Bruce H Wilkinson1,
Carl N Drummond2
(1) University of Michigan, Ann Arbor, MI
(2) Indiana-Purdue University, Fort Wayne, IN
Abstract: Areal and vertical extents of peritidal carbonate lithosomes: Statistical measures for quantitative reservoir modeling
Quantitative characterization of carbonate reservoirs requires knowledge of abundances, thicknesses, and lateral extents of different lithofacies, their lateral/vertical association tendencies, degree of preferred orientation, and of statistical nonstationarity in any of the above attributes. Abundances and thicknesses of peritidal rock types in Proterozoic and Phanerozoic successions yield negative exponential distributions wherein occurrence/thickness metrics exhibit a predictable proportionality. Thickness frequencies are in excellent agreement with those anticipated when horizons of lithologic change occur randomly, and are only dependent on number of stratal elements designated over some stratigraphic interval. Areal extents of Holocene peritidal lithotopes from Antigua, the Bahama Bank, and the Persian Gulf are the same as those anticipated for population of mosaic elements, where lateral extents define an exponential distribution of element sizes. Area frequencies of facies elements on Holocene (and presumably older Proterozoic/Phanerozoic) platforms are only dependent on the total area and number of facies elements that existed over regions of carbonate accumulation. These data suggest that carbonate accumulation occurs via largely Poisson processes, afford significant statistical predictability in lateral and vertical persistence of different rock types, and allow for the formulation of stochastic models that incorporate both spatial and temporal aspects of random accumulation seen in ancient and modern systems. Such numerical simulations can also incorporate more deterministic low-frequency attributes of carbonate reservoirs such as spatial association, preferred orientation, and statistical nonstationarity among different lithofacies elements. Stochastic models are perhaps superior to deterministic formulations predicated on generally unsubstantiated assumptions about higher-frequency manifestations of stratal order in peritidal successions.
AAPG Search and Discovery Article #90914©2000 AAPG Annual Convention, New Orleans, Louisiana