If one  further looks at the numbers that the EIA site provides. There is a slight discrepancy between the tabulated data and that plotted, and it took me a minute to realize that the curves are showing the rolling four-week averaged data, rather than the actual weekly for the imported gasoline data (and presumably also for the rest).  This is of significance since such an average, in this case with the changing situation, initially masks the fact that the imports for the past three weeks have gone from 0.938 to 1.207 to 1.423 mbd.  One may assume (at least I am) that about 500,000 bd of this is coming from the 20 mbd of gasoline that was put up by the IEA partners in response to our request following Katrina.  If this is the level of support that can be anticipated, then it will last some 40 odd days before that source is gone.