Comment on seismic electric signals (SES) and earthquakes: A review of an updated VAN method and competing hypotheses for SES generation and earthquake triggering by Daniel S. Helman, physics of earth and planetary interiors, 302 (2020)

Abstract

As a comment on Seismic Electric Signals (SES) and earthquakes: A review of an updated VAN method and competing hypotheses for SES generation and earthquake triggering, herein we present the analysis of time series gathered by three MT stations located in the Eastern Cordillera of Colombia, for evaluating the electric and magnetic fields during an interval of several days. Records allow estimating multi-temporal apparent resistivity profiles in each station. Then, we correlate possible resistivity anomalies with seismicity, assuming that large earthquakes generated those anomalies. As an example, we identified a change in the apparent resistivity 7-h previous to the Mesetas earthquake (Mw6.0, December 24th, 2019). We would like to highlight that the resistivity anomaly was identified at the same depth as the earthquake hypocenter, and by more than one station.

Introduction

Herein we comment the analysis done by D. S. Helman regarding VAN method, and show a related geophysical approach using multi-temporal apparent resistivity profiles. Arguments presented in the article highlight the main problems of the VAN method. We agree with D. S. Helman that VAN's team has not always shared the needed data or does not inform false positives. However, to balance judgments about issues and collateral contributions, we bring to light some critical elements that the VAN team presents. As we can find in the literature (see, e.g., Cicerone et al., 2009; Park, 2005), anomalous magnetic or electrical signals preceding earthquakes have been reported, mainly during the last seven decades, relating to events M > 4, epicentral distances ranging from 3 km to several thousand km, and precursory times between 0.5 h to 3 months (see, e.g., Petraki et al., 2015). In almost all of there purported cases, the anomalies' visual identification (i.g. without using computer rutine) corresponds to frequencies in the ULF – LF ranges (0.001 Hz to several kHz). However, HF signals were also reported (up to tens of MHz). The VAN's approach for detecting SES (ULF) follows the electrokinetic effect hypothesis based on the flow of electric currents in the ground in the presence of an electrified interface at solid-liquid boundaries, which in turn produce magnetic fields. Oriented long dipoles and representative values of the crust's electrical impedance are considered in the VAN's approach. It is also interesting to highlight that the VAN's team never reported complementary magnetic observations. Instead, electrical anomalies were purportedly associated with earthquakes, with precursory times 2 h to 11 days. Other observational approaches focused only on magnetic measurements, discarding electric dipoles (Cicerone et al., 2009). Hence, we consider that the VAN's team, and other teams, which work in the same field, were not far to evaluate the temporal variability of the apparent resistivity (instead of signals related to the electric or magnetic fields, independently), which may offer an alternative view to the problem. In this case, crustal fluids have mobility under changes in the stress field, which would generate changes in the resistivity.

Despite the alleged weaknesses, the VAN group's research focus has promoted new interest and potential ideas for understanding the electrical response during the precursory period of the earthquake process. We recognize that the VAN method was an inspiration for alternative ideas to explain purported SES. For instance, we propose applying multi-temporal 1D-magnetotelluric (MT) surveys to measure the ground's resistivity variability due to changes during the earthquake cycle. For our research, we use the time series of three MT stations located in the Eastern Cordillera of Colombia for evaluating the electric and magnetic fields over several days. Records allow estimating multi-temporal apparent resistivity profiles at each station. Then, we correlate possible resistivity anomalies with seismicity, assuming that large earthquakes generated those anomalies. As an example, we identified a change in the apparent resistivity 7-h previous to the Mesetas earthquake (Mw6.0, Lon = 74.184°W, Lat = 3.462°N, H = 13 km-depth, December 24th, 2019, UTC 19:03:55. With aftershocks). We want to highlight that the resistivity anomaly was identified at the same depth as the earthquake hypocenter and by more than one station (see Fig. 1). Although D. S. Helman dismisses the phenomenon's piezoelectric nature, we consider that electrokinetic or even piezoelectric effects may explain the origin of this anomaly. We do not discard other causes, such as those mentioned by D. S. Helman regarding the presence of volatiles that generates additional electric charge via groundwater or material motion. In this sense, If resistivity is correlated with a physical phenomenon such as mineral dissolution and precipitation in the preparatory seismic phase, a local observation of apparent resistivity may not present abundant false positives as VAN method presents.

The VAN method, as well as in the case of other approaches, uses a visual inspection for rejecting noise due to light storms, cultural activity, etc. We suggest that this practice is unnecessary to apply in time series of magnetic and electrical measurements, due to the possibility of using frequency domain's estimations. Besides, to discard the ionospheric noise, we include the Kp-index estimated by the GFZ Potsdam. In our case, the time windows of one (1) hour are used to overlap spectrums and compute the apparent electrical resistivity with electric and magnetic fields' orthogonal channels. Finally, this process is repeated along with a large time window (e.g., five days, as is presented in Fig. 1).

The percentage of the anomaly of apparent resistivity is used for detecting the intensity of perturbations. We are confident that our three stations' small radius coverage precludes detecting additional anomalies caused by far earthquakes. However, we highlight that the anomaly of interest related to the Mesetas' earthquake starts seven hours before the mainshock, and it vanishes one hour before the earthquake. The stations are located closer than 300 km to the epicentral zone, and the nearest one of them provides a more significant percentage anomaly value. In contrast to the VAN team's practices, our dataset and algorithms are open access by direct contact to authors of this letter. In any case, our laboratory is still compiling enough cases supported in a reliable dataset for presenting to the scientific community a robust hypothesis that correlates the ground's resistivity variability with earthquake activity.