Follow Dave on Twitter for real-time updates. And join the conversation on Facebook.
Don't forget to visit the Helpful Links page, the Helpful Information page, the Turn Table page, and the Map page. Also, I've added a newsletter service for those of you who want to get an e-mail notification when this website is updated. For more information, click here.
Follow Dave on Twitter for real-time updates. And join the conversation on Facebook.
The current proposal to "suspend" weekend MBTA Commuter Rail service is the extreme endpoint of a game that has been played for many years with the finances of the MBTA. A "draconian service cut" such as this has been predicted since at least 2009 - and maybe it's surprising that it took so long to finally get here.
The basic problem is the mandate that the MBTA balance its budget. Although that sounds reasonable, the devil is in the details. And the details in this case are a sudden reallocation of additional state assistance towards capital projects rather than to offset legacy debt costs.
In 2015, Governor Baker's own "Special Panel" recommended using $187 million of additional state funding to help the MBTA balance their budget. The latest budget proposal from the MBTA still asks for $187 million of additional state funding - but now that money won't go to operations (and debt service) but rather to capital improvements.
Without that money in the operations budget, the MBTA has to cut costs in order to balance the budget. Eliminating weekend Commuter Rail service and expanded "Ride" services are two of the biggest line items being cut to balance the budget - and supposedly "allow" the $187 million to go towards capital improvements and not operations.
The solution is obvious: take some of that $187 million and put it back into the operating budget. Only $17 million of the $187 million is needed in order to save both weekend Commuter Rail service AND the expanded Ride services!
THIS SOLUTION DOES NOT REQUIRE ANY ADDITIONAL FUNDING FROM THE STATE. It simply reallocates state budget money already being earmarked for the MBTA.
Skip down to the end of this blog post for the details on how you can help to make this happen.
How did we get here?
As described in the aptly titled "Born Broke" report from 2009, when the "Forward Funding" law was enacted in 2000 the MBTA was burdened with $3.3 billion in debt. Much of that debt was related to the Big Dig - and that debt is even called out in the recent presentation regarding the budget (see page 48 here).
The law also required that the MBTA balance its budget using "own source revenue" (like fares, advertising, etc.) and a fixed share of the state sales tax. In many fiscal years, that hasn't worked out, and the state has had to allocate additional funding beyond those sources in order for the MBTA to balance their budget.
The conclusion of the Born Broke report sums up the situation both then and now: "The underperformance of the sales tax as a principal financing source and too much debt are the causes of the T’s structural weaknesses. Until these factors are addressed, no amount of reorganization, efficiencies, or reforms will allow prevent deficits in FY10 or in the future."
Hence the prediction in the Executive Summary of the Born Broke report that "without external assistance in the form of debt relief or new revenue the Authority will be forced to make draconian service cuts and impose dramatic fare increases."
We've already seen the dramatic fare increases. With the loss of the previous 'external assistance,' we're now on to the draconian service cuts.
The oft-cited D'Alessandro Report from November 2009 reaches much of the same conclusions.
Those reports are from quite a few years ago, what about now?
For many of the intervening years since "Forward Funding" was enacted in 2000 (both prior to and after 2009), the MBTA has been able to cobble together enough 'reforms,' revenue increases (fare hikes), and additional state funding to avoid drastic service changes. Some of those reforms - especially recent reforms - have been wholly positive and beneficial. These include the investigation and move to improve the parking revenue stream and the restructuring of debt in a fiscally responsible way (see page 8 here) among others. Other attempts, especially in the early to mid-2000's, were essentially shell games to keep the budget balanced while avoiding the larger real issues. Many reforms have been controversial - privatization of the 'cash room' and other services. Other reforms haven't made the headlines, but have affected many - "30% reduction in Corporate Headquarters/Administrative positions" (see page 14 here) for example (and then we wonder how designs like Auburndale slip through?).
In April 2015, Governor Baker commissioned a Special Panel to investigate the MBTA - primarily as a direct result of the disastrous winter of 2015. The report that was produced explicitly "rejects the ‘reform vs. revenue’ debate because the MBTA needs both."
Most of their proposed solutions were implemented - the Fiscal and Management Control Board was implemented as a result of the report. That report also laid the groundwork for the fare increase in July 2016.
However, one of their recommendations that wasn't implemented was that the Legislature should "Limit future General Fund operating assistance by purpose to cover (1) debt service payments (2) employee costs for staff moving off of the capital budget."
Huh? That's some mumbo-jumbo. Let's break that down. Since "Forward Funding" was implemented, it has never worked perfectly. In many fiscal years, the legislature has allocated additional money to the MBTA so that the MBTA can balance their books. That's the "General Fund operating assistance" being talked about. [The "employee costs for staff moving off the capital budget" is another fascinating story - essentially the MBTA was using capital funds to pay operating expenses. That's now been fixed.]
The Governor's Special Panel recognized that the FY16 "General Fund operating assistance" was $187 million. The 2015 report included this slide:
Allow me to summarize that slide in one phrase: The State should take back the legacy and Big Dig debt from the MBTA. That would allow the MBTA to implement the goals of "Forward Funding" without the weight of that debt dragging the T's budget down.
Doesn't that sound familiar? IT'S THE SAME CONCEPT RECOMMENDED IN BOTH 2009 REPORTS.
But alas, that's not what is being proposed with this year's budget. This slide from the March 13, 2017 presentation to the FMCB still includes a mention of the $187 million supplemental assistance - but instead of being used to pay down the debt, it is now allocated to capital improvements:
That's from page 15 of the presentation - I've added the highlighting.
Although it is an admirable goal to set aside funding for capital improvements, reallocating the additional state assistance away from the operating budget means that SOMEBODY STILL HAS TO PAY THE BILLS - the "bills" being the debt service on the mountain of debt that the MBTA carries.
Enough Dave! I can't stand all this talk about debt and General Funds and State Assistance...
This is all backdrop to where we find ourselves today. The MBTA has set this goal of being able to balance the budget without using the $187 million towards operating expenses. How do they do they do that? The same way you balance your budget at home: reduce expenses.
So forget everything you're reading about operating subsides, ridership counts, and other factors related to weekend Commuter Rail Service. Elimination of weekend Commuter Rail service is intended to do one thing only: scrape up some cash to balance the budget and to make it appear that the $187 million in additional state assistance can be removed from the operating budget and allocated to capital improvements.
Let's dive into the details - but first, we need to make sure we're all on the same page:
Capital budgets vs. Operating budgets
Remember, capital expenses (building stations, the Green Line Extension, the purchase of red and orange line subway cars, etc.) are NOT part of the operating budget. The operating budget pays for operations - bus drivers, subway operators, and the Keolis subcontract among other things. So don't get confused when you hear that the MBTA isn't spending enough capital - they actually aren't spending all that they could in terms of capital funding. Some of that is due to the delay in the Green Line extension, but capital spending is a story for a different day (see slide 17 here if you want details on the capital spending problem).
Keolis is a subcontractor
Keolis, as the private operator of the Commuter Rail, has a contract to perform a service. As part of the contractual arrangement, they are paid on a monthly basis for the work performed. If the level of service that they are required to perform changes (for example, more or less trains on a schedule), then their compensation would change. This is an important fact that many folks don't understand. The MBTA sets the service level and implements policies. Keolis is essentially obligated to do what the MBTA tells them to do. Clearly they have nothing to do with the proposal to eliminate weekend service - it would reduce the payments they get!
Commuter Rail Revenue goes to the MBTA
Every penny of every fare collected by Keolis goes directly to the MBTA. Keolis does not 'keep' any of the revenue they collect - they are paid separately by the MBTA. Again, there are probably plenty of people who think that the compensation for Keolis is somehow tied to the fares collected. It just simply isn't true. The value of the payments to Keolis are independent of the value of the fares collected.
Most of the cost of weekend Commuter Rail service has already been paid for
Saving $10 million while cancelling 104 days of service doesn't seem like a great bargain because it isn't. The problem is that the only savings that can be recognized are the small incremental staff and fuel costs for weekend service. The other costs - rolling stock, infrastructure, and maintenance aren't affected enough by the reduction in weekend service. The rails and fleet of equipment still need to be maintained, and the reduction in 'wear and tear' isn't enough to change the price that the MBTA is paying for those long-term services.
Fine, just please tell me why Weekend Service is being cut!
By cancelling weekend Commuter Rail service, the MBTA projects that they can save $10 million. Presumably this value is from:
1) Lower cash payments to Keolis; but offset by
2) Lower revenue from weekend fare collection.
In other words, if you stop going out to dinner every weekend, you'll save the money that you would usually be paying the restaurant. But there is an offsetting cost - you still need to eat, so the cost of food that you're cooking at home offsets the savings. The savings CAN'T be the total of your restaurant bills. Same for Commuter Rail - there won't be any cost for running the trains on the weekend, but there also won't be any fares being collected.
There has been no explanation of how they are calculating the $10 million value - and we need it.
This page from the recent budget presentation shows how the $10 million works to close the projected $42 million budget gap:
That slide is a little confusing, but read it from right to left. Without any changes, and assuming the $187 million of additional state assistance does NOT go to the operating budget, then the MBTA predicts that they will have a $42 million structural deficit.
On top of that $42 million deficit, they want to spend $7 million on strategic operations hires (why are they needed? see: Auburndale). That means that they have to find $49 million in cost reductions or additional revenue to get to a balanced budget. Each column on that chart represents an incremental move from the total $49 million deficit towards the balanced budget on the left.
Clearly the solution is to follow the recommendations of the Governor's own Special Panel in 2015: maintain additional state assistance to the operating budget for the purposes of payment of debt service. That 'frees up' funding that can be used to avoid making some or all of the $49 million in cuts to 'balance the budget.'
What? It's that simple?
Well, yes and no. Clearly it's that simple with regards to the reallocation of the $187 million of additional state assistance. But once I discovered this whole story, I realized it leads down lots of other paths:
1) With the MBTA typically not spending all of their capital funding, the $187 million allocated to the MBTA under this budget proposal could easily be 'taken back' towards the end of the fiscal year if it is unspent. That could help the administration balance the overall state budget if needed. Another shell game with MBTA money.
2) The idea of using the $187 million towards capital improvements allows the administration to claim that they are 'investing in the MBTA.' But it avoids what the 2015 Special Panel found was the REAL solution: Balance the MBTA budget with the $187 million of additional state assistance to the operating budget AND find NEW sources of ADDITIONAL REVENUE to provide the MBTA with funding for capital projects. Another shell game with MBTA money.
3) Highlighting the "subsidy" that is needed for various weekend Commuter Rail trips is not a valid measurement of the savings that can be achieved. Those "subsidy" values include a percentage of the fixed cost of the Commuter Rail network, which, as noted above, isn't a fair way to calculate the savings. Steve Koczela does the math that matters in this tweet and shows that the actual savings averages out to $6.24 per weekend rider. The MBTA presentation of subsidies that aren't real is another shell game with MBTA money.
4) Reform before revenue has failed. Every report that has been commissioned to look at the MBTA - including Governor Baker's own 2015 Special Panel - has reached the conclusion that the MBTA needs both reform AND revenue. As noted above, some of the recent reforms have been good - and some of the controversial and/or painful reforms are probably also good. But the idea that the MBTA can continue to balance their budget without new revenue AND avoid draconian service cuts has finally been proven false. Another shell game with MBTA money.
What do we do?
This is where it becomes very simple. If you want a world class transit system that isn't the only major Commuter Rail system that shuts down on the weekend, then DEMAND IT.
Remember that the MBTA is not a completely independent state agency. Clearly the MBTA didn't come up with this budget proposal that includes the idea of moving the $187 million around without participation from the corner office in the State House. So calling the MBTA isn't really what needs to happen.
Call your legislator and the governor's office (617.725.4005 or 888.870.7770) and let them know what you think. Here are my ideas for what I'm going to tell them:
1) Stop playing shell games with the MBTA budget.
2) Allocate at least some of the $187 million back to the operating budget so that weekend service and the Ride can be preserved.
3) Actually, allocate the full $187 million back to the operating budget so the MBTA can restore operations staff and systems to where they need to be while paying down more of the legacy and Big Dig debt.
4) Find new additional revenue sources for capital improvements for the MBTA, as recommended by Governor Baker's Special Panel in 2015.
Use this link to find the phone numbers for your legislator.
Changing the Heart to Hub train to times that align with a 9-5 workday in Boston (as recently requested by Worcester politicians) will devastate the schedule for the non-Worcester ridership of the Framingham-Worcester line. It's just not possible with the current infrastructure that we have.
Recently, Worcester Mayor Joseph Petty sent a letter to the MBTA requesting that the times of the Heart To Hub train be changed to be better aligned with a 9-5 workday and that no stops be added to the train. While these might be reasonable requests that would add value to MBTA Commuter Rail passengers from the Worcester area, those requests need to be considered in the context of the entire Framingham-Worcester line ridership.
[I discuss the rationale of adding Framingham and Ashland as station stops to the evening Heart to Hub train at item #9 in this previous blog post. I won't repeat that explanation here.]
The current schedule of the Heart to Hub train is:
Depart Worcester: 8:05 AM
Arrive South Station: 9:07 AM
Depart South Station: 7:35 PM
Arrive Worcester: 8:40 PM
The Heart to Hub train also stops at Yawkey and Back Bay but currently does not stop at any other station between Yawkey and Worcester. The Heart to Hub train was originally announced in October 2015 and was added to the schedule in May 2016. There are many blog posts here about the Heart to Hub train and they can all be found using the "Heart to Hub" category link for this blog.
