HSR Phase One Yards

8 May 2009: In response to a complaint regarding slow downloads of the home page, I've modified this post such that previously embedded videos and maps are now hyperlinks. My apologies for any inconvenience this causes.

---

by Rafael

While there has been a lot of debate about the route CHSRA prefers and the throughput issues of San Francisco's new Transbay Terminal Center, the issue of maintenance and overnight storage of high speed trains has not received as much attention. However, given the distance between the end points of the starter line for the California network, the success of the service absolutely depends on having enough trains on hand to support operations in the first hours of each day, until the first trains from the other end arrive and are turned around.

Castle Airport: Primary Maintenance Hub

The starter line will be between San Francisco, Los Angeles and Anaheim. AB3034 doesn't explicitly mention San Jose, but the primary reason Pacheco Pass was selected was to ensure that all trains would pass through South City's Diridon station. In addition, phase I will include the wye at Chowchilla and spur up to Merced.

There's been talk of siting the primary maintenance hub at the nearby but underutilized Castle Airport (formerly Castle AFB), which is adjacent to the BNSF right of way. CHSRA wants to prioritize the construction of a test track in the Central Valley, both to help FRA draw up a "rule of special applicability" to enable operations of non-compliant equipment at 220mph and, to help narrow down the list of track and train technology vendors.

So far, CHSRA is holding out for UPRR to cede part of its ROW along CA-99 through Fresno and up to Sacramento to obtain a straight alignment and access to downtown areas, but it's an open secret that rival BNSF has been much more receptive to overtures from the Authority. If CHSRA ends up sticking with BNSF from Bakersfield all the way to Modesto, it would be possible to site the Merced county HSR station at Castle Airport and implement a bypass around the town.

HSR Phase One Yards: Castle Airport

Unfortunately, Merced county lacks a major population center. Without any commercial flights (high speed cargo and/or passenger), an HSR station at the airport would struggle to attract enough ridership to justify running more than a a small number of trains per day to SF or LA/Anaheim in phase 1. That means early construction of a spur up to Castle Airport would primarily be for testing and maintenance. A secondary objective would be a show of good faith that the planned extension to downtown Sacramento really will be built in phase 2.

Here are a couple of videos showing SNCF's and JR's maintenance yards in Villeneuve (near Paris) and Tokyo, respectively. The equipment required depends on the train technology selected. Note that JR has special "Dr. Yellow" trains that measure track geometry, the condition of the overhead catenary system and verify signaling performance while running at 270km/h (approx. 170mph).

• SNCF's Villeneuve maintenance yard
  
  
• JR's Tokyo maintenance yard
  
  
• JR's Dr. Yellow series 923 shinkansen

Overnight Parking: Northern California

In addition, it's common for railroads to operate yards for overnight storage and at least minor maintenance (incl. thorough cleaning and provisioning) near both ends of major lines because that where trains naturally end up at the end of the day. In the specific case of the starter line for California's system, that means San Francisco and Anaheim, but throughput constraints on the Fullerton-Anaheim section mean that many trains will have to terminate in Los Angeles instead.

Considering that non-stop express runs between SF and LA are expected to take 2h38m, it seems reasonable to say that the first three hours of operation on weekdays have to be supported by trainsets that were parked overnight. Businessmen and tourists looking to make day trips between these cities will likely generate above-average demand for express service in the early morning (departure 6-9am) and afternoon (departure 3-6pm), with additional demand for red-eye service in the evening (departure 8-9pm). CHSRA's target is to build an infrastructure with sufficient capacity to last for 50, perhaps 100 years. Inevitably, that means planning for ridership levels well beyond anything that can be achieved in the first decade of operations.

In particular, CHSRA insisted that it needs capacity to run 12 trains per hour (tph) each way through the new Transbay Terminal (TBT) in San Francisco. The issue was discussed extensively on this blog, leading to these general conclusions:

• SF wants to stick with the current, imperfect design to secure a slice of the HSR funds in the ARRA (aka H.R. 1-111th aka the stimulus bill).
  
