)

Extra transit use: 7,244 subway and commuter rail boardings in Quincy and Braintree beyond those accounted for by those two towns’ commuter market; 1,598 commuter ferry riders, mostly from Hingham

Mode share: 60% of the narrow market, 48% of the wide market

Providence and Stoughton Lines

Commute market (wider): all of Rhode Island, all of Massachusetts’ Bristol County, and the towns of Sharon, Canton, Stoughton, and Plainville in Norfolk County

Commuter market (narrower): the three Norfolk County towns omitting Plainville, all of Rhode Island’s Providence and Kent Counties, and Massachusetts’ Bristol County’s towns of Attleboro (including North), Mansfield, Easton, Norton, and Seekonk

Commuter volume (wider): 23,297

Commuter volume (narrower): 17,490

Inbound commuter rail ridership: 11,719

Extra transit use: none

Mode share: 67% of the narrow market, 50% of the wide market

Franklin Line

Commute market (narrower): Norfolk County’s towns of Dedham, Westwood, Norwood, Walpole, Norfolk, and Franklin

Commute market (wider): the above plus the towns of Medfield, Medway, and Wrentham

Commuter volume (wider): 14,899

Commuter volume (narrower): 12,747

Inbound commuter rail ridership: 7,043

Extra transit use: none

Mode share: 55% of the narrow market, 47% of the wide market

Worcester Line

Commute market (wider): all of Worcester County except Fitchburg, Harvard, Westminster, Gardner, and Leominster; Wellesley; and Middlesex County’s towns of Natick, Ashland, Framingham, Marlborough, Hudson, Hopkinton, Holliston, and Sherborn

Commute market (narrower): Wellesley; Middlesex’s above counties omitting Hudson, Holliston, and Marlborough towns; and including only Worcester County’s towns of Southborough, Westborough, Northborough, Shrewsbury, Grafton, Worcester, Auburn, and Millbury

Commuter volume (wider): 19,997

Commuter volume (narrower): 15,581

Inbound commuter rail ridership: 7,479 (west of Newton only)

Extra transit use: none, as all Green Line usage is accounted for by Newton commuters

Mode share: 48% of the narrow market, 37% of the wide market

Fitchburg Line

Commute market (wider): Worcester County’s Fitchburg, Harvard, Westminster, Gardner, and Leominster towns; Middlesex County’s Belmont, Waltham, Wayland, Weston, Concord, Sudbury, Acton, Maynard, Stow, Lincoln, Littleton, Boxborough, Ayer, and Townsend towns

Commute market (narrower): Worcester County’s Fitchburg and Leominster only; Middlesex County’s above towns, omitting Stow, Townsend, Wayland, and Sudbury

Commuter volume (wider): 16,544

Commuter volume (narrower): 13,358

Inbound commuter rail ridership: 5,883 excluding Porter

Extra transit use: two local bus lines (70, 70A) to Waltham with 4,343 trips in each direction and two more express lines (553-4) with a combined 681 trips (none serving just Waltham, but also Newton or Cambridge), but conversely much of the ridership Waltham generates is non-commuter student traffic

Mode share (assuming neutral Waltham effect): 44% of the narrow market, 36% of the wide market

Lowell Line

Commute market (narrower): Middlesex County only, towns of Winchester, Wilmington (half, shared with Haverhill Line), Woburn, Billerica, Lowell, Chelmsford, Tewksbury, and Dracut

Commute market (wider): the above, plus the Middlesex County towns of Dunstable, Westford, and Tyngsboro, and all of New Hampshire’s Hillsborough County

Commuter volume (wider): 15,551

Commuter volume (narrower): 11,912

Inbound commuter rail ridership: 5,586 excluding West Medford

Extra transit use: none – if anything, this line is missing more of its potential coming from its inability to serve Medford well under current commuter rail paradigms

Mode share: 47% of the narrow market, 36% of the wide market

Haverhill Line

Commute market (narrower): Middlesex County’s towns of Melrose, Stoneham, Wakefield, Reading, North Reading, and Wilmington (half, shared with Lowell Line); and Essex County’s towns of Andover, North Andover, Lawrence, and Haverhill

Commute market (wider): the above, plus Essex County’s towns of Boxford, Middleton, Groveland, and Lynnfield