The draft schedule that will be implemented in May 2017 proposes these changes for the Heart to Hub train:
Depart Worcester: 8:00 AM
Arrive South Station: 9:06 AM
Depart South Station: 7:35 PM
Depart Framingham: 8:12 PM (NEW STOP)
Depart Ashland: 8:18 PM (NEW STOP)
Arrive Worcester: 8:45 PM
Clearly neither the current nor the proposed times for the Heart to Hub train are ideal for the vast majority of commuters working a 'typical' 9:00 AM to 5:00 PM (or close) workday in Boston. In that context, the request by the Worcester politicians seems to make sense. But the solution - a schedule that answers this request - is fraught with problems.
Hypothetical "Ideal" Schedule for Heart to Hub train
In order to demonstrate the problems with a modified schedule for the Heart to Hub train, I have developed a hypothetical schedule with the Heart to Hub train at a 'more desirable' time. For simplicity, I have only modeled the evening commute, but the morning commute has similar problems.
In order to develop this schedule, I have applied the same constraints that currently exist for the infrastructure that exists on the Framingham-Worcester line. These constraints are:
1) Minimum of 10 minutes between departures from South Station.
2) Minimum of 12 minutes between any trains arriving at Framingham.
3) Minimum of 30 minutes between trains arriving at Worcester.
The reasons for these constraints are explained at the bottom of this blog post.
The other 'rules' I applied were not really constraints, but were used to guide the process of building the schedule:
1) All trains destined to points west of Framingham are express trains. These express trains skip Boston Landing and all stations before West Natick; and
2) All local trains terminate at Framingham and make all stops between Boston and Framingham.
These rules (and common sense) dictate that a local train can depart close behind an express train - but not the other way around. It is impossible to have a local train depart just before an express train - the local train would block the track!
NOTE: I am NOT proposing that the outbound Heart to Hub train stop in Framingham. I have included the time that it PASSES THROUGH Framingham because that is a constraint in scheduling it. Hence the italicized "strikethrough" for that time - it is NOT A STOP, but that time is the approximate time the train passes through Framingham.
The methodology used to develop this hypothetical schedule was as follows:
1) I aimed to have the Heart To Hub depart Boston at a 'reasonable' time for a worker leaving their office at 5 PM. Assigning a departure time of 5:10 PM seemed to be reasonable to account for walking to the station. It also allowed me to keep the three earlier trains (3:30 PM, 4:20 PM, and 4:30 PM departures) at the same times as they are proposed on the May 2017 draft schedule.
2) Having the Heart to Hub depart at 5:10 PM 'works' because it fulfills all three of the operating constraints:
a) It departs 40 minutes after the previous departure from Boston;
b) It passes through Framingham approximately 24 minutes after the previous train; and
c) It arrives in Worcester 31 minutes after the previous train.
These metrics also show that it is impossible to move those previous trains to LATER departure times:
d) if the 4:20 PM express train departed later, it would create a conflict at Worcester with less than 30 minutes between arrivals; and
e) if the 4:30 PM local train departed later, the Heart to Hub train could possibly 'catch up' to it somewhere prior to Framingham (although I will acknowledge that the 4:30 PM local train could probably move 5-10 minutes later).
3) For the remainder of the schedule, I kept the total number of outbound rush hour trips at 10, which is the same as what is proposed on the May 2017 draft schedule for the PM rush hour PLUS the Heart to Hub.
4) With the Heart to Hub departure and arrival times set and the earlier train times locked in, the next step is to build the remainder of the PM schedule. First, we can add an express train with a departure time from Boston of 5:20 PM. Again, this train just barely fulfills all three criteria:
a) It departs Boston 10 minutes after the previous train;
b) It passes through Framingham 16 minutes after the Heart to Hub train; and
c) It arrives in Worcester 29 minutes after the Heart to Hub train.
5) The next train needs to be a local train in order to offer service to stops before West Natick. To comply with the scheduling constraints, the local train can depart at 5:30 PM. This train then meets all three criteria:
a) It departs Boston 10 minutes after the previous train;
b) It arrives at Framingham 21 minutes after the previous train; and
c) Worcester is not a factor since it terminates in Framingham.
6) To maintain consistent service to all stations, the next train is an express. Assigning it a 6:00 PM departure from South Station allows it to fulfill all three criteria:
a) It departs Boston 30 minutes after the previous train;
b) It arrives at Framingham 18 minutes after the previous train; and
c) It arrives at Worcester 38 minutes after the previous train.
7) The 6:10 PM, 6:40 PM, and 6:50 PM departures from South Station follow the same pattern established for the 5:20 PM, 5:30 PM, and 6:00 PM departures from South Station.
What's the Problem?
The above hypothetical schedule for Heart To Hub passengers probably looks great if you want to get to Worcester after leaving your office at 5 PM. So where's the problem? First, let's take a look at the draft schedule proposed for May 2017.
A comparison of this proposed draft with the hypothetical schedule reveals the problems:
1) Passengers on P521 (the 5:00 PM express) have their commute pushed 20 minutes later. This train is the most popular train in the evening commute and carries an approximate passenger load of 1,400 passengers. For simplicity's sake, let's assume that those 1400 passengers are evenly distributed to all 7 destinations (which is NOT a valid assumption - data shows that Framingham and West Natick have much higher ridership than other stations). This assumption means that approximately 200 passengers are destined for Worcester - and those passengers would benefit from the new Heart to Hub times. BUT THE OTHER 1,200 PASSENGERS WOULD BE DELAYED 20 MINUTES. Even if you try to stretch the ridership count at Worcester to assume that Worcester has a disproportionate share of the ridership or assume that some Grafton passengers would drive to Worcester to take advantage of the Heart to Hub, there still is over one thousand passengers going to the other stations.
2) Passengers on P593, (the 5:10 PM local) have their commute pushed 20 minutes later. This train is the most popular local train in the evening commute and carries an approximate passenger load of 630 passengers. None of these passengers are destined for Worcester, so ALL 630 PASSENGERS WOULD BE DELAYED 20 MINUTES.
That is a total of approximately 1,800 passengers that have had their commute made 20 minutes later. That far exceeds the TOTAL number of passengers using Worcester Station for their commute.
3) With the hypothetical schedule above, there is a ONE HOUR GAP IN SERVICE FOR ALL STATIONS (except West Natick and Framingham). For example, if you want to go to Wellesley Square, you can only arrive there at 5:08 PM or 6:09 PM. Similarly, if you are destined to Westborough, you can only arrive there at 5:25 PM or 6:26 PM. This is at the height of the rush hour. The proposed May 2017 schedule has trains arriving at those stations every 40 minutes during the same period. This gap in service is a logical consequence of a super express train - the track needs to be clear so that the Heart to Hub can speed through directly to Worcester.
4) The modified arrival times for the 1,800 passengers delayed 20 minutes could have serious consequences for those passengers. During many recent public comment periods, passengers have noted that many after-school child care facilities close at 6 PM. On the current and proposed May 2017 schedule, passengers leaving work around 5 PM can get to their destination with enough time for child care pickup before 6 PM at all stations between Boston and Framingham (Ashland is too close for my comfort - but maybe some passengers try to do it). The hypothetical ideal Heart to Hub schedule prevents anyone from arriving at their destination with enough time to get somewhere else by 6 PM if they leave work around 5 PM.
5) The ~5 PM express train and ~5:15 local train have existed for at least 15 years. Many passengers have built their daily commute / work / life schedules based on these train times, and disrupting that schedule could inconvenience them or push them away from the Commuter Rail. I'm not normally a fan of doing something because "that's the way it's always been done," but when it comes to schedules such as these, there should be some deference given to historical precedent.
Changing the time of the Heart to Hub train to be "better" for Worcester bound passengers leaving their offices at 5 PM clearly presents too many insurmountable problems for passengers from every other station on the line. That's why it can't happen with the current infrastructure on the line.
What's the solution?
If the Heart to Hub train can't be moved around on the schedule to a 'better' time, than what can be done to improve the experience for Worcester (and all) passengers on the line? Better infrastructure is the answer...and in particular:
1) A second platform at Worcester Station and/or improved track layouts around Worcester. Once you start playing with the schedule, you quickly realize that the limitation of a minimum of 30 minutes between arriving trains at Worcester really handicaps any kind of creative scheduling. Eliminating that constraint by adding a second platform or reconfiguring the track layouts could eliminate or change that constraint, which could allow for better scheduling options. See footnote #3 below about Worcester station.
2) Upgrade the speed limit on the line (or sections of the line). Speed limits on a railroad are governed by many factors: track bed and rail conditions, track geometry (it's considered bad form for trains to go flying off of curves at high speeds), signal spacing, and safety considerations (density of grade crossings). Some of those constraints can't be corrected - the curvature of the track can't be changed without realigning entire sections of the railroad, which would most likely be impossible in the densely developed area around the Framingham-Worcester line. But some sections of track are straight or nearly straight, and those sections could have the speed limit changed. That might require upgrading the track bed and/or modifying the signal system, but those are changes which can be done.
3) Install a passing track somewhere on the line. In years past, the schedule included "leapfrog" moves where an express train would pass a local train - while both were travelling in the same direction. With only two tracks between Boston and Worcester, this required opposing traffic to be scheduled so that it would not conflict with the leapfrog moves. This resulted in very tight timing of the schedules for moving trains back and forth. Any upset or delay to the schedule would quickly cascade to multiple trains. Recent schedules have eliminated the "leapfrog" moves, since they introduced an unacceptable risk of cascading delays. I believe this is the correct answer given the infrastructure we have.
Adding a passing track (or two!) somewhere between Boston and Framingham would allow these "leapfrog" moves to be added back to the schedule. The real estate exists for additional tracks between Weston and Framingham - this section once had four tracks. Most bridge abutments accommodate four tracks, although most station platforms would need to be rebuilt. The ideal configuration would be to put the passing track(s) between the outer "local" service tracks, although having the local tracks in the middle serviced by center island platforms with the express tracks on the outside could also work.
Adding a passing track would allow the Heart to Hub train to move past trains making station stops - reducing or eliminating the long gaps in service for those stations while the track is kept clear for the Heart to Hub with the current infrastructure.
4) Installing high level platforms. Dwell time is the time a train spends in a station while passengers embark or disembark. Dwell time is a large factor in the overall duration of a train's trip. As ridership increases, dwell times increase - add 30 seconds of dwell time at 6 stations and you've added 3 minutes to the duration of a trip. High level platforms - where passengers step directly onto the train rather than needing to climb the stairs on the train from the platform into the train vestibule - dramatically reduce dwell time. Not only can passengers board more quickly (climbing stairs is slow), but ALL of the doors of a train can be remotely opened and closed (like a subway car) allowing MORE passengers to board at the same time. Both Yawkey and South Station have high level platforms for the entire length of every platform - and the soon to open Boston Landing station will also have full length high level platforms. But every other station (including Back Bay) has either all low level platforms or a combination of low level platforms and a "mini-high" platform that offers some handicapped accessibility.
Constructing high level platforms at every station would be expensive - but they don't all need to be done at once. Changing the highest ridership stations to high level platforms (especially Back Bay) would be money well spent.
5) Electrification. Now we're talking about very long term but very beneficial capital investment. Electric powered trains offer many advantages over the diesel locomotive trains that the MBTA now uses but the most relevant for this discussion is the quicker acceleration and deceleration (stopping time). Electric trains would decrease the overall duration of every trip - and would be compatible with the proposed North-South Rail Link tunnel through Boston.
6) Stop getting hung up on the marketing buzz of "one hour Worcester to Boston" (which isn't even true anyway). Obviously everyone wants a direct train from their station to Boston at the most convenient time for them. With only two tracks, it just isn't possible. So instead of trying to promote service that benefits one station at the expense of other stations, why not just focus on the really good express train service that Worcester DOES HAVE regardless of the Heart to Hub? Why not add more express trains (that stop at all stations from Framingham to Worcester) paired with local trains so nobody loses service but everyone gets better service?
Hopefully Worcester residents and politicians will understand that a more comprehensive and collaborative approach to the Framingham-Worcester Commuter Rail line schedule will result in benefits for everyone. That's more productive than pitting the ridership from one station against the ridership from other stations in an unwinnable civil war. Let's work together to make Commuter Rail great again!
Footnotes / Technical reasons for the three constraints:
1) Minimum of 10 minutes between departures from South Station. The block signalling system effectively requires about 8 minutes between trains on the Framingham-Worcester line so that a train behind another train can proceed on 'clear' signals. Rounding that up to 10 minutes allows for some contingency. For a more comprehensive explanation of block signalling, see "Automatic Block Signal System" in the Glossary. Watch the corny little video linked from that definition.
2) Minimum of 12 minutes between any trains arriving at Framingham. Again, this is due to the block signalling system. At Framingham, the signals and switches are densely spaced, which is a good thing - lots of options and room for trains to go to different tracks and get out of the way. So why 12 minutes instead of 8 or 10? Framingham is 21 track miles from Boston - and trains making station stops or just travelling over those 21 miles have more opportunities to encounter minor delays. Moving the spacing to every 12 minutes at Framingham allows for some contingency and reduces the chances of delays for a following train. This constraint actually doesn't come into play very much since the first and last constraints effectively govern the schedule.
3) Minimum of 30 minutes between trains arriving at Worcester. Worcester Union Station is served by a single platform on a siding track off the main line. This track is effectively a dead end for the MBTA - the track past Worcester station is owned and dispatched by CSX - and movements by MBTA / Keolis trains onto CSX property are complicated and difficult. In addition, the storage yard for MBTA trains is EAST of Worcester station. This means that after a train arriving at Worcester is unloaded of passengers, it has to reverse direction and move towards Boston then reverse direction AGAIN and move into the storage yard - all the while, blocking access into the station for any other train. The switch and track configuration in this area does not allow for parallel train movements or train movements around the storage track "entrance."
