Heterogeneity

Key Takeaways

• The Dykstra-Parsons coefficient (V), developed by H. Dykstra and R.L. Parsons in 1950 for characterizing permeability variation in layered systems, is calculated from the log-normal distribution of permeability values measured across multiple layers or core plugs — V = (k50 - k84.1) / k50, where k50 is the median permeability and k84.1 is the permeability at one standard deviation below the median on a log-normal probability plot; a V value below 0.25 indicates mild heterogeneity with relatively uniform permeability distribution; V between 0.25 and 0.75 represents moderate heterogeneity typical of many sandstone reservoirs; V above 0.75 indicates severe heterogeneity where a small fraction of the pore volume contains most of the flow capacity, commonly found in fractured carbonates or highly laminated turbidite sequences; the Dykstra-Parsons coefficient feeds directly into waterflood recovery predictions, because severe permeability heterogeneity causes injected water to breakthrough early in the high-permeability layers while leaving much of the lower-permeability rock unswept.
• Vertical heterogeneity — variation of permeability and porosity with depth within the reservoir interval — is the most directly measurable form of reservoir heterogeneity from routine core and log analysis, and it manifests in production as layered reservoir behavior where different intervals contribute different fractions of the total well productivity; spinner flowmeter surveys run on production logging tool strings identify which intervals are producing and in what proportions, and the comparison of the flow profile against the permeability profile from core or log analysis reveals whether the flow is coming from the highest-permeability layers (as Darcy's law predicts for infinite-conductivity flow from multiple layers) or whether there is crossflow between layers (which occurs when a producing layer is hydraulically connected to an adjacent lower-pressure layer); the practical consequence of vertical heterogeneity for production management is that commingled production from multiple layers in a single completion leaves low-permeability layers underperformed, often justifying selective perforation strategies or mechanical isolation of different intervals for separate production tracking and stimulation.
• Horizontal heterogeneity — lateral variation of reservoir properties perpendicular to the well trajectory — is the most difficult form to characterize because well data is inherently one-dimensional (each well samples only its own location) and the interpolation of properties between wells requires geostatistical methods that introduce uncertainty; the spatial correlation length of horizontal permeability variation (the range of the variogram) determines the minimum well spacing required to adequately characterize the reservoir — if the permeability correlation length is 500 meters, wells spaced 1,000 meters apart will often encounter completely uncorrelated permeability values and the reservoir model between them will be highly uncertain; if the correlation length is 5,000 meters, widely spaced appraisal wells may still provide reasonably accurate interpolated permeability maps; determining the correlation length from seismic attributes, horizontal well data, and tracer tests is a key objective of reservoir characterization programs in heterogeneous fields.
• Natural fractures create a specific form of heterogeneity — dual-porosity or dual-permeability behavior — where the reservoir consists of two overlapping flow systems: the matrix (low permeability, high storage capacity) and the fracture network (high permeability, low storage capacity); pressure transient analysis in naturally fractured reservoirs shows a characteristic double-porosity signature (a plateau in the Bourdet derivative followed by a second upward trend) that reflects the contrast between the rapidly depleted fracture system and the slower matrix-to-fracture transfer; the fracture heterogeneity is particularly important for EOR design because injected fluid (water, gas, or chemical) tends to flow preferentially through the fractures rather than entering the matrix, reducing sweep efficiency unless the injection process is specifically designed for dual-porosity systems (gravity-stable flooding, low-rate injection that allows capillary imbibition to transfer fluid from fractures to matrix); characterizing the fracture intensity, orientation, and connectivity from image logs, core, seismic attributes, and well interference tests is the primary challenge in developing naturally fractured carbonate reservoirs.
• Geostatistical reservoir modeling uses variogram analysis — the quantification of spatial correlation of reservoir properties — to generate multiple equiprobable realizations of the reservoir property distribution that honor both the well data (which constrains local values at the well locations) and the spatial statistics (which constrain the pattern of variation between wells); each realization represents one plausible version of the heterogeneous reservoir, and running flow simulations on multiple realizations generates a range of production forecasts that quantifies the production uncertainty arising from reservoir heterogeneity; the P10 (optimistic), P50 (base case), and P90 (pessimistic) production forecasts that reservoir engineers report to management are typically derived from this ensemble of heterogeneity realizations; fields where the P90/P10 ratio exceeds 3:1 (meaning the optimistic case produces three times the pessimistic case) are considered high-heterogeneity risks where more appraisal data is needed before development decisions are made.