Spatial and Temporal Properties of Noise from the Aquistore CCS Pilot Permanent Surface Array

Synthetic datasets are commonly used to aid interpretation, test hypothesis and as a benchmarking tool for evaluating the robustness of seismic imaging algorithms and defining confidence limits under which an algorithm will perform. Noise within these datasets is often modelled as white and/or Gaussian and therefore does not account for the spatial and temporal variations and trends observed in noise present within field data. This study defines a noise classification scheme that systematically represents these temporal and spatial variations and trends. Noise signals identified at the Aquistore injection site were classified using the scheme defined into the noise categories: stationary, non-stationary and pseudo-non-stationary noise. Future studies will focus on creating a mathematical description of the signals focussing on non-stationary and non-linear aspects with the aim to build this into a synthetic seismic dataset as realistic noise.