Scientists predict volcanic eruptions by tracking a combination of signals: earthquake swarms, ground deformation, gas emissions, and temperature changes. No single measurement is reliable on its own, but when several indicators shift simultaneously, volcanologists can often forecast an eruption days to weeks in advance. In some cases, the warning window stretches to months.
Earthquake Swarms: The Earliest Warning Sign
Rising magma cracks and fractures the rock above it, generating clusters of small earthquakes that seismometers can detect long before any visible changes appear at the surface. These “volcano-tectonic” earthquakes represent brittle rock failure, the same physical process that occurs along tectonic faults. At volcanoes, they signal that magma or volcanic fluids are forcing their way upward through new or existing fractures.
A second type of seismic signal, called long-period or low-frequency earthquakes, occurs when cracks resonate as magma and gases flow through them. These are often seen before eruptions, though they also show up as part of normal background activity at some volcanoes, so their presence alone doesn’t guarantee an eruption is coming. The combination of both types, especially when accompanied by harmonic tremor (a continuous, rhythmic ground shaking distinct from sharp earthquake jolts), is what raises the alarm.
Mount St. Helens in 1980 is the textbook example. A magnitude 4.2 earthquake on March 20 was the first real sign of awakening after 123 years of dormancy. Within five days, earthquake activity surged dramatically, with 174 shocks above magnitude 2.6 recorded per day at peak levels. By March 31, seismographs were also picking up harmonic tremor. That combination of strong earthquake activity and continuous tremor told scientists that magma was actively moving inside the volcano.
Ground Deformation: Watching the Volcano Change Shape
When magma accumulates inside a volcano, it physically pushes the ground outward. Scientists measure this swelling using GPS stations, satellite radar, and laser-based instruments. Even small changes in the shape of a volcano’s surface can reveal how much magma is moving and where it’s headed.
At Mount St. Helens, a bulge developed on the north face that grew outward at roughly five feet per day, measured with precise laser instruments starting in late April 1980. That rate of deformation was extraordinary, and it made clear that a massive volume of magma was intruding into the volcano’s interior. For underwater volcanoes like Axial Seamount in the Pacific, deformation tracking on the seafloor has made eruptions relatively predictable by revealing the same kind of inflation cycle that occurs on land.
Gas Emissions Signal Magma Rising
Magma carries dissolved gases, primarily carbon dioxide and sulfur dioxide, that escape as it rises toward the surface and pressure drops. Changes in the ratio between these two gases can signal that fresh, gas-rich magma is ascending from deeper in the Earth.
Carbon dioxide escapes from magma at greater depths than sulfur dioxide does. So a spike in the carbon dioxide-to-sulfur dioxide ratio often indicates that new magma is arriving from below, while a shift toward more sulfur dioxide suggests magma has reached shallower levels where sulfur dioxide can escape. At Stromboli volcano in Italy, substantial shifts in this gas ratio have appeared weeks before major explosive events. Scientists track these changes continuously using ground-based sensors and, in some cases, drones that fly directly through volcanic plumes.
Satellite Thermal Monitoring
Satellites equipped with infrared sensors can detect heat changes at volcanic surfaces that would be invisible to the naked eye. Modern algorithms can identify temperature increases as small as 0.5 degrees above the surrounding background, with a false alarm rate of roughly 1.8%. These satellites pass over volcanic regions multiple times per day at a spatial resolution of 375 meters, meaning they can catch subtle warming trends at remote or hard-to-access volcanoes long before anyone on the ground notices a change.
This kind of monitoring is especially valuable for the hundreds of volcanoes worldwide that don’t have ground-based instruments. A thermal anomaly picked up by satellite can trigger closer investigation with seismic and gas monitoring equipment.
What Crystals Inside Magma Reveal
One of the more remarkable forecasting tools comes from studying the minerals that grow inside magma as it moves. Crystals in volcanic rock form distinct layers, much like tree rings, and each layer records the chemical conditions at the time it grew. By analyzing sharp chemical boundaries in these crystals, scientists can reconstruct how long ago magma began its ascent toward the surface.
During the 2016-2017 eruption of Bogoslof volcano in Alaska, researchers found that sharp chemical boundaries in crystals correlated with known increases in seismicity and sulfur dioxide emissions. Diffusion modeling showed that these boundaries formed no more than 180 days before the final explosive event. This kind of analysis works after the fact rather than in real time, but it helps scientists understand the typical timelines of magma movement for a given volcano, which improves future forecasting.
How Machine Learning Is Changing Detection
Volcano monitoring generates enormous volumes of data, far more than human analysts can review in real time. Machine learning models are now being trained to recognize the seismic “fingerprint” of pre-eruptive activity. At Piton de la Fournaise volcano on Réunion Island, researchers built a model that could determine the eruptive state of the volcano from a single time window of raw seismic data recorded at a single station. The model isolated consistent frequency patterns present across the majority of eruptions in the dataset.
This approach could eventually be used to detect subtle precursory signals that human analysts miss, potentially extending warning times. Spectral clustering techniques can also distinguish different phases within eruptions and differences between eruptions at the same volcano, helping to build a more nuanced picture of what “normal” versus “dangerous” activity looks like.
Monitoring Underwater Volcanoes
Submarine volcanoes pose a unique challenge because their signals have to travel through water rather than air. Seismic waves from underwater eruptions convert into hydroacoustic waves that can be detected by ocean-bottom sensors or even distant hydrophone arrays. These eruptive signals typically appear as repetitive broadband pulses of energy that increase in amplitude before abruptly stopping. Individual pulses can last from tens of seconds to tens of minutes, and the activity can continue for months to years.
Axial Seamount, located off the coast of Oregon, is one of the best-monitored submarine volcanoes in the world, with real-time instruments on the seafloor. Its eruptions have proven relatively predictable based on precursory seismicity and the same inflation-deflation deformation cycle seen at land volcanoes.
How Much Warning Time Do You Actually Get?
Precursors to volcanic eruptions typically appear days to weeks before the event. For larger eruptions involving significant magma buildup, the warning window can stretch to months or even years. The USGS notes that the buildup to a catastrophic eruption at a well-monitored volcano would likely be detectable for weeks to months beforehand.
That said, the reliability of predictions depends heavily on how well a volcano is monitored and how well scientists understand its individual behavior. Volcanoes with decades of seismic records and known eruption histories are far easier to forecast than those with little monitoring infrastructure. Some volcanoes also produce false alarms: all the signs of an impending eruption appear, but the magma stalls underground and never reaches the surface. The science has improved dramatically over the past 25 years, but predicting exactly when and how large an eruption will be remains one of the hardest problems in earth science. What scientists can do reliably is identify when a volcano has shifted from quiet to potentially dangerous, giving communities time to prepare or evacuate.