The new design of a proposed change to the Auburndale Commuter Rail station moves the platform from track 2 to track 1. This doesn't sound like a big deal - trains can just switch tracks, right? It's not that simple, and this proposed design will either result in massive disruption to the entire schedule (for EVERYONE) or massive changes to service at the Auburndale station.
The simple problem is that moving trains back and forth from one track to another blocks both tracks for opposing traffic while the switches are aligned for the track change and while the train moves from one track to another. Properly implementing this would require precise timing for trains to meet (or actually NOT meet) at these locations. The current schedule is not designed to accommodate that timing. And with the AM and PM rush hour schedules jam packed with trains, tweaking the times of ANY train will require changing the times for ALL TRAINS. THIS IS WHY THIS ONE STATION DESIGN AFFECTS EVERYONE ON THIS LINE. It's not just a problem for the Auburndale passengers.
I'll dive into the details below and present some potential solutions at the bottom. Skip forward to those if you're not interested in the technical details.
100% Design Now Complete
On February 15, 2017, the MBTA held a public meeting to present the final design for the reconstruction of the Auburndale Commuter Rail Station. I joined about 50 people to listen to the presentation which was led by a combination of the MBTA Capital Delivery Department Project Manager, the Design Consultant Project Manager, and the Project Manager from the architecture firm. Much thanks to the Village Bank in Auburndale for their hospitality and for providing snacks!
About 17 people spoke to provide comments on the station design. Many comments focused on construction issues such as the potential loss of parking and road closures while the station is being rebuilt. Some relevant comments related to the design and final product included:
- Shelter has no walls to provide protection from the wind (multiple comments on this topic);
- Design does not include screening between the station and the Mass Pike; and
- Thanks and praise for the accessibility that the new station will provide.
I used my public comment time to highlight the switching platform problem and the potential schedule and/or service disruption problem.
Ari Ofsevit spoke after me and echoed the concerns about a single platform station on track 1. He picked up on a potential interesting solution to part of the problem - please read his blog post for his summary of the problem and his idea for a solution.
Why design a new station with a platform on one track only?
Since the Mass Pike was constructed along the railroad right-of-way in the 1960's, the three Newton stations have had a platform on the track 2 side only. This is the reason that those stations have no 'reverse peak' service: track 2 is used for inbound AM trains AND outbound PM trains. Similarly, track 1 is used for outbound AM trains and inbound PM trains - and without a platform on track 1 at any of those stations, none of those trains can stop there.
The concept of a redesign of the Auburndale station has been ongoing for many years - driven by the local community and local politicians. The major goal of the redesign appears to have been to provide an accessible station - but both a handicapped accessible station AND a more accessible station to the village center and the surrounding community.
Maintaining a one platform station (rather than constructing platforms on both sides of the station) is apparently a strategy that was adopted to minimize the cost of the project. It is important to note that the Americans with Disabilities Act essentially mandates that when any improvements or changes are made to a Commuter Rail station that exceed 30% of the assessed value of the station, all elements of the station must be made fully ADA-accessible. These accessibility requirements require full length high level platforms and the 'typical' accessibility elements that we are familiar with at newer stations like Yawkey and Boston Landing. This means that it is not legal nor acceptable to build a new platform on track 1 and keep the low level existing platform on track 2.
In 2013, a 30% design review meeting was held for the Auburndale station where a single rebuilt platform on track 2 was presented (at least as one of the options). This design required passengers to go "up and over" both tracks to get from the village center to the platform. According to reports, this design was widely criticized at that public meeting and the public advocated or agreed that having the station platform on track 1 was the correct option - where the "up and over" is not needed. There does not appear to have ever been a concept or plan to design or build a two-platform solution, nor did the project team convey the operational issues with a platform on the track 1 side only.
In my opinion, NO COMMUTER RAIL STATION on a two track line should ever be allowed to have a one platform station designed, constructed, or even talked about. Commuter Rail stations don't get built very often, and having a station with a platform on only one track potentially locks that station into reduced service for many years.
The current design for the Auburndale station (with the new platform on track 1) includes a new "universal interlocking" just east of the station. This interlocking is a set of switches that allows a train to switch from either track to the other track. The new interlocking ("CP 10") will mean that Auburndale will have interlockings on either side of the station (CP 11 already exists). The intent of adding a new interlocking was to provide a way for trains to switch from track 2 to track 1 to make a station stop at the new platform on track 1. It is apparent that everyone involved assumed that these interlockings and this switching back-and-forth concept would allow for either the same level of service or even better service. BUT THERE WAS NO MODELING OF THE SCHEDULE TO PROVE THAT.
During the public meeting, I described the problem as this: even with two interlockings on either side of the station, there are still essentially two one-way streets pointed at each other. Switching trains back and forth will require precise timing and probably will require some trains pausing to allow these switching moves to happen. This is less than ideal - well, actually, it's disastrous.
What does switching the platform from track 2 to track 1 do to the schedule?
Let's focus on the AM commute to see how disastrous the track change could be. First, we'll take a look at how things work on the proposed May 2017 schedule. As discussed above, track 2 is the inbound track and track 1 is the outbound track for the AM commute (actually from Framingham all the way to Boston). Keeping the tracks dedicated to these 'directions' for the AM commute allows for unimpeded flow. I've added the approximate times that EVERY train passes Auburndale to the schedule image below - the orange boxes with italicized times are NOT station stops, but rather the times that a train passes Auburndale without stopping.
If we focus on the Auburndale station stop times and rotate the data, we get the table below.
Track 1 in Boston: P587 moving westbound, approaching Boston Landing
Track 2 near Wellesley Farms: P584 departing the station stop.
Track 1 at Boston Landing: P587 departing the station stop.
Track 2 at CP 11: P584 passing inbound through CP 11 and switching to track 1.
Track 1 at Auburndale: P584 making the station stop.
Track 1 near West Newton: P587 approaching CP 10.
Track 1 at CP 10: P584 stopped inbound at CP 10 awaiting P587.
Track 1 at CP 10: P587 switching from track 1 to track 2.
Track 1 at CP 10: P584 stopped while switches are realigned to allow it to switch to track 2.
Track 2 at Auburndable: P587 passing P584 and moving outbound on track 2.
Track 1 at CP 10: P584 moving inbound from track 1 to track 2 towards the station stops at West Newton
Track 2 at CP 11: P587 moving outbound from track 2 back to track 1.
This scenario has taken less than 10 minutes but it involves about 8 miles of track along with each train passing through FOUR switches - two at each interlocking. Obviously that's less than ideal - and it doesn't even work (one train needed to wait at a signal / interlocking). And we've only tried to solve ONE of the conflicts.
Part of the problem is the one mile length of the "wrong direction" track at Auburndale. For these scenarios to work, it is almost impossible to get the timing exact so that the trains are both on the 'wrong' track at the same time - in other words, having the inbound train moving through CP 11 from track 2 to 1 and the outbound train moving from track 1 to 2 at CP 10 AT EXACTLY THE SAME MOMENT. If either of those movements doesn't happen at the same time, then it is likely one train will get through their set of switches but will arrive at the next interlocking before the other train has cleared through it. With only ONE mile between interlockings, there just isn't enough time for anything to go less than perfectly. The first train will have to stop at the 'blocked' interlocking and wait for the other train to clear the interlocking.
This can be illustrated by attempting to fix the CP 10 conflict in our model scenario above by moving the operation of P587 five minutes earlier. But this just moves the conflict to CP 11! P587 will arrive at CP 11 before P584 has had time to switch from track 2 to track 1 at CP 11.
This analysis proves that to have trains pass each other at Auburndale on the 'wrong' tracks requires precisely timed meets that would have to occur with the precision measured in seconds. Any delay of even a few minutes to one of the trains involved in the meet would most likely delay the other train. Most of you realize that keeping trains on time to the precision of under a minute is not a realistic goal on this line.
With this short distance between interlockings, the clear solution is to move one train completely through BOTH interlockings before the other train arrives. Then the timing does not have to be as precise, since the second train just needs to arrive after the first train has cleared both interlockings. But wait a minute...by doing that, THE RAILROAD IS EFFECTIVELY REDUCED TO A SINGLE TRACK AT THAT LOCATION! We've been waiting years for them to fix the single track bottleneck at Beacon Park, and now we're implementing a new one. That's one context to prove this won't work. Also remember that the new CP 10 interlocking can't be moved east to make the single track section longer - the West Newton platform is still on track 2 just east of new CP 10. Ugh.
But even if you accept the concept of an effective single track at Auburndale solution, the schedule consequences are massive. P587 would have to move 10-15 minutes earlier so that it could get past CP 11 before P584 arrived there... but once you start making changes that dramatic, the schedule completely falls apart for multiple different reasons:
1) Meets at Framingham (departure/arrival of local trains vs. expresses);
2) Arrival times of trains at Boston; and
3) The equipment cycle.
The equipment cycle problem is easy to illustrate since it isn't even possible to move P587 five minutes (let alone 10-15 minutes) earlier - it is using the equipment from P502 which arrives at South Station at 7:33 AM. With 15 minutes as the most reliable time to turn a train from inbound to outbound at South Station, moving the departure time of P587 from 7:48 AM to 7:43 AM would require moving P502 five minutes earlier... and you can see how the problems cascade exponentially (especially if you start moving departure times by 10 or 15 minutes). In fact, the AM schedule is completely jam packed at both Framingham and in Boston - so there isn't any way to tweak the times of any train without AFFECTING EVERY OTHER TRAIN FROM 6 AM TO 9 AM.
THIS IS WHY THIS ONE STATION DESIGN AFFECTS EVERYONE ON THIS LINE. It's not just a problem for Auburndale passengers.
But wait, it gets worse. The signal system of a railroad is designed to prevent collisions, and it does this by essentially warning a train crew about the condition of the rail and signals ahead. For example, when you're driving around town in your 2 ton car, you can see any traffic signal with plenty of time to stop. On a higher speed highway, there might be a warning sign that a traffic signal is ahead - and sometimes those even warn you of the CONDITION of the signal (i.e. signs which say "red signal ahead when flashing"). For a multi-ton train that isn't as easy to stop, this is exactly how the railroad signal system works. If a signal is red for stop, then signals BEFORE that red signal will require the train to start slowing down well before it reaches the red signal.
The implications of this for the CP 11 - CP 10 dance are clear. If our inbound P584 is switching from track 2 to track 1 to make the station stop at Auburndale, it is effectively occupying BOTH tracks in that area, and there will be stop signals facing an outbound train coming from Boston. The 'warning' signals that require the outbound P587 to slow down approaching the stop signals will stretch towards Boston for at least a few miles. This means that P587 is either going to have to slow down as it approaches the area (with the resultant negative schedule consequences) or the schedule will have to be adjusted to keep it away from that entire area until the signals can allow for the train to operate at normal full speed. Either way, the overall schedule is drastically affected. And remember - the schedule of an AM outbound train is critical for inbound service - those outbound trains have to get out to Worcester or Framingham in order to operate back inbound.
So what's the solution for this mess? There are a number of possibilities:
1) Build Auburndale with a platform on both tracks. This should be the ONLY solution. It allows for flexible scheduling with increased service for the reverse commute option. But it isn't funded and it isn't designed. Your first reaction might be that a two platform station would be much more expensive than the current plan, but that's not the case. The two-platform Yawkey station cost ~$13.5 million and the two-platform South Acton station recently cost ~$9.5 million. The average of those is $11.5 million - which is the amount budgeted / estimated for the current Auburndale design. We're getting a one platform station for about the same cost as a two platform station - because we're also getting a new interlocking and signal system upgrades (which don't really help us).
There are two sub-options under this solution:
a) Postpone implementation of the current design until a two platform solution can be designed, funded, and implemented. Obviously this delays accessibility for Auburndale station.
b) Modify the current design to incorporate elements that will allow for a two platform station in the future. For example, set aside space that can accommodate elevators, ramps, and other required elements to get across the tracks - even if they can't be built now.
2) Implement Ari's solution detailed in his blog - build new platforms on track 1 at all three Newton stations using the money budgeted for the new CP 10 interlocking (which Ari readily admits is inferior to building a two-platform station at Auburndale). This would presumably allow all three Newton stations to have the same rush-hour only service that they have now. It still may require some schedule changes, because Wellesley Hills and West Natick require rush hour service on track 2 only (see this blog post about that). So rush hour trains would still be required to switch tracks at CP 11. But it would be easier to manage ONE change of tracks rather than two for rush hour trains. Also note that this solution perpetuates the lack of reverse commute service throughout Newton. And with millions of dollars being spent on stations in Newton now, the second platform at each station will probably be delayed well into the distant future.
3) Build the station as designed and change the use of Auburndale. This concept eliminates rush hour 'normal' commute service at Auburndale but implements NEW 'reverse' commute service at Auburndale. In other words, since the 'reverse' commute trains are already using track 1, having them stop at the new Auburndale platform will not introduce the switching tracks CP 11 - CP 10 dance problem. Keep the trains traveling on the tracks they use today. Obviously the downside to this is the loss of the brand new Auburndale station to the ridership that uses the station the most - passengers commuting to and from Boston on a 'typical' schedule. And although the data is somewhat old, the indications are that Auburndale is the busiest of the three stations. This solution also means that the new CP 10 will be relatively unused (although more interlockings on a railroad are generally good, since they offer solutions to unforeseen problems).
4) Demand the MBTA develop a functional schedule AND solicit public input BEFORE construction proceeds. This should have been how the project started - isn't the schedule the most important aspect of a station? What does accessibility matter if the station has no service? Regardless of how we got to where we are now, this solution should be implemented in conjunction with any solution above or any other possible solution. Who knows, maybe they can come up with something that works... but I seriously doubt it.
