Seismic reservoir characterisation – rock and fluid prediction

The lateral and vertical distribution of reservoir properties such as fluid type, facies, lithology, porosity, and permeability away from well control is a key source of uncertainty in geological modelling. Heterogeneities of these properties within the reservoir occur at different scales and thus influence individual well performance and overall field productivity. Predictions of rock and fluid distributions at the seismic scale are essential elements of successful exploration and field development efforts.

Seismic predictions of rock and fluid property distributions rely on the pre-stack inversion of the amplitude versus angle (or offset) response to obtain estimates of P-wave and S-wave impedances and density (or derivatives such as lambda-mu-rho) along with their associated uncertainties. These elastic properties are then calibrated using well data and rock-physics models to predict the rock and fluid properties. The seismically derived properties are then depth-converted and used to guide the interpolation of well-log-derived properties between wells. Alternative calibration workflows use joint inversion for facies, reservoir properties and elastic properties, or deep learning methods such as neural networks.

Regardless of the approach used for seismic reservoir characterisation, there remain challenges. These include:

• prediction of fluid saturation, especially distinguishing low-saturation gas from higher gas saturations
• overlap in elastic properties among reservoir facies and rock types, which limits discrimination and classification
• limited spatial resolution of surface seismic data
• errors in time-depth conversion

To support the application of seismic reservoir characterisation, CRGC is working in the following areas:

• seismic signatures of patchy saturation
• experimental and theoretical study of partial saturation effect on elastic properties of sedimentary rocks
• broadband acoustic measurements: from seismic to ultrasonic frequencies
• seismic anisotropy from VSP and surface seismic data
• seismic attenuation from VSP and log data
• time-lapse full waveform inversion of vertical seismic profiles