Commuter volume (wider): 19,196

Commuter volume (narrower): 16,902

Inbound commuter rail ridership: 5,343 excluding Malden

Extra transit ridership: Malden’s two subway stations get 8,375 riders beyond the commute market, but they could be coming from people from Everett and Medford – let’s just add the number of parking spaces, which is 976, plus one half of the remaining ridership at Oak Grove, representing Melrose and Stoneham’s share, which is 2,603 (total bus ridership in Malden and the cities to its north is 3,394 in each direction, and I’m willing to believe 2,603 of this is from north of Malden)

Mode share: 53% of the narrow market, 46% of the wide market (update: a previous version of this post gave slightly lower numbers coming from forgetting to add the 976 subway parking spaces to the imputed ridership)

Newburyport and Rockport Lines

Commute market (wider): all of Essex County except the towns of Andover, North Andover, Lawrence, Haverhill, Boxford, Middleton, Groveland, and Lynnfield

Commute market (narrower): Essex County omitting, in addition to the above, the towns of Amesbury, Georgetown, Essex, West Newbury, Topsfield, and if one wants to be very narrow then also Saugus

Commuter volume (wider): 26,926

Commuter volume (narrower): 25,534 with Saugus, 22,999 without

Inbound commuter rail ridership: 8,821 excluding Chelsea

Extra transit use: no rail, but likely high bus ridership coming out of Lynn – total North Shore ridership is 4,409 in each direction excluding the 430, which is assigned to the Haverhill Line, and some lines that do not serve subway stations (429, 431, 435, 436, 451, 455, 465, 468)

Mode share: counting commuter rail only, 38% of the narrowest market and 33% of the wide market; counting also buses, 52% of the narrow market including Saugus and 49% of the wide market

Conclusion

The best-performing sectors capture about half the ridership of their general area of service and two-thirds of the ridership coming from the towns served by transit. Except for the Providence Line, this requires heavy use of subways and buses; the high contribution of commuter buses north of Boston suggests that better regional rail service, with more useful frequencies and integration between the CharlieCard and Charlie Ticket, really would make a difference, effectively railstituting the trip to Boston and allowing redirecting the buses to feeder service. The worst-performing lines only capture a third of the market, and for those there’s less that can be done; regional rail improvements would help, especially with regards to speed and reliability, but often those areas are too sprawled out.

Another conclusion is that for the purposes of constructing timetables for an electrified, FRA-free MBTA, we should ignore current ridership patterns, and instead look at the volume of commuting. The southwesterly slice serving Providence is not special; it just gets higher mode share on commuter rail since the service levels are higher than elsewhere. A 15/30 timetable on each line or branch (counting Kingston and Plymouth as one, but not any other split) would more or less approximate market demand; it would still underserve the South Shore, but the South Shore has ferry use and subway use, and therefore there’s no lower-level service to replace with good commuter rail.

A third conclusion is that it matters whether commuter rail can serve shorter trips or not, such as trips from Waltham, Medford, Malden, Lynn, and the Fairmount Line area. Those towns and neighborhoods have much larger volumes of commuters heading toward Boston than farther-out towns. This observation favors the regional rather than intercity interpretation of commuter rail, sacrificing speed in order to improve coverage in the innermost suburbs and in outer-urban neighborhoods.

On no line is express service a good proposition except when it’s downright intercity service, such as high-speed rail to New York via Providence, or even intercity trains to Maine or deep into New Hampshire. The problems coming from runtimes that aren’t a short turnaround time less than an even multiple of 15 minutes should not be fixed with express trains, but with closing some stations that aren’t necessary (such as River Works and Mishawum), and building more infill together with the North-South Rail Link, redistributing lines in such a manner as to maximize efficiency. For example, if the branch from Lynn to Marblehead is reactivated, then even with several additional infill stops, trip time from North Station to Marblehead would be about 22 minutes, which together with time gains from the tunnel (travel time through the rail link should be a hair lower than a turnaround time) would allow mixing with the Worcester Line. Currently it’d be difficult to do Worcester-Boston in 55 minutes even under best operating assumptions, and impossible with infill stations, but with the rail link and through-service to Marblehead, the limit would be closer to 1:01, and this allows a few infill stops in Brighton.

What’s driving all of this is the fact that there isn’t all that much demand mismatch. The South Side lines have a larger (wide) market than the North Side lines, but the difference is much smaller than the difference in ridership, and decreases even further if one lets the subway take care of certain parts of the South Shore. This means that, in terms of planning, all lines should be upgraded and should receive ample attention toward service levels, and in terms of operations, through-service should be based on infrastructure capabilities and scheduling constraints rather than on service demand matching.

