We've seen any number of calculations of energy efficiency and green house gases in transit modes. Some a bit out there because of crazy assumptions for autos. Even our favorite libertarian O'Toole played the game. The problem with all of these is that they don't consider the whole picture, or what happens when buildings are built closer together and transportation makes it easier for people to walk. But I digress.
At Rail~Volution I saw a presentation by Tina Hodges at the FTA that had some cool charts and comparisons of modes. The one I've seen before is the increase in VMT versus what CAFE standards will do. Now we've seen that there is a bit of a drop recently due to the economy but with gas prices as they are and no change in habits, I still believe this will happen.
Then here is the difference between current occupancies vs. all of the vehicles full and over the lifecycle of the vehicle in the second chart. These are based on a UC Berkeley study by Mikhail Chester that considered vehicle construction, guidway construction etc. The list of items lifecycled are at the link. Apparently buses off peak are the worst and peak are the best, even better than rail lines. Yet rail lasts longer and attracts more passengers overall so on average is better. I didn't really have time to read the 332 page tome, but if you're interested go for it.
But the most interesting in the presentation to me was the difference between the Heavy Rail modes. BART is the most efficient while Cleveland is almost as bad as a single occupancy vehicle. The relative inefficiency of the EL was surprising to me as well.
Thought this would be of interest to folks. I have to say again that its necessary to not just measure the lifecycle and modes but rather the land uses and transportation, but its interesting to learn that this work is being done.