Geoscientists rely on variability in seismic reflections to indicate and quantify flow-unit characteristics. Such quantification depends on the ability to detect and resolve subtle waveform characteristics. Seismic attributes are useful in subdividing the components (attributes) that make up the seismic waveform; components which play a key role in reservoir characterization. Attribute computation and analysis tools have become commonplace in seismic interpretation workstations. Effective and efficient use of seismic attributes depends on the geoscientist's familiarity with attributes and what they represent.
Two sets of models help define and illustrate these attributes. The first set contains seven models that incorporate no tuning and show the effect of variable energy, bandwidth, and phase on a single reflectivity spike. Each model consists of nine traces 150ms in length; each trace has one reflectivity spike at 75ms. In other words, these seven models vary in the processing applied to one reflectivity spike:
When examined as a set, these seven models provide insight into the sensitivity of attributes to any combination of energy, bandwidth and/or phase variability.
The second model set incorporates tuning. These three models exhibit variable bed thickness while maintaining constant energy, bandwidth and phase. Each model exhibits the same wedge geometry and consists of 50 traces, 250 ms in length. The top of the wedge is at time 100ms. The wedge thickens from 0ms on the left to 50ms on the right:
It is important to examine both model sets because seismic attributes are sensitive to both processing and rock properties. The first model set exhibits laterally consistent reflectivity, but laterally variable processing (I.e., variable scaling, filtering, and phase rotation; consistent rock properties). The second model set exhibits laterally variable reflectivity; but laterally consistent processing (I.e., consistent scaling, filtering, and phase rotation; variable rock properties). To model sub-resolution thicknesses, each wedge was filtered with a zero-phase 8-10-40-50 Hz Ormsby filter prior to attribute computation.