The Limits of Clockface Scheduling

This is morally the last post in my series on improving the MBTA: see here, here, and here for the three previous posts. However, it’s a more general principle concerning interlined regional rail services.

Good practice for running transit service that isn’t at show-up-and-go frequency – say, anything that comes every 10 minutes or more, certainly anything that comes every 15 minutes or more – is to have regular clockface intervals. This is memorable for passengers, and works as a baseline with which to work on providing extra connections. In addition, if there is interlining, then it makes it easy to schedule trains to come at a uniform frequency on the share segment. If service is uniform throughout the day, then this is very easy. The problems start when it is not.

Normally, if extra peak service is required, then rigid clockface systems, such as those found in the German-speaking world, will usually interpolate in the middle of the period. In other words, if a station gets inbound trains at :00 and :30 every hour, then in the peak it will also get them at :15 and :45. This is what’s done in Stuttgart on two of the S-Bahn lines, interpreting peak very liberally, and less rigidly on the TER in Nice. Many systems instead use similar peak and midday service, dropping service only in the evening, such as the Berlin S-Bahn, and BART.

To see where problems could occur, let us look at Berlin again. There are three services on the Stadtbahn: the S5, the S7, and the S75. At the peak, all three run every ten minutes, with westbound trains departing Ostkreuz at :02, :00, and :05 respectively. Off-peak, the S75 drops to 20-minute frequencies, introducing 8-minute gaps into a schedule whose average headway is 4 minutes.

For a cleaner, contrived example, let’s say we interline two services, each with a 15/30 frequency; a factor-of-2 difference in frequency is more or less the norm on commuter lines in Tokyo and Paris, which do not have rigid clockface schedules – more local lines have a slightly smaller gradient than more long-distance lines. There is an inherent tradeoff between uniform frequency at the peak and uniform frequency off-peak. It’d be much easier to do if both services were bound to have the same frequency but the frequency varied continuously, as it does on most subways; however, what works on a dedicated line when passengers show up and go fails when passengers consult schedules and when timed connections or overtakes are involved.

More concretely, if Line 1 leaves a station at :00 and Line 2 leaves at :08, providing uniform peak frequency of 7-8 minutes, then off-peak we will have a 22-minute gap when we reduce to half-hourly frequency on each line; and if Line 2 instead leaves nearly at :15 to provide uniform off-peak frequency, then there will be a gap of nearly 15 minutes at the peak. The sum of the largest peak and off-peak gaps is necessarily 30 minutes, whereas the ideal would be for the sum to be 22.5 minutes.

Extra constraints can force one choice of gaps. For example, the Providence and Stoughton Lines are (or should be) constrained by the need to fit faster intercity trains on the line, at least in the future; for details of those constraints, see my posts on MBTA-HSR compatibility. In short, if we choose the symmetry axis to be :00, then Providence Line trains are compelled to leave South Station at :02 to meet up with trains to Woonsocket, and high-speed trains leave South Station at :10-11 and begin to overtake Providence trains at Readville at :15. Stoughton trains should then leave immediately after the high-speed trains so that they can leave the line toward Stoughton just before they’d get overtaken (at Sharon, if they continued), which means at :12-13. Thus we obtain about a 10-minute gap at the peak and a 20-minute gap off-peak, which is an acceptable compromise.

In contrast, one thing that clockface scheduling does not limit is short-turns. Indeed the Berlin S-Bahn often short-turns every other train, without trouble. Moreover, it is not difficult to drop to half the peak frequency with short-turns. If a train that leaves at :00 runs all the way to the end and a train that leaves at :08 short-turns (both repeating every 15 minutes), then it is not difficult to change things so that in the off-peak, a train that leaves at :00 or :30 runs to the end and a train that leaves at :15 and :45 short-turns. People beyond the short-turn point would still only need to memorize at what minute between :00 and :29 the trains serve their stations, and it would be regular all day; people before the short-turn point would again only need to memorize one number, from :00 to :14. In the case of the MBTA, this means that the Fairmount Line (which should get a train that turns at Readville for every train that continues toward the Franklin Line) can get a perfectly regular timetable.

Improving the MBTA: Electronics and Concrete

Where improvements in New York and other very large cities can easily include multiple new subway lines, the same is not true of Boston. The concrete pouring would be wasted, since Boston’s existing subway lines are not at capacity. The busiest line, the Red Line, has a peak frequency of one train every 4.5 minutes, which could be doubled with appropriate signaling improvements and more rolling stock if necessary. The Green Line has bigger issues coming from branching – its core segment already runs close to 40 trains per hour – but this could be resolved by obtaining fully low-floor vehicles and lengthening trains to allow one or two extra branches.

