This explains a lot. Census data via Capitol Alert:
Just 50.1 percent of California's 27.4 million voting-age residents were registered to vote for last year's election, higher only than Hawaii's 48.3 percent, the state-by-state breakdown of political participation found. The national rate was 59.8 percent. The state's voter turnout, 39.2 percent of voting-age residents, was one of the lower rates, but not the lowest. Ten other states had lower voting levels with Texas, at 31.4 percent, the lowest; the national rate was 45.5 percent.
It's not quite as bad as it looks. One reason for the low participation is that many of the state's residents are not citizens (just 64.5 percent of voting-age Latinos, for example).
*Update: Hold everything - the Census Bureau seems to have gotten the numbers wrong. From Capitol Alert:
Although the Census Bureau had said 13.8 million Californians were registered to vote last year, the Field Poll's Mark DiCamillo noted that actual registration, as compiled by the secretary of state's office, was 17.3 million for the November 2010 election. State election officials also calculated the potentially eligible voter number at 23.6 million, nearly a million higher than the Census Bureau's number. So obviously the Census Bureau's figure was numerically wrong - by a lot. But was it statistically wrong as well?
The Census Bureau didn't use official registration data in its survey. Rather, it extrapolated the number from its monthly "current population survey." In other words, it asked those in the survey whether they were registered and had voted and then massaged the results to what its demographers believe to be an accurate cross-section of the state's residents.