Capstone 4: Tight-gas reservoir with fracture overprint

Tight-gas reservoirs (k_matrix < 0.1 mD) cannot produce economically from MATRIX alone. Their commercial viability depends entirely on a FRACTURE OVERPRINT — natural or induced. Geostatistical modeling must capture both the matrix-property uncertainty AND the fracture-distribution uncertainty.

Two-property, two-scale model

• Matrix porosity φ_m: mean 0.08, sd 0.02. Continuous Gaussian field via SGS.
• Fracture density F: cells per unit area, typically positively correlated with LOW matrix-porosity zones (geomechanically brittle rocks fracture more).
• Effective permeability: k_eff(c) = k_matrix + k_fracture · F(c). With k_matrix ≈ 0.01 mD and k_fracture · F ≈ 5-20 mD, fractures dominate by factors of 100-1000.
• Drainage area: cells with k_eff above a producible threshold (e.g., 5 mD) and CONNECTED to the well are drained. P10/P50/P90 drainage area is the development metric.

Why the matrix-fracture correlation matters

If the geomechanical relationship is correctly captured (low φ → high F), the fracture network is BIASED toward low-φ regions — but those low-φ regions are exactly where you wouldn't produce from matrix alone. The combined effect can be GOOD (fractures unlock low-φ rock) or BAD (fractures connect to nearby water-bearing zones). The widget lets you toggle this correlation and observe drainage-area sensitivity.

Try it

• Defaults (positive matrix-fracture correlation = 0.6, k_matrix = 0.01, k_fracture = 5). The drainage area is moderate; fractures pulled into low-porosity rock unlock production. P90/P10 is moderate.
• Set correlation = 0 (no correlation). Fracture distribution becomes independent of porosity. Drainage area becomes more uniform across realizations (less variability).
• Set correlation = -0.6 (negative — fractures in high-porosity rock). Counterintuitive but possible (e.g., if fractures were diagenetically enhanced in porous zones). Drainage area concentrates in already-good rock — less value-add.
• Drop k_fracture to 1 (weak fracture contribution). The matrix dominates → drainage area collapses to near zero (k_matrix = 0.01 mD is far below threshold). This is the "uneconomic without fractures" case.
• Crank k_fracture to 20 (vigorous fractures). Drainage areas balloon. Production rates per well rise dramatically. The well-count and well-spacing decisions become very different.

An asset team has matrix-porosity logs from 8 wells (mean 0.07, sd 0.015) and seismic-derived FMI fracture-density indicators across the reservoir. The correlation coefficient between matrix porosity and fracture density is +0.4 (low porosity → high fractures). The current development plan assumes 40-acre well spacing. What do realizations of drainage-area uncertainty tell you about whether 40-acre spacing is sufficient or wasteful?

What you now know

Tight-gas workflows must integrate matrix and fracture properties with realistic correlations. Drainage-area distributions across realizations directly inform well-spacing decisions. The key uncertainty is the matrix-fracture relationship, which can dramatically change drainage outcomes. Modern workflows combine geomechanical models (predicting brittleness from elastic properties) with discrete-fracture-network (DFN) realizations.

Mark capstone complete →

References

• Holditch, S.A. (2006). "Tight gas sands." JPT 58(6), 86–93.
• Dershowitz, W. (1998). Practical Applications of Discrete Fracture Network Modelling. Golder Associates.
• Cipolla, C., Lolon, E., Mayerhofer, M., Warpinski, N. (2009). "Fracture design considerations in horizontal wells drilled in unconventional gas reservoirs." SPE 119366.
• Pyrcz, M.J., Deutsch, C.V. (2014). Geostatistical Reservoir Modeling, 2nd ed. (Chapter on co-simulation.)
• Aguilera, R. (2014). "Flow units: from conventional to tight-gas to shale-gas to tight-oil to shale-oil reservoirs." SPE 165360.