Agro-economic modeling suggests that cellulosic feedstocks may be grown on Conservation Reserve Program (CRP) acreage to help meet targeted biofuel volumes.  However, the biomass production capacity of CRP lands is currently unclear, as well as the associated environmental trade-offs.  Employing detailed CRP data, SSURGO, and the EPIC model (Environmental Policy Integrated Climate), we simulated growing switchgrass and a perennial mixture for ethanol on Iowa CRP already in grass cover, and compared it to the perennial mixture without harvesting.  Results estimated yield, carbon (C) flux, sediment erosion, and nitrogen (N) loss at a 56 m resolution over a 12 year time-span. On ca. 270,000 hectares of CRP, an estimated 3.2 Tg of switchgrass biomass could be harvested per year, compared to 2.1 Tg of the perennial mixture.  Across CRP units, yields for switchgrass varied from 4.5 to 20.5 Mg/ha, with an average of 11.9. Switchgrass and perennial-mixture harvesting reduced soil C stocks by an average of 2.9 and 1.4 Mg C/ha/year, respectively, over the 12 year period.  Similar to yields, changes in C stocks varied substantially across CRP units. Simulated erosion rates and sediment N loss increased with switchgrass production, but other N losses decreased.  Overall, our results suggest that harvesting switchgrass and a perennial mixture on selected CRP lands in Iowa could produce 1,226 and 796 million liters of cellulosic ethanol per year, respectively—ca. 2.0% and 1.3% of the 16 billion gallon cellulosic ethanol goal in the Energy Independence and Security Act of 2007.  Our study found a wide variation in both yields and soil C stocks across CRP units.  Further investigation of C loss and other environmental tradeoffs of placing CRP into biofuel production is needed.