Geological Classification of Seismic-Inversion Data in the Doba Basin of Chad
Abstract
The Mangara field in Chad produces from Lower-Cretaceous Sequences of the Doba Basin. The reservoir units consist of interbedded sands and shales of different thicknesses, and the resulting variability in properties makes accurate characterization important for field development. The approach used in our characterization first involved a quantitative analysis
of
log
data and rock-physics models. This
analysis
established relationships between geological properties of interest, and elastic rock properties that could be obtained from seismic inversion. The results of the rock-property
analysis
were then mapped onto the seismic attributes to produce a geologically-classified volume. The first step was a statistical
analysis
of
log
-based elastic properties as they relate to other geological parameters. Significant shifts in impedance and vP:vS were observed between sand and shale points in the reservoir. For the sands, porosity also had an impact on the elastic properties. To incorporate multiple attributes, and to better identify classes of data, crossplots of the elastic attributes were used. Trends of both Vshale and porosity are evident on crossplots of P-impedance vs. vP:vS. These trends are roughly orthogonal, meaning that the two properties may be interpreted with some independence. Yet at the seismic scale, the thin nature of the sand/shale intervals (5-10m) is unlikely to be resolved, requiring the use of net-to-gross ratio rather than Vshale. A rock-physics model was calculated with different inputs for N:G and porosity so that the resulting trends could be used to guide the crossplot interpretation. The rock was modelled as a cemented, five-mineral composition, based on mineralogy from x-ray diffraction
analysis
. Variable N:G were modelled through a Backus average of the sand model and representative shale values. With the behaviour identified from
log
analysis
and rock-physics modelling, seismic attributes from AVO inversion were crossplotted in the same manner as the
well
data. Classification of the zone of interest followed the trends defining low, medium, and high N:G. The highest N:G points were further divided into low-, medium-, and high-porosity classes. The resulting lithology volume proved to be a good indicator of net reservoir, showing variations in continuity and thickness that were confirmed by additional drilling. The classified volume was used to produce sand probability maps that could be input into a reservoir model.
AAPG Datapages/Search and Discovery Article #90259 ©2016 AAPG Annual Convention and Exhibition, Calgary, Alberta, Canada, June 19-22, 2016