Differential attenuation (Δt*) measurements of teleseismic P and S waves are made in the frequency range 0.05--5.0 Hz across North America using broadband seismic data from the Department of Energy's Regional Seismic Test Network and Lawrence Livermore National Laboratory's Basin and Range network. Data from four events have been analyzed by first assuming frequency independent Q. The observed regional variations for (Δt*) are consistent with previous studies and show greater attenuation for stations in the Basin and Range relative to those in shield regions of North America.

A significant component of apparent P attenuation appears at frequencies above about 0.5 Hz in both the Basin and Range and shield which reduces the ts*/tp* ratio to values possibly as low as 2. The second part of the analysis involved an extension to an absorption band attenuation model for both P and S waves. A linearized inversion for Δt*(f) is then performed on the averaged spectral ratios for the Basin and Range and shield. Both S and P attenuation decrease with increasing frequency in both provinces, but the frequency dependence of Qs-1 is more pronounced. Attenuation is greater for the Basin and Range, and the data suggest enhanced S attenuation between about 0.2 and 1.4 Hz relative to the shield. The ratio of Qs-1/Qp-1(Qp/Qs) decreases from ~2.0 at 0.05 Hz to less than 1.0 at frequencies greater than 1 Hz with similar effects in both provinces. Comparisons of our observations with different types of geophysical data lead us to conclude that dissipative mechanisms involving partial melt may be acting beneath the Basin and Range.

The decrease in Qs-1/Qp-1 at high frequencies is difficult to explain on both the Basin and Range and shield using intrinsic absorption mechanisms. The broadband data set used in this study appears to span the frequency band where intrinsic attenuation mechanisms evident at low frequencies are dominated by effects due to scattering at higher frequencies. If P and S waves interact with the same heterogeneities, the frequency where the attenuation due to scattering will be strongest is proportional to the inverse of the wavelength (i.e., where ka~1, where k is the wave number and a is the scale length of heterogeneity). Thus, in the presence of scattering, a decrease in the apparent S to P attenuation should be observed at higher frequencies, which is consistent with our observations.