A quick back off the envelope calculation suggests that the current US coal production (responsible for roughly 20% of the US energy consumption vs. approx. 40% for oil) is indeed equivalent to about half of the oil consumption in the US.  Thus 10 million barrel/day sounds quite right to me. The actual number would be probably lower due to conversion losses. Does anybody has the EROEI numbers for Fischer-Tropsch?
Waldi's right, I'm wrong. I was remembering the graph for electric power sources where coal was as important as everything else put together, not the graph for coal as an energy source.
We would have to mine twice as much coal to replace oil imports. Also, since coal is CH and oil is CH2 you get a lot of byproduct CO from the H2 production, so you burn that in gas turbines (after you clean it, just like you clean H2S out of gas) and that is what we will use to replace gas production as it declines.
I should have remembered that.
Waldi
I don't know the EROI but I can guesstimate a greenhouse intensity figure.  Choren Industries say their SunFuel is 10%  as greenhouse intensive as petroleum diesel.  Princeton University give a range of figures for FT diesel from coal with the middle value about 150% of petrodiesel. If my arithmetic is correct that means FT diesel from coal is about 15 times more net GHG intensive than from biomass, assuming no CO2 capture.  I think a carbon tax would be good insurance to  ensure coal-to-liquids producers stayed 'clean'.