András Fall

Natural fractures in shale: A review and new observations

Natural fractures have long been suspected as a factor in production from shale reservoirs because gas and oil production commonly exceeds the rates expected from low-porosity and low-permeability shale host rock. Many shale outcrops, cores, and image logs contain fractures or fracture traces, and microseismic event patterns associated with hydraulic-fracture stimulation have been ascribed to natural fracture reactivation. Here we review previous work, and present new core and outcrop data from 18 shale plays that reveal common types of shale fractures and their mineralization, orientation, and size patterns. A wide range of shales have a common suite of types and configurations of fractures: those at high angle to bedding, faults, bed-parallel fractures, early compacted fractures, and fractures associated with concretions. These fractures differ markedly in their prevalence and arrangement within each shale play, however, constituting different fracture stratigraphies—differences that depend on interface and mechanical properties governed by depositional, diagenetic, and structural setting. Several mechanisms may act independently or in combination to cause fracture growth, including differential compaction, local and regional stress changes associated with tectonic events, strain accommodation around large structures, catagenesis, and uplift. Fracture systems in shales are heterogeneous; they can enhance or detract from producibility, augment or reduce rock strength and the propensity to interact with hydraulic-fracture stimulation. Burial history and fracture diagenesis influence fracture attributes and may provide more information for fracture prediction than is commonly appreciated. The role of microfractures in production from shale is currently poorly understood yet potentially critical; we identify a need for further work in this field and on the role of natural fractures generally.

Natural fractures in shale

Natural fractures have long been suspected as a factor in production from shale reservoirs because gas and oil production commonly exceeds the rates expected from low-porosity and low-permeability shale host rock. Many shale outcrops, cores, and image logs contain fractures or fracture traces, and microseismic event patterns associated with hydraulic-fracture stimulation have been ascribed to natural fracture reactivation. Here we review previous work, and present new core and outcrop data from 18 shale plays that reveal common types of shale fractures and their mineralization, orientation, and size patterns. A wide range of shales have a common suite of types and configurations of fractures: those at high angle to bedding, faults, bed-parallel fractures, early compacted fractures, and fractures associated with concretions. These fractures differ markedly in their prevalence and arrangement within each shale play, however, constituting different fracture stratigraphies—differences that depend on interface and mechanical properties governed by depositional, diagenetic, and structural setting. Several mechanisms may act independently or in combination to cause fracture growth, including differential compaction, local and regional stress changes associated with tectonic events, strain accommodation around large structures, catagenesis, and uplift. Fracture systems in shales are heterogeneous; they can enhance or detract from producibility, augment or reduce rock strength and the propensity to interact with hydraulic-fracture stimulation. Burial history and fracture diagenesis influence fracture attributes and may provide more information for fracture prediction than is commonly appreciated. The role of microfractures in production from shale is currently poorly understood yet potentially critical; we identify a need for further work in this field and on the role of natural fractures generally.

Testing the basin-centered gas accumulation model using fluid inclusion observations: Southern Piceance Basin, Colorado

The Upper Cretaceous Mesaverde Group in the Piceance Basin, Colorado, is considered a continuous basin-centered gas accumulation in which gas charge of the low-permeability sandstone occurs under high pore-fluid pressure in response to gas generation. High gas pressure favors formation of pervasive systems of opening-mode fractures. This view contrasts with thatofothermodelsoflow-permeabilitygasreservoirsinwhich gas migrates by buoyant drive and accumulates in conventional traps, with fractures an incidental attribute of these reservoirs. We tested the aspects of the basin-centered gas accumulation model as it applies to the Piceance Basin by determining the timing of fracturegrowth and associated temperature,pressure, and fluid-composition conditions using microthermometry and Raman microspectrometry of fluid inclusions trapped in fracture cement that formed during fracture growth. Trapping temperatures of methane-saturated aqueous fluid inclusions record systematic temperature trends that increase from approximately 140 to 185°C and then decrease to approximately 158°C over time, which indicates fracture growth during maximum burial conditions. Calculated pore-fluid pressures for methanerich aqueous inclusions of 55 to 110 MPa (7977–15,954 psi) indicate fracture growth under near-lithostatic pressure conditions consistent with fracture growth during active gas maturation and charge. Lack of systematic pore-fluid–pressure trends

Natural hydraulic fracturing of tight-gas sandstone reservoirs, Piceance Basin, Colorado

