ABSTRACT: Natural opening-mode fractures in Mississippian Barnett Shale cores from the Fort Worth Basin, Texas are commonly steep (> 75°), narrow (< 0.05 mm), and sealed with calcite. These fractures are important because they interact with fractures generated during hydraulic fracture treatments. We tested the effect of calcite-sealed fractures on tensile strength of shale with a bending test. Samples containing natural fractures are half as strong as those without, and always break along the fracture plane. In Barnett Shale fractures the junction between the wall rock and cement is weak because the cement grows mostly over noncarbonate grains and there is no chemical bond between cement and wall rock. Thus, even completely sealed fractures are prone to reactivation. The fractures observed in core are present in en echelon arrays, but data on fracture spatial distribution beyond the well bore are lacking. Geomechanical modeling, using the subcritical crack index as an input parameter, allows prediction of the fracture pattern, and where the index is high fractures are clustered. We measured the subcritical crack index of different shale facies and found it to be high, (> 92), with some variability with facies. Thus, with highly clustered natural fractures, interaction with hydraulic fractures will be heterogeneously distributed.
Gas in many shale-gas reservoirs is extracted using hydraulic fracture treatments. Evidence from microseismic monitoring of fracture growth shows that hydraulic fractures sometimes propagate away from the present-day maximum horizontal stress direction . One likely cause is that natural opening-mode fractures, which are present in most shales, act as weak planes that reactivate during hydraulic fracturing . Natural fractures in the Mississippian Barnett Shale from the Fort Worth Basin are commonly steep (> 75°) narrow (<0.05 mm), and are present in en echelon arrays . These narrow fractures are sealed with calcite and probably do not enhance reservoir permeability without being reactivated. This contribution addresses two aspects of the natural fractures that must be understood if their effect on hydraulic fracture treatments is to be predicted: namely, the spatial distribution of the fracture system and the likelihood that they will reactivate. The spatial distribution of natural fractures that have grown subcritically is predominantly a function of mechanical layer thickness and subcritical index. We measured the subcritical index of several shale samples to assess, for constant mechanical layer thickness, the degree of clustering. Reactivation of sealed fractures will be controlled in part by the strength of the junction between the fracture cement and the wall rock. We measured the tensile strength of shale samples with and without sealed fractures to determine whether sealed fractures are significantly weaker than the host rock. Samples for this study are from two cores of the Barnett Shale in the Fort Worth Basin, Texas (Mitchell Energy #2 T. P. Sims and Core A).
2.1. Subcritical Crack Index Measurements
We used a dual torsion beam apparatus at constant displacement to measure subcritical crack index (Fig. 1a,b) . Sample preparation, test and calculation procedures for subcritical testing are described fully by Holder et al.  and tests on two Barnett shale samples re reported by Gale et al. .