Another thing that Boston lacks and other US cities do is very busy bus lines to railstitute. Boston’s busiest bus line is the Silver Line to Dudley Square, which used to be the southern part of the Orange Line and should be light rail; unfortunately, the MBTA rejected it as cost-ineffective (see pp. 36-7) by applying a wrong cost-per-rider metric, as I will explain in a later post. But beyond that, the list of bus lines (p. 50 of the Bluebook) doesn’t contain anything nearly as juicy as New York’s bus lines: New York’s 50th busiest bus is roughly even with Boston’s top bus at 15,000 weekday riders, and its top routes have 50,000, making them obvious choices for subway extensions.

Since Boston does not have a capacity problem requiring more concrete pouring on its subway lines, nor high-productivity buses to railstitute, concrete pouring should focus on the other main reason to build rapid transit: to extend service to areas that do not have it. That’s the main reason to build the North-South Rail Link: it’s as much about direct service from suburbs north of Boston to downtown and maybe Back Bay as about rationalizing service and permitting through-running. As in Philadelphia and as should be the case in New York, through-running is primarily not about suburb-to-suburb service, but about access to job centers near the stations of the other half of the commuter network (in New York those would be at Newark, Brooklyn, and Jamaica; in Philadelphia, at Temple and 30th Street Station and in University City).

The list of concrete-pouring, lines-on-a-map extensions of the MBTA in or near Boston should therefore be limited to required Big Dig mitigations, and not much more. This is not just because they are legally mandated. They are also good transit by themselves – the North-South Rail Link for the aforementioned reasons, the Assembly Square stop on the Orange Line because of the TOD potential, the Red-Blue connection because of the East Boston-Cambridge service need, and the Green Line extensions because they provide much-needed transit service in Somerville that would otherwise need to be picked up by commuter rail, at the cost of good intercity service on the Lowell Line. Apart from these, the only major radial extension that should be pursued is the dismembering of the Needham Line outlined in my last post, in which the Orange Line would take over the portion within Boston and the Green Line would take over the portion in Needham.

What should be done instead of more expansive extension plans is very aggressive use of electronics to make regional rail more useful, recalling that its share of the suburbs-to-Boston market is about one third. This necessitates a lot of concrete pouring as well – on high platforms, on track repairs, on double-tracking some single-track segments, and on other things that do not show up easily on maps – but much less than adding tunnels.

The one difficult bit of concrete pouring that has to be done, in conjunction with the North-South Rail Link, is grade-separating the junctions that lead up to North and South Stations. Without the rail link, the South Station throat is such that, run right, it’s operationally at least two stations (one for lines serving Back Bay, one for the rest), and as many as four (Worcester, the lines feeding into Ruggles, Fairmount, and Old Colony and Greenbush); this allows for zero-conflict moves, higher capacity than the MBTA thinks, and a system in which delays on one line do not affect the others. With the rail link, those two to four systems need to feed into one track pair in a way that avoids opposite-direction flat junctions. The need for grade separations right in the station throats would add substantially to the cost of the rail link over a simple two-track tunnel; that’s why I’m not instantly dismissing it as something that at normal-world costs would take a relatively trivial $500 million.

Despite the rail link’s cost, the electronics are themselves substantial.  Signaling improvements are also required, to enable tighter overtakes. Moreover, full electrification should be non-negotiable – the MBTA’s stop spacing may not be as close as that of Metra or the LIRR or Metro-North or SEPTA, but it’s short enough that electrification would make a significant difference in performance. It also interacts interestingly with FRA waivers: on the one hand, without electrification, there are no good FRA-compliant trains – the Colorado Railcar DMUs have mediocre performance and are expensive and vendor-locked, and locomotive-hauled trains have terrible performance. With electrification, there exist decent FRA-compliant trains, but there also exist very good noncompliant trains. According to the Fairmount Line DMU document, current trains have a total acceleration-only penalty of 70 seconds to 60 mph, and Colorado Railcars shave that to 41 (see chart on p. 10); judging by timetable differences and dwell times, the best compliant EMUs lose about 20-25, and judging by YouTube videos FLIRTs lose 13.

