Studying volcanoes involves getting uncomfortably close to things that could kill you in about six different ways simultaneously. It’s not a field for people who make sensible life choices.
Modern volcanology uses satellites, drones, seismometers, gas sensors, thermal cameras. We’ve basically turned volcano monitoring into a high-tech surveillance operation. And yet volcanoes still surprise us regularly, because predicting geological events is harder than predicting weather, and weather forecasting already has a reputation for being wrong half the time.
The Foundation of Everything We Know Which Is Mostly Just Educated Guessing Based on Previous Eruptions
Seismology is the backbone of volcano monitoring. When magma moves underground, it fractures rock. Those fractures generate earthquakes—hundreds or thousands of tiny ones before eruptions. Seismometers detect these tremors.
The challenge is distinguishing volcanic earthquakes from tectonic ones, or from traffic, or from literally anything else that makes the ground vibrate.
Hawaii Volcano Observatory has been recording seismic data since 1912, which sounds impressive until you realize that’s only 110 years of observations on volcanoes erupting for hundreds of thousands of years.
Before the 2018 Kilauea eruption, seismicity increased dramatically in the Lower East Rift Zone. Thousands of earthquakes over several weeks. Scientists knew something was coming but couldn’t predict exactly what or where until fissures started opening in peoples backyards.
Ground deformation is another major indicator. Magma rising into a chamber pushes the ground upward. GPS stations and satellite radar measure these changes—sometimes just centimeters of uplift spread over months.
Mount St. Helens bulged outward by 150 meters before its 1980 eruption.
The north flank was expanding at 1.5 meters per day in the final weeks. That’s dramatic.
Usually deformation is more subtle—a few millimeters here, slight tilt there, nothing obvious without instruments.
InSAR—Interferometric Synthetic Aperture Radar—uses satellites to detect ground movements with millimeter precision. Scientists compare radar images taken weeks apart and create deformation maps showing which parts of a volcano are rising or sinking.
The Technology That Lets Us Pretend We Understand What Mountains Are Thinking
Volcanic gas monitoring might be the most dangerous job in volcanology. Someone has to hike up to active craters with spectrometers and sample plumes of sulfur dioxide, carbon dioxide, hydrogen sulfide—all the fun stuff that can asphyxiate you if wind direction changes.
Sulfur dioxide is particularly useful because it’s only significant source is magma. When SO2 emissions increase, magma is probably rising. When emissions drop suddenly, it might mean the conduit is blocked, which often precedes explosive eruptions.
Etna has a permanent gas monitoring network. The volcano emits about 5,000 tons of SO2 per day baseline. Before eruptions, that can spike to 20,000 tons.
Drones have revolutionized gas sampling recently. Instead of risking human lives, send a quadcopter into the plume with sensors attached. If it crashes, you’re out $5,000 instead of a graduate student.
Thermal monitoring uses infrared cameras and satellites to detect heat anomalys. Rising magma increases surface temperatures even before anything reaches the surface. MODIS and ASTER satellites scan for thermal hotspots globally.
When Nyiragongo’s lava lake level rises, satellite thermal data shows it days before visual confirmation.
The Problem with Trying to Schedule Appointments with Geological Processes That Operate on Timescales of Centuries
Historical records matter enormously. Mount Pinatubo hadn’t erupted for 500 years before 1991. Scientists had to dig through historical accounts, interview indigenous peoples, study deposits. They found evidence of massive eruptions every few centuries. That knowledge helped them recognize 1991’s precursors for what they were.
Japan has detailed eruption records going back over 1,000 years. Iceland has sagas describing eruptions from the 9th century. Italy’s records extend to ancient Rome. These historical accounts provide context modern instruments cant.
But most volcanoes lack good historical documentation. Indonesian volcanoes have maybe 400 years of records. We’re monitoring them with sophisticated equipment but without historical baselines to interpret the data.
Geological fieldwork remains essential. Someone has to physically map lava flows, sample deposits, date previous eruptions. This involves hiking up steep slopes, dodging rockfalls, camping in remote locations.
Radiometric dating tells us when eruptions occurred. Carbon-14 dating works for recent eruptions. Argon-argon dating works for older ones.
Why We’re Better at Detecting Eruptions Than Preventing Them From Destroying Everything
Modern monitoring has dramatically reduced eruption death tolls. The 1991 Pinatubo eruption could have killed tens of thousands. Because of monitoring and evacuation, fewer than 350 people died—mostly from roof collapses under ash weight.
But monitoring infrastructure is expensive. Indonesia has 130 active volcanoes and maybe 70 monitoring stations.
That’s insufficient coverage.
Many dangerous volcanoes have minimal monitoring—one seismometer if they’re lucky.
The USGS monitors 169 volcanoes in the U.S. Only 27 have real-time seismic monitoring. The rest get periodic check-ins. If one of the unmonitored ones decides to wake up, we might get weeks of warning.
Or we might get hours.
Yellowstone probably has the most sophisticated monitoring network of any volcano globally. Hundreds of seismometers, dozens of GPS stations, multiple gas sensors, thermal monitoring. And still scientists cant predict if or when it will erupt.
They can tell us magma chamber depth, size, temperature—but not timing.
We study volcanoes because we live near them and would prefer not to die horribly. The technology has improved dramatically in 50 years. Satellites, computers, drones, better instruments—we have tools previous generations couldn’t imagine.
But volcanoes remain fundamentally unpredictable in the ways that matter most.
We know more than ever before and still get surprised regularly. That’s geology.