I will be raising this issue in future meetings of the Worcester Working Group and I'll engage with local politicians and stakeholders. It's never too early to reach out to your legislators to sound off on this issue. I'll keep you updated with what I learn.
Congratulations to the Patriots! I wasn't convinced they would win until they got the second 2-point conversion. Then I knew it was destiny.
The MBTA & Keolis have announced a service change for Tuesday 2/7/2017 for the Framingham-Worcester line: the morning HeartToHub will operate as a full length local train making all station stops. The theory is that by having this train make extra stops, it increases the overall passenger capacity for all of the intermediate stops - potentially helping alleviate overcrowding on ALL trains.
I've attempted to make an UNOFFICIAL very rough ESTIMATE of the schedule for it here:
P552 - AM Heart To Hub:
Worcester: 8:05 AM
Grafton: 8:18 AM
Westboro: 8:22 AM
Southboro: 8:31 AM
Ashland: 8:35 AM
Framingham: 8:46 AM
West Natick: 8:51 AM
Natick Center: 8:56 AM
Wellesley Square: 9:01 AM
Wellesley Hills: 9:05 AM
Wellesley Farms: 9:08 AM
Auburndale: 9:13 AM
West Newton: 9:16 AM
Newtonville: 9:19 AM
Yawkey: 9:29 AM
Back Bay: 9:34 AM
South Station: 9:40 AM
This schedule assumes that P586 departs Framingham on-time and operates ahead of P552. The above schedule is NOT official and is probably VERY OPTIMISTIC - but it should be generally close to what might happen.
The official page with updates is here (http://www.mbta.com/events) although as of early Monday afternoon it didn't yet have details about the conversion of the HeartToHub to a local train.
Expect overcrowding on many trains - try to move your AM commute / trip earlier since earlier trains (pre-7 AM departures) will have more capacity on them. The weather certainly won't help with on-time performance...but I would rather be having some delayed trains if it means we're celebrating another Super Bowl win!
As discussed in this blog post below, the next phases in the roll out of the new May 22, 2017 schedule are the public hearings and the associated public comment period.
The MBTA has published the public hearing schedule:
Tuesday, January 31
6:30 p.m. - 8:30 p.m.
Newton City Hall
War Memorial Auditorium
1000 Commonwealth Avenue – Newton
Wednesday, February 1
6:30 p.m. - 8:30 p.m.
2 Washington Square – Worcester (Parking is available in the Union Station Garage at 225 Franklin St.)
Monday, February 6
6:30 p.m. - 8:30 p.m.
Natick Town Hall
13 East Central Street – Natick
The official MBTA web page with details on the public hearings is here.
The official 'reveal' of the new schedule is on this MBTA web page. There were no major changes to the draft schedule that I published back on December 16, EXCEPT that inbound train P520 was moved 5 minutes later. There were some other minor time tweaks, but no changes at all to any rush hour trains. I updated my PDF of the proposed schedule so that it includes all the changes.
Please make sure to fill out this survey and add your comments. That is the only way to get your written comments on the "official" record - tweets and comments written here will NOT become part of the official record.
But the best way to get your comments on the record will be to come to one of the public hearings! I hope to be able to get to some of them, so I'm looking forward to meeting more of my blog and Twitter followers.
Finally, the MBTA published a helpful comparison table to compare the existing schedule to the proposed schedule for service from each station to and from selected destinations.
[Edited 1/18/17: Images and PDF updated with most recent draft schedule. The only significant change is P520 was moved 5 minutes later. A few other trains had minor tweaks to some times. There were NO changes to any times for AM or PM rush hour trains between the 12/16/16 and 1/15/17 versions.]
On December 1, 2016, this blog post provided a preview of the draft morning Framingham-Worcester weekday schedule to be implemented on May 22, 2017. The entire weekday schedule is now available. The schedule is shown in the three images below - the entire schedule plus 'zoomed in' images of the AM and PM rush hour service. Click on any of the images below for a PDF of the entire schedule.
There have been a couple of changes to the morning schedule between December 1 and now, which are outlined below. My entire commentary from the previous blog post is also provided here, just for continuity.
Comments / thoughts:
1) Caveats: This is the official draft proposed schedule, which the MBTA and Keolis have agreed can be implemented in May 2017. But that also comes with a bunch of caveats:
a) This is the final draft developed as a result of the Worcester Line Working Group meetings. There might be some minor tweaks to some times as the MBTA completes simulations of the 'runs' over the next few weeks but the concepts shouldn't change - there won't be any stops added to or taken away from particular trains, locals and expresses will stay where they are, and trains won't be shuffled around before the final draft schedule is published by the MBTA.
b) The intent is that in a few weeks, this draft / proposed new schedule will be formally published by the MBTA. That will be followed by a formal public comment process. More on that below.
c) This is only a PROPOSED schedule, and the public comment process may result in changes being made before it is implemented in May 2017. There is no way to predict if those changes could be substantial or not.
2) Publication of this draft: All members of the Working Group agreed that this draft schedule could be published here on my blog to offer everyone a 'sneak preview' before it gets published officially by the MBTA and the public comment period starts. So if you see something you hate, don't get too fired up yet - remember, this isn't cast in stone.
Enough with the disclaimers...
Morning Rush Hour Inbound Service:
3) Our Petition worked! The AM service is essentially my proposed schedule from the petition that many of you signed, with two major changes:
a) the Heart To Hub train is NOT eliminated; and
b) P510 stops at Natick Center & 3 Wellesley stations.
3a) Heart to Hub: How did they keep the Heart to Hub? Easy - we're getting another train set. Currently, the AM peak service is serviced by eight train sets. With the addition of another set to increase the number of sets assigned to nine, the Heart To Hub can stay on the schedule. Or another way to look at it is new train P586 (7:14 AM departure from Framingham) can be added to the schedule with the ninth set.
As you can see, there will now be 12 inbound trips between the start of service and 10:00 AM arrival at South Station - 13 if you include the P512 'shoulder service.' It's relatively easy to see how 9 train sets can make 12 or 13 trips - the sets from the first four trips go back outbound and then come back inbound. Those turns would theoretically be:
P500 goes back out to Framingham and is 're-used' as new P586.
P502 goes back out to Framingham and is 're-used' as new P588.
P504 goes back out to Framingham and is 're-used' as new P590.
(P582 goes back out to Worcester and is 're-used' as P512, but P512 is outside of rush hour, so it doesn't count towards the 12 inbound peak trips, but it is the fourth of the four re-used sets to get to 13 trips from 9 sets.)
As previously noted on other blog posts, 4 train sets start the day in Worcester after spending the night in the layover yard. The other 5 sets come from storage in Boston.
Where do they get a ninth train set? How do they get more equipment when they have been struggling to maintain set sizes recently? The MBTA & Keolis have recently undertaken some relatively drastic steps to increase the availability of locomotives & coaches, and the expectation is those projects will be complete well before May 2017. This includes:
a) Off-site overhaul of 10 existing out-of-service locomotives. The request for proposals for this project was recently published and bids are due in mid-December.
b) Overhaul of additional out-of-service locomotives by Keolis and other local repair facilities.
Technically speaking these projects are a mix of true 'overhauls' and more modest comprehensive repairs or proactive component replacement, but it's easier for me to just say 'overhauls.'
c) Off-site execution of 4-year inspections for 19 coaches which have been out-of-service awaiting those inspections. This work will be done in Delaware by a railcar maintenance company.
When complete, these tasks will result in growth of the overall fleet size for the South Side Commuter Rail operations - allowing us to get a ninth set and allowing everyone to benefit from better equipment availability.
3b) P510 at Natick Center & 3 Wellesley stations: My original proposed schedule converted P510 to an express train with similar service as the other express trains - express from West Natick to Yawkey. However, it was pointed out that Natick Center and the 3 Wellesley stations have had inbound service with stops in the 8:20 to 8:30 AM time frame for many years. I went back and checked and it's true - the ~7:30 AM departure from Worcester making all stops to Wellesley Farms has been on the schedule for at least 16 YEARS. So there is something to be said about not removing long-established service that people have built their routine around. From a purely schedule fairness perspective, I'm not thrilled that these 4 stations end up with better service than any other stations, but keeping those 4 stations on the P510 schedule only adds 8 minutes to the duration of P510.
4) Boston Landing: Adding Boston Landing to the local Framingham-originating trains actually allows everyone to get to Boston Landing relatively easily without adding more time to the duration of the express trains. For anyone west of Framingham wishing to go to Boston Landing, they can de-board in Framingham and transfer to the next inbound local train. Each Framingham-originating local departs Framingham only about 8 minutes after the express, so the transfer wait is not onerous. This transfer process actually allows reasonable service for anyone west of Framingham going to any of the stations between Natick Center & Boston Landing.
4a) Boston Landing added to P510: Since 12/1/16, the decision was made to add Boston Landing to P510. This is because the Heart to Hub train prevents a Framingham originating local train from leaving right after P510 comes through Framingham. Passengers on P510 originating west of Framingham would face a 40 minute wait at Framingham to transfer to P590 to get to Boston Landing if P510 didn't stop at Boston Landing.
5) P506 now an express: The major element of my proposal and petition was the conversion of P506 to an express train, with a new local Framingham-originating train right behind it. That's what we wanted and that's what we got! Our hope is that this change moves some passengers from P508 to P506, thereby alleviating some of the overcrowding on P508. So if you are a P508 passenger, start thinking about what you can do to move to P506!
6) Addition of Ashland to P590: For the first time ever a local train will start in Ashland, rather than Framingham. There are a number of reasons we decided to do this:
a) The outbound set of equipment used for this service can continue west to Ashland without needing to depart Boston substantially earlier. In other words, the equipment & personnel are available and it's possible.
b) There are no conflicting train movements in the area which would preclude P590 making a 'turn' in Ashland.
c) Ashland has the largest parking capacity on the line, and both lots never come close to being filled at any point during the day.
d) A later morning inbound trip from Ashland could theoretically align with the addition of Ashland to the outbound evening "Heart to Hub" express train which is later in the evening (see below for details on PM schedule). In other words, perhaps some passengers will be able to take advantage of a slightly shifted workday schedule in Boston (~10 AM to ~7 PM?) by taking P590 in the morning and P552 (the Heart to Hub) in the evening.
e) Adding Ashland to P590 should not impact the crowding on the train, since it has excess capacity.
f) Adding Ashland to P590 does not impact riders from Framingham east since the added stop has not changed departure times from any other station for P590.
7) Renumbering of trains: The MBTA has renumbered the local trains (vs. the current schedule and the draft I published on 12/1/16). The elimination of using "P580" as a train number and starting with "P582" now means that the equipment turns have increasing numbers for each set of equipment. For example, on the current schedule, train P581 comes outbound from Boston and turns in Framingham to become P580. By skipping the number 580 on this new schedule, P581 will now come outbound from Boston and turn at Framingham to become P582, which is much more intuitive - all outbound trains will increment by one digit UP at Framingham for their inbound turn.
Evening Rush Hour Outbound Service:
8) Consolidation of P517 & P589: The current schedule includes express P517 & local P589. The proposed May 2017 schedule combines these trains into one full length local train (which would be called P517) departing South Station at 3:30 PM. Three reasons for this change:
a) Ridership: The ridership for P589 is quite low. The ridership for P517 isn't terrible, but it still isn't incredibly high.
b) Set utilization: By eliminating one set from the early rush hour 'rotation,' a set becomes available to serve as a spare for the rush hour commute. This increases the resiliancy of the equipment fleet for the entire rush hour service schedule.
c) Track utilization: By eliminating one set from going to Framingham and coming back into Boston early in the evening commute, the tracks have less trains running across them, allowing for less congestion and better resiliancy.
9) Addition of Framingham & Ashland to Heart to Hub train: The PM Heart to Hub train departing South Station at 7:35 PM has excess capacity (OK, that's the politically correct way of saying it has low ridership). Adding Framingham and Ashland does not substantially increase the transit time of the trip to Worcester, and may draw additional ridership onto this train. Why Framingham and Ashland?
a) Framingham has the highest ridership of any station on the line. Also, due to the signals, grade crossings, and physical characteristics of the line, trains already operate at reduced speed through Framingham, so adding the station stop is not eliminating a high speed segment of the trip. Finally, the construction of the new parking lot on the north side of the station will add over 200 parking spaces to the station area.
b) Ashland (as noted in #6 above) has excess parking capacity. Also, matching it up with the AM P590 trip offers a possible 'alternative' workday schedule for some commuters / potential passengers for the Heart to Hub. Finally, a new Transit Oriented Development project near the Ashland station may result in increased ridership to/from Ashland station - some of whom could potentially utilize the evening Heart to Hub train.
10) Standardized headways: From 4:20 PM to 5:50 PM, the schedule now operates on a fixed 40 minute headway for each local train and each express train, with 10 minutes between an express and a local train. In other words, with the exception of West Natick and Framingham (which are serviced by both express and local trains), a train departs for each destination west of Boston every 40 minutes. For example, Worcester bound passengers can depart South Station at 4:20 PM, 5:00 PM, and 5:40 PM. Passengers bound for Natick Center (as an example) can depart South Station at 4:30 PM, 5:10 PM, and 5:50 PM.
These standardized headways are the best compromise between frequent service and the spacing of trains to avoid conflicts. As many of you are aware, the 25 minute spacing between the current 5:15 PM local (P593) and the 5:40 express (P523) is just too close - P523 frequently 'catches up' to P593 before Framingham. Increasing that spacing to 30 minutes (along with changing the equipment rotation) should solve that problem.
