Test of Rock Physics Models for Prediction of Seismic Velocities in Shallow Unconsolidated Sands

In this paper various rock physics models are tested for the prediction of seismic parameters in shallow unconsolidated sands. For evaluation of the models we use well log data containing P- and S-velocities, porosity, water-saturation and mineralogy. Our main approach implies the use of various contact theories, as Hertz-Mindlin and Walton with rough or smooth grain contact points, to model the elastic properties of the high porosity end member. Then bounding models, such as the Hashin-Shtrikman lower bound or the Reuss bound, are used to estimate the properties for porosities within the high porosity and mineral end member. Additionally, we check the empirical models with parameters from laboratory measurements made by Zimmer et al (2007). Using either Hertz-Mindlin or Walton with rough bounds to define the high porosity end member gives an over prediction of the velocities: P-velocity by a factor of ~1.4 and S-velocity by a factor of ~1.7. Combining Walton with smooth grain contacts with Hashin-Shtrikman lower bound gave reasonable results, except for the low porosity interval. Walton with smooth grain contact points, combined with Reuss bound describes the data very well. The empirical models also gave a very good fit to data.