Seismic Texture Classification PDF Download

Seismic methods are fundamental to subsurface research and exploration. High resolution three-dimensional (3D) seismic reflection data is tremendously rich in subsurface information, yet can fail to map broader geological and or geotechnical contexts. We introduce a new suite of methods for volumetric classification or "domaining" of seismic reflectivity based on texture. The methods are quantitative, traceable and may reveal geological and or geotechnical domains within a seismic reflectivity images, that would otherwise remain hidden. These methods are sensitive to subtle variations in reflectivity texture and have particular application in hard rock settings where change in average velocity may be negligible across rock volumes that exhibit significant changes in texture. Dip steered seismic texture attributes like; contrast, entropy, and homogeneity, are used as input to cluster based geo-statistical techniques to recover new volume rendered image of the subsurface. Examples are provided from 3D seismic surveys in Nevada, USA and Kevitsa, Finland. For the Nevada data set, our technique differentiates textures in thick cover sequences to over 500 m below ground level and reveals changes in seismic textures across fault zones. Also the textural domains provide 3D boundaries that feed into cooperative inversion of seismic and electromagnetic data to yield subsurface conductivity distributions with detail not possible with standard magnetotelluric inversion. In Finland, our textural domaining points towards relationships between seismic texture and distribution of Ni concentration at the polymetallic Kevitsa mine site.