President Obama has proposed a $13 billion dollar high speed rail initiative. Based on the statements accompanying his announcement, it would appear that Obama's primary motivation for this program is economic stimulus, although he did state that "we must start developing clean, energy-efficient transportation that will define our regions for centuries to come."
So, how efficient is rail transportation? Surprisingly, not particularly. While the famous CSX claim that railroads can transport a ton of freight 436 miles on a gallon of fuel is factually correct, this outstanding efficiency would not seem to carry over to our passenger rail system. In 2005, Amtrak reported an energy intensity of 2,935 BTU/passenger mile. Using 129,500 BTU/gallon for diesel #2, this is about 44.1 passenger miles/gallon. If I'm on the freeway in my Land Rover LR3 with a passenger in my car, I do as well or better (net of the conversion from the energy equivalent of diesel fuel in automobile gasoline).
On the other hand, very high figures for passenger miles per gallon can be found. For example, on the French TGV Duplex on the Paris-Lyon route with two intermediate stops, the train consumes 17.65 kilowatt hours/train-kilometer over a distance of 427 kilometers, or a total consumption of 7536.55 kilowatt hours. This is 2.7132x10^10 J (joules). This train has 545 passenger seats and if we assume an occupancy factor of 0.8, we see that the train converts 0.0404817 kilowatt hours/passenger kilometer.
Finding an equivalent passenger miles/gallon figure is fraught with difficulty, but let's proceed undaunted. I'll assume a diesel generator with efficiency of 50%, we find that the generator would use 276.26 btu/passenger kilometer or 444.6 btu/passenger mile. This is the fuel in 3.4332*10^(-3) gallons of diesel fuel. This is 291 passenger miles per gallon of diesel fuel. Wow, big difference!
What explains this huge difference? Part of it is the load factor. In fiscal year 2007, Amtrak's load factor was 48.9%, so increasing this to 80% (though this article states that a load factor of 65% technically makes a long distance train sold out, because allowances have to be made for entraining/detraining passengers at intermediate station stops) would effectively increase the passenger miles/gallon to (80/48.9)*44.1 or 72.1 passenger m.p.g. This is still a sad comparison to the TGV Duplex and not much better than airline service. I haven't been able to explain the remaining discrepancy.
The geographical areas contemplated by the Obama plan would seem to make sense for the types of route on which the TGV, for example, excels. That would be routes of intermediate length connecting the center portions of relatively large concentrations of population. There are further advantages of new rail over new road. Most importantly, for a given capacity, the amount of land required for rail is much smaller. This assumes, of course, that that capacity is actually used but the same can be said for passenger cars.
On the other hand, evaluating only from an energy efficiency point of view, it would seem that intercity bus service with a high occupancy factor is easily the best option. And to the extent that people could be induced to leave their cars for buses, the need to build new infrastructure could be avoided. This may be one of the reasons why it doesn't happen.