13 April 2017

Core Capacity Math

With federal funding for Caltrain modernization on indefinite hold, a diverse flock of vultures have started circling the skies over the peninsula corridor--opponents of electrification, boosters of peninsula BART, opponents of high-speed rail, and detractors of rail or transit in general. Some opponents have latched on to a perceived vulnerability of the modernization project, using a legal parsing of federal regulations to allege that Caltrain's application for federal transit funding is fraudulent and illegal. This sensational claim merits closer examination.

FTA "Core Capacity"
The $647M of funding that Caltrain has pursued for several years, and came agonizingly close to obtaining, is allocated by the Federal Transit Administration and disbursed by congress under a competitive grant program known as a Capital Investment Grant or nicknamed "5309," after the section number of the United States Code under which the program is defined. One of the ways to obtain federal funding under this program is to increase the capacity of an existing transit system by at least ten percent. This is known as a "core capacity" grant, the type that Caltrain is pursuing.
The fracas is all about this ten percent, and whether Caltrain is actually meeting the criteria for eligibility.
To prevent gaming of the system, section 5309 policy guidance specifies how that ten percent increase is to be counted for a core capacity project to become eligible for funding. Importantly, the FTA makes an artificial distinction between "light rail / heavy rail" (where passengers traditionally sit and stand) and "commuter rail" (where passengers traditionally only sit) that fundamentally changes the metric used to measure capacity.
  • Light rail / heavy rail capacity is measured by "peak hour person capacity in the peak direction," a measurement that includes standees and is based on floor space.
  • Commuter rail capacity is measured by "peak hour peak direction seated load," a measurement that excludes standees and is based solely on seat count.
Under this uniquely American taxonomy, Caltrain is an odd duck: the modernization project seeks to transform it from a traditional diesel commuter train into a swift and frequent transit system whose attributes will bear a closer resemblance to a heavy rail system like BART than to a stereotypical American commuter train. Indeed, Caltrain's choice of a Swiss train design underscores the cultural disconnect with the rigid system of American train categories.

If you needed any more proof, BART is removing seats to increase capacity!

EMU seating capacity
As ordered, Caltrain's EMUs will have significantly fewer seats per train  than today's six-car diesels. EMUs are not a magical technology: they may lack a locomotive, but all the traction components that would normally be found in the locomotive still need to be accommodated elsewhere in the train. In a preliminary brochure of Caltrain's new EMUs, the traction bits show up as cabinets marked with a spark symbol. These things take up space, so a six-car EMU offers fewer seats than a traditional six-car train hauled by a diesel locomotive.

(In passing, some have blamed the lower seating capacity of the Stadler EMUs on the dual door configuration. A cursory review of proposed seating layouts can retire this fallacy.)

The seating capacity directly counted from the brochure was 573 seats per six-car EMU, although Caltrain's FTA grant application assumes only 558, averaging just 93 seats per car.

A Shifting Baseline
Ten percent, but compared to what? The baseline present-day capacity is also contentious, since Caltrain's record ridership has created the need to provide more seats today.
In Caltrain's September 2016 grant application materials and correspondence with the FTA (helpfully obtained under a Freedom of Information Act request by CARRD Morris Brown), the capacity baseline is tabulated as 3403 seats per peak hour per peak direction, with a net increase of 365 seats (barely squeaking by with a 10.7% increase) to a total of 3768 after the modernization project is completed.  Detailed tabulation is provided below, as extracted from the grant application.

Detail of Existing Operations Commuter Rail
Train # Train Line Reference
(e.g. Name/Color/Number)
Departure Time Number of Cars Seats per Car Seats Per Train
1 #217 6:57 5 121 605
2 #319 7:03 6 132 792
3 #221 7:18 5 120 600
4 #323 7:45 6 131 786
5 #225 7:50 5 124 620
Total During the Peak Hour


Detail of Operations At Project Opening Commuter Rail
Train # Line Reference Departure Time Number of Cars Seats per Car Seats Per Train
1 #305 7:00 6 134 804
2 #113 7:07 6 93 558
3 #115 7:12 6 93 558
4 #307 7:29 6 122 732
5 #117 7:36 6 93 558
6 #119 7:42 6 93 558
Total During the Peak Hour


Opponents have pointed out that well before the date of the FTA grant application, train 225 was converted to a six-car Bombardier consist seating about 790 passengers, as was train 217, thus increasing today's peak hour baseline by 170 + 185 = 355 seats, and cutting the future capacity increase fully in half from 10.7% to an ineligible zero percent 5%. Caltrain may have done itself in simply by serving its customers today.

