Comments on the paper “Two independent real-time precursors of the 7.8 M earthquake in Ecuador based on radioactive and geodetic processes – Powerful tools for an early warning system” by Toulkeridis et al. (2019)
Introduction
After a major earthquake hits populated areas, there are often individuals, either scientists, emergency professionals or laypeople who look for precursory signals that could have been recognized and communicated prior to the event, thus avoiding the number of deaths and injuries. The 2009 L’Aquila earthquake is one of the most recent and notorious examples for this (Jordan et al., 2011), and the amateur prediction and its aftermath has the scientific community erring on the right side (Kolbert, 2015).
The scientific literature abounds with success stories about earthquake precursors being retrospectively identified after major earthquakes (Geller, 1997; Uyeda et al., 2009), but only a handful of examples exist for genuine short-term forecasts such as the M 7.3 earthquake in Haicheng, China, in 1975 (National Research Council, 2003). The search for diagnostic precursors has not yet produced a successful short-term prediction scheme (Jordan et al., 2011) and any proposed forecast/prediction methodology must follow a rigorous and transparent process of evaluation (Peresan et al., 2012).
In the paper “Two independent real-time precursors of the 7.8 M earthquake in Ecuador based on radioactive and geodetic processes – Powerful tools for an early warning system” Toulkeridis et al. (2019) compare gamma radiation time series from a single sensor installed in the Andes at Lasso (Lat.: S 0.7898°, Long.: W 78.6152°) with the occurrence of earthquakes. They claim that almost all earthquakes with magnitude M ≥ 5, and located up to 250 km from the sensor, occurred few hours after a significant positive radiation anomaly, including the M 7.8 earthquake on Abril 16th 2016 whose epicenter was about 200 km from the sensor. Toulkeridis et al. (2019) further claim that they observe a ∼1 m transient displacement at all continuous GPS sites in Ecuador several minutes prior to the M 7.8 earthquake. They indicate that the whole GPS network recorded a northward instantaneous displacement exceeding 1 m at most GPS sites at the time of the earthquake. They conclude that real-time monitoring of radiation and of GPS displacement can be used to implement an early warning system for forecasting earthquakes in the medium and short terms.
Our comment includes the analysis of a 15 month-long radiation time series from the same sensor as Toulkeridis et al. (2019) that overlaps their study period. We show that the detection performance of earthquakes is very poor, while correlation with local rainfalls is high, as it is seen during the hours preceding the M 7.8 earthquake. When analyzing the GPS data, we find no transient displacement anomaly before the earthquake. We further show that their GPS results are: (1) inconsistent with the known physics of earthquakes; (2) of bad quality compared with the standard state-of-the-art of GPS analysis; and (3) inconsistent with independent estimates of displacements for the Ecuador earthquake.
Section snippets
Method, data availability, operating conditions and detection range
We solicited the time series to NOVACERO, the firm owning the Radiation Portal Monitor (RPM) used by Toulkeridis et al. (2019). We have been given about 15 months of radiation data overlapping their period of study. The LUDLUM model 4525 RPM used in this work is equipped with an EJ-200 plastic scintillator that reacts in the presence of gamma radiation and optionally to neutron emissions. It is designed to detect radioactive material in scrap metal for recycling purposes. Besides radioactive
GPS data
Toulkeridis et al. (2019) show GPS 1-sample-per second kinematic analysis result for sites located from a few tens to a few hundreds of kilometers from the rupture area. Summarized in their Figure 7, the general behavior of the calculated positions is: (1) a random apparent displacement confined within a ∼1 m wide ellipse during the hours before the M 7.8 earthquake; (2) a westward to southwestward transient motion of several tens of centimeters, exceeding a meter at a few sites during “several
Conclusion
Both analyses of radiation and GPS time series from Toulkeridis et al. (2019) show major flaws. We demonstrate that radiation anomalies are seen only for the three earthquakes shown in Toulkeridis et al. (2019) over a set of 19 M ≥ 5 earthquakes, while a total of 162 radiation anomalies occurred during the 15-month period of analysis. Therefore, their hypothesis that radiation anomalies are reliable earthquake precursors has to be rejected. We also show that radiation anomalies and rainfalls
Acknowledgment
We thank an anonymous reviewer that helped us to improve the present comment. Gamma radiation data were provided by NOVACERO and rain precipitation data by Aglomerados Cotopaxi S.A., with a special acknowledgment to María Gallardo and Roberto Neumann.
References (21)
- et al.
Strain budget of the Ecuador–Colombia subduction zone: a stochastic view
Earth Planet. Sci. Lett.
(2018) - et al.
Terrestrial gamma-ray flashes
Nucl. Instrum. Methods Phys. Res. A
(2013) - et al.
Short-term earthquake prediction: current status of seismo-electromagnetics
Tectonophysics
(2009) - et al.
Rupture process of the 2016 M 7.8 Ecuador earthquake from joint inversion of InSAR data and teleseismic P waveforms
Tectonophysics
(2018) - et al.
Comparison of observed ground‐motion attenuation for the 16 April 2016 m 7.8 Ecuador megathrust earthquake and its two largest aftershocks with existing ground‐motion prediction equations
Seismol. Res. Lett.
(2017) - et al.
Statistical tests for pre-earthquake ionospheric anomaly
Terr. Atmos. Ocean. Sci.
(2004) Earthquake prediction: a critical review
Geophys. J. Int.
(1997)- et al.
Instantaneous geodetic positioning with 10–50 Hz GPS measurements: noise characteristics and implications for monitoring networks
J. Geophys. Res. Solid Earth
(2006) - et al.
Coseismic slip in the 2016 Mw 7.8 Ecuador earthquake imaged from Sentinel‐1A radar interferometry
Seismol. Res. Lett.
(2017) - et al.
Operational earthquake forecasting. State of knowledge and guidelines for utilization
Ann. Geophys.
(2011)