  
• HSR trains can be actually turned around in less than 10 minutes, provided the TBT is considered a terminal station on a route that nominally begins and ends in Southern California, rather than an old-fashioned "Grand Central of the West" terminus station. Unlike steam and diesel locomotives, EMU trains do not need to be reprovisioned with fuel and water, nor do they need to be brought up to operating temperature before departure.
  
  
• the station throat and tunnel can be improved with relatively minor modifications, there is no need for a radically different solution. Given that Caltrain will also heavily use the TBT, headways inside the tunnel for both services combined could be as short as 3 minutes during peak periods such as morning rush hour. That implies a combined total of 20tph, though such a schedule would leave little room for recovering from a delay caused by an off-design condition.
  
  
• given the long walk to the nearest BART station, there is considerable doubt that the TBT could attract enough HSR ridership to justify operating 12tph. For long-term capacity planning purposes, a target of 8-10 HSR trains per hour (each way) was suggested, especially in light of the fact that HSR trains are up to 1320 feet (400m) long and that bi-level rolling stock officially rated at 200mph has been in commercial operation in France since 1996. Each full-length TGV Duplex offers 1090 seats, about 8-9 times the number on the Boeing 737s operated by Southwest Airlines. It will be a while before duplex HSR trains are needed in California.
  
  
• if need be, some northbound trains could terminate in Millbrae or else in San Jose and return south the next morning, without ever having reached the TBT.

The TBT itself will feature six full-length platforms, of which four are supposed to be reserved for HSR trains. This hard allocation relates primarily to differences in platform height, an issue the two operators ought to resolve by seeking the requisite waiver from CPUC. As things stand, this would imply the ability to store four full-length trains overnight at the downtown station.

Assuming a frequency of 10tph for the first 3 hours of each weekday and, that each train could one day be a full-length assembly of two trainsets, that implies a need for a yard near SF with space for at least 26 full-length trains - preferably 28 so one can be a spare while another undergoes maintenance. Caltrain's 4th & King station does not have enough capacity and anyhow none to spare.

There simply aren't that many locations where so many full-length trains could possibly be stored and secured near SF, so it makes sense to buy the land now to avoid a capacity bottleneck down the road. One obvious option is the huge old SP yard at Bayshore (h/t to Clem), which is undeveloped but part of a development project (h/t to DaveO) that does not yet account for HSR. The location is still being decontaminated but actually has space for at least 50 full-length trains. Part of the area could be used to store Caltrain equipment or, as a transshipment facility for High Speed Cargo. All this is assuming the City of Brisbane is interested in once again hosting a rail yard, perhaps with development of the airspace above.

A big issue is that CHSRA currently intends to switch from the inside tracks south of Bayshore to the outside tracks north of it because of the four Caltrain tunnels. However, accommodating UPRR may yet force CHSRA to switch to the two western tracks instead - the issue of track assignments along the Caltrain ROW is fundamental but not yet resolved.

HSR Phase One Yards: Bayshore yard

A second-best alternative to a Bayshore yard might be found in the relatively undeveloped southern part of Mission Bay.

Either way, HSR trains could make the short run from the yard to the TBT in the morning (and vice versa at night) or else, some trains could make Millbrae/SFO their first stop of the day. While CHSRA is pessimistic about boardings/alightings at that station, a number of commenters have pointed out that it would be cheaper to build and use multi-story car parks there than in downtown SF once the area around the TBT is converted into a transit-oriented district. In addition to customers hailing from the north peninsula and central East Bay, residents of western SF may well prefer to ride BART out to Millbrae rather than board at the TBT.

If San Jose Diridon is expanded to two levels, there should be space for an additional 4-6 full-length trains there. The Merced county maintenance facility could easily contribute another 12. All in all, Northern California would have more than enough overnight parking, maintenance and HSC capacity if CHSRA acquires the Bayshore yard. It should not matter that BART will usurp the entire Newhall yard in Santa Clara.

