Precursory Phenomena Prior to Earthquakes

1. Introduction

This editorial is a part of the Special Issue (SI) “Precursory Phenomena Prior to Earthquakes (2nd Edition)” [1] but also covers papers from three Special Issues (SIs) that were published in Geosciences under the title “Precursory Phenomena Prior to Earthquakes” and were opened sequentially from 3 January of 2023 up to 1 May 24; the current one closed on 25 March of 2025 [1,2,3].

2. Significance

Earthquakes are unavoidable dangerous phenomena, which often have catastrophic effects [4]. They can lead to significant damage, property loss, and human casualties [4]. Despite the tremendous efforts that have been made in diverse fields of seismic research, forecasting earthquakes is a challenging problem which remains unsolved [5,6,7]. Research on seismic forecasting has been ongoing for over fifty years [7]. Nevertheless, it still falls short in accuracy and reliability because the physical mechanisms of earthquake genesis have yet to be determined [8,9]. Despite the drawbacks, the primary objective of earthquake forecasting is still to minimize the uncertainty associated with the time, location, and magnitude of a forthcoming earthquake [5,6]. Seismic forecasting is categorized into three types [7]: long-term (10–100 years), intermediate-term (1–10 years) and short-term (weeks–hours). Nonetheless, it is important to understand that establishing a direct link between observations and earthquakes is challenging, particularly when it comes to short-term forecasting [5,6,7]. Despite the complicated history of short-term earthquake forecasting, it has yet to be resolved [9,10].

The short-term estimation of arrival parameters of a forthcoming earthquake from warnings obtained weeks, days, or hours prior to the event is extremely crucial, especially in seismically active areas [5,6,7,11]. The precursory activity based on ground-based observations of physical irregularities that occur in close proximity and within short time intervals before an earthquake allows for its short-term forecasting [5,6]. Although this short-term forecasting is vital, it is inherently more difficult than long-term forecasting. The related scientific evidence suggests that several observations can be recorded in various parts of the Lithosphere–Atmosphere–Ionosphere Coupling (LAIC) [10]. Examples include disturbances in electromagnetic fields [4,6]; uneven variations in radon concentrations in groundwater and surface water, soil and atmosphere [5,11]; anomalous gas emissions [5]; irregular surface deformations brought on by pressure differentials [11]; several types of ionospheric perturbations [6,10,11]; and anomalies detected by satellites and remote sensors [12].

Monitoring of electromagnetic disturbances is intriguing. Pre-seismic electromagnetic disturbances occur in a wide range of frequencies starting below 10 Hz (Ultra-Low Frequencies—ULFs), within the kHz range and up to several MHz (High Frequencies—HFs) and between 100 MHz and 300 MHz (Very High Frequencies—VHFs) (e.g., [5,7,9,11,12,13,14] and references therein). Radon precursors are also significant. Radon monitoring has rapidly expanded due to its significance [5,6,15] and because it can be detected at low levels and long distances from its host site [11]. Before earthquakes, abnormal radon concentrations changes occur in the soil, groundwater, and atmosphere (e.g., [5,11] and references therein). The range, length, number of radon anomalies, the precursory time, and the epicentral distance are examples of features that vary before earthquakes [11]. However, seasonal fluctuations, rainfall, and changes in barometric pressure have impacts on radon emissions and, consequently, radon time-series are preprocessed [11]. Most associations between radon and earthquakes involve small- and intermediate-magnitude earthquakes. There are also associations between radon and large-magnitude earthquakes (see [11] and references therein).

Catastrophic earthquakes have been anticipated through electromagnetic and radon precursors. For example, the Mw = 9.0 Chile earthquake of 22 May 1960 was forecasted by an 18 MHz radioastronomy receiver [16]. Several precursors of atmospheric radon were recorded before the great Mw = 6.9 Kobe, Japan earthquake of 17 January 1995 and checked with numerous techniques (see [15] and references therein). The Ms = 7.1 Mindoro Philippines earthquake of 11 April 1994 was predicted 7 days before it occurred through monitoring of soil-radon with the Baracol VDG probe [17]. The MjMA = 9.0 Tohoku earthquake of 11 March 2011 was forecasted by ionospheric measurements [18]. The Mw = 7.5 Indonesia earthquake of 28 September 2018 was forecasted based on the physical properties of Ne, Te ionospheric disturbances by sensors of the China’s seismo-electromagnetic satellites [19]. Several other earthquakes have also been forecasted by electromagnetic and radon precursors (see the Tables in [5,6,7,11,12,13,14] and references therein).

The above scientific evidence demonstrates that the three SIs on the subject of “Precursory Phenomena Prior to Earthquakes” [1,2,3] are very important, since they cover significant topics of earthquake forecasting. This has ramifications for both science and society.

3. Aims

Following the analysis in Section 1, the aims of the three SIs [1,2,3] are described by the following keywords: (1) electromagnetism; (2) radon; (3) earthquakes; (4) remote sensing; (5) satellite measurements; (6) non-linear dynamics and chaos; (7) fractals; (8) self-organized systems; (9) seismic source mechanisms; (10) design of experiments; data analysis: algorithms and implementation; data management; (11) modeling and simulation [1,2,3]. All aspects of earthquake precursors are considered.

4. Analysis of the Published Papers

The first SI comprises three published papers [3]; the second has six [2]; and the third is composed of another six [1]. A correction was also published in [3]. Hence, fifteen papers have been published in total. This is a good number of publications to align with the aims of Section 2.

Table 1. Distribution of paper per country-region, geographical area, SI keywords, and citations.

