Statistical Analysis of Free-Surface Variability's Impact on Seismic Wavefield

One of the main challenges in seismic monitoring is the repeatability of the experiment conditions. Among other reasons changes in surface topography over time may cause strong non-repeatability in acquired seismic data especially in desert environment where the sand dunes can move between surveys. In this paper we present a numerical study of the effects of free-surface variation, defined as homogeneous Gaussian random field, on the seismic data. We show that for homogeneous models repeatability metrics (such as NRMS and predictability) using the early arrivals as measured by buried receivers depend mainly on the perturbation, but not on the smooth trend of the free–surface topography. For models with complex near-surface velocity NRMS for the thin sand area (<5 m) is almost twice as high as in the thick sand area (>10 m). Moreover, we demonstrate that significant non-repeatability (NRMS up to 70%) can be caused by just surface elevation changes.