Extracting information from "noise"

In the context of an increasing world demand for oil and gas, the untapped reserves of the arctic are becoming nowadays of great interest. However, seismic exploration of areas located in artic regions is constrained by limited access and short operating time and weather window. One way to enlarge the time window is to extend the operations over the frozen sea during the winter season: in March 2007 Eni, Shell and Repsol conducted seismic acquisition tests in Alaska in order to assess the liability of such explorations.
Specific processing sequences must be implemented to address problems related to seismic acquisition in this environment: the tougher processing step is the suppression of the ice induced noise: Flexural Ice Wave - F.I.W. In fact, F.I.W. appears on seismograms as a noise fan, with a complicated pattern, and it is characterized by high energy (60-80 dB over useful reflection signals) and large dispersion, with strong aliasing artifacts. The combination of badly wrapped spectrum and high velocity dispersion makes standard filtering options to fail to attenuate FIW noise correctly.

ARESYS has been asked to design a novel approach to specifically attenuate F.I.W. noise, by investigating the physics that rules Flexural Ice Wave. ARESYS developed a noise attenuation algorithm (MADMAX - Multi Axes Double Model Adaptive subtraCTion) that overrules the results obtained with standard processing sequences. Moreover the deep understanding of the generating mechanism underlining F.I.W. guided waves allowed to derive an analytical representation of flexural waves. This compact representation became the engine of a powerful inversion scheme, able to extract information (such as local ice thickness) from what is usually treated as "noise".