The simulation of ice accretion on a wing or other surface is often required for aerodynamic evaluation, particularly at small scale or low Reynolds number. Although there are commonly accepted practices for ice simulation, there are no established and validated guidelines. The purpose of this paper is to report the results of an experimental study establishing a high-fidelity, full-scale, iced-airfoil aerodynamic performance database. This research was conducted as a part of a larger program with the goal of developing subscale aerodynamic simulation methods for iced airfoils. Airfoil performance testing was carried out at the ONERA F1 pressurized wind tunnel using a 72 in. (1828.8 mm) chord NACA 23012 airfoil over a Reynolds number range of 4.5x10⁶ to 16.0x10⁶ and a Mach number range of 0.10 to 0.28. The high-fidelity ice-casting simulations had a significant impact on the aerodynamic performance. A spanwise-ridge ice shape resulted in a maximum lift coefficient of 0.56 compared with the clean value of 1.85 at Re = 15.9x10⁶ and M = 0.20. Two roughness and streamwise shapes yielded maximum lift values in the range of 1.09 to 1.28, which was a relatively small variation compared with the differences in the ice geometry. The stalling characteristics of the two roughness ice simulations and one streamwise ice simulation maintained the abrupt leading-edge stall type of the clean NACA 23012 airfoil, despite the significant decrease in maximum lift. Changes in Reynolds and Mach numbers over the large range tested had little effect on the iced-airfoil performance.