Turnout in 2008 was D +4.6.
Turnout in 2012 was D +1.6.
ALL the polls are using a bad turnout model, some worse than others.
Your citation conflicted with others I have read, so I went to look at the report.
You are misreading the numbers on page 6. That is not a measurement of party ID.
That is the percentage of eligible voters that voted for the Democrat or Republican candidate.
THIS is the party ID breakdown, over the past couple of decades:
-Turnout in 2008 was D +4.6.
Turnout in 2012 was D +1.6.
You are misreading that pdf. That is % of Party turnout, not % of Electorate.
Dems had a +6% in share of electorate.
http://ropercenter.cornell.edu/polls/us-elections/how-groups-voted/how-groups-voted-2012/