How to Do Coordinated Public Transport Planning

I’d like to share an example of how to implement coordinated planning for public transportation, using an example of something I’ve been working on with TransitMatters in and around Boston. Right now we’re writing schedules and proposing concrete investments including electrification on each commuter line into Boston; the process is different for each line, but the first line we’ve launched the document for, the Worcester Line, is illustrative in itself. You can find the file here and the broader proposal here; the first link bundles two separate documents, of which the Worcester proposal is the second. I’ve harped a lot on using the Swiss model for better regional rail, and here is one example of how to get a city whose rail technology is stuck in the 1930s to have what Zurich has.

Slogans and principles

I’ve harped on a few Swiss and Swiss-adjacent slogans before:

Organization before electronics before concrete. Investments in more tracks, tunnels, and so on should come last as they are expensive, and beforehand agencies should improve signaling and electrify as it is much cheaper. Moreover, fixing organizational issues, for example writing good schedules and integrating planning between different agencies, should come before anything else, as it requires planners to do more work but is otherwise cost-free.

Takt and symmetry. If a train leaves your station going eastbound at 7:14 am and the schedule is every half hour, then a train leaves your station going eastbound at :14 and :44 all day, every day; this is also called a clockface schedule. If there’s additional service during rush hour, it should fit into the takt, e.g. more trains coming at :29 and :59 for 2 morning hours and 2 afternoon hours. By the same token, trains going westbound should serve your station at :16 and :46, since 60-14 = 46. This means the overtakes, meets on a single-track line, etc. all occur at consistent places.

The magic triangle of infrastructure, rolling stock, and timetable. The plan must account for all three sides of the triangle simultaneously, in order to optimize investment. For example, if additional tracks are required for timed overtakes, then the agency should know what trains it’s going to run and how frequent it’s going to run them in order to know where the overtakes are needed. With a takt, the overtakes will be at consistent location where the region can target investment.

Run trains as fast as necessary. Increases in speed should be designed around making timed connections and limiting train downtime. One refinement on a suburban line is that the stop spacing should depend on the schedule: if the one-way trip time is 52 minutes then a short turnaround makes an hour and additional stops are difficult to fit in, whereas if it is 46 minutes then the turnaround is longer and there is room for more stops.

The knot system: knots (or nodes, same word in German) occur at major stations at regular intervals – at a minimum an interval equal to half the systemwide takt frequency. If trains run half-hourly, then a station with service at :00 and :30 or with service at :15 and :45 will be served in both directions at the same time, so it’s a good place for bus and train connections. This works in both planning directions: if the schedule happens to place a knot at a station then buses should go there, and conversely if a city is a major node then the schedule should be written to a place a knot there.

What we propose for Worcester

The proposal as written calls for two service patterns, one express and one local. At rush hour, both run every 15 minutes. Off-peak, the express pattern drops to 30-minute frequency, but the local pattern stays at 15 minutes, as it serves Boston neighborhoods and Newton, close enough in that high off-peak ridership can be expected. With electrification and high platforms, the following schedule is feasible:

Station Km-point Local Express
Boston South 0 0:04 0:11
Back Bay 2.1 0:07 0:14
Lansdowne (Fenway Park) 4 0:09 0:16
West (Allston) 6.2 0:12
Boston Landing 7.5 0:14
Newton Corner 11.3 0:17 0:21
Newtonville 13.3 0:19
West Newton 15.2 0:21
Auburndale 16.9 0:23
Wellesley Farms 20.3 0:26
Wellesley Hills 21.8 0:28
Wellesley Square 23.8 0:30
Natick 28.5 0:34
West Natick 32.1 0:36
Framingham 34.3 0:39 0:32
Ashland 40.5 0:36
Southborough 44.2 0:40
Westborough Center 51.5 0:44
Grafton 58.7 0:49
Worcester 71.3 0:56

Express trains overtake locals at Wellesley Farms; there are plans for triple-tracking Wellesley (and farther west, but it’s not necessary). At Framingham, locals take 12 minutes to turn, which means there needs to be a non-revenue move around 0:41 westbound to a yard just west of Framingham to avoid getting in the way of express trains at :43 and :47 before getting back to Framingham at 0:49 to collect passengers; triple-tracking Framingham is also an option but is more expensive.

How it fits the principles

Let’s go over the Swiss principles one by one and see how this all fits.

Organization before electronics before concrete. As presented the plan includes elements of all three: organization is better-timed schedules and the potential use of the yard as a pocket track to avoid triple-tracking Framingham, electronics is electrification, concrete is the triple track. The electronics-concrete order is important – without the triple track but with electrification, EMUs can still do Boston-Worcester in around 57 minutes with the above stops, or 55 without infill at West Station and Newton Corner, either of which is faster than the fastest express trains today. The ultimate in concrete in the Boston area is the North-South Rail Link, which should come only after full electrification and related modernization steps, such as high platforms.

Takt and symmetry. The timetable is on a takt and symmetric, to ensure the overtake takes place at a manageable spot in Wellesley. It would be easier to change the offset slightly and overtake around West Newton, but there the tracks are in a constrained location where triple-tracking is prohibitively expensive. Note also that with the above timetable, the westbound overtake is at :11, :26, :41, :56, and the eastbound overtake is at :04, :19, :34, :49, which means it requires triple-tracking but not four-tracking.

The magic triangle of infrastructure, rolling stock, and timetable. The timetable is calibrated around the performance specs of the latest EMUs, like Coradias, Mireos, Talent 3s, and FLIRTs. The high acceleration capabilities of these trains let a local train leave Boston just 7 minutes ahead of the next express train, and still keep up through double-track narrows in Newton until Wellesley Farms, the sixth station skipped.