Overnight Parking: Southern California

The number of trains traveling north between 6 and 9am need not be the same as the number traveling south. Over the course of a staggered 15-hour day, any train can comfortably travel twice in each direction, such that it would always be parked overnight in the same yard. That means the division of trainsets between northern and southern California can be based on passenger demand during the peak periods. It also means crews can be operated in two shifts of 7.5 hours each. Depending on the contract, weekend and holiday service could be implemented via a scheduled 45-hour six-day work week and a total of six weeks paid vacation per year.

In the absence of directional ridership data, I will assume a requirement to operate 10 tph out of LA and Anaheim combined to SF during the morning peak period, which spans 3 hours in LA but 3h30m in Anaheim, incl. a 10 min layover at Union Station. LA Union Station will feature a second level with perhaps six full-length platform tracks for HSR plus run-through tracks on aerials. Anaheim ARTIC will have room for perhaps four full-length platform tracks. That means one or more yards for an additional 21+2 full-length parking spaces may be needed.

One option would be a new yard just west of ARTIC in a utility ROW between I-5 and South Claudina Way, sufficient for 13 tracks in addition to the electricity pylons and telephone poles already present. Normally, trains would only use the turnoff toward ARTIC. Note the brown line representing a potential extension of the Metrolink network in Orange County, linking LA US, Long Beach airport, Disneyland, ARTIC and John Wayne Airport (zoom out to see the route). Some tunnel sections would be involved, in those the locomotives should switch to third rail electric power (e.g. 1500VDC).

HSR Phase One Yards: Anaheim

However, note that the Fullerton-Anaheim section will be dual track and limited to about 3 HSR trains per hour. That means only 7+2=9 slots will actually be needed, leaving plenty of room for the utility poles and pylons plus two potential Metrolink tracks.

That implies a further 14 spaces would be needed in Los Angeles. One option would be a second level above the Metrolink yard at South Santa Fe, just south of Union Station. This could actually accommodate up to 16 full-length trains.

HSR Phase One Yards: South Santa Fe

The snag is that an elevated yard would be very expensive to construct. A better option may be an at-grade yard north of LA Union Station, next to North San Fernando Rd. It may be possible to park 14 full-length trains in that triangular patch of land.

HSR Phase One Yards: North San Fernando

An additional issue is that there will be local HSR trains within Southern California once the phase 2 spur to San Diego is built, so additional yard space will be needed. Palmdale airport would be a good candidate, especially if a new passenger terminal with an integrated HSR station is constructed approximately half-way in-between Palmdale and Lancaster. It should be easy enough to design platform + storage tracks for e.g. 20 full-length trains at this new transportation hub. In practice, fewer may be needed, this depends in part on if and how CHSRA and DesertXPress choose to integrate their services.

HSR Phase One Yards: Palmdale

Conclusions

Finding adequate yard space for Phase 1 looks feasible, though there are constraints in both Los Angeles and Anaheim. However, CHSRA doesn't have a whole lot of viable options if it wants to keep the total number of yards manageable. That means it should purchase the Bayshore yard near San Francisco plus land for a new North San Fernando yard and a third near South Claudina Way in Anaheim sooner rather than later.

Yard space at Castle and Palmdale airports should be cheap and plentiful, but its unclear how useful that would be early on. There are currently no firm plans to build a new passenger terminal with integrated HSR station at either one, though LAWA is on the hook to some extent because it used its influence to ensure the route would run past Palmdale rather than across the geologically more challenging Grapevine. Developing the currently-defunct Palmdale airport into a facility that attracts large numbers of passengers and airlines depends critically on a sound development plan. Right now, LAWA is looking at using part of the land for a solar thermal power plant, which could potentially cause glare problems. There are also plans to develop the Antelope Valley, i.e. north-east LA county, via conventional low-rise sprawl rather than transit-oriented high rise towers that are more easily cooled and supplied with non-potable recycled water in addition to a constrained supply of potable water.

