http://www.kgs.ku.edu/PRS/publication/2006/OFR06_06.html/p3-01.html

Last Modified April 2006

## Carbonate Lithofacies Prediction Using Neural Network and Geostatistical 3D Modeling of Oolite Shoals, St. Louis Limestone, Southwest Kansas |
Kansas Geological Survey KGS Open-file 2006-04 |

Facies proportion maps of oolitic complexes calculated from facies model built using object-based simulation. A) Facies proportion map of oolitic complexes for zone A, overlay of restored structural map of zone A; B) Facies proportion map of oolitic complexes for zone B, overlay of restored structural map of zone B; C) Facies proportion map of oolitic complexes for zone C, overlay of restored structural map of zone C; D) Facies proportion map of oolitic complexes for zone D, overlay of restored structural map of zone D. In all zones, occurrence of oolitic complexes is higher around the edges of the embayment, relative to structurally highest and lowest areas.

**A)** The distribution of length and width of oolitic deposits, measured from oolitic complexes on top of each St. Louis zone. The ratio of length and width shows a general trend of 3:1. The lateral extent of the oolitic deposits increased from zones D to zone C and B, and then decreased at the top of zone A; **B)** Geometric connectivity calculated in each modeling layer for oolitic complexes, displaying a general trend of increase from zone D to zones C and B. Within zones D, C, and B, the connectivity curves show a sigmoid shape. The rates of increase in connectivity are slow starting at the bottom, reach a maximum, then decrease toward the top of each zone. The geometric connectivity values are high and relatively stable at the lower part of zone A, and then drop dramatically near the top of zone A.

Variogram model of depositional facies in St. Louis Limestone computed from facies curves in the wells and fitted to the variogram model to obtain horizontal and vertical ranges of the facies . Data used to model distribution of facies using indicator simulation with simple kriging and Bayesian kriging.