Run trains as fast as necessary. Without onward connections beyond Worcester, transfers between trains are not really a factor. Thus, what matters is tight turnaround times to keep trains moving and earning revenue rather than loitering at the terminal. The local train spends 35 out of 45 minutes running, and the express train 45 out of 60.

The knot system. Knots occur wherever trains stop around :00, :15, :30, and :45. The word around includes a few minutes of wiggle time, especially at a terminal, where transfers are unidirectional. Thus Worcester is a knot at :00, with a few minutes of rail-to-bus and bus-to-rail transfer. Framingham is a knot at :30, as long as the buses get there before :28 to transfer to eastbound express trains and depart after :32 to accommodate transfers from westbound express trains. On local trains, Newton Corner may be a knot, with a connecting bus shuttle to Watertown.

Bus nodes

Framingham and Worcester already exist as bus nodes, and in both cities, the main city bus hub is already the train station. The next step is to integrate the schedules. The rule is generally that bus timetabling should follow rail timetabling, because trains require more infrastructure whereas buses can be moved more easily; there are exceptions, but not many.

The principles for bus design on the Zurich model aren’t as catchy as for rail design, but they are still useful and generally worth learning:

One ticket for all. Fares must be totally integrated. If a train makes two stops in the same zone (for example, Framingham and West Natick), it should charge the same as a bus. A train ticket should be valid within the entire zone traversed, which includes bus transfers. The same fare media should be used on all modes – and they should be paper tickets with no surveillance, not Boston’s ongoing smartcard disaster (“AFC 2.0”). Fare integration requires a mechanism for sharing revenue across agencies, but this is organization, and is doable under the aegis of the Massachusetts state government, with revenue allocated to agencies based on periodic counts to ascertain ridership (Berlin’s are every 3 years).

Timed transfers. Suburban buses should come every half hour, on a takt of course, with timed transfers to the trains at the relevant knot. Worcester’s bus agency, WRTA, does not do this at all – bus #1 runs on an hourly takt, but other routes may run every 50 minutes or every 75. Framingham’s, MWRTA, has 65-minute headways, and a route that runs to a Green Line station rather than much closer to a commuter rail station.

High vehicle utilization. If the bus takes around 25 minutes to reach its outlying destination, then two vehicles serve one route, and if it takes 40 minutes, then three vehicles do. Buses should run as fast as necessary as well, deleting meanders, installing queue jump lanes, and shortening the route in order to squeeze inside a timetable with short turnaround times.

Connections between different train stations. A bus can connect two different train stations, either on the same line or on different lines. It should be timed at both ends, though it if runs parallel to the train, then it’s fine to time only right-way connections (e.g. eastbound bus to eastbound train, westbound bus to westbound train), which do not require knots.


  1. Rover030

    If your New England high and low speed rail proposal was built, I assume you would want to extend the Worcester service to the Worcester HSR station to provide timed interchanges to/from Springfield (and further), Providence and Fitchburg. But then you need to arrive at Worcester HSR 0:57 at the latest, meaning you need to save 2/3 minutes somewhere to make the 3 minutes trip from Worcester Union station. Is there scope to save those 2/3 minutes in between Wellesley Farms and Worcester? If not, would you propose to depart Boston 3 minutes earlier and build an overtaking track in a more expensive location?

    I’m asking this because I’m curious how far you should go with planning ahead for these kinds of things. In this example operations and infrastructure are very well integrated, but that does seem to come at the cost of future flexibility.

    • Alon Levy

      High-speed rail (almost certainly) has to come equipped with bypass tracks in Newton, so the express regional trains can use these bypass tracks too and overtake locals in West Newton rather than Wellesley.

  2. Mike@London

    This is all common sense. The real question is why it is politically impossible to achieve. The U.K. is taking baby steps to proper integrated public transport after many false starts from the 1930’s onwards but the USA seems as though it will never get there. The lobbying is there and there are many intelligent and forceful advocates but the political mindset is just sclerotic.

  3. Pingback: Today’s Headlines – Streetsblog California
  4. SB

    Surprised that there were no references to upzoning and removing most of the parking lots next to the stations.

    • Alon Levy

      Look at the location of Westborough! But generally, TOD is not really part of the Zurich model regrettably, there are other models for that (e.g. Tokyo, Paris, Stockholm, Copenhagen).

      • Paul

        Westborough: What on earth? The historic depot still exists, why isn’t MBTA using that?

        Western Switzerland has decent TOD. The Lausanne-Geneva line serves the historic town centers pretty well, and there is some infill development happening.

        • F-Line to Dudley

          They whiffed…simple as that. Westborough’s the biggest perennial underperformer of the 4 intermediate stops that opened from 2000-02. Ashland, also moved away from its still-standing H. H. Richardson dead-center downtown depot to a new parking sink, is unsurprisingly the second-biggest laggard underperforming estimates.. Southborough, at its traditional station location, strongly overperformed its initial estimates and looks to scale bigger with each service increase. Grafton made a risky but someone reason-backed tradeoff from its former downtown site to the newer Tufts Univ. campus 1.5 miles east and has treaded water *near* expectations. Millbury succumbed to local NIMBY fever and outright rejected siting a stop at US 20 in downtown stone’s throw from the Mass Pike exit, which they now regret.

