Deep learning-based 3D microseismic event direct location using simultaneous surface and borehole data: an application to the Utah FORGE site

Microseismic monitoring is crucial for characterizing and assessing fracture systems developed during subsurface operations like hydraulic fracturing, geothermal reservoir development, and CO₂ sequestration. In enhanced geothermal systems, such as the Frontier Observatory for Research in Geothermal Energy (FORGE) project in Utah, direct location of microseismic events stimulated from fluid injection can offer vital insights about fracture propagation and reservoir connectivity for more informed production decisions. Surface and borehole seismic arrays are usually deployed for monitoring in these long-standing projects, each with its own strengths and limitations. Surface arrays provide a wide aperture for accurate epicenter location, but suffer from depth estimation uncertainty and poor signal-to-noise ratio. In contrast, borehole arrays offer better signal quality and improved depth estimations but with limited coverage and azimuth for epicenter location. To bridge the capacities of surface and borehole arrays, a deep learning method is proposed to directly locate microseismic events in 3D using recordings from both surface and borehole sensors based on an elastic medium assumption. The proposed method has potent compatibility for embracing diverse datasets and a strong ability to model complex dynamics and interactions between multiple datasets. Through an example application on field passive seismic data at the Utah FORGE site, this method demonstrates its potential to capitalize on the available data while avoiding their inherent shortcomings.