Approaches to defining reservoir physical properties from 3-D seismic attributes with limited well control: An example from the Jurassic Smackover Formation, Alabama

Much industry interest is centered on how to integrate well data and attrib utes derived from 3-D seismic data sets in the hope of defining reservoir p roperties in interwell areas. Unfortunately, the statistical underpinnings of the methods become less robust in areas where only a few wells are avail able, as might be the case in a new or small field. Especially in areas of limited well availability, we suggest that the physical basis of the attrib utes selected during the correlation procedure be validated by generating s ynthetic seismic sections from geologic models, then deriving attributes fr om the sections. We demonstrate this approach with a case study from Applet on field of southwestern Alabama. In this small field, dolomites of the Jur assic Smackover Formation produce from an anticlinal feature about 3800 m d eep. We used available geologic information to generate synthetic seismic s ections that showed the expected seismic response of the target formation; then we picked the relevant horizons in a 3-D seismic data volume that span ned the study area. Using multiple regression, we derived an empirical rela tionship between three seismic attributes of this 3-D volume and a log-deri ved porosity indicator. Our choice of attributes was validated by deriving complex trace attributes from our seismic modeling results and confirming t hat the relationships between well properties and real-data attributes were physically valid. Additionally, the porosity distribution predicted by the 3-D seismic data was reasonable within the context of the depositional mod el used for the area. Results from a new well drilled after our study valid ated our porosity prediction, although our structural prediction for the to p of the porosity zone was erroneous. These results remind us that seismic interpretations should be viewed as works in progress which need to be upda ted when new data become available.