Seismicity and Stress Associated With a Fluid‐Driven Fracture: Estimating the Evolving Geometry

Abstract

A coupled approach, combining the theory of rate-and state-dependent friction and methods from poroelasticity, forms the basis for a quantitative relationship between displacements and fluid leak-off from a growing fracture and changes in the rate of seismic events in the region surrounding the fracture. Poroelastic Green's functions link fracture aperture changes and fluid flow from the fracture to changes in the stress field and pore pressure in the adjacent formation. The theory of rate-and state-dependent friction provides a connection between Coulomb stress changes and variations in the rate of seismic events. Numerical modeling indicates that the Coulomb stress changes can vary significantly between formations with differing properties. The relationship between the seismicity rate changes and the changes in the formation stresses and fluid pressure is nonlinear, but a transformation produces a quantity that is linearly related to the aperture changes and fluid leak-off from the fracture. The methodology provides a means for mapping changes in seismicity into fracture aperture changes and to image an evolving fracture. An application to observed microseismicity associated with a hydrofracture reveals asymmetric fracture propagation within two main zones, with extended propagation in the upper zone. The time-varying volume of the fracture agrees with the injected volume, given by the integration of rate changes at the injection well, providing validation of the estimated aperture changes. These measurements are essentially point observations obtained during fracture movement (Guglielmi, Cappa, et al., 2015). Thus, there is a need for imaging fracture evolution at the field scale. Imaging the opening of a fracture is a difficult task due to the limited width of the feature, the rapid changes in properties, the significant depth of the event, and the complexity of the process, though there have been improvements in seismic imaging (Grechka et al., 2017). Much of the deformation within the fracture, and directly on the fracture surface, is aseismic or at frequencies that require broadband sensors (Guglielmi, Cappa, et al., 2015; Tary et al., 2014). One common feature of an opening fracture is an increase in the number of microseismic events in the surrounding region due to fluid flow and changes in the local stress field. This leads to changes in the rate of microseismicity around the fracture. Such fracture-related microseismicity is illustrated in Figure 1, where we plot events associated with the injection of fluid into a newly created hydrofracture that was monitored using microseismicity. The events detected during the monitoring are from a field experiment that we will analyze in section 3. The microseismic events resulting from the injection of fluid into the fracture were identified and located using 80 seismometers in four boreholes

Innovative Subsurface Learning and Hawaiian Exploration using Advanced Tomography (ISLAND HEAT) Phase 1 (Final Report)

“…Because of the uncertain relationship between intrusion-related stress changes and seismicity, we do not use the seismicity data to directly estimate aperture changes in the evolving fracture zone (Vasco, Smith, and Hoversten 2020). Rather, we hypothesize a causal link between the observed seismicity and aperture change, and design a regularization measure that will attempt to put the largest changes in fracture aperture at locations with abundant seismicity.…”

“…An alternative to regularization methods based upon injection location is to use natural seismicity as a proxy for locations that should correspond to fracture aperture changes. This approach can be based upon the theory rate-and state-dependent friction (Dieterich 1994) and the development of seismicity around a fluid-driven fracture (Vasco, Smith, and Hoversten 2020). As noted by Pedersen, Sigmundsson, and Einarsson (2007), some care must be exercised when applying Dieterich's theory in geothermal and volcanic areas due to the relationship between the current and pre-existing stressing rates.…”

“…This regularization term will favor solutions with aperture change near seismically active regions. It is possible to use the seismicity in a more direct fashion, as in Vasco, Smith, and Hoversten (2020), but that assumes that the fracture is the sole driver of the seismicity. This might not be true in our case due to the influence of nearby magmatic movement and the influence of the basal decollement (Dieterich, Cayol, and Okubo 2000;Montgomery-Brown et al 2015).…”