          Unfortunately 20 years feels more like last century than it literally is in light of the scruples of some of these decisions. Absolutely nothing TOD-wise is going on with Westborough, which sites in an industrial park that eschews any substantial enrichment of land use for mixed-use. Basically the only pivot that’s going to save it is to eat their losses and abandon it for a move back 2 miles east to the downtown depot location centered on the MA 30/MA 135 intersection. Trying to reboot TOD at the failed current location amounts to nothing more than throwing good money after bad. The western infill stations are all pre-designed such that by moving one set of platforms back beside the existing up-and-over pedestrian ramps the stations can be re-done in place as triple-trackers with full-high platforms and high-and-wide freights passing in the center. Which will be a merciful cost and time saver when they’re done backfilling all the full-highs to Framingham on the no-longer-freight-clearance inner half and can turn attention to closeout mode on the outer stops. In Westborough’s specific case, however, that’s the time to pull the plug and relocate rather than renovate.

          Ashland they are stuck with the current location because service increases + intercity service increases the priority on the already badly-needed and locally desired downtown grade separation project that would elevate the tracks over Main & Cherry Streets on rail overpasses. The inclines for the grade separation would prevent re-use of the old depot location, so their best bet is going to be to try to make lemonade from lemons at the current detached parking sink by striving for some more intrepid TOD. At least there’s nearby housing sprouting up, although it’s being infilled stupidly as gated-community cul de sacs without much nod to connectivity with surroundings. Par for the course with what still passes for suburban planning. At least things *can* get better here with the canvas they have to work with, unlike Westborough. I wouldn’t wager any money on local-yokel planners getting with the program in enough time, but it’s not an outright squander.

          Grafton’s not exactly firing on all cylinders, but at least the Tufts Campus affords a TOD blueprint they can work with. The Millbury infill would help them a lot as the Millbury location spaces kinda close to where the old Grafton station used to be and is easy couple miles reach for spanning any catchments the Tufts siting leaves in the cold. Revisiting Millbury after service increases are enacted and most existing station plant is brought up-to-spec pretty much fixes the glitch on anything that might’ve been left wanting with the Grafton station siting choice while plugging a station spacing gap.

          • adirondacker12800

            But the new mall at the old station in Westborough has 44 condos!! !! Surfing through on Google’s Streetview the architects did a half way decent job of putting Main Street siding on the more or less standard steel frames. I can’t tell if the Italian restaurant is in the former freighthouse or they did a nice job of doing faux freighthouse. It looks too new, from Streetview. I lean towards faux.
            Millbury would be a great place for the ginormous park-n-ride. Nobody is going to care if it’s in the interchange next to the sewer plant and railroad yard. Which would dampen the urge for residential. Or retail. Park-n-rides out in the middle of nowhere are good thing. People who are going to drive aren’t in downtown and downtown can have something other than parking lot.

  5. YDD

    Specific to the Worcester line, but what’s the impact of the following on the ability to make these schedules:
    – The CSX yard immediately next to Framingham station
    – The two level crossings east of Framingham, one of which is practically at the end of the platform and is on a fairly major north-south road link
    – The Amtrak Lakeshore Limited. Would it just replace one of the Express trains in each direction every day? If so, could Amtrak be persuaded to integrate the schedule (so it literally runs as the ‘express’) and ticketing?

    I’d love to have that 15 minute frequency (even though it means giving up my express).

    • Alon Levy

      – The CSX yard is helpful! One track there should be used to turn local trains at Framingham to avoid having to add tracks to the station.

      – The level crossings will presumably be closed when a train passes, like today. A grade-separation + four-tracking project may be prudent with further increases in service, say to 10-minute frequency for each of the two patterns at rush hour, or with the North-South Rail Link introducing new and exciting timetable constraints (now the Worcester Line gets to depend on whichever North Side line it pairs with).

      – The current scheduled Boston-Framingham trip times on the LSL is 35 minutes, same as what a local EMU would do; level boarding and higher superelevation can cut this somewhat, but most likely the train would trail an off-peak local that isn’t being overtaken, then continue to Worcester at comparable speed to an EMU making those intermediate Framingham-Worcester stops. This is how much diesel locomotive speed and Amtrak padding suck. What’s more interesting is East-West – if there’s strong enough infrastructure east of Worcester, then electrifying to Springfield and running East-West as a Worcester train that continues farther west would be useful.

      • adirondacker12800

        When there is high speed rail service all the way out to Minneapolis there isn’t going to be a LateForSure Limited. The Empire Builder may still exist as a social service to North Dakotans and Montanans who live in places so remote not even Greyhound goes there but it’s not going to go east of Minneapolis. There won’t be a Sunset Limited west of Tuscon, there won’t be a California Zephyr west of Sacramento or a Coast Starlight south of there. There won’t be a Maple Leaf or an Adirondack or a Pennsylvanian and the Capitol Limited if it survives is going to terminate in Indianapolis. The Silver Services may still exist but it will a high speed train through Atlanta.
        You aren’t going to get permission to build a four track grade crossing in the 21st century. If you are going to have ten minute frequencies on the Worcester expresses and ten minute frequencies on Framingham locals and have Springfield served by twice an hour intra New England kodama-ish things how many high speed trains are going to Albany in an hour-ish? Juggle four thoughts at a time.

        • Alon Levy

          First of all, if there’s high-speed rail, it’s going on a near-Turnpike alignment from Auburndale west, so it doesn’t touch Framingham.

          Second, I’m specifically not talking about a four-track grade crossing – when it’s time to four-track Framingham, it’s also time to elevate the tracks.