Natural fractures form preferred pathways for basinal fluid flow and associated heat and mass transport. In gas sandstone reservoirs with low matrix permeability, fractures provide flow pathways between organic-rich source and reservoir layers during gas charge, and between matrix pores, hydraulic fractures, and the well bore during production. While the formation of natural fractures has previously been associated with gas generation and pore-fluid pressure increase through a process referred to as natural hydraulic fracturing, other driving mechanisms such as stress changes by tectonic or exhumation processes remained viable alternatives. To test whether these mechanisms contributed to fracture development, we investigated the spatial and temporal distribution of fracture formation and its relationship to gas generation, migration, and charge in sandstone of the Cretaceous Mesaverde Group across the entire production interval on a basinwide scale. Using fluid inclusion microthermometry of crack-seal fracture cement formed concurrently with fracture opening, we observed temperature trends that, when compared with temperature evolution models of the formation, date fracture formation between 41 and 6 Ma in the northern and between 39 and 6 Ma in the southern Piceance Basin. The onset of fracture formation 20–30 m.y. prior to maximum burial eliminates changes in stress state associated with exhumation as a mechanism for triggering the onset of fracture formation. Instead, calculated paleo–pore-fluid pressures of 40–90 MPa (5800–13,000 psi) during fracture opening and the presence of methane-rich inclusions in fracture cement suggest that fracture formation was aided by high pore-fluid pressures during gas generation in organic-rich shales and coals and associated charging of adjacent and interlayered sandstone reservoirs. A 10–20 m.y. age progression in the onset of fracture formation from deeper to shallower horizons of the Mesaverde Group is consistent with gas generation and onset of fracture formation activated by burial temperature with limited upward migration of gas at this stage of reservoir evolution. This age progression with depth is inconsistent with fracture formation triggered by changes in stress conditions associated with tectonic or structural processes expected to affect the entire formation synchronously. Our observations are thus most consistent with fracture formation by natural hydraulic fracturing in response to gas generation in interbedded source layers and reservoir charge. Based on widespread observations of fractures with similar structural and diagenetic attributes, we consider natural hydraulic fracture formation in response to thermocatalytic gas generation to be a fundamental mode of brittle failure in otherwise structurally quiescent basins.

The effect of fluid inclusion size on determination of homogenization temperature and density of liquid-rich aqueous inclusions

Abstract Homogenization temperature variations of several degrees Celsius or more are often observed within a group of fluid inclusions that appear to have all trapped the same homogeneous fluid at the same time and presumably at the same PTX conditions. For inclusions that homogenize at T ≤ ≈230°C, much of the observed variation can be attributed to the size of the inclusions. Larger inclusions homogenize at higher temperatures compared to smaller inclusions with the same density. The relationship between inclusion size and observed homogenization temperature is predicted by the Young-Laplace equation that relates the stability of a vapor bubble to the surface tension and pressure differential across the vapor-liquid interface. Vapor bubbles instantaneously collapse when the vapor bubble radius becomes less than the critical radius. During heating the critical radius of the vapor bubble is achieved at a lower temperature in the smaller inclusions. The critical vapor bubble radius varies from about 0.01 to ~3 μm for low-temperature aqueous fluid inclusions. The Gibbs surface free energy associated with the growth and collapse of vapor bubbles in pure H2O inclusions with critical radii ranging from 0.01 to 1 μm ranges from about 10-18 to 10-13 J/m2 and increases with both increasing critical vapor bubble radius and homogenization temperature. As a result of surface tension effects, the highest measured homogenization temperature, obtained from the largest inclusion in the group of coeval inclusions, most closely approximate the homogenization temperature that would be expected based on the inclusion density. For inclusions ranging from a few to several tens of micrometers in diameter and having densities such that the homogenization temperatures are approximately <230°C, homogenization temperatures may vary by about 1-3°C, depending on the inclusion size

Fracture porosity creation and persistence in a basement-involved Laramide fold, Upper Cretaceous Frontier Formation, Green River Basin, USA