How to solve the FTA seating equation
To meet the FTA requirements without question, and to shoo those vultures away, here are some solutions Caltrain could reasonably pursue:
  • Build the train cars 3.2 meters wide with five-abreast seating on the upper deck (the lower deck would remain four-abreast with a wider aisle). A car width of 3.2 meters is within the AAR Plate F loading gauge that is cleared to operate on the peninsula corridor, and is a common width in East Coast commuter railroads (the LIRR M7, the Metro North M8, and the SEPTA Silverliner V are all 3.2 meters wide with sections of five-abreast seating). Why Caltrain hasn't already pursued this is baffling, because it is a low-cost and high-benefit change regardless of FTA rules. The Stadler KISS EMU that Caltrain ordered has previously been delivered in widths up to 3.4 meters. This design change is worth +64 seats per six-car EMU, or +256 seats/peak hour, or +7.5% core capacity.
  • Increase the size of the initial EMU order.  This is a tough sell, given how hard it has been to fund the modernization project, but the relative cost increment is minor when considered in proportion to the entire budget. The Stadler contract already includes an additional 96 cars under a fixed-price option, 32 of which (one third) could be exercised to make all 16 of the EMUs on order eight cars long.  The incremental cost would be another $130M (one third of $390M, or less than 7% of the entire value of the PCEP project), and the seating capacity would go up by +186 seats per EMU, or +744 seats/peak hour, or +21.9% core capacity.  That's right: for an extra 7% cost you can triple the capacity increase.
  • Both measures applied together would increase seating capacity by +272 seats per EMU, or +1088 seats/peak hour, or +32% core capacity (over and above the +10.7% capacity increase in Caltrain's FTA application).
The devil is of course in the details: changing the Stadler car shell is not free at this late stage of design, and lengthening trains to eight cars isn't just a financial headache but brings about awkwardness with certain platforms that are shorter than they ought to be.

Nevertheless, if it comes down to an existential issue of project eligibility, seat count nitpickers can undoubtedly be satisfied, and to everyone's benefit, by making a few basic adjustments. When a project's core capacity metric can be tripled for just 7% extra cost, it's a clear indication that it isn't an illegitimate fraud. Core capacity seat math either isn't an issue, or it can easily be resolved.


  1. Clem:

    It least you understand the problem. Your numbers are not right, since with correct existing numbers, the capacity increase after electrification, using current conditions works out to be 1% not 5%.

    The FOIA request was from me, not from CAARD, so don't blame them for what it disclosed. When Caltrain was asked about the non-compliance, they responded saying I had not included different after electrification conditions, and that indeed they do comply.

    I asked for a EXCEL Template showing this to be true, and Caltrain declined to provide one, saying the FTA only required for approval the spreadsheet obtained via the FOIA.

    You seem to be of the opinion that all of this makes no sense, but the needed 10% increase is codified in law; that is the application must meet this requirement or the FTA cannot approve the grant request.

    These are the ground rules, like them or not, that Core Grant applicants must meet. In this case as filed, Caltrain certainly did not meet them; Caltrain did not supply valid existing capacity data.

    1. I'm not sure where you got the idea that I "blame" anyone for the FOIA... it is quite helpful.

      Note that FTA uses Caltrain's spreadsheet as guidance only; they are the ultimate arbiters of the project ratings and of the validity of the calculations. They are unlikely to simply take Caltrain's word.

      My point stands that the numbers can be vastly improved with changes to train design and order structure. Even the supposed 1% figure can easily be rescued to well above the 10% eligibility criterion.

    2. I should also mention, I believe my numbers are correct. On the contrary, there are several factual errors in the Warren document.

