California HSR should use Tejon Pass

The current California HSR plan has significant flaws; a major mistake was selecting a detour via Palmdale instead of the direct route via I-5 across Tejon pass to connect Bakersfield to Los Angeles. The cost of the Palmdale route -- about $32 billion in current dollars -- is more than twice that of the direct link, and nobody wants to pay for it. The HSR project will be restricted to the Central Valley until the less expensive Tejon route is embraced.

Thursday, November 10, 2011

Low-cost HSR for California

The revised financial plan for implementing the first phase of HSR issued November 1, 2011, further demonstrates the folly of forging ahead with a very expensive project without realistic financial support in a reasonable time frame. If the project is going to produce a usable system within a 20 year horizon, a complete revaluation of the project must be undertaken with the goals of imposing a cap on costs, establishing a clear order for building the segments, and setting definite targets for their completion.

Some observations:

1. There is still no credible plan for financing the first phase, from San Francisco to Anaheim. Even assuming that private funding and/or system revenue will generate $20 billion at some point in the future, the remaining $80 billion has to come mainly from CA state bonds and federal grants. If the project is stretched out to 50 years perhaps this is possible, but that pushes the benefits from building the system far into the future as well. What is needed is a clearer understanding of what is financially possible in the next 20 years.

2. Speed is expensive, and the consultants have consistently underestimated the cost of an alignment that will support train operation at 200 mph plus. The obvious solution to run-away costs of the proposed system is to reduce the maximum speed, perhaps as low as 150 mph, which will greatly reduce construction costs in the readily available corridors (mainly existing railroad routes).

3. Reduced speed means increased travel times, and therefore makes HSR less attractive to airline passengers. Penetration into this market becomes negligible above a 5 hour running time for SF-LA, which is why HSR planners have pushed for a time below 3 hours (realistically, 3.5 hours is the best that could ever be achieved on the proposed alignment). On the other hand, any running time below 5 hours is competitive with driving, and further reduction only marginally increases HSR ridership from this market (see figure). A maximum speed of 150 mph appears to be a reasonable target, since construction costs are only marginally less for a modern rail system operating at lower speeds, and these costs rise sharply for higher speeds as recent estimates have revealed. The bottom line is that it is too expensive for HSR to try to compete with air travel in California, and improved rail service should concentrate instead on reducing auto travel as much as possible. (This argument is based partly on an FRA study, "High-Speed Ground Transportation for America," U.S. Dept. of Transportation, Sept. 1997.)

Figure based on data taken from "High-Speed Ground Transportation for America."

4. In line with this argument, the first segment should at a minimum run from Bakersfield to Merced (about 170 miles) with a reduced design speed in the range 150-160 mph. This is a long enough segment to make electrification useful, and possibly allow through running of trains from northern California to Bakersfield with a switch from diesel to electric power in Merced.

5. On the basis of cost and immediate usefulness, the second segment should be the Tehachapi crossing, from Bakersfield to Palmdale. Metrolink seems confident that it can quickly implement improvements to the Antelope Valley line (including electrification) to allow operation of HSR trains from Palmdale into Los Angeles in less than one hour, and probably on to Orange County as well. This option should allow a running time from Merced to Los Angeles of around 2.5 hours, which is competitive with driving and probably flying as well. Interim service from northern California would either require a transfer at Merced, or through-running with a change of locomotives.

6. The Pacheco pass route is far too expensive to consider building in the near future, if ever. The major justification for this route is to put San Jose on the main line to San Francisco, and this goal is unnecessary since San Jose is a secondary market that can adequately be served by a branch line. The alternative is of course Altamont pass (the sensible option from the beginning of the HSR program), with a Bay crossing to San Francisco by a bridge near Dumbarton or (perhaps better) eventually by a tunnel near Oakland. This decision how to reach San Francisco can be deferred for now.

7. Improvements to the Altamont pass route are currently being studied in parallel with the main HSR project. One promising suggestion coming out of this study is to run the first phase of HSR up the San Joaquin Valley to Manteca, and then west to Livermore. Transfer in Livermore to BART (extended from Dublin) could be used to access Oakland and San Francisco, and transfer to an improved ACE to get to San Jose. This routing is no slower to San Francisco than the proposed "Bay to Basin" (San Jose to San Fernando) interim service, and a lot less expensive.

8. These considerations lead to proposing an initial HSR rail system than is considerably different than the one advocated by the HSR Authority. (Its options are restricted by mandates written into Prop. 1A, many of which need to be changed to implement a useful HSR line in the near future.) New HSR trackage would be restricted to a trunk line from Livermore to Palmdale, continuing on to Los Angeles and beyond using improved Metrolink routes, and accessing San Francisco and Oakland via transfer to BART. The trunk line would operate at a maximum speed of around 150 mph, but be built to retain the option of increasing the maximum speed at a later date. An expanded (and improved) ACE line would be geared mainly toward regional rail service, but properly designed (and electrified) it could provide through running for HSR trains to San Jose and Sacramento. The relevant Metrolink lines (Antelope Valley, Orange County, and Riverside) likewise would be improved and electrified to permit through running of HSR trains to multiple destinations in the greater Los Angeles basin.

9. Construction cost for the proposed alternative should be under $15 billion (ignoring inflation and interest), about half of the cost of constructing the same route to a design speed of 220 mph. For comparison, the cost of the truncated "Bay to Basin" phase one is around $45 billion, and the full phase one around $65 billion (again ignoring inflation and interest). Ridership on the low-cost alternative should be more than 50 percent of the full phase one, and this can be achieved at one-quarter the cost. This demonstrates that it is possible to build a useful HSR line at an affordable price.

10. By keeping the investment in HSR as low as possible, it is possible to simultaneously expand and improve local and regional rail services. The cost of modernizing Caltrain, ACE, Capitol Corridor, Metrolink, and Pacific Surfliner, and restoring the Coast Daylight will exceed $10 billion, and concentration solely on HSR will detract from these equally important programs.