AI Enhances Earthquake Detection Using Data from Multiple Sensors

AI analysis of data from multiple sensors can significantly improve the detection of earthquakes, according to a new study published on July 2, 2026. Researchers from the Norwegian research foundation NORSAR combined readings from various seismometers to enhance seismic signal detection. This comprehensive approach allows for better identification of weak seismic signals, which is crucial for monitoring both natural and human-made seismic activities.

Improving Seismic Monitoring with AI

Traditional methods of earthquake detection often rely on single seismometer readings, which can be insufficient for accurate analysis. The new study, led by A. Köhler and colleagues, demonstrates the advantages of utilizing artificial intelligence to process data from multiple sensors located within a small geographic area.

The researchers analyzed 30 years of seismic data from various stations, training an AI model through three distinct methods to optimize signal detection. The first method involved training the model on data from individual stations, while the second method combined signals from multiple sensors before training. The third approach provided the model with comprehensive data from all stations, allowing it to autonomously determine the best way to combine the signals.

Key Findings of the Research

The study revealed that the second method, where signals were combined prior to training, yielded the most accurate detection of weak seismic signals. This approach effectively amplified weak signals, making it easier to identify seismic events such as earthquakes and underground nuclear tests. Conversely, the third method, which allowed the model to decide how to combine the data, proved to be more computationally efficient but fell short in accuracy compared to the second method.

• Method 1: Train on individual station data.
• Method 2: Combine signals before training; most accurate.
• Method 3: Allow model to decide combination; more efficient but less accurate.

The researchers recommend using the AI model's autonomous approach for real-time monitoring, while the combined signals method is preferable for slower, more detailed analyses. However, it is important to note that the model's effectiveness diminishes outside the regions for which it was trained, particularly affecting the detection of S waves.

Future Directions for Seismic AI Applications

As the study concludes, the findings underscore the potential of AI to transform seismic monitoring. Training the model on a more extensive, global dataset is expected to improve its performance and generalization capabilities, particularly for S-wave detection.

Overall, the results, published in the Journal of Geophysical Research: Machine Learning and Computation, highlight the role of AI in advancing our ability to detect seismic activities, ultimately enhancing public safety and scientific understanding.