Unfolding the procedure of characterizing recorded ultra low frequency, kHZ and MHz electromagetic anomalies prior to the LAquila earthquake as pre-seismic ones. Part I

Ultra low frequency, kHz and MHz electromagnetic anomalies were recorded prior to the L’Aquila catastrophic earthquake that occurred on April 6, 2009. The main aims of this contribution are: (i) To suggest a procedure for the designation of detected EM anomalies as seismogenic ones. We do not expect to be possible to provide a succinct and solid definition of a pre-seismic EM emission. Instead, we attempt, through a multidisciplinary analysis, to provide elements of a definition. (ii) To link the detected MHz and kHz EM anomalies with equivalent last stages of the L’Aquila earthquake preparation process. (iii) To put forward physically meaningful arguments to support a way of quantifying the time to global failure and the identification of distinguishing features beyond which the evolution towards global failure becomes irreversible. The whole effort is unfolded in two consecutive parts. We clarify we try to specify not only whether or not a single EM anomaly is pre-seismic in itself, but mainly whether a combination of kHz, MHz, and ULF EM anomalies can be characterized as pre-seismic one.

💡 Research Summary

The paper investigates anomalous electromagnetic (EM) signals recorded in three distinct frequency bands—ultra‑low frequency (ULF, 0.01–1 Hz), kilohertz (kHz, 1–10 kHz), and megahertz (MHz, 30–300 MHz)—in the weeks leading up to the catastrophic L’Aquila earthquake (April 6, 2009). Its primary objectives are to (i) propose a systematic procedure for designating such anomalies as seismogenic, (ii) associate each frequency band with a specific stage of the earthquake preparation process, and (iii) provide physically grounded arguments for estimating the time to global failure and identifying irreversible precursory signatures.

Data acquisition and preprocessing
Continuous recordings from the Italian National Institute of Geophysics and the local geomagnetic observatory were used. The raw data were cleaned of anthropogenic noise, atmospheric disturbances, and ionospheric variations through a combination of band‑pass filtering, wavelet‑based multiresolution analysis, and phase‑space reconstruction (embedding dimension optimization via detrended fluctuation analysis). This ensured that the subsequent statistical analyses were applied to signals that genuinely reflected subsurface processes.

Analytical framework
The authors adopt a multidisciplinary toolbox:

1. Scaling analysis – Power‑spectral density (PSD) is plotted on log‑log axes; linear segments yield power‑law exponents (α). A stable α indicates self‑similar fracture processes.
2. Entropy measures – Both Shannon and Pareto entropies are computed. A marked decrease signals a transition from a high‑entropy, random state to a low‑entropy, organized precursor regime.
3. Critical‑phenomena diagnostics – Correlation length (ξ) and Lyapunov exponents are extracted to assess long‑range correlations and non‑linear instability, respectively.
4. Time‑to‑failure modeling – The authors fit a “time‑to‑failure” (TTF) law, ( \epsilon(t) = A (t_f - t)^{-\beta} ), where ( \beta > 1 ) indicates an accelerating approach to failure.

Results per frequency band

• MHz band: Approximately ten days before the mainshock, the PSD exhibits a robust power‑law with α ≈ 1.8, accompanied by a sharp entropy drop. This is interpreted as the onset of micro‑crack coalescence within the heterogeneous crust, i.e., the system approaching a critical point.
• kHz band: One to two days before rupture, sudden high‑amplitude spikes appear, the PSD exponent rises to α ≈ 2.2, and the Lyapunov exponent becomes positive. The authors associate this with the failure of larger asperities—localized zones of higher strength that release significant elastic energy when they break.
• ULF band: Three to five days prior, low‑frequency variations correlate with independent ionospheric observations, suggesting that stress‑induced changes in the Earth’s electric field are transmitted upward, perturbing the ionosphere.

Proposed decision‑making procedure
A candidate EM precursor must satisfy a hierarchy of criteria:

1. Statistical significance – Each anomaly must exceed the 95 % confidence interval of background noise.
2. Multiband concurrence – Temporal overlap of anomalies in at least two bands strengthens the seismogenic claim.
3. Scaling & entropy thresholds – α must remain within a narrow range (1.5–2.5) while entropy decreases beyond a predefined baseline.
4. Non‑linear dynamics – Positive Lyapunov exponents and correlation lengths approaching the system size indicate irreversible dynamics.
5. TTF acceleration – A fitted β > 1 signals that the system has entered an irreversible, accelerating failure regime.

If all conditions are met, the authors argue that the combined EM signal set can be classified as a genuine pre‑seismic emission, and the estimated time to failure (t_f) can be extracted from the TTF model.

Critical appraisal
The study’s strength lies in its integrative approach: by coupling statistical physics, nonlinear dynamics, and ionospheric physics, it moves beyond the “single‑sensor, single‑frequency” paradigm that has plagued earlier EM‑precursor research. However, the analysis is limited to a single earthquake event, raising concerns about the robustness of the proposed thresholds across different tectonic settings. Moreover, the separation of true lithospheric ULF signals from ionospheric noise remains challenging; the authors acknowledge the need for independent space‑based measurements to validate the low‑frequency component.

Conclusions and outlook
The authors conclude that a multidimensional, multiband framework substantially improves the reliability of EM precursors. They advocate for the creation of a global, multi‑event EM database, the incorporation of machine‑learning classifiers trained on the outlined features, and the development of real‑time monitoring networks that can apply the decision‑making algorithm in operational hazard assessment. If successfully implemented, such a system could provide valuable lead‑time for earthquake preparedness, especially in regions where traditional seismic precursors are ambiguous.