The timetable examples I’ve put out – for the Providence Line in past posts, and for the Lowell Line in comments – are very ambitious, and require the signaling, electrification, and rolling stock to be perfect. The costs are not very high by US standards, but are nontrivial. Electrification costs a little more than a million dollars per kilometer (or about $2 million per mile), though it’s unclear whether this is based on route-km or track-km, as one citation I have is for a single-track line and another does not make it clear which one is under consideration. The cost is thus either about $750 million or about $1.5 billion, exclusive of rolling stock. But the benefit is commuter trains that can beat the freeways while also providing adequate regional service and connect to urban rail.

Improving the MBTA: Regional vs. Intercity Service

The MBTA commuter rail lines are laid in such a way that there’s an inherent tension between providing local service and providing longer-distance intercity service. It’s less apparent on the Providence Line because the intercity component, i.e. Boston-Providence, follows immediately just from serving the suburbs between Boston and Providence, but elsewhere there are greater problems. Good local service would have intense frequency in the inner portions of commuter lines; unfortunately, most lines only meet right next to the termini, reducing the opportunities to use interlining to create high-frequency inner segments.

Good local service also needs many infill stops, while good intercity service needs higher speeds. My proposals for the Providence Line essentially go with intercity service needs, justified by the facts that Providence is a major anchor, that high top speeds are possible on the line, and that the line should also host high-speed trains. Fortunately, the Providence Line has an opportunity for more intense local service using the Stoughton Line to add frequency; while this would end up overserving Canton Junction and Route 128, Readville and points north would get adequate peak service, and acceptable off-peak service. This is not as true on other lines, especially on the North Side, in which there’s a tradeoff between fast service to outlying cities and good service within Cambridge and Somerville.

Of course, the issues I’ve focused on in my previous post on the subject – electrification, high platforms, modern rolling stock – are useful for both. A fast-accelerating EMU could connect Boston with the various terminals at the same time as today’s express trains while making all stops as well as some extra infill stops. The problem comes from trying to fit trains into a clockface schedule. On a few lines, for example the Lowell Line, it’s actually easier to close very lightly used stations (Mishawum) or stations that are very close to other stations (Wedgemere).

Another issue is outbound extensions. With some, there’s so little traffic beyond the current terminus, or sometimes even beyond a point slightly closer than the current terminus, that the decision should be easy. This contrasts with the MBTA’s approach of proposing more and more outer extensions. With others, the intercity functions make extensions more reasonable, within certain bounds. I believe the following list of judgment calls would be reasonable:

1. Providence Line: no extension required – the line’s natural end is Providence. If Rhode Island wants to provide a low-frequency glorified parking shuttle from Wickford Junction and the airport to Providence, it’s its business, as long as it doesn’t muck up timetabling that’s based on Providence-Boston service.

2. Stoughton Line: an extension to Taunton would work, and possibly even to New Bedford. I’m iffier on Fall River, which has stronger commute ties to Providence; however, Providence-Fall River requires too much new infrastructure to be easy.

3. Franklin Line: either extend it to Milford (which may be easier to serve from the Worcester Line), or cut it back to Franklin. The Forge Park terminus is close to a lot of office park jobs, but the local road network is so sprawled out that it’s not worth the extra few minutes of travel time.

4. Fairmount Line: building infill stations is an excellent idea, though it should be coupled with increase in frequency and service level to make them more useful. One way to improve off-peak frequency is to route all Franklin Line trains along this line, and perhaps add supplementary trains that turn at Readville. The advantage of this is that the Fraknlin and Fairmount Lines used to be one railroad, with a grade-separated crossing over the Providence Line; in contrast, the junction at Readville is flat, making it more operationally cumbersome to have trains cross from one line to the other.

5. Needham Line: no extension necessary – the only possibilities would dismember the line in favor of much lower-density suburbs than Needham. Better would be to eliminate the line entirely and put Needham on a branch of the Green Line, and restore past plans to extend the Orange Line to West Roxbury. This would dismember the line too, but in favor of more service to dense areas rather than less. I don’t know what’s Needham’s commute tie to West Roxbury, but its commute tie to Newton and Brookline is fairly strong, 1,300 vs. 3,400 to Boston and another 3,400 in-town.

6. Worcester Line: Worcester is the natural terminus, so no extension should be entertained.

7. Greenbush Line: Greenbush is the natural terminus. The greatest urbanization is on the coast rather than along the railroad, and this limits the line’s usefulness.

