Uncertainties in Local Anisotropy Estimation from Multi-offset VSP Data

We have quantified the errors associated with VTI parameter estimation using multi-offset VSP data. Two common methods, P-wave slownesses only and slowness-polarization are investigated. Estimation errors are expressed in terms of the magnitude of the earth anisotropy, uncertainties related to first break pickings and maximum available source offset. For homogeneous overburden, P-wave slownesses technique can be used to estimate VTI parameters. We demonstrate that estimation errors of using only P-wave slownesses are significantly decreased as longer source offsets are included in the inversion algorithm. Larger offsets involve P-waves which propagate near horizontal at the receiver level and enhance the method's efficiency. An example synthetic VSP is presented next where P-wave slownesses technique successfully recovers VTI model parameters. In case of heterogeneous overburden, P-wave slowness-polarization technique seems to be a solution as there is no need to compute P-wave horizontal slownesses. However, we demonstrate that the errors of VTI parameter estimation using this technique are small only where the anisotropy is very weak (below 5%) and they are not improved by increasing the offset. Furthermore, wave interference effect on polarizations makes the method impractical even on noise free synthetic data.