      1) page 6 incorrectly speculates that two diesel trains are retained in the future scenario for the specific purpose of goosing the seat count. In fact, Caltrain is not ordering a big enough EMU fleet (only 16) to cover the peak-hour service. They are fleet-limited to 4 tph EMU, with the remaining 2 tph (Baby Bullets) operated using diesel equipment. The EMU contract documents show clearly that Caltrain intends to exercise the 96 option cars to eventually expand the EMU fleet to 8 x 24 = 192 cars, from the initial EMU fleet of 6 x 16 = 96 cars reflected in the FTA grant application materials. Warren's idea that Caltrain would, could or should ever operate 6 tph of six-car EMUs is baseless. The next logical capacity increment is to extend EMUs to 8 cars, and not to remove the diesel bullets as Warren envisions.

      2) top of page 8 incorrectly speculates that room for bikes, wheelchairs and bathroom will reduce the seat count of a six-car EMU below 558 seats. A simple review of the EMU layout reveals otherwise.

      3) bottom of page 8 argues they should have used six trains per hour in the current seat count, but one look at the timetable shows there are only five trains per hour per direction during the peak. I certainly hope that Warren isn't attempting a hair-splitting argument that there are six trains per peak hour. That would be gaming the numbers, the very thing he decries, and can be immediately rejected.

      The only numerically valid claim regarding today's capacity baseline is that a sixth car was added to one of the trains. Does Caltrain deserve to be punished for adding capacity today?

      The takeaway: Warren fails to understand how easily the future capacity number can be increased, should the FTA grant be so much as threatened by his analysis. It's a swing and a miss, one that could and should have been avoided by Caltrain if they had been competent enough to anticipate the issue.

    3. Clem:

      I don't deny that Caltrain might be able to submit an FFGA grant application that would meet the codified requirements for a grant; but this application does not meet the requirements.

      Please note, the 10% required increase in capacity must be met.

      Look at:


      and read:

      the proposed project corridor must be at or over capacity currently or will be within five years [5309(e)(2)(A)(iii)]
      the proposed project must be a substantial, corridor-based capital investment in an existing fixed guideway system that increases the capacity of a corridor by not less than 10 percent [5309(a)(2)]
      the proposed project does not include project elements designed to maintain a state of good repair [5309(a)(2)]
      the proposed project cannot include elements to improve general station facilities or parking, or acquisition of rolling stock alone [5309(b)(2)]

      The point I am trying to make is, the spread sheet must provide at least a 10% increase in capacity or the application fails. This application fails and the FFGA grant as submitted will not be approved.

      What I really don't understand is all those who want to support Caltrain, support this project, which gives away 40% of track capacity to HSR. That is capacity that will for sure be needed by Caltrain in the future.

    4. I cannot take issue with the law as written, and there is no need to quote the language. I agree the capacity numbers as submitted are dodgy at best. My point is the requirement could so easily be met as to make this line of attack completely ineffective. It is baffling to me that Caltrain and their FTA minders even left themselves open to it.

  2. 1. Couldn't they just schedule more trains? Or are they limited by the number of trains they're buying, or the bullet/regular timing issues?

    2. How would changing the width affect future compatibility with HSR?

    Anyway, it seems odd to say that the plans don't expand capacity enough when they obviously lay the necessary groundwork for a much greater capacity, which can be incrementally realized over the years, as demand warrants, through making changes to the schedule and buying additional trains.

    1. Yes, the number of trains that can be scheduled is limited by fleet size. Not to mention environmental clearance, which is required to add more than 6 tphpd.

      HSR hasn't settled on a width yet, but it could be up to 3.4 m (Shinkansen-like). So, 3.2 m Caltrain would have no compatibility issues that I am aware of.

      Your closing comment is spot on.

    2. Why do you think environmental clearance is required to run more than 6 trains per hour?

    3. Gate down time is a potentially significant environmental impact, covered extensively during the PCEP EIR process. Noise and vibration are another, especially horn noise at crossings, for which EMUs will be no quieter than diesels.

