Text Size
Print Page
Print This Page
E-mail Page
Email This Page

Noise Attenuation for Ocean Bottom Seismic Surveys

Ocean bottom seismic (OBS) acquisition offers several advantages over towed streamer by offering converted shear wave in addition to P-wave data for AVO analysis and reservoir characterization. OBS data include a variety of acquisition methods including ocean bottom cables (OBC) and ocean bottom nodes (OBN). While placing sensors on the seafloor greatly diminishes the effects of surface noise, some of the same noise types that degrade conventional marine streamer data also impact OBS data. Regardless of the acquisition type, data conditioning to enhance the signal-to-noise ratio is important to allow subsequent signal processing methods to work as effectively as possible.

Suppressing noise on OBS data presents its own unique challenges that require a rich suite of tools and a wide breadth of experience - all of which ION brings to bear. Noise removal algorithms and workflows are tailored to the specific challenges of OBS environments. Noise is attenuated with the parallel goals of preserving signal amplitude, controlling signal phase, and maintaining signal fidelity over as broad a bandwidth as possible; all are paramount to achieving imaging and reservoir characterization objectives.

ION has experience processing data from more than 50 OBS projects and provides rigorous noise analysis and classification. For OBS data hydrophone and geophone or accelerometer noise types are approached independently.

ION utilizes custom data processing modules to handle both coherent and random noise categories in order to deliver on your imaging objectives. Common OBS noise types include:

  • Scholte waves
  • Shear wave leakage
  • Swell noise
  • Poor coupling
  • Marine life
  • Source bubble events
  • Vessel noise

Typical noise analysis and removal tools for coherent noise include FK, FKK, and 2D and 3D Radon domain processing followed by adaptive matching and subtraction. ION exploits properties of orthogonal OBS shooting geometries with a radial filter algorithm that uses the actual offset and azimuth of each trace and maps the data to the radial domain (apparent velocity versus time). Coherent surface wave noise (or ground roll) represented by a very low-frequency trend, is filtered out then transformed back to the x-t domain. The method is well suited for 3D OBS data.

Our SWD noise application is a very effective tool for random noise and spikes recorded in OBS acquisition environments. This process works in the frequency domain on shot, receiver, or CMP ensembles, determined by which domain would best randomize the noise. This process can be run in multiple domains to attenuate various noise types and maintains the original amplitude characteristics of the signal.