At The Upshot Nate Cohn explains “How Black Turnout Could Decide Senate Control.”

Black voters will play an outsize role in this year’s fight for control of the Senate. Mississippi, Louisiana and Georgia are the three states where African-Americans represent the largest share of the population, and North Carolina isn’t too far behind, at seventh on the list.
According to the Census Bureau, the black share of the national electorate has dropped off in every midterm election since at least 1998. That could cripple Democrats in these states.
…But it is possible to imagine the black share of the electorate holding at 2012 levels in those states, or even increasing. It has happened before.

Cohn shows that Democrat Mary Landrieu “benefited from a small but surprising increase in the black share of the electorate, which rose by 0.6 of a percentage point between 2000 and the 2002 runoff…” and that she would have lost if it had dropped 2.7 percent, as it did in 2010. Landrieu also benefitted by a relative decline in turnout of white voters.
In Georgia, Cohn notes that “1998, black turnout in Georgia increased by 1.7 percentage points above 1996 levels.” Cohn speculates that the uptick was caused by “impeachment proceedings against President Clinton, who was popular among black voters.” Not so sure. The small increase could just as likely be attributed to better GOTV, or a combination of other factors. He notes also that the turnout record during the same period in NC was inconclusive — until Obama ran. Cohn believes the Obama factor might help Democrat Kay Hagan hold her senate seat, and in GA:

…There may still a narrow path for the black share of the electorate to rise beyond 2012 levels in 2014. In Georgia, the black share of the electorate could increase to 31.7 percent, up from 30 percent in 2012, if black turnout remains as high as it was in 2010 and if white turnout falls as it low as it did in 1998, when there was a competitive governor’s race. And although 2012 might represent an unusually high baseline, President Obama’s appeal among black voters also raises the possibility of an unusually high black turnout.

Cohn concludes that “there is room for high black turnout or low white turnout to upset the conventional wisdom. In a close contest in the racially polarized South, such a shift in racial turnout could easily be decisive.”
For Dems, the implications should be clear: Put more resources into registering and turning out African American voters, particularly in GA, NC, LA, AR and MS, but also try to get a little bite of the high turnout white seniors.