The Keys to the Magic Kingdom

One of the relatively late changes to the definition of the HSR starter line was the inclusion of the segment from LA Union Station to Anaheim ARTIC, OCTA's planned multimodal transit hub. Nothing wrong with that at all, after all Orange County is the second most populous in the state. Besides, OC needs the grade separations along BNSF's core "91" line to support future capacity growth for container freight out of the LA/LB harbors. A similar billion-dollar project is to improve grade crossings with UPRR's lines in the San Gabriel valley, dubbed Alameda Corridor East, is already underway. Grandiose plans for using fancy, unproven freight maglev technology to shuttle containers to a new Inland Port in Mira Loma appear to have bitten the dust and those for maglev to Vegas along with them.


Anaheim ARTIC on YouTube

The hoped-for quid pro quo for grade separation is that BNSF will offer CHSRA enough land to run dedicated tracks between Redondo Junction - the northern terminus of the successful Alameda Corridor - and Fullerton. To date, BNSF has been receptive to CHSRA overtures, but that was in the Central Valley. The company's ROW between Richmond harbor and Bakersfield is arguably not fully utilized, much of the alignment is single track.

The "91" line, named - this is Southern California for you - after the adjacent freeway through the Santa Ana river valley, is a whole different ball of wax. It is the only ROW out of California that BNSF actually owns outright. All its other routes out of the state depend on trackage rights on competitor UPRR's network. Up to a point, those could be restricted or canceled without falling foul of anti-trust laws. The price for the trackage rights is also subject to change.

Therefore, we should expect BNSF to take a much tougher line in negotiations regarding the sale of part of the "91" ROW. The Alameda corridor currently carries about freight 40 trains a day, but sooner or later that may double, even triple. With prescient investment in modern signaling, diligent track maintenance and professional operations, two track ought to be enough for BNSF to handle its share of that. But what if there is an accident or an earthquake. UPRR has multiple rights of way that it can use to re-route its trains. The companies may well have secured mutual emergency trackage rights on each other's lines. Still, BNSF risks putting itself at something of a competitive disadvantage if it sells part of its most precious asset in California. Once you give up a right of way, you never get it back. Not ever. BNSF may want to be a good corporate citizen and it knows that grade separation will reduce the risk of accident-related disruptions to its own service. But does that mean CHSRA will get to lay down tracks? We'll see.

Between Fullerton and the OC/SD county line near San Onofre, the right of way is owned by the Southern California Regional Railway Authority (SCRRA) which operates Metrolink. Amtrak Pacific Surfliner and some BNSF freight trains also use this alignment. BNSF owns the connector that runs north from Orange to the "91" line as well as the bit from National City to San Diego. The harbor down there is not a container terminal, it appears to be used for importing cars. All of these trains use FRA-compliant locomotives and rolling stock.

The SCRRA ROW in Orange County is quite narrow, really only wide enough for two tracks in most places. Widening it via eminent domain for the sake of laying down dedicated HSR tracks would be massively unpopular and ruinously expensive. The plan is therefore to seek FRA approval to operate mixed traffic, i.e. to add a limited number of non-compliant bullet trains to the mix via time separations guaranteed by signaling upgrades that prevent engineers from running red lights by accident. Top speed for this last stretch would be down to 79mph, but it's so short that presents no object. After last year's disaster at Chatsworth, HR 2095 mandated the installation of positive train control technology on busy rail corridor and those carrying hazardous materials by 2015.

The request for the mixed traffic exemption will be rolled into that for a "rule of special applicability" that CHSRA will anyhow need for its entire network. FRA currently has no rules at all for passenger train operations at speeds in excess of 150mph, nor any for the safe design and operation of 25kV AC overhead catenary systems. The agency had already begun to draft such rules for Florida HSR some years ago but they were shelved when Gov. Jeb Bush killed that project. They will soon be dusted off and completed, presumably many will be cribbed from the regulations Asian and European countries have already drawn up.