Fracture-hosted porosity and quartz distribution along with crack-seal texture and fluid inclusion assemblage sequences in isolated, bridging quartz deposits show that open fractures can persist through protracted burial and uplift in foreland basins. Fractures oriented at a high angle to current maximum compressive stress remain open and were weak mechanical discontinuities for millions of years even at great depth. Upper Cretaceous Frontier Formation sandstones in the basement-involved (Laramide) Table Rock anticline, eastern Greater Green River Basin, Wyoming sampled by two horizontal wells (cut parallel or nearly parallel to bedding and at a high angle to steeply dipping fractures) have 41.5 m of rock in four cores at depths of 4538–4547 m. Cores intersect older E-striking Set 1 fractures are abutted by or locally cross-cut by N-striking Set 2 fractures. Both sets contain quartz and porosity. Sequenced using quartz crack-seal cement texture maps, Set 1 fluid inclusion assemblage (FIA) trapping temperatures increase progressively from 140 to 165°C then decrease to c . 150°C, compatible with fracture opening over c . 15 Ma during rapid burial followed by uplift in Eocene–Oligocene time. Set 2 opened at c . 160°C, probably near maximum burial. After a period of quiescence, Set 2 reopened at c . 5 Ma at c . 140°C, on a cooling trajectory. Intermittent Set 2 movement could reflect local basement-involved fault movement, followed after a pause by further Set 2 reactivation in the modern stress field during uplift. Interpretations are sensitive to available burial/thermal histories, which have considerable uncertainty.

Fracturing and fluid flow in a sub-décollement sandstone; or, a leak in the basement

Crack-seal texture within fracture cements in the Triassic El Alamar Formation, NE Mexico, shows that the fractures opened during precipitation of quartz cements; later, overlapping calcite cements further occluded pore space. Previous workers defined four systematic fracture sets, A (oldest) to D (youngest), with relative timing constrained by crosscutting relationships. Quartz fluid inclusion homogenization temperatures are higher within Set B (148 ± 20°C) than in Set C (105 ± 12°C). These data and previous burial history modelling are consistent with Set C forming during exhumation. Fluid inclusions in Set C quartz have higher salinity than those in Set B (22.9 v. 14.2 wt% NaCl equivalent, respectively), and Set C quartz cement is more enriched in 18O (20.2 v. 18.7‰ VSMOW). Under most assumptions about the true temperature during fracture opening, the burial duration, the amount of cement precipitated and fluid-flow patterns, it appears that the fracture fluid became depleted in 18O and enriched in 13C. This isotopic evolution, combined with increasing salinity, suggests that throughout fracture opening there was a gravity-driven influx of fluid from upsection Jurassic evaporites, which form a regional décollement. Fracture opening amid downward fluid motion suggests that fracturing was driven by external stresses such as tectonic stretching or unloading, rather than increases in fluid pressure.

Diagenesis and its impact on a microbially derived carbonate reservoir from the Middle Triassic Leikoupo Formation, Sichuan Basin, China

The uppermost Middle Triassic Leikoupo Formation in the western Sichuan Basin of China has recently been shown to host as much as 5.3 tcf (1.5 × 1012 m3) of natural gas resources. The reservoir rocks, composed mainly of microbially derived dolomudstone (e.g., thrombolites and stromatolites), are characterized by low porosity (<8%) and permeability (<0.001 to 10 md). The limestone is commonly tight and not of reservoir quality because of abundant meteoric calcite cementation, whereas the dolostone has various types of pores dominated by solution-enlarged pores and vugs, microbial framework pores, and micropores. Breccias are well developed in places, probably because of dissolution of underlying evaporites (e.g., anhydrite) by an influx of low-salinity fluids (e.g., freshwater and seawater) during an early burial stage. Early dolomitization created micropores in the dolomudstone, and subsequent diagenetic events were dominated by calcite, dolomite, quartz cementation, pyrite replacement, compaction, fracturing, and development of stylolites. Localized hydrothermal activity has been evidenced by high homogenization temperatures (∼160°C–200°C) obtained from fluid inclusions in fracture-filling cements. Bacterial sulfate reduction probably resulted in H2S generation, pyrite precipitation, and solution-enlarged pore and vug formation, whereas part of the current H2S in these reservoirs may have been sourced from thermochemical sulfate reduction or an underlying formation (e.g., the Feixiangguan Formation). Development of microfractures and associated micropores was probably the final diagenetic event, which improved pore interconnectivity. This study confirms the effect of diagenesis on the development of a microbial dolomudstone reservoir, which may be applicable to other similar microbial carbonate reservoirs elsewhere, for example, Middle Triassic sections of the Tethys region and offshore Brazil.