See: http://www.renewableenergyworld.com/ugc/blogs/2016/04/rooftop_solar_could.html
This is a much higher estimate than previous studies came up with. I think this is because PV cells are more efficient and cheaper than they used to be, so it is more practical to use them more extensively. I believe 40% would be enough to retire all coal-fired generators and many natural gas generators. "Distributed" (rooftop) PV is difficult to account for. It produced ~0.3% of U.S. electricity at last year's peak (August 2015: 1,260 thousand megawatt hours out of 401,535 thousand megawatt hours): http://www.eia.gov/electricity/data/browser/#/topic/0?agg=2,0,1&fuel=vtvv&geo=g&sec=g&linechart=ELEC.GEN.SUN-US-99.M~ELEC.GEN.DPV-US-99.M~~~~~~~~~~~~~~~~&columnchart=ELEC.GEN.ALL-US-99.M~ELEC.GEN.COW-US-99.M~ELEC.GEN.NG-US-99.M~ELEC.GEN.NUC-US-99.M~ELEC.GEN.HYC-US-99.M~ELEC.GEN.WND-US-99.M&map=ELEC.GEN.ALL-US-99.M&freq=M&start=201212&end=201601&chartindexed=0&ctype=linechart<ype=pin&rtype=s&pin=&rse=0&maptype=0 Utility scale solar produced 2,834 thousand megawatt hours (0.7%). It seems reasonable to project that rooftop PV it could increase by a factor of 130, up to 40% of total generation. You do not see many PV roof installations in the U.S. See also: http://www.eia.gov/todayinenergy/detail.cfm?id=25512 http://www.eia.gov/forecasts/aeo/supplement/renewable/?src=home-b1 - Jed
