An empirical study of hydrocarbon indicators

Brian H. Russell, Hong Feng, John C. Bancroft

Numerous approaches have been published that derive fluid indicators, often called direct hydrocarbon indicators (DHI), from the amplitude variations with offset (AVO) method. These methods use linearized approximations to the Zoeppritz equations to extract physical parameters such as P-impedance, S- impedance, density, bulk modulus, shear modulus, Lam parameters and Poisson's ratio, and then infer the fluid content of a hydrocarbon reservoir from these parameters. Russell et al. (2003) used poroelasticity theory (Biot, 1941) to generalize several of these methods using a parameter dependent on the dry rock VP/VS ratio. The purpose of this study is to examine the generalized fluid method and compare this method with other fluid methods to see which indicator can most effectively discriminate between hydrocarbon sands and wet sands and which indicator is most sensitive to pore- fluid content estimation. To perform the sensitivity analysis we use an empirical dataset measured by Han et al. (1985), which covers saturated and dry sands with a wide range of porosities and clay contents at different pressures. As an indicator to evaluate the best method, we use the measure suggested by Dillon et al. (2003) in their analysis of fluid indicators. Using Han's dataset, we conclude that the generalized fluid method is in general the most successful fluid discriminant, given that we can get a reliable estimate of the dry rock VP to VS ratio. We are also able to make predictions about the dry rock VP to VS ratio as a function of porosity, clay content and pressure, which is related to depth.