Predicting volcanic eruptions early and alerting authorities and nearby communities remains one of the biggest challenges in volcanology. Research published in nature communications describes a new detection technique called “jerk” developed by researchers and engineers at the Paris Institute of Geophysics (IPGP) and the GFZ Helmholtz Center for Geosciences. This method utilizes a single broadband seismometer to detect highly subtle ground movements associated with magma intrusions deep underground.
These weak signals can reveal the early stages of volcanic activity in real time. The researchers tested their method over a 10-year period at a volcanic observatory on La Réunion Island. During this period, the system successfully predicted 92% of the 24 eruptions that occurred between 2014 and 2023. Warning times ranged from just a few minutes before the eruption to up to eight hours. Approximately 14% of warnings did not result in an eruption. However, these alarms still detected movement of magma beneath the volcano. Because jerk systems require relatively little equipment, they can be an important early warning tool, especially for volcanoes that are not closely monitored.
Why it is difficult to predict volcanic eruptions
Volcanoes often show warning signs before erupting. These may include increased seismic activity, ground deformation, and changes in the release and composition of volcanic gases. Although these signals are well known, they remain difficult to interpret accurately. Scientists still struggle to determine exactly when an eruption will occur, how long it will last, and how powerful it will be.
Misinformation also causes serious problems. False warnings can cause costly evacuations, economic disruption, and public distrust of surveillance systems. As a result, improving the reliability of eruption predictions has become a major goal for scientists studying volcanic hazards.
Detecting minute earthquake motions from rising magma
Many previous approaches to eruption prediction rely on probabilistic analysis, which means exploring statistical relationships in large monitoring datasets. In contrast, a new approach developed by a team led by Dr. François Baudussel of the Paris Institute of Geophysics and Dr. Philippe Jusset of the GFZ Helmholtz Geological Survey Center in Potsdam focuses on directly detecting the physical signals associated with magma movement.
The “jerk” method identifies extremely small seismic motions that occur when magma penetrates the Earth’s crust. These signals appear as very low frequency transients. That is, impulses such as transition or settling signals recorded with horizontal ground motion, including both acceleration and inclination. The researchers say these signals likely originate from the dynamic rock-fracturing process that occurs before an eruption.
Scientists first identified these signals more than a decade ago while analyzing an extensive dataset from past eruptions of the Piton de la Fournaise volcano on Réunion Island. The signal is very small and can only be measured at a few nanometers per second to the power cube (nm/s3). Still, it can be detected with a single, very broadband seismometer.
The system includes special data processing that corrects for factors such as tidal currents. If a characteristic signal exceeds a certain threshold, the automated system will immediately issue an alert.
10 years of real-time volcano monitoring
The system was installed in April 2014 at the Piton de la Fournaise Volcano Observatory, operated by the Cité University Paris Geophysics Institute (IPGP) (OVPF-IPGP, Réunion Island). This tool acts as an automated component of the WebObs monitoring system and uses data from a broadband seismic station belonging to the global Geoscope network located 8 km from the summit of the volcano (Rivière de l’Est).
The first alarm occurred on June 20, 2014. The system issued an alert one hour and two minutes before the eruption began.
For the next 10 years, the jerk detection system operated continuously. Of the 24 eruptions recorded between 2014 and 2023, 92% triggered automatic warnings. Depending on the event, warnings were issued anywhere from minutes to 8.5 hours before magma reached the surface.
Piton de la Fournaise is one of the most closely monitored volcanoes in the world, making it an ideal place to test new techniques. Scientists were able to confirm the jerk warning using other monitoring indicators, including measurements of seismic activity, ground deformation, and volcanic gases. These independent observations confirmed that magma intrusion had occurred and an eruption was likely. The system was also evaluated using historical data from 24 eruptions from 1998 to 2010. Jerk signals consistently appeared before eruptive events.
“The great originality of this study lies in the fact that the Jerk method was tested and validated in real time in an automated and unsupervised manner for more than a decade, rather than in post-processing of the data, as is the case with the majority of studies of eruption precursor phenomena published in the literature,” explains Dr. Philippe Jusset, co-author of the study and GFZ Section 2.2 Geophysical Imaging Scientist.
About false positive alerts
Although the system performed well overall, some alerts did not result in an eruption. These “false positives” occurred in 14% of the times the system raised an alarm. However, further analysis revealed that these events were not random errors. Instead, they corresponded to real magma intrusions that ultimately did not cause an eruption. Scientists sometimes refer to these events as “aborted eruptions.”
During these warning periods, additional observations, including measurements of seismic activity, ground deformation, and volcanic gases, confirmed the presence of magma beneath the volcano. “In addition to the effectiveness of the jerk warning against eruptions, this tool has proven to be able to perfectly and unambiguously detect magma intrusions,” continues Philippe Jusset.
A recent example occurred during the earthquake crisis in Piton de la Fournaise on December 5, 2025. In addition to small deformation changes and gas anomalies, the scientists recorded a weak jerk signal of only 0.1 nm/s3. This signal confirmed that magma had penetrated beneath the volcano.
Testing the method on other volcanoes
After more than a decade of continuous real-time monitoring on La Réunion Island, researchers believe the Jerk system could serve as a practical early warning tool for other volcanoes, especially those with limited monitoring infrastructure.
The research team plans to expand testing this method to other active volcanoes. One of the first targets will be Mount Etna in Italy. The project “POS4dyke” deploys a new network of broadband seismometers from the GIPP geophysical instrument pool in Potsdam to detect jerk signals. In cooperation with INGV (Italy), installation is scheduled to begin in 2026.
The research will also work in conjunction with the SAFAtor project, which is investigating how fiber optic cables can be used to improve early warning systems for earthquakes and volcanic eruptions. Together, these efforts could greatly improve scientists’ ability to detect and predict volcanic activity around the world.
