Porosity prediction using attributes from 3C-3D seismic data

Todor I. Todorov, Robert R. Stewart, Daniel P. Hampson

The integration of 3C-3D seismic data with petrophysical measurements at wells can significantly improve the spatial description of porosity. Using multi-regression analysis and neural networks, a functional relationship between measured porosity logs and some seismic attributes can be derived at the well locations. Once derived, the relationship is applied to the seismic data and a porosity cube is generated.

We test this procedure on a 3C-3D seismic survey recorded over the Blackfoot oil field in Southern Alberta. A functional relationship between porosity and P-P and converted to P-P time P-S seismic attributes is determined at thirteen well locations. A cross-validation test is used to determine the meaningful attributes. A porosity cubes are generated and high porosity anomalies are interpreted as sand channel.