Knowledge of the spatial distribution of bed lubrication regimes, i.e. frozen vs wet conditions, is crucial for understanding ice-sheet flow. Radar sounding can probe differing reflectivities between wet and frozen beds, but is limited by uncertainty in attenuation within the ice of bed echoes. Here we present two methods to estimate attenuation: (1) wide-angle radar sounding, in which source and receiver locations are varied so as to vary propagation path length, and thus echo amplitude; and (2) profiling, inwhich similar variations are obtained by sounding through varying ice thicknesses (assuming constant bed reflectivity). Siple Dome, West Antarctica, provides unusually favorable circumstances for application of these methods: the bed beneath Siple Dome is flat and uniform in its radar reflectivity, while ice thickness varies by several hundred meters. Wide-angle data 4 km from the summit yield an estimate for characteristic attenuation length of 124 m (35 dB km–1 loss), whereas profiling yields an estimate of 168 m.The difference between estimates is modest compared to the range of attenuation lengths reported in the literature. It may nonetheless prove informative by bounding effects of two ice properties to which the methods respond differently: (1) wide-angle sounding sampled relatively warm (lossy) ice beneath the summit, whereas the profiling method sampled relatively cold ice beneath the flanks as well; and (2) strain-induced crystal orientation fabrics and resulting dielectric anisotropy in the ice would vary from summit to flank, and may influence wide-angle sounding more strongly than profiling.