Separating Hydraulic Fracturing Microseismicity From Induced Seismicity by Bayesian Inference of Non‐Linear Pressure Diffusivity

Abstract

Understanding subsurface fluid propagation and containment is fundamental to safe and effective hydraulic fracturing, wastewater disposal (Elsworth and Spiers, 2016), carbon sequestration (Vilarrasa et al., 2019), and geothermal development (Zang et al., 2014). In the context of hydraulic fracturing, operators need to understand the stimulated reservoir (i.e., in-zone fluid propagation) to assess operational efficiency, while simultaneously mitigating out-of-zone fluid propagation that may result in induced seismicity (IS). Natural fractures, which include faults at a variety of scales (Nelson, 2001), are a major contributor to out-of-zone fluid propagation and IS. The activation of existing faults causes IS, which can result in inefficient energy or thermal reservoir stimulation and potentially hazardous ground shaking.Microseismic monitoring data acquired during hydraulic stimulation provide a wealth of information on fluid migration in the subsurface and are widely used to characterize hydraulic fracture (HF) propagation away from a wellbore in applications such as unconventional resource production and enhanced geothermal systems (Shapiro et al., 2006). The HF propagates as a Mode I tensile fracture and microseismic events are caused by secondary shear fracturing around the primary HF and fluid leakoff (Maxwell et al., 2009). Events can also be induced far from the HF due to poroelastic effects or triggering of shear failure of natural fractures due to propagation of pore fluid through natural fractures.It is advantageous to separate microseismic point clouds into HF and induced events. This allows HF-related microseismic events to be used to calculate the HF dimensions for each hydraulic fracturing stage (Hummel & Shapiro, 2012;Sayers & Calvez, 2010). The induced events can be used to identify tectonic structures and natural