SEPARATION OF SCATTERING AND INTRINSIC ATTENUATION IN SOUTHERN SPAIN AND WESTERN ANATOLIA (TURKEY)

We have obtained a separation of intrinsic and scattering attenuation in two tectonically active areas of the Mediterranean Basin, southern Spain and western Anatolia, using the Multiple-Lapse Time Window Metho d. In southern Spain, Q(s) is predominant over Q(i) for frequencies lo wer than 4 Hz; between 4 and 8 Hz, results show similar scattering and intrinsic attenuation. For frequencies higher than 8 Hz, intrinsic at tenuation is clearly dominant over scattering. L(e)(-1) does not chang e uniformly with frequency; it increases until 8 Hz and decreases at f requencies greater than 8 Hz. The integrated energy curves of western Anatolia do not show regular change with distance. For the first energ y window (0-15 s), there is a clear increase of the energy with distan ces between 0 and 80 km, and then a decrease after 80 km. The fit of t he experimental data in the full hypocentral distance range is not goo d. We divided the experimental energy curves into two different segmen ts, 0-80 and 80-170 km. In 0-80 km, B-0 is close to 1 and the L(e)(-1) values are low (0.004-0.013). In the second distance range (80-170 km ) intrinsic attenuation is predominant over scattering. Q(s)(-1) is mo re frequency-dependent than Q(i)(-1):f(-1.47) against f(-1.29). We obs erved that coda Q(-1) is equal to the observed Q(i)(-1) for intermedia te frequencies, in both regions. However, Q(c)(-1) is really similar t o Q(s)(-1) for short distances and is between Q(i) and Q(s) for long d istances in southern Spain at low frequencies. Q(c) is similar to the observed Q(t) in both regions at high frequencies. The poor fit of the simulated curves with experimental data can be explained by introduci ng geometrical spreading values different to those used here. It is ne cessary to improve the model used by taking into account phenomena lik e non-isotropic scattering, variation of attenuation with depth, the p resence of surface waves or a geometrical spreading value different to the theoretical one used.