Several alternate fuels have been suggested for gasoline, probably the most common are natural gas, ethanol, and hydrogen. I'm mostly going to ignore natural gas for the moment, because in the long-term it has the same problems as gasoline; net carbon dioxide production and finite supply. It also has other problems for vehicle fuel use (similar but not as extreme as some of the problems hydrogen has). Into the mix there's also vegetable oil, specifically soybean oil (because it's cheap). Mix it with a little diesel fuel oil and you've got biodiesel (but I'm not sure there's any compelling reason to mix it in the first place, other than additives to make it a little more engine-friendly or some such).
Of the set, soybean oil is the most attractive. It's reasonably cheap (about $3/gallon), easy to produce, easy to store and transport, and has decent energy density. It also has the advantage of being usable in diesel engines, which are more efficient than gasoline engines. The numbers:
|Fuel||Energy Density (MJ/kg)||Specific Gravity (kg/l)||Energy per Gallon (x gasoline)|
The thing that really stands out here is how lousy a fuel hydrogen is as a drop-in replacement for gasoline. Even ignoring the difficulty of storing and transporting it, the complete lack of existing infrastructure, and the difficulty of economically producing it, the energy density per volume is terrible. Pretty much every other fuel is better. Pure ethanol, while being fairly cheap, isn't a whole lot better. Methane (the primary component of natural gas) is a little better. After crunching the numbers, the obvious biofuel seems to be soybean oil, which has properties similar to heavier diesel fuels.
So the question is, how much would we need? Right now, the U.S. consumes about 130 billion gallons of gasoline annually, and about 45 billion gallons of diesel fuels. If we convert all of that to soybean oil (keeping in mind that diesel engines are approximately 40% more efficient than gasoline engines), that comes to something like 85 billion gallons of soybean oil annually. Given an approximate (conservative) yield of 50 gallons of soybean oil per acre, that would require about 1.7 billion acres of farmland dedicated to soybean production, or 2.66 million square miles.
That's a bloody awful lot of land, over two-thirds of the area of the United States (including Alaska). Which kind of blows the whole thing out of the water. By switching to rapeseed (which has approximately 2.5 times the yield per acre in terms of oil), you can get that down substantially, but the amount of land needed is still huge. Clearly traditional farming methods aren't going to cut it -- it's going to take genetic engineering or even more radical methods to fulfill that kind of demand. So much for all of the advantages of vegetable oil.
So, how about ethanol? Given the fact that this would be used in traditional gasoline engines (I don't know if it's possible to use ethanol in a diesel engine) and the general low energy density, it would take approximately 345 billion gallons(!) of ethanol annually to replace our current consumption of gasoline and diesel fuels. Approximate ethanol yields would probably come to about 500 gallons/acre (using corn), which would require about 690 million acres, or approximately 1.1 million square miles, which isn't any better than rapeseed oil.
Which gets us back to natural gas. To replace our current rate of petroleum fuel use, we'd need about 225 billion gallons of liquid natural gas annually in addition to the 22 trillion cubic feet (approximately 155 billion gallons in liquid form) we already use per year. Which doesn't actually address the practical difficulties with pressurized storage and transport given our current infrastructure.
The upshot -- it's not a pretty picture.
[Update: my math was a bit off converting from cubic feet to gallons, the actual number was not 800 billion gallons, but the corrected 155 billion gallons. Increasing production by 145% could conceivably also prove somewhat difficult.]