Many of you may recall the 20 minute headway between local trains on previous schedules - a 5:15 PM local departure from South Station was followed by a 5:35 PM local departure from South Station. As previously discussed on this blog, that was only possible with a 'leapfrog' move - the 5:30 PM express train went AROUND the 5:15 PM local train. While that was a nifty move, it was perilous for the schedule - if anything went wrong and there were any delays, the delays could rapidly become widespread and substantial. We just don't have good enough infrastructure to keep a leapfrog move on the schedule. A (quite long) third passing track is really needed to be able to reliably include a leapfrog move on our schedule.
We spent a fair amount of time exploring the options for evening headways and looking at the pros and cons of various options. The Worcester Working Group feels that the 40 minute headway system for the height of the rush hour is the most reliable schedule at the most frequent service interval possible. We spent time learning about the infrastructure constraints that if solved would allow for more frequent service. Hopefully there will be announcements in the near future about some infrastructure projects that will benefit us.
11) The 5:50 PM train no longer operates to Worcester: The current P525 departs South Station at 5:50 PM and operates all the way to Worcester. But P523 departs South Station at 5:40 PM and operates all the way to Worcester, so having the Ashland to Worcester stops on P525 is somewhat duplicative. Removing the Ashland to Worcester segment from P525 improves congestion at Worcester, which has been a problem in the ~7 PM to ~9 PM window on the current schedule. Service to stations from Ashland to Worcester is not substantially or realistically reduced.
12) Only minor changes to trains between 6 PM and 7 PM: The important factor for this later half of the evening rush hour is the constraint at Worcester: trains should arrive at Worcester no more frequently than every 30 minutes. This has been discussed on previous blog posts, and is the result of four factors:
a) The dead end track at Worcester station (MBTA doesn't have permission to go west of the station);
b) The single platform at Worcester station;
c) The arrangement of interlockings on the approach into Worcester station; and
d) The positioning of the layover yard EAST of Worcester station.
The current schedule includes a number of trains which arrive at Worcester less than 30 minutes after the preceding train. While that can sometimes work, it is only possible if everything happens precisely on time. Increasing the interval to 30 minutes for each arrival increases the resiliancy and reliability of the entire schedule.
So, how does this affect the 6 PM to 7 PM window of Boston departures? With the Heart to Hub arriving in Worcester at 8:45 PM, the latest arrival of the previous train is ~8:15 PM. This pushes it back to a 6:45 PM departure from South Station, which also squeezes the departure of the previous train between that and the 5:50 PM local. Taking a closer look at these timings, you can see that it is impossible to continue the 40 minute express / local pattern after 5:50 PM - everything would get jammed up and there would be a long interval of no service to particular stations. Therefore keeping the ~6 PM to ~7:30 PM Boston departures similar to what they are now offers reasonable service to the most stations.
The current 6:20 PM "limited" train departing South Station is moved earlier 5 minutes, but it can't become an express and skip the Wellesleys and Natick Center. If it did, then those stations would not have service for an entire hour between the 5:50 PM and 6:45 PM Boston departures.
13) Boston Landing: As with the morning commute, only local Framingham terminating trains will stop at Boston Landing. This does present a minor problem for Boston Landing passengers bound for destinations west of Framingham, since the express-local sequence at Framingham is not conducive to a transfer there (transfer times of ~20 minutes). There are a couple of solutions / mitigations for this issue:
13a) Boston Landing passengers can get to P521: P521, the 5 PM South Station express to Worcester, is the most popular evening train. Although it doesn't stop at Boston Landing, those passengers can take INBOUND P520 from Boston Landing at 4:59 PM and get off at Yawkey at 5:04 PM. They can then board outbound P521 at Yawkey at 5:11 PM.
13b) Boston Landing added to P525: Without a similar inbound-outbound trick to reasonably access P523, the next train servicing stations west of Framingham is P525. A stop at Boston Landing has been added to that train since there is also no corresponding inbound-outbound trick for Boston Landing passengers to access P525. Boston Landing passengers can use P593 to Framingham and transfer to P523, but that requires a ~18 minute wait at Framingham.
14) Public Comment Process: The current conceptual plan for implementing the new schedule is:
a) Worcester Line Working Group completes finalizing entire new weekday draft proposed schedule by Christmas (COMPLETE).
b) Early January: MBTA formally publishes draft schedule and begins 6 week public comment period. Public Hearings will be held, although the venues and schedule haven't been worked out yet.
c) Mid-February: End of 6 week public comment period.
d) Mid-February to Mid-March: MBTA, Keolis, and Worcester Line Working Group reconvenes to review public comments & finalize new schedule.
e) Mid-March to early April: Keolis builds final equipment & crew rotations for new schedule.
f) Early April to mid-May: Crews choose new assignments (this always happens for 6-8 weeks prior to every new schedule being published in May or November).
g) May 22, 2017: New schedule implemented.
ALL OF THAT IS DRAFT AND SUBJECT TO MAJOR REVISION
This is a pretty exciting example of activism and everyone coming together behind a petition with a government agency responding and making the requested change. Our voices did matter and we're on the way to a better schedule!
Even though we're not done, it's important to note that the Working Group included many people who all contributed and engaged very constructively. It has been a very collaborative process and I'm glad I got to be a part of it. Besides me, the Working Group includes Lt. Governor Karyn Polito, Rep. Alice Peisch (D - Wellesley), Rep. Carolyn Dykema (D - Southboro, Westboro), Rep. James O'Day (D - Worcester), Stuart Loosemore (Worcester Chamber of Commerce), Jessica Strunkin (495/Metrowest Partnership), Brian Shortsleeve (MBTA General Manager), Jody Ray (MBTA Assistant General Manager in charge of Commuter Rail Railroad Operations), two Deputy Directors of MBTA Railroad Operations, David Scorey (Keolis CEO/GM), the Keolis Manager of Operations Planning, and another general public member / commuter from Wellesley.
Stay tuned to this blog and I'll publish updates on the Public Comment process as soon as it is finalized. Subscribe to get e-mail updates when the blog is updated.
There have been a number of articles recently about the poor on-time performance of the Framingham-Worcester line.
Boston Herald article
Boston Globe article
One of our Twitter friends, Meghan, even had her letter published in the Boston Globe.
The longest article was in the Telegram & Gazette, and it does the best job at explaining the difference between the "Worcester Working Group" and the "Worcester Line Rapid Action Group" (and that's not only because the article quoted me!). More on the Rapid Action Group below.
With regards to the delays, the data doesn't lie, and kudos to the MBTA for being transparent with the on-time performance data. Take a look for yourself at:
In general, delays are unacceptable. In reality, it's a bit more complicated than that. I'll offer my thoughts on the on-time performance and some potential reasons for the delays. But first, I'll offer one criticism of the data being provided by the MBTA and reported in those articles: the data provides no information about the MAGNITUDE of the delay. A 5 minute delay is the same as a 90 minute delay for the purposes of reporting how many trains were on time. I think that is an important fact - I'm not happy when my train is 6 minutes late, but it also doesn't ruin my whole day. But a 30 or 60 minute delay has the potential to really screw up my work or evening schedule.
I'm not trying to minimize any delays in the recent past, but my own completely subjective feelings are that:
1) There have been too many epic meltdowns recently (see blog posts about 10/24, 11/4, and 11/8. I didn't even write blog posts about some of the severe equipment shortages around 11/4 and the week of 11/7).
2) Those epic meltdowns result in some huge delays for many of us.
3) But overall, those epic meltdowns don't represent 40% of the trips I'm taking. Maybe 10%? That would be one per week on average. Maybe it's 2 per week? The other 2-3 delayed trips per week (to get to ~40% delays) are probably less than 10 minutes. Again, that's just my experience, and feel free to open the debate in the comments below...
Why are the delays happening at all? I think there are a number of factors to blame.
1) Bad luck. A fatality in Wellesley relatively early in the rush hour one evening not too long ago delayed some trains by hours. The train involved in the buffer strike at South Station being a Worcester train was bad luck (the incident itself wasn't bad luck - I'm just saying it was bad luck it was one of OUR trains).
2) The slippery rail season seems to have been bad this year. With only one 'wash train' for all the lines on the South Side, combating slippery rail is difficult. And it isn't clear to me that the wash train operated as much as it could, either because of equipment issues or other factors. This was what I was alluding to in my quote in the T&G article - slippery rail is bad luck, but was everything possible being done with the wash train to "make our own luck" to mitigate the related delays?
3) Equipment availability & mechanical issues. It is no secret to anyone that the last few months have seen a periodic shortage of coaches, locomotives, and/or entire train sets. The extensive delays and cancellations around 11/4/2016 could only have been caused by equipment issues.
What's the good news? Believe it or not, Keolis actually cares about these delays. Perhaps it's our advocacy, perhaps it's the political pressure, or perhaps it's the fines for late trains, or perhaps it's them just doing their jobs, but they're trying to fix it. They have created the Worcester Line Rapid Action Group to work on short-term solutions for the delays.
David Scorey, the Keolis General Manager (and a fellow Framingham-Worcester commuter) recently spoke about the Worcester Line Rapid Action Group. Watch this video - it has some nifty footage from some of the recent construction projects.
The Rapid Action Group is an internal Keolis-only group, focused on creative solutions to reducing the delays for our trains. Some of the solutions they have implemented include:
1) Adding conductors / crew members to the trains with the highest ridership. P508 actually went from 4 assigned conductors to 5 on the new crew rotation which was implemented on 11/21/2016. Other morning trains have had extra staff added on a temporary basis. Adding staff allows for more doors to be opened, speeding up passenger loading. Dwell time (the time a train is stopped in a station), especially at our stations which don't have full height platforms, can be a significant source of delays. Just 1 minute longer at five stations is 5 minutes - delays add up quick.
2) This next one is a little technical, but it's a great idea and it has been very effective. The basic idea is to have inbound express trains switch tracks well before Boston Landing - therefore avoiding a slow switch there.
"CP 4" is the interlocking / switch just west of Boston Landing. This is the switch where the two tracks from Worcester become one to get past the Boston Landing construction. Switches are not all the same - there are different switches with different geometries that allow for different speeds through them. A switch with a shallower angle off the straight track allows for higher speeds through the switch. Unfortunately, the switch at CP 4 is only rated for 15 mph.
As many of you know, rush hour trains (both inbound and outbound) usually use track 2. There's a long story behind that, but here at CP 4 the basic story is that the Newton stations have platforms only on track 2 - so trains making the Newton stops have to be on track 2 (there are no crossovers between CP 4 and the Newton stations). Trains not making the Newton stops don't have to be on track 2.
As shown on the diagram, going to or from track 2 to the single "shoo fly" track requires going through the 15 mph CP 4 switch. The 15 mph speed limit through CP 4 is effectively enforced by the fact that the 'best' signal a train can get going to or from track 2 is a "slow clear" ("Proceed at Slow Speed [15 mph] until entire train clears all interlocking or spring switches, then proceed at Normal Speed [maximum authorized speed for the track]").
Passing through CP 4 to or from track 1 avoids the effective 15 mph speed limit. Trains on that routing remain on a straight path. There is a 30 mph speed limit adjacent to the Boston Landing construction zone, so trains still have to slow down there, but the slow down is not as dramatic.
The Rapid Action Group came up with the idea to route morning express trains onto track 1 through CP 4. Since they don't need to stop at the Newton stations, they can switch to track 1 at CP 11 (the "Weston Switch"). This is the closest interlocking that allows for changing tracks west of CP 4. This change can save ~2 minutes for those trains, and every minute matters. This routing has been used for trains P502, P504, P508, and P510 for the last couple of weeks, and it appears to have helped with on-time performance. Believe it or not, P508 was actually early on Tuesday 12/6/2016.
3) Another change that Keolis has made involves the sequencing of the tie replacement project. In addition to the limitations described in this blog post, they are also not allowing the construction work to occupy track 2 until P512 passes through. As many of you recall, the tie replacement project in May and June did not have this restriction placed on it, and P512 was frequently substantially delayed due to the construction work. P512 is not technically a rush hour train, but it does have relatively high ridership, so this change helps those passengers. Mid-day trains after P512 are also less affected by the construction work, since the schedule through the mid-day has been structured to allow for areas of single track operation without dramatically affecting the schedule.
I'm sure there will be those of you who claim that the Rapid Action Group is just making changes that should be normal operational corrections, and there is an element of truth to that. But regardless of how it's happening, they are trying to make changes to improve on-time performance.
There are probably a combination of factors at play, but on-time performance has improved over the past few weeks. Those factors include:
1) End of slippery rail season;
2) Better equipment cycles with the new 11/21/2016 schedule (even though they didn't change our schedule, the changes to other lines allows for better equipment rotations);
3) Better equipment availability (apparently); and
4) the work of the Rapid Action Group.
Specifically, here is the on-time performance for peak trains for the following weeks:
11/21/16: 65% (Thanksgiving week)
We're still a long way from 90%, but at least the numbers are going in the right direction.
DO NOT READ THIS BLOG POST.
IT HAS BEEN SUPERSEDED BY THE RELEASE OF THE ENTIRE DRAFT SCHEDULE.
This blog post has not been removed / deleted, since many of the comments are still applicable. But all of the information in this blog post is now repeated in the blog post detailing the entire new draft schedule.
CLICK HERE TO GO TO THE CURRENT BLOG POST FOR THE ENTIRE 5/22/2017 SCHEDULE.
Original blog post for historical reference only:
As many of you know, the Worcester Line Working Group was formed in late September. We've been meeting regularly since then with a focus on developing a new schedule for May 2017. I'm pleased to announce we're ready to release the draft proposed new schedule for the morning rush hour inbound commute. Click on the image below for a PDF.
