189 6. The use of passive seismic methods for Geothermal exploration and monitoring of seismic data efficiently. Neural networks trained on synthetic and real earthquake datasets can rapidly classify faulting styles and estimate moment tensors with high accuracy. The study of focal mechanisms provides critical information for understanding tectonic processes and seismic hazard. For instance, the analysis of focal mechanisms during aftershock sequences can reveal how stress is redistributed on faults after a major event. Additionally, comparisons of focal mechanisms across different earthquakes help map the orientations of active faults and infer the directions of regional tectonic stress. Modern focal mechanism analyses, with their increasing precision and automation, remain a cornerstone of seismology, linking the physics of faulting with broader geodynamic processes. 6.1.2 Ambient noise seismic interferometry (ANSI) At the heart of ANSI is the concept of a “diffuse field”, where energy from seismic waves is dispersed evenly in all directions through a medium like the Earth’s crust. This concept has its roots in statistical physics, where wave energy behaves in random but statistically predictable ways. By understanding how this energy propagates, researchers can use noise as a sort of “natural tomography”, revealing the Earth’s properties, such as wave speed and material composition, down to fine scales. More practically, the ANSI method designates a signal processing approach that allows to extract coherent seismic waves from the incoherent ambient seismic signal recordings. This reconstruction process is achieved through cross-correlation operations between the diffuse noise signals recorded at two different locations on the earth’s surface (Figure 6.2), yielding empirical Green’s functions (EGFs) that are estimates of the impulse response of the subsurface medium in between the two Figure 6.2 Illustration of the Empirical Green’s function emergence process. Two incoherent noise signals recorded at two different locations are cross-correlated, to extract the EGF of the subsurface in between the two sensors. (White frame) In this example, the reconstructed EGF is highly dominated by surface wave (real data).
RkJQdWJsaXNoZXIy NjA3NzQ=