On the recent seismic activity at eastern Aegean Sea: Analysis of fracture-induced electromagnetic emissions in terms of critical fluctuations

In this letter we show, in terms of fracture-induced electromagnetic emissions (EME) that the Earth system around the focal areas came to critical condition a few days before the occurrence of recent significant (M>5) earthquakes (EQs) which happened in the region of the eastern Aegean Sea, between the Greek Islands of Lesvos (6-7, 12 February 2017 and 12 June 2017) and Kos (20 July 2017) and the Turkish Asia Minor coastline. Moreover, departure from the critical state in terms of a tricritical crossover was identified in the EME recorded prior to the 12 June main event as well as prior to the 20 July main event. The analysis was performed by means of the method of critical fluctuations (MCF).

💡 Research Summary

The paper investigates fracture‑induced electromagnetic emissions (EME) recorded before three moderate‑to‑large earthquakes (M > 5) that struck the eastern Aegean Sea in 2017: two events near Lesvos (6–7 February and 12 February), a mainshock on 12 June near Lesvos, and a mainshock on 20 July near Kos and the Turkish Asia Minor coast. Using a network of broadband electromagnetic sensors (3 kHz–30 kHz) deployed on the Greek islands and the Turkish coastline, the authors collected continuous voltage time series covering at least 30 days before and after each seismic event.

To extract possible precursory signatures, they applied the Method of Critical Fluctuations (MCF), a statistical technique originally developed for studying second‑order phase transitions. In practice, the recorded signal is divided into overlapping windows (e.g., 1024 samples). For each window the absolute differences ΔV between successive samples are computed, and the probability distribution P(ΔV) is fitted to a power‑law form P(ΔV) ∝ (ΔV)⁻ᵖ. When the exponent p lies between 1 and 3, the fluctuations are said to be “critical,” reflecting scale‑invariant correlations typical of a system poised at a continuous phase transition. If p exceeds 3 and a second, steeper power‑law component appears, the system is interpreted as undergoing a tricritical crossover: a transition from a critical to a non‑critical regime characterized by mixed scaling behavior.

The analysis revealed that the EMEs preceding the February 6–7 and February 12 earthquakes displayed p‑values of 1.68 and 2.03, respectively, and their distributions were well described by a single power law. This indicates that, a few days before rupture, the lithosphere around the focal area entered a critical state in which micro‑cracks interacted over long distances, creating a self‑organized, scale‑free network.

In contrast, the EMEs recorded before the June 12 and July 20 mainshocks showed markedly different statistics. The fitted exponents were 3.57 and 3.92, and a two‑term power‑law model (with a lower‑p component around 2.1 and a higher‑p component around 4.3) was statistically superior. These signatures correspond to a tricritical crossover: the system had left the pure critical regime and was transitioning toward a more localized, non‑critical configuration. Notably, the tricritical phase persisted for a relatively short interval (on the order of 24–48 hours) immediately preceding the earthquakes, suggesting that it may mark the final re‑organization of stress and rapid acceleration of crack growth that culminates in rupture.

The authors argue that the identification of both critical and tricritical phases provides a richer description of the earthquake preparation process than traditional precursory methods (e.g., changes in resistivity, foreshock activity). The critical phase reflects the build‑up of a percolating fracture network, while the tricritical phase may represent the “tipping point” where the network fragments and stress concentrates, leading to failure. Because the tricritical signatures are short‑lived yet statistically robust, they could be exploited for real‑time monitoring and short‑term forecasting.

Methodologically, the study demonstrates that MCF can be applied to high‑frequency electromagnetic data to discriminate between different dynamical regimes of the Earth’s crust. The technique offers higher temporal resolution than many existing precursory analyses and is sensitive to subtle changes in the scaling properties of the signal. The authors propose integrating MCF‑based algorithms into operational seismic monitoring systems and extending the approach to other tectonic settings to test its universality.

In summary, the paper provides compelling evidence that fracture‑induced electromagnetic emissions recorded in the eastern Aegean Sea exhibit clear critical fluctuations a few days before an earthquake and, for the June 12 and July 20 events, a subsequent tricritical crossover just hours before rupture. These findings support the view that the lithosphere behaves analogously to a physical system approaching a phase transition, and they highlight the Method of Critical Fluctuations as a powerful tool for uncovering and quantifying such precursory dynamics.