Scientists have discovered a massive crack beneath one of the world’s most perilous volcanoes, Campi Flegrei, located near Naples, Italy. This groundbreaking finding was made possible by an artificial intelligence study that analyzed over 54,000 earthquakes since 2022, revealing a distinct ring-shaped fault that had previously gone unnoticed.
The caldera, spanning approximately 7 miles, is situated just west of Naples and is home to neighborhoods built across its floor and rim. Led by Xing Tan, a doctoral researcher at Stanford University, the study focuses on machine learning for seismic monitoring.
The ring fault, a circular fracture outlining a collapsed volcanic basin, plays a crucial role in stress distribution and earthquake occurrence. The USGS explains how these ring fractures shape calderas and influence future volcanic activity. At Campi Flegrei, the new map highlights a thin, well-defined ring that correlates with uplift zones and extends offshore.
The discovery significantly impacts the understanding of the volcano’s inner structure, providing scientists with a precise outline of stress zones that were previously invisible. Traditionally, seismologists relied on phase picking, but AI has now proven its effectiveness in recognizing tiny, overlapping signals that older methods missed.
The improved catalog showcases the alignment of faults and areas of stress accumulation. Researchers emphasize that the new system offers clearer views of underground activity, enabling real-time detection of subtle shifts and precise earthquake locations.
The study highlights two long faults converging under Pozzuoli, a town on the caldera’s north side, raising concerns about stronger shaking in a shallow, urban environment. An earthquake in the magnitude 5 range is not ruled out, and the area’s long-standing risks are underscored by the evacuation of a portion of the city in the 1980s.
The analysis also reveals a shallow band of activity above 2.5 miles, indicating pressurized fluids and faults as the primary drivers. Campi Flegrei experiences slow ground movements called bradyseism, with an average uplift of 15 millimeters per month since April 2025.
Italy’s Civil Protection maintains an emergency plan, dividing the area into red and yellow zones for eruption and ash hazards. This guidance helps local officials respond swiftly to potential changes. Residents are advised to know their zones, secure heavy furniture, and follow INGV updates during seismic events.
The new map significantly enhances the understanding of the caldera’s faults, allowing engineers to estimate the realistic range of shaking a fault can produce. This information is invaluable for building checks, emergency drills, and planning the placement of shelters, ambulances, and road clearances.
The study’s findings also narrow scientific questions, particularly regarding the link between pressure in the shallow system and earthquakes. As the tool monitors the system in near real-time, it can test the relationship between the ring fault and uplift acceleration. This approach can be applied to other restless volcanic systems with existing seismic networks.
The model’s success at Campi Flegrei is attributed to data coverage and careful retraining. With ongoing improvements and additional data, the model will continue to enhance its accuracy. While it doesn’t predict eruption timing, it identifies stress concentration pathways, which are vital for communities facing ground instability.
The study, published in Science, emphasizes the importance of AI in seismic monitoring and its potential to revolutionize our understanding of volcanic systems.