Paper	Country-Region	Geographical Area	SIs Keywords	Citations
[20]	Japan	East Asia	earthquakes, precursors; electromagnetism	24
	Russia	East Europe
	Taiwan	East Asia
[21]	Russia	East Asia	earthquakes, precursors; ionosphere	24
	Kazakhstan	Central Asia
[22]	Japan	East Asia	earthquake precursors; electromagnetism; ionosphere; remote sensing	39
	Russia	East Europe
	Taiwan	East Asia
[23]	Russia	Eastern Siberia, Asia	earthquakes; radon; data analysis	45
	Russia	East Europe
	Russia	Central and Eastern Siberia, Asia
[24]	Mexico	North America	earthquakes; ionosphere; remote sensing; seismic source mechanisms	45
	India	South Asia
[25]	Pakistan	South Asia	earthquakes; radon; fractals; self-organization; non-linear dynamics and chaos; seismic source mechanisms	6
	China	Asia
	Greece	Europe
[26]	Greece	Europe	earthquakes; self-organization; chaos; electromagnetism; seismic source mechanisms	11
	Japan	East Asia
	Mexico	North America
[27]	Italy	Europe	earthquakes; data analysis-management	0
[28]	Russia	East Europe	earthquakes; data analysis-management; seismic source mechanisms	1
	Russia	Central and Eastern Siberia, Asia
[29]	Austria	Europe	earthquakes; precursors; ionosphere; remote sensing	1
	France	Europe
	Italy	Europe
	Greece	Europe
[30]	Russia	East Asia	earthquakes; algorithms and implementation; modeling	0
	Kazakhstan	Central Asia
[31]	Russia	East Europe	earthquakes; modeling	0
[32]	Greece	Europe	earthquakes; satellites; remote sensing; models; seismic source mechanisms’ self-organization; chaos; electromagnetism	13
[33]	Greece	Europe	earthquakes; satellites; seismic source mechanisms; electromagnetism	13
[34]	Romania	Europe	earthquakes; modeling	0
	Hungary

presents the distribution data of the published papers.

Table 1 shows that the three SIs covered collaborations from (A) 14 countries–regions (France, Austria, Italy, Greece, Romania, Hungary, Russia, Kazakhstan, Pakistan, India, China, Taiwan, Japan, and Mexico) and (B) 6 geographical areas (Europe, Central, East and South Asia, Siberia–Asia, and North America). (C) The common keywords of the three SIs describing the aims (Section 2) were fully covered. (D) The total number of citations was 210, which is significant for the metrics of the journal, despite the fact that the last SI [1] was closed recently, on 25 March 2025. All SIs succeeded in terms of the covered countries, regions, and keywords and also sourced noteworthy citations.

5. Synoptical Paper Presentation

The third SI [1], to which this editorial belongs, published six papers: (a) three research papers; (b) two reviews; and (c) one research review. In (a), Boudjada et al. [29] reports sub-ionospheric VLF anomalies of the Turkey–Syria earthquake using the multi-terminator method and, specifically, the amplitude and phase variations of the TBB transmitter signal. The daily amplitude and minima–maxima of TBB signals from the Granz facility and the extrema list from [4] are presented. Then, the sunrise and sunset terminators and the TBB variations five days before the earthquake, together with the list of inflexion and jump spectral types, are given. The authors show two main precursors from 27 to 30 January, and from 31 January to 3 February with anomaly fluctuations similar to those at the epicenter. Finally, a model is proposed. Salikhov et al. [30] reports anomalies of gamma-ray flux, VLF (f = 7.5 kHz, 10 KHz), ULF (0.001–20 Hz), and the Doppler frequency shift of the ionospheric signal before a M = 6.1 nearby earthquake. Two high-amplitude sites of the Tien Shan Research Station of the Lebedev Physical Institute are used for the combined measurements. A 300 m deep borehole is used to assess the seismogenic effects. Figures 2, 3, 5, 6, 7, and 10 of this publication present the variations of all measurements. Significant anomalies are reported in all parameter variations, practically, 8 days before the main shock and at various distances from the epicenters. This combination enhances the possibility of forecasting. Gitis and Derendyaev [31] report an iteration method to estimate the minimum area of alarm (MMAA). Their model utilizes the detection probability and the probability of a successful forecast. Using a learning algorithm, information capacity in terms of alarm and likelihood is calculated, and a precursor vector f^k is also calculated. Via a computational machine learning algorithm, early warnings of potential seismic hazard are found. In (b), the review by Nikolopoulos et al. [32] reports data similar to those described in Section 2. The review by Tritakis [33] reports non-electromagnetic indicators, such as strange animal behavior. Potential oxidations, electromagnetics, and electrophonics are proposed as possible mechanisms. Then, disturbances in physical parameters are analyzed, such as ground stress and electric currents, temperature and gas emanations, ground deformation water levels, and seismic lights. Then, electromagnetic indicators via LAIC and satellite data are provided, followed by data from studies conducted using artificial intelligence. Lastly, the paper in (c) [34], proposes the concept of a “precursory fingerprint” which is a collection of various precursory signals (seismic, physical, chemical, and biological) that are detected by a proper sensory matrix of monitoring equipment installed on the ground at sensitive spots and on satellites. The “fingerprint” consists of patterns and space–time components. Spatial fingerprints are presented in Romania (Figure 5, [34]) and time ones with virtual experiments (Figure 6, [34]). Finally, the concepts are evaluated.

6. Conclusions

The three SIs (including the two published previously) have fully covered the topic in hand. The next step will be an SI on the modern topic of artificial intelligence and earthquakes.