8. Kingston/Plymouth Line: the natural terminus is downtown Plymouth, slightly farther out from the current Plymouth station, which should be renamed North Plymouth or just closed for lack of utility. In addition, Plymouth sends Boston 2,565 commuters, and Kingston only 797. Either the roles of Kingston and Plymouth should be switched – Plymouth would get served all day and Kingston would get only supplemental rush hour trains – or the Kingston branch should be closed, and replaced with a station on the main line.

9. Middleborough Line: for ordinary regional traffic, the line should be marginally cut back, to place the Middleborough station at the center of the town. In fact, there’s a dropoff in commute volume south of Brockton, and yet another south of Bridgewater; Middleborough is a fine terminus, but is not a proper anchor like Providence, Worcester, or especially Plymouth. On the other hand, there’s some potential for intercity traffic to Cape Cod, capturing some commuters as well as vacationers heading the other way.

10. Fitchburg Line: the MBTA’s proposed extension to Gardner looks weak to me, though not completely daft. That entire region of northern Worcester County has much stronger commute tie to Worcester than to Boston, in similar vein to the issue of Fall River’s connection to Providence. The commute tie to Framingham, as in the MBTA plan to have a branch leaving Framingham toward Leominster, is even weaker than that to Boston. It would be better to have a regional line connecting Gardner to Worcester, which would also have the advantage of taking a much more direct route than the freeway network; connecting Fitchburg and Leominster would require more work and compete with I-190 directly.

11. Lowell Line: here an outbound extension is natural and desirable, since Nashua and Manchester have a nontrivial commute tie to Boston and are significant cities in themselves, though as with Cape Cod this would be more of an intercity line. New Hampshire had a plan for such an extension, but it was killed by state Republicans early last year. This is unfortunate, since Nashua in particular has a less than great freeway connection to Boston, which a fast electric train could consistently beat.

12. Haverhill Line: Haverhill is a natural terminus. Although Rockingham County has a strong commute tie to Boston, the greatest part of it comes from very sprawled out towns near I-93, far from the line.

13. Newburyport/Rockport Line: the split at Salem allows natural interlining to give the towns with the strongest commute ties the most frequency. An additional branch to Marblehead would be prudent, providing even more frequency to Lynn, Chelsea, and additional infill stops in Revere. At the north end, Portsmouth looks like a fine intercity terminus, but in fact that part of Rockingham County is a marginal commute market to Boston, better than that feeding into Haverhill but much worse than the I-93 sprawl.

Not discussed above are station placement and infill stations. Station placement is relatively easy, since bad cases like Westborough and the aforementioned Middleborough and Kingston look obvious on a map. In addition, such office park stations with terrible ridership as Mishawum and River Works are already treated as such, so almost all trains skip them and their ridership is very low, making them clear candidates for closure.

Infill stations are harder. The problem is that on the North Side, the four lines split too early. This means that, while infill stations are possible, it’s hard to give them adequate frequency. Short-turning local trains could help somewhat, but is the most difficult on the two lines that serve Cambridge and Somerville, the Lowell and Fitchburg Lines. It’d be much easier to do this with Lynn (which already benefits from interlining and would benefit even more from a Marblehead branch) or Malden (which has the Orange Line).

That said, the Lowell Line might be able to support a local train to Winchester and an intercity train that makes zero or one intermediate stop between North Station and Winchester. The commute market is not great at this distance, though; Belmont has 3,100 Boston-bound commuters, and 290 inbound riders at its two commuter rail stations. A reroute of the Fitchburg Line along the Charles River Branch through Watertown might get more ridership; it would be slower, but it has zero intercity function, compared with strong potential at and east of Brandeis. To succeed, high frequency and short station spacing are required. For an example using the Charles River Branch, see here.

On the South Side, the Worcester Line begs for infill between Yawkey and Newtonville, but some of the people it would serve may already be riding the Green Line. The Green Line doesn’t perfectly parallel the line the way the Red Line parallels the Old Colony Line or the Orange Line parallels the Providence Line and the Haverhill Line, though, and there’s room for two or three stations serving Allston, Brighton, and Nonantum. On the other hand, some of these stations would compete with Watertown somewhat, and are less ideally placed in that the Worcester Line has an intercity function whereas the Fitchburg Line doesn’t.