    4. Ahhh, never mind, it appears I was mistaken. It was pointed out to me that there is a specific exemption in CEQA for rail traffic increases on the peninsula corridor. What about NEPA?

    5. Under SP era, there was more then 6 train/h during peak period. (I assume 8 trains)
      Why do you think environmental clearance are needed for increase train?

  3. Roland LeBrun pointed out to me that #217 is now also a six-car Bombardier set. That means four out of the five peak hour trains are six-car Bombardier sets, which entirely wipes out the future capacity increase. The post has been updated to reflect this.

  4. Legally, the FTA staff has complete discretion to evaluate "capacity increase" however it wants. Yes, there is a policy guideline but it is just that -- a guideline. If the project carries 10% more passengers than before (seated or otherwise), then I don't see the problem.

    1. Drunk Engineer: The "interim policy guidance" rules I posted above are not just a guideline. The enabling legislation from Congress for the FAST act put these into law. They must be followed/

    2. On one hand, the law can be satisfied, and then some, by purchasing a fleet of sixteen 8-car EMUs (7% PCEP cost increase).

      On the other hand, the law is completely irrelevant if the Republican congress goes along with the Trump administration's plans to defund transit.

    3. Morris,
      There is a huge difference between "policy guidance" and a regulation/law. The text of the enabling legislation just says "10% capacity increase". FTA can define that however they want. It is not at all uncommon for grant applications to fall within a grey area of a policy guidance doc.

  5. Clem:

    From your own previous posts, an 8 car EMU is too long at some stations, such as my city of Menlo Park. Without grade separation, which are decades at least in the future, are you just going to by-pass my city? Not going to be popular for sure around here.

    1. Menlo Park is a fine example of how easy and cheap it is to solve this problem. The existing platform is 645 feet long. An 8-car EMU is 690 feet long. The distance between Ravenswood and Oak Grove is 930 feet. Solution: ~$1 million project to extend platforms to the south by 150 feet, removing the pedestrian crossing at the south end of the platforms.

      Each station is its own microcosm and requires its own solution, but for the most part extending platforms is easy and cheap.

      I will have a post about this sometime, with case-by-case details.

    2. A much better solution is to give up on electrification. Running some kind of Tier 4 or hybrid which will not need electrification of the tracks and would save $1 to $1.5 billion. In addition would not need to cede 40% of corridor bandwidth to HSR.

      Finding funds to replace the present HSR funding should be a smooth ride, now that SB-1 will become law and a little of the pork included there can be extracted. Hell, the Governor had no problem finding $1 billion to bride legislators to get SB-1 passed.

    3. I would say that perhaps Cal Train, and Amtrak, need to look to the UK to see how rail is being revitalized and how diesel and electrified lines run together. Rail service near to London and into London is all electric, but non-electric can also run on those tracks. In some of the rural areas, particularly in the Midlands, many lines that were taken out of service in the 60's have been restored, and not all are tourist trains.

    4. You have a problem with public transportation.
      You "look to the UK".
      Now you two problems. In fact, now you have dozens of problems.

      Why not look at things that work well from funny-language parts of the world that only deliver catastrophic budget and schedule and scope failure projects sometimes, rather than nearly always?

    5. @Morris: I do appreciate the nakedness of your agenda, but bums in seats is not the only metric for what makes a good rail transit system. Average speed is another important metric, and Tier-whatever diesels will never fix that. Bear in mind that Caltrain locals average 29 mph on a good day. That must be greatly improved, and the best prescription for that is electrification (already being pursued) as well as level boarding (not quite on everyone's radar yet).

    6. If there is platform length issue, just run short train more frequently. Caltrain can make differentiate between local and express train in term of train consist. 8 car Express train vs 4-car local train. This is win-win solution for express train's capacity while maintaining frequency of local train. If City of Menlo Park don't willing to pay for grade separation, they will not get express train.

    7. Unfortunately there are many locations on the Peninsula, not just Menlo Park, that are not grade separated. As the number of trains per hour increases, so does the gate down times and resultant traffic delays. There is also a limitation in the number of locomotives and passenger cars available. Currently at peak hours, most of the existing fleet is on the road.