          • adirondacker12800

            If the intercity trains aren’t going to be anywhere near Framingham the MBTA can call SEPTA and ask for advice about how to run 12 trains an hour on two tracks. SEPTA. Not NJTransit or MetroNorth. SEPTA.
            Where is ShinWorcester going to be and since apparently the intercity tracks are going to crossing over the MBTA at I-495 why won’t ShinFramingcester or perhaps Worminham be there? Where are the trains from Providence? If they are where I think they will be the only NIMBY that might object is a sewer plant and the Providence-Worcester trains could share the north-south platform with the Boston-Worcester trains and downtown Worcester has higher frequency to ShinWorcester. Juggle four thoughts at a time.
            And if there are 12 MBTA trains an hour stopping at Framingham why do you care if the bus timed to any of them?

          • SB

            “MBTA can call SEPTA and ask for advice” AKA the blind leading the blind
            12 MBTA trains an hour won’t happen right away. It takes time to build NSRL and for demand to rise.
            Until then time transfers makes most sense.

      • YDD

        My query about CSX was more practical. Yes it could be used to allow expresses to pass turning locals at Framingham, but I’m guessing that CSX will also want to keep using their yard. I know that freight trains still use the line which diverges just east of Framingham station, and goes south towards Dover. Presumably those freight movements would impact the schedules?

        As for dealing with the level crossing, if you elevate the line through Framingham to eliminate it, wouldn’t that have a functionally similar effect to putting an elevated interstate through there? There are pedestrian friendly shopping streets on both sides of the line, and that would cut them off from one another. Putting the railway line in a cutting might be possible, but I’m not sure what the water table is like in that area.

        • Alon Levy

          There aren’t a lot of freight movements, so they can fit in between the passenger trains over a short distance.

          Elevating the tracks in Framingham wouldn’t be like a freeway, for a number of reasons:

          1. Four tracks is a much narrower structure than a six- or eight-lane freeway.
          2. Elevated trains are less noisy than cars; els in New York and Chicago are mostly steel boomboxes, but the few concrete els (like on Queens Boulevard on the 7) are quieter than the cars at street level, as are the concrete els in Paris and Berlin.
          3. Crossing the tracks is easier when they’re grade-separated, because there aren’t gates that stay down for a large fraction of rush hour.
          4. The station is a desirable spot for retail, so retail will keep clustering there, whereas a freeway is noisy and repels retail, which creates a negative spiral of decline.

          EDIT: I forgot to mention – this isn’t hypothetical, the crossings of the Ringbahn in Berlin are pretty desirable spots even when the Ring is elevated, as it is in my neighborhood. The trains are quiet enough I don’t hear them coming so I don’t know if I’m about to miss my train and need to run (I approach the station from the opposite direction as the trains I take from it). There’s a shopping mall adjacent to the station, as is the case at many other stations. The main retail drag crossing the Ring has a lot of retail on both sides, with only a brief interruption for the crossing.

          • Richard Mlynarik

            Four-track two-island-platform elevated station can be really really nice, as they’re two slimmer rail bridges separated by spreading space for the platforms (outside the station itself) and — at least outside the USA, where worst possible claustrophobic NY-Penn-is-an-ideal-not-an-antithesis station architecture is mandatory — the space under the platforms can be made comparaitively airy, with good ceiling heights and numerous permeations for plentiful and wide banks of stairs+escalators and elevator shafts. (Here’s Bijlmer Arena station near Amsterdam.)

            And even where it’s one big solid concrete deck (here Skøyen near Oslo), a four-track two-platform overbridge is way less oppressive and intrusive than an interstate freeway.

            But no, we can’t have nice things in the USA.

            PS Elevating Framingham is tricky because of the junction. Elevating passenger and leaving freight at ground-level — the awful route that is de rigeur in the US, becuase where “public transit” is involved, there’s no planning, no respect for human beings, and no concept of station accessibility — leads to awful results. (It’s PBQD/WSP’s plan for all the California HSR stations, of course. Of course.)

          • adirondacker12800

            Zillow says the median sales price in Framingham is $436,457. I suspect the grade crossings are still there because those people’s great grandmothers took off their white gloves to write letters protesting grade separation projects going back to 1898.
            Your chances of putting an El through downtown are slim and none. Especially since it would ruin the vistas to the historic landmark train station.

          • Richard Mlynarik

            Thank you, once again. Your valuable input has been noted. As it happens, I only have familiarity with grandmothers in Framingham, but I — and indeed all of us — must all defer to your inability to shut up on absolutely any topic, especially given my professed lack of great-grandmother Metrowest Boston expertise.

          • Nilo

            I’d add that the Fullerton station which uses concrete pillars is perfectly pleasant to be around, and has quite good pedestrian circulation via escalator and elevator, and wide platforms. I seem to remember the platforms at Belmont being plenty wide as well, but I haven’t spent nearly the amount of time there.

            Where else besides Chicago are there even four track two-platform elevated stations in the US?

          • adirondacker12800

            I spend 145 seconds reading stuff after asking the Google to look up “Framingham train station”. Around a dozen proposals over 122 years, there seems to be some resistance to the idea. It’s going to very very difficult to build an elevated hovering over a landmarked railroad station. Which I’m sure is beloved by one and all. Including from across the tracks. Where the view would be blocked by the El. I’m seeing trench to connect the parts that already are grade separated because they’ve been finding reasons to not grade separate for 122 years.

            There are four track stations scattered all over Pennsylvania, New Jersey and New York. There’s a bunch six track ones in Illinois and a few in Pennsylvania, New Jersey and New York. There is fun spot in Philadelphia where the six tracks of elevated trains pass over four tracks of train in a trench. All of them electrified, all of them with passenger trains and without much effort it could be the same train. And it might even be at times but not in revenue service. Because after decades of having advocates insist people in Lansdale had the urge to shop in Darby, they don’t and they trimmed back the through running to five stations in the center of Philadelphia.

