Deep inside a remote corner of Prince William Sound, a mountain is sending out signals that no one heard until a network of sensors was installed only a few years ago. The region is already known to hold one of the largest and most dangerous unstable slopes in the United States. The Barry Landslide towers over Barry Arm, a long, narrow fjord bordered by glaciers and steep fractured rock. If the slope collapses, hundreds of millions of cubic meters of debris would fall straight into the water. That scenario has long been recognized as a high consequence hazard because confined fjords amplify waves. A rapid failure would produce a violent displacement wave that could strike Whittier, Cordova, and any vessel in the area. For years scientists have tried to determine what might trigger a sudden collapse. What they found during the most recent study revealed something unexpected hidden beneath the ice.
Beginning in 2020, a series of seismic and infrasound stations were positioned across Barry Arm to watch for rockfalls, small earthquakes, glacier calving, and any motion along the unstable slope itself. One of these stations sits directly on the surface of the landslide. Another sits across the fjord. These instruments record every vibration that moves through the ground. The data volume is immense. To understand it, researchers manually reviewed an entire year of continuous waveforms. They cataloged every type of signal they could identify. Earthquakes were easy. Glacier calving was unmistakable. Rockfalls showed long rumbling signatures. Then a different class of signals began to stand out. They were extremely short and sharp, with very high frequencies. They had clear P and S waves, meaning they were true brittle fractures. Most importantly, they were recorded only on the station positioned on top of the landslide. The signals did not appear across the fjord. They did not show up in the infrasound array. They came from very close to the station, but not from the surface.
Once the team realized these strange events were consistent and abundant, they developed a method to detect them across the full three year record. The result was startling. A total of 32,877 of these short impulsive events occurred between 2021 and 2023. Many of them were nearly identical. They repeated with the same waveform shape and arrived from the same direction. By comparing P and S arrival times, and by measuring the direction the waveforms approached the station, the researchers constrained their source region. The signals came from the southwest of the landslide, roughly one to two kilometers away. That direction points directly at Cascade Glacier, a large ice mass that sits behind the unstable slope. The signals did not originate inside the slide. They did not originate from normal glacial calving. They came from something deeper.
What made the discovery more alarming was the seasonal pattern. These fracture events were rare in early summer. As late summer approached, they began to increase. They continued rising through autumn and reached their highest rate in December and January. Then they shut off abruptly toward the end of winter. This cycle repeated every year. It did not match the seasonal cycle of glacier movement. It did not match rainfall. It did not match temperature at the surface. It followed its own pattern independent of these common environmental drivers.
To understand the possible origin, the scientists combined the seismic data with radar measurements of the landslide surface. During 2022 and 2023 the unstable slope experienced periods of accelerated movement. In both years, the first signs of rapid motion occurred as the seasonal cracking began to rise. This meant the internal cracking beneath the glacier increased at the same time the surface of the landslide began to shift. The timing lined up closely at the start of each acceleration. The surface then slowed and stopped. The cracking did not. It continued rising for months afterward. In 2021 the landslide barely moved, yet the cracking still rose and peaked through the fall and winter. This showed the signals did not depend on visible slope activity. The internal process was independent and persistent.
This raised an important question. If the signals do not originate inside the slide, why do they correlate with the early stages of slope acceleration. The researchers considered several mechanisms. The first was crevassing within the glacier itself. Large ice masses produce cracking events as they deform. But the seasonal velocity of Cascade Glacier does not match the seismic cycle. The glacier speeds up in late spring and slows in early winter. The fracture signals rise months after the peak in glacier speed. The signals also have waveforms that differ from typical icequakes. Another possibility was sliding at the base of the glacier. Some glaciers experience episodes of basal slip triggered by meltwater pulses. Those events also do not match the timing seen at Barry Arm. The seasonal mismatch ruled out many simple explanations.
The most plausible explanation tied the seismic bursts to the behavior of water that flows beneath the glacier and through the fractured rock behind the landslide. During the warm season, meltwater travels downward through the ice and enters a network of cracks in the underlying bedrock. These cracks and conduits act as an underground plumbing system. The water that moves through them is relatively warm. Late in the year the supply of meltwater decreases. The cold season begins to dominate. As the water trapped within fractures begins to freeze, pressure increases. Ice expands as it forms. Expansion inside a confined rock fracture can split the rock further. Each fracture produces a sharp seismic pulse. The pulses repeat from the same directions because the same pathways freeze every year.
This interpretation matches both the waveforms and the seasonal pattern. The warm months fill the fractures. The cold months freeze them shut. Once the freeze is complete, the system becomes quiet until spring. The researchers note that this behavior has been observed in a different unstable slope in Norway. There, a similar seasonal crack pattern was tied to freeze and thaw cycles deep within the rock. The two sites share similar fracture physics even though the landscapes differ.
That explanation carries significant implications for Barry Arm. The fractures that freeze in winter sit behind the unstable slope. They exist inside the bedrock that controls the pressure and water delivery into the landslide. If the underground pathways behind the slide begin to freeze shut, the hydraulic pressure within the system can rise. Elevated pressure is a known driver of slope motion in steep fractured terrain. The fact that the seismic bursts rise at the same time the slope begins to accelerate suggests the freezing process may be related to the changes observed on the surface. The study does not claim these signals are direct precursors. The authors are clear that more research is required. They do point out a critical fact. The seasonal pattern is extremely consistent. Should a year arrive in which the cycle does not begin when expected, or begins far later, it could indicate an abnormal level of water pressure persisting into the fall. That scenario has the potential to increase landslide risk. This is not a forecast of collapse. It is a warning that the frozen pathways behind the slope may provide indirect information about the state of the system.
The deeper concern is that this cracking could be revealing internal stress changes in a part of the mountain that cannot be seen or measured with traditional tools. The unstable slope at Barry Arm has been creeping for decades. It is heavily fractured and lacks the ice support that once stabilized it. The glacier has retreated, leaving a steep unsupported wall of rock. The slope is capable of rapid failure. A collapse of five hundred to seven hundred million cubic meters of material would strike the fjord with enormous force. Historical events in Alaska have shown how destructive these slides can be. The 1958 Lituya Bay rockfall produced the tallest wave runup ever recorded on Earth. What makes Barry Arm especially concerning is the narrow geometry of the fjord and the large volume of the overhanging mass. These are the conditions that produce violent displacement waves.
The new seismic findings do not confirm that a collapse is imminent. They do not provide a countdown. They provide something different and equally important. They reveal a hidden process that repeats every year beneath the ice. They show that the interior of the mountain is active even when the slope looks motionless. They show that the freeze cycle behind the landslide is powerful enough to break rock tens of thousands of times per year. They show that this cracking begins as the slope enters its active phase. And they show that the cracking continues even after the surface settles.
The picture that emerges is a mountain with a deep internal rhythm. A system of water and rock behind the slope fills, freezes, cracks, and resets. Each cycle sends a message through the sensors placed on the ridge. Each crack is a sign of stress where no one can see. Scientists now have a dangerous landscape that speaks through these signals. The next challenge is to learn which changes in that pattern matter most.
Source:
Davy, G. K., West, M. E., Karasözen, E., and Lyons, J. J.
“Searching for Seismic Precursors: The Barry Landslide Hazard.”
Seismological Research Letters, 2025.
