Application of the unwrapped phase inversion to land field data with irregular topography

Phase wrapping effect (or cycle skipping) is a big obstacle for the success of the full waveform inversion when low frequency information is not available. The phase unwrapping process recovers the missing cycles of the data, thus its inversion has a potential to generate an accurate and long wavelength structure model without low frequency information. Recently, the unwrapped phase inversion with an exponential damping was developed to generate a long wavelength structure with somewhat high frequency data. In this study, we apply the unwrapped phase inversion to land field data with a quite high minimum frequency. We also employ a finite-element-based modeling to take account of the irregular topography of the land data. Despite the elastic nature of the data, we use an acoustic-based inversion algorithm for the land data. In the pre-processing, we pick the first arrival traveltime and mute unwanted signals, like Rayleigh waves. We show that the unwrapped phase inversion provides a good convergent long wavelength velocity model. Numerical examples demonstrate that our unwrapped phase inversion result can be a good starting model for the subsequent conventional phase inversion and reverse-time migration on this high frequency, limited offset land data.