            There’s going to be more rail freight. More containers moving on rail means there are less containers on the MassPike. and less containers on highways in Connecticut and less containers in New York and less containers in New Jersey, Pennsylvania, at least all the way out to Virginia and Ohio.
            There was much screeching when this happened.


            It goes with the intermodal CSX built in ……………Westborough……….. whether those two things are related would better be answered by F-Line-to-Dudley. They happened around the same time. And whether the two rather busy yards in Framingham play a part. There is going to be more rail freight. Or the Long Island Expressway, Connecticut Turnpike and the Massachusetts Turnpike have to be double decked. And it’s much more cost effective to ship the drywall, cement, gravel and sand in by rail and the garbage and sewage sludge out. Beer, toilet paper and frozen food in…. New automobiles. People who buy 400k houses buy new cars. That they drive to the mall next to where the train station used to be in Westborough. But not from Framingham because I doubt there is much difference between the Starbucks in Framingham and the Starbucks in Westborough.

          • Alon Levy

            There’s actually less rail freight in the US now that coal is in decline; the Class Is are hurting quite a lot. And on the Worcester Line specifically, the MBTA has been clawing back at CSX, buying out rights and telling CSX to redirect freight via secondary lines.

            I don’t know to what extent there’s NIMBYism in Framingham; I’ve asked around TransitMatters, will report if I get an on-the-record answer. My best guess from what I’ve seen is that the reason the MBTA hasn’t grade-separated the junction yet is cost rather than local opposition.

          • YDD

            I think that the original Framingham station could readily be used as the entrance to an elevated station. The tracks might have to be bumped to the north slightly (with the platforms possibly moved a bit to the west) to avoid overhang. However, with the park and ride, there’s plenty of space for that.

            The freight lines could be a killer, though. I’d forgotten that there’s one to the west of the station too, which appears to be for access to some sidings slightly further west (but too close for the commuter rail to descend again). Those freight lines would have to remain at their current elevation, which implies a rather high level for the commuter line. The nearby pond almost certainly makes a cutting impractical.

            And while not as big as an interstate, an elevated track is going to have a significant impact on what is a fairly flat area.

          • Alon Levy

            I bring up the width because it makes it easier to cross under – even Karl Marx Strasse in Neukölln, which is dead space under the tracks, has a short enough crossing that people do it and there’s healthy retail on both sides thanks to the access brought by the station. In contrast, a freeway is longer dead space because it’s wider, and it’s also noisier and thus less desirable, and there’s no access reason for retail to locate right there so it’ll locate away and the city will be cleaved in two.

          • Seungho Bak

            Why not elevate the roads instead of the railroad tracks?
            According to satellite images other rail separations in Framingham are elevated road crossings.
            So there are precedents to counter any NIMBY/Not-the-way-we-do-around-here arguments.

          • adirondacker12800

            When was the last time anyone in the Northeast outside of Pennsylvania burned coal? The only coal that gets shipped in intermodal containers is the stuff the home improvement center is selling by the bag. They didn’t rebuild Mechanicville and a new intermodal in Westborough circa 2012 in anticipation of the business dying off.

          • YDD

            Elevating the roads… I don’t think that is especially practical either. Rte 27 is a main street on both sides of the tracks. There’s a variety of low rise buildings with shops on both sides north and south. A little further to the north is the centre of Framingham. Getting the road back down to street level there would mean a noticeable climb.

            If it were just 27, then perhaps a combination of lowering the railway and raising the road (or vice-versa) could work. But the combination of the freight tracks and 135 (the east-west road) make that rather hard. While the buildings obviously aren’t immutable, demolishing a large section of the urban centre of Framingham in order to promote urbanism feels a bit off.

            These problems would get easier if CSX didn’t need those two tracks heading south from either side of the current station.

          • Alon Levy

            Yeah, generally you want to elevate the railway, not the street, for the reason you mention, and also because the railway needs more clearance above than the street.

          • Nilo

            I think you have to plan for elevated streets before you develop. Five Points in Atlanta is actually elevated streets above rail, so there’s one level of subway, a level of at grade rapid transit and mainline rail at ground level, and the streets above that. Also sections of downtown Chicago were built with elevated streets.

          • F-Line to Dudley

            I don’t know where this is going to get lost in WordPress’s crappy nesting, but answering a few queries. . .

            To Ad12800: Future freight outlook is like this. . .

            — Framingham yards are still ‘the’ classification yard for Eastern MA, as Worcester is intermodal-only and Westborough is a ‘transflo’ (i.e. liquids, aggregates) loading facility. For example, stuff that runs BACK out west to Worcester to interchange with Grafton & Upton RR, Providence & Worcester, and Pan Am still has to get manifested in Framingham because the other facilities are single-task.

