Single-track, surface running, downtown.

Over the last century pretty much all of the major passenger and freight rail lines in Osaka were elevated, trenched, or replaced with subways. In particular there was a double track freight bypass which branched off the main north-south Tokaido Line and ran into a large freight yard at Umeda. Then almost as an afterthought, the far end of the Umeda yard was connected back into the main Osaka rail loop – basically a large runaround track.

As befits a track designed for occasional use it’s not engineered to anything like the rest of the Osaka network. The single track winds around the edge of the yard, behind shops, across level crossings and under viaducts for other, more important rail lines. Just not a heavily-used track.

But there’s a big capacity constraint in this system. Traditionally, all the JR trains to the south and east – to Nara, or Wakayama – terminated at Tennoji or Nanba, while all the trains to the north and west terminated at Osaka/Umeda. Then if you wanted to transfer between those stations you used the Loop Line, a tram, or the subway. It’s very much like Gare du Nord / Gare de Lyon or New York Penn / New York Grand Central.

This is great for local service. But what if you want to run an express train through from Kyoto to the beach or perhaps the airport? To do this you have to switch from the northside tracks (which are geared towards a stub terminal) to the loop line tracks (which are set up for continuously-circulating orbital trains) and then to the southside tracks (more stubs). This means fouling basically every other train movement. What to do?

Well, if you were a US consultant, you’d propose a flyover or three. Eliminate the multi-track crossing movements, put everything on the right tracks upstream, bang. A project like this might cost $400 million. Or $600 million, for two. Lots of concrete. And indeed, you might expect the Japanese to do the same. Certainly they are no stranger to rail flyovers. But did they? No, they just ran the express trains down the freight runaround.

Both of these trains run 30-minute headways in the peak, so this means JR West is putting four trains per hour in each direction through a weedy singletrack designed for minor yard movements. Let’s all take a cab ride and see what we can observe.

The first thing to notice is at 0:43 where the train switches from the passenger tracks (at left) to the freight bypass (at right). You sort of have to marvel at how bad this track geometry is. First there’s a 25mph yank to the right, then another S-curve to join the southbound freight track, then another curve to line up with the existing underpass. You can also see how the southbound freight track has been bent and contorted into a “broken back” curve to allow for switches in this location; the northbound track (on the far right) retains its original, broader curvature. Then at 1:25 you see the crossover that allows expresses on the northbound freight track to wrong-rail back over said bad trackwork which our cameraman has just traversed.

It’s smooth sailing over the Yodo River to the throat of the Umeda yard before the train abruptly encounters a 33mph switch at 2:58. It’s worth noting here that the Japanese are willing to accept more jerk as their trains traverse turnouts, so that posted 45km/h back at the beginning would probably be 15mph in the Northeast Corridor, and the posted 55km/h you’re going over now is probably 20mph in the US. Winding around the yard, you merge with the northbound track at 3:44 and it’s all single track from there as the train crosses active yard switches at 3:53, 4:11, and 4:46 before crossing a major arterial at grade.

Then at 5:02 it’s under the viaduct that carries all 1067mm JR trains between Osaka and Kobe. The trackwork here is so kinky it’s speed-restricted to 55km/h, which is 33mph in direct translation but probably 20-25mph given US tolerances for runoff and unbalanced superelevation. Another kink at 5:14 is followed by another level crossing. This one has heavy pedestrian traffic, as it’s directly underneath Fukushima Station. Then it’s up the ramp to the Loop Line, past ye olde 201 series at 6:47, and over some more crossovers into the center track at Nishi-Kujo.

What relevance does this have… to anything? Well, it’s potent visual reminder that you don’t always need to pour a bunch of concrete. If you want to put 4tph through some weedy industrial trackage, be like Nike and just do it.

There’s a whole bunch of reasons why the US leans towards over-concreting rail projects. For one thing, it’s long been Federal policy to throw tons of money at capital projects while refusing to assist with operations. This leads to some seriously warped Cost-Benefit analyses. You can have a capital improvement that costs double or triple the present value of the expected operational savings, but if the Feds are paying 55% or 70% it’s a net benefit to you, the agency.

There’s also the freight railroads, who are very happy to add ridiculous stipulations for minor service improvements. Want to run one Coast Daylight a day? Pay for 350 miles of PTC installation and a whole bunch of double track. In Houston commuter rail planning, this has most often shown up as a 25′ minimum track spacing requirement.

But beyond these issues there’s also just a lot of intellectual laziness that is an outgrowth of the political context in which we build trains. In Japan you have a company like JR West which is privatized and publicly-traded and has a fiduciary duty not to waste money on limestone. In France you have an operator like SNCF which enjoys broad support among all political parties and has a mandate to provide quality service at low cost to the public – “TGV is for all,” etc.