Comments / thoughts:
1) Caveats: This is the official draft proposed schedule, which the MBTA and Keolis have agreed can be implemented in May 2017. But that also comes with a bunch of caveats:
a) This is the initial draft, so there might be some minor tweaks to some times as the final draft is finalized over the next few weeks. But the concepts shouldn't change - there won't be any stops added to or taken away from particular trains, locals and expresses will stay where they are, and trains won't be shuffled around before the draft schedule is published.
b) The intent is that in a few weeks, the entire draft / proposed new schedule for the entire weekday service will be formally published by the MBTA. That will be followed by a formal public comment process. More on that below.
c) This is only a PROPOSED schedule, and the public comment process may result in changes being made before it is implemented in May 2017. There is no way to predict if those changes could be substantial or not.
2) Other parts of total schedule: The Working Group is continuing to work on all other aspects of the schedule, including the PM rush hour outbound commute, the AM & PM "reverse commute," and the off-peak schedules. We're not ready to publish those yet - still some tweaking, negotiating, arguing, arm twisting, horse trading, and disagreeing to push through. Ok, actually it's just tweaking, but it sounds better to imagine us yelling and screaming at each other, right? It hasn't been like that at all...it's been very informal, open, and productive.
3) Publication of this draft: All members of the Working Group agreed that the draft schedule could be published here on my blog to offer everyone a 'sneak preview' before it gets published officially by the MBTA and the public comment period starts. So if you see something you hate, don't get too fired up yet - remember, this isn't cast in stone.
Enough with the disclaimers... more after the "read more" break:
I'll write another blog post about the overall poor on-time performance of the Framingham-Worcester line (as highlighted in some recent news articles). But for today, I wanted to focus on the tie replacement project.
This fall's tie replacement project is on track 2 between Boston and Southboro. Track 2 is the track used for both inbound and outbound rush hour trains between Boston and Framingham.
The tie replacement project is necessary maintenance and actually decreases the risk that restrictive speed limits will be imposed due to track and railbed condition. Every 6 months, an inspection train checks the condition of the tracks. This sophisticated "geometry train" uses a number of sensors to verify that the track and railbed are safe for the operation of the railroad at the highest possible speeds (daily inspections check for visual and other defects). If the geometry train detects a fault, there is a speed restriction imposed on that section of track until the defect can be repaired. Without regular replacement of ties and resurfacing of the railbed, the risk increases that the geometry train will find defects that will result in temporary speed restrictions.
As many of you recall, the previous major tie replacement project in May-June on track 1 resulted in significant delays for many trains. The MBTA & Keolis changed the way the current tie replacement project is being executed for track 2 in order to minimize delays.
The disturbance of the railbed from any construction work requires temporary speed restrictions after the work has been done. But how the work is sequenced can affect the length (distance) of those speed limits - and by minimizing the length of the speed limits, the overall impact to on-time performance can be mitigated.
For this fall's project, the productivity of the construction has been intentionally slowed down in order to minimize the length of the temporary speed restriction. In addition, a specialized piece of equipment - a track stabilizer - has been added to the track surfacing crew so that the speed restrictions can be further minimized.
For this tie replacement project, the sequence is:
1) close track
2) tie replacement crew / work
3) track surfacing crew / work
4) open track
5) temporary speed restriction for first / ONE train movement across entire length of work area (both tie replacement area & surface work area).
6) ~24 hour 30 mph temporary speed restriction for the ~1 mile of track that the tie replacement crew disturbed (or actually the disturbed area where the surface crew HASN'T worked yet).
The 'track surfacing' work is the distribution of new ballast (stone) and the tamping and stabilization of the stone. Normally the area where the surfacing crew works requires a temporary speed restriction for 24 hours. Adding the stabilizer to the surfacing crew reduces the temporary speed restriction from 24 hours to one train passage.
In addition to adding the track stabilizer to the surfacing crew, the tie replacement crew's productivity is being intentionally limited so that they do not get too far ahead of the surface crew. Normally, the productivity of the tie replacement crew is HIGHER than the surfacing crew. So over time (multiple days), the tie replacement crew could get way ahead of the surface crew, resulting in a 24 hour temporary speed restriction for their work area and the entire 'gap' between the two work areas.
In other words, for this project, the tie replacement crew is not allowed to go as far as they could if they were unconstrained. This limits the tie replacement crew to probably ~75% of their unconstrained productivity.
An MBTA manager mentioned to me that they can remember an instance when the Old Colony lines had tie replacement done and the temporary speed restriction between the crews got all the way up to 7 miles. This was a long time ago - but that is a long distance to be restricted to 30 mph.
The track stabilizer as a member of the surfacing gang and the constraint on the tie replacement crew were both NOT present during May-June when they worked on track 1. The delays resulting from having multiple miles of temporary speed restrictions were catastrophic to on-time performance for that period - and not only were there many delays, but the delays were much more substantial than the delays being caused by this fall's tie replacement project.
The overall scope of this tie replacement project is from Boston (actually CP 4 on the west side of Boston Landing station) to CP 28 in Southboro (see the map). 33,000 ties are scheduled to be replaced - an average of 1,600 per mile. That's about half the ties (US railroads typically have ~3,000 ties per mile).
As of today, the work has been completed from CP 4 in Boston to just west of Wellesley Farms station. Railbed construction and maintenance is usually halted in the winter when the ground freezes. This project got a slightly later start than was intended - I think it started around the beginning of October. The current plan is to continue with the project until approximately December 16th, with the remainder of the work presumably deferred until the spring.
The limitations on the track work productivity have been successful and have resulted in 24-hour 30 mph speed restrictions of no more than one mile since the project started in October. Many of you have commented on the recent "slow-downs" in Wellesley and Newton - and this is what you have been noticing. It is much more noticeable on an express train that would be operating at 60 mph for that whole stretch - a one mile speed restriction requires slowing down before the train reaches the start of the speed restriction (as opposed to a local train that may never even get up to 60 mph after its nearest station stop).
It is likely that this tie replacement project has had a impact on on-time performance for many trains over the past month. As noted, a one mile speed restriction for an express train may be enough to push it close to 5 minutes late (the threshold for "on-time"). In addition to the impact on rush hour trains, there have been more significant delays for mid-day trains when one track is out-of-service completely and train service is limited to one track only during the construction.
So in summary, we need to give credit where credit is due. I'm not suggesting that ~60% on-time performance is acceptable - but I am suggesting that without these changes to the way the tie replacement project is being executed, the delays related to the tie replacement project would have been much more extensive.
As I have noted in previous blog posts, the 'new' schedule for the Framingham-Worcester line implemented on 11/21 does not have any changes - just the notation of 'severe weather' trains.
The reason that there were no changes implemented on 11/21 were noted in this blog post, but I'll repeat them here since they haven't changed and are still true:
1) Over the next few months there will be major changes to the infrastructure of the line which will require / enable major changes for the May 2017 schedule. These are primarily the introduction of the station stops at Boston Landing and the addition of the second track through Beacon Park. So making major changes now - only to have to make major changes again in May 2017 - is less than ideal.
2) The level of interest, possible changes, and number of stakeholders warrants a more deliberate and public process than just implementing changes decided upon by MBTA planners and management. The "Worcester Line Working Group" is the first step of that process.
The Worcester Line Working Group continues to meet - our sixth meeting is coming up this week - and we're making good progress on a number of initiatives:
1) Developing & reviewing potential alternative schedules for the May 2017 schedule change;
2) Narrowing down those alternates to a handful of potential schedules that could be presented for public comment;
3) Developing a schedule for a public comment period (if needed) to review proposed schedule changes;
4) Hearing about shorter term strategies available to reduce the present-day delays & problems; and
5) Hearing about longer term ideas for achieving better schedules.
Once we get to a point where there is something to share, I'll be publishing it here. Stay tuned and subscribe for updates.
The main point of this blog post is to highlight two thoughts I have regarding the 11/21/16 schedule. I've made these points in previous blog posts and on Twitter, but they are important to note.
1) Severe weather trains: Many people have expressed dismay at the thought that Keolis will implement the 'severe weather' service frequently and without adequate notice. This is simply not true. The rationale behind the 'severe weather' service schedule is to have a pre-planned schedule available in the rare case that service should be curtailed. I can recall MANY times when rush hour trains were almost completely empty on days of very heavy snowfall - these are the days where the 'severe weather' service is designed to be implemented. Concerns of overcrowding are therefore unfounded - if the 'severe weather' service is being implemented, it is likely that most schools are being cancelled and many commuters will not be going to work anyway. The inconvenience of the reduced level of service when the 'severe weather' service is implemented will be far outweighed by the inconvenience of a severe storm. There are, therefore, two goals being achieved: 1) less trains are needed since there will be less ridership; and 2) less trains makes the schedule more reliable by allowing for more resiliency and less strain on the equipment fleet. In addition, having less trains operating in severe weather helps to preserve service when the weather gets BETTER - there is less of a risk of a substantial amount of equipment becoming disabled in the severe weather and preventing normal service from being restored after the storm.
It also should be noted that a version of this 'severe weather' schedule existed during the winter of 2015-2016 and it was NEVER activated. Granted it was a mild winter, but this should reinforce the idea that the 'severe weather' will be just that - SEVERE weather service. Some people have asked for clear and definitive guidelines regarding what the definition of 'severe' weather is - and I think it is correct that the MBTA & Keolis are reluctant to define it. There are too many variables across too large of a geography to expect a simple definition. Furthermore, offering a hard and fast definition will lead to second guessing and complaints about the definition ("Why are you implementing the severe weather service? There isn't 12 inches of snow in MY backyard!").
Finally, the pre-planned 'severe weather' schedule is more of a communication tool than a change in the policy regarding service. There have been many examples in the past where some trains would be cancelled due to severe weather - but without a pre-planned 'severe weather' schedule, the cancellations were communicated on an individual train-by-train basis. This new 'severe weather' schedule offers a consistent & clear communication method for letting us know what trains will be cancelled. We've wanted better communications - this is an example of it.
2) The departure times of certain evening trains at interim stations are not accurate on the new 11/21 schedule. Huh? What does that mean?
In July 2016, the schedules were modified for some trains to add time to the overall duration of those trips. However, instead of changing the entire schedule, the added duration was just added to the final destination. These July changes are maintained in the 11/21 schedule. An example illustrates this fact:
5/23/16 schedule: P523 departs Boston 5:40 PM, departs Grafton 6:46 PM, arrives Worcester 7:00 PM.
July 2016 & 11/21/16 schedule: P523 departs Boston 5:40 PM, departs Grafton 6:46 PM, arrives Worcester 7:10 PM.
There is nothing that has changed to make the trip from Grafton to Worcester take 10 more minutes. The modification is that the OVERALL time that P523 takes to get from Boston to Worcester is now 10 minutes longer than it was on the 5/23/16 schedule. This longer duration represents an acknowledgment that the modeling used to develop the 5/23 schedules was overly aggressive and optimistic. The 7:10 PM time is a more realistic arrival time at Worcester given the actual performance of the train since 5/23.
When the July modifications were implemented, the MBTA did NOT remodel the entire trip of P523 (and other trains affected by the changes in duration) to change the times at the 'interim' stations between the originating station and the train's final destination. On-time performance is only officially measured at the final destination, so the times at interim stations are not needed to measure on-time performance.
Another fact needed to understand this topic is that the time noted on any train schedule for any station (except the final destination) is the DEPARTURE time of a train. Trains are not supposed to depart from any station until the time shown so that passengers can get to the station and board the train before it departs. This is also why some times at some station stops have an "L" on the schedule for "trains may leave ahead of schedule."
Back to our example - Grafton is still shown as a 6:46 PM departure time on the 11/21 schedule. This is no longer accurate - the departure time from Grafton will be somewhere between 6:46 PM and 6:56 PM (and probably closer to 6:56 PM). But passengers arriving at Grafton for the 6:46 PM departure time will never be too late for the train - it will never leave earlier than that time. This is the same situation for all stations stops for P523 - the times for all station departures between Back Bay and Grafton are not accurate.
From a railroad operations point-of-view, having passengers arrive early to board a train that may be departing a few minutes later than what is shown on the schedule is conceptually acceptable, since it won't result in passengers being left behind. This is part of the reason the times shown on the revised 11/21/16 schedule have not changed - the trains now have the flexibility to leave from the interim stations as early as the time shown on the schedule, even if they probably won't.
The important takeaway for commuters is that it is now simply impossible to know precisely WHEN a train (that had its duration modified in July) is actually supposed to ARRIVE or DEPART from station stops between the originating station and the train's final destination. The trains affected by this (along with the overall duration change) are:
P522 (+10 minutes)
P588 (+10 minutes)
P526 (+5 minutes)
P517 (+5 minutes)
P519 (+7 minutes)
P521 (+7 minutes)
P593 (+7 minutes)
P523 (+10 minutes)
P525 (+5 minutes)
Simply put, if you are on one of the trains listed above and are getting off at a station BEFORE the final destination, your train is not officially late until the amount of time listed above is ADDED to the time shown in the schedule.
For example, the schedule for P521 shows West Natick at 5:37 PM. P521 may depart from West Natick as late as 5:44 PM and still be operating "on time."
Note that this issue also affects the predictions made by 'apps.' The real-time data feed from the MBTA does not include any specific information about 'lateness.' The data feed includes information about the predicted departure time for each train at each station, but the schedule that the apps use is the same schedule that is printed for us. Any 'lateness' displayed by an app or website is a comparison between those two values - by the actual app.
