From what I understand, food-miles is being used as a proxy for carbon output (and associated pollution), under the assumption that it takes fuel to move things, so food that travels a long way must use a lot of fuel.
A much more direct measurement of carbon output is "food cost", since the fuel used to move the food -- as well as the energy used to grow the food -- has to be paid for, and the people bringing the food to you price it high enough to cover those costs. It turns out, shipping food thousands of miles by train or barge doesn't use much energy relative to the amount of food shipped; as the article mentioned, most of the energy usage comes from the consumer, not the grower or shipper.
