The islands in the Pacific Ocean are shaken daily as scientists investigate the source of a mysterious swarm of earthquakes.

From late July to October 2022, residents of the Manua Islands in American Samoa experienced earth tremors several times a day, raising concerns about an impending volcanic eruption or tsunami, as reported by PHYS.org.

The earthquake catalog for the region provided little information, as the islands lacked a seismic monitoring system capable of measuring underground tremors and aiding seismologists in locating the source of the swarm of earthquakes.

However, the inhabitants of the Pacific islands of Ta‘ū, Ofu, and Olosega sought answers. To determine the cause of the underground tremors, Clara Yun and her colleagues from the U.S. Geological Survey found an alternative method to fill the seismic gaps. During their research, the scientists employed machine learning and another technique called pattern matching on data regarding the tremors recorded by a single seismic sensor located 250 kilometers from the swarm in American Samoa.

The team utilized data from one station along with reports of tremors from residents to track the swarm. This continued until August-September 2022, when no local permanent seismic stations were established in American Samoa.

It is known that the swarm of non-eruptive volcanic earthquakes began in July 2022, approximately 15 kilometers off the coast of Ta‘ū. The volcanic islands of Samoa are formed when the Pacific tectonic plate moves over a hotspot in the southern Pacific Ocean.

According to Yun, initially, the information was limited to reports from locals about frequent tremors occurring several times a day for a few seconds at a time. Subsequently, the scientists turned to a remote seismic station in Upolu (Samoa), which is part of the Global Seismographic Network, for data that could be downloaded in near real-time via the EarthScope data processing center.

The researchers noted that the seismic signal from the series of earthquakes in American Samoa was difficult to detect at the remote station. Yun and her colleagues utilized a deep learning model called EQTransformer, as well as a method known as pattern matching, to isolate these tiny earthquakes from the noisy seismic background.

The model identified a significant number of earthquakes with locations corresponding to the eastern part of American Samoa, the largest of which aligned with the timing of felt reports. As a result, the scientists were able to characterize the onset and peak activity of the swarm. Portable and inexpensive Raspberry Shake sensors deployed in August 2022 helped quickly identify the area of the swarm. The data also indicated that the largest earthquake in the American Samoa swarm had a magnitude of 4.5.

The data suggest that the swarm concluded in October 2022 without an eruption but was likely associated with volcanic magma movement. According to Yun, the method used in the study may be beneficial in other locations worldwide where continuous seismic monitoring is rare and seismic hazards are poorly understood.