--> --> Space - The Final Frontier: Nature and Implications of Multi-Scale Spatial Heterogeneity in Carbonate Reservoir Analogs, by Eugene C. Rankey; #90029 (2004)

Datapages, Inc.Print this page

Click to view figures in PDF format

Space - The Final Frontier: Nature and Implications of Multi-Scale Spatial Heterogeneity in Carbonate Reservoir Analogs

Eugene C. Rankey
Comparative Sedimentology Laboratory/RSMAS
University of Miami, Miami, FL 33149
[email protected]


Construction of realistic geologic and simulation models of subsurface reservoirs requires understanding interwell heterogeneity, but quantitative description in the subsurface is difficult. Modern carbonate systems provide vehicles for exploring quantitative facies patterns along a ‘time line.’ The purposes of this presentation are: 1) to discuss general patterns and scaling relationships within modern carbonate depositional systems to 2) provide a conceptual framework for understanding and quantifying facies patterns and heterogeneity across a range of scales that can 3) yield insights into patterns possibly present in analogous reservoir systems.

The primary input for the analyses are optical remote sensing data with spatial resolution ranging from 250 m (MODIS) to 30 m (Landsat) to 1 m (IKONOS) (Figure 1). Qualitatively, these data show the multiple scales of heterogeneity, from the scale of the Bahamas platform (Figure 1A), to changes from windward to leeward sides of an island (Figure 1B), to the facies mosaic present within a tidal flat (Figure 1C), to attributes of individual subfacies (Figures 1D-E). These different types and scales of variability should influence production strategies.

These data can be quantitatively analyzed also. The images are processed to create thematic maps of spectral lithotopes that can be interpreted in terms of sedimentary facies. These maps are then imported into ARC, and the sizes, shapes, and spatial configuration of the subfacies are analyzed. These metrics are distinct from, but complimentary to, classic geostatistical measures.

In one example, we explored the size and spatial configuration of subfacies on part of the Andros Island tidal flat. The area-probability distributions for all subfacies follow distinct power-law relationships, as does lacunarity-area (a measure of ‘gappiness’) (Figure 2). These indicate statistical scale-invariance at the landscape scale, and the lack of a characteristic scale of either size or spatial configuration. In contrast, the tidal creeks responsible for sediment redistribution have a distinct exponential length-frequency distribution (Figure 3). Both of these results suggest that scaling in spatial patterns of these systems is predictable stochastically, but not deterministically, although the nature of scaling varies. Several examples from different sedimentologic systems will be illustrated.

These data provide an overview of the complex patterns present at multiple scales in modern carbonate systems, the nature of heterogeneity at those scales, and the dynamics of change between scales. The fundamental results illustrate the spatial domains in which patterns are determistic or stochastic, and explore the nature of the ‘rules’ within those spatial domains. Collectively, these data provide conceptual insights and quantitative measures of heterogeneity, both of which can be used to condition or constrain geologic or simulation models. 


Figure 1. Multiple scales of heterogeneity, Bahamas platform. The nature and dynamics of heterogeneity at each of these scales is different. A. MODIS image, illustrating platforms and basins. B. Landsat image showing reefs on windward (east) side, tidal flats on leeward (west) flank of Andros Island. C. IKONOS image illustrating the shallow marine region to west, a channeled belt, and supratidal inland algal marsh to east. The channeled belt includes a complex mosaic of subfacies. D. IKONOS image illustrating details of the complex of creeks, levees, marshes. E. details of patterns in D. 1C-E are © Spaceimaging.com. 

Figure 2. Attributes of size and spatial distribution for the levee-beach ridge habitat: tidal flats, Andros Island. A. Area-Exceedance probability (E.P.), both on log scales. E.P. is equal to the ranking in size, from largest to smallest, divided by the (number of samples + 1), and it represents the cumulative probability P[X>=x] that a given patch of area X has an area larger than x. B. Lacunarity-Box size, both on log scales. Lacunarity describes gap size distribution between patches of the same habitat, for a range of box sizes. Both have general power-law scaling relationships (lines).

Figure 3. Plot of length of tidal creek versus exceedance probability (log scale). The generally linear trend is consistent with an exponential distribution in length frequency of tidal creeks.