For our example of P521 at West Natick, an app will be comparing the predicted departure time in the feed with 5:37 PM as printed in the schedule. If the predicted departure time in the data feed is 5:42 PM, the app may report the train as 5 minutes late, regardless of the fact that the train may actually be 'on-time' with regards to overall on-time performance.
This is a confusing and highly technical issue. Passengers depend on schedules for planning their lives, and having the trains with the majority of commuters during the evening commute performing to an unpublished schedule for most of their station stops is inconvenient and misleading to the customers. At the Worcester Working Group meetings, I have advocated for the remodeling of the entire schedules for those trains and the publication of revised schedules that show accurate times for the interim station stops. In fairness to the MBTA & Keolis, their attention has been on developing new schedules for the May 2017 schedule change - and each alternative we ask for at the Working Group meetings requires a fair bit of modeling and planning to be able to be discussed. But I do believe that fixing the 11/21/16 schedule is important - we're going to have it for 6 months, and it will be frustrating to everyone when it appears trains are continually late due to those inaccurate interim station stop times.
Another (rather late) post-mortem of a really bad morning commute.
At about 6:10 AM, outbound P503 (with locomotive 2009) died between Wellesley Farms & Wellesley Hills. As is usual, this train was on track 1 at this time. After a number of attempts to repair it and get it moving, it was determined that it was beyond the capabilities of the crew to rectify the failure. I never did hear what the actual problem was.
A decision was made to use inbound train P504 as the 'rescue set' - this was partly due to the lack of a 'rescue set' in Boston. The normal operations plan for the south side includes one standby ("protect") train set that can be used in these circumstances. For whatever reason, that protect set was unavailable.
P504 was already past Framingham on track 2 when the decision was made to use it as the 'rescue' set. There are no switches between Framingham and Wellesley, so P504 had to continue east past the broken down set in Wellesley, through the "Weston Switch" (CP 11) and then back west on track 1 to the dead P503. Once at the broken down P503, the crew requested "protection" on track 2 while they worked to hook up the trains. "Protection" means that the dispatcher stops trains on the adjacent track to avoid striking the crew walking around the trains. The dispatcher accomplished this "protection" by holding P506 in Framingham until the P503/P504 crew had completed the hook-up. P506 was delayed about 10 minutes in Framingham.
Timeline, for those of you keeping score:
P504 stopped eastbound past Weston Switch at 6:53 AM
P504 moving west on track 1 at 6:59 AM
P504 stopped and ready to hook-up to P503 at 7:08 AM
P504 hitched to P503 at 7:11 AM
P504-P503 double draft moving east at 7:20 AM
While the situation unfolded in Wellesley, the equipment that was designated to operate as P582 was sent to Worcester to operate as P510, since P510 is usually the 'turn' of the P503 equipment. P582 was cancelled.
The 20 passengers who were originally on P503 wanting to go west were now on an eastbound train (the P503-P504 double draft). So the double draft stopped in Wellesley Farms and discharged those 20 passengers. But with the cancellation of P582, the passengers waiting for P582 were directed to cross over from track 2 to track 1 (using the Glen Road bridge) to get onto the double draft. This further delayed the double draft, which didn't end up leaving Wellesley Farms until 7:31 AM.
The P582 cancellation also left inbound passengers waiting at all three Newton stations. Although I didn't hear the discussion on the scanner (there was at least one phone call from the crew to management during this event), it is probable that there was a concern that leaving all the P582 AND P506 passengers for P506 would have overloaded P506. If P506 became overloaded, P508 (an already notoriously overloaded train) would have had to stop and pick up passengers in the Newtons. So the decision was made to have the P504-P503 equipment double draft make all three Newton stops to pick up the stranded P582 passengers (and by that time, a fair share of P506 passengers also).
Eventually the double draft made it to the Boston stations about one hour late vs. the schedule of P504. Some P582 passengers who made it onto the double draft were about 30 minutes late. P506 arrived at the Boston stations about 12 minutes late, meaning the passengers intending to take P582 who were on P506 were 41 minutes late. P508 arrived in Boston close to on-time.
But that's not the end of our story, because with track 1 blocked by the rescue maneuver, delays cascaded for later trains.
As noted above, the P582 equipment was sent west to replace the dead P503/P510 equipment. P582 is actually the turn of outbound P583, and it usually travels well west of Framingham to make its 'turn' in Ashland. This is just an operational sequencing move to keep the traffic flow moving. In this case, it was somewhat advantageous, since the equipment was already a few miles closer to Worcester when it was ordered to keep going west to Worcester.
Once it arrived in Worcester, it was able to make the turn and depart on time as P510.
But back in Wellesley, we still had 20 passengers at Wellesley Farms wanting to go west but stranded after the death of P503. The next westbound train after P503 is actually not on the schedule - it is a deadhead / non-revenue move (that is called train 6501). That deadhead move starts the day in Worcester as P500, goes into Boston and then turns and goes outbound with no passengers. It usually goes express / direct from South Station to Worcester where it becomes P552, the infamous "Heart to Hub" super express train. It usually passes through Framingham at about 7:10 AM.
Train 6501 and P585 right behind it were stuck on track 1 at CP 11 (the Weston Switch) waiting for the double draft P504-P503 to move east off track 1 and onto track 2. P585 terminates in Framingham to become inbound P584. But since at least some of the stranded P503 passengers wanted to go west of Framingham, train 6501 was converted into a local train to pick up the stranded passengers at Wellesley Farms and any other westbound passengers waiting for P503. But this meant that P585 was stuck following close behind 6501 - further delaying it and therefore delaying P584.
In fact, with the on time performance of P510, it actually made it to Framingham BEFORE P585 could make the turn at Framingham to become P584. This created an out-of-sequence problem, where P510 was now gathering all passengers east of Framingham who were waiting for P510 AND P584.
P510 is also notoriously crowded, and even with the light ridership of election day, P510 rapidly became very crowded after Framingham. P510 usually skips all three Newton stops, but the initial decision was that P510 should make the Newton stops to gather stranded passengers waiting for P584. Text messages indicated this plan. As P510 became overcrowded, this decision was changed - and P510 went express from Wellesley Farms to Yawkey - skipping the Newtons. But this was only announced on the train, causing confusion for everyone - passengers on the train trying to figure out the discrepancy between the announcement and the text / official twitter alerts, and for passengers waiting at the Newtons. There were also numerous reports of problems with passengers on P510 not sitting in middle seats and not moving into the coaches - and both of those problems cause overall reduced train capacity.
P510 arrived at South Station at 9:11 AM - about 14 minutes late, while P584 arrived at South Station at 9:26 AM - about 42 minutes late. Passengers from Framingham to Wellesley Farms who intended to take P584 but got on P510 were 27 minutes late.
6501 eventually arrived in Worcester and was so late that it operated on the P512 full local schedule.
As I've previously mentioned, it is somewhat unfair of me to question the decisions that get made, but in this case, I'm not convinced that having P504 rescue P503 was the best decision. Another strategy could have been to have outbound 6501 come up to P503, hitch onto it, and push it to Worcester. It could have been dragged back to Boston with an inbound trip (delayed P510 for example). The inbound Heart To Hub would have been the cancelled trip in this scenario - which would be appropriate - cancel the trip with the least number of passengers. This strategy could have kept all inbound trips (except P510) somewhat close to schedule. I wonder if they thought that the P504-P503 rescue maneuver would go quicker?
As many of you are aware, the MBTA Rail Tracker app - created by a fellow Framingham-Worcester rider - has a feature that allows for passengers to share 'reports' (or actually any comments) regarding their trains. The 2 pages of comments from Tuesday morning are here - and they are presented unedited, so warn small children about the coarse language. The comments do provide some good insight into the frustrations that we all feel regarding the rather frequent recent delays. Some of the more interesting comments include:
Train 506, 7:19 am: So annoying. Late yet again. When will someone do something about this?
Train 504, 7:50 am: The conduct lady is being nice, let's try to not give her too much shit today.
Train 504, 7:51 am: they should change the name of this app to "Therapist for those who love to complain"
Train 510; 8:37 am: "For the last time, this train will be stopping at Wellesley Hills, Wellesley Farms, Yawkey, Back Bay, and South Station. Period."--grumpy conductor over intercom on p510
Hopefully service will improve and we won't have too many more mornings with 2 pages of irate comments.
<edited 11/8/2016 to incorporate info from comments>
<edited 11/16/2016 to incorporate news about why the train was damaged>
It's always something on the Framingham-Worcester line. The morning of Friday 11/4 brought a new reason for delays and problems - namely, a damaged plow.
P506 (the 6:30 AM local from Worcester) struck something <edit: see the new information at the end of this post for what really happened> at the Cherry Street and Main Street grade crossings in downtown Ashland at about 7:00 AM. There were reports of a loud bang and sparks flying. Upon arrival at Framingham, an inspection of the front of the control coach revealed damage to the plow on the front of the train. Rider photos:
Due to the damage, the decision was made to remove this original P506 equipment from service. All passengers were unloaded at Framingham. The damaged train set was moved west back onto the "4th iron" siding next to the Framingham Nevins Freight Yard.
Here's some photos of everyone waiting on track 1 at Framingham.
The normal operations plan for South Side commuter rail includes one 'standby' extra train. This train is usually kept in Boston with an assigned crew and is ready to be deployed in the case of a breakdown or other problem. It is known as the "protect" set. Since it is usually sitting in Boston, it is not always immediately helpful if there is a problem away from Boston.
However, due to the expected higher ridership on the Framingham-Worcester line as a result of the Mass Pike toll demolition project, the extra protect set had been sent to Framingham every morning since Monday 10/31. It was kept on standby at Framingham for use in case of severe overcrowding on any of the usual train trips. As many of you know, the higher ridership never materialized. Luckily for us, they kept the extra set going to Framingham each weekday morning.
Just as P506 was being cancelled on Friday morning, the extra train set was approaching Framingham from Boston. The train came through Framingham station and pulled west of the station so that the crew could 'switch ends' for the inbound trip. This extra set was much smaller than the usual 6 doubles of P506 - it had five coaches: 4 flats and 1 double.
After loading all of the stranded passengers plus some Framingham passengers intending to take P508, the 'new' P506 departed Framingham at 7:37 AM - only 26 minutes behind schedule. Immediately upon departing, some assistant conductors reported that their coaches were filled to capacity. This information was relayed to the dispatcher.
At Natick Center, the 'new' P506 was given instructions to proceed express to Yawkey, primarily due to the (perceived) overcrowding on it. P506 waited in Natick Center for a longer period of time than normal to be sure that the information was communicated and interzone passengers were able to get off at Natick Center. I was on the double on P506 and it actually wasn't SRO. The 'new' P506 may have been able to make one additional stop at Wellesley Square to squeeze on some more passengers, but it is hard to fairly armchair quarterback these decisions.
P508 was converted to a local and made all local stops from Framingham through the Wellesleys and the Newtons. Due to the overcrowding on P508, some passengers were denied boarding at one of the Newton stops (and maybe more than one). Those passengers would have been displaced to P584 which was relatively close behind P508. But those Newton passengers (originally intending to board P506 but denied boarding onto P508) would have ended up being over one hour late to Boston.
The final answers:
Replacement P506 arrived South Station at 8:18 - 11 minutes late (for passengers able to board at Natick Center or before). Passengers intending to be on P506 from any of the Wellesleys or Newtons were at least 50 minutes late.
P508 arrived South Station at 8:57 AM - 37 minutes late.
1) The pre-positioning of the extra set in Framingham clearly saved the day. Had that equipment not been near Framingham, the scenario would have been much more dire. It's also possible that the existence of the extra set influenced the decision to remove the original P506 set from service. In other words, if the extra set hadn't have been near Framingham, perhaps they would have tried to keep the original P506 set in service (maybe at reduced speed?).
The extra set at Framingham raises the larger operational issue that having a 'protect set' permanently stationed in Framingham during the AM commute would be a wonderful way to offer schedule and overcrowding resiliency. There are frequent examples when an extra set at Framingham could have mitigated serious meltdowns for the AM commute. But without enough equipment on the system, this remains an item for our wish list. The infrastructure is there - the "4th iron" siding is a perfect spot to keep a train on standby.
2) Friday ridership is MUCH lower than any other weekday. Had this been any other weekday, train capacity would have been a huge problem, with P506 passengers probably being deferred all the way to P584.
Overall, due to these first two items, we got very lucky. Had there been no extra set or had it been a regular weekday, the delays for trains P506, P508, P584, and maybe even P510 could have been more extensive & more severe. This is not to minimize the delay that P508 passengers and P506 passengers from the Wellesleys and Newtons had - I am simply saying that it could have been much worse.
3) The decision to send P506 express from Natick Center was a good decision and a creative solution to try and mitigate the delays for some passengers. P508 was already stuck close behind P506, so having P506 pick up passengers just ahead of P508 would have prevented P508 from getting to Boston any faster. Could P506 picked up at least some passengers from Wellesley Square? Perhaps, but it's hard to second guess decisions that are made in real-time with limited information.
4) Communications were a mixed bag. Some passengers reported adequate information, while other passengers reported no communications. Those situations are very difficult to manage with lots of moving parts. I was able to hear decisions being communicated to train crews as the dispatcher said them - meaning I was probably aware of some operational decisions before the @MBTA_CR customer service got the info. And that's not a dig at the customer service system - it is simply an example of how information has to move through many steps to get out to passengers via official channels. There are some things that could be done to improve communications, but overall I think they did OK on Friday morning.
<Original post text:>
I doubt we'll ever hear what really happened to the original P506 train and what it hit or why. When P508 passed over the area they did not observe any debris or damage, but they could see marks (probably on the pavement) where something had happened.