            — CSX isn’t using all 3 Framingham yards. The south one, “CP Yard”, used to be for autoracks until CSX consolidated 3 inefficiently separate old auto loading facilities into a unified new one in Brookfield. Right now “CP” is unstaffed and only used for overflow storage, while west (Nevins) and North Yard handle main duties. City & state really covet the North Yard property for downtown redev, with Framingham State Univ. having interest in expanding campus to dead-center downtown. CSX has long stated that it’s willing to entertain offers if all North Yard functions could be re-accommodated at CP Yard with some state fun bux for upgrading. So far no one has been willing to meet CSX’s steep asking price, but this is still a thing that probably will happen. If/when it does North Yard would get busted down to 2 mainline tracks for the Fitchburg Secondary and a runaround siding for CSX changing directions when it goes through the wye. A decades-derelict CP Yard track connection from the Framingham Secondary (crossing Irving St. & Loring Dr. south of downtown) would get reanimated, and in exchange for *slightly* higher freight usage of the grade crossings for Nevins-to-CP interyard swaps they’d be able to stage all Walpole/Boston freights straight out of CP and gerrymander use of the crossings to strictly off-peak. In the process the existing, seldom-used CP Yard lead crossings of MA 135, Clafin St., and MA 126, just west of the downtown clusterfuck could get abandoned to at least permanently clear all roads to the south and west from the emergency services liability of having active crossings.

            — CSX is in territory-shedding mode. It’s handing off Franklin-Milford to Grafton & Upton RR (which it interchanges with in North Grafton), ending all freight service on the lower Franklin Line between Walpole and Franklin Stations (with just a smidge of G&U overlap at Forge Park). It is widely believed to be interested in trading out its Attleboro-Middleboro-Braintree overnight to Mass Coastal RR, once MCRR resolves its shaky ownership situation. This would trim what’s currently 2 daily round trips on the Mansfield-South Attleboro portion of NEC to just 1, and trim what’s 2 daily trips through Foxboro to just 1. Braintree Yard pickups will get backfilled on an overnighter out of Readville Yard via the Fairmount Line after commuter traffic has ended, and they are seeking to buy out 1 zit of an intermittent customer in Stoughton so they can trim their lunchtime upper NEC jaunt to just 1 anchor customer behind Route 128 station. This won’t affect Framingham’s volumes at all…in fact, the shortlines will probably increase total carloads. But it’s a large labor reduction as the shortlines will do sorting for them, and allows them to pack everything originating from Framingham onto basically 4 daily round-trips (1 north, 1 west, 2 south) that will get longer but never more frequent, plus the yard restock jobs from Albany/Springfield that stay static at 2-3 per day.

            — Because less switching is going to need to be done (and also because CP Yard is way larger than North Yard if that land swap happens), the T should have little trouble getting an easement for a couple trains’ worth of layover storage in Nevins Yard. Which is where the layover used to be before commuter service was reinstated to Worcester. Wouldn’t be spacious by any means, but enough to do the job.

            — There are NO freight clearance increases needed inbound of Westborough. CSX has already signed off on a chop-down of its former autorack clearances to Framingham to plain old Plate F, and there are no IM sites in Greater Boston large enough to handle stacked cars. You have the Plate F route Framingham-Walpole-Readville and Framingham-Walpole-Attleboro (already under wire), and that’s it. If Port of Boston ever contributes any cube loads (a big if), lack of onsite sorting space would force Trailer-on-Flatcars to have to get brought to Framingham by default for sorting before it’s even possible to drop the second stack on-top. There’s no future tasks they would ever be handling outside of current OK-for-TOFC clearances to begin with. Eventually Pan Am will get cleared to stacks from Ayer to Portland so Portland can be an auxiliary distribution node, but the rest of Southern New England is well-covered by the current “all roads lead to Worcester County” freight orientation.

          • adirondacker12800

            F-Line, it’s real easy to find out how much the electric companies are burning in Massachusetts.
            It’s “none”. It’s not 1960 anymore, it’s not even 2005 anymore. That links says Mass. is one of three New England states that no longer burns coal. I don’t do this for living, I’m not going to go through the other New England states to discover which ones it is. There isn’t a lot of coal being slung around in New England any more. There hasn’t been for a quite some time.
            Thanks for the details. They are realigning things from when A&P sent a horse and wagon down to the team tracks to pick things up, next to the coal dealer and the lumber yard to moving whole cars of stuff?
            A mailbox at the end of the driveway has a few problems in places where it snows a lot. They get solved by having my mail delivered at the Post Office, in a P.O. Box. Which has all sorts of fun associated with it but it’s less effort. It’s annoying to open the box and find that I don’t have anything. So the Post Office, yes that Post Office, sends me photographs that the sorting machine took as the first class passes through in blur…. it’s not 1890 anymore or even 1960… it whole carloads of stuff… Portland is an important origin. All those frozen fries from the plants in rural Maine and New Brunswick get shipped by rail. To supermarkets that have frozen distribution center someplace different than the refrigerated distribution which is different from the dry goods…

      • F-Line to Dudley

        It’s actually silly-easy to add a third track to Framingham station as the current westbound platform can be raised/widened to 12 ft. island with third track trenched across the wye ‘infield’ on the backside. CSX no longer needs any high-and-wide passage through the station, as the abolishment (as of July 2019) of the very last Allston-bound freight job consolidates all remaining Greater Boston and SE Mass. freights to a lumped-together couple daily Framingham-Walpole yard feeder jobs on the Framingham Secondary that split at Walpole to their separate ways. Most get dispatched in/out of North Yard via the east wye except for as-needed lazy shortcutting through the station, and with volumes now low enough for the wye to suffice CSX has allowed the T to sunset the high/wide exemption through the station-proper. You might have a design pickle trying to shiv a full-high on that east wye when it comes time to build branchline service to Northborough on the Fitchburg Secondary because *that* particular track still carries high-and-wide loads on a curve…but the mainline platforms are now all-clear. It used to be thought that third passing track on the wye ‘infield’ was thought to be necessary for CSX, but now that it isn’t any platform rebuild for full-highs might as well double the westbound platform to island-width and throw down a cheapie extra pax track for the hell of it.

