Estimation of velocity and attenuation coefficient maps from crosswell seismic data

Abstract

Using a classic traveltime tomography and an attenuation tomography based on the centroid frequency shift method, velocity and attenuation coefficient tomograms are computed from two crosswell seismic data. Because of the limited frequency band of the data and the difficulties of picking the first transmitted events, velocity and attenuation coefficients are assumed independent of frequency during the processing, meaning that the attenuation linearly increases with frequency. This reduces the possibility of using a theoretical model to interpret the results. Moreover, attenuation results include both scattering and viscous attenuation. Nevertheless, the attenuation coefficient tomograms confirm some of the features seen in the velocity tomograms: The sand zones show a higher attenuation, although the shale zones and the seals of the reservoirs show a lower attenuation. With a Kramers-Kronig relation for the dispersion formula based on frequency-independent attenuation coefficient, the smooth part of the measured sonic velocity could be approximated from the seismic velocity and attenuation coefficient.