But in the US you have one political party which thinks all rail projects are “boondoggles” and part of the secret UN plot to herd us all into Agenda 21 housing blocks, while your other political party is less focused on cost/spending issues and is more concerned with environmental and social outcomes, which is great for getting lines built but not for getting them built cheaply. Finally, at the local level you have a basic indifference to cost on the part of local politicians which is quite bipartisan. Construction is jobs so hey more money more jobs. All of this serves to increase the cost of providing transit service which in turn reduces the amount of transit we all get.

Exceptions happen when the funding gets cut. Two examples:

In Houston, you had Culberson and DeLay waging war against trains, which forced METRO to build the Main Street line with local money. This lead to what is a pretty cost-effective LRT line by modern US standards. Among other things they repurposed an existing 60’s-era underpass to carry trains across Holcombe/Bellaire instead of building something new. This results in a traffic lane which is narrower-than-standard and wire clearance which is less than ideal, but it works. This same sort of creative cheapness is present at the 610 underpass where the space for one track was stolen from the southbound roadway and the other was swiped from the U-turn. This again results in not the greatest wire clearance, so METRO just hung a few crossing gates over the feeder to warn truck drivers they’re about to screw up. If you’re paying for it yourself, it’s cheaper to fix an occasional dewiring than it is to build some soaring 40ft concrete thing that goes over the freeway. Plus you can always go after the guy’s insurance.

In Portland, the South-North Rail Line was supposed to run down Interstate Avenue and then sail over I-5 on a long viaduct to connect with Good Sam/Emanuel before returning to the surface to cross the Steel Bridge. Certainly decent transportation planning; string all your major activity centers together along one line. But expensive. So when voters rejected the South-North proposal, TriMet asked “how can we build this without a tax increase?” They settled on making it nearly entirely street-running, including the same sort of underpass repurposing that you saw on Houston’s Main Street Line.

I’m not exactly sure how you force US transit generally to conform to the cost-efficient design seen on Portland’s Yellow Line, Houston’s Red Line, or JR West. In both US cases the designs came from a lack of anticipated funds; federal funds in Houston’s case, local funds in Portland’s. But cutting transport funds from their current meager levels is not a recipe for more transit.

One not-entirely-seriously option would be to have the Feds commit to 90% funding, wait a decade for everyone to do their long-range plans and EIS’s – and then turn around and reduce funding to 60%. Then you’d stick everyone with the need to find design compromises to cut costs, while simultaneously increasing funding over a 2008 or 2012 baseline.

I don’t think this would actually work, though. So how do we get more and cheaper trains?

4 thoughts on “Single-track, surface running, downtown.”

  1. There’s a lot going on here. For one thing, there’s the as-yet unexplained US-premium on construction costs, e.g. a mile of subway in NYC seems to cost way more than it does in London.

    The lack of federal funds has encouraged cities to be thriftier. If LA was building the Gold Line, Expo Line, Crenshaw, etc. in the era of 90% federal funding, they’d probably be totally grade separated. And the need to spread the wealth of Measure R funds to build enough support, but without making the tax so high that it was unappealing, also encouraged the county to try to advance as many projects as possible. However, as we’ve already established, continuing to underfund transit at the federal level is not a great approach.

    The capital vs O&M problem is compounded by the US’s illogical political divisions. This can be a the inter-state level (e.g. projects between NY and NJ will necessarily waste money somewhere because of the “need for balance”) or at the intra-state level. For example, the Green Line Extension in Boston is being funded by MassDOT and the FTA, so as far as the MBTA is concerned, it’s OPM and they might as well demand gold-plated everything. Hence the operationally unnecessary flyover where the branches split, while in LA, the Blue/Expo has a flat junction.

    When you start looking at commuter rail and intercity, the problem is that freight is a lot more important in the US than in Japan or Europe. Japan is small and has sea access everywhere, so you don’t bring stuff in at a port and ship it 2000 miles by rail. Likewise you don’t load unit coal trains and run them all over the country. The result is that for JR West, passenger service is an absolute necesseity; for UP or BNSF, it’s a nuisance. So UP demands 25′ track centers so that your cute little 3-car trains don’t mess things up, to the point where you can even work on your track without fouling their track by FRA regs. Then there’s the whole FRA crash-worthiness thing, which drives up vehicle weight, lowers speeds, and increases maintenance costs, etc.

    Okay, so this is a really a list of problems without any ideas, yet… so let’s keep thinking…

  2. Alon Levy blames this on NIMBYs and the common law legal system that, in his opinion, empowers them (here and in the UK, India, etc.)

Comments are closed.