However, there's always a chance that one or more domestic - or at least North American - would-be HSR vendors will emerge. They might seek to gain a competitive advantage by getting FRA to draft rules that would force more established global players to modify their designs. That would be foolish from a safety/reliability/cost point of view, but protectionism-by-red-tape is one of the oldest games in town. Bureaucrats, too, love to draft special rules under the guise of "special conditions" that supposedly exist in the territory they are responsible for. In the US, some freight operators may well seek some regulatory overkill to push back against states that suddenly want to acquire part of their ROW, just so they can benefit from the federal largesse that is now being heaped - ok, spooned - onto high speed rail projects. Well-connected organizations that oppose HSR in principle, a particular project or impacts on their back yard might also try to lobby FRA, in the hope that this might kill or at least delay HSR implementations. We should be vigilant and push back against such selfish efforts.

The rules for California will set a national precedent for both true bullet trains running at well over 125mph on dedicated tracks and, for rapid rail below that speed, with freight and passenger trains each using the equipment best suited to their business model yet sharing a ROW and even tracks. Let's hope President Obama and Secr. of Transportation LaHood realize the importance of FRA's rulemaking on HSR in the years to come.

Clem Tillier over at the Caltrain-HSR compatibility blog has published the following diagram in his analysis of SF Peninsula Rail Traffic in 2030. It shows statewide train traffic on the HSR network, based on CHSRA's 2008 business plan. It does not include non-HSR trains that HSR will share track with in sections such as Fullerton-Anaheim.


Hi-Resolution Version

For Anaheim, CHSRA is apparently forecasting 2-3 trains in each direction per hour during a six-hour peak window. Of those, 60% would head up to the SF peninsula and 40% to Sacramento. There appear to be no plans to terminate trains in Los Angeles. Today, Metrolink's Orange county line between LA and Oceanside is served by up to 3 trains per hour each way during rush hour, but far fewer during the day. Amtrak Pacific Surfliner adds one more. If there were a 100% increase in legacy train operations by 2030, we'd still only be talking about 10-11 passenger trains running on the existing dual track alignment between Fullerton and Anaheim. That's a 6-minute headway, quite busy but not extremely so, provided that all trains run pretty much on time. I don't know how many freight trains run through Anaheim station today, let alone in 2030. That's a wildcard.

In a comment on the March 10 post entitled LA - San Diego: Quo Vadis?, yeson1a suggested that CHSRA might want to keep it simple and stupid (KISS): continue on to San Diego by staying on the 91 corridor from Fullerton through Corona and switching to the I-15 median there. This option has not been studied by CHSRA and would not be considered unless securing a ROW past Ontario airport proves impossible or too expensive. Yeson1a's alternative would leverage the investment already made for LA Union Station - Fullerton in phase I. A connector from Anaheim to Atwood would permit a capacity-defined subset of HSR trains to detour from the main line to Anaheim. If desired, some of the rest could stop at a no-frills auxiliary station in Fullerton. Norwalk/Santa Fe Springs would then only be served by some of the trains that make the detour into Anaheim. The route would probably feature and intermodal with Metrolink in Corona, with the option for a regular secondary station in Lake Elsinore, though Murietta would be fairly close by.


View Larger Map

The principal downside of this idea is that neither Ontario airport not Riverside are on the route. However, a possible phase II/III spur up to Las Vegas could give the downtown areas of Riverside, San Bernardino and Victorville HSR stations. Passengers hailing from Northern California would likely prefer a shorter spur off the starter line at Mojave. Perhaps both will end up getting built someday, with tracks joining up in Barstow. Such a full build-out would, however, require quite a bit of magic dust from the money fairy.