The Union of Concerned Scientists points out that switching from an S.U.V. that gets 14 miles per gallon to one that gets 16 would save the same amount of fuel as swapping a 35-mile-a-gallon car for a 51-m.p.g. new generation gas-sipper. This is not an argument for more S.U.V.’s. It simply shows that we can wring savings from modest efficiency gains in products we already use.Since this is somewhat counter-intuitive, I thought a brief explanation might be useful.
The confusion stems directly from the units used to express fuel economy in America. "Miles per gallon," while perhaps clearer for consumers, is a difficult framework in which to think about resource consumption (since the resource - gallon of gas - is, by definition, fixed).
On the other hand, consider the alternative -- flip the term and talk about gallons of fuel used per mile traveled. In the example above, an SUV that gets 14 mpg burns 0.0714 gallons of fuel per mile traveled, or, to make things easier, 7.14 gallons of fuel per 100 miles traveled. The car that gets 35 mpg burns 2.86 gallons of fuel per 100 miles traveled.
First of all, from the perspective of limiting absolute energy consumption, the car is clearly a better choice, and that fact should not be lost in this discussion.
But consider the impact stemming from relative improvements to the two vehicles. If I improve the SUV's fuel economy from 14 to 16 mpg, my fuel consumption per 100 miles traveled drops by almost a gallon - from 7.14 to 6.25. In other words, I have saved one gallon of fuel as compared with my baseline scenario. To get the same improvement from the car's baseline, I have to decrease my fuel consumption from 2.86 gallons per 100 miles traveled to 1.96, corresponding to a fuel economy improvement to 51 mpg:
The optimal scenario would of course be to upgrade all your vehicles to 51 mpg (or more). However, given that such an ideal is often not practical or reasonable, comparative analysis as described here can be very beneficial from political, business, and personal perspectives.
The Energy Analysis blog has a similar post about this issue, and includes the following relevant graph showing the non-linear relationship between fuel consumption and mpg: