In reservoir studies in clastic environments, the two major problems that we encounter are reducing the risk of finding productive sands and defining the boundaries of these sands. A recent approach to solving these problems involves generating seismic attributes that are physically related to the reservoir properties and combining these attributes to predict the petrophysical properties of the reservoir (Hampson et al., 2001). The combination of attributes can be done using either multilinear regression or neural network analysis. Once we have derived a relationship between the attributes and the petrophysical parameters, these log properties can be extrapolated through the seismic volume. This allows us to infer the lithology, fluid content, and boundaries of the productive zones.