On Monday 11/14/2016, there was a regularly scheduled Fiscal Management & Control Board ("FMCB") meeting. These meetings are held generally every two weeks on Mondays and open to the public and the press. At this meeting, the Keolis Chief Safety Officer revealed that a train had struck a buffer at the end of a track at South Station in the early morning of November 4th (actually 12:25 AM). Boston Globe article here. This revelation appeared to have been in response to the concern that safety incidents were being withheld from the FMCB.
After some additional digging, the Boston Globe reporter was able to determine that the train involved in the buffer strike at South Station was the same train that was involved in the incident described in this blog post. A follow up article linked the early AM incident with the P506 issue.
The sequence of events now makes some sense to me. Without any evidence or inside knowledge, I can speculate as to what happened:
1. Strike occurs at 12:25 AM at South Station. Some trains are stored overnight at South Station, and perhaps this set was stored there. But even if it was moved to a different location, the sequence doesn't much change.
2. This set is then used for the P501 - P506 equipment rotation for the AM commute. P501 departs South Station at 4:45 AM. With the damaged plow being dragged at the rear of the train, any defects or irregularites may not have been noticed, especially with nobody at the back of the train during its transit.
3. Upon arrival in Worcester, the train 'turns' (not physically - the engineer just moves from the locomotive to the control coach) to become P506. With the damaged plow at the front of the train being pushed ahead, it is in a position much more likely to get 'caught' on something.
I'm thinking of a situation similar to dragging a rake across the lawn: drag the rake behind you and it will just skip over any rocks or changes in the surface. But push the rake ahead of you facing the wrong way, and it will easily 'catch' a rock or similar obstruction.
It also makes sense that P506 had some sort of related incident in Ashland. Travelling east from Worcester, Cherry Street in Ashland is the second grade crossing that the train would encounter. (The first one is Parmenter Road in Grafton - I can't explain why the train didn't have a problem there). A grade crossing is important to the context of this story because it is typically the only place where there is any sort of obstruction between and immediately adjacent to the rails that rises up to the top of the rail. The pavement for the road could have been enough to catch something hanging down from the damaged plow. This would also explain why there was no 'debris' or object found - the already damaged plow just struck something associated with the railroad.
The Boston Globe articles indicate that all involved are taking the incident and the failure to report the original incident very seriously. The Federal Railroad Administration is investigating along with the expected internal investigations by both the MBTA and Keolis.
The Monday morning commute was rough for many passengers. It appeared that there were a number of factors at play.
First, let's talk about slippery rail. It is a true problem for railroads around the world - it isn't just an excuse they're creating to justify delays. The advantage of moving trains by rail - low friction & higher efficiency - becomes a problem when conditions are not ideal. The leaves that fall during autumn get crushed on the rail head and create a slippery substance. Locomotives have trouble stopping and accelerating on these slippery rails. Natick Center is a notoriously bad spot, especially on track 2. Track 2 is depressed to maintain adequate clearance for tall freight cars under the roadway bridges in downtown Natick. This depression is just enough to cause problems for trains stopping at or trying to depart Natick Center. But the slippery rail can cause problems at any station stop (but note that slippery rail is less of a factor for express trains - it is primarily a problem for trains stopping and starting frequently, such as local trains).
The primary method that Keolis uses to mitigate the effects of the slippery substance from crushed leaves is the operation of two 'wash trains' - one train on the north side and one on the south side. Each train is essentially just a power washer that cleans the substance off the rails with a blast of 15,000 psi water. The power washer is mounted on a flatbed railcar and a tank car is included in the set to provide a water source.
Old announcement with photo about wash train here.
Article about world-wide troubles with slippery rail here.
Keolis video about slippery rail here.
Keolis flyer about slippery rail here.
I have inquired about the usage of the wash train on the Framingham-Worcester line, but I haven't received any confirmation of when it last operated on our line. Certainly the strong wind storm over the weekend which blew down many leaves didn't help, and it may have been too late to deploy the wash train to mitigate the damage done by the storm.
As you would guess, slippery rail prevents proper traction between the train wheels and the rail. In addition to preventing the train from stopping or starting normally, damage can occur to both the rail and the locomotive.
Wheel slip can actually occur in two different scenarios: stopping and starting. When a train attempts to stop and the wheels slip, some wheels STOP rotating and the train will seem to lurch as you enter a station. You may be familiar with this sensation - it is apparent in extreme cases. It is a very similar situation to a car's brakes 'locking up' when skidding on ice or snow. And also similar to a car, when traction motor wheels slip during acceleration and departure from a station, some wheels may SPIN after losing traction against the railhead. Here is a crazy video of extreme wheel slip on a freight train (although this is probably not related to slippery rail, it still illustrates what happens with wheel slip).
Below is a photo of damage to a rail at Grafton station on our own Framingham-Worcester line caused by wheel slip - a traction wheel on a locomotive spinning but not moving the train anywhere. Kudos to one of our alert Twitter users for spotting this! Note that this photo is from a year ago - not this season.
On Monday 10/24/16, P500 - the first inbound train of the morning - encountered serious wheel slip problems. Some crewmembers reported to passengers that 'leaves got stuck in the axles' of the train, but I'm not sure if that was the case. If leaves became ingested into the traction motors on the locomotive, that would be a difficult problem to remedy in the field. More likely the problem was that the locomotive computer managing the traction motors sensed the wheel slip as the train stopped at each station and interpreted the non-rotating wheels as "locked axles." The computer shut down the motors to prevent damage to the rail, wheels, or motors. This would require the crew to go to the locomotive at the back of the train and reset the computer to allow the motors to operate. Again, it may have been leaves in the axles - I wasn't there - but in any case, the root cause was slippery rail.
P500 ended up arriving in Boston about 60 minutes late. In the meantime, that jammed up the trains behind it. P502 was converted to a local (rather than an express) - although I'm not sure why they did this. But the conversion to a local had no major effect on the on-time performance of P502, since it was already blocked by the very late P500 directly ahead.
P580 & P504 were delayed by the traffic jam of P500 & P502. P506 and P508 both experienced minor delays, but only as a result of their own problems dealing with wheel slip and slippery rail - not because of the traffic jam from P500.
So by 7:00 - 7:30 AM, most of the residual delays from the very delayed P500 were being worked out. While inbound trains were still going to be delayed by the continued slippery rail conditions, those delays would have been relatively minimal (P500 effectively cleaned the track as the first train across the line).
This is where the other factors of the morning start to play a role in the whole story.
The second factor is the "equipment cycle" or "equipment rotation." The Worcester storage yard only has space for four train sets. The four trains that are stored there overnight are used for P500, P502, P504, and P508. Every other train on the schedule is a 'turn' of an outbound train. Long time readers are familiar with my frequent references to the "Turn Table" which explains this concept in detail.
With the new schedule implemented in May 2016, most Framingham-Worcester trains remain on the Framingham-Worcester line all day long - going back and forth between Worcester, Framingham, and South Station. So this is the second factor - with trains P500, P502, P580, and P504 all delayed on their inbound trips, their subsequent outbound trips - and then potentially the next subsequent inbound trips - were bound to be delayed. I refer to this as 'cascaded' delays.
The normal equipment rotation at South Station includes the following 'turns:'
P500 goes outbound as a 'deadhead' / non-revenue super express direct to Worcester, where it becomes the P552 "Heart to Hub" bullet train.
P502 goes outbound as P585 which then turns at Framingham to become P584.
P580 goes outbound as P505 which then turns at Worcester to become P512.
P504 goes outbound as P587 which then turns at Framingham to become P586.
Therefore, there was already a potential for P584, P552, P586, and P512 to be delayed.
Here is where the third factor comes into play. At about 7:45 AM, the Keolis dispatcher instructed train P585 to continue west as P505 and go all the way to Worcester and become P512. This actually made some sense to me - P585 operates ahead of P505, the usual equipment for P512. Having the P585 equipment get out to Worcester could mean less delays for P584 and P512 - assuming the P505 equipment would go out to Framingham and operate as P584.
Keolis sent out a message at 7:32 AM that P584 would operate 10-20 minutes late. This was NOT surprising to me, given the 'cascading delay' issues noted above. P584 is scheduled to originate at Framingham with a 7:50 AM departure - just after the passage of P508. It follows P508 making all the local stops. P584 typically has ridership of ~750 passengers.
However, equipment for P584 never appeared in Framingham. Multiple requests for information from @MBTA_CR on the status of P584 went unanswered. A message went out at 9:00 AM that P584 was cancelled - 1 hour and 10 minutes AFTER the scheduled departure of the train from Framingham. There were no official messages regarding P584 between 7:32 AM and 9:00 AM. This fourth factor in our list - failed communications - is baffling, since as much as they get dumped on by lots of riders and twitter users, they are usually pretty good at major messages such as those. I can only guess that the multiple changing circumstances forced important information to get lost. I hope they conduct a post-mortem / lessons learned inspection of this communication failure and take a look at any systems or processes that may be able to be improved.
If P584 had operated 10-20 minutes late, it still probably would have operated AHEAD of P510 - and probably would have swept up passengers waiting for P584 AND P510. Any resultant overcrowding on P584 could have been accommodated by P510 directly behind.
The fifth factor sealed many passengers' fate. P510 was operating with a consist of only 5 flats - a capacity of 570 seats. The normal load on P510 is 687 passengers, so it was already destined to become overcrowded. With P584 cancelled, P510 was now being asked to accommodate the load of both P584 and P510 - approximately 1,437 passengers.
As many passengers reported, P510 was filled to capacity at Natick Center and passengers waiting for P584 or P510 at Wellesley Square, Wellesley Hills, Wellesley Farms, Auburndale, West Newton, and Newtonville were unable to board P510 due to the overcrowding on it. Video from Wellesley Farms:
It is impossible to know why P585/P584 was cancelled. It could have been that the P500 equipment was so negatively affected by its battle with the slippery rail that it needed to be taken out of service and sent for maintenance. This could have been a reason for the Framingham-Worcester line being short one set of equipment by the end of the AM commute.
Fairmount train 750 (a prominent subject of Sunday's Boston Globe article) was briefly cancelled at 6:45 AM and then "un-cancelled" at 6:49 AM, but as I study the timing of these events, it doesn't look like that could have affected the Framingham-Worcester equipment rotation (especially the later turn of the P585 / P584 equipment).
If you haven't read Sunday's front page Boston Globe story about the Fairmount line, it is a very important story. A follow-up article has just been posted that confirms what many of us have long suspected - there is a shortage of 14 coaches versus what is needed to operate the entire commuter rail schedule. I honestly don't believe that there is any targeted denial of service to the passengers of the Fairmount line - I think that Keolis and the MBTA have been allocating equipment to the services with the highest ridership and cancelling the lowest ridership trains. But the optics of preserving service to wealthier suburbs while cancelling trains in urban areas is problematic. The bottom line remains that for at least a few weeks (perhaps months?) the MBTA and Keolis have been trying to operate a schedule with insufficient equipment. There are no indications that can be immediately or quickly remedied. The Fairmount story and the equipment shortage aspect of this story may draw more attention until sufficient equipment can be allocated to the current schedule. There is a potential that future cancellations of trains will rotate amongst all of the lines - and affect both lightly used trains and trains with heavier loads. And perhaps that is more 'fair.' But that's just my speculation.
In summary, many factors conspired to make the Monday morning commute problematic.
1) Slippery rail. Without this problem, everything probably would have been OK. It is unknown at this time if there was anything that Keolis could have done better to mitigate the slippery rail, since it is unknown if the wash train has been deployed appropriately.
2) Cascaded delays as a result of the equipment cycle. Add a spare set into the mix and cascaded delays may have been mitigated.
3) The decision to cancel P584. Utilizing a spare set for P585 / P584 could have alleviated this problem.
4) The communications failure regarding the cancellation of P584. Had Wellesley & Newton passengers been told at 7:45 AM (when the P585 equipment was re-purposed and sent to Worcester for P512) that P584 was cancelled and/or that the next train might not be able to accommodate them, they could have sought alternate transportation.
5) The small size of P510. Already under capacity for its usual load, it was forced to try to accommodate the load of P584 & P510. Even having a relatively large set for P510 probably would not have mitigated this problem - P584 + P510 is just too many people for all but the largest sets.
Finally, please remember that the crewmembers out on the trains are NOT the people that have anything to do with any of the issues or decisions above. There were many kudos given out to the crew of P500 for keeping their passengers informed about the problems that train experienced. But there were also reports of passengers becoming irate with crewmembers regarding the overcrowding or lateness issues. Keolis crewmembers are NOT our enemies. Let's bring back some civility and direct our anger at letter writing and advocacy for a better commuter rail system.
A little over a week ago, MBTA General Manager Brian Shortsleeve invited me to participate in a new group he wanted to convene to look at potential schedule changes for the Framingham-Worcester line. The first meeting of the "Worcester Line Working Group" was held on Thursday 9/29/2016 in his conference room. In attendance were Mr. Shortsleeve, me, another commuter from Wellesley, Rep. Alice Peisch (D - Wellesley), Lt. Governor Karyn Polito, Jody Ray (MBTA Assistant General Manager in charge of Commuter Rail Railroad Operations), the Keolis Manager of Operations Planning, David Scorey (Keolis CEO/GM), an MBTA policy analyst, and a Deputy Director of MBTA Railroad Operations. Both Mr. Scorey & Mr. Shortsleeve are regular Framingham-Worcester line commuters so they bring the experience of regular riders to the group - as well as their management perspective.
More details after the "read more."
Not an MBTA employee, not a Keolis employee, just a regular commuter (with a scanner and some knowledge of railroad operations). Everything here is my speculation, so up to you to judge the accuracy.