Samples of Ells River bitumen sand from Alberta, Canada were measured at low frequencies (0.2–205 Hz) to determine the temperature and frequency dependence of velocities and attenuations. The samples were first measured “as received” where the pore space is mostly filled with bitumen but also contains small amounts of air and water. With residual air in the pores, at 5°C, there is strong dispersion in the P-wave modulus and a peak in attenuation at seismic frequencies. The frequency-dependent moduli and attenuations shift by three orders of magnitude in frequency as temperature is increased from 5°C to 48°C, consistent with the bitumen viscosity. Samples were then saturated so any empty pore space is filled with water. After saturation, at 1 Hz, increasing temperature from 5°C to 49°C causes a 30% reduction in the saturated P-wave modulus, a 34% reduction in the saturated bulk modulus, and a 6% reduction in the shear modulus. This behavior can only be explained by the temperature-dependent bulk modulus of bitumen. The results enable predictions regarding the P-velocities that can be expected during seismic monitoring of thermal enhanced oil recovery processes. Velocities for cold bitumen sand are near at reservoir pressure and temperature. Following steam injection, velocities should be very low (near ) in heated zones more than 50°C with a free gas phase, which could be steam or gas. There will be a progressive reduction in velocities, i.e., at 25°C and at 49°C, in areas of formation heating, but without steam or gas in the pores. Albeit smaller than the effect of steam, the effect of formation heating alone is large enough to be easily detected by today’s 4D surveys. With local rock physics calibration, it should be possible to map the areal extent of formation heating using 4D seismic data.