Correlation Between Fracto-Emissions and Statistical Seismic Precursors in the Case of Low-Magnitude Earthquakes

Abstract

High-frequency pressure waves (up to TeraHertz, 10^12 Hz), produced by mechanical instabilities at the nano-scale, can trigger fracto-emission signals in the form of Acoustic Emission AE, Electromagnetic Emission EME, and Neutron Emission NE. The same phenomenon can take place during the early stages of a seismic event. Cracking is a multi-scale phenomenon within the earthquake preparation zone, nano-cracks coalesce to form larger ones, and the preparation area shrinks until coinciding with the quake epicentre. At the same time, pressure wave frequencies vary from THz for fracture at the nano-scale up to the simple Hz at the kilometre scale, which is the typical frequency of seismic oscillations. In this framework, fracto-emissions can represent a promising tool in seismology, not only for their monitoring capabilities during the earthquake, but also for their forecasting potentialities before the event. Another important seismic precursor is represented by the temporal variation of the b-value, a sort of statistical parameter deriving from the magnitude-frequency relation proposed by Gutenberg and Richter, that is also successfully applied to study the cracking process. In general terms, the fracture moves from nano to macrocracks and the b-value decreases from about 1.5 in the initial stages to 1.0 or less approaching the final collapse. In seismology, values of unity are commonly related to an incoming seismic event: i.e. an earthquake is often preceded by a b-value decrement in the weeks to few days before the quake occurrence. Since July 2013, an in-situ experimental campaign has started at a gypsum mine located in Northern Italy, revealing the strong seismic forecasting potentialities of the fracto-emissions by means of a dedicated monitoring platform and a multi-modal statistical analysis. In particular, AE, EME, and NE tend to anticipate the next seismic swarm peak with an evident and chronologically ordered shifting of about one day, three-four days, and one week, respectively. On the other hand, the b-value trend, estimated for all the major seismic swarms observed during the in-situ monitoring, shows a decrement to values below 1.0 approximately one-two days before the swarm peak occurrence.