    8. The typical and relatively infrequent RR crossing gate activation to allow a train to pass with several hundred (sometimes even a thousand!) people aboard lasts about a minute or so.

      The ubiquitous, very frequent and unavoidable red light activation at most any large and/or busy intersection to allow cars with far fewer people (dozens ... sometimes even a hundred!) aboard to pass almost always lasts much longer.

      Now can anyone explain why drivers complain loudly about the former while silently ignoring the latter? ... and, for extra credit, why so many intelligent seeming people, particularly transit advocates and elected officials, pay those complaints any attention without even a snicker (let alone with head-nodding agreement)?

    9. First of all, plenty of people do get irritated and complain about stoplights. All the time. How many times have I heard somebody say "If only those stupid traffic engineers would TIME THE LIGHTS on this road! They're clueless." Meanwhile these complainers are completely ignorant of the fact that they are not the only person on the road: although the stoplights may not be timed to optimize THEIR journey time, they are often quite well tuned for maximum network throughput.

      Nevertheless, I do agree that most people feel a certain deeper sense of dread when they see those crossing gates go down. When a stoplight turns red, you can be relatively assured that it will turn green again in 2, 3, or at the ABSOLUTE worst case maybe 5 minutes. Whereas, with a railroad crossing, most people have had the experience of being stuck at a railroad crossing for 5 or 10 minutes while a mile long train slowly trundles by, only to have it STOP across the tracks and sit there with no indication at all of when it's ever going to move. I think it's the hair-pulling frustration caused by that worst-case scenario from RR crossings, combined with the unpredictability element (Will I be stuck here for 30 seconds or 15 minutes??) that drives it.

      Not rational, by any means, but it is what it is.

    10. Agreed that people MIGHT have to wait longer at a railroad crossing, but on the Peninsula the number of freight trains is VERY low (and they're quite short), so it's much less of an issue there.

      Compare that to San Diego, where sometimes freight trains are being assembled while blocking several crossings at a time, leading idiot pedestrians to climb over and under freight cars to cross the tracks (Padres fans are notorious for doing so).

  6. Noting that Clem (correctly) tags incompetence to Caltrain for some of this, please note the Caltrain Board awarded CEO Jim Hartnett a very nice increase in compensation, a bonus, annual inflation adjustment and contract extension



    I guess signing vendor contracts before securing funding, which will at least result in $20 million in extra costs, doesn't faze the Caltrain Board.

    1. Morris, that's not correct. Your link is to the Samtrans Board meeting. According to the Joint Powers Agreement, Samtrans is the managing agent of Caltrain, and the General Manager of Samtrans is the General Manager of Caltrain. Jim Hartnett does not report to the Caltrain JPB. He reports to the Samtrans Board.

  7. We all know that PCEP only gives us 6-car EMUs, still some diesel trains and no level boarding. The root cause here is cost, and I think we can understand the pressure in keeping costs in line with funding that can be obtained. We might not like it, but it's the state of the world.

    With that said, I think I found the first estimates of level boarding, extended platforms (where needed) and 100% 8-Car EMU fleet. Buried deep in meeting notes are initial discussions for raising the bridge tolls as part of RM3. Caltrain is suggesting the following projects (with price tags) for possible consideration:
    * Full conversion to 100% EMU + capacity increase: $440m
    * Level boarding & platform extensions: $250M

    These are not fully scoped, but had they been included in original PCEP, the current $2.0 billion budget would be $2.7 billion, and larger projects are even harder to pass. Anyway, this is probably going too far into the political side of things.

  8. I see there is updated capacity information (dated 4/19/2017) at:


    Looks like current seated capacity has been increased to 3705 - close to Clem's estimate, and future 8-car all-electric seated capacity to 4560 or 95/car. Initial post-PCEP capacity (25% diesel/75% electric) has been listed as 4112, but not sure how they got to this given Clem's analysis above.

    Anyone have any ideas how they came up with this number?

    1. Caltrain's new numbers do need justification, although they're hammering the right message: there's more to capacity than counting seats.