Finally, another unmentioned issue is the effect of rapid transit extensions, especially of the Green Line. The extension plan to Somerville, which the state is obliged to build as one of many mitigations for the traffic induced by the Big Dig, is effectively a replacement for Lowell Line infill in Cambridge and Somerville; the line would only really need one infill stop to connect to the Green Line, and perhaps the Green Line would need to be extended to West Medford, if not to Winchester. That said, the interaction with rapid transit is more complex than this, and I will discuss it more in a future post.

Improving the MBTA

The MBTA has a problem. And I say this coming from New York, whose standards for good regional transit aren’t all that high, but now Metro-North looks like something to look up to from the MBTA. Ridership on the system is rising, but not very quickly; the MBTA moreover has no plans to modernize. Most of what I’m going to suggest will involve commuter rail, not because it’s the most important portion of Boston’s public transportation but because it’s the part I’m most familiar with and also the part that seems most direly in need of improvements. Put another way, I’m necessarily going to talk about the MBTA as perceived from Providence, rather than from within Boston.

The main difference with New York and past proposals for improvements, both subway extensions and regional rail, is size, and scope. In New York, practically everyone who works in Manhattan takes public transportation or walks. The transit mode share to Boston is lower and the car mode share is much higher. This seems especially true for people commuting from north of Boston.

The main prescriptions will not surprise people who have read my posts on best industry practices. In short, the MBTA commuter rail needs to do the following:

– Full electrification, starting from running EMUs rather than diesels under the catenary on the Providence Line, but also extending to all other lines.

– Level boarding along the entirety of all platforms, rather than just one car length, in order to shorten dwell times to no more than 30 seconds at outlying stations.

– Higher-quality rolling stock, with better-configured doors than the present cars as discussed in a DMU conversion study; all new EMUs available, both FRA-compliant and noncompliant, would be fine, though noncompliant trains with a waiver would have somewhat better performance and lower operating costs.

– Reasonable frequency all-day on a simple clockface schedule: ideally, all branches should have 4 trains per hour at the peak and 2 off-peak – the lowest-ridership lines tend to be the shortest-distance, for which frequency matters the most, whereas the highest-ridership lines (Providence, Worcester) are practically intercity, the higher demand balancing out a lesser need for frequency.

– A fare union with local buses and the subway, so that commuter train tickets are automatically valid without extra pay.

– Relocation of stations to walkable urban areas, away from park-and-rides that only serve to extend the suburbs into Boston rather than extending Boston into the suburbs.

– An end to outbound extensions, such as the ongoing project to extend the Providence Line to Wickford Junction, and instead a shift toward infill stations, especially in underserved Cambridge and Somerville.

In the longer term, a North-South Rail Link is unavoidable – North Station is too far from the CBD, some through-service from south of Boston toward Cambridge is advisable, and the rail link as proposed would give a direct connection to the Blue Line and thus to East Boston and the airport. Although the official cost estimate is $9 billion, for barely 2 kilometers of tunnel and associated connections, such an estimate would make the project more expensive km-per-km than any other I know of except perhaps East Side Access, and a more honest attempt at cost estimation yielded $3-4 billion, on a par with outsized American subway construction costs; at European costs, it would be less than a billion. Observe that electrification could reduce the cost by allowing steeper grades; the official proposal still uses heavy diesel locomotives. In either case, this is far more expensive than the points above; concrete costs much more than organization and electronics.

Let me now explain in more detail what’s happening in and around Boston – more precisely, what is wrong, and potentially what ridership level should be expected of good regional rail.

The main datasets I’ll be working with are the American Community Survey as of 2009, the town-to-town commuter flows as of the 2000 census, and the MBTA Blue Book, offering ridership numbers as of 2009 and going back to 1989. Bear in mind that most data from the 2009 ACS will be scrubbed from the net on January 20th, giving us only 2010 census-based numbers, which undercount immigrants and the poor and thus undercount cities; however, while the 2010 census gets magnitudes of change wrong, it’s very close in terms of absolute populations, absolute mode shares, etc. All numbers I cite here are from the 2009 ACS; you can verify that a source exists now, but not beginning a week from now.

The current background trends to observe are:

– Boston’s population is increasing, quickly. The 2000 estimate base, using a 2010 backdate that also depresses intercensal estimates to fit the 2010 undercount, was 692,745 for Suffolk County, which contains Boston and three small inner suburbs. By 2009, the county’s population was 753,580, a growth of 8.8%. Boston itself had 9.5% growth from the 2000 census, which is not directly comparable to the ACS and the estimate base but is extremely close in numbers. The metro area grew only about 4.5% over 2000 – a little less if one takes the full Combined Statistical Area, which includes slow-growing satellite metros like Providence.

