Preferred model (solid line) of frequency dependence of attenuation within the absorption band compared with constraints from laboratory studies (hachured region) and the frequency-independent assumption (dashed line). In our model, α is approximately 0.3 at periods shorter than 200 s, decreasing to 0.1 in the period range 300–800 s, and becoming negative (−0.4) at periods longer than 1000 s. We assume that the high-frequency corner occurs at 1 Hz where 1/q is 600, past which frequency α is 1 (Sipkin and Jordan, 1979). It is important to emphasize that we constrain the value of α and not of 1/q.
Constraining the frequency dependence of intrinsic seismic attenuation in the Earth is crucial for: 1. correcting for velocity dispersion due to attenuation; 2. constructing attenuation and velocity models of the interior using datasets with different frequency contents; and, 3. interpreting lateral variations of velocity and attenuation in terms of temperature and composition. Frequency dependence of attenuation q can be represented by a power law q ∝ q0ω–α. Despite its importance, efforts at determining α from surface wave and free oscillation data have been thwarted by the strong tradeoffs between the depth- and frequency dependence of attenuation. We develop and validate a new method that eliminates this tradeoff, allowing a direct estimation of effective frequency dependence of attenuation without having to construct a new depth-dependent model of attenuation. Using normal mode and surface wave attenuation measurements between 80 and 3000 s, we find that α varies with frequency within the absorption band. It is 0.3 at periods shorter than 200 s, it decreases to 0.1 between 300 and 800 s, and becomes negative at periods longer than 1000 s.
You can read more about this research in this article.