      4112 minus 4 tph of 6-car EMUs at 93 seats/car leaves (4112 - 4*6*93) = 1880 seats to spread across two diesel trains, which is not plausible.

      If the EMUs can seat 5-abreast on the upper deck, each 6-car EMU can seat another 64 passengers, leaving (1880 - 6*64) = 1492 seats to spread across two diesel bullets, or 746 each. That seems quite plausible.

      I do think they've gone five abreast!

    2. morris brown02 May, 2017 05:21

      With regards getting a FTA FFGA grant approved, the FTA is not interested in counting "standees". The FTA wants "seated passengers" increases and insists that there funding buy increases in "seated passengers"

  9. House transportation bill includes $100 million for electrification.


  10. While driving to Lowe's I spotted a freshly painted blue GATX locomotive parked in San Bruno. Do you know if this a recent thing (ie, involved with UP leaving) or was it always there?

    1. GATX is a leasing company- likely UP is using the locomotive to fill in for a motive power shortage.

    2. Makes sense. While I was sitting at the Jack-in-the-Box three UP 9-44s moved by at 1200 today.

  11. morris brown11 May, 2017 08:32

    At the May 4th, Caltrain Board meeting, CEO Jim Hartnett promised to send to the Board data to explain the passenger capacity issue.

    In a letter to the board, Seamus Murphy has sent out a memo, which can be viewed at:


    The problem remains since although this note gives numbers, it fails to say how these numbers are derived.

    The new website


    shows only 2 x 2 seating for the EMUs, thus presumably eliminating using 3 x2 seating to increase capacity.

    The EMUs have less capacity in a 6 unit train sets then the Metrolink cars they are replacing. How is this magic of increased capacity achieved.

    As Clem as postulated above, it would seem Caltrain will have to start running 7 or 8 car EMUs to get the increase they are showing in this memo.

    This then indeed brings up the funding issue; where is the extra funding going to be found?

    Caltrain has said they don't plan on submitting a new grant application showing these numbers, but maybe they have now changed their stance on this.

    All of this within the context that the grant application showed 6 trains per hour for EMUs, while limiting existing diesel powered trains to 5 trains per hour. I would think the FTA will find all of this very interesting.

    1. Hi Morris, indeed it would be nice if they showed their math behind their 3705/4112 figures.

      Another conceivable "cheat" to get to 4112 would be to add 1 car each to the 2tph of diesel bullets, for a total of 7 cars or ~920 seats/train, although that would start to exceed the platform lengths at several stations (not to mention taxing the poor diesel locomotives)

    2. Over the years, Caltrain has occasionally run specials as long as 10 cars long (and not always as "double-headers").

    3. Based on the Seamus Murphy Memo, it seems indeed most likely that 7 car diesel bullets is the plan. Any idea which current stations don't have platforms long enough to support this? Not much loss anyway as EMUs better acceleration means the difference in run time between limited and bullet will be a lot less than it is now. That's also the reason they can run 6tph vs. 5tph for diesel, while still maintaining reasonable service patterns.

    4. Lengths of all the Caltrain platforms are documented in this awesome California rail network schematic from SMA+partner, made with Swiss precision and information design.

      A seven-car bullet would need at least 600 feet of platform length, and mini-highs may need to be re-arranged. Probably figure a bit longer to account for stop point accuracy, which isn't a local specialty. My wag is 700 feet to make it work in daily practice.

    5. Indeed, 10-car trains are coming on Sunday for Bay To Breakers. In fact, Caltrain is running two of them in the morning. Since the pick up passengers only, the probably stagger their stop locations so either front-5 or rear-5 cars stop at alternating stations.

  12. Thanks Clem for the great SMA+partner link. Looks like of the existing bullet stops only 22nd St. is less than 600 feet, but plenty of stops are just over. Pretty sure they can make 2tph 7-car bullets work. Most likely the schedule would be quite different than the current one due to the very different acceleration profiles of the EMUs and 7-car Bombardiers.

    1. Of course, performance of post-electrification baby-bullets would improve as well if Caltrain uses the AEM-7 it plans to buy from Amtrak as back-ups on these trains, as they are approximately 1.5x more powerful than MP36s.