– Transit ridership has grown in the last 10 and 20 years, but by much less than in New York. The Red Line’s grown 50% in the last 20 years, but the other T lines barely grew. The commuter rail grew quickly as lines were put into service in the 1990s, but had little growth in the 2000s, despite high gas prices.

– The Silver Line BRT is very underused, despite the promise and branding as rapid transit on tires. Even for airport service, where the Silver Line gets to the terminals, it gets less than half the ridership of the Blue Line (2,600 vs. 6,900), which only serves a station connected to the terminals by free shuttle buses. The Washington Street branches (SL4, SL5) are more frequented, but their combined ridership is only about the same as that of a single subway station, and are just bus-plus.

– Boston is the opposite of a bedroom community – it has 520,000 jobs vs. 278,000 employed residents, all as of 2000. This 1.87 ratio is much higher than that of New York (1.18), which contains most of its bedroom communities, and is more comparable to that of Manhattan (2.75). The same is true of Cambridge, with 114,000 jobs and 55,000 employed residents, for a ratio of 2.08.

– Unlike New York, both Boston and Cambridge draw substantial numbers of commuters from suburbs outside urban transit range – Boston draws about 200,000, and Cambridge draws about 55,000. Inbound commuter rail ridership on the MBTA is 70,000. Cambridge is a lost cause under current operating paradigms – it has no stations, and if it did they’d be too poorly integrated with the top two employers.

– Total transit vs. car mode share is 26-52 for people working in Cambridge and 37-50 for people working in Boston; the corresponding numbers are 56-29 in New York (including bedroom communities like Queens) and 73-14 in Manhattan (which is more comparable to Boston in terms of workplace geography).

– There are about equally many suburban commuters into Boston from the north as from the south. People driving to the edges of the Orange and Red Lines cannot make too big a difference (Alewife has 2,700 parking spots, and Malden and Oak Grove have just under 1,000 between them), so the difference seems to be that more people are commuting into South Station than into North Station. Observe that South Station is right next to the Boston CBD, whereas North Station is a little farther out.

– Boston has built too much highway infrastructure for a kernel of a transit-oriented edge city to exist along Route 128 as it does in Stamford. 10% of people who work in Stamford take transit to work. There aren’t numbers for all edge cities near Boston, but where they exist, they’re much lower, e.g. 2% in Burlington. Furthermore, since Route 128 exists and is continually upgraded, there’s not much hope of serving these centers by commuter rail from suburbs on the opposite side of Boston.

The upshot of all this is that there’s room to more than triple MBTA commuter rail ridership, while also maintaining healthy urban rail ridership coming from population growth in Boston itself. However, this requires very good service from the suburbs to the city, and the MBTA isn’t providing it. The problem is that the MBTA relies too much on cars: Middleborough and SouthWestborough are particularly egregious for their poorly located stations, chosen for drivers’ convenience rather than for that of transit users. Even worse, Plymouth, a city that’s older than Boston, gets few trains, while most trains serving the Plymouth Line instead stop at a park-and-ride nearby, at Route 3.

Although the focus of all suburban rail is service to the urban core, this can only be done by treating it as longer-range, lower-frequency rapid transit, rather than by treating it as shuttles from parking lots to the CBD (or almost the CBD, in North Station’s case). People won’t use the trains if they’re too infrequent past rush hour; it’s not 1960 anymore, and people do not always work 9-to-5.

For an example of what the MBTA is doing wrong, let’s look at commuter flows in Rhode Island. There are 4,700 people living in Rhode Island working in Boston. The biggest single source of Boston-bound commuters is Providence, with 1,100; Providence Station has 2,000 inbound weekday riders, so it also draws people from some nearby suburbs – but not too many people. Cranston and Warwick have 700 between them – and they’re getting an airport stop with a very small number of trains. Even Washington County, with 170 commuters, is getting a station. Those two stations cost $336 million between them. Meanwhile, Pawtucket, with 600 commuters plus another 800 in suburbs to its northwest and in Woonsocket, is not getting an infill station.

I hope to discuss concrete schedules, possible changes to station placement, and ways to keep operating costs under control in a future post. For now all I’ll note is that the MBTA needs to stop pushing for extensions far out into suburbia. It’s not going to get ridership out of 9 roundtrips per weekday with a 5-hour service gap, which is what the T. F. Green Airport station gets. It’s going to get it out of reliable, frequent all-day service.