The YoY decline in February was blamed on the snow, and there might have been some extra driving in March once the weather improved.The map above shows the percentage changes in miles driven, by state, Feb 2010 over Feb 2009. The scale runs from -10% (red) to +10% (green). That's some strangely widespread snow...
Then here's the same thing for March 2010 over Mar 2009:
Apparently almost the whole country just up and decided to start driving again. Strange.
If we look at gasoline prices, there's no real help there either:
That's the year on year changes in price. Nothing particularly dramatic happens in Feb/May that could explain the change in driving.
So this remains a mystery. One has to wonder about some kind of data glitch.
And, yes, I made the maps myself as an exercise in using some Perl scripts to output KML, which I then read into Google Earth to make the screenshot. Pretty soon I'll know enough to be dangerous...