The detection of a sub-surface present-day ocean on Europa is of considerable interest. One possible method of detecting an ocean is by an orbiting radar sounder. The effects of a range of possible Europan ice chemistries on radar attenuation are investigated, using plausible Europa ice temperature profiles. Ice chemistries are derived from geochemical models of Europa predicting a sulfate-dominated ocean, a chloride-dominated ocean scaled from the Earth, and on experimental data on marine ice formed beneath ice shelves on Earth, on low-salinity sea ice and models of rock and ice mixtures. Chloride ions are expected to dominate the radar absorption because they are incorporated into the ice lattice, though if freezing rates are rapid or similar to sea ice, then brine pockets will dominate losses. In the case of an ocean being present underneath the ice, the range of attenuation found in the models is from about 5 dB/km for rock/ice mixtures up to 80 dB/km for sea ice models. However, perhaps the best model at present is for ice formed from a plausible sulfate-dominated ocean with the fraction of chloride incorporated into the ice set to the same as for low accretion rate Ronne Ice Shelf marine ice. This has a radar absorption of 9-16 dB/km for surface temperatures of 50-100 K. In the case of a convecting isothermal ice layer beneath a conducting ice lid, absorption in the conducting lid is lower for all the models than it is over an ocean as the convecting ice is modeled to be 250-260 K. Absorption in the isothermal layers is very high, but the interface between conducting and convecting ice may be marked by a reflection coefficient that enables it to be imaged. It is concluded that realistic ice-penetrating radars are likely to be able to penetrate some kilometers into the ice, though problems of interpretation caused by scattering are not considered here. © 2000 Academic Press.