If Katrina takes refinery capacity off-line, the effect at the pump will be immediate. Ther is no more capacity available to up output within the counry. Shortfalls will have to be made up from imported gasoline suppliers. They are unlikely to be moved by complaints of gauging from American congressmen.
At this point Americans may ask why no new refinery has been built in the United States since 1976. (There is one on the drawing board for Yuma, Arizona.)
This, it strikes me, is remarkably similar to the problems exposed in the US transmission system a couple of years ago. Running something at 97% of capacity is a recipe for catastrophic failure because there is no margin to cover failure. When one designs highly available systems (and if oil refining is not something that needs high availability it is hard to think what does) a safety margin is always required and usually that safety margin is around 10-20% of average use. For the US today that implies a sensible safety margin would be roughly equivalent to the entire Louisiana refining capacity whereas the actual margin would seem to be approximately that of just a couple is Louisiana refineries.
Hugh identifies legislators as the prime reason for the lack of investment. I believe he is omitting a couple of groups that helped remove the spine from the legislators and the oil companies. The first group are the ambulance chasers lawyers. They probably didn't actually say anthing but in America's sue happy society a company would be nuts to try and build something as potentially polluting as an oil refinery anywhere new because of the chances of being blamed (and thus sued) for a zillion different health, environmental etc. concerns by the residents. The second group are the NIMBY residents themselves who want the oil refinery to be somewhere else and their aiders and abettors the envirowackoes who act like modern-day prohibitionists by trying to remove the supply of something that has enormous demand.
The good news about Hurricane Katrina is that, if it really does take out a couple of refineries, we will see serious shortages of petrol in parts of the USA. Given Hawaii's attempt to rewrite the law of supply and demand I'm going to predict that Hawaii will be the first state to run out of petrol but it will not surprise me in the least if we see problems elsewhere too. I put this as good news because it looks like the only way to start getting US consumers behind a couple of required steps. The first is to actually build some more refineries and to distribute them so that a single event will not wipe out many of them. The second is that it may help provide impetus for alternative fuels instead of petrol. If the US reduced its oil consumption to (say) 15M barrels/day then a whole boatload of good effects would probably occur including a reduction in the oil price and a very healthy diversification of energy sources so that US is far less dependant on one sort of energy or one supplier.
BTW Roger L Simon writes about how he was effectively unaware of Typhoon 11's recent progress over Japan. This partially ties in to EU Rota's graph question. Japan has very good typhoon tracking technology and expects to be hit by half a dozen typhoons each year and Japan's TV broadcasters seem to spend significant parts of news broadcasts listing the current state of typhoons in the Pacific. The result is that (as with the tsunami warning system) everyone is prepared for a typhoon to arrive as well as used to being pounded and this building appropriately - it also helps that Japan's typhoons seem to be more rain than wind and that Japan has some honking storm drains (it could also be that the typhoons are more predictable because there is a lot less land in the path I don't know).