        Once City of Framingham stops dragging its feet on installing maximal quad-gate protection at the MA 126 and Bishop St. grade crossings and allows the T to install security fencing on the midblock to close off everyone’s favorite illegal pedestrian shortcut/a.k.a. Game of Chicken, the 2 Framingham crossings cease to be any concern for train traffic. Accel/decel out of a station stop on 100% of schedules mean any speed restrictions on the approach are a non-factor, and speed is never a factor with collision risk. And, with 28 fewer scheduled daily freight trains using the crossings than 12 years ago, total gates-down time is far LESS today than it was before and capable of increasing generously before overtopping what it used to be. The crossings are mainly a shitshow for car/ped traffic. The City has been navel-gazing at southerly street grid reconfigs to better load-spread traffic for decades, but cannot come to any actionable consensus. That’s on them. A MassHighway plan to sink MA 135 under MA 126 at the worst of the 2 intersections would have hugely solved the queue backups around the tracks and changed most turning patterns into frontage-ramp turns at a fraction of the volumes. Crossings would’ve remained, but they would’ve been effectively neutralized as a problem. Unfortunately, it costs a kajillion dollars and the state is in no rush to identify funding sources, so it languishes in the MassDOT project database.

        Outright grade separation is virtually impossible there. The triple-junction and need to do 1 more costly crossing elimination to plug the Framingham Secondary into an elevated structure is mind-boggling expensive, and the City already vomited all over the notion of erecting an elevated wall there. Trenching the tracks is no-go because of the adjacent Sudbury Aqueduct. But, again, it’s a least concern for every single train schedule pax or freight that’s making a Framingham stop so the solves are mostly about City vs. highway dept. and making up their damn minds on the set of ‘vision thing’ blanket solutions for the street grid. The Ashland crossings are much more consequential (not to mention way easier) eliminations for train traffic because of higher train speeds and all expresses, Amtraks, and freights blowing through there nonstop and needing to be inoculated from creeping speed restriction.

        RE: electrification west of Worcester…don’t put it on a clock. There are a total of 35 overhead bridges between Worcester and Springfield which would need to be cleared for 2-1/2 feet of extra vertical learance for double-stack freight under 25 kV wire in territory where intermodal stacks run 24/7 at 100 cars a pop. And untold more overhead structures from Springfield to Schodack, NY where freights and Albany pax diverge. Most of those bridges were maxed out with all available cheap trackbed-undercutting tricks when the double-stack clearance project was done a few years ago, so the cost for modifying bridges and changing road grades on the approaches is going to be the most extreme of any electrified line north of NYC. Mercifully, the vacated Plate F clearances Framingham-Allston (sunset 2018) and vacated autorack clearances Westborough-Framingham (sunset 2010) make clearances inside MBTA territory A-OK as-is, with only 6 total overhead structures Worcester-Westborough in double-stack/MBTA overlap (at least 3-4 of them verifiably tall enough as-is). No problems whatsoever electrifying to Worcester for the T…but in looking at the priority pile you won’t be touching those extreme Boston & Albany cross-state clearance costs until well after you’ve settled every single T line, the Springfield Line, and probably several others. Not a “never” by any means, as this has much more to do with what else is clogging up the backlog of needed electrification work…but you are definitely going through at least one entire procurement lifecycle of Amtrak diesels before working the electrification bucket list down to point where Worcester-Springfield (much less beyond to Albany) has got next-ups. It shouldn’t matter excessively to schedules in the interim as Palmer is the only planned infill between Worcester and Springfield and all schedules under any power plant are subject to where they can ID meaningful curve-easing exploits in the Worcester Hills. But there will realistically have to be *some* diesel interregnum of significant duration because of where the very high WOR-SPG-ALB costs slot it in the overall electrification bang-for-buck priority pile.

    • SB

      Why not elevate the roads instead of the railroad tracks?
      According to satellite images other rail separations in Framingham are elevated road crossings.
      So precedents exist.

      • F-Line to Dudley

        Problem is the tracks are 20 ft. to the side of the curb on MA 135/Waverley St., and Howard St. on the parallel block is only 75 ft. away from the tracks. It’s not nearly enough room to incline up/down and be able to interface with the huge thoroughfare Waverley or the block-spanning connector for a lot of development in Howard. Worse…both Bishop St. and MA 126/Concord fork into diverging thoroughfares a mere 150 ft. south of their intersections with Waverley. So it would royally screw up traffic patterns if the forks lost their access to Waverley and Howard because a bridge had to fly over them. And even then you still have a tight fit getting down the bridge incline in time for the splits.

        It’s not practical in the slightest. Boston & Albany did mass grade crossing eliminations in the first decade of the 20th century, and there’s very good reason why this one has had to malinger so long. Now, if the MassHighway plan to sink 135/Waverley in a duck-under underpass of 126/Concord ever comes to fruition you would majorly take the sting out of it all. Concord would no longer be at risk of traffic light queues backing up onto the tracks because the turning frontage ramps onto the sunken Waverley would only have a fraction of the volumes and there’d be more queue management space before the fork in the road Street grid changes under consideration to the south would then de-emphasize Bishop St. and take a lot of volumes off that crossing. So while no crossings would outright go away, the traffic problems around them are perma-solved.

        Unfortunately, it’s hella expensive…the conversion to duck-under + frontages would be hella ugly for dead-center downtown (if ultimately way more overall-beneficial than its looks…and Framingham is notoriously difficult to deal with on community input so it’s going to be an excruciating exercise in teeth-pulling to wrestle them to a consensus on what’s good for them. In the meantime MassDOT might as well just settle up the Ashland crossing pair, because that town is in agreement that they’re A-OK with a rail overpass over Main St. and half- raised rail / half- dipped road at Cherry St. That one’s Git ‘R Dun straightforward…and more consequential to train traffic than Framingham for whacking a minor speed restriction on any express schedules.

