DS0102 - Innovation technologique pour analyser, remédier ou réduire les risques environnementaux

Earthquake early warning using detection of gravity perturbations during fault rupture – E-GRAAL

Earthquake early-warning system based on the gravity perturbation

Study of the feasibility of an earthquake early-warning system based on the detection of gravity perturbation during the fault rupture.

Study of the feasibility of an earthquake early-warning system based on the detection of gravity perturbation during the first few seconds of fault rupture.

The objective of this project is to study the feasibility of an earthquake early-warning system based on the detection of gravity perturbation during the first few seconds of fault rupture. Earthquake Early Warning Systems (EEWS) are automatic devices (composed of seismometers and communication systems) capable of detecting the occurrence of an earthquake as soon as possible and to issue a regional alert prior to the arrival of strong seismic shocks. Conventional systems use the speed difference between P (faster and non-hazardous) waves and S (slower and destructive) waves. Each second of alert can have a great impact in terms of preserving human lives and reducing the consequences of the earthquake because it allows the activation of security procedures for people but also the launch of automatic systems (stopping trains, stopping elevators, controlling traffic lights, saving vital computer data and others). Since the perturbation of the gravity field occurs at the speed of light, a gravity warning system could be much faster in conventional systems. Moreover, this system could reduce the blind zone of the warning system (the zone too close to the hypocenter where it is not possible to issue an alert), but also to have a faster estimate of earthquake magnitude.

The project is structured in three parts: 1) Modeling the gravity perturbation during fault rupture based on realistic models of sources and the surrounding environment, and comparing the models with existing data from current gravimeters and seismometer, 2) Investigating the feasibility of a earthquake detector by gravity with a sufficient sensitivity to be used in an early warning system. In this part we study the feasibility of different concepts of gravity gradiometers with seismically isolated test masses as well as the influence of the local gravity fluctuations on the detectors, 3) Develop strategies in signal processing of a system based on gravity (extraction of appropriate parameters such as location, magnitude, false alarm probability) and implementation operational alert system and their linkage with existing EEWS.

1) The first part of the work was focused on the search for a gravity signal before the arrival of the seismic waves (never observed before) and the modeling of the gravity signals by analytical and numerical methods. A first analysis shows that a gravity signal is present during the Tohoku 2011 M=9.1 earthquake (with a 99% confidence level) in a superconducting gravimeter in the Kamioka mine ~ 400 km from the source, and before the seismic waves (Montagner et al., Nature Communications, 2016). This work demonstrate the existence of the signal, and made it possible to understand the limits of its detection with the existing instruments, and to compare the models with the data. In this work (Harms et al. 2015), a first first-order analytical calculation of the signal of gravity in an infinite and homogeneous space was used. This calculation was very ilportant l to understand the orders of magnitude of the gravity signals. Promising results with more realistic configurations (which will be published soon) have been obtained with the technique of the eigenmodes of the earth. 2) We have studied the impact of the local gravity disturbances on a earthquake gravity detector: the noise generated by seismic waves, by the air mass movements and by the objects moving around the detector. 3) A preliminary study of the implementation of a system for the early detection of seismic breaks based on gravity has been carried out. This study assumes the availability of a high-precision instrument. It seeks to optimize the detection performance of a network using these instruments.


Transient gravity perturbations induced by earthquake rupture, J.Harms et al. Journal of Geophysical Journal International, 201, 1416–1425 Prompt gravity signal induced by the 2011 TohokuLOki earthquake, J.4P.Montagner et al. – Nature Communicatio

The goal of this project is to demonstrate the feasibility of a new earthquake early-warning system (EEWS) based on the detection of the gravity perturbation from the mass redistribution during the first seconds of the fault rupture.
Earthquake early-warning systems are automatic devices able to detect the occurrence of an earthquake as soon as possible and to broadcast a regional alert before the arrival of strong seismic shaking. Conventional systems uses the difference in speed between P-waves (faster, harmless) and S-waves (slower, destructive). Every second of warning can have an important impact in terms of life preservation and earthquake mitigation, since it enables the implementation of safety procedures for people, but also the launching of automatic prevention systems (i.e. stop train, close gas pipelines, stop of elevators, control trafic lights by turning red on bridges and freeways entrances, save vital computer information, ...). Since the perturbation of the gravity field given by the earthquake is instantaneous, a gravity-based warning would be much faster with respect to conventional systems, allowing in general to increase the available time to take safety actions. Moreover, a gravity-based warning would in principle allow also a reduction of the "blind zone" (the zone around the hypocenter where no warning is available) and a faster estimation of the earthquake magnitude, since the magnitude information is completely contained in the duration of the rupture, contrarily to conventional system where the magnitude estimation can take several minutes, and in some cases tens of minutes.
According to preliminary results by the authors of this proposal and international Collaborators, the current seismometers and gravimeters are not usable in an gravity-based system, since they are not sensitive enough at the Fourier frequencies around ~ 0.1 Hz (corresponding to periods of tens of seconds) where the gravity signal is expected to peak during the fault rupture. New instruments with much better intrinsic sensitivity at 0.1 Hz are required. Moreover, due to the lack of seismic isolation, current instruments measure at 0.1 Hz the acceleration produced by the seismic noise rather than the perturbation of the gravitational field. The approach to improve sensitivity is to measure the relative gravitational acceleration between two seismically isolated test masses, a detector concept known as a gravity strain meter. The past two decades have shown rapid progress in the development of ultra-sensitive gravity strain meters, mainly driven by scientific community working on gravitational-wave detection. In particular new detector concepts as torsion bar antennas or atom interferometers, developed to observe gravitational-waves at sub-Hz frequencies can be used also for earthquake detection. Preliminary results show that a torsion bar antenna, with a strain sensitivity of h=10e-15 sqrt(Hz) at 0.1 Hz, can detect an M=6 earthquake at more than 100 km with a signal-to-noise ratio higher than 10. Experimental efforts to build a torsion bar gravitational-wave antenna with such a sensitivity are currently on-going in Japan.
The goal of this project is demonstrate the feasibility of a new gravity-based earthquake early warning system. In order to reach this goal three different tasks equally crucial and complementary should be performed: 1) the modelling of the gravity perturbation during the fault rupture based on realistic features for the source and surrounding medium, 2) the demonstration of the feasibility of a earthquake gravity detector, 3) the development of a signal processing strategy with this gravity-based system (to extract the relevant earthquake parameters, such as magnitude, position, false alarm probability) as well as the development of a practical implementation strategy and the coupling with existing earthquake early-warning systems.

Project coordinator

Monsieur Matteo BARSUGLIA (Astroparticule et Cosmologie)

The author of this summary is the project coordinator, who is responsible for the content of this summary. The ANR declines any responsibility as for its contents.


APC Astroparticule et Cosmologie
IPGP Institut de physique du globe de Paris

Help of the ANR 348,691 euros
Beginning and duration of the scientific project: September 2014 - 36 Months

Useful links