Global earthquake detection and warning using Android phones

Editor’s summary

Earthquake early warning systems depend on real-time analysis of waves detected at nearby seismic stations or by smartphone accelerometers. The latter approach, although less sensitive, is a promising way to offer alerts in areas without seismic stations. In a major step forward, Allen et al. describe the outcomes of an alert system rolled out to Android phones in 98 countries, which increased total warning system access 10-fold to 2.5 billion people over 3 years. Issuing more than 1200 alerts over that period, the system recorded low-magnitude errors and provided preshake warnings consistent with other detection systems but on a much larger scale. —Angela Hessler

Abstract

Earthquake early-warning systems are increasingly being deployed as a strategy to reduce losses in earthquakes, but the regional seismic networks they require do not exist in many earthquake-prone countries. We use the global Android smartphone network to develop an earthquake detection capability, an alert delivery system, and a user feedback framework. Over 3 years of operation, the system detected an average of 312 earthquakes per month with magnitudes from M 1.9 to M 7.8 in Türkiye. Alerts were delivered in 98 countries for earthquakes with M ≥4.5, corresponding to ~60 events and 18 million alerts per month. User feedback shows that 85% of people receiving an alert felt shaking, and 36, 28, and 23% received the alert before, during, and after shaking, respectively. We show how smartphone-based earthquake detection algorithms can be implemented at scale and improved through postevent analysis.

Earthquakes remain a constant challenge for many communities around the world. Our global understanding of where seismic events will occur is excellent (1), and we know how to engineer structures that can withstand shaking. However, we still face catastrophic earthquakes. Some kill and injure thousands of people such as in Türkiye and Morocco in 2023; many more injure hundreds, thereby disrupting many lives. Access to the knowledge and necessary resources to construct safe buildings remains a limiting factor in many parts of the world. Even with the resources, it would still take many decades to replace vulnerable buildings. Once earthquake-resilient buildings are in place, there remain hazards from falling objects inside buildings that can cause many injuries and some fatalities. We must therefore continue to look for additional strategies to reduce the impact of future earthquakes globally. Earthquake early-warning (EEW) alerts are one such strategy.

The global adoption of smartphone technology places sophisticated sensing and alerting capabilities in people’s hands, in both the wealthy and less-wealthy portions of the planet. Although the accelerometers in these phones are less sensitive than the permanent instrumentation used in traditional seismic networks, they can still detect the ground motions and building response in hazardous earthquakes (2–9). Traditional seismic networks have been used to develop EEW systems that rapidly detect an earthquake close to the epicenter and issue a warning across the affected region (10). EEW was first implemented in Mexico and Japan. The success of these systems led to implementation in Taiwan, South Korea, the United States, Israel, Costa Rica, and Canada (11–17). All these systems use permanent seismic stations deployed as part of regional or national seismic networks.

We present the methodology and performance of an algorithm that uses the existing network of Android smartphones to detect earthquakes globally and deliver early-warning alerts to smartphone owners ahead of hazardous shaking. The goal of the system is to deliver useful and timely alerts in as many earthquake-prone regions around the globe as possible. The system is supplementary to any existing national warning systems and includes the collection of uniform user feedback. We compare the accuracy of our early warnings against the global catalog of postshaking evaluations, and we also measured the utility of the alerts using surveys completed by more than a million alert recipients. We discuss the successes and challenges of this global sensing and alerting network called Android Earthquake Alerts (AEA) system.

Earthquake detection with smartphones

AEA uses the same principles of seismic wave propagation as traditional earthquake detection systems. When an Android smartphone is stationary, it uses the output of its accelerometer to detect a sudden increase in acceleration as would be generated by P and S waves in an earthquake. When an individual phone triggers, it sends a message to Google servers with acceleration information and an approximate location (location coarsened to preserve privacy). The servers then search for candidate seismic sources that are consistent with the time-space trigger distribution. When a candidate earthquake source satisfies the observed data with a high enough confidence, an earthquake is declared and its magnitude, hypocenter, and origin time (OT) are estimated on the basis of the arrival time and amplitude of the P and S waves. This detection capability is deployed as part of Google Play Services core system software, meaning it is on by default for most Android smartphones. As Android phones represent an estimated 70% of all smartphones globally (18), this system largely provides an earthquake detection capability wherever there are people, in both wealthy and less-wealthy nations.

From 1 April 2021 to 31 March 2024, AEA detected a total of 11,231 earthquakes (Fig. 1, fig. S1, and movie S1). Of these, 85% match an earthquake listed in a traditional earthquake catalog that our system was able to automatically ingest for detection validation (materials and methods). All nonmatching events were reviewed to verify whether they were earthquakes or not on the basis of user survey feedback, felt-shaking reports collected through Google Search (19), manual review of additional earthquake catalogs, and review of waveforms and triggers. For matching events, the traditional network magnitude estimates range from M 1.9 to M 7.8 (corresponding to an event in Japan and Türkiye, respectively). The ability of the system to detect earthquakes depends on the distribution of phones and other sources of cultural noise at the time of the earthquake. AEA does not detect earthquakes on most mid-ocean ridges but does detect subduction zone events that are tens to hundreds of kilometers offshore. Global Android detections and alerts are shown in Fig. 1 and fig. S1.

Read more   Science 17 Jul 2025 Vol 389, Issue 6757