AAPG ANNUAL CONFERENCE AND EXHIBITION
Making the Next Giant Leap in Geosciences
April 10-13, 2011, Houston, Texas, USA
Geostatistically Modeling Topographically Controlled Deposition of Sub-seismic Scale Sandstone Packages Within a Mass Transport Dominated Deep-Water Channel Belt
(1) School of Earth Sciences, Stanford University, Stanford, CA.
(2) Advanced Resources and Risk Technology, Sunnyvale, CA.
The irregular surface topography of subaqueous mass-transport-deposits (MTDs) plays a primary role in controlling the deposition of sand from turbidity currents in the vicinity of the Puchkirchen field in the Molasse Basin, Upper Austria. The MTD topography created mini-basins in which sandstone reservoirs accumulated. These accumulations are difficult to laterally correlate using well-log data due to variable and unpredictable shape and size. Prediction is further complicated because the sandstone accumulations are difficult to detect in seismic-reflection data. Geostatistical modeling methods, which provide a framework to characterize spatial uncertainty, have focused on channel-levee-overbank-lobe models in deep-water depositional settings, which do not apply in this case. To overcome these challenges related to the lack of reliable lateral prediction methods, a new geostatistical modeling approach is presented.
We introduce a methodology to stochastically model sandstone accumulations whose irregular shape is defined by the underlying topography of the upper surface of one or multiple MTD(s). The modeling process simulates debris flow lobes followed by turbidites that fill in the accommodation space between debris flow lobes. Extensive data analysis is first performed to understand the extent to which available seismic attributes can aid in interwell prediction of the presence or absence of each of the defined depositional facies, as well as their geometries and distributions. We then present a modeling approach by which each successive depositional body is modeled constrained to 1) preceding topography, 2) seismic attributes, and 3) well data.