  6. Max Wyss

    Is there a reason why you go for a flying overtake instead of a timed overtake in Framingham?

    This would create an “inner express, outer regional” train providing connections between all stations (some of them with changing trains on the same platform). Most prominent example for the S-Bahn Zürich would be the 15/45 node in Wetzikon.

    • Alon Levy

      Yeah, there’s room for three-tracking in Wellesley and Natick. In Framingham there’s room for four tracks, but it’s right next to a grade crossing, and the assumption is that four-tracking it would have to also involve grade-separating the station.

      The other issue is, a Framingham overtake wouldn’t be at a knot, because that would require an additional trainset relative to optimizing the turnaround time. So instead there’s an 8-minute wait time for connections from the local stations west toward Worcester.

  7. Peter Furth

    Great analysis. A couple of questions to enable others to do the same kind of analysis. You’re relying on good acceleration from EMUs. What acceleration (and deceleration) rates can be assumed? And is cruise speed the same on every segment? If not, how does one determine maximum speed by segment? And what are you assuming for dwell time?

    • Alon Levy

      1.25 m/s^2 initial acceleration, short-term power to weight ratio of 18 kW/t with passengers. The top speed ranges between 100 and 160 km/h depending on stop spacing and how curvy the segment is.

      • YDD

        What’s a good limit for gradients with a modern EMU? In the back of my mind I have 1-in-50 as a “limit to aim for” and 1-in-37 as “don’t want to do that” (with 1-in-11 being the physical limit). However, I don’t know what modern rolling stock can tackle.

        • Alon Levy

          EMUs can technically go up grades up to ~12% in adhesion mode at glacial speed if everything works out correctly (all motors are running as intended, no rain, etc.). In practice, 7% is about the limit, and is done on some Swiss mountain railways and on many light rail lines. On mainline tracks the standards are more commonly 3.5-4%, but 3.5% is pretty conservative, and honestly even 5% is feasible given 21st-century technology.

          • adirondacker12800

            You’ve been in the Northeast in the autumn and winter? The “wet” season is the summer.

          • F-Line to Dudley

            Maximum grade anywhere on MBTA commuter rail is 3.5%, at the Wellington tunnel on the Haverhill/Reading Line shared with the Orange Line. And that one will probably go away someday not-too-distant as the freight junction it underpasses (Medford Industrial Track) has been out-of-service 9 years and is a highly likely abandonment next time Pan Am does a company housecleaning of unused holdings. When they get around to it they’ll probably fill the tunnel in and run both modes at-grade on the surface. Only reason it was allowed that steep to begin with is because the Reading Line was intended to be an eventual conversion to Orange Line extension when that tunnel was built 45 years ago.

            Next-steepest is 3%: (1) Mystic River Bridge (Somerville-Everett) on the Eastern Route mainline…Everett side being couple tenths % steeper. (2) Neponset River Bridge (Dorchester-Quincy) on the Old Colony mainline. Both were fixed-span replacements built in last 30 years for former drawbridges, which made them special exceptions for 3%. Nothing else exceeds 2%. The T’s track structures design guide sets an uppermost limit of 2% for new construction except for extraordinary cases meriting special exemption like a drawbridge-for-fixed span replacement scenario (only potential case on the current system: Saugus Draw on the Eastern between Revere and Lynn, which won’t need to be nearly that steep if it’s redone fixed). 1% is the recommended max for new construction.

            Grades for the Boston & Albany mainline cross-state are available here from a 1950 track map: Max grade is Middlefield Hill at the base of the Berkshires at 1.52%. It’s a several-mile long 1.52%, such that CSX dispatches a helper loco job out of Pittsfield and Chester to help with the push up the grade for the monster IM trains. But the very modest climb speaks to how extremely overbuilt the B&A grading is for the terrain it passes through. Utterly trivial for passenger trains. Single-loco Lake Shore Ltd.’s don’t strain at all up Middlefield, and it’ll get better still when the aging GE Genesis power gets replaced by the in-progress order of long-distance configuration Siemens Chargers units within next 2-3 years. Curves are many, many times more a performance issue for equipment and schedules on the B&A than grades.

  8. Yom Sen

    (1st message sent on Friday seems to have been lost)

    I see that Boston – Worcester would be done in 45 min with 8 intermediate stops instead of currently 79 min with 9 stops and 66 min with only 2 stops. How come such improvement? You’re mentionning electrification and high platforms. My understanding is that electrification will mostly improve acceleration since diesel trains are also able to reach 160 km/h and high platforms reduce time spent in stations. We should see then little improvement for direct trains, however Worcester – Lansdowne in your project would be reduced from currently 55 min non stop to 42 min with 6 intermediate stops.

  9. Pingback: Resist the Urge to Start Small | Pedestrian Observations
  10. Pingback: Quick Note: Timed Orbital Buses | Pedestrian Observations
  11. Pingback: Integrated Timed Transfer Schedules for Buses | Pedestrian Observations
  12. Pingback: Pulses (Hoisted from Comments) | Pedestrian Observations
  13. Pingback: Austerity is Inefficient | Pedestrian Observations
  14. Pingback: One- and Two-Seat Rides | Pedestrian Observations
  15. Pingback: Can Intercity Trains into Boston Enter from Springfield? | Pedestrian Observations

Leave a Reply to F-Line to Dudley Cancel reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.