Shale Bodies as a GeoCritical Reservoir – Observing/Modeling Lognormal Permeability Structures in Shale Production

P. C. Leary¹, P. E. Malin¹, J. A. Pogacnik¹, J. Rugis¹, B. Valles¹, P. Geiser², A. Lacazette², and J. Vermilye^2
1Institute of Earth Science & Engineering, University of Auckland, Auckland, New Zealand
²Global Geophysical Services, Denver, CO, USA

Reservoir permeabilities are conventionally assumed to be normally distributed about a mean value within the limits of a standard deviation. The normal distribution approach favors the use of small-scale well-cores and well-logs to estimate reservoir model permeabilities. However, reservoir–scale data show that variations in reservoir permeability are lognormally rather than normally distributed. Lognormal distributions are extensively observed for in situ flow data such as well productivity, trace-element abundance, ore body grades, and induced reservoir seismic noise.

Well-log and outcrop data also show that reservoir spatial property variations obey power-law scaling for lengths from mm to km. The “1/k” inverse spatial frequency scaling law of spatial fluctuations of reservoir properties implies that spatial fluctuations are larger at large scale lengths and smaller at smaller scale lengths. The fracture physics that lies behind the inverse scaling law is related to the existence of lognormal reservoir permeability distributions.

Lognormally distributed permeabilities and well productivities imply the existence of spatial heterogeneity in flow systems. The “1/k” scaling law implies that heterogeneity increases with scale length. We thus see why the use small-scale permeability estimates of to predict large-scale reservoir flow properties are generally inaccurate, and why many oil/gas/geothermal wells are poor producers.

Evidence shows that these general heterogeneity features of standard reservoirs also apply to shale bodies and shale body hydrocarbon production. Figure 1 compares the lognormality of oil-field and geothermal well productivity distributions (left) to lognormality of shale body ‘frack’ productivity (right). Figure 2 shows that well-log spatial fluctuations scale inversely with spatial frequency, thus obey the “1/k” scaling law.

With multiple reservoir data sets showing that spatial variations in reservoir properties are stronger at larger versus smaller length scales and that well/’frack’ productivity is lognormally distributed, we argue that large-scale reservoir observations should be used to map in the significant large-scale permeability structures in reservoirs, including shale bodies. This large-scale permeability mapping can be achieved through monitoring ambient seismic emissions (e.g. Tomographic Fracture ImagingTM) along with observing induced and natural microseismic events.

We present our argument in terms of a ‘GeoCritical Reservoir’ paradigm. This paradigm explains large-scale reservoir heterogeneity through fractures and fracture-connectivity. The observed distributions of in situ fracture properties establish the existence of a grain-scale ‘critical state’ fracture density at the percolation threshold. The critical grain-scale fracture density lies at the base of the pyramid of multiplicatively-connected fracture permeability. Multiplicatively-connected fractures explain the origin of lognormally-distributed reservoir flow structures. A critical density of fractures explains how and why producing reservoirs emit a steady background of seismic energy that can map in situ fracture distributions.

We present our argument in terms of a ‘GeoCritical Reservoir’ paradigm. This paradigm explains large-scale reservoir heterogeneity through fractures and fracture-connectivity. The observed distributions of in situ fracture properties establish the existence of a grain-scale ‘critical state’ fracture density at the percolation threshold. The critical grain-scale fracture density lies at the base of the pyramid of multiplicatively-connected fracture permeability. Multiplicatively-connected fractures explain the origin of lognormally-distributed reservoir flow structures. A critical density of fractures explains how and why producing reservoirs emit a steady background of seismic energy that can map in situ fracture distributions.

Using Tomographic Fracture Imaging data recorded by seismic arrays over active reservoirs, we map reservoir-scale permeability fluctuation structures into ‘Open Porous Media’ computer meshes for modelling flow in heterogeneous permeability media. These data collection and computational tools combine with GeoCritical Reservoir concepts to improve reservoir development and management.

AAPG Datapages/Search and Discovery Article #90180©AAPG/SEPM/China University of Petroleum/PetroChina-RIPED Joint Research Conference, Beijing, China, September 23-28, 2013