and the train has the daily ridership of 2,500 people ...

2,500 x  2,784 = 6,860,000 BTU  (Train)

2,500 x  3,445 = 8,612,300 BTU  (Car)

This is what Bundy, et. al., is trying to obscure.  Differences don't seem so different when they are put on a per/unit basis.  To accurately assess the situation you need to look at the population (users) of the differing systems and establish the final figures.  

Here for more info regarding the Handbook.

Look at the supporting figures for the 90 year CO2 payback.  A lot of games can be played with long-term projections and extrapolations.  The most common is to load one side, such as light rail, with every conceivable cost and ignore costs for the side the argument supports.  Example, the cost of maintaining and repairing highways and other roadways is conveniently forgotten by the anti-rail people so if you are talking about establishing a light rail system it is legitimate to compare to the cost(s) of establishing alternative transportation systems - such as automobile/highway system.  (Don't forget to include inflation -- a billion don't buy as much as it used to.)

The way you write something is as important, so would say it is MORE important, as what is said.   For example, their own figures show the train has a 124% BTU efficiency compared to cars.  (Train BTU divided into Car BTU) using 0% efficiency as the baseline.)  

(ATinNM.  Sophist for Hire.  Have gun, will travel :-)